The cytotoxic T-lymphocyte response to wild-type simian virus 40 large tumor antigen (Tag) in C57BL/6 (H2 b ) mice is directed against three H2-D b -restricted epitopes, I, II/III, and V, and one H2-K b -restricted epitope, IV. Epitopes I, II/III, and IV are immunodominant, while epitope V is immunorecessive. We investigated whether this hierarchical response was established in vivo or was due to differential expansion in vitro by using direct enumeration of CD8 ؉ T lymphocytes with Tag epitope/major histocompatibility complex class I tetramers and intracellular gamma interferon staining. The results demonstrate that epitope IV-specific CD8 ؉ T cells dominated the Tag-specific response in vivo following immunization with full-length Tag while CD8 ؉ T cells specific for epitopes I and II/III were detected at less than one-third of this level. The immunorecessive nature of epitope V was apparent in vivo, since epitope V-specific CD8 ؉ T cells were undetectable following immunization with full-length Tag. In contrast, high levels of epitope V-specific CD8 ؉ T lymphocytes were recruited in vivo following immunization and boosting with a Tag variant in which epitopes I, II/III, and IV had been inactivated. In addition, analysis of the T-cell receptor ␤ (TCR␤) repertoire of Tag epitope-specific CD8 ؉ cells revealed that multiple TCR␤ variable regions were utilized for each epitope except Tag epitope II/III, which was limited to TCR␤10 usage. These results indicate that the hierarchy of Tag epitope-specific CD8 ؉ T-cell responses is established in vivo.Immunity to the large tumor antigen (Tag) of simian virus 40 (SV40) in C57BL/6 mice is characterized by the development of CD8 (23,34,50,52). An immunological hierarchy has been demonstrated among these four epitopes within Tag. Immunization of C57BL/6 mice with SV40, SV40 Tagtransformed cells, or a recombinant vaccinia virus (rVV) which encodes the full-length Tag leads to the induction of cytotoxic T lymphocytes (CTL) specific for epitopes I, II/III, and IV (26, 41, 51). Frequency estimates from limiting-dilution analysis of splenic lymphocytes obtained 9 days after immunization with SV40 Tag-transformed cells revealed that epitope IV-specific CTL represent 1 in 14,000 splenocytes while epitope I and II/III-specific CTL were less abundant (1 in 67,000) and epitope V-specific CTL were undetectable (41).Although epitope V-specific CTL are not detected following immunization with full-length SV40 Tag, immunization with syngeneic cells carrying inactivating mutations or deletions in Tag epitopes I, II/III, and IV leads to the induction of epitope V-specific CTL (41, 50). Accordingly, epitope V has been characterized as immunorecessive. Additional strategies which enhance the immunogenicity of epitope V include immunization with rVVs which express epitope V as a minigene linked to a secretory signal sequence (ES) or in which the epitope V sequence is inserted into a nonimmunogenic murine self protein, dihydrofolate reductase (26). Precise mechanisms which control the immun...

CD8(+) T lymphocytes (T(CD8)) responding to subdominant epitopes provide alternate targets for the immunotherapy of cancer, particularly when self-tolerance limits the response to immunodominant epitopes. However, the mechanisms that promote T(CD8) subdominance to tumor Ags remain obscure. We investigated the basis for the lack of priming against a subdominant tumor epitope following immunization of C57BL/6 (B6) mice with SV40 large tumor Ag (T Ag)-transformed cells. Immunization of B6 mice with wild-type T Ag-transformed cells primes T(CD8) specific for three immunodominant T Ag epitopes (epitopes I, II/III, and IV) but fails to induce T(CD8) specific for the subdominant T Ag epitope V. Using adoptively transferred T(CD8) from epitope V-specific TCR transgenic mice and immunization with T Ag-transformed cells, we demonstrate that the subdominant epitope V is weakly cross-presented relative to immunodominant epitopes derived from the same protein Ag. Priming of naive epitope V-specific TCR transgenic T(CD8) in B6 mice required cross-presentation by host APC. However, robust expansion of these T(CD8) required additional direct presentation of the subdominant epitope by T Ag-transformed cells and was only significant following immunization with T Ag-expressing cells lacking the immunodominant epitopes. These results indicate that limited cross-presentation coupled with competition by immunodominant epitope-specific T(CD8) contributes to the subdominant nature of a tumor-specific epitope. This finding has implications for vaccination strategies targeting T(CD8) responses to cancer.

GAL41, GAL411, and GAL4I61 are three forms of the yeast transcriptional activator protein that are readily distinguished on the basis of electrophoretic mobility during sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Phosphorylation accounts for the reduced mobility of the slowest-migrating form, GAL4m, which is found to be closely associated with high-level GALIMEL gene expression (L. Mylin, P. Bhat, and J.Hopper, Genes Dev. 3:1157Dev. 3: -1165Dev. 3: , 1989). Here we show that GAIA1, like GALAm, can be converted to GAL41 by phosphatase treatment, suggesting that in vivo GAL411 is derived from GAL41 by phosphorylation. We found that cells which overproduced GALA under conditions in which it drove moderate to low levels of GALIMEL gene expression showed only forms GAL41 and GAL411. To distinguish which forms of GALA (GAL41, GAL41, or both) might be responsible for transcription activation in the absence of GALAm, we performed immunoblot analysis on UASgal-binding-competent GALA proteins from four gal4 missense mutants selected for their inability to activate transcription (M. Johnston and J. Dover, Proc. Nati. Acad. Sci. USA 84: 2401-2405 Genetics 120;63-74, 1988). The three mutants with no detectable GAL) expression did not appear to form GALAI6 or GAL4m, but revertants in which GALA-dependent transcription was restored did display GAL46-or GALA4l-like electrophoretic species. Detection of GAAIA, in a UASgal-binding mutant suggests that neither UASgal binding nor GALIMEL gene activation is required for the formation of GAL4A. Overall, our results imply that GAL41 may be inactive in transcriptional activation, whereas GA1,46 appears to be active. In light of this work, we hypothesize that phosphorylation of GALA1 makes it competent to activate transcription.How eucaryotic transcriptional activators work is an unsolved problem. One eucaryotic transcriptional activator that has been subject to considerable study is the GAL4 protein, which activates transcription of the galactose/melibiose regulon genes (GALIMEL genes) in the yeast Saccharomyces cerevisiae (for a review, see reference 19). In cells grown in the absence of galactose and glucose, GAL4 binds specifically to sites (UASgal) located upstream of the GALI MEL gene promoters (2,10,11,14,27,42,48) but cannot activate transcription as a result of its association with the GAL80 protein (23,28,30,37,45,49). Presentation of galactose (in the absence of glucose) relieves GAL80 inhibition, allowing GAL4 to activate GALIMEL gene transcription. Glucose represses GAL4 protein-dependent transcription (even in the absence of the GAL80 protein [45, 49; reviewed in reference 13]) by mechanisms which include reduced GAL4-UASgal interaction (14,27,42).Learning how GAL4 activity is modulated by carbon source has become a central aim in the overall goal of understanding how it activates transcription. Recently, we discovered that GAL4 protein exists in multiple forms that can be distinguished on the basis of electrophoretic mobility during sodium dodecyl sulfate-p...