The ras superfamily proteins

Biochimie 70 (1988) 865- 868 (~) Soci~t6 de Chimie biologique / Elsevier, Paris

865

The r a s superfamily proteins Pierre C H A R D I N I N S E R M U-248, Facult6 de M6decine Lariboisi~re-Saint-Louis, 10 avenue de Verdun, 75010 Paris, France (Received 16-2-1988, accepted 18-3-1988)

Summary m

Several recent discoveries indicate that the ras genes, frequently activated to a transforming potential in some human tumours, belong to a large family that can be divided into-three main branches: the first branch represented by the ras, ral and rap genes; the second branch, by the rho genes; and the third branch, by the tab genes. The C-terminal end of the encoded proteins always includes a cystein, which may become fatty-acylated, suggesting a sub-membrane localization. The ras superfamily proteins share four regions of high homology corresponding to the GTP binding site; however, even in these regions, significant differences are found, suggesting that the various proteins may possess slightly different biochemical properties. Recent reports show that some of these proteins play an essential role in the control of physical processes such as cell motility, membrane ruffling, endocytosis and exocytosis. Nevertheless, the characterization of the proteins directly interacting with the ras or ras-related gene-products will be required to precisely understand their function. ms-related / GTP binding site / sub-membrane protein / oncogene / growth-factor

Extensive screenings for transforming sequences in human tumour DNA, using the N I H / 3 T 3 focus assay, have repeatedly led to isolation of one of three closely related genes: H-ras, K-ras (first found as the oncogenes of the Harvey and Kirsten sarcoma viruses) and N-ras. The three ras genes code for 21 kDa G T P / G D P ' binding proteins possessing a weak GTPase activity and transiently attached to the inner leaflet of the plasma membrane upon reversible c-terminal fatty-acylation. Acquisition of a transforming potential is usually due to a point mutation leading to the substitution of amino acids 12, 13 or 61. Expression of such activated proteins leads to an increase of membrane ruffling and cell motility, and to the loss of contact inhibition. Ras proteins have been highly conserved through evolution and are found in organisms as different as yeast and man, where they certainly play an essential role; however, in spite of intense research efforts, the precise biochemical function of the ras proteins is not yet known [1]. 'The fortuitous discovery of two proteins: rho

[2] and YPT [3] (from the marine snail Aplysia und baker's yeast, respectively) sharing ----30% homology with ras proteins suggested the existence of a large family. We noticed that a sequence of six amino acids was strictly conserved in ras, rho and YPT proteins and took advantage of this conserved sequence to isolate several additional ras-related genes. Using an original oligonucleotide strategy we first isolated the cDNA of the ral protein [4] sharing --50% amino acid identity with ras proteins, and more recently the cDNAs of the four tab proteins [5] sharing ~30% amino acid identity with ras proteins and 75% to 45% identity with the yeast YPT protein, the most homologous one, rabl, being a mammalian homolog of YPT. Other rasrelated genes have also been found in various organisms by low-stringency hybridization with the H-ras probe: R-ras in mammals [6], Apl-ras in Aplysia [7], D-ras2 [81 and D-ras3 [91 in Drosophila; or with the YPT probe: YPT2 and Y F I 3 in Schizosaccharomyces pombe [D. Gallwitz and M. Yamamoto, personal communications]. And

866

P. Chardin

surprisingly, it appears that a cDNA recently isolated from a yeast secretion mutant, SEC4 [ 10], belongs to the rab / YPT family.

Structure of !he ras suped~.mily In its present state, the ras superfamily might be divided into three main branches represented by ras, rho and rab (see the Fig. 1). In mammals, the ras branch includes the three classical ras proteins H-ras, K-ras and N-ras (-~85% amino acid identity), the ralA and ralB proteins (~85% amino acid identity, =50% identity with any of the ras proteins), the R-ras protein (~50% identity with ras or ral), the rap! and rap2 proteins (~65% identity, and =55% identity with ras or ral) which are human homologs of the Drosophila D-ras3 protein (V. Pizon et al., submitted). At present, two proteins found in different organisms: D-ras2 from Drosophila, and Apl-ras from Aplysia have no mammalian counterparts; however, it is likely that these counterparts do exist. This prediction and recent discoveries suggest that the ras branch may include at least a dozen different proteins. In mammals, the second branch is represented by the rho proteins: rhoA [11] rhoB and rhoC (P. Chardin et al., submitted) (=85% identity and ~30% when compared to ras). The yeast counterpart of these human proteins is RHO!, but a second protein, RHO2, possessing =50% identity with RHOI, has also been founu and might be considered as a yeast rho-like protein raplA raplg

rap~

rlJsII ri.t~K

v..j OrasJ

ru~R

rul,q

Expression Most of the ras superfamily genes are expressed in all tissues, although several of these genes are expressed at higher levels in one or a few organs [14, B. Olofsson et al., submitted]. High levels of expression are frequently observed in brain; moreover, one of the rab genes, rab3, appears to be expressed at high levets in brain, while it is t=UlUlff

U~tI~I~I.I~U

II1 O i l y

Ollll;;!

Oi~dll

btUUI~;U

I,D.

Olofsson et al., submitted). Two RNAs of clearly different length are commonly observed in most tissues, but for several of these genes an l'lltJl~ J'hlll.l rlme

r~ i H

"-V Oraa2'

[12]; thus it would be very surprising that such rho-like proteins did not exist in mammals, although they have not been found as yet. In mammals, the third branch is represented by the rab proteins: rabl, rab2, tab3 and rab4 (---40% identity, ~30% when compared to ras); rabl exhibits ~75% identity with YPT and is a mammalian homolog of this yeast protein. Rab2, 3 and 4 possess ~40% identity with rabl [13]. In yeast, three proteins showing =40% identity with YPT have recently been found: SEC4, YPT2 and YPT3, but none of them is very closely related to rab2, 3 or 4. Thus, it is likely that some of the rab2, 3 or 4 genes have yeast counterparts that remain to be discovered, while reciprocally, the yeast SEC4, YPT2 and YPT3 probably have mammalian counterparts. Therefore several arguments suggest that this rab branch is as large as the ras branch and includes more than a dozen different proteins.

riJhl

ri_d~k~ rlJlJ.5 r d l ] / ,

llraa 1 AI)I - t ' ~

rhu

RIIIJ I

RIIII~

/

Vl j

5EI.'4

VPI I

VPI_~

J

J

J

J Fig. 1. Schematic representation of the ras superfamily in its presenl state. Yeast genes are at the bottom, Aplysia and Drosophila genes in the middle, mammalian genes at the top.

T h e ras s u p e r f a m i l y p r o t e i n s

increased expression of the lower RNA is observed in testis and ovaries. It appears that every gene displays a specific pattern of expression in the different organs, suggesting a complex regulation. However, a detailed expression study in the different cells of a single tissue, by in situ hybridization a n d / o r by use of specific antibodies, would provide reformation of high interest.

Biochemical properties of the ras superfamily proteus All these proteins have at least one cysteine close to the c-terminal end, and - - like ras m are probably fatty-acylated and associated with plasma a n d / o r internal membranes. All the:;e proteins share four regions of high homolo,~;y corresponding to the GTP binding site of the tas protein, and it is very likely that all of them bind GTP. In fact, the GTP binding ability, and low GTPase activity, have been demonstrated for ras, rho, YPT, R-ras, ral and rab proteins [15, 16, 17, 18, P. Chardin et al., A. Zahraoui et at., in preparation]. However, it is noteworthy that many of these proteins have an amino acid difference in positions known to be 'critical' in the case of ras proteins. The r h o proteins have an alanine in position corresponding to ras glycine 13; t a b proteins do not possess a glycine in position corresponding to ras glycine 12 (and rab3 lld~t

llf&l I.l lt, l

~l~yq~l

i1~

l~-~

i 1 lk-nI

~lff

~i11~,

Iv].

~..4vv..

region 61, which had been considered as a hallmark of .,as proteins, is not strictly conserved: rap proteins possess a threonine instead of the glutamine found in ras position 61. As far as the conclusions drawn from the study of ras proteins might be transposed to the other protems of this family, we would predict that rho, rab and rap proteins are constitutively in a slightly more active state than ras proteins. Besides the four main GTP binding blocks, a few scattered amino acids are strictly conserved in all the proteins of this family; among the m , threonine 35 and tyrosine 71 are of special interest as potential substrates for serine / threonine kinases and tyrosine kinases.

Function of the ras superfamily proteins The first proteins characterized on the basis of their biological activity are H-ras, K - r a s and Nras: the activated proteins transform NIH/31"3 cells, and their most obvious effect, when

867

microinjected, is an increase in membrane ruffling and cell motility, which may explain the loss ot contact inhibition. None of the ras-related genes has ever been isolated from the large number of tumours studied by the N I H / 3 T 3 assay in many laboratories, indicating that ras-related genes are no~ frequently implicated in tumour formation, or cannot be detected by this approach. In fact: it seems that R-ras [i8] and r a l A [P. Chardin et al., in preparation] are not able to transform N I H / 3 T 3 cells, in the usual assay; therefore, the potential involvement of ras-related genes in transformation remains an open question. The second protein characterized on a functional basis in yeast is SEC4: this mutation impairs the fusion of post-Golgi vesicles with the plasma membrane. It appears also that YPT plays an important role in the assembly of the actin network and in microtubule assembly [19, 201. Therefore, it seems likely that the ras superfamily proteins play a critical role at the membrane/cytoskeleton interface and regulate some of the physical exchanges between the cell and its environment. However, a direct involvement of the ras superfamily proteins in these events remains to be demonstrated. It has been known for decades that the biochemistry of a transformed cell is deeply altered, consequently it is not surprising that the careful study of any biochemical parameter in a ras transformed cell demonstrates a difference from a normal cell. However, these kinds of data provide little information, if any, on ras function. What we would like to know is the nature of the proteins directly interacting with the ras proteins. R a s proteins possess several structural and biochemical analogies with the a subunits ef G proteins involved in many receptor/effector systems; and it was postulated that ras proteins might transduce signals from a growth factor receptor to an eflector such as phospholipase C or phospholipase A2. However, recent evidence saggests that the action of ras on these phospholipases is an indirect effect [21,22]. It is very likely that ras proteins are transiently activated by growth factor receptors, bu~ the biochemical basis of this activation is not understood. An elegant genetic approach has localized an essential external segment of the ras protein, which can never be deleted or mutated without loss of activity, even in a constitutively active protein ( v - H - r a s p21). This region, in position 32-42, is known as the 'effector' binding site. Recent e~ idence suggests

868

P. Chardin

that the only protein known to interact directly with ras proteins, named G A P (GTPase Activating Protein) [23], binds to this region [D. Lowy, F. McCormick, A. Hall, personal communications]. It is thus postulated that G A P is the ellector; but little is known of this protein, except that it is a 116 kDa cytoplasmic protein. The idea that G A P is the effector seems to be in contradiction to the conclusion of the first work [23], that G A P is not able to interact with mutants Aspl2 and Vail2, among the most active; thus we should also bear in mind the hypothesis that G A P might be a negative regulatory protein. The characterization of this G A P protein will certainly be an essential step in the understanding of ras function. It is noteworthy that at least one ras-related protein, rapl, has exactly the same am, ao acid sequence as ras in the 'effector' region, that rap2 and R-ras display only one difference, while ral has three. Obviously it would be of major interest to know whether rapl also interacts with GAP and whether rap2, R-ras, or ral interact with the same G A P or with proteins very closely related to GAP. It is likely that every protein of the ras superfamily possesses its own GAP-like interacting protein. It might seem discouraging that in spite of intense efforts of research teams all over the world for several years, the function of ras is not precisely understood. However we have two good reasons to be optimistic: first, most ras and ras-related proteins produced in E. coli may now be purified in large amounts so that good biochemistry may be done; second, for the general question 'what is the function' have been substituted more precise questions briefly raised in the course of this review; and certainly, more precise answers will soon be obtained.

Reterences 1 Barbacid M. (1987) Annu. Rev. Biochem. 56,

776-827 2 Madaule P. & Axel R. (1985) Cell 41, 31-40 3LGallwitz D., Donath C. & Sander C. (1983) Nature 306, 704-707 4 Chardin P. & Tavitian A. (1986) EMBO J. 5, 2203 -2208 5 Touchot N., Chardin P. & Tavitian A. (1987) Proc. Nad. Acad. Sci. USA 84, 8210-8214 6 Lowe D., Capon D., Delwart E., Sakaguchi A., Naylor S. & Goeddel D. (1987)Cell48, 137-146 7 Swanson M., Elste A., Greenberg S., Schwartz J., Aldrich T. & Furth M. (1986) J. CellBiol. 103, 485 -492 8 Mozer B., Marlor R., Parkhurst S. &Corces V. (1985) Mol. Cell. Biol. 5,885-889 9 Schejter E. & Shilo B.-Z. (1985) EMBO J. 4, 407-412 10 Salminen A. & Novick P. (1987) Cell 49, 527-538 11 Yeramian P., Chardin P., Madaule P. & Tavitian A. (1987) Nucleic Acids Res. 15, 1869 12 Madaule P., Axel R. & Myers A. (1987) Proc. Natl. Acad. Sci. USA 84, 779-783 13 Zahraoui A., Touchot N., Chardin P. & Tavitian A. (1988) Nucleic Acids Res. (in press) 14 Leon J., Guerrero I. & Pellicer A. (1987) Mol. Cell. Biol. 7, 1535-1540 15 Tucker J., Sczakiel G., Feuerstein J., John J., Goody R. & Wittinghofer A. (1986) EMBO J. 5, 1351-1358 16 Anderson P. & Lacal J. C. (1987) Mol. Cell. Biol. 7, 3620-3626 17 Wagner P., Molenaar C., Rauh A., Br6kel R., Schmitt H. & Gallwitz D. (1987) EMBO J. 6, 2373-2379 18 Lowe D. & Goeddel D. (1987) Mol. Cell BioL 7, 2845-2856 19 Schmitt A., Wagner P., Pfaff E. & Gallwitz D. (1986) Cell 47,401-412 20 Segev N. & Botstein D. (1987) Mol. Cell. Biol. 7, 2367-2377 21 Seuwen K., Lagarde A. & Pouyssegur J. (1988) EMBO J. 7,161 - 168 22 Yu C.-L., Tsai M.-H. & Stacey D. (1988) Cell52, 63-71 23 Trahey M. & McCormick F. (1987) Science 238, 542-545