Synthesis of N-(Methoxycarbonylthienylmethyl)thioureas and Evaluation of Their Interaction with Inducible and Neuronal Nitric Oxide Synthase

Two isomeric N-(methoxycarbonylthienylmethyl)thioureas were synthesised by a sequence of radical bromination of methylthiophenecarboxylic esters, substitution with trifluoroacetamide anion, deprotection, formation of the corresponding isothiocyanates and addition of ammonia. The interaction of these new thiophene-based thioureas with inducible and neuronal nitric oxide synthase was evaluauted. These novel thienylmethylthioureas stimulated the activity of inducible Nitric Oxide Synthase (iNOS).


Introduction
Nitric oxide (•NO) is biosynthesised from L-arginine in two steps catalysed by nitric oxide synthase (NOSs) giving L-citrulline as a co-product. There are three isoforms, endothelial NOS (eNOS) and neuronal NOS (nNOS) (constitutive Ca 2+ -dependent forms) and an inducible form (iNOS).
Underactivity and overactivity of each of these isoforms can be associated with disease states.
Excessive NO production by eNOS within blood vessel walls is thought to be the basis for conditions such as septic-and cytokine-induced circulatory shock [1]. However, if too little NO is produced, this can lead to conditions such as high blood pressure, angina and impotence [2]. An overexpression of nNOS in circulating neutrophils has been found in patients with Parkinson's disease [3] and nNOS activity is thought to be linked to migraine headaches. NO production by iNOS is essential for the defence mechanism of an organism; however, overactivity of iNOS has been related to several pathological conditions, including cancer, arthritis and diabetes [4][5]. Thus Selective inhibition of each of these three isoforms (eNOS, nNOS and iNOS) has potential applications in medicinal chemistry. Most potent inhibitors of NOS ligate to the haem iron at the active sites of the enzymes, through amidines and guanidines and through (iso)thioureas. Early inhibitors included close analogues of L-Arg, such as N δ -(iminoethyl)-L-ornithine (L-NIO) [6] and N ω -nitro-L-arginine methyl ester (L-NAME). [7] More recent iNOS-selective inhibitors contain cyclic amidines to bind to the haem iron but lack the amino-acid motif of L-Arg [8,9]. Most interesting are 1400W (1) [10] (Figure 1), an N-benzylacetamidine which is a highly selective inhibitor of rat iNOS, and its lower homologue 2 [11] (Figure 1), an N-phenylacetamidine which is selective for inhibition of nNOS. We recently reported [12] a series of N-benzyland N-phenyl-thioureas and analogous 2-(substituted-amino)-4,5-dihydrothiazoles; several of these were moderate inhibitors of iNOS but were without activity against nNOS. More interestingly, N-benzylthioureas 3 ( Figure 1) and 2-benzylamino-4,5-dihydrothiazoles 4 ( Figure 1) caused significant stimulation of iNOS activity when the agents were added to the enzyme preparation simultaneously with the substrate. Therefore, in this work, our aim was to synthesis a new target in which benzene has been deplaced with the approximately isosteric thiophene. This rigidity introduced into the target is designed to maximise interaction and, more importantly, to differentiate between various NOS isoforms. Information from potent but non-isoformselective NOS inhibitors suggested that 4,5-dihydrothiazole, thiourea and imidazole are good head groups for haem-iron-binding. So, This work extended our previous work into two exemplary Nthienylmethylthioureas.

Results and Discussion
Retrosynthetic analysis suggested that the commercially available 2-methylthiophene (5) would be an ideal starting material. Abstraction of the relatively acidic thiophene 2-proton by butyl lithium would give an appropriate nucleophile for the reaction with a reactive carbonyl electrophile. Therefore, lithiation of 5 with butyl lithium at 0 °C, followed by quench with ethyl chloroformate (a reactive carboxy electrophile equivalent) gave the required ester 6 in 68% optimised yield (Scheme 1 A similar lithiation of 5 with butyl lithium, followed by quench with methyl chloroformate, afforded mixture of products. Thus the precise nature of the trapping electrophile was important. Bromination using N-bromosuccinimide (NBS) in the presence of perchloric acid introduced functionality to the methyl group, although the bromomethyl analogue 7 was obtained in only 7% isolated yield; ethyl 4-bromo-5-methylthiophene-2-carboxylate and ethyl 4-bromo-5-bromomethylthiophene-2-carboxylate were also identified in the reaction mixture. Thiourea is nucleophilic at sulfur; thus direct displacement of the bromine of 7 would give the isothiourea, rather than the target thiourea 13, so the thiourea had to be introduced stepwise. Replacement of the bromine with the anion derived from trifluoroacetamide provided one nitrogen in 8. From here, no conditions were found which would remove the trifluoroacetyl masking group from the primary amine while leaving the ester unscathed.
Hence vigorous base-hydrolysis cleaved both the amide and the ester to give the highly polar 5-(aminomethyl)thiophene-2-carboxylate which could only be isolated from the reaction mixture after convertion in situ to the Cbz-protected derivative 9. Now, the methyl ester 10 was formed and the Cbz was removed quantitatively and selectively, using hydrogen bromide, to afford 11. From here, our previously developed method [12][13][14] was employed to convert the primary amine firstly to the isothiocyanate 12 (with thiophosgene) and then to the required thiourea 13. Oxidation of this mixture to the carboxylic acids 16a,b was achieved with silver(I) oxide but this also failed to allow isolation of the pure required regioisomer 16b. Similarly, quench of the lithiated species derived from 14 with methyl chloroformate gave a 1:4 mixture of the esters 17a and 17b, respectively. Unfortunately, this mixture was again inseparable. Radical bromination of the mixture was achieved by treatment with one equivalent of NBS, using dibenzoyl peroxide as radical initiator. However, a mixture of unreacted starting material, ring-brominated products, and side-chain brominated (and polybrominated) products were produced. Repeated attempts using two equivalents of NBS gave higher percentages of ring-brominated products. Separation of the mixture via chromatography afforded 4.5% of 18. As in the 2,5-disubstituted sequence, reaction with trifluoroacetamide anion gave a moderate yield of the secondary amide 19 but, despite the presence of a large excess of the nucleophile, a small quantity of the dialkylated product 20 was also isolated. Hydrolysis of the trifluoroacetamide and the ester functional groups of 19 was carried out simultaneously using sodium hydroxide. In the same reaction flask, a Schotten-Baumann reaction was carried out using benzyl chloroformate, giving, after acid work-up, the Cbz compound 21 in 82% yield. This was esterified with methanol and H 2 SO 4 to give the ester 22 in high yield. As before, the protecting Cbz group of 22 was cleaved quantitatively upon treatment with hydrogen bromide in acetic acid.

Biological Evaluation
Compounds 13 and 25 were evaluated for their effects on the activity of rat brain nNOS and of recombinant human iNOS (hiNOS) at a test concentration of 100 µM using the L-[U-14 C]-arginine to L-[U-14 C]-citrulline conversion assay. 1400W 1 [10] and N-(3-aminobenzyl)thiourea 3 (R = 3-NH 2 ) [12] were used for comparison. Compounds were evaluated either by adding the radiolabelled substrate simultaneously with the inhibitor or with a pre-incubation of 10 min before the substrate was added. This pre-incubation has been reported to be optimum for the inhibitory activity of 1400W 1 [10]. This pre-incubation was investigated since test compounds can be considered to be analogues of the slow inhibitor 1400W 1.
As expected, 1400W 1 was a potent inhibitor of iNOS activity, both with and without preincubation with the enzyme. N-(3-Aminobenzyl)thiourea 3 was inactive towards the rat nNOS isoform but, as noted previously [12], stimulated formation of L-[U-14 C]-citrulline from L-[U-14 C]-arginine when the agent was added simultaneously with the substrate. The N-(thienylmethyl)thioureas 13 and 25 showed similar activity to that of 3, stimulating human iNOS activity but not modulating the activity of this isoform at all when the compounds were pre-incubated with the enzyme before addition of the substrate. Neither 13 nor 25 affected the activity of rat nNOS, with or without pre-incubation. -

Enzyme study
Measurements of the inhibitory activity of the test compounds against nNOS was perfomed using an enzyme preparation from rat brain (in which the large majority of the NOS activity is nNOS), whereas the assay of activity against iNOS was performed using preparation of recombinant human iNOS (hiNOS) overexpressed in an HT1080 cell line as described previously by us [12].
The assay used the conversion of L-[U-14 C]-arginine to L-[U-14 C]-citrulline as described previously by us [16]. Compounds were evaluated at 100 μM and assays were performed in two modes, simultaneous addition of the test compound to the enzyme preparation and of L-[U-14 C]-arginine (to start the enzymic reaction) and pre-incubation of the test compound with the enzyme preparation for rpm for 5 min to pellet the resin. This process was repeated twice, after which the cytosol was treated as free of endogenous arginine and was used for assays of inhibition, using the usual protocol with and without pre-incubation of the test compounds with the preparation. The results are shown in Table 1 as the mean of triplicate experiments SEM.

Conclusions
In this work, we have extended the range of compounds that cause this apparent stimulation of iNOS activity from benzylthioureas [12] into thienylmethylthioureas. These results are significant and will be interesting for many scientists.
Several synthetic guanidines and N-hydroxyguanidines can act as artificial substrates for nitric oxide synthases [19][20]. However, since the assay used by us does not measure nitric oxide production but rather measures conversion of radiolabelled L-Arg to radiolabelled L-Cit, it is clear that the increased activity of the iNOS is a genuine stimulation. Therefore, one might speculate that these compounds interact with the enzyme through more than one binding site under different conditions; binding to one site would have a stimulatory effect while binding to the other would have inhibitory effect. The binding to the allosteric stimulatory site may be fast, whereas inhibitory binding of some known inhibitors (e.g. 1400W) to the catalytic haem site is known to be slow [10].
Recently, it was demonstrated that 1400W is not a slow-binding inhibitor but a time-dependent inactivator of iNOS [22]. Therefore our result could be also explained as follows: Binding is rapid, but initially reversible; the rate-determining step is inactivation of the enzyme by heme modification. Therefore, the "loss" of stimulation after 10 minutes is because the enzyme is becoming inactivated. This would then explain the stimulation effects reported without preincubation.
Although most effort in this area has gone into inhibiting biosynthesis of •NO, increasing biosynthesis may have some therapeutic value. Nitric oxide blocks the expression of NADP oxidase, an enzyme important in the etiology of cardiovascular disease [23]. Also, sildenafil augments •NO concentrations by inhibiting phosphodiesterase-5 (PDE-5) and is useful in treating erectile disfunction [24] and, possibly, pulmonary hypertension and acute respiratory distress syndrome (ARDS) [25].
Direct stimulation of NOS isoforms may provide an alternative approach to these beneficial effects.
Also, the thienylmethylthioureas and benzylthioureas reported here and previously [12] will be used in a future work to provide a pharmacophore from which more potent and more site-selective ligands may be designed by modeling and docking studies.