Two Nonidentical Forms of Subunit V Are Functional in Yeast

Proc. Nati. Acad. Sci. USA Vol. 82, pp. 2235-2239, April 1985 Biochemistry

Two nonidentical forms of subunit V are functional in yeast cytochrome c oxidase (Saccharomyces cerevisiae/nuclear genes/mitochondria/DNA sequence/gene disruption) MICHAEL G. CUMSKY, CHRISTINE Ko, CYNTHIA E. TRUEBLOOD, AND ROBERT 0. POYTON Department of Molecular, Cellular and Developmental Biology, University of Colorado at Boulder, Campus Box 347, Boulder, CO 80309 Communicated by David M. Prescott, November 19, 1984

ABSTRACT In Saccharomyces cerevisiae, the inner mito- ing the gene on a multicopy plasmid, (ii) the ability of the chondrial membrane protein cytochrome c oxidase is com- plasmid to select subunit V-specific mRNA, and (iii) the ap- posed of nine polypeptide subunits. Six of these subunits (IV, pearance of a new cross-reacting subunit V polypeptide in V, VI, VII, VIIa, VIII) are encoded by the nuclear genome, transformant mitochondria (5, 6). and the remaining three (I, II, III) are encoded by mitochon- In this paper, we report that the latter gene, now called drial DNA. We report here the existence of two nonidentical COXSb, encodes a subunit V polypeptide that is similar to, subunit V polypeptides, which are encoded by separate genes but not identical with, the subunit V polypeptide recently within the yeast genome. One gene, COX5a, encodes the poly- sequenced in this laboratory (7). We also report the molecu- peptide Va, normally found in preparations of holocytochrome lar cloning of another gene for subunit V, COX5a. This gene c oxidase. The other gene, COX5b, encodes the polypeptide encodes a polypeptide identical in sequence to the subunit V Vb, which cross-reacts with anti-subunit Va antiserum and re- polypeptide found in the holoenzyme. We show that both stores respiratory competency and cytochrome oxidase activity genes are expressed in yeast and their products can function in transformants of coxSa structural gene mutants. This poly- as bona fide subunits in the holoenzyme. peptide also copurifies with the holoenzyme prepared from these transformants. We have found that COX5b is expressed in vegetatively growing yeast cells, and that the Vb polypeptide MATERIALS AND METHODS can be detected in mitochondria from strain JM28, a cox5a Plasmids, Strains, and Growth Media. The following Sac- mutant. This mutant has 15%-20% residual cytochrome oxi- charomyces cerevisiae strains were used: D273-10B (mata, dase activity, and it respires at 10%-15% the wild-type rate. ATCC 24657); JM28 (mata, leu2-3 leu2-112 his4-580 ura3-52 By disrupting the COX5b gene in this strain, we show that this trpl-289 ade2 coxfa-J) was constructed as described (5). residual activity is directly attributable to the presence of a JM43 (mata leu2-3 leu2-112 his4-580 ura3-52 trpl-289) was chromosomal copy of the COX5b gene. Taken together, these constructed by crossing D273-1OB with AB35-13D (5), then results suggest that Va or Vb can function as cytochrome oxi- backcrossing to D273-10B five times. JM28-511, and JM28- dase subunits in yeast and that Vb may be used under some 552 are transformants of JM28, as described in Results. specific, as yet undefined, physiological conditions. JM28-GD5b2 and JM28-GD5b3 are strains carrying disrup- tions of COXSb also as described in Results. Parental and The biogenesis of a respiratory competent mitochondrion re- transformed yeast strains were grown at 28°C on either YPD, sults from the joint expression of two genomes-mitochon- YPGE, SD, or SCD medium (8), as described in the text. drial and nuclear (1). Over the last decade, studies from a Amino acids and nucleotides were added or deleted as re- number of laboratories have resulted in a fairly thorough un- quired. Solid media contained 2% bacto agar. derstanding of the organization and expression of the mito- Escherichia coli strains used for propagating plasmids chondrial genome (for review, see ref. 2). However, only re- were RR1 (F- pro leu thi lac Y St' r-m- endoI-) and HB101 cently have similar questions begun to be addressed for the (F- pro leu thi lacY Str' r-m- endoI- recA-). A RecA- de- nuclear genome. The hetero-oligomeric membrane protein rivative of strain JM103 (9) was used as a host for all phage cytochrome c oxidase provides a useful model for investigat- cloning. E. coli strains were grown at 37°C in LB or YT me- ing nuclear-mitochondrial interactions in yeast. It is com- dium (10). When grown selectively, transformed strains posed of nine nonidentical polypeptide subunits, the three were grown in either of the above media supplemented with largest of which are mitochondrial in origin (I, II, III), while 100 ,g of ampicillin per ml. the remaining six are nuclear encoded (IV, V, VI, VII, VIIa, The shuttle plasmid YEp13 (11) was used in this study. VIII) (3). The plasmids YEp13-511 and YEp13-552 were isolated from To facilitate studies on the nuclear-encoded subunits of the yeast library constructed by Nasmyth and Tatchell (12). cytochrome oxidase, we have begun to isolate the genes en- Nucleic Acid Hybridizations. High stringency nucleic acid coding these polypeptides. Previously, we reported the mo- hybridizations were done at 680C in 6x NaCl/Cit (lx NaCl/ lecular cloning of two of these genes. The structural gene for Cit is 0.15 M NaCl/0.015 M Na citrate, pH 7.0)/3x Den- subunit VI (COX6) was isolated by using two synthetic oligo- hardt's solution (13)/0.5% NaDodSO4/100 ,ug of sonicated nucleotide probes complementary to the DNA sequence, as salmon testes DNA per ml. Two to three washes were done predicted from its amino acid sequence (4). In addition, we at 680C in 2x NaCI/Cit/0.5% NaDodSO4, followed by an ad- have used a subunit V structural gene mutation to clone a ditional 2-3 washes in lx NaClI/Cit/0.5% NaDodSO4. Re- gene for subunit V, COXS (5). Evidence that this gene was duced stringency hybridizations were at 450C in the same COXS came from (i) the restoration of respiratory competen- hybridization mix; washes were done the same way as de- cy and cytochrome oxidase activity in mutant strains carry- scribed above, except at 450C. All DNA transfers were to nitrocellulose filters, while RNA was transferred to Gene- The publication costs of this article were defrayed in part by page charge Screen (New England Nuclear) after separation on 6% form- payment. This article must therefore be hereby marked "advertisement" aldehyde/1.5% agarose gels (14). in accordance with 18 U.S.C. §1734 solely to indicate this fact. DNA Sequence Analysis. Restriction fragments to be se-

2235 Downloaded by guest on September 25, 2021 2236 Biochemistry: Cumsky et al. Proc. NatL Acad Sci. USA 82 (1985) quenced were subcloned into the M13 phage vectors mp8 00 and mp9 (15) and were sequenced by using the dideoxy N.I method (16). Sequencing reactions were analyzed on 6% /) buffer gradient gels (17) as well as with multiple loadings on I, 6% sequencing gels. I<4R5/ B 1 / t>c, N. Purification of Cytochrome Oxidase. Holocytochrome c oxidase was prepared from various yeast strains, using a scaled-down version of our standard procedure (3), with the following modifications. Mitochondria were prepared from yeast spheroplasts and converted to submitochondrial parti- cles by sonication. Ten milligrams of submitochondrial parti- cles were then extracted with 0.3 vol of 20% Fisher cholic acid, followed by overnight precipitation with 50 mg of am- monium sulfate per ml at 40C. After centrifugation, the su- Vb_ Vb pernatant was subjected to octyl Sepharose chromatography Va- __-- -VVa (0.2-ml bed volume) as described (3), except that the cyto- Vm vm chrome oxidase was eluted in 1 ml of buffer containing 5%, rather than 3%, Triton X-100. The exchange centrifugation and desalting steps were accomplished by ultracentrifuga- tion of the enzyme through the appropriate buffers (3) in a microfuge tube. The tube was supported by an O-ring on the rim of a centrifuge tube that had been filled with the same buffer. A typical preparation yields 10-20 ,ug of cytochrome FIG. 1. Transformants regain cytochrome oxidase subunit V oxidase. polypeptides. Immunoblots of mitochondria (50 ,g) from the indi- Miscellaneous Methods. NaDodSO4/polyacrylamide gel cated strains were carried out as described (5), using anti-subunit V antisera. (A) Strains are D273-1OB and JM43, wild type; JM28, a electrophoresis; immunoblotting; yeast and E. coli transfor- cox5a mutant; and JM28-511, a JM28 transformant carrying the mations; restriction endonuclease analysis; cytochrome oxi- COXSb plasmid YEpl3-511. Positions of subunits Va, Vb, and Vm, dase assays; and the preparation of DNA, RNA, and mito- the mutant form of Va, are indicated. CO, 0.5 ,ug of purified holocy- chondria were as described (5). Poly(A)+ RNA was prepared tochrome c oxidase standard. (B) Same as A. Additional strains are from total yeast RNA by the method of Aviv and Leder (18). JM28-552, the mutant JM28 transformed with the COX5a plasmid YEp13-552; JM43-511, the wild-type strain JM43 transformed with RESULTS YEp13-511. S. cerevisiae Contains Two Nonidentical Genes for Cyto- the vector YEp13 (10), we obtained seven additional trans- chrome Oxidase Subunit V. Previously, we reported the use formants in which the ability to grow on media containing of a cytochrome oxidase subunit V structural gene mutation glycerol and on media lacking leucine cosegregated. To de- to clone the subunit V gene by complementation in yeast (5). termine which subunit V gene they contained, plasmid DNA This mutant, JM28, contains an aberrant form of subunit V, from the seven transformants was screened by Southern designated Vm. In these initial experiments, a hybrid plas- blotting against a COXSb probe. In addition, mitochondria mid, YEpl3-511, isolated from one transformant, JM28-511, was shown to contain a subunit V gene (COXS) on the basis COX 5a of its ability to restore respiratory competency and cyto- V _ 0- S .0 chrome oxidase activity to mutants harboring it, and its abili- CL & O mr(Lw CD ca0 0.m .Ia.4co.ILt tD x I X ty to select subunit V-specific mRNA. Curiously, the plas- ~~~~~~~~~..>1~ ( ( ) ")>1rII I ( mid-encoded subunit V present in JM28-511, had an appar- ent molecular weight on NaDodSO4/polyacrylamide gels that was slightly higher than the polypeptide normally observed in yeast mitochondria (Fig. 1A). Since eight different respiratory-proficient transformants obtained in our initial kbp' experiments contained this slower-migrating form of subunit 1 2 3 4 5 6 V, we speculated that aberrant processing-possibly due to overexpression of the plasmid-encoded gene-might be re- sponsible for this new form. COX5b We have now localized the COX5 gene on the YEp13-511 0 x insert and subjected the gene to DNA sequence analysis. I " W -0 r t4 ,_ c This analysis indicates that YEp13-511 contains a gene E 0 .2 * 8 V) S o WoI -E c x w CD w whose protein product, designated Vb, has an amino acid m 4 m_ I CD sequence similar to, but not identical with, that of the sub- AIrr unit V polypeptide previously sequenced in this laboratory (see Fig. 3). This surprising result led us to look for an additional gene for the subunit V polypeptide, designated Va, which is normally found in preparations ofholocytochrome c 1 3 4 oxidase. kbp 12 3 4 Molecular Cloning of the COX5a Gene. Because hybridizations under stringent conditions indicated that the COXS FIG. 2. Restriction maps and sequencing strategy for the COXSa gene on YEp13-511 (called COXSb) did not cross-hybridize and COXSb genes. Restriction maps for the respective inserts with any other sequences present in the yeast genome, we (COXSa, YEp13-552; COXSb, YEp13-511) within the vector YEp13 again transformed JM28, a leu- strain carrying the original are shown. Solid bar represents coding sequence for each gene. Ar- subunit V mutation in an attempt to clone the other subunit rows indicate individual clones in phage vectors mp8 and mp9 that V gene, called COX5a. Using a recombinant yeast library in were sequenced. Size of each insert is shown in kilobase pairs (kbp). Downloaded by guest on September 25, 2021 Biochemistry: Cumsky et aL Proc. Natl. Acad. Sci. USA 82 (1985) 2237

from each of the transformants was subjected to immunoblot Table 1. Either COX5a or COX5b can restore cytochrome c analysis using anti-subunit V antiserum. The results of these oxidase activity in a COX5a structural gene mutant experiments indicated that six of the transformants harbored Strain Specific activity* % wild typet plasmids that encoded the COXSb gene. The seventh, JM28- 552, contained a plasmid, YEp13-552, whose restriction map JM28 0.85 19 (Fig. 2) differed from YEp13-511 and that encoded a poly- JM28-511 3.09 69 peptide that migrated identically to that of the subunit V nor- JM28-552 4.7 113 mally found in yeast mitochondria (i.e., see subunit Va in *Calculated as K, the first-order velocity constant (,umol of cyto- Fig. 1B). chrome c oxidized per min per mg of mitochondrial protein). To show unequivocally that YEp13-552 contained the tValues for JM28 and JM28-511 are from ref. 5 and were calculated COXSa structural gene, we sequenced DNA from a region of using strain D273-1OB (K = 4.5) as wild type. Data for JM28-552 the plasmid that cross-hybridized with a COXSb probe under were calculated using strain JM43 (K = 4.2) as wild type. Indi- conditions of reduced stringency. As indicated in Fig. 3, the vidual strains are described in the text. DNA sequence from this region predicts an amino acid sequence in perfect agreement with our previously determined sidual cytochrome c oxidase activity observed in the COX5a subunit V amino acid sequence, confirming that we had, in structural gene mutant JM28 (Table 1) could be attributable fact, cloned the COX5a structural gene. Fig. 3 presents par- to Vb. Although it is conceivable that the Vb polypeptide tial DNA sequences for both COX5a and COXSb from that acts indirectly to restore cytochrome c oxidase function, region of the gene for which we have amino acid sequence several observations suggest that it can function as a bona data. In this region, the NH2-terminal 68 amino acids, the fide subunit of the holoenzyme. two subunit V genes share only 62% amino acid homology. First, COX5b is expressed in yeast and the Vb polypeptide COX5a or COX5b Can Function in Holocytochrome c Oxi- can be detected in mitochondria prepared from the coxfa dase. The original mutation used to clone the subunit V mutant JM28. Fig. 4A shows an RNA blot of total RNA from genes resides in the COX5a gene, because it is this polypep- the transformant JM28-511, and also of poly(A)+ RNA from tide, and not Vb, that is altered as a consequence of the mu- a wild-type yeast strain, D273-1OB. The blot was probed tation (5, 6). Because either the cloned COX5a gene itself or with a COX5b-encoding sequence-specific restriction frag- the nonidentical COX5b gene can complement the mutation ment, and as indicated in the figure, the three COX5b tran- and restore cytochrome oxidase activity to the holoenzyme scripts are visible in both lanes. While we expect to observe (Table 1), we wondered whether the COX5b polypeptide transcripts for COX5b in the transformant, since the plas- could, under some conditions, functions as a bona fide cyto- mid-encoded gene is being overexpressed, the presence of chrome oxidase subunit. We also wondered whether the re- the same transcripts in poly(A)+ RNA from a wild-type strain indicates that COX5b is expressed in vegetatively

1 31 GTT CM ACA AAG GCC CTT TCG AAG GCA ACA TTG ACA GAT CTG CCC GAA AGA A %- valZgnthr Iysaa Zeuer Iysaa thrZeu thr aspleeu pro glu arg B &la gln thr his ala leu aer asaalaa&la val mat asp leu gin ser arg P ala gin thr his ala leu ser asn ata ala vaZ met asp Zeu gin serarg 'b e~' GCT CAA ACA CAT GCT CTT TCC MC OCT GCT GTA ATG GAT CTG CM TCA CGA IrVl A~

61 91 TGG GAA AAT ATG CCA AAC TTA GAA CAG AM GAG ATT OCA GAT AAT TTG ACA trp g1-u asn met pro asn Zeu g~u gtn Iys giu i7e aZa asp asn e thr

trp glu asa met pro ser thr glu gin gin asp ile val (-) lys leuler

trp gZu asn met pro ser thr glu gZn gin asp ile val ser lye L s'er TGG GAG MC ATG CCC TCC ACT GAG CAG CAG GAT ATT GTC AGT AAG TTG AGT -Vb

121 151 -Vm GAA CGT CM MG CTT CCA TGG MA ACT CTC MT MC GAG GM ATC MA GCA giu arg gin Iys leu pro trp lys thr Zeu asn asn glu glu ile lye ala

glu arg gln Iys leu pro trp la glnaleu thr glu pro glu lys gln ala glu arg gin Iys leu pro trp ala gin leu thr glu proo Luj Iys gin ala GAA CGT CM MA TTA CCA TGG OCA CAG CTT ACT GAG CCT GAA AAG CAA OCT

181 GCT TGG TAC ATA TCC TAC GGC GAG TGG GGA CCT AGA AGA CCT GTA CAC GGA ala trp tyr ile aser tyr gly glu t'p gly pro arg arg pro val hie gly

val trp tyr ile ser tyr gly glu trp gly pro arg arg pro vai leu aan

vai trp tyr ile ser tyr gly glu trp giy pro arg arg pro val leu asn GTG TGG TAC TCT TAC GGA GAA TGG GGC CCA AGA AGA CCT GTA TTG MT FIG. 4. The COX5b gene is expressed in yeast. (A) RNA blot of 5 ATE ,ug of total RNA from the transformant JM28-511, or 20 ,ug of poly- FIG. 3. Partial DNA sequences of the COX5a and COX5b genes. (A)' RNA from the wild-type yeast strain D273-1OB. The probe Top line, DNA and predicted amino acid sequence of COXSb gene; used was a nick-translated 1.2-kilobase Bgl II restriction fragment middle line, determined amino acid sequence of subunit V from ref. containing nearly all the COXSb coding sequence (see Fig. 2). Ar- 7; bottom line, DNA and predicted amino acid sequence of COXSa rows denote the three COX5b transcripts. (B) Immunoblot of 50 jig gene. The first amino acid for the determined peptide sequence and of mitochondrial protein from the wild-type strain JM43, or the for the COX5a gene corresponds to the NH2-terminal amino acid of cox5a mutant JM28. Immunoreactive subunits were identified using the mature polypeptide. COX5b was aligned on the basis of amino anti-subunit V-specific antiserum and 125"-labeled protein A. Vb and acid sequence homology. Identical amino acids are boxed. Com- Va denote positions of their respective subunits. Vm corresponds to plete DNA and flanking sequence data for both COXSa and COX5b the mutant subunit Va polypeptide in JM28. The blot was overex- will be published elsewhere. posed to accentuate the subunit V polypeptides. Downloaded by guest on September 25, 2021 2238 Biochemistry: Cumsky et al. Proc. Natl. Acad. Sci. USA 82 (1985)

growing yeast. In Fig. 4B we present an immunoblot of wild- type and of JM28 mitochondria. While no subunit Vb is observed in wild-type mitochondria, the subunit Vb polypep- LEU2 tide is clearly visible in mitochondria prepared from JM28. < 3.0kb > Second, both COXSa and COXSb restore similar levels of Bg/11 Bg/11 oxidase activity in coxfa mutant mitochondria 8om cytochrome YEpI3 when present in yeast on high copy number plasmids. As prepared from strain JM28- shown in Table 1, mitochondria BgI 11 552 (the cox5a mutant JM28 transformed with a wild-type COXSa gene on plasmid YEp13-552) or JM28-511 (JM28 transformed with a wild-type COXSb gene on YEpl3-511), contain 100% and 70% of the wild-type amount of cytochrome oxidase activity, respectively. Moreover, cytochrome oxidase purified from JM28-511 contains the COX5b polypeptide (Fig. 5), indicating that COXSb is fully capable of assembling into the holoenzyme complex. Together, these two results-the restoration of a nearly wild-type level of cytochrome oxidase activity by COXSb and the copurifica- tion of this subunit with the holoenzyme-argue strongly against the possibility that the COXSb gene is able to complement a COXSa defect fortuitously. Finally, we have been able to show that the residual cytochrome oxidase activity present in the cox5a mutant JM28 Bom results from the presence of a good copy of COXSb in this strain. To do this, we used the one-step gene disruption tech- Bg/ 11 6g/ 11 nique of Rothstein (19) to inactivate the chromosomal disruption of the COX5b gene. A fragment strategy used for FIG. 6. Strategy for COXSb gene in JM28. The experimental containing the entire COXSb gene was isolated from YEp13-511 (see this disruption is shown in Fig. 6. After cloning the selecta- Fig. 2) by digestion with BamHI followed by partial digestion with ble yeast LEU2 gene into a BgI II restriction site within the Bgl lI. The fragment was then subcloned into the vector YIp5 (24) to COXSb coding sequence, the disrupted gene on a linear yield YIp5-5B. This plasmid was then opened at the unique Bgl II BamHI/Cla I fragment was transformed into JM28. LEU+ site, and a 3.0-kilobase (kb) Bgl II fragment containing the LEU2 transformants were screened by Southern hybridization and gene (from YEp13) was inserted, leaving a disrupted COXSb gene on immunoblot analysis to confirm that the proper integration the plasmid pCET-GD5B. pCET-GD5B was linearized with BamHI had occurred (data not shown). Two transformants, JM28- and Cla I prior to transformation. GD5b2 and JM28-GDSb3, were then analyzed for respiratory proficiency (and, hence, cytochrome oxidase activity), by DISCUSSION re- determining the rate of cyanide-sensitive respiration. The In this report, we have presented evidence that the yeast S. sults of these experiments are presented in Table 2. While cerevisiae contains two similar, but clearly different, genes the two JM28 respires at :'12% the wild-type (JM43) rate, for subunit V of cytochrome c oxidase. We have also shown no strains carrying the disrupted COXSb gene have detect- that the product of either gene is capable of functioning as a able cyanide sensitive respiration. Thus, when both the subunit of the holoenzyme complex. These findings are in- COX5a and COX5b genes are rendered inactive, yeast cells triguing for a number of reasons. First, they indicate that the have no detectable respiration. cytochrome c oxidase holoenzyme can tolerate the assembly of isologous forms of at least one subunit and still function catalytically. Second, they indicate that even proteins ex- 0b pressed at low levels in yeast can exist in multigene families. 40 /I While multigene families exist in Saccharomyces [e.g., gly- colytic pathway isozymes (21-23), ribosomal protein genes (20)], they usually encode polypeptides that are efficiently expressed and are needed in large amounts. 1S- t~~~~ These results also raise a number of questions regarding Table 2. Disruption of COXSb leads to total loss of respiration in IV a COXSa structural gene mutant Vb- * Cyanide-sensitive Strain respiration* % wild type

9- JM43 34 100 JM28 4 12 JM28-GD5b2 NDt <2 JM28-GD5b3 NDt <2 FIG. 5. Subunit Vb copurifies with cytochrome oxidase holoen- Respiration was measured as 02 consumption at 30'C, on 3 ml of a zyme in a COXSb transformant. Cytochrome oxidase was prepared an exponentially growing culture in SCD medium, using Yellow from the COXSb transformant JM28-511, using the small-scale pro- Springs Instruments Model 53 Oxygen Monitor. Values for each cedure described in Materials and Methods. Approximately 1 ,ug of strain represent the average of 3-5 independent measurements. this preparation, as well as 20 ,ug of total mitochondria from the Individual strains are described in the text. wild-type strain D273-10B were analyzed by immunoblotting. Anti- *Cyanide-sensitive respiration is presented as pmol of 02 per min serum raised against the holoenzyme was used to detect the sub- per ,g of dry weight. was units, whose positions are indicated. Mutant subunit V polypeptide, tND, not detectable. Standard deviation for this experiment Vm, is not seen because it comigrates with subunit VI (5). used to estimate level of detectability. Downloaded by guest on September 25, 2021 Biochemistry: Cumsky et aL Proc. Natl. Acad. Sci. USA 82 (1985) 2239

the biological role of COXSb. Although any hypothesis re- by a National Institutes of Health postdoctoral fellowship. C.E.T. garding the function of this gene is speculative at this point, was supported in part by a University of Colorado Graduate School it seems likely that there may be some specific growth or Fellowship. R.O.P. was an Established Investigator of the American physiological condition in which COX5b functions. This Heart Association. would necessarily predict the existence of a specific activa- 1. Poyton, R. O., Bellus, G. & Kerner, A. L. (1982) in Mem- tor of the COX5b gene, because, as is clear from the results branes and Transport, ed. Martonosi, A. (Plenum, New York), presented here, COX5b is not normally expressed at high Vol. 1, pp. 237-247. enough levels to make a cox5a mutant (JM28) respiratory 2. Dujon, B. (1981) in The Molecular Biology of the Yeast Sac- proficient. Indeed, COX5b must be present on a high copy charomyces cerevisiae: Life Cycle and Inheritance, eds. number plasmid (e.g., YEp13 or YEp24) (24) in order to Strathern, J. N., Jones, E. W. & Broach, J. R. (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), pp. 505-635. complement the COX5a defect. Mutants carrying the 3. Power, S. D., Lochrie, M. A., Sevarino, K. A., Patterson, COXSb gene on a low copy number centromere plasmid T. E. & Poyton, R. 0. (1984) J. Biol. Chem. 259, 6564-6570. grow poorly on media containing glycerol as the sole carbon 4. Wright, R. M., Ko, C., Cumsky, M. G. & Poyton, R. 0. source (unpublished observation). However, if a strain (1984) J. Biol. Chem. 259, 15401-15407. (JM43) carrying a wild-type COXSa gene is transformed with 5. Cumsky, M. G., McEwen, J. E., Ko, C. & Poyton, R. 0. a high copy number COXSb plasmid (YEpl3-511), both poly- (1983) J. Biol. Chem. 258, 13418-13421. peptide subunits are seen in mitochondria from the transfor- 6. McEwen, J. E., Cumsky, M. G., Ko, C., Power, S. D. & Poy- mant (Fig. 1B, JM43-511). This indicates that, if present at ton, R. 0. (1984) J. Cell. Biochem. 24, 229-242. sufficient levels, COX5b can be imported into the mitochon- 7. Power, S. D., Lochrie, M. A. & Poyton, R. 0. (1984) J. Biol. Chem. 259, 6575-6578. drion and, as the data in Table 1 suggest, function nearly as 8. Sherman, F., Fink, G. & Lawrence, C. W. (1979) Methods in well as COX5a. It is therefore the low level of expression of Yeast Genetics (Cold Spring Harbor Laboratory, Cold Spring COX5b, or possibly preferential import and/or assembly of Harbor, NY). subunit Va, that normally excludes subunit Vb from the ho- 9. Messing, J., Crea, R. & Seeburg, P. H. (1981) Nucleic Acids loenzyme. The hypothesis that COX5b has a specific role in Res. 9, 309-321. cytochrome oxidase biogenesis is further supported by our 10. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold recent finding that genes for both COX5a and COX5b are Spring Harbor Laboratory, Cold Spring Harbor, NY). present in many different Saccharomyces species (unpub- 11. Broach, J. R., Strathern, J. N. & Hicks, J. B. (1979) Gene 8, lished data). 121-133. 12. Nasmyth, K. A. & Tatchell, K. (1980) Cell 19, 753-764. The existence of two genes for a respiratory protein, while 13. Denhardt, D. J. (1966) Biochem. Biophys. Res. Commun. 23, surprising, is not unprecedented. In yeast, there are two 641-646. genes for apocytochrome c, CYCI and CYC7 (25), and the 14. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular product of either gene can function in respiration. These iso- Cloning: A Laboratory Manual (Cold Spring Harbor Labora- zymes share 85% amino acid sequence homology and consti- tory, Cold Spring Harbor, NY). tute 95% and 5% of total cytochrome c complement in yeast, 15. Messing, J. & Vieira, J. (1982) Gene 19, 269-276. respectively. Another, perhaps more analogous, situation is 16. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. found in the tissue-specific genes of cytochrome oxidase in Acad. Sci. USA 74, 5463-5467. higher eukaryotes (26, 27). At least one polypeptide from 17. Biggin, M. D., Gibson, T. J. & Hong, G. F. (1983) Proc. Natl. beef is differ- Acad. Sci. USA 80, 3963-3965. heart cytochrome oxidase, subunit VIa, clearly 18. Aviv, H. & Leder, P. (1972) Proc. Natl. Acad. Sci. USA 69, ent and immunologically unrelated to subunit VIa isolated 1408-1412. from beef liver. Each of the remaining 12 subunits appears to 19. Rothstein, R. J. (1983) Methods Enzymol. 101, 202-211. be the same. Moreover, tissue specificity seems to override 20. Abovich, N. & Rosbash, M. (1984) Mol. Cell. Biol. 4, 1871- species specificity, in that antibodies against subunit VIa 1879. from the same tissue (heart or liver) but from different spe- 21. Holland, M. J., Holland, J. P., Thill, G. P. & Jackson, K. A. cies cross-react, while antibodies against subunit VIa from (1981) J. Biol. Chem. 256, 1385-1395. the two different tissues within the same species do not 22. Holland, J. P., Labieniec, L., Swimmer, C. & Holland, M. J. cross-react (27). (1983) J. Biol. Chem. 258, 5291-5299. Clearly, the specific functions for each of the two cyto- 23. Young, T., Williamson, V., Taguchi, A., Smith, M., Sled- ziewski, A., Russell, D., Osterman, J., Denis, C., Cox, D. & chrome c isozymes, the tissue-specific polypeptides of high- Beier, D. (1982) in Genetic Engineering ofMicroorganisms for er eukaryotic cytochrome oxidase, and the two subunit V Chemicals, eds. Hollaender, A., DeMoss, R. D., Kaplan, S., polypeptides of yeast cytochrome oxidase remain to be de- Konisky, J., Savage, D. & Wolf, R. S. (Plenum, New York), termined. pp. 335-361. 24. Botstein, D., Falco, S. C., Stewart, S. E., Brennan, M., The authors gratefully acknowledge the work of Dr. Scott Power, Scherer, S., Stinchcomb, D. T., Struhl, K. & Davis, R. W. who provided the polypeptide sequences that led to the identifica- (1979) Gene 8, 17-24. tion of COX5b, and who also devised the cytochrome oxidase mi- 25. Montgomery, D. L., Leung, D. W., Smith, M., Shalit, P., cropurification protocol. We also thank Robin Adams and Cameron Faye, G. & Hall, B. D. (1980) Proc. Natl. Acad. Sci. USA 77, Serbu for assisting with that experiment and Dr. Joan McEwen for 541-545. many interesting discussions. This work was supported by Research 26. Kadenbach, B., Hartman, R., Glanville, R. & Buse, G. (1982) Grants GM29838 and GM30228 from the National Institutes of FEBS Lett. 138, 236-238. Health. M.G.C. was supported in part by a postdoctoral fellowship 27. Jarausch, J. & Kadenbach, B. (1982) Hoppe-Seyler's Z. Physi- from the Damon Runyon-Walter Winchell Cancer Fund, and in part ol. Chem. 363, 1133-1140. Downloaded by guest on September 25, 2021