Three-dimensional structure of the quorum-quenching N-acyl homoserine lactone hydrolase from Bacillus thuringiensis
- Dali Liu*,
- Bryan W. Lepore*,†,
- Gregory A. Petsko*,
- Pei W. Thomas‡,
- Everett M. Stone§,
- Walter Fast‡,§,¶, and
- Dagmar Ringe*,∥
- ‡Division of Medicinal Chemistry, College of Pharmacy, and §Graduate Program in Cell and Molecular Biology, University of Texas, Austin, TX 78712; and †Program in Bioorganic Chemistry and *Departments of Chemistry and Biochemistry and Rosenstiel Basic Medical Sciences Research Center, Brandeis University, Waltham, MA 02454-9110
-
Contributed by Gregory A. Petsko, June 28, 2005
Abstract
The three-dimensional structure of the N-acyl-l-homoserine lactone hydrolase (AHL lactonase) from Bacillus thuringiensis has been determined, by using single-wavelength anomalous dispersion (SAD) phasing, to 1.6-Å resolution. AHLs are produced by many Gram-negative bacteria as signaling molecules used in quorum-sensing pathways that indirectly sense cell density and regulate communal behavior. Because of their importance in pathogenicity, quorum-sensing pathways have been suggested as potential targets for the development of novel therapeutics. Quorum-sensing can be disrupted by enzymes evolved to degrade these lactones, such as AHL lactonases. These enzymes are members of the metallo-β-lactamase superfamily and contain two zinc ions in their active sites. The zinc ions are coordinated to a number of ligands, including a single oxygen of a bridging carboxylate and a bridging water/hydroxide ion, thought to be the nucleophile that hydrolyzes the AHLs to ring-opened products, which can no longer act as quorum signals.
Studying the sociomicrobiology of bacterial populations has illuminated the importance of quorum sensing in the virulence of both plant and animal pathogens (1). Over 50 species of Gram-negative bacteria are known to produce N-acyl-l-homoserine lactones (AHLs) as signaling molecules used in quorum-sensing pathways that indirectly sense cell density and regulate communal behavior (2). For example, the plant pathogen Erwinia cartovora produces N-3-(oxohexanoyl)-l-homoserine lactone to coordinate production of exoenzymes involved in soft rot disease and potato blackleg (3). Mixed biofilms containing Pseudomonas aeruginosa and Burkholderia cepacia often form on the lung surfaces of human patients with cystic fibrosis and produce AHLs thought to enable interspecies coordination of virulence factor production (4). AHL-mediated signaling is also a significant factor in the virulence of the biowarfare pathogen Burkholderia mallei, the etiologic agent of glanders (5). Because of their importance in pathogenicity, quorum-sensing pathways have been suggested as potential targets for the development of novel therapeutics (6).
Nature has already evolved several methods that can subvert quorum-sensing pathways by producing natural products that act as AHL antagonists (7) and by expressing “quorum-quenching” enzymes that can hydrolyze either the amide (8) or the lactone (9) moiety of AHL-signaling molecules. Currently, the best-characterized quorum-quenching enzymes catalyze AHL ring opening by hydrolyzing lactones to the corresponding γ-hydroxy acid, a product that can no longer be used for signaling. AHL lactonases are found predominantly in Gram-positive bacteria, although they are not entirely absent from Gram-negative species (10). The AHL lactonases from Bacillus sp. seem to display broad substrate specificity in regards to acyl-chain length and substitution (11) but show strict selectivity for the (S)-configuration found in naturally occurring AHLs (12). The utility of these enzymes in combating quorum-sensing organisms has been clearly demonstrated through the use of bacteria that naturally express AHL lactonase and with bacteria and transgenic plants engineered to heterologously express these lactonases (10).
Despite their promise, not much is known about the structure or catalytic mechanism of the AHL lactonases. Amino acid sequence alignments reveal a conserved HXHXDH motif that is also found in the metallo-β-lactamase (MBL) superfamily of proteins (9). This superfamily shares an αβ/βα protein fold, but its members display only low sequence identity and have a variety of diverse activities (13). Hydrolytic enzymes in this superfamily, such as RNase Z, MBL, and glyoxalase II, generally rely on 1 or 2 equiv of bound zinc to catalyze their reactions, but the redox-active enzymes, such as rubredoxin oxygen:oxidoreductase and nitric oxide reductase, instead use a diiron center and contain an additional flavin-binding domain (13). It has not yet been possible to predict the metal content of these enzymes from their primary sequence (14). Initial reports describing AHL lactonase and variants containing site-directed mutations of the proposed metal-binding motif suggest that these enzymes may not be metal-dependent (11), but a recent study (12) provides evidence that an AHL lactonase containing 2 equiv of zinc is an active catalyst of AHL hydrolysis, that these metal ions are required for activity, and that they bind close together, forming a dinuclear site resembling that of glyoxalase II. Here, we describe the x-ray crystal structure of an AHL lactonase.
Materials and Methods
Protein Expression and Purification. The dinuclear zinc form of the AHL lactonase from Bacillus thuringiensis subsp. kurstaki (accession number AF478059) was expressed as a maltose-binding protein (MBP) fusion, cleaved, and purified as described (12). For crystallization trials, the purified AHL lactonase was concentrated to 10 mg/ml and exchanged into a buffer containing 10 mM Hepes (pH 7.0) and 2 mM DTT.
Crystallization. AHL lactonase crystallization was carried out in hanging drops by using standard screens (Hampton, San Diego). Crystals were obtained from a solution containing 20% glycerol, 80 mM Tris·HCl, 24% polyethyleneglycol (PEG) 4000, and 160 mM MgCl2 at pH 8.5, by using a protein:well solution ratio of 4:1. Crystals appeared within 2 weeks at room temperature along with cloudy precipitates. The average size of the crystals has the smallest dimension around 0.3 mm. Crystals with the best morphological quality were transferred into fresh drops of well solution to get rid of attached precipitates before being frozen with liquid nitrogen. The native mother liquor was adequate as a cryo-protectant. Crystals were found to have the symmetry of space group P212121 with cell dimensions of a = 54.6, b = 55.6, and c = 80.1 Å. The data are not consistent with the tetragonal crystal system despite the near identity of the unit cell a and b axes. Crystals contain one molecule per asymmetric unit with a solvent content of 45% by volume (Table 1).
Data Collection. Taking advantage of the two zinc ions bound at the active site of AHL lactonase, a single-wavelength anomalous dispersion (SAD) data set was collected on beamline 14-BMD at BioCars, Advance Photon Source (APS), Argonne National Laboratory. A fluorescence scan determined the zinc absorption peak wavelength to be 1.2827 Å. One SAD data set was collected at the peak wavelength, by collecting 1° per frame, with inversion of the crystal by 180° every 10° to minimize radiation damage to Friedel pairs on an Advanced Detector Systems Corporation (ADSC) Q4 charge-coupled device (CCD) area detector. The SAD data set resolution was cut off at 2.5 Å based on merging statistics. A 1.6-Å data set was collected on beamline 11-1 at Stanford Synchrotron Radiation Lab (SSRL) at a wavelength of 0.8265 Å with 1° oscillation per frame on an ADSC Q315 CCD area detector (Table 1). Both data sets were integrated and scaled by using the hkl software suite (15).
Phasing and Refinement. The program solve (16) was used to locate the positions of the two zinc ions from an anomalous difference Patterson synthesis calculated by using the SAD data set (Table 1). Preliminary phase angles were then calculated, and statistical density modification was carried out in resolve (17). The iterative pattern-matching/model-building routines in resolve (18) coupled to refinement in refmac (19) generated the initial protein model with a free R factor of 44.0%. Interactive model building was then carried out with the programs xfit (20) and coot (21), assisted by software from the Uppsala Software Factory (ref. 22, http://xray.bmc.uu.se/usf). Map format conversion was carried out by using cns2fsfour (www.scripps.edu/~cdputnam/software/cns2fsfour.html). Simulated annealing torsion angle refinement and positional and B-factor refinement were carried out in cns (23) by using the maximum likelihood target on Hendrickson–Lattman (24) coefficients to construct the bulk of the protein chain, including the zinc atoms. Phase extension to 1.6-Å resolution using the native data set allowed gaps, connectivity, and registration errors in the model to be cleared up, and addition of bound solvent molecules. The entire protein, including the two zinc atoms, was defined as a single rigid body, the displacement of which was modeled with anisotropic displacement parameters (ADPs) derived from a single set of Translation-Libration-Screw (TLS) tensors (25). TLS tensors, all atomic positions, residual B-factors for ADPs, and the full B-factors for atoms excluded from the TLS group were refined (19). fft, mapmask, and other programs from the ccp4 suite (26) were used during refinement. Throughout, judgment of phase quality was assisted by use of the Model Bias Removal Server (http://tuna.tamu.edu), and ADP analysis was assisted by using the PAR-VATI server (ref. 27, www.bmsc.washington.edu/parvati/parvati.html). Molecular surface area was calculated with cns (23).
The final model includes all of the expected amino acid residues, two zinc ions, and eight glycerol molecules, including three in dual coexisting conformations, and 178 waters. Calculation of a Ramachandran plot with procheck (28) indicates that Ile-190 and Asp-50 lie in generously allowed and disallowed sectors of the plot, respectively; however, they exhibit outstanding omit map electron density (29). Asp-50 makes a 2.8-Å hydrogen bond to His-109, which coordinates one of the two zinc ions. Ile-190 is only one position before Asp-191, which contributes the bridging carboxylate between the zincs. Twenty-three other residues are in additionally allowed sectors of the Ramachandran plot; the rest are in core sectors. Five surface residues containing 17 atoms total were observed to have little to no electron density; therefore, their occupancies were set to zero (Glu-43, Thr-67, Phe-68, Glu-211, and Arg-244). Twenty-nine residues are modeled with dual conformations beyond Cα. The mean anisotropy for atoms in the TLS group is 0.53, with sigma 0.138, which is not significantly different from values of other high resolution structures in the Protein Data Bank. All ADPs have positive ellipsoids, and no atoms have anisotropy >0.1.
Figures were made by using stamp (30), molscript (ref. 31, www.avatar.se), povscript+ (ref. 32, www.stanford.edu/~fenn/povscript), weblab viewerlite (Molecular Simulations, Waltham, MA, www.msi.com), pymol (www.pymol.org), or povray (www.povray.org).
Sedimentation Velocity Analytical Ultracentrifugation. Sedimentation velocity data were collected in a Beckman Optima XL-I analytical ultracentrifuge. Double-sector cells, with 12-mm Epon centerpieces and quartz windows, were loaded with AHL lactonase samples (≈400 μl each) at three different protein concentrations (0.134 mg/ml, 0.101 mg/ml, and 0.067 mg/ml), all in Na2PO4 buffer (100 mM) at pH 7. Absorbance data were collected at 230 nm at a continuous rotor speed of 40,000 rpm for 10 h taking one scan every 10 min at 20°C. The sedimentation velocity absorbance profiles were analyzed to obtain the apparent distribution of sedimentation coefficients for all of the quaternary structures present in solution by using the program sedfit and van Holde–Weischet distribution plots (33, 34). The primary amino acid sequence of AHL lactonase was used to calculate both the s max (3.35 × 10-13 s) and, by using the Biology Workbench (http://workbench.sdsc.edu), the partial specific volume (0.73 cm3/g) of this protein. All experimental results were corrected to s 20,w according to standard procedures (33).
Results and Discussion
Overall Structure. The overall structure of AHL lactonase reveals a MBL-like (αβ/βα) fold (35). Two major β-sheets, which contain five and six strands, respectively, are in a parallel formation at the center of the structure. This β/β motif is surrounded by six α-helices (Fig. 2). The N terminus and C terminus of the protein are 16 Å from one another and are separated by the turn of an antiparallel strand that specifically coordinates the C-terminal carboxylate, giving the C terminus a relatively low B value and excellent omit map electron density. However, the N terminus is less well ordered. Two loop-rich regions are found at the two ends of the β/β motif. The active site is defined by the location of two zinc ions and is located at one end of this fold, where the largest number of loops are found.
The crystals contain one AHL lactonase subunit per asymmetric unit. The largest surface area between molecules in the asymmetric unit and any crystallographically related subunit is 2,340 Å2, which is 27% of the total surface area from all crystallographic contacts made by the monomer in the asymmetric unit. This result implies that AHL lactonase is a monomer under our crystal conditions.
Oligomeric State. The MBL superfamily contains examples of monomeric (36), dimeric (37), and tetrameric (38) proteins. To investigate the oligomerization state of AHL lactonase in solution, the sedimentation coefficient (s 20,w = 2.76 ± 0.04 × 10-13 s) was determined by sedimentation velocity analytical ultracentrifugation and found to be smaller than that calculated for a monomeric spherical protein with the same primary amino acid sequence (s max,calc = 3.35 × 10-13 s), an observation consistent with AHL lactonase also being a monomer in solution (Fig. 1). The vertical nature of the van Holde–Weischet distribution plots indicates that this monomeric species is a homogeneous one-component system with consistent sedimentation properties at all concentrations tested (34).
Sedimentation velocity analysis of AHL lactonase. Three concentrations of AHL lactonase, 0.067 mg/ml (▴), 0.101 mg/ml (□), and 0.134 mg/ml (▪), were analyzed by sedimentation velocity analytical ultracentrifugation, and van Holde–Weischet integral distribution plots from each are overlayed.
Two Twisted Antiparallel Strands. A striking aspect of the AHL lactonase structure is the third antiparallel β-sheet extending beyond the β/β motif formed by strand 1 (15 residues), strand 2 (13 residues), and a short third strand (strand 4, Fig. 2). Strands 1 and 2 are the longest β strands of the structure. The loop that joins these antiparallel strands folds over the active site and helps form the β-sheet that covers the active site. A feature of these antiparallel strands is their twisted conformation, corresponding to a single turn of ≈180° about the strand axis from one end of the strand to the other. The strands are also bent slightly off-axis toward the active site. The origin of this bending is the classic β bulge at positions Ala-11 and Gly-12 (Fig. 2; ref. 39). At one end of these antiparallel strands, the first of the two β-sheets in the αβ/βα fold is formed with four other shorter strands (strands 3, 5, 6, and 7).
The overall fold of AHL lactonase. Two zinc ions are shown as silver spheres, the bridging water species as a red sphere, and the N and C termini are labeled. The classic β bulge area is in the dashed circle and also shown in a zoomed-in image on the right. The antiparallel configuration of strands 1 and 2 breaks at residues Ala-11 and picks up again at Arg-13.
A DALI (distance matrix alignment) search (40) identified nine structures from the Protein Data Bank (PDB) with significant structural similarity. The two most similar proteins are methyl parathion hydrolase (MPH) from Pseudomonas Sp. Wbc-3 (PDB ID code 1P9E, Z score = 22.2) and a MBL from Bacteroides fragilis (PDB ID code 1A7T, Z score = 16.5, ref. 41, Fig. 3). Most parts of the AHL lactonase structure are superimposable with both of these structures (rms deviation 2.7 and 2.6 Å, respectively, on Cαs).
Comparison of the structures of the AHL lactonase (green; this work) with those of methyl parathion hydrolase (blue; PDB ID code 1P9E) and metallo-β-lactamase (red; PDB ID code 1A7T).
The structure of AHL lactonase is distinct from those of MPH and MBL by the presence of the three-stranded β-sheet composed of strands 2, 1, and 4 (Fig. 2 and 3). In particular, the MPH and MBL structures have only two antiparallel strands in the structurally equivalent region. In AHL lactonase, these two antiparallel β-strands hang over the active site (see below). In AHL lactonase, these three strands along with the loops attached to them also position a number of hydrophobic side chains toward the active site of the protein. We suggest that these strands may contribute to substrate recognition, formation of the active site, and/or desolvation. The substrates of AHL lactonase, which have hydrophobic acyl chains ranging from 4–12 carbons in length (11), will most likely require more buried surface area, closure, rigidification, and/or definition of the active site, relative to the substrates of MPH, MBL, and other similar enzymes.
Active Site. Differing reports have been published regarding the metal content of active AHL lactonase (11, 12). In this structure, we find that two zinc ions are bound at the active site and are located 3.3 Å apart in a bridged dinuclear site. The zinc ions in our structure are located in a loop-rich region on top of the αβ/βα fold, with the third β-sheet (strands 1, 2, and 4) reaching over the top of the zinc ions like a lid (Fig. 2). As shown in Fig. 4, seven residues are found coordinating the two zinc ions: His-104, His-106, and His-169 coordinate Zn1; Asp-108, His-109, and His-235 coordinate Zn2; and a single side-chain carboxylate oxygen from Asp-191 and either a water or a hydroxide ion form two bridges between the two zinc ions. The oxygen of the monodentate bridging Asp-191 is somewhat closer to Zn2 (2.0 Å) than Zn1 (2.6 Å, Fig. 5), consistent with more favorable interactions forming between cations and the syn lone pair, rather than the anti lone pair of a carboxylate (42). In total, each zinc ion has five ligands and is coordinated with distorted trigonal bipyramidal geometry. The constellation of the zinc-binding sites is consistent with that predicated by extended x-ray absorption fine structure (EXAFS) studies of AHL lactonase and sequence comparisons with other members of this superfamily of enzymes (12).
The wall-eyed stereo presentation of the AHL lactonase active site. Zinc ions are shown as gray spheres, the bridging water as a red sphere. Glycerol molecules are shown in ball-and-stick form with yellow bonds. One glycerol molecule is shown with two alternate conformations. The anomalous difference Fourier map, calculated without zinc phases, contoured at 35 σ, is rendered as a red mesh.
Schematic diagram of the active site showing coordination and distances in Å. The bridging water species is shown as a hydroxide for the convenience of making the figure.
Besides the high similarity existing in overall structure among the three proteins compared (AHL lactonase, MPH, and MBL; see below), the active site of AHL lactonase is remarkably similar to those of MPH and MBL (Fig. 6). Methyl parathion hydrolase coordinates its two zinc ions exactly the same way as AHL lactonase. However, MBL is distinct in that Cys-164 is structurally equivalent to the bridging carboxylate, yet only coordinates Zn2. Secondly, Cys-87 in MBL is structurally equivalent to His-109, which is a ligand to Zn2 in AHL lactonase. However, it is observed to be >5 Å from either of the two zinc ions in the structure of MBL and is not considered to be a zinc ligand. Tyr-194 is highly conserved in AHL lactonases from different organisms (12) and is located in the active site (Figs. 4 and 5), making it a potential catalytic residue in the AHL lactonase reaction. The existence of Y194 is unique for lactonases because the structurally equivalent residues are L258 in MPH (PDB ID code 1P9E) and K167 in MBL (PDB ID code 1A7T, Fig. 6).
Structural alignment of the active site regions of AHL lactonase (green), MPH (blue; PDB ID code 1P9E), and MBL (red; PDB ID code 1A7T). For clarity, zinc ions and the bridging water species are shown for the AHL lactonase structure only. Nonconserved residues are underlined.
As is so often the case in enzyme crystal structures, the residues that are Ramachandran plot outliers (I190 and D50) in AHL lactonase are found close to the active site, one adjacent to the bridging ligand Asp-191, and the second within van der Waals contact of the Zn2 ligand His-109 (Fig. 4). The Ramachandran plot outliers from seven di-zinc and di-iron enzymes of similar fold follow this trend: the outliers in three structures of rubredoxin oxygen oxidoreductase (PDB ID code 1E5D, ref. 43), flavoprotein Tm07551 (PDB ID code 1VME), and glyoxalase II (PDB ID code 1QH5, ref. 44) are adjacent to the bridging ligand, whereas the outliers in the others are in close proximity to the active sites. The residues adjacent to the bridging ligand tend to fall in the +φ/-Ψ sector of the Ramachandran plot, suggesting that this conformation is important to the energetics of bridged bimetallo enzymes of the MBL superfamily.
Glycerol Molecules Bound to AHL Lactonase. Electron density corresponding to a total of eight bound glycerol molecules was observed. Five glycerol molecules are hydrogen-bonded on the surface of the protein between crystallographically related protein molecules, and two others (glycerol 805 and glycerol 835) are in very close proximity to the zinc ions. Some of the glycerol moieties, including the one closest to the zinc ions, were modeled in dual conformations. The second hydroxyl of glycerol 835 has very little electron density; therefore, the occupancy was set to zero. Although AHL lactonase crystallizes readily without glycerol, the use of 20% glycerol in the crystallization conditions forms relatively larger crystals. Glycerol 805 sits above the bridging water or hydroxide ion, whereas the glycerol 835 in the active site region is further away (Figs. 4 and 7). Although glycerol is not a perfect analog of either the substrates or the products in the AHL lactonase reaction, it is reasonable to suggest that one of its hydroxyl groups mimics the hydroxyl group in the final product. Two of the hydroxyl groups of the glycerol molecule closest to the zinc ions are within hydrogen bonding distance of the potential catalytic residues. The 2-hydroxyl group is 2.6 Å from the bridging water or hydroxide ion, and 2.5 Å from the phenolic oxygen of Y194. The 1-hydroxyl group is 2.8 Å from the noncoordinating carboxylate of Asp-108 (Fig. 5).
Branched cavity. Zinc ions are shown as gray spheres, the bridging water as a red sphere. Tyr-194 and Asp-108 are shown in green. The rest of the zinc-coordinating residues are shown with standard atom colors. Glycerol molecules 805 and 835 are in magenta. The surface is colored according to the contributing atoms.
A Branched Cavity. The active site cavity has a rather interesting shape. It is branched and extends deep into the enzyme (Fig. 7). The two zinc ions, the bridging water/hydroxide, and the potential catalytic residues Y194 and Asp-108 are all sitting on the ridge between the two branches (Fig. 7). The easiest way for the lactone ring to bind to the potential catalytic groups would be in a head-in direction, with the hydrophobic acyl chain pointing out. But that would leave the two deep branches unoccupied during catalysis. If the branches are actually used to accommodate shorter hydrophobic acyl chains for some AHL substrates, then how the substrate chooses one branch over the other is a mystery. The active site cavity is covered on one side by two three-residue motifs (Cys-Met-Phe), the side chains of which form hydrophobic patches. One branch contains Cys-14, Met-53, and Phe-107. The other branch contains Cys-141, Met-138, and Phe-64. The sulfur atom of Cys-141 makes a hydrogen bond to an oxygen atom of one of the bound glycerol molecules. Those residues are not conserved in other AHL lactonases, nor is the functional significance of this repeated arrangement clear.
We propose three possibilities for the function of this unique branched cavity. First, different lactones varying in the length and the nature of the C3 position (11) of the acyl chains could occupy different parts of the cavity. Second, with different binding direction, the enzyme might be able to catalyze the reaction by means of different mechanisms. However, the fact that AHLs with varying length are hydrolyzed with similar kinetics (11) makes this possibility less likely. Third, evolution may not have fully optimized AHL lactonase for AHL substrates, and alternative substrates that completely occupy this branched cavity could exist.
Proposed Substrate Binding and Catalytic Mechanism. The distorted trigonal bipyramidal coordination geometry of each zinc ion leaves two empty coordination sites available for ligand binding at the active-site pocket, providing an excellent site for dual Lewis acid catalysis. AHL lactonase is also known to hydrolyze N-acyl-(S)-homoserine lactone substrates with acyl chains varying from 4–12 carbons long, all with similar kinetics (2, 3). In view of these constraints, we can evaluate several possible substrate-binding orientations. If the lactone carbonyl oxygen of an AHL is coordinated to Zn2 and the leaving group to Zn1, then substrates with long acyl chains would fit poorly and clash with the walls of the binding pocket, making this orientation unlikely. Placing the lactone carbonyl oxygen over Zn1 and the leaving group over Zn2 would allow longer acyl chains to fit better into the active-site groove. Although other possibilities of different binding orientations may still exist for the substrates with short acyl chains, we will not further speculate on them in this article.
Keeping the orientation of the long acyl chains, the lactone ring could then place either its si or re face toward the zinc ions. If the more-hindered si face of the lactone ring faces the metal center, the carbonyl and leaving group oxygens of the AHL would be oriented toward the conserved Asp-108 residue. However, Asp-108 is coordinated to Zn2 and is within hydrogen bonding distance of the bridging water/hydroxide (Fig. 5), suggesting that this residue is likely deprotonated in the resting state. An anionic aspartate would not serve as a suitable general acid to protonate the lactone's leaving group, making this orientation less likely unless Asp-108 serves as a proton shuttle, as proposed in the mechanism of isoaspartyl dipeptidase, an unrelated dinuclear hydrolase (45).
Alternatively, if the less-hindered re face of the lactone ring faces the metal center, the lactone's carbonyl and leaving group oxygens would be oriented toward the phenol sidechain of a conserved Tyr-194 residue. This orientation would also point the amide nitrogen of the chiral substrate away from the floor of the active site, providing a possible reason why AHL lactonase is selective for the (S)-enantiomer (12). Tyr-194 is not coordinated to either metal and is well positioned to serve as a hydrogen bonding partner to stabilize a tetrahedral adduct or as a general acid to protonate the leaving group during lactone hydrolysis. Alternatively, Zn2 may play a role in stabilizing the leaving group directly, as seen with other hydrolytic enzymes in this superfamily (46, 47). Finally, by analogy to bridged bimetallo-aminopeptidases (48), Zn2 may play a role in stabilizing the tetrahedral transition state such that the substrate carbonyl oxygen forms a transient ligand to Zn1 and the second oxygen of the carboxylate product, derived from the bridging water/hydroxide, forms a ligand to Zn2. If such a transition state forms, one would expect a base to abstract the proton from the bridging water/hydroxide when it attacks the carbonyl carbon of the substrate. Asp-108 is a candidate for such a role because its dangling carboxylate oxygen is only 2.8 Å away from the bridging water/hydroxide (Fig. 5). Such a role has been proposed for Glu-204 in methionine aminopeptidase (49).
The two glycerol molecules bound at the active site also support this proposed substrate-binding mode and act as a form of solvent mapping of this site (50). Glycerol 835 lies in the arm of the branched cavity that extends away from Zn1 and is also adjacent to a long hydrophobic groove suitable for accommodating acyl chains of longer AHL substrates. Glycerol 805 is bound directly over the dinuclear zinc site and positions its 2-hydroxyl group equidistant from each zinc ion and within hydrogen bonding distance of the phenol side chain of Tyr-194, consistent with proposing a catalytic role for this residue. Although Tyr-194 is conserved in the AHL lactonases, it is not found in the two proteins that have the closest overall protein folds. At this position, the Bacteroides fragilis metallo-β-lactamase has a Lys-167 that is not essential for catalysis (51) and the methyl parathion hydrolase has Leu-258 (PDB ID code 1A7T), presumably reflecting different catalytic requirements for each reaction (Fig. 6). Although functional studies are needed to verify the importance of Tyr-194, this conserved residue may be helpful in distinguishing AHL lactonases from other superfamily members based on primary sequence information.
Determination of the AHL lactonase protein fold firmly places this enzyme in the metallo-β-lactamase superfamily, which nature has evolved to catalyze a diverse set of reactions encompassing both hydrolytic and redox chemistry (13). Understanding the detailed structure of AHL lactonase will help us to understand how small variations in this motif lead to mechanistic differences and may help us to design more effective catalysts for disrupting the quorum-sensing pathways of pathogenic organisms.
Use of the Advanced Photon Source (APS) was supported by the U.S. Department of Energy, Basic Energy Sciences, Office of Science, under Contract No. W-31-109-Eng-38. Use of the BioCARS Sector 14 was supported by the National Institutes of Health, National Center for Research Resources, under Grant RR07707. Portions of this research were carried out at the Stanford Synchrotron Radiation Laboratory (SSRL), a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences. The SSRL Structural Molecular Biology Program is supported by the Department of Energy, Office of Biological and Environmental Research, and by the National Institutes of Health, National Center for Research Resources, Biomedical Technology Program, and the National Institute of General Medical Sciences. This work was supported by National Institutes of Health Grant GM26788 (to G.A.P. and D.R.), National Institutes of Health Grant K22 AI50692 (to W.F.), and Robert A. Welch Foundation Grant F-1572 (to W.F.).
Footnotes
-
↵ ¶To whom correspondence may be addressed at: Division of Medicinal Chemistry, College of Pharmacy and the Graduate Program in Cell and Molecular Biology, University of Texas, Austin, TX 78712. E-mail: waltfast{at}mail.utexas.edu. ∥To whom correspondence may be addressed at: Departments of Chemistry and Biochemistry and Rosenstiel Basic Medical Sciences Research Center, Brandeis University, Waltham, MA 02454-9110. E-mail: ringe{at}brandeis.edu.
-
Author contributions: D.L., B.W.L., P.W.T., and E.M.S. performed research; D.L., B.W.L., G.A.P., P.W.T., E.M.S., W.F., and D.R. analyzed data; G.A.P., W.F., and D.R. designed research; and D.L., B.W.L., W.F., and D.R. wrote the paper.
-
Abbreviations: AHL, N-acyl-l-homoserine lactone; SAD, single-wavelength anomalous dispersion; ADP, anisotropic displacement parameter; TLS, translation-libration-screw; MPH, methyl parathion hydrolase; MBL, metallo-β-lactamase.
-
Data deposition: The atomic coordinates and structure factors have been deposited in the Protein Data Bank, www.pdb.org (PDB ID code 2A7M).
- Copyright © 2005, The National Academy of Sciences