THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1985 by The American Society of Biological Chemists. Inc.

Vol. 260, No. 2, Issue of January 25, pp. llffi-l190,19ffi Printed in U. S. A.

Exposure of Antigenic Sites during Immunization MONOCLONAL ANTIBODIES TO MONOMER OF LACTOSE REPRESSOR PROTEIN*

(Received for publication, July 30, 1984)

Clarence F. Sams, Virginia B. Hemelt, Frederick D. Pinkerton, George J. Schroepfer, Jr., and Kathleen Shive Matthews From the Department of Biochemistry, Rice University, Houston, Teras 77251

Monoclonal antibodies specific for the lactose re- pressor protein have been purified from three mouse hybridoma cell lines, and ascitic fluids from five other cell lines producing repressor antibodies have been assayed for immunoglobulin subclass and antigenic specificity. The chymotryptic core region (amino acids 57-360) of the repressor reacted with all antibodies examined, while no reaction with the NHAerminal domain (1-56) could be detected. All of the purified antibodies and ascitic fluids reacted with the carboxyl- terminal fragment (amino acids 281-360) produced by cyanylation and base-catalyzed cleavage at the cys- teine residues. Although none of the purified antibodies associated with native, tetrameric lac repressor, reac- tion was observed with repressor which had been de- natured or dissociated into monomers by treatment with low levels of sodium dodecyl sulfate. Additionally, a mutant repressor which exists as a monomer in so- lution reacted with the antibodies in the absence of any denaturing treatments. These data indicate the car- boxyl-terminal region is inaccessible in the intact re- pressor tetramer and further suggest that denatura- tion/dissociation of a protein during the initial immu- nologic challenge may result in the production of mono- clonal antibodies to antigenic areas of the protein which are not exposed in the native conformation.

The lactose repressor protein controls the expression of the lactose metabolizing enzymes in Escherichia coli (1). This regulation involves specific recognition of the operator DNA sequence and consequent inhibition of transcription initiation by RNA polymerase; modulation of binding at this target site by small sugar molecules, inducers, provides the basis for regulation. Although inducer binding decreases the affinity of repressor for operator DNA by several orders of magnitude, the interaction of the protein with nonspecific sequences on the DNA is not altered by sugar ligands (1). Treatment of the tetrameric repressor protein (Mr = 150,000) with proteases under mild conditions results in chain scission in the region between amino acids 50 and 60; the products of proteolysis are a tetrameric core protein (M, = 125,000) and 4 NH2- terminal fragments (Mr 2: 6,000) (2). Different ligand binding activities are associated with these isolated domains of the repressor protein: specific and nonspecific DNA binding with

* This work was supported by grants from the Robert A. Welch Foundation (C-576 and C-583), the National Science Foundation (PCM 80-12048), and the National Institutes of Health (GM 22441, HL 22532, and HL 15376). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

the NH2 termini (2-6) and inducer and specific DNA binding with the core protein (2,7,8). The importance of this genetic control protein and its availability in large amounts has provided the impetus for production and isolation of mono- clonal antibodies to lac repressor. Fusion technologies have been used to produce eight cell lines which secrete specific, unique antibodies to the lactose repressor protein. This paper describes the characterization of these mouse monoclonal immunoglobulins and their interaction with the repressor molecule.

EXPERIMENTAL PROCEDURES

Materiuls-M.A. Bioproducts, Walkersville, MD, supplied Dulbec- CO’S modified Eagle’s medium (4.5 g of glucose/l), fetal bovine serum, polyethylene glycol 1500, L-glutamine, powdered phosphate-buffered saline (Dulbecco’s modification), Costar 96-well flat-bottom plates, and Lux 60-mm Petri dishes. Microtiter plates (96 wells) and 25-cm2 and 75-cm2 T-flasks were from Falcon. Hypoxanthine, aminopterin, thymidine, and 5-bromodeoxyuridine were purchased from Sigma. Calbiochem Behring supplied Freund’s complete and incomplete ad- juvants. Pristane was from Aldrich and SeaPlaque Agarose was purchased from FMC Corp., Marine Colloids Division, Rockland, ME. All immunochemicals were from Miles Laboratories, Elkhart, IN. 2,2’-Azino-di-[3-ethylbenzthiazoline]-6’-sulfonate was from Boehringer Mannheim. All other chemicals were reagent grade.

Isolation of Repressor, Core Protein, and NH2 Terminus-Lactose repressor protein was isolated from E. coli CSH 46 according to the methods described previously (8, 9). Cells were grown in 100-liter batches and stored frozen before use. The yield of pure repressor was 1-2 mg/g of frozen cells. The purity of the isolated protein (>95%) was assessed by SDS’-gel electrophoresis (10).

Core protein was purified according to the method described by Matthews (8) with the exception that the digestion was by chymo- trypsin in 1 M Tris-HC1, 30% glycerol, 0.1 mM dithiothreitol, pH 7.4, and both phosphocellulose columns were in 0.048 M potassium phos- phate, pH 7.4,5% glucose, 0.1 M dithiothreitol. The purity of the core protein (>98%) was assessed by SDS-gel electrophoresis (10) and binding activities were assayed RS previously described (8). NH, termini were isolated as described by Arndt et al. (11); SDS-gel electrophoresis indicated >95% purity.

Cells-The immunologically silent murine myeloma cell line S194/ 5.XX0.BU.1, originally from Dr. R. Hyman of the Salk Insititute, was a gift from Dr. David M. Miller, 111, Department of Neurology, Baylor College of Medicine. This cell line, which is TK- and resistant to 1 X lo“ M 5-bromodeoxyuridine, was maintained in Falcon flasks (25 cmZ) containing Dulbecco’s modified Eagle’s medium supple- mented with 2.5 mM glutamine and fetal bovine serum (20 m1/100 ml of medium). The cells were maintained in 1 X 10“ 5-bromodeoxyuri- dine for 96 h prior to fusion.

Zmmunization-Male BALB/C mice, 12-16 weeks old, received primary intraperitoneal injections of 50-80 Gg of lac repressor in Freund’s complete adjuvant. Secondary immunizations were carried out every 30 days with 80-100 fig of protein in Freund’s incomplete

The abbreviations used are: SDS, sodium dodecyl sulfate; PBS, phosphate-buffered saline; ELISA, enzyme-linked immunosorbent assay; IPTG, isopropyl-l-thio-8-D-galactopyranoside.

1185

1186 Monoclonal Antibodies to lac Repressor Monomer

adjuvant. Spleens were taken for fusion 3 days after a secondary immunization.

Fusions-All fusions were performed essentially as described by Galfr6 et al. (12); cells were inoculated into Costar 96-well flat-bottom plates (1 X lo6 cells/well) containing 2.5 X 10' mouse peritoneal macrophages (13). Hybrids were selected in HAT medium (14) and those populations producing antibodies to lac repressor protein were cloned into soft agar as described by Kennett (15) except that 60-mm Petri dishes were used instead of tubes and thymocytes (2 x IO6 cells/ ml) were used as feeder cells. Subcloning was continued until stable antibody-producing cells lines were obtained.

ELZSA-Antibodies to lac repressor protein were detected using the ELISA method essentially as described by Engvall and Perlman (16). Color generated was scored by eye or by Titertek Multiscan Plate Reader.

Antibody Production and Purification-Male BALB/C mice were injected intraperitoneally with 0.5 ml of Pristane and allowed to recover for 4-8 weeks. Pristane-primed animals were injected intra- peritoneally with 1.0-5.0 X 10' antibody-producing hybrid cells. Tu- mors appeared within 3-10 days. The ascitic fluid was periodically drained and fluid from several animals injected with the same hybrid was combined. IgG antibodies were purified by ammonium sulfate precipitation and passage through a DEAE-column. IgM antibodies were purified by molecular sieve chromatography on Ultrogel AcA34.

Subclass Specijicity-A modification of the ELISA method was used to determine immunoglobulin subclasses. Briefly, lac repressor protein (-2 pg) was bound to 96-well microtiter plates and treated with antibody or ascitic fluid. Control sera from repressor-immunized mice and nonimmunized mice were also tested. The wells were washed with PBS, then incubated with the desired rabbit anti-mouse immu- noglobulin (Miles, anti-K, -yl, -y2,, - ~ 2 ~ , -y3, or p ) . The excess immu- noglobulin was removed and peroxidase-conjugated goat anti-rabbit IgG was added. Following removal of excess peroxidase antibody and washing with PBS, antibody recognition was visualized by addition of peroxidase substrate. Positive reactions were scored by eye.

Electrophoresis and Transfer-SDS-gel electrophoresis was per- formed according to the procedure of Laemmli (17). Protein transfer to nitrocellulose was carried out by diffusion as described by Bowen et al. (18) or electrophoresis (19). After transfer to nitrocellulose, proteins were visualized by staining in 0.1% naphthol blue black (in 45% methanol, 5% acetic acid, 50% H2O) or used for antibody analysis. For this analysis, nitrocellulose transfers were washed with PBS, then incubated with 0.5% bovine serum albumin in PBS for 2 h to saturate available sites on the nitrocellulose. The transfers were rinsed free of bovine serum albumin with numerous PBS washes and treated with a 10 pg/ml antibody solution or ascitic fluid for approx- imately 3 h. They were again washed extensively with PBS, then incubated overnight with rabbit anti-mouse antibody linked with peroxidase (usually 0.02% in 0.1% bovine serum albumin/PBS). After an extensive PBS wash, the transfers were developed by immersion in 0.3 mg/ml 4-chloro-l-naphthol, 0.01% Hz02 in PBS. Stained transfers were rinsed in PBS and dried for storage.

Cyanylation-Cyanylation and cleavage at the cysteine residues were performed by a modification of the procedure of Jacobson et al. (20). Repressor was reacted with a 600-fold excess of 2-nitro-5- thiocyanatobenzoic acid in 0.2 M Tris-HC1, pH 8.0, at 32 "C. After 1 h, urea was added to 8 M final concentration and another 600-fold molar excess of reagent was added. The pH was increased to pH 9.0 by the addition of Tris base and incubation of the reaction mixture was continued overnight at 44 "C. The resulting repressor fragments were separated on a 2 X 110-cm Sephadex G-75 column equilibrated in 50 mM Tris-HC1, pH 8.8, containing 4 M urea. Alternatively, the fragments were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose for Western analysis.

Mutant Repressor Monomer-Monomer of the repressor mutant, 7'41 (21), was partially purified by an adaption of the procedure used for wild-type lac repressor. The E. coli were frozen and subsequently homogenized in buffer using a Waring Blendor at low speed. The resulting cell lysate was fractionated by precipitation with 37.5% ammonium sulfate. The precipitate was resuspended in 0.048 M potassium phosphate containing 5% glucose and M dithiothreitol and loaded onto a phosphocellulose column equilibrated with the same buffer. The column was eluted with 0.048 M potassium pbos- phate buffer and the monomer was located by optical density at 280 nm and by inducer assay. This partially purified monomer was used in further experiments.

RESULTS

Production and Purification of Monoclonal Antibodies- Hybrid cell lines secreting antibody against the lactose re- pressor protein were produced by fusion of myeloma cells with lymphocytes from immunized mice. Eight cell lines (B-1,2; M-1 through -6) were isolated and cloned into soft agar. After stabilization of the cell lines, intraperitoneal injection of cells was used to produce ascitic fluid in Pristane-primed mice. The ascitic fluids from several mice injected with the same cell line were pooled. Purified IgGs were obtained from two cell lines (B-1 and B-2) after ammonium sulfate precipitation of the ascitic fluid at 50% saturation followed by DEAE- cellulose chromatography. The purity of the antibody as well as the molecular weight of the individual chains were deter- mined by SDS-gel electrophoresis. The subunit molecular weights corresponded to those expected for mouse IgG. Ascitic fluid from one of the cell lines producing IgMs ("1) was subjected to molecular sieve chromatography to produce a purified IgM. The purified IgM M-1 was >95% electrophoret- ically pure.

Characterization of the Antibodies-The purified antibodies and ascitic fluids were characterized with regard to subclass specificity. The results shown in Table I established the presence of K-light chains and 7,-heavy chains in the purified IgG immunoglobulins; no reaction was observed for either of these monoclonal antibodies with anti-IgM antisera. The purified IgM M-1 antibody reacted with anti-r and anti-p antisera and did not cross-react with any anti-y antisera. Similar results were observed for "2, "3, "4, and " 6 , while M-5 cross-reacted with anti-y3 antisera; the latter result suggested the possibility that both IgG and IgM were produced by cell line "5.

Interaction of the Antibodies with Repressor, Core, and Amino Terminus-The repressor protein can be divided into two domains by treatment with chymotrypsin. NH2 terminus (amino acids 1-56) and core (amino acids 57-360) were easily separated on SDS-polyacrylamide gels and subsequent trans- fer to nitrocellulose was utilized to test each of the antibodies for recognition of these individual repressor domains. All of the purified antibodies as well as all of the ascitic fluids recognized the intact repressor and. the core protein as deter- mined by Western analysis (Fig. 1). None of the antibodies or ascitic fluids exhibited any binding to the NHn-terminal domain.

The binding of the purified IgG antibodies to the lactose repressor protein was assessed using protein bound irreversi- bly to microtiter plates. While this method does not yield a thermodynamic dissociation constant, it can be used to esti- mate the affinity of the interaction between the two protein molecules. The results were monitored by measuring the absorbance of dye produced by horseradish peroxidase in the enzyme-linked immunoassay in each well; by this assay, B-2

TABLE I Subclass specificity of monoclonal antibodies

Isolated monoclonal antibodies or ascitic fluid was reacted with the indicated antisera (results shown by plus and minus signs) in microtiter plates as described under "Experimental Procedures."

Monoclonal antibody

B-1 B-2 M-1 M-2 M-3 M-4 M-5 M-6

YI + + - - - - - - I& YZ. IgG yzb IgG 7 3 IgG K + + + + + + + +

- - + - IgG P " + + + + + +

_ " - - - " " _ " - - - " "

Monoclonal Antibodies to lac Repressor Monomer 1187

REPRESSOR - lL, _. - - CORE PROTEIN -

NH,-TERMINUS-

z

FIG. 1. Antibody recognition of lac repressor, core, and NH2 terminus. Approximately 5-10 pg of repressor, core, or NH, terminus were separated on a 10% polyacrylamide-SDS gel. After electrophoresis, the proteins were transferred to nitrocellulose by diffusion (18) and the transfers were treated with the indicated antibody. The transfers were developed as described under "Experimental Procedures." A, SDS-polyacrylamide gel of NHQ terminus, core, and lac repressor stained with Coomassie Blue. Western blots of NHZ terminus, core, and luc repressor treated with B-1 antibody ( R ) and B-2 antibody (C). Western blots treated with purified IgM M-1 and ascitic fluid 2CB1 ("3) are shown in D and E , respectively. The relative positions of NH, terminus, core. and repressor are reversed for these transfers.

exhibits a higher apparent affinity for the repressor protein than B-1. An approximate dissociation constant for B-1 is -10" M, while an approximate dissociation constant for B-2 is -lo+' M.

Attempts to immunoprecipitate antibody-repressor com- plexes under a wide variety of conditions were not successful. Native repressor filtered onto nitrocellulose did not bind antibodies. Fluorescence measurements on protein-antibody complexes (repressor, M; antibody, 1-6 X lo" M) dem- onstrated no alterations in tryptophan fluorescence emission intensity or wavelength maximum. Thus, antibody reaction with repressor was observed only under conditions of dena- turation or irreversible binding to plastic plates.

Determination of the Site of Interaction-The repressor protein was chemically divided into smaller fragments to determine the site of recognition by the antibodies more precisely. Repressor was reacted with 2-nitro-5-thiocyanato- benzoic acid to cyanylate the cysteine residues. Incubation of the modified repressor a t pH 9.0 overnight a t 44 "C resulted in the cleavage of the protein at the modified cysteine residues. The repressor was thereby divided into four fragments (Fig. 2): fragment I (amino acids 1-106), fragment I1 (amino acids 107-139), fragment 111 (amino acids 140-280), and fragment IV (amino acids 281-360). Oligopeptides resulting from in- complete cleavage a t each cysteine were also produced. These fragments were separated by fractionation on a Sephadex G- 75 column (Fig. 3), analyzed by electrophoresis in SDS- polyacrylamide gels, and tested for antibody recognition by Western blots. Alternatively, the mixture of fragments was separated directly on SDS gels and tested for antibody rec- ognition by Western analysis. All of the antibodies and ascitic fluids recognized fragment IV (amino acids 281-360, Fig. 4). Purified fragment IV was digested witb trypsin or chymotryp-

A 1 1 0 161 360

8 1 80 00 2 75 380

DNA Blndlng Hlnge I n d u c e r B l n d l n g A l a e m b l y

FIG. 2. Cleavage fragments of lac repressor. A . fragments resulting from the chemical cleavage of /or repressor at the cysteines as described in the text. The positions of the cysteine residues in the primary sequence are given above the schematic. H . functional do- mains of the repressor as defined by genetic studies (21, 28).

sin in an attempt to further refine the site recognized by the antibodies. Following digestion with the proteolytic enzyme, the incubation mix was filtered through nitrocellulose to bind the resulting peptides. Processing the nitrocellulose sheet for antibody reaction yielded no reaction for the treated fragment; these data indicated that treatment with either of these en- zymes resulted in the loss of antibody recognition.

Antibody RecoEnition of Tetrameric and Monomeric lac Re- pressor-We were unable to detect any interaction between the purified antibodies and native lac repressor tetramer by immunoprecipitation, dot blot analysis on nitrocellulose, or alterations in intrinsic fluorescence. The interaction of the antibodies with repressor as measured by ELISA on microtiter

1188 Monoclonal Antibodies to lac Repressor Monomer

I n 0.4 I

- a0 100 120 140 160

FRACTION NUMBER FIG. 3. Isolation of lac repressor fragments. Fragments of lac

repressor generated hy chemical cleavage with 2-nitro-5-thiocyana- tohenzoic acid were separated on a Sephadex G-75 column equili- hrated in 50 mM Tris-HCI, pH 8.8, containing 4 M urea. The elution of the fragments was monitored hy optical density a t 230 nm. Column fractions were collected by drop counting which results in deviations from linearity of the fraction number with time. Actual positions of the fractions were noted hy an event marker. Inset. electrophoresis of various fractions on an SDS-polyacrylamide gel. Aliquots from several points along the elution profile were taken and analyzed by SDS electrophoresis. Arrows A-F indicate the positions of the frac- tions corresponding to the indicated lanes in the SDS gel. The bar on the graph and the ? signs on the inset indicate the fractions contain- ing the lowest molecular weight fragment to give positive reaction with H-1 antihody hy Western analysis of the gel. The peak under arrow F corresponds to fragment IV (amino acids 281-360). The positions of molecular weight standards are shown in the first lane: I’HH, phosphorylase €3 ( M , = 92,000); RSA, bovine serum albumin ( M . = 65,000); OVA, ovalhumin ( M , = 45,000); CAR, carbonic anhy- drase ( M . = 31.000): STI. sovhean trypsin inhibitor ( M , = 21,500); L Y S , lysozyme ( M , = 14.400).

. I _

A

111 - I -

IV-

plates was generally positive, although some variability in these assays was observed if care was taken to facilitate repressor stability. In contrast, the antibodies exhibited high affinity for repressor after electrophoresis on SDS gels and transfer to nitrocellulose. These data suggested that the an- tibodies may recognize a denatured form of the repressor or some part of the repressor which is not exposed in the native tetramer. To test this hypothesis, repressor was treated with low percentages of SDS (0.01T to O . l T ) and then filtered through nitrocellulose. The nitrocellulose sheet was incubated with antibody and stained with peroxidase-linked second an- tibody. The antibodies recognized the repressor after incuba- tion with as little as 0.01% SDS (Table 11). Increasing the SDS concentration to 0.03% gave maximal antibody response. Although this amount of SDS dissociates the repressor to monomers, the preservation of IPTG binding activity indi- cates that no major denaturation has occurred (22).

To further explore the possibility that reaction with anti- body required dissociation of the tetramer, a mutant repressor, T41, which exists primarily as the monomer and has normal IPTG activity (21), was partially purified and tested for recognition by the antibodies. The crude T41 repressor was fractionated on Sephadex G-7.5 and the fractions were assayed for inducer binding and antibody reaction. The results are shown in Fig. 5. The IPTG activity and antibody reactivity co-elute a t a position which calibrates to a molecular weight in the range of 33,000-38,000 and corresponds well to the expected molecular weight for the repressor monomer (37,500). The monomer is well resolved from the tetrameric repressor as indicated by the relative positions of elution. The antibodies, therefore, interact with repressor which has been either dissociated into native monomers or denatured to ex- pose the antigenic determinants.

DISCUSSION

Monoclonal antibodies have been produced which are spe- cific for the lac repressor protein. It-is noteworthy that e k h of the eight different monoclonal antibodies produced (two IgGs and six IgMs) recognizes only the carboxyl-terminal region of the protein. This recognition occurs only when the protein is in the monomeric form or denatured completely, as there is complete absence of immunoglobulin reaction with the native molecule. Dissociation of the protein by chemical means yields a monomer which hinds to inducer (22) and which can be recognized by the antibodies. In addition, a mutant repressor which exists predominantly as monomer (21) exhibits a high degree of reactivity with the immunoglob- ulins. The region of the protein which serves as antigen for these antibodies is the carboxyl-terminal 80 amino acids; genetic studies (21, 23) have demonstrated that mutations in

B C

TAR1.E 11 Antibody recognition of diwociatpd/dpnaturpd lac rpprwsor

Repressor (25 pg) was incubated with the indicated amount of SDS and filtered through nitrocellulose. Dot blots were washed with PHS. treated with the indicated antihody, and processed as descrihed under “Experimental Procedures.” Antibody recognition is indicated hy pluci and minus signs.

~~

Monnclonnl nnti- -~

a sns M y FIG. 4. Antibody recognition of lac repressor fragments. A,

SDS-polyacrylamide gel of cyanylation fragments of lac repressor stained with Coomassie Blue. The positions of the different cleavage fragments (as numhered in Fig. 2) are shown to the left of lane A. LHI’ designates the position of lac repressor. Western blots of R, lac cyanylation fragment mixture and C, isolated fragment IV treated with H-2 antihodv.

. R- 1 R.2

”~

0.00 - - 0.01 + + 0.03 ++ ++ 0.06 ++ ++ 0.10 + +

~

Monoclonal Antibodies to lac Repressor Monomer 1189

400 t I' i6

FRACTION NUMBER FIG. 5. Antibody recognition of T41 repressor monomer.

Partially purified T41 repressor was fractionated on Sephadex G-75 and the resulting fractions were assayed for IPTG binding (c".) and antibody recognition (stippled shading). Inducer binding was assayed by ammonium sulfate precipitation assay (26). Antibody binding was measured by dot blot and color development was scored by eye on a scale of zero to four. The elution positions of several molecular weight markers are shown: LRP, lac repressor protein; BSA, bovine serum albumin; STI, soybean trypsin inhibitor; and LYS, lysozyme. The molecular masses are indicated in kilodaltons within the parentheses.

this segment of the molecule affect the assembly of the subunits into the native tetrameric state. The inaccessibility of this antigenic site in the native quaternary structure of the lac repressor corroborates the genetic results and suggests that this portion of the primary structure of the protein participates in intersubunit contacts. These studies provide the first chemical definition of the subunit interface in the lactose repressor protein.

Only antibodies to the carboxyl-terminal region of the repressor were obtained in these experiments and eight inde- pendent isolates yielded products which recognize the same 80-amino-acid fragment of the protein. This region of the repressor must be strongly antigenic, since the methods used for screening the hybridoma population were not designed to select for recognition of this portion of the protein. Production of antibodies to other areas of the protein may thus require methods which maintain the tetrameric structure or remove this region prior to immunization. Preliminary results indi- cate that cross-linking of the protein prior to initial challenge yields antibodies in whole animals (rabbits and mice) which recognize the native structure.' Alternately, specific repressor fragments (e.g. NH, terminus) might be used for immuniza- tion.

Production of antibodies in animals injected with proteins suspended in a carrier such as Freund's adjuvant (heat-killed Mycobacterium tuberculosis in 85% paraffin oil and 15% man- nide monooleate) is a generally useful procedure. However, significant alterations in the tertiary/quaternary structure may occur as a consequence of suspension in adjuvant, so that the protein is not in its native form when challenge occurs in the animal. Failure of antibodies obtained by these methods

D. Mandoli, T. J. Daly, R. Davis, and K. S. Matthews, unpub- lished results.

to react with the protein under native conditions when rec- ognition is observed by other standard assays (e.g. ELISA, Western) may indicate that the antigenic determinants to which antibodies have been generated are not available in the native structure. When the antibodies are to be used in applications where protein is denatured (e.g. Western blots), these considerations are not essential. However, antibodies used for binding perturbation, immunoprecipitation, or fluo- rescence studies must bind to the protein in its native state. Depending upon the intended use of the antibodies being produced, the tertiary/quaternary structure of the antigen to which the animal is exposed may be of considerable impor- tance. Difficulties in production of immunoglobulins to many native proteins are widely discussed but not often presented in the literature; these problems may derive from the exposure of antigenic sites not normally accessible in the native struc- ture during immunization.

Bartel and Campbell (24) observed almost a quarter century ago that polyclonal antibodies generated against a protein may exhibit different precipitation titrations for native and denatured antigen. Further experiments by Ishizaka et al. (25) demonstrated that the serum of animals challenged with an intact protein contained antibodies which reacted with pro- teolytic fragments of the antigen and did not recognize the intact molecule. Our findings corroborate these results and, additionally, indicate the immunization/selection protocol may influence the nature of the antibodies obtained. Disso- ciation of oligomeric proteins may expose sites which are at subunit interfaces, as in the case of the lac repressor, while denaturation will reveal residues which are buried in the native structure. This process may result in the production of antibodies (whether poly- or monoclonal) which exhibit high affinity for denatured/dissociated protein but are incapable of recognizing the antigen in the native protein. When a region exposed during immunization is highly antigenic com- pared to sites available in the native conformation and anti- bodies to the native protein are required, the procedure for selection of antibodies must be designed to discriminate be- tween antibodies which recognize native versus non-native protein.

In summary, we have isolated 8 cell lines which produce monoclonal antibodies to the lac repressor. All of the immu- noglobulins recognize the same region of the primary sequence and none reacts with native protein. Chemical dissociation of the tetramer into monomer and isolation of mutant protein which exists as monomer both result in species which bind well to the antibodies and maintain their tertiary structure as reflected by their inducer binding activity. These results dem- onstrate that the antigenic site recognized by the immuno- globulins is not accessible in the tetrameric form and may form a part of the subunit interface. The inability to isolate antibodies (polyclonal or monoclonal) to the native structure of many proteins may derive at least in part from the disrup- tion of tertiary/quaternary structure during immunization and consequent exposure of highly antigenic sites not avail- able in the native protein.

Acknowledgments-We thank Dr. L. C. Smith, Department of Medicine, Baylor College of Medicine, for use of the Titertek Multi- scan plate reader and their in-house software and Dr. David M. Miller, 111, for cell lines and invaluable advice and consultation.

REFERENCES

1. Miller, J. H., and Reznikoff, W . S. (1980) The Operon, 2nd Ed, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

2. Platt, T., Files, J. G., and Weber, K. (1973) J. B i d . Chern. 248, 110-123

1190 Monoclonal Antibodies to lac Repressor Monomer

3. Geisler, N., and Weber, K. (1977) Biochemistry 16,938-943 4. Jovin, T. M., Geisler, N., and Weber, K. (1977) Nature (Lord.)

5. Ogata, R. T., and Gilbert, W. (1978) Proc. Natl. Acad. Sci. U. S.

6. Ogata, R. T., and Gilbert, W. (1979) J. Mol. Biol. 132, 709-728 7. Butler, A. P., Revzin, A., and von Hippel, P. H. (1977) Biochem-

8. Matthews, K. S. (1979) J. Biol. Chem. 254, 3348-3354 9. Rosenberg, J . M., Kallai, O., Kopka, M. L., Dickerson, R. E., and

Riggs, A. D. (1977) Nucleic Acids Res. 4, 567-572 10. Weber, K., Pringle, J. R., and Osborn, M. (1972) Methods En-

11. Arndt, K. T., Boschelli, F., Lu, P., and Miller, J. (1981) Biochern-

12. GalfrC, G., Howe, S. C., Milstein, C., Butcher, G. W., and Howard,

13. De St. Groth, S. F., and Scheidegger, D. (1980) J. Immunol.

14. Littlefield, J. (1964) Science 145, 709-710 15. Kennett, R. H. (1980) in Monoclonal Antibodies (Kennett, R. H.,

269,668-672

A. 75,5851-5854

istry 1 6 , 4757-4768

~yrnol. 26,3-27

istv 20,6109-6118

J. C. (1977) Nature (Lond.) 266, 550-552

Methods 35, 1-21

McKearn. T. J.. and Bechtol. K. B.. eds) aa. 372-373. Plenum Press, New York

I . "

16. Enmall. E.. and Perlmann. P. (1972) J. Immunol. 109. 129-135 17. Laemmli, U . K. (1970) Nature (Lord.) 227,680-685 . 18. Bowen, B., Steinberg, S., Laemmli, U. K., and Weintraub, H.

19. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. A d .

20. Jacobson, G. R., Schaffer, M. H., Stark, G. R., and Vanaman, T.

21. Schmitz, A., Schmeissner, U., Miller, J. H., and Lu, P. (1976) J.

22. Hamada. F.. Ohshima. Y.. and Horiuchi. T. (1973) J. Biochem.

(1980) Nucleic Acids Res. 8, 1-20

Sci. U. S. A. 76, 4350-4354

C. (1973) J. Biol. Chem. 248, 6583-6591

Biol. Chem. 251, 3359-3366

73,1299-1302 , , , . .

23. Miller. J. H. (1980) in The ODeron (Miller. J. H.. and Reznikoff. W. S., eds) 2nd Ed, pp. 31-88, Cold Spring Harbor Laboratory: Cold Spring Harbor, NY

24. Bartel, A. H., and Campbell, D. H. (1959) Arch. Biochem. Biophys.

25. Ishizaka, T., Campbell, D. H., and Ishizaka, K. (1960) Proc. SOC.

26. Bourgeois, S. (1971) Methods Enzymol. 21D, 491-500

82,232-234

Exp. Biol. Med. 103, 5-9