We have shown that myosin VI and its interacting protein optineurin are essential for directed cell migration in mammalian cells. The loss of either optineurin or myosin VI inhibits the delivery of EGFR to the leading edge at the front of the cell and causes a dramatic defect in lamellipodia formation. Surprisingly, this does affect random migration of myosin VI or optineurin KD cells; however, both proteins are crucial for cells to perform directed migration towards a chemotactic gradient. As optineurin has been shown to mediate the function of myosin VI in post-Golgi membrane transport, our results indicate that the role of myosin VI in cell motility is linked to anterograde membrane trafficking pathways.
Numerous lines of evidence now suggest that in a motile polarized cell plasma membrane proteins are delivered into the leading edge in an anterograde direction from the Golgi complex and/or the endocytic recycling pathway. In this process, directed secretion towards the wound edge is dependent on Golgi positioning; however, very little is known about the other constituents of the machinery responsible for polarized delivery into the leading edge. Interestingly, in neither myosin VI or optineurin KD cells was a defect in Golgi reorientation towards the wound observed, indicating that myosin VI and optineurin are not involved in this process, but are most likely involved in polarized post-Golgi transport or sorting steps. Furthermore, as cdc42 is the key regulator of cell polarity and Golgi reorientation, our results suggest that myosin VI and optineurin are not major players in this cdc42-dependent signalling pathway. This hypothesis is supported by the finding that depletion of functional myosin VI and optineurin leads to a reduction of Rac-dependent lamellipodia but to an increase in finger-like protrusions known as filopodia, which in most cell types are formed after cdc42 activation. This replacement of broad lamellipodia and ruffles with finger-like filopodia can be observed both in myosin VI-depleted cells in scratch wound assays of cell monolayers () and also during single cell spreading in GFP-myosin VI tail-expressing cells. Interestingly, these dramatic morphological changes do not result in changes in the speed of spreading of myosin VI-depleted cells (data not shown).
In epithelial cells, distinct pathways for the selective sorting of cargo leaving the trans
Golgi network to either the apical or basolateral cell surface have been described. In polarized MDCK cells, we have shown a role for myosin VI and optineurin in the selective sorting of cargo such as vesicular stomatitis virus glycoprotein (VSV-G) to the basolateral domain (22)
. In a motile fibroblast, VSV-G is targeted to the leading edge (35)
, suggesting that similar pathways and machinery used for delivery of proteins to the basolateral domain are also required for targeting of proteins to the leading edge in motile cells. Interestingly in neurons, trafficking in the biosynthetic pathway is also targeted to distinct subdomains of the plasma membrane, e.g. to the dendritic surface of the cell body, which is similar to the basolateral domain and to the axonal membrane, which can be regarded as analogous to the apical domain of MDCK cells (24)
. Thus, in hypocampal neurons, exocytic trafficking of AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptor to the postsynaptic membrane at the somatodendritic cell surface, the basolateral analogue, has been shown to require myosin VI (33)
. These observations give further support to our observations that a complex of myosin VI and optineurin drives the local delivery of EGFR into the leading edge of a motile cell at the wound edge.
Previously, myosin VI and optineurin have been linked to constitutive exocytosis of reporter molecules such as secreted alkaline phosphatase (SEAP) and VSV-G; depletion of either myosin VI or optineurin resulted in a kinetic delay in delivery of the reporters to the cell surface (14
). In this study, we show that the steady-state surface levels of endogenous transmembrane proteins, such as the EGFR, are not affected in myosin VI or optineurin KD cells. The absence of myosin VI or optineurin, however, dramatically impairs the selective delivery of these membrane proteins into the leading edge. These results indicate that myosin VI and optineurin are required for the spatial delivery of membrane proteins in the anterograde trafficking pathway, but are not essential for steady-state secretion of every cargo to the plasma membrane. Transmembrane proteins are either delivered directly from the Golgi complex to the plasma membrane or via an intermediate recycling endosome. In polarized epithelial cells, myosin VI and optineurin are present on a specialized recycling compartment, a sorting station for proteins delivered to the basolateral domain (22)
. In epithelial cells, cargo en route
to the basolateral domain can pass through a possibly specialized type of recycling endosome, and also in macrophages this type of organelle is important for polarized secretion of cytokines (36)
. This proposed role of myosin VI and optineurin in anterograde trafficking pathways, which include the route from the recycling compartment to the plasma membrane, is supported by our finding that polarized delivery of β1-integrin towards the cell front is also affected in myosin VI or optineurin A549 KD cells (Figure S5
). Future work will confirm whether the myosin VI-dependent delivery of the EGFR to the leading edge involves a sorting step in a recycling compartment, whereas transport to the rest of the plasma membrane proceeds directly from the trans
Golgi network and therefore may not be dependent on myosin VI.
Interestingly, the absence of myosin VI or optineurin, which leads to a dramatic reduction in lamellipodia formation at the leading edge, does not lead to defects in random cell migration. In contrast, myosin VI or optineurin-depleted cells show an increase in motility, which is reflected in an increase in MSD as well as the velocity over time. Although it seems surprising that myosin VI and optineurin KD cells with less lamellipodia show increased random migration, this is not a new observation. In epithelial PtK1 cells, non-muscle tropomyosin inhibition causes a loss of lamellipodia, but a dramatic increase in random migration velocity (37)
, supporting our observation that indeed cells can migrate without a lamellipodium.
Although myosin VI or optineurin KD cells do not lack the ability to move, they are impaired in performing directed migration towards a stimulus gradient. This type of movement requires a signal perceived and amplified in a restricted area at the front of the cell. Although a polarized concentration of signalling receptors at the leading edge is not required in all cell types for chemotaxis (38)
, we clearly observed an accumulation of EGFR in the plasma membrane of protrusions at the front of A549 cells close to the coverslip using confocal microscopy. Therefore, we believe that the polarized localization of EGFR at the leading edge of A549 cells enhances localized signalling at the front, which is important for directed migration (39)
. Polarized secretion has also been shown in migrating fibroblasts, where secretory vesicles are preferentially delivered and inserted at the leading edge (7)
Our results are consistent with previous work in Drosophila that implicated myosin VI in epithelial cell migration required for dorsal closure during embryogenesis (10)
and in border cell migration in oocytes (11)
. In these border cells, inhibition of myosin VI also leads to a reduction in protrusion formation at the leading edge. From Drosophila oocytes, myosin VI can be isolated in a complex with Drosophila E-cadherin and it was therefore proposed that myosin VI promoted protrusion formation by moving towards the minus end of the polymerizing actin filaments while binding to a stationary E-cadherin complex and thereby pushing the actin filaments into the leading edge. At present, we cannot exclude this role for myosin VI in protrusion formation in mammalian cells; however, we did not observe any association of myosin VI with focal adhesion complexes at the leading edge in A549 cells or any other defects in focal adhesion formation or cell spreading in myosin VI-depleted cells. Our results, however, are supported by data showing that although vinculin is a downstream effector of myosin VI at E-cadherin-dependent cell–cell contact sites, no changes or defects at integrin-based cell-matrix adhesion sites were observed upon myosin VI depletion (15)
. In contrast, a recent study indicates that in endothelial cells myosin VI and its interacting protein GIPC are important for α5β1 integrin endocytosis and that the absence of either protein causes defects in cell adhesion (20)
. These different results suggest that there are probably cell-specific variations in the involvement of myosin VI in cell adhesion formation.
In addition, we cannot exclude a role for myosin VI in actin filament reorganization at the leading edge, although so far no interacting protein directly linking myosin VI to the machinery regulating actin filament dynamics has been identified in mammalian cells. However, we have previously shown the recruitment of myosin VI into the leading edge after EGF stimulation (28)
. Furthermore, in Drosophila myosin VI has been shown to stabilize actin filaments required for spermatid individualization (40)
. Therefore, future work is required to show whether myosin VI coordinates membrane delivery and actin organization at the leading edge.
In summary, we have clearly shown that myosin VI together with the adaptor protein optineurin is required for the delivery of specific cargo such as the EGFR in the anterograde pathway towards the front of a migrating cell. This delivery is essential for protrusion formation at the leading edge of the cell and for the generation of directionally persistent migration.