Structural and mutational studies of an electron transfer complex of maize sulfite reductase and ferredoxin

Abstract

The structure of the complex of maize sulfite reductase (SiR) and ferredoxin (Fd) has been determined by X-ray crystallography. Co-crystals of the two proteins prepared under different conditions were subjected to the diffraction analysis and three possible structures of the complex were solved. Although topological relationship of SiR and Fd varied in each of the structures, two characteristics common to all structures were found in the pattern of protein-protein interactions and positional arrangements of redox centres; (i) a few negative residues of Fd contact with a narrow area of SiR with positive electrostatic surface potential and (ii) [2Fe-2S] cluster of Fd and [4Fe-4S] cluster of SiR are in a close proximity with the shortest distance around 12 Å. Mutational analysis of a total of seven basic residues of SiR distributed widely at the interface of the complex showed their importance for supporting an efficient Fd-dependent activity and a strong physical binding to Fd. These combined results suggest that the productive electron transfer complex of SiR and Fd could be formed through multiple processes of the electrostatic intermolecular interaction and this implication is discussed in terms of the multi-functionality of Fd in various redox metabolisms.

sulfite reductase, ferredoxin, electron transfer complex, x-ray crystallography, protein–protein interaction

Assimilatory sulfite reductase (SiR) catalyses the six-electron reduction of sulfite to sulfide. In most organisms, cysteine is synthesized from sulfide and O-acetylserine, and SiR is one of key enzymes for biosynthetic assimilation of sulfur-containing amino acids from inorganic sulfur (1, 2). In addition to its biosynthetic role, SiR is known to play a role in protection against the toxicity of sulfite accumulation in plant leaves (3).

SiR exists in two forms in biosphere; SiRs from oxygenic photosynthetic organisms such as higher plants and algae require ferredoxin (Fd) as a physiological electron donor (hydrogen-sulfide:Fd oxidoreductase, EC 1.8.7.1) (4, 5), and SiRs from heterotrophic organisms such as enterobacteria and fungi utilize NADPH as a reductant (hydrogen-sufide:NADPH oxidoreductase, EC 1.8.1.2) (6, 7). NADPH-SiR from Escherichia coli has been extensively studied and a large body of biochemical and biophysical knowledge is accumulated (8–10). The enzyme is an oligomeric protein composed of eight 66-kDa flavoprotein (SiRFP) and four 64-kDa hemoprotein (SiRHP) subunits. SiRFP has four FAD and four FMN in the octameric state, and transfers electrons from the reduced pyridine nucleotide to SiRHP. SiRHP contains one [4Fe-4S] cluster and one siroheme in each subunit. Siroheme serves as a site of sulfite reduction. X-ray crystal structure of SiRHP was determined at 1.6 Å resolution, and the three-domain hold of the polypeptide backbone, cofactor assembly and catalytic mechanism were revealed (8). Fd-dependent SiRs (Fd-SiRs) from higher plants and algae are a monomeric enzyme with a molecular mass of around 65 kDa and contain the same cofactors as SiRHP, one [4Fe-4S] cluster and one siroheme (11, 12). In spite of its simple molecular architecture, the mechanism by which a total of six electrons is utilized via three one-electron redox centres, namely [2Fe-2S] cluster of Fd and the two cofactors of SiR, is not fully understood (11), since no crystal structure of plant SiR complexed with Fd has been reported yet.

In higher plants, SiR is present both in leaf and root plastids, where reduced form of Fd required for the SiR activity is formed by the photoreduction via Photosynstem I and by NADPH reduction via Fd:NADPH oxidoreductase (FNR), respectively (13). Oxidation-reduction properties of [4Fe-4S] cluster and siroheme of maize SiR, which have E_m values of −400 and −285 to −190 mV, respectively, indicated that the reduction of SiR by reduced Fd, which E_m value is around −420 mV, would be a thermodynamically favorable process (14). As is the case for other Fd-dependent enzymes, such as FNR (15), glutamate synthase (16), nitrite reductase (17) and Fd:thioredoxin reductase (FTR) (18), SiR forms an electrostatically stabilized complex with Fd (11,19, 20). Crystal structures of the complexes of maize leaf Fd/FNR (21), Anabaena Fd/FNR (22) and Synechocystis Fd/FTR (23) have given an insight into their protein-protein interactions between Fd and Fd-dependent enzymes; a set of positively charged residues of the partner enzymes is close to the acidic residues of Fd in a spatially matching manner, although molecular shapes of the sites for the interaction differ considerably among the partner enzymes.

The interaction of Fd with SiR was studied by NMR chemical shift perturbation and cross saturation experiments (19) and by site-directed mutagenesis of Fd (20), and it was proposed that the interaction site of Fd with SiR differs at least partly from that of Fd with FNR. However, structural data regarding the interaction site of SiR in the complex with Fd has not been elucidated. Such information would be important for understanding molecular mechanisms of Fd function as a multi-functional electron donor to various Fd-dependent enzymes. We have been focusing on the crystal structure of the complex of SiR and Fd to obtain further insight into the mechanism for protein–protein interactions of Fd with multiple partner enzymes.

Here, we co-crystallized maize leaf SiR and Fd under several conditions and succeeded in solving three crystal structures of the complex at resolutions of 2.1–2.2 Å. The topology of the two proteins was found to be unique in each of the three complexes, but, interestingly, all structures showed general characteristics of electrostatic intermolecular interaction and a common close proximity of the redox centers of Fd and SiR. We hypothesized that Fd could form functional electron transfer complex with SiR in a multiple fashion. This idea was supported by biochemical studies of site-directed mutagenesis of SiR, which proved that basic residues distributed at the interface of any of three complexes are crucial for an efficient intermolecular transfer of reducing equivalents from Fd to SiR.

Materials and Methods

Purification of recombinant maize SiR and FdI

E. coli JM109 cells were transformed with the expression plasmid carrying maize SiR cDNA as described previously in (24). The transformed E. coli cells were grown overnight at 37°C in 50 ml of Lurial-Bertani medium containing 50 µg/ml ampicillin. The seed culture was inoculated into 8 L of the same medium and grown with vigorous aeration at 37°C for 3 h. Then culture temperature was lowered to 27°C and isopropyl-β-D-thiogalactopyranoside was added to a final concentration of 0.5 mM. After further cultivation for overnight, the cells were collected by centrifugation at 6,000 rpm for 10 min and stored at −30°C.

The frozen cells were suspended in 50 mM Tris-HCl, pH 7.5 containing 150 mM NaCl, 1 mM MgCl₂, 1 mM ethylenediaminetetraacetic acid and 0.1% (v/v) β-mercaptoethanol. After adding phenylmethylsulfonyl fluoride at a final concentration of 0.5 mM, the cell suspensions were thoroughly disrupted by sonication on ice. The broken cell suspension was centrifuged at 10,000 rpm for 15 min and resulting supernatant was mixed with DE-52 resin (GE Healthcare) in a batch wise manner. SiR was recovered in a pass-through fraction and fractionated by an ammonium sulfate fractionation between concentrations of 35 and 60%. The precipitated proteins were dialyzed against 25 mM Tris-HCl buffer, pH 7.5 overnight at 4°C and chromatographed on a DEAE-Toyopearl column (Toso Corp) with a linear gradient with NaCl from 0 to 200 mM in 50 mM Tris-HCl, pH 7.5. SiR fraction was added with ammonium sulfate to a final concentration of 30%, applied on a Phenylsepharose column (GE Healthcare), and developed with a reverse linear gradient of ammonium sulfate from 30 to 0% in 50 mM Tris-HCl buffer, pH 7.5.

Maize FdI was prepared as a recombinant protein expressed in E. coli JM109 as described previously (25). Unless otherwise specified, Fd indicated Fd1.

Co-crystallization of SiR and Fd

Purified SiR and Fd were passed through a small column of Sephadex G-25 resin (GE Healthcare) equilibrated with 25 mM potassium phosphate buffer, pH 7.5. Resulting two proteins were mixed in a molar ratio of 1:1 and concentration of the mixed proteins was adjusted to be 18–20 mg/ml. Crystallization was performed using hanging-drop vapour diffusion method at 5°C. The volume of the reservoir solution was 500 µl and the droplets consisted of 2 µl each of protein and reservoir solutions. Three kinds of buffer, (i) 85 mM Tris-170 mM sodium acetate trihydrate, pH 8.5, (ii) 85 mM Tris-170 mM sodium acetate trihydrate, pH 9.0 and (iii) 100 mM Tris-50 mM sodium acetate trihydrate, pH 9.0, containing precipitant, 25.5% (w/v) PEG 4000 and cryo-protectant, 15% (v/v) glycerol, were used as a reservoir solution. Crystals with reddish green colour were obtained under the three conditions, Form-1, -2 and -3 crystals, respectively, in 2 weeks at 5°C. Form-1 and -2 crystals were plate-shaped and Form-3 crystals were rugby-ball shaped with maximum dimensions of 0.5 × 0.5 × 0.2 mm.

Data collection and structure calculation

X-ray experiments of Form-1 crystals were performed on beamline BL-5 A at Photon Factory, KEK, Tsukuba, Japan. To obtain the crystallographic phase information, three independent data sets from Form-1 crystals were collected with the wavelength near the absorption edge of Fe atoms. Datasets of Remote1 (λ = 1.00000 Å) and Peak1 (λ = 1.73739 Å) were collected from one crystal, and the other two of Peak2 (λ = 1.73114 Å) and Edge (λ = 1.74134 Å) were from another crystal. Diffraction images were collected at −173°C using an ADSC Q315 charge-coupled-device detector equipped on the beamline. The other two diffraction data sets from the Form-2 and the Form-3 crystals were collected at −173°C on beamline BL-44XU at SPring-8, Hyogo, Japan. Diffraction images were collected using a DIP2060 image plate detector. All diffraction images were processed and scaled with the HKL2000 programme package. The data collection statistics are summarized in Table I. Crystals in three different forms each belong to a different space group; P4₁2₁2 for Form-1, P6₁ for Form-2 and P2₁2₁2₁ for Form-3 crystals.

Table I.

Data collection, phasing and refinement statistics

	Form-3	Form-2	Form-1
			Remote1	Peak1	Peak2	Edge
Data collectiona
Beam line	SPring-8 BL44XU	SPring-8 BL44XU	PF BL5A	PF BL5A	PF BL5A	PF BL5A
Wavelength (Å)	0.90000	0.90000	1.00000	1.73739	1.73114	1.74134
Space group	P212121	P61	P41212
Cell dimensions??
a, b, c (Å)	103.414	176.219	103.346	103.373	103.107	103.102
	104.406	176.219	103.346	103.373	103.107	103.102
	122.434	195.315	255.048	255.23	254.972	254.952
α,β,γ(°)	90, 90, 90	90, 90, 90	90, 90, 90	90, 90, 90	90, 90, 90	90, 90, 90
Resolution (Å)	50.0–2.08 (2.15–2.08)	50.0–2.20 (2.28–2.20)	50.0–2.20 (2.28–2.20)	50.0–3.18 (3.29–3.18)	50.0–2.32 (2.41–2.32)	50.0–3.18 (3.29–3.18)
Rmergeb	0.099(0.479)	0.110(0.525)	0.083(0.169)	0.075(0.103)	0.103(0.351)	0.093(0.123)
I/σI	12.0	6.5	13.2	12.0	9.3	8.4
Completeness (%)	100.0(100.0)	100.0(100.0)	99.4 (95.8)	99.4(92.3)	99.8(98.7)	96.7(99.4)
Redundancy	7.3(7.1)	6.9(6.6)	9.0(3.9)	13.8(13.8)	20.2(16.4)	4.8(4.7)
Phasing (acentric/centric)				2.332/2.132	0.450/0.515	0.894/0.920
Phasing power Isomorphous			−/−	1.687	1.134	0.851
Phasing power anomalous			0.478	0.418/0.410	0.909/0.837	0.717/0.659
RCullis (Isomorphous)			–/–	0.627	0.823	0.905
RCullis (anomalous)			0.961
Heavy atom sites			3
Figure of merit (FOM)					0.377/0.302
Overall FOM after density modification					0.53
Refinement
Resolution (Å)	43.27–2.08	30.07–2.20	48.06–2.20
Rwork/Rfree	0.179/0.232	0.205/0.255	0.204/0.249
No. molecules
SiR	2	4	2
Fd	1	2
Water	381	392	255
B-factors (Å2)
SiR-A	39.201	35.385	15.658
SiR-B	43.172	35.606	14.093
SiR-C		40.426
SiR-D		44.435
Fd-C/E	23.059	52.336
Fd-F		67.954
Water	34.804	34.200	11.581
R.m.s deviations
Bond lengths (Å)	0.016	0.017	0.021
Bond angles (°)	1.74	1.769	1.967

^aValues in parentheses are for the highest resolution shells.

^bR_merge(I) = Σ |I (k) − <I>| / Σ I (k), where I (k) is the value of the k th measurement of the intensity of a reflection, <I> is the mean value of the intensity of that reflection and the summation is over all measurement.

Crystallographic phase problem was solved by the multi-wavelength anomalous dispersion (MAD) method with anomalous scattering of Fe atoms using the program SHARP (26). Phase angles were improved using density modification and non-crystallographic averaging with the program RESOLVE (27). Phasing statistics are shown in Table I. An initial atomic model of SiR molecules was built on the density-modified electron-density map using the programmes of O (28) and COOT (29). The crystal structures of Form-2 and -3 crystals were determined using the molecular replacement method with the atomic model of SiR from Form-1 crystal and the high resolution X-ray structure of maize Fd (PDB ID: 3B2F) as search models. Crystallographic refinement was performed by the programme REFMAC5 in the CCP4 package (30). Several rounds of refinement calculation and manual model correction were iterated until convergence. Flexible loops and side chains were not included in the final model. The final crystallographic refinement statistics are shown in Table I. Stereochemical geometry of each model was verified by the programme MolProbity (31).

Sire-directed mutagenesis of SiR

Site-specific mutants of SiR were generated with QuickChange site directed mutagenesis kit (Stratagene Corp.). The list of mutants generated in this study and oligonucleotides used for the mutagenesis is shown in Supplementary Table S1. The mutation sites and sequence integrity of coding regions of mutant SiRs were verified by nucleotide sequencing.

Assay of SiR activity

Fd- and methyl viologen (MV)-dependent activities of SiR were assayed with measuring sulfide formation after being converted to cysteine by cysteine synthase (CSase) (32). The reaction mixture included 10 µM Fd or 1 mM MV, 2 mM sodium sulfite, 200 nM SiR, 10 mM O-acetyl serine and 0.4 U CSase in 50 mM Tris-HCl buffer, pH 7.5 containing 100 mM NaCl. The reaction was initiated by adding 20 mM Na₂S₂O₄ and carried out at 30°C for 3 min, and stopped by an addition of trichloroacetic acid to a final concentration of 5% (w/v). After the precipitates were removed, the amount of cysteine formed was determined by an acid ninhydrin reaction (32). Kinetic measurements of Fd-dependent SiR activity were conducted by increasing the concentration of Fd from 0 to 40 µM. All measurements except for the result shown in Figure 2B were repeated at least three times.

Fig. 1

Three structures of the complex of SiR and Fd. (A) Overall structure of the complex of SiR and Fd. Complex AC is obtained from Form-3 crystal, and Complex AE and Complex BF are from Form-2 crystal. All structures are drawn by aligning SiR with the same topology. For details of crystal packing, see Supplementary Figure S1. (B) Relative arrangement of the redox centres in the complexes AC, AE and BF. The final σ-weighted 2|F_O|-|F_C| electron density map around the redox centres were contoured on the stick models at the 1.2σ level. The shortest distance between [4Fe-4S] cluster of SiR and [2Fe-2S] cluster of Fd is shown with dotted line. The backbone structure was also shown as transparent cartoon models. (C) Structure of the interface of the complexes AC, AE and BF. Published results of NMR study on the complex of Fd and SiR are included; amino acid residues of Fd with large chemical shift perturbation upon complex (19) are shown in red (>0.08 ppm) and yellow (0.06–0.08 ppm). Fd is aligned in the same topology.

Fig. 2

Electrostatic surface potentials of SiR and effect of site-directed mutagenesis on activities of SiR. (A) Electrostatic surface potential of SiR on the interaction side with Fd is shown. Positive potential is colored in blue and negative potential in red. [4Fe-4S] cluster of SiR is located at the inner surface of the center region with neutral potential. Seven basic amino acid residues distributed at the interface of the complex of SiR and Fd were substituted to glutamine. Resulting mutants were characterized by measuring Fd (hatched bars)- and MV(open bars)-dependent activities. The assay was repeated at least three times and SD of the obtained values is shown with deviation bars. (B) Seven non-charged amino acid residues distributed at the center region were substituted to glycine or alanine. The resulting mutants were assayed by a single measurement of Fd- and MV-dependent activities.

Results and Discussion

X-ray crystallographic analysis of the complex of SiR and Fd

The phase problem of Form-1 crystal data was successfully solved by applying the Fe-MAD method. The atomic model of SiR was built based on the clear electron density map modified by the density modification and the non-crystallographic averaging. In the crystallographic asymmetric unit of Form-1 crystal, there were two SiRs, designated as molecules A and B, aligned face-to-face; the structures of the polypeptide chain and two prosthetic groups were clearly revealed except for those of a flexible loop (Ser240 and Ala241 for molecule A and Gly236 to Ala241 for molecule B) and the terminal regions (the N-terminal ten residues and the C-terminal nine residues). One phosphate and two magnesium ions added in the crystallization solution were located at the distal side of siroheme and at the molecular surface near the prosthetic groups, respectively. In contrast to the clear interpretation of electron density map for SiR, Fd was hardly modeled by molecular replacement in reciprocal space and rigid body fitting in real space. Although there was a significant quantity of electron density near [4Fe-4S] cluster of SiR in difference map, the density was too noisy and non-continuous to be traced. It was presumed that two Fd molecules occupied this density rich region on the non-crystallographic two fold axis formed by the two SiR molecules. However, two Fds might not be uniformly bound to two SiRs, making the corresponding electron density to be averaged out. Consequently, no further analysis of Form-1 crystal was carried out.

Two new crystals of the complex of SiR and Fd, Form-2 and -3 crystals, prepared under different crystallization conditions as described in materials and methods, were successively solved by molecular replacement method using structure of SiR obtained from Form-1 crystal and PDB data of maize Fd (17) as a search model as described below. According to MolProbity Ramachandran analysis (31), 98, 94 and 97% of the total number of residues of the models in Form-1, -2 and -3 crystals were in the favoured region, respectively.

There were four SiR and two Fd molecules in the asymmetric unit of Form-2 crystal as shown in Supplementary Figure S1, panel of Form-2 crystal. Four molecules of SiR, labelled as molecules A–D, were aligned in tandem, central two, SiR-A and SiR-B, of which formed the complex with Fd, Fd-E and Fd-F, respectively. These two complex structures were aligned side-by-side. Peripheral SiRs, molecules C and D did not associated with Fd. Active site on the siroheme was occupied by chloride ion in all four SiR molecules instead of phosphate ion in Form-1 crystal. Crystallographically invisible regions of SiR and Fd molecules were different in each molecule due to the different packing environment. Judging from the average temperature factor of each molecule, a pair of SiR-A and Fd-E was more stable than that of SiR-B and Fd-F. The former pair of the complex included much longer polypeptides assigned in the model of SiR and Fd (Table I)

In the crystallographic asymmetric unit of Form-3 crystal, there were two SiR, molecules A and B, and one Fd, molecule C (Supplementary Fig. S1, panel of Form-3 crystal). SiR molecules A and B were aligned back-to-back in clear contrast to those in the other two crystal forms. Only SiR-A formed the complex with Fd, Fd-C, and the other SiR did not. This asymmetry was consistent with the average temperature factor of each molecule, SiR-A showed a relatively lower value of 39.201 and SiR-B showed a value of 43.172. Probably due to the crystallization conditions; SiR molecules in Form-3 crystal possessed a bound phosphate ion at the active site as similar as the case of Form-1 crystal.

Comparison of three complex structures of SiR and Fd

Based on the spatial arrangement of SiR with Fd in a unit cell of Form-2 and -3 crystals, three crystal structures could be chosen as candidates of complex of SiR and Fd, which were designated as Complex AC, Complex AE and Complex BF as shown in Figure 1A. Topological relationship of SiR and Fd was distinct in each individual complex. Compared to the Complex AE, Fd in Complex AC was rotated over 90° relative to SiR along an axis between [2Fe-2S] cluster and [4Fe-4S] cluster. Complex BF and Complex AE, both present in Form-2 crystal, were rather similar and Fd was bent by 10° relative to SiR. In all complexes, Fd biding regions were present inside a concave region of SiR, where [4Fe-4S] cluster is partially exposed to the solvent. It was noted that distances of [2Fe-2S] cluster of Fd and [4Fe-4S] cluster of SiR were quite similar in the three complexes (Fig. 1B). A close vicinity of the two redox centres with a distance of 12–13 Å seems to provide an advantageous situation for direct electron transfer between two redox centres in any of the three complex structures. [4Fe-4S] cluster was ligated with four cysteine residues, Cys494, Cys500, Cys540 and Cys544 and Cys544 bridged non-heme iron and siroheme iron as reported in bacterial SiRHPs (9, 33), facilitating the intramolecular electron transfer from [4Fe-4S] cluster to siroheme. No significant change in the main chain fold of SiR was observed between Fd-bound Form-2, 3 and unbound Form-1. In the case of the complex of FNR and Fd, an appreciable movement between FAD-binding and NADP-binding domains of FNR was induced upon complex formation with Fd (15, 21).

We have previously determined negatively charged surface of Fd involved in the interaction with SiR by 2D NMR spectroscopy (19). Interaction resides of ¹⁵N uniformly-labelled Fd, which showed large chemical shift perturbation of ¹H-¹⁵N heteronuclear single-quantum coherence correlation spectra upon forming a complex with SiR in solution, were well mapped onto the interfaces of the three crystal structures (Fig. 1C). It is hypothesized, at present, that SiR and Fd would interact each other probably in a non-unified and multiple fashion depending on the crystallization conditions. In order to gain further insights into the correlation of these multiple complex structures in crystals with those in solution, in-depth NMR studies of isotope-labelled SiR would be necessary.

Mutagenesis of basic and non-charged residues of SiR located at the interface of the complex

Three structures of the complex of SiR and Fd determined in this study gave important pieces of information on the protein–protein interaction. The surface of SiR at the interface was surrounded by a wide area with positive electrostatic potential and the center region of the interface, where [4Fe-4S] cluster was located, was covered by a non-charged narrow area as shown in Figure 2A and B. In order to investigate which area of the positively charged surface of SiR is crucial for formation of the electron transfer complex with Fd, site-directed mutagenesis of seven basic residues of SiR, which are distributed at the various region of the interface, was carried out. Lys66, Arg111, Arg114, Lys117, Arg324, Lys582 and Lys584 were separately substituted to glutamine and effects of the mutations on the activity of SiR were examined (Fig. 2A). The mutant SiRs were purified and showed absorption spectra comparable with wild-type enzyme (Supplementary Fig. S2), indicative of no significant effect on the stability or structural integrity. All mutants showed a remarkably lowered activity when Fd was used as an electron donor, while no such significant decrease was observed in the MV-dependent activity. This data clearly showed that almost all of the basic residues are crucial for an efficient electron acceptance from reduced Fd, but not for enzyme activity itself.

Next, seven non-charged residues with a large side chains at the center region of the interface were substituted to glycine or alanine and resulting mutants were assayed as earlier (Supplementary Fig. S2 and Fig. 2B). The results revealed that the mutations gave no significant effects on either Fd-dependent or MV-dependent activity, indicating that the non-charged residues located between the redox centers of the two proteins is dispensable at the level of single mutations. Thus, the electron transfer between redox centers of Fd to SiR with a distance of around 12 Å might not need any specific contribution of side chains of amino acids located near the two prosthetic groups, implying an intermolecular electron transfer by through-space mechanism.

Kinetic and gel-filtration analyses of the affinity of mutant SiRs with Fd

Kinetic properties of the mutant SiRs, which basic residues in a close vicinity with acidic residues of Fd were mutated (Fig. 3A and B), were further analysed by Fd-dependent assay in the presence of increasing concentrations of Fd as shown in Figure 3C. Double mutants, R111Q/R114Q and K582Q/K584Q and single mutants, R324Q, exhibited extremely low activity. K582Q and K584Q decreased activity to some extent compared to wild type (WT). These kinetical data suggested that the decreased ability of mutant SiRs in the Fd-dependent activity was due to weakened affinity of mutant SiRs for Fd.

Fig. 3

Analysis of SiR mutants with Fd-dependent activity decreased. (A,B) Electrostatic interaction at the interface of Complex AC is shown. Basic residues of SiR are in a close proximity with acidic residues of Fd as follows; R111 and R114 versus D60, R324 versusD34, K582 versus E29 and K584 versus E30. (C) Kinetic analysis of Fd-dependent activity of mutant SiRs. The activity was measured in the presence of increasing concentrations of Fd from 0 to 40 µM as described in ‘Materials and Methods’ section. (D) A size-exclusion chromatography of an equimolar mixture of SiR and Fd. A mixture of 100 µM SiR (WT or mutants indicated) and 100 µM Fd was subjected to a column of Superdex 75 and developed with 50 mM Tris-HCl, pH 7.5 containing 25 mM NaCl with monitoring absorbance at 330 nm. The complex of SiR WT and Fd was eluted as a major single peak at an elution time of 19.8 min, faster than free SiR (20.2 min) and free Fd (23.2 min). Mutant SiRs and Fd were separately eluted at the times corresponding to each of their free forms.

In this context, therefore, physical complex formation between wild type and mutant SiRs with Fd was investigated using a size exclusion chromatography as described previously (20). When a mixture of Fd and wild-type SiR in a molar ratio of 1:1 was loaded, SiR and Fd were eluted as a single major peak (Fig. 3D), and further addition of Fd over the 1:1 ratio was resulted in an elution of the excess Fd as a separate peak (data not shown), confirming the stoichiometric formation of the complex of SiR and Fd. The equivalent gel filtration of R111, R324Q and K582Q/K584Q with Fd showed that these mutations abolished the ability of SiR to form the stable complex with Fd (Fig. 3D). These combined data demonstrated that almost all of the basic amino acid residues of SiR at the interface of the complexes of SiR and Fd are generally contributed to the protein-protein interaction which governs an efficient acceptance of reducing equivalents from reduced Fd.

Conclusion

We have determined three structures of the complex of plant SiR and Fd. A significant difference in the orientation of the two proteins is seen among the three structures, but such difference does not result in a drastic change in topology and distance of prosthetic groups of two proteins. Therefore, we presume these multiple structures are not an artifact due to crystal packing, but would be present in solution under some equilibrium conditions. For example, binding of substrate or even small ligand compound to siroheme might cause a differential interaction of SiR and Fd.

Three complex structures of FNR and Fd from different origins, maize leaf, maize root and cyanobacterium, have been determined and they differ significantly each other (15). It is tempting to assume that such structural difference might result from a rational reason, such that each complex plays a predominant role in Fd-dependent redox metabolisms in their origin specific manner. This work would give an important example that Fd could interact with one certain partner enzyme in a multiple fashion. Apparently, more detailed studies are needed to understand precise mechanisms of the multi-functionality of Fd.

Abbreviations

Abbreviations
 
• CSase

  cysteine synthase

 
• Fd

  ferredoxin

 
• FNR

  ferredoxin:NADPH oxidoreductase

 
• FP

  flavoprotein

 
• FTR

  ferredoxin:thioredoxin reductase

 
• HP

  hemoprotein

 
• HSQC

  heteronuclear single-quantum coherence correlation

 
• MV

  methyl viologen

 
• MAD

  multi-wavelength anomalous dispersion

 
• SiR

  sulfite reductase

Funding

This work was supported in part by grants-in-aid from the Ministry of Culture, Education, Science and Sports of Japan (J.Y.K., G.K. and T.H.), the Funding Program for Next Generation World-Leading Researchers (GS016) from the Cabinet Office of Japan (G.K.) and CREST, Japan Science and Technology Agency (G.K.).

Conflict of Interest

None declared.

Author notes

*These authors contributed equally to this work.

The atomic coordinates and structure factors (accession code 5H8V, 5H8Y and 5H92 for the structure of Form-1, -2 and -3 crystals) have been deposited in the Protein Data Bank Japan, a member of the worldwide PDB (http://www.pdbj.org).