On-line version ISSN 0104-6632
Braz. J. Chem. Eng. vol.20 no.1 São Paulo Jan./Mar. 2003
Screening of lectins from South American plants used as affinity ligands to purify rhEPO
G.I.AmadeoI; R.MoreiraII; R.LimaII; D.TeixeiraII; R.KratjeI; M.EtcheverrigarayI
ILaboratorio de Cultivos Celulares, Facultad de Bioquímica y Ciencias Biológica, Universidad Nacional del Litoral. C.C 242, S3000ZAA, Santa Fe, Argentina
IILaboratório de Lectinas e Glicoconjugados, Dpto de Bioquímica e Biología Molecular, Universidade Federal do Ceará., Cx.P. 6020, CEP 60451-970, Fortaleza, CE, Brazil
Two groups of isoforms of rhEPO, at a concentration of 300 µg/ml, were tested as putative inhibitors of the lectinic hemagglutination reaction in order to obtain affinity ligand(s) for hormone purification: groups I (pI: 3.80; 3.89; 3.95; 4.07, 4.15 and 4.26) and groups II (pI: 4.15, 4.26; 4.38; 4.51; 4.72 and 4.93) Crude extracts from the vegetable materials Abrus precatorious (Abrin), Artocarpus incisa (Frutalin), Artocarpus integrifolia (Jacalin), Canavalia ensiformes (ConA), Canavalia brasiliensis (Conbr), Cratylia floribunda, Dioclea altissima (DAL), Dioclea grandiflora (DGL), Erythrina vellutina (EVL), Erythrina cristagalli, Lutaelburgia auriculata (lectin not fully characterized yet), Lycopersicum esculentum (LEA), Phaseolus vulgaris (PHA), Ricinus communis (Ricin) and Triticum vulgaris (WGA) were used. Only some of the galactose-specific lectins and the GlcNAc-specific lectins showed rapid full inhibition of the hemagglutination reaction for the less acidic isoforms and the total isoforms of rhEPO, respectively. On this basis, the selected lectins were purified by affinity chromatoghraphy and covalently coupled to cyanogen bromide activated Sepharose® (Amersham-Pharmacia). CHO.K1 cell culture supernatant containing rhEPO was loaded onto the lectin resins and the recoveries were calculated by using specific elutions.
Keywords: rhEPO, lectin affinity purification, isoform selection.
The extensive application of genetic engineering technologies to the large-scale preparation of biologically active proteins has substantially enhanced the prospects of obtaining them in quantities that can satisfy high demands. In several instances, recovery from the production medium is troublesome, cumbersome and expensive, particularly when the proteins are highly glycosylated. In this case, it is necessary to find affinity ligands to develop simple and efficient procedures suitable for large-scale production (Zanette et al., 2000). Natural proteins usually occur in a spectrum of glycosylated forms. They can have a variety of different biophysical and biochemical properties (Macmillan et al., 2001). Like other glycoproteic hormones, human erythropoietin (EPO) is a mixture of isoforms (Vliegenthart, 1994). In fact, its microheterogeneity is related to the charged carbohydrate moiety of the protein (Imai et al., 1990). Its specific activity and therapeutic profile is profoundly influenced by its glycosylation (Imai et al., 1990; Gokana et al., 1997). Of particular significance is the finding that the lack of terminal sialic acid (Neu5Ac) on the oligosaccharide structure results in high in vitro activity, but low in vivo activity (Delorme et al., 1992; Narhi et al., 1991).
Plant lectins have been used as ligands in glycoprotein purification. Lectins are glycoproteins or oligomeric proteins with one or more sugar-binding site(s) per subunit. These molecules are of nonimmune origin. They reversibly bind specific sugars as well as precipitated polysaccharides, glycoproteins and glycolipids bearing specific sugars, thus acting also as cell recognizers (Singh et al., 1999). The erythrocyte agglutinating activity of lectins is a major attribute of these proteins and is routinely used for their detection and characterization. This reaction is inhibited by those sugar molecules which are specific ligands of the lectins.
MATERIALS AND METHODS
Cell Lines and Cultures
A stable CHO.K1 cell clone transfected with the human EPO gene was obtained in our lab. The cells were adapted to suspension growth in spinner flasks. Dense cell cultivation was carried out during 52 days in a perfused 25 l stirred tank bioreactor (MBR AG, Switzerland). In all cases the media used consisted of a 1:1 mixture of DMEM and Ham's F12 (Gibco BRL, USA) supplemented with 0.2% fetal calf serum (Bioser, Argentina). Human recombinant EPO (rhEPO) was purified from 48 liters of cell culture supernatant obtained between days 13 and 15 of culture, which had a mean density of 1.15 x 106 viable cells/ml and a perfusion rate of 17 liters/day. Two isoform populations (Fig. 1) were isolated using a method previously described (Amadeo et al., 2000): group I (pI = 3.80, 3.89, 3.95, 4.07, 4.15 and 4.26) and group II (pI = 4.15, 4.26, 4.38, 4.51, 4.72 and 4.93).
All the other reagents were of analytical grade. Rabbit blood was obtained by puncture of the marginal vein of healthy animals. All the seeds were collected from plants grown in the state of Ceará, Brazil. Potatoes, tomatoes and unprocessed wheat germ were purchased at a local store.
Analysis of Samples
The protein concentration was measured by Bradford (1976). The rhEPO concentration was determined by a sandwich ELISA developed in our lab. Briefly, commercial monoclonal anti-EPO IgG (GENZYME, England) was adsorbed onto microtiter plates and then incubated with serial dilutions of rhEPO standard or samples. rhEPO was detected with a polyclonal rabbit anti-rhEPO (antiserum produced in our lab), followed by an HRP conjugated anti-rabbit IgG (DAKO), and o-phenylenediamine (Sigma Chemical Co., USA) was added as substrate. The optical density at 450 nm was measured in a microplate reader (Labsystem, Finland). The isoform pattern of the rhEPO eluted from lectin columns was determined by isoelectric focusing. Western blot analysis was used to evaluate the nature of the contaminants in the rhEPO eluates.
Seeds and other crude vegetal materials (potatoes and tomatoes) were finely ground and stirred with 0.15 M NaCl. Suspensions were maintained for 3 h at room temperature or overnight at 4ºC and then centrifugated at 10,000 g during 30 min. at 7ºC. The clear supernatants were used to determine the hemagglutinating activity.
Erythrocyte Agglutination and Inhibition Assays
The assays were carried out in tubes or 96-well microtiter plates. The extract containing the lectins dissolved in 0.15 M NaCl was submitted to serial two-fold dilutions and mixed 1:1 with a 2% rabbit erythrocyte suspension. The extent of hemagglutination was monitored visually after the tubes or plates were allowed to stand at 37ºC for 30 min. and at room temperature for an additional identical interval. The results are reported as hemagglutination titer (the reciprocal of the highest dilution giving visible hemagglutination). The screening of different lectins was assessed by the ability of rhEPO oligosaccharides to inhibit the agglutination of rabbit erythrocytes. Each group of rhEPO isoforms (300 µg/ml) was added to wells containing serial lectin dilutions. The tubes or plates were allowed to stand at 37ºC for 1 h, and after adding a 2% rabbit erythrocyte suspension, they were incubated at 37ºC for 30 min. and at room temperature for an additional identical interval. The results were expressed as hemagglutination inhibition titer (the reciprocal of the highest dilution giving full inhibition).
Purification of the Putative Lectins and Preparation of the Lectin Affinity Columns
All the selected lectins (potato lectin (PL), tomato lectin (TL) and Abrin) were purified by affinity chromatography (Isamu et al., 1983; Nachbar et al., 1979; Moreira, 1998).
Purified lectins were coupled to cyanogen bromide activated Sepharose 4B® (Amersham-Pharmacia, Sweden) following the dealer's specifications. For wheat germ lectin (WGA), we preferred the use of a commercial column (Amersham-Pharmacia, Sweden).
Chromatography on lectin Affinity Columns
Culture supernatants containing 200 mg of rhEPO were loaded into the four columns: WGA-Sepharose® (0.7 x 2.5 cm), PL-Sepharose (1 x 1.27 cm), TL-Sepharose (1 x 1.27 cm) and Abrin-Sepharose (1 x 1.27 cm). The columns were extensively washed with 20 mM Tris-HCl (pH 7.5). The bound rhEPO was eluted at a linear flow rate of 19 cm/h with a solution containing 0.5 M N-acetylglucosamine (GlcNAc) and 20 mM Tris-HCl (pH 7.5) for the first three cases and with another one containing 0.5 M galactose and 20 mM Tris-HCl (pH 7.5) for the Abrin-Sepharose chromatography.
RESULTS AND DISCUSSION
Table 1 summarizes the results of the hemagglutination and inhibition assays. We assayed distinct lectins with the same sugar-binding specificity because differences in their interactions with cell membranes and glycoconjugates had been previously described (Ramos et al., 1996). This screening method is inexpensive and rapid, it does not require lectin purification prior to assay and it can be used for any glycoprotein. In this way, the use of screening in blots was avoided. However, the latter approach is more tedious and time consuming, because it requires not only prior purification and labelling of the lectin, but also, the use of small test columns for the initial analysis.
Our previous studies showed both that the mixture of both groups of rhEPO isoforms have a reduced biological activity, which does not comply with the regulations established by the Pharmacopoeia, and also, that this effect is caused by the low activity of the less acidic glycoforms. Therefore, techniques that allow estimation and/or removal of these undesirable rhEPO isoforms from the culture supernatant are of great interest.
The inhibition titer shown in Table 1 ranged from 1 to 212 for both rhEPO fractions assayed. Low inhibition titer means that the avidity of that lectin for the corresponding rhEPO isoform population is high. Consequently, on account of their low inhibition titer and their similar value for both rhEPO isoform populations, tomato lectin, potato lectin and wheat germ lectin were selected as possible affinity ligands for the recovery of all rhEPO isoforms and Abrin was chosen for the recovery of the less acidic forms (Group II).
Fig. 2 shows the recoveries actually obtained using each lectin resin. The rhEPO eluted from the columns was further characterized by isoelectric focusing (data not shown). The results showed that rhEPO recovered from the chromatographic affinity columns with WGA-Sepharose, PL-Sepharose and TL-Sepharose have all the isoforms, while only the less acidic isoforms of rhEPO were eluted from the Abrin-Sepharose column.
In this way, the performance of the different lectin affinity chromatographies assayed for both rhEPO isoform populations could be predicted. Moreover, determination of recovery of rhEPO bound to Abrin columns might be useful to estimate the less acidic isoforms produced by the recombinant line under different culture conditions.
Using the WGA affinity chromatography made it possible to obtain rhEPO preparations of intermediate purity (70 %) in one single step (Fig. 3). Fig. 4 shows that the main eluate contaminant is a glycoprotein containing N-acetylglucosamine, whose M.W. is approximately 60,000 Da. Western blot analysis demonstrated it is neither bovine serum albumin nor cellular protein.
Downstream processing is considered a critical step in the commercial development of biotechnology. Of the different techniques available, affinity chromatography remains one of the most powerful methods. Lectins are glycoproteins or oligomeric proteins with one or more sugar-binding site(s) per subunit.
Inhibition assays of erythrocyte agglutination by sugar molecules that are specific ligands of lectins allowed us to screen a large lectin collection in order to select the appropriate ones for rhEPO purification. The screening method used has proved to be adequate, as the performance of lectin affinity chromatography could be predicted by the hemagglutination inhibition titer.
How EPO binds to wheat germ lectin, but without distinguishing among its different fractions, has been previously described (Rudzki et al., 1978). Here, taking into account both the low inhibition titer as well as the similar values for both rhEPO isoform populations, we have selected potato lectin, tomato lectin and wheat germ lectin for the recovery of all rhEPO isoforms present in culture supernatants. On the other hand, Abrin was selected for binding the less acidic rhEPO isoforms.
Considering the design of an industrial rhEPO purification protocol, it would be of great interest to directly select the acidic rhEPO population from the culture supernatant. But no lectin with a selective binding capacity for this low pI rhEPO fraction is available. Alternatively, another approach would be to remove the higher pI fraction from rhEPO preparations containing a mixture of all isoforms. For this purpose, Abrin cannot be used due to its high toxicity. However, it could be replaced by other nontoxic lectins with a similar sugar specificity and performance. Results from Table 1 show that Jacalin would be a probable candidate, but further studies have to be carried out with the aim of checking this assumption.
Amadeo, G. I., Pereira Bacci, D., Kratje, R. and Etcheverrigaray, M. (2000) Selecting the Correct Isoforms of rhEPO through the Production Process. Proceedings of the International Symposium of More Quality of Life by Means of Biotechnology. Hannover Braunschweig, Germany, 24. [ Links ]
Bradford, M. M. (1976) A Rapid and Sensitive Method for the Quantitation of Microgram Quantities of Protein Utilizing the Principle of Protein-Dye Binding. Anal. Biochem. 72: 248-254. [ Links ]
Delorme, E., Lorenzini, T., Giffin, J., Martin, F., Jacobsen, F., Boone, T. and Elliot, S. (1992) Role of Glycosylation on the Secretion and Biological Activity of Erythropoietin. [ Links ]Biochem 31: 9871-9876.
Gokana, A., Winchenne, J. J., Ben-Ghanem, A., Ahaded, A., Cartron, J.P. and Lambim, P. (1997) Chromatographic Separation of Recombinant Human Erythropoietin Isoforms. J. Chromatography A 791:109-118. [ Links ]
Imai, N., Higuchi, M., Kawamura, A., Tomonoh, K., Oh-Eda, M., Fujiwara, M., Shimonaka, Y. and Ochi, N. (1990) Physicochemical and Biological Characterization of Asialoerythropoietin: Suppressive Effects of Sialic Acid in the Expression of Biological Activity of Human Erythropoietin in vitro. European J. Biochemistry 194: 457-462. [ Links ]
Isamu, M., Akiko, J., Yuko, M. and Nobuko, S. (1983) Purification and Characterization of Potato Lectin. The Journal of Biological Chemistry 258: 5, 2886-2891. [ Links ]
Macmillan, D., Bill, R. M., Sage, K. A., Fern, D. and Flitsch, S. L. (2001) Selective in vitro Glycosylation of Recombinant Proteins: Semi-synthesis of Novel Homogeneous Glycoforms of Human Erythropoietin. Chem. Biol. 8 (2): 133-145. [ Links ]
Moreira, R. A. (1998) Thesis. Universidade Federal do Ceará. Fortaleza, Brazil. [ Links ]
Nachbar, M. S., Oppenheim J. D. and Thomas J. O. (1979) Isolation and Characterization of a Lectin from the Tomato (Lycopersicum esculentum). The Journal of Biological Chemistry 255: 5, 2056-2061. [ Links ]
Narhi, L. O., Arakawa, T., Aoki, K. H., Elmore, R., Rohde, M. F., Boone, T. and Strickland, T. W. (1991) The Effect of Carbohydrate on the Structure and Stability of Erythropoietin. J. Biol. Chem. 266: 23022-23026. [ Links ]
Ramos, M. V., Moreira, R. A., Oliveira, J. T. A., Cavada, B. S. and Rouge, P. (1996) Memorias do Instituto Oswaldo Cruz 91 (6) : 761. [ Links ]
Rudzki, Z., Lange, R. D., Andrews, R. B. and Dunn, C. D. (1978) The Use of Wheat Germ Lectin in the Purification of Erythropoietin. Haematologica 63: 4, 426-430. [ Links ]
Singh, R. S., Tiwary, A. K. and Kennedy, J. F. (1999) Lectins: Sources, Activities and Applications. Critical Reviews in Biotechnology 19 (2): 145-178. [ Links ]
Vliegenthart, J. F. (1994) Studies on Glycoprotein-Derived Carbohydrates. Biochem. Soc. Trans. 22, 370. [ Links ]
Zanette, D., Sarubbi, E. G., Soffientini, A. and Restelli, E. (2000) Process for the Purification of Glycoproteins like Erythropoietin. European Patent Office, EP 0 820 468 B1. [ Links ]
Address to correspondence
Received: March 5, 2002
Accepted: August 30, 2002