Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A key vascular endothelial cell (EC) function is to serve as a semiselective barrier between circulating blood and the interstitial fluid. Agonist-induced EC activation is tightly linked to the disruption of EC monolayer integrity, an increase in vascular permeability, and ultimately organ dysfunction. Although increased vascular EC permeability is a cardinal feature of inflammation and thrombotic processes, the biochemical basis for regulating this important vascular response remains poorly understood. Early studies by Majno and Palade (1) determined that a breach in the integrity of the EC monolayer induced by bioactive agents involves the formation of intercellular gaps, a morphologic change mechanistically linked to interstitial edema. Majno and colleagues (2) observed that the nucleus and cytoplasm of stimulated EC undergo changes characteristic of contraction including fibrillar band formation at regular intervals. Confluent EC form specific microfilamentous structures ("dense peripheral bands"), including -actin, myosin, and tropomyosin, that colocalize with F-actin (3). EC activation evoked by the serine protease thrombin results in significant alterations in F-actin filament distribution, loss of dense peripheral bands, an increase in the number of cytoplasmic stress fibers (4), and isometric force development (5), providing significant cellular shape changes and dramatic increases in EC monolayer permeability (6). In well-developed smooth-muscle (SM) model systems, contraction is initiated by an increase in cytosolic free Ca²⁺ that associates with calmodulin (CaM) to activate the Ca²⁺/CaM-dependent myosin light chain kinase (MLCK). This increase in myosin light chain (MLC) phosphorylation results in activation of Mg²⁺-adenosine triphosphatase actomyosin activity and finally force generation (7). In EC, centripetal tension generation appears to be similarly driven by an actomyosin molecular motor tightly linked to MLC phosphorylation catalyzed by a Ca²⁺/CaM-dependent MLCK (5, 8, 9). Reductions in cytosolic Ca²⁺_i or inhibition of MLCK enzymatic activity results in loss of isometric tension and marked attenuation of thrombin-stimulated EC gap formation and barrier dysfunction (5, 8). These data indicate that SM and nonmuscle cells such as EC share similar contractile machinery in which MLCK-catalyzed MLC phosphorylation is a key event (5, 8, 9, 11, 12). Activity of EC MLCK is a central determinant of basal and agonist-mediated permeability properties (8).

Despite numerous similarities between EC and SM contractile processes, distinct differences exist in the specific cytoskeletal and regulatory proteins involved in contractile events. Specific observations that illustrate distinct contractile regulation by EC and SM cells include: (1) unlike SM, increases in Ca²⁺ in EC are necessary but insufficient to increase EC MLC phosphorylation (8, 13); (2) unlike SM, two Ser/Thr phosphatases (types 1 and 2B) are involved in agonist-induced MLC dephosphorylation in endothelium (14); and (3) whereas the phosphorylation of calponin (a CaM-, actin-, and Ca²⁺-binding protein) is an important regulatory event in SM contraction, immunoreactive calponin is absent in both human and bovine EC (W. T. Gerthoffer and J. G. Garcia, unpublished observations). Recently, we have identified a novel high molecular-weight non-muscle MLCK isoform (214 kD) in cultured endothelium that is unique and distinct from its SM MLCK counterpart (130 to 150 kD). Molecular cloning of this enzyme from human endothelium identified an 8.1-kb complementary DNA (cDNA) with substantial homology (> 95%) to the coding region of the rabbit and bovine SM MLCK in amino acids #923-1914 (15). Sequence analysis also identified, however, a 5' stretch of novel sequence (amino acids #1-922) that is not contained in the open reading frame of mammalian SM MLCK (16). Given the important role of EC MLCK in the regulation of EC barrier properties, in this report we have further characterized the expression of this novel EC MLCK isoform in several adult and cultured EC tissues. Our data suggest that the EC MLCK isoform may represent a novel nonmuscle MLCK isotype that is preferentially expressed in endothelium.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Endothelial cell cultures were maintained in M-199 or DME media (GIBCO, Chagrin Falls, OH) supplemented with 20% (vol/vol) colostrum-free bovine serum (Irvine Scientific, Santa Ana, CA), 1% antibiotic and antimycotic (penicillin, 10,000 units/ml; streptomycin, 10 mg/ml; and amphotericin B, 25 µg/ml) (K.C. Biologicals, Lenexa, KS), and 0.1 mM nonessential amino acids (GIBCO). Unless specified, reagents were obtained from Sigma Chemical Company (St. Louis, MO). Phosphate-buffered saline (PBS) and Hanks' balanced salt solution without phenol red were purchased from GIBCO (Grand Island, NY).

Preparation of Human and Bovine Endothelial Cells

Human umbilical vein EC (HUVEC) were isolated from umbilical cords as previously described (17, 18) by treatment with collagenase (0.1%) to detach the endothelium. The cells were plated onto gelatin-coated 25-cm² flasks and cultured with M199 medium containing 20% colostrum-free bovine serum (Sigma), heparin (2 U/ml), endothelial cell growth factor (30 µg/ml; UBI, Lake Placid, NY), and antibiotics. Confluent cells were propagated twice by detachment with 1 ml trypsin/ethylenediaminetetraacetic acid (EDTA) (0.05%/0.02%) for 1 to 2 min at 25°C and replating. The purity of the cells was confirmed by the typical cobblestone morphology judged routinely by light microscopy and by the presence of von Willebrand factor as assessed by immunofluorescence microscopy. Human dermal microvascular EC (HDMVC) (CC-2505; Clonetics Corp., San Diego, CA) were obtained frozen at passage 3-4 and cultured 2 to 3 additional passages with microvascular endothelial cell growth media (CC-3125; Clonetics).

Bovine pulmonary artery endothelial cells (BPAEC) were obtained from the main pulmonary artery, secondary branches were immediately dissected en bloc from heart/lung/trachea preparations at the slaughterhouse and packed in ice-cold saline. The lumen of the vessel was carefully exposed by longitudinal slicing and the endothelium was removed by gentle scraping with minimal fibroblast and SM cell contamination. The cells were frozen and stored at 70°C. A small sample of the freshly isolated cells was plated and grown to confluent monolayers to demonstrate their purity by phase-contrast microscopy and staining for von Willebrand factor antigen. For routine studies of cultured BPAEC, cells (CCL-209) were obtained as cryopreserved cells at the sixteenth passage from American Type Culture Collection (Rockville, MD) and cultured as previously described (6) using complete medium (DMEM) with 10 µg/ml endothelial cell growth supplement (UBI) and 20% colostrum-free bovine calf serum (Sigma). Cells were used from passages 19-24 and were plated onto 100-cm² dishes.

Bovine aortic endothelial cells (BAEC) were obtained from the aorta dissected en bloc from the heart/lung/trachea preparation as described for BPAEC. Bovine lung microvascular endothelial cells (BMVC) were a kind gift from Dr. Peter Del Vecchio (Albany Medical College, Albany, NY) at passage 11. The microvascular cells were passaged and cultured as previously described (13) and were propagated only three times. Bovine aortic smooth muscle cells (BASMC) were obtained by the explant technique from fresh cow aorta as described (19), cultured in DMEM with 10% serum and antibiotics, and used during passages 2-6.

Preparation of Specific Bovine and Rat Tissues

Bovine tracheal SM strips were dissected from the noncartilaginous muscle span between the cartilaginous rings of the bovine trachea obtained from en bloc bovine heart/ lung/trachea preparations. Visible adipose and connective tissue were carefully removed to improve purity and 3-mm cubes of SM were snap-frozen in liquid nitrogen and stored at 70°C. Lung samples, obtained from bovine lung, were sectioned into 3-mm cubes and snap-frozen in liquid nitrogen. Skeletal muscle samples from the gastrocnemius muscle of an adult rat were sectioned into 3-mm cubes, and cubes selected with minimal connective and vascular tissue were frozen at 70°C. The inner epithelial lining and glandular tissue was removed from samples of stomach from an adult rat, and 3-mm cubes of smooth muscle were frozen for later analysis.

Western Immunoblotting

Proteins were extracted from cultured and freshly isolated EC and SM cell preparations using sodium dodecyl sulfate (SDS) sample buffer as previously described (20). Extracts were separated by 4 to 15%-gradient SDS-polyacrylamide gel electrophoresis (PAGE) (Bio-Rad, Hercules, CA), transferred to nitrocellulose (30 V overnight or 100 V for 2 h), and reacted with MLCK-specific antibodies as previously described (14, 21). Immunoreactive proteins were detected using an enhanced chemiluminescent detection system according to the manufacturer's directions (Amersham, Little Chalfont, Buckinghamshire, UK). For normalization of protein loading, tissue or cell extracts were initially subjected to SDS-PAGE and stained with Coomassie, and the relative amount of protein loaded in each lane was quantitated by densitometry on GS-670 Imaging densitometer (Bio-Rad). To obtain an equal amount of proteins for each probe, the sample volume was adjusted with SDS sample buffer and normalized proteins were subjected to Western immunoblotting. As noted in Table 1, four antibodies recognizing MLCK were used to study EC MLCK expression. Monoclonal antibody K36 (Sigma) was produced against purified chicken gizzard SM MLCK and recognizes SM MLCK from mammalian and avian species. Polyclonal antibody (D119), a generous gift from Dr. P. Gallagher (Indiana University, Indianapolis, IN) was produced in rabbits against a peptide representing the 12 amino acid repeat sequence (KPVGNAKPAETL) between amino acid residues #102-293 in the rabbit and bovine SM MLCKs (15, 22). Polyclonal antibody (V368) was raised in rabbits against a synthetic peptide corresponding to residues #368-374 of the newly cloned human EC MLCK (15). Polyclonal antibody raised in rabbits against highly conserved, independently expressed C-terminal fragment of SM MLCK known as kinase-related protein (KRP, or telokin) was a generous gift from Dr. D. M. Watterson (Northwestern University, Chicago, IL).

MLCK Immunoprecipitation under Denaturing Conditions

For immunoprecipitation, confluent EC monolayers in 60-mm tissue culture dishes were rinsed twice with 2 ml medium, further rinsed with 2 ml PBS, and scraped into 100 µl SDS/denaturing stop solution (PBS, pH 7.4; 1 mM EDTA; 1 mM ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid; 50 mM NaF; 10 mM NaPP; 0.2 mM orthovanadate; 1% SDS; and 14 mM -mercaptoethanol). The homogenate was prepared by passing the cell suspension several times through a 16-gauge needle. Homogenates were heat-treated at 110°C for 5 min, diluted 1:10 with 900 µl PBS and incubated with 50 µl of 10% Pansorbin suspension containing formalin-hardened and heat-killed Cowan 1 strain Staphylococcus aureus cells (Calbiochem, La Jolla, CA) for 30 min at room temperature. Samples were clarified by microcentrifugation (Eppendorf, 5 min) and supernatants were incubated with 10 µl anti-MLCK antibodies (60 min at room temperature or overnight at 4°C), then with 50 µl of 10% Pansorbin suspension for 60 min at room temperature. Immunocomplexes were pelleted by microcentrifugation for 5 min, washed three times in 1 ml PBS, boiled in 100 µl of SDS sample buffer (20) for 5 min, separated from Pansorbin by microcentrifugation, and subjected to SDS electrophoresis (20). After electrophoresis, proteins were transferred to nitrocellulose membranes (21) and signals were detected by immunostaining with anti-MLCK antibodies.

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)

Total RNA from HDMVC, adult BAEC, and BPAEC were extracted using guanidine isothiocyanate and CsCl₂ centrifugation according to Sambrook and coworkers (23). Approximately 1 µg of total RNA was reverse transcribed using Superscript II (GIBCO BRL, Gaithersburg, MD). The synthesized first-strand cDNAs were subjected to PCR amplification using oligonucleotide primers that flank base pairs 2781-2174. PCR amplifications were performed using 1.5 mM MgCl₂, 200 µM dNTP, and 1.25 units taq polymerase (Boehringer Mannheim, Indianapolis, IN) on Perkin-Elmer PCR Model 9600 (Perkin-Elmer, Foster City, CA). The cycle conditions include 94°C for 30 min, 60°C for 30 min, and 72°C for 1 min for 35 cycles.

Immunohistochemical Analysis

Immunocytochemical localization of MLCK was assessed in normal and transplanted human cardiac tissues with the D119 polyclonal MLCK antibody. Right ventricular endomyocardial biopsies from two donor hearts obtained before transplantation and from two cardiac allografts obtained 7 d and 3 yr after transplantation were embedded in Optimum Cutting Temperature medium compound (Miles Co., Elkhart, IN), snap-frozen in liquid nitrogen, and stored at 20°C. Tissues were sectioned (4 µm) on a Tissue-Tek cryostat (Miles), removed from the cryostat blade by flash condensation onto glass microscope slides, and air-dried at room temperature. The sections were washed in 10 mM PBS, incubated 30 min at room temperature with D119 antibody (1:100), washed three times in PBS, and incubated 30 min with fluorescein isothiocyanate-conjugated goat F(ab')₂ antirabbit antibody (Biomeda, Foster City, CA). To confirm the spatial localization of MLCK within the microcirculation, double antibody experiments (24, 25) with monoclonal antibodies to SM-specific -actin (IA4; BioMaker, Rehovot, Israel) to localize vascular SM cells or with PECAM-1 (CD31; Dako, Santa Barbara, CA) to localize EC (25, 26) were performed. Mouse monoclonal antibodies were followed by rhodamine-conjugated goat F(ab')₂ antimouse antibody (Protos Immunoresearch, San Francisco, CA). Control experiments were performed by exposing tissue sections to PBS, to secondary antibodies only, or to irrelevant isotype-matched primary antibodies. The sections were coverslipped with aqueous mounting media and examined with a Leitz microscope equipped for epifluorescence N2.1-type rhodamine and type A fluoroblue interference filters. Single- and double-exposure photomicrographs were taken by using a Leitz camera and 35-mm Ektachrome film (ASA 200).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of MLCK in Cultured Human and Bovine Endothelium

We have recently reported the cloning of a novel high molecular-weight MLCK isoform from a HUVEC library with an apparent molecular weight on SDS-PAGE of 214 kD (15). Although the EC MLCK isoform is clearly distinct from the SM MLCK isoform (130 to 150 kD) (15), as depicted in Figure 1, our initial studies demonstrated that EC MLCK contains all regional motifs previously reported in SM MLCK (16), including the putative actin-binding domain (AA #923-964), MLC-binding region, catalytic core region, CaM-binding region, and the C-terminal region corresponding to the sequence of the SM protein known as KRP or telokin. Unlike the SM MLCK isoform, however, N-terminal amino acids #1-922 from the deduced HUVEC MLCK sequence are not found in the reported SM MLCK structure and do represent the unique EC MLCK isoform protein sequence. Therefore, we hypothesized that HUVEC MLCK may represent a novel nonmuscle high molecular-weight MLCK isoform expressed in endothelium. To study species- and site-specific expression of the EC MLCK isoform we used a panel of antibodies of varying specificity (Table 1). We first used a monoclonal antibody (clone K36; Sigma) produced against purified chicken gizzard SM MLCK, and a polyclonal antibody (D119) directed against the repeated 12-amino-acid sequence downstream of the putative actin-binding region (AA #1-41) of rabbit SM MLCK (Figure 1). Our preliminary experiments and published data (15, 22) demonstrate that the K36 antibody preferentially recognizes SM MLCK, whereas D119 antiserum preferentially detects the high molecular-weight MLCK isoform. Figure 2 demonstrates that cultured HUVEC and BMVC express only the high molecular-weight MLCK isoform, whereas cultured BASMC appear to express both high and low molecular-weight MLCK isoforms. Consistent with these results, HUVEC freshly isolated from umbilical cords expressed only the high molecular-weight MLCK isoform (data not shown).

View larger version (20K): [in this window] [in a new window]   Figure 1.   Schematic representation of human EC MLCK and rabbit uterine SM MLCK protein structures. Both MLCK isoforms contain (from C- to N-terminal) the KRP region, the CaM-binding domain, the catalytic domain, the MLC-binding domain, and the repeated motif. In addition, EC MLCK includes the unique N-terminal fragment (shown as a closed rectangle), which is absent in the SM MLCK structure. Total amino acid numbers for each isoform and positions of sequences used for antibody generation are indicated.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Detection of the high molecular-weight MLCK isoform in cultured endothelial and SM cells. Proteins from cultured EC and SMC were solubilized in SDS sample buffer, separated by 4 to 15% gradient SDS-PAGE (Bio-Rad), transferred to nitrocellulose, and reacted with a monoclonal antibody directed against chicken gizzard MLCK (K36, A) or to the rabbit SM MLCK 12-amino acid repeat region (D119, B) as described in MATERIALS AND METHODS. In this immunoblot, lane 1 represents human macrovascular EC (HUVEC); lane 2, BMVC; and lane 3, bovine aortic SM cultured cells (BASMC). The amounts of total proteins were normalized by quantitative densitometry of Coomassie-stained gels. The positions of the molecular weight prestained markers as well as the 214-kD EC MLCK and the 150-kD SM MLCK bands are indicated. Passaged macro- and microvascular EC express only the high molecular-weight MLCK isoform, whereas passaged aorta SM cells express both high and low molecular-weight MLCK isoforms.

KRP is an independently expressed protein considered to be a marker for the SM phenotype (27). This sequence is also contained in the COOH terminus of SM MLCK and is highly conserved in the C-terminus of the EC MLCK with more than 90% homology with SM MLCK (15). When the expression of KRP was examined in cultured endothelium using specific anti-KRP antibodies (Figure 3), in contrast to SM, KRP was only weakly detected in EC. The reason for the reduced KRP expression in EC when compared with SM is not clear, but this reduction is likely due to variations in transcriptional regulation between the two tissues. However, we cannot exclude the possibility that the KRP antibody produced against purified SM KRP (Table 1) may not efficiently recognize KRP when expressed in endothelium.

View larger version (49K): [in this window] [in a new window]   Figure 3.   Comparison of KRP expression in SM and endothelium. Bovine tracheal SM and BPAEC homogenates were separated by 4 to 15% gradient SDS-PAGE (Bio-Rad), transferred to nitrocellulose, and reacted with anti-KRP antibodies. Each lane represents equivalent amounts of total protein as normalized by Coomassie-stained gel quantitation. The positions of the molecular weight markers and KRP are indicated. The KRP protein (the independently expressed C-terminal part of SM MLCK) is expressed only minimally in endothelium.

Expression of MLCK in Freshly Isolated Adult Bovine Endothelium

To determine MLCK isoform expression in adult EC, we isolated EC from BPAEC and BAEC and analyzed SDS cell extracts by Western immunoblotting with K36 and D119 antibodies. Figure 4 demonstrates that freshly prepared EC from adult bovine tissues express the 214-kD high molecular-weight MLCK isoform. Densitometric quantitation of MLCK detected by Western blot analysis demonstrates similar level of expression of EC MLCK in both cultured and freshly isolated cells (Figure 4C), whereas expression of the SM MLCK isoform was not detected in either EC preparation.

View larger version (21K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   Figure 4.   MLCK expression in adult endothelium. Proteins from bovine SM tissue (tracheal SM, lane 1), freshly isolated BPAEC (lane 2), and EC from adult aorta (BAEC, lane 3) were solubilized in SDS sample buffer, separated by 4 to 15% gradient SDS-PAGE (Bio-Rad), transferred to nitrocellulose, and reacted with either K36 anti-SM MLCK monoclonal antiserum (A) or D119 polyclonal anti-MLCK antiserum (B and C). C represents the comparison of the level of MLCK expression in freshly isolated and cultured BPAEC by Western immunoblotting of cell homogenates. Each lane represents an equivalent amount of total protein as normalized by Coomassie-stained gel quantitation. The positions of the molecular weight prestained markers as well as the EC MLCK and SM MLCK bands are indicated. Freshly isolated or passaged adult EC express only the high molecular-weight MLCK isoform, whereas SM tissues highly express the low molecular-weight MLCK isoform with only minimal expression of the high molecular-weight MLCK isoform.

Immunodetection of EC MLCK in Human, Bovine, and Rat Tissues

We next extended these biochemical observations by demonstrating that the 214-kD EC MLCK isoform is present in human endomyocardial biopsies from normal donor hearts (n = 2) and from patients who have undergone cardiac transplantation (n = 2). Immunohistochemical analysis of human endomyocardial biopsies with anti-MLCK D119 antibodies (seen as blue in Figure 5) reveals the presence of the kinase in capillary, venous, and arteriolar endothelium, but not in cardiac muscle tissues of normal hearts (Figure 5) or hearts from cardiac allograft recipients (data not shown). The localization of MLCK reactivity on endothelium was confirmed using the double-antibody technique with an antibody to the EC-specific monoclonal marker PECAM (CD31) (Figure 5).

View larger version (65K): [in this window] [in a new window]   Figure 5.   Expression of MLCK in human endomyocardial biopsies. Expression of the EC MLCK isoform was assessed by the immunocytochemical double-antibody technique using myosin light chain kinase (blue) and endothelial cell-specific CD31 (red) antibodies. Both antibodies identify protein present in capillary, venous, and arteriolar endothelium but not in cardiac muscle. Furthermore, these studies demonstrate clear colocalization of PECAM (CD31) and EC myosin light chain kinase in vascular endothelial cells, as shown by the pink reactivity signal (arrows). Immunofluorescence was not detected in tissue sections exposed to PBS alone, to conjugated secondary antibodies alone, or to irrelevant isotype-matched primary antibodies. Original magnification: ×400.

To study specific expression of the high molecular-weight EC MLCK isoform in different cell types, we used antibodies (V368) designed against the novel N-terminus of this enzyme (Figure 1) for Western blotting analysis of cell lysates. Figure 6 demonstrates that V368 antibodies cross-react only with the 214-kD MLCK isoform but not with the lower molecular-weight MLCK isoforms in all cell types examined. Examination of bovine EC and bovine and rat tissues (Figure 6) revealed that the 214-kD protein was highly expressed in endothelium and lung (highly enriched in endothelium), had very low expression in rat stomach SM, and was not expressed in rat skeletal muscle. It is interesting to note that the V368 antibody (which is specific to the EC MLCK isoform) does not recognize any protein in the tracheal SM sample (Figure 6, panel B), whereas the D119 antibodies (which preferentially recognize nonmuscle MLCK in different tissues) have weak cross-reactivity with a ~ 208-kD protein band in the same sample (Figure 4, panel B). The appeared discrepancy between the results with two different MLCK antibodies may reflect the existence of two high molecular-weight MLCK isoforms differentially expressed in SM and endothelial cells. The specific EC MLCK antibody (V368) weakly cross-reacts with a 214-kD protein band in rat stomach SM (Figure 6, panel A), which may reflect the limited nonmuscle contamination of this sample. Collectively, the immunologic data presented in Table 2 indicate that both cultured and in situ adult bovine and human endothelium express only the unique high molecular-weight 214-kD MLCK isoform.

View larger version (35K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Figure 6.   Endothelial cell-specific MLCK expression. Proteins from bovine cultured EC and bovine and rat tissues were assessed for immunoreactivity with V368 antibodies generated against a unique N-terminal region of human EC MLCK. (A) Lane 1, BPAEC; lane 2, rat stomach SM (SmM); lane 3, rat skeletal muscle (SkM). (B) Lane 1, BPAEC; lane 2, bovine lung; lane 3, bovine tracheal SM (trachea). The positions of the molecular weight prestained markers as well as the EC MLCK band are indicated. EC MLCK is highly expressed in cultured endothelium, minimally expressed in SM, and not expressed in skeletal muscle tissues.

Structural Relationship between Bovine and Human EC MLCKs

To characterize the degree of structural similarity between the previously cloned human EC MLCK and the bovine 214-kD MLCK isoform, we used several MLCK antibodies directed against different SM and EC MLCK epitopes (Figure 1) to immunoprecipitate MLCK from BPAEC under denaturing conditions. Western blotting analysis of immunoprecipitates with D119 anti-MLCK antibodies (Figure 7A) demonstrated that each antibody was able to immunoprecipitate the 214-kD protein in bovine endothelium, confirming a high degree of similarity between bovine and human EC MLCK in at least several antigenic epitopes. We next used RT-PCR to examine the similarity of EC MLCK in human and bovine endothelium using oligonucleotide primers which flank base pairs #2174-2781. As expected, only one amplified ~ 0.6-kb product was found in human (Figure 7B) and bovine (Figure 7C) EC, verifying that adult bovine EC express a high molecular-weight MLCK isoform that is highly similar to the human EC MLCK isoform.

View larger version (44K): [in this window] [in a new window]   View larger version (48K): [in this window] [in a new window]   View larger version (39K): [in this window] [in a new window]   Figure 7.   Structural similarity of bovine and human EC MLCK. (A) Anti-MLCK immunoprecipitates were prepared from BPAEC extracts under denaturing conditions with MLCK-specific antibodies and subjected to gradient SDS-PAGE. This was followed by Western immunoblotting with specific anti-SM MLCK repeat region D119 antibodies. In this immunoblot, lane 1 represents V368 antibody immunoprecipitates; lane 2, D119 antibody immunoprecipitates; and lane 3, K36 antibody immunoprecipitates. Positions of EC MLCK and molecular weight markers are indicated. The MLCK isoform from bovine endothelium strongly shares antigenic determinants with human EC MLCK and rabbit SM MLCK. (B and C) An approximately 600-bp PCR-amplified fragment (AF) was detected in cultured HDMVC (B), BAEC (C, faint band in the left lane), and BPAEC (C, right lane) using human EC MLCK-specific primers. The positions of base-pair markers are indicated. These data indicate substantial nucleotide similarity between the EC MLCK isoforms derived from bovine and human species.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To understand the regulation of EC MLCK by Ca²⁺/CaM and other regulatory elements, we have recently cloned and sequenced the novel, high molecular-weight MLCK isoform expressed in human endothelium (15). The deduced amino acid sequence between AA #923-1914 revealed substantial homology to bovine stomach SM MLCK (> 90%) and to rabbit uterine SM MLCK (> 85%). However, the much larger molecular mass and the unique 5' open reading frame sequence (AA #1-922) distinguishes the 214-kD EC MLCK from the SM kinase (15, 16). To characterize the EC MLCK isoform further, we examined its expression in different cell types and tissues using antibodies specific for SM and EC MLCK (Table 2). The 214-kD EC MLCK, but not the SM MLCK isoform, is detected in cultured micro- and macrovascular EC and, more importantly, in EC freshly isolated from adult tissues. Consistent with this observation, the nonmuscle high molecular-weight MLCK transcript was recently detected in adult liver (28). Examination of the expression of the 214-kD isoform with specific anti-EC MLCK antibodies revealed very low expression of this MLCK isoform in adult stomach SM, possibly reflecting nonmuscle contamination of this tissue sample with EC MLCK isoform expression completely absent in tracheal SM and skeletal muscle. Immunocytochemical studies further localized the EC MLCK isoform to endothelium from normal adult human hearts and human cardiac allografts but not in cardiac muscle. This pattern of expression is significantly different from the expression of the recently described 208-kD "embryonic" MLCK isoform, which appears to be preferentially expressed in undifferentiated embryonic cells and early embryonic tissues (22). The possibility continues to exist that the "embryonic" 208-kD MLCK isoform as well as the high molecular-weight nonmuscle MLCK isoform described by Fisher and Ikebe (28) are developmentally regulated in SM but not in nonmuscle tissues. It is worth noting that despite the similarity in the molecular weights, the embryonic 208-kD MLCK isoform appears to exhibit specific structural differences when compared with our recently cloned EC MLCK 214-kD isoform. For example, we have recently shown that anti-KRP antibodies recognize EC MLCK (15) but do not cross-react with the 208-kD embryonic MLCK isoform (22), suggesting structural differences in the C-terminal portions of these high molecular-weight MLCK isoforms. Our present data indicate that specific anti-EC MLCK V368 antibodies do not cross-react with the 208-kD MLCK isoform minimally present in adult tracheal smooth muscle (22), suggesting a structural difference in the N-termini of these isoforms. Consistent with these results, our recent observations indicate that the high molecular-weight MLCK from smooth muscles binds more avidly to the cytoskeleton than does EC MLCK (A. D. Verin and J. G. N. Garcia, unpublished data). Comparison of the 214-kD EC MLCK isoform with the recently revised nonmuscle avian fibroblast MLCK sequence (29, 30) revealed substantial homology in the C-terminal parts of the two molecules (amino acids #923- 1914), but much less homology between amino acids 1 through 922. For example, only 38% homology exists between the two isoforms in the region flanked by residues #200-500 (15). In contrast, comparison of the EC MLCK isoform present in both human and bovine EC, either by immunoreactivity or by RT-PCR analysis, fails to reveal significant structural differences in the EC MLCK from the two species.

Together, these data suggest that cultured and adult bovine and human endothelium express only the unique nonmuscle MLCK isoform which, despite significant homology, may be distinct from either the embryonic 208-kD MLCK isoforms or nonmuscle MLCK isoforms present in other cell types. We speculate that members of this family of nonmuscle MLCKs may exhibit distinct functional activities and regulatory properties. Further studies designed to examine the biochemical regulation of the unique EC MLCK isoform are under way.