We have mutated two regions within the yeast profilin gene in an effor
t to functionally dissect the roles of actin and phosphatidylinositol
4,5-bisphosphate (PIP2) binding in profilin function. A series of trun
cations was carried out at the C terminus of profilin, a region that h
as been implicated in actin binding. Removal of the last three amino a
cids nearly eliminated the ability of profilin to bind polyproline in
vitro but had no dramatic in vivo effects. Thus, the extreme C terminu
s is implicated in polyproline binding, but the physiological relevanc
e of this interaction is called into question. More extensive truncati
on, of up to eight amino acids, had in vivo effects of increasing seve
rity and resulted in changes in conformation and expression level of t
he mutant profilins. However, the ability of these mutants to bind act
in in vitro was not eliminated, suggesting that this region cannot be
solely responsible for actin binding. We also mutagenized a region of
profilin that we hypothesized might be involved in PIP2 binding. Alter
ation of basic amino acids in this region produced mutant profilins th
at functioned well in vivo. Many of these mutants, however, were unabl
e to suppress the loss of adenylate cyclase-associated protein (Cap/Sr
v2p [A. Vojtek, B. Haarer, J. Field, J. Gerst, T. D. Pollard, S. S. Br
own, and M. Wigler, Cell 66:497-505, 1991]), indicating that a defect
could be demonstrated in vivo. In vitro assays demonstrated that the i
nability to suppress loss of Cap/Srv2p correlated with a defect in the
interaction with actin, independently of whether PIP2 binding was red
uced. Since our earlier studies of Acanthamoeba profilins suggested th
e importance of PIP2 binding for suppression, we conclude that both ac
tivities are implicated and that an interplay between PIP2 binding and
actin binding may be important for profilin function.