Evidence that the fosfomycin target Cys115 in UDP-N-acetylglucosamine enolpyruvyl transferase (MurA) is essential for product release.

MurA (UDP-N-acetylglucosamine enolpyruvyl transferase, EC 2.5.1.7) is an essential enzyme in the biosynthesis of the peptidoglycan layer of the bacterial cell. It provides an attractive template for the design of novel antibiotic drugs and is the target of the naturally occurring antibiotic fosfomycin, which covalently attaches to Cys115 in the active site of the enzyme. Mutations of Cys115 to Asp exist in pathogens such as Mycobacteria or Chlamydia rendering these organisms resistant to fosfomycin. Thus, there is a need for the elucidation of the role of this cysteine in the MurA reaction. We determined the x-ray structure of the C115S mutant of Enterobacter cloacae MurA, which was crystallized in the presence of the substrates of MurA. The structure depicts the product state of the enzyme with enolpyruvyl-UDP-N-acetylglucosamine and inorganic phosphate trapped in the active site. Kinetic analysis revealed that the Cys-to-Ser mutation results in an enzyme that appears to perform a single turnover of the reaction. Opposing the common view of Cys115 as a key residue in the chemical reaction of enolpyruvyl transfer, we now conclude that the wild-type cysteine is essential for product release only. On the basis of a detailed comparison of the product state with the intermediate state and an unliganded state of MurA, we propose that dissociation of the products is an ordered event with inorganic phosphate leaving first. Phosphate departure appears to trigger a suite of conformational changes, which finally leads to opening of the two-domain structure of MurA and the release of the second product enolpyruvyl-UDP-N-acetylglucosamine.

Growth and survival of bacteria relies on the functionality of the enzyme MurA (UDP-N-acetylglucosamine enolpyruvyl transferase, EC 2.5.1.7), which catalyzes the first committed step in the biosynthesis of the bacterial cell wall (1,2). Because this pathway does not occur in mammalian organisms, en-zymes involved in peptidoglycan synthesis are ideal targets for the design of new antibiotic drugs (3)(4)(5)(6). MurA in particular has received attention in the past because the enzyme is the target of the naturally occurring broad spectrum antibiotic fosfomycin (7). Fosfomycin covalently attaches to the thiol group of a cysteine (position 115 in Escherichia coli numbering) in the active site of E. coli and Enterobacter cloacae MurA and thus irreversibly inhibits the enzymatic function (7)(8)(9).
MurA catalyzes the transfer of the enolpyruvyl moiety of phosphoenol pyruvate (PEP) 1 to the 3Ј-hydroxyl group of UDP-N-acetylglucosamine (UNAG) (Fig. 1). This chemically unusual reaction, in which the C-O bond of PEP is cleaved rather than the high energy P-O bond, runs as an addition-elimination process (10). After the substrates have docked into the active site (Fig. 2), adhering to the strict binding order UNAG before PEP, a proton is added to PEP, and the resulting oxocarbenium ion is, as a whole, added to the sugar nucleotide. This addition step of the reaction yields an intermediate with a covalent link between the 3Ј-hydroxyl group of UNAG and C-2 of PEP, which then adopts tetrahedral configuration (11,12). In the elimination step of the reaction a proton is removed from C-3 of the PEP moiety, leading to separation of inorganic phosphate (P i ) from the tetrahedral adduct and reformation of the double bond between C-2 and C-3. The only other enzyme known to catalyze a similar reaction is AroA (also called 5-enolpyruvylshikimate-3-phosphate synthase, EC 2. 5.1.19), which is vital for the shikimate pathway in plants and numerous microorganisms (13).
In addition to their strong mechanistic similarity, MurA and AroA share an unusual protein architecture: a single polypeptide chain folds into two similar globular domains, each built up by 3-fold repetition of the same ␤␣␤␣␤␤ motif (14 -16). Upon binding of UNAG and shikimate 3-phosphate to MurA and AroA, respectively, the two domains approach each other to establish the active site in the interdomain cleft (9,15,17,18). In concert with the domain movement in MurA, a 10-residue loop, Pro 112 -Pro 121 , from the top domain moves toward the active site and closes the interdomain cleft like a lid. This loop hosts Cys 115 in MurA from E. cloacae and E. coli. Cys 115 has been shown to be critical for MurA activity (8,19,20), because mutation of the cysteine to serine or alanine renders the enzyme inactive (20,21). However, aspartate is tolerated in position 115 (20) and is a natural mutation in MurA from Mycobacterium and Clamydiae species (22,23).
It has been suggested that Cys 115 participates in the proton transfer toward and from PEP during catalysis (24). Such a dual role for Cys 115 as an acid-base catalyst in the additionelimination reaction has recently been questioned by crystallographic analysis of the D305A mutant of MurA crystallized in the presence of UNAG and PEP (12). This structure revealed the genuine tetrahedral reaction intermediate of the enzyme, which is S-configured at the C2 of the attached PEP molecule. The S configuration was also found for the genuine tetrahedral intermediate of the AroA reaction using the corresponding mutant enzyme D313A. Bartlett and co-workers (25,26) deduced the same S configuration of the intermediate of the AroA reaction by chemical studies. However, the stereochemical course of the enolpyruvyl transfer reactions is a matter of debate, partly because of the lack of understanding the role of Cys 115 in MurA catalysis and because an active site residue such as Cys 115 is not present in AroA.
In this work, we present evidence that Cys 115 may not be involved in the transfer of the enolpyruvyl moiety from PEP to UNAG but in fact appears to be crucial for product release. We engineered a single-site mutant from E. cloacae MurA, which carries serine instead of the wild-type cysteine in position 115, and crystallized the mutant enzyme in the presence of the substrates of MurA: UNAG and PEP. Subsequent x-ray analysis to 2.3 Å resolution surprisingly revealed the product state of the enzyme with enolpyruvyl-UDP-N-acetylglucosamine (EP-UNAG) and P i trapped in the active site. Kinetic analysis confirmed that this mutant enzyme catalyzes a single turnover only. We discuss the implications from this structure for the catalytic mechanism and the possible role of Cys 115 in the domain movement of MurA associated with catalysis. EXPERIMENTAL PROCEDURES PEP (potassium salt) and UNAG were purchased from Sigma. Protein concentration was determined using the Pierce Coomassie reagent with bovine serum albumin as standard. The mutation C115S was introduced to wild-type E. cloacae MurA using a Kwik Change mutagenesis kit from Stratagene and appropriate primers. The pET-Vector 9d (Novagen) containing the open reading frame of the wild-type enzyme was used as template for the point mutation of MurA. C115S-MurA was overexpressed in a STBL2-DE3 E. coli cell. The overexpressed protein was purified as described (21).
Crystallography-The mutant enzyme was concentrated to ϳ100 mg/ml using Centricon 30 devices (Amicon) at 4°C. C115S-MurA was crystallized at 19°C from 10 mM MES, pH 6.4, 10% (w/v) polyethylene glycol 20,000 in the presence of 5 mM UNAG and 5 mM PEP. Diffraction data were recorded from a single flash frozen crystal on a R-axis IV 2ϩ image plate detector (Molecular Structure Corporation, The Woodlands, TX) using the rotation method (x-rays, CuK␣, focused by mirror optics (Molecular Structure Corporation); generator, RU300 (Rigaku, Molecular Structure Corporation)). Data processing, phasing and refinement, and model building were performed with the program packages HKL suite (27), CNS (28), and O (29), respectively. The structure was solved by molecular replacement, using unliganded MurA (Protein Data Bank code 1EJC) (28) as search model. Solvent molecules and residues 111-124 were omitted in the translational and rotational searches. Diffraction data were limited to low resolution reflections 20.0 -6.0 Å in the cross-rotation and translation search. From the 30 highest peaks of the cross-rotation function, eight solutions could be successfully translated into the asymmetric unit of the crystal. Exploiting the point group symmetry, the eight independent solutions were rearranged in the asymmetric unit. This rearrangement revealed that the asymmetric unit of the crystal contains two tetramers of MurA molecules. Rigid body refinement of the asymmetric octamer, with the two globular domains of each of the eight MurA molecules refined independently, lowered the initial R free from 50.3 to 36.1%. Refinement was performed using data to the highest resolution with no cut-off applied. Most of FIG. 2. Addition-elimination mechanism of MurA. enzyme-X, enzyme-Y, and enzyme-Z stand for side chains involved in the transfer reaction. Cys 115 was proposed to function as enzyme-Y and enzyme-Z in E. coli MurA (24). A proton is added to PEP, yielding a PEP oxocarbenium ion, whereas the target hydroxyl group of the first bound substrate UNAG gets deprotonated (A), followed by nucleophilic attack of the oxyanion on the C-2 atom of the PEP oxocarbenium ion (B), and leads to a covalent adduct of the two substrates, where the C-2 atom of the PEP moiety is in tetrahedral configuration. The elimination of phosphate may proceed through self-catalysis (C1) (26) or through the action of an enzyme residue such as Asp 305 (C2) (12). Either mechanism would result in the formation of the vinyl ether product, EP-UNAG, and P i (D). the protein regions were constrained exploiting the noncrystallographic symmetry between the eight molecules, except for the active site and intermolecular contact regions. Solvent molecules were added to the model at chemically reasonable positions. The ligands were modeled according to the clear electron density map. Residue 67 of each of the eight molecules was modeled as iso-aspartate (30). The data collection and refinement statistics are summarized in Table I Kinetic Analysis-The catalytic activity of E. cloacae C115S MurA was tested at 20°C with a Shimadzu 1650PC spectrophotometer. To measure C115S MurA activity, we made use of two assays; one uses MurB as a coupling enzyme to detect the product enolpyruvyl-UNAG (EP-UNAG). We have recently used the MurB coupled assay for a reevaluation of the mode of action of fosfomycin on E. cloacae MurA (17). The coupled reaction starts with MurA converting UNAG and PEP to UNAG-EP, which is then reduced to UDP-N-acetylmuramic acid by MurB, using one equivalent of NADPH. To compensate for the diaphorase activity of MurB, a glucose oxidase/glucose system was exploited. This resulted in a stable base line prior to initiation of the MurA reaction. The second assay, the Lanzetta assay, was used to detect the second product, P i (35). We have previously utilized the Lanzetta assay to characterize P i release from 5-enolpyruvylshikimate-3-phosphate synthase (36).
The purified C115S mutant MurA contains trace amounts of genomic E. coli MurA, which would interfere with the kinetic analysis, especially using the high concentrations of mutant enzyme used here. Therefore we selectively inactivated the wild-type enzyme using fosfomycin. Inhibition of MurA by fosfomycin is solely due to covalent attachment to Cys 115 , whereas the C115S mutant enzyme is not affected (8). To test C115S MurA activity using the MurB coupled assay, a reaction mixture containing 50 mM Tris, pH 8.0, 50 mM KCl, 2 mM dithiothreitol, 20 mM glucose, 20 units of glucose oxidase, 0.20 mM NADPH, 40 g of MurB, 5 mM UNAG, 10 M fosfomycin, and 1.8 mg of C115S MurA (0.04 mM), in a final volume of 1 ml, was incubated at room temperature for 10 min. A base-line time course was run for ϳ30 s prior to the addition of 5 mM PEP to start the reaction. The decrease in NADPH absorbance was recorded at 340 nm. In the Lanzetta assay the enzyme (0.05 mM) was incubated in 100 l of the same assay mixture for 10 min before the reaction was initiated by the addition of 5 mM PEP. The reaction was stopped by the addition of 800 l of the Lanzetta reagent. Color development was stopped after 5 min by the addition of 100 l of 34% (w/v) sodium citrate. The change in optical density was measured at 660 nm, and the amount of P i was determined by comparison with phosphate standards.
To determine whether the C115S MurA is capable of catalyzing the conversion of UNAG and PEP to UNAG-EP and P i but is unable to release the products (single turnover), several denaturants were used to test whether the trapped products could be detected after denaturation of the enzyme-ligand complex. This was accomplished by incubating a mixture containing reaction buffer and 40 mg/ml C115S MurA (0.9 mM) in a final volume of 400 l at room temperature for 10 min. The reaction was initiated by the addition of 5 mM PEP and allowed to proceed for 5 min. To stop the reaction and denature the mutant enzyme a final concentration of 6 M urea or 5% trichloroacetic acid was added and incubated at 4°C for 16 h. Alternatively, the reaction was heated at 100°C for 10 min. To test for the product UNAG-EP, the coupled assay with varied concentrations of the denaturation mixture (0.4 -2.9 mg/ 0.01-0.07 mM) was employed. A stable reading was established prior to the start of the reaction by the addition of 40 g of MurB, and the decrease in NADPH absorbance was recorded. Control experiments were conducted throughout with C115S MurA treated in the same way but omitting PEP during the reaction time.
Testing the amount of P i produced in the reaction failed because of the noncompatibility of the Lanzetta reagent with high protein concentrations. Notably, the mutant enzyme does not precipitate out of urea solutions and thus cannot be separated from EP-UNAG using centrifugation. Heat and trichloroacetic acid, which denature and precipitate the mutant enzyme, were not suitable; nonreacting PEP hydrolyzes rapidly at elevated temperatures to pyruvate and P i , thereby producing too high background levels, and as low as 0.2% trichloroacetic acid interferes with the Lanzetta reagent.

RESULTS AND DISCUSSION
Enzyme-Products Complex-Under the crystallization conditions used, C115S-MurA oligomerizes to densely packed tetramers (Fig. 3), of which two constitute the asymmetric unit of the crystal. X-ray analysis to 2.3 Å resolution revealed the products of the enzymatic reaction, EP-UNAG and P i , trapped in the active site of each of the eight crystallographically independent C115S-MurA molecules (Fig. 4). The products uniformly occupy all MurA molecules of the crystal. Refined at unit occupancy, and the mean temperature factor of the atoms  of the products lies with 17.0 Å 2 yet below the mean temperature factor of the protein of 23.9 Å 2 .
This liganded C115S-MurA, which in the following is denoted as MurA:P 1 P 2 , exhibits the closed enzyme conformation observed previously for MurA (9,12,24). Its overall fold is virtually identical with that of the tetrahedral reaction intermediate state of the enzyme (MurA:THI; Protein Data Bank entry 1Q3G) (12). The positions of equivalent pairs of the 419 C␣ atoms of MurA:P 1 P 2 and of MurA:THI deviate in average by only 0.23 Å with a maximum discrepancy of 1.7 Å at the C terminus of the polypeptide chain.
Detailed comparison of MurA:THI and MurA:P 1 P 2 suggests that the enzyme cleaves the covalent bond between C-2 and the enol oxygen atom of the PEP moiety without macroscopic changes in the active-site residues of the enzyme (Fig. 4). Residues that are in polar or hydrophobic interaction with the UNAG moiety of the tetrahedral adduct as well as the charged residues Lys 22 , Arg 120 , Asp 305 , Arg 331 , Arg 371 , and Arg 397 , which are responsible for PEP binding (12,15), adopt in MurA: P 1 P 2 the same conformation as in MurA:THI. Even the Pro 112 -Pro 121 loop and especially residue 115 retain their conformation in the transition from the intermediate state to the product state. Although the loop lies in the interface of MurA molecules in the noncrystallographic tetramer (Fig. 3), packing interactions only marginally involve residues of the loop; two hydrogen bonds are formed between the main chain of the loop in residues Leu 111 and Gly 114 to the elongated side chains of Arg 340 and Glu 337 , respectively, from a neighboring molecule.
Single Turnover Catalysis-With serine substituted for the wild-type cysteine in position 115, the enzyme is apparently still able to catalyze the enolpyruvyl transfer reaction but in a single turnover only, because both reaction products, EP-UNAG and P i , are trapped in the active site. Indeed, this mutant enzyme showed no catalytic activity using either the MurB coupled assay (detecting EP-UNAG) or the Lanzetta assay (detecting P i ), which is in agreement with the structure data. EP-UNAG produced by the mutant enzyme was detected only after denaturation of the enzyme with urea, followed by determination of the amount of EP-UNAG released using MurB, the second enzyme of peptidoglycan biosynthesis, which utilizes EP-UNAG as substrate. The product was observed in a close to 1:1 ratio with the C115S concentration (Fig. 5). Quantifying the amount of P i trapped in the active site using the Lanzetta reagent failed because of its incompatibility with high concentrations of protein and/or denaturant (see "Experimen- tal Procedures"). However, given that the generation of either product by the C115S mutant MurA is not detectable, but EP-UNAG is produced stoichiometrically with the mutant enzyme, it appears that both products of the reaction are formed but cannot dissociate from the active site.
Proposed Role of Cys 115 in the MurA Reaction-The present data provide evidence that Cys 115 is crucial for product release. The data also imply that this residue is not involved in the addition step of enolpyruvyl transfer. It is possible, however, that the onset of the reaction is perturbed as a result of the mutation to Ser. Because we have detected the enolpyruvylproduct after a 5-min reaction time, long enough to build up products by an even remotely active enzyme, we cannot rule out other roles that Cys 115 might exert during catalysis. However, two additional findings prompted us to reconsider the function of Cys 115 . First, a residue such as Cys 115 is not present in AroA. Although it has been suggested that Glu 341 , a strictly conserved AroA residue, might be the catalytic counterpart to Cys 115 in MurA (15,26,37), the location of these two residues in their respective active sites is different (12). Secondly, whereas Glu 341 in AroA is in short hydrogen-bonding distance to the PEP-moiety of the tetrahedral intermediate, the only apparent interaction that Cys 115 exerts in the MurA-THI structure is a long range bond (distance ϭ 3.7 Å) to NH 1 of the guanidinium group of Arg 120 (12). Given the nucleophilic nature of Cys 115 when reacting with fosfomycin, this interaction is likely to be electrostatic. In the MurA-products complex the distance between the Ser 115 hydroxyl group and NH 1 of Arg 120 is 3.5 Å (Fig. 6). Modeling a cysteine side chain into the Ser 115 position results in a distance of 3.3 Å between the sulfur anion and the NH 1 of Arg 120 , sufficiently small enough to establish a salt-bridge (38). This interaction probably weakens bonding of the phosphate ion to Arg 120 and appears to be a prerequisite for the dissociation of products from the enzyme. Serine in position 115 cannot fulfill such a role. Thus, the lack of a negative charge in the Ser 115 mutant enzyme may be sufficient to keep the enzyme in its closed form and prevent product release. On the other hand, aspartate as in position 115 of Mycobacterium MurA could substitute for the proposed Cys 115 function (Fig. 6). The side chain of aspartate is longer and flexible enough to bring one of its carboxyl oxygen atoms in close salt bridge distance to NH 1 of Arg 120 . This is reflected in the observation that C115D-MurA from E. coli retains enzymatic activity (20).
Together with Lys 22 and Arg 397 , the guanidinium group of Arg 120 coordinates the phosphate moiety of the tetrahedral intermediate, and after formation of the products, the phosphate ion (Figs. 4 and 6 and Table II). The charged side chain groups form hydrogen bonds or salt bridges to the phosphate oxygen atoms, depending on the protonation state of the phosphate ion. These interactions are principally identical with the coordination of the phosphate group in the tetrahedral intermediate. After cleavage of the covalent bond between the enolpyruvyl and phosphate moieties of the tetrahedral intermediate, the phosphate ion moves 0.6 Å away from its former location and slightly rotates (Fig. 4). Detachment of the phosphate group has been suggested to be either self-catalyzed (syn-elimination) (26) or through the action of a strictly conserved aspartate residue, Asp 305 MurA /Asp 313 AroA (anti-elimination) (12). Although the Pro 112 -Pro 121 loop in the closed enzyme state shields most of the active site from solvent, the phosphate site is situated in a solvent accessible cavity (Fig. 7). This cavity is narrowed toward the surface by the side chain of Arg 397 . Upon opening of the enzyme, the side chain of Arg 397 would swing toward the hinge region into a position parallel to the side chain of Lys 48 (16), whereby the guanidinium group of Arg 397 is displaced by about 9 Å (30). Such a shift of Arg 397 would leave a free exit route for the phosphate ion from the closed enzyme (Fig. 7, b and e). This is different for EP-UNAG. From the active site geometry in the product state, it becomes obvious that EP-UNAG cannot dissociate from MurA as long as the enzyme and the loop are closed. The only direct interaction between the loop and EP-UNAG is a hydrogen bond from the side chain of Arg 120 to an oxygen atom of the di-phosphate moiety (Fig. 4), but the loop sterically blocks dissociation of this product (Fig. 7).
From this scenario it is conceivable that product release is an ordered event, where the phosphate ion has to leave before the Pro 112 -Pro 121 loop opens to release the second product EP-UNAG. The shift of Arg 397 that accompanies the dissociation of the phosphate ion aligns the positively charged side chain of Arg 397 parallel to that of Lys 48 . This would result in repulsive forces in the hinge region between the bottom and top domains of the enzyme, which might initiate the opening of the cleft. Comparison of MurA:P 1 P 2 with the unliganded state of the enzyme (Protein Data Bank entry 1EJC) (28) reveals that the widening of the cleft dislodges the ␣-carbon atom of Asp 123 by 3.0 Å from its position in the product state. By this shift the side chain of Asp 123 would lose its interaction with the uridine moiety of EP-UNAG and would have enough space to rotate about 180°around its angle. Such rotation would swap the positions of the side chain of Asp 123 and the last turn of the loop-anchoring helix a2 of subdomain IIc of MurA (Fig. 7, b and  c). It has been reported previously that the opening of the last turn of helix a2 is a main feature in the transition from the  6. Coordination of the phosphate ion in the product state of MurA (stereo representation). Dashed lines designate hydrogen bonds. In position 115 a modeled cysteine (as in wild-type MurA from E. cloacae) and aspartate (as in MurA from Mycobacterium) are superimposed to the serine from the MurA:P 1 P 2 structure. The respective resulting hydrogen-bonding distances to Arg 120 are shown in magenta.
winded conformation of unliganded MurA observed in 1EJC into the U-shaped loop of liganded MurA (30). We propose that transposition of the last turn of the loop-anchoring helix and Asp 123 is the key event for winding up the loop and eventually releasing EP-UNAG.
Conclusions-Because both products are locked in the active site if residue 115 is a serine, the mutation of the wild-type cysteine apparently impedes the initial trigger for a suite of conformational changes in the enzyme that finally leads to the dissociation of P i and EP-UNAG. This finding, together with the kinetically detected single turnover, questions the previously suggested dominant role of the wild-type cysteine as proton donor in the course of product formation (24). The function of Cys 115 might rather be to keep binding of the phosphate group weak enough to allow dissociation of the ion from the enzyme as soon as the covalent bond to the enolpyruvyl moiety is lost. Such a weakening of the coordination of the phosphate ion in the product state of the enolpyruvyl transfer reaction could probably be realized by any acceptor group in hydrogenbonding distance to NH 1 of Arg 120 . In the mechanistically similar enzyme AroA, which is the only other known enolpyruvyl transferase besides MurA, Arg 124 corresponds in position and function to Arg 120 of MurA (12,15). In this enzyme, NH 1 of Arg 124 exerts a bifurcated hydrogen bond to the phosphate group of the tetrahedral intermediate and to the carbonyl oxygen atom of the main chain at Asn 94 . The fact that AroA does not posses a cysteine in the vicinity of Arg 124 but still catalyzes a comparable reaction as MurA corroborates our findings that Cys 115 of MurA is just one of several possible alternatives to facilitate product dissociation in enolpyruvyl transferases.
The proposed ordered mechanism for product release is in contradiction to the previous suggestion that MurA would only open under substrate depletion (30). The structure of the trapped product state shows that the Pro 112 -Pro 121 loop remains in the same conformation like during catalysis. As long as the phosphate ion is bound to the active site, there apparently is no exit route for the large EP-UNAG molecule from the closed enzyme state. Thus, MurA has to close and reopen with every catalytic turnover. Binding of UNAG to MurA initiates the catalytic cycle by triggering enzyme closure (17,18). Subsequently PEP would enter the closed enzyme on the same route that the phosphate ion later uses to exit and docks to its newly formed binding site. After the enolpyruvyl transfer is completed, the produced phosphate ion leaves the active site. It thereby triggers opening of the two-domain structure of MurA and winding up of the Pro 112 -Pro 121 loop, which in turn allows the dissociation of the second product EP-UNAG.
Although details of the conformational changes required for product release remain speculative, the presented structure expands the data base of x-ray structures that characterizes MurA in different states of its catalytic cycle. The presented structure depicts the moment right before the macro-conformational change from the liganded closed form to an open form of MurA sets in and hence contributes to the elucidation of the induced fit mechanism. Understanding the details of domain movement interwoven with the function of MurA would lay the groundwork for an alternate strategy for structure-based drug design. Identification of the prerequisites of the induced fit mechanism may provide new templates for the design of novel antibacterial agents that do not target the active site but block closure of the enzyme and formation of the catalytic center.

Susanne Eschenburg, Melanie Priestman and Ernst Schönbrunn
Enolpyruvyl Transferase (MurA) Is Essential for Product Release -acetylglucosamine N in UDP-115 Evidence That the Fosfomycin Target Cys