Eur. J. Biochcm. 203, 587-592 (1992) ic’: FEBS 1992

Primary structure and import pathway of the rotenone-insensitive NADH-ubiquinone oxidoreductase of mitochondria from Saccharomyces cerevisiae Simon DE VRIES, Richard VAN WITZENBURG, Leslie A. CRIVELL and Carla A. M. MARRES Section for Molccular Biology and Biotechnology Centre, Department of Molecular Ccll Biology, University of Amsterdam, The Nethcrlands

(Rcceived July S/September 20, 1991) - EJB 91 0875

The gene encoding the yeast mitochondrial rotenone-insensitive internal NADH : ubiquinone-6 oxidoreductase has been sequenced. The DNA sequence indicates the presence of an open reading frame of 1539 bp predicted to encode a protein of 513 amino acid residues (57.2 kDa). The NADH dehydrogenase is synthesized as a precursor protein containing a signal sequence of 26 residues. In vitro import experiments show that the precursor NADH dehydrogenase is cleaved to the mature size by the matrix processing peptidase. Both cleavage and translocation across the mitochondrial membrane(s) are dependent on the membrane potential component of the proton-motive force. Comparison ofthe protein sequence of the yeast NADH dehydrogenase with the data bank indicates that the enzyme from yeast is homologous to the NADH dehydrogenase of Escherichia coli (22.2% identical residues). Both NADH dehydrogenases contain in the central part of the protein a sequence predicted to fold into a Pap structure involved in the binding of NADH or FAD(H2). Various aspects of the protein structure are discussed.

The mitochondria of the yeast Saccharomyces cerevisiae contain at least two different NADH : ubiquinone-6 oxido- reductases, cach bound to the mitochondrial inner membrane, both being rotenone-insensitive and not coupled to site-1 phos- phorylation [I -51. The enzyme facing the intermembrane space (external NADH dehydrogenase), is involved in the oxidation of NADH produced in the cytosol and may thus considered to be the physiological equivalent of the malate/ aspartate shuttle present in higher eukaryotes, but apparently absent in S. cercrisiae [4,6]. The external NADH dehydrogen- ase is also present in plant mitochondria, which, like yeast, apparently lack a functional malate/aspartate shuttle system

The other type of dehydrogenase faces the matrix site (internal NADH dehydrogenase), oxidizes NADH produced by the Krebs cycle and the mitochondrial alcohol dehydrogen- ase and is the physiological counterpart of complex I [4], not present in S. cerevisiue. Neither of these rotenone-insensitive NADH dehydrogenases are present in mammalian mitochon- dria.

[71.

-___ Correspondence lo S. de Vrics, Section for Molecular Biology,

Department of Molecular Ccll Biology, University of Amsterdam, Kruislaan 318, NL-1098 SM Amsterdam, Thc Nethcrlands.

Abbreviation. DecylQ, 2,3-dimethoxy-5-methyl-6-decyl-1,4-ben- zoquinonc.

l k z y n w . NADH : ubiquinone-6 oxidoreductase, NADH dehydro- genase (EC 3.6.5.3).

N o f e . The novel nucleotide sequence data published here have bcen submitted to the EMBL sequencc data bank and are available under accession number X 61590.

We have described the purification of a NADH :ubiqui- none oxidoreductase from yeast mitochondria [3] and have subsequently established its identity as the internal NADH dehydrogenase by analysis of the properties of the mitochon- dria of a null mutant [5].

Rotenone-insensitive NADH : ubiquinone oxidoreduc- tases are not unique to yeast. For example, an enzyme similar to the single subunit, FAD-containing internal NADH de- hydrogenase is present in Esclzerichia coli [8] and possibly in other prokaryotes [9, 101. Likewise, plant mitochondria [I 13 and Neurosporn crassa [12- 141 seem to contain such an in- ternal NADH dehydrogenase. The mitochondrial inner mem- brane of the organisms listed above, except S. cerwisiae, con- tain both a rotenone-insensitive internal N ADH dehydrogen- ase and rotenone-sensitive and/or piericidine-sensitive proton- translocating complex I. The physiological significance of such a dual pathway is not clear [7]. With respect to S. wrrvisiuc, site-I phosphorylation is inducible under some conditions [I 5 , 161 and, since the internal NADH dehydrogenase is present under these conditions [3], the mitochondria of S. ccwvisiur might also contain a dual pathway to oxidize mitochondrial NADH. The nature of the enzyme in S. cercvisiac catalyzing site-I phosphorylation is obscure, except that its activity is not inhibited by rotenone (or piericidin) and that it apparently lacks Fe-S clusters [15].

To characterize the composition and properties of the NADH dehydrogenases in S. cerevisiae, we have first purified the internal NADH dehydrogenase and cloned its gene [3, 51. In this paper, we report the DNA sequence of this enzyme, the import pathway and discuss the properties of the primary structure in relation to those of the homologous NADH de-

588

hydrogenase of E. coli, purified and sequenced previously by Young et al. [8, 17, 181.

MATERIALS AND METHODS

Strains and growth conditions

E. coli strains JMlOl [I91 and HBlOl [2O] were used to propagate recombinant DNA constructs. S. cerevisiue strain YPIO2(a, his3,leu2, lysl, ura, ade, NDH) (wild type) was used to obtain the mutant lacking the intact NADH dehydrogenase gene, YP102(a, his3, lysl, ura, ade, NDH : : leu2) [5]. This mutant strain was transformed with the shuttle plasmids YCP5O or YEplacl95 [21] carrying various DNA inserts encoding the internal NADH dehydrogenase. Yeast cells were grown on complete or minimal medium supplemented with the appropriate amino acids as described previously [22]. Growth on lactate semisynthetic medium was as described in [l]. Solid growth media also contained 2% agar.

Enzyme purification

The NADH dehydrogenase was purified essentially in the same manner as that described previously [3], except that 1 mM phenylmethylsulfonyl fluoride was present in all stages of the purification. Consequently, the purified enzyme contained negligible amounts of the proteolytic breakdown product [3]. The higher-molecular-mass band was cut out after SDS/PAGE as described in [5] and used for determination of the N-terminus.

DNA sequence strategy

The DNA sequence of part of the Pstl - EcoRV fragment (see Fig. 2) was determined, both from plasmid pMV5 and pCS12 [5]. To this end, restriction fragments (Pstl -Sail, Xbul - HindII1, Hind111 - BamH1, BarnHl - EcoRV, several Suu3A and Mspl fragments) were purified, cloned into M13mp18 or Ml3mp19 and subsequently sequenced. In ad- dition, several specific oligonucleotide primers were used to complete the sequence.

In vitvo transcription, translation and import

In vitro transcription, translation and import was performed essentially as described in [23] employing the rabbit reticulosite lysate from Promega. The DNA construct used for transcription was prepared by digestion of the Sull- EcoRV fragment with Ba131 from the Sall restriction site. Digested fragments were cloned into the PEP40 transcription vector from which the Sphl site was removed (treatment with Sphl followed by blunting with T4 DNA polymerase and ligation), since this restriction site contains an ATG sequence which might function as an unwanted start for translation. DNA sequence analysis of various recombinants yielded a clone starting at nucleotide -37 (see Fig. 3), which was further used for transcription. Coupled mitochondria were prepared as in [22].

Miscellaneous

Published procedures were used for the following: manipu- lation of DNA [24]; transformation of yeast [25]; protein determination [26]; DNA sequencing [27]; SDSjPAGE; West- ern blotting [22]. Determination of enzyme activity [28] was

11 1 2 1 3 1 4 1

Fig. 1. Western blot analysis of wild-type, mutant and transformed- mutant strains using a monoclonal antibody directed against the internal NADH dehydrogenase. Spheroplasts were prepared as in [22] and treated for SDSjPAGE as in [5]. In each lane, 100 pg protein was applied. Lane 1, mutant transformed with YEplacl95/NDH; lane 2, mutant transformed with pMV5; lane 3, mutant; lane 4, wild type.

as in [3] except that 2,3-dimethoxy-5-methyl-6-decyl-l,4-ben- zoquinone (decylQ) was used as acceptor instead of ubiqui- none-2. Restriction and other enzymes used in DNA manipu- lation were purchased from Boehringer, Biolabs and Sigma and used as recommended by the manufacturers. Radioactive chemicals were obtained from New England Nuclear and Amersham. The YCP50 yeast-clone bank was obtained from the American Type Culture Collection. The monoclonal anti- bodies used were the same as those described previously [3, 51.

RESULTS AND DISCUSSION

DNA sequence analysis and predicted protein sequence

In a recent paper, we described the cloning of the gene encoding a mitochondrial NADH dehydrogenase [5]. By con- structing a null mutant using the one-step gene-disruption method, this gene was shown to encode the internal NADH dehydrogenase responsible for the oxidation of NADH prod- uced in the mitochondria1 matrix. Here we present the primary structure of the internal NADH dehydrogenase as determined from the DNA sequence.

The clone pMV5 selected from the YCP5O yeast-clone bank using a specific oligonucleotide probe and containing an insert of 12.5 kb was shown to contain the gene encoding the internal NADH dehydrogenase [5]. Reintroduction of pMV5, a low-copy CEN plasmid, into a null mutant in the internal NADH dehydrogenase restores the presence of the enzyme as determined by Western blotting using a monoclonal antibody (Fig. 1). The same is observed after reintroduction of the Pstl - EcoRV fragment (see Fig. 2) in a high-copy shuttle vector (YEplacl95/NDH, Fig. 1) or a low-copy shuttle vector (not shown). Table 1 indicates that the NADH + decylQ ac- tivity, which in the mutant is entirely due to the activity of the external NADH dehydrogenase, increases upon introduction of a plasmid containing the Pstl - EcoRV fragment. In par- ticular, the enzyme activity is increased by more than a factor of ten in the mutant carrying the multicopy plasmid YEplacl95/NDH, consistent with the increased level of the protein detected by Western blotting. Furthermore, growth on acetate which is absent in the null mutant [5], is restored upon introduction of a single-copy or multicopy plasmid carrying the Pstl - EcoRV restriction fragment (cf. Table 1). The Pstl - EcoRV fragment also contains the genetic infor- mation to repress synthesis of the internal NADH dehydro- genase during growth on glucose (data not shown).

Determination of the sequence of (part of) the Pst l - EcoRV fragment indicates the presence of an open reading

589

H3 X P Sal X H3 B H3 R5

I - 300bp

ATG TAG

Fig. 2. Restriction map of pMV5 and genomic DNA of S. cerevisiae. The region between ATG and TAG contains the sequence encoding the internal NADH dehydrogenase. The part of the DNA that was sequenced in both directions is indicated by the bar. The Pstl -EcoRV fragment contains all the genetic information for expression of the NADH dehydrogenase and was cloned in various plasmids (see text). H3, HindIII; X, Xhal; P. PstI; Sal, Sun; B, BurnHI; R5, EcoRV.

Table 1. Enzyme activities and growth properties of wild-type, mutant and transformed-mutant yeast strains. Growth on acetate was monitored on agar plates, 72 h after inoculation. Growth on glucose was similar for all strains. Mutant strains were transformed with the plasmids pMV5 and YEplacl95/NDH (see text). Reduced growth on acetate of the mutant strain carrying YEplacl95/NDH might be duc to the high copy number of the plasmid.

Strain Specific activity Growth of spheroplasts on acetate

pmol NADH . m i n - ' . mg-'

Wild type 3.46 + + + N D H O mutant 0.39 - pMV5 2.24 + + + Y Eplacl95/NDH 14.65 + +

frame of 1539 bp capable of encoding a protein of 513 amino acids (molecular mass, 57.2 kDa; Fig. 3). The sequence of the protein contains 44% apolar residues and lacks cysteine. Determination of the N-terminus of the isolated NADH de- hydrogenase indicates that the protein starts at Ser27, yielding a molecular mass of 54.2 kDa for the mature internal NADH dehydrogenase. The proteolytic breakdown products that are usually present in the purified preparation [3] start at Lys41 and Lys44. The molecular masses determined by SDSjPAGE for both the mature protein and the breakdown products (53 kDa and 51 kDa, respectively) are in good agreement with those calculated from the DNA sequence. The internal NADH dehydrogenase is thus synthesized as a precursor protein with a signal sequence of 26 amino acid residues, cleavage occurring after an arginine residue. The presequence contains no nega- tively charged residues, is relatively rich in serine/threonine and leucine, as observed for other mitochondria1 targeting and import signals [29, 301. Neither the Chou-Fasman [31] nor the Garnier algorithm [32] predict the presequence to fold into an cx-helix. If, however, the presequence folds into an CI-

helical structure then the helix is amphipathic [33]. The DNA sequence shown in Fig. 3 is that derived from

plasmid pMV5 isolated from the YCPSO yeast-clone bank. We noticed that this sequence differs in three nucleotides from that obtained from plasmid pCS12 isolated from the pFLl yeast-clone bank (5, 341. In pCS12, A29 is changed into G29, changing LyslO into ArglO. The other changes found, G87 + A87 and A192 + T192 do not alter the predicted amino acid sequence. Modification of LyslO into ArglO in the prese- quence probably does not interfere with the function of the presequence in the process of import and translocation of the NADH dehydrogenase.

In vitro import of the precursor NADH dehydrogenase To investigate how the NADH dehydrogenase is imported

into the mitochondrion, we performed in vitro import exper-

CCGGTATTTGWCGTCGACAGGGTATATATAAG -121 ATGTATAGGGCATAGTGGGGAAGGAGTTTTAACTACCAATTGAGGAGTTTCAGGTAGGGT -61 GTCAGTTTCATCACATCATCGAATTACACGTTTACCCAAGAAAAGAAACTAAAAACCACT -1

ATGCTATCGAAGAATTTGTATAGTAA~GAGGTTGCTCACCTCGACGAATACGCTAGTC 60 1 M L S K N L Y S N K R L L T S T N T L V

21 R F A S T R S T G V E N S G A G P T S F

41 K T M K V I D P O H S D K P N V L I L G AGATTCGCTTCCACCAGATCCACAGGGGTGGTG~CTCCGGAGCAGGTCCTACATCTTTT 120

~~~~~~~

A A G A C C A T G A A A G T C A T T G A C C C T C A G C A C A G C G A C A A A C C T G G G T 180

TCGGGGTGGGGAGCTATTTCGTTTTTAAAGCACATTGACACCAAGAAGTACAACGTTTCC 240 61 S G W G A I S F L K H I D T K K Y N V S

81 I I S P R S Y F L F T P L L P S A P V G ATCATCTCTCCTAGAAGCTATTTCTTATTTACGCCTTTGTTACCTTCTGCACCAGTTGGG 300

101 T V D E K S I I E P I V N F A L K K K G

121 N V T Y Y E A E A T S I N P D R N T V T ACAGTAGACGAAAAGTCAATTATTGAGCCCATCGTTAATTTTGCTCTCAAGAARAAGGGG 360

AACGTTACCTACTATGAGGCAGAA~CACCTCTATWTCCCGACAGGAATACCGTTACC 420 141

161

181

201

221

241

261

281

I K S L S A V S O L Y O P E N H L G L H ATAAAATCATTATCTGCCGTTAGCCAGCTATACCAACCTCCATCTAGGGCTGCAT Q A E P A E I K Y D Y L I S A V G A E P

CAAGCAGAACCTGCTGAAATTAAGTACGATTATTTAATCAGTGCTGTAGGTGCG~CCT N T E G I P G V T D Y G H F L K E I P N AACACATTTGGTATTCCTGGGGTCACTGATTACGGTCATTTCCTGAAG~TTCCCAAC S L E I R R T F A A N L E K A N L L P K

TCTTTGGAAATAAGAAGAACTTTTGCCGCCAATCTAGAGAAGGCTAACTTATTGCCAAAG G D P E R R R L L S I V V V G G G P T G

G G T G A T C C C G T A C T G T C C A T T G T C G T G G T T G G T G G T G G G C C T A C T G G T V E A A G E L O D Y V H O D L R K F L P GTAGAGGCCGCTGGTGAACTACAG~TTATGTTCACCAGGACCTGAGAAAGTTTCTCCCT A L A E E V Q I H L V E A L P I V L N M GCATTGGCCGAAGAAGTCCAAATTCACTTGGTCGAAGCTTGGTCGAAGCTCTGCCCATCGTTTTGAATATG F E K K L S S Y A Q S H L E N T S I K V

480

540

600

660

720

780

840

900 301 H L R T A V A K V E E K Q L L A K T K H

321 E D G K I T E E T I P Y G T L I W A T G CATCTGAGAACGGCTGTCGCCAAAGTTGAAGAAAAGWTTGTTGGCAAAGACCL'4iCAC 960

GAAGACGGTAAAATAACCGAAAAGAAACTATTCCATACGGTACTTTGATTTGGGCCACGGGT 1020 341

361

381

401

421

441

461

481

N K A R P V I T D L F K K I P E Q N S S A A C A A G G W G A C C G G T A A T C A C T T T T C A A G A A A A C T C G T C C K R G L A V N D F L Q V K G S N N I F A

AAGAGAGGATTGGCAGTGAATGACTTTTTGCAGGTGAAAGGCAGCAAWCATTTTCGCC I G D N A F A G L P P T A Q V A H Q E A

ATTGGTGA~TGCATTTGCTGGGTTGCCACCAACCGCCCAAGTAGCGCACCAAGAtiGCC E Y L A K N F D K M A Q I P N F Q K N L

S S R K D K I D L L F E E N N F K P F K TCTTCAAGAAAGGATAAAATTGATCTCTTGTTCGAGGAGAACAACTTTAAACCTTTCAAA Y N D L G A L A Y L G S E R A I A T I R

TACAACGATTTAGGTGCCTTAGCATACCTGGGATCCGAAAGGGCCATTGCAACCATACGT S G K R T F Y T G G G L M T E Y L W R I TCCGGTAAGAGAACATTTTAC9CCGGTGGTGGCTTAATGACCTTCTACTTATGGAGAAT~ L Y L S M I L S A R S R L K V F F D W I

T T G T A C T T G T C C A T G A T T C T A T C T G C A A G A T C G A W I T T A A T T

GAATATTTGGCCAAGAATTTTGATAAAATGGCTCAAATACCAAATTTCCWGAATCTA

1080

1140

1200

1260

1320

1380

1440

1500 501 K L A E F K R D F F K G L

AAATTAGCATTTTTCAAAAGAGACTTTTTTAAAGGATTATAGATGAAATTAACATGCCCT 1560 TTTCTGGAAAAAGGAAAAAAGGTGGTAGGCACCAGTTTTTTCCTGAGTTTGCATCCTTTT 1620 TTTTCTAAAACCCTCTAAACAAAACCTAACACACACACACACGCACL%A?@AATGCACAT 1680 GATGTTTTATTATTTATATATTCCCACTTTTTTCGAAATGATGCTTGACT~TGCA 1736

Fig. 3. DNA sequence of the gene encoding the internal NADH dehydro- genase and the predicted protein sequence. The underlincd sequences were also determined by directly sequencing the full-sized mature protein (STG . . . )and the proteolytic breakdown product (KTM . . .; cf. [3 ,5]) . The DNA sequence shown here is that derived from plasmid pMV5 isolated from the YCP50 yeast-clone bank (see text).

iments. Fig. 4 indicates that the in-vitro-synthesized NADH dehydrogenase is larger by about 3 kDa than the mature pro- tein isolated by means of native immune precipitation from cells uniformly labeled with [35S]methionine. Incubation of the in-vitro-synthesized protein with a matrix fraction from yeast mitochondria results in processing, albeit partially, of the precursor to the mature size (not shown).

Fig. 4 shows that in mitochondria supplied with a respira- tory substrate (a-glycerol phosphate) and ADP the precursor NADH dehydrogenase is cleaved to its mature size. The ma- ture protein, but not the precursor, is resistant to protease K treatment, indicating that the mature protein has been import- ed into the mitochondrion. Subfractionation of the mitochon-

590

1 2 3 4 5 6 7 8

L precursor

mature

Fig. 4. In vitvo import and processing of the internal NADH dehydrogen- ase. In-vitro-synthesized [35S]methionine-labeled precursor NADH dehydrogenase (lane 1) was incubated for 30 min a t 29°C with coupled mitochondria (200 pg; lanes 2 and 4-8) in the presence of a-giycerol phosphate (10 mM) and ADP (1 mM). Lane 1, in-vitro-synthesized prccursor; lane 2, precursor incubated with mitochondria; lane 3, mature NADH dehydrogenase prepared by means of native immune precipitation from cells grown in the presence of [35S]methionine; lane 4, precursor incubated with mitochondria and subsequently treated with protease K (75 pg); lanes 5 - 8, same as lane 4 but in the presence of 2 mM o-phenanthroline (lane 5), 1 pM valinomycin (lane 6 ) , 1 pM nigericin (lane 7), or 500 pM flavone, an inhibitor of CI-

glycerol phosphale oxidation (lane 8).

dria into a membrane fraction (inner membrane plus outer membrane) and a soluble fraction (intermembrane space plus matrix) indicated that the mature protein is membrane bound (not shown), most likely to the inner membrane. No precursor or mature protein was detected in the soluble fractions in these experiments.

In the presence of o-phenanthroline, which inhibits the matrix processing peptidase [35], the precursor NADH de- hydrogenase is not cleaved, yet it has reached a location where it is resistant to protease K. Fig. 4 further shows that import and subsequent processing of the internal NADH dehydro- genase is dependent on the membrane potential component of the proton-motive force, since flavone, inhibiting a-glycerol phosphate oxidation, or valinomycin prevented import and subsequent processing whereas nigericin was without effect. In the presence of carbonylcyanide nz-chlorophenylhydrazone, import and processing were also inhibited (not shown).

The experiments shown in Fig. 4 reveal, for the first time, the import characteristics of a membrane-bound respiratory- chain enzyme which is not part of a larger complex. The results obtained so far indicate that the pathway of import of the internal NADH dehydrogenase is very similar to for example that of the /3 subunit of ATP synthase [29, 30, 351. Whether the NADH dehydrogenase uses the same receptor as the /3 subunit, and how and when the prosthetic group FAD is inserted, remains to be established.

The protein structure of the NADH dehydrogenase

A search in the database indicated that the internal NADH dehydrogenase from yeast mitochondria is homologous to the NADH dehydrogenase from E. coli, cloned and sequenced by Young et al. [18]. Fig. 5 shows the optimal alignment of the two sequences in which 22.2% of the residues are identical after introduction of several gaps. Similarity between the two NADH dehydrogenases starts around residue 50 (unless stated otherwise, residue numbers refer to the sequence of the yeast protein), i.e. the enzyme from E . coli apparently lacks a (cleavable) presequence to direct the protein to the plasma membrane. The similarity between the two NADH dehydro- genases is greatest for the first 400 amino acid residues. No

1 0 20 30 40 50 60 S.cer. MLSKNLYSNRRLLTSTNTLVRFASTRST~PTSFKTMKVIDPQHSDWNVLILG

TTPLKKIVIVG E,coli --_----__--_---_---_----------------------------- . . . .

l o 70 80 90 100 110

S.cer. SGWGAISFLKHID-----TKKYNVSIISPRSYFLFTPLLPSAPVGTEKSIIEPIVNFA

E.coli GGAGGLEMATQLGHKLGRITLVDRNHSHLWKE'LLHEVATGSLDEGVDALSYIAf3A . . . . . . . . . . . . . . .. .. . . . .

20 30 40 50 60

120 130 140 150 160 1 7 0 S.cer. LKKKGNVTYYEAEATSINPDRNTVTIKSLSAVSQLYQPENHLGLH~P~IKYDYLISA

E.coli RNHG--FQFQLGSVIDIDRITIAELRDEKG----------ELL~E~IAYDTL~ 70 80 90 1 0 0 110 1 2 0

. . . . . . . . . . . . . . . . . .

180 190 200 210 220 230 S.cer. VGAEPNTFGIPGVTDYGHFLKEIPNSLEIP.RTF&WLE~LPKGDPERRF&LSIWG

E.coli LGSTSNDFNTPGVKENCIFLDNPHQA----RRFHQEMLNLFLKYSANLGANGKVNIAIVG . . . . . . . . . . . . . . . . . . . . . . . . . .

130 140 150 160 170

240 250 260 270 280 290 S.cer. GGPTGVEAAGELQDYVHQDLWLPALAEE-VQIHLVEALPIVLNMFEKKLSSYAQSHLE

E.coli GGATGVELSAELHNAVKQLHSYGYKGLTNEALNVTLVEAGERILPALPPRISAAAHNELT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 190 200 210 220 230

300 310 320 330 340 350 S.cer. NTSIKVHLRTAVAKVEEKQLLAKTKHEDGKITEETIPYGTLIWATG~VITDLF~I

E.coli KLGVRVLTQTMVTSADEGGLHTK--------DGEYIEADLMWAAGIKAPDFLKDIGGL- 240 250 260 270 280

360 31 0 380 390 400

. . . . . . . . . . . . . . . . . . . . . . . . . .

S.cer. PEQNSSKRGLAVNDFLQVKGSNNIFAIGDN~AGL------PPTAQVAHQEAEYIAKNF

E.coli - - E T N R I N Q L W E P T L Q T T R D P D I Y A I G D C A S C P R P E G G F V - - - - - - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

290 300 310 320 330

410 420 430 440 450 460 S.Cer. DKM?.QIPNFQKNUSRW)KIDLLFEENFFKRFKYNDLGALAYLGSERAIATIRSGKRTF

E.co1i --------------CAM"IIAQKNGKPLKNYQYKDHGSLVSLSNFSTVGSLMGNLTR- . . . . . . . . . . . .

340 350 360 370 380

470 480 490 500 510 S.Cer. YTGGGLMTFYLWRILYLSMILSAFSRLKVFFDWIIUAFFKRDFFKGL

E . coli --GSMMIEGRIARFVYISLYFMlQIALHGYFKTGLMMLVGSINRVIRPRLKLH . . . . . . . . . .

390 400 410 420 430

Fig. 5. Alignment of the amino acid sequences of the NADH dehydrogen- ases from S. cerevisiae and E. coli. See text for further details.

unique alignment could be assigned to the C-terminal part of the two proteins.

Since the two dehydrogenases both catalyze the same type of oxidoreduction reaction, contain the same prosthetic group (FAD) and are both membrane bound [3, 8, 181, one may expect to find conservation of the primary structure between the two proteins of domains involved in the binding of NAD(H), FAD(Hz), ubiquinone and a hydrophobic domain involved in the binding of these proteins to the membrane. As indicated in Fig. 5 , the two proteins are best conserved over residues 230 -275. This segment conforms to all of the 11 (10 in the case of the sequence of the E. coli enzyme) criteria discussed by Wieringa et al. [36, 371 required to fold into a Pa/3 structure and is involved in the binding of the ADP moieties of NAD(H), NADP(H) and FAD(H2). Accordingly, this segment contains three glycine residues, spaced GXGXXG, at the beginning, and a negatively charged resi- due, Glu272, at the end. Residues 230-234 are proposed to fold into a p structure, residues 238-250 into an a-helix, residues 251 -261 form a loop and residues 268 -212 fold into a /3 structure. The loop in these two NADH dehydrogen- ases seems to be much larger than in the dehydrogenases examined so far by X-ray crystallography 1371, which usually contain a loop sequence of about four residues. Since the flap domain binds the adenine moiety, we do not know whether this domain is involved in the binding of NADH or FAD(H2) in these two dehydrogenases.

591

2 . 0 0 I . o o > 0.00 .- - 1 . 0 0

c - 2 . 0 0 Q

1 . o o

- I . o o - 2 . 0 0

S . cerevisiae c .- - .-

2 2.00

E. coli z 0.00 I

100 2 0 0 3 0 0 4 0 0 5 0 0

Fig. 6. Hydropathy profiles of the NADH dehydrogenases from yeast and E. coli. The hydropathy profile was calculated as in I461 with thc MacVcctor v3.04 application (International Biotechnologies Inc., 1989, 1990) using a window of seven residues. To compare directly thc hydropathy profiles of the two scquences, the gaps that are shown in the alignment of Fig. 5 were introduced as blanks in thc calculation. This slightly affects the calculated hydropathy index in the boundary regions.

As noted previously by Young et al. [18], the enzyme from E. coli may have arisen from a gene-duplication event, i.e. residues 1 - 165 (E. coli sequence) are homologous to residues 266- 331. A possible gene-duplication event is not evident from the sequence of the yeast NADH dehydrogenase. The similarity between the first and the second stretch of 165 residues of the enzyme from E. coli is very clear when comparing residues 171 -215 (the Pa$ domain discussed above) with residues 6 - 40, which, in fact, have previously been proposed to fold into a PLXP domain [37]. The main difference between these two flap domains is the length of the loop, seven residues for the domain closest to the N-terminus compared to 18 for the other domain 1371. Thus two nucleotide-binding domains may be identified in the sequence of the E. coli NADH dehydrogenase. However, according to the alignment in Fig. 5, the 8.P domain of residues 6-40 of the E. coli sequence is apparently not represented as such in the sequence of the enzyme from yeast. Although the comparable sequence in yeast (residues 55 - 85/90) contains three glycine residues, their spacing deviates from the consensus and, in addition, no negatively charged residue at a position homolo- gous to Asp40 (E. coli) is present in the yeast sequence. It is therefore questionable whether the N-termini of the two NADH dehydrogenases are directly involved in the binding of the substrate, NADH, or the prosthetic group FAD.

Apart from the nucleotide-binding domains discussed above, the sequence of the two enzymes are relatively well conserved in residues 167 - 188, 337 - 343, 378 - 385 and 390 - 400. The function of these segments is unknown. Resi- dues 337 - 343 contain the only tryptophan residue conserved between the two sequences and, by analogy with the bacterial reaction center, this residue, might play a role in the binding of ubiquinone 138, 391.

The similarity between the hydropathy profiles of the two NADH dehydrogenases is quite low (Fig. 6). The E. coli NADH dehydrogenase contains a stretch of 20 uncharged amino acids at the C-terminus (residues 362 - 381), but this segment is probably not sufficiently hydrophobic to anchor the enzyme to the membrane. At a comparable position in the yeast enzyme, the residues 465 - 489 are fairly hydrophobic, but this stretch is interrupted by Arg479. Although argininc may be present at the hydrophobic half of an amphipathic a- helix [40], we do not consider it likely that Arg479 is buried in the membrane. For the same reason, residues 77- 102, interrupted by Arg85, are not likely to fold into a membrane- spanning a-helix. Moreover, the comparable sequence in the E. coli enzyme is not hydrophobic at all.

Fig. 6 also shows that the central /lap domains of the two NADH dehydrogenases are relatively hydrophobic. Likewise, residues 7-24 of the E. coli sequence are relatively hydro-

phobic, but interrupted by Glul8. The comparable sequence in the yeast enzyme (residues 54-69) is also quite hydro- phobic, but probably too short to span the membrane. In summary, there is no clear indication from the primary struc- ture that the NADH dehydrogenase from yeast or E. coli is anchored to the membrane by a membrane-spanning a-helix.

Furthermore, the primary structure and the known properties of the yeast enzyme are not consistent with the formation of a B-barrel structure important in the binding of the porins to the membrane [41,42]. In addition, the primary structure does not contain a consensus sequence to which a (phospho)lipid might be covalently linked [43, 441. Finally, the NADH dehydrogenase might be bound to the membrane by means of hydrophobic interaction of a non-transmembrane a-helical segment as has been recently proposed for the mode of binding of the Rieske Fe-S protein to the ubiquino1:cyto- chrome-c oxidoreductase [45]. However, the observation that both the enzyme from E. coli and from yeast are not removed by sonication, that they are solubilized by mild treatment with detergent and that both are purified as single subunit enzymes without loss of activity during purification, does not favour the suggestion that these NADH dehydrogenases form part of a larger membrane-bound respiratory-chain complex.

From the foregoing discussion, the question arises as to how these NADH dehydrogenases are bound to the mem- brane. This point, and the way in which FAD is inserted into the NADH dehydrogenase, are currently under investigation.

We thank W. C. van Heeswijk for his contribution to somc of thc experiments, Gist-Brocades for their financial support in part of this work. S. de V. thanks the Royal Dutch Academy of Scicnces (KNAW) for the award of a long-term fellowship.

REFERENCES 1, Ohnishi, T., Kawaguchi, K. & Hagihara, B. (1966) J . Biol. Chun.

2. Von Jagow, G. & Klingenberg, M. (1970) Eur. J. Bioniem. 12,

3. De Vries, S . & Grivell, L. A. (1988) Eur. J . Biochum. 176, 377 -

4. De Vries, S. & Marres, C. A. M. (1987) Biochim. Biophys. Acta

5. Marres, C. A. M., de Vries, S. & Grivell, L. A . (1991) Eur. ./.

6. Hollcnberg, C. P., Riks, W. P. & Borst, P. (1970) Biochim. Bio-

7. Douce, R. & Neuburger, M. (1989) Annu. Rev. Plant Physiol. 40,

8. Young, I. G., Rogers, B. L., Campbell, H. B., Jaworowski, A. &

241,1797 - 1806.

583 - 592.

384.

895,205 -239.

Biochem. 195, 857-862.

phvs. Acta 201, 13-19.

371 -414.

Shaw, D. C. (1981) Eur. J . Biochcm. 116, 165-170.

9. Bergsma, J., van Dongen, M. B. M. & Konings, W. N. (1982) Eur. J . Biochem. 128, 151 - 157.

10. Yagi, T. (1987) Biochem. 26,2822-2828. 11. Cook, N . D. & Cammack, R. (1985) Biochim. Biophys. Acta 827,

12. Schwitzguebel, J.-P. & Palmer, J. M. (1982) J . Bacteriol. 149,

13. Friedrich, T., Hofhaus, G., Ise, W., Nehls, U., Schmitz, B. &

14. Tuschen, G., Sackmann, U., Nehls, U., Haiker, H., Buse, G. &

15. Ohnishi, T. (1970) Biochem. Biophys. Res. Commun. 41, 344-

16. Mackler, B. & Haynes, B. (1973) Biochim. Biophys. Acta 292,

17. Jaworowski, A., Campbell, H. D., Poulis, M. & Young, I. G.

18. Young, I. G., Rogers, B. L., Campbell, H. D., Jaworowski, A. &

19. Yannish-Perron, C., Vieira, J. & Messing, J. (1984) Gene (Amst.)

20. Boyer, H. W. & Roulland-Dussoix, D. (1969) J . Mol. Bid. 41,

21. Gietz, R. D. & Sugino, A. (1988) Gene (Amst.) 74, 527-534. 22. Westerbeek-Marres, C. A. M., Moore, M. M. & Autor, A. P.

23. Gasser, S. M. (1984) Methods Enzymol. 97, 329-336. 24. Maniatis, T., Fritsch, E. F. & Sambrook, J . (1982) Molecular

cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

25. Klebe, R. J., Harries, J. V., Sharp, D. & Douglas, M. D. (1983) Gene (Amst.) 25, 333-341.

26. Lowry, 0. H., Roseborough, N. J., Farr, A. L. & Randall, R. J. (1951) J . Biol. Chem. 193,265-275.

27. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Null Acad.

30 - 35.

612 -619.

Weiss, H. (1989) Eur. J . Biochem. 180, 173-180.

Weiss, H. (1990) J . Mol. B i d . 213, 845-857.

352.

88-91.

(1981) Biochemistry 20, 2041 -2047.

Shaw, D. C. (1981) Eur. J . Biochem. 116, 165-170.

33,303-119.

459 -472.

(1988) Eur. J . Biochem. 174, 611 -620.

Sci. USA 74, 5463 - 5461.

28. Laemmli, U. K. (1970) Nature 227,680-685. 29. Pfanner, N. & Neupert, W. (1990) Annu. Rev. Biochem. 59,331 -

30. Grivell, L. A. (1988) Znt. Rev. Cytol. 3, 107-141. 31. Chou, P. Y. & Fasman, G. D. (1978) Annu. Rev. Biochem. 47,

32. Gamier, J., Osguthorpe, D. J. & Robson, B. (1978) J . Mol. Bid.

33. Von Heijne, G. (1986) EMBO J . 5, 1335-1342. 34. Chevallier, M.-R., Bloch, J.-C. & Lacroute, F. F. (1980) Gene

35. Bohni, P. C., Daum, G. & Schatz, G. (1983) J . Bid. Chem. 258,

36. Wieringa, R. K., De Maeyer, M. C. H. & Hol, W. G. J. (1985)

37. Wieringa, R. K., Terpstra, P. & Hol, W. G. J. (1986) J . Mol. Biol.

38. Michel, H., Epp, 0. & Deisenhofer, J. (1986) EMBO J . 5, 2445-

39. Deisenhofer, J. & Michel, H. (1989) EMBO J . 8, 2149-2170. 40. Cornette, J. L., Cease, K. B., Margalit, H., Spouge, J. L.,

Berzofsky, J. A. & DeLisi, C. (1987) J . Mol. Bid. 195, 659- 685.

353.

251 -276.

120, 97-120.

(Amst.) 11, 11-19.

4937-4943.

Biochemistry 24, 1346 - 1 357.

187, 101 -107.

2451.

41. Vogel, H. & Jahnig, F. (1986) J . Mol. Biol. 190, 191 - 199. 42. Paul, C. & Rosenbusch, J. P. (1985) EMBO J . 4, 1593- 1597. 43. Mumby, S. M., Casey, P. J., Gilman, A. G., Gutowski, S. &

Sternweis, P. C. (1990) Proc. Nut1 Acad. Sci. USA 87, 5873 - 5817.

44. Glomset, J. A., Gelb, M. H. & Farnsworth, C. C. (1990) Trend.s Biochem. Sci. 15, 139-142.

45. Gonzilez-Halphen, D., Vazquez-Acevedo, M. & Garcia-Ponce, B. (1991) J . Biol. Chem. 266,3870-3876.

46. Kyte, J. & Doolittle, R. F. (1982) J . Mol. Biol. 157, 105- 132.