DNAMOLECULES AND METHODS

The present invention relates to DNA molecules and methods using the molecules for introducing mutations into DNA in a Gram positive bacterial cell, particularly a cell of the class Clostridia.

The class Clostridia includes the orders Clostridials, Halanaerobiales and Thermoanaerobacteriales . The order Clostridials includes the family Clostridiaceae, which includes the genus Clostridium.

Clostridium is one of the largest bacterial genera. It is composed of obligately anaerobic, Gram-positive, spore formers. Certain members may be employed on an industrial scale for the production of chemical fuels, eg., Clostridium thermocellum and Clostridium acetobutylicum. This latter clostridial species, together with other benign representatives, additionally has demonstrable potential as a delivery vehicle for therapeutic agents directed against cancer. However, the genus has achieved greatest notoriety as a consequence of those members that cause disease in humans and domestic animals, eg, Clostridium difficile, Clostridium botulinum and Clostridium perfringens .

Despite the tremendous commercial and medical importance of the genus, progress either towards their effective exploitation, or on the development of rational approaches to counter the diseases they cause, has been severely hindered by the lack of a basic understanding of the organisms' biology at the molecular level. This is largely a consequence of an absence of effective genetic tools.

In recent years, the complete genome sequences of all of the major species have been determined from at least one representative strain, including C. acetobutylicum, C. difficile, C. botulinum and C. perfringens. In other bacterial species such knowledge can act as a springboard for more effective disease management or for the generation of strains with improved process properties. A pivotal tool in such undertakings is the ability to rationally integrate DNA into the

genome. Such, technology ma3' be employed: (i) to generate specific mutants as a means of ascribing function to individual genes, and gene sets, as an essential first step towards understanding physiology and pathogenesis; (ii) to insertionally inactivate regulatory or structural genes as a means of enhancing the production of desirable commercial commodities, and; (iii) to stably introduce genetic information encoding adventitious factors. However, there are currently no effective integration vectors for mutational studies in any Clostridium sp. and the ability to insertionally inactivate genes in the genus remains woefully inadequate.

Previous attempts to make mutants in Clostridium sp. relied on homologous recombination between an integration vector and the host chromosome In C. perfringens strain 13, C. beijerinclάi NCIMB 8052, C. acetobutylicum ATCC 824 and C. difficile CD37, replication-minus plasmids carrying regions of the host chromosome have been shown to integrate into the genome via homologous recombination (Shimizu et al (1994) J. Bacteriol. 176: 1616-23; Wilkinson and Young (1994) Microbiol. 140: 89-95; Green et al (1996) Microbiol. 142: 2079- 2086; Liyanage et al (2001) Appl. Environ. Microbiol. 67: 2004-2010). In the case of C. beijerincldi and C. difficile, vectors were mobilized from E. coli donors. In C. perfringens and C. acetobutylicum, plasmids were introduced by transformation. Integrants arose in C. beijerinckii at frequencies of 10 ^" to 10 ^" per recipient, which represented some two orders of magnitude lower than the transfer frequency observed (10 ^"4 to 10 ^"5 ) with replication proficient plasmids (Wilkinson and Young, 1994, supra). In the case of C. difficile, no indication of the frequencies attained was reported (Liyanage et al, 2001, supra). In C. acetobutylicum, integrants arose at a frequency of 0.8 to 0.9 'colonies' per μg of DNA (Green et al, 1996, supra). In the above integrants, plasmid sequences at the target site were flanked by two directly repeated copies of the DNA segment directing integration. As a consequence, they were segregationally unstable, e.g., losses per 30 generations of between 1.8 to 3.0 X 10 ^"3 for C. acetobutylicum (Green et al, 1996, supra) and between 0.37 to 1.3 X 10 ^"3 for C. beijerincldi (Wilkinson and Young, 1994, supra).

It follows that integrants resulting from allelic exchange are preferred. Accordingly, double crossover mutants were sought and obtained in C. perfringens (Awad et al (1995) MoI. Microbiol. 15: 191-202; Bannam et al (1995) MoI. Microbiol. 16: 535-551. However, allelic exchange only proved possible through the inclusion of rather long (3.0 kb) regions of homology on either side of the antibiotic resistance gene employed to inactivate the target gene. Furthermore, even with this provision, the isolation of mutants proved highly variable (i.e., pic mutants were only obtained in 2 of 10 independent experiments), and many mutants can take up to 6 months to isolate, while others may never be isolated at all. Rare integration events could be detected in C. perβ-ingens as a consequence of the high frequency with which DNA can be transformed into this organism. Attempts to generate double crossover mutants in other clostridial species have been unsuccessful.

To date the generation of mutants in a range of clostridial species, other than C. perfi'ingens, using classical homologous recombination has proven difficult. Thus, only five mutations have ever been made in C. acetobutylicum. Four (butK, CAC3075; pta, CAC1742; aad, CACPOl 62, and; solR, CACP061) were made by single cross-over integration of a replication deficient plasmids (Green et al., 1996, supra; Green and Bennett (1996) Appl. Biochem. Biotechnol. 213, 57-58; Harris et al (2002) J Bacteriol. 184, 3586-3597) while a fifth in spoOA (CAC2071) was isolated by a strategy which attempted, but did not succeed, in the generation of a mutant by reciprocal exchange using a replication-defective plasmid (Nair et al (1999) J. Bacteriol. 181, 319-330). Similarly, the generation of only three directed mutants has been reported in C. difficile. One mutant (gldA, CD0274) was generated using a replication-deficient plasmid (Liyanage et al, 2001, supra) although this event appeared to be lethal and mutant cells could not be propagated. The other two genes inactivated (rgaR, CD3255 and rgbR, CDl 089) arose following the introduction of a replication-defective plasmid carrying internal fragments of the two structural genes (O'Connor et al (2006) MoI. Microbiol. 61, 1335-1351)., These latter plasmids were apparently

introduced with "some difficulty" and whilst integrants were isolated, no isolation frequencies were noted. Indeed, an assessment of the efficiencies of the mutagenesis procedures previously used in both organisms is difficult to make, as no indication of the frequency with which mutants are generated is generally presented. In the case of C. acetobutylicum it is acknowledged (Thomas et al, (2005) Metabolic engineering of soventogenic Clostridia. In: Dϋrre, P. Handbook on Clostridia, CRC Press, pp 813-830) to be "less than one transformant per μg plasmid DNA". Moreover, as the majority of these mutants are made by single cross-over insertion, they are unstable due to plasmid excision. For example, Southern blotting of the C. difficile rgaR mutant revealed the presence of "looped out", independently replicating plasmid in some cells in the population (O'Connor et al, 2006, supra).

Increasingly, technologies are being devised which capitalise on the systems involving mobile genetic elements to bring about more effective modification of bacterial genomes. The Group II intron Ll XtrB of Lactococcus lactis is an element that mediates its own mobility through the action of an intron-encoded reverse transcriptase (LtrA) and the excised lariat RNA. Furthermore, it may be re-targeted to virtually any desired DNA sequence through modification of the intron RNA (Guo et al (2000) Science 289: 452-457; Mohr et al (2000) Genes Dev. 14: 559-573). Thus, by appropriately mutating individual bases in the 15 bp region of the intron involved in targeting, Karberg et al (Nature Biotech. (2001) 19: 1162-1167) were able to direct the insertion of the element into distinct, defined positions within several different E. coli genes at frequencies of between 0.1 to 22%. Disruption of one of these genes, thyA, gives rise to dories that are naturally trimethoprim resistant. Thus, integrants could be selected for by culturing in the presence of trimethoprim. Integrants in other genes were identified by screening individual colonies for the presence of the Ll XtrB intron. The plasmid used to disrupt the thyA gene in E. coli was also used to disrupt the thyA gene in S. flexneri and in S. typhimurium. Trimethoprim resistant colom ^' es were obtained at a frequency of 1% and 0.3% respectively.

The Group II intron Ll.LtrB of Lactococcus lactis was used to generate knockouts in the pic gene of C. peφingens (Chen et al (2005) Appl Environ Microbiol. 71: 7542-7). A chloramphenicol resistant plasmid containing, inter alia, a modified Ll.LtrB intron designed to target the pic gene was electroporated into C. peφingens. Transformants were selected on chloramphenicol and were tested for the presence of the insertion in the pic gene by PCR. Of 38 colonies tested, most were negative for the insertion but two colonies contained both wild-type and intron-inserted pic gene. The latter colonies were deemed to have arisen from a single transformed bacterium, which gave rise to progeny in which the insertion occurred and progeny in which the insertion did not occur. Bacteria from these mixed colonies gave rise to pure clones, 10% of which contained intron-inserted pic gene. Thus, insertion mutants were identified via two rounds of screening without the need for selection for growth on an antibiotic, other than selection on chloramphenicol for transformation. In fact, the lack of any introduction of an antibiotic resistance gene into the chromosome was identified as a particular advantage of the method. In particular, the authors envisaged that the method could be used to construct multiple gene disruptions in the same bacterial cell using the same shuttle plasmid carrying different modified Ll.LtrB introns. The frequency of transfer to C. perfringens is high, some two orders of magnitude greater than other Clostridial species. Moreover, the gene knockout (in pic) gives rise to an easily detected phenotype, which may be visualised readily on agar plates.

Yao et al (RKfA (2006) 12: 1-11) used Ll.LtrB to disrupt genes without selection in Staphylococcus aureus. A cadmium-induced promoter was used to direct expression of the Ll.LtrB intron in S. aureus; induction with cadmium was beneficial to obtaining insertion mutants in one gene. When mutants were made in another gene, all colonies tested positive for insertion of the intron in the absence of cadmium.

Zhong et al (Nucleic Acids Res. (2003) 31: 1656-64) described a method of positively selecting for re-targeting of the Group II intron involving inserting into

the Group II intron a "retrotraiisposition-activated selectable marker" or RAM consisting of a trimethoprim (Tp) resistance cassette containing the td intron of phage T4. The Tp resistance gene encodes a type II dihydrofolate reductase. The td intron is a Group I intron, i.e. a self-catalytic RNA-element which, in its correct orientation, can splice itself from an RNA transcript in which it is located. The orientation in which td is inserted into Tp ^R is such that when the gene is transcribed, the element is not spliced. Thus, the mRNA remains mutant, and the protein required for Tp resistance is not produced. ^" When the Group II element is transcribed into RNA, during re-targeting, the opposite strand of the RAM is now present in an RNA form. Under these circumstances the td element is orientated correctly, and is spliced. As a consequence, when the Group II element retargets to the chromosome, the Tp ^R gene has lost its td insertion, and is now functional. As a consequence, cells in which successful re-targeting has taken place are Tp resistant. They may therefore be directly selected. The method was used in Escherichia coli cells.

Clostridial species are frequently resistant to trimethoprim, making the use of a RAM based on a Tp resistance cassette unworkable. For instance in the study of Swenson et al (1980) Antimicrob. Agents Chemother. 18: 13-19 the vast majority of the isolates tested were resistant. Resistance is also common in the nonpathogenic, industrially useful strains. Indeed, the intrinsic resistance of C. cellulolyticum forms the basis of the conjugation method used for gene transfer experiments in Jennert et al (2000) Microbiology. 146: 3071-80.

A kit for performing gene knockouts (principally in E. coli) based on a RAM consisting of a kanamycin resistance (Km ¹ ) cassette is marketed as "TargeTron™ Gene Knockout System" by Sigma-Aldrich. Clostridium spp. are naturally resistant to kanamycin, so kanamycin resistance cannot be used as a selection marker in Clostridium.

The inability to make defined gene knock-outs in Clostridial genomes, by reciprocal marker exchange, is a major impediment to the commercial exploitation

of members of the ■ class Clostridia, and particularly the genus Clostridium. It impinges on all areas. Thus, the application of metabolic engineering to generate industrial stains with improved fermentation characteristics presently cannot be contemplated (eg C. acetobutylicum and the Acetone-Butanol fermentation process; strains carrying chromosomalfy located therapeutic genes useful in cancer therapy cannot be generated (a prerequisite for clinical trials, eg C. sporogenes and Clostridial-DivQcted Enzyme Prodrug Therap}'); and fundamental information on pathogenic mechanisms, an essential first step in the formulation of effective countermeasures, is being severely impaired (eg C. difficile and hospital-acquired infections).

The listing or discussion of a prior-published document in this specification should not necessarily be taken as an acknowledgement that the document is part of the state of the art or is common general knowledge.

The inventors have devised DNA molecules and methods which allow for the efficient insertion of DNA into the genome of Clostridium spp and other bacteria of the class Clostridia, thereby allowing the targeted mutation of genes in the genome.

A first aspect of the invention provides a DNA molecule comprising:

a modified Group II intron which does not express the intron-encoded reverse transcriptase but which contains a modified selectable marker gene in the reverse orientation relative to the modified Group II intron, wherein the selectable marker gene comprises a region encoding a selectable marker and a promoter operably linked to said region, which promoter is capable of causing expression of the selectable marker encoded by a single copy of the selectable marker gene in an amount sufficient for the selectable marker to alter the phenotype of a bacterial cell of the class

Clostridia such that it can be distinguished from the bacterial cell of the class Clostridia lacking the selectable marker gene; and

a promoter for transcription of the modified Group II intron, said promoter being operably linked to said modified Group II intron; and

wherein the modified selectable marker gene contains a Group I intron positioned in the forward orientation relative to the modified Group II intron so as to disrupt expression of the selectable marker; and

wherein the DNA molecule allows for removal of the Group I intron from the RNA transcript of the modified Group II intron to leave a region encoding the selectable marker and allows for the insertion of said RNA transcript (or a DNA copy thereof) at a site in a DNA molecule in a bacterial cell of the class Clostridia.

Group II introns are mobile genetic elements which are found in eubacteria and organelles. In nature, they use a mobility mechanism termed retrohoming, which is mediated by a ribonucleoprotein (RNP) complex containing the intron-encoded reverse transcriptase (IERT) and the excised intron lariat RJSfA. It is believed that the excised intron RNA inserts directly into one strand of a double-stranded DNA target site by a reverse splicing reaction, while the IERT also site-specifically cleaves the opposite strand and uses the 3 '-end of the cleaved strand for target

DNA-primed reverse transcription (TPRT) of the inserted intron RNA. As a result, the intron (and any nucleic acid carried in a modified intron) are inserted into the target DNA. The TPRT system requires only the IERT and the excised intron RNA (see Saldanha et al (1999) Biochemistry 38, 9069-9083). Details of Group II introns are found in Karberg et al (2001) Nature Biotechnology 19, 1162-

1167, incorporated herein by reference, and in references cited therein.

The IERT is also known in the art as the intron-encoded protein (IEP). The IEP (IERT) has reverse transcriptase activity as well as endonuclease and maturase activities which allow a copy of the intron to be inserted into DNA.

The process of cleaving the ^' DNA substrate and inserting nucleic acid molecules involves base pairing of the Group π intron RNA of the RNP complex to a specific region of the DNA substrate. Additional interactions occur between the intron-encoded reverse transcriptase and regions in the DNA substrate flanking the recognition site. Typically, the Group II intron RNA has two sequences, EBSl and EB S2, that are capable of ^' hybridizing with two intron RNA-binding sequences, IBSl and IBS2, on the top strand of the DNA substrate. Typically, the Group II intron-encoded reverse transcriptase binds to a first sequence element and to a second sequence element in the recognition site of the substrate. Typically, the Group II intron RNA is inserted into the cleavage site of the top strand of the DNA substrate. The first sequence element of the recognition site is upstream of the putative cleavage site, the IBSl sequence and the IBS2 sequence. The first sequence element comprises from about 10 to about 12 pairs of nucleotides. The second sequence element of the recognition site is downstream of the putative cleavage site and comprises from about 10 to about 12 nucleotides.

As denoted herein, nucleotides that are located upstream of the cleavage site have a (-) position relative to the cleavage site, and nucleotides that are located downstream of the cleavage site have a (+) position relative to the cleavage site. Thus, the cleavage site is located between nucleotides -1 and +1 on the top strand of the double-stranded DNA substrate. The IBSl sequence and the IBS2 sequence lie in a region of the recognition site which extends from about position -1 to about position -14 relative to the cleavage site.

Typically, EBSl is located in domain I of the Group II intron RNA and comprises from about 5 to 7 nucleotides that are capable of hybridizing to the nucleotides of the IBSl sequence of the substrate.

Typically, EBS2 is located in domain I of the Group II natron RNA upstream of EBSl and comprises from about 5 to 7 nucleotides that are capable of hybridizing to the nucleotides of 1BS2 sequence of the substrate.

In order to cleave the substrate efficiently, it is preferred that the nucleotide or sequence, which immediately precedes the first nucleotide of EBSl of the Group II intron RNA ₅ be complementary to the nucleotides at +1 in the top strand of the substrate.

The modified Group II intron contained in the DNA molecule of the invention does not express the IERT. Preferably, the Group II intron does not contain a functional open reading frame for the IERT. Preferably, domain IV of the Group II intron, which typically contains the IERT is partially deleted such that it does not contain the IERT.

Various Group II introns which may be useful in the practice of the invention are known. These include bacterial introns such as the eubacterial introns reviewed in Dia and Zimmerly (2002) Nucleic Acids Res. 30: 1091-1102, and also include mitochondrial and chloroplast introns referred to in Zimmerly, Hausner and Wu (2001) Nucleic Acids Res. 29: 1238-1250. It is preferred if the Group II intron is the Lactococcus lactis Ll .LtrB intron (Mohr et at (2000) supra). The IERT in this Group II intron is the LtrA protein. The all and aI2 nucleotide integrases of Saccharomyces cerevisiae are also suitable.

Another alternative is the Group II intron from the clostridial conjugative transposon Tn5397 (Roberts et al (2001) J. Bacteriol. 183: 1296-1299).

The LtrA RNP complex comprises an excised, wild-type or modified excised Group Ll.LtrB Group II intron RNA of the Lactococcus lactis LtrB gene, hereinafter referred to as the "Ll.LtrB intron" RNA ₅ and a wild-type or modified

Ll. LtrB intron-encoded reverse transcriptase, referred to as the LtrA protein. The

EBSl of the Ll.LtrB intron RNA comprises 7 nucleotides and is located at positions 457 to 463. The EBSl sequence of the wildtype Ll.LtrB intron RNA has the sequence 5'-GUUGUGG (SEQ ID No. 1). The EBS2 of the Ll.LtrB intron RNA comprises 6 nucleotides and is located at positions 401 to and

including 406. The EBS2 sequence of the wild-type Ll.LtrB intron RNA has the sequence 5 ¹ AUGUGU (SEQ ID No. 2).

The Group II intron in the DNA molecule of the invention has been modified to include a modified selectable marker gene. A selectable marker gene is any gene which confers an altered phenotype in a bacterial cell in which it is expressed, compared to the bacterial cell in which it is not expressed. The modified selectable marker gene is modified (compared to the unmodified selectable marker gene) by containing a Group I intron which disrupts the expression of the selectable marker. The term "unmodified selectable marker gene" includes a gene comprising a promoter and a coding region of a gene, where the promoter is not the promoter of the naturally occurring gene. "Unmodified selectable marker" also includes where the promoter is the promoter of the naturally occurring gene. Further details of the modification of the selectable marker gene are described below but, in essence, the presence of the Group I intron prevents the expression of the selectable marker but, upon excision of the Group I intron, the resulting nucleic acid (ie unmodified selectable marker gene) is able to express the selectable marker. Preferably, the selectable marker gene is located in domain IV of the Group II intron.

It will be appreciated that the Group I intron may be positioned at any location within the selectable marker gene as long as expression of the selectable marker is prevented by the presence of the Group I intron. It will be appreciated that the Group I intron may be positioned, for example, within the promoter, such as between the -10 and -35 elements of the promoter, between the promoter and the coding region or in the coding region.

The selectable marker gene containing the Group I intron (ie the modified selectable marker gene) may be considered to be a retrotransposition activated marker (RAM).

Group I introns are self-splicing introns which may or may not require auxiliary factors such as proteins in order to be excised. Various Group I introns which may be useful in the practice of the invention are known including bacteriophage introns (Sandegran and Sjδberg (2004) J. Biol. Chem. 279: 2221S-22227), and Tetrahymena Group I intron (Roman (1998) Biochem. 95: 2134-2139). It is preferred that the Group I introns do not require auxiliary factors in order to be excised. It is preferred if the Group I intron is the td Group I intron from Phage T4 (EhrenMan et al (1986) Proc. Natl Acad. Sd. USA 83: 5875-5879).

It will be appreciated that the orientation of the various components within the DNA molecule is very important. Thus, from Figure 2 it will be seen that the modified selectable marker gene is present within the Group II intron in the reverse orientation to the Group II intron. Also, the Group I intron which is present within the modified selectable marker gene in a reverse orientation to the selectable marker gene but in the same forward orientation as the Group II intron. If the Group I intron were in the same orientation as the selectable marker gene, the intron would be able to excise from the mRNA transcript of the selectable marker gene and the phenotype conferred by the selectable marker would be present irrespective of whether the Group II intron containing the selectable marker had retargeted to the chromosome. Therefore, the Group I intron and the selectable marker gene must be in opposite orientations.

If the selectable marker gene were in the same orientation as the Group II intron, following the above logic, the Group I intron would have to be in the opposite orientation to the Group II intron. However, in this orientation, it would not excise from the mRNA trancript and so, even if the Group II intron did retarget to the chromosome, there would be no selectable phenotype.

Only when the various components are orientated as shown in Figure 2 will retargeting of the Group II intron to the chromosome be necessary and sufficient for expression of the selectable marker phenotype.

When the DNA molecule of the invention is used to introduce a nucleic acid molecule into a site of a DNA molecule in a bacterial cell of the class Clostridia (as is described in more detail below), the Group I intron is removed from the RNA transcript produced from the modified Group II intron to leave a region encoding the selectable marker, and the RNA transcript (or a DNA copy thereof) is introduced into a site in a DNA molecule in a bacterial cell of the class Clostridia. In this way, the nucleic acid introduced into a DNA molecule in a bacterial cell of the class Clostridia has a selectable marker gene which is able to express the selectable marker in the bacterial cell.

hi a preferred embodiment, the modified Group II intron is flanked by exons, which exons allow splicing of an RNA transcript of the Group II intron.

The promoter of the selectable marker gene is capable of causing expression of the selectable marker when it is encoded by a single copy of the selectable marker gene in an amount sufficient for the selectable markers to alter the phenotype of a bacterial cell of the class Clostridia such that it can be distinguished from the bacterial cell of the class Clostridia lacking the selectable marker gene. For example, the promoter may be one which, when present in a single copy in the bacterial chromosome, and when in operable linkage with the coding region of the selectable marker, expresses the selectable marker in a detectable amount. The promoter of the selectable marker gene is one which is functional in a bacterial cell of the class Clostridia and causes adequate, expression when present in a single copy as described above. It is preferred that the promoter is functional in a Clostridium sp. Suitable promoters include the fdx gene promoter of C. perfiingens (Takamizawa et al (2004) Protein Expression Purification 36: 70-75); the ptb, thl and the αdc promoters of C. αcetobutylicum (Tummala et αl (1999) App. Environ. Microbiol. 65: 3793-3799) and the cpe promoter of C. perfringens (Melville, Labbe and Sonenshein (1994) Infection and Immunity 62: 5550-5558) and the thiolase promoter from C. acetobutylicum (Winzer et al (2000) J MoI. Microbiol. Biotechnol. 2: 531-541). Preferably, the promoter of the selectable marker gene is the promoter of the thl gene of C. acetobutylicum.

To test whether a promoter is likely to be effective as a promoter of a selectable marker of the invention, a spliced variant of the RAM (ie encoding the selectable marker since the Group I intron has been removed) may be placed under its transcriptional control and introduced into the Clostridia to be targeted at a low copy number, preferably equivalent to the copy number of the chromosome. This can be achieved by using a low copy number plasmid, such as the low copy number derivatives of plasmid pAMBl described in Swinfield et al (1990) Gene.

87:79-90 or more ideally using a conjugative transposon and the method described in Mullany et al {Plasmid (1994) 31: 320-323) and Roberts et al {J Microbiol

Methods (2003) 55: 617-624). To achieve the latter, the spliced RAM together with the promoter under evaluation may be cloned into a vector that is unable to replicate in a Gram-positive bacterium but which carries an antibiotic resistance gene (eg catP) and a segment of DNA derived from a conjugative transposon, such as TnPl 6. The plasmid is then transformed into & Bacillus subtilis cell that carries the appropriate conjugative transposon in its genome (TnPi 6), and transformants selected on plates containing chloramphenicol. As the plasmid cannot replicate, the only way that chloramphenicol resistant colonies can arise is if the plasmid integrates into the genome as a consequence of homologous recombination between TnPi 6 and the region of homology carried by the plasmid.

This results in a transposon: :plasmid cointegrate carrying the spliced RAM and promoter under test that is located in a single copy in the genome. The Bacillus subtilis transconjugant obtained may now be used as a donor in a conjugation with the Clostridia to be targeted. In these matings, transfer of the transposon: :plasmid cointegrate into the Clostridia recipient can be selected on the basis of acquisition of resistance to thiamphenicol. Once obtained, transconjugants may be tested for the resistance encoded by the RAM, eg., erythromycin.

The promoter for regulating the transcription of the modified Group II intron may be any suitable promoter which is functional in a bacterial cell of the class

Clostridia. The promoter may be a constitutive promoter or an inducible promoter. An inducible promoter may be derepressed such that it diives

expression in a constitutive fashion. In particular experiments described in the Examples, the inventors found that regulated expression of the modified Group II intron confers no advantage in allowing for a high intron insertion frequency compared to constitutive expression. However, in other situations, it may be useful to be able to regulate expression of the modified Group II intron. A person of ordinary skill can perform experiments to determine whether a particular promoter is suitable to allow for a satisfactory intron insertion rate.

Girbal et al (2003) Appl Environ. Microbiol. 69: 4985-4988 describe a preferred xylose-inducible promoter in C. acetobutylicum, which is based on the

Staphylococcus xylosus xylose operon promoter-repressor regulatory system.

Suitable inducible promoters are IPTG or xylose-inducible. Conveniently, for example when the DNA molecule is for use in Clostridial cells, the promoter is the promoter region of the C. pasteurianum ferredoxin gene under the control of the lac operator region of the E. coli lac operon. Conveniently, the DNA molecule further comprises the lad gene of E. coli.

A promoter for regulating the transcription of the modified Group II intron may be a constitutive promoter. The skilled person will appreciate that in general all promoters are regulated under one condition or another, even if such conditions are not known. Therefore, we intend "constitutive promoter" to be interpreted broadly to encompass a promoter that is active in the Clostridial cells under the normal culture conditions employed in the retargeting protocol, without the need for addition of an agent to activate expression driven by the promoter. Promoters of genes that are essential to primary metabolism may be suitable "constitutive promoters". For example, the thiolase promoter, thl, described in the Examples may be a suitable promoter. Other suitable promoters are the C. acetobutylicum promoters hbd, crt, etfA, etfB amd bed (Alsaker and Papoutsalds (2005) J Bacteriol 187:7103-7118). Promoters suggested as being suitable for driving expression of the modified selectable marker in the RAM may also be suitable.

The use of an inducible promoter allows transcription of the Group II intron containing the selectable marker gene interrupted by the Group I intron (which may be termed a RAM) to be switched off following retargeting of the RAM to the bacterial chromosome. When the RAM is transcribed from the inducible promoter, expression of the selectable marker is ineffective. This may be because of duplex formation between the transcripts of the coding strand transcribed from the chromosome and the non-coding strand transcribed from the DNA molecules.

The DNA molecule of the invention preferably is capable of replication in a bacterial cell of the class Clostridia. More preferably, it is capable of conditional replication. Conveniently, the DNA molecule contains a suitable origin of replication and any necessary replication genes to allow for replication in the Gram-positive bacterial cell (ie suitable rep genes). Preferably, the DNA is a plasmid. Alternatively, the DNA may be linear or it may be filamentous phage like Ml 3. Conveniently, the DNA molecule is a shuttle vector which allows for replication and propagation in a Gram-negative bacterial cell such as Escherichia coli and for replication in a Gram-positive cell, particularly a cell of the class Clostridia and more particularly of the genus Clostridium. Additionally or alternatively, the DNA molecule of the invention contains a region which permits conjugative transfer from one bacterial cell to a bacterial cell of the class Clostridia. It is particularly preferred if the DNA molecule contains a region which permits conjugative transfer between E. coli and a bacterium of the class Clostridia, and more particularly of the genus Clostridium. For example, the DNA molecule may contain the oriT (origin of transfer) region, including the traJ gene.

Methods of transformation and conjugation in Clostridia are provided in Davis, I, Carter, G, Young, M and Minton, NP (2005) "Gene Cloning in Clostridia ^' ", In: Handbook on Clostridia (Durre P, ed) pages 37-52, CRC Press, Boca Raton, USA.

The selectable marker may be any suitable selectable marker which can be expressed in and used to select a cell of the class Clostridia containing the

selectable marker. Suitable selectable markers include enzymes that detoxify a toxin, such as prodrug-converting enzymes. Selectable markers also include a prototrophic gene (for use in a corresponding auxotrophic mutant). Preferably, the selectable marker is one which gives a growth advantage to the bacterial cell of the class Clostridia in which it is expressed. Thus, typically, under a given growth condition the bacterial cell which expresses the selectable marker is able to grow (or grow more quickly) compared to an equivalent cell that does not express the selectable marker.

Convenient selectable markers include antibiotic resistance factors. Thus, suitably, the selectable marker gene is a gene which confers antibiotic resistance on a bacterial cell of the class Clostridia.

Not all antibiotic resistance genes can be used in all cells of the class Clostridia. For example, Clostridium sp. are naturally resistant to kanamycrα, and are frequently resistant to trimethoprim. Thus, it is preferred that the selectable marker gene is not a kanamycin resistance gene or a trimethoprim resistance gene particularly when the bacterial cell is of the genus Clostridium. Suitable antibiotic resistance genes for use in Clostridial cells, such as Clostridium sp., include erythromycin resistance genes (such as Erm) and chloramphenicol resistance genes (such as catP). Another suitable antibiotic resistance gene is tetM, for example tetM from the Enterococcus faecalis Tn916 conjugative transposon (Roberts et al (2001) Microbiol. 147: 1243-1251). Another suitable antibiotic resistance gene, widely used in bacteria of the class Clostridia, is spectinomycin adenyltransferase, aad (Charpentier et al (2004) Appl. Environ. Microbiol. 70, 6076-6085).

The methods and DNA molecules of the invention may also be used to investigate genes the function of which is not known. For example, the DNA molecule of the invention may be adapted to contain a unique oligonucleotide sequence referred to as a tag which will be introduced into the DNA in the cell of the class Clostridia. Conveniently, a plurality of DNA molecules of the invention are produced, each

containing a different tag sequence. When the DNA inserts into the bacterial chromosome;, the tag is present in the genomic DNA and may be detected for example by amplification by hybridising to a labelled oligonucleotide probe, a portion of which has a sequence complementary to a portion of the tag. Suitable tags, probes and methods of amplifying and hybridising are described in Hensel et al (1995) Science 269: 400-403. A plurality of mutants may be generated by the method of the invention in which each has the DNA inserted into a different gene, and each may be identified by its unique tag. Typically, each different retargeting nucleic acid contains targeting portions which direct it to a different gene in the DNA of the cell of the class Clostridia. The plurality of mutants may be introduced into an environment for a period of time. Mutants may then be recovered from the environment. The ability of individual tags to be detected in the recovered pool of mutants gives an indication of whether a particular mutant has been able to grow or survive as well as other mutants. In this way, genes that are required for growth or survival in the environment may be identified. Hensel et al (1995; supra) used a similar approach to identify virulence genes in Salmonella.

In a modification of the above method, DNA molecules of the invention having the same tag but different randomised Group II intron targeting portions and corresponding exon sequences may be generated, pooled and used to make bacterial mutants. Group II introns with randomised targeting portions are described in WO 01/29059. Many of the DNA molecules may be unable to insert anywhere in the bacterial genome. However, some may be able to insert at an unknown location in the bacterial genome governed by the sequence of the targeting portions. A sufficiently large pool of DNA molecules of the invention may be used in the method such that one or more colonies are obtained in which the DNA has inserted into the chromosome. A single clone may be selected. The process may be repeated for a pool of DNA molecules of the invention having a different unique tag, to obtain another single mutant bacterial clone with a unique tag. In this way, a plurality of bacterial mutants, each with a unique tag are generated. The plurality of mutants may be exposed to an environment as

described above, to identify particular mutants that are compromised for growth or survival in that environment. A mutant identified from such a screen may then be characterised to determine in which gene the DNA has inserted.

Further details of genes encoding modified selectable markers which contain a Group I intron which disrupts the expression of the selectable marker are given below.

The selectable marker gene or its coding region may be associated with regions of DNA for example flanked by regions of DNA that allow for the excision of the selectable marker gene or its coding region following its incorporation into the chromosome. Thus, a clone of a mutant Clostridial cell expressing the selectable marker is selected and manipulated to allow for removal of the selectable marker gene. Recombinases may be used to excise the region of DNA. Typically, recombinases recognise particular DNA sequences flanking the region that is excised. Cre recombinase or FLP recombinase are preferred recombinases.

Alternatively, an extremely rare-cutting restriction enzyme could be used,- to cut the DNA molecule at restriction sites introduced flanking the selectable marker gene or its coding region. A preferred restriction enzyme is I-Scel.

A mutant bacterial cell from which the selectable marker gene has been excised retains the Group II intron insertion. Accordingly, it has the same phenotype due to the insertion with or without the selectable marker gene. Such a mutant bacterial cell can be subjected to a further mutation by the method of the invention, as it lacks the selectable marker gene present in the RAM.

Although the modified Group II intron in the DNA molecule of the invention does not express the IERT ₅ conveniently the DNA molecule contains in another location a gene which is able to express the IERT.

Where the Clostridial cell into which the Group II intron is to be inserted uses a different genetic code from the Group II intron and its associated Group II intron-

encoded reverse transcriptase, it is preferred that the sequence of the Group II intron-encoded reverse transcriptase is modified to comprises codons that correspond to the genetic code of the host cell.

A particularly desirable embodiment of the invention is wherein the modified Group II intron comprises targeting portions. Typically, the targeting portions allow for the insertion of the RNA transcript of the modified Group II intron into a site within a DNA molecule in the Clostridial cell. Typically, the site is a selected site, and the targeting portions of the modified Group II intron are chosen to target the selected site. In a preferred embodiment, the selected site is in the chromosomal DNA of the Clostridial cell. Typically, the selected site is within a particular gene, or within a portion of DNA which affects the expression of a particular gene. Insertion of the modified Group II intron at such a site typically disrupts the expression of the gene and leads to a change in phenotype.

Genes may be selected for mutation for the purposes of metabolic engineering. For example, in organisms such as Thermoanaerobacterium. saccharolyticum, or other members of the class Clostridia which have a similar metabolism, deletion of lactate dehyrogenase and phosphotransacetylase to prevent formation of lactate and acetate, respectively, could be used to elevate levels of ethanol (Desai et al, (2004) Appl Microbiol Biotechnol. 65: 600-5). In solventogenic Clostridia, such as Clostridium acetobutylicum and Clostridium specific deletions may be made to the genes encoding the enzymes responsible for solvent and acid production as a means of maximising acetone and butanol (see Jones and Woods (1986) Microbiol Rev. 50: 484-524). Thus, strains could be generated that produce only acetone or butanol, by elimination of enzymes responsible for production of acetate (phosphotransacetylase and or acetate kinase), butyrate (phosphotransbutyrylase and or butyrate kinase), butanol (butanol dehydrogenase A and/or butanol dehydrogenase B) and/or acetone (acetyoacetate decarboxylase and/or acetoacetyl-CoA transferase). Moreover, the fermentative ability of such strains could be extended by gene addition into the chromosome, such that new

substrates could be degraded (sugars, lignocellulose, hemicellulose, etc.) and/or new end products made (isopropanol, 1,3-propanediol, etc.).

Genes may be selected for mutation in order to determine the role of their encoded products in virulence, a prerequisite to the development of vaccines and other countermeasures. In C. difficile, for example, the relative roles of toxin A and toxin B (CdtA and CdtB) remain to be established (Bongaerts and Lyerly (1994) Microbial Pathogenesis 17: 1-12) due to a previous inability to generate isogenic mutants. Certain strains (Perelle et al (1994) Infect Immun. 65: 1402-1407) also produce an actin-specific ADP-ribosyltransferase CDT (CdtA and CdtB). Other factors undoubtedly contribute to virulence, particularly the initial colonisation process. The participation of a number of gene products has been proposed (Tasteyre et at (2001) Infect Immun 69: 7937-7940; Calabi et al (2002) Infect Immun 70: 5770-5778; Waligora et al (2001) Infect Immun 69: 2144-2153), including those involved in adhesion, the S-layer proteins (SpIA) and motility (FIiC and FIiD). Definitive proof of the involvement of these factors in disease through the generation of mutants has until now not been possible.

The DNA sequences of the genomes of many bacteria of the class Clostridia are known. For example, the DNA sequences of the genomes of C. acetobutylicum

(ATCC 824 (GenBank Accession No AE001437), C. difficile (GenBank

Accession No AMI 80355), C. tetani E88 (GenBank Accession No AE015927) and C. perfringens strain 13 (GenBank Accession No BAOOOOl 6) and C, botulinum are known. The sequence of a C. sporogenes genome is partially known and is very similar to the sequence of the C. botulinum genome. From this information, sites for insertion are readily identified, for example within open reading frames. It is preferred if the DNA molecule of the invention contains a modified Group II intron which contains targeting portions which targets the 3RNA transcript of the modified Group II intron (or a DNA copy thereof) into a gene in the genome of one of these bacterial species.

As described above, Group II introns naturally contain regions which target the intron to a specified sequence in target DNA. Because the recognition site of the DNA substrate is recognized, in part, through base pairing with the excised Group II intron RNA of the RNP complex, it is possible to control the site of nucleic acid insertion within the DNA substrate. This may be done by modifying the EBSl sequence, the EB S2 sequence or the δ. sequence, or combinations thereof. Such modified Group II introns produce RNP complexes that can cleave DNA substrates and insert nucleic acid molecules at new recognition sites in the genome. For example, by reference to the Ll.LtrB Group II intron of Lactococcus lactis illustrated in Figures IA and IB the EBSl, EBS2 and δ are modified to permit base pairing of the RNA transcript of the modified Group II intron with a target site. Rules for DNA target-site recognition by Ll.LtrB Group II intron which enable retargeting of the intron to specific DNA sequences are described in Mohr et al (2000) Genes & Development 14, 559-573, incorporated herein by reference. Computer-aided design of targeting portions are also described in Perutka et al (2004) J. MoI Biol. 336, 421-429, incorporated herein by reference.

WO 01/29059 to the Ohio State University Research Foundation, incorporated herein by reference, describes a selection-based approach in which the desired DNA target site is cloned into a recipient vector upstream of a promoterless tet ^R gene. Introns that insert into that site are selected from a combinatorial donor library having randomized targeting portions (EBS and δ) and IBS exon sequences. The modified Ll .LtrB intron contains a heterologous promoter, such that when it inserts into the target site in the recipient vector, the tet ^R gene is transcribed and the bacterial cell containing the vectors may be selected for. The sequence of the modified intron may be determined by PCR. Thus, a modified Group II intron DNA may be isolated that allows for insertion into the target DNA site within a Clostridial cell.

In the case of the Ll .LtrB Group II intron, it is thought that the interaction of the δ region with a δ' region of the target DNA is not critical to efficient retrohoming of

the Group II intron. However, the interactions between EB S2 and EBSl in the intron RNA and IBS2 and IBSl in the target DNA are more important.

When the Group II intron excises from the RNA transcript, it is believed that it transiently base-pairs with portions of the flanking exon RNA. In particular, the

EBS2 and EBSl regions base-pair with the IBS2 and IBSl regions of the 5' exon respectively. Therefore, it is preferred that the IBS2 and IBSl region of the 5' exon is modified so as to promote base-pairing with the modified EBS2 and EBSl regions of the intron RNA. This facilitates efficient excision of the Group II intron from its RNA trancript.

Modification of the EB S2 and EBS lδ sites and the IBS2 IBSl site may conveniently be performed using any suitable site directed mutagenesis methods known in the art, for example oligonucleotide-directed mutagenesis or PCR-based methods.

Typically, the DNA molecule of the invention is able to express an antibiotic resistance marker which is different to the selectable marker. For example, if the selectable marker gene is a first antibiotic resistance gene the DNA includes a second antibiotic resistance gene. It is particularly preferred if both antibiotic resistance genes are ones which give rise to antibiotic resistance in. Clostridial cells. For example, the selectable marker gene in the DNA molecule may be an erythromycin resistance gene and the DNA molecule may further contain a chloramphenicol resistance gene (or vice versa). When the DNA molecule is for use in a Clostridium sp. it is particularly preferred that any antibiotic resistance genes are selected from erythromycin resistance genes (eg ermB) or chloramphenicol resistance genes (eg catP).

It will be appreciated that although it is convenient for the DNA molecule of the invention to itself contain a gene which is able to express the IERT, this may be provided on a separate DNA molecule. Thus, a further aspect of the invention

provides a kit of parts comprising a DNA molecule of the first aspect of the invention and a separate DNA molecule which is able to express the IERT.

^• Typically, the DNA molecules are plasmids, preferably compatible plasmids. It will be appreciated that the kit may further contain a DNA molecule (typically a plasmid) which is able to express the lac repressor protein. This is useful in the situation where the DNA molecule of the invention comprises an IPTG-inducible promoter which is operatively linked to the Group II intron, but when the DNA molecule of the invention does not include the lad gene.

A third aspect of the invention provides a method of introducing a nucleic acid molecule into a site of a DNA molecule in a bacterial cell of the class Clostridia, the method comprising the steps of:

(i) providing a bacterial cell of the class Clostridia with the DNA molecule of the invention and a DNA molecule capable of expressing a Group II intron-encoded reverse transcriptase; and (it) culturing the bacterial cell under conditions which allow for removal of the Group I intron from the RNA transcript of the modified Group II intron and the insertion of said RNA transcript containing the selectable marker gene (or a DNA copy thereof) into said site.

Preferably, the bacterial cell of the class Clostridia is cultured under conditions which allow for expression of the selectable marker. Typically, the bacterial cell of the order Clostridia into which nucleic acid has been introduced at a site of a DNA molecule within the cell (ie mutated cell) is selected based on an altered phenotype conferred by the selectable marker.

Conveniently, the selectable marker is an antibiotic resistance marker and the mutated Clostridial cell is selected on the basis of its ability to grow in the presence of the relevant antibiotic.

Conveniently, the selected cell is cloned and a single clone of cells is obtained. ^•

A further aspect of the invention provides a method of targeting a nucleic acid molecule to a selected site of a DNA molecule in a bacterial cell of the class Clostridia, the method comprising providing a bacterial cell of the class Clostridia with a DNA molecule of the invention in which the modified Group II intron comprises targeting portions and a DNA molecule capable of expressing a Group

II intron-encoded reverse transcriptase; and culturing the bacterial cell under conditions which allow removal of the Group I intron from the RNA transcript of the modified Group II intron and the insertion of said RNA transcript (or DNA copy thereof) containing the selectable marker gene into said selected site.

It will be appreciated that in this way it is possible to make site directed mutations in DNA (such as the genome) of a bacterial cell of the class Clostridia, such as a Clostridium spp.

Mutant bacterial cells of the class Clostridia obtained by the methods of the invention are also part of the invention.

It will be appreciated that with respect to all aspects of the invention it is preferred that the bacterial cell of the class Clostridia is a Clostridium spp. It is particularly preferred if the Clostridial cell is C. thermocellum or C. acetobutylicum or C. difficile or C. botulinum or C. perfringens or C. sporogenes or C. beijerincMi or C. tetani or C. cellulyticum or C. septicum. The Clostridial cell may alternatively by Thermoanaerobacteria saccharolyticum, an important species for industrial ethanol production. By the term "Clostridia", we also include Roseburia, such as Roseburia intestinalis, which is a probiotic bacterium. Thus, preferably, the selectable marker gene in the DNA molecule of the invention is a gene which can be used for selection in these species (eg an erythromycin resistance gene or a chloramphenicol resistance gene or a tetracycline resistance gene or a spectinomycin resistance gene). Also preferably, the DNA molecules of the

invention contain origins of replication and any necessary replication genes which allow for replication in these bacterial species.

A particular feature of the invention is that the modified selectable marker gene is one which contains a Group I intron which disrupts expression of the selectable marker. The selectable marker is one which may be expressed in and used for selection in a bacterial cell of the class Clostridia, particularly a Clostridium cell.

It is particularly preferred that the selectable marker is an antibiotic resistance gene which can be used for selection in a Clostridium spp.

A further aspect of the invention provides a DNA molecule comprising a modified erythromycin-resistance gene which contains a Group I intron.

A further aspect of the invention provides a DNA molecule comprising a modified chloramphenicol-resistance gene which contains a Group I intron.

A further aspect of the invention provides a DNA molecule comprising a modified tetracycline-resistance gene which contains a Group I intron.

A further aspect of the invention provides a DNA molecule comprising a modified spectinomycin resistance gene which contains a Group I intron.

The invention also includes these DNA molecules present in a host cell, for example an E. coli cell or a cell of the class Clostridia.

Preferably the Group I intron is present in the opposite orientation to the antibiotic resistance gene.

The Group I intron may be present anywhere within the antibiotic resistance gene, for example within the coding region thereby disrupting translation, or upstream of the coding region thereby disrupting transcription or translation.

The Group I intron is present within the antibiotic resistance gene in a form whereby when the intron is transcribed it is able to excise (splice) itself from the RNA transcript.

Any autocatalytic RNA which can self-splice out of a larger RNA in an orientation-dependent manner could substitute for a Group I intron in the present invention. Suitably, an "IStron" may be used, which is believed to be a fusion of a Group I intron and an IS element (Haselmayer et at (2004) Anaerobe 10: 85-92; Braun et al (2000) MoI. Microbiol. 36: 1447-1459).

For the avoidance of doubt, for the purposes of all aspects of the invention any autocatalytic RNA which can self-splice out of a larger RNA in an orientation- dependent manner is considered to he a Group I intron, whether or not it requires auxiliary factors. Preferably the Group I intron does not require auxiliary factors.

It is preferred that the Group I intron does not encode an intron-encoded protein such as an intron-encoded reverse transcriptase. This feature prevents the excised Group I intron RNA from re-inserting at another site within the bacterial genome.

It is noted that, typically, the splicing of Group I introns (such as the td intron of Phage T4) is reliant on exon sequences flanking the point of insertion. Thus, the modified selectable marker genes of the invention (and in particular the modified antibiotic resistance genes which encode erythromycin resistance and chloramphenicol resistance and tetracycline resistance and spectinomycin resistance of this aspect of the invention) contain the Group I intron inserted in a position whereby it is flanked by suitable exon sequences that allow the Group I intron to splice out of the RNA transcript and wherein the resulting spliced transcript (or DNA copy thereof) encodes a functional selectable marker (such as functional erythromycin resistance or functional chloramphenicol resistance). Suitable flanking sequences are known for Group I introns. For example, for the Phage T4 td Group I intron, the intron is typically preceded by a G residue (ie

present 5' of the intron) and the intron is typically followed by the sequence 5'-ACCCAAGAGA-3' (SEQ ID No. 3)(ie present 3' of the intron). Alternatively, the intron maybe followed by the sequence 5'-ACCCAAGAA-3' (SEQ ID No. 4).

In a preferred embodiment of the invention, the coding region of the selectable marker (such ^• as the erythromycin or chloramphenicol or tetracycline or spectinomycin resistance genes) contains suitable sequences which flank the intron. In relation to the td intron, and the combined 5' and 3' flanking sequence 5'-GACCCAAGAGA-S' (SEQ ID NO. 5) this is able to code for several amino acid sequences depending on the reading frame (as explained in more detail in the examples).

In Frame 1, it encodes the amino acid sequence DPRD/E (SEQ ID No. 6); in Frame 2 it encodes the amino acid sequence R7GPKR (SEQ ID No. 7) and in Frame 3 it encodes the amino acid sequence "X"TQE"Z" (SEQ ID No. 8) where X can be any of G, E, A, V, L, S, W, P, Q, R, M, T or K and "Z" can be any of K, S ₅ R ₅ I ₅ M ₅ T OrN.

Thus, in a preferred embodiment, the coding region of the selectable marker gene encodes a portion of peptide with the above amino acid sequence.

In a further preferred embodiment, the exon sequence 3' of the intron is present in an appropriate reading frame at the 5' end of the coding sequence of the selectable marker so that, in the absence of the intron, the coding sequence encodes a functional selectable marker which contains a linker peptide at the N-terminus of the selectable marker polypeptide.

The linker peptide is typically a peptide of 4 to 2O ₅ preferably 4 to 15, typically 4,

5, 6, 7, 8, 9, 10, 11, 12, 13, 14 or 15 amino acid residues, a portion of which are encodable by the exon coding sequences flanking the intron. The presence of the linker peptide does not interfere substantially with antibiotic resistance activity. In

other words, the polypeptide produced from expression of the nucleic acid molecule produced when the Group I intron has been excised has antibiotic resistance activity.

Alternatively, the Group I intron flanking sequence may be disposed so that the insertion of the Group I intron disrupts transcription of the selectable marker gene. For example, it may be located between the -35 and -10 elements of the promoter.

In a further alternative, the Group I intron flanking sequence may be disposed so that the insertion of the Group I intron disrupts translation of the selectable marker gene. For example, it may be located between the ribosome binding site and the start codon.

It will be appreciated that the DNA molecules of the invention may be made using standard molecular biological techniques as described in Sambrook et at, "Molecular cloning: A laboratory manual", 2001, 3 ^rd edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

The invention will now be described with reference to the following non-limiting Examples and Figures.

Figure 1. A: Secondary structure model of Ll.LtrB group II intron. The predicted secondary structure consists of six domains (I- VI). The EBS2/IBS2, EBS1/IBS1 and δ-δ' interactions between the intron and flanking exons in unspliced precursor RNA are indicated by broken lines. In the un-modified Ll .LtrB intron, the open reading frame encoding the LtrA protein is present in the non-structural loop indicated as domain IV. B: Mechanism of DNA target site recognition by Ll.LtrB group II intron. The LtrA protein binds to the Ll .LtrB group II intron RNA forming a ribonucleoprotein complex. The intron splices out of the pre-mRNA, liberating the ribonucleoprotein as a particle. The ribonucleoprotein particle locates target DNA sequences within the cell. The target DNA sequence of the unmodified ribonucleoprotein is an intronless copy of

the UrB gene, the sequence of which is depicted (SEQ ID No. 9). The intron RNA is inserted into the insertion site within the top strand (IS). The bottom strand is then cleaved at the cleavage site (CS) and the LtrA primes from the cut DNA and reverse-transcribes the intron RNA. Host repair activities complete the integration process. Recognition of the target is mediated by a combination of interactions between LtrA and nucleotides in the target sequence, and between EBS2 and EBSl in the intron RNA and complementary sequences IBS2 and IBSl in the target sequence. The most important of the nucleotides recognised by the LtrA protein are indicated by grey shading.

Figure 2. Positive Selection of retargeting nucleic acid-derived mutants. A. Transcription of the selectable marker gene from the plasmid-located retargeting nucleic acid does not result in resistance, because the mRNA produced retains the td group I intron insertion and the expression of the selectable marker gene is therefore disrupted. The td element cannot splice out of the mRNA because it has been transcribed in the wrong orientation. B. Ll. LtrA group II intron RNA production is induced by addition of IPTG, causing transcription from Clostridial promoter fac. The td group I intron within the selectable marker gene is transcribed in the correct orientation and the td RNA splices out of the RNA produced. C. The Ll.LtrA RNA and the selectable marker gene are inserted into the target site in the chromosome. The selectable marker gene does not contain the td group I intron and therefore the expression of the selectable marker gene is not disrupted. The cells therefore exhibit the phenotype associated with expression of the selectable marker, and may be selected accordingly.

Figure 3. Inducible expression from pMTL5401Fcat in C. sporogenes and C. acetobutylicum.

(a) The E. coli I Clostridium shuttle plasmid ρMTL5401Fcat. (b) A clone of C. sporogenes or (c) C. acetobutylicum containing pMTL5401Fcat was grown to early exponential growth phase and the CAT activity in cell lysates monitored after induction with 1 mM IPTG (B) or without induction (A).

Figure 4. Sequences suitable for a selectable marker gene for successful splicing of the td group I intron. The required amino acid sequences (SEQ ID Nos. 6-8, in any of the three translation reading frames, are shown above the nucleotide sequences (SEQ ID Nos. 132-134). Amino acids at position ^£ X' could be either G, E, A, V, L, S ₅ W, P, Q, R, M, T or K. At position 'Z' they could be . K ₅ S, R ₃ 1, M, T or N.

Figure 5. RAM functionality added to the ermB gene using a linker.

(a) A linker containing the td intron and its exons was inserted between the ermB ORF and its promoter (SEQ ID No. 10), preventing expression of erythromycin- resistance. Splicing of the td intron out of the reverse strand yields a modified ermB gene (SEQ ID No. 11) that encodes a functional protein with 12 additional amino acids at its N-terminus (SEQ ID No. 12). The ermB promoter of ErmBtdRAMl is replaced by the thl promoter in ErmBtdRAM2. (b) PCR using various templates and primers ErmB-Pro-F3 and ErmB-Rl, which flank the td intron in ErmBtdRAMl. Lane 1 : ErmBtdRAMl DNA; Lane 2: ErmBtdRAMl SE DNA; Lane 3: cDNA synthesised from RNA isolated from cells containing pMTL201acZTTErmBtdRAMl after IPTG induction; Lane 4: the same RNA preparation before cDNA synthesis. (c) PCR using various templates and primers Thio-Fl and ErmB-Rl, which flank the td intron in ErmBtdRAM2. Lane 1 : C. sporogenes spoOλ mutant genomic DNA; Lane 2: pMTL007::Csp-_?po0y4-249s plasmid DNA; Lane 3: C. sporogenes wild-type genomic DNA; Lane 4: water.

Figure 6. Features and sequence of ErinBftfftAMl

ErmBtdRAMl sequence (SEQ ID No. 13)

Figure 7. Direct evidence that the ErmBft? RAMI is spliced in E. coll

To test that the td group I intron has been spliced from ErmBtdRAMl following induction of the group II intron RNA expression, RNA was prepared from cells expressing pMTL201acZTTErrnB&/RAMl. RT-PCR was performed using primers that flank the td site of insertion. In control reactions, the same primers

were used to amplify ErmBfciRAMl and Spliced Equivalent SE DNA by PCR. Lane 1: DNA markers; lane 2, PCR of εxmBtd RAMI; lane 3, PCR of EτmBtd RAMI SE; lane 4, RT-PCR on total RNA from cells containing pMTL201acZTTErrnB£dRAMl, and; lane 5 RT-PCR negative control.

Figure 8. Construction of a multicloning site in pBRR3

Sequences of the cloning sites of pBRR3-LtrB (SEQ ID No. 14) and ρCR2.1- TOPO plasmids (SEQ ID No. 15). The multicloning site fragment depicted (SEQ ID No. 16) was inserted into a cleaved pBRR3-LtrB to make pBRR3-MCSl depicted (SEQ ID No. 17), containing restriction sites found in the ρCR2.1-TOPO plasmid.

Figure 9. Sequences of the thl and thl2 promoters

Sequences of thl (SEQ ID No. 18) and thl2 promoters (SEQ ID No. 19) are shown in comparison to a consensus promoter (SEQ ID No. 20). "x" indicates a nucleotide substitution compared to the consensus sequence. The spacing between the -10 and the -35 elements is indicated for each sequence.

Figure 10. Features and sequence ofErmBtdRAMl ErmBtdRAM2 sequence (SEQ ID No. 21 )

Figure 11. Features and sequence of pMTL007

Plasmid map of the final clostridial retargeting system (the illustrated example is a derivative modified to re-target lacZ) and sequence (SEQ ID No. 22)

Figure 12. Construction of pMTL5401F

Restriction sites used at each step are indicated. DNA end-blunting was performed using T4 DNA polymerase.

Figure 13. Construction of pMTL5402F and pMTL5402F- lacZTTErniBtdRAMl

Restriction sites used at each step are indicated. DNA end-blunting was performed using T4 DNA polymerase.

Figure 14. Construction of pMTL007 Restriction sites used at each step are indicated.

Figure 15. Examples of mutant screening and characterisation.

(a) Plasmid pMTL007. (b, c) PCR was used to initially screen for the presence of the intron insertion in the C. difficile spoOA gene using the intron-specific primer EBS Universal and gene-specific primer Cd-spoOA-R2 (small arrows). Lane 1: water; Lane 2: C. difficile parental strain genomic DNA; Lane 3: pMTL007::Cdi- spoOA-llSa. plasmid DNA; Lanes 4-6: DNA from three randomly-selected Em ^R C. difficile clones generated using pMTL007::Cdi-iyc>ft4-178a. (d) Southern blots of the spoOA and pyrF mutants of C. difficile using a probe to ermB. Hybridisation of this probe to the pre-existing (non-functional) chromosomal ermB ORF causes a second band, also visible hi the parental lanes. In the EcoKV digest of the spoOA mutant, both bands are a similar size, (e) Equivalent Southern blot for C. acetobutylicum and (f) C. sporogenes.

Figure 16. The spoOA mutants do not form spores.

Phase-contrast micrographs of the spoOA mutants and parental strains of C. difficile, C. acetobutylicum and C. sporogenes grown on solid media for 14 days, 4 days or 3 days respectively. Mean sporulation frequencies of three separate experiments are shown as percentages.

Example 1: Development of an IPTG-inducible 'fac' promoter

The use of a E. coli I Clostridium shuttle vector (pMTL540F) carrying the artificial promoter 'fac' has previously been described. It was derived by inserting the operator of the E. coli lacZ operon immediately downstream of the promoter of the C. pastueurianum ferredoxin gene (Fox et al (1996) Gene Ther. 3: 173- 178). Although this promoter element was used to direct the high level expression

of heterologous genes in Clostridia, regulated transcription has not been demonstrated. A new E. coli I Clostridium shuttle vector pMTL5401F was, therefore, constructed featuring the_/αc promoter, a lad repressor gene under the transcriptional control of the promoter of the C. acetobutylicum phosphotransbutyrylase (ptb) gene and the oriT region of plasmid RK2 to facilitate conjugative transfer to C. sporogenes, C. botulinum and C. difficile. To test pMTL5401F, we inserted a promoted ess copy of the pC194 cat gene, such that its transcription was under the control of the fac promoter in the resultant plasmid, pMTL5401Fcα£ (Figure 3). We then assayed for the enzyme activity of the cat gene product in the Iy sates of C. sporogenes or C. acetobutylicum cells carrying pMTL5401Fcαt grown in the presence or absence of exogenous IPTG. Induction was observed in both organisms, but while strong repression of transcription was evident in C. sporogenes in the absence of IPTG (Figure 3), a significant basal level of expression was observed in C. acetobutylicum (Figure 3). Although pMTL5401F could be introduced into C. difficile, the ρCB102 replicon functions relatively ineffectively in this clostridial host (Purdy et al (2002) MoI. Microbiol. 46: 439-452) and cannot support the growth of its transconjugants in antibiotic-supplemented liquid culture. Therefore, an equivalent induction experiment could not be performed.

Example 2: Development of ErmBztf as a selectable marker for Clostridia

Splicing of the td group I intron is reliant on exon sequences flanking the point of insertion. The target site recognised by the Phage T4 td group I intron is 5'-GACCCAAGAA-S ' (SEQ ID NO. 23) and the intron inserts after the intial 'G'. However the td group I intron will also insert at the site 5'-GACCCAAGAGA-3 ' (SEQ ID No. 5) (Sigma Aldrich TargeTron™ Gene Knockout System). Sequences of antibiotic genes currently in use in Clostridia were evaluated for the presence of these sequences but no genes incorporating either of these sequences were identified. If the splice site 5'-GACCCAAGAGA-S ' (SEQ ID NO. 5) were present in a protein coding region, the amino acid sequence it would encode would depend on its reading frame. Amino acid sequences (corresponding to the three

possible frames) that may be encoded by the splice site are shown in Figure 4. Screening of the protein sequences of all proteins known to confer resistance on Clostridia failed to identify a candidate protein containing any of the desired amino acid sequences.

Accordingly, a gene encoding a selectable marker was engineered such that it contained an insertion site for the td group I intron. This was to form the basis of a Clostridial RAM. The native ermB gene of the Enterococcus faecalis plasmid pAMβl, which confers resistance to erythromycin, was chosen as the selectable marker gene because this gene has been widefy used in the construction of E. coli / Clostridium shuttle vectors (Dϋrre, P. Handbook on Clostridia. 2005. Taylor and Francis, CRC Press.)

A linker sequence was designed which contained the required splice site, and fused to the 5 'end of the coding region of the ermB gene, in effect extending the

N-terminus of the protein by 12 amino acids (Figure 5). In the design of this sequence, the chosen reading frame was that which encoded amino acids that were as inert as possible, and soluble, to minimise risk of adversely affecting ErmB protein function. Frame 1 (DPRD; SEQ ID No. 6) was the best option as it includes three charged residues (Asp -ve, Arg +ve) which should favour solubility. It was hoped that the mixture of charges might help prevent a strong interaction with the rest of the protein. The rest of the linker was composed of the small, inert residues GIy and Ala, with a single Ser to avoid a long stretch of hydrophobic residues which might reduce the protein's solubility. In addition, the nucleotide sequence chosen incorporates clostridial codon usage, to minimise any potential expression problems.

Two constructs were assembled using SOEing PCR as described below, using oligonucleotide primers indicated in Table 1 below. The EτxήBtd RAMI (the modified ermB gene containing the td intron inserted at the indicated site in Fig. 6 (SEQ ID No. 13), and the spliced equivalent (SE), in which the td intron is absent. EmiBtdRAMl and EπnBføRAMl SE were each cloned into the high copy

plasmid pMTL5402F in the opposite orientation to the /αc promoter (so that any resistance conferred is due to the RAM or SE's own promoter).

Table 1 : Oligonucleotide primers

Primer Sequence (5' - 3 ') SEQ ID NO

ErmB-Pro-F3 CTACGCGTGGAAATAAGACTTAGAAGCAA 24 ACTTAAGAGTGTG

ErmB-Pro- CAGAAGCACCAGCATCTCTTGGGTCCATGT 25 RA AATCACTCCTTCTTAATTACAAATTTTTAGC

ATC linker 1- ACCCAAGAGATGCTGGTGCTTCTGGTGCTG 26

ErmB-Fl GTATGAACAAAAATATAAAATATTCTCAA

AACTTTTTAACGAGTG

ErmB-Rl GAACGCGTGCGACTCATAGAATTATTTCCT 27

CCCG

ErmB-Pro- GGGGTAAGATTAACGACCTTATCTGAACAT 28 RB AATGCCATGTAATCACTCCTTCTTAATTAC

AAATTTTTAGCATC tdGpI-Fl GCATTATGTTCAGATAAGGTCGTTAATCTT 29

ACCCC tdGpI-Rl CCAGAAGCACCAGCATCTCTTGGGTTAATT 30

GAGGCCTGAGTATAAG

Thio-Fl CTACTAGTACGCGTTATATTGATAAAAATA 31

ATAATAGTGGG

Thio-R-RAM CCTTATCTGAACATAATGCCATATGAATCC 32 CTCCTAATTTATACGTTTTCTC

The ErmBήiRAMl SE was made as follows. The erniB promoter was PCR- amplified from pMTL5402F using primers ErmB-Pro-F3 and ErmB-Pro-RA. The erniB ORF was PCR-amplified from pMTL5402F using primers linkerl -ErmB-Fl and ErmB-Rl . The PCR products were gel-purified and used as templates in a SOEing PCR using the outer primers EπnB-Pro-F3 and ErmB-Rl. The PCR product encoding ErmBtdRAMl SE was cloned into pCR2.1-TOPO. ErmBtdRAMl SE was excised from pCR2.1 "ErmBtdRAMl SE as a Hindlll I

MoI fragment and ligated into pMTL5402F linearised with the same enzymes. This placed ErmBtdRAMl SE in the opposite orientation to thefac promoter on the resulting plasmid ρMTL5402F::ErmBtdRAMlSE.

The ErrnBtaRAMl construct was made as follows. The ermB promoter was PCR- amplified from ρMTL5402F using primers ErmB-Pro-F3 and ErmB-Pro-RB. The ermB ORF was PCR-amplified from ρMTL5402F using primers linkerl-ErniB-Fl and ErmB-Rl . The attenuated td group I intron and its exons were PCR-amplified from ρACD4K-C using primers tdGpI-Fl and tdGpI-Rl. The PCR products were gel-purified and used as templates in a 3-way SOEing PCR using the outer primers ErmB-Pro-F3 and ErmB-Rl. The PCR product encoding ErmBtdRAMl was cloned into pCR2.1-TOPO. ErmBtdRAMl was excised from pCR2.1::ErmBtdRAMl as aiJmdIII / Xhol fragment and ligated into ρMTL5402F linearised with the same enzymes.

E. coli carrying pMTL5402F::EnriBή2RAMl was sensitive to erythromycin at 500 and 125 μg/ml (no growth overnight at 37°C). E. coli carrying pMTL5402F::ErmBή$lAMlSE was resistant to erythromycin at 500 and 125 μg/ml (grew overnight at 37 ⁰ C). These experiments demonstrated that the modified ermB gene conferred resistance to erythromycin in E. coli, and equally important, that the insertion of ft/ inactivates the gene.

Example 3: Validation of the ErmBtd selectable marker in E. coli

The retargeting nucleic acid component of pACD4K-C was sub-cloned as a λføel (blunt) fragment into pMTL20 (Chambers et al (1988) Gene 68: 139-149) between Hindlll and Smal sites and the lacZ re-targeting region again shown to be able to knock-out the lacZ gene in the E. coli host HMS174(DE3). Next the KanRAM in pMTL20/αcZTT was replaced with ErmBtd RAMI as Mlul fragment. To test that the td group I intron was being spliced from εxmQtd RAMI following induction of group II intron RNA expression, E. coli cells carrying pMTL201acZTTErøiBfc£RAMl were harvested and RNA prepared. RT-PCR

reactions were then undertaken using primers that flank the td site of insertion. As a control, standard PCR was performed on EτmBtd RAMI and ErmBtii RAMI SE (the spliced equivalent of ErniBtd RAMI). As can be seen in Figure 7, the predominant product obtained from the IPTG induced RNA samples was of the smaller size corresponding to the SE gene. This clearly demonstrates that td is being spliced from the RNA of the modified ermB gene in ετmBtd RAMI .

Despite the fact that demonstrable splicing of εnήBtd RAMI had been shown to occur, no erythromycin resistant colonies were obtained following plating of the IPTG induced cells on agar media supplemented with 500, 250 or 125 μg/ml erythromycin. It was not possible to reduce the concentration of antibiotic any further, as E. coli is naturally resistant to lower levels of the antibiotic.

Failure to obtain erythromycin resistant colonies may have been due to a copy number effect. Thus, a single copy inserted in the genome may have been insufficient to raise resistance to the antibiotic above the usual low level of resistance inherent to wild type E. coli To test this possibility, a DNA fragment fragment carrying ErrnBftiRAMl SE was ligated to cleaved pACYC184, and the ligation mixture transformed into E. coli and plated on 2YT containing either tetracycline or erythromycin at three different concentrations, 500, 250 and 125 μg/ml. ^' Similar numbers of colonies grew on Erml25 and Tet, but several-fold less grew on Erm250, and only a few grew on Erm500. This control experiment set the practical limit for the screening of the inheritance of ErmB&3RAMl SE when present on pACYC184 as being 125 μg/ml.

Having established the level of erythromycin needed to screen for EmαBfrfllAMl SE in E. coli, a region of lacZ encompassing the targeting region was PCR amplified with primers lacZ target-F (ACGAATTCCGGATAATGCGAACAGC- GCACGG; SEQ ID No. 33) and lacZ target-R (TGCGATCGCACCGCCGA- CGGCACGCTGATTG; SEQ ID NO. 34), cloned into pCR2. ITOPO, and then subcloned into pACYC184, which is present at several copies in the E. coli cell. The re-targeting experiment was then repeated by introducing

into E. coli cells carrying pACYCl 84 v.lacZ. Following induction with IPTG, the cells were plated onto media containing erythromycin. In contrast to the previous experiment, appreciable numbers of resistant colonies were obtained. The use of appropriate primers in a diagnostic PCR confirmed that re-targeting of the group II intron to the lacZ gene on pACYC184 had taken place. Therefore, when ErrnBtaflRAMl SE is present as a single copy, expression of ErmB is insufficient to confer resistance to erythromycin, but when present in multiple copies, ErmB is expressed at a sufficient amount to confer the resistant phenorype.

Example 4: Construction of a Clostridial retargeting system using the ErmBtd selectable marker

Having established that EπnBtdKAMl could substitute for the KanRAM in the Sigma- Aldrich group II intron, the entire element, together with the re-targeting region for lacZ, was subcloned from ρMTL201acZTTEπriB/tfRAMl (as

HinάllllSacl and Sacl/Nhel fragments) into the clostridial expression vector pMTL5402F (cleaved with Hinάlll-Nheϊ) to give pMTL5402FlacZTTErmBήiRAMl. As a consequence, expression of the group II intron was under the control of the fac promoter. Expression of the group II intron will be regulated by IPTG.

The ability of this vector to re-target the lacZ gene on pACYCl 84 wlacZ in E. coli was tested. Following IPTG induction and plating on erythromycin, successful re- targeting was demonstrated.

Example 5: Determination of the efficiency of ErmBftflRAMl in group II intron retargeting

To assess whether EnnBήiRAMl affects the frequency with which the group II intron can retarget, compared to KanRAM, we undertook some mobility assays

using a two-plasmid system developed Karberg et al (2001, supra). Retargeting of the group II intron from ρACD2, following IPTG-induction, to ρBRR3-LtrB (which carries its natural target, LtrB) results in activation of the Tet gene on the latter plasmid. Thus, individual retargeting events can be detected on the basis of acquisition of resistance to Tetracycline.

Plasmid pACD2 was therefore modified by the insertion of either the ErmBtaRAMl or the KaήRλM, into the vector's unique MIuI site. These two plasmids were then transformed into HMS174(DE3) cells containing pBRR3-LtrB — i.e. the recipient plasmid with the wild type target sequence. After selection for the donor plasmid, cells were induced with 500 μM IPTG for lhr, re-suspended in LB, allowed to recover for lhr, and then various dilutions were plated onto various selective plates. For those constructs containing a RAM, Tet ^R colonies were first re-streaked onto Tet plates and then again onto plates containing the appropriate antibiotic to test RAM splicing. Results are shown in Table 2.

This experiment demonstrated that the KanRAM and ErmBtdRAMl have a similar effect on intron efficiency - presumably mainly due to the increased size of the intron. Importantly, the data indicate that both RAMs splice at similar efficiencies.

Table 2: Results of mobility assays

* Intron mobility efficiency = Tet ^R colonies / Amp ^R Cm ^R colonies t RAM splicing efficiency = Kan ^R or Erm ^R re-streaked Tet ^R colonies / all re-streaked Tet ^R colonies

Splicing of neither RAM could be detected by antibiotic resistance initially, but only when re-streaked from Tet ^R colonies.

Example 6: Identification of effective clostridial re-targeting sequences

To evaluate retargeting of the eight different test genes were chosen from 3 different clostridial species. These were: Clostridium sporogenes pyrF, spoOA, codY, and SONO, Clostridium difficile pyrF (Genome Annotation No. CD3592) and spoOA (Genome Annotation No. CD1214), Clostridium aceotbutylicum pyrF (Genome Annotation No. CAC2652) and spoOA.

Each gene was analysed at http://www.signia-genosys.com/targetiOn/, and suitable changes to allow for re-targeting identified. Using appropriate primers, the generation of appropriately modified Group II introns was effected by performing a PCR as directed in the Sigma-Aldrich TargeTron™ Gene Knockout System User Guide. Each PCR required unique IBS, EBS2 and EBSId primers designed to modify the targeting portions of the Group II intron or its 5' exon, and the EBS Universal primer. The sequences of the target insertion sites for each gene and primers are given in Tables 3 and 4 below.

Table 3 : Predicted target insertion sites for retargeting nucleic acids

^a Target indicated as organism, ORP and insertion point. Target insertion sites were selected such that introns would be inserted after the indicated number of bases from the start of the ORF, in either the sense (s) or antisense (a) orientation.

. Table 4: Oligonucleotide primers used to generate PCR products for retargeting

To ensure that the modified group II introns were capable of retargeting to the selected clostridial genes, experiments were first undertaken in E. coli using plasmid systems that were known to function effectively. The system utilised is a two-plasmid system developed by Karberg et al (2001) as described in Example 5. Using this system, the engineered group II intron is placed on one plasmid (pACD2) and its target (in this case the cloned clostridial gene) is placed on a second plasmid (pBRR3). Retargeting of the group II intron from pACD2 to pBRR3 results in activation of the Tet gene on the latter plasmid. Thus, individual retargeting events can be detected on the basis of acquisition of resistance to Tetracycline. A portion of the bacteria are plated on non-selective agar plates to give an indication of total viable bacteria and a portion are plated on tetracycline- containing agar plates (Tet plates). The efficiency of retargeting is estimated based on the proportion of total viable bacteria that are resistant to tetracycline.

To facilitate subcloning of the target genes from ρCR2.1 / pCRII TOPO plasmids into pBR_R3, a multiple cloning site was introduced into pBRR3-LtrB to make pBRR3-MCSl. This was done by insertion of a multi cloning site fragment

between the AatII and EcόBl sites of pBRR3-LtrB ₅ containing restriction sites found in the pCR2.1-TOPO plasmid. Sequences of the cloning sites are given in Figure 8. The multi cloning site fragment depicted in Figure 8 was made from MCSIa oligonucleotide CTCGAGGTACCATGCATAGGCCTGAGCTCA- CTAGTGCGGCCGCG (SEQ ID NO. 68) and MCSIb oligonucleotide AATTC- GCGGCCGCACTAGTGAGCTCAGGCCTATGCATGGTACCTCGAGACGT (SEQ ID No. 69).

Four retargeting nucleic acids (each intended for insertion in one of C. sporogenes genes pyrF, spoOA, codY, and SONO) were evaluated using the two-plasmid intron mobility assay. All four permitted far more efficient retargeting than anticipated. Consequently the dilutions chosen for plating on Tet plates were not ideal and were only just in range for colony counts and therefore the efficiencies given may be less accurate than if fewer bacteria had been plated. The next four retargeting nucleic acids (each intended for insertion in one of C. difficile pyrF and spoOA genes or C. aceotbutylicum pyrF and spoOA genes) were evaluated for retargeting. Retargeting events were estimated by plating bacteria on Tet plates. In this initial experiment, SONO gave no Em ^R colonies. Results are shown in Table 5.

Table 5: Results of intron mobility assay

* Intron mobility efficiency = Tet ^R colonies / Amp ^R Cm ^R colonies, Cs - C. sporogenes, Ca - C. acetobutylicum, Cd - C. difficile. Numbers refer to the site of intron insertion relative to the start of the gene, in either the sense (s) or antisense (a) orientation. TR= retargeting nucleic acid

Example 7: Evaluation of retargeting nucleic acids in Clostridia

The first four new retargeting nucleic acids (each intended for insertion in one of C. sporogenes genes pyrF, spoOA, codY, and SONO) were sub-cloned into the prototype vector pMTL5402FTTErmB tdRAMl, and the resultant recombinant plasmids introduced into the E.coli donor CA434, and thence used in conjugation experiments with either C. sporogenes or C. difficile as the recipient. In the case of the latter, no trans conjugants were obtained. Transconjugants were obtained with both plasmids in the case of C. sporogenes. Single transconjugants were inoculated into 1.5ml of an appropriate growth medium supplemented with 250μg/ml cycloserine and 7.5μg/ml thiamphenicol (the latter of which ensures plasmid maintenance) and the culture was allowed to grow to stationary phase by anaerobic incubation at 37 ⁰ C overnight. 150μl of this culture was used to inoculate 1.5ml of fresh broth of the same type and containing the same supplements, which was then incubated anaerobically at 37°C. As soon as growth was visible in the culture, typically after lhr, the culture was induced with ImM IPTG and incubated for 1 hr.

2ml of the induced cells were harvested by centrifugation for 1 minute at 7000 rpm, washed by re-suspension in PBS and harvested as before. The pellet was re- suspended in an equal volume (2ml) of an appropriate growth medium without supplements, and incubated anaerobically at 37 ⁰ C for 1 hour. Serial dilutions of the culture were then plated onto an appropriate solid growth media supplemented with l-10μg/ml erythromycin, after 1 hr, 24 and 48 hr and incubated anaerobically at 37 ⁰ C.

No erythromycin colonies were obtained after two independent attempts.

Example 8: Evidence that the natural ermB promoter is too weak to drive expression of ErmB sufficient for it to act as a selectable marker

One explanation for the inability to detect retargeting of the retargeting nucleic acids as described in Example 7 is that the ermB promoter is too weak to allow a single copy of the gene in a cell's chromosome to confer resistance to erythromycin. Analysis of the ermB promoter sequence of Enterococcus faecalis plasmid pAMβl showed that the spacing . between the promoter's -35 and -10 regions is 21bp. The optimum for Gram-positive promoters is 17±lbp.

The EτmBtdRAMl SE was cloned in two different orientations in pMT5402F relative to fac. Only when the gene was under the control of fac was the plasmid carrying επήBtdRAMl SE capable of endowing the C. sporogenes host with resistance to erythromycin, hi the opposite orientation, transcription of the ermB coding region is reliant on its own promoter and expression was insufficient for resistance, despite ermB being present on a multi-copy plasmid.

Example 9: Development of a Clostridial ErmBfcflRAM with a strong promoter

The promoter of the thl gene of C. acetobutylicum is recognised as a strong and constitutive promoter. Primers were designed to replace the promoter with the thl promoter. These delete unnecessary sequences between the transcriptional start site and ribosome binding site, and insert an Ndel site at the start codon to allow the promoter to be easily changed again if necessary. To guard against the possibility that the thl promoter might be too strong, a mutant thl promoter HhIT was also designed, by changing the spacing between the -35 and - 10 to 16 nt, and making minor changes to the -35 and -10 elements. Sequences of the thl and thl2 promoters spanning the -35 and -10 elements compared to a consensus Gram positive vegetative promoter are shown in Figure 9. The sequence of the complete thl promoter is given at positions 15-84 of the

BrmBtdRAMl sequence in Figure 10. The sequence of the complete thl2 promoter differs from the thl promoter only in the region depicted in Figure 9.

The thl and thϊl promoters were fused to the ErmB^JRAMl and ErmBføfRAMl SE start codons using SOEing PCR and cloning steps, producing ErmBf JRAM2 and EπήBtdRAML SE, each containing the thl promoter, and EmxBtdRAM3 and

ErmBήiRAM3 SE each containing the thϊl promoter. Sequence-correct clones were obtained of both RAM2 and RAM3, and were sub-cloned into ρACD2 and pMTL201acZTT for evaluation. The features and sequence of ErmBϊJRAM2 is depicted in Figure 10.

The ability of the RAM2 and RAM3 portions to confer resistance to erythromycin on E. coli TOPlO cells was determined for plasmids containing these portions as indicated in Table 6 below.

Table 6: Erythromycin sensitivity of TOPlO clones bearing new constructs

RAM TOPO SE pACD2::RAM pMTL20- pMTL5402F-

TOPO lacZTTRAM TTRAM

EtmBtdKAM.2 R R S S S

EtmBtdBJMS S R S S S

In all but the SE TOPO plasmid, the ermB gene is disrupted by the group I intron, and therefore resistance was not expected. In the SE TOPO plasmid, the ermB gene is not disrupted, and so if the promoter of ermB is sufficiently strong, a resistance phenotype should be obtained. Unexpectedly, RAM2 TOPO clone conferred resistance to erythromycin in E. coli. It would appear that the very strong thl promoter and very high copy number seems to overcome the presence of the group I intron in this context, presumably by rare translation initiation at the. native ATG, This effect is only seen when the gene is present in TOPO. When inserted in a plasmid relevant for retargeting, such as pACD2 and ρMTL201acZTTRAM, E. coli cells are not resistant to erythromycin. The RAM3 resistance profile was

as expected. Therefore, either promoter appeared to be useful to drive expression of the selectable marker in the Clostridial retargeting nucleic acid.

Example 10: Evaluation of ErmB?tfRAM2 and 3 in pACD2/pBRR3 system

The retargeting efficiency of the ErmBfc£RAM2 and 3 were evaluated using the retargeting assay described in Example 5. Results are indicated in Table 7 below.

Table 7: Results of intron mobility assay

* Intron mobility efficiency = Tet ^R colonies / Amp ^R Cm ^R colonies t RAM splicing efficiency = Kan ^R or Erm ^R re-streaked Tet ^R colonies / all re-streaked Tet ^R colonies

Splicing of RAM2 and RAM3 was efficient (90%), and equivalent to the original RAM (RAMl).

Example 11: Evaluation of ErmBtdRAM2 and 3 in pMTL201acZTT system in E.coli

As previously, neither RAM could be used to detect retargeting of the Group II intron into lαcZ in the E. coli HMS 174(DE3) chromosome using either Erysoo or

Ery ₁₂₅ . RAM2 but not RAM3 gave numerous Ery ^R colonies, but they were shown by PCR not to contain a Group II intron retargeted to the lacZ gene. Presumably these colonies arose due to weak resistance conferred by the plasmid.

Example 12:

Figure 11 illustrates the essential components of the vector ρMTL007, also referred to as pMTL5402FlacZTTErmBtdRAM2. This Group II intron is modified to retarget the lacZ gene. It may be modified to retarget a gene of a bacterial cell of the class Clostridia.

The essential elements of the plasmid are:-

A clostridial promoter to bring about expression of the retargeting nucleic acid element, which in the illustrated example is the inducible fac promoter. Other promoters may be similarly employed which have been made inducible by provision of a lac operator, eg., fac2. To mediate induction the plasmid also carries the E.coli lad gene under the control of a clostridial promoter, in this instance the promoter of the ptb gene (encoding phosphotransbutyrylase) of Clostridium acetobutylicum. A constitutive promoter may be used instead of an inducible promoter. The plasmid also carries the CoIEl replicon of plasmid pMTL20E, to allow maintenance of the plasmid in E.coli and the replication region of the Clostridium butyricum plasmid pCB102, to allow maintenance in Clostridium species. Maintenance of the plasmid is also provided by the inclusion of the catP gene to enable selection of the plasmid in E.coli (through supplementation of the media with chloramphenicol) and Clostridia (through supplementation of the media with thiamphenicol). To provide the facility to conjugate the plasmid into clostridial recipients in addition to transformation, the vector also carries the oriT 'region of plasmid RP 4.

AU of these elements are interchangeable with other equivalent factors from other sources. Thus, CoIEl maybe exchanged with other replicons capable of

replication in E.coli, sucli as pi 5a, pVWOl or phage origins such as M13. The catP gene may be substituted with other appropriate antibiotic resistance genes, such as tetM or aad. Similarly, any replicon capable of replicating in the targeted clostridial or Gram-positive bacterial host may be employed, such as pIM13, pIP404, pAMBl, pCD6, ρC194 ₅ pE194, pTISl, pCBlOl, pBPl. Replicons which are defective for replication may also be employed, including replicons that may be conditional for replication, eg., temperature sensitive or reliant on an exogenous factor for replication. Plasmids may also be employed which lack any provision for replication in a Gram-positive plasmid, ie., a suicide vector carrying only a CoIEl replicon

Other combinations of operator and repressor gene may be emploj'ed. A promoter identified on the conjugative transposon Tn53P7 which is regulated by tetracycline (Tet), represents another candidate (Roberts, PhD thesis, UCL). Alternatively, a xylose-inducible promoter, derived from S. xylosus, has recently been shown to function in C. acetobutylicum (Girbal et al (2003) Appl Environ Microbiol. 69: 4985-8). Another candidate is a tet-τegalateά promoter developed in B. subtilis (Geissendorfer and Hillen (1990) Appl. Microbiol. Biotechnol. 33: 657-63). It was constructed by adding a tet operator itetO) sequence between the -35 and -10 of a strong xyl promoter (Geissendorfer and Hillen, 1990, supra). In the presence of a tetR gene (encoding the repressor), the derivatised promoter was 100-fold inducible by sub-lethal concentrations of Tet. The basal levels of expression obtained could be completely abolished by the addition of a second tet operator, although this addition caused an overall reduction in expression levels. Subsequently, this promoter has found wide application in S. aureus (Bateman et al (2001) Infect Immun. 69: 7851-7; Ji et al (2001) Science 293: 2266-2269), where only a single operator proved necessary.

A tet-regulated promoter makes an ideal alternative to our developed fac I lad system. Thus, we will be able to express tetR using the same promoter used to express lad (the C. acetobutylicum ptb promoter). In B. subtilis, the degree of induction was dose-dependent over the range tested. However, as B. subtilis was

sensitive to the antibiotic, high, concentrations of Tet could not be added. A similar constraint will not apply to Clostridia such as C. difficile, which are resistant to this antibiotic. To test the feasibility of the system, we will re-synthesise^c, replacing the region between the -35 and -10 with the tetO. Should high basal levels be observed in the absence of Tet, then a second operator can be added. Addition of further synthetic lacO sequences can also be used to enhance repression of promoters by Lad (Muller et al (1996) J MoI Biol. 257: 21-9).

Construction of pMTL007

Oligonucleotide primers used in the construction are indicated in Table 8 below.

Table 8: Oligonucleotide primers

Primer Sequence (5' - 3') SEQ ID No. lad-Pi GTGGTGCATATGAAACCAGTAACG 70

Iacl-P2 GAATTCCTAACTCACATTAATTGCG 71

TTGCG ptb-Pl GAATTCAGGGAATTAAAAGAATGT 72

TTACCTG ptb-P2 ACTCATATGTTGCACCTCTACTTTA 73

ATAATTTTTAAC tdGpI-Fl GCATTATGTTCAGATAAGGTCGTTAATCTT 29

ACCCC CatPFwd CAGCTGACCGGTCTAAAGAGGTCCCTAGC 74

GCC

CatPSOER CGGTCATGCTGTAGGTACAAGGTAC 75

CatPRev CAGCTGACCGGTCTCTGAAAATATAAAAA 16

CCACAGATTGATAC

CatPSOEF GTACCTTGTACCTACAGCATGACCG 77

Thio-Fl CTACTAGTACGCGTTATATTGATAAAAATA 78

ATAATAGTGGG Thio-R-RAM CCTTATCTGAACATAATGCCATATGAATCC 79

CTCCTAATTTATACGTTTTCTC

A 1.627 lcb LspI-HinάJω. fragment was isolated from the Clostridium butyricum plasmid pCB102 (Minton and Morris (1981) J Gen Microbiol 127: 325-33) and blunt-ended with Klenow polymerase. The replicon cloning vector pMTL21E (Swinfield et al (1990) Gene 87:79-89) was cleaved with MeI, blunt-ended with

Klenow polymerase and ligated with the isolated pCB102 replicon fragment. The resultant plasmid was designated pMTL540E (T Davis, PhD Thesis, The Open University, 1989), as shown in Figure 12

The inducible promoter element was derived from the promoter of the ferredoxin gene of Clostridium pasteurianum. Now termed fac, it was created by adding an E.coli lac operator immediately after the +1 of the ferredoxin gene promoter, and altering the sequence immediately preceding the ATG start codon of the ferredoxin structural gene to CAT, thereby creating a Ndel restriction site (CATATG) (Minton et al (1990) Vector systems for the genetic analysis of Clostridium acetobutylicum In: Anaerobes in Human Medicine and Industry (eds P Boriello & J Hardie), Wrightson Publishing, Petersfield, UK pp. 187-206). In this particular instance the lac operator inserted in the opposite orientation relative to transcription compared to the lac promoter. However, this does not affect functionality, and Lad protein will still bind and repress transcription from the promoter.

The fac promoter was then sub-cloned as an Ndel and EcoRI restriction fragment, between the equivalent sites of plasmid pMTL1003 (Brehm et al (1991) Appl. Biotechnol. 36, 358-363), generating plasmid pMTL1006. This sub-cloning step effectively removed the trp promoter of pMTL1003 and placed the expression of lacZ' under the control of the modified/^ promoter. Plasmid pMTL1006 was then subjected to a BgH digest and the larger of the two resultant fragments isolated. Plasmid pMTL500E (Oultram et al (1988) FEMS Microbiol Letts 56: 83-88) was similarly cleaved with BgU. and the larger of the two fragments isolated and ligated with the larger fragment isolated from pMTL1006. The plasmid obtained was designated pMTL500F (Fox et al, 1996, supra).

Although pMTL500F is replication proficient in Clostridia, we have found that the transfer frequency of pAMβl based shuttle vectors into Clostridia is relatively inefficient. We therefore elected to change the replicon to that of pCB102.

Accordingly, both pMTL540E and pMTL500F were cleaved with BgK, and the

larger fragment of pMTL540E ligated to the smaller fragment of pMTL500F. The plasmid obtained was designated pMTL540F (Fox et al, 1996, supra). For simplicity, the ligation of pMTL500E and pMTL1006 fragments, and the ligation of pMTL540E and pMTLSOOF fragments is represented as a single ligation of pMTL540E and pMTLl 006 in Figure 12.

To enable conjugative transfer of the plasmids for those instances where transformation has yet to be demonstrated, we elected to endow the plasmid with the oriT (origin of transfer) of plasmid RP4. As such, the RP4 oriT region was excised from pEoriT (Purdy et al, 2002) using EcoRV and Smal, and sub-cloned into the EcoRV restriction site of pMTL540F, generating pMTL5400F (see Figure 12).

To bring about the production of Lad repressor protein, a promoter-less copy of the E.coli lad gene was amplified from pNM52 (Gilbert et al (1986) J. Gen.

Microbiol. 132: 151-160) as an approx 1.0 fcb Ndel-EcoRI fragment using the

PCR primers lacI-Pl and Iacl-P2 In parallel, the promoter region of the

Clostridium acetobutylicum ptb (phosphotransbutyrylase) gene was PCR amplified using the primers ptb-Pl and ptb-P2. This localised the gene to a 578 bp EcoRl-Nάel fragment. The two fragments were isolated and ligated with EcoRI- cleaved pMTL20E, thereby placing the lad gene under the transcriptional control of the ptb promoter, and localizing the modified gene to a portable EcoRl fragment. This fragment was excised from the plasmid generated, blunt-ended with Klenow polymerase, and ligated with ϋeoRV-cleaved pMTL5400F. The plasmid obtained was designated pMTL5401F, as shown in Figure 12.

Plasmid pMTL5401F carries an erm gene as the selectable marker. It is, therefore, not compatible with the ErmRAM. The erm gene was therefore replaced with the catP gene of pJIR418 (Sloan et al (1992) Plasmid 27: 207-219). This was achieved by cleaving pMTL5401F with Midi I Tthllll, blunt-ending the DNA with Klenow polymerase, and then ligating to a 1.1 kb PvwII fragment carrying the pJIR418 catP gene to the larger of the two pMTL5401F fragments generated by

cleavage with Ahdl and Tthllll. This manipulation resulted in the complete deletion of ermB and removal of the majority of the bla gene. The plasmid obtained was designated pMTL5402F, as shown in Figure 13.

Prior to this substitution, a BsrGl site within the catP fragment was removed by mutating a sequence to destroy the BsrGl palindrome without changing the catP coding sequence. This was undertaking using Sewing Overlap Extension (SOE) PCR (Horton et al, 1990), using the primers CatPSOEF and CatPSOER. In addition, the flanking primers CatPFwd and CatPRev used were designed to encompass both a Pvull site and internal Agel sites. The former were incorporated for the subsequent insertion of the plasmid into pMTL5401F, whereas the latter were introduced to facilitate the subsequent substitution of catP in the final plasmid, pMTL007, with alternative markers at a future date.

pMLT007 was constructed as follows:

The Targetron™ plasmid pACD4K-C was purchased from Sigma, and re-targeted to the E. coli lacZ gene using the control primers provided in the kit according to the provided protocol, except that the PCR product was first cloned and its sequence verified before sub-cloning the Hindlll/BsrGI fragment into pACD4K- C.

The lacZ-retargeting nucleic acid region was excised as a 5099bp Nael fragment and ligated into a 2412bp fragment of pMTL20 which had previously been generated by digestion with HindIII and Smal, with T4 polymerase blunting of the HindIII end, as shown in Figure 13. A construct was chosen in the orientation in which the HindIII and Nhel sites flanked the retargeting nucleic acid region.

The KanRAM was excised using MM, and replaced with a 1259bp MIuI fragment containing ErmBtdRAM2, as shown in Figure 13.

The entire lacZ-retargeting nucleic acid region including the ErmRAM was then excised as a ~3.3kbp Hindlll/Sacl fragment and a ~1.8kbp Sacl/Nhel fragment, which were ligated together into pMTL5402F digested with HindIII and Nhel. The resulting plasmid was designated pMTL5402FLacZTTErmBtdRAMl.

The thl promoter of C. acetobutylicum ATCC 824 was PCR-amplified from pSOS95 (Tummala et al (2003) J. Bacterial 185: 1923-1934) using primers Thio- Fl and Thio-R-RAM. The PCR product was gel-purified and used, along with the td group I intron PCR product from the construction of ErmBtdRAMl, as template in a SOEing PCR using the outer primers Thio-Fl and tdGpI-Rl. The thl promoter and part of the td intron were excised from this PCR product as a 143bp Speϊ I Nspl fragment. The remainder of the td intron, and the ermB ORF from pCR2.1::ErmBtdRAMl were excised together as a Nspl I Notl fragment. These fragments were ligated in a three-way ligation into pCR2.1 "ErmBtdRAMl SE linearised with Spel and Notl, yielding plasmid pCR2.1 : :ErmBtdRAM2.

The MM / MM fragment of ρCR2.1::ErmBtdRAM2 containing the RAM was ligated with the larger Mlul / MM fragment of pMTL201acZTT to form ρMTL201acZTTErmBtdRAM2 as shown in Figure 14.

A BsrGI /BstBI fragment of pMTL201acZTTErmBtdRAM2 containing the RAM was subcloned into BsrGI / BstBI cleaved pMTL5402FlacZTTErmBtdRAMl to generate pMTL007 as shown in Figure 14.

ρMTL007 was initially designated ρMTL5402FlacZTTErmBtdRAM2, and sometimes referred to as ρMTL5402FlacZTTErmRAM2 or pMTL5402FlacZTTRAM2 or pMTL5402FlacZTTR2.

Once re-targeted, the plasmid was designated pMTL007 (or pMTL5402FTTEiτnBtdRAM2 or pMTL5402FTTErmBRAM2 or ρMTL5402FTTRAM2 or pMTL5402FTTR2) suffixed by an identifier for the

'Targeting Region' (TR). The TR is the entire region between the HindIII and

BsrGI sites of the sequence generated by the re-targeting PCR, For example, once the plasmid was re-targeted to the C. difficile 630 gene spoOA, at position 178 of the spoOA ORF, by cloning the appropriate TR fragment in as a Hindlll/BsrGI fragment, the plasmid was designated pMTL5402FTTErmBtdRAM2::Cd-sρo0A- 178aTR.

Example 13: Evaluation of ErmB2tfRAM2 and 3 in pMTL5402FlacZTT system in C. sporogenes against codY

Having generated new RAMs that were capable of giving erythromycin resistance in Clostridia, Clostridial retargeting nucleic acids comprising a modified Group II intron having targeting portions designed to target the intron to C. sporogenes against codY were constructed. Two plasmids were constructed bearing either the RAM2 or the RAM3 and named pMTL5402FCs-codY-417sTT::RAM2 and RAM3 respectively. The RAM2 version is identical to that depicted in Figure 10, except the retargeting portions of the Group II intron and the IBS sequence are designed to allow retargeting of C. sporogenes against codY instead of E. coli lacZ. Either plasmid was conjugated into Clostridium sporogenes. Transconjugants were verified by PCR and shown to be completely sensitive to Eryi. ₂₅ . A selected transconjugant of each RAM was then induced with IPTG, and after removal of inducer by centrifugation and washing, allowed 3 hours recovery before plating out on agar plates containing a range of concentrations of erythromycin.

The number of colonies obtained is shown in Table 9 below.

Table 9: Results of retargeting assay

These data demonstrate the importance of a sufficient induction period, and the superior efficiencies achieved using RAM2 compared to RAM3.

Example 14: Further mutant generation

We elected to target two genes whose inactivation would lead to easily detectable phenotypes. These were pyrF and spoOA. Inactivation of the former should lead to uracil auxotrophy, while disruption of the latter should lead to asporogeny.

pMTL007 was re-targeted to the C. sporogenes spoOA gene and hundreds of Em ^R colonies of C. sporogenes were readily obtained after IPTG induction. DNA was extracted from four random colonies, and used as a template in PCR. In all cases, primers specific to the RAM generated a DNA fragment of a size consistent with loss of the td intron (Figure 15).

Having demonstrated apparent functionality of the RAM with pMTL007::Csp- spoOA-249s, we proceeded to generate mutants in the two genes (pyrF and spoOA) in all three clostridial species using the protocols outlined in the methods section. PCR screening of Em ^R clones (Figure 15b, c) revealed very high frequencies of insertion into the intended chromosomal site (Table 10), demonstrating how easily integrants can be obtained using this method. After isolation, single colonies of

integrants were screened for plasmid loss by thiamphenicol-sensitive phenotype, and colonies cured of the plasmid were found to predominate in all these organisms without additional passaging. Insertion sites were verified by sequencing across the intron-exon junctions (Table 10) and Southern blotting with a probe for the RAM confirmed the presence of a single copy of the insertion element (Figure 15d, e and f).

IPTG induction of intron expression from pMTL007::Csp-.spo&4-249s in C, sporogenes increased the insertion frequency by over 100-fold (Table 10), in keeping with the reporter data from pMTL5401 V cat.

Table 10: Effect of regulated intron expression on insertion frequencies in C. sporogenes.

Plasmid IPTG ^a Insertion Relative insertion frequency ^b frequency ⁰ pMTL007: :Cs$-spoOA-249s - < 1.31 ± 0.34 X lO ^"9 1 pMTL007: :Csp-spoOA-249s + 1.63 ± 0.72 x lo- ⁷ 124 ρMTL007:. -Csp-spoOA- - 1.95 ± 0.54 x 10 ^"6 1489 249sAlacI ^a Intron expression was induced with 1 mM IPTG (+) or with water in place of IPTG (-). ^b After the recovery period, cells were spread onto TYG cycloserine plates with or without erythromycin supplementation. Insertion frequencies are expressed as Em ^R c.f.u./ml / total c.f.u./ml. ^c Relative insertion frequencies are normalised to the experiment with pMTL007::Csp-,spo 0,4 -249s and water in place of IPTG.

To establish whether regulated expression of the intron conferred any advantage over constitutive expression, we de-repressed the fac promoter by introducing a frameshift mutation into the lad gene of pMTL007::Csp-spoOA-249s. A further insertion frequency increase of over 10-fold was observed. (Table 10), indicating that regulated expression of the intron confers no advantage over constitutive expression. We performed an equivalent experiment in C. acetobutylicum with

pMTL007::Cac-,spo6L4-242a and observed no change in integration frequencies with the addition of IPTG (data not shown). Consistent with the pMTL5401Fcαt reporter data, basal intron expression from the fac promoter in this organism is apparently sufficient to achieve easily-detectable integration frequencies. Like pMTL5401F, pMTL007 is too unstable in C. difficile to support the growth of its transconjugants in antibiotic-supplemented liquid culture (Purdy et at (2002) MoI. Microbiol. 46: 439-452). Therefore no comparable IPTG-induction experiments to those undertaken in C. sporogenes could be performed. However, in both C. difficile and C. acetobutylicum, Em ^R integrants could be easily obtained by simply re-streaking transconjugant colonies onto growth media containing erythromycin with no addition of IPTG.

As anticipated, all the spoOA mutants were unable to form endospores (Figure 16). AU of the pyrF mutants were shown to be unable to grow on minimal media unless supplemented with 50 μg/L uracil. We attempted to select revertants to uracil prototrophy by growing all three clostridial mutants in rich liquid media lacking erythromycin selection and then plating them onto minimal agar medium with or without uracil. Revertants were never detected on media lacking uracil in at least three experiments. By comparison to the cell counts on media supplemented with uracil, reversion frequencies per cell were estimated to be less than 9.36 x 10 ^'9 in C. difficile, less than 9.60 x 10 ^"7 in C. acetobutylicum and less than 5.50 x 10 ^"9 in C. sporogenes. These findings are consistent with data in the literature (Frazier et al (2003) Appl. Environ. Microbiol. 69: 1121-1128) showing that intron integrants are extremely stable - a ^' highly desirable mutant characteristic.

Example 15: Evaluation of ErmBtdRAMl system against other targets

A standard protocol has been developed for retargeting in Clostridia, as follows. JL Intron re-targeting sequences to the gene of interest are generated essentially according to the method provided by Sigma with the Targetron™ kit:

The computer algorithm provided at the Sigma website |Tittp://www.sigrna- genosys.com/targetron/] is used to identify possible intron targets within the sequence of the gene of interest, and to design PCR primers. These primers are then used according to the Sigma Targetron™ protocol, and using PCR reagents provided in the Sigma Targetron™ kit, to generate a 353bρ PCR product which corresponds to part of the intron and includes modified IBS, EBSId and EB S2 sequences such that the intron can be re-targeted to the gene of interest. This PCR product is cloned into an appropriate cloning vector such as pCR2.1 and its sequence verified. Alternatively, it may be subcloned directly into pMTL007.

2. The prototype clostridial retargeting plasmid pMTL5402FlacZTTR2 is re-targeted essentially according to the method provided by Sigma with the

Targetron™ kit:

If the PCR product of step 1 was cloned into a cloning vector, the desired re- targeting sequence is excised from its plasmid by digestion with the restriction enzymes HindIU and 557'GI, and cloned into pMTL5402FlacZTTR2 digested with the same enzymes. In either case, the resultant constructs are verified by restriction analysis and / or sequencing.

3, The successfully re-targeted clostridial retargeting plasmid is transferred into the target organism:

Recombinant plasmids may be introduced into the clostridial hosts by standard DNA transfer methods based either on electrotransformation or conjugation. Methods for either are given in Davis I, Carter G, Young M and Minton NP (2005) "Gene Cloning in Clostridia", In: Handbook on Clostridia purre P, ed) pp. 37-52, CRC Press, Boca Raton, USA. In our experiments, plasmids were introduced into Clostridium difficile and Clostridium sporogenes by conjugation from E. coli donors. In contrast, plasmids were introduced into Clostridium acetobutylicum by transformation.

4. Retargeting nucleic acid expression and subsequent integration is achieved by induction of the transformant with IPTG:

An individual transformant colony is used to inoculate 1.5ml of an appropriate growth medium supplemented with 250μg/ml cycloserine and 7.5μg/rnl thiamphenicol (the latter of which ensures plasmid maintenance) and the culture is allowed to grow to stationary phase by anaerobic incubation at 37 ⁰ C overnight. 150μl of this culture is used to inoculate 1.5ml of fresh broth of the same type and containing the same supplements, which is then incubated anaerobically at 37 ⁰ C. As soon as growth is visible in the culture, typically after lhr, the culture is induced with ImM IPTG and incubated for 3hrs.

5, Retargeting nucleic acid integrants are detected and isolated using a recovery step followed by plating of cells onto selective solid media and incubation:

2ml of the induced cells are harvested by centrifugation for 1 minute at 7000rpm, washed by re-suspension in PBS and harvested as before. The pellet is re- suspended in an equal volume (2ml) of an appropriate growth medium without supplements, and incubated anaerobically at 37°C for 3hrs. Serial dilutions of the culture are then plated onto an appropriate solid growth media supplemented with 1-lOμg/ml erythromycin, and incubated anaerobically at 37°C. Erythromycin resistant colonies corresponding to retargeting nucleic acid integrant clones can be picked from these plates after 18-48hrs, depending upon the organism and erythromycin concentration used.

Optionally, serial dilutions of the culture can additionally be plated onto unsupplemented solid growth media or solid growth media supplemented with 15μg/ml thiamphenicol in place of erythromycin in order to determine the frequency of the integration event.

The standard protocol was used to make Clostridial mutants as indicated in Table 11.

Diagnostic PCR primers give a product of the expected size if the retargeting nucleic acid has inserted in the targeted gene. a Presence of desired mutant demonstrated in a pool of several clones b Several individual clones screened for desired mutation

Sometimes, retargeting is inefficient. Therefore, it is recommended to try more than one targeting portion to disrupt any given gene. Furthermore, colonies may be pooled before PCR screening of combined batches. If, say portions of 10 or 100 colonies were combined and a PCR product of the size expected for a retargeted mutant was generated, colonies could then be individually screened.

Example 16: Further mutant generation

To further establish the utility of the method, we selected several other genes from each of the three species, and repeated the mutagenesis procedure. The genes targeted are listed in Table 12 and the oligonucleotide primers used to generate PCR products according to the standard protocol for modification of the Group II intron of pMTL007 are shown in Table 4 or Table 13. In every case the desired integrant was obtained. Each insertion was confirmed by PCR screening and the insertion point verified by nucleotide sequencing.

Table 12: Intron insertion frequencies with erythromycin selection

Target Em ^F Desired Frequency of desired Insertion site verified by sequencing SEQ ID

(Organism, ORF and insertion point ¹ ) clones mutants mutant among Em ^R No screened* ¹ obtained ⁵¹ clones' ⁵

C. sporogenes spoOA 249s 4 4 100% TATGCCAAGG- - intron - GTAATTGTTT 80

C. difficile spoOA 11 '8a 3 3 100% ATTACATCTA- intron - GTATTAATAA 81

C. acetobiitylicum spoOA 242a 8 4 50% TATTCTTGGA- intron -AGGTTTTCTG 82

C. sporogenes pyrF 595s 2 2 100% ATAGGAGCAG- - intron - TAGTTGGATG 83

C. difficile pyrF 97a 96° 7-19 ^c 7-20% ^c ACCTTAAATA- intron - TGTCTACACT 84

C. αceiobutylicum pyrF '345s 8 2 25% CTTTGAAGGT- intron - GATTTTGAAG 85

C. αcetobutylicum CAC0081 141a 6 6 100% TTTTAATGAC- intron -ATAGTTTATA 86

C. αcetobutylicum CAC0080 121s 6 6 100% CTGAAATTAT- intron -TTCGTTAATA 87

C. αcetohutylicum CAC0078 385a 3 3 100% GTATCTCCAG- intron - GCGCATATCT 88

C. αcetobutylicum CAC2208 201s 4 3 75% TGTGGAGTAT- - intron - TCGGTACACA 89

C. difficile CDO153 IMa 5 5 100% CCAATAAGCC- -intron - CATCTCCAGA 90

C. difficile CD055275a 10 9 90% CTCTACAATA- - intron - TCTATCTTTA 91

C. difficile CD3563 226s 8 8 100% GAGGGACAGG - intron - TTGCTGTAGC 92

C. botulinum CBO0780671a 4 4 100% TTTATTATTT- intron -TCTTTTTTAA ^d 93

C. botulinum CBOl 120 670s 4 4 100% GAATTTTATG- intron - CTAATATATC ^d 94

C. botulinum CBO27 ^' 62 1014s 4 4 100% TTTAACATAT- intron -AGATTAGTTA ^d 95

C. botulinum spoOA 249s 4 4 100% TATGCCAAGG - intτon - GTAATTGTTT ^d 96

^a Introns were inserted after the indicated number of bases from the start of the ORF, in either the sense (s) or antisense (a) orientation. ^b Genomic DNA was extracted from Em ^R clones picked at random and used as template in PCR using primers which amplify across an intron-exon junction. One clone of each desired mutant was selected and the intron insertion site verified by sequencing. °Ninety-six Em ^R C. difficile pyrF mutant candidate clones were screened in pools. Exhaustive screening was not required to isolate the mutant, so a range of possible frequencies is given. ^d Predicted site of insertion, not verified by nucleotide sequencing.

Table 13: Oligonucleotide primers

Oligonucleotide Sequence (5' -3 ^f ) SEQ ID No

Cdi-CD0552-75a-IBS AAAAAAGCTTATAATTATCCTTATTCTCCACAATAGTGCGCCCAGATAGGGTG 97

Cdi-CD0552-75a-EBS 1 d CAGATTGTACAAATGTGGTGATAACAGATAAGTCACAATATCTAACTTACCTTTCTTTGT 98

Cdi-CD0552-75a-EBS2 TGAACGCAAGTTTCTAATTTCGGTTGAGAATCGATAGAGGAAAGTGTCT 99

Cdi-CD3563-226s-IBS AAAAAAGCTTATAATTATCCTTAATGAGCGACAGGGTGCGCCCAGATAGGGTG 100

Cdi-CD3563-226s-EBS 1 d CAGATTGTACAAATGTGGTGATAACAGATAAGTCGACAGGTTTAACTTACCTTTCTTTGT 101

Cdi-CD3563-226s-EBS2 TGAACGCAAGTTTCTAATTTCGGTTCTCATCCGATAGAGGAAAGTGTCT 102

Cac-CAC0081-141a-IBS AAAAAAGCTTATAATTATCCTTAATTTTCAATGACGTGCGCCCAGATAGGGTG 103

Cac-CAC0081-141a-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCAATGACATTAACTTACCTTTCTTTGT 104

Cac-CAC0081-141a-EBS2 TGAACGCAAGTTTCTAATTTCGGTTAAAATCCGATAGAGGAAAGTGTCT 105

Cac-CAC0078-385a-IBS AAAAAAGCTTATAATTATCCTTACTGTACCTCCAGGTGCGCCCAGATAGGGTG 106

Cac-CAC0078-385a-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCCTCCAGGCTAACTTACCTTTCTTTGT 107

Cac-CAC0078-385a-EBS2 TGAACGCAAGTTTCTAATTTCGATTTACAGTCGATAGAGGAAAGTGTCT 108

Cac-CAC0080-121s-roS AAAAAAGCTTATAATTATCCTTAAACTGCAATTATGTGCGCCCAGATAGGGTG 109

Cac-CAC0080-121s-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCAATTATTTTAACTTACCTTTCTTTGT 110

Cac-CAC0080- 121 S-EBS2 TGAACGCAAGTTTCTAATTTCGGTTCAGTTCCGATAGAGGAAAGTGTCT 111

Cac-CAC2208-201s-IBS AAAAAAGCTTATAATTATCCTTACATGTCGAGTATGTGCGCCCAGATAGGGTG 112

Cac-CAC2208-201 s-EBS 1 d CAGATTGTACAAATGTGGTGATAACAGATAAGTCGAGTATTCTAACTTACCTTTCTTTGT 113

Cac-CAC2208-201s-EBS2 TGAACGCAAGTTTCTAATTTCGGTTACATGTCGATAGAGGAAAGTGTCT 114

Cdi-CD0153-784a-IBS AAAAAAGCTTATAATTATCCTTATACCACTAAGCCGTGCGCCCAGATAGGGTG 115

Cdi-CD0153-784a-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCTAAGCCCATAACTTACCTTTCTTTGT 116

Cdi-CD0153-784a-EBS2 TGAACGCAAGTTTCTAATTTCGGTTTGGTATCGATAGAGGAAAGTGTCT 117

Cbo-CBO0780-671a-IBSl AAAAAAGCTTATAATTATCCTTAGCTTTCTTATTTGTGCGCCCAGATAGGGTG 118

Cbo-CBO0780-671a-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCTTATTTTCTAACTTACCTTTCTTTGT 119

Cbo-CBO0780-671a-EBS2 TGAACGCAAGTTTCTAATTTCGATTAAAGCTCGATAGAGGAAAGTGTCT 120

Cbo-CBOl 120-67Os-IBSl AAAAAAGCTTATAATTATCCTTACTGAACTTTATGGTGCGCCCAGATAGGGTG 121

Cbo-CBOl 120-67Os-EBSId CAGATTGTACAAATGTGGTGATAACAGATAAGTCTTTATGCTTAACTTACCTTTCTTTGT 122

Cbo-CBOl 120-670s-EBS2 TGAACGC AAGTTTCTAATTTCGGTTTTCAGTCGATAGAGGAAAGTGTCT 123

Gbo~CBO27624014s-IBSl AAAAAAGCTTATAATTATCCTTAGATTTCACATATGTGCGCCCAGATAGGGTG 124

Cbo-CBO27624014s-EBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCACATATAGTAACTTACCTTTCTTTGT 125

Cbo-CBO2762-1014s-EBS2 TGAACGCAAGTTTCTAATTTCGATTAAATCTCGATAGAGGAAAGTGTCT ^" 126

Cbo-spoft4-249s-IBSl AAAAAAGCTTATAATTATCCTTACCTATCCCAAGGGTGCGCCCAGATAGGGTG 127

Cbo-spo ⁽ M~249s~BBSld CAGATTGTACAAATGTGGTGATAACAGATAAGTCCCAAGGGTTAACTTACCTTTCTTTGT 128

Cbo-Ji?o&4~249s-EBS2 TGAACGCAAGTTTCTAATTTCGGTTATAGGTCGATAGAGGAAAGTGTCT 129

Example 17: Construction of a catP-based RAM

A RAM containing an alternative modified selectable marker gene, namely catP, which confers resistance to chloramphenicol or thiamphenicol, 5 was constructed.

The catP ORF was PCR-amplified from pMTL5402F plasmid DNA template using primers linker-catP-F S'-ATACTCAGGCCTCAATTAAC- g, CCAAGAGA-TGCTGGTGCTTCTGGTGCTGGTATGGTATTTGAAAAAATTGATAAAAATAG TTGGAACAG-S' (SEQ IDNO. 130) and catP-Mluϊ-Rl S'-ATACGC-GTTTAACTATTTATCAATTCCTGCAATTCGTTTACAAAACGGC-S' (SEQ ID NO. 131),which added a 10 small part of the tcl intron and linker to the 5' of the catP ORF using a primer extension.

The PCR product was digested with Stul and MMI. A portion of ErmBtdRAM2 containing the thl promoter, linker and most of the id group I intron was excised as a Spel I Stul fragment from pCR2.1::ErmBtdRAM2. These two restriction fragments were ligated together into pCR2.1::BrmBtdRAM2 linearised with Spel and MM, yielding the plasmid pCR2.1::RAM-Cl, which contains the new RAM element RAM- 15 Cl.

The sequence immediately preceding the catP ORF in RAM-Cl is identical to the sequence immediately preceding the ermB ORF in ErmBtdRAM2, containing the thl promoter, linker and td group I intron. The entire RAM-Cl element is flanked by MhA sites to facilitate its sub-cloning into the MIiA site of the Ll .LtrB intron for use as a RAM.

The RAM-Cl or a derivative thereof may be used as the RAM element in a plasmid analogous to pMTL007 to select for retargeting events in Clostridia on the basis of acquisition of thiamphenicol or chloramphenicol resistance. It will be appreciated that the selectable marker that is required to maintain the plasmid in the host must confer resistance to a different agent from the resistance conferred by the RAM. Therefore, pMTL007 will be modified by replacement of its catP selectable marker with a different selectable marker, such as ermB, which is effective in Clostridia. A plasmid modified in this way may be used for retargeting Clostridia.

As described herein, the promoter operatively linked to the region encoding the selectable marker must be capable of causing expression of the selectable marker encoded by a single copy of the selectable marker gene in an amount sufficient for the selectable marker to alter the phenotype of the Clostridial cell such that it can be distinguished from the Clostridial cell lacking the selectable marker gene. If the thl promoter in the RAM-Cl element fails to fulfil this criterion, it may be replaced or modified using methods disclosed herein. Similarly, if the positioning of the td group I intron is inappropriate either to prevent expression of the selectable marker when it is present in the RAM, or to permit expression of the selectable marker when it has spliced out of the RAM, its position may be modified. The function of the elements of the RAM may be tested using the two- plasmid system developed Karberg et al (2001) (see Example 5). Ultimately, RAM-Cl, or a derivative thereof, will be used to generate retargeting mutants in Clostridia.