Inhibitors of the Hydrolytic Enzyme Dimethylarginine Dimethylaminohydrolase (DDAH): Discovery, Synthesis and Development

Dimethylarginine dimethylaminohydrolase (DDAH) is a highly conserved hydrolytic enzyme found in numerous species, including bacteria, rodents, and humans. In humans, the DDAH-1 isoform is known to metabolize endogenous asymmetric dimethylarginine (ADMA) and monomethyl arginine (l-NMMA), with ADMA proposed to be a putative marker of cardiovascular disease. Current literature reports identify the DDAH family of enzymes as a potential therapeutic target in the regulation of nitric oxide (NO) production, mediated via its biochemical interaction with the nitric oxide synthase (NOS) family of enzymes. Increased DDAH expression and NO production have been linked to multiple pathological conditions, specifically, cancer, neurodegenerative disorders, and septic shock. As such, the discovery, chemical synthesis, and development of DDAH inhibitors as potential drug candidates represent a growing field of interest. This review article summarizes the current knowledge on DDAH inhibition and the derived pharmacokinetic parameters of the main DDAH inhibitors reported in the literature. Furthermore, current methods of development and chemical synthetic pathways are discussed.

In contrast, reduced ADMA concentrations have been demonstrated in other disease states, such as amyotrophic lateral sclerosis [51], multiple sclerosis [43], Alzheimer's disease and dementia [52,53]. ADMA has also been shown to have neuroprotective effects in a model of Parkinson's disease [54]. Since the identification of S-nitrosylation as an important mechanism of DDAH regulation, the discovery of L-homocysteine's ability to inhibit DDAH-1 led to the development of S-nitroso-L-homocysteine (HcyNO,9), an endogenous S-nitrosothiol and a source of endogenous NO [100][101][102]. Synthesis and characterisation of HcyNO as a bovine DDAH inhibitor was reported by Knipp et al. (2005) [102]. HcyNO was synthesized in near quantitative yield in a two-step reaction (Scheme 2) via alkaline hydrolysis of commercially available L-homocysteine thiolactone hydrochloride (7) to L-homocysteine (8) [103], followed by S-nitrosylation of L-homocysteine with sodium nitrite (NaNO 2 ) and acidification to afford HcyNO (9) [104]. Alternatively, S-nitrosylation can be undertaken directly on commercially available L-homocysteine (8). Mild reaction conditions are required, and excellent yields make HcyNO simple to prepare on a large scale. hypermethylation has the ability to suppress DDAH-2 transcription, while histone acetylation has the ability to up-regulate DDAH-2 transcription [83].

Pharmacological Induction of DDAH
As a result of diminished DDAH activity and/or accumulation of ADMA and L-NMMA in various disease states, several studies have investigated the effects of restoring or increasing DDAH expression. Statin treatment is associated with a significant reduction in ADMA concentrations [84][85][86], and an increase in both DDAH-1 and DDAH-2 expression and activity [87][88][89]. Another class of compounds enhancing ADMA metabolism via DDAH up-regulation is represented by the farnesoid X receptor (FXR) agonists. GW4064 upregulates both DDAH-1 and the cationic transporter CAT-1, leading to increased ADMA cellular uptake and metabolism, and decreased serum ADMA concentrations [15,90]. Another FXR agonist, INT-747, can increase DDAH-1 expression [91,92], however this is not associated with significant changes in ADMA and NO concentrations [91]. The lipid lowering drug probucol also lowers ADMA concentrations [93] via up-regulation of both DDAH expression and activity, and down-regulation of PRMT-1 expression [78,94]. Another lipid lowering drug, fenofibrate, decreases ADMA concentrations by restoring DDAH activity in conditions of oxidative stress [95]. Additionally, the oral antidiabetic drug pioglitazone and the β1 receptor blocker nebivolol both decrease serum ADMA concentrations by increasing DDAH-2 expression [96][97][98][99].

Pharmacological Inhibition of DDAH
Contrary to the comments in Section 3 above, there is a growing body of evidence to suggest that the modulation of DDAH enzyme activity may be utilized in conditions characterized by excessive NO production [36,67]. The remainder of this review will focus on highlighting the current knowledge associated with the main DDAH inhibitors reported in the literature to date.
HcyNO irreversibly inhibits DDAH by covalently binding to the putative catalytic cysteine with an inhibitory constant (Ki) of 690 μM [105]. Mass spectrometry analyses using different isotopic forms of S-nitroso-L-homocysteine (Hcy 15 NO and HcyN 18 O) and oxygen-18 labelled water (H2 18 O) demonstrated that HcyNO forms a N-thiosulfoximide adduct with cysteine 273 (Cys273) of bovine DDAH-1. The authors' proposed mechanism of inhibition requires the lone pair of electrons of the Scheme 3. Schematic of the reaction mechanism proposed by Knipp et al. [102] for the covalent inhibition of DDAH-1 by HcyNO (9) at Cys273. Adapted with permission from [102]. Copyright 2005 American Chemical Society.

2-Chloroacetamidine and N-But-3-ynyl-2-chloroacetamidine
Stone et al. [106] identified 2-chloroacetamidine (11) as a nonspecific inhibitor of two enzymes of the amidinotransferases superfamily, Pseudomonas aeruginosa DDAH (PaDDAH) and human peptydilarginine deaminase (PAD). 2-Chloroacetamidine (11, Scheme 4) is a commercially available compound bearing a 2-chloromethylene chain and an amidinium group, which partially resembles arginine. It can be synthesized via reaction of chloroacetonitrile (10) with sodium methoxide followed by treatment with ammonium chloride [107,108]. 2-Chloroacetamidine acts as a weak irreversible amidinotransferase inhibitor with greater affinity for PaDDAH, relative to PAD4 (Table 1). Data acquired by electrospray ionization mass spectrometry revealed the formation of an irreversible thiouronium-enzyme adduct whereby the loss of chlorine from the inhibitor is observed [106]. Cys273 SH to attack the S-nitroso group in HcyNO to eliminate a hydroxyl anion (Scheme 3). This mechanism is likely stabilized by neighbouring amino acids, namely His172. A further hydroxyl anion originating from water attacks the cationic sulfur of the conjugated intermediate, and subsequently rearranges to give the covalent thiosulfoximide adduct [102].

Scheme 3.
Schematic of the reaction mechanism proposed by Knipp et al. [102] for the covalent inhibition of DDAH-1 by HcyNO (9) at Cys273. Adapted with permission from [102]. Copyright 2005 American Chemical Society.
Amine module 15 was prepared in three steps: activation of the terminal hydroxyl group of 3-butyn-1-ol (12) through mesylation to afford 13 in 62% yield, followed by conversion to amine 15 via azide intermediate 14 in 49% yield across two steps. The imidic ester module 17 was obtained in moderate yields (approximately 48%) by reacting chloroacetonitrile (10) with anhydrous ethanol (16) in the presence of dry hydrochloric acid gas. Final coupling between amine 15 and the imidic ester 17 proceeds under basic aqueous conditions to afford probe N-but-3-ynyl-2-chloro-acetamidine (18) in reasonable yields (approximately 63%). To date, no data are available regarding the selectivity of N-but-3-ynyl-2-chloro-acetamidine for human DDAH-1 and the potential cross reactivity of this probe with PaDDAH.  (10) with anhydrous ethanol (16) in the presence of dry hydrochloric acid gas. Final coupling between amine 15 and the imidic ester 17 proceeds under basic aqueous conditions to afford probe N-but-3-ynyl-2-chloro-acetamidine (18) in reasonable yields (approximately 63%). To date, no data are available regarding the selectivity of N-but-3-ynyl-2-chloro-acetamidine for human DDAH-1 and the potential cross reactivity of this probe with PaDDAH.
High chemical and structural similarity between 4124W and the DDAH substrate L-NMMA exists (compound 21, Scheme 6, and compound 2, Figure 1), and although exhibiting weak DDAH inhibition (IC50 = 416 μM) [110], this prompted Rossiter et al. [112] to synthesize a further 34 compounds based on the 4124W structure. Furthermore, Rossiter et al. [112] reported an alternative synthetic approach, which although requiring three steps, is synthetically versatile and was applied to develop their library of 4124W analogues. Rossiter's strategy utilized protected intermediates to avoid complications associated with the purification of polar amino acids. Intermediate 24 (Scheme 7) is generated from commercially available N,N′-bis-tert-butoxycarbonylpyrazole-1H-carbox-amidine (22)
High chemical and structural similarity between 4124W and the DDAH substrate L-NMMA exists (compound 21, Scheme 6, and compound 2, Figure 1), and although exhibiting weak DDAH inhibition (IC50 = 416 μM) [110], this prompted Rossiter et al. [112] to synthesize a further 34 compounds based on the 4124W structure. Furthermore, Rossiter et al. [112] reported an alternative synthetic approach, which although requiring three steps, is synthetically versatile and was applied to develop their library of 4124W analogues. Rossiter's strategy utilized protected intermediates to avoid complications associated with the purification of polar amino acids. Intermediate 24 (Scheme 7) is generated from commercially available N,N′-bis-tert-butoxycarbonylpyrazole-1H-carbox-amidine (22)  High chemical and structural similarity between 4124W and the DDAH substrate L-NMMA exists (compound 21, Scheme 6, and compound 2, Figure 1), and although exhibiting weak DDAH inhibition (IC 50 = 416 µM) [110], this prompted Rossiter et al. [112] to synthesize a further 34 compounds based on the 4124W structure. Furthermore, Rossiter et al. [112] reported an alternative synthetic approach, which although requiring three steps, is synthetically versatile and was applied to develop their library of 4124W analogues. Rossiter's strategy utilized protected intermediates to avoid complications associated with the purification of polar amino acids. Intermediate 24 (Scheme 7) is generated from commercially available N,N 1 -bis-tert-butoxycarbonylpyrazole-1H-carbox-amidine (22) by Mitsunobu reaction with alcohol 23 containing the desired R 1 group, followed by reaction with (S)-tert-butyl 4-amino-2-((tert-butoxycarbonyl)amino)butanoate (25) to afford the protected amino acid analogue 26 (Scheme 7). Cleavage of the Boc protecting groups is achieved with a 4 M solution of HCl in 1,4-dioxane and produced 4124W (21) in moderate yields (approximately 44%) or analogues 27 in moderate to good yields (12%-67%).

N G -Substituted-L-Arginine Analogues
Rossiter et al. [112] also synthesized a library of N-substituted arginine analogues. A significant increase in inhibitor binding affinity was obtained for N G -(2-methoxyethyl)-L-arginine (commonly known as L-257 (compound 39; IC 50 = 22 µM; Scheme 9) and its corresponding methyl ester (also known as L-291 (compound 40; IC 50 = 20 µM; Scheme 9), relative to compound 27a (Scheme 7) that contains one less carbon within the main-chain. Interestingly, the benzyl ester of L-257 did not increase DDAH-1 inhibition as was observed with compound 29a. These data suggest that arginine analogues may bind differently within the putative DDAH-1 active-site. Moreover, neither the corresponding methyl amide of L-257, nor the (R)-isomer of L-257 (structures omitted) resulted in DDAH-1 inhibition.
The synthesis of L-257 (39) is represented in Scheme 9. The reaction sequence follows a similar approach to that described in the synthesis of the (S)-2-amino-4-(3-methylguanidino)butanoic acid analogues, using Boc protected amino acid analogues. Firstly, 2-methoxyethyl pyrazole carboxamidine (36) is prepared under Mitsunobu conditions [113]  Deprotection of the amino acid analogues 33 is carried out using the same acidic conditions described in the synthesis of Rossiter's 4124W analogues in Scheme 7 and provides analogues 34 in low to moderate yields (0.29%-41%). Each analogue's inhibitory potential was assessed based on the reduced conversion of [ 14 C]-NMMA to [ 14 C]-citrulline by rat DDAH. Screening experiments revealed N-monosubstitution to be more effective than disubstitution, and that a R 1 substituent comprising a N-2-methoxyethyl group increased inhibitor binding affinity (viz. compound 27a; IC50 = 189 μM; Scheme 7). Furthermore, a series of N-2-methoxyethyl guanidinobutanoate esters [112] was synthesized via thionyl chloride promoted esterification with alcohol containing the desired R 3 group (Scheme 7). The generated benzyl ester 29a resulted in substantial improvement in DDAH-1 inhibition (IC50 = 27 μM). A series of amide analogues based on structures 29 (structures omitted) exhibited poor DDAH-1 inhibitory potentials.

N G -Substituted-L-Arginine Analogues
Rossiter et al. [112] also synthesized a library of N-substituted arginine analogues. A significant increase in inhibitor binding affinity was obtained for N G -(2-methoxyethyl)-L-arginine (commonly known as L-257 (compound 39; IC50 = 22 μM; Scheme 9) and its corresponding methyl ester (also known as L-291 (compound 40; IC50 = 20 μM; Scheme 9), relative to compound 27a (Scheme 7) that contains one less carbon within the main-chain. Interestingly, the benzyl ester of L-257 did not increase DDAH-1 inhibition as was observed with compound 29a. These data suggest that arginine analogues may bind differently within the putative DDAH-1 active-site. Moreover, neither the corresponding methyl amide of L-257, nor the (R)-isomer of L-257 (structures omitted) resulted in DDAH-1 inhibition.
The synthesis of L-257 (39) is represented in Scheme 9. The reaction sequence follows a similar approach to that described in the synthesis of the (S)-2-amino-4-(3-methylguanidino)butanoic acid analogues, using Boc protected amino acid analogues. Firstly, 2-methoxyethyl pyrazole carboxamidine (36) is prepared under Mitsunobu conditions [113]   Both L-257 (39) and L-291 (40) inhibit DDAH-1 in vivo [112]. They are non-cytotoxic and selective for DDAH-1 against the three main isoforms of NOS [112,114] and arginase [115], an enzyme responsible for the metabolism of L-arginine into urea and ornithine [112,114,115]. L-257 (39) was utilized by Kotthaus et al. [114] as a validated active 'parent' compound for determining viable lead compounds, and to obtain initial structure-activity relationships (SARs) for those leads. In Both L-257 (39) and L-291 (40) inhibit DDAH-1 in vivo [112]. They are non-cytotoxic and selective for DDAH-1 against the three main isoforms of NOS [112,114] and arginase [115], an enzyme responsible for the metabolism of L-arginine into urea and ornithine [112,114,115]. L-257 (39) was utilized by Kotthaus et al. [114] as a validated active 'parent' compound for determining viable lead compounds, and to obtain initial structure-activity relationships (SARs) for those leads. In particular, the role of the N-substituent in DDAH inhibition was investigated as the guanidino moiety is considered to be important in the proposed catalytic mechanism and binding to the enzyme [116]. In the work undertaken by Kotthaus et al., the N G -(2-methoxyethyl) side-chain was varied along with the degree of guanidine substitution. Similarly, Tommasi et al. synthesized further L-257 derivatives with differing side-chains [117]. In both studies, the N G -(2-methoxyethyl) side-chain conferred inhibitors with activity toward DDAH-1. N-monosubstituted examples are represented in Table 2. Table 2. Inhibitory parameters for the N-monosubstituted L-arginine analogues with varied N-substituents (R groups) [114,117].
Molecules 2016, 21, 615 9 of 31 particular, the role of the N-substituent in DDAH inhibition was investigated as the guanidino moiety is considered to be important in the proposed catalytic mechanism and binding to the enzyme [116]. In the work undertaken by Kotthaus et al., the N G -(2-methoxyethyl) side-chain was varied along with the degree of guanidine substitution. Similarly, Tommasi et al. synthesized further L-257 derivatives with differing side-chains [117]. In both studies, the N G -(2-methoxyethyl) side-chain conferred inhibitors with activity toward DDAH-1. N-monosubstituted examples are represented in Table 2. Table 2. Inhibitory parameters for the N-monosubstituted L-arginine analogues with varied N-substituents (R groups) [114,117].
Kotthaus et al. [114] examined five N 5 -(1-iminoalk(en)yl)-L-ornithines with varying side-chain lengths and bond saturation for their ability to inhibit L-citrulline formation by human DDAH-1 using both high performance liquid chromatography and colorimetry. L-VNIO (54a, Figure 2) emerged as the best DDAH-1 inhibitor of this series of derivatives (97% inhibition at 1 mM, IC 50 = 13 µM, K i = 2 µM). However, increasing the main-chain length by one carbon atom from ornithine to lysine reduced its DDAH inhibitory potential.

Pentafluorophenyl (PFP) Sulfonate Esters
Vallance et al. [130] identified pentafluorophenyl (PFP) sulfonate esters 61 as nonspecific PaDDAH and bacterial arginine deaminase (ADI) inhibitors, after highlighting their structural similarity with the cysteine protease family of enzymes. This class of inhibitor can be obtained in moderate to good yields (44%-75%) via the 1,3-dipolar cycloaddition between nitrone 60 and the electron deficient vinyl-appended PFP sulfonate 57 (Scheme 13) [131]. The shelf-stable PFP vinylsulfonate 57 is synthesized in moderate yield (59%) from pentafluorophenol (55) and 2-chloroethane-1-sulfonyl chloride (56) [132], while various nitrones 60 can be generated from their corresponding aldehydes 58 by reaction with N-methyl hydroxylamine (59) under mild conditions and in good yield (for examples, please see [133,134]). The electron deficiency of the vinyl PFP sulfonate dipolarophile 57 affords the reversed regioselective 4C-substituted adduct of the isoxazolidine in compound 61, contrary to the 5C-substituted adduct typically observed with olefins. Although this was previously observed for electron deficient dipolarophiles [135][136][137][138], Caddick et al. [131] reported the 4C-substituted isoxazolidine 61 is increasingly favoured at higher temperatures. Each of the nitrones 60 reported by Caddick et al. [131] participates in cycloaddition with the vinyl PFP sulfonate 57, including C-aryland C-alkyl-N-methyl species with both electron-donating and electron-withdrawing substituents in addition to poly-and hetero-aromatic substituents. This robust synthetic methodology provides the ability to generate a structurally diverse library of PFP sulfonate esters 61 with moderate ease and high yields [130]. Several PFP sulfonate esters 61 have been screened for their activity as PaDDAH and ADI The electron deficiency of the vinyl PFP sulfonate dipolarophile 57 affords the reversed regioselective 4C-substituted adduct of the isoxazolidine in compound 61, contrary to the 5C-substituted adduct typically observed with olefins. Although this was previously observed for electron deficient dipolarophiles [135][136][137][138], Caddick et al. [131] reported the 4C-substituted isoxazolidine 61 is increasingly favoured at higher temperatures. Each of the nitrones 60 reported by Caddick et al. [131] participates in cycloaddition with the vinyl PFP sulfonate 57, including C-aryland C-alkyl-N-methyl species with both electron-donating and electron-withdrawing substituents in addition to poly-and hetero-aromatic substituents. This robust synthetic methodology provides the ability to generate a structurally diverse library of PFP sulfonate esters 61 with moderate ease and high yields [130]. Several PFP sulfonate esters 61 have been screened for their activity as PaDDAH and ADI inhibitors at two concentrations (500 µM and 50 µM), with five compounds emerging as relatively potent PaDDAH inhibitors (IC 50 = 16-58 µM, see [130] for structures). Vallance et al. revealed the mode of inhibition was competitive utilising time-dependence and reversibility experiments [130]. Interestingly, the PFP sulfonate esters 61 have been proposed as potential antibacterial agents, however no data has been reported identifying if PaDDAH inhibition modulates bacterial growth or viability. Likewise, no studies utilizing human DDAH have been reported identifying whether the PFP sulfonate esters exhibit specificity for bacterial PaDDAH. Non-specific PFP sulfonate ester binding to both human DDAH and PaDDAH would be counterproductive for their proposed application as antibiotics.

Indolylthiobarbituric Acid Derivatives
The indolylthiobarbituric acid derivatives 68 [139] were identified as PaDDAH inhibitors by a combination of virtual screening using a compound library comprising 308,000 structures and subsequent biological screening. Among the 109 highest scoring compounds, 90 were commercially available and used in colorimetric enzyme activity assays. Three compounds consisting of two chemical scaffolds were identified as potential PaDDAH inhibitors. One of these scaffolds, indolylthiobarbituric acid (68a), together with additional compounds 68b-c identified from a similarity search, emerged as promising inhibitors of PaDDAH (IC 50 = 2-16.9 µM). The most potent inhibitor of this series was SR445 (68c, IC 50 = 2 µM), and no member of this class of compounds showed any human DDAH inhibition [114].
This class of inhibitors is comprised of thiobarbituric acid module 64 and indolyl module 67 (Scheme 14). The thiobarbituric acid module 64 is synthesized using a modified version of the method described by Fisher and Mering in 1903, by reaction of N-substituted 2-thiourea 63 with diethyl malonate (62) [140,141]. The indolyl module 67 is alkylated by deprotonation of indole-3-carbaldehyde (65) and reaction with 2-chloro-N-furan-2-ylmethyl-acetimide (66) using a method previously reported (yield 83%) [142]. The aldehyde substituent of the indolyl derivative 67 is subsequently reacted with the thiobarbituric acid derivative 64 under acidic conditions to afford 68 (yields not reported) [139,143]. This produced a mixture of (E)-and (Z)-isomers about the linking double bond.  [130]. Interestingly, the PFP sulfonate esters 61 have been proposed as potential antibacterial agents, however no data has been reported identifying if PaDDAH inhibition modulates bacterial growth or viability. Likewise, no studies utilizing human DDAH have been reported identifying whether the PFP sulfonate esters exhibit specificity for bacterial PaDDAH. Non-specific PFP sulfonate ester binding to both human DDAH and PaDDAH would be counterproductive for their proposed application as antibiotics.

Indolylthiobarbituric Acid Derivatives
The indolylthiobarbituric acid derivatives 68 [139] were identified as PaDDAH inhibitors by a combination of virtual screening using a compound library comprising 308,000 structures and subsequent biological screening. Among the 109 highest scoring compounds, 90 were commercially available and used in colorimetric enzyme activity assays. Three compounds consisting of two chemical scaffolds were identified as potential PaDDAH inhibitors. One of these scaffolds, indolylthiobarbituric acid (68a), together with additional compounds 68b-c identified from a similarity search, emerged as promising inhibitors of PaDDAH (IC50 = 2-16.9 μM). The most potent inhibitor of this series was SR445 (68c, IC50 = 2 μM), and no member of this class of compounds showed any human DDAH inhibition [114].
This class of inhibitors is comprised of thiobarbituric acid module 64 and indolyl module 67 (Scheme 14). The thiobarbituric acid module 64 is synthesized using a modified version of the method described by Fisher and Mering in 1903, by reaction of N-substituted 2-thiourea 63 with diethyl malonate (62) [140,141]. The indolyl module 67 is alkylated by deprotonation of indole-3-carbaldehyde (65) and reaction with 2-chloro-N-furan-2-ylmethyl-acetimide (66) using a method previously reported (yield 83%) [142]. The aldehyde substituent of the indolyl derivative 67 is subsequently reacted with the thiobarbituric acid derivative 64 under acidic conditions to afford 68 (yields not reported) [139,143]. This produced a mixture of (E)-and (Z)-isomers about the linking double bond.

Ebselen
Ebselen (2-phenyl-1,2-benzisoselenazol-3(2H)-one, 73), a selenazole heterocyclic compound with well-known antioxidant properties [144], was recently identified as a potent DDAH-1 inhibitor. This was achieved by utilizing a high throughput screening (HTS) assay of two commercial libraries comprising 2446 compounds [145]. The HTS identified each compound's ability to inhibit the conversion of the synthetic substrate, S-methyl-L-thiocitrulline (SMTC) to citrulline and methanethiol by PaDDAH. Compounds exhibiting PaDDAH inhibitory potentials were validated for their ability to inhibit human DDAH-1 using ADMA as the substrate. The quantification of inhibition was

Ebselen
Ebselen (2-phenyl-1,2-benzisoselenazol-3(2H)-one, 73), a selenazole heterocyclic compound with well-known antioxidant properties [144], was recently identified as a potent DDAH-1 inhibitor. This was achieved by utilizing a high throughput screening (HTS) assay of two commercial libraries comprising 2446 compounds [145]. The HTS identified each compound's ability to inhibit the conversion of the synthetic substrate, S-methyl-L-thiocitrulline (SMTC) to citrulline and methanethiol by PaDDAH. Compounds exhibiting PaDDAH inhibitory potentials were validated for their ability to inhibit human DDAH-1 using ADMA as the substrate. The quantification of inhibition was measured using derivatized citrulline formation and colorimetric detection. Characterisation of PaDDAH and human DDAH-1 inhibition by ebselen was subsequently determined by dose-response, reversibility, and ESI-MS experiments.
While ebselen (73) is commercially available, it can be prepared via two main methods [146][147][148]. The method described in Scheme 15 [147,148] is preferred due to its shorter reaction times and elementary design. This method involves the treatment of benzoic acid (69) with thionyl chloride and aniline to afford benzanilide (70) followed by ortho-lithiation to dianion 71 using n-butyl lithium. Selenium is then inserted into the carbon-lithium bond of 71, forming dianion intermediate 72, whose cyclization to ebselen (73) is promoted by copper(II) bromide oxidant.  (73) is commercially available, it can be prepared via two main methods [146][147][148]. The method described in Scheme 15 [147,148] is preferred due to its shorter reaction times and elementary design. This method involves the treatment of benzoic acid (69) with thionyl chloride and aniline to afford benzanilide (70) followed by ortho-lithiation to dianion 71 using n-butyl lithium. Selenium is then inserted into the carbon-lithium bond of 71, forming dianion intermediate 72, whose cyclization to ebselen (73) is promoted by copper(II) bromide oxidant. Ebselen is reported to irreversibly inhibit both PaDDAH (IC50 = 96 ± 2 nM) and human DDAH-1 (IC50 = 330 ± 30 nM), forming a covalent selenosulfide bond with each enzyme's respective catalytic cysteine. In silico docking experiments suggest inhibitor binding is supported by π-stacking interactions with Phe76 in addition to other polar interactions involving His162. In vitro experiments demonstrated ebselen exhibits no effect on human DDAH-1 expression [145].
Ebselen is the most potent DDAH inhibitor with 'non-substrate-like' structure reported in the literature to date and demonstrates an anti-inflammatory response [149]. However, its application as a DDAH inhibitor is limited due to its poor DDAH selectivity. A wide range of molecular targets are modulated by ebselen including: NADPH oxidase 2 [150,151], H + /K + ATPase transporter [152], Ca 2+and phospholipid-dependent protein kinase C [151], lipoxygenase [153], thioredoxin reductase [154], and metallothionein [155]). Moreover, cellular toxicity in leukocytes and hepatocytes has been reported, further reducing the likelihood of ebselen's use in clinical practice [156,157].
Ebselen is the most potent DDAH inhibitor with 'non-substrate-like' structure reported in the literature to date and demonstrates an anti-inflammatory response [149]. However, its application as a DDAH inhibitor is limited due to its poor DDAH selectivity. A wide range of molecular targets are modulated by ebselen including: NADPH oxidase 2 [150,151], H + /K + ATPase transporter [152], Ca 2+ -and phospholipid-dependent protein kinase C [151], lipoxygenase [153], thioredoxin reductase [154], and metallothionein [155]). Moreover, cellular toxicity in leukocytes and hepatocytes has been reported, further reducing the likelihood of ebselen's use in clinical practice [156,157].

4-Hydroxy-2-Nonenal
4-Hydroxy-2-nonenal (4-HNE, 76) is a cytotoxic aldehyde generated from lipid peroxidation during inflammation [158,159] and is abundant in the body. 4-HNE is highly reactive toward proteinaceous nucleophilic groups, such as thiol and imidazole moieties. Nucleophilic attack affords Michael adducts, whilst Schiff bases are formed with primary amines (e.g., lysine residues) [160][161][162][163]. Different methods have been reported for the synthesis of 4-HNE (76) and its derivatives [164][165][166]. The method reported by Soulér et al. [166] is described here due to its simplicity, speed, and versatility (Scheme 16). The synthesis is completed in a single cross-metathesis reaction between octen-3-ol (75) and 2-propenal (74) in the presence of Hoveyda-Grubbs 2nd generation ruthenium catalyst under a nitrogen atmosphere and in dry oxygen-free solvent (Scheme 16). The reaction proceeds at room temperature for 25 h affording the (E)-isomer of 4-HNE (76) in 75% yield. 4-HNE is reported to react with cysteine and histidine residues and it was this reactivity that led Forbes et al. to investigate its use as a human DDAH-1 inhibitor [167]. In Forbes' experiments, DDAH activity was measured via both colorimetric and radioisotope activity assays, with 4-HNE showing dose-dependent inhibition (IC50 = 50 μM) and near complete inhibition at 500 μM. Mass spectrometry experiments elucidated the mechanism of inhibition, whereby the formation of a Michael adduct between 4-HNE and His173 of the DDAH-1 active-site occurs.

PD 404182
6H-6-Imino-(2,3,4,5-tetrahydropyrimido)[1,2-c]- [1,3] benzothiazine (PD 404182, 79), is a commercially available heterocyclic pyrimidobenzothiazine and a suitable lead compound for SAR studies due to its simple synthetic methodology [168]. PD 404182 (79) is a compound listed in the Library of Pharmacologically Active Compounds (LOPAC) and is reported to have antibacterial [168,169] and antiviral [170,171] activity. Ghebremariam et al. [61] utilized a colorimetric assay to screen a library of over 1200 compounds for their ability to inhibit the conversion of ADMA to L-citrulline and dimethylamine by recombinant human DDAH-1. Compounds with reasonable inhibitory potentials were subjected to kinetic characterisation via fluorimetric assay using the synthetic substrate, S-methylthiocitrulline. Reversibility and ESI-MS studies were additionally undertaken. PD 404182 was subsequently evaluated for its ability to reduce angiogenesis and to protect against lipopolysaccharide-induced NO production. PD 404182 (79) was found to be a potent irreversible competitive human DDAH-1 inhibitor (IC50 = 9 μM) and showed promise as an anti-inflammatory and antiangiogenic agent.
The synthesis of PD404182 can be achieved via two routes [172,173]. Both approaches generate reasonable to high yields (61%-88%) over several synthetic steps. The first approach involves the reaction of 2-(2′-haloaryl)-tetrahydropyrimidine 77 with carbon disulfide in the presence of sodium hydride, followed by hydrolysis of the carbamodithioate derivative 78 and subsequent treatment with cyanogen bromide to afford 79 (Scheme 17; Route A). Alternatively, the 2-(2′-haloaryl)tetrahydropyrimidine 77 can be reacted with tert-butylisothiocyanate (80)   4-HNE is reported to react with cysteine and histidine residues and it was this reactivity that led Forbes et al. to investigate its use as a human DDAH-1 inhibitor [167]. In Forbes' experiments, DDAH activity was measured via both colorimetric and radioisotope activity assays, with 4-HNE showing dose-dependent inhibition (IC 50 = 50 µM) and near complete inhibition at 500 µM. Mass spectrometry experiments elucidated the mechanism of inhibition, whereby the formation of a Michael adduct between 4-HNE and His173 of the DDAH-1 active-site occurs.

PD 404182
6H-6-Imino-(2,3,4,5-tetrahydropyrimido)[1,2-c]- [1,3]benzothiazine (PD 404182, 79), is a commercially available heterocyclic pyrimidobenzothiazine and a suitable lead compound for SAR studies due to its simple synthetic methodology [168]. PD 404182 (79) is a compound listed in the Library of Pharmacologically Active Compounds (LOPAC) and is reported to have antibacterial [168,169] and antiviral [170,171] activity. Ghebremariam et al. [61] utilized a colorimetric assay to screen a library of over 1200 compounds for their ability to inhibit the conversion of ADMA to L-citrulline and dimethylamine by recombinant human DDAH-1. Compounds with reasonable inhibitory potentials were subjected to kinetic characterisation via fluorimetric assay using the synthetic substrate, S-methylthiocitrulline. Reversibility and ESI-MS studies were additionally undertaken. PD 404182 was subsequently evaluated for its ability to reduce angiogenesis and to protect against lipopolysaccharide-induced NO production. PD 404182 (79) was found to be a potent irreversible competitive human DDAH-1 inhibitor (IC 50 = 9 µM) and showed promise as an anti-inflammatory and antiangiogenic agent.
The synthesis of PD404182 can be achieved via two routes [172,173]. Both approaches generate reasonable to high yields (61%-88%) over several synthetic steps. The first approach involves the reaction of 2-(2 1 -haloaryl)-tetrahydropyrimidine 77 with carbon disulfide in the presence of sodium hydride, followed by hydrolysis of the carbamodithioate derivative 78 and subsequent treatment with cyanogen bromide to afford 79 (Scheme 17; Route A). Alternatively, the 2-(2 1 -haloaryl)-tetrahydropyrimidine 77 can be reacted with tert-butylisothiocyanate (80) in the presence of sodium hydride to afford intermediate 81, followed by treatment with trifluoroacetic acid to afford 79 (Scheme 17; Route B). 4-HNE is reported to react with cysteine and histidine residues and it was this reactivity that led Forbes et al. to investigate its use as a human DDAH-1 inhibitor [167]. In Forbes' experiments, DDAH activity was measured via both colorimetric and radioisotope activity assays, with 4-HNE showing dose-dependent inhibition (IC50 = 50 μM) and near complete inhibition at 500 μM. Mass spectrometry experiments elucidated the mechanism of inhibition, whereby the formation of a Michael adduct between 4-HNE and His173 of the DDAH-1 active-site occurs.

PD 404182
6H-6-Imino-(2,3,4,5-tetrahydropyrimido)[1,2-c]- [1,3]benzothiazine (PD 404182, 79), is a commercially available heterocyclic pyrimidobenzothiazine and a suitable lead compound for SAR studies due to its simple synthetic methodology [168]. PD 404182 (79) is a compound listed in the Library of Pharmacologically Active Compounds (LOPAC) and is reported to have antibacterial [168,169] and antiviral [170,171] activity. Ghebremariam et al. [61] utilized a colorimetric assay to screen a library of over 1200 compounds for their ability to inhibit the conversion of ADMA to L-citrulline and dimethylamine by recombinant human DDAH-1. Compounds with reasonable inhibitory potentials were subjected to kinetic characterisation via fluorimetric assay using the synthetic substrate, S-methylthiocitrulline. Reversibility and ESI-MS studies were additionally undertaken. PD 404182 was subsequently evaluated for its ability to reduce angiogenesis and to protect against lipopolysaccharide-induced NO production. PD 404182 (79) was found to be a potent irreversible competitive human DDAH-1 inhibitor (IC50 = 9 μM) and showed promise as an anti-inflammatory and antiangiogenic agent.

Proton Pump Inhibitors
Proton H + /K + ATPase pump inhibitors (PPIs) are widely prescribed to inhibit gastric acid secretion and are characterized by a 2-pyridylmethylsulfinylbenzimidazole scaffold. The chemical structures of the commonly prescribed PPIs 82a-f are shown in Figure 3.

Proton Pump Inhibitors
Proton H + /K + ATPase pump inhibitors (PPIs) are widely prescribed to inhibit gastric acid secretion and are characterized by a 2-pyridylmethylsulfinylbenzimidazole scaffold. The chemical structures of the commonly prescribed PPIs 82a-f are shown in Figure 3. HTS experiments utilising a library of 130,000 compounds and recombinant enzyme identified PPIs as human DDAH-1 inhibitors [174,175]. Initial screening was undertaken using a colorimetric assay with ADMA as the substrate. Further characterization of compounds of interest was achieved via a fluorimetric assay using S-methylthiocitrulline [174,175]. Surface plasmon resonance (SPR) was additionally utilized to detect the binding of PPIs to amine-coupled DDAH-1 (CM5 sensor chip) [175]. All commercially available PPIs were found to reversibly inhibit DDAH-1 (IC50 = 51-63 μM). Interestingly, in vitro experiments using human endothelial cells and in vivo experiments in mice demonstrated that PPIs increase ADMA concentrations [175] further supporting the evidence that prolonged use of PPIs is associated with an increased risk of cardiovascular disease [176,177].
PPIs are comprised of pyridine and benzimidazole moieties with different substituents. Their syntheses are protected by current patents (with the exception of omeprazole (82a)). The original patent for omeprazole (82a) [178] details preparation of the 2-(chloromethyl)pyridine derivative 84. Condensation [179] of compound 84 with 2-mercaptobenzimidazole derivative 83 affords the thioether precursor 85, which is subsequently oxidized to the sulfoxide in omeprazole (82a) using conditions such as hydrogen peroxide and a vanadium catalyst [180] (Scheme 18; Route A).  HTS experiments utilising a library of 130,000 compounds and recombinant enzyme identified PPIs as human DDAH-1 inhibitors [174,175]. Initial screening was undertaken using a colorimetric assay with ADMA as the substrate. Further characterization of compounds of interest was achieved via a fluorimetric assay using S-methylthiocitrulline [174,175]. Surface plasmon resonance (SPR) was additionally utilized to detect the binding of PPIs to amine-coupled DDAH-1 (CM5 sensor chip) [175]. All commercially available PPIs were found to reversibly inhibit DDAH-1 (IC 50 = 51-63 µM). Interestingly, in vitro experiments using human endothelial cells and in vivo experiments in mice demonstrated that PPIs increase ADMA concentrations [175] further supporting the evidence that prolonged use of PPIs is associated with an increased risk of cardiovascular disease [176,177].
PPIs are comprised of pyridine and benzimidazole moieties with different substituents. Their syntheses are protected by current patents (with the exception of omeprazole (82a)). The original patent for omeprazole (82a) [178] details preparation of the 2-(chloromethyl)pyridine derivative 84. Condensation [179] of compound 84 with 2-mercaptobenzimidazole derivative 83 affords the thioether precursor 85, which is subsequently oxidized to the sulfoxide in omeprazole (82a) using conditions such as hydrogen peroxide and a vanadium catalyst [180] (Scheme 18; Route A).

Proton Pump Inhibitors
Proton H + /K + ATPase pump inhibitors (PPIs) are widely prescribed to inhibit gastric acid secretion and are characterized by a 2-pyridylmethylsulfinylbenzimidazole scaffold. The chemical structures of the commonly prescribed PPIs 82a-f are shown in Figure 3. HTS experiments utilising a library of 130,000 compounds and recombinant enzyme identified PPIs as human DDAH-1 inhibitors [174,175]. Initial screening was undertaken using a colorimetric assay with ADMA as the substrate. Further characterization of compounds of interest was achieved via a fluorimetric assay using S-methylthiocitrulline [174,175]. Surface plasmon resonance (SPR) was additionally utilized to detect the binding of PPIs to amine-coupled DDAH-1 (CM5 sensor chip) [175]. All commercially available PPIs were found to reversibly inhibit DDAH-1 (IC50 = 51-63 μM). Interestingly, in vitro experiments using human endothelial cells and in vivo experiments in mice demonstrated that PPIs increase ADMA concentrations [175] further supporting the evidence that prolonged use of PPIs is associated with an increased risk of cardiovascular disease [176,177].
PPIs are comprised of pyridine and benzimidazole moieties with different substituents. Their syntheses are protected by current patents (with the exception of omeprazole (82a)). The original patent for omeprazole (82a) [178] details preparation of the 2-(chloromethyl)pyridine derivative 84. Condensation [179] of compound 84 with 2-mercaptobenzimidazole derivative 83 affords the thioether precursor 85, which is subsequently oxidized to the sulfoxide in omeprazole (82a) using conditions such as hydrogen peroxide and a vanadium catalyst [180] (Scheme 18; Route A).  Further development of the synthetic method to yield omeprazole (82a) [180] identified that the coupling between the pyridine and benzimidazole moieties can be modified to simplify purification. Principally, this method avoids the formation of the thioether intermediate 85. Furthermore, the crude product containing omeprazole (82a) produced from the original synthesis is prone to discolouration during the oxidation step [180].  [117]. The synthesis of these derivatives was undertaken based on reports that identified binding of both the 4-halopyridine and benzimidazole fragments to human DDAH-1 [181]. A triazole linker was selected to connect these fragments via copper(I)-catalysed 'click' cycloaddition between the terminal alkyne 92 and azide 94 (Scheme 19). The alkyne 92 is synthesized by Sonogashira coupling of the alkyne derivative with the required halogenated species 91 [182,183]. The azide 94 is synthesized by reaction of diphenylphosphoryl azide with the required primary alcohol 93 [184].
Human DDAH-1 inhibition, albeit low (IC 50 = 422 µM, 57%, inhibition at 1 mM), was reported with triazole 95a. In general, 95a-b failed to show significant inhibitory potential and may be due to the triazole linker not facilitating a productive inhibitory binding conformation of the fragments in the enzyme active-site. Further development of the synthetic method to yield omeprazole (82a) [180] identified that the coupling between the pyridine and benzimidazole moieties can be modified to simplify purification. Principally, this method avoids the formation of the thioether intermediate 85. Furthermore, the crude product containing omeprazole (82a) produced from the original synthesis is prone to discolouration during the oxidation step [180].  [117]. The synthesis of these derivatives was undertaken based on reports that identified binding of both the 4-halopyridine and benzimidazole fragments to human DDAH-1 [181]. A triazole linker was selected to connect these fragments via copper(I)-catalysed 'click' cycloaddition between the terminal alkyne 92 and azide 94 (Scheme 19). The alkyne 92 is synthesized by Sonogashira coupling of the alkyne derivative with the required halogenated species 91 [182,183]. The azide 94 is synthesized by reaction of diphenylphosphoryl azide with the required primary alcohol 93 [184].
Human DDAH-1 inhibition, albeit low (IC50 = 422 μM, 57%, inhibition at 1 mM), was reported with triazole 95a. In general, 95a-b failed to show significant inhibitory potential and may be due to the triazole linker not facilitating a productive inhibitory binding conformation of the fragments in the enzyme active-site.  [117] with permission of the Royal Society of Chemistry.

Conclusions
Current evidence suggest that inhibiting the DDAH family of enzymes may be used as a therapeutic tool to indirectly inhibit the NOS enzymes via increasing concentrations of endogenous NOS inhibitors, such as ADMA and L-NMMA. Therefore the development and synthesis of DDAH inhibitors is emerging as a promising field of clinical research. Furthermore, engineering compounds that target bacterial DDAH may represent a suitable strategy for the development of new antimicrobial agents; however further investigations are required. This has led to different classes of DDAH inhibitors being reported within the last decade characterized by alternative features in terms of chemical structure, potency, enzyme selectivity, and mechanism of action. The importance placed on each of these features depends on the end application of the inhibitor. Table 3 [117] with permission of the Royal Society of Chemistry.

Conclusions
Current evidence suggest that inhibiting the DDAH family of enzymes may be used as a therapeutic tool to indirectly inhibit the NOS enzymes via increasing concentrations of endogenous NOS inhibitors, such as ADMA and L-NMMA. Therefore the development and synthesis of DDAH inhibitors is emerging as a promising field of clinical research. Furthermore, engineering compounds that target bacterial DDAH may represent a suitable strategy for the development of new antimicrobial agents; however further investigations are required. This has led to different classes of DDAH inhibitors being reported within the last decade characterized by alternative features in terms of chemical structure, potency, enzyme selectivity, and mechanism of action. The importance placed on each of these features depends on the end application of the inhibitor. Table 3 (40) are highly selective for DDAH-1, while ornithine derivatives such as L-VNIO (54a) or L-IPO (54b) are known to inhibit not only DDAH, but also arginase and the NOS family of enzymes. Furthermore, compounds such as ebselen (73) and 4-HNE (76), although very potent DDAH inhibitors, are well known to exhibit a range of off-target effects, thus are more suitable for experiments in vitro rather than in vivo.
In terms of the varied approaches to chemical synthesis of the DDAH inhibitors discussed here, many can be synthesized in moderate to excellent overall yields and use reaction sequences that can be easily modified to enable the design of a large number of analogues. The commercial availability of several of the compounds determined to be DDAH inhibitors (2-chloroacetamidine (11), ebselen (73) 4-HNE (76) and the PPIs 82) demonstrates the versatility of some of the synthetic routes for large scale production.
Based on the in vitro pharmacokinetic and/or enzyme selectivity data compiled in this review, there appear to be three promising human DDAH-1 inhibitors that have the potential to progress to clinical trial: ZST316 (K i = 1 µM), Cl-NIO (K i = 1.3 µM), and L-257 (K i = 13 µM) ( Table 3). Despite exhibiting the most potent inhibitory response in vitro, a paucity of data exists with respect to the in vivo pharmacokinetic and pharmacodynamic profiling of ZST316. Literature reports involving the enzyme selectivity of Cl-NIO shows considerable promise, with no measurable cross-reactivity in experiments specifically involving eNOS [129]. However, data are not available regarding this inhibitor's interactions with iNOS, nNOS, or the arginase family of enzymes. Contrary to this, data reported for L-257 by Kotthaus et al. [115] demonstrates that L-257 is highly selective for DDAH-1. The administration of L-257 in a rodent model of septic shock not only improved survival, but additionally improved haemodynamic and organ function [67], indicating a potential clinical role for L-257.
Interestingly, little is understood about the regulation and pharmacological modulation of human DDAH-2. As such, the development of improved metabolomic approaches for the sensitive measurement of DDAH-2 activity in vitro is an important step forward in profiling the endogenous substrates and biochemical role(s) of this important enzyme in humans.
Human DDAH-1 51-63 -- [175] Acknowledgments: The authors thank Flinders Medical Centre Foundation and Flinders Faculty of Health Sciences, the Scottish Universities Life Sciences Alliance (SULSA), and the NHS Grampian Endowment Fund for providing financial support which assisted with reading available evidence and conducting experimental work.
Author Contributions: The manuscript was conceived by A.A.M., R.B.M. and S.T. collected the primary data and compiled draft manuscripts. B.C.L. and A.A.M. supervised development of the manuscript, and assisted in data interpretation, manuscript evaluation and editing.

Conflicts of Interest:
The authors declare no conflicts of interest.

Abbreviations
The following abbreviations are used in this manuscript: