Identification and functional expression of a second humanb-galactoside a2,6-sialyltransferase, ST6Gal II Marie-Ange Krzewinski-Recchi1, Sylvain Julien1, Sylvie Juliant2, Me´lanie Teinte
Trang 1Identification and functional expression of a second human
b-galactoside a2,6-sialyltransferase, ST6Gal II
Marie-Ange Krzewinski-Recchi1, Sylvain Julien1, Sylvie Juliant2, Me´lanie Teintenier-Lelie`vre1,
Be´ne´dicte Samyn-Petit1, Maria-Dolores Montiel1, Anne-Marie Mir1, Martine Cerutti2,
Anne Harduin-Lepers1and Philippe Delannoy1
1
Unite´ de Glycobiologie Structurale et Fonctionnelle, UMR CNRS – USTL 8576, Universite´ des Sciences et Technologies de Lille, F-59655 Villeneuve d’Ascq, France;2Station de Recherche de Pathologie Compare´e, UMR CNRS – INRA 5087,
F-30380 Saint Christol-lez-Ale`s, France
BLASTanalysis of the human and mouse genome sequence
databases using the sequence of the human CMP-sialic
acid:b-galactoside a-2,6-sialyltransferase cDNA (hST6Gal I,
EC2.4.99.1) as a probe allowed us to identify a putative
sialyltransferase gene on chromosome 2 The sequence of
the corresponding cDNA was also found as an expressed
sequence tag of human brain This gene contained a
1590 bp open reading frame divided in five exons and the
deduced amino-acid sequence didn’t correspond to any
sialyltransferase already known in other species Multiple
sequence alignment and subsequent phylogenic analysis
showed that this new enzyme belonged to the ST6Gal
subfamily and shared 48% identity with hST6Gal-I
Consequently, we named this new sialyltransferase
ST6Gal II A construction in pFlag vector transfected in
COS-7 cells gave raise to a soluble active form of ST6Gal II Enzymatic assays indicate that the best acceptor substrate of ST6Gal II was the free disaccharide Galb1–4GlcNAc structure whereas ST6Gal I preferred Galb1–4GlcNAc-R disaccharide sequence linked to a protein The a2,6-linkage was confirmed by the increase of Sambucus nigra agglutinin-lectin binding to the cell surface of CHO transfected with the cDNA encoding ST6Gal II and by specific sialidases treatment In addition, the ST6Gal II gene showed a very tissue specific pattern of expression because it was found essentially in brain whereas ST6Gal I gene is ubiquitously expressed
Keywords: human; b-galactoside a2,6-sialyltransferase; molecular cloning
Sialylated sugar chains are present at the cell surface of
various animal species Due to their position, they are
thought to serve important roles in a large variety of
biological functions such as cell–cell and cell–substrate
interactions, bacterial and virus adhesion, and protein
targeting [1,2] Sialylated glycoconjugates exhibit
remark-ably diverse structures [3–5] and their expression has been
shown to change during development [6], differentiation,
disease and oncogenic transformation [7] In mammals,
sialic acids are found at the nonreducing terminal position
of glycoconjugates sugar chains, a2,3- or a2,6 -linked to a
b-D-galactopyranosyl (Gal) residue, or a2,6 -linked to a b-D-N-acetylgalactosaminyl (GalNAc) or a b-D -N-acetyl-glucosaminyl (GlcNAc) residue Sialic acids are also found a2,8-linked to sialic acid residues in gangliosides and in polysialic acid, a linear a2,8-homopolymer observed on several glycoproteins including the neural cell adhesion molecule N-CAM [8] In addition, a2,6-linked sialic acid is also present in free oligosaccharides such as 6¢-sialyllactose from human milk [9], monosialylganglioside of the human meconium [10]; Neu5Aca2–6GalNAcb1–4GlcNAc-R sequence has been described as a terminal sequence of the N-glycans of pituitary hormones [11]
The biosynthesis of sialylated oligosaccharides is cata-lysed by a family of enzymes named sialyltransferases [3,12] These enzymes are a subset of the glycosyltransferases family (family 29 in the CAZy database [13]) that use CMP-Neu5Ac as the activated sugar donor to catalyse the transfer
of sialic acid residues to the terminal position of oligosac-charide chains of glycolipids and glycoproteins Sialyltrans-ferases are a family of type II membrane-bound glycoproteins with a short NH2-terminal cytoplasmic tail, a 16–20 amino acid signal anchor domain that is involved with retention of the protein in the Golgi apparatus, a stem region, highly variable in length (from 20 amino acids to 200 amino acids), ending with a large COOH-terminal catalytic domain that resides in the Golgi lumen [14] The catalytic domain contains three highly conserved amino-acid sequences termed sialylmotifs L (large), S (small), and VS (very small)
Correspondence to M.-A Krzewinski-Recchi, Unite´ de Glycobiologie
Structurale et Fonctionnelle, UMR CNRS no 8576, Laboratoire de
Chimie Biologique, Universite´ des Sciences et Technologies de Lille,
F-59655 Villeneuve d’Ascq, France.
Fax: + 33 320 43 65 55, Tel.: + 33 320 43 69 23,
E-mail: Marie-Ange.Recchi@univ-lille1.fr
Abbreviations: EST, expressed sequence tag; Gal, b- D
-galactopyrano-syl residue; GalNAc, b- D -N-acetylgalactosaminyl; N-CAM, neural
cell adhesion molecule; EGT, ecdysone-S-glycosyltransferase.
Enzyme: CMP-sialic acid:b-galactoside a-2,6-sialyltransferase cDNA
(hST6Gal I, EC2.4.99.1).
Note: nucleotide sequence data are available in the DDBJ/EMBL/
GenBank databases under the accession number AJ512141.
(Received 5 November 2002, revised 7 January 2003,
accepted 10 January 2003)
Trang 2Sialylmotifs L and S are involved in the binding of donor
and acceptor substrates, respectively [15,16], whereas the
sialylmotif VS is involved in the catalytic process [17]
To date, 19 different sialyltransferases have been
identi-fied in mouse and humans but only one of these enzymes,
ST6Gal I (CMP-sialic acid:b-galactoside
a2,6-sialyltrans-ferase, EC 2.4.99.1) is known to mediate the transfer of a
sialic acid residue in a2,6-linkage to the galactose residue of
the type 2 disaccharide (Galb1–4GlcNAc) found as a free
disaccharide or as a terminal N-acetyllactosamine unit of
N- and O-glycans However, as reviewed previously [3],
ST6Gal I has been shown in in vitro assays to have a low
activity for transferring sialic acid onto other
oligosac-charide structures, such as lactose (Galb1–4Glc), type 1
disaccharide structure (Galb1–3GlcNAc) [18] or type 2
structure GalNAcb1–4GlcNAc [19,20] but not onto
type 3 structure (Galb1–3GalNAc) ST6Gal I has been
purified to homogeneity from animal livers and hepatoma
cells (reviewed in [3]) and cDNA has been cloned from rat
liver [21], human placenta [22], mouse liver [23], bovine
tissues [20] and chick embryo [24] Several mRNA isoforms
are generated from a unique gene encoding ST6Gal I
through the use of physically distinct promoters These
transcripts differ only in their 5¢-untranslated region and
share an identical ST6Gal I coding region These transcripts
are expressed in a tissue-specific manner and contribute to
the regulation of a2,6-sialylation in tissue and cells during
cell differentiation [25,26], inflammation [27] and oncogenic
transformation [28,29]
BLAST analysis of the mouse and human genome
databases allowed us to identify an unknown
sialyltrans-ferase gene encoding a second Galb1–4GlcNAc
a2,6-sialyltranferase that has been named hST6Gal II In this
report, we describe the functional analysis of a recombinant
human ST6Gal II, which shows slightly different substrate
specificity than ST6Gal I The expression pattern of the
gene was also examined in various human tissues and found
to be very restricted, mainly to brain and fetal tissues
Materials and methods
Materials
CMP-[14C]Neu5Ac (10.7 GBqÆmmol)1), Redivue stabilized
[a-32P]dCTP (110 TBqÆmmol)1) and Rediprime II DNA
labelling system were from Amersham Pharmacia Biotech
(Little Chalfont, UK) The DyNazyme EXT DNA
poly-merase was from Ozyme (Saint Quentin en Yvelines,
France) Oligonucleotides were synthesized by Eurogentec
(Seraing, Belgium) Dulbecco’s modified Eagle’s medium
(DMEM) containing 4.5 gÆL)1glucose and lacking
gluta-mine was from BioWhittaker Europe Alpha Eagle’s
minimal essential medium (aMEM), OPTIMEM,L
-gluta-mine and antibiotics used in cell culture were from Gibco
BRL (Cergy-Pontoise, France) Fetal bovine serum was
from D Dustscher (Issy-les-Moulineaux, France)
Lipofect-AMINE PLUS Reagent was from Invitrogen The
block-ing reagent and fluorescein labelled anti-digoxigenin Fab
fragments and DOTAP transfection reagent were from
Roche (Meylan, France) a1-Acid glycoprotein, fetuin,
Galb1–3GalNAca-O-benzyl, Galb1–3GlcNAc, Galb1–
4GlcNAc, the expression vector pFlag-CMV-1, the mono-clonal antibody (mAb) anti-Flag BioM2, alkaline phospha-tase conjugated goat anti-[mouse IgG (Fab specific)] Ig and CMP-Agarose beads were from Sigma (St Louis, MO, USA) Lacto-N-neotetraose and Lacto-N-tetraose were the generous gift of G Strecker and F Chirat (UMR CNRS
8576, Villeneuve d’Ascq, France) Galb1–3GlcNAcb-O-octyl and Galb1–4GlcNAcb-O-Galb1–3GlcNAcb-O-octyl were the generous gift
of C Auge´ (URA CNRS 462, Orsay, France) MTNTM multiple tissue Northern blot and MTETMmultiple tissue expression arrays were from Clontech (Palo Alto, CA, USA) Sialic Acid Linkage Analysis Kit was from Glyko Inc (Novato, CA, USA) The expressed sequence tag (EST) clone (GenBank/EBI accession number AB058780) was a generous gift from T Nagase, KAZUSA DNA Research Institute, Chiba, Japan
Preparation of asialo-glycoproteins Fetuin and a1-acid glycoprotein (10 mgÆmL)1) were incu-bated for 1 h in 0.05Msulfuric acid at 80C The asialo-products were then neutralized, dialyzed and lyophilized prior to use The carbohydrate content of asialo-glycopro-teins was analysed by GC-MS [30]
Construction of expression vectors of hST6Gal II and transfections
A truncated form of hST6Gal II lacking the first 33 amino acids of the open reading frame, was generated by PCR amplification using a 5¢ primer containing an EcoRI site, 5¢-CCGACAGGAATTCCGCTGAGCCTGTACCCAGC TCCC-3¢ (nucleotides 91–127, Fig 1A) and 3¢ primer containing a BamHI site, 5¢-ACATTGGATCCCAAG AAACCCTTTTTAAGAGTGTGG-3¢ (nucleotides 1577–
1614, Fig 1A) A full-length open reading frame of ST6Gal II was also prepared by PCR amplification using
a 5¢ primer containing an EcoRI site, 5¢-CCCTCTGA ATTCAGACACAAGGTGCTGACCGCAGAG-3¢ (nuc-leotide 1–35, Fig 1A) and the 3¢ primer described above The 25 lL of PCR mixture consisted of 1 unit of DyNazyme EXT, 0.3 lM of each primer, 0.2 mM dNTP and 1.5 ng of plasmid DNA Reactions were run using the following conditions: 1 min at 96C, 4 min at 72 C for 40 cycles Two amplification fragments of 1656 bp and
1522 bp, respectively, were obtained and subcloned in the Topo TA cloning vector (Invitrogen, USA) The inserted fragments were cut out by digestion with BamHI and EcoRI and inserted into BamHI and EcoRI sites of the pFlag-CMV-1 expression vector, which contains the preprotrypsin leader sequence Restriction enzymes digestions and DNA sequencing by Genoscreen (Lille, France) confirmed the cDNA sequence and the insert junctions The resulting plasmids encoded a soluble fusion protein consisting of the Flag sequence and a truncated form of ST6Gal II (pFlag-sST6Gal-II) or the whole coding region of ST6Gal II (pFlag-wST6Gal-II) COS-7 cells were grown in DMEM with 4.5 gÆL)1 glucose without glutamine supplemented with 10% fetal bovine serum,L-glutamine 20 mM, penicillin, streptomycin at 37C under 5% CO2
Twelve micrograms of Qiagen-purified pFlag-sST6Gal-II
or pFlag plasmids were transiently transfected into COS-7
Trang 4cells in a 100-mm diameter dish using LipofectAMINE
PLUS reagent, following the manufacturer’s instructions
The medium was harvested 48 h after transfection The
enzymatic protein expressed in the medium was used as the
enzyme source
Western blot analysis of a soluble ST6Gal II
Nine milliliters of media from COS-7 cells transfected with
the expression plasmid pFlag-sST6Gal II and from
mock-transfected cells were concentrated into 2 mL on Macrosep
30K and 1.5 mL of this preparation were incubated with
150 lL of CMP-Agarose beads (2.8 lmolÆmL)1 CMP)
After washing, the supernatants were discarded and beads
were boiled for 5 min in 100 lL SDS/PAGE loading buffer,
centrifuged and loaded on a 4–20% gradient
polyacryl-amide gel under reducing conditions After Western blotting,
the nitrocellulose membrane was incubated with 10 lgÆmL)1
anti-Flag BioM2 mAb Alkaline phosphatase-labelled goat
anti-(mouse IgG) Ig was used as the second antibody and
revealed by Nitro Blue
tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate and X-phosphate staining
Confocal microscopy
CHO cells were grown in aMEM with glutamax,
supple-mented with 10% fetal bovine serum, penicillin and
streptomycin at 34C under 5% CO2 CHO and COS-7
cells were transiently transfected with the pFlag plasmid or
pFlag plasmid containing the full-length ST6Gal II cDNA
(pFlag-wST6Gal-II) Briefly, 15 000 cells were seeded on
eight chamber slides (LAB-TEK Nalgen Nunc
Interna-tional) and cultured in standard conditions until
mid-confluence Then cells were transfected with 0.25 lg of
purified plasmid per well in 200 lL OPTIMEM, using the
Lipofectamine Reagent Plus kit After transfection, cells
were cultured for 24 h in fresh medium containing fetal
bovine serum Cells were then fixed for 30 min at 4C with
4% paraformaldehyde and quenched 30 min with 50 mM
NH4Cl in phosphate buffered saline (NaCl/Pi) Cells were
then saturated for 30 min at 4C with 2%
polyvinylpyr-olidone in Tris buffered saline (NaCl/Tris) and incubated
with digoxigenin-labeled Sambucus nigra agglutinin
(10 lgÆmL)1) After washing with NaCl/Tris, the cells were
further saturated for 30 min with the blocking reagent
S nigraagglutinin-DIG was revealed using anti-digoxige-nin-fluorescein Fab fragments diluted 1 : 100 in NaCl/Tris, 1% BSA Laser confocal microscopy analysis was per-formed using a Zeiss-instrument (Model LSM 510)
Construction of recombinant baculoviruses and production of soluble ST6Gal I and ST3Gal III in Sf9
In order to express soluble and His6-tagged enzymatic forms
of human ST6Gal I (GenBank accession number X17247) and rat ST3Gal III (clone ST3N-1 [31]), the 5¢ end of these genes were modified The cytoplasmic tail and the trans-membrane domain were deleted and the signal peptide sequence of a viral gene ecdysone-S-glycosyltransferase (EGT) was inserted [32] For this purpose, a 247 bp PCR fragment corresponding to the hST6Gal I amino acids 52–122 was amplified using a pflag/hST6Gal I plasmid, generated by PCR from HepG2cDNA library (C Baisez and A Harduin-Lepers, unpublished data), as the template and two specific primers For 6I 5¢-CGATGAATTC GTTAACGCTCATCACCATCACCATCACGGGAAA TTGGCCATGGGGT-3¢ containing a HpaI site and Back
and subcloned into pUC19 for further sequencing (Euro-gentec, Belgium) This plasmid was then digested with AvrII and BamHI and ligated to the remaining 909 bp AvrII– BamHI fragment containing the 3¢ end of the gene purified from the original construct pflag/hST6Gal I This construc-tion named pUC 6hisST6Gal I contained an addiconstruc-tional HpaI site, the last codon of the EGT signal peptide sequence and six histidine codons The modified hST6Gal I sequence was then excised as an 1104 bp HpaI–BamHI fragment and inserted into the HpaI–BglII sites of pUC-PSEGT The pUC-PSEGT plasmid was generated by inserting a 78 bp fragment encoding the signal peptide sequence of EGT gene [32] This 78 bp fragment was obtained after the annealing
of the two following synthetic oligonucleotides For EGT 5¢-GATCCGCCACCATGACCATCTTATGTTGGCTCG CTCTCCTGAGCACACTCACAGCTGTTAACGCTG ACATCA-3¢ and Back EGT 5¢-GATCTGATGTCAGCG TTAACAGCTGTGAGTGTGCTCAGGAGAGCGAG CCAACATAAGATGGTCATGGTGGCG-3¢ (In order
to avoid homologous recombination between the two EGT sequences, codons were degenerate.) For rST3Gal III, a
108 bp DNA fragment containing a HpaI site, the last codon of the EGT signal peptide sequence, six histidine codons and 19 codons corresponding to amino-acid residues 34–52 was reconstituted using a set of nine overlapping synthetic oligonucleotides and four unique restriction sites AvrII, MunI, HindIII and SacI The DNA fragment was subcloned in a pUC plasmid and sequenced The reconstituted fragment was HpaI–SacI cut and cloned
in HpaI–SacI sites of the pUC-PSEGT plasmid described above The resulting construct was digest with AvrII–MunI
to receive the 1900 bp AvrII–EcoRI fragment prepared from pBS SK ST3Gal III (clone ST3N-1 [31]) The plasmid pUC PS6HisST3Gal III was obtained The full length modified hST6Gal I gene was then excised after digestion with BamHI and HindIII and inserted at the BglII–HindIII sites of the p119 transfer vector designed for recombination into the p10 locus of the baculovirus giving rise to the
p119-PS ST6Gal I construct [33] The full-length modified
Fig 1 Nucleotide and predicted amino acid sequence (A) and
hydro-pathy profile (B) of human ST6Gal II, and (C) comparison of the
sia-lylmotifs L, S and VS of hST6Gal II with those of previously cloned
sialyltransferases (A) Numbering of the cDNA begins with the
initi-ation codon The amino-acid sequence is shown in single letter code.
The putative N-terminal transmembrane domain is boxed Putative
N-glycosylation site (N-X-S/T) are marked with asterisks (*) and
O-glycosylation sites (NetOGlyc 2.0) with back dots (d) Sialylmotifs
L, S and VS are underlined (B) The prediction of transmembrane
region has been determined by the Dense Alignment Surface method
according to Cserzo et al (1997) [49] The portions (positive numbers)
above the horizontal dotted line correspond to hydrophobic regions.
(C) The sialyltransferases protein sequences were aligned using the
CLUSTAL W algorithm Amino acid identities are marked with asterisks
and dots indicate a position that is well conserved.
Trang 5rST3Gal III gene was excised with BamHI and HindIII and
inserted at the BglII–HindIII site of the p119 transfer vector
giving the p119-PS ST3Gal III construct Sf9 cells (ATCC
CRL1711) were cotransfected by lipofection [34] using
DOTAP with the transfer vectors and purified viral DNA
The recombinant baculoviruses were plaque purified and
viral clones were tested for sialyltransferase activity as
described below
Sialyltransferase assays
Sialyltransferase assays were performed as described
previ-ously [30,35–36] In brief, enzyme activity was measured in
0.1M cacodylate buffer pH 6.2, 10 mM MnCl2, 0.2%
Triton CF-54, 50 lMCMP-[14C]Neu5Ac (1.85 KBq), with
one of the acceptor substrate (2 mgÆmL)1for glycoprotein
or 1 mMfor arylglycosides and oligosaccharides) and 23 lL
of the enzyme source in a final volume of 50 lL The
reactions were performed at 37C for 4 h Reaction
products were separated from CMP-[14C]Neu5Ac
depend-ing on the acceptor substrate For glycoproteins, the
reaction was terminated either by precipitation and
filtra-tion as previously described [35] or by SDS/PAGE After
Western blotting, the radioactive products were detected
and quantified by radio-imaging using a Personal Molecular
Imager FX (Bio-Rad, France) Quantification was
per-formed within the linear range of standard radioactivity
For arylglycosides, the reaction was stopped with the
addition of 1 mL H2O and products were isolated by
hydrophobic chromatography on C18 SepPak cartridges
(Millipore Corp., Milford, MA, USA) For free
oligosac-charides, the reaction mixture was heated at 100C for
5 min, centrifuged and subjected to a paper descending
chromatography (Whatman 3) in the following solvent:
pyridine/ethyl acetate/acetic acid/H2O (5 : 5 : 1 : 3, v/v/v/v)
and the radioactive products were detected and quantified
by radio-imaging Under these conditions, the product
formation from the individual acceptor substrates was
linear up to 8 h
For kinetic analysis, incubations were performed as
described above using various concentrations of acceptor
substrates: 0–500 lM of CMP-Neu5Ac, 0–500 lM of
Galb1–4GlcNAcb-O-octyl, or 0–5 mgÆmL)1 of asialo-a1
-acid glycoprotein Kinetic parameters were determined by
Lineweaver–Burk plots and the Km for asialo-a1-acid
glycoprotein is expressed in mM relative to the 18 mol
terminal Gal residues per mol of human of asialo-a1-acid
glycoprotein [37]
Linkage analysis by sialidase digestion
For linkage analysis, asialo-a1-acid glycoprotein sialylated
either with the soluble ST6Gal II or, for the control, with
soluble ST6Gal I and ST3Gal III, was precipitated with
ethanol and air dried, dissolved in water and then digested
with specific sialidase (sialic acid linkage analysis kit, Glyco
Inc., USA): NANase I (specific for a2,3-linked sialic acid,
0.5 mUÆlL)1), NANase II (specific for a2,3 and a2,6-linked
sialic acid, 1 mUÆlL)1), or NANase III (specific for a2,3,
a2,6and a2,8/9-linked sialic acid, 0.5 mUÆlL)1) at 37C for
1 h Digested materials were then analysed by SDS/PAGE
and the radioactive products were analysed by radio-imaging
Multiple tissue expression array and northern analysis
An EcoRI–BamHI 1.6kb fragment of the human ST6Gal
II cDNA and a 1.8 kb human b-actin cDNA (Clontech) used as a positive control for Northern were labelled with [a-32P]dCTP by random priming using the Rediprime II DNA labelling system The human multiple tissues array membrane and Northern blot were probed according to the manufacturer’s instructions and analysed by radio-imaging
Results
Identification and isolation of human ST6Gal II cDNA Similarity searches using the tBLASTn algorithm in the human expressed sequence tag (EST), high throughput genomic sequences (HTGs) and human genomic sequences divisions of the GenBankTM/EBI databases at the National Center for Biotechnology Information allowed us to identify nucleotide sequences with significant similarities
to hST6Gal I (X17247 [22]) These sequences (GenBank accession numbers AB058780, BC008660, AA385852; EST clones and AC108049, AC016994, AC005040 and NT_005429; genomic clones) were subsequently used to reconstitute a nucleotide sequence potentially encoding a sialyltransferase as yet not described Clone AB058780 represented a full-length cDNA sequence already cloned from hippocampus [38] whereas EST clones BC008660 and AA385852, found in ovary adenocarcinoma and in thyroid, respectively, represented truncated nucleotide sequences Oligonucleotides were designed and partial cDNA sequence (nucleotides 940–1629) was obtained by RT-PCR using neuroblastoma cells NSK total RNAs as template (data not shown) This nucleotide sequence (GenBank accession number AJ512141) was subcloned and sequenced and found to be 100% identical to the AB058780 corresponding sequence However, we failed to amplify the corresponding full-length open reading frame
as one fragment and thus we further worked with the clone AB058780 kindly provided by T Nagase, Kazusa DNA Research Institute (Kisarazu, Chiba, Japan) As shown in Fig 1A, the nucleotide sequence contains an open reading frame of 1586bp encoding a putative 529 amino-acid polypeptide with three putative N-glycosyla-tion sites and two putative O-glycosylaN-glycosyla-tion sites Hydro-pathy profile analysis of the predicted protein (Fig 1B) suggests that it has the structural organization of a membrane-bound type II glycoprotein, which is com-monly described for Golgi glycosyltransferases This polypeptide shows a short cytosolic region of 11 acid residues, a single hydrophobic segment of 20 amino-acid residues and a large luminal catalytic domain (498 amino-acid residues) Comparison of the primary structure
of this new sialyltransferase with that of the 18 other cloned human sialyltransferases indicates that there are significant similarities in the three sialyltransferases con-served regions named sialylmotifs L, S and VS (Fig 1C)
In particular, this protein shares with hST6Gal I a common motif YEXXP in the sialylmotif S where the glutamic acid residue (E) is present only in these two proteins This analysis strongly suggests that this protein represents a new sialyltransferase and since this
Trang 6polypep-tide shows 48% overall identity with the human ST6Gal I,
we have named it hST6Gal II
The gene organization of hST6Gal II was reconstituted
from the genomic clones previously identified and found to
localize on human chromosome 2 (2q11.2-q12.1) As
presented in Fig 2A, in a similar manner to the hST6Gal I
gene found on human chromosome 3 (3q27-q28), hST6Gal
II gene divides into five exons and spans over 38 kb of
human genomic sequence Sequence comparison of each
exon shows that these two genes share high similarities in E2,
E3, E4, E5 (Fig 2B) This analysis, as well as the
dendro-gram of the cloned human sialyltransferases (Fig 3),
suggests a common ancestral gene for the two ST6Gal genes
that have evolved independently
ST6Gal-II gene expression
In order to determine the expression pattern and the size of
hST6Gal II mRNA, Northern blotting was performed using
the ST6Gal II cDNA (1.6 kb fragment) as a probe As shown in Fig 4A, among the 12 human tissues examined, hST6Gal II mRNA was detectable only in brain as an 8.0 kb transcript An expression array of 72 different human tissues and eight different control RNAs and DNAs, was also probed with the 1.6kb hST6Gal II cDNA (Fig 4B) hST6Gal II gene appears to be expressed in lymph node, to a lesser extent in testis, thyroid gland, caudate nucleus, temporal lobe, hippocampus, and fetal tissues (brain, kidney, thymus, liver), and rather weakly in placenta, lung, aorta, amygdala, occipital and parietal lobe and salivary gland Almost no expression was observed in fetal lung and heart, uterus, bladder, kidney, duodenum, trachea, Burkitt’s lym-phoma, and colorectal adenocarcinoma These data lead us to the conclusion that hST6Gal II gene is weakly expressed in
a very restricted manner, which is in contrast to hST6Gal I which is expressed in most of human tissues [39]
Expression of a recombinant hST6Gal II
In order to facilitate functional analysis of the enzyme, a truncated cDNA of hST6Gal II lacking the first 33 amino acids of N-terminus region was generated by PCR from the human cDNA clone AB058780 The putative catalytic domain was fused to a Flag octapeptide (DYKDDDDK) and transiently expressed in COS-7 cells This construction including a preprotrypsin signal produced a soluble form
of the enzyme secreted from the cells This soluble Flag-ST6Gal II fusion protein produced in cell culture media was concentrated on CMP-agarose beads, subjected to SDS/PAGE and Western blotting and visualized as a
70 kDa band (Fig 5) A smaller 40 kDa band was also observed, probably as the result of proteolytic degradation
in the cell culture medium To monitor the activity of the soluble form of hST6Gal II, media from cells transfected with pFlag-sST6Gal II or control plasmid were collected after 3 days of transfection and assayed for sialyltrans-ferase activity, using various acceptor substrates (Table 1)
We also simultaneously carried out the same enzymatic
Fig 3 Dendrogram of the cloned human sialyltransferases The
deduced amino-acid sequences of the catalytic domain (starting 10
amino-acids upstream of the sialylmotif L) of the cloned human
sia-lyltransferases were aligned by CLUSTAL W and the corresponding
phylogenetic tree was constructed using the neighbour-joining method.
Fig 2 Comparison of the genomic structure
the human ST6Gal I and ST6Gal II genes.
(A) Exon structure of hST6Gal I and
hST6Gal II genes are represented by boxes
and are denoted E1 to E5 Darkened boxes
with their size (in bp) indicated above,
repre-sent the protein coding sequences and opened
boxes represent untranslated sequences Solid
lines between the exons represent the intron
sequences (not drawn to scale); their sizes are
indicated below (B) Comparison of the
deduced amino acid sequence of the
hST6Gal II gene with those of the hST6Gal I
gene.
Trang 7assay with a recombinant soluble form of hST6Gal I
produced in the Sf9 cells hST6Gal II was shown to be able
to transfer a sialic acid residue onto a terminal Gal residue
of asialofetuin and asialo-a-acid glycoprotein As shown
in Table 1, the best acceptor substrate of hST6Gal II was the free disaccharide Galb1–4GlcNAc, lacto-N-neotetraose and Galb1–4GlcNAcb-O-octyl whereas hST6Gal I pre-ferred Galb1–4GlcNAc-R disaccharide linked to a protein
as found in asialo-a1-acid glycoprotein No significant activity was observed towards sialylated glycoproteins such
as native fetuin and a1-acid glycoprotein, or type 1 containing structures such as Galb1–3GlcNAc, Galb1– 3GlcNAcb-O-octyl or lacto-N-tetraose
The kinetic parameters of hST6Gal II were determined using CMP-Neu5Ac as the donor substrate, and using Galb1–4GlcNAcb-O-octyl and asialo-a1-acid glycoprotein
as the acceptor substrates The apparent Km value of hST6Gal II for CMP-Neu5Ac (59 lM) was very close to those previously described for native or recombinant hST6Gal I which range from 33 to 50 lM [40,41] The apparent Kmvalue for asialo-a1-acid glycoprotein (0.12 mM) was also in the same range than that determined for the recombinant hST6Gal I (0.10 mM) [41] On the other hand, the apparent Kmvalue of hST6Gal II for Galb1–4GlcNAcb-O-octyl was 0.74 mM, which is significantly lower than the values determined for native (1.78 mM) or recombinant (2.38 mM) ST6 Gal I using Galb1–4GlcNAc [41]
Fig 4 hST6Gal II gene expression in various human tissues.
(A) Northern blot analysis Commercially prepared Northern blot
(Clontech) with 1 lg poly(A)+RNA from various adult human
tis-sues were probed with a 1.6kb 32P-random-labelled hST6Gal II
cDNA as described in the Materials and methods section and a 1.8 kb
human b-actin cDNA control probe (upper and lower panels,
respectively) RNA size marker bands are indicated on the left side of
the blot Sizes of the detected mRNA are indicated on the right.
(B) Expression array analysis of the expression of hST6Gal II in
various human tissues Commercially prepared Multiple Tissue
Expression (Clontech) array with poly(A + ) RNA from 72 different
human tissues and eight different control RNAs and DNAs was
probed with32P-random-labelled hST6Gal II cDNA.
Fig 5 Immunoblotting of hST6Gal II recombinant protein from transfected cell culture media Cell culture media from pFlag-sST6Gal II and mock-transfected cells (48 h after transfection) were incubated with CMP-Agarose beads The beads were washed, boiled and subjected to SDS/PAGE under reduced conditions and Western blotting using the BioM2 anti-Flag mAb The positions of the high range prestained SDS/PAGE standards are indicated in KDa on the left side of the figure Lane 1, 50 lL from mock transfected cells; lane 2, 50 lL from pFlag/ST6Gal-II transfected cells.
Trang 8Linkage analysis
To determine the incorporated sialic acid linkage, sialidase
digestions of asialo-a1-acid-glycoprotein, sialylated with
either hST6Gal II, hST6Gal I, or hST3Gal III, and
subse-quent electrophoresis of the digested products were
per-formed As shown in Fig 6, the incorporated14C-labelled
sialic acid was resistant to treatment with a2,3-specific
sialidase compared to asialo-a1-acid-glycoprotein sialylated
with ST3Gal III used as a positive control of the NANase I
specific action The radioactive material was completely
removed upon treatment with a2,3/6-sialidase or
a2,3/6/8-sialidase, indicating that the product formed is indeed NeuAca2–6Gal
ST6Gal II induces the expression of NeuAca2–6Gal structures at the cell surface of transfected cells
In order to visualize a phenotypic change in the hST6Gal II expressing cells, the full-length ORF of hST6Gal II was inserted in the pFlag expression vector and transfected into CHO or COS-7 cells Forty-eight hours after transfection, cells were incubated with digoxigenin-labelled S nigra agglutinin, a lectin that recognized NeuAca2–6Gal/GalNAc structures, and revealed with an anti-digoxigenin fluoresc-ein-labelled Fab fragment A negative control without lectin was also performed As observed by confocal microscopy (Fig 7), hST6Gal II induces the over-expression of Neu-Aca2–6Gal structures at the cell surface of transfected CHO cells whereas mock transfected cells weakly expressed NeuAca2–6Gal No fluorescence was detected in the control, indicating the specificity of the fluorescence detec-tion The same result was also obtained with COS-7 cells (data not shown)
Discussion
The ST6Gal subfamily Unlike all the other sialyltransferases cloned to date and characterized [36], ST6Gal I was so far the unique member
of the b-galactoside a2,6-sialyltransferase subfamily to be identified This unique gene is widely expressed in human tissues and the enzyme is mainly involved in the a6-sialy-lation of membrane and secreted glycoproteins However, other b-galactoside a2,6-sialyltransferases with different
Fig 6 Analysis of linkage specificity of hST6Gal II [14
C]Neu5Ac-labelled asialo-a 1 -acid glycoprotein was produced using soluble
recom-binant ST6Gal II, ST6Gal I or ST3Gal III The sialylated labelled
products were subjected to sialidase treatment with NANase I (specific
for a2–3-linked sialic acid, lane 2), or NANase II (specific for a2–3/
6-linked sialic acids, lane 3), or NANase III (specific for a2–3/6/8/
9-linked sialic acids, lane 4), or none (lane 1) The resulting products
were separated on SDS/PAGE and detected by radio-imaging.
Table 1 Comparison of the acceptor substrate specificity of ST6Gal I and ST6Gal II Acceptor substrates were used at a concentration of 1 m M for arylglycosides and 2 mgÆml -1 for glycoproteins Relative rates are calculated as a percentage of the incorporation of sialic acid onto asialo-a 1 -acid glycoprotein A value of 0 indicates less than 0.4 % d bn, benzyl; pNp, para-nitrophenol.
Acceptor Structures
Relative rate (%) hST6Gal II hST6Gal I Asialo-a 1 -acid glycoprotein Galb1-4GlcNAc-R a
100 (0.52)b 100 (13.86)b
a 1 -Acid glycoprotein NeuAca2-6Galb1-4GlcNAc-R 0 0 Fetuin NeuAca2-3Galb1-3GalNAca1-O-Ser/Thrc 0 1.4
NeuAca2-3Galb1-3[Neu5Aca2-6]GalNAca1-O-Ser/Thr c
NeuAca2-6(3)Galb1-4GlcNAc-R c
Asialofetuin Galb1-3GalNAca1-O-Ser/Thr 66 83
Galb1-4GlcNAc-R
LNnT: Galb1-4GlcNAcb1-3Galb1-4Glc 623 120 LNT: Galb1-3GlcNAcb1-3Galb1-4Glc 0 0
a
Rrepresents the remainder of the N-linked oligosaccharide chain.b Actual activities are shown in brackets in pmolÆh)1ÆlL)1.c Data from Spiro & Bhoyroo [50].
Trang 9substrate specificity or preferences were expected to exist to
account for the presence 6-sialylated oligosaccharides such
as 6¢-sialyl-lactose, sialyl-lactosamine or
sialyl-lacto-N-neo-tetraose found in human milk [42,43],
monosialylgan-glioside NeuAca2–6Galb1–4GlcNAcb1–3Galb1–4Glc-Cer
immunostained in human meconium or
NeuAca2–6Gal-NAcb1–4GlcNAc structures found on pituitary hormones
As the human genome is being deciphered, we gain access to
a large number of sialyltransferase-gene related sequences
through screening of the databanks This strategy allowed
us to identify a new human sialyltransferase gene located on
chromosome 2 with high similarity to hST6Gal I located on
chromosome 3, both in terms of gene organization and
sequence Our data clearly indicate that these two genes may
have a common ancestral gene and after dispersion in the
human genome would have evolved independently From
an evolutionary point of view, we could also identify the rat
homologue of this new gene (in genomic clones AC094827
and AC106335; data not shown), and also the mouse
homologue located on mouse chromosome 17 (XM140080)
Several EST (BU055532, BB651169, BB552328) expressed
either in the neonate cerebellum or in pregnant mouse
oviduct were assembled and the corresponding protein
sequence deduced A protein sequence comparison of the
two homologues is shown in Fig 8, which indicates 77%
identity between them Two cDNAs corresponding to
partial mRNAs were also found in the zebra fish databanks
(BE606075, BM103887) which suggest that this ST6Gal II
protein appeared early in the evolution
Sequence analysis
Sequence analysis of the deduced protein showed that this
protein has one of the longest stem region (around 200
amino acid residues) whereas the size of the catalytic
domain is conserved among the different sialyltransferases
It is interesting to note that hST6Gal II protein shared 48%
overall identity with hST6Gal I and even higher identities within the sialylmotifs, with 67%, 56% and 90% identity for the L, S and VS motif, respectively In the catalytic domain, hST6Gal II protein shows six cysteine residues that are strictly conserved in hST6Gal I protein Ma and Colley (1996) [44] have described formation of a disulfide-bonded dimer of ST6Gal I that is catalytically inactive but retains its ability to bind galactose The presence of a faint band around 140 kDa in Fig 5 suggests that this dimerization may also occur for hST6Gal II Human ST6Gal II protein shows also three potential N-glycosylation sites, two of which lie within the sialylmotif L (Fig 1) These two sites are conserved in the mouse ST6Gal II protein whereas the third-one, located in the stem region, is missing The influence of N-glycosylation on the activity and trafficking
of ST6Gal I has been previously investigated [45] It appears that these N-glycosylation sites are required for the activity and endoplasmic reticulum to Golgi transport of the soluble form of ST6Gal I However, the position of these glycosy-lation sites is not conserved in ST6Gal II In addition, we have shown that the soluble hST6Gal II recombinant protein is secreted in the culture medium of COS-7 cells as
a 70 kDa polypeptide (Fig 5) Taking into account the expected molecular mass of the nonglycosylated polypep-tide (58 kDa), we can predict that these N-glycosylation sites are occupied Further analyses will be required to determine whether or not the N-glycosylation could influ-ence ST6Gal II activity
Gene expression pattern The results of Northern and expression array analyses clearly indicated a restricted and low level expression of hST6Gal II gene as an 8 kb transcript mainly in brain, which is in accordance with the size of the cDNA clone AB058780 (6.782 kb) It was found also in specific regions
of the brain: hippocampus and amygdala of the limbic
Fig 7 S nigra agglutinin staining of CHO/hST6Gal II transfected cells CHO cells were transiently transfected with pFlag-wST6Gal II vector encoding the full length ORF of hST6Gal II cDNA or with the pFlag vector by means of LipofectAMINE-reagent plus Expression of NeuAca2– 6Gal was detected using digoxigenin-labelled S nigra agglutinin and an anti-digoxigenin-fluorescein Fab fragment, and the fluorescence was detected by confocal microscopy (A) CHO/pFlag-wST6Gal II cells, (B) CHO/pFlag cells, (C) CHO cells anti-digoxigenin-fluorescein Fab frag-ment.
Trang 10system, caudate nucleus and the cerebral cortex temporal
lobe Very low levels of mRNA were detected also in lymph
node, testis, thyroid gland and fetal tissues This low and
restricted level of expression is in agreement with the little
number of ST6Gal II EST in databanks This is in contrast
to the expression pattern of the ST6Gal I gene, which is
abundantly expressed in almost all human tissues examined,
including fetal tissues but with the notable exceptions of
testis and brain [39] where it is expressed at lower levels One
can postulate that expression of ST6Gal II in these two
tissues could compensate for the lower expression of
ST6Gal I or be responsible for the specific synthesis of
a6-sialylated glycoconjugates We also identified two
upstream untranslated exons, exon A (119 bp) found in
the EST BE613250 and located 42 941 bp upstream the first coding exon, and exon B (62 bp) found in the mRNA sequence AB058780 and located 42 058 bp upstream the first coding exon (data not shown) The hST6Gal II gene would thus drive the expression of at least two individual mRNAs through the use of two distinct promoters Very preliminary analysis of the two upstream genomic sequences identified in the databanks indicated the presence potential trans-acting factors binding sites such as SP1 and TBP binding sites found upstream exon B This finding would argue for a ubiquitous and low-level expression of exon B containing transcripts On the other hand, the presence of CREB and SREBP binding sites upstream exon A would indicate a specific expression of exon A containing
Fig 8 Comparison of the putative amino-acid sequence of hST6Gal II and mST6Gal II The two vertical dots indicate identical amino acid residues and single dots indicate similar amino-acid residues The underlined amino acid residues indicate the sialylmotifs L, S and VS Putative conserved N-glycosylation site (N-X-S/T) are marked with asterisks (*).