Zizimin1, a novel Cdc42 activator, reveals a new GEF domain for Rho proteins

articles

Zizimin1, a novel Cdc42 activator, reveals a new GEF domain for Rho proteins Nahum Meller, Mohammad Irani-Tehrani, William B. Kiosses, Miguel A. Del Pozo and Martin A. Schwartz* Division of Vascular Biology, Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla CA 92037, USA *e-mail: [email protected]

Published online: 31 July 2002; doi 10.1038/ncb835

The Rho family GTPases Rac, Rho and Cdc42 are critical in regulating the actin-based cytoskeleton, cell migration, growth, survival and gene expression. These GTPases are activated by guanine nucleotide-exchange factors (GEFs). A biochemical search for Cdc42 activators led to the cloning of zizimin1, a new protein whose overexpression induces Cdc42 activation. Sequence comparison combined with mutational analysis identified a new domain, which we named CZH2, that mediates direct interaction with Cdc42. CZH2-containing proteins constitute a new superfamily that includes the so-called ‘CDM’ proteins that bind to and activate Rac. Together, the results suggest that CZH2 is a new GEF domain for the Rho family of proteins.

T

he Rho family of low-molecular-weight GTPases, including Rac, Rho and Cdc42, are critical for multiple cellular processes including regulation of the actin-based cytoskeleton, cell–cell contacts, gene transcription, cell-cycle progression, apoptosis and tumour progression1,2. Cdc42 mediates filopodia formation via the effectors WASP and N-WASP, as well as gene expression, cellcycle progression, cell polarity and cell–cell contacts1. Like other GTPases, Rho proteins are active when bound to GTP and inactive when bound to GDP. Conversion of the GDP-bound proteins to the active state is catalysed by GEFs. This reaction involves several stages. First the GEF binds with low affinity to the GDP-bound small GTPase. GDP then dissociates from this complex, leading to formation of a higher-affinity intermediate. This intermediate is then dissociated by GTP binding. So GEFs catalyse nucleotide exchange by destabilizing the strong interaction with GDP and stabilizing the nucleotide-depleted state3. Eukaryotic GEFs for Rho-like GTPases share a common motif, designated the Dbl-homology (DH) domain, for which the Dbl oncogene is the prototype4. Deletion analysis of the Dbl protein demonstrated that the DH domain was necessary and sufficient for GEF activity5. Another family that activates Rac — termed the CDM (CED-5, DOCK180, Myoblast city) proteins6 — has been identified, although whether these are GEFs or adaptors that recruit Dbl proteins is not known. For example, overexpressing DOCK180 (ref. 7) or DOCK2 (ref. 8) in 293 cells increased the level of GTPRac, and dominant-negative Rac blocked DOCK180-induced local membrane spreading9. Mice lacking DOCK2 are deficient in lymphocyte migration and Rac activation in response to chemokines10. In addition, mutations in CED-5, the Caenorhabditis elegans orthologue of DOCK180 and DOCK2, can be rescued by overexpressing CED-10, the orthologue of Rac11. Importantly, although DOCK180 and DOCK2 were shown to associate with nucleotide-depleted Rac7,8,12, it is not known whether this interaction is direct. To identify new Cdc42 activators in fibroblasts, we used binding to nucleotide-depleted Cdc42 to purify GEF-like proteins. This led to the cloning of a Cdc42-activating protein we call zizimin1 that is related to the CDM proteins. This protein activates Cdc42 in vivo and in vitro and binds preferentially to nucleotide-depleted Cdc42. We also identified a 370-amino-acid domain that mediates direct interaction with Cdc42 and that is conserved in CDM and other protein families.

Results Cloning of zizimin1. GEFs can be distinguished from other GTPase-interacting proteins by their ability to bind preferentially to the nucleotide-depleted state compared with GTP- or GDP-bound states3,5. To search for major Cdc42 GEFs expressed in fibroblasts and other adherent cell types, we adopted the approach of Hart and colleagues13 who isolated the DH-domain-containing protein p115-RhoGEF based on its tight binding to nucleotidedepleted Rho. We have further refined this approach by eluting with specific nucleotides. To establish the method, glutathione-S-transferase (GST)– Cdc42 beads were depleted of nucleotides or loaded with GTPγS and incubated with cell lysate from NIH-3T3 cells expressing Dbl. For the nucleotide-depleted beads, lysates were supplemented with EDTA to chelate magnesium, an essential cofactor for nucleotide binding. For the nucleotide-bound Cdc42 beads, magnesium was added to lysates to maintain nucleotide binding. After washing the beads, potential GEFs were eluted in a buffer that contained magnesium and GTP or GDP to reverse GEF binding. Control reactions included elution in magnesium buffer with ATP or without nucleotides. In certain cases, a second, consecutive elution was performed using free glutathione to release the fusion protein and its associated proteins from the beads. The second elution provided an estimate of the specificity and efficiency of the GTP elution. As shown in Fig. 1a, Dbl was detected in the GTPγS eluate from the nucleotide-depleted Cdc42 column but not from the GTPγS-Cdc42 column. Dbl was not released when the nucleotide-depleted Cdc42 beads were eluted with buffer containing ATP or without nucleotides. Analysis of the second glutathione elution demonstrated that the majority of bound Dbl was eluted by GTPγS. We also tested binding of Dbl to D118A Cdc42, a mutant with reduced affinity for nucleotides. Dbl bound to D118A Cdc42 even in the presence of magnesium and was not eluted by GTPγS (Fig. 1a, lane 5), suggesting that binding to wild-type Cdc42 in the presence of EDTA and elution by GTPγS does not result from non-specific effects of these compounds. These results confirm the effectiveness and specificity of the approach. We then performed the assay using biotin-labelled NIH-3T3 cell lysate and then analysed the eluates by probing western blots with horse radish peroxidase (HRP)–avidin. We observed a protein with

© 2002 Nature Publishing Group NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

639

articles

WT Cdc42

Nucleotide loading: Nucleotide elution: anti-Dbl

– –

– G

– A

G G

– G

–

Sample:

1

2

3

4

5

1

3

4

5

avidin–HRP 118 –

Second elution

85 –

Nucleotide loading: – Mr(K) Zizimin1 187 –

G –

Cd c4 2 Rh o

d Cd c4 2 Ra c

c

2

G – – G G Glutathione

G –

61 – Sample:

–

1

2

3

1

2

3

First elution Second elution

e

avidin–HRP 118 – Nucleotide loading:

G

–

Ra c Rh o

First elution

– – G – Glutathione

Nucleotide loading: – – Nucleotide elution: G A Mr(K) Zizimin1 187 –

Cd c4 2

WT Cdc42

D1 18 A

b D1 18 A

a

–

–

85 – anti-HA

61 –

Figure 1 Specific binding of zizimin1 to nucleotide-depleted Cdc42. a, Specific binding of Dbl to nucleotide-depleted Cdc42 and its elution with GTPγS. Lysates of NIH-3T3 cells stably expressing oncogenic Dbl were supplemented with EDTA (samples 1–3) or MgCl2 (samples 4–5) and incubated with nucleotide-depleted (nucleotide loading: −) or GTPγS-loaded (nucleotide loading: G) Cdc42 beads, as indicated. The beads were first eluted in buffer supplemented with the indicated nucleotides (nucleotide elution: G, GTPγS; A, ATP; −, no nucleotides). A second elution was performed (second elution) using glutathione. Samples 1–4 contained wildtype Cdc42, and sample 5 contained D118A Cdc42. Eluates were analysed by western blotting with anti-Dbl antibody. The position of Dbl is indicated by an arrow. b, Specific binding of zizimin1 to nucleotide-depleted Cdc42 and elution with GTPγS. NIH-3T3 cell lysates were labelled with biotin and incubated with

nucleotide-depleted Cdc42 columns, or control GTPγS-loaded Cdc42 columns as indicated. Bound proteins were first eluted with either GTPγS (G) or ATP (A) followed by a glutathione elution. Eluted proteins were detected by western blotting using avidin–HRP. The position of zizimin1 is indicated by an arrow. c,d Selective binding of zizimin1 to Cdc42 but not Rac or Rho beads. Biotinylated NIH-3T3 cell lysates were incubated with nucleotide-depleted or GTPγS-loaded Cdc42, Rac (c) or Rho (d) columns. Bound proteins were eluted with GTPγS and detected as in b. The position of zizimin1 is indicated by an arrow. e, Recombinant zizimin1 binds nucleotidedepleted Cdc42 but not Rac or Rho. Cos-7 cells were transfected with pEF4-HAZizimin1. Cell lysates prepared 48 h later were incubated with either nucleotidedepleted or GTPγS-loaded Cdc42 Rac or Rho beads. Bound proteins were eluted in SDS sample buffer and detected by anti-HA western blotting.

an apparent molecular mass (Mr) of 220,000 (220K) that bound specifically to nucleotide-depleted but not GTPγS-loaded Cdc42 and was eluted specifically with GTPγS (Fig. 1b). We designated this protein p220. No other specific bands were evident above the background. Similar analyses of lysates from human umbilical vein endothelial cells (HUVEC), rat vascular smooth muscle cells (VSMC) and Cos-7 cells also identified a protein with Mr 220K that bound specifically to nucleotide-depleted Cdc42 (data not shown). This 220K protein did not bind to Rac or Rho (Fig. 1c, d). Rho bound a protein with slightly lower molecular mass and another protein with Mr 115K (Fig. 1d) that is probably p115-RhoGEF (ref. 13). p220 was purified in larger quantity from NIH-3T3 cells, its peptide sequence obtained and the complete human complementary DNA cloned (GenBank accession number: AF527605). A vector coding for haemagglutinin (HA)-tagged protein was prepared and transfected into Cos-7 cells. p220 bound to Cdc42 in the predicted nucleotide-dependent manner (Fig. 1e), confirming its resemblance to the protein isolated from NIH-3T3 cells. p220 was named zizimin1 (from the Hebrew word zizim meaning spikes) because of its ability to induce microspikes when overexpressed in fibroblasts (see below). zizimin1 messenger RNA is broadly distributed. Northern blotting using a probe made of the first 450 base pairs (bp) of zizimin1 detected a single band of approximately 7.5 kb that is broadly distributed in organs and tissues of adherent cells. The highest zizimin1

mRNA levels were detected in heart and placenta, with intermediate levels in the kidney, brain, lung and skeletal muscle, and low levels in liver, intestine and hematopoietic tissues (Fig. 2a). Cdc42 is activated by zizimin1 overexpression. The capacity of zizimin1 to induce GTP loading on Cdc42 in cells was tested using the ‘PBD pull-down assay’14. NIH-3T3 cells were transfected with HA-tagged Cdc42 plus either wild-type or truncated zizimin1 (zizimin11–1695; see below). Cdc42-GTP was precipitated from cell lysates using the p21-binding domain (PBD) of PAK1. Overexpression of wild-type zizimin1 significantly increased GTP loading on Cdc42, whereas the truncated mutant was inactive (Fig. 2b). The Cdc42 activation level in different experiments varied between twofold and eightfold depending on the expression level of zizimin1. To assess activation of endogenous Cdc42 by zizimin1, NIH3T3 cells were transfected with zizimin1 cDNA and cells were fixed and stained 30 h later using rhodamine-phalloidin to label F-actin. Overexpression of zizimin1 substantially increased filopodia in stably adherent cells (Fig. 3a). Time-lapse imaging demonstrated dynamic extension and retraction of microspikes (data not shown), indicating that these structures were true filopodia and not retraction fibres. To test the specificity of zizimin1-induced filopodia for Cdc42, cells were microinjected with zizimin1 together with dominant-negative mutants or specific inhibitors of Rho proteins. Overexpression of the Cdc42-binding domain (CBD) of WASP was used to inhibit endogenous Cdc42 (ref. 15). Filopodia induced by

© 2002 Nature Publishing Group 640

NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

leu ko cy te

articles

Size (kb)

Br ain He art Sk ele t Co al m us lon cle Th ym us Sp lee n K id ne y Liv er Sm all Pla intes tin ce e Lu nta ng Pe rip he ral blo od

a

(1–

in1

in1 ziz im

ziz im

Ve cto r

w il

b

16 95 )

d-t yp e

9.5 – 7.5 –

Pull down Cdc42 Total

zizimin1

Figure 2 Expression and function of zizimin1. a, Tissue distribution of zizimin1 mRNA. Northern blot of zizimin1 expression in human tissues with premade poly A+ RNA and 32[P]-labelled probe encompassing the first 450 bases of zizimin1. A single band at 7.5 kilobases (kb) was observed, and no other prominent bands between 0.24 kb and 11 kb were detected (data not shown). b, Activation of Cdc42 by zizimin1. NIH-3T3 cells were transfected with pCDNA3-Cdc42 and pEF4zizimin1-WT, pEF4-zizimin11–1695 or empty pEF4 (Vector), all HA-tagged. Cdc42-GTP was pulled down from cell lysates. The level of the transfected Cdc42 in pull downs or whole-cell extracts was detected by anti-HA western blotting. Densitometric quantification of relative Cdc42 activity (x-fold activation) was calculated from the amount of PBD-bound Cdc42 normalized to the amount of Cdc42 in total cell extracts. Expression of wild-type and truncated zizimin1 was confirmed by anti-HA western blotting (top and bottom arrows, respectively; lower panel). Similar results were obtained in six other experiments with wild-type zizimin1 and two other experiments with zizimin11–1695.

zizimin1 were blocked by coinjection of either CBD (Fig. 3b) or a dominant-negative Cdc42 mutant (N17-Cdc42; data not shown). Inhibiting Rho by Clostridium botulinum C3 exoenzyme or dominant-negative Rho (N19Rho) did not inhibit zizimin1-induced filopodia formation (Fig. 3b and data not shown), nor did dominant-negative Rac1 (N17-Rac; Fig. 3b). These inhibitors of Rho and Rac efficiently blocked stress fibres and membrane ruffles induced by their corresponding GEFs under the same conditions (data not shown). These results indicate that filopodia formation by zizimin1 is mediated by Cdc42. Taken together, the PBD and filopodia-formation assays demonstrate that zizimin1 activates Cdc42 in vivo. Analysis of the zizimin1 amino-acid sequence. The zizimin1 fulllength sequence reveals a protein with 2069 amino acids and a predicted Mr of 236K (Fig. 4). It contains a pleckstrin homology (PH) domain near its amino terminus (predicted by the Pfam program16), and an area with potential coiled-coil structure at its carboxyl terminus (predicted by the Coils and Paircoil programs17,18; Fig. 5b). Interestingly, no DH domain could be identified in the sequence, suggesting that the association with Cdc42 is either by a novel domain or indirect. Database searches indicated that zizimin1 belongs to a superfamily composed of at least four subfamilies (Fig. 5b; Table 1). The zizimin subfamily includes three or four human proteins and the CDM subfamily6 includes three human proteins. A third subfamily (represented by KIAA1395) shows ~30% sequence identity to zizimin1 but lacks a PH domain. A fourth family (represented by KIAA0299) shows ~40% sequence identity to the DOCK180 protein but lacks an SH3 domain. Database searches also identified what are very likely to be zizimin orthologues in Drosophila melanogaster and C. elegans (39% and 26% identical over their entire lengths, respectively). Dictyostelium discoideum and Arabidopsis thaliana each contain a protein that is similar to zizimin1 over its entire length except for the N-terminal and PH domains. The homology between zizimin1 and the CDM proteins is restricted to two regions that we named CZH1 and CZH2 (CDM-zizimin homology domains 1 and 2; Fig. 5b). Additional CZH domains were found in Trypanosoma brucei and Saccharomyces cerevisiae proteins of unknown function (Fig. 5b). A multiple alignment of CZH2 sequences from representative members of the zizimin superfamily is presented in Fig. 5a, revealing residues conserved in multiple subfamilies that span mammals, flies, worms and plants. A consensus sequence was constructed

Table 1 Structure of the zizimin superfamily. Subfamily zizimin

KIAA1395

DOCK180

KIAA0299

Human members

Chromosome

Drosophila homologue

C. elegans homologue

Specificity

zizimin1

13

CG6630

CAB02974

Cdc42

zizimin2

X

zizimin3

2

?

zizimin4*

14

?

KIAA1395

19

KIAA1771

1

FLJ00026

9

DOCK180

10

DOCK2

5

Un-named

8

KIAA0299

3

KIAA0716

7

?

CG11376

AAB37023

? ? ?

Myoblast city

CED-5

Rac Rac ?

CG11754

? ?

The proteins were identified with BLAST. Classification to subfamilies is based on identity scores. Percentage identity of amino-acid sequence between human members within the subfamilies are 50–65%. *Searches of the EST database identified transcripts for all the human genes, apart from zizimin4. Only partial sequence information is available for many of the proteins.

© 2002 Nature Publishing Group NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

641

articles a

b Filopodia per cell

GFP

Wild-type zizimin

Actin

30 25 20 15 10 5 0

Vector

zizimin1

zizimin1 + zizimin1 + CBD C3

zizimin1 + RacT17N

GFP

Control

Actin

c

Figure 3 Filopodia induction by zizimin1. a, NIH-3T3 cells were co-transfected with pEF4-zizimin1 WT (95% of total DNA) plus pEGFP (5% of total DNA). Control cells were transfected with empty pEF vector and pEGFP at the same proportion. GFP was used as a marker to identify expressing cells. Serum-starved cells on glass coverslips were fixed and the actin cytoskeleton stained using rhodaminephalloidin. b, NIH-3T3 cells were microinjected with HA-tagged zizimin1 together with GFP-WASP CBD, GFP-N17-Rac1, GFP-C3 or GFP. Control cells were injected

with GFP and empty vector to keep total DNA constant. Fixed cells were stained with rhodamine-phalloidin and scored for the number of filopodia per cell. The bar graph shows the number of filopodia per cell obtained from at least 50 cells per condition from two to three independent experiments. Values are mean ± s.e.m. c, Top row, rhodamine-phalloidin staining of injected cells; central row, magnification of boxed areas above; bottom row, GFP image of the cells in the top panel.

from the CZH2 domains (Fig. 6b) using the CTREE program (D. J. States Institute for Biomedical Computing, Washington University, St. Louis, MO). A GenBank BLAST search with the consensus sequence detected all of the proteins used to construct the sequence as well as the remaining zizimin family members. Zizimin1 interacts directly with Cdc42 via a new domain. To identify the Cdc42-binding sites in zizimin1, a series of deletion mutants was expressed in Cos-7 cells and their binding to Cdc42 was tested. Deleting the N-terminal portions of zizimin1 up to and including the CZH1 domain did not substantially affect Cdc42 binding. By contrast, C-terminal deletions of half or the entire CZH2 domain completely abolished Cdc42 binding (Fig. 6). In all cases, binding was specific to the nucleotide-depleted state of Cdc42 (data not shown). Importantly, a fragment containing only the CZH2 domain (zizimin11693–end) bound Cdc42 selectively and in a nucleotidedependent manner, albeit with lower efficiency than the longer constructs. This result indicates that the CZH2 domain is both necessary and sufficient for Cdc42 binding. Longer fragments containing sequences N-terminal to the CZH2 (such as zizimin11605–end) bound better than the CZH2 alone, suggesting that the additional sequences may stabilize or modulate Cdc42 binding. Consistent with the CZH2 domain being essential for Cdc42 binding, CZH2deleted zizimin1 (zizmin11–1695) failed to activate Cdc42 (Fig. 2b). To address whether the interaction is direct, bacterially expressed zizimin1 CZH2 domain was prepared. This protein also bound nucleotide-depleted but not GTP-bound Cdc42 (Fig. 7a). Because both proteins were prepared in bacteria, no other eukaryotic protein could mediate the interaction. Full-length zizimin1 expressed and purified from Drosophila S2 cells also bound efficiently to nucleotide-depleted but not GTP-loaded Cdc42 (data not shown). These results show that the CZH2 domain interacts directly with nucleotide-depleted but not GTP-bound Cdc42. GEF activity on Cdc42 in vitro. We noted that intermediate-length N-terminal deletions bound to Cdc42 somewhat better than did

the full-length protein (Fig. 6). Further mutagenesis identified a C-terminal fragment (zizimin11512–end) that is slightly longer than the minimal CZH2 domain but that bound to Cdc42 much better than the minimal sequence (Fig. 7b). When expressed in NIH-3T3 cells, zizimin11512–end had only low activity towards the coexpressed wild-type Cdc42 in the PBD assay (Fig. 7c), but this low activity might result from the lack of the PH domain to target zizimin11512–end to the membrane. We reasoned that cytoplasmic Cdc42 may not be accessible to GEFs because it is bound to Rho-GDI, the chaperone protein that keeps cytoplasmic Cdc42 soluble and inhibits nucleotide exchange19,20. So, to allow zizimin11512–end and Cdc42 to interact, we mutated Cdc42’s ‘CAAX box’, which is the isoprenylation site, to SAAX (Cys 188→Ser). This Cdc42 variant is entirely cytoplasmic but does not bind Rho-GDI. Expression of zizimin11512–end was decreased compared with full-length zizimin1, yet activated SAAX-Cdc42 in the PBD assay to the same extent (Fig. 7c). Zizimin11512–end therefore seems to be an activated mutant. We then used this fragment in GEF assays in vitro. Zizimin11512–end expressed in Cos-7 cells and purified by anti-HA immunoprecipitation was incubated with 3H-GDP-loaded Cdc42 and GDP release was monitored. As shown in Fig. 7d, zizimin11512–end increased the release rate of GDP from Cdc42 to an extent that was comparable to oncogenic, activated Dbl expressed and purified under the same conditions. Staining of the immunoprecipitated material with Coomassie blue dye showed no significant contaminating bands except for a degradation product with an Mr of 60K (Fig. S1 in the Supplementary Information). Attempts to detect GEF activity with bacterially expressed zizimin11512–E were unsuccessful, as the fragment was very poorly expressed and did not bind Cdc42. Although we cannot exclude the possibility that a very active GEF is present at low stoichiometry, the fact that activity is comparable to activated Dbl is inconsistent with this possibility. Together with the direct binding of the CZH2 domain to Cdc42, the results strongly suggest that this domain contains true GEF activity.

© 2002 Nature Publishing Group 642

NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

CZH2

CZH1

PH

articles

Figure 4 Sequence of zizimin1. The amino-acid sequence of human zizimin1. Bold, underlined sequences match the peptides from the purified zizimin protein.

Boxed areas denote the pleckstrin homology (PH) domain and the CZH1 and CZH2 domains.

Discussion

a new GEF for Cdc42. Precedent for a Rho-GEF lacking a DH domain comes from the Salmonella typhimurium protein, SopE, which functions as an efficient GEF for Cdc42 and Rac, even though its primary sequence bears no DH domain21,22. It is well established that CDM proteins are upstream of the Cdc42-related protein Rac in a number of systems7,10–12. DOCK180 immunoprecipitates show GEF activity towards Rac23. DOCK180GEF activity is stimulated by another family of proteins named ELMO, but these do not interact with Rac without DOCK180, suggesting they are modulators of DOCK180 activity. It may be of particular interest that the Arabidopsis genome contains homologues of Rac but completely lacks DH-domain containing proteins24. That Arabidopsis contains a zizimin homologue (Fig. 6) suggests that CZH-containing proteins may be the major Rho activators in plants. Taken together these results prompt us to propose a new superfamily of GEFs for Rho GTPases that use CZH2 sequences as their catalytic domains.

We describe here the cloning of a new protein, named zizimin1, that activates Cdc42 in vivo and in vitro. Zizimin1 binds preferentially to nucleotide-depleted Cdc42 and the binding is reversed by adding GTP. Similar parameters characterize steps in GEF–small-GTPase interactions where the GEFs stabilize the nucleotide-depleted form of the GTPase until GTP then displaces the GEF3. Zizimin1 binds nucleotide-depleted Cdc42 directly via a novel domain we have named CZH2. These domains are present in additional proteins, including DOCK180 and its homologues that have been shown to activate Rac8–11. DOCK180 proteins were also shown to associate with nucleotide-depleted Rac7,8,12, but whether this interaction was direct or was mediated by an associated Dbl protein was not clear. Our data show that a CZH2 domain can mediate a direct interaction with a Rho family protein. In addition, immunoprecipitates of a zizimin1 fragment containing the CZH2 domain (zizimin11512–end) have GEF activity towards Cdc42. These results strongly suggest that zizimin1 is

© 2002 Nature Publishing Group NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

643

articles

a 1

10

20

30

40

50

60

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

...............DEEASMMEDVGMQDVHFNEDVLME..LLEQCADGLWKAERYELIADIYKLIIPIYEKRR.DFERL .........EESAISDDILSPDEEGFCSGKHFTELGLVG..LLEQAAGYFTMGGLYEAVNEVYKNLIPILEAHR.DYKKL ..........................IQSYHFTEAGVIK..ILEKAFALLEKAQLYELLFPFSKIILKYYHATK.SYSRV ................DEQGLKLDAGAQDSQYTEQMLLE..QLKLCADFLDRAERFECLGELYKLILPMYERDR.SYQDL ..............ETEQSQGGDAGSVQPAGFTTDNLGA..KIDKTAAALMLAERFEAVGPLYRLIVPVLEKNM.NFTSL .................EASAAEVEGYGASKLTVDSAVK..YLQLANKLFSQAELYHFCASILELVIPVYKSRK.AYGQL EAAYTLLLHAKLLKWSEDVCVAHLTQRDGYQATTQGQLKEQLYQEIIHYFDKGKMWEEAIALGKELAEQYENEMFDYEQL ...................................................KGKQWECAIDMCRVLARQYEEEIFDYLKL .........................................LYHEILKYFDKGKCWEKGIPLCKELAHLYETRRFDYNKL .....................................................................YES.LYDYQSL .........................................LLKESARNFSRGQKPEKALAVYKDLIKAYDEINYDLNGL ........................................DLMEEAGNLFSKGEDWEDALIVYNQLVPVYQNIIMDYDKL ................ee.........q...ft.d......llee.a..f.kae.ye....vykeiipvye...fdye.l

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

AHLYDTLHRAYSKVTEVMHSGRRLLG.TYFRVAFFGQAAQYQFTDSETDVEGFFEDEDGKEYIYKEPKLTPLSEISQRLL AAVHGKLQEAFTKIMHQSSGWERVFG.TYFRVGFYG...............AHFGDLDEQEFVYKEPSITKLAEISHRLE SHTHKRLGIAADQIKETGEYYENQSD.AWISPLPGIDKRCFGSFFRVAFYGKLFGALNNAEFVYKESAFSKLNEISNRLE AHCYEHLTQAYNKIVEM...................................YFEEDHAIEFVYKEPKLTSLSEISERLA VSVYAELQQTYSRAAEVRSSGKRHLG.AYFRVRFNGE..............RHFGSEHNTDWIYREAGLTSLAAFALEIK AKCHTLLTNIYESILDQESNPIPFIDATYYRVGFYGE...............KFGKLDRKEYVYREPRDVRLGDIMEKLS SELLKKQAQFYENIVKVIRPKP.....DYFAVGYYGQ..............GFPTFLRGKVFIYRGKEYERREDFEARLL AELLNRMALFYEKIIKELRHNS.....EYFRVCFYGR..............GFPRFLQNRVYIFRGKEYERHSDFCARML SEILIQEAKFFQNILTQLRPEP.....EYFRVGFYGM..............GLPLFVRNKQFVYRGLEYERIGAFTQRLQ SWIRKMEASYYDNIMEQQRLEP.....EFFRVGFYGR..............KFPFFLRNKEYVCRGHDYERLEAFQQRML AFVHDQIAGIYTRLQSIDRLVP.....TYFKVSFMGF..............GFPKSLRNKSFVFEGLPFEHITSMHDRLL AGLLQKIAQLYTSISRTERAYF.....YYYLVAFYGQ..............GFPAYLNGHKFVFRSEKLEMHGEFMQRIM a.v.d.laq.ye.i.eq.r.........yfrvgfygq..............gff.fldnkef!yre..yerl.ei.qrll

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

KLY..SDKFGSENV.KMIQDSGKVNPKDLDSKYAYIQVTHVIPFFDEKE..LQERKTEFERSHNIRRFMFEMPFTQTGKR EFY..TERFGDDVV.EIIKDSNPVDKSKLDSQKAYIQITYVEPYFDTYE..LKDRVTYFDRNYGLRTFLFCTPFTPDGRA TFY..TNMYGEGNV.VVLKDSKPVQLEKLNPEKAYIQITFVDVYLSDNE..KMERTTYFTRRNNVNRFYFEAPYTMEGRA KQY..KEKFGADVV.KMIMDSSPVKVDELDAKLAYIQVTHVIPFFSKDE..LDQRLNEFEQNHDVDTFMYETPFTKSGAA EKC..QRQVGHDRV.Q.IEANEQLDLSKIDPTVAYVQITHVEPSIPAAAGIADQHRNDFLVHTNLSEFSYECATIENERK HIY..ESRMDSNHILHIIPDSRQVKAEDLQAGVCYLQITAVDAVMED.....EDLGSRRERIFSLST............. TQFPNAEKMKTTSPPGDDIKNSP.GQYIQCFTVKP...KLDLP.PKFH.RPVSEQIVSFYRVNEVQRFEYSRPIRKGEKN VQHPQAELMQTLEAPGDDITNSD.GQYIQVNKVEP...IMGQAFNKFNDKIINNEIVKYFTANNVQKFQFSRPFRDSTNG TEFPSAQILGNNSPPDNAILNAP.DQYIQISNVRP...VGDAQALKTAMVPVPEKIARFYEVNDVTRFIYDRPMYKGTVD SEFPQAVAMQHPNHPDDAILQCD.AQYLQIYAVTP...IPDYV.DVLQMDRVPDRVKSFYRVNNVRKFRYDRPFHKGPKD RSYHGSNIVHSQEEVDMLLMNPPMGKYIHVASVEPCLSISDNYNSSDKKSSINNKVRMYIENRDLRTFSNSRRLPGAKGV KMYDNPEKIMKTDPCPHLVDSP..GRYIQVFNIDPIGTGCSFENNPEVKPVIKKYFRYYNIQTFEYSKVEERKDTKWTSI .qy...e.mg.dnv.d.iidn.pvd.yilv..vayiqit.vepf..d.e..vderv..fer.nnv..f.yerpftk....

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

...QGGVEEQCKRRTILTAIHC...........FPYVKKRIPVMYQH.HTDLNPIEVAIDEMSKKVAE....LRQLCSSA ...HGELPEQHKRKTLLSTDHA...........FPYIKTRIRVCHRE.ETVLTPVEVAIEDMQKKTRE....LAFATEQD ...QGELAAQYKKRTILTVENS...........FPYIKTRLQVVNRS.VKDFSPIEVAIEDIEKKTRE....LSAAAQHK ...RGSVEEQWKRKTVI................................KKLSPIEVAIDEMQSKVSE....LEEIIL.P VSKEPAIHEQCLKRTVLRVSPSPVSEDSRAATGFPATRRRLPVISVH.FEQFSPLEFACQKLNTKAEQIRKTLNAASNGR ....GSVRARTEGS...................FPALVNRLLVTKSE.SLEFSPVENAIGMIETRTTALRNELEEPRSSD PD.NEFANMWIERTIYTTAYK............LPGILRWFEVKSVF.MVEISPLENAIETMQLTNDKINSMVQQ.HLDD GDRDDVRNLWLERTELRISYP............LPGILRWFPVVETN.TFKISPLERAVEIMKDTNRDIRQLVIL.HKSD KD.NEFKSLWIERTILEIASP............LPGILRWYEVKQKT.MQELTPVEYACEIISNAGKELSELIVQ.YKRD KE.NEFKSLWIERTTLTLTHS............LPGISRWFEVERRE.LVEVSPLENAIQVVENKNQELRSLISQ.YQHK TD......LWVEEYTYHTMNT............FPTLMNRSEIVKVT.KSKLSPLENAIRSLQVKIQELYGLENMCNKTL DPSSEFMRNWLVRRRIKTADS............LPTDLRFTEIVELSDPIYVTPLQNAVEQMRKKNKELNETAASAESNP .d.ne...lq.er.ti.t...............lp.ilrr.evv......elsPi#nAieemq.k..el..lleq....d

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

E...VDM...IKLQLKLQGSVSVQVNAGPLAYARAFLDDTNTK....RYPDNKVKLLKEVFRQFVEACGQALAVNERLIK P...PDA...KMLQMVLQGSVGPTVNQGPLEVAQVFLAEIPED....PKLFRHHNKLRLCFKDFCKKCEDALRKNKALIG N...P.....KMLSMFIQGSIGTTVNQGPLEIANVFLANAMLDDR.GRPVDRLQNKLRLSFRHLQCKAMEAIELCRQLIG P...ADV...KKLQLRLQGSVAVTVNAGPLAYAHAFLDAKVVN....NFSMDRVGDLKDVFRDFIVVCQKALFLNERIIS Q...LDV...KGLQLLLQGAVLPTVNAGPLAYAEVFTKEEQRE....RYGDDGLVKLRESFRNLMNSCQLAIEANASAIG GDHLPRL...QSLQRILQGSVAVQVNSGVLSVCTAFLSGEPAT....RLRSQELQQLIAALLEFMAVCKRAIRVHFRLIG PSL.PIN....PLSMLLNGIVDPAVMGGFANYEKAFFTDRYLQ..EHPEAHEKIEKLKDLIAWQIPFLAEGIRIHGDKVT ETL.HIN....PLSMKLNGIVDPAVMGGFAKYEEAFLTDDYLE..QNPDDKELVEELKELIANQIPLLDLAIQLHRLRAP PKR.NIN....PFSMRLQGTIDANVMGGISKYQEAFFSEQFLKSPQGAGQQANVQRLKVLILEQIQILEQALELHGQLAP QVHGNIN....LLSMCLNGVIDAAVNGGIARYQEAFFDKDYIN..KHPGDAEKITQLKELMQEQVHVLGVGLAVHEKFVH KDHGDVNDLFTELSTNITGTISAPVNGGISQY.KAFLEPSTSK....QFSTDDLGRLTLAFDELVAVLGRCLTLHAELLP NFDLKL......LSRDILGVVSAAVMGGVKNYEVFFTE.......................................... e....in.....lqm.lqGs!...VngGileyeeaFldde..d........d.vn.lke.f.e.i.vleqai.lh..li.

zizimin1 KIAA1395 Celeg.-KIAA1395 DM-zizmin Celeg.-zizmin SPIKE1 DOCK180 MBC Celeg.-KIAA0299 KIAA0299 S.cerevisiae Ced5 consensus>40

E.DQLEYQEEMKANYREMAKELSEIMHEQLG. P.DQKEYHRELERNYCRLREALQPLLTQRL.. E.DQKEYQRNVEENFESFVTHLKPMLSRQ... A.DQKEYHHVLKENYEKLCQALSELLDDE... S.DQQTYHEVLVSSFDAMHERL.......... E.EDQEFHTQLVNGFQSLTAELS......... E.ALRPFHERMEACFKQ.LKEKVEK....... D.SLKALQEHLERCFAD.MQQHVEQ....... S.GVQPLHNRLLERFSQ.LKQSLS........ P.EMRPLHKKLIDQFQM.MRASL......... SKDLKPSHDLLVRLFEENFAEEIERYSRTLSE ................................ e.dqkeyhe.l.enfeel..el.e........

70

80

150

160

220

100

170

230

280

350

90

180

240

290

360

110

300

310

120

190

250

130

200

260

320

140

210

270

330

340

370

© 2002 Nature Publishing Group 644

NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

articles b zizimin1 CC PH

CZH1

CZH2

Drosophila melanogaster zizimin (AAF52524)

39%

Caenorhabditis elegans zizimin (CAB07369)

26%

KIAA1395

30%

SPIKE1 Arabidopsis thaliana

22%

Dictyostelium discoideum zizimin (AAM08471)

26%

*

PH

PH 25%

*

* 25%

KIAA0299

23%

CED-5

23%

23%

SH3

21%

DOCK180

20%

SH3

20%

Myoblast city

23%

SH3

TB292 protein

24%

Trypanosoma brucei

Saccharomyces cerevisiae (NP_013526)

18%

21%

* Figure 5 Structure of zizimin1 and its homologues. a, Multiple sequence alignment of CZH2 domains. CZH2 sequences from representative members of the zizimin superfamily were aligned by the multalin program29. A BLAST search with zizimin1 amino acids 1693–end was used to identify potential CZH2 sequences. The sequences obtained by the search constituted the input for the multiple alignment. Red residues indicate >40% similarity. Red shading indicates identity in all of the sequences. C. eleg, C. elegans; DM, Drosophila melanogaster; MBC, Myoblast city. The sequences of the S. cerevisiae, Drosophila and C. elegans proteins (apart from MBC and CED-5) were predicted by computational translation. The names or GenBank accession numbers for

the proteins are listed in Table 1 and Fig. 5b. Residues that are isoleucine or valine are indicated by ‘!’, and # indicates residues that are any one of asparagine, aspartate, glutamine or glutamate. b, Homology to zizimin1 was estimated using the BLAST tool. Areas with homology to zizimin1 are blue and per cent identity in amino-acid sequence is indicated. Areas with no significant homology are white. Protein names or GenBank accession numbers (in parentheses) are listed. Truncated sequences are indicated by double lines. PH, pleckstrin homology domain; SH3, src homology 3 domain; CZH1 and CZH2, CDM-zizimin homology domains; CC, predicted coiled-coil structure; asterisk, predicted protein.

Methods

coverslips were coated with 20 µg ml−1 human fibronectin and washed three times in PBS. Cells were transfected and allowed to grow for 24 h. They were plated in medium containing 0.2% CS on the fibronectin coverslips for 7 h, then fixed and stained. For microinjection, cells were plated on glass coverslips in growth media. Nuclei were microinjected with 0.2 mg ml−1 pEF4-zizimin1 or empty pEF4 plasmid (control) together with 1 mg ml−1 pEGFP-WASP-CBD, pEGFP-N17-Rac1, pEGFP-C3 or empty pEGFP. Cells were fixed in 3% paraformaldehyde 3.5 h post-injection. They were permeabilized in 0.2% Triton X-100 in PBS and stained with rhodamine-phalloidin 30 min at 37 °C. Images were acquired using a Bio-Rad 1024 MRC laser-scanning confocal-imaging system.

Cell culture and transfections NIH-3T3 cells and Dbl-transformed NIH-3T3 cells25 were grown in Dulbecco’s modified Eagle’s Medium (DMEM) supplemented with 10% calf serum (CS), penicillin and streptomycin, Cos-7 cells were grown in DMEM supplemented with 10% fetal bovine serum (FBS) penicillin and streptomycin, all from Invitrogen (Carlsbad, CA). For transfection, cells in 100 mm tissue-culture dishes at approximately 50% of confluence were transfected with 1.3 µg of total DNA using the Effectene reagent (Qiagen, San Diego, CA) according to the manufacturer’s instructions.

Affinity chromatography GST–PBD pull-down assays GTPase activation assays were performed essentially as described previously14. Transfected NIH-3T3 cells at 24 h post-transfection were incubated in DMEM supplemented with 0.2% CS for 10 h. The cells were washed and lysed on the dish in 50 mM Tris pH 7.2, 500 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 10% glycerol, 10 mM MgCl2 10 µg ml−1 leupeptin and aprotinin, and 1 mM phenylmethylsulphonyl fluoride (PMSF). Cdc42–GTP was pulled down using the PBD of PAK1 immobilized on glutathione beads.

Immunofluorescence Microinjection, staining and confocal microscopy were performed as described previously26. Glass

Cells were lysed in 20 mM HEPES, pH 7.5, 500 mM NaCl, 1% Triton X-100, 10 µg ml−1 leupeptin and aprotinin and 1 mM of each sodium pyrophosphate, NaF, sodium orthovanadate, dithiothreitol (DTT) and PMSF. Lysates were cleared by centrifugation at 13,000g and supplemented with either 5 mM EDTA (nucleotide-depleted condition) or 5 mM MgCl2 (GTP-loaded condition). In certain experiments, lysates were incubated for 1 h with 0.2 mM N-hydroxysuccinimide–biotin (Calbiochem, La Jolla, CA ) to label the proteins, the reaction was quenched using 10 mM Tris pH 7.5 and DTT was added. Bacterially expressed GST–Cdc42 Rac1 or RhoA fusion proteins attached to glutathione agarose beads (Sigma, St Louis, MO) at 1 mg ml−1 protein were incubated for 5 min at 30 °C in 50 mM Tris pH 7.5, 1 mM DTT, 5 mM EDTA with 120 µM GTPγS (GTP-loaded condition) or without nucleotides (nucleotide-depleted condition). 50 µl aliquots of the beads were added to samples containing 1 mg cell lysates in 1 ml volume and incubated 1 h at 4 °C with shaking. After four washes, the beads were

© 2002 Nature Publishing Group NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

645

articles

118 –

Pull-down with nucleotide-depleted 85 – Cdc42 61 –

b Mr(K)

40 – Mr(K) 187 – Total

85 – 61 –

Total

c

Wild-type Cdc42

SAAX Cdc42

Pull down Total fold:

Mr(K) 187 –

118 –

–

Wi

187 – Pull-down with nucleotide- 118 – depleted 85 – Cdc42 61 –

53 –

Rac

Ve cto r

187 –

G

ldtyp ez 15 12 –E izimi n1 Ve cto r Wi ldtyp ez 15 12 –E izimi n1

Cdc42 Nucleotide loading: –

Fu ll-le 15 ngth 12 ziz – mi 16 E n1 05 –E

Mr(K)

Fu ll-l 28 engt hz 8– izm 63 E in1 9– 92 E 3– 16 E 05 16 –E 93 1– –E 18 7 1– 8 16 9 1– 5 92 1– 2 53 9 16 05 16 –E 93 –E

a

a

1 2.7 1.2

1 4.9 3.6

Wild-type zizimin1

118 – 85 –

1512–E

61 –

53 – 40 –

d Control Zizimin1 Dbl PH

CZH1

CZH2

Full-length zizmin1

CZH1

CZH2

288–E

CZH1

CZH2

639–E

CZH2

923–E

CZH2

1605–E

CZH2

1693–E

PH

CZH1

PH

CZH1

CZ

1–1878

GDP-bound Cdc42 (%)

b 100 90 80 70 60 50 0

5

10

15

Time (min)

1–1695

Figure 7 Binding and activation of Cdc42 by the C2H2 domain. a, Direct binding of zizimin11693–end to Cdc42. Soluble His6-tagged zizimin11693-end purified from PH 1–539 E. coli was incubated with nucleotide-depleted or GTPγS-loaded GST–Cdc42 or Rac beads. Binding was performed as in Fig. 1e. Binding was detected by western blotting using an antibody raised against the zizimin1 C terminus. The amount of bound Figure 6 The zizimin1 CZH2 domain is necessary and sufficient for Cdc42 zizimin11693–end was estimated to be 5% by comparing the amount bound with total binding. Zizimin1 fragments were expressed in Cos-7 cells and Cdc42 binding was zizimin11693–end. Similar results were obtained in two other experiments. tested as in Fig. 1e. a, Top panel: pull-down assay using a nucleotide-depleted b, Zizimin11512–end binding to nucleotide-depleted Cdc42. The experiment was perCdc42 column. Bottom panel: total cell extracts containing one ninth the amount of formed as in Fig. 6a. Similar results were obtained in two other experiments. cell lysate used for the pull-down assay. Binding of the zizimin1 fragments c, Activation of the SAAX Cdc42 mutant by zizimin11512–end. NIH-3T3 cells were trans1605–end and 1693–end was tested at two different amounts, either similar to the fected with wild-type Cdc42 or the C188S (SAAX) Cdc42 mutant in pCDNA3, togethother fragments or with threefold more (the last two lanes). None of the proteins er with HA-tagged zizimin1-WT, zizimin11512-end (1512–E) or empty vector. Cdc42-GTP binds GTPγS-loaded Cdc42 (data not shown). b, zizimin1 fragments used in the was pulled down from cell lysates as in Fig. 2b. The level of transfected Cdc42 in experiments. Similar results were obtained in three additional experiments. pull downs or whole-cell extracts was detected by western blotting with anti-HA. Densitometric quantification of Cdc42 activity relative to control was calculated from the amount of PBD-bound Cdc42 normalized to the amount of Cdc42 in total cell eluted for 5 min at room temperature in 50 mM Tris pH 7.5, 5 mM MgCl2, 1 mM DTT supplemented extracts. Expression of wild-type zizimin1 and zizimin11512–end (1512–E) was conby 40 µM of the indicated nucleotides. In some cases a second (consecutive) elution was performed firmed by anti-HA western blotting (lower panels). Similar results were obtained in using 20 mM reduced glutathione in 50 mM Tris pH 7.5. The eluates were resolved by SDS–polyacrylamide gel electrophoresis (PAGE) and analysed by western blotting. two other experiments. d, GEF activity of zizimin11512–end. 3H-GDP-loaded Cdc42 (at 100 nM) was incubated with immunoprecipitated zizimin11512–end (zizimin1; 56 nM), GEF assays onco-Dbl (Dbl; 47 nM) or control immunoprecipitation from non-transfected cells GST–Cdc42 expressed in Escherichia coli was purified on glutathione agarose beads (Sigma) and eluted (control). At the indicated time points, aliquots were removed and quantified by filtrausing free glutathione. For loading with 3H-GDP, Cdc42 at 500 nM in 50 mM Tris pH 7.5, 20 mM KCl, tion on nitrocellulose and liquid scintillation counting. Values at time zero of each 3 −1 2.5 mM EDTA, 5 µM H-GDP (12–14 Ci/mmol; Amersham-Pharmacia), 1 mM DTT and 1 mg ml sample were defined as 100%. Values are means of duplicate samples ± range. BSA was incubated for 8 min at 30 °C. To stabilize nucleotide binding, MgCl2 was added to a final conSimilar results were obtained in two other experiments. centration of 20 mM, incubated for 3 min at 30 °C and then placed on ice. HA-tagged oncogenic Dbl © 2002 Nature Publishing Group PH

646

CZH1

1–922

NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

articles or zizimin11512–E were expressed in Cos-7 cells and purified by anti-HA immunoprecipitation using protein G sepharose beads (Amersham Pharmacia Biotech, Piscataway, NJ). Precipitated Dbl or zizimin1 at the indicated concentrations were suspended in 50 mM Tris pH 7.5, 20 mM KCl, 5 mM MgCl2, 1 mM DTT, 1 mg ml−1 BSA and 125 µM GTPγS on ice. The nucleotide-exchange reaction was initiated by adding 3H-GDP-loaded Cdc42 (20% of the total volume) and transfer to 30 °C. Final concentrations in the reaction were 100 nM Cdc42, 1 µM 3H-GDP, 100 µM GTPγS, 0.5 mM EDTA, 8 mM MgCl2, 50 mM Tris pH 7.5, 20 mM KCl, 1 mM DTT and 1 mg ml−1 BSA. At the indicated time points, aliquots were removed, filtered on nitrocellulose and quantified by liquid-scintillation counting.

Cloning of zizimin1 NIH-3T3 cell lysate (80 mg) prepared as described above was supplemented with 5 mM EDTA and incubated with 5 mg nucleotide-depleted Cdc42 on beads for 1 h at 4 oC. After washing, the column was eluted for 5 min at room temperature in 1 ml of 50 mM ammonium carbonate pH 7.5, 1 mM MgCl2 and 10 µM GTPγS. The eluate was concentrated using a SpeedVac concentrator, and resolved by using SDS–PAGE. Proteins were visualized by ultraviolet illumination using SYPRO Ruby protein gel stain (Molecular Probes, Eugene, OR), according to the manufacturer’s protocol. The p220–zizimin1 protein band was cut from a gel and analysed by tryptic digestion and mass-spectrometry-based microsequencing of the derived peptides (Harvard Microchemistry Facility, Cambridge, MA). Sequences of 13 peptides were obtained and used to search the database at the NCBI (National Center for Biotechnology Information) using BLAST (Basic Local Alignment Search Tool). Three peptides matched an N-terminal truncated, 1534 amino-acid human protein sequence, named KIAA1058 (GenBank AB028981), which corresponds to the C-terminal two thirds of zizimin1. KIAA1058 was cloned from an adult brain cDNA library at the Kazusa DNA Research Institute (Yana Kisarazu, Chiba 292, Japan)27. Its nucleotide sequence contains 1260 bp 3′ non-coding sequence. Another peptide matched two human expressed sequence tag (EST) clones of different origin (GenBank AW070874: pooled fetal lung, testis, and B cell; GenBank W23667: fetal lung). The EST clones were obtained and sequenced, and they contained identical sequences, starting at the same position coding for the zizimin1 5′ untranslated region (UTR), translation start and an open reading frame (ORF). AW070874 was used to construct full-length zizimin1. To obtain sequence connecting AW070874 and KIAA1058, the polymerase chain reaction (PCR) was performed using human spinal-cord cDNA library as template DNA (#HL50011b, Clontech Inc., Palo Alto, CA) and the primers 5′-GCTGAGTGTGTTTTGGTATGCAGGGCAC-3′, and 5′-CAATGATTTATCTTTCAATTTGCAATGCTG-3′. A single product of ~900 bp was obtained, which was subcloned and four clones were sequenced, yielding identical results. This intermediate fragment, AW070874 and KIAA1058 were combined to construct the full-length zizimin1 using the endogenous DraII (1166) and XbaI (1753) restriction sites. The nine remaining peptides were derived from a second zizimin isoform.

DNA constructs D118A Cdc42 was obtained from R. A. Cerione (Department of Molecular Medicine, College of Veterinary Medicine, Cornell University, Ithaca, NY), and the Cdc42 insert was subcloned in pGEX-2T (Pharmacia) using the BamHI and EcoRI sites. Full-length and C-terminal truncated zizimin1 constructs. The HA tag (YPYDVPDYA) spaced by two glycine residues was added to zizimin1 on its N terminus. HA-zizimin1 (full-length or C-terminal truncations) were cloned into pEF4-myc-his-C (Invitrogen) using KpnI and NotI sites (see Supplementary Information). N-terminal truncated zizimin1 constructs. The HA tag, followed by a KpnI site, was cloned into pEF4myc-his-C KpnI site, to generate pEF4-HA-KpnI plasmid. DNA inserts coding for N-terminal truncated zizimin1 fragments were cloned into pEF4-HA-Kpn using the KpnI and NotI sites (see Supplementary Information). Bacterially-expressed zizimin11693-end. The insert coding for Zizimin11693-end was excised from pEF4- zizimin11693-end using KpnI and SfuI and subcloned in to pRSETB (Invitrogen) using these sites. The plasmid was transformed into E. coli BL21(DE3)pLysS cells (Novagen, Madison, WI), induced 8 h at room temperature with 1 mM isopropyl-β-D-thioglacatoside (IPTG) and purified on Nickel–nitrilo tri-acetic (NTA) agarose beads (Qiagen) according to the manufacturer’s protocol. Onco-Dbl. To generate HA-tagged onco-Dbl, oncogenic Dbl (GenBank AAA52172) was obtained from S. Gutkind. Onco-Dbl was amplified by using PCR, cloned into the pEF-HA-KpnI vector using the KpnI and BamHI sites and sequenced. C188S Cdc42. Cys 188 of wild-type Cdc42 (HA-tagged; cloned in pCDNA3; Invitrogen) was mutated to serine. GFP-botulinum-C3 has been described previously28. CBD. The sequence coding for hWASP amino acids 201–321 was amplified by PCR and cloned in pEGFP-C1 (Clontech) using the KpnI and BamHI sites.

Reagents Premade poly A+ RNA northern blot was from Clontech (cat# 7780-1). Rabbit anti-Dbl antibody was obtained from C. J. Der (Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, NC). For anti-zizimin1 antibody, rabbits were immunized against peptide bearing the last 12 amino acids using standard procedures. The derived sera were affinity purified on columns conjugated to the same peptide. RECEIVED 28 JANUARY 2002; REVISED 7 JUNE 2002; ACCEPTED 12 JULY 2002 ; PUBLISHED 31 JULY 2002.

1. Bishop, A. L. & Hall, A. Rho GTPases and their effector proteins. Biochem. J. 348, 241–255 (2000).

2. Aspenstrom, P. Effectors for the Rho GTPases. Curr. Opin. Cell Biol. 11, 95–102 (1999). 3. Cherfils, J. & Chardin, P. GEFs: structural basis for their activation of small GTP-binding proteins. Trends Biochem. Sci. 24, 306–311 (1999). 4. Cerione, R. A. & Zheng, Y. The Dbl family of oncogenes. Curr. Opin. Cell Biol. 8, 216–222 (1996). 5. Hart, M. J. et al. Cellular transformation and guanine nucleotide exchange activity are catalyzed by a common domain on the dbl oncogene product. J. Biol. Chem. 269, 62–65 (1994). 6. Wu, Y. C. & Horvitz, H. R. C. elegans phagocytosis and cell-migration protein CED-5 is similar to human DOCK180. Nature 392, 501–504 (1998). 7. Kiyokawa, E. et al. Activation of Rac1 by a Crk SH3-binding protein, DOCK180. Genes Dev. 12, 3331–3336 (1998). 8. Nishihara, H. et al. Non-adherent cell-specific expression of DOCK2, a member of the human CDM-family proteins. Biochim. Biophys. Acta 1452, 179–187 (1999). 9. Kiyokawa, E., Hashimoto, Y., Kurata, T., Sugimura, H. & Matsuda, M. Evidence that DOCK180 upregulates signals from the CrkII-p130(Cas) complex. J. Biol. Chem. 273, 24479–24484 (1998). 10. Fukui, Y. et al. Haematopoietic cell-specific CDM family protein DOCK2 is essential for lymphocyte migration. Nature 412, 826–831 (2001). 11. Reddien, P. W. & Horvitz, H. R. CED-2/CrkII and CED-10/Rac control phagocytosis and cell migration in Caenorhabditis elegans. Nature Cell Biol. 2, 131–136 (2000). 12. Nolan, K. M. et al. Myoblast city, the Drosophila homolog of DOCK180/CED-5, is required in a Rac signaling pathway utilized for multiple developmental processes. Genes Dev. 12, 3337–3342 (1998). 13. Hart, M. J. et al. Identification of a novel guanine nucleotide exchange factor for the Rho GTPase. J. Biol. Chem. 271, 25452–25458 (1996). 14. Del Pozo, M. A., Price, L. S., Alderson, N. B., Ren, X. D. & Schwartz, M. A. Adhesion to the extracellular matrix regulates the coupling of the small GTPase Rac to its effector PAK. EMBO J. 19, 2008–2014 (2000). 15. Tran Van Nhieu, G., Caron, E., Hall, A. & Sansonetti, P. J. IpaC induces actin polymerization and filopodia formation during Shigella entry into epithelial cells. EMBO J. 18, 3249–3262 (1999). 16. Bateman, A. et al. The Pfam protein families database. Nucleic Acids Res. 28, 263–266 (2000). 17. Berger, B. et al. Predicting coiled coils by use of pairwise residue correlations. Proc. Natl Acad. Sci. USA 92, 8259–8263 (1995). 18. Lupas, A. Prediction and analysis of coiled-coil structures. Methods Enzymol. 266, 513–525 (1996). 19. Olofsson, B. Rho guanine dissociation inhibitors: pivotal molecules in cellular signalling. Cell Signal. 11, 545–554 (1999). 20. Keep, N. H. et al. A modulator of the Rho family G proteins, RhoGDI, binds these G proteins via an immunoglobulin-like domain and a flexible N-terminal arm. Structure 5, 623–633 (1997). 21. Hardt, W. D., Chen, L. M., Schuebel, K. E., Bustelo, X. R. & Galan, J. E. S. typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 93, 815–826 (1998). 22. Rudolph, M. G. et al. Biochemical analysis of SopE from Salmonella typhimurium, a highly efficient guanosine nucleotide exchange factor for RhoGTPases. J. Biol. Chem. 274, 30501–30509 (1999). 23. Gumienny, T. L. et al. Ced-12/elmo, a novel member of the crkII/dock180/rac pathway, is required for phagocytosis and cell migration. Cell 107, 27–41 (2001). 24. Valster, A. H., Hepler, P. K. & Chernoff, J. Plant GTPases: the Rhos in bloom. Trends Cell. Biol. 10, 141–146 (2000). 25. Khosravi-Far, R. et al. Dbl and Vav mediate transformation via mitogen-activated protein kinase pathways that are distinct from those activated by oncogenic Ras. Mol. Cell. Biol. 14, 6848–6857 (1994). 26. Kiosses, W. B., Daniels, R. H., Otey, C., Bokoch, G. M. & Schwartz, M. A. A role for p21-activated kinase in endothelial cell migration. J. Cell Biol. 147, 831–844 (1999). 27. Kikuno, R. et al. Prediction of the coding sequences of unidentified human genes. XIV. The complete sequences of 100 new cDNA clones from brain which code for large proteins in vitro. DNA Res. 6, 197–205 (1999). 28. Del Pozo, M. A., Vicente-Manzanares, M., Tejedor, R., Serrador, J. M. & Sanchez-Madrid, F. Rho GTPases control migration and polarization of adhesion molecules and cytoskeletal ERM components in T lymphocytes. Eur. J. Immunol. 29, 3609–3620 (1999). 29. Corpet, F. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res. 16, 10881–10890 (1988).

ACKNOWLEDGMENTS We thank T. Nagase from the Kazusa DNA Research Institute for providing the KIAA1058 cDNA. We thank C. Der for providing Dbl expressing NIH-3T3 cells and R. Cerione for the D118A Cdc42. We are grateful to J. Han for advice about GEF assays and to N. Alderson for excellent technical assistance. This work was supported by grants from the U.S. Public Health Service (RO1 GM41721) to M.A.S., grants from the American Heart Association to N.M. and W.B.K. and from the Leukemia & Lymphoma society to M.A.P. Correspondence and requests for information should be addressed to M.A.S. Supplementary Information accompanies the paper on www.nature.com/naturecellbiology.

COMPETING FINANCIAL INTERESTS The authors declare that they have no competing financial interests.

© 2002 Nature Publishing Group NATURE CELL BIOLOGY VOL 4 SEPTEMBER 2002 www.nature.com/naturecellbiology

647