Enhancement of ELISAs for screening peptides in epitope phage display libraries

JOURNAL OF lMMUNOLOGlCAL METHODS Journal of Immunological

Enhancement

Methods

197 (1996) 17 I- 179

of ELISAs for screening peptides in epitope phage display libraries Philippe Valadon, Matthew D. Scharff Received

10 January

*

1996; revised 4 June 1996: accepted 20 June 1996

Abstract The complete analysis of epitope phage display libraries requires sensitive assays capable of detecting peptides expressed on phage that have a wide range of affinities for antibody. We have compared two ELISAs, a ‘direct’ assay where the phage is captured by an anti-phage antibody and the peptide detected by the antibody used for screening, and a ‘reverse’ assay where the antibody used for screening is first coated on the well and the binding of phage detected by the anti-phage antibody. We demonstrate. by comparing two fUSE5 derived phage bearing five peptides reacting with the anti-cryptococcal polysaccharide antibody 2Hl. that the reverse ELBA is the more sensitive assay. Further, there is a limit in affinity. here around I PM, above which phage clones are negative by the direct ELISA. We describe an enhancement of the direct assay by mixing 2H I with j-fold excess of anti-heavy or anti-light chain antibody. The resulting polymerization of 2Hl induces an increase in antibody avidity that is responsible for the enhancement. The enhanced direct ELISA allowed rapid and sensitive detection of positive clones and is easily inhibited by free peptide. while the reverse ELBA is not. The enhanced ELISA has also been used successfully for immunological screening of intermediate libraries, allowing detection of rare positive clones that would otherwise be lost. The combination of the three ELlSAs, reverse. direct, and enhanced direct, should provide a way to rank phage clones into three classes: very low, low, and high affinity, providing information previously obtained only by the synthesis and testing of many peptides. K~JWU~S: Peptide library: Phage display; ELISA; Immunological

1. Introduction

(Parmley

Epitope phage display libraries are constructed insertion of a random peptide into either the

by pII1

Abbreviations: anti-K antibody. Anti-K light chain antibody: BSA. bovine serum albumin: ELISA, enzyme-linked immunosorbent assay: CNPS, Cpptomccus neoformnns polysaccharide; lD,,, 50% inhibitory dose; Ig, immunoglobulin; LB, Luria-Bertani medium: OD, optical density; ~111. coat protein encoded by gene III of filamentous phage; PEG, polyethylene glycol. * Corresponding author. Tel.: (718) 430-3527; Fax: (718) 430. 8574. 0023- 1759/96/$15.00 Copyright PII SOO22-1759(96)00133-O

screening

and Smith,

199 1) coat protein

1988) or the pVII1 (Felici of filamentous

phage

et al.,

fd or m 13. 1 X lo7 and

Such libraries usually contain between 1 X IO” clones, each of which displays a different peptide. Library screening with a monoclonal antibody requires repetitive selection and amplification cycles that progressively enrich for phage which interact specifically with that antibody (Scott and Smith, 1990; Cwirla et al., 1990; Devlin et al., 1990). The affinity between phage and antibody can vary over a wide range, with K, from nM to mM, and several methods have been used to study this

0 1996 Elsevier Science B.V. All rights reserved

172

2. Materials

and methods

2.1. Morloclonal saccharide

phw

antibodies

and cryptococcal

poly-

ti-yl

///// Reverse

ELISA

Direct

ELISA

Fig. 1. Scheme showing the principle of reverse ELISA on phage (on the left) and direct ELISA on phage (on the right) (ALP for alkaline phosphatase).

interaction: dot-blot, micropanning (Parmley and Smith, 19SS>, immunological screening (Christian et al., 1992). and ELISA (Scott and Smith, 1990; Barrett et al., 1992). ELISAs that detect the interaction of antibody and peptide on phage may take two different forms: a direct ELISA in which the phage is first immobilized to the plate by adsorption or capture with an anti-phage antibody and the antibody used for screening is then added: or a reverse ELISA in which the antibody used for screening is first immobilized on the plate and then phage is added followed by anti-phage antibody (Fig. 1). There is a clear preference for the direct ELISA in the literature, although the signal is sometimes low and requires amplification (Scott and Smith. 1990). We have used monoclonal antibodies which protect mice from lethal infection with Cc~,ytococcus ne~fOmzan~ (Mukherjee et al., 1992) to define peptide mimotopes for the cryptococcal capsular polysaccharide (Valadon et al., 1996). In the course of studies we compared direct and reverse ELISAs using two fUSE5 derived phage displaying five peptides that react with either low (@601) or high affinity (@Al) with the 2Hl antibody. We found that the direct ELISA was negative for low affinity phage whereas the reverse ELISA was positive. In this paper we describe those results and repel-t the use of anti-mouse light or heavy chain antibody to enhance the signal of the direct ELISA. This modification has proved very useful in studying the binding of positive clones and inhibition of their binding by the original antigen. Moreover, a similar enhancement of immunological screening greatly simplified the analysis of intermediate libraries by allowing the rapid screening of hundreds of phage colonies at a time.

2Hl is an IgG 1, K mouse monoclonal antibody to Cyptococcus neoformclns polysaccharide that protects mice from infection (Mukherjee et al.. 1993). Antibody was produced either in hybridoma culture medium or in BALB/c ascites and was used unpurified or after protein A purification (in high salt concentration according to Harlow and Lane, 1988). 2Hl Fab’ was prepared by papain digestion using a kit from Pierce Chemical Co. (Rockford, IL). Cryptococcal polysaccharide A (CNPS) (strain ATCC 48183) was a gift from Arturo Casadevall of this institution. 2.2. Phage and peptides Phage a601 and phage @Al were selected by 2Hl from a pII1 hexapeptide library and a pII1 decapeptide library respectively. Both libraries were derived from fUSE5 vector (Valadon et al., 1996; Scott and Smith, 1990). Phage @673 was a randomly picked clone used as a control bearing the hexapeptide RKVWVI. Phage were grown on K9lkan (Parmley and Smith, 1988) at saturation in Luria-Bertani medium (LB) in the presence of tetracycline 20 pg/ml. After 2 PEG precipitations, virions were resuspended in Tris/HCl 10 mM pH 7.5, NaCl 150 mM at the final concentration of 2 X 10” virions/ml. Analysis of phage stock DNA by agarose electrophoresis showed an unique band of singlestranded DNA. Peptides P601E (DGASYSWMYEA) and PA1 (GLQYTPSWMLVG) were synthesized by the peptide facility of the Albert Einstein Cancer Center. 2.3. Direct ELlSA The sheep anti-ml3 antibody was from 5 prime, 3 prime (Boulder, CO). All other commercial antibodies were from Southern Biotechnology Associates (Birmingham, AL). Assays were done in 96 well microtiter polystyrene plates (Coming Glass Works, Coming, NY) at 37°C with a reaction volume of 50 ~1 except for the blocking step where 100 ,c~l were

P. Valadon.

qf Imnwdogical

M.D. ScharJr/Joumal

used. The ELISA buffer was Tris/HCl 10 mM, pH 7.2, NaCl 150 mM (TBS). BSA 1% W/V, Tween 20 0.05% v/v, and NaN, 0.02% w/v. Coating was done in TBS alone and blocking was with TBS plus 1% BSA. The washing buffer did not contain BSA and was used in a series of three washes with the help of an automatic washer (Titertek from Flow Laboratories). Wells were first coated with a l/ 1000 dilution of the anti-ml3 antibody for 2 h at 37°C or overnight at 4°C and then blocked. Phage (5 X 10 I0 virions when not specified) were incubated for 2 h; after washing, the antibody to be tested (2 pg/ml when not otherwise specified) was incubated for 2 h with or without the enhancing antibody (3 equivalents w/w of anti-mouse light chain antibody unless otherwise indicated). After washing, an anti-mouse yl conjugated to alkaline phosphatase was added and incubated for 1 h (Fig. 1). Lastly, the phosphatase substrate p-nitrophenylphosphate (1 mg/ml in MgCl? 1 mM, Na,CO, 50 mM, pH 9.8) was added, and, after 1 h of incubation at 37°C absorbance at 405 nm was measured in an EL 320 microplate reader (Biotek Instrument, Winoosky, VT). Each assay was done in duplicate and the signal obtained with the control phage was subtracted (except in Table 1). For peptide inhibition studies, the antibody was preincubated with the peptide for 2 h at 37°C prior to addition to the plate. 2.4. ReL:erse ELISA Unless indicated, materials and reagents were the same as above. Wells were coated overnight with 50

Table I Comparison

of the different

ELISAs and immunological

Phage

screening @Al

Methods

197 (1996) 171-I

79

173

~1 of 2Hl 5 pg/ml in TBS and then blocked. The sandwich (Fig. 1) was constructed sequentially with 5 X 1O’O virions (unless otherwise indicated) for 2 h at 37”C, anti-ml 3 antibody l/4000 for 1 h at 37°C and lastly a l/4000 dilution of a donkey anti-sheep IgG conjugated to alkaline phosphatase (Sigma A5187) for 1 h at 37°C. OD measurements at 405 nm were done after 30 min of incubation in the presence of the phosphatase substrate. Each assay was done in duplicate and the signal obtained with the control phage was subtracted (except in Table 1). For peptide inhibition studies, the assay included an additional step after blocking in which the peptide was incubated in the well for 1 h at 37°C before addition of phage with an identical concentration of free peptide. 2.5. lrwnur~ological screening Immunological screening method was adapted from Sambrook et al. (1989). Single phage preparations or intermediate libraries were diluted to a final concentration of between 1 X lo5 and 1 X IO6 transducing units (TU)/ml in TBS with 0.1% gelatin. 200 ~1 of these preparations were mixed with 200 ~1 of competent K91 kan. After 10 min at room temperature, infected cells were transferred to a 2059 Falcon tube containing 5 ml of LB and tetracycline 0.2 pg/ml. After 30 min incubation at 37°C with shaking, 5 ~1 of tetracycline 20 mg/ml was added and the incubation continued for 1 h. The Falcon tubes were then chilled and stored at 4°C. Prior to use, the concentration of transduced bacteria still alive was

on both phage @Al and @601 @601

(0673

I .66 0.08 0.08 I .38

0.37 0.08 0.09 0.15

ELBA

Immunological

reverse direct direct with 2H I Fab’ enhanced direct

1.71 1.31 0.29 2.09

screening with 2H I alone with 2H 1 plus 3 X anti-y 1

+++ +++

* f + +

0.02 0.00 0.02 0.02

+ 0.03 rt_0.00 + 0.01 f 0.04

Not detectable +++

* f f f

0.03 0.00 0.01 0.00

Not detectable Not detectable

@673 is the control phage for ELISAs. A secondary anti-K antibody was used for Fab’ detection. Numbers correspond to the OD at 405 nm after a 30 min incubation in the presence of alkaline phosphatase substrate for the reverse assay and 60 min for the direct assay.

174

P. V&don.

M.D. Scharjf’/Jounml

of Irnrmmological Methods I97 (1996) 171-179

determined. Plates for immunological screening were grown overnight at a density of 200 colonies per petri dish and chilled at 4°C just before transfer. Plates were overlaid with 82 mm dry nitrocellulose filters (BA85 from Schleicher and Schuell, NH), and three asymmetric locations marked with a needle attached to a syringe containing India ink. Filters were promptly removed with two pairs of forceps and immersed in a large tray containing 100 ml Tris/HCl 10 mM pH 7.5, NaCl 150 mM, Tween 20 0.05% v/v (TNT) and 20% calf serum (Alpha calf serum from Hyclone, UT). The tray was gently agitated for 30 min and the TNT/calf serum 20% exchanged every 10 min. Filters were transferred to petri dishes and processed independently. 2H1 was first added to a concentration of 2 pg/ml with or without 6 pg/ml of anti-71 antibody in 8 ml of TNT/calf serum 20% and incubated for 2 h. Washing was repeated three times for 5 min each with 10 ml of TNT alone. The secondary antibody was an anti-mouse K chain conjugated to horseradish peroxidase (Southern Biotechnology Associates, AL), diluted to l/4000 and incubated for 1 h. After washing, filters were developed by chemiluminescence (ECL kit from Amersham. England).

3. Results 3.1. Direct ELBA for low affinity phage is negatirle Phage @601 was selected with the 2Hl IgGl mouse monoclonal antibody from a hexapeptide library. It displays the peptide SYSWMYE where the glutamate, located one amino acid C-terminal to the peptide insert in the pII1 protein, is required for effective binding. Phage @Al was selected from a decapeptide library and displays the peptide sequence LQYTPSWMLV (Valadon et al., 1996). In the reverse ELISA, where 2Hl antibody was first adsorbed to the plate (Fig. 1). @601 or @Al gave strong signals when their binding was detected by an anti-phage antibody and anti-sheep antibody conjugated to alkaline phosphatase (Table 1). However, in the direct ELISA where phage were bound to the plate first and then reacted with 2H 1. only @A 1 was detected by an anti-71 antibody conjugated to alkaline phosphatase. Similarly 2Hl Fab’ binding when

assayed by the direct ELISA gave a signal with @Al but did not with @601. Immunological screening of colonies on agar plates was performed by direct detection of phage transferred to a nitrocellulose filter. The results obtained with this assay were identical to those observed with the direct ELISA: @Al gave intense spots by chemiluminescence and @60 1 infected colonies were indistinguishable from non-specific clones (Table 1). These preliminary observations suggested that the affinity of 2Hl to @601 was too low to give a positive signal by direct ELISA and that if 2Hl could be made into a higher avidity antibody it would achieve a positive signal with @601. The antibody 2Hl is an IgGl and can interact with no more than two peptides in the direct assay whereas in the reverse ELISA a total of five interactions between the phage and multiple 2Hl antibodies on the plate can occur since the phage displays five peptide copies on its head. Thus, the reverse ELISA confers a higher avidity on 2Hl than the direct assay. 3.2. Enhancemerlt ofdirect ELISA by an anti-mouse light chain antibody In order to increase the avidity of the 2Hl antibody in the direct assay, we incubated 2Hl together with a polyclonal anti-mouse K light chain. Our expectation was that the resulting antibody polymerization could stabilize the antigen-antibody complex and therefore increase antibody binding as, for example, was observed following the Fc dimerization of mouse IgG3s upon binding to the antigen (Greenspan et al., 1987). The results obtained with @601 and different amounts of anti-K antibody at different 2H1 concentrations are presented in Fig. 2. The direct ELISA was greatly enhanced at almost all 2Hl concentrations tested. The enhancement increased with the amount of anti-K light chain at a fixed 2HI concentration, and reached a maximum at a ratio between 2 and 4 anti-K for 1 2Hl w/w. This maximum was independent of the 2Hl concentration, as seen in Fig. 2 which shows curves corresponding to four different 2Hl concentrations. The enhancing effect disappeared progressively with excess of anti-K antibody. and was no longer visible at a ratio of anti-K to 2H 1 of 64 : 1.

P. Valadon, M.D. Schafl/

Journal of Immunological Mefhods 197 (1996) I71- 179

175

A 2

1.6 1.2 8 0.8

64

32

16

8

ratio of anti-

4 K

2

1

0.5

antibody

0.125

0.03

per 2Hl (w/w)

Fig. 2. Effect of anti-rc antibody on 2Hl binding to phage @601 by ELBA. The anti-light chain antibody is first serially diluted and. after mixing with a given amount of 2H1, is immediately transferred to a 96 well plate coated with the phage ((0) 2Hl 2 Fg/mI. 10) I pg/ml, (I) 0.5 ,ug/mI, and t 0) 0.25 &g/ml).

0.8

0.6

3.3. Comparison L:erse ELISA

of enhanced

direct ELISA and reQ 0.4

Due to higher avidity of the antigen-antibody interaction mediated by the anti-u antibody, it was now possible to detect 2Hl binding to the low affinity phage @601 by a direct or a reverse assay. We compared the two assays in terms of sensitivity, phage number, and inhibition by free peptides. The reverse ELBA was more sensitive than the enhanced

]

I

4 IX 1011

I

x 1010

1x16

1x 108

1 x 107

virions per well Fig. 3. Effect of virion concentration on reverse and direct ELBA on both phage ((0) @601 by enhanced direct ELISA. (W) @601 by reverse ELISA, (0) @Al by direct ELISA, and (0) @Al by reverse ELBA).

0 l--r17,T//I 0.1

1

10 bwfW

100

1000

(PM)

Fig. 4. Inhibition of binding of 2Hl to phage @601 by the free peptides PA1 (@) or P601E (0) detected by enhanced direct ELISA (A), and reverse ELBA (B).

direct ELBA, although the background was higher (Table 1). The effect of phage concentration is presented in Fig. 3. There is an apparent titration between 5 X 10’ and 5 X lo9 virions per well for both assays, except for the high affinity phage @Al, when detected by reverse ELISA. Above 1 X IO” virions per well, the signal of both ELISAs are relatively independent of virion concentration. Inhibition data from both assays by free peptides are presented in Fig. 4. The inhibition of the enhanced direct ELISA on phage @601 was complete for both peptides at high concentrations and sigmoid, with 50% inhibition at a lower concentration for peptide PAI than peptide P601E (Fig. 4A). On the other hand, inhibition of the reverse ELISA was not sig-

176

Fig. 5. Immunological screenin, 0 with 2H1 of an intermediate library obtained during the screening of a decapeptide library by 2HI (library 3. Valadon et al.. 1996). On the left (A), the filter was incubated with 2H1 alone and on the right (B). a filter taken on the same plate was incubated with 3 equivalents of anti-y1 antibody together with 2HI. The plate contained 121 clones; the negative clones are not visible; the arrows indicate some of the new clones revealed by enhancement.

moidal and much higher peptide concentrations were needed to see a significant effect (Fig. 4B). If one considers the ID,,, inhibition of the reverse ELISA required ten times more peptide than was required for the enhanced direct ELISA for both phage. 3.4. Use screening

of‘ the eidzanceme~~t in immunological

The need for an efficient screening assay for positive phage clones led us to test the effect of the enhancing antibody on immunological screening of colonies on plates. For this purpose we used an intermediate library obtained during the screening of a decapeptide library by 2H1 (library 3, Valadon et al., 1996). Since the last step of the immunological screening uses an anti-K antibody conjugated to horseradish peroxidase. we examined the enhancing effect of adding anti-71 antibody together with 2Hl (Fig. 5). We choose the optimum ratio of 3 anti-y] antibody per 2Hl derived from the titration in Fig. 2. The same plate was used for both assays and the 122 clones were re-grown before the second transfer onto a nitrocellulose filter. The first screening was done in the presence of the enhancing anti-y] antibody to eliminate the possibility that apparent enhancement was due to very small clones that only became detectable after regrowing the phage on plates. In the absence of anti-y 1 antibody, 60 clones were clearly

positive. In the presence of the anti-y 1 antibody, the signal given by these clones was more intense and additional 11 clones were identified. This result indicated that the enhancement obtained with the direct ELISA was also effective for increasing the sensitivity of immunological screening for positive colonies on plates.

4. Discussion We have identified many peptide motifs that are potential mimotopes for the C~ptococcus neqformarzs capsular polysaccharide by screening peptide libraries with a large family of anti-polysaccharide monoclonal antibodies (Valadon et al., 1996). One of our major difficulties was the detection of binding of antibodies to phage with a wide range of affinities. In particular. a sensitive assay was needed which was capable of detecting the binding of rare clones with a low affinity for a given antibody. We were also interested in whether the signal with a given assay would predict the relative affinity of the peptide for antibody. For these reasons, we tried to increase the sensitivity of direct assays that are routinely used to detect binding to displayed peptides but are often negative with low affinity phage. Two phage were selected, @Al and @601, as representatives of high affinity and low affinity clones isolated

P. Valudon. M.D. Schatff/ Jounral

of

by the antibody 2Hl. These two phage behaved differently by ELISA. Both were positive for 2Hl with the reverse ELISA but only @Al was positive by direct ELISA. Consistent with the apparent better binding of 2Hl to @Al, 2Hl Fab’ gives a positive signal with @Al but not with @601. The corresponding free peptides have an ID,, for the binding of 2Hl Fab’ to an immobilized peptide of 1.4 PM and 2.0 PM, respectively, as measured by plasmon surface resonance (Valadon et al., 1996). In the case of the fUSE5 vector that was used to construct the two phage, all copies of the pII1 coat protein are recombinants, and peptides are displayed on the phage head. When the reverse assay is used, several peptides of a single phage can be simultaneously bound by multiple antibodies to the antibody coated well, resulting in a dramatic increase of the avidity of the interaction between the phage and the antibody. As we observed, both phage of low and high affinity give a strong signal under these conditions. Several additional low affinity phage like @601 have been isolated and they were all negative by direct ELISA except when the antibody is IgM, IgA, or IgG3, all of which have higher avidities than the other IgG subclasses (results not shown). This observation points to avidity as the limiting factor in the sensitivity of the direct assay. Barrett et al. (1992) found a similar impact of avidity on phage binding. Their phage ELISA was analogous to our reverse ELISA and two phage displaying peptides with either low K, (7.1 nM) or high K, (8300 nM) gave positive signals. With monovalent binding by the use of the Fab’ alone, this group has not been able to detect direct binding for K,s greater than 500 nM by ELISA. Although our system is not directly comparable (different k,, , koff, antibody, and library), it presents a similar cut-off around 1 PM in the peptide affinities. Above a given K,, the antibody is unable to maintain its binding to the phage and the direct assay is negative. In an attempt to detect low affinity clones by direct ELISA by increasing the 2Hl avidity, 2Hl was polymerized with a polyclonal anti-light chain antibody. In the presence of the two antibodies, a signal appeared for low affinity clones like @601. The signal for @Al was also increased. Three attributes point to the formation of a large complex between the phage head, 2H1, and the anti-K anti-

imnwrologicnl

Methods I97 (1996) 171-179

177

body: (1) the presence of a maximum OD at a ratio of three anti-~ antibodies to one 2Hl; (2) the independence of 2Hl concentration at this maximum; and (3) the disappearance of enhancement when either anti-K antibody or 2Hl were in excess. There is presumably a particular molecular relationship between the three partners when the enhancement is at a maximum. Because the anti-K chain antibody is polyclonal, we cannot be certain what this relationship is. We have successfully enhanced the direct ELISA by other methods: 2Hl plus anti-heavy chain antibody revealed by an anti-K antibody conjugated to alkaline phosphatase; 2Hl plus biotinylated anti-K antibody followed by streptavidin conjugated to alkaline phosphatase, or 2Hl mixed with anti-K antibody conjugated with alkaline phosphatase. So far, enhancement of a direct ELISA has been achieved with a three fold excess of anti-light chain antibody with 20 different monoclonal antibodies (results not shown). These monodonal antibodies even include IgM, IgG3, and IgA isotypes. Examination of Fig. 2 shows that enhancement is significant when there is a 1-8-fold excess of anti-light chain per antibody. This wide range indicates that a systematic determination of the maximum ratio for each antibody is unnecessary. The enhancement not only makes it possible to study peptide inhibition on phage of low affinity (see below), but also makes it possible to screen colonies on plates. As seen in Fig. 5, immunological screening of colonies on plates was enhanced by the anti-y1 antibody. This proved to be crucial for the screening of the hexapeptide library with 2H1, where initially no high binding mimotope was detected, and for detecting phage positive with several antibodies (Valadon et al., 1996). Enhancement of immunological screening was especially useful during the analysis of early intermediate decapeptide libraries where peptide diversity was high but only a few clones were positive and many were of low affinity. Several findings demonstrate that the reverse ELISA is more sensitive than the direct ELISA: the shorter incubation time needed to reach an equivalent OD; a better detection of the highest affinity clones at very low virion concentration (Fig. 3); and the presence of positive clones only seen with the reverse assay (data not shown). Because the interaction between phage and antibody is of high valency,

it is particularly difficult to inhibit the reverse ELISA. Barrett et al. (1992) have made similar observations. Some aspects of the system may make the reverse assay more difficult to set up, in particular the requirement for purified antibody to coat ELISA plates. Although other reverse immunological screening techniques have been described, they require complicated sandwiches such as those described by Skerra et al. (1991). For these reasons, direct assays are often preferred. The enhancement mediated by the anti-heavy or anti-light chain provides an altemative direct assay which is capable of analyzing the binding of low affinity phage. Comparison of ID,, obtained by surface plasmon resonance (Valadon et al., 1996) and enhanced direct ELISA (Fig. 4A) shows differences in the direction of higher peptide concentrations required to inhibit the direct ELISA (for example. ID,, of 1.4 FM by plasmon surface resonance versus 1.5 PM by direct ELISA for @Al and 2.0 PM versus 20 PM respectively for @601). It is possible that the increased avidity of the complex between the anti-K chain and 2Hl is responsible for higher peptide concentration required. Differences in the loft. or the k,,, of @Al and @601 may also contribute to differences observed between the two phage and the two assays. Nevertheless, comparison of four different peptides shows that the concentrations required for 50% inhibition are in the same order with both techniques (Valadon et al.. 19961, suggesting that the direct ELISA is a reasonable assay for comparing phage in terms of relative affinity. Finally, observation of a large collection of 2Hl binding clones shows a clear correlation between the direct and the enhanced direct ELISA (Valadon et al., 1996). 2Hl binding phage can therefore be divided in three classes: very low affinity phage where only the reverse ELISA is positive, low affinity phage where enhancement is required, and high affinity phage which are positive by direct ELISA. In conclusion, we describe here a simple method to increase the sensitivity of the direct detection of antibody binding to phage. The addition of a three fold excess (w/w> of anti-light chain or anti-heavy chain antibody to the antibody preparation greatly increases the signal of direct ELISAs and of the immunological screening of colonies on plates. The enhanced assays allow for rapid screening and isola-

tion of rare clones among low affinity phage. These clones are of major importance in numerous areas such as epitope mapping or the characterization of mimotopes of non-peptidic substances.

Acknowledgements We thank Drs. David Beenhouwer and Nancy Green for editorial assistance. This study was supported by grants from the National Institutes of Health (CA39838 and POlAI33184 (M.D.S.)). and the Harry Eagle Chair for Cancer Research from the National Women’s Division of the Albert Einstein College of Medicine (M.D.S.). P.V. was partially supported by the Philippe Foundation.

References Barrett. R.W.. &it-la. S.E.. Ackerman, M.S.. Olxm. A.M.. Perer\. E.A. and Dower. W.J. (1991) Selective enrichment and characterization of high affinity ligands from collections of ~-andom peptides on filamentous phage. Anal. Biochrm. X-1. 357-364. Christian. R.B., Zuckermann. R.N., Kerr. J.M.. Wang, L. and Malcolm, B.A. (1992) Simplified methods for construction. assessment and rapid screening of peptide libraries in hactcriw phage. J. Mol. Biol. 227. 71 I-718. Cwirla. S.E.. Peter>. EA.. Barrett, R.W. and Dower, W.J. (1990) Peptides on phage: it vast library of peptides for identifying ligands. Proc. Natl. Acad. Sci. USA 87. 637X-6387. Devlin, J.J., Panganiban, L.C. and Devlin. P.E. (1990) Random :I source of specific protein binding peptide libraries: molecule\. Science 249. 101-306. Felici. F.. Castagnoli, L., Musacchio. A., Jappelli. R. and Cesareni, G. ( 1991 1 Selection of antibody ligands from a large library of oligopeptides expressed on :I multivalent exposition vector. J. Mol. Biol. 222. 301-310. Greenspan. N.S.. Monafo. W.J. and Davie. J.M. (1987) Interaction of IgC3 anti-strcptococcal group A carbohydrate (GAC) antigroup A vaccine: enhancing and body with streptococcal inhibiting effects of anti-GAC. anti-iaotypic, and anti-idiotypic antibodies. J. Immunol. 138. X-1-91-. Harlow. E. and Lane. D. (1988) Antihodirs: A Laboratory Manual. Cold Spring Harbor Laboratory Press. Cold Sprin: Hnrbar. NY. Mukherjre, J., Casadevall, A. and Scharff. M.D. (1993) Molecular characterization of the hutnoral response\ to Cr\‘/?/oc.oc~ctr.r wofw0nluul.tinfection and glu~uronoxylomannan-trtanu\ toxoid conjugate immunization. J. Exp. Med. 177. I lO5- I 116. Mukherjee, J., Scharff. M.D. and Caaadevall. A. (I Y92) PI-otective murine monoclonal antibodies to Cryptococcua neoformans. Infect. Immun. 60. 4534-354 I.

P. V&don.

M.D. Scharff/Joumai

of Imnnrrwlogicul Methods 197 (1996) 171-179

Parmley, S.F. and Smith, G.P. (1989) Filamentous fusion phage cloning vectors for the study of epitopes and design of vaccines. Adv. Exp. Med. Biol. 251, 215-218. Sambrook, J.. Fritsch. EF. and Mania& T. (1989) Molecular cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press. Cold Spring Harbor. NY. Scotr. J.K. and Smith, G.P. (1990) Searching for peptide ligands with an epitope library. Science 249. 386-390.

179

Skerra. A., Dreher, M.L. and Winter G. (1991 ) Filter screening of antibody Fab fragments secreted from individual bacterial colonies: specific detection of antigen binding with a twomembrane system. Anal. Biochem. 196. 151-155. Valadon. P., Nussbaum. G., Boyd L.F., Margulies D.H. and Scharff D. (1996) Peptide libraries define the fine specificity of anti-polysaccharide antibodies to C~pfococcus neoformnns. J. Mol. Biol., in press.