The total number of a cell surface glycoprotein molecules at the plasma membrane results from a balance between their constitutive internalization and their egress to the cell surface from intracellular pools and/or biosynthetic pathway. Constitutive internalization is a net result of spontaneous endocytosis and endocytic recycling (Subtil et al., 1994). In this study we have shown that the spontaneous internalization of MHC class I molecules can be followed by depletion of their egress from intracellular sources to the cell surface. Spontaneous endocytosis reduces number of molecules at the cell surface with time, although endocytic recycling prevents depopulation of the cell surface for a long time, indicating for a prevalence of endocytosis over recycling in the absence of protein synthesis. Recycling of class I molecules has been proved by many authors (Naslavsky et al., 2003; Abdel Motal et al., 1993; Reid and Watts, 1990; Grommé et al., 1999), and in our studies of MHC class I endocytic sorting (Mahmutefendić et al., manuscript in preparation). The same principle is operative on both, non-adherent (P815) and adherent (fibroblasts) cell lines.

Constitutive endocytosis and recycling have been described for a number of receptors and transporters (Royle et al., 2003). Cell surface receptors may be categorized into two types, based upon their pattern of internalization: constitutively internalized receptors (such are those for LDL or Tf) that internalize and recycle to the surface at essentially the same rate whether or not they are occupied by ligand (Moore et al., 1995; Anderson et al., 1982; Watts, 1985), and ligand stimulated receptors that activate signal transduction pathways (such as chemokine receptors or β₂-adrenergic receptors) and internalize at a much faster rate when occupied by ligand (Moore et al., 1995; Wiley et al., 1991).

CHX was relatively frequently used in order to determine if endocytosed molecules are subjected either to recycling or degradation, but usually in shorter kinetic studies (up to 8 hours). In some studies CHX did not inhibit recycling or the whole vesicular transport (Machy et al., 1990, Grady et al., 1995), but in other cases retarded these processes (Uhlin-Hansen and Yanagishita, 1995; Roseberry and Hosey, 1999). Although this approach has been used to determine constitutive internalization of some cell surface glycoproteins (e.g. IL2-R, CFTR) (Yu et al., 2000; Morelon et al., 1998; Subtil et al., 1994; Mahmutefendić et al., 2002; Sharma et al., 2004) it was not systematically used to track the spontaneous internalization of membrane glycoproteins. In a study on concanavalin A activated T lymphocytes it was found that in the presence of CHX 20-30% of surface H2^k molecules were internalized in 2 hours, followed by the plateau which lasted for additional 3 hours (Tse and Pernis, 1984). The observation that 70% of the cells did not internalize class I molecules in these conditions led to the assumption that CHX inhibits the synthesis of some short-lived polypeptide(s) important for the process of internalization. However, we have shown that the internalization of H2^d molecules occurs continuously in the presence of CHX on P815 cells (Figures 1, 2, 5 and 7), but also on other cell lines (L-L^d and Balb 3T3 cells, Fig. 1), even after 24 hours. Furthermore, CHX did not inhibit or retard endocytosis of H2^d molecules in the antibody internalization assay (Fig. 5) as well as endocytosis of Tf and CTxB (Fig. 4). It was also reported that other recycling receptors, such as the LDL receptor in fibroblasts, are not affected by inhibition of protein synthesis and their internalization is not influenced by treatment with CHX over the period of 48 hours (Pearse and Bretscher, 1981).

Spontaneous internalization of MHC class I proteins was usually studied by binding of specific mAb to the pre-existing cell surface proteins (Dasgupta et al., 1988; Capps et al., 1989; Machy et al., 1987). Although this method is technically convenient and reproducible, it is questionable whether this approach may address all issues related to constitutive internalization of MHC molecules, as well as other surface glycoproteins. For instance, mAbs may cause capping and cross-linking of surface glycoproteins and thereby induce endocytosis (Favoreel et al., 1999). Furthermore, antibody molecules may unbind from the class I proteins during longer incubations since the process of the association and dissociation is reversible. In addition, it is impossible to ensure that all surface MHC class I molecules are labeled with mAb. These types of studies may be appropriate for tracking of intracellular (vesicular) trafficking of internalized molecules. Usually a small proportion of surface proteins are internalized rapidly and they are distinguished in the intracellular compartment by visualization of an antibody rather than internalized protein. It has been shown in kinetics studies by mAbs that only a proportion (20% in an hour) of cell surface molecules was internalized (Davis et al., 1997). Namely, internalized molecules enter the endocytic pathway, including endocytic recycling compartments (Naslavsky et al., 2003; Radhakrishna and Donaldson, 1997) and subsequently could reappear at the cell surface even after labeling with mAbs (Weber et al., 2004; Sharma et al., 2004).

In order to rule out the possibility of induced endocytosis some authors argue that cross-linking does not increase internalization of MHC class I molecules. It is based on the same kinetics of internalization when either whole mAbs or Fab fragments of mAbs were used (Grady et al., 1995), and that the cross-linking of class I molecules by mAbs does not induce internalization of those molecules that are normally not endocytosed (Dasgupta et al., 1988; Tse and Pernis, 1984). On the other hand, it was reported that cross-linking by antibodies may induce internalization of MHC class I molecules in fibroblasts (Huet et al., 1980).

Internalization of MHC class I molecules determined by external binding of mAbs is faster compared with the spontaneous internalization after depletion of their egress, although overall internalization is rather slow (Fig. 2). External binding of antibodies does not induce endocytosis, but rather MHC class I glycoproteins carry bound antibodies during regular endocytic internalization. The difference observed may be a result of different recycling rates of MHC class I molecules and antibodies bound to them. Although antibodies internalized together with MHC class I molecules recycle back to the cell surface in association with MHC class I molecules (Mahmutefendić et al., manuscript in preparation), the contribution of endocytic recycling in retention of MHC class I molecules at the cell surface is difficult to estimate. However, the internalization kinetics in BFA treated cells is more similar to the internalization kinetics determined by mAbs binding (Figures 1 and 2), indicating that BFA treatment alters also endosomal trafficking, including recycling, in addition to the depletion of MHC class I egress to the plasma membrane. Similar observation that the cell surface half-life is longer when determined by treatment with CHX than by mAbs or ligand binding has been made for IL-2R and its subunits (Yu et al., 2000; Morelon et al., 1998; Subtil et al., 1994). It is interesting that the kinetic of internalization for some receptors (LDLR, IL-2R, D6 receptor) is essentially the same whether the ligand was added together with mAb or not (Subtil et al., 1994; Anderson et al, 1982; Galliera et al., 2004). On the other hand, some receptors (CCR2b, β₂-ARs) cannot be internalized without presence of specific ligand regardless the presence of specific mAb (Moore et al., 1995; Galliera et al., 2004).

MHC class I molecule internalization was also studied by external binding of glycopeptides (Abdel Motal et al, 1993) or fluorescently labeled β₂m (Hochman et al., 1991) on preexisting molecules at the cell surface. These models were convenient to distinguish both kinetics and the mechanisms of their endocytosis, however the acidic milieu of endosomes, that cause dissociation from the ligand, did not allow conclusion on spontaneous internalization rate of MHC class I alleles. Furthermore, Reid and Watts (1990) labeled the cell surface of B lymphoblastoid A46 cell line by DPSgtc reagents, containing iodinated tyrosine residues (¹²⁵I-Tyr) linked by disulphide bond to N-hidroxysuccinimide ester which reacts with amino groups at the cell surface. Using the reducing reagents, ¹²⁵I-Tyr can be detached enabling us to distinguish the surface (non-radioactive) from internalized (radioactive). This model has many advantages, but is technically demanding and includes the work with radioactive material.

MHC class I alleles differ in stability, both intracellulary and at the cell surface (Balendiran et al., 1997). However very little is known about mechanisms of their internalization and intracellular trafficking within vesicular compartments. They also exert qualitative differences in interaction with molecules that are included in peptide binding, such are tapasin, calnexin, calreticulin and Erp57 (Machold et al., 1996). Considering H2^d molecules, the differences are markedly expressed between K^d and D^d molecules from one side, and L^d molecules on the other side (Cook et al., 1996; Beck et al., 1986). In accordance to that, we have used two approaches to determine spontaneous internalization of murine H2^d alleles in different cell types. We have shown that K^dand D^d molecules have similar internalization kinetics in different cell lines, but D^d molecules have longer half life at the cell surface (more than 24 hours) comparing to 15-18 hours for K^d (Fig. 2 and Fig. 6). L^d molecules internalize much faster than K^d and D^d (Fig. 2 and Fig. 4), and their estimated half life at the cell is about 6-8 hours.

It has been reported that the mechanism and the kinetics of MHC class I internalization does not depend only on the allele but also on cell type (Dasgupta et al., 1988; Reid and Watts, 1990; Huet et al., 1980; Machy et al., 1987; Tse and Pernis, 1984; Anderson et al., 1982). Some cell lines show more tendency to internalize their class I molecules, like activated T lymphocytes (Machy et al., 1987; Tse and Pernis, 1984), monocytes and macrophages (Dasgupta et al., 1988), EBV transformed B lymphoblastoid cell line A46 (Reid and Watts, 1990) and B lymphocytes derived 721 221 cell line (Galliera et al., 2004). In contrast, it has been published that endocytosis of MHC class I molecules was not detected in B lymphocytes (Machy et al., 1987), monocyte tumor cell lines (Dasgupta et al., 1988) and L fibroblasts (Machy et al., 1982), and that fibroblasts internalize their MHC class I molecules only after strong cross-binding with multivalent ligands (Huet et al., 1980). However, we have shown that Balb 3T3 and L-L^d fibroblasts (Fig. 1) spontaneously internalize their H2^d molecules after CHX treatment and by mAb internalization assay. In these cell lines internalization kinetics and allele specificities are similar to that in mastocytoma P815 cell line.

Constitutively internalized fully conformed MHC class I molecules neither cross the plasma membrane via clathrin-coated vesicles nor via lipid raft-dependent endocytosis, since the internalization cannot be inhibited by inhibitors of clathrin endocytosis (chlorpromazine) or cholesterol depleting agent (filipin). Thus, we can conclude that fully conformed molecules internalize and enter into the endocytic compartments via the bulk pathway, although they crossroad with the transferrin and cholera toxin internalization pathways (Blagojević et al., manuscript in preparation). Constitutive internalization of non-conformed L^d molecules,, followed either in the CHX model or by mAb binding, is dependent on intact cholesterol-rich membrane microdomains (Fig. 7) indicating their sorting into the lipid rafts at the cell surface and their internalization by lipid-raft dependent endocytosis. This also could result in separation of fully conformed and non-conformed molecules into distinct intracellular vesicular compartments. In general, a majority of fully conformed MHC molecules recycle between the plasma membrane and early endosomal compartments, whereas non-conformed molecules accumulate within intracellular vesicles (Mahmutefendić et al., manuscript in preparation). It is possible that sorting into the lipid rafts is more general biological principle for separation of non-conformed and misfolded glycoproteins at the plasma membrane, and their rerouting into distinct intracellular vesicular pathway in order to separate from fully conformed glycoproteins. An argument supporting this assumption comes form observations that misfolded D^d and L^d, as well as misfolded human MHC class I alleles, are sorted into lipid rafts and internalized by lipid rafts-dependent endocytosis (Mahmutefendić et al., manuscript in preparation).

Understanding of internalization and endocytic processes of MHC class I molecules is very important due to their dynamic role in presenting foreign peptides to the immune system. Although they have been known for a long time to function in the presentation of endogenous antigens to CD8⁺ T lymphocytes, several studies indicate that MHC class I molecules can also present peptides derived from exogenous sources (Grommé et al., 1999; Schirmbeck and Reimann., 1996). Exogenous peptide loading should occur in the endocytic compartment and, therefore, for further understanding of antigen presentation it is essential to understand endocytic trafficking of MHC class I molecules.