Glycoconjugate Journal (1995) 1 2 : 3 7 1 - 3 7 9 Substrate specificity and inhibition of UDPGIcNAc:GlcNAc131-2Man(z1-6R 131,6-N-acetylglucosaminyltransferase V using synthetic substrate analogues INKA BROCKHAUSEN 1'2., F O L K E R T R E C K a, W I L L I A M K U H N S t, S H A H E E R K H A N 3., K H U S H I L. M A T T A 3, E R N S T M E I N J O H A N N S 4, H A N S P A U L S E N 4, R A J A N N. S H A H 5, M I C H A E L A. B A K E R 6 a n d H A R R Y S C H A C H T E R ~,2 1Research Institute, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada 2Department of Biochemistry, University of Toronto, Toronto, Ontario, Canada 3Department of Gynecologic Oncotogy, Roswell Park Memorial Institute, Buffalo, NY, USA 4Institut fiir Organische Chemic, UniversitgitHamburg, Hamburg, Germany 5Department of Medical Genetics and Biophysics, University of Toronto, Toronto Ontario, Canada 6Department of Medicine, Toronto Hospital, Toronto, Ontario, Canada Received, 4 November, 1994, Revised, 10 January, 1995 UDP-GtcNAc:GlcNAc 131-2MancO-6R (GtcNAc to Man) t]l,6-N-acetylglucosaminyltransferase V (GIcNAc-T V) adds a GIcNAc[31-6branch to bi- and triantennary N-glycans. An increase in this activity has been associated with cellular transformation, metastasis and differentiation. We have used synthetic substrate analogues to study the substrate specificity and inhibition of the partially purified enzyme from hamster kidney and of extracts from hen oviduct membranes and acute myeloid leukaemia leukocytes. All compounds with the minimum structure GIcNAcl31-2Manc~l-6Glc/Manl3R were good substrates for GlcNAc-T V. The presence of structural elements other than the minimum trisaccharide structure affected GlcNAc-T V activity without being an absolute requirement for activity. Substrates with a biantennary structure were preferred over linear fragments of biantennary structures. Kinetic analysis showed that the 3hydroxyl of the Man, l-3 residue and the 4-hydroxyl of the Manl]- residue of the MancO-6(Man~l-3)Manl3-R N-glycan core are not essential for catalysis but influence substrate binding. GlcNAc~l-2(4,6-di-O-methyl-)Man~l6Glcl3-pnp was found to be an inhibitor of GlcNAc-T V from hamster kidney, hen oviduct microsomes and acute and chronic myeloid leukaemia leukocytes. Keywords: GlcNAc-transferase V, substrate specificity, inhibition, leukaemia, N-linked glycans Abbreviations: all, allyl; AML, acute myeloid leukaemia; BSA, bovine serum albumin; CML, chronic myelogenous leukaemia; Gal, G, D-galactose; Glc, D-glucose; GlcNAc, Gn, N-acetyl-D-glucosamine; HPLC, high performance liquid chromatography; Man, M, D-mannose; mco, 8-methoxycarbonyl-octyl, (CH2)sCOOCH3; Me, methyl; MES, 2-(N-morpholino)ethanesulfonate; oct, octyl; pnp, p-nitrophenyl; T, transferase. Introduction Complex N-glycans have been implicated in many diverse functions [1], particularly in cell-cell adhesion [2, 3] and in diseases such as metastatic cancer [4, 5]. It has been known for many years that transformed and malignant cells usually present on their surfaces complex N-glycans that are larger than normal due primarily to a combination of increased branching, sialytation and poly-N-acetyllactosamines [4-8]. UDP-GlcNAc:GlcNAc[~I-2Man~xI-6R (GlcNAc to Man) $ Present address: Perkin-Elmer, Applied Biosystems Division, 850 Lincoln Center Drive, Foster City, CA 94404, USA. * To whom enquiries should be addressed. 0282-0080© 1995 Chapman & Hall [~l,6-N-acetylglucosaminyltransferase V (GlcNAc-T V), which adds a GlcNAc[~I-6 branch to bi-and triantennary N-glycans, plays a major role in this phenomenon. Comparatively higher amounts of tetraantennary chains, concomitant with higher GlcNAc-T V activity, have been reported after Rous sarcoma virus [9, 10] and Potyoma virus [11] infection of baby hamster kidney cells. The increase in activity was associated with an increase in Vm,x suggesting an induction of GlcNAc-T V [12]. T24H ras transformed rat fibroblasts and metastatic SP1 mammary carcinoma cells also exhibit higher GlcNAc-T V activity [13]. Induction of the ras gene with dexamethasone in a stable NIH 3T3 transfectant containing a normal N-ras proto-oncogene under the control 372 Brockhausen et al. of a glucocorticoid-inducible promoter resulted in increased branching of complex N-glycans and increased GtcNAc-T V activity [14]. An increase in GlcNAc-T V-dependent branching has been reported in malignant human breast tissue compared to benign hyperplastic lesions [15] and in other oncogene-transformed tissues [ 16, 17]. L-phytohaemagglutinin (L-PHA) reactivity was shown to correlate with GlcNAc-T V activity [15], suggesting that this lectin is useful in detecting N-glycan antennae initiated by [31-6-1inked GIcNAc. The N-glycan GtcNAcl31-6ManccI-6Man~- branch, initiated by GlcNAc-T V, favours the addition of poly-N-acetyllactosamine chains [18, 19]; highly branched complex N-glycans and poly-N-acetyllactosamines carrying cancerassociated antigens have been found in many animal models of tumour progression and acquisition of metastatic potential, and in human melanomas and carcinomas of breast, colon and baby hamster kidney (BHK) cells is 0.18 mM for GlcNAc[312Manocl-6Man[3-mco and 0.25 mM for UDP-GIcNAc [10]. GlcNAc-T V cannot act on substrates which contain a bisecting GlcNAc, or in which either antenna is substituted with a [31-4-1inked Gal residue [23, 26, 33, 34]. Removal (deoxy analogue) or substitution of the 4-OH of the terminal GlcNAc of the substrate GlcNAc~l-2Manocl-6Glcl3-octyl leads to an inactive compound [35]. Although a bisecting GlcNAc attached to the 4-OH of the 13-1inked Man residue on a biantennary substrate prevents enzyme action, 4-O-methyl or 4-deoxy linear substrate analogues aJ'e excellent substrates [31]. The GlcNAc-terminal triantennary compound is a better substrate than the biantennary compound [11, 26]. GlcNAc[312(4-deoxy)Manocl-6Glcl3-octyl is a good substrate but the (4-O-methyl)Man derivative is all inhibitor but not a substrate ovmes [4]. In this study, we investigated a panel of synthetic linear and biantennary compounds as substrates and inhibitors of GlcNAc-T V. GlcNAcl31-2(4,6-di-O-methyl)Manccl-6Glc~3pnp was found to be an inhibitor of GlcNAc-T V from hamster kidney, hen oviduct microsomes and acute myeloid (AML) and chronic myeloid (CML) leukaemia leukocytes. Kinetic analysis showed that the 3-hydroxyl of the Mantel-3 residue and the 4-hydroxyl of the Man[3- residue of the Manczl6(Manc~l-3)Man[~-R N-glycan core are not required for catalysis but influence substrate binding. A number of reports suggest the involvement of GIcNAc-T V in cellular activation and differentiation. Upon differentiation of human colonic adenocarcinoma Caco-2 cells, GlcNAcT V activity increased concomitantly with increased GlcNAc-T II, III, and IV activities and a loss of fucosylated poly-N-acetyllactosaminoglycan chains [20]. Mouse F9 teratocarcinoma cells acquired higher GlcNAc-T V activity upon retinoic acid induced differentiation [21]. Interleukin-6 was found to stimulate GIcNAc-T IV and V activities in a human myeloma cell line concomitant with a decrease in GlcNAc-T III [22]. GlcNAc-T V was first described in 1982 [23] and catalyses the conversion of bi- to tri-antennary and of tri- to tetraantennaxy N-glycans: GlcNAc~I-6 \ GlcNAc[~I-2 Mano~l-6 GlcNAc[31-2 Mantel-6 \ Manl3-R / [+/-GlcNAc[31-4] Mano¢1-3 / GlcNAc[31-2 \ ----> Manl3-R / [+/-GlcNAc[~ 1-4] Manor1-3 / GlcNAc[31-2 The enzyme has been purified from hamster kidney [24], rat kidney [25] and from the culture supernatant of human lung cancer cells [26]. The cDNA encoding the enzyme was isolated from rat and mouse [27] and human fetal liver [28, 29]. The human gene was mapped to chromosome 2q21 [28, 29]. GlcNAc-T V also acts on linear structures representing the Mano¢l-6 arm of the N-glycan core, i.e. GlcNAc[~I-2Manc~I6Manl3-R where R may be a hydrophobic group or BSA [30-32]. Like other [31,6-GlcNAc-transferases, the enzyme is fully active in the presence of EDTA [25]. The compound GlcNAc~l-2(6-deoxy)Manccl-6Glc[3-octyl was shown to be a competitive inhibitor of GlcNAc-T V with a Ki of 0.07 mM [12, 24]; this compound lacks the hydroxyl to which the enzyme attaches GlcNAc. The KM(app)of GlcNAc-T V from [36]. Materials and methods Materials" AGI-x8 (100-200 mesh, C1- %rm) and Bio-Gel P4 @400 mesh) were purchased from Bio-Rad. Bovine serum albumin and Triton X-100 were purchased from Sigma. Acetonitrile (190 UV cutoff) was from Fisher Scientific Co. or Caledon Laboratories. UDP-N-[1-14C] acetylglucosamine was synthesized as described previously [37] and diluted with UDPGlcNAc from Sigma. GlcNAc[~l-2Manc~l-6Glc[3-oct and GlcNAcl31-2Manotl-6Man~-mco were kindly provided by Dr. O. Hindsgaul, University of Alberta, Edmonton. The following compounds were chemically synthesized: GlcNAc~l-2Manocl-6Glc[3-pnp [38]; GlcNAc~l-2Manocl6Man[3-Me [391; GlcNAcl31-2Manocl-6Glc[3-all, GlcNAcl312Manocl-6Glc~31-4Glc~-all [38]; GlcNAc[31-2Man~I6Glc~t-4GlcNAc[40]; GlcNAc[31-2(4-O-Me)Mano~l-6Glc[~pnp [41]; G l c N A c l 3 1 - 2 ( 6 - O - M e ) M a n o c l - 6 G l c l 3 - p n p, G l c N A c [~1-2(4,6-di- O-Me)Mano~l-6Glc[3-pnp [40]: GtcNAc[~ 1-2Manoc-Me, GlcNAc[~ 1-4(GlcNAc[3 l-2)Mano~Me, and GlcNAc[~l-6Mano~-Me [39]: GIcNAc[31-2(6-O-Me) Manor-Me and GlcNAcl31-2(4,6-di-O-Me)Manoc-Me [42]; GlcNAc[31-2Manet 1-6Man~-oct, GtcNAc[31-2Man~l6(GlcNAc [~1-2Mano¢ 1-3)Man [3-oct, GlcNAc ~ 1-2Mano~ t6(GlcNAcl31-2[3-deoxy] Mano¢l-3) Man[3-oct, GIcNAc ~ 1-2M ano¢ t -6(G lcNAc[31-2 [4-deo xy] Mano~ 13)Manl3-oct, GlcNAc[3 t-2Mano~ 1-6(GlcNAcI31-2Manc¢ 1-3)4- Substrate specificity and inhibition of GlcNac-T V O-Me-Man[3-oct, GlcNAc~31-2Mana1-6(GlcNAc1312 M a n o ~ l - 3 ) 4 - d e o x y - M a n l 3 - o c t , GtcNAct31-2Manc~l6(GlcNAc[31-2[4-deoxy]Manod-3)4-O-Me-Man[3-oct [43]; and 3-O-pivaloyl-GlcNAcl3 l-2Manc~t-6Glcl3-pnp (K.L. Matta, unpublished). High performance liquid chromatography HPLC separations were carried out with an LKB or a Waters system [34, 39]. Acetonitrile/water mixtures were used as the mobile phase for all columns. Reverse phase C18, amine (NH2) and amine-cyano (PAC) columns were used, depending on the hydrophobicity of the aglycon. Elutions of compounds were monitored by measuring the absorbance at 195 nm and counting the radioactivity of collected fractions [34]. Preparation of enzyme Homogenates from rat, human and fetal human colon, pig stomach mucosa, leukocytes from patients with acute (AML) and chronic (CML) myelogenous leukaemia and normal human granulocytes, and hen oviduct microsomes were prepared as previously described [34, 37, 44]. GlcNAc-T V was partially purified from hamster kidneys according to Hindsgaul et al. [24] as follows. Fifty hamster kidneys (Keystone Biologicals) were homogenized in 50 ml acetone (-20°C) with several short pulses in a Polytron homogenizer. Homogenate was washed with a further 50 ml cold acetone on a Btichner funnel with standard filter paper. Homogenization, washing and filtration of residue were repeated. The acetone powder produced was homogenized briefly with 100 ml buffer A (0.1 M Na-acetate pH 6.0, 10 mM EDTA, 0.2 M NaC1) in a Teflon hand homogenizer, transferred with a further 100 ml buffer A to centrifuge tubes and centrifuged at 15 300 x g for 30 rain. The supernatant was discarded and homogenization/centrifugation was repeated with the pellet. Finally, the pellet was homogenized with 180 ml cold water and centrifuged as in the previous step. The pellet was homogenized in 60 ml buffer B (0.1 M Tris/HCl pH 7.6, 0.4 M KC1, 1 mM EDTA and 1% Triton X-100) using two passes with a teflon Dounce homogenizer, stirred 4 h, centrifuged at 23 700 x g for 45 rain and supernatant was removed. The pellet was re-homogenized with buffer B (60 ml), stirred overnight, and centrifuged. Both supernatants (crude extracts) were dialysed versus two exchanges of 1 1 buffer C (50 mM MES pH 6.5 with 10 mM EDTA and 0.1% Triton X-100). Dialysed crude extracts were loaded onto a UDP-hexanolamine-Sepharose column (10 ml, I0 ~mol UDPhexanolamine per ml gel) equilibrated with buffer C. The column was prepared by coupling UDP-hexanolamine (Sigma) with cyanogen bromide activated Sepharose (Pharmacia) according to the Pharmacia protocol. The column was washed with 30 ml buffer D (buffer C containing 0.25% Triton X-t00 and 20% glycerol). Enzyme was eluted with buffer D containing a step gradient of NaCI: 0.1 M, 0.25 M, 0.5 M and 1 M, 15 ml each, in 5 ml fractions. Most of the enzyme was in the 0.25 M NaC1 fractions. Enzyme 373 activity (46%) was also found in the flow-through fraction and was applied again to a UDP-hexanolamine-Sepharose column; 50% of enzyme activity was again found in the flowthrough fraction. The 0.25 M NaC1 fractions from both columns were pooled and dialysed against 50 mM Na-cacodytate buffer pH 6.5 containing l0 InM EDTA, 20% glycerol and 0.1% Triton X-100, and concentrated with potyethyleneglycol (molecular mass 10 000) to 50% volume. Dialysis and concentration were repeated and the enzyme was dialyzed 2 more times to yield a total of 1.27 mU in 12 ml (33% of the activity in the crude homogenate) at a specific activity of 0.06 mU mg -~ (1 mU = 1 nmol rain -I) Protein determination Protein was determined by the Bio-Rad method using IgG as the standard. Assay for fl],6-GlcNAc-transferase V partiatly purified from hamster kidney The standard assay mixture for measuring GleNAc-T V activity contained in a total volume of 30 gl: 1 mM UDP-N[ 1-14C]acetylglucosamine (5555 dpm nmol~I), 0.1% Triton X-100, 10 mM EDTA, 20% glycerol, 50 mM cacodylate buffer, pH 6.5, acceptor as indicated and 3.t3 gU enzyme preparation (3.t3 pmol rain -1, 30 ~tl, 53.1 lag protein). Incubations were carried out for 2 h at 37°C and stopped by the addition of 0.4 ml ice cold water. The mixtures were passed through Pasteur pipettes filled with AGI-x8, 100-200 mesh, C1- form, equilibrated in water. The columns were washed with 2.6 ml water and the eluates were counted in 17 ml scintillation fluid. The apparent Km and V,n~xvalues were determined by double reciprocal Lineweaver-Burk plots. Assay for fll,6-GIcNAc-transferase Vfrom microsomes and ceil homogenates The standard assay for measuring GlcNAc-T V activity contained in a total volume of 40 tal: 2 mM UDP-N-[1-1ac] acetylglucosamine (2200 dpm nmo1-1) or UDP-[6-3H] N-acetytglucosamine (1600-4200 dpm nmol-1), 0.25% Triton X-t00, 0.125 M GlcNAc, 10 mM AMP, 0.125 M MES buffer, pH 7, acceptor as indicated in the Tables and 10-20 IA microsomes or homogenates (0.12-0.5 mg protein). Incubations were carried out for 2 h at 37°C and stopped by the addition of 0.4 ml 20 mM Na-tetraborate - 1 mM EDTA or 100 pl water. The mixtures were passed through Pasteur pipettes containing AGt-x8, 100-200 mesh, CI- form, equilibrated in water. The columns were washed with water and the eluates were lyophilized and stored at -20°C. Aliquots were analysed by HPLC as described in the Tables. bzhibition of GlcNAc-transferase V The substrate was incubated at 0.05-0.3 mM concentration under standard assay conditions in the presence of an inhibitor as indicated in the Tables. Potential inhibitors were also preincubated for 10 minutes at room temperature with the enzyme 374 B r o c k h a u s e n et al. 16) had little effect on activity whereas the 3-deoxy analogue (compound 15) was less active (Table 2). Substitution of the 3-hydroxyl of the GlcNAc residue of GlcNAc~31-2ManeO-6Glcl3-pnp by a pivaloyl group (compound 9) prevented catalysis by hamster kidney GlcNAc-T V (Table 2). 3-O-pivaloyl-GlcNAc[31-2Man~l-6Glc[3-pnp (compound 9) showed 19% and 47% activity with the enzyme from pig stomach and AML cells, respectively, compared to GlcNAc[31-2Man~l-6Glcl3-pnp (compound 2), HPLC analysis showed that the pivaloylester was not destroyed by esterases during the incubation and product eluted as one peak before the substrate on reverse phase HPLC (Table 1). Upon incubation of hen oviduct microsomes with short linear compounds (Table 3), at least two products were usually formed which could be separated by HPLC. Product due to GlcNAc-T VI' action [39], GlcNAcI]I-4(GlcNAcl312)Manc~t-6R, always eluted earlier on HPLC than product due to GlcNAc-T V action, GtcNAcI31-6(GlcNAcl31-2)MancO-6R (Fig. 1). When Mn 2+ was added to the assay, GlcNAc-T VI' product was increased about 20-fold and GIcNAc-T V product was not detectable ( Fig. 1). Although there were variations in activity, derivatives with octyl, 8-methoxycarbonyloctyl, p-nitrophenyl, methyl or allyl aglycon groups were active as substrates for GlcNAc-T V from hamster kidney and hen oviduct (Tables 2 and 3). In contrast to GlcNAc-T VI', GlcNAc-T V acted well on the free reducing tetrasaccharide GlcNAc~l-2Man~l-6 GlcI314GlcNAc (compound 7, Tables 2 and 3). Activities towards linear tetrasaccharide substrates were within the same range as for trisaccharides but the disaccharide GlcNAc~ 1-2Man~-Me was significantly less active (compound 8, Table 3). before the addition of the substrate. Since previous work had shown that irradiation at 350 nm in the presence of nitrophenyl substrate derivatives greatly reduced the activity of core 2 136-GlcNAc-transferase acting on O-glycans [45], we tested the effect of UV light on inhibition by pnp-containing compounds. Inhibitors were irradiated at 30°C with UV light at 350 nm for 20 rain in the presence of the enzyme before incubation, using a Rayonet RPR 100 reactor equipped with 16 RPR 3500 ]k lamps. Results Substrate Specificity o f GlcNAc-T V Purified hamster kidney GlcNAc-T V was stable for several months at 4°C and acted on a number of compounds with the general structure GlcNAc131-2Man(xl-6Man/Glc[3-R. The HPLC separation conditions and elution times for various substrates and products are listed in Table 1. The nature of the aglycon group had a strong influence on the effectiveness of the substrate with the octyl compounds giving the highest activities (Tables 2 and 3). Compounds with a biantennary structure (14-19) were significantly better substrates than those with only the linear GlcNAc[~l-2ManoO-6 Man/Glc structure (compounds 1, 5-7, 10). The K m values of the biantennary derivatives ranged from 0.035 to 0.18 mM and enzyme activities were relatively high when compared to the linear substrates (Table 2). Methyl substitution of the 4hydroxyl of the internal Man[~ residue (compound 17) increased activity but omission of this hydroxyl (compound 18) increased the Km 3-fold. Omission of the 4-hydroxyl of the Manal-3 residue of the biantennary substrate (compound Table 1, HPLC conditions for separation of substrates and GlcNAc-T V products Flow rate (ml min -j ) % Acetonitrile in mobile phase C18 1 PAC PAC NH2 C18 PAC C18 PAC NH2 PAC NH2 C18 0.7 0.7 0.7 18 19 82 82 80 15 82 16 80 82 82 82 14 Compound No. Column 1 GlcNAc[31-2Manc~l-6Glc[~-oct 2 3 GIcNAc[31-2Man00-6Glc[~-pnp GIcNAcl31-2Man~1-6Manl3-mco GlcNAc[~1-2Man~ 1-6Man[~-Me GlcNAc[31-2Man~l-6Glc~-all 4 5 6 7 8 9 GtcNAc~ 1-2Manotl-6Glc[~l-4Glc~-alt GlcNAcI31-2Man~1-6Glc[~1-4GlcNAc GlcNAc[~1-2Manc~-Me 3-O-pivaloyl-GlcNAcl31-2Mane~l-6Glcl3-pnp 1 0.7 1 0.7 1 0.7 0.7 1 Elution time (min) Substrate Product 36 43 19 22 41 29 32 34 29 48 90 13 34 25 32 37 51 96 29 75 25 54 112 113 34 24 Compounds were separated by HPLC as described in Methods with acetonitrile/water mixtures as the mobile phase on reverse phase (C18), amine (NH2) or amine-cyano (PAC) columns. Substrate specificity and inhibition of GlcNac-T V Table 2. 375 Specificity of GIcNAc-T V from hamster kidney using linear and biantennary substrates Activity (pmol mg-1per h) Compound Linear Substrates (0.8 mM) t0 GIcNAcl31-2ManeO-6Manl3-oct 1 GIcNAcI31-2MancO-6Gtcl3-oct 7 GlcNAcI31-2ManoA-6Glc~1-4GlcNAc 5 GlcNAc~ 1-2Mano~l-6Glcl3-all 6 GlcNAc~ 1-2Mano~l-6Glc~ 1-4Glc~-all 2730 3180 840 660 < 300 Biantennary Substrates (0.3 raN) 14 GIcNAcl31-2Man~ 1-6(GlcNAc[~ 1-2 Mano~l-3)Manl3-oct 15 GlcNAcl31-2Mana 1-6(GlcNAc[31-2[3-deoxy-]Manc~ 1-3)Man~-oct 16 GlcNAcI31-2Manc~1-6(GIcNAc[~ 1-2[4-deoxy-]Mano~l-3)Manl3-oct 17 GlcNAc~ 1-2Mano~l-6(GlcNAc[31-2Man~l-3)4-O-Me-Man~-oct 18 GIcNAcl31-2Man~ 1-6(GlcNAcl31-2Man~l-3)4-deoxy-Man [~-oct 19 GIcNAc[31-2Mano~1-6(GIcNAcl]1-2[4-deoxy-]Manc~ 1-3)4-O-Me-Man[3-oct 5880 1800 5580 6300 5340 8040 Substrates (2.7 mM) 2 GlcNAcI~ I-2Mano~ 1-6Glcl3-pnp 9 3-O-pivaloyl-GlcNAc[31-2Mancd-6Glcl3-pnp 1460~ < 300~ K~ (mM) 0.09 0.06 0.13 0.08 0.035 0.18 0,06 Assays were carried out using purified hamster kidney GlcNAc-T V as described in the Methods Section without HPLC separation. Substrates were present in the assay at near saturating concentrations; the.activity values therefore represent the Vm,Xvalues. aCompounds were tested with a tess active enzyme preparation. Table 3. Compound Conditions A: 1 1 mM GlcNAc131-2Manc~1-6Glc[3-oct 13 2.5 mM GlcNAcl31-2(4,6-di-O-Me) Man~l-6Glcl3-pnp + 1 mM GlcNAcl31-2Manc~l-6Glcl3-oct 1 2 mM GlcNAc~l-2Manc~l-6Glcl3-oct 5 2 mM GlcNAc~l-2Manc~l-6Glcl3-all 6 2 mM GIcNAc~ 1-2Manod -6Glc[34Glc13-all 3 2.5 mM GlcNAc[31-2Man~l-6Manl3-mco 4 1 mM GlcNAcl31-2Man~l-6Man~-Me 8 Inhibition studies Specificity of GlcNAc-T V from hen oviduct Activity (pmol/h/mg) 485 120 664 359 313 176 69 2.3 m s GlcNAc~l-2Mana-Me 2.3 mM GlcNAc~l-4(GlcNAc[~l-2)Manc~-Me 1 mM GlcNAc[~l-6Man~-Me 60 0 0 Conditions B: 2 2 mM GlcNAc131-2Manc~1-6Glct3-pnp 5 2 mM GlcNAcl31-2Manc~l-6Glc[3-all 6 2 ~ GlcNAcl31-2Manc~l-6Glc[~l-4Gtcl3-all 7 2 mM GlcNAct31-2Man~t-6GIcI31-4GlcNAc 250 670 385 460 20 21 Enzyme assays were carried out as described in Methods by HPLC, Two different experiments (conditions A and B) were carried out using different enzyme preparations. Conditions A: 1 h incubation, 1 mM UDP-[~4C]GlcNAc,0.5 mg protein per assay. Conditions B: 2 h incubation, 0.84 ms UDP-[~aC]GlcNAc,0.12 mg protein per assay. Inhibition studies were carried out with hamster kidney GlcNAc-T V using 0.3 mM GlcNAc~l-2MancO-6Glc[3-oct as substrate (compound 1, Table 4), The activity was significantly reduced in the presence o f inactive substrate analogues with 4- or 6-O-methyl substitution o f the MancO-6 residue (compounds 11 and 12, Table 4). The best inhibitor was GlcNAcl31-2(4,6-di-O-Me)Manal-6 Glc~-pnp (compound 13) which inhibited GlcNAc-T V by 53%. GlcNAc-T V from hen oviduct showed a 75% inhibition by compound 13 (Table 3). Irradiation of the enzyme at 350 nm in the presence of the mono-methylated pnp derivatives resulted in slightly increased inhibition (Table 4). Several leukaemic leukocyte samples were used as sources of GlcNAc-T V to study the inhibition with compound 13. The activity was inhibited 3 7 - 6 7 % with A M L cell extracts and 2 9 - 8 8 % with C M L granulocyte extracts using a 2.5-fold molar excess of compound 13 over substrate. GlcNAc-T V in leukaemic leukocytes and other tissues GlcNAc-T V activity was also found in homogenates from rat colon, adult and fetal human colon, pig stomach mucosa, leukocytes from patients with A M L and C M L and normal human granulocytes (data not shown). The average GlcNAcT V activities (measured with 1 mM GlcNAcl31-2MancO6Glcl3-oct as substrate) of extracts from A M L (eight samples), C M L (four samples) and normal (three samples) leukocytes were 860, 670 and 430 pmol h -1 per mg respectively. Brockhausen et al. 376 Discussion III 3000- A I 2000 O001] "5 E 3000 - B c', "0 I 2000- 1000 - 01/ I 0 20 III I I 40 l" I IV I 60 I I 80 Elution Time (minutes) Figure 1. HPLC elution pattern of GlcNAc-T V product using GlcNAc~l-2Mancd-6Glc[~-allyl as the substrate and hen oviduct microsomes. HPLC was carried out on a PAC column, using acetonitrile:water 82:18 at 0.7 ml rain-I. Elution patterns were established with standard compounds GlcNAc[31-6(GlcNAc[31-2)Man-R and GlcNAc~t-4(GlcNAc~I-2)Man-R [39]; GlcNAc~I-6(GlcNAc~I2)Man-R (GlcNAc-T V product) always elutes after GlcNAc~I4(GlcNAc[31-2)Man-R (GlcNAc-T VI' product). Peak I, [14C]GIcNAc; peak II, unknown; peak III, GlcNAc-T VI' product, [14C]GlcNAc[~l-4(GlcNAc[~l-2)Man-R; peak IV, GlcNAc-T V product, [14C]GlcNAc[~l-6(GlcNAcl~l-2)Man-R. (A) In the presence of Mn z+ radioactivity due to GlcNAc-T VI' product was very high and no GlcNAc-T V product was detected. (B) In the absence of Mn 2+ in the assay, GlcNAc-T V product eluted at 75 min and GlcNAc-T VI' product eluted earlier at 65 rain. The latter product was not seen when mammalian cells were used as the source of GlcNAc-T V. GlcNAc-transferase V adds the GlcNAc~I-6 branch to the Maned-6 arm in the biosynthesis of complex N-glycans [23, 34, 46]. It has been suggested that this branch carries most of the long chain poly-N-acetyllactosaminoglycans which in turn may carry antigenic determinants and sialic acid residues [1 S, 19, 21]. Inhibition of GlcNAc-T V would therefore prevent the synthesis of highly branched N-glycan structures and possibly reduce overall sialylation of N-glycans. This inhibition may be beneficial in certain diseases such as cancer or metastasis where increased occurrence of complex structures has been reported [8, 15, 17, 47, 48]. Various cell types exhibit great differences in the expression of GlcNAc-T V activities consistent with regulation of GlcNAc-T V in a tissue-specific fashion and during development and differentiation [4, 11, 20-22]. Typical mucin secreting tissues such as colonic mucosa and pig stomach have comparatively high GlcNAc-T V activity. Our results suggest that GlcNAc-T V is increased in AML and CML leukocytes. This may reflect an altered stage of differentiation of these cells compared to normal granulocytes. We previously reported increased activities in AML and CML cells of another branching GtcNAc-T, core 2 ~6-GlcNAc-T [44], which in combination with other changes in the O-glycosytation pathways may be partly responsible for the increased cell surface sialytation in leukaemic cells [49-51]. High GlcNAc-T V activity would be expected to increase the total proportion of tetraantennary chains and thereby contribute to the overall sialylation of leukaemic cells even though sialyltransferase activities acting on N-glycans remain unchanged as previously reported [51]. Inhibition of GlcNAc-T V activity by inhibitors with the ability to penetrate membranes and act on Golgi-localized GlcNAc-T V may therefore be beneficial in reducing the abnormal cell surface sialylation of leukaemic cells. Knowledge of the substrate recognition mechanism used by GlcNAc-T V is an essential prerequisite in the design of effective inhibitors. We have shown in this study that the size of the substrate influences GlcNAc-T V activity. Three sugars are required for optimal activity [52]. GlcNAc-T VI' [39] and GIcNAc-T I [53, 54] also require a substrate with a minimum of three sugar residues whereas GIcNAc-T II requires a tetrasaccharide [55-57]. The influence of the peptide sequence of glycoprotein substrates on GlcNAc-T V activity has not yet been investigated. Although there is no absolute requirement for carbohydrate residues other than the GlcNAc[31-2Manc~l-6Man[~ structure, the biantennary compounds show higher activities for the hamster kidney enzyme than do the linear compounds. The biantennary compound GlcNAc[31-2Man(xl6(GlcNAcl31-2[4-deoxy-]Mancd-3)4-O-methyl-Manl3-octyl (compound 19, Table 2) is of special interest because it shows the highest activity of all the acceptors studied (Table 2) and it is a specific substrate for GlcNAc-T V from mammalian sources (which lack GlcNAc-T VI and VI') since GlcNAc-T I, II, III and IV cannot act on it. 377 Substrate specificity a n d inhibition o f G l c N a c - T V Table 4. Inhibition of GlcNAc-T V purified from hamster kidney Compound (0.8 mM) Activity a (pmol mg -1 per h) In the absence of 0.3 mM GIcNAc~I-2 Mano~l-6 Glc~-oct: 11 GlcNAc[31-2(4-O-Me)Man~ 1-6Glc[3-pnp 12 GlcNAc[31-2(6-O-Me)Man~x 1-6Glc~-pnp 13 GlcNAc[~ l-2(4,6-di-O-Me)Mano~l-6Glcl3-pnp 22 GlcNAc[~l-2(6-O-Me)Manc~-Me 23 GlcNAc~ 1-2(4,6-di-O-Me)Mano~-Me ND ND ND ND ND In the presence of 0.3 mM GIcNAc~ 1-2 Mano~l-6 Glc[~-oct: 11 GlcNAc~] 1-2[4-O-Me]Mano~l-6Glc~-pnp + UV UV - 12 GlcNAc[~1-2[6-O-Me]Manor t-6Glc[~-pnp + UV UV - 13 22 GlcNAc[31-214,6-di- O-Me]Mano~l-6Glcl3-pnp GIcNAcl31-2(6-O-Me)Manor-Me % Inhibition ~ 33 48 41 28 41 28 53 0 a Assays were carried out without HPLC, using purified hamster kidney GlcNAc-T V, as described in Methods with 0.3 mMGlcNAc[31-2Man(xl-6Glc[3-octas substrate, 1 mM UDP-[14C]GlcNAc,with and without the addition of 0.8 mM of the compounds under test. ND, activity was not detectable. + UV, compounds were preincubated with the enzyme and irradiated at 350 nm. UV, compoundswere preincubated with the enzyme without irradiation. bEnzyme activity in the absence of inhibitor is 3270 pmol hq per mg; % inhibition = 100 x (difference in activity of the enzyme in the absence and presence of inhibitor) + (activity in the absence of inhibitor). - Figure 2 displays the structure of a biantennary N-glycan substrate for GlcNAc-T V and the hydroxyls that were found to be important or essential for activity from this study and previous reports [ 12, 23, 24, 26, 31, 34-36, 40, 42]. Although the Mano~l-3 arm is not required, the 3-hydroxyl, but not the 4-hydroxyl, of the Manod-3 residue appears to be important O R Figure 2. GlcNAc-transferase V specificity towards GlcNAcI312Manc~l-6(GlcNAct31-2-Man~ 1-3)Manl3-R substrate. The hydroxyls found to be essential for activity are surrounded by squares. Hydroxyls that have an influence but are not essential for activity are surrounded by circles. The data are from this study and from previous reports [12, 23, 24, 26, 31, 34-36, 40, 42]. since the 3-deoxy derivative shows reduced activity and a higher Kin. The 4-hydroxyl of the internal Man~3 residue has been shown to be essential for GlcNAc-T I activity [53, 54, 58]. It is not essential, however, for GlcNAc-T II and V which act on the Mano~l-6 residue, although its removal appears to affect substrate binding to GlcNAc-T II [57] and V (Table 2). Methyl substitution of the 4-hydroxyl of Man~31-4 increases GlcNAc-T V activity. Similar results were reported using the linear trisaccharide [52]. Methyl substitution of the 4-hydroxyl of Man[31-4 inhibits binding to G l c N A c - T II [57] whereas substitution by the bisecting residue ( G l c N A c ~ t - 4 linked to Manl31-4 ) turns off all of the GlcNAc-T acting on N-glycans with the exception of GlcNAc-T VI [26, 34, 54]. The M a n , l - 4 residue may be replaced by Glc, indicating that the configuration of the 2-hydroxyl is not important for activity. However, when Glc was replaced by a hexane ring [31], the activity was significantly reduced. Using derivatives of GlcNAcl31-2Manc~l-6Glc[3-R, the crude hamster kidney enzyme was found to be more active when the 4-hydroxyl of Glc was substituted with a methyl group, but the activity was unchanged when this hydroxyl was removed [31]. The 4- and 6-hydroxyls of the GlcNAc residue attached to the Mano~l-6 residue are essential for substrate recognition by G l c N A c - T V [35, 59]. Galactosylation, removal or modification o f the 4-hydroxyl of the terminal G l c N A c residue results in loss of activity [35]. This is consistent with the finding that GlcNAc-T V from mouse lymphoma cells does 378 not act on galactosylated or sialylated biantennary substrates [23] and that the enzyme from human lung cancer cells is inactive towards biantennary substrates with Gal-substituted GlcNAc residues on both the Manc~l-3 and the Manc~l-6 arms [26]. Interestingly, GlcNAc-T I, II, III and IV are also known to be inhibited by galactosylation of GlcNAc [35, 53, 55, 56, 60, 61]. Results using 3-O-pivaloyl-GlcNAct31-2Manc~l6Glc[3-pnp (compound 9, Table 2) suggest that the 3-hydroxyl of this GlcNAc residue may also be important [59], or that the pivaloyl group causes steric hindrance or unfavourable electronic interactions which prevent binding. Both the 4- and the 6-hydroxyls of M a n a l - 6 are minimally involved in binding to the enzyme substrate binding site. The substrate analogue which is methylated at the 4-hydroxyl of the M a n , l - 6 residue is a competitive inhibitor and inhibition is probably due to steric hindrance of catalysis by the methyl group since the corresponding 4-deoxy compound is a substrate [36, 62]. The enzyme is competitively inhibited by the substrate analogue lacking the 6-hydroxyl of the M a n , l-6 residue, the site of enzyme action [24, 41]. Methylation of the 6-hydroxyl of the M a n ~ l - 6 residue of GlcNAc[31-2 ManoO-6 Glcl3-pnp produced a compound that was not active as a substrate although it inhibited enzyme activity (Table 4). Thus methylation of the Manc~t-6 residue does not interfere with binding to the enzyme substrate binding site. This finding is important for the design of GlcNAc-T V inhibitors. We therefore used the di-methylated derivative GlcNAc~l-2(4,6-di-O-Me)Mancxl-6 Glcl3-pnp as an inhibitor for GlcNAc-T V from hamster kidney, hen oviduct and leukaemia cells and found it to be effective in vitro. It remains to be shown if compounds with similar structures can be designed to penetrate biological membranes and thereby reach the intracellular site of glycosyltransferase action. Acknowledgements This research was supported by grants from the Medical Research Council of Canada to I. Brockhausen, from the Deutsche Forschungsgemeinschaft to H. Paulsen and NIAID No. AI 29326 to K. L. Matta. F. Reck was supported by the Deutsche Studienstiftung and the BASF Aktiengesellschaft. References 1. 2. 3. 4. Varki A (1993) Glycobiology 3:97-30. Varki A (1992) Curr Opin Cell Biol 4:257-66. Feizi T (1993) Curt Opin Struct Biol 3:701-10. Dennis JW (1992) In Cell Surface Carbohydrates and Celt Development, (Fukuda M, ed.) pp 161-94, Boca Raton, Florida: CRC Press, Inc. 5. Dennis JW, Laferte S (1988) In Altered Glycosylation in Tumor Cells (Reading CL, Hakomori S-I, Marcus DM, eds) pp. 257-67, New York, N.Y.: Alan R. Liss Inc. 6. Smets LA, Van Beek WP (1984) Biochim Biophys Acta 738:237-49. B r o c k h a u s e n et al. 7. Santer UV, DeSantis R, Hard KJ, Van-Kuik JA, Vliegenthart JFG, Won B, Glick MC (1989) Eur J Biochem 181:249-60. 8. Dennis JW, [ aferte S, Waghorne C, Breitman ML, Kerbel RS (1987) Science 236:582-85. 9. Pierce M, Arango J (1986) Y Biol Chem 261:10772-77. 10. Arango J, Pierce M (1988) J Cell Biochem 37:225-31. 11. Yarnashita K, Tachibana Y, Ohkura T, Kobata A (1985) J Biol Chem 260:3963-69. 12. Palcic MM, Ripka J, Kaur KJ, Shoreibah M, Hindsgaul O, Pierce M (1990) J BioI Chem 265:6759-69. 13. Yousefi S, Higgins E, Daoling Z, Pollex-KNger A, Hindsgaul O, Dennis JW (1991) J BioI Chem 266:1772-82. 14. Easton EW, Bolscher JGM, Van den Eijnden DH (1991) J Biol Chem 266:21674-80. 15. Dennis JW, Laferte S (1989) Cancer Res 49:945-50. 16. Easton EW, Blokland I, Geldof AA, Rao BR, Van den Eijnden DH (1992) Febs Lett 308:46-49. 17. Dennis JW, Kosh K, Bryce D-M, Breitman ML (1989) Oncogene 4: 853-60. 18. Cummings RD, Kornfeld S (1984) J Biol Chem 259:6253-60. 19. van den Eijnden DH, Koenderman AHL, Schiphorst WECM (1988) J Biol Chem 263:12461-71. 20. Brockhausen I, Romero PA, Herscovics A (1991) Cancer Res 51:3136-42. 21. Heffernan M, Lotan R, Amos B, Palcic M, Takano R, Dennis JW (1993) J Biol Chem 268:1242-51. 22. Nakao H, Nishikawa A, Karasuno T, Nishiura T, Iida M, Kanayama Y, Yonezawa T, Tarui S, Taniguchi N (1990) Biochem Biophys Res Commun 172:1260-66. 23. Cummings RD, Trowbridge IS, Kornfeld S (1982) J Biol Chem 257:13421-27. 24. Hindsgaul O, Kaur KJ, Srivastava G, Blaszczyk-Thurin M, Crawley SC, Heerze LD, Palcic MM (1991) J Biol Chem 266:17858-62. 25. Shoreibah MG, Hindsgaul O, Pierce M (1992) J Biol Chem 267:2920-27. 26. Gu J, Nishikawa A, Tsuruoka N, Ohno M, Yamaguchi N, Kanagawa K, Taniguchi N (1993) J Biochem 113:614-19. 27. Shoreibah M, Perng GS, Adler B, Weinstein J, Basu R, Cupples R, Wen D, Browne JK, Buckhaults P, Fregien N, Pierce M (1993) J Biol Chem 268:15381-85. 28. Saito H, Nishikawa A, Gu JG, Ihara Y, Soejima H, Wada Y, Sekiya C, Niikawa N, Taniguchi N (1994) Biochem Biophys Res Commun 198:318-27. 29. Saito H, Nishikawa A, Gu JG, Ihara Y, Soejima H, Wada Y, Sekiya C, Niikawa N, Taniguchi N (1994) Biochem Biophys Res Commun 200:668-69. 30. Hindsgaul O, Tahir SH, Srivastava OP, Pierce M (1988) Carbohydrate Res 173:263-72. 31. Srivastava OP, Hindsgaul O, Shoreibah M, Pierce M (1988) Carbohydrate Res 179:137-61. 32. Crawley SC, Hindsgaul O, Alton G, Pierce M, Palcic MM (1990) Anal Biochem 185:112-17. 33. Brockhausen I, Grey AA, Pang H, Schachter H, Carver JP (1988) Glycoconjugate J 5:419-48. 34. Brockhausen I, Carver J, Schachter H (1988) Biochem Cell Biol 66:1134-51. 35. Kanie O, Crawtey SC, Palcic MM, Hindsgaul O (1993) Carbohydr Res 243:139-64. Substrate specifici~ and inhibition of GlcNac-T V 36. Khan SH, Crawley SC, Kanie O, Hindsgaul O (1993) J BioI Chem 268:2468-73. 37. Brockhausen I, Matta KL, Orr J, Schachter H (I985) Biochemist~. 24:186674. 38. Khan SH, Abbas SA, Matta KL (1989) Carbohydr Res 193: t 25-39. 39. Brockhausen I, M6ller G, Yang JM, Khan SH, Matta KL, Paulsen H, Grey AA, Shah RN, Schachter H (1992) Carbohydr Res 236:281-99. 40. Khan SH, Matta KL (t 993) J Carbohydr Chem 12:335-48. 41. Khan SH, Abbas SA, Matta KL (1990) Carbohydr Res 205:385-97. 42. Khan SH, Matta KL (1993) Carbo,Sydr Res 243:29--42. 43. Paulsen H, Meinjohanns E, Reck F, Brockhausen I (1993) Liebigs Ann Chem 737-750. 44. Brockhausen I, Kuhns W, Schachter H, Matta KL, Sutherland DR, Baker MA (1991) Cancer Res 51:1257-63. 45. Toki D, Granovsky MA, Reck F, Kuhns W, Baker MA, Matta KL, Brockhausen I (1994) Biochem Biophys Res Commun 198:417-23. 46. Brockbausen I, Hull E. Hindsgaul O, Schachter H, Shah RN, Michnick SW, Carver JP (1989) JBiol Chem. 264:11211-21. 47. Kobata A (1988) Biochimie 70:1575-85. 48. Yamashita K, Hitoi A, Taniguchi N, Yokosawa N, Tsukada Y, Kobata A (1983) Cancer Res 43:5059-63. 379 49. Baker MA, Taub RN, Kanani A, Brockhausen I, Hindenburg A (1985) Blood 66:1068-71. 50. Baker MA, Kanani A, Brockhausen I, Schachter H, Hindenburg A, Taub RN (1987) Cancer Res 47:2763-66. 51. Kanani A, Sutherland DR, Fibach E, Matta KL, Hindenburg A, Brockhausen I, Kuhns W, Taub RN, Van den Eijnden DH, Baker MA (1990) Cancer Res 50:5003-7. 52. Linker T, Crawley SC, Hindsgaul O (1993) Carbohydr Res 245:323-31. 53. Vella GJ, Paulsen H~ Schachter H (1984) Can J Biochem Cell Bio162: 409-17. 54. MNler G, Reck F, Panlsen H, Kaur KJ, Sarkar M, Schachter H, Brockhausen I (I992) Glycoconjugate J 9:180-90. 55. Bendiak B, Schachter H (1987) JBiol Chem 262:5775-83. 56. Bendiak B, Schachter H (1987) J Biol Chem 262:5784-90. 57. Reck F, Meinjohanns E, Springer M, Wilkens R, Van Dorst JALM, Paulsen H, M011er G, Brockhausen I, Schachter H (1994) Glycoconjugate J 11:210-16. 58. Nishikawa Y, Pegg W, Paulsen H, Schachter H (1988) J Biol Chem 263:8270-81. 59. Kanie O, Crawley SC, Palcic M, Hindsgaul O (1994) Bioorg. Medicinal Chem 2:1231-41. 60. Gleeson PA, Schachter H (1983) JBiol Chem 258:6162-73. 61. Narasimhan S (1982) JBiol Chem 257:10235-42. 62. Khan SH, Duus JO, Crawley SC, Palcic MM, Hindsgaul O (1994) Tetrahedron Asymmeto, 5:2415-35.