Biology Reference
In-Depth Information
grounds that a positive feedback loop, in which PI-3-kinase activity and generation of
PI(3,4,5)P
3
encourages location of further active PI-3-kinase molecules to the area, might be
an important mechanism in sharpening the gradient.
1
Experimental evidence has shown
this not to be the case, however; cells treated with the PI-3-kinase inhibitor LY294002 still
show normal location of PI-3-kinase protein.
14
The activity of PI-3-kinase is opposed by the phosphatase PTEN, which dephosphory-
lates at the 3 position of PI(3,4,5)P
3
. PTEN shows a distribution complementary to PI-3-
kinase; in cells not treated with cAMP, it is located at the membrane all round the cell
and is active. In cells treated with homogenous cAMP, it leaves the membrane and loses
its activity. In those treated with a gradient of cAMP, it is located and active in all regions
of the membrane except the leading edge. Both the membrane location and activity of
PTEN are regulated by a PI(4,5)P
2
-binding domain.
14,15
The fact that PI-3-kinase removes
PI(4,5)P
2
from its locale, by phosphorylating it to make PI(3,4,5)P
3
, suggests one mechanism
by which PTEN may be excluded from PI-3-kinase rich areas. In the absence of a strong acti-
vation of PI-3-kinase, PTEN-rich areas may maintain themselves by undoing the work of
any PI-3-kinase that does happen to be around, again helping to separate areas of strong
PI(3,4,5)P
3
production, such as the leading edge of the cell, from areas in which it is virtually
absent. However useful these feedback loops are, they cannot be the whole reason for the
complementary distribution of PI3-kinase and PTEN, because blocking the activity of either
enzyme, by mutation in the case of PTEN or with the drug LY294002 in the case of PI-3-
kinase, fails to randomize the distribution of the proteins although their domains may
change a little in size.
14
LINKING THE INTERNAL REPRESENTATION OF THE EXTERNAL
GRADIENT TO MOTILITY
Work in D. discoideum and in other cells suggests a relatively simple link between the
signal transduction system of
Figure 9.3
and the production of a lamellipodium. There is
good evidence, from a large number of mutant and other studies, that D. discoideum assem-
bles its lamellipodium using Rac-mediated activation of actin filament branching, via SCAR
and Arp2/3, that is described in Chapter 8.
16
If homologues of these proteins exist in
D. discoideum, then the complete pathway may be relatively direct as depicted in
Figure 9.6
.
17
As this diagram would predict, mutations that block Rac activity block cell migration, while
mutations that result in hyperactive Rac decouple motility from sensing and cause the
appearance of lamellipodium-like membrane ruffles all round the cell, not just in the direc-
tion of a cAMP gradient.
18
Cells migrating by chemotaxis normally have only one protrusive leading edge, at the
place on the cell's edge that faces directly up the external gradient of cAMP. This implies
that other parts of the cell must be prevented from making a leading edge of their own.
The cell is unlikely to use PI(3,4,5)P
3
to repress protrusive activity elsewhere, because the
very short range of the PI(3,4,5)P
3
gradient that is so useful shaping activity at the leading
edge is useless for signalling across the entire cell. Without a cell-wide signal, though, how
is a point on the side of the cell receiving some cAMP to 'know' that another part of the
cell is receiving more and is already a leading edge?
