Interactions of a Family 18 Chitinase with the Designed Inhibitor HM508 and Its Degradation Product, Chitobiono-δ-lactone
2004; Elsevier BV; Volume: 279; Issue: 5 Linguagem: Inglês
10.1074/jbc.m310057200
ISSN1083-351X
AutoresGustav Vaaje‐Kolstad, Andrea Vasella, Martin G. Peter, Catharina Netter, Douglas R. Houston, Bjørge Westereng, Bjørnar Synstad, Vincent G. H. Eijsink, Daan M. F. van Aalten,
Tópico(s)Enzyme Production and Characterization
ResumoWe describe enzymological and structural analyses of the interaction between the family 18 chitinase ChiB from Serratia marcescens and the designed inhibitor N,N′-diacetylchitobionoxime-N-phenylcarbamate (HM508). HM508 acts as a competitive inhibitor of this enzyme with a Ki in the 50 μm range. Active site mutants of ChiB show Ki values ranging from 1 to 200 μm, providing insight into some of the interactions that determine inhibitor affinity. Interestingly, the wild type enzyme slowly degrades HM508, but the inhibitor is essentially stable in the presence of the moderately active D142N mutant of ChiB. The crystal structure of the D142N-HM508 complex revealed that the two sugar moieties bind to the -2 and -1 subsites, whereas the phenyl group interacts with aromatic side chains that line the +1 and +2 subsites. Enzymatic degradation of HM508, as well as a Trp → Ala mutation in the +2 subsite of ChiB, led to reduced affinity for the inhibitor, showing that interactions between the phenyl group and the enzyme contribute to binding. Interestingly, a complex of enzymatically degraded HM508 with the wild type enzyme showed a chitobiono-δ-lactone bound in the -2 and -1 subsites, despite the fact that the equilibrium between the lactone and the hydroxy acid forms in solution lies far toward the latter. This shows that the active site preferentially binds the 4E conformation of the -1 sugar, which resembles the proposed transition state of the reaction. We describe enzymological and structural analyses of the interaction between the family 18 chitinase ChiB from Serratia marcescens and the designed inhibitor N,N′-diacetylchitobionoxime-N-phenylcarbamate (HM508). HM508 acts as a competitive inhibitor of this enzyme with a Ki in the 50 μm range. Active site mutants of ChiB show Ki values ranging from 1 to 200 μm, providing insight into some of the interactions that determine inhibitor affinity. Interestingly, the wild type enzyme slowly degrades HM508, but the inhibitor is essentially stable in the presence of the moderately active D142N mutant of ChiB. The crystal structure of the D142N-HM508 complex revealed that the two sugar moieties bind to the -2 and -1 subsites, whereas the phenyl group interacts with aromatic side chains that line the +1 and +2 subsites. Enzymatic degradation of HM508, as well as a Trp → Ala mutation in the +2 subsite of ChiB, led to reduced affinity for the inhibitor, showing that interactions between the phenyl group and the enzyme contribute to binding. Interestingly, a complex of enzymatically degraded HM508 with the wild type enzyme showed a chitobiono-δ-lactone bound in the -2 and -1 subsites, despite the fact that the equilibrium between the lactone and the hydroxy acid forms in solution lies far toward the latter. This shows that the active site preferentially binds the 4E conformation of the -1 sugar, which resembles the proposed transition state of the reaction. Chitin, a structural component of invertebrate exoskeletons and fungal cell walls, is an abundant, rigid, linear carbohydrate polymer consisting of β(1, 4)-linked N-acetyl-glucosamine (GlcNAc) units. In nature, chitin is degraded by chitinases and β-N-acetylhexosaminidases belonging to families 18 and 19 and families 3 and 20 of glycoside hydrolase, respectively (1Henrissat B. Biochem. J. 1991; 280: 309-316Google Scholar, 2Henrissat B. Bairoch A. Biochem. J. 1993; 293: 781-788Google Scholar, 3Henrissat B. Bairoch A. Biochem. J. 1996; 316: 695-696Google Scholar). Chitinases occur in a variety of organisms from bacteria and fungi to plants and vertebrates. It has been shown that inhibitors of family 18 chitinases affect the life cycles of insects (4Arai N. Shiomi K. Yamaguchi Y. Masuma R. Iwai Y. Turberg A. Koelbl H. Omura S. Chem. Pharm. Bull. 2000; 48: 1442-1446Google Scholar, 5Cohen E. Arch. Insect Biochem. Physiol. 1993; 22: 245-261Google Scholar, 6Sakuda S. Isogai A. Matsumoto S. Suzuki A. J. Antibiot. (Tokyo). 1987; 40: 296-300Google Scholar) and human pathogens such as Candida albicans (7Izumida H. Nishijima M. Takadera T. Nomoto A.M. Sano H. J. Antibiot (Tokyo). 1996; 49: 829-831Google Scholar) and the human malaria parasite Plasmodium falciparum (8Vinetz J.M. Valenzuela J.G. Specht C.A. Aravind L. Langer R.C. Ribeiro J.M.C. Kaslow D.C. J. Biol. Chem. 2000; 275: 10331-10341Google Scholar, 9Vinetz J.M. Dave S.K. Specht C.A. Brameld K.A. Xu B. Hayward R. Fidock D.A. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 14061-14066Google Scholar, 10Tsai Y.-L. Hayward R.E. Langer R.C. Fidock D.A. Vinetz J.M. Infect. Immun. 2001; 69: 4048-4054Google Scholar). Thus, family 18 chitinases have been proposed as targets for the development of drugs and insecticides. Structural and enzymological studies of several family 18 chitinases have provided detailed insight into the catalytic center and mechanism of these enzymes (11Terwisscha van Scheltinga A.C. Armand S. Kalk K.H. Isogai A. Henrissat B. Dijkstra B.W. Biochemistry. 1995; 34: 15619-15623Google Scholar, 12Tews I. Terwisscha van Scheltinga A.C. Perrakis A. Wilson K.S. Dijkstra B.W. J. Am. Chem. Soc. 1997; 119: 7954-7959Google Scholar, 13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 14Bokma E. Rozeboom H.J. Sibbald M. Dijkstra B.W. Beintema J.J. Eur. J. Biochem. 2002; 269: 893-901Google Scholar, 15Bortone K. Monzingo A.F. Ernst S. Robertus J.D. J. Mol. Biol. 2002; 320: 293-302Google Scholar). Leaving group departure is promoted by a glutamate residue that acts as catalytic acid (Glu-144 in chitinase B (ChiB) 1The abbreviations used are: ChiBchitinase BWTwild type ChiBHM508N,N′-diacetyl-chitobionoxime-N-phenylcarbamateESI-MSelectrospray ionization mass spectroscopyBSAbovine serum albumin. from Serratia marcescens, the enzyme used in this study). The emerging positive charge on the anomeric carbon is stabilized by concomitant nucleophilic attack of the N-acetyl group of the -1 sugar on the anomeric carbon, which leads to formation of an oxazolinium ion intermediate (Fig. 1). Glu-144 is located near the end of β-strand 4 of the catalytic (βα)8 barrel and is preceded by other conserved acidic residues that are located in the core and are essential for catalysis (Asp-140 and Asp-142 in ChiB) (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 14Bokma E. Rozeboom H.J. Sibbald M. Dijkstra B.W. Beintema J.J. Eur. J. Biochem. 2002; 269: 893-901Google Scholar, 15Bortone K. Monzingo A.F. Ernst S. Robertus J.D. J. Mol. Biol. 2002; 320: 293-302Google Scholar, 16Watanabe T. Kobori K. Miyashita K. Fujii T. Sakai H. Uchida M. Tanaka H. J. Biol. Chem. 1993; 268: 18567-18572Google Scholar, 17Synstad B. Gaseidnes S. Vriend G. Nielsen J.E. Eijsink V.G.H. Peter M.G. Muzzarelli R.A.A. Domard A. Advances in Chitin Science. Vol. 4. University of Potsdam, Potsdam, Germany2000: 524-529Google Scholar). Asp-142 contributes to the correct positioning of the N-acetyl group of the -1 sugar, modulation of the pKa of Glu-144 during the catalytic cycle, and stabilization of the oxazolinium ion intermediate. Mutation of Asp-142 to Ala in ChiB almost completely abolishes catalytic activity, whereas mutation to Asn reduces activity ∼50-fold (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 17Synstad B. Gaseidnes S. Vriend G. Nielsen J.E. Eijsink V.G.H. Peter M.G. Muzzarelli R.A.A. Domard A. Advances in Chitin Science. Vol. 4. University of Potsdam, Potsdam, Germany2000: 524-529Google Scholar). chitinase B wild type ChiB N,N′-diacetyl-chitobionoxime-N-phenylcarbamate electrospray ionization mass spectroscopy bovine serum albumin. The best known inhibitor of family 18 chitinases is allosamidin (18Sakuda S. Isogai A. Matsumoto S. Suzuki A. Koseki K. Tetrahedron Lett. 1986; 27: 2475-2478Google Scholar) (Fig. 1), a pseudotrisaccharide that inhibits with Ki values in the 1 nm-1 μm range (19Sakuda S. Muzzarelli, R.A.A Chitin Enzymology. Vol. 2. Atec Edizioni, Ancona, Italy1996: 203-212Google Scholar). Structural studies have shown that allosamidin binds in the -3 to -1 subsites (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 15Bortone K. Monzingo A.F. Ernst S. Robertus J.D. J. Mol. Biol. 2002; 320: 293-302Google Scholar, 20Terwisscha van Scheltinga A.C. Kalk K.H. Beintema J.J. Dijkstra B.W. Structure. 1994; 2: 1181-1189Google Scholar, 21Papanikolau Y. Tavlas G. Vorgias C.E. Petratos K. Acta Crystallogr. Sect. D Biol. Crystallogr. 2003; 59: 400-403Google Scholar, 22Rao F.V. Houston D.R. Boot R.G. Aerts J.M.F.G. Sakuda S. van Aalten D.M.F. J. Biol. Chem. 2003; 278: 20110-20116Google Scholar). The moiety bound to the -1 subsite is an allosamizoline, which resembles the oxazolinium ion reaction intermediate (Fig. 1). The only other known high affinity inhibitors of family 18 chitinases are the naturally occurring cyclopentapeptides argadin (4Arai N. Shiomi K. Yamaguchi Y. Masuma R. Iwai Y. Turberg A. Koelbl H. Omura S. Chem. Pharm. Bull. 2000; 48: 1442-1446Google Scholar) and argifin (23Shiomi K. Arai N. Iwai Y. Turberg A. Koelbl H. Omura S. Tetrahedron Lett. 2000; 41: 2141-2143Google Scholar). Structural and enzymological analyses have shown that these peptides bind intimately to the active site of ChiB with affinities in the 10 nm (argadin) and 10 μm (argifin) range (24Houston D.R. Shiomi K. Arai N. Omura S. Peter M.G. Turberg A. Synstad B. Eijsink V.G.H. van Aalten D.M.F. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 9127-9132Google Scholar). Although total syntheses of allosamidin have been reported (reviewed in Ref. 25Berecibar A. Grandjean C. Siriwardena A. Chem. Rev. 1999; 99: 779-844Google Scholar), these syntheses are lengthy and do not offer a practical access to the inhibitor. Until recently, allosamidin was commercially available at a cost of $500 (United States currency) per milligram, but it is currently not being sold. No total synthesis of the cyclic pentapeptides argifin and argadin has been reported. Attempts to synthesize oligo-GlcNAc-based chitinase inhibitors have had limited success to date, mainly due to the fact that the affinity of chitinases for short, non-hydrolyzable chitooligosaccharides (e.g. GlcNAc2) is usually low and that longer oligosaccharides with higher affinities are too readily degraded. Thus, effective sugar-based inhibitors need to be based on the introduction of non-hydrolyzable bonds between the sugar moieties and/or on combining short chito-oligosaccharides with other structural elements that increase affinity while preventing degradation of the inhibitor. To the best of our knowledge, such designed inhibitors of family 18 chitinases with an affinity in the μm range have not yet been identified. Here, we study the interaction between the family 18 chitinase ChiB from S. marcescens and the previously designed and synthesized GlcNAc2-derivative N,N′-diacetyl-chitobionoxime-N-phenylcarbamate (HM508; Fig. 1) (26Beer D. Maloisel J.L. Rast D.M. Vasella A. Helv. Chim. Acta. 1990; 73: 1918-1922Google Scholar). We show that HM508 inhibits ChiB with a Ki in the 50 μm range. The crystal structures of complexes between HM508 and ChiB and ChiB-D142N reveal details of the interactions between the inhibitor and the enzyme. The importance of some of these interactions was probed by studying the effects of two site-directed mutations (M212A, W220A) in the active site of ChiB on the affinity for HM508. We also show that ChiB is capable of slowly hydrolyzing HM508 to chitobionolactone/chitobiono-hydroxy acid and that the enzyme's active site is optimized for binding the lactone form of this degradation product. Overexpression, Purification, and Crystallization of ChiB from S. marcescens—Wild type chitinase B and the D142N mutant were overexpressed in Escherichia coli and purified as described elsewhere (27Brurberg M.B. Nes I.F. Eijsink V.G.H. Microbiology. 1996; 142: 1581-1589Google Scholar). The pure protein of wild type and D142N was lyophilized, dissolved to 1.0 mg/ml in 25 mm Tris buffer, pH 8, dialyzed overnight in the same buffer, and concentrated to 10 mg/ml before it was used in hanging drop vapor diffusion crystallization experiments. Crystals for wild type appeared within 3 days in 100 mm HEPES, pH 7, 10% glycerol, and 1.5 m ammonium sulfate. Crystals for the D142N mutant appeared within 3 days in 100 mm HEPES, pH 7, 15% glycerol and 1.3 m ammonium sulfate. Wild type and D142N crystals were soaked with 100-fold molar excess of HM508 for 1 month and 3 h, respectively, before being frozen in a nitrogen stream. Data were collected at beamline ID14 at the European Synchrotron Radiation Facility in Grenoble, France. Further mutants of ChiB were produced by site-directed mutagenesis using the Stratagene (La Jolla, CA) QuikChange kit, as described previously (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 17Synstad B. Gaseidnes S. Vriend G. Nielsen J.E. Eijsink V.G.H. Peter M.G. Muzzarelli R.A.A. Domard A. Advances in Chitin Science. Vol. 4. University of Potsdam, Potsdam, Germany2000: 524-529Google Scholar). Mutants were overexpressed and purified as described above. Structure Determination—The data were processed using DENZO and scaled using SCALEPACK from the HKL suite (28Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326PubMed Google Scholar). The previously published structure of the wild type enzyme in complex with allosamidin (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar) was used as template for an initial rigid body refinement. Both structures were refined in crystallography and NMR systems software (29Brunger A.T. Adams P.D. Clore G.M. Gros P. Grosse-Kunstleve R.W. Jiang J.-S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Google Scholar), which included initial simulated annealing combined with iterative model building in O (30Jones T.A. Zou J.Y. Cowan S.W. Kjeldgaard M. Acta Crystallogr. Sect. A. 1991; 47: 110-119Google Scholar). The starting structure and molecular topology for HM508 were created using the PRODRG server (31van Aalten D.M.F. Bywater R. Findlay J.B.C. Hendlich M. Hooft R.W.W. Vriend G. J. Comput. Aided Mol. Des. 1996; 10: 255-262Google Scholar). Two forms of the inhibitor were refined. For the D142N-HM508 data set, the full ligand was observed in the unbiased |Fo| - |Fc|,ϕcalc density (Fig. 4) and included in the refinement. For the WT-HM508 data set, the unbiased |Fo| - |Fc|,ϕcalc density in the active site and ESI-MS data (see below) showed a product of hydrolysis of HM508, viz. chitobionolactone. Refinement statistics are shown in Table I.Table IRefinement and structure quality statistics of WT-HM508 and D142N-HM508 Values in parenthesis are in the outer resolution shell. Crystals were of space group P212121. No resolution or I/σ I cutoffs were applied to the data used for the refinement. R.m.s.d. is root mean square deviation.WT-HM508D142N-HM508Unit cell (Å) a55.3455.52 b103.48104.17 c186.02185.99Resolution (Å)30—1.90 (1.97—1.90)25—1.80 (1.86—1.80)Observed reflections321,320 (20,722)347,374 (15,937)Unique reflections84,877 (7748)97,967 (7746)Redundancy3.8 (2.7)3.5 (2.1)Mean I/σI9.3 (3.2)25.9 (2.7)Completeness (%)99.0 (91.7)97.1 (77.8)Rmerge0.058 (0.37)0.052 (0.30)Rcryst0.2090.208Rfree0.2450.255Rfree reflections1230983Protein atoms77267816Water molecules756892Glycerol molecules926SO4 molecules77Inhibitor molecules22R.m.s.d. from ideal geometry Bonds (Å)0.0140.010 Angles (°)1.581.48〈B〉 (Å2) protein29.832.2〈B〉 (Å2) HM50824.828.2 Open table in a new tab The crystals used for this study have two monomers in the asymmetric unit (named the A and B monomers) related by a 2-fold non-crystallographic axis (32van Aalten D.M.F. Synstad B. Brurberg M.B. Hough E. Riise B.W. Eijsink V.G.H. Wierenga R.K. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 5842-5847Google Scholar). As with previous complexes, comparison of the monomers did not reveal significant conformational changes or differences in the binding site. The monomer best defined by the electron density maps was used for further analyses (the B monomer for the ChiB-HM508 complex and monomer A for the ChiB-D142N-HM508 complex). Inhibition Studies—Before being used in enzyme assays, the lyophilized inhibitor was dissolved in doubly distilled water to yield a working solution of 1 mm. Chitinase activity was measured using 4-methylumbelliferyl-β-d-N,N′-diacetylchitobioside (Sigma) as a substrate. It has previously been shown that this assay permits accurate determination of kinetic parameters despite the fact that substrate inhibition necessitates the use of a relatively narrow range of (low) substrate concentrations (27Brurberg M.B. Nes I.F. Eijsink V.G.H. Microbiology. 1996; 142: 1581-1589Google Scholar). At first, inhibitor affinity was estimated by determination of IC50 values at 20 μm substrate concentration. Subsequently, Ki values were approximated by determining kinetic parameters (Km, kcat) in the presence of inhibitor at a concentration close to the IC50. Calculation of kinetic and inhibitory parameters was performed using the curve fit option in the Enzyme Kinetics!Pro software package (ChemSW, Fairfield, CA). Reaction mixtures for measurement of kinetic parameters contained 1.15 nm WT, 26 nm W220A, 10 nm D142N, or 16 nm M212A substrate at a concentration ranging from 5-20 μm (WT), 30-120 μm (W220A), 2.5-10 μm (D142N and M212A), and 0.1 mg/ml BSA in 50 mm citrate-phosphate buffer, pH 6.3. HM508 concentrations used were 100 μm (WT), 125 μm (W220A), 1 μm (D142N), and 1 μm (M212A). The reaction mixtures were incubated at 37 °C, and 50-μl samples were taken after 0, 2.5, 5, and 7.5 min and transferred immediately to 1.95 ml of 0.2 m Na2CO3 to stop the reactions. The amount of liberated 4-methylumbelliferyl was measured by fluorometry. Product formation was linear over time for all ChiB variants at all substrate concentrations. Analysis of HM508 Degradation—To investigate degradation of HM508, wild type ChiB, ChiB-M212A, ChiB-D142N, and ChiB-W220A were incubated for 17 days at room temperature with and without HM508 under normal assay conditions without substrate. To accelerate the rate of degradation, the enzymes were used at a concentration higher than that for the published standard assay. The reaction mixtures contained 3.6 nm WT, 19 nm M212A, 57 nm D142N, or 98 nm W220A and 0.1 mg/ml BSA, in 50 mm citrate-phosphate buffer, pH 6.3. HM508 concentrations used were 75 μm (WT), 2 μm (M212A), 5 μm (D142N), and 200 μm (W220A). After 10 min and 1, 3, 9, 13, and 17 days of incubation, 45-μl samples were removed from these enzyme-inhibitor mixtures and from parallel control mixtures (enzyme without inhibitor) for determination of enzyme activity. Activities were determined by the addition of 5 μl of substrate (final concentration 20 μm) and by using standard conditions (T = 37 °C; pH = 6.3; incubation time = 10 min). All measurements were performed in duplicate. Activities in the enzyme-inhibitor mixtures are expressed as a percentage of the activity in the corresponding control sample (no inhibitor). The control reactions did not show any significant loss of activity during the incubation period. ESI-MS was used to analyze degradation products of HM508 in the absence and presence of ChiB. 1 μm ChiB (wild-type or the D142N mutant) was incubated with 100 μm HM508 in 50 mm ammonium acetate buffer, pH 6 (the specific activity of ChiB in this buffer is similar to the specific activity in the citrate-phosphate buffer used in standard activity assays). We used high concentrations of enzyme and inhibitor to mimic the conditions during crystallization and soaking and to abolish the need for the addition of BSA as an enzyme stabilizer (BSA gave a dramatic increase in background noise in subsequent analyses). Samples were analyzed by ESI-MS at the start of the incubation and at various time points up to 30 days later. ESI-MS analysis was conducted using a Micromass Quattro LC triple quadrupole instrument. The sample was introduced directly into the source using a syringe pump at a flow rate of 10 μl/min. Mass spectra were obtained in positive mode, with cone voltage at 30 V, capillary voltage at 3 kV, and multiplier voltage at 650 V. The desolvation and source block temperatures used were 160 and 80 °C, respectively. The mass scan range was 50-650 m/z. Inhibitor Stability—Initial crystallographic soaking studies of wild-type ChiB with HM508 suggested that the inhibitor was degraded. Therefore, we studied the possible degradation of HM508 by ChiB and a less active variant of ChiB (D142N) (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 17Synstad B. Gaseidnes S. Vriend G. Nielsen J.E. Eijsink V.G.H. Peter M.G. Muzzarelli R.A.A. Domard A. Advances in Chitin Science. Vol. 4. University of Potsdam, Potsdam, Germany2000: 524-529Google Scholar) using relatively high concentrations of enzyme and inhibitor (thus mimicking to some extent the conditions during the crystal-soaking experiments). ESI-MS analyses of mixtures of ChiB and HM508 showed that a 557 peak corresponding to the intact inhibitor (557 for HM508 + H+) and the K+ adduct peak (595 for HM508 + K+) disappeared over time, whereas a new set of peaks (441 for M + H and 479 for M + K+) appeared (Fig. 2). As explained below, a hydrolytic reaction that would generate a compound with mass 440 would be the generation of a chitobionolactone (Fig. 1) in equilibrium with its hydroxy acid. The equilibrium between these compounds is expected to lie far toward the hydroxy acid at neutral pH (33Secemski I.I. Lienhard G.E. Lehrer S.S. J. Biol. Chem. 1972; 247: 4740Google Scholar). The degradation of the inhibitor was a slow process; after 17 days ∼50% of the inhibitor had been degraded by the enzyme (Fig. 3). Conversion of HM508 to a compound with a molecular mass of 440 was not observed in the absence of enzyme (Fig. 2) nor in the presence of the less active D142N mutant of ChiB (not shown).Fig. 3Effect of HM508 degradation on enzyme inhibition. Enzyme activities are expressed relative to the activity in the samples without inhibitor; ChiB wild type (triangles), ChiB-M212A (squares), ChiB-D142N (diamonds), and ChiB-W220A (crosses).View Large Image Figure ViewerDownload (PPT) To explore the significance of the enzymatic HM508 degradation in our standard enzyme inhibition assays, ChiB, ChiB-M212A, ChiB-W220A, and ChiB-D142N were incubated for 17 days with and without HM508 under normal assay conditions with no substrate. Fig. 3 shows that the ChiB-D142N HM508 mixture was the only one with stable enzyme activity throughout the incubation period. For the remaining enzyme-inhibitor mixtures, activities gradually increased until they reached levels close to what was observed in the corresponding samples without HM508. These results confirm that degradation of the inhibitor is a slow process. Because degradation of the inhibitor occurs on a significantly longer time scale (days) than experiments for determination of kinetic parameters (minutes), the Ki values determined in this study apply to intact HM508. Structural Analysis—Because of the degradation of HM508 in the wild type enzyme, the interaction of HM508 with ChiB was analyzed by solving the structure of HM508 in complex with the less active ChiB mutant (D142N) that did not show degradation of the inhibitor within the time scale of our experiments. In addition, we solved the structure of ChiB using a crystal that had been soaked with HM508 for several months to determine the structure of the degradation product. The ChiB-HM508 and ChiB-D142N-HM508 complexes were refined to 1.90 and 1.80 Å, respectively, with R-factors converging at R (Rfree) = 0.209 (0.245) and 0.205 (0.255), respectively. Statistics of the refinement and the final models are shown in Table I. Both structures showed well defined unbiased |Fo| - |Fc|,ϕcalc density in the -1 and -2 subsites that corresponded to the sugar moieties of the inhibitor (Fig. 4). In addition, the ChiB-D142N-HM508 maps showed density for the HM508 phenylcarbamate moiety, which interacts with the +1 and +2 subsites of the enzyme (Trp-97 and Trp-220; Fig. 4). In the D142N-HM508 structure, the -1 sugar lies toward the 1,4B conformation, close to the conformation seen for the -1 sugar in a complex between an inactive mutant of ChiB (E144Q) and GlcNAc5 (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar) (Fig. 4). In general, the active site architecture and the interactions between the disaccharide moiety and the enzyme are similar to those observed in the E144Q-GlcNAc5 complex (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar), as shown in Fig. 4 and Table II. One of two major differences concerns residue 142, which, in the E144Q-GlcNAc5 complex and in most other ChiB-ligand complexes, hydrogen bonds to Glu-144 (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar, 24Houston D.R. Shiomi K. Arai N. Omura S. Peter M.G. Turberg A. Synstad B. Eijsink V.G.H. van Aalten D.M.F. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 9127-9132Google Scholar). This hydrogen bond does not exist in the D142N-HM508 complex because the side chain of Glu-144 is in a different position (rotated 14° around χ1, 32° around χ2, and 86° around χ3 compared with E144Q-GlcNAc5) to hydrogen bond the N1 of the HM508 oxime (Fig. 4; Table II). A second difference concerns the position of the glycosidic oxygen in the E144Q-GlcNAc5 structure and the equivalent N1 nitrogen of the HM508 oxime (shift of 1.1 Å). The electron density and the planar conformational restraints for the sp2-hybridized C1 carbon place the oxime N1 0.9 Å closer to the conserved Met-212 than the equivalent glycosidic oxygen in the E144Q-GlcNAc5 structure (Table II; Fig. 4). Although this does not directly generate a steric clash (distance between N1 and Met-212-Sδ is 3.7 Å compared with 4.6 Å for the equivalent distance in the E144Q-GlcNAc5 complex), it does appear to distort the linker between the -1 sugar and the phenyl ring in HM508 from planarity (by as much as 27° over the oxime bond).Table IIInteractions observed in the -2, -1 and +1 subsites in the E144Q-GlcNAc5, D142N-HM508, and WT-HM508 structures Hydrogen bond donor-acceptor distances (D-A column) were calculated with WHAT IF (43Vriend G. J. Mol. Graph. 1990; 8: 52-56Google Scholar), along with the geometrical quality of the hydrogen bonds (HB2 column), using the HB2 algorithm (44Hooft R.W.W. Sander C. Vriend G. Proteins. 1996; 26: 363-376Google Scholar). The HB2 values range from 0 (no hydrogen bond) to 1 (optimal hydrogen bond). Weak hydrogen bonds (HB2 < 0.3) are not listed.D142N-HM508E144Q-NAG5WT-δ-lactoneSubsiteH-bondD-AHB2H-bondD-AHB2H-bondD-AHB2—2Tyr-9 N-O630.79Tyr-98 N-O63.10.88Tyr-98N-O630.79—2Trp-403 Nϵ1-O72.90.66Trp-403 Nϵ1-O730.66Trp-403 Nϵ1-O72.90.70—2Gln-407 Nϵ2-O73.20.51Gln-407 Nϵ2-O73.50.43Gln-407 Nϵ2-O73.40.48—1Glu-144 Oϵ2-N12.80.59Glu-144 Oϵ2-O430.90—1Trp-97 N-O330.87Trp-97 N-O33.10.90Trp-97 N-O33.20.89—1Tyr-214 Oη-O72.60.78Tyr-214 Oη-O72.60.75Tyr-214 Oη-O72.60.77—1Asp-215 Oδ2-O62.70.89Asp-215 Oδ2-O62.70.76Asp-215 Oδ2-O62.80.95—1Asp-142 Oδ1-N22.80.80Asp-142 Oδ1-N22.90.74Asn-142 Oδ2-N22.80.90+1Tyr-145 Oη-O23.30.51Tyr-145 Oη-O62.70.59 Open table in a new tab The phenylcarbamate group hydrogen bonds with Asp-215-Oδ2 through the Nγ nitrogen and with Tyr-145-Oη through the Oβ oxygen (Figs. 1 and 4; Table II). The hydrophobic phenyl ring is located close to Trp-97 and Trp-220 that line the +1 and +2 subsites of ChiB. These two aromatic residues are shifted toward the ligand (up to 0.5 Å compared with the apo-D142N structure) (34Kolstad G. Houston D.R. Rao F.V. Peter M.G. Synstad B. van Aalten D.M.F. Eijsink V.G.H. Biochim. Biophys. Acta. 2003; (in press)Google Scholar). A similar, but larger displacement (shifts up to 1 Å) has previously been observed upon substrate binding to E144Q (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar). Additionally, main chain movements of the loops covering the active site indicate a closure of the roof of the active site tunnel, equivalent to what was observed in the E144Q-GlcNAc5 structure (13van Aalten D.M.F. Komander D. Synstad B. Gåseidnes S. Peter M.G. Eijsink V.G.H. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 8979-8984Google Scholar). The phenyl group does not form the triple sandwich with Trp-97 and Trp-220, as was observed in the E144Q-GlcNAc5 structure for the +1/+2 sugars; rather, it interacts with these tryptophans in an orthogonal orientation. A similar interaction has been observed for the phenylalanine side chain in the cyclic pentapeptide inhibitor argifin (24Houston D.R. Shiomi K. Arai N. Omura S. Peter M.G. Turberg A. Synstad B. Eijsink V.G.H. van Aalten D.M.F. Proc. Natl. Acad. Sci.
Referência(s)