- Email: [email protected]
Immunological screening of standard cDNA libraries in pBR322 vectors: Detection of human fibrinogen and prothrombin cDNA clones
ANALYTICAL
BIOCHEMISTRY
142,271-276
(1984)
Immunological Screening of Standard cDNA Libraries in pBR322 Vectors: Detection of Human Fibrinogen and Prothrombin cDNA Clones HENRI PLAISAN&,* YVONNE ALEXANDRE,* GEORGES UZAN,~ CLAUDE BESMOND,* RICHARD BENAROUS,$MONIQUE FRAIN,~ Jose SALA TREPAT,~ JEAN-CLAUDE DREYFLJS,*AND AXEL KAHN*” *INSERM U.129, Institut de Pathologie Molkdaire, CHU Cochin, 75674 Paris Cedex 14, TINSERM lJ.143, Institut de Pathologie Cellulaire, Hopital de BicZtre, 94270 Le Kremlin Bi&tre, $INSERM U.15, Institut de Pathologie Molkculaire, CHU Cochin. 75474 Paris Cedex 14, and $Laboratoire d’Enzymologie du CNRS. 91190 Gif sur Yvette, France
Received February 6, 1984 The in situ immunological detection of antigens encoded by cDNA inserted into the PstI site of pBR322 plasmids was optimized. It was found that sensitivity of the detection was dramatically increased by in situ amplification of the recombinant plasmids on chloramphenicolcontaining medium followed by a brief incubation without chloramphenicol during which protein synthesis resumes. In addition, several modifications of the previously described methods which permit total suppression of background and false positives are described. These techniques allowed easy detection of cDNA clones for human B 6- and -y-fibrinogen and -prothrombin using a human liver double-stranded cDNA recombinant plasmid library in pBR322 vectors. 0 1984 Academic Press, Inc. KEY WORDS: cDNA libraries; pBR322 plasmids; immunological screening; expression vectors; human fibrinogen; human prothrombin.
Detection of cDNA recombinant clones by immunoassay of the antigen synthesized by the transformed bacteria colonies is a very powerful technique: provided that specific antibodies are available, it allows direct screening of a large number of clones for mRNA species which are expressed at a low level and whose sequence is unknown. Probably one of the best methods used today is to build a special expression library in the lysogen phage expression vector Xgt 11 (1). Most of the earlier cDNA libraries, however, were built using the standard technique of cloning poly(dC)-tailed cDNA into the poly(dG)-tailed PstI site of pBR322 plasmid (2-5). Kemp and Cowman described a simple in situ immunoassay for screening Escherichia coii colonies transformed by recombinant pBR322 (6). This method was subsequently I To whom reprint requests should be addressed.
used for detecting cDNA clones for rat dopamine P-hydroxylase (7) and ferritin (8). In this paper we show that some modifications of the previously published technique greatly increase its sensitivity, with suppression of background and nonspecific colonies. The main improvement is achieved by in situ amplification of the plasmid by chloramphenicol, followed by a brief reincubation without chloramphenicol to permit synthesis of new proteins. This technique should probably enable already existing cDNA libraries to be used for detecting any clone corresponding to products for which specific antisera are available. MATERIAL
AND
METHODS
cDNA library and antibodies. A human
liver ds* cDNA recombinant
plasmid library
* Abbreviations used: ds, double-stranded: CNBr, cyanogen bromide. 271
0003-2697184 $3.00 Copyright 0 1984 by Academic Press, Inc. All r&s of reproduction in any form reserved.
272
PLAISANCIE
in the pBR322 vector was constructed as reported elsewhere (9,lO) using the dC-tailing method. cDNA was inserted into the PstI site. Anti-human fibrinogen (11) and prothrombin (12) antisera were prepared by immunizing rabbits against denatured proteins; the antibodies were then purified by immunoaffinity chromatography. Anti-glucose phosphate isomerase and phosphofiuctokinase-purified antibodies used as controls were prepared against native enzymes (13). Fibrinogen cDNA clones. Cloning of several cDNA inserts for the three fibrinogen chains has been reported elsewhere (14). The inserts were all identified by positive hybridization selection and by verifying that their restriction map fitted well with the known nucleotide sequence of the different fibrinogen mRNAs ( 1518). Other previously identified cDNA clones, specifying aldolase B and transferrin (9,10,19), were used as controls. All the tests aimed at improving the immunological screening methods were performed using replicas from selected, ordered clones in a 96well microtitration plate. The plate contained three clones for the A a-fibrinogen chain, five for B p-, three for y-, two for aldolase B, and two for transfer-tin. In situ ampliJcation of plasmid DNA and restarting of protein synthesis. Colonies were replicated onto a nitrocellulose filter which was then placed on an L-broth agar plate containing tetracycline (10 pg/ml). After colony growth (overnight incubation at 37°C) the filter was peeled from the agar plate and ;.:. transferred onto a similar medium containing 10 &ml chloramphenicol in addition to 10 hg/ml tetracycline. After a further 12- 18 h incubation at 37”C, the filter was peeled again from the chloramphenicol medium and transferred onto a third medium identical to the first one for a 4-h period at 37°C. In some preliminary tests, the amplification step was omitted, or the reincubation without chloramphenicol after plasmid amplification was either omitted or stopped after 1, 2, 3, 4, or 12 h. The colony sizes ranged from 1
ET AL.
to 3 mm. Comparisons of the signal obtained with or without in situ amplification plasmid DNA were performed using colonies of similar sizes. Immobilization of lysed E. coli protein on CNBr-activated Whatman 540 jilters. This step was performed essentially as indicated by Kemp and Cowman (6) except that 1 mM diisopropylphosphofluoridate was added to the transfer buffer. The colonies grown on the nitrocellulose filters were lysed by contact (colony side up) with blotting paper saturated with 0.1 M NaHCOJl% (v/v) Triton X-100/ 2 mg of lysozyme per milliliter, for 20 min in an atmosphere saturated with CHCl, vapor. After an additional 20 min of contact with paper saturated with “binding buffer” (0.1 M NaHCOs, 0.1% Triton X-100), the lysed colonies on nitrocellulose filters were blotted on CNBr-activated Whatman 540 filters for 4-6 h, in a system similar to that described by Southern for DNA blotting (20) using “the binding buffer.” The free reactive groups of the CNBr filter were blocked by incubation for 3 h at room temperature in 2% (w/v) glycine, 1% bovine serum albumin in wash buffer 1 [lo mM Tris-HCl (pH 7.8) 1 M NaCl, 1% Triton X-100, 1% (v/v) Nonidet-P40, and 1% (w/v) sodium deoxycholate]. The filter was then treated for 1 h at room temperature by 2 pg/ ml DNase 1, 10 pg/ml RNase in a 10 mM Tris-HCl buffer (pH 7.4), 5 mM MgClz, 100 mM NaCl, and then washed again with wash buffer 1. All the buffers used were extemporarily filtered through 0.42~pm Millipore membranes. Preparation of pBR322-transformed E. coli protein whole extract. E. coli (C 600 or MC 106 1 strains) transformed by nonrecombinant pBR322 plasmids were grown in liquid medium up to an absorbance of 1.60 OD at 600 nm and then collected by centrifugation and washed in 10 mM Tris-HCl, 150 mM NaCl, and 50 mM sucrose. The bacterial cells were lysed by adding lysozyme (2 mg/ml) and Triton X- 100 (I%, v/v). Nucleic acids were precipitated by the progressive addition
IMMUNOLOGICAL
SCREENING
OF cDNA LIBRARIES
IN pBR322 VECTORS
273
of a 50 mg/ml solution of protamine sulfate and then eliminated by centrifugation. The supernatant was adjusted to 1 mrvt diisopropylphosphofluoridate, 10 mM +aminocaproic acid, 0.1 mM pepstatin, and 1% (v/v) aprotinin as antiproteolytic agents (13), and then stored frozen at -80°C.
Detection of antigens in immobilized colony lysates.The filter was incubated under continuous gentle agitation for 4 h at room temperature with diluted antiserum (1 to 5 pg/ml of purified antibodies) in about 2 ml/ filter of a 10 mM Tris-HCl buffer (pH 7.8), 150 mM NaCl, 1% (w/v) glycin, 1% (w/v) bovine serum albumin, Nonidet-P40, Triton X- 100 and sodium deoxycholate (1% each), E. coli and 2 mg/ml pBR322-transformed proteins. The filter was then washed several times with 20 ml/filter of, successively, wash buffer 1 (2-3 times over 1 h), wash buffer 1 without detergent (2-3 times over 1 h), and finally with an excess of 10 BIM Tris-HCl buffer (pH 7.8) plus 10 mM NaCI. This last wash usually lasted overnight. The filter was then treated for 2 h with ‘251-labeled protein A in the same buffer as that used for the antigen-antibody reaction, and then washed again exactly as indicated after antibody fixation. As mentioned above, all the buffers were filtered on Millipore membranes. In particular, antibody and protein A solutions could be used several times without apparent decrease of sensitivity, but should be filtered again each time. Autoradiography was performed at -80°C for from 15 h to several days using intensifying screens. RESULTS AND DISCUSSION
Three out of the fifteen fibrinogen clones studied were easily detected by the direct immunoassay screening reported here, without any background or false positives (Fig. 1). These clones were not detected when either nonimmune IgG or anti-prothrombin, anti-glucose phosphate isomerase, and antiphosphofructokinase antibodies were used
without
with amplification qk
1
2
FIG. 1. lmmunodetection of fibrinogen clones with or without amplification of recombinant plasmids by chloramphenicol and influence of E. cob protein extracts and glycine on the background. (I) After amplification, the colonies were deposited on a new medium without chloramphenicol for 8 h. The filters with “amplified” or “not amplified” plasmids were treated in a similar fashion, the autoradiograms being exposed for the same times (14 h at -80°C using intensifying screens). For both filters E. coli protein extract and glycine were added to the buffers to reduce background. (II, 1) Same conditions as in (I), with amplification but without addition of glycine. 14-h autoradiographic exposure. (II, 2) Idem, but without addition of E. coli protein extract and glycine. 6-h autoradiographic exposure. The diameters of the transferred colonies were about 3 mm in each of these experiments.
instead of anti-fibrinogen antiserum. Clones designated a and b contained 800- to lOOObp-long cDNA inserts specific for B P-fibrinogen chains while clone c contained a 1200bp-long, y-chain specific insert. All these inserts were oriented correctly for translation and, although they were not full length (B @-
274
PLAISANCIk
and y-mRNAs are about 1800-2 100 bases long (14-15)), they all corresponded to a large part of the coding region (15- 18). The low positivity of the y-chain clone (c) as compared to that of the B P-chain clones (a and b) could be explained either by a lower expression or by a lower reactivity of the antiserum against y than against B /3 chains. However, this last explanation is unlikely because the antibodies used, prepared against denatured fibrinogen, recognized cell-free synthesized y chains practically as well as A cx or B @chains (11). It is possible, however, that the portion of y chain synthetized under
ET AL.
direction of the insert does not contain the main antigenic epitopes of the proteins. Since we tested only 15 fibrinogen clones, we cannot determine if, as indicated by Kemp and Cowman (6), the proportion of detectable colonies is higher than the expected one in six sequences which would have a correct orientation and frame. As shown in Fig. 1, sensitivity of detection of the positive clones is dramatically increased by in situ amplification of recombinant plasmids in the transformed colonies. This is achieved using the classical chloramphenicol method (2 1); this antibiotic, however, blocks
FIG. 2. Immunodetection of fibrinogen clones on colonies lysed various times after amplification with chloramphenicol. After amplification, the colonies were lyscd either immediately or 1, 2, or 4 h after the colonies were transferred to agar plates without chloramphenicol. The diameter of the colonies did not significantly increase during the first hours of incubation without chloramphenicol. Their diameters ranged from 2.5 to 3.5 mm. All the filters were similarly treated and exposed for autoradiography.
IMMUNOLOGICAL
SCREENING
OF cDNA LIBRARIES
protein synthesis so that the antigen is undetectable immediately after amplification (Fig. 2). When the filter is put on a new medium without chloramphenicol, antigen synthesis starts again, is detectable after 1 h, reaches a maximum at the 4th-5th hour (Fig. 3), and then decreases slightly (not shown) while growth of the colonies resumes. The increase of the sensitivity of detection of the antigens is clearly not dependent upon growth of the colonies; the sizes of “amplified” and “nonamplified” colonies are indeed similar, and the increase in the size of “nonamplified” colonies results in an increase of the surface of the autoradiographic signal, not of its intensity. In addition, protracted growth without chloramphenicol after chloramphenicol amplification results, as indicated above, in a progressive decrease of the signal. Among the factors which contribute to decreasing background and false-positive colonies or spots, four proved essential in our tests: (i) protracted washes are needed especially to avoid the diffuse background which can occur after long autoradiographic expo-
IN pBR322 VECTORS
275
sure; (ii) filtering of buffers as well as antibody and protein A solutions permits elimination of the strong false-positive spots which can occur outside the colonies; and (iii) and (iv) adding both glycine and pBR322transformed E. coli total proteins to antibody and protein A solutions dramatically decreases nonspecific labeling of colonies (Fig. 1). In fact, negative colonies either are not seen even after several days of autoradiographic exposure (Fig. 3) or, in some cases can appear in negative on a slight background (Fig. 3). Using the method described in this paper, we have screened 7000 ordered clones from a human liver ds cDNA library with antiprothrombin antibodies. Three clones were so detected, each of them being firmly identified as containing a prothrombin cDNA insert by positive hybridization-selection and comparison of their restriction map with the known sequence of the mRNA (22-23). We recently found that this method is also fully convenient for screening nonordered colonies on high-density filters (unpublished results). In conclusion, this paper describes opti-
FIG. 3. Immunodetection of two human prothrombin clones among 192 ordered colonies. Seven thousand colonies ordered in 96-well microtitration plates were transferred to filters, the plasmids were amplified, and then the colonies were lysed and proteins were fixed to CNBr-activated Whatman 540 paper. This figure shows two Whatman filters on which positive colonies were detected with purified antibodies against sodium dodecyl sulfate-denatured human prothrombin. 36 h of autoradiographic exposure.
PLAISANCIE
276
mization of the immunological detection cDNA clones in usual pBR322 vectors. Using the methods described it is possible to obtain strong and specific signals, without any background or false positives. Investigators who possess large recombinant plasmid cDNA libraries may find these methods a valuable alternative to the construction of a new library in special expression vectors.
REFERENCES
ET AL. Tmpat, J., Marie, J., and Kahn, A. (1983) J. Biol. Chem. 258, 14576-14584. 10. Uzan, G., Frain, M., Park, I., Beamond, C., Maessen, G., Sala-Ttipat, J., Zakin, M. M., and Kahn, A. (1984) Biochem. Biophys. Res. Commun. 119, 273-281.
II. Uzan, G., Besmond, C., Kahn, A., and Marguerie, G. (1982) Biochem. Int. 4, 27 1-277. 12. Besmond, C., Benarous, R., Kahn, A. (198 1) Biochem. Biophys. Res. Commun. 103, 587-594. 13. Kahn, A., Cottreau, D., Daegelen, D., and Dreyfus, J. C. (1981) Eur. J. Biochem. 116, 7-12. 14. Uzan, G., Courtois, G., Besmond, C., Frain, M, &la-T&pat, J., Kahn, A., and Marguerie, G. (1984) B&hem. Biophys. Rex Commun. 120, 376-383.
1. Young, R. A., and Davis, R. A. (1983) Proc. Null. Acad. Sci. USA 80, 1194-l 198. 2. Kushner, S. R. (1978) in Genetic Engineering (Boyer, H. B., and Nicosia, S., eds.), pp. 17-28, Elsevier/ North-Holland, Amsterdam/New York. 3. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) in Molecular Cloning (Maniatis, T., ed.), pp. 3 12328, Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. 4. Rowekamp, W., and Firtel, R. A. (1980) Dev. Biol. 79,409-418.
5. Wickens, M. P., Buell, G. N., and Shimer, R. T. (1978) .I. Biol. Chem. 253,2483-2495. 6. Kemp, D. J., and Cowman, A. F. (1981) Proc. Nafl. Acad. Sci. USA 78,4520-4524. 7. O’Malley, K. L., Mauron, A., Riese, J., Barchas, J. D., and Kedes, L. (1983) Proc. Nat/. Acad. Sci. USA 80,2161-2165. 8. Brown, A. J. P., Leibold, E. A., and Munro, H. N. (1983) Proc. Natl. Acad. Sci. USA 80, 1265-1269. 9. Simon, M. P., Besmond, C., Cottreau, D., ChaumetRiffaud, P., Weber, A., Dreyfus, J. C., Sala-
15. Kant, J. A., Lord, S. T., and Crabtree, G. R. (1983) Proc. Natl. Acad. Sci. USA. 80, 3953-3957. 16. Rixon, M. W., Chan, W. Y., Davie, E. W., and Chung, D. W. (1983) Biochemistry 22, 32373244.
17. Chung, D. W., Que, B. G., Rixon, M. W., Mace, M., and Davie, E. W. (1983) Biochemisfry 22, 3244-3250.
18. Chung, D. W., Chan, W. Y., and Davie, E. W. (1983) Biochemistry 22, 3250-3256. 19. Besmond, C., Dreyfus, J. C., Gregori, C., Frain, M., &kin, M. M., Sala-Trepat, J., and Kahn, A. (1984) Biochem. Biophys. Res. Commun. 117, 601-609. 20. Southern, E. W. (1975) J. Mol. Biol. 98, 503-5 17. 21. Norgard, M. W., Emigholz, K., and Monahan, J. J. (1979) J. Bacterial. 138, 270-272. 22. Mac Gillivray, R. T. A., Friezner-Degen, S. J., Chandra, T., Woo, S. L., and Davie, E. W. (1980) Proc. Natl. Acad. Sci. USA. 77, 5153-5157. 23. Friezner-Degen, S. J., Mac G&ray, R. T. A., and Davie, E. W. (1983) Biochemistry 22,2087-2097.
BIOCHEMISTRY
142,271-276
(1984)
Immunological Screening of Standard cDNA Libraries in pBR322 Vectors: Detection of Human Fibrinogen and Prothrombin cDNA Clones HENRI PLAISAN&,* YVONNE ALEXANDRE,* GEORGES UZAN,~ CLAUDE BESMOND,* RICHARD BENAROUS,$MONIQUE FRAIN,~ Jose SALA TREPAT,~ JEAN-CLAUDE DREYFLJS,*AND AXEL KAHN*” *INSERM U.129, Institut de Pathologie Molkdaire, CHU Cochin, 75674 Paris Cedex 14, TINSERM lJ.143, Institut de Pathologie Cellulaire, Hopital de BicZtre, 94270 Le Kremlin Bi&tre, $INSERM U.15, Institut de Pathologie Molkculaire, CHU Cochin. 75474 Paris Cedex 14, and $Laboratoire d’Enzymologie du CNRS. 91190 Gif sur Yvette, France
Received February 6, 1984 The in situ immunological detection of antigens encoded by cDNA inserted into the PstI site of pBR322 plasmids was optimized. It was found that sensitivity of the detection was dramatically increased by in situ amplification of the recombinant plasmids on chloramphenicolcontaining medium followed by a brief incubation without chloramphenicol during which protein synthesis resumes. In addition, several modifications of the previously described methods which permit total suppression of background and false positives are described. These techniques allowed easy detection of cDNA clones for human B 6- and -y-fibrinogen and -prothrombin using a human liver double-stranded cDNA recombinant plasmid library in pBR322 vectors. 0 1984 Academic Press, Inc. KEY WORDS: cDNA libraries; pBR322 plasmids; immunological screening; expression vectors; human fibrinogen; human prothrombin.
Detection of cDNA recombinant clones by immunoassay of the antigen synthesized by the transformed bacteria colonies is a very powerful technique: provided that specific antibodies are available, it allows direct screening of a large number of clones for mRNA species which are expressed at a low level and whose sequence is unknown. Probably one of the best methods used today is to build a special expression library in the lysogen phage expression vector Xgt 11 (1). Most of the earlier cDNA libraries, however, were built using the standard technique of cloning poly(dC)-tailed cDNA into the poly(dG)-tailed PstI site of pBR322 plasmid (2-5). Kemp and Cowman described a simple in situ immunoassay for screening Escherichia coii colonies transformed by recombinant pBR322 (6). This method was subsequently I To whom reprint requests should be addressed.
used for detecting cDNA clones for rat dopamine P-hydroxylase (7) and ferritin (8). In this paper we show that some modifications of the previously published technique greatly increase its sensitivity, with suppression of background and nonspecific colonies. The main improvement is achieved by in situ amplification of the plasmid by chloramphenicol, followed by a brief reincubation without chloramphenicol to permit synthesis of new proteins. This technique should probably enable already existing cDNA libraries to be used for detecting any clone corresponding to products for which specific antisera are available. MATERIAL
AND
METHODS
cDNA library and antibodies. A human
liver ds* cDNA recombinant
plasmid library
* Abbreviations used: ds, double-stranded: CNBr, cyanogen bromide. 271
0003-2697184 $3.00 Copyright 0 1984 by Academic Press, Inc. All r&s of reproduction in any form reserved.
272
PLAISANCIE
in the pBR322 vector was constructed as reported elsewhere (9,lO) using the dC-tailing method. cDNA was inserted into the PstI site. Anti-human fibrinogen (11) and prothrombin (12) antisera were prepared by immunizing rabbits against denatured proteins; the antibodies were then purified by immunoaffinity chromatography. Anti-glucose phosphate isomerase and phosphofiuctokinase-purified antibodies used as controls were prepared against native enzymes (13). Fibrinogen cDNA clones. Cloning of several cDNA inserts for the three fibrinogen chains has been reported elsewhere (14). The inserts were all identified by positive hybridization selection and by verifying that their restriction map fitted well with the known nucleotide sequence of the different fibrinogen mRNAs ( 1518). Other previously identified cDNA clones, specifying aldolase B and transferrin (9,10,19), were used as controls. All the tests aimed at improving the immunological screening methods were performed using replicas from selected, ordered clones in a 96well microtitration plate. The plate contained three clones for the A a-fibrinogen chain, five for B p-, three for y-, two for aldolase B, and two for transfer-tin. In situ ampliJcation of plasmid DNA and restarting of protein synthesis. Colonies were replicated onto a nitrocellulose filter which was then placed on an L-broth agar plate containing tetracycline (10 pg/ml). After colony growth (overnight incubation at 37°C) the filter was peeled from the agar plate and ;.:. transferred onto a similar medium containing 10 &ml chloramphenicol in addition to 10 hg/ml tetracycline. After a further 12- 18 h incubation at 37”C, the filter was peeled again from the chloramphenicol medium and transferred onto a third medium identical to the first one for a 4-h period at 37°C. In some preliminary tests, the amplification step was omitted, or the reincubation without chloramphenicol after plasmid amplification was either omitted or stopped after 1, 2, 3, 4, or 12 h. The colony sizes ranged from 1
ET AL.
to 3 mm. Comparisons of the signal obtained with or without in situ amplification plasmid DNA were performed using colonies of similar sizes. Immobilization of lysed E. coli protein on CNBr-activated Whatman 540 jilters. This step was performed essentially as indicated by Kemp and Cowman (6) except that 1 mM diisopropylphosphofluoridate was added to the transfer buffer. The colonies grown on the nitrocellulose filters were lysed by contact (colony side up) with blotting paper saturated with 0.1 M NaHCOJl% (v/v) Triton X-100/ 2 mg of lysozyme per milliliter, for 20 min in an atmosphere saturated with CHCl, vapor. After an additional 20 min of contact with paper saturated with “binding buffer” (0.1 M NaHCOs, 0.1% Triton X-100), the lysed colonies on nitrocellulose filters were blotted on CNBr-activated Whatman 540 filters for 4-6 h, in a system similar to that described by Southern for DNA blotting (20) using “the binding buffer.” The free reactive groups of the CNBr filter were blocked by incubation for 3 h at room temperature in 2% (w/v) glycine, 1% bovine serum albumin in wash buffer 1 [lo mM Tris-HCl (pH 7.8) 1 M NaCl, 1% Triton X-100, 1% (v/v) Nonidet-P40, and 1% (w/v) sodium deoxycholate]. The filter was then treated for 1 h at room temperature by 2 pg/ ml DNase 1, 10 pg/ml RNase in a 10 mM Tris-HCl buffer (pH 7.4), 5 mM MgClz, 100 mM NaCl, and then washed again with wash buffer 1. All the buffers used were extemporarily filtered through 0.42~pm Millipore membranes. Preparation of pBR322-transformed E. coli protein whole extract. E. coli (C 600 or MC 106 1 strains) transformed by nonrecombinant pBR322 plasmids were grown in liquid medium up to an absorbance of 1.60 OD at 600 nm and then collected by centrifugation and washed in 10 mM Tris-HCl, 150 mM NaCl, and 50 mM sucrose. The bacterial cells were lysed by adding lysozyme (2 mg/ml) and Triton X- 100 (I%, v/v). Nucleic acids were precipitated by the progressive addition
IMMUNOLOGICAL
SCREENING
OF cDNA LIBRARIES
IN pBR322 VECTORS
273
of a 50 mg/ml solution of protamine sulfate and then eliminated by centrifugation. The supernatant was adjusted to 1 mrvt diisopropylphosphofluoridate, 10 mM +aminocaproic acid, 0.1 mM pepstatin, and 1% (v/v) aprotinin as antiproteolytic agents (13), and then stored frozen at -80°C.
Detection of antigens in immobilized colony lysates.The filter was incubated under continuous gentle agitation for 4 h at room temperature with diluted antiserum (1 to 5 pg/ml of purified antibodies) in about 2 ml/ filter of a 10 mM Tris-HCl buffer (pH 7.8), 150 mM NaCl, 1% (w/v) glycin, 1% (w/v) bovine serum albumin, Nonidet-P40, Triton X- 100 and sodium deoxycholate (1% each), E. coli and 2 mg/ml pBR322-transformed proteins. The filter was then washed several times with 20 ml/filter of, successively, wash buffer 1 (2-3 times over 1 h), wash buffer 1 without detergent (2-3 times over 1 h), and finally with an excess of 10 BIM Tris-HCl buffer (pH 7.8) plus 10 mM NaCI. This last wash usually lasted overnight. The filter was then treated for 2 h with ‘251-labeled protein A in the same buffer as that used for the antigen-antibody reaction, and then washed again exactly as indicated after antibody fixation. As mentioned above, all the buffers were filtered on Millipore membranes. In particular, antibody and protein A solutions could be used several times without apparent decrease of sensitivity, but should be filtered again each time. Autoradiography was performed at -80°C for from 15 h to several days using intensifying screens. RESULTS AND DISCUSSION
Three out of the fifteen fibrinogen clones studied were easily detected by the direct immunoassay screening reported here, without any background or false positives (Fig. 1). These clones were not detected when either nonimmune IgG or anti-prothrombin, anti-glucose phosphate isomerase, and antiphosphofructokinase antibodies were used
without
with amplification qk
1
2
FIG. 1. lmmunodetection of fibrinogen clones with or without amplification of recombinant plasmids by chloramphenicol and influence of E. cob protein extracts and glycine on the background. (I) After amplification, the colonies were deposited on a new medium without chloramphenicol for 8 h. The filters with “amplified” or “not amplified” plasmids were treated in a similar fashion, the autoradiograms being exposed for the same times (14 h at -80°C using intensifying screens). For both filters E. coli protein extract and glycine were added to the buffers to reduce background. (II, 1) Same conditions as in (I), with amplification but without addition of glycine. 14-h autoradiographic exposure. (II, 2) Idem, but without addition of E. coli protein extract and glycine. 6-h autoradiographic exposure. The diameters of the transferred colonies were about 3 mm in each of these experiments.
instead of anti-fibrinogen antiserum. Clones designated a and b contained 800- to lOOObp-long cDNA inserts specific for B P-fibrinogen chains while clone c contained a 1200bp-long, y-chain specific insert. All these inserts were oriented correctly for translation and, although they were not full length (B @-
274
PLAISANCIk
and y-mRNAs are about 1800-2 100 bases long (14-15)), they all corresponded to a large part of the coding region (15- 18). The low positivity of the y-chain clone (c) as compared to that of the B P-chain clones (a and b) could be explained either by a lower expression or by a lower reactivity of the antiserum against y than against B /3 chains. However, this last explanation is unlikely because the antibodies used, prepared against denatured fibrinogen, recognized cell-free synthesized y chains practically as well as A cx or B @chains (11). It is possible, however, that the portion of y chain synthetized under
ET AL.
direction of the insert does not contain the main antigenic epitopes of the proteins. Since we tested only 15 fibrinogen clones, we cannot determine if, as indicated by Kemp and Cowman (6), the proportion of detectable colonies is higher than the expected one in six sequences which would have a correct orientation and frame. As shown in Fig. 1, sensitivity of detection of the positive clones is dramatically increased by in situ amplification of recombinant plasmids in the transformed colonies. This is achieved using the classical chloramphenicol method (2 1); this antibiotic, however, blocks
FIG. 2. Immunodetection of fibrinogen clones on colonies lysed various times after amplification with chloramphenicol. After amplification, the colonies were lyscd either immediately or 1, 2, or 4 h after the colonies were transferred to agar plates without chloramphenicol. The diameter of the colonies did not significantly increase during the first hours of incubation without chloramphenicol. Their diameters ranged from 2.5 to 3.5 mm. All the filters were similarly treated and exposed for autoradiography.
IMMUNOLOGICAL
SCREENING
OF cDNA LIBRARIES
protein synthesis so that the antigen is undetectable immediately after amplification (Fig. 2). When the filter is put on a new medium without chloramphenicol, antigen synthesis starts again, is detectable after 1 h, reaches a maximum at the 4th-5th hour (Fig. 3), and then decreases slightly (not shown) while growth of the colonies resumes. The increase of the sensitivity of detection of the antigens is clearly not dependent upon growth of the colonies; the sizes of “amplified” and “nonamplified” colonies are indeed similar, and the increase in the size of “nonamplified” colonies results in an increase of the surface of the autoradiographic signal, not of its intensity. In addition, protracted growth without chloramphenicol after chloramphenicol amplification results, as indicated above, in a progressive decrease of the signal. Among the factors which contribute to decreasing background and false-positive colonies or spots, four proved essential in our tests: (i) protracted washes are needed especially to avoid the diffuse background which can occur after long autoradiographic expo-
IN pBR322 VECTORS
275
sure; (ii) filtering of buffers as well as antibody and protein A solutions permits elimination of the strong false-positive spots which can occur outside the colonies; and (iii) and (iv) adding both glycine and pBR322transformed E. coli total proteins to antibody and protein A solutions dramatically decreases nonspecific labeling of colonies (Fig. 1). In fact, negative colonies either are not seen even after several days of autoradiographic exposure (Fig. 3) or, in some cases can appear in negative on a slight background (Fig. 3). Using the method described in this paper, we have screened 7000 ordered clones from a human liver ds cDNA library with antiprothrombin antibodies. Three clones were so detected, each of them being firmly identified as containing a prothrombin cDNA insert by positive hybridization-selection and comparison of their restriction map with the known sequence of the mRNA (22-23). We recently found that this method is also fully convenient for screening nonordered colonies on high-density filters (unpublished results). In conclusion, this paper describes opti-
FIG. 3. Immunodetection of two human prothrombin clones among 192 ordered colonies. Seven thousand colonies ordered in 96-well microtitration plates were transferred to filters, the plasmids were amplified, and then the colonies were lysed and proteins were fixed to CNBr-activated Whatman 540 paper. This figure shows two Whatman filters on which positive colonies were detected with purified antibodies against sodium dodecyl sulfate-denatured human prothrombin. 36 h of autoradiographic exposure.
PLAISANCIE
276
mization of the immunological detection cDNA clones in usual pBR322 vectors. Using the methods described it is possible to obtain strong and specific signals, without any background or false positives. Investigators who possess large recombinant plasmid cDNA libraries may find these methods a valuable alternative to the construction of a new library in special expression vectors.
REFERENCES
ET AL. Tmpat, J., Marie, J., and Kahn, A. (1983) J. Biol. Chem. 258, 14576-14584. 10. Uzan, G., Frain, M., Park, I., Beamond, C., Maessen, G., Sala-Ttipat, J., Zakin, M. M., and Kahn, A. (1984) Biochem. Biophys. Res. Commun. 119, 273-281.
II. Uzan, G., Besmond, C., Kahn, A., and Marguerie, G. (1982) Biochem. Int. 4, 27 1-277. 12. Besmond, C., Benarous, R., Kahn, A. (198 1) Biochem. Biophys. Res. Commun. 103, 587-594. 13. Kahn, A., Cottreau, D., Daegelen, D., and Dreyfus, J. C. (1981) Eur. J. Biochem. 116, 7-12. 14. Uzan, G., Courtois, G., Besmond, C., Frain, M, &la-T&pat, J., Kahn, A., and Marguerie, G. (1984) B&hem. Biophys. Rex Commun. 120, 376-383.
1. Young, R. A., and Davis, R. A. (1983) Proc. Null. Acad. Sci. USA 80, 1194-l 198. 2. Kushner, S. R. (1978) in Genetic Engineering (Boyer, H. B., and Nicosia, S., eds.), pp. 17-28, Elsevier/ North-Holland, Amsterdam/New York. 3. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) in Molecular Cloning (Maniatis, T., ed.), pp. 3 12328, Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. 4. Rowekamp, W., and Firtel, R. A. (1980) Dev. Biol. 79,409-418.
5. Wickens, M. P., Buell, G. N., and Shimer, R. T. (1978) .I. Biol. Chem. 253,2483-2495. 6. Kemp, D. J., and Cowman, A. F. (1981) Proc. Nafl. Acad. Sci. USA 78,4520-4524. 7. O’Malley, K. L., Mauron, A., Riese, J., Barchas, J. D., and Kedes, L. (1983) Proc. Nat/. Acad. Sci. USA 80,2161-2165. 8. Brown, A. J. P., Leibold, E. A., and Munro, H. N. (1983) Proc. Natl. Acad. Sci. USA 80, 1265-1269. 9. Simon, M. P., Besmond, C., Cottreau, D., ChaumetRiffaud, P., Weber, A., Dreyfus, J. C., Sala-
15. Kant, J. A., Lord, S. T., and Crabtree, G. R. (1983) Proc. Natl. Acad. Sci. USA. 80, 3953-3957. 16. Rixon, M. W., Chan, W. Y., Davie, E. W., and Chung, D. W. (1983) Biochemistry 22, 32373244.
17. Chung, D. W., Que, B. G., Rixon, M. W., Mace, M., and Davie, E. W. (1983) Biochemisfry 22, 3244-3250.
18. Chung, D. W., Chan, W. Y., and Davie, E. W. (1983) Biochemistry 22, 3250-3256. 19. Besmond, C., Dreyfus, J. C., Gregori, C., Frain, M., &kin, M. M., Sala-Trepat, J., and Kahn, A. (1984) Biochem. Biophys. Res. Commun. 117, 601-609. 20. Southern, E. W. (1975) J. Mol. Biol. 98, 503-5 17. 21. Norgard, M. W., Emigholz, K., and Monahan, J. J. (1979) J. Bacterial. 138, 270-272. 22. Mac Gillivray, R. T. A., Friezner-Degen, S. J., Chandra, T., Woo, S. L., and Davie, E. W. (1980) Proc. Natl. Acad. Sci. USA. 77, 5153-5157. 23. Friezner-Degen, S. J., Mac G&ray, R. T. A., and Davie, E. W. (1983) Biochemistry 22,2087-2097.