S-substituted 2-mercaptoquinazolin-4(3H)-one and 4-ethylbenzensulfonamides act as potent and selective human carbonic anhydrase IX and XII inhibitors

Abstract We evaluated the hCA (CA, EC 4.2.1.1) inhibitory activity of novel 4-(2-(2-substituted-thio-4-oxoquinazolin-3(4H)-yl)ethyl)benzenesulfonamides (compounds 2–20) towards the isoforms I, II, IX, and XII. hCA Isoforms were effectively inhibited by most of new compounds comparable to those of AAZ. Compounds 2 and 4 showed interestingly efficient and selective antitumor (hCA IX and hCA XII) inhibitor activities (KIs; 40.7, 13.0, and 8.0, 10.8 nM, respectively). Compounds 4 and 5 showed selective hCA IX inhibitory activity over hCA I (SI; 95 and 24), hCA IX/hCA II (SI; 23 and 5.8) and selective hCA XII inhibitory activity over hCA I (SI; 70 and 44), hCA XII/hCA II, (SI; 17 and 10) respectively compared to AAZ. Compounds 12–17, and 19–20 showed selective inhibitory activity towards hCA IX over hCA I and hCA II, with selectivity ranges of 27–195 and 3.2–19, respectively, while compounds 12, 14–17, and 19 exhibited selective inhibition towards hCA XII over hCA I and hCA II, with selectivity ratios of 48–158 and 5.4–31 respectively, compared to AAZ. Molecular docking analysis was carried out to investigate the selective interactions among the most active derivatives, 17 and 20 and hCAs isoenzymes. Compounds 17 and 20, which are highly selective CA IX and XII inhibitors, exhibited excellent interaction within the putative binding site of both enzymes, comparable to the co-crystallized inhibitors. Highlights Quinazoline-linked ethylbenzenesulfonamides inhibiting CA were synthesised. The new molecules potently inhibited the hCA isoforms I, II, IV, and IX. Compounds 4 and 5 were found to be selective hCA IX/hCA I and hCA IX/hCA II inhibitors. Compounds 4 and 5 were found to be selective hCA XII/hCA I and hCA XII/hCA II inhibitors. Compounds 12–17, 19, and 20 were found to be selective hCA IX/hCA I and hCA IX/hCA II inhibitors. Compounds 12, 14–17, 19 were found to be selective hCA XII/hCA I and hCA XII/hCA II inhibitors. Graphical Abstract Compounds 4 and 5 are selective hCA IX and XII inhibitors over hCA I (selectivity ratios of 95, 23, and 24, 5.8, respectively) and hCA II (selectivity ratios of 70, 17, and 44, 10 respectively). Compounds 12–17, and 19–20 are selective hCA IX inhibitors over hCA I (selectivity ratios of 27-195) and hCA II (selectivity ratios of 3.2-19). Compounds 12, 14–17 and 19 are also selective hCA XII inhibitors over hCA I (selectivity ratios of 48-158) and hCA II (selectivity ratios of 5.4-31).


Introduction
Carbonic anhydrases (CAs; EC 4.2.1.1) constitute the superfamily of metalloenzymes that catalyse the CO 2 hydration and dehydration reactions. CAs are classified into eight genetically distinct families, named a-, b-, c-, d-, f-, g-, Õand i-CAs 1,2 . 15 a-class CA isozymes have been detected in humans, which are further classified into four different subsets on the basis of their subcellular localisation-CA I, II, III, VII, VIII, X, XI, XIIII are cytosolic proteins, CA IV is a glycosylphosphatidylinositol (GPI)-anchored protein, CA VA and VB are located in the mitochondrial matrix, CA VI is secreted, and CA IX, XII and XIV are trans-membrane isoforms [1][2][3] . Human CAs (hCAs) are spread in the human body, and are implicated in a plethora of essential physiological processes. Therefore, the dysregulated expression and/or activity of the CAs can lead to various pathological conditions 2 . CA II is the most physiologically relevant CA isoform, implicated in various disorders including cerebral oedema, glaucoma (such as CA XII), and epilepsy. It is conversely off-target, as CA I, when targeting tumours where CA IX and XII are overexpressed and represent validated targets to combat the growth of both primary tumours and metastasis 4,5 . The high structural similarities between various CA isoforms necessitate high selectivity in the design of small-molecule anti-CA drugs for the treatment of diseases associated with CA dysregulation, to minimise the side effects 3 . Benzene sulphonamides are one of the bestknown molecules clinically used as CA inhibitors. Additionally, "SLC-011 (Figure 1), a benzenesulfonamide, is a selective CA IX/XII inhibitor currently being evaluated in a Phase I trial for the treatment of solid, metastatic tumors" [6][7][8][9][10] . Sulphonamide derivatives are not only one of the most preferred CA inhibitor classes 9,11-23 , but also important COX-2 inhibitors and antitumor agents 17,19,[24][25][26] . The quinazolinone scaffold is also used widely across medicinal chemistry [27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43] . (6-Iodo or 7-flouro-2-merqapto-4-(3H)-quinazolinone3-yl)-benzenesulfonamides (A, Figure 1) have been shown to potently inhibit CA I, II, IX, and XII 44,45 . A number of 2-((3-benzyl-4-oxo-3,4-dihydroquinazolin-2-yl)thio)-N-(4-sulfamoylphenethyl)anildes (B, Figure 1) also showed potent inhibitory activity against different hCA isoforms 38 . The 2-mercapto-4(3H)quinazolinone derivatives containing ethylsulfonamide tail (C, Figure 1) showed strong inhibitory activity against different hCA isoforms with low-concentration inhibition constants Here, we studied 2-mercaptoquinazolinone, (C, Figure 1) a slightly polar and non-selective hCA inhibitor. Because the sulfhydryl group has been reported to be associated with various metabolic and pharmacological problems 46-49 , we used a 2-mercaptoquinazolinone scaffold bearing an ethylsulfonamide head with alkylation of the thione group with a terminal lipophilic moiety, so that it can interact selectively with CA through both, hydrogen and hydrophobic interactions. Here, we synthesised various derivatives of 2-mercaptoquinazolinone (2-20, Figure 1) with different selectivity criteria for the hCA inhibitors, particularly for the tumor-associated hCA IX and hCA XII. The role of alkyl substituent in 2-mercaptoquinazolinone was computationally analysed and the conserved residues responsible for the target selectivity were identified.

CA inhibition
The hCA I, II, IX, and XII isoenzyme inhibition assays were performed according to the reported method using the SX.18 MV-R stopped-flow instrument (Applied Photophysics, Oxford, UK) [52][53][54] . All CA isoforms were recombinant isoforms obtained in-house, as reported earlier 55,56 .

Molecular docking
3.4.1. Molecular docking of compounds 17 and 20 with CA IX and CA XII isoenzymes To further investigate the interactions between the selected active compounds 17 and 20 with the hCAs targets, we performed docking simulations into the binding pockets of the hCA isoforms, IX and XII, using the MOE Suite 65 (data are summarised in Figures 2 and 3).
Both the compounds 17 and 20 were shown to directly interact with the zinc ion of CA IX and CA XII isoenzymes, via the sulphonamide anion of the active sites of both enzymes. However, the contributions of the quinazoline scaffold and the terminal bulky thioether fragments interaction are different, based on the CA isoform. In CA IX, the quinazoline ring of compound 20 interacts with the Gln71 residue through a stable hydrogen bond, and gets accommodated in the hydrophobic pocket lined by the Val121, Val130, Leu134, and Leu91 residues, thereby stabilising the binding (Figure 2, lower panel). In addition, the terminal p-chlorobenzamide fragment formed a hydrophobic interaction with the Leu91 residue ( Figure 2, lower panel). In contrast, compound 17 was shown to bind similarly to the pocket of CA IX, except the unfavourable orientation of the quinazoline carbonyl moiety of compound 17 towards the hydrophobic pocket formed by Leu91 residue in CA IX (Figure 2, upper panel). Also, the benzamide core showed a polarnonpolar interaction with the Leu91 and Thr73 residues, as the bulky side chain causes steric hindrance, inducing conformational changes in the bulky thioether tail and the quinazoline groups Results also showed different interactions between CA XII and compounds 17 and 20 ( Figure 3). The carbonyl group on the quinazoline ring in compound 17 was stabilised by direct hydrogen bonding with the target residue Ser132 of CA XII (Figure 3, upper panel). In addition, the Lys67 residue showed favourable hydrophobic binding to the quinazoline core of compound 17. The trimethoxybenzamide group of compound 17 was accommodated in the polar pocket of CA XII that included Ser132 and Thr133 residues (Figure 3, upper panel). The placement of compound 20 within the CA XII pocket was not favoured, particularly because the quinazoline ring of compound 20 was trapped between the polar pocket of CA lined by the Ser135, Gln92, and Ser132 residues (Figure 3, lower panel). Therefore, this interaction causes an energetically unfavourable change in the terminal benzamide and quinazoline scaffold of compound 20, which could be responsible for the decreased inhibitory activity of compound 20 (Figure 3, lower panel).

Molecular orbital analyses
According to the frontier molecular orbital theory, HOMO and LUMO are the most important orbitals found in a molecule, as they can affect its biological activity, the molecular reactivity, the ionisation and the electron affinity [68][69][70] . The molecular orbital analysis of the representative compounds 4, 17, and 20 ( Figure 4) as an active and selective derivatives was done by exploring their structureselectivity relationship. The electron transition from HOMO to LUMO occurs freely when the energy gap is small. The HOMO-LUMO energy gap for the compounds 4, 17, and 20 was calculated to be À0.3125, À0.2834, and À0.28949 eV, respectively. The negative energy values are indicative of a stable structure and confirm the eventual charge transfer interactions. The distributions and energy levels of the HOMO-LUMO orbitals computed for the abovementioned compounds are represented in Figure 4. HOMO and LUMO orbitals are mainly delocalised in the carbon and nitrogen of the quinazoline scaffolds and the sulphur ether atoms in the active compound 4. While they are mainly delocalised in the S-linker of the benzamide moiety, ring substituents in the compounds 17 and 20 reverse their interactions with the enzyme isoforms. These results indicate that the affinity of the selective compounds for the CA IX and CA XII binding sites could be because of the involvement of the thioether moiety, and that the quinazoline moiety could mostly provide the structural basis and the lipophilic function, contributing strongly to their selectivity. In addition, the low HOMO-LOMO energy gap suggests that the molecules have high stability and are in their lowest energy conformation.