ANALYTICAL

BIOCHEMISTRY

205, 70-76 (19%)

A Strategy for Generating Monoclonal Antibodies against Recombinant Baculovirus-Produced Proteins: Application to the Bcl-2 Oncoprotein John C. Reed,l Shigeki Tanaka, Michael Cuddy, David Cho, Joseph Smith,* Roland Kallen,* H. Uri Saragovi, and Toshihiko Torigoe Departments Philadelphia,

of Pathology

Pennsylvania

Received January

& Laboratory 19104-6082

Medicine,

and *Biochemistry

University

of Pennsylvania,

15, 1992

A strategy is described for production of monoclonal antibodies against recombinant proteins that are produced using the baculovirus expression system and that requires no prior purification of the protein of interest. Crude lysates prepared from cultured Sf9 insect cells infected with recombinant or control baculoviruses are adsorbed to nitrocellulose filters and used in a dot-immunobinding assay for screening hybridomas. The monoclonal antibody-producing hybridomas are derived by immunization of mice with a synthetic peptide corresponding to a hydrophilic region in the recombinant protein of interest. By using the baculovirus-produced recombinant protein as the screening antigen and by comparing antibody binding to filters containing control Sf9 lysates, hybridomas are identified that produce monoclonal antibodies with specific reactivity for the recombinant protein of interest and that can then subsequently be used to assist in the large-scale purification of the recombinant protein from baculovirus-infected cells. We applied this method to recombinant 26kDa human Bcl-2 (B-cell lymphomalleukemia-2), an integral membrane oncoprotein that regulates programmed cell death (“apoptosis”) in hematolymphoid cells through unknown mechanisms. Two mouse monoclonal antibodies were produced that specifically bound the recombinant Bcl-2 baculoprotein in both solution o I992 Academic PWS, ILIC. and solid-phase assays.

The baculovirus expression system has been of tremendous benefit for the production of full-length, inte1 To whom correspondence should be addressed at La Jolla Cancer Research Foundation, Cancer Research Institute, 10901 N. Torrey Pines Rd., La Jolla, CA 92037. Fax: (619) 450-4693.

70

& Biophysics,

gral membrane proteins which are often difficult to produce in Escherichia coli (1). Unlike bacterial cells, however, the subsequent purification of the recombinant protein of interest from baculovirus-infected insect cells is often laborious, necessitating several chromatographic steps (2,3). Access to a monoclonal antibody (MAb)2 that specifically recognizes the recombinant protein can greatly simplify the purification (4,5). Unfortunately, a source of purified or at least highly enriched specific protein is required for immunizations of mice prior to hybridoma production. While the low complexity of E. coli with regards to the variety of proteins present in these bacteria typically permits partial purification of recombinant proteins by simple preparative SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (6), often baculoproduced recombinant proteins are insufficiently abundant to be visualized in Coomassie-stained gels containing crude insect cell lysates (2,3). An alternative approach is to utilize synthetic peptides as immunogens in an effort to produce MAbs that specifically bind the recombinant protein of interest. However, if the same peptide is used either directly or conjugated to a carrier protein for screening hybridomas, many of the resulting MAbs typically recognize the peptide, carrier protein, or crosslinking moiety used

’ Abbreviations used: MAb, monoclonal antibody; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; DIA, dot-immunobinding assay; m.o.i., multiplicity of infection; Bcl-2, Bcell lymphoma/leukemia-2; Bat-Bcl-2, baculo-Bcl-2, PBS-phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride; BSA, bovine serum albumin; KLH, Keyhole Limpet Hemocyanin; HA, hypoxanthine and azaserine; ELISA, enzyme-linked immunosorbend assay; AcMNPV, Autographa californiu multiply enveloped nuclear polyhedrosis virus. 0003-269’7192 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

A STRATEGY

FOR

GENERATING

to conjugate the peptide to the carrier protein, but fail to bind efficiently to the protein of interest (7). In contrast to using a synthetic peptide for screening hybridomas, we reasoned that if crude lysates of insect cells containing the recombinant protein of interest could be employed, the resulting MAbs would perforce be capable of recognizing the desired baculoprotein. Taking advantage of the high adsorptive capacity of nitrocellulose, therefore, we employed a dot-immunobinding assay (DIA) (8) wherein hybridoma culture supernatants were screened for differential binding to filters containing crude cell lysates derived from Sf9 cells infected with either specific or control recombinant baculoviruses. To test this strategy, we produced recombinant baculoviruses expressing the 26-kDa human Bcl-2 (B-cell lymphoma/leukemia-2) oncoprotein, an intracellular integral membrane protein that normally resides within mitochondria, shares no significant sequence homology with other known proteins, and blocks programmed cell death (“apoptosis”) through unknown biochemical mechanisms (9). Hybridomas derived from mice immunized with a synthetic peptide corresponding to a hydrophilic stretch of amino acids (61-76) within this protein were then screened initially with lysates containing the baculo-Bcl-2 protein (BacBcl-2). Of the 31 positive hybridomas, 2 were subsequently found to react specifically with the Bat-Bcl-2 protein in a secondary screen which compared MAb binding to Bat-Bcl-2 and a control lysate. Both of these MAbs were shown to react specifically with the BacBcl-2 protein in solution and solid-phase assays. MATERIALS

AND

METHODS

Production of recombinant baculouiruses. A 910-bp cDNA containing the complete open reading frame for the human p26-Bcl-2 protein was excised from the plasmid pB4 (10) using EcoRI, its ends blunted with Klenow fragment, and subcloned into the blunted BamHI site of the baculovirus transfer plasmid pAcYM1 (11). Recombinant plasmids containing the bcl-2 cDNA in the proper orientation were identified by restriction mapping and their structure was confirmed by DNA sequencing of the junctions between the plasmid and the cDNA insert. An additional recombinant transfer vector was similarly prepared using a 2040-bp EcoRI fragment from a plasmid pYT16 that contains a cDN,4 encoding the human p56-LCK kinase (12). The pAcYMl-BCL-2 and pAcYMl-LCK plasmids were banded twice in CsCl gradients and dialyzed extensively against 10 mM Tris/l mM EDTA (pH 8.0) prior to being cotransfected into Sf9 cells with DNA from the Autographa California multiply enveloped nuclear polyhedrosis virus (AcMNPV) as described previously (2,13). Recombinant viruses resulting from homologous recombination between the polyhedrin gene sequences within

MONOCLONAL

ANTIBODIES

71

the wild-type AcMNPV genome and the pAcYMlBCL-2 or pAcYMl-LCK transfer vectors were initially enriched by a limiting-dilution method employing a DNA dot-blot hybridization assay (14) and then twice plaque-purified (13,15). A recombinant baculovirus containing a cDNA encoding the human p72-Raf-1 kinase was provided by D. Morrison (NCI; Frederick, MD) (16). Production of recombinant baculoproteins. To produce recombinant baculoproteins for use in DIAs, 18 X lo6 Sf9 cells were grown in T-150-cm2 flasks and exposed to purified recombinant baculoviruses at a m.o.i. of ~10 for 1 h in 5 ml of Grace’s medium without serum. Virus was then removed and the cells were cultured in 20 ml of Grace’s complete medium [contains 10% fetal bovine serum (Hyclone), 0.33% (w/v) lactalbumin and yeastolate, 100 PgIml streptomycin, and 50 units/ml penicillin] for 3 days, which was the time empirically determined to be optimal for production of p26-BacBcl-2. Infected cells from one T-150-cm2 flask were washed once at 4°C with phosphate-buffered saline [PBS contains 12.5 mM NaPO, and 150 mM NaCl (pH 7.4)] containing 1 mM PMSF and then either lysed directly in 1 ml of SDS-lysis buffer [30 mM Tris (pH 6.6), 1% SDS, 5% glycerol, 0.5% 2-mercaptoethanol] or resuspended for 15 min in 0.5 ml of ice-cold RIPA lysis buffer [50 mM Tris (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1% SDS, 1% Na-deoxycholate] containing protease inhibitors (1 mM PMSF, 1 pM pepstatin, 10 pM leupeptin, 0.23 units/ml aprotinin, 1 mM benzamidine) and centrifuged at 16,000g for 15 min to remove nuclei and debris; the resulting supernatants were mixed with an equal volume of 2~ SDS-lysis buffer. All lysates were boiled for 5 min and then centrifuged at 16,OOOg for 5 min to remove debris prior to adsorption to nitrocellulose filters. DIA for hybridoma screening. Sf9 lysates were diluted into PBS containing 1 mM PMSF and 0.1% NaN, to achieve final concentrations for SDS of -0.05% and for protein of ~0.5 mg/ml. Nitrocellulose filters (11.5 X 9 cm) were wet in water, incubated in PBS for 5 min, and then incubated from 6 to 14 h with lo-20 ml of diluted Sf9 lysates. In some cases, nitrocellulose filters were similarly incubated with 50 /*g/ml of BSA-peptide conjugates for 3-6 h. Filters were then placed onto glass plates and allowed to air-dry. Typically, several filters were prepared in advance and stored between sheets of Whatman 3MM paper in heat-sealed plastic bags at -80°C. For DIAs, antigen-containing filters were wet in water, then preblocked in PBS containing 4% BSA and 0.1% NaN, for 2 h, placed onto two sheets of Whatman 3MM paper that was wet in PBS, and placed into a 96-well dot-blot apparatus (Minifold-I; Schleicher & Schuell, Inc.). One-hundred microliters of antibody

72

REED ET AL.

binding solution (PBS with 2% BSA and 0.1% NaN,) was added to the manifold wells followed by 100 ,~l of hybridoma supernatants. Alternatively, rabbit antisera diluted to 0.1-l% (v/v) in antibody-binding solution was added. In some cases, 2 ,ug of competing peptide was included in the manifold wells. Parafilm was placed across the apparatus to prevent evaporation, and after 2 h the residual solution was briefly filtered through by application of vacuum. Approximately 400 ~1 of antibody wash buffer [12.5 mM NaPO, (pH 7.4), 400 mM NaCl, 0.25% Tween 201 was added to all wells, and the solution was filtered through by brief application of vacuum without drying of the membranes. This washing procedure was repeated four more times before dissembling the manifold apparatus and washing each filter (200 ml per filter) for 5-10 min each. Filters were then preblocked for 0.5-l h, incubated with 1 /*g/ml affinitypurified rabbit anti-mouse IgG (CappeUOrganon Teknika, Inc.) for 2 h, washed five times for 5-10 min, preblocked again for 0.5-l h, incubated with 0.25 &i/ml affinity-purified ‘251-protein A (Amersham, Inc.) for 2 h, washed five times, and exposed to x-ray film for 1 h with intensifying screens. Antibody and hybridoma production. Polyclonal antisera were raised in rabbits using synthetic peptides conjugated to Keyhole Limpet Hemocyanin (KLH) as described previously in detail (17). All of the synthetic peptides employed here contained either amino- or carboxyl-terminal cysteine residues which were utilized for conjugations to maleimide-activated carrier proteins (Pierce, Inc.) by the method of Partis et al. (18). The antisera specific for peptides corresponding to amino acids 41-54 and 61-76 of the human p26-BCL-2 protein have been described previously (17). Additional antisera were similarly raised using synthetic peptides corresponding to amino acids 1-13 of p26-BCL-2 (MAHAGRTGYDNRE[C]) and to 315-326 of the v-Raf protein ([CITLTTSPRLPVF). For hybridoma production, 4-week-old female Balb/c mice were immunized six times at 2-week intervals with 200-400 pg of KLH-Bcl-2 peptide (61-76) using a combination of intraperitoneal and intradermal injections without adjuvants. Four days after the last boost, mice were sacrrficed and their splenocytes isolated and fused at a 1O:l ratio with SP/O2 myeloma cells using a solution containing 50% polyethylene glycol and 5% dimethylsulfoxide, essentially as described by Lane et al. (19). Selections were performed using media supplemented with hypoxanthine and azaserine (HA) after plating 2 X lo4 cells per well in 200 ~1 into 96-well round-bottom plates (Costar, Inc.). Hybridomas displaying reactivity in DIAs were transferred to 24-well plates containing 2 ml of fresh HA-containing medium and these culture supernatants were recovered 3-5 days later for use in secondary-screening DIAs. The resulting 4D7 and 2E8 hybridomas were further expanded in culture and

cloned twice by limiting dilution on splenocyte feeder layers, and their MAb-containing culture supernatants used for immunoblot and immunoprecipitation assays. 4D7 and 2E8 were determined to be IgG, and IgM MAbs, respectively, by an ELISA method (BoeringerMannheim). Immunoprecipitation and immunoblot assays. For immunoprecipitations, 50 pug of protein in 500 jd of a solution containing 50 mM Tris (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, and protease inhibitors (see above) was incubated with 30 ~1 of hybridoma supernatants overnight at 4”C, with or without 10 pg of competing peptide. Immune complexes were collected using fixed Staphyloccous aureus (Pansorbin; Calbiothem, Inc.) that had been coated with rabbit anti-mouse IgG or rabbit anti-mouse IgM [l pg per 30 ~1 of 50% (v/v) Pansorbin suspension in RIPA with 1% BSA] and washed extensively, and the antigens eluted using Laemmli buffer, boiled, and subjected to SDS-PAGE followed by blotting to nitrocellulose. Blots were incubated with 0.1% (v/v) anti-Bcl-2 rabbit antiserum (directed against peptide 61-76) and processed as described previously (17,20). In some cases, immunoblots were incubated with 5-10 ml of hybridoma supernatants followed by appropriate secondary rabbit antisera specific for mouse IgG or IgM, and then 0.25 &i/ml of ‘251-protein A. RESULTS

Production and characterization of recombinant 26kDa human Bcl-2 protein using the baculovirus system. Having produced recombinant BCL-2 baculoviruses (AcMNPV-BCL-2), we first immunologically confirmed that Sf9 cells infected with this virus produced the proper Bcl-2 protein. Figure 1 shows an immunoblot analysis of AcMNPV-BCL-2-infected Sf9 cells, using a polyclonal rabbit antisera that we raised previously against a synthetic peptide corresponding to amino acids 61-76 of the human Bcl-2 protein (17). As shown, a 26-kDa protein was detected in St9 cells infected with AcMNPV-BCL-2 that comigrated precisely in gels with the p26-BCL-2 protein derived from a human lymphoma cell line. These RS11846 lymphoma cells contain a t(14;18) translocation which causes overexpression of the Bcl-2 proto-oncogene (11). The lack of immunoreactivity with Sf9 lysates prepared from uninfected cells and cells infected with a control virus AcMNPV-LCK excluded the possibility that the protein we detected in AcMNPV-BCL-2-infected cells represented a fortuitous cross-reactivity with a baculovirus or Sf9 cell protein. Though the recombinant Bcl-2 protein was roughly 40-times more abundant in AcMNPVBCL-2-infected St9 cells than in t(l4;18)-containing human lymphoma and leukemia cell lines, no specific 26-kDa band corresponding to Bat-Bcl-2 could be de-

A STRATEGY

FOR

GENERATING

MONOCLONAL

1 123

73

ANTIBODIES

2

3

4

5

6

456 33 kD

A 33 Kd

28

28 p26Bcl-2

FIG. 1. Characterization of p26-Bat-Bcl-2 recombinant protein. Sf9 cells were infected for 3 days with either AcMNPV-BCL-2 or AcMNPV-LCK viruses at a m.o.i. of 10, and postnuclear, detergent lysates were prepared for analysis by immunoblot (A) and immunoprecipitation (B) assays. In (A), proteins were subjected to SDS-PAGE (12% gel) and transferred to nitrocellulose for incubation with 0.1% (v/v) and autoradiography. Lanes represent (l-3) 5,1, rabbit antiserum raised against Bcl-2 peptide 61-76, followed by iz51-protein A (0.25 &i/ml) and 0.2 pg of protein, respectively, from Sf9 cells infected with AcMNPV-BCL-2; (4) 100 pg protein from the RS11846 human lymphoma cell or uninfected Sf9 cells, respectively. In (B), 7.5 wg of protein prepared line; (5,6) 5 gg protein derived from control-infected (AcMNPV-LCK) from AcMNPV-BCL-2-infected Sf9 cells was subjected to immunoprecipitations using rabbit antisera raised against peptides corresponding to amino acids l-13 (lanes 1,2), 41-54 (lanes 3,5), and 61-76 (lanes 4,6), respectively, of the human p26-Bcl-2 protein. In some cases (lanes 2, 5, 6), 10 fig of the appropriate peptide was added to Sf9 cell lysates to compete with antibody. Immunoprecipitates were analyzed by SDSPAGE/immunoblotting using anti-Bcl-2 peptide (61-76) antiserum to detect the p26-Bat-Bcl-2 protein as described above. The RIPA extraction buffer used here was shown to extract essentially all the immunodetectable Bcl-2 protein from cells, based on comparisons with cells lysed in Laemmli buffer [(17), and data not shown].

tected by Coomassie staining of gels containing lysates from AcMNPV-BCL-2-infected Sf9 cells (not shown). That the 26-kDa protein detected by immunoblot analysis of Sf9 cells infected with AcMNPV-BCL-2 was the authentic Bcl-2 protein was demonstrated by immunoprecipitation assays using three different polyclonal antisera (Fig. 1B). These antisera were raised against synthetic peptides corresponding to amino acids l-13,41-54, and 61-76 of the human Bcl-2 protein, and all of them were capable of immunoprecipitating from AcMNPV-BCL-2-infected Sf9 cells a 26-kDa protein that reacted in immunoblots with the antisera raised against the 61-76 peptide. As an additional control for immunospecificity, the corresponding peptides also were added to Sf9 lysates in some cases, thus competing with the 26-kDa Bat-Bcl-2 protein for antibody binding and almost completely blocking immunoprecipitation of the recombinant protein (Fig. 1B). These data immunologically confirm that the AcMNPV-BCL-2 virus leads to production of the desired recombinant protein. Because no biochemical activity has yet been ascribed to p26-Bcl-2, however, we were unable to functionally evaluate this baculo-produced Bcl-2 protein. Development and application of a dot-immunobinding assay that utilizes crude lysates from Sf9 cells infected with recombinant baculoviruses. To test the feasibility of using crude lysates from Sf9 cells infected with recombinant baculoviruses for MAb screening, we first examined the immunoreactivity of polyclonal rabbit antisera raised against synthetic peptides in a DIA that employed nitrocellulose filters to which S&l lysates had been adsorbed. The antisera used for these preliminary investigations had been raised in rabbits using synthetic

peptides that corresponded to sequences within the human Bcl-2 and v-Raf oncoproteins and that were conjugated to the carrier protein KLH. For these DIAs, which were modeled after previous reports (6,8), sheets of nitrocellulose were incubated with crude lysates derived from Sf9 cells infected with either the AcMNPV-BCL-2 or AcMNPV-RAF viruses. In addition, BSA-peptide conjugates were adsorbed to nitrocellulose filters, as an alternative source of antigen. As shown in Fig. 2, when lysates derived from AcMNPV-BCL-2-infected Sf9 cells were used as the antigen, specific immunoreactivity was observed in that anti-Bcl-2 antiserum bound to these filters, but not anti-Raf serum and not normal rabbit sera. Conversely, when the antigen was lysates prepared from AcMNPVRAF-infected Sf9 cells, anti-Raf antisera bound but not anti-Bcl-2 or normal rabbit sera. As anticipated from previous reports (7), BSA-peptide conjugates proved to be poor reagents for these assays. When a BSA-Bcl-2 peptide (61-76) conjugate was applied to nitrocellulose filters, both anti-Bcl-2 and anti-Raf antisera bound. The specificity of this assay employing crude lysates from recombinant baculovirus-infected cells was further demonstrated by peptide competition experiments. Figure 2B, for example, shows that binding of two different anti-Bcl-2 peptide antisera to filters containing lysates from AcMNPV-BCL-2-infected Sf9 cells was completely blocked when 2 pg of specific competing peptide was included. In contrast, when BSA-Bcl-2 peptide (61-76) was used as the antigen, binding of antisera raised against this same 61-76 peptide was not competed by addition of excess free peptide. Furthermore, nonspecific immunoreactivity was found for anti-

74

REED

a-B&2(41) a-B&2(41) + pep cx-B&2(61) a-B&2(61) + pep cc-Raf

B

NRS FIG. 2. Dot-immunobinding assay using crude lysates from infected SKI cells. Sheets of nitrocellulose to which protein-containing lysates from Sf9 cells infected with AcMNPV-BCL-2 (Bat-Bcl-2) or AcMNPV-RAF (Bat-Raf) had been adsorbed were placed into a manifold apparatus and rabbit antisera applied for 2 h before washing and treatment with iz51-protein A. Filters were also prepared that contained the Bcl-2 peptide (61-76) conjugated to bovine serum albumin (BSA-[61-761). In (A), 0.1% (v/v) of antisera that was raised against either Bcl-2 peptide (61-76) or Raf peptide (315-326), or normal rabbit serum (NRS) was incubated with filters. In (B), antisera raised against Bcl-2 peptides 41-54 or 61-76 or against Raf peptide 315-326 were applied to filters at 0.1% (v/v) with or without preincubation for 0.5 h with 2 pg of appropriate competing peptide, as indicated.

ET AL.

specific peptide (61-76) or an irrelevant peptide (41-54) was added to aliquots of hybridoma supernatants prior to DIA using Bat-Bcl-2-containing filters. For both hybridomas, the specific peptide (61-76) completely blocked binding of the MAb, whereas the irrelevant peptide (41-54) had no effect in this DIA (Fig. 3). Demonstration that the 407 and 2E8 monoclonal antibodies specifically bind p26-Bat-Bcl-2. Culture supernatants from several positive subclones of the 4D7 and 2E8 hybridomas were tested by immunoblot analysis, comparing the immunoreactivity of MAbs on blots prepared by SDS-PAGE separation of proteins derived from Sf9 cells infected with either AcMNVP-BCL-2 or a negative control virus AcMNVP-LCK. Figure 4A, for example, shows the results obtained for the 4D7.6 and 2E8.1 subclones. Both of these MAbs specifically bound to a 26-kDa protein in Sf9 lysates derived from AcMNPV-BCL-2-infected cells, but not from cells infected with the control virus, AcMNPV-LCK. The ability of these MAbs to bind the Bat-Bcl-2 protein in a solid-phase (immunoblot) assay was anticipated, since a solid-phase approach (DIA) was used for hybridoma screening. Our original purpose in preparing MAbs to p26-Bat-Bcl-2, however, was to employ them for immunoaffinity purification of this recombinant baculoprotein. To test the capacity of the 4D7 and 2E8 1

sera raised against other peptides. All of these anti-peptide antisera were prepared using cysteine-containing peptides that were conjugated to maleimide-activated KLH. Thus, this nonspecific binding probably reflects the production of antibodies that recognize determinants involved in the crosslinking of peptide to KLH (7). We next employed this DIA to screen hybridomas that had been prepared from mice immunized with KLH-Bcl-2 peptide (61-76). For the initial screen, hybridoma supernatants were incubated only with nitrocellulose filters adsorbed with lysates from AcMNPVBCL-S-infected cells. Of the 942 hybridomas tested, 31 reacted positively in this primary screening assay. These 31 hybridomas were then rescreened using two types of filters that had been prepared by adsorption with lysates from Sf9 cells infected with either AcMNPV-BCL-2 or AcMNPV-RAF viruses. The latter type of filter served as a negative control to exclude hybridomas that fortuitiously produced antibodies against St9 or nonrecombinant baculovirus proteins. Of the 31 hybridomas positive in the primary screen, only 2 produced MAbs that displayed specific immunoreactivity with filters containing Bat-Bcl-2 protein. These 2 hybridomas (termed 4D7 and 2E8) were then subjected to a final tertiary screening assay, whereby competing

2 3 4

5 6

7

8

9101112

A B C D

FIG. 3. Hybridoma screening by DIA. Hydridoma supernatants (100 ~1) were applied to nitrocellulose filters that had been adsorbed with lysates prepared from AcMNPV-BCL-2-infected Sf9 cells using a 96-well manifold apparatus. For the figure at top, wells A6-D12 represent primary screening of hybridoma culture supernatants. Al, rabbit anti-Bcl-2 peptide (61-76) antiserum at 1% (v/v); A2,1% normal rabbit serum; A3, 1% anti-Raf antiserum; A4, 0.1% anti-Bcl-2 peptide (61-76) antiserum plus 2 fig of 61-76 peptide; and A5, 0.1% anti-Bcl-2 (61-76) antiserum. For the figure at bottom, supernatants from the 4D7 hybridoma (100 ~1) were applied to Bat-Bcl-2-containing nitrocellulose filters with or without 2 pg of competing Bcl-2 peptide 61-76 (pep61) or nonspecific Bcl-2 peptide 41-54 (pep41). Normal rabbit serum or rabbit antisera raised against Bcl-2 peptide 61-76 {anti-Bcl-2 (61)), or the Raf peptide were also applied to filters at 0.1% (v/v), with or without 2 pg of competing peptide as indicated. All filters were subsequently incubated with 1 ag/ml rabbit anti-mouse lz51-protein A. Blots were exIgG antibody, followed by 0.25 &i/ml posed to x-ray film for 1 h.

A STRATEGY

(4D7) B L

FOR

GENERATING

L - 80 k-Da - 50

- 33 p26 + Bac-Bcl-

ANTIBODIES

75

resulted in the production of MAbs capable of binding the p26-Bat-Bcl-2 recombinant protein in solution.

PW B

MONOCLONAL

-28

- 19

FIG. 4. Characterization of anti-Bat-Bcl-2 monoclonal antibodies. Culture supernatants derived from the 4D7 and 2E8 hybridomas were employed for immunoblot (A) and immunoprecipitation (B) analysis of the p26-Bat-Bcl-2 protein. In (A), 25 pg of protein prepared from Sf9 cells infected with either AcMNPV-BCL-2 or AcMNPV-LCK were subjected to SDS-PAGE (12% gels) and transferred to nitrocellulose and incubated with 4D7 or 2E8 MAbs followed by 1 fig/ml of rabbit anti-mouse IgG, in the case of 4D7, and ‘?-protein A. In (B), 50 pg of protein prepared from infected Sf9 cells with either AcMNPV-BCL-2 (lanes 1, 2, 4) or AcMNPV-LCK (lanes 3, 5) was subjected to immunoprecipitation in 500 ~1 of a Triton X-lOO-containing buffer (17) containing either 30 ~1 of 4D7 hybridoma supernatant (lanes l-3) or 2.5 ~1 of rabbit antiserum raised against Bcl-2 peptide 61-76 (lanes 4, 5). Competing peptide (10 pg) was added to some samples (lane 2) prior to addition of antibodies. Immune complexes were collected with fixed S. anre~s that had been coated with rabbit anti-mouse IgG for lanes l-3, washed, and analyzed by SDS PAGE/immunoblotting using rabbit anti-Bcl-2 antiserum as described previously (17).

MAbs to bind p26-Bat-Bcl-2 in solution, therefore, we next employed these MAbs in immunoprecipitation assays. Both the 4D7 (Fig. 4B) and the 2E8 (not shown) antibodies specifically immunoprecipitated a 26-kDa protein from AcMNPV-BCL-2-infected Sf9 cells that reacted with the anti-61-76 antiserum on immunoblots. No immunoreactivity was seen for immunoprecipitates prepared from AcMNPV-LCK-infected cells. Addition of competing peptide to the Sf9 lysates prior to immunoprecipitation completely blocked MAb binding, further demonstrating the specificity of these results (Fig. 4B). Thus, the approach taken here whereby crude Sf9 lysates were employed in a DIA for hybridoma screening

DISCUSSION The baculovirus expression system offers several advantages over bacterial approaches for the production of recombinant proteins, including the ability to express eukaryotic cDNAs without special attention to sequences required for efficient translation in prokaryotes, improved maintenance of protein solubility, ability to express full-length integral membrane proteins, high-yield production of secreted proteins, and greater likelihood of obtaining biologically active protein because of proper protein folding, processing, and posttranslational modification (1). Significant disadvantages of the baculovirus system, however, are the relatively long time (compared to E. coli) required to create recombinant baculoviruses carrying the gene of interest and the subsequent task of purifying the recombinant protein from infected Sf9 cells. Having a MAb that specifically binds the baculoproduced protein of interest can greatly simplify the purification (4,5). Unfortunately, recombinant proteins produced via the baculovirus system frequently are insufficiently abundant (~5% total protein) to be conveniently isolated for use as antigens, unlike the case with many bacterially expressed proteins. An alternative approach is to use synthetic peptides as immunogens, in an effort to produce MAbs capable of binding the recombinant baculoproduced protein. However, if the peptide is then subsequently employed during the screening of hybridomas, many of the MAbs obtained typically react with the peptide (or carrier protein and crosslinking group) but fail to recognize the protein of interest, presumably because portions of the peptide fail to assume conformations resembling the corresponding sequence within the context of the whole protein (7). Consequently, considerable effort may be wasted on the maintenance and characterization of hybridomas that secrete anti-peptide MAbs that are of no utility. We reasoned that if synthetic peptides could be used as immunogens and unpurified baculoproduced proteins employed for hybridoma screening, few anti-peptide MAbs would be obtained that failed to efficiently bind the recombinant protein. To test this strategy, we created recombinant baculoviruses expressing a cDNA from the human Bcl-2 proto-oncogene and produced hybridomas from mice immunized with a synthetic peptide corresponding to a hydrophilic, proline-rich sequence (amino acids 61-76) found within this oncoprotein. Crude lysates prepared from AcMNPV-BCL-2-infected Sf9 cells were then adsorbed to nitrocellulose and these filters employed in a DIA for screening hybridomas. Though alternative DIA methods have been described wherein nitrocellulose disks are placed into 96-well plates (6), we elected to adsorb Sf9 lysates to whole sheets of nitrocellulose and then apply MAb-containing

76

REED

hybridomas supernatants to these filters using a commercially available dot-blot manifold apparatus for convenience of subsequent washing and other processing steps. No other special equipment was required. Because the MAbs were in contact with the antigen-containing nitrocellulose filters for only 2 h, three rounds of screening (96 wells X 3 = 288 hybridomas) could be accomplished per day with only a single manifold apparatus. Since the growth of hybridomas in microtiter plates was variable and some reached the appropriate cell density faster than others, we were able to easily keep pace with the 942 wells that yielded viable hybridomas of the 1440 plated after cell fusions. Given the large numbers of hybridoma culture supernatants that must be tested, we omitted any assays using control antigens during the primary screening. Of the 942 hybridomas tested for immunoreactivity with Bat-Bcl-2-containing lysates on nitrocellulose filters, 31 (3.3%) were positive. These were then assayed for differential reactivity with Bat-Bcl-2-containing and control Sf9 lysates, yielding 2 hybridomas of the 942 total screened (0.2%) that displayed the appropriate specificity. The false-positive result obtained for 29 of the hybridomas presumably represents MAbs with fortuitous cross-reactivity with nonrecombinant baculovirus or Sf9 insect cell antigens. As a source of negative control antigen, we employed crude lysates prepared from another recombinant baculovirus (AcMNPVRAF), because the polyhedrin protein in Sf9 cells infected with wild-type AcMNPV typically accounts for ~50% of total cellular protein and thus creates a situation that less accurately resembles cells infected with recombinant baculovirus. Such control recombinant viruses can be readily obtained from other investigators, and thus does not represent a disadvantage of the approach described here. Preliminary experiments indicate that the MAbs obtained by this method can be applied for immunoaffinity purification of the p26-Bat-Bcl-2 recombinant protein (Fig. 4B and data not shown). Both of these MAbs also efficiently recognize the native human Bcl-2 protein derived from lymphoid cells in immunoblotting and immunoprecipitation assays (details to be published elsewhere), thus suggesting that this approach can lead to the generation of MAbs with multiple utilities. Admittedly, not all anti-peptide MAbs that recognize denatured recombinant baculoproteins in a solid-phase assay (DIA) may be capable of also binding the undenatured protein in solution. For this reason, it may be advisable to first test antisera generated in mice or rabbits against synthetic peptide immunogens for their ability to immunoprecipitate the protein of interest. Nevertheless, assuming that the results obtained for Bcl-2 are broadly applicable, the strategy described here permits one to start with a cDNA containing the com-

ET AL.

plete open reading frame of a protein of biological interest and, without any intervening protein purification steps, to arrive at multifunctional MAbs that can then be used to assist in the purification of recombinant baculoproduced proteins from infected insect cells. ACKNOWLEDGMENTS We thank D. Morrison for AcMNPV-RAF virus, D. Bishop and R. Compans for pAcYM1 plasmid, and Y. Kokai, J. Lambris, R. Kennett, and J. Mullins for helpful discussions. This work was supported in part by grants from the NIH (CA-47956, CA-54957) and the University Research Foundation and a Scholar Award from the Leukemia Society of American to J.C.R.

REFERENCES 1. Luckow, V. A., and Summers, M. D. (1988) &o/Technology 6, 47-55. 2. Kallen, R. G., Smith, J. E., Sheng, Z., and Tung, L. (1990) Biothem. Biophys. Res. Commun. 168,616-624. M., Nathan, M., Honegger, A., 3. Hsu, C.-Y. J., Mohammadi, Ullrich, A., Schlessinger, J., and Hurwitz, D. R. (1990) Cell Growth Di/f. 1, 191-200. 4. Morgan, D. O., Kaplan, J. M., Bishop, J. M., and Varmus, H. E. (1991) in Methods in Enzymology (Hunter, T., and Sefton, B. M., Eds.), Vol. 200, pp. 645-660, Academic Press, San Diego, CA. 5. Wang, N. P., Qian, Y., Chung, A. E., Lee, W.-H., Lee, E. Y.-H. P. (1990) Cell Growth II@. 1, 429-437. 6. Steinitz, M., Rosen, A., and Klein, G. (1991) J. Immunol. Methods 136, 119-123. 7. Lie, L. D., Jolivet, M., Audibert, F., Fernandez, A., Wickstrom, E., Chedid, L., and Schlesinger, D. H. (1989) Pept. Res. 2, 114-119. 8. Hawkes, R. (1986) in Methods in Enzymology (Langone, J. J., and Van Vunakis, H., Eds.), Vol. 121, pp. 484-499, Academic Press, San Diego, CA. D., Nunez, G., Milliman, C., Schreiber, R. D., and 9. Hockenberry, Korsmeyer, S. J. (1990) Nature 348, 334-336. 10. Tsujimoto, Y., and Croce, C. M. (1986) Proc. N&l. Acad. Sci. USA 83, 5214-5218. 11. Matsuura, Y., Possee, R. D., Overton, H. A., and Bishop, D. H. L. (1987) J. Gen. Viral. 68, 1233-1250. 12. Koga, Y., Caccia, N., Toyonaga, B., Spolski, R., Yanagi, Y., Yoshikai, Y., and Mak, T. W. (1986) Eur. J. Immunol. 16, 1643-1646. H. (1990) Current Protocols in Molecular Biol13. Piwnica-Worms, ogy (Ausbel, F., Brent, R., Kingston, R., Moore, D., Seidman, J., Smith, J., and Struhl, K., Eds.), Vol. 2, pp. 16.8.1-16.11.7, WileyInterscience, New York. 14. Goswami, B. B., Glazer, R. I. (1991) BioZ’echniques 10,626-630. 15. Summers, M. D., and Smith, G. E. (1987) A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Station Bulletin No. 1555. College Station, TX. 16. Morrison, D., Kaplan, K., Escobedo, J., Rapp, U., Roberts, T., and Williams, L. (1989) Cell 58, 649-657. 17. Reed, J. C., Meister, L., Tanaka, S., Cuddy, M., Yum, S., Geyer, C., and Pleasure, D. (1991) Cancer Res. 51, 6529-6538. 18. Partis, M., et al. (1983) J. Protein Chem. 2, 263-277. 19. Lane, R. D., Crissman, R. S., and Ginn, S. (1986) in Methods in Enzymology (Langone, J. J., and Van Vunakis, H., Eds.), Vol. 121, pp. 183-192, Academic Press, San Diego, CA. 20. Reed, J. C., Tanaka, S., and Cuddy, M. (1992) Cancer Res. 52, 2802-2805.