Epitope-Specific Antibody and Suppression of Autoantibody Responses Against a Hybrid Self Protein¹

An immune response against complex protein Ags is typically dominated by Abs and T cells specific for a limited number of epitopes. The identification of neutralizing epitopes of a pathogen suggests the development of vaccines incorporating such sites to induce a more efficient protective response. This strategy could also apply to vaccines that target vulnerable sites of a pathogen that may escape recognition in the natural immune response to infection. Sequences describing B and T cell epitopes of a polypeptide are best presented in the context of a carrier protein to define relevant conformations of the B cell epitope and to enhance uptake and processing of T cell epitopes. Traditional carrier proteins have been selected for their ability to elicit a potent Th response directed by autologous epitopes. However, these proteins are also likely to contain numerous B cell epitopes, which can account for competing anticarrier responses. The introduction of heterologous T cell epitopes into a protein or peptide by genetic engineering (1, 2) or synthetic coupling (3, 4) is a general strategy for boosting immunogenicity of a molecule. Recently, self Igs have been employed as carriers for heterologous epitopes to elicit Abs or T cells (5, 6). The Ig scaffold could confer stability or improve uptake of Ags, and it may also be effective for immunotargeting of APCs (7). Conversely, self Ags presenting foreign Th epitopes also have the potential to induce autoantibodies, suggesting breakdown of tolerance that could impede their use as carriers (8, 9, 10). The balance between immunity and autoimmunity in the response to hybrid self Ags could suggest a possible relationship of clonal selection to tolerance mechanisms.

Somatic mutations of V region genes during an immune response provide an opportunity for self-reactive B cells to arise. Several mechanisms of B cell regulation have been postulated to impose peripheral tolerance, including negative signaling by the self Ag and receptor editing (11, 12, 13). However, these mechanisms may not be available to cells recognizing self proteins that are sequestered or present at low concentration (14). Clonal dominance plays a significant role in the regulation of Ab diversity generated in response to antigenic stimulation. Epitope-specific B cells competing for diminishing Ag could establish a hierarchy for memory responses that includes self-reactive clones. The potential for expression of autoantibodies may therefore depend on the availability of alternative B cell epitopes and the delivery of T helper activity to the respective B cell clones.

The distinction of autologous and heterologous epitopes of a hybrid self protein by B lymphocyte clones could be investigated using a vaccine model in which an autoantibody response can be established. Ubiquitin is a highly conserved self protein found in all eukaryotic cells (15). It has a central role in regulating intracellular protein turnover and may also be functional at the cell surface (16, 17, 18). Recent studies have been reported suggesting bypass of tolerance to ubiquitin in mice immunized with hybrid proteins in which foreign Th epitopes were grafted into the ubiquitin sequence (9). We adapted this system by modifying the immunogen to superimpose a new heterologous B cell epitope on the antigenized self protein. It was necessary to define a second insertion site independent of the T cell epitope to permit comparison of the response in the presence and absence of the foreign B cell epitope. It was further necessary to maintain a stable, native-like folding of the ubiquitin hybrid protein to preserve the autologous epitopes. Fusion of short sequences to the ubiquitin C terminus is minimally disruptive to folding and provides a potential advantage for T cell epitope use, as only a single proteolytic cleavage is required for processing. We chose to include a universal Th epitope to provide a strong response independent of the MHC haplotype (19). Although processing of exogenous Ags occurs in lysosomal compartments, we anticipated the possibility that cleavage of artificial ubiquitin fusions by intracellular ubiquitin-specific proteases (UBPs)⁴ could also influence presentation (20). We therefore prepared Ags having a mutation at the ubiquitin C-terminal residue to inhibit cleavage by UBPs (21). The exceptional stability of the ubiquitin fold was further exploited for engineering of a unique B cell epitope. Insertion of a heterologous sequence in a conformationally restricted environment on the surface of a carrier protein is known to enhance its presentation as a B cell epitope (22). Accordingly, a surface loop on ubiquitin was replaced with a short loop sequence from HIV gp120, which elicits a good Ab response in the native and heterologous contexts (23, 24). We demonstrate that a significant autoantibody response is induced by ubiquitin fusions in which only the C-terminal Th epitope is included. However, hybrid molecules containing the gp120 loop insert elicited a specific response against the new B cell epitope. Remarkably, the autoantibody response was completely suppressed despite preservation of the ubiquitin epitopes and use of equivalent T cell help. These results support the use of self proteins as carriers for epitope-specific vaccines and suggest a possible mechanism to avert autoimmunity in the B cell response.

An expression plasmid pDSUb encoding the ubiquitin gene under the control of the lac promoter was provided by Dr. Martin Rechsteiner (University of Utah). Expression plasmid pJT184, used for co-expression of UBP1 (25), was a gift from Dr. Alex Varshavsky (California Institute of Technology, Pasadena, CA). Oligonucleotides used for mutagenesis and construction of genetic fusions were obtained by custom synthesis through Bioserve Biotechnologies (Laurel, MD). The 3′end of the ubiquitin gene was modified by cloning of annealed oligonucleotides 5′-TGTTGTTA AACTGTCTGACGCTCTGTAAGCTTCTGCA-3′ and 5′-GAAGCTTA CAGAGCGTCAGACAGTTTAACAACAGCCGGCGGCA-3′ combined with 5′-TAAGACTGCGTGGCGGCGACCAGGTTCACTTC CAGCCGCGCCGCCGGC-3′ and 5′-GCGGCTGGAAGTGAACCTG GTCGCCGCCACGCAGTC-3′ or 5′-TTAAGACTGCGTGGCGCTGAC CAGGTTCACTTCCAGCCGCTGCCGCCGGC-3′ and 5′-GCGGCTG GAAGTGAACCTGGTCAGCGCCACGCAGTC-3′ between AflII and PstII sites in pDSUb to obtain pDSUbgMT or PDSUbaMT, respectively. Modification of the ubiquitin gene at codon 35 was initially performed in pRSETUb, derived from pRSET (Invitrogen, San Diego, CA) by subcloning of the ubiquitin gene from pDSUb using the NdeI and HindIII restriction sites. Plasmid pPX153 containing BpmI and XcmI restriction sites spanning codon 35 in a noncoding ubiquitin gene was created by in vitro mutagenesis using synthetic oligonucleotide 5′-GCGAAAATCCAG GATAAAGCTGGAGGTTAACCGCCGGATCAGCAGCGTC-3′, pRSE TUb as template, and the MORPH mutagenesis kit (5 Prime-3 Prime Inc., Boulder, CO). Complementary oligonucleotides 5′-AAGAAATCCACATCGGTCCGGGTCGTGCTTTCTACACCACCATCCCGCCGGAT CA-3′ and 5′-ATCCGGCGGGATGGTGGTGTAGAAAGCACGACCCGGACCGATGTGGATTT- 3′ were annealed and cloned in pPX153 between BpmI and XcmI sites to obtain pRSETUbV3. A fragment in the UbV3 coding region between AflII and BglII sites in pRSETUbV3 was isolated and cloned at equivalent sites in pDSUb, pDSUbgMT, or pDSUbaMT to obtain expression plasmids pDSUbV3, pDUbV3gMT, and pDSUbV3aMT, respectively. Digested vector fragments were treated with calf intestinal alkaline phosphatase and purified from agarose gel slices after electrophoresis using the Geneclean kit (BIO 101, San Diego, CA). All restriction digests, ligations, and transformation of Escherichia coli were done according to standard manipulations (26). Commercial DNA modifying enzymes were used as instructed by the manufacturer (New England Biolabs, Beverly, MA). Correct cloning of the oligonucleotide insertions was confirmed by DNA sequencing using Sequenase kit version 2.0 (United States Biochemical, Cleveland, OH).

Cultures of E. coli (A61) harboring ubiquitin fusion expression plasmids were grown and induced by addition of isopropyl β-d-thiogalactoside (IPTG), as previously described (27). Cells were disrupted by sonication in lysis buffer (20 mM MES, pH 5.5, supplemented with 1 mM PMSF and 0.5 mM ZnCl₂) at 4°C twice for 4 min at a 50% cycle. Homogenates were centrifuged at 15,000 × g, 4°C for 45 min. Supernatants were diluted 3-fold with lysis buffer, filtered through a 0.45-μm filter, and loaded onto an SP-Sepharose HP ion exchange column at a linear flow rate of 1.4 cm/min. Fusion proteins were eluted over 16 column volumes with a 0–0.5 M NaCl gradient in 20 mM MES (pH 5.5). Fractions were assayed by SDS-PAGE on 10–20% tricine gels (Novex, San Diego CA), and protein concentration was determined by the BCA method (Pierce, Rockford IL) using bovine ubiquitin for a standard curve. Cleavage of hybrid proteins in vivo was determined by selection of recombinant E. coli cotransformed with pJT184 and a hybrid ubiquitin expression plasmid, induction of a 25 ml culture with IPTG for 2 h, collection of cells by centrifugation, lysis in gel loading buffer (Novex), and analysis by SDS-PAGE. An in vitro assay for cleavage of ubiquitin hybrid proteins was performed by incubating 100–200 μg of a ubiquitin hybrid and 2 μg of purified recombinant UCH-L3 (28) in 50 mM Tris, 5 mM DTT, 1 mM EDTA, pH 8 (reaction buffer) at 37°C. The reaction was monitored by SDS-PAGE or by reverse-phase (RP)-HPLC on a Vydac 218TP54 column eluting with a gradient of 20–50% acetonitrile in water containing 0.1% trifluoroacetic acid over 10 min at 1 ml/min flow rate.

Bovine ubiquitin (Sigma, St. Louis, MO) was dissolved in PBS and filtered through a 0.2-μm filter. Synthetic peptides 1–20, 13–30, 24–41, 32–49, 40–57, 52–71, 64–81, and 72–87 representing overlapping sequences of UbV3 were obtained in 70–90% purity by conventional automated solid phase synthesis through a commercial service (Genemed Synthesis, South San Fransisco, CA). Synthetic peptide KRIHIGPGRAFYTTK (V3) was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health. Peptide DQVHFQPLPPAVVKLSDAL (MT) was prepared by a large-scale digest of purified UbgMT with an aliquot of UCH-L3 in reaction buffer (28). The product was isolated by preparative C18 RP-HPLC, as described above.

Female BALB/c, C57BL/6, and C3H/HeN mice 6–8 wk old were obtained from Charles River Laboratories (Frederick, MD) or The Jackson Laboratory (Bar Harbor, ME) and housed in a dedicated facility under supervision from the Institutional Animal Care and Use Committee. Mice were inoculated in groups of three or five per Ag by an initial s.c. injection of 100 μg of proteins in PBS emulsified in an equal volume of CFA or IFA. Booster injections of proteins (100 μg) in IFA were delivered i.p. 3 wk and 7 wk later. Blood samples were collected at 4, 6, 8, and 10 wk from the initial injection. Serum was separated by centrifugation and stored at −20°C.

Proteins (10 μg/ml) or synthetic peptides (10 μg/ml) in PBS were dispensed at 0.1 ml/well into immunosorbent 96-well flat-bottom plates (Corning, Acton, MA) and incubated for 1 h at 37°C. Excess Ag was decanted, and nonspecific sites were blocked by addition of BSA (10 mg/ml in PBS, 0.2 ml/well) incubated 30 min at 37°C. Plates were washed three times with 10mM Tris buffer and 0.1% Tween 20 (pH 8), and serially diluted serum samples in PBS supplemented with 1% BSA were dispensed at 0.1 ml/well. After 1 h at 37°C, plates were washed as before and developed with 0.1 ml/well of affinity purified goat anti-mouse IgG-horse radish peroxidase conjugate (Promega, Madison, WI) diluted 1:2000 in PBS supplemented with 1% BSA. Washing was repeated, and bound enzyme was detected with 1 mg/ml o-phenylenediamine in 0.05 M phosphate-citrate buffer and 0.02% H₂O₂ (pH 5). Plates were read at 450 nm on a 96-well plate reader (Titertek, Huntsville, AL). Titers were expressed as the dilution of serum giving an absorbance reading of 0.3, or 15% of the maximum reading. Solution phase binding assays were performed by incubating peptides or proteins at concentrations ranging from 0 to 8 mg/ml with antiserum at fixed concentration in the range of its titer in 0.1 M potassium phosphate, 2 mM EDTA, 10 mg/ml BSA (pH 7.8). Samples were incubated at 37°C for 2 h, then applied to Ag-coated ELISA plates. The standard ELISA procedure was followed to determine the concentration of unbound Ab relative to samples containing no added ligand.

A previously described assay was used with the modifications noted (29). Six-week-old female mice (BALB/c or C3H/Hen) were immunized in groups of three by foot pad injection with 50 μg of proteins in PBS emulsified in IFA. Eight to 10 days later, mice were sacrificed and popliteal lymph nodes (LN) were removed aseptically. Pooled LN cells from three mice were washed with HBSS, suspended in RPMI 1640 (Life Technologies, Gaithersburg, MD) supplemented with 100 U penicillin/ml, 100 μg streptomycin/ml (Life Technologies), 1% syngeneic normal mouse serum (Sigma), 2 mM l-glutamine, 0.05 mM 2-ME (Sigma), and 1 mM sodium pyruvate (Life Technologies) and dispensed into 96-well cell culture plates in 200 μl aliquots of 5 × 10⁵ cells/well. Proteins, peptides (10–200 μg), or Con A (5 μg/ml) in 20 μl PBS were added in triplicate wells, and plates were kept in a CO₂ incubator at 37°C for 4 days. Wells were then pulsed with 1 μCi/well of [³H]thymidine (Amersham, Arlington Heights, IL) for 18 h, and cells were harvested on filter pads. Counts were read on a Wallac microbeta plate reader (Wallac, Gaithersburg, MD). Stimulation was expressed as the average signal of triplicate wells corrected for mean background cpm of cells incubated without peptide.

Distinct sites within the ubiquitin sequence for insertion of a linear B cell epitope and a T cell epitope were selected on the basis of structural considerations and precedents for preservation of the native fold. The flexible C terminus permits a fused Th epitope to be presented apart from other structural determinants, whereas a surface site within residues 34–40, previously shown to accomodate an enlarged loop (30), was deemed suitable for conformational display of the B cell epitope. A modified ubiquitin gene in which codon 35 is replaced with a sequence coding for residues 312–323 of HIV-1 gp120 MN was created in pRSETUbV3, obtained by insertion of complementary oligonucleotides between XcmI/BpmI restriction sites of pPX153. The sequence for residues 350–368 of the Mycobacterium tuberculosis 38-kDa protein was fused to the ubiquitin C terminus while reconstructing seven C-terminal ubiquitin residues by inserting one of two alternative sets of complementary oligonucleotides between AflII and PstI sites of pDSUb to provide either pDSUbgMT or pDSUbaMT. The latter provides a Gly→Ala mutation at codon 76 of ubiquitin. These plasmids were used to express linear fusions UbgMT and UbaMT having the native or mutant cleavage site for a UBP. A DNA fragment of pRSETUbV3 spanning the modified region of the ubiquitin gene was ligated to pDSUb, pDSUbgMT, or pDSUbaMT using the common BglII and AflII sites to generate expression vectors pDSUbV3, pDSUbV3gMT, and pDSUbV3aMT, respectively. These were used for expression of soluble proteins UbV3 and double-epitope hybrid proteins UbV3gMT and UbV3aMT in recombinant E. coli. Fig. 1 illustrates the linear arrangement of the foreign sequences with respect to the ubiquitin polypeptide in the different hybrid proteins. Fusion proteins were purified from lysates by ion exchange chromatography in yields ranging from 185 mg/l to 336 mg/l of culture. Purities were determined to be >98%, as judged by SDS-PAGE and RP-HPLC (Fig. 2).

Cleavage of ubiquitin fusions at the ubiquitin-polypeptide junction by UBPs was examined as a diagnostic test for the folded conformation of the ubiquitin hybrid protein (31). E. coli were transformed with a mixture of a recombinant ubiquitin expression plasmid and pJT184 to obtain recombinants coexpressing UBP1 and a ubiquitin hybrid protein. Cultures of bacteria expressing the ubiquitin hybrid protein alone or in combination with UBP1 were grown and induced under identical conditions. Expression of the desired proteins was confirmed by SDS-PAGE analysis of lysed cells. Cleavage of ubiquitin hybrids containing a C-terminal extension by UBP1 was apparent from the appearance of a product band of appropriate size (Fig. 2,A). The mutant ubiquitin hybrid UbaMT was completely resistant to cleavage, demonstrating stringent specificity of the UBP for the native recognition site (21). The hybrid UbV3gMT was also cleaved, although more slowly than UbgMT. Similar observations were made in an in vitro cleavage assay using purified ubiquitin hybrid proteins and purified recombinant UCH-L3 as the specific UBP (Fig. 2 B). Under these conditions complete cleavage of UbV3gMT and partial cleavage of UbaMT was observed, suggesting reduced specificity of this enzyme in the in vitro format. These results suggested that the ubiquitin tertiary structure is largely preserved in the modified proteins. The high level expression of these soluble hybrid proteins is also consistent with this conclusion as structural stability of the ubiquitin fold is thought to enhance expression yields (32).

The contribution of ubiquitin autoantibodies to the response elicited by ubiquitin hybrid proteins described in Fig. 1 was determined by ELISA, comparing the specific binding on native ubiquitin and on hybrid molecules presenting epitopes in the immunogen. Inbred mice (BALB/c) were immunized in groups of three by an initial injection and two booster injections delivered i.p. at 3 and 7 wk. Sera were collected 1 and 3 wk after each boost. Diluted antisera were assayed against immobilized UbV3, UbgMT, or unmodified bovine ubiquitin. Autoantibody was a significant component of the immune response in mice treated with UbaMT or UbgMT (Fig. 3,A). In contrast, the immune response to UbV3gMT, UbV3aMT, or UbV3 was highly specific for UbV3 (or UbV3gMT) but not for ubiquitin or UbgMT. A much lower anti-ubiquitin response was observable in some mice, but it declined and became negligible with further boosting (Fig. 3,B). This was also seen in mice immunized with recombinant or bovine ubiquitin and is consistent with a tolerogenic response (data not shown). ELISA titers of ubiquitin-specific autoantibodies and UbV3-specific Abs in mature antisera were compared. Autoantibodies could be detected at dilutions >1:32,000, whereas anti-UbV3 titers were in excess of 1:10⁵, suggesting a 5- to 10-fold greater avidity of Abs in the latter antisera (Fig. 4). Abs to UbV3 were shown to be IgG and IgM in the primary response, while in the mature antisera IgG2b was the predominant isotype.

Experiments in C57BL/6 (H-2^b) and C3H (H-2^k) mouse strains demonstrated analogous anti-ubiquitin responses in mice receiving UbaMT and anti-UbV3 responses in mice immunized with UbV3aMT. Antisera of mice receiving UbV3aMT demonstrated a high level of discrimination between UbV3 and ubiquitin, suggesting a dominant response against the heterologous B cell epitope as well as suppression of Abs against native epitopes of ubiquitin (Fig. 5). Reactivity against the single-epitope hybrid UbgMT was also minimal indicating that other potential B cell epitopes introduced with the MT sequence did not participate in the response. The single-epitope hybrid UbV3 and double-epitope hybrid UbV3gMT were significantly less effective at eliciting a response in these mice. Anti-V3 epitope-specific responses were <30% of the levels found in antisera generated to UbV3aMT.

To preclude possible bias in binding to solid phase-adsorbed Ags, the relative specificity for soluble Ags was investigated by indirect measurement of Ab binding. Diluted antisera were incubated with ubiquitin, UbV3, or V3 peptide at varying concentrations, and residual unbound Ab was measured by capture on ubiquitin or UbV3-coated ELISA plates. Autoantibodies elicited against UbaMT in BALB/c mice recognized epitopes of native ubiquitin as demonstrated by competition binding with soluble bovine ubiquitin. Binding to soluble UbV3 could also be observed, confirming the integrity and accessibility of at least some of the native ubiquitin epitopes on the internally modified ubiquitin molecule (Fig. 6,A). The specificity of high titer antisera to UbV3 for a unique epitope, suggested by the large difference in reactivity to immobilized UbV3 and ubiquitin (Fig. 4), as well as by binding to solid phase-adsorbed V3 peptides (data not shown), was supported by the solution phase binding. A high level of discrimination was apparent for an epitope or epitopes displayed by UbV3 but not ubiquitin (Fig. 6 B). Binding to the V3 peptide, which could represent part of a conformational or discontinuous determinant, was observed in the low micromolar range of peptide concentration. In contrast, no competition was seen in the presence of 100 μM soluble native ubiquitin.

To determine whether the autoantibody response and alternative immune responses are mediated by equivalent T cell help, we tested the ability of proteins or synthetic peptides providing putative Th epitopes to restimulate T cells of immunized BALB/c mice. Mice primed with UbV3 or UbaMT provided LN cells for in vitro proliferation assays in the presence of either recombinant protein or unmodified ubiquitin. A dose-dependent proliferative response was observed in cells incubated with the protein to which the mice were sensitized (Fig. 7). No obvious stimulation was seen with 100 μg/ml of a ubiquitin hybrid or recombinant ubiquitin, which lack the heterologous epitope of the immunogen. To control for sensitization to potential bacterial contaminants in the recombinant protein, mice were treated with similarly purified recombinant ubiquitin. No significant proliferation of LN cells was observed in the presence of either recombinant ubiquitin from the same stock or hybrid proteins UbaMT or UbV3, indicating that E. coli-derived mitogens do not account for T cell responses (data not shown).

Strong immunogenicity of UbV3 in BALB/c mice suggested that a Th response could be directed against peptides containing the V3 sequence or autologous sequences that are cryptic epitopes in native ubiquitin. A set of overlapping synthetic peptides representing the entire sequence of UbV3 was prepared to localize the T cell epitope. Serially diluted peptides in PBS were incubated with LN cells of UbV3-primed BALB/c mice, and proliferation was determined. Values at highest peptide concentrations were compared with stimulation afforded by intact UbV3. Peptide 40–57, and to a lesser extent, peptide 24–41, which map to sequences spanning the junction sites between V3 and flanking ubiquitin sequences, provided significant proliferative responses (Table I).

We compared T cell responses in C3H mice in which UbV3 was a poor immunogen to ascertain that the universal Th epitope can also contribute to B cell responses against the heterologous epitope. LN cells from C3H mice primed with UbV3aMT proliferated only weakly on incubation with UbV3. Incubation in the presence of increasing concentration of MT peptide, but not the V3 peptide, produced a significant proliferative response (Fig. 8). No stimulation was observed in the same experiment using LN cells from C3H mice immunized with UbV3. Collectively, these data provide evidence that T cell help that can break B cell tolerance is efficiently diverted to the B cell response against a dominant heterologous epitope.

The immunogenicity of genetically engineered hybrid proteins is of interest for development of subunit vaccines. Concerns in the choice of a carrier protein include the potential for irrelevant anticarrier responses and suppression specific to the heterologous epitope (33, 34). Presentation of foreign epitopes on self proteins offers certain advantages. Ig molecules are effective as carriers for foreign B and T cell epitopes (5, 6). However, the potential for autoimmune responses has not been addressed. This issue is further complicated by evidence that engineered Igs can also be used to impart tolerogenicity to heterologous epitopes (35). Autoantibody responses elicited by modified self proteins presenting foreign Th epitopes provide an experimental model for breakdown of B cell tolerance (9). We antigenized the self protein ubiquitin by successive modifications to provide T and B cell epitopes and observed a shift in immunogenicity from an autoantibody response to a response converging on the foreign epitope. The experiments demonstrate regulation of autoantibody responses in a system where breakdown of tolerance is possible. These findings could suggest a strategy for epitope-specific vaccine development.

Modifications to ubiquitin were designed to introduce the B and T cell epitopes independently while minimally affecting the native epitopes. Conjugation to the C terminus has been used in numerous studies in which the ubiquitin structure and function are preserved (21, 36, 37). A mycobacterial sequence fused to the flexible ubiquitin C-terminus was expected to provide a strong Th response while not contributing an epitope for Ab responses (38, 39). A surface-exposed loop (residues 34–40) that bridges well-defined secondary structures in the folded protein (40) was selected as an alternative site for presentation of a conformationally restricted B cell epitope. This site was suggested by studies on “split” ubiquitin in which association between an N-terminal and a C-terminal fragment was used to enforce proximity or other constraints on inserted sequences (30). We anticipated that similar insertion of a gp120 sequence, which has a natural loop conformation and is a neutralizing determinant in its native context, could provide a new target for B cell responses. Cleavage of a C-terminal ubiquitin fusion substrate by eukaryotic UBPs requires the native-like folding of the ubiquitin polypeptide (31). We established that the hybrid protein containing C-terminal fusion and V3 insert is cleaved by UBPs to support the view that the insert does not interfere with folding. A reduced rate of in vivo hydrolysis of UbV3gMT relative to that of UbgMT could indicate a steric effect due to the loop at the interface in the protease-substrate complex.

Autoantibodies of high titer to ubiquitin were induced in mice immunized with the ubiquitin hybrid containing only a C-terminal epitope. The autoantibodies also reacted strongly with UbV3, demonstrating that the internal insert does not interfere with binding to autologous epitopes (Fig. 3,A). By contrast, animals immunized with UbV3 or UbV3aMT produced strong antisera selective for UbV3 and a conspicuous absence of Ab reactive with native ubiquitin (Fig. 3,B). The lack of an autoantibody response was equivalent to the nonresponsiveness of mice immunized with unmodified ubiquitin. Divergent responses against these closely related Ags cannot be explained as the regulation of autoreactive B cells by endogenous ubiquitin. We considered the possibility that insertion of the V3 sequence alters Ag processing or provides alternative T cell epitopes that fail to elicit Th responses cooperative with autoreactive B cells. A strong response to UbV3 in BALB/c mice and stimulation of their T cells by peptides mapping to the junction between the V3 insert and the ubiquitin sequence implicated the presence of such epitopes. However, this explanation does not account for analogous deviation of autoantibody responses in other mouse strains where the V3 epitope-specific responses elicited with UbV3aMT and ubiquitin autoantibody responses elicited with UbaMT can be attributed to the common Th epitope. These observations suggest that both autoimmune and V3 epitope-specific responses can be directed by similar Th cell activity generated against the MT epitope. Proliferation assays with UbV3aMT-sensitized C3H mice confirmed stimulation of T cells by a peptide representing this epitope (Fig. 8). Furthermore, the results indicate that the heterologous epitope-specific response can be driven by other Th epitopes derived from processing of internal sequences in the Ag.

Competition by V3 peptides for binding of high titer antisera to UbV3 suggests that the linear sequence is a part of the dominant epitope (Fig. 6,B). The high peptide concentrations required for significant competition are indicative of a large difference in avidity for the constrained and unconstrained V3 sequences. Therefore, we concluded that the anti-UbV3 antisera recognize a conformational or discontinuous determinant comprised of residues in the V3 loop. The possibility that conformationally altered ubiquitin sequences also contribute to the response cannot be excluded. Low reactivity of anti-UbV3aMT antisera to UbMT shows that neither ubiquitin determinants nor sites introduced with the C-terminal fusion compete effectively with the constrained V3 loop as B cell epitopes (Fig. 5). The high-titer epitope-specific response and virtual absence of alternative “anticarrier” responses (Fig. 4) is remarkable for a soluble immunogen presenting a single copy of the new epitope. These data suggest unique advantages of ubiquitin as a carrier for construction of epitope-specific subunit vaccines.

Cleavage by UBPs is a plausible explanation for reduced immunogenicity of UbV3gMT relative to UbV3aMT in C3H and C57BL/6 mice (Fig. 5). One interpretation of this result is that UBP-mediated processing depletes the Ag by generating a substrate for rapid degradation by the N-end rule pathway (41, 42). Alternatively, the cleavage product may be less accessible to MHC class II presentation than peptides produced by lysosomal enzymes. Intracellular processing of ubiquitin conjugates and linear fusions has been implicated in the presentation of MHC class I-restricted epitopes (43, 44). However, exogenous Ags are generally excluded from this pathway. It is not known whether engineered ubiquitin fusion proteins could improve vaccines for an MHC class I-restricted response to exogenous Ags. Nonetheless, poor presentation of the C-terminal class II-restricted epitope due to processing by UBPs appears to account for reduced immunogenicity of the UBP-cleavable hybrid proteins.

Ubiquitin is highly conserved and universally expressed in eukaryotic cells. Tolerance in both T and B cell compartments may therefore be established early in development. Although autoreactive B cells can arise during vaccination, we have shown that heterologous epitope dominance imposes a counterselective pressure. The mechanism of this effect can speculatively be attributed to affinity-driven clonal competition for T cell help. Titers suggesting higher avidity of V3 epitope-specific Abs than ubiquitin autoantibodies are consistent with this possibility (Fig. 4). The Ag-presenting function of mature B cells has previously been suggested to play a role in carrier-induced epitopic suppression (45). Deletion of autoreactive B cells from the pre-immune repertoire could facilitate heterologous epitope-specific responses, as the affinity maturation of autoantibodies might then require more extensive somatic mutation than Abs to a foreign epitope. Clonal dominance in the immune response to foreign Ags could therefore provide another means to restrict autoantibody formation.

We thank James Ahn and Miriam Rogers for technical assistance, Mary Kuchera for assistance with T cell assays, Drs. John Kenten and Richard Massey (IGEN Inc.), and Dr. Moncef Zouali (Pasteur Institute) for helpful discussions and review of the manuscript.