Ag mannosylation represents a promising strategy to augment vaccine immunogenicity by targeting Ag to mannose receptors (MRs) on dendritic cells. Because fungi naturally mannosylate proteins, we hypothesized that Ags engineered in fungi would have an enhanced capacity to stimulate T cell responses. Using the model Ag OVA, we generated proteins that differentially expressed N- and O-linked mannosylation in the yeast Pichia pastoris and compared them to their unglycosylated counterparts produced in Escherichia coli. We found that yeast-derived OVA proteins containing N-linkages, extensive O-linkages, or both were more potent than the unmannosylated Ags at inducing OVA-specific CD4+ T cell proliferation. This elevated response to fungal Ags was inhibited by mannan, suggesting involvement of MRs. However, the macrophage MR (CD206) was not essential, because macrophage MR-deficient dendritic cells were fully competent in presenting yeast-derived OVA Ags. Thus, the use of fungal glycosylation to provide N-linked and/or extensive O-linked mannosylation increased the capacity of the model Ag OVA to stimulate Ag-specific T cell responses in an MR-dependent manner. These data have implications for vaccine design by providing proof of principle that yeast-derived mannosylation can enhance immunogenicity.
Currently, the majority of the vaccines licensed for human use induce protective Ab responses (1, 2). However, vaccine-mediated protection against many tumors and infections is thought to require the induction of T cell-mediated immunity (1, 2). Mannosylation of Ags has been reported to enhance MHC class I- and MHC class II-restricted Ag presentation and T cell stimulation by up to 200-fold compared with nonmannosylated proteins (3, 4, 5, 6, 7). Furthermore, in vivo induction of Th1 cytokines and Ag-specific CTL responses were observed in mice upon immunization with mannosylated vaccines (6, 8, 9). Thus, Ag mannosylation is an attractive strategy for boosting cell-mediated immune responses.
The effectiveness of mannosylated proteins is probably attributable to their ability to target mannose receptors (MRs),3 such as the macrophage MR (MMR; CD206) and dendritic cell (DC)-specific ICAM-3-grabbing nonintegrin (DC-SIGN; CD209), found on DCs. MMR binds oligosaccharides containing terminal mannose residues and traffics Ag into early endosomes (10), whereas DC-SIGN recognizes high mannose oligosaccharides and directs Ag into late endosomes or lysosomes for degradation (11, 12). Both the MMR and DC-SIGN have the capacity to direct internalized Ag into endocytic pathways that result in MHC presentation and subsequent T cell activation. Tumor Ags fused to a mAb specific for the MMR stimulated MHC class I- and II-restricted T cell responses (13), whereas IgG1 isotype Abs specific for DC-SIGN induced IgG1-specific CD4+ T cell proliferation (12).
Many studies on MR-targeted vaccines involved the use of chemical conjugation to provide mannosylation of Ags (6) and therefore were unable to test the effects of N- and O-linked mannosylation. Because fungi preferentially glycosylate their glycans with mannose residues, we wanted to examine the immunogenic potential of mannosylated proteins produced in a fungal system. We have previously established that mannosylation is essential to the capacity of mannoproteins (MP) from the pathogenic fungus, Cryptococcus neoformans, to stimulate a cell-mediated immune response (14). Furthermore, MRs on DCs bind to exposed mannose residues on the MP, resulting in efficient Ag uptake, processing, and presentation.4 Molecular characterization of cryptococcal MPs revealed serine/threonine (S/T)-rich C-terminal regions that serve as sites for extensive O-linked glycosylation and asparagine-X-S/T (N-X-S/T) sequences as potential N-glycosylation sites (15, 16).
Given the above findings, we hypothesized that the natural ability of fungi to mannosylate Ags could be exploited to make recombinant mannosylated vaccines that efficiently stimulate CD4+ T cell responses. To test this hypothesis, mannosylated and unglycosylated Ags were engineered in the yeast Pichia pastoris and the bacterium Escherichia coli, respectively. The model Ag was a portion of OVA that includes epitopes recognized by OVA-specific T cells and two potential N-linked glycosylation sites. Site-directed mutagenesis of the N-X-S/T sequences was performed to study the contribution of N-linked mannosylation, whereas the S/T-rich region from cryptococcal MP98 (15) was fused to OVA to study the effects of extensive O-linked mannosylation. We found that mannosylated OVA Ags were significantly more potent than unglycosylated OVA at stimulating Ag-specific CD4+ T cell responses. Moreover, both N- and O-linked mannosylation targeted Ag to MRs on DCs.
Materials and Methods
Materials and cell culture
All chemical reagents were obtained from Sigma-Aldrich unless otherwise noted. Tissue culture media were purchased from Invitrogen Life Technologies cell culture systems (Invitrogen Life Technologies). R10 medium is defined as RPMI 1640 containing 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 55 nM 2-ME. Tissue culture incubations were performed at 37°C in humidified air supplemented with 5% CO2.
Generation of mannosylated constructs in the P. pastoris expression system
The EasySelect Pichia Expression kit (Invitrogen Life Technologies) was used to develop all mannosylated recombinant proteins (i.e., P. pastoris-derived protein containing the OVA portion (ppOVA), protein consisting of the OVA portion fused to the S/T-rich region produced in P. pastoris (ppOVAST), P. pastoris-derived (counterparts lacking the N-linked glycan site (ppOVANQ and ppOVASTNQ)). The P. pastoris host strain used to express ppOVA and ppOVAST was KM71H (Muts, Arg+), whereas ppOVANQ and ppOVASTNQ were produced in the X-33 wild-type (WT) strain. All procedures and medium formulations for yeast maintenance and growth were performed as directed in the manual.
Construction of inserts into the P. pastoris expression vector was as follows. For ppOVA, OVA DNA encoding aa 230–359 (GenBank accession no. CAA23716) was amplified by PCR from plasmid pJR052 (gift from Dr. A. Marshak-Rothstein, Boston University, Boston, MA). To use this vector as a template, the vector was linearized with BglII (New England Biolabs). Primers (Invitrogen Life Technologies) included the restriction sites EcoRI in the forward primer and XbaI in the reverse primer for directional cloning into the expression vector (forward primer: pOVA-F, 5′-GGTGAATTCATCCTGGAGCTTCCATTTG; reverse primer: pOVA-R, 5′-ACCTCTAGACCGCTAGCAAATTC). ThermalAce polymerase (Invitrogen Life Technologies) was used to perform the PCR (25 cycles). The first cycle consisted of a 3-min denaturation at 95°C, followed by 25 cycles of a 30-s denaturation at 95°C, a 30-s annealing at 60°C, and a 40-s extension at 74°C. The last cycle was a 10-min extension at 74°C. This is the standard PCR protocol unless otherwise noted. All PCR products were purified using the MinElute PCR purification kit (Qiagen).
The OVA portion was joined to a S/T-rich region of C. neoformans MP98 by PCR fusion after amplification of each DNA fragment. The S/T-rich region from cloned MP98 cDNA (GenBank accession no. AF361369) (15) encodes a stretch of amino acids (380–432) containing 34 S and T residues. Primers were designed to produce products with complementary regions on the 3′ end of the OVA region and the 5′ end of the S/T-rich region to allow for a final round of fusion PCR to create the ppOVAST construct. The OVA portion was amplified by PCR as reported for ppOVA, except the reverse primer was changed to pOVAfusionMP98ST-R (5′-ACCATCGATACCGCTAGCAAATTC). To construct the S/T-rich region, a portion of the S/T-rich region from cryptococcal MP98 (15) was amplified by PCR from a pYES2.1/V5-His-TOPO vector containing MP98. This vector was linearized with HindIII (New England Biolabs). The following primers were used: forward primer: pOVAfusionMP98ST-F, 5′-ATCGATGGTAGCAGCTCTACC; and reverse primer: MP98ST-R, 5′-ACCTCTAGACCACTGCTAGCGTT. To create ppOVAST, the OVA portion and S/T-rich region PCR products were used as templates. The following primers were used: forward primer, pOVA-F; and reverse primer, MP98ST-R. The PCR protocol was the same as described above except that PCR cycles were increased to 35 cycles, and each extension cycle was changed to 1 min at 74°C.
Final PCR products for ppOVA and ppOVAST and the expression vector pPICZαA were digested with EcoRI and XbaI. pPICZαA was also treated with calf intestinal alkaline phosphatase (Promega). Inserts were then ligated into the vector using T4 DNA ligase (Invitrogen Life Technologies) and transformed into competent E. coli. (TOP10 F′; Invitrogen Life Technologies).
ppOVANQ and ppOVASTNQ.
Both constructs were developed by site-directed mutagenesis of the two potential N-linked glycosylation sites in the OVA sequence using the QuikChange II Site-Directed Mutagenesis Kit as directed by Stratagene. The two potential N-linked sites, NLT and NLS, were conservatively mutated to QLT and QLS. Mutations were introduced one at a time using the pPICZαA vector with either the ppOVA or ppOVAST insert as the template DNA. A PvuII site was created with each mutation. The mutagenic primers were designed using a QuikChange primer design online tool at 〈http://labtools.stratagene.com/QC〉. The following primers were used for the first and second round of mutagenesis: NLT1-F, 5′-TGAAGATGGAGGAAAAATACCAGCTGACATCTGTCTTAATGGC; NLT1-R, 5′-GCCATTAAGACAGATGTCAGCTGGTATTTTTCCTCCATCTTCA; NLS2-F, 5′-TGTTTAGCTCTTCAGCCCAGCTGTCTGGCATCTCC; and NLS2-R, 5′-GGAGATGCCAGACAGCTGGGCTGAAGAGCTAAACA. The cycle parameters for the PCR began with a 30-s denaturation at 95°C, followed by 16 cycles consisting of a 30-s denaturation at 95°C, a 1-min annealing at 55°C, and a 4-min and 20-s extension at 68°C. Mutagenized products were transformed into E. coli strain XL1-blue. All plasmids were screened for correct size inserts and orientation by restriction enzyme digestion, followed by sequencing at the Molecular Genetics Core Facility (Boston University School of Medicine). Plasmid DNA for all constructs were linearized with SacI and transformed into P. pastoris using the EasyComp transformation. Zeocin-resistant transformants were checked for proper integration of recombinant DNA by PCR according to the manufacturer’s protocols
Recombinant protein expression and purification from P. pastoris
Expression of protein constructs followed the protocol for scale-up expression in shaker flasks included in the Invitrogen Life Technologies manual. Briefly, cells were grown in buffered complex medium containing glycerol at 30°C with shaking at 250 rpm in baffled flasks for 72 h. Cells were harvested by centrifugation and resuspended in buffered complex medium containing methanol (BMMY) to induce protein expression. ppOVA and ppOVAST were resuspended in half the volume, whereas ppOVANQ and ppOVASTNQ were resuspended in an equal volume of medium. Cells were grown for 24 h, and the supernatant was harvested while the cells were resuspended in fresh BMMY and induced for another 24 h before harvesting. Supernatants were sterile-filtered and concentrated ∼70-fold using a 10-kDa cutoff regenerated cellulose tangential filtration cassette (Millipore). The concentrated supernatants were then diluted with 5 vol of buffer B (100 mM NaH2PO4, 10 mM Tris-Cl, and 8 M urea (pH 8.0)) and purified over an Ni-NTA (Qiagen) column using fast performance liquid chromatography (AKTA model; Amersham Biosciences). The column-bound fractions were eluted with buffer E (100 mM NaH2PO4, 10 mM Tris-Cl, and 8 M urea (pH 4.5)) and dialyzed against dH20 using a 7-kDa cutoff dialysis cassette (Pierce). Samples were lyophilized, resuspended in PBS, and filter-sterilized. The protein concentration was assessed using the bicinchoninic acid assay (Pierce).
Generation of unmannosylated constructs in an E. coli expression system
The pMAL Protein Fusion and Purification System Kit (New England Biolabs) was used to develop all unmannosylated recombinant proteins (i.e., E. coli-derived OVAST (ecOVAST) and ecOVA). All procedures and medium formulations for E. coli maintenance and growth were performed as directed in the manual.
Construction of inserts into pMALc2x expression vector
To produce ecOVA and ecOVAST, the constructs were amplified by PCR with the pPICZαA vector with the ppOVA and ppOVAST inserts, respectively, as the template DNA. Vectors were linearized with SacI to use as templates. The forward primer was pOVA-F. The reverse primer included a SpeI site, PICZR1 (5′-GGTACTAGTTCTAAGGCTACAAACTCA). Bio-X-Act polymerase (Bioline) was used to perform the PCR. The first cycle consisted of a 3-min denaturation at 95°C, followed by 25 cycles of a 30-s denaturation at 95°C, a 30-s annealing at 58°C, and a 1-min extension at 68°C. The last cycle was a 7-min extension at 68°C. PCR products were digested with EcoRI and SpeI. The pMALc2x vector was digested with EcoRI and XbaI and dephosphorylated with calf intestinal alkaline phosphatase. PCR products were then ligated into the pMALc2x vector and transformed into E. coli strain TOP10 F′. Colonies were screened for correct size and sequence by blue-white screening and restriction enzyme digestion with EcoRI and XbaI.
Recombinant protein expression and purification from E. coli
Expression of protein constructs followed the protocol included in the New England Biolabs manual. Briefly, 1 l of rich broth plus glucose and ampicillin was inoculated with a 10-ml overnight culture of cells containing either ecOVA or ecOVAST in the pMALc2x vector. Isopropyl β-d-thiogalactoside was added to the culture after reaching a density of 2 × 108 cells/ml to induce the production of protein. After 3 h of growth, the cells were harvested by centrifugation and resuspended in 50 ml of column buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, and 1 mM DTT). Samples were frozen overnight at −20°C. To lyse cells, samples were thawed in an ice-water bath and subjected to sonication. Samples were centrifuged, and the crude extract containing the protein of interest was diluted 1/5 with column buffer and purified over an amylase column using fast performance liquid chromatography. Proteins were eluted with column buffer and 10 mM maltose and dialyzed against dH2O using a 7-kDa cutoff dialysis cassette. The samples were lyophilized, resuspended in PBS without Mg2+/Ca2+, and filter-sterilized. The protein concentration was assessed using the bicinchoninic acid assay.
Enzymatic and chemical deglycosylation of recombinant proteins
To remove N-linked oligosaccharides, proteins were subjected to peptide:N-glycosidase F (PNGase F) treatment according to the recommended reaction conditions (New England Biolabs). Briefly, proteins were denatured by incubation for 10 min at 100°C in glycoprotein denaturing buffer (0.05% SDS and 0.1% 2-ME). A final concentration of G7 reaction buffer (50 mM sodium phosphate (pH 7.5) supplemented with 1% Nonidet P-40) was then added to the mixture. PNGase F or water (for mock-treated samples) was added and incubated at 37°C for 2 h.
Chemical deglycosylation was performed using the GlycoProfile IV chemical deglycosylation kit (Sigma-Aldrich) according to the manufacturer’s protocol. Briefly, a precooled solution containing 10% anisole in trifluoromethanesulfonic acid was added to a 1-mg sample of lyophilized protein. Trifluoromethanesulfonic acid nonselectively cleaves N- and O-linked oligosaccharides by hydrolysis, leaving the protein backbone intact, but degrading the carbohydrates. The sample was gently mixed until completely dissolved and then incubated on ice for 3 h with occasional shaking. Using bromophenol blue (0.2%) as a pH indicator, the sample was brought to pH 6 with precooled pyridine (60%) and then dialyzed against PBS using a 7-kDa cutoff dialysis cassette. After filter sterilization, the protein concentration was assessed using the bicinchoninic acid assay.
Western blot analysis
Proteins were resolved in 12% SDS-PAGE and transferred to nitrocellulose membranes. After blocking in TBST (25 mM Tris-HCl (pH 8.0), 125 mM NaCl, and 0.05% Tween 20) with 5% milk for 1 h at room temperature, membranes were probed with either 1/5000 c-Myc-HRP Ab (Invitrogen Life Technologies) in TBST with 1% milk overnight or 1/1000 polyclonal OVA Ab (ICN) overnight followed by a secondary Ab, 1/2000 swine anti-goat IgG-HRP Ab (Caltag Laboratories) in TBST with 1% milk. To detect mannosylation, membranes were blocked with TBST containing 1% BSA for 1 h at room temperature, probed with 1 μg/ml biotinylated Con A for 1 h, and then incubated with 31.2 ng/ml streptavidin conjugated to HRP for an additional hour at room temperature in TBST with 1% BSA. Blots were washed with TBST and then visualized using the Supersignal West Pico chemiluminescent substrate (Pierce).
C57BL/6 and OT-II mice were purchased from The Jackson Laboratory. OT-II TCR-transgenic (TCR-Tg) mice express an αβ TCR specific for OVA aa 323–339 in the context of MHC class II (I-Ab). DO11.10 TCR Tg mice (a gift from A. Marshak-Rothstein, Boston University, Boston, MA) express an αβ TCR specific for OVA aa 323–339 in the context of MHC class II (I-Ad). MMR−/− (H-2b) mice (17) were a gift from M. Nussenzweig (The Rockefeller University, New York, NY) and were backcrossed at least seven times before use. All animal procedures were reviewed and approved by the institutional review board at the Boston University School of Medicine animal facility.
Splenocytes and bone marrow-derived DCs (BMDCs)
Mouse splenocytes were isolated as in our previous studies (14, 15). BMDCs were generated following the protocol of Lutz et al. (18). Briefly, bone marrow cells obtained from tibiae and femurs of 8- to 16-wk-old mice were cultured in R10 medium supplemented with 10% GM-CSF supernatant from the GM-CSF-secreting J558L cell line (19). Cells were fed with fresh supplemented medium on days 3, 6, and 8 and were used on day 9.
T cell proliferation assays
Lymphoproliferation was assayed by [3H]thymidine incorporation and flow cytometric analysis of CFSE (Molecular Probes)-labeled cells. All Ags were preincubated with 20 ng/ml polymyxin B to minimize potential effects of endotoxin. Mitomycin C (0.5 mg/ml)-treated splenocytes or BMDCs, used as an APC source, were pulsed with Ags, washed three times, and added to U-bottom tissue culture plates at 1 × 105 cells/well. The source of naive CD4+ T cells was a preparation of brachial, axial, inguinal, and mesenteric lymph node (LN) cells isolated by Histopaque-1083 separation or magnetic cell-sorted (Miltenyi Biotec) CD4+ T cells from LN and spleen. LN or CD4+ purified T cells (1 × 105) were added to each well and cocultured with Ag-pulsed APCs for 96 h. For the thymidine incorporation assay, 1 μCi of [3H]thymidine was added per well for the final 18 h. Cells were harvested onto filters, and [3H]thymidine incorporation was assessed using a beta counter. The flow cytometric assay was identical, except Ag-pulsed BMDCs were incubated with CFSE-labeled LN cells. Briefly, LN cells were labeled with CFSE at 1 × 107 cells/ml in PBS with 0.1% FBS and 1.5 μM CFSE for 8 min at 37°C. Labeling was quenched with an equal volume of 37°C FBS for 10 min. Cells were washed three times with R10 medium before coculture. After 96 h, cells were harvested, stained with CD4-PE (Caltag Laboratories), and analyzed by flow cytometry (BD Biosciences; FACScan). The number of cell divisions on CD4+ gated cells was determined by measuring 2-fold decreases in CFSE fluorescence with WinMDI 2.8 software (The Scripps Research Institute, La Jolla, CA).
Means and SEMs were compared using Student’s t test; p < 0.05 was considered significant.
Construction of mannosylated and unmannosylated recombinant OVA proteins
To investigate the immunological effects of fungal mannosylation, our strategy was to compare the capacities of mannosylated and unmannosylated Ags to induce Ag-specific T cell proliferation. Six recombinant proteins were produced, four of which were heterogeneously mannosylated P. pastoris-derived proteins and two were unglycosylated E. coli-derived proteins (Fig. 1 A). The target Ag was a portion of OVA (aa 230–359) containing epitopes recognized by tg OVA-specific CD4+ and CD8+ T cells. This portion of OVA also includes two N-X-S/T motifs (where X denotes any amino acid except proline) that serve as potential sites for N-linked glycosylation as well as 22 serines and threonines, which are possible sites for O-linked glycans. Glycosylation of recombinant proteins was determined by Con A reactivity and PNGase F treatment, which specifically cleaves off entire N-linked oligosaccharides and results in SDS-PAGE mobility shifts.
Our data show that ppOVA appears larger in apparent molecular size than ppOVANQ, which harbors mutated N-X-S/T motifs (Fig. 1,B). This strongly suggests the presence of N-linked mannosylation on ppOVA. This is confirmed by the decrease in molecular size after PNGase F treatment (Fig. 1 C). As expected, given that it lacks sites for N-linked glycosylation, ppOVANQ is unaffected by PNGase F treatment. The apparent molecular sizes of ppOVANQ and PNGaseF-treated ppOVA were larger than the predicted unglycosylated molecular size of 17.2 kDa, indicating the presence of modest degrees of O-mannosylation. Lower Mr bands were also observed in the ppOVA and ppOVANQ preparations, the identities of which remain to be defined. These bands appear to contain variable segments of the OVA portion extending through the C terminus, because they are detected by both polyclonal Abs specific for OVA (data not shown) and mAbs specific for c-Myc. Hence, these bands could represent variably processed proteins or proteolytic fragments truncated at the N-terminal end.
To examine the contribution of extensive O-mannosylation, the S/T-rich region (containing 34 potential sites for O-mannosylation) from cryptococcal MP98 (15) was incorporated into the OVA constructs. Thus, ppOVAST displayed both N-linked and extensive O-linked mannosylation, whereas ppOVASTNQ only contained extensive O-linked mannosylation. Predictably, the apparent molecular sizes of ppOVASTNQ (Fig. 1,B) and PNGase F-treated ppOVAST (Fig. 1,C) were similar and less than that of ppOVAST. Moreover, PNGase F-treated ppOVASTNQ remained unchanged in size (Fig. 1,C). Both ppOVAST and ppOVASTNQ, regardless of PNGase F treatment, were positive for Con A reactivity (Fig. 1,D), indicating the presence of mannosylation on these proteins. The broad bands seen on the Western blots of ppOVAST and ppOVASTNQ (Fig. 1,B) Ags are typical of heavily glycosylated fungal proteins and are thought to be due in large part to glycan heterogeneity (14, 15, 20). Finally, ecOVA and ecOVAST were produced to serve as unglycosylated controls. Both ecOVA and ecOVAST were unaffected by PNGase F treatment (Fig. 1 E) and did not stain with Con A (data not shown).
Effect of fungal mannosylation on Ag-specific T cell proliferation
To determine the relative potencies of mannosylated and unmannosylated Ags at inducing Ag-specific T cell stimulation, APCs were pulsed with recombinant OVA Ags and then cocultured with naive OVA-specific CD4+ T cells from TCR-Tg DO11.10 (H-2d) or OT-II (H-2b) mice. In the first set of experiments, splenocytes were used as a source of APCs and were pulsed with OVA Ags for various periods of time (Fig. 2,A) or at varying concentrations (Fig. 2,B). Splenocytes were then cocultured with DO11.10 TCR-Tg LN cells, which were used as a source of CD4+ T cells. Mannosylated ppOVAST and ppOVA were significantly more potent at inducing Ag-specific lymphoproliferation than unmannosylated ecOVAST and ecOVA in an Ag pulse time-dependent manner (Fig. 2,A) and a dose-dependent manner (Fig. 2 B).
Next, we examined Ag-specific CD4+ T cell responses from OT-II mice in response to P. pastoris- and E. coli-derived recombinant OVA proteins. For these experiments, all six mannosylated OVA Ags were tested, and BMDCs were used as APCs. The use of a relatively pure population of DCs allowed us to better examine the mechanisms by which the mannosylated Ags enhanced immunogenicity as well as to determine whether DCs were the pertinent APC. To confirm that proliferation was due to CD4+ T cells and that it was occurring throughout the incubation period, in selected experiments the [3H]thymidine incorporation assay was complemented with a flow cytometric analysis of CFSE-labeled CD4+ T cells.
Once again, mannosylated ppOVA and ppOVAST induced stronger proliferative responses than unmannosylated ecOVA and ecOVAST, as measured by flow cytometry of naive CFSE-labeled OT-II TCR-Tg CD4+ T cells (Fig. 3). Interestingly, ppOVASTNQ, which contains extensive O-linked mannosylation, but lacks N-linked glycans, promoted an increase in T cell proliferation over ecOVAST. In contrast, ppOVANQ, which only displays a minimal amount of O-mannosylation, exhibited roughly the same degree of T cell stimulation as unglycosylated ecOVA. Similar results were obtained when OT-II TCR-Tg CD4+ T cell proliferation was measured by [3H]thymidine incorporation, in that ppOVAST, ppOVASTNQ, and ppOVA were significantly more potent than ppOVANQ and the unglycosylated E. coli-derived OVA proteins (Fig. 3 C).
To demonstrate that mannosylation was responsible for the immunopotentiating effects of the yeast-derived Ags, ppOVAST was chemically deglycosylated and tested for immunogenicity. Deglycosylation of ppOVAST was confirmed by the decrease in molecular size (Fig. 3,E, inset) and the loss of Con A reactivity (data not shown). Deglycosylated ppOVAST stimulated significantly less T cell proliferation than mannosylated ppOVAST (Fig. 3 E). T cell proliferation was unaffected when cocultures incubated with OVA and deglycosylated ppOVAST were compared with those incubated with OVA alone, demonstrating that deglycosylation of ppOVAST did not appear to harbor any toxic effects (data not shown).
Effect of competitive inhibition of mannose receptors on Ag-specific T cell responses
As mannosylated OVA Ags were able to augment T cell proliferation, we next sought to unravel the mechanism behind this process by addressing the contribution of MRs on DCs. Yeast mannans, which are composed of long chains of branched mannose residues, have been shown to block MRs such as MMR and DC-SIGN (12, 14, 21). Thus, competitive inhibition of MRs was examined by pulsing BMDCs with mannosylated OVA Ags in the presence or the absence of mannans and subsequently coculturing the treated BMDCs with naive CD4-purified OT-II TCR-Tg T cells. We found that mannans profoundly inhibited the ability of mannosylated P. pastoris-derived OVA proteins to stimulate T cell proliferation (Fig. 4). Similar results were obtained when CFSE-labeled naive OT-II CD4+ T were measured by flow cytometry (data not shown). Overall, these results support a dominant role for MRs in the uptake of mannosylated Ags by BMDCs.
Effect of T cell responses to mannosylated Ag-pulsed BMDCs on an MMR knockout (KO) background
The above experiments, demonstrating inhibition of T cell proliferation by yeast mannans, suggested a critical contribution of MRs on DCs, but did not address the issue of which MRs were most important. Therefore, in the final set of experiments, the Ag-presenting capacity of MMR-deficient BMDCs was compared with that of WT C57BL/6 BMDCs. To limit the variables contributing to experimental differences, the same pool of naive CD4+ T cells was used in the cocultures with DCs, making the only variable the source of DCs (either WT or MMR KO). Because the T cells are the responding population of cells that provide the assay readout, the T cell mitogen, Con A, was used as a positive control, whereas BSA served as a negative control to show the background level of proliferation. Both BMDC populations induced comparable levels of OT-II TCR-Tg CD4+ T cell proliferation after pulsing with P. pastoris-derived OVA Ags (Fig. 5). As expected, the number of T cell replications stimulated by ecOVA did not significantly differ in wells containing DCs from WT vs MMR-deficient mice (2.8 ± 0.03 and 2.4 ± 0.1, respectively). Due to the presence of other MRs, such as DC-SIGN, these data suggest that there are redundant roles for these receptors and illustrate the difficulty in teasing out the contributions of each. We conclude that although the MMR may play a role in the uptake of mannosylated Ag, other MRs are likely to be involved as well.
In the current study we examined the ability of heterogeneously mannosylated Ags produced from the yeast P. pastoris to enhance Ag-specific T cell responses. A substantial amount of evidence has accumulated in support of the capacity for mannosylation of Ags to augment Ag-specific CD4+ and CD8+ T cell-mediated immune responses. The basis for this enhancement of immune responses has been attributed to the ability to direct Ags to MRs on the surface of DCs. As mentioned previously, mannosylated Ags, such as BSA and Mycobacterium leprae HSP65 protein, have been shown to increase MHC class II-restricted T cell responses by 100- to 200-fold over their nonmannosylated counterparts (3, 7). In contrast, mice immunized with mucin 1 conjugated to oxidized mannan induced CTL responses and tumor protection (8, 22, 23, 24, 25). Furthermore, mannan-coated liposomes containing DNA or protein have been shown to elicit CTL responses and hallmarks of Th1-mediated immunity, such as delayed-type hypersensitivity responses and the secretion of cytokines (IL-12 and IFN-γ) (4, 5, 9, 26, 27). These studies incorporate the conjugation of mannose or mannan onto Ags, but did not examine the potential benefits of differential mannosylation using a system (e.g., fungi) that naturally mannosylates its glycoproteins. We have shown that cryptococcal mannoproteins elicit T cell responses that are dependent upon binding to MRs (such as MMR and DC-SIGN) on DCs (Ref.14).4 Because whole Saccharomyces cerevisiae expressing recombinant OVA protein has been shown to induce protective cell-mediated immunity (28), we were interested in testing the effects of mannosylated Ags produced from a fungal source.
Carbohydrate analysis of the four mannosylated recombinant proteins produced in P. pastoris (ppOVA, ppOVAST, ppOVANQ, and ppOVASTNQ) revealed differential levels of mannosylation, whereas, as expected, the two unglycosylated recombinant counterparts produced in E. coli (ecOVA and ecOVAST) displayed no mannosylation. N-linked oligosaccharides in P. pastoris have been reported to consist of eight to 14 mannose residues, whereas O-linked oligosaccharides are shorter, averaging two or three mannose residues (20, 29, 30). Furthermore, N-linked oligosaccharides are branched, whereas O-linked oligosaccharides are usually synthesized as a single chain. Importantly, mannose is exclusively found as the terminal sugar for both N- and O-glycans on glycoproteins secreted by P. pastoris strains. Thus, the variability in glycan structure may have an impact on the ability of the mannosylated Ags to bind certain MRs (e.g., MMR and DC-SIGN) and influence Ag presentation as each MR traffics Ag into its own respective pathway.
Our results demonstrate that fungal mannosylation, in the form of either N-linked or extensive O-linked glycans, can increase the efficiency of an Ag-specific T cell response by targeting MRs. Because DCs are thought to function as a critical link between innate and adaptive immune responses, we were interested in studying whether DC MRs, such as MMR and DC-SIGN, were involved in the uptake and presentation of mannosylated recombinant proteins. The experiment design, which used Ag pulsing of DCs, mimics the situation in vivo, where diffusion of Ags may limit the time that Ag is available extracellularly. Pulsing would be expected to bias uptake of the mannosylated Ags toward a receptor-mediated process as opposed to the less efficient constitutive process of macropinocytosis (21). Nevertheless, molecules endocytosed via macropinocytosis still traffic into lysosomes and MHC class II-enriched compartments (7), and this process probably accounted for the moderate amounts of lymphoproliferation stimulated by the E. coli-derived proteins.
P. pastoris has been recently used by the biopharmaceutical industry to produce glycoproteins for therapeutic purposes (31). This system has the benefit of yielding large scale preparations of glycosylated recombinant proteins at a lower cost than mammalian expression systems. However, because the oligosaccharides contain exposed mannose residues, they have the troublesome potential of being immunogenic, thus limiting long-term use. In an attempt to circumvent this problem, yeast have been genetically engineered to produce more humanized glycoproteins (31, 32, 33). In contrast, our approach, which exploits the mannosylation behavior of fungi to augment Ag immunogenicity, takes advantage of the immune recognition of exposed mannoses as pathogen-associated molecular patterns.
The heightened T cell responses seen with mannosylated Ags are presumably due to the efficient uptake through MRs on DCs. S. cerevisiae-derived yeast mannans serve as efficient competitive inhibitors by binding to MRs (12, 14, 21). Our data implicate MR involvement in the uptake of P. pastoris-derived mannosylated Ags, because inhibition of T cell proliferation was observed with the addition of yeast mannans. Moreover, although MMR may play a role in the uptake of mannosylated Ag, it is not solely responsible for this process, because T cell proliferation was not found to be significantly different when MMR KO BMDCs were compared with WT BMDCs. These results suggest the involvement of other MRs, such as DC-SIGN. The contribution of DC-SIGN may be difficult to determine in mice, because there are five murine homologues to human DC-SIGN, three of which (mDC-SIGN, SIGNR1, and SIGNR3) include a Glu-Pro-Asn motif that is presumed to recognize mannose residues (34, 35). Although langerin has been demonstrated to be a putative MR (36, 37, 38), the contribution of langerin on the BMDCs used in the current study is probably negligible. MHC class I and class II Ag presentation was not impaired in Langerhans cells from langerin KO mice (39). In addition, <5% of BMDCs on day 6 of culture have been reported to express langerin (40, 41).
Although fungal mannosylation increased T cell proliferation in our studies, there are some caveats that could be relevant to vaccine design. Studies have shown that although some ligand-MR interactions lead to strong Th1 and CTLs responses, others result in anti-inflammatory or anergic responses (6, 42). Human monocyte-derived DCs stimulated with an anti-MMR mAb induced the production of anti-inflammatory cytokines, including IL-10, and inhibited IL-12 (42, 43). Furthermore, treatment of DCs with anti-MMR mAb induced T cell anergy and caused them to behave as regulatory/suppressor T cells (42). Targeting through DC-SIGN has also been shown to elicit immunosuppressive responses and is used by some pathogens to evade the immune system. For example, upon binding of DC-SIGN to mannosylated lipoarabinomannan from Mycobacterium tuberculosis, DC maturation and IL-12 secretion were suppressed (44). However, recent work by Nagaoka et al. (45) suggested that SIGNR1 and TLR4 act together to capture Gram-negative bacteria via LPS and deliver a danger signal. β1,2-linked mannose has been reported to be immunogenic in vivo, and β1,2-mannobiose capped O-linked oligosaccharides have been detected on recombinant human bile salt-stimulated lipase expressed in P. pastoris (30). Thus, it will be interesting to determine whether our P. pastoris-derived mannosylated proteins contain components that can alert the immune system.
Finally, glycans can inhibit proteolysis by masking cleavage sites, thus shifting the pattern of peptide fragments processed by the DC (46). Moreover, O-glycans can remain on MHC-presented peptides, which can then interfere with peptide recognition by T cells (47). Thus, it will be important to determine the nature of the in vivo immune response that mannosylated fungal Ags stimulate, particularly with regard to whether differences exist between Ags rich in O-glycans and those with N-glycans. In conclusion, our data provide a proof-of-principle that by targeting MRs on DCs, fungal mannosylation can be exploited to boost immunogenicity of a model Ag. Multiple approaches were taken to demonstrate the link between mannosylation and immunogenicity, including the comparison between mannosylated P. pastoris and unglycosylated E. coli-derived proteins, site-directed mutagenesis of N-linkage sites, the inhibitory effect of yeast mannans, and the loss of immunogenicity after deglycosylation of P. pastoris-derived protein. Immunogenicity, as measured by T cell proliferation, does not necessarily correlate with protective immunity, and it will be important to determine whether mannosylation contributes to protective immunity using in vivo models. Nevertheless, the use of fungal systems to mannosylate vaccine candidates seems to provide a potent immunostimulatory platform for the development of vaccines and is worthy of further study.
We thank Jianmin Chen and Dr. Shuhua Nong for excellent technical support, Dr. Ann Marshak-Rothstein for the DO11.10 mice and reagents, and Dr. Michel Nussenzweig for the gift of the MMR KO mice.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by National Institutes of Health Grants RO1AI25780, RO1AI066087, RO1AI37532, and T32AI07642.
Abbreviations used in this paper: MR, mannose receptor; BMDC, bone marrow-derived DC; DC, dendritic cell; DC-SIGN, DC-specific ICAM-3-grabbing nonintegrin (CD209); ec, E. coli derived construct; KO, knockout; LN, lymph node; MMR, macrophage MR (CD206); MP, mannoprotein; N, asparagine; PNGase F, peptide:N-glycosidase F; ppOVA, P. pastoris-derived protein containing the OVA portion; ppOVANQ, P. pastoris-derived counterpart of ppOVA lacking the N-linked glycan site; ppOVAST, protein consisting of the OVA portion fused to the S/T-rich region produced in P. pastoris; ppOVASTNQ, P. pastoris-derived counterpart of ppOVAST lacking the N-linked glycan site; S/T, serine/threonine; Tg, transgenic; WT, wild type.
M. K. Mansour, E. Latz, and S. M. Levitz. Cryptococcus neoformans glycoantigens are captured by multiple lectin receptors and exclusively presented by dendritic cells. Submitted for publication.