Immunogenicity of Synthetic Peptides Corresponding to Flexible and ...

Vol. 264, No. 18, Issue of June 25, pp. 10513-10519,1389 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Immunogenicity of Synthetic Peptides Corresponding toFlexible and Antibody-accessible Segments of Mouse Lactate Dehydrogenase (LDH)-C4* (Received for publication, January 9, 1989)

Holly H. HogrefeS, Pravin T.P. Kaumaya, andErwin Goldberg8 From the Department of Biochemistry, Molecular Biology & Cell Biology, Northruestern Uniuersity, Euanston, Illinois 60208

The immunological properties of apanel of synthetic natural isozyme with a synthetic antigen would ensure availpeptides that represent the most accessible and mobile ability and homogeneity of vaccine preparations, as well as segments of the lactate dehydrogenase (LDH)-C4 mol- guarantee antigenic specificity for the testicular LDH isozyme ecule were characterized. Peptides corresponding to (Goldberg et al., 1983). mouseLDH-C4 amino acid sequences: 1-14b, 5-15, Peptides, coupled to carrier molecules, have been shown to 49-58,97-110, 211-220, 231-243, 274-286,elicit304the production of antibodies which will bind to the 316,and 318-330 were synthesized and compared in protein from which they were derived (Sutcliffe et al., 1983). terms of binding antibodies raised in rabbits against This observation has raised considerable interest in developthe intact protein. Six of these sequences were cova- ing synthetic peptide vaccines (Lerner, 1984; Arnon, 1986). lently coupled to diphtheria toxoid and used to immunize groups of rabbits. LDH-Co-specific antibodies For example, immunization with peptides encompassing sewere detectable in immune sera by enzyme-linked im- quences of human chorionic gonadotropin (Stevens, 19861, munosorbent assay, Western blotting, and immunopre- influenza virus hemagglutinin (Shapira et al., 1984), foot-andcipitation assays. The immunogenicity of the mouse mouth disease virus VP1 (Bittle et aZ., 1982; DiMarchi et al., LDH-C4 peptides in rabbits couldbe ranked in the 1986), hepatitis B surface antigen (Gerin et al., 1983; Itoh et following order: 5-15,304-316 z 211-220,274-286 al., 1986), herpes simplex glycoprotein D (Eisenberg et al., > 49-58, 97-110. The immunological properties of 1985), and the malaria circumsporozoite protein (Egan et al., these short synthetic peptides did not correlate with 1987) has been shown to protect experimental animals against features of the mouse LDH-C4structure except that the subsequent challenge with the intact antigen. most active sequences appeared to be those that difThe potential development of synthetic vaccines has fueled fered from the somatic isozymes to the greatest extent. renewed interest in defining and predicting the antigenic sites These results have direct bearing on the selection of of proteins, particularly those which can be represented by immunogenic LDH-C4peptides for contraceptive vac- linearpeptides (Berzofsky, 1985;Van Regenmortel, 1987). cine studies in humans and non-human model systems. Immunological self-tolerance dictates that antigenic sites are located in regions of the molecular surface that differ structurally from the host’s self proteins (Benjamin et al., 1984). In an attempt todevelop an immunocontraceptive vaccine Crystallographic analyses of antigen-antibody complexes re(Goldberg et al., 1983), the antigenic structure of the testes- veal that an antigenic site may consist of as many .as 17-20 specific isozyme of lactate dehydrogenase, LDH‘-C4, has been amino acids and can encompass as much as 750 A’ of the studied (Wheatand Goldberg, 1985). This isozyme is ex- protein surface (Mariuzza et al., 1987; Sheriff et al., 1987; pressed only by spermatogenic cells (Goldberg, 1977) and has Colman et al., 1987; Amit et at., 1986). A number of predictive approaches have been employed to been found on the surface of mature spermatozoa (Erickson identify potential antigenic determinants. Algorithms aimed et al., 1975). The numerous amino acid sequence differences between the testicular and somatic LDH isozymes (-70% at identifying continuous antigenic sites are based upon athomology) (Eventoff et al., 1977; Li et al., 1983; Millan et al., tempts to correlate peptide antigenicity with various struc1987) render them antigenically distinct (Goldberg, 1971; tural features of protein antigens. For proteins whose strucLiang et al., 1986). Immunization of female animals with tures are known at the amino acid level, some success in LDH-G purified from mouse testes results in the production identifying surface epitopes has been achieved from prediction of highly specific antisera and in the suppression of fertility based on hydrophilicity (Hopp and Woods, 1981; Kyte and (Goldberg et al., 1981; Goldberg, 1973). Replacement of the Doolittle, 1982) or the propensity to form a /3-turn (Gamier, et al., 1978). * This work was supported by National Institutes of Health grants In proteins of known three-dimensional structure, the ac(to E. G . ) and a National Institutes of Health training grant predoc- cessibility of peptides to antibody-sized molecules has been toral fellowship (to H. H,). The costs of publication of this article were defrayed in part by the payment of page charges. This article shown to correlate with antigenicity. In these studifs, expomust therefore be hereby marked “aduertisement” in accordance with sure was determined with a spherical probe of 10-A radius (Novotny et al., 1986) or through the use of protrusion indices 18 U.S.C. Section 1734 solely to indicate this fact. t Present address: Dept. of Biochemistry, University of Iowa, Iowa (Thornton et al., 1986). City, IA 52241. In addition to surface accessibility, a correlation has been ITo whom correspondence should be addressed. observed between antigenicity and atomic mobility deter’ The abbreviations used are: LDH, lactate dehydrogenase; ELISA, enzyme-linked immunosorbent assay; Boc, t-butoxycarbonyl; FMoc, mined from the temperature factors of highly refined crystal fluoren-9-ylmethoxycarbonyl; HPLC, high performance liquid chro- structures (Westhof et al., 1984; Tainer et al., 1984). An matography. extensive study employing myohemerythrin (Geysen et ai., ~

~~~~~~

~~~~~~~~

~

~

10513

Immunogenicity of LDH-G Sequences

10514

1987), a molecule which is completely foreignto the vertebrate immune system, showed that those epitopes which are best mimicked by shortpeptides (6-10 residues) correspondto segments of the polypeptide chain which exhibit high flexibility, a convex surface shape, and negative electrostatic potential. The parameters that dictate which peptides will be the most likely t o elicit protein-reactive antibodies (immunogenic) are less clear. In this publication,we describe the immunological properties of a panel of synthetic peptidesderived from mouse LDH-C4. These sequences were selected based upon the location of known antigenic sites and regions of high surface accessibility, segmental flexibility, and evolutionary variability. The relative immunogenicity of the peptides (asassessed by enzyme-linked immunosorbent assay (ELISA), immunoprecipitation, an9 Western blot) were compared with features of the refined 3-A structure of mouse LDH-C, (Hogrefe et al., 1987) to provide guidelines for selecting immunogenic LDHCq peptides for vaccine studies in humans and non-human model systems. MATERIALSANDMETHODS

All Boc-protected L-amino acids were purchased from Vega Biotechnologies, Inc. (Tucson, AZ) or Bachem (Torrance, CA). FMocamino acids werefrom Cambridge Research Biochemicals (Cambridge, Great Britain). p-Alkoxybenzyl alcohol polystyrene resins were purchased from Vega Biotechnologies, substituted with the appropriate Boc- or FMoc-amino acid (0.4-0.5 meq/g), respectively. Solvents and chemicals were obtained from the following sources: Sequanal-grade trifluoroacetic acid and diisopropylethylamine (Chemical Dynamics Corporation); gold label indole, l-hydroxybenzotriazole, 1,3-dicyclohexylcarbodiimide (99%), pentafluorophenol (99+ %), and piperidine (98%) (Aldrich); HPLC/Spectro-grade dichloromethane, and dimethylformamide (Pierce Shemica1 Co.). Anhydrous dimethylformamide was stored over 4-A molecular sieves, bubbled daily with Nz, and tested prior to use with dinitrofluorobenzene for the presence of secondary amines. Peptide Synthesis-Seven peptides of mouse LDH-C4were synthesized by solid-phase methodologies (listed in Fig. 1).Peptides 1, 3, 6, and 8 were assembled by the Boc-benzyl procedure (Barany and Merrifield, 1979; Stewart and Young, 1984).Peptides 2,7, and9 were synthesized using the FMoc-t-butyl strategy (Chang and Meienhofer, 1978; Athertonet ai., 1983). All couplings were carried out with preformed Boc- or preformed FMoc-amino acids, which were activated as their symmetrical anhydrides, hydroxybenzotriazole esters, and/or pentafluorophenyl esters and added manually to thereaction vessel. Double couplings were routinely carried out using 5-6 equivalents of activated amino acid for each cycle. Completeness of the coupling reaction was monitored with the Kaiser ninhydrin test (Kaiser et al., 1970). The peptide resin was washed and filtered using either a Vega coupler model 1000 peptide synthesizer or a custommade solvent delivery system. A cysteine residue was added to the amino terminus of each peptide to facilitate conjugation to a macromolecular carrier. In most cases, the S-ethylmercapto group, which remains intact during acid cleavages of peptides from the resin, was used for protection of the sulfhydryl moiety. Purification of crude cysteine-containing synthetic peptides are thus simplified as no dimerization and/or polymerization products contaminate the preparation. Before conjugation, the purified peptides are reduced in the presence of excess dithiothreitol, gel-filtered on Sephadex G-15 in 0.1 M acetic acid and lyophilized. Boc-benzyl Synthesis-At each synthesis cycle, the a-amino group was deprotected with 50% trifluoroacetic acid/dichloromethane, containing 10% indole (Stewart and Young, 1984), and neutralized with 5% diisopropylethylamine/dichloromethane. Boc-amino acids were coupled as activated symmetric anhydrides (Hagenmaier and Frank, 1972) and 1-hydroxybenzotriazole esters (Konig and Geiger, 1970) for the first and second couplings, respectively. Final cleavage of the peptide from the resin and side chain deprotection was carried out with HBr/trifluoroacetic acid, containing 10% anisole (Stewart and Young, 1984). FMoc t-Butyl-At each synthesis cycle, deprotection of the baselabile FMoc moiety was carried out with 20% piperidine/dimethylformamide. Activated FMoc-amino acids were coupled as symmetric

anhydrides (Chang and Meienhofer, 1978) and pentafluorophenol esters (Kisfaludy and Schon, 1983) for the first and second couplings, respectively. FMoc-glycine and FMoc-lysine were coupled as pentafluorophenol esters exclusively. Allcouplings were carried out for 1 h in dimethylformamide to minimize aggregation of the growing peptide chains(Pillai and Mutter, 1981). Final cleavage and side chain deprotection were carried out with 50% trifluoroacetic acid/dichloromethane, containing 10% anisole (Atherton et al., 1983). Purification and Characterization of Synthetic Peptides-The crude cleavage produce, after rotaryevaporation, was precipitated with cold anhydrous ethyl ether.The peptide mixture was extracted into acetic acid:water (3:1), washed several times with ether to remove anisole, and lyophilized. The desired peptide was purified from modified and deletion peptides by reverse-phase HPLC. Crude peptides were first chromatographed on an analytical reverse-phase C, column (Waters WBondapak) in the presence of 0.1% trifluoroacetic acid (Pierce Chemical Co.) and eluted with a gradient of 0-50% acetonitrile. Purification ona larger scale was accomplished using a semipreparative CIS column (Waters gBondapak 19 X 150 mm) with typical loadings of50-100mgof crude peptide. HPLC runs were carried out at a flow rate of 10-20 ml/min based upon the optimal separation conditions of the analytical runs. Fractions were collected, pooled, and lyophilized. With one exception, fractions used for immunochemical analysis were of greater than 95+ % purity, as judged by analytical reversephase HPLC and amino acid analysis. Chromatographic analysis of peptide 6 showed extensive dimerization and polymerization, and the desired sequence probably accounted for less than 50% of the total product. HPLC-purified peptide samples (50-100pmol)werehydrolyzed (110 "C, 24 h in uacuo) and derivatized with phenyl isothiocyanate (Pierce Chemical Co.). Phenyl isothiocyanate-amino acid derivatives were separated using the Waters PicoTag system (Cohen et al., 1984) and quantified relative to amino acid standards (Pierce Chemical Co.). The amino acid compositions of the HPLC-purified peptides were consistent with expected theoretical values. Preparation of Antigens-In addition to the seven sequences described above, synthetic peptides 4 and 5 (Fig. l) were purchased from Peninsula Laboratories (San Carlos, CA) and judged to be homogeneous, as described previously (Wheat etaZ., 1985).A cysteine residue was subsequently added to theNHZ terminus of peptide 4 for conjugation purposes (Wheat et al., 1985). An unrelated peptide corresponding to residues 92-103 of tobacco hornworm moth cytochrome c (peptide 10, Fig. 1) served as a control sequence. Peptides 2, 3, 4, 5, 7, and 8 were conjugated to a macromolecular carrier, using a previously described method (Wheat et al., 1985). Briefly, the reduced purified peptides werecoupled to diphtheria toxoid (Connaught Laboratories, Philadelphia, PA) through their amino-terminal sulfhydryl groups with the heterobifunctional reagent 6-maleimidocaproylN-hydroxysuccinimide (Sigma) (Lee et ai., 1980). The degree of substitution was measured using Ellman's reagent, as described (Gonzales-Prevatt etal., 1982) as well as amino acid analysis. Coupling efficiencies ranged from 12 to 20 peptides/carrier molecule. LDH-C4was purified to crystalline homogenity from random-bred mouse testes by affinity chromatography on 5'-AMP-Sepharose (Bachman and Lee, 1976). Preparation of Antisera-Six groups of female rabbits (Langshaw Farms, Augusta, MI), consisting of between one to three animals each, received primary immunizations at multiple intradermal sites of 2 mg of peptide-diphtheria toxin conjugate emulsified in complete Freund's adjuvant. Secondary immunizations consisted of 1 mgof conjugate in incomplete Freund's adjuvant and were administered in the same manner 4 weeks after the primaries. Sera were collected at regular intervals, heated at 56 "C for 30 min to destroy complement, and aliquots stored at -20 "C until assayed. Two +21 day sera were pooled from rabbits receiving identical formulations and tested for: ( a ) reactivity with the immunizing peptide and mouse LDH-Caby an ELISA, (b) reactivity with mouse LDH-C4 by immunoprecipitation of'251-labeledLDH-C,, and ( c ) specificity for the mouse LDH-C, protein by Western blotting of mouse tissue extracts. Rabbit antisera to affinity purified mouse LDH-C4 were prepared as described (Liang et al., 1986). The ELISA procedure was adapted from Liang et al. (1986). Results were expressed as the antiserum dilution whichgave anarbitrary ELISA reading on a Dynatech ELISA reader. A solution-phase radioimmunoassay was employed with polyeth-

-

Immunogenicity of LDH-C4 Sequences

-

10515

Eplircpn ylene glycol( M , 6000-7500) as the precipitating agentand ‘251-labeled a b mouse LDH-C4as tracer (Liang et al.,1986; Wheat et al., 1985).All c(SEt)-S-T-V-K-E-Q-L-I-Q-N-L-V-P-E-D 1 C(SEt)-l-l4b samples were assayed in triplicate. Peptide-specific antibody binding a b to LDH-C4was expressed as the total counts/min precipitatedby the C(SEt)-E-Q-L-I-Q-A-L-V-P-E-D-K 2 C(SEt)-5-15 anti-peptide antiserumminus the counts/minprecipitated in the 3 C(SEe)-C-49-58 C(SEt)-C-V-D-A-D-T-D-K-L-8-C presence of an excess amount of the homologous peptide. Results were also calculated as the mean micrograms of ‘251-labeled LDH-Cd 4 97-110 R-H-V-S-C-Q-T-B-L-D-L-L-9-R bound per ml of serum. 5 c-211-220 C-S-L-N-P-A-I-C-T-D-Y Western Blot Analysis-The LDH enzyme activity of crude mouse C-9-V-V-E-C-C-Y-E-V-L-D-H-K 6 C-231-243 testes and liver homogenates was measured spectrophotometrically as described (Hawtrey and Goldberg, 1970). Extracts (0.5 units of C(SXt)-C-C-C-L-V-K-G-F-H-C-I-K-E-E~V-F-C 7 C(SEt)-C3-276-286-C enzyme activity/Iane) were electrophoresed under native conditions C(SEt)-V-N-H-T-A-E-E-E-C-L-L-K-K 8 C(SEt)-3D4-316 using 7.5% polyacrylamide gel electrophoresis gels. Protein blotting wasdoneaccording to the method of Burnette (1981) except that C(SEt)-A-D-T-L-U-N-H-Q~~-N-L-E-L 9 C(SEt)-318-330 electrophoretictransfer to nitrocellulose (Schleicher and Schuell) was C(SEC)-A-D-L-I-A-Y-L~K~Q~A~~-K C(SEt)-cc92-103 10 performed with cooling in native running buffer (5 mM Tris, 40 mM glycine, pH 8.3). FIG. 1. Amino acid sequences of the assembled mouse LDHOne nitrocellulose strip was stained for LDH activity (Goldberg, C4 peptides. In the text, the synthetic peptides are referred to as 1964). Total protein was detected by Amido Black staining. numbers 1 to 9, as indicated.Controlpeptide10corresponds to The remaining strips were blocked with5%bovine serum albumin, sequence92-103oftobaccohornwormmothcytochrome c (cc). rinsed with Tris-buffered saline, incubated with peroxidase conjuC(SEt) refers to an S-ethylmercapto-protected NH,-terminalcysgated to goat anti-rabbit IgG (Boehringer Mannheim), and then in a teineresidues. The positions of glycine spacers (C) andcysteine substrate solution prepared with 4-chloro-1-naphthol and hydrogen linkers (C), not present in the natural mouse LDH-C4 sequence, are peroxide as described (Monroe, 1985). indicated.

RESULTS

Thethree-dimensionalstructure of mouse LDH-C4has + been determined (Musick and Rossmann, 1979; Hogrefe et .................. al., 1987). Temperature factor and surface accessibility pro+ files for the refined 3-A structure of mouse LDH-C4have been + I published (Hogrefe et al., 1987). Previous attempts to correlate 4 / ........................................................................................ features of the structure with the location of antigenic sites + were based upon the bindingof anti-mouse LDH-C4 antibod- e ies totrypticpeptides by solid-phaseradioimmunoassay 2 ....... . . . . . . ...................................................... ..*.......................... (Wheat et al., 1985). To obtain a more complete map of the antigenic determinants onLDH-C4, a panel of synthetic peptides was selected for synthesis based ona number of criteria 0 which have been shown t o correlate with antigenicity. 1 2 3 4 6 8 7 8 8 1 0 Structural analysisof mouse LDH-C4 showed that, ingeneral, amino acids which are accessible t o solvent inth? native LDX-C‘ PEPTIDES tetramer (determined with a spherical probe of 1.4-A radius FIG. 2. Antigenicity of mouse LDH-C4 synthetic peptides. tosimulate a water molecule; (LeeandRichards, 1971)) exhibit a high degree of flexibility, whereas those residues The amino acid sequences of the peptides are given in Fig. 1. The binding of rabbit anti-mouse LDH-C4 antibodies was measured by located in the protein interior or in the subunit interfaces ELISA. The results are expressed as the antiserum dilution which have lower temperature factors. For LDH-C4,nine major gave an absorbance of 0.2 units abovebackground (corrected by peaks of B value maxima are centeredat positions 8-20, 51subtracting outthe absorbance of preimmune sera at the same dilution). The three different serum pools tested exhibited the following 55, 79-82, 96-99, 115-124, 150-151, 207-228, 282, and 315rabbits); +, 330 (Hogrefe et al., 1987). In addition, regions of the tetramer LDH-C4-specific titers: 0, >> 21,870(poolfromtwo 240,000 (pool from two rabbits); *, 180,000 (pool from four rabbits). which are accessible to antibody-sized yolecules have been identified by calculation relative to a 10-A radius probe (Novotny et d . , 1986) and correspond to the following sequences: andStegnink, 1970) shows thatrabbitLDH-A4 is indeed 1-16, 51-57, 79-82, 97-117, 124, 150, 214-241, 276-284, and homologous to the known vertebrate LDH-AI molecules. 308-330 (Hogrefe et al., 1987). The relative antigenicity of the synthetic peptides (capacity Nine peptides encompassing the mostflexible and accessi- to bindantibodies raised againsttheintactprotein) was ble segments of the LDH-C4 tetramer, and hence the most determined by ELISA (Fig. 2). The binding by rabbit antilikely to be immunogenic, were synthesized (listed in Fig. 1). mouse LDH-C4 antibodies to the synthetic peptides can be Each of these peptides encompasses between 10 and 15 amino evaluated both in terms of the antibody titer and the number acids of mouse LDH-C4 and, in total, represent>50% of the of antisera which show cross-reactivity (Geysenet al., 1987). solvent accessible surface of the tetramer. In addition, these By using these two criteria, the synthetic peptides can be synthetic peptidescomprise sequences, whichspan the amino grouped according to their ability to bind rabbit anti-mouse and carboxyl termini, exhibita variety of secondary structural LDH-C4antibodies: 1)“highly antigenic” (react with 3/3 sera attributes and cover virtually all the prominent maxima of pools; mean antibody titer in the range of 1425-3300), pepthe hydrophilicity profile (Hopp and Woods, 1981; Kyte and tides 1, 2,6, and 8; 2) “moderately antigenic” (react with 1Doolittle, 1982). Sequencevariability in the synthetic peptides 3/3 serapools; mean antibody titerin the rangeof 640-1670), ranged from 40 to 60% and was calculated as the average peptides 3, 5, 7, and 9; 3) “nonantigenic” (no reactivity obnumber of variations in each peptide pairwise in comparisons served), peptides4 and 10 (controlsequence). with somatic LDH sequences from human, murine, and porto the The binding of rabbit anti-mouse LDH-C4 antibodies cine sources. Although the complete amino acid sequences of syntheticpeptides wasalso evaluatedin a solution-phase rabbit (the host’s) somatic LDHs are not known, the infora competitionassayusing lZ5I-labeled mouseLDH-C4as mation which is available (Taylor and Oxley, 1976; Brummel tracer. This assayallows an estimateof the relative affinities

f

q;y;*$L;;;*l *

*

Immunogenicity of LDH-C4 Sequences

10516

of antibodies for the peptide versus the native protein. Nonerabbits immunized with peptides 3 and 4. An identical hierof the synthetic peptides (at concentrations ranging from archy of reactivity with LDH-C4 was also observed with anti10"' to lo-, M ) were able to displace the binding of rabbit peptide antiseracollected after tertiary immunizations. anti-mouse LDH-C, antibodies from Iz5I-LDH-C4.No comAntibody titers were also determined by immunoprecipitapetition wasobservedeven whentheninepeptides were tion of '2sI-labeled LDH-C4 (Table I). Antisera from rabbits pooled together,usingconcentrations which were 104-fold immunized with peptides 2,5,7, and 8 specifically precipitated higher than the amount of unlabeled LDH-C4 required to '2sI-labeled LDH-C,, whereas antibodies to peptides 3 and 4 completely saturate all available antibody combining sites. did not. Antibody binding to LDH-C4could be absorbed with Taken together these results indicate that antibodies raised the homologous peptide, in a concentration-dependent mantothenative LDH-C, bindtopeptides whenassayed by ner (only the data with excess amount of competing peptide ELISA but have no or low affinity for the peptidesin solution. are shown in Table I), thereby demonstrating that the antiA reasonable conclusion is that these short peptides do not peptide antibodies can bind in solution to both the peptide faithfully mimic the conformation of the corresponding epi- immunogen and the corresponding epitope in the intact moltope in the native protein. Alternatively, the antibodies may ecule. The relative affinities of the different antisera for LDHbe detecting denatured antigenby solid-phase assay (as anti- C4 in the solution-phase assay were similar to thoseseen with gen attaches to the plate) or from the immunizing preparation the ELISA (Table I) and consistent with the hierarchy of (when antigen is emulsified in Freund's adjuvant). Indeed a immunogenicity established previously for peptides 2, 4, and recent study has questioned whether a protein antigen is in 5 (Wheat et al., 1985). its native state when assayed for antibody binding (JemmerAntibodies raised against LDH-C, and the peptide conjuson, 1987). The discrepancy between the ELISA and solutiongates bind to LDH-C, in solution with affinitieswhich differ phase assays emphasizes theneed to use a variety of immunochemical techniques when the antigenicity of peptides and by a t least four logs (Table I). A pool of the six different sitespecific sera could precipitate more of the protein than any proteins is measured (VanRegenmortel, 1987). of the sera individually, althoughthebinding was still a The relativeimmunogenicity of six mouse LDH-C4sequences was determined in rabbits (Table All I). of the rabbits fraction of that obtained with antibodies raised against the received diphtheria toxin-peptide conjugates that had been whole protein. The specificity of the anti-peptide antisera for mouse LDHprepared and administered in an identical fashion. ExaminaC4 was also evaluated with Western blotting techniques(Fig. tion of individual responses (Hogrefe, 1987; Wheat andGold3). Mouse testes and somatic tissue extracts were electrophoberg, 1984; Wheat et al., 1985) showed that although there was variability among animals receiving the same formula- resed under native conditions to separate the various LDH nitrocellulose. LDH activity tion,in all casesthemagnitude of the immune response isozymes andtransferredto appeared todiffer betweenanimals receiving different peptide staining of the strips showed that the testes extract lanes contained only the C4 isozyme, whereas the liver extract lanes conjugates. To compensate for variabilityintheimmune response and enablegeneral trends tobe examined, antiserum contained the A, isozyme (migrates close to mouse LDH-C4) was pooled from two or three rabbits (only one rabbit was and minor amountsof A/B heterotetramers. The stripswere immunized with peptide 3). Similar investigations aimed a t incubated with dilutions of anti-LDH-C4 and anti-peptide comparing the immune response to different peptides (Wes- antisera which bind mouse LDH-C4 similarly in the ELISA thof et al., 1984) have usedserum pools fromthe same number (Table I). Fig. 3 shows that each of the anti-peptide antisera of animals. binds tomouse LDH-C4. With the exception of antisera raised Antibody titers against the appropriate peptide immunogen to peptide 3, which has the lowest ELISA titer for mouse and against mouse LDH-C, were measured by ELISA (Table LDH-C, (Table I), binding to somatic LDHs or other unreI). All of the antisera contain antibodies which bind to the lated proteins was not observed a t these dilutions, thereby immunizing peptide and the intact LDH-C, molecule. Levels 1 2 3 4 5 6 7 8 9 of LDH-C4-specific antibody appear tobe highest in the sera 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 of rabbits immunized with peptides 2 and 8 and lowest in TABLE I Immunogenicity of selected mouse LDH-C4peptides in rabbits '"I-LDH-CI titeP

^l.?

Antiserum

LDH-C, Total pools Pools 2 8 5 7 4 3

'*"' pooled

Peptide LDH",,

immunoprecipitation . . Peptide'*'I-LDH-C, specificb bound

Cpll precrprtated

pglml serum

2 15