- Email: [email protected]
Structure of the cytochrome complex SoxXA of Paracoccus pantotrophus, a heme enzyme initiating chemotrophic sulfur oxidation
Journal of
Structural Biology Journal of Structural Biology 152 (2005) 229–234 www.elsevier.com/locate/yjsbi
Short communication
Structure of the cytochrome complex SoxXA of Paracoccus pantotrophus, a heme enzyme initiating chemotrophic sulfur oxidation Tresfore Dambe a,c,1, Armin Quentmeier b, Dagmar Rother b,c, Cornelius Friedrich b, Axel J. Scheidig a,c,* a
Max-Planck-Institut fu¨r Molekulare Physiologie, Abteilung Physikalische Biochemie, Otto-Hahn-Strasse 11, D-44225 Dortmund, Germany Lehrstuhl Technische Mikrobiologie, Fachbereich Bio- und Chemieingenieurwesen, Universita¨t Dortmund, D-44221 Dortmund, Germany c Department of Structural Biology, Saarland University, Universitaetsklinikum, Geb. 60 D-66421 Homburg/Saar, Germany
b
Received 13 July 2005; received in revised form 12 September 2005; accepted 12 September 2005 Available online 2 November 2005
Abstract The sulfur-oxidizing enzyme system (Sox) of the chemotroph Paracoccus pantotrophus is composed of several proteins, which together oxidize hydrogen sulfide, sulfur, thiosulfate or sulfite and transfers the gained electrons to the respiratory chain. The hetero-dimeric cytochrome c complex SoxXA functions as heme enzyme and links covalently the sulfur substrate to the thiol of the cysteine-138 residue of the SoxY protein of the SoxYZ complex. Here, we report the crystal structure of the c-type cytochrome complex SoxXA. The structure ˚ identifying the axial heme–iron coordination involving an could be solved by molecular replacement and refined to a resolution of 1.9 A unusual Cys-251 thiolate of heme2. Distance measurements between the three heme groups provide deeper insight into the electron transport inside SoxXA and merge in a better understanding of the initial step of the aerobic sulfur oxidation process in chemotrophic bacteria. 2005 Elsevier Inc. All rights reserved. Keywords: c-type cytochrome; SoxXA complex; Sulfur oxidation; X-ray crystal structure
1. Introduction Oxidation of reduced inorganic sulfur compounds like hydrogen sulfide, sulfur or thiosulfate to sulfuric acid represents one half of the global sulfur cycle and is mainly performed by specialized prokaryotes. The ability to oxidize inorganic sulfur compounds is found in aerobic chemotrophic and anaerobic phototrophic bacteria. Electrons derived from oxidation of inorganic sulfur compounds are used by chemotrophic and phototrophic bacteria for autotrophic carbon dioxide reduction (Brune, 1989; Friedrich et al., 2001; Kelly et al., 1997; Tru¨per and Fischer, 1982). The Sox enzyme system has been studied *
1
Corresponding author. Fax: +49 6841 1626251. E-mail address: [email protected] (Axel J. Scheidig). Present address: PSF Biotech AG, D-14059 Berlin, Germany.
1047-8477/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2005.09.002
in detail and in vitro assays of a reconstituted enzyme system have been described for the aerobic chemotroph Paracoccus versutus and Paracoccus pantotrophus (reviewed in Friedrich et al., 2005; Kelly et al., 1997). Four periplasmic proteins reconstitute the Sox enzyme system of P. pantotrophus that oxidizes hydrogen sulfide, sulfur, thiosulfate, and sulfite. The oxidation of thiosulfate yields eight electrons which are transferred in vitro to a small monoheme cytochrome c (Rother et al., 2001). The central protein within this enzyme system is the SoxYZ complex. In a first reaction, thiosulfate is covalently bound to the thiol of a conserved cysteine residue at the carboxy-terminal end of the SoxY subunit (Quentmeier and Friedrich, 2001) through interaction with the c-type cytochrome complex SoxXA yielding cysteine-S-thiosulfonate. However, the precise mechanism of this reaction is unknown. In a second step, the dimanga-
230
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
nese SoxB protein interacts with the modified SoxYZ (Quentmeier et al., 2003) and is proposed to function as sulfate thiohydrolase to release sulfate from the cysteineS-thiosulfonate to yield cysteine persulfide (Quentmeier and Friedrich, 2001). In a final step, the hetero-tetrameric molybdoprotein cytochrome c complex sulfur dehydrogenase Sox(CD)2 oxidizes the outer sulfane of the cysteine persulfide of SoxYZ to cysteine-S-sulfonate (Bardischewsky et al., 2005) and sulfate is again released through interaction with SoxB (Friedrich et al., 2001). According to this model SoxXA delivers two and the sulfur dehydrogenase Sox(CD)2 six electrons to horse heart cytochrome c in vitro (Bardischewsky et al., 2005; Rother et al., 2001) per molecule of thiosulfate oxidized to sulfate. The Sox system of P. pantotrophus does not only oxidize thiosulfate but also hydrogen sulfide, sulfur (or polysulfide), and sulfite (Rother et al., 2001). This substrate versatility is surprising as these inorganic sulfur compounds are neither isosteric nor isoelectronic and differ in redox potential and reactivity. This substrate versatility is directly linked to the initial reaction which is proposed to be performed by the coordinated action of SoxXA and SoxYZ. This implies a high flexibility of both proteins for the acceptance of the various sulfur substrates that differ substantially in mass, pKa, electron density, and reactivity (Friedrich et al., 2000; Rother et al., 2001). The c-type cytochrome complex SoxXA (45 kDa) is composed of the 16 kDa monoheme subunit SoxX and the 29 kDa diheme subunit SoxA (Friedrich et al., 2000; Rother and Friedrich, 2002). With respect to the primary structure, closely related SoxXA cytochrome complexes have been identified from other chemotrophic (Kappler et al., 2004; Mukhopadhyaya et al., 2000) and phototrophic bacteria (Appia-Ayme et al., 2001; Friedrich et al., 2001). The hemes of SoxA are designated according to the appearance of the CxxCH motif in the amino acid sequence, to which the heme is covalently connected, heme1 and heme2 (Bamford et al., 2002; Cheesman et al., 2001). The heme group of SoxX is designated heme3. Most SoxA proteins are diheme cytochromes (heme1 and heme2) while SoxA of Starkeya novella does not contain heme1 (Kappler et al., 2004). SoxX proteins contain a single heme (heme3). As SoxXA from S. novella and P. pantotrophus appear to have an identical function, this observation suggests that heme1 of SoxXA does not play a significant role in sulfur oxidation. Rhodovulum sulfidophilum is a marine phototrophic alpha proteobacterium which grows anaerobically with thiosulfate. This strain harbors a sox gene cluster and forms the cytochrome complex, designated SoxAX. The diheme subunit SoxA of R. sulfidophilum has an unusual axial coordination of heme2–iron by a thiolate, and SoxAX was suggested to function as sulfurtransferase (Bamford et al., 2002). Although the sox genetic determinants of the phototrophic R. sulfidophilum show significant identities to those of the chemotrophic P. pantotrophus (Appia-
Ayme and Berks, 2002; Appia-Ayme et al., 2001) no data are available on the biochemistry of the Sox system from R. sulfidophilum. For a detailed comparison of the biochemical data, the structures of the sox gene products for both organisms are desirable. Here, we report the X-ray crystal structure of the SoxXA complex of P. pantotrophus at a resolution of ˚ . The electron density map resolves a cysteine persul1.9 A fide for axial coordination of the heme2–iron of SoxA proposed as the active site located between the SoxX and SoxA subunits accessible from the solvent in a deep pocket. 2. Crystallization and structure determination Paracoccus pantotrophus was cultivated chemolithoautotrophically with thiosulfate as an energy source. The SoxXA complex was purified from these cells to homogeneity as described previously (Friedrich et al., 2000). For crystallization by the hanging drop vapour diffusion method, equal volumes (each 1 ll) of reservoir solution (100 mM MES, pH 6.5, 4–7% PEG 1500, 4–8 mM ZnSO4, and 75 mM [Co(NH3)6]Cl3) and protein solution (6 mg/ml SoxXA in 10 mM Bis–Tris, pH 6.5, 1 mM MgSO4, and 0.1 mM Na2S2O3) were mixed and equilibrated against 500 lL reservoir solution at room temperature. Thin plates of SoxXA crystals appeared after 3 days with the dimension of up to 700 · 400 · 50 lm3. All X-ray diffraction data were collected at 100 K. For this purpose, 5 lL cryo-solution (100 mM MES, pH 6.5, 7% PEG 1500, 4 mM ZnSO4, and 25% glycerol) was added to the crystallization drop. After incubation for several minutes, the crystals were mounted in nylon loops (Hampton Research) and flash frozen in liquid nitrogen. Data collection was carried out at beamlines BW6 (MPG) and BW7 (EMBL), respectively, at DESY (Deut sches Electronen Synchrotron, Hamburg, Germany) and at the X06SA beamline at SLS (Swiss Light Source, Villigen, Switzerland). Data were processed and scaled using the DENZO/SCALEPACK (Otwinowski and Minor, 1997) or XDS (Kabsch, 1993) program package. The first approach to determine the phases using the anomalous signal of iron failed. Whereas the position of the iron atoms could be well located, the anomalous signal was insufficient to solve the complete protein structure. Therefore, we used the structure of SoxAX of R. sulfidophilum [PDB entry 1H33 (Bamford et al., 2002)] for molecular replacement [program MOLREP (Collaborative Computational Project Number 4, 1994)] to obtain initial phases. First refinement steps using simulated annealing and energy minimization were carried out with the program package CNS (Brunger et al., 1998). For a final refinement at high resolution the program REFMAC5 from the CCP4 program package (Collaborative Computational Project Number 4, 1994) was used. After each round of refinement the model was checked and rebuilt with the program O (Jones et al., 1991). Final model analysis, imaging, and
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
ray tracing were performed using the program PyMOL (DeLano, 2002). 3. Overall structure SoxXA from P. pantotrophus crystallizes in the space ˚ , b = 180.0 A ˚, group P2(1) with cell parameters a = 42.9 A ˚ c = 117.9 A and b = 92.8. The model could be refined at ˚ resolution to Rcryst and Rfree values of 15.8 and 1.9 A 21.1%, respectively, with reasonable r.m.s.d. values for bond length and bond angles (Table 1). The final model comprises four hetero-dimeric complexes SoxXA per asymmetric unit. In each complex the residues 27–290 of SoxA and 21–157 of SoxX, respectively, could be traced in the electron density (lacking the leader peptides for secretion into the periplasma). The most favoured region of the Ramachandran plot covers 91.7% of the residues, and no residues are located within disallowed regions (Table 1). The approximate dimensions of the hetero-dimeric ˚ 3. This complex contains three enzyme are 34 · 40 · 77 A almost globular cytochrome c domains which are arranged in a linear fashion leading to its stretched form of which Table 1 Data collection and refinement statistics Data collection statistics Beamline Temperature (K) Detector ˚) Wavelength (A Space group Cell dimensions ˚) a (A ˚) b (A ˚) c (A b () ˚) Highest resolution (A Highest resolution shell Number of recorded reflections Redundancy Rsym (%)a,b I/r (I)a ˚ 2) Wilson B value (A Refinement statistics ˚) Resolution range (A Number of unique reflections Completeness (%)a Rcryst/Rfree (%)c,d Ramachandran plot Most favoured, favoured, allowed (%) ˚) r.m.s.d. on bond length (A r.m.s.d. on bond angles () a
BW6/DESY (Hamburg) 100 MarCCD 1.737 P21 42.9 180.0 117.9 92.8 1.9 2.0–1.9 631,082 5.2 8.5 (15.2) 12.6 (5.9) 23.8 19.5–1.9 120,519 93.7 (62.6) 15.8/21.1 91.7, 7.7, 0.6 0.014 1.5
Values in parentheses are for the high-resolution bin. P P P P Rsym ¼ 100 h i jI i ðhÞ hI i ðhÞij= h i I i ðhÞ where Ii(h) is the ith measurement and ÆIi(h)æis the mean of all measurements of I(h) for Miller indices hkl. P P c R ¼ ðkF obs j kjF calc kÞ= jF obs j where Fobs and Fcalc are observed and calculated structure factor amplitudes, respectively. d Rfree value is the R value obtained for a test set of reflections, consisting of a randomly selected 5% subset of the diffraction data not used during refinement. b
231
SoxX contributes one domain and SoxA two domains (Fig. 1). Each of these domains contains one heme linked to a CxxCH motif characteristic for c-type cytochromes. The imidazole ring of the histidine residue provides an axial ligand of the iron atom of the heme. The two axial ligands of heme1 are His-110 and Cys-143 while heme2 is coordinated by His-210 and Cys-251. At the Cys-251 residue additional electron density is observed, which is described below (see Section 4). Heme3 of SoxX is coordinated by His-65 and a conserved Met-111 as previously suggested from sequence alignments (Cheesman et al., 2001; Rother and Friedrich, 2002) (Fig. 2). The first heme domain of the P. pantotrophus SoxA, SoxAN, is close to the amino terminus (residues 82 and 181) and the second heme domain, SoxAC, is located towards the carboxy terminus (residues 182–280). The N-terminal region of SoxA (residue 27–81) does not display any known domain motif, but is in contact with the SoxAC domain and may stabilize this special fold of a diheme cytochrome c, as similarly observed for the respective complex of R. sulfidophilum (Bamford et al., 2002). The structures of the two cytochrome c domains of SoxA are closely related with a mean ˚ [60% matching Ca-atr.m.s.d. on Ca-atoms of around 1.5 A oms calculated with the program MAPS (Zhang et al., 2003)]. The arrangement of SoxAN and SoxAC includes a pseudo-twofold axis oriented between the two domains (Fig. 1). Sequence analysis demonstrates a high homology in primary structure to SoxAX of R. sulfidophilum with an identity of 53.4% (SoxA) and of 44.4% (SoxX), respectively. A structural homology analysis using the program DALI (Holm and Sander, 1993) is summarized in Table 2 and confirms the high homology in tertiary and quaternary structure of SoxXA of P. pantotrophus and SoxAX of R. sulfidophilum. Structural similarities to other proteins with known structure were not detected. 4. Active site The kinetics of the reconstituted Sox enzyme system (Rother et al., 2001) suggested that SoxXA functions as a heme enzyme and covalently links the sulfur substrate to the thiol of Cys-138 in SoxY (Friedrich et al., 2001). However, this first step of the reaction cycle is not understood in detail. The electron density map of SoxA of P. pantotrophus shows an electron density between the thiol of Cys-251 and the iron atom of heme2 (Fig. 3). This density fits well with a sulfur atom to give cysteine persulfide at position 251 (termed Css-251). The geometry of cysteine persulfide is almost identical with cysteine sulfenic acid (Cys-S-OH). However, the observed additional electron density is insufficient for an oxygen atom to allow interpretation as a cysteine sulfenic acid. This analysis is in accordance with that of SoxA of R. sulfidophilum and suggested as a post-translational modification (Bamford et al., 2002; Cheesman et al., 2001). The primary struc-
232
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
Fig. 1. Stereo representation of SoxXA. SoxXA is a hetero-dimer consisting of SoxX (blue) and SoxA. SoxA consists of two cytochrome c domains. The N-terminal SoxAN domain (orange) and the C-terminal SoxAC domain (green). SoxAN and SoxAC are related to each other by a twofold rotation axis (indicated with ¤). The N- and the C-terminus (both grey) of SoxA are not part of the typical cytochrome c domain. The heme groups are represented as sticks and are coloured according to their affiliation to the respective cytochrome c domain.
Fig. 2. Sequence alignment of SoxXA of P. pantotrophus (Pp) with SoxAX of R. sulfidophilum (Rs). On top of the sequence the secondary stucture of SoxXA from P. pantotrophus is shown. The heme binding CxxCH-motifs are assigned with green bars. The second axial ligand of heme1 and heme3 are assigned with (m) and the active site residue Cys251 is assigned with (q).
tures of SoxA of both organisms are 53.4% identical. Both proteins have an almost identical tertiary structure, both proteins appear to have the same novel axial coordination of the heme2–iron by cysteine persulfide and for both proteins heme1 appears to be EPR silent (Cheesman et al., 2001; P. Hellwig, E. Reijerse, M. Sommerhalter, unpublished data). These striking identities strongly
suggest that the first step is identical in aerobic chemotrophic and anaerobic phototrophic sulfur oxidation. However, at present no reconstituted enzyme system has been described and no kinetic data are available from phototrophic bacteria. Omission of the sulfur dehydrogenase Sox(CD)2 from the reconstituted Sox enzyme system which then contains
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
233
Fig. 3. Phase weighted electron density distribution (2mFo-DFc) around the proposed active site at heme2. (A) The electron density is not sufficiently described by the cysteine residue at position 251. (B) A post-translational modification of the cysteine at position 251 to a cysteine persulfide describes the electron density well.
Table 2 Comparison of the structures of SoxXA of P. pantotrophus and SoxAX from R. sulfidofilum with the program DALI (Holm and Sander, 1993)
SoxA SoxX
Z-Score
LALIa
r.m.s.d.b
%IDEc
41.5 23.3
259 133
0.9 1.0
56 49
a
Total number of equivalenced residues. Positional root mean square deviation of superimposed Ca atoms in angstroms. c Percentage of sequence identity over equivalenced positions. b
only SoxXA, SoxYZ, and SoxB of P. pantorophus allows oxidation of hydrogen sulfide, thiosulfate, and sulfite albeit with a yield of two moles of electrons per mole of substrate. Based on this reconstitution the two electron-yielding step has to be assigned to SoxXA, as SoxB is not redox active and SoxYZ does not contain a cofactor or metal. From the active site at heme2, the electrons could be transferred to a second heme group and further transmitted to the final electron acceptor. In an alternative mechanism the two electrons are stored temporarily in SoxXA at two different heme groups as each heme group is just able to store one electron. After SoxXA released the substrates, the electrons are likely to be transmitted via heme2 to the final electron acceptor. Two ways for the electron transfer from heme2 are possible. The first route goes to heme1, the second to ˚ ) a tunnelling of heme3. Over short distances (up to 14 A electrons through proteins without further co-factors has been described for different proteins (Page et al., 1987). ˚ between heme1 and heme2 With a distance of around 25 A it is unlikely, that the transport of an electron can take place without an additional co-factor. It is more likely, that the electrons are transferred to or stored within the heme groups of SoxX (distance heme2–heme3: approximately ˚ ). This interpretation is further supported by the 11.3 A comparison of SoxA of P. pantotrophus with the corresponding protein of Starkeya novella. SoxA of S. novella and P. pantotrophus are homologues and supposed to catalyze the same reaction. However, SoxA from S. novella is missing heme1 in its SoxAN domain and contains a protein disulfide instead (Kappler et al., 2004).
Acknowledgments We thank Sabine Vogt, Bjo¨rn Klink, and Georg Holtermann for technical assistance and our colleagues at the Max-Planck-Institut fu¨r molekulare Physiologie for helpful suggestions and discussions. We thank Gleb P. Bourenkov and Hans D. Bartunik for their assistance during the collection of the MAD data set at beamline BW6 and M. Weiss, P. Tucker, and E. Pohl for assistance during beamtime at beamlines BW7A/B (DESY, Hamburg, Germany), E. Mitchell and A. McCarthy for assistance in using beamline ID14-1 (ESRF, Grenoble, France). Part of this work was performed at the Swiss Light Source, Paul Scherrer Institut (Villigen, Switzerland). We would like to thank C. Schulze-Briese and T. Tomizaki for assistance. We gratefully acknowledge support by grants to A.J.S. (Sche 545/3-1) and C.G.F. (Fr 318/8-1) from the Deutsche Forschungsgemeinschaft. References Appia-Ayme, C., Berks, B.C., 2002. SoxV, an orthologue of the CcdA disulfide transporter, is involved in thiosulfate oxidation in Rhodovulum sulfidophilum and reduces the periplasmic thioredoxin SoxW. Biochem. Biophys. Res. Commun. 296, 737–741. Appia-Ayme, C., Little, P.J., Matsumoto, Y., Leech, A.P., Berks, B.C., 2001. Cytochrome complex essential for photosynthetic oxidation of both thiosulfate and sulfide in Rhodovulum sulfidophilum. J. Bacteriol. 183, 6107–6118. Bamford, V.A., Bruno, S., Rasmussen, T., Appia-Ayme, C., Cheesman, M.R., Berks, B.C., Hemmings, A.M., 2002. Structural basis for the oxidation of thiosulfate by a sulfur cycle enzyme. EMBO J. 21, 5599– 5610. Bardischewsky, F., Quentmeier, A., Hellwig, P., Rother, D., Kostka, S., Friedrich, C.G., 2005. Sulfur dehydrogenase of Paracoccus pantotrophus: the heme-2 domain of the molybdo-protein cytochrome c complex is dispensable for catalytic activity. Biochemistry 44, 7024– 7034. Brune, D.C., 1989. Sulfur oxidation by phototrophic bacteria. Biochim. Biophys. Acta 975, 189–221. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T., Warren, G.L., 1998. Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905–921.
234
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
Cheesman, M.R., Little, P.J., Berks, B.C., 2001. Novel heme ligation in a c-type cytochrome involved in thiosulfate oxidation: EPR and MCD of SoxAX from Rhodovulum sulfidophilum. Biochemistry 40, 10562– 10569. Collaborative Computational Project Number 4, 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50, 760–763. DeLano, W.L., 2002. DeLano Scientific. San Carlos, CA, USA,
Mukhopadhyaya, P.N., Deb, C., Lahiri, C., Roy, P., 2000. A soxA gene, encoding a diheme cytochrome c, and a sox locus, essential for sulfur oxidation in a new sulfur lithotrophic bacterium. J. Bacteriol. 182, 4278–4287. Otwinowski, Z., Minor, W., 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. Page, E., Goings, G.E., Power, B., Upshaw-Earley, J., 1987. Tunneling cell processes in myocytes of stretched mouse atria. Am. J. Physiol. 253, H432–H443. Quentmeier, A., Friedrich, C.G., 2001. The cysteine residue of the SoxY protein as the active site of protein-bound sulfur oxidation of Paracoccus pantotrophus GB17. FEBS Lett. 503, 168–172. Quentmeier, A., Hellwig, P., Bardischewsky, F., Grelle, G., Kraft, R., Friedrich, C.G., 2003. Sulfur oxidation in Paracoccus pantotrophus: interaction of the sulfur-binding protein SoxYZ with the dimanganese SoxB protein. Biochem. Biophys. Res. Commun. 312, 1011– 1018. Rother, D., Friedrich, C.G., 2002. The cytochrome complex SoxXA of Paracoccus pantotrophus is produced in Escherichia coli and functional in the reconstituted sulfur-oxidizing enzyme system. Biochim. Biophys. Acta 1598, 65–73. Rother, D., Henrich, H.J., Quentmeier, A., Bardischewsky, F., Friedrich, C.G., 2001. Novel genes of the sox gene cluster, mutagenesis of the flavoprotein SoxF, and evidence for a general sulfur-oxidizing system in Paracoccus pantotrophus GB17. J. Bacteriol. 183, 4499– 4508. Tru¨per, H.G., Fischer, U., 1982. Anaerobic oxidation of sulphur compounds as electron donors for bacterial photosynthesis. Philos. Trans. R. Soc. Lond. B Biol. Sci. 298, 529–542. Zhang, Z., Lindstam, M., Unge, J., Peterson, C., Lu, G., 2003. Potential for dramatic improvement in sequence alignment against structures of remote homologous proteins by extracting structural information from multiple structure alignment. J. Mol. Biol. 332, 127–142.
Structural Biology Journal of Structural Biology 152 (2005) 229–234 www.elsevier.com/locate/yjsbi
Short communication
Structure of the cytochrome complex SoxXA of Paracoccus pantotrophus, a heme enzyme initiating chemotrophic sulfur oxidation Tresfore Dambe a,c,1, Armin Quentmeier b, Dagmar Rother b,c, Cornelius Friedrich b, Axel J. Scheidig a,c,* a
Max-Planck-Institut fu¨r Molekulare Physiologie, Abteilung Physikalische Biochemie, Otto-Hahn-Strasse 11, D-44225 Dortmund, Germany Lehrstuhl Technische Mikrobiologie, Fachbereich Bio- und Chemieingenieurwesen, Universita¨t Dortmund, D-44221 Dortmund, Germany c Department of Structural Biology, Saarland University, Universitaetsklinikum, Geb. 60 D-66421 Homburg/Saar, Germany
b
Received 13 July 2005; received in revised form 12 September 2005; accepted 12 September 2005 Available online 2 November 2005
Abstract The sulfur-oxidizing enzyme system (Sox) of the chemotroph Paracoccus pantotrophus is composed of several proteins, which together oxidize hydrogen sulfide, sulfur, thiosulfate or sulfite and transfers the gained electrons to the respiratory chain. The hetero-dimeric cytochrome c complex SoxXA functions as heme enzyme and links covalently the sulfur substrate to the thiol of the cysteine-138 residue of the SoxY protein of the SoxYZ complex. Here, we report the crystal structure of the c-type cytochrome complex SoxXA. The structure ˚ identifying the axial heme–iron coordination involving an could be solved by molecular replacement and refined to a resolution of 1.9 A unusual Cys-251 thiolate of heme2. Distance measurements between the three heme groups provide deeper insight into the electron transport inside SoxXA and merge in a better understanding of the initial step of the aerobic sulfur oxidation process in chemotrophic bacteria. 2005 Elsevier Inc. All rights reserved. Keywords: c-type cytochrome; SoxXA complex; Sulfur oxidation; X-ray crystal structure
1. Introduction Oxidation of reduced inorganic sulfur compounds like hydrogen sulfide, sulfur or thiosulfate to sulfuric acid represents one half of the global sulfur cycle and is mainly performed by specialized prokaryotes. The ability to oxidize inorganic sulfur compounds is found in aerobic chemotrophic and anaerobic phototrophic bacteria. Electrons derived from oxidation of inorganic sulfur compounds are used by chemotrophic and phototrophic bacteria for autotrophic carbon dioxide reduction (Brune, 1989; Friedrich et al., 2001; Kelly et al., 1997; Tru¨per and Fischer, 1982). The Sox enzyme system has been studied *
1
Corresponding author. Fax: +49 6841 1626251. E-mail address: [email protected] (Axel J. Scheidig). Present address: PSF Biotech AG, D-14059 Berlin, Germany.
1047-8477/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2005.09.002
in detail and in vitro assays of a reconstituted enzyme system have been described for the aerobic chemotroph Paracoccus versutus and Paracoccus pantotrophus (reviewed in Friedrich et al., 2005; Kelly et al., 1997). Four periplasmic proteins reconstitute the Sox enzyme system of P. pantotrophus that oxidizes hydrogen sulfide, sulfur, thiosulfate, and sulfite. The oxidation of thiosulfate yields eight electrons which are transferred in vitro to a small monoheme cytochrome c (Rother et al., 2001). The central protein within this enzyme system is the SoxYZ complex. In a first reaction, thiosulfate is covalently bound to the thiol of a conserved cysteine residue at the carboxy-terminal end of the SoxY subunit (Quentmeier and Friedrich, 2001) through interaction with the c-type cytochrome complex SoxXA yielding cysteine-S-thiosulfonate. However, the precise mechanism of this reaction is unknown. In a second step, the dimanga-
230
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
nese SoxB protein interacts with the modified SoxYZ (Quentmeier et al., 2003) and is proposed to function as sulfate thiohydrolase to release sulfate from the cysteineS-thiosulfonate to yield cysteine persulfide (Quentmeier and Friedrich, 2001). In a final step, the hetero-tetrameric molybdoprotein cytochrome c complex sulfur dehydrogenase Sox(CD)2 oxidizes the outer sulfane of the cysteine persulfide of SoxYZ to cysteine-S-sulfonate (Bardischewsky et al., 2005) and sulfate is again released through interaction with SoxB (Friedrich et al., 2001). According to this model SoxXA delivers two and the sulfur dehydrogenase Sox(CD)2 six electrons to horse heart cytochrome c in vitro (Bardischewsky et al., 2005; Rother et al., 2001) per molecule of thiosulfate oxidized to sulfate. The Sox system of P. pantotrophus does not only oxidize thiosulfate but also hydrogen sulfide, sulfur (or polysulfide), and sulfite (Rother et al., 2001). This substrate versatility is surprising as these inorganic sulfur compounds are neither isosteric nor isoelectronic and differ in redox potential and reactivity. This substrate versatility is directly linked to the initial reaction which is proposed to be performed by the coordinated action of SoxXA and SoxYZ. This implies a high flexibility of both proteins for the acceptance of the various sulfur substrates that differ substantially in mass, pKa, electron density, and reactivity (Friedrich et al., 2000; Rother et al., 2001). The c-type cytochrome complex SoxXA (45 kDa) is composed of the 16 kDa monoheme subunit SoxX and the 29 kDa diheme subunit SoxA (Friedrich et al., 2000; Rother and Friedrich, 2002). With respect to the primary structure, closely related SoxXA cytochrome complexes have been identified from other chemotrophic (Kappler et al., 2004; Mukhopadhyaya et al., 2000) and phototrophic bacteria (Appia-Ayme et al., 2001; Friedrich et al., 2001). The hemes of SoxA are designated according to the appearance of the CxxCH motif in the amino acid sequence, to which the heme is covalently connected, heme1 and heme2 (Bamford et al., 2002; Cheesman et al., 2001). The heme group of SoxX is designated heme3. Most SoxA proteins are diheme cytochromes (heme1 and heme2) while SoxA of Starkeya novella does not contain heme1 (Kappler et al., 2004). SoxX proteins contain a single heme (heme3). As SoxXA from S. novella and P. pantotrophus appear to have an identical function, this observation suggests that heme1 of SoxXA does not play a significant role in sulfur oxidation. Rhodovulum sulfidophilum is a marine phototrophic alpha proteobacterium which grows anaerobically with thiosulfate. This strain harbors a sox gene cluster and forms the cytochrome complex, designated SoxAX. The diheme subunit SoxA of R. sulfidophilum has an unusual axial coordination of heme2–iron by a thiolate, and SoxAX was suggested to function as sulfurtransferase (Bamford et al., 2002). Although the sox genetic determinants of the phototrophic R. sulfidophilum show significant identities to those of the chemotrophic P. pantotrophus (Appia-
Ayme and Berks, 2002; Appia-Ayme et al., 2001) no data are available on the biochemistry of the Sox system from R. sulfidophilum. For a detailed comparison of the biochemical data, the structures of the sox gene products for both organisms are desirable. Here, we report the X-ray crystal structure of the SoxXA complex of P. pantotrophus at a resolution of ˚ . The electron density map resolves a cysteine persul1.9 A fide for axial coordination of the heme2–iron of SoxA proposed as the active site located between the SoxX and SoxA subunits accessible from the solvent in a deep pocket. 2. Crystallization and structure determination Paracoccus pantotrophus was cultivated chemolithoautotrophically with thiosulfate as an energy source. The SoxXA complex was purified from these cells to homogeneity as described previously (Friedrich et al., 2000). For crystallization by the hanging drop vapour diffusion method, equal volumes (each 1 ll) of reservoir solution (100 mM MES, pH 6.5, 4–7% PEG 1500, 4–8 mM ZnSO4, and 75 mM [Co(NH3)6]Cl3) and protein solution (6 mg/ml SoxXA in 10 mM Bis–Tris, pH 6.5, 1 mM MgSO4, and 0.1 mM Na2S2O3) were mixed and equilibrated against 500 lL reservoir solution at room temperature. Thin plates of SoxXA crystals appeared after 3 days with the dimension of up to 700 · 400 · 50 lm3. All X-ray diffraction data were collected at 100 K. For this purpose, 5 lL cryo-solution (100 mM MES, pH 6.5, 7% PEG 1500, 4 mM ZnSO4, and 25% glycerol) was added to the crystallization drop. After incubation for several minutes, the crystals were mounted in nylon loops (Hampton Research) and flash frozen in liquid nitrogen. Data collection was carried out at beamlines BW6 (MPG) and BW7 (EMBL), respectively, at DESY (Deut sches Electronen Synchrotron, Hamburg, Germany) and at the X06SA beamline at SLS (Swiss Light Source, Villigen, Switzerland). Data were processed and scaled using the DENZO/SCALEPACK (Otwinowski and Minor, 1997) or XDS (Kabsch, 1993) program package. The first approach to determine the phases using the anomalous signal of iron failed. Whereas the position of the iron atoms could be well located, the anomalous signal was insufficient to solve the complete protein structure. Therefore, we used the structure of SoxAX of R. sulfidophilum [PDB entry 1H33 (Bamford et al., 2002)] for molecular replacement [program MOLREP (Collaborative Computational Project Number 4, 1994)] to obtain initial phases. First refinement steps using simulated annealing and energy minimization were carried out with the program package CNS (Brunger et al., 1998). For a final refinement at high resolution the program REFMAC5 from the CCP4 program package (Collaborative Computational Project Number 4, 1994) was used. After each round of refinement the model was checked and rebuilt with the program O (Jones et al., 1991). Final model analysis, imaging, and
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
ray tracing were performed using the program PyMOL (DeLano, 2002). 3. Overall structure SoxXA from P. pantotrophus crystallizes in the space ˚ , b = 180.0 A ˚, group P2(1) with cell parameters a = 42.9 A ˚ c = 117.9 A and b = 92.8. The model could be refined at ˚ resolution to Rcryst and Rfree values of 15.8 and 1.9 A 21.1%, respectively, with reasonable r.m.s.d. values for bond length and bond angles (Table 1). The final model comprises four hetero-dimeric complexes SoxXA per asymmetric unit. In each complex the residues 27–290 of SoxA and 21–157 of SoxX, respectively, could be traced in the electron density (lacking the leader peptides for secretion into the periplasma). The most favoured region of the Ramachandran plot covers 91.7% of the residues, and no residues are located within disallowed regions (Table 1). The approximate dimensions of the hetero-dimeric ˚ 3. This complex contains three enzyme are 34 · 40 · 77 A almost globular cytochrome c domains which are arranged in a linear fashion leading to its stretched form of which Table 1 Data collection and refinement statistics Data collection statistics Beamline Temperature (K) Detector ˚) Wavelength (A Space group Cell dimensions ˚) a (A ˚) b (A ˚) c (A b () ˚) Highest resolution (A Highest resolution shell Number of recorded reflections Redundancy Rsym (%)a,b I/r (I)a ˚ 2) Wilson B value (A Refinement statistics ˚) Resolution range (A Number of unique reflections Completeness (%)a Rcryst/Rfree (%)c,d Ramachandran plot Most favoured, favoured, allowed (%) ˚) r.m.s.d. on bond length (A r.m.s.d. on bond angles () a
BW6/DESY (Hamburg) 100 MarCCD 1.737 P21 42.9 180.0 117.9 92.8 1.9 2.0–1.9 631,082 5.2 8.5 (15.2) 12.6 (5.9) 23.8 19.5–1.9 120,519 93.7 (62.6) 15.8/21.1 91.7, 7.7, 0.6 0.014 1.5
Values in parentheses are for the high-resolution bin. P P P P Rsym ¼ 100 h i jI i ðhÞ hI i ðhÞij= h i I i ðhÞ where Ii(h) is the ith measurement and ÆIi(h)æis the mean of all measurements of I(h) for Miller indices hkl. P P c R ¼ ðkF obs j kjF calc kÞ= jF obs j where Fobs and Fcalc are observed and calculated structure factor amplitudes, respectively. d Rfree value is the R value obtained for a test set of reflections, consisting of a randomly selected 5% subset of the diffraction data not used during refinement. b
231
SoxX contributes one domain and SoxA two domains (Fig. 1). Each of these domains contains one heme linked to a CxxCH motif characteristic for c-type cytochromes. The imidazole ring of the histidine residue provides an axial ligand of the iron atom of the heme. The two axial ligands of heme1 are His-110 and Cys-143 while heme2 is coordinated by His-210 and Cys-251. At the Cys-251 residue additional electron density is observed, which is described below (see Section 4). Heme3 of SoxX is coordinated by His-65 and a conserved Met-111 as previously suggested from sequence alignments (Cheesman et al., 2001; Rother and Friedrich, 2002) (Fig. 2). The first heme domain of the P. pantotrophus SoxA, SoxAN, is close to the amino terminus (residues 82 and 181) and the second heme domain, SoxAC, is located towards the carboxy terminus (residues 182–280). The N-terminal region of SoxA (residue 27–81) does not display any known domain motif, but is in contact with the SoxAC domain and may stabilize this special fold of a diheme cytochrome c, as similarly observed for the respective complex of R. sulfidophilum (Bamford et al., 2002). The structures of the two cytochrome c domains of SoxA are closely related with a mean ˚ [60% matching Ca-atr.m.s.d. on Ca-atoms of around 1.5 A oms calculated with the program MAPS (Zhang et al., 2003)]. The arrangement of SoxAN and SoxAC includes a pseudo-twofold axis oriented between the two domains (Fig. 1). Sequence analysis demonstrates a high homology in primary structure to SoxAX of R. sulfidophilum with an identity of 53.4% (SoxA) and of 44.4% (SoxX), respectively. A structural homology analysis using the program DALI (Holm and Sander, 1993) is summarized in Table 2 and confirms the high homology in tertiary and quaternary structure of SoxXA of P. pantotrophus and SoxAX of R. sulfidophilum. Structural similarities to other proteins with known structure were not detected. 4. Active site The kinetics of the reconstituted Sox enzyme system (Rother et al., 2001) suggested that SoxXA functions as a heme enzyme and covalently links the sulfur substrate to the thiol of Cys-138 in SoxY (Friedrich et al., 2001). However, this first step of the reaction cycle is not understood in detail. The electron density map of SoxA of P. pantotrophus shows an electron density between the thiol of Cys-251 and the iron atom of heme2 (Fig. 3). This density fits well with a sulfur atom to give cysteine persulfide at position 251 (termed Css-251). The geometry of cysteine persulfide is almost identical with cysteine sulfenic acid (Cys-S-OH). However, the observed additional electron density is insufficient for an oxygen atom to allow interpretation as a cysteine sulfenic acid. This analysis is in accordance with that of SoxA of R. sulfidophilum and suggested as a post-translational modification (Bamford et al., 2002; Cheesman et al., 2001). The primary struc-
232
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
Fig. 1. Stereo representation of SoxXA. SoxXA is a hetero-dimer consisting of SoxX (blue) and SoxA. SoxA consists of two cytochrome c domains. The N-terminal SoxAN domain (orange) and the C-terminal SoxAC domain (green). SoxAN and SoxAC are related to each other by a twofold rotation axis (indicated with ¤). The N- and the C-terminus (both grey) of SoxA are not part of the typical cytochrome c domain. The heme groups are represented as sticks and are coloured according to their affiliation to the respective cytochrome c domain.
Fig. 2. Sequence alignment of SoxXA of P. pantotrophus (Pp) with SoxAX of R. sulfidophilum (Rs). On top of the sequence the secondary stucture of SoxXA from P. pantotrophus is shown. The heme binding CxxCH-motifs are assigned with green bars. The second axial ligand of heme1 and heme3 are assigned with (m) and the active site residue Cys251 is assigned with (q).
tures of SoxA of both organisms are 53.4% identical. Both proteins have an almost identical tertiary structure, both proteins appear to have the same novel axial coordination of the heme2–iron by cysteine persulfide and for both proteins heme1 appears to be EPR silent (Cheesman et al., 2001; P. Hellwig, E. Reijerse, M. Sommerhalter, unpublished data). These striking identities strongly
suggest that the first step is identical in aerobic chemotrophic and anaerobic phototrophic sulfur oxidation. However, at present no reconstituted enzyme system has been described and no kinetic data are available from phototrophic bacteria. Omission of the sulfur dehydrogenase Sox(CD)2 from the reconstituted Sox enzyme system which then contains
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
233
Fig. 3. Phase weighted electron density distribution (2mFo-DFc) around the proposed active site at heme2. (A) The electron density is not sufficiently described by the cysteine residue at position 251. (B) A post-translational modification of the cysteine at position 251 to a cysteine persulfide describes the electron density well.
Table 2 Comparison of the structures of SoxXA of P. pantotrophus and SoxAX from R. sulfidofilum with the program DALI (Holm and Sander, 1993)
SoxA SoxX
Z-Score
LALIa
r.m.s.d.b
%IDEc
41.5 23.3
259 133
0.9 1.0
56 49
a
Total number of equivalenced residues. Positional root mean square deviation of superimposed Ca atoms in angstroms. c Percentage of sequence identity over equivalenced positions. b
only SoxXA, SoxYZ, and SoxB of P. pantorophus allows oxidation of hydrogen sulfide, thiosulfate, and sulfite albeit with a yield of two moles of electrons per mole of substrate. Based on this reconstitution the two electron-yielding step has to be assigned to SoxXA, as SoxB is not redox active and SoxYZ does not contain a cofactor or metal. From the active site at heme2, the electrons could be transferred to a second heme group and further transmitted to the final electron acceptor. In an alternative mechanism the two electrons are stored temporarily in SoxXA at two different heme groups as each heme group is just able to store one electron. After SoxXA released the substrates, the electrons are likely to be transmitted via heme2 to the final electron acceptor. Two ways for the electron transfer from heme2 are possible. The first route goes to heme1, the second to ˚ ) a tunnelling of heme3. Over short distances (up to 14 A electrons through proteins without further co-factors has been described for different proteins (Page et al., 1987). ˚ between heme1 and heme2 With a distance of around 25 A it is unlikely, that the transport of an electron can take place without an additional co-factor. It is more likely, that the electrons are transferred to or stored within the heme groups of SoxX (distance heme2–heme3: approximately ˚ ). This interpretation is further supported by the 11.3 A comparison of SoxA of P. pantotrophus with the corresponding protein of Starkeya novella. SoxA of S. novella and P. pantotrophus are homologues and supposed to catalyze the same reaction. However, SoxA from S. novella is missing heme1 in its SoxAN domain and contains a protein disulfide instead (Kappler et al., 2004).
Acknowledgments We thank Sabine Vogt, Bjo¨rn Klink, and Georg Holtermann for technical assistance and our colleagues at the Max-Planck-Institut fu¨r molekulare Physiologie for helpful suggestions and discussions. We thank Gleb P. Bourenkov and Hans D. Bartunik for their assistance during the collection of the MAD data set at beamline BW6 and M. Weiss, P. Tucker, and E. Pohl for assistance during beamtime at beamlines BW7A/B (DESY, Hamburg, Germany), E. Mitchell and A. McCarthy for assistance in using beamline ID14-1 (ESRF, Grenoble, France). Part of this work was performed at the Swiss Light Source, Paul Scherrer Institut (Villigen, Switzerland). We would like to thank C. Schulze-Briese and T. Tomizaki for assistance. We gratefully acknowledge support by grants to A.J.S. (Sche 545/3-1) and C.G.F. (Fr 318/8-1) from the Deutsche Forschungsgemeinschaft. References Appia-Ayme, C., Berks, B.C., 2002. SoxV, an orthologue of the CcdA disulfide transporter, is involved in thiosulfate oxidation in Rhodovulum sulfidophilum and reduces the periplasmic thioredoxin SoxW. Biochem. Biophys. Res. Commun. 296, 737–741. Appia-Ayme, C., Little, P.J., Matsumoto, Y., Leech, A.P., Berks, B.C., 2001. Cytochrome complex essential for photosynthetic oxidation of both thiosulfate and sulfide in Rhodovulum sulfidophilum. J. Bacteriol. 183, 6107–6118. Bamford, V.A., Bruno, S., Rasmussen, T., Appia-Ayme, C., Cheesman, M.R., Berks, B.C., Hemmings, A.M., 2002. Structural basis for the oxidation of thiosulfate by a sulfur cycle enzyme. EMBO J. 21, 5599– 5610. Bardischewsky, F., Quentmeier, A., Hellwig, P., Rother, D., Kostka, S., Friedrich, C.G., 2005. Sulfur dehydrogenase of Paracoccus pantotrophus: the heme-2 domain of the molybdo-protein cytochrome c complex is dispensable for catalytic activity. Biochemistry 44, 7024– 7034. Brune, D.C., 1989. Sulfur oxidation by phototrophic bacteria. Biochim. Biophys. Acta 975, 189–221. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T., Warren, G.L., 1998. Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905–921.
234
T. Dambe et al. / Journal of Structural Biology 152 (2005) 229–234
Cheesman, M.R., Little, P.J., Berks, B.C., 2001. Novel heme ligation in a c-type cytochrome involved in thiosulfate oxidation: EPR and MCD of SoxAX from Rhodovulum sulfidophilum. Biochemistry 40, 10562– 10569. Collaborative Computational Project Number 4, 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50, 760–763. DeLano, W.L., 2002. DeLano Scientific. San Carlos, CA, USA,
Mukhopadhyaya, P.N., Deb, C., Lahiri, C., Roy, P., 2000. A soxA gene, encoding a diheme cytochrome c, and a sox locus, essential for sulfur oxidation in a new sulfur lithotrophic bacterium. J. Bacteriol. 182, 4278–4287. Otwinowski, Z., Minor, W., 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. Page, E., Goings, G.E., Power, B., Upshaw-Earley, J., 1987. Tunneling cell processes in myocytes of stretched mouse atria. Am. J. Physiol. 253, H432–H443. Quentmeier, A., Friedrich, C.G., 2001. The cysteine residue of the SoxY protein as the active site of protein-bound sulfur oxidation of Paracoccus pantotrophus GB17. FEBS Lett. 503, 168–172. Quentmeier, A., Hellwig, P., Bardischewsky, F., Grelle, G., Kraft, R., Friedrich, C.G., 2003. Sulfur oxidation in Paracoccus pantotrophus: interaction of the sulfur-binding protein SoxYZ with the dimanganese SoxB protein. Biochem. Biophys. Res. Commun. 312, 1011– 1018. Rother, D., Friedrich, C.G., 2002. The cytochrome complex SoxXA of Paracoccus pantotrophus is produced in Escherichia coli and functional in the reconstituted sulfur-oxidizing enzyme system. Biochim. Biophys. Acta 1598, 65–73. Rother, D., Henrich, H.J., Quentmeier, A., Bardischewsky, F., Friedrich, C.G., 2001. Novel genes of the sox gene cluster, mutagenesis of the flavoprotein SoxF, and evidence for a general sulfur-oxidizing system in Paracoccus pantotrophus GB17. J. Bacteriol. 183, 4499– 4508. Tru¨per, H.G., Fischer, U., 1982. Anaerobic oxidation of sulphur compounds as electron donors for bacterial photosynthesis. Philos. Trans. R. Soc. Lond. B Biol. Sci. 298, 529–542. Zhang, Z., Lindstam, M., Unge, J., Peterson, C., Lu, G., 2003. Potential for dramatic improvement in sequence alignment against structures of remote homologous proteins by extracting structural information from multiple structure alignment. J. Mol. Biol. 332, 127–142.