Journal of Immunological Methods, 147 (1992) 201-210

201

C 1992 Elsevier Science Publishers B.V. All rights reserved 0022-1759/92/$05.00

JIM 06206

Monoclonal antibody binding to peptide epitopes conjugated to synthetic branched chain polypeptide carriers Influence of the carrier upon antibody recognition F_ Hudecz • and M.R. Price

b

• Remlrch Group of Peptide Chemistry, Hungarian Academy of Science, EotvOs L. University, Budapest 112, Hungary H-1518, and b Cancer Remlrch Campaign Laboratories, University of Nottingham, Nottingham NG72RD, UK (Received 15 August 1991, accepted 30 October 1991)

The peptide CAP 0 T R PAP G has been linked covalently to defined branched polypeptides with a polylysine backbone and side chains of oL-alanine or o-leucine-oL-alanine oligopeptides. The peptide was coupled via its N terminal cysteine to the side chains of the macromolecular carrier to ensure uniform orientation. The compounds were subjected to compositional analyses to characterise the degrees of substitution and secondary structural studies were performed using circular dichroism spectroscopy. The peptide selected for investigation contains the immunodominant sequence PDT R P A P which is expressed in the protein core of epithelial mucins. It is to this region that many anti-mucin monoclonal antibodies bind (Burchell et aI., 1989; Price et aI., 1990a,b). With these characterised constructs, it has been possible to evaluate the influence of secondary structure upon the binding of monoclonal antibodies which recognise short linear sequences in the synthetic antigenic peptide. The fmdings are relevant to the design and construction of synthetic immunogens and vaccines as well as to the production of synthetic analogues of clinically relevant antigens (in this case, epithelial mucins associated with breast and ovarian carcinomas). Key words: Carrier-dependent anb'body recognition of peptide epitopes: Branched chain polypeptide carrier

Correspondence to: M.R. Price, Cancer Research Campaign Laboratories, University of Nottingham, Nottingham NG7 2RD, UK. Abbreviations: XAK, poly(Lys-{Xi-DL-Ala.»): AK, po1y[LYS..(DL-Ala3.1»); o-LAK, po1y[Lys-{o-LeuO.98-oL-Ala3.1»);

DP , number average of the degree of polymerisation; CG, C

A lip D T R PAP G; SPDP. N-succinimidyl 3-{2pyridyldithio)propionate; PSS. 2-pyridyl-disulphide group: DTT. dithiothreitol; AK-CG, po1y[Lys-{CGj-OL-Ala3.1»); 0LAK-CG, po1y[Lys-{CG;-D-LeuO.98-oL-Ala3.1»); 'OS, averaae degree of substitution: NMP. N-methylpyrrolidone; TFA, trifluoroacetic acid; CD, circular dichroism; PBS, phosphatebuffered saline, pH 7.3; ABTS, 2,2'-azino-di-{3-ethyl-benzthiazoline-sulphonic acid) diammonium salt.

Introduction Many monoclonal antibodies against human milk fat globule membranes or breast carcinoma cells recognise epithelial mucins (Burchell et aI., 1983; Hilkens et aI., 1984; Price et aI., 1985; Lan et al., 1987). The monomeric protein core of these high molecular weight glycoproteins is largely composed of tandem repeats of a 20 amino acid sequence (Gendler et aI., 1988). Analysis of antibody-defined epitopes within the protein core has identified a major immunodominant domain, PDT R PAP within the repeated 20 residue

202

motif and synthetic peptides containing this sequence are recognised by several anti-mucin monoclonal antibodies with differing epitope specificities (Burchell et aI., 19S1J; Xing et aI., 1989; Price et aI., 1990a, b). It is of note that the epitope mapping tests performed in these studies have established that these mucin-specific monoclonal antibodies react with relatively short linear sequences of 3-5 amino acid residues. In the present investigation. these systems have been further developed to analyse the recognition of synthetic peptide epitopes by monoclonal antibodies. In the design of immunoassays for synthetic peptides, reliable immobilisation of small peptide antigens to a solid phase frequently presents difficulties and can result in a loss or inappropriate presentation of epitope(s). The methods of peptide immobilisation include covalent linking of the peptide to reactive groups on a solid phase (Larsson et aI., 1987) and physical adsorption of the ligand (Engvall, 1980) or its conjugated form (Lutz et aI., 1990; Verschoor et aI., 1990). Covalent coupling of antigenic determinants to protein or polypeptide carriers may significantly reduce the efficacy of antibody recognition by alteration of the accessibility of the epitope. Alternatively, hapten molecules correctly coupled to a rationally designed macromolecular carrier may provide more appropriate epitope orientation and density for application in solid-phase antibody binding techniques. In order to investigate more precisely the influence of the carrier effect on epitope recognition, two structurally related branched polypeptides were selected. Both peptides, with a general formula poly[Lys-(X,-DL-Ala m )] (XAK), where i < I and m - 3, have polycationic characteristics, but their conformational properties in solution differ significantly (Hudecz et aI., 1985). We now describe the preparation and secondary structure analysis of branched polypeptide conjugates linked covalently to a peptide, CAP D T R P A P G, which contains the immunodominant domain of human epithelial mucins. In the construction of conjugates, the antigenic peptide was coupled through its N terminal cysteine to give a uniform orientation of determinants linked to the polypeptide polymer (Fig. 1). The binding proper-

Fig. I. Schematic representation of hranched polypeptide. C A PDT R PAP G conjugates (A) AK-CG and (8) D-LeU (D-LAK-CG).

ties of various monoclonal antibodies with different epitope recognition capacities have been investigated and compared with the original mucin glycoprotein. These studies indicate that the secondary structure of the conjugate, determined by carrier properties, has a regulatory role in efficient recognition of epitopes in the conjugated, antigenic PDT R PAP peptide motif.

Materials and methods Nomenclature The abbreviations used in this paper follow the rules of the IUPAC-IUB Commission of Biochemical Nomenclature (1972) in accord with the recommended nomenclature of graft polymers (IUPAC-IUB Commission of Biochemical Nomenclature, 1984). Peptides Branched polypeptides were synthesised as previously described (Hudecz and Szekerke, 1980; Hudecz et aI., 1985). Briefly, poly(Lys) was produced by the polymerization of N"-carboxy-NE"(benzyloxycarbonyI}-lysine anhydride under conditions that allowed an average degree of polymerisation of approximately 400-500. After cleavage of the protecting groups short oligomeric DL-Ala side chains were introduced onto the Eamino groups of poly(Lys) with the aid of NcarboxY-DL-Ala anhydride and resulted in poly[Lys(DL-Ala m)] (AK). Poly[Lys-(D- Leu ;-DL-

203

Ala m )] (o-lAK) where i < 1 and m - 3 was prepared by reacting benzyloxycarbonyl protected 0leucine pentachlorophenyl ester with the a-amino groups of AK. Blocking groups were removed with HBr in glacial acetic acid, as confirmed by UV absorbance at 254 nm. The primary structure of polypeptides was characterised by amino acid analysis (to give the m and i values) and by the identification of leucine at the side chain terminal (Hudecz and Sz6kan, 1985). The enantiomer composition of the side chain proved that there had been no stereoselective polymerisation of N-carboxy oL-Ala anhydride and verified the presence of > 98% of o-Leu in D-LAK (Sz6kan et aI., 1988). The size of these compounds was defined by the average molar masses (Mn , M.£ ~), the relative molar mass distribution (Mz/Mw) and the average degree of polymerisation (DPn) of the poly(Lys) backbone (Hudecz et aI., 1984). The average molar mass of polypeptides was calculated from I>Pn ( = 450) and of the amino acid composition of the side chains (m = 3.1 and i = 0.98) and found to be Mw (AK):z 156,700 and Mw (o-LAK) = 206,600 (Clegg et aI., 1990). The peptide CAP D T R PAP 0 was prepared by solid phase synthesis using p-hydroxymethyl-phenoxymethyl polystyrene resin as the support. All amino acids were coupled as Fmoc derivatives. The following protecting groups were applied for the reactive side chains of individual amino acids: tertiary butyl groups for Asp and Thr, 2,2,5,7,8-pentamethylchroman-6-sulphonyl groups for Arg and trityl groups for Cys residues. Coupling was carried out by HOBt-DCC methodology in N-methylpyrrolidone (NMP). After completion of the synthesis, the peptide was deprotected by 20% piperidine in NMP and cleaved from the resin by TFA containing 5% of a mixture of ethanedithiol-anisole-ethyImethyl sulphide (1: 3 : 1, v/v/v). The synthesis was performed using an ABI Automatic Peptide Synthesiser (Model 431A). The peptide was purified by gel filtration and by reversed phase HPLC with an Aquapore RP-300 C 18 cartridge (3 em X 4.6 mm column packed with spherical 7 J.l.m silica (30 nm pore size) (Brownlee Labs, Santa Carla, USA}) using gradient elution, where buffer A = 0.1 % TFA in water, while buffer B = 0.075% TFA

acetonitrile-water (70: 30, v/v). The composition of CO was verified by amino acid analysis. Synthesis of CG-branched polypeptide conjugates The peptide, CAP D T R PAP 0 was

. ' the conjugated to branched polypeptides through N terminal cysteine using SPDP (obtained from Sigma Chemical Co., Poole, UK) as the coupling reagent (Carlsson et aI., 1978). . 10 mg (21:8-28.7 J.l.moI) of branched polypeptides were dissolved in 3 ml of 0.05 M carbonate buffer, pH 9.0. 1.7-3.1 mg (5.5-9.9 J.l.moI) SPDP in 170-310 J.l.1 dioxane (Le., at a concentration of 10 mg/mI) were added dropwise to the solutions of the polypeptides. The reaction mixtures were stirred for 30 min at room temperature and then applied to a Sephadex 0-25 M column (15 x 1.6 em) equilibrated with PBS. The appropriate macromolecular peak as defined by UV detection at 252 nm was collected, dialysed against distilled water and freeze-dried. To determine the extent of 2-pyridyl-disulphide group incorporation, 400 J.l.1 of PSS-polypeptide solution obtained after gel filtration was reacted with DTT (from Sigma Chemical Co., Poole, UK) at pH 4.5, and the absorbance of the released pyridine-2-thione was measured. 10 mg of PSS-polypeptide, containing 5.4-7.3 J.l.mol of the 2-pyridyl-disulphide group was dissolved in 3 ml distilled water and mixed with 7.9-10.8 mg (8.1-11 J.l.moI) of CO peptide in PBS, pH 8.0 (i.e., at a concentration of 10 mg/rnl). After stirring for 30 min, the reaction mixture was dialysed against distilled water for 48 hand freeze-dried. The absence of 2-pyridyl-disulphide gro~ps and pyridine-2-thione in CO-polypeptide conjugate samples was verified by recording the UV spectra of purified product dissolved in PBS. The average degree of substitution (D'S) was estimated by amino acid analysis of the conjugates. Amino acid analysis

. Amino acid analysis of CO, branched polypeptides and of conjugates was performed using a B~ckman 6300. Automatic Amino Acid Analyser pnor to analYSIS, samples were hydrolyzed in 6N HCI in sealed and evacuated tubes at 110"C for 24 h.

204

CD spectroscopy The conformation of conjugates in solution was studied by CD spectroscopy. CD spectra were recorded using a Roussel-Jouan model III dichrograph (Jobin-Yvon, France) in quartz cells of optical path 0.02 em at room temperature under constant nitrogen flush. The dichrograph was calibrated with D-( - )-pantoyllactone at 220 nm (Schippers and Dekkers, 1981). The samples were dissolved in PBS at pH 7.3. The concentration of solutions was approximately 0.5 mg/mi. CD band intensities were expressed as .1e values, representing the difference of molar absorptivity of the left and right circularly polarised components . .1e values were related to one lysine residue in the main chain including a whole side chain. Interpretation of CD spectra was based on the secondary structure analysis of the well characterized poly(Lys) (Hudecz et aI., 1985; Votavova et al., 1985; Hudecz et aI., 1988). Monoclonal antibodies NCRC-ll (IgM) was prepared using the spleen cells from a BALB/c mouse immunised with dissociated breast carcinoma cells (Ellis et aI., 1984) and C595 was prepared against purified urinary epithelial mucin (Price et aI., 1990b). The following anti-human milk fat globule membrane antibodies were employed: HMFG-1 (IgG 1) (Taylor-Papadimitriou et aI., 1981) and SM-3 (IgG1), prepared against deglycosylated milk mucin (Burchell et aI., 1987). IgG monoclonal antibodies were purified from tissue culture supernatants by their binding to, and elution from Sepharose-Protein A, and with C595 and NCRC-11 antibodies, purification was also performed using epitope affinity chromatography (Price et aI., 1991). Protein concentrations of purified antibodies were determined spectrophotometrically assuming E 280nm = 14.3 for IgG antibodies and E280nm = 11.9 for IgM antibodies. Mucin antigen preparation Urinary mucin was isolated from normal human urine by immunoadsorbent chromatography using Sepharose-linked NCRC-11 antibodies, as previously described (Price et at., 1990b).

Radioisotopic antiglobulin assay Purified human urinary mucin or peptide conjugates, in phosphate-buffered saline (PBS, pH 7.3) at 10 ILg/ml were adsorbed to the wells of Terasaki microtest plates (Nunc, Roskilde, Denmark) by incubation at 37°C for 18 h. After antigen adsorption, the wells were washed four times with PBS containing 1% casein (washing buffer). Monoclonal antibodies were dispensed at 10 ILl/well and incubation was continued for 60 min at room temperature. The plates were again washed with washing buffer. 125I-labelled F(ab')2 fragments of affinity purified rabbit anti-mouse immunoglobulins were added at lOS cpm/lO ILl/well (radioiodination of this reagent was performed using the chloramine T procedure of Jensenius and Williams (1974) using 18 MBq 12SI per 25 ILg protein). Incubation was continued for 1-2 h at room temperature. The contents of the wells were then aspirated and the plates washed six times, after which the radioactivity in each well was determined. The non-specific binding of radiolabelled antiglobulin to wells treated with washing buffer alone was subtracted from the values obtained with test samples. Double determinant immunometric assay Synthetic conjugates and urinary mucin were examined for their capacity to link biotin-labelled C595 antibody, added as tracer, to C595 antibody immobilised by adsorption to the wells of Falcon Microtest III Plates (Becton Dickinson, Oxnard, CA, USA). Antibody was adsorbed overnight at 50 ILl/well from a 10 ILg/ml solution in PBS. The test was performed using conventional 'sandwich assay' ELISA techniques with measurement of colour development at 405 nm following addition of streptavidin peroxidase and ABTS (2,2'-azino-di-(3-ethyl-benzthiazolinesui phonic acid) diammonium salt). Results

Preparation and chemical characterisation of CG. branched polypeptide conjugates The coupling of CO to branched polypeptides was achieved by the insertion of a disulphide bridge between the peptide carrying the epitope

lOS

and the carrier molecule. This approach provided covalent linkage in which the thiol group of the N terminal Cys of the peptide was attached to the SH group situated at the side chain end of the carrier polymer. First, protected thiol groups were introduced to the a-amino groups of the side chains of the macromolecule by reacting AK or D-LAK with SPDP in alkaline solution (pH 9.0) (Carlsson et ai., 1978). In the second step, CO with a free SH group was added to the PSS-polypeptides and the coupling reaction proceeded as indicated by the release of pyridine-2-thione. It should be noted that no precipitate was formed during the synthesis of conjugates under these conditions. The conjugates were then purified and characterised by UV spectroscopy and amino acid analysis. UV spectra of conjugates indicated no absorbance at 280 nm suggesting that all protected thiol groups were substituted during the second step of preparation. The data concerning epitope content and average relative molecular mass (Mw value) are summarized in Table I. The average molar substitution ratio was also expressed as a percentage of modified side chains of the carrier. These results showed no significant difference in the coupling efficiency of peptide epitope to either of the branched chain polymers studied (22.2% for AK-CO and 23.8% for DLAK.-CG). Conformation of branched polypeptide-CO conjugates

The conformation of the conjugates in water solution was studied by CD spectroscopy in the wavelength region 190-250 nm. The CD culVes

01-+------,-.......::===4

·5

·10

200

220

240

260 [nml

Fig. 2. CD spectra of branched polypeptide, CAP D T R P A P G conjugates at pH 7.3 in PBS: (a) AK-CG and (b) o-LAKCG.

of AK-CG and D-LAK-CG conjugates in PBS at pH 7.3 are shown in Fig. 2. In this wavelength region, the circular dichroic spectra indicate the

TABLE I CHARACTERISTICS OF CAP D T R PAP G-POLYPEPTIDE CONJUGATES

Conjulate

Code·

PoIy{Lys..(CGj OL-A1a 3.1)] Poly{Lys..(CGjo-Leuo.CJ8-0L-A1a3.1)]

AK-CG o-LAK-CG

Molar ratio of polypeptide: CO

USb

1: 99.9 1: 107.1

22.2 23.8

(%)

Mw e

±5%

253300 310100

• Based on one-letter symbols of amino acids and abbreviation of CAP D T R PAP G (CG). Capital letters denote the size of the carrier polypeptide where the average degree of polymerisation is 450. b Average degree of substitution, expressed as percentage of side chains modified with CG. e Calculated from the average degree of polymerisation of poIy(L-Lys) and of the side chain composition as described in the materials and methods section.

206

presence of a helical secondary structure for DLAK (with a positive maximum at 200 nm and negative maximum at 208 nm and shoulder at 221 nm). In contrast, the CD curve of the AK-CG conjugate correlates with that of unordered conformation (negative maximum at 199 nm). The CD signals in this wavelength range originate from the optical activity of amide bonds present in the carrier as well as in CG. In order to consider this, the ellipticity values (Je) are related to one lysine residue in the main chain including a whole CG substituted side chain. It should be noted, that the CO spectra of CG under identical conditions display unordered spatial arrangement (data not shown). At pH 7.3 in PBS the CD spectra of the unconjugated polypeptide D-LAK indicates a partially ordered helical conformation (Votavova et aI., 1985) while AK (Hudecz et aI., 1985) proved to be completely unordered.

Antibody binding to CG-branched polypeptide conjugates The antibody binding properties of two branched polypeptide conjugates containing the P o T R PAP epitope motif, together with a preparation of urinary mucin, were investigated using a solid phase radioimmunoassay with 125 1_ labelled affinity purified F{ab')2 fragments of

rabbit anti-mouse Ig as a developing agent. Three IgG monoclonal antibodies recognise distinct but partially overlapping determinants in the protein core of epithelial mucins. The minimum structure for binding for the HMFG-l antibody is PDT R. whereas C595 antibody recognises peptides containing the R PAP sequence. The antibody SM-3 defines an epitope of five amino acid residues (P 0 T R P) located within the same immunodominant domain of the protein core (Burchell et aI., 1989; Price et aI., 1990a). For comparison, the IgM antibody NCRC-ll (epitope: R P A) was also included. The results of binding studies are summarized in Table II. As shown, all four antibodies bound to branched polypeptide-CG conjugates, but marked differences could be observed. It should be noted, that the reaction of a preparation of urinary mucin with the antibodies is as reported earlier (Price et ai., 1990a). As shown in Table II, all anti-mucin antibodies reacted more strongly with the D-LAK-CG conjugate than the AK-CG conjugate (both adsorbed to the microtest plates from 10 "g/ml solutions). In this assay, the relative order of antigen reactivity with the antibodies HMFG1, C595 and NCRC-ll was D-LAK-CG > mucin> AK-CG while the order of SM-3 reactivity was o-LAK.CG > AK.-CG > mucin. The lower level of reac-

TABLE II BINDING OF ANTI·EPITHEUAL MUCIN ANTIBODIES TO PEPTIDE ANTIGENS CONJUGATED TO BRANOIED CHAIN POLYMERS Antibody binding (mean cpm) to:

Antibody Name (class)

Epitope

HMFG-l

PDTR

(lgG)

SM·3

PDTRP

(lgG)

C59S

RPAP

(lgG)

NCRC·ll (lgM)

Cone.

o-LAK-CG

AK-CG

(p.g/ml)

RPA

10 1 0.1 10 1 0.1 10 1 0.1 10 1 0.1

Urinary mucin

7029 2211 219 5094 3060 955 8851 3874 297 3872 3230 1083

2408 301 53 2137 350

44 3322 465 27 845

99

1

4450 1786 58S 926 159 34 3SS6 3268 1530 2858 3007 1231

TABLE III BINDING OF MURINE MONOCLONAL ANTIBODIES TO SYNTHETIC PEPTIDE EPITOPES CONJUGATED TO BRANCHED CHAIN POLYMERS Antibody binding (mean cpm ± SO) to:

Antibody Name (class)

o-LAK-CG

Antiaen

14 ±23 6781 ±S09 28 ±14 S ±7 38

PlSNO

CS9S (lgG)

C337

Urinary mucin CEA

(laG)

C161

NCA

(lgG)

791T/36

1J)72

±to

(laG)

tivity of SM-3 with mucin in these tests most probably reflects its requirement for a pentameric epitope to be accessible while the other antibodies define smaller determinants of three or four residues. In control experiments several IgG monoclonal antibodies against breast, ovarian and colorectal carcinomas, but recognizing irrelevant antigens, failed to react with both branched polypeptideCG conjugates as well as mucin antigen. In addition, the positive control anti-urinary mucin C595

TABLE IV IMMUNOMETRIC ASSAY USING SYNnIETIC PEP· TIDES CONJUGATED TO BRANCHED CHAIN POLY· MERS AntiJen cone. (Jol./mJ) 0 0.1 1.0 10

Absorbance (405 am)