Finally, de novo actin polymerization from a 'nucleus' of three associated
actin monomers could provide new filament templates (Pollard et al., 2000).
The formation of actin trimers is kinetically unfavourable though, and would
need a factor to accelerate and stabilize it. For many years nucleating factors
were postulated but never found. However, in 1994 the discovery of a
multiprotein complex, the Arp2/3 complex, and subsequent finding that this
complex promoted actin-based motility of Listeria monocytogenes bacteria,
has had a profound effect on the understanding of de novo actin
polymerization (Machesky et al., 1994; Welch et al., 1997a). Indeed it is
now apparent that the Arp2/3 complex is a major 'nucleator' of new actin
The Arp2/3 complex
Together, the actin-related proteins Arp2 and Arp3 and five unique proteins
p41-Arc/ARPC1 ( Ar p c omplex/accepted human genome nomenclature ¼
ARPC), p34-Arc/ARPC2, p21-Arc/ARPC3, p20-Arc/ARPC4 and p16-Arc/
ARPC5 form the Arp2/3 complex. Each subunit has homologues throughout
eukaryotes, including Drosophila melanogaster, Caenorhabditis elegans,
Acanthamoeba castellanii, Saccharomyces cerevisiae and Homo sapiens. The
structures of Arp2 and Arp3 greatly resemble actin, with the exception that their
central clefts are spread open approximately 158 further and that some of their
surface loops differ (Robinson et al., 2001). The other five subunits are novel
and do not share any significant homology with previously identified proteins.
However, ARPC1 contains seven WD (tryptophan and aspartate) repeat motifs
that fold to give a b-propeller conformation characteristic of protein-protein
interaction domains (Robinson et al., 2001; Westerberg et al., 2001).
In mammals, there appear to be isoforms of several of the Arp2/3 complex
subunits. Humans express two genes for ARPC1, p41-Arc and SOP2Hs.
SOP2Hs functionally complements loss of the 41-kDa subunit (SOP2), in the
fission yeast Schizosaccharomyces pombe as does p41-Arc (Balasubramanian
et al., 1996). Recent research also indicates the expression of two forms of
Arp3 (Arp3 and Arp3b) and ARPC5 (p16A and p16B) (Jay et al., 2000;
Millard et al., 2003). As yet the functional relevance and tissue expression
pattern of most of these different isoforms is unknown, but some examples are
emerging. In mouse, p16A and p16B show different expression patterns, with
p16A being highly abundant in haematopoietic cells and p16B being more
abundant in brain (Millard et al., 2003). It also appears based on peptide
sequences, that the Arp2/3 complex from human platelets contains p41-Arc
but not SOP2Hs and p16A but not p16B (Machesky et al., 1997).
The first evidence that the Arp2/3 complex was able to stimulate de novo
actin polymerization in vivo came in 1997 through studies with the motile