Involvement of a mate chaperone (TorD) in the maturation pathway of molybdoenzyme TorA.

As many prokaryotic molybdoenzymes, the trimethylamine oxide reductase (TorA) of Escherichia coli requires the insertion of a bis(molybdopterin guanine dinucleotide)molybdenum cofactor in its catalytic site to be active and translocated to the periplasm. We show in vitro that the purified apo form of TorA was activated weakly when an appropriate bis(molybdopterin guanine dinucleotide)molybdenum source was provided, whereas addition of the TorD chaperone increased apoTorA activation up to 4-fold, allowing maturation of most of the apoprotein. We demonstrate that TorD alone is sufficient for the efficient activation of apoTorA by performing a minimal in vitro assay containing only the components for the cofactor synthesis, apoTorA and TorD. Interestingly, incubation of apoTorA with TorD before cofactor addition led to a significant increase of apoTorA activation, suggesting that TorD acts on apoTorA before cofactor insertion. This result is consistent with the fact that TorD binds to apoTorA and probably modifies its conformation in the absence of cofactor. Therefore, we propose that TorD is involved in the first step of TorA maturation to make it competent to receive the cofactor.

As many prokaryotic molybdoenzymes, the trimethylamine oxide reductase (TorA) of Escherichia coli requires the insertion of a bis(molybdopterin guanine dinucleotide)molybdenum cofactor in its catalytic site to be active and translocated to the periplasm. We show in vitro that the purified apo form of TorA was activated weakly when an appropriate bis(molybdopterin guanine dinucleotide)molybdenum source was provided, whereas addition of the TorD chaperone increased apo-TorA activation up to 4-fold, allowing maturation of most of the apoprotein. We demonstrate that TorD alone is sufficient for the efficient activation of apoTorA by performing a minimal in vitro assay containing only the components for the cofactor synthesis, apoTorA and TorD. Interestingly, incubation of apoTorA with TorD before cofactor addition led to a significant increase of apoTorA activation, suggesting that TorD acts on apo-TorA before cofactor insertion. This result is consistent with the fact that TorD binds to apoTorA and probably modifies its conformation in the absence of cofactor. Therefore, we propose that TorD is involved in the first step of TorA maturation to make it competent to receive the cofactor.
The molybdenum cofactor is a ubiquitous molecule associated with a wide range of redox enzymes and is found in most organisms from bacteria to humans (1,2). Except for nitrogenase, the metal in molybdoenzymes is coordinated to a pterin derivative called molybdopterin (MPT) 1 to form the MPT-Mo cofactor (3). In bacteria, the basic form of the molybdenum cofactor is generally modified by the attachment of a nucleotide moiety to the phosphate group of MPT-Mo. In many cases, the final step of the cofactor synthesis is the linkage of GMP to MPT-Mo, giving rise to molybdopterin guanine dinucleotide (MGD-Mo) (3). In Escherichia coli, it is now clearly established that the conversion of MPT-Mo to MGD-Mo is catalyzed by the MobA protein and can occur in vitro in the presence of GTP and MgCl 2 , even in the absence of any apoenzyme (4,5). The determination of the crystal structure of several MGD-containing enzymes has revealed that the molybdenum atom is in fact coordinated to two MGD molecules, leading to the bis(MGD)Mo cofactor (6 -9).
The molybdoenzymes have been classified according to the composition of their molybdenum cofactor (1). Among these families, the Me 2 SO reductase family is made up of bis(MGD)-Mo-containing enzymes that share a conserved overall tertiary structure (6 -10). They are organized in four domains and form compact molecules with a deep funnel-like depression on one side where the bis(MGD)Mo is buried. The cofactor is associated to the protein by a single covalent linkage to a serine residue and a network of hydrogen bonds. The insertion of the molybdenum cofactor is a cytoplasmic event required for the translocation of periplasmic molybdoproteins. As with a variety of periplasmic redox enzymes, their leader sequence exhibits the twin arginine consensus motif (RRXFLK) specifically recognized by the twin arginine translocation system, exporting metalloenzymes through the membrane in a folded conformation (11,12). Consequently, in a molybdenum cofactor-deficient context, the apoenzymes remain in the cytoplasm of the bacteria where they can be subsequently degraded (13).
In E. coli, apart from biotin sulfoxide reductase BisC, molybdoenzymes are involved in anaerobiosis respiration processes. Among these respiratory pathways, the Tor system, leading to the specific reduction of trimethylamine oxide (TMAO), is one of the best studied with respect to the regulation and the biochemical aspects (14 -16). It consists of the periplasmic reductase TorA, belonging to the Me 2 SO reductase family described above, and TorC, a membrane-anchored pentahemic c-type cytochrome. A recent study showed that the interaction of the two proteins allows the electron transfer from TorC to the terminal reductase TorA (17). The genes encoding TorC and TorA are organized in the torCAD operon. The last gene of this operon codes for TorD, a 22-kDa cytoplasmic protein (18). The related operons, torECAD and dorCDA (also called dmsCBA), were characterized in Shewanella and Rhodobacter species, respectively, and they encode E. coli TorC, TorA, and TorD homologous proteins (19 -21).
We showed previously that the TorD proteins from E. coli and Shewanella massilia bind specifically to their cognate TorA enzymes (22,23). Moreover, the absence of TorD in E. coli leads to a significant decrease of the amount of TorA (22). Strikingly, in Rhodobacter capsulatus the absence of DorD results in a complete loss of DorA (24). Finally, in an E. coli MGD-deficient strain, an excess of TorD limits the proteolytic degradation of the TorA cytoplasmic apo form (22). All these results suggest that the proteins of the TorD/DorD family play the role of chaperones dedicated to molybdoenzymes of the Me 2 SO reductase family. However, they can act either passively by protecting the reductases against cytoplasmic proteases or more actively during the folding of the enzymes or even during the insertion of the bis(MGD)Mo cofactor. The x-ray structure of TorD of S. massilia was recently solved (25). It reveals that the protein presents an all helical architecture organized in a globular domain showing no similarity with any known protein structures. Based on sequence homologies, proteins of the TorD/DorD family could share a similar all helical three-dimensional fold. This new protein fold lets us imagine that this protein family might play an original role during molybdoenzyme biosynthesis. Moreover, DmsD, an E. coli TorD homologue, was proposed to act as an escort protein targeting periplasmic molybdoenzymes to the twin arginine translocation machinery through its binding to their leader sequence (26). This chaperone family could, therefore, possess a dual function in the maturation process of the enzyme and the targeting to the export machinery.
In this study, using an in vitro approach we showed that TorD is required for the efficient maturation of the apo form of TorA and that it probably acts just before the bis(MGD)Mo cofactor insertion. We thus propose that TorD binds the apo form of TorA to favor a conformation of the apoenzyme that is competent for acquiring the cofactor.
Constructions of Plasmids-To construct plasmid pTorA, allowing the synthesis of His 6 -tagged holo-and apoTorA, the complete torAcoding sequence was amplified by PCR, performed with MC4100 chromosomal DNA as template, and primer TorAN, which corresponds to an XbaI site followed by the Shine-Dalgarno motif of torA and its 5Ј-coding sequence, and primer TorAC, which corresponds to a HindIII site followed by a sequence encoding a C-terminal His 6 tag and the complementary sequence of the 3Ј end of torA. The purified PCR product was digested by the appropriate restriction enzymes and cloned into the corresponding sites of expression vector pJF119EH (30). The recombinant plasmid was then introduced into strains LCB436 and RK5208.
To construct plasmid pTorD, allowing the synthesis of His 6 -tagged TorD, the same cloning strategy was used except that one primer (TorDN310) corresponds to an XbaI site followed by a Shine-Dalgarno consensus motif and the torD 5Ј sequence, and the other (TorDC310) corresponds to a HindIII site followed by a sequence encoding a Cterminal His 6 tag and the complementary sequence of the 3Ј end of torD. The digested and purified PCR product was cloned into the corresponding sites of vector pET28 (Novagen). The recombinant vector pTorD was introduced into strain BL21(DE3). All PCR products and fusion sites were confirmed by sequencing. Transformations were carried out according to the method of Chung and Miller (31).
Purification of Recombinant Proteins-His 6 -tagged TorD was produced from strains BL21(DE3) harboring plasmid pTorD. The recombinant strain was grown aerobically until the optical density at 600 nm reached 0.8 unit. Overproduction of the protein was then induced with isopropyl ␤-D-thiogalactopyranoside (1 mM) for 4 h. TorD was purified from the soluble fraction after French press treatment of the cells in phosphate buffer (20 mM, pH 7.4) and two successive centrifugations (18,000 ϫ g for 10 min and 120,000 ϫ g for 90 min). The supernatant was equilibrated in phosphate buffer (20 mM, pH 7.4) containing NaCl (500 mM) and imidazole (5 mM) and loaded onto a HiTrap TM -chelating Ni 2ϩ column (Amersham Biosciences). Proteins were eluted with a step gradient of imidazole ranging from 20 to 500 mM. TorD was recovered in the 150 mM imidazole fraction mainly in a monomeric form. All steps were performed at 4°C. The purification of His 6 -tagged apoTorA was carried out from the soluble fraction of strain RK5208, harboring plasmid pTorA. Overproduction and purification of apoTorA were performed as described for TorD. Strain LCB436 containing plasmid pTorA, grown as described above, was used to produce the mature form of His 6 -tagged TorA (holoTorA). HoloTorA was purified from the periplasmic fraction of the cells prepared by the lysozyme-EDTA procedure (32) and dialyzed overnight against 20 mM phosphate buffer, pH 7.4. The periplasm was then diluted in phosphate buffer (20 mM, pH 7.4) containing NaCl (500 mM) and imidazole (5 mM) and loaded onto a HiTrap TM chelating Ni 2ϩ column to perform the purification as described for TorD. ApoTorA and holoTorA were eluted with 150 mM imidazole. For subsequent uses, apoTorA, holoTorA, and TorD were dialyzed against phosphate buffer 20 mM, pH 7. Purified MobA was a C-terminal His 6 -tagged recombinant protein kindly provided by the Giordano's group (Laboratoire de Chimie Bactérienne, CNRS).
Preparation of Supernatant Fraction (SN504) of Strain LCB 504 -Cells of strain LCB504 grown anaerobically were harvested at late exponential phase, resuspended in phosphate buffer (20 mM, pH 7), and disrupted by French press. The extract was centrifuged twice, and the soluble fraction obtained (SN504) was kept at a protein concentration of 40 mg/ml. All steps were performed at 4°C under nitrogen atmosphere.
Release of MPT-Mo from mobA Soluble Fraction-Cells were grown anaerobically in presence of KNO 3 (10 mM) to increase the amount of MPT-Mo (33). The supernatant fraction (40 mg/ml), obtained as described for SN504, was heated for 5 min at 100°C under a nitrogen atmosphere and centrifuged 10 min at 18,000 ϫ g.
Native Gel Experiments-ApoTorA (1.68 M) and TorD (21 M) in phosphate buffer (20 mM, pH 7, final volume of the sample, 20 l) were incubated together or separately for 90 min at 37°C. Before loading the samples onto a 10% polyacrylamide gel, both samples were treated with 10 l of either native sample buffer (200 mM Tris-HCl, pH 8.8, 5 mM EDTA, 1 M sucrose, 0.1% bromphenol blue) or denaturing-reducing sample buffer (66% native sample buffer, 6% SDS, 0.1 M dithiothreitol, and 1.7% ␤-mercaptoethanol). When indicated, the samples were heated at 70°C for 5 min before loading. The migration patterns were revealed by Western blots using TorA or TorD antisera.
Analytical Procedure-Purified proteins were subjected to 12.5% SDS-PAGE and stained with Coomassie Blue. For Western blotting, proteins were transferred to a Hybond ECL nitrocellulose membrane, and the ECLϩ Plus Western blotting System was used as recommended by the supplier (Amersham Biosciences). Protein concentrations were measured by the technique of Lowry et al. (34). TMAO reductase activity was measured spectrophotometrically at 37°C by following the oxidation of reduced benzyl viologen at 600 nm coupled to the reduction of TMAO (35).
Interaction Study by Biosensor Experiments-The surface plasmon resonance (BIAcore apparatus, Biosensor) was used to analyze the binding between TorD (M r 22,427) and apoTorA (M r 94,263). All experiments were carried out at 25°C. TorD in 10 mM sodium acetate buffer, pH 5, was immobilized on a sensor chip CM5 (BIAcore) through amine coupling. ApoTorA was diluted in phosphate buffer (20 mM, pH 7) and injected over the test and control (no protein immobilized) surfaces at a flow rate of 30 l/min. Analysis of interactions was carried out by using the BIAevaluation 3.0 software.
Chemical Cross-linking Studies-These experiments were carried out by using 1,6-bismaleimidohexane (BMH) as a cross-linker. Proteins (apoTorA, 1.6 M; TorD, 28 M) were incubated for 30 min at room temperature in phosphate-buffered saline with BMH (1 mM). The interactions were analyzed by SDS-PAGE. Western blot detection was performed with TorA antibodies.
Activation of Purified ApoTorA-In the assays apoTorA (0.55 M) was mixed with 100 l of SN504 (40 mg/ml) and various concentration of TorD (from 0 to 2.1 M), and phosphate buffer (20 mM, pH 7) was added to bring the total volume to 150 l. All solutions were oxygendepleted. The mixture was incubated from 0 to 240 min at 37°C under a nitrogen atmosphere. To monitor the extent of activation, the TMAO reductase activity recovered in the samples was measured as a function of incubation time and was expressed as a percentage of the activity measured with holoTorA (0.55 M). As a control, the same assays were performed using bovine serum albumin (BSA) (20 M) either instead of TorD or in addition of TorD (2.1 M). The experimental data that correspond to at least three assays for each point were fitted to exponential curves. As a control, the same assays were performed in the absence of either apoTorA or SN504, and no activity was detected.
ApoTorA and TorD Preincubation Experiments-Three assays were performed. ApoTorA (3.3 M, 25 l) and TorD (12.5 M, 2.5 l) were mixed together and incubated for 60 min at 37°C before the addition of SN504 (100 l, 40 mg/ml) and phosphate buffer (20 mM, pH 7) to bring the final volume to 150 l. As a control, the same assay was performed except that no TorD was added. In a last assay, apoTorA The samples were then incubated at 37°C, and the extent of activation was measured as described above. As a control, assays containing all the compounds except either apoTorA, MobA, or MPT-Mo source were incubated in the same conditions as indicated above and led to no TMAO reductase activity.

Purification of TorD and of the Holo and Apo Forms of
TorA-To investigate the putative role of TorD during the maturation process of TorA, we produced and purified each protein as C-terminal His 6 -tagged recombinant proteins. The different forms of His 6 -tagged TMAO reductase were obtained from an expression plasmid harboring the coding sequence of tagged TorA. The plasmid was introduced either into a tor strain or into a mobA strain, allowing production of the mature holo form (active) or the immature apo form (non-active) of His 6 -tagged TorA, respectively. Tagged-holoTorA was purified from the periplasmic fraction of the tor mutant. The specific activity (65 mol of TMAO reduced/min/mg) was close to that observed with untagged TorA, indicating that the addition of the His 6 tag modifies neither the activity of the enzyme nor its maturation. ApoTorA, the inactive tagged protein devoid of bis(MGD)Mo, was synthesized in the mobA strain and was purified from the cytoplasm. The TorD protein was also produced as a His 6 -tagged recombinant protein. HoloTorA, apo-TorA, and TorD were purified near to homogeneity, as checked by Coomassie Blue staining after SDS-PAGE of the proteins (Fig. 1).
TorD Interacts with ApoTorA-To characterize a possible binding between apoTorA and TorD, we studied the formation of an apoTorA-TorD complex using various complementary techniques. We first analyzed a potential interaction between the two proteins using a surface plasmon resonance (BIAcore) technique, and for this purpose, the purified TorD protein was coupled to the dextran matrix of a sensor chip, and apoTorA was injected over it. The sensorgram obtained reflected the association and dissociation of two proteins, indicating that apoTorA directly interacts with TorD ( Fig. 2A). The apoTorA-TorD interaction was also revealed by cross-linking experiments using BMH or 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide as chemical ligands (Fig. 2B and data not shown). TorD and apoTorA alone or together were incubated with BMH, and the mixtures were loaded onto an SDS-polyacrylamide gel. Western blotting was then carried out with TorA antibodies.
An additional band was detected in the track containing the two proteins, compared with those with TorD or apoTorA alone. The mobility of this band is consistent with the molecular mass of a TorD (22.5 kDa)-apoTorA (94 kDa) heterodimer complex (Fig. 2B).
Another approach to confirm the interaction was to analyze the behavior of these proteins by loading an apoTorA-TorD mixture previously incubated at 37°C onto a 10% polyacrylamide native gel. After electrophoresis under native conditions, the migration patterns were revealed by Western blot using antibodies specifically raised against each of the proteins (Fig.  3, A and B). When TorA-specific antibodies were used, a new band below that of apoTorA alone, appeared in the lane corresponding to the mixture of both proteins (Fig. 3A). In contrast, no additional band other than that corresponding to the migration pattern of TorD alone was revealed using TorD antibodies (Fig. 3B). This result suggests that the additional band corresponds either to a specific proteolysis of apoTorA mediated by TorD or to a new migrating form of apoTorA induced by TorD. To draw a distinction between the two possibilities, the samples were treated with a non-denaturing or a reducing-denaturing loading buffer and heated at 70°C before being loaded on native polyacrylamide gel (Fig. 3, C and D). The fact that the band was still present after heating the sample in non-reducing buffer but disappeared in reducing conditions indicates that this band corresponds to a new fold of apoTorA stabilized by at least one disulfide bridge and discards the possibility of apo-TorA proteolysis mediated by TorD. We suspect that the conformational change in apoTorA induced by TorD brings in vicinity two cysteine residues, allowing the formation of a disulfide bridge in some cases. Altogether, these results suggest that TorD not only interacts with apoTorA but also modifies the conformation of apoTorA. TorD Increases the Activation of ApoTorA in Vitro-The fact that TorD interacts with apoTorA and apparently induces apo-TorA conformational change prompts us to study whether TorD contributes to the maturation of this enzyme and possibly to the bis(MGD)Mo insertion. Because apoTorA has no TMAO reductase activity but can be activated upon bis(MGD)Mo insertion, we measured apoTorA activation in the presence and in the absence of TorD. The bis(MGD)Mo source used in this experiment was the supernatant fraction of strain LCB504 (referred as SN504). This strain, mutated on both tor and dms operons is devoid of TMAO reductase activity and does not synthesize TorD. On the other hand, this strain produces a complete bis(MGD)Mo and can, thus, be used as a convenient cofactor source for the assay. ApoTorA (0.55 M) was mixed with SN504 (100 l, 40 mg/ml of proteins) under anaerobic conditions and incubated at 37°C. The extent of apoTorA activation was determined by measuring the TMAO reductase activity generated in the sample as a function of the incubation time (Fig. 4A). In this assay, the amplitude value of the recovered activity reached about 20% of the activity of the same amount of holoTorA. Because the activity measured corresponds to the quantity of active enzyme obtained in the sample, we conclude that only one-fifth of apoTorA can be activated even after a long time of incubation. To check whether the amount of SN504 was or was not limiting in the assay, increasing it did not lead to a change in the recovered activity, whereas decreasing it led to a decrease in the level of apoTorA activation (data not shown).
To assess a possible role of TorD in the TorA maturation process, increasing amounts of TorD were added to the assay defined above. As shown in Fig. 4A, the recovered TMAO reductase activity increased from 20 to 80% that of holoTorA when TorD concentration was raised from 0 to 2.1 M. TorD concentration higher than 2.1 M did not further increase the apoTorA activation (data not shown). In conclusion, in the absence of TorD, only a fraction of apoTorA could be activated in the assay, whereas the presence of TorD allowed the maturation of most of the apoprotein. The fact that TorD seems to facilitate the bis(MGD)Mo insertion into apoTorA strongly suggests that it plays an active role in the maturation process of the apoenzyme. To verify that TorD has a specific action toward apoTorA, it was substituted by the BSA. As expected, the apoTorA activation was not increased in the presence of a high concentration of BSA compared with that obtained in absence of TorD (Fig. 4A). Moreover, when both TorD and BSA were added together in the same assay, no decrease of apoTorA activation was observed compared with that measured with TorD alone (data not shown). These experiments show that the increase of apoTorA activation is specifically due to the presence of TorD in the assay.
TorD Acts on ApoTorA before the Insertion of the Molybdenum Cofactor-So far we showed that TorD interacts with apoTorA, induces a conformational change of apoTorA, and greatly increases activation of the apoprotein in our in vitro assay. If the role of TorD is to modify the apoTorA conformation to allow a better cofactor insertion, preincubation of apoTorA with TorD before the addition of the cofactor source should lead to a better incorporation of bis(MGD)Mo cofactor. ApoTorA (0.55 M, concentration in the final assay) was incubated for 60 min at 37°C alone as well as with a limiting concentration of TorD (0.21 M, concentration in the final assay) before SN504 was added, and then the extent of apoTorA activation was measured as a function of time. Interestingly, preincubation of the two proteins together before the addition of the bis(MGD)-Mo source improves the maturation of apoTorA since the recovered activity was raised from 26 to 38% that of holoTorA (Fig. 4B). In a similar experiment, the increase of TorD concentration allowed the activation of apoTorA to reach about 80% of holoTorA activity but abolished the effect of the preincubation of both proteins, observed above (data not shown). In contrast, whatever the TorD concentration, preincubation of TorD with SN504 did not modify the apoTorA activation, and curves similar to that of Fig. 4A were obtained (data not shown). As a control, we also checked that the recovered activity was lower in the absence of TorD than in its presence (Fig.   FIG. 3. Effect of TorD protein on apoTorA migration pattern. A  and B, apoTorA (1.68 M), a mix of apoTorA (1.68 M) and TorD (21 M), or TorD (21 M) were incubated 90 min at 37°C. After 10% native PAGE of the proteins in native buffer sample, they were visualized by Western blotting with TorA (A) or TorD (B) antibodies. C, as in A except that the samples were heated at 70°C, 5 min before loading on native gel. D, as in C except that the samples were treated with a denaturingreducing buffer sample. Arrows indicate the additional band as revealed by TorA antiserum; red and non-red are for reducing and nonreducing conditions, respectively. 4B). These results are in agreement with our previous experiments showing that TorD interacts with apoTorA even in the absence of the cofactor and supports the idea that TorD induces a conformational change of apoTorA, turning it into a competent state to receive the molybdenum cofactor. Maturation of apoTorA could, thus, involve an additional step before bis(MGD)-Mo insertion.
Active Role of TorD during the ApoTorA Maturation Process-In the reconstitution assays we used so far, the bis(MGD)-Mo was provided by the supernatant fraction of strain LCB504. In this case, an indirect effect of TorD over one constituent of SN504, a passive role of TorD against proteases, or the involvement of another protein present in the extract could not be excluded. To check whether TorD does not need another partner to facilitate the apoTorA maturation, we performed a defined in vitro system. In this defined system the bis(MGD)Mo is synthesized directly in the sample from MPT-Mo, GTP, MgCl 2 , and the purified MobA protein, which catalyzes the conversion of MPT-Mo to bis(MGD)Mo cofactor. The MPT-Mo source was provided by a heated supernatant fraction of mobA strain (100°C, 5 min). As observed in the experiments performed with SN504, when TorD (2.1 M) was added to the defined system, the activation of apoTorA strongly increased compared with that obtained in the absence of the chaperone (Fig. 5). The amplitude value reached a maximum corresponding to 63% that of the activity of holoTorA after 4 h of incubation, whereas it corresponds to only 13% of holoTorA activity in the absence of TorD. The plateau value obtained in the presence of TorD is, thus, almost five times higher than that obtained in the absence of TorD. As a control, we verified by Western blot that the quantity of the TorA polypeptide in the samples was still the same in the presence or not of TorD (data not shown). The fact that the proteins of the supernatant fraction were previously denatured in our reconstitution system rules out the involvement of a protein different from TorD and MobA during the in vitro TorA maturation process and also excludes a passive role of TorD in apoTorA protection against proteases. Moreover, assuming that MobA is only involved in the biosynthesis of the bis(MGD)Mo, TorD alone is sufficient to allow an efficient maturation of apoTorA. DISCUSSION The E. coli TorA protein is a TMAO reductase containing a bis(MGD)Mo cofactor in its catalytic site, and we showed in this study that TorD is the mate chaperone of TorA, allowing its efficient maturation. Indeed, the TMAO reductase activity that is recovered in vitro from immature apoTorA mixed with a source of molybdenum cofactor, increased as a function of TorD concentration. It reached an amplitude value corresponding to the activation of 63-80% of the apoTorA present in the assays, whereas only 13-20% of apoTorA was matured in the absence of TorD (Fig. 4A and 5).
Because preincubation of apoTorA with TorD before the cofactor addition significantly increased the quantity of recovered TMAO reductase activity (Fig. 4B), we suggest that TorD acts in a first step onto apoTorA to favor insertion of the bis(MGD)-Mo cofactor (Fig. 6). This hypothesis is also supported by the fact that TorD binds to apoTorA in the absence of any other compound and probably modifies to some extent apoTorA conformation (Figs. 2 and 3). It would be interesting to define the TorD and apoTorA binding regions and to determine whether the conformational change induced by TorD is restricted or not to the region of the TorA catalytic site. To explain that only a small fraction of apoTorA was matured in the absence of TorD even after a long incubation time, we propose that a large fraction of apoTorA evolves to a stable non-activable conformation, whereas the remainder turns spontaneously into a competent form able to receive the molybdenum cofactor. In this line of thought, the presence of TorD could shift the equilibrium toward the competent state of apoTorA and, thus, allows the maturation of the majority of the apoenzyme (Fig. 6). However, our results do not exclude an additional role for TorD during or even after the cofactor insertion.
Using a defined in vitro system, we observed that the efficient activation of apoTorA is obtained in the presence of only two functional proteins, TorD and MobA (Fig. 5). Although this latter is required for the conversion of MPT-Mo into bis(MGD)-Mo (5), it remains possible that MobA plays some role during TorA maturation, for example by escorting the cofactor during its insertion into apoTorA. To clarify this point, it would be necessary to perform defined in vitro assays without MobA but unfortunately, pure bis(MGD)Mo is unstable, and MobA is always required for this kind of in vitro experiment (3).
It should be noted that the presence of E. coli-soluble proteins in the cofactor-containing extract does not significantly change the pattern of activation of apoTorA compared with the experiment carried out with the defined in vitro assays ( Fig. 4A  and 5). This observation indicates that, at least in vitro, no soluble protein can efficiently substitute for TorD during the maturation process of TorA. However, it was demonstrated that GroEL is required for the final insertion of the ironmolybdenum cofactor into the nitrogenase of Azotobacter vinelandii (36). Therefore, the possibility that a general chaperone may also be involved in the maturation process of apoTorA should be considered.
The TorD family is a large family of proteins (25), but except for TorD of E. coli, the putative role of these proteins in the maturation process of molybdoenzymes remains to be defined. Interestingly, disruption of the dorD gene, which encodes a TorD homologue in R. capsulatus, leads to the disappearance of the molybdoenzyme DorA (24). This result points out a relationship between DorD and DorA and suggests that immature DorA is rapidly degraded unless DorD binds to it and allows maturation. It was recently shown that 73% of apoDorA can be activated in vitro after 7 h of incubation (5). This experiment was carried out in the absence of DorD, and it would, thus, be interesting to test whether the presence of DorD will accelerate and/or increase the maturation of DorA.
There is little evidence for the implication of chaperones in the maturation process of molybdoenzymes, and in addition to TorD, only two proteins, NarJ and XDHC, have been proposed to play such a role toward their cognate molybdoprotein. Although both proteins do not belong to the TorD family, NarJ was shown to act as a specific chaperone of the bis(MGD)Mocontaining subunit of the membranous nitrate reductase of E. coli, and XDHC was suspected to be involved in the maturation of MPT-Mo containing xanthine dehydrogenase (XDH) of R. capsulatus (37)(38)(39). Because several homologues of these proteins are found in different organisms, we postulate that many molybdoenzymes need a mate chaperone during their maturation process to improve cofactor insertion.