THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 261, No. 14, Issue of May 15, pp. 6141-6144, 1986 0 1986 by The American Society of Biological Chemists Inc.
Printed in U.’S.A.
coding region into an expression vector based on the Ipp [outer membrane lipoprotein of E. coli) promotor (9). The lectin isolated from pea (Pisum sativum) seeds is synthesized as a single 275-amino acid residue preproprotein (Received for publication, January 22,1986) consisting of a signal sequence followed by the @ chain and Marlene E. Stubbs, Jeremy P. Carver, and then the a chain (8). The signal sequence is removed in the Robert J. Dunn endoplasmic reticulum, leaving a prolectin in which the B and From the Departments of Medical Genetics and Medical a chains are still joined (8). This prolectin i s transported to Biophysics, University of Toronto, Toronto, Ontario, the protein bohes where it is cleaved into a 187-amino acid Canada M5S lA8 residue 0 chain and a 58-amino acid residue a chain (10). Further shortening of the carboxyl termini of both polypepIn order to explore the molecular basis for the gly- tides occurs, probably in the protein bodies, givingrise to two copeptide specificity of legumelectins, we havedeveloped an experimental system in which specific amino isolectins (11).Preliminary evidence suggests that [email protected]
chain of 179 amino acid residues is common to both. isolectins,l but acid alterations can be introduced into the carbohydrate bindingsite of pealectin. This system is based on that thea chains differ, being 54 and 52 amino acid residues the production of pea lectin in Escherichia coli. The in isolectins I and 11, respectively.’ Higgins et al. (8) have plasmid coding for the lectin was constructed fromtwo established that the uncleaved prolectin can bind carbohylectin cDNA sequences isolated from Pisum sativum drate,asdemonstrated by its affinity for Sephadex. This seeds (Higgins, T.J. V., Chandler, P. M., Zurawski, result suggested that pea lectin, synthesized as a single polyG., Button, S. C., and Spencer, D. (1983)J. Biol. Chem. peptide chain in E. coli, should be functional. Furthermore, 258, 9544-9549) and an expression vector based on production of the uncleaved @-achain would eliminate many the gene for the outer membrane lipoprotein E. of coli of the problems associated with the expression of the lectin (Nakamura, K., and Inouye, M. (1982) EMBO J. 1, as a multichain protein. These problems include cleavage of 771-775). The lectin is produced as a single polypep- the @ and a! chains in the E. coli product, or assembly of tide chain and forms insoluble aggregates in E. coli separately synthesized @ and a polypeptides into the extencells (2-5 mg/liter). Functional lectin is recovered by sively interwoven @-sheetsthat form the functional protein. solubilization of the aggregates in guanidinium hydrochloride, renaturation in the presence ofMnCLand MATERIALS AND METHODS CaC12, and affinity purification on Sephadex. This proSources-Enzymes for molecular biologywere purchased from cedure yields a homogeneous 28,000-dalton protein. Comparisonof the recombinantlectin with natural pea Boehringer Mannheim, with the exception of Bal 31 whichwas purchased from Bethesda Research Laboratories (Rockville, MD). lectin in an inhibition of hemagglutination assay demPea lectin cDNA plasmids were provided by T. J. V. Higgins (CSIRO, onstrated that thereis no detectable difference in the Canberra) and PIN expression vectors were provided by M. Inouye carbohydrate binding properties of the two lectins. (SUNY, Stony Brook). E. coli strain JA22l(hsdM+, h d R-, leu B6, lacy, Atrp E5/F’lacZ*) was from M. Inouye, and W3110 ( P ,hsd R-,
Production of Pea Lectin in Escherichia coli*
The interaction between protein and carbohydrate is central toboth cell-to-cell and intracellular recognition phenomena. In animal cells, investigation of these interactions has been limited since the carbohydrate binding proteins have proven difficult to isolate and characterize. In contrast, the binding of soluble plant lectins to glycopeptides has provided a valuable model system. Lectins from the legume family are readily purified and have a high degree of homology as shown by crystallographic (1)and amino acid sequence (2-4) data. Despite such structuralsimilarities, these lectinshave distinct differences in oligosaccharide specificity (5-7). Our goal is to define the molecular basis for this specificity. To complement the biophysical and biochemical analysis of naturally occurring lectins, we wish to generate lectins containing specific amino acid substitutions. As a first step toward this end, we now report the recovery of functional pea lectin after expression in Escherichia coli cells. This was accomplished by fusing two overlapping pea lectin cDNAs (8)and inserting the lectin
* This work was supported by grants from the Medical Research Council of Canada and theNational Cancer Institute of Canada. The costs of publication of this article were defrayed in part by the payment of page charges. This articlemust therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
hsd M+) was from B. Seed (Harvard). General Methods-Plasmid DNAs were prepared by the alkaliSDS3 lysis procedure (12) and purified by centrifugation in cesium chloride-ethidium bromide. Residual RNA was removed by chromatography on Bio-Gel A-50m in 50 mM Tris (pH 8.0), 0.5 M NaCl, 2 mM EDTA. Restriction enzyme digestions and T4 DNA ligation reactions were carried out underthe conditions suggestedby Maniatis et al. (13). DNA fragments were purified from agarose using powdered glass as described by Vogelstein and Gillespie (14) except that lithium bromide was substituted for sodium iodide. Plasmid DNAs were introduced into various strains of E. coli by calcium-mediated transformation (15).All procedures involving recombinant DNAwere carried out at containment level A as specified by the Medical Research Council of Canada. Construction of the Expression Plusmid-The PstI fragment containing the entire pPS15-104 cDNA insert was purified, the ends were made blunt by incubation with T4 DNA polymerase, and Eco RI linkers were added. After digestion with BamHI, the EcoRIBamHI fragment, which codes for the 5‘ region of the lectin gene, was joined to EcoRI-BamHI-digested pBR322 by reaction with T4 DNA ligase and introduced into E. coli strain HB101. The resulting plasmid contained the 5’ region of the pea lectin gene which was recovered by digestion with PstI and BamHI. The DNA coding for the 3’ end of the pea lectin was prepared in an analogous manner J. Rini, personal communication. J. Rini, T. Hofmann, and J. P. Carver, manuscript in preparation. The abbreviations used are: SDS, sodium dodecyl sulfate; TBS, Tris-buffered saline; PBS, phosphate-buffered saline; Fuc, fucose; 1, liter.
Production of Pea Lectin in E. coli
from pPS15-50 by conversion of the RsaI restriction site in the 3' noncoding region of the pea lectin cDNA to HindIII, purification of the BamHI-Hind111 fragment, and ligation into HindIII-BamHIdigested pBR322. Following transformation and selection of the appropriate plasmid, the BamHI-HindIII fragment containing the 3' end of the cDNA, the PstI-BamHIfragment containingthe 5' end of the cDNA, and the large PstI-HindIII fragment from pBR322 were joined to form plasmid pMS-1. The 5'EcoRI terminus of the pea lectin gene was shortened to remove the G-C sequences and some of the signal sequence coding region present in the original cDNA (8).pMS-1 was digested with EcoRI and treated with exonuclease Bal 31. The ends of the DNA were repaired with T4 DNA polymerase and EcoRI linkers were added. The DNA was recircularized and used to transform E. coli strain HB101. Ampicillin-resistant colonies were screened for plasmids containing EcoRI-Hind111 fragments close to 780 base pairs in length. Thesefragments were inserted separately into EcoRIHindIII-digested PIN-11-A,, Az,and A, (9). The PIN-I1 plasmids carry the lac UV5 promotor-operator, allowing controlled expression of an inserted gene, as well as theIpp promotor. AI, Az,and As are different reading frames. E. coli strain JA221 was transformed with the plasmid mixture and 100 ampicillin-resistant colonies from each version of the plasmid were screened for pea lectin expression, following induction with 2 mM isopropylthiogalactoside. Colonies were screened with affinity purified rabbit anti-pea lectin antibodyand lZ5I-proteinA, as described by Young and Davis (17), except that 5% nonfat dry milk (18) was substituted for fetal calf serum. The EcoRI-Hind111fragment from the strongest producer, PIN-11-As-70, wastransferred to pIN-1A3, a vector identical to PIN-11-A3but lacking the lac-UV5 promotoroperator (9), to yield pMS-2 (Fig. 1).The sequence of the 5' end of the XbaI-Hind111fragment from pMS-2 was determined (16) and the amino acids expected at the amino terminus were deduced from the nucleotide sequence. Expression of Pea Lectin in E. coli-E. coli strain W3110(pIN-I-A3 or pMS-2) was grown in the presence of 50 pg/ml ampicillin in L broth at 37 "C to anAsso of 0.8. The cells were pelleted, washed with 0.15 M NaC1, 10 mM Tris-HC1, pH 7.4 (TBS), and lysed by high pressure (20,000 p.s.i.) in thepresence of 1mM phenylmethylsulfonyl fluoride. The lysate was centrifuged in an SS-34 rotor at 15,000 rpm for 30 min at 4 "C. The supernatantwas removed and thepellet was resuspended in TBS. Samples were analyzed on SDS-polyacrylamide gels (19) stained with Coomasie Blue or immunoblotted (20). Purification of Pea Lectin from E. coli-The pellet from E. coli grown to an Asso of 1.0-1.4, prepared as described above, was resuspended in 4 ml of denaturant (7 M guanidine hydrochloride, 0.3 M NaC1, 50 mM Tris-HC1, pH 6.9)/1 of cells and placed on a rotator overnight a t 4 "C. Following centrifugation at 87,000 X g for 30 min at 4 "C, the supernatant was passed over a DEAE-cellulose column (11.6 X 1.2 cm) equilibrated in thesame buffer. The unbound material was made 100 mM CaClZ,100 mM MnClz and diluted a t least 35-fold with 0.15 M NaC1, 100 mM CaCIZ,100 mM MnCl2, 10 mM Tris, pH 6.9 (TBS metals). Sephadex G-75 (10 ml/l of cells) was added and the mixture was placed on a rotator overnight at 4 "C. This material was poured into a column, washed with 10 column volumes of TBS metals and eluted with 200 mM a-methyl-D-mannopyranoside in TBS metals. After extensive dialysis against distilled water, the eluted material was lyophilized. Protein samples were analyzed by electrophoresis on a 12.5% SDS-polyacrylamide gel (19) stained with Coomassie Blue. Hemagglutination Assay-Assays were done in a total volume of 200 pl with 1%human erythrocytes inPBS, p H 6.6. Inhibition assays used the lowest concentration of lectin that gave complete agglutination and thesubstances listed in Table I. The concentration at which hemagglutination was 50% inhibited was used to calculate dissociation constants for the various saccharides or glycopeptides by using (22). the known dissociation constant for mannose ( K d = 7.1 X RESULTS
Construction of the Complete Pea Lectin cDNA-Plasmids pPS15-50 and pPS15-104 contain incomplete but overlapping cDNA inserts coding for pea lectin, inserted intothe PstIsite of pBR322(8). The pPS15-104 insert codes for 19 amino acid residues of the amino-terminal signal sequence, all of the p chain, and all but the last 6 amino acid residues of the a chain. The pPS15-50 insert codes for all but the first amino
'Met Lys Gly Giy Ile ProI Leu Thr ThrIle Leu Phe Phe Lys Val Asn Ser Thr Glu Thr Thr-
FIG. 1. The pea lectin expressionplasmid. pMS-2 consists of the DNA coding for a segment of amino-terminal signal sequence and the pea lectin @-achain precursor, inserted intothe PIN-I-As expression vector. At the amino terminus of the expressed protein are 6 amino acid residues from the Ipp expression vector and 11 amino acid residues from the pea lectin signal sequence. These are followed by the complete coding region for the @ and cy subunits of the lectin protein. The relevant restriction sites areindicated. Izb, kilobases.
acid residue of the /3 chain, the entire a chain, and contains nucleotides of the 3' untranslated sequence. These two inserts were fused to reconstruct the coding region for a single polypeptide chain as outlined under "Materials and Methods." pMS-2 was produced by splicing the lectin coding region into the expression vector PIN-I-A3 (9). This vector is based on the gene for the outer membrane lipoprotein of E. coli. It has a very strong promotor, A-T-rich regions surrounding and upstream fromthe promotor, and stem-loop structures inthe 3' region (9). These features are thoughtto be important for efficient transcription, translation, and stabilization of the mRNA. As shown in Fig. 1, pMS-2 directs the synthesis of a protein which contains the unprocessed 245-amino acid residue p and a chain sequences, 11amino acid residues from the signal sequence, and 6 amino acid residues derived from the vector. mRNA synthesis and polypeptide chain initiation are directed by signals present in the lpp gene sequence. The initiating methionine residue is situated at thenative position relative to the lpp ribosome binding site, thus insuring efficient translational initiation. Protein synthesis terminates at the UAG codon provided by the lectin gene itself. Expression of Pea Lectin in E. coli-The plasmid pMS-2 was introduced into E. coli strain W3110 bycalcium-mediated DNA transformation. As shown in Fig. 2, lectin synthesis was measured by polyacrylamide gel electrophoresis of cell extracts and subsequent immunoblotting with affinity purified anti-pea lectin antibody. Although the plant produced pea lectin and its precursor are soluble proteins (lo), therecombinant lectin formed insoluble aggregates in the E. coli cell. These aggregates fractionated with the cell debris following lysis (Fig. 2). The differential solubility may have been a function of overexpression in E. coli(23-26) or may have resulted from the presence of additional hydrophobic amino acids at theamino terminus. As shown in Fig. 2, the antibody reactive protein appears as a strong band with a relative molecular mass of about 28,000. This protein is not seen in a
Production of Pea Lectin in E. coli
FIG. 2. Expression of pea lectin in E. coli. A, 12.5% SDSpolyacrylamide gel stained with Coomassie Blue. Lane 1, 25pgof W3110(pIN-I-A3) supernatant. Lane 2, 25pg of W3110(pIN-I-A3) pellet. Lane 3, 25 pg of W3110(pMS-2) supernatant. Lane 4,25 pg of W3110(pMS-2)pellet. Lanes 5-7, 1, 2.5, and 5 pg, respectively, of the p chain of natural pea lectin. The arrow marks the position of the recombinant lectin. The positions of relative molecular mass markers are shown on the right. B, immunoblot of 12.5% SDS-polyacrylamide gel. Lane 1, 2.5 pg of W3llO(pIN-I-A3) supernatant. Lane 2, 2.5 pg of W3110(pIN-I-A3)pellet. Lane 3, 2.5 pg of W3110(pMS-2) supernatant. Lane 4, 2.5 pg of W3110(pMS-2) pellet. Lanes 5-8, 50,100,250, and 500 ng, respectively, of natural pea lectin. The arrow marks the position of the recombinant lectin.
FIG. 3. Purification of pea lectin from E. coli. Protein samples were analyzed by electrophoresis on a 12.5% SDS-polyacrylamide gel stained with Coomassie Blue. Lane 1, pellet from the initial lysate. Lane 2, unbound material from DEAE column. Lane 3, material eluted from Sephadex. The amount of material run in each lane is 1%of1-1 preparation of E. coli. Lane 4, relative molecular mass markers. Lanes 5-7, 2, 4, and 6 pg of concanavalin A, respectively. Concanavalin A is a similar size to the recombinant pea lectin and was used to estimate the amount of recombinant lectin a t various stages of the purification. PBS 32 16
.5 .25 l25.06 -03
E. COLI control preparationof E. coli containing the PIN-I-A3plasmid PRODUCED alone. We estimate that the bacteria produce approximately PEA LECTIN 2-5 mg of lectin/l of cells grown to late log phase or about NATURAL 10-20% of the insoluble E. coli protein. PEA LECTIN Purification of the Expressed Lectin-Aggregation may have I I prevented degradation of the lectin, but it has also necessiFIG.4. Hemagglutination assay comparing natural and E. tated development of procedures to solubilize the protein top two rows contain E. coliaggregates and torecover a functionalproduct. We have found coli-producedpealectins.The produced pea lectin, the bottom two rows contain natural pea lectin. that the aggregates can be solubilized by treatment with 7 M Column 1, PBS. Columns 2-12, doubling dilutions of lectin from 32 guanidinium hydrochloride. Dilution of this extract yields a pglwell. renatured lectin that can be isolated by affinity chromatography on the dextran polymer, Sephadex G-75. Analysis of conclude that the recombinant lectin is multivalent and has the purified lectin by SDS-polyacrylamide gel electrophoresis an affinity for red blood cell surface carbohydrates similar to reveals a single band of 28,000 Da (Fig. 3). The recovery of that of the naturalpea lectin. lectin by this procedure was 20-40%. Preliminary N-terminal Carbohydrate specificity can be established by comparing protein sequence data on the purified pea lectin (data not sugars on the basis of the minimal concentration required to shown) agrees with that deduced from the nucleotide sequence inhibit a hemagglutination reaction (28). The added sugars shown in Fig. 1. compete with the red blood cell surface carbohydrates for the Functional Studies-Although elution from Sephadex with binding sites on the lectin. The concentration at which the a-methyl-D-mannopyranoside is evidence that the E. coli- agglutination reactions are inhibited is proportional to the produced lectin can bind carbohydrate, additional information dissociation constant ( K d ) (29). Standardization with a subwas provided by agglutination of human erythrocytes. Lectin stance of known K d allows approximation of the Kd for a agglutination of red blood cells requires multivalent interac- given sugar using inhibition of agglutination. Table I shows tions with the cell surface carbohydrate (27). As each lectin the Kd values obtained for both natural pea lectin and the monomer has only one carbohydrate binding site, agglutina- recombinant lectin with a variety of sugars. The K d values tion requires that the lectin be in a multimeric form. Fig. 4 were determined by inhibition of hemagglutination using Dshows the results of ahemagglutinationassay comparing mannose as a standard. The inhibitors tested show comparanatural and E. coli-produced pea lectin. In this assay both ble Kd values for the natural pea lectin and for the recombiproteins displayed a titer of-2 pglwell or 10 pg/ml. We nant lectin. The 2-fold difference observed in the case of a-
Production of Pea Lectin inE. coli
TABLE I Dissociation constantsfM) of various substances for natural or E. coli-produced pea lectin The abbreviation used is: Me-. methvl-. Inhibitor
Natural pea lectin"
E. coli produced pea lectin"
D-Man 7.1 X 10-4 7.1 X 10-4 a-MeMan 6.5 -+ 1.6 x 10-4 5.3 f 0.7 X 10-4 D-Gal >1 x 10" >1 x 10-1 1.9 f 0.7 x IOV4 3.7 + 1.0 x 1 0 - ~ Mancul-&(Mancul+3) MannMeb GnGn(F)' 4.3 f 0.7 X 3.7 f 1.0 X "Each value is the mean from at least three experiments. bObtained from Toronto Research Chemicals. GlcNAc terminating fucosylated biantennary glycopeptide purified from human yl-IgG(Tem) myeloma protein (21).
binding site of the legume lectins is formed by fourpolypeptide loops. Alteration of the lectin within these carbohydrate binding loops will allowus to produce lectins with modified specificities. Analysis of such proteins should provide a powerful new approach to the study of the structural basis of carbohydrate bindmg specificity. Acknowledgments-We thank James Dunn for help with DNA sequencing, Thomas Higgins for the cDNA clones, The0 Hofmann for protein sequencing, Masayori Inouye for the expression vectors, Emmanuel Maicas and Gerald Proteau for advice on the use of Bal 31, Alex MacKenzie for concanavalin A, and JimRini for pea lectin. We also thank Robert Allore, Hagan Bayley, Karl Hardman, Alex MacKenzie, Jim Rini, and Doug Romans for helpful discussion.
1. Meehan, E. J., Jr., McDuffie, J., Einspahr, H.,Bugg, C. E. & methyltrimannoside is acceptable within the limits of the Suddath, F. L. (1982) J. Biol. Chem. 2 5 7 , 13278-13282 assay. Thus, in the hemagglutination assay, there is no de2. Cunningham, B. A., Hemperly, J. J., Hopp, T. P. & Edelman, G. tectable difference in the carbohydrate specificity when the M. (1979) Proc. Natl. Acad. Sci. U. S. A. 76,3218-3222 3. Foriers, A., Lebrun, E., Van Rapenbusch, R., de Neve,R. & recombinant lectin is compared to authentic pea lectin. Strosberg, A. D. (1981) J. Biol. Chem. 256,5550-5560 DISCUSSION 4. Olsen, K. W. (1983) Biochem. Biophys. Acta 743,212-218 5. Kornfeld, K., Reitman, M. L. & Kornfeld, R. (1981) J. Biol. The results reported above demonstrate that a functional Chem. 256,6633-6640 plant lectin can be expressed from a cloned gene in E. coli. 6. Baenzinger, J. U. & Fiete, D. (1979) J. Biol. Chem. 2 5 4 , 2400The level of expression is sufficient to permit detection of the 2407 lectin as a major band after electrophoresis of the particulate 7. MacKenzie, A. E. (1985) Ph.D. thesis, University of Toronto fraction of the cell lysate. In initial studies (see Fig. 4 and 8. Higgins, T. J. V., Chandler, P. M., Zurawski, G., Button, S. C. & Spencer, D. (1983) J. Biol. Chem. 258,9544-9549 Table I), lectin purified from this fraction exhibited carbo9. Nakamura, K. & Inouye, M. (1982) EMBO J. 1,771-775 hydrate binding properties comparable to those of natural pea 10. Higgins, T. J. V., Chrispeels, M. J., Chandler, P. M. & Spencer, lectin, although there are several differences between the two D. (1983) J. Biol. Chem. 258,9550-9552 proteins. Natural pea lectin, in its mature form, exists as a 11. Entlicher, G., Kostir, J. V., & Kocourek, J. (1970) Biochim. dimer with each monomer being composed of a /3 and a chain. Biophys. Acta 221,272-281 The lectin purified from E. coli is likely a dimer as well, but 12. Birnboim, M. & Doly, J. (1979) Nucleic Acids Res. 7 , 1513-1523 the p and a chains are uncleaved. The recombinant lectin is 13. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning:A Laboratory Manual, Cold Spring Harbor, New York not processed in the bacterial cell and as a result contains 14, Vogelstein, B. & Gillespie, D. (1979) Proc. Natl. Acad. Sci. U. S. extra amino acid residues. These occur at three sites: 17 A . 76,615-619 additional amino acid residues at the amino terminus, 8 that 15. Mandel, M. & Higa, A. (1970) J. Mol. Biol. 5 3 , 159-162 remain between the p and a sequences, and 2-4 that remain 16. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A . 7 4 , 5463-5467 at thecarboxyl terminus. On the basis of the conformational homology between natural pea lectin and concanavalin A (l), 17. Young, R. A. & Davis, R. W. (1983) Proc. Natl. Acad. Sci. U. S. A . 80,1194-1198 the additional amino acid residues present inthe recombinant 18. Johnson, D. A., Gautsch, J. W., Sportsman, J. R. & Elder, J. H. lectin are known to be at sites remote from the carbohydrate (1984) Gene Anal. Techn. 1,3-8 binding site. Our resultsdemonstrate that the additional 19. Laemmli, U. K.(1970) Nature 227,680-684 residues at these sites do not effect the carbohyrate binding 20. Burnette, W. N. (1981) Anal. Biochem. 112,195-203 21. Grey, A.A., Narasimhan, S., Brisson, J.-R., Schacter, H. & properties of this protein. Carver, J. P. (1982) Can. J. Biochem. 6 0 , 1123-1131 Although the pea lectin is considered to be mannose- or Trowbridge, 1. S. (1974) J. Biol. Chem. 249,6004-6016 glucose-specific, one characteristic which differentiates it 22. 23. Cheng, Y.4. E., Kwoh, D.Y., Kwoh, T. J., Soltvedt, B.C. & from other leguminosae lectins with this specificity (e.g. conZisper, D. (1981) Gene (Amst.)14,121-130 canavalin A) is the increased affinity for N-linked glycopep- 24. Williams, D. C., Van-Frank, R. M., Muth, W. L. & Burnett, J. P. (1982) Science 215,687-689 tides which have Fucal-6 linked to the Asn-GlcNAc (5, 30). For this reason we have compared the affinity of the recom- 25. Kenten, J., Helm, B., lshizaka, T., Cattini, P. & Gould, H. (1984) Proc. Natl. Acad. Sci. U. S. A . 81, 2955-2959 binant and natural pea lectins for such a ligand (Table I). 26. Sumi, S-i., Nagawa, F., Hayashi, T., Amagase, H. & Suzuki, M. Both lectins demonstratedan affinity approximately 100-fold (1984) Gene (Amst.) 29,125-134 higher for the fucosylated glycopeptide compared to a triman- 27. Wang, J. & Edelman, G.M. (1978) J. Biol. Chem. 253, 3000nosy1 core analogue. Thus, the two lectins are the same with 3007 respect to this important feature of the binding of pea lectin 28. Goldstein, 1. J. & Hayes, C. E. (1978) Adu.Carbohydr.Chem. Biochem. 35,127-340 to N-linked glycopeptides. From these results we conclude that thecarbohydrate specificity of the recombinant lectin is 29. Loontiens, F. G., Van Wauwe, J. P. & De Bruyne, C. K. (1975) Carbohydr. Res.44,150-153 very simlar, if not identical, to thatof natural pea lectin. 30. Debray, H., Decout, D., Strecker, G., Spik, G. & Montreuil, J. We are now in a position to alter either individual amino (1981) Eur. J. Biochem. 117, 41-55 acid residues or groups of amino acid residues in thepea lectin 31. Carver, J. P., MacKenzie, A. E. & Hardman, K. D. (1985) Biopolymers 24,49-63 gene. According to a proposed model (31), the glycopeptide