Proc. Natl. Acad. Sci. USAVol. 85, pp. 8673-8677, November 1988Medical Sciences

Biochemical localization of hepatic surface-membraneNa',K+-ATPase activity depends on membrane lipid fluidity

(Na+/KV-transporting ATPase/sinusoidal membrane/bile canalicular membrane)

E. SUTHERLAND*, B. S. DIXON*, H. L. LEFFERTt, H. SKALLYt, L. ZACCARO*, AND F. R. SIMON**Department of Medicine and Hepatobiliary Research Center, University of Colorado School of Medicine, Denver, CO 80262; and tDepartment ofPharmacology, University of California at San Diego, La Jolla, CA 92093

Communicated by Isidore Edelman, August 9, 1988

ABSTRACT Membrane proteins of transporting epitheliaare often distributed between apical and basolateral surfaces toproduce a functionally polarized cell. The distribution ofNa',K+-ATPase [ATP phosphohydrolase (Na+/K+-trans-porting), EC 3.6.1.37] between apical and basolateral mem-branes of hepatocytes has been controversial. BecauseNa',K+-ATPase activity is fluidity dependent and the physio-chemical properties ofthe apical membrane reduces its fluidity,we investigated whether altering membrane fluidity mightuncover cryptic Na',K+-ATPase in bile canalicular (apical)surface fractions free of detectable Na',K+-ATPase and glu-cagon-stimulated adenylate cyclase activities. Apical fractionsexhibited higher diphenylhexatriene-fluorescence polarizationvalues when compared with sinusoidal (basolateral) membranefractions. When 2-(2-methoxyethoxy)ethyl 8-(cis-2-n-octylcyclopropyl)octanoate (A2C) was added to each fraction,Na',K+-ATPase, but not glucagon-stimulated adenylate cy-clase activity, was activated in the apical fraction. In contrast,further activation of both enzymes was not seen in sinusoidalfractions. The A2C-induced increase in apical Na',K'-ATPaseapproached 75% of the sinusoidal level. Parallel increases inapical Na',K+-ATPase were produced by benzyl alcohol andTriton WR-1339. All three fluidizing agents decreased theorder component of membrane fluidity. Na',K+-ATPase ac-tivity in each subfraction was identically inhibited by themonoclonal antibody 9-A5, a specific inhibitor of this enzyme.These findings suggest that hepatic Na',K+-ATPase is distrib-uted in both surface membranes but functions more efficientlyand, perhaps, specifically in the sinusoidal membranes becauseof their higher bulk lipid fluidity.

Membrane proteins that polarize function across cell surfacesof transporting epithelia are generally distributed asymmet-rically at the cellular apical or basolateral pole (1-3). Forexample, leucine aminopeptidase is present on the apicalmembranes of enterocytes, proximal renal tubules, andhepatocytes (3-8), whereas IgA receptor (secretory compo-nent) is found in basolateral surfaces of most epithelial cells(9). On the other hand, Na',K+-transporting ATPase [ATPphosphohydrolase (Na+/K+-transporting); EC 3.6.1.37] isgenerally believed to be localized to the basolateral surface ofrenal (10), intestinal (11), and cultured MDCK (kidney) cells(12, 13) but on the apical surface of the choroid plexus (14)and possibly on both poles of the exorbital and parotid glandcells (15-17).The localization of Na',K+-ATPase in hepatocytes has

been controversial (18). With histochemical, biochemical,and cell-fractionation techniques, Na',K+-ATPase activityhas been seen in the basolateral membrane (19-22), whereasimmunocytochemistry using a Na',K+-ATPase-specific an-tibody has localized Na+/K+-pump sites to the bile canalic-

ular (or apical) surface, as well as the sinusoidal (or basolat-eral) surface (23, 24). Contradictory results also have beenreported with these techniques (25-27).

Activities ofmany intrinsic membrane enzymes depend onthe composition of lipids surrounding them and the generalphysical state of these membrane lipids (28-30). For exam-ple, Na',K+-ATPase activity has been directly related tomembrane fluidity (31-33). Because differences in the lipidand biophysical properties of hepatic apical and basolateralsurfaces result in a less-fluid apical surface, we wonderedwhether the biochemical localization of Na',K+-ATPaseactivity in sinusoidal surfaces might be due to the relativelydecreased fluidity of bile canalicular membranes (BCMs).The results of our investigation indicate that Na',K+-pumpa subunits are present in BCM fractions in a cryptic form andthat they become functionally active when canalicular mem-brane lipid fluidity is elevated in vitro to levels seen for thesinusoidal surface.

MATERIALS AND METHODSReagents. All inorganic chemicals or solvents were pur-

chased from Fisher and were the highest grade available.2-(2-Methoxyethoxy)ethyl 8-(cis-2-n-octylcyclopropyl)octa-noate (A2C) was obtained from Sigma and Triton WR-1339(oxyethylated tertiary octylphenol-polymethylene polymer)was obtained from Ruger (Irvington, NJ).Plasma Membrane Isolation. Male Sprague Dawley rats

(Harlan, Indianapolis, IN) weighing 180-220 g were usedafter overnight fasting. Liver plasma membrane (LPM)subfractions were prepared as described (34). Sinusoidalmembrane (SM) and BCM fractions were isolated concom-itantly from rat-liver homogenates by sucrose density cen-trifugation and Mg2+ precipitation (15 mM), respectively.Enzyme Studies. Na',K+-ATPase and Mg2+-ATPase were

measured after overnight freeze-thawing, by means of anenzyme-coupled kinetic assay with pyruvate kinase andlactate dehydrogenase, assuming ouabain (2.5 mM) inhibitionmeasures the Na+/K' pump (35). Other enzymes wereassayed according to standard procedures: alkaline phospha-tase (36), leucine aminopeptidase (37), ouabain-sensitiveK+-dependent p-nitrophenyl phosphatase (38), and adenyl-ate cyclase (39, 40). Protein was measured using bovineserum albumin (Sigma) as standard (41).

Fluorescence Polarization Studies. Fluorescence polariza-tion (P), angle of hindrance (roo), and time offluorescence (Tf)measurements were done on a model 4800 polarizationspectrofluorometer (SLM Industries, Urbana, IL) with fixed

Abbreviations: A2C, A2C membrane mobility agent, 2-(2-meth-oxyethoxy)ethyl 8-(cis-2-n-octylcyclopropyl)octanoate; BCM, bilecanalicular membrane; SM, sinusoidal membrane; LPM, liverplasma membrane; DPH, 1,6-diphenyl-1,3,5-hexatriene; P, polariza-tion; roo, angle of hindrance; Tf, time of fluorescence; mAb, mono-clonal antibody.

8673

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.

8674 Medical Sciences: Sutherland et al.

a

E

._u

A B C D

emission and excitation filters as described (34). Fluores-cence intensity was measured perpendicularly and parallel tothe polarization phase of the excitation light [excitationwavelength, 360nm; KV 389 emission filters (Schott, Duryea,PA)]. Measurements were made at 370C. 1-6-Diphenyl-1,3,5-hexatriene (DPH) (Molecular Probes, Junction City, OR) wasdissolved in tetrahydrofuran (0.3 gg of DPH/ml) and addedto LPM subfractions (72 ttg of protein) directly beforeaddition ofeither A2C (in dimethyl sulfoxide), benzyl alcohol,or Triton WR-1339 in a final total volume of 1.3 ml of 1 mMNaHCO3. Concentration of each test chemical is given inResults. Samples were mixed for no longer than 15 min, andmeasurements were done in triplicate.

Statistical Analysis. All results are expressed as mean +

SE. A two-tailed Student's t test was used to comparedifferences between mean values. Linear regression wascalculated by regression analysis.

RESULTS

Relative enrichment of putative plasma membrane markerenzymes for LPM subfractions is shown in Fig. 1. Na',K+-ATPase and glucagon-stimulated adenylate cyclase activitieswere enriched (43 ± 2)-fold and (17 + 3)-fold, respectively,in SM fractions. In contrast, these enzyme activities were notdetected in canalicular fractions. The BCM enzyme markersleucine aminopeptidase and Mg2+-ATPase were increased(34 + 4)- and (46 ± 5)-fold, respectively. Leucine aminopep-tidase and Mg2+-ATPase were modestly increased (8- and9-fold, respectively) in the SM fraction. Marker enzymes ofintracellular organelles were enriched 2-fold or less in bothfractions (data not shown). Total recovery of Na',K+-ATPase was 85 ± 5% with 32 ± 5% accounted for in the SMfraction. Thus, BCM fractions prepared under these condi-tions were devoid of detectable Na' ,K+-ATPase activity andanother predominantly SM enzyme marker-adenylate cy-clase (42-44).

FIG. 1. Relative enrichment of markerenzymes of LPM subfractions. m, SM; *,BCM. Enrichment equals plasma membraneactivity divided by homogenate activity.Specific activities of homogenates are:Na',K+-ATPase (A) (n = 16), 1.1 ± 0.3tmol/hr per mg of protein; glucagon-stimulated adenylate cyclase (B) (n = 3),14.6 ± 5.1 pmol/min per mg of protein;leucine aminopeptidase (C) (n = 16), 5.5 ±

1.1 ,umol/hr per mg of protein; and Mg2+-ATPase (D) (n = 16), 2.6 ± 0.4 ,umol/hr permg of protein.

Fluorescence-polarization values were significantly higherfor BCM fractions than for SM preparations (Table 1).Fluidity measured with DPH is determined largely by astructural component (28). Although Tf for BCM fractionswas longer than Tf for SM fractions, roo was significantlyincreased, indicating that the BCM fraction was significantlyless fluid than the SM fraction.

Fluidizing agents are known to expand the packing order ofmembrane lipids and thus increase the motional freedom ofthe phospholipid fatty acid acyl chains (45). Fig. 2 shows theeffect of additions of the synthetic ester A2C (46) on P valuesfor BCM and SM fractions. A2C increased membrane fluidityin a linear fashion in both LPM subfractions. Addition of 7.7,uM A2C to BCM fractions reduced DPH P values to thosemeasured in the basal state of the SM fraction (Table 1).Although time-of-fluorescence values in these studies wereminimally but significantly reduced, roo measurementsshowed that the BCM is less ordered after treatment withA2C. Benzyl alcohol and A2C similarly decreased Tf and roo

in BCM fractions. Measurements of the effect of TritonWR-1339 on membrane fluidity could not be made becausethe detergent markedly quenched fluorescence. However,previous electron spin resonance studies with LPM fractionshave shown that Triton WR-1339 also decreased orderparameter (47).

Fig. 3 shows the effect of in vitro addition of A2C onNa+,K+-ATPase activity in SM and BCM fractions. As BCMfractions were made more fluid by A2C, Na+,K+-ATPaseactivity increased progressively, reaching a plateau at 3.3,uM, corresponding to aP value of0.276 ± 0.003. In contrast,Na+,K+-ATPase activity in SM remained stable after theaddition of A2C over the same concentration range. Similareffects of A2C on Na+,K+-ATPase activity were also seenwhen this same enzyme was measured with the phosphate-release assay (47) (data not shown).To determine whether this increase in Na+,K+-ATPase

activity in the BCM fraction was related to the ability of A2C

Table 1. Fluorescence polarization studies of LPM subfractionsMembrane fractions (n) Tf, nsec P roo,0

SM (6) 8.5 ± 0.2 0.254 ± 0.006 0.162 ± 0.003BCM (7) 9.8 ± 0.1* 0.303 ± 0.0002* 0.210 ± 0.002*+ 7.7 ,um A2C (3) 9.1 ± Olt 0.258 ± 0.OOlt 0.172 ± 0.001t+ 95 mM benzyl alcohol (4) 9.0 ± 0.1t 0.231 ± 0.003t 0.143 ± 0.002tParentheses indicate number of separate determinations. Fluidity parameters were determined at

370C using DPH as the probe. The data are means ± SE.*Probability < 0.001 compared with SM.tProbability < 0.001 compared with BCM.

Proc. Natl. Acad. Sci. USA 85 (1988)

Proc. Natl. Acad. Sci. USA 85 (1988) 8675

T

T

BILE CANAUCULAR

T

9 ~~~~~~~~~~\SINUSOIDAL

i~.

0 2.0 5.0 8.0 11.0A2C, AM

FIG. 2. Effect of A2C on SM and BCM fluidity. DPH-fluorescence polarization was measured at 370C as described. Linearregression curves were as follows: BCM (e) (n = 5), y = -0.005x +0.2% and r = 0.994°, SM (o) (n = 5), y = -0.OOSx + 0.236 and r =0.969°.

to increase lipid fluidity or rather to nonspecific activation,other fluidizing agents were tested. At their optimal concen-trations, benzyl alcohol (95 mM) activated Na+,K+-ATPaseto 69 ± 7% and Triton WR-1339 (0.015%) to 77 ± 6% of SMenzyme activity. Neither one of these agents increasedNa',K+-ATPase activity in the SM fraction. Taken together,these results indicated that activation of cryptic Na+,K+-ATPase activity in BCM fractions was independent of thestructure offluidizing agents and acted somewhat specificallyon the BCM fraction.To investigate whether these increases in BCM Na+,K+-

ATPase activity induced by fluidizing agents might have beendue partly to detergent-like increases in the accessibility ofsubstrate, the effects of A2C were compared to the additionof 0.1% Triton X-100 on the activities of Na',K+-ATPase,Mg2+-ATPase, ouabain-sensitive K+-stimulated p-nitro-phenyl phosphatase, leucine aminopeptidase, and alkalinephosphatase (Table 2). The addition of A2C (6 AM) to BCMfractions selectively increased Na+,K+-ATPase and oua-bain-sensitive K+-p-nitrophenyl phosphatase, whereas otherenzyme activities were unaltered. In contrast, Triton X-100(0.1%) did not change either Na+,K+-ATPase or ouabain-sensitive K+-p-nitrophenyl phosphatase, but Mg2+-ATPasewas decreased.

Hormone-stimulatable adenylate cyclase activity is alsoinfluenced by the physical state of membrane lipids (48).Hepatic glucagon receptors are linked functionally to thecatalytic unit through membrane-associated GTP-bindingproteins (49); therefore, the effects ofGTP and glucagon werealso studied. The effects ofA2C on adenylate cyclase activityare presented in Table 3. In SM fractions, basal adenylatecyclase activity was increased 2.5-fold by glucagon alone and15-fold by glucagon plus GTP. This response was unalteredby A2C (6 uM) in the SM fraction. In contrast, in bothuntreated and A2C-fluidized BCM fractions, neither glucagon

FIG. 3. Effect of A2C on Na',K+-ATPase activity in LPMsubfractions. SM (0) and BCM (e) fractions were isolated asdescribed. Na',K+-ATPase activity was measured at 37TC 5 minafter addition of A2C to LPM subfractions (67 gg of protein/ml) withvigorous mixing.

nor GTP-stimulated adenylate cyclase activities were mea-surable to significant levels. In additional experiments, for-skolin (10' M) failed to activate the trace levels of catalyticsubunit in BCM fractions, even after addition of 6 uM A2C(data not shown).A monoclonal antibody (mAb) (9-A5), shown previously to

inhibit the a subunit of rat hepatic and renal Na',K+-ATPase, was used to examine whether BCM Na',K+-ATPase activity was immunochemically identical to theNa',K+-ATPase activity of a subunits seen in SM fractions(23, 50). In both SM and BCM fractions treated with A2C,mAb 9-A5 inhibited Na',K+-ATPase activity similarly (Fig.4). mAb 9-A5 had no effects on Mg2+-ATPase activity (datanot shown). The ID50 values for SM and BCM Na',K+-ATPase inhibition were 90 ± 7 and 93 ± 7 ng of mAb 9-A5per,4g of membrane protein, respectively.

DISCUSSIONThe purpose of this study was twofold. First, to find whetherdifferences in membrane lipid fluidity regulate, in part,Na',K+-ATPase activity. Because the apical domain ofepithelial cells is well known to be less fluid than thebasolateral surfaces (28), we proposed that the decreasedBCM lipid fluidity masks a cryptic Na+/K+-ATPase cata-lytic activity. Our second aim was to resolve the controversyregarding the polarized distribution of the Na+/K' pumps inhepatocytes. This report presents biochemical and immuno-chemical evidence that the catalytic subunits of Na',K+-ATPase are present in bile canalicular as well as in sinusoidalliver plasma membrane subfractions. These observationsindicate that previous controversial findings in situ showingimmunologically reactive Na',K+-ATPase a subunits aswell as enzyme cytochemistry showing no Na',K+-ATPasecatalytic activity in BCM are both correct.

Table 2. Effect of agents on enzyme-specific activities in bile canalicular membrane fractionsBCM enzyme-specific activity*

Enzymes Basal A2C (6 uM) Triton X-100 (0.1%)Na',K+-ATPase 0 20.6 0Mg2+-ATPase 109.2 100.5 31.1Ouabain-sensitive K+-p-nitrophenyl

phosphatase 0 2.23 0Leucine aminopeptidase 95.5 92.2 96.7Alkaline phosphataset 20.2 18.7 t

Values are reported as the means (n = 3-6); measurement errors differed by <10%o.*Enzyme activities reported as ymol of substrate hydrolyzed/hr per mg of protein.tMeasured routinely with 0.1% Triton.

9')

a,

c

5D

I

0.310 -

0.290-

0.270

0.250

0.230

0.210

0.190

0.170

30.0-

0.5Za E 20.0-cLEam I00-,,a

co 8+ E 10.0+Y z

z

4.0

A2C, AM

Medical Sciences: Sutherland et al.

8676 Medical Sciences: Sutherland et al.

Table 3. Adenylate cyclase-specific activity in LPM subfractionsAdenylate cyclase specific activity,

pmol/mg of protein per min

Hormone/ SM BCMnucleotide -A2C +A2C -A2C +A2CNone 9.2 ± 0.4 8.6 ± 0.9 2.6 ± 0.4 1.3 ± 0.3Glucagon 23.4 ± 3.6 19.8 ± 1.8 1.3 ± 0.4 0.7 ± 0.1Glucagon+ GTP 126.6 ± 7.4 119.4 ± 7.2 2.4 ± 0.7 1.1 ± 0.6n, Three separate determinations presented as mean ± SE.

Glucagon (10-6 M) and GTP (10-' M) were added to separate assays,

Three independent lines of evidence suggest Na+,K+-ATPase is located on the bile canalicular domain: (i) increas-ing BCM fluidity by different agents activates Na+,K+-ATPase; (ii) increases in fluidity in these membranes selec-tively activates Na+,K+-ATPase and in parallel ouabain-sensitive K+-p-nitrophenyl phosphatase activity; and (iii)monospecific anti-a subunit mAb 9-A5 inhibits BCM and SMNa+,K+-ATPase activity with similar kinetics.Our preparations of BCM were apparently devoid of

contaminating SM vesicles, which might be attributed to theisolation of these BCM fractions by Mg2+-precipitation aftervigorous homogenization (34). Such preparations yield aBCM fraction devoid of detectable Na+,K+-ATPase andadenylate cyclase (either glucagon- or GTP-stimulated) ac-tivity but are still highly enriched in enzyme markers locatedat the biliary pole (21). In addition, fluorescence p measure-ments on BCM fractions indicate that their lipid structurebehaves rigidly, unlike the more fluid SM fraction (28).

Addition ofA2C, benzyl alcohol, and Triton WR-1339 eachselectively increased Na+,K+-ATPase activity in the BCM.All these agents are structurally different; yet, their effectsupon the BCM fractions were similar. It would appear,therefore, that the decreased membrane fluidity of untreatedBCM fractions normally does mask a cryptic Na+,K+-ATPase activity. A cryptic Na+,K+-ATPase, by this crite-rion, seems absent from SM fractions, because when SMfluidity increases beyond its apparent physiological level, nostimulation of Na+,K+-ATPase occurred. Previous workalso demonstrated that benzyl alcohol increased Na+,K+-ATPase activity and decreased order parameter (51), butinterpretation ofthese results was complicated because oftheuse of mixed-LPM fractions.

Previous studies from our laboratory have shown thatdetergents increase Na+,K+-ATPase activity in freshly an-alyzed preparations, but do not increase this activity after

100.04

1ool*f 80.0

(a

EDui 60.0

a.

> 40.0-.-0)

*. 20.0

0

0.0*

0.01 0.10 1.00

Antibody added to assay, Ag

FIG. 4. Inhibition of Na',K+-ATPase activity by mAbs in LPMsubfractions. Affinity-purified mAb 9-A5 was added to SM and BCMfractions after treatment with 6 gM A2C. After 5 min at 22TC,Na',K+-ATPase activity was measured as described. o, SM; e,

BCM. Each point is the mean SD. (n = 3).

freeze-thawing (52). In experiments reported here, additionof 0.1% Triton X-100 to both membrane fractions afterfreeze-thawing did not further increase Na',K+-ATPaseactivity. Moreover, ouabain-sensitive K+-p-nitrophenylphosphatase, which is closely associated with Na',K+-ATPase and reflects an externally oriented enzymatic activ-ity (53), showed similarly increased ouabain-sensitive K+-p-nitrophenyl phosphatase activity with A2C but not TritonX-100 in BCM fractions. Such lipid fluidity regulation may besomewhat specific for Na',K+-ATPase, as leucine amino-peptidase and Mg2+-ATPase, BCM ectoenzymes, were notaltered by A2C (54).An alternative interpretation of these findings is that

Na',K+-ATPase undergoes lateral movement from its SMlocation to the BCM during membrane isolation (5, 55). Toinvestigate this possibility, basal and stimulated adenylatecyclase activities were measured in LPM subfractions. Nofunctional components (glucagon receptors, catalytic sub-units, or G proteins) of this complex were identified in theBCM fraction. Thus, glucagon-stimulated adenylate cyclasewas localized to the SM fraction, a finding indirectly sup-porting the conclusion that SM proteins do not redistributeduring the isolation procedure.A recent study, however, failed to identify Na',K+-

ATPase in BCM by either immunocytochemistry or immu-noblotting (27); the reasons for the negative results areunclear. With regard to the immunocytochemical findings(27), it is likely that the positive epitopes were limited by thestability, conformational state, or accessibility of their anti-bodies to Na+,K+-ATPase (18, 23, 50). Failure to demon-strate Na',K+-ATPase catalytic subunits in the BCM frac-tion by immunoblotting may be due to isolation of BCMfractions in low yield (21). Because apical enzymes may belocalized in microdomains (56, 57), these workers could haveselectively discarded the BCM fraction containing Na+,K'-ATPase. In addition to Na+,K+-ATPase their study alsofailed to identify secretory component, a receptor that hasbeen shown to reside, at least transiently, in the BCM beforespecific proteolysis (58).

Additional experimental findings in the present work alsodemonstrated similar inhibition curves for Na+,K+-ATPaseactivities in A2C-activated BCM and SM fractions usingmAbs against the Na+/K+ pump. Thus, immunologicallyidentical protein epitopes are present on both surfaces ofLPM subfractions, strongly suggesting that the Na+,K+-ATPase activities in each fraction are probably due, at leastin part, to similar a-subunit isoforms.These studies demonstrate the functional importance of

membrane lipid fluidity in governing hepatic Na+,K+-ATPase activity. Cryptic Na+,K+-ATPase activity in thecanalicular membrane domain raises the possibility thathepatic Na+/K+ pump activity is activated without any netincrease in number ofpump sites-i.e., by increasing plasmamembrane fluidity. Indeed, we recently demonstrated thischange with bile duct obstruction (59) and, in addition,hepatic regeneration is associated with increased Na+,K+-ATPase activity, bile flow, and plasma membrane fluiditywithout any increase in the number of enzyme units (60-62).

These studies were supported by Veterans Administration andPublic Health Service Grants AM-15851 and AM-34914 to F.R.S.;and AM-28215, John Simon Guggenheim Foundation, Mortin IGrossman Scholar Award, and National Science Foundation (DCB86-16740) to H.L.L.

1. Evans, W. H. (1980) Biochim. Biophys. Acta 604, 27-64.2. Almers, W. & Stirling, C. (1984) J. Membr. Biol. 77, 169-186.3. Simons, K. & Fuller, S. D. (1985) Annu. Rev. Cell Biol. 1, 243-

288.4. Vannier, C. H., Louvard, D., Maroux, S. & Desnuelle, P.

(1976) Biochim. Biophys. Acta 455, 185-199.

Proc. Natl. Acad. Sci. USA 85 (1988)

Proc. Natl. Acad. Sci. USA 85 (1988) 8677

5. Ziomek, C. A., Schulman, S. & Edidin, M. (1980) J. Cell Biol.86, 849-857.

6. Tanaka, K., Omori, K. & Tashiro, Y. (1986) J. Histochem.Cytochem. 34, 775-784.

7. Roman, L. M. & Hubbard, A. L. (1983) J. Cell Biol. 96, 1548-1558.

8. Roman, L. M. & Hubbard, A. L. (1984) J. Cell Biol. 98, 1488-14%.

9. Ahnen, D. J., Brown, W. R. & Kloppel, T. M. (1985) Gastro-enterology 89, 667-682.

10. Kyte, J. (1976) J. Cell Biol. 68, 287-303.11. Stirling, C. E. (1972) J. Cell Biol. 53, 704-714.12. Lamb, J. F., Ogden, P. & Simmons, N. L. (1981) Biochim.

Biophys. Acta 644, 333-340.13. Caplan, M. J., Anderson, H. C., Palade, G. E. & Jamieson,

J. D. (1986) Cell 46, 623-631.14. Ernst, S. A., Palacios, J. R., II & Siegel, J. J. (1986) J.

Histochem. Cytochem. 34, 189-195.15. Mircheff, A. K. & Lu, C. C. (1984) Am. J. Physiol. 247, G651-

G661.16. Contreas, C. N., McDonough, A. A., Kolowski, T. R., Hens-

ley, C. B., Wood, R. L. & Mircheff, A. K. (1986) Am. J.Physiol. 250, C430-C441.

17. Smith, Z. D. J., Caplan, M. J., Forbush, B., III & Jamieson,J. D. (1987) Am. J. Physiol. 253, G99-G109.

18. Leffert, H. L., Schenk, D. B., Hubert, J. L., Skelly, H.,Schumacher, M., Ariyasu, R., Ellisman, M., Koch, K. S. &Keller, G. A. (1985) Hepatology 5, 501-507.

19. Latham, P. S. & Kashgarian, M. (1979) Gastroenterology 76,988-996.

20. Blitzer, B. L. & Boyer, J. L. (1978) J. Clin. Invest. 62, 1104-1108.

21. Meier, P. J., Sztul, E. S., Reuben, A. & Boyer, J. (1984) J. CellBiol. 98, 991-1000.

22. Poupon, R. E. & Evans, W. H. (1979) FEBS Lett. 108, 374-378.

23. Schenk, D. & Leffert, H. (1980) Proc. Natl. Acad. Sci. USA 80,5281-5285.

24. Takemur, S., Omari, K. & Tanaka, K. (1984) J. Cell Biol. 99,1562-1570.

25. Toda, G., Oka, H., Oda, T. & Ikeda, Y. (1975) Biochim.Biophys. Acta 413, 52-64.

26. Kakis, G., Phillips, M. J. & Yousef, I. M. (1980) Lab. Invest.43, 73-81.

27. Sztul, E. S., Biemesderfer, D., Caplan, M. J., Kashgarian, M.& Boyer, J. L. (1987) J. Cell Biol. 104, 1239-1248.

28. Schachter, D. (1984) Hepatology 4, 140-151.29. Sandermann, H. (1978) Biochim. Biophys. Acta 515, 209-237.30. Spector, A. A. & Yorek, M. A. (1985) J. Lipid Res. 26, 1015-

1035.31. Sinensky, M., Pinkerton, F., Sutherland, E. & Simon, F. R.

(1979) Proc. Natl. Acad. Sci. USA 76, 4893-4897.32. Kimelberg, H. K. & Paphadjopoulos, D. (1974) J. Biol. Chem.

249, 1071-1080.33. Chong, P. L.-G., Fortes, P. A. G. & Jameson, D. M. (1985) J.

Biol. Chem. 260, 14484-14490.34. Rosario, J., Sutherland, E., Zaccaro, L. & Simon, F. R. (1988)

Biochemistry 27, 3939-3946.35. Schoner, W., von Ilberg, C., Kramer, R. & Seubert, W. (1967)

Eur. J. Biochem. 1, 334-343.36. Bessy, 0. A., Lowry, 0. H. & Brock, M. J. (1946) J. Biol.

Chem. 164, 321-329.37. Goldberg, J. A. & Ratenburg, A. M. (1958) Cancer 11, 283-

291.38. Rodriguez, H. J., Hogan, W. C., Sinha, S. K., Jacobson,

M. P. & Klat, S. (1981) Biochim. Biophys. Acta 641, 36-54.39. Iyengart, R., Mintz, P. W., Swartz, T. L. & Birnbaumer, L.

(1980) J. Biol. Chem. 255, 11875-11882.40. Salomon, Y. (1979) Adv. Cyclic Nucleotide Res. 10, 35-55.41. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.

(1951) J. Biol. Chem. 193, 265-275.42. Cheng, H. & Farquhar, M. G. (1976) J. Cell Biol. 70, 660-670.43. Blitzer, B. L. & Donovan, C. B. (1984) J. Biol. Chem. 259,

9295-9301.44. Reik, L., Petzold, G. L., Higgins, J. A., Greengard, P. &

Bannen, R. J. (1970) Science 168, 382-386.45. Seeman, P. (1972) Pharmacol. Rev. 24, 583-653.46. Kosower, E. M., Kosower, N. S. & Wegman, P. (1977) Bio-

chim. Biophys. Acta 471, 311-329.47. Davis, R. A., Kern, F., Jr., Showalter, R., Sutherland, E.,

Sinensky, M. & Simon, F. R. (1978) Proc. Natl. Acad. Sci.USA 75, 4130-4134.

48. Houslay, M. D. & Gordon, L. M. (1983) Curr. Top. Membr.Transp. 18, 179-281.

49. Casperson, G. F. & Bourne, H. R. (1987) Annu. Rev. Phar-macol. Toxicol. 27, 371-384.

50. Hubert, J. J., Schenk, D. B. & Leffert, A. L. (1986) Biochem-istry 25, 4156-4163.

51. Gordon, L. M., Suuerheber, R. D., Esgate, J. A., Dipple, R. J.& Houslay, M. D. (1980) J. Biol. Chem. 255, 4519-4527.

52. Molitoris, B. A. & Simon, F. R. (1985) J. Membr. Biol. 83, 207-245.

53. Ernst, S. A., Riddle, C. V. & Karnaky, K. J. (1980) Curr. Top.Membr. Transp. 13, 355-385.

54. Trams, E. G. & Lauter, C. J. (1974) Biochim. Biophys. Acta345, 180-197.

55. Jesaitis, A. & Yguerabide, J. (1986) J. Cell Biol. 102, 1256-1263.56. Matsuura, S., Eto, S., Kato, K. & Yashiro, Y. (1984) J. Cell

Biol. 99, 166-173.57. Kejaschki, D., Noronha-Blob, L., Sacktor, B. & Farquhar,

M. G. (1984) J. Cell Biol. 98, 1505-1513.58. Musil, L. S. & Baemziger, J. U. (1987) J. Cell Biol. 104, 1725-

1733.59. Nibel, D., Sutherland, E., Zaccaro, L. & Simon, F. R. (1988)

Hepatology, in press.60. Schenk, D. B., Hubert, J. J. & Leffert, H. L. (1984) J. Biol.

Chem. 259, 14941-14951.61. Bruscalupi, G., Caratola, G., Lenaz, G., Leoni, S., Mangian-

tini, M. T., Mazzanti, L., Spagnuolo, S. & Trentalance, A.(1980) Biochim. Biophys. Acta 597, 263-273.

62. Leong, G. F., Pessotti, R. L. & Brauer, R. W. (1959) Am. J.Physiol. 197, 880-887.

Medical Sciences: Sutherland et al.