Activation of carbonic anhydrases from human brain by amino alcohol oxime ethers: towards human carbonic anhydrase VII selective activators

Abstract The synthesis and carbonic anhydrase (CA; EC 4.2.1.1) activating effects of a series of oxime ether-based amino alcohols towards four human (h) CA isoforms expressed in human brain, hCA I, II, IV and VII, are described. Most investigated amino alcohol derivatives induced a consistent activation of the tested CAs, with KAs spanning from a low micromolar to a medium nanomolar range. Specifically, hCA II and VII, putative main CA targets when central nervous system (CNS) diseases are concerned, were most efficiently activated by these oxime ether derivatives. Furthermore, a multitude of selective hCA VII activators were identified. As hCA VII is one of the key isoforms involved in brain metabolism and other brain functions, the identified potent and selective hCA VII activators may be considered of interest for investigations of various therapeutic applications or as lead compounds in search of even more potent and selective CA activators.


Introduction
Activators of the metalloenzymes carbonic anhydrases (CAs; EC 4.2.1.1, CAAs) have been lately going through a second youth in drug discovery processes 1 . Early evidence of the CA activation efficacy of amines, such as histamine, dating back to the 1940s was thereafter long debated up, and even considered as an experimental artefact 2,3 . In the early 1990s, the combination of highly purified enzymes and precise techniques, such as the stopped flow assay, put an end to the long controversy, testifying the undeniable existence of CAAs [4][5][6][7] . CAAs intervene in the second and rate determining step of the catalytic mechanism of carbon dioxide reversible hydration (Equations 1 and 2), that is the regeneration of the catalytically active, zinc hydroxide species, by a proton transfer reaction from the Zn 2þbound water molecule to the external medium (Equation 2) [6][7][8] . This process is assisted by active site residues acting as proton shuttle, such as His residues placed in the middle or at the entrance of the active site cavity of the a-class human CAs 4 . EZn 2þ ÀOH À þ CO 2 À EZn 2þ ÀHCO À 3 þ H 2 O À EZn 2þ ÀOH 2 þ HCO À

3
(1) EZn 2þ ÀOH 2 À EZn 2þ ÀOH À þ H þ (2) EZn 2þ À OH 2 þ A À EZn 2þ À OH 2 À A Â Ã À EZn 2þ À OH À À AH þ ½ À EZn 2þ À OH À þ AH þ The activators non-competitively activate the CAs through the formation of a ternary complex consisting of the enzyme, the substrate, and the activator. CAAs bind apart from the metalcoordination system, namely at the middle edge of CAs active site cavity, where they assist the proton shuttling [9][10][11] . As a proof of this phenomenon, it was demonstrated that efficient activators possess pK a values in the range of 6-8, values similar to the pK a of the His imidazole moiety 7 .
It is seemingly odd that CAs, among the most efficient enzymes in Nature, may be activated for biomedical purposes 8 . However, genetic deficiencies of several CA isoforms (e.g. human CA I, II, IV, VA, XII and XIV) have been reported in the last decades, associated to diseases such as osteopetrosis, cerebral calcifications, retinal problems, hyperammonaemia, hyperchlorhydrosis [12][13][14][15] , and the loss of function of these enzymes would in principle be treatable with CA selective activators 8 . In addition, there is evidence that CAs activation improves memory deficits, cognitive performance and learning [16][17][18] , being nine of the fifteen known human CA isoforms present in brain 19,20 . Contrariwise, other evidence supported that CA inhibitors (CAIs) impair memory in human, according to studies on the CAIs topiramate and acetazolamide during acute high-altitude exposure 21,22 . Thus, CAs represent a crucial family of new targets for improving cognition, but also in therapeutic areas, such as phobias, obsessive-compulsive disorder, generalised anxiety, and post-traumatic stress disorders, for which few effective therapies are available. In fact, in a recent paper, one of our groups showed that the CAA D-phenylalanine and the CAI acetazolamide are respectively able to reinforce and impair extinction memory, that is, a new memory trace that inhibits the expression of the memory of a traumatic event 23 . In this golden period for CAAs, the discovery of new brain isoform selective CAAs (as well as CAIs) is invaluable to elucidate the role of CA isoforms in brain processes. In addition, CAAs are considered relevant in artificial tissues and in CO 2 capture and sequestration processes 1,24 .
Here, we extend the knowledge of CAA chemotypes by describing the synthesis and CA activating effects of a series of oxime ether based amino alcohols towards four hCA isoforms expressed in human brain.

Material and methods
2.1. Chemistry 1 H and 13 C NMR spectra were recorded on a Bruker Avance III HD 400 MHz spectrometer. Chemical shifts (d) are reported in parts per million and coupling constants (J) are reported in hertz (Hz). 13 C NMR spectra were fully decoupled. The following abbreviations were used to explain multiplicities: singlet (s), doublet (d), triplet (t), double doublet (dd), broad (br), and multiplet (m). Chromatographic separations were performed on silica gel columns by flash column chromatography (Kieselgel 40, 0.040À0.063 mm, Merck). Reactions were followed by thin-layer chromatography (TLC) on Merck aluminium silica gel (60 F254) sheets that were visualised under a UV lamp. Evaporation was performed in vacuo (rotating evaporator). Sodium sulphate was always used as the drying agent. Commercially available chemicals were purchased from Sigma-Aldrich.

General procedure for the synthesis of aminoalcohols 14-16
Epichlorohydrin (0.88 ml,11.2 mmol) was added dropwise to a stirred solution of N-hydroxy-5-norbornene-2,3-dicarboximide (2 g, 11.2 mmol) and Et 3 N (3.12 ml, 22.4 mmol) in anhydrous DMF (8 ml). After stirring for 18 h at rt the reaction mixture was poured into water (50 ml) and extracted with CHCl 3 (2Â50mL). The organic phases were combined, washed with water (2 Â 100 ml), dried (Na 2 SO 4 ) filtered and evaporated under reduced pressure. The crude epoxide 39 was purified by crystallisation with n-hexane. NMR data were in accordance with those reported in literature 25 .
A stirred solution of epoxide 39 (611 mg, 2.6 mmol) in dry EtOH (8 ml) was treated with an excess of iPrNH 2 (0.43 ml, 5 mmol) or t-BuNH 2 (0.52 ml, 5 mmol). The reaction mixture was stirred at 50 C for 4 h, and then the solvent was evaporated. The crude was constituted principally by the amino alcohol 40 (iPr) or 41(tBu) and it was used in the next reaction step without further purification.
A solution of the proper amino alcohol 40 or 41 (1.56 mmol) in 7.8 ml of NH 3 -MeOH 7 N was stirred rt for 2 h. The resulting mixture was filtered and evaporated. The crude was purified by flash chromatography with EtOAc/MeOH/Et 3 N (8: 1.5: 0.5) affording derivative 42 or 43 (60-63% yield). NMR data were in accordance with those reported in literature 25 .
A solution of p-, m-, or o-hydroxybenzaldehyde (6.75 mmol) in EtOH (11.4 ml) and the proper oxyamine 42 or 43 (6.75 mmol) was refluxed for 12 h. After cooling, the mixture was evaporated. The crude was dissolved in a mixture of MeOH/EtOH 3:7 (20 ml) and treated with 1.2 equivalent of the proper organic acid (oxalic, malic or fumaric acids) affording the amino alcohols (

Carbonic anhydrase activation
A stopped-flow method 34 has been used for assaying the CA catalysed CO 2 hydration activity with Phenol red as indicator, working at the absorbance maximum of 557 nm, following the initial rates of the CA-catalysed CO 2 hydration reaction for 10-100 s. For each activator, at least six traces of the initial 5-10% of the reaction have been used for determining the initial velocity. The uncatalyzed rates were determined in the same manner and subtracted from the total observed rates. Stock solutions of activator (0.1 mM) were prepared in distilled-deionised water and dilutions up to 0.1 nM were done thereafter with the assay buffer. The activation constant (K A ), defined similarly with the inhibition constant (K I ), was obtained by considering the classical Michaelis-Menten equation (Equation 4), which has been fitted by nonlinear least squares by using PRISM 3: where [A] f is the free concentration of activator. Working at substrate concentrations considerably lower than K M ([S](K M ), and considering that [A] f can be represented in the form of the total concentration of the enzyme ([E] t ) and activator ([A] t ), the obtained competitive steady-state equation for determining the activation constant is given by Equation (5): where v 0 represents the initial velocity of the enzyme-catalysed reaction in the absence of an activator [35][36][37] . Enzyme concentrations in the assay system were in the range of 9-12 nM.

Chemistry
As mimic of the main proton shuttling residue (i.e. histidine), histamine (Figure 1) is the main lead for designing CAAs 8 . The binding mode of histamine with hCA II was elucidated by X-ray crystallography (Figure 2) 6 . In the CAA/enzyme adduct, a complex network of H-bonds involve the Zn-bound water molecule, His64 and the imidazole ring of the activator, located far away from the metal ion ( Figure 2). Successive X-ray crystallographic studies showed that many other amines and amino acids (Figure 1) bind in this area and share flexible tails decorated with protonable moieties ( Figure 2) [38][39][40] . No isoform-selective CAAs were detected so far among these amines and amino acid derivatives, except for few exceptions (e.g. histamine shows a 10 nM K A s against hCAs VA and XIV but is a micromolar activator of the remaining isoforms) 1 . The molecule of histamine has been extensively modified (Figure 3), including substituents on the imidazole C atoms (A) 41,42 , replacing the imidazole ring with other heterocycles, such as 2,4,6-trisubstituted pyridinium (B), 1,3,4-thiadiazole (C) or a combination of these two ring systems (D) 43, 44 and functionalising the NH 2 group, as in carboxamides/ureas/thioureas (E) 45 , sulphonamides (F) 46 , arylsulfonylureas (G) 47 , bis-histamine (H) 48,49 , oligopeptides 49,50 , or imidazole/imidazoline derivatives of the alkaloyd spinaceamine 51,52 and drug clonidine 53 (Figure 3) were reported to possess improved CA activatory profile when potency and isoform selectivity are concerned. Interestingly, pharmacological agents born for other medical applications ( Figure 4) showed also to possess more or less relevant CA-activating effects, among which psychoactive compounds of the amphetamine and methamphetamine family 54 , the selective serotonin reuptake inhibitors fluoxetine, sertraline and citalopram 55 , the phosphodiesterase IV inhibitor sildenafil 56 , and the b-blocker amino alcohol derivative timolol 57 . Timolol was taken as lead compound in the present study to design a new series of uncommon CAAs, which do not bear imidazole like scaffolds, intesively explored in the field. The enzyme kinetic method showed that timolol noncompetitively activates hCA I and II by forming of a ternary complex consisting of the enzyme, the substrate, and timolol. Docking studies were used to point out that timolol also binds at the entrance of the active site cavity nearby the proton shuttle residue His64 57 . Thus, a series of aminoalcohol oxime ether derivatives previously shown to possess b-blocking or analgesic/ antiarrhythmic activity (compounds 1-16) 25-32 , was investigated for the activation of a panel of brain hCAs.
Compounds 1-6 and 9-16 were prepared as previously described (Schemes 1 and 2) 25-32 . Here, an updated synthetic procedure is reported. Moreover, new 1 H NMR and 13 C NMR spectra of final compounds are provided in the experimental section. A fully detailed description is provided for the synthesis of new compounds 7 and 8, following the same synthetic procedure.
The condensation between derivative 42 or 43 and o-, m-or phydroxybenzaldehyde afforded the desired amino alcohols 14-16, purified by crystallisation as organic salts.
All the final compounds have been achieved as organic salt oximes presenting E configuration. The configuration of the C¼N bound for compounds 1, 2, 4-8, 12 and 13 was assigned by analogy considering the E configuration of starting oximes 17, 18, 20-22, 25-27. As previously reported, the configuration of Eoximes 17, 18, 20-22, 25-27 was stable under the following reaction conditions. The E configuration of amino alcohols 14-16 was conferred by comparison between the 1 H signals of the iminic protons of compounds 14-16 with the same signals of final compounds 1, 2, 4-8, 12 and 13. The chemical shift of iminic proton signal resulted around d 8.05-8.50 ppm for all derivatives as usually reported for oximes with E configuration.

Carbonic anhydrase activation
Amino alcohols 1-16 were here assayed for their activating properties of 4 catalytically active and physiologically relevant hCA isoforms expressed in human brain, that are: the cytosolic hCA I, II, and VII, and the membrane associated hCA IV 58 . In the CNS context, hCA I is expressed in the motor neurons in human spinal cord 59 . The physiologically dominant isoform hCA II is located both in the choroid plexus, and in oligodendrocytes, myelinated tracts, astrocytes and myelin sheaths in the vertebrates brain 60 . hCA IV is located on the luminal surface of cerebral capillaries, associated with the blood-brain barrier, and expressed in layers III and VI in the cortex, thalamus and hippocampus 60,61 . CA VII is expressed in the cortex, hippocampus and thalamus 62,63 . CA VII might be considered a brain-associated CA as it is predominantly expressed in the brain, and absent in most other tissues. CA VII is also considered a key molecule in age-dependent neuronal pH regulation.
The CA activation data of these 4 isoforms with amino alcohols 1-16 and histamine as standard CAA are shown in Table 1. The following structure-activity relationship (SAR) can be worked out: i. A special mention should be done for phenol derivatives 14-16 which, uniquely among amino alcohols 1-16, did not produce any activation of the tested hCAs. In fact, it is worth stressing that CAI properties, rather than as CAAs, are commonly ascribed to the phenolic chemotype. Phenols have been thoroughly reported to anchor to the zinc-bound nucleophile (water molecule or hydroxide ion), that is, by one of the four CA inhibition mechanisms known to date. Surprisingly, neither a significant inhibition was detected by treating CA I, II, IV and VII with phenols 14-16. One could speculate that the CAI efficacy of 14-16 is counterpoised by their CAA action, hindering its detection by the Stopped Flow assay. ii. The cytosolic isoform hCA I was moderately activated by all remaining amino alcohols, that are 1-13, with K A values in the range of 0.92-12.1 mM. The tert-butylamino derivatives 1-8 induce a slightly greater activation of CA I than isopropylamines 9-13. In the 1-8 subset, it can be noted that the o-or p-substitution of the aromatic ring bearing the oxime ether group lower the K A values 4, 6, 8 up to fivefold with respect to 2 (K A s of 1.37, 1.26 and 0.92 vs 4.25 mM). In contrast, m-substitutions are deleterious for the binding to the    9-13 activate hCA II in a low micromolar range (K A s in the range 1.94-5.90 mM). All CAAs reported here were more active towards hCA II than histamine, which is a quite weak activator of this isoform with a K A of 125 mM. iv. No submicromolar K A values were measured for amino alcohols 1-13 as hCA IV activators. Indeed, all K A s are in a rather flat low micromolar range (K A s in the range 1.01-12.9 mM), making such a membrane associated isozyme the less activated by the assayed derivatives. Notably, the reference CAA histamine even less activates hCA IV with a K A of 25.3 mM. No significant differences exist between tert-butyl and isopropyl derivatives as for the activation of hCA IV. v. The other cytosolic isoform investigated here, hCA VII, was rather potently activated, that is chiefly in a submicromolar range, by most of the compounds reported in this work (K A s in the range 0.082-8.86 mM). All derivatives showed much better activation profile than the reference histamine (K A of 37.6 mM) towards hCA VII. Contrariwise to the other hCAs tested, the most efficient CAAs are the isopropylamino derivatives 9 and 10, ethers of secondary oxime with an aliphatic pendant showing medium nanomolar K A s values of 82 and 91 nM, respectively. Only the oxime ethers of p-anisaldehyde and acetone in the isopropylamino and tert-butylamino series, respectively, that are compounds 4 and 11, show K A s steadily in a low micromolar range. As this isoform is one of the most widely spread in the brain, and probably involved in crucial metabolic/pH regulation processes, these results are of interest in the search of more effective CA VII activators than the currently available such derivatives.

Conclusions
In the present study, we described the synthesis and CA activating effects of a series of oxime ether-based amino alcohols towards four hCA isoforms expressed in human brain, that are CA I, II, IV and VII. Except for the phenolic compounds 14-16, all amino alcohol derivatives from this study induce a consistent activation of the tested CAs, with K A s spanning from a low micromolar to a medium nanomolar range. Specifically, hCA II and VII, probable main CA targets when CNS diseases are concerned, 58 are most potently activated by these oxime ether derivatives. With K A s of 79 and 420 nM towards hCA II and VII, respectively and K A s for CA I and IV settling in a range 10-to 100-fold higher, derivative 5 from the tert-butylamines series turned out as the most potent and selective hCA II activator of the study. On the other hand, a multitude of selective hCA VII activators were identified. Specifically, a mention should be done for derivatives 9 and 10, from the isopropylamines series, showing a 10-to 100-fold selective hCA VII activation profiles with respect to all other assayed CAs. In this second youth period for CAAs, innovative pharmacological studies made these lately neglected agents to draw attention in the memory therapy and cognitive neurodegenerative disorders, as well as in therapeutic areas, such as phobias, obsessive-compulsive disorder, generalised anxiety, and post-traumatic stress disorders 1,20 . As hCA VII is a key isoform involved in brain metabolism, the here identified potent and selective hCA VII activators may be considered of interest for investigations of possible therapeutic applications or as lead compounds in search of more potent and selective CAAs. This work might bring new lights on the intricate relationship between CA activation and brain physiology.

Disclosure statement
No potential conflict of interest was reported by the author(s).