Identification of Key Amino Acids of the Mouse Mammary Tumor Virus Superantigen Involved in the Specific Interaction with T-Cell Receptor V_β Domains

Abstract

Mouse mammary tumor virus (MMTV) is a retrovirus encoding a superantigen that is recognized in association with major histocompatibility complex class II by the variable region of the beta chain (V_β) of the T-cell receptor. The C-terminal 30 to 40 amino acids of the superantigen of different MMTVs display high sequence variability that correlates with the recognition of particular T-cell receptor V_β chains. Interestingly, MMTV(SIM) and mtv-8 superantigens are highly homologous but have nonoverlapping T-cell receptor V_β specificities. To determine the importance of these few differences for specific V_β interaction, we studied superantigen responses in mice to chimeric and mutant MMTV(SIM) and mtv-8 superantigens expressed by recombinant vaccinia viruses. We show that only a few changes (two to six residues) within the C terminus are necessary to modify superantigen recognition by specific V_βs. Thus, the introduction of the MMTV(SIM) residues 314-315 into the mtv-8 superantigen greatly decreased its V_β12 reactivity without gain of MMTV(SIM)-specific function. The introduction of MMTV(SIM)-specific residues 289 to 295, however, induced a recognition pattern that was a mixture of MMTV(SIM)- and mtv-8-specific V_β reactivities: both weak MMTV(SIM)-specific V_β4 and full mtv-8-specific V_β11 recognition were observed while V_β12 interaction was lost. The combination of the two MMTV(SIM)-specific regions in the mtv-8 superantigen established normal MMTV(SIM)-specific V_β4 reactivity and completely abolished mtv-8-specific V_β5, -11, and -12 interactions. These new functional superantigens with mixed V_β recognition patterns allowed us to precisely delineate sites relevant for molecular interactions between the SIM or mtv-8 superantigen and the T-cell receptor V_β domain within the 30 C-terminal residues of the viral superantigen.

Superantigens (Sags) constitute a group of proteins with potent effects on the immune system. Although different Sags are expressed by a wide variety of microorganisms, they share the ability to stimulate a large number of T cells through similar mechanisms. Sags are presented in the context of major histocompatibility complex (MHC) class II molecules at the cell surface and interact with subsets of T cells expressing specific variable domains in the T-cell receptor (TCR) β chain (12, 23, 31, 43). The encounter with Sag leads first to the expansion and subsequently to the deletion of reactive mature T cells (30, 42, 43). When immature T cells interact with Sag during thymic development, they undergo intrathymic deletion (22, 23, 31).

Mouse mammary tumor virus (MMTV) is a retrovirus which exists either as an infectious viral particle transmitted from mother to offspring via milk (exogenous MMTV) or as a germ line-integrated provirus stably transmitted via genetic inheritance (endogenous MMTV) (24). In addition to the classical retrovirus genes gag, pol, and env, the MMTV provirus genome contains an open reading frame within the 3′ long terminal repeat which encodes the viral Sag (2, 10). The MMTV Sag plays a crucial role in the viral life cycle. MMTV preferentially binds to B cells (5), triggers their activation (3), and infects them. The Sag is then expressed at the surface of the cells in association with MHC class II molecules. Sag-reactive T cells are activated and accumulate locally, providing help to infected B cells through cognate T-cell–B-cell interactions, which lead to a large increase in the number of infected B cells (17, 27). This Sag-mediated increase in viral load has been shown elsewhere to be an essential step for the subsequent transport of the virus to the mammary gland and for its efficient transmission to the next generation of mice (16, 18).

The presentation of the Sag-MHC class II complex and its recognition by the TCR are very different from the recognition of classical antigens. The interaction of a T cell with Sags almost exclusively involves the V_β domain of the TCR. The other parts of the TCR do not contribute to the interaction, except for the V_α domain, which has a minor indirect influence (37, 40, 41). In particular, Sags are thought to interact specifically with the hypervariable region 4 (HV4) of the V_β domain as well as with elements of the complementarity-determining regions 1 and 2 (CDR1 and CDR2) (4, 6, 9, 11, 15, 19, 20, 35, 36). The HV4 is located on the lateral part of the molecule, far away from the central region implicated in the recognition of peptide-MHC complexes (14). In the case of bacterial Sags, the published crystal structure of a Sag-TCR complex shows a direct interaction between the Sag and the HV4, as well as with several amino acids of CDR1 and CDR2, of the TCR (13). This report, together with other studies (4, 8, 9, 15, 19, 33, 35, 36), indicates that residues of the HV4 as well as some residues outside this region are critical for the specific interaction of the TCR with Sag.

Comparisons of Sag proteins from several MMTV strains have shown a high degree of sequence identity between them (29, 46). Most sequence differences are located within two regions defined by amino acids (aa) 174 to 198 (region I) and aa 288 to the C terminus (region II). The C-terminal variability has been correlated with observations that certain Sag molecules interact with particular TCRs (46). For instance, the C3H and GR exogenous Sags reacted with T cells bearing V_β14 chains (10). Moreover, experiments performed by Yazdanbakhsh and colleagues showed that C-terminal replacement of endogenous mtv-1 Sag (V_β3 reactive) with mtv-7 Sag (V_β6 reactive) allowed the recombinant Sag to react with V_β6-expressing T cells in stable-transfection assays (46). The reciprocal experiment confirmed that a polymorphic Sag region (30 to 40 aa) at the C terminus is sufficient to specify interactions with certain TCR V_β chains (46). However, little is known about the precise requirements within this region for Sag function. Recently, a series of substitutions and deletions transferred into a cloned infectious MMTV provirus has been used for in vivo analysis (45). These mutant viruses induced tumors with lower incidence in mice, although all but one C-terminal amino acid substitution abolished Sag function. Interestingly, the one mutation affecting the C-terminal 3 aa that retained partial Sag function lost the ability to be transmitted through milk to susceptible offspring (45).

The aim of this work was to characterize the amino acids that determine the V_β specificity of the MMTV(SIM) Sag (32) and, indirectly, the mtv-8 Sag. Among the 39 sequenced viral Sags, MMTV(SIM) is the only one showing reactivity with V_β4 and V_β10^a/c TCRs (31, 32; reviewed in reference 29). Its sequence has greatest similarity with Sags interacting with V_β5, -11, and -12, such as mtv-8. MMTV(SIM) and mtv-8 Sags differ at only six positions and/or regions within the C-terminal 70 aa (Fig. 1). Four of them (positions 266, 273, 305, and 319-320; all amino acid positions refer to the mtv-8 sequence) can be found in Sags with V_β specificities other than that of mtv-8 and MMTV(SIM), whereas two (289 to 295 and 314-315) are unique to MMTV(SIM). We have addressed the importance of these two regions for SIM-specific reactivity to V_β4 and V_β10^a/c in vivo by using the recombinant vaccinia virus (RV) expression system. We have previously shown that RV can be used not only to assess the expression and posttranslational modifications of the MMTV Sag in cell cultures but also to monitor the specific Sag response in vivo (25, 26). Subsequently, we generated RVs expressing the complete Sag molecule of mtv-8 or MMTV(SIM), as well as a panel of chimeric and mutant mtv-8 Sag molecules. With this strategy, we aimed to simultaneously monitor the response in vivo to the mutant Sag in terms of gain of SIM-specific reactivity (V_β4 and -10^a/c) and of loss of mtv-8-specific reactivity (V_β5, -11, and -12). All mutants had functional Sag activity in vivo. We confirm that the C-terminal part of the viral Sag is necessary and sufficient to confer specific TCR V_β reactivity. In addition, we show that the exchange of 7 aa (VWGKIFH [289 to 295]) of the mtv-8 Sag with the SIM-specific 4 aa and the small deletion that they encompass (F*R---Y, designated “Δ”) established a partial SIM-specific V_β4 reactivity. Interestingly, only some of the original mtv-8-specific TCR V_β interactions were lost in parallel, thereby generating a Sag molecule with a mixed mtv-8 and MMTV(SIM) V_β reactivity pattern, i.e., V_β11 and V_β4. The full SIM-specific V_β4 reactivity was reached by the introduction in the mtv-8 Sag of two additional point mutations (NS [314-315, designated “M”]) together with the previously mentioned deletion and was paralleled by the total loss of mtv-8-specific TCR interactions. However, none of the examined mutations in the mtv-8 Sag restored the V_β10^a/c reactivity observed with the MMTV(SIM) Sag. In conclusion, we were able to identify residues important for the complex interactions between a viral (MMTV) Sag molecule and TCR V_β elements on T cells.

FIG. 1.

Comparison of amino acid sequences of mtv-8 and MMTV(SIM) Sag molecules. Asterisks indicate amino acid identity with the mtv-8 sequence. Dashes indicate gaps introduced to maximize amino acid identity. Open circles indicate the portion of the MMTV(SIM) Sag molecule used to produce the chimeric mtv-8/MMTV(SIM) Sag molecule.

MATERIALS AND METHODS

Mouse strains.

BALB/c mice (H-2^d TCR V_β^b ) were purchased from Harlan OLAC Ltd. (Bicester, United Kingdom). BALB/c mice bearing the TCR V_β^a haplotype were kindly provided by Alexandra Livingstone (Basel Institute of Immunology, Basel, Switzerland). The mtv-free mice have been previously described (8). BALB/c TCR V_β^a and mtv-free mouse strains were kept as breeding pairs in our animal facilities.

Antibodies.

The following antibodies were used in this study: fluorescein isothiocyanate (FITC)-labeled anti-V_β4 (KT4-10 [39]), FITC-labeled anti-V_β5 (MR9-4; Pharmingen, San Diego, Calif.), FITC-labeled anti-V_β11 (RR3-15; Pharmingen), FITC-labeled anti-V_β12 (MR11-1; Pharmingen), Cychrome-coupled anti-CD4 (RM4-5; Pharmingen), and phycoerythrin-coupled anti-CD69 (H1.2F3; Pharmingen). Anti-V_β10^a/c (KT10a [38]), kindly provided by K. Tomonari (Fukui Medical School, Fukui, Japan), was used as a hybridoma supernatant and was detected by an FITC-coupled goat anti-rat immunoglobulin G (IgG). A polyclonal rabbit anti-open reading frame-peptide serum (serum C [7]) raised against a 23-aa synthetic peptide of the MMTV(GR) Sag was used for immunoprecipitation of the viral Sags expressed by RVs.

Cloning and mutagenesis.

The mtv-8 and the MMTV(SIM) Sag coding sequences were amplified by PCR from genomic DNA of a mouse harboring mtv-8 as a single endogenous mtv (8) and from a pUC19 backbone containing the MMTV(SIM) Sag coding sequence (32), respectively, using the following primers: 5′-TT GGA ATT CCA CCA TGC CGC GCC TGC AG-3′ and 5′-CCA CGC GTT GGG AAC CGC AAG GTT GG-3′. Both PCR products were cloned into pCI-neo expression vector (Promega, Madison, Wis.) after EcoRI and AflIII digestion to generate pCIMtv-8 and pCISIM. The chimeric mtv-8–SIM Sag construct was generated by switching in the PpuMI/AflIII fragment from pCISIM into pCIMtv-8 to generate pCIMtv-8/S. These three Sag coding sequences upon digestion with EcoRI, blunted with Klenow enzyme and digested with XbaI, were shuttled into the vaccinia virus transfer vector pARO1 (25) digested with HindIII, blunted with Klenow enzyme, and digested with XbaI. The resulting plasmids, i.e., pAMtv-8, pASIM, and pAMtv-8/S, were then sequenced. To generate pAMtv-8M and pAMtv8-ΔM, the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, Calif.) was used with pAMtv-8 and pA8-ΔM, respectively, as templates, following the manufacturer's instructions. The following primers were used: 5′-GAA CAC ATT TCA GCT GAT ACT AAT AGC ATG AGC TAT AAT GG-3′ and 5′-CCA TTA TAG CTC ATG CTA TTA GTA TCA GTC GAA ATG TGT TC-3′. pAMtv-8Δ was generated using the ExSite PCR-based site-directed mutagenesis kit (Stratagene) following the manufacturer's instructions and using the following primers: 5′-TCT CCA AAC GTT CAT TCC TGT TCC-3′ and 5′-CCA TTA TAG CTC ATG CTA TTA GTA TCA GCT GAA ATG TGT TC-3′.

Preparation of RVs.

The procedure to generate RVs has been described previously (25). To generate virus stocks, HeLa cells were infected with RV for 48 h at a multiplicity of infection of 0.1. The cells were recovered and Dounce homogenized in cold 10 mM Tris-HCl, pH 9, and cell fragments were eliminated by centrifugation for 10 min at 4°C and 750 × g The supernatant (referred to as crude stock) was incubated for 30 min at 37°C with trypsin (0.25 mg/ml). The crude stock was then laid over a 36% sucrose cushion in 10 mM Tris-HCl, pH 9, and centrifuged for 80 min at 4°C and 25,000 × g The purified virus in the pellet was resuspended in a minimal volume of 10 mM Tris-HCl, pH 9, and titrated in a plaque formation assay on huTk⁻143B cells (ATCC CRL-8303).

Injection and sampling.

Purified virus (5 × 10⁷ PFU) in a 30-μl volume was injected subcutaneously into the hind footpads of naive mice. Twenty-four hours postinjection, the mice were sacrificed; the draining popliteal lymph nodes were isolated and homogenized to a single-cell suspension. For each experiment, we used the two hind feet of two to three mice per virus tested. All experiments involving mice were repeated three times.

Flow cytometric analysis.

Popliteal lymph node cells were triple stained in a single step with a mixture of a given anti-V_β antibody (V_β4, -5, -11, or -12), anti-CD4 antibody, and anti-CD69 antibody. V_β10^a-positive cells were labeled by incubation with KT10a antibody (hybridoma supernatant) and subsequently with FITC-coupled goat anti-rat IgG. After blocking with rat IgG (10 μg/ml), the cells were incubated with anti-CD4-Cychrome and anti-CD69-phycoerythrin. Analysis was performed on a FACScan (Becton Dickinson & Co., Mountain View, Calif.) cell analyzer using the Lysis II software for data evaluation. Dead cells were excluded on the basis of their forward and side scatter characteristics. From 30,000 to 50,000 cells were acquired, and results were analyzed with a Student t test assuming unequal variance.

Metabolic labeling and immunoprecipitation.

Confluent CV-1 cells in six-well plates were infected at a multiplicity of infection of 10 for 12 h. The cells were washed with phosphate-buffered saline and starved in 1 ml of methionine- and cysteine-free medium for 1 h at 37°C. The medium was then replaced by 0.3 ml of methionine- and cysteine-free medium containing 100 μCi of ³⁵S-Easy Tag express labeling mix (NEN Life Science Products, Boston, Mass.). After 1 h of incubation at 37°C, cells were washed with cold phosphate-buffered saline and lysed for 30 min on ice with 0.5 ml of RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% Na-deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 50 mM Tris-HCl, pH 8) containing protease inhibitors (0.2 mM phenylmethylsulfonyl fluoride, 10 μg of aprotinin/ml, 10 μg of leupeptin/ml). Clarified lysates (350 μl) were precleared with 35 μl of preimmune rabbit serum and 50 μl of protein G-agarose beads (Amersham Pharmacia Biotech, Uppsala, Sweden) for 6 h at 4°C. The resulting lysates were further incubated overnight at 4°C with 10 μl of anti-MMTV(GR) Sag serum C and 50 μl of protein G-agarose beads. After two washes in cold RIPA buffer, the protein G beads were resuspended in loading buffer (3% β-mercaptoethanol, 3% SDS, 0.3% bromophenol blue, 10% glycerol) and heated for 5 min at 95°C, and the samples were resolved on an SDS–8 to 12% polyacrylamide gel. The gel was fixed, soaked in Amplify (Amersham Pharmacia Biotech) for 30 min, dried, and exposed at −80°C with amplifying screens.

Statistical treatment of the data.

All data were analyzed for statistical significance with a t test (two-sample test assuming unequal variance). For stimulation data, we have compared the values obtained after injection of an RV with the data obtained from mock-infected animals. For CD69 upregulation, the data obtained for CD4⁺ T cells in a given V_β subset were compared to data for all other V_β CD4⁺ subsets.

RESULTS

Construction of RVs expressing a wild-type or a mutated MMTV Sag molecule.

Despite 90.9% sequence identity and 92% sequence homology (Fig. 1), mtv-8 and MMTV(SIM) Sags interact with different TCR V_β domains. The mtv-8 Sag is recognized by the V_β5, -11, and -12 regions of TCRs (29), whereas the MMTV(SIM) Sag has been shown previously to interact with the V_β4 and V_β10^a/c regions of TCRs (32, 33). Figure 1 shows the amino acid sequence alignment of the mtv-8 and MMTV(SIM) Sags, which differ in only 29 aa. Because previous results have shown that the C-terminal portion of a Sag dictates its reactivity with given V_β regions of TCRs (46) and because most of the sequence variability of viral Sags resides within these C-terminal 30 to 40 aa, we decided to examine the importance of this C-terminal domain for V_β reactivity. When all known Sag sequences were compared, only two regions within the C-terminal 70 aa were unique to the MMTV(SIM) and its unique V_β specificity. One region consists of three amino acid changes and a small deletion of 3 aa present in the MMTV(SIM) Sag molecule close to the C terminus (Fig. 1, “Δ”).The other region is characterized by the double mutation F to N and G to S compared to mtv-8 (Fig. 1, “M”). These few changes gave us an opportunity to map the amino acids responsible for the different V_β specificities of MMTV(SIM) and mtv-8 Sags. To do this, we constructed six RVs producing wild-type or mutant mtv-8/SIM Sags (Fig. 2A): RV-8 and RV-S produced the wild-type Sags of mtv-8 and MMTV(SIM), respectively, while RV-8/S expressed a chimeric Sag with the N-terminal portion of the mtv-8 Sag and the 106 C-terminal aa of the MMTV(SIM) Sag. We also produced three RVs to characterize the amino acids determining the V_β reactivity patterns of the MMTV(SIM) and mtv-8 Sag molecules. In all three cases, the mtv-8 Sag was used as a backbone to insert SIM-specific modifications. This strategy allowed us to diagnose the gain of the reactivity with SIM-specific V_β4 and V_β10^a/c TCRs and the loss of the mtv-8-specific interaction with TCRs V_β5, -11, and -12 of the newly generated mutant Sag. We inserted the deletion and the two flanking mutations present in the MMTV(SIM) Sag molecule (Fig. 1, aa 289 to 295 in mtv-8, “Δ”) to produce RV-8Δ. We also introduced the double mutation F to N and G to S (Fig. 1, aa 314-315, “M”) to produce RV-8M. We further combined both the deletion Δ and the mutation M to produce RV-8ΔM. All the other mutations (T to S [aa 266], I to M [aa 273], Q to L [aa 305], N to Y [aa 319], and G to D [aa 320]) were ignored because they are present in other Sags displaying different V_β specificities (reviewed in reference 29)

FIG. 2.

Scheme of the different Sag molecules constructed and their expression in mammalian cells infected with RVs. (A) The different Sag molecules constructed and used to produce RVs are shown. The amino acids that have been introduced in the mtv-8 backbone have been indicated: F*R---Y (aa 289 to 295, referred to as “Δ”) and NS (aa 314-315, referred to as “M”). The numbering refers to the mtv-8 Sag molecule (Fig. 1). (B) CV-1 cells were infected with Sag-expressing RVs and labeled with [³⁵S]Met-Cys at 12 h postinfection as described in Materials and Methods. The upper panel shows the immunoprecipitated Sag molecules including an uninfected sample (Mock), a wild-type-infected sample (WT), and a positive control (rm1 [25]). The lower panel shows the same samples before immunoprecipitation. The numbers at left are molecular masses in kilodaltons.

Expression of the different MMTV Sag molecules in mammalian cells infected with RVs.

Expression of the viral Sag molecules by the RVs was monitored in cultured CV-1 cells. Figure 2B shows the results of a representative immunoprecipitation of the different Sag molecules. We used noninfected cells (Mock) and wild-type vaccinia virus (WT)-infected cells as negative controls. Furthermore, we used the previously described RV-rm1 (25, 26), which expresses the MMTV(GR) Sag, as a positive control. Two specific bands were seen for each RV (upper panel), one with a size of 46 kDa as previously described (25) and one slightly smaller. The lower-molecular-weight band might reflect either an internal initiation or differential glycosylation. Overall, all RVs expressed a Sag molecule of the appropriate size at similar levels (maximally threefold difference as measured by densitometry) in tissue culture. The lysates before immunoprecipitation are shown in the lower panel.

Response of BALB/c mice to RVs expressing wild-type or mutant MMTV Sags: the Δ mutant induces weak reactivity, and the combination of Δ and M induces full reactivity, with TCR V_β4.

BALB/c mice were tested for their response to the six RVs expressing wild-type or mutant mtv-8/SIM Sags. Because BALB/c mice harbor, among others, an integrated mtv-8 virus deleting V_β5, -11, and -12 T cells (reviewed in reference 29) and because their TCR V_β domains are of the b haplotype, only SIM-reactive V_β4 CD4⁺ T cells can be analyzed. Figure 3A shows the percentage of large CD4⁺ lymphoblasts expressing the V_β4 TCR in the lymph node draining the site of injection. The results indicate that RV-8/S (P = 1 × 10⁻⁵), RV-8ΔM (P = 6 × 10⁻⁵), and to a lesser extent RV-8Δ (P = 2.6 × 10⁻⁴) or RV-8M (P = 3.4 × 10⁻²) induced the expansion of the V_β4-bearing CD4⁺ lymphoblasts like the control virus RV-S (P = 1.8 × 10⁻³). The control virus, RV-8, did not stimulate V_β4 CD4⁺ T cells as expected. The expansion by RV-8M observed in this experiment was not seen in two other experiments. Furthermore, no expansion in total V_β4⁺ CD4⁺ T cells was observed to be statistically significant in all experiments done. In conclusion, RV-8M only very weakly and inconsistently caused some expansion of V_β4⁺ CD4⁺ blasts.

FIG. 3.

Activation and expansion of V_β4⁺ CD4⁺ T cells upon injection of Sag-expressing RVs in BALB/c mice. BALB/c mice (TCR V_β^b) were not injected (Mock) or were injected with 5 × 10⁷ PFU of RV-8, RV-S, RV-8/S, RV-8M, RV-8Δ, or RV-8ΔM. (A) The response of V_β4⁺ cells to each RV tested is indicated as a percentage of large CD4⁺ blasts. ∗, P < 0.05; ∗∗, P > 0.05. (B) The percentage of CD69⁺ CD4⁺ T cells among V_β4-negative (Non-V_β4) or V_β4-positive (V_β4) cells is shown for the same experiment. ∗, P < 0.01. Data represent mean values ± standard errors of the means of four draining lymph nodes 24 h postinfection. The experiment is representative of three performed.

To corroborate our results, we measured the upregulation of the activation marker CD69, which can be used to determine early Sag-mediated cell activation that leads subsequently to the expansion of a given V_β population (3). Figure 3B displays the expression of the CD69 activation marker on non-V_β4 and V_β4 CD4⁺ T cells upon injection of the RVs. Very clearly, no significant difference in CD69 expression between non-V_β4 and V_β4 CD4⁺ T cells was measured when no RV (Mock, P = 0.2), RV-8 (P = 0.08), or, importantly, RV-8M (P = 0.232) was injected, although the CD69 expression level increased in both populations for RV-8 and RV-8M. This latter effect is most likely due to a non-Sag-related activation caused by the vaccinia virus itself, since RV-8 cannot trigger mtv-8 Sag-reactive CD4⁺ T cells due to their absence in BALB/c mice. In contrast, preferential activation among the V_β4-bearing CD4⁺ T cells was measured upon injection of RV-S (P = 2.5 × 10⁻⁴), RV-8/S (P = 1.1 × 10⁻⁶), RV-8Δ (P = 1.8 × 10⁻³), and RV-8ΔM (P = 1.1 × 10⁻⁵) in agreement with the expansion of V_β4 CD4⁺ lymphoblasts presented above. This upregulation of CD69 was not seen in non-SIM-relevant V_β6 CD4⁺ T cells (data not shown). Together, these data confirm that the C-terminal half of the viral Sag determines the V_β specificity of the TCR in our system. They further show that the small deletion with the surrounding mutations Δ (Fig. 1, aa 289 to 295) is sufficient to confer V_β4 reactivity on the mtv-8 Sag and that the double mutation M (aa 314-315) increases this effect.

Response of BALB/c TCR V_β^a mice to RVs expressing wild-type or mutant MMTV Sags: the Δ and/or M modifications do not confer V_β10^a specificity on the mtv-8 Sag.

The MMTV(SIM) Sag was reported previously to interact not only with V_β4-expressing T cells in mice with a TCR V_β^b haplotype (32) but also with V_β10-expressing T cells in BALB/c mice with the TCR V_β^a or the TCR V_β^c haplotype (33). We therefore injected our RVs into BALB/c mice with the TCR V_β^a haplotype in order to detect a Sag response in V_β4 and V_β10^a CD4⁺ T cells. As shown in Fig. 4, RV-S and RV-8/S triggered the expansion of both V_β4 and V_β10^a CD4⁺ T cells (P = 1.4 × 10⁻², P = 5.9 × 10⁻⁴, P = 1 × 10⁻³, and P = 1.4 × 10⁻⁴, respectively). RV-8Δ and RV-8ΔM did not stimulate V_β10^a CD4⁺ T cells (Fig. 4B), although they were clearly capable of inducing the expansion of V_β4 CD4⁺ T cells (P = 3.5 × 10⁻² and P = 3.7 × 10⁻², respectively) (Fig. 4A). RV-8 and RV-8M were unable to change the levels of V_β4 and V_β10^a CD4⁺ T cells. The expression of the activation marker CD69 (data not shown) correlated with these results, demonstrating that neither Δ, M, nor the combination of Δ and M residues mediates the interaction of the SIM Sag with the V_β10^a TCR and that other amino acids must be important for this V_β10^a TCR reactivity.

FIG. 4.

Expansion of SIM-reactive V_β4⁺ and V_β10⁺ CD4⁺ T cells upon injection of the different RVs into BALB/c (TCR V_β^a) mice. BALB/c mice bearing the TCR V_β^a haplotype were not injected (Mock) or were injected with 5 × 10⁷ PFU of RV-8, RV-S, RV-8/S, RV-8M, RV-8Δ, or RV-8ΔM. The response of V_β4⁺ (A) or V_β10⁺ (B) cells is indicated as a percentage of enlarged CD4⁺ T lymphoblasts. Data represent mean values ± standard errors of the means of four draining lymph nodes 24 h postinfection. The experiment is representative of three performed. ∗, P < 0.05.

Response of mtv-free mice to RVs expressing wild-type or mutant MMTV Sags: reactivity with V_β5, -11, and -12 TCRs.

The experiments presented so far have addressed the gain of SIM-specific V_β4 and V_β10^a reactivity by the mtv-8/MMTV(SIM) Sags expressed by RVs. We further determined whether the gain of function with SIM-specific TCRs was paralleled by the loss of mtv-8-specific reactivity with V_β5, -11, and -12. Since the response to RV-8 cannot be measured in BALB/c mice because of the presence of endogenous mtv genes, we tested all the RVs in MMTV-free mice (8). These mice have no integrated mtv genome and therefore have an intact V_β repertoire, making it possible to test for any V_β reactivity for a given Sag molecule. mtv-free mice were injected with the six RVs, and the draining lymph nodes were analyzed for the preferential activation (data not shown) and the expansion of a given V_β population among the CD4⁺ T cells (Fig. 5). When the V_β5 TCR was analyzed, only RV-8 (P = 4.8 × 10⁻²) and RV-8M (P = 3.0 × 10⁻²) were found to stimulate a low but significant expansion of this CD4⁺ T-cell population (Fig. 5A). The analysis of V_β11 CD4⁺ lymphoblasts indicated that RV-8 (P = 3.7 × 10⁻²), RV-8M (P = 2.2 × 10⁻³), and, interestingly, also RV-8Δ (P = 1.1 × 10⁻⁴) induced the expansion of this lymphoblast population (Fig. 5B). The stimulation of V_β11 CD4⁺ T cells by RV-8Δ was highly significant and was observed in other experiments. When we looked at the V_β12 CD4⁺ T-cell population, we observed that RV-8 (P = 4.4 × 10⁻⁹), RV-8M (P = 3.3 × 10⁻³), and, very weakly, RV-8Δ (P = 3.1 × 10⁻²) triggered their expansion (Fig. 5C). In all other experiments, the stimulation of V_β12 CD4⁺ T cells by RV-8Δ was undetectable (data not shown). Interestingly, the expansion induced by RV-8M and RV-8Δ was very low compared to that of the control virus RV-8. This decreased stimulation cannot be due to lower Sag expression by RV-8M and RV-8Δ since these viruses stimulated V_β11 CD4⁺ T cells as well as RV-8 (Fig. 5B). The analysis of V_β4-bearing T cells (data not shown) revealed that RV-S, RV-8/S, RV-8ΔM, and, very weakly, RV-8Δ triggered V_β4⁺ T cells, in agreement with the results obtained with BALB/c mice (Fig. 3 and 4). The expression of the activation marker CD69 measured in parallel confirmed the obtained V_β reactivity pattern of the tested RVs (data not shown).

FIG. 5.

V_β reactivity pattern in mtv-free mice observed after injection of Sag-expressing RVs. mtv-free mice were not injected (Mock) or were injected with 5 × 10⁷ PFU of RV-8, RV-S, RV-8/S, RV-8M, RV-8Δ, or RV-8ΔM. The response of V_β5-, -11-, and -12-positive cells is indicated as a percentage of large CD4⁺ blasts in panels A, B, and C, respectively. V_β4⁺ CD4⁺ T cells from the same lymph nodes were analyzed in parallel (data not shown). Data represent mean values ± standard errors of the means of four draining lymph nodes 24 h postinfection. The experiment is representative of three performed. ∗, P < 0.05.

Our results can be summarized as follows (Table 1): RV-S and RV-8 stimulate T cells with the same specificity as their corresponding mtv-8 or MMTV(SIM) viruses, i.e., V_β4 and -10^a and V_β5, -11, and -12, respectively. RV-8/S behaves like RV-S in accordance with the idea that the C-terminal portion of the MMTV Sag determines the V_β specificity (46). RV-8M interacts with TCR V_β5 and -11 and weakly with V_β12. RV-8ΔM has gained V_β4 reactivity, but not V_β10^a reactivity, in parallel with the complete loss of mtv-8-specific reactivity with V_β5, -11, and -12. RV-8Δ displays a dual mtv-8/SIM V_β reactivity pattern characterized by the interaction with V_β4, V_β11, and maybe V_β12.

TABLE 1.

Sag interaction with T-cell populations

Sag	Result for TCR Vβa:
	4	10a	5	11	12
mtv-8	−	−	+	++	++
mtv-8M	−	−	+	++	+
mtv-8Δ	+	−	−	++	+/−
mtv-8ΔM	++	−	−	−	−
mtv-8/S	++	++	−	−	−
SIM	++	++	−	−	−

^a−, no interaction detected; +/−, +, and ++, weak to strong stimulation and expansion of T-cell population, respectively. The results are based on all experimental results shown and on data not shown. 

DISCUSSION

MMTV/mtv Sags can be grouped into seven families based on the alignment of their 70 C-terminal aa (reviewed in reference 29). Within this C-terminal region, the highest sequence variability is found in the polymorphic region II containing the last 30 to 40 aa. Several studies have indeed implicated this part of the Sag molecule in the interaction with the V_β domain of the TCR (1, 44, 46).

Most of the knowledge about the interaction of the viral Sag with the TCR comes from studies analyzing TCR residues involved in Sag recognition. In these studies, the TCR polymorphism in wild-type mice and in rats, responsible for gain or loss of recognition of particular MMTV Sags, was used to map the key amino acids in TCR-Sag interactions (4, 8, 9, 15, 19, 33, 36). The few residues ascribed were mainly located in the HV4, but also in the CDR2 of the TCR. Although the same regions are also implicated in the recognition of bacterial Sags, the TCR binding sites for viral and bacterial Sags were shown elsewhere to be different (28).

Little is known about the precise amino acid requirements in the MMTV Sag for TCR interaction besides the already-mentioned importance of the 30 to 40 aa of the C terminus. This is undoubtedly due to the low abundance of the Sag and its lack of function in purified form. In this report, we sought to study the residues of the viral Sag that determine its interaction with the TCR by taking advantage of the closely related Sags of mtv-8 and MMTV(SIM). These Sags show 91% sequence identity in their C-terminal 70 aa but are recognized by different TCR V_β domains. We focused on two SIM-specific regions of the Sag because they are unique to MMTV(SIM) and its unique TCR V_β recognition pattern. We thus introduced into the mtv-8 backbone the MMTV(SIM)-specific mutations that we named M and Δ to be able to simultaneously monitor the gain of SIM-specific V_β TCR interaction and loss of mtv-8-specific V_β TCR interaction. RVs were used for the expression of wild-type or mutant mtv-8/SIM Sags in vivo, as we have previously shown that this system can be used to dissect the in vivo response to the retroviral MMTV(GR) Sag in the context of an unrelated viral infection (26).

By exchanging residues from two naturally occurring Sags, we obtained a panel of mutant proteins expressed by RVs which all had intact Sag function as monitored by the activation and expansion of Sag-reactive CD4⁺ T cells in vivo. This contrasts with the results of two recent studies addressing the impact of mutations and deletions-insertions on the function of Sag molecules (34, 45). In both studies, mutations not occurring in heterologous Sags were introduced in the polymorphic region II and led to the loss of Sag function, due in some cases to the abolition of transport to the cell surface.

In the present study, we provide the first evidence of a switch of V_β reactivity for a viral Sag when expressed in vivo. Thus, we confirm that the C-terminal portion of a given Sag is sufficient to confer its TCR V_β specificity on another Sag molecule when it is recognized in the mouse. Indeed, the exchange of the 106 C-terminal aa of the mtv-8 Sag with the 106 C-terminal aa of the MMTV(SIM) Sag results in a complete transfer of TCR V_β specificities, i.e., a switch from a V_β5, V_β11, and V_β12 to a V_β4 and V_β10^a reactivity pattern. This is in agreement with a previous report showing that the exchange of the polymorphic region II between mtv-1 Sag and mtv-7 Sag led to the exchange of the V_β specificity of the responding T cells in stable-transfection assays (46).

Here, we further report that the two small regions M and Δ of the MMTV(SIM) Sag, when transferred into the mtv-8 Sag, are sufficient to modify the V_β specificity of the parental Sag molecule. Importantly, the two regions of 2 and 6 aa, respectively, differentially affected the recognition by specific TCR V_β domains. The SIM-specific M mutation greatly decreased the V_β12 reactivity of the parental mtv-8 Sag, thereby indicating that amino acids F₃₁₄ and G₃₁₅ of the mtv-8 Sag play a role in the interaction with this TCR V_β domain. These mutations were, however, not sufficient for the induction of a SIM-specific V_β recognition pattern.

Interestingly, we also show that the Sag molecules containing the Δ region of MMTV(SIM) displayed a mixed phenotype with an acquired SIM-specific V_β4 reactivity while retaining an mtv-8-specific interaction with V_β11 and a strong reduction of mtv-8-specific V_β12 recognition. Importantly, the V_β12 reactivity was almost totally abolished despite the presence of the mtv-8 residues F₃₁₄ and G₃₁₅ shown above to be involved in its recognition. Structural changes induced by the Δ mutations may explain this apparent contradiction.

The combination of M and Δ conferred full SIM-specific V_β4 reactivity on the mtv-8 Sag molecule and completely abrogated its mtv-8-specific reactivity with V_β5, V_β11, and V_β12 CD4⁺ T cells. The two regions are, however, not sufficient to induce detectable SIM-specific V_β10^a reactivity in the mtv-8 Sag. Since the response to the MMTV(SIM) Sag of V_β10^a/c is normally higher than that of V_β4 (33), residues other than M and Δ must contribute to the interaction with V_β10^a/c. Only 5 aa are different between RV-8/S and RV-8ΔM. Since all these amino acid changes can be found in Sags with V_β specificities other than that of MMTV(SIM) and mtv-8, a combination of one or several of them with Δ and/or M may be required for the recognition by V_β10^a.

Altogether, our results show that only a few Sag residues are involved in the interactions with specific TCR V_β. A similar situation is found with bacterial Sags, where two to three residues can change the TCR V_β specificity (21). The lack of a three-dimensional model for the MMTV Sag molecule precludes the possibility of ascribing precise positions to the interaction. We can therefore not distinguish whether M and Δ residues are in direct contact with the TCR V_β chain or whether they impose structural constraints on the Sag molecule, thereby modifying the exposure of the contact points.

Our work presents a model system that makes it possible to screen in vivo for amino acids of Sags involved in the interaction with a given TCR V_β domain. It conserves one of the principal Sag functions encountered in MMTV infection, namely, the activation and expansion of the TCR V_β-reactive CD4⁺ T cells. Another Sag function, the deletion of the amplified T cells, cannot be addressed with this model system because of the lack of a sustained Sag expression due to the rapid elimination of the RVs in the mouse.

In conclusion, we were able to precisely delineate eight residues relevant for molecular interactions between the SIM, or mtv-8, Sag and the TCR V_β domain within the 35 C-terminal aa of the viral Sag.