We first determined the role of Apa mannosylation in T cell recognition and recall of cytokine responses in healthy, BCG vaccinated (BCG⁺) and BCG unvaccinated (BCG⁻) adults. The peripheral blood mononuclear cells (PBMCs) from 17/24 (70.8%) BCG⁺ individuals produced more than 50 IFN-γ spot forming units (SFU)/10⁶ cells (a positive response) after in vitro stimulation with purified nApa, as compared to those from only 3/24 (12.5%) individuals after stimulation with rApa (Figure 1A). Among BCG⁺ individuals, an IFN-γ response was predominantly observed in individuals with positive skin test reactivity to Mtb purified protein derivatives (PPD⁺), with 15/16 (93.7%) individuals showing nApa-specific positive response. Only 2/8 (25%) BCG⁺ individuals without PPD reactivity (PPD⁻) and 1/22 (4.5%) controls, i.e. BCG⁻PPD⁻ individuals, responded positively to nApa. This dominant nApa-specific T cell response was dose dependent, as demonstrated by the increased frequency of IFN-γ SFU when PBMCs of 5 healthy PPD⁺ (including 3 confirmed Mtb-exposed) individuals were stimulated with varying amounts of nApa, in contrast to different doses of rApa (Figure S1A). Overall, these results demonstrate that nApa is recognized preferentially over rApa in these individuals.

Next, we characterized the Ag-specific IFN-γ, TNF-α and IL-2 cytokine responses of 8 randomly selected BCG⁺PPD⁺ subjects by intracellular cytokine staining (ICS) and polychromatic flow cytometry and determined the percentages of CD4⁺ and CD8⁺ T cells expressing one (1+), any combination of two (2+), or all three (3+) cytokines. Significantly higher frequencies of IFN-γ, TNF-α or IL-2 producing CD4⁺ T cells were observed in nApa stimulated PBMCs than those stimulated with rApa (Figure 1B and 1C) and the response was characterized by higher percentages (but comparable proportions) of nApa-specific polyfunctional CD4⁺ T cells in the responding donors (Figure S1B). In contrast, the frequency of total cytokine producing CD8⁺ T cells was not statistically different after stimulation with the two forms of Apa (Figure 1C), although 4 donors had more nApa-specific cytokine producing CD8⁺ T cells. Collectively, these results suggest that mannosylation of Apa is important for the Ag-specific T cell recall responses in these individuals.

The heightened T cell responses seen with mannosylated Ags are presumably due to efficient uptake through mannose binding CLRs on the APCs such as dendritic cells (DCs) and higher efficiency of Ag presentation [20]–[22]. Therefore, we sought to determine whether Apa mannosylation influences its uptake by DCs. We found increased binding/uptake of FITC-labeled nApa as compared to labeled rApa by the blood monocyte derived dendritic cells (MoDCs) and by the CD11c⁺HLA-DR⁺ myeloid dendritic cells (mDCs) in the PBMCs of BCG⁺ and BCG⁻ individuals (Figure S1C and D). A higher percentage of myeloid (mDCs) but not plasmacytoid dendritic cells (pDCs) from human PBMCs expressed CD80 following in vitro stimulation with nApa than after rApa (Figure 1D and 1E), suggesting increased activation of mDCs by nApa. It is known that mannose binding CLRs such as DC-specific intracellular adhesion molecule-3 grabbing nonintegrin (DC-SIGN) and mannose receptor (MR) are highly expressed on mDCs but are absent on pDCs, and play an important role in the uptake and presentation of mannosylated Ags [23]. Corroborating these observations, we found that nApa but not rApa binds to recombinant human MR, DC-SIGN and DC-SIGNR using an in vitro CLR adhesion assay (Figure 1F).

To determine whether Ag presentation of nApa to specific T cells requires intracellular processing, APCs were fixed with glutaraldehyde before or after Ag pulsing. The MoDCs generated from the blood monocytes of three nApa-responding individuals (BCG⁺PPD⁺) were used as APCs, while T cells were purified from the PBMCs of respective individuals. The IFN-γ ELISPOT was used to investigate Ag presentation and activation of T cells after co-culture. Glutaraldehyde fixation of APCs before addition of nApa abrogated IFN-γ response (Figure S1E), whereas fixation of APCs after pulsing with nApa did not prevent IFN-γ production, although the response (SFU) was lower than pulsing of Ag alone without fixation. These results indicate a requirement for intracellular processing of nApa for presentation to T cells. In contrast, fixation of APCs before addition of the superantigen, staphylococcal enterotoxin B (SEB), did not prevent the IFN-γ response of T cells. Therefore, the lack of nApa-specific IFN-γ secretion by specific T cells, if fixation of APCs preceded Ag pulsing, is consistent with the presentation of cognate Ag (versus superantigen) [24], [25].

Next, to determine whether increased binding/uptake and activation of mDCs by nApa influences T cell reactivity and epitope recognition, we screened 32 sequential nonmodified synthetic overlapping peptides of Apa with PBMCs of BCG⁺PPD⁺ individuals who recognized both forms of Apa and those who recognized nApa over rApa. Several peptides induced positive IFN-γ responses in PBMCs from individuals who recognized both forms of Apa (Figure S1F), whereas no peptide induced >20 IFN-γ SFU/10⁶ PBMCs from individuals who recognized only nApa. These results suggest that mannosylation of Apa does not induce alternate Ag processing in APCs to produce unique peptide epitopes, but rather carbohydrate contributes to T cell epitopes and recognition of such mannosylated epitopes (for example glycopeptides) is probably responsible for elevated T cell responses to nApa in these individuals.

To determine whether Apa mannosylation is required for T cell antigenicity in animal models, BALB/c mice were infected with Mtb intranasally (i.n.) or with M. bovis BCG subcutaneously (s.c.). At different time points after infection, lung, spleen and draining inguinal lymph node (ILN) cells were isolated and stimulated with nApa or rApa. The frequency of IFN-γ, TNF-α and IL-2 producing CD4⁺ T cells after nApa stimulation of lung and spleen cells from BCG-administered mice was comparable (IFN-γ, IL-2) to 2-fold (TNF-α) more than that induced by nAg85B stimulation, whereas stimulation by rApa induced only a marginal increase in the frequency of cytokine secreting CD4⁺ T cells as compared to controls (Figure 2A and S2A). Similar results were also obtained using ILN cells (data not shown).

The time kinetics of CD4⁺ T cell response confirmed the presence of significantly higher frequency of nApa-specific total cytokine secreting cells in the lung (p = 0.0002) and spleen (p = 0.0021) as compared to those specific for rApa (Figure 2B) and revealed that the Apa responses peaked at 32–52 weeks (wks) after BCG administration, with a similar expansion and contraction trend for the nApa- and rApa-specific T cells. During the expansion phase, 3.9- and 8.6-fold more nApa-specific total cytokine secreting CD4⁺ T cells were observed at 12 wks in the lung and spleen, while at the peak (52 wks) the increase was 7.7- and 2.7-fold more in the lung and spleen, respectively. At individual cytokine levels, however, the difference between nApa- and rApa-specific cytokine producing CD4⁺ T cell frequency was more pronounced for IFN-γ and TNF-α (Figure S2B). These results suggest that mannosylation of Apa strongly influences the magnitude of Apa-specific CD4⁺ T cell responses in BCG administered mice and indicates that the CD4⁺ T cells of these mice predominantly recognize major histocompatibility complex-II (MHC-II) restricted epitopes present in nApa only. On the contrary, the frequency of nApa- and rApa-specific total cytokine secreting CD8⁺ T cells was not significantly different at the time points evaluated, except at the 32 wk time point in the lung (Figure 2B). This difference in the lung was mainly due to the presence of a higher frequency of nApa-specific IFN-γ producing CD8⁺ T cells (Figure S2C). These results indicate that both forms of Apa possess MHC-I restricted epitopes recognized by the CD8⁺ T cells of BCG administered mice.

Next, we compared the cytokine expression profiles of nApa- and rApa-specific CD4⁺ or CD8⁺ T cells. The lung CD8⁺ T cell response was dominated almost exclusively by 1+ cytokine (IFN-γ) producing cells (Figure 2C, histograms). In contrast, about 38% and 50% of the nApa-specific cytokine secreting CD4⁺ T cells produced more than 1 cytokine at the peak (52 wk) time point in the lung and spleen, respectively, and a significantly higher frequency of 2+ and 1+ cytokine producing CD4⁺ T cells was observed after stimulation with nApa than rApa in the lung (Figure 2C, histograms). When the Ag-specific cytokine producing lung CD4⁺ T cells were categorized into 7 distinct subpopulations based on cytokine expression profiles, IFN-γ and TNF-α co-producing cells dominated nApa-specific 2+ cytokine producing cells (Figure 2D), while IFN-γ or TNF alone producing cells were predominantly present among nApa-specific 1+ cytokine producers. Despite differences in the magnitudes, the proportions of nApa- and rApa-specific cytokine producing 3+, 2+ and 1+ cytokine producing CD4⁺ or CD8⁺ T cells was not significantly different in both organs at the peak time point (Figure 2C, pie charts).

The proportions of nApa- and rApa-specific IFN-γ, IL-4 and IL-17 SFU in cultured ELISPOT assay were also not significantly different when compared at 3 time points; however, statistically higher frequency of IFN-γ or IL-17 SFU was present in the lung and spleen cell cultures after nApa stimulation (Figure 2E). These results collectively suggest that Apa mannosylation may have only minor effect on the polyfunctionality and ratios of specific subpopulations of Apa-specific T cells in BCG inoculated mice. Similar dynamics were observed in Mtb Erdman infected mice (Figure S2D–F).

Thirty-two non-modified, synthetic overlapping peptides of Apa were screened for their capacity to induce IFN-γ response in mice at 12 and 32 wks post BCG infection. Only peptide p271-288 induced a positive response which only occurred 32 wks after infection (Figure 3A). These analyses support the indication that heightened T cell responses to nApa are mainly due to glycosylation. We synthesized rApa C-terminal peptide (residues 281–325; 45-mer) with di-mannosyl-threonine residue at position 316 (Fig. S3A), akin to that found in nApa, to confirm the role of Apa glycosylation in T cell antigenicity.

A higher frequency of IFN-γ SFU was observed in the spleen and lung cell cultures of BCG and Mtb infected mice after stimulation with the synthetic glycopeptide as compared to stimulation with the non-glycosylated control peptide (Figure 3B), confirming the role of Apa glycosylation in T cell antigenicity. A higher frequency of cytokine producing CD4⁺ T cells, but not CD8⁺ T cells, was found in the splenocytes of BCG mice after in vitro stimulation with glycopeptide compared to stimulation with control peptide (Figure 3C), suggesting that the carbohydrate modification of nApa C-terminus may constitute a CD4⁺ T cell epitope. The T cell epitope prediction analyses also indicated probable binding of a 15-mer encompassing Thr316 to MHC-II molecules (Figure S3B).

To determine whether T cells recognize carbohydrate (di-mannose) in the absence of the proper peptide context, we synthesized rApa C-terminal 45-mer peptide with di-mannosyl-threonine residue at its N-terminus (at an unnatural position) linked by the additional glycine residues (Figure S3C). We termed this Apa 45-mer with additional four N-terminal residues as a 49-mer glycopeptide (i.e., Apa non-glycopeptide residues 281–325 with Gly-Thr(di-man)-Gly-Gly extension). Significant cytokine response was observed in ELISPOT when the lung, spleen or ILN cells from the BCG mice were stimulated with the synthetic 45-mer glycopeptide with a di-mannosyl residue at its natural position (Thr316). On the contrary, no positive cytokine response was found after in vitro stimulation with the 49-mer glycopeptide, 49-mer nonglycopeptide (control) or the free di-mannose alone (Figure S3D). These results suggest that the carbohydrate (di-mannose) attached to the proper peptide backbone is likely required for the recognition of nApa C-terminal glycopeptide by specific T cells from BCG mice. Further characterization of MHC bound glycopeptide interactions with the TCR and the orientation of carbohydrate residues can be accomplished through crystallography studies.

To determine whether mannosylation of Apa also influences its ability to induce a protective immune response, mice were vaccinated s.c. with 3 doses of nApa or rApa (1 µg/dose) in the presence of dimethyl-dioctadecyl ammonium bromide (DDA) and monophosphoryl lipid A (MPL) adjuvants at 4 wk intervals, and the T cell responses were investigated 1 wk after the last dose. Stimulation of splenocytes or ILN cells of vaccinated mice with either nApa or rApa, resulted in a comparable frequency of individual cytokine producing CD4⁺ T cells, regardless of which form of the Ag was used for vaccination. This Apa-specific T cell response was characterized by more TNF-α and IL-2 than IFN-γ producing cells in an ICS assay (Figure 4A, 4B and S4A). The frequency of Ag-specific total cytokine producing CD4⁺ and CD8⁺ T cells induced in the spleen and lung by either Apa vaccine was also comparable (Figure 4C).

When the cytokine expression profiles of CD4⁺ T cells from the spleen and lung were evaluated in the vaccinated groups, IL-2 and TNF-α co-producing cells dominated 2+ cytokine producing cells, while TNF-α or IL-2 single-producers were predominantly present among 1+ cytokine producing cells (Figure 4A). In contrast, the cytokine expression profiles of CD8⁺ T cells in the spleen and lung were dominated by TNF-α or IFN-γ secreting 1+ cytokine producers, respectively. Higher proportions of total cytokine producing splenic CD4⁺ T cells of rApa vaccinated mice consisted of polyfunctional T cells than those of nApa vaccinated mice (Figure 4C). Otherwise, we found no significant difference in the quality of nApa- and rApa-specific response, when the proportions of 3+, 2+ and 1+ cytokine producing CD4⁺ and CD8⁺ T cells were determined in the two vaccinated groups. Significantly more immunogen-specific IL-17 and IFN-γ SFU were observed in the spleens of rApa vaccinated as compared to nApa vaccinated mice in cultured ELISPOT assay (Figure 4D, histograms). However, the proportions of specific IFN-γ, IL-17 and IL-4 SFU constituting the total ELISPOT response were not significantly different in nApa or rApa vaccinated mice (Figure 4D, pie charts).

Synthetic peptide screening in nApa and rApa vaccinated mice revealed that T cell cytokine responses were predominantly directed toward non-glycosylated p271-288 and p231-250 peptides (Figure S4B), regardless of the immunogen used, indicating that Apa mannosylation does not induce alternate Ag processing in vivo when administered in DDA-MPL. Next, we evaluated the lung and splenic T cell responses of C-terminal 45-mer glycopeptide (Gp-281-325) in nApa vaccinated mice. Comparable cytokine SFU after in vitro stimulation with 45-mer glycopeptide and its non-glycosylated control were found, regardless of dose of nApa (1 or 10 µg) used for vaccination (Figure 4E), and the p271-288-specific response was significantly more than that induced by Gp-281-325, collectively suggesting that the differences observed between nApa and rApa induced T-cell responses may be partially overcome by co-administration of appropriate adjuvants.

Since nApa is also glycosylated with complex mannose modifications at its N-terminus (Figure S3B), the presence of N-terminal glycopeptide(s)-specific T cells in our nApa vaccinated mice samples remained a possibility. Our preliminary attempt to synthesize and evaluate N-terminal glycopeptide (39-mer) was unsuccessful, and whether nApa glycopeptide-specific T cells are generated following subunit vaccination was not clear. Therefore, to identify nApa-specific T cell clones we employed a specialized protocol of T-cell hybridoma generation, using a single dose nApa vaccination of BALB/c mice in Freund's incomplete adjuvant (FIA). Individual T cell clones responding to nApa were purified by serial dilutions and tested for reactivity to nApa or rApa using an IL-2 capture ELISA after co-culture with APCs (i.e., syngeneic bone marrow derived dendritic cells) pulsed with Ag. Of the total 17 Apa-reactive hybridoma clones developed, 7 clones recognized both nApa and rApa while 10 clones recognized only nApa and not rApa (data not shown). Three of the 7 clones that recognized both nApa and rApa responded to a specific synthetic, nonmodified 15-mer peptide (two reactive to peptides spanning p271-288, and one to peptide p70-84, data not shown), while none of the 10 nApa reactive clones recognized any of the synthetic, overlapping, nonmodified peptides. Of the 10 clones only reacting to nApa, one clone, 4C3, reacted with a glycopeptide fraction of trypsin digested nApa fractionated by a reversed phase-HPLC column chromatography. Clone 4C3 produced significant amount of IL-2 when cultured with APCs pulsed with the whole digest of nApa but not with the digest of rApa (Figure 5A), indicating that the Ag presentation of nApa to 4C3 T cell clone was not inhibited by trypsin digestion. The RP-HPLC fractions consisting of the N-terminal 106 amino acid glycopeptide (residues p40-145) of nApa only demonstrated reactivity in IL-2 assay (Figure 5B, Figure S5A and S5B). Glycosylation of the N-terminus of nApa from Mtb likely explains specific recognition of the 4C3 clone to the nApa peptide. The lack of biological activity of T cell clone 4C3 when presented with rApa or synthetic nonglycosylated peptides collectively suggest that N-terminal glycopeptide-specific T cells are generated after nApa subunit vaccination. Further confirmation of these results will require identification of the precise epitope and synthesis of the N-terminal glycopeptide.

We investigated the protective efficacy of two Apa forms using DDA-MPL, a known Th1 adjuvant. BALB/c mice were vaccinated s.c. with nApa or rApa 3-times at 4 wk intervals using two different concentrations (1 or 10 µg), in DDA-MPL adjuvant and challenged with Mtb 4 wks after the last dose to assess their protective potential. Mice vaccinated with nAg85B, a known protective Ag, were included for comparison. Mice immunized once with live BCG at the start of vaccination were used as positive controls, while negative controls received adjuvant alone (3-times) or were left untreated (naïve). Both forms of Apa imparted significant protection compared to naïve and adjuvant control groups at 2 different doses used (Figure 6A), and the bacterial burden among the Apa vaccinated groups was not significantly different. Vaccination with nAg85B also induced protection, but only nAg85B-10 µg was statistically different from the negative control groups at the level of lung and spleen. However, the degree of protection afforded by BCG vaccination was greater than any of the subunit vaccinated groups. Vaccination with either nApa or rApa induces comparable levels of protection in our model and suggests that glycosylation of Apa is not critical for protection.

To ascertain whether comparable protective efficacy offered by nApa and rApa vaccination in mice correlated with an equal ability to recall cellular and humoral immune responses after Mtb challenge, T cell responses and serum Ab titers were measured before Mtb challenge (at 4 wks after last vaccination) and 6 wks post challenge. All Apa vaccinated groups presented with an equal vaccine immunogen-specific CD4⁺ T cell recall response in the lungs after challenge, characterized by a significantly higher frequency of total cytokine producing CD4⁺ T cells in the post-challenge versus pre-challenge group (Figure 6B). Significant recall from spleen derived T cells was only observed in rApa-1 µg group. No significant differences were observed when the immunogen-specific lung or spleen CD8⁺ T cell responses were compared between vaccine immunogens; however, the CD8⁺ T cell responses of all vaccinated groups were higher than those of control groups in the lungs. Both nApa and rApa were recognized equally well when used for stimulation of lung and spleen cells of vaccinated and challenged groups (Figure S6A). Comparable patterns of cytokine co-expression profiles were found when T cell responses of respective Apa vaccinated groups were compared before and after challenge (Figure 6B); characterized by immunogen-specific TNF-α and IL-2 producing 2+ and 1+ CD4⁺ T cells in the spleen and IFN-γ producing 1+ CD4⁺ and CD8⁺ T cells in the lungs. On the contrary, splenic CD4⁺ T cells responses of the challenged naïve controls were characterized by nApa-specific IFN-γ and TNF-α producing 1+ and 2+ CD4⁺ T cells only.

The cultured ELISPOT assay revealed that the frequency of immunogen-specific total IFN-γ, IL-17 and IL-4 SFU in the lung was significantly lower in vaccinated-challenged groups compared to respective vaccinated group pre-challenge (Figure 6C). These results suggest possible decrease in in vitro proliferation of immunogen-specific lung T cells in ELISPOT cultures after recent in vivo recall or regulatory responses suppressing cytokine release following Mtb challenge. Proof for any of these explanations will require further in-depth investigation. All Apa vaccinated groups produced comparable amounts of immunogen-specific serum IgG1 Abs before and after challenge (Figure 6D), and these Abs recognized nApa and rApa equally well (Figure S6B). However, rApa vaccination induced significantly higher amounts of IgG2a and IgG2b Abs than nApa vaccination, except at 10 µg dose post-challenge.

Of importance, when pathogen-specific immune responses were investigated in the lung and spleen of vaccinated and control groups of mice in vitro, higher frequency of Mtb short term culture filtrate (STCF)-specific total cytokine producing CD4⁺ and CD8⁺ T cells was observed in naïve and adjuvant only mice as compared to vaccinated mice post-challenge (Figure 7A), with a significant difference at the level of lung. As expected, STCF-specific responses of unvaccinated-challenged mice were dominated by IFN-γ and TNF-α 1+ and 2+ CD4⁺ T cells and IFN-γ 1+ CD8⁺ T cells. Furthermore, when STCF-specific IFN-γ, IL-17 and IL-4 SFU were compared among challenged groups, responses of unvaccinated mice were found to be highly skewed toward IFN-γ (87–96% of total response), with significantly more SFU compared to all vaccinated groups (Figure 7B).

Considering the fact that most individuals in the world have been vaccinated with BCG and new TB vaccines have to be considered in the context of prior BCG vaccination, development of strategies aimed at boosting and improving the protective immunity induced by BCG is considered to be one of the rational approaches to develop an effective vaccination regimen against TB. Since BCG induced protection wanes significantly by 18 months in mice (unpublished data), we investigated the booster effect of 2 doses of nApa or rApa administered 3 wks apart in DDA-MPL at 16 months after BCG vaccination. Both forms of Apa significantly boosted T cell responses compared to boosting with saline, and a significantly higher frequency of T cells from the Apa-boosted groups recognized nApa over rApa after in vitro stimulation (Figure 8A and B). Nonetheless, comparable nApa-specific responses were found in both Apa-boosted groups, except for splenic IFN-γ SFU. Of importance, boosting with either form of Apa, but not with saline, significantly reduced Mtb CFU in the lung and spleen of BCG mice compared to age-matched controls (Figure 8C). BCG mice boosted with nApa or rApa showed about 2.0-log reduction in bacterial counts in the lungs as compared to age-matched naïve mice, while 1.7-log to 2.0-log reduction was found in the spleens of nApa- and rApa-boosted mice, respectively. This protection is characterized by a significantly lower frequency of ESAT-6+CFP-10-specific total as well as CD4⁺ T cell IFN-γ responses in the spleens of Apa-boosted groups (Figure 8D).

The study was approved by the Institutional Review Board of Centers for Disease Control and Prevention (CDC), Atlanta, Georgia, USA, (approved protocol number 1652), and informed written consent was obtained from all human participants before withdrawal of venous blood. All animal experiments performed in BSL-II or BSL-III animal facilities were in strict accordance with the guidelines of the U.S. Public Health Service Policy on the Humane Care and Use of Animals and the Guide for the Humane Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committees of Centers for Disease Control and Prevention, Atlanta, Georgia, and the Colorado State University, Fort Collins, Colorado, USA reviewed and approved these animal protocols (approval numbers SABMOUC1664, 1847, SHIMOUC 1490 and CSU 98-026A).

The nApa was purified from Mtb H37Rv culture filtrate by traditional and reversed-phase chromatography as described previously [8]. The recombinant clone for Apa (pMRLB.17), WCL, STCF, nAg85B, rESAT-6 and rCFP-10 of Mtb H37Rv were obtained through the NIH Biodefense and Emerging Infection Research Resources Repository. The rApa was expressed in E. coli BL21 (DE3) and purified from lysates by Nickel chromatography with endotoxin removal using analogous methods described elsewhere [48] followed by DEAE-Sepharose chromatography. The overlapping peptides of Apa and the C-terminal 45-mer and 49-mer (glyco) peptides were synthesized by the Fmoc/tbu solid-phase peptide synthesis strategy [49], [50]. See Supplementary Methods (Text S1) for details.

Blood was obtained from 50 healthy adult individuals (26 BCG⁺ and 24 BCG⁻). All donors were HIV negative and were without any clinical signs of TB. All BCG⁺ donors received BCG as a neonatal vaccine. We excluded responses of 1 BCG⁻ and 2 BCG⁺ donors who tested positive for hepatitis. All 16 BCG⁺ donors with positive tuberculin skin test reactivity (>10 mm induration after intradermal injection) had normal chest radiographs. However, 1 BCG⁻ and 4/16 BCG⁺ donors demonstrated positive reactivity to ESAT-6+CFP-10 in PBMC ELISPOT assay. The Mtb exposure of these donors was further confirmed by commercial IFN-γ ELISPOT (Oxford Immunotech) and ELISA (Cellestis) tests (IGRAs). Remaining 22 BCG⁻ donors had negative tuberculin skin test and no known history of contact with individuals with TB. The exposure of donors to environmental mycobacteria is not known but all live in Atlanta area with very low incidence of atypical mycobacterial infections. All donors demonstrated reactivity to phytohaemagglutinin (PHA) in ELISPOT assay and none were anergic.

Specific-pathogen-free, 6–8 wk old female BALB/c (H-2^d) mice (Harlan Sprague Dawley) were used in the study. BCG vaccinated mice received 100 µl of 1×10⁶ CFU Copenhagen s.c. above the gluteus superficialis and biceps femoralis muscles of hind legs (50 µl/leg). For protein vaccination, groups of mice received 200 µl of nApa, rApa or nAg85B (1 or 10 µg) in DDA and MPL (from Salmonella minnesota Re 595) [51] s.c. on hind legs (100 µl/leg) either 3-(subunit) or 2-times (prime-boost vaccination). Mtb infections were performed i.n. with 5×10⁴ CFU Erdman strain. At 48 h after Mtb challenge, about 0.5% of the total CFU delivered could be cultured from the lungs. The bacterial burden was determined by plating organ homogenates onto Middlebrook 7H10 agar supplemented with OADC and TCH (2 µg/ml). CFU were enumerated after 4 wks of incubation at 37°C.

IFN-γ, IL-4 or IL-17A ELISPOT assay was performed using commercially available human or mouse ELISPOT reagent set (BD-Biosciences) according to the manufacturer's protocol as described previously [18], except that no DCs were added as supplemental APCs and cells were stimulated with 10 µg/ml of purified or complex Mtb Ags, synthetic peptides or di-mannose (unless mentioned otherwise) for 40 h in RPMI-1640. See Supplementary Methods (Text S1) for details.

Expression of cell surface markers and intracellular cytokine production by human PBMCs or mouse organ cells after in vitro stimulation with Mtb Ags (10 µg/ml) at 37°C for total 12 h was assessed as described previously [52]. See Supplementary Methods (Text S1) for details.

For fluorescent labeling of proteins, commercially available FITC labeling kit (Pierce Biotechnology) was used. The protein labeling was carried out in 50 mM sodium borate buffer (pH 8.5) using 2 mg/ml protein concentration. Unreacted excess florescent dye was removed using spin columns with purification resin as per the manufactures instructions. Protein concentration was assessed using the BCA assay (Sigma). The efficiency of labeling was determined by Nano Drop Fluorospectrometer and was comparable between nApa and rApa. Protein samples were stored in the dark at −20°C before use. Human blood monocyte derived DCs (MoDCs) were generated as described before [53] using recombinant human GM-CSF (80 ng/ml) and IL-4 (40 ng/ml) (PeproTech). Freshly isolated human PBMCs or MoDCs (day 6) were pulsed with FITC-labeled nApa, rApa or WCL (20 µg/ml) and incubated at 4°C or 37°C for the indicated amounts of time. Cells were washed with PBS containing 10% FBS. Samples were then stained with desired Abs to identify DC subsets and analyzed using flow cytometry.

Mtb nApa or rApa was coated onto ELISA plates (Nunc, Maxisorp) at 5 µg/well in 0.1 M carbonate-bicarbonate buffer (pH 9.6); and coating took place for 18 h at room temperature. Wells coated with 5 µg/well of purified BSA fraction-V (Fisher Scientific) were used as negative controls while those coated with Mtb H37Rv WCL or mannose caped LAM served as positive controls. Blocking was carried out with 1% BSA for 30 min at 37°C and additional 1.5 h at room temperature in TSM buffer (20 mM Tris-HCL (pH 7.4) containing 150 mM NaCl, 2 mM CaCl₂ and 1 mM MgCl₂) [54]. Soluble recombinant human DC-SIGN (CD209)-Fc chimera, DC-SIGNR (CD299)-Fc chimera or MR (CD206) (R&D Systems) (2.5 µg/ml in TSM buffer) was added and the adhesion was performed for 2 h at room temperature. Unbound CLR was washed away and the binding of DC-SIGN-Fc or DC-SIGNR-Fc was determined by an anti-IgG1-Fc ELISA using a peroxidase conjugate of goat anti-human-Fc. Binding of MR was determined by anti-human-MR (CD206) mouse mAb (clone 685641; R&D Systems) followed by peroxidase conjugate of anti-mouse IgG. The plates were developed using o-phenylenediamine dihydrochloride (Sigma-Aldrich) and absorption was measured at 492 nm. Specificity of CLR binding was determined by blocking the interaction in the presence of either 2 mg/ml mannan or 5 mM EDTA (OD 492 nm <0.1).

The PBMCs were isolated from the venous blood of healthy BCG⁺PPD⁺ human donors who responded to nApa in IFN-γ ELISPOT assay. Recombinant human GM-CSF and IL-4 developed MoDCs (day 6) were used as APCs while T cells purified from the PBMCs of respective donors using ‘Dynabeads Untouched Human T Cells’ purification kit (Invitrogen) and magnetic separation (>95% pure) were used as effector cells. APCs (2×10⁶) were pulsed with either nApa (10 µg/ml) or SEB (1 µg/ml) or without any Ag (complete RPMI-1640 media only) and incubated at 37°C. After 4 h, APCs were washed with RPMI-1640 and were either left untreated or fixed with 0.05% glutaraldehyde for 5 min at 22°C. APCs were washed and the fixation was stopped by incubation with 0.2M glycine in RPMI-1640 at 22°C. After 5 min, cells were washed 4 times with RPMI-1640. Alternatively, APCs were fixed and Ag pulse was carried out for 4 h. APCs were washed 4 times after Ag pulse and were co-cultured with purified T cells (1∶2 ratio) in ELISPOT plates pre-coated with anti-human IFN-γ capture Abs. After 40 hr of incubation at 37°C in the presence of 5% CO₂, ELISPOT plates were developed and SFU were counted.

T-cell hybridomas specific for Mtb nApa were generated as previously described [55]. Briefly, four BALB/c mice were each vaccinated with 40 µg of nApa in FIA by injecting 25 µl into each hind footpad and the remainder at the base of the tail. Five days after the vaccination, the draining lymph nodes (popliteal, inguinal and periaortic) were harvested to obtain lymph node (LN) cells. The primed LN cells were restimulated in vitro with syngeneic bone marrow derived dendritic cells (BMDCs) that had been pulsed with 10 µg/ml nApa per well. Approximately 1×10⁶ primed LN cells were added per well in two 24-well tissue culture plates in a total volume of 1 ml of complete RPMI 1640 medium (RPMI 1640 supplemented with 10% FCS, 5×10⁻⁵M 2-ME (Sigma), plus a nutrient cocktail as described [56]). After two days of culture, the cells were harvested and pooled from all 48 wells, washed, fused with the T cell fusion partner BWα⁻β⁻ [57], and plated out into ten 96-well plates. Clones that grew in individual wells were screened using either BMDCs, as above, or a BALB/c mouse-derived B cell lymphoma line, A20 [58], pulsed with 0.5 µg/well nApa. The selection of responding T cell hybridomas was performed by assaying 24 hr culture supernatants using paired rat monoclonal Abs specific for mouse IL-2 (Pharmingen/BD-Biosciences) in a capture ELISA. For Ag presentation studies, microculture wells were prepared containing 250 µl of culture medium, 5×10⁴ each of T cell hybridoma and APC, and a known amount of nApa or rApa (intact or trypsin digested), in flat-bottomed 96-well microtiter wells. Dose response curves with native and recombinant Ags were performed to determine which T cell hybridomas responded nApa alone, or to both the native and recombinant form of the protein. T cell hybridomas were further tested with a panel of synthetic, overlapping, nonglycosylated Apa peptides or trypsin digested and RP-HPLC separated nApa fractions to determine the peptide epitopes that were recognized. For details regarding trypsin digestion of Apa and characterization of RP-HPLC fractions please refer to Supplementary Methods (Text S1).

ELISA assays of mouse sera were performed as described previously [46], using 2 µg/ml of nApa or rApa for coating and HRP-conjugated anti-mouse IgG1, IgG2a (BD-Pharmingen) and IgG2b (Santa Cruz Biotech) Abs for detection.

Differences between groups were assessed by the parametric Student's t test or 1-way ANOVA followed by Bonferroni's test and the nonparametric 2-tailed Mann-Whitney U-test (the Wilcoxon test for matched pairs) or Kruskal-Wallis followed by the Dunn's post-test (GraphPad Prism program). Unless indicated, all immune response data are presented after subtracting no Ag control values. A value of p<0.05 was considered to be significant and * <0.05; ** <0.01; *** <0.001.

The accession/identification numbers of Mtb proteins used in the study are Mtb (H37Rv) alanine and proline-rich antigen (Apa), Rv1860, NCBI reference sequence: YP_177849.1; Mtb (H37Rv) Ag85B, Rv1886c, NCBI reference sequence: NP_216402.1; Mtb (H37Rv) 6 kDa early secretory antigenic target (ESAT-6), Rv3875, NCBI reference sequence: YP_178023.1; Mtb (H37Rv) 10 kDa culture filtrate protein (CFP-10), Rv3874, UniProtKB/Swiss-Prot reference sequence: P0A566.2.