Identification, Structural, and Functional Characterization of a New Early Gene (6A3-5, 7 kb): Implication in the Proliferation and Differentiation of Smooth Muscle Cells

Arterial smooth muscle cells (SMCs) play a major role in atherosclerosis and restenosis. Differential display was used to compare transcription profiles of synthetic SMCs to proliferating rat cultured SMC line. An isolated cDNA band (6A3-5) was shown by northern (7 kb) to be upregulated in the proliferating cell line. A rat tissue northern showed differential expression of this gene in different tissues. Using 5′ RACE and screening of a rat brain library, part of the cDNA was cloned and sequenced (5.4 kb). Sequence searches showed important similarities with a new family of transcription factors, bearing ARID motifs. A polyclonal antibody was raised and showed a protein band of 175 kd, which is localized intracellularly. We also showed that 6A3-5 is upregulated in dedifferentiated SMC (P9) in comparison to contractile SMC ex vivo (P0). This work describes cloning, structural, and functional characterization of a new early gene involved in SMC phenotype modulation.


INTRODUCTION
Migration and proliferation of smooth muscle cells (SMCs) into the intima plays a key role in the initiation and perpetuation of atherosclerotic lesions [1,2,3]. Indeed, arterial SMCs are a major component of atherosclerotic plaques and restenotic vessels. According to Ross [4], proliferation of SMCs in atherosclerotic lesions is the result of an excessive inflammatory fibroproliferative response to various forms of insult to the endothelium. In these diseased vessel walls, SMCs undergo a phenotypic modulation [5,6] where they change from a highly contractile, fully differentiated, state to a synthetic and/or proliferating dedifferentiated phenotype [4,7,8]. Subsequently, SMCs are transformed into foam cells by Correspondence  This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. accumulating lipids [9,10,11]. Harvested SMCs, under in vitro conditions, progressively lose their highly contractile phenotype to another phenotype that mimics synthetic SMCs present in diffuse intimal thickening [11,12]. In long-term cultures, aortic SMCs generate a proliferating transformed phenotype [13,14] with similarities to proliferating cells [15].
Differences have been observed, at the gene and protein level, between the contractile and the synthetic/proliferating phenotypes. However, at this stage, a greater understanding of the genes implicated in SMC phenotypic differentiation is vital to further understand the pathogenesis of atherosclerosis [16]. In the present study, rat SMCs showing synthetic (subcultures at passage 9) or highly proliferating (spontaneously growing V8 cells) phenotypes were compared with regards to their gene expression by differential display [17]. The rationale for comparing these cell cultures relies on the similar changes in SMC phenotypes that occur in the formation and progression of vascular lesions. Results obtained allowed the identification of a new transcription factor gene, bearing an ARID motif (AT-rich interaction domain), present at high levels in proliferating cultured SMCs. This gene may play an important role in SMC differentiation and proliferation.

Differential display analysis
Differential display was performed as previously described [17] (RNAimage, GenHunter). Briefly, (i) reverse transcription (RT) reaction: 0.2 µg of total RNA from each sample was reverse transcribed with 100 U MMLV reverse transcriptase in the presence of 250 µM dNTPs and 2 µM H-T 11 M (M can be either dA, dG, dC, and H is the Hind III restriction site). The 20-µL RT reaction was reverse transcribed for 1 hour at 37 • C, then the enzyme is denatured by heating at 75 • C for 5 minutes. (ii) PCR amplification: 2 µL of the single-strand (ss) cDNA mixture thus obtained was used for 8 different PCR reactions, each containing a different arbitrary primer from the 5 end. The 18-µL PCR mix included 2 µM of the H-T 11 M primer (same as RT), 2 µM of a specific arbitrary primer, 25 µM dNTPs with 0.25 µL of α-33 P dATP (2000 Ci/mmole, Amersham, UK) and 1 U of Taq DNA polymerase (Perkin Elmer Mass, USA). Thermal cycling amplification parameters (40 cycles) using GeneAmp PCR System 9600 (Perkin Elmer) were as follows: 94 • C (15 seconds), 40 • C (2 minutes), 72 • C (30 seconds), and a final 5 minutes extension step at 72 • C. (iii) Separation by electrophoresis: only 3.5 µL of the PCR products was separated on a 6% denaturing polyacrylamide gel in TBE buffer after addition of 2 µL loading dye (95% formamide, 10 mM EDTA, pH 8.0, 0.09% xylene cyanole, and 0.09% bromophenol blue). The gels were run for 4 hours at 1400 V, dried without fixation for 2 hours at 80 • C, exposed for 72 hours, and then visualized by autoradiography.

Band recovery, cloning, and sequencing
The fingerprinting of the amplified fragments was common between the two cell types under study. (i) Differentially expressed bands (up-or downregulated) were recovered under sterile conditions by excising the gel slice from the dried gel using a razor blade. Each gel slice was placed in 100 µL sterile water, boiled for 15 minutes to solubilize the DNA, and then ethanol precipitated in the presence of glycogen. The pellet was resuspended in sterile water after ethanol washing. (ii) The reamplification is done with 4 µL of purified fragment using the same primer pair and PCR parameters that gave rise to the band. (iii) Reamplified DNA fragments were run on a 1.5% agarose gel. Bands that succeeded to be reamplified were cloned into PCR II vector (TA cloning kit, Invitrogen, The Netherlands). (iv) For DNA sequencing, minipreps of plasmidic DNA were carried out [19], followed by the dideoxy sequencing method [20] (T7 Sequencing Kit, Pharmacia, France).

Bioinformatics
The sequences obtained were compared with known sequences by similarity searching in the different databases (GenBank, EMBL, EST, STS, etc.) using the BLAST [21] and FASTA [22] programs. The multiplesequence alignment was carried out using the Omiga 2.0 Software (Oxford molecular, UK).

Probes and northern blot
Total RNA was extracted as above, denatured, separated by electrophoresis in formaldehyde-MOPS-agarose gel and then transferred to a nylon membrane (Hybond, Amersham, UK). After capillary blotting performed overnight, the membrane was baked for 2 hours at 80 • C. Probes for northern blots were prepared following the random priming method (High Prime, Boehringer, Germany), using the PCR amplified inserts in the PCR II vector described above, and then purified using G-sephadex (Quick Spin Columns, Boehringer). Prehybridization and hybridization were done according to standard protocols [23]. Blots were exposed, at −70 • C, with intensifying screens against a Kodak film for one week. Similar loading of RNA was assessed by using the actin probe. The following primers (Eurogentec, Belgium) were used for the preparation of cdk2α probe by RT-PCR (see below): cdk2α up: ACGGAGTGGTGTACAAAGCC, cdk2 down: GAGTCTCCAGGGAATAGGGC.

rapid amplification of c-DNA ends (RACE)
To obtain the upstream 5 region of the new gene, the 5 RACE technique was carried out basically by applying the touchdown PCR principle [24] and by using Marathon cDNA amplification and Advantage KlenTaq polymerase kits (Clontech Calif, USA). (i) In the first step, ss cDNA is synthesized with 1 µg of V8 poly(A+) RNA, using 10 µM of the cDNA synthesis primer and MMLV-RT for 1 hour at 42 • C. DNA synthesis was verified by the addition of dNTPs among which one was radiolabeled α − 32 P dCTP (1 µCi/µL, NEN, France). (ii) The second step was the synthesis of ds DNA carried out at 16 • C for 3 hours in an enzyme mixture containing E coli DNA polymerase I, Rnase H, and E coli DNA ligase. These enzymes allow the synthesis of ds cDNA, RNA degradation, and the formation of blunt ends, respectively. A 1% agarose gel electrophoresis is done to estimate the quantity and quality of the ds cDNA synthesized. The gel is then dried and put in contact with a Kodak film at −70 • C in order to visualize the DNA smear. (iii) The third step allows us to obtain a library of ds cDNA, from V8 cells, by ligating an adapter to both ends of the ds cDNA, using a T4 DNA ligase at 16 • C overnight. (iv) In the last step, an aliquot of the library is subjected to PCR. The 50-µL PCR reaction contains 10 µM dNTP, 10 µM of the adapter primer (complementary to the cDNA adapter), 5 µL of the 50x KlenTaq polymerase, and 10 µM of gene-specific primer (GSP) complementary to the 3 differentially expressed fragment (6A3-5 GSP: 5 -GTATTACAGTTTTAGGGAAGTGAATTC-3 ). The mixture was subjected to a PCR step at 94 • C (1 minute); followed by 33 cycles of 94 • C (30 seconds), 60 • C (30 seconds), and 68 • C (2 minutes and 15 seconds); and a 5 minutes extension step at 72 • C. The amplified DNA fragments were cloned into the PCR II vector and purified using Qiagen Plasmid Midi Kit (Qiagen). The insert DNA is then sequenced commercially (Genome Express, France).

Screening of a rat brain cDNA library
A cDNA library originating from the rat brain and containing hard-to-clone 5 end of long cDNAs was purchased from OriGene Technologies, Md, USA. Screening was done according to manufacturer's guidelines. Briefly, the 96-well master plate was screened by PCR using genespecific primers that were constructed from the previously cloned 1.2 kb. The following primers were used: 6A3-5 U18: TTGGGGATCGCAAAAACC, 6A3-5 L21: TAGT-GAATGGGGCAGAGAAGC. The cycling conditions (40 cycles) were as follows: 94 • C (30 seconds); 94 • C (15 seconds), 60 • C (45 seconds), 72 • C (1 minute), and a final extension step of 5 minutes at 72 • C. After identification of a positive well, a 96-well subplate containing dilutions of the master positive well is then screened. The same genespecific primers are then used on the subplate and positive wells identified. Bacteria are then plated and a clone of interest is isolated by filter hybridization. The positive clone is then inoculated, purified, and sequenced after midiprep plasmid preparation.

Quantitative competitive RT-PCR
The quantitative competitive RT-PCR was performed as described [25,26]. Briefly, this technique is based on the addition of a known quantity of a serial dilution of an exogenous internal recombinant RNA (RcRNA) standard to a constant quantity of total RNA target RNA sample. Target and internal standard transcripts are reverse transcribed and amplified simultaneously with the same primers. These primers give rise to 2 bands of different molecular weights but of equal intensities when identical number of initial RNA molecules are present. (1) In the first step, synthesis of the RcRNA is done in a 4-step procedure: (i) amplification of the RcDNA using 2 Rc primers constructed by Oligo 5.0 Primer Analysis Software (MedProde, Norway). This is done in a 50-µL PCR reaction containing 10 mM dNTP, IU Taq polymerase, 100 ng of plasmid cDNA, and 10 µM of each Rc primer. (ii) The PCR product is run on a 1.5% agarose gel in order to purify the band by Jetsorb (Bioprobe Miss, USA). (iii) Transcription of the RcRNA (Riboprobe in vitro Transcription, Promega) is done in a 20 µL reaction containing 100 mM DTT, 4 µL rNTP, 20 U RNasine, and 10 U of T7 RNA polymerase for 2 hours at 37 • C. (iv) The product is treated with 0.5 U RQ1 RNase-free DNase for 30 minutes at 37 • C, to eliminate plasmidic DNA, and then the RcRNA concentration is measured by spectrophotometry. (2) In the second step, the RT reaction is carried out in a 10 µL total volume with 10 mM dNTP, 10 mM DTT,

Western blot
(i) Protein extraction: cultured cells are washed with Hanks, trypsinized, and centrifuged at 1200 g during 5 minutes. The cell pellet is then lysed in a lysis buffer containing 1% of 10 mM aprotinin, 10 mM leupeptin, 10 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride (inhibitor cocktail, ICN), 25 mM Tris pH 7.6, 150 mM NaCl, and 1% Triton X100. Cell lysate is then incubated during 40 minutes at 4 • C under agitation. When rat tissues are used, they were maintained at −180 • C in liquid nitrogen and pound in a mortar, then homogenized with a Polytron at 0 • C (two 10 seconds burst) in 50 mM Trisbuffered saline pH 7.6 containing 1% aprotinin, 2 mM ε-aminocaproic acid, and 0.5 mM phenylmethylsulfonyl fluoride. The homogenate is then centrifuged at 3000 rpm for 5 minutes to remove the unhomogenized fragment. Cell-or tissue lysate is then centrifuged at 14000 rpm for 5 minutes to remove cell debris and unlysed fragments, and the supernatant is retained. Quantification of proteins in the supernatant is realized by colorimetry (BCA kit, Pierce, France). (ii) SDS-PAGE: proteins in the supernatant are then diluted in Laemmli buffer, denatured for 5 minutes at 100 • C, and separated on 7% acrylamide SDS-PAGE gels. Migration is done under a constant voltage (100 mV) in a migration buffer (200 mM glycine, 25 mM Tris, 1% SDS). The gel is then equilibrated in the transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol) for 15 minutes. (iii) Transfer of proteins and revelation: nitrocellulose membranes (immobilon P, Millipore) are incubated in methanol for 30 seconds then rinsed in water and equilibrated in the transfer buffer for 15 minutes. The proteins are transferred to the membrane (100 V for 1 hour) then blocked for 2 hours with blocking solution containing 3% gelatin and 0.05% Tween 20 in Tris buffered saline pH 7.5 (TBS). After washing, the membrane is incubated overnight at 4 • C with the rabbit anti-rat 6A3-5 polyclonal antibody (2 µg/mL). Following 3 washing steps of 10 minutes in 0.05% TBS-Tween, a swine anti-rabbit-horse radish peroxidase-conjugated secondary antibody (DAKO) is incubated during 45 minutes at RT. The membrane is then washed (4×10 minutes) in 0.05% TBS-Tween and revealed by a chemiluminescent technique (ECL Kit, Amersham, UK) according to the manufacture's instructions. The ECL films are scanned with Sharp JX-330 scanner (Amersham) and 6A3-5 bands were quantified with appropriate software. In parallel, SDS-PAGE gels, containing identical sample volumes to those used for the western blot were coomassie stained. An electrophoretic band (220 kd) was scanned and used as a quantitative control to check for sample loading and 6A3-5 normalization.

FACScan analysis
Confluent SMCs are trypsinized, centrifuged at 1500 rpm for 5 minutes at 4 • C, and washed with 1% PBS/BSA. Cells are then fixed and permeabilized by 100% methanol at −20 • C during 10 minutes. After washing with 1% PBS/BSA, the rabbit anti-rat 6A3-5 polyclonal antibody (5 µg/mL) is incubated for 1 hour at 4 • C with agitation every 15 minutes. Cells are then washed and centrifuged for 10 minutes at 1200 g. They are then incubated with the secondary antibody (Goat anti-rabbit FITC-conjugated, DAKO) during 45 minutes at 4 • C. Following washing steps, cells are centrifuged during 10 minutes at 1200 g. The cell pellet is then suspended, fixed in 1% PBS/formaldehyde, and stocked away from light at 4 • C before analysis by the FACScan apparatus (Becton Dickinson, France).

Differential display
Fifty bands were differentially expressed between synthetic (P9) and highly proliferating cells (V8, Figure 1). Thirty-six, out of the fifty bands, were selected because of their high molecular weight. Only 22 bands, out of the 36, were successfully reamplified and cloned into PCR II plasmid. Reproducible sequence information, containing flanking sequences corresponding to the particular poly dT and arbitrary primers used for PCR, was obtained for only 16 bands.

Northern blot analysis and tissue distribution of 6A3-5 gene
Four genes (4G2, 5C1, 2A3-2, 6A3-5), showing no similarities in the databases, were confirmed by northern blots to be differentially expressed ( Figure 2a). Three of these genes (5C1, 2A3-2, 6A3-5) were upregulated in the highly proliferating cell line, compared with synthetic cells (Figure 2a). One of these genes (2A3-2) was cloned and characterized in our lab in a previous study [27]. In this work, another gene (6A3-5, 7 kb) upregulated in the highly proliferating cell line, compared with synthetic cells, was further analyzed. Indeed, a rat multiple-tissue northern blot, probed by the 6A3-5 cDNA band, showed this gene (7 kb) to be present in different organs ( Figure  2b). Some tissues such as brain, kidney, and testis showed a very high expression of the gene. Other tissues such as skeletal muscles and heart expressed the gene to a lesser extent. Testis had 3 independent mRNA's that might come from different polyadenylation sites [28]. The multiple northern blot shows that 6A3-5 gene is not an artifact induced by cell culturing but is present in vivo in different tissues.

RACE, screening of a rat brain library, cloning, and sequencing of the 6A3-cDNA
Part of the 5 coding region of 6A3-5 was obtained by 5 RACE using a cDNA library that we constructed from the V8 highly proliferating cells. The size of the 5 RACE-PCR product was 1.2 kb while the full length mRNA size was determined, by northern, to be 7 kb. The 5 RACE-PCR product ( Figure 3) was amplified, purified, cloned, and sequenced. This original 6A3-5 nucleotide sequence was then sent to GenBank and to the European Molecular Biology Laboratory (EMBL) to get an accession number (AJ005202). Gene-specific primers were then constructed and used on a cDNA library originating from the rat brain and containing long cDNAs. Screening of the brain library allowed the isolation of a specific clone that was fully sequenced (5.4 kb). This clone contained the previously identified 1.2 kb.

Characteristics of the 6A3-5 cDNA and protein
The open reading frame of the sequenced part of the gene (5410 bp) was identified and showed to contain 4708 bp running to a TGA stop codon ( Figure 4). This sequence contained poly-CAG repeats between nucleotides 3896 and 3913. The 5 untranslated region, as well as the uppermost 5 coding region, has not yet been cloned. The cDNA contained 681 bp in the 3 -untranslated region with a typical poly-A signal (AATAAA) that was determined 73 bp upstream of the poly-A tail [29]. On the protein level, 6A3-5 had an ARID domain (187-296), LXXLL motif (1177-1181), a Q-rich region (1298-1304), a serine-rich region (112-175), and a phenyl-rich region (1472-1481) ( Figure 5). Analysis of the 6A3-5 protein fragment revealed the presence of multiple glycosylation sites, phosphorylation sites, myristyl sites, and amidation sites. The hydropathy analysis data indicated that there were no significant hydrophobic transmembrane domains. , showing no similarities in the databases, were confirmed by northern blots to be differentially expressed. The 6A3-5 gene is upregulated in proliferating (V8) but not synthetic cells (P9). Quantification of 6A3-5 signals (n = 3), reported to 28S levels, showed a 3-fold increase in the V8 compared to the P9 cells. The 6A3-5 mRNA has a size of 7 kb as given by northern blot. The internal deposition control of the same RNA quantity is given by 28S. Lanes P9 and V8 correspond respectively to synthetic and rapidly proliferating cells. (b) Multiple-tissue northern blot analysis with the 6A3-5 cDNA band in the rat. The blot contained 20 µg of total RNA from various rat tissues and was probed with the 6A3-5 cDNA fragment isolated by DD. Transcripts of ∼ 7 kb could be observed in all rat tissues analyzed, but at different levels of expression. Indeed, brain, kidney, and testis tissues expressed this gene at very high levels. Two lower transcripts of ∼ 6 and ∼ 5 kb were also observed for testis. Nucleotide similarity search DNA FASTA search program was used to search for sequences showing relationships to rat 6A3-5 (see Table  2). Similarity searches revealed important similarities mainly with mouse, rat, and human ESTs. The highest similarities with 6A3-5 were with the following. (1) An EST (99% identity) coming from rat PC12 cells [30]. This EST clone could not be reproduced by the TIGR ΦX174 6A3-5 1.2 Kb Figure 3. 5 RACE-PCR amplification. A cDNA library (obtained from V8 proliferating cells) was used along with a primer coming from the 3 end of the 6A3-5 band. The other primer in the PCR comes from the adapter that is already ligated to the cDNA library. The touchdown PCR technique was used during the 5 RACE, which allowed us to obtain a part of the cDNA. The size of the 5 RACE-amplified 6A3-5 fragment is 1.2 kb. φX174 is given on the left as a molecular weight marker. The lanes represent different dilutions of the cDNA library.
institute due to contamination problems. (2) A newly identified human clone (92% identity, KIAA1235) originating from a brain library. This partially sequenced clone (5.3 kb) contains an ARID domain (AT-rich interaction domain). It is known that genes of the ARID family are important for binding to DNA [31,32]. (3) A cDNA product (72% identity, b120) whose coding sequence was cloned as part of a search for genes containing CAG repeats [33]. (4) p270 cDNA (72% identity) which is also a transcription factor of the ARID family. It is interesting to note that b120 sequence appears to be a portion of p270, but whose coding sequence contains a frame-shift that gives rise to a truncated p270.

Protein similarity search
FASTA program identified several proteins with statistically significant degree of relationship to rat 6A3-5 (Table 3). Proteins with significant similarity to r6A3-5 include the following. (1) A translated human brain KIAA1235 clone (99% identity). (2) p270, an ARID transcription factor (78% identity) which was first identified through its shared antigenic specificity with p300 and CREB binding protein (CBP). This protein (p270 or SWI1) is member of the SWI-SNF complex which is implicated in the regulation of the transcription by modifying the conformation of nucleosomes [34,35,36]. (3) b120 protein (78% identity) which is highly expressed in skeletal muscles and the brain. It was suggested to be implicated in lipid metabolism and could be responsible for Schnyder crystalline corneal dystrophy [37]. (4) Eyelid protein (eld, also referred to as OSA) which is another transcription factor of the ARID family. Our protein sequence had 52% identity with the eld protein. Eyelid is an ubiquitous expressed protein involved in embryonic growth, development, and differentiation of the eye in the drosophila (segmentation and photoreceptor differentiation) [38,39]. p270 and eyelid are large proteins with high degree of identity. (5) Finally, there were also interesting similarities to other transcription factors such as IkB epsilon, human BAT2, and APETALA-1. Sequences of KIAA1235, p270, b120, Eld, and Osa genes reveal shared motifs that are potentially functional. They bear a Q-rich region that might be implicated in transactivation functions [40]. They also contain the amino acid motif LXXLL, which has been shown to be critical for the binding of a variety of nuclear proteins to nuclear hormone receptors [41]. Finally, they contain an ARID domain that is implicated in the binding to the DNA. This ARID domain on 6A3-5 sequence runs over 105 aa and has 86% similarity with the other members of the ARID family.

Quantitative RT-PCR analysis of the 6A3-5 gene
Levels of mRNA expression, in synthetic and highly proliferating SMCs, were also measured using quantitative competitive RT-PCR. Recombinant and quantification primers used are given in the methods section. The quantitative competitive RT-PCR on the 6A3-5 gene (Figure 6a) showed its expression to be increased by at least five times in the proliferating (7.5-10 pg) compared to synthetic cells (1-2 pg). RT-PCR was also done using an actin control, on P9 and V8 cells. This was considered as an internal control in order to verify that the same amounts of RNA would give rise to the same number of actin molecules in both cell types (Figure 6b). This control gene was expressed at the same level in both cell types. These results further confirm those observed by northern.

Structural characterization of the 6A3-5 protein in vitro and in vivo
Polyclonal antibodies were raised by rabbit immunization of specific peptides from the predicted rat protein sequence. Antibodies revealed specifically, by western blot, a unique band of 175 kd in V8 SMCs ( Figure  7a). Moreover, the 175 kd protein band was also observed in different rat tissues, but at different levels of expression (Figure 7b). It is worth noting that brain tissues expressed the protein at very high levels. Furthermore, FAC-Scan analysis revealed the presence of this protein only when SMCs were permeabilized, but not in intact cells (Figure 7c). This suggests that 6A3-5 protein is not present on the cell membrane but has an intracellular localization.

6A3-5 gene expression of synthetic SMCs following stimulation by PMA or FCS
Functional characterization of 6A3-5 gene and its implication in SMC proliferation was then studied. Northern blots showed that 6A3-5 gene expression is reduced to a minimum in quiescent and synchronic SMCs after serum depletion (0 minutes), in comparison to levels of expression in standard cell culture conditions (Figure 9). In contrast, 10% FCS induced 6A3-5 gene transcription to a maximum after 1 hour of stimulation (12-fold increase). On the other hand, PMA induced a peak of 6A3-5 after 2 hours of stimulation (4-fold increase). These data indicate that 6A3-5 gene is induced at a very early stage in response to stimuli. Moreover, 6A3-5 mRNA levels decrease after 1-2 hours then increase, after 24 hours, to go back to its normal level observed prior to serum depletion and stimulation.

DISCUSSION
Using differential display, we have identified for the first time a new 7 kb transcription factor gene (6A3-5) that is overexpressed in proliferating, but not synthetic, rat smooth muscle cells. Several lines of evidence back the above statement: (1) differential display shows an upregulation of 6A3-5 in proliferating but not synthetic SMCs. These results were confirmed by northern blot and quantitative competitive RT-PCR. Moreover, rat multipletissue northern showed the presence of this 7 kb mRNA in different tissues. (2) 5 RACE technique, followed by screening of a rat brain library, allowed us to clone and sequence 5.4 kb of the cDNA. (3) This new gene shows, on database search, important similarities to different human EST clones. Strong similarities were observed with transcription factors of the ARID family (AT-rich interaction domain). The ARID motif, which runs over 105 aa and which had 86% similarity with other ARID family members, has been identified, sequenced, and localized on our protein sequence. 6A3-5 also had similarities with functional domains such as the LXXLL motif and a Q-rich region. (4) A polyclonal antibody, raised against a 6A3-5 peptide, showed a 175 kd unique protein band under in vitro and in vivo conditions. (5) FACScan analysis showed that the protein was only accessible after cell permeabilization. (6) 6A3-5 was upregulated, using northern and western blots, in dedifferentiated secretory SMCs in comparison to contractile quiescent phenotype. (7) This new gene was significantly upregulated, in synthetic P9 cells, 1-2 hours following stimulation by PMA or FCS.
Using differential display, we have identified a number of sequences (12) that showed either 80 to 100%, 50 to 80%, or no similarities to known genes. Five genes, in the 80-100% cluster, showed interesting similarities in databases. Indeed, the 4G3-2 sequence had a 93% similarity with the rat glia-derived nexin (GDN) or protease nexin I. This gene is implicated in neurogenesis and neurite growth [42]. In addition, a study showed that a nexin-derived serine protease (GdNPF) is implicated in the migration of the neuronal cells [43]. Moreover, the GDN has similarities [44] with plasminogen activator inhibitor (PAI), antithrombin III (ATIII), and α-1 proteinase inhibitor. The second interesting sequence, 3A2-7, had 97% similarity with autographa calcifornica nuclear polyhedrosis virus. This viral tyrosine/serine phosphatase gene [45] is used for the overexpression of eukaryotic genes [46]. The third sequence, 3A1-1, had a 98% similarity with the glucose-regulated protein GRP78 and rat immunoglobulin heavy chain binding protein. It was demonstrated that GRP78 is upregulated in cells in case of energy restrains [47]. Since this gene is overexpressed Table 2. Similarities of rat 6A3-5 gene with nucleotide sequences in the databases. After comparison to all databases, 6A3-5 had similarities mainly to human EST clones (established sequence tags). One interesting EST sequence was the rat PC-12 EST clone. This rat EST sequence could not be obtained and amplified. Another interesting EST clone was KIAA1235 which is thought to be a transcription factor expressed in the brain. The corresponding gene was estimated to be of 6.5 kb. in the synthetic cells, it can act as a suppressor of the migration and proliferation of synthetic cells. The fourth sequence, 4A1-4, had 93% homology with 2C9 gene which is activated after overexpression of c-fos and is implicated in cellular invasion [48] or metastasis. It is interesting to note that c-myc and c-fos are upregulated during restenosis, so it is possible that this gene is implicated in cellular proliferation. The fifth gene, 4C1-4, had a 92% similarity with mouse embryonal carcinoma F9 clone [49] and with rat assembly protein associated with clathrin vesicles [50]. It is difficult at this stage to identify the relationship between these genes. However, some of them may act in concert following SMC stimulation.
The 6A3-5 cDNA band, following identification by differential display and confirmation by northern blot, was selected for further study as it was observed to be upregulated in the rapidly proliferating SMCs. This gene did not show, at the initial stage of the study, any significant similarity to known genes. DNA database search showed that 6A3-5 has significant similarities to human, rodent, and fruit fly ESTs. Interestingly, 6A3-5 shares important homologies (90-100%) with ESTs originating from human fetal brain, testis, neuronal cells, and numerous cancerous cell line libraries. One of these similarities (99%) was with an EST present in a rat pheochromocytoma PC12 cell line [30] that differentiates into a neuronal Table 3. Similarities of rat 6A3-5 to known proteins in the databases. After comparison to all protein databases (SWISSPROT, TREMBL, PIR, . . .), we had only few similarities to known proteins. The best similarities were with a number of transcription factors. p270 and b120 had 78% identity (83% similarity matching) while eyelid had 52% identity (61% similarity matching) with 6A3-5. phenotype following stimulation by NGF. Another important similarity was with an ARID containing human brain clone called KIAA1235. In addition, ARID-motif bearing transcription factor genes (human p270, human b120, and drosophila eld), albeit with lower similarities to 6A3-5, have been obtained in similarity searches. It is of considerable interest that other transcription factors (IkB epsilon, human BAT2 and APETALA-1) share some similarities to 6A3-5.
On the other hand, when investigating the protein database for structural and functional relationships to 6A3-5, we come across a number of proteins having the new DNA binding motif termed ARID. It is important to note that ARID genes are transcription factors (activators, coactivators, or co repressors) strongly implicated in different physiologic processes such as the regulation of cell growth, development, and tissue-specific gene expression. The ARID domain, which runs over 105 aa and which had 86% similarity with other ARID family members, has been identified, sequenced, and localized on our protein sequence. The presence of an ARID motif on our protein significantly bolsters the role of 6A3-5 as a potential transcription factor since ARID domains are known to be implicated in the binding to DNA. We were particularly interested in the appearance of human p270, human eyelid, and drosophila eyelid in this list of proteins. genes including those required for the mating-type switch and sucrose fermentation pathways [51,52,53]. More recent studies suggested that SWI-SNF complex, in response to control by multiple steroid hormone receptors [54,55,56,57], also has a more general role in the regulation of gene expression during cell growth and development in all organisms [58,59]. Moreover, the complex has a general nucleosome-remodelling activity that can be upregulated in response to various signals. It is of interest to note that the drosophila eyelid [38] protein is implicated in embryonic development and is thought to be a transcription factor acting as an antagonist to the wingless (Wg) pathway. In fact, target genes in this pathway are activated in the absence of eyelid and inhibited in the presence of an excess of the gene. One should note that the rat homolog to human p270 is not yet known. Moreover, human and rat homologs to drosophila eyelid have yet to be identified. However, 6A3-5 appears to be a homolog of the human brain clone (KIAA1235). It is conceivable that both 6A3-5 and KIAA1235 are the homologs of drosophila eyelid gene. Other proteins with ARID regions, but with no similarities to 6A3-5, include human and murine bright, drosophila DRI and its human homolog DRILI, the CMV enhancer binding proteins MRF-1 and MRF-2, retinoblastoma binding proteins (RBP) 1 and 2, PLU-1, and yeast SWI1 [60,61,62,63]. None of the ARID genes have been reported to be implicated in differentiation and proliferation of SMCs. However, 6A3-5 and ARID nuclear proteins show similar high molecular weight (> 140 kd) and are differentially expressed in tissues [6,11,18,61]. Northern blot analysis showed substantial levels of 6A3-5 mRNA in brain, kidney, and testis. Moreover, western blot of 6A3-5 showed a unique band of a molecular weight of 175 kd, present in multiple rat tissues, albeit at substantially high levels in brain and testis. It is of interest to note that the possible role of 6A3-5 in the brain is supported by a 99% similarity with a rat cell line (PC12) sequence that differentiates into a neuronal phenotype following stimulation by NGF. Moreover, a human clone (KIAA1235), bearing an ARID nuclear domain, was isolated in the brain and showed also an important similarity (99%) to 6A3-5. Experimental data indicate that 6A3-5 may be a transcription factor implicated in the dedifferentiation and proliferation of SMCs. Indeed, the antibody directed against 6A3-5 confirmed, by FACScan, that 6A3-5 protein is not localized on the membrane but has a cytoplasmic or nuclear localization. Transcription factors are either permanently present in an inactive form in the nucleus, or translocated from the cytoplasm to the nucleus in response to a specific stimulus [64].
We have observed that at every stage when SMCs change phenotype, this affects the expression of 6A3-5. Our data suggest that this protein may be a potential factor involved in the processes of differentiation and proliferation of cells. First, the P9-V8 dedifferentiation model (synthetic versus proliferating cells) demonstrates that 6A3-5 is upregulated in the dedifferentiated V8 cells in comparison to P9 cells. Second, the P0-P9 differentiation model (contractile versus synthetic cells) demonstrates that 6A3-5 is upregulated in the dedifferentiated P9 cells in comparison to differentiated contractile P0 cells. These contractile quiescent cells (passage 0), in comparison to dedifferentiated SMCs (passage 9), show substantially lower mRNAs and protein levels of 6A3-5. Third, P9 synthetic cells stimulated by FCS or PMA after cell arrest (an in vitro model of cell proliferation) demonstrates that 6A3-5 is upregulated (1-2 hours after stimulation) in comparison to resting P9 cells. In fact, when dedifferentiated SMCs are synchronized in the quiescent G0 phase, 6A3-5 mRNA levels are significantly increased (within a period of 1-2 hours) following stimulation by PMA or FCS. Induction of SMC differentiation and proliferation by mitogenic agents results in a burst of 6A3-5 mRNA levels at a very early stage. Modulation of SMC phenotypes are known to induce the upregulation of a number of genes such as c-myc, c-myb, c-fos, p65 subunit of NF-kB, ras proteins, Osteopontin, mitogen-activated protein (MAP) kinases, angiotensin II, and cdk2 kinase [65,66,67,68,69,70,71,72,73]. Moreover, some new genes [74,75,76,77,78] were recently found to be upregulated in activated proliferating SMCs such as sgk (serum and glucocorticoidregulated kinase), type VIII collagen, nucleophosmin (a nuclear phosphoprotein implicated in the regulation of cell growth and protein synthesis), and Interferon inducible protein-10 (IP-10).
In conclusion, this work describes the structural and functional characterization of a new early gene. In essence, theses results, when taken together, suggest that the 6A3-5 gene may play a key role in genetic control of cellular differentiation and proliferation. The identification of 6A3-5 as a member of the emerging family of ARID proteins suggests that it might function as a coactivator or corepressor. Such activity may take place in combination with nuclear hormone receptors, as implied by the presence of the LXXLL motif. This takes place before activation complexes (including coactivators as p300 and CBP) are formed at specific promoter sites. Further work will be needed to delineate the role of this new gene in vascular lesions. Phenotypic modulation of SMCs from a contractile into a secretory and proliferate phenotypes is the result of changes in gene expression of multiple genes [79]. The 6A3-5 gene, identified in this study in SMCs, could conceivably be part of genes involved in modulating SMC phenotype. Carefully mapping the cascade of genes implicated in SMC migration and proliferation, in atherosclerosis and restenosis, may ultimately allow a better understanding of the SMC phenotypic modulation. It remains to be seen if the role of 6A3-5 in differentiation is limited to SMC or is implicated in other cellular or pathological models of differentiation.