Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Assembly of Cytochrome-c Oxidase in the Absence of Assembly Protein Surf1p Leads to Loss of the Active Site Heme

2005; Elsevier BV; Volume: 280; Issue: 18 Linguagem: Inglês

10.1074/jbc.c500061200

ISSN

1083-351X

Autores

Daniel J. Smith, Jimmy Gray, Larkin Mitchell, William E. Antholine, Jonathan P. Hosler,

Tópico(s)

ATP Synthase and ATPases Research

Resumo

Surf1p is a protein of the inner membrane of mitochondria that functions in the assembly of cytochrome-c oxidase. The specifics of the role of Surf1p have remained unresolved. Numerous mutations in human Surf1p lead to severe mitochondrial disease. A homolog of human Surf1p is encoded by the genome of the α-proteobacterium Rhodobacter sphaeroides, which synthesizes a mitochondrial-like aa3-type cytochrome-c oxidase. The gene for Surf1p was deleted from the genome of R. sphaeroides. The resulting aa3-type oxidase was purified and analyzed by biochemical methods plus optical and EPR spectroscopy. The oxidase that assembled in the absence of Surf1p was composed of three subpopulations with structurally distinct heme a3-Cu active sites. 50% of the oxidase lacked heme a3, 10–15% contained heme a3 but lacked CuBB, and 35–40% had a normal heme a3 -CuB active site with normal activity. CuA assembly was unaffected. All of the oxidase contained low-spin heme a, but the environment of the heme a center was slightly altered in the 50% of the enzyme that lacked heme a3. Introduction of a normal copy of the gene for Surf1p on an exogenous plasmid resulted in a single population of normally assembled, highly active enzyme. The data indicate that Surf1p plays a role in facilitating the insertion of heme a3 into the active site of cytochrome-c oxidase. The results suggest that maturation of the heme a3-CuB center is a step that limits the association of subunits I and II in the assembly of mitochondrial cytochrome oxidase. Surf1p is a protein of the inner membrane of mitochondria that functions in the assembly of cytochrome-c oxidase. The specifics of the role of Surf1p have remained unresolved. Numerous mutations in human Surf1p lead to severe mitochondrial disease. A homolog of human Surf1p is encoded by the genome of the α-proteobacterium Rhodobacter sphaeroides, which synthesizes a mitochondrial-like aa3-type cytochrome-c oxidase. The gene for Surf1p was deleted from the genome of R. sphaeroides. The resulting aa3-type oxidase was purified and analyzed by biochemical methods plus optical and EPR spectroscopy. The oxidase that assembled in the absence of Surf1p was composed of three subpopulations with structurally distinct heme a3-Cu active sites. 50% of the oxidase lacked heme a3, 10–15% contained heme a3 but lacked CuBB, and 35–40% had a normal heme a3 -CuB active site with normal activity. CuA assembly was unaffected. All of the oxidase contained low-spin heme a, but the environment of the heme a center was slightly altered in the 50% of the enzyme that lacked heme a3. Introduction of a normal copy of the gene for Surf1p on an exogenous plasmid resulted in a single population of normally assembled, highly active enzyme. The data indicate that Surf1p plays a role in facilitating the insertion of heme a3 into the active site of cytochrome-c oxidase. The results suggest that maturation of the heme a3-CuB center is a step that limits the association of subunits I and II in the assembly of mitochondrial cytochrome oxidase. Cytochrome-c oxidase (CcO) 1The abbreviations used are: CcO, cytochrome-c oxidase; Strep/Spec, streptomycin and spectinomycin. functions as the terminal member of the respiratory electron transfer chain in mitochondria. Electrons from soluble cytochrome c are first transferred to a dimeric copper center in subunit II, CuA, then on to six-coordinate heme a in subunit I, and finally to the buried heme a3-CuB site, also located in subunit I, where O2 is reduced to water (1Saraste M. Science. 1999; 283: 1488-1493Crossref PubMed Scopus (1030) Google Scholar). Some of the energy of electron transfer is used to pump protons through the protein, across its host membrane. Rhodobacter sphaeroides, a member of the α-subgroup of the proteobacteria that gave rise to mitochondria, synthesizes an aa3-type oxidase with high genetic and structural similarity to the subunits of the catalytic core (I, II, and III) of mammalian CcO (2Shapleigh J.P. Gennis R.B. Mol. Microbiol. 1992; 6: 635-642Crossref PubMed Scopus (57) Google Scholar, 3Cao J. Hosler J. Shapleigh J. Revzin A. Ferguson-Miller S. J. Biol. Chem. 1992; 267: 24273-24278Abstract Full Text PDF PubMed Google Scholar, 4Svensson-Ek M. Abramson J. Larsson G. Törnroth S. Brzezinski P. Iwata S. J. Mol. Biol. 2002; 321: 329-339Crossref PubMed Scopus (478) Google Scholar). R. sphaeroides has proven to be highly useful for producing engineered forms of its mitochondrial-like oxidase, thus providing experimental models for elucidating structure/function relationships for the catalytic core of CcO. Numerous proteins that are not members of the final complex are specifically required for the assembly of CcO in eukaryotic cells (5Tzagoloff A. Dieckmann C.L. Microbiol. Rev. 1990; 54: 211-225Crossref PubMed Google Scholar, 6Carr H.S. Winge D.R. Acc. Chem. Res. 2003; 36: 309-316Crossref PubMed Scopus (196) Google Scholar). The genome of R. sphaeroides encodes homologs of five of these assembly proteins, which are also present in yeast and human cells. The presence of these proteins in R. sphaeroides and related bacteria suggests that they play a fundamental role in the assembly of the catalytic core. Including this report, the evidence for a role in CcO assembly for at least four of these five bacterial proteins is compelling. Cox10p and Cox15p of R. sphaeroides form a membrane complex that coverts heme B to heme A (7Brown B.M. Wang Z. Brown K.R. Cricco J.A. Hegg E.L. Biochemistry. 2004; 43: 13541-13548Crossref PubMed Scopus (33) Google Scholar). Cox11p is a membrane-bound copper-binding protein (8Carr H.S. George G.N. Winge D.R. J. Biol. Chem. 2002; 277: 31237-31242Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar); analysis of a R. sphaeroides strain lacking cox11 established that Cox11p is required for the assembly of CuB (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). PrrC is another copper-binding protein of R. sphaeroides with homology to eukaryotic Sco1 (10McEwan A.G. Lewin A. Davy S.L. Boetzel R. Leech A. Walker D. Wood T. Moore G.R. FEBS Lett. 2002; 518: 10-16Crossref PubMed Scopus (48) Google Scholar), the copper protein proposed to function in the assembly of CuA in mitochondria (6Carr H.S. Winge D.R. Acc. Chem. Res. 2003; 36: 309-316Crossref PubMed Scopus (196) Google Scholar). However, PrrC is not required for the assembly of CuA in the aa3-type oxidase of R. sphaeroides. 2J. P. Hosler, unpublished observation. Here, we report the analysis of CcO synthesized in the absence of Surf1p of R. sphaeroides, the fifth of the conserved CcO assembly proteins present in this bacterium. It is well established that Surf1p is essential for the assembly of normal amounts of mitochondrial CcO (11Williams S.L. Valnot I. Rustin P. Taanman J.W. J. Biol. Chem. 2004; 279: 7462-7469Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar, 12Nijtmans L.G. Artal Sanz M. Bucko M. Farhoud M.H. Feenstra M. Hakkaart G.A. Zeviani M. Grivell L.A. FEBS Lett. 2001; 498: 46-51Crossref PubMed Scopus (59) Google Scholar, 13Barrientos A. Korr D. Tzagoloff A. EMBO J. 2002; 21: 43-52Crossref PubMed Scopus (141) Google Scholar, 14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar). At least 40 different mutations in Surf1p of humans lead to CcO deficiency, resulting in severe mitochondrial disease (14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar). In fact, human Surf1p has been linked with a greater number of defects leading to human disease than any of the other CcO assembly proteins (14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar). Surf1p is anchored in the inner mitochondrial membrane (or the bacterial cytoplasmic membrane) by two predicted transmembrane helices, one each at the N and C termini of the protein (15Mashkevich G. Repetto B. Glerum D.M. Jin C. Tzagoloff A. J. Biol. Chem. 1997; 272: 14356-14364Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar). The intervening extramembrane domain faces the intermembrane space in mitochondria (15Mashkevich G. Repetto B. Glerum D.M. Jin C. Tzagoloff A. J. Biol. Chem. 1997; 272: 14356-14364Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar), which is analogous to the periplasmic space of R. sphaeroides. The predicted sizes of human (GenBank™ accession number NP_003163) and R. sphaeroides Surf1p (AY918925) are similar (300 versus 262 residues). Using EMBOSS-Align (European Bioinformatics Institute), the DNA-predicted amino acid sequence of R. sphaeroides Surf1p is 31% identical and 45% similar to human Surf1p. Essentially all of this homology is found in the extramembrane domains of the two proteins; the sequences of the transmembrane helices are not conserved. Surf1p of R. sphaeroides is no less similar to human Surf1p than is Shy1p of Saccharomyces cerevisiae, the yeast homolog of Surf1p (15Mashkevich G. Repetto B. Glerum D.M. Jin C. Tzagoloff A. J. Biol. Chem. 1997; 272: 14356-14364Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar). Studies in yeast and cultured human cells indicate roles for Surf1p in promoting the passage of subunit I through the assembly process, particularly the association of subunit II with a protein subassembly containing subunit I plus one or more of the nuclear encoded structural subunits present in the mitochondrial oxidase (11Williams S.L. Valnot I. Rustin P. Taanman J.W. J. Biol. Chem. 2004; 279: 7462-7469Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar, 12Nijtmans L.G. Artal Sanz M. Bucko M. Farhoud M.H. Feenstra M. Hakkaart G.A. Zeviani M. Grivell L.A. FEBS Lett. 2001; 498: 46-51Crossref PubMed Scopus (59) Google Scholar, 13Barrientos A. Korr D. Tzagoloff A. EMBO J. 2002; 21: 43-52Crossref PubMed Scopus (141) Google Scholar, 14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar, 16Perez-Martinez X. Broadley S.A. Fox T.D. EMBO J. 2003; 22: 5951-5961Crossref PubMed Scopus (153) Google Scholar, 17Barrientos A. Zambrano A. Tzagoloff A. EMBO J. 2004; 23: 3472-3482Crossref PubMed Scopus (165) Google Scholar). Here, we present evidence that Surf1p assists in the assembly of heme a3 into the active site of R. sphaeroides CcO. The results also provide the likely explanation for previous proposals of Surf1p function in the assembly of mitochondrial CcO. Construction of the surf1 Deletion Strain DS003 and the Surf1p Expression Plasmid pDS306—A 1.8-kbp PCR product containing surf1 was obtained from genomic DNA isolated from R. sphaeroides 2.4.1 (18Choudhary M. Fu Y.X. Mackenzie C. Kaplan S. J. Bacteriol. 2004; 186: 2019-2027Crossref PubMed Scopus (15) Google Scholar) using primers that included EcoRI restriction sites (Surf1EcoRIforward (5′-3′) = GGGAATTCCGGCCTGGTTCTGGAGCTTCTTCAAGCACGCG; Surf1EcoRIreverse = GGGAATTCCCCGAGGTGGCGCCGACGATGGTGATCC). The gel-purified fragment was restricted with EcoRI and ligated into the EcoRI site of pUC18 to create pDS301. A 397-bp fragment internal to surf1 was removed by restricting pDS301 with EcoNI and NcoI, and blunt ends were created with T4 DNA polymerase. A 2.1-kbp SmaI/EcoRV fragment containing the Strep/Spec ο cassette from pUI1638 (19Eraso J.M. Kaplan S. J. Bacteriol. 1994; 176: 32-43Crossref PubMed Google Scholar) was ligated into blunt-ended pDS301 to create pDS304. A 3.5-kbp EcoRI fragment from pDS304 (the original 1.8-kbp PCR fragment, minus the 0.4-kbp deletion in surf1, plus the 2.1-kbp Strep/Spec resistance gene) was cloned into the suicide plasmid pSUP202 (20Simon R. Priefer U. Puhler A. Bio/Technology. 1983; 1: 784-791Crossref Scopus (5642) Google Scholar) to create pDS305. The pDS305 suicide plasmid was conjugated into R. sphaeroides 2.4.1 by established methods (21Donohue T.J. Kaplan S. Methods Enzymol. 1991; 204: 459-485Crossref PubMed Scopus (77) Google Scholar). The Strep/Spec-resistant exconjugate colonies were replica-plated onto media containing 1.0 μg/ml tetracycline to identify cells in which double crossover events had eliminated the plasmid from the genome leaving behind the modified surf1. Tetracycline-sensitive (double crossover) exconjugates were analyzed by PCR using primers just outside of the Surf1EcoRIforward and Surf1EcoRIreverse primers for the presence of an ∼3.6-kbp fragment containing modified surf1 plus the Strep/Spec resistance gene. Three such isolates were identified, and one was carried forward for use as the surf1 deletion strain, DS003. The 1.8-kbp EcoRI fragment containing surf1 from pDS301 was cloned into the broad host range vector pBBR1MCS-2 (22Kovach M.E. Elzer P.H. Hill D.S. Robertson G.T. Farris M.A. Roop II, R.M. Peterson K.M. Gene. 1995; 166: 175-176Crossref PubMed Scopus (2730) Google Scholar) to create the Surf1p expression plasmid pDS306. Other—Bacterial growth and oxidase purification (23Zhen Y. Qian J. Follmann K. Hayward T. Nilsson T. Dahn M. Hilmi Y. Hamer A.G. Hosler J.P. Ferguson-Miller S. Protein Expression Purif. 1998; 13: 326-336Crossref PubMed Scopus (72) Google Scholar, 24Mills D.A. Tan Z. Ferguson-Miller S. Hosler J. Biochemistry. 2003; 42: 7410-7417Crossref PubMed Scopus (38) Google Scholar) and activity measurements (25Bratton M.R. Pressler M.A. Hosler J.P. Biochemistry. 1999; 38: 16236-16245Crossref PubMed Scopus (69) Google Scholar) were as described previously. Analyses of heme A content were as described previously (26Hosler J.P. Fetter J. Tecklenburg M.M. Espe M. Lerma C. Ferguson-Miller S. J. Biol. Chem. 1992; 267: 24264-24272Abstract Full Text PDF PubMed Google Scholar), using the α-band of the reduced absolute spectra of the oxidase forms to determine oxidase concentrations. An extinction coefficient of 40 mm–1 cm–1 (27Vanneste W.H. Biochemistry. 1966; 5: 838-848Crossref PubMed Scopus (191) Google Scholar) was used for the wild-type oxidase and the oxidase isolated from DS020 (see below). The contribution to α-band absorbance by heme a3 is ∼20% (27Vanneste W.H. Biochemistry. 1966; 5: 838-848Crossref PubMed Scopus (191) Google Scholar). Once it became apparent that the oxidase that assembles in the absence of Surf1p lacked 50% of its heme a3 (see below), an extinction coefficient of 36 mm–1 cm–1 was then used for this oxidase form in order to more closely estimate the concentration of the protein. The binding of CO and CN– were performed as described (25Bratton M.R. Pressler M.A. Hosler J.P. Biochemistry. 1999; 38: 16236-16245Crossref PubMed Scopus (69) Google Scholar). EPR spectra were obtained as noted in the legend to Fig. 2. Biochemical and Spectroscopic Characteristics of the aa3-type Cytochrome-c Oxidase That Assembles in the Absence of Surf1p—In the R. sphaeroides genome, surf1 is located immediately adjacent to the gene for subunit III of the aa3-type oxidase, as the first gene in what appears to be a four-gene operon. Three downstream genes encode a putative threonine synthase, a peptidase of the M16 family, and a putative alanine acetyltransferase. In the surf1 deletion strain, DS003, a 0.4-kbp fragment encoding residues Trp-64 through Pro-195 of Surf1p (i.e. most of the conserved extramembrane domain) was deleted and replaced by a gene encoding resistance to streptomycin and spectinomycin. DS003 grew as well as wild-type strain 2.4.1 under aerobic conditions, indicating that the presence of the antibiotic resistance gene at this position was not cytotoxic. The growth rate of DS003 was the same whether or not the minimal growth media was supplemented with l-threonine. Thus, the expression of the downstream threonine synthase was sufficient for cell needs despite possible polar effects from the insertion of the antibiotic resistance gene. A strain (DS009) for over-expression of the aa3-type CcO in the absence of Surf1p was created by introducing the expression plasmid pRKpAH1H32 (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), which contains the genes for subunits I, II, and III as well as those for the assembly proteins Cox10p and Cox11p, into DS003. DS009 also grew at a normal rate. This indicates that expressing the aa3-type oxidase in the absence of Surf1p is not harmful to the cell but, other than that, it yields no information about the assembly of the oxidase. Even the complete absence of the aa3-type oxidase from R. sphaeroides does not affect the aerobic growth rate, because a cbb3-type CcO is also synthesized to high levels during aerobic growth (28García-Horsman J.A. Berry E. Shapleigh J.P. Alben J.O. Gennis R.B. Biochemistry. 1994; 33: 3113-3119Crossref PubMed Scopus (131) Google Scholar). The aa3-type oxidase was isolated from purified DS009 membranes via its histidine tag on the C terminus of subunit I (23Zhen Y. Qian J. Follmann K. Hayward T. Nilsson T. Dahn M. Hilmi Y. Hamer A.G. Hosler J.P. Ferguson-Miller S. Protein Expression Purif. 1998; 13: 326-336Crossref PubMed Scopus (72) Google Scholar, 29Mitchell D.M. Gennis R.B. FEBS Lett. 1995; 368: 148-150Crossref PubMed Scopus (144) Google Scholar) and termed ΔSurf1p CcO. Extensive use of this tag to isolate the wild-type oxidase and many mutant oxidase forms has shown that the tag itself does not alter the structural or functional properties of the oxidase. The cytochrome c-driven O2 reduction activity of purified ΔSurf1p CcO (Vmax = 706 ± 29 s–1) was ∼35% that of the wild-type enzyme (Vmax = 2074 ± 45 s–1). This indicated an overall decrease in activity or the presence of at least two oxidase forms, one active and one inactive. Densitometry of protein gels (not shown) showed that the subunit composition of the CcO complex that assembled in the absence of Surf1p was entirely normal. Inspection of the absolute spectrum of dithionite-reduced ΔSurf1p CcO showed a significant decrease in the ratio of the amplitudes of the Soret peak at ∼444 nm and the α peak at ∼605 nm (Fig. 1A; Table I). Because the α peak absorbance is primarily due to heme a (80%) and the Soret absorbance is more equally contributed by both hemes a and a3 (27Vanneste W.H. Biochemistry. 1966; 5: 838-848Crossref PubMed Scopus (191) Google Scholar), this decrease in the Soret/α amplitude ratio suggested a loss of heme a3. The Soret/α value of ΔSurf1p CcO is approximately half way between that of the wild-type oxidase and the theoretical value for an oxidase lacking all heme a3, calculated from the extinction values of Vanneste (27Vanneste W.H. Biochemistry. 1966; 5: 838-848Crossref PubMed Scopus (191) Google Scholar).Table ISpectroscopic characterization of CcO forms that assemble with and without Surf1pCcO formSoret/α from reduced spectrumHeme A content (heme A per monomer)Heme a3 contentNormal heme a3-CuB contentWild-type (normal aa3)5.3-5.5 (n = 5)2.04 ± 0.02 (n = 3)100% (81.6 ± 3.7; n = 3)100% (18.4 ± 1.3; n = 3)ΔSurf1p aa3 (from DS009)4.5-4.8 (n = 10)1.46 ± 0.06 (n = 3)51% (41.2 ± 5.0; n = 3)37% (6.8 ± 0.3; n = 4)DS020 (Surf1p complementation)5.4-5.5 (n = 3)1.95 ± 0.13 (n = 3)100% (81.7 ± 0.4; n = 3)99% (18.3 ± 1.9; n = 3) Open table in a new tab The heme A content of ΔSurf1p CcO was measured to be 75% that of the wild-type oxidase by pyridine hemochrome analysis (Table I). Because the content of six-coordinate heme a appears normal based on the amplitude of the α peak and the Soret/α value, this indicated the loss of 50% of the heme a3 of ΔSurf1p CcO. Measurements of the extent of carbon monoxide (CO) binding to ΔSurf1p CcO confirmed this loss of heme a3. The CO difference spectrum (a2+ a 2+–CO minus a32+ a3 2+) is solely due to five-coordinate heme a3, because six-coordinate heme a does not bind CO (30Rich P.R. Moody A.J. Graber P. Milazzo G. Bioenergetics. Birkhauser Verlag Basel, Basel, Switzerland1997: 418-456Crossref Google Scholar) and its absorbance contribution is subtracted. From the amplitude of the trough at 472–447 nm in the CO difference spectrum, it was determined that ΔSurf1p CcO bound ∼50% less CO than the same amount of wild-type CcO, consistent with a heme a3 content of 50% less than normal. The CO difference spectrum of wild-type CcO contains a peak at 430 nm due to the absorbance of the low-spin heme a3–CO adduct (Fig. 1B) (30Rich P.R. Moody A.J. Graber P. Milazzo G. Bioenergetics. Birkhauser Verlag Basel, Basel, Switzerland1997: 418-456Crossref Google Scholar). This peak is also present in the CO difference spectrum of ΔSurf1p CcO, but an absorbance shoulder ∼419 nm is also present. This indicates the presence of a small amount of heme a3 in ΔSurf1p CcO that resists reduction by sodium dithionite and does not bind CO (27Vanneste W.H. Biochemistry. 1966; 5: 838-848Crossref PubMed Scopus (191) Google Scholar). A sensitive optical assay for the presence of a normally assembled heme a3-CuB active site is the ability of CN– to bind to the reduced enzyme (25Bratton M.R. Pressler M.A. Hosler J.P. Biochemistry. 1999; 38: 16236-16245Crossref PubMed Scopus (69) Google Scholar, 30Rich P.R. Moody A.J. Graber P. Milazzo G. Bioenergetics. Birkhauser Verlag Basel, Basel, Switzerland1997: 418-456Crossref Google Scholar, 31Mitchell R. Moody A.J. Rich P.R. Biochemistry. 1995; 34: 7576-7585Crossref PubMed Scopus (26) Google Scholar). Cyanide moves to the active site as neutral HCN, where it dissociates to allow CN– to bind to heme a3. In the reduced active site, the presence of CuB is required to bind the resulting H+, presumably on a hydroxyl anion bound to the copper (31Mitchell R. Moody A.J. Rich P.R. Biochemistry. 1995; 34: 7576-7585Crossref PubMed Scopus (26) Google Scholar, 32Mitchell R. Rich P.R. Biochim. Biophys. Acta. 1994; 1186: 19-26Crossref PubMed Scopus (167) Google Scholar). The CN– difference spectrum (a2+ a32+–CN– minus a2+ a32+) of the reduced, normal oxidase shows a characteristic peak at 590 nm (resulting from the low-spin heme a3–CN– adduct) and a trough at 613 nm (Fig. 1C) (30Rich P.R. Moody A.J. Graber P. Milazzo G. Bioenergetics. Birkhauser Verlag Basel, Basel, Switzerland1997: 418-456Crossref Google Scholar, 31Mitchell R. Moody A.J. Rich P.R. Biochemistry. 1995; 34: 7576-7585Crossref PubMed Scopus (26) Google Scholar). The 590–613 nm absorption of ΔSurf1p CcO was measured to be 37% that of wild-type CcO (Table I), indicating that slightly more than one-third of the ΔSurf1p CcO contained a normal active site. This agreed well with the amount of normal oxidase (35%) indicated by activity measurements of ΔSurf1p CcO. As a negative control, the oxidase isolated from our bacterial strain lacking the copper chaperone Cox11p, an enzyme that contains heme a3 but lacks CuB (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), failed to bind any CN– in its reduced form (data not shown). Complementation of the Deletion—The Surf1p expression vector pDS306, which is compatible with the pRK415-based CcO expression plasmid pRKpAH1H32, was introduced into DS009 to create strain DS020. The aa3-type oxidase isolated from DS020 had wild-type activity (Vmax = 2195 ± 133 s–1), a normal heme A content, and a structurally normal active site based on CN– and CO binding (Table I). Thus, a single plasmid-borne copy of surf1 complements the deletion of the genomic copy of the gene. In a separate control experiment, the pRKpAH1H32 plasmid expressing ΔSurf1p CcO was extracted from DS009 and conjugated into YZ200, a strain of R. sphaeroides lacking the genomic copies of the genes for subunits II and III, as well as the assembly proteins Cox10p and Cox11p (23Zhen Y. Qian J. Follmann K. Hayward T. Nilsson T. Dahn M. Hilmi Y. Hamer A.G. Hosler J.P. Ferguson-Miller S. Protein Expression Purif. 1998; 13: 326-336Crossref PubMed Scopus (72) Google Scholar). The CcO isolated from this strain (DS021) via the histidine tag present on the subunit I expressed from the plasmid was completely normal with high activity (data not shown). This demonstrated that the defective CcO isolated from DS009 could not have resulted from spontaneous alteration of any of the five genes present in pRKpAH1H32. Together, these results confirmed that the absence of Surf1p, and not another gene product, was responsible for the defective assembly of the heme a3-CuB active site of the aa3-type CcO. EPR Spectroscopy—EPR spectroscopy provides direct information about the environment and quantity of CuA and heme a in CcO (26Hosler J.P. Fetter J. Tecklenburg M.M. Espe M. Lerma C. Ferguson-Miller S. J. Biol. Chem. 1992; 267: 24264-24272Abstract Full Text PDF PubMed Google Scholar, 33Stevens T.H. Martin C.T. Wang H. Brudvig G.W. Scholes C.P. Chan S.I. J. Biol. Chem. 1982; 257: 12106-12113Abstract Full Text PDF PubMed Google Scholar, 34Aasa R. Albracht P.J. Falk K.E. Lanne B. Vanngard T. Biochim. Biophys. Acta. 1976; 422: 260-272Crossref PubMed Scopus (135) Google Scholar). The structure and the quantity of the CuA center in subunit II of ΔSurf1p CcO appeared completely normal (Fig. 2). The gz component of the heme a signal of wild-type CcO appears as a sharp peak at g = 2.81 (Fig. 2) (26Hosler J.P. Fetter J. Tecklenburg M.M. Espe M. Lerma C. Ferguson-Miller S. J. Biol. Chem. 1992; 267: 24264-24272Abstract Full Text PDF PubMed Google Scholar). This signal was also present in ΔSurf1p CcO, but its amplitude was ∼50% that of the gz signal for heme a of wild-type CcO. An additional, broadened signal was observed immediately downfield of the gz signal in the spectrum of ΔSurf1p CcO. In a separate experiment, the g ≃ 3 region of wild-type and ΔSurf1p CcO was more highly resolved (Fig. 2, inset). Both spectra show a broad peak centered at g = 2.89 in addition to the g = 2.81 signal. However, although the g = 2.89 signal is a minor component in the spectrum of wild-type CcO, integration showed that the broad g = 2.89 signal and the more narrow g = 2.81 signals each contributed ∼50% of the total gz signal of heme a of ΔSurf1p CcO. In the normal heme a3-CuB center of oxidized CcO, the unpaired electrons of heme a3 and CuB are spin-coupled such that both centers are EPR-silent (35Tweedle M.F. Wilson L.J. Garcia-Iniguez L. Babcock G.T. Palmer G. J. Biol. Chem. 1978; 253: 8065-8071Abstract Full Text PDF PubMed Google Scholar). In the absence of CuB, however, high-spin heme a3 yields a signal at g ≃ 6 (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 36Hunter D.J. Moody A.J. Rich P.R. Ingledew W.J. FEBS Lett. 1997; 412: 43-47Crossref PubMed Scopus (10) Google Scholar). An elevated g ∼ 6 signal in the spectrum of ΔSurf1p CcO (Fig. 2) indicates the presence of some high-spin heme a3. Using our previously published spectra of the ΔCuB oxidase as a standard for 100% high-spin heme a3 (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), the slightly elevated g ≃ 6 signal in the spectrum of ΔSurf1p CcO indicates the presence of ∼15% high-spin heme a3. This strongly suggests that about 15% of ΔSurf1p CcO contains heme a3 but lacks CuB. The relationship of this finding to the spectroscopic results in Table I is discussed immediately below. The CcO population that assembles in the absence of Surf1p contains three distinct configurations of the heme a3-CuB active site. First, ∼50% of ΔSurf1p CcO lacks heme a3, as evidenced by the Soret/α value, measurements of heme A content and the amount of CO binding. Second, 35–40% of the CcO population is active and contains a normal heme a3-CuB active site, as shown by activity measurements and the extent of CN– binding to the reduced enzyme. Third, optical and EPR spectroscopy reveal a smaller population, corresponding to the remaining 10–15%, in which the active site contains heme a3 in the normal high-spin configuration but lacks CuB. The evidence for this is as follows. Carbon monoxide binding, which requires reduced heme a3 but not CuB (31Mitchell R. Moody A.J. Rich P.R. Biochemistry. 1995; 34: 7576-7585Crossref PubMed Scopus (26) Google Scholar), shows the presence of 50% heme a3 in ΔSurf1p CcO. However, CN– binding to the reduced enzyme, which requires both heme a3 and CuB (31Mitchell R. Moody A.J. Rich P.R. Biochemistry. 1995; 34: 7576-7585Crossref PubMed Scopus (26) Google Scholar, 32Mitchell R. Rich P.R. Biochim. Biophys. Acta. 1994; 1186: 19-26Crossref PubMed Scopus (167) Google Scholar), is only 35–40% that of normal CcO. Together this indicates that 10–15% of the ΔSurf1p CcO contains heme a3, but not CuB. This conclusion can be derived independently from the EPR spectra. The presence of ∼15% high-spin heme in ΔSurf1p CcO (Fig. 2) is best explained by the presence of a subpopulation that contains high-spin heme a3 but lacks CuB (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). In addition, preliminary metal analysis data are consistent with the absence of CuB in the 50% of ΔSurf1p CcO that lacks heme a3. It should be noted, however, that metal content measurements are complicated by the fact that a 50–60% loss of CuB in the presence of a normal CuA center leads to only a 17–20% loss of the total amount of copper in each CcO monomer. Assembly in the absence of Surf1p does not affect metallation of the heme a and CuA sites. Retention of the α band ∼605 nm in the optical spectrum of reduced ΔSurf1p CcO, along with the fact that all of the loss of heme A is accounted for by the loss of heme a3, indicates that each ΔSurf1p CcO monomer contains a heme a center. Two configurations of heme a were revealed in the EPR spectrum of ΔSurf1p CcO, however. Each accounted for half of the total amount of heme a. Structural disruption of the heme a3-CuB active site has previously been shown to lead to broadened gz signals in the g ≃ 3 region (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 37Mitchell D.M. Aasa R. Δdelroth P. Brzezinski P. Gennis R.B. Malmström B.G. FEBS Lett. 1995; 374: 371-374Crossref PubMed Scopus (27) Google Scholar, 38Bratton M.R. Hiser L. Antholine W.E. Hoganson C. Hosler J.P. Biochemistry. 2000; 39: 12989-12995Crossref PubMed Scopus (30) Google Scholar). Both hemes a and a3 are bound to transmembrane helix 10 in subunit I (4Svensson-Ek M. Abramson J. Larsson G. Törnroth S. Brzezinski P. Iwata S. J. Mol. Biol. 2002; 321: 329-339Crossref PubMed Scopus (478) Google Scholar, 39Hosler J.P. Ferguson-Miller S. Calhoun M.W. Thomas J.W. Hill J. Lemieux L. Ma J. Georgiou C. Fetter J. Shapleigh J. Tecklenburg M. Babcock G.T. Gennis R.B. J. Bioenerg. Biomembr. 1993; 25: 121-136Crossref PubMed Scopus (242) Google Scholar), and EPR spectroscopy is highly sensitive to structural changes in the environment of heme a, such as alterations in hydrogen bonds to the axial histidine ligands (26Hosler J.P. Fetter J. Tecklenburg M.M. Espe M. Lerma C. Ferguson-Miller S. J. Biol. Chem. 1992; 267: 24264-24272Abstract Full Text PDF PubMed Google Scholar). Therefore, we assign the g = 2.81 signal to the 50% of ΔSurf1p CcO that contains heme a3 and the broad g = 2.89 signal to the 50% of ΔSurf1p CcO that lacks heme a3. The EPR spectra also show that ΔSurf1p CcO contains a normal amount of properly assembled CuA in subunit II. This is the first demonstration that Surf1p plays a role in the assembly of the metal centers of CcO. The assembly role of Surf1p is clearly limited to the heme a3-CuB active site. Formally, Surf1p could be involved in heme a3 insertion, or it could be involved in the insertion of CuB if the assembly of CuB is a prerequisite for the assembly of heme a3. However, we have previously demonstrated that heme a3 insertion does not require the prior assembly of CuB (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). Thus, the sum of our previous and present results argues for a role for Surf1p in the assembly of the heme a3 center. The putative absence of CuB when heme a3 is also absent suggests that stable binding of this copper requires the presence of the active site heme. As yet, it cannot be discerned whether Surf1p plays a direct (e.g. as a heme chaperone) or an indirect role in the insertion of heme a3. Surf1p may interact with the proposed Cox10p-Cox15p complex that synthesizes and possibly inserts heme A (7Brown B.M. Wang Z. Brown K.R. Cricco J.A. Hegg E.L. Biochemistry. 2004; 43: 13541-13548Crossref PubMed Scopus (33) Google Scholar). Alternatively, Surf1p could interact with subunits I or II to facilitate the assembly of the heme a3-CuB center. In this light, an apparent association of Surf1p and subunit II has been reported for cultured human cells (12Nijtmans L.G. Artal Sanz M. Bucko M. Farhoud M.H. Feenstra M. Hakkaart G.A. Zeviani M. Grivell L.A. FEBS Lett. 2001; 498: 46-51Crossref PubMed Scopus (59) Google Scholar). The finding that some normally assembled CcO accumulates in the absence of Surf1p in R. sphaeroides is consistent with reports of the assembly of mitochondrial CcO in the absence of Surf1p. A constitutive knock-out of both surf1 alleles in mice is highly detrimental but not 100% lethal, indicating the production of some active CcO (40Agostino A. Invernizzi F. Tiveron C. Fagiolari G. Prelle A. Lamantea E. Giavazzi A. Battaglia G. Tatangelo L. Tiranti V. Zeviani M. Hum. Mol. Genet. 2003; 12: 399-413Crossref PubMed Scopus (71) Google Scholar). Mutations that lead to the absence of Surf1p in humans, causing Leigh syndrome, severely decrease the amount of CcO, but some CcO is correctly assembled and active (14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar). Work with the R. sphaeroides system has previously demonstrated that assembly of the heme a3-CuB active site requires the association of subunit II with subunit I (38Bratton M.R. Hiser L. Antholine W.E. Hoganson C. Hosler J.P. Biochemistry. 2000; 39: 12989-12995Crossref PubMed Scopus (30) Google Scholar). The various proposals for Surf1p function in mitochondria center on the maturation of subunit I and the subsequent association of subunit I with subunit II (11Williams S.L. Valnot I. Rustin P. Taanman J.W. J. Biol. Chem. 2004; 279: 7462-7469Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar, 12Nijtmans L.G. Artal Sanz M. Bucko M. Farhoud M.H. Feenstra M. Hakkaart G.A. Zeviani M. Grivell L.A. FEBS Lett. 2001; 498: 46-51Crossref PubMed Scopus (59) Google Scholar, 13Barrientos A. Korr D. Tzagoloff A. EMBO J. 2002; 21: 43-52Crossref PubMed Scopus (141) Google Scholar, 14Pecina P. Houstkova H. Hansikova H. Zeman J. Houstek J. Physiol. Res. 2004; 53: S213-S223PubMed Google Scholar, 16Perez-Martinez X. Broadley S.A. Fox T.D. EMBO J. 2003; 22: 5951-5961Crossref PubMed Scopus (153) Google Scholar, 17Barrientos A. Zambrano A. Tzagoloff A. EMBO J. 2004; 23: 3472-3482Crossref PubMed Scopus (165) Google Scholar). Our finding that Surf1p is involved in the assembly of heme a3 suggests that the assembly of the heme a3-CuB active site is the maturation step that makes it possible for subunits I and II to associate in mitochondria. The association of these subunits in R. sphaeroides does not absolutely require the assembly of the heme a3-CuB active site, as evidenced by this and two previous studies (9Hiser L. Di Valentin M. Hamer A.G. Hosler J.P. J. Biol. Chem. 2000; 275: 619-623Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 41Hiser L. Hosler J.P. J. Biol. Chem. 2001; 276: 45403-45407Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar). However, the level of expression of the R. sphaeroides oxidase is lower in the absence of full assembly of the heme a3-CuB active site, suggesting that subunit association is less stable. The prerequisites for subunit association are likely to be more stringent in mitochondria than in R. sphaeroides because the core subunits do not associate independently but as members of subassemblies that include nuclear encoded accessory subunits (11Williams S.L. Valnot I. Rustin P. Taanman J.W. J. Biol. Chem. 2004; 279: 7462-7469Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar, 42Nijtmans L.G. Taanman J.W. Muijsers A.O. Speijer D. Van den Bogert C. Eur. J. Biochem. 1998; 254: 389-394Crossref PubMed Scopus (203) Google Scholar). We thank Drs. Victor Davidson and Dennis Winge for valuable discussions and Dr. Paul Cobine for metal analysis assays.

Referência(s)