Photoaffinity Labeling of the Antimycin Binding Site in Rhodopseudomonas sphaeroides

Photoaffinity Labeling of the Antimycin Binding Site in Rhodopseudomonas sphaeroides

A special thanks to all of the members of the 11 Takemoto Lab   The major finding of this study was the identification of an 11,000 dalton polypeptide as the predominantly labeled protein.
Although this polypeptide was not exclusively labeled, it was consistently labeled and showed competition with antimycin. These results are consistent with a similar study performed by das Gupta and vii Rieske (1973)  INTRODUCTION Antimycin A is a potent and specific inhibitor of electron transport within the ubiquinone cytochrome b-c 1 oxidoreductase region · of the photosynthetic bacteria (Dutton & Prince, 1978, Gabellini et al, 1982, the analogous Complex III region of mitochondria , Wikstrom & Berden, 1912, and the cytochrome b 6 -f region of chloroplasts (Huber & Edwards, 1977). Because of the ubiquitous and highly specific nature of its inhibition, antimycin is a valuable tool Antimycin A was first reported in 1948 as a fungicide produced by a species of Streptomyces (Leben & Keitt, 1948). The first evidence that antimycin toxicity was due to inhibiton of mitochondrial respiration was reported by Ahmad et al in 1950. The site of action was further pinpointed by Chance and Williams (1956) who showed that when respiring mitochondria were treated with antimycin, a redox crossover point was observed between cytochromes b and c. This evidence indicated that the site of inhibition was between these cytochromes. Hatefi and coworkers (1962) demonstrated that antimycin was a potent inhibitor of purified mitochondrial reduced · ubiquinol-cytochrome c reductase (Complex III). Chance (1952) showed that antimycin caused oxidant-induced reduction of cytochrome b of the mitochondrial respiratory chain. This unusual property of antimycin has also been shown to occur in Rhodopseudomonas sphaeroides (Dutton & Prince,1978). This one effect of antimycin has been a "test" for models proposed for electron transport through the ubiquinonecytochrome b-c 1 region. If a model cannot justify this event then it is not a viable model of electron transport through this region.
Linear schemes of electron transport through this region have been abandoned because of failure to explain oxidant-induced reduction of cytochrome b (Bowyer & Trumpower, 1981).
The ubiquinone cytochrome b-c 1 regions of the respiratory chain of mitochondria and the cyclic photosynthetic electron transport chain of R. sphaeroides show a high degree of similiarity. Both systems contain similar redox components. There are two thennodynamically and spectrally distinct b cytochromes, a memhrane-bound cytochrome c 1 , a membrane associated cytochrome c or c 2 , and an iron sulfur protein which is characterized by an EPR resonance at g=l.90. There is also a bound ubiquinone in both ubiquinone-cytochrome b-c 1 regions. The similarity of electron transport components indicates that the 3 pathways of electron transport may also be similar if not identical in the b-c 1 complexes of mitochondria and the photosynthetic hacteria (Bowyer & Trumpower, 1981  There exist two kinds of light-harvesting antenna pigment-protein complexes.
The first is B875 with an absorption peak at 875 nm, and the second is B850 which has absorption peaks at 800 nm and 850 nm.
These pigment complexes are composed of bacteriochorophyll ~.
carotenoids, and protein and function to absorb and transfer light energy to the reaction centers.
The reaction center is responsible for converting the light energy of the photon to a form usable by the cell. The reaction center is composed of three protein subunits with a total molecular weight of approximately 100,000 daltons, (Okamura et al, 1974)  Ubiquinone then accepts the electron from the bacteriopheophytin with ferrous iron serving as a modulator. Two types of quinone are present, a 0 1 which is more tightly bound and the first of the quinones to receive an electron, and 0 11 which is more loosely bound a. nd receives the electron from 0 1 • 011 passes the electron to the ubiquinone-cytochrome b-c1 oxidoreductase.
The ubiquinone-cytochrome b-c 1 is composed of at least two types of b cytochromes, a cytochrome c 1 , a Rieske-type iron-sulfur protein (Rieske et al, 1964a} and a quinone (Oz) which functions as a reductant for cytochrome c. The exact mechanism for electron flow through this region is still controversial. The linear schemes which were originally proposed have been abandoned because of failure to describe a mechanism for oxidant-induced reduction of cytochrome bin the presence of antimycin A. Cyclic models are the likely candidates for the actual mechanism. The two pioneering cyclic models are the Q cycle as proposed by Mitchell, 1975 (Fig.Sa)   was determined by chemical degradation (van Tamelan et al, 1961, Dickie et al, 1963. Antimycin is composed of an acyl-and alkylsubstituted nine membered dil actone ring joined to 3-formamidosalicylic acid by an amide bond. Antimycin A 1 and A 3 differ by two methylene groups in the alkyl substituents. In an attempt to identify the structurally-important functional groups a number of synthetic homologs and analogs of antimycin have been synthesized. Kinoshiota and Umezawa (1971) replaced the dilactone ring with a 15 membered dilactone ring, and Dickie and coworkers (1963) replaced the dilactone ring with a eighteen carbon alkyl group. Both of these analogs retained high inhibitory activity.
These studies indicated that the dilactone ring functions in a supporting role, possibly providing the hydrophobicity necessary to penetrate the membrane. Structure-activity studies indicate that the other important functional groups necessary for inhibitory acivity are the formamido and the phenolic hydroxyl groups of the aromatic ring.
The formamido function is thought to he important as an electron withdrawing group because replacement by a nitre functional group in a 13 position either ortho or para to the hydroxyl, forming 3-nitro or 5-nitrosalicyl-N-(n-octadecyl)amide, also yields an active compound (Neft, 1971, Neft & Farley, 1971).

Effects of _Antimycin on Mitochondrial Respiration
Many phenomema are observed in the properties of Complex III treated with antimycin. Electron transport between ubiquinol and ferricytochrome c is inhibited by stoichiometric concentrations of antimycin , Rieske & Zaugg, 1962. Secondly antimycin strongly inhibits the cleavage of cytochromes b and c by dissociating reagents such as bile salts or guanidium salts . Thirdly, the b cytochromes are altered in their redox and spectral properties.
In the presence of antimycin the rate and extent of reduction of the b cytochormes by succinate or NADH and in purified Complex III by ubiquinol is greatly increased (Rieske, 1971, Pumphrey, 1962, Baum & Rieske, 1966. Other primary effects on cytochrome bare as follows: 1. Antimycin displaces the absorption peak of the ferro forms of both low and high potential cytochrome b-562 to longer wavelengths by 1.5-2.0 nm (red shift) (Chance, 1958, Pumphrey, 1962. It has no effect on the position of cytochrome b-566 or b-558 (Berden & Opperdoes, 1972).
4. Multiple splitting of the soret band (415 nm) of the oxidized complex and increased resolution of the absorption band in the visible region is seen with the addition of antimycin (Dervartanian et al, 1973).
These effects on cytochrome b indicate that antimycin affects the binding of one or both of the heme ligands sufficiently that it can be seen in the absorption spectrum of both ferro-and ferricytochrome b (Slater, 1973).
Two EPR-detectable semiquione radicals are present in the mitochonrial complex. Both signals are affected by the presence of antimycin. Antimycin enhances the intensity of the g=2.00 signal that is _ observable at high microwave power, and also causes a shift in the potentiometric titration toward a more positive potential. The g=2.00 signal observable at low microwave power disappears upon addition of antimycin (Ohnishi & Trumpower, 1980). This site (Oc) is thought to be closely linked to the antimycin site (Wraight, 1982).
Effects of Antimycin on R. sphaeroi des 15 Antimycin affects the bacterial system in a similar manner to mitochondria. The rapid reduction of flash oxidized cytochrome c 2 is abolished (Dutton & Jackson, 1972), and photoreduction of cytochrome b is stimulated. Oxidant-induced reduction of cytochrome b (Dutton & Prince, 1978) and the inhibition of the uptake of the second proton per electron after a short light flash have been shown (Petty & Dutton, 1976, Petty et al, 1977. Antimycin abolishes Phase III of the redshift of the absorption bands of endogenous carotenoid pigments. This shift registers electrogenic charge transfers across the membrane, and Phase III is closely related to ferricytochrome c 2 reduction (Jackson & Dutton, 1973, Bashford et al, 1979.

Factors Affecting the Binding of Antimycin A
The Binding of antimycin to mitochonddrial membranes and purified Complex III is specific and stoichiometric. The affinity of antimycin for its site is very high. The dissociation constant for antimycin in submitochondrial particles has been estimated to be 0.032 nM with the cytochromes band c in their oxidized state (Berden & Slater, 1972 antimycin than membrane-bound complex. Dissociation constants for the oxidized complex have been measured at 0.1 pM (Rieske & das Gupta, 1972).
The binding of antimycin to intact Complex III was found to be essentially irreversible.
No transfer of antimycin from treated complex to untreated complex was seen . However, up to 70% of the bound antimycin could be extracted using taurocholate or acetone . Conditions that gave cleavage of the complex also inhibited antimycin binding .

Shape of the Antimycin Inhibition Titration Curve
The shape of the inhibition titration curve of the effects of antimcyin in the mitochondria raised some discussion during the early seventies. Sigmoidal titration curves were found for the inhibition of respiration of submitochondrial particles with NADH or succinate as substrate (von Jagow & Baher, 1975) and for inhibition of the growth rate of yeast cells (Burger et al, 1976). The red shift of cytochrome b has been seen to follow curves which are hyperbolic or sigmoidal depending on the experimental conditions (Bryla et al, 1969a). The inhibition of cleavage of isolated Complex III is proportional to antimycin until saturation is reached. Bryla et al (1969b) proposed that the binding of antimycin was cooperative in nature. This proposal was further supported by Berden and Slater (1972). An alternate proposal was put forth by Potter and Reif (1952) and supported experimentally by others Kroger & Klinenburg, 1973a, Rieske & das Gupta,1972. This proposal maintains that the sigmoidal cruves of submitochondrial particle result from sequential electron transfer reactions.
The electron transfer capacity of the ubiquinol cytochrome b-c region is higher than that of the preceding region from NADH or succinate to ubiquione.
Only when the antimycin-sensitive region is sufficiently inhibited to limit the overall electron transport will respiration be significantly inhbited (Kroger & Klingenburg, 1973b).
van den Berg and coworkers (1979)  These results support the noncooperative model described for the mitochondrial system (Kroger & Klingenberg, 1973b).

Other Antimycin-Like Inhibitors
Several other antimycin-like inhibitors have been identified.
HQNO shows an inhibitory capability at least one order of magnitude less than antimycin. Also, the shape of the titration curve is different. Brandon et al (1972) compared inhibition curves for antimycin and HQNO using succinate oxidase, NADH oxidase, and ATP supported reverse electron transport in submitochondrial particles.
Antimycin produced a concave or a sigmoidal curve whereas HQNO produced either a convex or hyberbolic curve. HQNO, like antimycin, induces extra reduction of cytocrhome b-566 and a shoulder at 557-558 nm. The increase in absorbance at 557 nm is greater for HQNO than for antimycin. HQNO produces no red shift of cytochrome b 562 (Brandon et all, 1972). Results indicate that HQNO and antimycin compete for the same site (Van Ark & Berden, 1977).
Funiculosin is an antibiotic produced by Penicillium funicolosin. Moser et al (1977) proposed that its activity was much like that of antimycin. Nelson et al (1977) showed that funiculosin inhibits electron flow between cytochrome b and cytochrome c in submitochondrial particles, and also inhibits duroquinol-cytochrome c 19 reductase activity of isolated Complex III. Funicolosin also induces extra reduction of cytochrome bin aerobic steady state conditions.
The binding affinity of funiculosin is less than that of antimycin.
Mucidin, an efficient fungal antibiotic, has also been shown to induce a redox crossover point between cytochrome band cytochrome c+cl (Subik et al, 1974a (Convent & Briquet, 1978).
Diuron was developed as a potent herbicide and is known to inhibit photosystem II electron transport in chloroplasts. It is also a respiratory electron transport inhibitor of yeast motochondria.
Inhibition of electron flow is between cytochromes band c (Inoue et al, 1967). Diuron also causes the extra reduction of cytochrome b, properties (Tappel, 1960). However, Rieske et al (1964b) were able to split the iron-sulfur protein away from the mitochondrial complex III.
This iron-sulfurless complex was still sensitive to antimycin.
Therefore, the iron-sulfur protein was dismissed as a potential site for antimycin interaction.
Cytochrome b has long been considered the prime site of interaction of antimycin. As seen with both mitochondrial systems and the photosynthetic systems of R. sphaeroides, the b cytochromes show pronounced spectral changes when antimycin is added. Storey (1972) has proposed a direct binding of antimycin to mitochondrial cytochrome region .
Recent work indicates that antimycin may function as an inhibitor by displacing either a quinone or a quinol from the Oc site (Cramer & Crofts, 1982). The Qc site is proposed to function in the stabilization of the semiquione observable by EPR at g=2.00 (low microwave power) (Ohnishi & Trumpower,1980). It is speculated that the redox changes observable upon addition of antimycin may be caused by the loss or gain of electrons as a semiquinone species is converted to either a _quinol or a quinone, by electron transfer to or from neighboring redox centers before the inhibitor displaces the group (Velthuys, 1982).

Multivalent Model of Inhibitor Binding
Rieske ( 1980)  Different molecular shapes and functional moieties of the inhibitors will perturb to different degrees the subunits that are exposed to the inhibitor on the wall of the cavity (Fig. 6).   (Gilchrist & Rees, 1969). Also, carbenes have a tendency to undergo Wolff rearrangements to form ketenes (Charmovich et al, 1968).
The nitrogen analog of a carbene is the nitrene. Nitrenes can be generated from alkyl-and arylazido groups, isocyanates, and carbonylazides (Gilchrist & Rees, 1969). Nitrenes offer the advantage However, they are longer lived than carbenes (10-4 sec), and they are more selective. In many cases they act as electrophiles and react with many of the nucleophilic groups that are attacked by ordinary protein modifiers.
Several of the criteria that should be met before using a nitrene and its azide precursor as a photoaffinity precursor are as follows.
The nitrene precursor should contain functional groups which allows it to bind with the desired receptor site. The nitrene should be generated at wavelengths that will not damage the receptor site. The precursor should be chemically stable at the temperature used, and it should not react with the solvent (water) (Singer, 1967 3-aminosalicyl-N-(n-octadecyl)amide .l!..!.l.Compound I · was reduced to the amine derivative and recovered as the amine hydrochloride as described previously (Neft,1971, Neft & Farley, 1971. The product was recovered with 70% yield, the m.p. was 155 C, and an Rf value of 0.45 was found using silica gel TLC with a solvent system of chloroform: methanol :ammonia(99:1:0.l).
Alternately, 0.5 g (1.5 mmol) was dissolved in 20 mls of chloroform, and 0.2 g of 5% palladium on carbon was added.  (Crounse et al, 1963) was used for all membrane labeling work, and in all inhibition studies. Cultures were maintained on YEMG agar stabs (Lascelles, 1969).

Growth Medium and Conditions
R. sphaeroides strain Ga was grown in MG media (Lascelles,1969 with glacial acetic acid and heated at ?0°C for 5 minutes. Gels were cast and run as described by Anderson et al (1983). This method incorporates the use of 8% acrylamide and 8M urea in the separating gel and was useful in separating the bulk lipid molecules and detergent from low molecular weight polypeptides. Approximately 25-50 ug of protein was applied to each lane of the gel. Gels were run in a Hoefer SE 600 dual slab gel apparatus at 12°c with a constant current of 50 mamps. The gels were stained using Coomassie Blue stain (Fairbanks et al, 1971), and destained · using a 10% isopropyl alcohol, 10% acetic acid solution.

Fluorography
The gels were prepared for fluorography by either using a commercially available fluorographic solution (Enhance or Fluorohance) or acetic acid and PPO as described by Skinner and Griswold (1983).
Gels were dried onto Whatman I filter paper using a Biorad slab gel dryer model 224.
film at -7o 0 c The dried gels were exposed to Kodak X-0mat R x-ray for 2 weeks. The fluorograms were developed using Cronex developer for 5 minutes at 25 C and Kodak fixer.

Phospholipid Extraction and Quantitation
To determine the amount of protein bound [ 3 H]-III as compared to lipid bound label, chromatophores were labeled with the photoreactive species and lipids were extracted (Lascelles & Szilagyi,1965 (Dittmer & Lester, 1964). Sections of the Silica gel plate were scraped and counted in order to quantitate the radioactivity.

Synthesis of a Photoactivatable Antimycin Analog
A photoaffinity analog of antimycin, 3-azidosalicyl-N-(n-octadecyl)amide was synthesized as diagrammed in Fig. 7 • Each step in the synthesis was accomplished with yields exceeding 50%.
The published method for synthesis of 3-nitrosalicyl-N-(n-octadecyl)amide (Neft & Farley, 1971, Neft, 1971 used .2_-phenylenephosphorochloridite (Anshutz et al, 1.943) as the coupling agent for amide bond formation. The overall reaction for this step in the synthesis is as follows: Attempts to duplicate this work yielded a yellow compound with a melting point of over 200°c. This material may have been a phenolic salt as described by Neft (1971). Also, the published yields for the formation of the product were low. Therefore, the use of DCCD as the coupling agent for amide bond formation was investigated. Using DCCD as the coupling agent, a yellow crystalline product with a melting point of 87-88°C was obtained. This melting point compares favorably with the published values of 85-86°c (Neft & Farley, 1971). The yielrl of this reaction was greater than or equal to 50% using DCCD as 35 compared with the published yield of 38% using .2_-phenylenephosphorochloridite. The mechanism for the formation of an amide bond using DCCD is as follows:

RCNHR"+ RNHCHHK
The catalytic reduction (Pd/C, H 2 gas) of Compound I to 3-aminosalicyl-N-(n-octadecyl)amide was performed as published (Neft,1971) with good yield. Adaptations were made in the technique in order to accomodate smaller reactant quantities and to better contain the radioactive materials. The use of sealed septum bottles as described in the METHODS section alleviated both of these concerns.
The synthesis of 3-azidosalicyl-N-(n-octadecyl)amide (III) was performed in the dark because the product is sensitive to light.
Compound III is also thermally sensitive, and was concentrated by precipitation with water instead of flash evaporation in order to avoid decomposition. A spectrum was taken after each period of irradiation, and disappearance of signal intensity observed. Ji (1977) has published similar spectra as evidence of photoactivation of arylazides (Fig.   8).

Inhibition of Succinate-Cytochrome c Reductase Activity
The  9).
By comparison, the amount of 3-nitrosalicyl-N-(n-octadecyl)amide needed to inhibit succinate-cytochrome c reductase in mitochondria by 50% is 30 umoles I mg of protein and is also sigmoidal (Neft,1971). This is as would be predicted or the basis of the electron withdrawing capabilities of a nitro-function as compared to an azido-function. Neft and Farley (1971)  µmole Inhibitor/ mg Protein 43 3-azidosalicyl-N-(n-octadecyl)amide needed for 50% inhibition of succinate-cytochrome c reductase activity. However, the difference in activity may also be due to differing amounts of ubiquinone-cytochrome b-c 1 in the two systems. Antimycin A inhibits the same reaction in mitochondria by 50% at a concentration of 0.05 umoles/mg protein (Neft, 1971).

Photoaffinity Labeling of the Antimycin Binding Site
The strategy for labeling the antimyci~ site is based on two fundamental premises. These conditions are that the 3-azidosalicyl-N-(~-octadecyl)amide has a high affinity for a specific polypeptide in the membrane, and that the reactive arylnitrene, which is generated upon irradiation, will covalently bind to this same polypeptide. Incorporating a radiolabel into the analog allowed identification of a specific antimycin binding polypeptide by NaDodS0 4 PAGE analysis.
In both purified ubiquinone-cytochrome b-c 1 oxidoreductase ( Fig.   10) and chromatophores from R. sphaeroides, (Fig. 13)  Inhibition of analog binding by antimycin is interpreted to mean that the analog used in this study and antimycin occupy the same site.
Antimycin has a very high affinity for the site relative to 3-azidosalcyl-N-(n-octadecyl)amide (Rieske & das Gupta, 1972) Therefore, little labeling by the radioactive analog would be expected when precincubation with antimycin precedes addition of the analog (Figs. 10,13 lanes D,E,F).
Several other proteins were labeled in the chromatophore preparation (Figs. 13,14). These proteins were not consistently labeled by the analog, and were labeled to a lesser extent than the 11,000 dalton polypeptide. These proteins seem to correspond to bands within the purified protein complex (ubiquione-cytochrome b-c 1 oxidoreductase) (Fig.11), and show competition with antimycin under some conditions.  Lanes Band E were incubated five minutes prior to irradiation and l anes C and F were irrad i ated immediately. Gels wer e r un by the method of Anderson et al ( 1983) and prepared for fl~orography as de~cribed in the text.   Lanes D,E, and F were preincubated with antimycin before addition of the photoaffinity analog and irradiation. Aliquots of chromatophores containing 100 ul (100-150 ug of protein) were either incubated for 10 minutes with 10 ul of ethanol (lanes A,B, and C) or with 10 ul of (10 mM) an~imycin A (lanes D,E, and F). Subsequently, 10 ul of [ HJ-III (0.8 uCi) were added to each sample. Each sample was then irradiated immediately as described in the text. Gels were run by the method of Anderson et al (1983) and prepared for fluorography as described in the text Figure 14. NaDodS0 4 -PAGE of Chromatophores from R. sphaeroides.
Lanes A and Bare molecular weight stanaaras. Lanes C and D the chromatophore preparation.
Gels were run by the method of Anderson et al (1983 Takemoto, 1981), were scraped and counted.
There was no significant amount of radiation found in the phosphatidylglycerol, and only a small amount of the applied radiation was present in t~e phosphatidylethanolamine ( Table 2). The largest amount of radiation present in the organic extract comigrated with the unreacted 3-azidosalicyl-N-(n-octadecyl)amide The hydrophobic nature of the antimycin analog used in this work could · explain why this large amount of radioactive material did not wash out of the gel during the staining and destaining procedures.
The use of the NaDodS0 4 -PAGE system described by Anderson et al (1983) enhanced the separation of the labeled protein from the large radioactive band previously described. Numerous attempts to identify the antimycin binding site using a conventional Laemmli (1970) gel system with total acrylamide concentrations ranging from 12-16% did not resolve the 11,000 dalton polypeptide from the detergent-lipid-unreacted analog band.
The identification of a protein with a molecular weight of 11,000 daltons as the antimycin binding protein is the major finding of this work. This finding supports the work of das Gupta and Rieske (1973), who reported the binding of deformamidoazidoantimycin to an 11,500 dalton protein in beef heart mitochondria.   (1973), Capaldi (1974), Hare and Crane (1974), and Gellerfors and Nelson (1975) all contain components with molecular weights in the range of 10,000 to 12,000 daltons. Recent purifications of this protein complex in~-sphaeroides by Yu and Yu (1982), Takamiya et al (1982), and Menick (1984) all contain proteins of this general size.
Of the possible antimycin binding sites proposed earlier in the LITERATURE REVIEW of this thesis, only the proposal that antimycin binds directly to cytochrome b would be affected by the results of this work. The published molecular weight of cytochrome b in mitochondrial systems (beef heart) is approximately 31,000 daltons (von Jagow et al, 1978) and in R. sphaeroides ~round 40,000 daltons (Gabellini et al, 1982, Menick, 1984 Both of the photoaffinity labeling studies performed thus far seem to be in agreement as to the approximate size of the protein which most strongly interacts with these analogs. Since the weight of this protein does not correspond to the published molecular weights of cytochrome bin either system, it can be concluded that the mode of action of antimycin is not solely by direct binding to cytochrome b. 58 The multivalent binding model described by Rieske (1980) and the hypothesis that antimycin interacts with one of the quinone binding sites seem to be the most plausible models of antimycin interaction with the ubiquinone-cytochrome b-c 1 oxidoreductase. Yu and Yu (1980) have reported using a photoaffinity quinone analog to label the quinone binding site of mitochondrial ubiquinone-cytochrome b-c 1 oxidoreductase. The approximately 15,000 protein labeled has a molecular weight of daltons.
Attempts by these authors to show competition of antimycin for this quinbne binding site were negative.
This evidence does not rule out the possibility that antimycin interacts with a quinone binding protein because more recent work indicates that there are two semi-quinone species which interact at two different sites (Ohnishi & Trumpower, 1980). The multivalent binding model proposed by Rieske (1980) also seems to be a viable model for antimycin binding.
using chromatophores showed several proteins The binding studies labeled by the radioactive photoaffinity analog with the 11,000 dalton polypeptide being predominantly labeled (Fig. 10,13). T~e interpretation of this finding is not readily apparent, but these results could be used to support the multivalent binding model.
Although the specificity of 3-azidosalcyl-N-(n-octadecyl )amide for the has not been rigorously tested, several fundamental aspects of the analog would offer credence to this supposition. These properties are as follow:

1)
Preincl!bation of either purified ubiquinone-cytochrome b-c 1 complex or chromatophores with antimycin 59 prior to exposure to the photoaffinity analog inhibited analog binding.
2) [ 3 HJ~III inhibits electron transport in a manner similar to antimycin.
3) Both antimycin and this analog show sigmoidal titration curves under certain conditions. 4) The concentration of this analog necessary for inhibition is low enough to indicate a specific site of interaction. 5) The photoactivatable azido is located adjacent to the phenolic hydroxyl group which is considered to to be the functional group of primary importance in inhibition. ' 6) A study using deformamidoazidoantimycin as a photoaffinity label in a mitochondrial system yielded similar results. 7) Similar results were obtained in this study with both a purified system and chromatophores.
In summary, this study has shown that direct binding to cytochrome bis not the only site of interaction action of antimycin.
The primary site of binding for the analog used in this study was a polypeptide with a molecular weight of 11,000. This polypeptide was labeled both predominantly and consistently, but not exclusively. The other proteins labeled showed variability botri in the amount of label bound and consistency in binding the analog. Therefore, this study does not offer unequivocal evidence for the exclusive role of the 11,000 dalton polypeptide as the antimycin binding site. More research is necessary to distinguish between a single protein as the binding site and the multivalent binding theory.