key: cord-321834-n5w88l23
authors: Huang, Cheng-Yang
title: Inhibition of a Putative Dihydropyrimidinase from Pseudomonas aeruginosa PAO1 by Flavonoids and Substrates of Cyclic Amidohydrolases
date: 2015-05-19
journal: PLoS One
DOI: 10.1371/journal.pone.0127634
sha: 
doc_id: 321834
cord_uid: n5w88l23

Dihydropyrimidinase is a member of the cyclic amidohydrolase family, which also includes allantoinase, dihydroorotase, hydantoinase, and imidase. These metalloenzymes possess very similar active sites and may use a similar mechanism for catalysis. However, whether the substrates and inhibitors of other cyclic amidohydrolases can inhibit dihydropyrimidinase remains unclear. This study investigated the inhibition of dihydropyrimidinase by flavonoids and substrates of other cyclic amidohydrolases. Allantoin, dihydroorotate, 5-hydantoin acetic acid, acetohydroxamate, orotic acid, and 3-amino-1,2,4-triazole could slightly inhibit dihydropyrimidinase, and the IC(50) values of these compounds were within the millimolar range. The inhibition of dihydropyrimidinase by flavonoids, such as myricetin, quercetin, kaempferol, galangin, dihydromyricetin, and myricitrin, was also investigated. Some of these compounds are known as inhibitors of allantoinase and dihydroorotase. Although the inhibitory effects of these flavonoids on dihydropyrimidinase were substrate-dependent, dihydromyricetin significantly inhibited dihydropyrimidinase with IC(50) values of 48 and 40 μM for the substrates dihydrouracil and 5-propyl-hydantoin, respectively. The results from the Lineweaver−Burk plot indicated that dihydromyricetin was a competitive inhibitor. Results from fluorescence quenching analysis indicated that dihydromyricetin could form a stable complex with dihydropyrimidinase with the K (d) value of 22.6 μM. A structural study using PatchDock showed that dihydromyricetin was docked in the active site pocket of dihydropyrimidinase, which was consistent with the findings from kinetic and fluorescence studies. This study was the first to demonstrate that naturally occurring product dihydromyricetin inhibited dihydropyrimidinase, even more than the substrate analogs (>3 orders of magnitude). These flavonols, particularly myricetin, may serve as drug leads and dirty drugs (for multiple targets) for designing compounds that target several cyclic amidohydrolases.

Thus, whether the substrate and inhibitor of each enzyme in this family competitively inhibit dihydropyrimidinase remains unclear.

Few therapies are effective against the six antibiotic-resistant ESKAPE pathogens. The increased prevalence of beta-lactamases in P. aeruginosa and other bacteria has begun to reduce the clinical efficacy of beta-lactams against the most common opportunistic pathogen [16] . To date, over 800 beta-lactamases have been identified, of which at least 120 beta-lactamases have been detected in P. aeruginosa [17] . The development of clinically useful small-molecule antibiotics and identification of new targets in microorganisms are seminal events in the field of infectious diseases [18] .

Flavonols belong to flavonoids, the most common group of plant polyphenols that is responsible for most of the flavor and color of fruits and vegetables [19] . Over 5,000 different flavonoids have been identified, many of which display structure-dependent biological and pharmacological activities [20, 21, 22] , including antimicrobial agents [23, 24] . These natural products are safe as pharmaceuticals because they have fewer side effects for human use.

In this study, we investigated the effects of the substrates and inhibitors of allantoinase and dihydroorotase, including the flavonols myricetin, quercetin, kaempferol, and galangin, on inhibiting the catalytic activity of a putative dihydropyrimidinase from P. aeruginosa PAO1. The derivatives of myricetin, namely, dihydromyricetin and myricitrin, were further used to test the structure-inhibition relationship of dihydropyrimidinase.

Construction of the dihydropyrimidinase expression plasmid PA0441, the gene encoding the putative dihydropyrimidinase, was amplified by PCR using genomic DNA of P. aeruginosa genomic DNA as the template. The forward (5 0 -CGCGGCATA TGTTTGATTTACTCCTGC-3 0 ) and the reverse (5 0 -TCGCACTCGAGAAAATCGAAGGC ATGT-3 0 ) primers were designed to introduce unique NdeI and XhoI restriction sites (underlined), permitting the insertion of the amplified gene into the pET21b vector (Novagen Inc., Madison, WI, USA). The DNA fragment was then inserted into pET21b to produce the plasmid pET21b-PaDHT for P. aeruginosa dihydropyrimidinase expression. The expected gene product expressed by pET21b-PaDHT has a C-terminal His tag, useful for purifying the recombinant protein.

The dihydropyrimidinase mutants (H59A, H61A, K150A, Y155A, H183A, H239A, S289A, D316A, and N337A) were generated according to the manufacturer's protocol (Stratagene, LaJolla, CA). The presence of the mutation was verified by DNA sequencing in each construct. The oligonucleotide primers for the preparation of mutants are shown in Table 1 .

Recombinant wild-type and mutant dihydropyrimidinases were expressed and purified using the protocol described previously for hydantoinase [12] , allantoinase [11, 15] , dihydroorotase [11, 15] , PriB [25, 26] , DnaB [27] , DnaT [28, 29] , and SSB [30, 31] . The protein purified from the soluble supernatant by Ni 2+ -affinity chromatography (HiTrap HP; GE Healthcare Bio-Sciences, Piscataway, NJ, USA) was eluted with Buffer A (20 mM Tris-HCl, 250 mM imidazole, and 0.5 M NaCl, pH 7.9) and dialyzed against a dialysis buffer (20 mM HEPES and 100 mM NaCl, pH 7.0; Buffer B). Protein purity remained >97% as determined by SDS-PAGE.

A rapid spectrophotometric assay was used to determine the enzymatic activity according to a previously described protocol for hydantoinase [12] , allantoinase [11, 15] , dihydroorotase [11, 15] , and imidase [3, 4, 5] . Dihydrouracil, 5-propyl-hydantoin, and phthalimide were used as substrates. Unless explicitly stated otherwise, dihydrouracil (2 mM) was used as the substrate in the standard assay of dihydropyrimidinase. Briefly, the decrease in absorbancy at 230, 248, and 298 nm was measured upon hydrolysis of dihydrouracil, 5-propyl-hydantoin, and phthalimide as the substrate at 25°C, respectively. To start the reaction, the purified dihydropyrimidinase (10-70 μg) was added to a 2 mL solution containing the substrate and 100 mM Tris-HCl (pH 8.0). Substrate hydrolysis was monitored with a UV/vis spectrophotometer (Hitachi U 3300, Hitachi High-Technologies, Tokyo, Japan). The extinction coefficient of each substrate was determined experimentally by direct measurement with a spectrophotometer. The extinction coefficients of dihydrouracil, 5-propyl-hydantoin, and phthalimide were 0.683 mM -1 cm -1 at 230 nm, 0.0538 mM -1 cm -1 at 248 nm, and 3.12 mM -1 cm -1 at 298 nm, respectively. The initial rates of change were a function of enzyme concentration within the absorbance range of 0.01-0.18 min -1 . A unit of activity was defined as the amount of enzyme catalyzing the hydrolysis of 1 μmol substrate/min, and the specific activity was expressed in terms of units of activity per milligram of enzyme. The kinetic parameters K m and V max were determined from a non-linear plot by fitting the hydrolyzing rate from individual experiments to the Michaelis-Menten equation (Enzyme Kinetics module of Sigma-Plot; Systat Software, Chicago, IL, USA). Dissociation constants determined by fluorescence spectrophotometer

The dissociation constants (K d ) of dihydropyrimidinase was determined using the fluorescence quenching method as previously described for allantoinase [15] , dihydroorotase [15] , and DnaB helicase [27, 32] . An aliquot of each compound was added into the solution containing dihydropyrimidinase (0.8 μM), 50 mM HEPES at pH 7.0. The decrease in intrinsic fluorescence of protein was measured at 334.5 nm upon excitation at 279 nm and 25°C with a spectrofluorimeter (Hitachi F-2700; Hitachi High-Technologies, Tokyo, Japan). At least seven data points were used to calculate each K d . Each data point was an average of 2-3 determinants, and the difference of the determinants was within 10%. The 

The amino acid sequences of 497 sequenced dihydropyrimidinase homologues were aligned using ConSurf [33] . The model of P. aeruginosa dihydropyrimidinase was built using human dihydropyrimidinase (PDB entry: 2VR2) as a template by SWISS-MODEL (http://swissmodel. expasy.org) [34] and (PS) 2 (http://140.113.239.111/~ps2v2/docs.php) [35] . The coordinate and topology file of the flavonoids was found in DrugBank (http://www.drugbank.ca/) [36] . Myricetin and dihydromyricetin were computationally docked into the three-dimensional model of dihydropyrimidinase by using PatchDock (http://bioinfo3d.cs.tau.ac.il/PatchDock/) [37] . The dihydrouracil-complexed structure model of P. aeruginosa dihydropyrimidinase was directly constructed by superimposing the crystal structure of the dihydrouracil-yeast dihydropyrimidinase complex (the coordinate of 2FVK). The structures were visualized by using the program PyMol.

Expression and purification of a putative dihydropyrimidinase from P. aeruginosa PAO1

The gene PA0441 encoding putative dihydropyrimidinase was PCR-amplified using genomic DNA of P. aeruginosa PAO1 as a template. The amplified gene was then ligated into the pET21b vector for protein expression. P. aeruginosa dihydropyrimidinase was hetero-overexpressed in E. coli and purified from the soluble supernatant using Ni 2+ -affinity chromatography. Pure protein was obtained in this single chromatographic step with an elution of buffer A. Approximately 50 mg of purified protein was extracted from 1 L of a culture of E. coli cells. The mutant dihydropyrimidinases were also purified according to the same protocol used for the wild-type proteins, and yielded very similar purification results.

The catalytic activity of purified dihydropyrimidinase (without any metal supplement in the culture) was not high, so some metal ions were added to the reaction mixture. Table 2 shows that the addition of 1 mM CoCl 2 , ZnCl 2 , or MnCl 2 activated dihydropyrimidinase activity, and followed the order Co 2+ > Zn 2+ > Mn 2+ ; CdCl 2 , NiCl 2 , MgCl 2 , and CaCl 2 were not useful. We also added 1 mM CoCl 2 , the best supplement, into the bacterial culture for dihydropyrimidinase expression, and the resultant dihydropyrimidinase was purified and analyzed. The specific activity of this dihydropyrimidinase toward dihydrouracil was 5.9 ± 0.4 μmol/mg/min, a value very similar to that of the Co 2+ -activated enzyme (5.8 ± 0.5 μmol/mg/min). Thus, dihydropyrimidinase (1 mM CoCl 2 supplemented into the bacterial culture) was used for all analyses in this study, unless explicitly stated otherwise.

This enzyme from microorganisms is commonly known as hydantoinase. Although the gene product of PA0441 has been described as a dihydropyrimidinase, the substrate specificities of dihydropyrimidinase and hydantoinase may differ. For example, the recombinant hydantoinase from Agrobacterium radiobacter prefers 5-leucinyl-hydantoin to phthalimide and dihydrouracil (~two orders of magnitude), as revealed by the catalytic efficiencies [12] . To ensure that the gene product of PA0441 is suitably identified as a dihydropyrimidinase, we analyzed the substrate specificity of this dihydropyrimidinase toward dihydrouracil (Fig 1B) , phthalimide ( Fig 1C) , 5-propyl-hydantoin ( Fig 1D) , allantoin, and dihydroorotate, which are typical substrates for dihydropyrimidinase, imidase, hydantoinase, allantoinase, and dihydroorotase, respectively (Table 3) . Unlike A. radiobacter hydantoinase, the catalytic efficiency of this enzyme toward dihydrouracil was higher than that toward phthalimide (fourfold) and 5-propyl-hydantoin (100-fold). Therefore, this bacterial enzyme was suitably identified as a dihydropyrimidinase, not a hydantoinase.

Substrate analogs for any enzyme are usually potential inhibitors. Members of the cyclic amidohydrolase family have highly similar active sites, and may use the same catalytic mechanism for substrate hydrolysis [38] . This condition raises an interesting question as to whether the substrate and inhibitor of allantoinase and dihydroorotase can be a potential inhibitor of dihydropyrimidinase because of their structural similarity. Allantoin and dihydroorotate are not substrates for dihydropyrimidinase-catalyzed reactions (Table 3) . Allantoin and dihydroorotate are structurally similar to hydantoin and dihydrouracil, except for the 5' side chain ( Fig  1A) . To assess whether allantoin and dihydroorotate are dihydropyrimidinase inhibitors, allantoin and dihydroorotate were individually included in the standard assay using 5-propylhydantoin (Fig 2A) or dihydrouracil ( Fig 2B) as a substrate. Allantoin and dihydroorotate only slightly inhibited the activity of dihydropyrimidinase. Thus, substrates of other cyclic amidohydrolase may not be strong inhibitors of dihydropyrimidinase. To assess whether inhibitors or substrate analogs of allantoinase and dihydroorotase can be inhibitors of dihydropyrimidinase, 5-hydantoin acetic acid, acetohydroxamate, and orotic acid were used in this study. Similarly, 5-hydantoin acetic acid and acetohydroxamate (an inhibitor of allantoinase) could only slightly inhibit the activity of dihydropyrimidinase (Fig 2) . We also tested whether inhibitors of the metalloenzyme glutaminyl cyclase, namely, N-ω-acetylhistamine, 3,5-diamino-1,2,4-triazole, and 3-amino-1,2,4-triazole [39, 40] , are potential inhibitors of dihydropyrimidinase, in which 3-amino-1,2,4-triazole at 100-750 mM affected enzyme activity (Fig 2) . High concentrations of these compounds were required to obtain significant inhibition of dihydropyrimidinase. Thus, inhibitors or substrate analogs of allantoinase, dihydroorotase, and glutaminyl cyclase were not potent inhibitors of dihydropyrimidinase.

To compare the inhibitory capability of these compounds on dihydropyrimidinase, their IC 50 values, the inhibitory concentration required to reduce the activity of the enzyme by 50%, were determined and compared. The IC 50 values of dihydroorotate, 5-hydantoin acetic acid, and acetohydroxamate were around 30, 50, and 500 mM, respectively. The IC 50 values of these substrate and inhibitor analogs of other cyclic amidases for dihydropyrimidinase were within the millimolar range.

Substrate analogs for any enzyme are usually potential inhibitors, but this common assumption was not present in this study of dihydropyrimidinase. Given broad substrate specificity, the active site of dihydropyrimidinase can accommodate many different substrates. Therefore, substrate analogs cannot be specifically recognized and cannot significantly inhibit its activity. Although we found that dihydroorotate, 5-hydantoin acetic acid, and acetohydroxamate inhibited dihydropyrimidinase activity, their IC 50 values were at the millimolar range and higher than the K m values of dihydropyrimidinase, which were insufficient as potent inhibitors. To determine whether a naturally occurring product is an inhibitor of dihydropyrimidinase, the inhibitory capabilities of the flavonols myricetin (with three hydroxyl substituents on the aromatic ring), quercetin (with two hydroxyl substituents on the aromatic ring), kaempferol (with one hydroxyl substituent on the aromatic ring), and galangin (without hydroxyl substituent on the aromatic ring) were tested. Furthermore, to study the structure-inhibition relationship with dihydropyrimidinase, the derivatives of myricetin, namely, dihydromyricetin (the ring does not contain a double bond but has additional two hydrogen atoms) and myricitrin (the ring is fused to an additional sugar block), were further used and tested (Fig 3A) . Identification of flavonol inhibition of dihydropyrimidinase

Flavonols are known to have antioxidant [20] , antiradical [22] , and antibacterial properties [23, 24] . Myricetin, quercetin, kaempferol, and galangin at different concentrations were included in the standard assay, and their IC 50 values were determined and compared to determine whether flavonols inhibit dihydropyrimidinase. Generally, the activities of dihydropyrimidinase toward dihydrouracil ( Fig 3B) and 5-propyl-hydantoin (Fig 3C) continued to decrease as the concentrations of the compound increased. However, these compounds caused distinct effects when different substrates were used. When dihydrouracil was used as a substrate, the inhibitory effects of myricetin and kaempferol on the activity of dihydropyrimidinase were significant; however, kaempferol only slightly inhibited the activity of Acetohydroxamate and 3-amino-1,2,4-triazole solutions, whose pH values, were pre-adjusted to pH 8. When using dihydrouracil as a substrate, some compounds at high concentrations were difficult to determine the inhibitory effect on the activity of dihydropyrimidinase using the spectrophotometric assay at 230 nm. dihydropyrimidinase toward 5-propyl-hydantoin (Fig 3C) . The effect of quercetin was insignificant. Overall, myricetin was found to strongly inhibit the hydrolysis of both dihydrouracil and 5-propyl-hydantoin catalyzed by dihydropyrimidinase. We further tested whether dihydromyricetin and myricitrin (Fig 3A) , derivatives of myricetin, can be inhibitors of dihydropyrimidinase. No effect was found when myricitrin was added into the standard assay using dihydrouracil and 5-propyl-hydantoin as substrates of dihydropyrimidinase. However, opposite to myricitrin, dihydromyricetin exhibited a significant inhibitory effect on the activities of dihydropyrimidinase for both substrates, even more than myricetin did (Fig 3B and 3C) . The IC 50 values of dihydromyricetin for dihydropyrimidinase determined from the titration curves using dihydrouracil and 5-propyl-hydantoin were 48 ± 2 and 40 ± 2 μM, respectively; myricetin yielded higher values of 83 ± 3 and 63 ± 3 μM, respectively. Based on the IC 50 values, the order of the inhibitory capability of the flavonoids for dihydrouracil was as follows: dihydromyricetin > kaempferol > myricetin > galangin > quercetin > myricitrin. Moreover, the order of the inhibitory capability of the flavonoids for 5-propyl-hydantoin was as follows: dihydromyricetin > myricetin > galangin > kaempferol > quercetin > myricitrin. Thus, for the first time, flavonols, especially dihydromyricetin, were established as novel inhibitors of dihydropyrimidinase. In addition, the inhibitory capabilities of these flavonols (at the μM range) were significantly higher than those of the substrate and inhibitor analogs (at the mM range) of dihydropyrimidinase.

To study the structure-inhibition relationship between dihydropyrimidinase and the flavonoids in silico, the structure of dihydropyrimidinase was modeled and used for docking experiments. The crystal structure of P. aeruginosa dihydropyrimidinase has yet to be determined. We modeled the P. aeruginosa dihydropyrimidinase structure (S1 Fig) by homology modeling using SWISS-MODEL (http://swissmodel.expasy.org/) [34] . Human dihydropyrimidinase (PDB entry: 2VR2) was the first hit suggested as a template by the program. The amino acid sequences of human (with 519 aa) and P. aeruginosa dihydropyrimidinase (with 479 aa) shared 51% identity and 66% similarity (S2 Fig). We also used (PS) 2 , another bioinformatic tool, for structural modeling [35] . (PS) 2 (http://140.113.239.111/~ps2v2/docs.php) is an automatic homology modeling server that combines both sequence and secondary structure information to detect the homologous proteins with remote similarity and target-template alignment. Human dihydropyrimidinase was also the first hit suggested as a template by (PS) 2 because it had the highest score. Results obtained from (PS) 2 analysis indicated that 99.58% of the secondary structure was aligned, which implied that human and P. aeruginosa dihydropyrimidinase shared a highly similar structure (S3 Fig). Mutational analysis of the residues within the active site According to the crystal structure of Saccharomyces kluyveri (PDB entry: 2FVK) [41] and Sinorhizobium meliloti dihydropyrimidinases (PDB entry: 3DC8) [42] , residues H59, H61, K150, H183, H239, and D316 of P. aeruginosa dihydropyrimidinase were crucial for the assembly of the binuclear metal center within the active site; meanwhile, residues Y155, S289, and N337 were crucial for substrate binding (Fig 4A) . The amino acid sequences of 497 sequenced dihydropyrimidinase homologs aligned using ConSurf [33] indicated that these residues were well conserved (Fig 4B) . The importance of these residues was then probed by site-directed mutagenesis, in which alanine substitution was constructed and analyzed. As expected, the catalytic activities of these Ala-substituted mutant proteins were severely impaired. H59A, H61A, K150A, H183A, H239A, D316A, Y155A and S289A were inactive. Only N337A mutant protein was found to be active, but its activity was about 20-fold less than that of the wildtype dihydropyrimidinase.

To understand the inhibitory mechanism of dihydromyricetin and myricetin on dihydropyrimidinase, their structures found in the DrugBank (http://www.drugbank.ca/) [36] were computationally docked into the 3D model of P. aeruginosa dihydropyrimidinase using PatchDock (http://bioinfo3d.cs.tau.acil/PatchDock/) [37] . Docking was automatically performed after uploading the coordinates and topology files of the compound and dihydropyrimidinase. The docking models with the highest score in dihydropyrimidinase interacting with dihydromyricetin and myricetin are shown in Fig 5. Although dihydromyricetin and myricetin were docked Inhibition of Dihydropyrimidinase by Flavonoids into the active site pocket of dihydropyrimidinase, their binding modes differed. On one hand, dihydromyricetin interacted with I95, S289, and D316 (Fig 5A) , in which S289 and D316 were found to be crucial for the catalytic activity of dihydropyrimidinase. On the other hand, myricetin interacted with N157, Q194, R212, and N337, in which only N337 was crucial. This docking study showed that myricetin only partially occupied the dihydropyrimidinase active site. The ring structure of dihydromyricetin was more flexible than that of myricetin ( Fig 5) because of lacking the double bond on its ring (Fig 3A) .

We identified dihydromyricetin as a potent inhibitor with an IC 50 value of 48 μM on dihydropyrimidinase. To determine what type of inhibitor dihydromyricetin is, this compound (40 μM) was included in the standard assay for dihydropyrimidinase with different concentrations of dihydrouracil. As shown in Fig 6, inhibition of dihydropyrimidinase by dihydromyricetin resulted in Lineweaver−Burk plots with lines that crossed the y-axis at a similar point, which indicated that dihydromyricetin was a competitive inhibitor of dihydropyrimidinase. The presence of 40 μM dihydromyricetin yielded V max and K m values, kinetic constants, of 5.4 ± 0.4 μmol/mg/min and 1.6 ± 0.2 mM, respectively. Without dihydromyricetin, the kinetic constants V max and K m were 7.6 ± 0.4 μmol/mg/min and 0.7 ± 0.1 mM, respectively ( Table 3 ). The K m value obviously increased by twofold, whereas V max was only slightly affected. These findings indicated the competitive inhibition of dihydropyrimidinase by dihydromyricetin; Inhibition of Dihydropyrimidinase by Flavonoids specifically, dihydromyricetin could compete with dihydrouracil for the active site of dihydropyrimidinase [43] .

To determine whether the inhibitory capabilities of myricetin and dihydromyricetin are correlated with their binding abilities, the dissociation constants (the K d values) of dihydropyrimidinase for myricetin and dihydromyricetin were determined through fluorescence quenching. Quenching refers to the complex formation process that decreases the fluorescence intensity of the protein. The fluorescence emission spectra of dihydropyrimidinase were remarkably quenched with dihydromyricetin ( Fig 7A) and myricetin (Fig 7B) . Upon adding 50 μM myricetin and dihydromyricetin, the intrinsic fluorescence of dihydropyrimidinase was quenched by 56.2% and 89.1%, respectively. As shown in Fig 7B, adding different concentrations of myricetin resulted in a red shift (~7 nm; λ max = 333.5-340.5 nm) in the dihydropyrimidinase emission wavelength (λ em ). Adding dihydromyricetin also caused a red shift in the dihydropyrimidinase emission wavelength (Fig 7A) , which was much more significant (~25 nm; λ max = 333.5-358.5 nm). These observations indicated that myricetin and dihydromyricetin interacted with dihydropyrimidinase, thereby suggesting that myricetin and dihydromyricetin could form stable complexes with dihydropyrimidinase. The K d values of dihydropyrimidinase bound to myricetin and dihydromyricetin, as determined through their titration curves (Fig 7C) , were 28.8 ± 2.2 and 22.6 ± 2.6 μM, respectively.

Our in silico experiments showed that myricetin and dihydromyricetin may use different modes to bind to dihydropyrimidinase (Fig 5) . These different possible binding modes were also revealed by the distinct fluorescence quenching spectra, in which their λ max shifts differed significantly (Fig 7) . The docking study showed that N337 interacted with the hydroxyl group on the ring of myricetin, but not dihydromyricetin. Thus, fluorescence quenching of the N337A mutant was also carried out to test whether this residue is important for myricetin binding, but not for dihydromyricetin. Myricetin and dihydromyricetin quenched the intrinsic fluorescence of dihydropyrimidinase by 58.4% and 86.3%, respectively (Fig 8) . Adding myricetin and dihydromyricetin resulted in red shifts in λ max of the N337A mutant from 333.5 nm to 341 nm (~7.5 nm) and 357 nm (~23.5 nm), respectively. The λ max shifts of the N337A mutant by myricetin and dihydromyricetin were similar, but still slightly different, to those of the wildtype dihydropyrimidinase (Fig 7) . The K d value of the N337A mutant bound to myricetin ( Fig  8C) , compared with that of the wild-type dihydropyrimidinase, was reduced to 59.3 ± 8.6 μM (twofold). However, the K d value of the N337A mutant for dihydromyricetin binding was 24.1 ± 1.7 μM, a value nearly identical to that of the wild-type dihydropyrimidinase, which suggested that N337 was important for myricetin binding but not for dihydromyricetin. These observations indicated that myricetin and dihydromyricetin could still form stable complexes with the N337A mutant, but the binding environment of the active site within the N337A mutant was somewhat different from that of the wild-type protein.

Substrate analogs for any enzyme are usually potential inhibitors, but this common rule is not applicable for dihydropyrimidinase (Fig 2) . Combating diseases caused by infections resistant to all antibacterial options requires the development of clinically useful small-molecule antibiotics and identification of novel targets. DNA metabolism is one of the most basic biological functions that should be a prime target in antibiotic development. Considering that dihydropyrimidinase is a component in the chain of pyrimidine catabolism required for metabolizing the DNA base, blocking this enzyme's activity may be useful to limit bacterial growth and survival. Although some chelators are known to inhibit dihydropyrimidinase, they are harmful to humans. In this study, we showed that dihydromyricetin, a flavonol, significantly inhibited the catalytic activities of dihydropyrimidinase toward both the natural substrate dihydrouracil and xenobiotic substrate 5-propyl-hydantoin (Fig 3) . Furthermore, the metabolic effects and safety of the flavonols are well established, making flavonols beneficial for humans. Thus, some of these plant polyphenols may become the lead compounds for further antibiotic development and drug design.

We found that the flavonols myricetin, quercetin, kaempferol, and galangin, which contain different numbers of hydroxyl substituents on their aromatic rings, exerted different inhibitory effects on dihydropyrimidinase (Fig 3) . Although flavonols are known to have several hydroxyl groups, thought to have remarkable potential for binding any protein, the strength of the flavonols' inhibitory effect (IC 50 ) on dihydropyrimidinase was not correlated with the number of hydroxyl substituents on the flavonols' aromatic rings. In addition, the inhibitory effect (IC 50 ) of flavonols on dihydropyrimidinase was substrate-dependent. For example, as shown in Fig 3, the inhibitory effect of kaempferol on the activity of dihydropyrimidinase was significant only with dihydrouracil as a substrate (with IC 50 value of 50 ± 2 μM), but not with 5-propyl-hydantoin. Although our docking results (Fig 5) and the findings of the kinetic studies (Fig 6) both showed a competitive inhibition of the flavonol dihydromyricetin on dihydropyrimidinase, further structure-inhibition relationship studies are still needed for deeper understanding of and myricetin with dihydropyrimidinase. An aliquot amount of dihydromyricetin and myricetin was individually added to the enzyme solution for each K d . The Inhibition of Dihydropyrimidinase by Flavonoids the inhibition mechanism, particularly because we are still unsure whether each enzyme only contains a single dihydromyricetin within the active site of dihydropyrimidinase.

Flavonoids are the most common group of plant polyphenols with antioxidant, antiradical, anticancer, and antibacterial properties [19] . Some flavonoids are ATPase-inhibiting agents and competitors for ATP-binding proteins [27, 32] . For example, several flavonoid derivatives have been developed as therapeutic agents for cancer [21] . Myricetin and scutellarein strongly inhibit the ATPase activity of nsP13 of the SARS coronavirus helicase [44] . Myricetin, luteolin, and morin inhibit the hexameric replicative helicases, and myricetin inhibits Gram-negative bacteria growth [45] . Myricetin also inhibits bacterial allantoinase and dihydroorotase [15] . Given that myricetin is also a potent inhibitor of numerous DNA and RNA polymerases and telomerases [46] , these flavonoids, especially myricetin, may be a competent dirty drug or multi-target drug [43] .

Dihydropyrimidinase is a zinc enzyme. However, in this study, dihydropyrimidinase supplemented with Co 2+ ions exhibited the highest activity ( Table 2 ). Due to the low bioavailability of cobalt on earth and in cells, we thought that the possibility of a biological role playing in dihydropyrimidinase may be ruled out. The higher activity of the Co 2+ -substituted enzymes has been proposed. The higher activity followed by metal replacement in Co(II)-thermolysin than in native thermolysin may be due to the enhancement of stabilization of the transition state than those of the native thermolysin [47] . For imidase, the sizes of the metal are found as the important factors that affect the activity of imidase [4] . The position of nucleophilic attack on an imide substrate may fluctuate due to size of the coordinating metal ion. Because the crystal structure of the Co 2+ -dihydropyrimidinase is still lacking, both these possibilities cannot be ruled out.

The chemical mechanism of the binuclear metal center-containing amidohydrolase likely involves three steps: (1) the hydrolytic water molecule must be activated for nucleophilic attack, (2) the amide bond of the substrate must be made more electrophilic by the polarization of the carbonyl-oxygen bond, and (3) the leaving group nitrogen must be protonated as the carbon-nitrogen bond is cleaved [38, 48] . This catalytic mechanism is applicable to each member in the cyclic amidohydrolase family [41, 48, 49] . However, a question remains as to why a common mechanism-based inhibitor for these imide-hydrolyzing enzymes, such as imidase, hydantoinase, dihydropyrimidinase, allantoinase, and dihydroorotase, is difficult to find. New and undetermined factors should be further studied and discovered in this protein family, such as the third metal ion recently found in dihydroorotase, which was found to be highly important for catalysis [50, 51] . Further studies are also necessary to investigate the substrate specificity, selectivity, and catalytic mechanism of dihydropyrimidinase and hydantoinase, as well as other cyclic amidohydrolases. 2 analysis showed that 99.58% of the secondary structure is aligned, indicating a highly similar structure between human and P. aeruginosa dihydropyrimidinase. (TIF)

Amidohydrolases of the reductive pyrimidine catabolic pathway purification, characterization, structure, reaction mechanisms and enzyme deficiency

The purification and properties of hydropyrimidine hydrase

A novel cold-adapted imidase from fish Oreochromis niloticus that catalyzes hydrolysis of maleimide

The role of metal on imide hydrolysis: metal content and pH profiles of metal ion-replaced mammalian imidase

Rat liver imidase

Enzymatic hydrolysis of organic cyclic carbonates

Dispelling the myths-biocatalysis in industrial synthesis

Hydantoinases and related enzymes as biocatalysts for the synthesis of unnatural chiral amino acids

From sequence and structure of sulfotransferases and dihydropyrimidinases to an understanding of their mechanisms of action and function

An evolutionary treasure: unification of a broad set of amidohydrolases related to urease

Chemical rescue of the post-translationally carboxylated lysine mutant of allantoinase and dihydroorotase by metal ions and short-chain carboxylic acids

Effect of metal binding and posttranslational lysine carboxylation on the activity of recombinant hydantoinase

Mammalian pyrimidine biosynthesis: fresh insights into an ancient pathway

Carboxylated lysine is required for higher activities in hydantoinases

Allantoinase and dihydroorotase binding and inhibition by flavonols and the substrates of cyclic amidohydrolases

Alarming beta-lactamase-mediated resistance in multidrug-resistant Enterobacteriaceae

Beta-lactamases identified in clinical isolates of Pseudomonas aeruginosa

The challenge of new drug discovery for tuberculosis

Dietary flavonoids: bioavailability, metabolic effects, and safety

Structure-activity relationships of flavonoids in the cellular antioxidant activity assay

Flavonoids as RTK inhibitors and potential anticancer agents

Antioxidant and antiradical activities of flavonoids

Polyphenols as antimicrobial agents

Antimicrobial activity of flavonoids

Crystal structure and DNA-binding mode of Klebsiella pneumoniae primosomal PriB protein

Identification of a novel protein, PriB, in Klebsiella pneumoniae

Characterization of flavonol inhibition of DnaB helicase: real-time monitoring, structural modeling, and proposed mechanism

DnaT is a single-stranded DNA binding protein

The N-terminal domain of DnaT, a primosomal DNA replication protein, is crucial for PriB binding and self-trimerization

C-terminal domain swapping of SSB changes the size of the ssDNA binding site

Characterization of a single-stranded DNA-binding protein from Klebsiella pneumoniae: mutation at either Arg73 or Ser76 causes a less cooperative complex on DNA

Inhibition of Klebsiella pneumoniae DnaB helicase by the flavonol galangin

ConSurf 2005: the projection of evolutionary conservation scores of residues on protein structures

The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling

PS) 2 : protein structure prediction server

DrugBank 3.0: a comprehensive resource for 'omics' research on drugs

PatchDock and SymmDock: servers for rigid and symmetric docking

Structural and catalytic diversity within the amidohydrolase superfamily

Structural and functional analyses of a glutaminyl cyclase from Ixodes scapularis reveal metal-independent catalysis and inhibitor binding

Cloning, expression, characterization, and crystallization of a glutaminyl cyclase from human bone marrow: a single zinc metalloenzyme

The crystal structures of dihydropyrimidinases reaffirm the close relationship between cyclic amidohydrolases and explain their substrate specificity

Structure of dihydropyrimidinase from Sinorhizobium meliloti CECT4114: new features in an amidohydrolase family member

An introduction to medicinal chemistry

Development of chemical inhibitors of the SARS coronavirus: viral helicase as a potential target

Flavones inhibit the hexameric replicative helicase RepA

Discovering new medicines targeting helicases: challenges and recent progress

Structural analysis of zinc substitutions in the active site of thermolysin

Crystal structures of vertebrate dihydropyrimidinase and complexes from Tetraodon nigroviridis with lysine carbamylation: metal and structural requirements for post-translational modification and function

Dihydropyrimidine amidohydrolases and dihydroorotases share the same origin and several enzymatic properties

Getting CAD in shape: the atomic structure of human dihydroorotase domain

Structure, functional characterization, and evolution of the dihydroorotase domain of human CAD

Conceived and designed the experiments: CYH. Performed the experiments: CYH. Analyzed the data: CYH. Contributed reagents/materials/analysis tools: CYH. Wrote the paper: CYH.