Binding Forces of Hepatic Microsomal and PlasmaMembrane Proteins in Normal and Pancreatitic Rats:An AFM Force Spectroscopic StudyLORI A. SLEZAK,1 ANTHONY S. QUINN,2 KUMUDESH C. SRITHARAN,1 JIN PING WANG,1GUDRUN ASPELUND,1 DOUGLAS J. TAATJES,2 AND DANA K. ANDERSEN1*1Department of Surgery, Yale University School of Medicine, New Haven, Connecticut 065102Department of Pathology and Cell Imaging Facility, University of Vermont, Burlington, Vermont

KEY WORDS chronic pancreatitis; liver microsome; membrane binding; atomic force micros-copy

ABSTRACT The docking and fusion of membrane-bound vesicles at the cell plasma membraneare brought about by several participating vesicle membrane, plasma membrane, and solublecytosolic proteins. An understanding of the interactions between these participating proteins willprovide an estimate of the potency and efficacy of secretory vesicle docking and fusion at the plasmamembrane in cells of a given tissue. Earlier studies suggest that in chronic pancreatitis, glucoseintolerance may be associated with impaired exocytosis/endocytosis of hepatic insulin receptor andglucose transporter proteins. In this study, the binding force profiles between microsome membraneproteins and plasma membrane proteins in liver obtained from normal and pancreatitic rats havebeen examined using atomic force microscopy. The ability of a VAMP-specific antibody to alterbinding between microsome- and plasma membrane-associated membrane proteins was examined.In pancreatitic livers, a significant loss in microsome-plasma membrane binding is observed.Furthermore, our study shows that, in contrast to control livers, the microsome-plasma membranebinding in pancreatitic livers is VAMP-independent, which suggests an absence of VAMP participa-tion in membrane-microsome binding. In confirmation with our earlier findings, these studiessuggest altered membrane recycling in liver of rats with chronic pancreatitis. Microsc. Res. Tech.44:363–367, 1999. r 1999 Wiley-Liss, Inc.

INTRODUCTIONChronic pancreatitis is often accompanied by glucose

intolerance and diabetes. Clinical studies as well asmodels of chronic pancreatitis show a loss of hepaticsensitivity to insulin, with resulting unsuppressed he-patic glucose production and hyperglycemia (Seymouret al., 1988). Loss of hepatic sensitivity to insulinresults, in part, from decreased synthesis and availabil-ity of insulin receptors on liver plasma membrane(Spector et al., 1997). In addition, the hepatic glucosetransporter protein GLUT 2, which mediates the up-take and release of glucose by hepatocytes under thedirect regulation of insulin, was also found to be alteredin animals with chronic pancreatitis (Andersen et al.,1994). Alterations in vesicular traffic, regulating theavailability of insulin receptor (IR) and the GLUT 2transporter in hepatocytes of chronic pancreatitic ani-mals, is therefore hypothesized. We have recently shownimpairment in both fluid-phase and receptor mediatedendocytosis in chronic pancreatitic rats (Nathan et al.,1998). IR and GLUT 2, located on the plasma mem-brane, are internalized in response to feeding; however,feeding-induced internalization of IR and GLUT 2 isimpaired in pancreatitic rats (Nathan et al., 1997;Slezak et al., 1998). This study was undertaken toexamine the possible impairment of membrane recy-cling due to impaired membrane fusion in chronicpancreatitis. In a recent study in neurons (Sritharan et

al., 1998), the atomic force microscope (AFM) was usedto study the force interactions between synaptosomalmembrane proteins and synaptic vesicle proteins, inthe presence and absence of antibodies to selectedproteins thought to serve as important components ofthe vesicle fusion process. Using this approach, theinteractive binding force that exists between micro-somal membranes under such in vitro conditions wasexamined and appears to be contributed to primarily byVAMP1, syntaxin, and SNAP-25 proteins. In the pres-ent study, we have used a similar approach to examinethe binding interactions between liver microsomal andplasma membrane proteins obtained from control andpancreatitic rats.

MATERIALS AND METHODSSprague-Dawley rats were obtained from Camm

Research Laboratory Animals, Wayne, NJ. Polyclonal

Lori A. Slezak and Anthony S. Quinn contributed equally to the manuscript.*Correspondence to: Dana K. Andersen, MD, Department of Surgery, Yale

University School of Medicine, 333 Cedar Street, New Haven, CT 06520–8062.E-mail: Dana.andersen@Yale.edu

Received ; accepted in revised formAbbreviations used: AFM, atomic force microscope; NSF, N-ethylmaleimide-

sensitive fusion protein; SNAP, NSF attachment protein; VAMP, vesicle associ-ated membrane protein or synaptobrevin.

Contract grant sponsor: United States Public Health Service; Contract grantnumber: DK39950.

MICROSCOPY RESEARCH AND TECHNIQUE 44:363–367 (1999)

r 1999 WILEY-LISS, INC.

monospecific synaptobrevin antibody (VAMP1 and 2)were purchased from StressGen Biotechnologies Co,BC, Canada. All other biochemicals were obtained fromSigma, St. Louis, MO.

Fractionation of Rat Liver to Obtain PlasmaMembrane and Microsomal-Enriched Fractions

All steps of tissue preparation and fractionationswere performed at 4°C, without the presence of ATP orATP regeneration systems. The fractions were thenflash frozen in liquid nitrogen, lyophilized, and storedat -20°C to be used at a later time.

Preparation of Sample for AFM ForceMeasurements

All experiments were performed at room tempera-ture. Lyophilized plasma membranes and microsomalfractions were resuspended in phosphate buffered sa-line (PBS). Five microliters of a 4 mg/ml plasmamembrane protein in PBS was applied to the center of afreshly cleaved mica disk adhered to a glass slide.Similarly, 1 µl of a 0.5 mg/ml microsomal proteinpreparation in PBS was applied to the AFM tip. Eachfraction was allowed to adsorb to the respective sur-faces for 2 minutes at room temperature, after whichnonbound excess protein was removed from the AFMtip and the mica surface by washing five times using200 volumes of PBS for each wash (Fig. 1). For eachexperiment, AFM cantilevers were functionalized withprotein only after force calibration had been performedon each of them.

AFM Binding Profile Between Plasma Membraneand Microsomal-Enriched Fractions

Contact AFM force plots were acquired using thesettings for contact AFM imaging and force calibrationprocedures (Sritharan et al., 1998). Two hundred mi-crometer oxide-sharpened silicon nitride pyramidalprobes, with a nominal apex radii of 10 nm and a springconstant of 0.12 N/m were used in our study. Initialengagement to the surface of the specimen was accom-plished by adjusting the setpoint value. Upon engage-ment, disengagement is brought about by decreasing

the setpoint by small increments. The gain is adjustedto stabilize the interaction of the probe with the speci-men. Tip calibration was performed in PBS at roomtemperature, using a freshly cleaved mica surface.While in the force calibration mode, the z-scan size,z-scan start, and setpoint parameters were adjusteduntil the force curve for a flat surface is obtained. Theslope of the force plots were calculated using theBioscope software. Calibration force plots of non-functionalized probes on mica in PBS, displayed noadhesions. Single extension-retraction cycles were em-ployed with the scan period ranging from 700 msec to1.10 sec as defined by the z-scan rate parameter tocomplete one approach and retract cycle. Several ap-proach and retraction cycles were initiated to deter-mine stability of the functionalized tip preparation andthe site on mica. When identical force curves wereobtained from contact between the functionalized tipand the functionalized mica surface, the site was cho-sen to examine changes in interactive forces whenexposed to VAMP antibody. The total tracking (contact)force and the adhesion (unbinding) forces were thendetermined.

RESULTSVAMP Immunoblots of Liver Homogenate

VAMP immunoblots of liver homogenates obtainedfrom control fed rats and fed pancreatitic animals,demonstrated the presence of VAMP immunoreactivity(Fig. 2). A loss of VAMP immunoreactivity is observed inpancreatitic liver.

Binding Profile Between Liver Microsomeand Plasma Membrane Proteins

When liver microsome protein-functionalized cantile-ver tip was allowed to make contact at various points onthe surface of plasma membrane-adsorbed mica placedin PBS, and then retracted, binding pattern was ob-served on the retraction path trace. The strength ofbinding at different points on the mica surface wasfound to vary, perhaps due to differing concentrations ofinteracting proteins. Binding between liver microsome

Fig. 1. Study of binding interactions be-tween liver microsomes and liver plasmamembrane proteins by AFM. Schematic draw-ing of the association of v-SNARE in micro-somes, and t-SNARE in the plasma mem-brane. When microsomal protein-functionalizedAFM cantilever tip is brought close to plasmamembrane proteins adsorbed on freshlycleaved mica (A), and allowed to make contact(B), protein-protein interactions are estab-lished. Subsequent withdrawal of the AFMtip from the mica surface (C) requires excessforce due to binding interactions establishedbetween the two sets of proteins, which can bemeasured by the AFM.

364 SLEZAK ET AL.

and plasma membrane proteins was reduced in chronicpancreatitic rats (46% of control) (Fig. 3).

Participation of VAMP in the BindingContribution Between Liver Microsome and

Plasma Membrane ProteinsTo understand if VAMP contributed to binding ob-

served between microsomal and the plasma membraneproteins, AFM binding studies between the two groupsof membrane proteins described earlier were performedin the presence and absence of specific antibodiesagainst VAMP protein. All binding experiments wereperformed in PBS, at room temperature. The micro-

somal protein-functionalized cantilever tip was allowedto make contact at a point on the surface of plasmamembrane-adsorbed mica in PBS, and then retracted.The AFM tip is approached and retracted several timesto obtain the same binding profile, to ensure thatadhesion of proteins from the mica surface onto theAFM tip and vice versa, does not occur. To such a stablepreparation, preimmune serum or specific VAMP anti-body is added, and the binding profiles examinedfollowing such addition. The binding strengths or pro-files were found unchanged following addition of up to100 µg/ml preimmune serum to the interacting proteinsin PBS. However, addition of 2 µg/ml of VAMP specificantibody resulted in 32% decrease in binding in controlanimals. VAMP antibody had no significant effect in thebinding pattern between microsomal and the plasmamembrane proteins isolated from pancreatitic livers(Fig. 3). Increasing the concentration of the specificantibodies in the experiment had no further effect onbinding.

DISCUSSIONIn the last decade, remarkable progress has been

made in the identification and biochemical characteriza-tion of some of the key components in secretory vesicledocking and fusion. Although the regulatory compo-nents of fusion may differ between intracellular com-partments or cell types, it appears that the fundamen-tal machinery and mechanism of fusion of twomembrane bilayers is highly conserved in all livingcells. Studies of vesicle fusion between purified Golgistacks by Rothman and colleagues (Malhotra et al.,1988;Rothman, 1994) led to the isolation and identification ofsoluble proteins necessary for intracellular membranefusion. These include the N-ethylmaleimide-sensitivefusion protein (NSF) and the soluble NSF attachmentproteins (SNAPs), which comprise a small family ofhomologous isoforms, designated as a, b, and g-SNAP(Wilson et al., 1989; Clary et al., 1990). Rothman andcolleagues also found a SNAP family of NSF attach-ment proteins in neurons (Whiteheart et al., 1993).NSF and SNAP homologues have now been identified inyeast, and demonstrated to be indispensable in theirsecretory pathway (Ferro-Novick and Jahn, 1994). NSFis an ATPase that, in the presence of SNAP, has beenreported to promote membrane fusion between un-coated transport vesicles and their compartment (Mal-hotra et al.,1988; Orci et al., 1989; Whiteheart et al.,1994). NSF does not bind directly to membranes butinstead binds to a/b and g-SNAP, which attach tospecific membrane receptors (SNAP receptors orSNAREs) (Weidman et al., 1989; Wilson et al., 1992;Whiteheart, et al., 1992). In neurons, SNAP receptors,first identified by Rothman and colleagues, form acomplex that contains the synaptic proteins synaptobre-vin (or vesicle associated membrane protein: VAMP),syntaxin, and SNAP-25 (25 kDa synaptosomal associ-ated proteins) (Sollner et al., 1993). The observationthat selective cleavage of each of these three proteins byclostridia neurotoxin leads to a complete block inneurotransmission, provided strong evidence that theyare essential proteins participating in the docking/fusion of secretory vesicles at the presynaptic plasmamembrane in neurons (Schiavo et al., 1992; Bennett etal., 1992; Link et al., 1992; Blasi et al., 1993). Identifica-

Fig. 2. Immunoblot analysis of brain and liver homogenates usingVAMP-specific antibody, demonstrating the presence of the 65–68 kDaand SDS-resistant SNARE complex. Twenty micrograms of brainhomogenate and 20 µg of liver homogenates obtained from control andpancreatitic animals were resolved using 12.5% sodium dodecylsulfate poly acrylamide gel electrophoresis. The resolved proteinswere electrotransferred to 0.2 µm nitrocellulose membrane and immu-noblotted using VAMP specific antibody, followed by HRP-conjugatedsecondary antibody. The immunoblots were then processed for theenhanced chemiluminescence reaction and exposed to X-OMAT-ARfilm. The exposed films were then developed and photographed.

Fig. 3. Relative binding between liver microsome and liver plasmamembrane proteins, using AFM force spectroscopy. Notice the loss ofbinding between the two protein fractions in pancreatitic animals(47.3 6 5.5; mean 6 SE), compared to controls (103.4 6 10.5; mean 6SE). Addition of VAMP-specific antibody to the preparation results inloss of binding in the control (69.6 6 8.5; mean 6 SE). Binding betweenliver microsome and liver plasma membrane proteins obtained fromrats with chronic pancreatitis is little effected by exposure to VAMPantibody (37.9 6 3.25; mean 6 SE).

365IMPAIRED MEMBRANE FUSION IN CHRONIC PANCREATITIS

tion of synaptobrevin and syntaxin homologues in yeast(Aalto et al., 1993; Protopopov et al., 1993) furthersuggested a remarkable conservation of these compo-nents of the membrane docking/fusion machinery.

Using microtiter plate binding assays, synaptic vesicleprotein-protein interactions have been examined (Pevs-ner et al., 1994; Zhong et al., 1997). The binding ofSNAP-25 to syntaxin, and of VAMP to syntaxin havebeen examined by such binding studies (Pevsner et al.,1994). The Pevsner et al. (1994) study shows that in thepresence of SNAP-25, there is a 10-fold increase inaffinity of binding of VAMP to syntaxin1A (the majorneuronal isoform of syntaxin). Recent studies (Otto etal., 1997) demonstrate that, although syntaxin andSNAP-25 are primarily localized to the neuronal plasmamembrane and VAMP exclusively to synaptic vesicles,all three proteins are also present as a ternary SDS-resistant complex (v-SNARE) at the synaptic vesiclemembrane, and undergo disassembly in the presence ofNSF-ATP. The structure of NSF-SNARE complex usingpurified recombinant proteins, has recently been exam-ined by quick-freeze/deep-etch electron microscopy(Hanson et al., 1997). Studies (Otto et al., 1997; Walch-Solimena et al., 1995) also show that although themajority of vesicle syntaxin and SNAP-25 are as-sembled in a complex with VAMP, there is a 5 molarratio of excess free VAMP on purified synaptic vesicles.These studies would suggest that purified microsomesand plasma membrane-enriched fractions from liverwill have their v-SNARE and t-SNAREs. According tothe current model for vesicle docking/fusion, v-SNAREand the two plasma membrane proteins (t-SNARE),SNAP-25 and syntaxin, interact with each other aftertheir disassembly by NSF. It is suggested that theresultant interactions between the two SNAREs bringthe vesicle membrane and the plasma membrane closeto each other to bring about membrane-membranefusion (Hanson et al., 1997; Weber et al., 1998), analo-gous to certain viral protein catalyzed membrane fu-sion reactions. Results from this study demonstratethat the binding forces between liver microsome andliver plasma membrane proteins are VAMP-sensitive,with a 32% loss of force seen after introduction of ananti-VAMP antiserum. In addition, the study shows a54% loss of binding force between microsomal andplasma membrane proteins isolated from pancreatiticlivers, with no further decrease in binding force appar-ent when VAMP antibody was introduced. These stud-ies suggest for the first time impaired membrane fusionin liver of animals with chronic pancreatitis and sug-gest that a loss of expression or function of the VAMPprotein contributes directly to this abnormality. Theloss of microsomal and plasma membrane binding forcemay play a role in the altered hepatocyte vesicletrafficking seen in chronic pancreatitis.

ACKNOWLEDGMENTSThis work was supported by United States Public

Health Service grant DK39950 to D.K.A.

REFERENCESAalto MK, Ronne H, Keranen S. 1993. Yeast syntaxins Sso1p and

Sso2p belong to a family of related membrane proteins that functionin vesicular transport. EMBO J 12:4095–4104.

Andersen DK, Ruiz CL, Burant CF. 1994. Insulin regulation of

hepatic glucose transporter protein is impaired in chronic pancreati-tis. Ann Surg 219:679–687.

Bennett MK, Calakos N, Scheller RH. 1992. Syntaxin: a synapticprotein implicated in docking of synaptic vesicles at presynapticactive zones. Science 257:255–259.

Blasi J, Chapman ER, Linke E, Binz T, Yamasaki S, De Camilli P,Sudhof TC, Niemann H, Jahn R. 1993. Botulinum neurotoxin Aselectively cleaves synaptic protein SNAP-25. Nature 365:160–163.

Clary DO, Griff IC, Rothman JE. 1990. SNAPs, a family of NSFattachment proteins involved in intracellular membrane fusion inanimals and yeast. Cell 61:709–721.

Ferro-Novick S, Jahn R. 1994. Vesicle fusion from yeast to man.Nature 370:191–193.

Hanson PI, Roth R, Morisaki H, Jahn R, Heuser JE. 1997. Structureand conformational changes in NSF and its membrane receptorcomplexes visualized by quick-freeze/deep-etch electron microscopy.Cell 90:523–535.

Link E, Edelmann L, Chow JH, Binz T, Yamasaki S, Eisel U, BaumertB, Sudhof TC, Niemann H, Jahn R. 1992. Tetanus toxin action:inhibition of neurotransmitter release linked to synaptobrevinproteolysis. Biochem Biophys Res Commun 189:1017–1023.

Malhotra V, Orci L, Glick BS, Block MR, Rothman JE. 1988. Role ofN-ethylmaleimide-sensitive transport component in promoting fu-sion of transport vesicles with cisternae of the Golgi stack. Cell54:221–227.

Nathan JD, Zdankiewicz PD, Wang JP, Spector SA, Jena BP, GeibelJP, Seymour NE, Andersen DK. 1997. Internalization of hepatocyteglucose transport protein after feeding is impaired in chronicpancreatitis. Pancreas 15:448.

Nathan JD, Slezak LA, Zdankiewicz PD, Wang JP, Spector SA, JenaBP, Geibel JP, Seymour NE, Andersen DK. 1998. Impaired regula-tion of hepatocyte endocytosis in chronic pancreatitis. Dig Dis Sci43:1875.

Orci L, Malhotra V, Amherdt M, Serafini T, Rothman JE. 1989.Dissection of a single round of vesicular transport: sequentialintermediates for intercisternal movement in the Golgi stack. Cell56:357–368.

Otto H, Hanson PI, Jahn R. 1997. Assembly and disassembly of aternary complex of synaptobrevin, syntaxin, and SNAP-25 in themembrane of synaptic vesicles. Proc Natl Acad Sci 94:6197–6201.

Pevsner J, Hsu S-C, Braun JEA, Calakos N, Ting AE, Bennett MK,Scheller RH. 1994. Specificity and regulation of a synaptic vesicledocking complex. Neuron 13:353–361.

Protopopov V, Govindan B, Novick P, Gerst JE. 1993. Homologs of thesynaptobrevin/VAMP family of synaptic vesicle proteins function onthe late secretory pathway in S. cerevisiae. Cell 74:855–862.

Rothman JE. 1994. Mechanism of intracellular transport. Nature372:55–63.

Schiavo G, Benfenati F, Poulain B, Rosetto O, Polverino De Lareto P,DasGupta BR, Montecucco C. 1992. Tetanus and botulinum-Bneurotoxins block neurotransmitter release by proteolytic cleavageof synaptobrevin. Nature 359:832–835.

Seymour NE, Brunicardi FC, Chaiken RL, Lebovitz HE, Chance RE,Gingerich RL, Elahi D, Andersen DK. 1988. Reversal of abnormalglucose production after pancreatic resection by pancreatic polypep-tide administration in man. Surgery 104:119–129.

Slezak LA, Nathan JD, Zdankiewicz PD, Spector SA, Wang JP, JenaBP, Geibel JP, Seymour NE, Andersen DK. 1998. Impaired hepato-cyte insulin receptor internalization in chronic pancreatitis. SurgForum 49:152–153.

Sollner T, Whiteheart SW, Brunner M, Erdjument-Bromage H, Gero-manos S, Tempst P, Rothman JE. 1993. SNAP receptors implicatedin vesicle targeting and fusion. Nature 362:318–324.

Spector SA, Frattini JC, Zdankiewicz PD, Nathan JD, Wang JP,Andersen DK, Seymour NE. 1997. Insulin receptor gene expressionin chronic pancreatitis: the effect of pancreatic polypeptide. SurgForum 48:168–171.

Sritharan KC, Quinn AS, Taatjes DJ, Jena BP. 1998. Binding contribu-tion between synaptic vesicle membrane and plasma membraneproteins in neurons: An AFM Study. Cell Biol Int (in press).

Walch-Solimena C, Blasi J, Edelmann L, Chapman ER, Fischer vonMollard G, Jahn R. 1995. The t-SNARE’s syntaxin 1 and SNAP-25are present on organelles that participate in synaptic vesiclerecycling. J Cell Biol 128:637–645.

Weber T, Zemelman BV, McNew JA, Westermann B, Gmachl M,Parlati F, Sollner TH, Rothman JE. 1998. SNAREpins: minimalmachinery for membrane fusion. Cell 92:759–772.

366 SLEZAK ET AL.

Weidman PJ, Malancon P, Block MR, Rothman JE. 1989. Binding ofN-ethylmaleimide-sensitive fusion protein to Golgi membranesrequires both a soluble protein(s) and an integral membranereceptor. J Cell Biol 108:1589–1596.

Whiteheart SW, Brunner M, Wilson D W, Wiedmann M, Rothman J E.1992. Soluble N-ethylmaleimide-sensitive fusion attachment pro-teins (SNAPs) bind to a multi-snap receptor complex in Golgicomplex. J Biol Chem 267:12239–12243.

Whiteheart SW, Griff IC, Brunner M, Clary D O, Mayer T, Buhrow SA,Rothman JE. 1993. SNAP family of NSF attachment proteinsincludes a brain-specific isoform. Nature 362:353–355.

Whiteheart SW, Rossnagel K, Buhrow SA, Brunner M, Jaenicke R,

Rothman JE. 1994. N-ethylmaleimide-sensitive fusion protein: atrimeric ATPase whos hydrolysis of ATP is required for membranefusion. J Cell Biol 126:945–954.

Wilson DW, Wilcox CA, Flynn GC, Chen E, Kuang WJ, Hanzel WJ,Block MR, Ullrich A, Rothman JE. 1989. A fusion protein requiredfor vesicle-mediated transport in both mammalian cells and yeast.Nature 339:355–359.

Wilson DW, Whiteheart SW, Wiedmann M, Brunner M, Rothman JE.1992. A multisubunit particle implicated in membrane fusion. J CellBiol 117:531–538.

Zhong P, Chen YA, Tam D, Chung D, Scheller RH, Miljanich GP. 1997.An a-helical binding domain within the H3 domain of syntaxin isrequired for SNAP-25 binding. Biochemistry 36:4317–4326.

367IMPAIRED MEMBRANE FUSION IN CHRONIC PANCREATITIS