Figures in brackets are for outer shell data.
Crystal structures of Staphylococcus aureus type I dehydroquinase from enzyme turnover experiments
Article first published online: 14 MAY 2004
Copyright © 2004 Wiley-Liss, Inc.
Proteins: Structure, Function, and Bioinformatics
Volume 56, Issue 3, pages 625–628, 15 August 2004
How to Cite
Nichols, C.E., Lockyer, M., Hawkins, A.R. and Stammers, D.K. (2004), Crystal structures of Staphylococcus aureus type I dehydroquinase from enzyme turnover experiments. Proteins, 56: 625–628. doi: 10.1002/prot.20165
- Issue published online: 18 JUN 2004
- Article first published online: 14 MAY 2004
- Manuscript Accepted: 1 MAR 2004
- Manuscript Received: 27 JAN 2004
- Arrow Therapeutics
Staphylococcus aureus is a major nosocomial pathogen which causes a range of diseases, including endocarditis, osteomyelitis, pneumonia, toxic-shock syndrome, food poisoning, carbuncles, and boils. The acquisition and spread of a variety of resistance genes has led to the emergence of endemic S. aureus populations resistant to all current antibiotics, even vancomycin.1
A promising approach in the search for new antimicrobial drugs to combat resistant bacteria is the structural characterization of biosynthetic enzymes common to bacteria and microbial eukaryotes, with a view to rational design. Shikimate pathway enzymes make excellent potential targets as they have no mammalian homologues and pathogenic bacteria mutant in this pathway are attenuated for virulence.1 The enzyme 3-dehydroquinate dehydratase (DHQase; EC 22.214.171.124) catalyses the third step of the shikimate pathway, the dehydration of 3-dehydroquinate (DHQ) to 3-dehydroshikimate (DHS), a step also common to the catabolic quinate pathway.3
DHQases can be divided into two classes according to mechanism of action, stereochemistry, overall structure, and sequence homology.4 Type I enzymes, which are generally only involved in biosynthesis, occur either as homodimers of subunit Mw ∼ 27 KDa or as one domain of the multifunctional AROM complex.5 Type I DHQases catalyse a syn elimination of water with the loss of the pro-R hydrogen from C2 via an imine intermediate.6 By contrast, type II enzymes occur as homo-dodecamers with subunit Mw ∼ 16 KDa, catalyse an anti elimination reaction with the loss of the more acidic axial pro-S hydrogen from C2, via an enolate intermediate7 and may function anabolically in the shikimate pathway and/or catabolically in the quinic acid pathway. Structures have been solved for both type-I4, 8 and type-II4, 9, 10 enzymes.
Structures of Salmonella typhi type-I DHQase have a typical (α/β)8 (TIM barrel) fold, with an additional “N”-terminal anti-parallel β-sheet region blocking substrate ingress from one side and a flexible loop region that folds in to close off the other end when substrate binds. In this paper, we report the crystallization and structure determination of an additional type-I DHQase from S. aureus (SaDHQase) in unliganded and liganded forms, both resulting from enzyme turnover experiments. Addition of substrate was essential for growing usable crystals. Different substrate concentrations yielded both an apo enzyme and a binary product complex with DHS still covalently linked to the catalytic lysine (Lys160) allowing a comparative analysis of the active site structure and the conformational changes associated with the bound ligand.
Materials and Methods.
The S. aureus DHQase gene was cloned into the Escherichia coli expression vector pRSETb. DHQase was purified by ammonium sulfate fractionation followed by column chromatography using Sephacryl S-300, DEAE Sephacel, ceramic hydroxyapatite, and MonoQ. DHQase enzyme was located by direct assay of the fractions and the purity was monitored by SDS PAGE. This procedure yielded approximately 150 mg of protein suitable for crystallography from 9 l of culture.
For crystallization, protein was exchanged into 10 mM TRIS (pH 7.4), 40 mM KCl, filtered through Amersham NAP™ 25 columns and re-concentrated to 30 mgml−1. Crystallizations were set up using six standard sparse matrix and four Grid screening kits (Hampton Research), 1:1 mixing ratio with 6 μl initial droplet size for sitting-drop vapor-diffusion experiments, and incubation at 277 K. Optimization of initial hits was achieved by employing finer intervals of pH/precipitant concentration. The crystals used for final data collection were grown from 16–18% PEG8000, 0.1 M TRIS (pH 8.0).
Crystals were transferred to their well-solution and 15% DMSO and 20% ethylene glycol added prior to data collection at 100 K at ESRF, beamline ID14EH4 (λ = 0.933 Å). Indexing, integration and merging of data images were carried out with DENZO and SCALEPACK.11 Rotation function/translation searches and rigid-body Patterson correlation refinement were carried out using CNS.12 Molecular replacement solutions were checked visually using “O”, as described previously.13 Rigid body, positional/B-factor refinement, simulated annealing, and initial water picking were carried out in CNS. Manual rebuilding, including insertion of ions, ligands, and extra water molecules was carried out using the program “O”. The final models were overlaid with previously released type-I DHQase structures using Top3D14 and results compared visually in “O” & VMD.15 Final figures were prepared using Corel11.
Results and Discussion.
Sparse matrix screens of the “apo” enzyme yielded numerous hits in a wide variety of conditions but the quality of X-ray data obtained from such crystals was always poor. Attempts were therefore made to crystallize SaDHQase with a range of concentrations of its substrate DHQ, to assess the utility of an “enzyme-turnover” approach. This proved successful as many of the resultant crystals diffracted to better than dmin = 2.5 Å. However, a partitioning of data quality was observed with low (5 mM, Crystal I) and high (20 mM, Crystal II) concentrations of DHQ yielding better data than either apo or intermediate crystallization attempts. Dataset II was solved by molecular replacement, using the “A” chain monomer of PDB model 1QFE, yielding two unique solutions equivalent to a dimer in the asymmetric unit. Dataset I was solved by rigid body refinement using the rebuilt model from dataset II. Further refinement of the structures gave the statistics shown in Table I. Crystal I is an open-form “apo” structure [Fig. 1(b)] with a largely disordered lid domain (LD) flipped away from the active site. Crystal II has an ordered LD domain in the closed position and clear electron density for DHS indicating the product is still covalently attached to Lys160 [Fig. 1(b), DHS B-factor of 42.7] analogous to the borohydride-reduced product seen in S. typhi DHQS.8 The observed substrate concentration-dependent partitioning of S. aureus DHQase crystal quality may be due to conformational heterogeneity in the enzyme. At the higher substrate concentration, S. aureus DHQase crystallizes substantially as a product complex (resolution 2.4 Å, overall B-factor of 53.6) but may include some unliganded enzyme. The lower concentration of DHQ gives more ordered crystals (resolution 1.8 Å, overall B-factor of 38.3). The order of such crystals may result from the generation of a particular conformational state following enzyme turnover and dissociation of product. Intermediate concentrations of substrate are presumed to give more heterogeneous mixtures of unliganded and product-bound enzyme hence resulting in crystals that show poorer quality diffraction.
|Unit Cell a,b,c (Å)/α = β = γ (°)||77.1, 77.1, 175.2 / 90||76.6, 76.6, 172.4 / 90|
|Resolution range (Å)||30.00–1.80 (1.83–1.80)||30.00–2.36 (2.44–2.36)|
|Redundancy||10.8 (6.3)||4.0 (3.9)|
|Completion (%)||92.5 (70.4)||95.1 (95.8)|
|Rmergea||0.061 (0.467)||0.074 (0.671)|
|I/σI||37.7 (3.2)||13.3 (1.46)|
|Asymmetric unit (Subunits)||2||2|
|Residues in most favored regionsc (%)||89.5||91.7|
|Residues in additionally allowed regionsc (%)||10.5||8.3|
|Mean B-factors (Å2)|
|RMSD bond lengths (Å)||0.005||0.007|
|RMSD bond angles (°)||1.19||1.27|
Comparing the S. aureus and S. typhi DHQase structures8 shows that although the sequence identity is relatively low at 32% the tertiary structure is very similar (RMSD for CA atoms of 1.2 Å with a 90% match rate). The S. aureus DHQase structure however lacks the “N”-terminal anti-parallel β-sheet region (NT) seen with S. typhi enzyme. The orientation of the two monomers is also slightly different making the dimer more compact, perhaps increasing structural rigidity and offsetting the loss of cross-linking by the NT region [Fig. 1(a)]. Within the active site, the substrate contacting residues critical for catalytic competency; Glu35(46), Arg37(48), Arg70(82), Lys160(170), Arg202(213), and Gln225(236), are conserved and map very tightly between the two structures [numbers in parentheses indicate equivalent S. typhi residues, Fig. 1(c)]. The surrounding residues, which form the lining of the barrel, are less well-conserved with three substitutions relative to the S. typhi sequence [Fig. 1(d)]. The substituted residues are however of similar bulk and only interact weakly with the bound substrate as their function is to form part of the “cushion” of residues lining the barrel and preventing trans access by water, The functional significance of these changes would therefore be predicted to be small.
Overall there is a high degree of structural homology between the two species' DHQases and it would seem likely that inhibitors of the enzyme of one species would also be efficacious against the other. The use of enzyme turnover for crystallization has again been demonstrated as a useful technique, opening up new regions of “crystallization-space” that can give rise to crystals suitable for structure determination where previous methods failed.16
We are grateful to the staff at ESRF, Grenoble, France, for their assistance with data-collection. We acknowledge financial support from the BBSRC and Arrow Therapeutics.