Self-subunit Swapping Chaperone Needed for the Maturation of Multimeric Metalloenzyme Nitrile Hydratase by a Subunit Exchange Mechanism Also Carries Out the Oxidation of the Metal Ligand Cysteine Residues and Insertion of Cobalt^,

Abstract

The incorporation of cobalt into low molecular mass nitrile hydratase (L-NHase) of Rhodococcus rhodochrous J1 has been found to depend on the α-subunit exchange between cobalt-free L-NHase (apo-L-NHase lacking oxidized cysteine residues) and its cobalt-containing mediator (holo-NhlAE containing and Cys-SO^- metal ligands), this novel mode of post-translational maturation having been named self-subunit swapping, and NhlE having been recognized as a self-subunit swapping chaperone (Zhou, Z., Hashimoto, Y., Shiraki, K., and Kobayashi, M. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 14849–14854). We discovered here that cobalt was inserted into both the cobalt-free NhlAE (apo-NhlAE) and the cobalt-free α-subunit (apo-α-subunit) in an NhlE-dependent manner in the presence of cobalt and dithiothreitol in vitro. Matrix-assisted laser desorption ionization time-of-flight mass spectroscopy analysis revealed that the non-oxidized cysteine residues in apo-NhlAE were post-translationally oxidized after cobalt insertion. These findings suggested that NhlE has two activities, i.e. cobalt insertion and cysteine oxidation. NhlE not only functions as a self-subunit swapping chaperone but also a metallochaperone that includes a redox function. Cobalt insertion and cysteine oxidation occurred under both aerobic and anaerobic conditions when Co³⁺ was used as a cobalt donor, suggesting that the oxygen atoms in the oxidized cysteines were derived from water molecules but not from dissolved oxygen. Additionally, we isolated apo-NhlAE after the self-subunit swapping event and found that it was recycled for cobalt transfer into L-NHase.

Nitrile hydratase (NHase3; EC 4.2.1.84) (1, 2), which is composed of α- and β-subunits, contains either a non-heme iron (3, 4) or a noncorrin cobalt ion (5–7) in a ligand environment that includes two oxidized cysteine residues (CXLC(SO₂H)SC(SOH)) (8–12). The enzyme catalyzes the hydration of a nitrile to the corresponding amide followed by consecutive reactions: amide → acid → acyl-CoA by amidase (13) and acyl-CoA synthetase (14, 15), respectively. Rhodococcus rhodochrous J1 produces high and low molecular mass NHases (H-NHase and L-NHase), which exhibit different physicochemical properties and substrate specificities (1, 16). In both H- and L-NHase, cobalt acts as an active center for the production of acrylamide and nicotinamide. Acrylamide is manufactured at the industrial level not only in Japan but also in the United States and France (17, 18).

Metalloproteins have been characterized intensively for decades yet only recently have investigators focused on the mechanisms underlying biological metallocenter assembly (19). The synthesis of some metalloproteins has been found to require the participation of accessory proteins (19). An open reading frame, nhlE, which is just downstream of the structural genes of L-NHase (nhlBA), is necessary for functional L-NHase maturation (20). In contrast to the cobalt-containing cysteine-oxidized L-NHase (holo-α₂β₂) derived from nhlBAE, the gene product derived from nhlBA is a non-oxidized cobalt-free apo-L-NHase (20). An L-NHase maturation mediator, NhlAE (encoded by the nhlAE genes, and consisting of αe₂ containing the cobalt-containing cysteine-oxidized α-subunit of L-NHase), has been discovered, and the incorporation of cobalt into L-NHase has been found to depend on the α-subunit exchange between apo-L-NHase and NhlAE. This is a novel post-translational maturation process different from general mechanisms of metallocenter biosynthesis known so far. Thus, we named it self-subunit swapping (Fig. 1(i)) (20). Compared with activators acting as metallochaperones in Fe-NHases (21) from Rhodococcus sp. N-771 (22), Pseudomonas putida 5B (23), Rhodococcus sp. N-774 (24), and so on, NhlE acts as a self-subunit swapping chaperone (Fig. 1(i)), exhibiting novel behavior for a protein in a protein complex (20).

FIGURE 1.

Incorporation of a cobalt ion into l-NHase. (i) Self-subunit swapping maturation of L-NHase. (ii) Insertion of a cobalt ion into apo-NhlAE. The cobalt ion is shown as a closed circle. The holo-α-subunit is shown in gray. Dotted lines with arrows indicate α-subunit swapping.

Metal ions in both Fe-NHase and Co-NHase are located in their α-subunits, which share a characteristic metal binding motif (CXLC(SO₂H)SC(SOH)) containing two oxidized cysteine residues: cysteine-sulfinic acid (αCys-SO₂H) and cysteine-sulfenic acid (αCys-SOH) (8–12). The post-translationally oxidized Cys-SO₂H and Cys-SOH have deprotonated Cys-SO^-₂ and Cys-SO^- structures, respectively (3), and the deprotonated Cys-SO^-₂ and Cys-SO^- in the holo-α-subunit form salt bridges with two arginines of the β-subunit (which are conserved in all known Co-type and Fe-type NHases) in the holoenzyme (8–11). The apoenzyme is likely to be non-oxidized, judging from the results of previous studies on NHase (25) and the related enzyme thiocyanate hydrolase (26–28). Such cysteine oxidation has been demonstrated to occur in cobalt-containing NhlAE (holo-αe₂), but not in cobalt-free NhlAE (apo-αe₂), and the cysteine oxidation plays an essential role in self-subunit swapping maturation (20).

Cysteine oxidation in proteins is receiving increasing attention due to its important physiological properties. Oxidized cysteine residues Cys-SO₂H or Cys-SOH are known to play roles in diverse processes, including signal transduction, oxygen metabolism and the oxidative stress response, transcription regulation, and metal coordination in various proteins such as NADH peroxidase (29, 30), peroxiredoxins (31), hydrogenase (32, 33), and so on (8–12, 26–28, 34). Among these enzymes, NHase and thiocyanate hydrolase are intriguing ones because they possess both Cys-SO₂H and Cys-SOH as ligands of the metal center and neither residue plays any catalytic redox role at all (29). Although the oxidized cysteine residues in both enzymes are essential for their activities (20, 34), the mechanism(s) involved in the post-translational oxidation of both residues remains unclear.

Non-corrin cobalt enzymes like NHase are receiving increasing interest not only in bioinorganic chemistry but also in biotechnology, and their availability and remarkable chemical versatility make them invaluable catalysts in the chemical industry (6, 35–37). Although we have discovered a cobalt transporter (NhlF) that mediates cobalt uptake into the cell (38), the mechanism underlying the insertion of a cobalt into a noncorrin-cobalt-containing protein in the cell has never been clarified. In this study we discovered that cobalt was directly inserted into a non-oxidized apo-α-subunit on the addition of NhlE, which yielded a cobalt-containing cysteine-oxidized NhlAE in vitro, suggesting that NhlE acts as a metallochaperone that is crucial for both post-translational cysteine oxidation and cobalt incorporation into the α-subunit of L-NHase (Fig. 1(ii)).

EXPERIMENTAL PROCEDURES

Bacterial Strain and Plasmid Used—Rhodococcus fascians DSM43985 was used as the host for vector plasmid pREIT19, which was used for nhlBAE, nhlAE, and nhlBA-(α-A3G) expression as described previously (20).

Culture Conditions—The R. fascians DSM43985 transformants carrying pREIT-nhlAE for holo-αe₂ expression were grown at 28 °C for 72 h in 2YT medium containing CoCl₂·6H₂O (0.1 g/liter) and kanamycin (50 μg/ml), and 0.1% (v/v) of isovaleronitrile, as an inducer, was added to the medium after incubation for 12 h. R. fascians DSM43985 transformants carrying pREIT-nhlBAE were grown under the same conditions for 96 h except that the inducer was continuously added every 24 h for a total of 4 times to increase the amount of L-NHase expressed. The R. fascians DSM43985 transformants carrying pREIT-nhlAE, pREIT-nhlBAE, and pREIT-nhlBA-(α-A3G) used for apoprotein expression were incubated under the same conditions except that CoCl₂·6H₂O was not added to the medium.

Protein Purification—Proteins were purified in the same manner as that described previously (20) except that 10 mm potassium phosphate buffer (KPB) (pH 7.5) containing 0.5 mm dithiothreitol (DTT) was used in all purification steps. No influence on L-NHase activity was observed under these purification conditions. The apo-L-NHases purified under these conditions were mostly of the heterotetramer form (apo-α₂β₂), and only little heterodimer forms (apo-αβ) were detected. Thus, the purified apo-α₂β₂ was used as the apo-L-NHase in this study.

In Vitro Activation of Apo-α₂β₂—The routine apo-α₂β₂ activation buffer consisted of 10 mm KPB (pH 7.5), 10 μm cobalt, and 2 mm DTT unless otherwise noted. The final concentrations of apo-α₂β₂ and apo-αe₂ in the activation buffer were 0.1 and 0.4 mg/ml, respectively. A CoCl₂ solution was used as a cobalt donor unless otherwise noted. The mixtures were incubated at 28 °C.

Purification of Resultant Proteins—Hiload 16/60 Superdex 200-pg (GE Healthcare) and Resource Q (6 ml) (GE Healthcare) columns were used to purify the resultant L-NHases and NhlAEs from mixtures in the same way as described previously (20) except that 10 mm KPB (pH 7.5) containing 0.5 mm DTT was used.

Enzyme Assay—L-NHase activity was assayed in a reaction mixture (0.5 ml) comprising 10 mm KPB (pH 7.5), 20 mm 3-cyanopyridine, and 10 μl of enzyme containing activation buffer or an appropriate amount of the enzyme. The reaction was carried out at 20 °C for 20 min and stopped by the addition of 0.5 ml of acetonitrile. The amount of nicotinamide formed in the reaction mixture was determined as described previously (20). One unit of L-NHase activity was defined as the amount of enzyme that catalyzed the release of 1 μmol of nicotinamide per min at 20 °C.

Denaturation of NhlAE and Renaturation of the α-Subunit and NhlE—The denaturation and renaturation experiments were carried out at room temperature in 20 mm Tris-HCl buffer (pH 7.5). SDS was carefully added to NhlAE, which had been mixed with 2-mercaptoethanol. The concentrations of resultant apo-αe₂ (R-apo-αe₂) and SDS were both 0.2 mg/ml, and that of 2-mercaptoethanol was 10% (v/v) (final, respectively). After 2 h, the α-subunit and NhlE were purified by gel filtration on a Hiload 16/60 Superdex 200-pg column equilibrated with 20 mm Tris-HCl buffer (pH 7.5) containing 1% SDS. After dialysis, the SDS remaining with the α-subunit was eliminated with a Resource Q column (6 ml) equilibrated with 20 mm Tris-HCl buffer (pH 7.5), and the protein was eluted with 0.5 liters of the buffer by increasing the concentration of KCl linearly from 0 to 0.5 m. This step was carried out twice to isolate the α-subunit. These two procedures were carried out with an AKTA purifier (GE Healthcare). Gel filtration chromatograph analysis was performed for molecular mass determination, the flow rate being 0.5 ml/min. A Superose 12 HR 10/30 column (GE Healthcare) and 0.2 m KCl-containing buffer were used for estimation of the molecular masses of the purified α-subunits and NhlE.

Matrix-assisted Laser Desorption Ionization Time-of-flight Mass Spectrometry (MALDI-TOF MS) Sequencing—MALDI-TOF MS was typically performed using a Voyager-DE STR mass spectrometer (Applied Biosystems) in the linear positive mode. The purified α-subunit derived from R-apo-αe₂ was reduced with 1 mm DTT and carboxamidomethylated with 2.1 mm iodoacetamide and then diluted with 3 volumes of 50 mm Tris-HCl (pH 8.0) and treated with trypsin at 35 °C for 20 h. The details were given previously (20).

Other Analytical Methods—CD measurement, UV-Visible Measurement, and Cobalt Ion Determination Were Described Previously (20).

RESULTS

Post-translational Activation of Apo-L-NHase by Apo-αe₂ in the Presence of Cobalt in Vitro—Previous work (20) has shown that apo-L-NHase is converted into holo-α₂β₂ after mixing with holo-αe₂ through self-subunit swapping maturation. To determine whether cobalt could be directly incorporated into apo-L-NHase to yield holo-L-NHase, we mixed apo-L-NHase (apo-α₂β₂, see “Experimental Procedures”) with cobalt (20 μm, final) in 10 mm KPB (pH 7.5) and then measured the L-NHase activity in the mixture. As shown in Fig. 2, no significant increase in L-NHase activity was observed. However, we found that L-NHase activity increased when apo-αe₂ was mixed in additionally and that the L-NHase activity significantly increased when 0.5 mm DTT was further added to this mixture (Fig. 2). These findings indicated that apo-α₂β₂ was post-translationally activated by apo-αe₂ in the presence of cobalt and that the post-translational activation was enhanced by the addition of DTT in vitro. Subsequently, the effect of the DTT concentration on activation of apo-α₂β₂ by apo-αe₂ in the presence of cobalt was investigated. Mixtures were incubated with various concentrations of DTT, and then the L-NHase activity in each activation mixture at 12 h was measured. The L-NHase activity in the activation mixtures became higher as the DTT concentration increased, but more than 4 mm DTT caused a decrease in activation competence (see Fig. 3A and “Discussion”). These findings suggest that a certain amount of DTT is necessary for activation of apo-α₂β₂ by apo-αe₂ in the presence of cobalt and that the suitable concentration of DTT is 2 mm. Thereafter, the effect of the cobalt concentration on activation of apo-α₂β₂ was investigated with this suitable DTT concentration. The L-NHase activity in the activation mixtures reached a plateau with 10 μm cobalt added (Fig. 3B). Thus, an activation mixture containing 2 mm DTT and 10 μm cobalt provided the optimal conditions and, thus, was used in the following experiments.

FIGURE 2.

Apo-αe₂-dependent activation of apo-α₂β₂. Apo-α₂β₂ was incubated in 10 mm KPB (pH 7.5) containing 20 μm CoCl₂ and apo-αe₂, apo-αe₂ plus 0.5 mm DTT, or 0.5 mm DTT. Aliquots of the samples were removed at the indicated times and then assayed for L-NHase activity. Values represent the means ± S.D. for at least triplicate independent experiments.

FIGURE 3.

Activation of apo-α₂β₂ by apo-αe₂ under various conditions in vitro. A, effect of the DTT concentration on apo-α₂β₂ activation. Apo-α₂β₂ was mixed with apo-αe₂ in the activation buffer containing 20 μm CoCl₂ and 0, 0.5, 1, 2, 4, 6, 8 or 10 mm DTT. B, effect of the cobalt concentration on apo-α₂β₂ activation. Apo-α₂β₂ was mixed with apo-αe₂ in the activation buffer containing 2 mm DTT and 0, 5, 10, 20, 40, 60, or 80 μm CoCl₂. Aliquots of the samples were removed at 12 h and then assayed for L-NHase activity. Values represent the means ± S.D. for at least triplicate independent experiments.

Direct Incorporation of Cobalt into Apo-αe₂—Site-directed mutagenesis and N-terminal amino acid sequence analyses were performed to confirm the source of the α-subunit in the active L-NHase derived from the mixture of apo-α₂β₂, apo-αe₂, cobalt, and DTT. As a result, the source of the α-subunit in the active L-NHase was found to be the apo-αe₂ (supplemental Fig. S1 and Table S1), demonstrating that the activation of apo-α₂β₂ was dependent on self-subunit swapping. We have already demonstrated that self-subunit swapping occurs between apo-α₂β₂ and holo-αe₂ but not between apo-α₂β₂ and apo-αe₂ (20). Considering these findings, we speculated that cobalt should be first inserted into the cobalt-free α-subunit (apo-α-subunit) of apo-αe₂ to yield the cobalt-containing α-subunit (holo-α-subunit), resulting in the formation of holo-αe₂, and then self-subunit swapping occurs between apo-α₂β₂ and this holo-αe₂ to yield holo-α₂β₂. To test this hypothesis, we mixed apo-αe₂ with cobalt and DTT in the absence of apo-α₂β₂. The R-apo-αe₂ was purified after incubation for 4 h (Fig. 4A) and then compared with apo-αe₂ and holo-αe₂. The results of cobalt content determination showed that R-apo-αe₂ contained 0.98 ± 0.08 mol ion/mol of αe₂, similar to the content of holo-αe₂ (Table 1). Although R-apo-αe₂, holo-αe₂, and apo-αe₂ showed similar CD spectra (Fig. 4B), the UV-visible spectrum of R-apo-αe₂ was similar to that of holo-αe₂ but not to that of apo-αe₂ (Fig. 4C). It has been reported that the absorption in the 300–350-nm region for Co-NHase reflects S → Co³⁺ charge transfer (7, 20). As shown in Fig. 4C, an extra shoulder in the 300–350-nm region was found for R-apo-αe₂ as well as holo-αe₂, suggesting that the cobalt-ligand environment of R-apo-αe₂ is the same as those of holo-αe₂ and Co-NHases. On the other hand, we also mixed apo-α₂β₂ with cobalt and DTT in the absence of apo-αe₂. The resultant apo-α₂β₂ (R-apo-α₂β₂) was purified after incubation for 4 h and then compared with apo-α₂β₂ and holo-α₂β₂. The UV-visible spectrum of R-apo-α₂β₂ was similar to that of apo-α₂β₂, but not to that of holo-α₂β₂, and no extra shoulder in the 300–350-nm region was observed (supplemental Fig. S1). In contrast to the cobalt content of holo-α₂β₂ (0.88 ± 0.03 mol/mol of αβ), only 0.16 ± 0.03 mol cobalt ion/mol of αβ was detected in R-apo-α₂β₂ (Table 1). These findings demonstrated that a cobalt ion was directly incorporated into the apo-α-subunit of apo-αe₂ (Fig. 1 (ii)) but not into that of apo-α₂β₂ in vitro.

FIGURE 4.

Characterization of purified holo-αe₂, apo-αe₂, R-apo-αe₂, and R-(α+e₂). SDS/PAGE (A), far-UV CD (B), and UV-visible (C) spectra of the purified holo-αe₂, apo-αe₂, R-apo-αe₂ and R-(α+e₂).

TABLE 1.

Characterization of the purified NhlAEs, L-NHases, α-subunits, R-NhlAEs, R-L-NHases, and R-α-subunit Values represent the means ± S.D. for at least triplicate independent experiments.

Protein	Cobalt content	Specific activity
	mol of ions/mol of protein	units/mg
Apo-αe2	0.02 ± 0.01/αe2	0
R-apo-αe2	0.98 ± 0.08/αe2	0
Holo-αe2	0.85 ± 0.03/αe2	0
Apo-α2β2	0.02 ± 0.01/αβ	4.16 ± 0.42
R-apo-α2β2	0.16 ± 0.03/αβ	20.6 ± 3.4
Holo-α2β2	0.88 ± 0.03/αβ	345 ± 12
Holo-α	0.84 ± 0.03/α	0
Apo-α	0.02 ± 0.01/α	0
R-apo-α	0.14 ± 0.03/α	0
R-(α+e2)	0.90 ± 0.05/αe2	0
R-holo-αe2	0.05 ± 0.02/αe2	0
R-apo-α2β2a	0.95 ± 0.07/αβ	270 ± 10

^aThe resultant L-NHase purified from the mixture of R-holo-αe₂, apo-α₂β₂, cobalt, and DTT

Cysteine Oxidation in R-apo-αe₂—We discovered here that cobalt was directly inserted into apo-αe₂ (Fig. 4 and Table 1). This finding permitted us to speculate that the cysteine residues could be oxidized in R-apo-αe₂. We analyzed R-apo-αe₂ by MALDI-TOF MS. During such analysis cysteine residues and Cys-114-SO^- (if present) would be reduced and carboxamidomethylated (CAM) as well as disulfide-bonded cysteine residues. The α-subunit (R-apo-α-subunit) isolated from R-apo-αe₂ was treated with trypsin after reduction and carboxamidomethylation. The molecular mass of the tryptic peptide containing all metal ligand residues, ⁸³EMGVGGMQGEEMVVLENTGTVHNMVVCTLCSCYPWPVLGLPPNWYK¹²⁸ (EK46), was measured. The MALDI-TOF MS spectrum of EK46 (Fig. 5) of the R-apo-α-subunit was compared with those of the apo-α-subunit isolated from apo-αe₂ and the holo-α-subunit isolated from holo-αe₂ (20). Considering the mass peak with an m/z of 5242.4 (Fig. 5, inset) corresponding to the calculated m/z value of the [M+H]⁺ ion of EK46 with three CAM-cysteines and the mass peak with an m/z of 5217.9 (Fig. 5, inset) corresponding to the calculated m/z value of the [M+H]⁺ ion of EK46 with two CAM-cysteines and one Cys-SO₂H, Cys-112-SO^-₂ was suggested to exist in holo-αe₂ but not in apo-αe₂ (20). In the mass spectrum of EK46 of the R-apo-α-subunit, the magnitude of the m/z 5242 peak corresponding to EK46 with three CAM-cysteines showed a dramatic decrease, and an intense peak at m/z 5218 corresponding to EK46 with two CAM-cysteines and one Cys-SO₂H was observed (Fig. 5), suggesting that Cys-112 in apo-αe₂ was oxidized to Cys-112-SO^-₂ in R-apo-αe₂. Although the occurrence of Cys-114-SOH oxidation has not been confirmed because of the chemical instability (34), this finding strongly suggests that oxidized cysteine residues (αCys-112-SO^-₂ and αCys-114-SO^-) exist in R-apo-αe₂.

FIGURE 5.

MALDI-TOF MS spectra of the metal-binding peptide, EK46, of R-apo-αe₂. The mass peaks with an m/z value of 5242 correspond to the [M+H]⁺ ion of EK46 with three CAM-cysteines (calculated m/z value = 5243.1); the mass peaks with m/z values of 5218 correspond to the [M+H]⁺ ion of EK46 with one Cys-SO₂H and two CAM-cysteines (calculated m/z value = 5218.1). The inset shows the MALDI-TOF MS spectra of the metal-binding peptide, EK46, of apo-αe₂ (top) and holo-αe₂ (bottom) (20).

Necessity of NhlE for Cobalt Incorporation into the α-Subunit—To determine whether cobalt could be directly inserted into the apo-α-subunit in the absence of NhlE or not, we separated the α-subunit and NhlE through denaturation and renaturation (20). The resultant α-subunits (holo-α and apo-α, derived from holo-αe₂ and apo-αe₂, respectively) were found to each be a monomer, respectively, whereas the resultant NhlE was found to be a dimer (e₂) (Fig. 6A). We mixed the apo-α-subunit with cobalt and DTT as described above and then purified the R-apo-α after incubation for 4 h. As a result, the cobalt content (0.14 ± 0.03/α) and the UV-visible spectrum of R-apo-α were found to be similar to those of apo-α but not to those of holo-α (Table 1 and Fig. 6B). On the other hand, from the mixture containing apo-α and e₂ (molar ratio of apo-α to e₂, 1:1) in the presence of cobalt and DTT after incubation for 4 h, cobalt-containing NhlAE (R-(α+e₂)) was successfully isolated. The cobalt content (0.90 ± 0.05/αe₂) and CD and UV-visible spectra of R-(α+e₂) are similar to those of holo-αe₂ (Table 1 and Fig. 4, B and C), and R-(α+e₂) was also able to convert apo-L-NHase to holo-L-NHase (data not shown). These findings demonstrated that NlhE is necessary for the conversion of the apo-α-subunit into the holo-α-subunit.

FIGURE 6.

Characterization of purified holo-α, apo-α, NhlE, and R-apo-α. A, molecular mass and structure determination of the α-subunits and NhlE derived from apo-αe₂ and holo-αe₂ by gel filtration on a Superose 12 HR 10/30 column. Each of nhlA and nhlE (DDBJ accession numbers X64360 (nhlA) and D83695 (nhlE)) would encode proteins of 207 amino acids (22.8 kDa) and 148 amino acids (16.9 kDa), respectively. Marker proteins: i, glutamate dehydrogenase (yeast) (290 kDa); ii, lactate dehydrogenase (pig heart) (142 kDa); iii, enolase (yeast) (67 kDa); iv, myokinase (yeast) (32 kDa); v, cytochrome c (horse heart) (12.4 kDa). Values represent the means ± S.D. for at least triplicate independent experiments. B, UV-visible spectra of the purified holo-α, apo-α, and R-apo-α.

Effects of Aerobic and Anaerobic Conditions on Activation of Apo-α₂β₂ by Apo-αe₂ in the Presence of Co²⁺ and Co³⁺—To confirm the effect of dissolved oxygen on the post-translational activation of apo-α₂β₂ by apo-αe₂ in the presence of cobalt and DTT, the same experiment was carried out under anaerobic conditions, for which most of the dissolved oxygen was removed with argon. Although apo-α₂β₂ was activated by apo-αe₂ in the presence of cobalt and DTT under aerobic conditions (245 ± 6 units/mg), it was hardly activated under anaerobic conditions (48.3 ± 4.5 units/mg) (Table 2). On the other hand, apo-α₂β₂ activation by holo-αe₂ through self-subunit swapping under anaerobic conditions was similar to that under aerobic conditions, suggesting that self-subunit swapping between apo-α₂β₂ and holo-αe₂ occurred irrespective of the presence or not of dissolved oxygen. Subsequently, [Co(NH₃)₆]Cl₃(Co³⁺ donor) was used instead of CoCl₂ (Co²⁺ donor) to investigate the effect of the cobalt redox state on activation of apo-α₂β₂ by apo-αe₂ in the presence of DTT under aerobic and anaerobic conditions. As shown in Table 2, activation of apo-α₂β₂ by apo-αe₂ with Co³⁺ was observed under both aerobic and anaerobic conditions and was similar to that with Co²⁺ under aerobic conditions.

TABLE 2.

Effects of aerobic and anaerobic conditions on activation of apo-α₂β₂ by apo-αe₂ and holo-αe₂ The mixtures were incubated in KPB (pH 7.5) at 28°C under aerobic or anaerobic conditions. [Co(NH₃)₆]Cl₃ and CoCl₂ solutions were used as Co³⁺ and Co²⁺ donors, respectively. The concentrations of apo-α₂β₂ and holo-αe₂ in the mixture were 0.1 and 0.4 mg/ml, respectively. Aliquots of the samples were removed at 12 h and then assayed for L-NHase activity. Values represent the means ± S.D. for at least triplicate independent experiments.

Mixture	Conditions	L-NHase activity
		units/mg
Apo-α2β2 + apo-αe2 + Co2+ + DTT	Aerobic	245 ± 6
Apo-α2β2 + apo-αe2 + Co2+ + DTT	Anaerobic	48.3 ± 4.5
Apo-α2β2 + holo-αe2	Aerobic	335 ± 12
Apo-α2β2 + holo-αe2	Anaerobic	330 ± 8
Apo-α2β2 + apo-αe2 + Co3+ + DTT	Aerobic	246 ± 10
Apo-α2β2 + apo-αe2 + Co3+ + DTT	Anaerobic	243 ± 8

Recycling of Used NhlAE—After holo-αe₂ post-translationally activates apo-α₂β₂ through self-subunit swapping, during which the holo-α-subunit of holo-αe₂ and the apo-α-subunit of apo-α₂β₂ are exchanged, cobalt-containing and cysteine-oxidized holo-α₂β₂ and non-cysteine-oxidized cobalt-free αe₂ (R-holo-αe₂) are formed (20). Here, we have the following question: Can R-holo-αe₂ be recycled for cobalt insertion into L-NHase? When apo-α₂β₂ has been changed into holo-α₂β₂ completely after incubation with more than 2-fold of holo-αe₂, the αe₂ purified from the mixture after self-subunit swapping under these conditions comprises a mixture of apo-αe₂ (used holo-αe₂: R-holo-αe₂) and unused holo-αe₂ (20). To isolate only R-holo-αe₂, we mixed holo-αe₂ with apo-α₂β₂ in the ratio of 1:4 (0.1 and 0.4 mg/ml, respectively) in 10 mm KPB (pH 7.5) and then purified the resulting αe₂ in the mixture after 12 h of incubation at 28 °C (Fig. 7A). The results of cobalt content determination (0.05 ± 0.02 mol/mol αe₂) (Table 1) and UV-visible spectrum analysis (supplemental Fig. S2) indicated that almost all of the purified αe₂ consisted of R-holo-αe₂. Subsequently, post-translational activation of apo-α₂β₂ by R-holo-αe₂ was investigated. As shown in Fig. 7B, activation of apo-α₂β₂ was not observed in a mixture of apo-α₂β₂ and R-holo-αe₂, this being consistent with that self-subunit swapping does not occur between apo-α₂β₂ and apo-αe₂ (20). In the presence of cobalt and DTT, however, R-holo-αe₂ increased the L-NHase activity of apo-α₂β₂ to a level similar to that in the case of apo-αe₂ (Fig. 7B). The enzyme activity and cobalt content (270 ± 10 units/mg and 0.95 ± 0.07/αβ) of the resultant L-NHase purified from the mixture of R-holo-αe₂, apo-α₂β₂, cobalt, and DTT after 12 h were almost identical to those (275 ± 8 units/mg and 0.96 ± 0.07/αβ) with apo-αe₂ instead of R-holo-αe₂ (Table 1), demonstrating that almost all of the apo-α₂β₂ was converted to holo-α₂β₂. These findings suggested that NhlAE once used would be recycled for post-translational maturation of L-NHase.

FIGURE 7.

Recycling of NhlAE once used for cobalt incorporation into l-NHase. A, SDS/PAGE of the purified R-holo-αe₂ (lane 1) and a mixture of apo-α₂β₂ with holo-αe₂ in KPB (pH 7.5) at 12 h (lane 2) and 0 h (lane 3). B, activation of apo-α₂β₂ by R-holo-αe₂ and apo-αe₂. Aliquots of the samples were removed at the indicated times and then assayed for L-NHase activity. Values represent the means ± S.D. for at least triplicate independent experiments.

DISCUSSION

Reactive cysteine residues in proteins are well known to be critical components in redox signaling (29). Cys-SOH oxidation has been found in various redox sensor and regulating proteins (29–31). Depending on the environment, Cys-SOH in proteins sometimes provides a metastable oxidized form and at other times it is a fleeting intermediate giving rise to more stable disulfide, Cys-SO₂H, and sulfenyl-amide forms (29). Cys-SOH and Cys-SO₂H serve as catalytically essential redox centers, transient intermediates during peroxide reduction, sensors for the intracellular redox status, catalytic active sites, etc. (29, 30, 32, 33). Compared with various proteins containing reactive cysteine residues, NHase and thiocyanate hydrolase have both post-translational oxidized Cys-SOH and Cys-SO₂H in their active sites (8–12, 26–28); both cysteines play an important role in the catalytic reaction but do not play any catalytic redox role (29). Several research groups have presented chemical and crystallographic evidence that Cys-SOH and Cys-SO₂H are essential for Fe- and Co-NHases and thiocyanate hydrolase activity (8–12, 25, 26, 35). However, an enzyme catalyzing the cysteine oxidation reaction has never been reported. We discovered here that the non-oxidized cysteine residues in the apo-α-subunit of apo-αe₂ are oxidized during cobalt insertion (Fig. 5). Once the cobalt is incorporated into NhlAE, it can never be removed through dialysis, even during NhlAE denaturation to isolate the holo-α-subunit (20). These findings suggest that the insertion of cobalt into the apo-α-subunit of αe₂ is associated with cysteine oxidation; cobalt coordination is essential for the cysteine oxidation, and cysteine oxidation would increase the metal binding affinity. Taking the above together with the findings that the α-subunit in apo-L-NHase and the α-subunit itself were recalcitrant as to direct binding of cobalt (Table 1), we demonstrated here that αe₂ is a vital complex for cobalt incorporation into the α-subunit of L-NHase and that NhlE is involved in the oxidation of cysteine ligands in the α-subunit during αe₂ coordination with cobalt. To the best of our knowledge there have been no reports on protein-dependent post-translational cysteine oxidation in vitro. Although the Cys-SO₂H in peroxiredoxine has been reported to be a reversible intermediate that is reduced to Cys-SOH through sulfiredoxin (39), this redox reaction occurs through a protein repair system rather than a protein-dependent post-translational maturation system. We discovered that NhlE facilitates Cys-SOH and Cys-SO₂H oxidation post-translationally and for the first time succeeded in making cysteine residues in a protein undergo oxidation to stable Cys-SOH and Cys-SO₂H in a protein-dependent manner in vitro, whereas enhancement of the spontaneous oxidation of cysteine residues in the α-subunit of Fe-NHase in Rhodococcus sp. N-771 in the presence of ferric citrate in vitro has been observed (34).

Cobalt exists as a non-corrin low-spin Co³⁺ ion in H-NHase in R. rhodochrous J1 (40) and NHase in P. putida NRRL-18668 (7), suggesting that the holo-α-subunit of L-NHase (holo-α₂β₂) and the holo-α-subunit of holo-αe₂ (which is the source of that of holo-α₂β₂) should also have a non-corrin low-spin Co³⁺ ion. Therefore, the post-translational oxidation of the cysteine residues was associated with the oxidation of Co²⁺ to Co³⁺, because Co²⁺ was used as the cobalt donor in the in vitro experiments. Two possible sources of the oxygen atoms in the oxidized cysteine residues are considered: dissolved oxygen and a water molecule. Apo-α₂β₂ was activated by apo-αe₂ in the presence of Co³⁺ and DTT under both aerobic and anaerobic conditions (Table 2), suggesting that the cysteines were oxidized irrespective of the presence of dissolved oxygen. On the other hand, apo-α₂β₂ was activated by apo-αe₂ in the presence or not of Co²⁺ and DTT under aerobic conditions but not under anaerobic conditions (Table 2), demonstrating that dissolved oxygen was necessary for the oxidation of Co²⁺ to Co³⁺. Considering these findings, we would like to propose that the oxygen atoms in the oxidized cysteine residues are derived from water molecules, although the oxygen atoms in the cysteine residues modified through oxidation using an inhibitor (2-cyano-2-propyl hydroperoxide) for Fe-NHase and spontaneous oxidation are proposed to come from water molecules (41) and dissolved oxygen (34), respectively.

The crystal structure of the apo-NHase in Pseudonocardia thermophila JCM3059 has shown that the oxidation of cysteine residues in the active center does not occur, but electron density connecting Cys-108 and Cys-113 of the α-subunit was observed, indicating the formation of a disulfide bond between the two residues (25). Considering this finding, we speculate that a disulfide bond should also exist between the corresponding Cys-109 and Cys-114 in the α-subunit of apo-αe₂ and that the cobalt incorporation would be blocked by the disulfide bond. An unusually strong reducing agent, DTT, because of its high conformational propensity to form a six-member ring with an internal disulfide bond, would act as an antioxidant for the reduction of the disulfide bond between Cys-109 and Cys-114, and the reduced forms of the resultant cysteine residues might be a necessary state at the beginning of cobalt insertion into apo-αe₂. This hypothesis is supported by the finding that L-NHase activity was significantly increased in the presence of two other reducing agents, mercaptoethanol and glutathione, instead of DTT (data not shown). Because we were not able to quantify the holo-αe₂-derived from apo-αe₂ in the presence of cobalt and DTT, L-NHase activity was measured in place of the holo-αe₂ formation after the conversion (Fig. 1(ii)) coupled with self-subunit swapping (Fig. 1(i)). A low concentration (<2 mm) of DTT would not be able to reduce all of the disulfide bonds completely. On the other hand, a high concentration (more than 4 mm) of DTT would inhibit L-NHase activity during the L-NHase assay. In the cases of 4, 6, and 8 mm DTT used for activation of apo-α₂β₂, the final DTT concentrations in the L-NHase assay buffer were 0.08, 0.12, and 0.16 mm, respectively. In fact, the L-NHase activity was assayed using holo-α₂β₂ in the presence of DTT, the following levels being exhibited: 0.05 mm, 92%; 0.1 mm, 85%. These results strongly support inhibition of the L-NHase activity by a high concentration of DTT. In addition, a high concentration of DTT might also inhibit the cysteine residue oxidation.

The αe₂ complex once used was demonstrated to be recycled in vitro during post-translational maturation of L-NHase (Fig. 7 and Table 1). If the αe₂ complex is recycled in vivo, the recycling system involved would allow sufficient maturation of the L-NHase even in the case of a small amount of αe₂. In fact, the SDS/PAGE patterns of cell-free extracts in the previous study (20) showed that the amount of NhlAE is lower than that of L-NHase and that the molar ratio of the purified holo-αe₂ to holo-α₂β₂ is about 2:3 (data not shown) when nhlBAE (containing both the L-NHase and NhlAE genes) is used for L-NHase and NhlAE expression. In addition, the self-subunit swapping kinetics appear not to be fast; however, the rate of activation of apo-α₂β₂ by holo-αe₂ becomes faster as the holo-αe₂ concentration increases (20). The recycling of the used NhlAE for cobalt transfer would provide a sufficient quantity of holo-αe₂ for activation of apo-α₂β₂ in vivo, resulting in efficient self-subunit swapping maturation.

Because cobalt coordination and oxidation of cysteine residues, both of which are essential for L-NHase activation (1, 20), occur in NlhAE but not in apo-L-NHase, NhlAE is a vital complex in the process of functional L-NHase biogenesis. The structural genes for the α- and β-subunits of L-NHase are on the order of β and α; the NhlE gene (nhlE) is located just downstream that of the α-subunit gene (nhlA) of L-NHase, and NhlE lacks any known metal binding motifs but exhibits significant sequence similarity to the β-subunit of L-NHase (27% identity). This gene organization of L-NHase and NhlE allows NhlE to easily form a complex with the apo-α-subunit after translation.

In R. rhodochrous J1, a cobalt transporter (NhlF) (38) and Mg²⁺ uptake systems would be involved in the uptake of cobalt into cells (1). Taking the above together with the results described in this study, we propose a model for the overall process of cobalt incorporation into this noncorrin-cobalt-containing enzyme, L-NHase. As shown in Fig. 8, cobalt is transported into R. rhodochrous J1 cells by NhlF and Mg²⁺ uptake systems and then inserted into apo-αe₂, resulting in cobalt- and oxidized cysteine-containing holo-αe₂ (Fig. 1(ii)) followed by the formation of holo-α₂β₂ through replacement of the cobalt-free and non-oxidized α-subunit in apo-α₂β₂ with the cobalt-containing and oxidized cysteine-containing α-subunit in holo-αe₂ (Fig. 1(i)). During the biogenesis of L-NHase, αe₂ could be recycled for further biogenesis of L-NHase. In the present study, for the first time we demonstrated that NhlE acts not only as a chaperone for self-subunit swapping (Fig. 1(i)) but also as a metallochaperone that is crucial for both cobalt insertion and post-translational cysteine oxidation (Fig. 1(ii)). The discovery of this novel function of NhlE will help us to clarify the mechanism underlying this post-translational cysteine oxidation.

FIGURE 8.

A proposed model for the process of cobalt incorporation into l-NHase. Mg²⁺ uptake systems are designated as Mg²⁺ uptake systems (which transport cations such as Co²⁺, Zn²⁺, and Mg²⁺ into the cell). The cobalt ion is shown as a closed circle. MUS, Mg²⁺ uptake system.