HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 36, 28254-28260, September 8, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/36/28254    most recent M002182200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ouzzine, M. Articles by Fournel-Gigleux, S. Search for Related Content PubMed PubMed Citation Articles by Ouzzine, M. Articles by Fournel-Gigleux, S. Social Bookmarking                      What's this?

Structure/Function of the Human Gal1,3-glucuronosyltransferase

DIMERIZATION AND FUNCTIONAL ACTIVITY ARE MEDIATED BY TWO CRUCIAL CYSTEINE RESIDUES^*

Mohamed Ouzzine, Sandrine Gulberti, Patrick Netter, Jacques Magdalou, and Sylvie Fournel-Gigleux

From the UMR CNRS 7561-Université Henri Poincaré Nancy 1, Faculté de Médecine, BP 184, 54505 Vandoeuvre-lès-Nancy, France

Received for publication, March 15, 2000, and in revised form, May 18, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Gal1,3-glucuronosyltransferase (GlcAT-I) that catalyzes the transfer of a glucuronic acid residue onto the trisaccharide primer of the glycosaminoglycan-protein linkage region plays an essential role in the early steps of the biosynthesis of glycosaminoglycans. In order to gain insight into the structure/function of the enzyme, the human recombinant GlcAT-I was successfully expressed in the yeast Pichia pastoris, with an apparent molecular mass of 43 kDa. Analysis of the electrophoretic mobility of the membrane-bound protein in nonreducing and reducing conditions, together with cross-linking studies, indicated that the membrane-bound GlcAT-I formed active disulfide-linked dimers. GlcAT-I expressed without the predicted N-terminal cytoplasmic tail or secreted as a polypeptide lacking the cytoplasmic tail and transmembrane domain was similarly organized as dimers, suggesting that the structural determinants for the dimerization state are localized in the luminal domain of the protein. In addition, the role of Cys³³ and Cys³⁰¹ in that process was investigated by site-directed mutagenesis combined with chemical modification of GlcAT-I by N-phenylmaleimide. Replacement of Cys³³ with alanine abolished the formation of dimers with a concomitant decrease in the catalytic efficiency mainly due to a decrease in apparent maximal velocity and in affinity for UDP-glucuronic acid. On the other hand, N-phenylmaleimide treatment or alanine substitution of the Cys³⁰¹ residue inactivated the enzyme. Our study demonstrates that GlcAT-I is organized as a homodimer as a result of disulfide bond formation mediated by Cys³³ localized in the stem region, whereas the residue Cys³⁰¹ localized in a conserved C-terminal domain is strictly required for the functional integrity of the enzyme.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Glycosaminoglycans (GAG)¹ are linear polysaccharides of various length and nature, generally attached covalently to core proteins to form proteoglycans. They are ubiquitously distributed on the surface of most cells and in the extracellular matrix, playing a pivotal role in the assembly and structural integrity of connective tissues (1). GAG are increasingly implicated as important regulators of many biological processes such as cell adhesion and differentiation, cytokine action, and modulation of enzyme catalysis, owing their activities to interactions with various components of cell surfaces and extracellular matrices through specific saccharide sequences (for review, see Refs. 2-4). Since GAG structures vary considerably during normal embryonic development, growth, and aging, in a tissue-specific manner, GAG chain synthesis is thought to be strictly regulated. In addition, GAG chains vary in size and in number during pathological conditions, leading to the alteration of the structural and functional properties of the tissues.

GAG chains consist of repeating disaccharide units containing hexosamine and hexuronic acid (or in the case of keratan sulfate, galactose). The biosynthesis of hexuronic acid-containing GAG is initiated by the formation of a common carbohydrate sequence, GlcA1,3Gal1,3Gal1,4Xyl-, bound to specific serine residues in the core protein to form the so-called GAG-protein linkage region (1). This structure serves as a primer for chain elongation to form either [-4GlcA1,4GlcNAc1-]_n, the core polymer in heparan sulfate/heparin, or [-4GlcA1,3GalNAc1-]_n in chondroitin sulfate and dermatan sulfate. Subsequent modifications of GAG chains by O-sulfation on different positions and by C₅-epimerization of glucuronic acid to iduronic acid generate a bewildering complexity encoding considerable biological information. Several enzymes involved in the processing of GAG chains have been purified and cloned, such as GlcNAc N-deacetylase/N-sulfotransferases (5) and O-sulfotransferases (6). Relatively little is known about the structure and function of the glucuronosyltransferases that are responsible for the assembly of GAG chains. The final biosynthetic step of the common linkage region is catalyzed by a 1,3-glucuronosyltransferase termed GlcAT-I that transfers a glucuronosyl moiety from UDP-glucuronic acid onto the nonreducing end of the second galactose of the trisaccharide primer. Thus, this enzyme plays a gating role in the overall synthesis of hexuronic-GAG chains. Furthermore, Bai et al. (7) and Salimath et al. (8), using synthetic -D-xyloside precursors added to cultured Chinese hamster ovary cells, showed that once glucuronic acid is transferred to the nascent chain, the intermediate product is efficiently consumed by downstream enzymes in the pathway, suggesting that at least in these cells, GlcAT-I may be rate-limiting. A putative involvement of GlcAT-I in the biosynthesis of the carbohydrate epitope HNK1 (human natural killer cell carbohydrate antigen-1, 3OSO₃GlcA1,3Gal-R) has also been proposed (9, 10).

GlcAT-I activity was first detected in an embryonic chick cartilage extract (11) and was subsequently partially purified from various sources (12, 13). However, attempts to purify GlcAT-I to homogeneity have not been successful due to low concentrations and to the difficulty in solubilizing the enzyme. Recent cloning experiments have yielded human (14) and Chinese hamster (10) GlcAT-I cDNA sequences. The GlcAT-I sequence is highly homologous with GlcAT-P cDNA coding for a brain enzyme responsible for the formation of HNK-1 determinants (15). Subfractionation studies of microsomal membranes of chick embryo epiphyseal cartilage indicated that GlcAT-I activity is associated with Golgi network (12). Computer-based secondary structure prediction indicates that GlcAT-I shares the common topology of type II membrane proteins, consisting of a short N-terminal cytoplasmic tail, a single signal-anchor/transmembrane segment, and a stem region followed by a large luminal C-terminal catalytic domain, characteristic of many other Golgi glycosyltransferases cloned to date. In addition, data base searches allowed to define some general features of the Gal1,3-glucuronosyltransferase family, in particular the presence of four highly conserved motifs (I-IV) located in the luminal domain of the enzymes (15) (see Fig. 1). Substrate specificity studies showed that GlcAT-I is selective for substrates that resemble the linkage region trisaccharide, namely Gal1,3Gal-terminated oligosaccharides. Although the functional expression of human (9, 14) and hamster GlcAT-I (10) as a fusion protein with protein A has been achieved in mammalian cells, no information is yet available concerning the membrane organization and the molecular basis underlying GlcAT-I catalytic activity.

To provide further insight into the structure and function of GlcAT-I, we generated a powerful expression system for the expression of wild-type, truncated, and mutant forms of the enzyme in the methyltrophic yeast Pichia pastoris (P. pastoris), taking advantage of the absence of endogenous Gal1,3-glucuronosyltransferase activity in this host cell. Our results provide evidence that either the native membrane-bound or the secreted form of GlcAT-I is organized as functionally active disulfide-bonded dimers. Furthermore, a site-directed mutagenesis approach was combined with chemical modification and cross-linking studies to investigate the putative role of cysteine residues in dimer formation and/or in the catalytic activity of GlcAT-I.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Bacterial and yeast culture media were from Difco. Protein assay reagent was obtained from Bio-Rad. T4 DNA ligase and competent Escherichia coli JM109 cells were purchased from Promega (Charbonnières, France). The P. pastoris yeast expression system was from Invitrogen (Groningen, The Netherlands). Restriction enzymes, Vent DNA polymerase, and peptide N-glycosidase F were provided by New England Biolabs (Hitchin, United Kingdom). Uridine 5'-diphosphate-glucuronic acid (sodium salt) was purchased from Roche Molecular Biochemicals. Gal1-S-3Gal (Gal-Gal), D-saccharic acid 1,4-lactone (saccharonolactone), -glucuronidase (bovine liver), methanol, anti-rabbit alkaline phosphatase-conjugated immunoglobulins, and the sulfhydryl reagent N-phenylmaleimide were purchased from Sigma. Trifluoroacetic acid, trichloroacetic acid, glycine, and dimethyl sulfoxide were provided by Merck, and acetonitrile was from BDH (Poole, UK). The cross-linking reagent 1,6-bis(maleimido)hexane (BMH) was from Pierce.

Cloning of GlcAT-I cDNA and Plasmid Constructions-- The human GlcAT-I sequence was cloned by polymerase chain reaction (PCR) from a liver cDNA library (CLONTECH, Palo Alto, CA) using a sense primer (5'-CCATGAAGCTGAAGCTGAAGAACGTGTTTCT-3') together with an antisense primer (5'-CCATCACACCTCAATTGCTGGGTCTGA-3') corresponding to the 5'-end and 3'-end of the coding region of GlcAT-I described by Kitagawa et al. (14). The PCR was carried out using Vent DNA polymerase, and the PCR fragment was subloned into the SmaI site of pGEM-3Z and then sequenced on both strands. The cDNA sequence obtained was 100% identical to that previously described by Kitagawa et al. (14).

For the heterologous expression of human GlcAT-I in P. pastoris, the full-length cDNA sequence was modified by PCR to include an EcoRI site and a Kozak consensus sequence at the 5'-end and a XbaI site at the 3'-end, using appropriate oligonucleotides. The modified cDNA was then subcloned into the EcoRI-XbaI sites of the yeast expression vector pPICZB to produce pPICZ-GlcAT-I. A construct coding for GlcAT-I lacking the predicted N-terminal cytoplasmic tail (GlcAT-INT; see Fig. 1) was obtained by PCR amplification using a sense primer containing an EcoRI site, a Kozak sequence and nucleotides 22-45 of the GlcAT-I coding region and an antisense primer comprising a XbaI site, a stop codon, and nucleotides 1005-989 (coding for the last six amino acids of the C-terminus of GlcAT-I). The PCR fragment was subcloned into the EcoRI-XbaI sites of pPICZB to generate pPICZ-GlcAT-INT. A construct for the secretion of the human GlcAT-I in the culture medium of P. pastoris was designed by the fusion of the yeast prepro--factor secretion leader peptide (N-terminal signal peptide and proregion sequence) with the sequence coding for GlcAT-I lacking the predicted N-terminal cytoplasmic tail and transmembrane domain (GlcAT-INT/TMD; see Fig. 1) by simple overlapping ends PCR (16). Briefly, the sequence coding for the yeast -factor secretion signal was amplified using pPIC9 vector (Invitrogen) as template and a sense primer containing an EcoRI site, a Kozak consensus sequence, and nucleotides 1-23 of -factor coding region together with an antisense chimeric primer corresponding to nucleotides 93-76 of GlcAT-I followed by nucleotides 255-238 of -factor (corresponding to the last six amino acid residues of the prepro--factor secretion leader peptide). The sequence coding for GlcAT-I lacking 25 N-terminal amino acids (GlcAT-INT/TMD) was amplified using a sense primer containing nucleotides 76-96 together with an antisense primer comprising a XbaI site, a stop codon, and nucleotides 1005-989 of GlcAT-I. The chimeric sequence coding for the yeast prepro--factor secretion leader peptide fused to GlcAT-I deleted from the N-terminal signal anchor region (25 amino acid residues) was then obtained by PCR amplification using the two PCR products generated above as template and a sense primer coding for the N-terminus of the -factor secretion signal together with an antisense primer corresponding to the coding sequence for the last six amino acids of GlcAT-I. This PCR fragment was finally ligated into the EcoRI-XbaI sites of pPICZB to yield pPICZ-F-GlcAT-INT/TMD.

Site-directed Mutagenesis-- Cys³³ and Cys³⁰¹ were replaced with alanine by site-directed mutagenesis using a two-round PCR-based method as follows. Two separate PCRs were performed using Vent DNA polymerase. The first reaction was realized using a sense primer containing an EcoRI site, a Kozak consensus sequence, and nucleotides 1-18 of the GlcAT-I coding region with an antisense primer introducing the chosen mutation, in which the codon TGT was changed to GCT for Cys³³ and TGC was changed to GCC for Cys³⁰¹. The second PCR was performed with a sense primer complementary to the antisense primer introducing the mutation together with an antisense primer comprising a XbaI site, a stop codon, and nucleotides 1005-989 of GlcAT-I. After purification from agarose, the two PCR fragments were hybridized via the overlapping regions from the sense and antisense primers introducing the mutation and then used as template for the amplification of the full-length GlcAT-I coding region. For each mutation, the resulting PCR fragment was purified from agarose and subcloned into the SmaI site of pGEM-3Z. The recombinant vectors were then digested by EcoRI-XbaI, and the fragment obtained corresponding to each mutation was individually subcloned into EcoRI-XbaI sites of pPICZB yeast expression vector to generate pPICZ-GlcAT-I-C33A and pPICZ-GlcAT-I-C301A. All mutant clones were screened for Taq-introduced errors by dideoxysequencing (17).

Heterologous Expression in the Yeast P. pastoris-- pPICZ-GlcAT-I, pPICZ-GlcAT-INT, pPICZ-F-GlcAT-INT/TMD, pPICZ-GlcAT-I-C33A, and pPICZ-GlcAT-I-C301A were individually transformed into P. pastoris SMD 1168 (Invitrogen) by the lithium chloride method according to the recommendations of the supplier. Transformants were selected on YPD plates (1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose) containing 100 µg/ml of Zeocin. The cells were grown in BMGY medium (1% (w/v) yeast extract, 2% (w/v) peptone, 100 mM potassium phosphate (pH 6.0), 1.34% (w/v) yeast nitrogen base and 1% (v/v) glycerol). Expression was induced in a BMGM medium (BMGY with 1% (v/v) glycerol replaced by 2% (v/v) methanol) and carried out for 48 h at 30 °C in a rotary shaker (215 rpm). Yeast cells were harvested by centrifugation at 3,000 × g for 10 min and further analyzed as described below. In the case of yeast cells transformed by pPICZ-F-GlcAT-INT/TMD, secretion of the recombinant protein was analyzed in the culture medium after sedimentation of yeast cells as above and precipitation of the proteins by trichloroacetic acid (25% (v/v)) or ammonium sulfate at 40% saturation.

Subcellular Fractionation and Protein Analysis of Recombinant Yeast Cells-- Subcellular fractionation of yeast cell extracts was performed as described previously (18). Briefly, after harvesting, cells were washed once and suspended in cold breaking buffer (50 mM sodium phosphate, pH 7.4, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 5% (v/v) glycerol). The cells were then broken by vortexing with glass beads. The resulting homogenate was centrifuged at 5,000 × g for 15 min, and the supernatant was centrifuged at 12,000 × g for 20 min. Membranes were then pelleted from the supernatant for 1 h at 100,000 × g at 4 °C. The pellet fraction was resuspended by Dounce homogenization in sucrose-Hepes buffer (0.25 M sucrose, 5 mM Hepes, pH 7.4), and the supernatant is referred to as the cytosolic fraction. For the analysis of N-linked glycosylation, proteins were denatured in 0.5% (w/v) SDS, 1% (v/v) -mercaptoethanol at 100 °C for 10 min and then incubated with peptide N-glycosidase F in 50 mM sodium phosphate (pH 7.5) buffer supplemented with 1% (v/v) Nonidet P-40 and containing Complete Mini^TM protease inhibitors (Roche Molecular Biochemicals) at 37 °C for 2 h, according to the recommendations of the supplier (New England Biolabs). Peptide N-glycosidase F digestion was also performed under native conditions without prior heating in denaturating buffer, at 37 °C for 2 h as above except that Nonidet P-40 was omitted. Protein concentration was evaluated by the method of Bradford (19).

Antibody Generation, SDS-PAGE, and Immunoblot Analysis-- An anti-GlcAT-I antibody was raised in rabbits against the conserved peptide ³⁰¹CTRVLWHTRTEKPK³¹⁵ in the motif IV of Gal1,3-glucuronosyltransferases. Proteins from various fractions of recombinant yeast cells and from the culture media were separated by SDS-PAGE under reducing and nonreducing conditions (20). Immunoblot analysis was performed using the newly generated polyclonal antibody (1:1000) and alkaline phosphatase-conjugated anti-rabbit immunoglobulins as secondary antibodies, as described previously (18, 21).

Analysis of GlcAT-I Enzymatic Activity-- Activity of recombinant human GlcAT-I was evaluated using Gal-Gal as acceptor substrate. Reaction conditions were optimized in terms of time, protein concentration, divalent ions, and incubation pH. Standard incubations were performed in 100 mM acetate buffer (pH 5.0) with 10 mM MnCl₂, 10-50 µg of membrane protein, or 10 µg of protein from ammonium sulfate fractionation from yeast culture medium, 2 mM Gal-Gal, 5 mM saccharonolactone, and 2 mM UDP-glucuronic acid, in a total volume of 100 µl. The mixture was incubated at 37 °C for 30 min, and the reaction was terminated by placing the tube in ice and adding 10 µl of 6 N HCl. Proteins were precipitated by centrifugation, and the reaction product was then analyzed by HPLC after chromophore labeling by reductive amination as described previously (22). Briefly, 10 µl of reagent mixture (consisting of 1 nmol of aniline, 35 mg of sodium cyanoborohydride, 40 µl of acetic acid, and 350 µl of methanol) were added to 100 µl of the reaction mixture and incubated for 30 min at 80 °C. After cooling, 500 µl of water were added, and noncoupled aniline was extracted with chloroform. The labeled water-soluble product was analyzed by HPLC on a reverse phase C18 column (4.6 × 150 mm, 4 µm, Waters, Milford, MA) at a detection wavelength of 280 nm according to an adapted method (23). The mobile phase was composed of 5% (v/v) acetonitrile, 0.015% (v/v) formic acid, and 0.03% (v/v) triethylamine in water (apparent pH 4.0) and used at a flow rate of 0.5 ml/min. Control assays in which either the donor substrate UDP-glucuronic acid or the acceptor substrate Gal-Gal was omitted were simultaneously run under the same conditions. The addition of a glucuronic residue on the nonreducing end of Gal-Gal was verified by the susceptibility of the product to hydrolysis by -glucuronidase from bovine liver. -Glucuronidase (200 µl, 1200 units) dissolved in 200 mM acetate buffer (pH 5.0) was added to 100 µl of the medium after a 60-min incubation without saccharonolactone. The reaction was conducted for an additional 4 h at 37 °C.

Apparent kinetic parameters (K_m, V_max) were determined using linear least-squares regression analysis of double-reciprocal plots of initial velocity versus Gal-Gal concentration (0-2 mM) at a constant concentration of UDP-glucuronic (2 mM) or of the initial velocity versus UDP-glucuronic acid concentration (0-2 mM) at a constant concentration of Gal-Gal (2 mM).

Chemical Modification-- For N-phenylmaleimide inhibition studies, microsomal membranes from yeast cells expressing wild-type or mutant GlcAT-I were incubated in the presence of an increasing concentration of cysteine-directed reagent (0-20 mM) for various periods of time at 37 °C. The mixture was then diluted 10-fold in the incubation buffer, and the reaction was started by adding UDP-glucuronic acid and Gal-Gal as described above.

Reduction of disulfide bridges was performed by incubation of microsomal membranes from yeast cells expressing wild-type or mutant GlcAT-I and secreted GlcAT-I recovered from the culture medium by ammonium sulfate precipitation with increasing concentrations of DTT (0-25 mM) for 30 min on ice. The enzymatic assays were then performed as described above.

Cross-linking Studies-- BMH was dissolved in dimethyl sulfoxide, and the solution (50 mM) was added to 50 µl of microsomal (50 µg) or secreted proteins (20 µg) in 10 mM Tris-HCl (pH 7.5), 0.25 M sucrose to give a concentration of 0.02-1 mM BMH as described (24). The cross-linking reaction was carried out at room temperature for 60 min and quenched by the addition of the same volume of sample buffer containing -mercaptoethanol. 20-30 µg of protein were analyzed by SDS-PAGE and immunostained with anti-GlcAT-I antibodies.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Expression of Membrane-bound and Secreted Forms of GlcAT-I in P. pastoris-- An expression vector directing the synthesis of the full-length GlcAT-I polypeptide in the yeast P. pastoris was constructed. In addition, we designed a mutant lacking the seven N-terminal amino acids constituting the cytoplamic tail of the protein in order to analyze a possible influence of this positively charged peptide on the membrane association and orientation of GlcAT-I (see Fig. 1). Finally, a soluble form of GlcAT-I was produced by the fusion of the GlcAT-I sequence coding for the polypeptide lacking the 25 N-terminal amino acids (corresponding to the predicted cytoplasmic tail and transmembrane domain) with the yeast cleavable prepro--factor leader (Fig. 1). The resultant recombinant plasmids (pPICZ-GlcAT-I, pPICZ-GlcAT-INT, and pPICZ-F-GlcAT-INT/TMD) were individually transformed into the methyltrophic yeast P. pastoris. Upon methanol induction, expression of the full-length and truncated GlcAT-I polypeptides was successfully achieved. Western blot analysis of whole cell extracts showed that recombinant GlcAT-I and GlcAT-INT migrated on SDS-polyacrylamide gel electrophoresis as a polypeptide band of approximately 43 kDa (Fig. 2, lanes a and b).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the predicted GlcAT-I organization showing the common topology of type II membrane proteins. NT, N-terminal cytosolic tail; TMD, transmembrane domain. Motifs I-IV conserved among the Gal1,3-glucuronosyltransferase family are in boldface type. Alignment of the partial amino acid sequence of motif IV of human GlcAT-I, hamster GlcAT-I, rat GlcAT-P, rat GlcAT-S, and putative protein from Schistosoma mansoni (GenBankTM accession no. AAC4695.1) and from Caenorhabditis elegans (GenBankTM accession no. ZK1307.5) containing the invariant Cys301 residue is shown. Positions of the cysteine residues mutated in this study are indicated. A schematic representation of the truncated GlcAT-I deleted from the predicted N-terminal cytoplasmic tail (GlcAT-INT) and of the construct designed for the secretion of GlcAT-I (F-GlcAT-INT/TMD) is shown.

View larger version (56K): [in this window] [in a new window]   Fig. 2.   Comparison of the mobility of recombinant wild-type GlcAT-I, GlcAT-INT, and secreted GlcAT-I on SDS-gel electrophoresis. Samples of total extracts, of the cytosolic fraction, and of the membrane fraction of recombinant yeast cells expressing GlcAT-I (lanes a, c and d, respectively) and GlcAT-INT (lanes b, f, and g, respectively) and a sample of culture medium containing secreted GlcAT-I (lane i) were analyzed by 10% SDS-PAGE, transferred to nylon membranes, and probed with anti-peptide antibodies as described under "Experimental Procedures." Membrane fraction from yeast cells expressing GlcAT-I (lane e), GlcAT-INT (lane h), and secreted GlcAT-I (lane j) were treated by N-glycosidase F.

Subfractionation experiments showed that, as expected, full-length GlcAT-I was associated with membrane-enriched fraction of recombinant yeast cells, whereas no polypeptide was detected in the supernatant from 100,000 × g centrifugation (Fig. 2, lanes d and c, respectively). Interestingly, GlcAT-INT exhibited the same distribution upon subfractionation (Fig. 2, lanes g and f, respectively), suggesting that the deletion of the predicted N-terminal cytoplasmic tail did not compromise membrane targeting and association of GlcAT-I protein.

The apparent molecular mass of GlcAT-I and GlcAT-INT was reduced to approximately 39 kDa after N-glycanase treatment (Fig. 2, compare lane e with lane d and lane h with lane g), indicating that the one potential N-linked glycosylation site located in the C-terminal part of the protein is utilized, in agreement with previous experiments performed on a protein A-GlcAT-I fusion protein (9). The apparent molecular mass of the deglycosylated protein (about 39 kDa) was in agreement with the expected molecular mass of 37.08 kDa, calculated from the predicted primary sequence. Furthermore, the N-glycosylation of the cytoplasmic domain deletion mutant strongly suggested that this truncated polypeptide was translocated and had achieved an "N in/C out" orientation, as the wild-type protein.

The construct pPICZ-F-GlcAT-INT/TMD, encoding GlcAT-I lacking the cytoplasmic tail and transmembrane domain, efficiently directed the synthesis and secretion of a soluble form of the enzyme into the culture medium of recombinant yeast cells, as shown by Western blot analysis (Fig. 2, lane i). The apparent molecular mass of the secreted form (about 41 kDa) was, as expected, slightly lower than that of the wild-type protein, was compatible with the deletion of the cytoplasmic tail and transmembrane domain (Fig. 2, compare lane i with lane d). Interestingly, upon treatment by N-glycosidase F, the mobility of the secreted polypeptide was increased of about 4 kDa (lane j), thus indicating that it was glycosylated to a similar extent to the wild-type protein and that no hyperglycosylation occurred during the secretion pathway.

Functional Characterization of Recombinant Membrane-bound and Secreted Forms of GlcAT-I-- An optimized HPLC assay for the recombinant GlcAT-I was developed using the digalactoside derivative Gal-Gal as an acceptor substrate. Enzyme activity was undetectable in yeast membrane fraction, in the 100,000 × g supernatant, and in the culture medium of nonrecombinant yeast cells or in noninduced recombinant cells, thus emphasizing the usefulness of P. pastoris as a host cell for heterologous expression (not shown). By contrast, high level of enzyme activity was associated with membrane fraction of recombinant yeast cells expressing the wild-type GlcAT-I (74 pmol·min¹·mg¹ protein). We found that divalent cations were essential for the enzymatic reaction and that Mn²⁺ exhibited the highest activating effect at an optimal concentration of 10 mM. Moreover, the activity was found to be maximal between pH 5.0 and 6.0.

On the other hand, membrane fraction from yeast cells expressing GlcAT-INT exhibited a similar level of activity when compared with the wild-type protein (not shown). Interestingly, the secreted form of GlcAT-I was highly active with a specific activity of 120 pmol·min¹·ml¹ culture medium. In addition, the secreted and membrane-bound enzymes exhibited similar apparent affinity constants (not shown).

Oligomeric State of GlcAT-I-- To investigate the oligomeric structure of GlcAT-I, the wild-type protein as well as the truncated forms of the enzyme were analyzed by SDS-PAGE under reducing or nonreducing conditions and immunoblotted using the anti-peptide antibodies. Under nonreducing conditions, the membrane-bound native GlcAT-I or the protein lacking the N-terminal cytosolic tail migrated as polypeptides of about 88 kDa (Fig. 3A, lanes a and b), whereas the secreted form of the enzyme was detected as a slightly faster migrating band of about 85 kDa (Fig. 3A, lane c). Disulfide reduction of the wild- type and of the cytoplasmic deletion mutant generated the 43-kDa monomer (Fig. 3A, lanes d and e), consistent with the dissociation of disulfide-linked homodimers in the presence of DTT. On the other hand, the possibility that GlcAT-I may be disulfide-linked with other cellular yeast proteins cannot be ruled out. However, as for the wild-type protein, the apparent molecular mass of the secreted form in nonreducing conditions was twice that of the polypeptide produced upon disulfide reduction (Fig. 3A, compare lanes c and f), thus favoring the hypothesis of homodimer formation. Furthermore, the dimerization of GlcAT-I lacking the N-terminal cytosolic tail only or lacking both the cytoplasmic tail and the membrane-spanning segment indicated that the disulfide bond was probably located on the luminal domain of the protein.

View larger version (76K): [in this window] [in a new window]   Fig. 3.   Analysis of GlcAT-I dimerization by SDS-PAGE and immunoblotting. Membrane fraction of yeast cells expressing wild-type GlcAT-I (A, lanes a and d) or GlcAT-INT (A, lanes b and e) and culture medium from yeast expressing the secreted GlcAT-I (A, lanes c and f) were analyzed by SDS-PAGE under nonreducing (lanes a-c) and reducing conditions in the presence of 25 mM DTT (lanes d-f). Membrane fractions of recombinant yeast cells expressing the wild-type GlcAT-I were treated with BMH (B, lane a, 0 mM; lane b, 0.02 mM; lane c, 1 mM) as described under "Experimental Procedures" and analyzed by SDS-PAGE and immunoblotting.

To verify the putative formation of GlcAT-I dimers, we used the cross-linking reagent BMH, which is a homobifunctional cross-linker that reacts with sulfhydryl groups of proteins. Fig. 3B shows the immunostaining analysis of BMH-treated microsomes expressing the wild-type GlcAT-I. The cross-linker BMH consistently induced the appearance of an 88-kDa polypeptide (Fig. 3B, compare lanes b and c with lane a). The amount of cross-linked dimers increased with the cross-linker concentration. The same experiment performed on the secreted form GlcAT-INT/TM indicated a similar behavior to that of the wild-type protein (not shown). In addition, for both membrane-bound and secreted forms, the apparent molecular mass of the cross-linked product on SDS-PAGE was similar to that of the polypeptide observed under nonreducing conditions, supporting the hypothesis of homodimer formation.

Alanine for Cysteine Substitution in Human GlcAT-I-- We investigated the role of Cys³³ and Cys³⁰¹ in disulfide bond formation and catalytic activity by constructing mutants in which the cysteine residues were individually replaced by alanine (see Fig. 1). The alanine-substituted mutants were found to be expressed in P. pastoris at a similar level to that of the wild-type protein, as shown by immunoblot analysis (Fig. 4, compare lanes b and c with lane a). When the proteins were analyzed under reducing conditions, the GlcAT-I-C33A mutant ran as a single polypeptide of approximately 43 kDa (Fig. 4, lane b), whereas the GlcAT-I-C301A mutant migrated with an apparent molecular mass about 4 kDa less than the wild-type protein (Fig. 4, compare lane c with lane a). Cys³⁰¹ is located in the unique potential Asn-linked glycosylation site of GlcAT-I (Asn³⁰⁰-Cys³⁰¹-Thr³⁰²). Since the apparent molecular mass of the GlcAT-I-C301A mutant was similar to that of the N-glycosidase F-treated wild-type protein, it was tempting to speculate that the increased mobility of the mutant was due to the absence of N-glycosylation resulting from the creation of a less efficient N-glycosylation consensus site by cysteine to alanine substitution. In agreement with this, we showed that the GlcAT-I-C301A mutant was not sensitive to N-glycosidase F treatment, indicating that indeed this mutant was not N-glycosylated (not shown).

View larger version (96K): [in this window] [in a new window]   Fig. 4.   SDS-PAGE and immunoblot analysis of expressed GlcAT-I with cysteine for alanine mutants under reducing and nonreducing conditions. Membrane fraction of recombinant yeast cells expressing wild-type GlcAT-I, GlcAT-I-C33A, and GlcAT-I-C301A were analyzed by SDS-PAGE and immunoblotting under reducing (DTT, 25 mM, lanes a, b, and c, respectively) and nonreducing conditions (lanes d, e, and f, respectively).

When samples were analyzed by SDS-PAGE under nonreducing conditions, the GlcAT-I-C301A mutant, like the wild-type protein, migrated as a dimer (Fig. 4, lanes f and d). This result suggested that N-linked glycosylation was not a prerequisite for dimerization of the protein. Interestingly, the GlcAT-I-C33A mutant protein migrated at the position of a 43-kDa monomer either under reducing or nonreducing conditions (Fig. 4, lanes b and e), suggesting that this mutation abolished the ability of the protein to form dimers and that Cys³³ is involved in disulfide bond formation.

Catalytic Activity and N-Phenylmaleimide Sensitivity of Wild-type and Mutant GlcAT-I Enzymes-- We then examined the possible role of cysteine residues in the functionality of GlcAT-I by chemical modification using a specific sulfhydryl reagent. N-Phenylmaleimide treatment strongly inactivated GlcAT-I in a time-dependent manner (Fig. 5A). Up to 60% of the initial rate was inhibited after 20-min exposure to the sulfhydryl reagent. Moreover, a 20-min incubation of microsomal membranes from yeast cells expressing the wild-type GlcAT-I with N-phenylmaleimide resulted in a dose-dependent decrease of GlcAT-I activity (Fig. 5B), strongly suggesting that the catalytic activity of GlcAT-I relies on the presence of free thiol groups.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Sensitivity of GlcAT-I activity to N-phenylmaleimide. A, membrane fractions from yeast cells expressing wild-type GlcAT-I () or GlcAT-I-C33A () were incubated with 5 mM N-phenylmaleimide for various periods of time up to 60 min and assayed for activity as described under "Experimental Procedures." B shows the variation in activity as a function of the concentration of N-phenylmaleimide after 20 min of inactivation.

To gain further insight into the putative role of cysteine residues, chemical modification was combined with analysis of the catalytic function of the cysteine to alanine mutants. The apparent kinetic parameters for wild-type and mutant GlcAT-I are presented in Table I. The loss of Cys³³ in the mutant GlcAT-I-C33A resulted in a decrease of about 25% of V_max compared with the wild-type enzyme. In addition, apparent affinity constants toward the acceptor substrate Gal-Gal and toward UDP-glucuronic acid were increased by 80 and 200%, respectively, resulting in an important reduction in the efficiency of the enzyme as illustrated by the V_max/K_m values of the mutant compared with those of the wild-type (Table I). This decrease in the efficiency of the enzyme can be attributed to the absence of dimer formation described above for GlcAT-I-C33A. The role of disulfide bonds in maintaining enzyme function was further investigated by treating the wild-type GlcAT-I with DTT. The results obtained indicated that DTT produced a loss of activity of about 28% at the highest dose used (25 mM). These complementary observations suggested that dimerization was not a strict prerequisite for the activity but was necessary for the optimal function of the enzyme.

Finally, the effect of substituting Cys³⁰¹ with alanine on the catalytic activity of GlcAT-I was investigated. Remarkably, this mutation led to a completely inactive protein (see Table I) indicating that Cys³⁰¹ is probably the target of N-phenylmaleimide inactivation. In addition, deglycosylation of the wild-type protein under native conditions yielded a fully active enzyme, ruling out the possibility that the absence of glycosylation compromises the functionality of GlcAT-I. It is also noteworthy that the GlcAT-I-C33A mutant was equally sensitive to N-phenylmaleimide inactivation as the wild-type protein (Fig. 5, A and B), thus excluding the contribution of Cys³³ to N-phenylmaleimide sensitivity. Altogether, our results suggest that the strictly conserved Cys³⁰¹, located near the C-terminus of the protein, is an essential residue for catalytic activity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In this work, we provide new structural features on the organization of GlcAT-I, the only human isoform belonging to the Gal1,3-glucuronosyltransferase family yet cloned. We demonstrate, for the first time, that the enzyme exists as catalytically active dimers and that two crucial cysteine residues, Cys³³ and Cys³⁰¹, mediate dimer formation and functional activity.

Heterologous Expression and Functional Analysis of the Recombinant Human GlcAT-I-- The structure/function analysis of human GlcAT-I was successfully achieved by the exploitation of the unique capability of the yeast P. pastoris to express high levels of heterologous proteins. In addition, the absence of endogenous activity was a predominant criterion for the choice of P. pastoris as a host cell for the heterologous expression of this enzyme. Indeed, we were able to produce high levels of the native membrane-bound and of several truncation and mutant GlcAT-I forms in this yeast expression system. Human (9) and hamster GlcAT-I (10) and two rat isoforms of Gal1,3-glucuronosyltransferases (15, 25) have been previously functionally expressed in mammalian cells as soluble secreted protein A-GlcAT-I fusion proteins, but this is the first study performed on the wild-type membrane-bound human GlcAT-I and on a secreted protein. We report here that human GlcAT-I expressed in P. pastoris, upon N-glycosidase F treatment, runs as a polypeptide of about 39 kDa on SDS-polyacrylamide gels. This first estimate of the apparent molecular mass of the protein is in good agreement with the computer-based, predicted molecular mass for human GlcAT-I of 37.08 kDa.

Furthermore, we show that the expression system developed is particularly valuable for its ability to secrete human GlcAT-I with high efficiency when placed under the control of the yeast prepro--factor secretion leader. During passage of the heterologous protein through the secretory pathway, the prepro--factor leader was cleaved out, as judged by the apparent molecular mass of the secreted GlcAT-I, compared with the wild-type protein. It is noteworthy that the glycosylation state of the secreted form was similar to that of the membrane-bound enzyme, based on N-glycosidase F sensitivity, indicating that no hyperglycosylation occurred during the secretory pathway. Furthermore, similar kinetic parameters were found for the secreted GlcAT-I and membrane-bound enzyme. Altogether, the production and the secretion of high yields of functionally active human GlcAT-I opens promising perspectives for further structural studies.

Recombinant GlcAT-I fused to protein A was previously shown to catalyze the addition of glucuronic acid onto the linkage trisaccharide Gal1,3Gal1,4Xyl-O-Ser (14). In addition, Gal1,3-Gal derivatives were reported to be good acceptors for embryonic chick cartilage GlcAT-I (11, 12) as well as for the recombinant enzyme (10). Here, we developed a convenient assay for human GlcAT-I based on the use of a digalactosidic compound and validated its usefulness as a reporter substrate for the recombinant GlcAT-I. The kinetic parameters obtained in this study were similar to values previously reported for the native cartilage enzyme (12) and for a recombinant fusion protein (9).

On the other hand, our results showed that deletion of the N-terminal cytoplasmic tail did not prevent GlcAT-I targeting and association to membranes in recombinant yeast cells nor its activity. Moreover, the sensitivity of this mutant to N-glycosidase F showed that it was correctly oriented in an N in/C out fashion, as the wild-type protein. This provided evidence for the optional role of the highly positively charged N-terminal segment in membrane orientation of GlcAT-I. In contrast, deletion of the cytoplasmic tail and transmembrane domain allowed this protein to enter the secretory pathway, suggesting the absence of a retention signal in the stem and catalytic domain. Similarly to our observations, Teasdale et al. (26) concluded from localization studies of bovine 1,4-galactosyltransferase and of hybrid 1,4-galactosyltransferase molecules that the cytoplasmic tail was not required for the association of this glycosyltransferase to Golgi membranes, whereas its transmembrane domain contained a positive signal for retention within the Golgi complex. In contrast, the luminal domain of other glycosyltransferases such as 2,6-sialyltransferase can only be efficiently secreted when lacking the stem region (27). It thus appears that GlcAT-I resembles the galactosyltransferase families that rely primarily on their transmembrane domain for Golgi retention without special requirements for the transmembrane flanking region and luminal stem sequences (28).

Oligomeric State of GlcAT-I-- Our results provide compelling evidence that GlcAT-I consists of disulfide-linked dimers. This assumption was based in the first instance on the finding that the wild-type GlcAT-I migrated as a high molecular mass polypeptide of 88 kDa when analyzed on nonreducing SDS-gels. This oligomeric form was interpreted to be a homodimer based on its estimated size, which was twice that of the monomer. The observation that the secreted form also exists as an apparent dimer suggested that the high molecular mass polypeptide probably results from homodimer formation rather than from interaction with yeast intracellular proteins. In addition, we showed that the bifunctional cross-linking reagent BMH also shifted the size of recombinant GlcAT-I monomeric form from 43 to about 88 kDa. Similarly, the high molecular form was generated from the membrane-bound as well as from the secreted protein, thus strengthening the assumption of homodimer formation.

Furthermore, the capability of both the cytoplasmic tail-deleted mutant and the secreted protein to dimerize strongly suggested that the presence of the N-terminal cytoplasmic tail and transmembrane domain are not required for intermolecular disulfide bridge formation. From these results, it could be anticipated that cysteine residue(s) of the luminal domain are likely to be responsible for dimer formation. Replacement of the conserved Cys³⁰¹ localized in the motif IV by alanine yielded a mutant remaining competent for dimerization, thus ruling out the involvement of this residue in this process. It is worth noting that this mutation abolished N-glycosylation of the polypeptide as evidenced by its reduced apparent molecular mass together with its insensitivity to N-glycosidase F, thus indicating that absence of N-glycosylation did not prevent dimer formation.

Interestingly, we found that substitution of Cys³³ by alanine suppressed dimer formation, based on the identical apparent molecular mass of the mutant when analyzed by SDS-PAGE in reducing and nonreducing conditions. From these data, we can reasonably assume that Cys³³ is involved in an intermolecular disulfide bridge. Indeed, several Golgi membrane-bound glycosyltransferases exist as intermolecular disulfide-bonded species (29, 30), although some of them have been reported to be monomers such as -galactosyltransferase (31). Homodimerization of type II membrane proteins has been shown to be mediated either by the transmembrane domain or by the luminal region of these proteins. In this regard, GlcAT-I appears similar to 1,4-N-acetylgalactosaminyltransferase (30) or to -mannosidase II (32), which forms dimers via luminal domain disulfide bonds.

Functional Role of Cysteine Residues-- To further analyze the importance of dimerization on the enzymatic activity, we investigated the consequences of DTT treatment. We reported that DTT partially affected the enzymatic activity at about the same extent as the conversion of Cys³³ residue to alanine, which was shown to abolish dimer formation. Interestingly, this mutation led to a glucuronosyltransferase with an impaired efficiency as evidenced by the significantly higher apparent affinity constants compared with the wild-type protein. Conceivably, dimerization of GlcAT-I via Cys³³ could favor monomer-monomer interactions, promoting a better catalytic efficiency. On the other hand, chemical modification of GlcAT-I with the sulfhydryl-specific reagent N-phenylmaleimide caused a dramatic inactivation of the wild-type protein as well as of the mutant GlcAT-I-C33A. This result provided evidence that Cys³³ was not targeted by the chemical reagent and that other cysteine residues might be essential for catalysis and/or substrate binding. Interestingly, mutation of Cys³⁰¹, a highly conserved residue on domain VI of Gal1,3-glucuronosyltransferases, by alanine yielded a totally inactive unglycosylated enzyme. Experiments designed to address the effect of N-glycosylation on glucuronosyltransferase activity of human GlcAT-I showed that deglycosylation of wild-type protein under nondenaturing conditions did not affect the activity of the enzyme. These results suggested that Cys³⁰¹ residue plays a key role in enzyme function. A common acid/base mechanism has been proposed for glycosyltransferases. Some enzymes such as glutamate racemase (33) are known to employ active-site cysteine residues as acid/base catalysts. It is therefore tempting to postulate such a function for Cys³⁰¹. On the other hand, a possible role of Cys³⁰¹ in the binding of acceptor or donor substrates cannot be ruled out. Further structural studies are currently under way to test these mechanisms.

In conclusion, we demonstrate the key role of cysteine residues in both dimer formation and activity of GlcAT-I. Altogether, the picture that emerges from our results is that of a homodimeric protein of 88 kDa with an interdisulfide bridge involving Cys³³ of the N-terminal variable stem region. In addition, our data support the idea that the dimeric state of GlcAT-I may promote an optimal folding of the protein, leading to a fully active enzyme. Finally, we provide direct evidence for the crucial role of the conserved Cys³⁰¹ in the mechanism underlying Gal1,3-glucuronosyltransferase-catalyzed transfer of glucuronic acid onto the trisaccharide GAG-core protein linkage region.

	ACKNOWLEDGEMENT

Dr. D. Hulmes is gratefully acknowledged for critical reading of the manuscript.

	FOOTNOTES

^* This work was supported in part by Région Lorraine and by a grant from FR 42 "Protéines."The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: UMR CNRS 7561-Université Henri Poincaré Nancy 1, Faculté de Médecine, BP 184, 54505 Vandoeuvre-lès-Nancy, France. Tel.: 33 3 83 59 27 49; Fax: 33 3 83 59 26 21; E-mail: ouzzine@pharmaco-med.u-nancy.fr.

Published, JBC Papers in Press, June 6, 2000, DOI 10.1074/jbc.M002182200

	ABBREVIATIONS

The abbreviations used are: GAG, glycosaminoglycans; BMH, 1,6-bis(maleimido)hexane; DTT, dithiothreitol; Gal-Gal, Gal1-S-3Gal; GlcAT-I, UDP-glucuronic acid:Gal1,3-glucuronosyltransferase (EC 2.4.1.135); HPLC, high performance liquid chromatography; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				Y. Tone, L. C. Pedersen, T. Yamamoto, T. Izumikawa, H. Kitagawa, J. Nishihara, J.-i. Tamura, M. Negishi, and K. Sugahara 2-O-Phosphorylation of Xylose and 6-O-Sulfation of Galactose in the Protein Linkage Region of Glycosaminoglycans Influence the Glucuronyltransferase-I Activity Involved in the Linkage Region Synthesis J. Biol. Chem., June 13, 2008; 283(24): 16801 - 16807. [Abstract] [Full Text] [PDF]

				M. Fondeur-Gelinotte, V. Lattard, S. Gulberti, R. Oriol, G. Mulliert, M. W. Coughtrie, J. Magdalou, P. Netter, M. Ouzzine, and S. Fournel-Gigleux Molecular basis for acceptor substrate specificity of the human {beta}1,3-glucuronosyltransferases GlcAT-I and GlcAT-P involved in glycosaminoglycan and HNK-1 carbohydrate epitope biosynthesis, respectively Glycobiology, August 1, 2007; 17(8): 857 - 867. [Abstract] [Full Text] [PDF]

				M. Fondeur-Gelinotte, V. Lattard, R. Oriol, R. Mollicone, J.-C. Jacquinet, G. Mulliert, S. Gulberti, P. Netter, J. Magdalou, M. Ouzzine, et al. Phylogenetic and mutational analyses reveal key residues for UDP-glucuronic acid binding and activity of beta1,3-glucuronosyltransferase I (GlcAT-I). Protein Sci., July 1, 2006; 15(7): 1667 - 1678. [Abstract] [Full Text] [PDF]

				S. Gulberti, V. Lattard, M. Fondeur, J.-C. Jacquinet, G. Mulliert, P. Netter, J. Magdalou, M. Ouzzine, and S. Fournel-Gigleux Phosphorylation and Sulfation of Oligosaccharide Substrates Critically Influence the Activity of Human {beta}1,4-Galactosyltransferase 7 (GalT-I) and {beta}1,3-Glucuronosyltransferase I (GlcAT-I) Involved in the Biosynthesis of the Glycosaminoglycan-Protein Linkage Region of Proteoglycans J. Biol. Chem., January 14, 2005; 280(2): 1417 - 1425. [Abstract] [Full Text] [PDF]

				N. Venkatesan, L. Barre, A. Benani, P. Netter, J. Magdalou, S. Fournel-Gigleux, and M. Ouzzine Stimulation of proteoglycan synthesis by glucuronosyltransferase-I gene delivery: A strategy to promote cartilage repair PNAS, December 28, 2004; 101(52): 18087 - 18092. [Abstract] [Full Text] [PDF]

S. Gulberti, S. Fournel-Gigleux, G. Mulliert, A. Aubry, P. Netter, J. Magdalou, and M. Ouzzine The Functional Glycosyltransferase Signature Sequence of the Human {beta}1,3-Glucuronosyltransferase Is a XDD Motif J. Biol. Chem., Augus