Transactivation domains are not functionally conserved between vertebrate and invertebrate serum response factors Sonia Avila, Marie-Carmen Casero, Rocı ´ o Fernandez-Canto ´ n and Leandro Sastre Insitituto de Investigaciones Biome ´ dicas CSIC/UAM, C/Arturo Duperier, Madrid, Spain The transcription factor serum response factor (SRF) regu- lates expression of growth factor-dependent genes and muscle-specific genes in vertebrates. Homologous factors regulate differentiation of some ectodermic tissues in inver- tebrates. To explore the molecular basis of these different physiological functions, the functionality of human, Droso- phila melanogaster and Artemia franciscana SRFs in mam- malian cells has been compared in this article. D. melanogaster and, to a lesser extend, A. franciscana SRF co-expression represses the activity of strong SRF-depen- dent promoters, such as those of the mouse c-fos and A. franciscana actin 403 genes. Domain-exchange experi- ments showed that these results can be explained by the absence of a transactivation domain, functional in mam- malian cells, in D. melanogaster and A. franciscana SRFs. Both invertebrate SRFs can dimerize with endogenous mouse SRF through the conserved DNA-binding and dimerization domain. Co-expression of human and A. franciscana SRFs activate expression of weaker SRF- dependent promoters, such as those of the human cardiac a-actin gene or an A. franciscana actin 403 promoter where the SRF-binding site has been mutated. Mapping of A. franciscana SRF domains involved in transcriptional activation has shown that the conserved DNA-binding and dimerization domain is neccessary, but not sufficient, for promoter activation in mammalian cells. Keywords:SRF;Artemia; Drosophila; transcription; evolu- tion. The serum response factor (SRF) is a transcription factor initially isolated as responsible for activation of several immediate early genes, such as c-fos, in the response of quiescent cells to serum [1]. It is characterized by its binding to the minor grove of the DNA through a conserved domain, the MADS (MCM1-Agamous-Deficiens-SRF) box [2]. This domain is shared by an extensive family of transcription factors that also include the animal MEF-2 factor and a large number of plant transcription factors [3,4]. Immediately C-terminal to the MADS box there is a region of amino acids (SAM-domain) conserved between SRFs from different species but not with other members of the MADS box family [3]. SRF is assembled as a homodimer and the dimerization domain has been also mapped to the MADS box [1]. SRF does not seem to be active by itself and requires association with other tran- scription factors bound to the same promoter or directly interacting with SRF [5]. Many of these interactions also occur through the MADS box [6]. In addition to this functional domain, a transactivation domain has been located in the C-terminal region of vertebrate SRF [7,8]. This transcription factor has been also involved in the activation of several muscle-specific genes in vertebrates [9]. Besides, the generation of SRF-null mice showed that this factor is necessary for mesoderm induction and for the proper differentiation of several mesodermal tissues [10,11]. SRF binding sites found in the promoter of serum-induced and muscle-specific genes are very similar and contain consensus CArG boxes: CC(A/T) 6 GG [12]. Despite the similarity in their SRF-binding sites, some genes, such as c-fos, are activated in response to serum and others, such as a-actin genes, are induced during muscle differentiation that, in cell culture, usually implies serum withdrawal. Moreover, SRF-dependent immediately early genes are repressed after muscle differentiation [13]. Recent experi- ments have shown that this complex regulation is promoter- context dependent [14]. The transcriptional regulation of each SRF-dependent promoter seems to be determined by the binding of SRF cofactors. Some of these cofactors are tissue-specific, as the muscular Nkx2.5 [15], GATA-4 [16] or myocardin [17] transcription factors. Other cofactors are regulated by growth-factor transduction pathways, such as the classical ternary complex factors, that are activated by MAP-kinase pathways [18]. Several nonvertebrate SRF homologues have been described that are mainly involved in differentiation processes. Drosophila melanogaster SRF (DmSRF) is neces- sary for differentiation of terminal tracheal cells and cells of the wing’s intervein regions [19,20]. Artemia franciscana SRF (AfSRF) is specifically expressed in ectodermal tissues [21]. The gene srfA is necessary for several morphogenetic processes and terminal spore differentiation in the social amoeba Dictyostelium discoideum [22,23]. Therefore, it seems that, although vertebrate and invertebrate SRFs have in common the participation in differentiation processes, the tissues involved are different, suggesting that SRF must Correspondence to L. Sastre, Insitituto de Investigaciones Biome ´ dicas CSIC/UAM, C/Arturo Duperier, 428029, Madrid, Spain. Fax: + 34 91 5854587, Tel.: + 34 91 5854626, E-mail: lsastre@iib.uam.es Abbreviations: AEBSF, 4-(2-aminoethyl) benzene sulfonyl fluoride; DMEM, Dulbecco’s modified Eagle’s medium; SRE, serum response element; SRF, serum response factor. (Received 13 March 2002, revised 4 June 2002, accepted 18 June 2002) Eur. J. Biochem. 269, 3669–3677 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03077.x regulate different sets of genes in these species. In accord with this idea, no evidence was found for the involvement of SRF in the regulation of muscle actin genes in D. melano- gaster, as would be expected from the data obtained in vertebrates. The bases for these differences, despite the high similarity of the SRF amino-acid sequences in the MADS box and SAM-domain sequences, are not known. SRF regulatory pathways could have diverged in the different species, responding to different tissue-differentiation signals. Alternatively, SRF structure and regulation could be conserved but SRF expression patterns and SRF-dependent promoters could have changed during evolution. Further understanding of the evolution of SRF and SRF- dependent genes would require a better characterization of functional domains from this transcription factor in differ- ent species. MADS boxes and SAM domains have been very conserved during evolution. MADS-box regions are 98% identical and SAM-domains over 79% identical between vertebrate and invertebrate SRFs. No similarity has been detected outside these domains, except for a short region around vertebrate Serine 103 [21]. However, some important transcription factor functional domains, such as transactivation domains, are very variable in their amino- acid sequence and their location could not be predictable from amino-acid sequence comparisons. As an approach to the comparative study of these domains, the functionality of SRF proteins from an insect (D. melanogaster), a crustacean (A. franciscana) and humans, in mammalian cells, has been studied in the present article. The evidence obtained suggests that functions associated with the MADS box, such as DNA binding and dimerization, are well preserved in SRFs from different species. However, the transactivation domain is not conserved. D. melanogaster and A. franciscana C-terminal regions are devoid of transactivation activity in mammalian cells. A weaker transcriptional activation domain has been found in the N-terminal and MADS box regions from A. franciscana SRF. EXPERIMENTAL PROCEDURES Cell culture and transfection Myoblastoid C 2 C 12 [24] and monkey kidney epithelial cells, Bsc1 [25] and Bsc40 (a Bsc1 subline) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supple- mented with 10% fetal bovine serum and 2 m M glutamine. Embryonic fibroblastoid cells NIH3T3 were cultured in DMEM supplemented with 10% neonatal bovine serum and 2 m M glutamine. Cells were transfected by the calcium phosphate precipitation method [26,27]. Five micrograms of the reporter vector, 1 lg of the SRF expression vector and 1 lgoftheb-galactosidase expression vector pCMV-bgal (Clontech Laboratory Inc, Palo Alto, CA, USA) were used for each transfection, except in some experiments, that are specified in each case. Cells were harvested 72 h after transfection, and b-galactosidase activities were determined [28]. Luciferase activity was determined with a commercial kit (Promega) according to the manufacturer’s instructions. The luciferase/b-galactosidase ratio was determined for each sample and the relative activity was calculated in relation to samples transfected with the reporter vector alone or together with the pcDNA3-myc vector without any SRF, that were given the value of 100. Each experiment was repeated at least three times with duplicated samples. Media ± SD are represented in the figures. The statistical significance of the differences observed, in relation with the sample transfected with the reporter vector alone, was analyzed with the PRISM 2.0 program, using the one-way ANOVA and Turkey tests (*p < 0.05; **p <0.01; ***p < 0.001). Cells lines permanently expressing SRFs from the three different species were generated by transfecting 10 6 C 2 C 12 cells with 5 lg of each pcDNA3-myc-SRF vector by calcium phosphate precipitation. Sixteen hours after transfection, cells were changed to the same medium containing 1 mgÁmL )1 of geneticin for 24 h. Cells were washed and cultured in the same medium containing 0.5 mgÁmL )1 of geneticin for 15 days before collection and further culture. Generation of reporter and expression vectors The reporter vector containing the human cardiac a-actin promoter was generated by inserting a fragment of this promoter, generated by PCR, between SacIandKpnIsites of the pXP2 reporter vector [29]. The oligonucleotides 5¢- GGTACCCTGGCTGATCCTCTCCCC-3¢ and 5¢-GAG CTCGGGTGGCTGGCTCCAGGAGG-3¢, that amplify the region between nucleotides )170 and +1 of the cardiac a-actin gene [9] were used as primers in the PCR reaction. Reporter vectors containing the wild-type A. franciscana actin 403 promoter )176 to )38 region (Act403)andthe same region with the CArG box mutated (Act403mut) have been described previously [30]. The reporter vector contain- ing the c-fos promoter was kindly provided by U. Moe ¨ ns (University of Tromso, Norway) [31]. The expression vector pcDNA3-myc was generated by inserting a BamHI–EcoRI fragment that codes for six copies of myc epitope, obtained from pCS2+MT [32], in the plasmid pcDNA3 (Invitrogen). cDNA molecules coding for human [33], D. melanogaster [34] and A. franciscana [21] SRFs were cloned in the pcDNA3-myc vector, in the same reading frame as the myc epitope. An EcoRI site was inserted in front of the initiating methionine of each SRF by PCR for these constructions. Besides, each cDNA was isolated as an EcoRI–EcoRI fragment and cloned in the plasmid expression vector pSG5 [35]. All the constructions generated were sequenced to confirm that no mutation had been introduced in PCR reactions. Hybrid SRF molecules were generated using a BclIsite conserved in the same position, relative to the MADS box, in the cDNA clones from the three species. The N-terminal coding region from each species, including the MADS box, was isolated as an EcoRI–BclI fragment and ligated to the C-terminal coding region from the other species in the pcDNA3-myc vector. A. franciscana SRF N-terminal deletions were generated by PCR between a series of oligonucleotides containing a EcoRI site, designed to conserve the myc epitope reading frame, and a common oligonucleotide downstream of the MADS box coding region. The different deleted fragments were isolated as EcoRI–SnaIfragmentsandusedtoreplace the full-length EcoRI–SnaI fragment in the original pcDNA3-myc-AfSRF vector. The oligonucleotides uti- lized as PCR primers were N1: 5¢-GGGAATTCGGGT GGTCTTGAACCCGATATT-3¢;N2:5¢-GGGAATTCG 3670 S. Avila et al. (Eur. J. Biochem. 269) Ó FEBS 2002 TCTTATATGAATGCAGTTCTG-3¢;N3:5¢-GGGAAT TCGCTAGGCCACAGTTTGAATTTG-3¢;N4:5¢-GGG AATTCGACCTCTGAAAATGTAAAACAG-3¢;N5: 5¢-GGGAATTCGGATCCTCTAACTGGGTTAGAT-3¢; N6: 5¢-GGGAATTCGTCCCCGGACGAGGACAGG TCA-3¢;N7:5¢-GGGAATTCGCCTGCCAATGGTAAA AAGACA-3¢. TheC1deletionwasgeneratedbyPCRusingthe oligonucleotide containing the methionine initiation codon and an oligonucleotide (C1: 5¢-GCGGCCGCCTAAACG TTATATGTGAGTTCCG-3¢) where the codon of amino acid 220 was changed to a stop codon. The fragments generated by PCR were cloned in pcDNA3-myc. Fragment M was generated by PCR, using oligonucleotides N7 and C1asprimers,andinsertedinpcDNA3-myc.Allthe fragments generated by PCR were sequenced to check for possible mutations. Western blots and co-immunoprecipitations Expression of SRF-myc molecules in transiently transfected cells was analyzed by Western blot. After electrophoresis of 10 lg of the extracts, the samples were transferred to poly(vinylidene difluoride) membranes and incubated with polyclonal rabbit anti-(a-myc) Ig (A14, Santa Cruz Bio- technologies). Horseradish peroxidase-conjugated goat anti-(rabbit IgG) Ig (Santa Cruz Biotechnologies) was used as secondary antibody, and detected by chemiluminiscence (ECL, Amersham Pharmacia Biotech). Co-immunoprecipitation experiments were performed with nuclear extracts obtained as described previously [36]. Protein samples (50–100 lg) from nuclear extracts (200– 400 lgforAfSRF expressing cells) were immunoprecipitated with 0.5 lL of anti-SRF Ig (G20, Santa Cruz Biotechnol- ogies) or monoclonal anti-myc Ig (9E10, Santa Cruz Biotechnologies) in 20 m M Hepes, pH 7.0, 70 m M NaCl, 0.005% NP40, 0.05 mgÁmL )1 BSA, 2% Ficoll, 2 m M dithiothreitol, 0.5 m M phenylmethanesulfonyl fluoride, leu- peptin, AEBSF and aproteinin. Immunoprecipitates were analyzed by 10% SDS/PAGE and transferred to poly(viny- lidene difluoride) membranes. Control samples containing 5–10 lg of nuclear extracts (10–20 lgforAfSRF expressing cells) were also included in the electrophoresis as positive controls. Membranes were incubated with anti-SRF Ig (G20, Santa Cruz Biotechnologies) or anti-myc Ig (A14, Santa Cruz Biotechnologies) rabbit polyclonal antibodies. Horseradish peroxidase-conjugated goat anti-(rabbit IgG) Ig (Santa Cruz Biotechnologies) was used as secondary antibody and its binding was detected by chemiluminis- cence, as described above. RESULTS Effect of the expression of SRFs from different species on the activity of SRE-dependent promoters Expression vectors containing cDNAs coding for human SRF (HsSRF), DmSRF or AfSRF were transfected in mammalian cells. The activity of the encoded proteins was studied by cotransfecting reporter vectors where luciferase gene expression was placed under control of serum response element (SRE)-containing promoters. Four different pro- moters were used in these studies, corresponding to the c-fos genes from mice (c-fos) [18], human cardiac a-actin (CA)[9], A. franciscana actin 403 (Act403) and a mutated actin 403 promoter where two nucleotides of the SRF-binding site had been changed to decrease SRF binding (Act403mut) [30]. The reporter vectors containing c-fos and Act403 promoters presented similar high levels of expression in cultured mammalian cells while CA and Act403mut pro- moters presented much weaker activity, 10–20 times less active in the cell lines tested (Fig. 1A). Expression of the different SRFs was obtained by transfection of pcDNA-3 and pSG5 vectors containing the corresponding cDNA clones in Bsc40 cells. The expression from the pcDNA-3 vectors could be analyzed by Western blot as SRF molecules are fused to six repeats of the myc epitope. The results are shown in Fig. 1B. HsSRF and DmSRF were expressed at similar levels, higher than those of A. franciscana SRF, probably due to the higher A/T content of the cDNA coding for the later protein. An additional, faster migrating, band was observed after HsSRF transfection, that probably corresponds to a degra- dation product as it was present in very variable proportions in different experiments. The effects of SRF co-expression on the activity of the SRE-containing promoters are shown in Fig. 1C for pcDNA-3 expression vectors, similar results were obtained for pSG-5 vectors (data not shown). Co-expression of HsSRF increased expression from c-fos and, specially, from the weaker CA and Act403mut promoters. In the case of the strong Act403 promoter there was no effect or a slight inhibition (50%) by HsSRF co-expression. Co-expression of AfSRF also produced an increase in the activity of the weak CA and Act403mut promoters, but had no effect on c-fos and produced a small inhibition on the Act403 promoter. Activation was always smaller with AfSRF than with HsSRF. In contrast to these factors, co-expression of DmSRF produced a marked inhibition of the more active promoters, c-fos and Act403, and had no significant effects on the weaker CA and Act403mut promoters. The activity of the SRFs from these three species was also tested for c-fos and CA promoters on two more mammalian cell lines: the fibroblast cell line NIH3T3 and the mouse myoblastic cell line C 2 C 12 . The effects on the activity of c-fos and CA promoters, using the pcDNA-3 expression vectors, were similar to those described previously for Bsc40 cells (data not shown). As the differences in expression levels shown in Fig. 1B could affect the functional consequences of co-expression, the effect of transfecting different amounts of each expression vector was studied. The activity of HsSRF and AfSRF was studied on the CA promoter (Fig. 2A,B). The inhibitory effect on the c-fos promoter was studied for DmSRF (Fig. 2C). The results obtained showed that the effects observed are dose-dependent and were in agree- ment with the activating capacity of HsSRF and AfSRF, and with the inhibition obtained by D. melanogaster co-expression. Interaction between transfected and endogenous SRFs The stimulatory effects observed for HsSRF and AfSRF could be due to the increase of intracellular SRF concen- tration that would raise the amount of SRF bound to the promoters and therefore transcription of the reporter gene. Ó FEBS 2002 Functionality of arthropod SRFs in mammalian cells (Eur. J. Biochem. 269) 3671 Inhibition by DmSRF and, in the case of Act403 promoter, AfSRF could be due to a dominant negative effect where exogenous SRF dimerizes with the endogenous mammalian factor avoiding its productive interaction with activating cofactors. Alternatively, inhibition could be explained by diluting of these cofactors by an excess of nonDNA-bound SRF. This would be the more likely mechanism of the inhibition observed by co-expression of HsSRF with the Act403 promoter, given the high similarity of human and mouse SRFs. The proposed dominant negative mechanism requires dimerization between the endogenous SRF and the trans- fected SRFs. To check for these possible interactions, stably transformed C 2 C 12 cell lines that expressed HsSRF, DmSRF or AfSRF were established. The interaction between SRF molecules was assayed by co-immunopreci- pitation experiments. Exogenous SRFs were specifically recognized by antibodies directed against the c-myc epitope. Endogenous SRF was recognized by a monoclonal anti- body specific for vertebrate SRF. Extracts from C 2 C 12 cells, either nontransfected (C 2 C 12 ) or stably expressing human (C 2 -Hs), D. melanogaster (C 2 -Dm)orA. franciscana (C 2 -Af) SRFs were immunoprecipitated using anti-(c-myc) Ig (aMyc) or with no antibody (–). Immunopreciptates were analyzed by Western blot using an anti-(vertebrate-SRF) Ig and the results are shown in Fig. 3A. The migration and relative expression level of endogenous (mouse) and human SRFs is shown in the lanes labeled as E, where smaller amounts of cell lysates were subjected to Western blotting. The complementary experiment is shown in Fig. 3B where immunoprecipitation was carried with antivertebrate-SRF antibodies and the immunoprecipitates analyzed by West- ern blot with anti-(c-myc) Ig. The results obtained indicate co-immunoprecipitation and, therefore, interaction between Fig. 1. Effect of SRF co-expression on the activity of CArG-containing promoters. (A) Activity of CArG-containing promoters from the genes c-fos from mice (c-fos), A. franciscana Actin 403 (Act403), an Actin 403 promoter with the CArG box mutated (Act403 mut) and human cardiac a-actin (CA) in Bsc40 cells. Luciferase activity is expressed as times of induction over the activity of the reporter vector without promoter (pXP2). (B) Expression of exogenous SRFs in Bsc40 cells. Cells were transfected with 1 lg of pcDNA3-myc vector containing cDNAs coding for human (HsSRF), D. melanogaster (DmSRF) and A. franciscana (AfSRF) SRFs. Cellular extracts were analyzed by Western blotting using an antimyc antibody. The migration of the fusion proteins containing the myc epitope: human (HsSRF-myc), D. melanogaster (DmSRF-myc) and A. fran- ciscana (AfSRF-myc) SRFs, is indicated to the right. The migration of molecular weight markers is shown to the left. (C) Effects of SRF co-expression, from the pcDNA3 expression vectors, on the activity of CArG containing promoters. Bsc40 cells were transfected with 5 lgofpXP2 reporter vectors containing the promoters indicated in the upper right corner of each graphic and 1 lg of the pcDNA3 expression vectors. The effect of human (Hs), A. franciscana (Af) and D. melanogaster (Dm) SRFs is shown. The luciferase activity value 100 was assigned to the samples transfected with the reporter vector alone (C). Asterisks indicate statistically significant variations with respect to control activity (C). 3672 S. Avila et al. (Eur. J. Biochem. 269) Ó FEBS 2002 HsSRF, DmSRF and AfSRF and the endogenous factor in C 2 C 12 cells. These data also show similar levels of expression of HsSRF and DmSRF and lower levels of AfSRF, as previously observed in transient expression experiments (Fig. 1B). Functional analysis of SRF domains The results obtained in the experiments shown above indicate that HsSRF, AfSRF and DmSRF, expressed in mouse cells lines, interact with the endogenous SRF. The functional consequences of their expression are very differ- ent, however. Human and AfSRFs can increase the activity of SRE-containing promoters, while DmSRF co-expression produced no activation but, instead, strong repression of the more active promoters. The comparison between human and D. melanogaster factors is especially significant as both proteins are expressed at similar levels but produce opposing effects. These results raised the possibility that important functional regions of this transcription factor could be not conserved between these species. As MADS boxes are over 90% identical in their amino-acid sequences, differences could reside in other regions, possibly in the transactivation domain, mapped to the C-terminal region of vertebrate SRF [8]. A series of hybrid SRF molecules was constructed to test for this hypothesis. The existence of a conserved BclI site in equivalent positions in the cDNA clones of the three species, immediately downstream of the MADS-box coding region, was used for making these constructs. The hybrid molecules coded for the N-terminal region, including the MADS box, from one species and the C-terminal region from another. The consequences of the co-expression of hybrid SRFs on the expression of reporter vectors containing Fig. 2. Dependence of promoter activities on the amount of SRF expression vectors cotransfected. (A) The indicated amounts of the human SRF expression vector pcDNA3- myc-HsSRF (0–5 lg) were transfected in Bsc40 cells together with the reporter vector containing the human cardiac a-actin pro- moter. Luciferase activity value 100 was as- signed to the sample without expression vector (column 0). Statistical significance is referred to differences with the activity obtained with- out expression vector. (B) The indicated amounts of A. franciscana SRF expression vector (pcDNA3-myc-AfSRF) were cotrans- fected together with 5 lg of the reporter vector containing human cardiac a-actin promoter. Relative luciferase activities and statistical significances are as indicated in panel A.C The amounts indicated of D. melanogaster expression vector (pcDNA3-myc-DmSRF) were cotransfected with 5 lg of the reporter vector containing the c-fos promoter. Luciferase activity and statistical significances are expressed as indicated in (A.). Fig. 3. Association of expressed SRFs with endogenous SRF. Nuclear extracts were obtained from C 2 C 12 cells expressing A. franciscana SRF-myc (C 2 Af), human SRF-myc (C 2 Hs) or D. melanogaster SRF- myc (C 2 Dm), and from nontransfected C 2 C 12 cells. (A) Experiments where nuclear extracts were immunoprecipitated with monoclonal anti-myc Ig (a-myc) or without antibody (–) and the immunoprecipi- tates analyzed by Western blot using anti-SRF Ig that recognize human and mouse SRFs, as indicated to the right. One tenth of the nuclear extracts used for immunoprecipitation was loaded on samples E, as positive control of SRF expression. The experiment shown in (B) was made using anti-SRF Ig for immunoprecipitation (a-SRF), or no antibody (–), and anti-myc Ig for Western blotting. Anti-myc Ig rec- ognized human (HsSRF-myc), D. melanogaster (DmSRF-myc) and A. franciscana (AfSRF-myc) SRFs fused to the myc epitope whose migration is indicated to the right. Samples labeled as E contain one tenth of the nuclear extracts used for immunoprecipitation. Ó FEBS 2002 Functionality of arthropod SRFs in mammalian cells (Eur. J. Biochem. 269) 3673 c-fos (Fig. 4A) or CA (Fig. 4B) promoters was determined. Expression of the hybrid molecules was tested by Western blot using anti-(myc epitope) Ig (Fig. 4C). Only hybrid molecules containing the human C-terminal region (AH and DH constructs) were able to increase CA promoter activity, even if the N-terminal region and MADS box were from D. melanogaster (DH) or A. franciscana (AH). How- ever, the presence of A. franciscana or D. melanogaster C-terminal regions produced inhibition of c-fos promoter activity and did not activate CA promoter, even in the presence of the human N-terminal region and MADS box (HD and HA constructs). These results strongly suggest that the transactivation domain present in the C-terminal region of vertebrate SRFs is not functionally conserved in the invertebrate D. melanogaster and A. franciscana factors. Mapping of A. franciscana SRF transactivation domains The lack of the C-terminal transactivation domain could explain the results obtained for DmSRF. This protein would inhibit the activity of the endogenous SRF through a dominant negative effect, as previously described for truncated vertebrate SRFs [37]. The activation observed after cotransfection of AfSRF is, however, in apparent contradiction with the absence of a transactivation domain in the C-terminal region of this protein. One possible explanation would be the existence of a transactivation domain in a different region of this protein. This possibility was tested through the deletion of several regions of AfSRF. Seven progressive deletions of the region N-terminal to the MADS box (N1 to N7), a construct lacking the region C-terminal to the MADS box (C1), and a short construct containing only the MADSbox and the SAM domain, conserved between all known SRF proteins (M), were generated (Fig. 5A). The constructs were cloned in the pcDNA-3 vector, fused to a c-myc epitope, and cotrans- fected in Bsc1 cells together with reporter vectors containing CA and Act403mut promoters, that had been previously showntobeactivatedbyAfSRF co-expression. The results obtained, shown in Fig. 5B,C, confirmed that the C-terminal region of AfSRF is not necessary for transcrip- tional activity as the C1 construct has the same activity than the complete SRF. However, deletion of the N-terminal region did not abolish transcriptional activity completely, either. A partial, although significant, decrease in transcrip- tional activity could be observed between deletions N3 and N4 on the Act403mut promoter only (Fig. 5B). These results suggest that neither of the regions outside the MADS box, N- or C-terminal, are essential for transactivation by AfSRF. However, it appears that at least one of the regions must be present in combination with the MADS-box to achieve transcriptional activity, since co-expression of the MADS box alone (M construct) does not activate Act403mut promoter. The expression of the deleted mole- cules was tested by Western blotting using anti-(c-myc) Ig (Fig. 5D). Fig. 4. Functionality of interspecies hybrid SRF molecules. SRF hybrid molecules were generated that contained the cDNA region coding for the N-terminal and MADS-box domain from one species and the region coding for the C-terminal domain from another species. Hybrid molecules were cloned in the pcDNA3 vector, fused to the myc epitope. Expression vectors coding for human (Hs), D. melanogaster (Dm), A. franciscana (Af) and hybrids (H/D, D/H, H/A, A/H, D/A, A/D) were cotransfected in Bsc1 cells with reporter vectors containing c-fos (A) or human cardiac a-actin (B) promoters. Hybrid SRF molecules have been named according to their components: the species contributing the N-terminal part of the protein is indicated by the first letter and the species that provides the C-terminal region by the second letter (H, H. sapiens;D,D. melanogaster; A, A. franciscana). Columns C show the activity obtained without SRF co-expression and were assigned the relative luciferase activity 100. (C) The analyses of expression of human (Hs), D. melanogaster (Dm), A. franciscana (Af) and hybrid molecules (H/D, D/H, H/A, A/H, D/A, A/D) by Western blot using anti-myc Ig. Ten micrograms of the cell extracts obtained in the transfection shown in (A) were analyzed. The migration of molecular weight markers is indicated in the left margin. 3674 S. Avila et al. (Eur. J. Biochem. 269) Ó FEBS 2002 DISCUSSION In this article, we have compared the function of SRF molecules from a vertebrate, Homo sapiens,andtwo invertebrates, the insect D. melanogaster and the crustacean A. franciscana, in cultured mammalian cells. Complemen- tary studies in invertebrate cells were hindered by the absence of A. franciscana cultured cell lines. There are also no available cell lines derived from D. melanogaster tissues where SRF plays a relevant developmental function, such as terminal tracheal cells and wing intervein cells. SRF binds to DNA through SREs, whose consensus nucleotide sequence, named CArG box, is CC(A/T) 6 GG [38]. Promoters with very similar CArG boxes show very different SRF-dependent regulation, probably due to the presence of binding sites for other transcription factor around the CArG box [14]. We have therefore assayed the activity of the SRF molecules on two differently regulated vertebrate promoters: the promoter of the c-fos gene, stimulated by serum treatment of quiescent cells, and the promoter of the muscle-specific cardiac a-actin (CA)gene. Small fragments of both proximal promoter regions, that had been shown to be sufficient for SRF-dependent regula- tion, were used in these studies. The CA promoter contained the two more proximal CArG boxes that have been identified [9]. Besides, we have analyzed an invertebrate promoter, the A. franciscana Act03 gene promoter, which is probably regulated by SRF in this organism, as SRF and Act403 have the same pattern of tissue-specific expression [21]. A mutated Act403 promoter, with decreased affinity for SRF [30], was also analyzed because preliminary studies had shown that this promoter is activated by SRF co-expression. The activity obtained for these promoters was markedly different. C-fos and Act403 promoters showed high activity in the transfected cells while CA and the Act403mut promoter were much less active. The lower activity of the Act403mut promoter is due to decreased affinity to SRF. In the case of the cardiac actin promoter the reason for its lower activity is not known, although it seems to depend on sequences outside of the CArG box, as mentioned above. The lower activity could be due to binding of transcriptional repressor mole- cules to this promoter, to the absence of binding of SRF coactivators or both. As the cell lines tested are not of cardiac muscle origen, they possibly do not contain cardiac-specific Fig. 5. Mapping domains of A. francisc ana SRF required for transcriptional activation. (A) Diagram of the deletions generated to map putative transcriptional activation domains of A. franciscana SRF. The upper sketch shows the position of the conserved MADS-box and SAM domain in A. franciscana SRF, with the first and last amino acids indicated on top. The name given to each construct is indicated to the right (N1 to N7, C1 and M). Numbers to the left indicate the N-terminal amino acid encoded by each construct. Constructs C1 and M terminated translation at amino acid 219, as indicated to the right of the diagrams. All the constructs were cloned in pcDNA3-myc for cellular expression. (B) Luciferase activities obtained after co-expression of full-length A. franciscana SRF (Af) or the deleted proteins (N1 to N7 and C1) with reporter vectors containing human cardiac a-actin (CA) or Actin 403 mutated (Act403 mut) promoters in Bsc1 cells. Samples C show the activity obtained without any SRF co-expression, that was given the value 100. Asterisks indicate statistically significant differences in relation with the activity obtained after full- length SRF co-expression (Af). (C) Full-length A. franciscana SRF (Af) and deletions coding only for the conserved MADS-box and SAM domain (M) or the conserved regions and the C-terminal region (C1) were cotransfected with reporter vectors containing the Act403 mutated promoter. Sample C was transfected with the reporter vector alone and was assigned the relative activity 100. (D) Western blot analyses of full-length and deleted proteins expression. Ten micrograms of each cellular extract were analyzed using an anti-myc Ig. Migration of molecular mass markers is showntotheleft. Ó FEBS 2002 Functionality of arthropod SRFs in mammalian cells (Eur. J. Biochem. 269) 3675 cofactors that could participate in the activation of this promoter. The consequences of human SRF co-expression on the activity of these promoters were also different. The less active CA and Act403mut promoters were stimulated 12–20 times by human SRF co-expression in three different cell lines and using two expression vectors, pcDNA3 and pSG5. The levels of exogenous SRF expression were similar to those of endogenous SRF, at least in the stably expressing C 2 C 12 cells. Activation of cardiac a-actin promoter by SRF cotrans- fection has been described previously [39]. Other cardiac- and smooth muscle-specific gene promoters are also activated by SRF cotransfection, such as those of smooth muscle a-actin [40], atrial natriuretic factor [41] or SM22a genes [42]. There are several possible mechanisms that could explain promoter activation by an increase in SRF expres- sion. Higher SRF concentration might increase the amount of SRF bound to low affinity promoters, which could be the case for the Act403mut promoter. The increase in SRF concentration may also compete with inhibitory molecules bound to the CArG box or to SRF itself, either bound to DNA or in solution. Several inhibitory factors that bind to SRF or compete with SRF-binding to the promoter have been described [43]. This later mechanism could be respon- sible for the activation of the CA promoter. The repressor effect of DmSRF and, to a lesser extend, AfSRF could be explained by the absence of transactivation domains functional in mammalian cells. Vertebrate SRF molecules without transactivation domain have been shown to act as dominant negative transcription factors [37]. This possible explanation would require interaction between DmSRF and AfSRF, and the endogenous mouse SRF. This interaction was demonstrated by co-immunoprecipita- tion experiments. Hybrid proteins were expressed to further test for this possibility. These molecules contained the N-terminal region, including the MADS box and SRF- conserved region, from one species and the C-terminal region from another. The results obtained were in agreement with the hypothesis that the C-terminal transactivation domain is not conserved between species. Hybrid molecules containing D. melanogaster or A. franciscana N-terminal regions and human C-terminal region were able to activate transcription from the CA and c-fos promoters. The activity of these constructs on c-fos promoter was significantly higher than that of human SRF, which could be due to the absence of repressor domains localized in the N-terminal region of this molecule [7]. These results suggest the capacity of the N-terminal region of SRFs from these organisms to bind DNA and dimerize with the mammalian SRF and of the human C-terminal region to activate transcription. In contrast, hybrid molecules that contained either D. mela- nogaster or A. franciscana C-terminal regions did not acti- vate CA promoter and strongly repressed the c-fos promoter, independently of the origin of the N-terminal region. The lack of activity of the A. franciscana C-terminal region was unexpected as cotransfection of this SRF molecule stimulated CA and Act403mut promoters. Dele- tion experiments confirmed that AfSRF molecules lacking the C-terminal region still activated CA-promoter-depen- dent transcription. More extensive deletion experiments did not allow to unequivocally locate a transactivation domain in AfSRF. Molecules lacking the region N-terminal to the MADS box also activated transcription, which would suggest that the transcriptional activation domain is located in the MADS box region. However, the conserved MADS box and immediately C-terminal SAM domain, devoid of the rest of N- and C-terminal regions, did not activate transcription, suggesting that this region is necessary but not sufficient for transcriptional activation. It is possible that nonconserved N- or C-terminal regions could be necessary for the proper structure of the MADS box. Alternatively, cofactors implicated in activation could bind SRF through the conserved MADS box and other nonconserved N- or C-terminal Af SRF regions. Despite the results discussed above, a significant decrease in Act403mut promoter activation was observed in cells after deletion of the region corresponding to amino acids 45–60. These results suggest that this region can act as a transac- tivation domain in a promoter-dependent manner. This region includes a small evolutionary conserved domain that has been reported to be phosphorylated by several protein kinases in vertebrates [21]. Recently, Hanlon et al. [44] have proposed that phosphorylation of Ser103 of vertebrate SRF, located in this conserved region, promotes interaction between SRF and C/EBPa to activate transcription. In summary, the data obtained in this study indicate the conservation of the SRF DNA-binding and dimerization domains during evolution and the divergence of the transactivation domain. These studies have been carried out in mammalian cells and they might not be translated to the cellular environment of the other species. There could be transactivation domains functional in D. melanogaster and A. franciscana that have not conserved the capacity to interact with mammalian cofactors or with the basic transcriptional machinery. This lack of conservation would be in agreement with the very different patterns of expres- sion and physiological functions of SRF in these organisms, as described in the Introduction. It seems possible that each SRF molecule could have evolved to interact with different tissue-specific cofactors in the different species, which would explain the divergence of their transactivation domains. ACKNOWLEDGEMENTS We are indebted to Dr Rosario Perona for donation of several vectors and reagents and to Drs Rosario Perona and Ricardo Escalante for critical reading of the manuscript. This work was supported by Grant PB98-0517 from the Direccio ´ n General de Investigacio ´ n. REFERENCES 1. Johansen, F.E. & Prywes, R. (1995) Serum response factor: transcriptional regulation of genes induced by growth factors and differentiation. Biochim. Biophys. Acta 1242, 1–10. 2. Pellegrini,L.,Tan,S.&Richmond,T.J.(1995)Structureofserum response factor core bound to DNA. Nature 376, 490–498. 3. Shore, P. & Sharrocks, A.D. (1995) The MADS-box family of transcription factors. Eur. J. Biochem. 229, 1–13. 4. Theiben, G., Kim, J.T. & Saedler, H. (1996) Classification and phylogeny of the MADS-box multigene family suggest defined roles of MADS-box gene subfamilies in the morphological evo- lution of eukaryotes. J. Mol. Evol. 43, 484–516. 5. Treisman, R. (1995) Journey to the surface of the cell: Fos regu- lation and the SRE. EMBO J. 14, 4905–4913. 6. Ling, Y., West, A.G., Roberts, E.C., Lakey, J.H. & Sharrocks, A.D. (1998) Interaction of transcription factors with serum response factor. Identification of the Elk-1 binding surface. J. Biol. Chem. 273, 10506–10514. 3676 S. Avila et al. (Eur. J. Biochem. 269) Ó FEBS 2002 7. Johansen, F.E. & Prywes, R. (1993) Identification of transcrip- tional activation and inhibitory domains in serum response factor (SRF) by using GAL4-SRF constructs. Mol. Cell. Biol. 13, 4640– 4647. 8. Liu, S.H., Ma, J.T., Yueh, A.Y., Lees-Miller, S.P., Anderson, C.W. & Ng, S.Y. (1993) The carboxyl-terminal transactivation domain of human serum response factor contains DNA-activated protein kinase phosphorylation sites. J. Biol. Chem. 268, 21147–21154. 9. Minty, A. & Kedes, L. (1986) Upstream regions of the human cardiac actin gene that modulate its transcription in muscle cells: presence of an evolutionarily conserved repeated motif. Mol. Cell. Biol. 6, 2125–2136. 10. Arsenian, S., Weinhold, B., Oelgeschlager, M., Ruther, U. & Nordheim, A. (1998) Serum response factor is essential for mesoderm formation during mouse embryogenesis. EMBO J. 17, 6289–6299. 11. Weinhold, B., Schratt, G., Arsenian, S., Berger, J., Kamino, K., Schwarz, H., Ru ¨ ther, U. & Nordheim, A. (2000) Srf ) / ) ES cells display non-cell-autonomous impairment in mesodermal differ- entiation. EMBO J. 19, 5835–5844. 12. Boxer, L.M., Prywes, R., Roeder, R.G. & Kedes, L. (1989) The sarcomeric actin CArG-binding factor is indistinguishable from the c-fos serum response factor. Mol. Cell. Biol. 9, 515–522. 13. Trouche, D., Grigoriev, M., Lenormand, J L., Robin, P., Leibovitch, S.A., Sassone-Corsi, P. & Harel-Bellan, A. (1993) Repression of c-fos promoter by MyoD on muscle cell differentiation. Nature 363, 79–82. 14. Chang, P.S., Li, L., Mcanally, J. & Olson, E.N. (2001) Muscle specificity encoded by specific serum response factor-binding sites. J. Biol. Chem. 276, 17206–17212. 15. Lints,T.J.,Parsons,L.M.,Hartley,L.,Lyons,I.&Harvey,R.P. (1993) NKx-2.5: a novel murine homeobox gene expressed in early heart progenitor cells and their myogenic descendents. Develop- ment 119, 419–431. 16. Molkentin, J., Lin, Q., Duncan, S.A. & Olson, E.N. (1997) Requirement of the transcription factor GATA-4 for heart tube formation and ventral morphogenesis. Genes Dev. 11, 1061–1072. 17. Wang, D Z., Chang, P.S., Wang, Z., Sutherland, L., Richardson, J.A., Small, E., Krieg, P.A. & Olson, E.N. (2001) Activation of cardiac gene espression by myocardin, a transcrip- tional cofactor for Serum Response Factor. Cell 105, 851–862. 18. Treisman, R. (1996) Regulation of transcription by MAP kinase cascades. Curr. Opin. Cell Biol. 8, 205–215. 19. Guillemin, K., Groppe, J., Du ¨ cker, K., Treisman, R., Hafen, E., Affolter, M. & Krasnow, M.A. (1996) The pruned gene encodes the Drosophila serum response factor and regulates cytoplasmic outgrowth during terminal branching of the tracheal system. Development 122, 1353–1362. 20. Montagne, J., Groppe, J., Guillemin, K., Krasnow, M.A., Gehring, W.J. & Affolter, M. (1996) The Drosophila Serum Response Factor gene is required for the formation of intervein tissue of the wing and is allelic to blistered. Development 122, 2589– 2597. 21. Casero, M.C. & Sastre, L. (2000) A serum response factor homologue is expressed in ectodermal tissues during development of the crustacean Artemia franciscana. Mechanism Dev. 96, 229– 232. 22. Escalante, R. & Sastre, L. (1998) A serum response factor homolog is required for spore differentiation in Dictyostelium. Development 125, 3801–3808. 23. Escalante, R., Vicente, J.J., Moreno, N. & Sastre, L. (2001) The MADS-Box gene srfA is expressed in a complex pattern under the control of alternative promoters and is essential for different aspects of Dictyotelium development. Dev. Biol. 235, 314–329. 24. Yaffe, D. & Saxel, O. (1977) Serial passaging and differentiation of myogenic cells isolated from dystrophic mouse muscle. Nature 270, 725–727. 25. Hopps, H.E., Bernheim, B.C., Nisalak, A., Tjio, J.H. & Smadel, J.E. (1963) Biologic characteristics of a kidney cell line drived from the African green monkey. J. Inmunol. 91, 416–424. 26. Chen, C. & Okayama, H. (1987) High-efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745–2752. 27. Graham, F.L. & Eb, A.J. (1973) Transformation of rat cells by DNA of human adenovirus 5. Virology 54, 536–539. 28. Murguı ´ a, J.R., De Vries, L., Gomez-Garcı ´ a, L., Scho ¨ nthal, A. & Perona, R. (1995) Activation of C-Fos promoter by increased internal pH. J. Cell. Biochem. 57, 630–640. 29. Nordeen, S.K. (1988) Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6, 454–458. 30. Casero, M C. & Sastre, L. (2001) Characterization of a functional serum response element in the Actin403 gene promoter from the crustacean Artemia franciscana. Eur. J. Biochem. 268, 2587–2592. 31. Moe ¨ ns, U., Subramaniam, N., Johansen, B. & Aarbakke, J. (1993) The c-fos cAMP-response element: regulation of gene expression by a beta 2-adrenergic agonist, serum and DNA methylation. Biochim. Biophys. Acta 1173, 63–70. 32. Roth, M.B., Zahler, A.M. & Stolk, J.A. (1991) A conserved family of nuclear phosphoproteins localized to sites of polymerase II transcription. J. Cell Biol. 115, 587–596. 33. Norman, C., Runswick, M., Pollock, R. & Treisman, R. (1988) Isolation and properties of cDNA clones encoding SRF, a trans- cription factor that binds to the c-fos serum response element. Cell 55, 989–1003. 34. Affolter, M., Montagne, J., Walldorf, U., Groppe, J., Kloter, U., Larosa, M. & Gehring, W.J. (1994) The Drosophila SRF homolog is expressed in a subset of tracheal cells and maps within a genomic region required for tracheal development. Development 120, 743– 753. 35. Green, S., Issemann, I. & Sheer, E. (1988) A versatile in vivo and in vitro eukaryotic expression vector for protein engineering. Nucleic Acids Res. 16, 369. 36. Andrews, N.C. & Faller, D.V. (1991) A rapid micropreparation technique for extraction of DNA-binding proteins from limiting numbers of mammalian cells. Nucleic Acids Res. 19, 2499. 37. Belaguli, N. & Zhou, W. (1999) Dominant negative murine serum response factor: alternative splicing within the activation domain inhibits transactivation of serum response factor binding targets. Mol. Cell. Biol. 19, 4582–4591. 38.Pollock,R.&Treisman,R.(1990)Asensitivemethodforthe determination of protein-DNA binding specificities. Nucleic Acids Res. 18, 6197–6204. 39. Belaguli, N.S., Sepulveda, J.L., Nigam, V., Charron, F., Nemer, M. & Schwartz, R.J. (2000) Cardiac tissue enriched factors Serum Response Factor and GATA-4 are mutual coregulators. Mol. Cell. Biol. 20, 7550–7558. 40. Bushel, P., Kim, J.H., Chang, W., Catino, J.J., Ruley, H.E. & Kumar, C.C. (1995) Two serum response elements mediate transcriptional repression of human smooth muscle alpha-actin promoter in ras-transformed cells. Oncogene 10, 1361–1370. 41. Zhang, X., Chai, J., Azhar, G., Sheridan, P., Borras, A.M., Furr,M.C.,Khrapko,K.,Lawitts,J.,Misra,R.P.&Wei,J.Y. (2001) Early postnatal cardiac changes and premature death in transgenic mice overexpressing a mutant form of serum response factor. J. Biol. Chem. 276, 40033–40040. 42. Kemp, P.R. & Metcalfe, J.C. (2000) Four isoforms of serum response factor that increase or inhibit smooth-muscle-specific promoter activity. Biochem. J. 345, 445–451. 43. Gualberto, A., Lepage, D., Pons, G., Mader, S.L., Park, K., Atchison, M.L. & Walsh, K. (1992) Functional antagonism between YY1 and the serum response factor. Mol. Cell. Biol. 12, 4209–4214. 44. Hanlon, M., Sturgill, T.W. & Sealy, L. (2001) ERK2- and p90 Rsk2 - dependent pathways regulated the CCAAT/Enhancer-binding Protein–b interaction with Serum Response Factor. J. Biol. Chem. 276, 38449–38456. Ó FEBS 2002 Functionality of arthropod SRFs in mammalian cells (Eur. J. Biochem. 269) 3677 . Transactivation domains are not functionally conserved between vertebrate and invertebrate serum response factors Sonia Avila, Marie-Carmen. boxes and SAM domains have been very conserved during evolution. MADS-box regions are 98% identical and SAM -domains over 79% identical between vertebrate and