966

E. Wenisch et al.

Electrophoresis 1990. 11, 966-969

The observation that only some of the fractions from chloramphenicol and tetracycline had antibacterial activity was therefore understandable. It is interesting that the chloramphenicol fraction with p H 10.03 was more effective than the whole antibiotic for bacteria E . coli strains No. 184 and 185, as well as the Streptococcus strains. The same result was found with tetracycline, where its activity against E. coli strains No. 185,2024 and 505 1 was less than that offractions with pH at 7.90 and 9.17, while the composition of these fractions (namely pH 9.17) only differed slightly according to the HPLC analysis. Future work may bring the necessary results for a detailed explanation of this problem.

5 References [ll 121 131 141 151 161 171

I81

Sova, O.,J. Chromatogr. 1985,320, 15-22. Sova, O.,J. Chrornatogr. 1985,320,213-218. Dobransky, T. and Sova, O., J. Chrornatogr. 1986,358,274-278. Dobransky, T., Polivka, L., Haas, I. and Sova, O., J. Chromatogr. 1987,411,486-489, Dobransky, T., Sova, 0. and Teleha, M.,J. Chrornatogr. 1989,474, 430-434. Sova, 0. and Szilasi, J., Studia Biophys. 1987, 119, 207-210. Sova, O., in: Proceedings: Electroporation Symposium, Bielefeld, University Press, 1990, pp. 54-55. Sova, 0. and Boda, K., Czechoslovak Patent No. 234 801.

I am indebted to WHO Professor Dr. B. L. Toth-Martinez from the Debrecen Medical University, Institute of Pharmacology, Department of Chemotherapy and Pharmacobiochemistry, for his helpful discussions. Received June 5, 1990

Elisabeth Wenisch' Susanne Reiter' Susanne Hinged Franz Steindl' Christa Tauer Alois Jungbauer' Hermann Katinger' Pier Giorgio Righetti* 'Institute of Applied Microbiology, University of Agriculture and Forestry, Vienna 2Departmentof Biomedical Sciences and Technologies, University of Milano

Shifts of isoelectric points between cellular and secreted antibodies as revealed by isoelectric focusing and immobilized pH gradients*) Charge microheterogeneity of monoclonal antibodies, as revealed by isoelectric focusing incarrier ampholytes, has been known for along time. Here wedemonstrate, in the case of monoclonals against the gp-4 1of the HIV-1 virus, that this heterogeneity is already present within the cell sap of hybridoma cells during antibody synthesis. When the monoclonals are secreted extracellularly, the same isoelectric point (PI) spectrum is maintained, but there is a marked redistribution of the relative isoform abundance towards the lower plcomponents. This suggests in vivo processing of such forms, possibly via glycosylation or deamidation. The secreted antibodies are also analyzed by immobilized p H gradients (IPG), where they demonstrate an even more extensive heterogeneity, due to the marked increment in resolving power. Single bands are purified by preparative IPGs in a multicompartment electrolyzer and are shown to be stable with time. Thus, artefactual heterogeneity produced by the focusing technique is completely excluded and cellular processing is clearly established.

1 Introduction In 1970, in a widely quoted paper, Awdeh et al. [ 11 performed an extensive investigation on the immunoglobulins produced by a plasma cell tumor line which, being a single cell clone, should express a single molecular species of immunoglobulin G (IgG). Such cell types should therefore be expressing only a single structural gene for both the heavy and light chains. These authors did, in fact, isolate a single immunoglobulin band by isoelectric focusing (IEF) but only when monitoring IgG synthesis with a short radioactive pulse (10 min). Upon Correspondence: Prof. P. G. Righetti, University of Milano, Dept. Biomedical Sciences & Technologies, Via Celoria 2, Milano 20133, Italy

prolonged labeling (1 h) the I E F pattern became microheterogeneous, with formation of 3 major isoelectric species, the additional bands being to the acidic side of the early synthetic product. This was perhaps the first clear-cut demonstration of in vivo, postsynthetic processing of IgG. With the advent of hybridoma cells, monoclonal antibody production has been feasible on a large scale [21. However, even then, no single I E F form could ever be demonstrated. Thus, according to Hamilton et al. [ 3,4], the main difference between monoclonal and polyclonal IgGs is their relative heterogeneity under I E F conditions. Monoclonals are, in general, characterized by a banding pattern within narrow PI

* Abbreviations: IEF, isoelectric focusing; IPG, immobilized pH gradients; IgG, immunoglobulin G 0VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990

Dedicated to our friend Milan Bier, the Napoleon of electrophoresis, hoping he will never leave the electrophoresis community for the island of St. Helen. 0173-0835/90/1 I 11-0966 %3.50+.25/0

Electrophoresis 1990, 11,966-969

ranges ((0.6 pH units) while polyclonal IgGs generally cover a broad plspectrum, typically > 2 pH units (e.g., pH 5.5-8.0). It is suggested that these posttranslational modifications result in a microheterogeneous population of antibodies with the same specificity (variable region) but with minor differences in the constant region. Current understanding suggests that the heterogeneity in the light chains could be due to deamidation, while the extensive heavy chain heterogeneity most probably results from glycosylation differences. Such heterogeneity of monoclonal IgGs is a common finding and has been described in a host of communications (e. g., [5-71). Even immunoglobulin-binding factors have been reported to be heterogeneous [81. We have recently produced, and purified in a pilot scale, human monoclonal antibodies against the transmembrane protein gp-41 of human immunodeficiency virus HIV-1 [9]. Such antibodies were purified to homogeneity and shown to be heterogeneous in immobilized pH gradients (IPG), with a spectrum of at least 3 major and 3 minor bands, having p l valuesinthepHrangeof9.1-9.6 [lOl.Inthepresentreport,we demonstrate that this limited heterogeneity exists both in the intracellular and the secreted product. In addition, we show that the secreted product, while retaining the same number of IEF bands, dramatically shifts the relative abundance of the isoforms in favor of the most acidic fractions. The major isoforms were purified by preparative IPG and are shown to be stable with time.

2 Materials and methods 2.1 Hybridoma cell culture and antibody extraction A mouse/human heteromyeloma was fused with a peripheral B lymphocyte from an HIV-1 seropositive donor. A xenohybridoma, producing a human monoclonal antibody (IgG,) against the HIV-I transmembrane protein gp-41 and the gp161 envelope protein, was selected and cultivated in roller bottles in 500 mL of RPM 1 1640 medium (Gibco, GB) supplemented with 5 % fetal calf serum (Biological Industries, Israel). The secreted IgG concentration in the cell culture supernatants was 32 Fg/mL, while the intracellular IgG concentration was 330 ng/mL. Before harvesting, the cells were grown to a density of 2.5 x lo5cells/mL. In general the viable cell content was high(>95 %).For purificationofintracellular IgGs, the cells were washed 3 x in phosphate-buffered saline and then suspended in 0.8 M NaCl, 20 % glycerol, 0.5 % Tween 20 and the following mixture of proteolysis inhibitors: 2.5 mM EDTA, 2.5 mM phenylmethylsulfonyl fluoride (PMSF) and IO-'M pepstatin A. Cell lysis was accelerated by freezing and thawing. After centrifuging the cell debris, the cell lysate was equilibrated in 0.1 % Tween 20 and 30 % glycerol. For IEF and IPG analysis, 0.003-0.1 yg protein was applied to precast sample wells in a polyacrylamide gel.

IEF was performed in a flat-bed Multiphor I1 chamber in 5 %T, 3 % C gels containing 24 % glycerol and 5 % v/v Pharmalyte, pH 3-10, as described [ 111. The IEF run lasted 2 h at 2000 V, 16 "C with the sample applied at the anodic gel side. The focusing chamber was flushed with nitrogen and traces of COPwere removed by soda lime. Before and after blotting, the IEF gels were routinely silver-stained according to [121.

pf Shifts between cellular and secreted antibodie\

967

2.3 Analytical IPGs IPG runs were performed in a 4 YoT, 4 % C matrix containing Immobilines in the pH range of 8.5-10 and 0.5 % Ampholine, pH 8-10, as described 1131. The run was continued for up to 8 h at 3000 V, 10 "C. Samples were applied at the anodic gel side. When the gel was stained with Coomassie Brilliant Blue R-250 sample loads of 3-20 pg protein were adopted. 2.4 Preparative IPGs Preparative purification of single monoclonal IgG bands was performed in a multicompartment electrolyzer with Immobiline membranes, as previously reported 1 14, 151. The electrolyzer consisted of 6 chambers, 2 for electrolyte solutions and 4 samplechambers. Recycling was continuedup to 3 days, at 10 OC,with up to 1500 V.

2.5 Direct blotting Nitrocellulose membranes (0.45 ym) were prewetted in distilled water. Upon termination of the IEF run, the moist nitrocellulose sheet was overlaid on the gel surface, excluding air bubbles, and the gel-nitrocellulose assembly was incubated in a humid chamber overnight at 23 "C. After separating the membrane from the gel, the former was stained with alkaline phosphatase conjugated goat anti-human IgGs [41.

2.6 Materials Repel-Silane, GelBond PAG, as well as the Multiphor 2 chamber, Multitemp thermostat, Macrodrive power supply and the Immobiline chemicals were from LKB Produkter AB (Bromma, Sweden), while Pharmalyte carrier ampholytes were purchased from Pharmacia (Uppsala, Sweden). The multichannel peristaltic pump was from Ismatec (Zurich, Switzerland). The glass microfiber filters (GF/D) were purchased from Whatmann. Light paraffin oil (article 7 160) was from Merck (Darmstadt, FRG). Acrylamide, N,N'-methylenebisacrylamide (Bis), N,N,N',N'-tetramethylethylenediamine (TEMED), ammonium persulfate, and Coomassierilliant Blue R-250 were from Bio-Rad (Richmond, C A). Monoclonal antibodies against the gp-41 from the HIV-1 virus were prepared and purified as in [91.

3 Results Figure 1 shows the focusing profile of human monoclonals produced by our hybridomas. After focusing, the IgGs were blotted onto nitrocellulose and, after binding to conjugated goat anti-human IgGs, revealed by a specific alkaline phosphatase stain. The monoclonals focus as a spectrum of at least 3 major and 3 minor bands, with alkaline pls (ca.pH 8.8-9.2 interval). The group of 6 tracks to the left refers to intracellular antibodies, while the corresponding group of 8 tracks to the right represents secreted antibodies. While it can be appreciated that the total number ofbands and theplspectrum is the same in the two populations, there is nevertheless a marked shift in the relative abundance of the various isoforms. In the intracellular population, the 3 most alkaline bands are the most represented species, while in the corresponding secreted bands, the acidic isoforms are now the most abundant. Thus,

968

E. weniscli er at.

Electrophoresis 1990, 11,966-969

Figure 1. IEF pattern of intracellular (left) and secreted (right) monoclonal antibodies. IEF in a pH 3-10 Pharmalyte range, in a 5 %T. 3 %C polyacrylamide gel. After focusing, blotting to a nitrocellulose filter and binding to conjugated goat anti-human IgGs. the monoclonals were revealed by alkaline phosphatase staining. Amounts of IgG loaded for the intracellular monoclonal~:(1)5.5;(2) 1 1;(3)22;(4)44;(5)66; and (6) 88 ng/track; for the secreted monoclonals: (7) 3.2; (8) 6.4; (9) 13;(10) 26; (1 1) 51; (12) 77; (13) 103; and (14) 171 ng/ track. The samples were loaded in pockets precast at the anodic gel side (pH ca. 5).

there appears to be a further processing step in the cell just prior to extracellular transport. In order to obtain a deeper insight into the origin of such polydispersity, the secreted monoclonals were also subjected to analysis on an I P G p H 8.5- 10interval. Essentially the same spectrum of bands is seen (Fig. 2, left side), with an even more heterodisperse profile, due to the much higher resolving power of the IPG technique. We have attempted to purify single p l forms in order to check for the possibility of an interconversion among the different isoforms. Given the pldistribution of thesixmostprominentbands(p1s: 9.01, 9.15,9.29,9.42,9.54 and 9.67), the electrolyzer was assembled with 6 chambers delimited by the following membranes: p1s 9.1 I , 9.25, 9.36, 9.49 and 9.64. After a preliminary purification step, the two central major isoforms (pfs 9.42 and 9.29) were collected in chamber No. 3 (Fig. 2, last track to the right). This material was then used in a cascade fashion for a second purification step in the electrolyzer. In chamber 3 we now collected an es-

Figure 2. Cascade preparative run. The content of chamber 3 in a 5chambered multicompartment electrolyzer run was used as total sample material (50 mg monoclonals), applied to the first chamber of the same electrolyzer equipped with only three sample flow-chambers and allowed to migrate to equilibrium. All conditions for the analytical IPG gel as described in Section 2.3. Ctrl, control, unfractionated starting material. Start: protein profile of the starting material for this experiment, corresponding to the content ofchamber 3 inthe 5-chamber apparatus run. Numbers 1-3 refer to the protein content in the corresponding chambers of the Immobilineelectrolyzer.

sentially pure PI9.42 isoform (Fig. 2). This band was kept at 4 "C for a month and did not convert or degrade into any ofthe other isoforms (not shown). This excludes the possibility that the sample microheterogeneity could be due to in vitro aging.

4 Discussion It is generally accepted today that the charge microheterogeneity of monoclonals is due to postsynthetic processing in the cell sap. This phenomenon can also be appreciated by twodimensional (2-D) gel electrophoresis. In 2-D maps, both heavy and light chains show heterogeneity, with the heavy chains exhibiting the greatest number of IEF bands [ 16, 171. Two main phenomena are held responsible for microheterogeneity of an otherwise homogeneous genetic product: deamidation and glycosylation. In the latter case charged sugars are often incorporated (e.g., sialic acid); therefore, upon processing in the cell sap, new, lower plisoforms usually appear. While glycosylation is an enzyme-driven process, and thus cell-programmed, deamidation is generally considered to be an accidental event, occurring spontaneously during the life time of a protein, perhaps needed for its clearance from the living organism. Asparagine is generally the most likely candidate for deamidation because it can easily form an intramolecular cyclic imide intermediate that can brake down to replace the amide substituent with a carboxyl group; glutamine should thus undergo deamidation at a much reduced rate [181. However, even deamidation does not seem to be a random event, but is probably also cell programmed: according to Kossiakov [ 191, the tertiary structure is the principal determinant of protein deamidation, and thus different proteins with markedly different susceptibility to amide side chain modifications could be assembled. Our data fit in the general framework of this knowledge: there is clearly an intracellular processing of our monoclonals, in agreement with Awdeh et ul. II I and others 13-8 I. However, we have found an additional processing step: in preparation for excretion, IgGs are further processed to lower p1 components (Fig. 1). It could be argued that such amodification is accidental and occurs in the supernatant, after IgG secretion. This seems an unlikely phenomenon however, because our hybridomas are not grown in vivo in ascitic fluid, but in vitro in a synthetic medium. In addition, the pf distribution of our monoclonals has been remarkably stable over the years in

Elecrrophoresis 1990, 11,966-969

which they have been in production and has remained invariable upon hybridoma storage, freeze-thawing and growth in culture. Thus whatever postsynthetic modifications occur, they are produced in a remarkably consistent and reproducible manner within a given hybridoma. It might be argued that, if the additional processing of our monoclonals to lower PI species is a cellular event, such species should then remix with the intracellular IgG population to such an extent that the two families are indistinguishable. Not so, however, ifthe modification event occurs in a separate compartment in preparation for extracellular transport. It is common knowledge today that proteins to be secreted are glycosylated and directly extruded in the lumen of the Golgi cisterne. From there they emerge into the extracellular matrix, packaged in secretory vesicles, by the process of exocytosis [20].According this to mechanism, the two antibody populations would in fact behave as separate entities.

p l Shifts between cellular and sccreted antibodies

969

121 Kohler, G. and Milstein, C., Nature 1975,256.495-497.

[31 Hamilton, R. G., Roebber, M., Reimer, C. B. and Rodkey, L. S., Hybridoma 1987,6,205-217. 141 Hamilton, R. G., Roebber, M., Reimer, C. B. and Rodkey, L. S., Electrophoresis 1987,8, 121- 134. 151 Lowe, J., Hardie, D., Jefferis, R., Ling, N. R., Drysdale, P., Richardson, P., Raykundalia, C., Catty, D., Appleby, P., Drew, R. and MacLennan, I. C. M., Immunology 1981,42,649-659. 161 Suzan, M., Boyer, C.. Schiff, C., Trucy, J., Milili, M. and De Preval, C.,J. Immunol. 1982,19, 1051-1062. [ 7 ] McCue, J. P., Sasagawa, P. K. and Hein, R. H., BiolechnoL Applied Biochem. 1988,10,63-71. 181 Neauport-Sautes, C., Blank, U., Daeron. M., Galinha, A., Teillaud. J . L.. Moncuit, J., Amigorena, S. and Fridman, W. H., Molec. Immunol. 1986,23, 1 183- 119 1. 191 Jungbauer, A,, Tauer, C., Wenisch, E., Steindl, F., Purtcher, M., Reiter, M., Unterluggauer, F., Buchacher, A,, Uhl, K. and Katinger, H., J. Biochem. Biophys. Methods 1989,19, 223-240. I101 Wenisch,E.,Jungbauer,A.,Tauer,C..Reiter,M., Gruber,G..Steindl, F. and Katinger, H., J . Biochem. Biophys. Methods 1989, 18, 309-322. [ 1 I ] Righetti, P. G., Isoelectric Focusing: Theory, Methodology and Ap-

plications, Elsevier, Amsterdam 1983.

Supported in part by a grant from the ‘Innovations und Technologiefond’from the Federal Ministry of Science and Research of the Austrian Government (Project No. 7/62). PGR is supported by grants from the Agenzia Spaziale Italiana and by Progetto Finalizzato FA TMA ( C N R , Rorna).

I121 Heukeshoven,J.andDernick,R.,E/ectrophoresis 1985,6,103-1 12. I 13 I Righetti, P. G., ImmobilizedpH Gradients: TheoryandMethodology, Elsevier, Amsterdam 1990, pp. 66-67. 1141 Righetti, P. G., Wenisch, E. and Faupel, M., J . Chromatogr. 1989, 475,293-309.

Received May 23, 1990

I151 Righetti, P. G., Wenisch, E., Jungbauer,A., Katinger, H. and Faupel, M., J . Clzromatogr. 1990,500,68 1-696. 1 161 Tracy,R. P.,Currie, R. M.,Kyle,R. A. andYoung,D.S.,Clin. Chem.

5 References I11 Awdeh, 2. L., Williamson, A. R. and Askonas, B. A,, Biochem. J . 1970,116,241-248.

1982,28,900-907. 1171 Jellum, E. and Thorsrud, A. K., Clin. Chem. 1982,28, 876-883. 1181 Harding, J. J., Adv. Protein Chemistry 1985,37, 247-265. 1191 Kossiakof, A. A,, Science 1988,240, 191-194.

1201 Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson,J., Molecular Biology o f t h e Cell, Garland Publishing Inc., New York 1983, pp. 355-366.