APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Aug. 2010, p. 5312–5316 0099-2240/10/$12.00 doi:10.1128/AEM.00382-10 Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Vol. 76, No. 15
Xer Site-Specific Recombination, an Efficient Tool To Introduce Unmarked Deletions into Mycobacteria䌤† Alessandro Cascioferro,1‡ Francesca Boldrin,1 Agnese Serafini,1 Roberta Provvedi,2 Giorgio Palu `,1 and Riccardo Manganelli1* Department of Histology, Microbiology, and Medical Biotechnologies1 and Department of Biology,2 University of Padua, Padua, Italy Received 11 February 2010/Accepted 1 June 2010
Genetic manipulation of mycobacteria still represents a serious challenge due to the lack of tools and selection markers. In this report, we describe the development of an intrinsically unstable excisable cassette for introduction of unmarked mutations in both Mycobacterium smegmatis and Mycobacterium tuberculosis. Recently, a new sequence-specific recombinase system based on the endogenous Xer recombinases (Xer-cise) was shown to be amenable for genetic manipulation and construction of unmarked deletion mutants in Escherichia coli and Bacillus subtilis (4). In this system, the antibiotic resistance cassette, flanked by dif sites, is intrinsically unstable since the endogenous recombinases XerC and XerD recognize and resolve the dif sites that border the cassette. This method, relying on endogenous recombinases, does not require the introduction and the subsequent removal of replicating plasmids carrying exogenous genes, making it extremely simple and practical. E. coli XerC and XerD recombinases are essential for chromosome segregation during cell division, as their role is to resolve chromosome dimers to monomers recognizing the 28-bp dif sequence present at the replication terminus region (8). The Xer site-specific recombination system is very well conserved in prokaryotes with circular chromosomes, and homologues of XerC and XerD have been identified among Gram-negative and Gram-positive bacteria (16). In this report, we adapted the Xer-cise technique to mycobacteria, demonstrating that it can be employed as a practical and efficient genetic tool for manipulating both M. tuberculosis and Mycobacterium smegmatis. Mycobacterial dif sites. In a recent paper, Hendrickson and Lawrence (8) identified the putative dif sites in several eubacterial species. Comparing the dif sequence of Mycobacterium avium reported by these authors to the chromosomal sequences of M. tuberculosis and M. smegmatis, we were able to identify the putative dif sequences in these two organisms. As shown in Fig. 1, the M. tuberculosis sequence differs by 3 nucleotides from that of M. avium and by 6 nucleotides from that of M. smegmatis. Construction of an excisable Hyg resistance cassette. A 1.7-kb hygromycin (Hyg) cassette was PCR amplified from pIJ963 (10) using primers containing the M. tuberculosis putative dif sequence and a BglII restriction site at their 5⬘ termini. The resulting fragment was subsequently cloned in pCR-BluntII-TOPO (Invitrogen). In order to evaluate the potential for excision of this cassette from the mycobacterial chromosome, this cassette was subcloned in the single BclI site of the integrative plasmid pSM316, a pYUB178 (2) derivative based on the L5 mycobacteriophage (unpublished), to obtain pAL70 (see Fig. S1 in the supplemental material), which was
Mycobacterium tuberculosis causes about 2 million deaths worldwide every year (15). Over the last few years, M. tuberculosis pathogenesis characterization at a molecular level required the development of efficient genetic tools for recombination and mutagenesis. The employment of replicating temperature-sensitive and suicide plasmids (14), specialized transducing mycobacteriophages (1, 9), and a recombineering system based on two exogenous recombinases (24) improved the ability to obtain mycobacterial mutants by homologous recombination. However, the availability of only few selection markers represents a real problem when multiple knockouts are required for the study of redundant gene families in mycobacteria. A way to circumvent this problem is by the production of unmarked mutations, which can be obtained by homologous recombination and selection for sequential crossing-over events using both positive and negative markers for counterselection of the different allelic exchange events (13), or by a different approach relying on sequence-specific recombination systems allowing the excision of the positive selection marker after it has been used to select for the recombination event (19). Three different sequence-specific recombinase systems have been successfully used with mycobacteria: the TnpR/res system of the ␥␦ transposon (1), the Flp/FRT system of Saccharomyces cerevisiae (18, 20), and the LoxP/cre systems from bacteriophage P1 (11, 19). While the Flp/FRT and the Lox/cre systems were shown to work efficiently in both slow- and fast-growing mycobacteria, the Tnp/res system was proved to be efficient only in fast-growing species. All of these systems require a first step during which the expression of an exogenous resolvase or recombinase from a replicative plasmid allows the excision of the resistant marker and a second step to eliminate the replicative plasmid, making the procedure very time-consuming, particularly when working with slow-growing mycobacteria.
* Corresponding author. Mailing address: Department of Histology, Microbiology, and Medical Biotechnologies, University of Padova, Via Gabelli 63, 35121 Padova, Italy. Phone: 39 049 827 2366. Fax: 39 049 827 2355. E-mail:
[email protected]. ‡ Present address: Institut Pasteur, Integrated Mycobacterial Pathogenomics, Paris, France. † Supplemental material for this article may be found at http://aem .asm.org/. 䌤 Published ahead of print on 11 June 2010. 5312
VOL. 76, 2010
FIG. 1. Putative dif site sequences of M. avium, M. tuberculosis, and M. smegmatis. Dashes represent nucleotides identical to those of the M. avium sequence.
finally introduced in M. tuberculosis H37Rv and M. smegmatis mc2155 by electroporation. The resulting strains were named TB44 and MS100, respectively. Since Xer-mediated excision could result in a significant reduction in transformation frequency, we compared the electroporation frequency of pAL70 to that of pYUB413, a similar integrative plasmid encoding Hyg resistance without dif flanking sites (12). The electroporation frequencies of the two plasmids were comparable in both M. smegmatis and M. tuberculosis (for M. smegmatis pYUB413, 2.9 ⫻ 104 ⫾ 0.5 ⫻ 104 transformants/g; for M. smegmatis pAL70, 3.0 ⫻ 104 ⫾ 0.8 ⫻ 104 transformants/g; for M. tuberculosis pYUB413, 1.7 ⫻ 104 ⫾ 0.6 ⫻ 104 transformants/g; for M. tuberculosis pAL70, 1.6 ⫻ 104 ⫾ 0.7 ⫻ 104 transformants/g). However, pAL70contaning M. tuberculosis gave smaller colonies. Excision frequency at dif sites. TB44 and MS100 were inoculated in Middlebrook 7H9 without selection and grown up to the end of the exponential phase. Cultures were then diluted, and the growth step was repeated. After each passage, bacteria were plated on Middlebrook 7H10, and 50 single colonies were replica plated in the presence or absence of Hyg. As shown in Table 1, excision frequencies were very high in both species: in M. smegmatis after 4 rounds of cell division (first passage), 77% of the cells had lost the Hyg cassette (frequency of excision/ generation, 0.18), while after a few more generations, almost all cells become Hyg sensitive. Interestingly, in M. tuberculosis the excision of the Hyg cassette was even more efficient, with 100% of the cells losing the resistance cassette after only 3 rounds of division (first passage) (frequency of excision/generation, ⱖ0.33) (Table 1). These data demonstrate that M. smegmatis Xer recombinases are able to recognize the M. tuberculosis dif sequence (even with an efficiency lower than that of M. tuberculosis Xer recombinases), despite the 6 mismatches between the dif sites in the two species (Fig. 1). Four Hyg-sensitive strains derived from TB44 or MS100 were screened by PCR with primers flanking the excisable cassette to ensure that this phenotype was due to excision of the Hyg cassette and not to any other mutation. As shown in Fig. 2, all PCRs produced a fragment whose size was compatible with that expected from the excision (520 bp). The PCR fragments were then sequenced, confirming that precise excision of the cassette was obtained (data not shown). Construction of an unmarked mutation in M. smegmatis. To demonstrate the feasibility of introducing unmarked mutations on the mycobacterial chromosome by Xer-cise, we decided to introduce a 247-bp deletion in MSMEG_2694, encoding the transcriptional regulator ClgR (3). Two DNA fragments, comprising 456 and 565 bp (including the first 59 bp and the last 33 bp of the gene, respectively), were amplified using primer couples RP893-RP894 and RP895-RP896, respectively (see Table S1 in the supplemental material), and cloned at the borders of the Hyg excisable cassette. The resulting DNA
Xer-CISE IN MYCOBACTERIA
5313
fragment was purified and introduced by recombineering in an M. smegmatis mc2155 derivative containing pJV53, a replicative plasmid expressing two phage recombinases and conferring kanamycin (Km) resistance (1, 23). Two Hyg-resistant colonies were isolated and tested by PCR for the presence of the correct integration of the excisable cassette into the chromosome (not shown). Subsequently, they were grown for 4 generations in the absence of Hyg and Km to allow excision of the Hyg cassette and loss of pJV53. Hyg- and Km-sensitive colonies were finally recovered at the expected frequency. One of them was analyzed in parallel with a colony of the wild-type (wt) parental strain by PCR with primers flanking the region used for the recombination. As shown in Fig. 3A, the sizes of the fragments amplified from the two colonies were compatible with their expected sizes (1,295 bp for the mutant and 1,542 for the wt). The fragment amplified from the mutant bacteria was sequenced to verify the correctness of the excision: as expected, a copy of the dif sequence flanked by two BglII restriction sites was found in the correct position, replacing the 247 central nucleotides of the target gene (Fig. 3C). Construction of an excisable cassette for creation of inframe deletions. An important feature of an excisable cassette is that it can allow the creation of in-frame deletions. For this purpose, we analyzed the sequence of M. tuberculosis dif sites to detect open reading frames and stop codons. As shown in Fig. 4, open reading frames covering the entire length of the sequence are present in both directions, making it possible to design cloning strategies to allow the creation of in-frame deletions. In-frame deletions are useful for inactivating one gene laying inside an operon without affecting the transcription and translation of the downstream genes. However, in the cases of mutants obtained by one-step double crossover, using transducing phages or the recombineering technique (1, 23), the translation and transcription of the downstream genes are restored only after the excision of the cassette, making it impossible to select the mutant if these genes are essential. To solve this problem, we constructed two plasmids (pAL74 and pAL75), each containing an excisable cassette in which either the medium-strength or the weak mycobacterial promoters Pmpt64 and PRv1818c, originating from pMV2-33 (6) and pMV4-36 (7), respectively, were cloned at the 3⬘ end of the Hyg cassette, followed by an open reading frame including the
TABLE 1. Frequencies of gene excision by Xer site-specific recombination at dif sites in M. smegmatis and M. tuberculosis Valuea for: Characteristic
No. of generationsb % recombinantsc
M. smegmatis
M. tuberculosis
1st passage
2nd passage
1st passage
2nd passage
4.3 ⫾ 03 76.7 ⫾ 1.9
3.6 ⫾ 02 94.7 ⫾ 0.9
3.0 ⫾ 01 100
3.9 ⫾ 07 100
Results are means ⫾ standard deviations for three different experiments. Calculated as (lnNt ⫺ lnN0)/ln2, where Nt represents the final optical density and N0 the optical density of the inoculum (absorbance at 600 nm) at each passage point. c Percentage of cells that have undergone the recombination event to excise the Hyg cassette. a b
5314
CASCIOFERRO ET AL.
APPL. ENVIRON. MICROBIOL.
FIG. 2. Excision of the dif-hyg-dif cassette from the M. smegmatis MS100 and M. tuberculosis TB44 chromosomes. The chromosomes of 4 Hyg-sensitive MS100 derivatives and 4 TB44 Hyg-sensitive derivatives were analyzed by PCR with primers flanking the excisable cassette. The amplification of a 520-bp band indicated the excision of the cassette. (A) Lane 1, 1-kb ladder (New England); lanes 2 to 5, PCR products from TB44 derivatives; lane 6, M. tuberculosis H37Rv. (B) Lanes 1 to 4, PCR products from MS100 derivatives; lane 5, M. smegmatis mc2155; lane 6, marker VIII (Roche).
dif site sequence (see Fig. S2 in the supplemental material; also Fig. 5A and B). When this open reading frame is cloned in frame with the terminal portion of the gene subjected to inframe deletion, once the cassette is integrated into the chromosome using transducing phages or recombineering, the downstream genes will be correctly transcribed and translated even if translationally coupled with the previous gene. This allows the recovery of the mutant even in cases where these genes are essential (see Fig. S3 in the supplemental material). Similar mutants might be obtained (i) by constructing a partial diploid strain in which the entire operon missing the gene of interest is introduced at the L5 attB site and then the wt operon is deleted, but this is more time-consuming both for the complex cloning steps involving mutagenesis of large fragments of DNA and for the requirement of two rounds of electroporation, and (ii) by using single-stranded DNA recombineering, but the efficiency of this technique is low, especially in slow-
growing mycobacteria, and this technique requires very timeconsuming work for selection of the mutant strains (22). In order to verify the functionality of these constructs, we cloned these cassettes into the integrative plasmid pMV306 (21), upstream of a promoterless green fluorescent protein (GFP)-encoding gene originating from pMV10-25 (7) (see Fig. S4 in the supplemental material). The constructs were designed to obtain a translational fusion between the open reading frame downstream of Pmpt64 or PRv1818c and the GFPcoding sequence. The resulting integrative plasmids, pDario1 and pDario2, were introduced into M. smegmatis mc2155 by electroporation, and the two resulting recombinant strains were named MS140 and MS141, respectively. These two strains were cultured overnight in 7H9 without selection to allow the excision of the Hyg cassette and plated on 7H10 without selection. Finally, Hyg-sensitive derivatives of these two strains were selected and named MS143 and MS144, re-
FIG. 3. Analysis of mycobacterial mutants. Introduction of a 247-bp deletion in MSMEG_2694. (A) Agarose gel electrophoresis of PCR fragments obtained with primers flanking the regions used for the recombination. Lane 1, 1-kb ladder (New England); lane 2, mutant strain; lane 3, no-DNA control; lane 4, mc2155. (C) Partial sequence of the PCR product shown in lane 2. The first 59 bp of MSMEG_2964 is followed by the dif sequence (in bold) flanked by BglII restriction sites (italicized and underlined) and the last 33 bp of MSMEG_2964. Introduction of a 123-bp deletion in Rv3790. (B) Agarose gel electrophoresis of PCR fragments obtained with primers flanking the regions used for the recombination. Lane 1, merodiploid mutant strain; lane 2, H37Rv; lane 3, 1-kb ladder (New England); lane 4, no-DNA control. (D) Partial sequence of the lower PCR product shown in lane 1. The dif sequence (in bold) flanked by BglII restriction sites (italicized and underlined) was replaced 123 bp of Rv2790.
VOL. 76, 2010
FIG. 4. Sequence of the M. tuberculosis putative dif site in the positive strand (A) and in the negative strand (B). Amino acids encoded by the three reading frames are shown in capital letters. Termination codons are indicated by boxes.
spectively. Fluorescence microscopy showed that while MS140 and MS141 were fluorescent, their Hyg-sensitive derivatives MS143 and MS144 were not (see Fig. S5 in the supplemental material). To confirm these data, the same amounts of proteins from whole-cell lysates of the four M. smegmatis strains were analyzed by Western blotting using a monoclonal anti-GFP antibody. As shown in Fig. 5C, the Hyg-resistant strains MS140 and MS141 expressed GFP, while no GFP expression was observed for MS143 and MS144, confirming the functionality of the translational fusions obtained with the Hyg cassettes containing the two promoters and the loss of transcription of the GFP-coding gene after the excision of the Hyg cassette. The lower band visible in the gel probably represents a degradation product of GFP and is often detected when this protein is expressed in mycobacteria (unpublished observation). Construction of an in-frame deletion in M. tuberculosis. To show the possibility of creating an in-frame deletion in M.
Xer-CISE IN MYCOBACTERIA
5315
tuberculosis, we decided to introduce a 123-bp deletion in Rv3790, the first gene of an operon whose genes are all predicted to be essential (17). Two DNA fragments internal to Rv3790, comprising 506 bp (the 5⬘ fragment [5⬘-F]) and 519 bp (the 3⬘ fragment [3⬘-F]), were amplified using primer couples RP818-RP819 and RP820-RP821, respectively (see Table S1 in the supplemental material), and cloned at the borders of the Hyg excisable cassette of pAL74. The cloning strategy was designed to obtain an in-frame translational fusion between the 5⬘-F and the dif sequence at the 5⬘ end of the Hyg cassette as well as between the open reading frame downstream of Pmpt64 (including the 3⬘ dif sequence of the Hyg cassette) and the 3⬘-F in accordance with the scheme shown in Fig. S3 in the supplemental material. The resulting DNA fragment was purified and introduced by recombineering in a pJV53-containing merodiploid M. tuberculosis strain in which a second copy of Rv3790 was integrated at the att site and expressed from Pptr (5). Hyg-resistant colonies were isolated and tested by PCR for the presence of the correct integration of the excisable cassette into the chromosome (not shown). One colony with the right integration was selected and grown for 4 generations in the absence of Hyg and Km to allow excision of the Hyg cassette and loss of pJV53. Hyg- and Km-sensitive colonies were finally recovered at the expected frequency. One of them was analyzed in parallel with a colony of the wt parental strain by PCR with primers flanking the region used for the recombination. As shown in Fig. 3B, the sizes of the fragments amplified from the two colonies were compatible with their expected sizes (1,263 bp and 1,395 bp for the merodiploid mutant and 1,386 for the wt). The smaller fragment amplified from the merodiploid mutant was sequenced to verify the correctness of the
FIG. 5. (A) Schematic representation of the excisable hyg cassette carried by pAL74 (Pmpt64) and pAL75 (PRv1818c); dif sites are shown in gray. (B) Sequence of the open reading frame (orf) on the 3⬘ side of the cassette. The dif site is boxed, and the restriction sites are underlined. (C) Western blot analysis performed on M. smegmatis. Lane 1, MS140; lane 2, MS143; lane 3, empty; lane 4, MS141; lane 5, MS143. GFP was detected using a monoclonal antibody.
5316
CASCIOFERRO ET AL.
excision: as expected, a copy of the dif sequence flanked by two BglII restriction sites was found in the correct position, replacing the 123 central nucleotides of the target gene (Fig. 3D). Conclusions. We have adapted the Xer-cise system to mycobacteria, showing that it is extremely efficient in this genus. The excisable cassettes that we developed will be extremely useful for the construction of stable, unmarked recombinant strains in both fast- and slow-growing mycobacteria. This system will be useful for several applications: (i) construction of unmarked strains for the development of new vaccines, which could be accomplished both by introducing the gene encoding a protective antigen of a cytokine in the mycobacterial genome with an integrative plasmid carrying the removable cassette and by producing attenuated strains in which a gene essential for virulence has been deleted using the removable cassette; (ii) studying the function of essential genes embedded in operons, in which case it will be possible to construct a merodiploid strain in which the gene of interest is cloned at the L5 attB site and expressed by an inducible promoter and then to introduce an unmarked in-frame deletion into the wt copy of the gene, with the final product being a strain in which the only functional copy of the gene will be transcribed from the inducible promoter without affecting the transcription and translation of the other members of the operon, which will still be transcribed from their physiological promoter; and (iii) construction of multiple mutants.
APPL. ENVIRON. MICROBIOL.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
We thank Rocky Cranenburgh and Issar Smith for helpful discussion, Giovanni Delogu for pMV2-33, pMV4-36, and pMV10-25, and Dario Cascioferro for technical assistance. This work was supported by European Commission EC-VI Framework contract no. LSHP-CT-2006-036871 (INNOVAC) and contract no. LHSP-CT-2005-018923 (NM4TB). REFERENCES 1. Bardarov, S., S. Bardarov, Jr., M. S. Pavelka, Jr., V. Sambandamurthy, M. Larsen, J. Tufariello, J. Chan, G. Hatfull, and W. R. Jacobs, Jr. 2002. Specialized transduction: an efficient method for generating marked and unmarked targeted gene disruptions in Mycobacterium tuberculosis, M. bovis BCG and M. smegmatis. Microbiology 148:3007–3017. 2. Bardarov, S., J. Kriakov, C. Carriere, S. Yu, C. Vaamonde, R. A. McAdam, B. R. Bloom, G. F. Hatfull, and W. R. Jacobs, Jr. 1997. Conditionally replicating mycobacteriophages: a system for transposon delivery to Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. U. S. A. 94:10961–10966. 3. Barik, S., K. Sureka, P. Mukherjee, J. Basu, and M. Kundu. 2010. RseA, the SigE specific anti-sigma factor of Mycobacterium tuberculosis, is inactivated by phosphorylation-dependent ClpC1P2 proteolysis. Mol. Microbiol. 75:592–606. 4. Bloor, A. E., and R. M. Cranenburgh. 2006. An efficient method of selectable
16. 17.
18. 19.
20.
21.
22.
23. 24.
marker gene excision by Xer recombination for gene replacement in bacterial chromosomes. Appl. Environ. Microbiol. 72:2520–2525. Boldrin, F., S. Casonato, E. Dainese, C. Sala, N. Dhar, G. Palu, G. Riccardi, S. T. Cole, and R. Manganelli. 2010. Development of a repressible mycobacterial promoter system based on two transcriptional repressors. Nucleic Acids Res. doi:10.1093/nar/gkq235. Delogu, G., A. Bua, C. Pusceddu, M. Parra, G. Fadda, M. J. Brennan, and S. Zanetti. 2004. Expression and purification of recombinant methylated HBHA in Mycobacterium smegmatis. FEMS Microbiol. Lett. 239:33–39. Delogu, G., C. Pusceddu, A. Bua, G. Fadda, M. J. Brennan, and S. Zanetti. 2004. Rv1818c-encoded PE_PGRS protein of Mycobacterium tuberculosis is surface exposed and influences bacterial cell structure. Mol. Microbiol. 52: 725–733. Hendrickson, H., and J. G. Lawrence. 2007. Mutational bias suggests that replication termination occurs near the dif site, not at Ter sites. Mol. Microbiol. 64:42–56. Lee, M. H., L. Pascopella, W. R. Jacobs, Jr., and G. F. Hatfull. 1991. Site-specific integration of mycobacteriophage L5: integration-proficient vectors for Mycobacterium smegmatis, Mycobacterium tuberculosis, and bacille Calmette-Guerin. Proc. Natl. Acad. Sci. U. S. A. 88:3111–3115. Lydiate, D. J., A. M. Ashby, D. J. Henderson, H. M. Kieser, and D. A. Hopwood. 1989. Physical and genetic characterization of chromosomal copies of the Streptomyces coelicolor mini-circle. J. Gen. Microbiol. 135:941–955. Malaga, W., E. Perez, and C. Guilhot. 2003. Production of unmarked mutations in mycobacteria using site-specific recombination. FEMS Microbiol. Lett. 219:261–268. Manganelli, R., M. I. Voskuil, G. K. Schoolnik, E. Dubnau, M. Gomez, and I. Smith. 2002. Role of the extracytoplasmic-function sigma factor SigH in Mycobacterium tuberculosis global gene expression. Mol. Microbiol. 45:365– 374. Parish, T., and N. G. Stoker. 2000. Use of a flexible cassette method to generate a double unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement. Microbiology 146:1969–1975. Pelicic, V., M. Jackson, J. M. Reyrat, W. R. Jacobs, Jr., B. Gicquel, and C. Guilhot. 1997. Efficient allelic exchange and transposon mutagenesis in Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. U. S. A. 94:10955–10960. Raviglione, M. C. 2003. The TB epidemic from 1992 to 2002. Tuberculosis (Edinb.) 83:4–14. Recchia, G. D., and D. J. Sherratt. 1999. Conservation of xer site-specific recombination genes in bacteria. Mol. Microbiol. 34:1146–1148. Sassetti, C. M., D. H. Boyd, and E. J. Rubin. 2003. Genes required for mycobacterial growth defined by high density mutagenesis. Mol. Microbiol. 48:77–84. Song, H., and M. Niederweis. 2007. Functional expression of the Flp recombinase in Mycobacterium bovis BCG. Gene 399:112–119. Song, H., F. Wolschendorf, and M. Niederweis. 2008. Construction of unmarked deletion mutants in mycobacteria, p. 279–295. In T. Parish and A. C. Brown (ed.), Mycobacteria protocols. Humana Press, Totowa, NJ. Stephan, J., V. Stemmer, and M. Niederweis. 2004. Consecutive gene deletions in Mycobacterium smegmatis using the yeast FLP recombinase. Gene 343:181–190. Stover, C. K., V. F. de la Cruz, T. R. Fuerst, J. E. Burlein, L. A. Benson, L. T. Bennett, G. P. Bansal, J. F. Young, M. H. Lee, G. F. Hatfull, et al. 1991. New use of BCG for recombinant vaccines. Nature 351:456–460. van Kessel, J. C., and G. F. Hatfull. 2008. Efficient point mutagenesis in mycobacteria using single-stranded DNA recombineering: characterization of antimycobacterial drug targets. Mol. Microbiol. 67:1094–1107. van Kessel, J. C., and G. F. Hatfull. 2007. Recombineering in Mycobacterium tuberculosis. Nat. Methods 4:147–152. van Kessel, J. C., L. J. Marinelli, and G. F. Hatfull. 2008. Recombineering mycobacteria and their phages. Nat. Rev. Microbiol. 6:851–857.
