Journal of Immunological Methods, 147 (1992) 1-11 © 1992 Elsevier Science Publishers B.V. All rights reserved 0022-1759/92/$05.00
JIM 06181
Use of recombinant fusion proteins for generation and rapid characterization of monoclonal antibodies Application to the Kunitz domain of human f3 amyloid precursor protein D. Wunderlich, A. Lee, R.P. Fracasso, D.V. Mierz, R.M. Bayney and T.V. Ramabhadran Molecular Therapeutics, Inc. & Miles Research Center, 400 Morgan Lane, West Haven, CT 06516, USA (Received 1 February 1991, revised received 29 August 1991, accepted 4 September 1991)
Production of peptides by recombinant DNA techniques is an efficient alternative to chemical synthesis of peptides. Proteins and peptides produced by recombinant DNA methods in E. coli are routinely used as antigens for the production of antibodies. However, most small peptides are rapidly degraded within the E. coli cell, and therefore, must initially be expressed as components of larger, more stable fusion proteins. The peptide of interest must be cleaved from the fusion protein, and purified prior to immunization to eliminate epitopes contributed by the fusion partner. We have now established methods for the production and characterization of monoclonal antibodies using partially purified, uncleaved fusion proteins. We have also described a method for efficient production and detection of the fusion protein, an EIA for rapid differential screening of hybridoma supernatants, and a strategy for epitope mapping of the antibodies. These methods have been applied to the production and characterization of monoclonal antibodies specific for a 75-amino-acid internal segment of the Alzheimer amyloid precursor protein, and should be applicable to a wide variety of other peptides and proteins. Key words: Fusion protein; Monoclonal antibody; (3 Amyloid; Kunitz domain
Introduction Antibodies raised against defined segments of proteins and peptides are useful and highly specific reagents for a variety of cell biological inves-
Correspondence to: T.V. Ramabhadran, Laboratory of Molecular and Cellular Neuroscience, Rockefeller University, 1230 York Avenue, New York, NY 10021, USA. Abbreviations: AD, Alzheimer's disease; APP, amyloid plaque precursor protein; KID, Kunitz inhibitor domain; EIA. enzyme-linked immunoassay; PBS. phosphate-buffered saline; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; DTI, dithiothreitol.
tigations. Bacterial expression systems are wellsuited for the production of heterologous peptides required for immunization and assay, and provide an efficient alternative to chemical synthesis of peptides. However, high level expression of most small peptides by Escherichia coli usually requires expression of the peptide as a fusion with a larger bacterial carrier protein, in order to stabilize against proteolysis (Burnett, 1986; Goldberg and Goff, 1986). Presence of epitopes on the fusion partner complicates the task of producing immunological reagents specific for the peptide of interest. Use of the fusion protein as immunogen for antibody production can elicit antibodies
2
specific for epitopes present on the larger carrier protein, which may mask immunoreactivities to the peptide of interest. Production of peptidespecific monoclonal antibodies may necessitate the cleavage of the recombinant peptide from the carrier protein (Moks et aI., 1987; Sambrook et aI., 1989) for subsequent use as immunogen, or as antigen to test immunoreactivities of the resultant hybridoma supernatants. In addition, fusion proteins present new epitopes not found in either segment, notably the linear epitopes at the junctions of the fusions. In this report we describe methods using partially enriched fusion proteins for the production and characterization of monoclonal antibodies to a 75-amino-acid peptide derived from the human ~ amyloid plaque precursor protein (APP). A 40-42 amino acid peptide termed ~ I A4 is the major constituent of cerebral and vascular plaques associated with Alzheimer's disease. This peptide is expressed as an internal segment of a family of proteins known as APPs (see review by Muller-Hill and Beyreuther, 1990). Two members of this family of lengths 751 and 770 amino acids (APP75 I and APPno ) contain a 56-amino-acid domain homologous to the Kunitz family of protease inhibitors, termed here as KID 56 • APPno contains a stretch of 19 additional amino acids contiguous to the 56-amino-acid Kunitz domain insert (Kang et aI., 1987; Kitaguchi et aI., 1988; Ponte et al.. 1988; Tanzi et aI., 1988) resulting in KID 75 . The biological functions of APPs and the significance of isoform differences are not understood. Our goal in preparing these antibodies was to produce isoform-specific reagents for studying the levels of the individual APP isoforms in normal and disease states. To date, the analysis of this question has been restricted to RNA analysis (Tanzi et aI., 1988) for lack of appropriate antibodies. We have expressed the KID domains of human APPs as recombinant fusion proteins. As observed with many recombinant proteins (Schoner et aI., 1985; Wilkinson and Harrison, 1991), the fusion proteins segregated to the pellet fraction of bacterial cells. An enriched bacterial pellet was used as the immunogen to generate KID peptide-specific monoclonal antibodies. A differential EIA using the enriched bacterial pel-
let was developed for the rapid identification of antibodies reactive to the Kunitz domain. In addition, mutant forms of the' fusion protein were used to epitope map the anti-KID antibodies. These methods should be generally applicable to the production of monoclonal antibodies against other peptides.
Materials and methods Materials Rabbit anti-recA antibody was a kind gift from Dr. C. Radding, Yale University, New Haven. CT. The mouse monoclonal antibody 22Cll, directed against the N-terminal extracellular domain of APPs was obtained from Dr. K. Beyreuther, University of Heidelberg, Heidelberg, Germany. Methods Expression vectors. DNA manipulations were performed by standard methods described in Sambrook et al. (1989). pMB58 vector, containing Hinfl-Sst II fragment spanning the recA gene of E. coli (Horii et aI., 1980; Sancar et aI., 1980), was obtained from Dr. M. Bittner, Amoco Research Center, Napierville, IL. The BamHI to Ncol segment from pMB58 containing the recA promoter and coding information for the first 36 amino acids of the recA protein were cloned between the BamHI and Ncol sites of pKK233-2 (Pharmacia LKB Biotechnology, New Jersey) replacing the trc promoter present in pKK233-2. The K1Ds6 and the KID7s coding regions were synthesized using polymerase chain reaction (Ehrlich, 1989) from full-length APP770 cDNA. with primers containing the sequences of the 5' and 3' regions of the Kunitz domains. The 5' primer also contained the Nco I restriction site required for the fusion with the Nco I site within the coding region of the recA gene (Horii et. ai., 1980; Sancar et. aI., 1980), and the 3' primers contained a T AA stop codon and a HindIII restriction site. The coding regions were inserted between the Ncol and HindIII sites of the recApKK233 hybrid (described above) yielding vectors pMTI-3534 and pMTI-3529 shown in Fig. 1.
membrane 1
1\4 1
- - - - - -- - - - - x - - - - - - - --!\ili '""ct "',·
1
"ylop l:"1ll
6'.15
751
770
19
KID pMTI-3 534
56 B
pMTI -3529 19
ASN-GLY
pMTI -3538
Fig. 1. Vectors used for the expression of KID domain fusions in E.coll. The organization and topology of the KID containing APP proteins (top drawing). Expression vectors (bottom drawing): pMTI-3534. recA-KlDs6 fusion ; pMTI-3529. recA-KID 75 fusion ; pMTI-3538. mutant recA-KID fusion with the insertion of asparagine and glycine at the fusion junction. Prec denotes the recA promoter and recA denotes the first 36 amino acids of the recA protein. Amp' represents the gene for resistance to ampicillin. Restriction endonuclease sites: B. Bam HI; H. Hindl/l; N. Nco I; S. Sail.
For epitope mapping, codons for Asn and Gly were introduced at the junction of the recA leader and the KID sequences by site-directed mutagenesis as described by Kunkel et al. (1987) resulting in pMTI-3538 (Fig. 1). The BamHI-HindIII fragment in pMTI-3529 containing the fusion gene was cloned into the phagemid vector Bluescript KS (Stratagene, La Jolla, CA). Single stranded DNA was produced in E. coli CJ236 and used for site directed mutagenesis. Sequences of the PCR products and mutants were verified by DNA sequencing (Sambrook et aI., 1989). Expression, detection, and enrichment of the fusion proteins. Vectors pMTls-3529, -3534 and -3538 were introduced into E. coli JMI05. Transformed strains were grown at 37°C in Luria broth containing 100 J,Lg/ml of ampicillin. Cultures were induced with 50 J,Lg/ml of nalidixic acid, an inducer of the recA promoter (Sigma Chemical Co., SI. Louis, MO; freshly prepared stock solu-
tion 10 mg/ml in 0.1 N NaOH), during exponential growth. After 3 h incubation at 37°C, cells were collected by centrifugation and pellets were dissolved in SDS-PAGE gel sample buffer or processed for the enrichment of fusion proteins. Cell fractions enriched in the fusion proteins were obtained by detergent lysis of cells and repeated Triton X-lOO washes (Tamburini et aI., 1990). Resulting pellets were solubilized in 8 M urea and dialyzed against phosphate-buffered saline (PBS). The dialysate was emulsified in Freund's adjuvant and used to immunize mice for the production of antibodies. Hybridoma methodology and EIA. Somatic cell hybrids were prepared using the method of Herzenberg et al. (1978) with some modifications (Lerner et aI., 1980). Cell lines secreting antibodies of interest were cloned at least twice by the limiting dilution method using culture medium supplemented with 5 U jml human recombinant
4
IL-6 (Genzyme, Boston, MA). Culture fluids from growing hybridomas were tested for the presence of specific antibody using EIA (Harlow and Lane, 1988). Bacterial pellets enriched for the fusion protein were dissolved in PBS, 1% SDS, 10 mM Drr and diluted in PBS for EIA. Extracts containing IJLg of enriched fusion proteins were allowed to adsorb to each well of Immulon II EIA plates (Dynatech, Chantilly, VA). After blocking of nonspecific protein binding sites, wells were sequentially incubated with hybridoma culture supernatant and peroxidase-labeled affinity-purified goat anti-mouse IgG (Kirkegaard and Perry, Gaithersburg, MD). Bound peroxidase-labeled second antibody was detected using the peroxidase substrate tetramethylbenzidine (TMB, Kirkegaard and Perry, Gaithersburg, MD). Isotypes of positive hybrid culture fluids were determined using EIA in which anti-mouse Fab was adsorbed to each well of EIA plates followed by sequential incubations with culture supernatants and peroxidase-labeled antiserum specific for mouse IgG 1, IgG2a, IgG2b, IgG3, and IgM. Epitope mapping. The junction mutants of KID fusions were constructed and expressed as described under "Expression Vectors". The holoAPP molecules used in the characterization of the antibodies were obtained from cell culture supernatants of established cell lines or mouse C127 cells transfected with bovine papilloma virus vectors (Howley et aI., 1983) containing the cDNAS of individual APP isoforms (Kang et aI., 1987; Kitaguchi et aI., 1988; Ponte et aI., 1988; Tanzi et aI., 1988). Supernatants were concentrated by precipitation with an equal volume of 25% trichloroacetic acid. Precipitates were washed twice with cold acetone, dried, and dissolved in gel sample buffer for SDS-PAGE and immunoblot analysis (Sambrook et aI., 1989). Results and discussion
We have developed methods for the production and rapid characterization of monoclonal antibodies to small peptides using the 75-aminoacid Kunitz domain, KID 7s , of the 770 isoform of human APP as the model peptide. Fig. 1 (top drawing) shows the schematic structure of the
APP isoforms and the location of the KID,s region used in our studies. The N-terminal 56 amino acids of KID,s domain are common to APP'51 and APP770 and show a high degree of homology to the Kunitz family of protease inhibitors (Kitaguchi et aI., 1988; Ponte et aI., 1988; Tanzi et aI., 1988).
Expression and detection of Kunitz domain fusions Small peptides expressed in E. coli in free fonn are rapidly degraded. Expression of such peptides as parts of larger fusion proteins usually leads to their stable accumulation. Therefore, the 56 and the 75 amino acid domains of APP were expressed in E. coli as fusion proteins with the first 36 amino acids of E. coli recA protein (Horii et aI., 1980; Sancar et aI., 1980; Krivi et aI., 1985). DNA segments encoding the KIDS6 and the KlD'5 domains were obtained from the cDNA for the precursors by polymerase chain reaction (peR) (Ehrlich, 1989). The PCR products were subcloned and the nucleotide sequence was verified (Sambrook et aI., 1989). They were introduced into Nco I-HindIII digested E. coli expression vector to produce vectors pMTI-3534 and pMTI-3529 containing respectively the genes for KlDs6 and KID,s inserts. These expression vectors are shown schematically in Fig. 1 (bottom drawing). The expected molecular weight of the recA-KlD,s fusion produced by the expression vector is 12.3 kDa. A mutant form of the recA-KIDs6 used in the epitope mapping of the antibodies was constructed by in vitro mutagenesis. A sequence encoding the dipeptide Asn-Gly was introduced between the Met residue 36 of the recA protein and Glu residue at the amino tenninus of the KID to produce pMTI-3538, also shown in Fig. 1 (bottom). This insertion disrupts the linear epitope at the junction of the recA and the KID segments. Exponentially growing cultures of transformed E. coli JM105 were induced by the addition of nalidixic acid, an inducer of the recA gene transcription. Because no antibodies to the KID domain were available, expression of the fusion protein was detected on immunoblots of SDSPAGE using a polyclonal rabbit antiserum which reacted with epitopes in the first 36 amino acids of the recA protein. As shown in Fig. 2A, pro-
5
teins of sizes expected for the fusion proteins in the 12-14 kDa range were detected on immunoblots. These protein bands were also visible by staining of the gels with Coomassie blue stain (not shown). The level of expression of the recAK1D75 fusions was 5-1O-fold higher than that of the recA-K1D56 fusion .The difference in the levels between the two fusion proteins may be due to the relative instability of the smaller, recAK1D5h fusion protein over the recA-KID75 fusion. SmaIIer polypeptides in this size range are rapidly degraded within E. coli cells (Goldberg and Goff, 1986) and the longer proteins are generaIIy more stable. Results in Fig. 2 show that the recA
+
A L
CO Ll1 CD
2:
Cl.
L (lJ
.:!J:.
0
L
+
+
(.!) Lf)
Ll1
U (lJ L
U (lJ L
JIM 06181
Use of recombinant fusion proteins for generation and rapid characterization of monoclonal antibodies Application to the Kunitz domain of human f3 amyloid precursor protein D. Wunderlich, A. Lee, R.P. Fracasso, D.V. Mierz, R.M. Bayney and T.V. Ramabhadran Molecular Therapeutics, Inc. & Miles Research Center, 400 Morgan Lane, West Haven, CT 06516, USA (Received 1 February 1991, revised received 29 August 1991, accepted 4 September 1991)
Production of peptides by recombinant DNA techniques is an efficient alternative to chemical synthesis of peptides. Proteins and peptides produced by recombinant DNA methods in E. coli are routinely used as antigens for the production of antibodies. However, most small peptides are rapidly degraded within the E. coli cell, and therefore, must initially be expressed as components of larger, more stable fusion proteins. The peptide of interest must be cleaved from the fusion protein, and purified prior to immunization to eliminate epitopes contributed by the fusion partner. We have now established methods for the production and characterization of monoclonal antibodies using partially purified, uncleaved fusion proteins. We have also described a method for efficient production and detection of the fusion protein, an EIA for rapid differential screening of hybridoma supernatants, and a strategy for epitope mapping of the antibodies. These methods have been applied to the production and characterization of monoclonal antibodies specific for a 75-amino-acid internal segment of the Alzheimer amyloid precursor protein, and should be applicable to a wide variety of other peptides and proteins. Key words: Fusion protein; Monoclonal antibody; (3 Amyloid; Kunitz domain
Introduction Antibodies raised against defined segments of proteins and peptides are useful and highly specific reagents for a variety of cell biological inves-
Correspondence to: T.V. Ramabhadran, Laboratory of Molecular and Cellular Neuroscience, Rockefeller University, 1230 York Avenue, New York, NY 10021, USA. Abbreviations: AD, Alzheimer's disease; APP, amyloid plaque precursor protein; KID, Kunitz inhibitor domain; EIA. enzyme-linked immunoassay; PBS. phosphate-buffered saline; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; DTI, dithiothreitol.
tigations. Bacterial expression systems are wellsuited for the production of heterologous peptides required for immunization and assay, and provide an efficient alternative to chemical synthesis of peptides. However, high level expression of most small peptides by Escherichia coli usually requires expression of the peptide as a fusion with a larger bacterial carrier protein, in order to stabilize against proteolysis (Burnett, 1986; Goldberg and Goff, 1986). Presence of epitopes on the fusion partner complicates the task of producing immunological reagents specific for the peptide of interest. Use of the fusion protein as immunogen for antibody production can elicit antibodies
2
specific for epitopes present on the larger carrier protein, which may mask immunoreactivities to the peptide of interest. Production of peptidespecific monoclonal antibodies may necessitate the cleavage of the recombinant peptide from the carrier protein (Moks et aI., 1987; Sambrook et aI., 1989) for subsequent use as immunogen, or as antigen to test immunoreactivities of the resultant hybridoma supernatants. In addition, fusion proteins present new epitopes not found in either segment, notably the linear epitopes at the junctions of the fusions. In this report we describe methods using partially enriched fusion proteins for the production and characterization of monoclonal antibodies to a 75-amino-acid peptide derived from the human ~ amyloid plaque precursor protein (APP). A 40-42 amino acid peptide termed ~ I A4 is the major constituent of cerebral and vascular plaques associated with Alzheimer's disease. This peptide is expressed as an internal segment of a family of proteins known as APPs (see review by Muller-Hill and Beyreuther, 1990). Two members of this family of lengths 751 and 770 amino acids (APP75 I and APPno ) contain a 56-amino-acid domain homologous to the Kunitz family of protease inhibitors, termed here as KID 56 • APPno contains a stretch of 19 additional amino acids contiguous to the 56-amino-acid Kunitz domain insert (Kang et aI., 1987; Kitaguchi et aI., 1988; Ponte et al.. 1988; Tanzi et aI., 1988) resulting in KID 75 . The biological functions of APPs and the significance of isoform differences are not understood. Our goal in preparing these antibodies was to produce isoform-specific reagents for studying the levels of the individual APP isoforms in normal and disease states. To date, the analysis of this question has been restricted to RNA analysis (Tanzi et aI., 1988) for lack of appropriate antibodies. We have expressed the KID domains of human APPs as recombinant fusion proteins. As observed with many recombinant proteins (Schoner et aI., 1985; Wilkinson and Harrison, 1991), the fusion proteins segregated to the pellet fraction of bacterial cells. An enriched bacterial pellet was used as the immunogen to generate KID peptide-specific monoclonal antibodies. A differential EIA using the enriched bacterial pel-
let was developed for the rapid identification of antibodies reactive to the Kunitz domain. In addition, mutant forms of the' fusion protein were used to epitope map the anti-KID antibodies. These methods should be generally applicable to the production of monoclonal antibodies against other peptides.
Materials and methods Materials Rabbit anti-recA antibody was a kind gift from Dr. C. Radding, Yale University, New Haven. CT. The mouse monoclonal antibody 22Cll, directed against the N-terminal extracellular domain of APPs was obtained from Dr. K. Beyreuther, University of Heidelberg, Heidelberg, Germany. Methods Expression vectors. DNA manipulations were performed by standard methods described in Sambrook et al. (1989). pMB58 vector, containing Hinfl-Sst II fragment spanning the recA gene of E. coli (Horii et aI., 1980; Sancar et aI., 1980), was obtained from Dr. M. Bittner, Amoco Research Center, Napierville, IL. The BamHI to Ncol segment from pMB58 containing the recA promoter and coding information for the first 36 amino acids of the recA protein were cloned between the BamHI and Ncol sites of pKK233-2 (Pharmacia LKB Biotechnology, New Jersey) replacing the trc promoter present in pKK233-2. The K1Ds6 and the KID7s coding regions were synthesized using polymerase chain reaction (Ehrlich, 1989) from full-length APP770 cDNA. with primers containing the sequences of the 5' and 3' regions of the Kunitz domains. The 5' primer also contained the Nco I restriction site required for the fusion with the Nco I site within the coding region of the recA gene (Horii et. ai., 1980; Sancar et. aI., 1980), and the 3' primers contained a T AA stop codon and a HindIII restriction site. The coding regions were inserted between the Ncol and HindIII sites of the recApKK233 hybrid (described above) yielding vectors pMTI-3534 and pMTI-3529 shown in Fig. 1.
membrane 1
1\4 1
- - - - - -- - - - - x - - - - - - - --!\ili '""ct "',·
1
"ylop l:"1ll
6'.15
751
770
19
KID pMTI-3 534
56 B
pMTI -3529 19
ASN-GLY
pMTI -3538
Fig. 1. Vectors used for the expression of KID domain fusions in E.coll. The organization and topology of the KID containing APP proteins (top drawing). Expression vectors (bottom drawing): pMTI-3534. recA-KlDs6 fusion ; pMTI-3529. recA-KID 75 fusion ; pMTI-3538. mutant recA-KID fusion with the insertion of asparagine and glycine at the fusion junction. Prec denotes the recA promoter and recA denotes the first 36 amino acids of the recA protein. Amp' represents the gene for resistance to ampicillin. Restriction endonuclease sites: B. Bam HI; H. Hindl/l; N. Nco I; S. Sail.
For epitope mapping, codons for Asn and Gly were introduced at the junction of the recA leader and the KID sequences by site-directed mutagenesis as described by Kunkel et al. (1987) resulting in pMTI-3538 (Fig. 1). The BamHI-HindIII fragment in pMTI-3529 containing the fusion gene was cloned into the phagemid vector Bluescript KS (Stratagene, La Jolla, CA). Single stranded DNA was produced in E. coli CJ236 and used for site directed mutagenesis. Sequences of the PCR products and mutants were verified by DNA sequencing (Sambrook et aI., 1989). Expression, detection, and enrichment of the fusion proteins. Vectors pMTls-3529, -3534 and -3538 were introduced into E. coli JMI05. Transformed strains were grown at 37°C in Luria broth containing 100 J,Lg/ml of ampicillin. Cultures were induced with 50 J,Lg/ml of nalidixic acid, an inducer of the recA promoter (Sigma Chemical Co., SI. Louis, MO; freshly prepared stock solu-
tion 10 mg/ml in 0.1 N NaOH), during exponential growth. After 3 h incubation at 37°C, cells were collected by centrifugation and pellets were dissolved in SDS-PAGE gel sample buffer or processed for the enrichment of fusion proteins. Cell fractions enriched in the fusion proteins were obtained by detergent lysis of cells and repeated Triton X-lOO washes (Tamburini et aI., 1990). Resulting pellets were solubilized in 8 M urea and dialyzed against phosphate-buffered saline (PBS). The dialysate was emulsified in Freund's adjuvant and used to immunize mice for the production of antibodies. Hybridoma methodology and EIA. Somatic cell hybrids were prepared using the method of Herzenberg et al. (1978) with some modifications (Lerner et aI., 1980). Cell lines secreting antibodies of interest were cloned at least twice by the limiting dilution method using culture medium supplemented with 5 U jml human recombinant
4
IL-6 (Genzyme, Boston, MA). Culture fluids from growing hybridomas were tested for the presence of specific antibody using EIA (Harlow and Lane, 1988). Bacterial pellets enriched for the fusion protein were dissolved in PBS, 1% SDS, 10 mM Drr and diluted in PBS for EIA. Extracts containing IJLg of enriched fusion proteins were allowed to adsorb to each well of Immulon II EIA plates (Dynatech, Chantilly, VA). After blocking of nonspecific protein binding sites, wells were sequentially incubated with hybridoma culture supernatant and peroxidase-labeled affinity-purified goat anti-mouse IgG (Kirkegaard and Perry, Gaithersburg, MD). Bound peroxidase-labeled second antibody was detected using the peroxidase substrate tetramethylbenzidine (TMB, Kirkegaard and Perry, Gaithersburg, MD). Isotypes of positive hybrid culture fluids were determined using EIA in which anti-mouse Fab was adsorbed to each well of EIA plates followed by sequential incubations with culture supernatants and peroxidase-labeled antiserum specific for mouse IgG 1, IgG2a, IgG2b, IgG3, and IgM. Epitope mapping. The junction mutants of KID fusions were constructed and expressed as described under "Expression Vectors". The holoAPP molecules used in the characterization of the antibodies were obtained from cell culture supernatants of established cell lines or mouse C127 cells transfected with bovine papilloma virus vectors (Howley et aI., 1983) containing the cDNAS of individual APP isoforms (Kang et aI., 1987; Kitaguchi et aI., 1988; Ponte et aI., 1988; Tanzi et aI., 1988). Supernatants were concentrated by precipitation with an equal volume of 25% trichloroacetic acid. Precipitates were washed twice with cold acetone, dried, and dissolved in gel sample buffer for SDS-PAGE and immunoblot analysis (Sambrook et aI., 1989). Results and discussion
We have developed methods for the production and rapid characterization of monoclonal antibodies to small peptides using the 75-aminoacid Kunitz domain, KID 7s , of the 770 isoform of human APP as the model peptide. Fig. 1 (top drawing) shows the schematic structure of the
APP isoforms and the location of the KID,s region used in our studies. The N-terminal 56 amino acids of KID,s domain are common to APP'51 and APP770 and show a high degree of homology to the Kunitz family of protease inhibitors (Kitaguchi et aI., 1988; Ponte et aI., 1988; Tanzi et aI., 1988).
Expression and detection of Kunitz domain fusions Small peptides expressed in E. coli in free fonn are rapidly degraded. Expression of such peptides as parts of larger fusion proteins usually leads to their stable accumulation. Therefore, the 56 and the 75 amino acid domains of APP were expressed in E. coli as fusion proteins with the first 36 amino acids of E. coli recA protein (Horii et aI., 1980; Sancar et aI., 1980; Krivi et aI., 1985). DNA segments encoding the KIDS6 and the KlD'5 domains were obtained from the cDNA for the precursors by polymerase chain reaction (peR) (Ehrlich, 1989). The PCR products were subcloned and the nucleotide sequence was verified (Sambrook et aI., 1989). They were introduced into Nco I-HindIII digested E. coli expression vector to produce vectors pMTI-3534 and pMTI-3529 containing respectively the genes for KlDs6 and KID,s inserts. These expression vectors are shown schematically in Fig. 1 (bottom drawing). The expected molecular weight of the recA-KlD,s fusion produced by the expression vector is 12.3 kDa. A mutant form of the recA-KIDs6 used in the epitope mapping of the antibodies was constructed by in vitro mutagenesis. A sequence encoding the dipeptide Asn-Gly was introduced between the Met residue 36 of the recA protein and Glu residue at the amino tenninus of the KID to produce pMTI-3538, also shown in Fig. 1 (bottom). This insertion disrupts the linear epitope at the junction of the recA and the KID segments. Exponentially growing cultures of transformed E. coli JM105 were induced by the addition of nalidixic acid, an inducer of the recA gene transcription. Because no antibodies to the KID domain were available, expression of the fusion protein was detected on immunoblots of SDSPAGE using a polyclonal rabbit antiserum which reacted with epitopes in the first 36 amino acids of the recA protein. As shown in Fig. 2A, pro-
5
teins of sizes expected for the fusion proteins in the 12-14 kDa range were detected on immunoblots. These protein bands were also visible by staining of the gels with Coomassie blue stain (not shown). The level of expression of the recAK1D75 fusions was 5-1O-fold higher than that of the recA-K1D56 fusion .The difference in the levels between the two fusion proteins may be due to the relative instability of the smaller, recAK1D5h fusion protein over the recA-KID75 fusion. SmaIIer polypeptides in this size range are rapidly degraded within E. coli cells (Goldberg and Goff, 1986) and the longer proteins are generaIIy more stable. Results in Fig. 2 show that the recA
+
A L
CO Ll1 CD
2:
Cl.
L (lJ
.:!J:.
0
L
+
+
(.!) Lf)
Ll1
U (lJ L
U (lJ L
