Journal of Microbiological Methods 103 (2014) 29–36
Contents lists available at ScienceDirect
Journal of Microbiological Methods journal homepage: www.elsevier.com/locate/jmicmeth
Efﬁcient construction of unmarked recombinant mycobacteria using an improved system Feng Yang a,1, Yaoju Tan b,1, Jia Liu a, Tianzhou Liu a,c, Bangxing Wang a,d, Yuanyuan Cao a,d, Yue Qu e, Trevor Lithgow e, Shouyong Tan b, Tianyu Zhang a,⁎ a
State Key Laboratory of Respiratory Diseases, Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, China State Key Laboratory of Respiratory Diseases, Department of Clinical Laboratory, Guangzhou Chest Hospital, China School of Life Sciences, University of Science and Technology of China, China d School of Life Sciences, Anhui University, China e Department of Biochemistry and Molecular Biology, School of Medicine, Nursing and Health Sciences, Monash University, Australia b c
a r t i c l e
i n f o
Article history: Received 9 January 2014 Received in revised form 3 May 2014 Accepted 5 May 2014 Available online 27 May 2014
a b s t r a c t The genetic study of mycobacteria, such as Mycobacterium tuberculosis and Mycobacterium ulcerans, is hampered heavily by their slow growth. We have developed efﬁcient, versatile, and improved genetic tools for constructing unmarked recombinant mycobacteria more rapidly including generating multiple mutants using the same antibiotic marker in both fast- and slow-growing mycobacteria. © 2014 Elsevier B.V. All rights reserved.
Keywords: In-frame deletion Mycobacteria Recombineering Unmarked gene disruption Xer recombinase system
1. Introduction Tuberculosis (TB), an infectious disease caused by Mycobacterium tuberculosis (MTB), is one of the most serious single infectious diseases of mortality worldwide. The 8 million incident cases of TB, 1.45 million deaths from TB patients in 2011 alone (Nahid and Menzies, 2012), and the appearance of multidrug-resistant (MDR), extensively drug-resistant (XDR) and even totally drug-resistant (TDR) TB (Loewenberg, 2012) provide a striking reminder of the magnitude of destruction caused by TB. However, the laboratory research of MTB is highly impeded by the slow growth rate of the MTB strains and lack of convenient and efﬁcient genetic tools for recombination and mutagenesis (van Kessel et al., 2008). The development and application of genetic manipulation in mycobacteria have accelerated the study of mechanisms of TB Abbreviations: CST1, dif-ΩHYG-dif CASSETTE V1; CST2, dif-ΩHYG-dif CASSETTE V2; HYG, hygromycin B; Hyg, HYG-resistant gene; KAN, kanamycin; MTB, Mycobacterium tuberculosis; TB, tuberculosis; sacB, a Bacillus subtilis gene encoding the enzyme levansucrase. ⁎ Corresponding author at: Room A132, 190 Kaiyuan Ave, Science Park, Guangzhou, Guangdong ZIP code: 510530, China. Tel.: + 86 18819181735, + 86 2032015270; fax: +86 2032015270. E-mail address: [email protected]
(T. Zhang). 1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.mimet.2014.05.007 0167-7012/© 2014 Elsevier B.V. All rights reserved.
pathogenesis and have been used extensively in the generation of potential recombinant live vaccines and in understanding the mechanisms of drug action and resistance. Appropriate antibiotic resistance genes are required for direct selection of the subset of bacteria that have taken up the DNA materials used for constructing recombinant mycobacteria, such as plasmids, cosmids and phasmids (Jacobs et al., 1987). However, the use of antibiotic resistance markers has some drawbacks. Firstly, there are only a few antibiotic resistance genes, especially for mycobacteria. In fact, only kanamycin (KAN) and hygromycin B (HYG) resistance marker genes are used in genetic experiments with strains of the MTB complex (Malaga et al., 2003). Some MTB clinical isolates are already resistant to KAN, which makes KAN resistance gene not suitable for them. Secondly, if a recombinant mycobacterial strain already has a resistance cassette, this excludes the marker for further use in it. This can greatly hamper the study of the functions of redundant genes in mycobacteria in which genomic studies have revealed extensive gene duplications with highly conserved sequences (Garnier et al., 1998) since disruption of one gene may be compensated by its homologous gene (Puech et al., 2002; Agarwal et al., 2009). Therefore, multiple mutation and complementation experiments would be required for uncovering the biological roles of a gene or a series of genes having the same or compensatory functions. Thirdly, the insertion of a resistance gene into an operon could affect the expression of the
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
corresponding downstream genes and the polar effect can complicate the characterization of each gene's roles. Fourthly, our previous study found that HYG resistance gene (Hyg) may disrupt the virulence of Mycobacterium ulcerans. Recombinant M. ulcerans strains integrated with a plasmid containing Hyg could not cause mouse footpad swelling. However, M. ulcerans strains with KAN resistance gene caused the swelling (unpublished data). Therefore, HYG resistance gene might affect its virulence. Lastly, in some special situations, for example, the construction of reporter mycobacterial strains for screening and evaluating drug activity, the resistance marker is not desired because of the potential cross resistance or the inability to test the activity of combined drugs. The unmarked mutations can be obtained by a two-step strategy in which the ﬁrst step is to only select one single cross-over recombinant event while the second step is to select the cells in which the second cross-over has occurred. This involves the use of multiple markers for selection and counter-selection of the different allelic exchange events (Parish and Stoker, 2000). In this case, only some of the selected cells would be mutated with the others recovered as wild type. Thus, this method is problematic and more applicable where the mutant has a distinct phenotype, such as auxotrophy. A common way to circumvent this problem is by the production of unmarked mutations which can be realized by developing an antibiotic resistance cassette ﬂanked by two short DNA sequences in direct orientation which can be recognized and catalyzed by a recombinase or resolvase. The recombination of these two short sequences allows the removal of the resistance marker after it has been used for the positive selection and its use in another round of genetic manipulation. These site-speciﬁc recombination systems are usually derived from bacteriophages or transposons. Three such systems have been successfully used with mycobacteria and were summarized before (Cascioferro et al., 2010). However, all the systems require the expression of an exogenous resolvase or recombinase from a plasmid to allow the excision of the resistance marker which makes the procedure complicated and very time-consuming. Recently, a new sequence-speciﬁc recombinase system based on the endogenous Xer recombinases (Xer-cise) was successfully adapted to mycobacteria (Cascioferro et al., 2010). In this system, the antibiotic resistance cassette was ﬂanked by dif sites which can be recognized and resolved by the endogenous recombinases XerC and XerD. The system does not need the introduction and subsequent elimination of the extrachromosomal plasmids containing exogenous genes for removing the resistance cassette, which makes it extremely simple and practical (Cascioferro et al., 2010). However, the Hyg cassette they used was not optimized for the length, the promoter of Hyg and the restriction sites in and surrounding it. In addition, the reuse of the same cassette in a single strain was not demonstrated before. In this study, we constructed an artiﬁcial, versatile, dif-containing resistance cassette based on the shortened, optimized Hyg ﬂanked with artiﬁcial promoter and multi-cloning sites. It could be an efﬁcient, versatile tool for constructing unmarked mycobacteria especially combined with the recombineering technique. Recombineering is an important progress in mycobacterial genetic manipulation by inducible expression of recombination proteins derived from mycobacteriophage Che9c for facilitating allelic exchange through increasing the recombination rate (van Kessel and Hatfull, 2007). This novel cassette and the improved recombineering system will be helpful in stimulating mycobacterial genetic research. 2. Material and methods 2.1. Bacterial strains and culture conditions Escherichia coli strain DH5α or ET12567 was grown at 37 °C in Luria– Bertani (LB) broth or on LB agar. Mycobacterium smegmatis mc2 155 (MS) and MTB H37Ra were grown in Middlebrook 7H9 broth (Difco) supplemented with 10% oleic acid albumin dextrose catalase (OADC, Becton
Dickinson), carbenicillin 50 μg/ml, cycloheximide 10 μg/ml and 0.05% tween 80 as indicated, or on LB agar or on solid Middlebrook 7H11 medium (Difco) supplemented with OADC (or ADC for MS) or containing 10% sucrose if necessary. Ampicillin (Sigma) 100 μg/ml for E. coli, KAN (Invitrogen) 40 μg/ml for all species or HYG (Roche Diagnostics) 200 μg/ml for E. coli, 150 μg/ml for MS and 50 μg/ml for MTB were added to the agar plates when required. About 20 μg/ml KAN and 100 μg/ml HYG for MS and 10 μg/ml for MTB were added to liquid broth when required. 7H9 induction medium was 7H9 broth supplied with 0.2% succinate and KAN at 20 μg/ml prepared as described before (van Kessel and Hatfull, 2008). 2.2. Plasmid construction 2.2.1. The dif-ΩHYG-dif cassette V1 (CST1) and pTY95 Two copies of the dif site were synthesized as DNA oligonucleotides (Table 1) and were inserted into plasmid pUC19 at HindIII–KpnI sites by 3-fragment-ligation (Zhang et al., 2006a, 2006b) resulting in pTYd. A primer Hse (Table S1) was designed for sequencing the upstream of Hyg gene in pNBV1 (Howard et al., 1995). The Hyg gene was ampliﬁed successfully by PCR from pNBV1 (Howard et al., 1995) with primers Hygf2 and Hygr727 (Table S1), cut with XbaI and inserted into pTYd at XbaI site to give pTYdH. Its identity was veriﬁed by restriction digestion and sequencing. The promoters of the wild type Hyg cassette and the artiﬁcial promoter in the CST1 (Fig. 1A) were analyzed by the online Neural Network Promoter Prediction program (http://www.fruitﬂy. org/seq_tools/promoter.html). The plasmid pTYdHm was obtained by removing KpnI and EcoRI sites in the Hyg in pTYdH using site-directed mutagenesis (Generay, China). To test if the Hyg gene could cause resistance in mycobacteria and the excision efﬁciency of the CST1, we replaced the KAN resistance gene of the integrative plasmid, pMH94 (Lee et al., 1991), with CST1 at the HindIII sites to give pTY95 (Fig. 1B). 2.2.2. pJV53Ts and pTY46H We digested the pJV53 plasmid with XbaI–SpeI and inserted the larger fragment containing the inducible promoter and gp60/61 genes into thermosensitive pPR23 (Pelicic et al., 1997) which was cut with the same enzymes and dephosphorylated to avoid the self-ligation as XbaI and SpeI are isoaudamers. The recombinant plasmid having the
Table 1 DNA oligonucleotides used in this study except those related to Hyg gene ampliﬁcation. DNA oligo
Nucleotide sequence (5′–3′)
Dlup DlLow DrUP
(p)AGCTTCTCGAGTAAGCCGATAAGCGACATTATGTCAAGTCCCGGGT (p)CTAGACCCGGGACTTGACATAATGTCGCTTATCGGCTTACTCGAGA (p)CTAGATCGATTAAGCCGATAAGCGACATTATGTCAAGTCTCGAGAA GCTTGGTAC (p)CAAGCTTCTCGAGACTTGACATAATGTCGCTTATCGGCTTAATCGAT AGAGCACCAACCCCGTACTG GTGAAGTCGACGATCCCGGT TTCATGTGCGCTCGGATCAT TCACGCTGGAGGAGTACACC TAAGAATTCGCTAGCCGACCTTCGTGATGACTGGC GATCTGCAGAGGATCCTGGTACCAGGTGGCGAAACGGCTGA CCACTGCAGCTCGAGTGCGGGTCGGGGAACAAC ACCAAGCTTACTAGTTGCAGCCCATCGGACAACA GCAGGATCCGATATCGAATTCCATATGCCCAAGCTTCTCGAGACT GCTGGTACCGCTAGCGCTGCAGCCATGGCAAGCTTCTCGAGTAAG GGTCCATGGTGAGCAAGGGCGAGG GCGGGTACCTTACTTGTACAGCTCGTCCATGC GTTAGCTTAGGCATACATAAGG ATCTCGGAGGTCTTGACCAT CCTGGATTCGCGTGCCTACC TACGGGGTTGGTGCTCTCGC GCCGCCTATGTAGAGCTGGTCGTT CTTGCCGGTGGTGCAGATGA
DrLow Hyg72F Hyg72R Int72F Int72R Ms614f Ms614r Ms616f Ms616r HV2f HV2r GYFf GYFr a b c d Hygdf Hygdr
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
Fig. 1. Diagram of the dif-ΩHYG-dif cassettes, plasmid pTY95, pJV53Ts and pTY46H. A, The artiﬁcial dif-ΩHYG-dif cassettes. Hygf2 and Hygr727 are primers used to amplify Hyg gene. P1, P2, P3: Putative promoters of the natural Hyg gene; Pa: putative artiﬁcial promoter of Hyg gene covering dif1 in dif-ΩHYG-dif cassettes. dif: the putative MTB dif sequence. Useful enzyme sites: E1, EcoRI; Nt, NotI; Sp, SpeI; B, BamHI, E5, EcoRV; Nd, NdeI; H, HindIII; Xo, XhoI; C, ClaI; Xb, XbaI; S, SmaI; Nc, NcoI; P, PstI; Nh, NheI; K, KpnI. B, Plasmid pTY95 for testing the excision efﬁciency of CST1. C, The thermosensitive plasmid pJV53Ts for recombineering. D, The plasmid for production of linear fragment for deletion of Ms0615–0616. bla, ampicillin resistance gene; attP, mycobacteriophage L5 attachment site; int, integrase gene; oriE, origin region of E. coli; ts oriM, thermosensitive origin region of mycobacteria; Kan, KAN resistance gene; gp60/61 from mycobacteriophage Che9c for facilitating homologous recombination; Ind Pr, acetamidase promoter which is inducible; sacB, a Bacillus subtilis gene encoding the enzyme levansucrase, a counter-selectable marker; gm, gentamycin resistance gene. Commonly used restriction enzyme sites were indicated.
gp60/61 genes and the sacB gene in the same direction was selected and designated as pJV53Ts (Fig. 1C). A 641-bp DNA fragment (ArmL) upstream of Ms0615 and a 610-bp DNA fragment (ArmR) of Ms0616 were ampliﬁed using primer couples Ms614f–Ms614r and Ms616f– Ms616r, respectively (Table 1) and cloned into EcoRI–HindIII sites of the pbluescript II SK(+) by 3-fragment-ligation (Zhang et al., 2006a, 2006b) to give the plasmid p46MSYF which was veriﬁed by restriction digestion and sequencing. The CST1 was then introduced into its XhoI site to give the plasmid pTY46H (Fig. 1D). 2.2.3. pblDHC1m The CST2 was ampliﬁed from pTYdHm by PCR using primers HV2f and HV2r which contained multi-cloning sites at their 5′ ends (Table 1). The cassette was then digested with KpnI and BamHI and inserted into pbluescript II SK(+) cut with the same enzymes. The resultant plasmid containing the CST2 (Fig. 1A) was designated as pblDHC1m (Fig. 2A), and was veriﬁed by sequencing. 2.2.4. Construction of integrative plasmid pTYGi9 expressing eGFP The egfp gene was ampliﬁed from pEGFP-N1 using primers GYFf and GYFr (Table S1), digested with KpnI and NcoI, and then cloned into pblDHC1m (Fig. 2A) cut with the same enzymes to give plasmid pTYG18 which was veriﬁed by restriction digestion and sequencing. We obtained the 2.1-kb attP:int fragment from pblueINT (Zhang et al., 2012) digested with XbaI and EcoRI, and inserted it into pTYG18
digested with SpeI and EcoRI to give pTYGi9 (GenBank# KF957646, Fig. 2B). 2.3. Allelic replacement by recombineering A wild type MS was transformed with pJV53Ts (Fig. 1C) or pJV53 ﬁrst and designated as MS-TS53 and MS-53, respectively. The fragment containing the ArmL–CST1–ArmR was then cut from pTY46H (Fig. 1D) by EcoRI and HindIII and transformed into induced competent cells of MS-TS53 and MS-53, respectively. Several HYG-resistant colonies were isolated and allelic replacements of the wild type Ms0615– Ms0616 genes from MS-TS53 derived colonies were tested individually by PCR using appropriate primers (Table 1), respectively. 2.4. Counter-selection HYG-resistant MS or MTB H37Ra transformed with CST1 or CST2 (Fig. 1A) were cultured individually in liquid media until they reach late logarithmic phase without selection to allow excision of the HYG cassette by the endogenous mycobacteria XerC and XerD. Ten-fold serial dilutions of bacterial culture were plated on plain agar plates. The colonies were picked and replica streaked on both plain and HYG-containing plates. The HYG-sensitive colonies were veriﬁed further by PCR using primers hyg72f and hyg72r (Table 1) for ampliﬁcation of the fragment in the dif-ΩHYG-dif cassette.
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
Fig. 2. Plasmid pblDHC1m (A) and plasmid pTYGi9 (B). bla, ampicillin resistance gene; oriE, origin region of E. coli; attP, attachment site of mycobacteriophage L5; int, integrase gene of mycobacteriophage L5; dif-Hyg-dif is as indicated in CST2 in Fig. 1 with eGFP gene inserted into its SpeI and EcoRI sites. Commonly used restriction enzyme sites were indicated.
To further cure the plasmid pJV53Ts or pJV53, individual colonies were ﬁrst cultured in liquid broth without selection at 42 °C or 37 °C for 3 days. The bacterial culture was then diluted and plated onto plain plates to obtain discrete colonies which were further replica plated in the presence and absence of KAN or 10% sucrose. 3. Results and discussion 3.1. Construction of CST1 To obtain an improved excisable HYG-resistance cassette, we constructed a much shorter, modiﬁed Hyg ﬂanked by putative MTB H37Rv dif sites (Cascioferro et al., 2010) and an artiﬁcial promoter in the same direction and surrounded by appropriate restriction sites (Fig. 1A). The Hyg gene from Streptomyces hygroscopicus (Garbe et al., 1994) is widely used in mycobacterial genetic manipulations (Howard et al., 1995). According to the sequence analysis of Hyg cassette contained in pNBV1, three putative promoters of Hyg (Howard et al., 1995) were found (Table S2, Fig. 1A). A pair of primers for ampliﬁcation of the Hyg gene was designed manually. The sequence of the upstream primer (Hygf2, Table S1) was close to the start codon and could form an artiﬁcial promoter with the dif and the XbaI site in CST1 (Fig. 1A). Its sequence was ATCGATCTAGACCGGCC TCGAG which was predicted with an online program (http://www.fruitﬂy.org/ seq_tools/promoter.html). The score was 0.8 with the total score being 1.0 and the score ≥0.6 is believed to be signiﬁcant. The dif sequence is boxed and the XbaI site is italicized. The boldface T indicates the transcription starting site. The 5′ end of the downstream primer (Hygr727, Table S1) except for the enzyme sites annealed to the stop codon accurately. In this case, the functional Hyg fragment is the shortest and thus could avoid the potential 5′ and 3′ higher order structure which not only makes ampliﬁcation of the gene difﬁcult in identifying the insertion site of the recombinant strain by PCR, but also creates strong polar effect in gene disruption experiments as a hypothetical concern. For example, an earlier study reported difﬁculties in ampliﬁcation of the Hyg cassette by PCR, which was attributed to the possible higher order structure(s) in DNA containing the Hyg gene and the higher G + C content (Stratigopoulos and Cundliffe, 2002). To attest this, we tried several pairs of primers (Hygf0–Hygr0, Hygf2–Hygr2, Hygfnew–Hygrnew and Hygpf–Hygpr, Table S1) designed using the software Primer Premier 5.0
(Premier, Canada) for amplifying the unmodiﬁed Hyg using the pNBV1 as a template. We did ﬁnd that the Hyg in the original cassette was very hard to be ampliﬁed, because no correct fragments were obtained using different primer combinations. Furthermore, ampliﬁcation was not successful by touchdown PCR and with LA Taq used for amplifying long DNA fragments with high G + C content (Takara, Dalian, China). The inherent EcoRI site in the original Hyg gene (Fig. 1) and KpnI site introduced by an accident mutation in the Hyg gene by PCR were then removed by changing A78 into G and A282 into T by site-directed mutation. The CST1 in plasmid pTYdHm was constructed successfully with such an order: HindIII–XhoI–dif–ClaI–XbaI–Hyg–XbaI–SmaI–dif–XhoI– HindIII with the termination codon of Hyg gene close to the SmaI site (Fig. 1A). Many noncutters in the XbaI–Hyg–XbaI fragment were found: ApaI, AvrII, BamHI, BglII, ClaI, EcoRI, EcoRV, HindIII, HpaI, KpnI, NcoI, NdeI, NheI, NotI, NruI, PacI, PstI, PvuII, ScaI, SmaI, SpeI, and XhoI. The sites ﬂanking XbaI–Hyg–XbaI in CST1 are underlined. Only SalI, SacI and NarI were commonly used restriction sites found in the fragment XbaI–Hyg–XbaI. The Hyg gene in CST1 could possibly be expressed efﬁciently due to the artiﬁcial promoter, which was supported by the ﬁnding that although the direction of Hyg gene was opposite to all the other known promoters in pTYdHm, it could still cause high-level HYG resistance in E. coli. However, we could not exclude that some hitherto unknown vector promoter could be providing transcription. Additional elements inserted into the ClaI or SmaI site could possibly be excised together with the Hyg gene in between dif1 and dif2 when constructing unmarked mycobacteria. It needs to be pointed out that SmaI site is downstream of the Hyg gene and insertion of a fragment into this site would not inﬂuence its potential promoter. In addition, any bluntended fragments can be inserted into the SmaI site, as the cognate enzyme produces blunt ends. 3.2. The excision efﬁciency of CST1 at dif sites The number of resistance markers available for mycobacterial genetic manipulation is very limited, and thus construction of an unmarked mutant by the removal of the CST1 from the chromosome would be very useful. Traditionally, another plasmid expressing the recombinase or resolvase was needed, as they could recognize special sequences ﬂanking the resistance marker and remove the cassette. The transformation of such a plasmid, followed by the removal of the marker and the subsequent elimination of the plasmid made the creation of unmarked
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
mycobacteria complicated and time-consuming. This is especially so for the slow-growing mycobacteria which need weeks or even months of incubation to form colonies on agar plates (Zhang et al., 2010, 2013). Herein, the MTB dif sequence could be recognized and removed by the endogenous Xer recombinases expressed by XerC and XerD genes from the mycobacterial genome as reported recently (Cascioferro et al., 2010). The integrative plasmid pTY95 (Fig. 1B) containing CST1 was constructed and transformed into MS successfully and was designated as MS94H. The MS transformed with pck0601 (Manabe et al., 2005), an integrative plasmid containing a commonly used Hyg gene fragment, was used as a control. MS94H grew as fast as the control strain both in broth and on agar plates, which was observed by us for many times (data not shown). This demonstrated that the CST1 could cause enough HYG resistance for the MS to grow and the dif did not affect replication when inserted at this site in MS. About 84% of MS94H colonies lost their HYG resistance just after one passage. Those which could only grow on plain plates were designated as MS94HL and veriﬁed further by PCR to test if the Hyg gene was lost or not. The primers Hyg72f and Hyg72r (Table S1) were used to amplify a 580-bp fragment within the Hyg gene, while the primers Int72F and Int72R (Table 1) were used to amplify a 586-bp fragment within the integrase gene which would remain in the genome. MS94H and wild type MS were used as positive and negative controls respectively. As expected, no correct sized products were obtained with wild type MS whereas correct sized products were obtained with MD94H using either pairs of primers. Ampliﬁcation using MS94HL as template, ~ 580 bp products were obtained using primers Int72F and Int72R but no correct sized products were obtained using primers Hyg72f and Hyg72r. This indicated that the Hyg gene was already lost (data not shown), and that the CST1 (Fig. 1A) could be removed efﬁciently in MS. This is in accordance with reported ﬁndings that MS Xer recombinases are able to recognize the MTB dif sequence despite the 6 mismatches between the dif sites in the two species (Cascioferro et al., 2010). The CST1 could also be removed more efﬁciently in MTB H37Ra (see Section 3.5 below). 3.3. Construction of an unmarked gene deletion in MS To demonstrate the feasibility of introducing unmarked gene deletions in the mycobacterial chromosomes with the CST1, we introduced a 2081-bp deletion in Ms0615–0616 genes in the esx-3 locus, which is crucial for the pathogenicity of MS (Sweeney et al., 2011). A recombineering system to increase homologous recombination efﬁciency was developed recently by introducing the plasmid pJV53 expressing mycobacteriophage Che9c genes 60 and 61 (Mahenthiralingam et al., 1993; Parish et al., 1997). To ensure more efﬁcient removal of the recombineering plasmid after completion of recombination, we created a new plasmid, pJV53Ts (Fig. 1C), which has both advantages of the counter-selective properties of the sacB gene and the mycobacterial thermosensitive origin of replication, and can therefore be efﬁciently removed after the recombineering. The plasmids pJV53Ts and pJV53 were transformed into MS and the resultant strains were designated as MS-TS53 and MS-53, respectively. The linear recombineering fragment comprising ArmL, CST1 and ArmR was cut from the plasmid pTY46H with EcoRI–HindIII (Fig. 1D), puriﬁed and transformed into MS-TS53 competent cells which were induced with acetamide at 30 °C and prepared as described before (van Kessel and Hatfull, 2008). Several HYG-resistant colonies were isolated from 37 °C incubation and cultured in 7H9 broth without selection. Two of them were tested by PCR for the presence of the correct exchange of the excisable cassette into the chromosome and the correct ones were designated as MS-46H. The positive MS-46H strains with right recombination crossover were selected by the sizes of the fragments ampliﬁed from the colonies. The expected sizes were 2315 bp for the MS-TS53 and 1396 bp for MS-46H strains with primers a and
b; 1071 bp for MS-46H strain and no product for the MS-TS53 with primers c and d (Fig. 3A). We however obtained a ~240-bp band instead of 1396 bp for both of the 2 MS-46H broth samples using primers a and b (Fig. 3B, left panel lanes 4 and 5) as it was for the control plasmid pTY46H containing ArmL, CST1 and ArmR as a template (Fig. 3B, left panel, lane 2). It was very possible that during incubation without selection, most of the MS-46H cells lost the CST1 fragment and resulting into MS-D56 (see below) and the PCR products for such strains with primers a and b were only 236 bp and not 1396 bp. Furthermore, the shorter one can be ampliﬁed much easier in competition. This further indicated that the excision rate of the CST1 was very high. No PCR product should have been expected using pTY46H as a template with primers c + d, as primer c had no corresponding sequence on it and thus the band would be a nonspeciﬁc product with wrong size (Fig. 3B, right panel, lane 2). In theory, only MS with right homologous recombination could show the right size (1071 bp) with primers c + d, and the two MS-46H strains were veriﬁed to be positive by PCR (Fig. 3B, right panel lanes 4 and 5). Subsequently, the sub-cultured colonies from veriﬁed MS-46H were cultured individually in 5 ml 7H9 broth with tween80 for 1–2 passages without selection to disrupt the clumps and allow excision of the HYG cassette further. HYG-sensitive colonies were ﬁnally recovered at a high frequency as more than 70% of the colonies lost their HYG resistance just after one passage. The mutant strains which lost the CST1 but still contained pJV53Ts were designated as MS-D56. Four of them were analyzed parallel to MS-TS53 strain by PCR with primers a, b, c and d (Table 1) as indicated (Fig. 3A, C). The sizes of the fragments ampliﬁed were as expected: 236 bp for the unmarked mutant MS-D56 (Fig. 3C, left panel lanes 4–7) and 2315 bp for MS-TS53 (Fig. 3C, left panel lane 2) with primers a and b; 885 bp for MS-D56 (Fig. 3C, right panel lanes 4–7) and 2964 bp for MS-TS53 (Fig. 3C, right panel lane 2) with primers c and b. Only MS mutants obtained from the right homologous recombination and excision of CST1 could give the right size products with both primer pairs a + b and c + b. The fragment ampliﬁed from one of the mutants with primers a + b was sequenced to verify the correctness of the excision. A copy of the dif sequence ﬂanked by XhoI was found to be in the correct position and it replaced a large part of Ms0615 and a small part of Ms0616 genes as designed. Two MS-D56 colonies were cultured individually for curing of plasmid pJV53Ts (Fig. 1C). After two passages in plain 7H9 broth incubated at 42 °C, the MS-D56 culture was diluted and plated onto LB plates containing 10% sucrose at 42 °C. About 100 colonies were individually cultured in replica on plates containing 10% sucrose and plates containing KAN. All grew on 10% sucrose but not on KAN. Several colonies were individually subcultured in plain 7H9 broth and then replica plated onto plates with and without KAN. They were all veriﬁed to have lost the plasmid pJV53Ts. Finally, we obtained the real unmarked MS mutants with the Ms0615–0616 gene deletions and the cured pJV53Ts plasmid which was sensitive to both KAN and HYG, and was designated as MS-D56U. The plasmid pJV53Ts (Fig. 1C) could be eliminated efﬁciently just after the fragment containing ArmL, CST1 and ArmR was transformed into MS-TS53 competent cells, recovered for 4 to 6 h, and incubated on plates containing HYG and 10% sucrose at 42 °C. We checked at least 5 such colonies and all of them had lost pJV53Ts. So the plasmid pJV53Ts could be lost during the transformation step or after the Hyg cassette was removed. If pJV53Ts plasmid left in the strain after the Hyg cassette was excised, it is suitable for deleting another part of the genome of the same strain. We also compared the curing rate of pJV53 and pJV53Ts. By a similar way, several MS-53 strains were passed more than twice. They were cultured at 37 °C and replica plated in the presence and absence of KAN. The pJV53 plasmid could not be easily lost, as we did not obtain any strain with pJV53 cured after at least two selection runs. The MS strain containing pJV53Ts was also passed without selection at 37 °C (lower than 39 °C suggested before) more than twice and replica plated in the presence and absence of KAN, but no colony cured of pJV53Ts was found.
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
Fig. 3. Unmarked deletion of Ms0615–0616 genes. A: Diagram of introduction of an unmarked deletion in Ms0615–0616 genes in the esx-3 locus of MS. The dif stands for the putative MTB dif sequence. The triangles stand for the primers and their directions. The targeting substrate was from pTY46H. B: Identiﬁcation of Ms0615–0616 gene deletion mutants (MS-46H). PCR in left panel using primer pair a + b and right panel using primer pair c + d. Lane M, 1-kb ladder DNA marker (kb); lane 1, PCR products from water as controls (no template); lane 2, PCR products from plasmid pTY46H containing ArmL-dif-ΩHYG-dif-ArmR; lanes 3, 6, PCR products from MS-TS53 (MS containing pJV53Ts.); lanes 4, 5, PCR products from passed MS-46H broth culture. C: Further identiﬁcation of unmarked Ms0615–0616 gene deletion mutants (MS-D56). PCR in left panel using primer pair a + b and right panel using primer pair c + b. Lane M, 1-kb ladder DNA marker (kb); lane 1, PCR products from plasmid pTY46H containing ArmL-dif-ΩHYG-dif-cassette ArmR; lane 2, PCR products from MS-TS53 (MS containing pJV53Ts); lane 3, water; lanes 4–7, PCR products from MS-D615K colonies (Their genomes were supposed to be the same as MS-D56's.).
Additionally, in our pilot study, we found that a lot of MS-TS53 colonies which could grow on agar containing 10% sucrose were still resistant to KAN. This is in accordance with earlier reports that mycobacteria exhibited high frequency of spontaneous sucrose resistance (Parish and Stoker, 2000). Therefore, the thermosensitive characteristic of the plasmid pJV53Ts is very important for its removal. 3.4. The dif-ΩHYG-dif cassette V2 (CST2) and in-frame deletions using CST1 and CST2 As shown above, the CST1 could be removed by XhoI or HindIII from pTYdHm. However, the direction of the cassette is hard to control when
inserting it in between the up and down stream regions (ArmL and ArmR) of the target genes for gene deletion. Therefore, we created a new plasmid, pblDHC1m (Fig. 2A, GenBank# KF737912), to obtain the directed dif-ΩHYG-dif cassette by surrounding it with many useful restriction sites. The sequence of the CST2 (Fig. 1A) was conﬁrmed by sequencing. Enzyme sites ﬂanking the CST2 (Fig. 1A) were changed by introducing them through PCR. It is very important for an excisable cassette to be able to create inframe deletions so as to avoid disrupting the expression of its downstream genes. For this purpose, we analyzed the open reading frames and the termination codons in the sequence of putative MTB H37Rv dif site ﬂanked by XhoI and HindIII sites, which were supposed to be
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
the leftovers of the CST1 in the frame after excision. This similar analysis has been reported using the dif system (Cascioferro et al., 2010). As shown in Fig. S1A and B, the possible read-through open reading frames were present in both directions. It was thus possible to design cloning strategies to allow the creation of in-frame deletions with the CST1 cut with either XhoI or HindIII and inserted in either direction between the left and right arms of the target gene. For the CST2, it was mainly intended to achieve directional cloning in the in-frame deletion experiments. The leftover of the CST2 in the frame after excision was supposed to be the fragment containing SpeI, BamHI, EcoRV, EcoRI, NdeI, HindIII, XhoI, dif, XhoI, HindIII, NcoI, PstI, and NheI (Fig. 1A). As shown in Fig. S1C, we designed an open reading frame which could be read through in the forward direction. It is possible to design in-frame deletions when cloning the left and right arms of the target gene into the multi-cloning sites upstream and downstream of the CST2, respectively. It is also possible to ﬁrst construct a plasmid containing the left and right arms of the target gene with appropriate enzyme sites in between them and then insert the CST2 directionally into these sites. The open reading frames in the reverse direction are not shown because none of them could be read through. 3.5. No polar effect using CST2 in both fast- and slow-growing mycobacteria To test if CST2 can be used to create unmarked gene deletion without disrupting the expression of the corresponding downstream genes, we created an integrative plasmid pTYGi9 (Fig. 2B) in which the int, Hyg from CST2 and the eGFP gene were in the same direction. The plasmid pTYGi9 was introduced into both MS and MTB H37Ra by electroporation. HYG-resistant colonies were picked up, and tested individually by PCR for the integration of the plasmid into the MS or MTB H37Ra chromosomes. The positive MS and MTB colonies were designated as MSeGFH and MTBeGFH respectively and were grown for 1 passage in plain 7H9 broth to allow excision of CST2. HYG-sensitive MS and MTB colonies were ﬁnally recovered at the frequency as we expected (about 80% for MS and about 100% for MTB), which is consistent with ﬁndings mentioned above and the previous report (Cascioferro et al., 2010). This was probably because the dif sequence used in both studies was from MTB and there were only 6 mismatches in MTB dif out of 28 bp dif sequence in MS. MS Xer recombinases are able to recognize and remove the MTB dif sequence but with a lower efﬁciency. It could also be due to different expression levels and/or catalysis efﬁciency of the enzymes or perhaps the different growth rates. Four of the obtained MS colonies were analyzed parallel to the wild type MS strain as a control by PCR with the primers Hygdf and Hygdr (Table 1) ﬂanking CST2 in pTYGi9 (Fig. 2B). As shown in Fig. S2, all 4 HYG-sensitive MS and 5 MTB colonies counter-selected from MSeGFH or MTBeGFH gave a band with the expected size (0.5 kb) while the wild type MS and MTB did not give any products. We randomly selected, puriﬁed and sequenced the PCR products from lane 4 in the left panel and lane 2 in the right panel. Their sequences were as expected. The corresponding strains veriﬁed by sequencing were designated as MSeGF and MTBeGF. The MSeGFH, MSeGF, MTBeGFH and MTBeGF were tested for the expression of eGFP gene with the wild type MS strain as a control. As expected, eGFP was expressed by both MSeGFH and MSeGF (Fig. S3A and B) and both MTBeGFH and MTBeGF (Fig. S3A), which indicated that the eGFP gene can be transcribed and translated irrespective of whether the dif-ΩHYG-dif cassette is upstream of eGFP gene or not. This demonstrated that the dif-ΩHYG-dif cassette did not create polar effect when used to replace a gene in an operon, as the gene downstream of it could be expressed in the presence or absence of the cassette. It is possible that the putative artiﬁcial promoter in the front of the cassette and also in the remaining section in the genome following the cassette excision could initiate the transcription of the downstream gene(s) of the CST2 (Fig. 1A). In addition, subsequent translation was not affected, although we can not exclude the possibility that the transcription might be promoted partially by the promoter of int gene
because it was upstream of eGFP gene following integration of the pTYGi9 plasmid into the MS genome. In fact, the promoter of the int gene is similar to the promoter of the operon containing the target gene for disruption. It needs to be pointed out that we cannot exclude that a neutral effect may exist, depending on the strength and any regulation (e.g. growth phase, cell density) of the artiﬁcial promoter that is now positioned upstream of a gene (genes) within an operon after the CST deletion. If the gene(s) is naturally downregulated under a certain condition, it may now be artiﬁcially elevated by the artiﬁcial promoter. We also noticed that the eGFP expression level from MSeGFH and MSeGF strains were higher than that from MTBeGFH and MTBeGF, which may be related to their metabolism states (Fig. S3A). Even the same batch of recombinant strain in different states could express eGFP in very different levels according to our observations. 3.6. Sequential use of the dif-ΩHYG-dif cassette in a single strain The integrative plasmid pTYGi9 (Fig. 2A) was introduced into MSD56U, the unmarked MS mutants with the Ms0615–0616 disrupted using CST1. The positive transformants were veriﬁed by PCR using primers Hygdf and Hygdr (Table 1) and designated as MS-D56HGFP. By counter selection, we obtained the unmarked MS mutants with the Ms0615–0616 gene deletion and the eGFP gene integrated in the genome. This indicated that the CSTI cassette could be used more than once to create unmarked mutations in a single strain. It needs to be pointed out that integration of multiple genes into the mycobacterial genome could be accomplished in various ways which includes: i) delivering multiple genes in one plasmid; ii) integration of an integrative plasmid carrying the dif-ΩHYG-dif cassette, the target genes and an attB site of mycobacteriophage L5 for the following integration of an additional mycobacteriophage L5-based integrative plasmid carrying some other target genes (Saviola and Bishai, 2004); and iii) by serial integration of the dif-ΩHYG-dif cassette containing integrative plasmids derived from different mycobacteriophages, as they could recognize and integrate into different attB sites located at the same mycobacterial genome (Pham et al., 2007; Morris et al., 2008). 4. Conclusions and perspectives We have created more improved cassettes based on the Xer-cise system for constructing unmarked mycobacteria and demonstrated that they are very efﬁcient in this genus. 1) The modiﬁed dif-ΩHYG-dif cassettes (the gene and its surrounding regions used for genetic manipulation) are much shorter and easier for identiﬁcation by PCR ampliﬁcation than the original one. Its potential natural promoters were replaced by an artiﬁcial promoter fused with the enzyme sites and dif sequence. The region following the stop codon was also removed. So its length is only about 2/3 of that of the original one. The new cassettes can be manipulated easier because of not only the shorter length, but also its easiness for PCR ampliﬁcation due to elimination of the potential higher structure(s). 2) The EcoRI is a widely used enzyme in genetic engineering. The inherent EcoRI site in the original Hyg gene was removed, so it can be used easily in more cases. 3) Multi-cloning sites were added at both ends of the Hyg cassette to make it more suitable for cloning and for in-frame deletion even with directional cloning without polar effect. 4) The fragments produced by isocaudarners of ClaI or SmaI (producing blunt ends) can be inserted into the dif-ΩHYG-dif cassettes and can then be removed together with the Hyg gene from the mycobacterial genome. This is useful in the instance that some genes/fragments need to be eliminated after their integration into the genome. 5) Reuse of the same cassette in the same strain was demonstrated for the ﬁrst time. The artiﬁcial dif-ΩHYG-dif cassettes described here and their possible derivatives can be versatile, very useful, convenient and effective in construction of unmarked mycobacteria, for example, construction of unmarked reporter mycobacterial strains and unmarked in-frame deletion gene knock out strains without polar effect. It is also possible to
F. Yang et al. / Journal of Microbiological Methods 103 (2014) 29–36
construct unmarked live vaccines using mycobacteria by introducing some gene(s) encoding protective antigen(s) in the mycobacterial genome with an integrative plasmid containing the dif-ΩHYG-dif cassette or by disruption of the gene(s) essential for virulence for engineering attenuated live vaccine strains using this cassette or its derivatives. Conveniently, serial disruption of multiple genes at different loci of the mycobacterial genome with the dif-ΩHYG-dif cassette from a single strain was demonstrated, which is very important in construction of attenuated mycobacteria and in the study of synergistic gene action and the functions of redundant genes. PCR ampliﬁcation was used to conﬁrm that the CST1 or CST2 cassette was integrated into mycobaterial genome or eliminated. This technique is more deﬁnitive and convenient if designed carefully than Southern blot analysis (van Kessel and Hatfull, 2007; Stratigopoulos and Cundliffe, 2002). The KAN resistance gene could possibly be inserted in between dif sequences surrounded with multi-cloning sites, just like the dif-ΩHYG-dif cassette described herein. The thermosensitive plasmid pJV53Ts for recombineering can be more useful in mycobacterial gene knock-out and genome modiﬁcation because of its much easier removal from the host than its previous version. This has also just been demonstrated in a very recent paper (Shenkerman et al., 2014) using the similar method. The unmarked Ms0615–0616 gene deletion mutants could be good candidate vectors for developing potential vaccines against many diseases, as the esx-3 operon disrupted MS mutant was reported to have no virulence even in several types of immune deﬁcient mice, and still grew as well as its parent strain (Sweeney et al., 2011). Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.mimet.2014.05.007. Acknowledgments This work was supported by the Chinese Academy of Sciences ‘One Hundred Talents Program’ (Category A, to T.Z.), Main Direction Program of the Chinese Academy of Sciences (KSCX2-EW-J27), the National Great Research Program of China (2013ZX10003006) and the Key Program of the Chinese Academy of Sciences (KJZD-EW-L02). We are grateful to Professor Christophe Guilhot from Institut Pasteur, for providing us with thermosensitive plasmid pPR23. We thank Professor Riccardo Manganelli from University of Padua, Italy, for providing us the plasmid pJV53 as a gift. We thank Professor Eric Nuermberger, William Bishai and Jacques Grosset from the Johns Hopkins University for providing plasmids such as pNBV1, pCK0601 and pblueINT. We thank Professor Jiaoyu Deng at Wuhan Institute of Virology, Chinese Academy of Sciences for providing us with the M. smegmatis mc2 155. References Agarwal, N., Lamichhane, G., Gupta, R., Nolan, S., Bishai, W.R., 2009. Cyclic AMP intoxication of macrophages by a Mycobacterium tuberculosis adenylate cyclase. Nature 460, 98–102. Cascioferro, A., Boldrin, F., Seraﬁni, A., Prowedi, R., Palu, G., Manganelli, 2010. Xer sitespeciﬁc recombination, an efﬁcient tool to introduce unmarked deletions into mycobacteria. Appl. Environ. Microbiol. 76, 5312–5316. Garbe, T.R., Barathi, J., Barnini, S., Zhang, Y., Abu-Zeid, C., Tang, D., Mukherjee, R., Young, D. B., 1994. Transformation of mycobacterial species using hygromycin resistance as selectable marker. Microbiology 140, 133–138. Garnier, C., Churcher, D., Harris, S., Gordon, V., Eiglmeier, K., Gas, S.C.E., Barry, C.F., Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T., Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin, N., Holroyd, S., Hornsby, T., Jagels, K., Krogh, A., McLean, J., Moule, S., Murphy, L., Oliver, K., Osborne, J., Quail, M.A., Rajandream, M.A., Rogers, J., Rutter, S., Seeger, K., Skelton, J., Squares, R., Squares, S., Sulston, J.E., Taylor, K., Whitehead, S., Barrell, B.G., 1998. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544.
Howard, N.S., Gomez, J.E., Ko, C., Bishai, W.R., 1995. Color selection with a hygromycinresistance-based Escherichia coli-mycobacterial shuttle vector. Gene 166, 181–182. Jacobs, W.R., Tuckman, M., Bloom, B.R., 1987. Introduction of foreign DNA into mycobacteria using a shuttle phasmid. Nature 327, 532–535. Lee, M.H., Pascopella, L., Jacobs, W.R., Hatfull, G.F., 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. Loewenberg, S., 2012. India reports cases of totally drug-resistant tuberculosis. Lancet 379, 205. Mahenthiralingam, E., Draper, P., Davis, E.O., Colston, M.J., 1993. Cloning and sequencing of the gene which encodes the highly inducible acetamidase of Mycobacterium smegmatis. J. Gen. Microbiol. 139, 575–583. Malaga, W., Perez, E., Guilhot, C., 2003. Production of unmarked mutations in mycobacteria using site-speciﬁc recombination. FEMS Microbiol. Lett. 219, 261–268. Manabe, Y.C., Hatem, C.L., Kesavan, A.K., Durack, J., Murphy, J.R., 2005. Both Corynebacterium diphtheriae DtxR(E175K) and Mycobacterium tuberculosis IdeR(D177K) are dominant positive repressors of IdeR-regulated genes in M. tuberculosis. Infect. Immun. 73, 5988–5994. Morris, P., Marinelli, L.J., Jacobs-Sera, D., Hendrix, R.W., Hatfull, G.F., 2008. Genomic characterization of mycobacteriophage Giles: evidence for phage acquisition of host DNA by illegitimate recombination. J. Bacteriol. 190, 2172–2182. Nahid, P., Menzies, D., 2012. Update in tuberculosis and nontuberculous mycobacterial disease 2011. Am. J. Respir. Crit. Care Med. 185, 1266–1270. Parish, T., Stoker, N.G., 2000. Use of a ﬂexible cassette method to generate a double unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement. Microbiology 146, 1969–19755. Parish, T., Mahenthiralingam, E., Draper, P., Davis, E.O., Colston, E.O., 1997. Regulation of the inducible acetamidase gene of Mycobacterium smegmatis. Microbiology 143, 2267–2276. Pelicic, V., Jackson, M., Reyrat, J.-C., Jacobs, W.R., Gicquel, B., Guilhot, C., 1997. Efﬁcient allelic exchange and transposon mutagenesis in Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. U. S. A. 94, 10955–10960. Pham, T.T., Jacobs-Sera, D., Pedulla, M.L., Hendrix, R.W., Hatfull, G.F., 2007. Comparative genomic analysis of mycobacteriophage Tweety: evolutionary insights and construction of compatible site-speciﬁc integration vectors for mycobacteria. Microbiology 153, 2711–2723. Puech, V., Guilhot, C., Perez, E., Tropis, M., Armitige, L.Y., Gicquel, B., Daffe, M., 2002. Evidence for a partial redundancy of the ﬁbronectin-binding proteins for the transfer of mycoloyl residues onto the cell wall arabinogalactan termini of Mycobacterium tuberculosis. Mol. Microbiol. 44, 1109–1122. Saviola, B., Bishai, W.R., 2004. Method to integrate multiple plasmids into the mycobacterial chromosome. Nucleic Acids Res. 32, 1. Shenkerman, Y., Elharar, Y., Vishkautzana, M., Gur, E., 2014. Efﬁcient and simple generation of unmarked gene deletions in Mycobacterium smegmatis. Gene 533, 374–378. Stratigopoulos, G., Cundliffe, E., 2002. Expression analysis of the tylosin-biosynthetic gene cluster: pivotal regulatory role of the thlQ product. Chem. Biol. 9, 71–78. Sweeney, K.A., Dao, D.N., Goldberg, M.F., Hsu, T., Venkataswamy, M.M., Henao-Tamayo, M., Ordway, D., Sellers, R.S., Jain, P., Chen, B., Chen, M., Kim, J., Lukose, R., Chan, J., Orme, I. M., Porcelli, S.A., Jacobs, W.R., 2011. A recombinant Mycobacterium smegmatis induces potent bactericidal immunity against Mycobacterium tuberculosis. Nat. Med. 17, 1261–1268. van Kessel, J.C., Hatfull, G.F., 2007. Recombineering in Mycobacterium tuberculosis. Nat. Methods 4, 147–152. van Kessel, J.C., Hatfull, G.F., 2008. Mycobacterial recombineering. Methods Mol. Biol. 435, 203–215. van Kessel, J.C., Marinelli, L.J., Hatfull, G.F., 2008. Recombineering mycobacteria and their phages. Nat. Rev. Microbiol. 6, 851–857. Zhang, T., Wang, L., Xu, G., Chen, Y., Zhang, Y., Li, Y., 2006a. Disruption of ste23 gene affects composition proﬁle and bioactivity of exopolysaccharide produced by Streptomyces sp. 139. Lett. Appl. Microbiol. 42, 132–137. Zhang, T., Wang, L., Xu, G., Chen, Y., Zhang, Y., Li, Y., 2006b. Disruption of the ste22 gene encoding a glycosyltransferase and its function in biosynthesis of Ebosin in Streptomyces sp. 139. Curr. Microbiol. 52, 55–59. Zhang, T., Bishai, W.R., Grosset, J.H., Nuermberger, E.R., 2010. Rapid assessment of antibacterial activity against Mycobacterium ulcerans by using recombinant luminescent strains. Antimicrob. Agents Chemother. 54, 2806–2813. Zhang, T., Li, S.Y., Nuermberger, E.L., 2012. Autoluminescent Mycobacterium tuberculosis for rapid, real-time, non-invasive assessment of drug and vaccine efﬁcacy. PLoS One 7, e29774. Zhang, T., Li, S.-Y., Converse, P.J., Grosset, J.H., Nuermberger, E.L., 2013. Rapid, serial, noninvasive assessment of drug efﬁcacy in mice with autoluminescent Mycobacterium ulcerans infection. PLoS Negl. Trop. Dis. 7, e2598.