Molecular Mechanisms of Mitochondrial Genetic Activity EFFECTS OF ETHIDIUM BROMIDE ON THE DEOXYRIBONUCLEIC ACID AND ENERGETICS OF ISOLATED

When isolated, respiring mitochondria of Saccharomyces cereuisiae are exposed to ethidium bromide (EtdBr) they are capable of catalyzing the following discrete reactions. fragments

(Received for publication, April 12, 1974) ROBERTO  [Etdbr, .DNA'] + (ATP) nuclease(s) * fragments + . . . (2) (Reactions 1 + 2) mtDNA (3) + mEtdBr + (ATP) + fragments + . . . nATp lEtdBrm'DNA'l , nADp + ,#. (4) In this scheme (EtdBr,.DNA'] represents a novel, stable derivative of EtdBr, linked (probably covalently) to mitochondrial DNA fragments (DNA') which, as isolated, have a mean mass of 12.5 x lo6 daltons. This intermediate had previously been shown to be formed in vivo and in vitro with kinetics of appearance and degradation consistent with it being the first product in the mutagenic sequence known to be initiated by EtdBr. Experiments using respiratory inhibitors and uncouplers suggest that Reaction 1 depends on an energized state of the inner membrane, while Reactions 2 and 3 are driven by ATP, either added externally or generated by oxidative phos- K, disrupts this link by providing an alternate pathway for the activation of ATPase and leads to an immediate cessation of Reaction 2. Euflavine, a known antagonist to EtdBr in vivo has no effect on any of the reactions by itself, but blocks Reactions 2, 3, and 4 when added simultaneously with or subsequent to EtdBr.
Strains of S. cerevisiae known to be resistant to mutagenesis by EtdBr all exhibit a lowered rate of Reaction 3 but for different reasons: some exhibit a lowered capacity to form [EtdBr .DNA'], others to bring about its degradation. Obligately aerobic, petite negative yeasts are known not to be affected by EtdBr in a permanent fashion: mitochondria of all these strains show a complete inability to catalyze Reaction 1.
The phenanthridinium dye ethidium bromide was originally synthesized and studied as an effective trypanocide (1,2). More recently this compound has received an ever-increasing amount of attention as a specific probe and reagent for extrachromosomal genophores in bacteria (3), algae (4,5), fungi (6-g)) protozoa (1, 2, lo), and mammalian cells (11)(12)(13)(14)(15)(16)(17). We are interested in the highly specific and effective mutagenic action of EtdBrl in facultatively anaerobic yeasts such as Saccharomyces cerevisiae. This mutagenesis eventually produces stable, respiration-deficient mutant strains eshibitiug au estrachromosomal, non-Mendelian mode of inheritance, now known to have resulted from the deletion of variable, but large segments of their mtDNA (6-9, 18, 19). Although it is thus the molecule ultimately affected (and is known to be susceptible to conformational alterations by virtue of its ability t,o bind E;tdUr in an intercalative mode (20-24)) mtDKA need not be the exclusive or even its primary target in the mutagenic sequence. The dye is

RESULTS
EtdBr Elicits Novel Series of Reactions in Isolated Jlitochondria -Isolated mitochondria of a number of respiration sufficient strains of S. cerevisiae, whether haploid or diploid, can carry out both the formation and degradation of EtdBr'DNA (Reactions 1 and 2) (38). This is demonstrated with mitochondria of haploid strain IL-16, suspended in phosphate buffer, in Fig. 1A (thick lines). Since the latter reaction exhibits an absolute requirement for ATP, either applied exogenously or generated internally by oxidative phosphorylation (here with succinate), combination of Reactions 1 and 2 can be regarded as Reaction 3, a DNAse that depends on both, the presence of EtdBr and probably ATP.
Such ATP-dependent enzyme activities using DNA as a substrate are well known in eukaryotes where they form part of systems responsible for the replication, repair, recognition and recombination of their DNAs (e.g. Refs. 41-44). These systems are frequently also capable, as a corollary, to bring about the converse reaction, and thus act as DNA-dependent ATPases, usually with the amount of phosphate liberated exceeding the number of phosphodiester bonds cleaved (e.g. Refs. 45 to 49). These considerations have prompted us to search for this aspect of the reaction as well. As can be seen from the data represented by thin lines in Fig. 1, A and B, the presence of EtdBr s DNA, or more simply (in susceptible strains) EtdBr added to mtDNA elicits a stimulation of the intrinsic ATPase and brings about a substantial hydrolysis of added ATPase (Reaction 4).
This is also the range in which we have previously established a satisfactory correlation in Go . At 1 = !10 min, 100 rnM succinate or 1 mM ATP was added to the sample and both DNA-associated, acid-precipitable activity (thick lines) and ATI'ase activity (thin lines) were assayed. Total activity at t = 90 min correspond to 2.7 X lo3 dpm of "C and 1.7 X 10" dpm of 3H.
Base-line ATPase activity (ATPaseo) corresponds to 4.5 pmoles of Pi liberated X mg of protein-' X 10 mine*; protein concentration was 5.9 mg ml-'. B, same conditions as above, except that addition was either reversed in order (i.e. succinate at t = 0 and EtdBr at t = 60, 0, 0, A) or simultaneous (i.e. both succinate and EtdBr at 1 = 0, l , n , A).  (37) we had felt constrained to use 13H]EtdBr at a relatively high concentration in order to make certain that the product formed would be of sufficiently high specific radioactivity.
As shown in Table II even at 10 PM the dye is present in sufficient excess to permit Reactions 1 to 3 to proceed at maximal rates. With EtdBr in excess, the extent of Reaction 1, as well as the total amount of EtdBr taken up by mitochondria, is strictly proportional to the amount of mitochondrial protein (and hence of mtDNA) added. This is most clearly demonstrated by the experiment in the bottom half, which employed a set of isogenic strains (2180A and B (haploids), HOH (diploid)), previously used in this laboratory for precise determinations of the amount of mtDNA and mitochondrial mass per cell (50). As can be seen, the amount of EtdBr covalently bound to the between the kinetics of mutagenesis and of formation of [3H]-mtDNA of the diploid is twice that bound to either of its isogenic haploids.
In all the experiments the total amount of EtdBr taken up into mitochondria exceeded the amount bound covalently by about a factor of ten and amounted to about 2 x lo* dpm mg-' of protein.
Presumably this excess is bound in part to mtDNA in the intercalative mode, and, in part to membranes (see introductory section). The data of Table 1 I permit a calculation of m, the number of EtdBr molecules incorporated into mtDNA as a result of Reaction 1. Assuming a value for haploid cells of 4.2 pg of DNA X mg-i of mitochondrial protein (8, 9, 50), a molecular weight of 5.0 x 10' for mtDNA (8,9), and a specific activity of 5 mCi mmole-i for the EtdBr used, the values calculated for m are of the order of 1 EtdBr molecule bound for each 120 deoxynucleotides. This corresponds to 5 nmoles ml+ of suspension and suggests that even at 10 pM the dye is present in sufficient excess to saturate all available binding sites.
We have also determined the value of n, the number of molecules of externally added ATP that can become hydrolyzed in the course of Reaction 4. As shown in Table III, the ratio of n:m equals 800 and hence n = 6.7 per phosphodiester bond split. This type of "nonproductive" ATP hydrolysis appears to be the rule for the behavior of bacterial recombination and restriction nucleases (46, 49).

Requirements
for Reactions 1 through S-The data presented in Fig. 2 and Table IV clearly indicate that, with the standard, actively respiring, mitochondria used in this investigation, (i.e. obtained from early stationary cells, grown on a nonfermentable carbon and energy source, such as lactate) Reaction 1 does not require free ATP, either added to the medium or generated internally.
Furthermore, this reaction probably does not even depend on the small amount of intramitochondrial ATP present at the start of the experiment since it is unaffected by the addition of excess glucose plus hexokinase and is resistant to atractylate, Dio 9 and oligomycin3 However, since it is responsive to uncouplers, the reaction is not independent of the energy generating properties of the mitochondrial membrane. In contrast, Reaction 2, and in consequence Reaction 3 as a For the purpose of this discussion we assume the specificities and modes of action commonly postulated in mitochondrial energetics (e.g. Refs. 51 to 55). DNA-associated, acid-precipitable EtdBr was assayed each 30 min as described.
+oligomycin (20 pg per ml) +u, +Dio 9 (20 pg per ml). Total counts incorporated in control sample correspond to -1400 cpm. well, exhibits an absolute requirement for (a) a supply of internal ATP (or compounds in ready equilibrium with it; see below), capable of interact.ing with t.he membrane-bound system for energy transduction, and (b) EtdBr.DNA. This is shown by the results of experiments summarized in Fig. 3. The reactions in question can be initiated in two ways: (a) either with added ATP or ADP, presumably by virtue of the presence of adenylate kinase in the intermembrane space of yeast mitochondria (56), or with GTP (or GDP), due perhaps to phosphate transfer by nucleosidediphosphate kinase; the resultant adenine nucleotides will then have to be transferred across the inner mitochondrial membrane by the transporter system (57-59) ; or (b) these reactions can be made dependent on the ener,T generating oxidation of members of the citric acid cycle. In the former case (initiation with ATl') the reactions are blocked by inhibitorsa of the adenine nucleotide transporter (atractylate) ; the F1 AT1 ase (Dio 9) ; its link to the membrane (oiigomycin); and uncouplers (Cccl;), but they are not affected by inhibitors of substrate oxidation (malonate) or eiect.ron transport (antimycin A). Conversely, when succinate is used, the reactions are blocked by these two inhibitors, as well as by the inhibitors of the energy transducing system, but not by atractylate.
Identity of Requirements for Reactions 3 and d--Implicit in the presentation so far has been the assumption that Reaction 4, the relatively massive hydrolysis of added ATP represents in fact (or is at least activated by) another (i.e. the converse) aspect of Reaction 3. This assumption suggests that the ATP requirement of Reaction 2 is satisfied by and represents one part of Reaction 4, and that the mitochondrial ATP synthetase or ATPase complex is the enzyme system implicated in both reactions. That this may be a valid hypothesis is indicated by the identity of t.he requirements and inhibition patterns, including their quantitative aspects, for the two reactions.
As shown by a comparison of the data of Figs. 3B and 4, both of which deal with events driven by succinate-dependent oxidstive phosphorylation, exposure of mitochondria to the uncouplers CCCP and colicin which prevents the formation of the covalent modification product (and of the DNAse) probably does so by virtue of their prior activation of the ATPase; oligomycin, Dio 9, and antimycin A At this point (t = -2O), drugs were added to various aliquots of the sample and incubation continued for another 20 min, at which time 100 mM succinate was added to each sample (t = 0). ATPase activity was then assayed as described under "Experimental Procedures" using ATP as an added substrate.
Base-line ATPase activity (ATPaseo) was 4.5 pmoles of Pi liberated X mg of protein-l X 10 min-'; protein concentration was 3. block DNAse and prevent activation of the ATPase and finally, atractylate has no effect. It should also be emphasized that, alt'hough the concentrations of some of the inhibitors used, especially oligomycin and atractylate, appear high (even on a protein basis) compared to those commonly employed uit.h mammalian mitochondria, they are comparable to or less t,han the ones reported in the literature to be effective with and utilized for particles from yeast (39, 59-62).
There was no effect on any of the reactions of the solvents used to add the inhibitors.
However, frequently the strongest evidence concerning the validity of any biochemical hypothesis is provided by the use of appropriate mutants with a well defined phenotype. Below we present the results of studies on the various reactions with two classes of mutants: (a) those isolated for, and exhibiting, alterations in the mitochondrial ATPase; (b) those with alterations affecting DNA repair and recombination.
Oligomycin-resistant Mulunts Exhibit Same Pattern Also in Their EtdBr. DNAse as Well as in Their EtdBr. DNA-induced ATPase-We have examined two representative mutants, both resulting in oligomycin resistance of their mitochondrial ATPase.
While both also manifest a mitochondrial mode of inheritance of this trait, they differ in the nature of the primary lesion responsible.
The first is a representative of a particular class of mitochondrial mutants isolated and characterized among others by Avner and Griffiths (63) Table I); effect of inhibitors. 3 Mitochondria isolated from D243-4A (A), 73/l (B), D243-4A-On-4 (C), or 73/l/p2 (D) grown on 3% galactose were incubated in Buffer T with 100 JAM [$H]EtdBr for 90 min. At this point, each sample was divided into aliquots and to each one inhibitors were added in the presence of succinate or ATP. Degradation (Ihick lines) and ATPase activity (thin lines) were followed as described.
Its phenotype is similar to one described by Goffeau et al. (39), for a nuclear mutant of Schizosaccharomyces pombe, in showing a pleiotopic deficiency in cytochrome oxidase (cytochrome UQ), some of the cytochrome b, as well as of the susceptibility of its mitochondrial ATPase to oligomycin. Its inheritance, however, unlike the S. pombe mutant is mitochondrial by all the criteria ordinarily employed' (65).
The response of mitochondria isolated from these two mutants, their parent wild type, as well as of 73/1-p-Z, a rho0 (mtDNA deficient) derivative obtained from 73/l by means of prolonged treatment with ethidium bromide (6-9, 18, 66), are shown in with only the response to succinate blocked by the addition of antimycin A or of malonate; (e) Reaction 4, in both its base-line and activated levels, is sensitive to oligomycin and Dio 9; and (f) the same inhibition pattern is exhibited by Reaction 2. In the OR-4 mutant (Fig. 5C), the results are similar, except for a loss in oligomycin (but not Dio 9) sensitivity in all reactions. These results strongly implicate the mitochondrial ATPase complex as a participating entity. Mutant 73/l (Fig. 5B) exhibits a similar loss of oligomycin sensitivity, but as might have been predicted for a mutant deficient in its mitochondrial elect,ron transport system, Reactions 2 and 4 can be elicited only by ATP and not by succinate.
There is also some reduction in extent but not in rate of Reaction 1. Finally, the loss of mitJochondrial DNA in 73/1/p2 (Fig. 50) renders the organelle incapable of catalyzing any reaction other than a base-line ATPase (61,62).
Recombination and Repair Mutants Show Altered Responses Not Only in Function and Degradation of EtdBr. DNA but Also in Their Induced ATPase-We have, examined representatives of three classes of such mutants.
Those in the first class, isolated by Moustacchi (67)) exhibit a heightened suscept,ibility t.oward ultraviolet light in producing cytoplasmic p-mutants, without any effect on nuclear mutation rates. However, the two strains selected differ in both, the kinctics of their response to other mutagens such as berenil (68) or EtdBr" (27, 28), as well as in their mode of inheritance.
The first, ZIVS p5 evinces enhanced resistance to mutagenesis by these agents, as compared to N123, its wild type parent, and exhibits a normal (nuclear) pattern of segregation.
The second, uvs ~72 shows enhanced susceptibility to such mutagenesis and is abnormal, probably with a mitochon drial contribution to its inheritance.", 6 Mutants in the second class are all nuclear in inheritance and ret-(recombination deficient) for nuclear markers.
The particular examples (e.g. Z140-51C2C4) have been chosen for their enhanced resistance to mutagenesis by EtdBr6 (69). The last is an example of a class of nuclear mutants (rud), particularly susceptible to radiation induced damage both with x-ray and ultraviolet (70). The strain chosen was selected for enhanced resistance to EtdBr. 6 Results of some experiments are summarized in Fig. 6: Fig. 6A presents data relevant to the reactions with isolated mitochondria and Fig. 6B the rates of mutagenesis on intact cells under comparable conditions.
We observe in the first series that mitochondria of uvs p5 appear normal in rate and extent of Reaction 1, the formation of EtdBreDNA, the modification product. They are, however, completely incapable of degrading it and are also grossly deficient in eliciting the succinate or ATP-elicited stimulation of ATPase in response to its formation.
We conclude that, (a) a nuclear mutation produces a defect in the ability of mitochondria to excise a particular form of damage produced in their DNA by EtdBr, and (b) that this deficiency is reflected simultaneously by an inability to produce the resultant stimulation of, an otherwise normal, mitochondrial ATPase. The response of ~72 mitochondria is quite different and pleiotropic: it is expressed in the extent of Reaction 1, since the level of the modification product produced is almost tripled; in the rate of both Reactions 2 and 4, both of which are accelerated; as well as in the final level of stimulation achieved in Reaction 4, which approaches that obtained with the most efficient classical uncouplers.
The pattern exhibited by the mitochondria of the EtdBr resistant ret-mutant 2C4 (ret 5) appears to be the converse of that FIG. G. Formation and degradation of the EtdBr containing modification product of mtDNA and its relation to the mutagenic action by EtdBr (see Table I  EtdBr was assayed at t = 20,40, GO, and 90 min. At this point, 100 mM succinate was added to each sample and both degradation (heavy lines) and ATPase activity (thin lines) were followed as described.
Symbols are defined under A. shown by ~5. Like the latter it exhibits a lowered rate of Reaction 3 but for a different reason. Relative to its wild type (Z140-51C) it forms au abnormally low amount (<25%) of the modification product (see Fig. 6 legend).
However, both its rate of formation and of its degradation, as well as the rate and extent of the resulting AT1 ase stimulation are normal.
1 he results with rad 6' are even more spectacular.
Under the conditions studied it is completely resistant to mutagenesis and in this regard resembles uvs ~5. However, in its lack of ability to incorporate EtdBr into its DNA it resembles ret 5 but in an exaggerated fashion: the level found is < 10% that of the wild type (53 versus 580 cpm X mg-' of protein) These experiments lay the groundwork for an interpretation of the genetic results in molecular terms, and provide additional evidence of the tight coupling between the mtDKAse induced by Etdfir and the resultant activation of the mitochondrial ATPase complex. Eujkwin Prevents Reactions 3 and &As mentioned in the introduction, euflavine is an intercalating dye similar in its structural aspects to E;tdUr and like it an effective mut.agen (Ref. 29 and citations therein).
However, it must differ in its mode of action since its activity is restricted to the buds of dividing cells (711. Furthermore, with starved cells in buffer we have shown it to act as a potent competitor in preventing the expression of the mutagenic action of EtdUr (27-29).
With isolated mitochondria, euflavine, at concentrations used in these in vivo experiments IL-16 mitochondria were suspended in Buffer T. 3H-EtdBr (100 j&M) was added at t = 0 and DNA-associated, acid-precipitable EtdBr was assayed each 30 min. At t = 90, euflavine was added (25 PM) simultaneously with 1 mM ATP. Degradation (thick lines) and ATPase activity (thin lines) were followed as described (A--A).
A control without euflavine was run simultaneously (O--O). When euflavine was added at t = 0 (i.e. together with EtdBr), the results were identical with those shown for addition at 1 = 90. Total counts incorporated in the sample corresponded to 1400 cpm. Base-line ATPase activity (ATPaseo) was 4.4 wmoles of Pi liberated X mg of protein-l X 10 mine'.
(25 PM), is completely ineffective in eliciting either the breakdown of (labeled) mtDNA or the activation of ATl'ase ( Fig. 7). It is also completely incapable of preventing the formation of EtdBr,DNA.
However, once this compound is formed, the subsequent (or simultaneous) addition of euflavine leads to a complete protection against its ATP-induced breakdown, the ATE and UdlSr-dependent DNAse, and the concomit,ant activation of ATPase.
Colicin K Acts as Uncoupler-Certain bacteriocidins such as colicin K are believed to act by virtue of their ability to interfere with integrated membrane function of susceptible cells, perhaps by producing a configurational alteration that prevents proper energy coupling (53). This mode of action resembles t,he one postulated for lipophilic uncouplers of oxidat,ive phosphorglation such as CCCP.
It therefore appeared of interest to compare t,he action of these two inhibitors in the current system. The data of Figs. 2 to 4 indicate that these two agents are in fact qualitatively similar, both in preventing the formation of EtdlSr . DNA and its degradation.
This result is probably the consequence of an activation of t.he mitochondrial ATPase in a competitive fashion, presumably by altering the energy coupling device. Two corollaries emerge from these studies: (a) that colicin K appears to be a highly effective uncoupler of mitochondrial energy transduction, an interesting effect that warrants further investigation, and (b) that uncoupling by CCCPT (or colicin) and by E;tdUr.
' The fluoroderivative FCCP is effective at concentrations ten times lower than those used with CCCP. were added. DN.4-associated, trichloroacetic acidprecipitable total EtdBr (thick lines) and ATPase activity (thin lines) were assayed as described.
At t = 30 min, another aliquot was removed, 10-S M CCCP was added (A--A), and the same measurements were performed.
Control sample (O--O) was also assayed for these two parameters.
DNA requires the participation of the same mitochondrial entity and therefore appear to be antagonistic and Ferhaps even mutually exclusive.
Further corroboration of this hypothesis comes from two types of experiments also summarized in Fig. 8. In the first me show that when operational "uncoupling" by EtdBr.DNA has reached its final level, the addition of CCCP only produces a small additional effect; in the second, that the addition of CCCP, once degradation of EtdBr. DNA has already been initiated, but before it is completed, leads to an immediate cessation of this reaction, coincident with a substantial enhancement of the ATPase .
E$ects of Other Modulalors-In this section \\\e examine the effects of a number of agents that have either been claimed to affect t.he extent of mutagenesis by EtdBr or that may reasonably be expected to do so (Table V). Among the first group are cyclohesimide (72) and galactose8 (73)) both of which can decrease the level of mutagenic expression; among the second are caffeine, which is an inhibitor of certain reactions concerned with DNA repair in yeast (74)   0 The reduction in amount of EtdBr incorporated in galactosegrown cells of other strains examined also shows a 507, inhibition of Reaction 1. has been demonstrated in this laboratory to affect mitochondrial mutagenesis by berenil (68). We have also examined cyclic 3':5'-AMP, which has been implicated in the modulation of catabolite repression in yeast (7, 76, 77) as it does in E. coli (SO,81). With isolated mitochondria, only the addition of galactose produces a significant effect, by reducing the extent, but not the rate, of formation of EtdBr . DNA M ithout any influence, however, on any of the other reactions.
This effect of galactose has also been observed for strains D243-4A and 73/l, 11 hen u e compared mitochondria from glucose and galactose grown cells. Petite Negative Yeasts-Yeast species are frequently classified as either petite negative or petite positive (7-9, 83, 84). The latter include Saccharomyces and related l'acultative anaerobes. Among the former are obligate aerobes such as Can&da, Torulopsis, Kluyveromyces, and Hansen&a. In such strains htdUr not only fails to induce any &able respirat.ion deficient mutant,s, but,, in spite of being capable of exerting its cust,omary inhibition of mtDNA and RNA synthesis, the agent does not produce a permanent, irreversible loss of mtDNA even upon extended growth in its presence9 (83)(84)(85)(86).
One would therefore anticipate the mitochondria isolated from such strains to behave quite differently from those of S. cereoisiae. We have tested particles from Hansenula tingei, Torulopsis utilis, and Kkuyveromyces l&is.
The results obtained confirm earlier, less complete experiments with intact cells and suggest the complete absence of any of the Reactions 1 through 4. This failure is not due to some obvious alteration in the mitochondria obtained from these cells, either intrinsic or as a result of isolation, since the particles take up EtdBr and exhibit base-line mitochondrial ATPase levels as high or higher than those described earlier for S. cerevisiae. We are therefore probably justified in concluding that (a) occurrence of Reaction 1 is a prerequisite for all the other three reactions and EtdBr'DNA is an obligatory intermediate, and (b) the resistance of these cells to EtdBr is the direct consequence of this inability t.o form the intermediate and the resultant lack in the degradation of their mtDNA.

DISCUSSION
Proposed Scheme-The hypothesis we wish to propose in order to account for the findings of our investigations is the following: (a) That when mtDNA of a sensitive strain of S. ceretisiae, in its natural environment within the inner membrane space of the mitochondrion, is exposed to EtdRr it is converted to large (MT = 12.5 x 106) fragments that contain the dye in covalent linkage (about 1 molecule for each 100 nucleotides).
(b) That this product is recognized and degraded by specific nuclease(s) with an absolute requirement for intramitochondrial ATP and which requires some components of the mitochondrial ATP synthetase (ATPase) for its function.
And (c) that coincident wit'h or as a consequence of b the mitochondrial ATl'ase becomes activated (and can be assayed with externally added ATl') in a manner reminiscent of the response of this system upon exposure to lipophilic uncouplers.
The coupling device (or if preferred, inner membrane integrity) is also implicated in the first step, and t.he most parsimonious, although by no means t.he only possible, assumption to rcconcilc b and c is that (with ATI' as the driving force) they represent different aspects of the same molecular event, also linked to or through inner membrane components; b would then correspond to the actual reaction involving AT1 in stoichiometric amounts, and c to its maximal capacity, measurable only in the presence of added ATP.
The evidence for point c, that the enzyme complex responsible for reaction is the mitochondrial ATl'ase, rest,s on both, experiments with specific inhibitors and with mutants of sufficiently well defmed phenotype. This evidence appears reasonable and utilizes a standard reaction involving the liberation of inorganic phosphate from added ATl' by mitochondria previously or simultaneously subjected to the various treatments described.
The evidence for its implication in point b is somewhat weaker.
We know that externally added ATI' can drive the reaction, but so can oxidative phosphorylation using succinate, phosphate, and t,he adenine nucleotide pool inside the mitochondria.
The degradation of EtdUr 'DNA is linked to the energy coupling device; the experiments with uncouplers and inhibitors, especially in the appropriate mutants leave little doubt of that. Rut what is not certain is whether this link is solely by virtue of a continuing and stoichiomet-ic requirement for ATI' for the excision of the EtdBr and the cleavage of phosphodiester bonds from mtDNA, or whether in addit'ion or alternatively, it requires a more act,ive participation of the whole membrane-integrated complex; this remains to be established. In preliminary experiments with mitoplasts, which have lost their adenylate kinase (56), addition of ADl' but not of ATl' can drive the reaction, which has retained its sensitivity to atractylate (in complete accord with the properties expected for the transporter system (57)).
The behavior of the mitochondria of respiration deficient mutants such as 73/l and the p-petites tested earlier (34) is puzzling within this context.
They are respiration deficient because they (a) lack the ability to both reduce and oxidize cytochrome c and (b) eshibit a lesion in the coupling device (determined operationally by the oligomyciu resistance of their ATPase).
Yet they are capable of performing the whole sequence of reactions, albeit at reduced levels. These observations suggest that such mitochondria retain in their membranes all the components (at least in a rudimentary form) necessary for the four reactions and can perform them at the expense of ATP generated by glycolysis.
Relations to Biological Phenomena-The reactions described in this and our two related previous publications (34, 37) appear to provide a satisfactory molecular basis for the mutagenic action of EtdBr described in the introduction.
All the events and steps observed in uiuo have now been duplicated with isolated mitochondria.
The results with the different mutation-prone and mutation-resistant strains of S. cereaisiue, as well as with the resistant, obligately aerobic yeast species, not only provide essential confirmation of the central hypothesis, but account in a consistent manner for their susceptibility to mutagenesis by EtdBr: resistant strains are charact.erized by a deficiency in the rate or extent of breakdown of mtDNA due to Reaction 3 (sometimes because of their inability to bring about Reaction 1) and hypermutable strains carry out the degradation at an accelerated rate.
The hypothesis also provides a framew-ork wit,hin which to find an explanation of one ot.her puzzling set of observat,ions: that lesions in the coupling device, either as a result of mutation, e.g. op.1 (pet 9) (87-89) or manipulation (90) affect the rate of mutation from p+ to p-. These results have frequently beeu discussed (87,88,90) in terms of an alteration of the inner mem brane, and this may, iu fact, const,it,utc t,heir ult.imatjc causp. Rut it is equally reasonable to assume that a mom immediate manifestation may be a change in the substrate requiremcut for Reaction 2 (and 3), resulting in a degradation of mtDNA IIO longer dependent on its prior interaction with EtdRr.
,4 similar explanation may also be advanced for mutagenesis by eutlavine which is knowu to require cellular (29, 71), and hence mitochondrial, division and may at some stage of this process bring the mtDNA into a configuration similar to the one found in EtdlSr S DNA.
These questions, and whether an analogous compound can also be formed (in an AT1 requiring reaction) by bereuil arid ally1 proflavine, mutagens known to require an energy source in vivo (29), are susceptible to further inquiry.
Relation to Isolated Enzymes-The most mystifying part of the scheme is Reaction 1 since it represents an unexpected and hitherto unrecognized reaction.
What. nil1 have to be determined is whether it reflects the action of one or more enzymes or of a peculiar configuration of mtDKA perhaps conditioned by its attachment to an appropriate membrane site (91). Such sites are known to be determiuative for the rcplicat,ion of circular bacterial chromosomes and plasmids (92,93). The possibilit,y also exists that ribonucleotide tracts iu mtDNA (identified so far unambiguously only in animal cells (91,(94)(95)(96)) actually coustitute the reaction sites; EtdlSr has been reported to be capable of becoming attached to tRKA in vitro (97). The absence of the reaction in rad 6 and in petite-negative straius opens this problem to ready inquiry.
There is no lack of precedent in prokaryotes for ATFdependent nucleases, capable of recognizing specific sequences or configurations in their DNA (e.g. Refs. 41-49).
Many of these enzymes are of large size and complexity and (when functioning as ATPases) are capable of hydrolyzing A'l'l' in great excess over the number of phosphodiester bonds cleaved (i.e. t,hey carry out reactions analogous to Reactions 3 and 4). No such activity has been explicitly reported for mitochondria.
The DNAse activity studied by Faoletti et al. (98) responds not only to EtdRr but also to euflavine, and it lacks the requirement for ATP. Clearly these properties exclude it from considerat,ion as the sole enzyme responsible for Reaction 2 or 3, but not as a possible participant within a complex that may result in modification of both its specificity and its requirements.
The results with the mitochondria of uvs p5 suggest that the prot.ein affected by this lesion constitutes (part of) the DNAse in question.
If that is so this mitochondrial protein is of nuclear specification.