A Simplifed Functional Version of the Escherichia coli Sulfite Reductase*

Escherichia coli sulfite reductase (SiR) is a large and soluble enzyme with an a 8 b 4 quaternary structure. Pro- tein a (or sulfite reductase flavoprotein) contains both FAD and FMN, whereas protein b (or sulfite reductase hemoprotein (SiR-HP)) contains an iron-sulfur cluster coupled to a siroheme. The enzyme is set up to arrange the redox cofactors in a FAD-FMN-Fe 4 S 4 -Heme sequence to make an electron pathway between NADPH and sulfite. Whereas a spontaneously polymerizes, we have been able to produce SiR-FP60, a monomeric but fully active truncated version of it, lacking the N-terminal part (Zeghouf, M., Fontecave, M., Macherel, D., and Cove`s, J. (1998) Biochemistry 37, 6114–6123). Here we report the cloning, overproduction, and characterization of the b subunit. Pure recombinant SiR-HP behaves as a monomer in solution and is identical to the native protein in all its characteristics. Moreover, we demonstrate that the combination of SiR-FP60 and SiR-HP produces a functional 1:1 complex with tight interactions retaining about 20% of the activity of the native SiR. In

In Escherichia coli, sulfite reductase (SiR) 1 catalyzes the transfer of six electrons from NADPH to sulfite to produce sulfide, which then is used during the synthesis of L-cysteine (1). SiR is a fascinating system not only because of the multielectron transfer reaction that it controls but also because of its complex organization. The enzyme is a 780-kDa soluble complex composed of two proteins, SiR-FP and SiR-HP (2). SiR-FP is a homo-octameric ␣ 8 flavoprotein (530 kDa), in which each ␣ protein binds one FAD and one FMN and contains a NADPHbinding site (3)(4)(5). SiR-FP thus clearly belongs to a family of proteins that includes NADPH-cytochrome P450 oxidoreductase, nitric oxide synthase, and cytochrome P450 BM-3, which have in common the presence of two flavin-binding domains, one ferredoxin:NADP ϩ -oxidoreductase-like domain for FAD and one flavodoxin-like domain for FMN (6 -9). SiR-HP is a homotetrameric ␤ 4 metalloprotein (250 kDa), in which each ␤ protein contains a Fe 4 S 4 cluster linked to a siroheme through a cysteine bridge (3,10). It has been shown that electrons are transferred from NADPH to FAD and subsequently transferred to FMN in SiR-FP, from which they are transferred to the metal center of SiR-HP, where they serve to reduce sirohemebound sulfite (11,12). With regard to such an electron transfer pathway, the proposed holoenzyme ␣ 8 ␤ 4 quaternary structure is rather intriguing. However, thus far, no three-dimensional structure is available to confirm such a stoichiometry, probably because of the size of the system.
Although SiR-FP and SiR-HP interact very strongly to form the ␣ 8 ␤ 4 holoenzyme, they can be separated. However, this requires urea-dependent dissociation of the holoenzyme, which results in significant denaturation of both proteins and the loss of certain prosthetic groups (4,13). When isolated, SiR-FP retains its octameric structure, whereas SiR-HP behaves as a monomer. Such a SiR-HP preparation was successfully crystallized and used to solve SiR-HP three-dimensional structure at 1.6 Å resolution (10).
SiR-FP has resisted crystallization thus far. However, during our investigation of SiR-FP, we discovered that deletion of the first 51 N-terminal amino acids of the ␣ polypeptide resulted in an ␣Ј protein, which no longer polymerized to an octamer but instead behaved as a monomer in solution (14). This observation demonstrated that the N terminus contained all the determinants required for the polymerization. The truncated protein, named SiR-FP60, retained both FAD and FMN as cofactors and displayed all the catalytic diaphorase activities of SiR-FP. SiR-FP60 could be crystallized, and its three-dimensional structure was determined at 1.94 Å resolution (15). Unfortunately, whereas we had confirmation that the FAD domain was folded as in ferredoxin:NADP ϩ -oxidoreductase or as in cytochrome P450 reductase, the whole FMN domain remained invisible, most likely as a consequence of a high degree of flexibility.
With the objective of obtaining more detailed information on the unique electron transfer pathway in SiR at the molecular level, we have studied the structural and catalytic properties of mixtures of SiR-FP60 and SiR-HP. For that purpose, we have prepared and characterized a recombinant SiR-HP protein.
Here we demonstrate that recombinant SiR-HP interacts very tightly and specifically with SiR-FP60 to generate a functional ␣Ј␤ complex that behaves as a simplified version of sulfite reductase, showing that an ␣ 8 ␤ 4 structure is not absolutely required. This new complex may prove to be a useful tool to study SiR-HP/SiR-FP interactions and electron transfer reactions as well.
Cloning of SiR-HP-Standard recombinant DNA techniques were used to generate the construct as described previously (17). The plasmid pJYW613, encoding for SiR, was used as a template to generate a PCR product corresponding to the cysI gene that encodes for the hemoprotein (␤ chain). The sense primer, 5Ј-ACGAATTCATATGAGCATGAGC-GAAAAACATCCAGGGCC, was designed to be complementary to the 5Ј end of cysI and to incorporate an EcoRI site (italic letters) followed by a NdeI site (underlined) that provides the start codon for the recombinant protein expression. The antisense primer, 5Ј-CGGGATCCTTAATC-CCACAAATCGCGCGCCGGGC, was designed to complement the 3Ј end of the cysI gene and to incorporate a BamHI site (italic letters) downstream of the stop codon. The amplification was performed by PCR using the following scheme: 60 s at 94°C, 90 s at 64°C and 150 s at 68°C for 12 cycles. This PCR product was first cloned into vector pET-3a (Novagen) between the NdeI and the BamHI sites, leading to the construct pET-apoHP. Another PCR product corresponding to the cysG gene, encoded for the siroheme synthase, was generated with pJYW613 as a template. The sense primer, 5Ј-GCGGATCCGGCTGCT-GCCAATAATTAAGGGGC, was designed to include the Shine-Dalgarno motif of the cysG gene and the antisense primer, 5Ј-CGGGATC-CTTAATGATTAGAGAACCAATTT, annealed the 3Ј end. Both are designed to incorporate a BamHI site (italic letters). The amplification reaction was carried out as described above, except that an annealing temperature of 54°C was used. The PCR product was cloned in the BamHI site of pET-apoHP, and a ligation product containing the cysG gene in the same orientation as cysI was selected by PCR. This construct, named pET-SiR-HP, allowed for the overproduction of the recombinant SiR-HP used in this study. The authenticity of all PCR products was confirmed by sequencing on both strands (Genome Express, Grenoble, France).
Overexpression and Purification of Proteins-SiR (the holoenzyme), SiR-FP (the flavoprotein), and SiR-FP60 (the monomeric form of the flavoprotein) were expressed and purified as described previously (4,14). Overexpression of SiR-HP was achieved by using a B834(DE3)pLysS E. coli strain (Novagen) transformed with pET-SiR-HP. The culture was grown for 2 h at 37°C to an A 600 of 0.6 and chilled on ice for 10 min before induction with 0.2 mM isopropyl-1-thio-␤-Dgalactopyranoside. After 3.5 h of incubation at 25°C, cells were harvested by centrifugation, and soluble extracts were prepared as described previously for SiR-FP (4). All subsequent steps were performed at 4°C. The pellet obtained after ammonium sulfate precipitation (60% final saturation) was dissolved in a minimal volume of 20 mM potassium phosphate (pH 7.4, buffer A) and dialyzed extensively against the same buffer. The protein solution was loaded at a flow rate of 1 ml min Ϫ1 onto a Q-Sepharose fast-flow column (2.6 ϫ 9.5 cm; 50 ml; Amersham Pharmacia Biotech) equilibrated with buffer A. The column was washed at the same flow rate with 100 ml of buffer A, and then elution was performed with a linear gradient from 0 to 1 M KCl during a 500-min period. Fractions (4 ml) were collected and analyzed for protein (A 280 ) and for SiR-HP content (A 389 ). The SiR-HP-containing fractions, eluting at about 90 mM KCl, were pooled, concentrated using a Diaflo cell equipped with a YM-30 membrane (Amicon Co.), and subjected to filtration on a Superdex-75 column (Amersham Pharmacia Biotech) equilibrated with 50 mM potassium phosphate buffer, pH 7.4. After concentration to approximately 10 mg ml Ϫ1 , aliquots were flash-frozen in liquid nitrogen and stored at Ϫ80°C until further use.
Cofactor Analysis, Protein Determination, and Enzymic Assay-FMN and FAD content was quantitated by fluorometric analysis (18) using a PerkinElmer Life Sciences LS450 apparatus after extraction from protein samples denatured by boiling for 10 min in the dark. An extinction coefficient of 10.4 mM Ϫ1 cm Ϫ1 at 450 nm was used to quantify the oxidized flavins in SiR-FP60 (14). Siroheme content of SiR-HP was determined as reported previously (19). Briefly, the cofactor was extracted with 0.015 N HCl-acetone and rapidly transferred into pyridine (33% final concentration). Then, quantitation was done spectrophoto-metrically using ⌬⑀ 557-700 ϭ 1.57.10 4 M Ϫ1 cm Ϫ1 . Iron and acid-labile sulfur contents were also quantitated by colorimetric assays as described previously (20,21). Measurements were repeated at least three times for each protein sample.
Protein concentration was determined using the Bradford method with bovine serum albumin as a standard (22). Amino acid composition analysis of SiR-HP was performed by Dr. J. Gagnon (Institut de Biologie Structurale, Grenoble, France) as reported previously (4). Denatured molecular masses and purity of proteins were estimated by 0.1% SDS-12% polyacrylamide gel electrophoresis (23). The native molecular mass of the ␣Ј␤ complex was checked by electrophoresis run under nondenaturing conditions, using a Bio-Rad precast 4 -20% polyacrylamide gel.
Sulfite reductase activity was monitored spectrophotometrically at 25°C from the loss of absorbance at 340 nm due to the oxidation of NADPH by sulfite as described previously (24), except that all reactions were carried out in a 50 mM potassium phosphate buffer, pH 7.4, and initiated by the addition of NADPH. NADPH concentrations were determined using an extinction coefficient of 6.22 mM Ϫ1 cm Ϫ1 at 340 nm. Blanks in the absence of sulfite in the reaction mixture were systematically run. One enzymatic unit corresponds to the oxidation of 1 nmol NADPH/min. For the determination of apparent dissociation constants, kinetic data were treated as detailed for the tight-binding two-component system (25). For these calculations, the amount of the flavoprotein component (SiR-FP60 or SiR-FP) in the system was kept constant to always subtract the same blank obtained in the absence of sulfite (due to the low NADPH oxidase activity of SiR-FP).
Isothermal Titration Calorimetry-Isothermal titration calorimetry experiments were performed by Dr I. Jelesarov (Department of Biochemistry, University of Zurich, Zurich, Switzerland) using an OMEGA titration microcalorimeter (Microcal Inc., Northampton, MA) as described elsewhere (26). SiR-HP (1.36 ml; 15.7 M) was titrated with 10 l of successive additions of SiR-FP60 (280 M) at 25°C using Tris-HCl or potassium phosphate as buffer, both at pH 7.4, and I ϭ 0.1. The heat of binding, ⌬H, was obtained from the best nonlinear least-square fit of the data, according to a 1:1 binding model, using the data analysis software of the instrument. All data were corrected for the nonspecific heat released during a similar titration of SiR-HP with the respective buffer only.
Spectroscopic Studies-UV-visible spectra of aerobic samples were recorded at 25°C in a quartz cell (3 or 10 mm light path) using a Cary 1 Bio (Varian) spectrophotometer. For recording the absorption spectra of anaerobic samples, a cuvette holder inside a glove box under overpressure of nitrogen was connected with optical fibers to a Hewlett-Packard 8452A diode array spectrophotometer. EPR spectra were recorded with a Brucker EMX spectrometer operating at 9.45 GHz and equipped with an ESR 900 helium flow cryostat (Oxford Instruments). Neutral semiquinone radical on flavin was quantitated as described previously (17) using TEMPO as a standard. For quantitation, EPR first-derivative spectra of ferriheme or reduced Fe 4 S 4 cluster were recorded at 10 K at a nonsaturating microwave power of 1 milliwatt or 10 microwatts, respectively. 2 Spin contents were determined either by comparison with the signal of a standard solution of oxidized SiR-HP or by using a 100 M solution of Cu-EDTA as a standard (27).
Reduction Experiments-All anaerobic experiments were performed in a glove box under nitrogen overpressure and maintained at 17°C throughout. Photochemical reductions in the presence of deazaflavin-EDTA were achieved in a 50 mM potassium phosphate buffer, pH 7.4, containing 10 mM EDTA, 5% glycerol, and a 0.3 deazaflavin:protein ratio. Photoreduction of the protein sample was achieved with a slide projector and monitored spectrophotometrically as described previously (14). For EPR analysis, 200 l of protein solution were transferred in 4-mm calibrated EPR tubes. All of them were capped and flash-frozen within the glove box before storage in liquid nitrogen until analysis.

Expression and Purification of Recombinant SiR-HP-From
the pJYW613 plasmid containing the cysJIG operon, the cysI gene was cloned as a NdeI-BamHI fragment under the control of the T7 promoter into vector pET-3a. The cysG gene was cloned downstream of cysI as a BamHI-BamHI fragment containing its own Shine-Dalgarno motif. cysG, encoding for the siroheme synthase, was introduced to achieve optimal incorporation of siroheme into the polypeptide (28). The resulting 2 Unpublished observations. plasmid, named pET-SiR-HP, was then used to transform the E. coli B834(DE3)pLysS strain.
Maximal expression of soluble SiR-HP was obtained when cells were grown at 25°C for 2 h after the addition of 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside to the culture medium. Under these conditions, an intense band at 63 kDa is clearly visible after SDS-polyacrylamide gel electrophoresis performed on soluble extracts. However, because a large amount of protein was retained as insoluble material, the yield of recombinant SiR-HP protein was relatively low.
After anion-exchange chromatography on Q-Sepharose and filtration on Superdex-75, SiR-HP was obtained in a nearly pure form (more than 95% purity as deduced from SDS-polyacrylamide gel electrophoresis analysis; data not shown). Typically, 14 mg of that preparation could be obtained from 5 liters of culture. Furthermore, the behavior of the protein on Superdex-75 confirmed that SiR-HP was, in the absence of SiR-FP, a monomer in solution (data not shown).
Spectroscopic Properties of Recombinant SiR-HP-Solutions of SiR-HP are brownish, and their light absorption spectrum displays in the visible region the bands at 389, 590, and 712 nm characteristic for the ferrisiroheme. The spectrum shown in Fig. 1 is in perfect agreement with previously reported spectra of native SiR-HP (13,29,30). The siroheme content was quantitated spectrophotometrically after extraction with acetone-HCl, as described previously (19). On the basis of a careful determination of the concentration of the protein by amino acid composition analysis, recombinant SiR-HP was found to contain 0.9 mol siroheme/mol protein. The extinction coefficient at 590 nm was thus estimated at 18.5 mM Ϫ1 cm Ϫ1 , in full agreement with that of native SiR-HP (29). Finally, recombinant SiR-HP was found to contain an average of 4.1 mol of iron and 2.7 mol of acid-labile sulfur per mole of protein. After subtraction of the contribution of the siroheme iron, SiR-HP appeared to contain a substoichiometric amount of Fe 4 S 4 clusters (0.7 cluster/protein) and probably contained a small amount of adventitiously bound iron. Fig. 2A shows the EPR spectrum of the SiR-HP. It exhibits a signal characteristic of a high spin S ϭ 5/2 ferrisiroheme with features at g ϭ 6.65, 5.23, and 1.98, accounting for 0.9 iron/ protein. Then, evidence for a cluster came from the EPR spectrum of reduced SiR-HP. After anaerobic reduction with photoreduced deazaflavin in the presence of KCN (to generate a low spin S ϭ 0 ferrosiroheme), the features due to ferrisiroheme were no longer present in the EPR spectrum, and a new signal was observed with g values at 2.04, 1.94, and 1.91, characteristic for a S ϭ 1/2 (Fe 4 S 4 ) ϩ cluster (Fig. 2B). The temperature dependence and microwave power saturation properties of this signal are consistent with a (Fe 4 S 4 ) ϩ center. Again spin quantitation confirmed the presence of 0.7 cluster/ protein. Both spectra in Fig. 2 are similar to those reported previously for native oxidized and reduced SiR-HP (29,30).
All these results show that recombinant SiR-HP, although slightly depleted in iron sulfur cluster, contains both metal centers (one cluster and one siroheme) and closely resembles the native protein.
SiR-HP and SiR-FP60 Form a Tight 1:1 Complex-After incubation of a fixed amount of SiR-HP (␤) with increasing amounts of SiR-FP60 (␣Ј), the solution was analyzed by native gel electrophoresis. As shown in Fig. 3, a discrete band with an apparent molecular mass of 127 kDa appeared, along with a band corresponding to SiR-FP60 when this protein was in excess in the incubation mixture (lane 5). This is in agreement with the formation of a complex resulting from the association of 1 mol of SiR-HP (theoretical mass, 63,843 Da) with 1 mol of SiR-FP60 (theoretical mass, 60,706 Da). During filtration on Superdex-75, the complex, eluted in the dead volume of the column, could be separated from free SiR-FP60 (data not shown) and analyzed for its spectroscopic properties. FP60/SiR-HP complex. The spectrum of SiR-HP (or that of SiR-FP60) could be obtained by subtracting the spectrum of SiR-FP60 (or that of SiR-HP) from the spectrum of Fig. 4 (data not shown). This demonstrated that the tight association of the two proteins did not significantly affect the spectroscopic properties of the redox cofactors and that the spectrum of the complex is just the sum of the light absorption spectra of each of the components. Similarly, the EPR spectrum shown in Fig.  2A, characteristic for the ferrisiroheme, was not changed during the addition of SiR-FP60 to SiR-HP (data not shown). Furthermore, the spectrum in Fig. 4 is identical to that of the native sulfite reductase holoenzyme (3). This is true for both the position of the absorption bands and their relative intensities. This observation makes the long-agreed proposal that SiR is a ␣ 8 ␤ 4 complex inconsistent.
Finally, titration of SiR-HP by SiR-FP60 was followed by microcalorimetry (data not shown). However, this method did not allow us to determine an accurate value for the dissociation K d constant associated with the complex due to the limitations of the method with our experimental conditions, i.e. the release of heat upon binding of the two partners was not high enough to get very precise results. In fact, it showed that this constant was below 20 nM, thus confirming the tight association of the two proteins.
SiR-HP and SiR-FP60 Form a Functional Complex-Sulfite reductase activity was assayed during oxidation of NADPH by sulfite in the presence of a mixture of SiR-HP and of an excess of the flavoprotein component. The reaction is monitored spectrophotometrically from the decay of the light absorption at 340 nm, characteristic of the oxidation of NADPH. As shown in Table I, the native holoenzyme has a specific activity of 6,500 nmol min Ϫ1 mg Ϫ1 of SiR-HP, in agreement with other reports (4,11). As observed previously (13), it is possible to reconstitute a fully active enzyme by incubating SiR-HP and the octamer SiR-FP (␣ 8 ) prepared independently (Table I). Now we show that the combination of SiR-HP with the monomeric SiR-FP60 affords an active catalyst for sulfite reduction by NADPH. As shown in Table I, the specific activity accounts for 20% of the native holoenzyme. Fig. 5A shows the dependence of the activity on increasing amounts of SiR-FP60 for various amounts of SiR-HP. The system became saturated with respect to SiR-FP60, and saturation was achieved for larger amounts when the amount of SiR-HP was increased. Fig. 5B shows a similar experiment, in which the sulfite reductase activity is measured as a function of increasing concentrations of SiR-HP, with a fixed concentration of SiR-FP60. Again, a saturation behavior was observed. From both experiments, one can calculate that a stoichiometric amount of the complementing protein was required for saturation. Furthermore, considering that the ratelimiting step is the electron transfer between SiR-FP and SiR-HP (29) and thus assuming that the activity is a function of the concentration of the complex, one can estimate from Fig. 5B an apparent K d value of about 10 nM. All these results can thus be explained by a very tight interaction between the two proteins that generates a 1:1 SiR-HP/SiR-FP60 complex.
When the same titration experiments were carried out with SiR-FP (␣ 8 ) instead of SiR-FP60, saturation with respect to SiR-HP was also observed (data not shown). It appeared that saturation was obtained for a ␣:␤ ratio close to 1 rather than 2, suggesting that the holoenzyme SiR is more likely described as an ␣ 8 ␤ 8 complex. Under these conditions, an apparent K d value of about 28 nM has been estimated. DISCUSSION One of the confusing features of the complex enzyme system sulfite reductase is the stoichiometry of its quaternary structure. Since 1970, on the basis of analytical ultracentrifugation experiments carried out by L. Siegel and collaborators, it has been accepted without further discussion that SiR has an ␣ 8 ␤ 4 structure, with protein ␣ (SiR-FP) carrying flavin prosthetic groups (one FAD and one FMN, as more recently demonstrated), and protein ␤ (SiR-HP) carrying the iron-sulfur/sirohemecoupled center, where reduction of sulfite takes place. Considering the well-established fact that the electrons are shuttled, through FAD, from NADPH to FMN in protein ␣, it is thus difficult to understand how the electron transfer between FMN in protein ␣ and the metal center in protein ␤ proceeds. Indeed, there is an inconsistency between the proposed linear scheme for the electron transfer pathway (NADPH-FAD-FMN-FeSsiroheme-sulfite) and the above stoichiometry, which suggested that SiR was set up to make two FMNs converge on the same iron-sulfur/siroheme center within an ␣ 2 ␤ functional unit. Looking at 20 years of reported studies on SiR, we observe that this inconsistency has never been raised.
What we now show is that a 1:1 stoichiometric combination of FAD ϩ FMN and of the iron-sulfur/siroheme center is functional, revealing that there is no need for a 2-fold excess of the SiR-FP with regard to SiR-HP. This could be nicely demonstrated by using two original proteins. The first one, named SiR-FP60, is a monomeric form of SiR-FP (protein ␣Ј) that can be obtained by deleting the first 51 N-terminal amino acids, as shown previously. SiR-FP60 contains both FAD and FMN and retains both FAD-and FMN-dependent diaphorase activities of SiR-FP. The second one is a recombinant form of SiR-HP, which contains both the Fe 4 S 4 center and the siroheme with spectroscopic properties (light absorption and EPR spectroscopies) similar to those of the corresponding centers in the native SiR. It consists of only one ␤ polypeptide chain. This confirms that, in the absence of protein ␣, protein ␤ does not polymerize, and the presence of several ␤ polypeptides in SiR is a consequence of protein ␣ polymerization. This had already been shown, but at that time, the experiments were carried out with a slightly denatured SiR-HP preparation obtained from a ureatreated SiR holoenzyme solution.
The results reported here demonstrate that SiR-FP60 (␣Ј) and SiR-HP (␤) bind tightly to each other to form a 1:1 ␣Ј␤ complex, which is able to catalyze the reduction of sulfite by NADPH. Although the activity is only 20% of the activity of the SiR holoenzyme, this strongly suggests that ␣␤, and not ␣ 2 ␤, is the actual functional unit in SiR, consistent with the proposed linear electron transfer pathway discussed above. It thus follows that the currently proposed ␣ 8 ␤ 4 structure is probably incorrect, and in all probability, SiR is an ␣ 8 ␤ 8 complex, with equal amounts of the ␣ and ␤ components. This is in agreement with our titration experiments carried out with SiR-FP (␣ 8 ) and recombinant SiR-HP, which best fitted with such a stoichiometry. More direct evidence came from the observation that a 1:1 mixture of SiR-FP60 with SiR-HP gave a light absorption spectrum just superimposable to that of the native enzyme. Because this spectrum contains signatures for the cofactors of both ␣ and ␤, this result makes totally inconsistent the idea that the native enzyme contains a 2:1 ␣:␤ mixture, as is still currently accepted. More experiments are required to further support this new scheme. Nevertheless, it should be noticed that direct determination of the size of SiR might be a difficult project because the native molecular weight of this protein is in the million-Da range.
As demonstrated previously in the case of the cytochrome P450 reductase activity of SiR-FP (31), the N-terminal part of this protein seems to play an important role. As indicated above, deletion of the first 51 amino acid residues results in a different quaternary structure of the holoenzyme and in decreased activity. Additional experiments, particularly structural investigations, are required to understand the molecular basis of these effects.
Finally, because the SiR-FP60/SiR-HP complex behaves as a simplified version of the SiR system with a much smaller size and retaining both cofactors and enzyme activity, it is an excellent candidate for crystallization and structural studies of the complex electron transfer pathway resulting in the 6-electron reduction of sulfite. SiR and the octameric SiR-FP have resisted crystallization thus far, and the only structural information available is the three-dimensional structure of SiR-HP, on one hand, and, on the other hand, that of SiR-FP60. It should be noted that, unfortunately, the SiR-FP60 crystal structure does not show any electron density related to the FMN binding domain. This disorder was interpreted as a consequence of a functional flexibility of that domain which probably serves to catch SiR-HP to locate the FMN cofactor close to the metal center. Thus, binding of SiR-HP to SiR-FP60 might stabilize the FMN domain, and the structure of the SiR-FP60/ SiR-HP complex is expected to both allow determination of the structure of the FMN domain and reveal the molecular details of the NADPH to sulfite electron transfer pathway. This is our present goal.