A Glycosyltransferase Inhibitor from a Molecular Fragment Library Simultaneously Interferes with Metal Ion and Substrate Binding

Fragment (logic)
DOI: 10.1002/anie.201108345 Publication Date: 2012-03-08T14:14:53Z
ABSTRACT
A fragmented approach: 3-Phenyl-5-piperazino-1,2,4-thiadiazole (designated “compound 382” in the Ro5 Maybridge Fragment Library, see scheme) is demonstrated to be an effective inhibitor of human blood group glycosyltransferase B. The compound interferes with both acceptor and donor binding also displaces Mn2+ ion pocket. Glycosyltransferases (GTs) are ubiquitous enzymes that form glycosidic bonds by transfer monosaccharides from donors, typically nucleotide sugar carbohydrate, protein, lipid, nucleic acid, or natural product acceptors.1, 2 As each linkage requires a different enzyme, GT class enormous encoded approximately 1 % genes have been sequenced date. GTs implicated normal abnormal biological processes, such as cellular adhesion, cell signaling, carcinogenesis, wall biosynthesis pathogens.3–6 linked pathological conditions, therefore, potent specific inhibitors long sought potential therapeutics. However, synthetic chemistry for production often difficult.7, 8 Inhibitor design complicated complex reaction mechanisms GTs, which undergo conformational changes upon substrate throughout catalytic cycle.1, Unlike other enzyme classes, potency most analogues substrates relatively limited, majority Ki values comparable greater than KM corresponding substrate.8 site glycosyl shared multiple it challenging based on donor-site ligands.7–9 very well defined, fairly low (μM) dissociation constant, makes further improvements difficult. specificity likely result recognition substrate, pocket shallow easily accessible. affinity less is, possible can obtained linking derivative acceptor-site binder reported bisubstrate-type sialyltransferases,10 β-1,4- α-1,3-galactosyltransferases.11, 12 Recently, 3-phenyl-5-piperazino-1,2,4-thiadiazole (1), heterocycle contains benzene piperazine substituents (Scheme 1) designated 382”, Library purchased 2006 predecessor Ro3 library, was galactosyltransferase B (GTB).13 Saturation difference (STD)-NMR spectroscopy experiments showed competes H-disaccharide (KM=88 μM) IC50 value 790 μM against acceptor, determined surface plasmon resonance (SPR).13. (1). Furthermore, docking suggest occupies part site.13 GTB uses uridine diphosphate galactose (UDP-Gal) transfers Gal α-L-Fucp-(1→2)-β-D-Galp (H-antigen) acceptors, produces structures. closely related N-acetylgalactosaminyltransferase (GTA) UDP-GalNAc GalNAc H-antigen GTA highly homologous differ only four amino acid residues (Arg/Gly176, Gly/Ser235, Leu/Met266, Gly/Ala268). To establish basis selectivity we carried out studies using GTB, GTA, GTA/GTB chimeric three GTs. We crystal structure bound enzyme. weakly inhibited 1. elucidate mode this compound, tested displacement SPR STD-NMR spectroscopy. data show GTB. unprecedented inhibition. Compound 1, previously identified library screen competitive site, evaluated its panel enzymes. Standard radiochemical assay formats were adapted competition co-incubating various concentrations 800 Dixon plots (Table 1). inhibits bovine α-1,3-GalT fivefold more 6.3 mM, but did not inhibit α-1,4-GalT Neisseria meningitidis β-1,4-GalT has 5 mM weak AAGlyB (Arg176, Gly235, Gly266, Ala268), 15 mM. murine genome14 dual utilizing UDP-Gal donors equal efficiency produce For all effect leveled off at above 3–4 caused problems evaluations AAGlyB. This attributed solubility Similar findings reduced inhibition higher reported.15 0.7 μM16 compared 27 GTB.17 When UDP-Gal, 4.5 obtained, indicates binding. Enzyme [mM][a] Acceptor [mM] Donor 0.8 0.088[b] 0.027[b] >15.0 0.021[c] 0.0007[c] 5.1 0.010[b] 0.009[b] α-1-3-GalT 0.19 0.077[c] n.i. >0.25 0.0005[c] 0.13 0.022[c] better understand differential potencies AAGlyB, Both structures solved molecular replacement refined X-ray diffraction sets collected 1.6 Å resolution AAGlyB–1 1.85 GTB–1 S1 Supporting Information). overall conformation isomorphous described (Figure A).17, 19 internal loop (residues 176–199) “open” 175–181 chain 175–184 being disordered. Crystal 382=1. A) Structure GTB–1. Hydrogen indicated green dotted lines. B) AAGlyB–1. ion=purple. open disordered seen apo-GTB (PDB entries 2RIT 1LZ7). significant between published C terminus 346–353) ordered adopts position points away UDP leaves completely (see Section previous AAGlyB.16 “semi-closed” AAGlyB–UDP structure.16 significantly backbone atoms residue side chains defined. kink α helix (α3) around Arg198, earlier “closed” structures,17 shifted slightly towards Cys196. usual conformation, last nine 346–354) excellent electron density entire active A). There strong peak corresponds sulfur five-membered ring inhibitor, confirms origin orientation compound. N4 atom forms hydrogen Asp211 Asp213 DXD motif, placed almost exactly where metal normally positioned. remainder predominantly hydrophobic, yet negatively charged pocket, consists Met214, Pro234, Ser235, Met266, Ala268, Leu324, Asp 326, Leu329, Leu330, Ala343 3). Electron maps site. GTB; Fo−Fc omit map Blue contoured 4 σ whereas red 20 emphasize atom. AAGlyB; UDP. 3.5 σ. In there ion. suggests inability displace may explain weaker mutant Residues surrounding Thus, directly overlaps fucose moiety AABB–UDP-Gal–acceptor Figure S3 Information comparison), prevents unambiguous same place B). AAGlyB–1, outer nitrogen equally tip ring, able fully considerable, occupied, molecule used during final elution step purification trace amounts remain even after extensive dialysis. With space ion, clashes severely β-phosphate Therefore, unlikely stay together will either displaced or, least, relocate accommodate no indication alternative map. By refining occupancy ligands Phenix refinement program, 56 63 B, 37 %, poor density, α-phosphate ribose visible. Correspondingly, 2Fo−Fc map, clearly visible, (σ=6.5, B-factor 20.9 Å2) relative (σ=7.5, 19.5 Å2). limited presence supports occupancy. Biacore probe immobilized CM5 chips constants analysis curves increasing results support 4; Table S2 Mg2+ ions present, 0.2 0.6 GTA. when comparing one-site model two-site model, yields fit experimental data. adheres portions protein surface, albeit much lower Figures S4–S7 mutants interpreted unspecific (low affinity) overlays inhibitory dramatically present. UDP–Mn2+ complexes bind tightly UDP–Mg2+ complexes, Dissociation (white), (light gray), (dark gray) derived experiments. Measurements performed MOPS (50 pH 6.7), NaCl (100 mM), Mg2+/Mn2+ (5 (1 (0.25 μM–7 mM). See details. derive epitopes atomic resolution, spectra measured absence S8 receives saturation phenyl ring. UDP, intensity STD signals reduced. kinetic influenced nature concentration. high hardly any detectable mutant. enzymological profile rather unexpected because chimera. Binding found interfere disaccharide coordinated motif pyrophosphates nucleotide. structural thereby making two-substrate inhibitor. could make candidate monophosphate (UMP) develop differs predicted Autodock program.13 predicts whereupon stacking interaction Trp300, bond Asp302. Interestingly, two shape dependent 235 266. Ser235 Met266 play key role forming fitting A; S9A 266 glycine although does seem change significantly, generates opening bottom B; S9B supported interface calculations respectively S9 Information), buried area It leucine instead methionine, change; however, Leu266 maintains closed S9C Structures showing overlapped experiments, KD increases binds structurally similar distantly α-1,4- β-1,4-GalT. reason difficult perspective. superimposing analogue entry 1GX4) onto several sterically S10 appear resembles α-1,3-GalT. Taken together, these ligand ions, sensitive environment especially 266, seems preferentially Residue known important discrimination transfer, properties good development inhibitors, proving useful trying mechanism All cloned expressed described20–23 apart (Sigma–Aldrich). UDP-GalNAc, α-L-Fucp-(1→2)-β-D-Galp-O-(CH2)7-CH3 blood-group previously16, 18 standard assay.17, linear regression (KM concentration, average duplicate lowest calculated R2 fitted lines experiment (except GTA) 0.97. scattering among measurements, resulted fits values. up five times. Crystals vapor diffusion reservoir solution PEG-3350 (23 %), ammonium sulfate (300 3-(N-morpholino)propanesulfonic (MOPS, 50 7), MnCl2 mM) (6–9 (50–200 After drop (6 μL) cryo-solution (reservoir MnCl2, glycerol (30 Maybridge) added crystals, crystals flash frozen liquid N2. European Synchrotron Radiation Facility (ESRF), Grenoble, France, Berliner Elektronenspeicherring-Gesellschaft für Synchrotronstrahlung (BESSY II). Berlin, Germany. integrated scaled XDS package.24 CCP4i module Phaser25 2RIT) search model. Initial REFMAC526 before iteratively rebuilding Coot,27 finally Phenix28 including translation/libration/screw motion (TLS) determination. quality statistics given Information. coordinates factors deposited Protein Data Bank accession codes, 3U0X AAGlyB-1 3U0Y GTB-1. BIAcore 3000 instrument (GE Healthcare) 298 K. sensor (Biacore) Amine Coupling Kit (Biacore). Different mm) injected flow rate 10 μL min−1 3 min reach equilibrium, then waiting time injections. Two triplicate performed, analyzed evaluation software Origin 7.0. equation under assumption measurements mm NMR tubes. Bruker Avance DRX 500 spectrometer equipped TCI cryogenic head. Spectra recorded K pulse sequence stddiff.3 ms, 7 kHz spin-lock pulse29 reduce background resonances. Additional details Dedicated Professor Hans Paulsen occasion his 90th birthday Detailed facts importance specialist readers ”Supporting Information”. Such documents peer-reviewed, copy-edited typeset. They made available submitted authors. Please note: publisher responsible content functionality supporting information supplied Any queries (other missing content) should directed author article.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (32)
CITATIONS (19)