Electron Paramagnetic Resonance-detectable Electron Acceptors in Beef Heart Mitochondria REDUCED REDUCTASE SEGMENT OF THE ELECTRON TRANSFER SYSTEM*

EPR-detectable of

through technical developments in the application of EPR spectroscopy to biological systems a number of additional, thus far unknown, oxido-reductively active components of mitochondria have been recognized.
The majority of these components belongs to the family of iron-sulfur compounds, and it indeed appears that there are more iron-sulfur groups present in mitochondria than cytochromes, which were previously thought to be the principal type of electron carrier in mitochondria.
Recent work also implicates iron-sulfur centers in energy conservation coupled to electron transfer (2, 3). Although there is now general awareness of these observations and notions derived therefrom, lit,tle systematic EPR work has been published which would make this knowledge more generally accessible and could provide a firm foundation for further studies.
The present series of papers represents an attempt to provide basic information on those electron carriers in mammalian mitochondria that are detectable by EPR spectroscopy at the present state of technical developments.
In the resolution of the EPR signals and assignment of these signals to the corresponding components mainly six approaches were used: (a) physical separation by fractionation of mitochondria into their constituent complexes and soluble proteins; (b) reductive titration, by which the signals of species of different oxidation-reduction potential can be made to appear in sequence; (c) rapid kinetic studies, by which electron acceptors may be functionally separated if their rates of reaction differ; (d) use of inhibitors to modulate the results obtained in (b) and (c) above; (e) variation of observation temperature, which will allow one to resolve the signals from components with different electron spin relaxation rates or different strength of magnetic coupling; (f) variation of incident microwave power to resolve signals from species with different spin relaxation rate.
The present papers report on attempts using these approaches to obtain spectra of individual components, identify and characterize them, and explore their gross functional roles. The first paper is concerned with the DPNH-ubiquinone reductase segment of the electron transfer system. A preliminary report of this work has appeared (4). The accompanying paper (5) will deal with the next segment of the electron transfer system, viz. that of ubihydroquinone-cytochrome c reductase. The accompanying paper also contains a discussion of the behavior and relationships of components dealt with in both papers and a discussion of some implications of the fact that the electron transfer system consists of, or is in equilibrium with, a much larger number of electron carriers than previously thought.
In the case of Centers 3 + 4, which are never seen alone, three approaches were used: (a) a reasonable line shape for the center line was assumed by analogy to those of other iron-sulfur proteins,   proceeds ( Fig. 1B) additional resonances emerge; one at g = 2.00, due to a free radical, and one with components at g = 2.036 and 1.986. These latter resonances disappear again on further reduction, a behavior typical of resonances from an intermediate state of a two electron acceptor. These three species are unidentified, although it is likely that the free radical signal largely represents flavin semiquinone. The unidentified species will be considered in the accompanying paper (5) since their signals are more prominent in the materials studied there.

Resolution of Resonances from Iron-Sulfur
Centers 1 and Z-The other resonances observed in these spectra of Complex I were assigned to four iron-sulfur centers as follows. In Spectrum B of Fig. 1 the resonance at g = 2.05 is approximately 50% of its maximal value (cf. Spectra D and E) and in Spectrum C 70%. The line at g = 1.92 increases in a parallel manner up to Spectrum C but then increases further as the lines at g = 2.02 and 1.94 develop. It appears, therefore, that two iron-sulfur centers are reduced in sequence, one with axial symmetry showing lines at g/l = 2.05 and gl = 1.92, and one with rhombic symmetry and lines at g = 2.02 (gJ; 1.94 (gY), and 1.92 (g,), such that gl of the former overlaps with g, of the latter. This assignment was confirmed by variation of temperature and microwave power. Fig. 2, A and B, shows the spectra of samples analogous to those of Fig. 1, B and D, at 42 K. The line at g = 2.00 is due to a free radical, presumably flavin. It is obvious that the lines of the iron-sulfur center wibh gll = 2.05 and gL = 1.92 are severely broadened at this temperature, whereas those of t.he other center are still quite distinct with gz and g, resolved. This latter iron-sulfur center observed in Spectrum .S'B is therefore presumably the same that had been observed in previous EPR studies conducted at 95 K (15). This center has been called Center 1 and the center with the more t.emperature sensitive spectrum, Center 2 (4). In a similar manner the same relationship of the various lines can be demonstrated by increasing the microwave power at very low temperature. In Fig. 3 the spectrum of the sample of Fig. 2B, in which the lines of both Centers 1 and 2 are well developed, is shown at 2.7 and 270 microwatts of power at 7.7 K. It can be clearly seen that the line at g = 2.05 of Center 2 is somewhat diminished with high power while the lines at g = 2.02 and 1.94 are increased. Since the line at g = 1.92 also increases it can be concluded that it has a contribution from Center 1. conditions of ePR spectroscopy were as in Fig. 1, excgpt that the modulation amplitude was 7.5 G, the temperature 7.7 K, and for A the microwave power was 2.7 microwatts.
The amplifier gain for A was 2.56 times that for B.  4. EPR difference spectrum of completely reduced Complex I (Center 1 + 2) minus partly reduced Complex I (Center 2). The samples used were those of Fig. 2, A and B, and their spectra were analogous to those of Fig. 1, C and E. The difference spectrum was obtained by use of a digital computer.
Subtraction of a spectrum showing mainly the lines of Center 2 (e.g. Fig. 1, B or C), suitably scaled, from a spectrum exhibiting the lines of both Centers 1 and 2 (e.g. Fig. 1, D or E) should yield the spectrum of Center 1, which at these temperatures is only observed with the more readily reduced Center 2. Thus, a spectrum approximating that of Center 1 was obtained by computer subtraction (Fig. 4; the lines outside the area of g = 2.05 to 1.90 are ignored for the present purposes). In turn, admixture of this spectrum wit.h that of Cent.er 2 at appropriate intensities yielded spectra simulating different degrees of reduction or presence of Centers 1 and 2 (Fig. 5). These simulated spectra were used in the quantitative evaluation of spectra obt.ained from experimental samples, where the centers were reduced to a different extent.

Resolution of Signals from Iron&&r
Centers 3 and &Thus far the resonances toward low and high field have not been considered. Their amplitude iucreases roughly in parallel during the tit.ration (Fig. 1)  FIG. 5. Digital computer synthesized EPR spectra of mixture of reduced iron-sulfur Centers 1 and 2 in the ratios indicated in the figure. The spectrum used for pure Center 1 was the difference spectrum of Fig. 4; the spectrum used for pure Center 2 was a spectrum of the sample of Fig. 2A. This same spectrum had been used to obtain Fig. 4. The contribution of Centers 3 and 4 and radical signals at g = 2 should be ignored in the spectra presented.
readily saturated at high microwave power.
Thus these resonances do not show the behavior of Center 1 or 2 and must be due to one or more additional species. On close inspection of Fig. 1 one can see that, as reduction proceeds, the line at g = 2.10 is slightly shifted downfield (g = 2.100 to g = 2.103); similarly, the line at g = 1.862 is shifted upfield.
While these shifts occur, the intensity of the resonances at g = 2.10 and 1.862 increases, but that at g = 1.886 remains constant. Such behavior may be interpreted either as indicative of two separate, sequentially reduced iron-sulfur centers with partly overlapping signals, or as an indication that there is a change in the environment (e.g. change of magnetism in an adjacent acceptor) of the iron-sulfur center under observation.
We conclude that there are two distinct but similar iron-sulfur centers represented by the lines at g = 2.10, 1.886, and 1.862, for the following reasons. (a) The shifts of line positions do not occur after the signal has reached its full intensity, but when the signal is approximately at 50% of its maximal size. (6) During the second stage of the reduction (>50%) when the shift of the lines occurs, only the lines at g = 2.10 and g = 1.862 increase, not the line at g = 1.886. (c) Double integrations of these lines account for as many iron-sulfur centers as do the lines of Centers 1 and 2 together, i.e. approximately two per flavin (cf. Table II). (d) When the position of the line at g = 2.10 is carefully measured and plotted against the quantity of reductant added and-the combined signal height at g = 2.10 as in Fig. 6, sequential reduction of two components with different oxidation-reduction potentials is clearly indicated.
Since in these measurements we were dealing with small changes in g factor, we also measured for every sample the line position of the resonance at g = 2.054. For this resonance we found the field position to be constant to within &0.0003 g. The center reduced first in the experiment of Fig. 6 we call Center 3, the other 4. The explanation that a shift of lines in a single center occurs could only be upheld, if gross heterogeneity in the preparation is assumed, Sequence of Reduction of Iron&&r Centers 1 to 4 in Titrations--The relationship of the degree of reduction of Complex I and the development of the lines of the indivdiual iron-sulfur centers is shown in Fig. 7 for all four centers. From this plot it appears that in Complex I, Center 3 has the highest oxidation- The experiment is that of Fig. 1. Each point on the abscissa corresponds to a separate sample. Percentage increase of the g = 2.10 peak, taken to represent reduction of iron-sulfur Centers 3 + 4 (0) and measured field position of this peak (A), are plotted on the ordinate and electron neq of DPNH added per mg of protein on the abscissa. The degree of reduction of Center 1 was evaluated by comparing the observed spectra with a sequence of computer synthesized EPR spectra, of which Fig. 5 shows examples. The notation used is: (C), iron-sulfur Center 1; (O), iron-sulfur Center 2; (A), iron-sulfur Centers 3 and 4. reduction potential closely followed by Center 2. Center 4 is reduced before Center 1. If the contributions of Centers 3 and 4 are not resolved, the sequence 2 > 3 + 4 > 1 is obtained, as previously reported (4). This order is seen with more complex preparations when the poorer signal to noise of the spectra does not permit the resolution of Centers 3 + 4 (cf. Ref. 5).
Quantitative Relationships between Components of Complex I- Table II presents the results of integrations of the EPR signals as outlined under "Materials and Methods" and as indicated in the last column of this table. In a number of such integrations the ratio of total centers to flavin was found to be between 2.5 and 4 to 1 with the concentrations of the individual centers being approximately equal. Thus it appears that there is present one iron-sulfur center of each kind per flavin. If the centers were of the plant ferredoxin type, i.e. with 2 iron atoms per center, this would mean that we can account for close to one-half of the iron found by chemical analysis in Complex I, 12. 16 to 20 g atoms per mole of flavin; if there were 4 iron atoms per center, as in ferredoxins of the clostridial type, the centers represented in the signals would account for 80 to 100% of the iron.a Iron-sulfur center This refers to one of the figures of this paper which shows a spectrum typical of the one integrated.
b The outer lines at high and low field assigned to Centers 3 and 4 were omitted in this integration.

Kinetic Studies on Reduction of Iron-Suljur
Centers-It was not possible to determine the time sequence of reduction of the iron-sulfur centers by an excess of DPNH, even when the reaction was carried out at 4", because all centers were reduced within 6 ms (not corrected for freezing time), the earliest reaction time available with our apparatus.
However, with APDPNH the sequence of signal appearance of Centers 1, 2, and 3 + 4 can be resolved, particularly when the reaction is carried out at 4". The data from such an experiment, are plotted in Fig. 8. A sample incubated with DPNH for 1 min was used as the 100% reduced reference. It is apparent that the potential of Center 1 is sufficiently low that APDPNH (E'o = -248 mv) is unable to reduce this center more than 50%. The sequence of signal appearance is: Center 2, 3 + 4, and 1. We have attempted the resoIution of the contributions from Centers 3 and 4. The data suggested that the signal of Center 3 appeared before that of Center 4, but the signal to noise ratio was not sufficient to consider this as established.
Resonances of Subfractions of Complex I-The EPR spectra after reduction of the three fractions obtained by resolution of Complex I according to the procedure of Hatefi and Stempel (8) are shown in Fig. 9, B to D together with the spectrum of reduced Complex I (A). Apparently the spectrum of the ironprotein fraction is largely that of a single species, similar to Center 2. The spectrum of the flavoprotein fraction (Fig. 9D), however, is that of at least three species as suggested by studies of both power and temperature dependence of the individual lines. One of these species is the iron-protein as represented in Fig. 9C. The fractions were analyzed for iron and labile sulfide and double integrations of the EPR signals were carried out. The number of unpaired electrons accounted for in the signals were 5 and 15$&, respectively, of the iron or labile sulfur atoms found in the iron-protein and the flavoprotein fraction. Thus, the iron-protein fraction seems to contain much more EPR undetectable labile sulfur and iron than the starting material, Complex I, in which the ratio of total labile sulfur or iron atoms to the number of unpaired electrons accounted for in the EPR signal lies between 4 and 6.5. Inasmuch as these findings point to the possibility that fractionation of particles may introduce artifacts, we have examined the EPR spectra of whole heart tissue frozen immediately after the death of the animal. Since these experiments also have a bearing on the material presented in the accompanying paper, these spectra are shown there in Fig. 6 (5). The major resonances in the range of g = Complex I, 94 mg of protein per ml and dissolved as for Fig. 1, was mixed with 39 mM APDPNH at a ratio of 5:l. The APDPNH was dissolved in the same buffer and 2 M Tris base was added to bring the Tris concentration to 0.2 M and the pH to a value of 9.1. Each point on the abscissa corresponds to a separate sample. The data are plotted as percentage reduction (ordinate) versus time elapsed between mixing of Complex I with APDPNH and freezing of the sample (abscissa). The degree of reduction of the centers is evaluated as in Fig. 7. The notation used is that of Fig, 7 The protein concentrations were 13.0, 19.8, and 10.1 mg per ml for B, C, and D, respectively.
The samples for B to D were reduced with an excess of solid dithionite. The conditions of EPR spectroscopy were as in Fig. 1. ubiquinone-reductase.
Understandably, there are additional resonances present as will be discussed (5).
Apparent Lability of Iron-rSulj%r Center I--In earlier work on Complex I unexplained EPR signals between g = 2 and 1.94 were seen at 90 K, when the preparation was exposed to DPNH for several minutes (6). It was originally thought that, t,his 2.2 to 1.8 are indeed those of the iron-sulfur centers of DPNH might be due to partial reoxidation, but Kawakita and Ogura FIG. 10. EPR spectra of Complex I exposed anaerobically to DPNH for various lengths of time. Complex I, 18.2 mg of protein per ml and dissolved as for Fig. 1, was reduced with 130 neq of DPNH and frozen at 1 min for-A. B, the sample of A was kept anaerobically at 0" for 10 min and refrozen.
C, the sample of B was thawed and left at 0" for another 30 min. D, the sample of C was thawed and left at 0" for a further 120 min. E, the spectrum of the sample at the stage of C above recorded at 42 K. The conditions of EPR spectroscopy were as for Fig. 1 except that for E the temperature was 42 K.
(16) showed that these signals probably arose on deterioration of the protein. It may be recalled that Beinert et al. (17) had shown that soluble DPNH dehydrogenase deteriorates when exposed to DPNH for long periods. An alternative suggestion (18) that the unexplained signals were due to molybdenum in the enzyme was not substantiated (19). By spectroscopy at 13 K we have essentially confirmed the suggestion of Kawakita and Ogura (16) that the enzyme deteriorates when exposed to DPNH. Typical records are shown in Fig. 10. The spectrum of reduced Complex I exposed to DPNH for only one min before the sample was frozen is shown in A. When the sample ,was thawed and further exposed to DPNH (Curves B to E) the resonances due to Center 1 disappeared and new resonances appeared. Contrary to the resonances from the intact iron-sulfur 1927 centers, these resonances are readily saturated with high microwave power and can easily be detected at 90 K (6). It thus appears that Center 1 is the most labile structure in Complex I in the presence of substrate and that the previously observed sensitivity of DPNH dehydrogenase toward DPNH (17) is probably due to this lability. A possibly related observation is, that Complex I cannot be effectively reduced with dithionite. We found that only Centers 2 and 3 are reduced; we are not certain about Center 4, but the signal of reduced center 1 was never seen after addition of dithionite, so that the reduction appeared to go only to the state of Fig. 1C.