purification and characterization of a rotenone-insensitive nadh: q6 oxidoreductase from...

Eur. J. Biochem. 176,377 - 384 (1988) 0 FEBS 1988

Purification and characterization of a rotenone-insensitive NADH : Q6 oxidoreductase from mitochondria of Saccharomyces cerevisiae Simon de VRIES and Leslie A. GRIVELL Section for Molecular Biology, Laboratory of Biochemistry, University of Amsterdam

(Received April 1/May 5 , 1988) - EJB 88 0378

A mitochondrial NADH : Q6 oxidoreductase has been isolated from cells of Saccharomyces cerevisiae by a simple method involving extraction of the enzyme from the mitochondrial membrane with Triton X-100, followed by chromatography on DEAE-cellulose and blue Sepharose CL-6B. By this procedure a 2000-fold purification is achieved with respect to whole cells or a 150-fold purification with respect to the mitochondrion. The purified NADH dehydrogenase consists of a single subunit with molecular mass of 53 kDa as indicated by SDS/polyacrylamide gel electrophoresis. The enzyme contains FAD, non-covalently linked, as the sole prosthetic group with

= - 370 mV and no iron-sulphur clusters. The enzyme is specific for NADH with apparent K , = 31 pM and was found to be inhibited by flavone ( Iso = 95 pM), but not by rotenone or piericidin. The purified enzyme can use ubiquinone-2, -6 or -1 0, menaquinone, dichloroindophenol or ferricyanide as electron acceptors, but at different rates. The greatest turnover of NADH was obtained with ubiquinone-2 as acceptor (2500 s- ’). With the natural ubiquinone-6 this value was 500 s-’. The NADH:Q2 oxidoreductase activity shows a maximum at pH 6.2, the NADH : Q6 oxidoreductase activity is constant between pH 4.5 - 9.0. The amount of enzyme in the cell is subject to glucose repression; it increases slightly when cells, grown on glucose or lactate, enter the stationary phase. The experiments performed so far suggest that the enzyme purified in this study is the external NADH : Q6 oxidoreductase, bound to the mitochondrial inner membrane and that it is involved in the oxidation of cytosolic NADH. The relation of this enzyme with respect to various other NADH dehydrogenases from yeast and plant mitochondria is discussed.

Unlike mammalian mitochondria, mitochondria from plants, fungi and yeast are capable of oxidizing externally added NADH. Since NADH cannot pass the mitochondrial inner membrane [l], the NADH-binding site on the NADH dehydrogenase responsible for the oxidation of external (or cytosolic) NADH must face the intermembrane space in these organisms. In contrast, the mammalian NADH dehydrogenase (complex I) is involved in the oxidation of intra-mitochondrial NADH, produced by the action of the Krebs cycle, and its NADH-binding site faces the mitochondrial matrix.

In general, the oxidation of cytosolic NADH via the mitochondrial respiratory chain is not coupled to site I phosphorylation, nor is it inhibited by rotenone or piericidin indicating that the external NADH dehydrogenase is structur- ally different from complex I. Furthermore, the external NADH dehydrogenase from Saccharomyces cerevisiae, S . carlsbergensis and Candida utilis lacks Fe-clusters [2 - 41. Un- like C. utilis, S. cerevisiae and S. carlsbergensis are not capable of expressing a complex I type of NADH dehydrogenase as evidenced from the finding that, under all growth conditions so far examined, the oxidation of NADH (intra- or extra- mitochondrial) is insensitive to rotenone or piericidin, even

Correspondence to L. A. Grivell, Sectie Moleculaire Biologie, Universiteit van Amsterdam, Kruislaan 318, 1098 SM Amsterdam, The Netherlands

Abbreviations. Qz, Qs, or Q L o , ubiquinone-2, -6 or -10; Em,,, standard midpoint potential at pH = x ; FCCP, carbonylcyanide p-triflouromethoxyphenylhydrazone.

Enzyme. NADH: Q6 oxidoreductase (EC 1.6.5.3).

though under some conditions site I phosphorylation is in- duced [2, 51.

In the yeast C . utilis and fungus Neurospora crassa the presence of complex I (measured as piericidin-sensitive site I phosphorylation) is dependent on growth phase. For instance, in C. utilis growing exponentially only the external NADH dehydrogenase seems to be present, whereas in cells harvested in the stationary phase this enzyme is absent and only complex I is present [4, 6, 71 (but see [3, 81). The properties of complex I from C. utilis or N . crassa are in many respects similar to those of the mammalian enzyme [9- 131.

In plant mitochondria the situation with respect to the number and identity of the NADH dehydrogenases is even more complex than in yeast. In addition to complex I, these mitochondria contain two piericidin-insensitive (‘non-phosphorylating’) NADH dehydrogenases located on opposite sides of the mitochondrial inner membrane. The possible roles played by the three dehydrogenases in vivo are discussed elsewhere [14, 151.

From the points mentioned above, one may infer that, dependent on strain and growth phase, mitochondria from yeast may contain a complex I type of NADH dehydrogenase, an ‘external’ NADH dehydrogenase (piericidin-insensitive; non-phosphorylating) or an ‘internal’ NADH dehydrogenase (piericidin-insensitive, site I phosphorylation). In addition, other data [l] are suggestive of the presence of an ‘internal’ NADH dehydrogenase (piericidin-insensitive, non-phosphorylating). Which combination of these four NADH dehydrogenases occurs in the mitochondrial membrane is not

378

known precisely. However, there is good indication that in exponentially growing yeast both an 'internal' and 'external' NADH dehydrogenase (and both piericidin-insensitive, nonphosphorylating) are present [I, 4, 6, 161.

In order to clarify this complex situation we started to purify the mitochondrial NADH dehydrogenase from S. cerevisiae. We report here the purification and (partial) characterization of the external NADH : Q6 oxidoreductase.

MATERIALS AND METHODS

Yeast strains

The NADH dehydrogenase was purified from commercial bakers yeast. For some experiments S. cerevisiae D273-10B (a) or S. carlsbergensis was used. Cells were grown on semi- synthetic medium with lactate as carbon source [16] or on a medium containing 1 YO yeast extract, 2% bactopeptone and any of the following carbon sources: glucose (0.4- 10?40), glycerol (2%, mass/vol.) or ethanol (1 YO, mass/vol.). Mito- chondria capable of oxidative phosphorylation were prepared according to [16], except that in the case of D273-10B the concentration of sorbitol in the buffer used for preparation of protoplasts was raised from 1.3 M to 2 M.

Materials

Column material was from Whatman (DEAE-cellulose DE-52) and Pharmacia (Blue Sepharose CL-6B): flavone and deamino NADH were from Sigma; Piericidin was purified from Streptoverticillium mobaerensis as described in [17].

Purification scheme

About 10 kg yeast was broken mechanically using a Dyno- Mill apparatus (type KDL, W. A. Bachofen Maschinenfabrik, Basel) and crude mitochondria were obtained by differential centrifugtion essentially as described in [18]. They were stored at - 70°C. An amount of crude mitochondria, equivalent to about 0.5 kg yeast, was thawed and suspended in 20 mM Tris/HCl, 1 mM EDTA, pH 7.6 to a concentration of 15 mg protein/ml. Phenylmethylsulfonyl fluoride (0.5 mM) was then added. The mitochondrial suspension was sonicated four times for 10 s at O"C, centrifuged, washed once and suspended in the same buffer (also in the presence of 0.5 mM phenylmethylsulfonyl fluoride) to a concentration of 15 mg protein/ml. All further steps were performed at 0-4°C. Tri- ton X-100 and NaCl were subsequently added to final concentrations of 0.18% and 200 mM, respectively, and incubated for 15 min. After centrifugation (40 min at 40000 x g), the supernatant was dialyzed overnight versus 20 vol. 20 mM Tris/HCl, 1 mM EDTA, 0.02% Triton X-100, pH 7.6. About 250 g DEAE-cellulose (DE-52) was washed once with 0.1 M NaOH, once with 0.1 M HC1 and with water until pH = 4. Thereafter the column material was wahed with elution buffer (20mM Tris/HCl, 1 mM EDTA, 0.05% Triton X-100, pH 7.6) until pH = 7.6 was obtained, poured into a column (5 x 30 cm) and equilibrated for 24 h before use.

After application of the dialyzate, the column was washed with 1.5 vol. elution buffer before the gradient (0-250 mM NaCl, 5 column volumes) was started. NADH : Q2 reductase activity eluted from the column between 30 - 80 mM NaCl. Fractions containing more than 50% of the activity of the peak fraction (35-70 mM NaCI) were pooled and applied batchwise to an amount of blue Sepharose (washed five times

with 20 mM Mops/KOH, 100 mM NaCl, 1 mM EDTA, 0.05% Triton X-100, pH 7.6), sufficient to bind 90% of the activity. Subsequently the blue Sepharose was poured into the column (1 x 20 cm) washed with 5 column volumes and a gradient (0.1 - 1.0 M NaCl, 4 vol.) was applied, under which conditions contaminating proteins elute. The NADH dehydrogenase could be eluted from the column after a passage of 5 - 10 column volumes of 1 M NaCl. Fractions with activity were pooled, concentrated on an Amicon diaflow apparatus, diluted ten times with 20 mM Mops/KOH, 1 mM EDTA, pH 7.6 to reduce the concentration of NaCl and Triton X-100, and concentrated again. The purified NADH : Qs oxidoreductase was frozen and stored in liquid nitrogen.

Enzyme assays

The standard assay medium consisted of a buffer containing 20 mM Mops/KOH, 1 mM EDTA, 0.03% Triton X-100, 200 mM KCl, 100 pM NADH, 100 pM Q2, 5 pM antimycin, pH 6.2. For assay with Q6 (100 pM), dichloroindophenol (50 pM) or ferricyanide (0.5 mM) the same medium (without Q2) was used. Stock solutions of Q2 (25 mM) were made in ethanol, those of Qs (15-20 mM) in 10% Triton X-100. Oxidation of NADH was monitored spectrophotometrically at 340 nm, reduction of dichloroindophenol and ferricyanide at 600 nm and 420 nm, respectively. Concentrations of Q2 and Qs used in the assay are saturating for the enzymic activity.

Procedures

Determination and characterization of flavin was performed by the fluorometric method described in [19]. The redox state of Q6 was determined with a rapid-mixing rapid- quenching device in combination with HPLC as in [20]. Mono- clonal and polyclonal antibodies were raised against purified one- or two-subunit preparations according to [21, 221. The two types of antibodies were found to be immunologically fully cross-reactive against the two kinds of purified enzyme. Samples used for SDS/polyacrylamide gel electrophoresis and Western blotting were denatured in the presence of 10 mM phenylmethylsulfonyl fluoride in order to minimize proteolytic degradation (cf. [23]). Cell lysates and mitochondria were first precipitated with 5% trichloroacetic acid. In purified preparations phenylmethylsulfonyl fluoride could be omitted. Protein was determined according to Lowry et al. [24]. This method overestimates the protein content of the purified enzyme as determined by amino analysis by 24%. Iron was determined by flameless atomic absorption as in [25] using a Hitachi 180-80 polarized Zeeman atomic absorption spectro- photometer.

RESULTS

In Table 1 the results of the purification of the NADH dehydrogenase are shown. The procedure, outlined in detail in Materials and Methods, involves breakage of the yeast cells mechanically followed by differential centrifugation. The crude mitochondrial fraction thus obtained is subsequently sonicated and washed. Thereafter NADH dehydrogenase activity is almost completely solubilized, using a 0.18% Triton X-100/200 mM NaCl. Under these conditions the much more hydrophobic cytochrome c reductase and cytochrome c oxidase are not solubilized. After extraction and centrifugation the supernatant is dialyzed overnight to remove the salt. The dialyzate is applied to a DEAE-cellulose column. Activity elutes at about 50 mM NaCl. Subsequently, NADH dehydro-

3 79

Table 1. Pur(fi'c~ution of'the N A D l f : Q , o.uidoreductuse.from S. cerevisiae n.d. = not determined

Fraction Protein Total Specific activity Purification Y ield

Q z Q6

~~~~ ~~ ~~ activity

m& pmol NADH/min pmol NADH mg- ' min- ' -fold '%I

Broken cells 72906 57057 0.78 n.d. 1 .0 I00 Supernatant 1000 x g 44 179 52 302 1.18 n.d. 1 .5 91.7 Mitochondria 13515 38533 2.85 0.095 (30)h 3.6 67.5 Son ica ted mi t oc hond ia 9 592 34 320 3.60 0.119 (30) 4.6 60.2 Supcrnatant TritonlNaCI 3 306 32 364 9.79 0.306 (32) 12.5 56.7 DE A E-cellulo~e 125 9 240 73.7 2.95 (25) 94.1 16.2 Blue Scpharosc 2.5 4125 1671 61 .9 (27) 2134 7.2

a The mitochondria from which the NADH :Qh oxidoreductase was purified did not show piericidin (rotenone)-sensitive or inscnsitivc site I phosphorylation (cf. [2, 51 and Introduction).

Numbers in parentheses indicate the ratio of activities with Qz and Qb as acceptor in the standard assay medium.

Fig. 1 . PoIj ,p~pt idi~ cotnposition of' wrious puri fkd prepurutions of' the .VA D H tk,lij.d~og,.entr.sc. Lane 1. molecular mass standards (values in kDa on left): lanes 2 - 5 , various purified preparations consisting of ii singlc smaller band (lanes 2 and 4) and of two bands in different proportions (lanes 3 and 5); lane 6. pooled and concentrated fractions from DEAE-ccllulosc; lane 7. supcrnatant after extraction of mitochondria with Triton X- l 00 and NaCI. Electrophoresis was performed on ii 12.5"/0 (0.075'1/0 bisacrylamide) SDS/polyacrylamide slab gel

genase is bound t o a blue-Sepharose column. The enzyme is eluted at 1 M NaCI and concentrated.

By this procedure the NADH dehydrogenase is purified by a factor of more than 2000 from whole cells based on the increase in specific activity using Qz as electron acceptor. I t is also seen in Table 1 that the ratio of the activities with Qz and Q6 (the natural acceptor) is almost constant (about 30) at the various stages of the purification. This finding indicates that the purified enzyme is a genuineYmitochondria1) NADH :Q6 oxidoreductase.

As shown in Fig. 1 , the polypeptide composition of the various purified preparations is variable. Some preparations were found to contain only a single band of molecular mass 51 kDa; other preparations consisted of two bands with molecular masses of 51 and 53 kDa, though in variable stoichiometries. All these preparations are enzymically active (see below).

444 nm

o ! 300 400 500 600

wavelength (nm)

Fig. 2. O p t i d pr.opiJrriiJs of'thc. purifi'id N A D H dc.li~drogc,na.sr,. Top: optical spectrum of oxidized and reduced (0.1 m M NADH) enzyme. Bottom: fluorescence excitation (left part) and emission (right part) spectrum

Optical spectroscopy (Fig. 2) shows a flavin-like absorption spectrum with a maximum at 444 nm. This peak disappears after reduction by NADH or dithionite. The fluorescence excitation and emission spectra are also typical of flavin. Flavin analysis by the fluorescence method described [ 181 indicated that the prosthetic group is FAD, non-covalently bound. The most pure preparations contain about 17 nmol FAD/mg protein (cf. Table 2). Assuming a single molecule of FAD per subunit, the best preparation is at least 90% pure.

A potentiometric titration employing the NADiNADH redox couple yielded (FAD/FADH2) = -370 mV, i.e. about 30 mV lower than that of the NAD/NADH couple at pH 7.6. No FADH intermediate could be detected during the titration either optically or by EPR (data not shown). In addition, EPR studies performed between 6 - 100 K failed to reveal a semi-quinone signal of Qh (added to the purified preparation poised at different redox potentials) or any other signal except a small resonance at g = 4.3 from adventitious Fe3', in agreement with the amount of iron present in the enzyme (cf. Table 2).

380

Table 2. Properties of the purified N A D H : Q6 oxidoreduztase from S . cerevisiae The purity was calculated from the FAD content assuming one molecule of FAD per subunit. Preparations 2, 3 and 5 correspond to the gel patterns shown in lanes 4, 3 and 5, respectively, of Fig. 1 . Numbers in parentheses are the approximate ratio of the two bands. Maximal turnover in mitochondria was determined by the rapid mixingpapid quenching/HPLC method (cf. Fig. 5 )

Preparation Bands in F A D Purity Maximal electrophoresis content turnover

nmol/mg s -

1 . 1 7.8 39.8 1753 _. 3 1 14.8 75.4 1713 3. 2 (1 :1 ) 7.7 39.1 2614 4. 2 (3: 1) 10.1 53.7 2574 5. 2 ( > 5 : 1 ) 17.2 91.0 2592

Other properties

Molecular mass according to SDS-PAGE: band 1 band 2

53 f 1.5 kDa 51 f 1.5 kDa

Kinctic properties: K , (NADH) = 31 pM Maximal turnover with Qh as acceptor in:

(a) Triton assay medium: 80- 100 s - ' b) mitochondria: 500-550 s - l

L~~~~~ (ox.) = 9.2 mM ~ cm- Em, , = -370 mV

Properties FAD: non-covalently linked

Metal content: 0.18 mol Fe/mol flavin before dialysis 0.08 mol Fe/mol flavin after dialysis

The turnover number of the various preparations was determined from the FAD content and the specific activity. In addition, the gel pattern was analyzed with respect to the ratio ofthe smaller (51-kDa) and larger (53-kDa) polypeptide. Inspection of Table 2 indicates that the turnover number of preparations containing only the smaller polypeptide is about 30-40%, lower than that of preparations containing pre- dominantly the larger polypeptide. This suggests that the smaller polypeptide is a proteolytic degradation product of the larger (native) polypeptide (cf. [23]). This conclusion is corroborated by the results from a Western blot shown in Fig. 3. The monoclonal antibody recognizes in whole cells and mitochondria a single polypeptide corresponding to the larger peptide found in purified preparations. Upon extraction with Triton X-100 the smaller fragment is also present. The extent of conversion from the larger into the smaller polypeptide was found to be only partially sensitive to the addition of the protease inhibitor phenylmethylsulfonyl fluoride during the purification. Extensive washing of the sonicated mitochondria also decreased the amount of the smaller polypeptide present in the final preparation, which itself was apparently free of proteolytic activity. So far, however, we have not been able to prepare a purified enzyme containing exclusively the larger, native, polypeptide.

The amount of the NADH dehydrogenase in exponentially growing yeast cells relative to the total amount of protein is dependent on the carbon source used (Fig. 3). With non- fermentable carbon sources like lactate (glycerol, ethanol, data not shown) the amounts is highest. With glucose the amount of the NADH dehydrogenase is thus subject to catabolite repression. When cells are harvested in the stationary (24 h) phase the amount of the NADH dehydrogenase is

Fig. 3. Wcstcvn blot ana1ysi.s of' the NADH dehydrogenase u s present at vurious stages of' the purifi'cation ( A ) and as present in whole ~ ~ ~ 1 1 s obtained ,from different grow,th ronditions ( B ) . Note that since the monoclonal antibody recognizes a single band in cell lysates the smaller bands seen in Triton-solubilized fractions are probably degradation products. (A) Lane 1 , whole cells grown on lactate, harvested i n the logarithmic phase; lane 2, mitochondria, sonicated and extensively washed in the presence of phenylmethylsulfonyl fluoride; lane 3, supernatant after extraction of mitochondria (treated as in lane 2) with Triton X-100 and NaCl; lane 4, pooled DEAE-cellulose fractions; lanes 5-8, purified NADH dehydrogenasc (lanes 5.4.3 and 2 in Fig. 1 , respectively) extracted from sonicated mitochondria treated as in lane 2 (lane 5) or from mitochondria treated as in lane 9 (lanes 6, 7 and 8); lane 9, supernatant after Triton extraction of mitochondria, sonicated in the absence of phenylmethylsulfonyl fluoride and washed once. (B) Lanes 1 -4, cell lysates from cells harvested in the stationary phase; lane 1 , lactate; lane 2, 0.4% glucose: lane 3, 2% glucose; lane 4. 10% glucose; lane 5, as lane 9 in A; lanes 6-9, cell lysates from cell harvested in the logarithmic phase; lane 6, lactate; lane 7, 0.4% glucose; lane 8, 2% glucose; lane 9, 10% glucose. All lanes, except lane 5 , contain the same amount of protein. Molecular mass values are shown in kDa on left

similarly dependent on the carbon source used for growth and the amount is slightly higher than that found in exponentially growing cells. This is in marked contrast to the behaviour of the counterpart of this NADH dehydrogenase in C'. utilis [6].

Kim. t ic proper t irs

The NADH dehydrogenase as present in either mitochondria of fractions obtained during the purification, or the final

381

100 -

80 -

60 -

40 -

20 -

14 -I

>

12

10

8

6

4

2

0 6 1 0 20 30 40 50

VIS Fig. 4. Eadie-Hofsteeplot showing the equivalence of the K, f o r NADH for the NADH:Q2 oxidoreductase and NADH oxidase activities in various stages of purity of the NADH dehydrogenase. NADH:Q2 oxidoreductase activity was measured in the standard assay medium with purified enzyme (+), supernatant after extraction of mitochondria with Triton X-100 and NaCl (0) and in sonicated (uncoupled) mitochondria (U). NADH oxidase activity was measured with intact mitochondria in the presence of 2 pM FCCP in 0.65 M sorbitol (cf. [16]) (0). Activities in the different preparations and assay media were normalized to the value measured at 90 pM NADH. The slope of the line corresponds to K , = 31 pM

purified preparation (either as a one-polypeptide or as a two- polypeptide preparation) shows simple Michaelis-Menten kinetics with respect to NADH with K, = 31 pM in the standard assay using Q2 as acceptor (Fig. 4). Furthermore, a similar value of the K , (33 pM) was obtained in intact mitochondria (i.e. capable of oxidative phosphorylation) for the NADH-oxidase activity. Other substrates like NADPH or deamino NADH showed rates that were at least 250 times lower than obtained with NADH at any pH between 5.5 and 8.5. The enzyme is thus specific for NADH.

As electron acceptors the enzyme can use Q2, Q6, Qlo, dichloroindophenol, menaquinone or ferricyanide. The ratio of the activities with Q2, dichloroindophenol, ferricyanide and Q6 equals 30: 8:6: 1, respectively (pH 6.2; NADH = 100 pM). The short-chain analogue Q 2 is thus much more active than the natural Q6. This is partly due to the presence of Triton X-100 in the assay medium, an absolute prerequisite to solubilize Q6 or Qz at concentrations above 20 pM (cf. [26 - 281). When the initial rate of reduction of Q6 was determined in mitochondria, in the absence of Triton X-100, using a rapid-mixing/rapid-quenching apparatus followed by extraction and chromatography on HPLC [20], a rate fivefold higher was found, but, nevertheless, still fivefold lower than with Qz as acceptor (cf. Fig. 5 and Table 2). Also shown in the figure is the effect of the inhibitor flavone. It inhibits the steady-state NADH : Q2 oxidoreductase or NADH oxidase activities, half-maximal inhibition being obtained at 95 pM. Furthermore, flavone inhibits the reduction of Q6 in the pre- steady state (Fig. 5). Other compounds like NAD, MgADP, MgATP, Ca2' (all at 2 mM) or piericidin, rotenone (both at 10 pM) or adenosyl-3'-0-{ 3-[N-(4-azido-2-nitrophenyl)- amino]propionyl}-NAD (50 pM) were without inhibitory or stimulatory effects.

In Fig. 6 the effect of pH on the activity of the enzyme as present in mitochondria or as in the final preparation with Q2 or Q6 as electron acceptor is shown. With Q6 as acceptor the activity is almost constant between pH 4.5 - 9.5, whereas with

A A loo k,,

20 If 0 -D 0 100 200

1 " 8

0 100 200 300

pM flavone added Fig. 5. The effect ofjlavone on the activity of the NADHdehydrogenase. (A) Pre-steady-state kinetics of oxidation/reduction of Q6 in sonicated mitochondria as obtained by the rapid-mixinglrapid-quenching/ HPLC method. In the traces A and B sonicated mitochondria (10 mg/ ml) without or with cytochrome c ( S . cerevisiae, 2 pM), respectively, were incubated with 2 m M NADH for 10min and subsequently mixed with an aerobic buffer. In the traces C and D mitochondria (freed of cytochrome c) without or with flavone (1 mM), respectively, were mixed with a buffer containing 2 mM NADH. 100% (reduction) of Q6 corresponds to 14.9 mol Q6(H2)/mol bcl (or NADH dehydrogenase). (B) Plot of the NADH:Q2 oxidoreductase activity of sonicated mitochondria as a function of the concentration of flavone (duplicate experiment). 50% inhibition was obtained at 95 pM flavone

7

x > c .- .- c

2

/ o l . , . , . , . , . , . I . , 4

3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

PH Fig. 6. ActivitylpH profile of the NADH:Q, and Q6 oxidoreductase activities in the purified enzyme and in sonicated mitochondria. ( A ) and (A): Q6 was used as acceptor in purified enzyme and mitochondria, respectively. 100% activity corresponds to 51 and 0.37 pmol NADH mg-' min-I, respectively. (0) and (U): Q2 was used as acceptor in purified enzyme and mitochondria, respectively. 100% activity corresponds to 1530 and 9.9 pmol NADH mg-I min-I, respectively. Concentrations of Q2 and Q6 used were saturating for the enzymic activity in the pH range studied

382

Qz as acceptor a sharp maximum at pH 6.2 is obtained. Note, however, that at all pH values the absolute rate with Q2 is higher than that with Q6. At pH values lower than 4.5 and higher than 9.5, enzymic activity is irreversibly destroyed.

DISCUSSION

The method described in this paper to isolate the mitochondrial NADH : Q6 oxidoreductase results in a 2000- fold purification of the enzyme with respect to whole cells as indicated by the increase of the specific activity in the NADH : Qz (or Qs) oxidoreductase activity and yields a preparation that is approximately 90% pure based on the FAD content. The finding that the ratio of the activities with Qz or Q6 is constant during purification of the enzyme (Table l), combined with the observation that flavone inhibits the steady-state reduction of Q2 or the NADH oxidase activity and the pre-steady state reduction of Q6 (Fig. 5) , indicates that the enzyme purified in this work is the mitochondrial (cf. Fig. 3) NADH : Qs oxidoreductase. The polypeptide pattern of the pure enzyme or, as inferred from the data presented in Fig. 3, of the enzyme as present after extraction with Triton X-100, is, however, variable. Only the larger of the two subunits, with an apparent molecular mass of 53 kDa, is present in whole cells or mitochondria. We presume that, as is more often observed for the purification of proteins from yeast (cf. [23, 29]), the smaller polypeptide, which still retains 60 - 70% activity, is a proteolytic degradation fragment of the larger polypeptide.

Like many other mitochondrial enzymes from S. cerevisiae the NADH dehydrogenase is subject to glucose repression. In derepressed (lactate) exponentially growing cells, the amount of the NADH dehydrogenase (Fig. 3) appears to be higher by a factor of about 5 - 10 compared to almost completely repressed cells (10% glucose). The level of the NADH dehydrogenase in cells harvested in the stationary phase is some- what higher than that in cells harvested in the logarithmic phase (Fig. 3) for the various carbon sources tested. This is also reflected in the values of the specific activities of the NADH dehydrogenase in various batches of mitochondria prepared from cells harvested in the stationary or logarithmic phase (Table 3). The observation that the amount of NADH dehydrogenase increases slightly as the cells enter the stationary phase is in sharp contrast to the behaviour of the equivalent NADH dehydrogenase present in C. utilis [3, 6, 71 or N . crassa [30]. In the latter two organisms this NADH dehydrogenase activity disappears when the cells enter the stationary phase, whilst, simultaneously, complex I is synthesized. So it is evident than in S. cerevisiae (and S. carlsbergensis [2]) the mechanism of regulation of the level of expression of the NADH dehydrogenase is different from that in C. utilis or N . crassa. This difference may be related to the finding that S. cerevisiue does not (and/or is not genetically competent to) express an enzymically active NADH dehydrogenase of the complex I type (cf. [3]).

The absolute amount of the NADK: Qs oxidoreductase present in mitochondria cannot be determined directly from the flavin content due to the presence of other flavin-containing proteins. Assuming no change of turnover upon purification of the enzyme, its amount can be estimated from the values of the specific activity and flavin content of the purified enzyme (Table 3). Such a calculation shows that the NADH dehydrogenase, be, complex and cytochrome c oxidase are all present in approximately the same amount. For comparison,

Table 3. Some properties of mitochondria f rom S. cerevisiae grown on lactate

Enzyme Growth Specific Amount phase activity

NADH : Q z

pmol nmol/mg min-' mg-'

NADH dehydrogenase log. 7.0- 8.1 0.072-0.084" NADH dehydrogenase stat. 9.2- 11.2 0.095 -0.1 15" QH,: c reductdse stat. - 0.109 -0.1 20 Cytochrome c oxidase stat. - 0.101 -0.123b NADH dehydrogenase purified 1671 17.2

Enzymic Specific activity in activity

mitochondria" sonicated mitochondria inhibition byd

monoclonal polyclonal

pmol min-' mg-' yo

NADH 0 2 0.84 0.77 52.2 69.6 Glycerol-3-

Succinate:02 0.52 0.49 4.7 - NADH: Q2 9.2 9.8 47.5 67.5

- phosphate: O2 0.48 0.43 5.1

a Calculated from the data obtained with the purified enzyme, since the amount of FAD determined in mitochondria is much too high to be accounted for by the NADH dehydrogenase alone.

Determined optically. Activities in the presence of 2 pM FCCP and in the case of

sonicated mitochondria, 5 pM cytochrome c (S . cerevisiae) was also present.

Samples were incubated, 25°C for 15 min, with a fivefold molar excess antibody over NADH dehydrogenase. The monoclonal antibody inhibits the purified enzyme by 85%, the polyclonal by 95%. Neither antibody inhibits the NADH-oxidase activity of mitochondria.

mammalian mitochondria contain about 0.2 mol NADH dehydrogenase (complex I)/mol bcl complex and 0.1 mol/mol cytochrome c oxidase.

Is the NADH : Q6 oxidoreductase described in this paper the external NADH dehydrogenase involved, in vivo, in the oxidation of cytosolic NADH, or the internal NADH dehydrogenase responsible for the oxidation of NADH in the matrix (cf. [I])? We believe that it is the external enzyme, for the following reasons. First, the value of K,,, for NADH for the NADH oxidase and NADH : Q2 oxidoreductase activities in mitochondria are similar (cf. Fig. 4). Second, antibodies that inhibit the activity of the purified enzyme do not inhibit either activity in mitochondria, because the antibody does not cross the outer membrane (cf. [31]), but they inhibit substantially after sonication of the mitochondria (Table 3). It is important to note that mitochondria from S. cerevisiue, in contrast to those of for example bovine heart, apparently retain the right-side-out orientation after sonication (see also [32]). This is concluded from the finding (see Table 3) that the glycerol-3-phosphate oxidase activity, catalyzed by the glycerol-3-phosphate dehydrogenase, located on the outer face of the mitochondrial inner membrane [I], is the same before and after sonication. A similar observation has been made by Schatz and Racker [33] but was not interpreted as such. Sonication of mitochondria leads to partial disruption

383

of the outer membrane. In addition, (most of) the enzymes of the Krebs cycle are released (data not shown). Hypotonic treatment has the same effect as sonication (results not shown).

There are a few reports in the literature dealing with the purification of rotenone-insensitive NADH dehydrogenases from S. cerevisiae and C. utilis [34-471. According to [35] and [37] these two species of yeast possess an external NADH dehydrogenase with a high molecular mass (1.2 - 1.5 MDa), containing FMN as prosthetic group but, unlike complex I, consisting of only two different subunits. In later work [38] these authors state that the enzymes purified in [35] and [37] are not the external NADH dehydrogenase but the cytosolic fatty acid synthetase. We have followed the purification procedure outlined in [34] (see also [39]) and could confirm the properties of the purified enzyme as described. This enzyme does not show NADH : Q6 oxidoreductase activity (results not shown), but only ‘diaphorase’ activity. Mackler et al. [36] have recently reported the purification of the internal NADH dehydrogenase from C. utilis. No experimental data were shown, however, with respect to the localization of the enzyme purified. Nevertheless, this enzyme is similar in some respects to the one purified in this work, i.e. it contains FAD as prosthetic group, reacts with dichloroindophenol, ferricyanide and Q1, is not inhibited by rotenone and is inactive towards NADPH and lipoic acid. The estimated molecular mass only 37 kDa and the K, for NADH equals 16 pM. These latter differences may be due to difference in species of origin.

Mitochondria from plants possess a rotenone-insensitive NADH dehydrogenase that has been partially purified [40, 411. The properties of this enzyme resemble those of the enzyme described in this work. Similarly, the NADH dehydrogenases from Escherichia coli and Bacillus subtilis [42,43] may be related to the one described here. A comparable enzyme has not been found in mammalian mitochondria or, more generally, in mitochondria of multicellular organisms, except in plants. What then is the physiological importance of the NADH : Q6 oxidoreductase in yeast? Clearly, since the enzyme is responsible for the oxidation of cytosolic NADH via the respiratory chain, the ATP formed may be used for all kinds of energy-requiring reactions. Further, the enzyme may play a role in the regulation of the cytoplasmic NADiNADH redox balance [44]. In mammalian cells shuttle systems play an important role in the regulation of the intracellular redox balance (cf. [45]), but the significance of similar systems in S. cerevisiue is presently unknown (cf. [44]). The role of the external NADH dehydrogenase is perhaps best illustrated by consideration of the following. In S. cerevisiae growing aerobically on glucose (NADH dehydrogenase repressed), NADH is formed by glycolysis and in the net formation of biomass, since glucose is more ‘reduced’ than biomass [44]. Excess NADH is removed by forming ethanol and glycerol as end products [44,46] (and references therein). During aerobic growth on galactose (NADH dehydrogenase partially derepressed (cf. [47]) less ethanol and less glycerol was formed compared to growth on glucose, i.e. excess cytosolic NADH may now be rapidly oxidized by the NADH dehydrogenase via the respiratory chain. In addition, the mitochondria1 and cytosolic glycerol-3-phosphate dehydrogenases [48] may also contribute to the removal of cytosolic NADH, in the same way as occurs in the glycerolphosphate shuttle present in mammalian cells.

The data presented in [l] and [16] imply the presence of a rotenone-insensitive, ‘nonphosphorylating’ internal NADH dehydrogenase in mitochondria of S. cerevisiae. The existence

of such an enzyme is the logical consequence of the fact that the Krebs cycle is active in yeast. Nevertheless, we have, as yet, not been able to produce new and more direct evidence for the existence of an internal NADH dehydrogenase for the following reasons. Mitochondria retain their orientation after sonication so that only the properties of the external enzyme can be measured in media without detergents; the only com- pound that was found to inhibit the NADH:Q2 oxidoreductase activity substantially, flavone, is known to inhibit other NADH dehydrogenase also [17]; the Triton used in the assay medium might be inhibitory to the internal NADH dehydrogenase (the complex I from beef-heart mitochondria shows no activity in the assay medium used in this study). The finding that the antibodies inhibit to a lesser degree in sonicated mitochondria than in the purified preparations (cf. Table 3) is suggestive for the presence of an internal NADH dehydrogenase which is active under the assay conditions, but other more trivial explanations cannot be ruled out, yet.

In conclusion, we have reported the first successful purification of the external NADH : Qs oxidoreductase from mitochondria of S. cerevisiae. The enzyme belongs to a novel class of NADH dehydrogenases of which the physiological role has yet to be investigated in more detail. Studies are in progress to elucidate this latter point, to characterize the enzyme further and to define the factor(s) involved in the regulation of its expression.

We thank Messrs H. L. Dekker, J. K. P. Post and A. N. van Hoek for their kind assistance with some of the experimental work described in this paper and Drs C. A. M. Marres and R. Benne for critical reading of the manuscript.

REFERENCES 1. von Jagow, G. & Klingenberg, M. (1970) Eur. J . Biochem. 12,

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

3. Ohnishi, T. (1973) Biochim. Biophys. Acta 301, 105-128. 4. Grossman, S., Cobley, J. G. & Singer, T. P. (1974) J . Biol. Chem.

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

6. Katz, R., Kilpatrick, L. &Chance, B. (1971) Eur. J . Biochem. 21,

7. Cobley, J . G., Grossman, S. & Singer, T. P. (1975) J . Biol. Chem.

8. Ohnishi, T. (1972) FEBS Lett. 24, 305-309. 9. Tottmar, S. 0. C. & Ragan, C. I. (1971) Biochem. J . 124, 853-

10. Albracht, S. P. J. & Subramanian, J. (1977) Biochim. Biophys.

11. Ohnishi, T. (1979) in Membrane proteins in energy transduction

12. Beinert, H. & Albracht, S. P. J. (1982) Biochim. Biophys. Acta

13. Isa, W., Haiker, H. & Weiss, H. (1985) EMBO J . 4, 2075-2080. 14. Moore, A. L. & Rich, P. R. (1980) Trends Biochem. Sci. 5, 284-

15. Palmer, J . M. & Moller, I. M. (1982) Trends Biochem. Sci. 7,

16. Ohnishi, T., Kawaguchi, K. & Hagihara, B. (1966) J . Bid. Chem.

17. Singer, T. P. (1979) Methods Enzymol. 55, 457-458. 18. Mason, T. L., Poyton, R. O., Wharton, D. C. & Schatz, G. (1973)

19. Siegel, L. M. (1978) Methods Enzymol. 53, 419-429. 20. van Hoek, A. N., van Gaalen, M. C. M., de Vries, S. & Berden,

J. A. (1987) Biochim. Biophys. Acta 892,152-161.

583 - 592.

352.

249, 3819-3826.

88-91.

301 - 307.

250, 211-217.

865.

Acta 462, 36-48.

(Capaldi, R. A., ed.) pp. 1-87, Marcel Dekker, New York.

683,245 - 277.

288.

258 - 261.

241, 1797 - 1806.

J . Biol. Chern. 248, 1346-1354.

21. Hum, B. A. L. & Chantler, S. M. (1980) Methods Enzymol. 70,

22. Galfre, G. & Milstein, C. (1981) Methods Enzymol. 73, 1-46. 23. Ryrie, I. J. & Gallagher, A. (1979) Biochim. Biophys. Acta 545,

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

25. Veillon, C. & Vallee, B. L. (1978) Methods Enzymol. 54, 446- 484.

26. Engel, W. D., Schagger, H. & von Jagow, G. (1979) in Metalloproteins (Weser, U., ed.) pp. 185- 193, Thieme, Stuttgart.

27. Weiss, H. & Wingfield, P. (1979) Eur. J . Biochem. 99, 151 - 160. 28. von Jagow, G., Engel, W. D., Schagger, H. & Becker, W. F. (1982)

in Function of quinones in energy-conserving systems (Trumpower, B. L., ed.) pp. 351 -364, Academic Press, New York.

104- 142.

1-14.

29. Ross, E. & Schatz, G. (1976) J. Biol. Chem. 251, 1991 - 1996. 30. Schwitzguebel, J.-P. & Palmer, J. M. (1982) J. Bacteriol. 149,

31. Daum, G., Bohni, P. C. & Schatz, G. (1982) J. Biol. Chem. 257,

32. de Vries, S., van Hoek, A. N., ten Bookum, A. & Berden, J. A. (1987) in Cytochrome systems: molecular biology and bioener- getics (Papa, S., Chance, B. & Ernster, L., eds) pp. 517- 522.

33. Schatz, G. & Racker, E. (1966) Biochem. Biophys. Res. Commun.

34. Duncan, H. M. & Mackler, B. (1966) Biochemistry 5,45-50.

61 2 - 619.

13028 - 13033.

22, 579 - 584.

35. Mackler, B., Haynes, B., Person, R. & Palmer, G. (1980) Biochim. Biophys. Acta 591, 289-297.

36. Mackler, B., Bevan, C., Person, R. &Davis, K. A. (1981) Biochem. Int. 3, 9- 17.

37. Davis, K. A., Mackler, B., Haynes, B., Person, R., Bevan, C., Grace, R., Palmer, G. & Carter, K. (1982) Biochem. Int . 5 , 49 - 54.

38. Evans, T. C., Mackler, B. &Grace, R. (1985) Arch. Biochern. 243, 492 - 503.

39. Singer, T. P. & Edmundson, D. E. (1978) Methods Enzymol. 53,

40. Cook, N. D. & Cammack, R. (1984) Eur. J . Biochem. 141,573-

41. Cook, N. D. & Cammack, R. (1985) Biochim. Biophys. Acta 827,

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

43. Bergsma, J., van Dongen, M. B. M. & Konings, W. N. (1982)

44. van Dyken, J. P. & Scheffers, W. A. (1986) FEMS Microbiol.

45. Dawson, A. G. (1979) Trends Biochem. Sci. 4 , 171 - 176. 46. Fraenkel, D. G. (1982) in The molecular biology of the yeast

Saccharomyces (Strathern, J. N., Jones, E. W. & Broach, J. R., eds) vol. 1, pp. 1-37, Cold Spring Harbor Press, New York.

47. Polakis, E. S., Bartley, W. & Meek, G. A. (1965) Biochem. J . 97,

48. Sprague, G. F. Jr & Cronan, J. E. (1977) J . Bacteriof. 129,1335-

397 - 41 8.

577.

30-35.

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

Eur. J . Biochem. 128, 151 - 157.

Rev. 32, 199 - 224.

298 - 302.

1342.

purification and characterization of a rotenone-insensitive nadh: q6 oxidoreductase from...

Documents