In most mammalian cells, damaged, denatured or superfluous proteins are degraded through the ubiquitin proteasome pathway to small peptides, subsequent to release and complete breakdown to amino acids (Fig. 1) 19, 20. A fraction of the peptides released from the proteasome are, however, not degraded further and are instead used for immune surveillance purposes 19–21. Such peptides are taken up into the ER through an ABC family transport system that involves the transporters associated with antigen processing 1 (TAP1) and TAP2 proteins (Fig. 1). TAP1 and TAP2 form a complex that transports peptides across the ER membrane and delivers them to MHC class I protein complexes 22, 23. Peptides of suitable size and sequence are then bound by MHC class I, transported to the cytoplasm via the vesicular system and displayed on the cell surface where they are subject to surveillance by cytotoxic, CD8+ lymphocytes 24. The purpose of such cell surface display of antigens via the MHC class I pathways permits identification of non-self foreign antigens in microorganism-infected cells, which are then targeted for lysis by CTL. However, this pathway is a potential target in cancer therapy if tumor antigens can be targeted for recognition by CTL 12, 25–27. HSP70 family members have the potential to participate in this pathway for antigen processing and presentation due to their ability to bind peptides using the C-terminal polypeptide binding domain 11, 28–30. It has been assumed that HSP70 proteins might bind peptides released into the cytoplasm from the proteasome in a similar way to their acquisition by the TAP1/TAP2 complex. This is largely inferred from the fact that HSP70 extracted from tumors can be used to cross-present (see Fig. 1) tumor antigens on APC, which are recognized by specific clones of cytotoxic CD8+ lymphocytes 6, 12, 31. The peptides that bind to the TAP complex and MHC class I, although of varying sequence, have some common properties; such peptides are 8–10 amino acids in length 32, 33. Most MHC class I ligands bind in extended conformation to the MHC binding groove and have an anchor residue at the C terminus that is either hydrophobic or basic, and such sequences are also preferred by TAP 26, 33. MHC class I-peptide binding is of fairly low affinity (KD 10–6 M), but is almost irreversible in the intact MHC class I-peptide complex. The binding domain of HSP70 has common properties including a similar affinity for peptides (KD approx 10–6 M; L. Mannheim-Rodman and S. K. Calderwood, unpublished) and accommodates peptides in extended conformation of 7–15 amino acids 34–36. A number of studies have addressed the peptide sequence binding preferences of HSP70 proteins and indicated a similar peptide binding preference for MHC class I, indicating roles for hydrophobic and basic amino acids 30, 34–36. Inferences derived from these studies, however, are complicated by the findings that substrate preferences in HSP70 family proteins are determined not solely by the properties of the peptide binding domain of HSP70 itself, but by J-domain proteins, co-factors that localize HSP70 to target molecules and effect binding by modulating the ATPase activity of HSP70 (Fig. 1) 37, 38. A range of J domain proteins with different substrate preferences have been found 11, 39–41. Thus, peptide binding by HSP70 proteins may be determined at least partially by the J domain protein partner 38. There is an interesting confluence in properties between MHC class I, MHC class II and HSP70 in that each protein requires the assistance of a co-factor for peptide association; TAP1 is required for MHC class I-peptide binding, HLA-DM for MHC class II and J domain proteins for HSP70. It is currently unknown whether HSP70 binds preferentially to any class of intracellular peptides. However, in an in vivo proteomic study, Grossmann et al. 42 suggest that HSP70 binds to peptides of generally 8–26 amino acids in length. Furthermore, HSP70 favorably interacts to a 5-amino acid core sequence, which likely contains some acidic residues 42. One compelling hypothesis is that a fraction of the intracellular HSP70 binds peptides released into the cytoplasm from the proteasome, protects them from further degradation and passes them to the TAP1/TAP2 complex. It is known that HSP70 proteins function “upstream” in the processing of proteins. HSP70 can bind to damaged target proteins, while associated with CHIP (C terminus of HSC70-interacting protein), which functions as an HSC70 binding ubiquitin E3 ligase, marking the target protein for proteasomal degradation by ligation of its side chains with ubiquitin 43–45. This complex is then transported to the proteasome in the company of another HSP70 binding protein Bag-I 44. HSP70 may then perform an “end-around” and accept and chaperone peptides extruded from the exit tunnel of the proteasome.
In addition, HSP70 family member HSC73 takes part in another form of antigen processing; HSC73 is evidently required for the processing of external antigens in the lysosome proximal endosome compartments, and is thus involved in the processing of peptides that associate with MHC class II proteins in APC and activate immune response through CD4 T lymphocytes 46. HSC73 co-associates with MHC class II in spherical organelles in macrophages and leads to the presentation of external antigens to MHC class II-restricted T lymphocytes 46. A role for HSC73 in protecting peptides generated by proteolysis in MHC class II-containing endosomes from further breakdown has been suggested, which is similar to the role proposed for the HSP70 family in protecting cytoplasmic peptides 47–49. The presence of HSC73 in clathrin-coated pits present in endosomes and lysosomes is well established and HSC73 has been shown by Dice and co-workers 47–49 to participate in targeting proteins for degradation through recognition of a consensus sequence (KFERQ). A common theme for HSP70, of mobilizing target proteins towards the sites of degradation while sparing a fraction of the partially digested peptide for immune surveillance, is suggested for both main pathways of protein degradation in the cytoplasm (proteasome) and lysosome. As the lysosomal pathway can be stimulated by serum starvation of tissue culture cells, it may thus play a role in generation of tumor immunity in nutritionally deprived tumor cells, which could potentially be captured in the production of anti-tumor HSP70 based vaccines 48. A major unknown in these studies is, whether, HSP70 proteins capture a nonspecific sample of intracellular peptide or actively select classes of peptides along structural lines, guided perhaps by associated J-domain proteins. In this context it has been shown recently that immunosuppressive drug 15-deoxyspergualin (DSG) specifically binds with high affinity to HSC70 and inhibits interactions that do not require DnaJ co-chaperone activity 18, 50. Although these studies are not definitive, as the properties of DSG are not completely understood, they do indicate a need to understand the mechanisms of molecular chaperone-peptide interactions in antigen processing.