Portal de Periódicos da CAPES

Structure-Function Analysis of the Human Sialyltransferase ST3Gal I

2004; Elsevier BV; Volume: 279; Issue: 14 Linguagem: Inglês

10.1074/jbc.m311764200

1083-351X

Charlotte Jeanneau, V. Chazalet, Claudine Augé, Dikeos Mario Soumpasis, Anne Harduin‐Lepers, Philippe Delannoy, Anne Imberty, Christelle Breton,

Galectins and Cancer Biology

All eukaryotic sialyltransferases have in common the presence in their catalytic domain of several conserved peptide regions (sialylmotifs L, S, and VS). Functional analysis of sialylmotifs L and S previously demonstrated their involvement in the binding of donor and acceptor substrates. The region comprised between the sialylmotifs S and VS contains a stretch of four highly conserved residues, with the following consensus sequence (H/y)Y(Y/F/W/h)(E/D/q/g). (Capital letters and lowercase letters indicate a strong or low occurrence of the amino acid, respectively.) The functional importance of these residues and of the conserved residues of motif VS (HX4E) was assessed using as a template the human ST3Gal I. Mutational analysis showed that residues His299 and Tyr300 of the new motif, and His316 of the VS motif, are essential for activity since their substitution by alanine yielded inactive enzymes. Our results suggest that the invariant Tyr residue (Tyr300) plays an important conformational role mainly attributable to the aromatic ring. In contrast, the mutants W301F, E302Q, and E321Q retained significant enzyme activity (25–80% of the wild type). Kinetic analyses and CDP binding assays showed that none of the mutants tested had any significant effect in nucleotide donor binding. Instead the mutant proteins were affected in their binding to the acceptor and/or demonstrated lower catalytic efficiency. Although the human ST3Gal I has four N-glycan attachment sites in its catalytic domain that are potentially glycosylated, none of them was shown to be necessary for enzyme activity. However, N-glycosylation appears to contribute to the proper folding and trafficking of the enzyme. All eukaryotic sialyltransferases have in common the presence in their catalytic domain of several conserved peptide regions (sialylmotifs L, S, and VS). Functional analysis of sialylmotifs L and S previously demonstrated their involvement in the binding of donor and acceptor substrates. The region comprised between the sialylmotifs S and VS contains a stretch of four highly conserved residues, with the following consensus sequence (H/y)Y(Y/F/W/h)(E/D/q/g). (Capital letters and lowercase letters indicate a strong or low occurrence of the amino acid, respectively.) The functional importance of these residues and of the conserved residues of motif VS (HX4E) was assessed using as a template the human ST3Gal I. Mutational analysis showed that residues His299 and Tyr300 of the new motif, and His316 of the VS motif, are essential for activity since their substitution by alanine yielded inactive enzymes. Our results suggest that the invariant Tyr residue (Tyr300) plays an important conformational role mainly attributable to the aromatic ring. In contrast, the mutants W301F, E302Q, and E321Q retained significant enzyme activity (25–80% of the wild type). Kinetic analyses and CDP binding assays showed that none of the mutants tested had any significant effect in nucleotide donor binding. Instead the mutant proteins were affected in their binding to the acceptor and/or demonstrated lower catalytic efficiency. Although the human ST3Gal I has four N-glycan attachment sites in its catalytic domain that are potentially glycosylated, none of them was shown to be necessary for enzyme activity. However, N-glycosylation appears to contribute to the proper folding and trafficking of the enzyme. The sialyltransferase (ST) 1The abbreviations used are: ST, sialyltransferase; CMP-Neu5Ac, CMP-N-acetylneuraminic acid; ST3Gal I, CMP-Neu5Ac:Galβ1–3GalNAc α2,3-sialyltransferase I (EC 2.4.99.4). gene family represents a group of enzymes that transfer sialic acid from CMP-Neu5Ac to carbohydrate groups of various glycoproteins and glycolipids. To date, 20 distinct protein members have been cloned and characterized (for a recent review, see Ref. 1Harduin-Lepers A. Vallejo-Ruiz V. Krzewinski-Recchi M.A. Samyn-Petit B. Julien S. Delannoy P. Biochimie (Paris). 2001; 83: 727-737Google Scholar). They are classically split into 4 groups, depending on the type of linkage formed and the nature of the sugar acceptor (ST6Gal, ST6GalNAc, ST3Gal, and ST8Sia) (2Tsuji S. Datta A.K. Paulson J.C. Glycobiology. 1996; 6: v-viiGoogle Scholar). The sialyltransferases are localized in the Golgi apparatus and they share with the other Golgi-resident glycosyltransferases a typical type II architecture consisting in a short N-terminal cytoplasmic tail, a transmembrane domain followed by a stem region, and a large C-terminal catalytic domain facing the luminal side (3Paulson J.C. Colley K.J. J. Biol. Chem. 1989; 264: 17615-17618Google Scholar). Comparison of peptide sequences strongly indicates that the length and amino acid composition of catalytic domains are relatively well conserved and variations in protein sizes are generally attributable to differences in the length of the stem region. The stem region can be defined as the peptide portion after the transmembrane domain that can be removed without altering the activity. The most striking differences are observed in the ST6GalNAc subfamily where ST6GalNAc I exhibits the longest stem region (about 200 amino acids) and ST6GalNAc III appears to be devoid of stem region. In this group, it appears that the broader the acceptor specificity, the longer the stem region (1Harduin-Lepers A. Vallejo-Ruiz V. Krzewinski-Recchi M.A. Samyn-Petit B. Julien S. Delannoy P. Biochimie (Paris). 2001; 83: 727-737Google Scholar). The stem region often displays high variability in amino acid composition and little secondary organization and was therefore predicted to be flexible (4Donadio S. Dubois C. Fichant G. Roybon L. Guillemot J.C. Breton C. Ronin C. Biochimie (Paris). 2003; 85: 311-321Google Scholar). However, this peptide portion often contains cysteine residues as well as several N- and O-glycosylation sites, which could contribute to a local conformation. Recent data suggested that the stem portion could also modulate the in vivo acceptor specificity (5Legaigneur P. Breton C. El Battari A. Guillemot J.C. Auge C. Malissard M. Berger E.G. Ronin C. J. Biol. Chem. 2001; 276: 21608-21617Google Scholar). All eukaryotic sialyltransferases share a unique feature, the presence, in their catalytic domain, of several conserved peptide regions referred to as sialylmotifs L and S (6Datta A.K. Paulson J.C. Ind. J. Biochem. Biophys. 1997; 34: 157-165Google Scholar) and VS (7Geremia R.A. Harduin-Lepers A. Delannoy P. Glycobiology. 1997; 7: v-viiGoogle Scholar). The functional significance of the sialylmotifs L and S has been assessed by site-directed mutagenesis, using ST6Gal I as a model. The mutagenesis of the most conserved residues led to the conclusion that the L-sialylmotif is mainly involved in donor substrate binding (8Datta A.K. Paulson J.C. J. Biol. Chem. 1995; 270: 1497-1500Google Scholar), whereas mutations in the S-sialylmotif were shown to affect both donor and acceptor binding (9Datta A.K. Sinha A. Paulson J.C. J. Biol. Chem. 1998; 273: 9608-9614Google Scholar). Considering all known ST sequences, the number of invariant residues in sialylmotifs L and S is 5 and 2, respectively, one cysteine residue being conserved in each region. Mutation of the two conserved cysteines yielded inactive enzymes (8Datta A.K. Paulson J.C. J. Biol. Chem. 1995; 270: 1497-1500Google Scholar, 9Datta A.K. Sinha A. Paulson J.C. J. Biol. Chem. 1998; 273: 9608-9614Google Scholar, 10Drickamer K. Glycobiology. 1993; 3: 2-3Google Scholar) and recent data suggest that they participate in the formation of an intramolecular disulfide linkage that is essential for maintaining an active conformation of the enzyme (11Datta A.K. Chammas R. Paulson J.C. J. Biol. Chem. 2001; 276: 15200-15207Google Scholar). Similar observations were made with the polysialyltransferase ST8Sia IV (12Angata K. Yen T.Y. El-Battari A. Macher B.A. Fukuda M. J. Biol. Chem. 2001; 276: 15369-15377Google Scholar). In the latter case, a second intramolecular disulfide bond, which brings the sialylmotifs and the C terminus within proximity, has been evidenced. Formation of dimer through disulfide bonds has also been demonstrated in the case of ST6Gal I (13Ma J. Colley K.J. J. Biol. Chem. 1996; 271: 7758-7766Google Scholar). This dimer comprises ∼20–30% of the enzyme found in liver Golgi and exhibits reduced catalytic activity because of its lower affinity for the sugar nucleotide donor, CMP-Neu5Ac. Recent data demonstrated that the mutation of a single Cys residue (Cys24) in the transmembrane domain of ST6Gal I abolishes dimerization of the enzyme (14Qian R. Chen C. Colley K.J. J. Biol. Chem. 2001; 276: 28641-28649Google Scholar). Two other Cys residues located downstream the S-sialylmotif could also be critical for in vivo enzyme activity (14Qian R. Chen C. Colley K.J. J. Biol. Chem. 2001; 276: 28641-28649Google Scholar). The enzymatic properties of sialyltransferases have been extensively studied in terms of substrate specificities toward synthetic acceptors as well as their glycoprotein and glycolipid acceptor preference, which revealed the exquisite specificity of some of them. However, at the present time, there is no structural information for sialyltransferases and the detailed mechanism of action remains unclear. The crystal structures of sixteen glycosyltransferases belonging to distinct sequence-based families have been recently solved and among them eight are of mammalian origin. Although belonging to different glycosyltransferase families showing no primary sequence identity, these protein structures fall into only two different structural superfamilies named GT-A (or SpsA fold) and GT-B (or BGT fold) (recently reviewed in Refs. 15Breton C. Heissigerova H. Jeanneau C. Moravcova J. Imberty A. Biochem. Soc. Symp. 2002; 69: 23-32Google Scholar and 16Coutinho P.M. Deleury E. Davies G.J. Henrissat B. J. Mol. Biol. 2003; 328: 307-317Google Scholar). In addition, both topologies exhibit the same class of fold; that is the threelayer α/β/α sandwich that resembles the "Rossmann fold." These structural data have begun to shed light on the role played by short conserved peptide motifs, such as the DXD motif. This motif has been identified in many different glycosyltransferase families (17Breton C. Bettler E. Joziasse D.H. Geremia R.A. Imberty A. J. Biochem. 1998; 123: 1000-1009Google Scholar, 18Breton C. Imberty A. Curr. Opin. Struct. Biol. 1999; 9: 563-571Google Scholar, 19Wiggins C.A. Munro S. Proc. Natl. Acad. Sci., U. S. A. 1998; 95: 7945-7950Google Scholar) and was shown, in several crystal structures, to interact mainly with the phosphate groups of nucleotide donor through the coordination of a metal cation. It is always flanked by apolar amino acids and located in a loop (β-turn) connecting two β-strands, and enzymes sharing this motif have an absolute requirement of divalent cation to be active. In the absence of crystal structures, alternative methods can be used to get insight in the folding properties of other GT families. The use of "fold recognition" or "threading" methods suggests that many other GT families could fall into one of these two structural families (15Breton C. Heissigerova H. Jeanneau C. Moravcova J. Imberty A. Biochem. Soc. Symp. 2002; 69: 23-32Google Scholar, 20Wrabl J.O. Grishin N.V. J. Mol. Biol. 2001; 314: 365-374Google Scholar). The sialyltransferase family was among those families for which it was not possible to predict with a high level of confidence which of the two currently known folds they could adopt. However, threading analyses favored the existence of a Rossmann fold (15Breton C. Heissigerova H. Jeanneau C. Moravcova J. Imberty A. Biochem. Soc. Symp. 2002; 69: 23-32Google Scholar). In the present study, we performed an extensive sequence analysis of all the known animal STs to gain further insight into the structure/function relationships of this large and biologically important glycosyltransferase family. During this work we evidenced a new peptide motif located between the sialylmotifs S and VS. To address the importance of this new motif and of the VS-sialylmotif in the enzymatic properties of STs, a site-directed mutagenesis was performed using the human ST3Gal I, as a protein model. ST3Gal I transfers Neu5Ac to the galactose residue of type 3 disaccharide found on glycolipids or O-glycosyl proteins. Kinetic studies have shown that ST3Gal I exhibits high transfer efficiency and high affinity (Km of 51 μm) toward the core 1 mucin type disaccharide Galβ1–3GalNAcα- (21Kono M. Ohyama Y. Lee Y.C. Hamamoto T. Kojima N. Tsuji S. Glycobiology. 1997; 7: 469-479Google Scholar). Therefore, this enzyme was considered as a good candidate for structure-function studies and for further crystallization trials. Our results contributed to delineate more precisely the nucleotide sugar and acceptor binding regions in the protein. Additional experiments allowed to determine the impact of each of the four predicted N-glycan attachment sites present in the catalytic domain of hST3Gal I on the glycosylation status and enzyme activity of the protein. Materials—Cytidine 5′-monophosphono-N-acetyl neuraminic acid (CMP-Neu5Ac), peroxidase-conjugated goat anti-mouse secondary antibody, gentamicin, l-glutamine, fetal bovine serum, and, AEC staining kit were obtained from Sigma. CMP-[14C]Neu5Ac (289 mCi/mmol) was obtained from Amersham Biosciences (Orsay, France), Galβ1–3Gal-NAcα-sp-biotin was from Lectinity Holdings, Inc (Moscow, Russia). Oligonucleotides for PCR were purchased from MWG Biotech (Courtaboeuf, France). Anti-Xpress antibody, restriction enzymes, insect cells and culture media were from Invitrogen (Cergy Pontoise, France), Escherichia coli XL1-Blue cells and the QuickChange site directed mutagenesis kit from Stratagene (Amsterdam-Zuidoost, The Netherlands). CDP-fractogel (15 μmol of CDP/ml of gel) was from Calbiochem and His-Bind Resin from Novagen, both purchased from VWR international S. A. S (Fontenay-sous-Bois, France). Pfu DNA polymerase was from Promega (Charbonnières-les-Bains, France) and BaculoGold expression system from BD Pharmingen. The nitrocellulose membranes were from Pall Gelman Laboratory (St Germain-en-Laye, France) and C18 SepPak cartridges from Millipore Corp. CDP-hexanolamine-Sepharose (4 μmol of CDP/ml of gel) was a gift from Dr C. Augé (University of Paris-Sud, France). Overexpression of hST3Gal I in Insect Cells—Baculovirus-mediated insect cell expression was used to express native and mutant soluble forms of human hST3Gal I with an N-terminal His6 tag and X-Presss™ epitope in order to facilitate the detection and further purification of the recombinant protein. Two cDNA fragments lacking the first 25 (ST3G-Δ25) or 56 (ST3G-Δ56) amino acids were generated by PCR using as template the plasmid pFlagST3G harboring the human ST3Gal I gene (22Vallejo-Ruiz V. Haque R. Mir A-M. Schwientek T. Mandel U. Cacan R. Delannoy P. Harduin-Lepers A. Biochim. Biophys. Acta. 2001; 1549: 161-173Google Scholar). The 951-bp and 860-bp coding regions corresponding to the desired truncated soluble forms of hST3Gal I were obtained using the forward primers (5′-AGACGCGGCCGCTAACTACTCCCACACCATGG-3′) for hST3G-Δ25 and (5′-AAGGGAATTCGCGTCATCTCCCCTTGAAG-3′) for hST3G-Δ56 and the same reverse primer (5′-ACTGGCGGCCGCCAGGCCTTGCACCTGCACCC-3′). Primers were designed to create NotI and EcoRI restriction sites at each end of the gene. The PCR products were obtained by using the Pfu DNA polymerase in 30 cycles, with each cycle comprising 45 s at 94 °C, 1 min of annealing at 50 °C, and 3 min of elongation at 70 °C. The PCR fragments were excised with NotI and EcoRI and cloned into the NotI/EcoRI sites of the baculovirus transfer vector, pVTBac-His (23Sarkar M. Pagny S. Unligil U. Joziasse D. Mucha J. Glossl J. Schachter H. Glycoconj. J. 1998; 15: 193-197Google Scholar) to give the plasmids pVT-ST3G-Δ25 and pVT-ST3G-Δ56. Recombinant proteins encoded by these vectors will contain a melittin cleavable signal peptide at their N terminus and be secreted from baculovirus-infected cells. Spodoptera frugiperda cells (Sf9) were used for the production and amplification of recombinant baculoviruses. The cells were cultured at 27 °C in Grace's medium supplemented with 10% fetal bovine serum, and 50 μg/ml gentamicin. The co-transfection of Sf9 cells with the transfer vectors and the BaculoGold linear DNA was done according to the manufacturer's instructions. Viral stocks of 108 plaque-forming units (pfu)/ml were prepared by repeated amplification. Trichoplusia ni (High Five™) cells grown at 27 °C in the protein-free Express Five™ SFM medium, supplemented with 1 mm l-glutamine and 50 μg/ml gentamicin, were generally used for the production of recombinant native and mutant forms of human ST3Gal I. Recombinant baculoviruses were used to infect High Five™ cells at a multiplicity of infection of 5 pfu per cell. Medium was collected 96 h after infection, clarified by centrifugation, and the supernatants were stored at –70 °C until they were used. Site-directed Mutagenesis—Mutant forms of hST3Gal I were prepared by using the plasmid pVT-ST3G-Δ56 as the template. PCR-based mutagenesis was used for all mutations. The primers used to create mutants of hST3Gal I are described in Table I. All of the recombinant plasmids were propagated into E. coli XL1-Blue cells. Mutants were systematically checked by sequencing.Table ISequence of the primers used for site-directed mutagenesisMutantSequencePrimerH299A5′-GCAAAGGGAACTGGCACGCCTACTGGGAGAACAACCC-3′(Sense)5′-GGGTTGTTCTCCCAGTAGGCGTGCCAGTTCCCTTTGC-3′(Antisense)H299Y5′-GCAAAGGGAACTGGCACTACTACTGGGAGAACAACCC-3′(Sense)5′-GGGTTGTTCTCCCAGTAGTAGTGCCAGTTCCCTTTGC-3′(Antisense)Y300A5′-GGAACTGGCACCACGCCTGGGAGAACAACCC-3′(Sense)5′-GGGTTGTTCTCCCAGGCGTGGTGCCAGTTCC-3′(Antisense)Y300F5′-GGAACTGGCACCACTTCTGGGAGAACAACCC-3′(Sense)5′-GGGTTGTTCTCCCAGAAGTGGTGCCAGTTCC-3′(Antisense)W301A5′-GGAACTGGCACCACTACGCGGAGAACAACCCATCC-3′(Sense)5′-GGATGGGTTGTTCTCCGCGTAGTGGTGCCAGTTC-3′(Antisense)W301F5′-GGAACTGGCACCACTACTACTTTGAGAACAACCC-3′(Sense)5′-GGGTTGTTCTCAAAGTAGTGGTGCCAGTTCC-3′(Antisense)E302Q5′-GGCACCACTACTGGCAGAACAACCCATCCGC-3′(Sense)5′-GCGGATGGGTTGTTCTGCCAGTAGTGGTGCC-3′(Antisense)H316A5′-GCAAGACGGGGGTGGCCGATGCAGACTTTG-3′(Sense)5′-CAAAGTCTGCATCGGCCACCCCCGTCTTGC-3′(Antisense)E321Q5′-CGATGCAGACTTTCAGTCTAAAGTCTGCATCG-3′(Sense)5′-CCGTCACGTTAGACTGAAAGTCTGCATCG-3′(Antisense)N79D5′-CGATGAGAGGTTCGACCAGACCATGCAGCCG-3′(Sense)5′-CGGCTGCATGGTCTGGTCGAACCTCTCATC-3′(Antisense)N114S5′-GCCCAATAACTTGAGTGACACCATCAAGG-3′(Sense)5′-CCTTGATGGTGTCACTCAAGTTATTGGGC-3′(Antisense)N201Q5′-CGGGAGCTGGGAGATCAGGTCAGCATGATCCTGG-3′(Sense)5′-CCAGGATCATGCTGACCTGATCTCCCAGCTCCCG-3′(Antisense)N323H5′-GCAGACTTTGAGTCTCACGTGACGGCCACCTTG-3′(Sense)5′-CAAGGTGGCCGTCACGTGAGACTCAAAGTCTGC-3′(Antisense) Open table in a new tab Western Blot Analysis—Aliquots (1 ml) of crude High Five™ supernatants were incubated for 1 h at 4 °C with 50 μl of affinity adsorbent (CDP-Fractogel, or CDP-Agarose or His-Bind) in order to trap the recombinant proteins. Beads were then washed twice with phosphatebuffered saline, resuspended into 50 μl of Laemmli denaturing buffer. Protein samples were separated on sodium dodecyl sulfate-10% polyacrylamide gels and electrotransferred to a nitrocellulose membrane. The blot was developed by adsorption of the anti-Xpress antibody (1: 4000) followed by peroxidase-conjugated goat anti-mouse secondary antibody (1:2000), and the protein bands were visualized using the AEC staining kit. Sialyltransferase Assay—Standard reactions were conducted at 37 °C for 20 min in a final volume of 50 μl in the presence of 50 μm of donor substrate CMP-Neu5Ac, 110,000 cpm of CMP-[14C]Neu5Ac, 50 μm of acceptor substrate Galβ1–3GalNAcα-sp-biotin in 0.1 m cacodylate buffer, pH 6.5. The reaction was initiated by addition of the recombinant enzyme source and stopped by addition of 450 μl of cold water. Reaction products were applied on C18 SepPak cartridges and eluted with methanol. The radioactivity was measured by scintillation counting. The apparent Km value for CMP-Neu5Ac was obtained using 1–200 μm of CMP-Neu5Ac with 0.5 mm of acceptor, and for the acceptor, using 2–500 μm of Galβ1–3GalNAcα-sp-biotin with 200 μm of CMP-Neu5Ac. For comparison of mutant and wild type activity, ST activity was normalized for protein expression (relative enzyme mass) assessed by Western blotting of wild-type and mutant proteins. All enzyme assays were done in triplicate. Sequence Analysis—Protein sequences were retrieved from Gen-Pept or SwissProt and analyzed using BLAST (24Altschul S.F. Madden T.L. Schaffer A.A. Zhang J. Zhang Z. Miller W. Lipman D.J. Nucleic Acids Res. 1997; 25: 3389-3402Google Scholar), LALIGN (25Huang X. Miller W. Adv. Appl. Math. 1991; 12: 337-357Google Scholar), and ClustalW (26Thompson J.D. Higgins D.G. Gibson T.J. Nucleic Acids Res. 1994; 22: 4673-4680Google Scholar) programs. The sensitive Hydrophobic Cluster Analysis method (HCA) was used to compare protein sequences with very low level of sequence identity (27Gaboriaud C. Bissery V. Benchetrit T. Mornon J.P. FEBS Lett. 1987; 224: 149-155Google Scholar). Sequence Analysis of Sialyltransferases—All eukaryotic sialyltransferases share a unique feature, the presence in their catalytic domain of several conserved peptide regions commonly referred to as sialylmotifs L, S, and VS. Except for these peptide motifs there are few sequence similarities between the various groups of STs (ST6Gal, ST6GalNAc, ST3Gal, and ST8Sia). Examination of all the known ST protein sequences reveals another conserved motif located between the the sialylmotifs S and VS (Fig. 1). The sequence consensus of this motif, rich in aromatic residues, is as follows (H/y)Y(Y/W/F/h)(D/E/q/g) (capital letters and lowercase letters indicate a strong or low occurrence of the amino acid, respectively). Of striking interest is the presence of the invariant Tyr residue at the second position. Taken together, the sialylmotifs S, VS, and the new one cover a large part of the C terminus of the catalytic domain (55–70 amino acids). For better clarity in the text, the sialylmotifs will be numbered as indicated in Fig. 1. The 20 distinct ST sequences that have been cloned to date have in common the presence of 10 invariant residues and about 30 conserved or semiconserved positions located in these motifs. The function of the most conserved residues in sialylmotifs L and S (motifs 1 and 2) has already been investigated, using ST6Gal I as a model (8Datta A.K. Paulson J.C. J. Biol. Chem. 1995; 270: 1497-1500Google Scholar, 9Datta A.K. Sinha A. Paulson J.C. J. Biol. Chem. 1998; 273: 9608-9614Google Scholar). To elucidate the functional relevance of residues of motifs 3 and 4, we constructed a series of mutants of the human ST3Gal I. Each residue of motif 3 was mutated (His299-Tyr300-Trp301-Glu302) as well as the two invariant amino acids of motif 4 (His316 and Glu321). Expression of hST3Gal I and Mutants in Insect Cells—For expression of the wild-type human ST3Gal I and its mutants, a baculoviral expression plasmid containing an N-terminal signal sequence, (His)6 tag and X-Press tag was used for expression of a soluble form of the cDNA for human ST3Gal I. Two different soluble forms of the native enzyme were generated: ST3G-Δ25 (truncated just after the transmembrane domain after the residue 25) and ST3-Δ56 (deleted of the first 56 amino acids), which was recently shown to be the minimal catalytically active form when expressed in COS-7 cells (22Vallejo-Ruiz V. Haque R. Mir A-M. Schwientek T. Mandel U. Cacan R. Delannoy P. Harduin-Lepers A. Biochim. Biophys. Acta. 2001; 1549: 161-173Google Scholar). The level of protein expression in the culture medium of insect cells was monitored by Western blotting using an anti-X-Press antibody and upon trapping of the recombinant proteins secreted in the culture supernatants with different affinity gels (CDP-beads and His-Bind resin). At this stage, we noticed that CDP-Fractogel was the most efficient affinity system to catch the recombinant proteins since almost 100% of enzyme activity was recovered (only 80–90% for the other affinity sorbents) (data not shown). However, we also observed a tight, almost irreversible, and nonspecific binding of the recombinant proteins on CDP-Fractogel beads (binding was only poorly inhibited with a large excess of CDP, and the enzyme binds with the same efficiency to GDP-Fractogel beads). In contrast, a specific and reversible binding was obtained using homemade CDP-agarose beads. Therefore, CDP-Fractogel beads were used to quantify the amount of secreted recombinant proteins whereas CDP-Agarose beads were used for the functional tests (i.e. capability of mutants to bind to CDP). The two isoforms ST3G-Δ25 and ST3G-Δ56 were equally produced in the culture medium upon transfection of Sf9 or Hi-5™ cells with the recombinant baculoviruses. Although they differ in the number of N-glycosylation sites (5 for ST3G-Δ25 and 4 for ST3G-Δ56) they display an apparent similar pattern of glycosylation since 3 major polypeptide bands with diffuse borders (glycoforms) can be distinguished by Western blotting whatever the construct and the cell line utilized (Fig. 2). As well, enzyme activity measurements did not reveal major differences between the two constructs, a slightly higher specific activity being observed for ST3-Δ56 expressed in Hi-5™ cells (data not shown). Therefore, the ST3-Δ56 construct overexpressed in Hi-5 cells was selected to further explore the function of the most conserved residues of motifs 3 and 4. Characterization of Native and Mutant Enzymes—As seen in Fig. 3A, eight of the nine mutants were efficiently secreted in the culture medium and most of them are produced at protein levels roughly equivalent to the native enzyme. Their similar pattern of migration (molecular mass and glycosylation status) indicated that the introduced point mutation did not affect the overall structure of the enzyme. However one mutant (Y300A) was repeatedly poorly expressed in the culture medium of insect cells. The mutation of this Tyr residue in Phe resulted in a better expression level. The relative amounts of proteins were determined for a comparative activity assay. Medium from cells mock-transfected using only the vector (pVTBac) showed no detectable sialyltransferase activity (data not shown). The native ST3G-Δ56 construct yielded an average transferase activity of 400 nmol·min–1·mg protein–1. As shown in Fig. 3A, the substitution in motif 3 of the highly conserved His residue (H299A) and of the invariant Tyr residue in Ala (Y300A) resulted in a complete loss of enzyme activity. The mutation H299Y was dictated by the fact that ST6Gal I is the only ST displaying a Tyr instead of a His residue at this amino acid position (see Fig. 1). A low but detectable activity (2% relative to the wild-type) is observed for the mutant H299Y, thus suggesting a possible role of the aromatic ring at this position. Interestingly, if the mutation Y300A yielded a completely inactive enzyme, the mutation Y300F restored substantial enzyme activity (30% activity relative to the native enzyme). These results combined to the expression levels of the mutants strongly suggest that the invariant Tyr residue of motif 3 plays an important conformational role mainly attributable to the aromatic ring. The third position of motif 3 is a conserved aromatic residue (Tyr, Phe, Trp, or His). The substitution W301A in ST3G-Δ56 reduced enzyme activity to a low but significant level of activity (2%) whereas the conservative mutation W301F had little effect on enzyme activity (30%). The fourth position of motif 3 is less conserved and the mutation of this residue in Gln (E302Q) did not really affect enzyme activity (80% of the wild type). These results demonstrated the importance of aromatic amino acids in motif 3 in the catalytic activity. Motif 4 comprised 2 invariant residues (His316 and Glu321), separated by four residues. The mutation H316A yielded a totally inactive enzyme, but the conservative substitution E321Q only slightly reduced enzyme activity. Effect of Mutations on Binding to CDP-beads—To determine the role of conserved residues in motifs 3 and 4 in nucleotide binding, we compared the ability of the wild type and ST3Gal I mutants to bind to CDP-beads (Fig. 3B). Four mutant proteins that were shown to be inactive or very poorly active (H299A, H299Y, W301A, and H316A) demonstrated the same or similar binding capacity relative to the wild type enzyme. The mutant proteins that still retained significant enzyme activity (W301F, E302Q, and E321Q) also demonstrated significant binding to CDP-beads. For all of these mutants the binding was inhibited upon addition of free CDP (data not shown). Only the mutation of the invariant residue Tyr300 strongly impaired CDP binding, but as mentioned above, a conformational role is postulated for this residue and therefore it could play a central role in maintaining the active conformation of the enzyme without establishing any contact with the nucleotide donor. Altogether, these results suggest that none of the conserved amino acid positions in motifs 3 and 4 plays a crucial role in nucleotide binding. Kinetic Analysis of the Native and Mutant Enzymes—Kinetic parameters for the donor and acceptor were measured for the native enzyme and for the four mutants that retained enough enzyme activity (Y300F, W301F, E302Q, and E321Q). The apparent Kms for CMP-Neu5Ac and the acceptor Galβ1–3GalNAcα-sp-biotin of native hST3Gal I are 8.5 μm and 15 μm, respectively (Table II). These values are in good agreement with the data previously obtained with the same enzyme expressed in COS cells (21Kono M. Ohyama Y. Lee Y.C. Hamamoto T. Kojima N. Tsuji S. Glycobiology. 1997; 7: 469-479Google Scholar, 22Vallejo-Ruiz V. Haque R. Mir A-M. Schwientek T. Mandel U. Cacan R.