60/037,464 filed February 6, 1997.

STATEMENT OF RIGHTS TO INVENTIONS MADE UNDER FEDERALLY SPONSORED RESEARCH Not applicable TECHNICAL FIELD This invention is in the field of agents useful in the targeting and modification of host genes involved in Human Immunodeficiency Virus (HIV) infection. More particularly, it is in the field of triplex-forming oligonucleotides capable of inducing modification of a target sequence at a specific site.

BACKGROUND HIV is known to be responsible for causing acquired immune deficiency syndrome (AIDS). HIV infects various cells of the immune system, in particular, macrophages and T-cells. Infection of macrophages is typical of early stages of infection while, at later stages, infection of T-cells predominates. It has been known for some time that, although the T-cell-specific cell-surface CD4 molecule serves as a receptor for HIV, CD4 alone is not sufficient for viral entry. Hence, it was believed that cofactors existed that, along with CD4, participated in the entry of HIV into the cell. Recently, such cofactors have been identified and partially characterized. For T-cell tropic HIV isolates the fusin protein acts as a co-receptor. Feng et al. (1996) Science 272:872-877. For those isolates of HIV which preferentially infect macrophages (M-tropic), the cofactor is a chemokine receptor known as CCR-5 or CKR-5. Samson et al. (1996) Biochemistry 35:3362-3367. Evidence for the participation of CCR-5 in HIV entry is derived from experiments in which excess (CCR-5 ligand has been shown to block infection by M-tropic HIV. Cocchi et al. (1995) Since 270:1811 - 1815. It has also recently been noted that a naturally-occurring mutation in the

CCR-5 gene, denoted A32, results in protection from HIV infection. Liu et al. (1996) Cell 86:367-377; Samson etna!. (1996) Nature 382:722-725; and Dean et al. (1996) Science 273:1856-1862. The A32 mutation is a 32-nucleotide deletion, causing a shift in the reading frame of the CCR-5 mRNA which leads to premature translation termination to form a truncated protein (see Figure 1).

All patents and publications mentioned herein, both supra and infra, are hereby incorporated by reference.

DISCLOSURE OF THE INVENTION The present invention provides compositions and methods for targeted alteration of the CCR-5 (CKR-5) gene in living cells, using triplex-forming oligonucleotides with attached crosslinking agents. Such alterations can cause heritable nucleotide sequence changes, some of which may result in altered cells that are resistant to infection by HIV.

The compositions and methods of the invention may be used prophylactically, to prevent HIV infection by blocking entry of the virus into macrophages, or therapeutically, to prevent further viral spread in an already-infected individual.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1: The upper portion of the figure shows part of the nucleotide sequence of the CCR-5 (CKR-5) gene (SEQ ID NO: 1). Samson et al., (1996) Biochemistry 35:3362- 3367. The portion shown begins at nucleotide 745. The 12-nucleotide polypurine stretch which serves as a target for the modified ODNs of the invention is boxed. Enclosed within a dashed box and labeled "A32" is the sequence deleted in the A32 mutation, which is characteristic of certain individuals that are resistant to HIV infection. The primer for first strand synthesis for LM-PCR (see Examples E and F) is shown by a horizontal arrow. The EcoRI site used as an internal standard for LM-PCR analysis is indicated. A portion of the complementary strand, written in lower case, shows the sequence surrounding the targeted reaction site and indicates the principal site of reaction (circled G residue).

The lower portion of the figure shows the sequence and structure of some of the TFOs used for modification of the CCR-5 gene. When n=l, the electrophilic moiety is a phenylacetate mustard; when n=3, it is a chlorambucil mustard.

Figure 2: Efficiency and specificity of triplex-directed alkylation of a 60 base-pair model duplex (corresponding to nucleotides 891-950 of the CCR-5 sequence shown in Figure 1) by TFO 1 and TFO 2. The noncoding (polypyrimidine-containing) strand was 32P-labeled. Figure 2A shows alklylation products resulting from reaction of the model duplex with TFO 1 (containing normal G residues); Figure 2B shows alklylation products resulting from reaction of the model duplex with TFO 2 (all G residues substituted by ppG). Results in 2A and 2B are shown as a function of coralyne concentration. Figure 2C shows the location of alkylated bases resulting from reaction with TFO 1 or TFO 2 at 8pM coralyne, determined by heat/piperidine treatment of the reactions, compared to an A+G sequencing ladder of the target sequence.

Figure 3: Targeted modification of HT-29 cell genomic DNA by a modified triplex-forming oligonucleotide. TFO 2 (SEQ ID NO: 6), containing a 5'-terminal chlorambucil moiety and having all G residues replaced by ppG, was incubated with HT- 29 cell genomic DNA as described in Example E. The leftmost lane is a control reaction lacking TFO, showing the product of LM-PCR defined by the upstream EcoRI site. The center lane shows the results of modification at 20"C and the right lane shows results of modification at 37"C. The arrow indicates bands resulting from modification of the target DNA.

Figure 4: Targeted modification of the CCR-5 gene in permeabilized HT-29 cells.

A triplex forming oligonucleotide (SEQ ID NO: 6) with a 5' phenylacetate mustard moiety and a 3' hydroxyhexylphosphate moiety, with all G resisues substituted by ppG, was introduced into HT-29 cells as described in Example F. LM-PCR analysis was performed on DNA extracted from these cells. Lane 1 shows analysis of DNA from cells to which no TFO was added; lane 2 shows analysis of DNA from cells to which were added a non- targeted TFO conjugated to a phenylacetate mustard, lane 3 shows analysis of DNA from cells to which 5 ,uM modified TFO was added, and lane 4 shows analysis of DNA from cells to which 20 ,uM modified TFO was added.

MODES FOR CARRYING OUT THE INVENTION The present invention encompasses modified oligonucleotides (ODNs) capable of forming a triple-stranded structure with a region of the CCR-5 gene. The ODNs are modified by attachment of a cross-linking agent. As shown herein, upon formation of a

triple-stranded complex between the modified ODN of the invention and DNA encoding the CCR-5 gene, the attached crosslinking agent reacts with targeted nucleotides in the DNA of the CCR-5 gene. As a result, a stable change is introduced into the CCR-5 gene.

A. Triplex forming anti-gene oligonucleotides A variation of the "antisense" approach to rational drug design is termed "anti- gene." Whereas antisense ODNs target single stranded mRNA, anti-gene ODNs hybridize with and are capable of inhibiting the function of double-stranded DNA. More specifically, anti-gene ODNs form sequence-specific triple-stranded complexes with a double stranded DNA target and thus interfere with the replication or transcription of selected target genes. Except for certain RNA viruses, nucleic acid-free viroids and prions, DNA is the repository for all genetic information, including regulatory control sequences and non-expressed genes, such as dormant proviral DNA genomes. In contrast, the target for antisense ODNs, which is mRNA, represents a very small subset of the information encoded in DNA. Thus, anti-gene ODNs have broader applicability and are potentially more powerful than antisense ODNs that merely inhibit mRNA processing and translation.

Anti-gene ODNs in the nuclei of living cells can form sequence-specific complexes with chromosomal DNA. The resultant triplexes can inhibit replication and/or transcription of the target double stranded DNA. The DNA binding properties of the ODNs of the invention are strengthened and enhanced by virtue of the crosslinking agents attached to the ODNs, providing a greater likelihood that covalent linkages between the anti-gene ODN and the target DNA sequence will be generated. Such covalent linkages may often result in the generation of one or more mutations at or near the target site, endowing the anti-gene ODNs of the invention with the ability to cause longer lasting effects than those achieved by corresponding antisense inhibition of mRNA function.

Mammalian cell DNA does not undergo turnover; in fact, cells possess sophisticated pathways capable of repairing lesions in DNA that may arise from environmental insults or from spontaneous rearrangements. In contrast, mRNA is transient and may exist only for minutes within a cell. The constant turnover of an mRNA species and the potentially high copy number of such mRNA species suggest that anti-sense ODNs will provide relatively short term effects. By contrast, anti-gene ODNs have the potential to generate more lasting changes and, once within the cell, the ODNs naturally concentrate in the nucleus, their site of action. Cellular uptake of ODNs may be enhanced by attachment of an ODN to carrier,

including but not limited to a lipophilic or lysosomotrophic carrier, with or without an intervening cleavable peptide linker, as disclosed in U.S. Patent 5,574,142.

Anti-gene therapy is based on the observation that certain DNA sequences are capable of forming triple-stranded complexes. In these triple-stranded complexes, the third strand resides in the major groove of the Watson Crick base-paired double helix, where it hydrogen bonds to one of the two parental strands. The triplets indicated below demonstrate the binding code which governs the recognition of base pairs by a third base.

In each case, the third strand base is presented first and is followed by the base pair; hydrogen bonding between the first two bases maintains the third base interaction.

A-A-T G-G-C T-A-T C-G-C Certain limitations of this base pair recognition code are apparent from the allowed triplets. First, there is no capability for the recognition of T-A and C-G base pairs; hence, triple strand formation is restricted to runs of homopurine bases on one strand and homopyrimidine bases on the other strand of the duplex. Second, if the third strand contains cytosine (C), the cytosine residues must be protonated to be able to hydrogen bond to the guanine of a G-C base pair. The pKa for protonation of cytosine is 4.6, suggesting that at physiological pH the stability of C-G-C triads is likely to be impaired.

The substitution of 5-methyl cytosine or the use of polyvalent cations (such as spermine or spermidine) may stabilize the C-G-C triplets at pH 7.0. Third, in all cases triads are maintained by two hydrogen bonds between the third strand base and the purine residue of the base pair. Hence, triple-stranded complexes are generally less stable than the parental double-stranded DNA, which is maintained by a combination of two (A-T) or three (G-C) hydrogen bonds between purine and pyrimidine pairs.

Cytosine/thymidine-, guanine/adenine- and guanine/thymidine-containing ODNs can bind, in a sequence-specific fashion, to homopurine runs in double-stranded DNA.

These recognition motifs are based on Hoogsteen or reverse Hoogsteen base pairing. In the CIT recognition motif, the ODN is parallel to the homopurine strand of the duplex; in the G/A recognition motif, the ODN is anti-parallel to the homopurine strand; in the G/T recognition motif, the ODN may bind parallel or anti-parallel to the homopurine strand of the duplex, depending on the G content of the third strand. These recognition motifs may be sequence-dependent. The sequence specificity of anti-gene ODNs using the CIT

recognition motif permits hybridization of such ODNs to homopurine runs in plasmid DNA and in yeast chromosomes. Since ODN binding is restricted to homopurine runs, it would be advantageous to identify additional heterocyclic bases or base analogues that can recognize the remaining two base pairs, i.e., C-G and T-A. While guanosine can be used in the third strand to recognize T-A base pairs, this interaction involves only one hydrogen bond and is relatively unstable.

By analogy to anti-sense ODNs, anti-gene ODNs can be modified with a variety of pendant groups designed to augment their activity. Hence, the present invention encompasses ODNs modified with intercalating groups, cleaving agents, reporter groups, reactive groups and/or crosslinking moieties appended within or at the termini of anti-gene ODNs. Upon triplex formation, these groups may interact with the adjacent duplex.

Further, in the C/T recognition motif, substitution of 5-methyl cytosine for cytosine in the third strand ODN significantly stabilizes triplexes formed with guanosine-rich homopurine runs. In addition, ODNs with modified backbones, such as oligonucleoside methyl- phosphonates and phosphorothioates, are capable of forming triple-stranded complexes and are encompassed in the present invention.

B. Sequence of CCR-5 The nucleotide sequence of the human CCR-5 gene has been determined. Samson et al., (1996) Biochemistry 35:3362-3367. A portion of the sequence is shown in Figure 1.

Within this sequence is the site ofthe A32 mutation (dashed box in Figure 1), a 32- nucleotide deletion resulting in a change of reading frame which generates a stop codon to produce a truncated CCR-5 protein. Also located within the region of the CCR-5 gene shown in Figure 1 is a segment containing 12 purine residues on one strand and 12 pyrimidines on the complementary strand (boxed in Figure 1). It has now been found that this polypurine string serves as a target for triplex-forming ODNs (TFOs), such TFOs binding to the polypurine string by Hoogsteen or reverse Hoogsteen base pairing. This binding can result in heritable modifications of the CCR-5 gene. Since the triplex-forming region of the CCR-5 gene is in the vicinity of the region encompassed by the A32 mutation, modifications at the triplex-forming region may be expected, in some cases, to have consequences similar to those of the A32 mutation, namely potential resistance to HIV infection.

C. Synthesis and Properties of ODNs As is known in the art, ODNs comprise a chain of nucleotides which are linked to one another by phosphate ester linkages. Each nucleotide typically comprises a heterocyclic base, a sugar moiety attached to the heterocyclic base, and a phosphate moiety which esterifies a hydroxyl function of the sugar moiety. The principal naturally-occurring nucleotides include uracil, thymine, cytosine, guanine or adenine as the heterocyclic base, and ribose or 2'-deoxyribose as the sugar moiety.

The ODNs of the present invention may comprise ribonucleotides, deoxyribonucleotides, or modified sugars or sugar analogues such as are known to those of skill in the art. For example, pentose, deoxypentose, hexose, deoxyhexose, glucose, arabinose, xylose and lyxose are non-limiting examples of sugar moieties which are encompassed by the present invention. The sugar moiety may be in a pyranosyl or a furanosyl form and may be attached to the heterocyclic base in either the a or P anomeric configuration. Preferred sugar moieties of the present invention include the furanosides of ribose, deoxyribose, arabinose and 2'-O-methyl ribose.

The phosphorus derivative or modified phosphate group which may be attached to the sugar or sugar analogue moiety may be a monophosphate, diphosphate, triphosphate, alkylphosphate, alkanephosphate, phosphorothioate, phosphorodithioate or the like. The internucleotide linkages will be those normally found in polynucleotides or those with similar steric properties including, but not limited to, phosphodiester, phosphotriester, phosphorothioate and methyl phosphonate. Preferably, the phosphate group in the ODNs of the present invention will form a phosphodiester. Alternative backbones, such as peptide nucleic acids, are also contemplated by the present invention.

The heterocyclic bases of the modified ODNs of the invention may be those normally found in nucleic acids, as well as those naturally-occurring and synthetic modified bases that are known to those of skill in the art. Examples include, but are not limited to, uracil, thymine, cytosine, 5-methyl cytosine, guanine, adenine, hypoxanthine, pyrrolo [2, 3-d] pyrimidines and pyrazolo [3, 4-d] pyrimidines. Particularly preferred in the present invention are the guanine analogue 6-aminopyrazolo[3,4-clpyrimidin-4(3H)- one (ppG) and the adenine analogue 4-amino-1H-pyrazolo[3,4-d]pyrimidine, both of which facilitate triplex formation by oligonucleotides into which they are incorporated.

See co-owned U.S. Patent Application Serial No. 08/848,373, filed April 30, 1997.

However, one of skill in the art will be aware that a large number of synthetic nucleosides and nucleotides comprising various modified heterocyclic bases and/or sugar analogues are known in the art and that the modified ODNs of the present invention may include one or more modified base and/or sugar moieties, as long as other criteria of the invention (such as ability to form a triple helix) are satisfied. One of skill in the art would appreciate that certain modified nucleotides might even enhance the properties of the ODNs, with respect to certain criteria.

ODNs can be prepared by automated chemical synthesis, using any one of a number of commercial DNA synthesizers, such as those provided by Applied Biosystems.

During automated synthesis, reactive groups such as amino groups can be attached at either end of, or internally to, the ODN, using commercially available reagents, such as are available from Applied Biosystems, Clontech and Glen Research, among others. Such reactive groups serve as a point of attachment for the linker arm A of the crosslinker, to which the leaving group L is attached (see below). Attachment of crosslinking agents to ODNs is accomplished according to the procedures outlined in PCT publication WO 94/17092.

The length and sequence of the ODNs of the invention are such that triplex formation with a target sequence will proceed under normal experimental or physiological conditions.

D. Modification of ODNs with crosslinking agents The ODNs of the invention are modified by the attachment of a cross-linking agent.

The cross-linking agent can be attached at either end of the ODN, or at internal positions.

Upon formation of a triplex, the attached cross-linking agent reacts with nearby nucleotides in the target sequence, leading to modification of the target sequence at a specific site, the site of modification depending on the nature of the cross-linking agent and its position in the ODN.

Any suitable cross-linking agent known in the art can be incorporated into the ODNs of the present invention, provided they meet certain requirements. First, each cross- linking agent must be covalently bonded to a site on the ODN. Second, the length and steric orientation of the cross-linking agent must be such that it can reach a suitable

reaction site in the target DNA sequence after the ODN is hybridized with the target.

Third, the cross-linking agent must have a reactive group which will react with a reactive group of the target DNA sequence. The cross-linking agents can be covalently attached to the heterocyclic bases or base analogues, to the sugar or modified sugar residues, or to the phosphate or modified phosphate functions of the ODNs by any method known in the art.

Any combination of the attachment of one or more cross-linking agents to the ODN is within the scope of the present invention.

Typically, a cross-linking agent comprises two groups or moieties, namely the reactive group, which is typically and preferably an electrophilic leaving group (L), and an "arm" (A) which attaches the leaving group L to the respective site on the ODN. The leaving group L may be chosen from, for example, such groups as chloro, bromo, iodo, SO2R"', or S+R"'R"", where each of R"' and R"" is independently C16 alkyl or aryl or R"' and R"" together form a C16 alkylene bridge. Chloro, bromo and iodo are preferred. Within these groups haloacetyl groups such as -COCH2I, and are 2' bifunctional "nitrogen mustards", such as -N-[(CH2)2-C1]2 are preferred. The leaving group will be altered by its leaving ability. Depending on the nature and reactivity of the particular leaving group, the group to be used is chosen in each case to give the desired specificity.

Although as noted above, linker arm A can be regarded as a single entity which covalently bonds the ODN to the leaving group L, and maintains the leaving group L at a desired distance and steric position relative to the ODN, in practice A can be constructed in a synthetic scheme where a bifunctional molecule is covalently linked to the ODN (for example by a phosphate ester bond to the 3' or 5' terminus, or by a carbon-to-carbon bond to a heterocyclic base) through its first functionality, and is also covalently linked through its second functionality (for example an amine) to a hydrocarbyl bridge (alkyl bridge, alkylaryl bridge or aryl bridge, or the like) which, in turn, carries the leaving group.

A general formula of the cross linking function is thus -A-L, or -A-L2 where L is the above defined leaving group and A is a moiety that is covalently linked to the ODN.

The A moiety itself should be unreactive (other than through L) under the conditions of hybridization of the ODN with the target DNA sequence, and should maintain L in a desired steric position and distance from the desired site of reactions such as an N-7 position of a guano sine residue in the target DNA sequence. Generally speaking, the

length of the A group should be equivalent to the length of a normal alkyl chain of approximately 2 to 50 carbons.

An exemplary more specific formula for a class of preferred embodiments of the cross-linking function is -(CH2)q - Y - (CH2)m - where L is the leaving group, defined above, each of m and q is independently 0 to 8, inclusive, and where Y is defined as a "functional linking group". A "functional linking group" is a group that has two functionalities, for example -NH2 and -OH, or -COOH and - OH, or -COOH and -NH2, which are capable of linking the (CH2)q and (CH2)m bridges. An acetylenic terminus (HCEC-) is also a suitable functionality for Y, because it can be coupled to certain heterocycles, as described, for example, in PCT Publications WO 94/17092 and WO 96/40711.

Other exemplary and more specific formulas for a class of preferred embodiments of the cross-linking function are -(CH2)q - NH - CO - (CH2)m - (X)n - N(R1) - (CH2)p-L and -(CH2)qt - O - (CH2)q., - NH - CO - (CH2)m - (X)n - N(R) - (CH2)p-L where q, m and L are defined as above, q' is 3 to 7 inclusive, q" is 1 to 7 inclusive, X is phenyl or simple substituted phenyl (such as chloro, bromo, lower alkyl or lower alkoxy substituted phenyl), n is 0 or 1, p is an integer from 1 to 6, and R1 is H, lower alkyl or (CH2)p-L. Preferably p is 2. Those skilled in the art will recognize that the structure - N(R,) - (CH2)p-L describes a "nitrogen mustard", which is a class of potent alkylating agents. Particularly preferred within the scope of the present invention are those modified ODNs wherein the cross-linking agent includes the functionality - N(R,) - (CH2)p-L where L is halogen, preferably chlorine; even more preferred are those modified ODNs where the cross linking agent includes the grouping - N-[(CH2)2-L]2 (a "bifunctional" N-mustard).

Particularly relevant to the present invention are those mustards having the structures 4-(p- (bis-(2-chloroethyl)amino)phenyl)butyryl (chlorambucil mustards) and 2-(p-(bis-(2- chloroethyl)amino)phenyl)acetyl (phenylacetate mustards).

In accordance with one aspect of the present invention, a bifunctional N-mustard (or other cross linking function having two reactive groups) is included in the cross-linking agent. One such cross-linking agent attached to the ODN is sufficient, as there is evidence

in accordance with PCT Publication WO 94/17092 that, after hybridization, the modified ODN attaches to both strands of the target double stranded DNA sequence.

Further details regarding the crosslinking agents described above may be found in PCT Publication WO 94/17092; PCT Publication WO 96/40711 and Kutyavin et al. (1994) J. Amer. Chem. Soc. 115:9303-9304.

E. Administration of modified ODN to living cells The modified ODNs of the invention are administered to cells by any method of nucleic acid transfer known in the art, including, but not limited to, transformation, co- precipitation, electroporation, neutral or cationic liposome-mediated transfer, microinjection or gene gun. The modified ODNs may be attached to carriers and/or connected to carriers by cleavable linkers, such carriers and linkers including, but not limited to, those disclosed in co-owned U.S. Patent 5,574,142. The modified ODNs ofthe invention are suitable for in vitro, in vivo and ex vivo therapy and may be administered parenterally, intravenously, subcutaneously, orally or by any other method known in the art. The modified ODNs of the invention can be combined with a pharmaceutically acceptable excipient for administration to a mammalian subject. The formulation of such pharmaceutically acceptable excipients is well within the skill of one in the art. A pharmaceutically acceptable excipient is usually nontoxic and nontherapeutic. Examples of such excipients are water, saline, Ringer's solution, dextrose solution, and Hank' s solution. Nonaqueous vehicles, such as fixed oils, sesame oil, ethyl oleate, or triglycerides may also be used. Parenteral vehicles may also take the form of suspensions containing viscosity-enhancing agents, such as carboxymethylcellulose, sorbitol or dextran. The excipient will also usually contain minor amounts of substances that enhance isotonicity and chemical stability. Examples of buffers include, but are not limited to phosphate buffer, bicarbonate buffer and tris buffer; while examples of preservatives include, but are not limited to thimerosal, m- or o-cresol, formalin and benzyl alcohol. Standard formulations will be either liquids or solids which can be dissolved in a suitable liquid medium as a suspension or a solution. Thus, in a non-liquid formulation, the vehicle may comprise dextrose, human serum albumin, preservatives, etc., to which sterile water, saline, buffer or other solvent can be added prior to administration.

F. Demonstration of specific modification in CCR-5 gene It was previously suggested that modified triplex-forming ODNs might be capable of directing a crosslinking reaction to a specific site in a target sequence within a living cell. Fedorova et al. (1988) FEBSLett. 228:273; Vlassov et al. (1988) Gene 72:313; Povsic and Dervan (1990) J. Am. Chem. Soc. 112:9428; and Shaw et al. (1991) J. Am.

Chem. Soc. 113:7765. It was hypothesized that, if such a crosslinked nucleotide were recognized by cellular repair enzymes, a change in nucleotide sequence might result.

Takasugi et al. (1991) Proc. Natl. Acad. Sci. USA 88:5602-5606. It has now been found that such changes can, in fact, be conferred upon the genome of a living cell, using the modified triplex-forming ODNs described herein.

Several ODNs, capable of triplex formation at the polypurine stretch in the CCR-5 gene, were synthesized. These were modified by the attachment of a crosslinking group (or groups) and the modified ODNs were tested for their ability to generate targeted modification at or near the CCR-5 gene polypurine stretch. The sequences of these ODNs are as follows: 5'-GGAGAAGAAGAG-3' SEQ ID NO: 2 5'-GGAGAAGAAGAGXAAG-3' SEQ ID NO: 3 where X is G or a heterocyclic base that can bind to a pyrimidine.

ODN 1 consists of SEQ ID NO. 2 modified by the attachment of the following crosslinking group at the 5' terminus: where n = 0 or 1. When n = 0, the structure depicts a phenylacetate mustard; when n = 1, the structure depicts a chlorambucil mustard.

The compositions of the invention thus encompass modified ODNs of the sequence depicted in SEQ ID NO: 3 or portions thereof sufficient to effect triple-strand binding to the CCR-5 gene. Pr 'erably, the ODNr have the sequence depicted in SEQ ID NO: 2.

The modified ODNs were tested in three systems: 1) in vitro with a double- stranded ODN target; 2) in vitro with naked genomic DNA as target and 3) in living cells with the cell genome as target. Experimental procedures and results of these tests are presented in the Examples below.

EXAMPLES A. Abbreviations: ClAmb is chlorambucil, 4-(p-(bis-(2-chloroethyl)amino)phenyl)butyryl, attached to either the 5'- or 3'- end of the oligo via an aminohexyl linker. Kutyavin, et al. (1993) J.

Amer. Chem. Soc. 115:9303-04; also described in PCT Publications WO 94/17092 and WO 96/40711.

PhAc mustard is 2-(p-(bis-(2-chloroethyl)amino)phenyl)acetyl, described in PCT Publications WO 94/17092 and WO 96/40711.

B. Oligonucleotide Synthesis All ODNs were prepared from 1 ,umol of the appropriate CPG support on an ABI 394 (Perkin-Elmer) using the protocol provided by the manufacturer. Protected - cyanoethyl phosphoramidites of 2' -deoxynucleosides, CPG supports, deblocking solutions, capping reagents, oxidizing solutions and tetrazole solutions were obtained from Glen Research. The guanine residues in certain ODNs were replaced by 6-aminopyrnzolo[3,4 d]pyrimidine-4(3H)-one, (or ppG), using a phosphoramidite prepared as described by Seela et al. (1988) Helv. Chim. Acta 71:1191-1199. The aminohexyl modification at the 5'-end of certain ODNs was introduced by using a N-(4-monomethoxytrityl)-6-arnino- 1 -hexanol phosphoramidite linker from Glen Research. A 3-hydroxyhexyl phosphate was incorporated into certain ODNs using a modified CPG support as described by Gamper et al. (1993) Nucleic Acids Res. 21:145-150. All other general methods for deprotection, HPLC purification, detritylation and butanol precipitation of ODNs were carried out using standard procedures as described. Gamper et al., supra. ODNs were >95% pure as determined by C-18 HPLC and detection of a single major band by capillary electrophoresis.

C. Preparation of conjugates ODNs containing a 5' aminohexyl tail were conjugated with the 2,3,5,6- tetrafluorophenyl esters of either chlorambucil or phenylacetate mustard as described by Reed et al. (1998) Bioconjugate Chem 9:64-71, and isolated from reaction mixtures by HPLC with 50-70% yield. All manipulations with collected HPLC fractions were performed in ice-cold solutions. Conjugated ODNs were dissolved in water and stored at - 70"C. The integrity of the conjugated nitrogen mustard was assessed as described by Reed et al., supra. Purified conjugates were analyzed by C-18 HPLC (column 250 x 4.6 mm) in a gradient of 5-45% acetonitrile in 0.1 M triethylamine acetate buffer (pH 7.0) over 20 min at a flow rate of 1 ml/min. The reactive conjugates were >90% pure by C-18 HPLC.

D. Modification of a CCR-5 ODN sequence The two strands of the target sequence (see Table 1) were synthesized separately and one ofthe strands was 5'-32P-labeled and mixed in buffer (140 mM KCl, 10 mM MgCl2, 1 mM spermine, 20 mM HEPES pH 7.2) with a two-fold excess of unlabeled complementary strand. The mixture was heated for 1 min at 950C and then incubated at 37"C for 30 min. Then the modified triplex-forming ODN (Table 1) was added. The final concentration of 32P-labeled strand in the mixture was 2 x 10.8 M and of the unlabeled strand was 4 x 10-8 M. The concentration of triplex-forming ODN added to the mixture was 2 x l O-6 M. Prior to the addition of triplex-forming ODN, coralyne was added to the mixture to a final concentration of 10 x 10.6 M. The mixture was incubated at 3 70C for 15 hr, then analyzed by denaturing gel electrophoresis. After electrophoresis, the gel was transferred onto paper, dried down and imaged using a Bio-Rad molecular imager (with its associated computer software) to measure the ratio between crosslinked product and overall radioactivity in each lane of the gel. Results are shown in Table 1.

Table 1 % modification Target Sequence: 3'-CACAGCTTTACTCTTCTTCTCCGTGTC-5' (SEQ ID NO: 4) 1.5% 5'-CTGTCGAAATGAGAAGAAGAGGcAcAG-3' (SEQ ID NO: 5) 63% Triplex-forming ODN: 3'-HO (CH2)6O-pGAGAAGAAGAGGp-O (CH2)6NH-ClAmb (SEQ ID NO: 2) In a separate experiment, a synthetic 60 base-pair duplex target was used. The sequence of this target extended from nucleotides 891-950 as shown in Figure 1. The targeting reaction mixture contained 20 nM labeled duplex 60-mer, 2pM conjugated ODN, 8 M coralyne, 20 mM HEPES, pH 7.2, 140 mM KCl, 10 mM MgCl2, and 1 mM spermine, incubated at 37"C for 4 hours.

Two 12-mer triplex-forming oligonucleotides (TFOs) were used. Both had a sequence antiparallel to the polypurine stretch in the coding strand of the CCR-5 gene (SEQ ID NO: 2) with a chlorambucil moiety conjugated to the 5' terminus through an aminohexyl linker. The two TFOs differed in the nature of the guanine bases present: TFO 1 contained normal guanines, while in TFO 2 all guanines were replaced with 6- aminopyrazolo[3 ,4-d]pyrimidin-4(3H)-one (ppG) (SEQ ID NO: 6). ODNs containing the guanine analogue ppG have increased rates of triplex formation and greater triplex stability. See co-owned U.S. Patent Application Serial No. 08/848,373, filed April 30, 1997. TFO 1 and TFO 2 were designed to alkylate the N7 position of the guanine residue on the noncoding strand opposite C-930.

The results are shown in Figure 2. Alkylation occurred predominantly on the noncoding strand of the target 60-mer (i. e., the strand opposite that containing the homopurine stretch) and the preferred sites of reaction for TFO 1 and TFO 2 were guanines, in the pyrimidine-rich strand, adjacent to the triplex site (Figure 2C). Alkylation of the coding strand did not exceed 2-6%. Figure 2A shows that TFO 1 provided 25% crosslinking in the presence of 8,uM coralyne, 14% at 5 ZM coralyne and 1% at 1 ptM coralyne. No reaction with TFO 1 was observed in the absence of coralyne. Results obtained using TFO 2 show that substitution of guanine residues with ppG had a

significant positive effect on reaction efficiency (Figure 2B). TFO 2 provided 74% alkylation at 8 ,uM coralyne and 2% target alkylation in the absence of coralyne. The effects of ppG on efficiency of alkylation of the CCR-5 gene are consistent with its ability to facilitate triplex formation, as disclosed in co-owned U.S. Patent Application Serial No.

08/848,373, filed April 30, 1997.

E. Modification of the CCR-5 gene in naked genomic DNA 1. Reaction of triplex-forming ODNs with Genomic DNA Genomic DNA from HT29 adenocarcinoma cells (ATCC #HTB-38) was prepared with a Wizard Genomic DNA Purification Kit (Promega), using the protocol provided by the manufacturer. To 5-10 ,ug of genomic DNA in 90 ,uL of 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 140 mM KCl, 10 mM MgCl2, 10 mM HEPES, pH 7.2, and 1 mM spermine was added 10 ptL of a 10X stock of TFO 2 (or controls) to give a final ODN concentration of 2xl O-6 M. After mixing and incubation overnight at 370C, the DNA was pelleted by addition of 10 ptL of 3MNaOAc, pH 7.0, and 300 pL of ice-cold 100% ethanol, chilling (-70"C) and centrifugation at 12,000 rpm for 15 min at 40C, then washed with ethanol and dried.

2. Quantitative Ligation-Mediated PCR (LM-PCR).

Most steps of this technique were performed according to the method of Pfeifer et al. (1993) Meth. in Mol. Biol. 15:153-168; and Mueller etal. (1994) In: Current Protocols in Molecular Biology, ed. F.M. Ausubel et al., John Wiley & Sons, Inc., vol. 2, pp. 15.5.1- 15.5.26, with the modifications described herein.

The first modification was to generate an internal control site by restriction digestion of the DNA after treatment with the reactive TFO, to allow quantitation of the amount of site-specific alkylation. Selection of a restriction endonuclease was based on the enzyme having a recognition site upstream (5') of the alkylated base, and no recognition sites between that base and the downstream (3') sequence complementary to the first strand primer. An EcoRI site upstream of (on the 5' side of) the targeted G residue on the noncoding strand was used. Initial experiments confirmed completion of digestion at this site, using Southern blot hybridization and a PCR-generated probe, and also confirmed the lack of dependence of the ratio of target site cleavage to restriction site cleavage on amount of treated DNA (0.5,ug-5!1g tested) used in the LM-PCR protocol.

The second modification to the protocol was heating the TFO-treated DNA to 950C in the first strand syntheses buffer, at pH 8.9 for 10 min prior to annealing and extending the first primer. This is slightly longer than the standard 3 min of the protocol (Mueller et al., supra), and causes quantitative depurination and cleavage of the DNA at any site of alkylation with the generation of the 5'-phosphate required for the ligation step. All chlorambucil alkylation sites in this work were designed to be on N-7 of guanines.

The procedure for LM-PCR was as follows. TFO-treated DNA samples were digested to completion with EcoRI by incubation for 3 hr under optimal conditions (according to the manufacturer) with a three-fold excess of restriction enzyme. The volume was adjusted to 100 ptL with water and the DNA was precipitated with ethanol.

The pellet was resuspended in 10 mMTris-HCl, pH 7.5, 1 mMEDTA, to give a DNA concentration of about 0.5 ,ug/pl. To 5 A of this chilled solution in a PCR tube was added 25 A of the first strand synthesis solution (Mueller et al., supra). First strand synthesis was primed with the following oligonucleotide: 5'-TCCATACAGTCAGTATCAATTCTGG-3' (SEQ ID NO: 7) Ligation of the universal linker (Mueller et al., supra) was performed as described by Mueller and coworkers supra, except that the 95° heating step was extended to 10 min.

Nested PCR was then performed as described by Pfeifer et al., supra. The second primer had the sequence: 5'-TCCAGACATTAAAGATAGTCATCTTGG-3' (SEQ ID NO: 8) The third primer was labeled with 32P at its 5' end and had the sequence: 5'-CATTAAAGATAGTCATCTTGGGGCTGG-3' (SEQ ID NO: 9) Phosphorimaging was used to analyze the'polyacrylamide gel electropherograms of the LM-PCR results, and the efficiency of targeted alkylation was calculated from the ratio of intensities for the restriction- and alkylation- induced bands. For each experiment, an aliquot of the genomic DNA was treated with dimethlysulfate and amplified along with the rest of the samples to provide a G-ladder (Pfeifer et al., supra).

3. Results LM-PCR analysis of the reaction of TFO 2 (SEQ ID NO: 6) containing a chlorambucil moiety conjugated to the 5'-terminus through a hexylamine linker, with isolated human genomic DNA, is shown in Figure 3. An image of an electrophoretic gel shows LM-PCR products of unmodified (leftmost lane) and TFO 2-modified target DNA

(center lane, modification at 200C; rightmost lane, modification at 370C). Since one of the CCR-5 alleles of the HT-29 cells used in this study has the A32 mutation, and the primer overlaps the site of the A32 deletion, only a single EcoRI band is obtained, corresponding to the wild-type CCR-5 allele. The sequences of the two strands of the target are shown in Table 1 (SEQ ID NOs: 4 and 5) above the sequence of the modified triplex-forming ODN (SEQ ID NO: 2). The modified strand corresponds to the upper of the two target strands in Table 1(SEQ ID NO: 4) and is modified primarily at the G residue five nucleotides from the 5' end of the sequence of that strand shown in Table 1. Almost quantitative modification of the targeted genomic DNA was obtained with 2 piM TFO 2 in the presence of 8 piM coralyne. Only trace target alkylation was detected by LM-PCR when modification was conducted in the absence of coralyne.

F. Modification of CCR-5 gene in living cells HT-29 cells were plated into 6-well 35 mm plates at 4.0 x 105 cells per well and were allowed to adhere for 4 hr at 370C in complete medium. The cells were then washed with phosphate-buffered saline (PBS) and treated for 5 min at 370C with 350 A of permeabilization buffer (137 mM NaCl, 100 mM PIPES, pH 7.4, 5.6 mM glucose, 2.7 mM KCl, 2.7 mM EGTA, 1 mM ATP, 0.1% bovine serum albumin) containing 500 Units/ml Streptolysin O (preactivated for 15 min at room temperature in the presence of 2.5 M DTT), 8 piM coralyne and 5 piM or 20 !1M of TFO 2 (SEQ ID NO: 6) with a phenylacetate mustard moiety conjugated to the 5' terminus through a hexylamine linker. The phenylacetate mustard has a slightly longer half-life in cells than does the chlorambucil mustard. This TFO was also modified with a 3'-hydroxyhexyl phosphate (as described in Example B) to slow the rate of nuclease digestion within cells. Streptolysin 0 was used to render cells permeable to ODNs. Spiller et al. (1995) Antisense Res. Dev. 5:13-21. After treatment, 5 ml complete medium was added to each well and cells were incubated for another 6 hr at 370C. After incubation, cells were washed with PBS, trypsinized and DNA was isolated as described in Example E. DNA modification was determined by LM-PCR, as described in Example E.

The results obtained in this experiment are depicted in Figure 4. The image of an electrophoretic gel shows LM-PCR products of target DNA from cells that were untreated (lane 1), treated with a nontargeting TFO with a conjugated phenylcetate mustard (lane 2),

or treated with 5 RM (lane 3) or 20 piM (lane 4) phenylacetate-modified TFO. The efficiency of site-directed DNA modification was quantitated by comparison of the PCR product obtained from modification at the targeted G residue to the PCR product defined by the EcoRI site. Analysis of these and similar results indicate that the efficiency of site- directed DNA modification, obtained using the procedure described above, was as follows: 5 piM reactive ODN - 5 % modification 10 piM reactive ODN - 12 % modification 20 pM reactive ODN - 24 % modification Both the presence of coralyne and substitution of guanine residues by ppG in this TFO were important for obtaining the levels of genomic modification that were observed.

These results demonstrate site-directed covalent modification of a native mammalian gene in intact cells by a phosphodiester ODN with an electrophilic reactive group attached, providing direct evidence for efficient formation of a triplex in an endogenous cellular gene in its native chromatin structure. It is likely that repair and/or replication of a gene so modified would result in mutation, possibly leading to loss or inhibition of function of the CCR-5 gene. Accordingly, the methods and compositions of the invention provide a unique approach for the study of the CCR-5 co-receptor in the initiation and progression of HIV infection, as well as new approaches toward modification of CCR-5 function in the prevention and treatment of HIV infection.

Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be apparent to those skilled in the art that certain changes and modifications may be practiced. Therefore the foregoing descriptions and examples should not be construed as limiting the scope of the invention, which is delineated by the appended claims.

SEQUENCE LISTING (1) GENERAL INFORMATION: (i) APPLICANT: Meyer, Rich B.

Kutyavin, Igor V.

(ii) TITLE OF INVENTION: TARGETED MODIFICATION OF THE CCR-5 GENE IN LIVING CELLS (iii) NUMBER OF SEQUENCES: 9 (iv) CORRESPONDENCE ADDRESS: (A) ADDRESSEE: MORRISON & FOERSTER (B) STREET: 755 PAGE MILL ROAD (C) CITY: PALO ALTO (D) STATE: CA (E) COUNTRY: USA (F) ZIP: 94304-1018 (v) COMPUTER READABLE FORM: (A) MEDIUM TYPE: Floppy disk (B) COMPUTER: IBM PC compatible (C) OPERATING SYSTEM: PC-DOS/MS-DOS (D) SOFTWARE: PatentIn Release #1.0, Version #1.30 (vi) CURRENT APPLICATION DATA: (A) APPLICATION NUMBER: US (B) FILING DATE: (C) CLASSIFICATION: (viii) ATTORNEY/AGENT INFORMATION: (A) NAME: Brennan, Sean M.

(B) REGISTRATION NUMBER: 39,917 (C) REFERENCE/DOCKET NUMBER: 34469-20001.40 (ix) TELECOMMUNICATION INFORMATION: (A) TELEPHONE: (650) 813-5600 (B) TELEFAX: (650) 494-0792 (C) TELEX: 706151 MRSNFOERS SFO (2) INFORMATION FOR SEQ ID NO:1: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 360 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:1: TCCCAGGAAT CATCTTTACC AGATCTCAAA AAGAAGGTCT TCATTACACC TGCAGCTCTC 60 ATTTTCCATA CAGTCAGTAT CAATTCTGGA AGAATTTCCA GACATTAAAG ATAGTCATCT 120 TGGGGCTGGT CCTGCCGCTG CTTGTCATGG TCATCTGCTA CTCGGGAATC CTAAAAACTC 180 TGCTTCGGTG TCGARATGAG AAGAAGAGGC ACAGGGCTGT GAGGCTTATC TTCACCATCA 240 TGATTGTTTA TTTTCTCTTC TGGGCTCCCT ACAACATTGT CCTTCTCCTG AACACCTTCC 300 AGGAATTCTT TGGCCTGAAT AATTGCAGTA GCTCTAACAG GTTGGACCAA GCTATGCAGG 360 (2) INFORMATION FOR SEQ ID NO:2: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 12 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2: GGAGAAGAAG AG 12 (2) INFORMATION FOR SEQ ID NO:3: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 17 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ix) FEATURE: (A) NAME/KEY: misc feature (B) LOCATION: 13 (D) OTHER INFORMATION: /note= "G or any heterocyclic base that can bind to a pyrimidine" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:3: GGAGAAGAAG AGNAAAG (2) INFORMATION FOR SEQ ID NO:4: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 27 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: CTGTGCCTCT TCTTCTCATT TCGACAC (2) INFORMATION FOR SEQ ID NO:5: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 27 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:5: GTGTCGAAAT GAGAAGAAGA GGCACAG (2) INFORMATION FOR SEQ ID NO:6: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 12 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ix) FEATURE: (A) NAME/KEY: modified~base (B) LOCATION: group(l..2, 4, 7, 10, 12) (D) OTHER INFORMATION: /mod~base= OTHER /note= "6-aminopyrazolo[3,4-d]pyrimidin-4(3H)-one (ppG)" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:6: NNANAANAAN AN 12 (2) INFORMATION FOR SEQ ID NO:7: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 25 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:7: TCCATACAGT CAGTATCAAT TCTGG 25 (2) INFORMATION FOR SEQ ID NO:8: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 27 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:8: TCCAGACATT AAAGATAGTC ATCTTGG (2) INFORMATION FOR SEQ ID NO:9: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 27 base pairs (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:9: CATTAAAGAT AGTCATCTTG GGGCTGG