Binding of monoclonal antibody to protein antigen in fluid phase or bound to solid supports

Journal of Immunological Methods, 55 (1982) 1-12

1

© Elsevier Biomedical press

Binding of Monoclonal Antibody to Protein Antigen in Fluid Phase or Bound to Solid Supports Stephen J. Kennel Biology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37830, U.S.A.

(Received 12 February 1982, accepted 30 April 1982)

Rat monoclonal antibody (MoAb) to fragment D (FgD) of human fibrinogen was used to characterize the direct binding of antibody to protein in solution or bound to solid supports. Purified IgG, F(ab')2 and Fab' were prepared from ascites fluid of hybridoma 104-14B which is a fusion product of spleen cells from a rat immunized with FgD and the mouse myeloma cell line, P3-X63-Ag8. Two-dimensional electrophoresis of radioiodinated antibody preparations demonstrated the presence of hybrid immunoglobulin molecules, but only structures having rat heavy and rat light chains had active antibody combining sites. The affinity constant for IgG as well as F(ab') 2 and Fab', 6× 10 9 M - 1 , was identical when tested using fluid phase antigen (125I-labeled FgD). Affinity constants determined for direct binding of iodinated IgG using FgD immobilized on solid supports showed a slight dependence on the antigen concentration used in the measurement. These values ranged from 0.5)<10 9 M - ! at high antigen concentrations (1.3× 10 -7 M) to 9× 109 M -1 at low antigen concentration (1.3× 10 -I° M). Binding constants for F(ab')2 and PUb' gave similar results indicating that binding was homogeneous and univalent. The capacity of solid state antigen to bind antibody varied with the method used to bind FgD to the solid support. FgD bound directly to polystyrene plates was least efficient at binding labeled antibody; FgD bound to plates through intermediate carriers poly(L-lysine) was ordy slightly more efficient, while antigen bound to Sepharose beads by cyanogen bromide activation was the most active. Key words: monoclonal antibody - - hybrid molecules - - binding constants - - solid state antigen

Introduction Identification of hybridomas producing monoclonal antibody (MoAb) of defined s p e c i f i c i t y r e q u i r e s e x t e n s i v e s c r e e n i n g o f l a r g e n u m b e r s o f cell cultures. M a n y l a b o r a t o r i e s use p r o t e i n a n t i g e n s i m m o b i l i z e d o n solid s u p p o r t s to e x p e d i t e w a s h i n g p r o c e d u r e s e m p l o y e d in a n t i b o d y b i n d i n g tests. W h i l e c o m p a r i n g s c r e e n i n g tests for M o A b we n o t e d t h a t cell m e m b r a n e b o u n d a n t i g e n was m u c h m o r e e f f i c i e n t at b i n d i n g a n t i b o d y , o n a m o l a r basis, t h a n w e r e p r o t e i n a n t i g e n s b o u n d to p o l y s t y r e n e

i Research sponsored by the Office of Health and Environmental Research, U.S. Department of Energy, under Contract W-7405-eng-26 with the Union Carbide Corporation.

plates. While interaction of antibody with cell membrane antigens follows predicted theoretical binding patterns (Accolla et al., 1980; Mason and Williams, 1980), the binding of MoAb to immobilized protein is complex (Pincus and Rendell, 1981) and has not been characterized in detail experimentally. Several investigators have noted that enzyme-linked assays (ELISA) and competitive radioimmunoassay are somewhat less sensitive than fluid phase assays (Ansari et al., 1978; Lehtonen and Villjanen, 1980; Elleman and Raison, 1981). Many ELISA type assays require as much as 1/~g of solid state antigen to produce assays with sensitivity in the nanogram range (Kalmakoff et al., 1977). These indirect assays indicate that much of the solid state antigen is inactive. Purified MoAb allow direct measurement of the amount of active antigen present in a preparation. In this paper, we use MoAb to FgD of human fibrinogen to show that the efficiencies of binding of antibody to immobilized protein depends on the way that antigen is bound to the support.

Methods and Materials

Protein preparations Hybridoma 104-14B was produced by fusion of spleen cells from a Fischer 344 rat, immunized repeatedly with 100 #g of FgD, with mouse myeloma P3-X63-Ag8 (Kennel et al., 1981). Ascites fluid was collected after i.p. injection of 107 104-14B cells into pristane treated n u / n u mice. IgG was purifed from ascites fluid by ammonium sulfate precipitation and ion exchange chromatography with a 0-1 M NaC1 gradient in 0.01 M phosphate buffer, pH 8.0, on DE52 resin (Whatman Chemical Separation, Maidstone, Kent). To prepare F(ab')2, IgG (10 mg/ml) was dialyzed versus 0.12 M sodium acetate buffer, pH 4.0, containing 0.05 M NaC1. Three hundred micrograms pesin (Worthington Biochemical Corporation, Freehold, N J) per 10 mg of IgG was added and the reaction mixture incubated at 37°C for 18h. The digested product was collected by precipitation with 50% saturated ammonium sulfate, redissolved and dialyzed in 0.01 M sodium phosphate and 0.15 M NaCI, pH 7.4 (PBS). Fab' was prepared by reduction of F(ab')2 at pH 5.0 with 0.02 M/3-mercaptoethanol at 37°C for 2 h. Free thiol groups were reacted with 0.02 M iodoacetamide and the product purified by gel filtration through AcA 44 resin (LKB, Rockville, MD) in PBS. Preparations were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis on 10% acrylamide gels (Laemmli, 1970). Purified FgD was a generous gift from J.P. Chen (University of Tennessee Memorial Research Center, Knoxville, TN) and was prepared from grade L fibrinogen (AB KABI, Stockholm) by the method of Chen et al. (1974).

Labeling and analysis of antibody Radioiodination was performed with carrier free 1251 or 131I (New England Nuclear) in small volumes (< 100 #1) using limiting chloramine T (5#g) and as previously described (Kennel et al., 1981). Labeled antibody and antibody fragments were diluted into PBS containing 5 mg/ml bovine serum albumin (BSA-PBS) and

further purified by gel filtration on AcA 34 (IgG) or AcA 44 (F(ab')2 and Fab'). Labeled antibody preparations were tested for purity by immunoprecipitation with goat antiserum to rat IgG (Kennel et al., 1981) and 2-dimensional gel electrophoresis as previously described (Kennel et al., 1981). To separate Fab' molecules having antibody activity from the whole Fab' population, 125I-labeled Fab' in BSA-PBS was incubated for 4 h at 4°C with Sepharose beads containing 100 #g/ml covalently bound FgD. After washing in BSA-PBS the bound antibody was eluted using 3 M KSCN containing 5 mg/ml BSA and dialyzed versus PBS. Bead assay

Purified FgD diluted in BSA-PBS was coupled to crosslinked, desulfonated (Porath et al., 1971) Sepharose 4B beads (Pharmacia, Uppsala) using cyanogen bromide (Cuatrecasas, 1970). Trace amounts of 131-labeled FgD were added to monitor the extent of coupling (usually 80-90%). Beads were then sonicated in a 1 : 1 suspension in BSA-PBS (Branson sonicator with microprobe) until 80-90% of the beads were fractured (assessed by light microscopy of trypan blue stained samples). For binding experiments, beads were washed in BSA-PBS and finally suspended at a final dilution of 5% (v/v). Beads were distributed (50 #1) into 96-well microtest plates using a multichannel pipettor and pipette tips cut off about 5 mm to provide a bigger hole for easy passage of beads. Antibody dilutions (50 #1) in BSA-PBS were added to the wells, the plate sealed with an adhesive cover (Linbro, Hamden, CT) and incubated for 18 h at 4°C on a 60 ° angled rotating disc at 20 rev/min (Beckton Dickinson and Company, CockeysviUe, MD) to provide mixing. In later assays, normal rat serum and NP40 were added to final concentrations of 5% and 0.05%, respectively, to lower background binding of antibody to beads. Plates were centrifuged in plate carriers at 1000 rev/min for 5 rain to remove beads and solution from the adhesive cover and the beads were harvested onto glass fiber filters using a multichannel harvester (Otto Hiller, Madison, WI) and PBS wash. Plate assay

Microtest wells in 96-well plates (Falcon, Oxnard, CA) were treated with 50 #1 intermediate carrier at 1 mg/ml for 6 h at room temperature on a plate agitator (Titertek, Flow Laboratories, McLean, VA). The different carriers tested were poly(L-lysine) (PLL), phosphatidyl ethanolamine, or phosphatidyl choline (Sigma, St. Louis, MO). To couple FgD to PLL, wells treated with PLL were washed with PBS using a cell harvestor and treated with 1% glutaraldehyde in PBS for 5 min to provide for crosslinking between FgD and PLL. Plates were washed and 50 #1 of FgD with trace labeled ~3~I-labeled FgD was added for 18 h at room temperature on the plate agitator. BSA (25 mg/ml in PBS) was added (100 #l/well) for at least 2 h prior to washing in PBS and addition of radiolabeted antibody (50 #1 in BSA-PBS) for 18 h at room temperature with agitation. After washing, wells were removed from the plate by drilling out the well with a bit or by cutting the plate with a hot nichrome wire.

Calculations

Hybrid molecules that do not bind to antigen represent impurities in the labeled antibody. Calculations of amount of antibody added must be normalized to active antibody present to correct for the component due to hybrid molecules. For all calculations, total antibody was normalized to the maximum fraction of antibody that bound at high antigen concentrations (1.3 × 10 -7 M) as suggested by Trucco and De Petris (1981). This normalization is valid for situations in which antibody Ka is much greater than the reciprocal of antigen concentration (Trucco and De Petris, 1981). Data for 125I-labeled FgD was normalized according to the fraction bound by excess rat antiserum to FgD (Kennel et al., 1981). All labeled preparations were used within 2 weeks of radioiodination. For plate or bead assays, background binding of antibody to beads or plates coated with BSA and no FgD was subtracted from each experimental value. Molecular weights of 150,000, 110,000, and 55,000 were used for IgG, F(ab')2 and Fab', respectively, to calculate values in molar quantities. Affinity constants were determined graphically by extrapolation of double reciprocal plots (abscissa intercept = - 1 / K a and ordinant intercept-- 1/Ab max) (Accolla et al., 1980).

Results

A n tibody purity

Direct binding studies require well characterized antibody reagents. Hybridoma 104-14B was produced using P3-X63-Ag8 as the mouse myeloma parent. This parent line secretes yi,K immunoglobulin (Ig) and these Ig chains might be present in antibody preparations. Purified IgG, F(ab')2 and Fab' from hybridoma 104-14B were examined for purity by SDS-PAGE (Fig. 1). The IgG shows little (<5%) contamination on either reduced or unreduced samples (slots 1 and 4). F(ab')2 and Fab' preparations appear to be nearly pure heavy and light chains in the reduced sample profiles (slots 2 and 3); however, unreduced samples show that F(ab')2 (slot 5) contains a number of unidentified bands and in addition the Fab' sample (slot 6) contains 5-10% F(ab')2. The low molecular weight bands in unreduced samples (slots 5 and 6) have been cut out, reduced with mercaptoethanol, and rerun on SDS-PAGE. The reduced bands move to the normal light chain position, indicating that the fast moving bands are actually light chains with intrachain disulfide bonds in oxidizing conditions that hold the molecules in conformations which causes faster mobility on SDS-PAGE. The banding pattern of reduced IgG (Fig. 1, slot 1) demonstrated the presence of 2 distinct heavy chain bands as well as 2 light chain bands. The faster migrating heavy chain and the slower light chain have been identified as P3 y, and K, respectively (results not shown). The majority of P3 IgG could be separated from the rat antibody activity by ion exchange chromatography. The fact that P3 heavy and light chains were still present in the purified antibody indicated that hybrid Ig molecules could be present. 125I-labeled Fab' was used to determine if hybrid molecules had antibody activity.

Fig. 1. SDS-PAGE of 104-14B immunoglobulinpreparations on 10% acrylamide gel stained with Coomassie blue. Lanes 1-3 contains samples of IgG, F(ab')2 and Fab' reduced with mercaptoethanoi. Lanes 4-6 containedunreducedsamples. Molecules having antibody activity were purified from the population by affinity chromatography on Sepharose bound antigen. Analysis by 2-dimensional electrophoresis shows that P3 immunogiobulin chains were not present in the purified antibody preparation (Fig. 2B) whereas the total Fab' preparation had both P3 and rat chains. This data indicates that only molecules with rat heavy and light chains have antibody activity whereas hybrid molecules containing both rat and mouse chains are inactive. Spots in the light chain regions of the gels indicate that multiple molecular weight forms of light chains are present. This is probably the result of oxidative formation of intrachain disulfide bonds during the e!ectrophoresis procedure. The pattern for P3 Fab' (Fig. 2C) appears to be deficient in heavy chain fragments. Gel staining patterns show that this is not the case and the light band in the labeled preparation is probably due to selective radioiodination of P3 light chains.

Binding to fluid phase antigen Binding constants were determined for 104-14B IgG and its F(ab')2 and Fab' fragments using 125I-labeled FgD (Kennel et al., 1981). A constant m o u n t (100 ng)

I

L Chains

HFab' Chains

I

I

I

I

I

B

II

! I

I

I

I

I

I

I

l

C

I

4x5

5,5

I

6.5 pH

t

I

I

7.5

8~5

Fig. 2. Two-dimensional electrophoresis of ~25I-labeled Fab' preparations. A: autoradiograms of dried gels show 1251-labeled Fab' from 104-14B; B: 125I-labeled Fab' from 104-14B purified by immunoaffinity chromatography; and C: 1251-labeled Fab' from parent myeloma P3-X63-Ag8.

of antibody was added to FgD at different specific activities and antigen-antibody complexes separated from free antigen using a secondary antibody (Kennel et al., 1981). Results derived from double reciprocal plots of the data (Table I) show that TABLE I B I N D I N G C O N S T A N T S F O R A N T I B O D Y TO F L U I D PHASE 125I-LABELED F g D Ab form

Ab concentration (M)

M a x i m u m Ab binding capacity

Affinity constant (M- l )

(M) IgG F(ab') 2 F(ab')

6.7 X l 0 - I ° 9.14 X 10- l0 18.3 x 10 -10

2 . 1 x 1 0 ~0 3.3 x 1 0 - l0

6.1x10 9 6.0 X 10 9

5 . 7 X 10 -10

6 . 0 X 10 9

IgG as well as Fab' and F(ab')2 fragments have nearly identical affinity constants (6 × 10 9 M - 1 ) . For IgG, the maximum antigen binding capacity was roughly 1/3 of

the molar IgG concentration present or 1/6 of the possible antibody sites since IgG has 2 potential combining sites per molecule. The ratios of maximum binding capacity to theoretical available antibody sites for the 3 preparations were 0.16, 0.18, and 0.31 for IgG, F(ab')2 and Fab', respectively. These data are consistent with about 20% of the antibody sites being capable of binding fluid phase FgD. Thus inactive immunoglobulin is present in all 3 preparations. However, it does not affect the affinity constant determination, since in this case, the label is on the antigen (FgD). Binding to FgD on Sepharose beads 125I preparations of 104-14B IgG, F(ab')2 and Fab' were Jested for binding to FgD (10 ng) coupled to Sepharose beads. As little as 10 ng 125I-labeled IgG were sufficient to saturate available antigen sites on the FgD conjugated beads (Fig. 3). On a molar basis, IgG, F(ab')2 , and Fab' bound 0.22, 0.20, and 0.30 molecules per molecule of FgD present. A series of experiments were done using beads derivatized with different amounts of FgD. A typical set of data is depicted in Fig. 4. Fig. 4A shows binding of 125I-labeled 104-14B IgG over a 100-fold concentration range for 10 ng of FgD bound to 2/xl of Sepharose beads. Data on either side of this range are unreliable due to technical problems of specific activity or nonspecific binding of antibody to beads. This data replotted in double reciprocal fashion (Fig. 4B) gives a linear plot (correlation coefficient 0.997) indicating an apparent affinity constant of 2.9 × 10 9 M -1. Affinity constants and maximum antibody binding levels for experiments

4.0.

"~&O. O Z 0

m 2,0. >FZ

<

1.0.

o

Ib

2'0

i5

3'6 4'o

go

ANTIBODY ADDED(ng)

Fig. 3. Binding of 12SI-labeled IgO ( t I ) , F(ab') 2 (A A), or Fab' (D n ) from hybridoma 104-14B to 10 ng of FgD covalently coupled to Sepharose beads. The amount of antibody added was normalized to the maximum fraction of antibody bound. Background binding levels, a constant fraction of the input antibody, was subtracted in each case, IgG, 4.0%; F(ab')2, 2.0%; and Fab', 1.2%.

4.0- A A

.Y

3.0z onrl

2.0-

,b

IbO

,obo

ANTIBODY ADDED (ng)

2.0. o o x I

.o

1,0

.//

JI'" ....... 0.2

0.6 1.0 1/f (M - l ) x 10 IO

1,4

Fig. 4. Binding of mSI-labeled IgG from hybridoma I04-14B to l0 ng of FgD covalently coupled to 2/.tl

Sepharose beads. Binding was for 18 h at 4°C in 100 #1 final volume. A: saturation curve of avialable antigen sites by added antibody; B: double reciprocal plot of data in A.

using different concentrations of FgD are summarized in Table II. Affinity constants for all experiments range from 1 × 109 to 1.5 X 101° M - I with a general trend of higher values obtained when low antigen concentrations were used. The average value for I g G of 4 × 109 is nearly the same as that determined for FgD in the fluid phase (Kennel et al., 1981). The fraction of antigen which is capable of binding antibody varies with the total amount of FgD present and the antibody preparation used for the measurement (compare data columns I and 2, Table II). I g G yields the highest values for antigen activity followed by F(ab')2 and Fab'. Higher antigen concentrations gave lower values for fraction of active antigen, possibly due to steric effects.

Binding to FgD on polystyrene plates Initial studies had indicated that binding of antibody to F g D immobilized on plates was inefficient. For example, when screening hybridomas for antibody to

TABLE II BINDING CONSTANTS F O R ANTIBODY TO FgD COUPLED TO SEPHAROSE BEADS Ab form

FgD present (ng)

IgG IgG IgG

1 10 20

IgG

Max active FgD (ng)

Affinity constant ( X 109 M - l )

Affinity constant (mean -+- S.D. X 10-9 M - I )

5.1 6.3 4.5 2.9 1.0

3.96 -+ 2.1

100

1.25 3.7 8.3 12.7 35.7

F(ab')2 F(ab')2 F(ab')2 F(ab')2

1 10 20 100

0.30 1.70 7.1 18.2

15.0 8.3 8.3 1.0

8.15 -4-5.7

Fab' Fab' Fab' Fab'

1 10 20 100

0.25 1.0 4.0 8.5

7.4 5.0 2.2 1.7

4.1 -----2.7

FgD, an antigen concentration of 1 # g / w e l l was required whereas screens for antibody to tumor antigens (Kennel et al., 1981) could be done with less than 50 ng of cell bound antigens. Several potential 'carriers' or 'spacer molecules' were tested to determine if a different antigen presentation could enhance antibody binding. The

160-

~120. Z ::D 0

80

-v

< 40

-o

I--Z

o

0

0

~6o

26o 36o ANT~BOOY AODED C~I

4()0

5do

Fig. 5. Binding of 125I-labeled IgG from hybridoma 104-14B to 1 #g FgD immobilized on polystyrene plates. FgD was bound directly to plates ( 0 O) or to plates pretreated with 1 m g / m l PLL ([] O); phosphatidyl choline ( A A); phosphatidyl ethanolamine ( A &); CHCI 3 (V ~7); or PLL fixed with 1% glutaraldehyde ( I I ) . BSA attached to plates as a control hound less than 1 ng of antibody at all concentrations tested.

10 TABLE III BINDING CONSTANTS FOR ANTIBODY TO FgD BOUND TO POLYSTYRENE PLATES Carrier

FgD present (ng)

Max active FgD (ng)

-

10 100 1000 10 100 1000

0.24--+0.23 7.5 + 6 . 8 43.3 +6.7 1.34±0.9 11.9 -+8.9 102 -+91

-

PLL+G PLL+G PLL+G

Affinity constant

(XIO9M 1) 8.4+0.84 1.9 0.9-+0.3 8.3-+ 1.1 2.9± 1.8 1.6+1.1

2.6 +

total amount of FgD bound to plates was tested using trace labeled 131I-labeled FgD. These controls indicate that nearly 100% of the F g D added to wells remained bound to the plate after the washing procedures. When FgD levels were raised above 1 #g/well, a lower percentage of input F g D was bound to the plates indicating saturation of binding sites. Data in Fig. 5 indicate that only about 40 ng of ~25I-labeled 104-14B could be bound to 1/~g of FgD in a polystyrene plate. Coating the well first with lipids (phosphatidyl ethanolamine or phosphatidyl choline) gave no improvement of antibody binding and in fact binding was slightly less than for the CHC13 solvent control. The small amount of CHC13 present (1%) probably interacts with the plastic plate to give this effect. Pre-coating the plate with poly(L-lysine) (PLL) actually decreased the amount of antibody bound. However, if PLL coated plates were treated with 1% glutaraldehyde, to provide for subsequent crosslinking of FgD to PLL, the maximum amount of antibody that could be bound was increased more than 3-fold. Table III presents a summary of data for binding experiments using F g D bound directly to plates or bound using the PLL-glutaraldehyde 'spacer'. Affinity constants ranged from 0.9 to 8.4 × 109 M -~ with a direct correlation between the calculated value and the concentration of antigen used. These results are similar to those calculated for FgD presented on Sepharose beads in that the affinity constants varied somewhat with antigen concentration and the average values for the constants are in the same range (4-8 X 109 M -1) as for the fluid phase reaction. The major difference in antibody binding to antigen on different solid supports is the fraction of antigen present that is capable of binding antibody. The fraction of active FgD present as determined by I g G binding to the same amount of solid state F g D (100 ng) was about 7.5% for FgD bound to plates; 12% for FgD bound through PLL, and 36% for FgD bound to beads. The corresponding numbers for 10 ng of F g D were 2.4%, 13%, and 37%, respectively. These values indicate that F g D bound directly to plates is slightly less active than F g D bound through PLL and glutaraldehyde which in turn is less active than FgD coupled to Sepharose beads.

II Discussion

MoAb have introduced a new dimension in the understanding of antigen-antibody interactions. Previous attempts to determine affinity constants of reactions relied on indirect methods (Stankus and Leslie, 1976; Goidi et al., 1979). Binding of different antibodies in an antiserum gave a range of affinity constants (Steward et al., 1979). Analysis of binding of MoAb to antigen in fluid phase (AccoUa et al., 1980; Mason and Williams, 1980; Trucco and De Petris, 1981) has shown that, as predicted, MoAb gave binding patterns mathematically consistent with a homologous binding event. Binding of MoAb to antigens attached to a solid support has the added complications of possible bivalent interaction, steric effects, and differences in diffusive and reactive rate constants (Delisi, 1981). The question of bivalent binding has been addressed theoretically (Bell, 1978; Karush, 1976) and practically (Mason and WilLiams, 1980) as it relates to the problem of MoAb binding to cell surfaces and to immobilized proteins (Pincus and Rendell, 1981). Steric effects are unlikely to be significant factors at antigen concentrations lower than 10-6M (Delisi, 1981). Excluded volume effects could account for the trend of lower apparent affinity constants at higher FgD concentrations (Pincus and Rendell, 1981), however these effects should result in nonlinear double reciprocal plots which were not observed over the concentration ranges that were experimentally practical (see Fig. 4). Differences in diffusive and reactive rate constants have been described mathematically (Delisi, 1981). The major factor here is that ligand (antibody) encounters with solid state antigen may not result in reaction but that reorientation of the original or adjacent antigen on the solid support may occur before the ligand is completely dissociated from the solid particles (Delisi, 1981). These considerations require information about the ability of the antigen to reorient on the solid support and can only be approached experimentally. In this study MoAb to FgD of human fibrinogen has been used to compare the reaction of antibody with fluid phase versus solid state antigen. Binding constants using antigen presented in fluid phase (6 × 109 M - I ) varied little from those using solid state antigen ( ~ 1-8 × 109 M-l). The fluid phase reaction is monovalent since each FgD molecule has only one site reactive with monoclonal antibody. A significant amount of bivalent reaction in solid state binding should cause an increase the affinity constant (Karush, 1976), however, this is not always observed. No definite conclusion about valency of the solid state reaction can be drawn from these data. A comparison of binding to antigen attached to solid supports in different ways shows differences in the amount of antigen capable of binding antibody. That is, a large fraction of the solid state antigen was inactive for antibody binding. Studies using binding of 125I-labeled protein A to assess a solid state competition radioimmunoassay (Gee and Langone, 1981) have shown that the sensitivity of the assay depends on both the amount of solid phase antigen and its density. FgD attached to polystyrene plates either directly or through a poly(L-lysine)glutaraldehyde intermediate was 3-10-fold less active in binding MoAb than was FgD bound to Sepharose beads. The reason for this difference is not known. Steric factors, selective

12 d e s t r u c t i o n or shielding of the epitope from the a n t i b o d y due to n o n - r a n d o m a t t a c h m e n t of the antigen to the solid s u p p o r t m a y play a role. The physical p r e s e n t a t i o n of the antigen, i.e., o n beads or a flat surface, m a y also be a factor as could be the n a t u r e of the b o n d between a n t i g e n a n d the support. If the results o b t a i n e d here can be generalized, serious thought must be given to the way a n t i b o d i e s are used in solid state assays. Currently, m a n y assays use p r o t e i n a n t i g e n b o u n d directly to plates. E L I S A assays having sensitivities in the n a n o g r a m range require m i c r o g r a m quantities of a n t i g e n per well ( K a l m a k o f f et al., 1977). Assays using a n t i g e n attached to Sepharose beads m a y allow a more efficient use of valuable a n t i g e n reagents.

Acknowledgements I t h a n k Drs. Leaf H u a n g a n d G a r y Braslawsky for valuable criticism of the m a n u s c r i p t a n d N o r m a K w a a k for help in its preparation. M a n y of the a n t i b o d y reagents were prepared b y L i n d a F o o t e a n d Trish Lankford. F g D was generously provided b y Dr. J.P. Chen.

References Accolla, R.S., S. Carrel and J.-P. Mach, 1980, Proc. Natl. Acad. Sci. U.S.A. 77, 563. Ansari, A.A., L.M. Bahuguna and H.V. Mailing, 1978, J. Immunol. Methods 23, 219. Bell, G.I., 1978, Science 200, 618. Chen, J.P., H.M. Shurley and M.F. Vickory, 1974, Biochem. Biophys. Res. Commun. 61, 66. Cuatrecasas, P., 1970, J. Biol. Chem. 245, 3059. Delisi, C., 1981, Mol. Immunol. 18, 506. Elleman, T.C. and R.L. Raison, 1981, Mol. Immunol. 18, 655. Gee, A.P. and J.J. Langone, 1981, Anal. Biochem. 116, 524. Goidi, E.A., A. Cusano, R. Redner, J.B. Innes, M.E. Weksler and G.W. Siskind, 1979, Cell. Immunol. 47, 293. Kalmakoff, J., A.J. Parkinson, A.M. Crawford and B.R.G. Williams, 1977, J. Immunol. Methods 14, 73. Karush, F., 1976, in: Contemporary Topics in Molecular Immunology, Vol. 5, eds. H.N. Eisen and R.A. Reisfeld (Plenum Press, New York) p. 217. Kennel, S.J., J.P. Chen, P.K. Lankford and L.J. Foote, 1981, Thromb. Res. 22, 309. Kennel, S.J., L.J. Foote and P.K. Lankford, 1981, Cancer Res. 41, 3465. Laemmli, U.K., 1970, Nature (London) 227, 680. Lehtonen, O.-P. and M.K. Villjanen, 1980, J. Immunol. Methods 34, 61. Mason, D.W. and A.F. Williams, 1980, Biochem. J. 187, 1. Pincus, M.R. and M. Rendell, 1981, Proc. Natl. Acad. Sci. U.S.A. 78, 5924. Porath, J., J.-C. Janson and T. Laas, 1971, J. Chromatogr. 60, 167. Stankus, R.P. and G.A. Leslie, 1976, J. Immunol. Methods 10, 307. Steward, M.W., M.C. Reinhardt and N.A. Staines, 1979, Immunology 37, 697. Trucco, M. and S. De Petris, 1981, in: Immunological Methods, Vol. 2, eds. I. Lefkovits and P. Benvenuto (Academic Press, New York) p. 1.