The Human Growth Factor-Inducible Immediate Early Gene, CYR61, Maps to Chromosome 1P

SHORT REPORT The human growth factor-inducible immediate early gene, CYR61, maps to chromosome 1p

P Jay, J L BergeÂ -Lefranc1, C Marsollier, C MeÂ jean, S Taviaux and P Berta

Centre de Recherche de Biochimie MacromoleÂcularie, CNRS ERS 155-INSERM U249, route de Mende, BP5051, 34033 Montpellier Cedex; 1INSERM U406, FaculteÂ de MeÂdecine, boulevard Jean Moulin, 13385 Marseille Cedex, France

Complementary DNA encoding the human CYR61 respective well known or putative functions. Some of protein was isolated from human embryonic tissues and them for example encode transcriptional regulators mapped to chromosome 1p22-p31. We show that CYR61 (jun, fos, myc, fra-1, SRF, krox 20/egr-2, egr-1/TIS8), encodes a 381 amino acid protein rich in cysteine and whilst other immediate early genes encode proteins proline residues that is strongly conserved with the mouse involved in paracrine intercellular communication homologue. Sequence analysis reveals the presence of (such as cytokines), in cytoskeleton or in extracellular several distinct protein domains which confer a mosaic matrix. In the latter group, several highly related structure to this protein and makes human CYR61 a proteins have been recently characterized using member of a recently described growth regulator family dierent approaches and diverse models. Cef10 is the that includes several proto-oncogene products. From our ®rst gene of this group to have been described after it results we hypothesize that this new immediate early was cloned from chicken embryonic ®broblasts upon gene may play a role in cell commitment during induction by the viral oncogene pp60v-src (Simmons et embryogenesis and more generally in the control of cell al., 1989). Later, a close relative, Cyr61 (also called proliferation. bIG-M1), was cloned among a set of immediate early genes rapidly induced by serum, by puri®ed platelet- Keywords: CYR61; immediate early gene; CCN family; derived growth factor or by TGF-b1 in mouse mosaic protein ®broblasts (O'Brien et al., 1990; Brunner et al., 1991). A third member, CTGF (Connective Tissue Growth Factor gene), was detected both by screening a human umbilical vein endothelial cell cDNA library with anti- Characterization of genes controlling cell growth and PDGF antibodies (Bradham et al., 1991) and in human dierentiation during embryogenesis or of genes foreskin ®broblasts under TGF-b stimulation (Igarashi encoding potential tumor suppressor proteins that et al., 1993). The mouse orthologue was termed ®sp-12 inhibit cell proliferation in favor of cell dierentiation and was also shown to be induced by serum (Ryseck et is obviously fundamental in order to understand the al., 1991). Finally, a chicken gene, nov, normally molecular basis of development and oncogenesis. Cell arrested in adult kidney was shown to be over- stimulation by diverse growth factors as well as by expressed in myeloblastosis-associated virus type 1- transforming oncogenes commits the cell to a complex induced nephroblastomas (Joliot et al., 1992). Its genetic program. Very rapidly during this program a human orthologue, novH, was also identi®ed and was set of genes, usually termed immediate early genes, found to be associated with dierent human tumors have been shown to be transcriptionally activated in such as Wilm's tumor as suggested by its chromosomal response to stimulation (Herschman, 1991). This localization (Martinerie et al., 1992). Finally the transcriptional activity which takes place within Xenopus laevis nov gene has recently been cloned minutes after stimulation does not require de novo (Ying and King, 1996). The emergence of this family protein synthesis. We focused on this family of genes of growth regulators led Bork to propose to group assuming that their characterization will be helpful in these genes under the denomination `CCN family' (for order to further the understanding of some of the CTFG,Cef10/Cyr61,nov family) (Bork, 1993). In this cellular signal transduction pathways activated follow- communication we report the isolation from an ing stimulation of the cell by growth factors, or embryonic source of a cDNA encoding the complete throughout cell determination/dierentiation processes sequence of the human CYR61 protein, its conserva- occurring during human embryogenesis. We cloned tion with the other members of the CCN family, several of these genes by screening a human embryo primary analysis of its expression pattern in human cDNA library and selecting low abundance cDNAs, tissues as well as its mapping to chromosome 1p22-31. instead of the diverse subtractive or dierential Screening of 46103 clones from a 6 week-old human hybridization approaches used in many research teams embryo cDNA library with a probe derived from the (Lau and Nathans, 1985; Simmons et al., 1989). same cDNA mixture that was used to construct the Immediate early genes that have been identi®ed to library allowed us to isolate low abundance cDNAs date can be grouped together according to their that were systematically partially sequenced (Jay et al. in press). The partial nucleotide sequence of clone HE6WCR113 revealed signi®cant homology with mouse Cyr61 mRNA and the HE6WCR113 sequence Correspondence: P Berta Received 20 August 1996; revised 5 December 1996; accepted was subsequently completed. This cDNA is 2014 5 December 1996 nucleotides long, including a polyadenylation signal The human CYR61 cDNA PJayet al 1754 but excluding the poly(A)+ tail (Figure 1). The members were de®ned as proteins with a mosaic nucleotide sequence of the 3' untranslated region is structure because they all contain several motifs very well conserved when compared to that of mouse, shared by various functionally unrelated proteins especially in its second half where AU-rich elements are (Bork, 1993). These domains are strongly conserved perfectly identical (not shown). The mRNA instability in the human CYR61 gene product (Figure 2). From conferred by these AU sequences is now well the N- to the C-terminal part of the human CYR61 documented and is one of the mechanisms known to protein, the ®rst striking feature is the presence of a increase the rate of mRNA degradation (Shaw and hydrophobic N-terminal stretch of 24 residues with a Kamen, 1986). AU elements are present in numerous predicted a-helix structure in good agreement with the mRNAs encoding secreted growth factors or other function of a signal peptide which probably directs transiently expressed genes. The predicted translation CYR61 to the secretory pathway for subsequent export product is a 382 amino acid polypeptide with a (Gorr and Darling, 1995). The ability of the mouse calculated molecular weight of about 42 kDa and homologue to be secreted was con®rmed by its with a calculated isoelectric point of 8.06. The association with both the extracellular matrix and the dierent features which helped to characterize CYR61 cell surface where it may act as a cell adhesion protein are shared by the other members of the CCN molecule (Yang and Lau, 1991). It is worth noting family. Human CYR61 is particularly cysteine-rich that these properties (export, protein localization, high (10% of all residues) and displays a remarkable degree cysteine content) are also shared with the proto- of evolutionary conservation (92.8% identity with the oncogene product Int-1/Wnt-1 (Brown et al., 1987). equivalent mouse gene product) (Figure 2). The A second important feature is the sequence of conserved cysteines are clustered in two segments and GCGCCKVC which begins at residue 49 (Figure 1) are totally absent from the central acidic domain of the and conforms to the GCGCCXXC motif found in protein which is furthermore highly variable when insulin-like growth factor binding proteins (Kiefer et compared with that of other members of the CCN al., 1991). It remains to be established if this cysteine- family (Figure 2). Twenty-®ve prolines are scattered rich domain really acts as a growth factor binding throughout much of the CYR61 protein and most domain. It is however noteworthy that in the case of probably prevent a-helix formation. CCN family nov, an amino-terminal-truncated variant deleted for

Figure 1 Nucleotide sequence and predicted amino acid sequence of the human CYR61 cDNA. The amino acid sequence is presented in the single-letter code under the ®rst position of the corresponding codons. The GenBank accession number for this sequence is U62015. Cysteine residues are encircled. The ®rst in-frame stop codon is indicated by an asterisk. A consensus polyadenylation signal in the 3' non coding region at position 1996-2001 is boxed. AU-rich elements are underlined. For molecular cloning of human CYR61 cDNA, total RNA was extracted from 6 week-old embryos following the method of Chomczynski and Sacchi. Poly(A)+ mRNAs were isolated by one cycle of binding elution on oligo(dT)-cellulose. Five micrograms of poly(A)+ RNA served as template in the preparation of size fractionated (4400 basepairs) cDNAs using the lZAP-cDNA synthesis kit (Stratagene, La Jolla, California). Around 4000 clones were screened using the original cDNA mixture as a probe after labeling with a random priming kit (Megaprime, Amersham). The same ®lter was then rehybridized with a vector-speci®c probe to detect all the clones present. Over 250 clones were not detected with the cDNA probe and were subsequently considered as low abundance cDNAs. The 3' ends of these clones were systematically sequenced using double stranded sequencing strategy with Taq DyeDeoxyTerminator kit and an Applied Biosystems 373A automated DNA sequencer (Foster City, California). Sequences were analysed and examined for homology using the Winconsin Genetics Computer Group (GCG) programs and ®tted together with GAP function from this program package The human CYR61 cDNA PJayet al 1755 this motif is sucient to induce ®broblast transforma- molecules such as sulphate glycoconjugates in cell tion (Joliot et al., 1992). The third domain (VWC adhesion. These domains are also partially shared by a domain, Figure 2) was originally found as a repeat in family of mucin-like proteins and by the recently the large multifunctional Von Willebrand factor cloned Norrie disease gene (Chen et al., 1993). protein, but is also present in several mucins where it Tissue distribution of the human CYR61 mRNA may contribute to form oligomers (Bork, 1993). was determined by Northern blotting (Figure 3). It Finally, the last two cysteine-rich regions, termed revealed a rather wide expression of CYR61 mRNA in TSP1 and CT domains (Figure 2) were postulated to adult as well as in fetal human tissues and the level of be involved in the binding of both soluble and matrix expression varied largely between the tissues tested.

Figure 2 Amino acid comparison of the members of the CCN family with the human CYR61 sequence using the PILEUP program of the GCG program package. The four distinct conserved modules as delineated by P Bork are indicated

fetal tissues adult tissues brain lung liver kidney heart brain placenta lung liver skeletal muscle kidney pancreas spleen thymus prostate testis ovary small intestine colon

—4.4—

—2.4— L Cyr61

—1.35—

—2.4— L actin

Figure 3 Northern blot analysis of human CYR61 mRNA expression in adult and fetal organs. The same ®lters were labeled with full-length CYR61 cDNA or with human b-actin as a control. Positions of size markers (kilobases) are indicated. Northern blots were commercial (Clontech fetal blot and adult blots I and II). Each lane contained 2 mg of puri®ed mRNA and blots were handled as recommended by the supplier. The ®nal washes were done at 558C in 0.16SSC; 0.1% SDS, followed by 4 days (CYR61 probe) or 1h(b-actin) exposure The human CYR61 cDNA PJayet al 1756 cartilaginous structures and circulatory system (O'Brien and Lau, 1992). It is therefore likely that CYR61 also contributes to such cell commitment events during early human development as is suggested by distantly related genes like twisted gastrulation in Drosophila (Mason et al., 1994). Two distinct mRNAs were detected with the CYR61 probe, a major one at 2.5 kb and a minor one at about 4 kb. Since the relative amount of both mRNAs is similar among the dierent tissues tested, it is likely that they represent transcripts from the same locus that were dierentially spliced or that used dierent polyadenylation signals. Another possibility is that the two transcripts represent dierent members of the CCN family but this is rather unlikely since only one signal was detected by ¯uoresence in situ hybridization (FISH) analysis. The human CYR61 gene was mapped to chromosome 1 at position p22-p31 by ¯uorescence in situ hybridization (Figure 4). Screening of genetic databases revealed the possible location on chromosome 1 of a locus with a malignant transformation suppression activity (Stoler and Bouck, 1985). Because of syntheny Figure 4 Fluorescence in situ hybridization on R-banded with mouse chromosome 4, the locus identi®ed by chromosome using the full length CYR61 cDNA. Chromosome Stodler and Bouck is likely to map on 1p, distal to metaphases were obtained from phytohemagglutinin-stimulated 1p22.1 (Zabel et al., 1985). More recently, abnormal- blood lymphocytes of a healthy donor after thymidine ities of chromosome 1p were also shown to correlate synchronization and bromodeoxyuridine incorporation (Viegas- Pequignot and Dutrillaux, 1978). The probe was labeled with with breast cancer (Hainsworth et al., 1992), neuro- biotin-16 dUTP by random priming. In situ hybridization was blastoma (Gehring et al., 1995), pheochromocytoma performed according to Pinkel et al. (1986). After one (Shin et al., 1993) or even with primary hepatomas ampli®cation, R bands were obtained as described (Lemieux et (Simon et al., 1991). Besides this chromosomal al., 1992). Idiogram of chromosome 1 showing the distribution of ¯uorescent spots on the 1p22-p31 bands for the 10 most resolutive localization diverse characteristics previously reported metaphases for the mouse homologue contribute in making CYR61 an attractive candidate for a proto-oncogene. Due to the very high level of sequence similarity among the members of the CCN family, these secretory proteins, Human CYR61 expression was not detected in including a putative oncoprotein as well as a leukocytes (not shown), whilst moderate amounts of connective tissue growth factor probably share most messenger were found in fetal or adult brain and liver, of their molecular functions. Transcriptional induction adult lung, adult kidney and adult thymus. The pattern of mouse Cyr61 was obtained via diverse stimuli, of expression that we obtained con®rms previous data including serum, PDGF, FGF, TPA, oncogene v-src from mouse which shows that expression of CYR61 (O'Brien et al., 1990), ionomycin (Sells et al., 1994) or mRNA is not speci®c to one particular tissue but even during liver regeneration following partial rather distributed in several organs (O'Brien et al., hepatectomy (Nathans et al., 1988). Mouse Cyr61 1990). The tissue distribution reported by O'Brien et al. was shown to be a secreted protein with the ability to is partially divergent from the one that we describe in interact with both cell surface and extracellular matrix this study but several factors explain this discrepancy. but also with heparin (Yang and Lau, 1991), and its First, the methods used in both studies are dierent primary structure also revealed its putative growth since O'Brien et al. used RNAse protection instead of factor binding capacity. Taken together, these proper- Northern blotting in our case. Second, as can be seen ties suggest a role for Cyr61 in cell-cell and cell- in Figure 3, the level of expression of Cyr61 mRNA extracellular matrix interactions acting as a mediator in can dier signi®cantly between fetal and adult tissues the growth response and subsequently in the control of (in kidney for instance) and the use of samples that do cell proliferation. Further analysis of the human not originate from the same developmental stage may CYR61 gene sequence and localization will be very therefore also contribute to the dierence in CYR61 helpful in order to determine whether an alteration of mRNA expression. Finally, interspeci®c dierences CYR61 sequence or expression is associated with between orthologous gene expression patterns can malignancy. also occur and do not presume dierent functions since in any case, the presence of the messenger does not necessarily imply that of the functional corresponding protein. On the other hand, cloning of human Acknowledgements CYR61 from an 6 week-old embryonic cDNA library This work was supported by the Centre national de la was enabled by the early expression of this gene during Recherche Scienti®que, Institut National de la SanteÂ et de embryogenesis, consistent with previous data in mouse la Recherche MeÂ dicale, and by grants GREG 43 to PB and describing the expression of Cyr61 in midgestation 30/95 to PB and JLBL. Authors thank May Morris for embryos, and essentially correlated with the developing critical reading of the manuscript. The human CYR61 cDNA PJayet al 1757 References

Bork P. (1993). FEBS Lett., 327, 125 ± 130. Mason ED, Konrad KD, Webb CD and Marsh JL. (1994). Bradham DM, Igarashi A, Potter RL and Grotendorst GR. Genes & Dev., 8, 1489 ± 1501. (1991). J. Cell Biol., 114, 1285 ± 1294. Nathans D, Lau LF, Christy B, Hartzell S, Nakabeppu Y Brown AM, Papko J, Fung YK, Shackleford GM and and Ryder K. (1988). Cold Spring Harbor Symp. Quant. Varmus HE. (1987). Mol. Cell Biol., 7, 3971 ± 3977. Biol., 53, 893 ± 900. Brunner A, Chinn J, Neubauer M and Purchio AF. (1991). O'Brien TP, Yang GP, Sanders L and Lau LF. (1990). Mol. DNA Cell Biol., 10, 293 ± 300. Cell Biol., 10, 3569 ± 3577. Chen Z-Y, Battinelli EM, Hendriks RW, Powell JF, O'Brien TP and Lau LF. (1992). Cell Growth Dier., 3, 645 ± Middleton-Price H, Sims KB, Breake®eld XO and Craig 654. IW. (1993). Genomics, 16, 533 ± 535. Pinkel D, Straume T and Gray JW. (1986). Proc. Natl. Acad. Chomczynski P and Sacchi N. (1977). Anal. Biochem., 162, Sci. USA, 83, 2934 ± 2938. 156 ± 159. Ryseck RP, Macdonald-Bravo H, Mattei MG and Bravo R. Gehring M, Berthold F, Edler L, Schwab M and Amler LC. (1991). Cell Growth Dier., 2, 225 ± 233. (1995). Cancer Res., 55, 5366 ± 5369. Sells SF, Wood DP Jr, Joshi-Barve SS, Muthukumar S, Gorr S-U and Darling DS. (1995). FEBS lett., 361, 8 ± 12. Jacob RJ, Crist SA, Humphreys S and Rangnekar VM. Hainsworth PJ, Raphael KL, Stillwell RG, Bennett RC and (1994). Cell Growth Dier., 5, 457 ± 466. Garson OM. (1992). Br. J. Cancer, 66, 131 ± 135. Shaw G and Kamen R. (1986). Cell, 46, 659 ± 667. Herschman HR. (1991). Annu. Rev. Biochem., 60, 281 ± 319. Shin E, Fujita K, Takami K, Kurahashi H, Kurita Y, Igarashi A, Okochi H, Bradham DM and Grotendorst GR. Kobayashi T, Mori T, Nishisho I and Takai S. (1993). Jpn (1993). Mol. Biol. Cell, 4, 637 ± 645. J. Cancer Res., 84, 402 ± 408. Jay P, Diriong S, Taviaux S, Roeckel N, MatteÂ iM-G,Audit Simmons DL, Levy DB, Yannoni Y and Erickson RL. M, BergeÂ -Lefranc J-L, FonteÁ s M and Berta P. (1996). (1989). Proc. Natl. Acad. Sci. USA, 86, 1178 ± 1182. Genomics, in press. Simon D, Knowles BB and Weith A. (1991). Oncogene, 6, Joliot V, Martinerie C, Dambrine G, Plassiart G, Brisac M, 765 ± 770. Crochet J and Perbal B. (1992). Mol. Cell Biol., 12, 10 ± 21. Stoler A and Bouck N. (1985). Proc. Natl. Acad. Sci. USA, Kiefer MC, Masiarz FR, Bauer DM and Zapf J. (1991). J. 82, 570 ± 574. Biol. Chem., 266, 9043 ± 9049. Viegas-PeÂ quignot E and Dutrillaux B. (1978). Ann. Genet. Lau LF and Nathans D. (1985). EMBO J., 4, 3145 ± 3151. (Paris), 21, 121 ± 125. Lemieux N, Dutrillaux B and Viegas-PeÂ quignot E. (1992). Yang GP and Lau LF. (1991). Cell Growth Di., 2, 351 ± 357. Cytogenet. Cell Genet., 59, 311 ± 312. Ying Z and King ML. (1996). Gene, 171, 243 ± 248. Martinerie C, Viegas-Pequignot E, Guenard I, Dutrillaux B, Zabel BU, Eddy RL, Lalley PA, Scott J, Bell GI and Shows Nguyen VC, Bernheim A and Perbal B. (1992). Oncogene, TB. (1985). Proc. Natl. Acad. Sci. USA, 82, 469 ± 473. 7, 2529 ± 2534.