Cell Motility and the Cytoskeleton 55:147–155 (2003)
Views and Reviews
IQGAPs: Integrators of the Cytoskeleton,
Cell Adhesion Machinery, and Signaling
Networks
Scott C. Mateer,1 Ningning Wang,1 and George S. Bloom1,2*
1
Department of Biology, University of Virginia, Charlottesville
Department of Cell Biology, University of Virginia, Charlottesville
2
Key words: actin; Cdc42; Rac1; calmodulin; CLIP-170; ␤-catenin; E-cadherin; microtubules
INTRODUCTION
Cell motility and morphogenesis are the end products of highly coordinated cellular activities involving
cytoskeletal structures, the machinery for cell adhesion,
and signaling networks. Among the hundreds of proteins
that contribute to cellular movements and shape changes,
only a few lie at the hub of these activities by virtue of
interacting directly with cytoskeletal, cell adhesion, and
signal transduction proteins. One such protein is
IQGAP1, the first known member of a protein family that
has a widespread, but curiously sporadic phylogenetic
distribution. IQGAP1 can bind directly to an impressive
collection of other proteins, including F-actin; the microtubule interacting protein, CLIP-170; the cell adhesion
factor, E-cadherin; ␤-catenin, which couples E-cadherin
to cortical actin networks and also regulates gene expression through the wnt signalling pathway; calmodulin;
and activated forms of the small G proteins, Rac1 and
Cdc42, which are intimately involved in regulating cortical actin networks. IQGAP1 is widely expressed among
mammalian cell types, is most concentrated in cortical
regions rich in actin filaments, and has been implicated in
the control of cellular motility and morphogenesis
through regulation of actin organization, cell-cell adhesion, microtubule stability, and small G proteins. Mammals express at least two different IQGAPs, which are
⬃50% identical in amino acid sequence. The functions of
IQGAP1 at the cellular and organismal levels are in the
process of being unraveled, and very little is known
about other mammalian IQGAPs. In contrast, genetic
approaches using Dictyostelium and yeast have revealed
© 2003 Wiley-Liss, Inc.
specific and essential roles in cytokinesis for IQGAPs
expressed in those organisms. We present this overview
of IQGAPs now in light of growing interest in these
proteins among cell biologists, and the fact that, until
now, no comprehensive reviews dedicated to this protein
family have been published in peer-reviewed journals.
MORPHOGENESIS ⴙ PARTIAL DETACHMENT ⴝ
CELLULAR MOTILITY
The ability of cells to change shape and move
underlies biological processes as diverse as organismal
development, synaptogenesis, and wound healing. Consider cells of the human nervous system, for example.
During embryonic development, spherical neuronal precursor cells undergo stereotypic migrations through the
developing brain and spinal cord. After reaching their
final destinations and completing their terminal mitoses,
The supplemental material referred to in this section can be found at
http://www.interscience.wiley.com/jpages/0886-1544/suppmat/index.
html
Contract grant sponsor: NIH; Contract grant numbers: NS30485,
HD007323.
*Correspondence to: George S. Bloom; Department of Biology; University of Virginia; 229 Gilmer Hall; Charlottesville, VA 22903.
E-mail: gsb4g@virginia.edu
Received 6 February 2003; Accepted 19 February 19 2003
148
Mateer et al.
these morphologically featureless cells abruptly develop
long extensions that become axons and dendrites. The net
result in an adult human may be a neuron with an
⬃100-␮m-wide perikaryon that is attached to a single
axon exceeding a meter in length, and to an elaborate
dendritic tree extending nearly a millimeter away from
the cell body. Paralleling and influencing this motile and
morphogenetic behavior of neurons are similar migrations and shape changes for glial cells. Synaptic connections form continuously through life, and depend on the
morphological plasticity of axon terminals and potential
post-synaptic targets, such as dendritic spines. When
peripheral nerves are injured, the proximal stumps of
affected axons may sprout new growth cones that allow
the axons to elongate and reestablish contacts with their
original targets. Accompanying this neuronal regeneration may be migration into the affected area of new
Schwann cell processes.
The molecular mechanisms for cellular morphogenesis and motility are intimately related. Indeed, it may
be useful to think of motility as resulting from repetitive
cycles of morphogenetic expansion of one side of a cell,
coupled to detachment of the opposite side of the cell
from its surroundings and a subsequent cytoplasmic contraction. The engine for moving cells and contracting the
cytoplasm is the actin cytoskeleton, and many mechanisms for actin-based motility and morphogenesis may
be common to such distinct structures as neuronal growth
cones, glial cell processes, and dendritic spines. Because
growth cones have been studied so much more extensively than the other structures in this context, however,
we will focus on growth cones as examples of how actin
dynamics underlie cell motility and changes in cell morphology.
The morphological and cytoarchitectural features
of growth cone advance have been well characterized
[Letourneau and Ressler, 1983]. The initial event is the
extension of multiple finger-like projections, or filopodia,
along the leading edge of the growth cone. Located
within the core of each filopodium and oriented parallel
to its long axis is a bundle of actin filaments that are
uniformly polarized with their barbed (plus) ends pointing away from the cell. Next, the filopodia that point in
the preferred direction for migration of the growth cone
become stabilized by mechanisms that involve sensing
directionally biased environmental clues. The subsequent
step is the formation of veil-like cellular extensions
known as lamellipodia, that grow toward the distal tip of
the growth cone and thereby fill the gaps between adjacent stabilized filopodia. The principal cytoplasmic component of lamellipodia is a gel of branched actin filaments. The net outcome of this sequence of events is
migration of the growth cone away from the neuronal
perikaryon, and concomitant elongation of the neurite to
which it is attached. The final step in the process is to
stabilize the newly extended neurite against retraction.
This step evidently depends upon the growth of microtubules into the proximal portion of the advancing
growth cone.
REGULATION OF CELLULAR MORPHOGENESIS
AND MOTILITY BY SMALL G PROTEINS
During the past decade, it has become clear that
members of the Rho family of small G proteins, in
particular Cdc42, Rac1, and RhoA, are among the most
important regulatory molecules for growth cone extension in neurons, and similar morphogentic and motile
phenomena in non-neuronal cells. The first evidence for
this was the demonstration that microinjection of quiescent, serum-starved Swiss 3T3 cells with activated (GTPassociated) forms of Rac1 or RhoA led rapidly and
dramatically to the formation of lamellipodia and stress
fibers, respectively [Ridley and Hall, 1992; Ridley et al.,
1992]. Subsequent work showed that activation of Cdc42
similarly induced filopodia [Kozma et al., 1995; Nobes
and Hall, 1995], and furthermore, led downstream to
activation of Rac1 followed by activation of RhoA [Nobes and Hall, 1995].
The realization that the assembly and organization
of actin filaments are profoundly regulated by Cdc42,
Rac1, and RhoA inspired many labs to search for mechanisms that activate these small G proteins and account
for their downstream effects on actin. Constitutively active, oncogenic Ras was found to stimulate Rac1 activation, so yet another low molecular weight GTPase influences actin organization, in this case by acting upstream
of Rac and Rho, but independently of Cdc42 [Ridley and
Hall, 1992]. Further upstream, GTPase activation was
found to result indirectly from binding of extracellular
ligands to cell surface receptors. For example, plateletderived growth factor (PDGF) or insulin act through PI-3
kinase to activate Rac1 [Ridley et al., 1992], lysophosphatidic acid acts through a tyrosine kinase to activate
RhoA [Ridley and Hall, 1992], and bradykinin triggers
activation of Cdc42 by a mechanism that remains poorly
understood [Kozma et al., 1995].
Some events that occur downstream of activated
Rho family GTPases and affect actin have also been
described. For instance, activated Cdc42 or Rac1 regulate the Arp2/3 complex, which stimulates branched assembly of actin filaments. This activity of Arp2/3 is not
constitutive, however, but relies instead on a variety of
either normal cellular proteins or pathologic co-factors.
The normal proteins include those of the WASP and
WAVE families, which must associate with activated
Cdc42 or Rac1 to interact effectively with Arp2/3 and
IQGAPs: Cytoskeleton, Adhesion and Signaling
thus stimulate formation of branched actin filament networks.
The recently gained knowledge about Arp2/3 represents an important foothold on how activated Cdc42
and Rac1 regulate cortical actin, but falls far short of
providing a complete explanation for the formation and
dynamics of filopodia or lamellipodia. For example, interactions between activated Cdc42 and WASP help to
explain how actin gels are formed in lamellipodia, but do
not provide much insight into an earlier effect of Cdc42
activation, namely the induction of actin bundle-containing filopodia. This should not be regarded as surprising in
light of the large number of proteins that have been
localized to filopodia and lamellipodia. Indeed, a recent
review article focused on more than twenty actin-binding, signaling or motor proteins that are concentrated in
lamellipodia, filopodia, or both [Small et al., 2002]. Curiously, Cdc42 and Rac1 were conspicuously missing
from the list, as was IQGAP1, which co-localizes with
actin filaments in lamellipodia [Bashour et al., 1997; Hart
et al., 1996].
IQGAPs
The IQGAP family comprises a limited number of
known proteins, but they are widely distributed phylogenetically. Published IQGAP sequences have been reported for mammals [Brill et al,. 1996; Weissbach et al.,
1994], frogs [Yamashiro et al., 2003], Dictyostelium discoideum [Adachi et al., 1997; Faix and Dittrich, 1996],
budding yeast [Epp and Chant, 1997; Lippincott and Li,
1998; Osman and Cerione, 1998], fission yeast [Eng et
al., 1998], Hydra [Venturelli et al., 2000], and the filamentous fungus, Ashbya gossypii [Wendland and Philippsen, 2002]. The Caenorhabditis elegans and Neurospora crassa genome projects have revealed potential
IQGAPs in those organisms, but their existence at the
protein level has not been confirmed. Curiously, IQGAPs
have not been detected in Drosophila, even at the level of
genomic DNA that potentially could encode such proteins.
As shown in Figure 1, the defining structural traits
of IQGAPs are the sequential arrangement of some or all
of the following functional motifs, each of whose sequences is at least modestly conserved across phylogenetic boundaries. Near the N-terminal of all known
IQGAPs except those in Dictyostelium, is an F-actin
binding calponin homology (CH) domain [Fukata et al.,
1997]. This is followed in vertebrate IQGAPs by multiple IQGAP tandem coiled-coil repeats (IRs) that cause
dimerization of the proteins that contain them [Mateer et
al., 2002]. Next, the vertebrate IQGAPs also contain a
WW domain, a small protein-protein interaction module
whose significance to the IQGAPs is unknown. Further,
toward the C-terminal are two additional types of func-
149
tional domains. The first to be encountered is the one to
four IQ motifs, which interact with calcium-binding EF
hand proteins like calmodulin and are found in every
known IQGAP. Finally, all known IQGAPs contain a
long C-terminal region (ICT) with a minimum of 19%
identity and 37% similarity at the amino acid level to the
C-terminal half of IQGAP1. Within this region, all
IQGAPs except the Saccharomyces cerevisiae protein,
Iqg1p/Cyk1p, contain a GAP-related domain (GRD) that
is at least 25% identical to the catalytic regions of
GTPase activating proteins, or GAPs, for small G proteins of the Ras superfamily. Also found within or near
the ICT of all known IQGAPs is a RasGAP C-terminal
domain (RGCT) with a minimum of 25% identity and
42% similarity among IQGAP family members. As a
group, the IQGAPs have been implicated in cellular
motility and morphogenesis through regulation of actin
dynamics [Bashour et al., 1997; Faix et al., 1998; Kuroda
et al., 1996; Venturelli et al., 2000] and cell adhesion
[Kuroda et al., 1998; Li et al., 1999], and in cytokinesis
[Adachi et al., 1997; Eng et al., 1998; Epp and Chant,
1997; Faix and Dittrich, 1996; Lippincott and Li, 1998;
Osman and Cerione, 1998] regulation of the small G
protein, Cdc42 [Sokol et al., 2001; Swart-Mataraza et al.,
2002], and regulation of the microtubule interacting protein, CLIP-170 [Fukata et al., 2002]. Despite their family
name and structural resemblance to GAPs, however, no
IQGAPs have yet been demonstrated to exhibit GAP
activity.
The published history of the IQGAPs began in
1994 with a report of the cloning and sequencing of
IQGAP1, which was discovered by serendipity using a
PCR-based screen designed to reveal novel matrix metalloproteases [Weissbach et al., 1994]. A similar screen
also led to the cloning of another closely related protein,
which was aptly named IQGAP2 [Brill et al., 1996;
McCallum et al., 1996]. The acronyms by which these
new proteins came to be known reflected the fact that
each contained four contiguous IQ motifs and an identifiable GAP-related region. Efforts to demonstrate RasGAP or even Ras-binding activities for either IQGAP1
[Hart et al., 1996; Weissbach et al., 1994] or IQGAP2
[Brill et al., 1996] have been unsuccessful, although both
proteins clearly can bind calmodulin [Bashour et al.,
1997; Brill et al., 1996; Hart et al., 1996; Joyal et al.,
1997]. A potential third close relative of IQGAP1, tentatively called IQGAP3, is inferred from a human
genomic DNA sequence that appeared very recently in
the Ensembl database, and from several human expressed
sequence tags (ESTs) that were entered into the GenBank
and Ensembl databases. The fact that these ESTs were
prepared from poly-A⫹ RNA is consistent with the notion that IQGAP3 is expressed at both the mRNA and
150
Mateer et al.
Fig. 1. Domain structure of IQGAPs. Domain arrangements of IQGAP
family members expressed in humans (HsIQGAP1 and HsIQGAP2);
Hydra vulgaris (HvIQGAP1); the budding yeast, Saccharomyces cerevisiae (Sclqg1p/Cyk1p); the fission yeast, Schizosaccharomyces pombe
(SpRng2p), the filamentous fungus, Ashbya gossypii (AgCyk1p), and the
slime mold, Dictyostelium discoideum, (DGAP1/RasGAP1 and GAPA)
are shown here. CHD, calponin homology domain, binding to F-actin; IR,
IQGAP repeats, dimerization; WW, protein-protein interaction domains
of unknown significance to IQGAPs; IQ, binding to calmodulin and
related calcium-binding EF hand proteins; ICT, IQGAP C-terminal region; GRD, GAP catalytic-related domain; RGCT, RasGAP-like C-terminal region.
protein levels, but detection of native IQGAP3 protein
has not been reported yet.
The initial published report of IQGAP1 [Weissbach
et al., 1994] seemed unremarkable until it became clear
that a number of other laboratories had discovered the
same protein independently and apparently concurrently
in several different contexts. The significance of the
GAP-related domain became evident when IQGAP1 was
shown to be the major cytosolic protein that bound to
activated, GST-tagged Cdc42 or Rac1 in pull down assays [Hart et al., 1996; Kuroda et al., 1996; McCallum et
al., 1996]. The GAP-related domain of IQGAP1 was
found to be necessary, albeit insufficient, for binding to
the activated small G proteins [Hart et al., 1996].
IQGAP2 was also found to bind Cdc42 and Rac1, but in
contrast to IQGAP1, it did not show a preference for their
activated vs. GDP-associated forms [Brill et al., 1996;
McCallum et al., 1996]. Furthermore, in defiance of the
“GAP” portions of their respective names, both IQGAP1
[Hart et al., 1996] and IQGAP2 [Brill et al., 1996] were
reported to suppress the intrinsic GTPase activities of
Cdc42 and Rac1.
IQGAP1 was also discovered as one of the major
proteins in adrenal cytosol to associate with actin filaments in vitro [Bashour et al., 1997]. Purified adrenal
IQGAP1 was found to be a homodimeric protein that can
bind directly to actin filaments, and cross-link them into
gels and bundles in vitro [Bashour et al., 1997]. In
cultured mammalian cells, IQGAP1 is most concentrated
in the cortex, where it co-localizes with F-actin [Bashour
et al., 1997; Hart et al., 1996; Kuroda et al., 1996], and is
particularly conspicuous in lamellipodia at the leading
edges of motile cells (see Fig. 2 and the corresponding
on-line video). Finally, IQGAP1 was also reported to be
IQGAPs: Cytoskeleton, Adhesion and Signaling
Fig. 2. IQGAP1 is concentrated at the leading edge of motile cells.
NIH-3T3 cells were infected with a modified adenovirus encoding an
IQGAP1-YFP fusion protein, and fluorescence micrographs were
taken at 30-sec intervals. Shown here are three of the images, which
were captured at 0, 13.5, and 27 min, respectively. Note that IQGAP1YFP remained highly concentrated at the leading edge of this motile
cell as it moved ⬃10 ␮m during the time interval shown here. A
QuickTime movie from which these images were obtained is available
on-line.
a major calmodulin binding protein in both normal and
cancerous breast cell lines, and to be the only protein in
such cells to demonstrate appreciable calmodulin binding
activity in the absence of calcium [Joyal et al., 1997]. It
seems striking in retrospect that so many research groups
working on such a diversity of issues encountered
IQGAP1 independently of each other. The collective
151
initial findings of these groups, namely that IQGAP1 is a
major binding partner for activated small G proteins,
actin filaments, and calmodulin, pointed to roles for
IQGAP1 in cellular activities that involve coordination
of signalling networks and the actin cytoskeleton.
Indeed, there is evidence that both calmodulin and
Cdc42 influence interactions between IQGAP1 and actin
filaments. Calmodulin was found to inhibit F-actin binding by IQGAP1 in a concentration-dependent manner
that was potently stimulated by calcium, and IQGAP1YFP was reversibly dissociated from cortical actin networks in live cells by raising intracellular calcium levels
[Mateer et al., 2002]. In contrast, activated Cdc42 was
reported to cause oligomerization of IQGAP1 and thus
increase its F-actin gelation activity [Fukata et al., 1997].
These latter results must be interpreted cautiously, however, because they were obtained using GST-Cdc42, the
dimeric structure of which might have artifactually induced aggregation of IQGAP1.
The evidence that IQGAP1 is an F-actin binding
protein in vivo seems well established, but actin filaments are not the only structural elements of the cell with
which the protein can interact. Several lines of evidence
have implicated IQGAP1 in cell adhesion, and within the
past year, a possible connection between IQGAP1 and
microtubules was revealed as well. Cell-cell adhesion is
typically mediated by homotypic binding of the extracellular domains of transmembrane cadherin molecules
originating from adjacent cells. It has long been known
that the cytoplasmic domain of E-cadherin binds ␤-catenin, which binds ␣-catenin, which in turn binds cortical
actin filaments. This complex of E-cadherin, ␤-catenin,
and ␣-catenin thus couples the actin cytoskeleton to the
molecular machinery for cell-cell adhesion. It is of considerable interest, therefore, that evidence obtained from
both cell biological and in vitro biochemical approaches
indicates that IQGAP1 can bind directly to E-cadherin
and ␤-catenin, and when doing so, disrupts interactions
between ␣-catenin and E-cadherin/␤-catenin [Kuroda et
al., 1998; Li et al., 1999]. In light of this evidence, it is
not surprising that overexpression of IQGAP1 in cultured
cells has been reported to cause decreased cell-cell adhesion [Kuroda et al., 1998].
These findings may be directly relevant to gastric
carcinogenesis. The gene for IQGAP1 is located within a
region of human chromosome 15 that is amplified in
many gastric carcinomas [Pujana et al., 2001] and gastric
tumor cell lines [Sugimoto et al., 2001]. Moreover,
IQGAP1 protein is overexpressed in both the cell lines
and the tumors [Sugimoto et al., 2001], and immunohistochemistry has demonstrated that in the tumors it is
especially abundant in invasion fronts of metastatic cells
[Nabeshima et al., 2002]. Taken together, these data
imply that overabundant IQGAP1 compromises cell ad-
152
Mateer et al.
Fig. 3. Regulation of IQGAP1. Biochemical and
cell biological effects of the binding to IQGAP1 of
calcium/calmodulin, and activated Cdc42 and Rac1
are summarized.
hesion in the gastric epithelium, and thus contributes to
the oncogenic phenotype of gastric carcinoma cells. It is,
therefore, curious that the only reported phenotype in a
strain of IQGAP1 knockout mice was gastric hyperplasia
[Li et al., 2000], but whatever the explanation may be, it
seems clear that IQGAP1 can have profound effects on
cell adhesion.
Provocative evidence for a connection between
IQGAP1 and microtubules was published in mid 2002,
based on the finding that C-terminal fragments of
IQGAP1 bound to CLIP-170, a protein that accumulates
at microtubule plus ends in cells [Fukata et al., 2002].
Intracellular expression of such IQGAP1 fragments disrupted the CLIP-170 association with microtubules, and
eventually led to partial disassembly of microtubules
from their plus ends. Additional evidence implied that
activated Rac1 or Cdc42, acting through IQGAP1, promoted capture of CLIP-170-capped microtubules in lamellipodia or filopodia, respectively. This might explain,
at least in part, the selective stabilization of microtubules
that project to sites of plasma membrane stimulation by
hormones. For example, one could imagine how a cell
encountering a gradient of PDGF would selectively activate Rac1 in the cell cortex near the site of maximal
PDGF concentration, which would then allow binding of
the GTP-associated small G protein to IQGAP1, followed by capture and stabilization of CLIP-170 capped
microtubules by IQGAP1.
The Rac1/Cdc42 chapter of the CLIP-170 story
serves as an example of how the binding of one reversibly associating ligand for IQGAP1 can influence the
binding properties of another. In this case, the effect
seems positive: binding of activated Rac1 or Cdc42 to
IQGAP1 reportedly promoted the association of CLIP170 with IQGAP1. In the other cases that have been
studied, however, competitive binding has emerged as
the more general rule (Fig. 3). As mentioned earlier,
calcium promotes binding of calmodulin to IQGAP1,
which inhibits binding of IQGAP1 to F-actin [Mateer et
al., 2002]. Calcium/calmodulin has also been reported to
downregulate binding of activated Cdc42 to IQGAP1
[Ho et al., 1999; Joyal et al., 1997], and to compete with
E-cadherin for binding to IQGAP1 [Li et al., 1999].
Similarly, activated Cdc42 or Rac1 [Fukata et al., 1999],
and calmodulin were reported to inhibit binding of
IQGAP1 to ␤-catenin. The fact that binding of calcium/
calmodulin to IQGAP1 inhibits the latter’s association
with most of its binding partners suggests that calcium,
acting through calmodulin, acts as a universal “off” signal for IQGAP1 activity. The formal possibility that the
calcium/calmodulin/IQGAP1 complex targets an unidentified effector molecule, however, emphasizes the need to
define in more detail the hierarchy of interactions between IQGAP1 and its numerous ligands.
One of the most intriguing aspects of IQGAP1 is
that it can bind indiscriminately to F-actin in vitro,
whereas in intact cells it co-localizes with actin filaments
in the cortex, but not in stress fibers [Bashour et al.,
1997]. This suggests that IQGAP1 is specifically targeted
to or near the cytoplasmic face of the plasma membrane,
but begs the question of why IQGAP1 should be targeted
to the cell cortex in the first place. A potential answer
IQGAPs: Cytoskeleton, Adhesion and Signaling
with far-reaching implications was recently reported
[Swart-Mataraza et al., 2002]. Overexpression of wildtype IQGAP1 in tissue culture cells was found to cause
an increase in cellular levels of activated Cdc42, and to
induce formation of filopodia. In contrast, overexpression of an IQGAP1 mutant lacking the GAP-related
region led to decreased levels of activated Cdc42, and a
failure of the cells to respond to bradykinin or constitutively active Cdc42, or to translocate activated Cdc42 to
membranes. These results raise the possibility that
IQGAP1 is a cortical “docking station” for activated
Cdc42, which evidently must be associated with the
plasma membrane in order to exert its effects on the
cytoskeleton and the machinery for cell-cell adhesion.
They also stress the importance of learning how IQGAP1
is targeted to the cell surface. There is evidence that
E-cadherin expression is required for this targeting [Kuroda et al., 1998; Li et al., 1999], but the minimal
targeting region on IQGAP1 has not been defined yet.
IQGAPs and Cytokinesis
Several microorganisms have been found to express IQGAPs that are intimately involved in cytokinesis.
The best understood example, by far, is the Saccharomyces cerevisiae protein, Iqg1p [Epp and Chant, 1997;
Osman and Cerione, 1998], which is also known as
Cyk1p [Lippincott and Li, 1998]. It is found at the
cytokinetic ring in fission yeast and is required for cytokinesis. Iqg1p appears to be necessary for recruitment to
the cytokinetic ring of other proteins, such as actin [Lippincott and Li, 1998] and a collection of factors that have
been proposed to serve as “landmark proteins” for coordinating axial budding and cytokinesis [Osman et al.,
2002]. Included in this collection are the GTP-binding
protein, Bud4p, the septin, Cdc12p, and the secretion site
determinant, Sec3p [Osman et al., 2002]. There is evidence that Mlc1p, a yeast myosin light chain, binds
directly to Iqg1p and recruits the latter to the nascent
cytokinetic ring [Shannon and Li, 2000], and that Iqg1p
is necessary for recruitment of Cdc42p to that region
[Osman and Cerione, 1998]. Curiously, Iqg1p has been
shown to interact with Cdc42p in a two-hybrid screen
[Osman and Cerione, 1998], but it does not contain a
readily recognizable equivalent of the GAP-related domain that is necessary, albeit insufficient [Hart et al.,
1996], for binding of activated Cdc42 to mammalian
IQGAP1. A protein that appears to be a homologue of
Iqg1p has been detected in the filamentous fungus, Ashbya gossypii [Wendland and Philippsen, 2002], and another potential homologue has been revealed by the Neurospora (N. crassa) Genome Sequencing Project.
The fission yeast, Schizosaccharomyces pombe,
also contains an IQGAP that is required for cytokinesis.
This protein, known as Rng2p, is localized to the cyto-
153
kinetic ring in S. pombe [Eng et al., 1998], and like
Iqg1p, binds a myosin light chain [D’Souza et al., 2001].
Finally, the slime mold, Dictyostelium discoideum, has
been found to express a pair of IQGAPs that are involved
in cytokinesis: GAPA [Adachi et al., 1997] and DGAP1
[Faix and Dittrich, 1996]. The two Dictyostelium proteins are ⬃50% identical in primary sequence, but apparently have non-identical functions. GAPA appears to
be required for the completion of cytokinesis [Adachi et
al., 1997]. In contrast, there are conflicting reports about
the necessity of DGAP1 for cytokinesis [Faix and Dittrich, 1996; Faix et al., 2001], but its elimination is
acknowledged to cause developmental defects during
starvation-induced cellular aggregation [Faix and Dittrich, 1996; Faix et al., 2001]. Although the exact role of
DGAP1 in cytokinesis remains to be determined, activated Rac1 has been reported to promote its interaction
with cortexillins I and II, which are targeted to the
cleavage furrow in dividing cells [Faix et al., 2001].
So far, vertebrates have not been found to contain
IQGAPs that are involved in cytokinesis. The possibility
that IQGAP1 or IQGAP2 serves a role in this phase of
cell division cannot be strictly dismissed yet. Nevertheless, the absence of both proteins from cleavage furrows
(unpublished observations) indicates that unlike the yeast
proteins, Iqg1p and Rng1p, neither mammalian protein
recruits other proteins to the constriction site that ultimately separates daughter cells at the end of mitosis. It is
possible that mammalian homologues of cytokinesis
IQGAPs exist, but they have not been reported at the
protein level, and predicted homologues have not been
revealed by analysis of the human genome.
Unsolved Mysteries of Vertebrate IQGAPs
Given the unlikelihood that vertebrate IQGAPs are
involved in cytokinesis, their roles in interphase cells
warrant much further attention. What initially captured
our fancy about IQGAP1 was the possibility that it represents a major “missing link” in the molecular machinery that couples activation of Cdc42 and Rac1 to reorganization of cortical actin networks. The basis for this
suggestion was straightforward: IQGAP1 binds directly
in vitro to F-actin and the activated forms of both small
G proteins, and co-localizes in cells with cortical actin
filaments. Although nothing that has been learned so far
about IQGAP1 would justify discarding the missing link
hypothesis, neither do the available data provide a satisfactory explanation to verify the theory. Even if verification can be obtained, it will be interesting to learn
whether the relevant data point to connections among
IQGAP1’s abilities to interact with actin filaments, microtubules, and the machinery for cell-cell adhesion. For
instance, can IQGAP1 bind simultaneously to actin filaments, CLIP-170 and ␤-catenin, or are at least some of
154
Mateer et al.
these interactions mutually exclusive? Another issue that
remains completely unresolved is the degree, if any, to
which IQGAP1, IQGAP2, and IQGAP3 (assuming it
exists) are functionally redundant. Solving that problem
will require knowledge of the cellular and subcellular
distributions of these proteins, manipulations of the proteins in live cells, comparative in vitro biochemical studies, and perhaps transgenic animal models as well. In
light of IQGAP1’s apparent position at the interface of
the cytoskeleton, cell adhesion, and signalling, such efforts will be amply justified.
REFERENCES
Adachi H, Takahashi Y, Hasebe T, Shirouzu M, Yokoyama S, Sutoh
K. 1997. Dictyostelium IQGAP-related protein specifically involved in the completion of cytokinesis. J Cell Biol 137:891–
898.
Bashour A-M, Fullerton AT, Hart MJ, Bloom GS. 1997. IQGAP1, a
Rac- and Cdc42-binding protein, directly binds and cross-inks
microfilaments. J Cell Biol 137:1555–1566.
Brill S, Li S, Lyman CW, Church DM, Wasmuth JJ, Weissbach L,
Bernards A, Snijders A. 1996. The Ras GTPase-activatingprotein-related human protein IQGAP2 harbors a potential actin binding domain and interacts with calmodulin and Rho
family GTPases. Mol Cell Biol 16:4869 – 4878.
D’Souza VM, Naqvi NI, Wang H, Balasubramanian MK. 2001. Interactions of Cdc4p, a myosin light chain, with IQ-domain containing proteins in Schizosaccharomyces pombe. Cell Struct
Funct 26:555–565.
Eng K, Naqvi NI, Wong KC, Balasubramanian MK. 1998. Rng2p, a
protein required for cytokinesis in fission yeast, is a component
of the actomyosin ring and the spindle pole body. Curr Biol
8:611– 621.
Epp JA, Chant J. 1997. An IQGAP-related protein controls actin-ring
formation and cytokinesis in yeast. Curr Biol 7:921–929.
Faix J, Dittrich W. 1996. DGAP1, a homologue of rasGTPase
activating proteins that controls growth, cytokinesis, and
development in Dictyostelium discoideum. FEBS Lett 394:
251–257.
Faix J, Clougherty C, Konzok A, Mintert U, Murphy J, Albrecht R,
Muhlbauer B, Kuhlmann J. 1998. The IQGAP-related protein
DGAP1 interacts with Rac and is involved in the modulation of
the F-actin cytoskeleton and control of cell motility. J Cell Sci
111:3059 –3071.
Faix J, Weber I, Mintert U, Kohler J, Lottspeich F, Marriott G. 2001.
Recruitment of cortexillin into the cleavage furrow is controlled
by Rac1 and IQGAP-related proteins. Embo J 20:3705–3715.
Fukata M, Kuroda S, Fujii K, Nakamura T, Shoji I, Matsuura Y,
Okawa K, Iwamatsu A, Kikuchi A, Kaibuchi K. 1997. Regulation of cross-linking activity of actin filament by IQGAP1, a
target for Cdc42. J Biol Chem 272:29579 –29583.
Fukata M, Kuroda S, Nakagawa M, Kawajiri A, Itoh N, Shoji I,
Matsuura Y, Yonehara S, Fujisawa H, Kikuchi A and others.
1999. Cdc42 and Rac1 regulate the interaction of IQGAP1 with
beta-catenin. J Biol Chem 274:26044 –26050.
Fukata M, Watanabe T, Noritake J, Nakagawa M, Yamaga M, Kuroda
S, Matsuura Y, Iwamatsu A, Perez F, Kaibuchi K. 2002. Rac1
and Cdc42 Capture Microtubules through IQGAP1 and CLIP170. Cell 109:873– 885.
Hart MJ, Callow MG, Souza B, Polakis P. 1996. IQGAP1, a calmodulin-binding protein with a RasGAP-related domain, is a potential effector for Cdc42Hs. EMBO J 15:2997–3005.
Ho YD, Joyal JL, Li Z, Sacks DB. 1999. IQGAP1 integrates Ca2⫹/
calmodulin and Cdc42 signaling. J Biol Chem 274:464 – 470.
Joyal JL, Annan RS, Ho Y-D, Huddleston ME, Carr SA, Hart MJ,
Sacks DB. 1997. Calmodulin modulates the interaction between IQGAP1 and Cdc42. J Biol Chem 272:15419 –15425.
Kozma R, Ahmed S, Best A, Lim L. 1995. The Ras-related protein
Cdc42Hs and bradykinin promote formation of peripheral actin
microspikes and filopodia in Swiss 3T3 fibroblasts. Mol Cell
Biol 15:1942–1952.
Kuroda S, Fukata M, Kobayashi K, Nakafuku M, Nomura N,
Iwamatsu A, Kaibuchi K. 1996. Identification of IQGAP as a
putative target for the small GTPases, Cdc42 and Rac1. J Biol
Chem 271:23363–23367.
Kuroda S, Fukata M, Nakagawa M, Fujii K, Nakamura T, Ookubo T,
Izawa I, Nagase T, Nomura N, Tani H, et al. 1998. Role of
IQGAP1, a target of the small GTPases Cdc42 and Rac1, in
regulation of E-cadherin- mediated cell-cell adhesion. Science
281:832– 835.
Letourneau PC, Ressler AH. 1983. Differences in the organization of
actin in the growth cone compared with the neurites of cultured
neurons from chick embryos. J Cell Biol 97:963–973.
Li S, Wang Q, Chakladar A, Bronson RT, Bernards A. 2000. Gastric
hyperplasia in mice lacking the putative Cdc42 effector
IQGAP1. Mol Cell Biol 20:697–701.
Li Z, Kim SH, Higgins JM, Brenner MB, Sacks DB. 1999. IQGAP1
and calmodulin modulate E-cadherin function. J Biol Chem
274:37885–37892.
Lippincott J, Li R. 1998. Sequential assembly of myosin II, an
IQGAP-like protein, and filamentous actin to a ring structure
involved in budding yeast cytokinesis. J Cell Biol 140:355–
366.
Mateer SC, McDaniel AE, Nicolas V, Habermacher GM, Lin M-JS,
Cromer DA, King ME, Bloom GS. 2002. The mechanism for
regulation of the F-actin binding activity of IQGAP1 by calcium/calmodulin. J Biol Chem 277:12324 –12333.
McCallum SJ, Wu WJ, Cerione RA. 1996. Identification of a putative
effector of Cdc42Hs with high sequence similarity to the RasGAP-related protein IQGAP1 and a Cdc42Hs binding partner
with similarity to IQGAP2. J Biol Chem 271:21732–21737.
Nabeshima K, Shimao Y, Inoue T, Koono M. 2002. Immunohistochemical analysis of IQGAP1 expression in human colorectal
carcinomas: its overexpression in carcinomas and association
with invasion fronts. Cancer Lett 176:101–109.
Nobes CD, Hall A. 1995. Rho, Rac and Cdc42 GTPases regulate the
assembly of multimolecular focal complexes associated with
actin stress fibers, lamellipodia and filopodia. Cell 81:53– 62.
Osman MA, Cerione RA. 1998. Iqg1p, a yeast homologue of the
mammalian IQGAPs, mediates Cdc42p effects on the actin
cytoskeleton. J Cell Biol 142:443– 455.
Osman MA, Konopka JB, Cerione RA. 2002. Iqg1p links spatial and
secretion landmarks to polarity and cytokinesis. J Cell Biol
159:601– 611.
Pujana MA, Nadal M, Gratacos M, Peral B, Csiszar K, GonzalezSarmiento R, Sumoy L, Estivill X. 2001. Additional complexity
on human chromosome 15q: identification of a set of newly
recognized duplicons (LCR15) on 15q11-q13, 15q24, and
15q26. Genome Res 11:98 –111.
Ridley AJ, Hall A. 1992. The small GTP-binding protein rho regulates
the assembly of focal adhesions and actin stress fibers in
response to growth factors. Cell 70:389 –399.
IQGAPs: Cytoskeleton, Adhesion and Signaling
Ridley AJ, Paterson HF, Johnston CL, Diekman D, Hall A. 1992. The
small GTP-binding protein rac regulates growth factor-induced
membrane ruffling. Cell 70:401– 410.
Shannon KB, Li R. 2000. A myosin light chain mediates the localization of the budding yeast IQGAP-like protein during contractile
ring formation. Curr Biol 10:727–730.
Small JV, Stradal T, Vignal E, Rottner K. 2002. The lamellipodium:
where motility begins. Trends Cell Biol 12:112–120.
Sokol SY, Li Z, Sacks DB. 2001. The effect of IQGAP1 on Xenopus
embryonic ectoderm requires Cdc42. J Biol Chem 276:48425–
48430.
Sugimoto N, Imoto I, Fukuda Y, Kurihara N, Kuroda S, Tanigami A,
Kaibuchi K, Kamiyama R, Inazawa J. 2001. IQGAP1, a negative regulator of cell-cell adhesion, is upregulated by gene
amplification at 15q26 in gastric cancer cell lines HSC39 and
40A. J Hum Genet 46:21–25.
155
Swart-Mataraza JM, Li Z, Sacks DB. 2002. IQGAP1 is a component
of Cdc42 signaling to the cytoskeleton. J Biol Chem 277:
24753–24763.
Venturelli CR, Kuznetsov S, Salgado LM, Bosch TC. 2000. An
IQGAP-related gene is activated during tentacle formation in
the simple metazoan Hydra. Dev Genes Evol 210:458 – 463.
Weissbach L, Settleman J, Kalady MF, Snijders AJ, Murthy AE, Yan
Y-X, Bernards A. 1994. Identification of a human rasGAPrelated protein containing calmodulin-binding domains. J Biol
Chem 269:20517–20521.
Wendland J, Philippsen P. 2002. An IQGAP-related protein, encoded
by AgCYK1, is required for septation in the filamentous fungus
Ashbya gossypii. Fungal Genet Biol 37:81– 88.
Yamashiro S, Noguchi T, Mabuchi I. 2003. Localization of two
IQGAPs in cultured cells and early embryos of Xenopus laevis.
Cell Motil Cytoskeleton 55:36 –50.