Structure of homoserine O-acetyltransferase from Staphylococcus aureus: the first Gram-positive ortholog structure

The structure of homoserine O-acetyltransferase (HTA) from the human pathogen Staphylococcus aureus has been determined. Despite a similar overall fold and active site architecture to other α/β-hydrolases, this more compact HTA structure has a more narrow access to the active site than can confer important specificity differences.

Abstract

Homoserine O-acetyltransferase (HTA) catalyzes the formation of l-O-acetyl-homoserine from l-homoserine through the transfer of an acetyl group from acetyl-CoA. This is the first committed step required for the biosynthesis of methionine in many fungi, Gram-positive bacteria and some Gram-negative bacteria. The structure of HTA from Staphylococcus aureus (SaHTA) has been determined to a resolution of 2.45 Å. The structure belongs to the α/β-hydrolase superfamily, consisting of two distinct domains: a core α/β-domain containing the catalytic site and a lid domain assembled into a helical bundle. The active site consists of a classical catalytic triad located at the end of a deep tunnel. Structure analysis revealed some important differences for SaHTA compared with the few known structures of HTA.

1. Introduction  

Homoserine O-acetyltransferase (HTA; EC 2.3.1.31) lies at a branch point in the aspartate metabolic pathway (Viola et al., 2011 ▶; Umbarger, 1978 ▶) which leads to the biosynthesis of the essential amino acids methionine and lysine in most plants, bacteria and fungi. The branch leading to the synthesis of methionine, as well as the essential methyl-transfer cofactor S-adenosylmethionine (AdoMet), starts with the acetylation of homoserine to form O-acetylhomoserine (OAH) catalyzed by HTA. AdoMet is also the substrate for a group of enzymes that synthesize acylhomoserine lactones (AHLs), which are virulence signaling molecules in Gram-negative bacteria (Hanzelka & Greenberg, 1996 ▶; Schaefer et al., 1996 ▶). The reaction catalyzed by HTA involves the acetylation of the γ-hydroxyl of homoserine through an acetyl-CoA-dependent acetylation via a double-displacement mechanism (Born & Blanchard, 1999 ▶) facilitated by a classic Ser–His–Asp catalytic triad which is located at the bottom of a narrow tunnel (Mirza et al., 2005 ▶). This reaction represents a critical control point for cell growth and viability. Genetic studies in bacteria and yeast have shown that deletion of the met2 gene that encodes HTA is lethal in minimal medium, and methionine supplementation is required for cell viability (Pascon et al., 2004 ▶; Nazi et al., 2007 ▶). Additionally, the met2 gene has been established to be essential for virulence in human pathogens (Nazi et al., 2007 ▶). Since the aspartate pathway is absent in mammals but is necessary for bacterial survival, the enzymes in this pathway are potential targets for the development of new antibiotics (Pavlovsky et al., 2012 ▶).

In the present study, we report the structure of HTA from Staphylococcus aureus (SaHTA). S. aureus is a Gram-positive bacterium and is the most common species of Staphylococcus to cause staph infections. MRSA is a widespread strain of S. aureus that has evolved resistance to β-lactam antibiotics, increasing the urgency for the development of new drugs against this human pathogen. Because HTA catalyzes an important step in an essential metabolic pathway in S. aureus, this enzyme can potentially serve as a novel target (De Pascale et al., 2011 ▶) for the development of S. aureus-selective drugs. This paper reports the structure of SaHTA determined to 2.45 Å resolution with an overall conserved secondary structure of the α/β-hydrolase superfamily. This study also shows differences in the structure from previously reported structures of Haemophilus influenza HTA (HiHTA; Mirza et al., 2005 ▶) and Leptospira interrogans HTA (LiHTA; Wang et al., 2007 ▶).

2. Materials and methods  

2.1. Expression and purification of recombinant SaHTA  

The sequence of the putative met2-encoded HTA gene from S. aureus has been reported. Two oligonucleotide primers, 5′-AGAAAGCTGGGTCCTTAGCTTAAAA-3′ and 5′-AAAAAGCAGGCTTCGAAGGAGATAGA-3′, were designed and synthesized (IDT) to be complementary to the amino-terminal and carboxyl-terminal ends of the gene, respectively. These primers were used to amplify the SaHTA gene from S. aureus genomic DNA using standard PCR conditions. The gene was cloned into pDEST42 vector containing a carboxyl-terminal hexahistidine tag by Gateway Technology (Invitrogen). The BL21 (DE3) Escherichia coli cell line transformed with the plasmid DNA was used for the expression of recombinant native SaHTA. Transformed cells were grown at 310 K until the culture density reached an OD₆₀₀ of 0.7–0.8. Cells were then cooled to 289 K followed by the addition of IPTG to a final concentration of 1 mM and were allowed to grow for an additional 16 h at 289 K (De Pascale et al., 2011 ▶). Cells were harvested by centrifugation at 17 000g for 10 min. The cell paste (5 g) was resuspended in 25 ml of a buffer consisting of 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 500 mM NaCl, 20 mM imidazole, 5 mM β-mercaptoethanol, 10% glycerol pH 7.5 and the cells were homogenized using a hand homogenizer. The cells were lysed by sonication for a total of 8 min at 30 s intervals with a sonic dismembrator and the cell debris was removed by centrifugation at 12 000g for 30 min. The supernatant was filtered through a 0.8 µm membrane syringe filter and loaded onto an Ni–IMAC column for initial purification. The enzyme was further purified on a Source 30Q anion-exchange column using an ÄKTA chromatographic system (GE Biosciences). The fractions that showed catalytic activity and contained a single band on a Coomassie-stained polyacrylamide gel were pooled and dialyzed overnight in a buffer consisting of 50 mM HEPES pH 7.5, 100 mM NaCl, 1 mM dithiothreitol (DTT), 0.1 mM ethylenediaminetetraacetic acid (EDTA). The protein was then concentrated using 10 kDa cutoff ultracentrifugal concentrators (Millipore) and the concentration of the protein sample was determined using a NanoDrop 2000 spectrophotometer. All purification steps were carried out at 277 K.

2.2. Biochemical assay  

HTA catalytic activity was determined by following the change in the absorbance owing to hydrolysis of the thioester bond of acetyl-CoA at 232 nm (∊ = 4.5 mM ⁻¹; Born & Blanchard, 1999 ▶) using a Cary UV–visible spectrophotometer. Assays were performed in 1 ml quartz cuvettes using an assay buffer consisting of 50 mM HEPES pH 7.5, 1 mM l-homoserine, 0.5 mM acetyl-CoA. The decrease in the absorption corresponds to the cleavage of the thioester bond of acetyl-CoA with subsequent formation of the product O-acetylhomoserine. The extinction coefficient of SaHTA was calculated to be 39 070 cm⁻¹  M ⁻¹.

2.3. Crystallization and data collection  

Initial screening was carried out using commercial sparse-matrix kits from Jena Bioscience with 10 mg ml⁻¹ protein concentration. Crystals of the SaHTA apoenzyme were grown at 293 K by the hanging-drop vapor-diffusion method using a 1:1 ratio of protein solution (5 mg ml⁻¹) to well solution with a total drop size of 2 µl. Hexagonal-shaped diffraction-quality crystals were obtained in 5–7 d when using 0.7 M ammonium formate, 100 mM imidazole–HCl pH 6.5 as the well solution. In preparation for data collection, the crystals were briefly soaked in a 25% ethylene glycol cryosolution and flash-cooled. These crystals were found to diffract to ∼2.5 Å resolution using an in-house R-AXIS IV image plate mounted on a Rigaku FR-E rotating-anode X-ray generator. Full crystallographic data were measured to 2.45 Å resolution on beamline 23-ID-B of GM/CA CAT at the Advanced Photon Source (Argonne National Laboratory), with the resolution limited by the increased crystal-to-detector distance needed to obtain good separation of the diffraction data. Data were processed using the HKL-2000 software suite (Minor et al., 2000 ▶). Details of the data-collection statistics are provided in Table 1 ▶.

Table 1. Data-collection and refinement statistics for SaHTA.

Values in parentheses are for the highest resolution shell.

Data collection
Wavelength ()	1.033
Space group	P6122
Unit-cell parameters (, )	a = 49.16, b = 49.16, c = 481.47, = = 90, = 120
Resolution ()	502.45 (2.542.45)
Observed reflections	73534
Unique reflections	21299
Multiplicity	3.5 (2.8)
Completeness (%)	88.7 (54.9)
R merge (%)	0.148 (0.370)
I/(I)	7.3 (2.6)
Wilson B factor (2)	39.9
Refinement
R work	0.24
R free	0.29
No. of protein atoms	2660
No. of waters	31
R.m.s.d., bond lengths ()	0.008
R.m.s.d., bond angles ()	1.359
Average B factor (2)	59.8
Ramachandran plot statistics†, residues in (%)
Favored region	93.5
Allowed region	5.6
Outlier region	0.9

^†As reported by Coot (Emsley et al., 2010 ▶).

2.4. Structure determination and refinement  

Initial attempts to use either of the known HTA structures (PDB entries 2b61 and 2pl5; Mirza et al., 2005 ▶; Wang et al., 2007 ▶) as search models did not yield a useful starting point for determination of the SaHTA structure. The structure of SaHTA was solved by a search model which was generated by BALBES, a fully automatic molecular-replacement pipeline (Long et al., 2008 ▶). The initial model from BALBES was further refined by Buccaneer (Cowtan, 2006 ▶), an automated protein model-building program in the CCP4 suite (Winn et al., 2011 ▶). The program automatically located most of the protein residues and the model generated had an overall figure of merit (FOM) of 0.60 with R _work = 0.30 and R _free = 0.35. The model was further refined and diffraction phases to 2.45 Å resolution were calculated and improved using Coot (Cowtan, 2006 ▶; Emsley et al., 2010 ▶) and REFMAC5 (Murshudov et al., 2011 ▶). Weighting of the X-ray and geometric components in the minimization function in REFMAC5 was carried out by choosing a geometric weighting factor of 0.03. A series of cycles of manual building and iterative restrained refinement with the data resolution extended to 2.45 Å produced the final structure with R _work = 0.24 and R _free = 0.29. The refined model had an overall FOM of 0.73. The stereochemical quality of the final model was checked by PROCHECK (Laskowski et al., 1993 ▶). The final structure was visualized and analyzed using PyMOL (DeLano, 2002 ▶). Atomic coordinates and structure factors have been deposited in the Protein Data Bank (Berman et al., 2003 ▶) with accession code 4qlo. The statistics of the structure refinement are summarized in Table 1 ▶.

3. Results and discussion  

3.1. Structure overview  

The full-length SaHTA enzyme consists of 322 amino acids, while the structural model accounts for residues 3–322 as well as the additional seven amino acids (DPAFLYK) which were derived from the vector. The structure of SaHTA is a member of the α/β-hydrolase fold family (Ollis et al., 1992 ▶; Heikinheimo et al., 1999 ▶; Nardini & Dijkstra, 1999 ▶), sharing the same overall structural features as HiHTA and LiHTA. SaHTA is organized into two distinct domains: a core α/β-domain (residues 3–161 and 239–322) and a helical bundle lid domain (residues 162–238) forming a canopy over the core domain (Fig. 1 ▶ a). The core domain consists of an eight-stranded, predominantly parallel β-sheet (β1, β4–β10), with the connectivity of the β-sheet achieved by five α-helices (αA–αE) on one side and one (αF) on the opposite side (Fig. 1 ▶ b). Two short antiparallel β-strands, β2 and β3, provide the connection between β1 and β4. The lid domain is organized as five α-helices (αL1–αL5). This domain is connected to the core domain via links between αL1 and β8 and between αL5 and αD (Fig. 1 ▶ b), with the intervening residues lacking defined secondary structure. The electron density for residues 197–208 is significantly weaker and similar observations were made for the LiHTA structure (Wang et al., 2007 ▶), in which the electron density for residues 252–268 was also quite poor. This is the same region in both structures, constituting the αL4 of the lid domain. The previously proposed flexibility in this region that would allow structural changes in response to substrate binding is supported by our current structure.

Figure 1.

(a) Stereoview of SaHTA in a ribbon representation. The β-sheets colored yellow and the α-helices colored red compose the core domain, while the α-helices of the lid domain are colored blue. The catalytic triad residues Ser131, Asp267 and His296 are shown in stick representation. (b) Schematic diagram of the topology of SaHTA showing the two domains and the catalytic residues.

Similar to the HiHTA (PDB entry 2b61; Mirza et al., 2005 ▶) and LiHTA (PDB entry 2pl5; Wang et al., 2007 ▶) structures, SaHTA crystallizes with only one HTA molecule in the asymmetric unit. However, both HiHTA and LiHTA have been shown by gel-filtration and sedimentation-velocity analysis to exist as dimers in solution (Born et al., 2000 ▶; Wang et al., 2007 ▶). Helices L1 and L2 from each subunit form a classical antiparallel four-helix bundle with the hydrophobic core, while helices L3 and L4 are involved in dimer stabilization flanking the four-helix bundle. In SaHTA the four-helix bundle is found to be left-handed, similar to that seen in HiHTA, while in the LiHTA structure this bundle is right-handed.

3.2. Active-site residues  

The catalytic triad consisting of a serine, a histidine and an aspartic acid is the most conserved feature of the α/β-hydrolase family. In the SaHTA structure, the catalytic triad is composed of the Ser131 nucleophile, the His296 base and the Asp267 acidic residue. Ser131 is located on a loop between β7 and αC, Asp267 is on a loop between β9 and αE, and His296 is on a loop between β9 and αE. This structural arrangement is conserved among all members of the HTA family.

In the LiHTA apoenzyme structure the active-site histidine was found to exist in two conformations with similar occupancy (Wang et al., 2007 ▶). In our SaHTA structure His296 is somewhat disordered, with a relatively high temperature factor (B = 72 Ų) for the imidazole ring relative to the other active-site functional groups, while in the HiHTA structure this histidine has well defined density and exists in a single conformation. This observed conformational flexibility of this active-site histidine in HTAs is consistent with its proposed role in catalysis. The k _cat/K _m values for acetyl-CoA and l-homoserine with HiHTA were reported to be 6.5 × 10⁵ and 7.1 × 10⁵  M ⁻¹ s⁻¹, respectively (Born et al., 2000 ▶). For SaHTA, the k _cat/K _m values for acetyl-CoA and l-homoserine were found to be 4.0 × 10⁵ and 3.3 × 10⁵  M ⁻¹ s⁻¹, respectively. These values are each within a factor of two of those reported for the Gram-negative ortholog, consistent with the highly conserved active site in this enzyme family.

3.3. Secondary- and tertiary-structure comparison among the HTAs  

Despite the many similarities of SaHTA to LiHTA and HiHTA, there are some significant structural differences. SaHTA is a shorter protein (322 amino acids) compared with LiHTA (366 amino acids) and HiHTA (377 amino acids) (Fig. 2 ▶). These deletions are primarily located in the loop regions. The loop connecting β6 and αB (amino acids 81–105) is shorter in SaHTA. Similarly, the 15-amino-acid loop connecting αL1 and αL2 has essentially been eliminated, as has the 15-amino-acid loop between αL3 and αL4, when compared with these regions in LiHTA and HiHTA (Fig. 2 ▶). Helix αE is also shorter in SaHTA. Overall, SaHTA can be viewed as a more compact version of HiHTA, with several truncated loops in this new structure but with no significant changes in the orientation of the secondary-structural elements (Fig. 3 ▶ a). In contrast, there are several differences in the secondary-structural arrangement between SaHTA and LiHTA. For example, in SaHTA the loop connecting β6 and β7 is flanked by a short helix which is not seen in LiHTA (Fig. 3 ▶ b).

Figure 2.

Sequence alignment of HiHTA, LiHTA and SaHTA with the assigned secondary structures. The catalytic triad residues Ser131, Asp267 and His296 are highlighted with asterisks.

Figure 3.

Ribbon diagram showing (a) superimposition of SaHTA (cyan) and HiHTA (green) and (b) SaHTA (cyan) and LiHTA (magenta).

3.4. Surface view and active-site tunnel of SaHTA  

An important common structural feature of HTA is the presence of a tunnel formed at the junction of the two domains (Fig. 4 ▶ a). This tunnel leads to the active-site nucleophile, Ser131, and appears to act as a channel for substrate access to the active site. The cross section of the tunnel reveals the size and shape of this channel leading to the active site. When compared with the tunnel present in HiHTA (Fig. 4 ▶ b) and LiHTA (Fig. 4 ▶ c), this feature in SaHTA is much narrower and more compact (Fig. 4 ▶ d), suggesting the possibility of some specificity differences in the S. aureus enzyme.

Figure 4.

(a) Cross section of SaHTA highlighting the active-site tunnel. The catalytic triad residues Ser131, Asp267 and His296 are shown as green sticks. Comparison of the active-site tunnel between (b) HiHTA, (c) LiHTA and (d) SaHTA, with the active-site serine colored red.

The natural product ebelactone A inhibits HiHTA with a K _i value of 203 ± 12 µM (De Pascale et al., 2011 ▶). This compound was shown to form a covalent adduct with the active-site serine by accessing it through the catalytic tunnel. When this compound was tested with SaHTA no inhibition was observed at concentrations as high as 1 mM. This greater than 50-fold discrimination in binding between these two homologous enzymes is likely to be due to the narrower catalytic tunnel in SaHTA which restricts access to the active-site serine.

4. Conclusions  

The HTA from S. aureus has a similar overall structure to the HTA orthologs from two Gram-negative bacteria with the same active-site catalytic triad. Flexibility in the linkages between the two domains would allow protein conformation changes in response to substrate binding. Differences have been found in the dimensions of the tunnel that potential substrates must traverse to enter the active site. This unique characteristic tunnel in the S. aureus enzyme can serve to limit the size of compounds that can access the active site, allowing this enzyme form to potentially avoid inhibition from known inhibitors of other HTA orthologs (De Pascale et al., 2011 ▶).

Supplementary Material

PDB reference: SaHTA, 4qlo