Nitrous anhydrase activity of carbonic anhydrase II: cysteine is required for nitric oxide (NO) dependent phosphorylation of VASP in human platelets

Abstract The carbonic anhydrase (CA) family does not only catalyse the reversible hydration of CO2 to bicarbonate, but it also possesses esterase and phosphatase activity. Recently, bovine CA II and human CA II have been reported to convert inorganic nitrite (O=N-O−) to nitric oxide (NO) and nitrous anhydride (N2O3). Given the ability of NO to mediate vasodilation and inhibit platelet aggregation, this CA II activity would represent a bioactivation of nitrite. There are contradictory reports in the literature and the physiological role of CA II nitrite bioactivation is still disputed. Here, we provide new experimental data in support of the nitrous anhydrase activity of CA II and the key role L-cysteine in the bioactivation of nitrite by CA II. Using washed human platelets and by measuring VASP phosphorylation we provide evidence that exogenous nitrite (10 µM) is bioactivated to NO in a manner strongly depending on L-cysteine (100 and 200 µM). The process is not inhibitable by acetazolamide, a potent CA inhibitor. The contradictory results of recently published studies in this area are thoroughly discussed.


Introduction
The "inherent" catalytic activity of the carbonic anhydrase (CA, EC 4.2.1.1) family is the reversible hydration of CO 2 to bicarbonate (HCO 3 À ) (R1) which is inhibited by several classes of drugs, among which sulphonamides (RSO 2 NH 2 ) such as acetazolamide. In addition, CA isoforms are known for long time to possess esterase and phosphatase activity and to exert these activities by the same mechanism. Recently, bovine and human CA II have been reported to convert inorganic nitrite (O¼N-O À ) to nitric oxide (NO) 1 and to nitrous anhydride (N 2 O 3 ) 2,3 . While authentic NO can be directly detected, there is no analytical method to detect native N 2 O 3 in biological samples thus far. N 2 O 3 may exist in four very labile forms: one form is ON-NO 2 (ON ᭹ þ N ᭹ O 2 $ ON-NO 2 ) without the nitrous anhydride structure N-O-N; the other three forms are isomeric O¼N-O-N¼O 4 . The results of our previous work 2,3 suggest that N 2 O 3 has the structure O¼N-O-N¼O. This structure is used in the present work throughout. The conversion of O¼N-O À to NO is a redox reaction and implies the provision of one electron by the CA II. Yet, such a mechanism is difficult to reconcile for redox-inactive Zn 2þ -containing CA II. A nitrous anhydrase activity of CA II seems more realistic, because such a reaction requires protons which can be easily provided by the enzyme (R2). Given the ability of NO to mediate vasodilation and inhibit platelet aggregation, and the potential of N 2 O 3 to nitrosate different functionalities of biomolecules to form for instance S-nitrosothiols (RSNO; (R3)), N-nitrosoamines (RNHNO; (R4)), the nitrous anhydrase CA II activity would represent a bioactivation of nitrite.
Nitrite and nitrate are metabolites of NO endogenously produced from L-arginine by the catalytic action of NO synthase (NOS). Exogenous nitrite and nitrate are also present in foods and drinking water and represent a considerable source of NO bioactivity via bacterial nitrate reductase activity in mouth and gut flora. Pharmacological nitrite, an NOS-independent source of NO, has protective effects in the cardiovascular system by decreasing blood pressure and inhibiting platelet activation 5 . Administration of nitrite has been reported to lower blood pressure and to inhibit platelet aggregation in vivo in humans and mice 6,7 . In contrast, nitrite added to washed human or mice platelets in vitro was found not to activate soluble guanylyl cyclase (sGC) and not to inhibit platelet activation 7,8 . We supposed that the different effects of nitrite in vivo and in washed platelets in vitro might be connected with the absence of extra-cellular L-cysteine in washed platelet preparations. In our experiments with washed human platelets we tested this hypothesis and used L-cysteine at physiological concentrations.
The NO/sGC/protein kinase G (PKG) system is one of the most powerful mediators of platelet inhibition. Activation of PKG results in phosphorylation of multiple substrates involved in platelet inhibitory mechanisms. Vasodilator-stimulated phosphoprotein (VASP) is one of the well-established PKG substrates and is often used as a marker of the activation of this kinase 9 . In our experiments with washed human platelets, PKG activation was monitored by measuring VASP phosphorylation at Ser239 (P-VASP Ser239 ) using phospho-specific antibodies. It is worth mentioning that human platelets contain CA I and CA II, but they do not contain NOS 10 .
There are contradictory reports in the literature regarding the above mentioned newest CA II activities and even a disputation of a physiological role of CA II in the bioactivation of nitrite [11][12][13][14] .
Here, we provide new experimental data in support of the nitrous anhydrase activity of CA II.

Materials and methods
Experiments with isolated CA II NaNO 2 was obtained from Riedel-de-Ha€ en (Seelze, Germany). 2,3,4,5,6-Pentafluorobenzyl bromide (PFB-Br), L-cysteine, bovine CA II and 15 N-labelled nitrite (Na 15 NO 2 ; declared >98 atom % 15 N) were from Sigma-Aldrich (Steinheim, Germany). H 2 18 O (declared isotopic purity >98 atom % 18 O) was purchased from Medical Isotopes (Pelham, New Hampshire, USA). The GC column Optima-17 (15 m Â 0.25 mm i.d., 0.25-mm film thickness) and the conical glass vials were obtained from Macherey-Nagel (D€ uren, Germany). GC-MS analyses were performed on a ThermoElectron DSQ mass spectrometer coupled directly to a Thermo-Electron Focus gas chromatograph (GC) equipped with an autosampler AS 3000 (ThermoElectron, Dreieich, Germany). Quantification was performed by selected-ion monitoring (SIM) of ions with mass-tocharge (m/z) ratios of m/z 46 for 16  The previously reported GC-MS method for nitrite and nitrate had been originally validated for 100-mL sample aliquots 15,16 . This method was adapted to 10-mL aliquots and validated for the CA microassay which involves the use of H 2 18 O needed to prepare the aqueous buffer 3 . For highest derivatization yield of nitrite, the incubation time was 5 min. Under these conditions, nitrate cannot be quantified accurately 15 . We compared in parallel the methods using 100-mL (in quadruplicate) and 10-mL aliquots (in duplicate) of nitrite solutions in 100 mM Tris-HCl buffer, pH 7.4, in H 2 16 O at added nitrite concentrations of 0, 2.5, and 5.0 mM. These concentrations were chosen because they were expected to be formed in experiments in H 2 18 O-Tris buffer. The relative standard deviation values were 2.9% and 2.4%, 0.5% and 1.0%, and 3.1% and 4.3%, respectively. The regression equations obtained from linear regression analyses of measured nitrite (y) versus added nitrite (x) were y ¼ 0.752 þ 0.645x, r 2 ¼0.9945 for the 100-mL samples and y ¼ 0.800 þ 0.668x, r 2 ¼0.9998 for the 10-mL. The y-axis intercepts reveal nitrite concentrations in the Tris-HCl buffer of 0.75 mM and 0.80 mM, respectively. The slope values of the regression equations of 0.645 and 0.668 are very close indicating almost complete agreement (96.5%) between the procedures in the investigated concentration range.

Experiments with washed human platelets
Human washed platelets were prepared as reported previously 17 with small modifications. Blood was obtained from healthy volunteers, after they gave informed consent, according to our institutional guidelines and the Declaration of Helsinki. Studies using human platelets were approved by the Sechenov Institute of Evolutionary Physiology and Biochemistry of the Russian Academy of Sciences (IEPHB RAS) (Study No.3-03; 02.03.2019). Blood was collected into ACD solution (12 mM citric acid, 15 mM sodium citrate, 25 mM D-glucose, all final concentrations). Platelet rich plasma (PRP) was obtained by 5 min centrifugation at 330 Â g; then collected PRP was centrifuged for 10 min at 430 Â g, the pelleted platelets were washed in CGS buffer (120 mM sodium chloride, 12.9 mM trisodium citrate, 10 mM D-glucose, pH 6.5), and suspended in modified HEPES buffer (140 mM sodium chloride, 20 mM sodium bicarbonate, 5 mM potassium chloride, 1 mM magnesium chloride, 5 mM D-glucose, 10 mM HEPES, pH 7.4) to a final cell concentration of 3 Â 10 8 platelets/mL. After 15 min rest in a 37 C water bath, washed platelets (100 mL/tube) were used in the experiments. Platelets were incubated with 10 mM sodium nitrite, at 100 mM or 200 mM concentrations of L-cysteine. The NO donor, sodium nitroprusside (SNP, 1 mM) was used as a positive control. Acetazolamide (200 mM) was used for inhibition of CA activity, sGC activity was inhibited by 20 mM ODQ (both inhibitors were preincubated for 10 min). All chemicals in these experiments were obtained from Sigma-Aldrich (Steinheim, Germany).
For Western blot analysis, after incubation with the mentioned compounds, 100 mL of 3x Laemmli buffer were added to each tube, proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were incubated with Phospho-VASP Ser239 Nanotools (Teningen, Germany) or actin (Cell Signalling, Frankfurt/am Main, Germany) primary antibodies overnight at 4 C. For visualisation of the signal, goat anti-mouse (for Phospho-VASPSer239) or anti-rabbit (for actin) IgG conjugated with horseradish peroxidase were used as secondary antibodies followed by electrochemiluminescence detection (GE Healthcare). Blots were analysed densitometrically using NIH Image J software for uncalibrated optical density.

Experiments with purified S-[ 15 N]nitrosoalbumin
In order to investigate the effects of protein-bound Cu 2þ and free L-cysteine on the formation of nitrite from a high-molecular-mass S-nitrosothiol we performed investigations with freshly prepared and purified S-[ 15 N]nitrosoalbumin without bound Cu 2þ (i.e. S 15 NALB) and with Cu 2þ bound to S 15 NALB (i.e. Cu 2þ -S 15 NALB), as well as with native albumin without bound Cu 2þ (i.e. ALB) and with Cu 2þ bound to albumin (i.e. Cu 2þ -ALB) as described elsewhere [18][19][20] . For this, each 6-ml aliquots of freshly prepared plasma obtained by centrifugation (10 min, 800 Â g, 4 C) of EDTA-anticoagulated blood donated by a healthy young volunteer were used. These samples were used to prepare ALB and S 15 NALB. Aliquots (2.5 ml) of the plasma samples were incubated for 10 min at room temperature with 1 mM CuSO 4 to prepare Cu 2þ -ALB and Cu 2þ -S 15 NALB. All preparations were extensively purified using first Sephadex PD10 columns. Then eluates were extracted using 5-ml HiTrapBlue Sepharose affinity columns and washed with 100 ml of 67 mM potassium phosphate buffer, pH 7.0 (buffer A) in order to eliminate remaining free Cu 2þ . Albumin species were eluted with 5 ml of 67 mM potassium phosphate buffer, pH 7.0, that contained 1.5 mM KCl (buffer B), the eluates were reduced to about 1-ml using a stream of nitrogen gas and used subsequently without delay.
Each 50-mL aliquots of S 15 NALB and Cu 2þ -S 15 NALB were incubated for 2 min at room temperature in buffer A in the absence (-Cys) or in the presence (þCys) of L-cysteine at a final concentration of 100 mM. Each 50-mL aliquots of S 15 NALB were incubated with ALB or Cu 2þ -ALB followed by their incubation in the absence (-Cys) or in the presence (þCys) of L-cysteine at a final concentration of 100 mM as well. In order to quantify 15 N-nitrite in the samples, they were spiked with 14 N-nitrite as the internal standard at a final concentration of 10 mM and derivatized as described previously 15 . Each four 100-mL aliquots of all samples were treated with 10 mL of PFB-Br and 400 mL of acetone. Subsequently, the samples were incubated for 5 min at room temperature to generate the PFB derivatives of 15 N-nitrite and 14 N-nitrite with the latter serving as the internal standard. PFB derivatives were extracted with toluene (300 mL) and 1-mL aliquots thereof were analysed by GC-MS in the SIM mode: m/z 47 for 15 N-nitrite and m/z 46 for 14 Nnitrite. The concentration of 15 N-nitrite in the samples was calculated by multiplying the peak area ratio of m/z 47 to m/z 46 by 10 mM, which was the concentration of the internal standard in the samples.

Results
We demonstrated that bovine CA II (bCA II) catalyses the incorporation of 18 O into nitrite from H 2 18 O used to prepare the pH-neutral buffered bCA II solution ( Figure 1). Previously, we found that acetazolamide does not affect bCA II-catalyzed formation of NO from nitrite in the presence of exogenous L-cysteine at pH 7.4 3 . In that study we also found that acetazolamide itself did not affect the 18 O-incorporation from H 2 18 O into nitrite in the absence of bCA II, 3 suggesting exclusive involvement of bCA II in this process. In the absence of L-cysteine, acetazolamide can modulate bCA IIcatalyzed incorporation of 18 O from H 2 18 O into nitrite in a concentration-dependent manner, leading to a higher 18 O-incorporation at 50 mM acetazolamide and to a lower 18 O-incorporation at 250 mM acetazolamide compared to no use of acetazolamide ( Figure 1). The acetazolamide concentration of 50 mM is pharmacologically relevant in human plasma (mean C max , 120 mM 21 ). The acetazolamide concentration of 250 mM occurs in human urine upon ingestion of a single 250-mg acetazolamide tablet by  The experiment was performed as described elsewhere 3 . Data are shown as mean with standard deviation from four independent incubations. Statistical analysis was performed by one-way ANOVA and Mann-Whitney test.
healthy subjects 22 . The results of Figure 1 suggest that acetazolamide may modulate the nitrous anhydrase activity of bCA II in a non-linear concentration-dependent manner. A similar phenomenon was observed for the esterase activity of human CA II (hCA II) towards carbohydrate-based sulfamate inhibitors 23 . By using this method we found that bovine and human erythrocytic CA II catalyses the incorporation of 18 3 . The presence of exogenous L-cysteine at a concentration of 100 mM was found to potentiate the nitrous anhydrase activity (incorporation of 18 O) 3 . In present study we also measured the concentration of remaining (unlabeled) nitrite, i.e. 16 O 2 -nitrite (m/z 46). The results of these investigations are illustrated in Figure 2. To our surprise, we observed a strong decrease in the concentration of 16 O 2 -nitrite (m/z 46), i.e. from 56.9 ± 3.35 mM without L-cysteine to 3.85 ± 1.28 mM in the presence of 100 mM L-cysteine (P ¼ 0.029). bCA II-mediated consumption of nitrite was observed in the absence of external L-cysteine, albeit to a smaller extent (e.g. 70.3 ± 4.8 mM vs. 54.2 ± 11.4 mM; by 23%). In the absence of bCA II no such decrease or formation of m/z 48 and m/z 50 in the 100 mM Tris-HCl buffer, pH 7.4, was observed. These findings have not been reported previously and suggest a considerable consumption of nitrite in a manner dependent on bCA II and L-cysteine. Possible reactions could involve nitration of certain CA II amino acid residues including Cys 205 , the single L-cysteine residue of CA II, and various L-tyrosine residues (8 Tyr residues in CA II 24 ). As tris(hydroxymethyl)aminomethane was used at 100 mM to prepare the Tris buffer, we cannot exclude a reaction of nitrous anhydride with tris(hydroxymethyl)aminomethane (pK a , 7.81) to form tris(hydroxymethyl)-N-nitroso-aminomethane which would simulate nitrite consumption. In the absence of L-cysteine the molarity of Tris buffer (10 À 100 mM) had not effect of nitrite consumption 3 . In this context it is notable that tris(hydroxymethyl)aminomethane can also react with p-nitrophenylacetate 25 which is commonly used to measure the esterase activity of CA.
Incubation of washed human platelets with 10 mM sodium nitrite or 100 mM L-cysteine alone did not result in an increase of the P-VASP/actin ratio as compared to control (Figure 3). The intra-platelet P-VASP/actin ratio increased almost 6-fold upon of co-incubation of platelets with 10 mM sodium nitrite and 200 mM L-cysteine ( Figure 3). This increase was very similar to that observed by incubating the washed human platelets with 1 mM sodium nitroprusside (SNP, Na 2 [Fe(CN) 5 NO]). The P-VASP/actin ratio was dependent upon the added L-cysteine concentration. At 100 mM acetazolamide, the P-VASP/actin ratio observed from 10 mM sodium nitrite/100 mM L-cysteine was almost the same as in the absence of acetazolamide, apparently suggesting no involvement of platelet CA. In contrast, pre-incubation of washed human platelets with ODQ, an inhibitor of sGC 26 , inhibited the phosphorylation of VASP, suggesting conversion of sodium nitrite to NO.
Human serum albumin (ALB) has a single cysteine moiety that does not form intra-molecular disulphide bridges. This cysteine moiety (Cys34) is accessible to modifications including S-nitrosation by nitrous anhydride and alkyl nitrites to form S-nitrosoalbumin (SNALB). SNALB itself is not an NO donor, but can release NO indirectly by means of free L-cysteine and intermediate formation of S-nitrosocysteine (CysSNO) via S-transnitros(yl)ation reactions. [18][19][20] In the present study we investigated whether Cu 2þ ions bound to albumin (Cu-ALB) and S 15 NALB (Cu-S 15 NALB) may modulate the L-cysteine-dependent release of NO from SNALB in buffer of neutral pH. Because nitrite (i.e. 14 N-nitrite) is abundantly present as a contamination we used commercially available 15 Nnitrite, i.e. the nitrite analog labelled with the stable isotope 15 N, of which the natural abundance is 0.36%. The results of this experiment are shown in Figure 4.
Addition of L-cysteine at 100 mM to virtually Cu 2þ -free S 15 NALB alone or in the presence of Cu 2þ -free ALB in potassium phosphate buffer did not result in elevated formation of 15 N-nitrite. In contrast, addition of L-cysteine at 100 mM to Cu 2þ -bound S 15 NALB (Cu-S 15 NALB) alone or in the presence of Cu 2þ -containing ALB (Cu-ALB) did result in elevated formation of 15 N-nitrite by 37% and 34%, respectively. These results indicate that free L-cysteine interacts with Cu-S 15 NALB and S 15 NALB À Cu-ALB. However, the results of Figure 4 do not exclude the possibility that the incubation of S 15 NALB with Cu-ALB have resulted in formation of Cu-S 15 NALB prior to the addition of L-cysteine.

Discussion
In the sections that follow below, we discuss the results of the present study and previous studies from our groups in the context of the results reported by other groups with respect to CA II functions as nitrous anhydrase or nitrite reductase. The discussion addresses mechanistic aspects and results from in vitro and in vivo studies. Based on missing differences of single iNaNO 2 and combined administration of iNaNO 2 and ivACZ with respect to various parameters, the authors concluded that CA is not a relevant nitrite reductase or nitrous anhydrase in the lung 13 . We agree with these authors that there is no convincing structure within the active site of the Zn 2þ -containing hCA II (i.e. Zn-hCA II), notably including the central redox-inactive Zn 2þ , or elsewhere in this protein capable of supporting one-electron reduction of nitrite (N oxidation state, þ3) to form NO (N oxidation state, þ2):  27 reported that Zn-CA II-induced generation of NO from nitrite, which was even increased by the CA inhibitor dorzolamide stronger at pH 5.9 compared to pH 7.3. A possible explanation for these findings could be NO formation from the chemical reaction of nitrite with the amine (NH 2 ) group of the sulphonamide functionality of dorzolamide at pH 5.9. Such reactions have been reported for many sulphonamides of which the amine group is quite acidic and easily nitrosable by nitrous acid (ONOH) and N 2 O 3 to finally decompose to NO 28 . An alternative explanation could be that the Zn-CA II preparation used in that study contained contaminating Cu 2þ (see below). On the other hand, we do not agree with Pickerodt et al. 2019 13 that CA II does not exert nitrous anhydrase activity to produce N 2 O 3 from the labile intermediate nitrous acid: The conversion of ONOto N 2 O 3 is not a redox reaction and does not require any electrons, but solely protons that can be provided by CA II even in the absence of CO 2 /HCO 3 -(R1). The results by Aamand et al. 1 are also supportive of a nitrous anhydrase activity of bCA II leading to the gaseous thus far undetectable N 2 O 3 which may in part decompose to NO 4 , thus producing signals when measured by a NO-electrode or by chemiluminescence. Andring et al. 11 reported that Zn-CA II, a cytosolic CA isozyme, does not exhibit nitrate reductase or nitrous anhydrase activity. We agree that Zn-CA II is not a nitrate reductase, but we disagree that Zn-CA II is not a nitrous anhydrase. Our groups 2,3 found by liquid chromatography-tandem mass spectrometry (LC-MS/MS) that bCA II and hCA II generated in the presence of glutathione (GSH) S-nitrosoglutathione (GSNO); we also found that in the presence of exogenous L-cysteine, bCA II and hCA II mediated NO formation as measured by a NO-sensitive electrode most likely via intermediate formation of S-nitrosocysteine (CysSNO), a potent NO donor. GSNO and CysSNO exert both cGMP-dependent (via NO release) and cGMP-independent (via S-transnitrosylation) biological activities, including vasodilation and inhibition of platelet aggregation. [29][30][31] The results reported in the present study using washed human platelets, nitrite and L-cysteine are in support of intraplatelet Zn-CA II-mediated conversion of nitrite to NO via intermediate formation of CysSNO from N 2 O 3 and L-cysteine, but not via reduction of nitrite to NO. We have demonstrated by mass spectrometry-based proteomic studies that CA II is physiologically present in platelets of healthy humans (35,000 copies/platelet). 10 Our studies suggest that upon hCA II-induced conversion of nitrite to N 2 O 3 , this anhydride has the potential to develop NOrelated bioactivity by subsequent redox-independent reactions. One mechanism may involve S-nitrosylation by N 2  activity, it is considered important for the stabilisation of the protein 32,33  Nitrite reacts both with water and CO 2 , and these reactions can be modulated by Zn-hCA II 35 . The pK a value of H 2 CO 3 during its formation in aqueous solution is 3.5 36 , i.e. it is much lower than the generally assumed value of 6.4. Zn-hCA II-dependent N 2 O 3 formation from nitrite and CO 2 could occur via nitritocarbonate/nitritocarbonic acid (O¼N-O-COO -/O¼N-O-COOH) across the proton transfer shuttle (R10). hCA Potential nitrite-dependent physiological and pharmacological roles of carbonic anhydrase Besides mechanistic aspects, other important issues should also be considered in the context of nitrite-dependent physiological and pharmacological roles of Zn-hCA and are discussed below. In the study by Pickerodt et al. 13 , sodium nitrite (NaNO 2 , 450 mg, 6.5 mmol) was inhaled by juvenile male 24-kg weighing pigs. This corresponds to a dose of 18.8 mg/kg body weight and is about 16 times higher than the reported maximum tolerated dose of iNaNO 2 in healthy humans 37 . After 1 h of hypoxia, arterial blood nitrite concentration increased 35-fold after iNaNO 2 alone and even 48-fold after co-administration with ivACZ 13 . After 2 h of hypoxia, arterial blood nitrite concentration was only 8 mM in both cases, suggesting a short elimination half-life of nitrite. Mean arterial blood pressure decreased almost equally with iNaNO 2 alone (79 to 60 mmHg) and in co-administration of acetazolamide (79 to 56 mmHg) after 1 h of hypoxia 13 . The average fractional excretion of NO (FENO) increased from 6 ppb at baseline to 46 ppb at 1 h, and fall to 27 ppb at 2 h, and to 20 ppb at 3 h upon iNaNO 2 alone; the corresponding FENO values for iNaNO 2 þivACZ were 6 ppb, 22 ppb, 13 ppb, and 11 ppb, i.e. constantly (almost by 50%) lower 13 . In that study, considerable differences between iNaNO 2 and iNaNO 2 þivACZ were also found for the respiratory rate. Despite considerable differences in some pharmacodynamic parameters, for instance 1) the lower FENO values (by 24, 14 and 9 ppb after 1 h, 2 h and 3 h of hypoxia) and 2) the mean pulmonary arterial pressure (MPAP) of the combination of iNaNO 2 þivACZ compared to iNaNO 2 alone (by 9 and 8 mmHg after 2 h and 3 h of hypoxia), the authors concluded that CA is not relevant to the bioactivation of nitrite 13 . We think that these observations may suggest that at the very high dose of iNaNO 2 used in the study 13 , iNaNO 2 may have attenuated/inhibited the nitrous anhydrase CA activity in the lungs. Acetazolamide has been demonstrated to increase the excretion of endogenous and exogenous nitrite and to a lower degree of nitrate in humans 16,22,38,39 . Although nitrite, nitrate and bicarbonate had not been measured in the urine, Pickerodt et al. 13 concluded, solely on the basis of the plasma nitrite concentration course, which in fact differed between the groups, that acetazolamide did not alter renal function including CA-dependent reabsorption of nitrite and nitrate. As far as we are informed, plasma nitrite concentrations of the order of 35 mM and 48 mM as found at 1st h of hypoxia 13 have not been reported until the present day. At such high concentrations the activity of many enzymes including erythrocytic catalase are likely to be inhibited to a considerable degree thus increasing oxidative stress and methemoglobinemia 40 .
In healthy subjects (mean weight, 60.7 kg), iNaNO 2 (75 mg, 1.09 mmol) for 10 min resulted in a 20-fold increase of the mean blood nitrite concentration (from 0.2 mM to 4 mM) which decreased with an elimination half-life of 0.6 h 41 . The average FENO increased from 10 ppb at baseline to 40 ppb and fell to the baseline levels 10 min after NaNO 2 inhalation 41 . No significant changes of systolic and diastolic pressure were observed in the human study, in which the iNaNO 2 dose was 1.2 mg/kg body weight, i.e. much lower compared to the dose of 18.8 mg/kg body weight used in the study by Pickerodt et al. 13 . The results of the studies by Pickerodt et al. 13 and Sirirat et al. 41 indicate a very rapid elimination of iNaNO 2 in accordance with an elimination half-life of about 0.5 h in humans 37 . The nitrite bioavailability of iNaNO 2 in healthy subjects was reported to be 18% 41 . Unfortunately, neither nitrite nor nitrate excretion rates in the urine were reported in the studies by Pickerodt et al. 13 and Sirirat et al. 41 . Such data would have allowed evaluating the expected drastically elevated excretion of nitrite due to inhibition of renal CA activity by acetazolamide in the animal study, as observed by us in humans at a pharmacological dose of acetazolamide 16,22,39 .
In 16 healthy young subjects (mean age, 23 years) who received intravenous infusion of NaNO 2 (approximate dose, 0.48 mg/kg body weight) for 2 h after a 4-days pharmacological treatment period with placebo (only NaNO 2 , no drug) as control, allopurinol to inhibit xanthine oxidoreductase (XOR) activity, enalapril to inhibit angiotensin converting enzyme (ACE) activity, or acetazolamide (250 mg thrice a day) to inhibit CA activity, consistent results with respect to many parameters were obtained with all medications 6 . The authors concluded that the effects observed in their study, including the decrease of mean arterial pressure (À1.88 mmHg by placebo; À1.84 mmHg by allopurinol; À2.07 mmHg by enalapril; À2.32 mmHg by acetazolamide) suggest that XOR, ACE and CA are not essential for nitrite bioactivation to NO 6 . Yet, this conclusion is not convincing for the following reasons: the NaNO 2 dose was low and the infusion period was relatively long when compared with the short elimination half-life of nitrite of 0.6 h; also the greatest effect of acetazolamide on nitrite excretion is quite rapidly exerted 16,22,39 .
A sophisticated animal study was performed by Wang et al. 14 . In CA II þ/þ , CA II ± and CA II -/mice (26 À 32 g body weight), infusion of NaNO 2 (30,50,100, 500, 2500 nmol over 5 min corresponding to doses of up to 3.8 mg nitrite/kg body weight) lowered mean arterial blood pressure equally from about 70 mmHg down to 40 mmHg in CA II þ/þ , CA II ± and CA II -/-, suggesting a mechanism independent of CA II 14 . Unfortunately, the authors did not report on the possible formation of S-nitrosothiols, which could have been an indication of hCA II involvement. It is worth mentioning that increase in plasma S-nitrosothiol concentration was observed upon iNaNO 2 for 10 min in doses of 0.06 to 2.2 mg NaNO 2 /kg body weight in healthy subjects 37 . Our group provided unequivocal evidence of the formation of 15 N-labelled S-nitrosoglutathione (GS 15 NO) from 15 N-labelled nitrite and GSH by means of a commercially available recombinant human erythrocytic CA II (heCA II) in Tris buffer at pH 7.4 in the absence of externally added bicarbonate 2 . In washed human platelets, the activity of sGC upon incubation with 100 mM nitrite, 20 mM bicarbonate and bovine erythrocytic CA II was comparable with that observed with 1 mM S-nitrosocysteine (CysSNO), one of the strongest endogenous inhibitors of platelet aggregation 29 . In the present study, comparable effects were obtained with 10 mM nitrite and 100 mM L-cysteine. This may suggest that 10 mM nitrite/100 mM L-cysteine is about 10 less active that SNP regarding NO formation.
Although CAs are investigated for several decades by many groups from various perspectives including biology and pharmacology, the CA family conceals many secrets and surprises 23 that remain to be revealed. While there is solid evidence of the participation of renal CA II and CA IV in the reabsorption of endogenous and exogenous nitrite 16,39 , the mechanisms underlying the bioactivation by CA II of nitrite to species such as N 2 O 3 with the potential of developing NO-related activity are still elusive and warrant further research. In vitro, haemoglobin species, XOR, ACE and CA have been shown to bioactivate nitrite under some specific conditions which included hypoxia and slightly acidic pH. The use of inhibitors of hCA II, XOR and ACE in studies on the bioactivation of nitrite is problematic for several reasons. The nitrous anhydrase activity of CA II and its importance in the bioactivation of inorganic nitrite to NO and S-nitrosothiols warrants further characterisation. Pharmacological acetazolamide exerts dilatory effects by not yet well-understood mechanisms, presumably in part independent of NO/cGMP 23 . The underlying mechanisms need elucidation before acetazolamide and other CA inhibitors can be used to test the involvement of hCA II in the bioactivation of nitrite. In this context the possibility should be considered that the use of hCA II inhibitors may result in loss of nitrite due to elevated renal excretion in the urine thus counteracting its bioactivation in other tissues. Measurement of urinary nitrite excretion in in vivo studies investigating the bioactivation of inorganic nitrite by CAs is mandatory. Lack of elevated urinary nitrite excretion and of other measures in the CA II -/mice 14 could be due to alterations of other CA isozymes and proteins such as renin and aquaporin [42][43][44] .
Cu 2þ ions were found to bind to Zn 2þ -CA II isolated from human erythrocytes at a site other than the active site and to inhibit the exchange of water from the enzyme without affecting the equilibrium rate of hydration of CO 2 by Tu et al. 45 almost four decades ago. This observation may suggest that classical CA inhibitors such as acetazolamide may inhibit the carbonic anhydrase activity of CA by tightly binding to CA II-bound Zn 2þ , but not to the second Cu 2þ -binding site. This may be an explanation for our previous 2,3 and present ( Figure 3) observations that neither acetazolamide nor dorzolamide inhibited the nitrous anhydrase activity of isolated bovine and human CA II.
The results of the present study observed with Cu-S 15 NALB and Cu-ALB suggest that L-cysteine is required for the bioactivation of the S-nitroso group of S-nitrosoalbumin (SNALB) either containing firmly bound Cu 2þ or in the presence of non-S-nitrosylated but Cu 2þ bound to albumin (Cu-ALB). It is likely that similar effects and mechanisms may also occur in ZnCu-CA II and CuCu-CA II. Yet this remains to be demonstrated by experiments analogous to those performed with Cu-S 15 NALB and Cu-ALB in the present study.
Zn-hCA II that additionally contains Cu 2þ arising from contaminations or from Cu 2þ -containing proteins/enzymes such as ceruloplasmin could act as a nitrite reductase. The dependence of the nitrous anhydrase/nitrite reductase activity of CA II upon L-cysteine would be compatible with these particular activities. Reaction of L-cysteine with N 2 O 3 in the bulk would form CysSNO (R8) of which the S-nitroso group is reduced to NO by Cu 1þ ions (R11) formed from the reaction of Cu 2þ with L-cysteine (R12). The socalled copper carbonic anhydrase has been in the past 46 52 . It is also unknown whether L-cysteine or ascorbic acid is required as reductants for Cu 2þ bound to hCA II.
Human serum albumin (HSA) contains two specific Cu 2þ binding sites, the N-terminal site (NTS) and the multimetal binding site (MBS). The NTS binds Cu 2þ ions much more strongly than the MBS, and it is therefore considered to be the only HSA site ever occupied by Cu 2þ ions in blood serum. In vitro, the incorporation of Cu 2þ into HSA is fast. The affinity of Cu 2þ to the NTS of HSA is 1 pM 53 . Under physiological conditions only about 1% of HSA molecules are estimated to carry a Cu 2þ ion. About 25% of HSA molecules derived from human blood have their N-terminal dipeptide clipped and do not possess an NTS. The NTS is composed of the first three amino acid residues of the HSA sequence: Asp-Ala-His. In our study we did not determine the concentrations of Cu-S 15 NALB and Cu-ALB. We consider that they correspond to the measured protein concentrations. Externally added Cu 2þ ions are likely to be bound on the NTS of our synthesised and purified Cu-S 15 NALB and Cu-ALB preparations, with no appreciable mutual hindrance and fully accessibility for L-cysteine. It is possible that Cu 2þ -containing CA may act as a nitrite reductase analogous to Cu 2þ -bound HSA, but this remains to be demonstrated in forthcoming experiments.

Conclusions and perspectives
Several endogenous and exogenous sources contribute to inorganic nitrite. Do bovine and human carbonic anhydrase isoforms, notably CA II, possess nitrous anhydrase activity or nitrite reductase activity like bacterial nitrite reductases or possibly both? These questions are currently in the focus of scientific research due to the potential bioactivation of inorganic nitrite to nitric oxide (NO) and nitrous anhydride (N 2 O 3 ). NO is a potent vasodilatator and inhibitor of platelet aggregation, which are considered to be beneficial in certain cardiovascular diseases. N 2 O 3 can nitrosate numerous biomolecules thus potentially contributing to health via S-nitrosation and subsequent NO release, or to disease such as cancer via N-nitrosation. Existing results from in vitro and in vivo experiments in animals and humans including the effects of established and novel carbonic anhydrase inhibitors, including those for treatment of hypoxic tumours, 54 are contradictory, often misinterpreted (discussed in References 52,55 ), and difficult to reconcile. The lack of an inhibitory effect of acetazolamide on nitrous anhydrase/nitrite reductase activity of carbonic anhydrase is often interpreted as a CA-independent effect. The potentiation of the effects of CA on nitrite by L-cysteine seen in human washed platelets in our study is compatible with nitrous anhydrase activity of intra-platelet CA accompanied by subsequent NO release. This can be concluded because measuring of VASP phosphorylation is considered to reflect platelet-derived NO 56 . This effect may simulate "nitrite reductase" activity of CA. Yet, due to the ability of CA to bind Cu 2þ on a site distinctly different from the Zn 2þ site, further investigations using sophisticated analytical approaches, such as GC-MS in combination with stable-isotope labelled water and nitrite 57 , and proteomics, are required to reveal hidden biochemical properties of CA II and other CA-isoforms beyond their inherent carbonic anhydrase activity (R1). S-Glutathionylation of CA VII was found not to be associated with loss of catalytic activity and affinity for sulphonamide inhibitors 58 . Such investigations are required for Zn-hCA II regarding the involvement of Cys 205 in the nitrous anhydrase activity, and for Cu-hCA II regarding the involvement of Cys 205 in its proposed nitrite reductase activity.

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