Tobacco Documents Online

0 E c . i I C 11AWW It/ ElecVophoresis . 1-) (+l Electrophnresis tutive MLC20 phosphorylation. In addition, the amount of iso- metric tension would be expected to increase in proportion to the increase in phosphorylation that results from receptor acti- vation of inositol phospholipid hydrolysis (29). We found that histamine activation ofHU V EC inositol phospholipid hydroly- sis increased MLC20 phosphorylation 0.18 mol/mol MLC,, a 90% increase. Thrombin, which also activates HU V EC inositol phospholipid hydrolysis, increased HUVEC centripetal ten- sion from 0.65 to 1.3 x 105 dyn/cm2 in Kolodney and Wysol- merski's report (30). Similarly, Boswell et al. (3 1) and Morel et al. (32) have observed that unstimulated bovine pulmonary P-Sar : 0 n 0 ~ P-Thr ~ Peptide 2 Peptide I Control HUVEC Figure 8. Autoradio- , gram of phosphoamino acids from peptide 1 and peptide 2 from control HUVECs. See text for explanation. Figure 7. (A) Map of tryptic digests of MLCzo phosphorylated in situ in control HUVECs. Two peptides, peptide I and peptide 2 were observed. (B) When the digest used in A was mixed with digest from MLCp phosphorylated in vitro with MLCK as in Fig. 6 a, peptide 1 comigrates with monophosphorylated ser" peptide standard. (C) Map of tryptic digest of MLC,I, phosphorylated in situ in HUVECs exposed to histamine (10-' mol/liter) for 30 s. (D) When the digest used in A was mixed with the digest in C, peptide maps are super- imposed. E. Map of tryptic digests of MLC,, phosphorylated in situ in HUVECs exposed to 8-Br-cAMP (10-" mol/liter) for 10 min. F. When the digest used in E was mixed with di- gest from the monophosphorylated ser19 in vitro standard, as in Fig. 6 a, peptide I comi- grates with the monophosphorylated serl'. See text for explanation for maps generated under di8erent conditions. microvascular endothelial cells display constitutive wrinkling of silicone rubber matrix and further wrinkle when cells are exposed to vasoactive mediators. Therefore, the change in MLC,, phosphorylation and isometric tension is consistent with the hypothesis that activation of signal transduction may mediate an acute increase in isometric tension through the cyto- skeleton. In our experiments peak phosphorylation of MLC20 was observed 30 s after adding histamine. Kolodney and Wysol- merski (30) observed an increase in tension exerted by HU- VEC monolayers 30 s after adding thrombin. These relation- ships are consistent with MLC20 phosphorylation contributing to the initiation of increased tension development in HUVECS. However, Kolodney and Wysolmerski (30) found that peak tension developed after 5 min and persisted for ? 40 min after adding thrombin. We found that MLC,, phosphorylation had returned to basal levels by 5 min and remained there at 15 min. Maintaining tension even after MLC20 phosphorylation ap- proaches basal levels may be analogous to smooth muscle cells (29). Other effects of activation of the signal transduction cas- cade, such as recruitment of the actin cytoskeleton, may also contribute to the persistent tension (17 ). Similar to what Fox and Phillips (33) found in platelets, Carson et al. (34) recently found that histamine increased the fraction of actin that was present as filamentous actin in HUVECs. To explain the transient increase in MLCzo phosphoryla- tion in HUVECs exposed to histamine, we asked whether MLC,0 phosphorylation returns to baseline because cell cAMP levels increase. Hekimian et al. (35) recently demonstrated that histamine increased cell cAMP levels in bovine aortic en- dothelial cells through a Hz receptor. We did not observe any 1204 A. B. Moy, S. S. Shasby, B. D. Scott, and D. M. Shasby

(•) Electrophoresis (-) d 0 L a 0 u w w I (-) B Figure 6. Maps of tryptic digests of tur- key gizzard MLC phosphorylated in vitro by MLCK (a) or by rat brain PKC (b). See text for explanation of maps and phosphoamino acid analysis. this suggests that peptide I is a mixture of monophosphory- lated peptides that contained monophosphorylated ser19 and monophosphorylated thr1e peptides. Peptide 2 contains phos- phoserine and phosphothreonine. As noted above, peptide 2 has been previously demonstrated to be a diphosphorylated ser19 and thr's peptide (24). This phosphoamino acid analysis pattern is similar in histamine-stimulated cells (data not shown). The presence of phosphothreonine in peptide 1 sug- gest that, unlike smooth muscle cells, monophosphorylation at thr1e by MLCK significantly occurs in the HUVEC. Discussion The intent of these experiments was to further explain how inflammatory molecules such as histamine cause adjacent en- dothelial cells to retract from each other ( t-8 ) and how cAMP agonists oppose this response (2, 25). Although the actin cyto- skeleton of permeabilized endothelial cells contract after the addition of 100 µM calcium, ATP, MLCK, and calmodulin (10, 11, 13), very little contraction and MLC~ phosphoryla- tion occurs at 1-5 µM free calcium in these permeabilized cell models (13 ). The limited contraction of endothelial cell cyto- skeletons at 1-5 µM calcium is especially relevant, since cell calcium remains in the high nanomolar range in stimulated endothelial cells. Histamine increased calcium in HUVECs to ^- 650 nM in our earlier report, and this is consistent with other reports on histamine stimulation of HUVECs and the response of endothelial cell calcium to other inflammatory mol- ecules such as bradykinin (2, 26, 27). In the experiments described in this manuscript we found that doses of histamine that increased HUVEC monolayer per- meability to albumin by 50-100%, increased MLCzO phosphor- ylation by - 0.2 mol phosphate/mol MLC20 (approximately a 90% increase from baseline). Unstimulated HUVECs demon- strated constitutive phosphorylation. Histamine caused a rapid but transient increase in MLCzo phosphorylation and achieved maximal levels by 30 s. Maximal levels persisted at 90 s but returned to basal levels by 5 min. Of note is that the maximal levels of MLCzo phosphorylation in histamine-stimulated HU- VECs is consistent with observed maximal changes in cell cal- cium. We and others have observed that histamine-stimulated increases in calcium are maximal by 15-30 s in HUVECs (2, 28). The close temporal relationship between change in cell calcium and MLCzo phosphorylation in HUVE suggest a cal- cium dependence for MLCzo phosphorylation. It does not seem likely that the modest increase in MLC, phosphorylation we observed represents a methodological problem. Exposing the cells to 5 µM ionomycin produced an appropriately greater increase in MLCze phosphorylation (0.5 mol phosphate/mol MLCp), and one that is similar to that observed in smooth muscle cells exposed to 10 pM ionomycin forthe same time (0.6 mol phosphate/mol MLC20, 90 s) (16). Similarly, the levels of phosphorylation achieved in stimulated smooth muscle cells at time points earlier than 30 s were pre- served at 30 s, so it is not likely that we missed higher levels of phosphorylation at earlier time points (16). The increase in cell calcium observed in HUVECs stimulated with histamine is also preserved at these times (26). Since the level of MLC, phosphorylation should reflect the amount of isometric tension exerted on the underlying extra- cellular matrix, one would predict that unstimulated HUVECs should exert resting isometric tension since they display consti- Histamine and cAMP Atter Myosin Phosphorylation in Endothelial Cells 1203

Control cAMP cAMP + Histamine Figure 2. (A) Autoradiograph of immunoprecipitated MLC isoforms from control "S-labeled HU V ECs. (B) Autoradiograph of immunopre- cipitated MLC isoforms from "S-la- beled HUVECs exposed to hista- mine (10-s mol/liter X 30 s). (C) Autoradiograph of immunoprecipi- tated MLC isoforms from "S-la- beled HUVECs exposed to 86r- cAMP (10'0 mol/liier X 10 min). (D) Autoradiograph of immunopre- cipitated MLC isoforms from J5S-la- beled HU VECs pitiexposed to 8Br- cAMP (10-4 mol/liter X 10 min) and then exposed to histamine (10-s mol /liter) for 30 s. See text for ex- planation. sponded to the unphosphorylated isoforms A and B (data not shown). The identity of the mono- and diphosphorylated iso- forms were confirmed by separation of the isoforms from cells labeled with 32P (Fig. I B). Quantitating phosphorylation of myosin light chain The effect of histamine on MLC20 phosphorylation. Histamine (10-5 M) induced.a rapid but transient increase in MLCp phosphorylation. This was evident as a.relative increase in the intensity of MLCzo that migrated as mono- and diphosphory- lated isoforms (Fig. 2, A and B). In unstimulated cells, the intensity of unphosphorylated MLCzo isoforms (A and B) greatly exceeds the corresponding intensity of the monophos- phorylated isoforms (A' and B'). In contrast, in HUVECs ex- posed to histamine, the intensity of the unphosphorylated MLC20 isoforms is almost equal to the intensity of the corre- sponding monophosphorylated isoforms. This indicates that increased MLCa,o phosphorylation occurs when HUVECs are exposed to histamine. Expressed as stoichiometry, histamine induced a modest increase in the number of mol phosphate/mol MLCzO in HU- VECs (Fig. 3). 30 s after exposure to histamine MLCzn phos- phorylation was 0.16±0.05 (isoform B) and 0.19±0.05 (iso- form A) mol phosphate/mol light chain greater in histamine- exposed cells than in control cells. This represented the peak increase in histamine-stimulated phosphorylation. The in- creased phosphorylation of MLCzo persisted through 90 s when MLCZO phosphorylation in histamine-exposed cells was 0.16±0.04 (isoform B) and 0.17±0.04 (isoform A) mol phos- phate/mol light chain greater than it was in control cells. How- ever, by 5 min MLCzo phosphorylation in histamine-exposed cells was no longer different than it was in control cells, and it remained at the control level 15 min after histamine exposure. lonomycin produced a much greater increase in MLCo phosphorylation than did histamine (Fig. 4). 90 s after initiat- ing exposure to 5 X 10-6 M ionomycin MLCio phosphoryla- (p <0.05) y- IsolartnB ~ ItolartnA / Biaeform Figure 3. This figure demonstrates the effect of histamine on MLC,S phosphorylation over time. Data are expressed as a change in stoichi- ometry (mol P/mol MLCzS) from control HUVECs. Stoichiometric change is reported for control HUVECs and HUVECs exposed to histamine for 30 s, 90 s, 5 min, and 15 min. Histamine ( l0-' mol/ liter) increased the amount of MLCzS phosphorylation at 30 and 90 s(P < 0.05). n a 15 for each time point and condition. 2 6 oaJ . ® ~ Aisoform group Figure 4. MLCm phosphorylation in control HUVECs and HUVECs treated with 5 µM ionomycin (90 s). lonomycin increased MLCza phosphorylation by 0.5 mol phosphate/mol MLC, (P < 0.05, n = 6 for each). Histamine and cAMP Alter Mlrosin Phosphorylalion in Endothelial Cells 1201

tion in HUVECs increased 0.51±0.06 (isoform B) and 0.50±0.04 (isofornt A) mol phosphate/mol light chain (n = 4 control and 6 ionomycin monolayers). . The ejfect of increasing intracellular cAMP on MLCOphos- phorylation. Increasing cAMP in HUVECs decreased MLC~ phosphorylation and prevented histamine from increasing phosphorylation above levels found in control cells. cAMP alone shifted the isoform distribution towatd the unphosphor ylated isoforms and, in histamine-eXposed cells, cAMP pre- treatment prevented the expected shift towards the mono- and diphosphorylated isoforms (Figs. 2, C and D). This shift in the isoforni distribution represented a decrease in MLCm phos- phorylation of 0.19±0.02 (isoform B) and 0:17±0.02 (isoform A) mol phosphate/mol light chain in cAMP-treated cells (Fig. 5). Similarly, cells treated with cAMP and then histamine had 0.09±0.02 (isoform B) and 0.08±0.02 (isoform A) fewer mol phosphate/mol light chain than did control cells (Fig. 5). The effect of hi.stamine on cellular cAMP. The rapid return of MLC phosphorylation to control levels would be enhanced if histamine increased cellular cAMP as well as increasing cell calcium. Basal cAMP levels (2.9±0.89 pmol/4.9 cm2 of con- fluent cells) did not change at 30 s(2.9±0.40 pmol/4.9 cm2), 90 s (3.9±0.84 pmol/4.9 cm2), or 5 min (2.3±0.07, pmol/4.9 em') after exposure to histamine. However, forskolin (2 X 10-5 M) and aminophyline (10-' M) increased HUVEC cAMP (18.8±2.6 pmol/4.9 Cm2) 5 min after exposure. Determining the kinase that mediates MLC20 phosphorylation Two-dimensional TLC peptide mapping. Histamine increases both calcium and DAG in HUVECs. Conceivably, histamine could initiate MLC~ phosphorylation by MLCK (calcium de- pendent) and/or by PKC (calcium and DAG dependent). To directly determine which kinase phosphorylated MLCZn in HUVECs exposed to histamine we compared maps of tryptic digests of MLCzn standards phosphorylated in vitro by MLCK and by PKC to peptide maps of tryptic digests of MLC~O phos- phorylated in situ in control and histamine-exposed HUVECs. control (p <0.05) cAMP cAMP-Hisl . . group . .. Figure 5. Number of moles of phosphate per mole. of i4ILC,(mol P/mol MLCzo) in control HUVECs (Con), in HUVECs exposed to 8-Br-cAMP (10-° mol/ liter) for 10 min (cAMP),and in HUVECs . exposed to 8-Br-cAMP (10-4 mol/liter for 10 min) aod then hista- mine (10-' mol/liter) for 30 s. 8-Br-cAMP alone decreased the amount of phosphorylation from controlcells. (P. < 0.05). Pretreat- ment with 8-Br-cAMP prevented an increase in-MLC~ phosphoryla- tion when cells were exposed to histamine. n? 15 for each condition. ln vitro standards were generated by phosphorylating tur- key gizzard sniooth muscle MLC or human platelet MLC with smooth muscle MLCK or rat brain PKC (each kinase dis- played the same map whether smooth muscle or nonmuscle MLC was used). Fig. 6 shows the two-dimensional TLC pep- tide map of tryptic fragments from turkey gizzard MLC20 phos- phorylated in vitro by MLCK and PKC. The map of-the MLCK standard (Fig. 6 a) consisted of a single dominant phosphorylated peptide fragment (peptide A), which con- tained only phosphoserine. The migration of this fragment was consistent with that of a fragment previously identified as con- taining phosphorylated ser19 (14). The map of the PKC stan- dard (Fig. 6 b) demonstrated three peptides. Peptide B, a minor peptide, contained only phosphoserine. Peptide C, a major peptide, also contained only phosphoserine. The migration of these fragments is consistent with that of fragments.previously identified as containing monophosphorylated ser' or serz (14 ). Peptide I) is a major peptide that contains only phos- phothreonine, and its migration is consistent with previous re- ports of a fragment containing monophosphorylated thr9 (14), Tryptic fragments of MLC,o phosphorylated in situ in HU- VECsdemonstmted a phosphorylated tryptic peptide (peptide 1, Fig. 7 A) similar to the one obtained with the MLCK stan- dard. The peptide map of digest from control HUVECs also demonstrated another phosphorylated tryptic peptide (peptide 2, Fig.. 7 A): Peptide I migrated to a similar position as the monophosphorylated ser19 in vitro standard. Peptide 2 was po- sitioned near the origin, a migration pattern previously identi- fied for a tryptic fragment diphosphorylated on thr's and ser19 (24). This pattern is unaltered even if repeated digestion is extended for 48 h. When digest from control cells was mixed with digest from the MLCK standard and separated by the same two-dimensional TLC procedures; peptide 1 comigrated with the monophosphorylated ser19 standard (Fig. 7 B), con- firming that peptide I represented a monophosphorylated pep- tide phosphorylated in situ by MLCK. Maps of MLC2e tryptic fragments from HUVECs exposed to histamine demonstrated the same two peptides as control cells (Fig. 7 C). When digests from control cells and histamine-stimulated cells were mixed, the maps showed comigration of both peptides, indicating that MLCK mediates both basal and histamine-stilmulated phos- phorylation (Fig. 7 D). Monophosphorylated ser'°a peptides were not observed under resting and histamine-stimulated con- ditions. This indicatesthat PKC does not mediate phosphory- lation in HUVECs under basal and histamine-stimulated con- ditions. In addition, phorbol dibutyrate PDBU (10-6 M) -ex- posed cells did not generate significant ser" fragments (data not shown). The peptide map pattern did not change in cells exposed to cAMPagonists (Fig. 7 E). Two-dimensional TLC maps of digests of cAMP-stimulated cells revealed a peptide (peptide 1) that comigrated with the monop}iosphorylated ser'9 standard (Fig. 7 F). Since cAMP decreased phosphorylation, and the peptide map was unaltered, cAMP decreased phosphorylation by decreasing MLCK activity and/or increasing dephosphory- lation of MLC20. Phosphoamino acid analysis of in situ phosphorylated pep- tides. When peptides from two-dimensional TLC maps of un- stimulated cells are eluted from the silica TLC plates and sub- jected to phosphoamino acid analysis, peptide 1 contained phosphoserine and phosphothreonine (Fig. 8). Since peptide 1 also comigrates with the monophosphorylated ser" standard, 1202 A. B. Moy S. S. Sha.sby. B. D. Scott, and D. M Shasby

monitored by the movement of acid fuschin dye toward the anode. Plates were then dried and subjected to ascending chromatography using 1-butanol/acetic acid/pyridine/water (49:8:38:30). Plates were then dried, wrapped in cellophane, and autoradiographs were devel- oped using Kodak X-Omat AR film Phosphoamino acid analysis. Peptides that were generated from two-dimensional TLC maps were scraped from the TLC plate and eluted-with 50 mM NH4HCO3. The peptide was then lyophilized, washed with water, and lyophilized again to remove the ammonium birarbonate. Peptides were subjected to acid hydrolysis in 6 N HCl at 120°C for 3 h and then relyophilized. Dry residue was washed with water and relyophilized. Labeled phosphoamino acids were resus- pended in electrophoresis buffer (acetic acid/formic acid/water, 9:3:88), which contained 2 µg/pl of phosphothreonine and phospho- serine, and were then spotted onto a TLC plate. Amino acids were subjected to electrophoresis at 700 V for 2 h and 50 min at 4°C. La- beled phosphoamino acids were detected by autoradiography and iden- tified by comparing the corresponding migration of the phosphoamino acid standards, which were stained with ninhydrin. Measurement ofintracelltdmcAMP. HUVECs plated on polycar- bonated filters were exposed to control media (M199 without serum), histamine (10-' M), or cAMP mixture (forskolin [2 X 10-' M] and theophylline [ I0-' M]). Afterthe indicated time intervals, the medium was aspirated and the cells were extracted twice with I ml of ethanol. The extracts were combined and centrifuged at 3,000 g for 15 min at 4°C. The supernates were dried under nitrogen and reconstituted in the sample buffer provided (cAMP'H Assay System; Amersham Corp., Arlington Heights, IL). The rest of the analysis system was conducted as described in the directions for the assay system. In vitro phosphorylation assays. Standards for tryptic peptide maps of MLCio were generated in vitro from turkey gizzard MLCm or plate- let MLCzo phosphorylated with either smooth muscle MLCK or rat brain PKC. Phosphorylation of MLC by MLCK was carried out in a volume of 200 µl using 20 mM Tris-HCl (pH 7.6 ), 5 mM MgCh, 200 µM CaClr, 1 mM [y-uP]ATP (400-1,000 cpm/pmol), 2 pM calmo- dulin, 7 µg/mi smooth muscle MLCK, and 3.25 mg/ ml turkey gizzard smooth muscle MLCro or 50 µg/ml human platelet MLCio. The reac- tion was initiated with ATP at 25°C for 10 min and terminated by precipitating the assay mixture with 40 µl 100% TCA. Samples were microcentrifuged, and the precipitate was washed three times with I ml of 10% TCA. The pellet was resuspended with 200 pl of SDS-sample buffer and 0.1'R, bromphenol blue and made alkaline with I N NaOH. Phosphorylation of MLCp by PKC was carried out in a final vol- ume of 400 µl using 20 mM Tris-HCI (pH 7.6 ), 5 mM MgClz, 200 µM CaC7z, 1 mM [Y-"PIATP (400-1,000 cpm/pmol), 50 µg/ml phos- phatidylserine, 0.8 pg/ml 1,3 diolein, and 2.7 µg/ml PKC at 30°C for 60 min. Phospholipids were stored in chloroform and evaporated under a constant stream of nitrogen. Phospholipids were then resus- pended in Tris buffer and sonicated at 4`C for 5 min using a sonicator (model 200; Branson Ultrasonics Corp., Danbury, CT). The reaction was initiated with the labeled ATP. Calculating the stoichiometry ofMLC~ phosphorylation. Quantita- tive phosphorylation of MLCio was determined using laser densitome- try of autoradiographs made from the two-dimensional gels of 35S-la- 35S-Ia6elled beled MLCm immunoprecipitated from the cells. There were two phos- phorylated isoforms of MLCp (designated A and B) and each phosphorylated isoform could exist as an unphosphorylated (A or B), monophosphorylated (A' or B'), or diphosphorylated (A" or B') iso- form. The volume (V) of each phosphorylation state of each isoform was integrated on the densitometer. The volume of each phosphoryla- lion state for each of the isoforms was then expressed as a fraction ( fxn) of the total radioactivity for that isoform, for example, fraction A = V(A)/ V(A) + V (A ') + V(A"). Fmctions for B isoforms were calcu- lated similarly. Phosphorylation stoichiomefry was then calculated by the following formula: mol phosphate/mol isoform A = Fxn A' + 2 x Fxn A' and mol phosphate/mol isoform B = Fxn B'+ 2 X Fxn B". Statistical analysis. Data are reported as means±standard errors (SE). Comparisons between two groups were made using the Student's t test Comparisons between more than two groups were made using analysis of variance and individual groups comparisons were done us- ing a Tukey honest significance difference test for post hoc compari- sons of ineans. Differences were considered significant at the P s 0.05 level. Results Identification ofMLC isoforms in HUVECs MLC isoforms from HUVECs were separated by two-dimen- sional gel electrophoresis. Fig. 1 A is an autoradiograph of a two-dimensional gel of immunoprecipitated MLC from un- stimulated HUVECs labeled with ['SS]methionine. Isoforms were observed at three molecular masses, 20, 19.7, and 16 kD. The most basic of the 20- and 19.7-kD isoforms were unphos- phorylated as they were not detectable in3zP-labeled celis (Fig. I B). Similarly, the 16-kD isoforms were unphosphorylated as they were also not observed in 32P-labeled cells (Fig. I B). Kawamoto et al. (14) also found 20- and 19.7-kD phosphory- lated isoforms and a 16-kD isoform in platelets. Unlike human platelet and smooth muscle MLC, the unphosphorylated A and B isoforms of HUVECs are focused adjacent to each other (14, 22). Also, the 16-kD isoform existed at only one isoelec- tric point in human platelets whereas in HU V ECs there are two 16-kD isoforms with distinct isoelectric points (Fig. I A). The 20- and the 19.7-kD isoforms each focused at three distinct isoelectric points, corresponding to the unphosphory- lated, monophosphorylated, and diphosphorylated states of each isoform. The unphosphorylated isoforms were focused toward the basic end where isoform B had a higher molecular mass and more acidic isoelectric point than isoform A. The corresponding mono- and diphosphorylated isoforms of A and B were focused toward the acidic end and were arbitrarily desig- nated A', B' and A", B', respectively (Fig. 1 A). We confirmed the identity of the unphosphorylated isoforms in'SS-labeled cells by depleting cells of ATP using antimycin A and deoxyglu- cose (23). ATP depletion generated two isoforms, which cone- Figure 1. (A) Autoradiograph of immunoprecipitated MLC isoforms from 3JS-labeled HUVECs. The iso- forms were separated by two-di- mensional gel electrophoresis. Iso- forms existed in unphosphorylated (A and B), monophosphorylated (A' and B'), and diphosphorylated (A' and B") states with distinct iso- electric points. (B) Autoradiograph of immunoprecipitated MLC iso- forms from 32P-labeled HUVECs. IEF See text for explanation. 1200 A. B. Moy, S. S. Shasby, B. D. Scott, and D. hf. Shasby

The Effect of Histamine and Cyclic Adenosine Monophosphate on Myosin Light Chain Phosphorylation in Human Umbilical Vein Endothelial Cells Alan B. Moy, Sandra S. Shasby, Brooks D. Scott, and D. NNchael Shasby Department ofMedicine, University of Iowa College of Medicine, and Veterans Administration Hospital, Iowa City, Iowa 52242 Abstract Histamine causes adjacent endothelial cells to retract from each another. We examined phosphorylation of the 20-kD myosin light chain (MLC,) in human umbilical vein endothe- lial cells (HUVECs) exposed to histamine to determine if we could find evidence to support the hypothesis that retraction of these cells in response to histamine represents an actomyosln- initiated contraction of the endothellal cytoskeleton. We found that MLCm in HUVECs was constitutively phosphorylated with ^- 0.2 mol phosphate/mol MLCm. Histamine increased MLC,o phosphorylation by 0.18±0.05 mol phosphate/mol MLCm. This peak increase in phosphorylation occurred 30 s after initiating histamine exposure, persisted through 90 s, and returned to control levels by 5 mio. Agents that increase HU- VEC cAMP prevent cell retraction in response to histamine. An increase in HUVEC cAMP decreased MLCm phosphoryla- tion by 0.18±0.02 mol phosphate/mol MLCm and prevented the increase in M" phosphorylation after exposure to hista- mine. Tryptic peptide maps of phosphorylated myosin light chain indicated that myosin light chain kinase phosphorylated MLC, in HUVECs under basal, cAMP-, and histamine-sti- mulated conditions. Phosphoaminoacid analysis of the mono- phosphorylated peptide indicated that, in contrast to smooth muscle cells, ser19 and thr's monophosphorylation occurs in HUVECs. On the basis of our results, modulation of myosin light chain kinase activity may be an important regulatory step in the control of endothelial barrier function. (J. Clin. Invest. 1993. 92:1198-1206.) Key words: myosin • phosphorylation • endothelial • histamine • cAMP Introduction Edema caused by molecules such as histamine is associated with the separation of adjacent endothelial cells from one an- other (1-8). Although there are multiple reports documenting this phenomenon, the mechanism of cell separation remains incompletely understood. Agents that disrupt actin filaments or that chelate calcium and breakdown calcium-dependent cell-cell and cell-substrate adhesion cause edema and separation of adjacent endothelial cells (6, 7, 9). These data demonstrate that separation may Address correspondence and reprint requests to Dr. Michael S. Shasby, Department of Internal Medicine, University of Iowa College of Medi- cine, Iowa City, IA 52242. Received for publication 12 October 1992 and in revised form 2 April 1993. result from loss of tethering between cells and between cells and substrate. They also suggest that constitutive centripetal intra- cellular cytoskeletal tension opposes these normal tethering forces, and unopposed expression of a constitutive tension may account for cell retraction when tethering forces are released. Others have suggested that the separation of adjacent endo- thelial cells from each other involves initiation of an active contraction mediated by actin and myosin (10, 11). Support for this hypothesis comes from two observations. First, Wysol- merski and Lagunoff ( 12) found that retraction of endothelial cells exposed to histamine was prevented if the cells were de- pleted of ATP before exposure to histamine. Second, l00 µM calcium, ATP, myosin hght chain kinase (MLCK),' and cal- modulin caused phosphorylation of the 20-kD myosin light chain ( MLCzo) and contraction of the cytoskeleton ofendothe- lial cells after their cell membranes were removed with deter- gents(10, I1, 13). Endothehal cell retraction in response to histamine is well documented (4). When histamine binds to an H, receptor on an endothelial cell it initiates a signal transduction cascade that results in an increase in cell calcium and diacyiglycerol (DAG) (2). The increase in cell calcium and DAG could stimulate an acute increase in MLCm phosphorylation by MLCK (calcium dependent) or by protein kinase C(PKC) (calcium and DAG dependent). The acute increase in light chain phosphorylation could then increase actomyosin contraction and centripetal tension (14-16). In the experiments described in this manuscript we investi- gated the effects of histamine on phosphorylation of MLCo in human umbilical vein endothelial cells (HUVECs). We were interested to determine if MLC2p was constitutively phosphor- ylated, if histamine acutely increased MLCs, phosphorylation in HUVECs, and, if so, how much phosphorylation of MLCxo increased. Because MLCw phosphorylation is mediated by both MLCK and PKC in other nonmuscle cells, we also con- structed tryptic peptide maps of the phosphorylated MLCp to directly determine which kinase was rgsponsible for MLCN phosphorylation in HUVECs, constitutively and after stimula- tion with histamine (15 ). We previously found that increasing cell cAMP prevented HUVECs from retracting in response to histamine, but the increase in cAMP did not prevent the in- crease in cell calcium (2). In other ce0s, one of the effects of increased cell cAMP is to inhibit MLC,o phosphorylation by MLCK (17). Hence, we were also interested in determining if increased cell cAMP prevents increased MLCa, phosphoryla- tion in histamine-stimulated HUVECs. If it does, it would, in part, explain how cAMP prevents the response of HUVECs to histamine. J. Clin. Invest. © The American Society for Clinical Investigation, Inc. 0021-9738/93/09/1198/09 $2.00 Volume 92, September 1993, 1198-1206 1. Abbreviations used in this paper: DAG, diacylglycerol; HUVEC, human umbilical vein endothelial celLr, MLCp, 20-kD myosin light chain kinase; MLCK, myosin light chain kinaw, PKC, prolein kinase C. 1198 A. B. Moy, S. S. Shasby, B. D. Scott, and D. M. Shasby

sium ions and edemaionnation in isolated lungs. Acm Physiol. Scand. 81:325- 339. 10. Schnittler, H., A. Wilke, T. Gress, N. Sutturp, and D. Drenckhahn. 1990. Role ofactin and myosin inthe control of pamcellularpermeability in pig, mt and human vascular endothelium. J. Physiol. 431:379-40I. 11. Wysolmerski, R., and D. Lagunoff. 1990. Involvement of myosin light chain kinase in endothelial cell retraction. Proc. Natl. Acad. SeL USA. 87:16-20. 12. Wysolmerski, R., and D. Lagunoff. 1988. Inhibition of endothelial cell retmction by ATP depletion. Am. J. Pathnl. 132:28-37. 13, Wysolmerski, R., and D. Lagunoff. 1991. Regulation of permeabilized endothelial cell retraction by myosin phosphorylation. Am. l Physlol. 261:C32- C40. 14 Kawamoto, S., A. Bengur, J. Sellers, and R. Adelstein. 1989. In situ phos- phorylafron ofhuman platelet myosin heavy and light chains by protein kinase C. J. Biol. Chem. 264:2258-2265. 15. Ludowyke, R., I. Peleg, M. Beaven, and R. Adelstein. 1989. Antigen-in- duced secretion of histamine and the phosphorylation of myosin by protein ki- nase C in rat basophilic leukemic cells. J. Biol. Chem. 264:12492-12501. 16. Taylor, D., and J. Stull. 1988. Calcium dependence of myosin light chain phosphorylation in smooth muscle cells. J Biol. Chem. 263:14456-14462. 17. Lamb, N., A. Fernandez, M. Conti, R. Adelstein, D. Glass, W. Welch, and J. Feramisco. 1988. Regulation of actin microftlament integrity in living non- muscle cells by cAMP-dependent protein kinase and the myosin light chain ki- nase. J. CellBiot. 106:1955-1971. l8. Daniel, J., and R. Adelstein. 1976. Isolation and properties of platelet myosin light chain kinase. Biochemistry 15:2370-2377. 19. Sellers, l., M. Sobceiro, K. Fmust, A. Bengur, and E. Harvey.1988. Prepa- mtion and characterization of heavy meromyosin and subfmgment I from verte- brate cytoplasmic myosin. Biochemistry. 27:6977-6982, 20. Isaaes, W., and A. Fulton. 1987. Cotranslational assembly of myosin heavy chain in developing cultured skeletal musele. Proc. Natl. Acad. Sci. USA. 84:6174-6178. 21. Gracy, R. 1977. Two dimensional thin layer methods. Methods Enzymo(. 47:195-204. 22. Singer, H. 1990. Protein kinase C activation and myosin light chain phos- phorylation in 32P-labeled arterial smooth muscle. Am. J. Physiol. 259:C631- C639. 23. Wilson, J., M. Winter, and D. Shasby. 1990. Oxidants, ATP depletion, and endothelial permeability to macromolecules. Blood. 76:2578-2582. 24. Haeberle, J., T. Sutton, and B. Trockman. (988. Phosphorylation of two sites on smooth muscle myosin: effects on contraction of glycerinated vascular smooth muscle. J. Biol. Chem. 263:4424-4429. 25. Stelzner, T., J. Weil, and R. O'Brien. 1989. Role of cyclic adenosine monophosphate in the induction of endothelial barrier properties. J. Cell. Phys- iol. 139:157-166. 26. Jacob, R., J. Meritt, T. Hallam, and T. Rink. 1988. Repetitive spikes in cytoplasmic calcium evoked by histamine in human endothelial cells. Nature (Lond.). 335:40-45. 27. SchilBng, W. 1989. Effect of membrane potential on cytosolic calcium of bovine aortic endothelial cells. Am. J. Physiol. 257:H778-H784. 28. Rotrosen, D., and J. Gallin. 1986. Histamine type I receptor occupancy increases endothelial cytosolic calcium, reduces F-actin, and promotes albumin diffusion across cultured endothelial monolayers. J. Cell. Biot. 103:2379-2387. 29. Hai, C., and R. Murphy. 1988. Sr" activates cross-bridge phosphoryla- tion and latch state in smooth muscle. Am. J. Physiol. 255:C401-C407. 30. Kolodney, M., and R. Wysolmerski. 1992. Isometric contraction by hbro- blasts and endothelial cells in tissue culture: a quantitative study. J. Cell Bio1. 117:73-82. 3 L Boswell. C., G. Majno, I. Joris, and K. Ostrom. 1992. Acute endothelial cell contraction in vitro: a comparision with vascular smooth muscle cells and fibroblasts. Microvaso. Re.c 43:178-191. 32. Morel, N., A. Dodge, W. Patton, I. Herman, H. Hechtman, and D, Shepro. 1989. Pulmonary microvascularendothelial cell contractility on silicone mbber substmte. J. Cell. Physiol. 141:653-659. 33. Fox, J., and D. Phillips. 1982. Role of phosphorylation in mediating the association of myosin with the cytoskeletal structures of human platelets. J. Biot. C'hem. 257:4120-4126. 34. Carson, M., S. Shasby, S. Lind, and D. M. Shasby. 1992. Histamine, actin-gelsolin binding and polyphosphoinositides in human umbilical vein endo- thelial cells. Am. J. PhysioL (Lung Cell Mof. Physiol. 7) 263:L664-L669. 35. Hekimian, G., S. Cote, 1. V. Sande, and J. M. Bceynaems. 1992. Ht receptor-mediated responses of aortic endothelial cellsto histamine. Am. J. Phys- iol. 262:H220-11224. 36. Delanerolle, P., M. Nishikawa, D. Yost, and R. Adelstein. 1984. In- creased phosphorylation of myosin light chain kinase after an increase in cyclic AMP in intact smooth muscle. Science (Wash. DC). 223:1415-1417. 37. Conti, M., and R. Adelstein. 1980. Phosphorylation by cyclic adenosine 3':5'-monophosphate dependent protein kinase regulates myosin light chain ki- nase. Fed. Proc.39:1569-1573. 38. Langeter, E., and V. Vanflinsbergh. 1991. Norepinephrine and iloprost improve harrier function of human endothelial cell monolayers: role of cAMP. Am. J. PhysioL 260:CI052-C7059. 39. Ikebe, M., and D. Hartshorne. 1985. Phosphorylation of smooth muscle myosin at two distinct sites by myosin light chain kinase. J. Biol. Chem. 260:10027-10031, 40. Umemmo, S., A. Bengur, and J. Sellers. 1989. Effect of multiple phus- phorylation of smooth muscle and cytoplasmic myosin in an in vitro motility assay. J. BioL Chem. 264:1431-1436. 41. Ikebe, M. 1989. Phosphorylation of a second site for myosin light chain kinase on platelet myosin. BiochemieTry 28:8750-8755. 42. Itoh, K., T. Hara, and N. Shibata. 1992. Diphosphorylation of platelet myosin by myosin light chain kinase. Biochim.-Biophys Actn. 1133:286-292. 43. Kumar, C., S. Mohan, P. Zavoday, S. Narula, and P. Leibowitz. 1989. Characterization and differential expression of human vascular smooth muscle myosin light chain 2 isoform in nonmuscle cells. Biochemistry. 28:4027-0035. 1206 A. B. May, S. S. Shasby, B. D. Scott, and D. M. Shasby

significant increase in cell cAMP from baseline in HUVECs exposed to histamine. As noted above, histamine increases both calcium and dia- cylglycerol in HUVECs. Hence, histamine could activate cal- cium-calmodulin-dependent kinases (MLCK) or DAG-de- pendent kinases (PKC). Since PKC-mediated MLC phosphor- ylation has been associated with platelet and basophil activation, it was of interest to us to determine which kinase was responsible for phosphorylation of MLC,, in HUVECs, both at rest and after stimulation with histamine (14, 15). We found that both basal and histamine-stimulated MLC,n phos- phorylation in HUVECs were mediated by MLCK, and we found no evidence for MLC phosphorylation by PKC in HU- VECs, even when the cells were stimulated with PDBU. We previously found that increasing HUVEC cAMP de- creased basal permeability and prevented histamine from in- creasing the permeability of HUVEC monolayers but did not prevent the increase in cell calcium (2). In the current experi- ments we found that increasing HUVEC cAMP markedly re- duced phosphorylation of MLCzp and prevented an increase with histamine. Hence, the effect of cAMP on histamine stimu- lation of MLC,, phosphorylation correlates with our earlier report on in vitro permeability (2). We also observed that MLCK mediates MLC,, phosphorylation under cAMP-stimu- lated conditions. Our data suggest that A kinase modulates MLC, phosphorylation indirectly by either decreasing the ba- sal activity of MLCK or enhancing phosphatase activity (17, 36). Conti and Adelstein (37) has shown that A kinase can phosphorylate and decrease MLCK activity. Phosphorylated MLCK binds poorly to the calcium-calmodulin complex, and contraction is prevented even when signal transduction is acti- vated. On the basis of our results, it is conceivable that modula- tion of MLCK activity may be an important regulatory step in the control of endothelial barrier function. In reports from others and ourselves, cAMP reduces the basal permeability of monolayers of endothelial cells (2, 25, 38). Whether a decrease in the level of MLC,, phosphorylation enhances endothelial barrier function is unclear. Using silicon rubber matrix, other investigators have shown that increasing cell cAMP induces matrix relaxation (31, 32). If centripetal isometric tension opposes tethering forces that link cells and substrate together, then, perhaps, a decrease in MLC2p phos- phorylation could decrease opposing tension on tethering forces which, in turn, could enhance adhesive forces between adjacent cells and cells to substrate. This could enhance barrier function. However, other effects of A kinase activation inde- pendent of MLC phosphorylation could also contribute to the enhanced barrier function, and it is not yet certain if cAMP prevents the response to histamine only by reducing MLC,o phosphorylation or whether there are also other effects of cAMP that may prevent the increase in permeability (2). Earlier reports of MLCK-mediated phosphorylation of MLCzO had identified ser"as the preferred site of MLCK phos- phorylation both in situ and in vitro in smooth muscle (24, 39, 40). The thr1e site has been shown to be diphosphorylated only under extreme in vitro conditions (39). However, phosphoa- mino acid analysis of the monophosphorylated peptide from HUVECs demonstrated phosphothreonine, suggesting that thr1e may be an acceptable monophosphorylation site in HU- VECs. It is unclear what effect monophosphorylation at thr1e would have on the actin-stimulated myosin ATPase activity compared with monophosphorylation at ser19. Diphosphory- lation at thr'a has been shown to augment and to have no effect on actomyosin contraction (24, 40-42 ). Protein sequencing of the threonine phosphorylated fragment will be necessary to confirm the identity of the monophosphorylated thr1e peptide. The isoelectric focusing patterns of the MLC,, isoforms from HUVECs were similar to, but not the same as, those re- ported in human platelets and smooth muscle (14, 22). HU- VECs demonstrated two 16-kD isoforms whereas in platelets there was only one I6-kD isoform. Also the isoform pattern of the 20-kD isoforms is unique. Other work has suggested that MLCs of nonmuscle cells may be encoded by different genes than those in smooth muscle cells (43). Our own observations would suggest that there are subtle differences among the iso- forms in smooth muscle, platelets, and endothelial cells, and these differences could contribute to differences in functional response. In summary, doses of histamine that increase the permeabil- ity of monolayers of HUVECs caused a modest increase in MLC,, phosphorylation in HUVECs, and the phosphorylation was mediated by MLCK. The amount of histamine-stimulated phosphorylation was consistent with the increase in HUVEC calcium that occurs with histamine stimulation, and further increases in cell calcium with ionomycin caused an appropri- ately greater increase in MLCzo phosphorylation. Increases in cell cAMP reduced basal phosphorylation of MLC20 and pre- vented the histamine-stimulated increase in MLC,, phosphor- ylation. Although these data are consistent with the hypothesis that MLCp phosphorylation contributes to retraction of HU- VECs stimulated with histamine, it does not rule out other mechanisms that may regulate cell retraction during inllamma- tion independent of actomyosin contraction. Acknowledgments Dr. Moy is a recipient of a Training Fellowship Grant from the Ameri- can Heart Association-Iowa Affiliate and the National Heart, Lung, and Blood Institute. This work was completed during Dr. D. M. Shasby's tenure as a Clinical Investigator and Dr. Scott's tenure as a Research Associate of the Veterans Administration. The work was also supported by National Institutes of Health grant HL-33540 and Ameri- can Lung Research grant 35131. References 1. Albertine, K. H., J. Weiner-Kronish, K. Koike, and N. C. Staub. 1984. Quantification of damage to lung microvessels in anesthetized sheep. I App(. Physial. 57:1360-1368. 2. Carson, M., S. Shasby, and D. M. Shasby. 1989. Histamine and inositol phosphate accumulation in endothelium: cAMP and G-protein. Am. J. Phy.siol. 257:L259-L264. 3. Laposata, M., D. Dovnatsky, and H. Shin. 1983. Thrombin-induced gap forntation in confluent endothelial cell monolayers in vitro. Blood 62:549-556. 4. Maino, G., and G. Palade. 1961. Studies on inllammation. 1. Effect of histamine and semtonin on vascularpermeability: an electron microscopic study. J Braphys. Bioclim. CytoL 11:571-605. 5. Shasby, D. M., S. Lind, S. Shasby, J. Goldsmith, and G. Hunninghake. 1985. Reversible oxidant-induced increases in albumin transfer across cultured endothelium: alterations in cell shape and calcium homeos[asis. Blood. 65:605- 614. 6. Shasby, D. M., S. Shasby-1. Sullivan, and M. Peach. 1982. Role of endothe- tial cell cytoskeleton in control of endotheGal permeability Circ. Res. 51:657- 661. 7. Shasby, D. M., and S. Shasby. 1986. Effects of calcium on transendothelial albumin transfer and electrical resistance. J Appl. Phyriol. 60:71-79. 8. Wysolmerski, R., and D. Lagunofl: 1985. The effect of ethchlorvynol on cultured endothelial eells-a model for the study of the mechanism of increased vascular permeability. Am. J Parhol. 119:505-512. 9. Nicolaysen, G. 1971. Intravascular concentrations of calcium and magne- Histamine and cAMPAlter.+Nyosin Phosphorylation in Endotheliaf Cells 1205

Methods Materials. Tissue culture supplies were obtained from the Cancer Center, University of Iowa. Fetal bovine serum was obtained from Hyclone Laboratories, Inc. (Logan, UT). Polyclonal rabbit IgG anti- myosin antibody against human platelet whole myosin was obtained from Biomedical Technologies, Inc. (Stoughton, MA). Protein A, Staphylocaucus aureus cell suspension was obtained from Calbiochem Corp. (San Diego, CA). Rat brain PKC was obtained from Calbio- chem Corp. Histamine, theophylline, 8-bromo cAMP, DL-histidine, DL-glutamic acid, phosphoserine, and phosphothreonine were ob- tained from Sigma Chemical Co. (St. Louis, MO). L-1-p-tosylamino- 2-phenylethyl chloromethyl ketone (TPCK) trypsin was from Worth- ington Biochemical Corp. (Freehold, NJ). ["P]Orthophosphate and ry-['2P]ATP were obtained from New England Nuclear (Boston, MA). ['''S]Methionine was obtained from ICN Radiochemicals (Irvine, CA). Transwells were purchased through Costar Corp. (Cambridge, MA ). Smooth muscle MLCK and turkey gizzard MLC were gifts from Dr. James Sellers, National Heart, Lung and Blood Institute (Be- thesda, MD). All other chemicals were reagent grade. Cell culturing. Cultured HUVECs were prepared by collagenase treatment of freshly obtained umbilical veins as described (2). For experiments designed to measure the amount of phosphorylated MLC,a and for some of the experiments designated for peptide map- ping, harvested cells were plated on 25-mm diameter polycarbonate filters (0.8-µm pore size) precoated with 30 µg/ml of fibronectin as described (2). Alternatively, harvested cells were plated on 60-mm diameter tissue culture plates (Costar Corp.) precoated with I% gelatin for some of the peptide mapping experiments. There was no difference in the peptide maps of cells grown on micropore filters when compared with cells grown on tissue culture plastic. However, basal phosphoryla- tion was higher ( - 0.2 mol phosphate/mol light chain) in cells grown on tissue culture plastic than in cells grown on micropore filters (n > 25 for each). All cells were cultured in medium 199 and supplemented with 20% fetal bovine serum, basal medium Eagle vitamins and amino acids, glucose (5 mM), glutamine (2 mM), penicillin (100 µ/ml), and streptomycin (100 pg/ml). All studies were performed on primary cultures that were 2 d postconfluent at the time of study. Cultures were identified as endothelial cells by their characteristic uniform morphol- ogy, uptake of acetylated LDL (- 99%of cells), andby indirect immu- nofluorescent staining for factor VIII (- 97% of cells). Because the cells are not visible on the polycarbonate filters, for each experiment, a control well was plated and examined for morphology to be certain that the morphology was consistent with that of cells identified as endothe- lial cells by these staining techniques. Purijication of human platelet myosin MLCp. Human platelet MLCm was purified from whole platelet myosin using the procedure described by Daniel and Adelstein (18) . Preparations were used if con- taminating kinases were not present. W hole-platelet myosin was pur{- fied according to Sellers et al. (19). Whole-platelet myosin was not a suitable substrate for generating in vitro peptide map standards because it contained contaminating endogenous kinases. Isotopic labeling and stimulation. Cells designated for quantimting MLC20 phosphorylation were grown on micropore filters and were la- beled with 1.5 ml ["S]methionine (555 uCi/ml) in M199 with 10% fetal bovine semm for 48 h at 37°C and 5% COz. Labeling in < 10% fetal bovine serum increased basal phosphorylation of MLCm. Cells designated for peptide mapping were grown on either micro- pore filters or tissue culture plates. Confluent cells were washed three times with a phosphate-free buffer with the following composition (mM): 119 NaQ, 5 KQ, 5.6glucose, 0.4 MgQi, I CaCh, 25 Pipes(pH 7.2), and were then labeled with 3 ml of 300-400 µCi/ml [32P]- orthophosphate for 2 h at 37°C without COi. Labeled cells were exposed to either buffer, histamine (10-' mol/ liter), or a mixture of agents used to increase cell cAMP (8-bromo- cAMP, 10-4 mol/liter, forskolin, 2 X] 0-' mol /liter, and theophylline, 10' mol/liter) with or without histamine. There was no difference in MLCm phosphorylation between cells exposed to the mixture of agents to increase cAMP versus 8-bromo-cAMP alone, and these groups were combined for analysis, Reactions were terminated by snap freezing on a dry ice methanol bath and the cells were lysed by thawing in I ml of a buffer containing 1% NP-40, 100 mM sodium pyrophosphate, 250 mM NaCl, 50 mM NaF, 5 mM EGTA, 0. 1 mM PMSF, 10 µg/ml leupeptin, 15 mM ft- mercaptoethanol, and 20 mM Tris-HCl (pH 7.9). The lysate was sedi- ment at 100,000 g for 5 min at 4°C, and the supernatant was incubated with 20 µl of rabbit anti-human platelet myosin antibody (2 mg/ml) at 4°C for 1 h. 50 µl of a prewashed S. aureus cell suspension (Pansor- bin) was then added to the suspension, which was incubated for an additional 30 min at 4°C (20). The mixture was centrifuged (100,000 g for I min ) and the pellet was washed with 0.5 ml of the lysing buffer and recentrifuged. The pellet was then washed with 0.5 ml of a 50:50 mixture of lysing buffer and PBS. The pellet was resuspended in 200 µ1 of SDS sample buffer for SDS-PAGE or 35 µI of urea-lysing buffer for two-dimensional electrophoresis. Isoelectric jocu.cing. The unphosphorylated and phosphorylated MLCz,a isoforms were separated by two-dimensional electrophoresis as described by Ludowyke et al. (15) but with the following modifica- tions. The sample was suspended in 35 µl of a buffer containing 9.5 M urea (ultrapure urea; Boehringer Mannheim Corp., Indianapolis, IN), NP-40, 0.04% pharmalyte (pH 4.5-5.4), 0.04% pharmalyte (pH 4-6.5 ) and 0.1 M DTT. Tube gels were prepared using 1.5 mm i.d. x 180-mm glass tubes. For 10 tubes, 2.78 g of ultra pure urea was mixed with 1.5 ml of HiO, 0.83 ml of 30%:0.8% acrylamide/bisacrylamide, 0.28 ml of pharmalyte (pH 4-6.5 ), 0.28 ml of pharmalyte (pH 4.5- 5.4 ), and 0.28 ml of CHAPS buffer (100 mg CHAPS, 0.3 ml HzO, and 0.03 ml NP-40). Gels were drawn to a height of 150 mm. Gels were polymerized with 5 µl of TEMED and 10 td ammonium persulfate. Isoelectric focusing was performed using a cell (model 175; Bio-Rad Laboratories (Richmond, CA). The anolyte contained 0.01 M glutamic acid and the catholyte con- tained 0.01 M histidine. Gels were prefocused at 25 °C at 200 V for 2 h followed by 500 V for 5 h. 30-µ1 samples were loaded onto the gel, covered with 10 pl of an overlaying buffer, which contained 8 M urea, 2% pharmalyte (pH 3-10), and 0.5%pharmalyte ( pH 4.5-5.4). Catho- lyte buffer was layered over the overlaying buffer. The samples were then electrophoresed at 2,500 V for an additional 32,000 V h. Second Dimension S'DS-PAGE. Tube gels were extracted and rinsed with transfer solution (3% SDS, 0.07 M Tris-Base (pH 6.7), and 0.03 mg/mI bromphenol blue), layered onto a polyacrylamide/bisac- rylamide slab gel (15%:0.4%), and sealed to the slab gel with hot agar (2.3% SDS, 0.062 M Tris-Base (pH 6.8 ), 0.5% agarose, and 5% p-mer- captoethanol). The gels were run at 4°C for 18 to 20 h at 200 V. Gels were treated with En'hance (Dupont Pharmaceuticals Inc., Wilming- ton, DE) to accelerate autoradiograph development. Autoradiographs were developed using Kodak X-OMAT AR film. The relative isoform distribution was quantitated using an Image Quant laser densitometer (Molecular Dynamics, Sunnyvale, CA). Twa-dimensional peptide mapping of myosin tryptic peptides. Im- munoprecipitate from "P-labeled HUVECs was dissolved in SDS sam- ple buffer as above and separated by an 8-t 5% polyacrylamide gradient gel, and electrophoresed at 200 V for 4 to 5 h at 21 °C. The "P-labeled MLC, was cut from the wet gel into small pieces and washed three times for 30 min with 40% methanol, 10% acetic acid, and 50% water, and then an additional three times with 10% methanol for the same time period. The washed gel slices were subsequently dried under nitro- gen and then digested with I ml of 60 pg/ml of TPCK-trypsin in 50 mM ammonium bicarbonate at 37°C for 4 h. The digest was collected and stored at 4°C. The remaining gel was further digested with I ml of 30 pg/ml trypsin for an additiona120 h. Digests were pooled and lyoph- ilized. Peptides were resuspended in I ml of water and lyophilized. Digest was resuspended with 30 ul of electrophoresis buffer (acetic acid/formic acid/water, 9:3:88), spotted on a plastic 20 x 20 x 0.20 mm silica P 60 TLC plate (EM Science, Gibbstown, NJ). Peptides were electrophoresed toward the cathode at 900 V at 4°C for 45 min using the technique described by Gracy (21). Peptide migration was Histamine and cAMP Alter:Ltyosin Phosphorylation in Endothelial Cells 1199

---