Melanosomes made from recycling (endosomes): A tubule-stabilizing function revealed for Biogenesis of Lysosome-related Organelles Complex-1.

Abstract

The melanosomes found in pigment-producing cells from humans and other vertebrates are morphologically unique organelles that have long been recognized as related to, but distinct from, conventional lysosomes. Several components of the molecular machinery involved in the biogenesis of melanosomes—as well as of other lysosome-related organelles—have been identified through genetic analyses of syndromic forms of albinism such as Hermansky–Pudlak syndrome (HPS). These components include the proteins dysbindin, BLOS3 and pallidin, which are encoded by the genes mutated in HPS types 7, 8 and 9, respectively, and—together with another five gene products—assemble into Biogenesis of Lysosome-related Organelles Complex-1 (BLOC-1). Despite intense research for over a decade, our understanding of the function of this stable protein complex had remained quite limited. For example, we had learned that membrane proteins that normally localize to melanosomes, such as tyrosinase-related protein 1 (Tyrp1), are mislocalized in BLOC-1-deficient melanocytes (Setty et al., 2007), and that other membrane proteins could likewise be mislocalized in other types of BLOC-1-deficient cells. However, the mechanism by which BLOC-1 would mediate the correct intracellular targeting of these proteins had remained obscure. Shedding light on this issue, a new study by Delevoye et al. in this year's first issue of Current Biology describes the role of BLOC-1 in coordinating the action of a molecular motor and other cytoskeleton components on the elongation and detachment of recycling endosomal tubules. The ‘textbook’ model of membrane protein trafficking mediated by small spherical vesicles that bud and detach from, travel through the cytoplasm between, and fuse with, stable vacuolar compartments, has been a powerful framework to understand the mechanisms of action of well-characterized regulators of protein sorting. However, it has been inadequate to understand several protein trafficking events that take place in the endosomal system, which morphologically can be characterized as a network of tubules emanating from vacuolar compartments. Previous work by Cédric Delevoye at Graça Raposo's laboratory, in collaboration with the laboratory of Michael S. Marks and those of other colleagues, had uncovered a direct role for a subset of endosomal tubules in the delivery of membrane proteins to maturing melanosomes (Delevoye et al., 2009). Biogenesis of these endosomal tubules, which were characterized as part of the recycling endosomal system based on the presence of internalized transferrin and the small GTPase Rab11, required the function of a microtubule-associated motor named Kinesin Family member 13A (KIF13A). Accordingly, depletion of KIF13A from a human pigmented melanoma cell line, MNT-1, by small interference RNA (siRNA), resulted in partial hypopigmentation and mislocalization of Tyrp1 to vacuolar endosomes—the compartment from where these tubular recycling endosomes would originate (Delevoye et al., 2009). Intriguingly, these hypopigmentation and protein mislocalization phenotypes of KIF13A-depleted cells were reminiscent to those previously observed for BLOC-1-deficient cells (Setty et al., 2007). These observations, together with the known localization of BLOC-1 to transferrin-positive endosomal tubules, led these authors—this time in collaboration with the groups of Victor Faundez and Etienne Morel—to test for a role of BLOC-1 in the biogenesis of recycling endosomal tubules. “[in BLOC-1-depleted cells] vacuolar sorting endosomes displayed an increased number of budding profiles” Even when KIF13A is a motor that functions on microtubule tracks, actin cytoskeleton rearrangements are known to contribute to the formation and/or detachment of tubules from vacuolar endosomes. Using well-characterized pharmacological agents, Delevoye et al. observed that disruption of actin fibers, or inhibition of the actin-branching activity of the ARP2/3 complex, resulted in decreased number and length of KIF13A-positive tubules emanating from transferrin-positive structures in live HeLa cells. At least at the level of fluorescence microscopy, these effects were quite similar to those observed for BLOC-1-depleted cells. Could there be a functional link? Although the very long list of potential binding partners described for BLOC-1 includes the Wiskott–Aldrich syndrome homologue (WASH) complex, which is a known activator of the ARP2/3 complex, Delevoye et al. were able to rule out a non-redundant role for WASH in the formation of KIF13A-positive tubules. Instead, the authors observed that siRNA-mediated depletion of Annexin A2, a protein which had been previously shown to initiate actin polymerization on vacuolar endosomes, resulted in reductions in number and length of KIF13A-positive tubules that were similar to those observed for BLOC-1-depleted cells. A key difference, however, was noted upon correlative light and electron microscopy: while BLOC-1 depletion resulted in an increased number of morphologically normal budding profiles associated with vacuolar sorting endosomes, depletion of Annexin A2 resulted in accumulation on these organelles of membrane profiles that were described as ‘elongated buds’ (because they were deemed too asymmetrical to be considered ‘buds’ and too short to be considered ‘tubules’). These observations have led Delevoye et al. to propose that Annexin A2 may function at a step subsequent to that of BLOC-1, for instance, in the scission of tubules formed by KIF13A and stabilized by BLOC-1. Are these observations made using HeLa cells, which lack lysosome-related organelles, of relevance to melanosome biogenesis? Delevoye et al. provide evidence to argue that this is indeed the case. Significantly shorter tubules containing the endosomal protein syntaxin 13 (fused to green fluorescent protein) were observed upon live imaging of a cell line derived from skin melanocytes of BLOC-1-deficient ‘pallid’ mice (which lack the pallidin subunit), as compared to the same cells ‘rescued’ by transient expression of the missing subunit. In addition, siRNA-mediated depletion from MNT-1 cells of BLOC-1, KIF13A, Annexin A2, any pairwise combination of them, or the three of them combined, resulted in very similar hypopigmentation and melanosomal maturation phenotypes, with the lack of additive effects being taken as evidence for BLOC-1, KIF13A and Annexin A2 functioning in the same pathway. Undoubtedly, this new study by Delevoye et al. represents a major advance in our understanding of the mechanism of action of BLOC-1. Its function in promoting the action of a kinesin on intracellular organelles is akin to that of BLOC-one-related complex (BORC), a recently described multimeric protein complex that shares with BLOC-1 three common subunits (BLOS1, BLOS2, and snapin) and promotes the action of kinesin-1 on lysosomes (Pu et al., 2015). Granted, the fact that snapin is a subunit shared by BLOC-1 and BORC raises the technical issue that the siRNA treatments using a cocktail of reagents targeting snapin and two BLOC-1-specific subunits might have resulted in depletion of not only BLOC-1 but also BORC, but future experiments should be able to address this straightforwardly. Another issue that merits future research is the possibility that other kinesins besides KIF13A may function in melanosome biogenesis with some degree of redundancy, at least in mouse skin melanocytes, as Kif13a-knockout mice have been shown to display apparently normal coat color (Zhou et al., 2013). Finally, it will be interesting to examine whether BLOC-1 is required for the biogenesis of all recycling endosomes or only a subset of them. As mentioned in previous paragraphs, BLOC-1 is expressed not only in cells that produce lysosome-related organelles (e.g., melanocytes) but also in virtually all other cell types—including those of the central nervous system. Delevoye et al. cite literature that argues for variations in the genes encoding BLOC-1 subunits (e.g., dysbindin) and KIF13A being associated to the risk of developing schizophrenia. To be clear, neither KIF13A nor any of the genes encoding BLOC-1 subunits have been validated as schizophrenia susceptibility genes. The human gene encoding the dysbindin subunit of BLOC-1, DTNBP1, was for several years considered a ‘plausible’ candidate risk gene for the disease, but despite numerous genetic studies—using different approaches and patient populations—the proposed association has systematically failed to reach ‘genomewide’ statistical significance (Farrell et al., 2015). Yet, during the same period it has become apparent that BLOC-1 somehow plays an important role in the mouse brain, as BLOC-1-deficient mice display several behavioral and electrophysiological phenotypes. With regards to the KIF13A gene, even when no positive genetic association with schizophrenia has ever been reported (http://www.szgene.org/geneoverview.asp?geneID=377), Kif13a-knockout mice have been shown to display anxiety-related behavioral phenotypes (Zhou et al., 2013). Interestingly, neurons isolated from these mouse models of BLOC-1 or KIF13A deficiency display altered surface expression of certain membrane proteins, which in light of this new work by Delevoye et al. could be interpreted as a consequence of impaired biogenesis of tubular recycling endosomes. Needless to say, these are exciting ideas that deserve future investigation.