Proc. Nati. Acad. Sci. USAVol. 81, pp. 6164-6168, October 1984Medical Sciences

Ca2` causes active contraction of bile canaliculi: Direct evidencefrom microinjection studies

(actin/calmodulin/cell motility/liver cell/intercellular communication)

SUMIO WATANABE* AND M. JAMES PHILLIPStResearch Institute and Departments of Pathology and Pediatrics, Hospital for Sick Children, and University of Toronto, Toronto, Ontario M5G 1X8

Communicated by Hans Popper, June 4, 1984

ABSTRACT Cytoplasmic microinjection of Ca2+ triggerscontraction of bile canaliculi in freshly isolated monolayer cul-tures of rat hepatocytes. Unseparated paired hepatocytes wereused to demonstrate this motility-based phenomenon. Onlyone cell of the pair was injected, but fluorescein spread fromthe target cell to the opposite cell; also, the contractions werealways uniform, equally involving both hepatocytes that formthe canaliculus, indicating that communication exists betweenthe cell pairs. Inhibitors of calmodulin and actin filaments,trifluoperazine and cytochalasin B, respectively, inhibited theCa2+-induced contractions. Hence, the mechanism of contrac-tion has features in common with actin-myosin based cyto-plasmic motility behavior found in other non-muscle cells.

Both actin and myosin are present in many non-muscle cellsincluding hepatocytes (1-5). In the liver cell, actin is presentthroughout the cytoplasm but is most evident under the plas-ma membrane, especially in the region of bile canaliculi; thishas been shown by many investigators (2, 6-9). It has beendemonstrated recently that in non-muscle cells the mecha-nism of actin-myosin based contractile events is differentfrom that observed in skeletal muscle; for instance, in non-muscle cells, contraction is myosin- rather than actin-based.Present evidence indicates that the initial event is an in-crease in intracellular calcium ion concentration from 0.1 to1 ,uM; that calcium combines with calmodulin, forming cal-cium-calmodulin complexes, which are involved in the acti-vation of inactive myosin light chain kinase; and results, fi-nally, in myosin activation by actin. The precise details ofthis activation process are a matter of intense current re-search (for recent reviews, see refs. 10-12). Structurally,there are a number of close parallels between hepatic cana-liculi and the intestinal brush border. The brush border is ofspecial interest to this study, because like the bile canaliculi,it can be induced to move in vitro (13, 14). The biochemistryof the contraction mechanism in brush borders has been ex-tensively studied (13-17) and present evidence is that, likeother non-muscle cells, it is a myosin-based mechanism.

In the liver, this mechanism has not been studied. Indeed,it is only very recently that cytoplasmic motility events havebeen observed in liver cells (18). Cytoplasmic microinjectionis a new technique, which permits the direct study of someaspects of hepatic cell function. In the present report, wecompare spontaneous bile canalicular contractions withthose that occur after intracellular microinjection of Ca2+ inisolated hepatocytes. The effects of inhibitors of actin (cyto-chalasin B) (19) and calmodulin (trifluoperazine) (20, 21) onthe contractions are also reported.

MATERIALS AND METHODSExperimental Materials. Collagenase (type I) and trifluo-

perazine dihydrochloride were purchased from Sigma. Cyto-chalasin B was purchased from Aldrich.

Measurement of Calcium. Direct measurements of the ion-ized calcium concentration of the injected solutions and cul-ture media were carried out using a Nova 2 ionized calciumanalyzer (Nova Biochemical, Newton, MA) (22).

Preparation of Primary Cultured Hepatocytes. Monolayercultures of rat hepatocytes were obtained from female Wis-tar strain rats (body weight, 200 g) by using a collagenasedigestion technique (23, 24) described in detail previously(25). The cultured hepatocytes were maintained in L-15 me-dium, supplemented with 10% fetal bovine serum/10 mMHepes, pH 7.2/penicillin (100 units/ml)/streptomycin (100,ug/ml). The culture medium contained 0.92 ± 0.01 (mean ±SEM) mM ionized calcium.The cell-separation technique is designed to prepare iso-

lated single hepatocytes, but in such preparations =10% ofthe cells remain in small incompletely separated clumps.Amongst them are cell pairs in which bile canaliculi are dilat-ed and easily visible; these cell pairs were selected for evalu-ation.

Viability was assessed by trypan blue exclusion and bytime-lapse cinephotomicrography as described (25). Afterisolation, viability was 82%-93%. In the trifluoperazine andcytochalasin B experiments of the hepatocytes that were via-ble at the start of treatment, viability was 94%-100% at theend of treatment with trifluoperazine, and 95%-100% withcytochalasin B. There were no significant differences be-tween controls (95%-100%) and these groups. Experimentswere performed 4-8 hr after cell isolation.

Microinjection Method. Microinjections were done ac-cording to the modified methods of Stockem et al. (26) andGraessmann and Graessmann (27). Micropipettes were pre-pared from glass capillaries (outer diameter, 2 mm) (W-P In-struments, New Haven, CT), using a micropipette puller(PG-1, Narishige Scientific Instruments, Tokyo, Japan). Thediameter of the tip of the micropipette was 0.5-1.0 ,um. Afterthe capillary was pulled, the solution was delivered throughits rear open end by means of a 20 GA Angiocath (Deseret,Sandy, UT). Then the micropipette was connected to thehub of the pipette holder (HI-1, Narishige), which was thenattached to the microinjector. Microinjections were done us-ing a micromanipulator (MP-1, Narishige) and a microinjec-tor (IN-4A, Narishige) on the stage of an inverted micro-scope (Nikon, Tokyo, Japan) equipped with phase-contrastand fluorescence optics. The temperature was maintained at37°C by using a plastic housing with an incubator (NP-2, Ni-kon).

In all instances, the dose of injected solution was -10% ofthe volume of the recipient cell (26). The microinjectiontechnique itself sometimes caused cell damage, such as blebformation; in all such cases, these experiments were reject-ed. The rate of successful microinjection was 40%-100% inthe present study, and no significant difference in successful

*Present address: Juntendo University, Department of Gastroenter-ology, 2-1-1 Hongo Bunkyo-ku, Tokyo, Japan.tTo whom reprint requests should be addressed at: 555 UniversityAve., Room 3112, Toronto, Ontario M5G 1X8 Canada.

6164

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

6, 2

020

Proc. NatL Acad Scd USA 81 (1984) 6165

microinjections was found between pretreated (cytochalasinB or trifluoperazine) and nontreated groups.

In other experiments, hepatocyte cultures were pretreatedwith cytochalasin B or trifluoperazine and then microinject-ed. In the cytochalasin experiment, 50 AuM cytochalasin Bwas administered for 3 hr before microinjection; 0.1% di-methyl sulfoxide (solvent for cytochalasin B) did not affectthe viability of the cells (survival rate, 95%-100%).

In the trifluoperazine experiment, the hepatocytes werepretreated with 20 ,uM trifluoperazine for 1 hr before micro-injection.Experimental Design. The essential design of this study

was to permit comparison of bile canalicular contractionsthat occur spontaneously with contractions in cells microin-jected with Ca2+ (standard solutions) and to assess the ef-fects of the inhibitors trifluoperazine and cytochalasin B.Two standard calcium solutions were used in the main partof this study: a calcium solution containing 0.1 mM CaCl2,designated the standard calcium solution; and a calcium-freesolution that contained 1 mM EGTA (Ca2+ < 10 nM), desig-nated the standard calcium-free solution. The Ca2+ concen-tration of all solutions was directly measured as describedabove, and all solutions contained 5 mM Hepes (pH 7.2) and0.05% sodium fluorescein (28) as a marker of cell injection.In each experiment, 100 cell pairs with visibly dilated bilecanaliculi were examined. The cell pairs examined were de-rived from 5 separate isolated liver cell preparations, 20 cellpairs per preparation.

In other experiments, to assess the physiological signifi-cance of the results with the standard calcium solutions (0.1mM) and <10 nM, the effects of other calcium-containingsolutions were also examined. In these experiments, 10 ,uM,1 AM, and 10 nM Ca2+ solutions buffered with EGTA wereused under similar experimental conditions.Other experiments were carried out to assess the dose-

response effects of cytochalasin B and trifluoperazine onspontaneous and on Ca2'-induced bile canalicular contrac-tions, using the standard calcium solution. Cytochalasin B in0, 5 ,uM, 15 AM, and 50 ,uM solutions was assessed. In simi-larly designed experiments, trifluoperazine in 0, 5 uM, 10AM, and 20 AM concentrations was examined.Monitoring Bile Canalicular Contractions. Bile canalicular

contraction, defined as a visible decrease in diameter of thecanalicular lumen, was recorded by taking a series of photo-graphs at 1-min intervals beginning immediately prior to in-jection and then for 10 min. Time-lapse cinephotomicro-graphs were also taken, using previously described methods(25). Fluorescence micrographs were taken 1 min and/or 10min after the microinjection for evidence of successful cellinjection and for assessment of cell-cell communication.The wavelength of the excitation filter was 410-485 nm, andthe wavelength of the cutting filter was 515 nm. Ilford XP1400 film was used.

Statistical Analysis. Statistical analysis was done usinganalysis of variance. The results are expressed as the mean± SEM. Since 100 canaliculi were assessed in each experi-ment, the results are also expressed as percentages for easeof comparison.

RESULTSThe main results are summarized in Table 1, which com-pares spontaneous contractions with those occurring in cellsmicroinjected with the standard calcium solutions. The tablealso summarizes the effects of the two inhibitors cytochala-sin B and trifluoperazine.Spontaneous Bile Canalicular Contractions in Noninjected

Cells. In normal untreated hepatocytes, 2.8 + 0.4 spontane-ous contractions were seen in each of the five groups of 20bile canaliculi observed over the 10-min experimental period

Table 1. Effects of Ca2+ on bile canalicular contractions innormal, cytochalasin B-, and trifluoperazine-treated hepatocytes

Spontaneous Injected with Injected withcontraction standard standard

(non- calcium-free calciumTreatment injected) solution solution

None 2.8 + 0.4 3.0 ± 0.5 16.8 ± 0.4(14%) (15%) (84%)

Cytochalasin B 0.4 ± 0.4 0.4 ± 0.2 1.6 ± 0.7(50 AM) (2%) (2%) (8%)

Trifluoperazine 1.2 ± 0.4 1.0 ± 0.3 7.2 ± 0.9(2 AM) (6%) (5%) (36%)Each value indicates the number of bile canalicular contractions

during the 10-min observation period. Results are expressed as themean ± SEM of five groups, each comprised of 20 bile canaliculi.Percentages are in brackets. Note that the microinjection with stan-dard calcium solution causes bile canaliculi to contract (85% com-pared to 14%; P < 0.001). Calcium-triggered bile canalicular con-tractions are inhibited by pretreatment with cytochalasin B (P <0.001) and with trifluoperazine (P < 0.01).

(14% of the canaliculi contracted). In the 50 ,uM cytochalasinand 20 ,uM trifluoperazine pretreated groups, there were 0.4± 0.4 and 1.2 ± 0.4 contractions, respectively (2% and 6% ofcanaliculi contracted). The difference between normal (nopretreatment) controls and pretreated noninjected groupswas significant: P < 0.01 (cytochalasin group) and P < 0.05(trifluoperazine group).

Bile Canalicular Contractions After Microinjection withStandard Solutions. Standard calcium free solution. In nor-mal (no pretreatment) hepatocytes microinjected with calci-um-free solution, the number of contractions was 3 ± 0.5(15%). Interestingly, there was no significant difference be-tween the incidence of contractions in this group and thatwhich occurred spontaneously in normal untreated controls.In the cells that were pretreated with 50 1AM cytochalasin Bor 20 AuM trifluoperazine, microinjection of calcium-free so-lution resulted in bile canalicular contractions in 0.4 ± 0.2(2%) and 1.0 ± 0.3 (5%) instances. These results are not sig-nificant compared to the spontaneous occurrence of contrac-tions.Standard calcium solution. After microinjection of the

standard calcium solution in normal (no pretreatment) hepa-tocytes, the number of contractions was 16.8 ± 0.4 (84%).There was a significant difference between this group andthe cells injected with the standard Ca2+-free solution (P <0.001) and between this group and the cells that were notinjected (P < 0.001). In the cells that were pretreated with 50AM cytochalasin B, the number of contractions after the mi-croinjection with the standard calcium solution was signifi-cantly less (1.6 ± 0.7; 8%) than in those cells that were notpretreated with cytochalasin B (P < 0.001). Similarly, thenumber of contractions after the microinjection of the stan-dard calcium solution decreased significantly in the 20 /1Mtrifluoperazine pretreated cells (7.2 ± 0.9; 36%) as comparedto the no pretreatment group.

Representative results of the bile canalicular contractionsare shown in Figs. 1-4. In these figures, contraction is repre-sented by a decrease in the caliber of the canalicular lumen;this is best seen with phase-contrast optics. In each instance,the bile canaliculi contracted uniformly. As shown by thefluorescein marker in Figs. 1, 2, and 3, the hepatocyte on theright was injected; in Fig. 4, it was the cell on the left.Dose-Response Experiments. The microinjection of calci-

um solutions of different concentrations resulted in contrac-tions as follows: 0.1 mM, 16.8 ± 0.4 (84%); 10 uM, 17.2 +0.6 (86%); 1 AM, 3.0 ± 0.5 (15%); 10 nM, 3.0 ± 0.5 (15%).These results show a marked difference between free calci-um concentrations of 10 ,uM and above, and of 1 uM and

Medical Sciences: Watanabe and Phillips

Dow

nloa

ded

by g

uest

on

Aug

ust 1

6, 2

020

6166 Medical Sciences: Watanabe and Phillips

FIG. 2. Normal hepatocytes (no pretreatment). (x440.) (a) Justprior to injection of the standard calcium-free solution. Open lumenof the bile canaliculus (arrow) can be seen between two hepatocytes.Phase-contrast micrograph. (b) Hepatocytes 10 min after the stan-dard calcium-free solution. The bile canaliculus has not contracted.(c) Injected fluorescent dye has spread into the neighboring cell.Fluorescence micrograph.

using standard calcium-free solution in the controls (no pre-treatment), 17.2 ± 1.0 (86%); cytochalasin B, 17.6 ± 1.2(88%); trifluoperazine, 17.2 ± 1.0 (86%). There is no statisti-cal difference between these results. This spread of fluores-cent marker is an indication of cell-cell communication.

FIG. 1. Normal hepatocytes (no pretreatment) (x440.) (a) Openlumen of the bile canaliculus (arrow) can be seen between two (cou-pled) hepatocytes. Phase-contrast micrograph. (b) The micropi-pette, which contains the standard calcium solution, is penetratingthe hepatocyte on the right. (c) The solution, which contains sodiumfluorescein, has been microinjected into the hepatocyte. Note thatonly the target hepatocyte is visualized. Fluorescence micrograph.(d) Hepatocyte 10 min after the injection of the standard calciumsolution. The lumen of the bile canaliculus has closed (i.e., canalic-ular contraction has occurred). (e) Fluorescent dye has spread intothe neighboring cell. Fluorescent intensity of the recipient cell isstronger than intensity of the neighboring cell.

below. Since the injected volume was -10% of the cell vol-ume, the results appear to be close to those reported by oth-ers, showing that intracellular Ca>2 of - 1 ,uM triggers con-tractions in other non-muscle cells, although exact intracel-lular Ca>2 contraction after the microinjection is not known(10, 11).The effects of cytochalasin B and trifluoperazine on calci-

um-induced bile canalicular contractions were dose depen-dent. The results are shown in Tables 2 and 3.

Cell-Cell Communication. In each instance, as shown inFigs. 1-4, after microinjection, the fluorescent markerspread into the neighboring noninjected hepatocyte of thecell pair. The incidence of spread of the marker to the oppo-site cell was as follows: using standard calcium solution, inthe controls (no pretreatment), 16.6 ± 0.5 (83%); cytochala-sin B, 18.0 ± 1.2 (90%); trifluoperazine, 16.8 ± 0.5 (84%):

FIG. 3. Hepatocytes pretreated with cytochalasin B. (x440.) (a)Just prior to injection of the standard calcium solution. Lumen of thebile canaliculus shows dilation (arrow). Phase-contrast micrograph.(b) Hepatocytes 10 min after injection of the standard calcium solu-tion. The dilated bile canaliculus has not contracted. (c) Injectedfluorescent dye has spread into the neighboring cell. Fluorescencemicrograph.

Proc. NatL Acad Sci. USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Aug

ust 1

6, 2

020

Proc. Natl. Acad. Sci. USA 81 (1984) 6167

FIG. 4. Hepatocytes pretreated with trifluoperazine. (x440.) (a)Just prior to injection of the standard calcium solution. Bile canalic-ulus is dilated (arrow). Phase-contrast micrograph. (b) Hepatocytes10 min after injection into the hepatocyte on the left. Bile canaliculushas not contracted. (c) The injected fluorescein has spread into theneighboring hepatocyte. Fluorescence micrograph.

DISCUSSIONMicroinjection of cells is a powerful technique, because itovercomes the difficulty of the impermeability of the plasmamembrane to the probes needed to study vital cell processes.The technique permits the direct study of an almost limitlessnumber of agents in living cells and is an excellent tool forthe evaluation of cytoplasmic motility events, such as bilecanalicular contraction. In the study reported here, it pro-vides decisive evidence that microinjection of Ca2+ ions intothe liver cell cytoplasm results in active bile canalicular con-traction in 84% of instances, compared to 14% in controls.Previous studies using time-lapse cinephotomicrographyhave shown that spontaneous canalicular motility with open-ing and closing of canaliculi is intermittent (18, 25). The con-tractions may be related to bile secretion, because they aresignificantly increased by the addition of taurocholate, acholeretic bile acid, to the medium (29). It is not known ifcontractions occur in vivo, but they may be a mechanism for

Table 2. Dose-response effects of cytochalasin B and oftrifluoperazine on spontaneous bile canalicular contractions

Cytochalasin B

0 5 ,uM 15 ,uM 50 AMNo. of 2.8 ± 0.4 3.2 ± 0.8 2.0 ± 0.6 0.4 ± 0.4*

contractions (14%) (16%) (10%) (2%)

Trifluoperazine

0 5 ,M 10 AM 20 /,M

No. of 2.8 ± 0.4 2.4 ± 0.4 2.0 ± 0.6 1.2 ± 0.4tcontractions (14%) (12%) (10%) (6%)

Each number indicates mean ± SEM of five groups of 20 bilecanaliculi.*P < 0.01.tp < 0.05.

Table 3. Dose-response effects of cytochalasin B and oftrifluoperazine on Ca2-triggered bile canalicular contractions

Cytochalasin B

0 5ouM 15 gM 50 ,uM

No. of 16.8 ± 0.4 16.8 ± 1.2 12.0 ± 1.3* 1.6 ± 0.7tcontrac- (84%) (82%) (66%) (8%)tions

Trifluoperazine

0 5 ILM 10 Mm 20,uMNo. of 16.8 ± 0.4 14.4 ± 0.4 11.6 ± 0.8* 7.2 ± 0.9t

contrac- (84%) (72%) (58%) (36%)tions

Each number indicates mean ± SEM of five groups of 20 bilecanaliculi. Results were obtained using microinjections of the stan-dard calcium solution.*P < 0.05.tp < 0.001.

the forward movement of bile into the next upstream regionof the canalicular network, hence facilitating bile flow. Inprevious work on spontaneous contractions, we have inter-preted the closure of the canaliculus as an active process,because it appeared forceful, was inhibited by cytochalasinB (30), and was associated with the expulsion of material(canalicular bile) from the lumen (25). However, an alterna-tive possibility is that with bile secretion into the canaliculus,there is a buildup of pressure, which at a critical point resultsin forceful opening and then collapse of the lumen in a man-ner similar to a pressure valve. The finding that the microin-jection of Ca2+ ions into liver cells results in the prompt con-traction of bile canaliculi is strong evidence that the con-tracting phase (luminal closure) is an active process.21Regarding the mechanism, it is interesting that Ca ions

can trigger bile canalicular contractions, because an increasein intracellular calcium ions is the basis of actomyosin-basedintracellular motility events in other non-muscle cells (15,17, 31, 32). Using presently available techniques, it is notpossible to determine the exact intracellular free calcium ionconcentration in the single cells injected or in the doubletsbefore and after microinjection. Indirect measurements byothers of free calcium ions in hepatocyte suspensions showthis value to be 0O.1 ,uM (33). In the experiments reportedhere in which we microinjected 10 uM, 1 ,uM, and 10 nMCa2+ solutions (buffered with EGTA), the results suggestthat there is a threshold for contraction. The inhibitory ac-tion of trifluoperazine on the contraction process suggeststhat calmodulin is involved, as is also the case in other non-muscle cells (34). It is well known that calmodulin is an im-portant regulator of many calcium-mediated cellular pro-cesses. Calmodulin has been found in most animals andplants, and calmodulin has many known functions, includingactivation of phosphodiesterase, adenylate cyclase, Ca2+-Mg2+ ATPase, myosin light-chain kinase, glycogen synthe-tase kinase, and cyclic nucleotide phosphodiesterase (34-36). Calmodulin apparently plays a role in processes such asprostaglandin synthesis, smooth muscle contraction, intesti-nal secretion, insulin secretion, disassembly of microtu-bules, and it may possibly be involved in ciliary motility andaxonal flow (34, 36). In virtually every case, the phenothi-azine antipsychotic drugs (including trifluoperazine) havebeen shown to inhibit these calmodulin-dependent enzymeactivities and processes (36). Therefore, trifluoperazine hasdiverse effects, but an inhibition of bile canalicular contrac-tions would be expected, as was found, if the contractionprocess is calcium-calmodulin mediated.

In previous reports, it has been shown that the actin fila-ment inhibitors cytochalasin B (and D) (30) and phalloidin(37) inhibit spontaneous canalicular contractions. Here, we

Medical Sciences: Watanabe and Phillips

Dow

nloa

ded

by g

uest

on

Aug

ust 1

6, 2

020

6168 Medical Sciences: Watanabe and Phillips

report that cytochalasin B also inhibits calcium-triggeredcontractions, presumably because of the effects of cytocha-lasin B on actin filaments (19).

It is generally believed that the intercellular communica-tion is regulated mainly by gap junctions and that high con-centrations of intracellular Ca2 (50-100 ,uM) disturb thiscommunication (38). In our present results, it is shown thatintercellular communication was well preserved in the pri-mary cultured hepatocytes after the microinjection of 100tkM Ca2+. Although we did not measure the intracellular cal-cium concentration after the microinjection with calcium so-lution, it is not unreasonable to assume that the intracellularcalcium concentration after the microinjection with 100 ,uMCa2+ rapidly falls to levels lower than 50-100 AM.

It is concluded from this study that calcium microinjectiontriggers bile canalicular contraction in vitro and that themechanism involves calcium, calmodulin, and actin fila-ments. Many questions remain: for instance, do contractionsoccur in vivo and, if so, are they triggered by an increase incytoplasmic free Ca2+ ions, and if this is the case, what is thephysiological stimulus and mechanism of this increase in in-tracellular calcium? The answer to these and other questionsmust await further study.

The authors thank Professors E. Farber, A. Rothstein, I. C.Radde, and T. Namihisa for their support and encouragement. Dr.Radde, Department of Endocrine Research, assisted with the calci-um measurements. The authors also thank Dr. S. Grinstein for criti-cally reviewing this manuscript. Mr. Vernon Edwards, Mrs. JanetRowley, and Mrs. Wandee Vetvutanapibul provided excellent tech-nical assistance. This research was supported by a Grant-in-Aidfrom the Medical Research Council of Canada (MT-0785) and by theResearch Institute of the Hospital for Sick Children, Toronto.

1. Ishikawa, H. (1979) in Cell Motility: Molecules and Organiza-tion, eds. Hatana, H., Ishikawa, H. & Sato, H. (Univ. of To-kyo Press, Tokyo), pp. 417-444.

2. French, S. W. & Davies, P. K. (1975) Gastroenterology 68,765-774.

3. Brandon, D. L. (1976) Eur. J. Biochem. 65, 139-146.4. Toh, B. H., Tildiz, A., Sotelo, J., Osung, O., Holborow, E. H.

& Fairfax, A. (1976) Cell Tissue Res. 199, 117-126.5. Yousef, I. M. & Murray, R. K. (1978) Can. J. Biochem. 56,

712-721.6. Oda, M., Price, V., Fisher, M. M. & Phillips, M. J. (1974)

Lab. Invest. 31, 314-323.7. Tamura, J., Kuroda, H., Watanabe, S. & Yokoi, Y. (1981)

Acta Histochem. Cytochem. 14, 661-669.

8. Gabbiani, G., Montesano, R., Tuchweber, B., Salas, M. &Orci, L. (1975) Lab. Invest. 33, 562-569.

9. Elias, E., Hruban, Z., Wade, J. B. & Boyer, J. L. (1980) Proc.Natl. Acad. Sci. USA 77, 2229-2233.

10. Adelstein, R. S. (1982) Fed. Proc. Fed. Am. Soc. Exp. Biol.41, 2873-2878.

11. Pollard, T. D. (1981) J. Cell Biol. 91, 156s-165s.12. Korn, E. D. (1982) Physiol. Rev. 66, 672-737.13. Mooseker, M. S., Bonder, E. M., Grimwade, B. G., Howe,

C. L., Keller, T. S. C., Wasserman, R. H. & Wharton, K. A.(1982) Cold Spring Harbor Symp. Quant. Biol. 46, 855-870.

14. Burgess, D. R. (1982) J. Cell Biol. 95, 853-863.15. Broschat, K. O., Stidwill, R. P. & Burgess, D. R. (1983) Cell

35, 561-571.16. Matsudaira, P. T. & Burgess, D. R. (1982) Cold Spring Harbor

Symp. Quant. Biol. 46, 845-854.17. Keller, T. C. S. & Mooseker, M. S. (1982) J. Cell Biol. 95,

945-959.18. Oshio, C. & Phillips, M. J. (1981) Science 212, 1041-1042.19. McLean-Fletcher, S. & Pollard, T. D. (1980) Cell 20, 329-341.20. Levin, R. M. & Weiss, B. (1976) Mol. Pharmacol. 12, 581-589.21. Osborn, M. & Weber, K. (1980) Exp. Cell Res. 130, 484-488.22. Simons, T. J. B. (1982) J. Membr. Biol. 66, 235-247.23. Seglen, P. 0. (1976) in Methods in Cell Biology, ed. Prescott,

D. M. (Academic, New York), Vol. 8, pp. 39-81.24. Laishes, B. A. & Williams, G. M. (1976) In Vitro 12, 521-532.25. Phillips, M. J., Oshio, C., Miyairi, M., Katz, H. & Smith,

C. R. (1982) Hepatology 2, 763-768.26. Stockem, W., Weber, K. & Wehland, J. (1978) Cytobiologie

18, 114-131.27. Graessmann, M. & Graessmann, A. (1976) Proc. Natl. Acad.

Sci. USA 73, 366-370.28. Hanzon, V. (1952) Acta Physiol. Scand. 28, Suppl. 101, 1-268.29. Miyairi, M., Oshio, C., Watanabe, S., Smith, C. R., Yousef, I.

& Phillips, M. J. (1984) Gastroenterology, in press.30. Phillips, M. J., Oshio, C., Miyairi, M. & Smith, C. R. (1983)

Lab. Invest. 48, 205-211.31. Yerna, M. J., Debrowska, R. M., Hartstone, D. L. & Gold-

man, R. D. (1979) Proc. Natl. Acad. Sci. USA 76, 184-188.32. Hathaway, D. R. & Adelstein, R. S. (1979) Proc. Natl. Acad.

Sci. USA 76, 1653-1657.33. Charest, R., Blackmore, P. F., Berthon, B. & Exton, J. R.

(1983) J. Biol. Chem. 258, 8769-8773.34. Means, A. E. & Dedman, J. K. (1980) Nature (London) 285,

250-252.35. Cheung, W. Y. (1980) Science 207, 19-27.36. Weiss, B., Prozialeck, W., Cimino, N., Barnette, M. S. &

Wallace, T. L. (1980) Ann. NYAcad. Sci. 356, 319-345.37. Watanabe, S., Miyairi, M., Oshio, C., Smith, C. R. & Phillips,

M. J. (1983) Gastroenterology 85, 245-253.38. Rose, B. & Loewenstein, W. R. (1975) Nature (London) 254,

250-252.

Proc. NaM Acad Sci. USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Aug

ust 1

6, 2

020