UNLABELLED: Specialized transduction has proven to be useful for generating deletion mutants in most mycobacteria, including virulent Mycobacterium tuberculosis. We have improved this system by developing (i) a single-step strategy for the construction of allelic exchange substrates (AES), (ii) a temperature-sensitive shuttle phasmid with a greater cloning capacity than phAE87, and (iii) bacteriophage-mediated transient expression of site-specific recombinase to precisely excise antibiotic markers. The methods ameliorate rate-limiting steps in strain construction in these difficult-to-manipulate bacteria. The new methods for strain construction were demonstrated to generalize to all classes of genes and chromosomal loci by generating more than 100 targeted single- or multiple-deletion substitutions. These improved methods pave the way for the generation of a complete ordered library of M. tuberculosis null strains, where each strain is deleted for a single defined open reading frame in M. tuberculosis. IMPORTANCE: This work reports major advances in the methods of genetics applicable to all mycobacteria, including but not limited to virulent M. tuberculosis, which would facilitate comparative genomics to identify drug targets, genetic validation of proposed pathways, and development of an effective vaccine. This study presents all the new methods developed and the improvements to existing methods in an integrated way. The work presented in this study could increase the pace of mycobacterial genetics significantly and will immediately be of wide use. These new methods are transformative and allow for the undertaking of construction of what has been one of the most fruitful resources in model systems: a comprehensive, ordered library set of the strains, each of which is deleted for a single defined open reading frame.
UNLABELLED: Specialized transduction has proven to be useful for generating deletion mutants in most mycobacteria, including virulent Mycobacterium tuberculosis. We have improved this system by developing (i) a single-step strategy for the construction of allelic exchange substrates (AES), (ii) a temperature-sensitive shuttle phasmid with a greater cloning capacity than phAE87, and (iii) bacteriophage-mediated transient expression of site-specific recombinase to precisely excise antibiotic markers. The methods ameliorate rate-limiting steps in strain construction in these difficult-to-manipulate bacteria. The new methods for strain construction were demonstrated to generalize to all classes of genes and chromosomal loci by generating more than 100 targeted single- or multiple-deletion substitutions. These improved methods pave the way for the generation of a complete ordered library of M. tuberculosis null strains, where each strain is deleted for a single defined open reading frame in M. tuberculosis. IMPORTANCE: This work reports major advances in the methods of genetics applicable to all mycobacteria, including but not limited to virulent M. tuberculosis, which would facilitate comparative genomics to identify drug targets, genetic validation of proposed pathways, and development of an effective vaccine. This study presents all the new methods developed and the improvements to existing methods in an integrated way. The work presented in this study could increase the pace of mycobacterial genetics significantly and will immediately be of wide use. These new methods are transformative and allow for the undertaking of construction of what has been one of the most fruitful resources in model systems: a comprehensive, ordered library set of the strains, each of which is deleted for a single defined open reading frame.
Tuberculosis (TB) is a major global health challenge, with approximately 8 million new cases and almost 2 million deaths per year (1). Genetic analysis of the causative organism Mycobacterium tuberculosis is essential to furthering the prevention and treatment of TB. A powerful genetic approach to characterize gene function is to construct defined null alleles in an organism, so that isogenic strains, differing at the defined locus, can be compared (2, 3). M. tuberculosis has a doubling time of approximately 24 h, and colony formation from a single cell takes 3 to 4 weeks. Slow growth means that it takes at least a month for each step of multistep genetic manipulations. Thus, there is particular value in methods that decrease the number of steps that require the growth of a single cell into a colony. The approaches taken with this challenging system may also interest those working with other “noncanonical” organisms whose genetic manipulation and laboratory domestication will require new approaches.In this study, we identified and improved upon several rate-limiting steps for the construction of deletion-substitution and unmarked deletion alleles in mycobacteria. First, the multistep process of engineering allelic exchange substrates (AES) for homologous recombination (HR) has been shortened to a single step. The new AES vector also contains both a hygromycin resistance gene as a selection and a sacB gene as a counterselection flanked by gamma delta resolvase sites for simplified unmarking. Second, an improved shuttle, phAE159, with an increased cloning capacity and efficient DNA delivery in comparison to phAE87, was constructed. Third, we have developed a phage-based transient expression of γδ resolvase methodology that allows recovery of an antibiotic-sensitive deletion strain in a single step.We have used these methods to construct a fully drug-sensitive double auxotroph of M. tuberculosis H37Rv (mc2 6206) which displays less virulence in mouse studies than the current human vaccine strain Bacillus Calmette-Guérin (BCG). The mc2 6206 strain is an excellent substrate for further genetic manipulation, as this M. tuberculosis derivative has been reclassified for biosafety level 2 (BSL2) containment. These methods have also been used to generate over 100 additional strains with defined deletion-substitution in virulent M. tuberculosis H37Rv to demonstrate that the methods apply to all classes of genes and at all chromosomal locations.
RESULTS
One-step construction of AES.
The left homology sequence (LHS) and right homology sequence (RHS) flanking the region targeted for deletion-substitution need to be cloned on either side of a selectable/counterselectable cassette [γδ(sacB-hyg)γδ] to generate an AES for HR (Fig. 1A). Current directional cloning approaches to construct AES require different restriction enzymes at both the ends of each flanking sequence (4, 5). Besides primer design complexity and the number of consecutive steps, this approach is often confounded by the presence of restriction sites, needed for cloning, within the genomic region to be cloned. We have exploited the property a type IIP restriction enzyme Van91I (isoschizomers PflmI, AccB7I, and BasI) (6), which recognizes a discontinuous palindrome interrupted by a segment of 5 bases of unspecified sequence (CCAN_NNN^NTGG), to accomplish a 4-fragment oriented ligation. Using this approach, we were able to construct the AES in a single reaction (Fig. 1A). Van91I cleaves within the degenerate sequence, allowing us to dictate the sequence of sticky ends. This methodology to generate one-step AES is applicable to all the genes in M. tuberculosis and other mycobacterial species. DraIII, BstAp1, and AlwNI restriction enzymes can also be used in place of Van91I to construct AES in one step using the same methodology (Fig. 1B).
FIG 1
High-throughput construction of AES via four-fragment oriented ligation. (A) Regions of 650 to 1,000 bases that flank the region to be deleted (Your Favorite Gene [yfg], M. tuberculosis recD in this case) were amplified with primers harboring the Van91I site at the 5′ end to obtain left homology sequence (LHS) and right homology sequence (RHS). (i) Van91I recognizes a 6-bp symmetric bipartite sequence bisected with 5 bp of degenerate sequence (CCAN_NNN^NTGG). Van91I digestion results in 3′ overhangs of three bases, which are shown at each end of the two PCR products. The sequences of the overhangs depend on the intervening degenerate sequences and are specified in the primer shown below. (ii) Complementary 3′ overhangs at the end of the 3.6-kb linear fragment [contains γδ(sacB-hyg)γδ cassette] and a 1.6-kb linear fragment (contains the lambda cos site, ColE1 origin of replication [oriE], and an 8-bp recognition site for PacI) were obtained after Van91I digestion of pYUB1471 to specify the site and orientation for the four fragments. (iii) Ligation of four different DNA fragments, Van91I-digested LHS, RHS, and 3.6-kb and 1.6-kb linear fragments, results in AES (pYUB1363, AES for recD). All the AES are approximately 6- to 7.5-kb plasmids which are propagated as plasmids in E. coli and cloned into mycobacteriophage vectors via their unique PacI site for generating STPs. (B) The 5′ sequence for primers depends upon the choice of restriction enzyme. If the sequence within the PCR amplicon contains a Van91I recognition site, BstAPI, AlwNI, or DraIII is used instead. The overhangs used in each case (LL, LR, RL, RR) are shown for all four restriction enzymes. The gene-specific sequence is indicated by N18-20 at the ends of primers. Any combination of LL/LR and RL/RR primers (i.e., LL/RR with Van91I overhangs and RL/RR primers with DraIII overhangs) can be used to generate AES.
High-throughput construction of AES via four-fragment oriented ligation. (A) Regions of 650 to 1,000 bases that flank the region to be deleted (Your Favorite Gene [yfg], M. tuberculosis recD in this case) were amplified with primers harboring the Van91I site at the 5′ end to obtain left homology sequence (LHS) and right homology sequence (RHS). (i) Van91I recognizes a 6-bp symmetric bipartite sequence bisected with 5 bp of degenerate sequence (CCAN_NNN^NTGG). Van91I digestion results in 3′ overhangs of three bases, which are shown at each end of the two PCR products. The sequences of the overhangs depend on the intervening degenerate sequences and are specified in the primer shown below. (ii) Complementary 3′ overhangs at the end of the 3.6-kb linear fragment [contains γδ(sacB-hyg)γδ cassette] and a 1.6-kb linear fragment (contains the lambda cos site, ColE1 origin of replication [oriE], and an 8-bp recognition site for PacI) were obtained after Van91I digestion of pYUB1471 to specify the site and orientation for the four fragments. (iii) Ligation of four different DNA fragments, Van91I-digested LHS, RHS, and 3.6-kb and 1.6-kb linear fragments, results in AES (pYUB1363, AES for recD). All the AES are approximately 6- to 7.5-kb plasmids which are propagated as plasmids in E. coli and cloned into mycobacteriophage vectors via their unique PacI site for generating STPs. (B) The 5′ sequence for primers depends upon the choice of restriction enzyme. If the sequence within the PCR amplicon contains a Van91I recognition site, BstAPI, AlwNI, or DraIII is used instead. The overhangs used in each case (LL, LR, RL, RR) are shown for all four restriction enzymes. The gene-specific sequence is indicated by N18-20 at the ends of primers. Any combination of LL/LR and RL/RR primers (i.e., LL/RR with Van91I overhangs and RL/RR primers with DraIII overhangs) can be used to generate AES.
Construction of shuttle phasmid phAE159 and specialized transducing phages (STPs).
Mycobacteriophage vector phAE159 was constructed by deletion of an ~6-kb region encompassing genes gp48 to gp64 from a temperature-sensitive mutant of TM4, PH101 (7, 8) (see Fig. S1 in the supplemental material). Similar to parent PH101, shuttle phasmid phAE159 propagates as a phage at 30°C but only injects its DNA and does not propagate in the infected mycobacterial cell at 37°C. The shuttle phasmid phAE159 delivers up to 10 kb of recombinant DNA into mycobacterial cells.Mycobacteriophage-based STPs harboring AES at the nonessential region in phAE159 were used to disrupt a specific gene in the mycobacterial chromosome via genomic deletion-substitution (4, 9, 10). For each STP construction, the carbenicillin cassette between the two PacI sites in phAE159 phasmid DNA was replaced by the desired AES targeting a specific region of the bacterial genome (Fig. 2). All the AES were hygromycin resistant, and the process of STP construction was simplified by the selection for an antibiotic marker (hygromycin) different from the original phAE159 marker (carbenicillin) and the use of in vitro lambda packaging to size select phasmid DNA (Fig. 2). Phasmid DNA was electroporated into Mycobacterium smegmatis mc2 155 to obtain STP plaques at the permissive temperature of 30°C. STPs were picked and amplified at 30°C to obtain high-titer allele-specific STPs.
FIG 2
Construction of gene-specific STPs and high-throughput amplification of phage stocks to generate deletion-substitution mutants of M. tuberculosis by specialized transduction. (i) phAE159 DNA is digested with PacI to release the ampicillin resistance insert and was ligated to PacI-digested AES (pYUB1363). The resulting ligation product is hygromycin resistant. (ii) The ligated products are packaged in vitro via the lambda cos site present in the AES. (iii) E. coli HB101 was transduced with lambda-packaged STP and selected for hygromycin resistance. (iv) STP phasmid DNA was isolated from HB101 as plasmid, transfected into mc2155 at 30°C to obtain mycobacterial phage plaques for STP. (v) High-titer phage stocks of each STP were grown at 30°C in a 24-well high-throughput format on a lawn of mc2155. Phage stocks were grown on an agar pad of 1 ml, essentially a miniaturized plate stock as described before (21). Transduction-competent M. tuberculosis or other recipient cells were incubated with the high-titer STP stock and incubated for 24 h at 37°C. (vi) Transduced M. tuberculosis recipients were plated on 7H10 supplemented with hygromycin (75 µg/ml) and incubated 3 to 4 weeks at 37°C to obtain deletion-substitution mutants. The colonies obtained were confirmed for HR using PCR primers upstream of LHS and downstream of RHS (shown as arrow). Expected-size products were obtained for WT (~2.9 kb) and knockout (~5.0 kb). The Invitrogen 1Kb Plus DNA ladder was used for the agarose gel, and all the bands differ by 1 kb.
Construction of gene-specific STPs and high-throughput amplification of phage stocks to generate deletion-substitution mutants of M. tuberculosis by specialized transduction. (i) phAE159 DNA is digested with PacI to release the ampicillin resistance insert and was ligated to PacI-digested AES (pYUB1363). The resulting ligation product is hygromycin resistant. (ii) The ligated products are packaged in vitro via the lambda cos site present in the AES. (iii) E. coli HB101 was transduced with lambda-packaged STP and selected for hygromycin resistance. (iv) STP phasmid DNA was isolated from HB101 as plasmid, transfected into mc2155 at 30°C to obtain mycobacterial phage plaques for STP. (v) High-titer phage stocks of each STP were grown at 30°C in a 24-well high-throughput format on a lawn of mc2155. Phage stocks were grown on an agar pad of 1 ml, essentially a miniaturized plate stock as described before (21). Transduction-competent M. tuberculosis or other recipient cells were incubated with the high-titer STP stock and incubated for 24 h at 37°C. (vi) Transduced M. tuberculosis recipients were plated on 7H10 supplemented with hygromycin (75 µg/ml) and incubated 3 to 4 weeks at 37°C to obtain deletion-substitution mutants. The colonies obtained were confirmed for HR using PCR primers upstream of LHS and downstream of RHS (shown as arrow). Expected-size products were obtained for WT (~2.9 kb) and knockout (~5.0 kb). The Invitrogen 1Kb Plus DNA ladder was used for the agarose gel, and all the bands differ by 1 kb.
Specialized transduction.
Specialized transduction and subsequent incubations were performed at 37°C, a temperature nonpermissive for the amplification of these STPs. At this temperature, STPs inject their DNA into mycobacterial cells but do not propagate and thus provide AES for HR. The transduced M. tuberculosis cells were selected on hygromycin plates to obtain colonies from M. tuberculosis cells that had undergone HR (Fig. 2). On average, we obtained 10 to 30 colonies on hygromycin plates after transduction, and >10 or no colonies appeared only in few cases. In these latter cases, transductions were repeated after retitration of the STP. Three colonies per transduction were analyzed either by PCR or Southern blotting to differentiate HR from illegitimate recombination. In most transductions, either all three or two-thirds of the colonies were the result of HR. Transductions for Rv0404 and Rv3801c, however, yielded no colonies despite repeated attempts (n = 5) and could be deleted only when transductions (n = 1) were performed over merodiploid strains expressing Rv0404 (mc2 7212) or Rv3801c (mc2 7241), suggesting their essentiality (see Fig. S2 in the supplemental material). To show the ability of STPs to provide substrate for HR at different genomic locations, we have generated more than 100 Southern blotting/PCR-confirmed deletion-substitution mutants in M. tuberculosis H37Rv by specialized transduction (Fig. 3A). The alteration of a desired locus by specialized transduction is not restricted to genes belonging to one functional group. The majority of the genes deleted in this study have been annotated as lipid metabolism (34%) and cell wall and cell processes (21%), with the rest of the genes showing a diverse distribution of functional classes (Fig. 3B).
FIG 3
High-throughput generation of gene deletion in mycobacteria. (A) List of 104 mutants generated by specialized transduction in M. tuberculosis H37Rv. (B) Functional categories of genes in virulent M. tuberculosis H37Rv, which were subjected to deletion-substitution via specialized transduction. List of primers to generate the AES are listed in Table S1 in the supplemental material.
High-throughput generation of gene deletion in mycobacteria. (A) List of 104 mutants generated by specialized transduction in M. tuberculosis H37Rv. (B) Functional categories of genes in virulent M. tuberculosis H37Rv, which were subjected to deletion-substitution via specialized transduction. List of primers to generate the AES are listed in Table S1 in the supplemental material.
Phage-based transient expression system for gamma-delta resolvase.
There are practical and scientific needs to remove the antibiotic resistance marker from the mutant strains. The practical need is that there is only one other class of antibiotic resistance marker (kanamycin) available for mycobacteria, which has an unacceptable spontaneous resistance rate (11). Therefore, the hygromycin cassette needs to be excised out of the first mutant strain so that the hygromycin-marked AES can again be used for the iterative construction of multiple mutant strains. The scientific need is that insertion of a selectable marker cassette could also be like a transposon insertion in terms of polarity, so being able to specifically remove the selection cassette to generate an in-frame deletion would reduce the polar effects of the manipulation carried out to make a deletion. Removal of the antibiotic resistance cassette in a deletion-substitution strain requires the expression of a site-specific resolvase (4, 12). We hypothesized that the temperature-sensitive mycobacteriophage vector phAE159 at 37°C would provide a platform for the transient expression of cloned genes into the transduced mycobacterial cells and could be used for the excision of γδ(sacB-hyg)γδ cassette from deletion-substitution mutants generated as described above. Because the phAE159 is a temperature-sensitive phage, it is readily lost at 37°C and thus does not require curing after unmaking, contrarily to the resolvase-expressing plasmids (4, 12). To test our hypothesis, the tnpR gene (γδ resolvase) from the γδ transposon was cloned into phAE159 to generate phAE280 (Fig. 4A). Mycobacterial strains harboring the γδ(sacB-hyg) γδ cassette were incubated with phAE280 and plated on sucrose-containing plates to counterselect for the loss of the sacB-hyg region between two γδ sites (Fig. 4B). Colonies obtained on sucrose plates were scored for hygromycin sensitivity, and resolution of γδ(sacB-hyg)γδ by the γδ resolvase were confirmed by PCR and sequencing. Figure 4C shows one such example of this analysis from M. tuberculosis. The results for this particular strain to excise the sacB-hyg cassette by γδ resolvase were similar to other unmarked knockouts in different classes of genes (data not shown). The residual sequence after recombination between the two γδ sequences flanking sacB-hyg is shown in Fig. 4D. The primers can be designed to generate an in-frame deletion irrespective of the restriction enzyme used to generate the AES. The amino acid sequence of the residues for an in-frame deletion is indicated below the nucleotide sequence (Fig. 4D). These methods to generate unmarked deletion mutants are also applicable to M. smegmatis, Mycobacterium bovis, and various M. tuberculosis strains. More than 80% of colonies recovered from sucrose plates after transduction by phAE280, in more than 40 independent experiments, were sensitive to hygromycin and had the expected residual excision sequence in all the cases, regardless of the targeted genomic locus. The remaining 20% were bypass mutants (i.e., sacB mutants or sucrose resistance mutants).
FIG 4
Phage-mediated unmarking. (A) Schematic of the mycobacteriophage expressing γδ resolvase (phAE280); (B) flow chart description of single-step excision of the γδ(sacB-hyg)γδ cassette using phAE280. (C) mc27204 [M. tuberculosis H37Rv ΔrecD::γδ(sacB-hyg)γδ] was transduced with phAE280, using the protocol listed above, to excise the selection/counterselection cassette and yield mc27205 (ΔrecD::γδ). Hygromycin sensitivity was confirmed by picking and patching (left); deletion was confirmed by PCR with primers that flank the targeted region (right, blue arrows). Lane 1, PCR on M. tuberculosis cells (wild type [WT]); lane 2, PCR on mc27204 [ΔrecD::γδ(sacB-hyg)γδ; hygromycin resistant]; lanes 3 to 9, PCR on hygromycin-sensitive colonies from pick and patch plate, unmarked ΔrecD mutant (mc27205 [ΔrecD::γδ]). The excision of the sacB-hyg cassette was also confirmed by sequencing of PCR products. (D) Generation of the in-frame deletions. The two γδ recombine to excise the sacB-hyg cassette after the expression of γδ resolvase from phAE280. The residual nucleotide sequence after the excision is shown. The encoded amino acids are shown below the nucleotide sequence. The amino acid at the junction depends upon the restriction enzyme used as shown. The primer sequence of LR and RL should match the coding frame to maintain the reading frame to generate in-frame deletion. The amino acid at the RL-restriction enzyme junction depends upon the gene-specific primer sequence (N18-20) and is indicated by X.
Phage-mediated unmarking. (A) Schematic of the mycobacteriophage expressing γδ resolvase (phAE280); (B) flow chart description of single-step excision of the γδ(sacB-hyg)γδ cassette using phAE280. (C) mc27204 [M. tuberculosis H37Rv ΔrecD::γδ(sacB-hyg)γδ] was transduced with phAE280, using the protocol listed above, to excise the selection/counterselection cassette and yield mc27205 (ΔrecD::γδ). Hygromycin sensitivity was confirmed by picking and patching (left); deletion was confirmed by PCR with primers that flank the targeted region (right, blue arrows). Lane 1, PCR on M. tuberculosis cells (wild type [WT]); lane 2, PCR on mc27204 [ΔrecD::γδ(sacB-hyg)γδ; hygromycin resistant]; lanes 3 to 9, PCR on hygromycin-sensitive colonies from pick and patch plate, unmarked ΔrecD mutant (mc27205 [ΔrecD::γδ]). The excision of the sacB-hyg cassette was also confirmed by sequencing of PCR products. (D) Generation of the in-frame deletions. The two γδ recombine to excise the sacB-hyg cassette after the expression of γδ resolvase from phAE280. The residual nucleotide sequence after the excision is shown. The encoded amino acids are shown below the nucleotide sequence. The amino acid at the junction depends upon the restriction enzyme used as shown. The primer sequence of LR and RL should match the coding frame to maintain the reading frame to generate in-frame deletion. The amino acid at the RL-restriction enzyme junction depends upon the gene-specific primer sequence (N18-20) and is indicated by X.
An exemplar for a multigenic deletion mutant: a BSL2-safe M. tuberculosis H37Rv derivative.
WHO guidelines mandate that laboratory work with live M. tuberculosis can be carried outside BSL3 containment only with strains that harbor two unlinked nonreverting attenuating mutations (13). We have previously shown that the deletion of panCD, which results in pantothenate auxotroph, or deletion of leuD, which results in a leucine auxotroph in M. tuberculosis, leads to bacterial attenuation in mice (5, 14) and that immunization with these strains of M. tuberculosis elicits some protective immunity against virulent M. tuberculosis. First, we deleted the panCD gene in the M. tuberculosis genome using specialized transduction. The resulting strain was unmarked using phAE280 as described above. The antibiotic-sensitive M. tuberculosis ΔpanCD strain was then used as a substrate for the deletion of the leuCD locus. The sacB-hyg cassette inserted at the ΔleuCD locus was excised again using phAE280, and the M. tuberculosis strain was designated mc26206. This strain is auxotrophic for both leucine and pantothenate and is fully antibiotic sensitive. The kinetics of persistence of this attenuated strain in C57BL/6 mice was similar to that of the currently used human vaccine BCG (Fig. 5A). The safety of mc26206 was further assessed by infecting severe combined immune-deficient (SCID) mice intravenously with a high dose (105 CFU) of bacteria. SCIDmice succumbed to virulent M. tuberculosis H37Rv infection 25 days after infection, while mice challenged with BCG died by day 230. Those challenged with mc26206 had a survival rate of 100% by day 250 (Fig. 5B), showing that the mc26206 is safer than M. bovis BCG in SCIDmice. These observations have led to the approval from the Albert Einstein College of Medicine Institutional Biosafety Committee for reclassification of this strain for BSL2 containment. This strain is being used as a base strain to develop hybrid vaccines against TB and human immunodeficiency virus (HIV) or to study M. tuberculosis physiology (see Discussion).
FIG 5
ΔleuCD ΔpanCD deletions attenuate virulence of M. tuberculosis H37Rv. (A) Kinetics of persistence of M. bovis BCG SSI Danish and the M. tuberculosis H37Rv ΔleuCD ΔpanCD strain (mc26206) in immunocompetent mice. C57BL/6 mice were intravenously (i.v.) infected with approximately 106 CFU of M. bovis BCG or mc26206. Organs were harvested at different time points as indicated. (B) Ten mice were i.v. infected with approximately 105 CFU M. bovis BCG SSI or mc26206 and their survival times were recorded. Results are representative of three independent experiments.
ΔleuCD ΔpanCD deletions attenuate virulence of M. tuberculosis H37Rv. (A) Kinetics of persistence of M. bovis BCG SSI Danish and the M. tuberculosis H37Rv ΔleuCD ΔpanCD strain (mc26206) in immunocompetent mice. C57BL/6 mice were intravenously (i.v.) infected with approximately 106 CFU of M. bovis BCG or mc26206. Organs were harvested at different time points as indicated. (B) Ten mice were i.v. infected with approximately 105 CFU M. bovis BCG SSI or mc26206 and their survival times were recorded. Results are representative of three independent experiments.
DISCUSSION
Rate-limiting steps for strain construction in mycobacteria, including virulent M. tuberculosis, have been identified and partially ameliorated. More than a hundred distinct loci at diverse chromosomal locations have been engineered in virulent H37Rv. The methods described are not restricted to any particular region of the chromosome or any functional category of genes and appear broadly applicable to different species and strains of mycobacteria. M. tuberculosis H37Rv, M. tuberculosis CDC1551, M. tuberculosis clinical isolates, and several M. bovis and M. smegmatis strains have all proven amenable to these methods. Multiple, up to 14, adjacent reading frames, which span 16.5 Mb and two operons, have been deleted in a single step (C. Vilchèze, unpublished data). The same AES generated by one-step cloning can be used for recombineering (15) or cloned into phAE159 to obtain STP. The STPs can be amplified by growing a phage stock. In most cases, the same STP can be used in multiple strains or species of mycobacteria which often harbor 95% or more sequence similarity (e.g., between M. bovis and various M. tuberculosis strains). The ability to use the same STP in different species will facilitate study of the subtle variations in M. bovis and BCG and M. tuberculosis physiology (16).Transposon mutagenesis has been conducted at a large scale appropriate for saturation for various M. tuberculosis strains (17). TnHimar1 insertion mutants are available for many genes from the Target website (http://webhost.nts.jhu.edu/target/) and are valuable tools for mycobacterial genetics. As informative as transposon insertions are, construction of defined null deletion by the methods described in this work has advantages. (i) The endpoints of these deletion-substitutions are precisely engineered and do not include the pseudo-random aspect of transposon insertions. (ii) Transposons must be mapped after the fact, whereas AES used in specialized transduction are mapped before by the choice of sequences for replacement and are verified by a simple PCR (Fig. 2). (iii) The antibiotic resistance cassette can be removed in a single step (Fig. 4); there is no easy way to remove the drug marker from transposon insertions. (iv) About half the genes in M. tuberculosis are in operons (18), and transposon insertions are known to often have polar effects (19), which can be easily avoided by generating in-frame deletions (Fig. 4D).In addition to the larger cloning capacity of 10 kb, phAE159 is also deleted for TM4 gene 49, which has been implicated in superinfection exclusion (20). Expression studies with green fluorescent protein (GFP) are consistent with the interpretation that phAE159 is superior to parental TM4 in allowing independent phage to infect and express in the same host cell (21). The ability to transiently express heterologous genes is a very effective tool to generate in-frame unmarked deletion mutants in mycobacteria and implies that phAE159 is a general and effective vector for transient gene expression in mycobacteria.Vaccination protection results obtained with the fully antibiotic-sensitive double-deletion strain mc26206 are consistent with the recently published results of a hygromycin-resistant ΔleuD and ΔpanCD deletion strain (22). Thus, mc26206 appears to be significantly safer than M. bovis BCG (a BSL2 organism) (Fig. 5B). This strain has been cleared for BSL2 work by the biosafety committee of Albert Einstein College of Medicine and is available for further biochemical and genetic studies. The mc26206 strain is fully antibiotic sensitive and amenable to further genomic manipulations. M. tuberculosismc26206 has been used as a base strain to delete secA2 (mc26208) and has been unmarked to generate the antibiotic-sensitive triple-deletion ΔpanCD ΔleuCD ΔsecA2 mutant (mc26209). It has been observed that the deletion of secA2 in mc26206 makes the strain proapoptotic (U. D. Ranganathan, M. H. Larsen, W. R. Jacobs, Jr., G. J. Fennelly, unpublished data). Another mc26206 derivative, M. tuberculosis mc26435 (ΔpanCD ΔleuCD ΔsecA2, expressing simian immunodeficiency virus [SIV] Gag), was found to be safe for oral or intradermal administration to non-SIV-infected and SIV-infectedinfant macaques (23), suggesting mc26206 as a platform for a hybrid TB and HIV vaccine.Strain construction in M. tuberculosis remains inherently difficult because of the organism’s slow growth and pathogenicity. The apparently low rate of homologous recombination in mycobacteria means that even these new methods do not bring mycobacterial genetics to the level of the most facile microbial systems, such as Escherichia coli, Bacillus subtilis, and Saccharomyces cerevisiae. Nonetheless, these new methods allow undertaking construction of what has been one of the most fruitful resources in model systems: a comprehensive set of strains, each of which is deleted for a single defined open reading frame.
MATERIALS AND METHODS
Generation of pYUB1471.
The AES cloning vector, pYUB1471, was generated by directional cloning of four PCR fragments via primers with a unique variable region at the 5′ end of their Van911 site. Initially, hygromycin and ColEI origin of replication (oriE) fragments were PCR amplified using primer pairs pJSC347HL/pJSC347HR and pJSC34OL/pJSC34OR, respectively, from the pJSC347 (5) template. LHS and RHS fragments of Rv0004 were PCR amplified with primer pairs 0004LL/0004LR and 0004RL/0004RR, respectively, from H37Rv genomic DNA. The four fragments, hygromycin, oriE, LHS, and RHS, were Van91I digested and ligated directionally to construct pYUB1165. pYUB1471 was generated by ligating the Van91I digested pYUB1165 to DraIII and AlwNI digested PCR product obtained using primer pair sacBL/sacBR and pYUB870 as template (4). The sequence of pYUB1471 is provided in Text S1 in the supplemental material. Sequences of all the primers are listed in Table 1.
TABLE 1
List of primers used to generate pYUB1471
Primer name
Primer sequence[a]
pJSC347HL
TTTTTTTTCCATGAATTGGCAGGTCCTGTATCCTAAATC
pJSC347HR
TTTTTTTTCCATTCTTTGGCTAGAGTCCTGTCCGAAATA.
pJSC34OL
TTTTTTTTCCATAAGTTGGCAGGTTTGACAGCTTATCCAT
pJSC34OR
TTTTTTTTCCATTTTTTGGAGTGAGTCGTATTACGATCCT
0004LL
TTTTTTT CCATAAATTGGCCGCACCGTGACGACCTAA
0004LR
TTTTTTTTCCATTTCTTGGAGCGGTTGTGGATCACGAATG
0004RL
TTTTTTTTCCATAGATTGGCGCGACACCTACGGATAACA
0004RR
TTTTTTTT CCATCTTTTGGGCGTTCACTTGCCGACTT
sacBL
TTTTTTTTCACGCGGTGCTTTTTAACCCATCACATATAACC
sacBR
TTTTTTTTCAGCGCCTGAGATCGGCATTTTCTTTTGCGTT
The variable overhangs in the Van91I recognition sequences at primer ends are in bold.
List of primers used to generate pYUB1471The variable overhangs in the Van91I recognition sequences at primer ends are in bold.
Construction of phAE280.
Plasmid carrying resolvase enzyme was constructed by digestion of plasmid pYUB870 (4) with XbaI and SpeI, and a 5.7-kb fragment carrying tnpR (γδ resolvase), a kanamycin resistance gene, and the sacB gene was ligated to a 1.9-kb XbaI and SpeI fragment from plasmid pYUB854 (4), which has oriE, a lambda cos site, and PacI recognition sequences. The resulting plasmid, pYUB1672, was cloned into phAE159 and electroporated into mc2155 to obtain temperature-sensitive phage phAE280.
Construction of AES.
Amplicons between 650 and 1,000 bases flanking the gene were PCR generated with primer sets (LL/LR and RL/RR) to generate gene-specific LHS and RHS. Plasmid pYUB1471 was digested with Van91I to release four fragments. Two vector fragments (3.6 kb and 1.6 kb) which correspond to the sacB-hyg cassette and oriE-cos fragment, respectively, were cut out of the agarose gel and were prepared in bulk for all AES constructions. These two fragments were ligated in one step to Van91I-digested LHS and RHS fragments corresponding to the gene of interest (Fig. 1A). If a Van91I site is present in LHS or RHS, DraIII, BstAPI, or AlwNI restriction sites (Fig. 1B) can be used for the generation of AES. The ligation mix was transformed in E. coli DH5α, and the clones were confirmed by Van91I digestion. It is important that the E. coli strain used for plasmid and phasmid propagation does not express γδ resolvase. All the primers used to generate the AES are listed in Table S1 in the supplemental material. The sequence of the plasmid pYUB1471 is provided in Text S1 the supplemental material.Construction of temperature-sensitive TM4 shuttle phasmid phAE159 is described in Text S2 the supplemental material. The sequence of the phasmid phAE159 is provided in Text S1 in the supplemental material.
High-throughput generation of STPs and transduction.
A stepwise protocol for the generation of deletion-substitution mutants by transduction is shown in Fig. 2. A detailed protocol is given in the supplemental material.
Unmarking of the γδ(sacB-hyg)γδ cassette from deletion-substitution mutants.
The protocol is summarized in Fig. 4B. Briefly, the log-phase M. tuberculosis culture, obtained by subculturing at least once at an optical density at 600 nm (OD600) of 0.6 to 0.8, was grown to an OD600 of 0.6 to 0.8. Three milliliters of culture was centrifuged at 4,000 relative centrifugal force (RCF) at room temperature (RT). The pellet was washed twice by resuspending in 10 ml MP buffer followed by centrifugation at 4,000 RCF at RT. The pellet was suspended in MP buffer (100 to 200 µl) to obtain an OD600 of 1.0 to 1.5. A total of 50 µl of washed cells was incubated with 500 µl of phAE280 (titer of ~1010) at 37°C for 3 days. A total of 100 µl of sample was plated directly on 10% sucrose, and the remaining 450 µl of sample was diluted with 1.5 ml of 7H9 medium with 0.05% Tween. A total of 20 µl, 100 µl, 300 µl, and the remaining sample (after centrifugation) were plated on 10% sucrose. The plates were incubated at 37°C for 3 weeks. Colonies obtained were analyzed for the excision of sacB-hyg, after recombination between the two γδ sites in the cassette γδ(sacB-hyg)γδ, by both colony PCR and replica patching on plates with and without hygromycin (75 µg/ml) to determine the frequency of unmarking.
Colony PCR and Southern analysis of M. tuberculosis mutants.
Approximately half the colony was mixed with 100 µl of Bio-Rad DNA matrix with Triton X-100 (90 µl matrix [catalog number 7326030] and 10 µl of 0.1% Triton X-100) and incubated at 98°C for 40 min in a PCR machine. PCR tubes were briefly centrifuged at ~3,000 rpm. A total of 10 µl of the sample was used to set up a PCR. Southern blot analysis was performed as described previously (4).Nucleotide sequence of pYUB1471 and phAE159 in GenBank format. DownloadText S1, DOC file, 0.1 MBA stepwise protocol for the generation of unmarked deletion-substitution mutants by specialized transduction. DownloadText S2, DOC file, 0.2 MBConstruction of temperature-sensitive TM4 shuttle phasmid phAE159. DNA of PH101, a temperature-sensitive mutant of TM4, was self-ligated to form concatemers and partially digested with Sau3AI to generate fragments 40- to 50-kb long. pYUB328 was digested with XbaI, alkaline phosphatase treated, and digested with BamHI to generate arms (24, 25). Each arm contained a lambda cos site, and one of the ends was compatible with the Sau3AI-digested phage DNA. Partially digested phage DNA was ligated to pYUB328 arms at a 1:10 M ratio of insert to arms, ligated, packaged with the MaxPlax lambda packaging extract, and transduced into E. coli HB101 (Epicentre Biotechnologies). Cosmid DNA was extracted from a pool of 50,000 ampicillin-resistant colonies, representing a library of random pYUB328 insertions into the TM4 genome. The insertions of pYUB328 into the phage genome were expected to be accompanied by deletion of the nonessential region between the two Sau3AI sites. The pooled cosmid DNA was electroporated into M. smegmatis (8), where only functional TM4 phage derivatives were able to form phage infectious centers on an M. smegmatis lawn. The majority of these phages were wild type and did not contain the cloning vehicle pYUB328 (screened by plaque hybridization using the ECL direct nucleic acid detection system). Twenty-one recombinant phages were detected by hybridization against the pYUB328 vector DNA. Based on the restriction analysis, phAE159 harbored the largest deletion with respect to wild-type TM4. Phage phAE159 was sequenced along with PH101 and phAE87 (8). DNA sequencing and alignment of all three phages showed that phAE87 and phAE159 have overlapping deletions of 0.3 kb (gp62 to gp64) and 5.8 kb (gp48 to gp64). The new phasmid vector phAE159 has a cloning capacity of ~10.0 kb of recombinant DNA and was used as a backbone for the construction of all recombinant mycobacteriophages in further studies. DownloadFigure S1, PDF file, 0.1 MBConfirmation that the genes which could not be deleted by specialized transduction are essential to M. tuberculosis survival. Transduction with STP for the Rv0404 (fadD30) gene or the Rv3801c (fadD32) gene failed to yield any hygromycin-resistant colonies after repeated transductions. H37Rv was transformed with pMV261/Rv0404 (fadD30) or pMV261/Rv3801c (fadD32) to generate merodiploid strains mc2 7212 and mc2 7241, respectively. Southern blotting confirming the deletion of the Rv0404 (fadD30) gene and the Rv3801c (fadD32) gene, in their respective merodiploid strains. (A) The adjoining maps showed the chromosomal region of Rv0404 (fadD30) of H37Rv and ΔRv0404 (fadD30). The genomic DNA digested with PvuII. Lane 1, H37Rv with pMV261/Rv0404 (fadD30); lane 2, ΔRv0404 (fadD30) with pMV261/Rv0404 (fadD30); lane 3, pMV261/Rv0404 treated with PvuII. (B) The adjoining maps showed the chromosomal region of Rv3801c (fadD32) of H37Rv and ΔRv3801c (fadD32). The genomic DNA digested with BglII. Lane 1, H37Rv with pMV261/Rv3801c (fadD32); lane 2, ΔRv3801c (fadD32) with pMV261/Rv3801c (fadD32); lane 3, pMV261/Rv3801c with BglII. The regions used as probes are indicated. DownloadFigure S2, TIF file, 8.7 MBList of primers used to generate the AES for the deletion-substitution mutants in M. tuberculosis H37Rv.Table S1, DOC file, 0.2 MB.
Authors: Paras Jain; Travis E Hartman; Nell Eisenberg; Max R O'Donnell; Jordan Kriakov; Karnishree Govender; Mantha Makume; David S Thaler; Graham F Hatfull; A Willem Sturm; Michelle H Larsen; Preshnie Moodley; William R Jacobs Journal: J Clin Microbiol Date: 2012-01-25 Impact factor: 5.948
Authors: V Balasubramanian; M S Pavelka; S S Bardarov; J Martin; T R Weisbrod; R A McAdam; B R Bloom; W R Jacobs Journal: J Bacteriol Date: 1996-01 Impact factor: 3.490
Authors: Vasan K Sambandamurthy; Xiaojuan Wang; Bing Chen; Robert G Russell; Steven Derrick; Frank M Collins; Sheldon L Morris; William R Jacobs Journal: Nat Med Date: 2002-09-09 Impact factor: 53.440
Authors: Michael B Harbut; Catherine Vilchèze; Xiaozhou Luo; Mary E Hensler; Hui Guo; Baiyuan Yang; Arnab K Chatterjee; Victor Nizet; William R Jacobs; Peter G Schultz; Feng Wang Journal: Proc Natl Acad Sci U S A Date: 2015-03-23 Impact factor: 11.205
Authors: Olivia Vergnolle; Hua Xu; JoAnn M Tufariello; Lorenza Favrot; Adel A Malek; William R Jacobs; John S Blanchard Journal: J Biol Chem Date: 2016-08-26 Impact factor: 5.157