Biochemical and structural characterisation of dehydroquinate synthase from the New Zealand kiwifruit Actinidia chinensis

Biochemical and structural characterisation of dehydroquinate synthase from the New Zealand kiwifruit Actinidia chinensis

Archives of Biochemistry and Biophysics 537 (2013) 185–191 Contents lists available at ScienceDirect Archives of Biochemistry and Biophysics journal...

1MB Sizes 0 Downloads 72 Views

Archives of Biochemistry and Biophysics 537 (2013) 185–191

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi

Biochemical and structural characterisation of dehydroquinate synthase from the New Zealand kiwifruit Actinidia chinensis Gerd Mittelstädt a, Leonardo Negron b,1, Linley R. Schofield b,2, Ken Marsh c, Emily J. Parker b,⇑ a

Biomolecular Interaction Centre and Department of Chemistry, University of Canterbury, PO Box 4800, Christchurch 8140, New Zealand Institute of Fundamental Sciences, Massey University, PO Box 11-222, Palmerston North 4442, New Zealand c The NZ Institute of Plant and Food Research, Mount Albert Research Centre, P.O. Box 92-169, Auckland, New Zealand b

a r t i c l e

i n f o

Article history: Received 4 June 2013 and in revised form 17 July 2013 Available online 2 August 2013 Keywords: Shikimate Quinic acid metabolism Kiwifruit proteins

a b s t r a c t One of the novel aspects of kiwifruit is the presence of a high level of quinic acid which contributes to the flavour of the fruit. Quinic acid metabolism intersects with the shikimate pathway, which is responsible for the de novo biosynthesis of primary and secondary aromatic metabolites. The gene encoding the enzyme which catalyses the second step of the shikimate pathway, dehydroquinate synthase (DHQS), from the New Zealand kiwifruit Actinidia chinensis was identified, cloned and expressed. A. chinensis DHQS was activated by divalent metal ions, and was found to require NAD+ for catalysis. The protein was crystallised and the structure was solved, revealing a homodimeric protein. Each monomer has a NAD+ binding site nestled between the distinct N- and C-terminal domains. In contrast to other microbial DHQSs, which show an open conformation in the absence of active site ligands, A. chinensis DHQS adopts a closed conformation. This is the first report of the structure of a DHQS from a plant source. Ó 2013 Elsevier Inc. All rights reserved.

Introduction The shikimate pathway consists of seven enzyme-catalysed reactions that generate chorismate [1,2]. Chorismate is the precursor to a range of aromatic compounds including the aromatic amino acids Trp, Phe and Tyr, which are required for protein biosynthesis. The shikimate pathway and the enzymes of aromatic amino acid biosynthesis are found in plants, fungi, bacteria and apicomplexan parasites, but not in animals. In plants, aromatic amino acids are the precursors for a range of diverse compounds which are important for plant growth, development, reproduction, defense and environmental responses [3]. The shikimate pathway also intersects with quinate metabolism, sharing two common metabolites dehydroquinate (DHQ3) and dehydroshikimate (DHS). The quinate pathway is responsible for the catabolism of quinic acid to protocatechuate (Fig. 1). Quinate is an abundant source of carbon for many fungi and plants. The production of DHQ is catalysed by the enzyme DHQ synthase (DHQS, EC 4.2.3.4) in the second reaction of the shikimate ⇑ Corresponding author. Address: Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch, New Zealand. Fax: +64 3 364 2110. E-mail address: [email protected] (E.J. Parker). 1 Current address: Callaghan Innovation Research Limited, Gracefield, Lower Hutt, New Zealand. 2 Current address: AgResearch Ltd, Grasslands Research Centre, Palmerston North, New Zealand. 3 Abbreviations used: DHQ, dehydroquinate; DHS, dehydroshikimate; NAD+, nicotinamide adenine dinucleotide; ORF, open reading frame; IMAC, immobilised metal affinity chromatography. 0003-9861/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.abb.2013.07.022

pathway. This reaction is the nicotinamide adenine dinucleotide (NAD+) dependent conversion of the sugar, 3-deoxy-D-arabinoheptulosonate 7-phosphate (DAH7P), into the first carbocycle of the pathway. In plants and bacteria DHQS exists as a monofunctional enzyme. In contrast, for fungi, the DHQS activity is part of the large pentafunctional AROM complex, which also catalyses the following four steps of the pathway [4–8]. The overall conversion of DAH7P to DHQ takes place in several steps, involving oxidation, elimination, reduction, ring opening and internal aldol reactions. There has been a considerable effort into defining the nature of the catalytic role the enzyme plays in each of these reactions steps [9–11]. The first structure of this enzyme was determined from an eukaryote, Aspergillus nidulans [12,13]. A. nidulans DHQS (AniDHQS) is homodimeric with subunits which comprise a Rossmann fold at the N terminus, and an a-helical Cterminal domain. The active site was located by the presence of a substrate analogue and a Zn2+ ion in a deep cleft between the Nand C-terminal domains. Subsequent structures of DHQS suggest that there is significant movement between open and closed forms of the enzyme on substrate binding [14]. The interactions between the bound ligands and the protein indicate that substrate binds in position to allow the protein to conduct multi-step catalysis and guide the formation of the correct product, DHQ. Structures of DHQS from the bacterial sources Staphylococcus aureus [15], Thermus thermophilus [16] and Helicobacter pylori [17] have also been described. In addition, the structures of the DHQS from Vibrio cholera and Mycobacterium tuberculosis are also available in the protein data bank (PDB, codes: 3OKF, 3QBD, 3QBE).

186

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

Fig. 1. Intersecting shikimate and quinate metabolic pathways.

Whereas the DHQS from bacterial and fungal sources has been investigated in some detail, there is very limited information about the enzyme from plant sources. Plant DHQSs were purified from Vigna mungo [18] and Pisum sativum [19], and the gene encoding the enzyme from tomato (Solanum lycopersicum) was able to complement a DHQS-deficient Escherichia coli strain [20]. Quinic acid is found in high concentrations in mature fruit from Actinidia species and is a distinguishing factor in fresh kiwifruit [21], hence shikimate and quinate metabolism is of particular interest. In these studies we characterise a plastidic isoform of DHQ synthase from the New Zealand kiwifruit Actinidia chinensis. This is the first description of the molecular structure of a plant DHQS. Materials and methods Cloning of A. chinensis dehydroquinate synthase gene An A. chinensis ‘Hort16a’ fruit library containing 7500 ESTs [22] was BLAST-searched [23] for sequences with homology to known DHQSs. A full-length cDNA of DHQS was identified by sequencing and by homology to the DHQS from S. lycopersicum. A truncated open reading frame (ORF), based on prediction by Target-P [24] and including the stop codon for DHQ synthase, was amplified by PCR using Pwo DNA Polymerase (Roche). The resulting product was subcloned into the pET200/D Topo vector (Promega) and sequence verified. Purification of A. chinensis DHQS (AchDHQS) Escherichia coli BL21(DE3)-Rosetta2 cells, transformed with vector pET200 containing the A. chinensis DHQS gene, were grown overnight at 37 °C in 50 mL Luria–Bertani (LB) medium supplemented with kanamycin (50 lg/mL) and chloramphenicol (25 lg/ mL). The preculture was used to inoculate 1 L of fresh LB medium, and growth continued at 37 °C. Isopropyl-b-D-thiogalactopyrano-

side was added at 0.5 mM to mid-logarithmic phase cultures. The temperature was reduced to 23 °C. E. coli cells were harvested by centrifugation (30 min, 14000g, 4 °C) 16 h after induction. Cell pellets were resuspended in 50 mM phosphate buffer containing 100 mM NaCl and 1 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0 and lysed by sonication. The cell lysate was centrifuged (30 min, 24000g, 4 °C). The supernatant containing AchDHQS was loaded onto an immobilised metal affinity chromatography (IMAC) column (5 mL HisTrap column (GE Healthcare)) with lowimidazole buffer (50 mM phosphate, 100 mM NaCl, 20 mM imidazole, pH 8.0) and eluted by applying a linear gradient to high-imidazole buffer (50 mM phosphate, 100 mM NaCl, 500 mM imidazole, pH 8.0) over 100 mL. AchDHQS containing fractions were pooled and mixed with TEV protease in a 1:100 ratio for overnight cleavage at 4 °C. Using IMAC his-tag free AchDHQS was obtained in the unbound flow through. Removing the tag did not alter the activity. For further purification size exclusion chromatography was used (26/600 Superdex 200 column (GE Healthcare)). The protein was frozen in liquid nitrogen and stored at 80 °C.

Activity assays The in vitro activity of AchDHQS was measured spectrophotometrically using a coupled enzyme continuous assay [25]. The AchDHQS-mediated conversion of DAH7P (prepared as described by Negron et al. [26]) into DHQ was followed by the E. coli dehydroquinase (EcoDHQase)-mediated conversion of DHQ to DHS. EcoDHQase (purification details are provided in the supplementary material) was added in excess. The standard reaction mixture contained DAH7P (0.78– 62.4 lM), CoCl2 (100 lM), NAD+ (29 lM) and EcoDHQase (680 nM) in 50 mM Bis–tris propane (BTP), 10 lM EDTA, pH 7.5 at 25 °C in a total volume of 1 mL. The reaction was initiated by the addition of AchDHQS (60 nM). KM and kcat values were determined by fitting the data to the Michaelis–Menten equation using GraFit

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

(Erithacus Software). CoCl2 was employed in the assays, to aid comparison to other DHQS enzymes. Altering the concentration of NAD+ (range) was observed to have no effect on reaction rates, in line with this cofactor being replenished in the DHQS mechanistic cycle.

187

in Phaser [30] to solve the structure of AchDHQS by molecular replacement. Refinements were conducted with Refmac5, and electron density maps were analysed with COOT [31]. The validation tools of COOT and MolProbity [32] were used to check the structures. All diagrams were drawn with PyMOL.

Effect of temperature on AchDHQS activity Results and discussion Buffer (50 mM BTP) was adjusted to pH 7.5 at the desired temperatures. The reaction mixture was prepared as described above with a constant DAH7P concentration of 40 lM and incubated for 5 min before the addition of EcoDHQase and AchDHQS. Control experiments were carried out to ensure that the rate of the reaction observed was limited by the rate of the DHQS reaction: conditions with double or half the concentration of DHQase showed no variation in rate. Differential scanning fluorometry Aliquots (25 lL) of 0.01 mg/mL protein (AchDHQS), 50 lM metal chloride (either CoCl2 or ZnCl2), 1 mM BTP buffer at pH 7.0, and 1X Sypro Orange dye (Invitrogen) were added into a 96-well microplate. The plates were sealed and heated in an iCycler iQ Real Time PCR Detection System (Bio-Rad) from 20 to 95 °C in increments of 0.5 °C, with 30 s dwell time. Experiments were carried out in triplicate. Metal dependency AchDHQS and EcoDHQase were pre-treated for 10 min at 21 °C with EDTA (100 lM) with the exception of one AchDHQS sample as control. All other assay components were prepared in pre-treated water (Chelex100 resin (Bio-Rad)). Reaction mixtures contained final concentrations of DAH7P (46 lM) and NAD+ (29 lM) and the divalent metal ion salt (100 lM) in 50 mM BTP, 10 lM EDTA (included to remove any adventitious metal binding), pH 6.8. Reaction mixtures were pre-incubated at 25 °C for 5 min followed by addition of EcoDHQase (330 nM). Assays were initiated by the addition of AchDHQS (11 nM). The metal salts used were: CoCl26H2O (Sigma), BaCl22H2O (BDH), FeSO47H2O (Sigma), MgSO4H2O (May and Baker), CaCl2 (Prolabo), MnSO4H2O (Sigma), CrCl2 (Aldrich), HgCl2 (Aldrich), CdCl2 (May and Baker), NiCl26H2O (May and Baker), CuSO4H2O (May and Baker), and ZnCl2 (BDH). The free metal ion concentration in the assay solution will be 90 lM due to the presence of EDTA. Crystallisation Crystals were grown using the hanging-drop vapor diffusion method at 20 °C. The drop sizes were 2 lL, and the volume of the reservoir solution was 500 lL. The crystals were obtained in 0.1 M SPG buffer (pH 7.0), and 25% (w/v) PEG 1500 in a 1:1 ratio of 10 mg/mL AchDHQS solution and crystallisation condition in the presence of 0.3 mM NAD+. Immediately before data collection, the crystals (Fig. S5) were transferred briefly into a cryoprotectant composed of 20% (w/v) PEG 400 in the reservoir solution. Determination and refinement of structure Data sets were collected at the Australian Synchrotron using the MX2 beamline and were processed using XDS and SCALA (CCP4 suite) [27,28]. The AchDHQS crystallised in the monoclinic space group P1211 and diffracted to 2.4 Å, with the following unit cell dimensions: a  58, b  151, c  85, a = 90°, b  101° c = 90°. The monomeric structure of A. nidulans DHQS (PDB code 1NR5), after modification by CHAINSAW [29], was used as the search molecule

Identification of a plastidic dehydroquinate synthase in A. chinensis A BLAST search of the cloned EST against Swissprot and NCBI databases showed the ORF to have high homology (>80% identity) to several plant, bacterial and fungal DHQSs (Fig. S1, supplementary material). The ORF encodes a protein of 377 amino acids in length. It is truncated from the original EST (1724 nucleic acids, Fig. S2) to remove a predicted chloroplast targeting sequence of 68 amino acids [24]. Unsurprisingly AchDHQS shows the highest identity (86%) with the enzyme from Populus trichocarpa, another plant DHQS and very similar identities to proteins from other plant sources, including the enzyme from S. lycopersicum [20,33]. The lowest identity is observed with the well-characterised DHQS from A. nidulans [13]. As for most fungi A. nidulans expresses the DHQS as part of the pentafunctional AROM polypeptide, therefore it is expected to differ from the monofunctional plant enzymes. AchDHQS was overexpressed in E. coli and purified by affinity chromatography. The relative molecular mass of the purified recombinant AchDHQS was 41,097.5 Da as determined by mass spectrometry (see supplementary material), which is in close agreement with the value calculated from the sequence (41,098.2 Da). A single peak was observed by size exclusion chromatography as well as light scattering experiments, corresponding to a molecular mass of 81.0 kDa and 79.2 kDa respectively (Figs. S3 and S4). These results are consistent with recombinant AchDHQS existing as a homodimer in solution. Both A. nidulans and T. thermophilus DHQS enzymes are also reported to be homodimeric [13,16]. Thermostability measurements by DSF showed a standard melting curve for a single species, presenting a melting point of 41 °C with either Zn2+ or Co2+ present in excess. A similar stability was also recorded for the E. coli enzyme, whereas Pyrococcus furiosus DHQS (PfuDHQS) has a melting temperature of 90 °C [26]. The catalytic activity of AchDHQS was found to increase with temperature to 55 °C (Fig. 2). This upper limit on temperature tolerance may arise due to protein denaturation in accordance with the relatively low melting temperature for AchDHQS. The lower melting temperature for the protein in comparison to the temperature for maximum catalytic activity likely arises due to the different experimental protocols for these measurements. The catalytic activity is determined from the initial rates of enzyme, which is previously held at 4 °C, whereas the melting temperature is determined from the observations of structural changes associated with slow ramping of the temperature. The normal growth temperature for A. chinensis is well below 41 °C [34]. AchDHQS was found to have a pH optimum of 7.5 (Fig. 2), which is in accordance with the reported pH optima of 7.0 for PfuDHQS [26] and 8.0 for EcoDHQS [10]. The enzyme exhibits standard Michaelis–Menten kinetics. The kinetic constants were KM 1.3 ± 0.1 lM and kcat 20.1 ± 0.7 s1, at 25 °C and pH 7.5. In comparison to DHQS enzymes from other sources including M. tuberculosis (KM 6.3 ± 1.1 lM and kcat 0.63 ± 0.03 s1) [35], E. coli (KM 6.3 ± 0.2 lM and kcat 16.0 ± 0.2 s1, at pH 6.8 and 25 °C) [26], P. furiosus (KM 3.7 ± 0.2 lM and kcat 3.0 ± 0.1 s1, at pH 6.8 and 60 °C) [26], B. subtilis (KM 132 lM at pH 6.5 and 37 °C) [36], Neurospora crassa (KM 1.4 lM and kcat 6.9 s1, at pH 7 and 25 °C) [6] and A. nidulans (KM 9.3 lM and kcat 5.1 s1, pH 7 and 25 °C) [37], AchDHQS has a

188

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

Fig. 2. AchDHQS activity (A) pH optimum of AchDHQS, (B) Effect of temperature on specific activity of AchDHQS.

slightly lower or similar Michaelis constant for DAH7P and turnover rate. Limited comparisons can be made to other plant DHQSs, for which only specific activities are reported for the enzymes from V. mungo (48.6 U/mg) [18], P. sativum (0.59 U/mg) [38] and the partially purified Sorghum seedling extract (5.5  107 U/mg) [39]. The corresponding specific activity of the recombinant AchDHQS was 29.9 ± 0.9 U/mg (25 °C, pH 7.5). AchDHQS required the presence of NAD+ for catalytic activity. In contrast to B. subtilis DHQS [36], NADP+ was not able to activate AchDHQS, which was also found previously for Sorghum DHQS [39]. In accordance with other characterised DHQSs, the activity of AchDHQS was found to require a divalent metal ion (Fig. 3). The highest activity was observed in the presence of Fe2+, although Co2+ was also found to be a significant activator of the enzyme. Co2+ has been reported to activate DHQS from microbial as well as plant sources [33,39]. Metal free enzyme preparations from M. tuberculosis DHQS were reactivated to 94% of the initial activity by the addition of 1 mM Co2+, and the highest rate of reactivation for the E. coli enzyme was also found with Co2+ [25,35]. On the other hand as for the E. coli and M. tuberculosis enzymes, Zn2+, which is reported to activate N. crassa DHQS [6], demonstrated limited ability to activate AchDHQS. DHQS activity of V. mungo and Sorghum preparations was shown to increase with CoCl2 at concentrations of 0.1 mM and 0.5 mM respectively, although V. mungo

DHQS reaches the same level of activity with only 0.01 mM Cu2+ and low concentrations of Cu2+ were found in the natively expressed enzyme [18].

Crystal structure of AchDHQS The structure of AchDHQS was solved by molecular replacement using the monomer of a single chain of the A. nidulans DHQS (PDB code 1NR5) as a search model. The resulting structure was then reTable 1 Crystal parameters, data collection and refinement statistics. AchDHQS A. Data collection Crystal system; space group Unit cell parameters (Å) a, b, c, a, b, c Resolution range (Å) Measurements Unique reflections Redundancy Completeness (%) I/r(I) Rmerge Wilson B-value (Å2) B. Refinement Resolution (Å) Rcryst Rfree Chain length Observed number of residues Water molecules Other

Fig. 3. Activation of EDTA-treated AchDHQS by various divalent metal ions relative to cobalt.

P 1 21 1 58.06, 150.97, 84.75, 90.00, 101.25, 90.00 50.32–2.40 (2.53–2.40) 246710 55631 4.4 (4.5) 99.6 (99.6) 9.2 (3.1) 0.122 (0.422) 33.4 2.40 0.188 0.252 1460 (11038 atoms) 181 4 NAD+, 14 PO43-, 6 glycine, 1 octaethylene glycol, 1 tetraethylene glycol, 1 triethylene glycol, 5 di(hydroxyethyl)ether

Mean B (Å2) Protein Water Other

33.7 27.3 32.4

r.m.s.d. from target values Bond lengths (Å) Bond angles (°) Dihedral angles (°)

0.01 1.66 6.25

Ramachandran Preferred (%) Allowed (%) Outliers (%) PDB entry

97.3 2.4 0.3 3ZOK

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

fined at 2.4 Å resolution, to a final R factor of 0.19 and Rfree of 0.25 (Table 1).

189

The final model for AchDHQS comprises 11038 protein atoms from the four molecules (A, B, C and D) in the asymmetric unit of

Fig. 4. Structure of AchDHQS. (a) AchDHQS dimer, with one NAD+ and three phosphate molecules bound to each monomer, (b) NAD+ binding site, (c) Stereo view Ca-overlay of the monomers of AchDHQS (yellow) and AniDHQS (1DQS, closed, green, N-terminal domains aligned); H1: active site b-hairpin, L1 + L2: flexible active site loops, (d) Stereo view Ca-overlay of AchDHQS (yellow) and AniDHQS (1NRX, open, magenta, N-terminal domains aligned), (e) Stereo view active site overlay of AchDHQS (yellow), ligands: glycine (cyan) and two free phosphates, and AniDHQS (1DQS, closed, green), ligands: carbaphosphonate (magenta) and Zn2+ (grey), NAD+ is shown in the background. Labels correspond to the AchDHQS protein. Arg1340 is contributed by the adjacent monomer. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

190

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

the crystal, together with four molecules of NAD+, 14 molecules of phosphate, six molecules of glycine, eight molecules of polyethylene glycol of various chain length, and 181 water molecules. All of these compounds are present in the crystallisation condition. In all molecules, no electron density is visible for the first 13 residues including the remaining glycine from the cleaved affinity tag at position 0. All four molecules are well restrained to standard bond distances and angles, and their main-chain torsion angles correspond well with expected values; 1301 of non-glycine residues are in the most favoured regions of the Ramachandran plot, as defined in MoLProbity, with a single outlier (G252) in all four molecules. All four molecules of the asymmetric unit have the same overall fold and have similar conformations (RMSD for Ca range between 0.27 and 0.31 Å for each pairwise chain comparison). Each molecule consists of two domains, an N-terminal domain that includes a Rossmann-type fold, and a C-terminal domain (Fig. 4a). This architecture closely resembles that found for DHQSs from other organisms (Table S1). Molecule pairs (A and B, and C and D) associate to form two homodimers, which share very close overall structural similarity (RMSD of Ca atoms is 0.26 Å). The dimer is formed by associations between N-terminal domains and buries 12.5% of the surface area of each molecule as calculated by PDBe PISA [40]. A single NAD+ is bound in the groove between the N- and C-terminal domains in each of the four molecules. The binding mode for this ligand is very similar to that observed for other DHQSs. Residues of the conserved GGXXXD motif (residues 118–123) make hydrogen-bonding contacts to the NAD+ phosphate ester functionalities and to the nicotinamide ring amide. Other NAD+-binding site residues include Asn54, Asp85, Glu87, Lys90, Thr143, Thr144, Lys156, Lys165, Asn166, Thr183, Thr186 and Lys240 of which all residues except Asp85 and Thr183 are conserved between DHQSs from different sources (Fig. S1). Asp85 and Thr183 form interactions with NAD+ via their main chain amide functionalities, interacting with the ribose and adenine moieties respectively, and these interactions are conserved for the currently known DHQS structures. Two phosphate molecules and one glycine molecule are found near the nicotinamide ring of NAD+ in all four monomers. One of the phosphates is found in an analogous position to the binding site of the phosphate group of the high affinity inhibitor carbaphosphonate (a non-hydrolysable DAH7P analogue), in the active site of the DHQS from A. nidulans [13]. The carboxylate of the free glycine molecule found in the AchDHQS active site overlays very well with the carboxy group of the carbaphosphonate molecule in AniDHQS, showing the importance of this moiety for the binding and orientation of the substrate DAH7P. All residues forming the active site are highly conserved between DHQS from all analysed species (Fig. S1) including residue Arg134, which is recruited into the active site of the adjacent monomer. Three side chains (Glu 198, His261 and His278) form the potential coordination site for the metal ion and a pair of molecular tweezers is formed on either side of the substrate binding pocket involving residues Lys156 and Asn166 (bound to phosphate), and Arg254 and Asn258 (bound to glycine) respectively. The same side chain arrangement is found in the closed AniDHQS structures where it holds the substrate analogue in place (Fig. 4e). Although a divalent metal ion is essential for the activity of AchDHQS, no metal ion was observed in the active site. There are other examples of metal free DHQS structures reported [14,16] and given the absence of any divalent metal ions in the crystallisation condition for AchDHQS, this finding is unsurprising. While the only metal found in all available DHQS structures to date is Zn2+ [13,15,17] crystallisation of AchDHQS with either Co2+ or Zn2+ was unsuccessful.

Comparisons of the structure of AchDHQS to structures from other organisms reveal close structural resemblance (Table S1). Despite the low sequence identity, the AchDHQS structure overall was found to be most similar to the original DHQS structure from A. nidulans complexed to the substrate analogue, Zn2+ and NAD+, and to the DHQS from prokaryote S. aureus (with substrate analogue and Zn2+ present). The DHQSs from both of these sources have been shown to undergo conformational adjustment on substrate analogue binding to form a functional active site [41]. The conformation change is evident by a twist between the N- and Cterminal domains. Analysis of previously reported structures reveals that a closed conformation is associated with a smaller relative angle between the two domains (1–4°), whereas the open forms, found in the absence of substrate analogue display larger angles between the two domains (11–14°). The angle of twist calculated in an analogous manner for AchDHQS was found to be 1– 2°, clearly revealing that this plant enzyme is in the closed form. This conformation is evident by the displacement of the C-termini in a direct comparison of AchDHQS with closed and open structures from A. nidulans (Fig. 4c and d). Given the absence of substrate analogues or metal ion this conformation was unexpected, and reveals that in the absence of a DAH7P analogue this closed conformation is indeed accessible. The positioning and side chain interactions of the free phosphate and glycine molecules in the active site of AchDHQS strongly resemble those found for the phosphonate and carboxy functionalities of carbaphosphonate in AniDHQS. The binding of carbaphosphonate has been shown to cause a rearrangement of three structural elements, a b-hairpin and two flexible loops, between the open and closed form leading to the overall conformational twist of the enzyme [14]. Thus it appears that the phosphate and glycine mimic the presence of substrate in the active site and lead to the observed closed conformation for AchDHQS. While the closed form of DHQS has not been observed when the active site is not fully occupied, this observation for AchDHQS may explain some of the unusual biochemical features of this enzyme. It has previously been demonstrated that 3-fluoroDAH7P is converted into two reaction products catalysed by EcoDHQS, the expected 3-fluorodehydroquinate and the 1-epi-3-fluorodehydroquinate [42,43]. This latter product is thought to be formed by premature release of the reaction intermediate from the enzyme. In contrast to EcoDQS, AchDHQS produces only a very small amount of the epimeric product. Thus, the conformational restrictions that favour the closed conformation for AchDHQS, may also prevent premature substrate release.

Conclusion This is one of very few reports on the structure of an enzyme purified from kiwifruit, and the first report on the structure of a plant DHQS. This study demonstrates that despite low sequence identity the plant DHQS adopts a similar structure and has similar properties to DHQS enzymes from microbial sources. One of the novel aspects of kiwifruit is the presence of a high level of quinic acid (0.4–1% w/w) which contributes to the flavour, and sugar/acid balance of the fruit. DHQS is highly expressed in A. chinensis at times when fruit are accumulating quinic acid (early fruit development) and expression is also high at later stages when it is assumed fruit are converting dehydroquinate to shikimate for onwards metabolism [21]. The shikimate pathway is an important contributor to secondary metabolism in kiwifruit which have been shown to be high in folate, chlorogenic acid, phenolic acids, and Vitamin K [44].

G. Mittelstädt et al. / Archives of Biochemistry and Biophysics 537 (2013) 185–191

Acknowledgment This work was funded by the New Zealand Foundation for Research, Science and technology (CO6X0403). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.abb.2013.07.022. References [1] R. Bentley, Crit. Rev. Biochem. Mol. Biol. 25 (1990) 307–384. [2] K.M. Herrmann, L.M. Weaver, Rev. Plant Physiol. Plant Mol. Biol 50 (1999) 473– 503. [3] K.M. Herrmann, Plant Cell 7 (1995) 907–919. [4] J.R. Coggins, M.R. Boocock, M.S. Campbell, S. Chaudhuri, J.M. Lambert, A. Lewendon, D.M. Mousdale, D.D.S. Smith, Biochem. Soc. Trans. 13 (1985) 299– 303. [5] D.D.S. Smith, J.R. Coggins, Biochem. J. 213 (1983) 405–415. [6] J.M. Lambert, M.R. Boocock, J.R. Coggins, Biochem. J. 226 (1985) 817–829. [7] K. Duncan, R.M. Edwards, J.R. Coggins, Biochem. J. 246 (1987) 375–386. [8] L.D. Graham, F.M. Gillies, J.R. Coggins, Biochim. Biophys. Acta 1216 (1993) 417–424. [9] T. Widlanski, S.L. Bender, J.R. Knowles, Biochemistry 28 (1989) 7572–7582. [10] S.L. Bender, T. Widlanski, J.R. Knowles, Biochemistry 28 (1989) 7560–7572. [11] P.A. Bartlett, K. Satake, J. Am. Chem. Soc. 110 (1988) 1628–1630. [12] J.D. Moore, J.R. Coggins, R. Virden, A.R. Hawkins, Biochem. J. 301 (1994) 297. [13] E.P. Carpenter, A.R. Hawkins, J.W. Frost, K.A. Brown, Nature 394 (1998) 299– 302. [14] C.E. Nichols, J. Ren, H.K. Lamb, A.R. Hawkins, D.K. Stammers, J. Mol. Biol. 327 (2003) 129–144. [15] C.E. Nichols, J. Ren, K. Leslie, B. Dhaliwal, M. Lockyer, I. Charles, A.R. Hawkins, D.K. Stammers, J. Mol. Biol. 343 (2004) 533–546. [16] M. Sugahara, Y. Nodake, M. Sugahara, N. Kunishima, Proteins: Struct., Funct., Bioinf. 58 (2004) 249–252. [17] J.-S. Liu, W.-C. Cheng, H.-J. Wang, Y.-C. Chen, W.-C. Wang, Biochem. Biophys. Res. Commun. 373 (2008) 1–7. [18] E. Yamamoto, Phytochemistry 19 (1980) 779–781. [19] R.K. Deka, I.A. Anton, B. Dunbar, J.R. Coggins, FEBS Lett. 349 (1994) 397–402.

191

[20] M. Bischoff, J. Rösler, H.R. Raesecke, J. Görlach, N. Amrhein, J. Schmid, Plant Mol. Biol. 31 (1996) 69–76. [21] K.B. Marsh, H.L. Boldingh, R.S. Shilton, W.A. Laing, Funct. Plant Biol. 36 (2009) 463. [22] R.N. Crowhurst, A.P. Gleave, E.A. MacRae, C. Ampomah-Dwamena, R.G. Atkinson, L.L. Beuning, S.M. Bulley, D. Chagne, K.B. Marsh, A.J. Matich, M. Montefiori, R.D. Newcomb, R.J. Schaffer, B. Usadel, A.C. Allan, H.L. Boldingh, J.H. Bowen, M.W. Davy, R. Eckloff, A.R. Ferguson, L.G. Fraser, E. Gera, R.P. Hellens, B.J. Janssen, K. Klages, K.R. Lo, R.M. MacDiarmid, B. Nain, M.A. McNeilage, M. Rassam, A.C. Richardson, E.H. Rikkerink, G.S. Ross, R. Schröder, K.C. Snowden, E.J. Souleyre, M.D. Templeton, E.F. Walton, D. Wang, M.Y. Wang, Y.Y. Wang, M. Wood, R. Wu, Y.-K. Yauk, W.A. Laing, BMC Gen. 9 (2008) 351. [23] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990) 403–410. [24] O. Emanuelsson, S. Brunak, G. von Heijne, H. Nielsen, Nature Protocols 2 (2007) 953–971. [25] S.L. Bender, S. Mehdi, J.R. Knowles, Biochemistry 28 (1989) 7555–7560. [26] L. Negron, M.L. Patchett, E.J. Parker, Enzyme Res. 134893 (2011) 134810. [27] W. Kabsch, Acta Crystallogr. Sect. D: Biol. Crystallogr. 66 (2010) 125–132. [28] P. Evans, Sect. D: Biol. Crystallogr. D62 (2006) 72–82. [29] N. Stein, J. Appl. Crystallogr. 41 (2008) 641–643. [30] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.J. Read, J. Appl. Crystallogr. 40 (2007) 658–674. [31] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D: Biol. Cryst. 60 (2004) 2126– 2132. [32] J.J. Arendall III, D.A. Headd, R.M. Keedy, G.J. Immormino, L.W. Kapral, J.S. Murray, D.C. Richardson, Acta Crystallogr. Sect. D: Biol. Crystallogr. 66 (2010) 12–21. [33] H. Maeda, N. Dudareva, Ann. Rev. Plant Biol. 63 (2012) 73–105. [34] S.M. Lionakis, W.W. Schwabe, Ann. Bot. 54 (1984) 485–501. [35] J.D. de Mendonça, O. Adachi, L.A. Rosado, R.G. Ducati, D.S. Santos, L.A. Basso, Mol. BioSyst. 7 (2011) 119. [36] N. Hasan, E.W. Nester, J. Biol. Chem. 253 (1978) 4999–5004. [37] A. Park, H.K. Lamb, C. Nichols, J.D. Moore, K.A. Brown, A. Cooper, I.G. Charles, D.K. Stammers, A.R. Hawkins, Protein Science 13 (2004) 2108–2119. [38] D.L. Pompliano, L.M. Reimer, S. Myrvold, J.W. Frost, J. Am. Chem. Soc. 111 (1989) 1866–1871. [39] R. Saijo, T. Kosuge, Phytochemistry 17 (1978) 223–225. [40] E. Krissinel, K. Henrick, J. Mol. Biol. 372 (2007) 774–797. [41] K.A. Brown, E.P. Carpenter, K.A. Watson, J.R. Coggins, A.R. Hawkins, M.H.J. Koch, D.I. Svergun, Biochem. Soc. Trans. 31 (2003) 543–547. [42] E.J. Parker, J.R. Coggins, C. Abell, J. Org. Chem. 62 (1997) 8582–8585. [43] L. Negron, E.J. Parker, Org. Biomol. Chem. 9 (2011) 2861–2867. [44] USDA National Database Release 25, , 2011.