Biology Reference
In-Depth Information
FAK and zyxin, but not visibly organized a-actinin, localize in these dynamic
paxillin adhesions. The turnover is a property of the adhesion as a whole since
other components turn over as well. For example, when paxillin-DsRed2 is
co-expressed with either GFP-FAK or zyxin-GFP, all three components turn
over at the base of the protruding lamellipodium. Thus, the adhesions appear
to turn over in protruding regions of the cell. This disassembly of adhesions
followed by formation of new adhesions appears to be important since we
consistently see a coupling between the rate of adhesion turnover and
protrusion frequency. The fate of the adhesion components after they turn
over is presently unknown; however, one possibility is that some of this
material is utilized in the formation of new adhesions.
Defects in migration of fibroblasts from mice lacking some tyrosine kinases
and phosphatases have been speculated to arise from impaired turnover of
adhesions (Angers-Loustau et al., 1999; Ilic et al., 1995; Klinghoffer et al.,
1999; Yu et al., 1998). This points to the phosphorylation/dephosphorylation
of their substrates as key regulators of adhesion dynamics. We investigated
this by expressing paxillin-GFP and zyxin-GFP in fibroblasts from mice
lacking FAK. We determined the rate constants for disassembly of adhesion
components by measuring the total fluorescent intensity in specific adhesions
as a function of time. The rate constant for paxillin disassembly decreased
14-fold in FAK null fibroblasts when compared with wild-type fibroblasts. A
similar inhibition of disassembly was also observed with zyxin suggesting that
it is a property of the entire adhesion. Src, another tyrosine kinase, showed a
similar inhibition of adhesion turnover. Co-expression of kinase-deficient Src
with paxillin-GFP decreased the rate constant for disassembly by 18-fold.
Taken together, these results demonstrate that both FAK and Src regulate
adhesion turnover.
Adhesion disassembly at the cell rear
Adhesion disassembly at the cell rear is not a simple reversal of the
mechanisms of formation. Several studies have shown that a fraction of the
integrin can be left behind on the substrate of migrating cells (Palecek et al.,
1996, 1998; Regen and Horwitz, 1992). In CHO cells and fibroblasts, we have
made similar observations using a5 integrin-GFP (Figure 19.2). By contrast,
neither paxillin nor a-actinin is detected in the tracks behind the cell indicating
that the cleavage of the integrin-cytoskeletal linkage occurs at or very close to
the integrin. Paxillin and a-actinin remain in clusters within the cell. These
clusters move along the edge of the cell and are stable for over 30 min, but
they eventually disperse. This contrasts the adhesion disassembly at the cell
front in which paxillin, zyxin and FAK rapidly turn over with similar kinetics.
Interestingly, in FAK null fibroblasts, paxillin, like integrin, is left behind on
