MgcRacGAP, A New Human GTPase-activating Protein for Rac and Cdc42 Similar to Drosophila rotundRacGAP Gene Product, Is Expressed in Male Germ Cells*

In a search for new partners of the activated form of Rac GTPase, we have isolated through a two-hybrid cloning procedure a human cDNA encoding a new GTPase-activating protein (GAP) for Rho family GTPases. A specific mRNA of 3.2 kilobases was detected in low abundance in many cell types and found highly expressed in testis. A protein of the predicted size 58 kDa, which we call MgcRacGAP, was detected in human testis as well as in germ cell tumor extracts by immunoblotting with antibodies specific to recombinant protein. In vitro, the GAP domain of MgcRacGAP strongly stimulates Rac1 and Cdc42 GTPase activity but is almost inactive on RhoA. N-terminal to its GAP domain, MgcRacGAP contains a cysteine-rich zinc finger-like motif characteristic of the Chimaerin family of RhoGAPs. The closest homolog of MgcRacGAP is RotundRacGAP, a product of the Drosophila rotund locus. In situhybridization experiments in human testis demonstrate a specific expression of mgcRacGAP mRNA in spermatocytes similar to that of rotundRacGAP in Drosophila testis. Therefore, protein sequence similarity and analogous developmental and tissue specificities of gene expression support the hypothesis that RotundRacGAP and MgcRacGAP have equivalent functions in insect and mammalian germ cells. Since rotundRacGAP deletion leads to male sterility in the fruit fly, the mgcRacGAP gene may prove likewise to play a key role in mammalian male fertility.

cently been described (4). GTPase-activating proteins (GAPs) 1 for Rho GTPases constitute a class of regulatory proteins that can bind activated (or GTP-bound) forms of Rho GTPases and stimulate GTP hydrolysis (5,6). Through this catalytic function, Rho GAPs negatively regulate Rho-mediated signals. However, it has long been speculated that GAPs may also serve as effector molecules and play a role in signaling downstream of Rho and other Ras-like GTPases (5,6). This idea has been recently substantiated in the case of n-chimaerin, a Rho GAP mainly expressed in the central nervous system. More specifically, n-chimaerin was shown to cooperate with Rac1 and Cdc42 in inducing specific changes in cytoskeletal morphology i.e. the formation of lamellipodia and filopodia, respectively (7). Interestingly, this effect of n-chimaerin requires G protein binding capacity but not GAP activity; it also requires the non-GAP N-terminal portion of the protein (7), which includes a cysteine-rich region, shared by all chimaerins, and a specific N-terminal extension (5,8,9).
These data support the idea that members of the Chimaerin family may mediate Rho-dependent signals in various cell types, as most of them show tissue-specific expression (5,9). Analysis of these Rho GAPs may therefore help in defining Rho-dependent functions in differentiated cells.
We report here the cloning and characterization of a new human chimaerin-like protein showing GAP activity toward Rac and Cdc42 and mainly expressed in male germ cells that we call MgcRacGAP (Male Germ Cell RacGAP). It shows a strong structural similarity with the Drosophila protein Rotund (Rn)RacGAP, the product of a spermatocyte-specific gene essential for male fertility in the fruit fly (10). We propose that MgcRacGAP operates in a Rac/Cdc42-dependent signaling pathway involved in the development and/or function of male germ cells in mammals.

EXPERIMENTAL PROCEDURES
cDNA Cloning-The two-hybrid system used in this study has already been described (11,12). The Jurkat cell cDNA library in plasmid pGAD was generously provided by Dr. J. Camonis (13). The bait in our screen was a fusion between LexA and a human Rac2L61 mutant previously described (14). This fusion protein was expressed in L40 yeast from plasmid pVJL10, a derivative from pBTM116 (12) with modified polylinker. LexA fusion protein expression was checked on Western blots with anti-LexA antibodies (a gift from P. Moreau, Gif/Yvette).
Yeast culture conditions and the two-hybrid procedures were carried out according to published methods (12,15). Library plasmids from transformed yeast colonies were recovered using HB101 as a recipient * This work was supported in part by INSERM and by grants from the Ligue Nationale Contre le Cancer and the Association pour la recherche sur le cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ A recipient of a fellowship from the Ministère de l' éducation nationale, de  To obtain the full-length sequence of the mgcRacGAP cDNA, overlapping cDNA clones were isolated from a human placenta ExLox cDNA library (kindly provided by J. M. Garnier and P. Chambon) using a 450-bp EcoRI 5Ј fragment of the two-hybrid clone as a probe.
Sequence Analysis-Clones of interest were sequenced using the Sanger dideoxy termination method adapted to ABI 373A automated sequencer. Sequence analysis was performed using computer facilities provided by the Centre Interuniversitaire de Traitement Informatique (CITI2) and Groupement d'Intérêt Scientifique (GIS) Infobiogen.
Gene expression was analyzed on Multiple Tissue Northern blots (CLONTECH) using mgcRacGAP cDNA as a probe. A ␤-actin cDNA was used as a control.
Immunodetection of MgcRacGAP-Antibodies to recombinant MgcRacGAP (amino acids 238 -513) were raised in chicken, affinitypurified on a MgcRacGAP/Affi-Gel column (Bio-Rad), and used to detect MgcRacGAP protein by Western blot. Proteins from human testis extracts were resolved on 12% SDS-polyacrylamide gel electrophoresis and electroblotted onto nitrocellulose. Filters were incubated sequentially with anti-MgcRacGAP-purified antibodies and peroxydase-labeled rabbit anti-chicken immunoglobulins (Dako). Blots were developed using the ECL detection system (Amersham Corp.).
GTP Hydrolysis Assays-The [␥ 32 P]GTP-bound form of small GTPases were prepared by incubating 10 pmol of protein with 2 Ci of [␥-32 P]GTP (30 Ci/mmol, NEN Life Science Products) in a 50-l volume of 25 mM Tris, pH 7.5, 5 mM EDTA, 0.2 mM MgCl 2 , 0.1 mg/ml bovine serum albumin, 1 mM dithiotreitol, 4 M GTP for 15 min at room temperature. GTP hydrolysis was initiated by raising MgCl 2 and GTP to a final concentration of 20 mM and 200 M, respectively, and was stopped after 3 min by the addition of 2 ml of ice-cold 50 mM Tris, pH 8, 35 mM MgCl 2 , 1 mM dithiotreitol, 150 mM NaCl. Amounts of [␥-32 P]GTP bound to GTPases were determined by radioactivity counting after rapid vacuum filtration of the samples on BA 85 nitrocellulose (Schleicher and Schuell). Typically, 100% of [␥-32 P]GTP bound to GTPases is around 20,000 cpm. GAP assays were conducted in the presence of 20 pmol of MgcRacGAP or Bcr GAP domains during the GTP hydrolysis step, and similar data were obtained at least three times. In an additional set of experiments, substrate specificity of MgcRacGAP toward Rho GTPases was determined as described previously (16), using decreasing amounts (20 pmol to 0.4 pmole) of GAP domain during GAP assays; measurements were done in duplicate after 5 min of GTP hydrolysis. GTP exchange reactions during the GTP hydrolysis step were checked using [␣-32 P]GTP-preloaded GTPases under similar conditions and were found not to be significantly affected by the presence of MgcRacGAP.
In Situ Hybridization Analysis-An 35 S-UTP-labeled antisense mgcRacGAP cRNA probe was generated by linearizing with KpnI a pExLox plasmid containing a 1837-bp insert of mgcRacGAP cDNA and transcribing with SP6 RNA polymerase. An SGP-2 antisense RNA probe used in control experiments was synthesized by SP6 transcription of a 1,197-bp subclone of rat SGP-2 cDNA in the pGEMblue vector, kindly donated by Dr. Michael W. Collard of Utah State University.
6-m paraffin sections of formalin-fixed adult human testis were hybridized overnight at 55°C, washed as described (17) and exposed to Ilford K5 emulsion for 6 days. After development of the emulsion, the sections were stained with toluidine blue.

RESULTS AND DISCUSSION
In a search for new targets of Rac GTPase, we have used a two hybrid procedure in yeast (11,12) to screen a Jurkat cDNA library (13) with a constitutively activated form of Rac. RacL61 variant rather than RacV12 was chosen as a bait, since it had been shown previously to interact with a higher affinity with target and GAP molecules (14,18). A 1,981-bp cDNA potentially encoding a new Rho GAP protein was isolated. Screening of a cDNA library from human placenta with the 5Ј end of this cDNA as a probe allowed the isolation of an overlapping 1,837-bp cDNA insert. Combination of the two sequences yielded a 2,840-bp sequence, showing an open reading frame of 1,581 bp. This open reading frame starts with an ATG codon in a Kozak consensus sequence (19,20) and encodes a putative protein of 527 amino acids with a molecular mass of 58,277 kDa. Stop codons are found in all three reading frames upstream of the start codon. In addition to the three blocks conserved among Rho GAPs (Fig. 1), the predicted protein exhibits a large N-terminal region including a cysteine-rich sequence with a C6H2 structure resembling a zinc finger and previously described in several isoforms of protein kinase C (PKC) (21).Therefore, this protein, which we call MgcRacGAP, appears clearly related to chimaerins that consist of a Rho GAP domain coupled to a PKC-like cysteine-rich motif with variable N-terminal sequences (8,21).
FASTA and BLAST programs were used to search for homologies in protein data banks. By both methods, the best scores were obtained with Drosophila RnRacGAP and human and rat n-and ␤-chimaerins. Despite the phylogenetic distance between human and Drosophila, the closest homolog to human MgcRacGAP is the Drosophila RnRacGAP protein encoded by the rotund locus (10). The two proteins show 40% identity (60% similarity) in the region encompassing the cysteine-rich and GAP domains (Fig. 2), whereas identity scores of MgcRacGAP or RnRacGAP with other mammalian members of the Chimaerin family in this region do not exceed 30%.
In PKC and n-chimaerin, cysteine-rich domains have been shown to bind phorbol esters and to mediate lipid-dependent regulation of enzymatic activity (21,22). Comparative analysis of cysteine-rich domains in those proteins have indicated a 14-amino acid consensus sequence for phorbol ester binding (23,24). However, low matching scores for both MgcRacGAP (10/14) and RnRacGAP (9/14) make them unlikely as phorbol ester receptors. Moreover, in a putative loop (amino acids 198 -210 in MgcRacGAP) of zinc finger-like domains, MgcRacGAP and RnRacGAP exhibit very similar basic sequences not found in other mammalian chimaerins (see Fig. 2B). This may indicate a conserved and specific function for this region in both proteins. Analysis of N-terminal sequences of MgcRacGAP and RnRacGAP did not reveal an SH2 domain as was found in ␣2 and ␤2 isoforms of chimaerins (25,26).
To confirm that the protein encoded by mgcRacGAP cDNA actually displays a GAP activity, we expressed the GAP domain (amino acids 238 -513) of MgcRacGAP in E. coli and examined its ability to stimulate Rho GTPases. The GAP domain of MgcRacGAP exhibited strong GAP activities toward Cdc42 and Rac1 (or Rac2) but appeared much less active on RhoA (Fig. 3). In another set of experiments, under rate-limiting conditions of GAP concentration, MgcRacGAP was estimated to be equally active on Rac1 and Cdc42, although being 30 times less active on RhoA (Table I). In addition, MgcRacGAP did not significantly stimulate GTPase activity in Rap2a nor in Rab6 (Fig. 3). Non-chimaerin mammalian Rho GAPs such as Bcr and Abr proteins have been previously reported to have the same specificity as MgcRacGAP toward Cdc42 and Rac (5,27); on the contrary, MgcRacGAP appears functionally distinct from other mammalian chimaerins, which are active primarily on Rac (8,9,28). The substrate specificity of RnRacGAP toward either mammalian or Drosophila Rho GTPases has not yet been reported.
Northern blot analysis revealed that an ϳ3.2-kilobase mgcRacGAP transcript is present in low abundance in most tissues and is highly expressed in testis (Fig. 4A). Antibodies to the GAP domain (amino acids 238 -513) of MgcRacGAP were raised in chicken, affinity purified, and found to recognize recombinant GAP domain (Fig. 4B) without showing crossreaction with two other recombinant mammalian RhoGAPs, Bcr and RLIP (29) (not shown). Western blot experiments using these antibodies allowed the detection of a major protein of ϳ60 kDa in testis extracts (Fig. 4B). This protein was not present in detectable amounts in several other tissues tested (except for a  trace amount in fetal muscle) (not shown), thus confirming at the protein level that MgcRacGAP is preferentially expressed in testis. MgcRacGAP was also detected at a high level in human germ cell tumors (Fig. 4B).
In situ hybridization of an antisense mgcRacGAP cRNA probe to sections of mature human testis showed that expression of mgcRacGAP mRNA was restricted to germ cells, as the grain density over somatic cells was not detectably above background (Fig. 5). This was in contrast to control experiments on sections from the same samples, in which a probe for the Sertoli cell marker SGP-2 (sulfated glycoprotein 2) specifically labeled the Sertoli cells but not the germ cells (Fig. 5). Moreover these experiments demonstrated that the level of mgcRacGAP transcript varied dramatically and reproducibly among the different types of germ cells present, indicating that expression of mgcRacGAP mRNA is quantitatively modulated during the process of spermatogenesis. The highest level of mgcRacGAP expression was found in spermatocytes, with a residual low level detectable in some spermatid stages. Grain densities over spermatogonia and mature spermatozoa were not above background. The developmental-and tissue-specific expression of the mgcRacGAP gene further supports the view of a close similarity with the rnRacGAP gene that was found by in situ hybridization to Drosophila testis specifically expressed in primary spermatocytes (10).
The sequence data reported here indicate that the mgcRac-GAP gene encodes a Rho GAP belonging to the Chimaerin family of proteins, one most similar to Drosophila RnRacGAP. Although mgcRacGAP mRNA can be detected in many tissues, it is preferentially expressed in germ cells, specifically in spermatocytes. Taken together, these features distinguish MgcRac-GAP from other known mammalian chimaerins and conversely, show a striking resemblance to Drosophila RnRacGAP, thus suggesting that MgcRacGAP is a mammalian homolog of rnRacGAP gene product.
In Drosophila, RnRacGAP is encoded by one of two transcripts of the rotund locus (10,30). Genetic analyses of the rn locus have defined the rn null mutant phenotype as male and female sterility and an absence of structures in the subdistal part of the appendages (31). It was recently shown that a RnRacGAP transgene can rescue male sterility but not appendage morphogenetic function nor female sterility in flies deficient for the rn locus (32), 2 implying that RnRacGAP is essential to development and/or function of male germ cells. Assuming that MgcRacGAP and RnRacGAP operate in equivalent pathways in mammalian and insect spermatocytes, respectively, the mgcRacGAP gene may prove to play a key role in spermatogenesis and male fertility in mammals as well.
Recent studies with n-chimaerin suggest that Rho GAPs of this family act as effectors for Rho GTPases and that this function requires the N-terminal non-GAP portion of the protein (7). Therefore, isolation of partners of the MgcRacGAP 2 R. Griffin, personal communication. Results obtained with the two specific probes are shown at 20 ϫ magnification under bright-field and dark-ground illumination. Using the mgcRac-GAP-specific probe, the density of silver grains (black dots in bright field and white dots in dark field) was insignificant over the interstitial (intertubular) tissue and outer perimeter of the tubules indicating low or undetectable expression of mgcRacGAP mRNA in somatic cells. In the germ cell population, strong signals were evident over spermatocytes (arrows), and weaker expression was detected in some spermatids (arrowheads) (spermatocytes were identified by their large round nuclei, which stained weakly with toluidine blue compared with the spermatids, which have smaller, more strongly staining nuclei). As expected, SGP-2-specific probe showed a strong staining within tubules corresponding to Sertoli cells. N-terminal region in germ cells should provide new insights into Rac/Cdc42 signaling pathways in those cells.