Unlike wild-type mouse melanocytes, where melanosomes are concentrated in dendrites and dendritic tips, melanosomes in dilute (myosin Va−) melanocytes are concentrated in the cell center. Here we sought to define the role that myosin Va plays in melanosome transport and distribution. Actin filaments that comprise a cortical shell running the length of the dendrite were found to exhibit a random orientation, suggesting that myosin Va could drive the outward spreading of melanosomes by catalyzing random walks. In contrast to this mechanism, time lapse video microscopy revealed that melanosomes undergo rapid (∼1.5 μm/s) microtubule-dependent movements to the periphery and back again. This bidirectional traffic occurs in both wild-type and dilute melanocytes, but it is more obvious in dilute melanocytes because the only melanosomes in their periphery are those undergoing this movement. While providing an efficient means to transport melanosomes to the periphery, this component does not by itself result in their net accumulation there. These observations, together with previous studies showing extensive colocalization of myosin Va and melanosomes in the actin-rich periphery, suggest a mechanism in which a myosin Va–dependent interaction of melanosomes with F-actin in the periphery prevents these organelles from returning on microtubules to the cell center, causing their distal accumulation. This “capture” model is supported by the demonstration that (a) expression of the myosin Va tail domain within wild-type cells creates a dilute-like phenotype via a process involving initial colocalization of tail domains with melanosomes in the periphery, followed by an ∼120-min, microtubule-based redistribution of melanosomes to the cell center; (b) microtubule-dependent melanosome movement appears to be damped by myosin Va; (c) intermittent, microtubule-independent, ∼0.14 μm/s melanosome movements are seen only in wild-type melanocytes; and (d) these movements do not drive obvious spreading of melanosomes over 90 min. We conclude that long-range, bidirectional, microtubule-dependent melanosome movements, coupled with actomyosin Va–dependent capture of melanosomes in the periphery, is the predominant mechanism responsible for the centrifugal transport and peripheral accumulation of melanosomes in mouse melanocytes. This mechanism represents an alternative to straightforward transport models when interpreting other myosin V mutant phenotypes.

No abstract

Melanocytes that lack the GTPase Rab27a (ashen) are disabled in myosin Va-dependent melanosome capture because the association of the myosin with the melanosome surface depends on the presence of this resident melanosomal membrane protein. One interpretation of these observations is that Rab27a functions wholly or in part as the melanosome receptor for myosin Va (Myo5a). Herein, we show that the ability of the myosin Va tail domain to localize to the melanosome and generate a myosin Va null (dilute) phenotype in wild-type melanocytes is absolutely dependent on the presence of exon F, one of two alternatively spliced exons present in the tail of the melanocyte-spliced isoform of myosin Va but not the brain-spliced isoform. Exon D, the other melanocyte-specific tail exon, is not required. Similarly, the ability of full-length myosin Va to colocalize with melanosomes and to rescue their distribution in dilute melanocytes requires exon F but not exon D. These results predict that an interaction between myosin Va and Rab27a should be exon F dependent. Consistent with this, Rab27a present in detergent lysates of melanocytes binds to beads coated with purified, full-length melanocyte myosin Va and melanocyte myosin Va lacking exon D, but not to beads coated with melanocyte myosin Va lacking exon F or brain myosin Va. Moreover, the preparation of melanocyte lysates in the presence of GDP rather than guanosine-5Ј-O-(3-thio)triphosphate reduces the amount of Rab27a bound to melanocyte myosin Va-coated beads by approximately fourfold. Finally, pure Rab27a does not bind to myosin Va-coated beads, suggesting that these two proteins interact indirectly. Together, these results argue that Rab27a is an essential component of a protein complex that serves as the melanosome receptor for myosin Va, suggest that this complex contains at least one additional protein capable of bridging the indirect interaction between Rab27a and myosin Va, and imply that the recruitment of myosin Va to the melanosome surface in vivo should be regulated by factors controlling the nucleotide state of Rab27a.