Selection of phage-display library peptides recognizing ethanol targets on proteins

Alcohol 25 (2001) 201 – 209

Selection of phage-display library peptides recognizing ethanol targets on proteins Helen Annia,*, Olga Nikolaevaa, Yedy Israela,b a

Thomas Jefferson University, Jefferson Medical College, Department of Pathology-Anatomy & Cell Biology, and Alcohol Research Center, 275 Jefferson Alumni Hall, Philadelphia, PA 19107, USA b University of Chile, Department of Pharmacological & Toxicological Chemistry, and Millennium Institute, Santiago 1111, Chile Received 15 February 2001; received in revised form 2 May 2001; accepted 3 May 2001

Abstract There is a forthcoming link between chronic alcohol consumption and proteins covalently modified by ethanol metabolites and their antibodies. To identify sensitive probes of protein – ethanol conjugates, we screened for the ethanol-altered protein domains with a phagedisplay combinatorial peptide library. In principle, recognition of the epitopes by the library peptides occurs through protein – protein interactions. A general screening, M13-based library with 109 random sequences of linear heptameric peptides was used. The peptides were displayed in five copies each, as fusion proteins with phage’s minor coat protein III. They were located on one end of the surface of the phage particles. The targets were a model protein, streptavidin, and protein – ethanol conjugates (hydroxyethyl radical- or acetaldehyde-modified bovine serum albumin). They were either immobilized on a surface by direct coating or affinity captured on floating beads. An enriched library of phages with the tightest peptide binders for each target was selected and amplified in a multiple-cycle biopanning in vitro procedure. Binders were characterized by DNA sequencing of the corresponding phages and by counter-screening with positive and negative targets in either an enzyme-linked immunosorbent assay or plaque assay. We obtained the HPQ motif for streptavidin and two unique subsets of peptides that recognized each ethanol target with a selectivity of two orders of magnitude above the carrier protein and controls. The application of biopanning processes, coupled with phage-display peptide libraries on biological fluids and tissues, could provide a systematic mapping of protein – ethanol conjugates and supply a means for early diagnosis and prognosis of chronic alcohol consumption in human beings. D 2002 Elsevier Science Inc. All rights reserved. Keywords: Ethanol; Hydroxyethyl radicals; Acetaldehyde; Bovine serum albumin; Protein conjugates; Phage-display peptide libraries; Heptameric peptides; Biopanning; Screening; Disease-specific epitopes; Biomarkers; Alcoholism

1. Introduction Antibodies against various covalent adducts of ethanol metabolites with proteins, as well as authentic ethanolrelated protein adducts, have been discovered in different levels in ethanol-fed animals and human beings consuming alcohol acutely or chronically. We showed that antibodies were generated by chronic ethanol administration to mice that recognized acetaldehyde (AcH) epitopes in synthetic AcH – protein adducts (Israel et al., 1986). Reversibly, adducts of AcH –hemoglobin were detected by using rabbit * Corresponding author and requests for reprints. Tel.: +1-215-5035064; fax: +1-215-923-9263. E-mail address: [email protected] (H. Anni). Editor: T.R. Jerrells

antibodies against synthetic AcH – protein adducts in the blood of human beings after acute ethanol ingestion (Niemela et al., 1990) and of alcohol abusers (Niemela & Israel, 1992). Aldehyde-derived protein modifications were also observed by using antibodies in the liver of ethanol-fed rats, micropigs, and human alcoholics, importantly even before any apparent pathological changes on histological examination (Niemela et al., 1994). Along the same lines, chronically ethanol-fed rats produced antibodies distinguishing the modified epitopes in adducts of another ethanol metabolite, hydroxyethyl radical (HER) (Moncada et al., 1994). Proteins classified as early targets for AcH binding in vivo include, among others, hemoglobin and the microsomal liver ethanolinducible cytochrome P4502E1 isozyme (Behrens et al., 1988). Acetaldehyde adducts found in liver and plasma of ethanol-fed rats had respectively a relatively long adduct

0741-8329/01/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved. PII: S 0 7 4 1 - 8 3 2 9 ( 0 1 ) 0 0 1 6 4 - 1

202

H. Anni et al. / Alcohol 25 (2001) 201–209

half-life of 2 –5 weeks (Nicholls et al., 1994). Consequently, such adducts were considered as possibly useful markers of ethanol consumption. Furthermore, knowledge about the proteins that carry covalent modifications of ethanol metabolites, the corresponding antibodies, or both would be important in creating a set of ethanol-targeted proteins for evaluation of ethanol intake (Niemela & Israel, 1992) and probably their role in hepatotoxicity (Clot et al., 1996). Attempts have been made in enzyme-linked immunosorbent assays (ELISAs) to correlate antibody reactivity against AcH – proteins, amounts of AcH – albumin, and conventional markers of alcohol intake in human beings (Holstege et al., 1994; Svegliati-Baroni et al., 1994; Worrall et al., 1994, 1996, 1998). However, the titers of antibodies against ethanol– protein adducts were mostly low in vivo, and quantitation of adducts was difficult. This might be due to different affinity and accessibility of the in vivo adducts to antibodies produced usually with differing synthetic antigens and, by the same token, of the in vivo antibodies to synthetic antigens in the immunochemical and immunohistochemical assays. It is important to note that because intestinal bacteria also produce ethanol, ethanol-modified proteins may be recognizable as ‘‘self’’, rather than as new antigenic entities, thus generating antibodies with low titer and specificity. The technology of phage-display combinatorial peptide libraries was applied initially in epitope-mapping studies, in which specific binders to a molecule of interest (target) were sought after (for a review, see Cortese et al., 1995). In the simplest format, a plate was directly coated with the target and incubated with the library. After unbound and nonspecifically bound phage was washed away, specifically bound phage was eluted and amplified. This procedure, so-called biopanning, was repeated multiple times under more stringent conditions until 10– 20 top-scoring strong binders were identified in an enriched pool. In the end, individual clones were isolated, characterized by DNA sequencing and ELISA. The platform of phage-display peptide libraries has been used for the identification of both disease-specific epitopes (Folgori et al., 1998) and antigens of viral infections (BirchMachin et al., 2000), as well as for the discovery of peptide immunogens for production of protective antibodies (Shakib et al., 2000). In recent years, emphasis has been given to novel uses of libraries for the characterization of unique protein – protein interactions (Bottger et al., 1997; Dedman et al., 1993; Han & Kodadek, 2000), detection of enzyme inhibitors (Hyde-DeRuyscher et al., 2000; Ke et al., 1997; Ohkubo et al., 1999; Sato et al., 1999; Sperinde et al., 2001; Zdanovsky et al., 2001), and development of receptor antagonists (Norris et al., 1999). Uses of the libraries extend to gene therapy with the discovery of cell-specific ligands for targeted gene delivery with altered adenoviral vectors (Nicklin et al., 2000; Romanczuk et al., 1999). Phagelibraries were exploited even by nonbiological systems (for a review, see Johnsson & Ge, 1999) to recognize metal, metal oxide, and semiconductor surfaces for material assembly purposes (Whaley et al., 2000).

The aim of this study was to isolate phageborne peptides that are able strongly and specifically to recognize ethanolmodified protein epitopes. Synthetic peptide sequences could substitute for phage-peptides as baits in unbiased screening of serum samples for existing and novel antigens, antibodies, or both and serve as diagnostic tools of chronic alcohol consumption in human beings.

2. Materials and methods 2.1. Materials A filamentous coliphage-display peptide library Ph.D-7 kit was purchased from New England BioLabs (Beverly, MA); M13 wild-type phage, from American Type Culture Collection (Rockville, MD); protein A-coated polystyrene microspheres, from Bangs Laboratories (Fischers, IN); and ethanol, from Pharmco (Brookfield, CT). Antibodies were obtained as follows: horseradish peroxidase (HRP)conjugated anti-major phage coat protein VIII (pVIII) M13 monoclonal antibody, from Amersham Pharmacia Biotech (Piscataway, NJ); sheep polyclonal anti-M13 antibody, from 5 Prime ! 3 Prime (Boulder, CO); and rabbit anti-bovine serum albumin (BSA), from Fitzgerald (Concord, MA). The HRP-conjugate of donkey anti-sheep immunoglobulin G (IgG) whole molecule, 2,2’-azino-bis(3-ethylbenz-hiazoline-6-sulfonic acid) (ABTS), dimethyl pimelimidate (DMP), ethanolamine, triethanolamine, fraction V BSA g-globulin free, acetaldehyde, ferrous ammonium sulfate, sodium cyanoborohydride, and hydrogen peroxide 30% (wt./ wt.) solution were purchased from Sigma Chemical Co. (St. Louis, MO). 5-bromo-4-chloro-3-indoyl-b-D-galactoside (Xgal), isopropyl-b-D-thiogalactoside (IPTG), Luria Broth (LB) base, LB agar, and Select agar were obtained from Gibco BRL (Rockville, MD). Polyethylene glycol-8000 (PEG), Tween-20, Costar 96-well ELISA plates, and 60  15 mm Becton Dickinson Petri dishes were supplied from Fisher Scientific (Pittsburgh, PA). Protein concentration of BSA adducts was measured with the Micro BCA Protein Assay kit obtained from Pierce (Rockford, IL). All targets were prepared by using sterile technique, and biopanning solutions and plates were sterile. 2.2. Preparation of adducts The BSA –HER (referred to as target A) was prepared by incubation of 8 mg/ml of BSA in phosphate-buffered saline (PBS), pH 7.4, for 1 h at room temperature, with 100 mM ethanol, 1 mM ferrous ammonium sulfate, 2 mM ethylenediamine tetraacetic acid (EDTA), and 1 mM hydrogen peroxide added slowly with stirring. The solution was dialyzed against PBS overnight at 4C. Reduced BSA – AcH (target B) was prepared by incubation of 8 mg/ml of BSA in PBS, pH 7.4, for 1 h at room temperature, with 1 mM AcH. Reduction of BSA – AcH was done by addition of

H. Anni et al. / Alcohol 25 (2001) 201–209

0.2 M sodium cyanoborohydride, and the solution was incubated further for 4 h. Unused reactants were removed by dialysis. The BSA –HER control #1 is an oxidized BSA sample that was prepared as BSA –HER, with ethanol, but not iron and peroxide, omitted. Likewise, BSA – AcH control #2 is a reduced BSA sample, with no AcH added in the incubation mixture but with cyanoborohydride. The BSA control #3 was processed in parallel with other samples without any additions, whereas BSA control #4 was taken from the bottle. Targets A and B used in solution (beads) biopanning were prepared with some modifications in incubation time (72 h), temperature (37C), BSA concentration (2 mg/ml), and AcH concentration (10 mM). Also, the reaction was done with mild shaking, and dialysis was substituted by ultrafiltration. All targets were aliquoted and stored at – 80C until used. 2.3. Biopanning procedure for selection of streptavidinspecific phage peptides Two wells of a 96-well ELISA plate were coated with streptavidin (150 ml per well of 100 mg/ml of streptavidin in 0.1 M carbonate buffer, pH 8.6) by nonspecific hydrophobic and electrostatic interactions (direct binding). After incubation overnight at 4C in a humidified container, excess streptavidin was shaken out. Immobilized streptavidin was blocked (350 ml per well of 0.5% BSA in 0.1 M carbonate buffer, pH 8.6, containing 0.1 mg/ml of streptavidin) for 2 h at 4C. Streptavidin was added to the blocking solution to complex any biotin in BSA. Excess blocking solution was discarded, and the wells were washed six times with Trisbuffered saline (TBS) containing 0.1% (vol./vol.) Tween (TBST-0.1%). Phages (100 ml per well, 10 ml of 2  1011 virions in TBST-0.1%) were added to the blocked wells and incubated for 1 h at room temperature. Detergent was included in the solution to reduce phage – phage interactions. Unbound phages were removed, and nonspecifically bound phages were washed away 10 times with TBST-0.1%. Specifically bound phages were eluted competitively by incubation for 1 h with a streptavidin ligand, biotin (100 ml per well of 0.1 mM biotin in TBS). The eluate from first biopanning round (EL1) (100 ml) was amplified 2  106 times and purified (aEL1). Both eluates (1 ml) were plated out for titer determination. In subsequent rounds, the Tween concentration was raised to 0.5%. As input for the second and third round of selection, respectively, 2  1010 aEL1 virions and 1  1012 aEL2 virions were used. An enrichment of 300-fold was calculated in one cycle, EL2/EL1. After three cycles, 22 individual clones were picked, along with aEL3 mixture, and analyzed. 2.4. Phage titers Eluates (10 ml, 10-fold serial dilutions in LB in the range 101 –107 for unamplified eluates, or 103 – 1014 for amplified eluates and plaques) were mixed with a plating

203

fresh culture (200 ml, 10 ml of LB inoculated with a single bacterium colony in midlog phase with OD600  0.5) and poured with agarose top (3 ml) onto LB agar XGal/IPTG plates. Library phages with a random peptide insert were derived from the vector M13mp19 that carries the lacZa gene and were blue in color once plated in LB XGal/IPTG medium. The infected cells were incubated overnight at 37C, and blue plaques expressed as plaque-forming units (pfu) were counted on lawns of bacterial cells. 2.5. Amplification and purification of phages and plaques Eluates (  100 ml), the entire pool of selected phages in EL3 (10 ml), and individual plaques were added to a bacterial culture that had been grown overnight and diluted 100-fold in LB. The infected cells were incubated in a rotating incubator with vigorous shaking for up to 5 h at 37C. After the bacteria were removed by centrifugation, the phages in the supernatant were precipitated twice with 20% (wt./vol.) PEG containing 2.5 M NaCl, and the purified pellet was dissolved in 200 ml of TBS and used for sequencing. The phage supernatant was stored at 4C, or at 20C in 50% glycerol for long periods. For ELISA analysis, phage supernatant (5 ml) was grown in a 100-fold diluted overnight bacterial culture and purified by PEG precipitation. 2.6. Purification of sequencing templates Single-stranded phage DNA was precipitated from amplified plaques or aEL3 (500 ml) by incubation for 10 min with iodide buffer and ethanol (100 ml of 10 mM Tris – HCl buffer, pH 8.0, with 1 mM EDTA [TE] and 4 M sodium iodide; and 250 ml of ethanol). The pellet was washed with 70% ethanol and resuspended in TE buffer. It was used as the sequencing template with a 96 gIII primer by the automated dideoxynucleotides method (ABI 377 sequencer with dye primer chemistry). 2.7. ELISA of selected peptides for streptavidin binding One row per selected clone in the plate was coated with streptavidin (200 ml per well of 100 mg/ml of streptavidin in 0.1 M carbonate buffer, pH 8.6) and blocked (300 ml per well of 0.5% BSA in 0.1 M carbonate buffer, pH 8.6), as in the biopanning procedure. A second uncoated row was blocked to test the binding of clones to BSA-coated plastic (nonspecific binding). By increasing the BSA concentration in the blocking buffer to 2.5% –5%, nonspecific background binding was diminished. Blocked wells were washed extensively with TBST-0.5%. Phages (200 ml per well, 1  1012 or 2.5  1011 virions in TBST-0.5%) were first diluted (fourfold serial dilutions) in separate blocked rows to prevent adsorption to the target during dilutions, transferred, and incubated for 2 h at room temperature with the target. After unbound phages were poured off, the wells were

204

H. Anni et al. / Alcohol 25 (2001) 201–209

washed six times with TBST-0.5%. The immobilized phages were incubated for 1 h at room temperature with an antipVIII HRP conjugate (1:5,000 dilution in 200 ml of blocking buffer). Excess antibody was washed away six times with TBST-0.5%. Bound phages were detected after a 30-min incubation at room temperature with HRP substrate (200 ml per well, 22 mg ABTS/100 ml of 50 mM sodium citrate, pH 4.0, and 0.05% hydrogen peroxide) following absorbance at 405 nm.

2.8. Biopanning procedure for selection of ethanol adductsspecific phage peptides Ethanol adducts were added to ELISA wells (150 ml per well of 3.3 mg/ml of BSA – HER or 2.7 mg/ml of BSA – AcH in 0.1 M carbonate buffer, pH 8.6) and blocked (350 ml per well of 3% BSA in 0.1 M carbonate buffer, pH 8.6), as described in detail in the biopanning for streptavidin peptides in section 2.3. Blocked wells were washed with TBST-0.1% and incubated with the phage library overnight at 4C. Bound phages from washed wells were eluted by nonspecific disruption of binding (100 ml of 0.2 M glycine –HCl, pH 2.2, and 1 mg/ml of BSA) for 10 min at room temperature and quickly neutralized (15 ml of 1 M Tris – HCl, pH 9.1). Eluted phages (EL1) from each target (90 ml) were amplified 105 times and input in the second biopanning round (3.6  108 aEL1 virions). The two following biopanning cycles were done in the presence of TBST-0.5%. Individual plaques picked from titration of eluate from all three rounds, alongside with the corresponding eluate mix, were amplified for sequencing and ELISA analysis, as described in sections 2.5 and 2.6. Alternatively, a molecular scaffold was devised for binding of phages to targets in solution. Protein A –coated beads were washed and covalently cross-linked by means of a DMP spacer for Fc-directed attachment to a polyclonal anti-BSA antibody (capture antibody), according to the manufacturer’s instructions. The selector antigen (0.5 mg/ml of BSA – HER or BSA –AcH) was incubated with the beads (bead –protein A-DMP-anti BSA antibody) overnight with rocking at 4C. The bead complex (bead –protein A-DMP-anti BSA antibody– BSA conjugate) was washed five times with TBST-0.1%, and the phage library was added and incubated for 3 h with rocking at room temperature. After exhaustive washing for 15 times, bound phages were eluted (0.2 M glycine – HCl, pH 2.2, containing 1 mg/ml of dry milk) and neutralized. The EL1 phages were amplified and mixed with preincubated BSA carrying beads (bead – protein A-DMP-anti BSA antibody– BSA) for a cross-adsorption cleaning step. The nonbound phages in the supernatant were combined with a single wash of absorbed phages and used for the second biopanning round performed in TBST-0.5% with dry milk blocking. The EL2 mix and individual plaques were grown for sequencing and plaque assay.

2.9. ELISA of selected peptides for binding to ethanol adducts Wells were coated with a panel of positive (BSA – HER, BSA –AcH) and negative (BSA controls #1– #4, buffer) targets. The assay was performed as described for streptavidin in section 2.7, with minor modifications. Wells were coated (150 ml per well of 100 mg/ml of the positive or negative targets in 0.1 M carbonate buffer, pH 8.6), blocked (300 ml per well of 3% BSA in 0.1 M carbonate buffer, pH 8.6), and exposed to phages (150 ml per well of 1  1011 virions in TBST-0.5% and twofold serial dilutions thereafter) for 2 h at room temperature. Differential binding of selected phage peptides with known DNA sequence to a variety of targets was measured by following the absorbance at 405 nm. 2.10. Plaque assay of selected peptides for binding to ethanol adducts Phages (100 ml per well, 1011 – 1012 virions in TBST0.5%) were added to various positive and negative targets (150 ml per well of 500 mg/ml of the target in 0.1 M carbonate buffer, pH 8.6). Instead of measuring the number of bound phages to the different targets by absorbance in ELISA, phages were recovered after a 3-h incubation at room temperature with rocking, acid-elution, and neutralization. Phage particles (10 ml, 100 – 105 serial dilutions) were mixed with a fresh bacteria culture (200 ml) and plated out overnight at 37C to determine phage numbers. For phages selected by biopanning on beads, plaque assay was also performed in solution (beads). 2.11. Statistical analyses The streptavidin binding experiment was performed in duplicate. All ELISA measurements were done in triplicate, and average values with standard deviation are presented. Plaque assays numbers are expressed as average values with standard deviation from five plating dilutions.

3. Results 3.1. Identification of motif containing streptavidin-specific phage peptides A phage display library of linear heptameric peptides was probed in a control experiment with a model target, streptavidin, for streptavidin-binding peptides. Streptavidin was directly coated onto ELISA wells, and in the presence of 0.1% Tween specifically captured phages from the library were eluted with a streptavidin ligand, biotin, and amplified. This biopanning process was repeated twice by using each time the enriched pool of phages obtained from the previous cycle. More stringent conditions were applied in subsequent

H. Anni et al. / Alcohol 25 (2001) 201–209 Table 1 Amino acid sequences of peptides displayed by selected phages for BSA – HERa Peptide sequence EL1A-p1b EL1A-p2 EL1A-p3 EL1A-p4 EL1A-p5 EL1A-p6

Q S L Q L L

G L P N N T

A R S P Y N

A D Y H H H

P T Q S R G

F P A P P T

S L I R S A

EL2A-p1 EL2A-p2 EL2A-p3 EL2A-p4 EL2A-p5 EL2A-p6

F W N F G Y

G A I H Q N

N M D E L L

P S N N P P

G P P W S W

A T N P S L

A K K S P N

EL3A-p

F

H

E

N

W

P

S

a,b

EL1A, EL2A, and EL3A stand for eluates from the three biopanning rounds against target A, BSA – HER (bovine serum albumin – hydroxyethyl radical). A total of 27 plaques, p, were picked: 6 plaques from each of the first two rounds and 15 from the third round. All 15 plaques from EL3 had the same sequence. Consensus residues are in bold boxed letters.

cycles by increasing the concentration of detergent from 0.1% to 0.5% Tween to identify phage peptides that were strong binders of streptavidin. Because the DNA encoding the fusion protein resides within the virion, DNA sequencing can easily identify each selected phage peptide. After three biopanning cycles against the same target, streptavidin, the peptide sequence of all selected individual clones, as well as the eluate mix, were characterized. All phage peptides exhibited the consensus sequence with the strongly conserved H P Q motif (Devlin et al., 1990; Katz, 1999). In

205

ELISA analysis it was validated that all discovered clones recognized streptavidin with a specificity of about two orders of magnitude above that for nonspecific binding to BSA-coated plastic (data not shown). Nonspecific background binding of these phage peptides to streptavidin was comparable to that of M13 phage lacking a peptide insert (data not shown). The high affinity recognition between streptavidin and the selected phage peptides afforded a light coating of 10 mg/ml of streptavidin on ELISA wells and a strong signal (OD > 1.0) was obtained (data not shown). 3.2. Identification of phage peptides that recognize the AcH or HER modification on BSA Shifting to the search for phageborne peptide sequences that identify preferentially ethanol epitopes created by ethanol metabolites on proteins, we used as paradigm BSA – HER and BSA – AcH. Both were biopanned, as streptavidin, for three cycles by directly coating ELISA wells, incubation with library phages, and elution of the strongest binders with low pH. Phage– peptides selected by biopanning for BSA – HER (target A) displayed no motif after the first two biopanning rounds. A single sequence, F H E N W P S, emerged after the third round for all 15 independent clones, which appeared in one clone (EL2A-p4) in the second biopanning round (Table 1). Screening of the EL3A clone against an array of positive and negative targets to determine binding efficiency showed high readings (OD>1.5) in ELISA analysis. However, there were no differences between either positive versus negative target-coated or target-coated versus BSAblocked uncoated wells by ELISA (data not shown), nor

Fig. 1. Direct binding biopanning. Binding profile of clones selected for target A, BSA – HER (bovine serum albumin – hydroxyethyl radical) or target B, BSA – AcH (bovine serum albumin – acetaldehyde) to a panel of positive and negative targets. Only plaques picked from eluate in the third biopanning round (EL3) were assayed. For EL3A peptides (derived by biopanning for target A, BSA – HER), BSA – HER was the positive target and BSA – AcH, BSA – AcH control #2 (reduced BSA), BSA – HER control #1 (oxidized BSA), and unmodified BSA were the negative targets. Correspondingly, for EL3B phage peptides (derived by biopanning for target B, BSA – AcH) the positive target was BSA – AcH, whereas negative targets were BSA – HER, BSA – HER control #1 (oxidized BSA), BSA – AcH control #2 (reduced BSA), and unmodified BSA. Recovered phage numbers were expressed as average plaque-forming units per milliliter (pfu/ml).

206

H. Anni et al. / Alcohol 25 (2001) 201–209

by a more sensitive method of plaque assay (Fig. 1). We concluded that most probably the selected EL3A clone was a super-binder of polystyrene surfaces that grew preferentially in the third biopanning round, and a different biopanning strategy for BSA – HER in solution was developed (see Fig. 2). In the same way, 23 phage-peptides were selected by biopanning for BSA – AcH (target B) (Table 2). However, no clear consensus sequence became evident after three rounds of biopanning. Rather, two unrelated motifs appeared to be more prominent: X P R A Y X X and X G T T A S L. Binding of all ELB peptides to targets was weak in ELISA (data not shown). Neither increasing target coating (200 ml per well of

1 mg/ml), adding more phages (1012 virions), and increasing the incubation time for phages (3 h at room temperature), nor switching detection system (polyclonal anti-M13 antibody 1:1,500 dilution combined with the corresponding HRPconjugate 1:10,000 dilution), improved appreciably the absorbance readings for phage peptides selected for BSA – AcH. For this reason, the ELB phage – peptides were also screened by plaque assay, which is a more sensitive method and even a single phage could be calculated. Fig. 1 shows the comparative binding profile of all EL3B clones. The EL3Bp5 and EL3B-p10 clones shared the same sequence, G P R A Y H E, and distinguished clearly the AcH epitope on BSA (dark bar) from the HER modification and a variety of

Fig. 2. Solution binding biopanning. Binding profile of clones selected for target A, BSA – HER (bovine serum albumin – hydroxyethyl radical) or target B, BSA – AcH (bovine serum albumin – acetaldehyde) to both positive and negative targets. A. Specificity of ELB clones (biopanning for target B, BSA – AcH) to its positive BSA – AcH target and negative targets BSA – AcH control #2 (reduced BSA) and unmodified BSA. B. Specificity of ELA clones (biopanning for target A, BSA – HER) to its positive BSA – HER target and negative targets BSA – HER control #1 (oxidized BSA) and unmodified BSA.

H. Anni et al. / Alcohol 25 (2001) 201–209 Table 2 Amino acid sequences of peptides displayed by selected phages for BSA – AcHa Peptide sequence EL1B-p1b EL1B-p2 EL1B-p3 EL1B-p4

S S L I

P L Q P

R M P S

P A T I

P N Y P

Y L S T

P I L M

EL2B-p1 EL2B-p2 EL2B-p3 EL2B-p4 EL2B-p5 EL2B-p6

N N A T W S

M M S P P L

E E F T T P

R R A T M P

V V Y T L N

P P T R S W

V V T S K P

EL3B-p1 EL3B-p2 EL3B-p3 EL3B-p4 EL3B-p5 EL3B-p6 EL3B-p7 EL3B-p8 EL3B-p9 EL3B-p10 EL3B-p11 EL3B-p12 EL3B-p13

L I W S G A Y L Y G T D T

R G T P P A G P G P P V N

H L W D R P T A T R Q E N

P P P S A M T P T A T L K

T L P R Y A A P A Y W R P

L R A L H L S F S H P P I

P S Y R E F L R L E L L F

a,b

EL1B, EL2B and EL3B stand for eluates from the three biopanning rounds against target B, BSA – AcH (bovine serum albumin – acetaldehyde). A total of 23 plaques, p, were picked: 4 plaques from the first round, 6 plaques from the second round, and 13 plaques from the third round. Consensus residues are in bold boxed letters.

control samples that were not modified by ethanol metabolites (oxidized BSA, reduced BSA, unmodified BSA). The plaque assay results showed the successful pulling out of a peptidic bait for AcH epitopes and reconfirmed the nonspecific binding capability of the EL3A clones (see Fig. 1). The cause of the failed biopanning attempt with the EL3A clones could be an inaccessible protein modification site in BSA – HER. This might be due to target denaturation along the surface by hydrophobic interactions of the protein to plastic, steric blocking of the epitope by protein binding, or both. To avoid target inaccessibility and improve binding kinetics, a new biopanning strategy was devised to select phage – peptides for BSA –HER and with increased sensitivity for BSA – AcH, as described in the Materials and Methods section 2.8 – alternative protocol. First, in this approach, a target captured on an affinity matrix was prebound to library phages in solution. Second, BSA – HER and BSA –AcH conjugates, with an increased number of epitope sites of ethanol metabolites to BSA, were used. Third, blocking was done with non-fat dry milk, as an alternative for BSA, to decrease nonspecific binding. Fourth, peptide sequences successfully bound to target in the first round of biopanning were eluted, amplified, and cross-adsorbed on BSA in an additional cleaning step. The nonbinders went through a second round of biopanning with

207

the same target as in the first round, but under more stringent conditions. Selected phages were isolated, sequenced (data not shown), and analyzed by plaque assay (Fig. 2). After three rounds of selection for BSA –AcH, including the cross-adsorption step, phage –peptide sequences were identified that could discriminate between an AcH modification on BSA and its controls. Clone 9 showed such a differential binding to BSA – AcH versus unmodified BSA or reduced BSA by a factor of 100 (Fig. 2A). Similarly, several peptide sequences were fished out of the library for BSA – HER biopanning that were specific for HER (Fig. 2B). Clone 12 was 300 times more selective for HER modification compared with findings for unmodified BSA or oxidized BSA. In neither case did a consensus sequence surface. More biopanning cycles and additional biopanning protocols should be carried out to ensure that a motif exists, before in vivo experiments are attempted.

4. Discussion 4.1. Streptavidin-specific sequences Phage-display peptide libraries are a source of binding molecules that can recognize specifically through protein – protein interactions any particular target, biological or nonbiological. They could be especially useful in mapping studies of protein modifications in providing baits for small protein domains for which antibody titers are restrictively low. Our findings initially confirmed the ability of a linear heptameric phage library to (1) identify by molecular recognition in iterative biopanning procedures the strongest binders to a model protein, streptavidin, (2) characterize the sequences of the binding phage –peptides, and (3) validate by ELISA analysis the specificity of the interaction to the binding domain of biotin, a streptavidin ligand. 4.2. Ethanol metabolites –specific sequences By analogy to streptavidin biopanning, the modifications of ethanol metabolites on a carrier protein such as BSA were treated as targeting moieties for phage-display peptides. In the case of biopanning for BSA – HER, a single peptide sequence of a super-binder EL3A clone was detected after three biopanning rounds in ELISA directbinding format (see Table 1, Fig. 1). It has been reported that biopanned phage peptides that bound to BSA-blocked plastic surfaces had no sequence similarity but were enriched in Y/W and devoid of C residues (Adey et al., 1995), or had a W-rich consensus sequence with a motif of natural adhesive protein fibrinogen, W _ X XW _ X X XW _ (Gebhardt et al., 1996). The EL3A sequence, F H E N W _ P S, had no C but exhibited a single W residue. Only EL2Ap6 (1 of 27) had both W and Y residues. A total of 18 of 27 clones in this biopanning had a single W residue, and 3 of

208

H. Anni et al. / Alcohol 25 (2001) 201–209

27 clones had a single Y residue. Our plastic binding peptide sequence points to a dissimilar plastic binding mechanism for EL3A from the fibrinogen type invoked earlier. We noted that with the exception of one peptide, EL1A-p6, all other 26 sequences in this biopanning contained at least one P residue. Biopanning by direct-binding for BSA –AcH did not provide a consensus sequence; instead, an abundance of P residues in 19 of 23 selected binding clones and positions X2 –X7 was observed (see Table 2). Two clones, EL3B-p5 and EL3B-p10, had the same sequence, G P R A Y H E. These clones were specific for the AcH epitope on BSA and recognized neither the HER modification nor unmodified carrier BSA, oxidized BSA, reduced BSA in a plaque assay (see Fig. 1). Interestingly, this differential binding of EL3B-p5/p10 to BSA – AcH and recognition of the AcH modification was not detected by anti-M13 antibody-based ELISA analysis. The choice of biopanning method and strategy is crucial in the selection of binding sequences that are specific for a particular domain and lack nonspecific binding properties (Barbas et al., 2001; D’Mello & Howard, 2001). In the previous experiments, biopanning was performed by direct ELISA binding of the target. By switching biopanning procedures to solution binding (beads) and combining it with plaque assay successful phage peptide sequences were pulled out of the library. Clone 9 (see Fig. 2A), which was selected as a strong binder to BSA –AcH, had a sequence A H P L M L Y and an ability to recognize specifically the AcH modification on BSA by two orders of magnitude above its controls. With the exception of residues L and M, the same residues participated in the binding of BSA – AcH in the direct-binding biopanning for BSA – AcH, G _P R A _Y _ H _ E (see Fig. 1). This indicates that heptameric phagepeptides function rather as mimotopes of the AcH domain. In this biopanning, a novel sequence S A L Y R H S in clone 12 (see Fig. 2B) was also selected that was a strong binder to BSA – HER and selective for HER modification compared with findings for its controls. In summary, a phage display peptide library was used for the first time as a source of specific-binding molecules to modifications by ethanol metabolites (HER, AcH) on proteins (BSA). By using this powerful heptameric library and biopanning optimizations for our system, specific phage peptides were discovered with a selectivity of more than two orders of magnitude typically achieved for a particular ethanol modification. Although other carrier proteins besides BSA have not been used, phage peptides selected by biopanning for ethanol metabolites did not bind to BSA, nor to its oxidized and reduced forms. Furthermore, peptides biopanned for BSA –HER did not recognize BSA – AcH and vice versa. These small heptameric peptides adopt no secondary structure and are expected to be in coil conformation and useful probably for the recognition only of small or partial domains. Ethanol epitopes might be more specifically recognized by the use of libraries of longer

peptides such as 12-mers that assume a secondary structure. Phage-display combinatorial peptide library biopanning could be applied not only to HER- or AcH-modified proteins in vitro, but also to serum samples obtained from ethanol-fed animals and alcoholics, where freedom from biases toward any particular protein or ethanol modification is sought.

Acknowledgments This study was supported by National Institute on Alcohol Abuse and Alcoholism (NIAAA) grants R37 AA10630, P50 AA07186, and T32 AA07463. We thank Mr. Pavlo Pristatsky for his technical assistance on solution biopanning and plaque assays.

References Adey, N. B., Mataragnon, A. H., Rider, J. E., Carter, J. M., & Kay, B. K. (1995). Characterization of phage that bind plastic from phagedisplayed random peptide libraries. Gene 156, 27 – 31. Barbas, C. F. III, Burton, D. R., Scott, J. K., & Silverman, G. J. (2001). Phage display. A laboratory manual. Cold Spring Harbor, MD: Cold Spring Harbor Laboratory Press. Behrens, U. J., Hoerner, M., Lasker, J. M., & Lieber, C. S. (1988). Formation of acetaldehyde adducts with ethanol-inducible P450IIE1 in vivo. Biochem Biophys Res Commun 154, 584 – 590. Birch-Machin, I., Ryder, S., Taylor, L., Iniguez, P., Marault, M., Cegile, L., Zientara, S., Cruciere, C., Cancellotii, F., Koptopoulos, G., Mumford, J., Binns, M., Davis-Poynter, N., & Hannant, D. (2000). Utilisation of bacteriophage display libraries to identify peptide sequences recognised by equine herpesvirus type 1 specific equine sera. J Virol Methods 88, 89 – 104. Bottger, A., Bottger, V., Garcia-Echeverria, C., Chene, P., Hochkeppel, H. K., Sampson, W., Ang, K., Howard, S. F., Picksley, S. M., & Lane, D. P. (1997). Molecular characterization of the hdm2-p53 interaction. J Mol Biol 269, 744 – 756. Clot, P., Albano, E., Eliasson, E., Tabone, M., Arico, S., Israel, Y., Moncada, C., & Ingelman-Sundberg, M. (1996). Cytochrome P4502E1 hydroxyethyl radical adducts as the major antigen in autoantibody formation among alcoholics. Gastroenterology 111, 206 – 216. Cortese, R., Monaci, P., Nicosia, A., Luzzago, A., Felici, F., Galfre, G., Pessi, A., Tramontano, A., & Sollazzo, M. (1995). Identification of biologically active peptides using random libraries displayed on phage. Curr Opin Biotechnol 6, 73 – 80. Dedman, J. R., Kaetzel, M. A., Chan, H. C., Nelson, D. J., & Jamieson, G. A. Jr. (1993). Selection of targeted biological modifiers from a bacteriophage library of random peptides. The identification of novel calmodulin regulatory peptides. J Biol Chem 268, 23025 – 23030. Devlin, J. J., Panganiban, L. C., & Devlin, P. E. (1990). Random peptide libraries: a source of specific protein binding molecules. Science 249, 404 – 406. D’Mello, F., & Howard, C. R. (2001). An improved selection procedure for the screening of phage display peptide libraries. J Immunol Methods 247, 191 – 203. Folgori, A., Luzzago, A., Monaci, P., Nicosia, A., Cortese, R., & Felici, F. (1998). Identification of disease-specific epitopes. Methods Mol Biol 87, 195 – 208. Gebhardt, K., Lauvrak, V., Babaie, E., Eijsink, V., & Lindqvist, B. H.

H. Anni et al. / Alcohol 25 (2001) 201–209 (1996). Adhesive peptides selected by phage display: characterization, applications and similarities with fibrinogen. Pept Res 9, 269 – 278. Han, Y., & Kodadek, T. (2000). Peptides selected to bind the Ga180 repressor are potent transcriptional activation domains in yeast. J Biol Chem 275, 14979 – 14984. Holstege, A., Bedossa, P., Poynard, T., Kollinger, M., Chaput, J. C. Houglum, K., & Chojkier, M. (1994). Acetaldehyde-modified epitopes in liver biopsy specimens of alcoholic and nonalcoholic patients: localization and association with progression of liver fibrosis. Hepatology 19, 367 – 374. Hyde-DeRuyscher, R., Paige, L. A., Christensen, D. J., Hyde-DeRuyscher, N., Lim, A., Fredericks, Z. L., Kranz, J., Gallant, P., Zhang, J., Rocklage, S. M., Fowlkes, D. M., Wendler, P. A., & Hamilton, P. T. (2000). Detection of small-molecule enzyme inhibitors with peptides isolated from phage-displayed combinatorial peptide libraries. Chem Biol 7, 17 – 25. Israel, Y., Hurwitz, E., Niemela, O., & Arnon, R. (1986). Monoclonal and polyclonal antibodies against acetaldehyde-containing epitopes in acetaldehyde – protein adducts. Proc Natl Acad Sci USA 83, 7923 – 7927. Johnsson, K., & Ge, L. (1999). Phage display of combinatorial peptide and protein libraries and their applications in biology and chemistry. Curr Top Microbiol Immunol 243, 87 – 105. Katz, B. A. (1999). Streptavidin-binding and -dimerizing ligands discovered by phage display, topochemistry, and structure-based design. Biomol Eng 16, 57 – 65. Ke, S. H., Coombs, G. S., Tachias, K., Corey, D. R., & Madison, E. L. (1997). Optimal subsite occupancy and design of a selective inhibitor of urokinase. J Biol Chem 272, 20456 – 20462. Moncada, C., Torres, V., Varghese, G., Albano, E., & Israel, Y. (1994). Ethanol-derived immunoreactive species formed by free radical mechanisms. Mol Pharmacol 46, 786 – 791. Nicholls, R. M., Fowles, L. F., Worrall, S., de Jersey, J., & Wilce, P. A. (1994). Distribution and turnover of acetaldehyde-modified proteins in liver and blood of ethanol-fed rats. Alcohol Alcohol 29, 149 – 157. Nicklin, S. A., White, S. J., Watkins, S. J., Hawkins, R. E., & Baker, A. H. (2000). Selective targeting of gene transfer to vascular endothelial cells by use of peptides isolated by phage display. Circulation 102, 231 – 237. Niemela, O., & Israel, Y. (1992). Hemoglobin – acetaldehyde adducts in human alcohol abusers. Lab Invest 67, 246 – 252. Niemela, O., Israel, Y., Mizoi, Y., Fukunaga, T., & Eriksson, C. J. (1990). Hemoglobin – acetaldehyde adducts in human volunteers following acute ethanol ingestion. Alcohol Clin Exp Res 14, 838 – 841. Niemela, O., Parkkila, S., Yla-Herttuala, S., Halsted, C., Witztum, J. L., Lanca, A., & Israel, Y. (1994). Covalent protein adducts in the liver as a result of ethanol metabolism and lipid peroxidation. Lab Invest 70, 537 – 546.

209

Norris, J. D., Paige, L. A., Christensen, D. J., Chang, C. Y., Huacani, M. R., Fan, D., Hamilton, P. T., Fowlkes, D. M., & McDonnell, D. P. (1999). Peptide antagonists of the human estrogen receptor. Science 285, 744 – 746. Ohkubo, S., Miyadera, K., Sugimoto, Y., Matsuo, K., Wierzba, K., & Yamada, Y. (1999). Identification of substrate sequences for membrane type-1 matrix metalloproteinase using bacteriophage peptide display library. Biochem Biophys Res Comm 266, 308 – 313. Romanczuk, H., Galer, C. E., Zabner, J., Barsomian, G., Wadsworth, S. C., & O’Riordan, C. R. (1999). Modification of an adenoviral vector with biologically selected peptides: a novel strategy for gene delivery to cells of choice. Hum Gene Ther 10, 2615 – 2626. Sato, A., Yamamoto, S., Ishihara, K., Hirano, T., & Jingami, H. (1999). Novel peptide inhibitor of ecto-ADP-ribosyl cyclase of bone marrow stromal cell antigen-1 (BST-1/CD157). Biochem J 337, 491 – 496. Shakib, F., Hooi, D. S., Smith, S. J., Furmonaviciene, R., & Sewell, H. F. (2000). Identification of peptide motifs recognized by a human IgG autoanti-IgE antibody using a phage display library. Clin Exp Allergy 30, 1041 – 1046. Sperinde, J. J., Choi, S. J., & Szoka, F. C. Jr. (2001). Phage display selection of a peptide DNase II inhibitor that enhances gene delivery. J Gene Med 3, 101 – 108. Svegliati-Baroni, G., Baraona, E., Rosman, A. S., & Lieber, C. S. (1994). Collagen – acetaldehyde adducts in alcoholic and nonalcoholic liver diseases. Hepatology 20, 111 – 118. Whaley, S. R., English, D. S., Hu, E. L., Barbara, P. F., & Belcher, A. M. (2000). Selection of peptides with semiconductor binding specificity for directed nanocrystal assembly. Nature 405, 665 – 668. Worrall, S., de Jersey, J., Wilce, P. A., Seppa, K., Hurme, L., & Sillanaukee, P. (1996). Relationship between alcohol intake and immunoglobulin A immunoreactivity with acetaldehyde-modified bovine serum albumin. Alcohol Clin Exp Res 20, 836 – 840. Worrall, S., de Jersey, J., Wilce, P. A., Seppa, K., Hurme, L., & Sillanaukee, P. (1998). Comparison of carbohydrate-deficient transferrin, immunoglobulin A antibodies reactive with acetaldehyde-modified protein and acetaldehyde-modified albumin with conventional markers of alcohol consumption. Alcohol Clin Exp Res 22, 1921 – 1926. Worrall, S., Jersey, J. D., Wilce, P. A., Seppa, K., & Sillanaukee, P. (1994). Studies on the usefulness of acetaldehyde-modified proteins and associated antibodies as markers of alcohol abuse. Alcohol Alcohol 2S, 503 – 507. Zdanovsky, A. G., Karassina, N. V., Simpson, D., & Zdanovskaia, M. V. (2001). Peptide phage display library as source for inhibitors of clostridial neurotoxins. J Protein Chem 20, 73 – 80.