Solution Structure of Sco1

Structure, Vol. 11, 1431–1443, November, 2003, 2003 Elsevier Science Ltd. All rights reserved.

DOI 10.1016/j.str.2003.10.004

Solution Structure of Sco1: A Thioredoxin-like Protein Involved in Cytochrome c Oxidase Assembly Erica Balatri, Lucia Banci, Ivano Bertini,* Francesca Cantini, and Simone Ciofi-Baffoni Magnetic Resonance Center CERM and Department of Chemistry University of Florence Via Luigi Sacconi 6 50019, Sesto Fiorentino, Florence Italy

Summary Sco1, a protein required for the proper assembly of cytochrome c oxidase, has a soluble domain anchored to the cytoplasmic membrane through a single transmembrane segment. The solution structure of the soluble part of apoSco1 from Bacillus subtilis has been solved by NMR and the internal mobility characterized. Its fold places Sco1 in a distinct subgroup of the functionally unrelated thioredoxin proteins. In vitro Sco1 binds copper(I) through a CXXXCP motif and possibly His 135 and copper(II) in two different species, thus suggesting that copper(II) is adventitious more than physiological. The Sco1 structure represents the first structure of this class of proteins, present in a variety of eukaryotic and bacterial organisms, and elucidates a link between copper trafficking proteins and thioredoxins. The availability of the structure has allowed us to model the homologs Sco1 and Sco2 from S. cerevisiae and to discuss the physiological role of the Sco family. Introduction Copper is an essential trace element that plays a pivotal role in cell physiology, as it constitutes the functional site of important cuproenzymes. The transport and cellular metabolism of copper depend on a series of membrane proteins and smaller soluble proteins (copper chaperones) (Culotta et al., 1997; Rosenzweig and O’Halloran, 2000; Solioz and Vulpe, 1996; Valentine and Gralla, 1997). Copper chaperones exchange copper with the membrane-bound copper-transporting enzymes or incorporate the copper directly into copper-dependent enzymes. All of these proteins constitute a functionally integrated system for controlling cellular copper homeostasis and preventing the release of free copper ions that will cause damage to cellular components (Banci and Rosato, 2003; Huffman and O’Halloran, 2001; Puig and Thiele, 2002). A relevant example of copper-dependent enzymes is constituted by cytochrome c oxidase (COX), which is the terminal enzyme of the respiratory chain, embedded in the inner mitochondrial membrane of all eukaryotes and in the plasma membrane of some prokaryotes. COX contains several cofactors, including two copper sites, *Correspondence: [email protected]

two heme groups (heme a and a3), and one magnesium and one zinc ion in different subunits of the enzyme (Harrenga and Michel, 1999; Soulimane et al., 2000; Tsukihara et al., 1996; Poyton et al., 1995). In particular, the copper ions are located in subunits COX I and COX II, which contain the CuB and CuA centers, respectively. The former is formed by one copper ion while the latter is constituted by a dinuclear copper center, where, in the oxidized state, the electron is totally delocalized over the two copper ions (Kroneck et al., 1988). In eukaryotes, a large number of nuclear genes are required for the proper assembly and function of this enzyme (Carr and Winge, 2003; Tzagoloff and Dieckmann, 1990). Among these, the Sco1 protein was suggested to be involved in copper ion delivery to the COX complex (Glerum et al., 1996; Krummeck and Rodel, 1990; Schulze and Rodel, 1988). It has been also proposed that Sco1 specifically delivers copper to the CuA site in the COX II subunit (Dickinson et al., 2000; Mattatall et al., 2000). Yeast Sco1 is a 30 kDa protein anchored by a single transmembrane segment of about 17 amino acids to the inner mitochondrial membrane. A BLAST search of the Bacillus subtilis genome identified a potential homolog of the Sco1 protein from yeast, YpmQ gene (Mattatall et al., 2000). The sequence alignment between the soluble domain of YpmQ (hereafter BsSco1) and yeast Sco1 shows 23% identity, which increases to 47% when conservative substitutions are considered. B. subtilis also expresses a cytochrome c oxidase that, like the canonical mitochondrial enzyme, has both CuA and CuB centers within the homologs of subunits COX II and COX I, respectively (Saraste et al., 1991; Winstedt and Von Wachenfeldt, 2000). Protein sequence analysis has identified a potential metal binding region, CXXXCP, that is conserved in all mitochondrial Sco proteins and their bacterial homologs. Both cysteines in yeast and bacterial Sco1 are essential for the function (Mattatall et al., 2000; Rentzsch et al., 1999). X-ray absorption spectroscopy, combined with functional studies on yeast Sco1, suggested that the two cysteines present in the conserved motif and a conserved histidine in the C-terminal end are involved in copper(I) binding (Nittis et al., 2001). In the eukaryotic genome there is also a highly homologous protein, Sco2, which conserves the CXXXCP metal binding motif and whose role remains unresolved (Carr and Winge, 2003; Winstedt and Von Wachenfeldt, 2000). No solution and/or X-ray structural data are yet available for this class of proteins. In the light of the similarities between B. subtilis and eukaryotic Sco1 systems, we have expressed BsSco1 and solved its 3D solution structure as it is purified, i.e., without metal ions, to further address the role of Sco1 in the assembly of CuA center in cytochrome c oxidase. The structure reported in this paper, which is the first of this class of proteins, has a thioredoxin-like fold similar to that of the functionally unrelated peroxiredoxin proteins, but with some peculiar differences. Mobility studies showed that the high disorder around the poten-

Structure 1432

tial metal binding region is determined by enhanced backbone flexibility. Structural models of Sco1 and Sco2 from S. cerevisiae were calculated and the potential energy surfaces have been analyzed. The interaction of the BsSco1 with copper(I) and copper(II) are also reported. From the solved structure and the comparison with similar structures of unrelated proteins, an intriguing picture on the involvement of Sco1 in the assembly of COX comes out. Results Expression, Purification, and Characterization of BsSco1 The YpmQ gene codes for a protein of 193 amino acids with a molecular weight of 21,646 Da. The sequence has a potential metal binding motif CXXXCP and a histidine at position 154, which are completely conserved in both prokaryotic and eukaryotic organisms. A hydropathy analysis indicates that this protein contains a single amino-terminal transmembrane-spanning helix, suggesting that it is a membrane-bound protein and that a water soluble form can be produced if the N-terminal transmembrane segment is taken out. Thus, the recombinant protein BsSco1 lacking the first 23 residues was overexpressed in an E. coli host. The 15N and 13C/15N labeled and unlabeled protein was produced in a culture medium not supplemented with CuSO4, thus obtaining after the purification process a protein without bound metal ions, as it was checked through atomic absorption spectroscopy. This protein was then structurally characterized (see below). In the expressed protein, four amino acids (HMLE), corresponding to the restriction enzyme recognition site, are further present at the N terminus as a consequence of the cloning, thus producing a final construct of 174 amino acids with a theoretical molecular weight of 19,834 Da, as confirmed through ES-MS analysis. Analytical gel filtration analysis showed that BsSco1 eluted in fractions corresponding to a monomeric state for the protein. The far-UV circular dichroism (CD) spectrum of BsSco1 has features characteristic of a folded protein with a large content of ␤ strands and ␣-helical structures, the latter indicated by a negative band at 224 nm. Analysis of the CD spectrum indicated a 23% ␣-helical content, consistent with the solution structure, which has a 26% ␣-helix content (see below). NMR Spectra and Resonance Assignment The 15N and 13C HSQC spectra of BsSco1 show welldispersed resonances indicative of an essentially folded protein. The backbone resonance assignment was obtained from the analysis of CBCANH and CBCA(CO)NH and of HNCO and HN(CA)CO spectra. One hundred and fifty-six out of the expected one hundred and sixty-three 15 N backbone amide resonances were assigned. The amide resonances are missing for residues Met 2, Lys 126, Glu 130, Val 133, His 135, Ser 137, and Ser 138. The assignment of the aliphatic side chain resonances was performed through the analysis of 3D CC(CO)NH and (H)CCH-TOCSY spectra, together with 15N-NOESYHSQC and 13C-NOESY-HSQC spectra. The assignment

of the aromatic spin systems was performed with 2D NOESY and TOCSY maps acquired in D2O solvent and a 3D 13C-NOESY-HSQC spectrum with the carrier centered in the aromatic region at 130 ppm. The ring NHs of the His residues are not detected, possibly due to their fast exchange with the solvent. The nonexchangeable His ring protons were assigned through a 1H-15N HSQC experiment tailored to the detection of 2J1H-15N couplings and from the analysis of the 2D NOESY maps. For two histidines (His 55 and 135, the only other His of the sequence being His 1), all the nonexchangeable protons were assigned. From the pattern of 2J1H-15N couplings the protonation state of the His ring nitrogens was also characterized (Pelton et al., 1993). His 55 has a positively charged imidazole ring, while His 135 exists predominantly in the N⑀2-H state. In total, the resonances of 95% of carbon atoms, 98% of nitrogen atoms, and 93% of protons were assigned, leaving only His 1 and Lys 126 completely unassigned. The 1H, 13C, and 15N resonance assignments of the BsSco1 are reported in Supplemental Table S1 available in the Supplemental Data online at http://www.structure.org/cgi/content/full/ 11/11/1431/DC1. Chemical shift index analysis (Wishart and Sykes, 1994) on H␣, CO, C␣, and C␤ resonances, 3JHNH␣ coupling constants, d␣N(i-1,i)/dN␣(i,i) ratios (Gagne et al., 1994), and NOEs patterns indicated the presence of eight ␤ strands, four ␣ helices, and a 310 helix. Structure Calculation and Analysis By analyzing 3D 15N-edited and 13C-edited NOESY spectra and 2D NOESY spectra, 5994 NOE crosspeaks were assigned and transformed into 4839 unique upper distance limits, of which 3492 were meaningful. The number of meaningful NOEs per residue is reported in Supplemental Figure S1A (Supplemental figures available online at http://www.structure.org/cgi/content/full/11/11/ 1431/DC1), being more than 20 in average. Two regions (44–51 and 124–138) clearly experienced a very low number of NOEs. A total of 88 proton pairs were stereospecifically assigned using the program GLOMSA (Gu¨ntert et al., 1991) and the HNHB experiment (Archer et al., 1991). Ninety dihedral φ angle constraints were obtained from the analysis of the HNHA spectrum, and ninety-one dihedral ␺ angles were obtained from the 15 N-edited NOESY spectrum (see Experimental Procedures). Distance constraints, stereospecific assignments, and dihedral angles used in structure calculations are reported in Supplemental Tables S2–S5. After restrained energy minimization (REM) on each of the 30 conformers of the family, the RMSD to the mean structure (for residues 5–170) is 1.21 ⫾ 0.24 A˚ for the backbone and 1.74 ⫾ 0.21 A˚ for all heavy atoms. If the segments 45–49 and 124–138 are not included, the RMSD values drop to 0.74 ⫾ 0.14 and 1.22 ⫾ 0.14 A˚ for the backbone and heavy atoms, respectively. The final REM family has a target function of 0.49 ⫾ 0.04 A˚2 and of 0.23 ⫾ 0.05 rad2 for the NOEs and the dihedral angle constraints, respectively. The results of the conformational and energetic analysis of the REM family of BsSco1 are reported in Table 1. The RMSD values per residue to the mean structure of the final REM family

Solution Structure of Sco1 from B. subtilis 1433

Table 1. Statistical Analysis of the Energy-Minimized Family of Conformers and the Mean Structure of BsSco1 RMS Violations per Meaningful Distance Constraint (A˚)b

⬍REM⬎a

REMa 30 Structures

Intraresidue (686) Sequential (926) Medium range (788)c Long range (1092) Total (3492)

0.0162 0.0084 0.0086 0.0091 0.0107

⫾ ⫾ ⫾ ⫾ ⫾

0.0010 0.0015 0.0012 0.0009 0.0005

Mean 0.0175 0.0080 0.0071 0.0091 0.0107

RMS Violations per Meaningful Dihedral Angle Constraints (Deg)b Phi (90) Psi (91) Average number of constraints per residue

2.23 ⫾ 0.45 1.57 ⫾ 0.35 21

1.83 1.41

31.4 ⫾ 3.5 13.0 ⫾ 3.0 10.6 ⫾ 2.4 16.0 ⫾ 2.6 71.1 ⫾ 6.9 8.4 ⫾ 1.6 3.4 ⫾ 1.2 0.0 ⫾ 0.0 7.5 ⫾ 2.0

35 13 7 15 70 6 5 0 7

67.8 (73.6) 24.8 (20.8) 5.7 (4.5) 1.7 (1.1) 3.18 ⫾ 0.2 ⫺0.42 ⫾ 0.03

65.8 (72.6) 26.5 (22.2) 7.1 (5.2) 0.6 (0.0) 2.90 ⫺0.46

68%

61%

50%

44%

Average Number of Violations per Structure Intraresidue Sequential Medium range Long range Total Phi Psi Average number of NOE violations larger than 0.3 A˚ Average number of NOE violations between 0.1 A˚ and 0.3 A˚ Structural Analysisd Percentage of residues in Percentage of residues in Percentage of residues in Percentage of residues in H bond energy (kJ mol⫺1) Overall G factor

most favorable regions allowed regions generously allowed regions disallowed regions

Experimental Restraints Analysise Percentage of completeness of experimentally observed NOE up to 4 A˚ cut-off distance Percentage of completeness of experimentally observed NOE up to 5 A˚ cut-off distanc

a REM indicates the energy-minimized family of 30 structures; ⬍REM⬎ is the energy-minimized mean structure obtained from the coordinates of the individual REM structures. b The number of meaningful constraints for each class is reported in parenthesis. c Medium range distance constraints are those between residues (i,i⫹2), (i,i⫹3), (i,i⫹4), and (i,i⫹5). d As it results from the Ramachandran plot analysis; in parenthesis the results of analysis (excluding the following residues: 1–4, 45–49, 124–137) is also reported. For the PROCHECK statistics, an average hydrogen bond energy in the range of 2.5–4.0 kJ mol⫺1 and an overall G factor larger than ⫺0.5 are expected for a good quality structure. e As it results from the AQUA analysis.

are shown in Supplemental Figure S1B. The most disordered regions are those involving residues 44–51 (located in loop 3) and residues 124–138 (located in loop 8), as a consequence of the low number of NOEs and, for the latter segment, also of the lack of backbone and/or side chain assignments. The solution structure of BsSco1 is shown in Figure 1. The structure of BsSco1 is characterized by the following secondary structure elements: ␤1 (11–13), ␤2 (16–20), ␤3 (24–28), helix 310 (␪1) (29–31), ␤4 (36–41), ␣1 (52–67), ␤5 (72–77), ␣2 (84–94), ␤6 (100–104), ␣3 (108– 119), ␤7 (139–143), ␤8 (146–152), and ␣4 (160–173). The eight ␤ strands are organized with parallel and antiparallel alignments and surrounded by four helices of various lengths (Figure 2A). The potential copper ligands Cys 45 and Cys 49 are located in loop 3, which is structurally close to loop 8 containing the other potential copper ligand His 135.

Mobility Studies on BsSco1 Reliable 15N R1, R2, and 1H-15N NOE values, which provide information on internal mobility, have been obtained for 141 of the 156 assigned backbone NH resonances. The others are too weak or are too overlapped to be integrated accurately. R1, R2, and 1H-15N NOE average values of BsSco1 are 1.36 ⫾ 0.05 s⫺1, 15.9 ⫾ 0.6 s⫺1, and 0.78 ⫾ 0.02, respectively, at 700 MHz and 1.40 ⫾ 0.06 s⫺1, 13.8 ⫾ 0.6 s⫺1, and 0.83 ⫾ 0.02, respectively, at 600 MHz. The experimental relaxation data are reported in the Supplemental Data (Supplemental Figures S2A and S2B. R1, R2, and 1H-15N NOE values were essentially homogeneous along the entire polypeptide sequence, with the exception of the 42–51 and 124–138 loops where some residues show larger R2 values. Moreover, the N terminus and the 125–134 segment show a significant decrease in 1H-15N NOE values. The correlation time for molecule reorientation (␶m),

Structure 1434

Figure 1. Solution Structure of BsSco1 The radius of the tube is proportional to the backbone RMSD of each residue. The secondary structure elements are shown: ␤ strands in cyan and helices in red. Inset (A) shows the backbone atoms in the vicinity of the potential metal binding region for the REM family of 30 structures. Cys 45, 49, and His 135 side chains are also depicted in blue. Inset (B) shows the backbone atoms in the vicinity of the potential metal binding region for 30 members of the DYANA family obtained applying on the apo structure upper distance limits between a copper ion and the sulfur of Cys 45, 49, and the N␦1 of His 135. Cys 45, 49, and His 135 side chains and copper ions are also depicted in blue and in gold, respectively.

estimated from the R2/R1 ratio, is 9.3 ⫾ 0.9 ns, as expected for a protein of this size in a monomeric state (Einstein, 1956), consistent with gel filtration analysis. From the spectral density function analysis (Figure 3), it appears that residues in the 125–134 region (loop 8) and at the N terminus have higher J(␻H) than average, consistent with lower heteronuclear NOE values. This behavior indicates the presence of local motions in ns-ps timescale, i.e., faster than the overall protein tumbling rate. In addition to these fast internal motions, slow conformational exchange processes on the ms-␮s timescale affect part of loop 8, as Jeff(0) and R2 values higher than the average are observed for residues Gly 129, Asp 131, and Ile 134 (Figure 3). Exchange contributions to relaxation have also been observed for the region 40–49 and Asp 78. The presence of exchange contributions to transverse relaxation rates was independently evaluated through R2 measurements as a function of the ␶CPMG length. The residues experiencing exchange contribution to R2 rates are mostly located around the potential metal binding region: Ala 38, Asp 39, Phe 40 (strand ␤4), Thr 43, Cys 45, Glu 46, Ile 48, Cys 49 (loop 3), Met 52 and Ala 54 (helix ␣1), Ser 76, Val 77, Asp 78 (strand ␤5) and Gly 129, and Asp 131 (loop 8). Residues experiencing conformational exchange processes in BsSco1 as a function of residue number are shown in Supplemental Figure S3. The NH of Ile 134 has a R2 value higher than average but does not experience a dependence of R2 with the ␶CPMG length, suggesting the presence of ex-

change processes occurring at rates faster than those accessible with the present experimental conditions (see Experimental Procedures). The occurrence at the same time of conformational exchange processes and ns-ps motions for residues belonging to loop 8 can explain both the disappearance of amide resonances for some residues and the paucity of NOEs for others, so determining their higher RMSD values. Backbone NH protons whose signals are still observable, with full or partial intensity, days after the protein was dissolved in D2O, are not accessible to the solvent. Fifty-one residues are still observable after 3 days and are all located in ␤ strands and ␣ helices (Supplemental Figure S3) consistent with the presence of a corestranded ␤ sheet with flanking helices. On the contrary, in the regions 160–173 (helix ␣4) and 11–13 (strand ␤1), the amide protons of almost all residues are completely exchanged. This suggests that the latter two secondary structure elements are more solvent accessible than the others. Residues displaying fast amide proton exchange, as detected in a 15N-(CLEANEX-PM)-FHSQC experiment, are located in loop regions and in the solvent-exposed side of strand ␤2 and ␤3 (Supplemental Figure S3). Modeling of Yeast Sco1 and Sco2 The soluble domains of Sco1 and Sco2 from S. cerevisiae show a sequence similarity of 47% and 42% with BsSco1, respectively. Structural models for the se-

Solution Structure of Sco1 from B. subtilis 1435

Figure 2. Comparison between the Structures of BsSco1, TlpA, and HBP23 BsSco1 (A), TlpA (B), and HBP23 (C). The side chains of the potential copper ligands Cys 45, Cys 49, and His 135 in BsSco1; of Cys 72 and Cys 75 in TlpA; and of Cys 52 and Cys 173 in HBP23 are shown. The topology of each protein is shown in the right panels. The dashed lines indicate the formation of a sheet between two ␤ strands.

quence of yeast Sco1 and Sco2 were calculated with the program MODELLER version 4.0 (Sali and Blundell, 1993), using the solution structure of BsSco1 as a template. Overall backbone RMSD values of 0.49 A˚ and 0.60 A˚ were found between yeast Sco1/Sco2 and BsSco1, respectively. The electrostatic potential surface of yeast Sco1 and Sco2 are strongly similar, in accordance with the high sequence similarity (85%). Both yeast Sco1 and Sco2 show an extended negatively charged patch around the potential copper binding ligands, generated by conserved Asp and Glu residues (Figure 4A). In addition, Lys and Arg residues in yeast

Sco1 and Sco2 produce positive electrostatic potential in the regions that comprise helix ␣3 and ␣4 (Figure 4A). The opposite side of the protein shows scattered negative and positive charges. Interaction with Copper BsSco1 is able to bind copper(I). Atomic absorption spectroscopy showed that purified unlabeled cleaved His tag BsSco1, from cells grown in rich medium with the presence of copper, binds ⵑ0.7 copper(I) atom per monomer. EXAFS data on yeast Cu(I)-Sco1 indicates that the conserved two cysteines and one histidine are

Structure 1436

Figure 3. Spectral density functions J(␻H), J(␻N), J(0) versus the residue number as obtained from the 15N relaxation data measured at 600 MHz

the ligands (Nittis et al., 2001). Therefore, BsSco1 structure indicates that Cys 45, Cys 49, and His 135 might be the possible ligands. BsSco1 is also able to bind copper(II). The optical spectra show an absorption band at ␭max ⫽ 349 nm (⑀ ⫽ 1813 M⫺1 cm⫺1), characteristic of a type 2 copper cysteinate coordination (Andrew et al., 1994; Canters and Gilardi, 1993; Lu et al., 1996), and two weaker absorption bands at 460 and 547 nm, which are responsible for the red color of the protein solution. The EPR spectra of Cu(II)-BsSco1 at 298 K and 120 K show the presence of two species with similar intensity, both with type 2 copper site parameters (g|| ⫽ 2.15, A|| ⫽ 178 ⫻ 10⫺4 cm⫺1, and g|| ⫽ 2.24, A|| ⫽ 196 ⫻ 10⫺4 cm⫺1, respectively). Binding of copper(II) to BsSco1 has sizable effects on nuclear relaxation and consequently on the NMR signal linewidths (Banci et al., 1991; Bertini et al., 2001). Through 1H-15N HSQC spectra, the effect of copper binding on each residue of the protein can be selectively followed during a titration, unless the overlap of the peaks prevents this analysis. Addition of up to one

equivalent of copper(II) resulted in the disappearance of the 1H-15N crosspeaks of the amide protons of residues 38–40, 42–46, 48, 49, 53–56, 76–78, 81, 82, 134, 139, and 153. For residues 57, 83, 86, 87, 91, 105, 109, 128, 140, and 152 a second crosspeak is observed during the titration whose intensity increases as the intensity of the original peak decreases. Further additions of copper do not produce other relevant spectral variations (i.e., line broadening and/or signal disappearance). The residues which disappear include the potential CXXXCP copper binding region, His 135 and His 55, thus indicating that these residues might be involved in the interaction with copper(II) or are located close to it. Discussion Description of the Structure The overall structure of BsSco1 comprises a typical thioredoxin fold (Martin, 1995), which is formed by a central four-stranded ␤ sheet (␤4, ␤5, ␤7, ␤8) and three flanking ␣ helices (␣1, ␣3, ␣4) (Figure 2A). In addition to the thioredoxin fold, a two-stranded ␤ sheet (␤2 and ␤3),

Solution Structure of Sco1 from B. subtilis 1437

Figure 4. Rotated Views of the Electrostatic Surfaces of the Modeled Sco2 from S. cerevisiae, BsSco1, and the Modeled COX II from B. subtilis The modeled structure of Sco2 from S. cerevisiae (A), the solution structure apoSco1 from B. subtilis (B), and the modeled structure of COX II from B. subtilis (C). The positively charged, negatively charged, and neutral amino acids are represented in blue, red, and white, respectively. The position of Cys 41, Cys 45, and His 132 of yeast Sco2, Cys 45, Cys 49, and His 135 of BsSco1, and of Cys 213 and Cys 217 of COX II are also indicated. Structural models for the sequences of yeast Sco2 and COX II from B. subtilis were calculated using the solution structure of BsSco1 and the X-ray structure of COX from P. denitrificans (Harrenga and Michel, 1999) as templates, respectively.

forming a ␤-hairpin structure and a short 310-helix (␪1), are present at the N terminus (Figure 2A). Another helix (␣2) and a strand (␤6), the latter forming a parallel ␤ sheet with strand ␤5, are inserted between strand ␤5 and helix ␣3 (Figure 2A). Considering the above, it follows that this fold belongs to a subset of the thioredoxin superfamily, present in peroxiredoxin proteins (Choi et al., 1998). An additional property of the BsSco1 fold is

the presence, at its N-terminal region, of a short strand ␤1, which forms an antiparallel ␤ sheet with strand ␤8 (Figure 2A). The core of the protein is characterized by several hydrophobic interactions which involve aromatic and aliphatic residues: stacking contacts are present among aromatic residues (Phe 40 [␤4], Phe 75 [␤5], Phe 120 [loop 8], Phe 115 [␣3], and Tyr 140 [␤7]), which are all

Structure 1438

Figure 5. Mapping of Conserved Residues in the Sco1 Family Identical and similar residues in 90% or more aligned sequences of bacterial and eukaryotic organisms are shown on BsSco1 structure in dark gray and gray, respectively. The conserved residues are also indicated. The secondary structure elements are shown: ␤ strands in black and helices in light gray.

clustered on one side of the central ␤ sheet and which determine the spatial vicinity of helix ␣3 to this region, particularly as a consequence of contacts between Phe 120 (loop 8) and Phe 115 (␣3). Aliphatic side chains of residues Val 70 (␤5), Leu 63 (␣1), Ile 72 (␤5), Leu 37 (␤4), and Leu 141 (␤7) form, on the other side of the central ␤ sheet, several hydrophobic interactions. All the abovementioned residues, with the exception of Phe 115, are completely conserved in Sco1 homologous proteins (Figure 5). Helix ␣4 is spatially distant from the central ␤ sheet and does not have conserved hydrophobic residues (Figure 5). A strong hydrogen bond network links all the secondary structure elements present in BsSco1, while other relevant hydrogen bonds stabilize some external loops and/or well-structured regions of the protein. In particular, the hydrogen bonds between NH of Asp 22 (loop 1) and the CO of the conserved (Figure 5) Asn 20 (␤2) and between the NH of Lys 24 (␤3) and the OD1 of Asn 20 (␤2) are crucial for capping the ␤-hairpin turn. The peculiar, short strand ␤1, antiparallel to strand ␤8, is stabilized by the presence of two H bonds, between NH of Val 13 and CO of Asp 146 and between NH of Val 148 and CO of Tyr 11. Finally, helix ␪1 is stabilized by two H bonds between NH of Ser 30 (␪1) and CO of Ser 27 (␤3) and between NH of Lys 32 (␪1) and CO of Leu 28 (␤3). Loop 3, containing the potential metal binding ligands Cys 45 and Cys 49, loop 8, containing the conserved

His 135, and loop 10, all structurally close, have the highest RMSD values in the protein. In particular, loop 8 spans a large number of different conformations (inset A of Figure 1) and the few backbone NHs detected in this region are characterized by both conformational exchange processes and fast internal motions. It appears, therefore, that the three conserved residues, which were shown to be essential for the function in eukaryotic and bacterial organisms (Mattatall et al., 2000; Rentzsch et al., 1999), are located in structurally close regions, which are peculiarly characterized by enhanced backbone flexibility. Calculations of solvent accessibility show that the sulfur atoms of Cys 45 and Cys 49 are remarkably highly solvent exposed (51% and 88% of their surface, respectively). Using EXAFS data (Nittis et al., 2001), we performed structure calculations introducing a copper(I) atom linked to the sulfur atoms of Cys 45 and 49, with upper distance limits of 2.5 A˚, and to the N␦1 of His 135, with an upper distance limit of 2.3 A˚. No structural violations of the experimental constraints of the apo form are observed, indicating that the loop 8 containing His 135 can move toward copper, where a well-ordered His 135 is coordinated to the copper ion. This is shown in insets A and B of Figure 1. Moreover, the RMDS values of loop 8 (residues 126–139) decrease from 2.82 A˚ to 1.72 A˚ when passing from the apo structure to the Cu(I) model. Similarly, it has been recently suggested (Cavet et al., 2003) for the copper-metallochaperone Atx1 from Syn-

Solution Structure of Sco1 from B. subtilis 1439

echocystis PCC 6803 that His 61 moves from a remote mobile loop for the apo form into the coordination sphere of copper and in so doing influences loading and unloading of the protein. This interaction is also proposed to regulate the interaction with the two copper(I) ATPase protein partners CtaA and PacS. The electrostatic potential surface of BsSco1 is shown in Figure 4B. The side of the protein containing the putative copper binding ligands has several negatively charged residues, including the conserved Asp 78 and Asp 82 (Figure 5), thus producing a negative region similar to what it was observed for the yeast homolog surfaces. The other side of the protein shows predominantly a negative patch with some scattered positive charges (Figure 4B). The exclusively negative electrostatic potential surface is consistent with the total protein charge of ⫺10 and with its estimated pI of 4.75. It is interesting to note that the extended positive region comprising helix ␣3 and helix ␣4 in yeast Sco2 is replaced by negative residues in the BsSco1 electrostatic surface (Figures 4A and 4B). Subunit II of B. subtilis COX has been modeled with the program MODELLER version 4.0 (Sali and Blundell, 1993), using the X-ray structure of COX from Paraccoccus denitrificans bacterium in its completely oxidized state (Harrenga and Michel, 1999). The sequence identity and similarity are 27% and 52%, respectively. The modeled subunit II, shown in Figure 4C, has a positively charged patch quite close to the metal binding site, thus suggesting a possible electrostatic interaction between BsSco1 and COX II. This is consistent with the complementary high pI value of the soluble region of COX II, which is estimated to be 8.2.

Comparison between the Structures of BsSco1 and Thioredoxin-like Proteins A search for structurally related proteins in Protein Data Bank (PDB) (Berman et al., 2000) identified hORF6 (PDB ID 1PRX) and HBP23 (PDB ID 1QQ2), both members of the peroxiredoxin family of peroxidases, and the thioredoxin protein TlpA (PDB ID 1JFU) as the closest structurally homologs to BsSco1. BsSco1 shares with HBP23 and TlpA proteins 19% and 24% identity, respectively. In Figure 2, the X-ray structures of TlpA (B) (Capitani et al., 2001) and HBP23 (C) (Hirotsu et al., 1999) are compared to the solution structure of BsSco1 (A). The RMSD values to BsSco1 mean structure, calculated on the C␣ atoms of aligned residues excluding gaps, are 2.9 and 3.1 A˚ for TlpA and HBP23, respectively. TlpA is a protein present in the nitrogen-fixing soil bacterium Bradyrhizobium japonicum and is required for the assembly and stability of COX (Loferer et al., 1993; Loferer and Hennecke, 1994). This protein is anchored to the cytoplasmatic membrane through a hydrophobic N-terminal transmembrane domain, with the hydrophilic thioredoxin-like domain oriented to the periplasm (Loferer et al., 1993). TlpA has a CXXC motif as redox active site, as found for all the periplasmic and cytoplasmic members of the thioredoxin superfamily, and can be functionally described as thiol:disulphide oxidoreductase (Capitani et al., 2001). HBP23 and its bacterial and eukaryotic homologs be-

long to a family of antioxidant enzymes known as peroxiredoxins, which show peroxidase activity (Ellis and Poole, 1997) using the redox properties of a single or a pair of highly conserved cysteine residues (Choi et al., 1998; Wood et al., 2002). The protein core constituted by the central ␤ sheet (␤4–␤8) and the flanking ␣ helices (␣1–␣4) is conserved among all the three structures (Figure 2), even if HBP23 structure has an additional helix 310 after helix ␣2. The topology differences observed between the three structures are located around the N-terminal region. Indeed, the HBP23 structure lacks strand ␤1. On the contrary, TlpA conserves strand ␤1 but it has a longer N-terminal region encompassing a further helix (␣0), which is absent in BsSco1 and HBP23 (Figure 2). Both cysteines of the CXXC motif of TlpA and the N-terminal cysteine of HBP23 are aligned with Cys 45 and Cys 49 of BsSco1, respectively (Figure 6). However, while Cys 45 and Cys 49 are located in a loop of BsSco1, the CXXC active site of TlpA is shared by helix ␣1 and by its preceding short loop (Figures 2 and 6). In the BsSco1 structure, helix ␣1 is interrupted by the presence of two sequential conserved proline residues, one of which is conserved (Figures 5 and 6). In HBP23, the conserved Cys residue is located in a loop region and is preceded by a proline, as in BsSco1 (Figures 2 and 6). The secondary structure elements among the three structures are quite well aligned, even if, in some cases, they have different lengths (Figure 6). A relevant difference between BsSco1 and the two structurally related proteins involves loop 8 located after helix ␣3, which is much longer in BsSco1 (Figure 6). Interestingly, this loop contains the His residue, which is conserved in Sco1 proteins and which is critical for the function of BsSco1 (Mattatall et al., 2000).

Biological Implications Cytochrome c oxidase is the terminal copper-dependent enzyme of the respiratory chain, constituted of 2–4 subunits in bacteria, 12 in yeast, and 13 in mammals (Harrenga and Michel, 1999; Poyton et al., 1995; Soulimane et al., 2000; Tsukihara et al., 1996). A series of accessory proteins are critical for the correct assembly and function of COX (Carr and Winge, 2003). Among them, the ubiquitous protein Sco1 has been implicated in the copper ion loading of COX (Glerum et al., 1996). Interestingly, pathogenic mutations in Sco1 and Sco2 genes were recently described in several cases of fatal infantile encephalopathy, which determine partial or incorrect assembly of COX (Papadopoulou et al., 1999; Valnot et al., 2000). The present solution structure of Sco1 from B. subtilis is the first of this class of proteins and serves as a model for eukaryotic and bacterial homologs, given the high residue similarity within this class of proteins. The structure shows that a thioredoxin-like fold, generally devoted to a thiol:disulphide oxidoreductase activity or a peroxidase activity, is able to host a potential copper binding site. Our data as well as EXAFS on yeast Sco1 (Beers et al., 2002; Nittis et al., 2001) have indeed shown that Sco1 protein is able to bind copper(I) and possibly copper(II) through the sulfur atoms of the CXXXCP motif. In particular, the copper binding motif

Structure 1440

Figure 6. Sequence Alignment of BsSco1, TlpA, and HBP23 The transmembrane segments present in the TlpA and BsSco1 proteins were removed in the sequence alignment. Amino acid numbering is reported at the top, according to the sequence of BsSco1. Secondary structure elements are indicated for all three proteins. Identical or highly conserved residues are indicated by the symbols “*” and “•”, respectively, below the sequences. The conserved Cys and His residues are gray shaded.

is similar to that of the CuA center of COX, which receives copper ions through the Sco1 pathway. This motif is not present in any thioredoxin-like proteins characterized so far. Notably, the above mentioned pathogenic mutations occur within the CXXXCP motif and near the conserved histidine (Papadopoulou et al., 1999). These data suggest that Sco1 can be directly implicated into the copper transfer to CuA site of COX II or can be involved in the COX II maturation, carrying out a function of thiol:disulphide oxidoreductase. In conclusion, the structure of BsSco1 will help to direct future biological studies for the understanding of the COX complex assembly process. Experimental Procedures Sequence Analysis Sequence alignments were done with CLUSTALW (Thompson et al., 1994). HMMTOP and TMpred programs were used to obtain a prediction of the transmembrane helices and membrane topology of BsSco1 (Hofmann and Stoffel, 1993; Tusnady and Simon, 1998). BsSco1 Cloning and Purification The YpmQ gene, lacking the first 69 base pairs (bp), was amplified by polymerase chain reaction (PCR) from B. subtilis genomic DNA template (extracted by phenol-chlorophorm method; Sambrook et al., 1989) using the following primers: forward, 5⬘-CCGCTCGAGat taaagatccgctcaattacg-3⬘; reverse, 5⬘-CAATACGCGGATCCttacttga gtgtactggctgac-3⬘. The 536 bp PCR product was cloned into the His-tag fusion expression vector pET16b (Novagen), creating pETBsSCO1. The plasmid was sequenced with a ABI PRISM 310 Genetic Analyzer (Applied Biosystem). B. subtilis cells (strain 168 obtained from Coleccio´n Espan˜ola de Cultivos Tipo [CECT]) were grown in Luria-Bertani medium containing 2% glucose at 37⬚C. The sequenced plasmid was used to transform E. coli [DH5␣ for plasmid propagation and BL21(DE3) Gold pLysS (Novagen) for protein expression]. Escherichia coli cells were grown in 2⫻ YT

medium for unlabeled protein preparation at 37⬚C in the presence of the ampicilline and chloroamphenicol antibiotics. To prepare unlabeled Cu(I)-BsSco1, the culture medium was supplemented with 1.4 mM of CuSO4. For labeled protein, the culture media contained 13 C glucose and/or (15NH4)2SO4 as only sources of carbon and nitrogen and were not supplemented with CuSO4. After the isopropyl␤-D-thiogalactopyranoside induction time, the cells were harvested, resuspended, and then lysed by repeated sonication. The protein was purified through affinity chromatography. The poly-His tag was cleaved with Factor Xa protease (Roche) and another affinity chromatography was then performed. Protein expression and purification was monitored by sodium dodecyl phosphate-polyacrylamide gel electrophoresis in 17% polyacrylamide gels stained with Coomassie brilliant blue R-250 against Perfect Protein marker (Novagen). Protein Analysis To prevent disulfide formation, which might occur because of the presence of two cysteines in the CXXXC motif, the protein samples were kept in anaerobic conditions and reduced by the addition of 2 mM dithiothreitol (DTT). Protein concentrations were determined by using the experimentally determined extinction coefficient of 15815 M⫺1 cm⫺1 at 280 nm. ES-MS spectra of BsSco1 were taken with an Applied Biosystems ESI-TOF Mariner mass spectrometer. CD spectra were collected on a JASCO J-500C spectropolarimeter with a fused quartz cuvettes with 0.1 cm path length (Merck). The CD spectrum of BsSco1 was analyzed, without applying any data smoothing routine, with the DICROPROT V2.5 application package (Deleage and Geourjon, 1993) to estimate secondary structure composition. NMR samples were prepared in 100 mM potassium phosphate buffer (pH 7) containing 10% D2O, with a protein concentration of 1–2 mM in the presence of 2 mM DTT. The metal content was measured using a Perkin Elmer 2380 atomic absorption spectrophotometer. NMR Measurements and Structure Calculations The NMR spectra were acquired on Avance 800, 700, 600, and 500 Bruker spectrometers operating at proton nominal frequencies of 800.13 MHz, 700.13 MHz, 600.13 MHz, and 500.13 MHz, respectively. All the triple resonance (TXI 5-mm) probes used were

Solution Structure of Sco1 from B. subtilis 1441

Table 2. Acquisition Parameters for NMR Experiments Performed on BsSco1

Experiments 1

1

b

[ H- H]-NOESY [1H-1H]-TOCSYb 1 H-15N-HSQCc 1 H-13C-HSQCc CBCA(CO)NHd CBCANHd HNCOd HN(CA)COd 13 C (H)CCH-TOCSYd CC(CO)NH-TOCSYd 15 N-edited [1H-1H]-NOESYc 13 C-edited [1H-1H]-NOESYc HNHAb HNHBb

Dimension of acquired data (Nucleus)

Spectral width (ppm)

t1

t3

F1

F2

2048(1H) 2048(1H) 1024(1H) 1024(1H) 1024(1H) 1024(1H) 1024(1H) 1024(1H) 1024(1H) 1024(1H)

15 15 40 70 88 88 16 16 88 88 15 15 15 15

15 15 7 14 40 40 40 40 88 40 40 80 40 40

t2 1

1024( H) 1024(1H) 512(15N) 256(13C) 136(13C) 136(13C) 80(13C) 80(13C) 272(13C) 128(15C) 368(1H) 320(1H) 256(1H) 128(1H)

1

2048( H) 2048(1H) 1024(1H) 2048(1H) 56(15N) 56(15N) 40(15N) 40(15N) 96(13C) 48(15N) 64(15N) 80(13C) 40(15N) 32(15N)

F3

na

Reference

12 12 12 12 12 12 15 15 15 15

64 64 16 16 16 16 8 16 4 16 16 8 16 32

Wider et al., 1984 Bax and Davis, 1985 Bodenhausen and Ruben, 1980 Bodenhausen and Ruben, 1980 Kay et al., 1990 Kay et al., 1990 Kay et al., 1990 Kay et al., 1990 Kay et al., 1993 Gardner et al., 1996 Wider et al., 1989 Wider et al., 1989 Vuister and Bax, 1993 Archer et al., 1991

a

Number of acquired scans. Data acquired on 800 or 600 MHz spectrometers. Two-dimensional NOESY and TOCSY experiments were also recorded on a sample exchanged in D2O. c Data acquired on a 700 MHz spectrometer. A 13C-edited NOESY was also recorded with a spectral width of 15 ppm ⫻ 25 ppm ⫻ 15 ppm, centered at 130 ppm to assign the aromatic proton and carbon resonances. d Data acquired on a 500 MHz spectrometer. b

equipped with pulsed field gradients along the z axis. The 500 MHz machine was equipped with a triple resonance cryoprobe. The NMR experiments recorded on 13C/15N and 15N enriched and unlabeled BsSco1 samples are summarized in Table 2. All 3D and 2D spectra were collected at 298 K, processed using the standard Bruker software (XWINNMR), and analyzed through the XEASY program (Eccles et al., 1991). The backbone assignment was initially performed using the automatic assignment program GARANT (Bartels et al., 1997) on the 3D spectra CBCANH, CBCA(CO)NH, HNCO, and HN(CA)CO without an input structure. The program assigned 60% of the backbone resonances; the assignment was then manually completed. Distance constraints for structure determination were obtained from 15 N-edited and 13C-edited 3D NOESY-HSQC experiments and from 2D NOESY (see Table 2). Stereospecific assignments of diastereotopic protons were obtained from the analysis of the HNHB experiment and with the program GLOMSA (Archer et al., 1991; Gu¨ntert et al., 1991). 3JHNH␣ coupling constants, determined through the HNHA experiment, were transformed into backbone dihedral φ angles through the Karplus equation (Vuister and Bax, 1993). Backbone dihedral ␺ angles for residue (i-1) were determined from the ratio of the intensities of the d␣N(i-1,i) and dN␣(i,i) NOEs obtained from the 15 N-edited NOESY-HSQC spectrum (Gagne et al., 1994). The elements of secondary structure were determined on the basis of the chemical shift index (Wishart and Sykes, 1994) of the 3 JHNH␣ coupling constants and of the backbone NOEs. An automated CANDID (Herrmann et al., 2002) approach combined with the fast DYANA torsion angle dynamics algorithm was used to assign the ambiguous NOE crosspeaks and to have a preliminary protein structure. Structure calculations were then performed through iterative cycles of DYANA (Gu¨ntert et al., 1997) followed by restrained energy minimization (REM) with AMBER 5.0 (Pearlman et al., 1997) applied to each member of the family. The quality of the structures was evaluated using the programs PROCHECK-NMR (Laskowski et al., 1996) and AQUA (Laskowski et al., 1998). Structures were represented with the program MOLMOL (Koradi et al., 1996). Structurally related proteins were searched through the DALI server (Holm and Sander, 1996). Relaxation Data Acquisition and Analysis All NMR experiments for 15N relaxation rates were recorded at 298 K on a 1.0 mM sample at pH 7 on Bruker Avance spectrometers operating at 600 and 700 MHz. 15N R1, R2, and steady-state heteronuclear NOEs were measured with pulse sequences as described by Farrow et al. (1994). R2 were measured as a function of refocusing

times (␶CPMG) ranging from 450 to 1150 ␮s with the Carr-PurcellMeiboom-Gill (CPMG) sequence (Mulder et al., 1999; Peng and Wagner, 1994). In all experiments, the water signal was suppressed with “water flipback” scheme (Grzesiek and Bax, 1993). A 1H-15N HSQC experiment was collected 3 days after BsSco1 had been dissolved in D2O to characterize NH exchangeability with water. Fast amide proton exchange were monitored using a 15 N-(CLEANEX-PM)-FHSQC experiment (Hwang et al., 1998) with a mixing time of 100 ms. The experimental relaxation rates were used to map the spectral density function values J(␻H), J(␻N), and J(0) following a procedure available in literature (Peng and Wagner, 1992). The R2 dependence on the ␶CPMG length was analyzed as previously described (Mulder et al., 1999). The overall rotational correlation time ␶m value and the rotational diffusion tensor were estimated from the R2/R1 ratio with the program quadric_diffusion (Bru¨schweiler et al., 1995; Lee et al., 1997), using the 3D structure of BsSco1. Titration of CuSO4 A Varian Cary 50 spectrophotometer and an Avance 600 Bruker Spectrometer were used to follow the binding of copper(II) to BsSco1 through NMR and electronic spectroscopy, respectively. To a 1 mM 15N BsSco1 sample up to 1 equivalent of copper(II) as CuSO4 was added stepwise and UV-visible and 2D 1H-15N HSQC spectra were acquired. X-band EPR spectra were acquired at 120K and 298K on a BRUKER 200D SRC EPR spectrometer. Acknowledgments We thank Prof. R. Udisti for atomic absorption spectroscopy; Prof. G. Moneti for carrying out MS experiments; and Prof. R. Pogni for EPR experiments. Prof. Nigel J. Robinson is acknowledged for his insightful comments. This work was supported by the European Community (SPINE n⬚QLG2-CT-2002-00988, “Structural proteomics in Europe”), by Ente Cassa Risparmio di Firenze, and by MIUR. Received: June 28, 2003 Revised: August 1, 2003 Accepted: August 1, 2003 Published: November 4, 2003 References Andrew, C.R., Yeom, H., Valentine, J.S., Karlsson, B.G., Bonander, N., Van Pouderoyen, G., Canters, G.W., Loehr, J.S., and Sanders-

Structure 1442

Loehr, J. (1994). Raman spectroscopy as an indicator of Cu-S bond length in type 1 and type 2 copper-cysteinate proteins. J. Am. Chem. Soc. 116, 11489–11498. Archer, S.J., Ikura, M., Torchia, D.A., and Bax, A. (1991). An alternative 3D NMR technique for correlation backbone 15N with side chain H␤ resonances in larger proteins. J. Magn. Reson. 95, 636–641. Banci, L., and Rosato, A. (2003). Structural genomics of proteins involved in copper homeostasis. Acc. Chem. Res. 36, 215–221. Banci, L., Bertini, I., and Luchinat, C. (1991). Nuclear and Electron Relaxation. The Magnetic Nucleus-Unpaired Electron Coupling in Solution (Weinheim, Germany: VCH). Bartels, C., Gu¨ntert, P., Billeter, M., and Wu¨thrich, K. (1997). GARANT - a general algorithm for resonance assignment of multidimensional nuclear magnetic resonance spectra. J. Comp. Chem. 18, 139–149. Bax, A., and Davis, D.G. (1985). MLEV-17-based two-dimensional homonuclear magnetization transfer spectroscopy. J. Magn. Reson. 65, 355–360. Beers, J., Glerum, D.M., and Tzagoloff, A. (2002). Purification and characterization of yeast Sco1p, a mitochondrial copper protein. J. Biol. Chem. 277, 22185–22190.

complexed Src homology 2 domain studied by 15N NMR relaxation. Biochemistry 33, 5984–6003. Gagne, S.M., Tsuda, S., Li, M.X., Chandra, M., Smillie, L.B., and Sykes, B.D. (1994). Quantification of the calcium-induced secondary structural changes in the regulatory domain of troponin-C. Protein Sci. 3, 1961–1974. Gardner, K.H., Konrat, R., Rosen, M.K., and Kay, L.E. (1996). An (H)C(CO)NH-TOCSY pulse scheme for sequential assignment of protonated methyl groups in otherwise deuterated 15N, 13C-labeled proteins. J. Biomol. NMR 8, 351–356. Glerum, D.M., Shtanko, A., and Tzagoloff, A. (1996). SCO1 and SCO2 act as high copy suppressors of a mitochondrial copper recruitment defect in Saccharomyces cerevisiae. J. Biol. Chem. 271, 20531– 20535. Grzesiek, S., and Bax, A. (1993). The importance of not saturating H2O in protein NMR. Application to sensitivity enhancement and NOE measurements. J. Am. Chem. Soc. 115, 12593–12594. Gu¨ntert, P., Braun, W., and Wu¨thrich, K. (1991). Efficient computation of three-dimensional protein structures in solution from Nuclear Magnetic Resonance data using the program DIANA and the supporting programs CALIBA, HABAS and GLOMSA. J. Mol. Biol. 217, 517–530.

Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H., Shindyalov, I.N., and Bourne, P.E. (2000). The Protein Data Bank. Nucleic Acids Res. 28, 235–242.

Gu¨ntert, P., Mumenthaler, C., and Wu¨thrich, K. (1997). Torsion angle dynamics for NMR structure calculation with the new program DYANA. J. Mol. Biol. 273, 283–298.

Bertini, I., Luchinat, C., and Parigi, G. (2001). Solution NMR of Paramagnetic Molecules (Amsterdam, Holland: Elsevier).

Harrenga, A., and Michel, H. (1999). The cytochrome c oxidase from Paracoccus denitrificans does not change the metal center ligation upon reduction. J. Biol. Chem. 274, 33296–33299.

Bodenhausen, G., and Ruben, D.J. (1980). Natural abundance nitrogen-15 NMR by enhanced heteronuclear spectroscopy. Chem. Phys. Lett. 69, 185–188. Bru¨schweiler, R., Liao, X., and Wright, P.E. (1995). Long-range motional restrictions in a multidomain zinc-finger protein from anisotropic tumbling. Science 268, 886–889. Canters, G.W., and Gilardi, G. (1993). Engineering type 1 copper sites in proteins. FEBS Lett. 325, 39–48. Capitani, G., Rossmann, R., Sargent, D.F., Grutter, M.G., Richmond, T.J., and Hennecke, H. (2001). Structure of the soluble domain of a membrane-anchored thioredoxin-like protein from Bradyrhizobium japonicum reveals unusual properties. J. Mol. Biol. 311, 1037–1048.

Herrmann, T., Gu¨ntert, P., and Wu¨thrich, K. (2002). Protein NMR structure determination with automated NOE assignment using the new software CANDID and the torsion angle dynamics algorithm DYANA. J. Mol. Biol. 319, 209–227. Hirotsu, S., Abe, Y., Okada, K., Nagahara, N., Hori, H., Nishino, T., and Hakoshima, T. (1999). Crystal structure of a multifunctional 2-Cys peroxiredoxin heme-binding protein 23 kDa/proliferation-associated gene product. Proc. Natl. Acad. Sci. USA 96, 12333–12338. Hofmann, K., and Stoffel, W. (1993). TMbase - A database of membrane spanning proteins segments. Biol. Chem. Hoppe Seyler 347, 166.

Carr, H.S., and Winge, D.R. (2003). Assembly of cytochrome c oxidase within the mitochondrion. Acc. Chem. Res. 36, 309–316.

Holm, L., and Sander, C. (1996). Mapping the protein universe. Science 273, 595–603.

Cavet, J.S., Borrelly, G.P., and Robinson, N.J. (2003). Zn, Cu and Co in cyanobacteria: selective control of metal availability. FEMS Microbiol. Rev. 27, 165–181.

Huffman, D.L., and O’Halloran, T.V. (2001). Function, structure, and mechanism of intracellular copper trafficking proteins. Annu. Rev. Biochem. 70, 677–701.

Choi, H.J., Kang, S.W., Yang, C.H., Rhee, S.G., and Ryu, S.E. (1998). Crystal structure of a novel human peroxidase enzyme at 2.0 A˚ resolution. Nat. Struct. Biol. 5, 400–406.

Hwang, T.L., Mori, S., Shaka, A.J., and Van Zijl, P.C.M. (1998). Application of phase-modulated CLEAN chemical exchange spectroscopy (CLEANEX-PM) to detect water-protein proton exchange and intermolecular NOEs. J. Am. Chem. Soc. 119, 6203–6204.

Culotta, V.C., Klomp, L.W., Strain, J., Casareno, R.L., Krems, B., and Gitlin, J.D. (1997). The copper chaperone for superoxide dismutase. J. Biol. Chem. 272, 23469–23472. Deleage, G., and Geourjon, C. (1993). An interactive graphic program for calculating the secondary structures content of proteins from circular dichroism spectrum. Comput. Appl. Biosci. 9, 197–199. Dickinson, E.K., Adams, D.L., Schon, E.A., and Glerum, D.M. (2000). A human SCO2 mutation helps define the role of Sco1p in the cytochrome oxidase assembly pathway. J. Biol. Chem. 275, 26780– 26785. Eccles, C., Gu¨ntert, P., Billeter, M., and Wu¨thrich, K. (1991). Efficient analysis of protein 2D NMR spectra using the software package EASY. J. Biomol. NMR 1, 111–130. Einstein, A. (1956). Investigations on the Theory of the Brownian Movement (New York: Dover). Ellis, H.R., and Poole, L.B. (1997). Roles for the two cysteine residues of AhpC in catalysis of peroxide reduction by alkyl hydroperoxide reductase. Biochemistry 36, 13349–13356. Farrow, N.A., Muhandiram, R., Singer, A.U., Pascal, S.M., Kay, C.M., Gish, G., Shoelson, S.E., Pawson, T., Forman-Kay, J.D., and Kay, L.E. (1994). Backbone dynamics of a free and phosphopeptide-

Kay, L.E., Ikura, M., Tschudin, R., and Bax, A. (1990). Three-dimensional triple-resonance NMR spectroscopy of isotopically enriched proteins. J. Magn. Reson. 89, 496–514. Kay, L.E., Xu, G.Y., Singer, A.U., Muhandiram, D.R., and FormanKay, J.D. (1993). A gradient-enhanced HCCH-TOCSY experiment for recording side-chains 1H and 13C correlations in H2O samples of proteins. J. Magn. Reson. Ser. B 101, 333–337. Koradi, R., Billeter, M., and Wu¨thrich, K. (1996). MOLMOL: a program for display and analysis of macromolecular structure. J. Mol. Graph. 14, 51–55. Kroneck, P.M., Antholine, W.A., Riester, J., and Zumft, W.G. (1988). The cupric site in nitrous oxide reductase contains a mixed-valence [Cu(II),Cu(I)] binuclear center: a multifrequency electron paramagnetic resonance investigation. FEBS Lett. 242, 70–74. Krummeck, G., and Rodel, G. (1990). Yeast SCO1 protein is required for a post-translational step in the accumulation of mitochondrial cytochrome c oxidase subunits I and II. Curr. Genet. 18, 13–15. Laskowski, R.A., Rullmann, J.A.C., MacArthur, M.W., Kaptein, R., and Thornton, J.M. (1996). AQUA and PROCHECK-NMR: programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 8, 477–486.

Solution Structure of Sco1 from B. subtilis 1443

Laskowski, R.A., MacArthur, M.W., and Thornton, J.M. (1998). Validation of protein models derived from experiment. Curr. Opin. Struct. Biol. 8, 631–639. Lee, L.K., Rance, M., Chazin, W.J., and Palmer, A.G., III. (1997). Rotational diffusion anisotropy of proteins from simultaneous analysis of 15N and 13C alpha nuclear spin relaxation. J. Biomol. NMR 9, 287–298. Loferer, H., and Hennecke, H. (1994). Expression, purification and functional properties of a soluble form of Bradyrhizobium japonicum TlpA, a thioredoxin-like protein. Eur. J. Biochem. 223, 339–344. Loferer, H., Bott, M., and Hennecke, H. (1993). Bradyrhizobium japonicum TlpA, a novel membrane-anchored thioredoxin-like protein involved in the biogenesis of cytochrome aa3 and development of symbiosis. EMBO J. 12, 3373–3383. Lu, Y., Roe, J.A., Bender, C.J., Peisach, C.J., Banci, L., Bertini, I., Gralla, E.B., and Valentine, J. (1996). New type 2 copper-cysteinate proteins. Copper site histidine-to-cysteine mutants of yeast copperzinc superoxide dismutase. Inorg. Chem. 35, 1692–1700. Martin, J.L. (1995). Thioredoxin––a fold for all reasons. Structure 3, 245–250. Mattatall, N.R., Jazairi, J., and Hill, B.C. (2000). Characterization of YpmQ, an accessory protein required for the expression of cytochrome c oxidase in Bacillus subtilis. J. Biol. Chem. 275, 28802– 28809. Mulder, F.A., Van Tilborg, P.J., Kaptein, R., and Boelens, R. (1999). Microsecond time scale dynamics in the RXR DNA-binding domain from a combination of spin-echo and off-resonance rotating frame relaxation measurements. J. Biomol. NMR 13, 275–288. Nittis, T., George, G.N., and Winge, D.R. (2001). Yeast Sco1, a protein essential for cytochrome c oxidase function is a Cu(I)-binding protein. J. Biol. Chem. 276, 42520–42526. Papadopoulou, L.C., Sue, C.M., Davidson, M.M., Tanji, K., Nishino, I., Sadlock, J.E., Krishna, S., Walker, W., Selby, J., Glerum, D.M., et al. (1999). Fatal infantile cardioencephalomyopathy with COX deficiency and mutations in SCO2, a COX assembly gene. Nat. Genet. 23, 333–337. Pearlman, D.A., Case, D.A., Caldwell, J.W., Ross, W.S., Cheatham, T.E., Ferguson, D.M., Seibel, G.L., Singh, U.C., Weiner, P.K., and Kollman, P.A. (1997). AMBER 5.0 (San Francisco, California: University of California). Pelton, J.G., Torchia, D.A., Meadow, N.D., and Roseman, S. (1993). Tautomeric states of the active-site histidines of phosphorylated and unphospohorylated IIIglc, a signal transducing protein from Escherichia coli, using two dimensional heteronuclear NMR techniques. Protein Sci. 2, 543–558. Peng, J.W., and Wagner, G. (1992). Mapping of spectral density function using heteronuclear NMR relaxation measurements. J. Magn. Reson. 98, 308–332. Peng, J.W., and Wagner, G. (1994). Investigation of protein motions via relaxation measurements. Methods Enzymol. 239, 563–596. Poyton, R.O., Goehring, B., Droste, M., Sevarino, K.A., Allen, L.A., and Zhao, X.J. (1995). Cytochrome-c oxidase from Saccharomyces cerevisiae. Methods Enzymol. 260, 97–116.

der Oost, J. (1991). The Bacillus subtilis cytochrome-c oxidase. Variations on a conserved protein theme. Eur. J. Biochem. 195, 517–525. Schulze, M., and Rodel, G. (1988). SCO1, a yeast nuclear gene essential for accumulation of mitochondrial cytochrome c oxidase subunit II. Mol. Gen. Genet. 211, 492–498. Solioz, M., and Vulpe, C.D. (1996). CPx-type ATPases: a class of P-type ATPases that pump heavy metals. Trends Biochem. Sci. 21, 237–241. Soulimane, T., Buse, G., Bourenkov, G.P., Bartunik, H.D., Huber, R., and Than, M.E. (2000). Structure and mechanism of the aberrant ba(3)-cytochrome c oxidase from thermus thermophilus. EMBO J. 19, 1766–1776. Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., and Yoshikawa, S. (1996). The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. Science 272, 1136–1144. Tusnady, G.E., and Simon, I. (1998). Principles governing amino acid composition of integral membrane proteins: application to topology prediction. J. Mol. Biol. 283, 489–506. Tzagoloff, A., and Dieckmann, C.L. (1990). PET genes of Saccharomyces cerevisiae. Microbiol. Rev. 54, 211–225. Valentine, J.S., and Gralla, E.B. (1997). Delivering copper inside yeast and human cells. Science 278, 817–818. Valnot, I., Osmond, S., Gigarel, N., Mehaye, B., Amiel, J., CormierDaire, V., Munnich, A., Bonnefont, J.P., Rustin, P., and Rotig, A. (2000). Mutations of the SCO1 gene in mitochondrial cytochrome c oxidase deficiency with neonatal-onset hepatic failure and encephalopathy. Am. J. Hum. Genet. 67, 1104–1109. Vuister, G.W., and Bax, A. (1993). Quantitative J correlation: a new approach for measuring homonuclear three-bond J(HNH␣) coupling constants in 15N enriched proteins. J. Am. Chem. Soc. 115, 7772– 7777. Wider, G., Macura, S., Kumar, A., Ernst, R.R., and Wu¨thrich, K. (1984). Homonuclear two-dimensional 1H NMR of proteins. J. Magn. Reson. 56, 207–234. Wider, G., Neri, D., Otting, G., and Wu¨thrich, K. (1989). A heteronuclear three-dimensional NMR experiment for measurements of small heteronuclear coupling constants in biological macromolecules. J. Magn. Reson. 85, 426–431. Winstedt, L., and Von Wachenfeldt, C. (2000). Terminal oxidases of Bacillus subtilis strain 168: one quinol oxidase, cytochrome aa(3) or cytochrome bd, is required for aerobic growth. J. Bacteriol. 182, 6557–6564. Wishart, D.S., and Sykes, B.D. (1994). The 13C chemical shift index: a simple method for the identification of protein secondary structure using 13C chemical shift data. J. Biomol. NMR 4, 171–180. Wood, Z.A., Poole, L.B., Hantgan, R.R., and Karplus, P.A. (2002). Dimers to doughnuts: redox-sensitive oligomerization of 2-cysteine peroxiredoxins. Biochemistry 41, 5493–5504.

Puig, S., and Thiele, D.J. (2002). Molecular mechanisms of copper uptake and distribution. Curr. Opin. Chem. Biol. 6, 171–180.

Accession Numbers

Rentzsch, A., Krummeck-Weiss, G., Hofer, A., Bartuschka, A., Ostermann, K., and Rodel, G. (1999). Mitochondrial copper metabolism in yeast: mutational analysis of Sco1p involved in the biogenesis of cytochrome c oxidase. Curr. Genet. 35, 103–108.

Resonance assignments and the derived atomic coordinates for a family of acceptable structures with all restraints used in structure determination are available at the BioMagResBank (accession number 5742) and the Protein Data Bank (PDB ID 1ON4), respectively.

Rosenzweig, A.C., and O’Halloran, T.V. (2000). Structure and chemistry of the copper chaperone proteins. Curr. Opin. Chem. Biol. 4, 140–147. Sali, A., and Blundell, T.L. (1993). Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual, Second Edition (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press). Saraste, M., Metso, T., Nakari, T., Jalli, T., Lauraeus, M., and Van