Molecular cloning and characterization of SmLIM, a developmentally regulated LIM protein preferentially expressed in aortic smooth muscle cells.

Differentiated, quiescent vascular smooth muscle cells assume a dedifferentiated, proliferative phenotype in response to injury, one of the hallmarks of arteriosclerosis. Members of the LIM family of zinc-finger proteins are important in the differentiation of various cells including striated muscle. We describe here the molecular cloning and characterization of a developmentally regulated smooth muscle LIM protein, SmLIM, that is expressed preferentially in the rat aorta. This 194-amino acid protein has two LIM domains, and comparisons of rat SmLIM with its mouse and human homologues reveal high levels of amino acid sequence conservation (100 and 99%, respectively). SmLIM is a nuclear protein and maps to human chromosome 3. SmLIM mRNA expression was high in aorta but not in striated muscle and low in other smooth muscle tissues such as intestine and uterus. In contrast with arterial tissue, SmLIM mRNA was barely detectable in venous tissue. The presence of SmLIM expression within aortic smooth muscle cells was confirmed by in situ hybridization. In vitro, SmLIM mRNA levels decreased by 80% in response to platelet-derived growth factor-BB in rat aortic smooth muscle cells. In vivo, SmLIM mRNA decreased by 60% in response to vessel wall injury during periods of maximal smooth muscle cell proliferation. The down-regulation of SmLIM by phenotypic change in vascular smooth muscle cells suggests that it may be involved in their growth and differentiation.

In their normal state, vascular smooth muscle cells (VSMCs) 1 regulate vessel tone and blood pressure. VSMCs are not terminally differentiated, in contrast with skeletal muscle and cardiac muscle cells. In response to mechanical, chemical, or immunologic injury (1)(2)(3)(4)(5) the VSMC phenotype changes rapidly from that of a differentiated, quiescent cell to that of a dedifferentiated, proliferating cell. Although VSMC proliferation is a hallmark of arteriosclerosis, the leading cause of death in developed countries, little is known about the molecular mechanisms regulating this phenotypic change. Progress in this area has been limited by the lack of VSMC-specific markers and precursor cells that can be differentiated into VSMCs in vitro (6).
So far there are three classes of LIM proteins. Class 1 proteins (LIM-HD) contain two LIM domains and a homeodomain. Lin-11, Isl-1, and Mec-3 (17)(18)(19), the first LIM proteins to be identified, belong to this group. Class 2 proteins (LIM-only) contain one or more LIM domains, but lack the homeodomain (14). Class 3 proteins (LIM-K) contain LIM domains and a protein kinase domain (20 -22).
Two members of the LIM-only class, RBTN2 (12) and MLP (11), have been shown to play important roles in cellular differentiation. Originally identified in childhood T cell acute lymphoblastic leukemia (23,24), RBTN2 is essential for erythroid cell development; a homozygous null mutation in RBTN2 leads to failure of yolk sac erythropoiesis and embryonic death (12). MLP is expressed only in the heart and skeletal muscle of rats. Overexpression of sense MLP in C2 myoblasts potentiates myogenic cell differentiation. In contrast, expression of antisense MLP retards myoblast differentiation and withdrawal from the cell cycle. Although these observations suggest that MLP could be involved in regulating skeletal and heart muscle cell-specific gene expression (11), MLP mRNA is not expressed in VSMCs in vitro or in vivo (data not shown). We hypothesized that a related but heretofore unidentified LIM protein may play an analogous role in VSMC differentiation.
We report the identification and characterization of a LIM-* This work was supported in part by a grant from the Bristol-Myers Squibb Pharmaceutical Research Institute and by National Institutes of Health Grants RO1 GM53249 (to M.-E. L.), KO8 HL03274 (to N. E. S. S.), and KO8 HL03194 (to M. A. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U44948 and U46006.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-Aortic smooth muscle cells were harvested from the thoracic aorta of adult male Sprague-Dawley rats (200 -250 g) by enzymatic digestion according to the method of Gunther et al. (25). COS-7 and 10T1/2 cells were obtained from the American Type Cell Culture Collection. Embryonic stem cells (D3) were kindly provided by R. Rosenberg (Massachusetts Institute of Technology, Boston, MA). Rat aortic smooth muscle cells were grown in DME (JRH Biosciences, Lenexa, KS) supplemented with 10% fetal calf serum, penicillin (100 units/ml), streptomycin (100 g/ml), and 25 mM Hepes (pH 7.4) in a humidified incubator (37°C, 5% CO 2 ). COS-7 and 10T1/2 cells were grown similarly, with the exceptions that DME was supplemented with Serum Plus (Hyclone, Logan, Utah) for the former and basal medium Eagle (JRH Biosciences) was substituted for DME for the latter. Cells were cultured and maintained in an undifferentiated state with leukemia inhibitory factor as described by Doetschman et al. (26). PDGF-BB was purchased from Collaborative Biomedical Products (Bedford, MA).

Cloning and Sequencing of Rat (r)-SmLIM and Human (h)-SmLIM-
The full-length rat MLP cDNA was amplified from rat heart RNA by the reverse transcriptase PCR (27). Forward (5Ј-GAGTCTTCACCATGC-CGAAC-3Ј) and reverse (5Ј-CTCTCCCACCCCAAAAATAG-3Ј) primers, designed according to the published rat MLP sequence (11), were used to amplify an 801-base pair fragment. The PCR fragment was then subcloned and sequenced by the dideoxy chain termination method (27). The r-MLP fragment was used to screen a rat neonatal aortic cDNA library in gt11 (27). Approximately 1.6 million phage clones were plated, transferred to nitrocellulose paper, and screened at low stringency. One out of nine isolated clones encoded the partial sequence of a novel LIM protein, r-SmLIM. This partial clone was then used to screen a rat smooth muscle cDNA library in ZAP (Clontech) to obtain the full-length clone. The same partial rat clone was also used to screen a human aortic gt11 cDNA library to obtain the human sequence. The sequences of several partially overlapping clones were compiled to obtain the full-length h-SmLIM sequence. Both strands of the entire r-SmLIM and h-SmLIM cDNAs were sequenced at least once by the dideoxy chain termination method or on an automated DNA Sequencer (Licor, Lincoln, NE) according to the manufacturer's instructions.
Cellular Localization of r-SmLIM-To construct the expression plasmid Myc-SmLIM/pCR3, we added in frame a c-Myc peptide tag (EQKLI-SEED) to the r-SmLIM open reading frame at the N terminus by PCR techniques. This hybrid DNA fragment was then cloned into the eukaryotic expression vector pCR3 (Invitrogen). COS-7 and 10T1/2 cells were transiently transfected with the Myc-SmLIM/pCR3 plasmid by the DEAE-dextran method (27) with minor modifications (28). The transfected cells were grown on chamber slides and fixed with 4% paraformaldehyde in phosphate-buffered saline. Immunostaining was performed 48 h after transfection with an anti-c-Myc monoclonal antibody (9E10, Oncogene) followed by a rhodamine-conjugated goat antimouse IgG secondary antibody. Nuclear counterstaining was performed with Hoechst 33258 as recommended by the manufacturer.
Chromosomal Localization of h-SmLIM-We localized h-SmLIM with the BIOSMAP somatic cell hybrid blot (BIOS Labs, New Haven, CT), which contains DNA from 20 somatic cell hybrid cell lines plus three control DNAs (human, hamster, and mouse). A full-length h-SmLIM fragment was randomly primed (27) and hybridized as recommended by the manufacturer. This blot was washed according to the manufacturer's recommendations and then exposed to Kodak XAR film at Ϫ80°C.
RNA Extraction and RNA Blot Analysis-We isolated total RNA from cultured cells, rat organs, embryonic stem cells, and mouse embryos by guanidinium isothiocyanate extraction and centrifugation through cesium chloride (29). The mouse embryo samples (days 7-10) included placenta and yolk sac. Carotid artery total RNA was obtained by the RNA-Zol method (Cinna/Biotecx Laboratories International, Houston, TX) from adult male Sprague-Dawley rats that had been subjected to balloon injury (Zivic-Miller Co., Zelienople, PA). Human poly(A) ϩ RNA was purchased from Clontech Laboratories (Palo Alto, CA). All RNA was fractionated on a 1.3% formaldehyde-agarose gel and transferred to nitrocellulose filters. The filters were then hybridized with the appropriate 32 P-labeled, random primed cDNA probes as described elsewhere (27,30,31). The hybridized filters were washed in 30 mM sodium chloride, 3 mM sodium citrate, and 0.1% SDS at 55°C, and autoradiographed on Kodak XAR film at Ϫ80°C. To correct for differences in RNA loading, the blots were hybridized with an 18 or 28 S oligonucleotide probe. The filters were scanned, and radioactivity was measured on a PhosphorImager running the ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
In Vitro Transcription and Translation-The complete r-SmLIM open reading frame was cloned into the eukaryotic expression vector pCR3 (Invitrogen). In vitro transcription and translation was performed in the TNT-coupled wheat germ extract system (Promega, Madison, WI) according to the manufacturer's instructions. The transcribed and translated products were resolved on a 10% SDS-PAGE Tricine gel (32), and autoradiography was performed with Kodak BMR film at room temperature.
In Situ Hybridization-r-SmLIM mRNA was hybridized in situ as described elsewhere (33) with minor modifications. Adult male Sprague-Dawley rats were perfused with 4% paraformaldehyde. Organs were then postfixed with 4% paraformaldehyde, soaked in 30% sucrose until the tissue had sunk, embedded in optimum cutting temperature compound, and stored in isopentane at Ϫ80°C. Tissue sections were cut at a thickness of 5 m. SmLIM mRNA was detected by hybridization with a 35 S-UTP-labeled antisense cRNA probe synthesized with the SP6 RNA polymerase from HindIII-linearized r-SmLIM in Bluescript II SKϩ. For control experiments, a 35 S-UTP-labeled sense cRNA probe was synthesized under the same conditions. RNA probes were degraded to a length of approximately 100 -200 nucleotides by partial hydrolysis for 15 min at 60°C in 80 mM NaHCO 3 and 120 mM Na 2 CO 3 . After hybridization the tissue sections were washed under moderately stringent conditions as previously described (33). The dried tissue sections were then dipped into Kodak NTB2 emulsion (Eastman Kodak Co.) and exposed for 2-4 days at 4°C. Counterstaining was performed with hematoxylin-eosin. We then cloned the entire r-SmLIM cDNA into the PCRIII eukaryotic expression vector. In vitro transcription and translation (Promega) of this expression plasmid with wheat germ lysate revealed a protein product of 21 kDa (Fig. 1B).

Isolation and Characterization of r-SmLIM and h-
To determine whether SmLIM was conserved across species, we obtained the human ( Fig. 2A) and mouse (m) 2 homologues. A comparison of the h-SmLIM and r-SmLIM open reading frames revealed 93% identity at the cDNA level and 99% identity at the amino acid level (Fig. 2, A and B). Comparison of the open reading frames of m-SmLIM and r-SmLIM revealed 97% identity at the cDNA level and 100% identity at the amino acid level (Fig. 2B). A GenBank TM search indicated that SmLIM shares homology with the cysteine-rich protein (CRP) family (13,15,(35)(36)(37). Fig. 2A compares r-SmLIM and h-SmLIM with their rat and human CRP counterparts and rat MLP. Although an amino acid sequence comparison of r-SmLIM versus h-SmLIM shows 99% identity (Fig. 2B), a comparison of r-SmLIM with r-CRP shows 79% identity. These data indicate that SmLIM and CRP are related but different genes.
Cellular and Chromosomal Localization of SmLIM-The r-SmLIM deduced amino acid sequence contains the putative nuclear localization signal KKYGPK, suggesting that SmLIM may be a nuclear protein. To determine the cellular localization of SmLIM, we first generated a plasmid that would express a fusion protein of the c-Myc tag and r-SmLIM. This plasmid and the control vector alone were transfected into COS cells and immunostained with an anti-c-Myc antibody. Detection of the immunofluorescent signal in the nuclei of COS cells transfected with the c-Myc-r-SmLIM fusion plasmid but not the control vector alone localized the SmLIM protein to the nucleus (Fig.  3). We performed the same experiment in 10T1/2 fibroblasts and found that SmLIM localized to the nucleus in these cells as well (data not shown). We also mapped the chromosomal location of h-SmLIM with the BIOS somatic cell hybrid blot. h-SmLIM localized to chromosome 3 (Fig. 4, arrow).
Tissue Distribution of r-SmLIM and h-SmLIM (Northern Analysis)-Total RNA were isolated from 15 tissues of adult male and female rats and analyzed for SmLIM expression by Northern blot analysis (Fig. 5A). A single, intense, 1.0-kb band was detected in aorta. A much weaker signal was detected in kidney, thymus, and intestine. SmLIM expression was not detectable in heart and skeletal muscle and was barely detectable in brain, testis, esophagus, lung, liver, aortic adventitia, vena cava, and uterus. Thus, r-SmLIM appears to be expressed in tissue containing smooth rather than striated muscle. Furthermore, because expression of SmLIM was much greater in aorta than in intestine or uterus it would appear to be expressed preferentially in VSMCs. Even among vascular RNAs, r-SmLIM expression was more robust in arterial tissue (aorta) than in venous tissue (vena cava). Consistent with the r-SmLIM expression pattern, h-SmLIM was expressed to a high degree in aorta but not in heart or skeletal muscle (Fig. 5B).
Tissue Distribution of r-SmLIM (in Situ Hybridization)-For each antisense experiment with the r-SmLIM riboprobe (Fig. 6, left) a corresponding sense (control) experiment (Fig. 6, right) was performed to localize r-SmLIM expression within the vessel wall. Fig. 6, top left, shows intense staining of r-SmLIM in

FIG. 3. Cellular localization of r-SmLIM. COS cells were transiently transfected with the c-Myc-r-SmLIM hybrid construct or vector alone (not shown).
Protein expression was assayed 48 h after transfection with an anti-c-Myc monoclonal antibody (9E10) followed by rhodamine-conjugated secondary antibody (red, right). Nuclear counterstaining was performed with Hoechst 33258 (blue, left). Magnification, ϫ 600. both the aorta (Ao) and a small artery (Ar) nearby. Consistent with our Northern analysis, minimal expression of r-SmLIM was visible in the vena cava (V). A view of the aorta at higher magnification reveals that r-SmLIM expression was limited to smooth muscle cells in the medial layer (Fig. 6, bottom left). SmLIM signal expression was absent in skeletal muscle cells (data not shown). These observations agree with our Northern analyses and indicate that r-SmLIM was expressed preferentially in arterial smooth muscle cells.

Down-regulation of r-SmLIM Expression in VSMCs by Growth Factors and Arterial Wall Injury-PDGF-BB is unique
among the smooth muscle cell mitogens in its ability to selectively suppress in vitro the expression of differentiation markers such as ␣-actin, smooth muscle myosin heavy chain, and ␣-tropomyosin (6). Therefore we evaluated the effect of PDGF-BB on SmLIM expression in cultured VSMCs. r-SmLIM mRNA levels decreased gradually in response to PDGF-BB stimulation (Fig. 7A). A decrease in r-SmLIM expression appeared as early as 4 h after treatment, and a maximal decrease of 80% was obtained at 32 h after treatment.
In response to vessel wall injury, VSMCs undergo a phenotypic change from a differentiated, contractile state to a dedifferentiated, proliferative state. Balloon injury of the rat carotid artery is a well characterized model for studying this change in phenotype in vivo. Previous work on cellular proliferation after arterial injury showed that smooth muscle cell proliferation reaches a maximum in the medial layer at 48 h and a maximum in the intimal layer at 96 h and declines thereafter (38). We therefore studied r-SmLIM mRNA levels in rats at 2, 5, and 8 days after balloon injury of the carotid artery (Fig. 7B). SmLIM mRNA levels decreased by more than 60% after day 2 in comparison with control and remained at this level through day 8. These data suggest that r-SmLIM mRNA decreases in response to smooth muscle cell proliferation and dedifferentiation both in vitro and in vivo.
Developmental Regulation of r-SmLIM mRNA Expression-The data described so far suggest that SmLIM is expressed preferentially in vascular tissue and that its levels are affected by the differentiation state of VSMCs. To determine whether SmLIM expression is regulated during development, we isolated total RNA from undifferentiated embryonic stem cells and whole mouse embryos at days 7.5-16.5 post coitum (p.c.). We found that SmLIM expression was indeed regulated developmentally (Fig. 8). Expression was highest during the late primitive streak stage (day 7.5 p.c.), the point at which the embryonic and extraembryonic circulations begin to develop (39,40). SmLIM expression decreased rapidly at subsequent time points. By normalizing the data to the hybridization signal value at 7.5 days p.c., we found that relative mRNA expression decreased by 40% at 8.5 days p.c. and by approximately 80% at 9.5-16.5 days p.c. DISCUSSION We have isolated a developmentally regulated nuclear LIM protein, SmLIM, from a rat smooth muscle cell library. SmLIM is expressed preferentially in arterial smooth muscle cells, and in response to external cues that promote smooth muscle cell proliferation and dedifferentiation, SmLIM mRNA is down-regulated.
SmLIM is a highly conserved, two-LIM-domain nuclear protein of the LIM-only class (Figs. 1-3). Other members of this class include RBTN2, MLP, and CRP. Like SmLIM, RBTN2 and MLP are nuclear proteins with two LIM domains, and they are highly conserved across species (11,12,41). CRP proteins also have two LIM domains and show high cross-species conservation (37,42). Sequence comparisons of SmLIM and CRP suggest that the two gene families are related yet distinct (Fig.  2). In contrast with SmLIM, which is a nuclear protein (Fig. 3), CRP has been localized to the cytoskeletal adhesion plaques (13,15). Moreover, h-SmLIM localizes to chromosome 3 (Fig. 4), whereas h-CRP localizes to chromosome 1 (43). Finally, Northern analysis of r-CRP tissue distribution showed that the size of its mRNA and pattern of expression were distinct from those of r-SmLIM (data not shown). Taken together, these data indicate that SmLIM and CRP are distinct LIM proteins. While this manuscript was in preparation, Weiskirchen et al. (42) reported the cloning of the chicken CRP2 gene. Sequence comparisons suggest that CRP2 is the avian homologue of SmLIM.
Although SmLIM is highly expressed in smooth muscle cells, it is not expressed in striated muscle cells (Fig. 5). This pattern is in contrast with that of MLP, which is expressed only in the heart and skeletal muscle (11). When a full-length MLP probe was hybridized to total RNA from aorta and cultured VSMCs, we were unable to detect a message (data not shown). Thus, the expression of the two LIM proteins is distinct within the myogenic cell lineage. Arber et al. (11) have shown that MLP may play an essential role in striated muscle differentiation. Perhaps SmLIM plays an analogous role in VSMCs.
SmLIM mRNA is expressed preferentially in tissue containing vascular smooth as opposed to nonvascular smooth muscle cells (Fig. 5). As such it joins two other recently identified genes, SM22␣ and gax (44,45), expressed highly in VSMC. However, some differences exist in their patterns of expression in tissue. For example, in addition to aorta, SM22␣ is highly expressed in uterus and intestine (45), whereas SmLIM is not.
Gax expression is not detected in intestine but is detected to a high degree in heart (44). By comparison, SmLIM expression appears to be more restricted to aorta. Furthermore, SmLIM is expressed preferentially in arterial as opposed to venous tissue. Arteries and veins have been shown to respond differently to injury (46) and various pharmacological manipulations (47)(48)(49); these observations suggest that smooth muscle cells may be fundamentally different in the two tissue types. To our knowledge the pattern of preferential expression in arterial but Rat aortic smooth muscle cells were made quiescent by incubation in low serum medium (DME plus 0.4% calf serum) for 48 h. Cells were then treated for the indicated times with PDGF-BB (20 ng/ml). Northern analysis was performed with 10 g of total RNA/ lane. After electrophoresis, RNA was transferred to nitrocellulose filters and hybridized with a 32 P-labeled r-SmLIM probe. A single r-SmLIM transcript is visible at 1.0 kb. Filters were hybridized with 18 S to verify equivalent loading. B, decrease in r-SmLIM mRNA expression after balloon injury in rat carotid arteries. Northern analysis was performed with 20 g of total RNA/lane at 2, 5, and 8 days after injury. A single r-SmLIM transcript is visible at 1.0 kb. Filters were hybridized with 18 S to verify equivalent loading.
FIG. 8. Developmental regulation of SmLIM mRNA expression. Total RNA isolated from undifferentiated embryonic stem cells (ES) and mouse embryos days 7.5-16.5 p.c. Northern analysis was performed with 10 g of total RNA/lane. After electrophoresis, RNA was transferred to nitrocellulose filters and hybridized with a 32 P-labeled r-SmLIM probe. A single r-SmLIM transcript is visible at 1.0 kb. Filters were hybridized with 28 S to verify equivalent loading. not venous smooth muscle cells is unique to SmLIM.
Smooth muscle cells differ from striated muscle cells in their ability to reenter the cell cycle. This reentry is accompanied by a change from a quiescent, differentiated phenotype into a proliferative, dedifferentiated phenotype (3,50). Genes important for maintaining the differentiated state may require down-regulation or inactivation to permit this phenotypic modulation. In this study we evaluated r-SmLIM expression in response to two different systems that model VSMC dedifferentiation. First, we found that r-SmLIM expression was downregulated in response to PDGF-BB stimulation (Fig. 7). Second, we found an analogous decrease in SmLIM mRNA expression after balloon injury to the rat carotid artery, with a brisk down-regulation at 2 days after injury. In both aspects SmLIM is similar to the growth arrest-specific homeobox gene gax (44,51). During this 2-8-day period after injury, smooth muscle cells dedifferentiate and assume a highly proliferative phenotype (3). Thus, both in vitro and in vivo, SmLIM expression is down-regulated as smooth muscle cells undergo phenotypic change.
SmLIM expression also appears to be regulated developmentally. Expression is highest at day 7.5 p.c. in mouse embryos (Fig. 8) and plateaus by day 9.5 p.c. These early stages represent important points in the development of the mouse heart and vascular systems. At the late primitive streak stage (day 7.5 p.c.), discrete blood islands make their first appearance and amalgamate shortly thereafter to form the yolk sac vasculature. Within the embryo one also sees the early formation of a vasculature at 8.0 days p.c. and amalgamation of the embryonic and extraembryonic circulations at 8.5 days p.c. (39,40). Given that SmLIM expression is highest in the adult aorta and correlates with the level of smooth muscle cell differentiation, it is interesting that its embryonic expression is highest during periods critical for vascular development.
The LIM domain functions as a modular protein-binding interface (52). For example, the LIM-only protein RBTN2 binds to the basic helix-loop-helix protein tal-1 (53), an interaction thought to be critical in regulating red blood cell development. Homozygous deletion of either RBTN2 or tal-1 results in absence of red blood cell formation (12,54). Similarly, it has been suggested that the effect of the LIM-only protein MLP on myoblast differentiation may be as a cofactor regulating muscle-specific gene expression. Identification of the interaction partner(s) of SmLIM may yield important information about other factors involved in smooth muscle cell development and differentiation.  We suggest that subscribers photocopy these corrections and insert the photocopies at the appropriate places where the article to be corrected originally appeared. Authors are urged to introduce these corrections into any reprints they distribute. Secondary (abstract) services are urged to carry notice of these corrections as prominently as they carried the original abstracts.