EXPEHIMENTAL
CELL
RESEARCH
190,
1
10 (19%))
Nonmuscle and Smooth Muscle Myosin lsoforms in Bovine Endothelial Cells ANNA C. RORRIONE,* ANNA MARIA C. ZANELLATO,* LUCA GIURIATO,* GIANLIJICI SCANNAPIEW,~ PAOLO PAULETTO,$ AND SAVERIO SARTORE*+’ *Institute
of General Pathology, $Institute
of Clinical Medicine, University Muscle Biology and Physiopatholugy,
A panel of monoclonal antibodies, specific for human platelet (NM-AS, NM-F6, and NM-G2) and for bovine smooth muscle (SM-E7) myosin heavy chains (MHC), were used to study the composition and the distribution of myosin isoforms in bovine endothelial cells (EC), in Using indirect and double immunofluvivo and in vitro. orescence techniques, we have found that in the intact aortic endothelium there is expression of nonmuscle MHC (NM-MHC), exclusively. By contrast, hepatic sinusoidal endothelium as well as cultured bovine aortic EC (BAEC) in the subconfluent phase of growth show coexistence of NM- and smooth muscle MHC (SM-MHC) isoforms. SM myosin immunoreactivity disappears when cultured BAEC become confluent. In this phase of cell growth, NM-MHC isoforms are localized differently within the cells, i.e., in the cytoplasm around the nucleus or in the cortical, submembranous region of EC cytoplasm. A third type of intracellular distribution of NM-MHC immunoreactivity was evident in the cell periphery of binucleated, confluent BAEC. These data indicate that (1) several myosin isoforms are differently distributed in bovine endothelia; and (2) SM myosin expression and the specific subcellular localization of NM myosin isoforms within EC might be regulated by cellcell interactions. 2 1990 Academic Press, Inc.
INTRODUCTION The cytoskeleton plays a crucial role in controlling cell movement and migration as well as in the maintenance of the cell shape and in the anchorage of cells to each other and to the underlying extracellular matrix 11-S]. It is assumed that the contractile force which is necessary for cell motility is provided by an actomyosin system. Actin and myosin [7-lo], as well as accessory contractile proteins (such as a-actinin and tropomyosin) have been demonstrated in NM cells both in vitro 1 To whom reprint requests should he addressed at Institute of General Pathology, IJniversity of Padova, Via l’rieste, ‘i5, 35131 Padova, Italy.
of Padova, and $National
Research !‘owwil
1rnit for
Padova, Ita1.v
and in situ [ 11-141. Recent studies have shown that the bundles of microfilaments which give rise LO the stress fiber system have probably a structural role rather than an involvement in cell motility 131. For example, in the vascular endothelium, stress fibers 115, 161 have been observed predominantly in those EC subject to hemodynamic forces: they may apply tension to resist the shear forces acting on the cells, allowing the-m to maintain their flattened shape and to remain firmly attached to the substratum [ 11, 16,171. In the cell periphery of confluent EC grown in vitro, microfiiaments can arrange differently compared with the stress fibers, giving rise to a distinct, F-actin-containing, microfilamentous structure, the so-called dense peripheral bands (DPB, Ref. [18]). The same authors [ 181 have found that a-actinin, myosin, and tropomyosin cc-localize with F-actin, suggesting that DPB may have the capacity to contract. The presence of DPR might be associat,ed with the ability of EC to form and maintain the cell monolayer. It is not known whether the peculiar organization of cytoskeletal apparatus in EC is due to a different distribution of the same set of cytoskeletal proteins or to the existence of structural variants. As concerns myosin, the different isoforms found in the sarcomeric muscles have been correlated with specific enzymatic and functional patterns (reviewed in Ref. 1191). in NM cells, such as platelets, myosin seems to consist of two isoforms [20-221 differently distributed in the cytoplasm and in the soiubilized membranes. Similarly, in the brush border of intestinal cells the two different subsets of myosin are localized in different manners [23]. Recent studies carried out on the eukaryotic amoebae, Dictiostelium discoideum have shown surprisingly that locomotion does not require myosin; conversely, this protein plays a crucial role in cytokinesis and in developmental morphogenesis [24, 25]. However, it could also he possible that amoeboid iocomotion is more linked to a minor 117-kDa myosin-like component (discussed in Ref. [26]). The aim of this study was to evaluate the myosin isoform composition and distribution in intact endothelia of aorta and hepatic sinusoids as well as in cultured
BORRIONE
2
BAEC. Our results indicate: (1) the existence of a different pattern of NM and SM myosin isoforms in the two endothelia, and (2) that SM myosin expression and the different distribution of NM myosin isoforms in BAEC are dependent upon the migrating or resting state of the cells. MATERIALS
AND
ET AL.
MHC-
METHODS
Preparation of monoclonal anti-myosin antibodies. Monoclonal anti-SM myosin antibodies (SM-E7) were obtained in our laboratory using bovine aortic actomyosin as immunogen [27]. Monoclonal antihuman platelet myosin antibodies (NM-F6, NM-AS) were produced following the immunization protocol and the cell fusion procedure described for the NM-G2 antibody [27]. The selected hybridomas were cloned by limiting dilution, and then used as culture supernatants, ascitic fluid, or purified IgG. Anti-Von Willebrand factor was purchased from Behringwerke AG, Federal Republic of Germany. Phalloidin conjugated with tetramethylrhodamine isothiocyanate (RITC) was obtained from Molecular Probes, Inc., Oregon, U.S.A. Western blotting. Human and bovine platelets, bovine aorta, and cultured BAEC were used in these experiments. Pellets of platelets, fibroblasts, or BAEC as well as small fragments of bovine aorta were extracted for 3 min in boiling Laemmli’s sample buffer solution [28]. After centrifugation at 10,OOOg for 10 min, the supernatants were collected and stored at -70°C until use (within a few days). Myosin extracts were electrophoresed in 10 or 7.5% gels in the presence of sodium dodecyl sulfate (SDS) and then stained subsequently with Coomassie brilliant blue R. Western blotting was performed essentially according to the procedure described in Ref. [29], using rabbit anti-mouse IgG conjugated with horseradish peroxidase (Dako, Dakopatts a/s, Glostrup, Denmark) to reveal the anti-myosin antibodies bound to blotted antigens. Cell cultures. FCNL 8112 bovine fibroblasts were a generous gift of Dr. Bressan (Institute of Histology and Embryology, University of Padova). BAEC were obtained essentially as described in Ref. 1301. Cells were cultured at 37’ (5% CO, atmosphere) in Dulbecco’s moclfied Eagle’s medium supplemented with 10% fetal calf serum (FCS). For Western blotting, cell monolayers were removed from plastic petri dishes by treatment of cultures of fibroblasts and BAEC with
FIG. 2. Specificity of anti-NM myosin antibodies as determined by Western blotting using NM tissue extracts. Crude antigens from human platelets (lane l), bovine platelets (lane 2), cultured bovine FCNL 81/2 fibroblasts (lane 3), and BAEC (lane 4) were electrophoresed in 7.5% SDS slab gels and blotted onto nitrocellulose paper. NM-AS (B), NM-F6 (C), and NM-G2 (D) antibodies were reacted with blotted NM tissue antigens. A, Coomassie blue staining of parallel gel. MHC, myosin heavy chains.
0.1% trypsin in phosphate-buffered saline (PBS). The activity of trypsin was blocked by adding FCS. Cells were collectedby centrifugation at 1OOOgfor 10 min and then washed twice with PBS. For immunofluorescence experiments, cells were cultured on glass coverslips and fixed at sub- and confluent stages of growth in cold acetone (-20°C) for 5 min. Zmmunofluorescence. Indirect and double immunofluorescence assays were performed on fresh frozen sections (lo-pm thick) of bovine aorta and on acetone-fixed cell cultures, according to the procedure described by Borrione et al. [31]. Briefly, after incubation with the appropriate dilutions of monoclonal antibodies (ascitic fluid or purified IgG) in 1% bovine serum albumine (BSA) in PBS, pH 7.2, for 30 min at 37°C in a humidified chamber, sections and cell monolayers were washed with PBS and then incubated with rabbit anti-mouse IgG conjugated with fluorescein isothiocyanate (FITC) or RITC (Dako). For the double-labeling procedure, one monoclonal antibody was conjugated with FITC and the other was indirectly revealed by rhodaminated anti-mouse IgG [31]. After fixation in 1.5%p-formaldehyde in PBS for 10 min, cryosections or cell cultures were mounted in Elvanol and anti-myosin binding patterns visualized using a Zeiss Axioplan microscope equipped with an epifluorescence apparatus and HBO 100-W high-pressure light. The following controls were performed: omission of the first or the second antibody, or use of an unrelated antibody instead of the first one.
RESULTS FIG. 1.
Western blotting of SM and NM myosin extracts with anti-myosin antibodies. Crude antigens from bovine aortic SM (lanes l,l’, and 1”) and from human platelets (lanes 2,2’, and 2”) were electrophoresed in 10% SDS slab gels and then blotted onto nitrocellulose paper. A, Coomassie blue staining of a parallel gel; B and C, Western blotting with NM-AS (B) and NM-F6 (C) antibodies. MHC, myosin heavy chains.
Characterization
of Anti-Myosin
Antibodies
Monoclonal SM-E7 is specific for myosin heavy chain 1 (MHC-1, M, 205 kDa) and 2 (MHC-2, M, 200 kDa) isoforms of SM-type [27]. This antibody can stain spe-
MYOSIN
ISOFORMS
IN ENDOTHELIAL
FIG. 3. Double immunofluorescence assay with anti-NM myosin wall c01itai ining a thin, continuous layer of cells (arrows) positive with and E) dirl ectly labeled with FITC and with NM-F6 (D), NM-G2 (C), Note th Iat: (1) NM-G2 antibody stains a large number of cells present present in the subendothelial region (F). Bar, 70 km.
CELLS
antibodies on intact endothelium of bovine aorta. Cryosec :tion:is ofac rtic anti-Von Willebrand factor antibodies were treated with NM -A9 (Pb J$ or SM-E7 (F) indirectly stained by anti-mouse IgG coup1 edw ith RI TC. in the elastic regions of aortic media (C), and (2) very rap ‘e SPVI cells are
4
BORRIONE
cifically the bovine vascular and nonvascular SM tissues (not shown). Characterization of anti-human platelet myosin antibodies (NM-AS, NM-F6, and NM-G2) is presented in Figs. 1 and 2. Western blotting analysis performed using 10% SDS gels shows that NM-AS and NM-F6 antibodies can bind to a 200-k.Da component present in the human platelet extract (Figs. 1B and 1C) which corresponds to the MHC (data not shown). The very weak reactivity with a bovine aortic extract is likely to be due to the presence of trace amounts of NM myosin [29, 321. NM-G2 antibody is specific for the heavy chains of human platelet myosin (Ref. [27]; see also Fig. 2D). In immunofluorescence tests, this antibody is able to recognize the EC of bovine pulmonary capillaries and the interstitial cells of renal parenchyma (not shown). Specificity of the three monoclonals (NM-AS, NM-F6, and NM-G2) for MHC from different bovine NM cells was determined by Western blotting with 7.5% SDS gels (Fig. 2). While all the three anti-myosin antibodies can react with human platelet MHC (lane l), two of them (NM-AS and NM-F6) are able to react with extracts from bovine platelets, FCNL 81/2 fibroblasts, and BAEC. NM-G2 displays a very weak immunoreac-
ET AL.
tivity with these NM antigens. The distinct binding properties of NM-AS and NM-F6 on the one hand and NM-G2 on the other were also evident in the intact aortic endothelium (see also Fig. 3) and in cultured FCNL 8 bovine fibroblasts (not shown). Immunolocalization
of Myosin Antigens in the Bovine EC
NM myosin antigenicity in aortic endothelium was analyzed by double immunofluorescence assay and the results of these experiments are shown in Fig. 3. NM-AS and NM-F6 antibodies stained brightly an aortic structure (Figs. 3A, 3B, 3D, and 3E) which was also recognized by anti-Von Willebrand factor antibody (not shown). In contrast to this immunofluorescence pattern, NM-G2 antibody was not able to label the aortic endothelium (Fig. 3C), but gives a strong reactivity with a population of SM cells which are mainly localized within the elastic fibers of aortic SM (Fig. 3C; in preparation). SM-E7 anti-SM myosin antibody do not label the intact endothelium of bovine aorta (Fig. 3F). In the hepatic parenchyma, NM-F6 immunoreactivity is distributed according to two different patterns (Fig. 4A), i.e., a fluorescence localized among the he-
FIG. 4. Double immunofluorescence on cryosections from the bovine liver with NM-F6 indirectly labeled with RITC (A) and SM-Ei’, directly labeled with FITC (B). Two distinct patterns of immunofluorescence can be seen in A, i.e., a honeycomb-like structure (asterisks) and a lilamentous material running at the level of hepatic sinusoids (arrows). Note that the sinusoidal fluorescence is also stained by SM-E’7 (B). Bar, 70 Km.
MYOSIN
ISOFORMS IN ENL)O’~HEI>IAI,
patic plates or in the cell periphery of hepatocytes. On the basis of desminivimentin and phalloidin immunoreactivity, as well as on morphological criteria (in preparation), the two hepatic structures recognized by anti-NM myosin antibodies have been identified as the sinusoidal endothelium and the bile canaliculi, respectively. NM-AS gives the same immunofluorescence pattern of NM-F6 whereas NM-G2 is completely negative with the hepatic parenchyma (not shown). In contrast to the bile canaliculi, the sinusoidal endothelium appears to be double labeled with SM-E7 anti-SM myosin antibody (Fig. 4R). Distribution of myosin isoforms was investigated in Ljitro using subconfluent and confluent cultures of BAEC (Figs. 5-7). These cells, which were positive with anti-Von Willebrand factor antibody (not shown), showed a fibroblast-like morphology in the subconfluent stage. In this phase, BAEC displayed a diffuse staining of cytoplasm with NM-AS (Fig. 5A) with almost no evidence for an organization in bundles of filaments, both at the cell periphery and in the central area of EC. Conversely, NM-G2 was completely unreactive (Fig. 5B) with these cells in this phase of growth. When cells reached confluency, they achieved a characteristic polygonal morphology typical of epithelial cells and EC, with reorganization into a single-cell monolayer composed of closely apposed and nonoverlapping cells. Among the mass of mononucleated cells, the presence of some binucleated cells which are reactive with anti-Von Willebrand factor antibody (not shown) was evident. Under these conditions, NM-AS showed a staining pattern similar to that found in subconfluent cells with the submembranous, cytoplasmic region almost devoid of any fluorescence (Fig. 5C). Interestingly, NM-G2 was found to be negative with the large majority of BAEC except for the larger, binucleate cells in which it was seen at the submembranous level often as short segments, rarely along the entire membrane and never across the cell (Figs. 5D and 5F). The filamentous structure recognized by NM-G2 was also labeled by the toxin phalloidine (Fig. 5E) which binds with high affinity to F-actin. The majority of the cells, which were negative with NM-G2 antibody, displayed a well-defined stress fiber labeling with phalloidine (not shown). We have also performed double-labeling experiments on BAEC monolayers utilizing NM-AS and NM-F6 antibodies (Fig. 6). In addition to giving a diffuse immunofluorescence pattern in the cytoplasm, NM-F6 antibody decorates the cytoplasmic filaments corresponding to the stress fiber system (Fig. 6B). Local differences in the distribution of the immunofluorescence pattern with the two monoclonal antibodies are visible in the cortical region of BAEC cytoplasm (Figs. 6A and 6B). Interestingly, distribution of myosin immunoreactivity was quite different in the sub- and confluent phases of BAEC growth. In cultures of BAEC close to confluency
5
CELLS
or at contluency, NM-AS binding was localized mainly in the cytoplasm around the nucleus with the corticai region, beneath the plasma membrane, almost negative with this antibody (Fig. 6C). This cellular region, possibly corresponding to the area in which DPB have been reported [18], is, however, strongly labeied by NM-F6 (Fig. 6D). Heterogeneity in the cytoplasmic reactivity with NM-F6 was occasionally evident in confluent cultures (Fig. 6D). Cells which were negative or weakly stained with NM-F6 antibody were found to be always positive with NM-AS (Fig. 6C). The potential presence of a SM myosin isoform in BAEC has been assayed utilizing the SM-E7 monoclonal anti-SM myosin antibody j27j. Double immunofluorescence tests carried out using NM-AS and SM-E7 (Fig. 7) in subconfluent and confluent cuitures show that SM-E7 is able to stain the subconfluent BAEC exclusively (compare Figs. 7B and 7D). But, more interestingly, SM-myosin isoform is distributed in a heterogeneous manner among and within BAEC. While NM antibody labels the whole cytoplasm around the nucleus (Fig. 7A), SM-E7 antibody staining is restricted to part of cytoplasm surrounding the nucleus (Fig. I;IB), giving a punctuate pattern of fluorescence (Fig. 7B: see also the inset). DISCUSSION
Taking advantage of a panel of monoclonal anti-myosin antibodies, we have found that bovine endothelia are heterogeneous with respect to myosin isoform composition and distribution. The two intact endothelia examined in this study show a different myosin content; i.e., while in the sinusoidal endothelium both SM and NM myosin isoform coexist, in the aortic endot.helium SM myosin immunoreactivity is absent. However, concomitant expression of SM and NM myosin isoforms can be induced in cultured subconfluent BAEC. In vivo and in vitro coexistence of SM and NM myosin isoforms in EC bears some resemblance to developing and cultured vascular SM cells [27, 331 as well as to the microvascular pericytes [34, 351. Our data may be indicative of a special contractile function of hepatic sinusoidal endothelium (“SM cell-like” EC?) which could be related to a specific embryological origin [36] and/or to some peculiar structural and functional characteristics [37] of this tissue. Since SM myosin antigenicity is distributed according to a punctuate pattern in the BAEC cytoplasm, it is likely that this isoform is not a component of the microfilament system. Herman et al. 131, working on embryonic chick and HeLa cells, have suggested that the large actomyosin filament bundles are associated with nonmotile cells, whereas diffuse actin and myosin distribution can be related to cells in active movement. The functional significance of a SM myosin isoform in the
Fl 1GI. 5. Double immunofluorescence with NM-AS, NM-G2 antibl odies, and phalloidin on subconfluent (A, B) and confluent (C- -F) OfB Al XC. Double immunofluorescence assays with NM-AS coupled d irectly with FITC (A, C) and NM-G2 indirectly stained with R .IT( InE a1nd F giant cells double-labeled (arrow) with rhodaminated phal loidin (E) and NM-G2 (F) indirectly stained with anti-mouse 1W with Ii LITC are visible. Arrow in D indicates the cytoplasmic periphe !ral structure recognized by NM-G2 in binucleate cells. Bar, 12
:ult1 nes (B, D). :oq Iled m.
MYQSIN
ISQFORMS
with NM-AS and NM-F6 FIG. 6. Double immunofluorescence The exI ,eri .ment was performed using NM-AS (A, C) directly labeled distinct .flU lorescence patterns can be observed in the cell periphery of NM-AS and NM-F6 immunostaining intracel !lul ar distribution NM-F6 (arstterisk in C and D). Bar, 12 pm.
IN ENDOTHELIAL
CELLS
antibodies on subconfluent with FITC and NM-F6 (B, of subconfluent cells (arrow (arrow in C and D). Some
(A, B) and confluent (C, D) culture SO fBP dx. D) indirectly stained with 53%. O(:ca sion ally, in A and B). Confluent cultures sb OR7disl :inct cells positive w&h N gal ive with
EVG. 7. Double immunofluorescence assay with anti-NM and anti-SM myosin antibodies on subconfluent (A, B) and confluent (C, D) cul ltures of BAEC. NM-AS directly labeled with FITC (A, C) and SM-E7 indirectly stained with RITC (B, D). In subconfluent cultures, SM-E7 the cytoplasm of BAEC, giving a punctuate pattern of fluorescence in SM-positive cells (arrow; see also inset fc )r major sta iins heterogenously de1Lails). Some cells, positive with NM-AS, are negative or weakly reactive with SM-E7 (star). In other cells, part of the cytoplasm is I iegative wit ;h SM-E7 (asterisk). Confluent cultures are negative with SM-E7 (D). Bars: A-D, 12 pm; inset, 6 pm. a
MYOSIN
ISOFORMS
IN ENDOTHELIAL
BAEC is not clear, but the fact that isoform is not expressed in intact endothelium and in confluent cultures of EC would suggest that SM myosin may be a marker for proliferating/migrating EC. The second important finding reported in this paper concerns the existence of three NM myosin isoforms and their different subcellular localization in cultured aortic EC. In confluent cultures of BAEC one isoform is mainly localized to the cortical cytoplasm and the other is more linked to the cytoplasmic region surrounding the nucleus. These findings are in agreement with (1) the data of Groschel-Stewart et al. [20-221, which indicate the existence of structural variants of NM myosin in membrane and cytoplasm of NM cells, and (2) genetic experiments w&h indicate the presence of two distinct NM-MHC mRNAs in different chicken tissues [38]. A third NM-MHC variant can be identified in binucleate, confluent BAEC at the level of peripheral microfilament bundles. These particular giant-type EC might correspond to dividing cells or to permanent binucleate cells. Confluent EC cultures from adult bovine aorta show a marked reduction of in i3H] thymidine incorporation compared with subconfluent cultures [39]; thus, it is more likely that these giant cells are EC in which nuclear duplication was not followed by cytokinesis. Similar giant cells have been found in human EC from adult thoracic aorta [40] and umbelical vein [41]. The function of this NM myosin isoform is presently unknown. One might speculate that this isoform is more directly involved in the process of cytokinesis of EC and that the perturbation of this event may determine the appearance of this myosin isoform at the submembranous level. There is some evidence which indicates that myosin differently in vascular EC in situ [ 1,5] and in vitro [ 18, 421. In fact, cytoskeletal myosin can be found in both DI’B and in the central microfilamentous bundles where, in association with actin and other accessory proteins [I$], it may be involved in contractile functions. Using a polyclonal anti-SM myosin antibody, Hormia et al. [43] and Gottlieb and co-workers [l&42] have also reported that myosin is present in two structurally distinct areas of EC. We have now identified a protein marker which permits one to distinguish the central and peripheral microfilamentous region. The specific distribution of NM-MHC isoforms within EC is to be achieved by cell-cell contact since in subent cultures of BAEC there is only limited evidence for this process. It is interesting that the NM myosin isoform present in the periphery of EC cytoplasm is also present at the level of the junctional eomplex in the bile canalieuli of bovine liver (Fig. 4; in preparation). Thus, it might be possible that this isoform is better suited for stabilizing cortical microfilament associations which migbt be responsible for cell-cell ad-
CELLS
9
hesion and for maintaining layer Cdl].
the integrity
of
We thank Dr. Elisabetta Deiana, ““M. Negri” Institute, Milan, Italy, for preparing BAEC and Mr. Maurizio MoretLo for technical assistance. EFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17.
18. 19. 20. 21.
22. 23. 24. 25. 26. 27. 28. 29.
Drenckhahn, D., and Wagner, J. (1986) J. @a11Viol. 1102, 738-1747. Spooner, B. S., Yamada, K. M., and Wessels, N. K. (1971) 9: Cell Biol. 49, 595-613. Herman, I. M., Crisona, N. J., and Pollard. T. D. (1981) J. Cell Biol. 90, 84-91. Korn, E. D. (1978) Proc. NatE. Acad. Sei. USA 7 White, 6. E., Gimbrone, M. A., Jr., and Fujiwa Cell &al. 97, 416-424. White, 6. E., and Fujiwara, K. (1986) J. Gek! Biol. 103, 63-70. Becker, 6. G., and Murphy, 6. E. (1969) Amer. ,P. Pathol. 55, 6-37. Lazarides, E., and Weber, K. (1974) Proc. N&l. dcad. Sci. USA 71,2268-2272. Fujiwara, K., and Pollard, T. D. (1976) J. &II BioZ, 71,848-875. Weber, K., and Groschel-Stewart, U. (1974) Proc. N&E. Acad. Sci. USA 71,4561-4564. Byers, H. R., and Fujiwara, K. (1982) J. Celb Viol. 93, 804-811. Byers, H. R., White, G. E., and Fujiwara, K. (1984) in Cell and Muscle Motility (Sbay, J., Ed.), Vol. 5, pp. 83-137. Plenum, New York. Herman, J. M., Pollard, T. D., and Wong, A. 3. (1982) Ann. iv. Y. Acad. Sci. 401, X-60. Rogers, K. M., and Kalnins, V. I. (B983j L&l. Invest. 49, 650-654. Hammersen, F. (1980) Adv. Gabbiani, G., Gabbiani, F., (1983) Proc. Natl. Acad. Sci. Franke, R. P., Grlfee, WI., mayer, C., and Drenckhahn, D. (1984) Nature U!mndoni 307, 648-649. Wong, M. K. K., and Gottlieb, A. I. (1986) Rrteriosclerosis 212-219. Swynghedauw, B. (1986) Whys. Rev. Peleg, I., Kahane, I., Eldor, A., Grosc wart, U., Mestan, J., and Mublrad, A. (1983) J. Biol. Chen. Groschel-Stewart, U., Rakousky, C., Franke, R., Peleg, I., Kahane, I., Eldor, A., and Muhlrad, A. (1985) Cell Tissue IZes. 241, 399-404. Peleg, I., Kabane, I., Eldor, A., Groschel-St dano, S., and Muhlrad, A. (1989) Thrornb. Keller, T. C. S., Conzelman, K. A., Chasa M. S. (1985) 9. Biol. Chem. 100, 1647-1655. Knecht, D. A., and Loomis, W. F. (1987) Science 23 1081-1086. De Lozanne, A., and Spudicb, J. A. (19$7) Science 236, 108661091. Mooseker, M. S. (1985) Annu. Rev. &El Bioi. P, 26-293. Borrione, A. C., Zanellato, A. M. C., Scannapieco, G., Paulette, P., and Sartore, S. (1989) Eur. J. Biochem Laemmli, U. K. (1970) Nature (London) 2: Sartore, S., De Marzo, N., Borrione, A. C., Zanellato, A. M. C.,
10
30. 31. 32.
33. 34.
BORRIONE Saggin, L., Fabbri, L., and Schiaffino, S. (1988) Eur. J. Biochem. 179, 79-85. Balconi, G., and Dejana, E. (1986) Med. Biol. 64, 231-245. Borrione, A. C., Zanellato, A. M. C., Saggin, L., Mazzoli, M., Azzarello, G., and Sartore, S. (1988) Differentiation 38, 49-59. Gaylinn, B. D., Eddinger, T. J., Martino, P. A., Monical, P. L., Hunt, D. F., and Murphy, R. A. (1989) Amer. J. Physiol. 257 [Cell Physiol. 261, C997-C1004. Kawamoto, S., and Adelstein, R. S. (1986) J. Biol. Chem. 262, 7282-7286. Joyce, N. C., Haire, M. F., and Palade, G. E. (1985) J. Cell Biol.
ET AL.
38. 39. 40.
Antonov, A. S., Nikolaeva, M. A., Klueva, T. S., Romanaov, Y. A., Babaev, V. R., Bystrevskaya, V. B., Perov, N. A., Repin, V. S., and Smirnov, V. N. (1986) Atherosclerosis 59, 1-19.
41.
Shirinsky, V. P., Antonov, A. S., Birnkov, K. G., Sobolevsky, A. V., Romanov, Y. A., Kabaeva, N. V., Antonova, G. N., and Smirnov, V. N. (1989) J. Cell Biol. 109,331-339.
42.
Gottlieb, A. I., Spector, W., Wong, (1984) Atherosclerosis 4,91-96.
43.
Hormia, M., Badley, R. A., Lehto, B. P., and Virtanen, Exp. Cell Res. 15’7, 116-126.
100,1387-1395. 35. 36. 37.
Herman, I. R., and D’Amore, P. (1985) J. Cell Biol. 101,43-52. Wagner, R. C. (1980) Adu. Microcirc. 9,45-75. Brouwer, A., Wiesse, E., and Knook, D. L. (1987) in The Liver: Biology and Pathobiology (Arias, I. M., Jacoby, W. B., Popper,
Received November lo,1989 Revised version received April
11, 1990
H., Schacter, D., and Shafritz, D. A., Eds.), 2nd ed., pp. 665-682. Raven Press, New York. Katsuragawa, Y., Yanagisawa, M., Inoue, A., and Masaki, T. (1989) Eur. J. Biochem. 184,611-616. Savion, N., Vlodavsky, I., Greenburg, G., and Gospadorowicz, D. (1982) J. Cell Physiol. 110, 129-141.
M. K. K., and Lacey,
C.
I. (1985)
CELL
RESEARCH
190,
1
10 (19%))
Nonmuscle and Smooth Muscle Myosin lsoforms in Bovine Endothelial Cells ANNA C. RORRIONE,* ANNA MARIA C. ZANELLATO,* LUCA GIURIATO,* GIANLIJICI SCANNAPIEW,~ PAOLO PAULETTO,$ AND SAVERIO SARTORE*+’ *Institute
of General Pathology, $Institute
of Clinical Medicine, University Muscle Biology and Physiopatholugy,
A panel of monoclonal antibodies, specific for human platelet (NM-AS, NM-F6, and NM-G2) and for bovine smooth muscle (SM-E7) myosin heavy chains (MHC), were used to study the composition and the distribution of myosin isoforms in bovine endothelial cells (EC), in Using indirect and double immunofluvivo and in vitro. orescence techniques, we have found that in the intact aortic endothelium there is expression of nonmuscle MHC (NM-MHC), exclusively. By contrast, hepatic sinusoidal endothelium as well as cultured bovine aortic EC (BAEC) in the subconfluent phase of growth show coexistence of NM- and smooth muscle MHC (SM-MHC) isoforms. SM myosin immunoreactivity disappears when cultured BAEC become confluent. In this phase of cell growth, NM-MHC isoforms are localized differently within the cells, i.e., in the cytoplasm around the nucleus or in the cortical, submembranous region of EC cytoplasm. A third type of intracellular distribution of NM-MHC immunoreactivity was evident in the cell periphery of binucleated, confluent BAEC. These data indicate that (1) several myosin isoforms are differently distributed in bovine endothelia; and (2) SM myosin expression and the specific subcellular localization of NM myosin isoforms within EC might be regulated by cellcell interactions. 2 1990 Academic Press, Inc.
INTRODUCTION The cytoskeleton plays a crucial role in controlling cell movement and migration as well as in the maintenance of the cell shape and in the anchorage of cells to each other and to the underlying extracellular matrix 11-S]. It is assumed that the contractile force which is necessary for cell motility is provided by an actomyosin system. Actin and myosin [7-lo], as well as accessory contractile proteins (such as a-actinin and tropomyosin) have been demonstrated in NM cells both in vitro 1 To whom reprint requests should he addressed at Institute of General Pathology, IJniversity of Padova, Via l’rieste, ‘i5, 35131 Padova, Italy.
of Padova, and $National
Research !‘owwil
1rnit for
Padova, Ita1.v
and in situ [ 11-141. Recent studies have shown that the bundles of microfilaments which give rise LO the stress fiber system have probably a structural role rather than an involvement in cell motility 131. For example, in the vascular endothelium, stress fibers 115, 161 have been observed predominantly in those EC subject to hemodynamic forces: they may apply tension to resist the shear forces acting on the cells, allowing the-m to maintain their flattened shape and to remain firmly attached to the substratum [ 11, 16,171. In the cell periphery of confluent EC grown in vitro, microfiiaments can arrange differently compared with the stress fibers, giving rise to a distinct, F-actin-containing, microfilamentous structure, the so-called dense peripheral bands (DPB, Ref. [18]). The same authors [ 181 have found that a-actinin, myosin, and tropomyosin cc-localize with F-actin, suggesting that DPB may have the capacity to contract. The presence of DPR might be associat,ed with the ability of EC to form and maintain the cell monolayer. It is not known whether the peculiar organization of cytoskeletal apparatus in EC is due to a different distribution of the same set of cytoskeletal proteins or to the existence of structural variants. As concerns myosin, the different isoforms found in the sarcomeric muscles have been correlated with specific enzymatic and functional patterns (reviewed in Ref. 1191). in NM cells, such as platelets, myosin seems to consist of two isoforms [20-221 differently distributed in the cytoplasm and in the soiubilized membranes. Similarly, in the brush border of intestinal cells the two different subsets of myosin are localized in different manners [23]. Recent studies carried out on the eukaryotic amoebae, Dictiostelium discoideum have shown surprisingly that locomotion does not require myosin; conversely, this protein plays a crucial role in cytokinesis and in developmental morphogenesis [24, 25]. However, it could also he possible that amoeboid iocomotion is more linked to a minor 117-kDa myosin-like component (discussed in Ref. [26]). The aim of this study was to evaluate the myosin isoform composition and distribution in intact endothelia of aorta and hepatic sinusoids as well as in cultured
BORRIONE
2
BAEC. Our results indicate: (1) the existence of a different pattern of NM and SM myosin isoforms in the two endothelia, and (2) that SM myosin expression and the different distribution of NM myosin isoforms in BAEC are dependent upon the migrating or resting state of the cells. MATERIALS
AND
ET AL.
MHC-
METHODS
Preparation of monoclonal anti-myosin antibodies. Monoclonal anti-SM myosin antibodies (SM-E7) were obtained in our laboratory using bovine aortic actomyosin as immunogen [27]. Monoclonal antihuman platelet myosin antibodies (NM-F6, NM-AS) were produced following the immunization protocol and the cell fusion procedure described for the NM-G2 antibody [27]. The selected hybridomas were cloned by limiting dilution, and then used as culture supernatants, ascitic fluid, or purified IgG. Anti-Von Willebrand factor was purchased from Behringwerke AG, Federal Republic of Germany. Phalloidin conjugated with tetramethylrhodamine isothiocyanate (RITC) was obtained from Molecular Probes, Inc., Oregon, U.S.A. Western blotting. Human and bovine platelets, bovine aorta, and cultured BAEC were used in these experiments. Pellets of platelets, fibroblasts, or BAEC as well as small fragments of bovine aorta were extracted for 3 min in boiling Laemmli’s sample buffer solution [28]. After centrifugation at 10,OOOg for 10 min, the supernatants were collected and stored at -70°C until use (within a few days). Myosin extracts were electrophoresed in 10 or 7.5% gels in the presence of sodium dodecyl sulfate (SDS) and then stained subsequently with Coomassie brilliant blue R. Western blotting was performed essentially according to the procedure described in Ref. [29], using rabbit anti-mouse IgG conjugated with horseradish peroxidase (Dako, Dakopatts a/s, Glostrup, Denmark) to reveal the anti-myosin antibodies bound to blotted antigens. Cell cultures. FCNL 8112 bovine fibroblasts were a generous gift of Dr. Bressan (Institute of Histology and Embryology, University of Padova). BAEC were obtained essentially as described in Ref. 1301. Cells were cultured at 37’ (5% CO, atmosphere) in Dulbecco’s moclfied Eagle’s medium supplemented with 10% fetal calf serum (FCS). For Western blotting, cell monolayers were removed from plastic petri dishes by treatment of cultures of fibroblasts and BAEC with
FIG. 2. Specificity of anti-NM myosin antibodies as determined by Western blotting using NM tissue extracts. Crude antigens from human platelets (lane l), bovine platelets (lane 2), cultured bovine FCNL 81/2 fibroblasts (lane 3), and BAEC (lane 4) were electrophoresed in 7.5% SDS slab gels and blotted onto nitrocellulose paper. NM-AS (B), NM-F6 (C), and NM-G2 (D) antibodies were reacted with blotted NM tissue antigens. A, Coomassie blue staining of parallel gel. MHC, myosin heavy chains.
0.1% trypsin in phosphate-buffered saline (PBS). The activity of trypsin was blocked by adding FCS. Cells were collectedby centrifugation at 1OOOgfor 10 min and then washed twice with PBS. For immunofluorescence experiments, cells were cultured on glass coverslips and fixed at sub- and confluent stages of growth in cold acetone (-20°C) for 5 min. Zmmunofluorescence. Indirect and double immunofluorescence assays were performed on fresh frozen sections (lo-pm thick) of bovine aorta and on acetone-fixed cell cultures, according to the procedure described by Borrione et al. [31]. Briefly, after incubation with the appropriate dilutions of monoclonal antibodies (ascitic fluid or purified IgG) in 1% bovine serum albumine (BSA) in PBS, pH 7.2, for 30 min at 37°C in a humidified chamber, sections and cell monolayers were washed with PBS and then incubated with rabbit anti-mouse IgG conjugated with fluorescein isothiocyanate (FITC) or RITC (Dako). For the double-labeling procedure, one monoclonal antibody was conjugated with FITC and the other was indirectly revealed by rhodaminated anti-mouse IgG [31]. After fixation in 1.5%p-formaldehyde in PBS for 10 min, cryosections or cell cultures were mounted in Elvanol and anti-myosin binding patterns visualized using a Zeiss Axioplan microscope equipped with an epifluorescence apparatus and HBO 100-W high-pressure light. The following controls were performed: omission of the first or the second antibody, or use of an unrelated antibody instead of the first one.
RESULTS FIG. 1.
Western blotting of SM and NM myosin extracts with anti-myosin antibodies. Crude antigens from bovine aortic SM (lanes l,l’, and 1”) and from human platelets (lanes 2,2’, and 2”) were electrophoresed in 10% SDS slab gels and then blotted onto nitrocellulose paper. A, Coomassie blue staining of a parallel gel; B and C, Western blotting with NM-AS (B) and NM-F6 (C) antibodies. MHC, myosin heavy chains.
Characterization
of Anti-Myosin
Antibodies
Monoclonal SM-E7 is specific for myosin heavy chain 1 (MHC-1, M, 205 kDa) and 2 (MHC-2, M, 200 kDa) isoforms of SM-type [27]. This antibody can stain spe-
MYOSIN
ISOFORMS
IN ENDOTHELIAL
FIG. 3. Double immunofluorescence assay with anti-NM myosin wall c01itai ining a thin, continuous layer of cells (arrows) positive with and E) dirl ectly labeled with FITC and with NM-F6 (D), NM-G2 (C), Note th Iat: (1) NM-G2 antibody stains a large number of cells present present in the subendothelial region (F). Bar, 70 km.
CELLS
antibodies on intact endothelium of bovine aorta. Cryosec :tion:is ofac rtic anti-Von Willebrand factor antibodies were treated with NM -A9 (Pb J$ or SM-E7 (F) indirectly stained by anti-mouse IgG coup1 edw ith RI TC. in the elastic regions of aortic media (C), and (2) very rap ‘e SPVI cells are
4
BORRIONE
cifically the bovine vascular and nonvascular SM tissues (not shown). Characterization of anti-human platelet myosin antibodies (NM-AS, NM-F6, and NM-G2) is presented in Figs. 1 and 2. Western blotting analysis performed using 10% SDS gels shows that NM-AS and NM-F6 antibodies can bind to a 200-k.Da component present in the human platelet extract (Figs. 1B and 1C) which corresponds to the MHC (data not shown). The very weak reactivity with a bovine aortic extract is likely to be due to the presence of trace amounts of NM myosin [29, 321. NM-G2 antibody is specific for the heavy chains of human platelet myosin (Ref. [27]; see also Fig. 2D). In immunofluorescence tests, this antibody is able to recognize the EC of bovine pulmonary capillaries and the interstitial cells of renal parenchyma (not shown). Specificity of the three monoclonals (NM-AS, NM-F6, and NM-G2) for MHC from different bovine NM cells was determined by Western blotting with 7.5% SDS gels (Fig. 2). While all the three anti-myosin antibodies can react with human platelet MHC (lane l), two of them (NM-AS and NM-F6) are able to react with extracts from bovine platelets, FCNL 81/2 fibroblasts, and BAEC. NM-G2 displays a very weak immunoreac-
ET AL.
tivity with these NM antigens. The distinct binding properties of NM-AS and NM-F6 on the one hand and NM-G2 on the other were also evident in the intact aortic endothelium (see also Fig. 3) and in cultured FCNL 8 bovine fibroblasts (not shown). Immunolocalization
of Myosin Antigens in the Bovine EC
NM myosin antigenicity in aortic endothelium was analyzed by double immunofluorescence assay and the results of these experiments are shown in Fig. 3. NM-AS and NM-F6 antibodies stained brightly an aortic structure (Figs. 3A, 3B, 3D, and 3E) which was also recognized by anti-Von Willebrand factor antibody (not shown). In contrast to this immunofluorescence pattern, NM-G2 antibody was not able to label the aortic endothelium (Fig. 3C), but gives a strong reactivity with a population of SM cells which are mainly localized within the elastic fibers of aortic SM (Fig. 3C; in preparation). SM-E7 anti-SM myosin antibody do not label the intact endothelium of bovine aorta (Fig. 3F). In the hepatic parenchyma, NM-F6 immunoreactivity is distributed according to two different patterns (Fig. 4A), i.e., a fluorescence localized among the he-
FIG. 4. Double immunofluorescence on cryosections from the bovine liver with NM-F6 indirectly labeled with RITC (A) and SM-Ei’, directly labeled with FITC (B). Two distinct patterns of immunofluorescence can be seen in A, i.e., a honeycomb-like structure (asterisks) and a lilamentous material running at the level of hepatic sinusoids (arrows). Note that the sinusoidal fluorescence is also stained by SM-E’7 (B). Bar, 70 Km.
MYOSIN
ISOFORMS IN ENL)O’~HEI>IAI,
patic plates or in the cell periphery of hepatocytes. On the basis of desminivimentin and phalloidin immunoreactivity, as well as on morphological criteria (in preparation), the two hepatic structures recognized by anti-NM myosin antibodies have been identified as the sinusoidal endothelium and the bile canaliculi, respectively. NM-AS gives the same immunofluorescence pattern of NM-F6 whereas NM-G2 is completely negative with the hepatic parenchyma (not shown). In contrast to the bile canaliculi, the sinusoidal endothelium appears to be double labeled with SM-E7 anti-SM myosin antibody (Fig. 4R). Distribution of myosin isoforms was investigated in Ljitro using subconfluent and confluent cultures of BAEC (Figs. 5-7). These cells, which were positive with anti-Von Willebrand factor antibody (not shown), showed a fibroblast-like morphology in the subconfluent stage. In this phase, BAEC displayed a diffuse staining of cytoplasm with NM-AS (Fig. 5A) with almost no evidence for an organization in bundles of filaments, both at the cell periphery and in the central area of EC. Conversely, NM-G2 was completely unreactive (Fig. 5B) with these cells in this phase of growth. When cells reached confluency, they achieved a characteristic polygonal morphology typical of epithelial cells and EC, with reorganization into a single-cell monolayer composed of closely apposed and nonoverlapping cells. Among the mass of mononucleated cells, the presence of some binucleated cells which are reactive with anti-Von Willebrand factor antibody (not shown) was evident. Under these conditions, NM-AS showed a staining pattern similar to that found in subconfluent cells with the submembranous, cytoplasmic region almost devoid of any fluorescence (Fig. 5C). Interestingly, NM-G2 was found to be negative with the large majority of BAEC except for the larger, binucleate cells in which it was seen at the submembranous level often as short segments, rarely along the entire membrane and never across the cell (Figs. 5D and 5F). The filamentous structure recognized by NM-G2 was also labeled by the toxin phalloidine (Fig. 5E) which binds with high affinity to F-actin. The majority of the cells, which were negative with NM-G2 antibody, displayed a well-defined stress fiber labeling with phalloidine (not shown). We have also performed double-labeling experiments on BAEC monolayers utilizing NM-AS and NM-F6 antibodies (Fig. 6). In addition to giving a diffuse immunofluorescence pattern in the cytoplasm, NM-F6 antibody decorates the cytoplasmic filaments corresponding to the stress fiber system (Fig. 6B). Local differences in the distribution of the immunofluorescence pattern with the two monoclonal antibodies are visible in the cortical region of BAEC cytoplasm (Figs. 6A and 6B). Interestingly, distribution of myosin immunoreactivity was quite different in the sub- and confluent phases of BAEC growth. In cultures of BAEC close to confluency
5
CELLS
or at contluency, NM-AS binding was localized mainly in the cytoplasm around the nucleus with the corticai region, beneath the plasma membrane, almost negative with this antibody (Fig. 6C). This cellular region, possibly corresponding to the area in which DPB have been reported [18], is, however, strongly labeied by NM-F6 (Fig. 6D). Heterogeneity in the cytoplasmic reactivity with NM-F6 was occasionally evident in confluent cultures (Fig. 6D). Cells which were negative or weakly stained with NM-F6 antibody were found to be always positive with NM-AS (Fig. 6C). The potential presence of a SM myosin isoform in BAEC has been assayed utilizing the SM-E7 monoclonal anti-SM myosin antibody j27j. Double immunofluorescence tests carried out using NM-AS and SM-E7 (Fig. 7) in subconfluent and confluent cuitures show that SM-E7 is able to stain the subconfluent BAEC exclusively (compare Figs. 7B and 7D). But, more interestingly, SM-myosin isoform is distributed in a heterogeneous manner among and within BAEC. While NM antibody labels the whole cytoplasm around the nucleus (Fig. 7A), SM-E7 antibody staining is restricted to part of cytoplasm surrounding the nucleus (Fig. I;IB), giving a punctuate pattern of fluorescence (Fig. 7B: see also the inset). DISCUSSION
Taking advantage of a panel of monoclonal anti-myosin antibodies, we have found that bovine endothelia are heterogeneous with respect to myosin isoform composition and distribution. The two intact endothelia examined in this study show a different myosin content; i.e., while in the sinusoidal endothelium both SM and NM myosin isoform coexist, in the aortic endot.helium SM myosin immunoreactivity is absent. However, concomitant expression of SM and NM myosin isoforms can be induced in cultured subconfluent BAEC. In vivo and in vitro coexistence of SM and NM myosin isoforms in EC bears some resemblance to developing and cultured vascular SM cells [27, 331 as well as to the microvascular pericytes [34, 351. Our data may be indicative of a special contractile function of hepatic sinusoidal endothelium (“SM cell-like” EC?) which could be related to a specific embryological origin [36] and/or to some peculiar structural and functional characteristics [37] of this tissue. Since SM myosin antigenicity is distributed according to a punctuate pattern in the BAEC cytoplasm, it is likely that this isoform is not a component of the microfilament system. Herman et al. 131, working on embryonic chick and HeLa cells, have suggested that the large actomyosin filament bundles are associated with nonmotile cells, whereas diffuse actin and myosin distribution can be related to cells in active movement. The functional significance of a SM myosin isoform in the
Fl 1GI. 5. Double immunofluorescence with NM-AS, NM-G2 antibl odies, and phalloidin on subconfluent (A, B) and confluent (C- -F) OfB Al XC. Double immunofluorescence assays with NM-AS coupled d irectly with FITC (A, C) and NM-G2 indirectly stained with R .IT( InE a1nd F giant cells double-labeled (arrow) with rhodaminated phal loidin (E) and NM-G2 (F) indirectly stained with anti-mouse 1W with Ii LITC are visible. Arrow in D indicates the cytoplasmic periphe !ral structure recognized by NM-G2 in binucleate cells. Bar, 12
:ult1 nes (B, D). :oq Iled m.
MYQSIN
ISQFORMS
with NM-AS and NM-F6 FIG. 6. Double immunofluorescence The exI ,eri .ment was performed using NM-AS (A, C) directly labeled distinct .flU lorescence patterns can be observed in the cell periphery of NM-AS and NM-F6 immunostaining intracel !lul ar distribution NM-F6 (arstterisk in C and D). Bar, 12 pm.
IN ENDOTHELIAL
CELLS
antibodies on subconfluent with FITC and NM-F6 (B, of subconfluent cells (arrow (arrow in C and D). Some
(A, B) and confluent (C, D) culture SO fBP dx. D) indirectly stained with 53%. O(:ca sion ally, in A and B). Confluent cultures sb OR7disl :inct cells positive w&h N gal ive with
EVG. 7. Double immunofluorescence assay with anti-NM and anti-SM myosin antibodies on subconfluent (A, B) and confluent (C, D) cul ltures of BAEC. NM-AS directly labeled with FITC (A, C) and SM-E7 indirectly stained with RITC (B, D). In subconfluent cultures, SM-E7 the cytoplasm of BAEC, giving a punctuate pattern of fluorescence in SM-positive cells (arrow; see also inset fc )r major sta iins heterogenously de1Lails). Some cells, positive with NM-AS, are negative or weakly reactive with SM-E7 (star). In other cells, part of the cytoplasm is I iegative wit ;h SM-E7 (asterisk). Confluent cultures are negative with SM-E7 (D). Bars: A-D, 12 pm; inset, 6 pm. a
MYOSIN
ISOFORMS
IN ENDOTHELIAL
BAEC is not clear, but the fact that isoform is not expressed in intact endothelium and in confluent cultures of EC would suggest that SM myosin may be a marker for proliferating/migrating EC. The second important finding reported in this paper concerns the existence of three NM myosin isoforms and their different subcellular localization in cultured aortic EC. In confluent cultures of BAEC one isoform is mainly localized to the cortical cytoplasm and the other is more linked to the cytoplasmic region surrounding the nucleus. These findings are in agreement with (1) the data of Groschel-Stewart et al. [20-221, which indicate the existence of structural variants of NM myosin in membrane and cytoplasm of NM cells, and (2) genetic experiments w&h indicate the presence of two distinct NM-MHC mRNAs in different chicken tissues [38]. A third NM-MHC variant can be identified in binucleate, confluent BAEC at the level of peripheral microfilament bundles. These particular giant-type EC might correspond to dividing cells or to permanent binucleate cells. Confluent EC cultures from adult bovine aorta show a marked reduction of in i3H] thymidine incorporation compared with subconfluent cultures [39]; thus, it is more likely that these giant cells are EC in which nuclear duplication was not followed by cytokinesis. Similar giant cells have been found in human EC from adult thoracic aorta [40] and umbelical vein [41]. The function of this NM myosin isoform is presently unknown. One might speculate that this isoform is more directly involved in the process of cytokinesis of EC and that the perturbation of this event may determine the appearance of this myosin isoform at the submembranous level. There is some evidence which indicates that myosin differently in vascular EC in situ [ 1,5] and in vitro [ 18, 421. In fact, cytoskeletal myosin can be found in both DI’B and in the central microfilamentous bundles where, in association with actin and other accessory proteins [I$], it may be involved in contractile functions. Using a polyclonal anti-SM myosin antibody, Hormia et al. [43] and Gottlieb and co-workers [l&42] have also reported that myosin is present in two structurally distinct areas of EC. We have now identified a protein marker which permits one to distinguish the central and peripheral microfilamentous region. The specific distribution of NM-MHC isoforms within EC is to be achieved by cell-cell contact since in subent cultures of BAEC there is only limited evidence for this process. It is interesting that the NM myosin isoform present in the periphery of EC cytoplasm is also present at the level of the junctional eomplex in the bile canalieuli of bovine liver (Fig. 4; in preparation). Thus, it might be possible that this isoform is better suited for stabilizing cortical microfilament associations which migbt be responsible for cell-cell ad-
CELLS
9
hesion and for maintaining layer Cdl].
the integrity
of
We thank Dr. Elisabetta Deiana, ““M. Negri” Institute, Milan, Italy, for preparing BAEC and Mr. Maurizio MoretLo for technical assistance. EFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17.
18. 19. 20. 21.
22. 23. 24. 25. 26. 27. 28. 29.
Drenckhahn, D., and Wagner, J. (1986) J. @a11Viol. 1102, 738-1747. Spooner, B. S., Yamada, K. M., and Wessels, N. K. (1971) 9: Cell Biol. 49, 595-613. Herman, I. M., Crisona, N. J., and Pollard. T. D. (1981) J. Cell Biol. 90, 84-91. Korn, E. D. (1978) Proc. NatE. Acad. Sei. USA 7 White, 6. E., Gimbrone, M. A., Jr., and Fujiwa Cell &al. 97, 416-424. White, 6. E., and Fujiwara, K. (1986) J. Gek! Biol. 103, 63-70. Becker, 6. G., and Murphy, 6. E. (1969) Amer. ,P. Pathol. 55, 6-37. Lazarides, E., and Weber, K. (1974) Proc. N&l. dcad. Sci. USA 71,2268-2272. Fujiwara, K., and Pollard, T. D. (1976) J. &II BioZ, 71,848-875. Weber, K., and Groschel-Stewart, U. (1974) Proc. N&E. Acad. Sci. USA 71,4561-4564. Byers, H. R., and Fujiwara, K. (1982) J. Celb Viol. 93, 804-811. Byers, H. R., White, G. E., and Fujiwara, K. (1984) in Cell and Muscle Motility (Sbay, J., Ed.), Vol. 5, pp. 83-137. Plenum, New York. Herman, J. M., Pollard, T. D., and Wong, A. 3. (1982) Ann. iv. Y. Acad. Sci. 401, X-60. Rogers, K. M., and Kalnins, V. I. (B983j L&l. Invest. 49, 650-654. Hammersen, F. (1980) Adv. Gabbiani, G., Gabbiani, F., (1983) Proc. Natl. Acad. Sci. Franke, R. P., Grlfee, WI., mayer, C., and Drenckhahn, D. (1984) Nature U!mndoni 307, 648-649. Wong, M. K. K., and Gottlieb, A. I. (1986) Rrteriosclerosis 212-219. Swynghedauw, B. (1986) Whys. Rev. Peleg, I., Kahane, I., Eldor, A., Grosc wart, U., Mestan, J., and Mublrad, A. (1983) J. Biol. Chen. Groschel-Stewart, U., Rakousky, C., Franke, R., Peleg, I., Kahane, I., Eldor, A., and Muhlrad, A. (1985) Cell Tissue IZes. 241, 399-404. Peleg, I., Kabane, I., Eldor, A., Groschel-St dano, S., and Muhlrad, A. (1989) Thrornb. Keller, T. C. S., Conzelman, K. A., Chasa M. S. (1985) 9. Biol. Chem. 100, 1647-1655. Knecht, D. A., and Loomis, W. F. (1987) Science 23 1081-1086. De Lozanne, A., and Spudicb, J. A. (19$7) Science 236, 108661091. Mooseker, M. S. (1985) Annu. Rev. &El Bioi. P, 26-293. Borrione, A. C., Zanellato, A. M. C., Scannapieco, G., Paulette, P., and Sartore, S. (1989) Eur. J. Biochem Laemmli, U. K. (1970) Nature (London) 2: Sartore, S., De Marzo, N., Borrione, A. C., Zanellato, A. M. C.,
10
30. 31. 32.
33. 34.
BORRIONE Saggin, L., Fabbri, L., and Schiaffino, S. (1988) Eur. J. Biochem. 179, 79-85. Balconi, G., and Dejana, E. (1986) Med. Biol. 64, 231-245. Borrione, A. C., Zanellato, A. M. C., Saggin, L., Mazzoli, M., Azzarello, G., and Sartore, S. (1988) Differentiation 38, 49-59. Gaylinn, B. D., Eddinger, T. J., Martino, P. A., Monical, P. L., Hunt, D. F., and Murphy, R. A. (1989) Amer. J. Physiol. 257 [Cell Physiol. 261, C997-C1004. Kawamoto, S., and Adelstein, R. S. (1986) J. Biol. Chem. 262, 7282-7286. Joyce, N. C., Haire, M. F., and Palade, G. E. (1985) J. Cell Biol.
ET AL.
38. 39. 40.
Antonov, A. S., Nikolaeva, M. A., Klueva, T. S., Romanaov, Y. A., Babaev, V. R., Bystrevskaya, V. B., Perov, N. A., Repin, V. S., and Smirnov, V. N. (1986) Atherosclerosis 59, 1-19.
41.
Shirinsky, V. P., Antonov, A. S., Birnkov, K. G., Sobolevsky, A. V., Romanov, Y. A., Kabaeva, N. V., Antonova, G. N., and Smirnov, V. N. (1989) J. Cell Biol. 109,331-339.
42.
Gottlieb, A. I., Spector, W., Wong, (1984) Atherosclerosis 4,91-96.
43.
Hormia, M., Badley, R. A., Lehto, B. P., and Virtanen, Exp. Cell Res. 15’7, 116-126.
100,1387-1395. 35. 36. 37.
Herman, I. R., and D’Amore, P. (1985) J. Cell Biol. 101,43-52. Wagner, R. C. (1980) Adu. Microcirc. 9,45-75. Brouwer, A., Wiesse, E., and Knook, D. L. (1987) in The Liver: Biology and Pathobiology (Arias, I. M., Jacoby, W. B., Popper,
Received November lo,1989 Revised version received April
11, 1990
H., Schacter, D., and Shafritz, D. A., Eds.), 2nd ed., pp. 665-682. Raven Press, New York. Katsuragawa, Y., Yanagisawa, M., Inoue, A., and Masaki, T. (1989) Eur. J. Biochem. 184,611-616. Savion, N., Vlodavsky, I., Greenburg, G., and Gospadorowicz, D. (1982) J. Cell Physiol. 110, 129-141.
M. K. K., and Lacey,
C.
I. (1985)
