Use of a flexible cassette method to generate a double unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement

Tanya Parish1 and Neil G. Stoker1

Department of Infectious & Tropical Diseases, London School of Hygiene & Tropical Medicine, Keppel Street, London WC1E 7HT, UK1

Author for correspondence: Tanya Parish. Tel: +44 20 7927 2425. Fax: +44 20 7637 4314. e-mail: Tanya.Parish{at}lshtm.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Progress in the field of mycobacterial research has been hindered by the inability to readily generate defined mutant strains of the slow-growing mycobacteria to investigate the function of specific genes. An efficient method is described that has been used to generate several mutants, including the first double unmarked deletion strain of Mycobacterium tuberculosis. Four mutants were constructed: a marked deletion of the plcABC cluster, which encodes three phospholipases C; separate unmarked deletions in plcABC and tlyA (encoding a haemolysin); and a double unmarked mutant tlyA{Delta} plcABC{Delta}. To accomplish this, two series of vectors were designed, the first of which, named pNIL, allows manipulation of the target gene sequence at a variety of convenient restriction sites. The second series, named pGOAL, contains marker cassettes flanked by PacI restriction enzyme sites. The final suicide plasmid vectors were then obtained by cloning a marker cassette from a pGOAL vector into the single PacI site of the pNIL vector with the modified gene of interest. Finally, a two-step strategy was employed whereby single cross-over events were first selected, then screening for the second cross-over was carried out to yield the mutant strains. This technique will now allow the construction of potential vaccine strains without the inclusion of antibiotic resistance markers, the ability to make multiple defined mutations and the possibility of making more subtle defined mutations, such as point mutations.

Keywords: homologous recombination, haemolysin, phospholipase C, rapid cloning system

Abbreviations: hyg, hygromycin; kan, kanamycin; suc, sucrose; MCS, multiple cloning site


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Tuberculosis remains a serious public health problem in many parts of the world and is responsible for approximately 2 million deaths per year (Dye et al., 1999 ). New drugs and vaccines are required to control the disease. However, progress has until recently been hampered by the lack of suitable tools with which to manipulate the genome of the causative pathogen, Mycobacterium tuberculosis. In particular, it has not been possible to generate defined mutants of M. tuberculosis by homologous recombination routinely despite early successes (Balasubramanian et al., 1996 ; Berthet et al., 1998 ; Pelicic et al., 1997 ). By making such mutants, the function of individual genes may be analysed, revealing potential targets for novel chemotherapies. Furthermore, now that the completed genome sequence is available (Cole et al., 1998 ), rationally attenuated strains which are potential vaccine candidates can be constructed.

For these reasons, we have been developing an efficient system for introducing specific mutations in M. tuberculosis. We have previously reported the construction of five M. tuberculosis strains with defined mutations in genes encoding a haemolysin (tlyA) (Hinds et al., 1999 ) and enzymes from amino acid biosynthesis pathways (hisD, metB, proC and trpD) (Parish et al., 1999 ). Our method uses a suicide plasmid vector to deliver the recombination substrate to the recipient cell. The plasmids lack a mycobacterial origin of replication (oriM) and are thus unable to replicate in M. tuberculosis. Plasmid DNA is pretreated with UV light or alkali to stimulate homologous recombination and abolish illegitimate recombination in the recipient cells. Construction of the suicide delivery vector has often been problematic as several cloning steps are required to include the relevant markers and finding appropriate restriction sites for inserting these genes has become limiting. In this paper, we describe two series of vectors that we developed to overcome the cloning bottleneck. By separating the construction of the mutated version of the target gene from that of the inclusion of the required marker genes, vector construction is rapid and flexible.

The genes that we chose for this study encode a haemolysin (tlyA) and three phospholipases C found in a cluster (plcABC) on the M. tuberculosis chromosome. Homologues of these genes have been shown to be important for virulence in other organisms (Hyatt et al., 1994 ; Songer, 1997 ; Titball, 1993 ). To study the role of these genes in virulence, we selected them as candidates for the construction of isogenic mutant strains of M. tuberculosis H37Rv. We have demonstrated the utility of the pNIL and pGOAL cloning vectors in M. tuberculosis by generating both marked and unmarked deletions of the plcABC gene cluster and an unmarked deletion in the tlyA gene. The tlyA mutant was then used to generate a double unmarked mutant when the plcABC gene cluster was deleted in this strain.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and plasmids.
Plasmids and strains used in this study are listed in Tables 1 and 2. M. tuberculosis H37Rv was grown in Middlebrook 7H9 liquid medium or on 7H10 agar plates supplemented with 10% (v/v) OADC (Becton Dickinson) plus 0·05% (w/v) Tween 80 for liquid cultures. Kanamycin (kan) was used at 20 µg ml-1, hygromycin (hyg) at 100 µg ml-1, X-Gal at 50 µg ml-1 and sucrose (suc) at 2% (w/v).


View this table:
[in this window]
[in a new window]
 
Table 1. Plasmids used

 

View this table:
[in this window]
[in a new window]
 
Table 2. Marker gene cassettes available in the pGOAL series

 
Construction of cloning vectors.
DNA manipulations were carried out according to standard techniques (Sambrook et al., 1989 ). The pNIL vectors were derived from pJBS633 which contains no lac sequences (Broome-Smith & Spratt, 1986 ): the 1·9 kb EcoRV–PvuII fragment was deleted and the 450 bp EcoRI fragment, from Tropist4, containing two PacI sites, was inserted in either orientation. Finally, a HindIII deletion was carried out to give the vectors p1NIL and p2NIL (Fig. 1a). The pGOAL vectors were derived from pBR322 by first deleting the 1·9 kb EcoRV–PvuII fragment, then the 450 bp EcoRI fragment from Tropist4, containing two PacI sites, was inserted. Other marker genes were then introduced between the PacI sites. The Phsp60-sacB gene was derived from p6.11 (Selwyn Quan, Glaxo Wellcome Research and Development, Stevenage, UK), PAg85-lacZ from pEM37 (Edith Machowski, South African Institute for Medical Research, Johannesburg, South Africa), hygromycin resistance gene (hyg) from pAGAN40 (Hinds et al., 1999 ) and lacZ from pATB10 (Martin Everett, Glaxo Wellcome Research and Development, Stevenage, UK) (Fig. 1b and Table 2).



View larger version (24K):
[in this window]
[in a new window]
 
Fig. 1. General features of the pNIL and pGOAL vector series. (a) p1NIL and p2NIL cloning vectors. Each vector carries the kan gene for selection in both E. coli and mycobacteria, an oriE for replication in E. coli, the f1 origin for generation of phagemid ssDNA if required, a small MCS and a single PacI site. (b) pGOAL vector series. pGOAL vectors carry different combinations of marker genes flanked by two PacI sites. Each vector also carries an oriE and amp for selection in E. coli.

 
Construction of delivery vectors.
The cloning strategy for generating suicide delivery vectors is illustrated in Fig. 2. For the tlyA deletion, a 2·5 kb SmaI fragment from pCIG19 (Wren et al., 1998 ) was cloned into the PmlI site of p2NIL and a subsequent 207 bp PmlI–EcoRV deletion was made (Fig. 3a). The PacI cassette from pGOAL15 (hyg, lacZ, Phsp60-sacB) was then cloned into the single PacI site to generate the suicide delivery vector pTLY5. For the unmarked plcABC deletion, two separate fragments were cloned into p2NIL. First a 2·0 kb PstI fragment (Fig. 3b) was cloned into the PstI site of p2NIL followed by a 2·0 kb SmaI fragment (Fig. 3b) in the PmlI site. This construct has the two fragments separated by a small region of the cloning vector. The marked deletion was constructed by cloning the hyg gene into the ScaI site in the middle of the two fragments (pPLC8). The final delivery vectors were generated by adding the PacI cassette from pGOAL15 (hyg lacZ Phsp60-sacB) to the plcABC{Delta} construct to give pPLC9, the PacI cassette from pGOAL13 (lacZ Phsp60-sacB) to the plcABC{Delta}::hyg construct to give pPLC10 and the PacI cassette from pGOAL19 (hyg, PAg85-lacZ Phsp60-sacB) to the plcABC{Delta} construct to give pPLC11 (Fig. 2b).



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 2. Cloning strategy for generating suicide delivery vectors. (a) The target gene is cloned into one of the pNIL vectors and the required mutation generated. (b) The PacI cassette containing the desired marker genes is then excised from the appropriate pGOAL vector and cloned into the unique PacI site of the pNIL/mutated gene vector, resulting in the final suicide delivery vector. (c) The final vector thus contains oriE, the kanamycin resistance gene (kan) and the f1 origin (f1 ori).

 


View larger version (10K):
[in this window]
[in a new window]
 
Fig. 3. Delivery vectors used to generate M. tuberculosis mutants. (a) Unmarked tlyA mutation: a 207 bp PmlI–EcoRV deletion was made in the tlyA gene (black) and the hyg, lacZ and Phsp60-sacB cassette from pGOAL17 was added to generate pTLY5. (b) plcABC deletion mutations. The SmaI and PstI flanking regions of the plcABC gene cluster were cloned into p2NIL. The hyg, lacZ and Phsp60-sacB cassette was added to generate pPLC9 and the hyg, PAg85-lacZ, Phsp60-sacB cassette was added to generate pPLC11. The marked mutation was made by cloning the hyg gene into the middle of the two flanking regions and the lacZ, Phsp60-sacB cassette added to make the final delivery vector pPLC10. The region absent from the delivery vectors is shown in black. E, EcoRV; P, PmlI; Ps, PstI S, SmaI.

 
Isolation of mutants.
Vector DNAs were pretreated with 100 mJ UV light cm-2 (Hinds et al., 1999 ) and used to electroporate M. tuberculosis (Parish et al., 1999 ). hygR kanR transformants (single cross-overs) were streaked out onto plates containing either hygromycin (for marked mutations) or plates without selection (for unmarked mutations). A loopful of cells was then resuspended in liquid medium by rigorous vortexing with 1 mm glass beads and serial dilutions plated onto sucrose plates (plus hygromycin and X-Gal where appropriate). Plates were incubated for 3–4 weeks. sucR colonies were then streaked onto plates with and without kanamycin (plus hygromycin where appropriate) to identify kanS colonies.

Genotypic analysis.
DNA was prepared from kanS colonies by the methods of Belisle or Santos (Belisle & Sonnenberg, 1998 ; Santos et al., 1992 ). Both Southern blotting and PCR analyses were carried out to confirm the expected genotype. For the tlyA{Delta} mutant, genomic DNA was digested with XhoI and hybridized to a probe corresponding to the entire tlyA gene. Strains carrying the deletion lost the wild-type bands of 3·3 and 0·8 kb and gained a band of 3·9 kb, reflecting the loss of an XhoI site. In addition, PCR using primers which flanked the deletion site was carried out, producing a 448 bp PCR fragment in the wild-type and a 241 bp fragment in the deletion strains. Genomic DNA from the plcABC mutant strains was digested with EcoRI and hybridized to the PstI fragment shown in Fig. 3(b) (partial plcA gene and flanking region): the wild-type possessed a 3·2 kb band, the deletion strain had a 5·7 kb band and the marked mutation had a 4 kb band. PCR analysis was also carried out using several sets of primers: set one was designed to amplify a 0·7 kb fragment from the wild-type which was absent from both mutant strains, set two amplified a 1·4 kb fragment from the unmarked deletion strain only and set 3 amplified a 1·1 kb fragment from the marked deletion strain only.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
A rapid cloning strategy
We have been developing an efficient system for introducing defined mutations in M. tuberculosis, but have found that the construction of the suicide delivery vectors has been problematic. This is firstly due to the lack of suitable restriction enzyme sites when multiple markers are already present, making gene cloning and manipulation difficult. Second, we often wish to change the marker genes used and this can entail a completely new cloning exercise. To overcome this bottleneck we have developed a flexible cloning strategy based on two different vector series to separate out the cloning and manipulation of the target gene from the inclusion of the required marker genes.

The first series of plasmids (pNIL) is for manipulating the gene of interest (Fig. 1a). The plasmids consist of a simple cloning vector with an origin of replication for Escherichia coli (oriE), a kanamycin resistance gene (kan) and multiple cloning sites (MCSs) with different restriction enzyme sites, but including a single PacI restriction enzyme site. The second series (pGOAL) is used to generate and store a number of cassettes of marker genes (Fig. 1b and Table 2). These consist of vectors containing oriE, an ampicillin resistance gene (amp) and different combinations of marker genes (hyg, lacZ and sacB) flanked by two PacI sites. Thus the marker genes can be excised as a PacI cassette and inserted into any of the pNIL series in a one-step cloning process. The benefit of using PacI sites is that none is present in the M. tuberculosis genome (Cole et al., 1998 ) or in the marker genes used in this study.

Strategy for construction of delivery vectors
The general strategy for constructing suicide delivery vectors is shown in Fig. 2. One approach is for the gene of interest to be cloned into a pNIL vector, the choice depending on which restriction sites are suitable. The required mutation is then generated, either by insertion of an antibiotic resistance gene, such as hyg, or by constructing an unmarked mutation (e.g. pTLY5; Fig. 3a). Alternatively, two separate gene fragments and flanking DNA can be brought together, effectively resulting in a deletion (e.g. pPLC10; Fig. 3b). The other marker genes such as lacZ, sacB and hyg are then cloned in as a PacI cassette from the appropriate pGOAL vector (Table 2).

Construction of M. tuberculosis plcABC::hyg, plcABC{Delta} and tlyA{Delta} mutants
To test the vectors out, constructs were made carrying marked and unmarked plcABC deletions (pPLC10 and pPLC9, respectively) and an unmarked tlyA deletion (pTLY5) (Fig. 3). Mutants were then generated by a two-step process for both marked and unmarked mutants. Single cross-overs were initially obtained and the second cross-over event (to generate double cross-over mutants) was subsequently selected and/or screened for by using the lacZ, sacB and kan genes. Vectors were pretreated with UV light and electroporated into M. tuberculosis and transformants were selected on hygromycin/kanamycin/X-Gal plates. Ten transformants were isolated after electroporation with pTLY5 and 50 each for pPLC9 and pPLC10. Unexpectedly, all the colonies were white, suggesting that the lacZ gene was not being expressed in M. tuberculosis. Southern blotting confirmed that representative colonies were single cross-overs.

The purpose of the lacZ gene had been to distinguish single cross-overs from double cross-overs in the second stage; in the absence of adequate expression, we used kanamycin sensitivity as a screen. One single cross-over transformant was picked from each plate and streaked out onto plates lacking antibiotics. Following growth, a loopful of cells was resuspended in liquid medium and plated onto sucrose plates to select for cells which had lost the integrated plasmid through a second cross-over. A reduction in c.f.u. of approximately 104 was seen on plates containing sucrose as compared to plates containing no sucrose. sucR colonies were streaked onto plates with and without kanamycin and scored for growth after 2–3 weeks to distinguish double cross-overs from single cross-overs which had acquired spontaneous resistance to sucrose. DNA was prepared from kanS colonies and analysed by Southern blotting and PCR.

For the marked plcABC::hyg mutation, colonies which have the phenotype hygR kanS sucR should all be double cross-over mutants; 12/45 sucR colonies were kanS – six of these were analysed by Southern blotting and PCR and all had the expected mutant double cross-over genotype.

For the unmarked mutations, a second cross-over can result in either the wild-type allele or the mutant allele remaining in the chromosome, depending on where the second event occurs. Thus, kanS hygS sucR colonies isolated can carry either the wild-type or mutant allele. The relative frequency of the two will be dependent on the relative frequency of recombination on each side of the gene and the relative fitness of the mutant and wild-type strains. For tlyA{Delta}, 12/48 sucR colonies were hygS kanS – 6/6 analysed by Southern blotting and PCR were all mutant double cross-overs. For plcABC{Delta} 20/40 sucR colonies were hygS kanS – 1/8 was the mutant double cross-over, with the remainder being wild-type.

Second generation pGOAL vectors
Since the lacZ gene was not expressed in M. tuberculosis in the existing pGOAL cassettes, we used the mycobacterial antigen 85A promoter to drive expression of this gene. We replaced the lacZ gene in pGOAL13 and pGOAL15 with the PAg85-lacZ gene fusion from pEM37 to create pGOAL17 and pGOAL19, respectively.

Construction of an unmarked M. tuberculosis tlyA{Delta} plcABC{Delta} mutant
One of the benefits of using unmarked mutations is that multiple mutations can be made in one strain by sequential mutagenesis using the same selection system. To show this to be a practical proposition, we used the improved pGOAL vectors to introduce a plcABC mutation into the unmarked tlyA mutant described above.

The PacI cassette in pPLC9 was replaced by the hyg, PAg85-lacZ, Phsp60-sacB cassette from pGOAL19 to make pPLC11. UV-pretreated DNA was used to transform the M. tuberculosis tlyA{Delta} mutant. The hygR kanR single cross-over transformants obtained were blue on X-Gal plates, confirming that lacZ was being expressed from the Ag85 promoter. A single cross-over colony was streaked out and a loopful of cells resuspended in medium and plated onto sucrose X-Gal plates. Approximately 10% of the sucR colonies were blue (and were therefore spontaneous sucrose mutants) and 90% were white. These white colonies should all be double cross-overs and this was verified by patch testing them for kanS 40/40 were kanS, confirming plasmid loss. kanS colonies were analysed by Southern blotting and PCR and 4/18 were shown to be mutant double cross-overs. Thus, this mutagenesis strategy had successfully generated a double unmarked tlyA{Delta} plcABC{Delta} mutant.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
An efficient, reliable system for making mutants
The system that we have used to generate mutants is designed to be highly efficient. The use of the two plasmid cloning systems allows the easy construction of the suicide delivery vectors. A two-step strategy means that only one single cross-over recombinant is required from the electroporation and the inclusion of multiple markers allows for the selection or screening for cells in which the second cross-over has occurred. This is particularly important where the mutant does not have a screenable phenotype. Recently, the generation of an unmarked mutation of lysA has been reported for M. tuberculosis (Pavelka & Jacobs, 1999 ). This strain was generated using a two-step strategy, with the double cross-over mutant strain being identified by virtue of its auxotrophy. However, this method is not applicable where the mutant has no obvious phenotype, whereas the use of multiple markers in the delivery vector enables screening for the second cross-over event.

We have now isolated mutant strains using a range of delivery vectors, all of which have been effective. The use of a system such as this, together with the efficiency and reliability of the technique should greatly facilitate the rapid generation of new mutant strains.

Flexibility of the two-step system
The cloning system described here is extremely flexible. Marker cassettes can be easily switched in the final delivery suicide vector if different combinations are required. In addition, new marker genes can be readily incorporated into the pGOAL PacI cassettes. Although these vectors were created specifically for mycobacteria, the general approach should be applicable to all bacterial species. The important features of the two-vector system are the flanking of marker genes by a restriction site not found in the genome of the target organism, a simple MCS and the use of different antibiotic resistance markers in the two separate vector series.

Advantage of the two-step strategy
The use of a two-step strategy enables the isolation of unmarked mutants, which cannot be isolated using a one-step strategy. This has allowed us to make deletions of several genes, plcABC and tlyA, separately and in combination. A one-step strategy can be used to generate marked mutations, but in some cases the required mutants will not be obtained, for example if the gene is essential or the recombination frequency at that locus is low. In these cases a two-step strategy will be required, so it may be easier to start with this strategy, regardless. In addition, the two-step method for generating marked mutants is simpler, since all the double cross-overs isolated should have the mutant allele.

As we are using a two-step process only one single cross-over transformant is required. Thus, the use of larger vectors (containing several marker genes) which will reduce transformation efficiency does not pose a problem.

We have previously shown that the lacZ gene is a useful marker for distinguishing single cross-overs from spontaneous hygromycin resistance in M. tuberculosis (Parish et al., 1999 ). Alternatively, using two antibiotics (kanamycin and hygromycin), as with the initial vectors (pTLY5, pPLC9 and pPLC10), abolished the problem of spontaneous antibiotic resistance. In either case, all the transformants that were recovered and analysed were single cross-overs.

The use of both lacZ and sacB together reduced the number of colonies to be tested for double cross-overs. When used individually there are problems with both of these markers. Although the sacB gene has previously been used for negative selection in mycobacteria (Pelicic et al., 1996 ), problems have been reported with the high frequency of spontaneous sucrose resistance (Papavinasasundaram et al., 1998 ; Pavelka & Jacobs, 1999 ; Pedulla & Hatfull, 1998 ). Our results show that 10–90% of sucR colonies are not double cross-overs, so a second selection or screening marker is desirable. In our early experiments where lacZ was not expressed, we used the loss of the kan gene as a second quick screen. The later vectors with PAg85-lacZ added another way of immediately identifying those cells which were still single cross-overs. This minimizes the number of colonies to be analysed by Southern blotting and significantly reduces the overall workload. The expression of lacZ can be variable and some colonies will be pale blue, so again it cannot be relied on alone as an indication of double cross-overs. In addition, it cannot be used to select against single cross-overs and so needs to be used in conjunction with a negative selection marker. The kan gene provides a further way of rapidly screening out single cross-overs and any mixed single cross-over/double cross-over colonies.

Importance of unmarked mutations
The ability to generate unmarked mutations is of great importance in the drive towards generating new potential vaccine strains, where antibiotic resistance markers cannot be left in the chromosome. Unmarked mutations have several advantages. Secondary mutations can be made without the problem of finding another resistance marker, as we have shown by making a double mutant (plcABC{Delta} tlyA{Delta}). In addition, more subtle mutations can be made, such as point mutations to study gene function.

Mutant phenotypes
We previously described TlyA from M. tuberculosis (Wren et al., 1998 ), which shows homology to a pore-forming haemolysin of Serpulina hyodysenteriae, which has been shown to play a role in virulence of the latter organism (Hyatt et al., 1994 ). Three genes have been identified in the M. tuberculosis H37Rv genome which encode phospholipases C (Cole et al., 1998 ) and haemolytic phospholipase C activity has been demonstrated for M. tuberculosis (Johansen et al., 1996 ). Phospholipases C have been shown to be virulence factors for other pathogens (Songer, 1997 ; Titball, 1993 ). For example, in Listeria monocytogenes, PlcCs, together with a haemolysin (listeriolysin), are required for the escape of the organism from the phagosome and extracellular spread. In Clostridium perfringens, phospholipase C mutants show reduced virulence. Pseudomonas aeruginosa also possesses two phospholipases C which show homology to two of the three M. tuberculosis plc genes. We are currently investigating the virulence of the mutant strains in both animal and macrophage model systems.

Conclusions
We have developed a rapid cloning strategy to facilitate the construction of delivery vectors and have used these vectors to generate four mutant strains of M. tuberculosis, including a double unmarked mutant strain. The mutant strains generated in this study are now being analysed phenotypically in virulence studies.


   ACKNOWLEDGEMENTS
 
T.P. was funded by the Glaxo Wellcome Action TB Initiative. We thank Valerie Mizrahi, Bhavna Gordhan, Ruth McAdam and Jason Hinds for helpful discussions; Martin Everett, Selwyn Quan and Edith Machowski for providing plasmid vectors; and Ken Duncan for critically reading the manuscript.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Balasubramanian, V., Pavelka, M. S., Bardarov, S., Martin, J., Weisbrod, T., McAdam, R. A., Bloom, B. & Jacobs, W. R. (1996). Allelic exchange in Mycobacterium tuberculosis with long linear recombination substrates.J Bacteriol178, 273-279.[Abstract]

Belisle, J. T. & Sonnenberg, M. G. (1998). Isolation of genomic DNA from mycobacteria. In Mycobacteria Protocols, pp. 31-44. Edited by T. Parish & N. G. Stoker. Totowa, NJ: Humana Press.

Berthet, F. X., Lagranderie, M., Gounon, P. & 9 other authors (1998). Attenuation of virulence by disruption of the Mycobacterium tuberculosis erp gene. Science 282, 759–762.[Abstract/Free Full Text]

Broome-Smith, J. K. & Spratt, B. G. (1986). A vector for the construction of translational fusions to TEM ß-lactamase and the analysis of protein export signals and membrane protein topology.Gene49, 341-349.[Medline]

Cole, S. T., Brosch, R., Parkhill, J. & 39 other authors (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544.[Medline]

De Smet, K. A. L., Jamil, S. & Stoker, N. G. (1993). Tropist3: A cosmid vector for simplified mapping of both G+C-rich and A+T-rich genomic DNA.Gene136, 215-219.[Medline]

Dye, C., Scheele, S., Dolin, P., Pathania, V. & Raviglione, M. C. (1999). Global burden of tuberculosis. Estimated incidence, prevalence, and mortality by country.J Am Med Assoc282, 677-686.[Abstract/Free Full Text]

Hinds, J., Mahenthiralingam, E., Kempsell, K. E., Duncan, K., Stokes, R. W., Parish, T. & Stoker, N. G. (1999). Enhanced gene replacement in mycobacteria.Microbiology145, 519-527.[Abstract]

Hyatt, D. R., ter Huurne, A. A., van der Zeijst, B. A. & Joens, L. A. (1994). Reduced virulence of Serpulina hyodysenteriae hemolysin-negative mutants in pigs and their potential to protect pigs against challenge with a virulent strain.Infect Immun62, 2244-2248.[Abstract]

Johansen, K. A., Gill, R. E. & Vasin, M. L. (1996). Biochemical and molecular analysis of phospholipase C and phospholipase D activity in mycobacteria.Infect Immun64, 3259-3266.[Abstract]

Papavinasasundaram, K. G., Colston, M. J. & Davis, E. O. (1998). Construction and complementation of a recA deletion mutant of Mycobacterium smegmatis reveals that the intein in Mycobacterium tuberculosis recA does not affect RecA function.Mol Microbiol30, 525-534.[Medline]

Parish, T., Gordhan, B. G., McAdam, R. A., Duncan, K., Mizrahi, V. & Stoker, N. G. (1999). Production of mutants in amino acid biosynthesis genes of Mycobacterium tuberculosis by homologous recombination.Microbiology145, 3497-3503.[Abstract/Free Full Text]

Pavelka, M. S. & Jacobs, W. R. (1999). Comparison of the construction of unmarked deletion mutations in Mycobacterium smegmatis, Mycobacterium bovis Bacillus Calmette-Guerin, and Mycobacterium tuberculosis H37Rv by allelic exchange.J Bacteriol181, 4780-4789.[Abstract/Free Full Text]

Pedulla, M. L. & Hatfull, G. F. (1998). Characterization of the mIHF gene of Mycobacterium smegmatis.J Bacteriol180, 5473-5477.[Abstract/Free Full Text]

Pelicic, V., Reyrat, J. M. & Gicquel, B. (1996). Positive selection of allelic exchange mutants in Mycobacterium bovis BCG.FEMS Microbiol Lett144, 161-166.[Medline]

Pelicic, V., Jackson, M., Reyrat, J. M., Jacobs, W. R., Gicquel, B. & Guilhot, C. (1997). Efficient allelic exchange and transposon mutagenesis in Mycobacterium tuberculosis.Proc Natl Acad Sci USA94, 10955-10960.[Abstract/Free Full Text]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Santos, A. R., Demiranda, A. B., Lima, L. M., Suffys, P. N. & Degrave, W. M. (1992). Method for high yield preparation in large and small scale of nucleic acids from mycobacteria.J Microbiol Methods15, 83-94.

Songer, J. G. (1997). Bacterial phospholipases and their role in virulence.Trends Microbiol5, 156-161.[Medline]

Titball, R. W. (1993). Bacterial phospholipases C.Microbiol Rev57, 347-366.[Abstract]

Wren, B. W., Stabler, R. A., Das, S. S., Butcher, P. D., Mangan, J. A., Clarke, J. D., Casali, N., Parish, T. & Stoker, N. G. (1998). Characterization of a haemolysin from Mycobacterium tuberculosis with homology to a virulence factor of Serpulina hyodysenteriae.Microbiology144, 1205-1211.[Abstract]

Received 15 February 2000; revised 18 April 2000; accepted 28 April 2000.