Baxter 2006 - Paper

active site formation in an enzyme?3A Trojan horse transition state analogue generated by MgF H. Williams Wolfgang Bermel, G. Michael Blackburn, Florian Hollfelder, Jonathan P. Waltho, and Nicholas Nicola J. Baxter, Luis F. Olguin, Marko Golicnik, Guoqiang Feng, Andrea M. Hounslow, doi:10.1073/pnas.0604448103 2006;103;14732-14737; originally published online Sep 21, 2006; PNAS This information is current as of November 2006. & Services Online Information www.pnas.org/cgi/content/full/103/40/14732 etc., can be found at: High-resolution figures, a citation map, links to PubMed and Google Scholar, Supplementary Material www.pnas.org/cgi/content/full/0604448103/DC1 Supplementary material can be found at: References www.pnas.org/cgi/content/full/103/40/14732#BIBL This article cites 35 articles, 11 of which you can access for free at: www.pnas.org/cgi/content/full/103/40/14732#otherarticles This article has been cited by other articles: E-mail Alerts . click hereat the top right corner of the article or Receive free email alerts when new articles cite this article - sign up in the box Rights & Permissions www.pnas.org/misc/rightperm.shtml To reproduce this article in part (figures, tables) or in entirety, see: Reprints www.pnas.org/misc/reprints.shtml To order reprints, see: Notes: A Trojan horse transition state analogue generated by MgF 3 H11546 formation in an enzyme active site Nicola J. Baxter ? , Luis F. Olguin ? , Marko Golic?nik ?§ , Guoqiang Feng ¶ , Andrea M. Hounslow ? , Wolfgang Bermel H20648 , G. Michael Blackburn ¶ , Florian Hollfelder ??? , Jonathan P. Waltho ?,?? , and Nicholas H. Williams ¶?? ? Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield S10 2TN, United Kingdom; ? Department of Biochemistry, University of Cambridge, Cambridge CB2 1GA, United Kingdom; ¶ Centre for Chemical Biology, Department of Chemistry, University of Sheffield, Sheffield S3 7HF, United Kingdom; and H20648 Bruker BioSpin GmbH, Silberstreifen 4, 76287 Rheinstetten, Germany Edited by Perry A. Frey, University of Wisconsin, Madison, WI, and approved July 26, 2006 (received for review May 30, 2006) Identifying how enzymes stabilize high-energy species along the reaction pathway is central to explaining their enormous rate acceleration. H9252-Phosphoglucomutase catalyses the isomerization of H9252-glucose-1-phosphate to H9252-glucose-6-phosphate and appeared to be unique in its ability to stabilize a high-energy pentacoordi- nate phosphorane intermediate sufficiently to be directly observ- able in the enzyme active site. Using 19 F-NMR and kinetic analysis, we report that the complex that forms is not the postulated high-energy reaction intermediate, but a deceptively similar tran- sition state analogue in which MgF 3 H11546 mimics the transferring PO 3 H11546 moiety. Here we present a detailed characterization of the metal ion?fluoride complex bound to the enzyme active site in solution, which reveals the molecular mechanism for fluoride inhibition of H9252-phosphoglucomutase. This NMR methodology has a general application in identifying specific interactions between fluoride complexes and proteins and resolving structural assignments that are indistinguishable by x-ray crystallography. enzyme mechanism H20841 fluoride inhibition H20841 NMR structure H20841 phosphoryl transfer H20841 isosteric isoelectronic H20841 transition state analogue P hosphate transfer reactions play a central role in metabolism, regulation, energy housekeeping and signaling (1). As phos- phate esters are kinetically extremely stable, efficient catalysis is crucial for the control of these cellular processes. Although model studies have taught us much about the intrinsic chemical mechanisms (2), our understanding of the origins of the enor- mous enzymatic rate accelerations involved, up to a factor of 10 21 (3), is far from complete (4). A snapshot of an enzyme in a high-energy state would be immensely useful, as it would allow the very interactions that bring about catalysis to be observed (5). However, is this realistic given how elusive high-energy intermediates and transition states (TSs) inevitably are? The direct observation of TSs for simple organic reactions has required ultrafast lasers with femtosecond resolution (6) and no physical or spectroscopic method is available to observe the structure of TSs of enzymatic reactions directly. Thus transition state analogues that bind tightly in an enzyme active site have been of paramount importance in defining the structural and energetic framework for catalysis (7,8). An observation that appears to challenge this paradigm arises from structural studies with H9252-phosphoglucomutase (H9252-PGM, EC 5.4.2.6): namely, that a high-energy phosphorane on the reaction pathway has been observed directly by x-ray crystallog- raphy, demonstrating how the enzyme interacts with a very high-energy, metastable species (9). The latter also apparently demonstrated that the enzyme catalyzed reaction proceeds through an addition?elimination mechanism, a reaction pathway not observed in solution for phosphate monoester anions. How- ever, the observation of an enzyme ?caught in the act? is surprising: the demands of turnover mean that the enzyme would gain no apparent advantage in evolving to stabilize such a high energy intermediate to the extent that it is more stable than the enzyme?product complex. The evolutionary driving force for catalysis is to stabilize the transition states for the formation and breakdown of any such intermediate, and very short lifetimes are expected for any high-energy intermediate. H9252-PGM is a member of the haloalkanoic dehalogenase super- family, and catalyses the relocation of a phosphate group on glucose in a ping-pong mechanism as shown in Fig. 1 (10). The physiological substrate H9252-glucose-1-phosphate (H9252-G1P) accepts a phosphate group from the phosphoenzyme (H9252-PGM*) to give the intermediate, H9252-glucose-1,6-bisphosphate (H9252-G16BP), which is dephosphorylated to yield the final product, glucose-6- phosphate (G6P). This reaction has an equilibrium constant of 28, which lies in favor of G6P (11). The active site cleft, containing a coordinated magnesium ion cofactor, is formed at the interface of the helical cap domain and the H9251H20862H9252 core domain. If H9252-PGM is crystallized in the presence of H9252-G1P or G6P a pentacoordinate species can be observed by x-ray analysis, and it is this that has been interpreted as a high-energy metastable intermediate (INT) on the reaction pathway, closely related to the TS (9). However, this interpretation is not unique. The electron density map is also consistent with a TS analogue (TSA) formed from a five-coordinate magnesium surrounded by two oxygen and three fluoride ligands (MgF 3 ?TSA) that mimics the phos- phoryl group in flight (12, 13). The crystallization buffer needs to contain magnesium and fluoride ions for crystals to form (14, 15), so all of the components necessary for the formation of the MgF 3 ?TSA are available. Furthermore, ab initio quantum- mechanical calculations indicate that MgF 3 ?TSA is stable, whereas INT is not and spontaneously rearranges to a phosphate monoester covalently bound to either the substrate or the enzyme, as would be expected of a transition state (16). This reassignment would account for the long bond lengths around the central atom (they are typical coordinate bond lengths) and would also remove the requirement that the hitherto unobserved trianionic phosphorane INT be stable for several days at 18°C for crystallization (9, 17). We have probed the identity of the pentacoordinate species by using NMR methodology and kinetic Author contributions: N.J.B., L.F.O., and M.G. contributed equally to this work; N.J.B., L.F.O., M.G., A.M.H., G.M.B., F.H., J.P.W., and N.H.W. designed research; N.J.B., L.F.O., M.G., A.M.H., W.B., J.P.W., and N.H.W. performed research; G.F. and W.B. contributed new reagentsH20862analytic tools; N.J.B., L.F.O., M.G., F.H., J.P.W., and N.H.W. analyzed data; and N.J.B., L.F.O., M.G., G.M.B., F.H., J.P.W., and N.H.W. wrote the paper. The authors declare no conflict of interest. This paper was submitted directly (Track II) to the PNAS office. Abbreviations: TS, transition state; TSA, TS analogue; H9252-PGM, H9252-phosphoglucomutase; H9252-G1P, H9252-glucose-1-phosphate; H9252-G16BP, H9252-glucose-1,6-bisphosphate; G6P, glucose-6- phosphate; INT, intermediate. Data deposition: The NMR chemical shifts have been deposited in the BioMagResBank, www.bmrb.wisc.edu (accession nos. 7234 and 7235). § Present address: Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, SI-1000 Ljubljana, Slovenia. ?? To whom correspondence may be addressed. E-mail: fh111@cam.ac.uk, j.waltho@ sheffield.ac.uk, or n.h.williams@sheffield.ac.uk. © 2006 by The National Academy of Sciences of the USA 14732?14737 H20841 PNAS H20841 October 3, 2006 H20841 vol. 103 H20841 no. 40 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.0604448103 analysis to establish whether MgF 3 H11002 , which cannot readily be observed in solution, can form in an enzyme active site as an isoelectronic and isosteric mimic of PO 3 H11002 . Results We expressed and purified H9252-PGM according to published procedures (9, 17, 18). Like the related H9251-PGM enzyme (19), H9252-PGM had been postulated to be in its phosphorylated form (H9252-PGM*) when isolated (18), but electrospray mass- spectrometric analysis only shows a peak at 24,207.8 H11006 1.8 Da corresponding to the predicted mass of unphosphorylated H9252-PGM (Supporting Text and Fig. 6, which are published as supporting information on the PNAS web site). This is to be expected from the estimated half-life of a carboxyl phosphate (t 1/2 H11015 20 h) for acetylphosphate (20) and is consistent with more recent data on H9252-PGM (10), but differs from the original report (9). However, H9252-PGM also exhibits phosphodismutase activity so the enzyme does not need to be initially phosphorylated nor require the addition of the bisphosphorylated intermediate for mutase activity. Phosphorylation of H9252-PGM by the substrate itself can generate H9252-PGM* (10) to initiate the catalytic cycle. Initial NMR experiments confirmed the lack of phosphate in the native enzyme by showing no signal for phosphorus in the 31 P NMR spectrum. We then added H9252-G1P (5 mM) to H9252-PGM (1 mM) to see whether we could directly detect the proposed phosphorane intermediate using 31 P NMR. However, we only observed full turnover of H9252-G1P (H9254 31 P 3.00 ppm) into G6P (H9254 31 P 4.95 and 5.02 ppm for H9251 and H9252 anomers, respectively) within the dead time of the experiment followed by slower hydrolysis over a few hours to leave inorganic phosphate (H9254 31 P 2.62 ppm) as the final product. Hence, there is no evidence for an accumulation of an enzyme bound phosphorane in solution. On the other hand, in the presence of fluoride and G6P, a new 31 P resonance corresponding to a phosphate was observed (H9254 31 P 5.75 ppm). This signal remained unchanged over at least 2 months, and possessed a 1:1 stoichiometry to the enzyme concentration and a line-width appropriate for an enzyme-bound ligand. We then used 19 F NMR spectroscopy to establish whether the pentacoordinate species could be MgF 3 ?TSA, i.e., whether fluoride is present in the G6PH20862H9252-PGM complex. This approach allows us to distinguish MgF 3 H11002 from the central PO 3 H11002 moiety in INT unambiguously. When H9252-PGM is added to a solution containing MgCl 2 and a 10-fold excess of NH 4 F, 19 F resonances are observed for free fluoride and for MgF H11001 in solution. As soon as either H9252-G1P or G6P are added, three intense new 19 F resonances appear (Fig. 2), each with a 1:1 stoichiometry with the enzyme concentration. When substoichiometric concentra- tions of fluoride are added into the G6PH20862H9252-PGM system, the three 19 F resonances appear simultaneously in a 1:1:1 ratio. The appearance of these resonances correlates with a substantial conformational change in the protein, as shown by changes in chemical shift of the backbone amide 1 H and 15 Nin2D 15 N-TROSY spectra (Fig. 3), corresponding to a transition from an open to a closed conformation (10). NMR resonance assign- ment of uniformly 2 H, 13 C, 15 N-labeled H9252-PGM both in the open conformation and in the G6PH20862H9252-PGM complex was achieved by using methods described in ref. 21. The regions of H9252-PGM Fig. 1. The reaction mechanism of H9252-PGM and the potential enzyme stabilized species. (a) H9252-PGM catalyses the interconversion of H9252-glucose-1-phosphate (H9252-G1P) and H9252-glucose-6-phosphate (G6P) via H9252-glucose-1,6-bisphosphate (H9252-G16BP) and the phosphorylated form of the enzyme (H9252-PGM*). (b) The TS for the latter part of the H9252-PGM catalyzed reaction (i.e., H9252-G16BP to G6P), the proposed phosphorane INT and the TSA (MgF 3 ?TSA) formed from H9252-G6P, magnesium, and fluoride. Fig. 2. 19 F NMR spectra of the MgF 3 H11002 H20862H9252-PGM system. (a) The peaks labeled F A-C (H11002147, H11002152, and H11002159 ppm) correspond to MgF 3 H11002 bound to H9252-PGM in the presence of G6P. [NH 4 F]H1100510 mM, [MgCl 2 ]H110055 mM, [H9252-PGM]H110051 mM, [G6P]H11005 5 mM. Conditions: pH 7.2, [K H11001 Hepes] H11005 50 mM, 5°C. (b) Control with G6P omitted. (c) Control with G6P and H9252-PGM omitted. The peak at H11002119 ppm is free fluoride in solution (F H11002 ), and the peak atH11002156 ppm is MgF H11001 . The three 19 F resonances (H11002134, H11002135, and H11002154 ppm), with H1101510% of the intensity of the major peaks, most likely correspond to a minor conformer of the MgF 3 ?TSA complex that exchanges with the major conformer more rapidly than the complex dissociates, because these resonances correlate via saturation trans- fer with specific resonances of the major conformer (resonances F A ,F B , and F C , respectively). Baxter et al. PNAS H20841 October 3, 2006 H20841 vol. 103 H20841 no. 40 H20841 14733 BIOCHEMISTRY CHEMISTRY exhibiting the largest chemical shift changes correspond to those residues which comprise the active site loops, or else residues located in the ??hinge?? region between the ??cap?? and ??core?? domains (10). These NMR observations are consistent with the suggestion that these domains move as rigid bodies relative to each other to enclose the substrate (10). The new 19 F resonances observed indicate that there are three distinct sites for fluoride binding in the G6PH20862H9252-PGM complex. Each site is occupied with a lifetime well in excess of 10 s, as determined by the lack of saturation transfer between 19 F resonances after selective irradiation of each resonance. The absence of exchange on this time scale with free fluoride in solution and with each other allows these resonances to be used as spectroscopic probes to examine their individual interactions with the protein. Selective irradiation of each fluoride resonance in turn estab- lished the existence of different { 19 F} 1 H-NOE distributions between the three enzyme-bound fluorine nuclei and the protein backbone amide protons. Resolution of the amide proton res- onances involved in these NOEs was achieved according to the frequency of their attached 15 N nuclei using selective { 19 F} 1 H, 15 N-HSQC spectra of uniformly 2 H, 15 N-labeled H9252-PGM in complex with G6P and fluoride. In total, 11 NOEs were identified between amide protons and the enzyme bound 19 F nuclei (Table 1, which is published as supporting information on the PNAS web site). These were used as restraints to move three fluoride ions into the closed conformation of H9252-PGM (Protein Data Bank ID code 1O08) from random starting coordinates by using standard solution structure determination procedures within the program CNS (22). The resulting positional distribu- tion of fluoride ions (Fig. 4) places them as predicted for the MgF 3 ?TSA interpretation of the x-ray structural data (9, 12). The new enzyme-bound 31 P resonance therefore arises from the phosphate group of G6P when bound to the enzyme (Fig. 4). The combined NMR data show clearly that in the presence of millimolar concentrations of fluoride, the major species present in solution on addition of substrate or product to H9252-PGM is MgF 3 ?TSA. The trigonal-planar MgF 3 H11002 moiety in MgF 3 ?TSA serves as a mimic for PO 3 H11002 ; if this is a good model of the TS, the enzyme should be efficiently inhibited. Fig. 5 shows that under the crystallization (9, 17) and NMR conditions, the reaction is clearly inhibited by fluoride with an apparent K i in the low mM range. This observation appears to contradict previous reports that H9252-PGM is not inhibited by added fluoride (14, 15). We note Fig. 3. Histogram of backbone amide proton chemical shift changes (H9004H9254) plotted against residue between H9252-PGM in the open form and H9252-PGM in MgF 3 ?TSA. Positive H9004H9254 represent up-field changes for the open to closed transition. Large H9004H9254 are expected in regions of H9252-PGM involved in binding substrate in the active site. However, no data were obtained for residues in the active site loops (D8?T16, L44?L53, S114?N118, V141?A142, and S171?Q172) in the open conformation, because the corresponding 15 N-TROSY peaks were broadened beyond the limits of detection. This line-broadening behavior is indicative of a conformational dynamics process between two (or more) similarly populated forms, and the difference in 1 H chemical shift of H110152 ppm between these interconverting conformations equates to conformational dynamics occurring in the millisecond timescale (i.e., dynamics in the inter- mediate exchange regime for 1 H). These residues are depicted with open bars. Further significantH9004H9254involve residues A17?Q43 positioned in twoH9251-helices of the ??cap?? domain and N77?S88, which locate to the C-terminal portion of the S65?I84 H9251-helix and the ??hinge?? region (Q85?Y93). For the remainder of H9252-PGM, small H9004H9254 indicate that, on formation of MgF 3 ?TSA, there is little change in the local protein fold outside of these regions. Fig. 4. The structure of the active site in MgF 3 ?TSA. Positions of fluoride bound to the enzyme were docked according to the 11 NOEs to the backbone amide protons. The protein structure (PDB ID code 1O08) was used as a template, and the fluorides were assigned zero van der Waals radii during their movement so that they could locate the optimum positions in the structure based solely on the NOE restraints. (a) The pale blue spheres show the results of 50 separate minimizations. (b) The active site of MgF 3 ?TSA. In a and b, the magnesium ion essential for catalysis is on the left. 14734 H20841 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.0604448103 Baxter et al. that kinetic assays of H9251- and H9252-PGM are usually run in the presence of either H9251-orH9252-G16BP, which accelerate the reaction by saturating the enzyme at the intermediate stage or maintain- ing its activated, phosphorylated state (10, 15, 19). Only H9252-G16BP can contribute by maintaining a high concentration of the H9252-G16BPH20862H9252-PGM complex, but other cofactors can stimu- late the enzyme by acting as phosphorylating agents, and com- mercially available H9251-G16BP is used commonly (10, 23?25). Interestingly H9251-G16BP is not turned over to G6P by H9252-PGM, but is converted to H9251-G1P to generate H9252-PGM* (10). If the cofactors H9251-orH9252-G16BP are used at low concentrations, the enzymatic reaction is still clearly inhibited by fluoride with a similar apparent K i (Fig. 5). However, the published inhibition study was conducted under the unusual conditions of a 2-fold excess of intermediate H9252-G16BP (200 H9262M) to substrate H9252-G1P (100 H9262M) (15). When we repeated our experiments in the presence of high concentrations of H9252-G16BP, we also observed that even 10 mM fluoride did not inhibit the reaction (Fig. 5). Under these conditions, pre-steady-state kinetics with H9252-PGM and H9252-G16BP show a burst of product G6P release followed by steady-state hydrolysis of phosphoenzyme (data not shown), which indicates that high concentrations of H9252-G16BP maintain H9252-PGM in a phosphorylated state. With the enzyme phosphorylated on the nucleophilic aspartate, MgF 3 ?TSA simply cannot form. Discussion All of the structural and kinetic observations described above, and the previous x-ray data, are consistent with the formation of MgF 3 ?TSA when G6P, magnesium, and fluoride are present with H9252-PGM. Reassigning the pentacoordinate species as a transition state analogue necessarily limits the inferences that can be drawn from the structure as the bond lengths and electrostatic contacts will be governed by the intrinsic ground state properties of the analogue, and thus do not necessarily reflect the bond lengths in the TS or potential INT. However, the central MgF 3 H11002 moiety is both isoelectronic and isosteric with metaphosphate PO 3 H11002 and so provides a closer portrayal of the expanded TS for phosphate monoester transfer in solution than aluminum and beryllium fluoride complexes which have been more widely utilized to date (13, 26?28). The intrinsic binding constant of MgF 3 H11002 to the enzyme-substrate complex will be considerably lower than the observed K i as MgF 3 H11002 has a very low formation constant in aqueous solution [MgF H11001 has log H9252 1 H11005 1.7 and MgF 2 has log H9252 2 H11349 3.2 (29); MgF 3 H11002 has not been observed in solution], which means that it is only observable in the enzyme active site. This low solution stability suggests that magnesium trifluoride will be a sensitive probe for identifying active sites that strongly stabilize such trigonal?bipyramidal species with a similar charge distribution, because productive interactions with the enzyme pocket are required for assembly of the transition state analogue. Consistent with this expectation, the simulta- neous appearance of the fluoride resonances in the same ratio as the final complex together with a concomitant transition of H9252-PGM from the open to closed conformation on titration of the G6PH20862H9252-PGM system with fluoride is conclusive evidence that MgF 3 ?TSA assembles cooperatively and is stoichiometric with H9252-PGM concentration. These observations have broad implications for understanding the basis of and potential for fluoride inhibition of enzymes that catalyze phosphoryl transfer in the presence of magnesium ions. Fluoride inhibition of certain enzymes has been linked to its mammalian toxicity and to its therapeutic action in dental health. The growth of Treponema, Actinomyces, Fusobacterium, and Bacteroides species is inhibited by millimolar concentrations of fluoride and this has been linked to fluoride inhibition of acid and alkaline phosphatases (30). The proton translocating en- zyme F o ATPases of oral bacteria are also inhibited by fluoride at millimolar concentration (31). In general, such fluoride inhibition has been attributed to the replacement of an essential water or hydroxide ion in the active complex by fluoride (32), but it is likely that formation of MgF 3 H11002 in the active site may be the true inhibitor in some cases; and it may be particularly important in affecting the most proficient enzymes which have exception- ally strong formal transition state affinities (3, 7). Here, we present a detailed characterization of a metal ion?fluoride complex bound to an enzyme active site in solution, and reveal the molecular mechanism for fluoride inhibition of H9252-PGM. These results show that it is possible for fluoride ions to be located in an enzyme?analogue complex in solution with high structural accuracy. Hence, MgF 3 ?TSA has the potential to allow us to define the solution structure of the active site as it is arranged around a mimic for the transferring phosphoryl moiety. The high susceptibility of 19 F NMR chemical shifts to the local electronic environment provides a sensitive probe of subtle changes within the enzyme that would be invisible to other structural biology methods. These methods can also be applied to the study of other complexes that contain metal?fluoride species as analogues for phosphates, such as aluminum (27, 33) and beryllium (26, 34) fluorides, and will enable the accurate assignment of their identity in future. The potential utility of using fluoride as a probe of enzyme structure is especially powerful when used in combination with an electronically and geometrically accurate transition state mimic for phosphoryl transfer, MgF 3 H11002 , that only exists within the confines of the active site and places the spectroscopically active atoms right into the catalytic cavity. Materials and Methods Expression and Purification of H9252-PGM. The H9252-PGM gene was am- plified from a Lactococcus lactis strain obtained from DSMZ (Braunschweig, Germany) (no. 20481, which is identical to ATCC strain 19435). This strain was grown at 30°C on plates with corynebacterium agar containing casein peptone tryptic digest (10 g), yeast extract (5 g), glucose (5 g), NaCl (5 g), agar (15 g), and distilled water (1,000 ml) adjusted to pH 7.2?7.4. A tooth- pick-scrape of a colony was used as a template in a PCR to Fig. 5. Fluoride inhibition of catalysis by H9252-PGM. The reaction is inhibited by fluoride (0?10 mM) unless a high concentration ofH9252-G16BP is present. Relative initial rates (%) are shown for the conversion of H9252-G1P to G6P. Reactions contained: green circles, [H9252-G1P]H11005250H9262M, [H9252-PGM]H11005200 nM; black triangles, [H9252-G1P]H11005250H9262M, [H9251-G16BP]H1100550H9262M, [H9252-PGM]H110055 nM; red squares, [H9252-G1P]H11005 250 H9262M, [H9252-G16BP] H11005 0.5 H9262M, [H9252-PGM] H11005 5 nM; blue inverted triangles, ([H9252-G1P] H11005 50 H9262M, [H9252-G16BP] H11005 200 H9262M, [H9252-PGM] H11005 5 nM. Baxter et al. PNAS H20841 October 3, 2006 H20841 vol. 103 H20841 no. 40 H20841 14735 BIOCHEMISTRY CHEMISTRY amplify the gene. Primers (5H11032-GAA TTC CAT ATG TTT AAA GCA GTA TTG-3H11032 and 5H11032-CCG CTC GAG TTA TTT TTG CTT TTG AAG-3H11032) were used to introduce a NdeI and a XhoI restriction site at the beginning and end of the gene, respectively. After digestion, the PCR product was cloned in the pET-22b(H11001) expression vector (Novagen, San Diego, CA) giving the expected published sequence (35). Transformed Escherichia coli BL21(DE3) with this plasmid were used to overexpress H9252-PGM. The expression and purification of the protein were carried out exactly as described in ref. 17. Expression and Purification of Labeled H9252-PGM. For 15 N-, 15 NH20862 2 H-, and 15 NH20862 13 CH20862 2 H-labeled protein samples, cells were grown in minimal M9 media as described in ref. 21 except that 15 NH 4 Cl was used in all three samples; glucose was the carbon source for the 15 N-labeled protein, and 1 mM IPTG was used to induce expression. The purification procedure followed was the same as for unlabelled H9252-PGM (17). NMR Methods. The samples prepared for the NMR analysis of MgF 3 ?TSA contained: 1 mM H9252-PGM, 5 mM H9252-G1P or G6P, 5 mM MgCl 2 , and 10 mM NH 4 Fin50mMK H11001 Hepes buffer (pH 7.2) containing 15% volH20862vol D 2 O,2mMNaN 3 , and EDTA-free Com- plete protease inhibitor mixture. The 1D 19 F NMR spectra were recorded at 5°C on a Bruker Avance 500 MHz spectrometer (operating at 470.59 MHz for fluorine) equipped with a 5-mm dual 1 HH20862 19 F probe. The 2D frequency selective { 19 F} 1 H, 15 N-HSQC NOE difference spectra of 2 H, 15 N-labeled H9252-PGM in MgF 3 ?TSA (sample prepared as above except with 20 mM G6P) were acquired at 25°C on a Bruker Avance 600 MHz spectrometer (operating at 564.69 MHz for fluorine) equipped with a 1 HH20862 15 NH20862 19 F probe and z axis gradients. Selective 19 F irradiation was achieved with a continuous wave at a power level of 40 dB applied over the 1-s recycle delay and the solvent signal was minimized with water flip-back pulses. The spectra were acquired as four interleaved 1 H, 15 N-HSQC experiments with defined selective irradiation fre- quencies of H11002147, H11002152, and H11002159 ppm and one at an off- resonance position. Three 2D frequency selective { 19 F} 1 H, 15 N- HSQC spectra were obtained by subtracting the spectra recorded with defined 19 F irradiation frequencies from the off-resonance spectrum. For the backbone resonance assignment of H9252-PGM in MgF 3 ?TSA, NMR spectra were acquired at 25°C on a sample of 2 H, 15 N, 13 C-labeled H9252-PGM (sample prepared as above except with 20 mM G6P) using a Bruker Avance 600 MHz spectrometer equipped with a 5-mm 1 HH20862 15 NH20862 13 CH20862 2 H cryoprobe and pulse-field z-gradients. Backbone resonance assignments for HN, N, C H9251 ,C H9252 , and CH11032 nuclei were obtained from 2D 15 N-TROSY, 3D TROSY ct-HNCA, 3D TROSY ct-HN(CO)CA, 3D TROSY HN(CA)CB, 3D TROSY HN(COCA)CB, 3D TROSY HN(CA)CO, and 3D TROSY HNCO. The data were acquired, processed, and analyzed and backbone sequential assignment was performed as in ref. 21. Backbone sequential assignment of H9252-PGM in the open conforma- tion was performed as for H9252-PGM in MgF 3 ?TSA using a sample containing 1 mM H9252-PGM, 5 mM MgCl 2 , and 10 mM NH 4 Fin50 mM K H11001 Hepes buffer (pH 7.2) with 15% volH20862vol D 2 O,2mMNaN 3 , and EDTA-free Complete protease inhibitor mixture. Proton chemical shifts were referenced relative to the methyl signals of internal DSS at 0.0 ppm. 15 N, 13 C, and 19 F chemical shifts were calculated indirectly by using the following gyromagnetic ratios: 15 NH20862 1 H H11005 0.101329118, 13 CH20862 1 H H11005 0.251449530, and 19 FH20862 1 H H11005 0.940940080. The backbone 1 H, 13 C, and 15 N chemical shifts for H9252-PGM in both the MgF 3 ?TSA complex and in the open confor- mation together with the three 19 F chemical shifts of the MgF 3 H11002 species have been deposited in the BioMagResBank under accession codes 7234 (MgF 3 ?TSA complex) and 7235 (open conformation). H9252-PGM Kinetic Assays. All H9252-PGM assays were carried out in a Molecular Devices (Sunnyvale, CA) Microtitreplate Spectra- max Plus Reader at 25°C in 50 mM K H11001 Hepes (pH 7.2) with 100 H9262l total volume containing 2 mM MgCl 2 and5nMH9252-PGM unless stated otherwise. All substrates were obtained com- mercially from Fluka (St. Louis, MO) and Sigma (St. Louis, MO) except for H9252-G16BP which was synthesized from H9251-D- glucose (Scheme 1, which is published as supporting informa- tion on the PNAS web site). Enzyme concentrations were measured by using H9255 280 H11005 20.6 H9262M H110021 H18528cm H110021 determined by the Biuret method (36) with commercial BSA standards (Sigma) used for the calibration. The appearance of G6P was moni- tored in a coupled assay with glucose-6-phosphate dehydro- genase (obtained from Sigma, 5 unitsH20862ml G6PDH and 0.5 mM NAD H11001 ) following the absorbance at 340 nm. Reactions were initiated by the addition of the substrate. The steady-state rates were calculated from the linear portions of time courses and when less than 20% of substrate was converted. Fluoride inhibition of glucose-6-phosphate dehydrogenase could not be detected below 100 mM in our experimental setup. Any inhibitory effects can therefore be ascribed to effects on H9252-PGM alone. We thank Steve Gamblin and Steve Smerdon for helpful discussion and Tony Kirby for helpful comments on the manuscript. M.G. is a European Union Marie Curie fellow, and L.F.O. holds a Consejo Nacional de Ciencia y Tecnolog?´a (Mexico) predoctoral fellowship. This research was supported by the European Union network ENDIRPRO, the Biotech- nology and Biological Sciences Research Council, and the Medical Research Council. 1. Dzeja PP, Terzic A (2003) J Exp Biol 206:2039?2047. 2. Schramm VL (1998) Annu Rev Biochem 67:693?720. 3. Lad C, Williams NH, Wolfenden R (2003) ProcNatlAcadSciUSA 100:5607?5610. 4. Benkovic SJ, Hammes-Schiffer S (2003) Science 301:1196?1202. 5. Knowles J (2003) Science 299:2002?2003. 6. Zewail AH (2000) Angew Chem Int Ed Engl 39:2586?2631. 7. Wolfenden R (2003) Biophys Chem 105:559?572. 8. Schramm VL (2005) Curr Opin Struct Biol 15:604?613. 9. Lahiri SD, Zhang G, Dunaway-Mariano D, Allen KN (2003) Science 299:2067? 2071. 10. Zhang G, Dai J, Wang L, Dunaway-Mariano D, Tremblay LW, Allen KN (2005) Biochemistry 44:9404?9416. 11. Marechal LR, Belocopitow E (1974) Eur J Biochem 42:45?50. 12. Blackburn GM, Williams NH, Gamblin SJ, Smerdon SJ (2003) Science 301:1184c. 13. Graham DL, Lowe PN, Grime GW, Marsh M, Rittinger K, Smerdon SJ, Gamblin SJ, Eccleston JF (2002) Chem Biol 9:375?381. 14. Allen KN, Dunaway-Mariano D (2003) Science 301:1184d. 15. Tremblay LW, Zhang G, Dai J, Dunaway-Mariano D, Allen KN (2005) JAm Chem Soc 127:5298?5299. 16. Webster CE (2004) J Am Chem Soc 126:6840?6841. 17. Lahiri SD, Zhang G, Radstrom P, Dunaway-Mariano D, Allen KN (2002) Acta Crystallogr D 58:324?326. 18. Lahiri SD, Zhang G, Dunaway-Mariano D, Allen KN (2002) Biochemistry 41:8351?8359. 19. Ray WJ, Peck EJ (1972) Enzymes 6:407?477. 20. Di Sabato G, Jencks WP (1961) J Am Chem Soc 83:4393?4400. 21. Reed MAC, Hounslow AM, Sze KH, Barsukov IG, Hosszu LLP, Clarke AR, Craven CJ, Waltho JP (2003) J Mol Biol 303:1189?1201. 22. Brunger AT, Adams PD, Clore GM, Delano WL, Gros P, Grosse-Kunstleve RW, Jiang J-S, Kuszewski J, Nilges M, Pannu NS, et al. (1998) Acta Crystallogr D 54:905?921. 23. Qian NY, Stanley GA, Hahn-Hagerdal B, Radstrom P (1994) J Bacteriol 176:5304?5311. 24. Mesak LR, Dahl MK (2000) Arch Microbiol 174:256?264. 25. Lahiri SD, Zhang G, Dai J, Dunaway-Mariano D, Allen KN (2004) Biochem- istry 43:2812?2820. 26. Kagawa R, Montgomery MG, Braig K, Leslie AG, Walker JE (2004) EMBO J 23:2734?2744. 27. Yu YW, Morera S, Janin J, Cherfils J (1997) Proc Natl Acad Sci USA 94:3579?3583. 14736 H20841 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.0604448103 Baxter et al. 28. Grigorenko BL, Nemukhin AV, Cachau RE, Topol IA, Burt SK (2005) J Mol Model 11:503?508. 29. Fovet Y, Gal JY (2000) Talanta 53:617?626. 30. Li L (2003) Crit Rev Oral Biol Med 14:100. 31. Ahmad Z, Senior AE (2006) FEBS Lett 580:517?520. 32. Lebioda L, Zhang E, Lewinski K, Brewer JM (1993) Proteins: Struct Funct Genet 16:219. 33. Schlichting I, Reinstein J (1999) Nat Struct Biol 6:721?723. 34. Cho H, Wang W, Kim R, Yokota H, Damo S, Kim SH, Wemmer D, Kustu S, Yan D (2001) Proc Natl Acad Sci USA 98:8525?8530. 35. Qian N, Stanley GA, Bunete A, Radstrom P (1997) Microbiology 143: 855?865. 36. Lowe CR, Thomas JA (1996) in LabFax Enzymology, ed Engel PC (Academic, New York), pp 34?40. Baxter et al. PNAS H20841 October 3, 2006 H20841 vol. 103 H20841 no. 40 H20841 14737 BIOCHEMISTRY CHEMISTRY