Adecuaciones en el laboratorio

As discussed in the Introduction, the molecular function of BLOC-1 is not known. What is clear is that BLOC-1-deficiency causes a buildup of Tyrp1 in vacuolar endosomal domains and a dramatic increase in the flux of Tyrp1 through the endocytic recycling pathway to the cell surface (Setty et al., 2007). Transferrin-positive tubules that are depleted of Tyrp1 are detectable emanating from endosomes in wild-type melanocytes as well as in BLOC-1-deficient melanocytes, suggesting that BLOC-1 activity primarily directs recruitment of cargo and/or effectors into the tubules, although since tubules also mediate recycling to the plasma membrane and the Golgi, the destination of the tubules observed in BLOC-1-deficient cells is not clear and thus a role for BLOC-1 in the formation or stabilization of melanosome-bound tubules cannot be ruled out.

Figure 4.1

Figure 4.1. New model of endosome to melanosome trafficking.

Previous research suggested a model of endosome to melanosome trafficking with two independent pathways, one mediated by AP-3 and used by the majority of tyrosinase, and one mediated by BLOC-1 in conjunction with AP-1 and used by Tyrp1. The research presented here suggest that OCA2 utilizes a distinct pathway that is both BLOC-1- and AP-3-dependent.

The similarity of the AP-1 knockdown and BLOC-1-deficiency phenotypes on Tyrp1 trafficking suggest that BLOC-1 and AP-1 functionally cooperate in cargo recruitment to the endosomal tubules (Setty et al., 2007; Delevoye et al., 2009). Localization of OCA2 has not been investigated in melanocytes in which AP-1 has been knocked down, but in wild-type cells, the OCA2 proline mutants, deficient in AP-1 binding but competent to bind AP-3, are still able to traffic to melanosomes, albeit less efficiently. This suggests that BLOC-1 can additionally work with AP-3 to facilitate cargo entry into the

endosomal tubules (Figure 4.2). The role of the AP complex in BLOC-1 cargo selection might be in the regulation of BLOC-1 membrane recruitment. The percentage of BLOC-1 that cofractionates with membranes relative to cytosol from mouse skin fibroblasts is reduced when AP-3 is genetically deficient (Di Pietro et al., 2006). Furthermore, AP-3 and BLOC-1 interactions appear to be regulated by cargo. For example, AP-3 and BLOC-1 can be co-immunoprecipitated with epitope-tagged PI4KIIalpha, an AP-3 cargo protein, from HEK293T cells in which the cargo has been transiently expressed (Salazar et al., 2009), but coprecipitation of both is reduced in cells that express a PI4KIIalpha with a mutation in its dileucine motif; this suggests that BLOC-1 recruitment requires an AP-3/PI4KIIalpha interaction. An alternative explanation for this result is that both BLOC-1 and AP-3 interact simultaneously with cargo proteins via the dileucine motif, but it is difficult to imagine how two complexes could simultaneously engage such a small peptide, especially considering the tight fit of the acidic and leucine residues into pockets within the AP complex. Moreover, BLOC-1 has never been shown to bind to the cytoplasmic domains of any of its cargoes. Finally, the OCA2 LL1 signal, like the acidic

Figure 4.2

Figure 4.2 Model of endosome-to-melanosome trafficking of OCA2.

Transferrin-positive tubules emanating from recycling endosomal domains making physical connections with melanosomes have been visualized in a human melanoma cell line. Tyrp1 has also been visualized on AP-1 coated endosomal tubules, and deficiency in BLOC-1 or AP-1 causes mislocalization of Tyrp1 to vacuolar endosomal domains, suggesting that BLOC-1 and AP-1 together mediate Tyrp1 entry into tubules. Our results from studies of OCA2 trafficking in mouse melanocytes are consistent with a model in which entry of OCA2 (blue) into tubules is also BLOC-1-dependent. Final delivery of OCA2 to melanosomes requires interaction with AP-3, predicted to take place in AP-3- coated buds analogous to the AP-1-coated buds that are proposed to traffic Tyrp1. (Modified from (Delevoye et al., 2009))

dileucine-based signal of tyrosinase (Theos et al., 2005), interacts with both AP-1 and AP-3, but unlike tyrosinase, OCA2 is highly dependent on BLOC-1 for its trafficking. This reliance on BLOC-1 is maintained even when the LL1 signal on OCA2 is replaced by that of tyrosinase, as indicated by the lack of overlap of OCA2-AA23 hTyr with PMEL- containing melanosomes in BLOC-1-deficient cells. This indicates that the sorting signal itself is not sufficient to determine whether or not the cargo uses a BLOC-1-dependent or –independent pathway.

Other determinants must distinguish the choice of whether to utilize the BLOC-1 pathway (Figure 4.1). I can only speculate on the identity of such determinants, but one possibility is suggested by the exclusion of transferrin from melanosomes during the formation of transient, tubular connections between recycling endosomes and

melanosomes. Transferrin remains associated with transferrin receptor—a dimeric, single transmembrane domain containing type II integral membrane protein—throughout the endocytic system (Dautry-Varsat et al., 1983). One might speculate that a “gating” mechanism that prevents transferrin/transferrin receptor accumulation in the melanosome is more generally used to prevent similarly composed integral membrane proteins from indiscriminately diffusing into the melanosome. Such a gate might function by excluding proteins of a certain size, membrane spanning domain content, or lipid microdomain. OCA2 is a 12-transmembrane domain protein and is therefore much larger than tyrosinase, which is a single-pass protein like transferrin receptor. Although Tyrp1

traverses the tubules and is approximately the same size as tyrosinase, previously published (Setty et al., 2008) and unpublished evidence suggests that Tyrp1 traffics to melanosomes together with ATP7A; indeed, ATP7A depletion results in the mistargeting of Tyrp1 to lysosomes and in its degradation (unpublished results). ATP7A is a P-type ATPase with 8 membrane-spanning domains (Lutsenko et al., 2007), and thus might have properties more in common with OCA2. Finally, when the OCA2 N-terminal

cytoplasmic domain was fused to the transmembrane and lumenal domains of human TfR and expressed in non-melanocytic Chinese hamster ovary cells, the chimera was able to localize to lysosomes in a dileucine motif-dependent manner (Sitaram et al., 2009), as expected for a melanosomal protein expressed in a non-melanocytic cell. However, in melanocytes the chimera localized to an unidentified intracellular compartment that did not colocalize with pigment, perhaps falling prey to the same mechanism that prevents melanosomal accumulation of native transferrin receptor. These examples suggest the existence of a gating mechanism through the BLOC-1-dependent pathway that might only permit entry to proteins or protein complexes with multiple membrane spanning

domains.

The gating mechanism might not function at the level of the proteins themselves, but rather through lipid interactions. Tyrp1 and tyrosinase both require glycosphingolipids for proper intracellular trafficking to melanosomes (Sprong et al., 2001; Groux-Degroote et al., 2008). Furthermore, experiments with chimeric proteins consisting of lumenal and

cytoplasmic domain swaps between Tyrp1 and the lysosomal protein, LAMP1, or between Tyr and LAMP1 showed that the determinants for the lipid-dependent behavior reside in the lumenal domains of the proteins. The failure of the OCA2 cytoplasmic domain chimera to traffic to melanosomes could thus be explained if similar interactions between lipids and OCA2 lumenal determinants regulated trafficking and passage through the endosome/melanosome “gate”. It will be interesting to see whether OCA2 trafficking similarly depends on glycosphingolipids.