9 1993 by Humana Press Inc. All fights of any nature whatsoever reserved. 0163-4992/92/21/69-79/$3.20

Regulated Secretion and Purification of Recombinant Antibodies in E. coU S. DtiBEL, F. BREITUNG, I. KLEWlNGHAOS,AND M. LrrrLE* Recombinant Antibody Group, German Cancer Research Center, FSP 04, lm Neuenheimer Feld 506, D-69120 Heidelberg, Germany

ABSTRACT A plasmid for optimized l~rotein expression of recombinant Fv antibodies (pOPE) in E. coli was used to express the variable domains of the murine monoclonal antibody HD39 specific for the human B-cell surface antigen CD22. The production of Fv antibodies by pOPE can be regulated over a wide range by varying the IPTG concentration. Antibodies that can discriminate between secreted and nonsecreted Fv antibody fragments were used to show that secretion is the limiting step for the production of functional Fv antibodies. IPTG concentrations above 20 W~d increased the total antibody production, but did not yield larger amounts of secreted Fv antibodies. The addition of five histidines to the C terminus facilitates an easy single-step enrichment procedure based on immobilized metal affinity chromatography. Index Entries: Recombinant antibody; monoclonal antibody; Fv antibody; IMAC; B-lymphocyte; CD22; immunoglobulin; E. coli.

*Author to whom all correspondence and reprint requests should be addressed. Cell Biophysics


Volume 21, 1992


Diibel et al.

INTRODUCTION Recombinant antibodies are gaining increasing interest for cancer diagnosis and therapy in order to minimize side effects for the patients and to reduce costs. "Fv antibodies" consisting of only the variable regions connected covalently by a peptide linker can be used as a minimal binding unit (1). These Fv antibodies should minimize the immune response of the patient against the antibody and provide an increased clearance rate. Chemical coupling to produce immunoconjugates with toxins or enzymes can be circumvented by gene fusion (2--4). In recent years, methods were developed to mimic the basic aspects of the mammalian immune response in E. coli (5,6). Large "naive" recombinant antibody gene libraries have been prepared from human peripheral blood lymphocytes. To select specific antibody clones from those large libraries (e.g., from more than 107independent clones [7]), vectors were developed for imitating the process of clonal selection in the immune system by physically connecting the antibody with a plasmid encoding its gene. This was achieved by fusing the antibody binding domain to surface proteins of bacteriophages (8-10) or bacteria (11). Completely "human" Fv antibodies for diagnostic or therapeutic purpose may therefore soon be available using this new technology. Alternatively, a large panel of well-defined hybridomas that secrete murine monoclonal antibodies has already been established. The production of these antibodies as Fv fragments in E. coli could therefore also be very interesting as a means of reducing side effects and costs. In both cases, a system for the small-scale production and purification of Fv fragments has to be established for analysis of specificity prior to mass production. Here we describe optimal expression conditions and a simplified purification procedure for an Fv antibody constructed from the murine monoclonal antibody HD39 (12) specific for the human B-cell surface antigen CD22.

MATERIALS AND METHODS E. coli JM109 cells transfected with pOPEHD39 were cultivated in LB medium (13) containing 120 mM glucose and 100 ~tM ampicillin at room temperature on a shaker. Production of Fv antibodies was induced by the addition of IPTG (isopropyl-~-D-thiogalactopyranosid) to the cultures. Total cell extracts for Western blots were Cell Biophysics

Volume 21, 1992

Recombinant Antibodies in E. Coli


prepared by pelleting the bacteria for I min at 8000g. The pellet was directly resuspended in SDS sample buffer containing mercaptoethanol and boiled for 10 min. The supernatants were precipitated with trichloroacetic acid (TCA, final concentration 20% w/v) prior to electrophoresis. The samples were analyzed on SDS-polyacrylamide gels (14). Western blots were performed according to Towbin et al. (15) using second antibodies coupled to horseradish peroxidase with diaminobenzidene as substrate in the presence of 0.02% (W/V) CoC12 to enhance the staining. The rat monoclonal antibody Yo11/34 was purchased from Serotec, Oxford, UK, and the mouse monoclonaI andbody 9E10 from Cambridge Research Biochemicals, Cambridge, UK. Proteins on Western blots were stained with 0.1% (w/v) Ponceau S. For IMAC (immobilized metal affinity chromatography [16,17]) cells harvested by centrifugation as described above were sonicated in 80 ~tL/mL bacterial culture of IMAC column buffer (50 mM sodium phosphate, pH 7.0 containing 1M NaC1) on ice. The supernatant of a 30min centrifugation at 40,000g and 4~ was directly applied on a column of chelating sepharose (Pharmacia, Freiburg, Germany) preloaded with Zn or Ni ions. Preloading was achieved by successive washes of the column with 2 bed vol of 50 mM ZnC12or NiC12 followed by 6 bed vol of H20 and 4 bed vol of IMAC column buffer. After the application of the sample, a stepwise gradient of 1-150 mM imidazole in a total of 12 bed vol of IMAC column buffer was applied to the column followed by an elution with 50 mM EDTA (2 bed vol of each concentration). All fractions were precipitated with TCA (final concentration 20% w/v) and washed with I mL 80% (v/v) ice-cold ethanol. SDS gel electrophoreses on 12% polyacrylamide gels and Western blots were performed as described above.

RESULTS AND DISCUSSION The isolation, PCR amplification, and cloning of the HD39 antibody variable domain genes were essentially carried out as described by Dfibel et al. (7). The primer design, PCR conditions, vector construction, and detailed procedures for these steps will be published in a separate paper, pOPE was derived from pSEX (8), a phagemid expressing a fusion protein of Fv with the phage docking protein pill, by removing the cat gene and substituting five histidine residues for pHI (Fig. 1B). Cell Biophysics

Volume 21, 1992


Di)bel et al. Xhol Pvull Pstl


Heavy Chain P/O




pOPE 3916bp


Amp r



xba I B a m HI ..GGA TCC CAT CAC CAT CAT CAT T A A TCT AGA... ..Gly Ser His His His His His Stop

Fig. 1. A. Plasmid for optimized larotein expression (pOPE). The expressed Fv antibody fusion protein is boxed. AmpR, Ampicillin resistance gene; P/O, promoter/operator; leader, signal sequence of pectate lyase; linker, 18 amino acids containing the epitope for the tubulin monoclonal antibody Yo11/34. B. Sequence of the histidine tag at the C terminus of the Fv antibody gene. The genes for the heavy- and light-chain variable domains are linked by a sequence containing the epitope ... EEGEFSEAR ... recognized by the monoclonal antibody Yoll / 34 (18,19). Yoll/34 can therefore be used to detect the fusion protein in bacterial preparations. The antibody gene is preceded by a pelB bacterial leader sequence for secretion into the periplasmic space. The 5 C-terminal histidines were added for the purification by "immobilized metal affinity chromatography" (IMAC [16,17]). The complete construct is shown in Fig. 1A. Different concentrations of IPTG can be used to adjust the amount of Fv antibody produced (Fig. 2A). One hundred micromolars of IPTG induced the production of Fv antibody to a level where it represented Cell Biophysics

Volume 21, 1992

Recombinant Antibodies in E. Coli




0 2 5 102050100



:: :i

_...45 36



~ ~ I

424 20

I" I I I I . . . . ~::~:~:§ ~i~ --..--45 2:::!i 36 ......... ~-29 -~24 --20



Fig. 2. Induction of pOPEHD39 with IPTG. E. coli JM109 transfected with pOPEHD39were grown to an OD600nm of 0.9. Aliquots were induced with different concentrations of IPTG. Cells were harvested by centrifugation, boiled in SDS sample buffer (100 ~tL/1OD600~m to achieve equal amounts of protein in each lane, controlled by Ponceau S staining of the Western blot) and subjected to SDS-PAGE. Western blots of 12% polyacrylamide gels were stained either by Yo11/34 recognizing total FvAb or by serum A recognizing only processed FvAb. (A) Yo11/34 stain of total protein preparations. (B) Serum A stain of the same total protein preparation used in (A). (C) Serum A stain of proteins from the culture supernatant of the cells analyzed in (A) and (B). Fractions and IPTG concentrations are indicated above each lane. the most prominent band in the total bacterial protein pattern without a significant inhibition of the growth rate and viability of the cells (Fig. 3). Concentrations above 100 W~I IPTG or induction at temperatures above 30~ led to a substantially decreased growth rate and finally to cell death. This wide range of Fv antibody production is facilitated by the promotor/operator region from pSEX. Its tight regulation helps to Cell Biophysics

Volume 21, 1992


Dhbel et al.


"0" - -'0

"~60 o

~ 40 g E















. . . .


' I "i


. . . . . . . .



fun IPTG

Fig. 3. Effect of IPTG concentration on the growth of bacteria transfected with pOPEHD39. Aliquots of a stock culture were induced at an OD6~nm of 0.9 (same aliquots as used in Fig. 2). Closed circles, 3 h after induction with IPTG, open circles, 24 h after induction with IPTG. avoid selective pressure against clones carrying antibodies. In the absence of IPTG, only very little Fv antibody could be detected. After induction with IPTG, however, sufficient amounts of fusion protein for analysis can be easily produced.

Optimized Expression of Fv Antibodies Bacterial secretion vectors like pOPE target fusion protein through the inner membrane by means of a bacterial leader sequence, which is cleaved during translocation. To determine the efficiency of this process, we developed a panel of rabbit sera raised against amino terminal peptides of antibody variable chains. Western blots of bacterial extracts containing different fusion proteins were used to characterize the epitopes. One of the sera (serum A) raised against the dodecapeptide QVQLQQSGGGAC representing the N terminus of a human heavy chain was shown to react with the core epitope NH2-QVQLQ, but not with QVQLQ as an internal epitope (Table 1). This core sequence is present in all pOPE constructs since the PstI site used for the insertion of the heavy-chain variable region gene is placed five amino acids downCell Biophysics

Volume 21, 1992


Recombinant Antibodies in E. Coli

Table 1 Serum A Recognizes the Core Sequence QVQLQ Only with a Free N Terminus



Original peptide (used for immunization) Fusion protein pSEX2c Fusion protein pre-pSEX8a Fusion protein pSEX4fY


Detection of the protein by serum A on Western blots a

+ +

aTotal bacterial lysate was separated by 12% PAA-gels and Western blotted. Correct expression of the fusion protein was controlled on a parallel set of Western blots stained with the monoclonal antibody Yo11/34. bThe carboxyterminal amino acids AC were added to facilitate coupling to the carrier protein for immunization. CDescribed by Breitling (20). eDescribed by Breitling et al. (8). qgerived from pSEX2 (8) by removal of the chloramphenicol resistance gene and replacing the Fv antibody gene with a polylinker.

stream from the N terminus of the processed Fv antibody. Since the leader peptide preceding the antibody sequence is cleaved by the signal peptidase on transport through the inner membrane (21,22) this serum can be used to discriminate between processed and unprocessed Fv antibodies. In contrast, Yo11/34 recognizing the epitope ... EEGEFSEAR ... in the linker between heavy- and light-chain domains can be used to detect both processed and unprocessed Fv antibodies. Using these two antibodies, we can therefore determine the relative amount of both processed and unprocessed Fv antibodies. To determine the IPTG concentration for optimal secretion, Western blots of total bacterial lysates and medium supernatants were analyzed with serum A and Yo11/34 (Fig. 2). In bacteria induced with IPTG concentrations above 20 WVI IPTG, no further increase and even a slight decrease of the amount of processed Fv antibodies were observed (Fig. 2B), whereas the Yo11/34 stain demonstrated a further increase in total Fv antibody production (Fig. 2A). The yield of processed Fv antibodies therefore seems to be limited by the capacity of the transmembrane transport system. This was confirmed by Western blots of the Cell Biophysics

Volurne 21, 1992


Di)bel et al.

culture supernatants. No further increase in the amount of secreted antibodies was found in the medium of cells induced with IPTG concentrations above 20 gM (Fig. 2C). We also tested the effect of prolonged induction times on the yield and processing of Fv antibodies. After induction for up to 24 h instead of 3 h, only a slight increase in the yield of total as well as processed and secreted Fv antibodies was observed (about a factor of 2, data not shown). The IPTG dependence of production and processing was identical to the results after 3 h. Furthermore, no increase in the occurrence of smaller proteolytic fragments was observed on Western blots. It is possible, however, that other antibody sequences are more susceptible to proteolysis, in which case long induction times would be a disadvantage. To determine whether the five histidines at the C terminus of the Fv antibody influences processing and secretion, we performed all the exeriments described above using a pOPE derivative in which the DNA coding for the histidine tag was exchanged for DNA encoding an epitope recognized by the anti c-myc antibody 9E10 (... EEKLISEEDL ... [231). No substantial differences concerning growth rate, cell viability and production, processing, and secretion of Fv antibodies were observed, with the exception of a slight decrease in the optimal IPTG concentration for processing and secretion from 20 to 10 W~I.

Enrichment of Fv Antibodies by IMAC The five hisfidines at the C terminus of the Fv antibody should facilitate purification by affinity to immobilized metal ions (16,17). To test whether our fusion protein could be purified by this method, a crude preparation of total soluble proteins of induced E. coli transfected with pOPEHD39was used as starting material (Fig. 4). Using a chelating sepharose column loaded with nickel, most of the proteins were found in the effiuate, whereas most of the Fv antibodies were bound to the column. Washes with increasing concentrations of imidazole removed most of the proteins that had initially bound to the column. The last wash containing 150 rnM imidazole followed by 50 mM EDTA eluted most of the Fv antibodies. The Ni ions on the affinity column can be substituted by Zn ions. Most of the bacterial proteins elute from these Zn columns below 20 mM imidazole, but at these concentrations, the Fv antibodies also start to elute (data not shown). We found no subCell Biophysics

Volume21, 1992

Recombinant Antibodies in E. CoU


< [..., mM Imidazole

1 20 30 40 50 150




I I I I I I I 64 45 36 .... ~ ~ 2 9 ~24 ~20 I I I I I I I

B 64 45 36 29 24 20

Fig. 4. Purification of FVHD39 from a crude bacterial extract by IMAC on chelating sepharose columns loaded with Nickel ions. (A) Western blot of 12% polyacrylamide gels stained with serum A; (B) total protein stain of the blot shown in (A). Fractions and imidazole concentrations are indicated above each lane. Equivalent fractions of the original preparation representing 10 mL of original bacterial culture were applied onto each lane. Arrow: Fv antibody.

stantial differences, however, between purity of Fv antibodies isolated from both column types. The above data show that the Fv antibodies produced by pOPE can be purified from a total soluble protein extract in one step by IMAC to yield a preparation of the fusion protein that contains only a few contaminants. To improve the purity of the Fv antibodies still further, preparations of periplasmic extracts could be employed instead of total CeUBiophysics

Volurne 21, 1992


Diibel et al.

soluble protein and may be a simple step to increase further the purity of the Fv antibodies, pOPE should also be useful for secretion and purification of other proteins in E. coli.

REFERENCES 1. Huston, J. S., Levinson, D., Mudgett-Hunter, M., Tai, M. S., Novotny, J., Margolies, M. N., Ridge, R. J., Bruccoleri, R. E., Haber, E., Crea, R., and Oppermann, H. (1988) Protein engineering of antibody binding sites: recovery of specific activity in an anti-digoxin single-chain Fv analogue produced in Escherichia coli. Proc. Natl. Acad. Sci. USA 85, 5879-5883. 2. Chaudhary, V. K., Queen, C., Junghans, R. P., Waldmann, T. A., FitzGerald, D. J., and Pastan, I. (1989) A recombinant immunotoxin consisting of two antibody variable domains fused to Pseudomonas exotoxin. Nature 339, 394--397. 3. Rybak, S. M., Hoogenboom, H. R., Meade, H. M., Raus, J. C. M., Schwartz, D., and Youle, R. J. (1992) Humanization of immunotoxins. Proc. Natl. Acad. Sci. USA 89, 3165-3169. 4. Seehaus, T., Breitling, F., Dhbel, S., Klewinghaus, I., and Little, M. (1992) A vector for the removal of deletion mutants from antibody libraries. Gene 114, 235-237. 5. Winter, G. and Milstein, C. (1991) Man made antibodies. Nature 349, 293-299. 6. Fuchs, P., Dfibel, S., Breitling, F., and Little, M. (1993) Recombinant human monoclonal antibodies: the basic principles of system transferred immune into E. coli. Cell Biophys. 21, 81-91. 7. Diibel, S., Breitling, F., Seehaus, T., and Little, M. (1992) Generation of a human IgM expression library in E. coli. Meth. Mol. Cell. Biol. 3, 47-52. 8. Breitling,F., Diibel, S., Seehaus, T., Klewinghaus, I., and Little, M. (1991) A surface expression vector for antibody screening. Gene 104, 147-153. 9. Marks, J. D., Hoogenboom, H. R., Bonnert, T. P., McCafferty, J., Griffiths, A. D., and Winter, G. (1991) By-passing immunization. Human antibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222, 581-597. 10. Kang, A. S., Barbas, C. F., Janda, K. D., Benkovic, S. J., and Lerner, R. A. (1991) Linkage of recognition and replication functions by assembling combinatorial antibody Fab libraries along phage surfaces. Proc. Natl. Acad. Sci. USA 88, 4363-4366. Cell Biophysics

Volume 21, 1992

Recombinant Antibodies in E. Coli


11. Fuchs, P., Breitling, F., Diibel, S., Seehaus, T., and Little, M. (1991) Targeting recombinant antibodies to the surface of E. coli: fusion to a peptidoglycan associated lipoprotein. Bio/Technology 9, 1369-1372. 12. D6rken, B., Moldenhauer, G., Pezzutto, A., Schwartz, R., Feller, A., Kiesel, S., and Nadler, L. M. (1986) HD39,(B3), a B lineage-restricted antigen whose cell surface expression is limited to resting and activated human B lymphocytes. J. Immunol. 136, 4470 4479. 13. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) in Molecular Cloning: A Laboratory Manual, 2nd. Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 14. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227, 680-685. 15. Towbin, H., Staehelin, T., and Gordon, I. (1979) Electrophoretic transfer of protein from polyacrylamide gels to nitrocellulose sheets: procedures and some applications Proc. Natl. Acad. Sci. USA 76, 4350--4354. 16. Sulkowski, E. (1985) Purification of proteins by IMAC. Trends in Biotechnology 3, 1-7. 17. Lilius, G., Person, M., Bhlow, L., and Mosbach, K. (1991) Metal affinity recipitation of proteins carrying genetically attached polyhistidine tails. Eur. J. Biochem. 198, 499-504. 18. Kilmartin, J. V., Wright, B., and Milstein, C. (1982) Rat monoclonal antitubulin antibodies derived by using a new nonsecreting rat cell line. J. Cell Biol. 93, 576--582. 19. Breitling, F. and Little, M. (1986) Carboxy-terminal regions on the surface of tubulin and microtubules. Epitope locations of Yo11/34, DMIA and DMIB. J. Mol. Biol. 189, 367-370. 20. Breitling, F. (1992) Konstruktion einer humanen Antik6rperbibliothek in Bakteriophagen zur klonalen Selektion in vitro. PhD Thesis, University of Heidelberg, Germany. 21. Chang, C. N., Blobel, G., and Model, P. (1978) Detection of procaryotic signal peptidase in E. coli membrane fraction: endoproteolytic cleavage of nascent fl pre coat protein. Proc. Natl. Acad. Sci. USA 75, 361-365. 22. Lei, S.-P., Lin, H.-C., Wang, S.-S., Callaway, J., and Wilcox, G. (1987) Characterization of the Erwinia carotovaora pelB gene and its product pectate lyase. J Bacteriol. 169, 4379-4383. 23. Evan, G. I., Lewis, G. K., Ramsay, G., and Bishop, M. (1985) Isolation of monoclonal antibodies specific for human c-myc proto-oncogene product. Mol. Cell. Biol. 5, 3610-3616.

Cell Biophysics

Volume 21, 1992

Regulated secretion and purification of recombinant antibodies in E. coli.

A plasmid for optimized protein expression of recombinant Fv antibodies (pOPE) in E. coli was used to express the variable domains of the murine monoc...
803KB Sizes 0 Downloads 0 Views