Sage Journals: Discover world-class research

Abstract

Genome-wide association studies have linked polymorphisms in the gene G72 to schizophrenia risk in several human populations. Although controversial, biochemical experiments have suggested that the mechanistic link of G72 to schizophrenia is due to the G72 protein product, pLG72, exerting a regulatory effect on human D-amino acid oxidase (hDAAO) activity. In an effort to identify hDAAO inhibitors of novel mechanism of action, we designed a pLG72-directed hDAAO activity assay suitable for high-throughput screening (HTS). During assay development, we confirmed that pLG72 was an inhibitor of hDAAO. Thus, our assay employed an IC₂₀ pLG72 concentration that was high enough to allow dynamic pLG72-hDAAO complexes to form but with sufficient remaining hDAAO activity to measure during an HTS. After conducting an approximately 150,000-compound HTS, we further characterized a class of compound hits that were less potent hDAAO inhibitors when pLG72 was present. Focusing primarily on compound 2 [2-(2,5-dimethylphenyl)-6-fluorobenzo[d]isothiazol-3(2H)-on], we demonstrated that these compounds inhibited hDAAO via an allosteric, covalent mechanism. Although there is significant interest in the therapeutic potential of compound 2 and its analogues, their sensitivity to reducing agents and their capacity to bind cysteines covalently would need to be addressed during therapeutic drug development.

Keywords

covalent inhibitors D-serine schizophrenia neuropathic pain

Introduction

Human D-amino acid oxidase (hDAAO) is a flavoenzyme expressed in liver, kidney, and brain.¹ Using flavin adenine dinucleotide (FAD) as a cofactor, hDAAO stereospecifically catalyzes the degradation of D-amino acids.¹ Several D-amino acids, most notably D-serine, are full agonists at the glycine site of the NMDA receptor (NMDAR),² a key excitatory receptor that is critical for several physiological processes such as cognition, learning, and brain structural plasticity.³ Alterations in NMDAR activation states are proposed to underlie various pathological conditions, such as schizophrenia, neuropathic pain, and stroke.³ By controlling the levels of the NMDAR coagonist D-serine, DAAO has been shown to be important for regulating cognition and learning.⁴ Because DAAO inhibitors have activity in rodent assays relevant to disease states such as cognitive impairment associated with schizophrenia,⁴ multiple industrial and academic groups have sought development of hDAAO inhibitors for potential utility in treating central nervous system (CNS) disorders.

One prominent hDAAO binding partner potentially relevant for schizophrenia is pLG72, the largest protein product of the G72 gene.⁵ More than 10 years ago, an influential genome-wide association study (GWAS) identified G72 as an anthropoid-specific gene linked to schizophrenia; this study further presented yeast two-hybrid and biochemical data demonstrating that pLG72 bound to and regulated hDAAO activity.⁵ Because of prior evidence linking hDAAO pathways to schizophrenia, the G72 genetic link to schizophrenia had an immediate mechanistic explanation. Following that initial publication, GWAS studies and meta-studies in several populations confirmed that G72 polymorphisms influence risk for schizophrenia and perhaps other psychiatric diseases.⁶

Despite genetic evidence linking the G72 gene and the pLG72 protein to hDAAO function and psychiatric disease, many mysteries remain regarding the precise function of pLG72 and its relationship to hDAAO. The most relevant issue concerns the biochemical outcome of pLG72 association with hDAAO. While the original report provided evidence that pLG72 activated hDAAO,⁵ subsequent publications presented careful data demonstrating hDAAO inhibition by pLG72.^7,8 Although a potential model synthesizing these competing observations is that pLG72 activates hDAAO over short periods of time but then initiates a slow, minutes-long conversion of hDAAO to an inactive conformation,⁷ this controversy has yet to be convincingly settled. Furthermore, pLG72 protein is expressed weakly in whole-brain homogenates, perhaps because it is expressed only in specific brain regions such as the cerebral cortex⁷ and amygdala.⁹ Although the original publication detected pLG72 messenger RNA (mRNA) in brain, spinal cord, and testes,⁵ there are reports of negative results in which the pLG72 mRNA and protein have not been detected in brain.¹⁰ Finally, experiments in cell lines and in transgenic mice suggest the pLG72 gene primarily affects mitochondrial function and has psychiatric effects independent of pLG72 association with hDAAO.^9,11

We initiated an effort to (1) generally identify hDAAO inhibitors of novel mechanism of action, (2) develop tool compounds to potentially explore hypotheses surrounding the function of the pLG72-hDAAO interaction, and (3) design compounds with potential utility for treating schizophrenia by targeting the pool of pLG72-hDAAO in the cerebral cortex. To accomplish these ambitious goals, we designed a pLG72-directed hDAAO activity assay (hDAAO/pLG72 assay) to screen for compounds that inhibited hDAAO when pLG72 was present. In our hands, pLG72 was a dose-dependent inhibitor of hDAAO. Therefore, we designed our assay with an IC₂₀ concentration of pLG72. Under these conditions, we hypothesized that pLG72 would be present at a high enough concentration to interact dynamically with hDAAO but not at a concentration high enough to fully inactivate the enzyme. In these balanced conditions, sufficient hDAAO activity would remain to be assayed in a high-throughput screen (HTS). During hit confirmation, we identified several classes of compounds that differentially inhibited hDAAO depending on the presence of pLG72. Because of their structural homogeneity, availability from commercial sources, and rapid opportunity for exploring structure-activity relationships, we describe here the characteristics of one class that were less active as hDAAO inhibitors when pLG72 was present. As we show here, these “class C compounds” were not active site hDAAO inhibitors but instead bound covalently to hDAAO cysteines to inhibit hDAAO activity. Although these compounds may not be effective as in vivo tool or therapeutic compounds (goals 2 and 3 above), our results confirm that our HTS strategy was capable of identifying hDAAO inhibitors of novel biochemical mechanism of action (goal 1 above).

Materials and Methods

Compound Synthesis and Protein Production

Compound 1 [4H-furo[3,2-b]pyrrole-5-carboxylic acid] was synthesized as described previously.¹² Compounds 2 [2-(2,5-dimethylphenyl)-6-fluorobenzo[d]isothiazol-3(2H)-on], 3 [2-(3-chloro-2-methylphenyl)-6-fluorobenzo[d]isothiazol-3(2H)-one], 4 [2-(4-fluorophenyl)-6-nitro-4-(2,2,2-trifluoroethoxy)benzo[d]isothiazol-3(2H)-one], and 5 [2-(p-tolyl)benzo[d]isothiazol-3(2H)-one] were purchased from ChemDiv (San Diego, CA). For more in-depth characterization, an additional batch of compound 2 was synthesized using established procedures.¹³ Compounds 6 [3-(7-hydroxy-2-oxo-4-phenyl-2H-chromen-6-yl)propanoic acid] and 7 [(4-hydroxy-6-(2-(7-hydroxy-2-oxo-4-phenyl-2H-chromen-6-yl)ethyl)pyridazin-3(2H)-one] were synthesized as described previously.¹⁴ Compound 8 [6-methylbenzo[d]isoxazol-3-ol] was prepared as described.¹⁵ Compounds 9 [6-fluoro-2-phenylbenzo[d]isoxazol-3(2H)-one], 10 [6-fluoro-2-(p-tolyl)benzo[d]isoxazol-3(2H)-one], and 11 [6-fluoro-2-(o-tolyl)benzo[d]isoxazol-3(2H)-one] were prepared based on previously described protocols.¹⁶ The structures of all synthesized compounds were verified using standard mass spectrometry and nuclear magnetic resonance (NMR) analytical techniques. Compound structures are presented in Table 1 .

Table 1.

Summary of Enzymatic Inhibition Data.

		IC50 (µM)
Compound	Structure	hDAAO (Amplex)	Counter (Amplex)	hDAAO (LC-MS/MS)	hDAAO (High FAD)	rDAAO (Amplex)
Compound 1		0.0030*	>1*	0.0076*	0.0013 ± 0.0004 (4)	0.086*
Compound 2		0.36 ± 0.14 (6)	>100 (4)	0.066 ± 0.025 (3)	2.0	0.85
Compound 3		0.19 ± 0.05 (4)	>100 (2)	0.074	1.5	0.87
Compound 4		0.27 ± 0.09 (4)	>100 (2)	0.11	2.0	0.66
Compound 5		0.36 ± 0.29 (4)	>100.0	0.074	2.9 ± 1.0 (2)	0.24 ± 0.04 (2)
Compound 6		0.037*	>10*	0.23*	0.038	4.0*
Compound 7		0.020*	>10*	0.11*	0.015	0.047*
Compound 8		0.014 ± 0.005 (4)	ND	ND	0.013	0.28
Compound 9		>50	>50	ND	ND	ND
Compound 10		>100	>100	ND	ND	ND
Compound 11		>100	>100	ND	ND	ND

IC₅₀ values (in µM) are presented for the different compounds in the following assays: human D-amino acid oxidase (hDAAO) activity determined by the Amplex Red assay system; the counterassay, which tests for compound interference in the Amplex Red system; hDAAO activity assayed using the liquid chromatography/tandem mass spectrometry (LC-MS/MS) platform; hDAAO activity in the high flavin adenine dinucleotide (FAD) (50 µM) condition; and rat D-amino acid oxidase (rDAAO) in the Amplex Red system. IC₅₀ values are followed by standard deviation (SD) with the number in parentheses when n is greater than 1. ND, not determined. Values marked with an asterisk have already been published¹⁴ but are presented in this table for comparison with novel findings.

hDAAO protein (NP_001908.3, 347 amino acids, 40 kDa) was expressed using an N-terminal 6-His-tag in BL21(DE3) Escherichia coli cells and purified by elution on a nickel column followed by size exclusion chromatography.¹⁷ pLG72 was prepared as reported in Molla et al.¹⁸ and was stored in 20 mM Tris-HCl (pH 8.5), 100 mM NaCl, 5% glycerol, 5 mM 2-mercaptoethanol (βME), and 0.1% N-lauroylsarcosine (NLS). Rat DAAO (rDAAO) was purified and stored as previously described.¹⁹ Glucose oxidase protein was purchased from Life Technologies (Carlsbad, CA; Amplex Red Glucose/Glucose Oxidase Assay Kit). Bovine D-aspartate oxidase (DASPO) was purchased from Gabrielle Tedeschi (University of Milan, Milan, Italy).^20,21

HTS

All enzymatic assays were conducted at room temperature (20–23 °C). For all concentration-response curves, data were analyzed by Prism (GraphPad Software, La Jolla, CA) using a standard, four-parameter equation:

y = Bottom + (Top - Bottom) / (1 + 10 (({LogIC}_{50} - x) * Hill slope)) .

In the primary screen, approximately 150,000 compounds were screened in the hDAAO/pLG72 assay. To eliminate nonspecific compounds, all compounds in the primary HTS were counterscreened for inhibition of the flavoenzyme glucose oxidase (GO assay) in parallel. The GO assay was chosen for counterscreen because GO is a well-characterized flavoprotein for which an easy-to-use kit exists for enzyme assay (Life Technologies). The general assay format was a coupling assay in which hydrogen peroxide produced by hDAAO or GO enzymatic activities was used by horseradish peroxidase (HRP; Sigma-Aldrich, St. Louis, MO) to oxidize Amplex Red (Life Technologies) and convert it into resorufin. The quantity of resorufin was then determined by a fluorescence plate reader.

More specifically, test compounds (10-mM stock solution in DMSO; Sigma-Aldrich) were diluted in the assay buffer (50 mM phosphate buffer, 150 mM NaCl, and 0.06% human serum albumin [Sigma Aldrich], pH 7.4) and added to the assay plate (low-volume 384-well plate; Corning, Corning, NY) using a CyBi-Well pipettor (CyBio, Jena, Germany) at 6 µL/well for a final concentration of 10 µM. In the hDAAO/pLG72 assay, the 5× enzyme mixture containing hDAAO, pLG72, FAD (Sigma Aldrich), and NLS in assay buffer and the 5× start solution containing HRP, Amplex Red, and D-serine in assay buffer were each added at 2 µL/well using FlexDrop PLUS (PerkinElmer, Waltham, MA). The final concentrations of each reagent were as follows: 6.25 nM hDAAO, 0.7 µM pLG72, 4 µM FAD, 0.005% NLS, 0.07 U/mL HRP, 0.035 mM Amplex Red, and 7.5 mM D-serine. In the GO assay, a similar procedure was used in which enzyme and start solution were both added as 5× master mixes. The final reagent concentrations in the GO assay were as follows: 1.25 mU/mL GO, 0.07 U/mL HRP, 0.035 mM Amplex Red, and 40 mM D-(+)-glucose. The assay plates were incubated for 1 h, and the fluorescence of resorufin was measured using an EnVision plate reader (PerkinElmer) with 535 nm excitation and 590 nm emission. In all experiments, hDAAO and pLG72 were incubated for a minimum of 30 min prior to activity assessment.

Primary hits were defined as compounds with >50% inhibition of hDAAO in the presence of pLG72, <20% inhibition of GO, and ≥50% purity of compound (as assessed by standard liquid chromatography/mass spectrometry analysis). As control, the IC₅₀ of compound 1 in the hDAAO/pLG72 assay and the IC₅₀ of Radicut (Wako Pure Chemical Industries, Osaka, Japan) in the GO assay were calculated in each assay plate.

The reproducibility and concentration dependency of primary hits were evaluated in the hDAAO/pLG72 assay, the GO assay, and the assay of hDAAO activity in the absence of pLG72 (hDAAO solo). Confirmed hits were defined as compounds that inhibited hDAAO/pLG72 with an IC₅₀ <10 µM and were negative for GO inhibition (IC₅₀ >10 µM).

Additional Enzymatic Assays

Enzymatic assays other than the HTS and hit confirmation activities described above were performed in 96-well black-walled plates (Corning Costar, Corning, NY). DMSO was used as solvent for compound solutions, except where specifically indicated. In all experiments, DMSO vehicle concentration was carefully maintained such that all controls and experimental groups had precisely the same DMSO concentration. No [DMSO]_f exceeded 1.5%.

The hDAAO and rDAAO activity assays in the Amplex Red platform were performed as described previously¹⁴ using roughly the same coupled assay format as described above for the HTS. The hDAAO reaction was allowed to incubate for 1 h, and accumulated fluorescent resorufin reaction product was measured with a FlexStation II (Molecular Devices, Sunnyvale, CA). While the final FAD in this standard protocol was 450 nM, for the “high FAD” condition, 50 µM FAD was used. In the counterassay, hDAAO was omitted and replaced with 800 nM hydrogen peroxide (Sigma-Aldrich) to produce the fluorescent product. For assaying rDAAO, the hDAAO protocol was followed with the following alteration: rDAAO was used instead of hDAAO, and D-serine concentration was increased to 30 mM.

DASPO activity and inhibition by test compounds also was assayed using the Amplex Red format. The assay protocol was the same as described above with the following exceptions: 0.36 nM DASPO was used as enzyme, 1.6 mM D-aspartate (Sigma-Aldrich) was used as substrate, and the reaction was allowed to proceed for 30 min prior to measurement of fluorescent resorufin product. The positive control inhibitor was potassium sodium tartrate solution (Sigma-Aldrich), which was diluted in aqueous buffer solution.

For an assay of hDAAO activity using an alternative detection method, liquid chromatography/tandem mass spectrometry (LC-MS/MS) was used to directly measure the benzoylformic acid product of the hDAAO reaction, when D-phenylglycine was used as substrate. This assay was described in greater detail previously.¹⁴

For determining hDAAO inhibition in the presence of reducing agents, the LC-MS/MS platform was used, as the Amplex Red platform was sensitive to high amounts of reducing agents. In these experiments, the standard LC-MS/MS–based activity assay was conducted, except that 5 mM glutathione (Sigma-Aldrich) or 5 mM βME (Pierce, Rockford, IL) was added to buffers as indicated in Results and figure legends.

As described previously,¹⁴ saturation experiments were used to determine compound mechanism of action with respect to substrate (D-serine) and cofactor (FAD). In the Amplex Red system, D-serine was maintained at a constant concentration, while the cofactor was varied over seven different concentrations or vice versa. These saturation curves were repeated in the presence of varying amounts of inhibitor. After a 1-h reaction, product was measured in endpoint mode using the FlexStation II. To determine V_max,obs and K_M,obs values at each inhibitor concentration, data of enzymatic activity (y) at given substrate concentrations (x) were plotted and fit with a one-site, specific binding equation (GraphPad Prism):

y = V_{\max, obs} * x / (K_{M, obs} + x) .

Several experimental designs were used to estimate rate of compound dissociation from hDAAO (off-rate). In each case, the assay used the Amplex Red assay platform described above. In the jump-dilution protocol, high concentrations of inhibitor were incubated with hDAAO and FAD for 30 to 90 min to form inhibited complexes. Reaction mixture (Amplex Red, D-serine, HRP) was then added, and fluorescent reaction product was measured kinetically using the FlexStation II. Examination of concentration-response curves was used such that at the “high” concentration, hDAAO was 80% to 100% inhibited, and at the “low” concentration, hDAAO was 0% to 20% inhibited. Utilization of this assay for hDAAO inhibitor quantitative off-rate determination has been described previously.¹⁴ Accumulated product formed (P) at a given time (t) is fit with equation (3) to determine initial reaction velocity (v_i), steady-state reaction velocity (v_s), and off-rate (k):

[P] = V_{s} t + \frac{V_{i} - V_{s}}{k} (1 - e^{- k t})

An alternative assay for off-rate determination with longer time scale was developed. Here, 200 nM hDAAO was incubated for 30 min with 4 µM inhibitor in 0.24 µM FAD. Samples then were diluted 200-fold, such that 1 nM hDAAO and 20 nM inhibitor were present. In this dilution, the FAD concentration was 61.5 µM. After a time of recovery that could last up to 24 h to allow for compound dissociation, 130 µL of diluted mixture was added to 70 µL of the Amplex Red reaction mix for hDAAO activity determination. The final [D-serine] was 50 mM, and the final [FAD] was 40 µΜ. Activity was assessed over 10 min kinetically on the FlexStation II, and activity was defined by the slope of the linear progress curve of product formation. For a given time point, all results were normalized to activity of DMSO-treated hDAAO. A 20-nM inhibitor condition was used to ensure that this concentration was not inhibitory.

For determination of compound reversibility in the presence of βME, 2 µM hDAAO was incubated with 20 µM inhibitor for 30 min. hDAAO inhibitor mixes were diluted 2-fold into 5 mM βME. After variable time periods, a small aliquot was diluted 1000-fold into Amplex Red reaction volume. With this dilution to 5 µM, βME did not interfere with the Amplex Red fluorescent detection system. [D-serine]_f was 50 mM, [FAD]_f was 40 µΜ, [compound]_f was 10 nM, and [hDAAO]_f was 1 nM. Activity was assessed over 10 min kinetically on the FlexStation II. Activity was defined by the slope of the linear progress curve of product formation, and results were normalized to hDAAO plus DMSO controls.

LC-MS/MS Analytical Methods

Detection of degradation products from the reaction mixtures of compounds 1 or 2 with βME was performed using positive or negative ion liquid chromatography/mass spectrometry (LC/MS). To determine those breakdown products, a 2-mL aliquot of sample was directly injected onto an ultra-high-performance liquid chromatography (UPLC) system (Waters, Milford, MA) equipped with a API 5500 QTrap mass spectrometer detector (Applied Biosystems/MDS Sciex, Framingham, MA) operated in positive TurboIonSpray mode for compound 2 and negative TurboIonSpray mode for compound 1. Instrument conditions were adjusted and optimized for the parent compounds 2 and 1, which were monitored using single-ion resonance (SRI) at m/z 274 and 150, respectively. Four additional masses around those two parent masses were also monitored as potential degradation product masses: m/z 270, 276, 278, and 280 for compound 2 and m/z 148, 152, 154, and 156 for compound 1. The gradient separation of the reaction mixture from extracted matrix materials was accomplished with an overall runtime of 1.5 min using a Waters Acquity BEH C-18 1.8-micron column (50 × 2.1 mm) maintained at 25 °C. The mobile phases used consisted of 1.0 mM ammonium formate with 0.2% formic acid (v/v) in water (A) and 1.0 mM ammonium formate with 0.2% formic acid (v/v) in acetonitrile (B) at a total flow rate of 0.6 mL/min. Wash solvent 1 was 3% formic acid in acetonitrile, and wash solvent 2 was 3% formic acid in water.

LC-MS/MS Proteomic Methods

Covalent binding of compound 2 to cysteine residues of hDAAO was determined by using the PTM Profiling plus protocol by MS Bioworks (Ann Arbor, MI). In brief, replicate gel segments were subjected to digestion with trypsin (Promega, Madison, WI), chymotrypsin, and elastase (Worthington, Lakewood, NJ). Digests were analyzed by nano LC/MS/MS with a NanoAcquity HPLC system (Waters) interfaced to an LTQ Orbitrap Velos tandem mass spectrometer (ThermoFisher, San Jose, CA). Peptides were loaded on a trapping column and eluted over a 75-µm analytical column at 350 nL/min; both columns were packed with Jupiter Proteo resin (Phenomenex, Torrance, CA). A 30-min gradient was employed for each segment. The mass spectrometer was operated in data-dependent mode, with MS performed in the Orbitrap at 60,000 FWHM resolution and MS/MS performed in the LTQ at unit resolution. The 15 most abundant ions were selected for MS/MS from each MS scan. Dynamic exclusion and repeat settings ensured each ion was selected only once and excluded for 30 s thereafter.

Product ion data were searched against the combined forward and reverse custom protein database comprising the hDAAO sequence using a locally stored copy of the Mascot search engine v2.3 (Matrix Science, London, UK) via Mascot Daemon v2.3. Peak lists were generated using Proteome Discoverer v1.4 (ThermoFisher). The database was appended with common background proteins. Search parameters were precursor mass tolerance 10 ppm, product ion mass tolerance 0.6 Da, two missed cleavages allowed (trypsin digest only), no fixed modifications, variable modifications of oxidized methionine, and deamidation on asparagine and glutamine and compound 2 on cysteine. Mascot search result flat files (DAT) were parsed to the Scaffold software v4.1 (Proteome Software, Portland, OR), with the following cutoff values used: 90% protein and 50% peptide level probability (probabilities were assigned by the Protein Prophet algorithm) and a minimum of two unique peptides per protein.

Results

pLG72-Directed hDAAO Activity Assay HTS

The principle on which the proposed assay is based is that pLG72 binds to hDAAO,^5,7,8 alters hDAAO structure in a time-dependent manner,^7,22 and regulates hDAAO activity. This regulation has been reported as activation^5,23 by some researchers and inhibition by others.^7,8 In our hands, following a 30-min preincubation protocol and assaying hDAAO reaction product over the course of 60 min, pLG72 inhibited hDAAO with an IC₅₀ of approximately 0.9 µM (Suppl. Fig. S1A).

As the subset of hDAAO protein relevant for schizophrenia could potentially be found in complex with pLG72,⁷ we sought to screen compounds capable of inhibiting hDAAO when pLG72 was present. We screened for hDAAO inhibition in the presence of 700 nM pLG72, which was an approximate IC₂₀ pLG72 concentration (Suppl. Fig. S1A). Under these conditions, we hypothesized that pLG72 would be in a dynamic on/off equilibrium with hDAAO that could potentially shield or expose novel small-molecule binding sites on either protein that would affect hDAAO activity. This screening assay, established in 384-well format, had a high Z′ factor value (Suppl. Fig. S1B), indicating suitability for identifying true positives during a screening campaign.²⁴ We observed consistent ability of the assay to report inhibition by the well-characterized hDAAO inhibitor compound 1 (Suppl. Fig. S1C). We further demonstrated consistency with two batches of pLG72 and no effect of pLG72 on hDAAO inhibition by the active site inhibitor compound 1 (Suppl. Fig. S1D). Additional experiments demonstrated that several known hDAAO active site inhibitors had similar IC₅₀ values with or without addition of pLG72 (data not shown).

In this hDAAO/pLG72 assay, we screened approximately 150,000 compounds from the Sumitomo Dainippon Pharma small-molecule corporate library. In the primary screen, compounds were screened in the hDAAO/pLG72 assay and counterscreened for inhibition of the flavoenzyme GO. Compound hits (>50% inhibition of hDAAO/pLG72 with <20% inhibition of GO) were subjected to purity analysis. Compounds with >50% purity were tested in concentration-response experiments in the two screening assays (hDAAO/pLG72 and GO assays) plus assay of hDAAO activity in the absence of pLG72 (hDAAO solo assay). This workflow is described in Figure 1A . Confirmed hits were defined as compounds that inhibited hDAAO in the presence of pLG72 with an IC₅₀ <10 µM that were negative for GO inhibition (>10 µM IC₅₀). Among the 102 confirmed hits from this screening workflow, we separated hDAAO inhibitors into three categories based on their pLG72 dependency: class A compounds were more potent in the presence of pLG72, class B compounds (the largest class) were equipotent inhibitors of hDAAO in the absence or presence of pLG72, and class C compounds were less potent inhibitors of hDAAO when pLG72 was present ( Fig. 1B ).

Figure 1.

Results from the high-throughput screen (HTS). (A) This schematic describes the general protocol along with the number of compounds passing through each step in the workflow. (B) Compounds were assayed in the human D-amino acid oxidase (hDAAO)/pLG72 assay (squares), the hDAAO solo assay (circles), and the glucose oxidase (GO) assay (triangles; the counterscreen). Confirmed HTS hits were placed into three categories: class A compounds were more potent hDAAO inhibitors in the presence of pLG72, class B compounds were equipotent (within 2-fold IC₅₀) inhibitors in both assays, and class C compounds were less potent hDAAO inhibitors in the presence of pLG72. The number of compounds in each class is indicated to the left of the curves. Compounds were defined as hits only if their GO inhibition IC₅₀ >10 µM.

Biochemical Characterization of Class C Compound hDAAO inhibitors

Because the active-site hDAAO inhibitor compound 1 was an equipotent hDAAO inhibitor in the presence or absence of pLG72 (Suppl. Fig. S1D), we hypothesized that class C compounds, which were consistently 5- to 10-fold less potent hDAAO inhibitors when pLG72 was present, would not be inhibiting hDAAO via the canonical, active-site mechanism. Furthermore, class C compounds were chemically interesting, as all seven members of this class had a substructure of 2-phenyl-2,3-dihydro-1,2-benzothiazol-3-one; this compound substructure is also called ebsulfur.^25,26

Using compound 1 as a positive control hDAAO inhibitor, we characterized, in a panel of biochemical assays, properties of commercially available compounds with the class C substructure. Although these compounds were not the exact compounds emerging from the HTS, because of their chemical similarity, they are furthered referred to as “class C compounds.” As shown in Figure 2A and summarized in Table 1 , four of the class C compounds inhibited hDAAO in the standard Amplex Red assay with IC₅₀ values between 0.19 and 0.36 µM. In a separate assay format in which D-phenylglycine was used as substrate and the product of the hDAAO reaction was directly measured by LC-MS/MS,¹⁴ the four class C compounds again inhibited hDAAO activity ( Fig. 2B ) with sub-µM potency ( Table 1 ). This result, along with the lack of effect of class C compounds in the Amplex Red counterassay ( Table 1 ), suggested that the hDAAO inhibition by class C compounds was not an artifact of the hydrogen peroxide–dependent Amplex Red system. The four class C compounds were sub-µM rDAAO inhibitors, albeit with unusually shallow Hill slopes ( Fig. 2C and Table 1 ). We tested inhibitor specificity by assaying inhibition of DASPO,²⁰ the mammalian flavoenzyme closest in amino acid sequence to hDAAO. As shown in Figure 2D , neither compound 1 nor any of the class C compounds up to 100 µM inhibited DASPO. Tartrate, a low-affinity but well-characterized inhibitor of DASPO,²⁰ was used as a positive control ( Fig. 2D ), at concentrations that had no effect on hDAAO activity ( Fig. 2A ).

Figure 2.

Class C compounds were specific D-amino acid oxidase (DAAO) inhibitors. (A) Compound 1 (positive control) and compounds 2 to 5 (class C compounds) inhibited human DAAO (hDAAO), as determined by the hDAAO activity assay in the standard Amplex Red platform. Tartrate, a positive control D-aspartate oxidase (DASPO) inhibitor,²⁰ did not inhibit hDAAO in this assay. (B) In the liquid chromatography/tandem mass spectrometry (LC-MS/MS) assay, hDAAO activity was determined using D-phenylglycine as substrate and with a direct measure of reaction product. In this assay, compounds 1 to 5 inhibited hDAAO activity. (C) Activity assay performed by using the Amplex Red platform, but substituting rat DAAO (rDAAO) for hDAAO, showed that compounds 1 to 5 inhibited rDAAO. (D) Compounds were tested for DASPO inhibition, using the Amplex Red platform for activity determination. Tartrate (the positive control compound) inhibited DASPO, but compounds 1 to 5 were ineffective as DASPO inhibitors up to the concentrations tested.

To determine if class C compounds inhibited hDAAO via a novel mechanism, we tested the effect of FAD on inhibition. In these experiments, we changed FAD concentration in the Amplex Red assay from the standard condition (450 nM FAD, near the K_M value) to a “high FAD” condition (50 µM FAD). In this assay, several compounds were marginally more potent in the latter, high FAD, conditions ( Fig. 3A , left): compound 1; two novel, “active-site lid-open” hDAAO inhibitors, compounds 6 and 7¹⁴; and compound 8, a previously described active-site hDAAO inhibitor.²⁷ Because the slight left shifts in this assay are within the minimum significance ratio (MSR) of our assay (which is 1.8; data not shown), it is unclear if this difference is significant. However, greater potency in the presence of higher FAD concentration is consistent with an FAD-uncompetitive mechanism of action, which has been documented for compounds 1, 6, and 7.¹⁴ On the other hand, all four class C compounds were 5- to 10-fold less potent as hDAAO inhibitors in the presence of the high FAD condition ( Fig. 3A , right and Table 1 ), suggesting FAD cofactor-competitive behavior.

Figure 3.

Class C compounds were flavin adenine dinucleotide (FAD) and D-serine mixed human D-amino acid oxidase (hDAAO) inhibitors. (A) Concentration-response curves for compound inhibition of hDAAO were conducted in the standard condition (0.45 µM FAD, approximate hDAAO K_M for FAD, solid lines) and in the “high FAD” condition (50 µM FAD, dashed lines). When comparing inhibition potency in the two assays run in parallel, compounds 1 and 6 to 8 were slightly left-shifted in the high FAD condition (left panels), while compounds 2 to 5 (class C compounds, right panels) were right-shifted in the high FAD condition. (B) K_M,obs and V_max,obs in the presence of several concentrations of inhibitor were determined by using nonlinear curve fitting. Left panels, FAD saturation experiments performed at high and fixed D-serine concentration. Right panels, D-serine saturation experiments performed at high and fixed FAD concentration.

To test this hypothesis more thoroughly, we performed substrate and cofactor saturation experiments in which hDAAO relative ligand affinity (K_M,obs) and maximum velocity (V_max,obs) were measured in the presence and absence of various concentrations of inhibitor. In FAD saturation experiments, compound 8 lowered both the K_M,obs and the V_max,obs of hDAAO, consistent with this compound being an FAD-uncompetitive hDAAO inhibitor ( Fig. 3B ). In D-serine saturation experiments, compound 8 increased K_M,obs with modest effect on V_max, consistent with active-site competitive inhibition ( Fig. 3B ). In contrast, compound 2 and compound 3 (class C compounds) increased K_M,obs and decreased V_max,obs when either FAD or D-serine saturation experiments were performed. This mixed-inhibition mechanism of hDAAO inhibition by compounds 2 and 3 is consistent with the FAD-competitive behavior described in Figure 3A and demonstrates that class C compounds inhibit hDAAO via a novel biochemical mechanism.

Class C Compounds Inhibit hDAAO Covalently and Reversibly

To further define the mechanism of action of the class C hDAAO inhibitors, we used jump-dilution assays to determine compound off-rate. In these experiments, a high, inhibitory dose of test compound was incubated with hDAAO and then rapidly diluted, such that the test compound was then at a noninhibitory concentration. By observing the kinetic recovery of enzymatic activity from an initial velocity (v_i) to a steady-state velocity (v_s), the compound off-rate can be determined, as we have done previously for compound 1 ( Fig. 4A and Terry-Lorenzo et al.¹⁴). Surprisingly, in these experiments, two of the class C compounds, compound 2 and compound 3, had no detectible off-rate over the course of a 2-h jump-dilution experiment ( Fig. 4B ). To differentiate between no off-rate and an extremely slow off-rate, a different protocol was used such that preformed inhibitory complexes were diluted and then allowed to reequilibrate for extended periods of time prior to activity assay. In these experiments, even 24 h after dilution, there was no measurable recovery of hDAAO activity ( Fig. 4C , solid lines).

Figure 4.

Compounds 2 and 3 have no detectible off-rate. In jump-dilution experiments, compound (cpd) at an inhibitory concentration is incubated with human D-amino acid oxidase (hDAAO) to form an inhibited complex. This complex is rapidly diluted 40-fold, such that the inhibitor is at a “low,” relatively noninhibitory concentration, and reaction product is measured kinetically. The slope of product accumulation curves defines enzymatic velocity. Under these conditions, initial velocity (v_i) is low, and over time, a steady-state velocity (v_s) is achieved as inhibitor is progressively released. The “low” concentration alone is used to define v_s. In these experiments, inhibitor concentrations are chosen to maximize the hDAAO activity difference between the initial high concentration and the diluted noninhibitory concentration (see Fig. 2A ). (A) Control data with the reversible inhibitor compound 1, demonstrating the time-dependent transition from v_i to v_s. (B) Over the course of 2 h of monitoring activity, bound complexes of hDAAO with either compound 2 or compound 3 show no evidence of recovering activity after dilution (25 nM final concentration, open symbols). In these experiments, 25 nM compound displayed no inhibitory activity (shaded triangle and diamond symbols; superimposable with DMSO-treated controls), demonstrating that if compound 2 or 3 had dissociated from hDAAO, activity would have recovered. (C) To differentiate between no off-rate and an extremely slow off-rate, 4-µM compounds 2 and 3 (inhibitory concentrations) were independently incubated with hDAAO for 30 min. The bound complexes were then diluted 200-fold to a noninhibitory final concentration (20 nM). After extended recovery postdilution, hDAAO activity was assayed kinetically over 10 min. Controls included conditions in which DMSO-incubated hDAAO was incubated in the noninhibitory, 20-nM compound concentration (dashed line). For each time point, activity was normalized to the value for hDAAO that was not exposed to any inhibitor.

Due to the lack of detectible off-rate ( Fig. 4 ), we hypothesized that the class C compounds were inhibiting hDAAO via a covalent modification. This hypothesis was informed by literature reports that other isothiazolone compounds, via S-S thiol crosslinking,²⁸ are capable of forming covalent bonds with cysteines in target proteins.²⁹ Consistent with this model, we synthesized and tested for hDAAO inhibition compounds 9, 10, and 11 ( Table 1 ). These compounds were similar in structure to class C compounds but contained an O instead of S in the heterocyclic ring. All three of these compounds were completely impotent as hDAAO inhibitors ( Table 1 ), suggesting that the S in the class C compounds was essential for hDAAO inhibition. As an additional test of our hypothesis, we assayed the impact of a reducing environment on hDAAO inhibition by the prototype class C compound, compound 2. As shown in Figure 5A , in 5-mM glutathione (GSH) conditions that would mimic the cellular environment, compound 2 was completely impotent as an hDAAO inhibitor. In contrast, compound 1 was an equipotent inhibitor in the presence or absence of GSH, demonstrating that the GSH-induced reducing environment did not have a general effect on hDAAO inhibition by active-site ligands. As a further test, we determined compound stability under reducing conditions. As shown in Figure 5B , while compound 1 was fully stable in βME, compound 2 was rapidly degraded. The LC-MS spectra of βME-treated compound 2 revealed a prominent m/z +2 ion (Suppl. Fig. S2) consistent with S-N bond reduction and addition of two H (proposed reaction in Fig. 5B , bottom).

Figure 5.

Compound 2 inhibition of human D-amino acid oxidase (hDAAO) is reversible in reducing environments. (A) By using the liquid chromatography/tandem mass spectrometry (LC-MS/MS) hDAAO activity assay platform, compound inhibition of hDAAO was tested in the absence (solid lines) and presence (dashed lines) of 5 mM glutathione (GSH). While compound 1 was equipotent in both conditions, when GSH was added, compound 2 was completely impotent to 10 µM. (B) Compounds 1 and 2 were incubated with hDAAO reaction mixture ±5 mM 2-mercaptoethanol (βME). The incubation duration was 1.5 h to mirror the conditions used in the hDAAO activity assay protocol. After incubation, LC-MS was used to assay the amount of parent compound. Data are normalized to control, non-βME conditions. Analysis of the compound 2 + βME mixture demonstrated both a loss of parent compound 2 and the formation of a molecule with an m/z ratio of +2. This result is consistent with the conclusion that, in the presence of βME, the S-N bond of compound 2 was reduced and two hydrogens were added to the molecule (see the proposed reaction, bottom panel). (C) To determine if hDAAO-compound 2 (cpd2) binding was reversible following exposure to a reducing agent, we performed the experiment described in the schematic on the right. Briefly, after a 30-min incubation of compound 2 (or DMSO) + hDAAO, the complex was diluted 2-fold. At indicated times, complexes were diluted 1000-fold and hDAAO activity was assayed kinetically. As indicated in the schematic, 5 mM βME was present or absent. No hDAAO inhibition was observed when compound 2 was present at 10 nM (solid symbols) or when 5 mM βME was present continuously (triangle symbols). Notably, when the hDAAO–compound 2 inactive complex was diluted into 5 mM βME, hDAAO activity recovered over the course of 1 to 2 h (open circles). In the absence of 5 mM βME, hDAAO remained inhibited (open squares).

Because compound 2 was sensitive to βME and unable to inhibit hDAAO in reducing environments, we determined if the compound 2-hDAAO inactive complex could be liberated by addition of a reducing agent. In these experiments, we repeated the experiment from Figure 4C , with the only difference that in this case, 5 mM βME was added during the dilution steps. In these experiments, dilution into βME restored hDAAO activity over the course of approximately 1 to 2 h ( Fig. 5C ). This result demonstrates that compound 2 inhibition of hDAAO is reversible upon exposure to a reducing environment.

Compound 2 Covalently Binds to All Five Cysteines of hDAAO

For a direct method to determine if compound 2 was binding covalently to hDAAO and to identify the amino acids modified, we incubated hDAAO with 100 µM compound 2, digested the protein–compound 2 complex with proteases, and performed LC-MS/MS. We observed evidence that all five cysteines of hDAAO were covalently bound to compound 2 (Suppl. Fig. S3). In each case, a 376 mass unit step in the y- or b-ion series provided confirmation of the cysteine plus compound 2 adduct. For C181, C263, and C264, only the +273 Dalton-modified cysteines were identified, while for C18 and C322, both modified and unmodified forms were identified. This result suggested that the compound 2 covalent modification of hDAAO cysteines was at high stoichiometry.

Discussion

In this study, we designed an HTS to identify hDAAO inhibitors sensitive to the presence of the hDAAO binding partner, pLG72. Approximately 85% of confirmed hDAAO inhibitors emerging from the screen were insensitive to pLG72 (class B compounds; Fig. 1B ). However, the remaining inhibitors were either more potent (class A) or less potent (class C) hDAAO inhibitors when pLG72 was present. Because of their structural homogeneity, availability from commercial sources, interest in the wider scientific community (see below), and rapid opportunity for exploring structure-activity relationships, we focused our efforts on characterizing the class C compounds. These compounds were both FAD and D-serine competitive hDAAO inhibitors, diverging from the known mechanism of previously described, active-site hDAAO inhibitors. Our biochemical characterization of the prototypic class C compound 2 indicated that this compound inhibited hDAAO via covalent binding of hDAAO cysteines.

Although the role of pLG72 remains enigmatic, our study provides further data that may help elucidate pLG72 function in relation to hDAAO. Our prior reports demonstrating that pLG72 was an hDAAO inhibitor used relatively high concentrations of hDAAO and pLG72 in near stoichiometric concentrations.⁷ In this study, to facilitate an assay suitable for HTS, our experimental design was different: we employed a classical pharmacological assay design (treating hDAAO as receptor and pLG72 as ligand) in which hDAAO was held at a relatively low (6.25 nM) concentration, and pLG72 concentration was increased until a functional outcome on hDAAO oxidative enzymatic activity could be measured. As a further difference, the newly designed assays were carried out over longer periods of time, with a 30-min hDAAO and pLG72 preincubation, followed by 1-h incubation after addition of substrates; reaction product was then measured in endpoint mode. Past assay formats from our laboratory included the 30-min preincubation but measured initial reaction rate⁷ rather than accumulation of product over time. Despite these differences, we confirmed that the primary effect of pLG72 was hDAAO inhibition. Furthermore, the general affinity of pLG72, based on our IC₅₀ determination of approximately 0.9 µM ( Fig. 1A ), was consistent with previous measurements of pLG72-hDAAO binding affinity using surface plasmon resonance.⁷ Although we cannot conclusively resolve the discrepancy between our experiments demonstrating hDAAO inhibition with prior experiments demonstrating hDAAO activation by pLG72, the time dependency of the interaction is critical for understanding the complex interaction between pLG72 and hDAAO. Under short incubation periods, pLG72 has no effect or even an activating effect on hDAAO. However, over longer periods of incubation (many minutes to hours), pLG72 acts as a type of “negative chaperone,” placing hDAAO in a conformation that renders the enzyme inactive and promotes hDAAO degradation in cellular environments.³⁰

Although it is unclear if class C compounds will be useful for clarifying the biological function of pLG72-hDAAO complexes, by seeking hDAAO inhibitors with differential sensitivity to the presence of pLG72, our approach revealed novel, nonactive site-directed hDAAO inhibitors. Unlike canonical active site hDAAO inhibitors, class C compounds displayed a mixed-inhibition mechanism of action with respect to both D-serine and FAD ( Fig. 3 ). With their immeasurably long off-rates ( Fig. 4 ), we hypothesized that the class C compounds were binding to hDAAO covalently. Indeed, compound 2 bound to all five hDAAO cysteines (Suppl. Fig. S3). This apparently nonspecific, pan-cysteine-binding effect is somewhat surprising in light of the observation that class C compounds displayed inhibitory specificity: they did not inhibit glucose oxidase ( Fig. 1 ), nor did they inhibit DASPO ( Fig. 2D ), an enzyme closely related in structure and enzymatic activity, but not in substrate specificity, to hDAAO.

Conceivably, flavoenzymes bound to FAD have cysteines buried within the protein structure inaccessible to solvent. Consistent with this model, crystal structures of hDAAO (e.g., protein data bank ID: 4QFC¹⁴) reveal that four of the five cysteines have <10% solvent-accessible surface (SAS; determined with Discovery Studio software [Accelrys, Burlington, MA]), while the remaining cysteine (C264) has <15% SAS. Under conditions of FAD association, these cysteines would have limited access to a modifying compound in solution. However, hDAAO, rather unique among flavoenzymes, samples the non-FAD-bound, apoprotein structure at physiological FAD concentrations, and previous work demonstrated that the apoprotein structure is an open conformation with reduced tertiary structure.³¹ Because FAD diminished the inhibitory potency of class C compounds ( Fig. 3 ), class C compounds may have better, or perhaps exclusive, access to hDAAO cysteines when hDAAO is in the apoprotein (FAD-unbound) conformation. Consistent with this model, stripping DASPO of FAD is very difficult,²⁰ and DASPO was fully active without adding FAD to the reaction. These observations suggest that DASPO rarely adopts the open, apoprotein conformation, perhaps forming the basis of its resistance to class C compounds. Finally, the defining feature of class C compounds, reduced inhibitory hDAAO potency in the presence of pLG72, is consistent with a model in which the pLG72-hDAAO interface shields hDAAO cysteines from class C compound access, either directly or via altered protein conformation. Further experiments would be necessary to test these biophysical hypotheses and determine the structure of the hDAAO-pLG72 complex.

The core substructure of the class C compounds, 2-phenyl-2,3-dihydro-1,2-benzothiazol-3-one, has been termed ebsulfur.²⁶ Biological studies of various analogues of these compounds have revealed general features consistent with our results. In experiments performed more than two decades ago, benzisothiazolone was shown to inhibit growth of various bacterial strains²⁸ and, consistent with our results in Figure 5 , glutathione reduced the biological potency of these compounds.²⁸ In more recent experiments, ebsulfur has been explored as a potential treatment for African trypanosomiasis. Ebsulfur inhibited the trypanosome enzyme trypanothione reductase via covalent inhibition, and trypanosome strains higher in glutathione levels were more resistant to ebsulfur treatment.²⁶ In unrelated studies, Trevillyan et al.²⁹ demonstrated that isothiazolone compounds, similar in structure to class C compounds, inhibited kinase targets via S-N bond reduction and covalent binding to cysteines in the kinase active site; these effects were blocked by the reducing agents βME and dithiothreitol. Taken together, our results are consistent with this general class of redox-sensitive compounds inhibiting protein targets via covalent binding to cysteine residues.

The identification of a class of compounds inhibiting hDAAO via a novel mechanism was initially quite exciting, since most hDAAO inhibitors occupy the enzyme’s active site. However, although covalent inhibition can be advantageous therapeutically, chemical reactivity generally is considered a negative attribute for a compound undergoing drug development. Because of their propensity to bind cysteines, isothiazolone compounds recently have been placed on lists of pan-assay interference compounds (PAINS), with advice to remove them from screening libraries.³² Despite these concerns, ebsulfur and other class C-like compounds have been pursued as potential therapies for a variety of indications. As described above, various studies have proposed ebsulfur as a potential antimicrobial agent. A recent study suggested that ebsulfur and the related compound ebselen act as lithium mimetics and, therefore, could potentially treat bipolar disorder.³³ A report from the Sanford-Burnham Institute described the development of class C compounds as inhibitors of the enzyme phosphomannose isomerase (PMI), which could have utility for the treatment of the congenital disorder of glycosylation type Ia.³⁴ In a subsequent report, the authors characterized the compound we termed compound 2 in cellular experiments, demonstrating that, although compound 2 was toxic at high concentrations, at nontoxic concentrations, it inhibited cellular PMI.³⁵ In our cellular experiments (not shown), we were unable to demonstrate a therapeutic window for compound 2; specifically, no compound 2 concentrations inhibited hDAAO in the absence of confounding toxicity. We observed this result even under conditions in which we used buthionine sulfoximine to deplete intracellular glutathione to potentially preserve compound 2 stability. Overall, while class C and class C–like compounds have appeal as inhibitors of several disease-relevant protein targets, opportunities to develop these compounds into drug-like molecules may be limited by their cysteine reactivity and sensitivity to reducing environments.

Footnotes

Acknowledgements

For preparing compounds 2, 9, 10, and 11, the authors thank the chemists at Shanghai ChemPartner: Yinghua Yang, Jun Wang, and Sufang Zhao. We also thank Richard Jones at MSBioworks for expertly conducting the proteomic experiment identifying cysteines modified by compound 2, as well as Michele Heffernan and Larry Hardy for critical review of the manuscript.

Supplementary material for this article is available on the Journal of Biomolecular Screening Web site at .

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Sacchi

Caldinelli

Cappelletti

. Structure-Function Relationships in Human D-Amino Acid Oxidase. Amino Acids 2012, 43, 1833–1850.

Papouin

Ladepeche

Ruel

. Synaptic and Extrasynaptic NMDA Receptors Are Gated by Different Endogenous Coagonists. Cell 2012, 150, 633–646.

Zhou

Sheng

NMDA Receptors in Nervous System Diseases. Neuropharmacology 2013, 74, 69–75.

Hopkins

S. C.

Campbell

U. C.

Heffernan

M. L.

. Effects of D-Amino Acid Oxidase Inhibition on Memory Performance and Long-Term Potentiation. Pharma. Res. Per. 2013, 1, e00007, 1-9.

Chumakov

Blumenfeld

Guerassimenko

. Genetic and Physiological Data Implicating the New Human Gene G72 and the Gene for D-Amino Acid Oxidase in Schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 13675–13680.

Drews

Otte

D. M.

Zimmer

Involvement of the Primate Specific Gene G72 in Schizophrenia: From Genetic Studies to Pathomechanisms. Neurosci. Biobehav. Rev. 2013, 37(Pt. 1), 2410–2417.

Sacchi

Bernasconi

Martineau

. pLG72 Modulates Intracellular D-Serine Levels through Its Interaction with D-Amino Acid Oxidase: Effect on Schizophrenia Susceptibility. J. Biol. Chem. 2008, 283, 22244–22256.

Sacchi

Cappelletti

Giovannardi

. Evidence for the Interaction of D-Amino Acid Oxidase with pLG72 in a Glial Cell Line. Mol. Cell. Neurosci. 2011, 48, 20–28.

Kvajo

Dhilla

Swor

D. E.

. Evidence Implicating the Candidate Schizophrenia/Bipolar Disorder Susceptibility Gene G72 in Mitochondrial Function. Mol. Psychiatry 2008, 13, 685–696.

10.

Benzel

Kew

J. N.

Viknaraja

. Investigation of G72 (DAOA) Expression in the Human Brain. BMC Psychiatry 2008, 8, 94.

11.

Otte

D. M.

Sommersberg

Kudin

. N-Acetyl Cysteine Treatment Rescues Cognitive Deficits Induced by Mitochondrial Dysfunction in G72/G30 Transgenic Mice. Neuropsychopharmacology 2011, 36, 2233–2243.

12.

Sparey

Abeywickrema

Almond

. The Discovery of Fused Pyrrole Carboxylic Acids as Novel, Potent D-Amino Acid Oxidase (DAO) Inhibitors. Bioorg. Med. Chem. Lett. 2008, 18, 3386–3391.

13.

Bhakuni

B. S.

Balkrishna

S. J.

Kumar

. An Efficient Copper Mediated Synthetic Methodology for Benzo[d]isothiazol-3(2H)-ones and Related Sulfur–Nitrogen Heterocycles. Tetrahedron Lett. 2012, 53, 1354–1357.

14.

Terry-Lorenzo

R. T.

Chun

L. E.

Brown

S. P.

. Novel Human D-Amino Acid Oxidase Inhibitors Stabilize an Active-Site Lid-Open Conformation. Biosci. Rep. 2014, 34, e00133.

15.

Fang

K. Q.

Hopkins

Jones

Benzo[d]isoxazol-3-ol DAAO Inhibitors. Patent WO2005066143A2. October 20, 2005.

16.

Shi

The First General Synthesis of N-Substituted 1,2-Benzisoxazolin-3-ones. Tetrahedron Let. 2000, 41, 2295–2298.

17.

Molla

Sacchi

Bernasconi

. Characterization of Human D-Amino Acid Oxidase. FEBS Lett. 2006, 580, 2358–2364.

18.

Molla

Bernasconi

Sacchi

. Expression in Escherichia coli and In Vitro Refolding of the Human Protein pLG72. Protein Expression Purif. 2006, 46, 150–155.

19.

Frattini

L. F.

Piubelli

Sacchi

. Is Rat an Appropriate Animal Model to Study the Involvement of D-Serine Catabolism in Schizophrenia? Insights from Characterization of D-Amino Acid Oxidase. FEBS J. 2011, 278, 4362–4373.

20.

Negri

Tedeschi

Ceciliani

. Purification of Beef Kidney D-Aspartate Oxidase Overexpressed in Escherichia coli and Characterization of Its Redox Potentials and Oxidative Activity towards Agonists and Antagonists of Excitatory Amino Acid Receptors. Biochim. Biophys. Acta 1999, 1431, 212–222.

21.

Simonic

Duga

Negri

. cDNA Cloning and Expression of the Flavoprotein D-Aspartate Oxidase from Bovine Kidney Cortex. Biochem. J. 1997, 322(Pt. 3), 729–735.

22.

Caldinelli

Molla

Bracci

. Effect of Ligand Binding on Human D-Amino Acid Oxidase: Implications for the Development of New Drugs for Schizophrenia Treatment. Protein Sci. 2010, 19, 1500–1512.

23.

Chang

S. L.

Hsieh

C. H.

Chen

Y. J.

. The C-Terminal Region of G72 Increases D-Amino Acid Oxidase Activity. Int. J. Mol. Sci. 2014, 15, 29–43.

24.

Zhang

J. H.

Chung

T. D.

Oldenburg

K. R.

A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. J. Biomol. Screen. 1999, 4, 67–73.

25.

Vlamis-Gardikas

Kandasamy

. Inhibition of Bacterial Thioredoxin Reductase: An Antibiotic Mechanism Targeting Bacteria Lacking Glutathione. FASEB J. 2013, 27, 1394–1403.

26.

Vodnala

S. K.

Gustavsson

A. L.

. Ebsulfur Is a Benzisothiazolone Cytocidal Inhibitor Targeting the Trypanothione Reductase of Trypanosoma brucei. J. Biol. Chem. 2013, 288, 27456–27468.

27.

Ferraris

Duvall

Y. S.

. Synthesis and Biological Evaluation of D-Amino Acid Oxidase Inhibitors. J. Med. Chem. 2008, 51, 3357–3359.

28.

Collier

P. J.

Ramsey

Waigh

R. D.

. Chemical Reactivity of Some Isothiazolone Biocides. J. Appl. Bacteriol. 1990, 69, 578–584.

29.

Trevillyan

J. M.

Chiou

X. G.

Ballaron

S. J.

. Inhibition of p56(lck) Tyrosine Kinase by Isothiazolones. Arch. Biochem. Biophys. 1999, 364, 19–29.

30.

Cappelletti

Campomenosi

Pollegioni

. The Degradation (by Distinct Pathways) of Human D-Amino Acid Oxidase and Its Interacting Partner pLG72—Two Key Proteins in D-Serine Catabolism in the Brain. FEBS J. 2014, 281, 708–723.

31.

Caldinelli

Molla

Sacchi

. Relevance of Weak Flavin Binding in Human D-Amino Acid Oxidase. Protein Sci. 2009, 18, 801–810.

32.

Baell

J. B.

Ferrins

Falk

. PAINS: Relevance to Tool Compound Discovery and Fragment-Based Screening. Aust. J. Chem. 2013, 66, 1483–1494.

33.

Singh

Halliday

A. C.

Thomas

J. M.

. A Safe Lithium Mimetic for Bipolar Disorder. Nat. Commun. 2013, 4, 1332.

34.

Dahl

Bravo

Sharma

. Potent, Selective, and Orally Available Benzoisothiazolone Phosphomannose Isomerase Inhibitors as Probes for Congenital Disorder of Glycosylation Ia. J. Med. Chem. 2011, 54, 3661–3668.

35.

Sharma

Ichikawa

. Phosphomannose Isomerase Inhibitors Improve N-Glycosylation in Selected Phosphomannomutase-Deficient Fibroblasts. J. Biol. Chem. 2011, 286, 39431–39438.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.76 MB

0.00 MB

High-Throughput Screening Strategy Identifies Allosteric,Covalent Human D-Amino Acid Oxidase Inhibitor

Abstract

Keywords

Introduction

Materials and Methods

Compound Synthesis and Protein Production

HTS

Additional Enzymatic Assays

LC-MS/MS Analytical Methods

LC-MS/MS Proteomic Methods

Results

pLG72-Directed hDAAO Activity Assay HTS

Biochemical Characterization of Class C Compound hDAAO inhibitors

Class C Compounds Inhibit hDAAO Covalently and Reversibly

Compound 2 Covalently Binds to All Five Cysteines of hDAAO

Discussion

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

References

Supplementary Material