Redox Control of Calcineurin by Targeting the Binuclear -Zn Center at the Enzyme Active Site*

The interaction of protein serine/threonine phosphatase calcineurin (CaN) with superoxide and hydrogen peroxide was investigated. Superoxide specifically inhibited phosphatase activity of CaN toward RII (DLD-VPIPGRFDRRVSVAAE) phosphopeptide in tissue and cell homogenates as well as the activity of the enzyme purified under reducing conditions. Hydrogen peroxide was an effective inhibitor of CaN at concentrations several orders of magnitude higher than superoxide. Inhibition by superoxide was calcium/calmodulin-depend-ent. Nitric oxide (NO) antagonized superoxide action on CaN. We provide kinetic and spectroscopic evidence that native, catalytically active CaN has a Fe 2 (cid:1) -Zn 2 (cid:1) binuclear center in its active site that is oxidized to Fe 3 (cid:1) -Zn 2 (cid:1) by superoxide and hydrogen peroxide. This oxidation is accompanied by a gain of manganese dependence of enzyme activity. CaN isolated by a conventional purification procedure was found in the oxidized, ferric enzyme form, and it became increasingly dependent on divalent cations. These results point to a complex redox regulation of CaN phosphatase activity by superoxide, which is modified by calcium, NO, and superoxide dismutase. EPR Spectroscopy— EPR spectra were recorded at 9.5 GHz (X-band) microwave frequency on a Bruker ESP300E spectrometer with a dual-mode resonator and peripheral equipment described elsewhere (20). Purified porcine CaN (with equivalent amounts of calmodulin added) was concentrated to 20 mg/ml in an Ultrafree-4 centrifugal device (Millipore). After recording the spectrum, the sample was thawed and incubated with 10 m M H 2 O 2 in the presence of 1 m M CaCl 2 5 min at room temperature, and the spectrum was recorded again.

The interaction of protein serine/threonine phosphatase calcineurin (CaN) with superoxide and hydrogen peroxide was investigated. Superoxide specifically inhibited phosphatase activity of CaN toward RII (DLD-VPIPGRFDRRVSVAAE) phosphopeptide in tissue and cell homogenates as well as the activity of the enzyme purified under reducing conditions. Hydrogen peroxide was an effective inhibitor of CaN at concentrations several orders of magnitude higher than superoxide. Inhibition by superoxide was calcium/calmodulin-dependent. Nitric oxide (NO) antagonized superoxide action on CaN. We provide kinetic and spectroscopic evidence that native, catalytically active CaN has a Fe 2؉ -Zn 2؉ binuclear center in its active site that is oxidized to Fe 3؉ -Zn 2؉ by superoxide and hydrogen peroxide. This oxidation is accompanied by a gain of manganese dependence of enzyme activity. CaN isolated by a conventional purification procedure was found in the oxidized, ferric enzyme form, and it became increasingly dependent on divalent cations. These results point to a complex redox regulation of CaN phosphatase activity by superoxide, which is modified by calcium, NO, and superoxide dismutase.
Calcineurin (CaN, 1 also called protein phosphatase 2B) is a protein serine/threonine phosphatase regulated by calcium/calmodulin. The list of physiological CaN substrates is growing and its important role in such processes as T-cell activation, heart and skeletal muscle hypertrophy, synaptic plasticity, and apoptosis has been recognized (reviewed in Refs. 1 and 2). The significance of CaN for the activation of T-cells by dephosphorylation of NFAT (nuclear factor of activated T cells) was revealed after the discovery of its inhibition by the immunosuppressive drugs cyclosporin A and FK506 (3). Recent interest has been focused on CaN regulatory mechanisms beyond calcium/calmodulin, and redox events have been suggested as modulators of phosphorylation/dephosphorylation pathways (4,5). Among reactive oxygen species (ROS), superoxide (O 2 . ) (6) and hydrogen peroxide (H 2 O 2 ) (7-10) had been suggested as inhibitors of CaN activity, but the complex enzymology of CaN has led to conflicting conclusions about the mechanism of inhi-bition. With regard to the H 2 O 2 effect, we have postulated a dithiol-disulfide transformation, since inhibition of the isolated enzyme was accompanied by the disappearance of two thiol groups, and reactivation was achieved by thioredoxin and dithiols (10). Furthermore, CaN activity could be blocked by phenylarsine oxide, a selective agent for vicinal dithiols at low concentrations. However, the H 2 O 2 concentrations required for inhibition were well above 100 M and, hence, out of a physiological range. Therefore, the previously suggested inhibition by superoxide (6) seemed to be a more relevant mechanism under physiological conditions, especially because we had obtained evidence that O 2 . alone or in combination with the ⅐NO radical can participate in a variety of signaling mechanisms (11).
In their pioneering report Klee and co-workers (6) propose that the metal catalytic center in CaN is targeted by superoxide. Similarly to related phosphatases PP1, PP2A, and purple acid phosphatases, CaN contains a binuclear metal center necessary for catalysis (12). Metal analysis indicated that iron and zinc are the components of this center (13,14), but the iron oxidation state in native CaN is controversial (1,2). Klee and co-workers (6) suggest the existence of superoxide-sensitive ferrous iron, but on the basis of spectroscopic and kinetic studies on isolated bovine CaN it was later suggested by Rusnak and co-workers (14,15) that native CaN exists in a redoxinsensitive Fe 3ϩ -Zn 2ϩ form. A redox-sensitive Fe 3ϩ -Fe 2ϩ form of CaN resembling the well characterized binuclear center in bovine spleen purple acid phosphatase was artificially obtained by metal exchange (14,16). It was hypothesized that this form could also contribute to redox sensitivity of native CaN in tissue homogenates (9,17). However, no evidence for the existence of such form in vivo has yet been obtained.
In the present work we attempted to clarify the mechanisms of CaN redox control with emphasis on the possible sensitivity of the enzyme concentrations were determined by using ⑀ 260 ϭ 2086 M Ϫ1 cm Ϫ1 . CaN Isolation-Bovine brain CaN was isolated according to purification procedure of Sharma et al. (18) as described in Bogumil et al. (10). Porcine brain CaN purification involved the following important modifications; ammonium sulfate precipitation and Affi-Gel blue chromatography steps were omitted, and calmodulin affinity chromatography peak fractions were subjected to gel filtration on a Superdex 200 col-* This work was supported by grants from Deutsche Forschungsgemeinschaft (DFG). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Preparation of Cell and Tissue Extracts-Jurkat or RAW 264.7 cells (0.5-1ϫ10 6 ) were pelleted by centrifugation (12,000 ϫ g, 5 s), washed with cold phosphate-buffered saline, and lysed in 75 l of lysis buffer (50 mM Tris-HCl, pH 7.5, 0.5 mM TCEP, 0.2 mM EGTA, 50 g/ml phenylmethylsulfonyl fluoride, 50 g/ml soybean trypsin inhibitor, and 10 g/ml leupeptin) by three cycles of freezing-thawing in liquid N 2 . The suspension was centrifuged for 10 min at 14,000 rpm and 4°C, and the supernatant was collected. After measuring the protein with BCA reagent (Pierce), its concentration was adjusted to 0.5-1 mg/ml with the lysis buffer. Bovine cerebral cortex was homogenized in the lysis buffer using Polytron tissue dispersion (Kinematica) and centrifuged for 10 min at 14,000 rpm and 4°C. The supernatant was diluted after protein determination to 1 mg/ml with the lysis buffer.
RII Phosphorylation-RII peptide (DLDVPIPGRFDRRVSVAAE, Bachem), corresponding to a portion of the regulatory RII subunit of protein kinase A, was serine-phosphorylated with [␥-32 P]ATP by using the catalytic subunit of protein kinase A. 20 l of 7.8 mM RII stock solution was added to a kinase reaction mixture containing 50 mM triethanolamine-HCl, pH 7.5, 5 mM MgCl 2 , 0.3 mM ATP, 1 mCi [␥-32 P]ATP, and 200 units of protein kinase A in a volume of 1 ml. The reaction was allowed to proceed for 2 h at 30°C, and the phosphorylated peptide was purified on 1-ml Sep-Pak C18 columns (Waters) pre-equilibrated with 5 ml of 30% acetonitrile, 0.1% trifluoroacetic acid followed by 8 ml 0.1% trifluoroacetic acid. After loading of the reaction mixture, the column was washed excessively with 0.1% trifluoroacetic acid. The peptide was eluted with 30% acetonitrile, 0.1% trifluoroacetic acid, then lyophilized and re-suspended in a total volume of 0.5 ml of 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1 mg/ml bovine serum albumin.
Phosphatase Assay-The assay was performed essentially according to Fruman et al. (19). The assay buffer consisted of 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.5 mM TCEP, 1 mg/ml bovine serum albumin, 6 M RII peptide, 1 M okadaic acid, and 0.4 mM CaCl 2 . To 20 l of assay buffer an equal volume of protein diluted in lysis buffer was added, and the reaction was conducted at 30°C for 5-10 min. The reaction was stopped by adding 50 l of 20% DOWEX 50-X8 (200 -400 mesh, Bio-Rad) in 10% trifluoroacetic acid. The mix was centrifuged 6 min at 14,000 rpm, and 40 l of supernatant were taken for Cherenkov counting. The amount of protein was adjusted so that the substrate consumption did not exceed 25%. Duplicate cpm values from the phosphatase assay were averaged, and the resulting value was adjusted by subtracting the counts in blanks lacking cell lysate. This was divided by the specific activity of the substrate to give picomoles of phosphate released. Finally, this was divided by the reaction time and the amount of protein to give values expressed as nmol phosphate/min/mg of protein. In the experiments with isolated CaN, 300 nM calmodulin was present in the assay, and TCEP was omitted from the buffers with no effect on CaN activity.
Limited Proteolysis of CaN-Purified porcine CaN (1 M) was incubated in 50 mM Tris-HCl buffer containing 100 mM NaCl, 0.5 mM CaCl 2 , 0.5 mM TCEP, and 5 mM ascorbate, pH 7.5, with 0.05 g of trypsin for 5 min at 30°C. The reaction was stopped by adding 0.2 g of soybean trypsin inhibitor. The reaction was diluted into lysis buffer before incubation with XO. Parallel samples were subjected to SDS-PAGE to control the extent of proteolysis.
Quantitation of Superoxide Production-The amount of superoxide generated by the XO/hypoxanthine system was assayed by measuring the rate of SOD-inhibitable cytochrome c reduction. Phosphatase assay buffer (50 mM Tris-HCl, 50 mM NaCl, 0.1 mM EGTA, 0.2 mM CaCl 2 , 0.5 mg/ml bovine serum albumin, and protease inhibitors, pH 7.5) containing 50 M cytochrome c was used, and the increase of A 550 was followed spectrophotometrically at 30°C after the addition of various XO amounts in the presence of 100 M hypoxanthine and 200 units/ml catalase. Under these conditions 1-5 milliunits/ml XO generated 0.2-1 M O 2 . /min.

Metal Analysis-Porcine
CaN was buffer-exchanged on a Sephadex G-25 fast desalting fast protein liquid chromatography column into 50 mM Tris-HCl buffer, pH 7.4 (pre-treated with Chelex 100 chelating resin (Bio-Rad), to remove excess of metal ions). The metal content of the samples was determined by inductively coupled plasma mass spectrometry (Spurenanalytisches Laboratorium Dr. Baumann, Maxhü tte, Germany).
EPR Spectroscopy-EPR spectra were recorded at 9.5 GHz (X-band) microwave frequency on a Bruker ESP300E spectrometer with a dual-mode resonator and peripheral equipment described elsewhere (20). Purified porcine CaN (with equivalent amounts of calmodulin added) was concentrated to 20 mg/ml in an Ultrafree-4 centrifugal device (Millipore). After recording the spectrum, the sample was thawed and incubated with 10 mM H 2 O 2 in the presence of 1 mM CaCl 2 5 min at room temperature, and the spectrum was recorded again. When bovine CaN was isolated as reported (10), its activity in initial experiments, contrary to the homogenate, was found to be insensitive to inhibition by XO/hypoxanthine (data not shown). During the calmodulin affinity chromatography step in the standard procedure of CaN isolation the protein mixture was in contact with high calcium concentrations. These conditions, according to Klee and co-workers (6) could result in loss of CaN activity due to O 2 . -dependent oxidation, and this might be responsible for the insensitivity of the isolated bovine CaN to XO/hypoxanthine system. Because ascorbate was found to be protective against oxidative inactivation (6), we carried out the purification procedure with 5 mM ascorbate addition to all buffers involved in CaN isolation. Porcine brain homogenates served as a starting material since bovine tissues were not available due to the bovine spongiform encephalopathy crisis in Germany. We noticed that the newly isolated porcine CaN behaved differently toward stimulation with Mn 2ϩ ions (Table  I). In the presence of calmodulin the bovine enzyme was activated about 10-fold by Mn 2ϩ in the assay with 32 P-labeled RII-peptide. In contrast, Mn 2ϩ activated the porcine enzyme only about 3-fold. The gain of Mn 2ϩ dependence was accompanied with a loss of XO sensitivity; thus, the bovine enzyme retained about ϳ25% of the activity after incubation with XO, and the porcine enzyme was inhibited to 5-10% that of controls (data not shown). 2 Fig. 1B shows the concentration dependence of purified porcine CaN inactivation by the XO/hypoxanthine system. The IC 50 for XO at 0.05 milliunit/ml is 10-fold lower than IC 50 observed in brain homogenate, possibly reflecting O 2 . scavenging by other homogenate constituents.

Superoxide Inhibition of CaN Activity in Cellular Extracts and in Isolated
Although the XO/hypoxanthine system is widely used for O 2 . generation, it also co-generates substantial amounts of  (21,22). A similar activation could be achieved by limited proteolysis of CaN (23,24), which removes the autoinhibitory and calmodulin-binding domains and renders the enzyme constitutively active in the absence of calmodulin. To examine if the removal of these domains also relieves the calcium/calmodulin dependence of O 2 . inhibition, we subjected porcine CaN to limited proteolysis by trypsin. A five-min incubation reduced the apparent molecular mass of the CaN A subunit from 60 to 42 depending on the ligand field, either reduce ferric iron or oxidize ferrous iron with subsequent loss of catalytic activity. As mentioned before, the redox state of iron in CaN is under dispute (6,9,15,17). Metal content analysis of porcine CaN by inductively coupled plasma mass spectrometry showed the presence of equivalent amounts of iron and zinc in the protein (data not shown). To investigate the involvement of the metal center in the redox sensitivity of CaN, we treated the enzyme with 1 mM univalent oxidant K 3 Fe(CN) 6 or with the same concentrations of K 4 Fe(CN) 6 and KCN (Fig. 6A) inhibited CaN activity to 17% that of controls, whereas K 4 Fe(CN) 6 and KCN had only slight inhibitory effects, suggesting that the inhibitory effect of K 3 Fe(CN) 6 is due to enzyme oxidation. We also tested the effect of 5 mM ascorbate with or without XO/hypoxanthine on CaN activity (Fig. 6B). Ascorbate increased basal phosphate ester hydrolysis by CaN and protected the enzyme against inhibition by XO/hypoxanthine. Thus, CaN is inhibited by one-electron oxidizing agents and protected against O 2 . inhibition by the reductant ascorbate.
This indicates that CaN is likely to require Fe 2ϩ for optimal activity. The strong reductant dithionite was previously shown to inhibit CaN activity, interpreted as a requirement for Fe 3ϩ in the catalytically active enzyme (14,15). In our hands, however, direct addition of up to 10 mM dithionite to the assay did not inhibit the enzyme; rather, it caused apparent activation to 154 Ϯ 6% of control, again indicating the presence of reduced iron in the active CaN.
Next, we investigated whether the O 2 . -inactivated enzyme could be reactivated. In previous work of Wang et al. (6) it was found that the treatment of the calcium-inactivated enzyme with an Fe 2ϩ /ascorbate/dithiothreitol mix restored the enzyme activity in brain homogenates. We tested several combinations of metal ions for the ability to reactivate CaN after XO/hypoxanthine treatment (Fig. 7A). Ascorbate alone restored the activity of the enzyme to ϳ70% of the control value, and addition of 0.5 mM (NH 4  Manganese II ions were shown earlier not only to increase CaN activity but also to restore it after calcium/calmodulinmediated inactivation, although the mechanisms involved were poorly understood (13). We attempted to determine if Mn 2ϩ could restore CaN activity after inhibition by O 2 . . Fig. 7B shows that after incubation with 1 milliunit/ml XO in the presence of 100 M hypoxanthine, porcine CaN was inhibited to 17% of control, and this inhibition could be reversed by 1 mM Mn 2ϩ with activities being 65% of control values (in the presence of Mn 2ϩ ). Remarkably, the addition of the Fe 3ϩ chelator desferrioxamine further increased the ability of Mn 2ϩ to restore enzyme activity up to 89% of control. Desferrioxamine alone had no effect on the enzyme activity or on inhibition by O 2 . . The effect of desferrioxamine likely may be facilitation of iron removal from the cluster after its oxidation to Fe 3ϩ , and it indicated that Mn 2ϩ could probably substitute for iron in the enzyme binuclear cluster. The loss of iron upon enzyme oxidation was confirmed by metal analysis (inductively coupled plasma mass spectrometry), which showed that approximately half of the enzyme iron is lost after O 2 . treatment in the presence of calcium/calmodulin (data not shown). EPR Spectroscopy of Porcine CaN-The low temperature EPR spectrum of native porcine CaN (0.25 mM) is shown in Fig.  8A. Its main feature consists of a weak signal around g ϭ 4.3. Upon incubation of the CaN sample with 10 mM H 2 O 2 in the presence of calcium/calmodulin this signal was significantly increased, and appeared as a doublet (Fig. 8B). Additional features centered at g ϭ 6.5, 5.6 and 5.0 were present, of which the first two were previously assigned to a Fe 3ϩ species with near axial symmetry (10,15). The absence of these signals in the native sample provides further evidence for the presence of a Fe 2ϩ -Zn 2ϩ center in native, catalytically active CaN. DISCUSSION The results of this study provide kinetic and spectroscopic evidence that native CaN has a catalytically active Fe 2ϩ -Zn 2ϩ binuclear cluster. This cluster is particularly sensitive to O 2 . , and its inhibition is calcium-and calmodulin-dependent. Furthermore, NO antagonizes the O 2 . inhibitory action, adding another player to the CaN regulatory network. A role of ROS as modulators of CaN activity has been proposed lately (17). O 2 . and H 2 O 2 are considered major ROS generated in living cells; however, the potential of O 2 . as a physiological oxidant until recently remained underestimated due to its instability and, therefore, low endogenous levels. Superoxide must also be regarded as a ROS with low reactivity. However, some new findings indicated that O 2 . could be a much more effective and selective oxidant as H 2 O 2 for several key components of intracellular signaling (11,26,27). Comparison of O 2 . and H 2 O 2 as CaN inactivators in our work showed that the former is effective at concentrations at least 3 orders of magnitude lower than the latter. Superoxide sensitivity of CaN is high enough to be of physiological importance and is in agreement with an earlier report of Klee and co-workers (6) on a requirement for Cu,Zn-SOD as a CaN protective agent. Later it was reported that exogenous application of H 2 O 2 caused CaN inhibition and a consequent block of downstream signaling in various cell types (7)(8)(9). In support of these data, we found that treatment of Jurkat cells by submillimolar concentrations of exogenous H 2 O 2 caused inhibition of CaN activity and NFAT1 dephosphorylation and also blocked the activity of a NFATdriven reporter gene. 3  The link between redox regulation of CaN activity and the status of the enzyme binuclear metal center had recently become a matter of intensive discussion (2,17,33). Among the protein serine/threonine phosphatases, the nature of the metals in the binuclear center was clearly defined only for CaN, as iron and zinc (13,14). The question of which oxidation state of iron is present in the native, active form of CaN remained unsolved. This study provides several lines of evidence in support of a binuclear Fe 2ϩ -Zn 2ϩ center in native CaN. First, 3 D. Namgaladze, U. Ruegg, and V. Ullrich, unpublished results. native CaN was inhibited by Fe 3ϩ , whereas Fe 2ϩ had no effect. Second, ascorbate could partly reactivate the O 2 . -inhibited enzyme, and Fe 2ϩ /ascorbate could fully reverse it. The loss of iron after enzyme oxidation could explain the discrepancy between ascorbate and Fe 2ϩ /ascorbate effects. Third, dithionite was not inhibitory when added directly into the assay in concentrations up to 10 mM. It also did not prevent the reactivation by Fe 2ϩ / ascorbate (data not shown). Inhibition of CaN by dithionite provided the basis for assuming an Fe 3ϩ -Zn 2ϩ cluster as the active enzyme form (14,15). We found that under certain conditions, e.g. after aerobic preincubation, dithionite could indeed inhibit CaN activity. This could be partly attributed to radical generation during dithionite oxidation. Furthermore, we found that sulfite in concentrations of 0.1-1 mM also inhibited CaN. Because sulfite is unavoidably generated from dithionite, this factor should be taken into account when considering dithionite effects. The other difference between our data and previously published data is the use of RII versus p-nitrophenyl phosphate as phosphatase substrates. Last, EPR spectroscopic data do not support the presence of ferric iron in native CaN. A weak signal around g ϭ 4.3, corresponding to highly rhombic high spin Fe 3ϩ , was present in the native sample. This feature has been described in previous reports and was attributed either to adventitious iron (14,15) or to iron in a binuclear Fe 3ϩ -Zn 2ϩ center (10). The EPR spectrum of the oxidized sample was in general similar to the previously published EPR spectrum of bovine CaN (10). The two new features centered at g ϭ 6.5 and 5.6 after H 2 O 2 oxidation have been previously assigned to a Fe 3ϩ species with a near axial symmetry (10,15) and most likely originate as primary oxidation products in the binuclear center. The previously described feature at g ϭ 8.1-8.5 only weakly appears in the present spectrum, probably since a higher recording temperature was used, and this signal was reported to rapidly decrease with rising temperature (15).
We therefore agree with the hypothesis first proposed by Qin et al. (34) that the active CaN is likely to have a Fe 2ϩ -Zn 2ϩ metal center. Drawing analogy to the Fe 3ϩ -Zn 2ϩ state of purple acid phosphatases appears to be not valid in this case since their iron ligand sphere contains a tyrosine, which tends to stabilize Fe 3ϩ , whereas CaN contains a histidine, which favors the divalent iron state.
With regard to the mechanism of the O 2 . action one also should consider the Lewis acid properties of the neighboring zinc ion, which can stabilize a peroxo-intermediate arising from the Fe 2ϩ -O 2 . interaction. This additional bridging would allow the ferrous species to acquire an oxidation potential high enough not to react with molecular oxygen. We therefore propose that the resulting species from O 2 . inactivation would be a peroxo-bridged Fe 3ϩ -O 2 2Ϫ -Zn 2ϩ complex, which we now try to identify by magnetic circular dichroism spectrometry.
Our results also provide a link between manganese dependence of CaN activity and O 2 . -mediated inactivation of the enzyme. We suggest that the enzyme in its native state is manganese-independent. During purification, particularly affinity chromatography on calmodulin-Sepharose, the enzyme undergoes oxidative inactivation, which is accompanied by partial iron loss. The addition of ascorbate in enzyme purification buffers allowed the isolation of the reduced enzyme form. Because dithiothreitol was not protective against enzyme inactivation by O 2 . , the use of this agent as reductant in established purification procedures (18) is probably not effective. Some conventional bovine CaN preparations could be activated up to 100-fold by manganese and were not sensitive to superoxide. We observed that after inactivation by O 2 . , Mn 2ϩ could reactivate the enzyme, and this reactivation increased in the pres-ence of the Fe 3ϩ chelator desferrioxamine, thus suggesting that Mn 2ϩ could substitute for iron in the enzyme active center. The enzyme activity in the presence of Mn 2ϩ was still higher than that of the control even for the intact enzyme or for the iron/ ascorbate-reactivated enzyme, probably due to initial inactivation of the enzyme or to the presence of iron-free enzyme. In addition, some zinc loss could also take place. The reasons for such losses are unknown, although a similar O 2 . -dependent zinc release was documented for endothelin-converting enzyme (35). The gain of CaN Mn 2ϩ dependence during purification was noticed as early as in 1982 (36), and the interaction of Mn 2ϩ with CaN was investigated in detail (13,37), but the present study is to our knowledge the first to connect Mn 2ϩ dependence to a previous oxidative inactivation of the enzyme. In summary, the isolation of the redox-sensitive Fe 2ϩ -Zn 2ϩ form of CaN confirms the initial hypothesis of Huang and co-workers (34) that this form represents the native enzyme. Our data provide evidence that native CaN could be the subject of complex redox regulation involving superoxide, calcium, NO, and SOD. Further studies are needed to shed more light onto the mechanism of phosphoester hydrolysis by CaN, which should be readapted to ferrous versus ferric iron in the enzyme binuclear metal center.