A Critical Role for Calponin 2 in Vascular Development*

Calponin 2 (h2 calponin, CNN2) is an actin-binding protein implicated in cytoskeletal organization. We have found that the expression of calponin 2 is relatively restricted to vasculature from 16 to 30 h post-fertilization during zebrafish (Danio rerio) development. Forty-eight hours after injecting antisense morpholino oligos against calponin 2 into embryos at the 1-4-cell stage, zebrafish demonstrated various cardiovascular defects, including sluggish axial and head circulation, absence of circulation in intersegmental vessels and in the dorsal longitudinal anastomotic vessel, enlarged cerebral ventricles, and pericardial edema, in addition to an excess bending, spiraling tail and twisting of the caudal fin. Knockdown of calponin 2 in the Tg(fli1:EGFP)y1 zebrafish line (in which a fli1 promoter drives vascular-specific enhanced green fluorescent protein expression) indicated that diminished calponin 2 expression blocked the proper migration of endothelial cells during formation of intersegmental vessels. In vitro studies showed that basic fibroblast growth factor-induced human umbilical vein endothelial cell migration was down-regulated by knockdown of calponin 2 expression using an antisense adenovirus, and overexpression of calponin 2 enhanced migration and hastened wound healing. These events were correlated with activation of mitogen-activated protein kinase; moreover, inhibition of this pathway blocked the promigratory effect of calponin 2. Collectively, these data suggest that calponin 2 plays an important role in the migration of endothelial cells both in vivo and in vitro and that its expression is critical for proper vascular development.

Calponin 2 is a member of the calponin family, which includes two other distinct members, basic calponin 1 (h1 calponin, CNN1) (1) and acidic calponin 3 (h3 calponin, CNN3) (2). These calponins share high sequence homology at the amino-terminal two-thirds of the molecule, which contains the calponin homology (CH) 3 domain followed by three calponin repeats (3,4). The carboxyl-terminal sequences beyond Cys-273 are unique for the three variants, accounting for differences in their isoelectric points.
Calponin 1 is specific to smooth muscle cells and serves as a marker for smooth muscle cell differentiation. It inhibits the actin-activated ATPase activity of myosin (5) and may play a role in regulating smooth muscle contraction. As an actin-, tropomyosin-and calmodulin-binding protein, it interacts in vitro with cytoskeletal components, such as myosin (6,7), caldesmon (8,9), desmin (10), and tubulin (11). Calponin 1 has been shown to play an important role in agonist-induced signal transduction in smooth muscle cells (12). It binds to extracellular regulated Ser/Thr kinases (ERK1 and ERK2) (13,14) through the aminoterminal CH domain and protein kinase C (PKC) (15).
Less is known about the function of calponin 2. Although the sequence conservation between calponin 2 and calponin 1 points to a similar function, they have diverged during vertebrate evolution, and the sequence diversity, especially in the carboxyl-terminal tail, suggests that they may have different functions. Calponin 2 has been shown to play a role in cytoskeletal organization (16). Overexpression of calponin 2 in a cell line of smooth muscle origin (SM3) lacking calponin 2 inhibits cell proliferation (17), likely mediated via the actin cytoskeleton. In a subcellular localization study, a green fluorescent calponin 2 fusion protein is shown to localize to the ends of stress fibers and in the protrusions of spreading cells (18).
The zebrafish is a valuable model system for studying vascular development, a complex and tightly regulated process. Blood vessel formation and expression patterns of several genes of importance in zebrafish embryonic vessel formation are similar to those in other vertebrates (19 -21). This conservation points toward similar vasculogenic and angiogenic signaling pathways (22). Importantly, zebrafish embryos can survive without a functioning circulatory system during the first three days of development (23), during which organogenesis is occurring. Thus vascular effects can be studied in a relatively pristine fashion. Moreover, fluorescent imaging methods, such as confocal microangiography, have been used to image vascular patterning in the fish. Transgenic zebrafish with vessel-specific expression of enhanced green fluorescent protein (EGFP) under the control of the fli1 promoter makes it possible to further visualize fine cellular features of vascular endothelial cells (ECs) (24). Finally, gene-specific "knockdown" through the use of antisense morpholino oligonucleotides (25) provides a rapid way to ascertain in vivo function of a gene. Our strategy to identify novel genes relevant to vascular development was as follows. A whole mount in situ hybridization screen was performed using an adult zebrafish kidney cDNA library, and it identified 44 transcripts from 4000 with vessel-enriched expression during early zebrafish development. One of these transcripts (GA3396) encoded a possible calponin 2 orthologue and was found to be primarily expressed in vessels at 24 h post-fertilization (hpf). Morpholino oligonucleotides (MOs) specifically targeting zebrafish calponin 2 were then used to knock down calponin 2 expression. The observed vasculature defects were studied in detail using a combination of several techniques. Signaling pathways were further defined using human umbilical vein endothelial cells (HUVECs) as a mammalian model in vitro.

MATERIALS AND METHODS
Isolation of Zebrafish cDNA-GA3396 (calponin 2) was identified by an in situ hybridization screen (26) and was isolated from an adult zebrafish kidney cDNA library (Dr. J. Rast, Children's Hospital, St. Petersburg, FL).
Animals-Zebrafish were maintained at 28.5°C as described previously (27). Developmental stages were determined by embryo morphology and hpf (28). For in situ hybridization, embryos were incubated in 0.003% 1-phenyl-2-thiourea after 24 hpf to prevent pigment formation.
Whole Mount RNA in Situ Hybridization-Whole mount RNA in situ hybridization was carried out as described previously (29,30). Fulllength antisense and sense probes for zebrafish calponin 2 were synthesized. For histological analysis, specimens were fixed in 4% paraformaldehyde overnight at 4°C, dehydrated with ethanol, and embedded in JB-4 resin (Polysciences, Inc., Warrington, PA). Serial sections were examined with a Nikon Optiphot 2 Microscope.
MOs were initially dissolved in sterile water to a concentration of 4 mM. After dilution with injection buffer (100 mM NaCl, 50 mM KCl, 0.5 mM Hepes, pH 7.6) to the desired concentration, 1 nl of diluted MO solution was injected into the 1-4-cell embryos mounted in agarose slots.
Digital Image Analysis-Similar digital image analysis was done as described by Lee et al. (31). Briefly, ϳ100 sequential phase contrast images of the embryos were taken using 100ϫ magnification on the microscope at ϳ1 frame/s. Thereafter, a z projection of the standard deviation of the whole image sequence was constructed using ImageJ (47). The standard deviation measured the variance of the gray value from the mean at a given position caused by moving objects, in effect depicting the flow of erythrocytes in the embryos.
Microangiography-Microangiography was performed as described previously (32), except that 0.02-m Fluorospheres with red fluorescence (Molecular Probes) in 1% bovine serum albumin were used during the injection.
RT-PCR and Sequencing-After injection of the MOs, embryos were collected at 48 hpf and stored in RNAlater (Ambion) at Ϫ80°C before extracting the total RNA with RNAqueous-4PCR kit (Ambion). Two g of total RNAs were reverse-transcribed into cDNA using SUPER-SCRIPT II (Invitrogen) and amplified with the Advantage-GC 2 PCR Kit (BD Biosciences). The PCR products were then cloned into pCRII-TOPO vector using the TOPO cloning kit (Invitrogen). Eight clones were randomly picked and sequenced.
Plasmid Construction-Full-length calponin 2 cDNA was PCR (Advantage-GC cDNA PCR kit from Clontech)-amplified from HUVEC RT product using forward primer 5Ј-ACCATGGGTAAGCC-TATCCCTAACCCTCTCCTCGGTCTCGATTCTACGAGCTCCA-CGCAGTTCAACAAGG-3Ј (containing the V5 tag) and reverse primer 5Ј-CAGATGGGAAGACGATGTGGGGAGAG-3Ј and TAcloned into the pCRII vector (Invitrogen). The insert was excised with EcoRI and ligated into the pCS2ϩ expression vector and confirmed by full-length sequencing.
Adenoviruses were amplified using QBI-HEK293 cells (Q Biogene, Inc., Irvine, CA). Amplification, CsCl purification and virus storage were done according to the protocol (48), and viral titers were determined using Adeno-X TM rapid titer kit (BD Biosciences). HUVECs (Cascade Biologics, Inc., Portland OR, passage number Ͻ5) cultured in 6-well tissue culture plates to 60% confluency were infected with different concentrations of adenoviruses for 6 h, at which point fresh EGM TM -2 MV medium (Cambrex) was added and cells incubated for 24 -48 h before further experiments. For the MEK-DD (constitutively active MEK construct, a gift from Dr. Eileen O'Leary, Harvard Institute of Medicine) rescue experiment, HUVECs were infected with empty or calponin 2 antisense adenoviruses for 6 h, changed to EGM TM -2 MV medium for 20 h, and then infected with MEK-DD adenovirus at a multiplicity of infection of 350 in EBM TM -2 (Cambrex) with 0.5% serum for 2 days.
Migration Assay-Migration assay was done as previously described (34).
Wound-healing Assay-HUVECs infected with adenoviruses were grown on 6-well tissue culture plates, and wounds of defined size were made into the confluent cell layer. Phase contrast micrographs were used to assess wound closure. Quantification was done by dividing the cell number within the wounds by the area of the wounds and then normalizing to the control for each experiment.
Luciferase Assay-Plasmids were transiently transfected into HUVECs using SuperFect transfection reagent (Qiagen). The cells were incubated for 20 h after transfection in EGM-2 MV medium, and cell lysates were collected with passive lysis buffer. The Dual Luciferase Reporter Assay System (Promega) was used to measure luciferase activities, determined using a luminometer normalized with sea pansy luciferase activity under the control of the thymidine kinase promoter.

RESULTS
Cloning of Calponin 2 from Zebrafish-The zebrafish GA3396 cDNA consisted of 2213 nucleotides after it was fully sequenced (GenBank TM accession number DQ104245). The predicted protein contained 307 amino acids with ϳ78% overall amino acid identities to pig, mouse, Xenopus, and human calponin 2 ( Fig. 1, A and B). A phylogenetic tree indicates that zebrafish GA3396 evolved from the same ancestor as other calponin 2 genes (Fig. 1C). After the conserved Cys-273, the carboxyl-terminal tail of this zebrafish protein was composed of 35 amino acids and had a pI of 2.97, closer to human calponin 2 (35 amino acids and pI of 2.56) than to human calponin 1 (24 amino acids and pI of 5.60). Lys-151 is conserved in mammalian calponin 1 but is not present in human calponin 2 nor is it present in our zebrafish sequence (Fig. 1A). Ser-175 and Thr-184, the PKC and Ca 2ϩ /CaM kinase II phosphorylation sites in human calponin 1, which are conserved among all cal-ponins, are also present in our zebrafish protein at residues 176 and 185. Together, these data suggest that the GA3396 clone is zebrafish calponin 2.
After blast searching in the latest zebrafish whole genome shotgun assembly Zv5 (Sanger Institute), the zebrafish calponin 2 cDNA was found to correspond to a gene region on chromosome 2. This gene has 7 exons spanning over 6.1 kb of genomic DNA, whereas the 7 exons of the human calponin 2 distribute over 12.2 kb of genomic DNA.
Expression Pattern of Calponin 2 in Early Zebrafish Development-Because clone GA3396 had been selected on the basis of its vascular expression in a whole mount in situ screen conducted at 9 hpf to 5 days postfertilization. We proceeded to first analyze its expression in detail during early zebrafish development. No specific expression of calponin 2 could be detected until 16 hpf (data not shown). At 16 hpf, expression of calponin 2 was more restricted to the posterior midline in presumed angioblast cells known to form axial vessels (Fig. 2, A and B). From 24 to 30 hpf, expression was observed throughout the vasculature, including the aorta, the caudal and posterior cardinal vein, blood vessels in the head (middle cerebral blood vessel, eye blood vessel) ( Fig. 2; C, D, G, and H), the duct of Cuvier (Fig. 2I, arrowhead), and intersegmental vessels (ISVs) (Fig. 2L). Staining was strong in the aortic wall, but the hypochord was not labeled (Fig. 2K). Labeling was also observed in the epidermis (mainly in the median fin fold), gut, dorsal spinal cord (Fig. 2, C and G), forebrain ventricular zone (Fig. 2I), and at rhombomere boundaries (Fig. 2J).
Knockdown of Calponin 2 Expression by MOs Leads to Vascular Defects-MOs were designed to knock down calponin 2 expression. The translation blocker (MOATG) was found to be efficacious in an in vitro transcription and translation assay (data not shown). The second MO (MOs13589) targeted the third exon-intron junction. The advantage of using MOs13589 for the knockdown of calponin 2 expression was that we could confirm its effectiveness by RT-PCR. Subsequent sequencing of the PCR product derived from RNA collected from the injected fish showed that exon 2 was directly linked to exon 4 after blockage of the third donor splice site by MOs13589, resulting in a smaller mRNA, which had two stop codons at 184 and 193 nucleotides downstream of the ATG (Fig. 3, A-C). The predicted protein would comprise the first 60 of its 307 amino acids and lack most of its known functional domains.
The phenotypes caused by different doses of MOATG, MOs13589, and Mis13589 were carefully evaluated and compared at 48 hpf (Table  1). Although 0.2 mM Mis13589 had no effect on the development of embryos, both MOATG-and MOs13589-injected embryos showed dramatic and identical changes in vascular development at 48 hpf. MOs13589 was approximately twice as potent as MOATG. MOs13589 at 0.2 mM gave very similar phenotypes to those obtained by 0.4 mM MOATG. Because 0.2 mM MOs13589 resulted in the absence of intersegmental circulation in almost every embryo without causing dramatic changes in overall morphology and massive death of embryos, this concentration was used in all of the subsequent experiments.
After injecting 0.2 mM MOs13589, no obvious morphological alterations in MO-injected fish in comparison with wild type embryos could be observed prior to 18 hpf. At ϳ27 hpf, circulating erythrocytes were present in control embryos but not in most of the morphants. In some of the morphants, pools of red blood cells could be found in the caudal vein. At 48 hpf, most of the morphants developed hindbrain ventricular edema and pericardial edema in addition to bending and spiraling of the tail and twisting of the caudal fin (Fig. 3, D and E). A digital motion analysis in control embryos showed vigorously circulating erythrocytes throughout the embryos, including axial vessels, ISVs, and head vessels. While in the morphants, we found a significantly reduced number of erythrocytes circulating sluggishly in the trunk vessels, no intersegmental circulation of blood cells, and weak and slowly pumping hearts (Fig.  3, F and G; Table 1).
Knockdown of Calponin 2 Prevents the Proper Migration of ECs to Form the Dorsal Portion of ISVs-These results suggested that knockdown of calponin 2 could lead to multiple vascular defects in the zebrafish. We therefore carried out the knockdown experiments in the Tg(fli1:EGFP) y1 line, in which robust expression of EGFP in virtually all ECs and their angioblast precursors (24) would provide us a particularly powerful tool for dissecting the vascular anomalies noted.  MARCH 10, 2006 • VOLUME 281 • NUMBER 10

JOURNAL OF BIOLOGICAL CHEMISTRY 6667
Vascular development was closely monitored with fluorescent microscopy after injection of MOs13589. Early vasculogenesis, the de novo formation of major trunk vessels by co-migration and coalescence of angioblast progenitor cells originating in the trunk lateral mesoderm (35), was largely unaffected in the morphants in comparison with that in wild types. Well formed axial vessels could be found in both controls and morphants at 22 hpf (Fig. 4, A and B), consistent with our fli1 in situ hybridization staining results (data not shown).
Starting from 20 hpf, the ECs started to sprout and migrate to form the ISVs in control embryos, a phenomenon not observed in morphants until 22 hpf (Fig. 4, A and B). Strikingly, although all of the ECs in the control embryos migrated dorsally across the midline and further dorsally at ϳ23 hpf, most of the ECs in the morphants did not migrate further after they reached the midline at ϳ25 hpf. This resulted in a partial primary ISV network in the morphants (Fig. 4D) in contrast to the delicate EC network in control embryos (Fig. 4C) at 30 hpf. Even at 37 hpf, we could see bright green dots, the cell bodies of ECs in ISVs, still lined up at the embryo midline throughout the length of the fish in almost every morphant (Fig. 4, F and G). Thus, this difference between normal and morphant fish was unlikely to merely represent a delay in an otherwise normal developmental program. At 52 hpf, a small percentage of these ECs remained exactly as they were at 36 hpf; some of the ECs branched horizontally at the midline and connected to each other. Most of the ECs did sprout dorsally during this time period but followed crooked paths to form fragmentary dorsal longitudinal anastomotic vessels (Fig. 4, H and I). These ISVs and dorsal longitudinal anastomotic vessels were not functional for blood flow, as shown previously (Fig. 3, F  and G), and were further confirmed by microangiography with 0.02-m Fluorospheres (Fig. 4, J and K).
The relatively high expression of calponin 2 in ECs and its important role in the proper formation of ISVs during zebrafish development prompted us to further study the function of calponin 2 in primary ECs in vitro and to probe signaling events. We focused on a study of endothelial cell migration, as this appeared to be the defect noted in vivo.

EC Migration Is Augmented by Overexpressing Calponin 2 and Inhibited by Calponin 2 Antisense in Vitro-
We generated adenoviral constructs of full-length calponin 2 and its antisense counterpart, and protein levels were checked 30 h after infection in HUVECs (Fig. 5A). The expression level of calponin 2 was ϳ25-fold higher compared with that of endogenous calponin 2 in our empty virus control. The calponin 2 antisense virus knocked down endogenous calponin 2 expression in HUVECs to undetectable levels, whereas the expression of calponin 1 was unaffected. To investigate the functional effects of modulating calponin 2 levels, HUVEC migration assay was done, which showed increased cell migration by calponin 2 (Ϸ7-fold) compared with base line in the empty virus control or in the antisense virus-infected cells (Fig. 5B). Because migration in vivo is often stimulated by gradients of a growth factor, we redid our experiments under stimulation of basic fibroblast growth factor (bFGF), a potent and well studied promigratory agent for ECs. In the migration assay with 20 ng/ml bFGF (Fig.  5C), calponin 2 augmented HUVEC migration by only 50%, suggesting that calponin 2-induced downstream signaling might overlap that induced by bFGF. The antisense adenovirus decreased HUVEC migration by ϳ40% compared with that obtained with empty virus, suggesting that calponin 2 was a component of the downstream events elicited by bFGF.
To further study the effect of calponin 2 on migratory responses, we used a wound-healing assay to compare the motility of the HUVECs infected with our different viruses. HUVECs infected with calponin 2 sense, antisense, and empty virus control were grown to confluence on 6-well plates, and wounds were made in the cell monolayer. EBM-2 with 0.5% serum was used in the experiments. Phase contrast pictures, taken at 0 and 16 h post-injury, revealed faster wound closure for the cells expressing calponin 2 sense ( Fig. 6; A, B, and E), whereas cells infected with calponin 2 antisense virus behaved similarly to empty virus-infected control cells, data consistent with our migration assay result. In wound-healing experiments done in the presence of bFGF, calponin 2 sense virus promoted wound closing and antisense virus delayed it (Fig.

TABLE I Frequencies (%) of calponin 2 morphant circulation defects at 48 hpf after injection of different doses of MOs
Frequency was calculated as the number of embryos with a specified phenotype divided by total live embryos. 6, C-E), data consistent with those observed for bFGF-stimulated migration.

Calponin 2 Increases ERK Phosphorylation and Functions Upstream of MEK-Because MAPK signaling affects cell migration in ECs and
other cell lines and bFGF is known to use this pathway in promoting cell migration, we sought to determine the effect of calponin 2 on MAPK signaling. Moreover, a previous report (13) suggests that ERK can directly bind to the CH domain in calponin 1 (Fig. 1A), a region of high sequence homology in all three calponins. SRE-luciferase assay was done using pCS2ϩ calponin 2 construct. Luciferase activity was increased by the calponin 2 construct in a dose-dependent fashion (Fig.  7A), suggesting that calponin 2 is an activator of MAPK. Next, we checked ERK1/2 phosphorylation using our adenovirus-infected cells (Fig. 7B). The expression of calponin 2 increased ERK1/2 phosphorylation by 1.5-fold, and its antisense decreased ERK1/2 phosphorylation by 2-fold. To check whether calponin 2 influenced bFGF-stimulated ERK signaling, we infected HUVECs with empty, calponin sense and antisense adenoviruses. ERK phosphorylation was up-regulated by calponin  Cell lysates were collected from HUVECs infected with empty, calponin 2 sense, or antisense adenoviruses. Western blot was done with calponin 2 antibody CP21, calponin 1 antibody CP3, and glyceraldehyde-3-phosphate dehydrogenase antibody, respectively. B, migration assay showing the effect of calponin 2 on HUVEC migration. HUVECs infected with empty, calponin 2 sense, or antisense adenoviruses were added in the upper chamber, and Dulbecco's modified Eagle's medium with 0.5% FBS was added in the lower chamber. Migrated cells were stained and counted 16 h later. C, the effect of calponin 2 on HUVEC migration stimulated by bFGF. HUVECs infected with the same adenoviruses as above were used, and bFGF (20 ng/ml) was added in the Dulbecco's modified Eagle's medium plus 0.5% FBS in the lower chamber. **, p Ͻ 0.01 compared with control as assessed by Student's t test; the difference between control and antisense was not significant. overexpression by 2-fold and down-regulated by antisense by 2-fold 10 min after bFGF stimulation (Fig. 7C). To dissect where in the MAPK pathway calponin 2 might act, we assessed ERK1/2 phosphorylation in cells overexpressing calponin 2 with or without U0126 (10 M), a MEK inhibitor of the MEK-ERK pathway. U0126 reduced the augmentation of the ERK1/2 phosphorylation in calponin 2 overexpressing HUVECs to the base line (Fig. 7D), indicating that calponin 2 increased the phosphorylation of ERK through action(s) upstream of MEK or at MEK.
Calponin 2 Influences HUVEC Migration through ERK Signaling-Finally, to tie signaling data to function, we conducted the HUVEC migration assay in the presence or absence of the MEK inhibitor U0126 (Fig.  8A). Calponin 2 expression increased HUVEC migration, which was down-regulated dose-dependently by U0126, suggesting that calponin 2 up-regulated HUVEC migration through activation of ERK signaling. To further confirm these data, MEK-DD (a constitutively active form of MEK) was used to attempt to rescue the effect of calponin 2 antisense on bFGF-induced EC migration (Fig. 8B). Our data showed that MEK-DD could partially rescue the effect of calponin 2 antisense, indicating that calponin 2 worked through MEK. MEK-DD alone and MEK-DD plus bFGF increased EC migration compared with their corresponding controls, which is consistent with the notion that MEK-ERK are important players in EC migration. It is worth noting that calponin 2 antisense did not have any effect on the ability of MEK-DD to enhance HUVEC migration (Fig. 8C), demonstrating also that calponin 2 acts upstream of MEK. Our model is summarized in Fig. 9.

DISCUSSION
Our studies in zebrafish and mammalian cell culture point to a critical role for calponin 2 in vessel development. We began by studying the zebrafish developmental expression of this gene, which emerged as part of a whole mount in situ screen for vessel-restricted cDNAs. Previous information on mammalian calponin 2 has pointed to a more widespread expression in numerous organs in the adult; it is detectable in non-smooth muscle organs such as the heart (36), as well as in smooth muscle cells (17) and non-muscle cells, such as fibroblasts (17) and endothelial cells (37). However, the largely vascular restricted expression of calponin 2 that we saw in early zebrafish development was likely due to its high expression level in ECs, because we could not find evidence for smooth muscle-like cells by electron microscopy near the axial vessels of zebrafish embryos from days 1 to 5. 4 In addition, one report showed that calponin 2 was expressed in porcine aortic ECs by RT-PCR (37). Moreover, by Western blotting, calponin 2 was expressed in HUVECs, human microvascular ECs, calf pulmonary artery ECs, and bovine aortic ECs, all at levels comparable with those in cultured human smooth muscle cells (data not shown). Collectively, these data suggest that calponin 2 expression in endothelial cells may be of functional importance, and we sought to explore this hypothesis by studying early zebrafish vessel development.
The MOs directed against calponin 2 caused striking cardiovascular defects. The effects of the two MOs phenocopied each other, providing evidence that the defects seen were specific to knockdown of calponin 2 expression. Moreover, the EC-restricted expression of calponin 2 in the fish strongly suggested that the effects noted were not secondary to those arising in other organs. An exception to this might be the heart, given the expression of calponin 2 in mice embryonic heart (36). Therefore, we cannot rule out the possibility that the impaired cardiac function noted in the morphants (slow heart pumping) contributed to some 4 I. Drummond and V. P. Sukhatme, unpublished data. Western blot was done with pERK1/2 and total ERK polyclonal antibody, respectively. C, the cells were treated as described in B, except that bFGF (10 ng/ml) was added in the medium for 10 min before collecting cell lysates. D, in the presence of calponin 2, the MAPK phosphorylation level was determined with or without the MEK inhibitor U0126 (10 M). FIGURE 6. Wound-healing assays using HUVECs. A and B, HUVECs infected with empty, sense, or antisense adenoviruses were grown to confluence. Smaller wounds were made in the cell lawn using 200-l pipette tips, and EBM-2 with 0.5% FBS was replaced. Pictures were taken at 0 and 16 h. C and D, HUVECs infected with the same adenoviruses were grown to confluence. Bigger wounds were made using 1-ml pipette tips after which EBM-2 with 0.5% FBS and 20 ng/ml bFGF was replaced. Pictures were taken at 0 and 60 h post-injury. E, quantification of wound-healing assays. *, p Ͻ 0.05 compared with control as assessed by Student's t test; #, p Ͼ 0.05 compared with control as assessed by Student's t test. Data are presented as mean Ϯ S.E. The microphotographs are the representative pictures of three independent experiments.
of the phenotypes, such as the sluggish blood flow we observed. However, because the sprouting of ECs to form ISVs normally occurs at 20 hpf, when circulation is not established, it is unlikely that cardiac dysfunction caused the sprouting defect noted in ECs, namely their inability to form the dorsal portion of the ISVs. In fact, each of the ISVs is composed of only three linked ECs, whose cell bodies are located approximately 1) at the dorsal longitudinal anastomotic vessel-primary segment junction (EC D ), 2) at the level of the parachordal vessels (EC P ), and 3) at the dorsal aorta-primary segment junction (EC A ) (38,39) (Fig.  4E). The formation of these ISVs in zebrafish comprises two steps, both requiring the sprouting and migration of endothelial cells. The first step, which occurs from 16.5 to 36 hpf, requires the sprouting and migration of endothelial cells from the dorsal aorta to form a primary network of vascular segments, whereas the secondary sprouting of endothelial cells from the posterior cardinal vein interfaces with this primary network to form the ISV network (38). The EC D was absent in every single primary segment in most of the morphants we examined at 36 hpf, whereas the EC p and EC A were present. This appears to be a unique phenotype not previously noted in the zebrafish vascular biology literature. We did see loss of EC P and EC A cells when we increased the dose of the MOs13589 against calponin 2, but because the morphology of the whole embryo was also severely affected, we were not sure that it was a direct effect of diminished calponin 2 expression in the ECs. It is believed that all three ECs composing the primary segment are derived from the dorsal aorta through migration and not proliferation (39). Thus, the absence of EC D in the calponin 2 morphants most likely resulted from failure of migration of the ECs destined to become EC D , which is the first in vivo evidence that calponin 2 may be critical in EC migration during early vessel development.
Importantly, the impaired migration of endothelial cells in zebrafish is consistent with the results from our in vitro experiments, in which knockdown of calponin 2 expression in HUVECs by antisense adenovirus inhibited migration and wound closing in response to bFGF, whereas overexpression of calponin 2 promoted migration in both assays (Fig. 5, B and C). However, Danninger and Gimona (18) show that green fluorescent calponin 2 fusion protein did not change the motility of NIH3T3 cells in wound assay, perhaps because 3T3 cells express high levels of endogenous calponin 2 (40). These data could also result from a different cell context. Collectively, our results suggest an important and previously unrecognized role for calponin 2 in the migration of endothelial cells in vivo and in vitro.
Calponin 1 has been found to play a role in signal transduction. Its CH domain is shared among cytoskeletal and signaling molecules, and this CH domain may help localize signaling molecules to the actin cytoskeleton (4). Calponin 1 has been shown to bind ERK1/2 directly (13,14), although ERK does not phosphorylate calponin 1 in vitro. Thus, previous studies have asked whether MAPKs could influence calponin 1, because MAPKs are known to participate in the phosphorylation of cytoskeletal proteins and cause cytoskeletal rearrangements. The reverse question, whether calponin 1 can affect the MAPK pathway, has not been addressed, and nothing is known about calponin 2-induced signaling. We have found that calponin 2 can up-regulate MAPK signaling via increased phosphorylation of ERK1/2 and increased activity of SRE-luciferase target; moreover, calponin 2 antisense reduced the phosphorylation of ERK1/2. A similar result was obtained in the presence of bFGF (10 ng/ml) stimulation, demonstrating that calponin 2 can influence bFGF signaling, which is consistent with our migration and wound assay data.
Our results are consistent with the notion of bidirectional signaling between the cytoskeleton and MAPK-signaling cascade components. were used in the experiment. ** indicates the difference between EV and sense without U0126 was significant (p Ͻ 0.01) as analyzed by Student's t test. Significant difference (p Ͻ 0.01) was also found between each concentration of U0126 treatment as assessed by single factor analysis of variance. B, MEK-DD can partially rescue the effect of calponin 2 antisense on bFGF-induced HUVEC migration. HUVECs infected with empty control, calponin 2 antisense, empty control plus MEK-DD adenoviruses, and antisense plus MEK-DD adenoviruses were used. Migration assay was done with or without 20 ng/ml bFGF. All comparisons were tested by Student's t test. Significant difference (p Ͻ 0.01) is labeled as **. C, calponin 2 antisense did not influence the effect of MEK-DD on HUVEC migration. HUVECs infected with empty control, empty control plus MEK-DD adenoviruses, and antisense plus MEK-DD adenoviruses were used. Migration assay was done without bFGF. All comparisons were tested by Student's t test. Significant difference (p Ͻ 0.01) is labeled with **, and no significant difference (p Ͼ 0.05) is labeled as #.  MARCH 10, 2006 • VOLUME 281 • NUMBER 10 For example, cytoskeletal changes caused by cytochalasin D, a drug that blocks actin polymerization, inhibit the activation of ERK and p38 MAPK, suggesting that the actin cytoskeleton plays a role in MAPK signaling (41). Certain cytoskeletal motor proteins (for example, kinesins) can activate MAPK members (42). Our finding provides new evidence that cytoskeletal proteins can have a direct impact on MAPK activity.

Function of Calponin 2 in Endothelial Cells
ERK plays an important role in the regulation of cell migration (43,44), and ERK activity has been reported to be necessary and sufficient in inducing EC migration (45). In particular, ERK plays a crucial role in bFGF-stimulated EC migration, because inhibition of ERK activity by the MEK1 inhibitor PD98059 blocks bFGF-induced bovine aortic endothelial cell migration (37). In bFGF-deficient ECs, there is a lack of ERK activation and cell migration (46). Based on these data and our findings that lowering calponin 2 expression in ECs (via the antisense adenovirus) causes a blunting of bFGF-stimulated EC migration, we suggest that calponin 2 expression is critical for bFGF action in endothelial cells.
The molecular mechanism of how calponin 2 influences ERK1/2 activity is not fully understood. Our MEK inhibitor studies suggest that it is acting upstream of or at MEK1/2, because the inhibitor U0126 almost completely blocked calponin 2-induced EC migration, and calponin 2 overexpression did not rescue the effect of the inhibitor on ERK1/2 phosphorylation. The partial rescue of MEK-DD on the effect of calponin 2 antisense adenovirus on bFGF-stimulated HUVEC migration provides further evidence for this hypothesis. One possibility is that it could act through PKC, similar to calponin 1, which is known to activate PKC autophosphorylation (15), and PKC is able to phosphorylate Raf, an upstream activator of MAPK/ERK. Future studies will determine whether calponin 2 can also bind to PKC and activate it. If so, it may up-regulate MAPK by promoting the formation of a signaling complex consisting of PKC, Raf, MAPK, and ERK.
In summary, we have provided evidence that calponin 2 expression is critical for normal vessel development in the zebrafish. The phenotype seen is most consistent with the defect in EC migration, which we have corroborated via two in vitro assays. Moreover, we have implicated calponin 2 as an intermediary in mediating the in vitro promigratory actions of bFGF.