Characterization of the cytochrome d terminal oxidase complex of

THEJOURNAL OF BIOLOGICAL CHEMISTRY (01984 by The American Society of Biological Chemists, Inc

Vol. 259,No. 12, Issue ofJune 25,pp. 7998-8003,19& Printed in U.S.A.

Characterization of the Cytochrome d Terminal Oxidase Complex of Escherichia coli Using Polyclonaland Monoclonal Antibodies* (Received for publication, December 28, 1983)

Robert G. KranzS and RobertB. Gennisg From the Departments of Chemistry and Biochemistry, Uniuersity of lllinois, Urbana, Illinois 61801

The cytochrome d terminal oxidase complex is one (8) fromEscherichia coli. This is one of the two terminal of two terminal oxidases in the aerobic respiratory oxidase complexes in the aerobic respiratory chain of E. coli chain of Escherichia coli. Previous work has shown by (for a review see Ref. 9). The cytochrome d complex has been dodecyl sulfate-polyacrylamide gel electrophoresis purified to homogeneity and shown to contain two subunits that this enzyme contains two subunits (I and 11) and (I and 11) by sodium dodecyl sulfate-polyacrylamide gel electhree cytochrome components, bS6s,a l , and d. Recon- trophoresis (8).The purified enzyme solubilized in detergents stitution studies have demonstrated that the enzyme such as Triton X-100 rapidly oxidizes ubiquinol-1 and also functions as a ubiquinol-8 oxidase and catalyzes an will utilize TMPD’ as a substrate (8). The oxidase has also electrogenic reaction, i.e. turnover is accompanied by a charge separation across the membrane bilayer. In been reconstituted into proteoliposomes using a detergentthis paper,monoclonal and polyclonal antibodieswere dialysis technique and shown to function as aubiquinol-8 used to obtain structural information about the cyto- oxidase (10). Turnover of the oxidase in this artificialsystem chrome d complex. It is shown that antibodies directed generates a transmembrane potential across the bilayer of about 180 mV, negative inside (10). It is, thus, proposed that against subunit I effectively inhibit ubiquinol-1 oxithe cytochrome d complex functions in vivo as a ubiquinol-8: dation by the purified enzyme in detergent, whereas antibodies which bind to subunit I1 have no effect on oxygen oxidoreductase, and is a coupling site in E. coli. In this paper, polyclonal and monoclonal antibodies are quinol oxidation. The oxidation rate of N,N,N’,N’d the tetramethyl-p-phenylenediamine,in contrast, is unaf- used to obtain structural information about cytochrome fected by antisubunit I antibodies, but is inhibited by complex. It is shown that thequinol oxidation site isprobably antibodies against subunit 11. It is concluded that the on subunit I (Mr= 57,000), but that TMPD is oxidized at a quinol oxidation siteis on subunitI, previously shown different site, located on subunit 11. The antibodieswere also to be the cytochrome b66s component of the complex, used to analyze the products of protein cross-linking experiand that N,N,N’,N’-tetramethyl-p-phenylenediamine ments. It is clearly shown that subunitI1 exists as a dimer in oxidation occursat a secondary site on subunit 11. the active, detergent-solubilized form of this enzyme. Finally, The antibodies were alsoused to analyze the results it is shown that lipopolysaccharide is a tightly bound compoof a protein cross-linking experiment. Dimethyl sub- nent of the purified enzyme. erimidate was used to cross-link the subunits of purified, solubilized oxidase. Immunoblot analysis of the EXPERIMENTALPROCEDURES products of this cross-linking clearly indicate that subunit 11 probably exists as a dimer within the complex. Muterials-Ubiquinol-1 was from Hoffmann-LaRoche.Purified Finally, it is shown that the purified enzyme contains lipopolysaccharides were from Sigma. tightly bound lipopolysaccharide. This was revealed Monoclonal Antibody Production and Purificution-Balb/c mice after discovering thatone of the monoclonal antibodies were immunized intraperitoneally with 200 pg of the purified cytoraised against the purified complex is actually directed chrome complex emulsified in complete Freund‘s adjuvant. Mice were againstlipopolysaccharide. The significance of this given a boost of 50 pgof the cytochrome complex in incomplete Freund’s adjuvant 1month later. Splenic lymphocytes were harvested finding is not known.

for cell fusion onthe 4thday following a second booster immunization with 25 pg of the cytochrome. Procedures detailing cell fusion and cloning have been described previously (11).Briefly, an Sp2/0-Ag14 myeloma cell line was fused Polyclonal and monoclonal antibodies canbe used as probes with immune splenocytes using polyethylene glycol. Following hypoto study structural and functional domains of macromolecules xanthine/aminopterin/thymidine selection, hybridomas were cloned (1-3). These include studies of ligand binding sites (4), topo- in 0.2%agar. Hybrids secreting cytochrome d antibodies were propagated in pristane-primed Balb/c mice to induce ascites fluid. Ascites graphical arrangements ( 5 , 6), protein conformational fluid was treated using sodium dextran sulfate and ammonium sulfate changes (7), and primary sequence (1).Polyclonal and mono- precipitation as previously described (12). Antibodies were then puclonal antibodies have been generated to study the antigenic rified over a protein A-Sepharose CL4B column and eluted with 0.1 determinants of the cytochrome d terminal oxidase complex N acetic acid followed by neutralization with 2 mM Tris base (pH 8.8) ~. and overnight dialysis in PBS, pH 7. These antibodies are called * This work was supported by Grant DEAC-02-80ER10682 from purified monoclonal antibodies. Monoclonal antibodies were isotyped using affinity purified rabbit the Department of Energy and Grant HL16101 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must The abbreviations used are: TMPD, N,N,N‘,N’-tetramethyl-ptherefore be hereby marked “aduertisement” in accordance with 18 phenylenediamine; PBS, phosphate-buffered saline (10 mM sodium U.S.C. Section 1734 solely to indicate this fact. phosphate, 0.14 M NaC1, pH 7); SDS-PAGE, sodium dodecyl sulfate$ Present address, Department of Biophysics and Theoretical Bipolyacrylamide gel electrophoresis; mAb, purified monoclonal antiology, University of Chicago, 920 East 58th St., Chicago, IL 60637. body; LPS, lipopolysaccharide. 5 T o whom inquiries should be addressed.

7998

Characterization of Terminal Oxidase Complex Using Antibodies anti-mouse reagents. Monoclonal antibodies were precipitated from culture supernatants with 50% ammonium sulfate and resoluhilized in PBS (pH 7)with 10-fold concentration.Ouchterlony diffusion indicated that mAbs B20-4 and A14-5 were isotype y2a and mAB A16-1 was y l . mAb B3-3 was too diluted to type but was of the y type as determined by protein A binding characteristics. Monoclonal Antibody Screening-Two assays were used to screen for anti-cytochrome d secretinghybrids. A solid phase assayemployed Dynatech Immulon wells coated with 1to 5 pg of purified cytochrome complex. When a solution containing radiolabeled cytochrome (50 pl/well) was incubated in the coated wells for 1 h at 37 "C and then removed, less than 0.1% of the counts adhered to wells. the Therefore, the purified (unlabeled) cytochromecomplex was dried onto thewells. The wells were then incubated with fetal calf serum in PBS (pH 7) containing 0.05% Tween 20 (Buffer A) to block the remaining binding sites. Hybridoma supernatants (100 p1 in Buffer A) were then added to the wells. After incubation for 1 h at 37 "C and overnight a t 4 "C, wells were washed three times with Buffer A. Approximately 200,000 cpm of radioiodinated protein A in Buffer A were incubated for 1 h at 37 "C in each well. Wells were washed three times with Buffer A and counted. Backgrounds variedfrom 1,000 to 10,000 cpm depending on the amount of protein dried on the wells. Positive supernatants ranged from 5,000 to 40,000 cpm above background. Bovine serum albumin (1%)could replace fetal calf serum with little effect on the assay. Liquid phase assays were performed using radioiodinated cytochromed complex. The cytochrome (10 pg) was iodinated with Iodogen and 0.1 mCi of K'T. Radiolabeled cytochrome (100,000 cpm) was incubated with hybridoma supernatants in Buffer A (450 p l ) for 1 h at 37 "C. Rabbit anti-mouse serum (50 pl) and 5 pg of mouse IgG were added and tubes were incubated at 37 "C for 1 h and overnight a t 4 "C. After centrifugation for 3 min in an Eppendorf centrifuge, precipitates were washed two times and counted. Measurements of Antibody Inhibition of Ubiquinol-I and TMPD Oxidase Activities and Antibody Binding Measurements-Purified cytochrome dcomplex (approximately 15 pg) in 100 pl of 1.0% Triton X-100,O.l M sodium phosphate, pH 7 (Buffer B), was incubated with the selected antibody preparation for 1 h at 37 "C. Oxygen uptake measurements were made using a Clark-type oxygen electrode. Oxygen uptake was initiated by adding the oxidase preparation to Buffer B, which was preincubated with either TMPD (2 mM) and dithioerythritol (1 mM) or ubiquinol-1 (0.14 mM) and dithioerythritol (2 mM). To determine the amount of antibody hound to the oxidase, 50 p1 of Pansorbin(protein A-coated Staphlococcus aureus which was washed three times in Buffer B) were added and, after incubation for 1 h at 37 "C, tubes were centrifuged in an Eppendorf centrifuge for 3 min. Ubiquinol-1 oxidase measurements were then performed with the supernatants. All measurements were compared to controls. For rabbit antisera, normal serum was used as a control. For mAbs, a previously described mAb directed against fluorescein (13) was used as a control. Affinity Purification of Polyclonal Antibody against Subunit II of the Cytochrome Complex-The cytochrome complex was covalently attached to a Bio-Rad Affi-Gel 10 column (1 ml) in 4-(2-hydroxyethyl)-1-piperazineethanesulfonicacid buffer according to the manufacturer. After washing with 10% fetal calf serum in PBS (pH 7) and 1 N acetic acid, followed by equilibration in 10% fetal calf serum in PBS (pH7), 50 mg of subunit I1 antiserum preparationwere passed through thecolumn. After washing with30 ml of 10% fetal calf serum in PBS (pH 7) followed by 20 ml of PBS (pH 7), bound antibodies were eluted with 1.0 N acetic acid and neutralized with 2 M Tris base. Antibodies were then dialyzed overnight against PBS (pH 7) and 0.1% azide at 4 "C. Other Methods-Immunoblotting, SDS-PAGE, and protein determinations have been described previously (14). Isoelectric focussing procedures, bacterial strains, and growth conditions have also been described previously (12). Purification of the cytochrome d complex has been described previously (8).Native and subunit antiseratoward cytochrome doxidase were obtained and characterized in earlier work (15).

mined using the same solid phase radioimmunoassay which was employed for initial hybridoma screening (Fig. 1).The relative binding affinity of the four mAbs are A14-5 > B20-5 > B3-3 > A16-1. Immunoblots were carried out to determine whether the four mAbs bind to the native cytochromed complex (i.e. oxidase active) and to determine the subunits they are directedagainst (Fig. 2).Thenative cytochromed complex isoelectrically focussed to a PI of 5.3 (Fig. 2A, Lanes 1 and 2). The complex was shown to have TMPD oxidase activity and was the only protein present when stained with Coomassie brilliant blue (Fig. 2 A , Lanes 1 and 2). This testified to the homogeneity of the cytochrome preparation and showed that the purified cytochrome d complex had the same isoelectric point as the TMPDoxidase in E. coli membranes solubilized in Triton X-100 (12). All four of the mAbs bound to the isoelectrically focussed native cytochrome d oxidasecomplex. An immunoblot using mAb B3-3 isshown in Fig. 2 A , Lane 3. Immunoblots of the purified cytochrome d complex using SDS-PAGE showed that mAb B20-4 bound to a component that migrated with a molecularweight of approximately 6000, estimated from molecular weight standards (Fig. 2B). This component stained very intensely with the silver stain and was extractable with ch1oroform:methanol (1:4). Also, mAb B20-4 was completely adsorbed to intact E. coli (not shown). These are properties expected of lipopolysaccharide (16).This antibody also immunoreacted with purified lipopolysaccharide from E. coli with certain serotypes (not shown). The data show that the purifiedcytochrome contains asignificant amount of LPS, which remains bound to the protein even after isoelectric focussing. The significance of this is not known. Each of the three other mAbs (A14-5, B3-3, and A16-1) were shown to bind to subunit I of cytochrome d complex. Both the purified cytochrome complex (Fig. 2B, Lanes 1)and crude E. coli membranes (Fig. 2B, Lanes 2) were tested. The minor bands observed around 30,000 daltons are proteolytic fragments of subunit I which transferred and/or immunoreacted very efficiently. The relative amounts of these fragments varied from preparation to preparation (not shown). Initial results using monospecific polyclonal antisera toward the native cytochrome complex showed that at concentrations where 100% precipitation occurred, very little inhibition of ubiquinol-1 oxidase activity was observed prior to centrifugation (17). In those studies, even the immunoprecipitates possessed high quinol oxidase activity. Similar quinol oxidase inhibition-precipitation experiments were performed using the purified mAbs. In these experiments purified cyto-

2

1 0

L .A14-5

-

X

Monoclonal Antibody Characterization-Fourmonoclonal antibodies directed against the cytochrome d complex were obtained. Binding curves of the purified mAbs were deter-

x--+ o

.- o

820-4 83-3

" - " - -x- -

:....,.

c . A16-1 L . . .c o n t r o l

-

>,

> -

4 -

c

0001

RESULTS

7999

001

01

1

10

Anttbody ( u g )

FIG. 1. Binding curves of monoclonal antibodies against the cytochrome d complex. Assays were carried out using the solid phase assay with purified mAhs as described in the text. 1 pg of pure cytochrome d was dried in each well.

Characterization of Terminal Oxidare Complex Using Antibodies

8000

FIG. 2. Isoelectric focussing and SDS-PAGE immunoblots of the cytochrome d complex with mAbs. A, purified cytochrome(approximately 10 p g ) was focussed using LKR Ampholine PAC plates (pH 3.5 to 9.5) which were equilibrated for 30 min with 1% Triton X-100. Focussing was carried out as described previously (12). Lane I , the gel was stained with TMPD; Lane 2, gel stained withCoomassie brilliant blue; IAne 3 , proteins were transferred to nitrocellulose and immunblottingwas performed using mAb R3-3. R, SDS-PAGE immunoblots with theindicated mAb. Lane 1 , purified cytochrome d complex (5 pg); Lane 2, solubilized E. coli crude membranes (100 p g of protein).Blots were incubated with approximately 100 pI of ascites fluid/lO mlof fetalcalf serum in PBS.

chrome d complexwasused instead of crudemembranes solubilized in Triton X-100. In addition, precipitationwith S. aureus coated with protein A (Pansorbin) was used to determine antibody binding by precipitation of the antibody-antigen complexes. Monoclonal antibody B20-4, which is directed toward LPS, showed no inhibition of the ubiquinol-1oxidase activity, but was able to precipitate nearly100% of the cytochrome complex (Fig. 3A). This verified the previous results the which suggestedthe LPS was tightly bound to cytochrome. Monoclonal antibody B3-3 inhibited and bound very poorly to the native cytochromed complex in solution (Fig. 3B). At the concentrationsshown, both mAbs A14-5 and 1416-1 were able to inhibit ubiquinol-1 oxidase activity up to 60% and 40%,respectively(Fig. 3, C and D). The relative order of binding to the cytochrome complex in solution was A14-5 > B20-4 > A16-1 > B3-3. When mAbs A14-5 and A16-1 were

\ \\

820-4

83-3

A16-1

A14-5

i5K-

IO”.. 1

21

1

2

2

1

2

added in equimolarconcentrations more than 95% of the quinol oxidase activity was inhibited (Fig. 3E). When a ratio of mAb A14-5 to A16-1 of 1:4 was used, 100% of the quinol oxidase activity was inhibited, even a t very low concentrations of antibody (see Fig. 4C). Using specific activities estimated from the liquid phase precipitation assays, approximately one molecule of mAb A14-5 to onemolecule of mAb A16-1 to one molecule of the cytochrome d complex yielded 100% inhibition. Combinations of mAb B3-3 withA14-5 or A16-1 showed no increase in inhibitionover mAb A14-5 or A16-1 alone (not shown). Localization of Ubiquinol-1 and TMPD Binding Sites-In an attempt to localize the ubiquinol-1 and TMPD oxidation

I-

” ” 4

20

60

1 0 0

20

60

100

Anlnbody (pll

pg ontibody/pg cytochrome d

3. Inhibition of the ubiquinol-1 oxidase activity of the cytochrome d complex using purifiedmAbs. Ubiquinol-1 oxidase activity (c”.) was measured with increasing antibody concentrations. mAh binding to cytochrome d (O---O) was assayed by precipitation with Pansorbinandmeasuringquinol oxidase activity in supernatants as described under “Experimental Procedures.” FIG.

FIG. 4. Inhibition of the ubiquinol-1 and TMPD oxidase activities of the cytochrome d complex using polyclonal and monoclonal antibodies. Ubiquinol-1oxidaseactivity (M), TMPD oxidase activity (A---A), and binding to the cytochrome d complex (O---O) were measured as described under “Experimental Procedures.” mAb binding was measured by precipitation with Pansorbin followed by measuringthe quinoloxidaseactivity in the supernatant. All purified mAb preparations were 1 mg/ml and polyclonal antisera were approximately 10 mg/ml. Approximately 20 pg of cytochrome d complex were used per assay.

Characterization of Terminal Oxidase Complex Using Antibodies sites on theenzyme, the purified mAbs (A14-5 to A16-1, ratio 1:4) were incubated with the cytochrome complex followed by measurements of the ubiquinol-1 and TMPDoxidase activities. The results in Fig. 4C show that ubiquinol-1 and TMPD do not interact at the same site. Although the quinol oxidase activity was inhibited loo%, the TMPDoxidase activity was not inhibited at any concentration of antibody tested. T o study the siteof interaction of TMPD with the cytochromed complex, monospecific polyclonal antibodies were used. These antisera were previously shown to be monospecific toward their specific antigen (15).Polyclonal antisera toward subunit I was shown to behave nearly identically with the mAb A145:A16-1 mixture (Fig. 4B).The ubiquinol-1 oxidase activity was inhibited 100% whilenone of the TMPDoxidase activity was inhibited. Polyclonal antisera toward the native cytochrome dcomplexinduceda slightactivation of TMPD oxidase activity (150%), while partially inhibiting the quinol oxidase activity (Fig. 4A).Polyclonal antisera toward subunit I1 was shown to be the only antibody preparation able to inhibittheTMPD oxidase activity(approximately60%), while having little effect on the ubiquinol-1 oxidase activity (Fig. 40). These results suggest that the ubiquinol-1 binding site is located on subunit I, and the TMPD binding site is located on subunit 11. Proteolytic Peptide Mappingof the Monoclonal AntibodiesTheubiquinol-1 oxidase inhibitionexperimentsusingthe mAbs suggested that mAbs ,414-5 and A16-1 may bind at or close to the quinol binding site while mAb B3-3 is elsewhere. If this is the case, then mAb A14-5 and mAb A16-1 should be locatedin close proximitytoeachother while mAb B3-3 should bind toa more distant site. Thisview is supported by the proteolyticdigestion and immunoblotting experiments on the cytochrome d complex shown in Fig. 5. All three mAbs bound to the30,000-dalton polypeptide fragment of subunit I of the trypsin-treated cytochrome complex. When the cytochrome was digested with chymotrypsin, mAb B3-3 binding was completely lost, while mAbs A14-5 and A16-1 bound to four identical chymotrypticpolypeptides derived fromsubunit

I.

8001

cross-linking of subunits to determine the stoichiometry of native oligomeric proteins (e.g. Refs. 18 and 19). The stoichiometry of the subunits within the cytochrome complex was investigated using this bifunctional reagent, and the antibodies were used to analyze the cross-linked products. The cytochrome d complex in 0.1% Triton X-100 was cross-linked with varying concentrations of dimethyl suberimidate. The cross-linked products were then analyzed by SDS-PAGE using 8%polyacrylamide gels. Immunoblots of the cross-linked products were then carried outusing affinity purified subunit I1 antiserum (Fig. 6A) and a mAb A14-5 and B3-3 mixture (Fig. 6B) specific for subunit I. At higher cross-linker concentrations, more product with a molecular weight of 60,000 was formed. This product containedonly subunit I1 (see Fig. 6, A and B, Lanes 1 through 5 ) . If the cytochrome complex was solubilized in SDS before dimethyl suberimidate was added, no 60,000-dalton product was formed (Fig. 6A, Lane 6). When the antibodies toward subunitI1 were removed and the same nitrocellulose sheet was probed for subunit I, no major products significantly larger than 57,000 daltons were observed (Fig. 6B, Lanes 2 through 5 ) . Instead, intramolecular crosslinking was observed. This was apparent from the new bands observed in the immunoblotsdirectly around subunitI. These 4 or 5 new bands increasedin intensity withincreasing dimethyl suberimidate concentration (see Fig. 6B, Lanes 1 through 5 ) . These were also observed when SDS was added priorto cross-linking (Fig. 6B, Lane 6). The minor band present a t approximately 100,000 daltons (Fig. 6B, Lanes 1 through 5 ) may represent a small amount of subunit I dimer which was also present even when no cross-linker was added (Fig. 623, Lane 1). This minor component may reflect the tendency of the subunit toaggregate, even in the presence of SDS (8). The preference of subunit I to cross-link intramolecularly when dimethyl suberimidate was used complicates the analysis of this component, but it is clear that subunit I is not involved inmajorcross-linkedproducts either with other molecules of subunit I or with subunit 11.

Cross-linking Studies on the Cytochrome d Terminal Oxidase-Dimethyl suberimidate hasbeen used previously inthe 1 2 3 4 5 6 A 14-5 A

BA

A 16- I BA

B 3-3

B

200K92K68KW

43 55K -

K. " "

30K -

A

FIG. 5. Immunoblotting of partiallyproteolyzedcytoB3-3.Lane chrome dcomplex using mAbs A14-5. A16-1, and I , cytochrome d (400 pg in 0.2 ml of 10 mM Tris, pH 7.4, buffer) was proteolyzed for 1 min with40 pg of trypsin at room temperature. Idone 2, cytochrome d (400 pg in 0.2 ml of 50 mM Tris HCI, 5 mM CaCI,, pH 8, buffer) was proteolyzed with 4 pg of chymotrypsin for 5 min at room temperature. Reactions were stopped by adding S D S sample buffer, and 20 pg of the protease-treated cytochromed complex were loaded in each lone. Immunohlots with the indicated mAbs were then carried out as described previously (14).

B

FIG. 6. Cytochrome d complex cross-linking and immunoblotting. Cytochrome d (1 mg/ml of protein) was cross-linked according to Ref. 18 in 0.2 M triethanolamine HCI, 0.1% Triton X-100, pH 8.5, buffer. Cross-linking was carried out for 3 h using dimethyl suberimidate. Dimethyl suberimidate concentrations were: Lane 1, 0 mg/ml; Lane 2,0.05 mg/ml; Lane 3,1 mg/ml; Lane 4 , 3 mg/ml; Lane 5 , 12 mg/ml; Lane 6, 3 mg/ml. In Lane 6,cytochrome d was treated with 2% SDS before the cross-linker was added. The samples were then analyzed by SDS-PAGE (8%gels) and, after transfer to nitrocellulose, immunoblotting was carried out with ( A ) affinity purified suhunit I1 antibody (approximately 50 p g ) or ( R ) a mixture of mAbs A14-5 and R3-3which bind to subunit1. Approximately 6 pg of crosslinked cytochrome d complex were loaded in each lone. Standards used were: myosin, 200,000; phosphorylase B, 92,500; bovine serum albumin, 68,000; ovalbumin, 13,000.

8002

Characterization of Terminal Oxidase Complex Using Antibodies

polyclonal antibodies directed against denatured subunit 11 bind much more poorly to the native enzyme, indicated by One of the major results of this work istheapparent the fact that the antibody preparation does not precipitate localization of the sites for ubiquinol-1 and TMPD oxidation the cytochrome complex solubilized in Triton X-100. This on subunits I and 11, respectively, of the cytochrome d comsuggests that the majority of subunit I1 is inaccessible in the plex. These conclusions arebasedonantibodyinhibition native structure.It is also noted that subunit I1 is considerably studies using the various polyclonal and monoclonal antibodmore resistant to proteolysis by trypsin and chymotrypsin ies against this enzyme. Presumably, the site for ubiquinol-1 compared to subunitI (not shown), supporting theview that oxidation is physiologically relevant and corresponds to the I1 is exposed to solution in the native form of site of ubiquinol-8 oxidation, although this has yet to be little of subunit the oxidase. Studies using antibodies against subunit I1 sugdemonstrated. The application of other techniques will be gest that the TMPD site of oxidation is on subunit11, but do required toconfirmthe model suggestedby theantibody not indicate whether ubiquinol-8 can also be oxidized at this inhibition studies. It is noted that quinone binding sites have site. been examined by various methods in other systems (e.g. see The antibodiesdirected against subunitI1 have provedvery Refs. 20 and 21). Yu and Yu (20) have synthesized functionuseful in helping to define the quaternary structure of the ally active 14C-labeled arylazide-ubiquinonederivatives to oxidase in TritonX-100. Cross-linking studiesclearly showed identify ubiquinone-bindingproteins. Debus et al. (22) rethat subunit I1 ispresentin more thanone copy in the cently used monospecific polyclonalantibodies tolocalize the I was complex. No evidencefor cross-linkswithsubunit quinone binding sites to individual polypeptides in the Rhoobserved, nor were there indicationsof aggregates higherthan dopseudomonas sphaeroides reaction center. the dimer of subunit I1 forming upon cross-linking. This The antigenic determinants on the cytochrome d complex suggests a minimal active unit of one copy of subunit I and can be summarized asfollows: two copes of subunit 11. Molecular weight studies' are conClass I-Antibodies which bind to the purified, detergentsistent with thismodel. solubilized cytochrome d complex but which do not inhibit The antibodies described in this work have been essential ubiquinol-1 oxidase activity. This class includes polyclonal forobtaining significant structuralinformationaboutthe antibodies raised against the native enzyme, and the mono- cytochrome d complex. Future studies will include defining clonal antibodies which bind to the LPS component of the the topology of the enzyme subunits with respect to the E. oxidase. Most of the antigenic sites(790%) of the polyclonal coli membrane and defining the regions of the protein directly antibody preparation are located on subunit I as determined involved in quinone binding. The gene for the cytochrome d by immunoblottinganalysis(notshown).Thesedetermicomplex has been located (24) and cloned: and work is in nants, although readily accessible on the native oxidase, are progress to determine the DNA sequence of the gene. The apparently not located a t or near the quinol oxidation site. continued use of immunological methods in combination with Theimmunoprecipitated enzyme obtainedwiththis polyinformation derived through techniques in molecular biology clonal antibody preparation still possesses ubiquinol-1 oxidase will be very useful for defining the architectureof this interactivity. mAb B20-4 binds to the LPS component of the native esting enzyme. oxidase, and also has little inhibitory effect. This antibody was used to demonstrate thatLPS binds tightly to the native Acknowledgments-We thank Dean Ballard in Dr. Edward Voss's oxidase, even following isoelectric focussing. It is yet to be laboratory forhis expert advice and help in obtaining the monoclonal shown whether this association is adventitious or whether the antibodies. We thank Pat Porter for supplying purified cytochrome d. LPS is bound to the enzyme in situ. Class II-Antibodies which bind to the denatured oxidase REFERENCES but bind onlypoorly to the native enzyme. The antigenic 1. Weldon, S. L.,Mumby, M. C., Beavo, J. A., and Taylor, S. S. determinant on subunit I of mAb B3-3 is readily accessible (1983) J . Biol. Chem. 258, 1129-1135 when the protein is denatured (e.g. immunoblotting or solid 2. Gullick,'W.J., Tzartos, S., and Lindstrom, J. (1981) Biochemistry phase assay), but isless accessible when the nativeenzyme is 20,2173-2180 in solution. 3. Rosen, D., Okamura, M. Y., Abresch, E. C., Valkirs, G. E., and Class III-Antibodies which bind to the native enzyme in Feher, G. (1983) Biochemistry 22, 335-341.4. Watters; D., and Maelicke, A. (1983) Biochemistry 2 2 , 1811solution and inhibit ubiquinol-1 oxidase activity. These in1819 clude monoclonal antibodies A14-5 and A16-1 as well as 5. Froehner, S. C., Douville, K., Klink, S., and Culp, W. J. (1983) I. polyclonal antibodies directed against denatured subunit J. Biol. Chem. 258,7112-7120 All antibodiesinthis class bind specifically to subunit I. 6. Freedman, J. A., and Chan, S. H. P. (1983) J. Biol. Chem. 2 5 8 , Antibody bindingtothe oxidasedoes notinhibitTMPD 5885-5892 oxidase activity. It is, thus, strongly suggested that the ubi7. Rogers, D. H., and Rudney, H. (1982) J. Biol. Chem. 257,1065010658 quinol-1 oxidase site is on subunit 1 and that the site for 8. Miller, M. J., and Gennis, R. B. (1983) J. Biol. Chem. 258,9159TMPD oxidation is located at a different site. Both A144 9165 and A16-1 bind to the same tryptic and chymotryptic poly9. Bragg, P. D. (1979) in Diversity of Bacterial Respiratory Systems peptide fragments, indicating that their determinants are in (Knowles, C. J., ed) Vol. 1, pp. 115-136, CRC Press, Boca close proximity to each other. It has previously been shown Raton, Florida that subunit I is the cytochrome bbS8component of the cyto- 10. Koland, J. G., Miller, M. J., and Gennis, R. B. (1984) Biochemistry 23,445-453 chrome d complex (23). Thus, it appears that the b-heme is a t or near the siteof quinol oxidation. None of the antibodies, 11. Kranz, D. M., Billing, P. A., Herron, J. N., and Voss, E. W. (1980) Zmmunol. Commun. 9,639-651 however, cause a perturbation of the reduced minus oxidized 12. Kranz, R. G., and Gennis, R. B. (1982) J. Bacteriol. 150,36-45 difference spectrum. 13. Kranz, D. M., and Voss, E. W., Jr. (1981) Proc. Natl. Acad. Sci. Class IV-Antibodies which bind to native and denatured U. S. A. 78, 5807-5811 subunit I1 and which inhibit TMPD oxidase activity but not ubiquinol-1 oxidase activity. No monoclonal antibodies have M. J. Miller and R. B. Gennis, unpublished data. been characterized as being directed against this subunit. The G. N. Green and R. B. Gennis, unpublished data. DISCUSSION

Characterization of Terminal Oxidase Complex Using Antibodies 14. Kranz, R. G., and Gennis, R. B. (1983)Anal.Biochem. 127,247257 15. Kranz, R. G., Barassi, c. A., Miller,M. J., Green, G. N.3 and Gennis, R. B. (1983)J. Bacteriol. 156, 115-121 Tsail '.* and Frasch9 ' E' (lgg2) Biochem' '19' 115-119 17. Kranz, R. G., and Gennis, R. B. (1983)J . Biol. Chem. 258, 10614-10621 18. Davies, G. E., and Stark, G. R. (1970)pmc. Natl. Acad. Sci U. S.A . 66,651-656 (1984) 19. LaPorte, D. C., Toscano, W. A., Jr., and Storm, D. R. (1979) 1275 Biochemistry 18,2820-2825

8003

20. Yu, C. A., and Yu, L. (1980)Biochem. Biophys. Res. Commun. 96,286-292 21. Trumpower, B. L., and Haggerty, J. G . (1980)J. Bioenerg. ~ i ~ membr. 12, 151-164 22. Debus, R. J., Valkirs, G. E., Okamura, M. Y., and Feher, G. (1982) Biochim. Biophys. Acta 682,500-503 23. Green, G . N., Kranz, R. G., Lorence, R. L., and Gennis, R. B. J . Biol. Chem. 259, 7994-7997 24. Green, G. N., and Gennis, R. B. (1983)J . Bacteriol. 154, 1269-

.