Effect of Sulfonamides and Their Structurally Related Derivatives on the Activity of ι-Carbonic Anhydrase from Burkholderia territorii

Carbonic anhydrases (CAs) are essential metalloenzymes in nature, catalyzing the carbon dioxide reversible hydration into bicarbonate and proton. In humans, breathing and many other critical physiological processes depend on this enzymatic activity. The CA superfamily function and inhibition in pathogenic bacteria has recently been the object of significant advances, being demonstrated to affect microbial survival/virulence. Targeting bacterial CAs may thus be a valid alternative to expand the pharmacological arsenal against the emergence of widespread antibiotic resistance. Here, we report an extensive study on the inhibition profile of the recently discovered ι-CA class present in some bacteria, including Burkholderia territorii, namely BteCAι, using substituted benzene-sulfonamides and clinically licensed sulfonamide-, sulfamate- and sulfamide-type drugs. The BteCAι inhibition profile showed: (i) several benzene-sulfonamides with an inhibition constant lower than 100 nM; (ii) a different behavior with respect to other α, β and γ-CAs; (iii) clinically used drugs having a micromolar affinity. This prototype study contributes to the initial recognition of compounds which efficiently and selectively inhibit a bacterial member of the ι-CA class, for which such a selective inhibition with respect to other protein isoforms present in the host is highly desired and may contribute to the development of novel antimicrobials.


Introduction
Carbonic anhydrases (CAs, EC 4.2.1.1) are essential metalloenzymes in nature, which speed up a fundamental reaction for all living organisms, the hydration of a molecule of carbon dioxide (CO 2 ) into bicarbonate (HCO 3 − ) and proton (H + ), according to the following chemical reaction [1][2][3][4][5][6][7]: CO 2 + H 2 O HCO 3 − + H + . The CA superfamily is grouped into eight CA classes indicated with the Greek letters (α, β, γ, δ, ζ, η, θ and ι), whose distribution is very variegated from the most complex organisms (plants and animals) to the simplest ones (bacteria and archaea) [1][2][3][4][5]. The genome of mammals, for example, encodes only for the α-CA class, of which 15 isoforms have been identified, which accomplish specialized functions in various tissues and organs [8][9][10][11][12]. Mammal breathing is a compelling example of the importance of these enzymes. In mammals, at the peripheral tissue level, the CO 2 produced by aerobic metabolism leaves the cells and enters the bloodstream due to a pressure gradient effect [13]. About 90% of CO 2 flows into red blood cells and is converted into bicarbonate by CAs. Then, the produced HCO 3 − comes out from the red cell through an anion exchanger (AE) protein and is transported by the bloodstream to the lungs [13]. At the alveolar level, the concentration of CO 2 is lower than in peripheral tissues, while there is a higher concentration of bicarbonate that is pumped into the red blood cell. Here, through the action of the inverse reaction catalyzed by CAs, the bicarbonate is transformed into water and CO 2 [13]. The CO 2 produced in this way is released into the bloodstream and, passing through the alveolus walls, is exhaled [13]. These reactions can also occur without the enzyme but carbonic anhydrase increases the reaction speed up to a million times (k cat = 10 6 s −1 ) [7]. At a physiological pH value, the spontaneous reversible CO 2 hydration reaction in the absence of the catalyst has an effective first-order rate constant of 0.15 s −1 , while the reverse reaction shows a rate constant of 50 s −1 [7]. Mammalian CAs are also involved in other important physiological processes, such as gluconeogenesis, lipogenesis and ureagenesis; transport of CO 2 /bicarbonate; electrolyte secretion in a variety of tissues/organs; bone resorption; calcification; renal and male reproductive tract acidification, signal transduction and formation of gastric acid [8][9][10][11][12].
The significant progress made in the DNA sequencing approach allowed the identification of genes encoding for CAs in pathogenic and non-pathogenic microorganisms [21]. Intriguing, the understanding of the function of the bacterial CAs has increased significantly [6,22], confirming that the activity of CAs is connected to the survival as well as the virulence of pathogens because the metabolic reaction catalyzed by these enzymes is essential for supporting numerous physiological functions involving dissolved inorganic carbon [6,22]. Based on these considerations and in the time of emerging antibiotic resistance, targeting bacterial CAs towards a new generation of antibacterial drugs might represent a valid alternative to reinforce the pharmacological arsenal against these pathogens. Thus, scientists heterologously produced in vitro the CA-classes encoded by bacteria as well as other pathogens. Many data are available on kinetic parameters and inhibition profiles of the α-, βand γ-CAs encoded by pathogenic and non-pathogenic bacteria, while very few data are available on the latest discovered bacterial CA-class, the ι-CA. This CA class was identified for the first time in the marine diatom Thalassiosira pseudonana [23] and, surprisingly, showed to prefer Mn 2+ to Zn 2+ as metal ion cofactor [23]. Generally, bacterial αand β-CAs use the Zn 2+ ion as catalytic metal, while γ-CAs are Fe 2+ -dependent enzymes but they are functional enzymes also with bound Zn 2+ or Co 2+ ions [24][25][26][27].
In a previous work, we have cloned, expressed and purified the first recombinant bacterial ι-CA (acronym BteCAι) identified in the genome of Burkholderia territorii, a Gram-negative bacterium found in soil and water, which is often resistant to common antibiotics [28,29]. The recombinant BteCAι was shown to be a suitable catalyst for the hydration of CO 2 to bicarbonate and proton, with a k cat of 3.0 × 10 5 s −1 and k cat /K M of 3.9 × 10 7 M −1 s −1 and was also sensitive to inhibition by the sulfonamide acetazolamide [29]. Here, we carried out an extensive study on the inhibition profiles of BteCAι using the substituted benzene-sulfonamides and clinically licensed drugs, which, among the groups of the classical CAIs, generally inhibit other CAs in the nanomolar range and have been clinically used for decades as antiglaucoma [30], diuretic [15], antiepileptic [11], antiobesity [8,31] and anticancer agents [10]. Besides, a comparative analysis of the sulfonamide inhibition profiles was performed comparing BteCAι results with those obtained from CAs belonging to different classes, such as the two human α-CA isoforms (hCA I and hCA II) and the β-and γ-CAs from Escherichia coli. This study gives useful information for designing new antibacterials to disarm the pathogen or bypass their resistance to con-ventional antimicrobials by inhibiting enzymes involved in the CO 2 /HCO 3 − balancing pathway.

Production, Validation and Qualitative Activity Assessment of BteCAι
The ι-CA encoded by the genome of B. territorii was heterologously expressed in bacteria with the aim to produce a sufficient amount of protein for the determination of the corresponding inhibition profile with substituted benzene-sulfonamides and clinically licensed drugs. Sodium dodecyl-sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and protonography were used for the evaluation of the enzyme homogeneity and purity, as well as to verify the enzyme activity of BteCAι. As shown in Figure 1, SDS-PAGE analysis showed a protein migration corresponding to a molecular mass of 19.0 kDa. The BteCAι band was also active when subjected to protonography analysis, as demonstrated by the corresponding yellow color developed by the production of ions (H + ) during the CO 2 hydration reaction ( Figure 1). Moreover, the protein concentration of the recombinant BteCAι was obtained from the SDS-PAGE, analyzing the gel by Image J and using a known concentration of the commercial bovine CA (bCA). Protein quantification by densitometry revealed that the E. coli cells overexpressed BteCAι with a yield of about 0.8 mg/mL from a bacterial culture of 1 L. The protein concentration calculated by densitometry resulted about 10% less with respect to that ascertained by the Bradford method. The BteCAι activity of the purified protein was also measured in solution and expressed in Wilbur-Anderson Units (WAU). BteCAι solution showed a specific activity of 80 ± 8.0 WAU mg −1 .

Comparison of the BteCAι Kinetic Parameters with Those of Other Bacterial CA Classes
The CO 2 hydratase activity and the kinetic constants of the purified BteCAι were determined using the stopped-flow technique. The enzyme had a high catalytic activity (k cat 3.0 × 10 5 s −1 ) for the physiological reaction of CO 2 hydration to bicarbonate and protons. The enzyme was also well inhibited by the sulfonamide acetazolamide (K I = 64.9 nM), a classical CA inhibitor. Table 1 shows a comparison of the BteCAι kinetic behavior (k cat and k cat /K M ) with those of other CAs belonging to different classes, such as two α-CA isoforms from Homo sapiens (hCAI and hCAII) and the βand γ-CAs from E. coli. In Table 1, it is readily apparent that considered enzymes showed a catalytic constant (k cat , which indicates the maximum rate of the reaction at saturating substrate concentration) in the same order of magnitude, except for the human isoform CA II. Intriguing, the K M value of BteCAι, the substrate concentration at which the reaction rate is half of V max , is one order of magnitude lower than that shown by the other two bacterial enzymes belonging to a different class (EcoCAβ and EcoCAγ). K M is a measure of the enzyme's affinity for its substrate and a lower K M indicates that the enzyme accomplishes its function at a lower substrate concentration. Considering this aspect, the affinity for the substrate (CO 2 ) of BteCAι is higher than that of the two human isoforms (hCA I and hCA II) as well as of the bacterial enzymes (β and γ). In particular, the affinity of BteCAι is 1.3-and 3.0-fold higher than that of hCA I and hCA II, respectively, whereas it is 4.2 to 26-fold higher compared with that of the other two bacterial enzymes ( Figure 2).

Figure 2.
Graphical representation of the enzyme activity. The enzyme affinity for the CO 2 was reported as 1/K M to evidence that more remarkable is the bar height, more potent is the affinity of the enzyme for the substrate.
In a previous work, we carried out a phylogenetic analysis to better evidence the amino acid sequence relationship of bacterial ι-CAs with other microorganism CA-classes (α, β and γ-CAs), demonstrating that the ι-class is closer to the γ-CAs [29]. In the literature, it has been reported that ι-CA from the marine diatom Thalassiosira pseudonana shows a promiscuous esterase activity, which is a feature of the CAs belonging to the α-class [23]. Generally, the enzyme catalytic pocket is small for the γ-CAs and gets more ample for the β-CAs, being quite large in the α-CAs [6]. Based on these observations, it is intriguing to hypothesize that the ι-CAs have an active site spatial organization similar to that of the α-CA, for example, hCA I. However, the BteCAι spatial organization of the catalytic pocket probably allows a better entry of CO 2 in the catalytic pocket as well as its tightly binding and thus a higher affinity for this molecule, as demonstrated by the low K M value of BteCAι with respect to that of the human and the other two bacterial enzymes (β and γ-CAs) ( Table 1 and Figure 2). Interestingly, although the ι-CA might have a catalytic pocket similar to the human CAs, the classical CA inhibitor, acetazolamide, inhibited BteCAι with a K I of 519 nM, which is a value higher than those determined for other CAs considered in this study. All these findings suggest that even though these enzymes catalyze the same reaction, there are differences in the catalytic pocket, explaining the distinctive K M and the diverse degree of inhibition shown by the various CA classes. Therefore, we are trying to crystallize the BteCAι with the aim to compare its three-dimensional structure with that reported in the literature for other known CAs. This analysis will allow the identification of the ι-CA structural elements responsible for such differences with respect to other CA classes.

Effects of Simple Aromatic/Heterocyclic Sulfonamide Inhibitors on BteCAι Activity
The first sulfonamide showing a significant antibacterial activity was Prontosil, a sulfanilamide prodrug, which is isosteric/isostructural with p-aminobenzoic acid (PABA), the substrate of dihydropteroate synthase (DHPS) [32,33]. After sulfanilamide was demonstrated to be an effective antibacterial agent, a range of molecular analogs constituting the so-called sulfa drugs entered in clinical use [30]. The presence of primary sulfonamide moieties in sulfanilamide characterizes most investigated CA inhibitors (CAIs) [8,[34][35][36][37]. Figure 3 shows some of these sulfonamide inhibitors (simple derivatives 1-24 and clinically used drugs) [3,[38][39][40][41][42][43][44][45][46][47][48][49][50][51][52]. Acetazolamide (AAZ), methazolamide (MZA), ethoxzolamide (EZA) and dichlorphenamide (DCP) are systemically acting antiglaucoma CAIs. Dorzolamide (DZA) and brinzolamide (BRZ) are antiglaucoma agents that function topically; benzolamide (BZA) is an orphan drug of this pharmacological class. Zonisamide (ZNS), sulthiame (SLT) and topiramate (TPM) are widely used antiepileptic drugs. Sulpiride (SLP) and indisulam (IND) also belong to this class of pharmacological agents, together with the COX-2 selective inhibitors celecoxib (CLX) and valdecoxib (VLX). Saccharin (SAC) and the diuretic hydrochlorothiazide (HCT) are also known to act as CAIs [11,53,54]. Famotidine (FAM) is a competitive histamine H 2 -receptor antagonist [53] and epacadostat (EPA) is an inhibitor of the heme-containing enzyme, indoleamine 2,3-dioxygenase-1 (IDO1) but they also act as CAIs [54]. The aromatic/heterocyclic part of the inhibitor interacts with the hydrophilic and hydrophobic residues of the catalytic cavity. Its −SO 2 NH 2 group binds in a tetrahedral geometry to the Zn 2+ ion in the deprotonated state, with the nitrogen atom of the sulfonamide moiety coordinated to Zn 2+ and an extended network of hydrogen bonds, involving amino acids of the enzyme, also participating in the anchoring of the inhibitor molecule to the metal ion, as shown by X-ray crystallographic data of enzyme-inhibitor adducts [11,53,54]. Numerous pieces of evidence support the involvement of CA activity in the survival, pathogenicity and virulence of several species of human pathogens, such as Helicobacter pylori [55][56][57], Vibrio cholerae [58], Brucella suis [52,[59][60][61], Salmonella enterica [62] and Pseudomonas aeruginosa [63]. For example, ethoxzolamide (EZA) was demonstrated to inhibit the V. cholerae virulence in vivo by blocking the cholera toxin gene expression, induced by the bicarbonate produced by the Vibrio CA activity [58]; this inhibitor can also prejudice the virulence of M. tuberculosis [64]. Recently, it has been demonstrated that AAZ and some AAZ-based sulfonamides act as potent inhibitors of vancomycin-resistant Enterococcus [65], which are the leading causes of drug-resistant healthcare-associated infections. All these findings explain the proof of concept that bacterial CAs are promising targets for developing new drugs. Since CAs are valuable targets for compromising the microbial vitality or virulence, the BteCAι inhibition profile with these compounds can be considered an initial information for recognizing efficient and selective inhibitors of bacterial members of the ι-CA class in pathogens, with respect to other protein isoforms present in the host. This information is highly desired for obtaining new pharmacological agents, which may impair the bacterial growth/virulence. Table 2 compares the inhibition profiles of three classes of bacterial CAs and those obtained for the two human isoforms (hCA I and hCA II). From the data of Table 2, the following results can be observed: 1.
Among the sulfonamides and sulfamate used to determine the BteCAι inhibition profile, only three inhibitors resulted in a K I lower than 100 nM. This is the case of the compounds 5, 9 and 19. These results confirm how different is the spatial organization of the catalytic pocket of the different CA classes. For the same compounds, the other two bacterial enzymes showed K I values in the range 160-2840 nM, while the corresponding K I values of two human isoforms were between 33-25,000 nM. The two human isoforms, hCA I and hCA II, resulted in a variegate behavior since hCA I was very susceptible to the inhibitors 9 and 19 with a K I values of 40 and 33 nM, respectively. Again, both human isoforms showed many nanomolar inhibitors with a K I below 100 nM, such as compounds 20, 21, 24, MZA, EZA, BZA, ZNS, IND. Once more, there were compounds such as 3, 7, 10, 11, 12, 13, 14, 15, 16, DCP, DZA, BRZ, SLP, VLX and CLX, which were potent inhibitors of hCA II with K I values in the range 3-94 nM but were very week inhibitors of hCA I (K IS > 5500 nM). Only the study of the three-dimensional structures of BteCAι (not available at this moment) will explain the structural factors responsible for the K I variations.

2.
Most of the inhibitors considered in Table 2  The results reported in Table 2 showed substantial differences in the inhibition constants between the various CA-classes. The inhibition pattern differences must be considered useful for the development of new specific drugs since it means that the spatiality and the amino acids surrounding the catalytic pocket of the CA enzymes are different. This diversity allows the possibility to design efficient and selective inhibitors of the bacterial enzymes without interfering with the activity of the human CAs, even if they show a high percentage of amino acid sequence identity.

Chemicals and Instruments
All the chemicals used in this study were of reagent grade and purchased from Sigma. The Affinity column (His-Trap FF) and the AKTA-Prime purification system were bought from GE Healthcare. SDS-PAGE apparatus was procured from Bio-Rad (Hercules, CA, USA).

Heterologous Expression and Purification of the Recombinant Enzyme
The synthetic B. territorii gene encoding for the BteCAι was cloned, expressed and purified, as described by Del Prete et al. [29]. Briefly, the synthetic gene contained in the expression vector pET100D-Topo/BteCAι was heterologously overexpressed, transforming the competent E. coli BL21 (DE3) codon plus cells (Agilent), using as Isopropyl β-D-1thiogalactopyranoside (IPTG) as inducer. The produced cytoplasmic protein was purified by using a resin functionalized with Ni 2+ , which has a high affinity for the polyhistidinetag (His-Tag) added by genetic engineering to the amino terminus of the neo-synthetized recombinant protein. The protein concentration was determined using the Bradford method (Bio-Rad) [66] and by densitometry using the Gel Analyzer tool of ImageJ [67].

Enzyme Activity, SDS-PAGE and Protonography
Throughout the purification, the proteolytic activity of BteCAι was achieved as described by Capasso et al. [68]. Twelve % SDS-PAGE performed as described by Laemmli [69] and protonography [70][71][72][73] carried out as reported by Del Prete et al. [29] were used to mon-itor the apparent molecular mass of the purified recombinant protein on the polyacrylamide gel and the corresponding hydratase activity, respectively.

Kinetic Parameters and Inhibition Constants
The CO 2 hydration activity exerted by BteCAι was monitored using an Applied Photophysics stopped-flow instrument [74]. Phenol red (at a concentration of 0.2 mM) was used as an indicator, working at the absorbance maximum of 557 nm, with 20 mM TRIS (pH 8.3) as a buffer and 20 mM NaClO 4 (for maintaining constant the ionic strength), following the initial rates of the CA-catalyzed CO 2 hydration reaction for a period of 10-100 s. To determine the kinetic parameters by Lineweaver-Burk plots and the inhibition constants, a concentration of CO 2 between 1.7 to 17 mM was used. For each inhibitor, at least three measurements were used to assess the initial velocity at all inhibitor concentrations tested. The uncatalyzed rates were identically determined and detracted from the total observed rates. Stock inhibitor solutions (10 mM) were prepared in distilled-deionized water and dilutions up to 0.1 nM were done with the buffer test. Inhibitor and enzyme solutions were preincubated together for 15 min, at room temperature, before the assay to allow for the formation of the E-I complex or the eventual active site mediated hydrolysis of the inhibitor. The inhibition constants were obtained by non-linear least-squares method using PRISM 6 and the Cheng-Prusoff equation, as reported earlier [75][76][77] and represent the mean from at least three different determinations. hCA I, hCA II and the bacterial enzymes were recombinant proteins obtained in-house. Their concentrations in the assay system were of 5-14 nM.

Conclusions
In this context, a broad range of substituted benzene-sulfonamides and clinically licensed drugs were used to determine the inhibition profile of BteCAι and the possible off-targets hCA I and hCA II. Among the sulfonamides and the one sulfamate used as inhibitors, only three of them resulted in having a K I value lower than 100 nM (compounds 5, 9 and 19). All the other inhibitors had K Is > 100 nM. Surprisingly, the results reported showed substantial differences in the inhibition constants between the various CA-classes considered in this study (α, β, γ and ι). For example, for some compounds EcoCAγ showed K Is > 2000 nM, evidencing that this enzyme form or others E. coli froms were less or not inhibited by some of the substituted benzene-sulfonamides and clinically licensed drugs. These differences in the sulfonamide inhibition pattern represent an aspect of the CA inhibition useful for the development of specific and selective drugs versus the bacterial enzymes. It means that the spatiality and the amino acids surrounding the catalytic pocket of the CA enzymes are different. This diversity will allow the possibility to design inhibitors of the bacterial enzymes, which are efficient and selective versus the bacterial enzymes without interfering with the activity of human CAs, even if they show a significant percentage of amino acid sequence identity.

Data Availability Statement:
The data presented in this study are openly available in https:// pubmed.ncbi.nlm.nih.gov.