Identification of Polyketide Inhibitors Targeting 3-Dehydroquinate Dehydratase in the Shikimate Pathway of Enterococcus faecalis

Due to the emergence of resistance toward current antibiotics, there is a pressing need to develop the next generation of antibiotics as therapeutics against infectious and opportunistic diseases of microbial origins. The shikimate pathway is exclusive to microbes, plants and fungi, and hence is an attractive and logical target for development of antimicrobial therapeutics. The Gram-positive commensal microbe, Enterococcus faecalis, is a major human pathogen associated with nosocomial infections and resistance to vancomycin, the “drug of last resort”. Here, we report the identification of several polyketide-based inhibitors against the E. faecalis shikimate pathway enzyme, 3-dehydroquinate dehydratase (DHQase). In particular, marein, a flavonoid polyketide, both inhibited DHQase and retarded the growth of Enterococcus faecalis. The purification, crystallization and structural resolution of recombinant DHQase from E. faecalis (at 2.2 Å resolution) are also reported. This study provides a route in the development of polyketide-based antimicrobial inhibitors targeting the shikimate pathway of the human pathogen E. faecalis.


Introduction
The shikimate pathway comprises seven catalytic steps that convert the metabolic substrates D-erythrose 4-phosphate and phosphoenolpyruvate to chorismate in all microorganisms, including common enteric bacteria such as Escherichia coli and pathogenic microbes such as the vancomycin-resistant Enterococci (VRE) [1]. Chorismate is a necessary precursor in the biosynthesis of folate, aromatic amino acids and ubiquinone. The pathway is exclusive and crucial for microorganisms, plants and fungi.
The recent emergence of multi-drug resistant pathogenic microbes demonstrates a pressing need to develop new antibiotics. The absence of the shikimate pathway in humans presents an attractive target in the development of antimicrobials. Numerous studies have attempted to target enzymes in the shikimate pathway [2]. Currently, N-phosphomethylglycine is the only commercially available compound that targets one of the enzymes in the pathway; it targets 5-enolpyruvate shikimate-3-phosphate synthase [3,4,5].
3-Dehydroquinate dehydratase (DHQase) is the third enzyme in the shikimate pathway. DHQase catalyzes the dehydration of 3-dehydroquinate to 3-dehydroshikimate ( Figure 1). There are two types of DHQase: type I enzymes catalyze a Schiff base mechanism using a catalytic lysine residue; type II DHQase catalyze the dehydration reaction via an enolate intermediate. DHQase from Enterococcus faecalis is a type I enzyme. Other organisms that have type I DHQases include Eschericia coli [6], Salmonella typhi [7] and Shigella dysenteriae [8].
Flavonoids are produced by plants as secondary metabolites, and include polyketide-based natural compounds. Polyketides are a large class of biomolecules that are naturally produced by bacteria, fungi and plants, and include many clinically important biomolecules with anti-cancer, anti-microbial, anti-oxidant and anti-inflammatory activities. Currently, there are about 3000 flavonoids extracted and characterized from plants [9]. The main function of these compounds is to provide colors attractive to plant pollinators [10]. Other physiological functions include defense against fungal pathogens and UV-B radiation [10]. They are also involved in energy transfer, growth regulation, respiration, photosynthesis, morphogenesis, and sex determination [10]. The basic structural feature of flavonoids is the 2-phenyl-benzo[a]pyr-ane or flavane nucleus, which consists of two benzene rings (A and B) linked through a heterocyclic pyrane ring (C) (Figure 2).
Herein, we report the identification of various flavonoids as potential inhibitors for DHQase from E. faecalis (efDHQase). The study also elucidated the structure of E. faecalis DHQase to a resolution of 2.2 Å . This study provides significant biochemical and structural information that will facilitate the future development of polyketide-based antimicrobial inhibitors targeting the shikimate pathway of the nosocomial pathogen E. faecalis.

Materials and Methods
Cloning, expression and purification of 3-dehydroquinate dehydratase from Enterococcus faecalis (efDHQase) The aroD gene encoding 3-dehydroquinate dehydratase (efDHQase, 3-dehydroquinate dehydratase from Enterococcus faecalis V583 strain) (GI: 29376281) was amplified via PCR from genomic DNA isolated from E. faecalis V583 strain using Platinum Pfx DNA polymerase (Invitrogen). The PCR mixture (100 mL) contained 1 ng of plasmid DNA, 10 mL of 106 Pfx amplifi cation buffer, 1 mM MgSO 4 , dNTPs (0.4 mM each), 40 pmol of each primer (forward primer 59-GAGGAGGGCG-CATATGAGCAACCATCTTTTCGACG-39 and reverse primer 59-GAGAGCGCGGGATCCTTACGTCCTGGTATAAA-GATC-39), and 5 units of Platinum Pfx DNA polymerase. The gene was amplified using a PTC-0200G Thermal Cycler (Bio-Rad Laboratories), with the following parameters: 94uC for 2 min followed by 40 cycles of 94uC for 1 min, 55uC for 1 min and 15 s, and 68uC for 3 min, and a final extension of 68uC for 10 min. The amplified gene was cloned into a modified pET-15b vector (Novagen) in which the N-terminus contained 10 His residues (kindly provided by Professor John Gerlt, University of Illinois, Urbana, IL) [12]. The protein was expressed in E. coli aroD 2 pTara, an auxotroph aroD 2 negative E. coli mutant strain in which the aroD gene was deleted from the genome. Transformed cells were grown at 37uC in LB broth (supplemented with 100 mg/ mL of ampicillin, 15 mg/mL of chloramphenicol and 50 mg/mL of kanamycin) to an OD 600 of 0.6, and IPTG (0.1 mM) was added to induce protein expression for 16 h. The cells were harvested by centrifugation and resuspended in binding buffer [5 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl (pH 7.9)] and lysed by sonication. The lysate was clarified by centrifugation, and the Histagged protein was purified using a column of chelating Sepharose Fast Flow (GE Healthcare Bio-Sciences Corp.) charged with Ni 2+ ion. The cell lysate was applied to the column in binding buffer, washed with buffer containing 154 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9, and eluted with 100 mM Lhistidine, 0.5 M NaCl, and 20 mM Tris-HCl, pH 7.9. The Nterminal His tag was removed with thrombin (GE Healthcare Bio-Sciences Corp.) according to the manufacturer's instructions, and the proteins were purified to homogeneity on a Q Sepharose High Performance column (GE Healthcare Bio-Sciences Corp.) equilibrated with binding buffer [25 mM Tris-HCl, pH 7.9] and eluted with a linear gradient from 0 to 0.5 M elution buffer [1 M NaCl and 25 mM Tris-HCl, pH 7.9].
Cloning, expression and purification of shikimate dehydrogenase from Enterococcus faecalis (efSHD) The aroE gene encoding shikimate dehydrogenase (efSHD) (GI: 29343586) was amplified via PCR from genomic DNA isolated from E. faecalis V583 strain using Platinum Pfx DNA polymerase (Invitrogen). The PCR mixture (100 mL) contained 1 ng of plasmid DNA, 10 mL of 106 Pfx amplification buffer, 1 mM   GATCCTTATCTATTTTCAATTTTACGTTTTACG-39), and 5 units of Platinum Pfx DNA polymerase. The gene was amplified using a PTC-0200G Thermal Cycler (Bio-Rad Laboratories), with the following parameters: 94uC for 2 min followed by 40 cycles of 94uC for 1 min, 55uC for 1 min and 15 s, and 68uC for 3 min, and a final extension of 68uC for 10 min. The amplified gene was cloned into the modified pET-15b vector (Novagen) [12]. The protein was expressed in E. coli aroE 2 pTara, an auxotroph aroE 2 negative E. coli mutant strain in which the aroE gene was deleted from the genome. Transformed cells were grown at 37uC in LB broth (supplemented with 100 mg/mL of ampicillin, 15 mg/ mL of chloramphenicol and 50 mg/mL of kanamycin) to an   OD 600 of 0.6, and IPTG (0.1 mM) was added to induce protein expression for 16 h. efSHD was purified from the harvested cells as previously described for efDHQase.

Crystallization and structure solution of efDHQase
Freshly prepared efDHQase was concentrated to 30 mg/mL. Crystals grew within 5 days at 25uC by the sitting-drop vapor diffusion method from a 1:1 (v/v) mixture with precipitant solution comprised of 100 mM HEPES sodium salt, pH 7.5, 30% (w/v) PEG 4000, and 0.2 M calcium chloride. Crystals were soaked in the precipitant solution supplemented with 20% (v/v) glycerol, and flash frozen in liquid nitrogen. X-ray diffraction data were collected at 100 K on a Bruker X8 PROTEUM system consisting of a MICROSTAR micro-focus X-ray generator, a PLATI-NUM135 CCD detector, and a 4-circle KAPPA goniometer. Data reduction was carried out using SAINT, SADABS, and XPREP, within the Bruker Proteum2 program suite [13,14].
Structural determination of efDHQase was initiated by molecular replacement using the structure of dehydroquinate dehydratase from Clostridium difficile (PDB ID: 3JS3, chain A) as the model. The model, together with the processed data and the amino acid sequence of efDHQase, was fed into MrBUMP [15], an automated molecular replacement pipeline in the CCP4 suite [16]. CHAINSAW [17], PHASER [18] and REFMAC5 [19] were used within this pipeline to automatically trim the model, search for its correct rotation and translation, and refine it, respectively. The solution was then subjected to five rounds of automated model building using ARP/wARP [20], followed by restrained refinement in PHENIX [21] and manual building in COOT [22]. TLS refinement was included in the later rounds of PHENIX refinement. All the structure-related figures were prepared with the PyMOL Molecular Graphics System (DeLano Scientific LLC). Data collection and refinement statistics are listed in Table 1. Attempts at obtaining inhibitor-liganded structures of efDHQase were unsuccessful.

Kinetic parameters of efDHQase
The activity of efDHQase was measured using a continuous coupled-enzyme spectrophotometric assay monitored at 25uC. The rate of dehydroshikimate formation was coupled to the oxidation of NADPH (340 nm, e = 6220 M 21 cm 21 ) (Figure 3). The 200 mL assay contained 50 mM HEPES buffer, pH 7.5, 0.25 mM NADPH, 0.25 mM 3-dehydroquinate, 10 nM DEF and 10 nM shikimate dehydrogenase (efSHD). The kinetic parameters were determined by fitting the initial rates to the Michaelis-Menten equation.
The efDHQase coupled-enzyme assay was used to identify potential inhibitors from the library. Flavonoids that decreased the rate of dehydroshikimate formation by at least 50% were considered as potential efDHQase or efSHD inhibitors.
Flavonoids that inhibited the efDHQase coupled-enzyme assay may inhibit efSHD instead of efDHQase. Thus, efSHD was assayed independently. efSHD was assayed using a continuous coupled-enzyme spectrophotometric assay (Figure 4) monitored at 25uC. The rate of 3-dehydroshikimate formation was coupled to the reduction of p-iodonitrotetrazolium (INT) (500 nm, e = 17800 M 21 cm 21 ). The 200 mL assay contained 50 mM HEPES buffer, pH 7.5, 0.15 mM NADP + , 1.5 mM shikimate, 5 units of diaphorase (Sigma), 1 mM INT, and 10 nM efSHD. Flavonoids that did not affect the rate of dehydroshikimate formation were considered as potential efDHQase inhibitors.

Determination of minimal inhibition concentration against E. faecalis
A minimal inhibition concentration (MIC) assay using the flavonoid library was performed to screen for compounds targeting the growth of E. faecalis. The assay was based on the standard guidelines from the National Committee for Clinical Laboratory Standards (NCCLS), sixth edition. Briefly, turbidity-adjusted bacterial suspensions from fresh overnight cultures were added to LB broth to conform to an inoculation density of 5610 5 cfu/ mL. Each 100 mL bacterial culture contained 2 mM flavonoid and was incubated for 17 h at 37uC. After incubation, the cell density at OD 600 was measured for each culture. If the cell density of the flavonoid containing culture is less than 50% compared to the cultures that did not have any flavonoid, these flavonoids were chosen as potential inhibitors.
Flavonoids used at 2 mM that inhibited E. faecalis growth by at least 50% were further tested. The concentration of each flavonoid was continually decreased until the lowest concentration of the compound that inhibited E. faecalis growth by at least 50% was obtained. This concentration is calculated to be the MIC for each flavonoid.

Results and Discussion
Crystal structure of efDHQase The crystal structure of efDHQase was solved using the Protein Crystallography X-Ray Facility hosted by the Institute of Molecular and Cell Biology, A*STAR Singapore. The efDHQase monomer has a (b/a) 8 -barrel scaffold ( Figure 5). The majority of the molecule is well ordered and with clear density in the electron density map, with the exceptions of the first and last two residues at the N-and C-terminus, respectively, and a disordered portion (Ala228-Gln236) within the loop connecting the 8th b-strand and the 8th a-helix (b8-a8 loop). There are two monomers in the asymmetric unit of the crystal, forming a homodimer. Using the Protein Interfaces, Surfaces and Assemblies (PISA) service at the European Bioinformatics Institute (https://www.ebi.ac.uk/  msd-srv/prot_int/cgi-bin/piserver) [23], the interface area between the two monomers (colored purple in Figure 5) was calculated to be 990.5 Å , or about 9% of one monomer's solvent-accessible surface area. On each monomer, there are a total of 94 atoms belonging to 24 residues that are involved in the formation of the dimer interface. The non-crystallographic symmetry (NCS) differences for most of the residues are well below 0.3 Å (as reported by COOT [22]), indicating close conformational similarity between the two protomers.
To gain further insight into efDHQase's enzymatic activity, we compared the efDHQase structure with homologous DHQase structures from Salmonella enterica (seDHQase) and Clostridium difficile (cdDHQase). These homologous enzymes share about 46% sequence identity with efDHQase, and have been characterized structurally for their ligand binding and Schiff base formation. The magnified view in Figure 5 shows the overlaid active centers of efDHQase and its homologs, seDHQase (PDB ID: 3M7W) and cdDHQase (PDB ID: 3JS3), that have been crystallized in the presence of covalently-bound pre-and postdehydration intermediates, respectively [24]. Key residues, which are responsible for ligand coordination in the homologous structures, are well conserved in efDHQase (sticks representation). Moreover, these residues, and in particular Lys170, are positioned in nearly identically positions in these three homologous structures, despite the absence of bound ligand in the efDHQase structure. Such similarities strongly suggest that efDHQase catalyzes the dehydration reaction via a Schiff base mechanism whereby the terminal amine of efDHQase's Lys170 reversibly reacts with the carbonyl of the substrate [24,25,26,27]. To reinforce this notion, the pre-dehydration covalent intermediate bound to seDHQase is also shown in Figure 5 (green sticks), with its carbonyl carbon in close proximity to efDHQase's Lys170. Among the key ligand-binding residues of the three overlaid structures, the side chain of His143 exhibits the largest conformational deviations, which is in line with the hypothesis that this residue utilizes its conformational flexibility to exert dual catalytic roles [24]. In the absence of bound ligand, the b8-a8 loop of efDHQase adopts the ''open'' conformation commonly seen in other DHQases [26,27], and becomes partially disordered. To model the ''closed'' conformation of the b8-a8 loop, the undefined portion of the loop was taken from seDHQase and is shown in Figure 5 (light green). Since the residues in this portion are highly  conserved among DHQases [27], the model is likely to have good credibility.

Kinetic parameters of efDHQase
The turnover number (k cat ) of efDHQase is 110 s 21 and the Michaelis constant (K M ) for 3-dehydroquinate is 65 mM. This is comparable to DHQases from other type I DHQases such as Escherichia coli [28], Salmonella typhi [7], and Streptococcus pneumoniae [29] characterized under similar conditions ( Table 2).

Flavonoid inhibitors that inhibit efDHQase and/or E. faecalis growth
From a library of 139 commercially available flavonoid compounds, nine efDHQase inhibitors were found ( Table 3). The flavonoid inhibitors exhibited either mixed or uncompetitive inhibition. Representative double-reciprocal plots to determine the type of inhibition are shown in Figure 6. The K I 's that were determined ranged from 20-350 mM, which is in the range of 3dehydroquinate's Michaelis constant (65 mM).
Compounds that showed mixed inhibition are usually a mixture of competitive and non-competitive inhibition ( Figure 6A and 6B, respectively). We propose that these compounds bind to either the enzyme or enzyme-substrate complex but with higher affinity for the former [30]. Some flavonoids showed uncompetitive inhibition ( Figure 6C). We propose that these inhibitors only bind to the enzyme-substrate complex and delay the time for the substrate and product to leave the active site, thus causing a decrease in V max [30].
Four compounds were identified as potential inhibitors with moderate MIC values, ranging from 500-1000 mM ( Table 4).
The 139-compound library may be classified into flavonoids and its derivatives (anthocyanidin, chalcone, flavan, flavone, flavanone, isoflavone), aromatic acids (benzoic acids and cinnamic acids), and coumarins. Some of the flavonoids and coumarins are glycosides, with one or two sugar units. The compounds were chosen based on previous antimicrobial studies [11,31,32] From this library, six flavonoid glycosides inhibited efDHQase while two flavonoids inhibited the growth of E. facaelis. The six flavonoid glycosides can be classified further based on their aglycone -a) gossypin and rutin are quercetin glycosides; b) homoorientin, luteolin-39,7-diglucoside, and luteolin-7-glucoside are luteolin glycosides; and, c) rhoifolin is an apigenin glucoside. None of the six flavonoid glycosides inhibited the growth of E. facaelis. Two flavonoids, datiscetin and naringenin, inhibited the growth of E. faecalis. Datiscetin, a flavonol from Datisca cannabina (hemp), was previously found to inhibit the growth of methicillin-resistant S. aureus (MRSA) (MIC = 512 mg/mL) [33]. Naringenin, in combination with other flavonoids extracted from Centaurea alexanderina, was found to inhibit the growth of P. aeruginosa [34].
These observations are consistent with previous studies that found flavonoids inhibited bacterial growth but not their corresponding glycosides. Rauha found that quercetin inhibited the growth of Gram-positive bacteria such as Bacillus subtilis, Micrococcus luteus, Staphylococcus aureus and S. epidermidis, but not quercetin glycosides such as rutin [35]. Funayama also showed that flavonoids without the sugar moiety have antimicrobial activity on B. subtilis, S. aureus, and M. luteus, but their corresponding glycosides did not [32]. Pereira found that a combination of flavonoid glycosides is needed to inhibit bacterial growth; extracts containing luteolin-7-glucoside, rutin, apigenin-7glucoside and luteolin-49-glucoside from Olea europaea L. (olive tree) were found to inhibit a range of Gram-positive bacteria such as B. cereus, B. subtilis, and S. aureus [36].  Coumarins are a large class of phenolic substances made of fused benzene and a-pyrone rings, which constitute the core skeleton of many flavonoid compounds. The library contained eight coumarins (Figure 7) but only 7-hydroxycoumarin and aesculin showed significant inhibition against efDHQase. It is possible that small perturbations to the coumarin structure greatly affected the binding to efDHQase. Both compounds did not inhibit the growth of E. facaelis. 7-Hydroxycoumarin was previously found to have antimicrobial properties against Paecilomyces fulva, a plant pathogen that causes Byssochlamys rot on strawberries [37]. Aesculin was previously shown not to have antimicrobial activity against S. aureus, E.coli, and P. aeruginosa [38].
A review by Venugopala postulated that long chain hydrocarbon substitutions on coumarins are needed for antibacterial activity [39]. Compounds such as ammoresinol and ostruthin showed activity against Gram-positive bacteria such as B. megaterium, M. luteus, M. lysodeikticus, and S. aureus [40]. Anthogenol, from green fruits of Aegle marmelos, showed activity against Enterococcus [39]. Chartreusin and its derivatives were found to be active against methicillin-resistant S. aureus (MRSA), vancomycin-resistant Enterococcus (VRE), and Mycobacterium vaccae but these compounds are highly toxic [41].
Chalcones are open-chain flavonoids with a,b-unsaturated carbonyl groups and is one of the important compound groups of flavonoids derived from nature. Out of the eight chalcones in the library, two of them, phloretin and marein, inhibited the growth of E. facaelis (MIC#512 mg/mL). Phloretin, a chalcone found in leaves of Malus domestica (apple), is known to have antitumor and antioxidant properties. A study by Osório found that it inhibited fourteen strains of MRSA (MIC = 250 mg/mL) [42].
Marein is the only compound that inhibited both efDHQase activity and E. facaelis growth. It is a chalcone glucoside found on Coreopsis maritime, functioning as one of the anthochlor pigments of this flower. It was recently found that it inhibited histone deacetylase enzymes and NF-kB, a protein complex that controls the transcription of DNA, in the 100 mM range [43]. It was also observed to have antihyperglycemic activity [44].
Various studies have shown that although chalcones weakly inhibits bacterial growth, adding conventional antibiotics greatly increases their antimicrobial activity [45]. A study by Tran observed better antibacterial activity when a combination of chalcones and antibiotics such as ciprofloxacin and doxycycline were used [46].

Summary
This study has demonstrated the utilization of flavonoids as potential drug leads. We have observed inhibition against efDHQase activity and E. facaelis growth using flavonoids and its glycosides, chalcones, and coumarins. From a 139-compound polyketide-based library, we found that flavonoids inhibited growth while its glycosides inhibited crucial metabolic enzymes such as efDHQase. In addition, we observed that small perturbations in the coumarin structure greatly affected the ability of the polyketide-based small molecule compounds to bind to the enzyme. Lastly, chalcones inhibited bacterial growth but a synergy with more conventional antibiotics may be needed for exerting potent inhibition. We have also elucidated the structure of efDHQase. Our work provides a route in the development of polyketide-based antimicrobial inhibitors targeting the shikimate pathway of the human pathogen E. faecalis.