Structure-Function Analysis of the Human Sialyltransferase ST3Gal I

Abstract

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. C. J. A. Biochem. and B. J. Biol. 2003; Scholar). In both the of that is the that the structural data have to on the role by short conserved peptide as the This motif has been in different glycosyltransferase families C. R.A. A. J. Biochem. 1998; Scholar, C. A. Biol. Scholar, S. S. A. 1998; Scholar) and was in several crystal to mainly with the groups of nucleotide donor through the of a is by amino and located in a two and enzymes this motif have an of to be In the of crystal can be used to in the folding properties of other The of or that other families could fall into one of these two structural families C. Heissigerova H. C. J. A. Biochem. Scholar, J. Biol. 2001; Scholar). The sialyltransferase family was among families for which it was to with a high of which of the two known they could However, analyses the of a C. Heissigerova H. C. J. A. Biochem. Scholar). In the present an sequence analysis of all the known to into the of this large and important glycosyltransferase this a new peptide motif located between the sialylmotifs S and To the importance of this new motif and of the in the enzymatic properties of a site-directed mutagenesis was using the human ST3Gal I, as a protein model. ST3Gal I to the residue of type found on or Kinetic have shown that ST3Gal I exhibits high transfer and high affinity of toward the type M. S. Glycobiology. 1997; 7: Scholar). this enzyme was as a for and for Our results to the nucleotide sugar and acceptor binding regions in the to the of each of the four predicted N-glycan attachment sites present in the catalytic domain of I on the and enzyme activity of the acid secondary were from was from was from for were from and were from and the mutagenesis from The of of was from and from both from S. A. S was from and from The were from and from of of was a from C. of of I in was used to and mutant of human I with an N-terminal and in to the and of the Two the or amino were by using as template the the human ST3Gal I gene V. R. R. Delannoy P. Harduin-Lepers A. Biophys. 2001; Scholar). The and regions to the of I were using the for and for and the were to and sites at each of the The were by using the in with each at of at and of at The were with and and cloned into the sites of the transfer M. S. J. J. H. J. 1998; Scholar) to the and proteins by these a peptide at their terminus and be from were used for the and of The were at in with and The of with the transfer and the was to the of were by at in the with and were generally used for the of and mutant of human ST3Gal I. were used to at a of of 5 was after by and the were at they were of I were by using the as the mutagenesis was used for all The used to mutants of I are in I. All of the were into were by of the used for site-directed in a new of were for at 4 with of affinity or or in to the were with into of were on and to a The was by of the followed by secondary and the protein were using the were at for 20 in a of in the presence of of donor substrate CMP-Neu5Ac, of of acceptor substrate in The was by of the enzyme and by of of were on and with The was by The for CMP-Neu5Ac was using of CMP-Neu5Ac with of and for the using of with 200 of CMP-Neu5Ac. of mutant and wild type ST activity was for protein enzyme assessed by of and mutant All enzyme assays were in sequences were from or and using J. 1997; Scholar), Scholar), and Scholar) The was used to protein sequences with low of sequence C. V. Scholar). of eukaryotic sialyltransferases share a unique feature, the presence in their catalytic domain of several conserved peptide regions referred to as sialylmotifs L, S, and for these peptide there are sequence between the various groups of (ST6Gal, ST6GalNAc, ST3Gal, and of all the known ST protein sequences conserved motif located between the the sialylmotifs S and VS The sequence consensus of this motif, in aromatic residues, is as letters and lowercase letters indicate a strong or low occurrence of the amino acid, striking is the presence of the invariant Tyr residue at the second the sialylmotifs S, and the new one a large of the C terminus of the catalytic domain amino in the the sialylmotifs be as in The 20 distinct ST sequences that have been cloned to have in common the presence of invariant residues and conserved or located in these The of the most conserved residues in sialylmotifs L and S and has been using ST6Gal I as a (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 the functional of residues of and a of mutants of the human ST3Gal I. residue of motif was as well as the two invariant amino of motif 4 and of I and in of the human ST3Gal I and its a an N-terminal and was used for of a of the for human ST3Gal I. Two different of the enzyme were after the transmembrane domain after the residue and of the amino which was recently shown to be the active in V. R. R. Delannoy P. Harduin-Lepers A. Biophys. 2001; Scholar). The of protein in the of was by using an and of the proteins in the with different affinity and this that was the most affinity to the proteins since of enzyme activity was for the other affinity However, also observed a and binding of the proteins on was only with a large of and the enzyme with the to In contrast, a and binding was using were used to the of proteins whereas were used for the functional of mutants to to The two and were in the of or with the Although they in the number of N-glycosylation sites for and 4 for they an of since with can be by the and the enzyme activity differences between the two a activity being observed for in the in was to the of the most conserved residues of and of and in eight of the mutants were in the and most of them are at protein to the enzyme. of and that the mutation affect the of the enzyme. one mutant was in the of The mutation of this Tyr residue in in a The of proteins were for a activity from using only the showed no sialyltransferase activity The yielded an activity of shown in the substitution in motif of the highly conserved residue and of the invariant Tyr residue in in a of enzyme activity. The mutation was by the that ST6Gal I is the only ST a Tyr of a residue at this amino acid low activity to the is observed for the mutant a role of the aromatic at this the mutation yielded a inactive the mutation enzyme activity activity to the results to the of the mutants strongly suggest that the invariant Tyr residue of motif plays an important conformational role mainly attributable to the aromatic ring. The of motif is a conserved aromatic residue or The substitution in reduced enzyme activity to a low significant of activity whereas the mutation had little effect on enzyme activity The of motif is conserved and the mutation of this residue in affect enzyme activity of the wild type). results demonstrated the importance of aromatic amino in motif in the catalytic activity. 4 comprised invariant residues and by four The mutation yielded a inactive the substitution E321Q only reduced enzyme activity. of on to the role of conserved residues in and 4 in nucleotide the of the wild type and ST3Gal I mutants to to mutant proteins that were shown to be inactive or active and demonstrated the or binding to the wild type enzyme. The mutant proteins that retained significant enzyme activity E302Q, and also demonstrated significant binding to all of these mutants the binding was of CDP the mutation of the invariant residue Tyr300 strongly CDP as a conformational role is for this residue and therefore it could a role in maintaining the active conformation of the enzyme without any with the nucleotide these results suggest that none of the conserved amino acid in and 4 plays a role in nucleotide binding. Kinetic of the and for the donor and acceptor were for the enzyme and for the four mutants that retained enzyme activity W301F, E302Q, and The for CMP-Neu5Ac and the acceptor of I are and are in with the data previously with the enzyme in M. S. Glycobiology. 1997; 7: Scholar, V. R. R. Delannoy P. Harduin-Lepers A. Biophys. 2001; Scholar). toward the donor substrate have been for the four mutants tested the of these mutants in CDP binding In contrast, the for the acceptor substrate were for the and E321Q mutants with in the of these results strongly suggest that the conserved sialylmotifs and 4 are involved in the binding of the acceptor The in activity observed for the mutant can be to a altering the catalytic of the enzyme. This residue participate in donor and acceptor substrate it has a in maintaining an active enzyme conformation as suggested analysis of wild type and mutants of in a new The of in and of of the primary sequence of I indicates the presence of N-glycosylation sites of them and are conserved in all of the known ST3Gal I The N-glycan attachment located after the transmembrane domain is for activity since the demonstrated the activity as the longest The importance of each of the four N-glycan attachment sites in enzyme was N-glycosylation sites were by in the by the amino in I. This was by the nature of the substitution or by the presence of the amino acid in the other ST3Gal I sequences and or in the ST3Gal II sequences mutations were to have on the conformation of the of mutants with single and N-glycosylation has been by site-directed mutagenesis using the The mutations of one of the two conserved N-glycan sites and as well as the mutation affect the protein or the activity All mutant proteins can be by in the However, the and the mutant the N-glycan sites, and were mutants were to a which to the protein the N-terminal their low of these mutants significant enzyme activity. results suggest that N-glycosylation is necessary for enzyme that it to the proper folding and to the of the Comparison of the indicates a of the which that each of the N-glycan attachment predicted in the catalytic domain, is potentially The of the present was to into in the large ST to into to which the sequence between the different members of this large family which by from an was by the to transfer sialic acid to a large of acceptor with different of Comparison of the peptide sequences showed that there is little sequence between the different ST with the of consensus sequences, the sialylmotifs L, S, and VS (6Datta A.K. Paulson J.C. Ind. J. Biochem. Biophys. 1997; 34: 157-165Google Scholar, R.A. Harduin-Lepers A. Delannoy P. Glycobiology. 1997; 7: v-viiGoogle Scholar). the presence of an motif located between sialylmotifs S and VS with the peptide as The presence of this motif in in sialyltransferases has been with a different that the a different sequence shown in the in this region J. Scholar). the human ST3Gal I as a results demonstrated the importance of aromatic amino in this region of its catalytic Kinetic suggested their involvement in acceptor and to catalytic efficiency. an important structural role is for the invariant Tyr residue In a recent a of mutants of the human ST6Gal I was by mutagenesis P. R. Glycobiology. 2001; Scholar) and using a high enzyme with as Two of these mutants were shown to be affected in residues of the motif in I, see ST6Gal I is the only enzyme to a at the of motif The following and in I in which significant activity toward and as well as a to to the glycoprotein acceptor results that aromatic residues are at these two for catalytic efficiency. the strong of aromatic residues in motif in the ST family with detailed analysis strongly suggest an important functional and/or structural role of these residues in enzyme The invariant and residues of VS 4 in were previously to be important for catalytic activity (7Geremia R.A. Harduin-Lepers A. Delannoy P. Glycobiology. 1997; 7: v-viiGoogle Scholar). The functional role of residue has been recently in the ST8Sia II and IV S. S. M. J. Biol. Chem. 2001; 276: Scholar). In both enzymes the of this by was shown to be critical for their catalytic activity since they no enzyme activity. the proteins showed binding to as wild type enzyme. In the present results were with the mutation of the invariant residue of motif 4 in I the of its involvement in In contrast, the of the conserved residue by the enzyme activity with the wild type). the of motif in all the characterized to date, as well as its predicted occurrence as a of an strongly suggest that both residues are necessary for catalytic of the transfer these observations and it is that motif 4 is of the active mainly located on the acceptor side or at the of both donor and acceptor sugar substrates. the present and mutagenesis (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, P. R. Glycobiology. 2001; Scholar, S. S. M. J. Biol. Chem. 2001; 276: Scholar) it is that the C-terminal of the catalytic domain and 4 and of motif is to the of the acceptor substrate whereas the N-terminal is involved in nucleotide sugar binding However, the that residues of L could be involved in acceptor (or modulate the acceptor as suggested in (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). Recent data the presence of an intramolecular disulfide the two conserved cysteine of and (11Datta A.K. Chammas R. Paulson J.C. J. Biol. Chem. 2001; 276: 15200-15207Google Scholar, K. Yen T.Y. El-Battari A. Macher B.A. Fukuda M. J. Biol. Chem. 2001; 276: 15369-15377Google Scholar). the occurrence of disulfide is a common to all remains to be this in sialylmotifs L and S and residues at the terminus of the catalytic domain the catalytic The catalytic domain has been defined for a or by sequence (4Donadio S. Dubois C. Fichant G. Roybon L. Guillemot J.C. Breton C. Ronin C. Biochimie (Paris). 2003; 85: 311-321Google Scholar, V. R. R. Delannoy P. Harduin-Lepers A. Biophys. 2001; Scholar, C. Colley K.J. Glycobiology. Scholar). The to the stem region from the catalytic domain generally for each with the of a region in the N-terminal of the S. Dubois C. Fichant G. Roybon L. Guillemot J.C. Breton C. Ronin C. Biochimie (Paris). 2003; 85: 311-321Google Scholar). The catalytic domain is therefore to residues from motif residues for the which to the ST6GalNAc III to can the of the catalytic domain residues, with the sialylmotifs located as in are often The role of N-glycosylation sites and their has been studied in several glycosyltransferases and it that it from protein to the it was shown for ST6Gal I that N-glycosylation the protein is for activity of the enzyme in vivo C. Colley K.J. Glycobiology. Scholar). In the case of the were the and N-glycosylation and proper to be critical for proper trafficking of to the Golgi J. Biol. Chem. 1998; 273: Scholar). this for the case of human ST3Gal I primary sequence N-glycosylation Recent data demonstrated that each of the N-glycan attachment sites is potentially in the enzyme in V. R. R. Delannoy P. Harduin-Lepers A. Biophys. 2001; Scholar). using the active of that none of the is essential for enzyme activity and that the mutant enzyme activity. these mutants are and as the wild type a role of N-glycosylation in the proper folding and trafficking of the enzyme as shown for other in the and of some of the active residues, no has been for a of the ST share the presence of four conserved the catalytic domain, a is for the ST The sialylmotifs are highly for this family of eukaryotic and were to any sequence or structural and the with other protein and also In structural and functional information a protein can be predicted by of using or through the of In the has that the of glycosyltransferase families fall only into two structural families GT-A (or SpsA fold) and GT-B (or BGT fold) C. Heissigerova H. C. J. A. Biochem. Scholar, B. J. Biol. 2003; Scholar, J. Biol. 2001; Scholar). two superfamilies have different and different active sites, in both the nucleotide binding domain was shown to a or a structural motif found in nucleotide binding Recent analyses also the of at one domain in the catalytic domain of C. Heissigerova H. C. J. A. Biochem. Scholar). is highly to the of ST by to the various and is in to of I. a be for detailed and for the for donor and acceptor substrate