Pigment organelles, or melanosomes, are transported by kinesin, dynein, and myosin motors. As such, melanosome transport is an excellent model system to study the functional relationship between the microtubule-and actin-based transport systems. In mammalian melanocytes, it is well known that the Rab27a/melanophilin/myosin Va complex mediates actin-based transport in vivo. However, pathways that regulate the overall directionality of melanosomes on the actin/microtubule networks have not yet been delineated. Here, we investigated the role of PKA-dependent phosphorylation on the activity of the actin-based Rab27a/melanophilin/myosin Va transport complex in vitro. We found that melanophilin, specifically its C-terminal actin-binding domain (ABD), is a target of PKA. Notably, in vitro phosphorylation of the ABD closely recapitulated the previously described in vivo phosphorylation pattern. Unexpectedly, we found that phosphorylation of the ABD affected neither the interaction of the complex with actin nor its movement along actin tracks. Surprisingly, the phosphorylation state of melanophilin was instead important for reversible association with microtubules in vitro. Dephosphorylated melanophilin preferred binding to microtubules even in the presence of actin, whereas phosphorylated melanophilin associated with actin. Indeed, when actin and microtubules were present simultaneously, melanophilin's phosphorylation state enforced track selection of the Rab27a/melanophilin/ myosin Va transport complex. Collectively, our results unmasked the regulatory dominance of the melanophilin adaptor protein over its associated motor and offer an unexpected mechanism by which filaments of the cytoskeletal network compete for the moving organelles to accomplish directional transport on the cytoskeleton in vivo.melanophilin | myosin Va | intracellular transport | transport regulation

Cross-talk between the microtubule and actin networks has come under intense scrutiny following the realization that it is crucial for numerous essential processes, ranging from cytokinesis to cell migration. It is becoming increasingly clear that proteins long-considered highly specific for one or the other cytoskeletal system do, in fact, make use of both filament types. How this functional duality of “shared proteins” has evolved and how their coadaptation enables cross-talk at the molecular level remain largely unknown. We previously discovered that the mammalian adaptor protein melanophilin of the actin-associated myosin motor is one such “shared protein,” which also interacts with microtubules in vitro. In a hypothesis-driven in vitro and in silico approach, we turn to early and lower vertebrates and ask two fundamental questions. First, is the capability of interacting with microtubules and actin filaments unique to mammalian melanophilin or did it evolve over time? Second, what is the functional consequence of being able to interact with both filament types at the cellular level? We describe the emergence of a protein domain that confers the capability of interacting with both filament types onto melanophilin. Strikingly, our computational modeling demonstrates that the regulatory power of this domain on the microscopic scale alone is sufficient to recapitulate previously observed behavior of pigment organelles in amphibian melanophores. Collectively, our dissection provides a molecular framework for explaining the underpinnings of functional cross-talk and its potential to orchestrate the cell-wide redistribution of organelles on the cytoskeleton.

1Myosin-V (MyoV) is a ubiquitous motor protein that transports an astonishingly diverse 2 set of cargos on the actin network in eukaryotes. Phosphorylation-dependent 3 processes often regulate MyoV-mediated cargo transport, molecular details of which 4 remain largely unknown. We previously showed that phosphorylation regulates MyoV's 5 switching from microtubules onto actin filaments, not its motor activity. Regulation of 6 switching at reconstituted microtubule-actin-crossings in fact sufficed to recapitulate 7 the MyoVa-driven redistribution of pigment-organelles in amphibian melanophores. 8 However, in those cells, MyoVa also encounter many actin-actin crossings. Here, we 9 show that isolated MyoVa motors switch with equal probabilities at reconstituted actin-10 actin-crossings. Under the control of its adaptor-protein melanophilin (Mlph), however, 11 the motor differentiates between the actin filaments at crossing points in a 12 phosphorylation-regulated manner. Whereas phosphorylation of Mlph forced about 13 ~2/3 of MyoVa to ignore the intersections, dephosphorylation completely reversed this 14 behavior and forced ~2/3 to switch. We show that the filament-binding domain (FBD) 15 of Mlph controls this switching behavior. This property evolved in amphibians, but not 16 in the early vertebrate zebrafish. By protein engineering, we demonstrate that changes 17 of a few residues are sufficient to impart actin-binding capability onto the zebrafish 18 Mlph. We thus unmask the molecular beginnings of dual filament binding in Mlph that 19 allow it to control the switching behavior of MyoVa at cytoskeletal crossings. We 20 therefore propose a direct link between intracellular phosphorylation activity and the 21 adaptor-protein, not to regulate MyoVa activity, but to navigate the motor through the 22 entire cytoskeletal maze for correct positioning of cargo. 23Significance statement 25 In virtually all eukaryotic cells, numerous myosin motors have to navigate through an 26 elaborate actin network for timely transport of intracellular cargo. Here, we unmask an 27 unintuitive regulation of the myosin-Va motor that is involved in pigment organelle 28 transport. We demonstrate that myosin-Va differentiates between the same actin 29 filaments and displays regulated switching at reconstituted actin-actin crossings, an 30 unexpected behavior that has been predicted from previous theoretical work. We trace 31 this regulation back to the adaptor protein of the myosin-Va motor and show that this 32 regulation was present in amphibian but had not evolved in the early vertebrate 33 zebrafish. Notably, we demonstrate that the evolution of actin-binding capability is 34 achieved by changing a few residues in the adaptor protein.