sign in sign up
<<
Home , Amino acid, Glutaminase

BNL-108431-2015-JA

Human Gamma-Glutamyl Transpeptidase: Inhibitor Binding and Movement within the

Simon S. Terzyan, Anthony W. G. Burgett, Annie Heroux, Blaine H.M. Mooers, and Marie H. Hanigan

Submitted to Journal of Biological Chemistry

July 2015

Photon Sciences Department

Brookhaven National Laboratory

U.S. Department of Energy USDOE Office of Science, Basic Energy Sciences

Notice: This manuscript has been authored by employees of Brookhaven Science Associates, LLC under Contract No. DE- SC0012704 with the U.S. Department of Energy. The publisher by accepting the manuscript for publication acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or any third party’s use or the results of such use of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof or its contractors or subcontractors. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. BNL-108431-2015-JA

Human Gamma-Glutamyl Transpeptidase: Inhibitor Binding and Movement within the Active Site

Simon S. Terzyan1, Anthony W. G. Burgett2, Annie Heroux3, Blaine H.M. Mooers4,6 and Marie H. Hanigan5,6

1Macromolecular Crystallography Laboratory Department of and Molecular University of Oklahoma Health Sciences Center Stanton L. Young Biomedical Research Center, Rm 472 975 N.E. 10th Street Oklahoma City, OK 73104, U.S.A.

Energy Sciences Directorate/Photon Science Division Brookhaven National Laboratory Building 745, Box 5000 Upton, NY 11973-5000, U.S.A.

4 Department of Biochemistry and Molecular Biology University of Oklahoma Health Sciences Center Stanton L. Young Biomedical Research Center, Rm 468 975 N.E. 10th Street Oklahoma City, OK 73104, U.S.A.

5Department of Biology University of Oklahoma Health Sciences Center Stanton L. Young Biomedical Research Center, Rm 264 975 N.E. 10th Street Oklahoma City, OK 73104, U.S.A.

6Stephenson Cancer Center University of Oklahoma Health Sciences Center Oklahoma City, OK 73104, U.S.A.

ABSTRACT: Gamma-glutamyl transpeptidase (GGT1) is a cell surface, Ntn- that cleaves , glutathione-conjugates and other gamma-glutamyl compounds. Its expression is essential to and its induction has been implicated in the pathology of asthma, reperfusion injury and cancer. In this study, we report the of the human apo- and human GGT1 bound to a series of inhibitors: GGsTop, diazonorleucine, and -borate. These glutamate analogs inhibit GGT1 activity and have been used to investigate the role of the enzyme in normal physiology and disease. Analysis of the structures reveals movement within the active site during binding of the glutamate including rotation of the of the catalytic . These data are the first structures reported for these inhibitors bound to GGT1 from any . The of the diazonorleucine-hGGT1 complex reveals that the mechanism by which diazonorleucine binds to and inhibits GGT1 differs from previously proposed mechanisms. These structures provide critical insights into inhibition of GGT1 which will aid in the further development of new classes of inhibitors that can be used clinically. INTRODUCTION

Gamma-glutamyl transpeptidase (GGT1, a.k.a. gamma-glutamyl ) is a cell surface enzyme that is expressed on the apical surface of ducts and glands throughout the body (Hanigan and Frierson 1996). The highest level of enzyme activity is on the surface of the proximal tubules of the where it cleaves glutathione in the glomerular filtrate preventing its excretion from the body and thereby conserving cysteine (Lieberman, Wiseman et al. 1996). The enzyme cleaves the gamma-glutamyl bond of any substrate in which the glutamate moiety is unfettered (Fig. 1). Substrates include oxidized and reduced glutathione, glutathione S- conjugates, Leukotriene C4 and glutathione S- (Hogg, Singh et al. 1997, Wickham, West et al. 2011). hGGT1 is induced and mislocalized in a variety of diseases resulting in access to substrates in serum and in interstitial fluid. GGT1 activity has been shown to potentiate damage in ischemia-reperfusion injury, to contribute to airway hyper-reactivity in asthma, to activate a series of glutathione-S drug conjugates to nephrotoxins and to increase resistance of tumors to alkylating agents (Anders and Dekant 1998, Hanigan, Gallagher et al. 1999, Townsend, Deng et al. 2003, Pompella, De Tata et al. 2006, Yamamoto, Watanabe et al. 2011, Tuzova, Jean et al. 2014).

Inhibitors of GGT1 that have been evaluated clinically, diazonorleucine (DON) azaserine and acivicin, are all glutamate analogs, and in humans they are extremely toxic (Ahluwalia, Grem et al. 1990). To better understand the interaction between glutamate-based inhibitors and the enzyme we have prepared a series of inhibitors with human GGT1 (hGGT1) and solved the structures. We recently reported the crystal structure of glutamate-bound hGGT1, the first structure reported for any eukaryotic GGT (West, Chen et al. 2013). The structures of the hGGT1-bound GGsTop, diazonorleucine (DON), and serine-borate are the first structures reported for any of these inhibitors bound to GGT1. In this study we also report the structure of apo-hGGT1. Comparison of the apo-structure to the inhibitor bound structures has revealed movement within the active site upon inhibitor binding. To pursue development of less toxic inhibitors it is essential to understand the interaction between inhibitors and the enzyme.

MATERIALS AND METHODS

hGGT1 Expression and Purification: For crystallization studies hGGT1 was expressed in strain X-33, purified and deglycosylated as described previously (West, Chen et al. 2013).

Add activity assay to methods for reporting malate inhibition data

Thermoflour Study: The sample consisted of 0.1 mg/ml hGGT1 alone or complexed with one of two inhibitors (GGsTop, Waco Chemicals, Richmond, VA or DON, Sigma-Aldrich, St. Louis, MO) in 10 mM Hepes buffer pH 7.5, 150 mM NaCl and 5x SYPRO orange. To each well of a 96 well plate, 12 µl of the protein sample and 4 µl of screening buffer were added. Nine buffers at 12 different were used. The plate was spun for 5 minute at 1000 rpm to remove air bubbles and then placed in an thermocycler 7500 RT-PCR. The temperature of the samples was increased from 25 to 95 oC at 1 oC per minute. At each degree the fluorescence of the protein-bound SYPRO orange was measured.

Crystallization Conditions: Crystals of hGGT1 were grown by vapor diffusion with the hanging drop method. The protein stock solution contained 4.3 mg/mL hGGT1 in 50 mM Hepes pH 8.0, 0.5 mM EDTA and 0.02% sodium azide. Crystallization drops contained 2 µl of protein solution, 1.7 µl of H2O and 2 µl of reservoir solution. Drops were equilibrated against 500 µl of one of two reservoir solutions. Solution A contained 20-25% PEG3350, 0.1 M Na Cacodilate buffer pH 6.0 and 0.1 M ammonium chloride. Reservoir solution B contained 16% PEG6000, 0.1 M MES buffer pH 6.3 and 0.1 M ammonium chloride. Two days after setting the drops, microcrystals of previously grown crystals were added to the drops to facilitate crystal growth. Crystals appeared in one or two days after seeding. After an additional week the crystals grew to a final size of approximately 0.05 mm x 0.1 mm x 0.5 mm.

Crystals of the apo-enzyme were grown against reservoir solution B. Data for glutamate-bound hGGT1 were collected from a crystal of hGGT1 grown in presence of 2.5 mM glutamate against reservoir Solution A. The crystal was soaked for 2.5 hours in reservoir Solution A supplemented with 10 mM glutamate and 1 mM OU749 inhibitor prior to cryopreservation.

Crystals of hGGT1-GGsTop complex were prepared with hGGT1 preincubated in 1 mM GGsTop. Two µl of 0.1 M GGsTop in 0.1 N HCl was added to 100 µl of the protein solution. The mixture was incubated overnight at 4oC prior to preparing the crystallization drops against reservoir Solution B. Crystals of the hGGT1-DON complex were prepared with hGGT1 preincubated in DON. Two µl of 0.5 M DON (in 50 mM Hepes, pH 8.0, 0.5 mM EDTA and 0.02% sodium azide) was added to 50 µl of protein solution. The mixture was incubated for 48 hours at 4oC prior to preparing the crystallization drops against reservoir Solution A. Crystals of GGT1 with serine-borate were prepared by soaking crystals of the apo form of hGGT1 (grown against reservoir solution A) for 15 min in reservoir Solution A supplemented with 10 mM L-serine borate. The stock serine borate solution contained 0.5 M tris-borate, pH 7 and 0.5 M L-serine.

Freezing Crystals: Crystals were quickly dragged through a cryoprotectant solution (the reservoir solution with 15% PEG1500 and the corresponding inhibitor molecules). The crystals were flash frozen in liquid .

Data Collection: Crystals were screened at the Macromolecular Crystallography Laboratories at The University of Oklahoma in Norman, OK and The University of Oklahoma Health Sciences Center in Oklahoma City, OK. The majority of the X-ray diffraction data were collected at 100 K at beam line X25 at the National Synchrotron Light Source, Brookhaven, NY. The beamline was equipped with a Pilatus 6M detector and data collected at 1.1 Å at 100 K.. Data for the serine- borate bound-GGT1 were collected at Stanford Synchrotron Radiation Light Source beamline 14-1,λ = 0.979 Å. The data were processed by HKL-2000 suite (Otwinowski and Minor 1997).

Structural Determination and Refinement: The unit cell parameters of crystals were isomorphous to ones from our published structure (4GDX) of hGGT1 in complex with glutamate. Rigid body refinement was used for positioning of the new structures in the unit cell. Difference 2Fo-Fc and Fo-Fc maps were used for detection of bound inhibitor molecules. The 4GDX structure without alternative conformations, and molecules served as a starting model. The structures were refined using REFMAC and were visually corrected using the graphic program COOT (Murshudov, Vagin et al. 1997, Emsley and Cowtan 2004). In the last stages of refinement, cofactor atoms (Cl, Na) and water molecules were added to structure using ARP/wARP or/and COOT programs. The refinement statistics and PDB accession are listed in Table 1.

RESULTS AND DISCUSSION

Movement within the Active Site: We have solved the structure of apo-hGGT1. In addition, we collected new high resolution data for glutamate-bound hGGT1 from a crystal that was soaked in a solution containing a higher concentration of glutamate than was used for our original structure. In our previous structure of glutamate-bound hGGT1 (4GDX1), the α-carboxyl and α- amino groups of the bound-glutamate showed significant electron density, but the electron density observed for the side chain atoms was poor. As a result the glutamate side chain atoms were not modelled into the structure. Our new structure of the glutamate-bound hGGT1 includes all of the glutamate side chain atoms. In comparing the apo-enzyme and the new glutamate- bound hGGT1 structures we have observed displacement of the amino backbone of the enzyme and rotation of side chains upon glutamate binding.

Purified hGGT1 was crystalized, complete x-ray diffraction data were collected and the structure of the apo-enzyme was solved (Table 1, Fig. 2). The crystal was isomorphous to the glutamate- bound hGGT1 crystal structure (4GDX1) (West, Chen et al. 2013). Therefore, the initial Fourier maps for the apo-enzyme were calculated using the coordinates of the glutamate-bound hGGT1 with glutamate, water, Cl and Na molecules removed. Residues with alternative positions were modeled in the higher occupancy conformation. The structure at 2.3 Å resolution was refined to Rwork=14.4 % and Rfree=19.5%.

Data collected for the glutamate-bound hGGT1 extended to 2.1 Å resolution. The structure was refined to Rwork and Rfree values of 0.144 and 0.1812, respectively. The initial phases for the structure factors were evaluated using only the coordinates of the atoms from protein and molecules within the original glutamate-bound structure (4GDX1). The new glutamate-bound structure (PDB#) showed clear electron density for all atoms of the glutamate molecule bound in the active site of the enzyme (Fig. 3). In both this new glutamate-bound hGGT1 structure (PDB#) and our original glutamate-bound structure (4GDX1), the α-carboxyl group of the bound glutamate formed bonds with Arg-107 NH1, Ser-451 OG, Ser-452 N, and bound to Ser-452 OG via a water molecule. The α-amino group of the bound glutamate formed bonds with Asn-401 OD1, Gln-420 OE1, and Asp-423 OD2. In the new glutamate-bound structure one of two atoms of the ɣ-glutamyl carboxyl group form hydrogen bonds with side chain OG atom of Thr381 and the main chain nitrogen of Gly474, while the second does not form bonds with any of the atoms in the enzyme. The data for the new glutamate-bound hGGT1 structure was collected from a crystal soaked in a solution containing 1mM OU749, an uncompetitive inhibitor of the enzyme, but in calculated Fourier maps there was no electron density that could be interpreted as a bound inhibitor molecule.

Overall the structures of apo and glutamate-bound forms of hGGT1 were similar. When the structures of the new glutamate-bound and apo forms were superimposed the rms deviation for 500 CA atoms with maximal distance cutoff of 1.5 Å was 0.23. When the CA atoms of large and small subunits were superimposed separately the rms deviations were 0.221 and 0.246, respectively, indicating larger flexibility for small subunit, which we also observed in comparing the apo-enzyme with the original glutamate-bound structure. The rms deviation between 500 CA atoms of apo and original glutamate-bound structures was 0.241.

A comparison of the structures of the apo-enzyme, original and new glutamate bound form of hGGT1 indicated movement in the active site during binding and catalytic cleavage of the substrate. The first region in which we observed conformational change was the side chain of Thr381. Kinetic analyses of the GGT1 reaction indicated that during cleavage of gamma- glutamyl substrates a bond forms between δ- of the gamma-glutamyl substrate and a nucleophile within the active site (Curthoys and Hughey 1979, Allison 1985). Affinity labelling and structural studies of bacterial GGTs identified the Thr at the N-terminus of the small subunit as the catalytic nucleophile (Inoue, Hiratake et al. 2000, Boanca, Sand et al. 2007). studies of the H. pylori GGT and hGGT1 indicated that Thr381 was the catalytic nucleophile in hGGT (Boanca, Sand et al. 2007). The crystal structure of the apo-form of hGGT1 showed that the side chain of Thr381 had two conformations (Fig. 4). These were similar to the two confirmations of Thr381 previously observed in the glutamate-bound hGGT1 (West, Chen et al. 2013). In the new glutamate-bound structure (PDB#) the catalytic Thr381 was detected in only one conformation (with chi=-169o). The structure of the glutamate- bound hGGT1 (PBD#) reveals the orientation of the Thr381 side chain in the enzyme-substrate complex. In the apo-enzyme, the two conformations of the Thr381 indicate unrestrained movement of the side chain when the active site is not occupied. The observation of both conformations in the original glutamate-bound structure could be explained by partial occupancy of glutamate molecule resulting in a mixture of glutamate-bound and apo-forms of the protein in crystal.

The largest shift within the active site was detected in the , which is formed by a loop composed of residues 473 to 475. The Fourier maps for our original glutamate-bound hGGT1 structure (AGDX1) showed an ill-defined electron density for residues 473 to 475, and therefore two conformations were modelled to better interpret the electron density (West, Chen et al. 2013). One of these conformations was similar to the single conformation observed in the apo-hGGT1 structure, although in the apo-hGGT1 structure, the loop consisting of amino 473 to 475 was shifted down towards Thr381. The second conformation that was modelled was almost identical to the single conformation of this loop detected in our new glutamate-bound structure (PDB#). These data further suggest that the original glutamate-bound structure consisted of a mixture of glutamate-bound and apo-forms of the protein in crystal. Comparing the position of the loop in the apo-enzyme (PDB#) to its position in the glutamate-bound structure (PDB#) reveals that with glutamate bound, the loop is pushed away from active site. The CA atom of Thr475 shows the largest shift of 1.17 Å. Upon glutamate binding, the shifts in CA atoms of the catalytically important residues G473 and G474 were 0.97 Å and 0.99 Å, respectively. These results clearly demonstrate plasticity in the oxyanion hole forming loop that switches from a closed to an open conformation upon substrate binding.

The other regions of the enzyme in which noticeable shifts were detected upon glutamate binding were the lid loop region (residues 428-439), the loop connecting β16 and β17 strands (residues 505-514) and the C terminal portion of α14. These elements of secondary structure form the top of a large cavity above glutamate- within the active site of the enzyme (Fig. 4). Within these regions, the largest displacements observed, when the CA atoms of the apo-enzyme and the new glutamate-bound structures were superimposed, were 0.9 Å for N431, 0.59 Å for K326 and 0.26 Å for Q507. A comparison of the apo-structure to the original glutamate-bound structure also showed a shift in the lid loop with the largest displacement of 1.51 Å in the CA atom of residue N431. The largest displacement within the loop composed of residues 505-514 was a 0.9 Å shift of the CA atom of L509. Among residues 323-330 of helix α14 the largest shift (0.8 Å) was detected for the CA atom of K326. Despite the displacement of these residues, the overall conformations of these elements were similar.

The structures of the GGT homolog from several have been reported. The binding mode of the α-carboxyl and α-amino groups of glutamate in the active site of GGT is almost identical in human, E-coli, and GGT structures. However, the conformation of side chains of the bound glutamate differ. Compared to the E-coli structure (2DBX), the plane of the ɣ-glutamyl carboxyl in the human enzyme is rotated about 45o around CG_CD bond, facilitating the formation of a between its oxygen and the OG1 atom of Thr381 (Okada, Suzuki et al. 2006). In Helicobacter pylori GGT (2QM6), the ɣ-glutamyl carboxyl is shifted away from the threonine towards the oxyanion hole and is wedged (drawn?) deeper into the active site {Morrow, 2007 #51}. Should we include some info about differences between the kinetics of human and bacterial GGT? I don’t want to get into a discussion of transpeptidation in this paper

A molecule of the MES buffer was observed within the crystal structure of the apo-hGGT1. MES was incorporated below the active site of the enzyme, almost parallel to side chain of Tyr403 and next to the side chains of Asn180, Lys181 and Thr539. One of three oxygen atoms of SO3 group of MES molecule formed a hydrogen bond with the main chain nitrogen of Tyr403 (2.82A), while the second oxygen interacts with the NZ atom of Lys407 through a water molecule. The oxygen in the six member ring of MES interacts with the side chain Oᵧ of Thr539. The MES molecule partially covered the entrance to the cavity located below Thr381. This cavity extends the active site cleft towards the membrane binding site of the enzyme. (Marie needs to repeat the enzyme assays with MES to include the data in the ms)

hGGT1 Bound to Inhibitors

Thermaflor studies: We conducted thermofluor studies to determine whether the inhibitors were stabilizing the enzyme (Table 2). Analysis of data from nine buffers +at 12 different pHs showed that hGGT1 is most stable at pH 7. Inactivating hGGT1 with either GGsTop or DON further stabilized the structure of the enzyme. GGsTop had the largest effect. It increased the melting temperature of the enzyme at pH 7 from 58 to 75 oC. As shown in the structures described below, the binding of GGsTop or DON to the apo-enzyme resulted in the formation of a network of new interactions which stabilized the structure and restrict movements within the enzyme.

Structure of hGGT1-GGsTop complex: GGsTop [2-amino-4-([3-(carboxymethyl) phenyl] (methyl)phosphono)-butanoic acid)] (Fig. 1) is a phosphonate-based inhibitor of GGT1 that was synthesized as an active site probe (Han, Hiratake et al. 2007). Prior to crystallization, we inactivated the enzyme by incubating it with a 20-fold molar excess of GGsTop overnight at 4oC. The data was collected at NSLS and the complex structure was solved at 2.2 Å resolution (Rwork=16.04 % and Rfree=21.87 %). The 2Fo-Fc and Fo-Fc maps were calculated after 10 cycles of rigid body and restrained refinement by Refmac5 with the coordinates of the apo-enzyme (PDB#) as a model. A clear density attached to Thr381 was observed in the difference Fourier maps (Fig. 5). We modelled GGsTop into the density and found that the density corresponded to a cleaved molecule of GGsTop. The α-amino and α-carboxy groups of the butanoic acid portion of the GGsTop molecule formed the same network of bonds with hGGT1 that were detected for the α-amino and α-carboxy groups of the glutamate molecule in the glutamate-hGGT1 bound structure (PDB#). The density in the GGsTop hGGT1 data indicated that upon binding in the active site, the GGsTop had been cleaved between the phosphorous and the oxygen atom attached to phenol ring. The structure showed a covalent bond between the phosphorous and the side chain OG of Thr381. There was no electron density corresponding to the leaving group of GGsTop, The second oxygen of the PO3 group of GGsTop formed hydrogen bonds with the two main chain nitrogen atoms of the oxyanion hole. The distances between the oxygen and the main chain of Gly473 and Gly474 were 2.67 Å and 2.6 Å, respectively. The methyl group attached to the third oxygen atom of PO3 group, mimicking the side chain of Cys in the molecule of glutathione, protruded out towards the solvent and did not make any interactions with the enzyme. This observation is consistent with our knowledge of hGGT1 substrates some of which have large bulky groups bound to the cysteine of glutathione, yet they are cleaved at the same rate as glutathione (Wickham, West et al. 2011).

This is first published structure of any GGT with GGsTop. In the development of GGsTop, Han and colleagues initially synthesized phosphono mono which were poor inhibitors of hGGT1 (Han, Hiratake et al. 2006). Addition of a methyl group to the second oxygen of phosponyl group increased the inhibitory activity of new compounds 2 orders of magnitude (Han, Hiratake et al. 2007). These authors hypothesized that three factors could contribute to this enhanced inhibitory activity. Our structure provides support for the first factor, but not the other two. The first factor proposed was the higher electrophilicity of neutral phosphonate diesters compared to mono-anionic phosphonates towards nucleophilic substitution, which occurs in the interaction between the inhibitor and the side chain oxygen of Thr381 (Behrman, Biallas et al. 1970). Our results showed that the phosphonyl group of GGsTop, one of best inhibitors of hGGT1 among the neutral phosphonate diesters, not only binds to active site of the enzyme but also undergoes a nucleophilic attack and forms a covalent bond with the Thr381 residue. This demonstrates that the electrophilicity of the phosphonyl group towards nucleophilic substitution plays an important role in inhibition of the enzyme. The second factor proposed by Han and colleagues was a higher affinity of neutral phosphonates to the active site of hGGT1 compared to negative ones (Lherbet and Keillor 2004). However, the structure of GGsTop- bound hGGT1 shows the side chain of Thr381 and the α-nitrogen atoms of Gly473 and Gly474 (that interact with phosphonyl moiety) are positively charged groups which prefer interaction with groups of opposite charge. Therefore, the negatively charged monoesters of phosphonates would have higher affinity to the active site of the GGT1 compared to more neutral diesters. The fact that the neutral diesters are better inhibitors of GGT1 than the phosphonate monoesters demonstrates that the electrophilicity of the group is more important for inhibitory activity than the affinity of the group towards the active site of the enzyme. The third factor was a possible interaction of the added methyl group with the hGGT1 protein molecule. Our structure shows no interaction between the methyl group and hGGT1, thus negating the hypothesis that such an interaction occurs and enhances inhibitory activity.

Structure of hGGT1-DON complex. DON (6-diazo-5-oxo-L-) is an irreversible inhibitor of GGT1 (Inoue, Horiuchi et al. 1977, Tate and Meister 1977). No structures have been reported for DON bound to any GGT, prokarytoic or eukarytotic. We solved the structure of the DON- hGGT1 complex at 2.2 Å resolution. The apo-form of human enzyme with water molecules and metal atoms removed was used as the initial model for the 2Fo-Fc and Fo-Fc difference Fourier maps. The maps calculated after 10 cycles of rigid body refinement revealed a clear density for a molecule bound in the active site of the hGGT1 (Fig. 6). The density indicated that DON had been cleaved between the diazo carbon and diazo nitrogens. So a molecule of DON without the diazo nitrogen atoms was modeled into the density and the structure refined to a final Rwork and Rfree values of 15.8% and 21.8%, respectively. In the refined structure the α-carboxy and the α- amino groups of DON molecule occupy exactly the same position as the corresponding groups of glutamate in glutamate-bound structure (PDB#), and they participate in the same extensive net of hydrogen bonds and charge interactions with enzyme. The OG atom of Thr381 formed a covalent bond with the carbonyl carbon (C5) of DON. The oxyanion (O3 atom of DON) hydrogen bonds with the main chain nitrogens of Gly473 and Gly474, at distances 2.66 Å and 2.55 Å, respectively. The electron density maps clearly demonstrates the position of diazo carbon CE. In the structure of hGGT1-DON complex the catalytic residue Thr381 was detected in only one conformation that resembles one of two conformations of these residue (CHI=-169 o) in apo-form structure, but is further rotated around C-CA bond changing corresponding torsion angle from -40.9o to -31.06o. There was a small change in CHI angle of side chain as well from - 169 o to -175.7 o.

The structure of DON-bound hGGT1 shows a novel interaction between the inhibitor and the enzyme. The structure is contrary to that predicted by Tate and Meister in which the C6 carbon of DON would form a covalent bond with the side chain oxygen of a nucleophile in the active site (Tate and Meister 1978). Our structure clearly shows the OG atom of Thr381 formed a covalent bond with the carbonyl carbon (C5) of DON forming a tetrahedral adduct. It also shows a covalent bond between the diazo carbon of DON and the N-terminal nitrogen of small subunit. The mechanism that we propose for the formation of this structure is derived from a mechanism initially proposed, but then rejected for azaserine inhibition of GGT in E-coli. Azaserine and DON are similar molecules both containing diazo nitrogens. The structures of the two compounds differ by only a single atom, with the C4 of DON substituted with an oxygen in azaserine. Wada and colleagues determined the structure of E. coli GGT1 bound to azaserine (Wada, Hiratake et al. 2008). When the crystals were prepared in the dark, the diazo nitrogens were present in the structure. But, when the crystals were prepared in the light, the diazo nitrogens were not observed in the structure. The authors discussed two possible mechanisms which would lead to these results. The mechanism, favored by the authors was that GGT1 acts on azaserine as a substrate and cleaves the bond between carbonyl (C5) and diazo (C6) . This cleavage had also been proposed by others for some serine that hydrolyze peptiddyldiazomethanes to corresponding (Zumbrunn, Stone et al. 1988). Hartman and McGrath had also proposed this mechanism for the degradation of DON by , a reaction which releases glutamate and diazomethane (Hartman and McGrath 1973). According to the mechanism proposed by Wada and colleagues, the nucleophile of GGT1 initially forms a tetrahedral intermediate with an azaserine molecule yielding the structure of the E.coli GGT and azaserine prepared in dark. Cleavage of the C-C bond and removal of diazomethyl (CH2N2) is proposed to guide the reaction toward formation of an O- carboxyseryl enzyme intermediate with three C-O bonds. The attack of a water molecule on the carbonyl carbon would add an OH group to the carbon and result in a tetrahedral adduct as detected in the structure of the complex prepared in the light. In this scenario the diazo carbonyl of azaserine will be replaced with an OH group as part of the reaction. They favored this mechanism as an oxygen at the position of the due to the interactions that the OH group would form with the N-terminal nitrogen of the catalytic Thr residue and the main chain oxygen of Asn411 in E. coli GGT. However, there are several facts that argue against this mechanism. Azaserine is an irreversible inhibitor of GGT1 (Tate and Meister 1977). The mechanism proposed by Wada and colleagues would result in the formation of an acyl bond between the C5 of the original azaserine molecule and the nucleophile. The acyl bond would be liable. However, in the inhibitor-E.coli GGT structure, the enzyme was detected with a tetrahedral adduct of the inhibitor molecule. Our DON-bound hGGT1 structure also showed a tetrahedral adduct of the inhibitor bound to the enzyme. The fact that azaserine and DON are irreversible inhibitors disproves the mechanism proposed by Wada and colleagues because their mechanism proposes the formation of a rapidly cleaved acyl bond rather than the stable tetrahedral structure observed in the crystal structures. In case of E-coli GGT-Azaserine complex prepared under visible light, the detection of tetrahedral adduct could be explained by short soaking time and data collection at 100K. However, for our hGGT1-DON complex which was prepared by cocrystallization after 48 hour of preincubation of enzyme with DON, if the mechanism had included acyl bond formation, all of the DON molecules would have been converted to glutamate prior to crystallization.

The alternative proposal, discussed but then dismissed by Wada and colleagues, was that upon exposure of the GGT1-azaserine complex to visible light the diazo group decomposed with generation of a highly reactive divalent carbon (carbene). A similar reaction had been reported for a and diazo compound complex (Shafer, Baronows.P et al. 1966). However, Wada and colleagues dismissed this mechanism, stating that the existence of a carbene in the structure was impossible. However, our data suggest that this is the mechanism by which DON forms a complex with hGGT1. We proposed that when the GGT1-DON complex was exposed to visible light the diazo group decomposed with generation of carbene. A highly reactive carbene is very short lived and would form a covalent bond with N-terminal nitrogen of small subunit (Thr381), thus forming a six member ring consisting of the N, CA, CB, OG of Thr381 and the C5 and C6 atoms of DON. Such a complex is stable and would irreversibly inhibit the GGT1. This mechanism is also consistent with our observation that the distance between the main chain nitrogen of Thr381 and the C6 atom to which it is bound in the DON complex is very short.

LSQ superposition of the CA atoms of the enzyme in apo-form and in complex with DON showed no significant differences between two structures (rmsd deviation was 0.2 for the 519 CA atoms of whole molecule, 0.175 for large subunit and 0.247 for small subunit). The largest movement was detected in the loop of oxyanion hole. Upon binding of the DON molecule the loop moved up displacing CA atoms of Gly 473 and Gly474 by 0.9 Å and 1.15 Å respectively. The other noticeable movements in the main chain of the complex included shifts in the same regions of the molecule detected in the glutamate-bound form of hGGT1.

Structure of hGGT1-serine-borate complex. In 1959, Revel and Ball showed that serine in the presence of borate buffers inhibited mammalian GGT1, but neither serine nor borate buffer alone inhibited (Revel and Ball 1959). Tate and Meister proposed that in solution a serine- borate complex forms that mimics gamma-glutamyl-bound glutamate and that the complex binds in the active site of GGT1 (Tate and Meister 1978). Additionally, they suggested that the serine-borate complex interacts with hydroxyl group of the nucleophile in the active site of enzyme forming an enzyme-inhibitor complex that replicates the tetrahedral transition state of the enzyme. Indeed, the structure of this complex solved at 2.4 Å resolution (final Rwork=19.37 and Rfree=24.32) confirmed their hypothesis. The difference Fourier electron density calculated using coordinates of apo-form structure (only protein and carbohydrate atoms) clearly showed a serine-borate complex bound in active site of human GGT1 (Fig. 7). The serine-borate complex was formed by a covalent bond between side chain OG atom of serine and boron. The backbone atoms of serine occupy positions almost identical to main chain atoms of glutamate in the active site of the enzyme and form the same set of hydrogen bonds and bridges. The complex was linked to enzyme by an bond between boron and side chain oxygen of Thr381. The conformation of the side chain of Thr381 is similar to one of two conformations of this residue in the apo-form structure (with Chi=-169 o) with a slight shift of 0.19 Å for the CA atom away from the inhibitor molecule. This is the same orientation observed in the new glutamate-bound hGGT1 structure (PDB#). The boron was tetrahedral, with the two additional forming hydrogen bonds with the enzyme stabilizing the serine-borate complex in the active site of hGGT1. One oxygen(oxyanion) formed hydrogen bonds with the α-amino groups of Gly473 and Gly474 which form the oxyanion hole (distances were 2.88 Å and 2.87 Å, respectively). The second oxygen interacted with main chain nitrogen of Thr381 which is the N- terminal nitrogen of small subunit of hGGT1 (distance 2.55 Å), main chain oxygen of Asn401 (2.63 Å) and through a water molecule with α-oxygen of His 81.

There were no significant changes in overall structure when CA atoms of the serine-borate complex were superimposed on the structure of apo-form of hGGT1 (rmsd deviation for 519 CA atoms=0.273, for 338 CA atoms of large subunit=0.258 and for 181 CA atoms of small subunit=0.273), although shifts between these two structures were larger and included some additional regions compared to already described complex structures. The main shift was detected in the oxyanion forming loop. It was moved up and back compared to its position in apo-enzyme structure resulting in displacement of 1.61 Å for the CA atom of Gly474, 1.30 Å for the CA atom of Gly 473 and 1.55 Å for CA atom of Thr475. The additional shifts detected included helices α6, α7, α8, α15 and α16, C-terminal part of α13, n-terminal part of α14, β6 strand and residues connecting helices α13 and α14 (residues 300-305).

SUMMARY

Five new structures are reported in this study. The structure of the apo-enzyme provided insight into flexibility within the active site of the enzyme. The glutamate-bound structure revealed novel information on the orientation of the side chain of Thr381, displacement of the main chain atoms that form the oxyanion hole and movement of the lid loop region upon glutamate binding. The structure of the GGsTop-bound hGGT1 showed its interactions with the enzyme and provided data which showed that the neutral phosphonate diesters that are more potent inhibitors than mono-anionic phosphonates predominantly due to their higher electrophilicity towards nucleophilic substitution. The DON-bound hGGT1 structure revealed a novel mechanism of DON binding following a reaction of DON with visible light. Finally, serine-borate hGGT1 confirmed the hypothesis that serine in the presence of borate buffer forms a serine-borate complex that occupies the active site of enzyme resulting in an enzyme-inhibitor complex that replicates the tetrahedral transition state of the enzyme. These structures show the changes within the conformation of the active site as the enzyme progresses from the free enzyme to its substrate-bound form and finally to its transition state. Among these structures are the first structures for any eukaryotic GGT in which a molecule in the active site forms a covalent bond with the catalytic Thr381. These data provide new insights into the mechanism of GGT- catalyzed reactions and will be invaluable in the development of new classes of GGT inhibitors.

ACKNOWLEDGEMENTS Research reported in this publication was supported in part by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health (NIH) under grant number P20GM103640. Funding was also provided by insert other investigator’s funding X-ray diffraction data were collected at the National Synchrotron Light Source for which financial support comes principally from the US Department of Energy Offices of Biological and Environmental Research and of Basic Energy Sciences and from NIH grants P41RR012408 and P41GM103473.(Bob Sweet is checking these grant numbers)

Insert acknowledgement for Stanford Synchrotron Radiation Light Source

TABLE 1 Data Collection and Refinement Statistics

Name Apo Glu GGsTop DON SerineBorate

Unit cell (Å) 105.58 105.24 105.70 105.51127.3 105.48 125.29 125.23 123.55 4 103.07 122.00 103.99 104.06 104.19 103.62 Resolution (Å) 50-2.3 (2.38-2.3) 50-2.1 (2.14-2.1) 47-2.18 50-2.2 (2.24-2.2) 50-2.1 (2.14- (2.22-2.18) 2.1) N reflections 28718(1892) 37862(1046) 34573 (1490) 34792(1661) 38155 (1511) Completeness (%) 93.1(62.4) 93 (52.4) 96.6 (84.1) 99.7(97.6) 95.5 (77) Redundancy 5.4(3.3) 10.5 (2.6) 5.8(3.7) 5.5(4.0) 6.3 (4.8) I/σ 18.9(4.2) 21.9(2) 24(2.7) 12.6(2.25) 9 (1.7)

Rmerge(%) 8.9 (31.9) 9.5(45.2) 7.6(37.4) 11.5(64.0) 11.1 (57.0)

Refinement Resolution high(Å) 2.20 2.1 (2.14-2.1) 2.18 (2.23- 2.2 (2.26-2.2) 2.1 (2.15-2.1) (2.26-2.2) 2.18) N reflections Work set 29189(987) 35939(1490) 32828(2035) 33034(1912) 36179 (2090) Rfree set 1556 (54) 1901(61) 1715(99) 1738(95) 1939 (113) overall 30745(1041) 34543(2134) 34772(2007) 38118 (2203) N of atoms 4645 4687 4562 4549 4528 Rwork (%) 14.43(20.5) 14.28 (23.7) 16.05 (27.6) 15.74 (34.8) 17.50(28.9) Rfree(%) 19.6 (28.3) 18.13 (24.7) 21.87 (33.7) 21.26(36.3) 21.97 ( 31.4) ???? 14.7 14.48 16.33 15.74 17.72 Fig. of merit 85.35 88.28 83.13 82.44 76.52 Correl coeff. 97.2 97.1 96.08 97.2 95.8 Estim coord err 0.125 0.105 0.127 0.132 0.137 (likel) Estim B value 5.06 4.27 5.074 5.376 5.67 error RMS from ideal values Bonds 0.011 0.011 0.010 0.013 0.010 angles 1.473 1.43 1.41 1.56 1.47

B factor 41.00(31.7) 35.15(23.09) 40.63 (32.08) 44.83(38.5) 28.75(18.81)

Table 2. Melting temperatures from thermofluor graphs for deglycosylated hGGT1

Buffers\ protein hGGT1 hGGT1_GGsTop hGGT1-DON A1. NaAc pH5.5 53 73 69 A2. BisTris pH5 55 70.5 70 A3. Na Citrate pH6.7 58 75 71 A4. Tris-HCl, pH7.0 56 74 71 A5. Tris-HCl, pH7.9 54 74 71 A6. Tris-HCl, pH8.5 53 74 - A7. 2M K/NaPO4 pH7.0 55 75 71 A8. Hepes pH6.5 56 75 71 A9. Hepes pH7.0 57 75 72 B1. Hepes pH8.6 53 - - B2. 0.5M CAPS, pH10.2 53 73 - B3. pH7.5 55 74 - B4. Imidazole pH8.5 54 73 - B5. Bicine pH9.0 53 72 - B6. Bicine pH10.0 51 71 -

FIGURE LEGENDS

Figure 1. Chemical structure of gmma-glutamyl bond (arrow) and the inhibitors co-crystallized with hGGT1.

Figure 2. Ribbon presentation of Apo-structure. Large subunit is in blue and small subunit is in green. N and C terminals of both subunits are marked.

Figure 3. Glutamate bound structure: A) Stereo presentation of the model of Glu fitted into Fo- Fc density map calculated after initial rigid body refinement. Carbon atoms of the enzyme are in yellow and of the glutamate in orange, oxygens are in red and nitrogens in blue. B) Stereo view of solvent accessible surface of the active site of hGGT1_Glu complex with a model of glutamate. Orientation of the GGT molecule and colors of atoms are the same as in A. C) Schematic representation of interactions of glutamate molecule in the active site of corresponding complex.

Figure 4. LSQ superposition of Cα atoms of Apo, original (4GDX) and new glutamate bound structures. The apo enzyme is red, the original structure is green and new glutamate bound structure is blue. The numbers show the first and last residues of corresponding structure. Figure 5. GGsTop bound hGGT: A) Molecule of GGsTop, B) Stereo presentation of the model of cleaved GGsTop fitted into 2Fo-Fc density map. Atoms are colored as in Fig 3. C) Schematic representation of interactions of cleaved GGsTop in the active site of corresponding complex.

Fig. 6 DON-bound hGGT A) Molecule of DON, B) Stereo presentation of the model of DON with removed diazo nitrogen atoms fitted into initial Fo-Fc density map. Atoms are colored as in Fig. 3. C) Schematic representation of interactions of DON in the active site of corresponding complex.

Fig. 7 Serine-borate bound hGGT: A) Molecule of Serine Borate, B) Stereo presentation of the model of Serine Borate fitted into initial Fo-Fc density map. Atoms are colored as in Fig. 3. C) Schematic representation of interactions of Serine Borate in the active site of corresponding complex.

Suggested reviewers – Joe Barycki, Rebecca Hughey (be sure and cite their refs)

REFERENCES

Ahluwalia, G. S., J. L. Grem, Z. Hao and D. A. Cooney (1990). " and action of amino acid analog anti-cancer agents." Pharmacol Ther 46(2): 243-271.

Allison, R. D. (1985). "gamma-Glutamyl transpeptidase: kinetics and mechanism." Methods Enzymol 113: 419-437.

Anders, M. W. and W. Dekant (1998). "Glutathione-dependent bioactivation of haloalkenes." Annu Rev Pharmacol Toxicol 38: 501-537.

Behrman, E. J., M. J. Biallas, H. J. Brass, J. O. Edwards and M. Isaks (1970). "Reactions of Phosphonic Acid Esters with Nucleophiles .1. ." Journal of 35(9): 3063-&.

Boanca, G., A. Sand, T. Okada, H. Suzuki, H. Kumagai, K. Fukuyama and J. J. Barycki (2007). "Autoprocessing of Helicobacter pylori gamma-glutamyltranspeptidase leads to the formation of a threonine-threonine catalytic dyad." J Biol Chem 282(1): 534-541.

Curthoys, N. P. and R. P. Hughey (1979). "Characterization and physiological function of rat renal gamma-glutamyltranspeptidase." Enzyme 24(6): 383-403.

Emsley, P. and K. Cowtan (2004). "Coot: model-building tools for molecular graphics." Acta Crystallogr D Biol Crystallogr 60(Pt 12 Pt 1): 2126-2132.

Han, L., J. Hiratake, A. Kamiyama and K. Sakata (2007). "Design, synthesis, and evaluation of gamma-phosphono diester analogues of glutamate as highly potent inhibitors and active site probes of gamma-glutamyl transpeptidase." Biochemistry 46(5): 1432-1447.

Han, L., J. Hiratake, N. Tachi, H. Suzuki, H. Kumagai and K. Sakata (2006). "Gamma- (monophenyl)phosphono glutamate analogues as mechanism-based inhibitors of gamma- glutamyl transpeptidase." Bioorg Med Chem 14(17): 6043-6054.

Hanigan, M. H. and H. F. Frierson, Jr. (1996). "Immunohistochemical detection of gamma- glutamyl transpeptidase in normal human tissue." J Histochem Cytochem 44(10): 1101-1108.

Hanigan, M. H., B. C. Gallagher, D. M. Townsend and V. Gabarra (1999). "Gamma-glutamyl transpeptidase accelerates tumor growth and increases the resistance of tumors to cisplatin in vivo." Carcinogenesis 20(4): 553-559. Hartman, S. C. and T. F. McGrath (1973). "Glutaminase A of . Reactions with the substrate analogue, 6-diazo-5-oxonorleucine." J Biol Chem 248(24): 8506-8510.

Hogg, N., R. J. Singh, E. Konorev, J. Joseph and B. Kalyanaraman (1997). "S- Nitrosoglutathione as a substrate for gamma-glutamyl transpeptidase." Biochem J 323 ( Pt 2): 477-481.

Inoue, M., J. Hiratake, H. Suzuki, H. Kumagai and K. Sakata (2000). "Identification of catalytic nucleophile of Escherichia coli gamma-glutamyltranspeptidase by gamma- monofluorophosphono derivative of : N-terminal thr-391 in small subunit is the nucleophile." Biochemistry 39(26): 7764-7771.

Inoue, M., S. Horiuchi and Y. Morino (1977). "Affinity labeling of rat-kidney gamma-glutamyl transpeptidase." Eur J Biochem 73(2): 335-342.

Lherbet, C. and J. W. Keillor (2004). "Probing the stereochemistry of the active site of gamma- glutamyl transpeptidase using derivatives of l-glutamic acid." Org Biomol Chem 2(2): 238- 245.

Lieberman, M. W., A. L. Wiseman, Z. Z. Shi, B. Z. Carter, R. Barrios, C. N. Ou, P. Chevez- Barrios, Y. Wang, G. M. Habib, J. C. Goodman, S. L. Huang, R. M. Lebovitz and M. M. Matzuk (1996). "Growth retardation and cysteine deficiency in gamma-glutamyl transpeptidase-deficient mice." Proc Natl Acad Sci U S A 93(15): 7923-7926.

Murshudov, G. N., A. A. Vagin and E. J. Dodson (1997). "Refinement of macromolecular structures by the maximum-likelihood method." Acta Crystallogr D Biol Crystallogr 53(Pt 3): 240- 255.

Okada, T., H. Suzuki, K. Wada, H. Kumagai and K. Fukuyama (2006). "Crystal structures of gamma-glutamyltranspeptidase from Escherichia coli, a key enzyme in glutathione metabolism, and its reaction intermediate." Proc Natl Acad Sci U S A 103(17): 6471-6476.

Otwinowski, Z. and W. Minor (1997). "Processing of X-ray diffraction data collected in oscillation mode." Macromolecular Crystallography, Pt A 276: 307-326.

Pompella, A., V. De Tata, A. Paolicchi and F. Zunino (2006). "Expression of gamma- glutamyltransferase in cancer cells and its significance in drug resistance." Biochem Pharmacol 71(3): 231-238. Revel, J. P. and E. G. Ball (1959). "The reaction of glutathione with amino acids and related compounds as catalyzed by gamma-glutamyl transpeptidase." J Biol Chem 234(3): 577-582.

Shafer, J., Baronows.P, R. Laursen, F. Finn and Westheim.Fh (1966). "Products from Photolysis of Diazoacetyl Chymotrypsin." Journal of Biological Chemistry 241(2): 421-&.

Tate, S. S. and A. Meister (1977). "Affinity labeling of gamma-glutamyl transpeptidase and location of the gamma-glutamyl binding site on the light subunit." Proc Natl Acad Sci U S A 74(3): 931-935.

Tate, S. S. and A. Meister (1978). "Serine-borate complex as a transition-state inhibitor of gamma-glutamyl transpeptidase." Proc Natl Acad Sci U S A 75(10): 4806-4809.

Townsend, D. M., M. Deng, L. Zhang, M. G. Lapus and M. H. Hanigan (2003). "Metabolism of Cisplatin to a nephrotoxin in cells." J Am Soc Nephrol 14(1): 1-10.

Tuzova, M., J. C. Jean, R. P. Hughey, L. A. Brown, W. W. Cruikshank, J. Hiratake and M. Joyce-Brady (2014). "Inhibiting lung lining fluid glutathione metabolism with GGsTop as a novel treatment for asthma." Front Pharmacol 5: 179.

Wada, K., J. Hiratake, M. Irie, T. Okada, C. Yamada, H. Kumagai, H. Suzuki and K. Fukuyama (2008). "Crystal structures of Escherichia coli gamma-glutamyltranspeptidase in complex with azaserine and acivicin: novel mechanistic implication for inhibition by antagonists." J Mol Biol 380(2): 361-372.

West, M. B., Y. Chen, S. Wickham, A. Heroux, K. Cahill, M. H. Hanigan and B. H. Mooers (2013). "Novel insights into eukaryotic gamma-glutamyl transpeptidase 1 from the crystal structure of the glutamate-bound human enzyme." J Biol Chem 288(44): 31902-31913.

Wickham, S., M. B. West, P. F. Cook and M. H. Hanigan (2011). "Gamma-glutamyl compounds: Substrate specificity of gamma-glutamyl transpeptidase ." Anal Biochem 414: 208-214.

Yamamoto, S., B. Watanabe, J. Hiratake, R. Tanaka, M. Ohkita and Y. Matsumura (2011). "Preventive effect of GGsTop, a novel and selective gamma-glutamyl transpeptidase inhibitor, on ischemia/reperfusion-induced renal injury in rats." J Pharmacol Exp Ther 339(3): 945-951.

Zumbrunn, A., S. Stone and E. Shaw (1988). "Synthesis and properties of Cbz-Phe-Arg-CHN2 (benzyloxycarbonylphenylalanylarginyldiazomethane) as a proteinase inhibitor." Biochem J 250(2): 621-623.