ANALYTICAL BIOCHEMISTRY 226, 342-548 (1905) Phospholipid/Alkanethiol Bilayers for Cell-Surface Receptor Studies by Surface Plasmon Resonance Anne L. Plant,* Michael Brigham-Burke,+ Eugene C. Petrella,t and Daniel J. O’Shannessyt “Biotechnology Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, Maryland 20899; + Department of Protein Biochemistry, SmithKline Beecham Pharmaceuticals, 709 Swedeland Road, King of Prussia, Pennsyleania 19406; and Department of Cell Biology and Anatomy, Johns Hopkins School of Medicine, 725 Narth Wolfe Street, Baltimore, Maryland 21205 Received October 20, 1994 Supported hybrid bilayer membranes (HBM) com- posed of a monolayer of phospholipid and a monolayer of alkanethiol associated with a thin gold film on glass ¢ useful as model lipid bilayer membranes for study- ing membrane receptor-ligand and cell-cell binding events by surface plasmon resonance (SPR). Measure- ments of specific binding of proteins and lipid vesicles to well-defined HBMSs have been performed under con- ‘tions of continuous flow using a commercial SPR in- strument (BIAcore). HBMs are shown to be stable in flow and to block nonspecific adsorption of proteins to the alkanethiol/gold surface. The use of such supported lipid bilayers in flow provides a means of conducting equilibrium and kinetic studies of models of ligand-cell and cell-cell interactions with receptors or ligands in ‘4 membrane environment. Compared to the extended dextran polymer layer that is currently used for sur- face modification of BlAcore “sensor chips,” the de- scribed HBMs provide a well-defined surface that will permit less ambiguous modeling of these important bio- logical interactions. ©1995 Academie Prost. ne. ‘The determination of equilibrium binding and kinetic rate constants between receptors and ligands can be problematic because it frequently requires the use of sec- ondary labels for detection. Surface plasmon resonance (SPR) is one of the few applicable techniques that cir- ' Abbreviations used: SPR, surface plasmon resonance; SAM, self assembled monolayer; HBM, hybrid phospholipid-containing bilay membrane; TBS, Tris-buffered saline; BSA, bovine serum album WGA, wheat germ agglutinin; NA, neatravidin; IgG, immunoglobulin G; POPC. 1-palmitoyl, 2-oleoyl phosphatidylcholine; Bt-x-PE, N-((6. biotinylbaminojhexanoyl-1,2-hexadecanoyl-sn-glycerophosphoeth: anolamine. ae cumvents this problem. The angle at which incident light will be in resonance with surface plasmons is sensi- tive to the refractive index of the medium near a metal surface. The binding of ligands to surface-immobilized receptors alters the interfacial refractive index, allowing ‘one to directly monitor a change in the molecular composition of the interface by monitoring the change in angle of the reflectivity minimum (1). BlAcore® is a ‘commercial SPR instrument which incorporates mi crofiui ing automation of sample injections, and real-time monitoring of both binding and dissocia- tion events, permitting direct quantitation of equilib- rium binding and kinetic rate constants. Applications of continuous-flow SPR have up to now been limited to water-solubilized receptors and ligands. ‘The modification of commercial SPR ‘sensor chips” with receptors or ligands has involved the covalent at- tachment of water-soluble ligands or receptor fragments onto an extended, hydrophilic dextran polymer (2,3) which is covalently immobilized to the gold surface via an alkylthiol. Here we present an alternative approach to surface derivatization which allows the use of lipid ‘membrane-bound species and is readily applicable to the study of cell-surface receptor/ligand binding, protein/ membrane interactions, and cell-cell interactions. Formation of the model membrane involves coating the gold plasmon resonance surface with a hydrophobic self-assembled monolayer (SAM) of octadecanethiol, to which phospholipid vesicles containing lipophilic agents of interest are allowed to fuse. Strong chemisorption of the thiol sulfur to the gold results in a highly stable al- kanethiol SAM; hydrophobic interactions between the ° Certain commercial products are identified in order to adequately specify the experimental procedure; this doos not imply endorsement ‘or recommendation by NIST. 003-2607/95 $8.00 Copyright © 1995 by Academic Pre ne Alleght of reproduction in any form reservedHYBRID BILAYER MEMBRANES FOR SURFACE PLASMON RESONANCE, ak FIG. 1. Schematic of the HBM and the SPR experiment. It is en visaged that the vesicles “unfold” onto the SAM to generate the HEM. ‘The thin gold coating is optically transparent, and light coupled into it generates plasmons. The reflected light is detected at a photodiode array. The angle of the reflectance minimum isa funetion of the re fractive index atthe interface alkyl chains of the alkanethiols and the phospholipid molecules provide a large stabilizing force to the bilayer. Itis expected that exclusion of water from the hydropho- bie monolayer is what provides the thermodynamic driv- ing force for bilayer formation, and formation of bilayers by fusion of phospholipid vesicles with hydrophobic surfaces has been described before (4). In our ease, the result of vesicle fusion with the alkanethiol monolayer is a hybrid phospholipid-containing bilayer membrane (HBM), consisting of a hydrophobic interior of long acyl chains, and polar headgroups at the membrane/solution, interface (see Fig. 1). Previous work (5~7) has indicated that the electrical, structural, and biomimetic properties of HBMs are consistent with those of other model bi layer membranes. In this report we provide evidence that such lipid bi- layer structures can be used in the BlAcore. Using mod- ified commercial sensor chips we provide evidence for the formation of a HBM on the sensor chip, the stability of these bilayers in flow, and the ability of these HBMs to provide reactive groups for specific molecular interae- tions. Using secondary lipid vesicles as models of cells, wwe use this approach to study models of cell-cell interac- tions under continuous-flow conditions. MATERIALS AND METHODS We have performed this study using a commercial plasmon resonance instrument (BIAcore, Pharmacia 343 Biosensor, Piscataway, NJ), which incorporates auto- mated microfluidics. All SPR experiments were per. formed under conditions of constant flow at 2 yl min” in TBS (20 mM Tris, 150 mM NaCl, pH 7). Bovine serum, albumin (BSA) and wheat germ agglutinin (WGA) were purchased from Sigma Chemical Co. (St. Louis, MO}: neutravidin (NA) was purchased from Pierce Chemical Co. (Rockford, IL); immunoglobulin G (IgG) was pre- pared in-house and was biotinylated with sulfosuccini- midyl 6-(biotinamido) hexanoate (Pierce) ‘Sensor Chip Reconditioning and Formation of Self- ‘Assembled Monolayer ‘To maximize compatibility of the components, com- mercial sensor chips which had been modified with an, alkylthiol-dextran coating by the manufacturer (Phar- ‘macia) were reconditioned for use in these studies. Each chip, composed of a thin (approx 450 A) layer of gold deposited on borosilicate glass, was removed from its plastic holder by soaking in ethanol, and a bare gold sur- face was regenerated by irradiation for 45 min in a uv/ ozone cleaner (Boekel Industries, Philadelphia, PA). Ul- traviolet irradiation has been used to destroy the thiol- gold interaction for the production of patterned SAMs (8). The irradiated chips were rinsed with ethanol and immersed immediately into a 1 mM ethanolic solution of octadecanethiol (Aldrich Chemicals, Milwaukee, WI) for at least 16 h in order to form a SAM. For SPR ex- periments, the SAM-coated chips were reattached to the plastic holders using silicon adhesive. Preparation of Lipid Vesicles 1-Palmitoyl, 2-oleoy! phosphatidylcholine (POPC) was purchased from Avanti Polar Lipids (Alabaster, AL). N-((6-Biotinyl)amino)hexanoyl-1,2-hexadecanoyl- sn-glycerophosphoethanolamine (Bt-x-PE) was pur- chased from Molecular Probes (Eugene, OR). Vesicles ‘were prepared by the injection method (9). Phospholipid solutions (POPC only or a mixture of POPC plus Bt-x- PE in CHCl) were taken to dryness under N, and des- iccated overnight to remove residual solvent. Dried lipid was then resolubilized in dry isopropanol and injected into TBS while vortexing. The final lipid concentration was 2 mM. Characterization of SAMs and HBMs Characterization of the alkanethiol monolayer on rep- resentative chips included contact angle and electrical impedance measurements. Advancing contact angles for water were determined under ambient conditions by ap- plying a 2-u1 drop onto the gold surface and measuring the angle between the drop and the surface using a goni- ometer (Rame-Hart, Inc., Mountain Lakes, NJ). The344 integrity of representative monolayers and bilayers was determined by electrical impedance measurements as has been previously described (5). Impedance measure- ments were made with a Solartron electrochemical in- terface (Model 1250, Schlumberger, Hampshire, En- gland) and a Solartron frequency generator (Model 1286). A three-electrode cell was used, with a Ag/AgCl reference electrode, a Pt counter electrode, and the gold- coated sensor chip as the working electrode. The work- ing gold electrode area was nominally 0.19 em?. Mea- surements were made in TBS at a de potential of 0 V and an applied sinusoidal ac potential of +10 mV. A simple equivalent circuit model (5) of a solution resistance in series with a capacitance was used to analyze the re- sponse a where |Z| is the total impedance magnitude, R is the solution resistance, w is the frequency, and C is the ca- acitance of the dielectric layer. After making imped- ance measurements on a SAM sample, a HBM was pre pared by adding vesicles directly to the cell in a final concentration of 0.2 mM total lipid in TBS. Formation of HBMs for SPR Experiments For SPR experiments, the SAM-modified chips were reattached to their plastic holders using silicon adhesive. Subsequently, HBMs were formed by applying a drop of lipid vesicle suspension to a monolayer-coated chip, allowing the chip to remain overnight in a humidified environment, and then rinsing off the excess vesicles briefly with water. Sensor chips were inserted into the BIAcore instrument and equilibrated with TBS at 2 ul/ min for at least 1 h before use. RESULTS Characterization of Monolayers and Bilayers ‘The removal of the dextran/alkylthiol layer from the gold surface of the sensor chip and its replacement with a layer of octadecanethiol were confirmed by advancing contact angle and electrical impedance measurements made on representative chips. These data are found in ‘Table 1. A contact angle of about. 6° was measured for the dextran-coated sensor chip prior to irradiation and treatment with octadecanethiol, indicating an extremely hydrophilic surface. After irradiation of the sensor chips and incubating with octadecanethiol, the contact angle measured for water on these surfaces was 105°, indicat- ing extremely poor wettability, as is consistent with the formation of a layer of alkane chains. This extremely PLANT ET AL. TABLE 1 Advancing Contact Angle and Impedance Measurements of Sensor Chips before and after Ultraviolet Irradiation and Exposure to Octadecanethiol Contact Capacitance angle (deg) Fem) Dextran-coated chip 628 nd Octadecanethiol monolayer Bateh 1 0.94 002 he 0.95 + 0.02 POPC /octadecanethiol bilayer Batch 1 nd 071 + 0.008 Batch 2 0.69 = 0.007 ‘Note. Two batches of chips were prepared and results are the aver ‘age from three chips from batch Land two chips from batch 2, Contact angles were measured for water. nd, not determined hydrophobic surface is essential for the subsequent as: sembly of the phospholipid layer. ‘The strong interaction between thiol sulfur and gold (10) causes the spontaneous formation of a well-ordered and complete monolayer of alkanethiol on a planar gold surface. Using octadecanethiol-modified gold chips as working electrodes, we measured the electrical imped- ance of the insulating alkane layer as a function of ac frequency and, from these data, calculated the mono- layer capacitance. Impedance measurements, which are extremely sensitive to the molecular details of the elec- trode surface, indicate that chips which have been treated with octadecanethiol are covered by a dielectric barrier with a specific capacitance of 0.94 uF/em*, which is consistent with that of a well-ordered octadecanethiol monolayer (5,11). Impedance measurements also allow us to follow changes in the surface coverage when lipid vesicles are added to the monolayer-coated gold, The thickness, d, of the dielectric layer is related to the specific capacitance (C,., the capacitance normalized for the electrode area) by the relationship (21 where «9 is the permittivity in free space and «, the di- electric constant of the alkanethiol, is taken to be 2.26 (AD). With time after addition of POPC vesicles to the electrochemical cell containing the SAM-modified sen- sor chip, the capacitance decreases, indicating an in- crease in the thickness of the dielectric layer. The de- crease in capacitance over time is shown in Fig. 2. The average limiting capacitance of 0.7 uF/em* compares fa- vorably with previous observations of POPC/octadec- anethiol HBMs prepared on 2000-nm-thick gold filmsHYBRID BILAYER MEMBRANES FOR SURFACE PLASMON RESONANCE 345, Be 5 HS Time (min) FIG. 2. Time course for formation of a HEM by fusion of POPC vesicles with an octadecy] mercaptan monolayer in an unstirred cell at room temperature. Capacitance was determined as described in the text. The different symbols indicate different data sets Capacitance (nF) & ® (6). Figure 2 also shows that alter adding POPC vesicles to the SAM- modified sensor chips in an unstirred cell at ambient temperature, a limiting capacitance is reached in about 1h. Surface Plasmon Resonance Experiments Integrity of HBMs. For SPR experiments, HBMs were formed by incubating alkanethiol-coated sensor chips in their holders with phospholipid vesicles over: night at ambient temperature prior to placing them in the instrument. The observation that bilayers were sta- ble under conditions of continuous flow is of great im- portance. No loss of the bilayer coating was observed for HBM.-modified sensor chips exposed to a flow rate of 2 l/min for at least 18 h. ‘The phospholipid-derivatized surface also showed tle nonspecific adsorption of protein, as shown in Fig. 3A. Relatively high concentrations (0.5 mg/ml) of BSA, IgG, and WGA were injected. These protein solutions have a significantly different refractive index compared to the buffer solution, and this is detected as an increase in the response units (Y-axis) as the concentrated pro- tein solutions flowed past the HBM-coated gold surface. As shown, little of the injected protein associated with the HBM surface. As the protein solutions passed out of the flow cell and were replaced by buffer solution (at approximately 1250, 2900, and 4500 s), the resonance angle returned to a value close to the initial value. This indicates almost no net change in the refractive index of the surface layer and therefore little binding of these proteins to the POPC-containing bilayer (see Fig. 3). In contrast, significant binding of BSA was observed when injected across an octadecanethiol-coated sensor chip (Fig. 3B). A comparison of Fig. 3A and Fig. 3B indi- cates that approximately 30 times more BSA bound to the SAM-coated surface than to the phospholipid bi- layer surface. This is not unexpected considering the hy- drophobic nature of the octadecane layer and the avail- ability of hydrophobic moieties in BSA. Since proteins are well known to adsorb onto gold, the lack of binding of BSA, WGA, or IgG to the HBMs (Fig. 3) demon strates the effectiveness of the POPC layer at blocking nonspecific protein adsorption onto either the octadeca- nethiol layer or the gold. Building specificity into HBMs. To confer binding specificity to the bilayer, POPC vesicles containing a otinylated phospholipid, Bt-x-PE, at a molar ratio of 2 ‘oF 10% (mol/mol) of total lipid were used for the prepa ration of HBMs. Figure 4A shows that BSA does not bind to these layers, but that NA does. Subsequent to NA binding to these layers, addition of a biotinylated IgG (Bt-IgG) (Fig. 4B) resulted in an interaction with the bound NA. The change in response units indicating a change in resonance angle due to the binding of NA to 110. & Response Units (x 103) 954 90. © ~~ 1000-2000 30004000 Time (s) a0. = FP S10. 2 Bsa = 5100. 3 8 § 95. 3 & 9.0. oO 500 1000 1500 Time (s) FIG. 3._ Integrity ofthe HBMs as determined by protein adsorption ‘Arrows indicate time of sample injection. The injected volume was 40 tl. The Y-axis, response units (RU), is directly proportional to the ‘change in resonance angle ofthe evanescent wave which is sensitive tothe thickness of the layer atthe gold surface. (A) SPR response of @ HBM composed of POPC. Injection of BSA (0.5 mg/ml), RU change = Als injection of WGA (0.5 mg/ml), RU change = 11; injection of IgG (0.5 mg/rall, RU change = 50.(B) Injection of BSA (0.5 mg/ml) onto an octadecanethiol monolayer, RU change ~ 1140.346 PLANT ET AL “4 im e ». exe I 2. ia le oN 2 % 104 asa 88h g a 2 4+— T — ¥ § 3 5 & Bee na bh © 100920 300040005000 1000 200 3000 S000 Time (s) FIG. 4. SPR response of HBMs containing 2 ‘mg/tal) (two injections), neutravidin (NA} (125 ug/ml), biotinylated (A, B) of 105 (C, D) Bact Conditions are as described in the legend to Fig. 8. BSA (0.5 (BU-IgC) (0.5 ma/ml), and POPC vesicles containing 10% Bt-x-PE. {20 gat cota lipid) were injected as indicated. RU changes due to BSA binding were (A) 244, (B) 154, (C) 249, (D) 279, RU changes due to NA, binding were (A) 1723, (B) vesicles were (B) 3663, (D) 36 the 10% Bt-x-PE surface (Figs. 4C and 4D) is greater than when NA binds to the 2% Bt-x-PE layer, indicating. that more NA binds to the surface containing more Bt- x-PE, as might be expected. Also shown in Fig. 4B is the result of injecting new Bt- x-PE-containing vesicles onto the Bt-x-PE- derivatized surface with which NA had interacted. The Bt-x-PE vesicles bound to the NA at the bilayer surface, simulat- ing a membrane-membrane interaction. Control exper- iments (data not shown) indicated that injection of POPC-only vesicles or vesicles containing 10% Bt-x-PE to the Bt-x-PE bilayer resulted in no significant change in the response. Similarly, in the absence of NA, Bt-x- PE-containing vesicles did not interact with the Bt-x. PE-containing HBM. Thus, lipid vesicles do not alter the preformed bilayer surface by nonspecific adsorption or fusion with it. Also, injection of NA or avidin in the presence of a 1000-fold molar excess of biotin onto a Bt- x-PE-containing bilayer resulted in no significant bind- ing of the protein and, subsequently, no binding of Bt-x- PE liposomes, indicating further that the binding that is observed is specific to the biotin-neutravidin interac- tion. 7, (C) 2490, (D) 3647, RU changes due to Bt-lgG binding were (A) 1103, (C) 132 RU changes due to Bt-x-PE, ‘The binding of the Bt-x-PE vesicles to the HBM pro- vided approximately a threefold increase in signal com- pared to the response from IgG binding. It is also appar- ent that the kineties of these two events are significantly different, with the increase in response due to vesicles occurring more slowly than that for IgG as would be ex- pected due to the significantly larger size and smaller diffusion constant of the vesicles. ‘One important advantage of monitoring the SPR re- sponse under continuous-flow conditions and in such small volumes is the ability to measure rapid binding events in real time (12). This is demonstrated in Fig. 5 where different concentrations of vesicles containing Bt-x-PE were injected onto a bilayer prepared from 10% Bt-x-PE to which NA was bound. The binding of Bt- x-PE vesicles was concentration- and time-dependent. Figure 5 demonstrates, as expected for a first-order ki- netic process, that the rate of vesicle binding is a func- tion of vesicle concentration. Alkanethiol monolayers can be easily regenerated from HBMs. The phospholipid portion of the bilayer can be removed without disturbing the underlying alkane- thiol monolayer, thus regenerating a monolayer and al-HYBRID BILAYER MEMBRANES FOR SURFACE PLASMON RESONANCE aw °S BPE 2 +0 =e Bis ° Bex PE Bae ote g 6 8 5 1s Bare oa 1500 2 T T T T © 04000000000 TIME (s) FIG. 5. SPR response to injections of Bt-x-PE vesicles when the bilayer was prepared with Bt-x-PE previously exposed to NA. Vesicle concentrations were 40, 20, and 4 4M total ipi lowing formation of a new bilayer containing the same or a different lipid. Regeneration of the monolayer was, achieved by applying 2 N NaOH to the surface overnight to remove residual proteins and then rinsing with etha- nol to completely remove the phospholipid layer. This procedure left the monolayer intact and capable of form- ing a new bilayer (data not shown). DISCUSSION ‘The strength of the interaction between gold and thiol sulfur molecules has been extremely useful in the modi fication of gold surfaces with alkylthiols for SPR studies. Commercially available sensor chips are prepared with an alkylthio layer that is derivatized with a large dex- tran polymer. We have successfully modified these com- mercial chips with a monolayer of hydrophobic al kanethiol with excellent reproducibility. Contact angle measurements indicate that after uv irradiation and in- cubation in octadecanethiol solution, chips are highly hydrophobic with a measured contact angle for water of 105°. This value is somewhat smaller than the 112° re- ported for octadecanethiol monolayers by Bain et al. (13), probably because our measurements were made un- der ambient conditions, and their measurements were made in a water-saturated atmosphere. The thickness of the hydrophobic layer as inferred from capacitance measurements is within 10% of that previously reported for octadecanethiol on 1000- to 2000-A gold films. The small absolute change in thickness of the layer when converting a monolayer to a bilayer by adding phos- pholipid vesicles precludes direct measurement of the 347 phospholipid layer thickness with the Biacore. However, the decrease in capacitance which we measure upon exposure of POPC vesicles to the hydrophobic mono- layer indicates an increase in the thickness of the dielec- tric layer, the magnitude of which is consistent with the formation of a second layer of lipid. The capacitance as- sociated with the phospholipid layer corresponds to the acyl chain region only, since the electrical capacitance of phospholipid membranes is dominated by the nonpolar acyl chain region (14). The capacitance of the phospho- lipid layer can be calculated from 1 1 ——l (3) Twphowphaiia Cmtleer Cyrimonlayr c From Ea. [3] we calculate a capacitance of about 2.8 uF/ ‘em: for the phospholipid layer, which is consistent with previous observations on 1000- to 2000-A gold-film elec- trodes (5,7). Using Eq. [2] and a dielectric constant of 2.7 (6,7), we calculate the thickness of the acyl chain re- gion of the POPC layer to be about 8.5 A. This is some- what shorter than the 12- to 13-A thickness that can be estimated from X-ray diffraction data of phospholipid bilayers (15). This could indicate that our POPC layers, are at a greater tilt from normal or that their acyl chains interdigitate with those of the alkanethiol monolayer. Other evidence for formation of a complete outer layer of phospholipid is the blockage of nonspecific adsorption, of BSA and the specific binding activity of NA to B-x PE in the layer. We have shown here that phospholipid/alkanethio! HBMs can be prepared on commercial sensor chips and that such HBMs are stable and apparently effective re- ceptor matrices under continuous-flow conditions. A more complete characterization of kinetic and equilib- rium constants is planned to determine that protein binding at this hybrid model membrane is indeed rele- vant to biological systems. Recently, a number of groups have demonstrated the response of thin monolayers of functionalized alkylthiols (16-18) and phospholipid/al- kanethiol bilayers (19) in nonflowing systems. Based on this work, we expect that a wide variety of surface chem- istries involving thiol-containing monolayers and layers will prove to be applicable to continuous-flow SPR. The attraction of using SPR in flowing systems to allow kinetic analysis of binding to cell membrane com: ponents has inspired one group to immobilize to the dex- tran layer membrane vesicles containing toxin receptors (20). We anticipate that planar hybrid bilayers, which will allow the characterization of membrane receptors in astable, well-defined lipid matrix, will prove to be a more useful and suitable matrix for such studies. ‘The use of a monolayer or bilayer will in most cases result in a smaller concentration of reactants at the sur-348, face compared to the relatively large reactive volume that is achieved by derivatization of the extended dex- tran layer. Since the magnitude of the change in reso- nance angle is a function of the change in concentration and refractive index of the species which associate with the reactive surface, we might expect to sacrifice signal intensity by using this surface derivatization approach. On the other hand, since the energy of the evanescent wave decays exponentially with distance from the sur- face, the relatively closer proximity of the reactive spe- cies to the resonance surface should result in a larger signal change. The magnitude of angle change that we detected in these experiments suggests that these two effects partially compensate for one another. It is ex- pected, therefore, that SAM- or HBM-modified surfaces will prove to be very useful for kinetic studies using com- mercial SPR flow instruments such as BIAcore. CONCLUSION ‘The phospholipid/alkanethiol HBM provides the means of incorporating membrane lipids, proteins, or protein fragments into a physiologically relevant micro- environment, eliminating the need for aqueously soluble fragments. The HBM may prove to be an ideal surface for studying kinetics and equilibrium constants for cell membrane receptors, even when the interaction is weak. For example, in preliminary studies we have observed the specific binding of profilin to HBMs containing PIP and PIP;; there was no binding of profilin to POPC-only HBMs, The combined use of planar HBMs with lipid vesicles as model cells provides a relevant experimental model of cell-cell interactions. Mobility of membrane- bound agents in the plane of the bilayers should allow for migration of interacting groups in both membrane compartments. Evidence for the mobility of HBM com- ponents includes the activity of melittin in this model membrane (5,7). Migration of reactive groups would be necessary to’ optimize multivalent interactions and mimie natural membrane conditions; this would not be achieved by covalent immobilization. This work demon- strates the feasibility of using HBMs with commercial PLANT. ET AL. SPR instrumentation and continuous-flow SPR detec- tion and opens the way for in-depth kinetic analyses of membrane receptor/ligand and cell-cell interactions, REFERENCES 1. Kooyman, R. P. H., Kolkman, H., Van Gent, J., and Greve, J (1988) Anal Chim. Acta 218, 35-48. 2. Brigham-Burke, M., Bawards, J, R., and O'Shannessy, D. J (1992) Anal Biochem. 208, 125-181 8, O'Shannessy, D. J, Brigham-Burke, M.. and Peck, K. (1992), Anal Biochem. 208, 132-138, 4. Kalb, E, Frey, S, and Tamm, L. (1992) Biochim, Biophys, Acta 1108, 907-316, 5, Plant, AL. (1988) Langmuir 8, 2764-2767, 6, Plant, A. L., and Gueguetchkeri, M. (1984) Proceedings of the 13th Symposium of the Biomedical Engineering Conference, pp. 610-613. 7. Plant, 8. Ly, Guegeutchkeri, M., and Yap. W. (1984) Biophys. J 67, 1126-1183, lov, M. J, Burgess, D.R. S., and Gil Chem. Soe. 118, 5308-8306. 8. Batzri, S, and Korn, E, D. (1973) Biochim. Biophys. Acta 901, 137. Nuzzo, R.G., Fusco, F.A., and Al Soe. 108, 2558-2968, Chidsey, C.F. D., and Loiscono, D. N. (1990) Langmuir 6, 682: 691, O'Shannessy,D. 4 ., (1993) J. Am. 10. D.L. (1987) J. Am. Chem. nh. 1, Brigham: Burke, M., Soneson, K.K., Hensley, P.,and Brooks, I (1993) Anal Biochem. 212, 457-468. Bain, C. D,, Troughton, B. B., Tao, YT, Eval, J. Whitesides, G.M, and Nuzz0, R. G. (1989) J. Am, Chem. Soe. 111, 321-836. Benz, R., Frohlich, O., Lauger, P, and Monta, M. (1975) Biochim. Biophys. Acta 394, 323-394 A, and Engelman, D, M, (1988) J, Mol Biol 168, 211 13, 4. Muller, W., Ringsdorf, H., Rump, E, Wildburg, G., Abang, X., ‘Angermaier,L., Knoll, W. Liley, M, and Spinke, J, (1998) Se fence 262, 1706-1708, van den Heuvel, D. J., Kooyman, R. P. H., Drjfhout, J. W., and Walling, G. W. (1983) Anal. Bigehem. 218, 223-230, Lang, H., Duschl,C., and Vogel, H. (1994) Langmuir 10, 197-216. ‘Terrettaz,S, Stora, T, Dusehl,C.,and Vogel, H. (1988) Langmuir 9, 1361-1368, Masson, L., Mazz, A.,and Brousseau, R. (1994) Anal. Biochem. 218, 405-412,