METHOD AND REAGENT FOR INHIBITING VIRAL REPLICATION

Background of the Invention This invention relates to reagents useful as inhibitors of viral replication, infection and gene expression. Summary of the Invention

The invention features novel enzymatic RNA molecules, or ribozymes, and methods for their use for inhibiting viral replication, e.g.. the replication of picornaviruses (poliovirus, coxsackievirus, hepatitis A virus, echovirus, human rhinovirus, cardioviruses (e.g. , mengovirus and encephalomyocarditis virus) and foot-and- mouth disease virus) ; immunodeficiency virus (e.g.. HIV-1, HIV-2 and related viruses including FIV-1 and SIV-l) ; hepatitis B virus (HBV) ; papillomavirus; Epstein-Barr virus (EBV) ; T-cell leukemia virus, e.g.. HTLV-I, HTLV-II and related viruses, including bovine leukemia virus (BLV) and simian T-cell leukemia virus (STLV-I) ; hepatitis C virus (HCV) ; cytomegalovirus (CMV) ; influenza virus; herpes simplex virus (HSV) . Such ribozymes can be used in a method for treatment of diseases caused by these viruses in man and other animals, including other primates. Indeed one ribozyme may be designed for treatment of many of the diseases caused by these viruses.

Ribozymes are RNA molecules having an enzymatic activity which is able to repeatedly cleave other separate RNA molecules in a nucleotide base sequence specific manner. Such enzymatic RNA molecules can be targeted to virtually any RNA transcript, and efficient cleavage has been achieved in vitro . Kim et al., 84 Proc. Natl. Acad. Sci. USA 8788, 1987; Haseloff and Gerlach, 334 Nature 585, 1988; Cech, 260 JAMA 3030, 1988; and Jefferies et al., 17 Nucleic Acids Research 1371, 1989.

Ribozymes act by first binding to a target RNA. Such binding occurs through the target RNA binding portion of a ribozyme which is held in close proximity to an enzymatic portion of the RNA which acts to cleave the target RNA. Thus, the ribozyme first recognizes and then binds a target RNA through complementary base-pairing, and once bound to the correct site, acts enzymatically to cut the target RNA. Strategic cleavage of such a target RNA will destroy its ability to direct synthesis of an encoded protein. After a ribozyme has bound and cleaved its RNA target it is released from that RNA to search for another target and can repeatedly bind and cleave new targets.

The enzymatic nature of a ribozyme is advantageous over other technologies, such as antisense technology (where a nucleic acid molecule simply binds to a nucleic acid target to block its translation) since the effective concentration of ribozyme necessary to effect a therapeutic treatment is lower than that of an antisense oligonucleotide. This advantage reflects the ability of the ribozyme to act enzymatically. Thus, a single ribozyme molecule is able to cleave many molecules of target RNA. In addition, the ribozyme is a highly specific inhibitor, with the specificity of inhibition depending not only on the base pairing mechanism of binding, but also on the mechanism by which the molecule inhibits the expression of the RNA to which it binds. That is, the inhibition is caused by cleavage of the RNA target and so specificity is defined as the ratio of the rate of cleavage of the targeted RNA over the rate of cleavage of non-targeted RNA. This cleavage mechanism is dependent upon factors additional to those involved in base pairing. Thus, it is thought that the specificity of action of a ribozyme is greater than that of antisense oligonucleotide binding the same RNA site. Ribozymes targeting any of the specific viral regions described in detail below should be able to cleave

the RNAs in a manner which will inhibit the translation of the molecules.

Thus, in a first aspect, the invention features an enzymatic RNA molecule (or ribozyme) which specifically cleaves the mRNA, hnRNA, or gene of any one of the following viruses: picornavirus, HIV-1, HBV, papillomavirus, EBV, T-cell leukemia virus, e.g.. HTLV-1, HCV, CMV, influenza virus, and HSV.

By "gene" is meant to refer to either the protein coding regions of the cognate mRNA, or any regulatory regions in the RNA which regulate synthesis of the protein or stability of the mRNA; the term also refers to those regions of an mRNA which encode the ORF of a cognate polypeptide product, and the proviral genome. By "enzymatic RNA molecule" it is meant an RNA molecule which has complementarity in a substrate binding region to a specified gene target, and also has an enzymatic activity which is active to specifically cleave RNA in that target. That is, the enzymatic RNA molecule is able to intermolecularly cleave RNA and thereby inactivate a target RNA molecule. This complementarity functions to allow sufficient hybridization of the enzymatic RNA molecule to the target RNA to allow the cleavage to occur. One hundred percent complementarity is preferred, but complementarity as low as 50-75% may also be useful in this invention. By "equivalent" RNA to a virus is meant to include those naturally occurring viral encoded RNA molecules associated with viral caused diseases in various animals, including humans, cats, simians, and other primates. These viral or viral-encoded

RNAs have similar structures and equivalent genes to each other.

In preferred embodiments, the enzymatic RNA molecule is formed in a hammerhead motif, but may also be formed in the motif of a hairpin, hepatitis delta virus, group I intron or RNaseP-like RNA (in association with an

RNA guide sequence) . Examples of such hammerhead motifs

are described by Rossi et al., 8 Aids Research and Human Retroviruses 183, 1992; of hairpin motifs by Ha pel and Tritz, 28 Biochemistry 4929, 1989 and. Hampel et al., 18 Nucleic Acids Research 299, 1990; an example of the hepatitis delta virus motif is described-- by Perrotta and Been, 31 Biochemistry 16, 1992; of the RNaseP motif by Guerrier-Takada et al., 35 Cell 849, 1983; and of the group I intron by Cech et al., U.S. Patent 4,987,071. These specific motifs are not limiting in the invention and those skilled in the art will recognize that all that is important in an enzymatic RNA molecule of this invention is that it has a specific substrate binding site which is complementary to one or more of the target gene RNA regions, and that it have nucleotide sequences within or surrounding that substrate binding site which impart an RNA cleaving activity to the molecule.

In particularly preferred embodiments, the RNA which is cleaved in viral RNA is selected from one or more of the sequences shown in Figs. 6 to 15. In a second related aspect, the invention features a mammalian or vertebrate cell which includes an enzymatic RNA molecule as described above. Preferably, the cell is a human cell, for example, a T4 lymphocyte having a CD4 receptor molecule on its cell surface, or other primate cell.

In a third related aspect, the invention features an expression vector which includes nucleic acid encoding the enzymatic RNA molecules described above, located in the vector, e.g. , in a manner which allows expression of that enzymatic RNA molecule within a mammalian or vertebrate cell. Such a vector may include the ribozyme under control of a virus promoter, e.g.. an EBV promoter.

In a fourth related aspect, the invention features a method for treatment of a virus-caused disease by administering to a patient an enzymatic RNA molecule which cleaves virus or virus encoded nucleic acid.

In other related aspects, the invention features treatment of cats or simians with ribozymes of this invention. Such ribozymes may be identical to those able to cleave HIV-1 RNA, or may be modified to target analogous locations in FIV and SIV virus RNAs.

In other related aspects, the invention features treatment of cattle or simians with ribozymes of this invention. Such ribozymes may be identical to those able to cleave HTLV-I or II RNA, or may be modified to target analogous locations in BLV and STLV virus RNAs.

The invention provides a class of chemical cleaving agents which exhibit a high degree of specificity for the viral related RNA in virus-infected cells or virion particles. If desired, such ribozymes can be designed to target equivalent single-stranded DNAs by methods known in the art. The ribozyme molecule is preferably targeted to a highly conserved sequence region of a virus such that all types and strains of these viruses can be treated with a single ribozyme. Such enzymatic RNA molecules can be delivered exogenously or endogenously to infected cells. In the preferred hammerhead motif the small size (less than 40 nucleotides, preferably between 32 and 36 nucleotides in length) of the molecule allows the cost of treatment to be reduced compared to other ribozyme motifs.

The smallest ribozyme delivered for treatment of HIV infection reported to date (by Rossi et al., 1992, supra) is an in vitro transcript having a length of 142 nucleotides. Synthesis of ribozymes greater than 100 nucleotides in length is very difficult using automated methods, and the therapeutic cost of such molecules is prohibitive. Delivery of ribozymes by expression vectors is primarily feasible using only ex vivo treatments. This limits the utility of this approach. In this invention, small ribozyme motifs (e.g. , of the hammerhead structure, shown generally in Fig. 1) are used for exogenous delivery. The simple structure of these molecules also

increases the ability of the ribozyme to invade targeted regions of the mRNA structure. Thus, unlike the situation when the hammerhead structure is included within longer transcripts, there are no non-ribozyme flanking sequences to interfere with correct folding of the ribozyme structure or with its complementary binding of the ribozyme to the mRNA target region.

The enzymatic RNA molecules of this invention can be used to treat virus infections including those caused by both HIV-1 and HIV-2, HTLV-I and HTLV-II. Such treatment can also be extended to other related viruses which infect non-human primates including the simian and feline immunodeficiency viruses and bovine leukemia viruses. Infected animals can be treated at the time of productive or latent infection. This timing of treatment will reduce viral loads and gene expression in infected cells and disable viral replication in any subsequent rounds of productive infection. This is possible because the preferred ribozymes disable those structures required for successful initiation of viral protein synthesis.

For treatment of transformed cervical epithelia or keratinocytes, the method of this invention will inhibit the expression of viral genes known to cause cell immortalization. For treatment of latent viral infections, this invention will inhibit gene expression required for the maintenance of the viral episomal genome. For treatment of transformed hepatocytes, the methods of this invention allow inhibition of the expression of viral genes thought to cause cell transformation. The preferred targets of the present invention are advantageous over other targets since they act not only at the time of viral absorption or genomic replication (reverse transcription) during the productive infection, but also in latently infected cells and in virally transformed cells. In addition, viral particles which are released during a first round of infection in the presence of such ribozymes will still be immunogenic

by virtue of having their capsids or virions intact. Thus, one method of this invention allows the creation of defective but immunogenic viral particles, and thus a continued possibility of initiation of an immune response in a treated animal.

In addition, the enzymatic RNA molecules of this invention can be used in vitro in a cell culture infected or transfected with viruses to produce defective viral particles, which may have intact capsids and defective genomic RNA or exhibit structurally intact but biologically deficient virions, i.e. , the virion may be stable but deficient in a protein required for adsorption or penetration. These particles can then be used for instigation of immune responses in a prophylactic manner, or as a treatment of infected animals.

In yet another aspect, the invention features use of viruses as vectors for carrying an enzymatic RNA molecule to a cell infected with another virus. Such vectors can be formed by standard methodology. The invention also features immunization preparations formed from defective viruses, e.g.. HIV-1, EBV, or HTLV-I particles (or related particles) , created by a method of this invention, and methods for immunization or vaccination using these defective particles, e.g.. with DNA or vectors encoding a ribozyme of this invention under the control of a suitable promoter.

Ribozymes of this invention may be used as diagnostic tools to examine genetic drift and mutations within diseased cells. The close relationship between ribozyme activity and the structure of the target RNA allows the detection of mutations in any region of the molecule which alters the base-pairing and three- dimensional structure of the target RNA. By using multiple ribozymes described in this invention, one may map nucleotide changes which are important to RNA structure and function in vitro , as well as in cells and

tissues. Cleavage of target RNAs with ribozymes may be used to inhibit gene expression and define the role (essentially) of specified gene products in the progression of disease. In this manner, other genetic targets may be defined as important mediators of the disease. These experiments will lead to better treatment of the disease progression by affording the possibility of combinational therapies (e.g. , multiple ribozymes targeted to different genes, ribozymes coupled with known small molecule inhibitors, or intermittent treatment with combinations of ribozymes and/or other chemical or biological molecules) .

Other features and advantages of the invention will be apparent from the following description of the preferred embodiments thereof, and from the claims. Description of the Preferred Embodiments

The drawings will first briefly be described. Drawings

Fig. 1 is a diagrammatic representation of a hammerhead motif ribozyme showing stems I, II and III

(marked (I) , (II) and (III) respectively) interacting with a target region. The 5' and 3' ends of both ribozyme and target are shown. Dashes indicate base-paired nucleotides. Figs. 2A and 2B are diagrammatic representations of the various genes and gene regions in HIV-1 and HIV-2.

Figs. 3A - 3C are diagrammatic representations of three ribozymes of this invention.

Figs. 4A - 4G are diagrammatic representations of chemically modified ribozymes of this invention (solid circles indicate modified bases) .

Fig. 5 is a diagrammatic representation of a chemically modified ribozyme of this invention (circled bases are modified with 2'-0-methyl) . Figs. 6 - 15 are preferred targets for picornavirus, HIV, HBV, papilloma, EBV, HTLV, HCV, CMV, influenza, and HSV, respectively.

Figs. 16, 17 and 18 are diagrammatic representations of various vectors of the present invention.

Fig. 19 is a diagrammatic representation of a method for formation of a vector of this invention.

Figs. 20, 21 and 22 are diagrammatic representations of hammerhead, hairpin and HDV ribozymes, respectively.

Figs. 23 and 24 are copies of HPLC results showing purification of ribozyme from failure sequences (Fig. 23) and other contaminants (Fig. 24) .

Fig. 25 is a diagrammatic representation of a hammerhead ribozyme in equilibrium between a tightly constrained structure and an undesirable loose structure. Fig. 26 is a general scheme for generation of useful substrate sets.

Fig. 27 shows an example of a standard kinetic analysis of ribozyme activity with enzyme concentration in excess over substrate concentration. 5'-[ ³² P]-labeled substrate (typically at a concentration of 1-2 nM) is incubated with ribozyme at varying concentrations (typically 5-100 nM) at 37°C in a buffer containing 75 mM Tris-HCI, pH 8.0, 10 mM MgCl ₂ . At various times, aliquots are removed and mixed into a standard formamide/EDTA/dyes gel-loading buffer to stop the reaction and prepare the samples for gel-electrophoresis. The products are separated by denaturing polyacrylamide gel electrophoresis and quantitated by a beta emission scanner (Ambis Systems, San Diego, CA) . Specifically, Fig. 27A shows a plot of the log of the fraction of substrate molecules remaining uncleaved as a function of time. In enzyme (e) excess conditions, substrate cleavage follows the equation:

S _t = s^- ¹ "* ⁸ ' where S, is the concentration of substrate at time, t, S ₀ is the starting concentration of substrate, and kobs is the observed cleavage rate. Thus:

log, (S, / s _o ) = log, (F) = -kobs t where F = S _t / S ₀ is the fraction of substrate at time t. So the slope of the lines in Fig. 27A gives kobs for each ribozyme concentration shown at the right of the graph. Fig. 27B shows a plot of kobs vs. enzyme concentration for five different ribozymes. Under enzyme excess conditions (and low enzyme concentration) the curves follow the equation: kobs = (kcat / Km) [Enzyme] so the slope of each line gives the (kcat / Km) value for that enzyme. Lines with steeper slopes have higher kcat/km values and indicate more active ribozymes. In this figure, ribozyme #4 is the most active and ribozyme #2 is the least active. Fig. 28 is a general scheme for a method of this invention.

Fig. 29A is a diagrammatic representation of a method of the invention.

Fig. 29B is a graph showing expected coupling efficiencies vs. time.

Figs. 30A, B, and C are diagrammatic representations of several vectors of this invention. The three vectors shown as general examples in Fig. 30 all contain a T7 promoter upstream from a region encoding a ribozyme. The vectors may also include a sequence known to encode a T7 terminator located downstream of the ribozyme-encoding sequence. The simplest of the vectors, shown in Fig. 30A, includes a T7 promoter, a ribozyme encoding sequence and a T7 terminator. As shown in Fig. 30B, additional sequence encoding hairpin structures may also be included on both the 5' and 3' sides of the ribozyme-encoding sequence. In Fig. 30C, a poly(A) polymerase processing site is included downstream of the ribozyme-encoding sequence and may or may not be followed by a T7 terminator sequence. The T7 promoter is not limiting to this invention. If other promoters are employed (e.g. , HSV ICP4-responsive promoters, or the

thymidine kinase gene promoter) then the T7 termination sequence may not be effective in terminating transcription. In that case, the poly(A) processing site would be valuable in ensuring proper termination of the transcript.

Figs. 31A-E show examples of ribozyme sequences with 5' and 3' flanking hairpin sequences. Specifically, Fig. 31A shows ribozyme alone, showing that the ribozyme forms only stem II and no other secondary structures in the absence of substrate. The ribozyme is targeted against a site (E) in the HSV ICP4 gene having the sequence UGGCGUCGGAGAA. The six nucleotides at each end of the ribozyme are complementary to the 6 nucleotides on each side of the central C in this sequence. Fig. 3IB shows the HSV ICP4 site E ribozyme of

Fig. 30 bound to a small RNA substrate containing the target sequence. All three stems (I, II, and III) of the hammerhead ribozyme are now formed.

Fig. 31C shows a folded structure for one possible sequence for a 5' flanking hairpin. This sequence is compatible with transcription from a T7 promoter as it begins with the preferred sequence GGGAG. Other possible sequences would include, but are not limited to, GGGCGAAAGCCC, GGAGGAAACUCC, and GGGACGAAAGUCCC. For efficient T7 transcription the occurrence of U nucleotides should be avoided in the first 9 nucleotides as U promotes undesirable abortive transcription.

Fig. 3ID shows a folded structure of the Tφ termination sequence from phage T7 (Dunn and Studier, 166 J. Mol. Biol. 477, 1983) as one possible 3'-flanking hairpin. Sequences 5' to the hairpin might be dispensable, in order to reduce excess unstructured RNA sequences, without reducing termination efficiency of the RNA polymerase.

Fig. 3IE shows the structure of the ribozyme shown in Fig. 31A coupled with the 5' and 3' hairpins shown in Figs. 31C and 3ID.

Figs. 32A-F are diagrammatic representations of a hammerhead having various hairpin sequence designs, with a summary of the cleavage activities obtained from said ribozymes.

Fig. 33 is a diagrammatic representation showing a general method of this invention. Specifically, a ribozyme population is formed including analogues at various sites. This population is 5' end-labeled and incubated under cleavage conditions. The active and inactive ribozymes are gel purified by size, treated with an agent which allows determination of the site of the analog (e.g.. H ₂ 0 ₂ ) , and then determination of those sites. Fig. 34 is a diagrammatic representation of the ribozyme-coding strand of a DNA fragment prepared for insertion into an expression vector.

Figs. 35A and 35B are a schematic representation of an RNaseH accessibility assay. Specifically, Fig. 35A is a diagram of complementary DNA oligonucleotides bound to accessible sites on the target RNA. Complementary DNA oligonucleotides are represented by broad lines labeled A, B, and C. Target RNA is represented by the thin, twisted line. Fig. 35B is a schematic of a gel separation of uncut target RNA from a cleaved target RNA. Detection of target RNA is either by autoradiography of body-labeled target RNA, by northern blotting with labeled target cDNA or by RNase protection with labeled RNA. The bands in each lane represent uncut target RNA. The remaining bands in each lane represent the cleaved products.

Fig. 36 is a copy of an autoradiogram of an RNaseH assay performed on in vitro transcribed target RNA using five probes to a portion of the HSV ICP27 mRNA. Probes labeled J, K, L, M, and N respectively represent five 13-mer DNA oligonucleotides complementary to five different sites on ICP27 (see Table 1) . The end lanes

labeled M are DNA size markers (DNA sizes are indicated at the left of the figure) . Lanes labeled 0 and H represent target RNA immediately after transcription (0) or after incubation with RNaseH but without a DNA oligonucleotide (H) . Lanes marked 1000, 100 and 10 represent digestion by RNaseH in the presence of 1000 nM, 100 nM and 10 nM respectively of DNA probe. Lanes marked S-RAN: UUC, GUC, GUA and AUC represent digestion of target RNA by RNaseH in the presence of 10 μM of the semi-random DNA oligonucleotides containing 12-nucleotides total with the middle three comprising sequences complementary to UUC, GUC, GUA and AUC, respectively.

Table 1. Sequence and position of ICP27 oligos used in RNaseH assays.

Position: The nucleotide position of the target sequence in ICP27. The transcription start is position 1. The actual positions of RNaseH cleavage promoted by the semirandom probes has not been determined.

Sequence: The sequence of DNA oligonucleotides used as RNaseH probes to accessibility at the given site. The probe sequenc4es are complementary to the target sequence. N represents a population of molecules with all four nucleotides (A, C, G, & U) represented.

Fig. 37 is a computer generated folding pattern of the sequence of HSV ICP27 from nucleotides 952-1376. Three of the sites that were targeted for cleavage in Fig.

36 are indicated by broad lines and labels (sites J, K and L) .

Fig. 38 is a copy of an autoradiogram of an RNaseH assay performed on HSV ICP27 mRNA contained in a cellular extract, using four DNA oligonucleotide probes. Target RNA was produced by infection of Vero cells with Herpes Simplex Virus (HSV) . Cell extracts were separated into nuclear, cytoplasmic and membrane fractions. ICP27 RNA in the cytoplasmic fraction was probed for accessibility to RNaseH cleavage in the presence of 15 μM DNA oligonucleotides complementary to site D, E, El or U. Cleaved and full-length target RNA was detected by RNase protection assay using a 400 nucleotide antisense RNA. Full-length message was detected as a 360 nucleotide band, while cleaved message was detected as bands in the 200-330 nucleotide size range. Each site was probed twice without added RNaseH (labeled "-") and twice in the presence of RNaseH (labeled "+") . Cleavage products in the "-" lane for site D confirms the presence of endogenous RNaseH in the extracts.

Fig. 39 is a graph showing quantitation of RNaseH cleavage rates at different sites on HSV ICP27. Sites K and T show efficient levels of cleavage at this DNA probe concentration; sites D and L show only low levels of cleavage.

Fig. 40 is a graph illustrating the effect of DNA probe length on RNaseH cleavage activity at an accessible site (K) in ICP27 target RNA. A plot of cleavage activity (% target RNA cleaved) as a function of DNA probe concentration for DNA probes of length 7, 9, 11 and 13 nucleotides is shown. The 7 nucleotide probe shows no activity even at 1 μm concentrations, suggesting that probe lengths greater than 7 nucleotides are preferred.

Fig. 41 is a diagrammatic representation of a method for producing target nucleic acid.

Fig. 42 is a schematic representation of a method of this invention; the dashed line represents a

splint strand to aid joining and ligation of the two RNA molecules (solid lines) .

Figs. 43 and 44 are diagrammatic representations of ribozymes formed by a method of this invention showing the 5' and 3' fragments and the substrate.

Fig. 45 is a schematic representation of an alternative method of this invention in which no splint strand is required for ligation but rather the two fragments hybridize together at stem II prior to ligation. Fig. 46 is a diagrammatic representation of the preferred half ribozyme structure prior to ligation.

Figs. 47A-D are diagrammatic representations of standard base modifications for adenine, guanine, cytosine and uracil, respectively; modifications on the upper line of each figure represent structures with stronger base pairing ability compared to those on the lower lines.

Fig. 48 is a diagram of the interaction between the releasing ribozyme (RR) and the therapeutic ribozyme (TR) cassette. (A) is an unpaired nucleotide between the TR concatamers which is required for effective RR cleaving activity. (TR)-l and (TR)-2 are the first and second TR monomers released by RR cleavage.

Fig. 49 is a diagrammatic representation of hammerhead ribozyme:substrate complex. N = G, A or C. M = U, A, C. Vertical lines represent intramolecular paired nucleotides between ribozyme and substrate molecules.

Fig. 50 is a flow diagram for cloning of quasi- random ribozyme cassettes. Target Sites Ribozymes targeting selected regions of a virus genome are preferably chosen to cleave the target nucleic acid in a manner which inhibits expression or translation of the nucleic acid. Genes are selected such that such inhibition of translation will occur or viral replication is inhibited, e.g.. by inhibiting protein synthesis or genomic replication and packaging into virions. Such inhibition of translation will generally inhibit viral

replication, e.g. , by inhibiting protein synthesis. Selection of effective ribozymes to cleave target sites within these critical regions of viral nucleic acid entails testing the accessibility of the target nucleic acid to hybridization with various oligonucleotide probes. These studies can be performed using nucleic acid probes and assaying accessibility by cleaving the hybrid molecule with RNaseH (see below) . Alternatively, such a study can use ribozyme probes designed from secondary structure predictions of the nucleic acids, and assaying cleavage products by polyacrylamide gel electrophoresis (PAGE) , to detect the presence of cleaved and uncleaved molecules.

The following are examples of methods by which suitable target sites can be identified and are not limiting in this invention. Generally, the methods involve identifying potential cleavage sites for a hammerhead ribozyme, and then testing each of these sites to determine their suitability as targets by ensuring that secondary structure formation is minimal. Example 1: Picornavirus

Picornaviruses include those referred to as poliovirus, coxsackievirus, hepatitis A virus, human rhinovirus, ECHO virus, the cardioviruses, ^{^} (e.g. , mengovirus, encephalomyocarditis virus) and foot-and-mouth disease virus. These viruses cause many harmful diseases in humans and other animals. See for example, Pelletier and Sonenberg, 334 Nature 320, 1988.

The picornaviruses comprise one of the largest and most important families of human and agricultural pathogens. The family is currently divided into 4 genera: the enteroviruses, the cardioviruses, the rhinoviruses and the aphthoviruses (foot-and-mouth disease viruses) . Sequence analysis has shown that these classifications may be too superficial. The rhinoviruses share many sequence similarities to the enteroviruses, while enterovirus 72 (hepatitis A virus) is markedly different and should be included in its own subgenus. Finally, the murine

polioviruses (Thieler's viruses) which are classified as enteroviruses are more closely related to the cardioviruses.

The enteroviruses, so-called because they infect the alimentary tract, include three serotypes of poliovirus, 23 serotypes of coxsackievirus, 32 serotypes of echoviruses, hepatitis A virus and 37 serotypes of non- human enteric viruses. The human rhinoviruses are important etiologic agents of the cold and other nasopharyngeal infections. There are over 100 serotypes of rhinovirus. Aphthoviruses infect cloven-footed animals, especially cattle, goats, pigs, sheep, and rarely, even humans. Seven immune types of aphthoviruses have been isolated. The cardioviruses (Columbia SK virus, encephalomyocarditis [EMC] virus, ME virus, MM virus and mengovirus) all belong to a single serotype and are considered to be strains of EMC virus. These viruses are primarily murine, but will infect humans, pigs, elephants and squirrels. Some viruses such as the equine rhinovirus 1 and 2 and a collection of insect picornaviruses have not been assigned to genera. One of these, cricket paralysis virus cross-reacts serologically with EMC virus.

The picornaviral genome consists of a single strand of mRNA which can be extracted from the virions. Isolated RNA has a specific infectivity about one- millionth that of the intact virion. Two reasons for this drop in infectivity are: (i) the loss of the vial capsid eliminates receptor mediated uptake of the RNA and (ii) a single break in the RNA, whether free or inside the virus particle, is sufficient to destroy infectivity. Viral RNA isolated from the capsid is particularly susceptible to nicks and breaks.

The small protein VPg (virion protein, genome) is attached to the 5' terminal -pUpUp of the genomic RNA through a phosphodiester linkage to the phenolic hydroxyl group of a tyrosine residue. The length of VPg varies only slightly in different species of picornaviruses. VPg

plays a important role in initiation of picornaviral RNA synthesis and possibly also in the packaging of the RNA into virions.

The 5' non-translated region (NTR) of the RNA is unusually long compared to the homologous region of cellular mRNAs; it ranges from 624 bases in HRV14 to 1200 in aphthovirus. Computer analyses predict the presence of stable stems and loops in the 5'-NTR and the folding patterns correlate with each virus group. For example, the first 600 bases of the rhinovirus 5' NTR align well with those of the enteroviruses but lack an additional 140 base loop which is found in the enteroviruses. The role of these structures in the control of internal initiation of translation has been shown for poliovirus and others. Cardioviruses, aphthoviruses and Thieler's viruses each display distinctive folding patterns which probably play an important role in the translational control and/or tissue-trophic expression of those viruses. A single C to U base change in nucleotide 472 of the Sabin strain of poliovirus 3 results in increased neurovirulence.

The 3' NTR is relatively short, ranging from 47 bases in HRV14 to 126 for EMC virus. The function is unknown but an 8 base insertion in this region of poliovirus RNA produces a temperature sensitive phenotype. The poly(A) sequence at the 3' end of the RNA is heterogeneous in length and the mean length is genetically determined. The range for poly(A) length is from 36 bases in EMC virus to 100 in aphthovirus. The RNA contains a single long ORF encoding a long polypeptide chain, called the polyprotein which is cleaved during translation. A cascade of cleavages, performed by three viral proteases, ultimately generates 11 viral protein products, 12 in the case of viruses encoding a leader (L) protein. Protease 3C, or its precursor 3CD, mediates most of the cleavages. Protein 3D is an enzyme capable of elongating nascent RNA chains

from an RNA template. Protein 2C is also involved in RNA synthesis, perhaps in the initiation process. The 2C sequence is the most strongly conserved sequence in the polyprotein. Mutation analysis has shown that protein 2B and 2C are involved in RNA synthesis while protein 2A is involved in the shut-off of host protein synthesis.

Recombination takes place in RNA viruses, including the picornaviruses. It has been proposed that recombination occurs when viral RNA polymerase detaches from its original template and finishes synthesis of the nascent chain from a second template. Studies supporting this model suggest that strand switching occurs most frequently during synthesis of the minus strands. The error frequency of RNA synthesis is one substitution per genome while the rate of recombination may occur in as much as 10-20% of all viral RNA molecules. Thus, the recombination rate may partially correct for an unusually high error frequency during RNA synthesis.

In humans, poliovirus infects cells of the nasopharynx and gut although the paralytic form of poliovirus also has a high specificity for the anterior horn cells of the spinal cord. Coxsackieviruses also infect nasopharyngeal cells have a propensity for skeletal and heart muscle cells. The mechanisms underlying cell tropisms are not understood. - The most plausible hypothesis for explaining cell tropism is that cells carry specific receptor molecules on their surfaces. For example, the ability of coxsackie A to produce myositis only in baby mice correlates with specific receptors present in differentiating myoblasts but not in other mouse tissues. The receptors for poliovirus, coxsackie B virus, echovirus and the major sero group of human rhinoviruses all map to chromosome 19 and those of poliovirus and rhinovirus have been identified as members of the immunoglobulin super-family of molecules. The major poliovirus receptor is a protein of 45 kd and the

rhinovirus receptor has been identified as ICAM-1 (intercellular adhesion molecule-1) .

Enteroviruses are associated with clinical manifestations which range from a mild febrile illness to permanent, sometimes fatal, paralysis. Among the infections associated with enteroviruses are: poliomyelitis, abortive poliomyelitis, meningitis, aseptic meningitis, paralytic poliomyelitis, epidemic myalgia, herpangina, hand-foot-and-mouth disease, various respiratory illnesses, eye disease, acute myocardiopathy, chronic cardiovascular disease, neonatal disease and fetal defects, gastrointestinal disease, hepatitis, post-viral fatigue syndrome and diabetes.

Treatment of picornaviral disease has concentrated on prophylactic vaccination, primarily with poliovirus. Vaccines are not available for the other picornaviruses due to an inability to target all of the multiple serotypes known to exist. Interferon treatment has been used with rhinovirus infections, but some people experienced more morbidity from the interferon treatment than from the rhinovirus infection. A number of candidate treatments for certain picornaviral diseases are currently in clinical trials. Most of these compounds work by binding the viral capsid and inhibiting the adsorption or penetration/uncoating of the viral particle. Other alternatives which do not depend upon the three dimensional structure of the viral capsid are preferable. Despite the availability of two effective polio vaccines, susceptible individuals still exist even in well- vaccinated populations. World-wide, there occur more than

200,000 cases of polio annually.

Many regions of the RNA associated with picornaviruses are appropriate targets for ribozyme attack. The untranslated regions of the genomic RNA, however, are of particular interest. The positive strand genomes of picornaviruses are uncapped and have untranslated regions with between 600 and 1300 bases.

These regions contain extensive secondary structure which correlates well with the virus groupings in genera.

Extensive regions of sequence homology _. exist within the

5'-NTR of the human picornaviruses. For example, see Table 2 which compares the RNA sequences of 8 human picornaviruses in a region known to influence initiation of translation and poliovirus virulence. Thus, this region is a good ribozyme target for all human picornaviruses. A ribozyme targeted to this region in these picornaviruses may be useful for targeting an equivalent region in related viruses whose sequences are unknown.

Table 2. Comparison of human picornavirus RNA sequences around nucleotide 440 in the genomic RNA. _

base virus

441 AGAGUCCUCCGGCCCCUGAAUGCGGCUAAUCCUA 437

442 GU 442 GU

429 GU C

448 G C U 438 U CU U C U

440 U C U

CxA21=coxsackie A21, CxBl=coxsackie Bl, CxB3=coxsackie B3 , CxA9=coxsackie A9, Pl=poliovirus 1, P3=poliovirus 3, HRV89=human rhinovirus 89, and HRVA=human rhinovirus A.

Only the nucleotides which are different from the prototypic sequence are listed.

There is a high degree of nucleotide sequence homology within the secondary structures of the human and non-human picornavirus 5' non-translated regions.

Particularly useful regions of homology for the cardioviruses is between nucleotides 450-834, and for the human viruses are regions between nucleotides 125-164, 236-538 and 564-828 of poliovirus, and their homologous positions in other genomes.

Preferred cleavage for picornaviruses is at regions required for viral replication (e.g. , protein synthesis, such as a 5' non-translated region), the poliovirus 2A or 3C proteases, and the 3'-NTR and their equivalent in other viruses. Alternative regions may also be used as targets for ribozyme-mediated cleavage of these viral genomes. Each target can be chosen as described below, e.g. , by a study of the secondary structure of the RNA, and the individual role of such RNA in the replication of the virus. If the targets are contained within the open reading frame of the polyprotein in regions which encode proteins essential to the replication of the virus, then these other targets are preferred candidates for cleavage by ribozymes, and subsequent inhibition of viral replication. The regions encoding the poliovirus 2A or 3C proteases and their homologous proteins in the other viruses are examples of such preferred targets.

The genome of picornaviruses may be subject to rapid genetic drift by virtue of its RNA content and the nature of errors in genomic replication. Those regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e. , are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of picornaviruses, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all picornavirus. We have selected several such RNA regions of these viruses, and examined their nucleotide sequences for the presence of conserved areas which may be cleaved by ribozymes targeted to those regions. Three regions analyzed in detail are the 5' non-translated region, the 3' non-translated region and the 3A and 2C protease genes; other genes can be analyzed in a manner similar to that described below.

Predictions of effective target sites in the viral genes, based upon computer generated sequence comparisons, were obtained and identified as SEQ. ID. NOS. 1-10 in Fig. 6. Example 2: HIV

Acquired immunodeficiency syndrome (AIDS) is thought to be caused by infection with the virus HIV-1. At present, it is treated by administration of the drug azidothymidine (AZT) , which is thought to slow the progress of, but not cure, the disease. AZT resistant strains of HIV-1 are found to develop after a year of treatment. In some patients AZT has limited efficacy and may be found intolerable. More recently, drugs such as dideoxyinosine (DDI) and dideoxycytidine (DDC) have been tested as treatments for AIDS. None of these compounds reduce the viral load in patients, but they do treat the disease symptoms.

Rossi et al., 8 Aids Research and Human Retroviruses 183, 1992, provide a review of the use of ribozymes as anti-HIV-l therapeutic agents. They state:

An emerging strategy in the treatment of viral infections is the use of antisense DNA or RNA to pair with, and block expression of viral transcripts. RNA, in addition to being an informational molecule, can also possess enzymatic activity. Thus, by combining anti-sense and enzymatic functions into a single transcript, it is now possible to design catalytic

RNAs, or ribozymes, which can specifically pair with virtually any viral RNA, and cleave the phosphodiester backbone at a specified location, thereby functionally inactivating the viral RNA. In carrying out this cleavage, the ribozyme is not itself altered, and is thus capable of recycling and cleaving other molecules, making it a true enzyme. There are several different catalytic motifs which possess enzymatic activity, and each one of these can be incorporated into an

enzymatic antisense with site-specific cleavage capabilities.

Rossi et al. also state that studies have demonstrated that a hammerhead ribozyme .targeted to the gag gene RNA in the vicinity of the translational initiation codon is capable of specifically cleaving that target in a complex milieux of total cellular RNA. With reference to identification of ribozyme targets in HIV-1 they state that mRNAs for the two regulatory proteins tat and rev are clearly targets of choice, and that they are examining potential ribozyme cleavage sites in the tat mRNA, as well as in the exon shared by tat and rev. In addition, they state:

A rational approach to the problem of target selection involves the following criteria. First, one should select a functionally important target, such as tat, rev, int. psi (packaging site) , or the tRNA ^lys priming site. Once a gene or target region has been decided upon, the nucleotide sequence should be assessed for strong conservation of sequence among the various isolates. Within these conserved regions, the potential cleavage sites, preferably GUC or GUA (others will suffice, but appear to be less efficiently cleaved) should be chosen. The region should be examined for potential secondary structures, and then the most promising sites chosen. Finally, before testing the ribozyme in cell culture, it is advisable to carry out a series of in _. vitro cleavage reactions (preferably kinetic analyses) using long (at least 100 nucleotides in length) substrates to verify that the chosen sites are truly structurally favorable for cleavage. [Citation omitted.]

Rossi et al. further state that a target which deserves further consideration and testing as a potential ribozyme cleavage site is the viral packaging signal or psi sequence.

Sioud and Drlica, 88 Proc. Natl. Acad. Sci. USA 7303, 1991, describe ribozymes designed to cleave the integrase gene of HIV. They state that when the ribozyme is transcribed from a plasmid in E . coli it leads to destruction of the integrase RNA and complete blockage of integrase protein synthesis. They state that the HIV-1 integrase gene may be a useful target for therapeutic ribozymes.

Heidenreich and Eckstein, 267 J. Biol. Chem. 1904, 1992, describe three ribozymes targeted to different sites on the long terminal repeat (LTR) RNA of HIV-l. They also describe the influence of chemical modifications within the ribozyme on the cleavage of the LTR RNA, including 2 '-Fluorocytidine substitutions and phosphorothioate internucleotidic linkages.

Weerasinghe et al., 65 J. Virology 5531, 1991, describe ribozymes designed against a conserved region within the 5' leader sequence of HIV-1 RNA.

Chang et al., 2 Clinical Biotechnology 23, 1990, describe ribozymes designed to target two different sites in the HIV-1 gag gene, and a single site in the viral 5'- LTR region.

Lorentzen et al., 5 Virus Genes 17, 1991, describe a ribozyme targeted to the virion infectivity factor (vif) of HIV-1.

Sarver et al., 247 Science 1222, 1990, describe ribozymes in the hammerhead family targeted to HIV-1 gag transcripts. They state that cells challenged with HIV-1 showed a substantial reduction in the level of HIV-1 gag RNA relative to that in nonribozyme-expressing cells, and that the reduction in gag RNA was reflected by a reduction in antigen p24 levels. They state that the results suggest the feasibility of developing ribozymes as therapeutic agents against human pathogens such as HIV-1. Hampel et al., "RNA Catalyst for Cleaving

Specific RNA Sequences", filed September 20, 1989, which is a continuation-in-part of U.S. Serial No. 07/247,100,

filed September 20, 1988, describe hairpin ribozymes, and provides an example of such a ribozyme apparently specific to the gag gene of HIV-1. Hampel and Tritz, 28 Biochemistry 4929, 1989 and Hampel, et al., 18 Nucleic Acids Research 299, 1990, also describe hairpin catalytic RNA models and state that one target site is the tat gene in HIV-1.

Goldberg et al., WO 91/04319, and Robertson and Goldberg, WO 91/04324, describe ribozyme expressed within a hepatitis delta vector and state that the genome of the delta virus may carry a ribozyme against the env or gag mRNA of HIV. Rossi et al., WO 91/03162, describe chimeric DNA-RNA catalytic sequences used to cleave HIV-1 gag transcript or the 5' LTR splice site. The invention features cleavage of the RNA of these viruses by use of ribozymes. In particular, the ribozyme molecules described are targeted to the nef, vif , vpr and rev viral genes. These genes are known in the art, see, e.g.. Matsukura et al., 86 Proc. Natl. Acad. Sci. USA 4244, 1989, Cheng-Mayer et al., 246 Science 1629, 1989, Viscidi et al., 246 Science 1606, 1989, Malim et al., 86 Proc. Natl. Acad. Sci. USA 8222, 1989, Terwilliger et al., 88 Proc. Natl. Acad. Sci. USA 10971, 1991, and Bartel et al., 67 Cell 529, 1991 and Figs. 2A and 2B. Preferred cleavage for HIV is at regions required for viral replication (e.g.. protein synthesis) , e.g.. the vif , nef , vpr or rev gene regions, or at structures known to regulate viral gene expression, e.g.. tar, rre or 3'-LTR regions. As above, the genome of HIV-1 is subject to rapid genetic drift by virtue of its RNA content and the nature of errors in reverse transcription. Those regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved

between different types or strains of immunodeficiency viruses, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all HIV-1 virus, as well as all HIV-2, SIV and FIV viruses. We have selected several such genes of HIV-1, and examined their nucleotide sequences for the presence of conserved regions which may be cleaved by ribozymes targeted to those regions. The 2.5 kb subgenomic mRNA was analyzed in detail especially in the regions encoding the vif and nef proteins; the tat , vpr, vpu and rev proteins could also be synthesized from this mRNA. Nucleotide sequences were acquired from the Los Alamos HIV gene bank.

The HIV-1 genomic sequences of the Los Alamos data bank were compared in the regions encoding the vif and nef genes. Fifteen putative ribozyme cleavage sites were found to be highly conserved between the 11 strains of virus sequence. These sites represent the preferable sites for hammerhead ribozyme cleavage within these two target RNAs. Two of the nef gene sites overlap regions within the 3'-LTR of the HIV-1 genome, which represents another target of potential therapeutic value. All of the nef targets are present in all known HIV-l mRNAs and may represent targets for cleavage which would disrupt 3' terminal control regions of the mRNA which may be required for efficient translation or export of the RNAs. In a similar manner, a number of the vif target sites are present in the pol , tat and vpr mRNAs (see Fig. 2).

Using the comparative sequence analyses and an analysis of the predicted RNA structure across this region, the following predictions of effective target sites in the 2.5 kb subgenomic mRNA of HIV were obtained (and identified as SEQ. ID. NOS. 1-58 in Fig. 7) . The HIVPCV12 sequence was used as a prototypic strain for analysis. The nucleotide numbers given in the table correspond to the 5' residue within the target sequence in the 2.3 kb mRNA sequence of Arya and Gallo, 83 Proc. Natl. Acad. Sci. USA 2209, 1986.

Example 3 : HBV

The following is a discussion of relevant art, none of which is admitted to be prior art to the pending claims. Examples of art discussed below include Kato et al., 266 J. Biol. Chem. 22071, 1991; Price et al., 86 Proc. Natl. Acad. Sci. USA 8541, 1989; and Nassal et al., 63 Cell 1357, 1990.

Although acute hepatitis following perfusion exposure to human serum or blood was recognized more than 100 years ago, a causative agent was not identified until the 1960s when hepatitis B surface antigen (HBsAg) was discovered and subsequently shown to be a viral antigen. Many liver infections are asymptomatic, but the virus is hepatotrophic and may persist for years. Persistent infections are often associated with minimal liver damage, but the risk of cirrhosis or hepatocellular carcinoma in patients carrying HBV can be more than 200 times greater than in non-carriers.

The Hepatitis B virus (HBV) is the prototypic member of the family of viruses known as the hepadnaviridae. The 3.2 kb genome is characterized by extensively overlapping reading frames which express different proteins through the use of internal initiator codons (AUGs) . Three major transcripts are synthesized from the HBV genome, but the major pregenomic mRNA (3200 nucleotides in length) is probably the template for DNA synthesis and the translation of the C and P gene products. The transcript from which the X gene is translated is unclear. In principle, any of the three transcripts can function as a polycistronic mRNA and be utilized for the translation of proteins, including the X protein. Although all RNAs contain an 85 nucleotide encapsidation signal near their 5' ends, only the shortest mRNA is encapsidated into core viral particles and used for reverse transcription into viral genomes. Recently, it was demonstrated that the progression of the 80s ribosome through the region containing the encapsidation

signal would preclude the functioning of the signal and prevent the encapsidation of an RNA into the core particles. Thus, disruption of this signal and the prevention of ribosomal progression could be utilized to alter protein expression and viral maturation.

Virions of HBV are spherical particles which are surrounded by a lipid envelope and contain the small circular DNA genome of approximately 3200 nucleotides. The virion also contains a DNA polymerase (reverse transcriptase) activity and the viral core protein (HBcAg) . The HBsAg protein is found in the viral envelope and may exist as three different species. The antigenic complexity of the HBsAg is much greater than expected from the three species synthesized. Greater than eight subtypes of HBsAg can be found in clinical isolates. The P gene (polymerase) shares sequence homology with sequences in many retrovirus pol genes. The X gene encodes a polypeptide of unknown function during viral replication, but which can serve as a transactivator of transcription, and thus is a candidate oncoprotein in the morphogenetic transformation of hepatocytes infected with HBV.

Much of the hepatocellular injury which is associated with mild cases of hepatitis is thought to be caused by immunologic defenses, particularly cytotoxic T- cell responses. It has additionally been shown that chronic carriers have impaired ability to induce interferon protection in infected cells. The role of HBV in progression to hepatocellular carcinoma (HCC) is supported epidemiologically, but no gene product has been demonstrated to be essential. In a number of cases the X gene has been found in an integrated form, often with deletions, but the duck HBV contains no X gene and thus the role of X in duck HCC is not supported. For some disease syndromes there is evidence that HBsAg-anti-HBs complexes play a role in the pathogenesis, and usually these extrahepatic

manifestations occur in the form of skin rash, glomerulitiε, arthritis, or necrotizing vasculitis. One third of the patients with biopsy proven polyarteritis nodosa have persistent HBV infection. A fraction of the membranous glomerulonephritis cases are associated with chronic acute HBV and persistent HBV infections. Immune complex deposits can be found along the subepithelial surfaces of the glomerular basement membrane. Pathogenesis appears to be tissue damage mediated by immune complexes.

In general, treatment of chronic hepatitis B has been relatively unsuccessful. Corticosteroids and adenine arabinoside have not been beneficial in the treatment of chronic hepatitis B although some patients do respond to treatment with steroids. After cessation of treatment, these patients show a rebound in serum aminotransferase levels due to an immune response to the increased levels of Hepatitis B which accumulate in the blood during steroid treatment. Interferons have been used to treat chronic

Hepatitis B. Alpha-interferon appears to be more effective than beta- or gamma-interferons although its effectiveness as a single drug has yet to be determined. Although vaccines for HBV exist, attempts to recruit individuals at risk and institute pre-exposure prophylaxis have been unsuccessful. Adequate control of the disease will require vaccination of newborn infants with a subunit or killed virus vaccine. Approval of such a vaccination regimen is distant. Post-exposure prophylaxis using vaccine or vaccine and Hepatitis B immunoglobulin is sufficient to prevent HBV infection following percutaneous inoculation, oral ingestion or direct mucous membrane contact with HBsAG positive material. The timing of the initial dose of immune globulin is crucial to the success of the treatment. There is no precedent for recommending immunoprophylaxis if exposure to HBV has occurred more than 7 days earlier.

Thus a need exists for effective treatment of this disease which utilizes antiviral inhibitors which work by mechanisms other than those currently, utilized in the treatment of both acute and chronic ^' hepatitis B infections.

Ribozymes of this invention exhibit a high degree of specificity for only the viral mRNA in infected cells. Ribozyme molecules targeted to highly conserved sequence regions allow the treatment of many strains of human HBV with a single compound. No treatment presently exists which specifically attacks expression of the viral gene(s) which is responsible for transformation of hepatocytes by HBV.

The methods of this invention can be used to treat human hepatitis B virus infections, which include productive virus infection, latent or persistent virus infection and HBV-induced hepatocyte transformation. The utility can be extended to other species of HBV which infect non-human animals where such infections are of veterinary importance.

Preferred cleavage sites are genes required for viral replication, e.g.. protein synthesis, such as the 5' most 1500 nucleotides of the HBV pregenomic mRNAs. For sequence references, see Renbao et al., 30 Sci. Sin. 507, 1987. This region controls the translational expression of the core protein (C) , X protein (X) and DNA polymerase (P) genes and plays a role in the replication of the viral DNA by serving as a template for reverse transcriptase. Disruption of this region in the RNA results in deficient protein synthesis as well as incomplete DNA synthesis (and inhibition of transcription from the defective genomes) . Cleavage 5' of the encapsidation site will result in the inclusion of the cleaved 3' RNA within the core virion structure and cleavage 3' of the encapsidation site will result in the reduction in protein expression from both the 3' and 5' fragments.

Alternative regions outside of the 5' most 1500 nucleotides of the pregenomic mRNA also make suitable targets of ribozyme-mediated inhibition of HBV replication. Such targets include the mRNA regions which encode the viral S gene. Selection of particular target regions will depend upon the secondary structure of the pregenomic mRNA. Targets in the minor mRNAs can also be used, especially when folding or accessibility assays in these other RNAs reveal additional target sequences which are unavailable in the pregenomic mRNA species.

As with other viruses, the genome of HBV may be subject to rapid genetic drift by virtue of its RNA content and the nature of errors in reverse transcription. Those regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of HBV viruses, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all HBV viruses. We have selected several such genes of these viruses, and examined their nucleotide sequences for the presence of regions which may be cleaved by ribozymes targeted to those regions. One region analyzed in detail is the 5' 1500 base region; other genes can be analyzed in a manner similar to that described below.

The mRNA sequences of the virus are folded using RNAfold computer analyses. The regions of the mRNA, in this case the 5' most 1500 nucleotides of the HBV pregenomic mRNA, which are predicted to have weak or non- base paired nucleotides are searched for putative ribozyme recognition motifs. These sites represent the preferred sites for hammerhead or other ribozyme cleavage within these target RNAs.

Using such analyses, predictions of effective target sites in the genes, based upon computer generated RNA structure analysis, were obtained (see SEQ. ID. NOS. 1-33 in Fig. 8) . The target nucleotides are noted as the pregenomic RNA nucleotide numbers (in the HBV sub-type adr-1 pregenomic mRNA) given to the left of the sequence. Flanking nucleotides are given for reference. These regions are selected from genes encoding polymerase, core protein, and X protein. Example 4: Papilloma

Examples of art discussed below include Scheffner et al. , 63 Cell 1129, 1990, Hausen, 184 Virology 9, 1991, Jones et al., 267 J. Biol. Chem. 908, 1992, McCance et al., 85 Proc. Natl. Acad. Sci. USA 7169, 1988, Storey et al., 19 Nucleic Acids Research 4109, 1991, Greenfield et al., 88 Proc. Natl. Acad. Sci. USA 11217, 1991.

The papillomaviruses are small viruses containing double-stranded DNA genomes, whose gene expression has been etiologically implicated in some malignant tumors. The link between papillomavirus replication and some benign squamous epithelial cell tumors was the first recognized example of virally induced cellular transformation. There are over 60 different human papillomaviruses associated with a variety of clinical lesions and over 20 of these have been closely linked to lesions of the anogenital tract. Some of these viruses are associated only with benign proliferative lesions (e.g. , HPV-6 and HPV-11) while others are considered to be associated with a high risk of progression from preneoplastic lesions to carcinomas of the anogenital region (e.g.. HPV-16, HPV-18, HPV-31, HPV- 33, HPV-35) .

The papillomaviruses are highly species specific and have a specific tropism for squamous cells. Viral replication is restricted to the most terminally differentiated keratinocytes of the epithelium, although

viral DNA can be found within basal and parabasal cells of a papilloma. Viral transcripts have been found in the basal cells of the epidermis, but late gene expression occurs only in the terminally differentiated cells. Thus, viral replication is latent in basal cells and vegetative (productive) in the more terminally differentiated cells within the upper layers of the epithelium..

Papillomaviruses encode trans-acting factors that modulate viral transcription. The first of these factors to be described was the E2 protein. The E2 ORF contains domains that are relatively well-conserved among the papillomaviruses. E2 proteins are DNA-binding proteins which consist of two highly conserved domains joined by a more degenerate central region which is considered to be a linker region. Limited homology of E2 protein with the cellular oncogene c-mos has been noted. The E7 genes of HPV-16 and HPV-18 and the E6 gene of HPV- 18 have been reported to exhibit transactivational properties. The E5 ORF of bovine papillomavirus (BPV) encodes a 44 amino acid protein which is composed of two domains: (i) a strongly hydrophobic domain at its amino terminus and (ii) a hydrophilic carboxy-terminal domain that has the capacity to induce cellular DNA replication in quiescent cells. The E5 ORF is highly conserved among the papillomaviruses which induce fibropapillomas and has structural analogues in many other papillomaviruses, including HPV-16.

Expression of the HPV E6 and E7 genes is key to the development of neoplastic lesions. The E6 protein binds to the cellular p53 protein, which is a native tumor suppressor, and causes the degradation of the p53 protein. This degradation results in the loss of tumor suppression which is mediated by p53. Elimination of E6 expression may result not only in loss of viral replicative ability, but also in the maintenance or renewal of the normal

cellular inhibition of tumor induction in pre alignant cells.

Expression of the HPV E7 protein results in the binding and methylation of the retinoblastoma growth suppressor protein (RB) . Because the RB protein is also implicated in the control or regulation of cellular growth, this interaction may be responsible for some aspect of the transformation in HPV infected epithelial cells. As was the case above with the E7/p53 interaction, reduction of E7 expression will inhibit an important function of HPV gene expression and result in increased levels of RB protein control of cell growth. It is currently believed that the expression of E6 protein is a prerequisite for E7 expression. The expression of both E6 and E7 may be required for the two step changes required to transform a normal epithelial cell to a cell which is capable of causing metastasizing tumors after the introduction into appropriate animal systems.

The development of a model system for the study of HPV-induced neoplasia has been hampered by the absence of both a permissive tissue culture system for viral replication and an easy method for growing normal stratified epithelium in vitro . The implantation of HPV- infected tissue under the renal capsules of nude mice results in features of HPV infection in vivo , but the technique is not amenable to the genetic manipulation of the virus. Most in vitro transformation studies have been done in rodent cells and do not accurately mimic the transformation process in differentiating epithelial cells.

An in vitro system has been developed which allows stratification and production of differentiation- specific keratins. Kopan et al., 105 J. Cell Biol. 427, 1987. This system uses human epithelial cells containing transfected copies of HPV-16 DNA and the cells retain the ability to stratify and express differentiation-specific keratins but lose their ability to differentiate

morphologically. The histopathological lesions produced by transfection of HPV DNA in these keratinocytes is typical of intraepithelial neoplasia.

For the most part warts are benign, self- limiting proliferative lesions. All warts consist of localized epithelial hyperplasia with a well-defined boundary and an intact basement membrane. All of the normal layers of the skin are represented in a wart. Since only the basal layers of the epithelium have the potential for multiplication, a wart must result from the infection, transformation and multiplication of a basal cell. Studies of such clonal expansion in genital warts have suggested other scenarios. Infection probably occurs as a result of exposure of basal cells to infectious virus particles after mild trauma to the epithelium (e.g. , trauma such as that which occurs during sexual intercourse, skin abrasions or passage through the birth canal) . The virus stimulates the proliferation of basal cells, initiating the formation of a wart. The HPVs fall naturally into two classes, cutaneous HPVs (about 33 types) and mucosal HPVs (about 25 types) . About 17 of the cutaneous types have been recovered from patients suffering from epidermodsplasia verruciformiε (EV) . The genital tract is the main reservoir for the mucosal HPVs, although HPV-6 and HPV-11 will infect the respiratory tract, the oral cavity and the conjunctiva. HPV-13 and HPV-32 exclusively infect the oral cavity. The clinical significance of warts is determined primarily by their location, the viral type and host factors such as immune competence. Skin warts are often transient and self-limiting. A wart on the vocal cords can be life-threatening.

There is no "one-time" treatment that will cure all warts. Effectiveness of therapy is difficult to assess because on the one hand spontaneous regression can be observed with recurrence on the other hand. The various kinds of therapy include application of caustic

agents (e.g. , podophyllin) , cryotherapy, use of DNA replication inhibitors (e.g. , 5-fluorouracil) and surgical therapy. Immunotherapy has been unsuccessful. Interferon treatment has given mixed results depending upon the type of wart and the immune status of the patient. There is no immediate hope for effective immunization. Elimination of the latent viral state is not possible using current non- invasive therapies.

These ribozymes exhibit a high degree of specificity for only the virally encoded mRNA in infected cells. Ribozyme molecules targeted to highly conserved sequence regions will allow the treatment of many species of human papilloma virus with a single compound. There is no acceptable treatment which will give a broad spectrum of activity with no toxic side effects. No treatment exists which specifically attacks viral gene expression which is responsible for the transformation of epithelial cells by HPV, for the maintenance of the episomal genome in latently infected cells or for the vegetative replication of the virus in permissive cells.

The methods of this invention can be used to treat human papilloma virus infections, which include warts and epithelial cell transformation. The utility can be extended to other papilloma viruses which infect non- human primates where such infections are of veterinary importance.

Preferred cleavage sites are at genes required for viral replication, e.g.. protein synthesis, such as in the HPV 16, E6 and E7 mRNAs. Alternative regions outside of the E6 and E7 genes make suitable targets of ribozyme-mediated inhibition of HPV replication. Such targets include the viral E2 transactivator gene and the E5 gene which has been shown to be important in some animal virus systems. As above, those regions (genes) of the genome which are essential for papilloma virus replication are expected to maintain a constant sequence (i.e.. are

conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of papilloma viruses, and thus oniy one ribozyme is needed to destroy all viral RNA. Thus, one ribozyme may be used to target all papilloma virus encoded RNA. We have selected several such genes of these viruses, and analyzed the structure of the encoded mRNAs using RNAfold computer analyses. We have identified open structures in the RNAs which may be cleaved by ribozymes targeted to those regions. Two regions analyzed in detail are the HPV- 16, E6 and E7 mRNAs; other genes can be analyzed in a manner similar to that described below.

The mRNA sequences of the viruses are analyzed throughout the regions discussed above. Putative ribozyme cleavage sites are found to be weak or non-base paired regions of the virus mRNA. These sites represent the preferred sites for hammerhead or other ribozyme cleavage within these target RNAs. Using such analyses, predictions of effective target sites in the viral mRNAs, based upon computer generated sequence comparisons, were obtained (see SEQ. ID. NOS. 1-18 in Fig. 9) . The target nucleotide is listed with the genomic nucleotide number (in the HPV-16 genome) noted as base number. Flanking nucleotides are given for reference. Example 5: EBV

Epstein-Barr virus belongs to the subfamily gammaherpesvirinae. These viruses are distinguished by their in vitro host range restriction to B-lymphocytes, their ability to establish latent infections, and their propensity to oncogenicity. Relatives of EBV can be found only in Old World primates. The details of gene expression in the other lymphocryptoviruses is not well understood although some proteins of these other species are cross-reactive with EBV-specific antibodies. At least two EBV types have been identified in most human

populations. These are called EBV-1 and EBV-2 to parallel the nomenclature of HSV-1 and HSV-2, although the EBV types are more closely related in terms of sequence homology and restriction enzyme profiles. EBV usually initiates infection in the oropharyngeal epithelium, which is permissive for viral replication. A persistent active infection ensues which continues for many years. Early in the course of the primary infection EBV infects B-lymphocytes, which pass close to the epithelial basement membrane of the oropharynx. EBV does not replicate in B lymphocytes but sets up a latent infection. Events associated with the establishment of this latent infection transform the lymphocytes and prevent the normal course of cell death or apoptosis. It is the study of these latently infected B lymphocyte populations which has provided much of the in vitro experimentation into understanding the molecular bases of EBV gene expression.

At least nine EBV genes are expressed during the latent infection. Two encode small, non-polyadenylated RNAs (EBER-1 and EBER-2) . Six genes encode nuclear proteins (EBNA-1, -2, -3A, -3B, -3C or -LP) . One encodes an integral membrane protein (LMP-1) which is known to interact with the cellular gene bcl-2 in a way which leads to the immortalization of cells. While the EBNA and LMP genes are transcribed by cellular polymerase II, the EBER genes are transcribed by cellular polymerase III even though the EBER promoter regions also contain elements known to be recognized by the polymerase II molecules. The EBNA-3 and EBNA-1 mRNAs are less abundant than the EBNA-2 or EBNA-LP mRNAs. All of the EBNA regions share the 5' leader sequence. While the EBNA-LP is encoded by the 5'-most ORF, the other proteins are encoded by ORFs which are located 3' to this region. EBNA-2 is translated more efficiently from a mRNA which is lacking the 5' leader sequence, suggesting that EBNA-2 mRNA with the long leader sequence can moderate internal initiation of

translation. EBNA-1 and EBNA-3 proteins are also translated from a 3'-ORF and contain glycine-alanine repeat motifs of variable lengths within their ORFs.

Although most of the EBNA proteins bind to DNA cellulose, only EBNA-1 binds with high affinity to specific sites within the EBV genome. The cognate sequences are located at three sites within the genome, two of which comprise the oriP domain. EBNA binding to the oriP domain allows viral DNA to replicate and persist as an episome in latently infected cells. In latently infected cell nuclei, EBNA-1 is associated with chromatin and may interact with cell proteins to ensure that EBV episomes are partitioned to progeny nuclei. The abundance of EBNA-1 in nuclei is disproportionately high considering the low abundance of the EBNA-1 mRNA, suggesting that there is a translational up-regulation of EBNA-1 synthesis.

Little is known of the EBNA-2 role in latent infection although there is evidence that the EBNA-2 proteins are direct or indirect transactivators of latent infection and viral gene expression.

Despite having weaker promoters, the LMP gene produces almost 10 times as much mRNA as the EBNA gene. LMP mRNA has two introns and encodes an integral membrane protein. LMP is phosphorylated and associates with the cytoskeleton. Nascent LMP has a half-life of between 2-5 hours and represents the conversion time to the form of LMP associated with the cytoskeleton. Once in the cytoskeleton, LMP has a half-life of 16-20 hours. LMP appears to interact with itself or other cell membrane components to form large, non-covalently linked, membrane complexes. More recently, LMP was shown to regulate the expression of the cellular oncogene bcl-2 which in turn is responsible for the loss of apoptosis in transformed cells.

The 1 kb BZLF1 mRNA and the 2.8 kb BRLF1 mRNA are key immediate-early type transactivators of early EBV

gene expression in the replicative phase of the virus gene expression. A unique fusion mRNA, the RAZ ORF mRNA which is composed of regions from both the BZLF1 and BRLF1 ORFs as well as the BZLF1/BZLF2 intergenic region, encodes a unique tyrosine kinase protein which may also be an important regulator of EBV gene expression. In transient transfection studies, both the BZLF1 and the BRLF1 proteins transactivate the strong EBV early promoters. The BZLF1 mRNA includes an exon which has homology to the c-fos/c-jun family of transcriptional activators. Like c- fos, BZLF1 will bind directly to API-like DNA segments.

DNA replication of EBV is catalyzed by a viral DNA polymerase which resembles the situation observed for replication of the other herpesvirus DNAs. The BALF2 ORF encodes the 110 kd DNA polymerase protein. The EBV polymerase is unlike the HSV polymerase in its salt sensitivity and resistance to amphidicolin. This basic difference between the HSV and EBV polymerases may account for the relative differences in the viral sensitivities to acyclovir and other inhibitors of HSV DNA polymerase activity. A number of other EBV genes are known to encode viral proteins which are accessory factors in the viral DNA replication.

Infectious mononucleosis (IM) is a polyclonal transformation of B cells which is initiated by EBV. Autoantibodies are the hallmark of IM.EBV infection. IM results in polyclonal B-cell activation, including the activation of many cells which are not even infected with EBV. It is likely that many of the clinical manifestations of IM are due to immune responses to the proliferating B-cells.

A number of recent reports have described the occurrence of diffuse lymphoproliferative disease with features of malignant lymphoma occurring in the wake of infectious mononucleosis. The infected patients are a heterogeneous group, including both girls and boys, but half of the cases are males with the X-linked

Lymphoproliferative Syndrome (XLPS) . The lymphomas contain EBV DNA and express EBNA. Boys with XLPS are not globally immunodeficient although they do express a deficiency in natural killer cell activity. Although the association of EBV with Burkitts

Lymphoma (BL) has been known since the discovery of the virus, the exact role which the virus plays in the etiology and pathogenesis of BL is still controversial. More than 90% of the tumors contain EBV DNA and express EBNA. In lymphoblastoid cell lines there is expression of all identified latent nuclear products and LMP. However, in BL cells there is repression of EBNA-2 and LMP, although the cells still produce EBNA-1. This could mean that the same genes which are required for immortalization are not required in BL cells, which may be immortalized by some other mechanism, perhaps induction of c-myc activity by the chromosomal translocations which are known to take place in BL cells.

The association of EBV with nasopharyngeal carcinoma (NPC) is strong. EBV DNA is regularly found in all cases of undifferentiated NPC and in a number of cases of differentiated NPC. Expression of EBV genes in NPC differs from that observed in BL and IM. The two most notable features are the limited expression of the latent proteins to EBNA-1 and LMP, and expression of several genes which are thought to be part of the early-antigen complex.

EBV induces fatal lymphoproliferative disease, sometimes with the features of frank lymphoma. There are thought to be several components of these progressive lymphoproliferative diseases in immunodeficient patients. The tumors which arise are often polyclonal initially, which implies that the tumors arise from several independent transformation events. With time the tumors evolve to an oligoclonal or even monoclonal state. Most EBV-associated lymphoproliferative lesions do not exhibit the same chromosomal translocations observed with BL.

Patients which exhibit these types of tumors include post- transplant recipients and AIDS patients. AIDS patients develop three types of EBV-associated lesions: (i) diffuse polyclonal lymphomas, often at extranodal sites such as the gut and central nervous system, (ii) lymphocytic interstitial pneumonitis (LIP) , and (iii) hairy cell leukoplakia of the tongue. LIP occurs mainly in children with AIDS and is not a progressive lymphoproliferative disease. Hairy leukoplakia is associated with chronic production of EBV in epithelial cells of the tongue. In the basal layers, the BZLF1 gene product is found; in the upper layers, early and late viral gene products are expressed.

The invention features the use of ribozyme inhibitors of EBV replication, especially the productive infection of epithelial cells or the establishment of latent EBV infections in B-lymphocytes. This invention may also be used to reverse the immortalization of B- lymphocytes or latent EBV infections. The cleavage of EBV mRNA or hnRNA using this new class of chemical inhibitors, ribozymes, allows specific interruption of viral replication or maintenance of EBV latency by cleavage of the viral RNA. This class of chemical cleavers will exhibit a high degree of specificity for only the viral RNA in infected cells. A ribozyme molecule targeted to a highly conserved sequence region will allow the treatment of all strains of EBV with a single compound. No current treatment can impact either the latent state or the transformed phenotype associated with EBV infections. All treatments with nucleoside analogues are virustatic and the effects are reversed upon withdrawal of the drug. Ribozymes offer a non-reversible treatment for EBV infections, and thus are virucidal. This invention allows treatment of Epstein-Barr virus infections in humans. Different ribozyme preparations can be used in treating productive,

persistent and latent EBV infections. Moreover, ribozymes targeted to the LMP mRNA are useful in reversing the immortalization event in lymphocytes and thus in allowing apoptosis in affected cells. Ribozymes targeted to the EBNA-1 and RAZ mRNAs can be used to disrupt the latent and productive stages of EBV infection, respectively.

Preferred cleavage sites are at genes or regions required for viral replication, e.g.. protein synthesis, e.g.. those genes and regions described above, e.g.. EBNA- 1 and RAZ.

Ribozymes targeting any of the above regions of these mRNAs can be designed to cleave the mRNAs in a manner which will inhibit translation of the molecules. Alternative regions outside of the EBNA-1 and RAZ regions also make suitable targets for ribozyme-mediated inhibition of EBV gene expression. In particular, other targets include mRNAs encoding the EBNA-2 proteins, LMP protein, the viral DNA polymerase and the BZLF1, BRAF1 and BMRF1 proteins. Secondary structure analyses of the mRNAs or hnRNAs encoding these proteins can be used to determine ribozyme-targeting regions, to aid selection of these other regions of viral RNA which are good candidates for cleavage by ribozymes and subsequent inhibition of viral replication, disruption of latent viral infection, or termination of B cell immortalization.

As above, those genes of the EBV genome which are essential for virus replication are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of EBV, and thus only one ribozyme is needed to destroy all such viruses. We have selected several such genes of EBV, and examined their nucleotide sequences for the presence of weak or non-base paired regions in the mRNA which may be cleaved by ribozymes targeted to those regions. Two genes

analyzed in detail are the EBNAl and RAZ genes; the other genes noted above can be analyzed in a manner similar to that described below.

The EBV mRNA sequences are analyzed using PC fold in the regions encoding the EBNAl and RAZ genes. Putative ribozyme cleavage sites are found which are weak or non-base paired. These sites represent. the preferable sites for hammerhead ribozyme cleavage within these two target RNAs. Using the above analyses, the following predictions of effective target sites in the chosen mRNA of the EBV genome, based upon computer generated sequence comparisons, are obtained (see Fig. 10) . The target sequence is listed first with the 5'-most nucleotide number, for reference.

Preliminary predictions of effective target sites in the EBNA-1 and RAZ genes, based upon computer generated RNA structure, are listed. Example 6: T-Cell Leukemia Adult T-cell leukemia (ATL) was first described by workers in Japan. Takatsuki et al., 9 Jpn. J. Clin. Oncol. 312, 1979. Since then, the disease has been found in other areas of the Orient, the Caribbean basin. South America and central Africa. The epidemiology has shown a clear association with the presence of a retrovirus known as human T-cell leukemia virus type I (HTLV-I) . HTLV-I infection of lymphocytes in vitro results in immortalization of the same cell type as tumor cells, and HTLV-I has proven to be oncogenic in animal model systems, e.g. , the rabbit.

HTLV-I was first identified in the HUT 102 cell line (Gallo et al., 43 Cancer Res. 3892, 1983), which was established from a patient diagnosed with cutaneous T-cell lymphoma. Yoshida et al., 79 Proc. Natl. Acad. Sci. USA 2031, 1982, demonstrated that another cell line, known as MT-1, derived from a patient with ATL, also harbored a C- type retrovirus which was later named HTLV-I.

Since the initial description of ATL and discovery of HTLV-I, the virus has been shown ' to be associated with other human diseases. The most notable of these is a neurologic disorder known as tropical spastic parapesis (TSP) , or HTLV-I associated myelopathy (HAM) .

HTLV-II was first described in a T-cell line derived from a patient diagnosed with hairy-cell leukemia. The cell line (Mo-T) was derived from splenic tissue and later shown to harbor a virus which was related to (but distinct from) HTLV-I, and subsequently named HTLV-II. Like HTLV-I, HTLV-II will immortalize lymphocytes. However, limited numbers of HTLV-II associations with disease has precluded epidemiologic demonstration of an etiologic role of the virus in human malignancy. HTLV is a member of the class of retroviruses that includes bovine leukemia virus (BLV) and simian T- cell leukemia virus (STLV-I) . The viruses are grouped within the oncovirus subfamily and are distinct from the other human pathogenic retroviruses, e.g.. human immunodeficiency virus (HIV) , which belong to the subfamily of lentiviruses. Although the leukemia viruses are oncogenic, they do not harbor oncogenic sequences derived from cellular genomes and their replication is regulated by at least three other genes, known as tax, rex, and pro .

Three mRNA species have been identified for HTLV. The full length RNA, transcribed from the U3-R junction in the 5'LTR and terminating in the R-U5 region of the 3'LTR, is used for the synthesis of gag and pol gene products, and serves as the genomic RNA packaged into virions. A subgenomic RNA with one intron removed serves as the template for env protein synthesis, and a second subgenomic RNA with a second intron removed serves to encode the tax and rex products. The tax initiator codon is the same as that of the env protein and is encoded within the second exon. The rex initiator codon is also

within the second exon but is located 59 nucleotides 5' to that of the tax and env .

Unlike the protease genes of the other retroviruses, the HTLV protease gene is encoded by a reading frame which spans the 3 ' part of the gag region and the 5' part of the pol region, but is distinct from both of the other ORFs. In some of the clones of HTLV-I, mutations in the protease region may be responsible for the lack of infectivity of the clones. The protease is responsible for processing mature gag products and autocatalyzed self-cleavage to produce the mature protease molecule.

Two unique genes are found in HTLV ( tax and rex) which encode nonvirion proteins translated from a single, sliced mRNA with three exons. Deletion of sequences within these two genes render infectious clones of HTLV-II noninfectious, providing evidence that the expression of these genes is critical for viral replication.

The HTLV-I and HTLV-II tax genes encode proteins of 40 and 37 kd, respectively. The tax proteins are trans-acting transcriptional activators which increase the rate of transcription initiation from the proviral 5'LTR promoter. Using co-transfection techniques, it has been shown that the tax protein is capable of trans-activating heterologous promoters, including the promoters for IL-2 and GM-CSF, c-fos and c-sis .

The rex gene of HTLV-I/HTLV-II encodes proteins of 27 and 21 kd for HTLV-I, and 26 and 24 kd for HTLV-II. The relationship between the two protein species is uncertain. Evidence suggests that the smaller rex protein is synthesized from an internal initiation codon within the same open reading frame. Both rex proteins are phosphorylated and localized in the nucleus of infected cells. Like the product of the tax gene, the rex protein is essential to the replication of HTLV. Unlike the tax protein, the rex product appears to act post-

transcriptionally to regulate viral gene expression. For HTLV-I, rex has been shown to increase the ratio of non- spliced RNA to the fully spliced mRNA which encodes tax and rex. In HTLV-II, rex protein increases the overall level of trans-activation in concert with tax protein. Additionally, in HTLV-II rex has a negative regulatory role which decreases viral mRNA levels. The ultimate effect of rex upon HTLV-I and HTLV-II infected cells is to regulate the levels of expression of genes encoding virion components, thereby determining the production of infectious virions.

The HTLV R and U5 regions are unusually long in comparison to other retroviruses. These regions encode the non-translated leader sequences at the 5' ends of all viral mRNAs, exhibit extensive secondary structure and probably play a role in the control of translation of those mRNAs.

HTLV-I and HTLV-II share about 65% overall sequence homology at the nucleotide level. The homology is lowest in the LTR and non-translated regions and is highest in the tax and rex genes. Among different HTLV-I isolates, there is 96-99% homology in these two regions. Sequence homology is equally high among the HTLV-II isolates. Direct cell-to-cell contact appears to be required for efficient HTLV infection. In vitro infection is usually accomplished by co-cultivation of target cells with X-ray-irradiated or itomycin C-treated virus producing cells. Infection by HTLV of susceptible cells is inefficient, with a slow course, as compared to other retrovirus infections. Infection of human cord or peripheral blood lymphocytes by HTLV results in virus production, and eventual immortalization of the cells. Only T-cells have been shown to be transformed by HTLV. Viral replication can occur in other cell types under certain circumstances, such as EBV-transformed B-cells. Productive infection has been observed in a few cells of

non-lymphoid origin, most notably human endothelial cells and diploid human fibroblastε.

Infection of cells with HTLV in vitro can usually be accomplished only by co-cultivation of the cells to be transformed with virus-infected cells. Proliferating transformed cells predominate in the cultures after 4 or more weeks. Transformed lymphocytes are mostly of the CD4 ⁺ phenotype which reflects the observation that tumors in ATL patients are almost invariably of CD4 ⁺ type. However, occasional CD8 ⁺ tumors have been reported and HTLV can transform CD8 ⁺ cells in vitro . Cells transformed in vitro transcribe RNA and produce virions, and these cells can be used to transform other cells. The lack of HTLV gene expression in ATL tumor cells suggests that in vitro transformation does not directly parallel the formation of a leukemic clone in vivo . These differences may reflect differences in selection pressure between in vitro and in vivo transformed clones. Virion preparations from supernatants of HTLV- infected cells are itogenic for quiescent human T-cells. Activated T-cells are more easily transformed by HTLV than are quiescent cells. This mitogenic activity of HTLV may be an important evolutionary adaptation because 95% of the T-cells in the body are generally, in a quiescent state.

The mitogenic activity of virions suggests a mechanism for HTLV transformation whereby continual stimulation of T-cells via a T-cell receptor results in continuous cell proliferation. The corequisite of tax protein expression resembles the situation observed in transformation of primary rodent fibroblasts in vitro , where co-expression of two oncogenes, one acting in the nucleus and the other at the cell membrane, are required to effect transformation of the cell. Thus, unlike with other retroviruses, it appears that T-cell transformation is a normal course of events for infection with HTLV.

Several categories have been described as stages in ATL. These stages include: (i) asymptomatic carriers; (ii) preleukemic state (pre-ATL) ; (iii) chronic/smoldering ATL; and (iv) acute ATL. In some cases, these stages may be temporally related. The only significant prognostic factor in acute ATL is the presence of ascites, which is associated with shorter survival. Despite aggressive chemotherapy, the median survival in acute ATL is measured in months. The form of disease in some^ATL patients is that of a T-cell lymphoma rather than a leukemia. The differential diagnosis of ATL includes other T-cell malignancies such as non-Hodgkins lymphoma, mycosis fungoides, Sezary'ε syndrome and T-cell chronic lymphocytic leukemia (CLL) . One malignancy distinct from ATL in which HTLV-I may play a role is B-cell chronic lymphocytic leukemia (B-cell CLL) . There is some evidence for functional impairment of the immune response in ATL patients, as well as some HTLV-I carriers. The role of HTLV-II in the few caεeε of leukemia associated with viral infection is unclear, since insufficient cases are available for study.

ATL is a highly malignant disease where survival of subacute or acute disease is measured in months. Since only about 1% of patients progress from asymptomatic to acute disease, treatment is reserved for more advanced disease. Standard combination therapies which are used for other aggressive lymphomas and leukemias have not worked for ATL. Other agents such as deoxycoformycin, beta interferon, gamma interferon and anti-Tac antiboby have not been of value in treating ATL.

The invention also features cleavage of the RNA of these viruses by use of ribozymes. In particular, the ribozyme molecules described are targeted to the nef, vif, tat and rev viral genes. These genes are known in the art, see, e.g. , Matsukura et al., 86 Proc. Natl. Acad. Sci. USA 4244, 1989, Cheng-Mayer et al., 246 Science 1629,

1989, Viscidi et al., 246 Science 1606, 1989, Malim et al., 86 Proc. Natl. Acad. Sci. USA 8222, 1989, Terwilliger et al., 88 Proc. Natl. Acad. Sci. USA 10971, 1991, and Bartel et al., 67 Cell 529, 1991. Preferred cleavage sites are in regions required for viral replication, e.g.. protein synthesis, e.g.. the tax, rex, or pro gene regions, or at structures known to regulate viral gene expression, e.g.. the 5'-LTR or 3'-LTR regions. The genome of HTLV-I is subject to genetic drift by virtue of its RNA content and the nature of errors in reverse transcription. Those regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of immunodeficiency viruses, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all HTLV-I viruses, as well as all HTLV-II, BLV and STLV viruses. We have selected two genes of HTLV-I which are known to be highly conserved, and examined their nucleotide sequences for the presence of regions which may be cleaved by ribozymes targeted to those regions. Two genes analyzed in detail are the tax and rex genes; the pro gene, the 5'-LTR and other genes noted above can be analyzed in a manner similar to that described below.

The HTLV-I mRNA sequence of the HS-35 clone was analyzed in the regions encoding the tax and rex proteins. Twenty-three putative ribozyme cleavage sites were found to be present in open (weakly base-paired) regions of the mRNA. These sites represent the preferable sites for hammerhead ribozyme cleavage within these two target RNAs. Using such analyses, the following predictions of effective target sites in the tax and rex genes of the HTLV-I genome, based upon computer generated sequence

comparisons, were obtained (see Fig. 11) . The target sequence is listed first with the 5'-most nucleotide number, for reference. Example 7: HCV The Hepatitis C virus group iε responsible for most cases of non-A, non-B hepatitis. Although non-A, non-B hepatitis may be caused by many different viruses, the principal cause is HCV. The infection is transmitted through blood transfusions (25-50% of hemophiliacs have chronic hepatitis) , percutaneous infusion from illicit drug usage (25-50% of all drug users have non-A, non-B hepatitis) , renal organ transplant (where most late mortality is due to non-A, non-B hepatitis) , and maternal- neonatal transfer. Up to 20% of sporadic outbreaks of hepatitis are caused by HCV.

The viral particles are 30-60 nM in diameter, enveloped, and consist of a linear RNA genome which is plus polarity and contains approximately 10 kb of coding information. Not much is known about the virus family in terms of the gene expression or encoded proteins. The viral genome is a positive-stranded RNA molecule which resembles that of the Flavivirus group in composition of open reading frames and overall size. There is no nucleotide homology between the HCV genome and that of the other Flaviviruses or Togaviruses (another positive stranded RNA virus) . It is believed that the RNA is translated from a single open reading frame and then the polypeptide precursor is processed into the individual proteins required for viral replication. Aside from this superficial resemblance to the Flaviviruses, the HCV family also resembles the picornavirus family of RNA viruses. As in the picornaviruses, the 5' non-translated region of the HCV genome controls the translational expresεion of the viral RNA. The replication of HCV in hepatocyteε is integrally intertwined with the replication of other hepatitis viruses. Although hepatitis delta virus (HDV)

is quite distinct from HCV, the life cycles of the two viruses can be interdependent. Normally HDV depends upon the replication of Hepatitis B Virus (HBV) to supply capsid proteins and other functions necessary for the assembly of infectious HDV particles. Because HCV inhibits HBV and Hepatitis A virus (HAV) growth in co- infected cells, the replication of HCV can also inhibit the maturation of HDV. The use of HDV as a gene vector for treatment of HCV infections would not be influenced by this interference because gene expression of the HDV genome iε apparently uninhibited by HCV.

This clasε of chemical cleaverε will exhibit a high degree of εpecificity for only the viral mRNA in infected cells. A ribozyme molecule targeted to a highly conserved sequence region will allow the treatment of many strains of human HCV with a single compound. No treatment exists which specifically attacks expression of the viral gene(s) of HCV. Current treatment protocols (interferon and ribavirin) consist of treating the symptoms of the disease, not the viral replication. Upon cessation of current treatment protocols, a rebound of disease ensues because the treatment does not impact the latency of the virus or actual viral replication. The cleavage of viral RNA offers a novel treatment which will reduce virus replication and establishment of latency in newly infected cells.

The methods of this invention can be used to treat human hepatitis C virus infections, which include both acute and chronic virus infection and HCV-induced hepatocyte transformation. The utility can be extended to other species of HCV which infect non-human animals, where such infections are of veterinary importance.

Treatment will occur at the time of active viral infection and will reduce virus loadε in the infected cellε and also disable viral replication. Ribozyme treatment may also be used as a means of creating defective genomes which can be used in vaccines.

Preferred cleavage sites are within regions required for viral replication, e.g. , protein synthesiε, εuch aε the 5' non-tranεlated region of the HCV genome. This region is believed to control the translation of viral proteins in a manner which iε reminiεcent of the cap independent tranεlational control of picornaviruses. Disruption of this region in the RNA results in deficient protein synthesis as well as incomplete DNA synthesis (and inhibition of transcription from the defective genomes) . Alternative regions outside of the 5' non- translated region also make suitable targets of ribozyme- mediated inhibition of HCV replication. Such targets include the mRNA regions which encode the viral structural proteins. Selection of particular target regions will depend upon the secondary structure of the genome. Targets in the minor mRNAs can also be used, especially when folding or accessibility assays in these other RNAs reveal additional target sequences which are unavailable in the genome. Two or more different ribozymes can be used in this invention for therapeutic treatment.

Aε above, the genome of HCV may be subject to rapid genetic drift by virtue of its RNA content and the nature of errors in genomic replication. Those ^' regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of HCV viruses, and thus only one ribozyme is needed to destroy all such viruseε. Thuε, one ribozyme may be uεed to target all HCV viruεeε. We have selected several such genes of these viruses, and examined their nucleotide sequences for the presence of regions which may be cleaved by ribozymes targeted to thoεe regionε. One region analyzed in detail is the 5' non-translated region; other genes can be analyzed in a manner similar to that described below.

The mRNA sequences of the virus are folded using RNAfold computer analyses. The regions of the mRNA, in this case the 5' nucleotides of the HCV genome, which are predicted to have weak or non-base paired nucleotides are searched for putative ribozyme recognition motifs. These sites represent the preferred sites for hammerhead or other ribozyme cleavage within these target RNAs.

Using such analyses, the following predictions of effective target sites in the HCV genomic RNA, based upon computer generated RNA structure .analysiε, were obtained (εee Fig. 12) . The target nucleotide iε listed with the genomic nucleotide number (in the HCV genome) given to the left of the sequence. Flanking nucleotides are given for reference. Example 8: CMV

The following is provided as a general outline of relevant CMV biology.

Cytomegaloviruses (CMVs) belong to the family of betaherpeεviruε and are species-specific viruses for replication and pathogenesis. Even in the natural host, not all cells are equally susceptible to productive viral infection. In the case of human CMV (HCMV) the abortive infections of lymphocytes may even contribute to the invasion of the host immune surveillance. CMV infections proceed much more slowly than those of alphaherpesviruseε (e.g.. HSV and VZV) both in cell culture and in the animal hoεt. Infectious virus may persist for an extended period of time, after which the virus may establish a latent infection. Although these viruseε have at least 50% more sequence complexity than the alphaherpesviruses, there is still a dependence on and close asεociation with the host cell. These include a growth requirement for differentiated cells in vitro and a stimulation of host cell macromolecular synthesis. Although both families of the herpesviruεes enter the host and express their first εet of geneε very rapidly (1-3 hr) , the betaherpeεviruεes require 12-24

hours to initiate viral DNA syntheεis while the alphaherpesviruεeε initiate viral DNA synthesis within 4-6 hours. Similarly, the alphaherpesviruεeε form plaques in cell monolayers at 24-48 hour poεtinfection while the betaherpesviruses require between 7-14 dayε to form detectable plaqueε. Aε theεe results suggest, the control of genetic expresεion is quite different in the CMVs when compared to that of HSV or VZV.

Although CMVε replicate in many of the same cell types as the other herpesviruses in their natural hosts, this group of viruses is restricted in vitro to replication in fibroblastε. The reason for this restriction is unknown.

Because the CMV genome can remain in host cells in a productive, persistent or latent state, there must be a number of features which regulate the expression of viral genes. The expression of the immediate early genes of the CMVs plays an important role in the replication of the virus because the products of these genes are known to influence the expression of other viral genes, their own genes and even cellular genes. The iel gene spans approximately 2.8 kb between 0.739 and 0.751 map units and codes for a 1.95 kb RNA which consists of a 5' leader exon of 121 nucleotides, followed by three exons of 88, 185, and 1,341 nucleotides. The open reading frame (ORF) of the mRNA begins in the second exon and encodes a 491 aa polypeptide. The size of the iel protein varieε εomewhat between strains of HCMV yet some iel nucleotide sequence homology is conserved even between CMVε of different species.

By an alternate splicing pattern, exons 1, 2, and 3 of the iel region can be ligated onto the ie2 mRNA resulting in two mRNAs of 2.25 and 1.70 kb. The proteins encoded by these mRNAs contain 85 amino acids derived from the exons 2 and 3 of iel but the majority of the protein is derived from region ie2a. The 2.25 kb and 1.70 kb mRNAs differ by a small splice in ie region 2a and they

are predicted to encode proteins of 71 and 53 kd, respectively. A ε aller mRNA of 1.4 kb alεo originates from ie region 2a. This mRNA iε transcribed under independent promoter control only late in the infectious cycle.

The HCMV DNA polymerase iε physiologically distinct from host cell polymerases. The viral polymeraεe iε a 140-kd protein which iε associated with a 51-58 kd polypeptide, referred to as the polymerase aεsociated DNA binding protein. The polymerase exhibits regionε of homology with other herpesvirus polymeraεes but also has regions which are unique to the CMV polymerases. Most antiviral agents are targeted to the inhibition of the CMV polymeraεe. Another viral protein (UL97 gene product) haε been found to be linked genetically with the CMV polymerase in antiviral resistance studies. The UL97 gene has been observed to be mutated in a number of viable strains which exhibit simultaneous mutations in the viral DNA polymerase which produce a gancyclovir resistant phenotype. The interactions of these two proteins in the replication of CMV genomic DNA is uncertain, but homology of the UL97 gene with other herpesvirus genes (e.g.. the UL13 gene of HSV-1) and the general conεervation of the transcription unit which encodes the mRNA among herpesviruses suggeεtε that this gene may be important to the efficiency of viral replication. Numerous protein kinaεe motifs exist within the putative ORF of the UL97 gene, and the HSV UL13 protein is known to posεeεε protein phoεphorylating activity. Thuε, it may be assumed that the UL97 gene of HCMV helps to control gene expresεion by phoεphorylation of some CMV protein, possibly the DNA polymerase.

Many of the structural proteins of CMV are expressed at late times after infection. Although all of the structural proteins are assumed to play a role in the formation or integrity of the virion particle, the most

relevant among these are the viral glycoproteins. The gB ho olog has homology to the gB genes of other herpesviruεeε such as HSV, VZV, EBV and other animal CMVε. Thiε viral protein is encoded by an unspliced mRNA of approximately 3.9 kb that codes for a protein 906 aa in length. The amino acid sequence contains a typical eucaryotic signal sequence, and as many as 17 potential N- linked glycosylation sites and transmembrane domains near the carboxy terminus. It has been proposed that the HSV gB functions to facilitate membrane fuεion during viral infection. Since the gB genes are so highly conserved among the herpesviruεeε, it is reasonable to assume that the CMV gB protein may alεo participate in membrane fusion events during infection. As in the case for HSV, other glycoproteins of HCMV may be involved in the viral infection, including the gH homolog, gel, gcll, gcIII and the HXLF family of glycoproteins.

Humans are believed to be the only reservoir for HCMV and transmission occurs by direct or indirect person- to-person contact. Sources of virus include nasopharyngeal secretions, urine, cervical and vaginal excretions, spermatic fluids, breast milk, tears, feceε and blood. Viruε excretion persistε for yearε after congenitally, perinatally and early postnatally acquired infections. Prolonged replication of virus can follow infection in adolescents and young adults, and recurrences are generally aεεociated with periodε of intermittent εhedding in adults.

Oral and reεpiratory εpread appear to be the dominant routeε of tranεmisεion during childhood and adult life. Blood transfuεionε are associated with increased risk of both primary and recurrent CMV infection. Transfuεion of leukocyte-enriched fractions further increases the risk. CMV infectionε are even more markedly increaεed in patientε which are immunoεuppressed, such as those undergoing chemotherapy for cancer or preparing for organ transplantε. Perhaps the most striking risk group

are those patients suffering from AIDS. The occurrence of CMV infection is so great among AIDS patients that HCMV was once considered as the etiologic agent.

Studies on the pathology of CMV infections has focused upon deεcribing the distribution ^' of cytomegalic inclusion bodies in various infected tissueε. With severe disseminated disease, evidence of CMV infection can be found in virtually all organ systems. In _. the human, CMV primarily infects the ductal epithelial cells and seldom involves fibroblaεtε, unlike the growth in cell culture. Viruria is a common feature of CMV infection in all age groups and results from renal involvement. The cytomegalic inclusion bodies are most often pronounced in the proximal tubules near cortical areas. With the possible exception of renal transplant patients, CMV renal infections seldom result in kidney dysfunction. In the renal transplant patients however, CMV may play a role in dysfunction and even graft rejection. In adults, CMV hepatitis is manifest by the cytomegalic bodies in hepatocytes. In lung infections, the bodies are most concentrated in the alveolar and broncheolar epithelium. The mucous membraneε of the trachea are less often involved and most lung infections occur in older immunosuppressed patients. CMV pneumonia is one of the leading causeε of death in AIDS patientε.

In the paεt, CMV infection of the CNS was rare except in fetal infections that resulted in cytomegalic inclusion disease (CID) . In recent years, CNS involvement haε increased in frequency in adults with AIDS. Permanent, usually severe generalized brain damage is characteristic of CID of the newborn. In both children and adults glial and neuronal cells may be either productively or abortively infected. Inclusion-bearing cells in capillary epithelium suggests blood-born spread in most cases.

Gastrointestinal involvement is also on the increase. CMV may coinfect with other pathologic agents

associated with AIDS and may be present with other lesions, such as Kaposi'ε sarcoma. Necrotizing lesions carry a poor prognosiε.

Senεineural hearing loss is a feature of both congenital and neonatal infections. Cytomegalic cellε have been seen in a number of cells including epithelial cells of the cochlea and semicircular canal. Other organs with a reasonable degree of occurrence include adrenals, ovaries, boneε, pancreaε and skin. CMV shares properties with other oncogenic viruses, such as the ability to stimulate host cell macromolecular synthesis, and to stimulate the synthesiε of ornithine decarboxylase. Genetic sequences of HCMV with transforming capabilities have been described. In addition, CMV or its products have been found in human tumors, including prostatic carcinoma, adenocarcinoma of the colon, cervical carcinoma and Kaposi's sarcoma.

The invention also features cleavage of the mRNA or hnRNA encoded by these viruses by use of ribozymes. In particular, the ribozyme molecules described are targeted to the iel and ie2 viral genes. These genes are known in the art, see, e.g. _r Stenberg et al., 56 J. Virol. 665, 1985, Alerigg et al., 2 Virus Reε. 107, 1985, Stenberg et al., 63 J. Virol. 2699, 1989, Her iston et al. , 61 J^ Virol. 3214, 1987.

Preferred cleavage sites are at regions required for viral replication, e.g.. protein syntheεiε, e.g.. regionε in the intron/exon regions of the region encoding the immediate early (ie) 1 and 2 mRNAs. This region controlε the subsequent timing of viral gene expression and therefore the succeεεful initiation of viral replication. Diεruption of thiε region in the mRNA reεultε in deficient protein synthesiε aε well aε incomplete DNA εyntheεiε. Alternative regionε outεide of the HCMV ie region alεo make uεeful targetε for ribozyme-mediated inhibition of HCMV replication. Other targetε which are

preferred in this invention include the viral DNA polymeraεe, gB homolog and UL97 genes. Targets in the other HCMV genes can also be designed, and accessibility assays (see below) in these other mRNAs and hnRNA will reveal additional target sequenceε which are not predicted to be accessible from RNA folding programs. Those genes which encode RNA essential for viral replication in humans can be targeted for ribozyme cleavage.

The use of ribozyme inhibitors of cytomegalovirus replication, especially human CMV (HCMV) , allows ribozyme specific inhibition of gene expresεion which causes CMV-induced cellular transformation, CMV- induced immunodeficiency, and maintenance of latent CMV infections. The invention provides a novel method of inhibiting human cytomegalovirus replication and provides methods for the treatment of diseases caused by these viruses in both man and other animal species.

These ribozymes exhibit a high degree of specificity for only the viral RNA in infected cells. A ribozyme molecule targeted to a highly conserved sequence region allows the treatment of many strains of HCMV with a single compound. No treatment exists which specifically attacks expression of the viral gene(s) other than those involved in the replication of viral genomic DNA. The drugs currently used for clinical treatment of HCMV infectionε present serious εide effectε which preclude uεage in all patientε. For example, Gancyclovir is known to cause non-reversible gonadal toxicity, and phosphonoformic acid (PFA) causes neutropenia and renal disfunction. Both of these latter effects are significant problems in organ transplant patients.

This invention can be used to treat human cytomegalovirus infections, which include both productive virus infection and latent infectionε or HCMV-induced cellular transformation. The utility can be extended to other species of cytomegalovirus which infect non-human

animals, where such infections are of veterinary importance.

Treatment will occur at the . time of active, latent or persistent viral infection and will reduce virus loads in the infected cells and also disable viral replication. For treatment of transformed epithelial cells, the use of this invention will inhibit the expression of viral genes or related RNAs thought to cause cell transformation. The genome of CMV is subject to. rapid genetic drift by virtue of errors in DNA replication. Those genes which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. ^" These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of CMV, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all CMVs. We have selected several such genes of CMV and examined their nucleotide sequences for the presence of open mRNA regions which may be cleaved by ribozymes targeted to those regions. Two genes analyzed in detail are the iel and ie2 genes; other genes noted above can be analyzed in a manner similar to that described herein.

The CMV mRNA sequences were analyzed for weak or non-base paired regions in the sequences encoding the iel and ie2 proteins. Putative ribozyme cleavage εites were found which represent the preferable sites for hammerhead ribozyme cleavage within these two target RNAs.

Using such analyses, the following predictionε of effective target sites in the iel and ie2 genes of the CMV genome, based upon computer generated sequence comparisons, were obtained (see Fig. 13) . The target sequence is listed first with the 5'-most nucleotide number, for reference. The target nucleotide is underlined with reference nucleotide numbers given at the

beginning of each line. Flanking nucleotides which will constitute all or part of the region complementary to the ribozyme binding arms are shown. The sequences were taken from the DNA sequence of the AD169 strain of HCMV. Example 9: Influenza Virus

Three types of influenza viruseε (A, B, and C) are distinguishable by antigenic reactivities of their internal antigens. There are other biological properties which characterize the three types: (a) Influenza A viruses have been isolated from many animal species in addition to humans while influenza B and C viruseε are mainly human pathogenε; (b) the surface glycoproteins of influenza A exhibit much greater variability than their homologueε in the B and C viruεeε; (c) morphological and molecular featureε of C viruses are distinctive from those of the A and B viruses.

The morphological characteristics of influenza viruses are a genetic trait, but spherical morphology dominates after passage in chicken embryos or tissue culture. The genes that specify morphology are uncertain, but segregate separately from the hemagglutinin (HA) and neuraminidase (NA) envelope surface proteins. Within the lipid envelope lies the matrix protein (M) , which plays a structural function. Within the matrix shell are eight single-stranded RNA molecules of negative sense associated with the nucleoprotein (NP) and three large proteins (PB1, PB2, and PA) required for RNA replication and transcription. At least three viral encoded nonstructural proteins (NS1, NS2 and M2) are formed in infected cells. The organization of the eight RNA segments within the virion has not been completely resolved. Although each segment may exist in vivo as a nucleoprotein complex, electron microscopic studies have shown that the internal component from disrupted virions iε a εingle large helix. The virion particle does not seem to be a tight protective coat around the RNA because ribonucleaεe digeεtion of virion reduces the RNA segments to

nucleotide. Genomic RNAs of influenza virus are held in a circular conformation in a virion and in infected cells by a terminal panhandle that plays a role in viral replication. The panhandle structure is present in all segments of genomic RNA.

The HA accountε for 25% of viral protein and iε distributed evenly on the virion surface. It is responsible for attachment of the virus to cells and penetration of virus into cells early in infection. The HA monomer is encoded by the fourth largest RNA segment and is synthesized as a single polypeptide chain which undergoes posttranslational cleavage at a minimum of three siteε. Cleavage of the HA polypeptide into HA1 and HA2 iε necessary for virus particle infectivity. A sequence of 25-32 hydrophobic amino acids at the C-terminus of HA2 saves to anchor HA in the virus membrane. In spite of functional domain conservation in HA, the amino acid or nucleotide sequences of the proteins vary considerably between isolates of different subtypes. The NA is the second subtype-specific glycoprotein of the virion and is composed of a single polypeptide chain. The NA is not evenly distributed on the surface of the virion but is found in patches. The role of the NA in the life cycle of the virus is unclear. No posttranslational cleavage of the NA polypeptide occurs. The nucleotide sequences of different NA gene isolates varies conεiderably between εubtypeε (e.g.. A and B virus amino acid homology is 26-29%) . The NA gene of influenza B encodes two proteinε, NA and NB. The NA iε thought to be structurally and functionally similar to the type A NA. The NB protein is a glycoprotein of unknown function which is 100 amino acids in length.

The nucleoprotein (NP) is one of the type- specific antigens of influenza viruses that distinguisheε among the influenza type A, B, and C viruses. The NP is a multifunctional protein having a structural role in forming the nucleoprotein complex and a putative role in

transcription and replication. Genetic analysis of a large number of influenza strains has revealed that the NP genes can be placed into one of five different groupings. All avian strains fall within two groups, equine strains fall within two more groups and all human and swine strains form the final group. The restriction of certain species strains to these groups suggests that the NP gene may influence species-specificity or host range.

RNA segment 7 encodes the two M proteins (Ml and M2) . The mRNA encoding Ml is colinear with RNA segment 7, whereas M2 is encoded by a spliced mRNA. The two proteins share the same initiation codon for protein synthesis and the eight amino acid residues before the 5' splice junction of the M2 mRNA. The remaining 88 amino acids of M2 are encoded in the +1 reading frame from nucleotides 740-1104. This organization of RNA segment 7 is present in all influenza A and B viruses sequenced.

The Ml protein is a virion structural protein that is intimately associated with the lipid bilayer in close proximity to both glycoproteins and the ribonucleoprotein complex. It is also believed to have a role in the down-regulation of the virion transcriptase activity. Passively transferred monoclonal antibodies to this protein do not confer resistance to infection by influenza virus.

The M2 protein of influenza A is an integral membrane protein that iε expressed at the surface of infected cells. The M2 protein may be a virion associated protein with between 14 and 68 molecules per virion. Amantadine-resistant mutants of influenza virus contain mutations in the transmembrane domain of the M2 protein. Becauεe amantadine alters viral penetration into cells, it is likely that M2 is in the virion.

Comparison of RNA segment 7 sequences of the H3N2 (Udorn) and HlNl (PR8) strains show that the M protein coding sequences of these viruses (isolated 38 years apart) are highly conserved. Lamb, "The genes and

proteins of influenza viruses," in. Krug. ed. The Influenza Viruses NY, Plenum. 1989. Comparison of 230 nucleotides of RNA segment 7 from 5 human HlNl, H2N2 and H3N2 strains isolated over a 43 year period suggestε that the same segment 7 was retained throughout the antigenic shiftε of HA and NA. Hall and Air, 38 J. Virol. 1, 1981. Studieε have shown that RNA segment 8 of influenza A and B encodes two nonstructural _. proteins which are translated from separate mRNAs. NS1 and NS2 polypeptides of influenza A share 9 amino acids at their N termini, after which the NS2 mRNA has a 423 nucleotide deletion; then, the NS2 mRNA rejoins the NS1 3' region in the +1 reading frame. NS1 is synthesized in large amounts early in infection. NS2 is made only late in infection. Both proteins share a nuclear localization signal and can be found in the nuclei of infected cells. Large deletions occur in the carboxyl termini of the NS1 proteins of field isolates from humans or birds, which indicates that a high degree of variation can be tolerated in this polypeptide without affecting its function.

The three largest proteins of the virion (PB1, PB2 and PA) are found associated with NP and virion RNA and carry the polymerase activity which transcribes invading viral RNA. The PB1 and PB2 proteins form a complex when expressed in the absence of other virion proteins or RNA and are probably required for complementary RNA εynthesis. PA and NP are required for virion RNA synthesis. The PB1 gene of influenza B virus showε 61% homology with that of the influenza A viruε. Influenza virus produces an acute febrile infection of the respiratory tract characterized by abrupt onεet prominent myalgiaε, headache and cough. Pneumonia iε the most frequent complication; it may be primary viral (due to invasion of lung parenchyma) , secondary bacterial, or mixed viral and bacterial pneumonia. It may be severe and progreεsive or mild and segmental. Other complications which occur with less frequency include

Reye's syndrome, myocarditis, pericarditis, myositiε, encephalopathy and transverεe myelitiε. It has been estimated that the direct costs of influenza exceed $1 billion per year and may reach $3 to $5 billion. Total costs may be two to three times higher.

Two typeε of vaccineε are available for influenza. The "split" vaccines are chemically treated to reduce pyrogenic components and are the only type given to children under 13 years of age. The "whole" vaccine iε generally given to adultε. Protective antibody titerε are preεent in more than 90% of normal subjects after vaccination with influenza A antigens, but there is much lesε response to influenza B antigens. Additionally, elderly subjects and patients with renal failures or immunosuppression are at much greater risk to infection even with vaccination. The 70-80% efficacy of the vaccine is only observed when strain matches are good. Lower efficacy is observed when the match is not close, and when patients are immunocompromised, or in institutional situations in which viruε is readily transmitted.

Two drugs, amantadine and rimantadine, are as effective as influenza vaccine in preventing influenza A infections. Unfortunately, they are not as active against influenza B, which is responsible for 20% of all influenza epidemics and in a given year may be the only virus circulating. Amantadine is approved in the United States; rimantadine is not. Both drugs appear to impair the uncoating of viral RNA in infected cells by blocking the acidification process required to open the viral particles.

Resistance to amantadine and rimantadine is easily produced in the laboratory by serial passage of strains of influenza A virus in low concentrations of the drug, and such isolates are cross-resistant to both drugs. Drug resistant strains of influenza virus are able to initiate infection of cellε as effectively as their wild type progenitors. Resistance is associated with the

presence of point mutations in the RNA sequence coding for the M2 protein. This occurs most frequently at amino acid 31, but may also occur at positionε 27 to 34, which encompaεε the tranεmembrane domain of the protein. It has been hypothesized that M2 protein may act as an ion channel to facilitate the acidification of the virus particle, and that amantadine and rimantadine block this viral-envelope pore.

Many regions of the RNAε associated with influenza viruses are appropriate targets for ribozyme attack. A ribozyme targeted to these regions in influenza viruses may be useful for targeting an equivalent region in related viruses whose sequenceε are unknown.

There is a high degree of nucleotide sequence homology within the sequences of the human and non-human influenza virus genomes and mRNAs. Particularly useful regions include sequences in the 5' and 3' regions of the genomic RNA segments and complementary regions at the 3' ends of all influenza A and B virus mRNAs and their homologous positions in other genomes. In addition, the 3' sequence which contributes to the panhandle (CCUGCUUUUGCU) is highly conserved among all influenza virus isolates and contains two putative ribozyme cleavage sites in a relatively accessible region of the RNA. The panhandle sequence (AGUAGAAACAAGG) at the 5' ends of the RNAs also contains a suitable ribozyme target site. The complementary sequence (CCUUGUUUCUACU) of the 5' panhandle is present at the 3' terminus in all mRNAs and contains an additional 4 ribozyme target sites which occur within accessible regions of these molecules (i.e.. the loosely base-paired terminal regions) . Detailed information of the sequence of the eight RNA segmentε of influenza (εeven in influenza C εtrainε) and the molecular weight of the encoded proteinε is known. Preferred cleavage is at regions required for viral replication (e.g.. protein synthesiε, such as the regions in RNA segment 7 which regulate or encode the M

proteins) the conserved 5' and 3' regions of the genomic RNA segmentε which are involved in the panhandle structures, as well as the regions at the 3' ends of all mRNAs which are complementary to the 5' panhandle structure and their equivalent in other viruses. Alternative regions may also be used aε targetε for ribozyme-mediated cleavage of these viral genomes. Each target can be chosen as described below, e.g.. by a study of the secondary structure of the RNA, and the individual role of such RNA in the replication of the virus. If the targetε are contained within the open reading frameε of regions which encode proteins essential to the replication of the virus, then these other targets are preferred candidates for cleavage by ribozymes, and subsequent inhibition of viral replication. The regions encoding the influenza A virus PB1, PB2 and PA proteins and their homologous proteins in the other viruses are examples of such preferred targets.

The genome of influenza viruses may be subject to rapid genetic drift by virtue of its RNA content and the nature of errors in genomic replication. Those regions (genes) of the genome which are essential for virus replication, however, are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. Theεe regions are preferred target siteε in this invention since they are more likely to be conserved between different types or strains of influenza viruses, and thus only one ribozyme is needed to destroy all such viruses. Thus, one ribozyme may be used to target all influenza viruseε. We have selected several such RNA regions of theεe viruεes, and examined their nucleotide sequences for the presence of conserved areas which may be cleaved by ribozymes targeted to those regions. Three regions analyzed in detail are the 5' and 3' panhandle regions, the complementary 3' regions present in all mRNAs, and the mRNAs encoding the M proteins; other

genes can be analyzed in a manner similar to that described below.

The genomic sequenceε of the viruses are compared in their potential target regions. Putative ribozyme cleavage sites are found to be highly conserved between strains and species of viruε sequence. These sites represent the preferred sites for hammerhead or other ribozyme cleavage within these target RNAs.

Using such analyses, predictions of effective target sites in the viral genes, baεed upon computer generated sequence comparisons, were obtained. These are identified as SEQ. ID. NOS. 1-32, shown in Fig. 14. Example 10: HSV

Human herpesviruses cause a wide variety of diseases which result in significant levels of morbidity and mortality worldwide. The HSV group accounts for about one million new cases of infection each year in the United States. These infections are maintained for the lifetime of the patient as latent viral infections, which can be stimulated to reactivate by a variety of factors. The manifestations of HSV infection range from mild infectionε of herpes labialis to more serious infections such as herpes encephalitis.

HSV contains a double-stranded DNA genome within its central core, has a molecular.weight of approximately 100 million, and a genome encoding at least 70 polypeptides. The DNA core is surrounded by a capsid constructed from capsomers arranged in icosapentahedral symmetry. Tightly adherent to the capsid is the tegument, which appears to consiεt of amorphouε material. Looεely εurrounding the capsid and tegument is a lipid bilayer envelope containing polyamines, lipids, and the viral glycoproteinε. These glycoproteins confer distinctive properties to the virus and provide unique antigenε to which the host is capable of responding. Glycoprotein G (gG) , for example, confers antigenic specificity to HSV,

and therefore results in an antibody response that can be used to distinguish HSV-1 (gG-1) from HSV-2 (gG-2) .

Replication of HSV is a multi-step procesε. Following the onεet of infection, DNA iε uncoated and tranεported to the nucleuε of the hoεt cell. Tranεcription of immediate-early geneε encoding variouε regulatory proteinε followε. Expression of immediate- early gene products is then followed by the expresεion of proteinε encoded by early and then late geneε, including structural proteins as well as proteinε necessary for viral replication. Asεembly of the viral core and capsid takes place within the nucleus. This is followed by envelopment at the nuclear membrane and transport out of the nucleus through the endoplasmic reticulum and the Golgi apparatus, where viral envelope proteins are glycosylated. Mature virons are transported to the outer membrane of the hoεt cell, and release of progeny virus is accompanied by cell death. Replication for all herpesviruses is considered inefficient, with a high ratio of noninfectious to infectious viral particles.

The complete sequence of the HSV-1 genome is known. McGeoch et al., 69 J. Gen. Virol. 1531, 1988; McGeoch et al., 14 Nucleic Acidε Research 1727, 1986; and the elucidation of the HSV-2 genome sequence is underway in laboratories worldwide. The two εubtypeε of HSV, HSV-1 and HSV-2, are 60-80% homologouε at the DNA level, but intragenic variation, where known, is less.

Antiviral drugs including acyclovir have been used to effectively treat HSV infections, although with limited succesε. For example, chronic treatment with acyclovir haε reεulted in the development of acyclovir- reεiεtant strains. Nucleoside analogueε, εuch as acycloguanosine and tri-fluorothymidine are currently used for treatment of mucosal and ocular HSV infections, but theεe compoundε have little if any effect upon recurrent or secondary infections (which are becoming more prevalent as the number of HIV-immunosuppreεεed patients rises) . In

addition, nucleoside analogues are poorly soluble in aqueous solutions, are rapidly catabolized intracellularly, and can be extremely toxic.

Ribozymes of this invention exhibit a high degree of specificity for only the virally encoded mRNA in infected cellε. Ribozyme moleculeε targeted to highly conεerved sequence regions will allow the treatment of many species or subtypes of HSV with a single compound. There iε no acceptable treatment which will give a broad spectrum of activity with no toxic side effects. No treatment exiεtε which specifically attacks viral gene expression which is responsible for the transformation of epithelial cells by HSV, for the maintenance of the episomal genome in latently infected cells or for the vegetative replication of the virus in permisεive cells.

The methods of this invention can be used to treat HSV infections, which includes these diseases noted above. The utility can be extended to other HSV-like virus which infect non-human primates where such infections are of veterinary importance.

The ribozymes of the invention are capable of specifically cleaving particular viral mRNA targetε, resulting in the destruction of mRNA transcript integrity required for translation, and therefore preventing the syntheεis of the encoded protein. More specifically, the ribozymes of the invention are targeted to and prevent the translation of mRNAε encoding proteins required for viral genomic replication, virion structure, and viral infectivity, maintenance of the latent state, etc., and therefore interfere with critical events required for viral survival. Thus, diseaεeε cauεed by HSV may be effectively treated by ribozyme-mediated interruption of the viral life-cycle.

Preferred cleavage εiteε are at genes required for viral replication, e.g.. protein synthesis, such as in the immediate early geneε (ICPO, ICP4, ICP22 and ICP27) , genes required for nucleic acid metabolism (UL13, 39, 40,

50) , host shut-off (UL41) , control of late viral protein synthesis (7 34.5) , DNA replication (UL5, 8, 9, 29, 30, 42, 53) and structural protein encoding.genes (gB and gC) . Alternative regions make suitable targets of ribozyme-mediated inhibition of HSV replication. Most preferred targetε include ICP4(IE3) , ICP27(UL54), UL39, UL40, UL5, 7 34.5 and UL27 (gB) geneε. Below are provided exampleε of ribozymes targeted to the ICP4 gene but other such ribozymes are expected to have utility at these other geneε.

In particularly preferred embodimentε, the RNA which iε cleaved iε HSV ICP4 or UL5 mRNA regionε (for UL5, nucleotide #1 iε defined by preliminary mapping of the cap site in our laboratory and nucleotide 2800 is the site of translation initiation for the UL4 protein whose open reading frame is also included in the UL5 mRNA) selected from one or more of the sequences shown in Fig. 15.

Those regions (genes) of the genome which are essential for HSV replication are expected to maintain a constant sequence (i.e.. are conserved) over extensive periods of time. These regions are preferred target sites in this invention since they are more likely to be conserved between different types or strains of HSV, and thus only one ribozyme is needed to destroy all viral RNA. Thus, one ribozyme may be used to target all HSV-encoded RNA. We have selected several such genes of these viruseε, and analyzed the structure of the encoded mRNAs using RNA-fold computer analyses. We have identified open structureε in the RNAs which may be cleaved by ribozymes targeted to thoεe regionε. Two regionε analyzed in detail are the HSV-1 ICP4 and UL5 mRNAε; other geneε can be analyzed in a manner similar to that described below.

HSV specific ribozymes may be designed to also inhibit other herpesviruseε, particularly varicella zoster virus (VZV) , which exhibits significant intragenic nucleotide homology with a number of analogous HSV genes. For example, the VZV immediate-early protein IE62 exhibits

considerable amino acid homology to the HSV immediate- early regulatory protein ICP4/IE3 and is a major component of VZV particles (Kinchington et al., 66 J. Virol. 359, 1992) . Also, some transcription units and their encoded proteins are conserved within the genomeε of evolutionarily divergent human herpeεviruεeε, and ribozyme targets within these genetic units may therefore provide the basis for designing ribozymeε with broad therapeutic applicationε. The mRNA εequences of the viruses are analyzed throughout the regions discuεεed above (using the nucleotide sequence of the HSV-1 ICP4 gene (EMBL No. HEHSV1G3, nucleotideε 1176-5435) and UL5 gene). Putative ribozyme cleavage sites are found to be weak or non-base paired regions of the viruε mRNA. Theεe sites represent the preferred sites for hammerhead or other ribozyme cleavage within these target RNAs.

Using such analyses, predictions of effective target sites in the viral mRNAs, based upon computer generated structural analyses, were obtained (see SEQ. ID. NOS. 1-115) . The target region is listed with the nucleotide number (in the HSV-1 ICP4 or UL5 mRNA) noted as baεe number.

In each of the above examples, short RNA substrates corresponding to each of the gene sites are designed. Each substrate is composed of two to three nucleotides at the 5' and 3' ends that will not base pair with a corresponding ribozyme recognition region. The unpaired regions flank a central region of 12-14 nucleotides to which complementary arms in the ribozyme are designed.

The structure of each substrate sequence iε predicted uεing a εtandard commercially available PC fold computer program. Sequences which give a positive free energy of binding are accepted. Sequences which give a negative free energy are modified by trimming one or two bases from each of the ends. If the modified sequenceε

are still predicted to have a strong secondary structure, they are rejected.

After substrates are chosen, ribozymeε are designed to each of the RNA substrateε. Ribozyme folding iε alεo analyzed uεing PC fold.

Ribozyme moleculeε are εought which form hammerhead motif stem II (see Fig. 1) regions and contain flanking arms which are devoid of intramolecular base pairing. Often the ribozymes are modified by trimming a baεe from the endε of the ribozyme, or by introducing additional base pairs in stem II to achieve the desired fold. Ribozymes with incorrect folding are rejected. After substrate/ribozyme pairs are found to contain correct intramolecular structures, the molecules are folded together to predict intermolecular interactions. A schematic representation of a ribozyme with its coordinate base pairing to its cognate target sequence is shown in Fig. 1.

Those targets thought to be useful as ribozyme targetε can be tested to determine accessibility to nucleic acid probes in a ribonuclease H assay (see below) . This assay provides a quick test of the use of the target site without requiring synthesis of a ribozyme. It can be used to screen for sites most suited for ribozyme attack. Synthesis of Ribozymes

Ribozymes useful in this invention can be produced by gene transcription as described by Cech, supra , or by chemical synthesis. Chemical synthesis of RNA is similar to that for DNA synthesis. The additional 2'-OH group in RNA, however, requires a different protecting group strategy to deal with selective 3'-5' internucleotide bond formation, and with RNA susceptibility to degradation in the presence of bases. The recently developed method of RNA synthesis utilizing the t-butyldimethylsilyl group for the protection of the 2' hydroxyl is the most reliable method for synthesiε of ribozymes. The method reproducibly yields RNA with the

correct 3 '-5' internucleotide linkages, with average coupling yields in excess of 99%, and requires only a two- step deprotection of the polymer.

A method based on H-phoεphonate chemiεtry exhibits a relatively lower coupling efficiency than a method baεed upon the phoεphoramidite chemiεtry. Thiε is a problem for synthesis of DNA as well. A promiεing approach to εcale-up of automatic oligonucleotide synthesis has been described recently for the H- phosphonates. A combination of a proper coupling time and additional capping of "failure" sequences gave high yields in the synthesiε of oligodeoxynucleotides in scales in the range of 14 μmoles with aε little as 2 equivalents of a monomer in the coupling step. Another alternative approach is to use soluble polymeric supportε (e.g.. polyethylene glycolε) , inεtead of the conventional solid supports. This method can yield short oligonucleotides in hundred milligram quantities per batch utilizing about 3 equivalents of a monomer in a coupling step. Various modifications to ribozyme structure can be made to enhance the utility of ribozymes. Such modifications will enhance εhelf-life, half-life in vitro , stability, and ease of introduction of such ribozymes to the target site, e.g.. to enhance penetration of cellular membranes, and confer the ability to recognize and bind to targeted cells.

Exogenous delivery of ribozymes benefits from chemical modification of the backbone, e.g.. by the overall negative charge of the ribozyme molecule being reduced to facilitate diffusion across the cell membrane.

The present strategies for reducing the oligonucleotide charge include: modification of internucleotide linkages by ethylphosphonates or methylphosphonates, use of phosphoramidites, linking oligonucleotides to positively charged molecules, and creating complex packages composed of oligonucleotideε, lipidε and εpecific receptorε or effectorε for targeted cells. Examples of such

modifications include εulfur-containing ribozymes containing phosphorothioates and phosphorodithioates aε internucleotide linkages in RNA. Synthesis of such sulfur-modified ribozymes is achieved by use of the sulfur-transfer reagent, ³ H-1,2-benzenedithiol-3-one 1,1- dioxide. Ribozymes may also contain ribose modified ribonucleotideε. Pyrimidine analogueε are prepared from uridine using a procedure employing diethylamino sulphur trifluoride (DAST) as a starting reagent. Ribozymes can alεo be either electroεtatically or covalently attached to polymeric cationε for the purpose of reducing charge. The polymer can be attached to the ribozyme by simply converting the 3'-end to a ribonucleoside dialdehyde which iε obtained by a periodate cleavage of the terminal 2',3'- ciε diol system. Depending on the specific requirements for delivery systems, other possible modifications may include different linker arms containing carboxyl, amino or thiol functionalities. Yet further examples include use of methylphoεphonates and 2'-0-methylribose and 5' or 3' capping or blocking with m ₇ GpppG or m ₃ ² ' ^2,7 GpppG.

For example, a kinased ribozyme is contacted with guanosine triphosphate and guanyltransferase to add a m ³ G cap to the ribozyme. After such synthesis, the ribozyme can be gel purified using standard procedure. To ensure that the ribozyme has the desired activity, it may be tested with and without the 5' cap using standard procedures to assay both its enzymatic activity and its stability.

Synthetic ribozymeε, including thoεe containing variouε modifierε, can be purified by high preεεure liquid chromatography (HPLC) . Other liquid chromatography techniqueε, employing reverεe phase columns and anion exchangerε on εilica and polymeric supports may also be used. Expression Vector

While synthetic ribozymeε are preferred in thiε invention, those produced by expresεion vectorε can alεo

be used. In designing a suitable ribozyme expression vector the following factors are important to consider. The final ribozyme must be kept as small as possible to minimize unwanted secondary structure within the ribozyme. A promoter (e.g.. the human cytomegalovirus immediate early region promoter or keratin promoters from human keratin genes) should be chosen to be a relatively strong promoter, and expressible both in vitro and in vivo . Such a promoter should expresε the ribozyme at a level εuitable to effect production of enough ribozyme to destroy a target RNA, but not at too high a level to prevent other cellular activities from occurring (unleεε cell death itεelf is desired) .

A hairpin at the 5' end of the ribozyme iε uεeful to protect the ribozyme from 5'-3' exonucleases. A selected hairpin at the 3' end of the ribozyme is useful since it acts as a protection from 3'-5' exonucleases. The T7 RNA polymerase termination signal is about 30 nucleotides in length (Fig. 3ID) and may serve as a good hairpin structure to stabilize ribozymes from exonuclease digestion, although other 3' hairpins of shorter length can be used to provide RNA stability. Such hairpins can be inserted within the vector sequences to allow standard ribozymes to be placed in an appropriate orientation and expressed with such sequences attached.

Poly(A) tails are also useful to protect the 3' end of the ribozyme. These can be provided by either including a poly(A) signal site in the expression vector (to signal a cell to add the poly(A) tail in vivo) , or by introducing a poly(A) sequence directly into the expression vector. In the first approach, the signal must be located to prevent unwanted εecondary structure formation with other parts of the ribozyme. In the second approach, the poly(A) stretch may reduce in size over time when expressed in vivo , and thus the vector may need to be checked over time. Care must be taken in addition of a poly(A) tail which binds poly(A) binding proteins which

prevent the ribozyme from acting upon their target sequence.

High Yield Ribozyme Production

There follows a description of a method for expresεion of a ribozyme in a bacterial or eucaryotic cell, and for production of large amountε of such a ribozyme. In general, the method involves preparing multi-copy cassettes encoding a defined ribozyme structure for production of a ribozyme at a decreased cost. A vector is produced which encodes a plurality of ribozymes which are cleaved from an RNA transcript produced from the vector by another ribozyme. The system is useful for scaling up production of a ribozyme, which may be either modified or unmodified, in situ or in vitro . Such vector systems can be used to express a desired ribozyme in a specific cell, or can be used in an in vitro system to allow production of large amounts of a desired ribozyme. The vectors allow a higher yield synthesis of a ribozyme in the form of an RNA transcript which is cleaved in situ or in vitro before or after transcript isolation.

The stability of the ribozyme produced in this method can be enhanced by proviεion of sequences at the termini of the ribozymes as deεcribed herein.

The method iε advantageouε since it provides high yield synthesiε of ribozymes by use of low cost transcription-baεed protocolε, compared to exiεting chemical ribozyme εynthesiε, and can uεe iεolation techniqueε currently uεed to purify chemically εynthesized oligonucleotides. Thus, the method allows synthesis of ribozymes in high yield at low cost for analytical, diagnoεtic, or therapeutic applications.

The method is also useful for syntheεis of ribozymes in vitro for ribozyme structural studies, enzymatic studies, target RNA accessibility studies, transcription inhibition studies and nuclease protection studies, much is described herein.

The method can also be used to produce ribozymes in situ either to increase the intracellular concentration of a desired therapeutic ribozyme, . ^' • or to produce a concatameric transcript for subεequent in ^' vitro iεolation of unit length ribozyme. The deεired ribozyme can be uεed to inhibit gene expreεεion in molecular genetic analyεes or in infectious cell syεtemε, and to teεt the efficacy of a therapeutic molecule or treat afflicted, cellε.

Thuε, in general, the vector includes a bacterial, viral or eucaryotic promoter within a plasmid, cosmid, phagmid, virus, viroid or phage vector. Other vectors are equally εuitable and include double-stranded, or partially double-stranded DNA, formed by an amplification method such as the polymerase chain reaction, or double-stranded, partially double-stranded or single-stranded RNA, formed by site-directed homologous recombination into viral or viroid RNA genomes. Such vectors need not be circular. Transcriptionally linked to the promoter region is a first ribozyme-encoding region, and nucleotide sequences encoding a ribozyme cleavage sequence which is placed on either side of a region encoding a therapeutic or otherwise desired second ribozyme. Suitable restriction endonuclease siteε can be provided to ease construction of this vector in DNA vectors or in requisite DNA vectors of an RNA expression system. The desired second ribozyme may be any deεired type of ribozyme, εuch aε a hammerhead, hairpin or other catalytic center, and can include group I and group II intronε, aε discussed above. The first ribozyme is chosen to cleave the encoded cleavage sequence, and may alεo be any deεired ribozyme, for example, a hepatitis delta viruε sequence, a Tetrahymena derived ribozyme, which may, for example, include an imbedded restriction endonuclease site in the center of a self-recognition sequence to aid in vector construction. This endonuclease site is useful for construction of the vector, and subεequent analysis of the vector.

When the promoter of such a vector is activated an RNA transcript is produced which includes the first and second ribozyme sequences. The first ribozyme sequence iε able to act, under appropriate conditions, to cause cleavage at the cleavage sites to release the second ribozyme sequenceε. Theεe εecond ribozyme sequences can then act at their target RNA sites, or can be isolated for later use or analysiε.

Thus, the vector includes a first nucleic acid sequence (encoding a first ribozyme having intramolecular cleaving activity) , and a second nucleic acid sequence (encoding a second ribozyme having intermolecular cleaving enzymatic activity) flanked by nucleic acid sequences encoding RNA which is cleaved by the first ribozyme to release the second ribozyme from the RNA transcript encoded by the vector. If deεired, the first ribozyme may be encoded on a separate vector and may have intermolecular cleaving activity.

As discussed above, the first ribozyme can be chosen to be any self-cleaving ribozyme, and the second ribozyme may be chosen to be any desired ribozyme. The flanking sequences are chosen to include sequences recognized by the first ribozyme. When the vector is caused to express RNA from these nucleic acid sequences, that RNA has the ability under appropriate conditionε to cleave each of the flanking regionε and thereby release one or more copies of the second ribozyme. If desired, several different second ribozymes can be produced by the same vector, or several different vectorε can be placed in the εame vessel or cell to produce different ribozymes.

Preferably, the vector includeε a plurality of the nucleic acid εequenceε encoding the second ribozyme, each flanked by nucleic acid sequenceε recognized by the firεt ribozyme. Moεt preferably, εuch a plurality includeε at leaεt εix to nine nucleic acid sequences. In other preferred embodiments, the vector includes a promoter which regulates expreεεion of the nucleic acid

encoding the ribozymes from the vector; and the vector is chosen from a plasmid, cos id, phagmid, virus, viroid or phage. More preferably, the plurality of nucleic acid sequenceε are identical and are arranged in sequential order εuch that each haε an identical end nearest to the promoter. If desired, a poly(A) εequence adjacent to the εequence encoding the first or second ribozyme may be provided to increase stability of the RNA produced by the vector; and a restriction endonuclease site adjacent to the nucleic acid encoding the first ribozyme iε provided to allow inεertion of nucleic acid encoding the second ribozyme during construction of the vector.

A method for formation of a ribozyme expression vector includes providing a vector including nucleic acid encoding a first ribozyme, as discussed above, and providing a single-stranded DNA encoding a second ribozyme, as discussed above. The single-stranded DNA is then allowed to anneal to form a partial duplex DNA which can be filled in by a treatment with an appropriate enzyme, such as a DNA polymeraεe in the presence of dNTPs, to form a duplex DNA which can then be ligated to the vector. Large vectors resulting from this method can then be selected to insure that a high copy number of the single-stranded DNA encoding the second ribozyme is incorporated into the vector.

A method for production of ribozymes involveε providing a vector aε described above, expresεing RNA from that vector, and allowing cleavage by the firεt ribozyme to releaεe the εecond ribozyme. Specifically, referring to Fig. 16, a vector 10 iε provided with a promoter 12 which regulates expreεεion of a downεtream nucleic acid region 14 encoding a firεt ribozyme, followed by DNA encoding an RNA cleavage εequence 16 which iε recognized by the firεt ribozyme and can be cleaved by that ribozyme in the corresponding RNA at the location εhown by the daεhed line 18. Several εuch εequenceε 16 are provided in the vector, each of which can

be cleaved at the line marked 18. Cleavage at line 18 causes release of a second ribozyme encoded by DNA 20 provided between each of the repeated sequences 16.

Referring to Fig. 17, in a preferred embodiment the repeated regions 20 are provided with- identical endε, each of which has the same end (shown as b in the figure) nearest to the promoter 12. In this way,. each unit has a sequence a/b, the corresponding RNA of which can be recognized and cleaved at line 18 by the first ribozyme. Referring to Fig. 18, the vector, may be provided with a promoter (PRO) followed by a first releasing ribozyme (RR) and then a restriction endonuclease site (RE) . Optionally, there may be provided a poly(A) εequence (AAA) which will provide εtability to the releasing ribozyme. Repeating structureε are then provided with flanking regionε (1 and 2) εurrounding a εecond therapeutic ribozyme (TR) .

The actual εequence of the deεired second ribozyme unit will depend upon the target it will recognize and cleave. This in turn will have some impact on the acceptable sequence of the cleavage site 18 chosen, as well aε the reεtriction endonucleaεe εite (RE) chosen. In all cases, minimization of any secondary interactions between regions of the RNA transcript encoded by vector 10 will be sought, and other intermediate sequences may be required to space the active components at critical distances to allow the necesεary folding of the RNA. An alternative approach includeε provision of a second transcription site within the vector, or within another vector, to synthesize other RNA with ribozyme activity which can cause appropriate cleavage of the transcript of a vector of this invention.

The actual cleavage of RNA to free the active desired therapeutic ribozyme may occur within a bacterial or eucaryotic cell in which the vector is grown, or can be accomplished after isolation of the transcript RNAs by addition of critical salt and/or other requisite co-

factors. Iεolation of unit length therapeutic ribozyme from the RNA mixture can be achieved by a εequence εpecific affinity chromatography, or charge or other componentε of the ribozyme may be used in other suitable isolation techniques. Example 11:

A preferred arrangement of sequences in the vector of this invention is to have the cleaving ribozyme- encoding region 5' in the vector, and the therapeutic second ribozyme-encoding region 3' in the vector (as shown in Figs. 16-18) . Each sequence encoding a εecond ribozyme iε εurrounded by flanking sequences encoding cleavage siteε. Any number of such flanking sequences can be added. Other flanking structures, such as the poly(A) sequence shown in Fig. 18, can be used and flexibility may be increased by provision of useful restriction endonuclease sites, such as AU rich restriction endonuclease sites, which would give regions convenient for cloning that would not have extensive secondary structure at physiological temperatures in cellε or in vitro extracts. The minimization of extensive RNA structure in regions flanking the therapeutic ribozyme are essential to optimizing the catalytic activity of the ribozyme. It is possible, for example, to provide the desired second ribozyme encoding sequence flanked by a chosen restriction endonuclease site. DNA encoding such a sequence can then be readily ligated to other such sequenceε to provide the plurality of repeated sequences desired in a DNA construct. In this sort of procedure, flanking inserts with two different reεtriction endonucleaεe εiteε make the construction of the vector easier.

For example, HindiII/SalI fragments encoding a releasing ribozyme for a desired ribozyme (VAHR36) are cloned into a plasmid pTZ19U (equivalent vectors are known in the art) . The sequenceε of these fragments are:

5'-CCCAAGCTTGTCGCCTGATGAGGTCCGAAAGGA Primer RRHS 3 '-CTCCAGGCTTTCCTGGCTTTCTGGCAGCTGCCC RRSH

Theεe will give a clone fragment after restriction endonuclease treatment, and insertion into the vector, having the following sequence:

AGCTTGTCGCCTGATGAGGTCCGAAAGGACCGAAAGACCGTCGA tfindlll Sail

\ / releasing ribozyme

This segment encodes a releasing first ribozyme εurrounded by HindiII/SalI reεtriction endonucleaεe εiteε. Thiε releaεing ribozyme will cleave flanking sequenceε in a concatamer having the following εequence:

ACGGUCUCGCGUCU tcleaved base

The desired second ribozyme, VAHR36, has the following sequence:

GCGUCUCUGAUGAGGUCCGAAAGGACCGAAACGGUCU [C] Flanking Flanking t Region Region extra base after concatamer cleavage

which has the two flanking sequences GCGUCU and ACGGUCU. Two primers used to create a DNA insert encoding thiε εequence are εyntheεized having the following εequenceε:

ACGGTCTCGCGTCTCTCATGAGGTCCGAAAGGACCGAAACGGTCT (primer TRS (Senεe) ) rG1AGACCGTTTCGGTCCTTTCGGTCCTCATCAGAGACGCGAGACCGT (primer TRC (Complement) )

These two primers can be uεed for concatamer formation as shown in Fig. 18. Hybridization of the primers, syntheεis

of the second DNA strand to make double-stranded inserts and insertion of these fragments into a suitable vector can be accomplished by methods known to others in the art. If desired, the sequence GGGGTCG can be added to the 5' end of the TRS primer to allow cloning into the Sail site, and GGGGGTACC added to the 5' end of the TRC primer to allow cloning into a BamHl site of pTZ19U or other suitable vector. These primers can be combined as discuεεed above to allow synthesis of the desired ribozyme vector. Similar constructε can be obtained for RNA vectorε by utilizing theεe methodε to make expreεεion vectorε, which will be used to syntheεize the appropriate RNA or DNA molecules for homologous recombination of ribozyme sequences into the viral, viroid or phage vectors. If delivery of the constructed vector to a eucaryotic cell is contemplated, cellular splicing machinery within the eucaryotic cell can be incorporated to cleave out the multimeric therapeutic second ribozyme by encoding recognition sequenceε for the ribozyme εegmentε within the flanking εequenceε of the expressed transcript. Multiple copies of the releasing first ribozyme may be provided to enhance release of the desired or therapeutic second ribozyme if turnover rate is slower than the degradation rate of the therapeutic second ribozyme. Thus, for these vectors cytoplasmic signalε or poly(A) tails may be provided to enhance transport to an appropriate cellular compartment. In bacterial cellε, in vitro modificationε and certain cell modifications can be enhanced by provision of appropriate sequenceε within the vector, and are uεeful to enhance ribozyme turnover rate, cleavage activity, and nucleaεe εtability. Example 12:

Referring to Fig. 19, a preferred method for formation of a vector of the preεent invention iε to provide a promoter εequence linked to a releaεing firεt ribozyme (RR) with an appropriate reεtriction endonucleaεe εite (RE) . A single-stranded oligonucleotide is then

provided which encodes the two flanking regions and a therapeutic second ribozyme. Oligonucleotides such as those designed in Example 1 are allowed to form partial duplexes by hybridization at the flanking regions. The single-stranded regions are then filled in, using a DNA polymerase and dNTPs, to form a double-stranded DNA which can then be ligated to the restriction endonuclease site to form the desired vector.

The preferred method of isolating therapeutic ribozyme is by a chromatographic technique. The HPLC purification methods and reverse HPLC purification methods described herein can be used. Alternatively, the attachment of complementary oligonucleotides to cellulose or other chromatography columns allows isolation of the therapeutic second ribozyme, for example, by hybridization to the region between the flanking arms and the enzymatic RNA. This hybridization will select against the short flanking sequences without the desired enzymatic RNA, and against the releasing first ribozyme. The hybridization can be accomplished in the presence of a chaotropic agent to prevent nuclease degradation. The oligonucleotides on the matrix can be modified to minimize nuclease activity, for example, by provision of 2'-0-methyl RNA oligonucleotides. Such modifications of the oligonucleotide attached to the column matrix will allow the multiple use of the column with minimal oligo degradation. Many such modifications are known in the art, but a chemically εtable non-reducible modification iε preferred. For example, phoεphorothioate modificationε can alεo be uεed.

The expressed ribozyme RNA can be isolated from bacterial or eucaryotic cellε by routine procedureε such as lysis followed by guanidine isothiocyanate isolation. Example 13 : The current known self-cleaving site of

Tetrahymena can be used in an alternative vector of this invention. If desired, the full-length Tetrahymena

sequence may be used, or a shorter sequence may be used. It iε preferred that, in order to decrease the superfluouε sequences in the εelf-cleaving εite at the 5' cleavage end, the hairpin normally preεent in the Tetrahymena ribozyme εhould contain the therapeutic εecond ribozyme 3' sequence and itε complement. That iε, the first releasing ribozyme-encoding DNA is provided in two portions, separated by DNA encoding the desired second ribozyme. For example, if the therapeutic second ribozyme recognition sequence iε CGGACGA/CGAGGA, then CGAGGA iε provided in the εelf-cleaving site loop such that it is in a stem structure recognized by the Tetrahymena ribozyme. The loop of the stem may include a restriction endonuclease site into which the desired second ribozyme- encoding DNA is placed.

If desired, the vector may be used in a therapeutic protocol by use of the systems described by Lechner, "Regulation of Nucleic Acid Translation", International Patent Publication WO 92/13070, to allow a timed expression of the therapeutic second ribozyme, as well as an appropriate shut off of cell or gene function. Thus, the vector will include a promoter which appropriately expresses enzymatically active RNA only in the presence of an RNA or another molecule which indicates the preεence of an undeεired organiεm or state. Such enzymatically active RNA will then kill or harm the cell in which it exists, aε described by Lechner id . , or act to cause reduced expression of a desired protein product.

A number of suitable RNA vectors may alεo be uεed in thiε invention. The vectors include plant viroidε, plant viruεeε which contain single or double- stranded RNA genomes and animal viruseε which contain RNA genomes, such as the picornaviruses, myxoviruses, paramyxoviruεeε, hepatitiε A viruε, reovirus and retroviruses. In many instances cited, use of these viral vectors also results in tisεue specific delivery of the ribozymes.

Cloning of a Concatameric Ribozyme Construct The following iε one specific example showing the cloning of multiple therapeutic ribozymes (TRs) within a transcription unit. The release of the TRε from the tranεcript neceεεitateε the cloning of an additional releasing ribozyme (RR) which can cleave between TR units in the nascent transcripts.

To test the concatameric ribozyme vector construction, we chose a ribozyme targeted to the HSV-1 ICP4 mRNA. This ribozyme (RPI 1197, same aε VAHR36) was chosen because it haε been extensively characterized enzymatically using both εhort and long RNA εubstrates (K _cal /K _m = 4 x lθ ⁷ and 3 x 10 ⁴ mole ^"1 /min ^"1 , reεpectively) and has been used in tiεεue culture εtudieε to inhibit the replication of herpes simplex virus. These data afford uε important catalytic and therapeutic values to which we can compare ribozyme activities observed with a concatameric vector.

After choosing the TR, RR must be deεigned which iε capable of cleaving tail-to-head concatamers of the TR. As shown in Fig. 48, the sequence of RPI 1197 made design of a RR very easy because insertion of one nucleotide (A) between the concatamers gave a good ribozyme target site. We have found that extension of the armε of hammerhead ribozymeε by one or two nucleotideε does not reduce the catalytic activity significantly (usually <30%) . Thiε allowε some flexibility in designing the TR monomers to accommodate extra nucleotideε which optimize the εequence of the RR cleavage εite. If the εequence cannot be manipulated uεing nucleotideε preεent in the gene, it iε preferable to place a UM sequence at the 3' end of the TR, where M = C, U or A. This modification of the sequence will give a UM trinucleotide of unpaired bases when the TR hybridizes to target RNA. Unpaired nucleotides at the 3' termini of hammerhead ribozymes usually reduce catalytic activity only when they force the ribozyme into altered structureε which are more

stable than the enzyme-substrate complex. A preliminary check to detect such unfavorable structures can be performed using computer-based RNA folding algorithms.

If additional sequenceε εuch ^' aε cytoplaεmic localization εignals or polyadenylation signals are desired in the TR, these sequences should be considered when designing the RR. Our cloning protocol may be used to incorporate these structures at either terminus of the TR, but we have found that many additional sequences interfere with proper folding of the TR. Exceptions to this observation are the sequenceε which have εtrong internal baεe-pairing, εuch aε hairpin loopε or tRNA motifε.

After deciding upon the sequence of the RR motif, it is preferable but not necessary to determine if the RR will cleave the tail-to-head concatameric TR junction. This may be accomplished by synthesizing substrate and ribozyme RNA, as described previously, then performing cleavage reactions in the presence of 10 mM Mg ⁺² . After verifying the catalytic activity of the RR, the RR motif is cloned into an appropriate vector which contains multiple cloning siteε 3' of a phage promoter. For thiε cloning step, we chose to use PTZ19R (USB) . By uεing a directional cloning protocol, the screening required consiεtε of reεtriction digeεtε of the iεolated plaεmid DNA. Without a directional protocol, the εmall εize of the RR inserts would necessitate sequencing of the plasmid DNA to chooεe correct clones, except in the rare caseε where the stem I or stem III regions (see Fig. 48) contain a convenient, diagnostic restriction endonuclease recognition site.

Cloning of the RR fragment into PTZ19R used the t w o o l i g o n u c l e o t i d e ε , 5 ' - CCCAAGCTTGTCGCCTGATGAGGTCCGAAAGGA-3 ' and 5'- CCCGTCGACGGTCTTTCGGTCCTTTCGGACCTC-3' , which will create the RR coding εequence flanked by Hindlll and Sail reεtriction endonucleaεe εites (bold faced type) at the 5'

and 3' termini, respectively. Four microgram samples of each oligonucleotide were added to a reaction mixture containing 4 μl of lOx SEQUENASE buffer, 4 μl of 2.5 mM dNTP, 4 μl of 0.1 M DTT, and 1.5 μl of SEQUENASE polymerase (20U, USB) in a total volume of 40 μl. After the mixture was incubated at 37°C for 1 hour, a 5 μl aliquot was removed and added to 5 μl of loading buffer prior to electrophoresis in 10% polyacrylamide gels. Bands were visualized by staining with Stains-all (USB) . The remainder of the sample was frozen at -20°C. After verification that the appropriate fragment was generated, 5 μl of the frozen mixture was added to a 15 μl mixture containing 10 U each of Sail and Hindlll restriction endonucleases, 1.3 x restriction endonuclease buffer, and 10 U of shrimp alkaline phosphataεe (USB) and the final mixture waε incubated at 37°C for 1 hour before phenol/ chloroform/isoamyl alcohol extraction and precipitation of the DNA with ethanol. The pellet was dissolved in water and 30 ng of the insert DNA fragment was added to 300 ng of PTZ19R DNA fragment which had been digested with findlll and Sail enzymes then gel isolated in low melting point agarose. This plasmid (PTZ19R) :insert concentration gave an approximate molar ratio of 1:10, respectively. The DNA samples were added to ligation buffer and ligated with T4 DNA ligase (USB) at room temperature for 2 hours, then at 4°C for 18 hours. One-half of the ligation mixture (8 μl) was transfected into 100 μl of competent DH50.-.E:. coli cellε (GIBCO/BRL) and plated onto LB agar plateε containing ampicillin. Six cloneε were picked from the plateε and grown in 10 ml overnight preparationε for the extraction and reεtriction enzyme digeεtion of the plasmid DNA. The insertion of the RR DNA was verified by the increaεe in plaεmid DNA εize when digested with either Hindlll or Sail enzymes and by the disappearance of the Pstl site which was replaced in the PTZ19R vector with the RR coding DNA. A 250 ml preparation of bacteria transformed with the new RR-containing plasmid was grown

and the plasmid DNA isolated was digested with Sail and -BamHl restriction enzymes prior to use as the vector for subsequent cloning of the concatameric TR inserts.

Four oligonucleotides were synthesized for use in the TR concatamer cloning. The oligoε were:

TRS 5'GCGTCTCTGATGAGGTCCGAAAGGACCGAAACGGTCTA

TRC 5'CATCAGAGACGCTAGACCGTTTCGG

TRSS 5'GGGTCGACGCGTCTCTGATGAGGTCCGAAAGGACCGAAACGGTCTA

TRCB 5'GGGGATCCCATCAGAGACGCTAGACCGTTTCGG

where the bold faced sequences in TRSS and TRCB are the restriction endonuclease recognition sequenceε for the Sail and BamHl enzymeε, respectively. The oligonucleotides, in a 5:1:0.1:0.1 molar ratio of TRS (5 μg) :TRC:TRSS:TRCB, respectively, were added to a mixture containing 15 U of SEQUENASE polymerase and lx SEQUENASE buffer in a total volume of 15 μl and incubated at 37°C for 2 hours. To optimize the fill-in reaction, another 10 U of SEQUENASE polymerase was added after the first hour of incubation. The mixture waε diluted to 150 μl and deεalted by εpin chromatography through G-50 SEPHADEX columns and reconcentrated to a volume of approximately 15 μl by lyophilization in a Speed-Vac. Nine microliters of the DNA mixture was transferred to another tube containing 1 μl of lOx ligase buffer and 2 U of T4 DNA ligaεe and incubated at 37°C for 2 hourε to ligate gapped DNA pieceε. Two microliters of lOx Sail digestion buffer, 1 μl (5U) each of Sail and BamHl enzyme and 4 μl of water were added and the mixture was incubated at 37°C for 3 hours. After the reεtriction digestion, the mixture was εubjected to electrophoresis in 0.8% agaroεe gel. The electrophoreεis was performed until the bromphenol blue marker waε 1.5 incheε into the gel. Thiε εeparated the cleaved end fragments from the concatameric DNA fragments. After electrophoresis the gel lane containing the DNA was cut from the area just above the bromphenol blue dye band to

the area approximately 1 mm below the bottom of the gel loading area. The DNA was electroeluted from the lane and concentrated by precipitation with ethanol. This protocol gives a variety of concatameric sizes which can also be size εelected from the agarose gel, if desired. Alternatively, the insert size can be selected after the concatamers are cloned into the RR vector DNA.

Concatameric DNA was cloned into the RR- containing vector deεcribed above by mixing the eluted concatameric DNA with a gel-purified fragment of the RR vector which had been digeεted previouεly with BamHl and Sail reεtriction enzymeε. The molar ratio of the vector DNA to insert DNA (as calculated from the original TR concentrations) used in our experiments was 1:100. Thiε gives approximately a 1:2 ratio if the concatamer size averages 2.3 kb. We have found this ratio to give multiple inserts of concatamer and recommend that the ratio be increased to contain at least a 3:1 ratio of vector: insert. We have used this protocol to clone fragments of concatamer which are between 1200 and 1500 nucleotides in length, as judged by restriction analysiε and transcription of BamHl linearized DNA fragments.

After the concatameric casεette is developed, the "self"-cleavage of the TR concatamer by the RR can be checked by incubation of ³² P-labeled, in vitro transcribed ribozyme in the preεence of 10 mM Mg ⁺² as described previously. After appropriate times, aliquots εhould be analyzed by gel electrophoreεiε. A εucceεεful digeεtion reaction iε exemplified by the disappearance of full- length transcript and accumulation of monomer or multimer length ribozymeε over time. After verification of RR activity in vitro the cassette can be excised from the plasmid vector and inεerted into a eucaryotic expression vector for intracellular analyseε of , ribozyme tranεcription and stability and the cleavage activities of the RR and TR can be assesεed.

We have found it preferable to limit the εize of the TRC and TRCB oligonucleotides because longer annealing arms decrease the extent of concatamer formation. Thiε cloning protocol may be modified for uεe with any ribozyme motif, but uεe of larger inεertε limits the number of ribozymes which may be encoded by any transcript because of limitations upon the εize of the encoding DNA which may be effectively introduced into cellε. If one deεireε to use multiple ribozymeε in the caεsettes, then separate mini-cassettes containing the TR concatamer (or monomer) and a corresponding RR εhould be prepared for incorporation into an expreεεion vector. Uεe of different 5' and 3' restriction endonuclease siteε flanking the individual cassettes would facilitate the cloning into vectors. Because the efficiency of ribozyme release within a cell will depend upon the intracellular salt concentrations, it is preferable to use the same ribozyme motif for the RR as is used for the TR. In this manner, both ribozymes will be active in the same cell types. A preliminary assay of ribozyme activity in cellular extracts is recommended to identify the best cleaving motif for a given target cell type. Other Vectors

In designing suitable ribozyme expresεion vectorε the following factorε are important to conεider.

The final ribozyme muεt be kept as εmall aε poεεible to minimize unwanted secondary structure within the ribozyme and to prevent binding of protein factors which may bind to longer RNAs. A promoter (e.g.. the CMV immediate early promoter) should be chosen to be a relatively strong promoter to which RNA polymerase binds with a high binding constant (K _B > 10 ⁸ M ^"1 ) and rapidly enterε the active mode for transcription (rate > 10 ^"2 sec ¹ ) . Such a promoter εhould expreεε the ribozyme at a level suitable to effect production of enough ribozyme to destroy a target RNA, but not at too high a level to

prevent other cellular activities from occurring (unless cell death itself is desired) .

A hairpin (i.e. , an RNA molecule which base pairs with itεelf) at the 5' end of the ribozyme iε uεeful to protect the ribozyme from 5'-3' exonucleaεeε. A εelected hairpin at the 3' end of the ribozyme iε uεeful to protect from digeεtion by 3'-5' exonucleaεeε. Such hairpinε can be inεerted within the vector εequenceε to allow εtandard ribozymeε to be placed in an appropriate orientation and expreεεed with such sequences attached.

Poly(A) tails are also useful to protect the 3' end of the ribozyme. This can be done by either including a poly(A) signal site in the expresεion vector (to εignal a cell to add the poly(A) tail in vivo) , or by introducing a poly(A) -coding sequence directly into the expression vector. In the first approach the signal must be located to prevent unwanted secondary structure formation with other parts of the ribozyme. For polyadenylation, the εequence AAUAA muεt be present in the correct context at/near the 3' end of the transcript in order to be cleaved by a specific eucaryotic endonuclease. Poly(A) polymerase then adds A residueε (about 250) onto the 3' end of the transcript. In the second approach, the poly(A) stretch may reduce in size over time when expresεed in vivo , and thuε the vector may need to be checked over time. Care must be taken in addition of a poly(A) tail which binds poly(A) binding proteins which prevent the ribozyme from acting, or which transport the ribozyme from the nucleus. Such selected transport may be advantageouε for certain ribozyme applicationε.

The correct combination of 5' and 3' protecting groupε can be determined without actually making expression vectors of this invention. This is advantageous because it takes time to produce such vectorε. Thus, the hairpins can be added to the ribozyme by either chemical RNA synthesis or by transcription using the system described by Milligan et al., 15 Nucleic Acids

Research 8783, 1988. Poly(A) tails can be added directly to the ribozyme using poly(A) polymerase or polynucleotide phosphorylase.

Preparation of Purified Ribozvmeε The following iε a method for preparation of pure enzymatically active ribozymeε (of size between 28 and 70 nucleotide bases) in sodium, potasεium or magnesium salt form by a two step purification method. Generally the method is applicable to both synthetically and enzymatically produced ribozymeε, and entailε use of high performance liquid chromatography (HPLC) techniques on reverse phase columns. Unlike gel purification, HPLC purification as described in thiε application can be applied to virtually unlimited amountε of purified material. This allows generation of kilogram quantities of ribozymes in each purification batch.

Thus, the method involves purification of an enzymatic RNA molecule of 28-70 nucleotide baseε by passing that enzymatic RNA molecule over a high presεure liquid chromatography column. Surpriεingly, applicant haε diεcovered that enzymatically active ribozymeε can be purified in the deεired salt form by the described HPLC methodology.

Preferably, the method includeε paεεing the enzymatically active RNA molecule -over a reverse phase HPLC column; the enzymatically active RNA molecule is produced in a εynthetic chemical method and not by an enzymatic proceεε; and the enzymatic RNA molecule iε partially blocked, and the partially blocked enzymatically active RNA molecule iε paεsed over a reverse phase HPLC column to separate it from other RNA molecules.

More preferably, the enzymatically active RNA molecule, after passage over the reverse phase HPLC column is deblocked and passed over a second reverse phase HPLC column (which may be the same as the reverεe phase HPLC column) , to remove the enzymatic RNA molecule from other components. In addition, the column is a εilica or

organic polymer-based C4 or C8 column having a porosity of at least 125 A, preferably 300 A, and a particle size of at least 2 μm, preferably 5 μm.

Pure ribozyme in a Na ⁺ , K ⁺ , or Mg ²⁺ salt form is produced. By "pure" is meant that the ribozyme is preferably provided free of other contaminants, and is at least 85% in the deεired salt form.

The method generally features HPLC purification of ribozymes. An example of such purification is provided below in which a synthetic ribozyme produced on a εolid phaεe is blocked. This material is then released from the solid phase by a treatment with methanolic ammonia, subεequently treated with tetrabutylammonium fluoride, and purified on reverse phase HPLC to remove partially blocked ribozyme from "failure" sequences (Fig. 23) . Such "failure" sequenceε are RNA moleculeε which have a nucleotide base sequence shorter to that of the desired enzymatic RNA molecule by one or more of the desired baseε in a random manner, and poεεeεε free terminal 5'-hydroxyl group. This terminal 5'-hydroxyl in a ribozyme with the correct sequence is εtill blocked by lipophilic dimethoxytrityl group. After εuch partially blocked enzymatic RNA iε purified, it iε deblocked by a standard procedure, and passed over the same or a similar HPLC reverse phase column to remove other contaminating components, such as other RNA molecules or nucleotides or other molecules produced in the deblocking and synthetic procedureε (Fig. 24) . The resulting molecule is the native enzymatically active ribozyme in a highly purified form.

Below is provided one example of such a method. This example can be readily scaled up to allow production and purification of gram or even kilogram quantitieε of ribozymeε. Example 14:

In thiε example solid phase phosphoramidite chemiεtry waε employed for εyntheεiε of a ribozyme.

Monomers used were 2 '-tert-butyl-dimethylsilyl cyanoethyl- phosphoramidites of uridine, N-benzoyl-cytosine, N- phenoxyacetyl adenosine, and guanosine (Glen Research, Sterling, VA) . Solid phaεe syntheεis was carried out on either an ABI 394 or 380B DNA/RNA εyntheεizer using the standard protocol provided with each machine. The only exception was that the coupling step was increased from 10 to 12 minutes. The phosphoramidite concentration waε 0.1 M. Synthesis was done on a 1 mmole scale using a 1 mmole RNA reaction column (Glen Research) . The average coupling efficiencies were between 97% and 98% for the 394 model and between 97% and 99% for the 380B model, aε determined by a calorimetric meaεurement of the released trityl cation.

After synthesis, the blocked ribozymes were cleaved from the CPG support, and the bases and diphosphoester moiety deprotected in a sterile vial by incubation in dry ethanolic ammonia (2 mL) at 55°C for 16 hours. The reaction mixture was cooled on dry ice. Later, the cold liquid was transferred into a sterile screw cap vial and lyophilized.

To remove the 2'-tert-butyl-dimethylsilyl groups from the ribozyme the obtained residue was suspended in 1 M tetra-n-butylammonium fluoride in dry THF (TBAF) , using a 20-fold exceεε of the reagent for every εilyl group, for 16 hourε at ambient temperature. The reaction waε quenched by adding an equal volume of a εterile 1 M triethylamine acetate, pH 6.5. The εample waε cooled and concentrated on a SpeedVac to half the initial volume.

The ribozymeε were purified in two steps by HPLC on a C4 300 A 5 μm DeltaPak column in an acetonitrile gradient.

The first step, or "trityl on" step, was a separation of 5'-DMT-protected ribozyme(ε) from failure εequenceε lacking a 5'-DMT group. Solvents used for this step were: A (0.1 M triethylammonium acetate, pH 6.8) and

B (acetonitrile) . The elution profile was: 20% B for 10 minutes, followed by a linear gradient of 20% B to 50% B over 50 minutes, 50% B for 10 minutes, a linear gradient of 50% B to 100% B over 10 minutes, and a linear gradient of 100% B to 0% B over 10 minutes.

The second step was a purification of a completely deblocked ribozyme by a treatment of 2% trifluoroacetic acid or 80% acetic acid on a C4 300 A 5 μ DeltaPak column in an acetonitrile gradient. Solvents used for this second step were: A (0.1 M Triethylammonium acetate, pH 6.8) and B (80% acetonitrile, 0.1 M triethylammonium acetate, pH 6.8). The elution profile waε: 5% B for 5 minutes, a linear gradient of 5% B to 15% B over 60 minutes, 15% B for 10 minutes, and a linear gradient of 15% B to 0% B over 10 minutes.

The fraction containing ribozyme, which is in the triethylammonium salt form, was cooled and lyophilized on a SpeedVac. Solid residue was dissolved in a minimum amount of ethanol and ribozyme in sodium salt form was precipitated by addition of sodium perchlorate in acetone. (K ⁺ or Mg ²⁺ salts can be produced in an equivalent manner.) The ribozyme was collected by centrifugation, washed three times with acetone, and lyophilized. Preferred Targets An improved method for designing active hammerhead ribozymes is provided herein. Such ribozymes are chosen to include sequenceε which form "εtrong" baεe- pairing interactionε at the positions closest to the internal loop formed by the ribozyme-subεtrate complex (parameterε for εequence-εpecific duplex εtability are found in Freier .et al . , "Improved Parameterε for Predictionε of RNA Duplex Stability", 83 Proc. Natl. Acad. Sci. USA 9373, 1986) . Thuε, such ribozymes have a substrate sequence XXUH/XX (where XX are "strong" duplex forming bases, H is U, A or C, and the slaεh represents the cleavage site) . "Strong" duplex forming bases are chosen to be those which have dinucleotide free energies

of helix propagation of -2.0 kcal/mole or less (i.e. , GA, UC, GU, AC, CG, GC, GG, CC; see, Freier et al . ) . Such ribozymes are more active than other ribozymes having "weak" duplex forming bases at the positions XX. These ribozymes are more efficient at cleaving the substrate RNA molecules since the ribozyme is able to bind to the substrate tightly enough to allow the catalytic components to be positioned for cleavage, but is still weakly enough bound to allow the reaction to be driven to completion by dissociation of the product RNA. Provision of strong ribozyme-substrate interactions at these positions allows for maximal cleavage activity by forming a rigid structure close to the central loop of the ribozyme. Thiε preventε the interior portionε of εtemε I and III of the ribozyme from εeparating prematurely due to local thermal fluctuationε. Fig. 25 iε a diagrammatic representation of the equilibrium between a tightly constrained structure ("Proper configuration") and an undesirable loose structure ("Improper configuration") . The method of this invention is to select ribozymes and substrates which maximize the proper configuration.

Stem III is limited in this regard since the innermost base-pair muεt be AU with the U ^" in the substrate. Stem I, however, has no such constraints and ribozymeε can be chosen which include "strong" duplex forming baseε immediately adjacent to the internal loop.

Thuε, the method for providing a more active hammerhead ribozyme includes modifying the ribozyme to have "strong" duplex forming baseε at the variable positions closeεt to the internal loop.

In addition, a method for εelecting a substrate or target site at which a hammerhead ribozyme will be more active is provided. The method includes the step of εelecting a εubεtrate εite having the sequence XXUH/XX (where XX are "strong" duplex forming bases, H is U, A or C, and the εlaεh represents the cleavage site) .

Example 15: HIV-1 and HSV Targeted Ribozymes

The following is data obtained from various ribozymes demonstrating the deεirability of ribozymes having strong duplex forming baseε a the variable poεitionε cloεeεt to the internal loop.

Tables 3 and 4 contain sequence and activity information for ribozymes designed againεt targets found in the HIV-1 and HSV-1 genomes. Activities (as kcat/k _M ) were determined by incubating 2-100 nM ribozyme with approximately 1 nM radioactively labeled substrate for different timeε in 75 mM Triε-HCl pH 8.0, 10 mM MgCl ₂ . The products are separated by denaturing polyacrylamide electrophoresis and quantitated by a beta emission scanner (Ambis Syεtemε, San Diego, CA) . The data are treated per Michaelis-Menton for excesε enzyme, εpecifically, the negative log of the fraction substrate remaining is plotted as a function of time. The resulting slope is then plotted as a function of ribozyme concentration. The εlope of this line is then kcat/k _M . Ribozymes which meet the strong base-pair criteria for stem I are marked with a "I" in the left margin. The average activity of these ribozymes is 43 x 10 ⁶ /M«min, which is above the average for ribozymes not meeting this criteria (23 x 10 ⁶ /M ^« min) . Ribozymes which meet the strong base-pair criteria for stem III are marked with a "III" in the left margin. The average activity of these ribozymes is 28 x 10 ⁶ /M«min, and for those not meeting thiε criteria iε 8 x 10 ⁶ /M*min. Six ribozymes have the appropriate strong inner baεeε in both εtemε. Their average activity iε 37 x 10 ⁶ /M«min.

TABLE 3

Summary of HIV ribozyme activity

Substrate Sequence

HCS r29ch GGAGCAUCCAGGAAGUCAGCCUAAMCUO

HBS-r21 UQCGAGAQCGUCADUAUUAAQ

HDS-T13 GGCAGUCAGACUC

HES-r13 CAGACUCAUCAAG.

HFS-M3 CUCOAUCAAAGCA

TABLE 4 HSV Summary

< _. nτ

ID

TO-

/

(x10 ^Λ 6)

Substrate Sequence

VAS-r17- (}AGACCX3UC^GACGCUC

VCS- 7 GUCGG0GUCCGUGQ3UC

VES-r17 GAIX3G03UCGGA3AACA

VFS-r15 G XXXX3UCXJCGUC

VGS-r16 GUCCUCGUCOICCGCA

VIS-r17 CX33Q9GGUCUUCGCG0G

VJS-M4 GGUOOUGIWCGGG

Arm Lengths for Optimal Enzyme Activity

A rapid assay for determining optimum length of ribozyme/substrate interaction regions iε provided. Applicant has determined that it is important to know the optimum length of the interaction regions, e.g. , stem I & stem III, of a hammerhead ribozyme targeted against a specific RNA sequence. Similarly, it is important to determine such parameters for other ribozymes, such as hairpin ribozymes, and those based upon the hepatitis delta virus structure.

Generally, a series of ribozymes that vary in the length of these regions has been made, the ribozymes purified, and kinetic assays performed on such ribozymes to determine the best configuration. Synthesis and testing of three ribozymes takes about two weeks. Thus, testing many different target sites with 6-8 ribozymes per site becomes a lengthy and expensive proposition.

Applicant has discovered that a more rapid determination of optimum arm lengths can be achieved by evaluation of many different arm-length combinations in one set of experiments. It is difficult to assay multiple ribozymes (targeted to the same substrate) in a single experiment because it is hard to distinguish which ribozyme is responsible for each cleavage event. However, it is possible to make a nested set of substrate RNAs that differ in the number of base-pairs that are made with a single ribozyme. Thiε reverse experiment provides approximately the same results as the desired experiment (changing the length of the ribozyme) as long aε the ribozyme used is long enough to interact with the longest substrate. Preferably, neither the ribozyme nor the longest εubεtrate form alternative structures by themselves. Applicant recognizes that the interaction free energies between a long ribozyme and a truncated εubstrate will not be identical to the interaction free energies for long substrates and truncated ribozymes, even if the base-pairs formed are the same and there are no

competing secondary structures. This is becauεe the terminal stacks contribute from 0 to -1.7 kcal/mole to the interaction free energy, and the difference in interaction free energies can be as much as 2.0 kcal/mole. Thuε, it may be necessary to make and test a few ribozymes with arms of varying lengths to ensure that these differences are not overlooked. The number of ribozymeε that will need to be made and teεted, however, will be significantly lesε than that number when each different ribozyme muεt be εynthesized rather than each substrate.

Thus, a method is provided for asεaying optimal arm lengthε in an enzymatic RNA molecule by contacting the molecule with a plurality of substrate RNAs. Each substrate RNA differs from another in the length or number of base pairs which each substrate RNA forms with the molecule. The method is performed under RNA subεtrate cleaving conditionε, and the rate of cleavage of the substrate RNAs is determined.

Preferably, the ribozyme is a hammerhead, hairpin or hepatitis delta virus ribozyme; each substrate RNA has an identical 5' or 3' end; and the kinetics of cleavage of each substrate is determined.

By "kinetic assays" or "kinetics of cleavage" is meant an experiment in which the rate of cleavage of target RNA is determined. Often a series of assays are performed in which the concentrations of either ribozyme or substrate are varied from one asεay to the next in order to determine the influence of that parameter on the rate of cleavage. By "rate of cleavage" iε meant a measure of the amount of target RNA cleaved as a function of time.

By "nested set of RNA" is meant a set of RNA molecules having a common sequence but varying in length through truncation of the sequence from either the 3' or 5' ends (or both) . The set of sequenceε AU, AUG, AUGC, AUGCG, AUGCGA defineε a nested set RNA sequenceε having

common 5' ends, while AUGCGA, UGCGA, GCGA, CGA, GA defineε a nested set of RNA sequences having common 3' ends.

By "ribozyme/substrate interaction regionε" is meant sequences within a ribozyme which are intended to bring ribozyme and target RNA together through complementary base-pairing interactions; e.g. , ribozyme sequences within stems I and III of the hammerhead ribozyme make up the ribozyme/substrate interaction regions (see Fig. 29) . A method iε also provided for production of a εet of RNA εubstrates. Thiε involveε synthesizing an RNA molecule with a limiting level of phosphoramidite or using shortened coupling times under conditions in which RNA sequences are capped when no nucleotide base is added at a given step. Example 16:

In one example, a set of substrates is synthesized by separate chemical synthesis of each individual substrate. Because test substrates are shorter in length than ribozymes, this approach uses less reagent and time than syntheεiε of an equivalent set of ribozymes and is consequently lesε expensive. Synthesis of shorter RNA molecules is also more efficient and the resulting products are easier to purify. Example 17:

A time and cost efficient way to generate a nested set of εubstrateε is to uεe standard chemical synthesiε with high coupling efficiency for the initial portion of a εynthesis, and then switch to low efficiency coupling for syntheεis of the last few (5-9) nucleotideε of the εubεtrate RNA. Low efficiency coupling can be achieved by lowering the concentration of phoεphoramiditeε or by decreasing the time allowed in the coupling reaction. This method is outlined in McSwiggen, supra . The end result is a set of substrateε sharing a common 3' end but that differ in length at the 5' end.

After chemical synthesis of the set of RNAε, individual RNAs of the set can be isolated (e.g. , by gel- purification) . The isolated RNAs are then 5' end-labeled with [γ- ² P]ATP and T4 polynucleotide kinase. To generate molecules in which the 3' end is alεo truncated, the individual 5' end-labeled substrates are subjected to limited base hydrolysiε or exonuclease treatment and those substrates in which 0 to 5 nucleotides have been removed from the 3' end are gel-purified. This procedure yields setε of RNAs that have a common 5' end, but are shortened on their 3' end. This general scheme is outlined in Fig. 26.

Together theεe two procedures, chemical syntheεiε with low efficiency coupling during 5' end synthesiε and subsequent base hydrolysis or exonuclease treatment of isolated RNAs, yields a full set of RNA substrates with a common central sequence but with all possible combinations of truncations in the last 5-9 nucleotides at both ends. Since only 'one ⁷ chemical syntheεiε iε performed, thiε method of generating a full set of substrates is greatly advantageouε over chemical synthesis of each individual subεtrate. Less time and reagent is required to generate the full set of substrates to be tested. Example 18:

The substrates are tested in a standard kinetic asεay to determine which one provides the best substrate for a particular ribozyme (see Fig. 27) . Each substrate is mixed with a constant amount of ribozyme and incubated over a particular time course. The resulting products are electrophoresed on a denaturing polyacrylamide gel and the extent of cleavage determined for each time point. Kinetic analyεiε reveals on which substrate the ribozyme is most active. Example 19:

The individual subεtrateε need not be separated from each other in order to be tested. Rather, a set of

substrates that, for example, differ only in length at their 5' end can be 5' end-labeled with [7~ ³² P]ATP and T4 polynucleotide kinase. Incubation of this set of 5' labeled εubεtrateε with a ribozyme allows the same kinetic analysiε as incubation with a single εubstrate. The cleavage productε generated alεo vary at their 5' end, and thus will be distinguishable from each other and the full- length molecule on a denaturing polyacrylamide gel (see Fig. 28) . This method of kinetic analysis of entire sets of subεtrateε alεo workε with εubstrates that have a common 5' end, but are different at their 3' end. The latter can be 3 ' end-labeled with [ ³² P]-cordycepin and poly(A) polymerase or [ ³² P]-labeled pCp and T4 RNA ligase using standard methods. Example 20:

The method includes determination of the maximum allowable length for the ribozyme substrate interaction. The candidate is folded using a standard computer program, e.g.. one described by Zucher, 244 Science 48, 1989, along with the full-length ribozymes and substrate oligonucleotides, and examined for secondary structureε which might affect activity. The length of the ribozyme/substrate interaction is adjusted to avoid such secondary structures. Full-length ribozyme is made and purified substrate oligonucleotides containing 5' truncations are syntheεized, e.g.. on an ABI εyntheεizer. A nested set of subεtrate RNAs containing the same 3' end but differing 5' ends can be produced by reducing the phosphoramidite concentration and ensuring that failure sequences get fully capped (see McSwiggen, supra) . The 5' truncated substrates are labeled at their 5' end with [7~ ³² P]ATP and T4 polynucleotide kinase and gel purified. Excesε ATP iε uεed to enεure complete 5' phoεphorylation εo that the reεulting RNA does not contain unlabeled n+1 substrate, where N is the desired length. Alternatively, the set of RNA molecules may be gel purified and then 5' end labeled with [7~ ³² P]ATP.

The 5' truncated substrates are checked by RNA sequencing to ensure that the expected RNA has been obtained. 3' truncated substrates are generated for each 5'-truncated substrate by limited base-hydrolyεiε or 3' nucleaεe digeεtion (e.g. , Nucleaεe P) . The base- hydrolyεiε reaction is neutralized and used directly in kinetic reactions, or 3' truncated substrates are gel purified for more detailed analysis.

Kinetics are then analyzed in terms of fraction of initial subεtrate remaining vε. time, εince in thiε caεe the product band will be the same for a given set of 3' truncated substrateε that are ³² P-end-labeled at their common 5' end. Synthesizing Sets of RNA Molecules Applicant has recognized that a 1 μmole scale standard chemical synthesiε of a ribozyme produceε a lot of ribozyme compared to the amount needed to carry out initial activity tests. Thus, applicant has devised a method by which such syntheses can be optimized to produce multiple ribozymes or other RNA molecules in a single syntheεiε. Thiε allows tremendous savingε of both material costs and purification time.

Test sets of ribozymes are routinely made that are targeted to the same target sequence but which contain varying lengths of substrate-binding arms. Using the method, setε of ribozymes containing variable-length 5' εubεtrate-binding armε can be chemically εyntheεized. In one embodiment, thiε iε achieved by εimply reducing the concentration of phoεphoramidites used during the last few stepε of RNA synthesiε (i.e.. near to the 5' end) in order to promote production of completed εequenceε of varying lengthε at each synthetic step.

Thus, a method for synthesiε of a set of enzymatic RNA molecules each member of the set differing in length is provided. In the method, syntheεiε of the moleculeε entailε uεe of a first concentration of reagent (preferably in excesε) and employment of an incubation

time sufficient to allow at least 95% coupling efficiency per nucleotide position. After a few synthesis steps (e.g.. during synthesizing of the 5' substrate binding region) a second concentration of reagent iε uεed which is insufficient to allow the high coupling efficiency. Alternatively, the time allotted to each coupling step iε reduced such that incomplete coupling is achieved.

Preferably, the limiting reagent is phosphoramidite or time is made limiting; the limiting concentration or time iε sufficient to allow between 80- 90%, or even 70-80% coupling efficiency; at least 3 (most preferably 5-9) different RNA molecules are εynthesized which differ in length by at least one nucleotide base each, and those RNA moleculeε have different 5' ends (see Fig. 29) . Example 21;

Referring to Fig. 29, the following is one example of this invention. Those in the art will recognize that equivalent strategieε can be used to attain an equivalent result.

Two RNA synthesizers (or two setε of reagentε on the εame synthesizer) are set up, one contains the standard concentrations of phosphoramidites (e.g. , 0.1 M) , while the other containε phosphoramidites at a lower concentration designed to give only 80% coupling efficiency under standard syntheεiε conditionε. Such a concentration can be determined empirically for each synthesizer or set of reactantε.

RNA synthesis iε εtarted with the εtandard concentrationε of phoεphoramiditeε and continued until the RNA sequence position at the 5' end where failure sequenceε are deεired.

The RNA εynthesis columnε are transferred at this point to the second synthesizer and syntheεiε continued for up to εeven more poεitionε at 80% coupling efficiency per poεition. A εet of ribozymes containing n, n-1, n-2, n-3, n-4 , n-5, n-6 nucleotides at the 5' end

will be generated in concentrations of 26%, 7%, 8%, 10%, 13%, 16% and 20% of the total, respectively. Capping is done by standard procedure (ABI Bulletin 53 _. :1-7 (1988)). The RNA in each column is deprotected and deεalted in the uεual manner. Small failure εequenceε can be diεcarded if deεired by gel purification, but n, n-1, n-2, n-3, n-4, n-5, and n-6 RNAε are not εeparated from each other.

A base composition analysiε is run on the deεalted RNAs (one base-compoεition analysis will suffice for up to 7 RNAs) . Individual RNAε are purified uεing either ion exchange HPLC or gel purification. In εome instances, purification of individual RNAs may not be necessary as the entire set can be utilized together. If separation of n from n-1 is difficult, synthesis can be extended one more nucleotide so that the relatively large amount of full-length RNA (26% in this case) is more easily separated from the largest desired failure sequence (at 7% in this case) . Purified ribozymes are then labeled at the 5' end with [7~ ³² P] ATP using T4 polynucleotide kinase. The ³² P-labeled ribozyme is then checked for purity and sequenced to ensure that the desired sequence has been obtained. Example 22:

A more uniform diεtribution of n, n-1, n-2, n-3, n-4, etc. RNAε can be obtained if the coupling efficiency iε adjusted at each of the n-4 through n coupling stepε. The moεt direct way of varying coupling efficiency at will is to change the coupling time while keeping the phoεphoramidite concentration conεtant. The standard coupling reaction employs 0.1 M phosphoramiditeε and a coupling time of 12-15 minutes to obtain 97-99% coupling efficiency. By reducing the phosphoramidite concentration, a longer time will be required to obtain the same level of coupling (0.1 M phosphoramidite is about 20-fold molar excesε over growing RNA chains in the

standard reaction; it is important that the phosphoramidite concentration not be dropped below a 1:1 molar ratio with the growing RNA chainε) . Fig. 30 εhowε a hypothetical εituation in which the phoεphoramidite concentration iε reduced εufficiently to require 43 minuteε to obtain 99% coupling efficiency. The coupling efficiency iε assumed to follow a curve of the form Coupling Efficiency = ι-3-< ^coulιn s ^,in>e > In such a case, the coupling efficiency would be 50% at 6.5 minuteε, 66.7% at 10.3 minuteε, 75% at 12.9 minuteε, and 80% at 15 minutes. If 5 RNAs are desired in equal concentrations then the coupling times and yields could be achieved as shown below:

One or more stabilizing RNA hairpins may be provided to a hammerhead or hairpin ribozyme, to provide an improved ribozyme. These RNA hairpins provide protection from RNaεeε within a cell with minimal increaεeε in the toxicity of the RNA molecule. Theεe additionε are readily added to exiεting or deεigned ribozymeε, εince they can εimply be appended to the RNA εequence making up a ribozyme. Such hairpinε may be added either by chemical εyntheεiε, by gene tranεcription, or by RNA ligation procedures. A 5' hairpin protects the ribozyme from 5'-3' exonucleaεeε and a 3' hairpin can be choεen to protect from digeεtion by 3'-5' exonucleaεeε.

Thus, the hammerhead or hairpin ribozyme may have a 5' or 3' RNA hairpin which does not significantly reduce the enzymatic activity of the ribozyme.

By "hairpin" is meant a nucleotide base sequence of RNA that may be ligated or otherwise attached to one end of a hammerhead ribozyme. Thiε RNA will baεe pair with itεelf to form a folded εtructure, commonly referred to aε a hairpin. Thiε εtructure may alεo fold onto nearby ribozyme εequenceε εo long aε such folding does not reduce the activity of the ribozyme significantly. Such a hairpin may be spaced from one end of the ribozyme by provision of one, two or more non-pairing nucleotide bases.

Preferred hairpins include GC rich εtemε with εtable tetranucleotide loops (e.g. , GAAA, UUCG, GNRA or UNCG where N = A, U, C or G and R = A or G) ; most preferred are hairpins including a sequence selected from GGCCGAAAGGCC, GCGCUUCGGCGC, and GGAGGAAACUCC, with the hairpin consiεting of between a 4 and 20 baεe pair stem and between a 4 and 7 nucleotide loop.

Hairpins can be readily added to 5' or 3' ends of a hammerhead ribozyme. Any particular hammerhead or hairpin ribozyme may be chosen and its stability enhanced by provision of such a hairpin. The optimum hairpin for any particular ribozyme may differ, but enhanced stability can be provided by use of one or more universal hairpins. Such universal hairpins can be chosen based upon an analysiε of the kineticε of enzymatic activity, and the εtability of a ribozyme uεing standard assays.

Initially, in choosing an optimal hairpin, computer folding of various ribozy e-hairpin combinations will indicate whether a chosen hairpin will promote misfolding of the ribozyme. Such hairpins should be avoided. Design of the hairpins which do not cause misfolding is readily performed. Steric hindrance of the folding of the ribozyme should be avoided as well, or the enzymatic activity of the ribozyme may be reduced. In a model syεtem, the length of the hairpin and the length of non-pairing εpacer sequences between the ribozyme subεtrate binding site or sequence and the

hairpin may be varied to determine which hairpin or non- pairing spacer sequence provides optimum stability, without significantly reducing enzymatic activity. Hairpins which promote in vivo binding to cellular factors should be avoided since these may lead to a reduction of ribozyme activity. Such a poεεibility may be aεεayed uεing a routine assay in which a cellular extract is added to an aεsay mixture to determine ribozyme activity in the presence of such a cellular extract. Thus, for example, recognition sequenceε, such aε the tat recognition εequence, should be avoided in the design of such hairpins.

After εelection of a hairpin which doeε not εignificantly diminiεh the enzymatic activity of the ribozyme when provided at the 3 ' or 5' end, the εtability of the ribozyme can be tested by addition of specific RNases using a standard asεay or by teεting in cell extracts. Larger hairpins are more effective at stabilizing ribozymes against exonuclease attack, but shorter hairpins are more readily synthesized and are lesε likely to reduce ribozyme activity. Thuε, the optimum εize of such a hairpin will need to be discovered empirically for any chosen ribozyme.

There follows exampleε of a procedure for aεεaying for a suitable 3' or 5' hairpin. These examples are not limiting in the invention and thoεe of ordinary εkill in the art will recognize that many equivalentε to theεe exampleε exiεt. Example 23: A ribozyme having εtrong enzymatic activity such as one termed HSV, ICP4 site E (shown in Fig. 31A) was εelected aε a model ribozyme. Referring to Figε. 32A-F, the core ribozyme was synthesized, and other modifications of the ribozyme synthesized using standard procedures. These included the core ribozyme (Fig. 32A) , core ribozyme including a minimal 3' hairpin (e.g.. GCGCGAAAGCGG, Fig. 32B) , a core ribozyme having the same minimal hairpin

except including a one nucleotide non-pairing spacer between the ribozyme εequence and the hairpin (Fig. 32C) , a core ribozyme including a minimal 5' hairpin which includeε the T7 transcription start sequence (but avoiding uracil nucleotides in the first nine nucleotides to minimize abortive initiation, e.g. , GGACGAAAGUCC, Fig. 32D) , the core ribozyme with a 5' hairpin which includes the T7 tranεcription εtart sequence but with a one nucleotide spacer between the hairpin sequence and the ribozyme (Fig. 32E) , and a core ribozyme flanked on its 5' side with the hairpin which includes the T7 transcription start site, a one nucleotide spacer and on its 3' side with the 3' hairpin and a one nucleotide spacer. Other hairpins, or any combination of the above variations which might be readily εyntheεized can alεo be tested and used. All of these ribozymes can then be tested for their enzymatic activity in vitro (shown in Figs. 32A-F) and then in vivo , and then tested for their stability using routine assays such as thoεe provided below. Example 24: Enzymatic Assay

To test the effect on activity in vitro small stem-loops to the termini of a ribozyme, the ribozymes shown diagrammatically in Figs. 32A-F were made and tested for cleavage activity on a short (17 mer) synthetic substrate.

Activities (aε kcat/k _M ) were determined by incubating 2-100 nM ribozyme with approximately 1 nM radioactively labeled εubstrate for different times in 75 mM Tris-HCI pH 8.0, 10 mM MgCl ₂ . The productε were εeparated by denaturing polyacrylamide electrophoreεiε and quantitated by a beta emisεion scanner (Ambis Systems, San Diego, CA) . The data were treated per Michaelis-Menton for excess enzyme, specifically, the negative log of the fraction substrate remaining waε plotted aε a function of time. The reεulting εlope waε then plotted aε a function of ribozyme concentration. The εlope of thiε line iε kcat/k _M .

Theεe ribozyme conεtructs all have activities within a factor of 2. This demonstrates that the addition of terminal stem-loops is not inhibitory to the activity of thiε ribozyme. These ribozymes were further tested for activity in vitro against a long (496 nucleotide) T7 transcript containing the target site. In this experiment, 4 μM of each ribozyme was incubated with 47 nM body-labeled transcript in the above-mentioned buffer for various times. The products were separated by a 5% denaturing gel and visualized by autoradiography. The result showed that cleavage by each ribozyme (except that in Fig. 32C which was εhown to be degraded prior to the experiment) occurred at virtually identical rates. Example 25: Stability Assay

Ribozyme constructs such as those described in Example 23 and Figs. 32A-F can be tested for stability against ribonuclease degradation in cellular extracts. The ribozyme construct is first labeled (e.g.. with ³² P) . This can be accomplished by incubation with [7~ ³² P]ATP and T4 polynucleotide kinase to label the 5' end, incubation with ³² P-cordycepin and poly(A) polymerase to label the 3 ' end, or by inclusion of a [ ³² P]NTP during the transcription reaction if the construct is such that it can be made by in vitro transcription. Various concentrationε of ³² P- labeled ribozyme are then incubated with cellular extractε. Aliquotε are removed at variouε timeε, mixed with standard stop buffer containing EDTA and formamide, and loaded onto a denaturing polyacrylamide gel. Full- length ribozymes are diεtinguished from smaller degradation products by size. Alternatively, complete degradation is followed as a disappearance of ³² P-labeled full-length ribozyme.

The modified ribozymes of this invention may be formed using a DNA template which is transcribed or chemically synthesized using εtandard procedureε. Such ribozymeε can be synthesized for RNase experiments in

which the ribozymes are labeled and tested in εtability aεεays using cell extracts.

Method for Forming Active Ribozymes

Once a ribozyme or enzymatic RNA molecule has been designed to a particular target RNA site, it is often useful to modify various nucleotide bases in the ribozyme to determine those bases which can be modified without losε of enzymatic activity, and those bases which can be modified to enhance enzymatic activity. Individual modification of each base position in a ribozyme is a slow method for determining those bases in the RNA molecule which are suited for such modifications. This invention features a method by which rapid screening of various modificationε can be achieved in a random or non-random manner.

In this method, the ribozyme is synthesized such that it contains chemical modifications at a given base or linkage in such a manner that that chemical modification can be later detected. The ribozyme is synthesized so that it is linked to or contains a substrate portion, such that the ribozyme may act in an intramolecular manner to cleave the substrate in cis , to generate at least two RNA fragments. In this way, the activity of each modified ribozyme can be tested by incubating the modified ribozymes in vitro under conditions which allow intramolecular cleavage to occur. Any modification which prevents the ribozyme from causing such cleavage will result in the εubεtrate remaining bound to that ribozyme molecule during εuch an incubation. Such bound εubstrate molecules are readily separated from those in which cleavage has occurred, and the position of the modified baεe or bases in the cleaved or uncleaved RNA molecules can be determined using standard technology.

Thuε, the method involves syntheεiε of an enzymatic RNA molecule in which the ribozyme molecule iε formed with a modified nucleotide baεe or a non-nucleotide baεe which can be detectably located relative to

unmodified nucleotide baεes. The RNA molecule iε formed with a εubεtrate portion which the enzymatic portion of the RNA molecule can cleave in an enzymatic manner.

By "unmodified nucleotide" is meant one of the bases adenine, cytosine, guanoεine, uracil joined to the 1' carbon of β-D-ribo-furanoεe. The sugar alεo has a phosphate bound to the 5' carbon. These nucleotides are bound by a phosphodiester between the 3' carbon of one nucleotide and the 5' carbon of the next nucleotide to form RNA.

By "modified nucleotide/RNA" is meant to include nucleotides or RNA which have modifications of: (1) the base moiety (e.g. , 2-Bromo-uracil) ; (2) the sugar moiety (e.g.. 2'-0-methyl) ; or (3) the phosphate moiety (e.g.. thiophosphates) .

Preferably, the method involves synthesis of a hammerhead, hairpin or hepatitis delta virus (HDV) ribozyme with the substrate attached to the 3' or 5' end (for example, in a manner that allows end-labeling of the products of the enzymatic reaction) , in which intramolecular cleaving of the ribozyme and substrate is achieved. For example, the HDV may include an extraneous sequence (not normally associated with the HDV ribozyme) on the 5' side of the cleavage site so that active or inactive ribozymes can be separated from each other dependent on size.

More preferably, the chemical modification involves use of thiophosphate. A modified base linkage formed in this manner can be detected by preferential cleavage with hydrogen peroxide or periodate cleavage, by protection from digestion with snake venom phosphodiesterase, aε well as some base specific RNaseε. Such cleavage allowε location of the modified base(s) relative to non-modified baseε. The baεeε may be modified by use of a methyl phosphonate, and can be cleaved and thereby detected by use of base specific RNaεeε; or the baseε are modified by

use of 2'-0-methyl, 2'-fluoro, 2 '-amino nucleotides or 2'- deoxynucleotides. The position of these modifications can be detected by resistance to base hydrolysis.

In addition, a method is provided for screening for nucleotide positions in a ribozyme which can be modified without eliminating ribozyme activity. The method involves providing a chemically synthesized ribozyme having a subεtrate attached in a manner that allows intramolecular cleaving of the subεtrate by the enzymatic portion of the ribozyme, and synthesizing the ribozyme such that a modification is made to one or more baseε in the ribozyme. That ribozyme iε then incubated under intramolecular cleaving conditionε, and the location of thoεe modificationε which eliminate or have no effect on the intramolecular cleaving reaction determined.

Preferably, the method further featureε εyntheεizing further modified ribozymeε which may include modificationε at those nucleotide positionε which were previously found not to affect the intramolecular cleaving activity. This procedure may be repeated until all those positions that can be modified have been identified.

Finally, the modified ribozymes having intramolecular cleaving activity can be tested with or without an attached substrate and compared in their stability and kinetic activity with an unmodified or parental ribozyme molecule. In this way, not only may modification siteε be determined which may later be modified in the ribozyme, but alεo ribozymeε which have increaεed activity due to εuch modifications can be determined. Example 26:

Referring to Fig. 33, an unmodified hammerhead ribozyme having a substrate attached at the 3' end is synthesized and purified using standard procedures (ABI User Bulletins 1988, 42:1-9; 1989, 51:1-7) . This εyntheεiε iε then repeated in a manner that cauεes about 3% of the ribozyme baseε to be chemically modified during

synthesis (to give about one modification per 35 nucleotide residueε) . Such chemical modification may include the uεe of thiophoεphate or methylphoεphonate among other reagentε. Thiophoεphateε may be introduced by limited treatment with 3H-1, 2-Benzodithiol-3-one-l, 1- dioxide at εpecific baεeε during automated εynthesis (Iyer et al., 11 JACS 121, 1990).

The ribozyme-substrate population of molecules is then 5' end labeled and incubated under RNA cleavage conditionε and both the productε and the εtarting material are gel purified to iεolate intramolecularly-cleaved ribozymeε from non-cleaved ribozymeε. Baεe poεitions which prevent intramolecular cleaving are determined by comparing the reacted and unreacted ribozyme populations by using a technique that is sensitive to the presence of the chemical modification, for example, for thiophosphate, periodate, snake venom phosphodieεterase, or base specific RNases can be utilized. For methyl phosphonate, a base specific RNase can be used. For 2 ^, -0-methyl, 2'- fluoro, 2'-amino, or 2'-deoxy modifications base hydrolysis can be used.

The above procedure can be repeated by synthesizing a ribozyme in such a manner that each ribozyme position has about a 6 to 20% chance of the bases being chemically modified (i.e.. between 2 to 7 modifications per 35 nucleotide ribozyme) and those positionε which were found in the previouε εtep to inhibit intramolecular cleaving are not modified. The poεitionε of chemically modified nucleotideε which prevent intramolecular cleaving can then be determined aε deεcribed above. Thiε procedure can be repeated aε many timeε aε desired.

The method of making and testing ribozymes with increased levels of modifications at allowed sites, while not modifying at the disallowed εiteε, will ultimately provide a population of active ribozymes that are fully modified at all allowed sites and completely unmodified at

the disallowed sites. These ribozymes can be made and tested with or without substrate attached using standard aεsays to determine their activity in kinetic experimentε aε well aε their εtability in vitro and in vivo . See, McSwiggen, "Improved Ribozymeε", U.S. Serial No. 07/884,521, filed May 14, 1992 and aεεigned to the same asεignee aε the preεent application, the whole of which (including drawingε) iε hereby incorporated by reference herein. Catalytic activitieε of enzymatically active modified ribozymeε can next be determined by teεting the specific ribozyme with a separate εubstrate such that the ribozyme can cleave in the separate subεtrate. Specifically, the modified ribozyme of intereεt and itε εubstrate are syntheεized and purified separately. Single turnover activities are determined by incubating 2-100 nM ribozyme with approximately 1 nM radioactively labeled substrate for different times in 75 mM Tris-HCI pH 8.0, 10 mM MgCl ₂ . The products are separated by denaturing polyacrylamide electrophoreεiε and quantitated by a beta emiεεion εcanner (Ambiε Syεtemε, San Diego, CA) . The data are treated per Michaelis-Menton for excesε enzyme; εpecifically, the negative log of the fraction substrate remaining is plotted as a function of time. The resulting slope is then plotted as a function of ribozyme concentration. The slope of this line is then kcat/k _M .

These ribozymeε can also be tested for stability against endogenous cellular nucleaseε by incubating end labeled ribozyme in cell extractε. Alternatively, the labeled ribozyme can be delivered to cells in culture or to animals to determine εtability. Thoεe ribozymeε which are enzymatically active and εtable are most useful in therapeutic and other applications.

This section alεo relateε to production of enzymatic RNA moleculeε or ribozymeε having enhanced or reduced binding affinity and enhanced enzymatic activity for their target nucleic acid εubεtrate by incluεion of

one or more modified nucleotideε in the εubstrate binding portion of a ribozyme εuch as a hammerhead, hairpin or hepatitis delta virus derived ribozyme. Applicant has recognized that only small changes in the extent of base pairing or hydrogen bonding between the ribozyme and subεtrate can have εignificant effect on the enzymatic activity of the ribozyme on that εubεtrate. Thuε, applicant has recognized that a subtle alteration in the extent of hydrogen bonding along a substrate binding arm of a ribozyme can be used to improve the ribozyme activity compared to an unaltered ribozyme containing no such altered nucleotide. Thus, for example, a guanosine base may be replaced with an inosine to produce a weaker interaction between a ribozyme and its substrate, or a uracil may be replaced with a bromouracil (BrU) to increase the hydrogen bonding interaction with an adenosine. Other exampleε of alterationε of the four εtandard ribonucleotide baεeε are εhown in Figε. 47A-D with weaker or stronger hydrogen bonding abilities shown in each figure.

Thuε, a modified ribozyme iε provided having one or more substrate binding arms including one or more modified nucleotide baseε; and in a related aspect, the invention features a method for production of a more active modified ribozyme (compared to an unmodified ribozyme) by inclusion of one or more of such modified nucleotide baseε in a εubεtrate binding arm.

The invention provideε ribozymeε having increased enzymatic activity in vitro and in vivo as can be measured by standard aεεayε. Thuε, the kinetic featureε of the ribozyme are enhanced by selection of appropriate modified baseε in the substrate binding arms. Applicant recognizeε that while εtrong binding to a substrate by a ribozyme enhances specificity, it may also prevent separation of the ribozyme from the cleaved substrate. Thus, applicant provides means by which optimization of the base pairing can be achieved.

Specifically, the invention features ribozymes with modified baseε with enzymatic activity at leaεt 1.5 fold (preferably 2 or 3 fold) or greater than the unmodified correεponding ribozyme. The invention alεo featureε a method for optimizing the kinetic activity of a ribozyme by introduction of modified bases into a ribozyme and screening for those with higher enzymatic activity. Such selection may be in vitro or in vivo .

By "enzymatic portion" is meant that part of the ribozyme esεential for cleavage of an RNA substrate.

By "substrate binding arm" is meant that portion of a ribozyme which is complementary to (i.e. , able to base-pair with) a portion of its substrate. Generally, such complementarity is 100%, but can be lesε if desired. For example, as few as 10 bases out of 14 may be base- paired. These arms contain sequences within a ribozyme which are intended to bring ribozyme and target RNA together through complementary base-pairing interactions; e.g. , ribozyme sequences within stems I and III of a standard hammerhead ribozyme make up the substrate-binding domain (see Fig. 1) .

Preferably, the modified ribozyme iε a hammerhead, hairpin or hepatitis delta virus derived ribozyme, and the hammerhead ribozyme includes between 32 and 40 nucleotide baseε. The εelection of modified baεeε iε most preferably chosen to enhance the enzymatic activity (as observed in standard kinetic assays designed to meaεure the kineticε of cleavage) of the εelected ribozyme, i.e. , to enhance the rate or extent of cleavage of a εubεtrate by the ribozyme, compared to a ribozyme having an identical nucleotide base sequence without any modified base.

There iε a narrow range of binding free-energieε between a ribozyme and its subεtrate that will produce maximal ribozyme activity. Such binding energy can be optimized by making ribozymeε with G to I and U to BrU εubεtitutions (or equivalent subεtitutionε) in the

substrate-binding arms. This allows manipulation of the binding free-energy without actually changing the target recognition sequence, the length of the two substrate- binding arms, or the enzymatic portion of the ribozyme. The shape of the free-energy vs. ribozyme activity curve can be readily determined using data from experimentε in which each baεe (or εeveral bases) is modified or unmodified, and without the complication of changing the size of the ribozyme/εubstrate interaction. Such experiments will indicate the most active ribozyme εtructure. It is likely that only one or two modifications are necessary since a very small change in binding free energy (even one base-pair interaction) can dramatically affect ribozyme activity; the use of modified baseε thus permits "fine tuning" of the binding free energy to assure maximal ribozyme activity. In addition, replacement of such bases, e.g. , I for G, may permit a higher level of substrate specificity when cleavage of non-target RNA is a problem. Method

Modified substrate binding arms can be synthesized using standard methodology. For example, phosphoramidites of inosine and 5-bromouracil can be used. Generally, a target site that has been optimized for stem I and III lengths (in a hammerhead ribozyme — other ribozymes can be treated in a similar manner) , and that has G and/or U in the ribozyme portion of stem I and III, is selected. Modified ribozymes are made by replacing various G and U residueε with I and BrU, reεpectively, during εyntheεiε of the ribozyme. The modified ribozymeε are then teεted to determine kinetic parameters using standard procedures. The binding affinities for the ribozymes can alεo be determined by standard procedures, e.g.. by T-melt, gel-binding, or by competition kinetics asεays. By comparison of binding affinity and ribozyme activity the optimum binding affinity of a ribozyme can then be found. Other combinations of G, I, U BrU, and

other bases can then be tested with nearly identical binding free energy, but different base sequence, to determine whether factors other than simple binding free- energy play a role. It is preferred to perform routine experiments of thiε type to εelect a deεired ribozyme εubεtrate binding sequence by use of an unmodified ribozyme with a modified subεtrate (which contains the modified bases) . That is, the reverse experiment to that described above is performed. Such an experiment is more readily performed since the subεtrate iε generally εhorter than the ribozyme, and can be readily synthesized without concern about its secondary εtructure. Thus, a single ribozyme can be tested against a plurality of modified substrateε in order to define which of the εubεtrateε provideε better kinetic results. Once a preferred substrate is identified, the ribozyme can then be modified in a way which mirrors the selected subεtrate, and then teεted against an unmodified substrate. Such experiments will define useful ribozymes of this invention in which one or more modified baseε are provided in the substrate binding arms with greater enzymatic activity in vitro and in vivo than comparable unmodified ribozymes. Such modifications may also be advantageous if they increase the resiεtance of a ribozyme to enzymatic degradation in vivo . In Vivo Selection of Ribozyme Targetε

Thiε method iε for conεtructing and selecting ribozymes, and in particular, amplifying vectors which expreεε ribozymeε in tiεεue culture systemε. Theεe ribozymeε are εelected for their ability to cleave a given target nucleic acid (e.g. , RNA) , and to inhibit the biological function of that molecule or any protein encoded by it. It is not necessary to know either the mRNA sequence of the target mRNA, or the intracellular localization of the ribozyme in order to select for active ribozyme constructε in thiε cellular εyεtem. Thiε in vivo

screening protocol (i.e. , one which takes place inside a cell) offers many advantages over extracellular εyεtemε, becauεe the synthesiε of large quantitieε of RNA by enzymatic or chemical methodε prior to aεεeεεing the efficacy of the ribozyme iε not neceεεary.

Applicant provideε an in vivo system for selection of ribozymes targeted to a defined RNA target. The system allows many stepε in a εelection process for desired ribozymes to be bypassed. In this syεtem, a population of ribozymeε having different εubεtrate binding armε (and thuε active at different RNA εequenceε) iε introduced into a population of cellε including a target RNA molecule. The cellε are deεigned εuch that only thoεe cellε which include a useful ribozyme will survive, or only those cells including a uεeful ribozyme will provide a detectable εignal. In thiε way, a large population of randomly or non-rando ly formed ribozyme molecules may be tested in an environment which is close to the true environment in which the ribozyme might be utilized as a therapeutic agent.

Thus, the method is for in vivo selection of a ribozyme active at a specific RNA target. In this method, a first population of ribozymes with differing substrate binding arms is introduced into a second population of cellε which includeε or iε later caused to include the RNA target, and those cellε which contain a ribozyme active at the target RNA identified.

Preferably, the ribozymes are encoded by an expresεion vector, and the method includeε cauεing thoεe ribozymeε to be expreεεed from the vector in the cellε. More preferably, the identified ribozyme iε iεolated and teεted to determine activity in vivo ; the RNA target iε a viral RNA and the identifying step includes determining cell survival after infection by the viral RNA; the RNA target is a mammalian RNA and the cells are mammalian cellε; the vector includes sequenceε encoding a 5' or 3', hairpin or a poly(A) tail for the ribozymes; the ribozyme

is a hammerhead ribozyme with at least one substrate binding arm having between five and eight nucleotides; and the identifying εtep includes detecting expreεεion of a reporter RNA attached to or regulated by the RNA target. More preferably, a ribozyme active at the RNA target iε recloned into an amphitrophic retroviruε vector, and a retrovirus vector having or encoding the ribozyme is selected for ribozyme activity active at the RNA target. Also provided is a population of nucleic acid molecules of the formula 5' NCNA 3', wherein each N is, or encodes, a ribozyme substrate binding arm, C iε, or encodeε, an enzymatic portion of a ribozyme, and A iε a polyadenylation sequence; wherein at least one of the Ns in the population differs in each vector. Preferably, the formula is 5' RNCNAS 3', wherein

R and S are restriction endonuclease siteε; the formula iε 5' RLNCNLALS 3', wherein each L iε nothing, or an inεertion region which is or encodes a nuclear procesεing εignal, an RNA εtability signal, a splicing signal, or another nucleotide signal which effectε tranεport or εtability of the ribozyme molecule; the formula is 5' DRLCNLALSE 3', wherein D and E are defined nucleotide sequences which can be used to prime amplification of the nucleic acid molecule, or the complement of which can be used to prime amplification of the molecule.

Also provided is a population of expression vectors each having one member of the population of nucleic acid molecules deεcribed above located within each vector, and a population of cellε each having one member of the population of expreεεion vectors present within each cell; and a ribozyme expresεion vector encoding a ribozyme and a hepatitis delta virus ribozyme motif, wherein the motif is able to cleave RNA including the ribozyme. A method iε provided for conεtruction of expreεεion vectorε for ribozymeε for uεe in their selection in eucaryotic cells. These vectors are εelected

for their ability to express inhibitory ribozymes; a pluripotent cell line which is preselected for the maintenance of DNA containing ribozyme expreεεion vectors; a method for the amplification of deεired ribozyme- expressing regions, and the use of these regions or their equivalent for the therapeutic, diagnostic or prophylactic administration of ribozymes; and a method for targeting accessible sites within an mRNA which can be treated with either chemically synthesized ribozymes or ribozyme- containing transcriptε.

Ribozvme Population

In general, to practice a method of thiε invention, it iε necessary to generate a population of RNAs that have a high probability of containing the active ribozyme that is sought. It is also necessary to provide a method for selecting such an active ribozyme out of that population of RNAs. If desired, the population of RNAs can be formed with stabilizing structures, such as 3' and 5' hairpins, to protect the RNA molecules from nucleases. In addition, the RNA may be capped or a poly(A) tail provided. The ribozyme to be tested should be as small as possible to ensure that the possibility of misfolding of the ribozyme, or secondary structure formation deleterious to enzymatic activity, is reduced. Thus, the ribozyme should be formed with as little flanking sequence as possible, and any such flanking sequences εhould be provided in a form that ensures that it forms a hairpin at the end of the ribozyme sequence.

The length of a substrate binding arm of the ribozyme should be optimized so that it is not so short that an optimum ribozyme can never be found, and not so long that any extra εequence might fold upon itεelf and prevent binding. Long substrate binding arms are also lesε useful because they are more laborious to randomize. A shorter binding arm permits use of a suitable size population of ribozymes to be formed, εcreened, and teεted. Subεtrate binding arms for a hammerhead ribozyme

having between six and eight nucleotides in length is optimal; thus, for a hammerhead ribozyme if the subεtrate binding arm iε eight nucleotides in length, with one invariant nucleotide at the 3' end of stem I, then only 15 positionε need to be randomized to provide about IO ⁹ ribozyme permutationε. Vector Svεtem

Once the generic εtructure of the ribozyme is chosen the position and identity of nucleotide variations can be selected by standard procedure. Generally, such ribozymes will be provided in a plasmid or other vector system. For example, such a vector may be a plasmid having the human cytomegalovirus immediate early promoter region to give strong constitutive expreεεion or an inducible promoter to allow control of expression of the ribozyme. The promoter may be followed by a hairpin structure, a ribozyme-encoding DNA sequence and a minimal T7 terminator hairpin to provide stability at the 3 ' end of the encoded ribozyme transcript. The 3 ' termination sequences or a hepatitis delta virus "cis-cleaving" ribozyme may be inεerted after the 3' terminal hairpin structure to create the 3' terminuε of the ribozyme transcript. Such a plasmid may be provided with a suitable restriction endonuclease site(s) to allow ready insertion of a population of random or otherwise synthesized ribozymes at the appropriate location. Care must be taken to ensure that such sequences do not interfere with ribozyme activity, which can be accomplished by including such sequences in a defined hairpin εtructure. Alternatively, the restriction endonuclease sequences can be omitted and cloning can be achieved by standard PCR techniques.

In order to aid in generation of a population of ribozymes, it may be preferable to make a series of libraries that differ in the length of the substrate binding arms, and then mix those libraries together prior to transformation of a test cell. For example, 16

different libraries can be formed which have randomized subεtrate binding arms from 4 to 8 nucleotides on each side.

Selection The selection εtrategy may take many formε.

Preferably, the RNA that iε targeted for cleavage εhould be in a configuration that iε aε close to "natural" as posεible, so that the natural state of the target site will be accessible. If possible, selection should be carried out against full-length target RNA. with no foreign flanking sequence. The target RNA also preferably is expressed in a cell that is its natural environment. For example, if the RNA is a mammalian RNA, a bacterial expreεεion cell should be avoided since it is unlikely to contain the splicing factors and nuclear transport factors that bind to natural target eukaryotic mRNA. The selection system may have to be specifically chosen for each specific RNA target.

If the RNA target is a viral RNA, then the selection system can be devised to select for survival of a host cell after infection by that viral RNA. That is, a population of ribozymes are introduced into a host cell and thoεe cellε infected by the viruε. Any cell which containε a ribozyme able to cleave the viral RNA will allow that cell to survive and grow. Thus, any growing cell potentially contains a desired ribozyme. It may be difficult to identify which viral gene is being cleaved by the selected ribozyme. While it may be important at some point to identify the target gene, it does not really matter which gene is targeted aε long as the ribozyme actually workε.

For other target RNAε it iε possible to attach a reporter RNA to the end of the target RNA and to select for reduction in reporter-gene activity after ribozyme introduction. That is, a reporter DNA is ligated to DNA encoding the RNA target so that the two RNAs are produced as one RNA molecule. Cleavage of the RNA target can be

deεigned to either activate or inactivate expreεεion of the reporter RNA. If it is activated then a signal iε provided in the cell (e.g. , an enzyme activity such as β- galactoεidaεe) ; if it iε inactivated then the εignal may be cell εurvival when the RNA encodeε a toxic cell product (e.g. , cholera toxin) . Thiε method permitε ribozyme εelection for one particular target RNA.

Any εelected ribozymeε will need to be εcreened to enεure that the reporter RNA iε not the target of the ribozyme attack. In addition, they will need to be εcreened to enεure that the preεence of any reporter RNA did not alter acceεεibility to the target RNA. That poεεibility can be readily checked in a εubsequent screen involving the target RNA not attached to the reporter RNA. If the RNA encoded by a regulatory gene is the

RNA target, it is posεible to place a reporter gene under the control of the target gene product. In that way a ribozyme can be selected against one particular RNA without worry about reporter RNA disrupting target RNA accesεibility.

There followε examples useful in thiε invention by way of illuεtration. Theεe exampleε are not limiting in the invention. Example 27: Syntheεiε and Conεtruction of Random Ribozyme Coding Insert

The random ribozyme coding insert can be constructed by any number of methods which are appreciated by those skilled in the art. The following is but one method of synthesis which illustrateε various essential and non-essential features of the insert.

The DNA fragments used for the insert are synthesized on an ABI 38OB synthesizer according to the protocols supplied by the manufacturer. For the insertion of random nucleotides, a bottle which contain equimolar mixtures of monomers is sequentially εampled and the coupling and deprotection timeε are increaεed. Theεe precautions will ensure the addition of the correct

numbers of bases at these critical points in the construct. After syntheεiε, the DNA εequence of the mixture is end-labeled with ³² P and analyzed by Maxam and Gilbert analysis, or primed with radiolabeled complementary fragments which are extended and sequenced using a dideoxy sequencing method.

Referring to Fig. 1, there iε shown one example of a ribozyme of the general hammerhead structure. The ribozyme population desired to select a useful ribozyme, however, has a randomized sequence in the regions which base pair with the target, i.e. , the substrate binding arms.

The DNA sequence of the random ribozyme coding εtrand iε εyntheεized first. One example is shown in Fig. 34, where D _π is any defined nucleotide sequence which can be used to give effective priming for amplification of the sense strand of the DNA fragment; E, ₇ is the inverse complement of a defined nucleotide sequence (F ₁₇ ) which can also be used to prime synthesiε, but thiε time primeε synthesis of the antisense strand of the fragment; R and S ₆ are restriction endonuclease recognition sequences for RE1 and RE2 enzymes, respectively; CUGAUGAGGCCGAAAGGCCGAAA is just one example of a core hammerhead ribozyme sequence for a hammerhead motif ribozyme; N ₈ is a random substrate binding arm sequence of 8 nucleotides in length; AAAAAAAAAAAAAAA,. is a polyadenylation region of undefined length; and each LLL iε an optional inεertion region which may be uεed to add nuclear proceεsing signalε, RNA εtability εignals, splicing signals or other such nucleotide signals which may effect the transport or stability of an expresεed ribozyme molecule. Terminal εequenceε are chosen to have minimal homology between the 3' end of D ₁₇ and the 5' end of E _π . RE1 and RE2 are any restriction endonuclease enzymes required for subεequent cloning procedures. The exact lengths of any of these regions may vary in different constructε, but the general

structure and arrangement of the various regionε iε uεeful in thiε invention.

Becauεe the terminal 3' nucleotide of the oligonucleotide iε attached to the support for εyntheεiε, it would be preferable to begin with an A-containing CpG bead and then extend the next 8 nucleotideε of the 3' random binding armε, proceed with the core ribozyme (enzymatic) εequence, the 8 nucleotideε of the 5' random binding armε, a restriction enzyme (RE1) recognition sequence (R ₆ ) , and end with a primer εequence (D ₁₇ ) at the 5' terminuε. Capping, uncoupling and deprotection would proceed by εtandard procedureε.

The 3 ' termini of the random ribozyme coding sequences can be polyadenylated using terminal deoxynucleotidyltranεferaεe according to standard protocols. After the addition of the poly(A) tails at the 3' termini of the DNA fragments, the fragments can be partially purified by extraction, heat denatured and annealed to a nucleotide primer composed of the F sequence at the 5' end, followed by a predetermined 6 nucleotide RE2 recognition sequence (S ₆ ) and terminating in a (dT) _2. s εequence.

Second strand syntheεis is achieved by the addition of Taq DNA polymerase, appropriate saltε and nucleotide εubεtrateε, and incubation for 16 minuteε at 65°C. The reεultant mixture containε double-εtranded random ribozyme inεertε which are compoεed of blunt-ended termini. The coding εtrands of these fragments contain the D, ₇ primer sequence, followed by the RE1 site, the random ribozyme coding εequence, a variable length poly(A) region, the RE2 εite and an annealing region at the 3' terminuε for the F ₁₇ primer. Theεe double-stranded fragments are amplified using Taq polymerase and appropriate DNA primer fragments (D ₁₇ and F ₁₇ ) prior to cloning into a desired vector.

Before ligation into the expresεion vector, the DNA mixture iε digeεted with the reεtriction enzymeε to

create fragments containing defined 5' and 3' termini to allow cloning into the vector in a directional manner. The mixture iε phenol extracted and purified by column chromatography to deεalt and remove undeεirable nucleotide fragmentε which are cleaved from the termini. Alternatively, the termini may contain a blunt end and εubεequent cloning into an expression vector can be achieved after treating the DNA fragments with Klenow polymerase according to establiεhed protocolε. If desired, the polyadenylation sequence may be omitted, and the restriction enzyme primer or blunt end primer-containing regions complementary to the 3 ' regions of the random ribozyme coding strands can be used to initiate the synthesiε of the εecond DNA εtrand. If either of the latter approacheε are preferred, then the initial randomized ribozyme coding sequences should contain a common DNA sequence of 15-25 nucleotides at the 3' termini to facilitate second strand synthesis. This common DNA sequence may be complementary to a subset of the random sequences and may limit the utility of the described invention by base pairing with the ribozyme binding sequences and preventing the proper formation of ribozyme/substrate interactions.

It may be appreciated that the generation of random sequenceε in the binding regions of the ribozymes will also generate restriction endonuclease cleavage sites used to clone the DNAs into the vectorε. Cleavage with the cognate reεtriction endonuclease may therefore reεult in the loεε of any ribozyme constructε containing the given reεtriction enzyme recognition εite. For thiε reaεon, each generation of conεtruct librarieε uεeε at leaεt two diεεimilar reεtriction enzyme sequences at the 5' and 3' termini (e.g.. recognition sequenceε for Sail and PvuII are GTCGAC and CTGCAG, respectively) . A serieε of fragments containing .EcoRl sites at the 5' ends and either PvuII or Sail siteε at the 3' ends, aε well aε fragmentε containing EcoRl εiteε at the 3' endε and either

Sail or PvuII sites at the 5' ends creates ribozyme coding fragments with enough degeneracy in the coding regions to represent all posεible nucleotide sequences in the expreεεion vectorε. Example 28: Conεtruction of Ribozyme Expreεsion Vectors A number of suitable vectors for the expresεion of mRNA constructs in tisεue culture and animals have been described in the literature, and may be used as effective vectors for the therapeutic expresεion of ribozymeε. Stably transformed cell lines present the best test system for establishing parameters of expression which will be necesεary for the development of optimal ribozyme- expreεεing vectorε. Fragmentε of DNA can be eaεily introduced into theεe types of cells (e.g. , HeLa cells) and selected for the maintenance of expression from the new fragments of DNA (e.g.. by means of using the co-expression of a second gene which lends a selective advantage to the survival of the cell) . Thus, for the effective construction of ribozyme expresεion vectors, a number of features may be built into the initial plasmid DNA construct; these include an eucaryotic transcriptional enhancer/promoter complex located immediately 5' to the ribozyme coding sequence, the presence of a self-cleaving (e.g.. hepatitis delta virus) ribozyme motif located immediately 3' of the ribozyme coding sequence, the presence of a second coding sequence whose transcription is controlled by eucaryotic promoter and polyadenylation signal sequenceε, a bacterial origin of replication and an antibiotic reεiεtance gene. The bacterial origin of replication and the antibiotic reεiεtance geneε are preεent to aid in growing the plasmid constructε prior to introduction into cell lineε.

Many human transformed cell typeε express transacting factors which are capable of activating the SV40 early promoter region. For this reason, it is preferable to use an expression vector which contains the selection gene under transcriptional control of the SV40

early promoter. In cell types where the appropriate transcriptional activators are not ordinarily present, certain chemicals can be added exogenously to activate this promoter. Another preferred promoter is the metallothionein promoter which can be induced to activity by the addition of either glucocorticoids or heavy metal salts to the cells. Additional promoters which can be induced by endogenous or extracellular substances which are readily available to the cells are conεidered aε alternative pro oterε in this invention. The ribozyme coding εequenceε may alεo be induced using the same promoter sequences as the selective gene, but the ribozyme construct can also utilize any promoter which is activated by a deεired target disease condition (e.g.. ribozymes targeted to inhibit specific viral gene expresεion, such as the herpes simplex virus ICP27 gene may utilize the same promoter as the viral gene, e.g. , the ICP27 promoter) . The preferable situation would involve a ribozyme gene operating under the promoter control of a very strong constitutive promoter such as the human cytomegalovirus immediate early region promoter (CMV iel promoter) or any of a number of housekeeping gene promoters (e.g. , beta actin promoter) . It is appreciated that many preferred promoters contain enhancer sequenceε within the DNA sequences required for optimal promoter activity. It is preferable to utilize promoterε which contain natural enhancer sequences because the addition of certain enhancer-like regions upstream of some promoter regions may suppreεε the promoter activity. Because extra nucleotides outεide of the catalytic region in the ribozyme may cauεe the formation of εecondary structures which would inactivate the enzymatic function of the molecule, the vector construct does not always contain intramolecular signaling sequences which are necesεary for the correct polyadenylation of the tranεcript or termination of transcription. The poly(A) sequence which waε introduced into the ribozyme construct

will substitute for the polyadenylation εignal, but the termination εignal iε an undefined εequence which iε thought to be no leεs than 25 nucleotideε in length. Instead of the termination signal sequence, each vector construct can have a εelf-proceεsing (e.g. , hepatitis delta virus) ribozyme sequence incorporated immediately 3' of the restriction enzyme cloning site at. the 3 ' terminus of the ribozyme coding fragment deεcribed previously. Thiε εequence iε able to cleave the ribozyme tranεcript aε transcription is proceeding past the region and will function aε an artificial tranεcription terminator for the ribozyme tranεcript. The hepatitis delta ribozyme cleavage will leave one extra nucleotide 3' of the restriction enzyme insertion site (RE2) . With certain restriction enzyme recognition sequences, it is posεible to engineer the hepatitiε delta ribozyme to cleave within the reεtriction εite or at the 3'-terminus of the poly(A) stretch. The appropriate hepatitis delta virus (or hairpin or other ribozyme) ribozyme can be engineered into the expression vector immediately 3' of the RE2 site by standard insertion mutagenesis techniques.

After the appropriate expresεion vector is constructed and the random ribozyme insert iε in place, the constructs must be stably introduced into cell lineε for expreεεion. The εtable integration and expreεsion is driven by the selective presεure imposed upon the cells by treatment with toxic chemicals, such as G418 (gentamicin) or hypoxanthine/puromycin/thymidine (HPT) mixtures. The expression of neomycin resistance or thymidine kinase, respectively, will allow the transformed, expressing cells to survive and simultaneouεly afford the εurviving cellε the DNA needed to express transcripts from the random ribozyme genes introduced during transformation. Example 29: Stable Introduction of Ribozyme Expreεεion Vectors Into Cell Lines

A number of methods to stably introduce DNA into cells are known in the literature, these include but are

not limited by Ca ₂ P0 ₄ precipitation, DEAE dextran precipitation and electroporation. Generally, one microgram of DNA is applied to a 100 mm tisεue culture dish containing subconfluent cultures of cells (approximately 3-4 million cells) . Initially, at least 10 disheε will be transfected for each ribozyme DNA preparation. After an appropriate time is given for DNA uptake, the cellε are rinsed in buffered medium and then refed with supplemented medium. In some cases, it may be preferable to include a glycerol shock step or the addition of inducing chemicalε (e.g.. glucocorticoid hormoneε or heavy metalε to induce metallothionein promoter expreεεion) to initiate the εyntheεiε of εelective gene products (e.g. , thymidine kinase or neomycin resistance protein) .

After 10-18 hours, the cells are trypsinized and replated in selective medium containing the toxic chemicals used for selection (e.g. , 250-350 μM G418 in HeLa cells) against viable non-transfected cellε. Thiε medium may also contain chemicals required for the continued expression of any selective gene products necesεary for cell survival. The continued presence of selective gene induction in the surviving cells may be required to maintain the ribozyme-containing plasmid because the transfected DNA iε uεually inherited aε a unit which iε maintained under the selective presεure uεed in the protocol. Often, when the εelective preεεure iε removed, the foreign DNA is lost or deregulated in transformed cells. Because the cellε will be maintained in the preεence of chemical agents, care should be taken to avoid the use of agents which will interfere with subεequent ribozyme activity or introduction of infectiouε agentε into the cell εyεtem.

Selected cellε will be grown with medium changeε every three dayε until confluent cultures are obtained. As the cells fill the tiεεue culture veεεelε, they will be trypεinized and tranεferred to larger veεεelε until

cultures of between IO ^8"9 cells can be harvested for the preparation of frozen stockε. It iε eεtimated that if each cell contains 1 nanogra of DNA, this will be 500 different ribozyme conεtructε if the average size of the plaεmidε iε 7000 nucleotideε. If the random ribozyme armε contain eight flanking nucleotideε at each end for hybridization, thiε will give 4 ¹⁶ or 4.29 x 10 ¹⁰ unique molecules. A total representation of these molecules should easily be found in 10 ⁸ cellε if no more than 100 ribozyme conεtructε per cell iε realized during the tranεfection/εelection procedureε. The expreεsion of 4 ¹⁶ ribozymes within this cell culture εhould give εequenceε capable of hybridizing and cleaving virtually any unique mRNA εequence which can be targeted. It haε been eεtimated that there are only 4 ¹² unique RNA sequences expressed from the human genome. Clearly, any infectious agent would express a great deal less unique RNA sequences than can be targeted with these pluripotent ribozyme- expressing lineε. Example 30: Selection of Therapeutic Ribozy e-Expreεεing

Cellε There are many wayε in which cellε expreεεing therapeutically active ribozymeε can be selected. In every case, the cells which express the desired ribozymes must be εcreened, cloned, and expanded for εubcloning of the constructs into new vectors which allow the introduction of single copies of vectors into new cell lineε. An example of an expreεεion/εelection protocol iε given. For thiε example, a ribozyme vector which iε tranεcriptionally controlled by the HCMV iel promoter was used. This promoter is a very strong constitutively acting promoter which is controlled by transcriptional factors which are present in all cells. By using this promoter or any εimilar promoter, preinduction of tranεcription by hormoneε or chemicalε iε generally not required. Additionally, the HCMV iel promoter iε

tranεcriptionally activated by many factorε which control genetic expression during infection of cells with heterologous viruεeε (e.g. , the iel promoter iε activated by the ICP4 and ICPO proteinε which control expreεεion of herpeε simplex virus [HSV] genes) . Selective presεure which will kill all but the ribozyme expreεεing cellε iε presented as a viral infection (e.g.. HSV infection of HeLa cells) . This infection should be performed at a tissue culture infectious dose which will kill approximately 50% of the cells in one replicative cycle. The cells are treated with viruε for three infectiouε cycleε of the viruε (three dayε for HSV-1) and then the RNA iε harveεted from a portion of the εurviving cellε to verify the expreεεion of ribozyme molecules from the transcription vectors. Surviving cells are expanded for the preparation of frozen stocks. Stocks are prepared by standard methods. A culture of the resistant cells is grown for the preparation of DNA and a verification that the cells are resiεtant to infection with the target virus (e.g. , verification of reεiεtance will be determined by reinfecting the εurvivorε with HSV-1) .

Example 31: Recloning of Ribozyme-Expreεεing Vector DNA Reεistant ribozyme-expressing cells selected by protocols described in the previous section will be expanded for the purpose of extracting the ribozyme casεetteε which are expreεsed in these cells. Total cellular DNA will be isolated and purified from the cells by standard methods. The DNA preparation will be subjected to restriction enzyme digestion using enzymes capable of recognizing siteε preεent in the expreεεion vectorε which are at leaεt 50 nucleotideε 5' and 3' of the ribozyme cassetteε. By uεing thiε approach, the DNA εequenceε present upstream and downstream of the ribozyme casεettes are used to hybridize primer DNA fragments. After primer hybridization to the restricted DNA preparation, the ribozyme cassettes can ba amplified by incubation with appropriate salts, nucleotide precursorε

and Taq polymeraεe (United Stateε Biochemical Corp.) and incubation at 65°C. The reεultant amplified DNA fragmentε can be electrophoretically separated from the other DNA fragments (unwanted cellular DNAs) in 2% agarose or other suitable matrix. The DNA casεetteε are releaεed from the amplified DNAε by digeεtion with restriction endonucleases which recognize the REl and RE2 sites used in the original cloning. A second agarose gel iε run to purify the deεired caεsettes from the unwanted ends of the amplified DNA which correspond to expression vector sequences. The ribozyme casεetteε are then cloned into appropriate amphitrophic retroviruε vectorε using standard protocolε.

Theεe vectorε εhould contain an inducible/εelectable gene εuch aε neomycin resistance and restriction enzyme cloning siteε REl and RE2 which are downεtream of an inducible or strong constitutive promoter. A protocol for packaging the recombinant DNAs into amphitrophic virus particles and transfection into cells is well known. Infection and selection of cells with these vectors will result in the incorporation of only one ribozyme vector per cell.

Example 32: Reselection for Therapeutic Ribozv e- Expressing Cassettes After cell lines are prepared which contain ribozyme expreεεion caεεettes within retrovirus vectors, the cells are again selected for the expression of ribozymes which protect the cells from infectious agents, or prevent expression of an undesirable genetic trait. The selection protocol is similar to that described for the first selection (i.e. , infection with HSV-1 at appropriate concentrations of virus, growth for three viral replicative cycles, and clonal expansion of any surviving cells) . After stock cells are prepared, the cells are rechecked for viral resistance and the retroviral vector is induced to replicate by standard methods (e.g. , chemical treatment) . Viral particles are isolated and the sequences of the therapeutic ribozymes

are determined by direct εequencing of the viral genomeε. Molecular methodε of inducing retroviral replication, isolation of particles and sequencing of the genomes are known to those in the art. Once the εequence of a therapeutic ribozyme haε been determined by the methodε deεcribed herein, the ribozyme can either be incorporated into an expreεεion vector for in vitro synthesis, or genetic therapy, or the ribozyme can be syntheεized chemically for use with exogenous delivery techniques. Descriptions of useful systems are provided in the art cited above, all of which is hereby incorporated by reference herein. Quasi random methodε

The following are examples methods of cloning quasi-random ribozymes into expresεion vectorε, in vitro assays for optimizing ribozyme structural features which yield favorable catalytic activity, and methods of cloning concatameric ribozymes into an expresεion vector.

The techniqueε deεcribed below may be uεed to find effective therapeutic ribozymeε to any deεired target RNA even if the exact location of the tranεcript within a gene iε uncertain. Methodε are deεcribed whereby the selected ribozyme may be optimized for catalytic activity using cellular extracts for in vitro assays. Four vectors are described which allow the cloning of quasi-random ribozymes into plasmids and facilitate the excision of selected ribozyme cassettes for cloning into a eucaryotic expression vector for intracellular asεays of ribozyme cleavage activity. A method iε described which allows for the cloning of concatameric ribozyme caεεetteε into expreεsion vectors and thereby resultε in a more effective ribozyme-expressing vector for uεe in cellular applicationε of genetic or therapeutic interest. Cloning of Quasi-random Ribozvmeε Four novel cloning vehicles were constructed for the generation of quaεi-random ribozymeε. The herpeε εimplex viruε thymidine kinaεe (TK; Wagner et al, 78 Proc.

Natl. Acad. Sci. USA 1441, 1981) gene waε uεed aε a prototype gene target for the cloning. Methodε for the generation of the cloning vehicleε, inεertion of the TK DNA fragmentε and incorporation of a hammerhead ribozyme motif into the conεtructε follow. Uεe of thiε method allowε one to clone target-specific ribozymes from any fragment of DNA, even when the primary DNA sequence or direction and dimensions of mRNA coding regionε are uncertain. The subsequent use of biological selection and amplification of the desired cloneε will generate ribozyme constructs which have potentially advantageouε genetic or therapeutic activities.

The basic requirement of the cloning vector in this protocol is that the restriction endonuclease site located immediately 5' of the cloned target DNA sequences be recognized by a restriction enzyme which cleaves the target DNA at a distance (x nucleotides) 3' of the recognition sequence. The distance (x) between the recognition and cleavage sites (along with the size of DNA fragments selected for cloning) will dictate the length of the ribozyme binding arms in the quasi-random constructs. The cloning of hammerhead ribozyme catalytic sequences into the inserted DNA fragments is facilitated greatly by using restriction enzymes that cleave to yield overhanging 3 ' ends which are 2 nucleotideε in length. The cloning protocol deleteε these two bases and insertε a ribozyme catalytic core at the εite of the reεtriction enzyme cleavage. Uεe of reεtriction sites which give 5' overhangs, blunt ended fragments or 1 nucleotide 3' overhanging ends result in the generation of flanking sequences which no longer reflect the original DNA sequence or ribozymes which do not allow for the unpaired base (M) in the RNA substrate which is required for effective hammerhead ribozyme cleavage. See Figs. 49 and 50.

Cloning sites were constructed so that the ribozyme expresεion cassetteε could be exciεed from the

original plasmid vectorε and inεerted into eucaryotic expreεεion vectorε. Because restriction endonuclease siteε will exist in the random fragments of target DNA, multiple releasing and core insertion cleavage sites should be available for each series of constructε. By uεing constructs with different releasing and core insertion restriction endonuclease sites, one can be relatively confident of acquiring all random ribozyme recognition sites with only four cloning vehicles. Two cloning inserts were designed for use in PGEM4 and pSP72 (Promega Corp; Madison, I) vectors. For the PGEM4 vectors, Bpml/Smal and Bsgl/Smal cloning sequences were flanked at the 5' and 3 ' ends by findlll and _BcoRI restriction endonuclease sites, respectively. For the pSP72 vectors, the 5' and 3' flanking restriction endonuclease sites were Bgrlll and Xhol, respectively. Each of these vectors contains a T7 and SP6 promoter region at the 5' and 3' terminus of the cloning sites, respectively. The phage RNA polymerase promoters allow in vitro transcription of the ribozyme inserts and testing of ribozyme activity prior to the insertion into eucaryotic expression vectors.

The four pairs of DNA oligonucleotides (Midland certified Reagent Co.; Midland, TX) used to clone into the PGEM4 and pSP72 vectors were:

(A) 5' AGCTTCTGGAGCCCGGG 3' 3' AGACCTCGGGCCCTTAA 5'

(B) 5' GATCTCTGGAGCCCGGGC 3' 3' AGACCTCGGGCCCGAGCT 5'

(C) 5' AGCTTGTGCAGCCCGGG 3'

3' ACACGTCGGGCCCTTAA 5'

(D) 5' GATCTGTGCAGCCCGGGC 3 ' 3 ' ACACGTCGGGCCCGAGCT 5'

Pairs A and C were cloned into PGEM4 and pairs B and D were cloned into pSP72. To incorporate the cloning inserts into the plasmids, 10 μg of each oligonucleotide was combined with its paired oligonucleotide in a 1:1 ratio and incubated in TMN solution (50 mM Triε-HCl, pH 8; 10 mM MgCl ₂ , and 100 mM NaCl) at 90°C for 2 minuteε, then slowly cooled to 25°C. One-tenth (2 μg) of the annealed mixture was ligated to the appropriate linearized vector fragments. Prior to addition to the ligation reaction, vector DNA was digested with the appropriate restriction enzymes (Hindlll/EcoRI for PGEM4 and Bglll/Xhol for pSP72) and the linearized fragments were isolated from low melting point agarose as described by Maniatis et al, 1982, Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Ligations were performed at 14°C for 18-24 hours in 20 μl total volume. Buffer was as suggested by the manufacturer (United States Biochemical Corporation (USB) , Cleveland, OH) . After ligation of inserts into the vectors, bacteria (HB101) were transformed at 37°C for 20 minutes with approximately 100 ng of vector DNA (one-tenth of the reaction volume, 2 μl) . Colonies of bacteria were selected on agarose plates containing ampicillin and then amplified in 3 ml cultures for identification of appropriate cloning vehicles. The presence of cloning inserts into PGEM4 and pSP72 vectors was analyzed by restriction endonuclease digestion of the plasmid DNA preparations. Samples of appropriately transformed bacteria were identified and 250 ml cultures of the bacteria were grown. Plasmid DNA was purified from the 250 ml cultures using QIAGEN columns (QIAGEN, Inc.; Chatsworth, CA) and protocolε εuggeεted by the manufacturer. Each of the four plasmid vectors generated in this manner were used in a similar protocol to clone random DNA inserts of the herpes simplex virus thymidine kinase (TK) gene.

For cloning of the TK gene fragments, 50 μg of the TK DNA (2.7 kb .EcoRl fragment N) was excised by

EcoRI/Scal digestion of plasmid DNA containing this fragment. The Seal digestion was incorporated to remove the plasmid bands of pSP72 which iε also 2.7 kb in length. .EcoRl fragment N waε iεolated by electrophoreεiε in low melting point agarose and subsequent elution by standard protocols (Maniatis et al., 1982). After isolation and concentration by ethanol/salt precipitation, the EcoRl fragment N DNA was fragmented by limited digestion with DNasel. The digestion was performed by combining 20 μg of TK DNA and 2 units of DNasel (USB) in 200 μl of total volume containing 50 mM Tris-HCI, pH 7.5; 10 mM MgCl ₂ , and 10 μg bovine serum albumin (BSA) . The mixture was incubated at 37°C and 50 μl aliquots were removed and combined with 5 μl of 0.5 M EDTA after 1, 2.5, 5 and 10 minuteε. All reaction aliquotε were electrophoresed in 8% polyacrylamide gels to size fractionate the pieces. Fragments 20-40 nucleotides in length were cut from the gels and eluted according to the methods of Maniatis et al., 1982. The digestion protocol is dependent upon the base composition of the DNA and should be optimized for each fragment which is to be digested. We found that the optimum digestion time for the HSV TK gene was approximately 2.5 minutes. The termini of the eluted DNA fragments were repaired with Klenow fragment of DNA polymerase to generate blunt ended fragments for cloning into the plasmid cloning vehicles described above. The repair reactions were performed in 20 μl volumes at room temperature for 1 hour according to protocols described by the manufacturer (USB) . After repair of the termini, the DNA fragments were dephosphorylated by incubation with calf intestinal phosphatase enzyme (CIP, United States Biochemical Corp.) and the fragments were prepared for insertion into cloning vectors by extraction in phenol/chloroform/isoamyl alcohol mixtures (24:24:1) and ethanol precipitation. Ten micrograms of each cloning vehicle was linearized by digestion with Smal restriction endonuclease and the

linearized fragments were electrophoresed in 1% low melting point agarose gels to remove the fraction of undigested vectors. Elution from the agarose was aε described by manufacturer. One microgram of vector was combined with one microgram of DNA fragment preparation (1:100 molar ratio of vector to insert, respectively) and then fragments were ligated in a total volume of 20 μl at 25°C for 18-24 hour. Prior to transformation of bacteria, the ligation mixture was digested with Xmal restriction endonuclease (5 units of enzyme in 20 μl volume for 3 hourε) to relinearize vector pieceε which had cloεed during ligation. (Xmal iε an isoschizimer of Stnal . ) The digested preparations were then transformed into HB101 cellε aε described earlier. The resultant populations of ampicillin resistant bacteria were grown in a batch culture to recover all plasmid vehicles for the insertion of the ribozyme catalytic core sequences into the TK DNA fragments.

Prior to cloning the ribozyme catalytic core into the TK DNA fragments, plasmids were linearized by digestion with appropriate enzymes (Bpml or Bsgl; New England Biolabs; Beverly, MA) and the linearized plas idε were isolated after electrophoresis in low melting point agarose gels. The linearized preparations were digested with T4 DNA polymerase to generate blunt ends and the DNAs were dephosphorylated by treatment with CIP enzyme. After phenol/chloroform/isoamyl alcohol extraction, the DNA fragments were precipitated with ethanol and reεuεpended in TE buffer (10 mM Tris-HCI, pH 7.5; 1 mM EDTA). The catalytic core sequence was inserted into the linearized plasmid DNA by annealing the following oligonucleotides as described above:

5' CTGATGAGGCCGAAAGGCCGAAA 3' 3' GACTAACTCCGGCTTTCCGGCTTT 5' After annealing, 0.1 μg of the oligonucleotide mixture was added to 1 μg of the blunt-ended, linearized plasmid preparation (gives a 10:1 molar ratio of insert:vector

sequences) in ligation buffer to give a final reaction volume of 20 μl. The mixture was incubated overnight at 16°C and then one-tenth of the total volume (2 μl) was added to 100 μl of transformation solution containing HB101 cells. After transformation of the cells, 10 ml preparations of the bacteria were grown and plasmid DNA was isolated from 5 ml of the preparation as described in a previous section. The remaining 5 ml of bacteria was divided among three tubes, combined with an equal volume of 40% glycerol and frozen at -80°C for stock solutions.

Use of the quasi-random ribozyme cloning generates gene fragments and catalytic cores which are present in both senεe and antisense orientations. Because ribozymes will be created which cleave transcripts from both strands of the DNA fragment, this mixture of inserts is advantageous when cell selection of ribozyme activity is sought and the direction of transcription within the target DNA is unknown. The random insertion method should result in a 1:1 mixture of sense strand and complementary strand cleaving ribozymes. The presence of RNA polymerase promoterε at both endε of the DNA insertions allows versatility in the choice of transcription protocols (because roughly equal distributions of ribozyme structures and specificities should be obtained with either series of transcripts) and compensateε for non- random directional insertions into the vector.

For a cursory examination of catalytic activity within the quasi-random ribozyme populations, in vitro tranεcriptε were εyntheεized from the plasmid DNA templates prepared above. The plasmid DNA preparation was split into two fractions (the yield of DNA was approximately 10 μg in each fraction) for digestion with restriction endonucleaseε and in vitro transcription with either T7 or SP6 RNA polymerase. The PGEM4 based plasmids were cleaved with EcoRI for T7 transcriptions and with Hindlll for the SP6 transcriptionε. The pSP72 baεed plaε idε were cleaved with Xhol for T7 transcriptions and

with Bgrlll for the SP6 transcriptions. One microgram of linearized DNA template was added to a transcription mixture containing 20 μl of 5x buffer (200 mM Tris-HCI, pH 7.5; 50 mM MgCl ₂ ; 10 mM spermidine and 50 mM NaCl), 10 μl of 100 mM dithiothreitol (DTT); 5 μl of RNAsin (20,000 U/ml; GIBCO/BRL, Gaithersburg, MD) ; 5 μl of NTP mixture (20 mM each of ATP, GTP, CTP and UTP) and enough water to bring the final volume to 95 μl. 100 units of T7 or SP6 RNA polymerase (USB) was added and the mixture was incubated at 37°C for 1 hour, after which RNase-free DNasel was added and the reaction waε incubated an additional 15 minutes at 37°C. The solution was brought to a volume of 150 μl, extracted once with phenol/chloroform/isoamyl alcohol and εubjected to εpin chromatography through a 1 ml bed volume of G-50 Sephadex (Pharmacia, Upεalla) . A typical tranεcription yieldε 10- 15 μg of RNA per μg of DNA template.

For the εynthesis of ³² P-labeled TK substrate RNA, template DNA was prepared by PCR amplification using primers

5' TAATACGACTCACTATAGGGACCGAGCGACCCTGCAG 3' and 5' GCCGTCATAGCGCGGGT which resulted in the generation of double-stranded DNA template containing a T7 RNA polymerase promoter region located immediately 5' of the TK mRNA coding sequences (1262 nucleotides) . The transcription reaction was essentially as described for ribozyme synthesis except that the total reaction volume was decreased to 20 μl, contained only 20 units of T7 polymerase and the 1 μl of 20 mM NTPs was substituted with 1 μl of 20 mM ATP, GTP and UTP, 1 μl of 2 mM CTP and 2 μl of α- ³² P-CTP (80 μCi; 800 Ci/mmol; Dupont/NEN; Boston, MA) . TK RNA was purified by εpin chromatography through G-50 Sephadex columnε. When ill-defined target DNA iε uεed aε the source of insert fragments, template for the transcription of target RNA may be prepared by blunt-end ligation of the DNA into pSP72, PGEM4 or other suitable transcription vectors.

For analysis of ribozyme cleavage of the TK transcript, the in vitro transcripts were mixed with 50,000 cpm of in vitro transcribed TK RNA and incubated at 37°C in a total volume of 50 μl containing 5 μl of lOx ribozyme cleavage buffer (750 mM Tris, pH 7.5, and 100 mM MgCl ₂ ) and a final ribozyme concentration varying between 0.1 and 10 μM. Control reactions contained 10 ng of yeast tRNA instead of the ribozyme mixture. At given time intervals, aliquots were removed and quenched by addition to a εolution containing (20 mM EDTA, 95% formamide and 0.1% Bromphenol Blue and Xylene Cyanol) . Sampleε were heated at 90°C for 10 minutes, snap cooled in ice and cleavage products of the TK RNA were identified by electrophoresis of the samples in 3% polyacrylamide gels and subsequent visualization of cleaved RNA fragments by autoradiography. DNA preparations whose transcriptε exhibited ribozyme activities specific for the TK RNA were amplified by growing 250 ml cultures of the parent bacterial stocks. The plasmid DNA isolated from these stocks was digested with either HindiII /EcoRI or Bglll/Xhol restriction enzyme combinations and the released ribozyme coding sequences were isolated by gel electrophoresis and cloned into appropriate eucaryotic expresεion vectors for use in tissue culture experiments. The εcreening and identification of ribozyme vectorε which cleave target RNA intracellularly is tedious and depends upon the choice of target mRNA whose expresεion lends a negative selective advantage to the cellε (e.g. , HSV TK expreεεion will render cells sensitive to treatment with acycloguanosine [ACV]). By growing TK- expreεεing cellε in the preεence of ACV and ribozyme- expreεεion vectorε, one may select and expand only the cells containing inhibitory ribozymes. The DNA encoding effective ribozymes may be extracted and PCR amplified by standard methods because the flanking regions within the eucaryotic expresεion vectorε will be known. After identification of effective ribozyme εequenceε, the

ribozyme εtructures may be optimized by the following in vitro methods prior to insertion into appropriate therapeutic expreεεion vectorε.

In Vitro Ribozyme Optimization An adjunct approach to cell-baεed εelection of effective quaεi-rando ribozyme constructs useε a modification of the in vitro ribozyme activity aεεay described in the previous section. When uniformly labeled transcript is used to check ribozyme activity, multiple cleavages within the transcript prevent the identification of susceptible RNA structures. For any given quaεi-random population of ribozymes, the number of posεible ribozyme molecules is approximately 2(x-y)/4, where x is the length of one strand of DNA and y is the average length of the DNA fragments inserted into the vectors. Because the length of ribozyme binding arms iε small compared to the sizes of most genes, the number of ribozymes screened will be approximately one-half the number of nucleotides in the coding strand of the DNA. For the 2.7 kb piece of TK DNA, one expects approximately 1350 ribozymes (approximately 632 ribozymes nucleotides in the TK coding region) . Many of these ribozymes will target inaccesεible structureε within the target mRNA or possess incompatible secondary structures which will reduce their catalytic activity. To determine the limited sites within the RNA which are accessible to active ribozymes, it is preferable to asεay ribozyme activity uεing εubεtrateε with uniquely labeled siteε (5' or 3' end-labeled). Analyεiε of the cleaved substrate fragment size will give an approximate location within the εubεtrate where ribozyme cleavage occurε. By using different sizes of substrateε εpanning the target RNA, one may localize most of the preferred ribozyme target siteε to defined distances from the 3' or 5' ends of the target RNA. Sequencing of the DNA template in these regions will reveal the probable ribozyme recognition sequence, which can be verified by synthesizing 20-40 mers of the εubεtrate RNA and teεting

for cleavage in the preεence of the quasi-random ribozyme preparation. Syntheεiε of RNA moleculeε in thiε εize range may be acco pliεhed enzymatically or chemically by eεtablished protocols. After the target region has been identified, ribozymes to the target sequence may be syntheεized which contain sequence modifications or additions which increase catalytic rates or nuclease stability in cellular extracts and/or alter intracellular localization when synthesized in transient asεayε from eucaryotic expreεεion vectorε. A reproducible, simple technique for assaying ribozyme cleavage of target RNAs in cellular extracts iε deεcribed below.

DNA templateε for the synthesis of antisenεe RNA probeε spanning the TK gene were amplified from 10 ng of HSV-1 (KOS strain) DNA, using the following primers: PRIMER A 5' TAATACGACTCACTATAGGGCTTCTTGCTGCCGGCGA 3' PRIMER B 5' CATCCCCGTGGCCCGTT 3' Primer A contains a T7 RNA polymerase promoter. Probes with a specific activity of 1.8 x IO ⁶ dpm/ng were syntheεized directly from PCR-amplified template DNA uεing T7 RNA polymeraεe (USB) as described by the manufacturer, in the presence of ³² P-labeled CTP (Dupont/NEN) . Cytoplasmic extracts (40 μl) derived from HSV-infected Vero cells (MOI =1, HSV-1, strain KOS) were combined with 5 μl of lOx ribozyme cleavage buffer (750 mM Tris, pH 7.5, and 100 mM MgCl ₂ ) and ribozyme (diluted in 5 μl water, final ribozyme concentration varied from 0.1 to 10 μM) , and incubated at 37°C for varying amountε of time. Following ribozyme cleavage, extractε were diluted to 400 μl with 4 M guanidine thiocyanate lyεiε buffer and RNA waε extracted aε deεcribed (Chomεzynεki and Sacchi, 162 Anal. Biochem. 156, 1987) . RNA waε resuspended in 45 μl lysiε buffer, antiεenεe TK probe (1-5 x 10 ⁵ cpm in 5 μl lyεiε buffer) εpanning the cleavage εite waε added, then hybridization reactionε were overlaid with 30 μl of mineral oil and incubated at 55°C for 18-20 hours. After

hybridization, 500 μl of RNase solution (0.4 M NaCl, 20 U/ml RNaεeA, 2 U/ml RNaεeTl and lOmM Triε/HCl pH 7.5) waε added, and reactions were incubated at 37°C for 30 minutes. Following nucleaεe digeεtion, 10 μl of 20% SDS and 10 μl of 20 mg/ml proteinase K were added, the reactions were heated an additional 30 minuteε at 37°C, and hybridε were precipitated with 500 μl of isopropanol. Pellets were resuspended in 10 μl of solution containing 95% formamide, 10 mM EDTA, 0.1% bromophenol blue and 0.1% xylene cylanol (USB) , heated to 95°C for 4 minutes, then loaded onto 5% polyacrylamide/7 M urea gelε and electrophoreεed at 65 W conεtant power. Gelε were dried and the radioactivity correεponding to protected RNA fragmentε waε viεualized by autoradiography and quantified uεing an AMBIS detector.

Cloning of a Concatameric Ribozv e Construct After ribozymes have been selected for effective cleavage of target RNA in cells or extracts and optimized for catalytic rate, intracellular εtability and εubcellular localization, the final εtep in designing efficient ribozyme-expression vectors involves the cloning of multiple therapeutic ribozymes (TRs) within the transcription unit. The release of the TRs from the transcript necessitates the cloning of an additional releasing ribozyme (RR) which can cleave between TR units in the naεcent tranεcriptε. Thiε protocol haε been deεcribed in a previouε εection. Assay for Ribozyme Target Site

Applicant has devised a method by which the accesεibility of any particular target εite for a ribozyme can be determined in vitro or in vivo . Thiε aεεay is important to ensure that time and energy is not expended on developing highly active ribozymes which are unable to locate or bind to a target sequence in vivo because it is later found that the target sequence is inaccessible to the ribozyme. Such inacceεεibility may be caused by the target RNA secondary structure, or by protein factors

which bind to the target sequence. The asεay for εuch target site acceεεibility can be performed without the need to form many ribozymes against each potential target site. In general, the method involves uεe of an enzyme which iε able to recognize a DNA-RNA duplex, and cleave the RNA in that duplex. For example, RNaεeH can be uεed to cleave RNA bound to a DNA molecule. Applicant recognizeε that the target εite for a ribozyme iε an RNA εubεtrate molecule and that an RNA-DNA duplex can be formed with εuch a target εite by proviεion of εuitable complementary DNA. The method involveε providing one or more such DNA oligonucleotides or molecules (which may be εyntheεized quickly and cheaply on a DNA εyntheεizer) , allowing those DNA molecules to contact the target RNA, and determining whether the RNA target is rendered susceptible to cleavage by RNaseH. Such cleavage may be measured by gel electrophoreεis of labeled target RNA, or by northern blotting, as well as by other methodology well known in the art.

The method of this invention allows DNA probes of variouε sizes to be chosen such that they approximate the εize and εhape of the ribozyme binding εite εo that acceεsibility of the entire ribozyme structure can be evaluated. In addition, εince many eukaryotic cellε contain endogenouε RNaεeH, it iε possible to use the method to evaluate target site accesεibility in whole living cellε by delivering the DNA oligonucleotides to those cellε under conditionε in which endogenouε RNaεeH will cleave any RNA-DNA duplexes formed in the cell.

Thus, the invention featureε a method for determining RNA target site acceεεibility to a ribozyme by providing a DNA oligonucleotide complementary to the RNA target εite and contacting that DNA oligonucleotide with the RNA target in the presence of a DNA-RNA specific cleaving agent, such aε RNaεeH, εo that the RNaεeH will cleave the RNA target when the DNA and RNA form a duplex.

The occurrence of cleavage is detected by any standard methodology, e.g. , northern blot or gel electrophoresiε. Those DNA oligonucleotides which provide a detectable level of cleavage indicate that the complementary RNA target site is acceεsible to a ribozyme. Ribozymes to these target siteε can then be εyntheεized, uεing εtandard methodology, and checked to determined that the RNA target εite iε truly accessible to a ribozyme.

By "target εite" iε meant a εequence within a target RNA that iε "targeted" for cleavage: (i) by a ribozyme which containε εequenceε within its substrate- binding domain that are complementary to the target sequence, or (ii) by a DNA/RNA cleaving agent, such aε RNaεeH, following the binding of a complementary DNA oligonucleotide to the target εequence.

By "substrate-binding domain" is meant sequenceε within a ribozyme which are intended to bring ribozyme and target RNA together through complementary base-pairing interactions; e.g. , ribozyme sequenceε within εtemε I and III of a standard hammerhead ribozyme make up the subεtrate-binding domain (εee Fig. 35) .

By "DNA oligonucleotide" iε meant a short sequence of DNA; in this application, the sequence is in the range of 7-20 nucleotides, and preferably in the range of 9-13 nucleotides.

By "RNaseH" is meant an endoribonuclease that degradeε the RNA portion of RNA-DNA hybrid duplexeε. RNaεeH is found in many organismε; E . coli RNaεeH is available from many commercial sources. By "detectable level of cleavage" is meant cleavage of target RNA (and formation of cleaved product RNAs) to an extent sufficient to discern cleavage products above the background of RNAs produced by random degradation of the target RNA. Production of cleavage products from 1-5% of the target RNA is sufficient to detect above the background for most methods of detection.

Preferably, the DNA oligonucleotides are between 9-15 baseε in length, and completely complementary to the target εite.

More preferably, the DNA oligonucleotides are of the form NGAN, NUAN or NXYZN where each N is any nucleotide εequence from 0-12 nucleotideε and XYZ iε εelected from the group conεiεting of GAA, GAC, GAG, GAU, UAA, UAC, UAG and UAU; the DNA oligpnucleotideε are εingle-stranded fragments of a gene encoding the target RNA; and the DNA oligonucleotides are obtained by asymmetric amplification followed by fragmentation by use of a procedure selected from: (a) thiophosphate incorporation followed by peroxide degradation, (b) chemical cleavage, and (c) DNasel digestion. Method

The method of the invention is generally described above. Examples of methods of the invention are provided below, and in Figs. 35 - 40 discussed above, which are not limiting to this invention. Those of ordinary skill in the art will recognize that many equivalents to such methods can be used.

The asεay of thiε invention doeε not provide a final test of target site accesεibility, but rather provides a preliminary test. A number of factors may cause the test to distort the picture of which siteε are acceεεible and which are not. For example, the enzyme uεed, e.g. , RNaεeH, may cut the target RNA at a εite that iε only partially bound by a DNA oligonucleotide. Such a εite may thuε be only partially εuεceptible to RNaεeH cleavage, and not accessible at all to a ribozyme. The occurrence of such an erroneous false positive result can be reduced by testing each RNA target site with a set of overlapping DNA oligonucleotides. In addition, a DNA oligonucleotide does not have exactly the same binding affinity as an RNA oligonucleotide, so that a εite that iε inacceεεible to a DNA probe εtill may be acceεεible to an RNA probe, or vice versa . For the most part such

differences are small, and false negative reεultε are rare. Again, misguidance by the occurrence of such falεe results can be avoided by use of several different DNA oligonucleotides for each test site. The DNA oligonucleotide used may completely span the ribozyme binding εite in juεt one helix (i.e. , with no baεe mismatches, bulges or loops) . In contrast, an equivalent ribozyme binding at the same site may have two binding regionε εplit by a core sequence, e.g. , this is the case for a hammerhead ribozyme. The presence of such a central core may reduce the binding affinity of the ribozyme compared to a DNA oligonucleotide by aε much as 7.0 kcal/mole; thus, a DNA oligonucleotide may bind many times more strongly than the equivalent ribozyme. The possibility of this discrepancy can be reduced by use of short DNA oligonucleotides (9-13 baεeε) . Since εhort oligonucleotides will not necessarily bind to the same site as the ribozyme, a number of shorter overlapping DNA should be used in each set of experiments. In general, the strategy involves the formation of DNA oligonucleotides complementary to any number of selected RNA target siteε. The target RNA can be tranεcribed in vitro uεing labeled nucleotide triphoεphateε, and added directly to an RNaseH assay mixture. The target RNA is preferably not purified in order to keep its configuration aε cloεe as possible to that in a natural in vivo environment. The DNA oligonucleotides are mixed in variouε concentrationε with an excess of RNaseH and the labeled target RNA and incubated at 37°C. Samples are removed at various timeε and loaded onto εequencing gels. The most useful RNA target εiteε are thoεe that εhow the largeεt extent of cleavage in the shorteεt amount of time.

The RNaseH experiments can alεo be performed as described above, except that the target RNA need not be labeled, and can be provided within a cell extract.

Northern blotting (or other means, e.g. , RNaεe protection) of the RNA can then be used to detect cleavage of the RNA. In one type of experiment, in vitro transcribed, body-labeled target RNA iε digeεted by addition of high concentrationε of RNaεeH in the preεence of varying concentrationε of 9-15 mer DNA oligonucleotideε complementary to εpecific εiteε on the target. The high concentration of RNaεeH reduceε the probability that cleavage by RNaεeH will be the rate-limiting εtep in the reaction. Thuε, differenceε in RNA cleavage rates (and extents) are interpreted to represent differences in rates (and extentε) of association between the DNA oligonucleotides and their respective siteε on the target RNA. Siteε which are acceεεible to cleavage by RNaεeH are more likely to be targetable by ribozymeε.

In a further experiment, we have employed semi- random DNA oligonucleotideε aε RNaseH probes of target RNA accessibility. The semi-random probes are of the form, NXYZN, where N representε a randomized poεition (i.e. , A, C, G, or T; or each N iε individually between 0 and 12 nucleotideε, and the oligonucleotide containε at leaεt 8 -

12 nucleotideε) and XYZ repreεentε a εpecific sequence complementary to one of the preferred ribozyme cleavage sequences (e.g. , GUC, GUA, CUC, CUA) . Probing with fully randomized DNA provideε unεatisfactory results; cleavage of the target RNA occurs only at high probe concentrationε and without the formation of diεcernable bandε. The εemi- random probeε deεcribed above permit a large number of potential target εiteε to be evaluated from a εingle reaction. The number of different εequenceε represented in the semi-random probe, however, iε only a fraction (4 ^"3 or 1.6%) of εequenceε present in a completely random probe. This has the advantage of increasing the concentration of any given sequence while the number of bands to be interpreted is kept to a manageable number. Example 33: RNaseH in vitro

The following reagents were used:

5X buf fer :

100 mM Tris pH 7.9 500 mM KC1 50 mM MgCl ₂ 0.5 mM EDTA

0.5 mM DTT RNaseH:

Obtained from Bethesda Research Labε. (BRL) 2.5 U/μl Formamide dye εtop:

95% Formamide 20 mM EDTA

0.1% bromophenol blue 0.1% xylene cyanol DNA oligonucleotide (typically 13 nucleotide long) at a 10X concentration of 10, 1, and 0.1 μM. The method involved the following steps:

1. Dilute a standard transcription reaction (including the target RNA formed by transcription in a cell lysate or in a test tube) to the appropriate counts per minute (cpm) per μl to a 10X concentration;

2. Set up the reaction mix: (typical reaction)

5X Buffer 2.0 μl Body labeled Target

RNA 50,000 (cpm/λ) 1.0 μl

RNaseH (2.5 u/λ) 0.32 μl

H ₂ 0 5.68 μl

9.0 μl 3. Prewarm reaction mix to 37°C, then add 1 μl of DNA oligonucleotide. Allow to incubate at 37°C for 10 minuteε, then add 5 μl of formamide εtop εolution. Brief spin to remove condensation from the tubes wall, then heat at a minimum of 90°C for 3' before loading onto a denaturing acrylamide gel.

Example 34: RNaseH in Cell Extracts

The following reagents were uεed:

5X buffer:

100 M Tris pH 7.9 500 mM KC1 50 mM MgCl ₂ 0.5 mM EDTA

0.5 mM DTT RNaseH:

BRL 2.5 U/μl Cell Ivsiε buffer: 4 M guanidine thiocyanate

0.5% εarcoεyl

25 mM sodium citrate pH 7.4 Formamide dye stop: 95% Formamide 20 mM EDTA

0.1% bromophenol blue 0.1% xylene cyanol RNase Solution: 0.4 M NaCl 10 mM MgCl ₂

20 U/ml RNaseA (United Stateε Biochemical Co. Ohio; USB)

2 U/ml RNase Tl (USB) 4 U/ml DNase 1, RNaεe free (USB) 20% SDS

Proteinaεe K: 20 mg/ml

Phenol: chloroform: iεoamyl alcohol 25:25:1 Iεopropanol tRNA: 2 μg/μl DNA oligonucleotide (typically 13 nucleotide long) at a 10X concentration of 150 μM.

After preparing a cell lysate (from a cell including a target RNA) using standard procedures the method involved the following stepε: 1. In a 0.5 ml tube add:

5x buffer 3 μl cell lyεate 10 μl

RNaseH 0 . 3 μ l

H ₂ 0 1 . 68 μl

14 μl

2. Prewarm to 37°C and add 1 μl of 10X stock of DNA oligonucleotide, incubate at 37°C for 10 minutes and stop the reaction by adding 40 μl of lysis buffer.

3. Add 5 μl of labeled RNase protection probe RNA (200,000 cpm/λ) , overlay with 25 μl of mineral oil and hybridize overnight at 55°C. 4. Transfer to a 1.5 ml tube containing 500 μl of RNase solution and incubate at 37°C for 30 minutes.

5. Add 10 μl of 20% SDS and 10 μl of 20 mg/ml proteinase K, incubate at 37°C for 30 minuteε.

6. Extract once with 500 μl of 25:25:1 mixture of phenol: chloroform: isoamyl alcohol, brief vortex and transfer the aqueous phase to a freεh 1.5 ml tube containing 500 μl of iεopropyl alcohol with tRNA carrier.

7. Pellet at room temperature (about 20°C) for 8 minuteε. 8. Reεuεpend in 10 μl of formamide dye εtop.

9. Heat above 90°C for over 3 minutes and load onto a denaturing acrylamide gel.

Results of the above two examples are shown in the figures. These resultε were obtained uεing the εtandard procedureε deεcribed above and other εtandard protocolε well known in the art.

Aε noted above, it iε important in thiε invention to provide a plurality of DNA oligonucleotideε.

Selection of appropriate oligonucleotideε iε important to the outcome of the experiment. Several exampleε of such selections are provided below. Each selection is useful dependent upon the data required, e.g. , whether a target site in a genome is required, or a target site in a gene.

Example 35: Thus, in one example completely randomized oligonucleotideε can be made on a DNA syntheεizer. Thiε iε the eaεieεt approach to try εince it takeε only one

synthesis and one test to select a target site. Random 6- and 11-mers have been tried without succeεs. The lack of succeεs, however, may be due to too much overlap between the various oligonucleotides, or becauεe the oligonucleotideε lightly coat the RNA and make all εiteε εlightly εuεceptible to cleavage. Example 36:

An alternative approach involveε making many different DNA oligonucleotideε, and mixing them together prior to teεting in the aεεay. The approach εaveε time over teεting each oligonucleotide εeparately. Oligonucleotideε are beεt εelected εo that they are not εo close together that it is difficult to determine which εite iε accessible. Example 37:

Another approach is to cut up a portion of a chosen target gene and use that mixture as DNA oligonucleotide material in the asεay. This approach offers a complete set of oligonucleotides that match the target RNA with a minimum of unwanted sequence. The DNA εtrandε firεt need to be separated so that only the complementary strand is used to make DNA oligonucleotide (since that other strand will act as a competitor for that DNA oligonucleotide) . It is also best to cut the DNA relatively uniformly to produce a distribution of fragments in a narrow size range. Aεymmetric polymerase chain reaction (PCR) is the beεt way of making εingle- εtranded DNA. Cleavage of the DNA can be accompliεhed by: (i) incorporation of thiophoεphate nucleotideε into the DNA followed by H ₂ 0 ₂ degradation, (ii) chemical cleavage of the DNA uεing DMS (at G) or DEPC (at A) or Hydrazine (at U or C) , or (iii) digeεtion by DNaεel.

Asymmetric PCR may not yield sufficient quantities of single-εtranded DNA complementary to the target εequence. An alternative iε to clone the target gene or sequence into an M13 vector and then isolate large quantities of single-εtranded DNA vector containing the

target gene (M13 iε a εingle-εtranded DNA phage which haε both a single-stranded and a double-stranded component to itε life-cycle) . Removal of the extra vector εequence could be achieved (if neceεεary) by annealing two complementary DNA oligonucleotideε to regionε flanking the target gene or εequence and then cutting the DNA within theεe duplex regionε uεing reεtriction enzymeε with sites located within those duplex regions (see Fig. 41). Example 38: Another method for producing suitable DNA oligonucleotides iε to make a partially-random oligomer. The εtrategy iε baεed on the fact that it is preferred to study accesεibility of ribozymeε only at thoεe εites that may be cleavable by such ribozymes. That is, siteε that contain a UC or a UA in the sequence. Thus a series of oligonucleotides can be made of the form 5'-NGAN or 5'- NUAN (these oligonucleotides are complementary to the UC or UA εiteε and where each N iε any nucleotide εequence from 0-12 nucleotideε) to be used in the assay. Even greater specificity can be achieved if many different syntheses are performed with the central three nucleotides being exactly specified (e.g.. GAA, GAC, GAG, GAU, UAA, UAC, UAG, UAU, see below) . Each of the oligonucleotide probes can be tested in separate asεays. The advantages of this method are: (i) the oligonucleotides can be made on the DNA εyntheεizer, (ii) three fewer sites are randomized so that each oligonucleotide sequence is preεent at a 64 fold (4 ³ ) higher concentration, (iii) poεsible problems with overlapping oligonucleotides are minimized εince each central εequence appearε at random only about every 64 (4 ³ ) nucleotideε, and (iv) identification of the accessible sites is easier since part of the DNA oligonucleotide sequence will be known and the position of cleavage thus identifiable within about 64 nucleotides.

The assay of thiε invention does not provide a final test of target site accesεibility, but rather

provides a preliminary test. A number of factors may cause the test to diεtort the picture of which εites are accessible and which are not. For example, the enzyme used, e.g. , RNaseH, may cut the target RNA at a site that iε only partially bound by a DNA oligonucleotide. Such a εite may thuε be only partially εuεceptible to RNaεeH cleavage, and not accessible at all to a ribozyme. The occurrence of such an erroneous false positive result can be reduced by testing each RNA target site with a set of overlapping DNA oligonucleotideε. In addition, a DNA oligonucleotide doeε not have exactly the same binding affinity as an RNA oligonucleotide, so that a site that is inaccesεible to a DNA probe εtill may be acceεεible to an RNA probe, or vice versa . For the oεt part εuch differenceε are small, and false negative resultε are rare. Again, miεguidance by the occurrence of such false results can be avoided by use of several different DNA oligonucleotides for each test site.

The DNA oligonucleotideε uεed may completely span the ribozyme binding siteε in juεt one helix (i.e. , with no baεe miεmatcheε, bulgeε or loops) . In contrast, an equivalent ribozyme binding at the same site may have two binding regionε εplit by a core εequence, e.g. , thiε is the case for a hammerhead ribozyme. The presence of such a central core may reduce the binding affinity of the ribozyme compared to a DNA oligonucleotide by as much aε 7.0 kcal/mole; thuε, a DNA oligonucleotide may bind many timeε more εtrongly than the equivalent ribozyme. The poεεibility of thiε diεcrepancy can be reduced by uεe of short DNA oligonucleotides (9-13 bases) , or by reducing the concentration of DNA probes in the acceεεibility assay.

In general, the strategy involves the formation of DNA oligonucleotideε complementary to any number of εelected RNA target εiteε. The target RNA can be transcribed in vitro Using labeled nucleotide triphosphates, and added directly to an RNaseH assay

mixture. The target RNA is preferably not purified in order to keep its configuration as close aε posεible to that in a natural in vivo environment. The DNA oligonucleotides are mixed in various concentrations with an excesε of RNaseH and the labeled -target RNA and incubated at 37°C. Samples are removed at various times and loaded onto sequencing gels. The most useful RNA target sites are those that show the largest extent of cleavage in the shortest amount of time. . The RNaseH experiments can alεo.be performed aε deεcribed above, except that the target RNA need not be labeled, and can be provided within a cell extract. Northern blotting or RNaεe protection of the RNA can then be uεed to detect cleavage of the RNA. In one type of experiment, in vitro tranεcribed, body-labeled target RNA iε digeεted by addition of high concentrationε of RNaseH in the presence of varying concentrations of 9-13 mer DNA oligonucleotides complementary to specific sites on the target. The high concentration of RNaseH reduces the probability that cleavage by RNaseH will be the rate-limiting step in the reaction. Thus, differences in RNA cleavage rateε (and extentε) are interpreted to represent differences in rates (and extents) of association between the DNA oligonucleotides and their respective siteε on the target RNA. Siteε which are acceεεible to cleavage by RNaεeH are more likely to be targetable by ribozymeε. Example 39:

In a further experiment, we have employed semi- random DNA oligonucleotides as RNaεeH probeε of target RNA accessibility. The semi-random probes are of the form, NXYZN, where N represents a randomized sequence position as above and XYZ represents a specific sequence complementary to one of the preferred ribozyme cleavage sequences (e.g.. GUC, GUA, CUC, CUA) . Probing with fully randomized DNA provides unsatiεfactory results; cleavage of the target RNA occurs only at high probe concentrations

and without the formation of discemable bands. The semi- random probes described above permit a large number of potential target sites to be evaluated from a single reaction. The number of different sequences represented in the εemi-random probe, however, iε only a fraction (4" ³ or 1.6%) of εequenceε preεent in a completely random probe. Thiε haε the advantage of increaεing the concentration of any given εequence while the number of bands to be interpreted is kept to a manageable number. Ribozyme Synthesiε

A method is provided for automated chemical syntheεiε of a ribozyme by εeparately εyntheεizing two or more portionε of the ribozyme and then ligating them to form an enzymatically active ribozyme. Thiε method increases the yield of the final ribozyme over other methods, and makes manufacture of such ribozymes cheaper. The prior inefficient synthesis of long RNA molecules is reduced or obviated in this method. One or more ribozyme portions may be used as one portion of several different ribozymes, and combined with other εynthetic portionε. In addition, each ribozyme fragment is more likely to be free of secondary structure and therefore can be purified by HPLC with greater resolution. The full length ribozyme, in contrast, may elute as multiple peaks. Thus, purity of the ribozyme resulting from two fragments can be greater than that when the ribozyme is formed as a single oligonucleotide.

Specifically, the invention features automated chemical synthesis to make ribozymes as two "half ribozymes" that can be joined by RNA ligase (e.g.. one derived from the bacteriophage T4) to produce a "full length" enzymatically active ribozyme molecule. Thiε method will minimize any impact of inefficient εyntheεiε of long RNA on final yield. Thiε εcheme can alεo be used to mix and match different half ribozyme units allowing the characterization of many half ribozyme combinations without neceεεitating additional costly syntheses.

Thuε, a method is provided for syntheεiε of an enzymatically active ribozyme by providing two or more portionε of a ribozyme and ligating thoεe portions together to form the active ribozyme. The ribozyme may be formed from RNA only, RNA and DNA, or may include modified nucleotide baεeε.

By the phraεe "portionε of a ribozyme" is meant an RNA (or modified RNA) molecule which can be used to form an enzymatically active RNA molecule by providing at least one other RNA molecule to which it is ligated or otherwise juxtaposed. For example, a portion is not enzymatically active until joined with such another RNA molecule. Preferably, only two or three such portions are required to form a fully enzymatically active ribozyme. Such enzymatically active ribozymes are active to cleave specific target RNAs or to cleave intramolecularly.

Preferably, the ribozyme is a hammerhead ribozyme having between 34 and 40 nucleotides and each portion includes 15 or more nucleotides; the portions are hybridized to a DNA strand complementary to each portion and are juxtaposed by that hybridization; the juxtaposed portions are ligated together by an RNA ligase, e.g. , T4 RNA ligase; one portion is phosphorylated; the ligated portions are purified after ligation; and more than two portions are provided and a plurality of different ribozymes simultaneously formed in the method.

Such juxtapositioning may be achieved by allowing each half-ribozyme to pair at stem II of the ribozyme (see, Fig. 20, which may include between 4 and 7 base pairε, preferably 6 baεe pairε) . Such pairing can be enhanced by increaεing the length of εtem II. The term "juxtapoεe" iε used merely to indicate that the ribozyme portions are sufficiently close to either be ligated or to have enzymatic activity. There follows examples of the invention. These examples are not limiting in this invention, and those in the art will recognize that other equivalent procedures

can be used. Generally, any ribozyme can be synthesized by the method of this invention, and may have either intra- or inter-molecular cleaving enzymatic activity. Each portion of a ribozyme will generally have only limited or no such activity, and the activity will increase substantially (by at least 5-10 fold) when all portions are ligated (or otherwise juxtaposed) together. The specific examples that follow feature hammerhead ribozyme εyntheεis, but the invention is equally applicable to hairpin ribozymeε, group I intron type ribozymeε and hepatitiε delta type ribozymeε among otherε. Example 40: Syntheεiε of Hammerhead Ribozyme

Prior to thiε invention, ribozyme εyntheεis involved εyntheεizing a full length hammerhead ribozyme chemically, e.g. , one having the structure εhown in Fig. 20, where stems I and III are about 5-8 nucleotides in length, and the full length ribozyme includes between 32 and 38 nucleotides.

RNA synthesis is less efficient than DNA synthesis. Stepwiεe coupling constantε for RNA synthesis are only about 96 to 98% (dependent upon the type of phosphoramidite used) . Thus, the maximum yield of full length ribozyme for a given synthesis (assuming 97% coupling efficiency) is (0.97) ³² or about 38% for a 32 mer, and (0.97) ³⁸ or about 31% for a 38 mer.

A method of this invention is εhown generally in Fig. 42, and featureε εyntheεis of two (or more) shorter "half" ribozymes and ligation of them together using T4 RNA ligaεe. Referring to Fig. 42, two 17 merε are εyntheεized, one iε phosphorylated, and both are gel purified. These purified 17 merε are annealed to a DNA or RNA or other nucleic acid εplint strand complementary to the two 17 mers. Thiε εplint haε a sequence designed to locate the two portions with one end of each adjacent each other. The juxtaposed RNA molecules are then treated with T4 RNA ligase in the presence of ATP. The 34 mer RNA so formed is then HPLC purified.

The specific stepε in thiε method will now be discussed: Synthesiε of RNA 17 mers iε practically trivial. Maximum yield for a 17 mer iε about (0.97) ¹⁷ or 60% (of. about 30% for a 34 mer) . Gel purification of a 17 mer can eaεily be accompliεhed on a 15% denaturing polyacrylamide gel with εingle nucleotide resolution, and the separated 17 mer easily removed from a gel slice. HPLC purification of a 17 mer from its failure sequences can be accomplished with much higher resolution than purification of a 34 mer. Reεolution of the 34 mer iε further reduced by secondary structure inherent to the ribozyme molecule.

After purification of the 3' segment or portion of the ribozyme it is phosphorylated in order to provide the 5' phosphate necessary for ligation. T4 RNA ligase can join the purified RNA molecules after alignment by a third "splint" strand as described by Linne et al., 42 Eur. J. Biochem. 157, 1974, in a manner analogous to T4 DNA ligase (Fareed et al., 246 J. Biol. Chem. 925, 1978) . The arrangement shown in Fig. 42 juxtaposes the only 5' phosphate (on the 3' RNA fragment) with the 3' OH of the 5' fragment. The splint strand can be chosen to be generally useful, and can be used for any ribozyme with the desired core sequence, e.g.. that shown in Fig. 42 (see, Forster and Symonε, 50 Cell 9, 1987) . The length of the DNA can be adjusted if necessary. Ligation with T4 RNA ligase can be accomplished in the same buffer as that of T4 polynucleotide kinase (i.e.. 50 mM HEPES (pH 7.5) , 20 mM MgCl ₂ , 3 mM DTT, 10 μg/ml nuclease free BSA) . This facilitates synthesiε since the 3' fragment can be used directly from the phosphorylation reaction (Beckett and Uhlenbeck, Ijn Oligonucleotide Synthesis: A Practical Approach, Chapter 8, pp. 185-197).

After ligation there will be RNA product, DNA splint, RNA starting materials and protein and salt compoundε. Theεe can be εeparated by reverεed phaεe HPLC,

or by gel purification. Denaturants may be used to help remove the splint strand.

Example 41: Synthesis of Several Ribozymes Simultaneouεly It is sometimes useful to syntheεize ribozymes having varying arm length to one target site. Such ribozymes can be studied to identify the "optimum" arm length at this target site. To vary each arm to include lengths of 5, 6, 7, or 8 nucleotides on each arm (i.e.. 16 combinations of arm lengths) would require syntheεizing 16 full length ribozymeε.

Uεing the method of thiε invention, however, one need only εyntheεize four 5' fragmentε and four 3' fragmentε, which produceε conεiderable εavingε in coεt and time over εynthesis of 16 different ribozymeε. With theεe two εetε of four half ribozymeε, one can mix and match to get the deεired combinationε of arm lengthε. One can also combine arms from various ribozymes to teεt various sequenceε of the armε to determine which form highly active ribozymeε. One εcheme for rapid synthesiε of four such ribozymes involves ligation of each of four phosphorylated 3' fragments with a choεen 5' fragment uεing an appropriate DNA splint molecule. The resulting products can be separated by polyacrylamide gel electrophoresiε (εince each product differε by one nucleotide baεe in length) . This procedure can be repeated with the other three 5' fragments. Thus, in four reactions and one gel, it iε poεεible to purify enough material to provide the deεired 16 different ribozymeε. Example 42: Juxtapoεed RNA Portionε Two-part or three-part ribozymeε can be formed without a ligation εtep. For example, such ribozymes can be uεed to εtudy the activity of a ribozyme aε a function of arm length. Referring to Fig. 43, there iε shown a ribozyme formed without the stem II loop closed. In this example, it iε necessary only to mix and match arm lengths in equimolar amountε. Stem II can be made longer to allow

the formation of a εignificant amount of 5' fragment: 3' fragment complex at 37°C. Example 43: No Splint Procedure

Referring to Fig. 45, there iε shown an alternative scheme by which a ribozyme can be ligated without use of a splint strand. Specifically, the two half ribozymes are constructed so that they hybridize together to enhance ligation. This procedure is advantageous over that described above since the splint strand need not be later removed. Ligation with such a procedure is almost quantitative and complete within an hour.

Example 44: Catalytic Activity of Ligated HIV Half- Ribozvmeε RNA molecules were chemically synthesized having the nucleotide base sequence shown in Fig. 44 for both the 5' and 3' fragments. Three identical 5' fragments were synthesized except that rather than having all 3 Us at their 5' end they had 0, 1 or 2 such Us. Three identical 3' fragments were also synthesized except that rather than having CUG at their 3' ends, they had only CU, C or nothing. As is evident from the figure the 3' and 5' fragments can hybridize together at stem II, and cannot be ligated because the hybridized complex is blunt-ended. When this structure hybridizeε to a εubεtrate, a ribozyme εtructure is formed.

Forty nanomolar mixtures of the appropriate 3' and 5' fragmentε were incubated with 1-5 nM εubεtrate for different ti eε in 75 mM Triε/HCl pH 8.0, 10 mM MgCl ₂ and shown to cleave the subεtrate appropriately with an activity about half that of full length εyntheεized ribozyme. Example 45: Preferred Half Ribozv eε

Referring to Fig. 46, there iε εhown the general εtructure of two half ribozymeε where N iε any desired nucleotide base. Thus, both the 3' and 5' fragmentε have about 21 baεes each with stem II having a base paired

region of at least 2 bases and preferably between 3 and 5 bases with a loop region containing at least 4 baseε (e.g. , on the 3' fragment). Stemε I and III are designed to have at least 3 base pairs, and preferably between 4 and 8 baεe pairε. Ribozyme Teεting

Once the deεired ribozymeε are selected, synthesized and purified, they are tested in kinetic and other experiments to determine their utility. An example of such a procedure is provided below. Preparation of Ribozyme

Crude εynthetic ribozyme (typically 350 μg at a time) iε purified by εeparation on a 15% denaturing polyacrylamide gel (0.75 mm thick, 40 cm long) and visualized by UV shadowing. Once excised, gel slices containing full length ribozyme are soaked in 5 ml gel elution buffer (0.5 M NH ₄ OAc, 1 mM EDTA) overnight with shaking at 4°C. The eluent is desalted over a C-18 matrix (Sep-Pak cartridges, Millipore, Milford, MA) and vacuum dried. The dried RNA iε reεuεpended in 50-100 μl TE (TRIS 10 mM, EDTA 1 mM, pH 7.2). An aliquot of thiε solution is diluted 100-fold into 1 ml TE, half of which iε used to spectrophotometrically quantitate the ribozyme solution. The concentration of this dilute stock is typically 150- 800 nM. Purity of the ribozyme iε confirmed by the preεence of a εingle band on a denaturing polyacrylamide gel.

A ribozyme may advantageouεly be εyntheεized in two or more portions. Each portion of a ribozyme will generally have only limited or no enzymatic activity, and the activity will increase subεtantially (by at leaεt 5-10 fold) when all portionε are ligated (or otherwiεe juxtaposed) together. A specific example of hammerhead ribozyme εyntheεiε iε provided below. The method involves synthesiε of two (or more) shorter "half" ribozymes and ligation of them together using T4 RNA ligaεe. For example, to make a 34 mer

ribozyme, two 17 mers are syntheεized, one iε phoεphorylated, and both are gel purified. These purified 17 mers are then annealed to a DNA splint strand complementary to the two 17 merε. This DNA splint has a sequence designed to locate the two 17 mer portions with one end of each adjacent each other. The juxtaposed RNA molecules are then treated with T4 RNA ligase in the presence of ATP. Alternatively, the DNA splint strand may be omitted from the ligation reaction if the complementary binding affects favorable ligation of the two RNA molecules. The 34 mer RNA so formed is then HPLC purified.

Preparation of Substrates

Approximately 10-30 pmoles of unpurified substrate is radioactively 5' end-labeled with T4 polynucleotide kinase using 25 pmoles of [7~ ³² P] ATP. The entire labeling mix is separated on a 20% denaturing polyacrylamide gel and visualized by autoradiography. The full length band is excised and soaked overnight at 4°C in 100 μl of TE (10 mM Tris-HCl pH 7.6, 0.1 mM EDTA) . Kinetic Reactions

For reactionε using short substrateε (between 8 and 16 bases) a substrate solution is made IX in assay buffer (75 mM Tris-HCI, pH 7.6; 0.1 mM EDTA, 10 mM MgCl ₂ ) such that the concentration of subεtrate iε leεs than 1 nM. A ribozyme solution (typically 20 nM) is made IX in assay buffer and four dilutions are made using IX aεεay buffer. Fifteen μl of each ribozyme dilution (i.e. , 20, 16, 12, 8 and 4 nM) iε placed in a separate tube. These tubes and the substrate tube are pre-incubated at 37°C for at least five minutes.

The reaction is started by mixing 15 μl of εubεtrate into each ribozyme tube by rapid pipetting (note that final ribozyme concentrationε are 10, 8, 6, 4, 2 nM) . Five μl aliquotε are removed at 15 or 30 εecond intervalε and quenched with 5 μl stop solution (95% formamide, 20 mM EDTA xylene cyanol, and bromphenol blue dyes) . Following

the final ribozyme time point, an aliquot of the remaining substrate is removed as a zero ribozyme control.

The samples are separated on either 15% or 20% polyacrylamide gels. Each gel is visualized and quantitated with an Ambiε beta εcanner (Ambis Systemε, San Diego, CA) .

For the moεt active ribozymeε, kinetic analyses are performed in substrate excess to determine K,„ and K^ values. For kinetic reactions with long RNA substrates

(greater than 15 baseε in length) the substrates are prepared by transcription using T7 RNA polymerase and defined templates containing a T7 promoter, and DNA encoding appropriate nucleotides of the viral RNA. The substrate solution is made IX in assay buffer (75 mM Tris- HCI, pH 7.6; 0.1 mM EDTA; 10 mM MgCl ₂ ) and contains 58 nanomolar concentration of the long RNA molecules. The reaction is started by addition of gel purified ribozymes to 1 μM concentration. Aliquotε are removed at 20, 40, 60, 80 and 100 minuteε, then quenched by the addition of 5 μl εtop εolution. Cleavage products are separated using denaturing PAGE. The bands are visualized and quantitated with an Ambis beta scanner. In one example, a long substrate RNA transcript corresponding to nucleotides 1- 493 of the HSV-l ICP4 mRNA is synthesized in vitro with T7 RNA polymerase from a defined template containing T7 promoter and DNA encoding nucleotideε 1-493 of ICP4. Kinetic Analysiε A εi ple reaction mechanism for ribozyme- mediated cleavage is: k, k ₂

R + S ⁱ → [R:S] [R:P] R + P k, where R = ribozyme, S = subεtrate, and P = productε. The boxed εtep iε important only in εubstrate excesε. Because ribozyme concentration iε in excess over substrate

concentration, the concentration of the ribozyme-substrate complex ([R:S]) is constant over time except during the very brief time when the complex is being initially formed, i.e., :

where t = time, and thus:

(R) (S)k, = (RS) (k ₂ + k,). -

The rate of the reaction is the rate of diεappearance of substrate with time:

Substituting these expresεionε:

(R) (S)k, = l/k ₂ -d(S) (k ₂ + k,) dt or:

-d(S) = k,k ₂ (R) dt

S (k ₂ + k,)

Integrating thiε expression with respect to time yields: -In S_ = k,k ₂ (R) t

S ₀ (k ₂ + k,) where S ₀ = initial substrate. Therefore, a plot of the negative log of fraction substrate uncut versus time (in minutes) yields a straight line with εlope: εlope = k,k ₂ (R) = k^

(k ₂ + k,) where k _obs = obεerved rate conεtant. A plot of εlope (k^ versus ribozyme concentration yields a εtraight line with a slope which is: slope = k,k ₂ which is k^,

(k ₂ + k,) K _m

Using these equationε the data obtained from the kinetic experimentε provides the necesεary information to determine which ribozyme teεted iε moεt uεeful, or active.

Such ribozymeε can be selected and tested in in vivo or ex vivo syεtemε. An example of preparing ribozymeε in a lipoεo e delivery vehicle is given below. Liposome Preparation There followε an example of the entrapment of ribozyme molecules within a liposome drug delivery vehicle. Lipid molecules were dissolved in a volatile organic solvent (CHC1 ₃ , methanol, diethylether, ethanol, etc.). The organic solvent was removed by evaporation. The lipid was hydrated into suεpension with O.lx phosphate buffered saline (PBS) , then freeze-thawed 3x uεing liquid nitrogen and incubation at room temperature. The suspenεion waε extruded sequentially through 0.4 μm, 0.2 μm and 0.1 μm polycarbonate filters at maximum pressure of 800 psi. The ribozyme was mixed with the extruded liposome suεpenεion and lyophilized to dryness. The lipid/ribozyme powder was rehydrated with water to one-tenth the original volume. The suspension was diluted to the minimum volume required for extrusion (0.4 ml for 1.5 ml barrel and 1.5 ml for 10 ml barrel) with IxPBS and re-extruded through 0.4 μm, 0.2 μm, 0.1 μm polycarbonate filters. The liposome entrapped ribozyme was separated from untrapped ribozyme by gel filtration chromatography (SEPHAROSE CL-4B, BIOGEL A5M) . The liposome extractions were pooled and sterilized by filtration through a 0.2 μm filter. The free ribozyme waε pooled and recovered by ethanol precipitation. The lipoεome concentration waε determined by incorporation of a radioactive lipid. The ribozyme concentration waε determined by labeling with ³² P. Roεεi et al., 1992, supra (and referenceε cited therein), deεcribe other methodε εuitable for preparation of lipoεomeε.

In experimentε with a lipoεome formulation compoεed of a synthetic lipid derivative disteraoyl- phoεphatidylethylamidothioacetyl εuccinimide (DSPE-ATS) co-formulated with dipalmitoylphoεphatidyl choline and choleεterol we observed uptake of 100 and 200 nm diameter

liposomes with similar kinetics. The larger particles accommodated a larger number of entrapped molecules, or larger molecular weight moleculeε, εuch aε an expression plasmid. These particles showed a linear relationship between the lipid dose offered and the mean log fluorescence (calcine was used to follow liposome uptake) . No cytotoxicity was observed even with a 200 μM dose. These liposomes are particularly useful for delivery to CD4 cell populations. In Vivo Aεεay

The efficacy of action of a choεen ribozyme may be teεted in vivo by uεe of cell cultureε εensitive to a selected viruε, uεing εtandard procedureε. For example, monolayer cultureε of virus-senεitive or virus-expressing cells are grown by established procedures. Cells are grown in 6 or 96 well tissue culture plates or in raft cultures of virus-permissive cells as described previously. Kopan et al., 105 J. Cell Biol. 427, 1987. Prior to infection with virus cultures are treated for 3 to 24 hours with ribozyme-containing liposomes. Cells are then rinsed with phosphate buffered saline (PBS) and viruε added at a multiplicity of 1-100 pfu/cell. After a one- hour adεorption, free viruε is rinsed away using PBS, and viral RNA is harvested for analysis. Alternatively, the cells are treated for three to five days with appropriate liposome preparations and fresh medium changes. Cells may then be re-fed with fresh medium and re-incubated. Virus is harvested from cells into the overlying medium. Cells are broken by three cycles of incubation at -70°C and 37°C for 30 minutes at each temperature, and viral titers determined by plaque asεay uεing eεtabliεhed procedureε. Theεe procedureε can be modified for each εpecific viruε to be teεted. Viral RNA can be harvested using the guanidinium isothiocyanate procedure and the resultant RNA preparation is subjected to analysis using the RNase protection assay.

Alternatively, transformed cell types (e.g. , HeLa cells for expression of HPV-18 E6 and E7 mRNA or Caski cellε for expreεεion HPV-18 E6 and E7 mRNA / Vero cellε) are uεed to aεεeεε the ability of ribozyme preparationε to cleave HPV/HSV mRNA and reduce the expreεεion of target proteinε. Cell monolayerε are treated with ribozyme containing lipoεome preparationε for between 3 and 96 hours. For RNA analyses, the cells are lysed in guanidine isothiocyanate solutions and the RNA are analyzed using the RNase protection asεay. For protein quantitation, cell lyεateε are prepared according to methodε known to thoεe in the art and proteinε are quantified by immunoanalyεiε uεing anti-HPV E6 and E7 / anti-HSV protein specific antibodies. For HBV, cells are rinsed with phosphate buffered saline (PBS) and HBV DNA is transfected into the cells by the calcium phosphate technique. The cells are treated for three to five days with appropriate lipoεome preparationε in fresh changes of medium. For HCV total cellular RNA iε harveεted by the guanidine isothiocyanate technique and the amount of virus mRNA iε quantified and aεsessed for ribozyme-mediated cleavage using the ribonuclease protection aεεay. Alternatively, cellular lysates can be prepared and the core particles of virus could be immunoprecipitated uεing polyclonal anti-core antiεerum adεorbed to protein A-Sepharoεe according to manufacturerε directionε. The precipitated coreε are treated with ribonucleaεe to digeεt any non-encapsidated RNAε and the core protein iε digeεted with proteinaεe K/phenol extraction. Uεually one-half of the total RNA and one-half of the extracted RNA iε analyzed uεing the ribonucleaεe protection aεεay.

Ribozyme Stability

The choεen ribozyme can be teεted to determine itε εtability, and thuε itε potential utility. Such a teεt can alεo be uεed to determine the effect of variouε chemical modificationε (e.g. , addition of a poly(A) tail)

on the ribozyme stability and thus aid selection of a more stable ribozyme. For example, a reaction mixture contains 1 to 5 pmoles of 5' (kinased) and/or 3' labeled ribozyme, 15 μg of cytoεolic extract and 2.5 mM MgCl ₂ in a total volume of 100 μl. The reaction iε incubated at 37°C. Eight μl aliquotε are taken at timed intervalε and mixed with 8 μl of a stop mix (20 mM EDTA, 95% formamide) . Samples are separated on a 15% acrylamide sequencing gel, exposed to film, and scanned with an Ambis. A 3'-labeled ribozyme can be formed by incorporation of the ³² P-labeled cordycepin at the 3' OH using poly(A) polymerase. For example, the poly(A) polymerase reaction contains 40 mM Tris, pH 8, 10 mM MgCl ₂ , 250 mM NaCl, 2.5 mM MnCl ₂ , ; 3 μl P ³² cordycepin, 500 Ci/mM; and 6 units poly(A) polymerase in a total volume of 50 μl. The reaction mixture is incubated for 30 minutes at 37°C. Effect of Base Subεtitution upon Ribozyme Activity To determine which primary structural characteristicε could change ribozyme cleavage of εubεtrate, minor baεe changes can be made in the subεtrate cleavage region recognized by a εpecific ribozyme. For example, the substrate sequences can be changed at the central "C" nucleotide, changing the cleavage site from a GUC to a GUA motif. The K _cat /K _1H values for cleavage using each substrate are then analyzed to determine if such a change increases ribozyme cleavage rates. Similar experiments can be performed to addresε the effectε of changing baεeε complementary to the ribozyme binding armε. Changeε predicted to maintain strong binding to the complementary subεtrate are preferred. Minor changeε in nucleotide content can alter ribozyme/εubεtrate interactions in ways which are unpredictable based upon binding strength alone. Structures in the catalytic core region of the ribozyme recognize trivial changeε in either εubεtrate structure or the three dimensional structure of the ribozy e/substrate complex.

To begin optimizing ribozyme design, the cleavage rateε of ribozymes containing varied arm lengths, but targeted to the same length of short RNA εubεtrate can be teεted. Minimal arm lengthε are required and effective cleavage varieε with ribozyme/εubεtrate combinationε.

The cleavage activity of εelected ribozymeε can be asseεεed using virus or virus-homologouε RNA εubεtrates or virus genomic RNA. The asεays are performed in ribozyme excess and approximate K _C!lt /K ₀ _ _m values obtained. Comparison of values obtained with short and long subεtrateε indicateε utility in vivo of a ribozyme.

Intracellular Stability of Liposome-Delivered

Ribozymes

To test the stability of a chosen ribozyme in vivo the following test is uεeful. Ribozymeε are ^P- end labeled, entrapped in liposomes and delivered to virus senεitive cellε for three hours. The cells are fractionated and ribozyme iε purified by phenol/chloroform extraction. Cellε (lxlO ⁷ , T-175 flask) are scraped from the surface of the flask and washed twice with cold PBS. The cells are homogenized by douncing 35 times in 4 ml of TSE (10 mM Tris, pH 7.4, 0.25 M Sucrose, mM EDTA). Nuclei are pelleted at lOOxg for 10 minuteε. Subcellular organelleε (the membrane fraction) are pelleted at 200,000xg for two hourε using an SW60 rotor. The pellet is resuspended in 1 ml of H buffer (0.25 M Sucrose, 50 mM HEPES, pH 7.4). The supernatant contains the cytoplasmic fraction (in approximately 3.7 ml). The nuclear pellet is resuεpended in 1 ml of 65% εucroεe in TM (50 mM Triε, pH 7.4, 2.5 mM MgCl ₂ ) and banded on a sucrose step gradient (1 ml nuclei in 65% sucrose TM, 1 ml 60% sucrose TM, 1 ml 55% sucroεe TM, 50% εucroεe TM, 300 μl 25% εucroεe TM) for one hour at 37,000xg with an SW60 rotor. The nuclear band iε harvested and diluted to 10% sucroεe with TM buffer. Nuclei are pelleted at 37,000xg uεing an SW60 rotor for 15 minutes and the pellet resuspended in 1 ml of TM buffer. Aliquots are size fractionated on denaturing

polyacrylamide gels and the intracellular localization determined. By compariεon to the migration rate of newly εyntheεized ribozyme, the variouε fraction containing intact ribozyme can be determined. To inveεtigate modificationε which would lengthen the half-life of ribozyme molecules intracellularly, the cells may be fractioned as above and the purity of each fraction assessed by asεaying enzyme activity known to exiεt in that fraction. The variouε cell fractionε are frozen at -70°C and uεed to determine relative nucleaεe reεiεtanceε of modified ribozyme moleculeε. Ribozyme molecules may be syntheεized with 5 phoεphorothioate (ps) , or 2'-0-methyl (2'-OMe) modificationε at each end of the molecule. Theεe moleculeε and a phoεphodieεter verεion of the ribozyme are end-labeled with ³² P and ATP uεing T4 polynucleotide kinaεe. Equal concentrationε are added to the cell cytoplaεmic extractε and aliquotε of each taken at 10 minute intervals. The samples are size fractionated by denaturing PAGE and relative rates of nuclease resiεtance analyzed by εcanning the gel with an Ambis 0-scanner. The resultε show whether the ribozymes are digested by the cytoplasmic extract, and which verεionε are relatively more nuclease resiεtant. Modified ribozymeε generally maintain 80-90% of the catalytic activity of the native ribozyme when εhort RNA εubεtrateε are employed.

Unlabeled, 5' end-labeled or 3' end-labeled ribozymeε can be used in the assayε. Theεe experimentε can alεo be performed with human cell extractε to verify the obεervationε.

Adminiεtration of Ribozyme

Selected ribozymeε can be adminiεtered prophylactically, or to viruε infected patientε, e.g.. by exogenouε delivery of the ribozyme to an infected tiεsue by means of an appropriate delivery vehicle, e.g. , a liposome, a controlled release vehicle, by use of iontophoresis, electroporation or ion paired molecules, or

covalently attached adducts, and other pharmacologically approved methodε of delivery. Routes of administration include intramuscular, aerosol, oral (tablet or pill form) , topical, syεtemic, ocular, intraperitoneal and/or intrathecal. Expreεεion vectorε for immunization with ribozymeε and/or delivery of ribozymeε are alεo εuitable. Of particular value iε the uεe of the CMV immediate early promoter in expreεεion vectorε becauεe thiε promoter iε known to exhibit very good activity in a number of cell typeε, including epithelial cells.

The specific delivery route of any selected ribozyme will depend on the use of the ribozyme. Generally, a specific delivery program for each ribozyme will focus on naked ribozyme uptake with regard to intracellular localization, followed by demonstration of efficacy. Alternatively, delivery to these same cells in an organ or tissue of an animal can be pursued. Uptake studies will include uptake asεayε to evaluate cellular oligonucleotide uptake, regardleεs of the delivery vehicle or strategy. Such asεayε will alεo determine the intracellular localization of the ribozyme following uptake, ultimately establishing the requirements for maintenance of steady-state concentrations within the cellular compartment containing the target sequence (nucleus and/or cytoplasm) . Efficacy and cytotoxicity can then be tested. Toxicity will not only include cell viability but also cell function.

Some methods of delivery that may be used include: a. encapsulation in liposomeε, b. tranεduction by retroviral vectorε, c. conjugation with choleεterol, d. localization to nuclear compartment utilizing antigen binding or nuclear targeting εites found on moεt εnRNAε or nuclear proteinε,

e. neutralization of charge of ribozyme by uεing nucleotide derivativeε, and f. uεe of blood εtem cellε to diεtribute ribozymeε throughout the body. At leaεt three typeε of delivery εtrategieε are uεeful in the preεent invention, including: ribozyme modificationε, particle carrier drug delivery vehicles, and retroviral expresεion vectorε. Unmodified ribozymeε like moεt εmall moleculeε, are taken up by cellε, albeit εlowly. To enhance cellular uptake, the ribozyme may be modified eεεentially at random, in wayε which reduce its charge but maintains specific functional groups. Thiε reεultε in a molecule which iε able to diffuse acrosε the cell membrane, thuε removing the permeability barrier. Modification of ribozymeε to reduce charge iε juεt one approach to enhance the cellular uptake of these larger molecules. The random approach, however, is not advisable since ribozymes are structurally and functionally more complex than εmall drug moleculeε. The εtructural requirementε necessary to maintain ribozyme catalytic activity are well understood by those in the art. These requirements are taken into consideration when designing modifications to enhance cellular delivery. The modifications are alεo deεigned to reduce susceptibility to nuclease degradation. Both of these characteristics should greatly improve the efficacy of the ribozyme. Cellular uptake can be increased by several orders of magnitude without having to alter the phosphodiester linkages necesεary for ribozyme cleavage activity. Chemical modificationε of the phosphate backbone will reduce the negative charge allowing free diffusion across the membrane. This principle has been succeεsfully demonstrated for antiεenεe DNA technology. The εimilaritieε in chemical compoεition between DNA and RNA make thiε a feaεible approach. In the body, maintenance of an external concentration will be neceεεary to drive the diffuεion of the modified ribozyme into the cellε of

the tiεsue. Adminiεtration routes which allow the diseased tisεue to be expoεed to a tranεient high concentration of the drug, which iε εlowly diεεipated by εyεtemic adεorption are preferred. Intravenouε adminiεtration with a drug carrier deεigned to increaεe the circulation half-life of the ribozyme can be uεed. The εize and compoεition of the drug carrier reεtrictε rapid clearance from the blood εtream. The carrier, made to accumulate at the εite of infection, can protect the ribozyme from degradative proceεεeε.

Drug delivery vehicleε are effective for both systemic and topical administration. They can be designed to serve as a slow release reservoir, or to deliver their contents directly to the target cell. An advantage of using direct delivery drug vehicles iε that multiple molecules are delivered per uptake. Such vehicles have been shown to increase the circulation half-life of drugs which would otherwise be rapidly cleared from the blood stream. Some exampleε of εuch εpecialized drug delivery vehicleε which fall into thiε category are lipoεomeε, hydrogels, cyclodextrins, biodegradable nanocapεuleε, and bioadheεive microsphereε.

From thiε category of delivery εyεtemε, lipoεomes are preferred. Liposomeε increaεe intracellular εtability, increaεe uptake efficiency and improve biological activity.

Liposomes are hollow spherical vesicles composed of lipids arranged in a similar fashion as those lipidε which make up the cell membrane. They have an internal aqueous space for entrapping water soluble compounds and range in size from 0.05 to several microns in diameter. Several studies have shown that liposomeε can deliver RNA to cellε and that the RNA remainε biologically active.

For example, a lipoεome delivery vehicle originally designed as a reεearch tool, Lipofectin, haε been εhown to deliver intact mRNA moleculeε to cellε yielding production of the correεponding protein. In

another study, an antibody targeted liposome delivery system containing an RNA molecule 3,500 nucleotides in length and antisense to a structural protein of HIV, inhibited viruε proliferation in a εequence εpecific manner. Not only did the antibody target the liposomeε to the infected cellε, but it also triggered the internalization of the liposomeε by the infected cells. Triggering the endocytosiε iε useful for viral inhibition. Finally, liposome delivered synthetic ribozymes have been shown to concentrate in the nucleus of H9 (an example of an HIV-sensitive cell) cells and are functional as evidenced by their intracellular cleavage of the sequence. Liposome delivery to other cell types using smaller ribozymeε (leεε than 142 nucleotideε in length) exhibit different intracellular localizations.

Liposomeε offer several advantages: They are non-toxic and biodegradable in composition; they display long circulation half-lives; and recognition molecules can be readily attached to their surface for targeting to tissues. Finally, cost effective manufacture of liposome- baεed pharmaceuticalε, either in a liquid εuspension or lyophilized product, has demonstrated the viability of thiε technology as an acceptable drug delivery system.

Other controlled release drug delivery systems, such as nonoparticles and hydrogelε may be potential delivery vehicleε for a ribozyme. Theεe carrierε have been developed for chemotherapeutic agents and protein- based pharmaceuticalε, and conεequently, can be adapted for ribozyme delivery. Topical administration of ribozymes is advantageous since it allows localized concentration at the εite of adminiεtration with minimal systemic adsorption. This simplifieε the delivery εtrategy of the ribozyme to the diεeaεe site and reduces the extent of toxicological characterization. Furthermore, the amount of material to be applied iε far leεε than that required for other adminiεtration routes. Effective delivery

requires the ribozyme to diffuse into the infected cells. Chemical modification of the ribozyme to neutralize negative charge may be all that .' is required for penetration. However, in the event that charge neutralization is insufficient, the modified ribozyme can be co-formulated with permeability enhancers, such as Azone or oleic acid, in a liposome. The liposomeε can either represent a slow release presentation vehicle in which the modified ribozyme and permeability enhancer transfer from the liposome into the infected cell, or the lipoεome phoεpholipidε can participate directly with the modified ribozyme and permeability enhancer in facilitating cellular delivery. In εome caεeε, both the ribozyme and permeability enhancer can be ^~ formulated into a εuppoεitory formulation for εlow releaεe.

Ribozymeε may alεo be εystemically administered. Systemic absorption refers to the accumulation of drugs in the blood stream followed by distribution throughout the entire body. Administration routes which lead to syεtemic abεorption include: intravenouε, subcutaneous, intraperitoneal, intranasal, intrathecal and ophthalmic. Each of theεe adminiεtration routeε expoεe the ribozyme to an accessible diseaεed tiεεue. Subcutaneouε administration drains into a localized lymph node which proceeds through the lymphatic network into the circulation. The rate of entry into the circulation has been εhown to be a function of molecular weight or εize. The uεe of a lipoεome or other drug carrier localizeε the ribozyme at the lymph node. The ribozyme can be modified to diffuεe into the cell, or the liposome can directly participate in the delivery of either the unmodified or modified ribozyme to the cell. This method is particularly useful for treating AIDS/T-cell leukemia using anti-HIV/HTLV ribozymes of thiε invention. A lipoεome formulation which can deliver ribozymeε to lymphocyteε and macrophageε iε alεo uεeful for the initial εite of enteroviruε replication iε in

lymphoid tisεue of the nasopharynx and gut. Coating of lymphocytes with lipoεomes containing ribozymes will target the ribozymes to infected cells expressing viral surface antigens. Whole blood studieε εhow that the formulation iε taken up by 90% of the lymphocyteε after 8 hourε at 37°C. Preliminary biodistribution and pharmacokinetic studies yielded 70% of the injected dose/gm of tissue in the spleen after one hour following intravenous administration. Also preferred in AIDS therapy iε the uεe of a lipoεome formulation which can deliver oligonucleotideε to lymphocyteε and macrophageε. Thiε oligonucleotide delivery εyεtem inhibitε HIV proliferation in infected primary immune cellε. Whole blood εtudieε εhow that the formulation iε taken up by 90% of the lymphocyteε after 8 hourε at 37°C. Preliminary biodiεtribution and pharmacokinetic studies yielded 70% of the injected dose/gm of tisεue in the spleen after one hour following intravenous administration. This formulation offers an excellent delivery vehicle for anti-AIDS ribozymeε for two reaεonε. Firεt, T-helper lymphocyteε and macrophageε are the primary cellε infected by the viruε, and second, a subcutaneous administration delivers the ribozymes to the resident HIV-infected lymphocytes and macrophages in the lymph node. The liposomes then exit the lymphatic system, enter the circulation, and accumulate in the spleen, where the ribozyme is delivered to the reεident lymphocytes and macrophages.

A liposome formulation which can asεociate ribozymeε with the surface of lymphocytes and macrophages iε also useful. Thiε will provide enhanced delivery to HPV-infected cellε by taking advantage of the εpecificity of macrophage and lymphocyte immune recognition of infected cellε. Whole blood studies show that the formulation is taken up by 90% of the lymphocytes after 8 hours at 37°C. Preliminary biodistribution and pharmacokinetic εtudieε yielded 70% of the injected

dose/gm of tisεue in the spleen after one hour following intravenous administration.

Preferred in EBV therapy is the use of a liposome formulation which can deliver oligonucleotides to lymphocytes and macrophages. This oligonucleotide delivery system inhibitε EBV proliferation in infected primary immune cells. Whole blood studieε show that the formulation is taken up by 90% of the lymphocytes after 8 hours at 37°C. Preliminary biodistribution and pharmacokinetic studieε yielded 70% of the injected doεe/gm of tiεεue in the spleen after one hour following intravenous administration. Thiε formulation offerε an excellent delivery vehicle for anti-EBV ribozymeε for two reaεonε. Firεt, T-helper lymphocyteε and macrophageε are the primary cellε infected by the viruε, and εecond, a subcutaneous administration delivers the ribozymes to the resident EBV-infected lymphocytes and macrophages in the lymph node. The liposomes then exit the lymphatic syεtem, enter the circulation, and accumulate in the spleen, where the ribozyme is delivered to the resident lymphocytes and macrophages.

Preferred in T-cell leukemia therapy is the uεe of a lipoεome formulation which can deliver ribozymes to lymphocytes and macrophages. This ribozyme delivery syεtem inhibitε HTLV proliferation in infected primary immune cellε. Whole blood studies show that the formulation is taken up by 90% of the lymphocytes after 8 hours at 37°C. Preliminary biodistribution and pharmacokinetic εtudieε yielded 70% of the injected doεe/gm of tiεεue in the εpleen after one hour following intravenouε adminiεtration. Thiε formulation offerε an excellent delivery vehicle for anti-HTLV-I or II ribozymeε for two reasons. First, T-helper lymphocytes are the primary cells infected by the viruε, and εecond, a subcutaneous administration delivers the ribozymes to the resident HTLV-infected lymphocytes in the lymph node. The liposomes then exit the lymphatic εyεtem, enter the

circulation, and accumulate in the spleen, where the ribozyme is delivered to the resident lymphocytes and macrophages.

A liposome formulation which can deliver ribozymeε to lymphocyteε and macrophageε is also useful for the initial site of influenza virus replication is in tissueε of the nasopharynx and respiratory system. Coating of lymphocytes with liposomes containing ribozymes will target the ribozymes to infected cells expressing viral surface antigens. Whole blood studies show that the formulation is taken up by 90% of the lymphocytes after 8 hourε at 37°C. Preliminary biodiεtribution and pharmacokinetic εtudieε yielded 70% of the injected dose/gm of tisεue in the εpleen after one hour following intravenouε administration.

A liposome formulation which can asεociate ribozymeε with the εurface of lymphocytes and macrophages is also useful. This will provide enhanced delivery to HSV-infected cells by taking advantage of the specificity of macrophage and lymphocyte immune recognition of infected cells. Whole blood studieε show that the formulation is taken up by 90% of the lymphocytes after 8 hours at 37°C. Preliminary biodistribution and pharmacokinetic studies yielded 70% of the injected dose/gm of tissue in the spleen after one hour following intravenouε adminiεtration.

Intraperitoneal adminiεtration alεo leadε to entry into the circulation, with once again, the molecular weight or εize of the ribozyme-delivery vehicle complex controlling the rate of entry.

Lipoεomeε injected intravenouεly εhow accumulation in the liver, lung and spleen. The composition and size can be adjusted εo that thiε accumulation repreεentε 30% to 40% of the injected doεe. The reεt iε left to circulate in the blood stream for up to 24 hours.

The chosen method of delivery will result in cytoplaεmic accumulation in the afflicted cellε and moleculeε εhould have εome nucleaεe-resistance for optimal doεing. Nuclear delivery may be uεed but iε leεε preferable. Most preferred delivery methods include liposomeε (10-400 nm) , hydrogelε, controlled-releaεe polymers, microinjection or electroporation (for ex vivo treatments) and other pharmaceutically applicable vehicles. The dosage will depend upon the disease indication and the route of administration but should be between 100-200 mg/kg of body weight/day. The duration of treatment will extend through the course of the diseaεe symptoms, usually at least 14-16 days and possibly continuously. Multiple daily doseε are anticipated for topical applicationε, ocular applicationε and vaginal applicationε. The number of doεeε will depend upon diεease delivery vehicle and efficacy data from clinical trials.

Establishment of therapeutic levels of ribozyme within the cell is dependent upon the rate of uptake and degradation. Decreasing the degree of degradation will prolong the intracellular half-life of the ribozyme. Thus, chemically modified ribozymes, e.g.. with modification of the phosphate backbone, or capping of the 5' and 3' ends of the ribozyme with nucleotide analogues may require different dosaging. Descriptions of useful systemε are provided in the art cited above, all of which is hereby incorporated by reference herein.

The claimed ribozymes are also useful as diagnostic tools to specifically or non-εpecifically detect the preεence of a target RNA in a sample. That iε, the target RNA, if preεent in the sample, will be specifically cleaved by the ribozyme, and thus can be readily and specifically detected as smaller RNA species. The presence of such smaller RNA specieε is indicative of the presence of the target RNA in the sample.

Example 46: HIV tat Ribozymes

The 5' exon of tat containε a number of potential cleavage sites. Eighteen of these: 2 GUC siteε,

3 GUA εiteε, 5 AUC εiteε, 3 UCC εiteε, and 5 CUC εites were examined by computer folding and RNaseH asεay to aεεeεε their acceεεibility for ribozyme cleavage.

A measure of the accesεibility of each site to binding by a 13-mer oligonucleotide was preferred in the following way: (a) The first 425 nucleotides- of the clone V sequence (this clone was made available by Dr. Roεεi [City of Hope Medical Reεearch Center; Duarte, CA] and includes the 5' tat exon at nucleotideε 151-366) waε folded on RNAFOLD 4.0 (a generally available program) and examined for the presence of folding domains, i.e. , self-contained structures closed by a stem.

(b) For each potential cleavage site, the domain containing that site was folded to confirm that it folded as in part (a) ; then the domain was refolded while forcing the cleavage site and surrounding nucleotides (11- 16 nucleotides in all) to remain unpaired. The difference in these two folding free energies was taken aε the coεt of melting out that region for base-pairing by a ribozyme or DNA oligonucleotide. (c) The lengths of DNA oligonucleotideε were adjuεted to give predicted delta-G (binding) of -17 to -18 kcal/mole. Thuε, differenceε in overall binding energy waε predicted to be reflected in the free energy differenceε calculated in part (b) . The calculationε are εhown in Table 5.

Eighteen DNA oligonucleotides were made to target the 18 target siteε liεted in Table 5. RNaεeH experimentε were performed uεing approximately 100 nM body labeled RNA tranεcript (clone V), 0.08 U/μl RNaseH (excess RNaseH) and a 2x dilution of 1 mM, 10 μM or 1 μM DNA oligonucleotide. The resultε are shown in Table 6. Eight out of 18 siteε had greater than 40% cleavage at 5 μM

after 10 minutes incubation. Three ribozymes were deεigned against the three most active sequences (H332, H337b, H352, see Table 6) .

Table 6

These three ribozymeε are shown in Figs. 3A, B and C, labeled respectively HDH, HEH and HFH.

Example 47: Ribozymes Containing Thiophosphate

The purpose of this example waε to evaluate the activity of ribozymeε containing εubεtitutionε of thiophoεphate for phoεphate at εome backbone poεitions. Ribozymeε to the HIV-1 tat gene were εyntheεized on an ABI εyntheεizer using standard phosphoramidite chemistry, however, at εtepε where thiophoεphate waε to be incorporated in the backbone the standard oxidation step (involving Iodine) was replaced with an oxidation step utilizing the Beaucage sulfur transfer reagent. Deprotected and deεalted RNA waε gel purified, eluted,

kinaεed and εequenced to enεure that the εequence waε correct. The end-labeled RNA was also treated with H202 which preferentially promotes cleavage at poεitionε containing thiophoεphate. Ribozyme activity waε teεted againεt cleavage of a εhort (12 nucleotides) end-labeled substrate RNA. Substrate concentration was approximately 1 nM; ribozyme concentration was 5-100 nM; incubation was at 37°C in 75 mM Tris (pH 7.5) , 0.1 mM EDTA, 10 mM MgCl ₂ for 2 - 40 minuteε. Cleavage extentε were determined by gel electrophoresis (PAGE) followed by quantitation on the AMBIS.

The ribozymes are shown in Figs. 4A - 4G. The following ribozymes showed esεentially 100% activity: r37 (unmodified) , r37s2 (1 thio on 5' arm, 2 on 3' arm) , r37s4

(3 thio modificationε on each arm) , s37A (3 modifications on 3' arm) , s37B (3 modifications alternating on 3' arm) .

Ribozymes that are fully modified showed some decrease in activity, for example, s37C (almoεt completely modified on substrate binding arms) showed a 3x reduction in activity relative to unmodified, s37D (thio modificationε in stem I, II and III) showed about a 9x reduction in activity relative to unmodified. Example 48: 2'O-methyl-containing Ribozymes Ribozymes were made on an ABI εyntheεizer uεing εtandard phoεphoramidite chemiεtry with 2'-0-methyl phosphoramidite nucleotides used in place of standard nucleotides. Ribozyme activity was determined in the manner described above with PAGE analyεiε. Referring to Fig. 5, a ribozyme containing 2'-0- methyl at all poεitionε that baεe-pair with εubεtrate (except for the A at the 5' side of stem III) was εyntheεized. The Kcat/Km for the 2'-0-methyl ribozyme waε 44 x 10 ⁶ M ^"1 min ^"1 compared to 41 x 10° m ^"1 min ^"1 for the unmodified, and 32 x 10 ⁶ for a thiophoεphate modified ribozyme. Thuε, the 2'-0-methyl ribozyme retainε 100% activity.

Example 49: Cleavage of Short Subεtrate RNAs Correεponding to ICP4 Gene Targetε

Substrate/ribozy e pairs were evaluated for predicted εtructural characteriεticε aε deεcribed above. Nine candidate εubεtrate/ribozyme pairε were tested for their capacity to cleave substrate RNAs in vitro in a ribozyme cleavage assay described above. Of these nine ribozymes, seven were targeted to the 5' end of the ICP4 mRNA and two ribozymes were targeted to the 3' end of the mRNA. Values for the k _cat , k.., and k _cat /k _m (the "cleavage constant") were experimentally determined for each of these ribozymes based upon the asεay reactionε. The reεults are presented in Table 7. The k _cat /k _m values for several of these ribozymes are higher than thoεe recently obεerved for HIV-specific ribozymes in our laboratory. Ribozyme G was extremely active in this assay, with a calculated k _cat /k _m of approximately 1 x 10 ⁸ M" ¹ min' ¹ , the highest level of activity observed using this assay. The generally high level of activity observed suggestε that factorε other than correct folding and the GUC cleavage region influence the structural characteristicε of ribozyme-substrate interaction.

fc.

1.73 min-' 34nM

5.7 min" ¹ 127nM

2.2 min" ¹ 69nM

1/ Nucleotide number relative to EMBL HSV-1 ICP4 gene sequence. Numbers in parentheεeε correεpond to target SEQ ID NOS.

Example 50: Cleavage of Long Subεtrate RNA'ε Corresponding to ICP4 Gene The ability of three ribozymeε to catalyze the cleavage of an approximately 490 nucleotide long ICP4 RNA substrate (corresponding to nucleotideε 1-493 of ICP4 and prepared aε deεcribed above waε determined uεing the ribozyme cleavage assay described above. The assays were preformed in ribozyme excesε and approximate k _cat /k _m values were calculated. The results are presented in Table 8. The cleavage constantε for short subεtrate cleavage by each ribozyme are alεo included for compariεon.

The reεults indicate that, generally, ribozyme cleavage of long subεtrateε proceeds at a slower rate compared to cleavage activities on short subεtrates, although cleavage activity is not altogether eliminated by the use of the longer substrate. The effect of using a

longer substrate varied among the five ribozymes tested, indicating that the accesεibility and εtructural parameterε of different cleavage εites change unpredictably when presented in the context of the longer RNA subεtrateε.

All three of the ribozymeε exhibited a low capacity to catalyze the long substrate relative to their activities on the shorter substrates. Two of the ribozymeε, ribozymes A and E, were relatively more effective at catalyzing cleavage of the longer subεtrate RNA.

Theεe reεultε demonstrate that structural characteristicε of the substrate RNA may influence ribozyme catalytic activity.

17 6 x 10' 100 490 5 X 10 ² 8

17 5.3 X 10 ⁷ 100 490 2 X 10 ³ 17

In nucleotides,

Example 51: Delivery of Stable Ribozyme Into Vero Cellε Ribozymes were end-labeled, encapsulated in liposomeε and delivered to Vero cellε aε deεcribed above. Intracellular ribozymes were purified from cell fractions by phenol/chloroform extraction, and aliquotε were εize fractionated along with newly εynthesized ribozyme samples on denaturing polyacrylamide gels to determine intracellular location of intact ribozymes. By comparison with the migration rate of newly synthesized ribozymes, only the cytoplasmic fraction was observed to contain intact ribozyme. Although intact ribozymes recovered from the cytoplasmic fraction represented a relatively small percentage of the ribozymes encapsulated into the liposome preparationε, theεe results demonstrate that ribozymes can be delivered to and retain εtability within the cell cytoplasm via a liposome delivery vehicle. Example 52: Svnthesiε of Modified Ribozymes with Enhanced Resistance to Nuclease Digeεtion Modified ribozymes were synthesized with 5 phosphorothioate (PS) or 2' O-methyl (2' O-Me) modifications at each end of the molecule, and end-labeled with ³² P~7ATP uεing T4 polynucleotide kinaεe. Unmodified (phosphodiester) ribosomes were synthesized as described above, and similarly end-labeled. Cultured Vero cells were fractionated as described above. The purity of each fraction was assessed by assaying enzyme activitieε known to reεide in that fraction. Aε εhown in Table 10, all enzyme activitieε teεted were found predominantly in the appropriate cell fractionε, indicating an acceptable level of purity. Cell fractionε were frozen at -70° C prior to uεe in a nucleaεe reεiεtance aεεay.

TABLE 9

ENZYME % OF TOTAL ENZYME ACTIVITY

Membrane Nuclear Cytoplasmic Fraction Fraction Fraction

Lactate 3 3 94

Dehydrogenase (cytoplasmic marker)

Calcine 81 3 16

(membrane marker) Lysoεome

Hexoεaminidaεe 75 5 20

Glucocerebrosidase 96 (membrane marker)

Endosome Alkaline 97 Phosphodiesterase (membrane marker)

Equal concentrationε (1-2 pmoleε/100 microliterε) of unmodified and PS or 2' O-Me modified ribozymeε were added to Vero cell cytoplasmic extracts and incubated at 37°C for 0-60 minutes. Aliquots of each were taken at 10 minute intervals, and size fractionated by electrophoresis on denaturing polyacrylamide gels (15%) . Relative rates of ribozyme degradation were analyzed by scanning the gels with an Ambis beta scanner as measure of relative nuclease resistance.

We found that the unmodified ribozyme is quickly digested by the cytoplasmic extract, but that the PS and 2' O-Me modified ribozymes are not digested as rapidly and are thus relatively more nucleaεe reεiεtant. The PS and 2' O-Me modified ribozymes retain 80-90% of the catalytic activity observed with their unmodified ribozyme counterparts, are better able to reεiεt nucleaεe digeεtion, and therefore may have increaεed half-liveε in vivo . In that connection, PS and 2' O-Me modified

ribozymes may have advantages over unmodified ribozymeε for uεe in therapeutic applications.

In addition, we found that nearly equivalent nuclease resiεtance aεεay reεultε were obtained for unlabeled, 5' end-labeled and 3' end-labeled unmodified ribozymeε indicating that the obεerved loss of radioactive label in the PS and 2' O-Me modified ribozymes was not due to dephosphorylation of the 5' end labelε, but inεtead reflects degradation by cytoplasmic nuclease. This suggests that a significant component of Vero cell nuclease activity is endonucleolytic in nature, and it may therefore be deεirable to introduce εimilar modificationε to the internal regionε of the ribozyme molecule in order to beεt optimize nucleaεe reεiεtance. Example 53 : Ribozyme Cleavage of Single Baεe-Subεtituted

Substrates Jn Vitro Unmodified ribozymeε I and E, εpecific for targetε within the ICP4 gene of HSV-1, were synthesized and purified as deεcribed above. A single nucleotide change was incorporated into the substrate cleavage site recognized by ribozyme I. Specifically, the substrate sequence CCGGGGGUCUUCGCGCG was changed at the central C nucleotide to become CCGGGGGUAUUCGCGCG, so that the GUC cleavage site became GUA. This modified subεtrate was syntheεized and tested for in vitro cleavage by ribozyme I using the ribozyme cleavage asεay described above.

Single nucleotide changes were incorporated into the regions complementary to the ribozyme binding arms in the substrate RNAs recognized by ribozymeε I and E aε indicated in Table 11, infra . Computer analyεiε of ribozyme/εubεtrate folding predicted that theεe changeε would result in the maintenance of ribozyme/substrate binding strength. Four such modified subεtrateε were εyntheεized and teεted alongεide their "native" εubεtrate counterpartε for in vitro cleavage by ribozymeε I or E.

Ribozyme I was tested for its ability to cleave two subεtrate εequenceε differing by a εingle baεe in the εubεtrate cleavage εite, one εubstrate correεponding to nucleotides 3550 - 3566 of the ICP4 mRNA εequence, and the other correεponding to the εame εequence with a εingle baεe εubεtitution in the cleavage εite. The k _cat /k _m valueε for cleavage of both subεtrateε are 5.3 x IO ⁷ M ^"1 min" ¹ and 7.3 x IO ⁷ M ^"1 min ^"1 for the native and modified site reεpectively. The reεultε indicate that ribozyme I iε εignificantly more efficient at cleaving the modified εubstrate containing the GUA site. Thus, ribozyme I is not only capable of recognizing and cleaving the "native" ICP4 mRNA target, but is also capable of recognizing and cleaving (more efficiently) that same ICP4 mRNA target with a point mutation in the cleavage site. Natural mutations to particular cleavage sites in viral RNA within infected host cells are posεible. Ribozyme I, therefore, may be a particularly good therapeutic ribozyme given itε target recognition versatility. The resultε alεo εuggest that ICP4 mRNA targets containing a GUA cleavage site would be acceptable if not preferred ribozyme substrateε.

Ribozymeε I and E were alεo teεted for their ability to cleave subεtrateε with single nucleotide modifications in the complementary regions. The resultε, εhown in Table 11, indicate that both ribozymes can recognize and efficiently cleave substrateε with nucleotide point mutationε. The reεultε alεo show that relatively minor changeε in εubεtrate sequence can influence ribozy e/subεtrate interactionε in wayε that were not predicted baεed on binding εtrength alone.

TABLE 10

EFFECT OF SUBSTRATE COMPLEMENTARY REGION MODIFICATIONS ON RIBOZYME CATALYTIC ACTIVITY

Single base changes are indicated in boldface.

Example 54: Effect of Ribozyme Arm Length on Cleavage Activity Variations of ribozymeε E and I having longer binding arm lengths were designed, synthesized and tested together with their prototype ribozymes for cleavage of the substrate recognized by the prototypes using a ribozyme cleavage asεay.

The reεultε are εummarized in Table 11, and indicate that the minimal binding arm lengthε required for effective cleavage vary from one ribozyme to another.

Thiε indicateε that each choεen ribozyme will need to be optimized in its structure on an individual basiε.

TABLE 11

EFFECT OF RIBOZYME BINDING STEM LENGTH UPON SUBSTRATE CLEAVAGE

k-at -lSiii

3.8 x io ⁷ M ^"1 min ^"1

2.8 x IO ⁷ M ^"1 min ^"1

[NA]

3 x IO ⁷ M ^"1 min ^"1

1.6 x IO ⁷ M ^"1 min ^"1

«1 x IO ⁶ M ^"1 min ^"]

1 x IO ⁶ M" ¹ min ^"1

5 x IO ⁷ M ^"1 min ^"1

Example 55:

We have used the ribonuclease protection assay to determine the localization of the HSV ICP4 and ICP27 mRNAs within infected Vero cells. We found that > 80% of the viral mRNAs were situated in the cytoplasm of the infected cells at 4 hours post infection. We have chosen a site within the HSV ICP4 mRNA which is accesεible in RNaεeH aεεayε for further εtudy. We have examined the effectε of altering ribozyme binding arm lengthε on catalytic activity and choεen a ribozyme which haε binding armε of 6 nucleotideε. We call thiε molecule RPI 1197. We have checked the ability of RPI 1197 to cleave the ICP4 mRNA in cell lyεates from infected cellε and found that the amount of ICP4 cleaved by added RPI 1197 increaεed over time and gave the anticipated cleavage productε. Uεing a lipoεome formulation which delivers ribozymeε to

the cytoplasm of cells, we pre-treated cells with two ribozyme molecules (RPI 1197 and RPI 1200 [a 2'-0-Methyl εubεtituted version of 1197]) and tested them for their ability to reduce intracellular levels of the viral ICP4 mRNA and inhibit viral replication. We- found that RPI 1197 and RPI 1200 reduced the mRNA levelε by 10 and 36%, reεpectively, and viral load by 6 and 18%., reεpectively.

Early teεting of the acceεsible sites of the UL5 transcript using the RNaseH asεay haε demonεtrated that a number of regionε within the tranεcript are acceεεible to binding of ribozymeε. We have uεed a 622 nucleotide region (region D) of the UL5 RNA to check the cleavage ability of ribozymeε targeted to both weakly and εtrongly acceεεible εites. We have found that the ribozyme activity in this fragment of RNA εhowed a good correlation between acceεεibility of εiteε and the ability of ribozymeε to cleave the target RNA at thoεe sites. In preliminary experiments, we have observed that the results of the RNaseH assay can also be mimicked using a gel- binding assay in which pieces of complementary nucleic acid (either RNA or DNA) are bound to target RNA and the mixtures are electrophoresed through gels to observe the formation of oligo/target complexes. By varying the concentration of the oligonucleotide, one can determine the accesεible regionε of the RNA aε well aε the binding affinity of the oligonucleotide for that region. We have found a good correlation between the εtrength of the binding complexeε at particular εites and the activity of ribozymes at those sites within the UL5 RNA molecules. Other embodimentε are within the following claims.