rho gtpase control of creb and animal size

Developmental Cell, Vol. 2, 1–20, May, 2002, Copyright 2002 by Cell Press

Modulation of CREB Activity by the Rho GTPaseRegulates Cell and Organism Sizeduring Mouse Embryonic Development

GTPases elicit their biological actions through directinteractions with a variety of so-called effector targets.

The biological function of the RhoGAPs is somewhatless clear. These negative regulators of Rho GTPaseshave been postulated to play a role in precisely modulat-

Raffaella Sordella,1,4 Marie Classon,1,4

Kang-Quan Hu,1 Stephen F. Matheson,1

Madeleine R. Brouns,1 Barry Fine,1 Le Zhang,1

Hiroya Takami,2 Yoshihiko Yamada,2

and Jeffrey Settleman1,3

1MGH Cancer Center and ing Rho-mediated signaling in vivo, and there are severalreports indicating that these proteins, when overex-Harvard Medical School

Charlestown, Massachusetts 02129 pressed, can inhibit GTPase-mediated signaling path-ways (Symons and Settleman, 2000). However, there2 National Institute of Dental

and Craniofacial Research has been relatively little analysis of these proteins in atruly in vivo context. Among the various RhoGAPs thatNational Institutes of Health

Bethesda, Maryland 20892 have been identified is a family of widely expressedpotent Rho-specific GAPs, referred to as the p190 Rho-GAPs. These GAPs are unusual in that they containan amino-terminal GTPase domain in addition to their

Summarycarboxy-terminal RhoGAP catalytic domain (Burbelo etal., 1995; Settleman et al., 1992). In mammals, there are

Rho GTPases regulate several aspects of tissue mor-two p190 RhoGAPs, referred to as p190 RhoGAP and

phogenesis during animal development. We found thatp190-B RhoGAP. Closely related orthologs of these pro-

mice lacking the Rho-inhibitory protein, p190-B Rho-teins are present in Drosophila and C. elegans as well.

GAP, are 30% reduced in size and exhibit develop-Recently, RNA interference studies have revealed a role

mental defects strikingly similar to those seen in micefor the Drosophila p190 RhoGAP in neurite retraction in

lacking the CREB transcription factor. In p190-B Rho-the developing nervous system (Billuart et al., 2001),

GAP-deficient mice, CREB phosphorylation is sub-although a loss-of-function mutant of this gene has yet

stantially reduced in embryonic tissues. Embryo-to be reported. As Rho activation has previously been

derived cells contain abnormally high levels of activedemonstrated to promote neurite retraction in cultured

Rho protein, are reduced in size, and exhibit defects inneurons (Jalink et al., 1994), this result suggests that

CREB activation upon exposure to insulin or IGF-1.loss of p190 RhoGAP activity in vivo leads to deregu-

The cell size defect is rescued by expression of consti-lated Rho GTPase function. Disruption of the p190 Rho-

tutively activated CREB, and in wild-type cells, expres-GAP gene in mice by gene targeting revealed that it is

sion of activated Rho or dominant-negative CREB re-an essential protein that mediates an adhesion signaling

sults in reduced cell size. Together, these resultspathway downstream of Src kinases that functions in

suggest that activity of the Rho GTPase modulates aneural development (Brouns et al., 2000, 2001).

signal from insulin/IGFs to CREB that determines cellAlthough the biological function of p190-B RhoGAP

size and animal size during embryogenesis.is not clear, it has been reported that this RhoGAP ac-counts for the majority of Rho-inhibitory activity in cul-tured fibroblasts, suggesting that it is a major down-Introductionmodulator of Rho GTPase activity in vivo (Vincent andSettleman, 1999). To determine the biological require-The Rho GTPases mediate signaling pathways that reg-

ulate a variety of cellular activities, including cytoskeletal ment for p190-B RhoGAP, we have disrupted the genein mice. Surprisingly, we found that these mice exhibitreorganization, gene expression, and cell cycle progres-

sion (Van Aelst and D’Souza-Schorey, 1997). They are phenotypes that are remarkably similar to those seenin mice lacking the CREB transcription factor, includingalso essential regulators of tissue morphogenesis during

embryonic development (Settleman, 1999). Like all a substantial reduction in animal size (Rudolph et al.,1998). Additional studies revealed that cell size is re-GTPases, the Rho proteins cycle between active, GTP-

bound and inactive, GDP-bound states. This cycling is duced in these animals and that Rho and CREB areimportant novel regulators of cell size in a developinglargely regulated by two classes of proteins, the guanine

nucleotide exchange factors (GEFs) and the GTPase mouse. The results also suggest a cytoskeleton-inde-pendent role for Rho GTPase signaling in embryonicactivating proteins (GAPs; Symons and Settleman,

2000). The RhoGEFs function to promote the conversion development.of inactive Rho to active Rho, while the RhoGAPs stimu-late the intrinsic GTP hydrolysis activity of Rho proteins, Resultsthereby converting them to a GDP-bound, inactive form.Accumulating evidence suggests that the RhoGEFs are Mice Lacking p190-B RhoGAP Closely Resemblerapidly activated upon ligand-receptor interactions at Mice Lacking CREBthe cell surface, and that the subsequently activated To explore p190-B RhoGAP function in vivo, we dis-

rupted the mouse gene using homologous recombina-tion in embryonic stem (ES) cells (Figure 1A). Two lines3 Correspondence: [email protected]

4 These authors contributed equally to this work. of properly targeted ES cells were used to establish

Developmental Cell2

Figure 1. Mice Lacking p190-B RhoGAP Resemble CREB-Deficient Mice

(A) Targeting strategy for disruption of the p190-B RhoGAP gene. Schematic representations of the 5� genomic region of the p190-B RhoGAPgene (upper line), the targeting vector (middle line), and the mutated allele (lower line) (gray bar depicts exon I). Bar under the upper linerepresents the probe used for the identification of heterozygous ES lines. S, StuI; H, HindIII; B, BamHI; N, NotI; NEO, neomycine resistancegene; TK, thymidine kinase gene.(B) An immunoblot of embryonic tissue from wild-type (�/�), heterozygous mutant (�/�), and homozygous mutant (�/�) tissue probed withantisera that specifically recognize either p190-B RhoGAP (top panel) or p190 RhoGAP (bottom panel).

Rho and CREB Regulate Embryonic Cell Size3

germline transmission of the targeted allele in mice, and significantly increased in the mutants (Figure 2A). For-ward light scatter FACS analysis, which indirectly mea-the resulting homozygous mutant mice fail to express

any detectable p190-B RhoGAP (Figure 1B). Homozy- sures changes in cell volume, revealed that this differ-ence reflects an approximately 30% reduction in cellgous mutant mice die immediately after birth, and,

grossly, they always exhibit a substantial and uniform size in the mutants (Figures 2B and 5D). Thus, they packmore densely in a monolayer. This result was indepen-reduction in size (30% reduction in mass) relative to

their littermates (more than 100 examined; Figure 1C). dently confirmed by comparing protein:cell number andprotein:DNA ratios between wild-type and mutant cellsHistological analysis revealed several completely pene-

trant (30/30 examined) defects in the mutant mice. At (Figures 2C and 2D), as well as by morphometric analysis(not shown). Changes in cell size observed by these firstbirth, they exhibit a severe reduction in thymus size

(disproportionate to the reduction in body size; Figure three criteria would not be influenced by changes in thecytoskeleton, indicating that the reduced cell size in1D), their lungs fail to inflate properly (not shown), and

in brain, they exhibit axon defects that include a severe mutants is not due to Rho-induced effects on actin as-sembly.reduction in the major midline forebrain crossing tracts

(corpus callosum and anterior commissure), as well as The small size of mutant animals could potentiallyreflect this cell size phenotype seen in 3T3 lines, ora thinner cortex, associated with partial loss of cortical

axon tracts (Figures 1E and 1F). The lateral ventricles it could be due to either reduced cell proliferation orincreased apoptosis in developing tissues. Indeed,in brain are substantially enlarged in mutants (Figures

1E and 1F). Aside from a small decrease in overall size, CREB has previously been implicated as a cell survivalfactor (Walton and Dragunow, 2000). However, in mutantall other major organs appear to be unaffected in the

mutants. embryonic tissues (E12–E18) analyzed at several levelsof sectioning, we were not able to detect any significantThe observed spectrum of defects in mutant animals

strikingly resembles defects previously reported in mice increase in apoptosis by TUNEL staining (data notshown). Moreover, no evidence of reduced cell prolifera-lacking the CREB transcription factor (Rudolph et al.,

1998). CREB-deficient mice also die immediately after tion was seen by BrdU incorporation into embryonictissues (Figures 2E and 2F). Then, by comparing pro-birth, are 30% reduced in size, fail to inflate their lungs,

have a small thymus, and exhibit brain defects that affect tein:DNA ratios, we found that in most tissues there isa significant reduction in cell size in animals lackingthe corpus callosum, anterior commissure, and lateral

ventricles (Rudolph et al., 1998). CREB, like its closely p190-B RhoGAP (Figure 2G). Together with the observa-tion that mutant 3T3 cells are reduced in size, theserelated family members CREM and the ATFs, undergoes

phosphorylation of serine 133 when activated (Shaywitz results suggest that the small size of the mutant micereflects a cell-autonomous defect in cell size regulation.and Greenberg, 1999). Significantly, we determined that

while CREB protein levels are unaffected in mutant tis- Significantly, p190-B RhoGAP is expressed in all tissuesduring embryonic development (Figure 2H).sues, phospho-CREB and phospho-ATF-1 are virtually

absent in mutant brain, are substantially reduced in thy- To determine whether the cell size defect that resultsfrom loss of p190-B RhoGAP is due to impaired CREBmus, and are also reduced in stomach and heart (Fig-

ure 1G). activation, we generated p190-B RhoGAP-deficient 3T3lines that stably express a constitutively activated formof CREB (CREBFY; Du et al., 2000). These cells are

Loss of p190-B RhoGAP Causes a Cell Size significantly larger than the parental p190-B RhoGAP-Reduction that Is Rescued by Activated CREB deficient lines, indicating that CREB activation is suffi-To determine the role of p190-B RhoGAP in CREB regu- cient to rescue the cell size defect in mutants, and thatlation, we established stable fibroblast lines from wild- CREB is probably functioning downstream of p190-Btype and mutant embryos. All results were confirmed RhoGAP (Figures 3A–3C). In addition, we determinedwith three independently derived litter-matched wild- that CREBFY rescues the cell size defect in both G1-type and mutant pairs of spontaneously immortalized and G2-selected cell populations, indicating a cell cycle-fibroblast (3T3) lines. Morphologically, wild-type and independent role for CREB in cell size regulation (Fig-mutant cells appear similar, although mutant cells, as ures 3D–3G).would be expected as a consequence of losing a Rho-inhibitory protein, contain significantly more polymer-ized actin (data not shown). By comparing the growth Rho and CREB Activity Modulate Cell Size

To address the mechanism by which loss of p190-Bproperties of these cells, we determined that while theyexhibit equivalent doubling times, saturation density is RhoGAP affects CREB activation, we considered the

(C) Newborn wild-type (�/�) and homozygous mutant (�/�) animals illustrating the size defect seen in animals lacking p190-B RhoGAP.(D) Thymuses dissected from newborn wild-type (�/�) and homozygous mutant (�/�) animals revealing the substantial reduction in the sizeof mutant thymus.(E and F) Coronal forebrain sections from wild-type (E) and mutant (F) embryos at E18.5 stained with Nissl reveals the reduction in the thicknessof the corpus callosum (black arrows) in the mutant as well as the enlargement of lateral ventricles in these animals. Note that the thicknessof the cortex is also reduced in the mutants, and this appears to reflect a reduction of the major forebrain cortical axon tracts (white arrows).(G) Immunoblots of various tissues from wild-type (�/�) and mutant (�/�) E18.5 animals with antibodies that specifically recognize thephosphorylated form of CREB and ATF-1 (upper panels) or total CREB protein (lower panels).

Developmental Cell4

Figure 2. Cell Size Is Reduced in Fibroblasts and Tissues from Mice Lacking p190-B RhoGAP

(A) Proliferation curve of wild-type (�/�) and p190-B RhoGAP mutant (�/�) 3T3 cells, indicating similar proliferative rates in the two populationsand increased saturation density of the mutant cells.(B) Cell size distribution of wild-type (�/�) and p190-B RhoGAP mutant (�/�) 3T3 cells determined by forward light scatter FACS analysis,indicating the mean reduction in cell size in a population of G1-selected (2N content) mutant cells.(C) Ratio of total protein to cell number in a population of confluent wild-type (�/�) and p190-B RhoGAP mutant (�/�) cells confirming thereduced cell size of mutant cells.(D) Ratio of total protein to total DNA content in a population of confluent wild-type (�/�) and p190-B RhoGAP mutant cells, again confirmingthe reduced cell size of mutant cells. Values in (C) and (D) reflect the average of three independent experiments, and error bars representstandard error of the mean.(E and F) BrdU incorporation of intestinal sections of wild-type (E) and p190-B RhoGAP mutant (F) E18.5 embryos (following 4 hr intrauterineBrdU injection), indicating approximately equivalent levels of cell proliferation. Red, BrdU positivity by anti-BrdU antibody fluorescence; blue,DAPI staining.(G) Ratio of total protein to total DNA content in various tissues derived from E18.5 wild-type (�/�) and p190-B RhoGAP mutant (�/�) animals,indicating that tissues from mutant embryos contain abnormally small cells. Values reflect the average of tissue measurements taken fromseven different mice, and error bars represent the standard error of the mean.(H) Immunoblot of p190-B RhoGAP with the indicated tissues dissected from an E18.5 wild-type embryo.


Figure 3. Activated CREB Rescues the Size Defect in Cells Lacking p190-B RhoGAP

(A) Forward light scatter analysis of G1-selected p190-B RhoGAP mutant (�/�) cells and a pool of mutant cells stably transfected with anexpression vector expressing activated CREB (CREBFY), indicating that the reduced cell size of mutant cells can be rescued by activatedCREB.(B) Immunoblot of protein lysates prepared from p190-B RhoGAP mutant 3T3 cells (left lane) and individual clonally derived cell lines of p190-B RhoGAP mutant cells stably expressing transfected CREBFY (FLAG epitope-tagged; anti-FLAG blot).(C) Relative cell size of G1-selected wild-type cells (�/�), p190-B RhoGAP mutant cells (�/�), and mutant clonal cell lines stably expressingCREBFY (FY1, FY2, and FY3), indicating the rescue of reduced cell size by CREBFY.(D and F) FACS gating of G1-selected (D) and G2-selected (F) populations of wild-type 3T3 cells.(E and G) Forward light scatter profile of wild-type, mutant, and CREBFY-rescued mutant 3T3 cells in both G1 (E) and G2 (G). Note thatCREBFY restores cell size in both G1 and G2 populations.

known biochemical action of p190-B RhoGAP as a dow- active CREB are in fact reduced in cells lacking p190-B RhoGAP, we used a CRE-luciferase reporter to deter-nmodulator of Rho. As would be expected, asynchro-

nously growing mutant cells were found to contain ab- mine that in mutant cells, CRE reporter activity is re-duced by 6-fold relative to wild-type cells (Figure 4B).normally high levels of activated, GTP-bound Rho.

Moreover, these levels are maintained in the CREB “res- Moreover, introduction of constitutively activated Rho(RhoV14) into wild-type cells results in a similar degree ofcue” lines, consistent with a role for CREB downstream

of Rho (Figure 4A). To verify that levels of biologically decreased CRE reporter activity (Figure 4B). Analogous

Developmental Cell6

Figure 4. Rho Antagonizes CREB-Mediated Transcriptional Activity

(A) Anti-Rho immunoblot of total cell lysates (lys.) and GST-Rhotekin “pull-downs” (GTP-Rho) from wild-type (�/�), p190-B RhoGAP mutant(�/�), and CREBFY-transfected mutant cells, indicating that levels of activated Rho are substantially increased in the mutant cells and thatthe proportion of Rho in the active, GTP-bound state is similarly increased in the CREBFY-transfected cells.(B) CRE-mediated transcriptional activity (CRE-luciferase reporter plasmid) in transiently transfected wild-type (�/�) and p190-B RhoGAPmutant (�/�) 3T3 cells, and in mutant cells cotransfected with activated Rho (RhoV14). The results indicate that the transcriptional activityof CREB is reduced by approximately 6-fold in mutant cells and is similarly reduced in wild-type cells transfected with RhoV14.(C) GAL4-CREB activation of a UAS-luciferase reporter confirming the requirement for ser133 phosphorylation and demonstrating the specificinhibitory effect of Rho on wild-type GAL4-CREB.(D) Immunoblot of cytochrome c in wild-type, mutant, and two CREBFY-rescued mutant 3T3 lines, indicating reduced expression in mutantcells that is restored in the CREBFY lines. Normalization is with an anti-actin blot.(E) Immunoblot of cytochrome c in wild-type and mutant brain (Br.), stomach (St.), and heart (Hrt.), indicating reduced expression in mutanttissues (from E18.5). Normalization is with an anti-actin blot.


experiments with an E2F-luciferase reporter as a control activation, and that this pathway can be specificallyrevealed similar levels of transcriptional activity in wild- downmodulated by increased levels of Rho GTPase ac-type and mutant cells (data not shown). To address tivity.specifically the ability of Rho to regulate CREB transcrip- One of the major downstream targets of insulin/IGFtional activity via phosphorylation of ser133, we used a receptor activation is the insulin receptor substrate IRS,GAL4-CREB reporter assay. In those experiments, Rho which plays a role in cell size regulation in Drosophilaantagonizes transcriptional activity of wild-type GAL4- (Bohni et al., 1999). Interestingly, one of the major RhoCREB, but does not affect the residual activity of GAL4- targets, Rho-kinase (ROK; Leung et al., 1995), can re-CREB containing a ser133 substitution (Figure 4C). We portedly antagonize insulin signaling through an associ-also determined that expression of one of the well-docu- ation with the major insulin receptor substrate IRS (Be-mented CREB-regulated genes, cytochrome c (Herzig gum et al., 2001; Farah et al., 1998) and an inhibitoryet al., 2000), is reduced in mutant 3T3s and embryonic phosphorylation (Paz et al., 1997). Therefore, we hypoth-tissues, and is restored in mutant 3T3s that express esized that increased ROK activity in mutant cells couldactivated CREB (Figures 4D and 4E). Together, these account for the observed defects. Indeed, we found thatobservations indicate that at least some of the conse- when mutant cells are exposed to a specific pharma-quences of hyperactivation of Rho in cells lacking p190- cological inhibitor of ROK (Y-27632), levels of IGF-1-B RhoGAP are likely to be mediated through specific induced CREB phosphorylation are restored to levelsinhibitory effects on CREB-dependent transcriptional seen in wild-type cells (Figure 5C). In addition, whenactivity. mutant cells are exposed to Y-27632, cell size is in-

These results implicate both Rho and CREB in cell creased nearly to wild-type size (Figure 5D). Thus, ROKsize regulation. However, in the context of the p190-B inhibition can “rescue” defects in cell size and IGF-1RhoGAP mutant cells, it is difficult to make this conclu- responsiveness in mutant cells. In addition, exposuresion definitively because it requires the assumption that of Rat-1 cells to Y-27632, in the absence of other treat-increased Rho activity is the only significant conse- ments, is sufficient to promote increased cell size (dataquence of losing p190-B RhoGAP. Therefore, to analyze not shown). This suggests that ROK is an importantmore directly the role of Rho and CREB in cell size Rho target in regulating cell size, and that the observedregulation, we examined transiently transfected Rat-1 growth inhibition seen upon Rho activation is not simplyfibroblasts. We observed by FACS analysis that cells a consequence of cell toxicity caused by active Rho,transfected with RhoV14 are significantly smaller than but rather that this reflects a physiological consequencenontransfected cells, while transfection with CREBFY of modulating Rho activity.results in cells that are abnormally large (Figure 5A). In To further establish the role of ROK and IRS in a celladdition, transfection of wild-type cells with a dominant- size pathway, we examined the IRS status in primarynegative form of CREB (Walton et al., 1992) gives rise MEFs lacking p190-B RhoGAP. After IGF-1 treatment,to cells that are reduced in size (Figure 5A). As before, levels of tyrosine phosphorylation of the IGF-1 receptorthese effects appear to be independent of potential ef- are normal, as expected; however, in mutant cells, therefects on cell cycle because the analysis was restricted is a significant decrease in tyrosine phosphorylation ofto cells in G1. These results, which suggest that Rho IRS-1, suggesting that IRS-1 does not interact as effi-activation inhibits cells growth and that CREB activation ciently with the activated IGF receptor in these cells (Fig-promotes cell growth, are consistent with the observa- ure 5E). Next, we tested the hypothesis that this reductiontion that cell size and CREB phosphorylation are both is associated with an increase in the “inhibitory” phosphor-reduced in mice lacking p190-B RhoGAP. ylation of IRS-1 (ser612) in mutant cells. This was found

to be the case (Figure 5F), and additionally, we foundRho Regulates Insulin Signaling to CREB via that treatment of mutant cells with Y-27632 “rescues” thean Inhibitory Effect of Rho-Kinase on IRS

excessive levels of serine phosphorylation of IRS-1, con-Previous studies in both vertebrate and invertebrate or-

sistent with the postulated role for ROK in antagonizingganisms revealed that cell size and animal size are regu-

insulin/IGF signaling through an inhibitory IRS phosphory-lated by components of the insulin/IGF signaling path-lation in mutant cells (Figure 5F).way (Stocker and Hafen, 2000). Significantly, we found

that wild-type fibroblasts undergo a substantial induc-tion of CREB phosphorylation following exposure to

Reduced MAP Kinase Signaling Links Rho-MediatedIGF-1, and that this response is diminished in cells lack-Inhibition of Insulin Signaling to CREBing p190-B RhoGAP (Figure 5B). Similar results were ob-We next tested the possibility that known downstreamserved with insulin (data not shown). We then determinedcomponents of insulin/IGF signaling that could poten-that expression of RhoV14 (Sullivan and Finkel, 2000) intially provide a link to CREB are affected in cells lackingwild-type cells is sufficient to reduce levels of IGF-1-p190-B RhoGAP. For these studies, we compared pri-induced phospho-CREB (Figure 5B). Significantly, levelsmary MEFs and embryonic tissues from wild-type andof IGF-1-induced phospho-ERK and phospho-MEK1 aremutant animals. In MEFs, cell size is also substantiallynot reduced in mutant cells, indicating that the observedreduced in the mutants (data not shown). In these pri-effect is specific (data not shown). In addition, we foundmary cell cultures, we confirmed the defect in CREBthat introduction of the activated Rac GTPase into wild-activation in response to IGF-1 in mutants, and addition-type cells does not inhibit insulin-induced CREB phos-ally, we found that phosphorylation of several down-phorylation, and in fact, causes an increase in CREBstream pathway components, including two kinases thatphosphorylation (data not shown). Together, these re-

sults indicate that insulin/IGF signaling promotes CREB can promote CREB activation (JNK, not shown and p38;

Developmental Cell8

Figure 5. A Rho-ROK (Rho-Kinase) Pathway Modulates Cell Size and IGF-Induced CREB Phosphorylation

(A) Relative cell size of G1-selected, GFP-positive Rat-1 fibroblasts transiently transfected with either RhoV14, CREBFY, or CREB-DN, indicatingthat expression of activated Rho or dominant-negative CREB results in reduced cell size, and expression of activated CREB promotesincreased cell size.(B) IGF-1-induced phospho-CREB (after 15 min) as revealed by immunoblotting with a phospho-CREB antibody (upper panel) of protein lysatesprepared from wild-type (�/�) and p190-B RhoGAP mutant (�/�) 3T3 cells, and wild-type cells infected with an adenovirus that expressesRhoV14 (48 hr before IGF-1 treatment). RhoV14 expression blocks approximately 50% of the IGF-induced CREB phosphorylation seen inwild-type cells. All non-RhoV14-expressing cells were infected with a GFP-expressing adenovirus as a negative control. Middle panel representsan immunoblot of the same lysates probed with anti-CREB antibody as a normalization control. Lower panel represents an immunoblot ofRho protein, indicating the size-shifted band corresponding to RhoV14 in the right-most lane.(C) IGF-induced CREB phosphorylation (after 15 min) similar to that shown in (B), and rescue of IGF-induced CREB phosphorylation in mutantcells by pretreatment with the Rho-kinase inhibitor Y-27632.(D) Relative cell size (by forward light scatter) of G1-selected wild-type (�/�) and p190-B RhoGAP mutant (�/�) 3T3 cells, and mutant cellstreated for 48 hr with Y-27632, indicating the partial rescue of cell size in mutant cells by inhibition of Rho-kinase. Values in (A) and (D) reflectthe average of three independent experiments, and error bars represent standard error of the mean.(E) Immunoprecipitation of IGF-1 receptor and IRS-1 from wild-type and mutant 3T3 cells before and after 15 min of IGF-1 treatment, followedby immunoblot with anti-phosphotyrosine antibodies. Samples were also immunoblotted for total IGF-1 receptor and IRS-1 for normalization.(F) Immunoprecipitation of IRS-1 from wild-type and mutant 3T3 cells before and after 15 min of IGF-1 treatment, followed by immunoblot withphospho-IRS-1 (phospho-ser612) antibodies. Y-27632 was included where indicated. Immunoblot of total IRS-1 is included for normalization.

Shaywitz and Greenberg, 1999), as well as Akt, is sub- are consistent with a model wherein increased ROKactivity in mutant cells downmodulates insulin/IGF sig-stantially reduced in mutant MEFs (Figure 6A). More-

over, expression of RhoV14 in wild-type MEFs is suffi- naling to several downstream pathway components, in-cluding MAP kinases that can promote CREB activitycient to inhibit IGF-1-induced phosphorylation of these

proteins, and treatment of mutant MEFs with Y-27632 (Figure 7A). Consistent with this model, we find thatlevels of phosphorylated p38, JNK, and Akt are substan-can restore normal levels IGF-1-induced phosphoryla-

tion of these kinases (Figure 6A). Together, these results tially reduced in embryonic tissues (Figures 6B and 6C),


Figure 6. Downstream Components of the Insulin/IGF Signaling Pathway Are Affected in Mutant Primary MEFs and Tissues

(A) Immunoblots with the indicated antibodies of lysates from wild-type and mutant MEFs upon treatment with IGF-1 (15 min), Y-27632, and/or adenovirus RhoV14 infection, as indicated.(B) Immunoblots with the indicated antibodies of tissue lysates from wild-type and mutant embryos (E18.5).(C) Immunoblot of phospho-Akt in the indicated wild-type and mutant tissues (E18.5).

and, as would be expected, there is a significant in- be uniformly reduced in size, while phospho-CREB lev-els are not substantially reduced in every tissue, raisescrease in ser612 phosphorylation of IRS (data not

shown). The expected reduction in levels of active Rac the possibility that there is additional complexity in-volved in determining organism size.GTPase in mutant cells was also confirmed (data not

shown). It is possible, for example, that modulation of CREBis a more significant determinant of cell size at earlierstages of development in some tissues (we focused onDiscussionE18.5). In addition, there may be both cell-autonomousand -nonautonomous mechanisms involved in vivo. In-Taken together, our observations indicate that activity

of the Rho GTPase and the CREB transcription factor deed, PI-3 kinase, which appears to play a role in thepathway proposed here, has been previously implicatedare important determinates of cell and animal size during

embryonic development, thereby defining a novel bio- in cell size regulation, and has been reported to haveboth cell-autonomous and -nonautonomous functionslogical function for both of these widely expressed regu-

latory proteins. In fibroblasts derived from mice lacking in C. elegans (Hsin and Kenyon, 1999). In addition, whileIRS plays a well-documented cell-autonomous role inp190-B RhoGAP, it appears that Rho is performing a

largely cell-autonomous role that can influence respon- regulating cell size in Drosophila (Bohni et al., 1999),studies in mice reveal an additional cell-nonautonomoussiveness to insulin/IGFs through activation of the Rho

target ROK. However, the fact that mutant fibroblasts role in controlling animal size (Withers, 2001). It is alsoworth noting that both insulin and IGF-1 gene expressionare not completely defective in insulin/IGF-1 respon-

siveness suggests that at least some aspect of insulin- have been found to be CREB responsive (Philippe andMissotten, 1990; Thomas et al., 1996), raising the possi-promoted growth is Rho insensitive. Moreover, the ob-

servation that mice lacking p190-B RhoGAP appear to bility that impaired CREB activity can also contribute in

Developmental Cell10

Figure 7. Model Depicting the ProposedRole of p190-B RhoGAP in Modulating Insu-lin/IGF-1 Signaling to CREB

a cell-nonautonomous manner to growth regulation by CREB-related protein provides some redundant func-affecting levels of secreted and circulating insulin/IGF. tion for cell survival in this setting, or that there is aSuch findings point to a complex regulatory system for phosphorylation-independent function for CREB in pro-the control of cell, organ, and animal size that involves moting survival. Interestingly, it has been reported thatboth cell-autonomous and -nonautonomous mecha- while mutational loss of one of the CREB homologs innisms that may require an overlapping set of proteins. Drosophila, dCREB2, results in a lethal phenotype, theIn fact, there may be additional compensatory mecha- few “escapers” that survive to adulthood are substan-nisms that serve to uniformly adjust the size of organs tially reduced in size (Belvin et al., 1999). This suggeststo accommodate changes that may be limited to a sub- that a role for CREB in determining animal size has beenset of organs in the mutant animals. evolutionarily conserved. It is also notable that CREB is

Rho GTPase signaling has not previously been directly a direct phosphorylation target of PKA (Colbran et al.,implicated in cell size regulation. In fact, the vast majority 1992), which, as described above, regulates animal sizeof biological functions associated with the Rho GTPase in Drosophila. Possibly, crosstalk between Rho- andin vivo appear to reflect the ability of Rho proteins to CREB-mediated pathways occurs at multiple levels.regulate the actin cytoskeleton (Settleman, 2000). How- The vast majority of studies of cell size regulation inever, there is indirect evidence to suggest a potential both vertebrates and invertebrates have focused on therole for Rho in cell size regulation. The human tumor insulin/IGF cell size pathway, and accumulating evi-suppressor TSC1 encodes a protein that reportedly pro- dence suggests that an important “endpoint” of thismotes Rho activation (Lamb et al., 2000) and which, pathway is the upregulation of components of the pro-when disrupted in Drosophila, results in tissue out- tein translational machinery (Stocker and Hafen, 2000).growth due to a substantial increase in cell size (Tapon However, there is also evidence that transcriptional reg-et al., 2001). Significantly, overexpression of TSC1, like ulation plays a role in regulating cell size, and it appearsRho activation, reportedly causes reduced cell growth that one of the relevant transcription factors is Myc,by inhibiting insulin signaling (Gao and Pan, 2001; Potter whose expression can promote cell growth in a cellet al., 2001). In addition, the PKA kinase, which, when cycle-independent manner (Iritani and Eisenman, 1999).disrupted in Drosophila, results in a small fly phenotype

Our results indicate that CREB also plays a significant(Lane and Kalderon, 1993), can directly phosphorylate

role in regulating cell growth, presumably through itsRho, resulting in Rho inactivation (Dong et al., 1998).

ability to regulate transcriptionally the expression of oneBoth of these results are consistent with the role of Rho

or more target genes. Thus far, the relevant transcrip-in cell size regulation that has been proposed here, andtional target(s) have not been identified, although, assignificantly, this role for Rho appears to be independentdiscussed, insulin and IGF-1 may be important CREBof its ability to regulate the cytoskeleton.targets in this context. Significantly, many of the identi-Previous studies to address the in vivo function offied CREB targets are regulators of metabolism (MayrCREB have highlighted its role in cellular differentiationand Montminy, 2001), raising the possibility that CREB-and protection form apoptosis (survival), as well as ininduced cell growth is due to increased expression oflearning and memory (Mayr and Montminy, 2001). Con-several essential metabolic enzymes. Thus, these col-sistent with this, we also see evidence of differentiationlective results suggest a model wherein cellular growthdefects in brain and thymus of mice lacking p190-Brequires increased protein synthesis together with in-RhoGAP (S.F.M., R.S., and J.S., unpublished observa-creased metabolism. Significantly, CREBFY-rescuedtions). However, we do not see evidence of any signifi-mutant cells are resistant to the ability of the p70 S6cant increase in the level of apoptosis in our mutantkinase inhibitor rapamycin to cause cell size reduction,animals, despite extensive analysis of embryos at sev-

eral developmental stages. Thus, it is possible that a suggesting that CREB may be additionally required


TCTTCACTTAGAACGG-3�), and primer 3, complementary to se-downstream of the translational machinery for cell sizequences at the 5� end of the PGK-neo promoter (5�-dCGGTGGATGTregulation (M.C. and J.S., unpublished observations).GGAATGTG-3�). Primers 1 and 2 were used to identify a 392 bpRho and CREB have also both been previously impli-fragment of the wild-type allele, whereas primers 2 and 3 were used

cated in the differentiation of thymocytes and neurons to amplify a 521 bp fragment of the mutated allele. PCR products(Henning et al., 1997; Luo, 2000), and, as described were generated after a 5 min hot start at 94�C by 30 cycles at 94�C

(1 min), 60�C (1 min), and 72�C (1 min) using Taq polymerase in theabove, we found that the brain and thymus defects inpresence of 1.25 mM MgCl2.p190-B RhoGAP-deficient mice are associated with a

failure in cell differentiation, as revealed by immuno-Immunodetection of Proteinsstaining with specific markers of differentiation. Signifi-To detect p190 RhoGAP proteins, lysates from embryo-derived cellscantly, in CREB-deficient mice, the small thymus is simi-were prepared in ice-cold lysis buffer (50 mM HEPES [pH 7.4], 150

larly associated with an excess of undifferentiated cells mM NaCl, 1.5 mM MgCl2, 5 mM EGTA, 10% glycerol, 1% Triton(Rudolph et al., 1998). Thus, our results are consistent X-100) containing protease inhibitors. Debris was removed by cen-with a role for Rho in regulating CREB function both for trifugation in a microfuge at 12,000 � g for 10 min at 4�C. Clarified

lysates were boiled in gel loading buffer and separated by 7.5%differentiation as well as for determining cell size andSDS-PAGE. Proteins were electrotransferred to nitrocellulose andanimal size. It is therefore interesting to note that indetected with monoclonal antibodies directed against p190 Rho-a cell culture model of neuronal differentiation (serumGAP or p190-B RhoGAP (hybridoma supernate diluted 1:5) followed

withdrawal from neuro-2a neuroblastoma cells), we ob- by incubation with horseradish peroxidase-conjugated secondaryserve, by FACS analysis, a substantial increase in cell goat anti-mouse antibody (Bio-Rad; 1:5000) and development withsize during differentiation (M.C. and J.S., unpublished enhanced chemiluminescence (DuPont NEN) and autoradiography.

For immunoprecipitation, cells that were treated or not treated withobservations). Such cell growth is particularly notableIGF-1 for 15 min were extracted with RIPA buffer, clarified by centrif-when considering that it takes place in the absenceugation, and subjected to immunoprecipitation with IGF-1 or IRS-1of any serum-containing factors. In similar cell cultureantibodies and protein A agarose for 2 hr. After washing the beads

models of neuronal differentiation, CREB has been several times with RIPA buffer, samples were subjected to immu-found to play an important role in promoting differentia- noblotting. For immunoblotting of CREB, JNK, p38, Akt, IRS-1, IGF-1tion, and it will certainly be of interest in the future to receptor, cytochrome c, actin, and the various phosphoproteins,

extracts were prepared from cells and tissues in RIPA buffer anddetermine whether CREB utilizes a distinct or overlap-lysates were clarified by centrifugation. Fifteen micrograms of pro-ping set of transcriptional targets for regulating cell sizetein of each sample were resolved by 12.5% SDS-PAGE, and theand differentiation.proteins were then electrotransferred to an Immobilon P membrane.Immunoblotting was performed as described for p190-B RhoGAP.

Experimental Procedures All antibodies were obtained from New England Biolabs.

Generation of p190-B RhoGAP Mutant Mice HistologyTo construct the p190-B targeting vector, a PGK-neo cassette was Newborn mice were sacrificed by CO2 asphyxiation and either dis-inserted in reverse transcriptional orientation relative to the p190- sected immediately or fixed by immersion in Bouin’s fixative. ForB coding sequence in a BamHI site (bp 1196) of a 4.5 kb p190-B Nissl staining, E18.5 wild-type and mutant forebrains were fixed bySV129 5� genomic clone, disrupting the gene immediately upstream immersion (4% paraformaldehyde) and then cryoprotected in PBS/of the GTPase domain (Burbelo et al., 1998). The complete insert 20% sucrose. Coronal cryostat sections (10–14 �M) were stainedof the resulting plasmid was subcloned as a KpnI-NotI fragment with cresyl violet acetate (0.5% in H2O) for 10 min.into a PGK-neo-free pPNT vector (Tybulewicz et al., 1991), resultingin the addition of the PGK-tk gene at the 5� end of the p190-B

Cell Size and Cell Cycle Analysisgenomic DNA. NotI-linearized targeting vector (25 �g/ 8 � 106 cells)

For determination of protein:cell number and protein:DNA ratios,was electroporated (400 V, 25 �F) into D3 ES cells of the mouse

cells were counted and lysed in RIPA buffer, and DNA and proteinSV129 strain. ES cells were maintained on a feeder layer of neor

content were determined as previously described (Franch et al.,embryonic fibroblasts in DMEM supplemented with 15% fetal bovine

1995). For FACS analysis (forward light scatter), 3T3 cells were la-serum (FBS) in the presence of 500 U/ml LIF (GIBCO). Transfected

beled with 0.1 mg/ml BrdU (cell labeling reagent; Amersham) forES cells were selected for growth in G418 at 300 �g/ml for 10

30 min, harvested, and analyzed according to the manufacturer’sdays and gancyclovir at 2 mM the first 5 days (initiated 36 hr after

instructions (Becton Dickinson). Briefly, cells were fixed in 80%electroporation). Single colonies were expanded for DNA isolation

ethanol, and the DNA was denatured by treatment with 2 N HCl/and Southern blot analysis was performed using standard methods,

0.5% Triton X-100 at room temperature for 30 min followed by ato identify correctly targeted p190-B heterozygous ES clones (14

neutralization step in 0.1 M Na2B4O7 � 10H2O (pH 8.5). For immuno-out of 130 screened). A 500 bp HindIII fragment (bp 1088–1586) was

fluorescence staining, anti-BrdU antibodies (Becton Dickinson) andused to distinguish the mutated allele from the wild-type allele based

FITC-conjugated horse anti-mouse antibodies (Vector) were used.on an altered StuI restriction pattern after homologous recombina-

The immunofluorescence step was followed by incubation in propid-tion. 5� and 3� external probes were used to confirm the absence

ium iodide and RNase prior to two-dimensional FACS analysis usingof rearrangements 5� and 3� of the targeted region. ES cells from

Cell Quest software (Becton Dickinson). To measure the relative cellthree p190-B heterozygous clones were injected into the blastocoel

size of asynchronously growing cells corresponding to the wild-cavity of 3.5-day-old C57BL/6J blastocysts, which were then trans-

type and mutant 3T3 lines, single cells with a DNA content of 2Nplanted into the uteri of pseudopregnant females to generate chime-

(the G1 population) were gated and forward scatter height (FSC-H)ric mice. Germline transmission was determined by agouti color

histograms were plotted. The mean FSC-H was calculated usingoffspring of chimeric males mated with C57BL/6J females and con-

Cell Quest software. The analysis was performed with similar resultsfirmed by Southern blot analysis of StuI-digested tail DNA.

on three independently derived pairs of wild-type and mutant 3T3cell lines.

For FACS analysis of transiently transfected Rat-1 cells, cells (60GenotypingGenotype analysis was performed by PCR using a three-primer reac- hr posttransfection) were fixed in 1% formaldehyde in PBS for 30

min, and permeabilized with 0.1% Triton X-100 in PBS followed bytion containing primer 1, complementary to p190-B sequences adja-cent to the PGK-neo gene insertion site (bp 1185–1202; 5�-dTAATG propidium iodide staining in the presence of RNase A, as described

above. The forward scatter height of the GFP-positive transfectedATAGGCGGATCCC-3�), primer 2, complementary to p190-B codingsequence 3� of the region of gene disruption (bp 1559–1577; 5�-dGGT cells with a 2N DNA content was analyzed as described above.

Developmental Cell12

Fibroblast Cell Culture Studies Reporter AssaysTo prepare embryo-derived fibroblasts, E13.5 embryos were ob- Reporter assays were performed following transfection (Fugene;tained from timed matings of p190-B RhoGAP heterozygotes. DNA Boehringer Mannheim) of 3T3 cells with the CRE-luciferase reporterwas extracted from tail tissue for genotype analysis. After removal plasmid (Stratagene; 100 ng) or the UAS-luciferase reporter plasmidof the head and internal organs, embryos were washed in PBS under (Stratagene; 250 ng). Cotransfections were with plasmids express-sterile conditions, followed by trituration and 30 min trypsinization ing either RhoV14 (100 ng), wild-type GAL4-CREB, or GAL4-CREB(using 5� trypsin) at 37�C. Trypsin was inactivated by reconstituting containing an ala133 substitution (100 ng). GAL4-CREB fusions con-the cells in DMEM containing 15% fetal calf serum (FCS) followed tain the entire CREB coding sequence. Twenty hours after transfec-by centrifugation at 1500 rpm for 5 min at room temperature. Mouse tion, the cells were switched to media containing 1% serum and 40embryonic fibroblasts (MEFs) were then resuspended in DMEM (plus hr later luciferase assays were performed (according to the manu-15% fetal calf serum) and plated at high density (cells derived from facturer’s instructions; Promega). RSV-LacZ (200 ng) was used asone embryo in one 10 cm dish). 3T3 cell lines were established using an internal control for transfection efficiency. CRE-luciferase activ-standard procedures (Todaro and Green, 1963). Briefly, cells were ity, as indicated in graph form, represents activity normalized totrypsinized, counted, and reseeded in DMEM supplemented with �-galactosidase activity from the same extracts. Results were quan-10% calf serum at a density of 106 cells/10 cm dish every 3 days tified using a luminometer from EG&G Berthold for luciferase andwith a media change the subsequent day. After 30 to 50 passages, an ELISA plate reader for �-galactosidase.stable lines of immortalized cells were established. Saturation den-sity was determined by plating 104 wild-type or mutant cells in a 60

Acknowledgmentsmm dish and then counting in triplicate the number of cells per dishevery day for 6 consecutive days following replating. 3T3 and Rat-1

We thank Drs. Daniel Haber, Iswar Hariharan, and Sofie Salama forfibroblast lines were maintained in DMEM containing 10% calf serumhelpful comments on the manuscript. We appreciate the generousor 10% fetal bovine serum, respectively, and antibiotics.assistance of Dr. Nadia Carlesso with GFP FACS analysis, Dr. FredTo generate “CREB rescue” 3T3 cells, homozygous mutant 3T3

cells were cotransfected (Fugene; Boehringer Mannheim) with plas- Preffer with interpretation of FACS data, Dr. En Li of the MGH mousemids encoding constitutively active CREB (CREBFY) and the puro- transgenic/knockout facility with the production of mutant mice, Dr.mycin resistance protein (pBabe-puro), then selected with puromy- Thaddeus Stappenbeck with the analysis of lung tissue, and Ms.cin (Sigma; 2 �g/ml) for 2 weeks, and pooled. The mean FSC-H of Tracy Huang with assistance with immunoblotting. We are gratefulthe G1 population was compared to that of the homozygous mutant to members of the Settleman laboratory for advice throughout theparental 3T3 cells. Cells were plated at 1 � 106 cells/10 cm dish course of these studies. The CREB expression plasmids were kindlyand the G1 and G2 populations of cells were analyzed for forward provided by Dr. Marc Montminy (Salk Institute), Dr. Linda Kobierskilight scatter FACS analysis as described above. In addition, three (MGH), and Drs. Jon Kornhauser and Michael Greenberg (Harvardindividual clonal lines of stably transfected CREBFY-expressing Medical School), and the RhoV14 adenovirus was a generous giftcells were established, and their relative cell sizes were also deter- of Drs. Enzo Calautti and Paolo Dotto (MGH). These studies weremined by forward light scatter FACS analysis. supported by NIH and ACS awards to J.S.

For experiments with Rat-1 cells, 106 asynchronously growingcells were plated per 10 cm dish in DMEM containing 10% FBS the

Received: February 15, 2002day before transfection. The next day, the cells were transfectedRevised: March 19, 2002with 10 �g of DNA (5 �g vector control, CMV-RhoV14, CMV-CREBDN

[dominant-negative], or CMV-CREBFY [activated], together with 5�g CMV-GFP vector) and 10 �l of Lipofectamine 2000 transfection Referencesreagent (Life Technologies). Twelve hours posttransfection, the me-dium was replaced with serum-free DMEM. After an additional 12 Begum, N., Sandu, O.A., Ito, M., Lohmann, S.M., and Smolenski,hr, the medium was replaced with DMEM containing 2.5% FBS and

A. (2002). Active Rho kinase (ROK-alpha) associates with insulin2 �g/ml aphidicolin (to enrich for a G1 population of cells; Urbani

receptor substrate-1 and inhibits insulin signaling in vascularet al., 1995). Thirty-six hours later, the cell size of the G1 cell popula-smooth muscle cells. J. Biol. Chem. 277, 6214–6222.tion was determined by FACS as described above.Belvin, M.P., Zhou, H., and Yin, J.C. (1999). The Drosophila dCREB2For experiments with IGF-1 treatment, 106 asynchronously grow-gene affects the circadian clock. Neuron 22, 777–787.ing 3T3 cells were plated in 10 cm dishes in DMEM with 10% calf

serum. After 3 days, the cells reached confluency and the medium Billuart, P., Winter, C.G., Maresh, A., Zhao, X., and Luo, L. (2001).was replaced with serum-free DMEM. After an additional 24 hr, cells Regulating axon branch stability: the role of p190 RhoGAP in re-were incubated with medium containing DMEM alone or DMEM and pressing a retraction signaling pathway. Cell 107, 195–207.80 ng/ml of IGF-1 (Life Technologies). After 15 min, cells were lysed

Bohni, R., Riesgo-Escovar, J., Oldham, S., Brogiolo, W., Stocker, H.,in RIPA buffer and 15 �g of protein was analyzed by immunoblottingAndruss, B.F., Beckingham, K., and Hafen, E. (1999). Autonomousfor the indicated proteins as described above. Adenoviral infectionscontrol of cell and organ size by CHICO, a Drosophila homolog ofwith virus expressing both RhoV14 (myc epitope-tagged) and GFPvertebrate IRS1–4. Cell 97, 865–875.or GFP alone (negative control) were performed 48 hr before IGF-1

treatment as previously described (Sullivan and Finkel, 2000). For Brouns, M.R., Matheson, S.F., Hu, K., Delalle, I., Caviness, V.S.,treatments with the Y-27632 compound (Welfide), the drug (10 �M) Silver, J., Bronson, R.T., and Settleman, J. (2000). The adhesionwas added to culture media 15 min prior to IGF-1 treatment in signaling molecule p190 RhoGAP is required for morphogeneticphospho-CREB/p38/Akt experiments, and for cell size studies, it processes in neural development. Development 127, 4891–4903.was added once per day on 2 consecutive days in the presence of

Brouns, M.R., Matheson, S.F., and Settleman, J. (2001). p190 Rho-2.5% FBS and 2 �g/ml aphidicolin prior to collecting G1 cells forGAP is the principal Src substrate in brain and regulates axon out-forward light scatter FACS analysis as described above.growth, guidance and fasciculation. Nat. Cell Biol. 3, 361–367.

Measuring Rho Activation Burbelo, P.D., Miyamoto, S., Utani, A., Brill, S., Yamada, K.M., Hall,Levels of activated, GTP-bound RhoA were determined using the A., and Yamada, Y. (1995). p190-B, a new member of the Rho GAPRhotekin “pull-down” assay as previously described (Ren and family, and Rho are induced to cluster after integrin cross-linking.Schwartz, 2000). Briefly, asynchronously growing wild-type and mu- J. Biol. Chem. 270, 30919–30926.tant cells were collected, lysed, and following clarification, lysates Burbelo, P.D., Finegold, A.A., Kozak, C.A., Yamada, Y., and Takami,were incubated with GST-Rhotekin on beads to capture GTP-bound

H. (1998). Cloning, genomic organization and chromosomal assign-Rho proteins. Bound proteins were washed several times and eluted

ment of the mouse p190-B gene. Biochim. Biophys. Acta 1443,with sample buffer before being subjected to SDS-PAGE and immu-203–210.noblotting with an anti-RhoA antibody (Santa Cruz). Whole-cell ly-Colbran, J.L., Roach, P.J., Fiol, C.J., Dixon, J.E., Andrisani, O.M.,sates (prior to incubation with GST-Rhotekin) were analyzed in paral-and Corbin, J.D. (1992). cAMP-dependent protein kinase, but notlel. The amount of whole-cell lysate loaded in each case represents

approximately 5% of the input material in binding assays. the cGMP-dependent enzyme, rapidly phosphorylates delta-CREB,


and a synthetic delta-CREB peptide. Biochem. Cell Biol. 70, 1277– lethality in mice lacking the cAMP response element binding protein.Proc. Natl. Acad. Sci. USA 95, 4481–4486.1282.

Settleman, J. (1999). Rho GTPases in development. Prog. Mol. Sub-Dong, J.M., Leung, T., Manser, E., and Lim, L. (1998). cAMP-inducedcell. Biol. 22, 201–229.morphological changes are counteracted by the activated RhoA

small GTPase and the Rho kinase ROKalpha. J. Biol. Chem. 273, Settleman, J. (2000). Getting in shape with Rho. Nat. Cell Biol. 2,22554–22562. E7–E9.Du, K., Asahara, H., Jhala, U.S., Wagner, B.L., and Montminy, M. Settleman, J., Narasimhan, V., Foster, L.C., and Weinberg, R.A.(2000). Characterization of a CREB gain-of-function mutant with (1992). Molecular cloning of cDNAs encoding the GAP-associatedconstitutive transcriptional activity in vivo. Mol. Cell. Biol. 20, 4320– protein p190; implications for a signaling pathway from ras to the4327. nucleus. Cell 63, 539–549.Farah, S., Agazie, Y., Ohan, N., Ngsee, J.K., and Liu, X.J. (1998). Shaywitz, A.J., and Greenberg, M.E. (1999). CREB: a stimulus-A rho-associated protein kinase, ROKalpha, binds insulin receptor induced transcription factor activated by a diverse array of extracel-substrate-1 and modulates insulin signaling. J. Biol. Chem. 273, lular signals. Annu. Rev. Biochem. 68, 821–861.4740–4746. Stocker, H., and Hafen, E. (2000). Genetic control of cell size. Curr.Franch, H.A., Shay, J.W., Alpern, R.J., and Preisig, P.A. (1995). Opin. Genet. Dev. 10, 529–535.Involvement of pRB family in TGF beta-dependent epithelial cell Sullivan, D.M., and Finkel, T. (2000). Expression of Rho GTPaseshypertrophy. J. Cell Biol. 129, 245–254. using adenovirus vectors. Methods Enzymol. 325, 303–314.Gao, X., and Pan, D. (2001). TSC1 and TSC2 tumor suppressors Symons, M., and Settleman, J. (2000). Rho family GTPases: moreantagonize insulin signaling in cell growth. Genes Dev. 15, 1383– than simple switches. Trends Cell Biol. 10, 415–419.1392.

Tapon, N., Ito, N., Dickson, B.J., Treisman, J.E., and Hariharan, I.K.Henning, S.W., Galandrini, R., Hall, A., and Cantrell, D.A. (1997). The (2001). The Drosophila tuberous sclerosis complex gene homologsGTPase Rho has a critical regulatory role in thymus development. restrict cell growth and cell proliferation. Cell 105, 345–355.EMBO J. 16, 2397–2407.

Thomas, M.J., Umayahara, Y., Shu, H., Centrella, M., Rotwein, P.,Herzig, R.P., Scacco, S., and Scarpulla, R.C. (2000). Sequential se- and McCarthy, T.L. (1996). Identification of the cAMP response ele-rum-dependent activation of CREB and NRF-1 leads to enhanced ment that controls transcriptional activation of the insulin-likemitochondrial respiration through the induction of cytochrome c. J. growth factor-I gene by prostaglandin E2 in osteoblasts. J. Biol.Biol. Chem. 275, 13134–13141. Chem. 271, 21835–21841.Hsin, H., and Kenyon, C. (1999). Signals from the reproductive sys- Todaro, G.J., and Green, H. (1963). Quantitative studies of the growthtem regulate the lifespan of C. elegans. Nature 399, 362–366. of mouse embryo cells in culture and their development into estab-Iritani, B.M., and Eisenman, R.N. (1999). c-Myc enhances protein lished lines. J. Cell Biol. 17, 299–313.synthesis and cell size during B lymphocyte development. Proc. Tybulewicz, V.L.J., Crawford, C.E., Jackson, P.K., Bronson, R.T.,Natl. Acad. Sci. USA 96, 13180–13185. and Mulligan, R.C. (1991). Neonatal lethality and lymphopenia inJalink, K., van Corven, E.J., Hengeveld, T., Morii, N., Narumiya, S., mice with a homozygous disruption of the c-abl proto-oncogene.and Moolenaar, W.H. (1994). Inhibition of lysophosphatidate- and Cell 65, 1153–1163.thrombin-induced neurite retraction and neuronal cell rounding by Urbani, L., Sherwood, S.W., and Schimke, R.T. (1995). DissociationADP ribosylation of the small GTP-binding protein Rho. J. Cell Biol. of nuclear and cytoplasmic cell cycle progression by drugs em-126, 801–810. ployed in cell synchronization. Exp. Cell Res. 219, 159–168.Lamb, R.F., Roy, C., Diefenbach, T.J., Vinters, H.V., Johnson, M.W., Van Aelst, L., and D’Souza-Schorey, C. (1997). Rho GTPases andJay, D.G., and Hall, A. (2000). The TSC1 tumour suppressor hamartin signaling networks. Genes Dev. 11, 2295–2322.regulates cell adhesion through ERM proteins and the GTPase Rho.

Vincent, S., and Settleman, J. (1999). Inhibition of RhoGAP activityNat. Cell Biol. 2, 281–287.is sufficient for the induction of Rho-mediated actin reorganization.

Lane, M.E., and Kalderon, D. (1993). Genetic investigation of cAMP- Eur. J. Cell Biol. 78, 539–548.dependent protein kinase function in Drosophila development.

Walton, M.R., and Dragunow, I. (2000). Is CREB a key to neuronalGenes Dev. 7, 1229–1243.

survival? Trends Neurosci. 23, 48–53.Leung, T., Manser, E., Tan, L., and Lim, L. (1995). A novel serine/

Walton, K.M., Rehfuss, R.P., Chrivia, J.C., Lochner, J.E., and Goodman,threonine kinase binding the Ras-related RhoA GTPase which trans-

R.H. (1992). A dominant repressor of cyclic adenosine 3�,5�-mono-locates the kinase to peripheral membranes. J. Biol. Chem. 270,

phosphate (cAMP)-regulated enhancer-binding protein activity in-29051–29054.

hibits the cAMP-mediated induction of the somatostatin promoterLuo, L. (2000). Rho GTPases in neuronal morphogenesis. Nat. Rev. in vivo. Mol. Endocrinol. 6, 647–655.Neurosci. 1, 173–180.

Withers, D.J. (2001). Insulin receptor substrate proteins and neuro-Mayr, B., and Montminy, M. (2001). Transcriptional regulation by the endocrine function. Biochem. Soc. Trans. 29, 525–529.phosphorylation-dependent factor CREB. Nat. Rev. Mol. Cell. Biol.2, 599–609.

Paz, K., Hemi, R., LeRoith, D., Karasik, A., Elhanany, E., Kanety, H.,and Zick, Y. (1997). A molecular basis for insulin resistance. Elevatedserine/threonine phosphorylation of IRS-1 and IRS-2 inhibits theirbinding to the juxtamembrane region of the insulin receptor andimpairs their ability to undergo insulin-induced tyrosine phosphory-lation. J. Biol. Chem. 272, 29911–29918.

Philippe, J., and Missotten, M. (1990). Functional characterizationof a cAMP-responsive element of the rat insulin I gene. J. Biol.Chem. 265, 1465–1469.

Potter, C.J., Huang, H., and Xu, T. (2001). Drosophila Tsc1 functionswith Tsc2 to antagonize insulin signaling in regulating cell growth,cell proliferation, and organ size. Cell 105, 357–368.

Ren, X.D., and Schwartz, M.A. (2000). Determination of GTP loadingon Rho. Methods Enzymol. 325, 264–272.

Rudolph, D., Tafuri, A., Gass, P., Hammerling, G.J., Arnold, B., andSchutz, G. (1998). Impaired fetal T cell development and perinatal

rho gtpase control of creb and animal size

Documents