and GlcNAcPl-3GalNAc-R to mucin core 4, GlcNAc~l-3(GlcNAc~1-6)GalNAc-R. Substrate specificity studies indicate that the enzyme attaches. GlcNAc to either ...
Eur. J. Biochem. 157,463-474 (1986) 0FEBS 1986
Mucin synthesis Conversion of RI-P1-3Gal-R2to R1-~1-3(GlcNAcf11-6)Gal-Rz and of R1-~1-3GalNAc-R2to R1f11-3(GlcNAcPl-6)GalNAc-R2by a P6-N-acetylglucosaminyltransferasein pig gastric mucosa Inka BROCKHAUSEN', Khushi L. MATTA', Jeanne ORR3, Harry SCHACHTER3, Anky H. L. KOENDERMAN4 and Dirk H. van den EIJNDEN4 Department of Medical Genetics, University of Toronto Roswell Park Memorial Institute, Buffalo, New York Hospital for Sick Children, Toronto Department of Medical Chemistry, Vrije Universiteit, Amsterdam (Received December 9, 1985/February 10, 1986) - EJB 85 1333
A UDP-GlcNAc:R,-p1-3Gal(NAc)-R2 [GlcNAc to Gal(NAc)] P6-N-acetylglucosaminyltransferaseactivity from pig gastric mucosa microsomes catalyzes the formation of GlcNAcfil-3(GlcNAcB1-6)Gal-R from GlcNAcpl-3Gal-R where -R is -jl-3GalNAc-or-benzyl or -~l-3(GlcNAc~l-6)GalNAc-a-benzyl. This enzyme is therefore involved in the synthesis of the I antigenic determinant in mucin-type oligosaccharides. The enzyme also converts Galj1-3Galpl-4Glc to Gal~l-3(GlcNAcfl1-6)Galfll-4Glc. The enzyme was stimulated by Triton X-100 at concentrations between 0 and 0.2% and was inhibited by Triton X-100 at 0.5%. There is no requirement for MnZ+and the enzyme activity is reduced to 65% in the presence of 10 mM EDTA. Enzyme products were purified and identified by proton NMR, methylation analysis and j-galactosidase digestion. Competition studies suggest that this pig gastric mucosal p6-GlcNAc-transferase activity is due to the same enzyme that converts Galfil-3GalNAc-R to mucin core 2, Gal~l-3(GlcNAc~l-6)GalNAc-R, and GlcNAcPl-3GalNAc-R to mucin core 4, GlcNAc~l-3(GlcNAc~1-6)GalNAc-R. Substrate specificity studies indicate that the enzyme attaches GlcNAc to either Gal or GalNAc in p(1-6) linkage, provided these residues are substituted in p(1-3) linkage by either GlcNAc or Gal. The insertion of a GlcNAcpl-3 residue into Galpl-3GalNAc-R to form GlcNAcpl3Galj1-3GalNAc-R prevents insertion of GlcNAc into GalNAc. These studies establish several novel pathways in mucin-type oligosaccharide biosynthesis. These core structures may carry a number of elongation sugars, some of which form blood-group antigenic determinants, e.g. the developmentally regulated i and I antigens, which are represented by R5-GalP1-4GlcNAcfil-3Gal-R6 and R5-Galj1-4GlcNAcfll -3(R7-GalPl-4GlcNAcpl -6)Gal-R6 sequences respectively. The GlcNAcPl-3(GlcNAc~1-6)Gal-R structure is found on various mucin oligosaccharides with core 1 : R1-Gal~l-3GalNAca-R2 core classes 1, 2 and 3 [l-71. The i and I structures are also found on N-linked oligosaccharides (GlcNAcfl-Asn linked) core 2 : R1-Galp3-3GalNAca-Rz and glycolipids. The synthesis by a pig gastric mucosal j?3-GlcNAc-transIPl-6 ferase of the GlcNAcpl-3Gal linkage in mucin-type oligosaccharides with core classes 1 and 2 has been described R3-GlcNAc recently [8, 91. A similar enzyme has been found in serum and in Novikoff tumour cells and ascites fluid; it attaches GlcNAc core 3: R4-GlcNAc~1-3GalNAccr-R2 in j1-3 linkage to the terminal Gal of GalPl-4GlcNAc-R and GalP1-4Glc-R [lo- 141. The synthesis of the I antigen using core 4: R4-GlcNAcfil-3GalNAcc-Rz human serum as the enzyme source and lactose as an acceptor has been reported [15]. Recently, Piller et al. [16] have deifil-6 scribed a UDP-GlcNAc: GlcNAcB1-3Gal(-R) (GlcNAc to Gal) P6-N-acetylglucosaminyltransferasefrom hog gastric Rs-GlcNAc mucosa. This enzyme was shown to convert GlcNAcjlCorrespondence to I. Brockhausen, Department of Medical Ge- 3Gal~1-4Glc~-O-methyl to GlcNAc~l-3(GlcNAc~l-6)Galjlnetics, University of Toronto, Toronto, Ontario, Canada M5S 1A8 4Glc-O-methyl. Pig gastric mucosal microsomes were shown This is paper I in a series on Mucin synthesis; paper 6 appeared to add GlcNAc to a variety of oligosaccharide and glycolipid elsewhere [I71. acceptors with the GlcNAcpl-3Gal-R structure but mucinAbbreviutions. NeuAc, N-acetylneuraminic acid; ACS, aqueous counting scintillant; HPLC, high-performance liquid chromatog- type acceptors were not studied. In this paper, we describe the synthesis by a pig gastric raphy; GC-MS, gas chromatography/mass spectrometry. Enzyme. P6-N-Acetylglucosaminyltransferase(EC 2.4.1 .lo2 and mucosal P6-GlcNAc-transferase of the GlcNAcPl-6Gal structure in acceptors with mucin-type oligosaccharidescontaining EC 2.4.1.148).
Mucin-type oligosaccharides (GalNAcor-Ser/Thr-linked) occur in mucins and in membrane-bound and secreted glycoproteins. These sugar chains may vary in length and complexity. Their core structures can be classified into four major classes:
464 core classes 1 and 2. By comparing this enzyme activity to other j6-GlcNAc-transferase activities in pig gastric mucosa, we conclude that the GlcNAc-transferase can act on subterminal Gal or GalNAc residues which are substituted in j(13) linkage by Gal or GlcNAc and that, at least in the pig gastric mucosa, a single enzyme is responsible for these activities.
Enzymepreparations
Pig gastric and colonic mucosal microsomes and dog submaxillary gland microsomes were prepared as described previously [171. Microsomal suspensions were stored in o.25 at - 7o "c, Analytical methody
EXPERIMENTAL PROCEDURES Materials
The following materials were purchased from commercial sources : Mes from Calbiochem-Behring ; ACS from Amersham ;acetonitrile, ultraviolet grade, from Caledon Laboratories or Lichrosolv grade from Merck AG, Darmstadt; AG 2 x 8 (200 -400 mesh), AG 1 x 8 (100 - 200 mesh), Bio-Gel P4 (minus 400 mesh) and AG 50W-X2 (H', 200-400 mesh) from Bio-Rad; Triton X-100, UDP-Gal and UDP-GlcNAc from Sigma; UDP-[3H]GlcNAc (6.6 Ci/mol) from New England Nuclear, Boston, MA. UDP-N-[l-'4C]acetylglucosamine (5 Ci/mol) was synthesized as previously described [17]. Bovine testicular fl-galactosidase was a generous gift from Dr G. W. Jourdian (Ann Arbor, MI, USA). The following compounds were prepared by organic synthesis: GlcNAc/31-3Galj-O-methyl [I 81, GlcNAcj13GalNAca-benzyl and GlcNAcjl-6GalNAca-benzyl [19], Galjl-3(GlcNAcj1-6)GalNAca-benzyl [20], GlcNAcjl3Galj1-3GalNAccc-benzyl[21],Galjl-3GalNAca-benzyl[22] and GlcNAcj1-3Gal~l-3(GlcNAcfl1-6)GalNAca-benzyl (K. L. Matta, unpublished). The following compounds were kindly provided by D. Whitfield and J. Krepinsky (Ludwig Cancer Institute, Toronto): GlcNAcjl-6Galp-0-CD3, Fucctl-3Ga1p-0-CD3, GalPl -4GlcNAcj-O-CD3 and 3-0-methyl-Galj-0-CD3. Galj1-3Gal~l-4Glcfrom marsupial milk was generously provided by Dr Michael Messer (University of Sydney, Australia). Galfll-4GlcNAc was donated by Dr A. Veyrieres (Universite Paris-Sud, Orsay, France). GlcNAcjl-3Gal was a gift from Dr L. Anderson (University of Wisconsin, Madison, WI, USA). 3-0-Methyl-Gal was synthesized by methylation 1231 followed by acid treatment (90% trifluoroacetic acid, 10 min at 20°C) of 1,2: $6 di-0-isopropylidene-D-galactofuranose, which was a gift of Dr G. J. F. Chittenden (University of Nijmegen, Nijmegen, The Netherlands). Galal3Galjl-4Glc was synthesized using a purified preparation of UDP-Gal:N-acetyllactosaminide al-3-galactosyltransferase [24] kindly donated by Dr D. H. Joziasse (Vrije Universiteit, Amsterdam, The Netherlands). Details of the procedure will be published elsewhere. UDP-GlcNAc :N-acetyllactosaminide ~1-3-N-acetylglucosaminyltransferase, partially purified from Novikoff tumour cell ascites fluid [14], was used to synthesize GlcNAcjl-3Galjl-4GlcNAc and GlcNAc~l-3Galjl-4Glc. The synthesis was carried out in a system which contained in 800 pl: 0.1 M sodium cacodylate, pH 7.2, 20 mM MnCI2, 4 mM ATP, 80 mM GlcNAc, 5 mM UDP-GlcNAc, 150 mM Galfll4Glc or 17 mM GalPl4GlcNAc and 2 mU enzyme preparation. The mixture was incubated for 24 h at 37°C and then passed through a column of 1 ml Dowex 1-X8, Cl--form. After washing, the combined eluates were fractionated on a column (1.6 x 200 cm) of Bio-Gel P-4 (200-400 mesh), equilibrated and eluted with 0.05 M ammonium acetate pH 5.2. The product peak was pooled and lyophilized. NeuAca2-3Galfl1-4Glc was isolated by HPLC as described before [25] from sialyllactose (Sigma), which consists of a mixture of the 3'- and 6-sialyllactose isomers.
Protein was determined by the method of Lowry et al. [26], using bovine serum albumin as the standard. High-performance liquid chromatography (HPLC)
Enzyme products were isolated by HPLC with an LKB system containing a 21 52 HPLC controller, 2150 pumps and a 2151 variable wavelength monitor. Elution of compounds was followed by monitoring the absorbance at 195 nm and counting radioactive fractions in ACS with a liquid scintillation counter. Guard columns packed with C18 (Whatman) were used for protection of HPLC columns. Benzyl derivatives were separated on a Partisil PXS 5/25 PAC column (Whatman) with an acetonitrile/water mixture (83 :17) at a flow rate of 1 ml/min and pressures of 9 - 15 MPa. Products formed with Galj1-3Galp1-4Glc were separated using a 5-pm Alltech NH2 column (250 x 4.6 mm) with acetonitrile/water (67: 33) at a flow rate of 0.7 ml/min and pressures of 10 15 MPa. For substrate specificity studies, HPLC was also performed as described for assay system 3. Proton nuclear magnetic resonance spectroscopy
Samples were prepared for proton NMR by D 2 0exchange and analyzed at room temperature with a Nicolet 360-MHz spectrometer as described previously [17] or at 400 MHz with a Bruker 400 MSL NMR spectrometer as indicated in Table 7. Proton resonances were assigned by comparison with spectra of known compounds and by spin-decoupling experiments. Methylation analysis
Samples of 150 nmol, containing 150 nmol inositol as a carrier, were permethylated, hydrolyzed, reduced with NaBD, and acetylated as described previously [27]. Methylated and acetylated derivatives were analyzed by gas chromatography/ mass spectrometry (GC-MS) as described previously [17]. The product formed with Gal~1-3Galj1-4Glcwas reduced prior to methylation analysis by reaction with 1 M NaBH4 for 4 h at room temperature, passage through AG 50W-X2, repeated flash evaporations from methanol to remove borate and purification on a Bio-Gel P-4 (minus 400) column (1.6 x 84 cm) equilibrated in H20. GlcNAcfll-6GalNAcabenzyl was subjected to methylation analysis as a parallel control. GlcNAc-transferase assays Assay system 1. This contained in a total volume of 50 pl: 0.1 M Mes pH 7,lO mM MnC12,0.2% Triton X-100,2.8 mM UDP-N-[l-'4C]acetylglucosamine(476 dpm/nmol), 20 p1 enzyme (0.12 - 0.32 mg protein) and acceptor. After incubation at 37°C for 2 h the reaction was stopped by the addition of 400 1 1 20 mM borate/l mM EDTA and the solution was passed through a Pasteur pipette filled with AG 2-X8 (200400 mesh) and washed with 2.6 ml H 2 0 . 17 ml ACS were
465 added to the eluate and the radioactivity counted with a scintillation counter. Results from control incubations lacking acceptor were routinely subtracted. Assay system 2. This contained in a total volume of 50 pl: sodium cacodylate pH 2.6 mM UDP-N-[1-14C1O.' acetyklucosamine (740dpm/mol), 20 Pig gastric Illucosal microsomes (0.12 mg protein) and acceptor. The assay was carried out as described for assay system 1 except that AG 1X8 (100-200 mesh) was used. Assay system 3 . This contained in a total volume of 50 pl: 0.1 M sodium cacodylate pH 7, 2.8 mM UDP-[3H]GlcNAc (2200 dpm/nmol), 4 mM ATP, pig gastric mucosal microsomes (0.35 mg protein) and acceptor. The mixtures were incubated for 2 h at 37°C. The incubation was stopped by passing the mixture through a I-ml column of Dowex l-XS, CI- form. The column was washed twice with 0.5 ml H20, and the combined eluates were lyophilized. After dissolving the material in 50 pl HzO, samples were analyzed by HPLC. Products were identified by comparison with standards. The product was clearly separated from t3H]GlcNAc and was used to calculate transferase activity. 77
Product identification
Table 1. Tissue distribution of GIcNAc-transferase activities Assays were carried Out as described in Experimental Procedures for GlcNAc-transferdse assay system 1. PGMic, pig gastric mucosal microsomes (0.12 mg protein/assay); P C M k Pig colonic mucosal microsomes (0.32 mg proteinlassay); CSMic, canine submaxillary gland microsomes (0.14 mg protein/assay); Bz, benzyl; the results are averages of at [east two separate determinations Substrate (2 mM)
Enzyme activity PGMic
PCMic
CSMic
nmol h-' mg-' 26.1
17.1
GalP,~3~GlcNAcB1~6~Ga,NAca~Bz 68.4 7,3
Galpl -3GalNAcu-Bz
238.5 43.9
1.3
1.4 4.7
GlcNAcpl-3Gal~1-3GalNAca-Bz5.4
20.5
16.4
GlcNAcpl-3GalB1-3~alNAca-Bz Ratio ofGalP1-3GalNAcu-B~to
Table 2. Effict of metal ions on branching ~6-GlcNAc-transferasse activity Enzyme activity was measured as described in Experimental Procedures for GlcNAc-transferase assay system 2 with 0.5 mM GlcNAcPl-3Gal/31-3GalNAca-benzyl as the acceptor. The results are averages of at least two separate determinations
Enzyme products were prepared on a large scale using the following three substrates: GlcNAc~l-3Gal~1-3GalNAmbenzyl, GIcNAc~l-3Gal~l-3(GlcNAc~1-6)GalNAccr-benzyl Additions to the assay Enzyme activity (cf. control) and GalP1-3Gal/?l-4Glc. nmol h-' mg-' (%) 600 nmol of each of the three acceptors were incubated in a total volume of 1.2 ml containing 2.6 mM UDP-N- None(contro1) 31.2 (100) [l-'4C]acetylglucosamine (740 dpm/nmol), 0.1 M sodium 10 mM MnCI2+ 29.5 (95) 26.0 (83) cacodylate pH 7, 0.1% Triton X-100 and pig gastric mucosal 10 mM MgCIZ+ 34.5 (111) microsomes (2.88 mg protein). The mixtures were incubated l o mM cac12+ (10) 3.3 at 37°C for 20-23 h. After 6-8 h, another 2.04 mg enzyme 0 (0) protein was added. Reactions were stopped by freezing and mM EDTA 20.2 (65) the addition of 6 ml 1 mM EDTA/ZO mM borate. Each enzyme product was divided into six aliquots and each aliquot was passed through a column (4 ml each ) of AG 1-X8 (100 200 mesh) and washed with 11 ml H 2 0 . The eluates were flash-evaporated and applied to Bio-Gel p-4 columns or o-nitrophenyl) as acceptor has been identified as GalP1-3(1.6 X 88 cm) equilibrated in H 2 0 . Pooled radioactive (GIcNAcP1-6)GalNAc-R for pig gastric mucOSa [8, 91 and fractions were further purified on HPLC and analyzed by canine submaxillary glands [28]. Galpl-3(GlcNAcpl-6)proton NMR and methylation analysis. Product formed with GalNAca-benzyl is a specific acceptor for the elongation 83Galbl-3GalPl-4Glc was also analyzed by HPLC before and GlcNAc-transferase which attaches GlcNAc in pl-3 linkage after digestion with bovine testicular P-galactosidase. to Gal [8]. The activity detected with G l c N A c ~ l - 3 G a l ~ l 3GalNAccr-benzyl as an acceptor is a novel branching (GlcNAc to Gal) p6-GlcNAc-transferase activity (see below). P-Gulactosidase digestion Pig gastric mucosa is the richest source for all three enzyme The incubation mixture contained in a total volume of activities. 40 pl: 8.3 nmol enzyme product (6135 dpm), 10 pl buffer The 86-GlcNAc-transferase acting on GlcNAcpl-3GalpH 4.3 (5.7 p10.1 M citric acid, 4.3 pl0.2 M Na2HP04),40 pg PI-3GalNAca-Bz was further characterized using pig gastric bovine serum albumin and 9 mU bovine testicular P-ga- mucosa as an enzyme source. Enzyme assay 2, lacking Mn2+, lactosidase (3 U = amount catalyzing transfer of 1 pmol/ was used since these conditions are inhibitory for the p3min). In control incubations, enzyme was omitted. The GlcNAc-transferase which adds GlcNAc in PI-3 linkage to mixtures were incubated at 37 "Cfor 4 h, frozen and subjected terminal Gal residues. The activity was linear with time for to HPLC. 2 h and a pH optimum was found at 7 (data not shown). The effect of metal ions and EDTA is shown in Table 2. None of the metal ions showed significant stimulation of enzyme RESULTS activity. 10 mM M n 2 + , Mg2+ or Ca2+ had little effect; 10 mM Co2+reduced the activity by 90% and 10 mM Zn2+ Assay conditions inhibited the enzyme totally. 10 mM EDTA was a relatively Table 1 shows GlcNAc-transferase activities obtained weak inhibitor. with three different acceptors and microsomes prepared from The effect of Triton X-100 was studied with three pig gastric and colonic mucosa, and dog submaxillary glands. substrates (Fig. 1): GlcNAc~l-3Gal,!W-methyl, GlcNAcPlThe product formed with Gal/?1-3GalNAcor-R (R = benzyl 3Gal~1-3GalNAca-benzyl and GlcNAcpl-3Gal~l-3(Glc-
i: E:
466 NAcPl-6)GalNAca-benzyl. With all three substrates, there was inhibition of enzyme activity at 0.5% Triton X-100, but significant stimulation around 0.1YO.Freezing and thawing twice and storage at -20°C for several months reduced the activity towards GlcNAc~l-3Gal/3-O-methylby 18%. Substrate specijicity studies
Pig gastric mucosal microsomes transferred GlcNAc to several acceptors as listed in Table 3. Competent acceptors have a subterminal Gal substituted in p(1-3) linkage by a GlcNAc (compounds 2, 7, 8, 9, 17, 18) or Gal (compound 3)
0.1 0.2 0.5 TRITON X-100 CONCENTRATION ( o h )
or have a subterminal GalNAc which is p(1-3) substituted by a GlcNAc (compound 16) or Gal (compound 15). Compounds 15 and 16 have previously been reported to be good acceptors for the /?6-GlcNAc-transferases which synthesize Galj?l-3(GlcNAcfi1-6)GalNAc-R [28] and GlcNAcpl-3(GlcNAc~l-6)GalNAc-R[17, 291, respectively. Detailed product identifications were carried out for compounds 3,17 and 18 as described below. In all three cases, the major product was due to the incorporation of a GlcNAc in p(1-6) linkage to the subterminal Gal residue. HPLC studies on the products of compounds 2, 7, 8 and 9 suggest that these were also due to the incorporation of GlcNAc into the subterminal Gal residue (data not shown). Compounds with a terminal Gal residue but lacking a 3-substituted subterminal Gal or GalNAc (compounds 1, 6, 11) were ineffective as acceptors (Table 3), as were compounds with a subterminal Gal 3-substituted in a-linkage (compare compound 3 with compounds 4 or 14) or 3-substituted with a methyl group (compounds 10 and 12). Compound 5 (3'-sialyllactose), however, showed 10% activity. GlcNAcp1-6Galj-O-CD3 (compound 13) is an ineffective acceptor whereas GlcNAcPl -3Galp-O-CH3 (compound 9) is a good acceptor, using either assay conditions 1 (containing Mn2+, data not shown) or assay conditions 2 (no Mn", Table 3). Assay conditions 1 permit action of the B3-GlcNActransferase incorporating GlcNAc in p(1-3) linkage to Gal [8, 91. These data indicate the pathway to the GlcNAcP13(GlcNAc/?l-6)Gal-R structure to be via GlcNAcpl-3Gal-R and not GlcNAcpl-6Gal-R.
Fig. 1. Dependence of GlcNAc-transferuse activities on Triton X-100 concentration. GlcNAc-transferase was measured using assay system 2 with 0.5 mM acceptor at various Triton X-100 concentrations. Competition experiments (.-a) GlcNAcfll-3Galfll-3GalNAca-benzyl;(0-0) Mixed substrate experiments using assay system 2 were CTlcNAc/?l-3Galfl-O-methyl; ( A__ A) GlcNAc~l-3Gal/?1-3(Glccarried out to investigate whether the activities shown in NAcfll-6)GalNAca-benzyl Table 3. Substrate specificity of pig gastric mucosal branching p6-GlcNAc-transferuse Enzyme activities were measured as described for GlcNAc-transferase assay system 2; 100% activity corresponds to 130 nmol h- ' mg- '. For results marked with an asterisk, assays were carried out with a less active enzyme preparation, which had been stored at -2O"C, using GlcNAc-transferase assay system 3; 100% activity corresponds to 9.8 nmol h-' mg-'. K,,, values were calculated from Lineweaver-Burk plots. n.d. = not determined; Bz, benzyl; Fuc = L-fucose Acceptor (2 mM)
1. Gal~1-4Glc 2. GlcNAcfll-3Gal/?l-4Glc 3. Gal/?1-3Galfll-4Glc 4. Gala1-3Galfl1-4Glc 5. NeuAccr2-3Gal~l-4Glc 6.Galfll-4GlcNAc 7. GlcNAcflI-3Galfll-4GlcNAc 8. GlcNAcfl1-3Gal 9. GlcNAcf11-3Galj?-O-CH3 10. 3-0-Methyl-Gal 11. Galf11-4GlcNAcfl-O-CD3 12. 3-O-Methyl-Gal/?-O-CD3 13. GlcNAcf11-6Galfl-O-CD3 14. Fuccrl-3Galfl-O-CD3
GlcNAc incorporation
Km
vm,,
%
mM
nmol h- mg -
< 1*
n.d. 3.4
n.d. 150
n.d. n.d. n.d. n.d. n.d. n.d. 2.1
n.d. n.d. n.d. n.d. n.d. n.d. 118
n.d. n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d. n.d.
0.9 1.8 1.6
286 262
0.9
73
43 55* 100 2* 10*
< 1* 25* 18* 39 14* l /+[ l S1/K1 + WKZI and v = Wd~dK2)1/[1+ S1/KI and 4 [17, 291 from Galfil-3GalNAca-benzyl and GlcNAcP- [+~ lS(z~/ lK/ zK] ,lwhere S1 and S2 are the concentrations of the two sub1-3GalNAca-benzyl, respectively, appeared to be due to the strates, V1 and V , are the V,,, values and K1 and K2 are the K,,, same enzyme. values. The kinetic values used in the calculations are: K, = 1.1 mM In addition, competition was shown between GlcNAcb1- and V,,, = 22.5 nmol h-'/O.l2mg for 'di'; K, = 2.75 mM and 3GalNAca-benzyl and GlcNAc~l-3Galfil-3GalNAca-benV,,,,, = 17.2 nmol/h-'/0.12 mg for 'tetra' zyl; between GlcNAcPl-3GalNAca-benzyl and GlcNAcPlSubstrate conc Transferase activity 3Galb-0-methyl; between GlcNAc~l-3Galfil-3GalNAcabenzyl and GlcNAcfil-3Gal~-O-methyl; and between tetra expt calculated GlcNAcPl-3GalNAcpa-benzyl and GlcNAcfll-3Gal#?l- di 3(GlcNAcfll-6)GalNAca-benzyl. This indicates that the same di tetra 2E 1E enzyme in pig gastric mucosa acted on all of the acceptors used in this experiment, i.e. the same fi6-GlcNAc-transferase di tetra di tetra appears to be responsible for the synthesis of core 2, core 4 and the GlcNAc~l-3(GlcNAc~1-6)Gal-structure. mM nmol h-'/0.12 mg ~
Product identification
Products formed with acceptors 3, 17 and 18 (Table 3 ) were purified and analyzed. In large-scale product preparations with all three substrates more than 80% conversion of substrate to product was achieved.
2 6
-
2 6
-
16.9 20.9
-
3 4 3 4
-
7.5 8.1 3.0 1.7
-
11.7 17.1
16.9 20.9
7.5 8.1
10.5 15.5
4.8
3.2
468 1
8 1 s17
i
D
>
-
'Ti
F
N
10
B X
6 5 1
E
k
4
F
3
0 Q
2
0
a 1
0 Q
>
0
t a Q
0 IT
CT
0
20
40
60
80
100
120
- , - I
140
160
ELUTION VOLUME (ml)
Fig. 2. Purifcation on Bio-Gel P-4 of enzyme product using GlcNAcfll3GalbI-3GalNAcu-henzyl as the acceptor. The column (88 x 1.6 cm) was eluted with water; aliquots of fractions were counted in ACS for radioactivity
r
I
1
I
I
5 10 15 20 25 30 35 TIME(rnin)
Fig. 3. HPLC purification of enzyme product using GIcNAcBI-3Galfll3GalNAcu-benzyl as the acceptor. A PAC column was used with acetonitrile/water (83:17) at a flow rate of 0.7 ml/min. Elution of standard compounds is indicated by arrows: (1) GlcNAc; (2) GlcNAcfll-3Gal/?-O-methyl; (3) GlcNAcPl-3Ga1/31-3GalNAcabenzyl; (4) Gal~l-3(GlcNAc~1-6)GalNAcct-benzyl; (5) GlcNAcbl3Galfl1-3(GIcNAcfil-6)GalNAca-benzyl
GlcNAcpI-3Ga1~1-3GalNAca-benzyl (compound 17) as the substrate
hexitol (eluting at 15.3 min) due to 3-substituted GalNAc; After the large-scale incubation using compound 17 as the and 2,4-di-O-methyl-l,3,5,6-tetra-O-acetylhexitol (eluting at substrate, the mixture was passed through AG 1 and Bio-Gel 13.0 min) due to 3,6-disubstituted Gal. The molar ratios P4 columns. Two radioactive peaks were obtained (Fig. 2, calculated from the total ion current were 2: 0.5 : 1.5 respecpeaks I and 11). The smaller peak I1 contained radioactive tively. No derivatives from 3- or 4-monosubstituted or 3,4- or GlcNAc from breakdown of UDP-GlcNAc, as determined 2,3-disubstituted Gal or 3,4- or 3,6-disubstituted GalNAc on HPLC by comparison with standard GlcNAc. The first were detectable. The data prove the structure of enzyme prodpeak from Bio-Gel P-4 eluted between 102 ml and 112 ml uct to be GlcNAc~1-3(['4C]Gl~NAc~l-6)Gal~l-3GalNAca(Fig. 2, peak I) and was further purified on HPLC using benzyl. a PAC column (Fig. 3). Radioactive product eluted in the tetrasaccharide region at 19 min in one sharp peak. The radioactive fractions were pooled and subjected to proton NMR GlcNA cfi I -3Galp 1-3 (GlcN Acfi 1-6)GalNAca-benzyl (compound 18) as the substrate spectroscopy and methylation analysis. Proton NMR. Fig. 4 shows parts of the 360-MHz proton Enzyme product using compound 18 (Table 3) as the subNMR spectra of GlcNAc~l-3Gal~1-3GalNAcct-benzyl sub- strate eluted as one major peak on Bio-Gel P-4 between 88 ml strate (Fig. 4A) and enzyme product (Fig. 4B). The NMR and 96 ml (data not shown). This pooled peak was further parameters are listed in Table 6. The addition of a fiGlcNAc purified on HPLC using a PAC column and eluted at 33 rnin residue in the product is indicated by an extra N-acetyl signal (data not shown). This chromatographic behaviour of the and a new doublet at 4.540 ppm (J1,z = 7.9 Hz), characteristic compound is consistent with that of a pentasaccharide-benzyl for the H-1 of a fl6-linked GlcNAc [3]. Major changes are (Fig. 3). seen in the H-3 resonances of GalNAc (from 4.027 ppm to Proton NMR. The NMR spectra and parameters of 3.982 ppm), as well as the H-4 resonances of Gal (from enzyme substrate and product are shown in Fig. 6A and 6B, 4.117 ppm to 4.086 ppm), indicating that the GlcNAc is linked respectively, and in Table 6. In the product spectrum there is to the 6 position of the 3-linked Gal rather than GalNAc. an additional resonance arising from the H-1 of a P6-linked Significant changes are also observed on the N-acetyl signal GlcNAc. It is difficult to assign the two P6-GlcNAc H-1 of the P3-linked GlcNAc, on H-1 of Gal, and H-4, H-5 and doublets (at 4.526 ppm and 4.560 pprn), since apparently the H-6 (but not the N-acetyl signal) of GalNAc, probably due H-1 signal of the GlcNAc residue pl-6 linked to GalNAc has to conformational changes in the molecule. Another indica- shifted significantly from 4.540 ppm. A new N-acetyl signal tion for the addition of GlcNAc to Gal is the nonidentity of also indicates addition of GlcNAc in the product. Significant the product spectrum with that of GlcNAcjl-3Galpl-3- changes are observed in the chemical shifts of GalNAc H-3 (GlcNAcPl-6)GalNAca-benzyl (Table 6 and Fig. 6A). (4.006 ppm to 4.15 ppm) and Gal H-4 (4.114ppm to To confirm the structure and exclude the possibility of a 4.077 ppm), indicative of a new linkage between GlcNAc and GlcNAcPl-4/2Gal linkage in the enzyme product, methyla- Gal. The chemical shifts of the benzyl methylene hydrogens, H-1 of the B3-linked GlcNAc and H-1 of Gal are also affected tion analysis was undertaken. Methylation analysis. After permethylation, hydrolysis, re- to a lesser degree. The H-4 signal of GalNAc is unchanged, duction and acetylation, compounds were analyzed by GC- indicating that GlcNAc has not been linked to the Chydroxyl MS. Fig. 5C shows the total ion current of enzyme product of GalNAc. formed from GlcNAc~l-3Gal~1-3GalNAca-Bz. The mass Methylation analysis. Partially methylated derivatives of ions of all major peaks were analyzed; in addition, selected enzyme product were analyzed by GC-MS. Fig. 5B shows the ion monitoring helped to identify sugar derivatives. GC-MS total ion current of the derivatives. We detected the presence of enzyme product derivatives showed the presence of 3,4,6- of 3,4,6-tri-0-methyl-1,5-di-O-acetyl-2-deoxy-2-N-methyltri-O-methyl-l,5-di-O-acetyl-2-deoxy-2-N-methylacetamidoacetamidohexitol (eluting at 13.8 min) due to terminal hexitol (eluting at 13.8 min) due to terminal GlcNAc; 4,6-di- GlcNAc; 4-O-methyl-l,3,5,6 tetra-O-acetyl-2-deoxy-2-N-meth0-methyl- 1,3,5-tri-O-acetyl-2-deoxy-2N-methylacetamido- yl-acetamidohexitol (eluting at 16.6 min) due to 3,6-di-
469 A.
GlcNAcrSl-3 Galpl3GalNAc CY benzyl
H-1 GaINAco
A
6.
GlcNAcPl-3 Galol-3 GalNAc benzyl 1/31-6 GIcNAcl
