Crystallization and preliminary X-ray diffraction analysis of the high molecular weight ketoacyl reductase FabG4 complexed with NADH

The cloning, overexpression, purification, crystallization in a triclinic space group and preliminary X-ray diffraction analysis of the high-molecular-weight ketoacyl reductase FabG4 complexed with NADH are reported.

Abstract

FabG4 from Mycobacterium tuberculosis belongs to the high molecular weight ketoacyl reductases (HMwFabGs). The enzyme requires NADH for β-ketoacyl reductase activity. The protein was overexpressed, purified to homogeneity and crystallized as a FabG4–NADH complex. A mountable FabG4:NADH complex crystal diffracted to 2.59 Å resolution and belonged to space group P1, with unit-cell parameters a = 63.07, b = 71.03, c = 92.92 Å, α = 105.02, β = 97.06, γ = 93.66°. The Matthews coefficient suggested the presence of four monomers in the unit cell. In addition, a self-rotation function revealed the presence of two twofold NCS axes and one fourfold NCS axis. At χ = 180° the highest peak corresponds to the twofold NCS between two monomers, whereas the second peak corresponds to the twofold NCS between two dimers.

1. Introduction  

Complex circuits of lipid metabolism in Mycobacterium tuberculosis enable the organism to synthesize several unusual lipid derivatives for its virtually impenetrable cell envelope. Nonpolar lipid derivatives resist the entrance of deleterious molecules generated by the host into the bacterial cell. The bacteria also keep changing their cell-envelope composition in response to environmental conditions (Makinoshima & Glickman, 2005 ▶). Lipid metabolism, including fatty-acid metabolism, thus plays a significant role in this organism.

Fatty-acid synthesis in M. tuberculosis is classified into two types: FAS-I and FAS-II (Takayama et al., 2005 ▶). The eukaryotic-like single enzyme multi-domain system FAS-I synthesizes fatty-acyl chains of up to 20 carbons in length. FAS-II, which is a set of enzymes, subsequently elongates the fatty-acyl chain up to 70–80 carbons in length. The elongation module of the FAS-II pathway consists of the enzymes ketoacyl synthase (KasA/KasB; Mdluli et al., 1998 ▶; Schaeffer et al., 2001 ▶; Gao et al., 2003 ▶), ketoacyl reductase (FabG1 or MabA; Marrakchi et al., 2002 ▶), hydroxyacyl dehydratase (HadAB/HadBC; Sacco et al., 2007 ▶), enoyl reductase (InhA; Dessen et al., 1995 ▶) and an acyl carrier protein (AcpM; Wong et al., 2002 ▶). FabG1, which catalyses the second step in the FAS-II module, is also known as β-ketoacyl reductase. The M. tuberculosis genome encodes more than one copy of the β-­ketoacyl reductase gene. The annotated FabG genes are fabG1 (Rv1483), fabG2 (Rv1350), fabG3 (Rv2002), fabG4 (Rv0242c), fabG5 (Rv2766c) and fabG6 (Rv3502c). Ketoacyl reductases belong to the short-chain dehydrogenase/reductase (SDR) family and mostly have a core structure of 250–350 amino acids in length (Oppermann et al., 2003 ▶). All of the FabG gene products of M. tuberculosis are 247–317 amino acids in length, with the exception of FabG4. FabG4 is unusually long (454 amino acids) and contains two distinguishable domains (Dutta et al., 2011 ▶). The N-­terminal domain is a flavodoxin-type domain and the C-terminal is a ketoreductase domain. Thus, this protein, along with its sequence homologues, is categorized as a high molecular weight FabG (HMwFabG). Although the role of FabG4 in mycobacteria is yet to be addressed, some reports have recently identified FabG4 as an essential gene in mycobacteria (Beste et al., 2009 ▶; Gurvitz, 2009 ▶; Sharma et al., 2010 ▶).

The C-terminal domain (215–454) of FabG4 shares 32% sequence identity with M. tuberculosis FabG1 (Cohen-Gonsaud et al., 2002 ▶), 35% with Plasmodium falciparum FabG (Wickramasinghe et al., 2006 ▶), 36% with Brassica napus FabG (Fisher et al., 2000 ▶), 38% with Staphylococcus aureus FabG1 (Dutta et al., 2012 ▶) and Rickettsia prowazekii FabG (Subramanian et al., 2011 ▶), 39% with Bacillus anthracis FabG (Zaccai et al., 2008 ▶), 40% with Aquifex aeolicus FabG (Mao et al., 2007 ▶) and 41% with Escherichia coli FabG (Price et al., 2001 ▶). Unlike all of these FabGs, FabG4 utilizes NADH for catalysis. In the cases of E. coli FabG and M. tuberculosis FabG1, NADPH has been hypothesized to reorient the catalytic residues in position for catalysis (Price et al., 2004 ▶; Cohen-Gonsaud et al., 2005 ▶). These holo structures also maintain two loops near the catalytic residues in stable conformations. In some FabG structures the stability of these two loops does not depend on the coenzyme (Wickramasinghe et al., 2006 ▶). It has been proposed that the amino-acid residue preceding the catalytic serine is responsible for holding the catalytic loop in its proper position (Poncet-Montange et al., 2007 ▶). We have recently determined the crystal structure of apo FabG4^17–448, truncating the six C-terminal residues ⁴⁴⁹QAMIGA⁴⁵⁴ (Dutta et al., 2011 ▶). The C-­terminal residues are involved in a hydrophobic interaction with the corresponding catalytic loop in FabG4, thereby stabilizing it. This clearly indicates the role of the conserved C-terminus in HMwFabGs. Intriguingly, a conserved C-terminus is also observed for FabG1s among mycobacterial species (Cohen-Gonsaud et al., 2005 ▶).

Although much study has been carried out on FabGs, the nature of the HMwFabGs has remained obscure, including their coenzyme specificity. The present work reports the overexpression, purification, crystallization and preliminary X-ray diffraction studies of the FabG4^17–454:NADH complex.

2. Materials and methods  

2.1. Purification of FabG4^17–454  

The cloning of the FabG4 gene from the M. tuberculosis H37Rv genome in pQE30 expression vector and the overexpression of recombinant His₆-FabG4^17–454 in E. coli M15 (pREP4) cells have previously been reported (Dutta et al., 2011 ▶). The entire FabG4 gene was not soluble. The protein became soluble upon truncation of the first 16 amino-acid residues.

The cells from 2 l culture were resuspended in buffer A (10 mM Tris–HCl pH 7.9, 300 mM NaCl, 10 mM imidazole, 5% glycerol) containing 0.1 mM each of leupeptin, aprotinin and pepstatin and 0.02 mM phenylmethylsulfonyl fluoride. The following steps were carried out at 277 K. The suspension was lysed by ultrasonication and the resulting lysate was centrifuged at 22 000g for 30 min. The supernatant was passed through Ni–NTA Sepharose high-performance affinity matrix (GE Healthcare Biosciences) pre-equilibrated with buffer A. The column was then extensively washed with buffer A. An intermediate buffer B (10 mM Tris–HCl pH 7.9, 300 mM NaCl, 100 mM imidazole, 5% glycerol) was passed through the column to remove any nonspecifically bound contaminant. Finally, the protein was eluted with buffer C (10 mM Tris–HCl pH 7.9, 300 mM NaCl, 300 mM imidazole, 5% glycerol). The eluted protein was subjected to gel-filtration chromatography using Superdex 200 prep-grade matrix in a 16/70C column (GE Healthcare Biosciences) pre-equilibrated with buffer D (10 mM Tris–HCl pH 7.9, 200 mM NaCl, 5% glycerol) on an ÄKTAprime Plus system. The flow rate was set to 1 ml min⁻¹. The fractions containing the desired protein were pooled together and kept for further use. The purity of the protein sample was assessed by SDS–PAGE and the protein concentration was measured using both the Abs₂₈₀ and the Bradford method (Bradford, 1976 ▶). The absence of NAD⁺ cofactor was confirmed by the OD₂₈₀/OD₂₆₀ ratio as described previously (Krimsky & Racker, 1963 ▶).

2.2. Crystallization of the FabG4^17–454:NADH binary complex  

Purified FabG4 was concentrated to 20 mg ml⁻¹ using a 10 kDa cutoff Vivaspin 20 concentrator (GE Healthcare). Concentrated protein was mixed with NADH (Sigma) in a 1:5 molar ratio followed by an incubation period of 30 min at 300 K. The mixture was subjected to crystallization using the previously reported crystallization condition for the C-terminally truncated mutant of apo FabG4 (Dutta et al., 2011 ▶). Despite several trials, no crystals appeared from this condition. The mixture was thus subjected to preliminary crystallization trials using Crystal Screen, Crystal Screen 2 and Index from Hampton Research by mixing 2 µl protein/NADH mixture with an equal volume of reservoir solution using the sitting-drop vapour-diffusion method. Initial hits were only obtained using a condition consisting of 0.1 M bis-Tris pH 6.5, 45%(v/v) polypropylene glycol P400. The single crystals that appeared from the initial crystallization condition were too tiny to mount (Fig. 1 ▶ a). The initial crystallization condition was further fine-screened using the hanging-drop vapour-diffusion method to optimize the crystal size and quality. After a number of screening trials with different buffering reagents and precipitating reagents, mountable crystals were obtained from the condition 0.1 M MES pH 6.5, 45%(v/v) polypropylene glycol P400 (Fig. 1 ▶ b).

Figure 1.

Crystals of the FabG4^17–454:NADH complex. (a) Initial crystals obtained from the preliminary crystallization screening. (b) Mountable crystals obtained from 0.1 M MES pH 6.5, 45%(v/v) polypropylene glycol P400.

2.3. Data collection and processing  

For data collection, an FabG4^17–454:NADH crystal was directly mounted from the mother liquor and flash-cooled in a nitrogen stream at 100 K. X-ray diffraction data were collected using an in-house X-ray diffraction facility. The facility was equipped with a Rigaku MicroMax-007 HF rotating-anode generator as a Cu Kα X-­ray source and a Rigaku R-AXIS IV⁺⁺ image-plate detector. The crystal-to-detector distance was maintained at 200 mm and the crystal was rotated 360° with 1° rotation per frame. The crystal diffracted to a resolution of 2.59 Å. The images were processed with XDS (Kabsch, 2010 ▶) in space group P1 and scaled with SCALA (Evans, 1993 ▶) from the CCP4 suite (Winn et al., 2011 ▶). The final statistics of data collection and processing are tabulated in Table 1 ▶.

Table 1. Data-collection and processing statistics.

Values in parentheses are for the highest resolution shell.

Wavelength (Å)	1.54
Space group	P1
Unit-cell parameters (Å, °)	a = 63.07, b = 71.03, c = 92.92, α = 105.02, β = 97.06, γ = 93.66
Resolution (Å)	19.78–2.59 (2.73–2.59)
Total No. of observations	179404 (23917)
No. of unique reflections	45759 (6394)
Completeness (%)	95.0 (90.7)
Multiplicity	3.9 (3.7)
Average I/σ(I)	7.7 (1.7)
R merge † (%)	15.0 (84.2)
Molecules per asymmetric unit (Z)	4
Matthews coefficient (Å3 Da−1)	2.21

^† R _merge = , where I _i(hkl) is the ith observed intensity of reflection hkl and 〈I(hkl)〉 is the mean intensity over all i measurements.

3. Results and discussion  

Recombinant FabG4^17–454 was successfully purified to homogeneity. The N-terminal residues (1–16) were found to hinder solubility and hence were truncated without altering the activity (Dutta et al., 2011 ▶). FabG4^17–454:NADH complex crystals were obtained from 0.1 M MES pH 6.5, 45%(v/v) polypropylene glycol P400 and diffracted to 2.59 Å resolution. The crystals belonged to space group P1, with unit-cell parameters a = 63.07, b = 71.03, c = 92.92 Å, α = 105.02, β = 97.06, γ = 93.66°, which differed from those of the FabG4^17–448 crystals. The Matthews coefficient (2.21 ų Da⁻¹) confirmed that there were four monomers in the unit cell (Matthews, 1968 ▶). The presence of a single fourfold NCS axis was also indicated by the self-rotation function plot at χ = 90° (Fig. 2 ▶ a). The NCS axis makes an ∼30° angle with the crystallographic y axis. The self-rotation function at χ = 180° (Fig. 2 ▶ b) also indicated two twofold NCS axes. The highest peak (peak 1; Rf = 508.2) is attributed to the dimeric axis between two monomers, whereas the second peak (peak 2; Rf = 338.7) is attributed to the dimeric axis between two dimers.

Figure 2.

Self-rotation function plot obtained from the FabG4^17–454:NADH complex data for (a) χ = 90° and (b) χ = 180°. Two peaks (peak 1 and peak 2) are indicated by red arrows.

The structure was solved by the molecular-replacement method with MOLREP (Vagin & Teplyakov, 2010 ▶), using a monomer of the apo FabG4 structure (PDB entry 3lls; Seattle Structural Genomics Center for Infectious Disease, unpublished work) as a search model. A promising result with a homotetrameric assembly has been found with a final score of 0.71; the resulting R factor was 37.4%. The model was subsequently subjected to rigid-body refinement followed by restrained refinement using REFMAC5 (Murshudov et al., 2011 ▶). Structural analysis is currently in progress.