Biochem. J. (1991) 273, 719-724 (Printed in Great Britain)

Topography of succinate dehydrogenase in the mitochondrial innermembraneA study using limited proteolysis and immunoblotting

George H. D. CLARKSON, James NEAGLE and J. Gordon LINDSAY*Department of Biochemistry, University of Glasgow, Glasgow G12 8QQ, Scotland, U.K.

The arrangement of the large (70 000-M,) and small (30 000-Mr) subunits of succinate dehydrogenase in the mitochondrialinner membrane was investigated by immunoblot analysis of bovine heart mitochondria (right-side-out, outer membranedisrupted) or submitochondrial particles (inside-out) that had been subjected to surface-specific proteolysis. Both subunitswere resistant to proteinase treatment provided that the integrity of the inner membrane was preserved, suggesting thatneither subunit is exposed at the cytoplasmic surface of the membrane. The bulk of the small subunit appears to protrudeinto the matrix compartment, since the 30000-Mr polypeptide is degraded extensively during limited proteolysis ofsubmitochondrial particles without the appearance of an immunologically reactive membrane-associated fragment;moreover, a soluble 27000-Mr peptide derived from this subunit is observed transiently on incubation with trypsin.Similar data obtained from the large subunit suggest that this polypeptide interacts with the matrix side of the innermembrane via two distinct domains; these are detected as stable membrane-associated fragments of 32000 Mr and27000 Mr after treatment of submitochondrial particles with papain or proteinase K, although the 27000-Mr fragmentcan be degraded further to low-Mr peptides with trypsin or a-chymotrypsin. A stable 32000 34000-Mr fragment isgenerated by a variety of specific and non-specific proteinases, indicating that it may be embedded largely within the lipidbilayer, or is inaccessible to proteolytic attack owing to its proximity to the surface of the intact membrane, possiblyinteracting with the hydrophobic membrane anchoring polypeptides of the succinate: ubiquinone reductase complex.

INTRODUCTION

The simplest mitochondrial component that catalyses theconversion of succinate to fumarate is succinate dehydrogenase(SDH, EC 1.3.99. 1), an active fragment of the succinate:ubiquinone reductase complex (complex II) of the mitochondrialrespiratory chain isolated by treatment of intact complex II withchaotropic agents.

Bovine heart SDH consists of two non-identical hydrophilicsubunits: a small 30000-Mr iron-sulphur protein (S) and a larger70000-Mr polypeptide (L) that contains one molecule of co-valently attached FAD (Walker & Singer, 1970; Davis & Hatefi,1971). Amino acid sequencing of the bovine heart small subunit(Yao et al., 1986) in conjunction with similar information on thestructural genes for SDH from Escherichia coli (Wood et al.,1984; Darlison & Guest, 1984) and Bacillus subtilis (Phillipset al., 1987), deduced from the DNA sequence of the clonedgenes, has revealed the presence of three conserved cysteinearrangements capable of accommodating all three distinct iron-sulphur centres on the small subunit.SDH as isolated requires the employment of artificial electron

acceptors such as ferricyanide or dichlorophenol-indophenol formeasurement of activity. It utilizes physiological electron accep-tors, ubiquinone or menaquinone, only as part of the nativesuccinate: ubiquinone reductase complex, in which it is bound totwo small hydrophobic polypeptides with Mr values of 15 500and 13 500 in bovine heart (Hatefi & Galante, 1980). Complex IIpreparations also contain approx. 1 mol of b-type haem boundto these latter proteins. The equivalent hydrophobic subunits ofthe bacterial succinate: ubiquinone reductase complexes fromE. coli (Wood et al., 1984) and B. subtilis (Magnusson et al., 1986)have been cloned and sequenced as well as the corresponding

components of the analogous fumarate reductase complex fromE. coli (Cole, 1982; Cole et al., 1982; Grundstrom & Jaurin,1982). Examination of the predicted amino acid sequences ofthese two polypeptides indicates the presence of three hydro-phobic segments in each, long enough to form transmembranea-helices with these integral proteins serving as anchoring sitesfor the catalytic subunits.SDH appears to be exposed exclusively on the matrix side of

the mitochondrial inner membrane. Such an arrangement isinferred from the facts that antimycin-insensitive succinate:ferricyanide reductase activity, inhibition of enzymic activityby anti-SDH antibodies and labelling ofSDH by the membrane-impermeant probe diazobenzene[35S]sulphonate are observedonly in submitochondrial particles, where the matrix side of theinner membrane is accessible to exogenously added reagents(Klingenberg & Buchholz, 1970; Merli et al., 1979).

In the present study we have investigated the topographicalarrangement of the large and small subunits of SDH in themitochondrial inner membrane in greater detail, employingimmune-replica analysis in combination with surface-specificproteolysis to detect membrane-associated or released proteolyticfragments of the two subunits.

MATERIALS AND METHODS

MaterialsLow-Mr marker proteins were obtained from Pharmacia.

Nitrocellulose paper (0.45 ,um pore size) was from Schleicher undSchiill, Dassel, Germany.

Phenylmethanesulphonyl fluoride, leupeptin (acetyl-L-leucyl-L-leucyl-L-argininal), Protein A, papain (type III), tosyl-L-lysyl-

Abbreviations used: SDH, succinate dehydrogenase; SMP, submitochondrial particle.* To whom correspondence should be addressed.

Vol. 273

719

G. H. D. Clarkson, J. Neagle and J. G. Lindsay

chloromethane ('TLCK ')-treated a-chymotrypsin (type VII) andtosyl-L-phenylalanylchloromethane ('TPCK ')-treated trypsin(type XIII) were obtained from Sigma Chemical Co. Proteinase Kwas from Boehringer Mannheim.

lodogen was purchased from Pierce and Warriner, Chester,U.K., Na1251 (carrier-free) was obtained from Amersham In-ternational, and X-ray film (X-Omat S or XAR-5) was fromKodak. All other reagents were of the highest grade availablecommercially.

Preparation of antiseraBovine heart SDH was purified as described by Clarkson et al.

(1987). Antisera to the purified holoenzyme and its constituentlarge and small subunits (anti-L and anti-S sera respectively)were raised by using previously described regimes (De Marcucciet al., 1985; Hunter & Lindsay, 1986).

Isolation of bovine heart mitochondriaBovine hearts were obtained from a local abattoir, transported

to the laboratory on ice, and used within 2 h of animal death.Mitochondria were isolated by the method of Smith (1967), withthe following modifications. Tris base (2 M, 3 ml/300 g of startingmaterial) was added to the suspension of minced tissue, whichwas then blended for 15 s in a Philips food blender. The materialwas reblended for 5 s after a second identical addition of Trisbase, and, after adjustment of the pH to 7.8, the procedure wascontinued as described previously (Smith, 1967). Mitochondriawere either suspended at a protein concentration of 4 mg/ml andused within 24 h of isolation or stored in pellet form at -80 °Cuntil required.

Preparation of submitochondrial particles (SMPs)Mitochondria were thawed from -80 °C, resuspended in

50 mM-sodium phosphate, pH 7.5 (phosphate buffer), to approx.10 mg/ml, and repelleted by centrifugation for 15 min at 6500 gat 4 'C. All further manipulations were performed at 4 'C. Theresuspension/centrifugation step was repeated once, and theresulting pellet was resuspended in phosphate buffer at approx.10 mg/ml.To generate SMPs, the mitochondrial suspension was subjected

to three 15 s bursts of sonication, with 15 s cooling intervals, withan MSE sonicator at amplitude setting 3 (high power). Unbrokenmitochondria were removed by centrifugation at 10000g for10 min; then the supernatant fraction was centrifuged at 100000 gfor 1 h.The final pellet was resuspended in phosphate buffer at 4 mg

of protein/ml concentration and stored in 1.0 ml portions at- 80 °C.

Protease treatment of mitochondria and SMPsPortions of a suspension of freshly isolated mitochondria were

incubated for 60 min at 30 °C with a-chymotrypsin or trypsin ata proteinase/mitochondrial protein ratio of4% (w/w). A controlincubation (no proteinase added) was performed in parallel.Samples (1 mg of mitochondrial protein) were taken at 0, 15, 30and 60 min, and these were dissociated in an equal volume of hot(70 °C) Laemmli sample buffer [62.5 mM-Tris/HCI, pH 6.8, con-taining 2% (w/v) SDS, 10% (w/v) sucrose, 10 mM-dithiothreitoland 0.001 % (w/v) Pyronin Yl. The dissociated samples wereboiled immediately for 5 min before analysis by SDS/PAGE andimmunoblotting. An identical protocol was followed for pro-teinase treatment of freeze-thawed mitochondria.SMPs were subjected to incubation with a-chymotrypsin,

trypsin, papain or proteinase K. Incubations were performed at30 °C for periods of up to 90 min. Specific conditions for eachexperiment are given in the appropriate Figure legend.

In some cases, proteolysis was terminated by mixing sampleswith 0.33 vol. of 2 x Laemmli sample buffer containing 0.8 mgof leupeptin/ml (for inactivation of papain) or 2 mM-phenyl-methanesulphonyl fluoride (for inactivation of a-chymotrypsin,trypsin or proteinase K). After 5 min at room temperature,samples were boiled for 5 min before analysis by SDS/PAGEand immunoblotting. Alternatively, samples of proteinase-treated SMPs were diluted with 3 vol. of ice-cold 20 mM-phos-phate buffer containing 0.15 M-NaCl and 0.8 mg of leupeptin/mlor 2 mM-phenylmethanesulphonyl fluoride. Subsequent mani-pulations were performed at 4 'C. The samples were centrifugedat 27000 g for 20 min, then the pelleted membranes were eachresuspended in 0.5 ml of the same buffer and re-centrifuged asabove. The wash procedure was repeated once with buffer, oncewith distilled water, and the final pellets were dissociated inLaemmli sample buffer and boiled for 5 min.

Other methodsProcedures for SDS/PAGE and immunoblotting were de-

scribed previously (Clarkson et al., 1987, and references citedtherein).

Protein concentrations were determined by a modification ofthe Lowry procedure, with BSA as a standard (Markwell et al.,1978). 125I-labelled Protein A and low-M, markers were preparedby the method of Salacinski et al. (1981), as described by DeMarcucci et al. (1985).

RESULTS

It has been shown previously that in preparations of freshlyisolated bovine heart mitochondria the cytoplasmic surface ofthe inner membrane is exposed to labelling agents such aslactoperoxidase, indicating that the outer membrane has beendisrupted by the isolation procedure (Boxer, 1975). A highdegree of removal of the outer membrane is also indicated byabsence of monoamine oxidase activity, a marker for the outermembrane. Therefore, by immunoblot analysis of proteinase-treated fresh mitochondria, it was possible to probe for exposedregions of subunits of SDH on the cytoplasmic surface of theinner membrane.

Samples of freshly isolated mitochondria were incubated in theabsence of proteinase or after addition of 40% (w/w) oc-chymo-trypsin or trypsin. The immunoblot obtained by probing thesesamples with anti-SDH serum (Fig. la) shows that both subunitsof SDH were essentially stable during a control incubation [(i),lanes 2-5]. Inclusion of a-chymotrypsin [(ii), lanes 6-9] or trypsin[(iii), lanes 10-13] caused no significant change in the stability ofthe 70000-Mr and 30000-Mr polypeptides. Similar results wereobtained with proteinase K and Pronase (results not shown).To test the possibility that SDH in its membrane environment

is inherently resistant to proteolysis under the conditions de-scribed above, the experiment was repeated with freeze-thawedmitochondria, where fracturing of the inner membrane hasoccurred (Pettit et al., 1978). Both polypeptides were again stableduring a control incubation [Fig. l(b), (i), lanes 2-5], but wererendered proteinase-sensitive (at least partially) by the freeze-thaw procedure. In the presence of ac-chymotrypsin, the largesubunit was degraded with the concomitant appearance of a34000-Mr proteolytic fragment [(ii), lanes 6-9] and a weaklyreacting species that migrated with an Mr value of 17000. Incontrast, there was no discernible decrease in the amount ofintact small subunit.

During incubation of freeze-thawed mitochondria with trypsin[(iii), lanes 10-13] degradation of both subunits was paralleled bythe appearance of large fragments of Mr 32000, as well as a

transient 27000-Mr immunoreactive polypeptide. The stable

1991

720

Intramembrane organization of mitochondrial succinate dehydrogenase

(a)

Time (min)10-3XM, F97-67-

45-

21-

Time (min10-3 X M,

97-67

45-

30-

21 -

(a)(1 ) (ii) (iii)

o 15TO60'0 1l5 30 60 0 15 30 60 10-3 X M

97-

67-

45-L

30-

21-30 - w_ _ _-_ s

1 2 3 4 5 6 7 8 9 10 1112 13 14

(b)

(i) (ii) (iii)) 0 15 3060 0153060 '0 1530 60

iS F~~~~~~~i

..~~~~~~~~~~a

1 2 3 4 5 6 7 8 9 101112 13 14

10-3 x M,

97

67-

45--L

30-

-S

Fig. 1. Imnmunoblot analysis of the effect of proteinases on SDH in bovineheart mitochondria

Portions of freshly isolated (a) or freeze-thawed (b) mitochondriawere incubated for 60 min at 30 °C with a-chymotrypsin or trypsin[4% (w/w) proteinase/mitochondrial protein] or with no proteinaseaddition (control incubation). Samples were removed from eachincubation at the indicated times, mixed with 0.33 vol. of 2 xLaemmli sample buffer and subjected to electrophoresis on 100%(w/v) SDS/polyacrylamide-slab gels. Resolved polypeptides wereelectrophoresed on to nitrocellulose paper for detection of immuno-reactive species with anti-SDH serum and '25I-labelled Protein A.Lanes 1 and 14, "251-labelled low-Mr markers; (i) lanes 2-5, controlsamples; (ii) lanes 6-9, incubation with oc-chymotrypsin; (iii) lanes10-13, incubation with trypsin. Volumes corresponding to 40 ,ug oftotal mitochondrial protein were loaded in lanes 2-13 inclusive.

trypsin-derived 32000-Mr species appeared to be less antigenicthan the corresponding fragment, Mr 34000, observed afterincubation with chymotrypsin, although this was rather variablefrom experiment to experiment. The 27000-Mr intermediate wasshown subsequently (Fig. 2b) to be a product of the smallsubunit, which was no longer associated with the inner mem-brane. Taken together, the above data suggest that both subunitsof SDH are exposed exclusively on the matrix side of the innermembrane.

Bovine heart SMPs generated by sonication of mitochondriaare predominantly vesicles of inner membrane with an invertedorientation (Godinot & Gautheron, 1979; Smith & Ragan,1980). Therefore the exposure of SDH on the matrix side of theinner membrane was investigated by immunoblot analysis ofcontrol (i) or proteinase-treated [(ii) and (iii)] SMPs (Fig. 2) witheither anti-L serum (Fig. 2a) or anti-S serum (Fig. 2b). Resultssimilar to those observed in freeze-thawed mitochondria were

obtained, indicating that disruption of the inner membrane bysonication in the preparation of SMPs does not radically alter

21 ----

721

(i) (ii) (iii)I0 1 5 30 6d0 0 15 30 6 0 1 5 30 60' Time (min)

(b) 0) (ii) (iii)I0 15 30 6d 0 15 30 60 o0 15 30 60 X Time (min)

-s

Fig. 2. Immunoblot analysis of the effect of proteinases on the large andsmall subunits of SDH in bovine heart SMPs

SMPs were incubated for up to 60 min in the absence of proteinaseor in the presence of a-chymotrypsin or trypsin as described in Fig.1 legend. Samples were processed in an identical manner for detectionof immunologically reactive fragments with subunit-specific anti-[large subunit (L)] (a) or anti-[small subunit (S)] (b) sera. Lanes 1 and14, '25l-labelled low-Mr markers; (i) lanes 2-5, control incubations;(ii) lanes 6-9, digestion with a-chymotrypsin; (iii) lanes 10-13digestion with trypsin.

the accessibility of SDH to surface-specific proteolysis. Degrad-ation of the large subunit with either a-chymotrypsin (ii) or

trypsin (III) yields stable immunologically reactive 34000-M,and 32000-Mr species respectively, but rapid degradation of thesmall subunit once again is achieved only with trypsin, yieldinga transient 27000-Mr polypeptide. The minor chymotrypticproduct of the large subunit, Mr 17000, is again also justdetectable.

Fig. 3 illustrates the immunoblot patterns obtained whencontrol or proteinase-treated SMPs were probed with anti-Lserum (Fig. 3a) or anti-S serum (Fig. 3b) after being washed with0.15 M-NaCl to test the interaction of these proteolytic fragmentswith the mitochondrial inner membrane. In the absence ofproteinase, both subunits were stable [Fig. 3(a) and 3(b), lanes 2and 3, (i)], whereas in the presence of a-chymotrypsin [Fig. 3(a),(ii), lanes 4-7] or trypsin [Fig. 3(a), (iii), lanes 8-11] the largesubunit was degraded with the parallel appearance of a mem-brane-bound fragment of Mr approx. 32000-34000. Incubationwith papain generated two large subunit fragments; one exhibiteda similar electrophoretic mobility to the a.-chymotryptic andtryptic fragments, and the second had an Mr value of approx.27000 [Fig. 3(a), (iv), lanes 12-15]. Long-term incubation withpapain [Fig. 3(a), (iv), lane 15] appeared to affect the integrity ofSMPs, which tended not to pellet so readily on centrifugation.

Vol. 273

--~ ~~~~~~~~~~~~~.^ - -7

1 2 3 4 5 6 7 8 9 10 111213 14

*

*^ 4

1 2 3 4 5 6 7 8 9 10 11 121314

n

- I

G. H. D. Clarkson, J. Neagle and J. G. Lindsay

(a) (a)

Time (min)

10-3 x M,

67-45

30-

21-

(i) (ii) (iii) (iv)IooX091 30 609011 03060 901030 60901

1 2 3 4 5 6 7 8 9 10 1112 131415

(b)

Time (min)

10-3 X M,

97-67-45-

30-

21 -

Mi (ii)-- -(iii) (iv)-

[09011030609011O3060 9010306090

2 34567 891 _

..

Time (min)

10-3 X M, I

L67 -

45-

30-

L

(i) (ii)ro 30'| 5 10 20 301

2 3 4 5 6 7

(b)

- S

Fig. 3. Inmunoblot analysis of salt-washed proteinase-treated submito-chondrial particles with anti-Ilarge subunit (L)l and anti-Ismalisubunit (S)l sera

Submitochondrial particles were incubated at 30 °C in the absenceofproteinase, or after addition ofa-chymotrypsin, trypsin or papain[40% (w/w) protease/particle protein]. In each case, samples were

removed at the indicated times, and mixed with 3 vol. of ice-coldphosphate buffer containing 0.15 M-NaCl and 2 mM-phenyl-methanesulphonyl fluoride or 0.8 mg of leupeptin/ml. Samples were

processed as described in the Materials and methods section, resolvedby electrophoresis on 15% (w/v) SDS/polyacrylamide slab gels,and analysed by immunoblotting with anti-[large subunit (L)J (a) oranti-[small subunit (S)] (b) serum. (i) Lanes 2 and 3, control samples;(ii) lanes 4-7, digestion with a-chymotrypsin; (I1I) lanes 8-11,digestion with trypsin; (iv) lanes 12-15, digestion with papain.Lanes I (a and b) and 16 (b) contain "25I-labelled low-Mr markers.

Fig. 4. Immunoblot analysis of proteinase K- or papain and a-chymo-trypsin-treated submitochondrial particles with anti-Ilargesubunit (L)l serum

SMPs were incubated for up to 30 min at 30 °C in the presence ofproteinase K (a) or papain (b) [4% (w/w) proteinase/particleprotein]. In (a), samples were taken at the indicated times, mixedwith 0.33 vol. of 2 x Laemmli sample buffer containing 2 mm-phenylmethanesulphonyl fluoride and boiled for 5 min. Sampleswere analysed by electrophoresis on a 10% (w/v) SDS/poly-acrylamide slab gel followed by immunoblotting with anti-[largesubunit (L)] serum. In (b), incubation of SMPs with papain was

terminated by the addition of a 5-fold excess of leupeptin (by weight)before further incubation with 4% (w/w) a-chymotrypsin. Sampleswere then processed as described above. (a) Lane 1, "25I-labelledlow-Mr markers; (i) lanes 2 and 3, control samples; (ii) lanes 4-7,digestion with proteinase K for the indicated times. Equal amountsof protein were loaded in lanes 2-7 inclusive. (b) Lanes I and 2,control incubation, 0 and 60 min; lane 3, 30 min digestion withpapain showing 32000-Mr and 27000-Mr species; lane 4, as lane 3plus extra 20 min after papain inactivation; lane 5, as lane 4 plus10 min incubation with a-chymotrypsin; lane 6, as lane 3 plus10 min incubation after a second addition of a-chymotrypsin.

Treatment of SMPs with a-chymotrypsin had no discernibleeffect on the small subunit [Fig. 3(b), (ii), lanes 4-7]. In contrast,both trypsin and papain caused extensive degradation of the30000-M, polypeptide [Fig. 3(b), (iii) and (iv), lanes 8-15].However, no membrane-associated fragment of the small subunitwas detected in the trypsin- or papain-treated samples, indicatingthat the 27000-Mr product of tryptic digestion observed pre-

viously was not bound to the inner membrane.Proteinase K treatment of SMPs also produced a large-subunit

fragment pattern similar to that produced by papain (Fig. 4a).Essentially complete digestion of the intact large subunit was

achieved during 5 min of incubation with proteinase K; this wasaccompanied by the appearance of large-sublinit fragments ofMr 32000 and 27000 respectively [Fig. 4(a), (ii), lane 4]. Solubleimmunologically reactive material corresponding to the residual11 000-Mr region of this polypeptide cleaved from the intact largesubunit could be detected at the dye front. During the remainderof a 30 min incubation the ratio of the two fragments remainedalmost constant [Fig. 4(a), (ii), lanes 5-7]. The 32000-Mr and27000-Mr species were subsequently shown to remain membrane-

associated when proteinase treatment was followed by saltwashing (results not shown).No precursor-product relationship could be observed between

the 32000-Mr and 27000-Mr fragments in a series of time courses

designed to produce controlled degradation of the large subunitwith papain or proteinase K. The separate identity of these twomembrane-associated fragments was confirmed in double-diges-tion experiments (Fig. 4b) in which papain-treated SMPs were

incubated subsequently with a-chymotrypsin (or trypsin; resultsnot shown). This led to the rapid disappearance of the 27000-Mrpolypeptide (lanes 5 and 6), whereas the 32000-Mr componentwas essentially unaffected by this treatment. Incomplete recoveryof samples was often observed after washing and re-centri-

fugation of SMPs subjected to double proteinase digestions,indicating that significant disruption of the membrane vesiclescan occur under these conditions.The stability of the principal 32000 34000-Mr fragment of the

large subunit (L) may result from its protection by the lipid

1991

J .~~~~~~~0 I'

......~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..

Z _o.~~~~~~~~~~~~~~~~.: ..

722

Intramembrane organization of mitochondrial succinate dehydrogenase

10 xM,

67-

45

30-

21-

-L

1 2 3 4 5 6 7 8 9 10 11 12

Fig. 5. Immunoblot analysis of the effect of non-ionic detergent on thestability of the major proteolytic fragment of the large subunit inbovine heart SMPs

SMPs were incubated at 30 °C for 30 or 60 min in the presence oftrypsin or a-chymotrypsin as described in Fig. 1 legend. At thispoint, Triton X-100 was added to some samples (final concentrationof 0.1 0%, v/v) and incubations were continued for a further 30 min.All samples were then subjected to SDS/PAGE and immuno replicaanalysis with anti-[large subunit (L)J serum. Lanes 1 and 2, controlincubations minus proteinase for 0 and 60 min respectively; lane 3,blank; lane 4, as lane 2 except that 0.1% (v/v) Triton X-100 wasadded at 30 min; lanes 5 and 6, incubations for 60 and 30 min withtrypsin; lanes 7 and 8, as lanes 5 and 6 except that 0.1 % (v/v)Triton X- 100 was added 30 min before termination ofthe incubation;lanes 9 and 10, as lanes 5 and 6 except incubated with a-chymotrypsin; lanes 11 and 12, as lanes 8 and 9 except incubatedwith a-chymotrypsin.

bilayer or alternatively relate to its organization into a compactdomain structure that is inherently resistant to further proteolysis.To test this latter possibility, the proteolytic sensitivity of the32000-34000-Mr fragment was tested after solubilization of theinner membrane in 0.1 % (v/v) Triton X-100. As illustrated inFig. 5, the major tryptic (lanes 5 and 6) and chymotrypticproducts (lanes 9 and 10) of the large subunit were both degradedrapidly to smaller non-immunogenic fragments (lanes 7 and 8and lanes 11 and 12 respectively) after solubilization of thesubmitochondrial particles in non-ionic detergent. Thus a majorsegment of the large subunit is apparently protected fromproteolytic attack by virtue of the intactness of the mitochondrialinner membrane.

DISCUSSION

In this investigation we have confirmed that the large andsmall subunits of SDH are protected from proteolysis in intactmitochondria, in agreement with previous ferrocyanide inter-action, antibody binding and chemical labelling studies (Kling-enberg & Buchholz, 1970; Merli et al., 1979). Both polypeptidesare sensitive to digestion by a variety of proteinases in 'inside-out' SMPs. The absence of a small-subunit fragment in pro-teinase-treated salt-washed SMPs suggests that most of thissubunit is released from the membrane under these conditions.An arrangement of the above type is further suggested by the

following lines of evidence. Firstly, we were able to detect a

soluble 27000-Mr species corresponding to the bulk of the intactsmall subunit among total tryptic digestion products. Secondly,a hydropathy plot of the amino acid sequences of the smallsubunit (252 residues) (Yao et al., 1986) indicates that it isgenerally very hydrophilic. A small hydrophobic region con-

taining several cysteine residues is located between leucine-155and proline-168 and may provide a non-aqueous pocket for one

Vol. 273

of the iron-sulphur centres. Interestingly the C-terminus ishydrophilic in character while the N-terminus consists of a

stretch of nine uncharged residues. In light of this evidence, it ispossible that the small subunit is associated with the innermembrane via a short N-terminal membrane anchor. A numberof potential tryptic-cleavage sites are located in the appropriateregion near to the N-terminus of this polypeptide.Each of the four proteinases used in this study produced a

stable large-subunit fragment of Mr value 32000-34000. In eachcase this species was shown to be tightly bound to the innermembrane, since it was not removed by repeated salt washes.The generation of a common fragment, even by low-specificityproteinases, suggests the presence of an extremely stable mem-brane-associated domain in the large subunit. As this fragment isrendered susceptible to proteolysis by solubilization of the innermembrane in non-ionic detergents, it is clear that its stability isthe result of its close association with an intact lipid bilayer. It ispossible that this segment of the large subunit is largely embeddedwithin the lipid bilayer, although a strong interaction with thesmall hydrophobic anchoring proteins of the native suc-cinate:ubiquinone reductase complex at the membrane surfacemay be responsible for shielding it from exogenous proteinases.In this regard, there is no evidence for the presence of extensivehydrophobic stretches in the equivalent bacterial subunit (Woodet al., 1984; Phillips et al., 1987), which is similar in size to itsmammalian counterpart.The reproducible appearance of an additional membrane-

associated fragment of the large subunit (27000 Mr) in samplesof papain- and proteinase K-treated SMPs has two possibleexplanations. The smaller fragment may be derived from thelarger by removal of an additional 5000-Mr fragment from the32000-Mr membrane-associated domain. Alternatively, it mayrepresent a separate membrane-associated region of the largesubunit. In this case, digestion with trypsin or a-chymotrypsingenerates additional large-subunit fragments of low M, value(Fig. lb) that are released as soluble polypeptides and are onlyweakly immunogenic (a-chymotrypsin) or non-immunogenic(trypsin). We favour the latter arrangement for two main reasons.Firstly, since the appearance ofthe 27000-Mr species in proteinaseK-treated SMPs is rapid, subsequent conversion of the 32 000-Mrspecies into its 27000-Mr form should also be rapid if the firstexplanation is correct: such an effect was not observed. Asimilar precursor-product relationship was not evident eitherduring incubation of SMPs with papain. Secondly, double-digestion experiments, in which papain treatment of SMPs wasfollowed by treatment with trypsin or a-chymotrypsin, suggestthat the 27000-Mr fragment is preferentially degraded by theseproteinases. This effect indicates that the 27000-Mr domain,although stable to proteinase K and papain, is held in themembrane by a relatively small (non-immunogenic) anchoringregion (Mr 10000), since it is susceptible to externally addedproteinases of appropriate specificity.

Girdlestone et al. (1981) have reported on the intramembraneorganization of complex II after incorporation into syntheticphospholipid vesicles in the presence of photoreactive radio-labelled (arylazido) phospholipids. They concluded that thesmall subunit was partially inserted into the lipid bilayer whereasthe large subunit was shielded from or raised above the surfaceof the membrane. Therefore the proteinase-resistant regionsof the large subunit may be bound to the membrane viainteractions with the small membrane-anchoring proteins ofcomplex II (Xu et al., 1987) as suggested previously, rather thanthrough direct contact of the large subunit with the hydrophobicinterior of the bilayer. However, comparison of data, obtained

with photoaffinity probes on complex II incorporated intoartificial lipid vesicles, is difficult to reconcile with this immuno-

411 *A 4Ub

723

G. H. D. Clarkson, J. Neagle and J. G. Lindsay

logical study conducted on intact membranes, which is likely.to be of greater physiological relevance. Unfortunately, theamino acid sequence of the large subunit of mammalian SDHis not available as yet to permit our topographical analysisto be compared with its predicted membrane association fromhydropathy profiles.

We thank Dr. A. Phelps for helpful discussions during the course ofthis wtrk. G. H. D. C. was the recipient of a Science and EngineeringResearch Council postgraduate studentship. J. G. L. acknowledges thecontinued financial support of the Science and Engineering ResearchCouncil.

REFERENCES

Boxer, D. H. (1975) FEBS Lett. 59, 149-152Clarkson, G. H. D., King, T. E. & Lindsay, J. G. (1987) Biochem. J. 244,

15-20Cole, S. T. (1982) Eur. J. Biochem. 122, 479-484Cole, S. T., Grundstrom, T., Jaurin, B., Robinson, J. J. & Weiner, J. H.

(1982) Eur. J. Biochem. 126, 211-216Darlison, M. G. & Guest, J. R. (1984) Biochem. J. 223, 507-514Davis, K. A. & Hatefi, Y. (1971) Biochemistry 10, 2509-2516De Marcucci, 0. L., Hunter, A. & Lindsay, J. G. (1985) Biochem. J. 226,

509-517Girdlestone, J., Bisson, R. & Capaldi, R. A. (1981) Biochemistry 20,

152-156

Godinot, C. & Gautheron, D. C. (1979) Methods Enzymol. 55, 112-114Grundstr6m, T. & Jaurin, B. (1982) Proc. Natl. Acad. Sci. U.S.A. 79,

1111-1115Hatefi, Y. & Galante, Y. M. (1980) J. Biol. Chem. 255, 5530-5537Hunter, A. & Lindsay, J. G. (1986) Eur. J. Biochem. 155, 103-109Klingenberg, M. & Buchholz, M. (1970) Eur. J. Biochem.13, 247-252Magnusson, K., Phillips, M. K., Guest, J. R. & Rutberg, L. (1986)

J. Bacteriol. 166, 1067-1071Markwell, M. A. K., Haas, S. M., Beiber, L. L. & Tolbert, N. E. (1978)

Anal. Biochem. 87, 206-210Merli, A., Capaldi, R. A., Ackrell, B. A. C. & Kearney, E. B. (1979)

Biochemistry 18, 1393-1400Pettit, F. H., Yeaman, S. J. & Reed, L. J. (1978) Proc. Natl. Acad. Sci.

U.S.A. 75, 4881-4885Phillips, M. K., Hederstedt, L., Hasnain, S., Rutberg, L. & Guest, J. R.

(1987) J. Bacteriol. 169, 864-873Salacinski, P. R. P., McLean, C., Sykes, J. E. C., Clement-Jones, V. W.& Lowry, P. J. (1981) Anal. Biochem. 117, 136-146

Smith, A. L. (1967) Methods Enzymol. 10, 81-86Smith, S. & Ragan, C. I. (1980) Biochem. J. 185, 315-326Walker, W. H. & Singer, T. P. (1970) J. Biol. Chem. 245, 4224-4225Wood, D., Darlison, M. G., Wilde, R. J. & Guest, J. R. (1984) Biochem.

J. 222, 519-534Xu, J.-X., Yu, L. & Yu, C.-A. (1987) Biochemistry 26, 7674-7679Yao, Y., Wakabayashi, S., Matsubara, H., Yu, L. & Yu, C.-A. (1986) in

Iron-Sulphur Protein Research (Matsubara, H., Katsube, Y. & Waka,K., eds.), pp. 240-244, Japan Scientific Societies Press, Tokyo, andSpringer-Verlag, Berlin

Received 15 February 1990/3 August 1990; accepted 9 August 1990

1991

724