The creation of Mtb ΔfbpA strain and its characterization in macrophages and mice have already been described [13], [14], [46]. To generate an additional sapM gene deletion in ΔfbpA, the plasmid construct pTBSAPM5 was electroporated into this strain and cultures were plated initially on 7H10-TW-OADC agar with hygromycin and X-gal to obtain hygromycin resistant blue colonies. This selection resulted in several blue colonies out of which one colony designated as DKO/C was subjected to further screening to obtain sucrose resistant colonies lacking β-galactosidase activity. This screening resulted in three white colonies namely DKO/C1, DKO/C2 and DKO/C7. To verify if deletion of sapM gene had occurred in these colonies, a Southern blot was made with genomic DNA from these strains and also with Mtb H37Rv and Mtb ΔfbpA. Upon hybridization with 3.3 kb radiolabeled DNA probe containing sapM region, and subsequent autoradiography, the blot showed two signals (2.1 and 2.9 kb) for Mtb H37Rv and Mtb ΔfbpA strains and only one signal (4.4 kb) for DKO strains (DKO/C1 is shown here) (Fig. 1a). These signals were on the predicted line, based on restriction sites in this region of the genome (Fig. S1), and indicate that sapM gene is deleted by allelic replacement in DKO strain. To further confirm the deletion of sapM in DKO strain, we also performed PCR using primers specific for this region (please see Fig. S1 for the location of the primers). While primers located at the 5′end (RV3310EX1) and 3′ end (RV3310EX2) of the sapM gene yielded the expected sizes of 900 bp and 165 bp DNA, respectively for the Mtb H37Rv and DKO strain (Fig. 1b), an internal primer RV3310RT2 with the primer at the 5′end of sapM (RV3310EX1) failed to amplify a 530 bp product in the DKO strain (Fig. 1c), again reinforcing the deletion of sapM gene in this strain. Finally, we determined whether the deletion of sapM gene in DKO strain led to the disruption of the expression of this gene by RT-PCR, using the internal primers RV3310RT1 and RV3310RT2. This revealed that only cDNA obtained from Mtb H37Rv and Mtb ΔfbpA strain only yielded the expected size DNA fragment (350 bp) but not the DKO strain (Fig. 1d), thus confirming the absence of sapM expression in DKO strain. This concluded the generation of Mtb fbpA/sapM double knockout (DKO) strain.

The ability of mycobacterial strains to grow inside macrophages is a virulence trait, and macrophages are routinely used to determine their virulence [14], [46]. Fig. 2a & d illustrate the growth of strains within mouse bone-marrow derived macrophages (BMs) and human THP1 macrophages. The DKO strain was relatively attenuated in both BMs and THP1 macrophages, compared to ΔfbpA and ΔsapM mutants or the wild type Mtb H37Rv (Fig. 2a,d). To determine if the enhanced attenuation of DKO was due to sapM, we complemented the DKO mutant strain with sapM gene and named the strain as DKOcom. BMs infected with the DKOcom showed a growth curve similar to that of its parental strain ΔfbpA (Fig. S2), indicating that the increased attenuation of DKO was due to deletion of sapM.

Since, the intracellular death of mycobacteria is partly due to oxidative radicals like reactive oxygen species (ROS) and nitric oxide (NO), macrophages and culture supernatants were evaluated respectively for ROS using a fluorescent probe and NO derived nitrite with Griess reagent. The BMs showed no significant differences in ROS responses but a marginally elevated NO response was induced by the DKO strain (Fig. 2b,c). In contrast, the ROS response was elevated for THP1 macrophages infected with DKO compared to other strains (Fig. 2e). There was no significant NO response observed among THP1 macrophages infected with either wild type or mutants (not shown). This observation was consistent with the notion that human macrophages produce barely detectable NO during mycobacterial infection [49].

ROS cascade begins with the generation of superoxide by phagocyte oxidase [43]. Superoxide and inducible nitric oxide synthase derived NO have bacteriostatic and bactericidal activity, respectively against mycobacteria. To confirm the susceptibility of DKO to oxidants, the superoxide and NO donor, 3-morpholinosydnonimine (SIN-1) was used to treat a highly viable culture of wild type and mutants in broth culture. Fig. 2f shows that DKO strain was again more susceptible to oxidants compared to others. These data suggest that the decreased growth of DKO was attributable in part to elevated oxidant responses in macrophages. Similar studies were done using macrophages infected with DKO, although it was difficult to rule out artifacts arising due to the dose-dependent toxic effects of oxidants on macrophages.

Although Mtb H37Rv resists phago-lysosomal fusion [37], [41], [50], certain mutant Mtb strains have a decreased ability to prevent PL fusion, and this decreased ability correlates with reduced intracellular viability for these mutants [46], [51]. Since lysosomes present an acidified hostile environment, and the parent ΔfbpA mutant was partially PL fusion competent, we reasoned that PL fusion is one additional mechanism through which, intracellular viability of DKO strain could be reduced. To test this hypothesis, BMs and THP1 macrophages were infected with either GFP tagged Mtb wild type (H37Rv) or Oregon green stained mutants (ΔfbpA, ΔsapM and DKO).The PL fusion was monitored using microscopic colocalization of lysosomal markers like CD63 and rab7 (Fig. 3a). An antibody to lysosome associated membrane protein-1 (LAMP-1), was used as a positive control since LAMP-1 is present on all mycobacterial phagosomes but at varying levels between virulent and avirulent bacteria [46]. The DKO strain extensively colocalized with LAMP-1 followed by ΔsapM, ΔfbpA and H37Rv (Table, Fig. 3b). Significantly, CD63 and rab7 were found to be more enriched on DKO phagosomes compared to either ΔsapM or ΔfbpA or H37Rv. Since CD63 and rab7 are definitive markers of lysosomes, and LAMP1 is present both on late endosomes and lysosomes, these data indicated that DKO phagosomes are lysosome fusion competent. It was also significant to note that ΔsapM and ΔfbpA stained for CD63 and rab7 more densely than wild type H37Rv, indicating that deletion of either fbpA or sapM renders these mutants comparatively more lysosome fusion competent than wild type H37Rv (Fig. 3b).

Since DKO strain showed increased PL fusion and enhanced susceptibility to killing within macrophages, we hypothesized that DKO could be more immunogenic since PL fusion leads to degradation of mycobacterial antigens facilitating their presentation through the MHC-II pathway. When mycobacteria infect macrophages, an Ag85B derived peptide-25 epitope is rapidly presented to T cells [52]. We demonstrated that, in vitro presentation of Ag85B is a measure of immunogenicity of mycobacteria in macrophages [43] and PL fusion of mycobacteria within macrophages enhances Ag85B production predicting vaccine efficacy against tuberculosis [45]. In vitro antigen presentation using dendritic cells and macrophages, indicated that the DKO strain was more efficiently processed since overlaid T cells secreted more IL-2 (>2500 pgs/mL in 4 hrs) than either ΔsapM (1000 pgs/mL) or ΔfbpA (950 pgs/mL) infected APCs (Fig. 4a).

Finally, in vitro immunogenicity was correlated with the ability of DKO to prime Th1 immunity in vivo. Mice were vaccinated subcutaneously with 10⁶ CFU of mutants and analyzed for Ag85B specific T cells using Elispot. The DKO strain again induced an increase in the number of Ag85B specific T cells in mouse spleens (Fig. 4b). These data indicate that the DKO (ΔfbpAΔsapM) has a highly immunogenic phenotype in macrophages and DCs as well as in mice. It may be noted that the numbers of spot forming cells (SFU) increased over 14 days followed by a decline. This is consistent with the splenic immune response following mycobacterial vaccines. For example, SFUs for Ag85B increase after a single vaccination with BCG over 2 weeks, and then decline [45].

This study was carried out in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. The protocol was approved by the Institutional Animal care and Use Committee of the University of Texas Health Science Center, Houston (Protocol number AWC-09-175 under Animal Welfare Assurance Number A3413-01).

All bacterial strains, unless specified, were grown at 37°C. Escherichia coli strain DH5α was used to subclone M. tuberculosis DNA. LB broth or LB agar plates were used to grow E. coli. The above media with the antibiotic ampicillin (100 µg/mL) or kanamycin (25 µg/mL) or hygromycin (100 µg/mL) were used to grow E. coli strains containing plasmids. Mtb H37Rv (27294) is from ATCC. M. tuberculosis ΔfbpA strain is a derivative of H37Rv, which we published earlier [13]. M. tuberculosis ΔsapM strain was from NIH TB resources Center at Colorado State University (currently BEI, ATCC). Middlebrook 7H9 broth medium containing OADC (10%) and Tween (0.05%) (7H9 broth), or Middlebrook 7H10 agar medium containing OADC (10%) and Tween (0.05%) (7H10 Agar) or Middlebrook 7H11 agar medium containing OADC (10%) (7H11 Agar) was used to grow M. tuberculosis strains. Kanamycin (25 µg/mL) or hygromycin (50 µg/mL) was added to the Middlebrook media to grow M. tuberculosis strains harboring plasmids.

To disrupt sapM (Rv3310) gene in Mtb ΔfbpA strain, we first downloaded the DNA sequences of sapM gene and its adjacent region from NCBI databases. Based on the sequences, we synthesized four oligonucleotide primers namely RV3310A, RV3310B, RV3310C and RV3310D (Table 1). Synthesis of oligonucleotide primers were performed at the Center for DNA Technology, University of Texas Health Science Center at San Antonio. While the primers RV3310A, RV3310B were designed to amplify the 5′ prime region of sapM and its upstream 1306 bp fragment (Frag I), primers RV3310C and RV3310D were designed to amplify the 3′ region of sapM and its downstream 934 bp fragment (Frag II). Also primers Rv3310B and RV3310C were engineered to have StuI sites in them. Using these primers and Mtb H37Rv DNA, we amplified fragments I and II in PCR and these fragments were cloned into pCR2.1 Vector (Invitrogen) to create plasmids pTBSAPM1 and pTBSAPM2, respectively. The fragment II from plasmid pTBSAPM2 was released by cutting with StuI and BamHI and cloned into the pTBSAPM1 cut with the same restriction enzymes. The resulting plasmid pTBSAPM3 has a DNA fragment, which has upstream and downstream regions of sapM but with substantial deletion in sapM coding region (Fig. S1). The DNA fragment in pTBSAPM3 was released from the plasmid by digesting the plasmid with restriction enzymes HindIII and NotI and the released fragment cloned into p1NIL to create plasmid pTBSAPM4. A 7939 bp PacI fragment, that contains sacB and lacZ genes and a gene for hygromycin resistance, was then isolated from the plasmid pGOAL19 [61] and cloned into the PacI site of the plasmid pTBSAPM4. The resulting plasmid pTBSAPM5 was used to generate a markerless sapM mutant in Mtb ΔfbpA strain. Plasmid DNA from E. coli was isolated by using a Qiaperp kit (Qiagen Inc.,Valencia, Calif.).

To obtain ΔfbpAΔsapM double mutant, we used ΔfbpA strain reported earlier [13]. This mutant strain was grown to mid-logarithmic phase (OD₆₀₀ = 0.8–1.0) in 7H9 broth and competent cells prepared according to Jacobs et al. [62]. Four hundred microliter of Δfbp cells were mixed with 5 µg pTBSAPM5 DNA, linearized with NaOH treatment, in 0.2 mm cuvettes (BioRad) and electroporated using standard protocols. After electroporation, 1 mL 7H9 medium without any antibiotic was added to each cuvette and left overnight at 37°C. Then, the cell suspension from the cuvettes was plated on 7H10 agar plates containing the antibiotic hygromycin (50 µg/mL) and X-gal (40 µg/mL). Transformants showing blue color, resulting from single crossover event, were selected after three weeks of incubation at 37°C. Further screening of the transformants for the deletion of sapM region was performed by a two-step selection method [61]. First, the blue colonies were streaked onto 7H10 agar plates containing no antibiotics. Following growth, a loop-full of cells were resuspended into liquid medium, diluted serially to several folds and plated onto 7H10- agar plates containing 2% sucrose. Sucrose resistant colonies resulting from double crossover event were streaked onto plates with or without hygromycin. Colonies showing no resistance to hygromycin were finally streaked onto plates containing kanamycin, since fbpA mutant is kanamycin resistant. The DNA from these colonies were further examined in Southern and PCR to confirm the deletion of sapM region.

To complement the DKO strain with sapM gene, we constructed an integration plasmid carrying sapM gene as follows. First, we amplified the whole sapM gene and its upstream promoter region (2116 bp) by PCR using primers RV3310A and RV3310EX2 (Table 1) and Mtb H37Rv genomic DNA. The fragment was cloned in pCR2.1 vector to result in plasmid pTBSAPMA. This plasmid was digested with KpnI and XbaI to release the fragment which was cloned in KpnI and XbaI digested pMV306H, a derivative of pMV306 in which kanamycin resistant marker is replaced with hygromycin marker, to get plasmid pM306SAPMA. This plasmid was transformed into DKO/C1 (ΔfbpA/ΔsapM double mutant) strain by electroporation. The colonies were selected in 7H10 hygromycin plates and the integration of the plasmid was confirmed by Southern blot (data not shown). The resulting strain was named as DKOcom.

To confirm the deletion of sapM gene in Mtb ΔfbpA strain, we performed Southern analysis [63]. We isolated chromosomal DNA from Mtb H37Rv, Mtb ΔfbpA and Mtb ΔfbpAΔsapM strains using CTAB (cetyltrimethylammonium bromide) method [64] and 3 µg of each DNA was digested with NdeI and BamHI. The digested fragments were separated on 1% agarose gel and Southern transferred to nitrocellulose membrane (BioRad). The blot was hybridized with a [α-³²P]dCTP labeled probe generated by random primer method using the 3300 bp DNA fragment (generated by primers Rv3310A and Rv3310D) as template. Southern hybridization and final washing of the blot were performed at 68°C. In addition to Southern, we also carried out PCR to confirm the deletion of sapM region in Mtb ΔfbpAΔsapM strain using standard protocol [63] with Taq polymerase (Perkin-Elmer, Foster City, Calif.). We used primers Rv3310EX1, Rv3310EX2 and Rv3310RT2 (Table 1) for this analysis.

Total RNA from Mtb strains was isolated using Tri reagent as described previously [65]. cDNA from total RNA was synthesized using Superscript (Invitrogen) and random hexamers. PCR analysis was performed with primers RV3310RT1 and RV3310RT2 (Table 1) and using the cDNA from the previous step as the template for the reaction.

Bone marrows from C57BL/6 (4–8 weeks) mice were cultured for 7–10 days in Iscove's modification of Dulbecco's modified Eagles medium (IDMEM) with 10% FBS and 10 ng/mL GM-CSF (Cell Sciences, USA). The macrophages (BMs) and dendritic cells (DCs) were purified using CD11c bead fractionation kit (Miltenyi Inc, USA) yielding >95% pure DCs and >95% adherent BMs (effluent of bead fractionation) [46]. They were plated onto 24 well plates (for colony counts, oxidant measurement) or 8-well slide chambers (for antibody stains) and were rested in IDMEM without GM-CSF before infections. In addition to mouse derived cells, phorbol mystyl acetate (PMA) (10 nM) activated human THP1 cells were used to test the intracellular growth of mycobacteria. The T cell hybridoma (BB7) cell line specific for an epitope (241–256) of Ag85B of Mtb was used in antigen presentation assays (kindly provided by Dr. C. Harding, Case Western University, Cleveland, OH). The culture of these cell lines were described earlier [43], [46].

To determine the growth curves in macrophages, the wild type and mutants were used to infect (MOI of 10) naïve BMs and PMA activated THP1 cells using procedures described earlier [14], [43], [46]. Phagocytosis was allowed to take place for 4 h, after which the monolayers were washed to remove the non-phagocytosed bacteria. One mL of fresh IDMEM medium was added to each well and the culture plates were incubated at 37°C in the presence of 5% CO₂. After each time point, macrophages were harvested and lysed with 0.05% sodium dodecyl sulfate (SDS) for 15 min at room temperature. Tenfold dilutions of the lysed macrophage suspensions were made in PBS and 100 µL of each dilution was examined for Mtb growth on triplicate 7H11 agar plates that were incubated for 3 weeks at 37°C. Each Mtb strain was examined in quadruplicate wells and three independent experiments were performed.

BMs and THP1 macrophages infected with mycobacteria were tested for ROS using the fluorescent probe H₂-dichlorodihydrofluorescein diacetate (H₂DCFDA)(Invitrogen, USA) that is cleaved within the cells by esterases and oxidized into fluorescent DCF reactive oxygen species (ROS) induced by infection [43]. Fluorescent DCF was measured by reading BMs or THPs of 24 well plates using in Ascent fluoroscan at 485 nm/530 nm. Triplicate wells of macrophages were read for each mutant at different time intervals in two separate experiments and plotted as average fluorescence units (AFU). The enzyme inducible nitric oxide synthase generated nitric oxide (NO) was measured in the supernatants of similar cultures using Griess reagent and expressed as µM nitrite in the medium. The susceptibility of mycobacteria to superoxide and NO released by the donor 3-morpholinosydnonimine N-ethylcarbamide (SIN-1)(Invitrogen, USA) was determined by incubating 10⁵ CFU/mL of mycobacteria in a broth culture with 10 mM of SIN-1 for 24 and 72 hrs and plating organisms on 7H11 agar for viable counts.

Macrophages were infected with GFP expressing Mtb H37Rv and Oregon green stained ΔfbpA, ΔsapM and DKO (ΔfbpAΔsapM) strains. These strains were sonicated slightly to disperse the bacteria, centrifuged at 500 rpm and the supernatant containing the single CFUs were used for infection. BMs of slide chambers were infected with Mtb strains (MOI, 1) for 4 h at 37°C in the presence of 5% CO_2, washed, and incubated with fresh medium up to 72 h. Washing, fixing of the cells, staining for different markers and mounting of the slides were performed as reported earlier [66]. Texas red conjugated antibodies to primary antibodies were from Jackson Immunochemicals (West Grove, PA). Colocalization was examined and scored using a Nikon fluorescence microscope equipped with a Metaview deconvolution software as described [67].

Monolayers of BMs were infected with Mtb strains for 4 hrs, (MOI 1∶5) washed and overlaid with Ag85B specific BB7 T cells (1∶20 ratio) as described earlier [43]. Four or 18 hrs later the supernatant was assayed for IL-2 using a sandwich ELISA kit (R and D systems, CA) and expressed as pg/mL/10⁶ T cells. This assay has been validated earlier by the lack of antigen presentation when macrophages are infected with Mtb strains deleted for Ag85B antigen (ΔfbpB) or when MHC class II deficient PMJ2-R macrophages are used as antigen presenting cells [43]. Two or four separate experiments were performed for each Mtb strain using triplicate wells for each assay and the data were averaged for p value calculations.

C57Bl/6 mice (3 per group per time point per strain) were immunized s.c. with 10⁶ CFU of Mtb strains once and after 7,14 and 21 days, the splenic cells were depleted of non-T cells using a pan-T cell kit (Miltenyi Inc). A 96 well plate coated with anti-mouse IFNγ (Elispot kit, Ebiosciences) was washed and overlaid with 10⁵ enriched T cells from spleens of mice along with T cells from naïve mice. The wells were added with 100 ng of Ag85B purified protein per well (BEI,ATCC). Control wells received 100 ng/mL of KLH protein and Elispots from this control were subtracted from those induced by Ag85B restimulation and expressed as spot forming cells per 10⁵ T cells. Three separate experiments were performed and data averaged.