A units-of-evolution perspective on the endosymbiont theory of the origin of the mitochondrion.
Mereschkowsky (`10), in an entertaining fantasy, has developed the hypothesis that the dualism of the cell in respect to nuclear and cytoplasmic substance resulted from a symbiotic association of two types of primordial microorganisms.... More recently, Wallin (`22) has maintained that chondriosomes may be regarded as symbiotic bacteria whose association with other cytoplasmic components may have arisen in the earliest stages of evolution. To many, no doubt, such speculations may appear too fantastic for present mention in polite biological society; nevertheless it is within the range of possibility that they may some day call for more serious consideration.
--E. B. Wilson, The Cell in Development and Heredity
Mitochondria, chloroplasts, and other cytoplasmic entities contain nucleic acids, and this strongly suggests that they have genotypes that direct their development. It is commonly believed that they originated as independent organisms competing for a common pool of resources. If so, their original induction into complex cooperative systems is an anomalous event, not to be expected on the basis of our usual understanding of the evolutionary process. The subsequent stability of these eukaryotic cell lineages through geologic time, despite potential disruption from selection among cellular components, presents an evolutionary problem that deserves detailed attention.
--G. C. Williams, Natural Selection
In this century, biologists' perspectives on the endosymbiont theory of the origin of the mitochondrion have changed dramatically. Once characterized as "too fantastic for present mention in polite biological society," this theory is now generally accepted, largely because of extensive compilations of supporting evidence (e.g., Margulis 1970, 1981; Schwartz and Dayhoff 1978; Baltscheffsky and Baltscheffsky 1981; Gray and Doolittle 1982; McCarroll et al. 1983; Spenser et al. 1984; Cavalier-Smith 1987a; Ballantyne and Chamberlin 1988; Gray 1992; Knoll 1992). Corresponding to such broad acceptance, a shift of emphasis is also occurring; presently, the focus is on questions of not merely whether mitochondria arose from endosymbionts, but how, and by what evolutionary mechanisms. Most generally, what are the implications of the endosymbiont theory in terms of the nature of the evolutionary process and the history of eukaryotic lineages? As suggested by Williams (1992), these implications are profound and only partially explored.
A units-of-evolution framework is particularly useful for assessing the evolutionary implications of the endosymbiont theory, because such a framework explicitly allows for the occurrence of synergistic or antagonistic effects between host and symbiont (e.g., Lewontin 1970; Buss 1987; Maynard Smith 1991). The endosymbiont theory has in fact been considered extensively in the context of intragenomic conflict, that is, in terms of the evolutionary dynamics of mitochondria and their hosts in a more or less modern form (e.g., Eberhard 1980; Cosmides and Tooby 1981; Hurst 1992; Hurst and Hamilton 1992). Surprisingly, the costs and benefits of the host-mitochondria association are less frequently viewed from the level of the cells that initiated this interaction, the protomitichondrion and the primitive host cell. Yet all evidence suggests that these building blocks of the eukaryotic cell were themselves distinct cells with a long period of independent descent and with strikingly different ecologies and physiologies (Margulis 1970, 1981; Schwartz and Dayhoff 1978; Yang et al. 1985; Zillig et al. 1985; Cavalier-Smith 1987a; Ballantyne and Chamberlin 1988; Gogarten et al. 1989; Iwabe et al. 1989; Woese et al. 1990; Rivera and Lake 1992; Rowlands et al. 1994). At its inception, the interaction between protomitochondria and the cells that ultimately became their hosts may have involved not just intra- or intergenomic conflict, but crucial intercellular interactions as well. This inception thus may have entailed a fabric of interactions much richer than merely gene selection; cellular-level features, particularly ecology and physiology, should be considered to assess the full implications of the endosymbiont theory.
I elaborate on this theme of intercellular interactions, and in doing so, I characterize the implications of the endosymbiont theory in the context of the cellular mechanics of modern eukaryotes. My central premise is that the cellular and molecular natural history of modern eukaryotes will remain obscure as long as the selective forces that shaped this natural history remain unexplored (for a compelling statement of this premise, see Buss 1987). To this end, I present a units-of-evolution perspective on the endosymbiont theory, focusing on the probable ecological and physiological aspects of the host and symbiont cells and the possible nature of their initial interactions. I suggest that a key part of this initial interaction was the manipulation of the host's life history by mitochondria. Mitochondria may have produced ATP and oxidants in such a way as to trigger asexual and sexual reproduction of the host, thus ensuring the transmission of mitochondria.
It is axiomatic that the establishment of the stable association that characterizes modern eukaryotic cells necessitated certain sequelae to this initial interaction (Buss 1987; Maynard Smith 1991). I propose that features of modern eukaryotic cells can be interpreted in this light. Specifically, key features of these cells may be derived from an evolutionary ritualization (e.g., see Kessel 1955) of aspects of the initial host-mitochondria interaction, channeling eukaryotic evolution in certain consequential ways. In this context, I draw attention to the role of phosphorylation cascades in cell-division cycles and to the ubiquitous calcium fluctuations in the eukaryotic cytosol. I suggest that these mechanisms may represent a ritualized form of mitochondrial control over the cell's life history. At the same time, features of modern mitochondrial genomes (the lack of replication and transcription factors, the lack of adenine nucleotide carrier genes, and other characteristics, see Gabriel et al. 1993) constrain the expression of mitochondrial variation in the cytoplasmic environment in favor of the host cell, particularly with regard to energy allocation.
These and other putatively derived cellular mechanisms, established early in the history of eukaryotic cells, may have been invoked in later episodes of eukaryotic evolution, for instance, in mediating cell-cell conflicts as multicellular animals evolved and in modulating the life histories of these multicellular organisms. The effects of extracellular ATP, the mechanisms of second messenger systems, and the role of mitochondria in the germ plasm deserve attention in this regard. Finally, I point out that "free-radical" theories of development and aging are suggestive of evolutionary vestiges of the initial host-mitochondria interaction as envisioned here.
Discussions of the endosymbiont theory sometimes focus on suites of character states, or even particular extant taxa, as ancestor models for mitochondria or host cells, or both. Although the role of ancestor models in evolutionary theory is subject to debate (e.g., Archibald 1994), it is generally accepted that an ancestral taxon must lack uniquely derived characters. Given the billion or so years that have elapsed since mitochondria became established in eukaryotic cells, it is highly unlikely that any extant taxon will meet this criterion. Further, constructing a hypothetical ancestral form by assembling suites of shared primitive characters introduces an element of circularity and can have unintended results such as erecting paraphyletic taxa (e.g., see discussion of the "hypothetical ancestral mollusk" in Brusca and Brusca 1990).
Potentially more informative results are based on comparisons between the subkingdoms of protists, the Archezoa and Mitozoa. Mitozoa usually have mitochondria, but Archezoa consistently lack them. It has been suggested that the Archezoa comprise eukaryotic lineages that never acquired mitochondria (Cavalier-Smith 1987b; Vossbrinck et al. 1987; Muller 1988; Sogin et al. 1989). However, the Archezoa are defined on the basis of shared primitive rather than shared derived characters (e.g., the lack of mitochondria); consequently, they are not monophyletic (Cavalier-Smith 1987b; Muller 1988; Sogin et al. 1989). Further, archezoan groups may have secondarily lost their mitochondria (see discussion in Fenchel and Finlay 1994). For instance, the archezoan trichomonad flagellates share a number of biochemical characteristics with rumen ciliates, and these latter protists have clearly lost their mitochondria secondarily (Muller 1988). Finally, the energy metabolism of the Microsporidia, possibly one of the most primitive archezoan groups (Cavalier-Smith 1987b; Vossbrinck et al. 1987; Sogin et al. 1989), has not yet been studied (Muller 1988). At present, the Archezoa are thus too ill-defined and too poorly characterized to provide insight into the nature of the primitive host cell.
Clearer insight has been provided by well-supported sister-group relationships. Phylogenies based on biochemical data indicate that eukaryotic mitochondria are a sister-group of the [alpha]-purple eubacteria, while eukaryotic cells are a sister-taxon to the archaebacteria (Yang et al. 1985; Gogarten et al. 1989; Iwabe et al. 1989; Woese et al. 1990; Gray 1992; Knoll 1992; Rivera and Lake 1992; Rowlands et al. 1994). These relationships imply that (1) biochemical shared derived characters of [alpha]-purple eubacteria and mitochondria can provide insight into the nature of protomitochondria, (2) nevertheless, given the billion or so years that have elapsed since mitochondria became established in eukaryotic cells, mitochondria likely exhibit numerous uniquely derived characters, (3) biochemical shared derived characters of eukaryotic cells and archaebacteria can provide insight into the nature of primitive host cells, but again (4) eukaryotic cells are likely to have numerous uniquely derived characters. Given (2) and (4), sister-group comparisons will be of limited value. This analysis will use such comparisons only to establish the broadly accepted shared derived (and in some cases shared primitive) characters of the relevant taxa. In fact, the uniquely derived characters of mitochondria and their hosts are of greater interest here. As discussed below, these characters detailed by the extensive literature on the biochemistry of eukaryotic cells and mitochondria, suggest that the adaptive responses of host and symbiont to each other have been a major selective force in the evolution of eukaryotes.
A Units-of-Evolution Perspective
There are three central tenets of Darwin's theory of evolution: organisms vary, this variation is heritable, and this variation is selected (Lewontin 1970). Further, any biological entity that meets these criteria will evolve (Lewontin 1970 Buss 1987; Maynard Smith 1988, 1991). Several terminologies have been developed for discussing these essential ideas (e.g., Sober 1984; Hull 1988; Lloyd 1988, 1992). I will use the term "unit of evolution" in preference to "unit of selection," thus emphasizing that evolution by natural selection requires heritable variation and not just selection (Maynard Smith 1988, 1991).
As clearly articulated by Buss (1987), the history of life is a history of an elaboration of units of evolution: molecules, molecules within cells, cells within cells, and cells within organisms. At each transition between units of evolution novel selective scenarios prevail. For instance, before the endosymbiosis, the protomitochondrion was selected by the external environment alone; following the endosymbiosis, the protomitochondrion was selected by traits expressed by the host cell. Once engaged in the endosymbiosis, variant protomitochondria could influence not only their own replication rate, but the replication rate of their hosts. Further, the host could develop mechanisms to constrain the expression of protomitochondrial variation. A units-of-evolution perspective focuses on such synergistic and antagonistic effects of the two evolutionary units on each other.
From this perspective, the endosymbiont theory must take into account not only the benefits of the interaction between the host and symbiont but the costs as well, and this cost-benefit analysis must consider the relevant units of evolution, in this case the two types of cells. This approach is particularly appropriate in light of the renewed emphasis on the deleterious effects of endogenous oxidants. Because of their extreme reactivity with biomolecules, such oxidants are strongly implicated in protein, lipid, polysaccharide, and nucleic acid damage, putatively leading to senescence, various cancers, and other pathologies (e.g., Richter 1988; Ames and Gold 1990a,b; Floyd 1991; Halliwell and Aruoma 1991; Hassan and Schiavone 1991; Stadtman 1992). In extant eukaryotic cells, mitochondria are a source of endogenous oxidants; for instance, in a resting state, mitochondria release superoxide radicals, that is, [O.sub.2]- (Chance et al. 1979).
Margulis (1981; see also Margulis and Sagan 1986) proposed several models for the evolution of the intricate symbiotic relationship between mitochondria and their hosts. Using the terminology of Chapman and Gest (1983), these models generally assume that the host was anaerobic, but may have been aerotolerant. In other words, although the host may have been able to survive and grow in the presence of oxygen, it did not use [O.sub.2] for energy conversion. The symbiont, on the other hand, is assumed to have been aerobic, that is, it used [O.sub.2] as the terminal electron acceptor for energy conversion. Based on this distinction, one such model envisions a rather benign interaction between a large, anaerobic host and smaller, aerobic protomitochondria, whereas an alternative model focuses on the possible parasitic nature of the initial interaction, and the need for the host cell to protect its genetic material from superoxides generated by mitochondria. The highly deleterious nature of endogenous oxidants produced from [O.sub.2]- in modern, relatively sophisticated eukaryotic cells strongly supports this latter interpretation, suggesting that the initial "infection" of host cells by aerobic mitochondria necessarily had costs, possibly including even lethal effects on some hosts.
Considerable uncertainty surrounds the timing of the mitochondrial endosymbiosis (e.g., 2400-2800 mya, Knoll 1992; 1400 mya, Jenkins 1991; 850 mya, Cavalier-Smith 1987b). In part, this uncertainty reflects different assumptions as to the factor(s) limiting the endosymbiosis. If the limiting factor was atmospheric oxygen, the first date may be correct, for at this time oxygen attained the 1%-2% present-day atmospheric level necessary to sustain aerobic metabolism. On the other hand, if the limiting factor was the origin of a phagotrophic eukaryote, the last date may be correct. However, although these and other factors may have been necessary for an initial host-symbiont association to form, they are not sufficient to produce or maintain an endosymbiosis. A successful endosymbiosis would depend on a successful resolution of units-of-evolution conflicts.
Once contact between host cells and protomitochondria occurred, a variety of selective dynamics between host and symbiont lineages may have ensued. Some of these associations may have been detrimental to both units of evolution, some associations may have benefited both, and some may have benefited one but not the other. From a units-of-evolution perspective, it is axiomatic that persistence of any host-symbiont association would have been facilitated if selection on the higher-level unit (the host cell, including the clonal mitochondrial population) guided the development of mechanisms to convert any initially detrimental effects of the symbiosis into benefits (or neutral effects) and to protect against the origin of any new detrimental effects (from genetically variant symbionts). Persistence of this association would be further enhanced if selection on the symbiont produced mechanisms by which mitochondria could perceive, and perhaps even regulate, major features of the host cell's life history. Such communication between the host and symbiont would diminish the chances of detrimental host-symbiont competition. In general, the symbiont would be exposed to potentially conflicting selection pressures. Group selection would act on the entire population of mitochondria in a single host, whereas individual selection might favor certain genetically variant strains of mitochondria within a given cytoplasmic environment.
Thus, once extrinsic and intrinsic factors (e.g., atmospheric oxygen and cell-membrane characteristics) were permissive, a number of experiments in host-symbiont association may have occurred (and in fact may continue to occur; Fenchel and Bernard 1993). The experiment that became the modern host and mitochondria did so because of a particular suite of derived features that best resolved the units-of-evolution conflicts inherent to the endosymbiosis. These derived features and their consequences for selection on the eukaryotic cell (the host and the clonal mitochondrial population) and perhaps on even higher level units (eukaryotic populations and species) determined which host-symbiont association succeeded and which failed.
AN EVOLUTIONARY SCENARIO
Sequence data suggest that superoxide dismutase (SOD) is a shared primitive character for host and mitochondrion (Loomis 1988). The dismutation of superoxide radicals occurs by the following reaction, which occurs spontaneously and is also catalyzed by SOD:
2[O.sub.2]- + 2H+ [right arrow] [H.sub.2][O.sub.2] + [O.sub.2]
The potentially toxic hydrogen peroxide product is converted to water by either catalase or glutathione peroxidase (Chance et al. 1979). Both host and symbiont likely possessed these mechanisms to alleviate the deleterious effects of oxidants; in anaerobic cells such mechanisms may have protected against external sources of oxygen and ultraviolet radiation. Nevertheless, the levels of activity of these enzymes may differ dramatically between cells; for instance, extant anaerobic bacteria exhibit negligible catalase activity compared to aerobes (Chance et al. 1979).
Ecological considerations suggest that a similar difference in enzyme activity existed in the anaerobic host and the aerobic symbiont. Sequences of 16S ribosomal RNA suggest that mitochondria are descended from the [alpha]-purple eubacteria (Yang et al. 1985; Gray 1992). Some extant representatives of this group have a mitochondrial type of respiratory chain when grown aerobically in the dark (e.g., Dutton and Wilson 1974; Baltscheffsky and Baltscheffsky 1981). Further, the evolution of aerobic metabolism in this group is often associated with photosynthetic phenotypes (Fox et al. 1980). Protomitochondria thus probably incurred damaging endogenous oxidants from respiration and may have been photosynthetic as well, although sulfur, not oxygen, was the likely electron donor. Regardless of the electron donor, photosynthetic protomitochondria must have inhabited the photic zone. At the time that protomitochondria became endosymbionts, the Earth's stratospheric ozone shield may have been substantially less than at present (Jenkins 1991; Runnegar 1991; Knoll 1992); cells inhabiting the photic zone may thus have been subject to considerable ultraviolet radiation and continuously exposed to UV-generated oxidants (Schopf et al. 1983). Such an ecology (photosynthesis and aerobic respiration) would have placed a premium on active mechanisms to scavenge and detoxify oxidants formed both endogenously and exogenously and to repair oxidative damage to genetic material effectively (see Avise 1993).
The primitive host cell, however, is thought to have been an anaerobic, but possibly aerotolerant, heterotroph with a functional glycolytic pathway (e.g., Margulis 1981; Ballantyne and Chamberlin 1988; Fenchel and Bernard 1993). An anaerobic metabolism would shield the host from endogenous oxidants. Further, since it was not photosynthetic, the host cell need not have inhabited the photic zone and may have been protected from the deleterious effects of ultraviolet radiation by seawater (see Schopf et al. 1983). The host cell likely suffered minimal oxidative damage, and despite some activity of SOD, likely did not have well-developed mechanisms to detoxify superoxide radicals and related oxidants, nor did it maintain effective mechanisms to repair oxidative damage to genetic material.
Proposed initial associations (e.g., Margulis et al. 1976; Margulis 1981; Margulis and Sagan 1986; Cavalier-Smith 1987a) should be evaluated in terms of the central postulate that both host cells and protomitochondria could produce heritable variants on which natural selection could act and that selective pressures on the host and symbiont would not necessarily have been similar. To emphasize the point that ecological and physiological features likely characterized these selective dynamics, I will consider an initial host-mitochondria interaction in the absence of oxygen and then suggest how this interaction would change fundamentally in the presence of oxygen. This is probably an artificial dichotomy; nevertheless, it illustrates the central role of oxygen in this host-symbiont interaction.
As a consequence of their oxidant-rich ecology (respiration and photosynthesis), protomitochondria would be exposed to high mutation rates. Inevitably, these mutation rates may have overwhelmed the repair mechanisms of some protomitochondrial individuals. Although damage could occur anywhere in the genome, by chance, some aerobic protomitochondria may have had their glycolytic enzyme-coding regions damaged by high levels of oxidant-induced mutations. These damaged forms, and others, would have been generated constantly as a feature of the oxidant-rich, protomitochondrial ecology. Protomitochondria with damaged glycolytic mechanisms, but with respiratory chains still intact, would be predisposed to occupy microhabitats rich in reduced carbon compounds, that is, microhabitats already supporting an array of anaerobic, glycolytic, potential host cells. If such habitats were entirely dark and anaerobic, protomitochondria might have processed some of these compounds using the fumarate reductase complex, that is, using fumarate as the terminal electron acceptor and reducing fumarate to succinate by the membrane-bound electron-transfer system (e.g., see Behm 1991). The presence of anaerobes may have "rescued" such damaged protomitochondria in nonphotic environments, and the initial association between these damaged cells and the anaerobic host cells may have been more or less obligatory. As this biochemical association developed, the anaerobic host may have received benefit in the form of the oxidation of the reduced cofactor NADH. There are considerable biochemical advantages to freeing the reoxidation of reduced cofactors from the pathway of oxidation of substrate; for instance, a net oxidation of substrate is permitted and a portion of the pyruvate pool is made available for biosynthesis (Gest 1980; Behm 1991). Thus, if the host cell could exchange strict glycolysis (in which NADH is reoxidized by lactate production), or similar substrate-linked phosphorylation, for a symbiotic metabolism in which the reoxidation of NADH is separated from the oxidation of substrate, significant advantages would accrue to both the host and the protomitochondria (fig. 1).
This exchange of metabolites may have occurred while both cells remained free-living. However, in nonphotic environments, protomitochondria with damaged glycolytic enzymes would have exhibited intense competition among themselves for substrate. Excretion of reduced carbon compounds by the host cell may have quickly become limiting. A variant protomitochondrion that could survive inside the host cell would have had a selective advantage in obtaining these compounds. This establishment of a benign endosymbiosis, in which both host and mitochondria benefit, could have occurred (and remained benign) as long as the local environment was free of oxygen.
Clearly, this is only one of many possible scenarios; an anaerobic stage to this endosymbiosis was not necessary. Nevertheless, it is of note that anaerobiosis need not have been a major impediment to the formation of an association between host cells and protomitochondria (for further illuminating discussion of this point, see Fenchel and Bernard 1993). Moreover, the presence of oxygen likely produced a fundamentally different kind of association.
The Role of Oxygen
Once sufficient oxygen was available in the microhabitats of the host cells, even intermittently, the host-symbiont interaction may have changed dramatically. Given their history as photosynthetic aerobes, protomitochondria likely could tolerate to a greater degree than host cells the endogenous oxidants produced by respiration. Further, aerobic respiration would produce roughly an order of magnitude more ATP than either glycolysis or anaerobic respiration (e.g., Bryant 1991). These dual imbalances would have had a powerful effect on subsequent selective dynamics (fig. 2). The protomitochondria may have used their aerobic metabolism to manipulate their hosts' life history to their own advantage. A mutant that was "leaky" for ATP would have increased the availability of ATP within the host cell. Host cells would have been selected to increase their rates of cell division, thereby providing additional environments for the symbionts (fig. 3a). In this way perhaps, a risky host-to-host dispersal stage of the protomitochondrion life history was eliminated and a competitive advantage over other free-living protomitochondrial lineages was maintained. On the other hand, deleterious mutations resulting from oxidative damage to the hosts' DNA would have selected for an increase in the frequency with which the host cells indulged in genetic recombination (Margulis and Sagan 1986). In a finite population reproducing by asexual fission, sexual recombination coupled with selection allows deleterious mutations to be eliminated from the population (e.g., Bell 1988; Avise 1993). A higher mutation rate would select for a higher rate of sexual recombination (i.e., a lineage would be selected to initiate sexuality after fewer asexual generations). In this case, sexual recombination of the host would benefit the symbiont by providing novel genetic clones to infect (fig. 3b). In this way, protomitochondrial lineages would be buffered from environmental changes that could force whole clones of host cells to become extinct (see Williams 1975; Maynard Smith 1978).
Features of the respiratory system may have predisposed protomitochondria toward such manipulation of their hosts. When modern mitochondria exhibit maximal phosphorylation, they generate small amounts of [O.sub.2]- (Chance et al. 1979). Cells with high metabolic demands (e.g., rapidly dividing cells) thus may engender minimal oxidative damage. However, under physiological conditions in resting "state 4" mitochondria, the respiratory carriers are characterized by a high degree of reduction and are quite reactive toward oxygen, and superoxide production is maximal (Chance et al. 1979). This pattern suggests that protomitochondria would have triggered minimal oxidative damage in a host cell that was able to maintain a rapid rate of cell division (and thus a high ATP demand) in a given environment. In a host cell that was unable to maintain a rapid rate of cell division, ATP demand would be lessened and protomitochondria would enter the resting state, with maximal production of [O.sub.2]-. Sexual recombination of hosts might soon ensue, and a genetically novel daughter cell with a higher rate of cell division might result.
At the same time, selection would favor genetic variants of the host cells that were better able to exploit the opportunities inherent in the effects of the mitochondria. Host cells that could adapt to higher levels of ATP with high energy demands and rapid rates of cell division would come to predominate. Endogenous oxidants would cause a higher mutation rate (Ames and Gold 1990a,b). Higher rates of senescence and sexual recombination would be expected to result (Bell 1988; Avise 1993). Sexual recombination would produce some mutationally "least-loaded" lines that would be favored by selection (see Bell 1988), and faster rates of evolution could result. In this way, the host cell could develop mechanisms to convert the initially detrimental effects of the mitochondria into evolutionary advantages, facilitating the persistence of the endosymbiosis.
Equally crucial, however, would be the evolution of host-symbiont communication. A host cell that resisted mitochondrial control by, for instance, greater SOD production could provoke an "arms race" with the mitochondria that would likely lead to the disintegration of the association. On the other hand, host cells that could detect mitochondrial signals to initiate sexual recombination prior to significant oxidative damage would have higher probabilities of survival. Host cells would be selected to respond to more subtle mitochondrial signals, and in this way the initial selective conflicts between these biological units would begin to be alleviated, in fact ritualized, and possibilities for a lasting association would be enhanced.
POSSIBLE SEQUELLAE: FROM DISCORD, HARMONY
The recognition of the nature of endogenous oxidants can thus recast the endosymbiont hypothesis in terms of a framework of intercellular conflict and benefits. Often, the resolution of units-of-evolution conflicts can lead to evolutionary innovations (Buss 1987), and such a result is suggested here. To achieve a stable association, intercellular conflict between the host and symbiont must give way to intracellular harmony (i.e., harmony within the evolving eukaryote, the host cell and the clonal population of mitochondria). To achieve a stable association, a chimeric nucleus (containing both host and symbiont genes) must counter selection for variant mitochondria. Such variants could subvert the aerobic energy production process for a "selfish" replicatory advantage to the detriment of the cell as a whole (including the other, clonal mitochondria). At the same time, intracellular communication is equally important for the success of the endosymbiont population; the mitochondria and the cell must "speak the same language." In other words, the mitochondrial population must be able to interpret, perhaps even signal, the major events in the life history of the cell. These dual evolutionary themes of nuclear control and within-host communication can be elucidated by the requirements for a stable interaction: on the one hand, cellular control to resist the evolution of variant mitochondria and, on the other, mitochondrial ability to perceive, respond to, and perhaps even regulate, cellular behavior.
Cellular Control of Mitochandria
Early in the evolution of the endosymbiosis, a variant mitochondrion could destabilize the host-symbiont association in two related ways. First, it could fail to export ATP to the cell. An individual mitochondrion that contributed no ATP to the cell would have an energetic advantage that could be translated into a replicatory advantage. Within the population of mitochondria in the cell, such a "selfish" variant would be favored and would leave more descendants. Ultimately, only such selfish mitochondria would remain, and the host cell, now possessing a life history requiring higher ATP levels, would likely die. This in turn might have a negative impact on the mitochondria; they would now have to colonize new hosts and such colonization might be risky or difficult. Second, a variant mitochondrion could consume metabolites and replicate at a faster rate than other mitochondria in the cell. The disadvantages to the cell are more subtle in this case. Unrestricted intracellular competition by mitochondria for metabolites could quickly lead to shortages in key substrates to the detriment of the cell and ultimately to the detriment of the mitochondrial population as well.
In this light, it is informative to view the uniquely derived aspects of the cell-mitochondria interaction in modern eukaryotes. Biochemical control of mitochondrial respiration rate is complex, but it is clear that a number of cytoplasmic factors regulate this rate (Erecinska and Wilson 1982; Brand and Murphy 1987). Mitochondria contain a small genome; a small number of mtDNA-encoded mRNAs specify essential polypeptide components of the mitochondrial electron-trans port system. The proteins produced by the mitochondrion function only in complexes with proteins produced by the nucleus and never in isolation (e.g., Gray 1989; Wallace 1992). Transcription or replication factors are not part of the mitochondrial genome; nuclear genes thus control these processes (Clayton 1991, 1992). Mitochondrial adenine nucleotide carriers catalyze the one-for-one exchange of adenine nucleotides, preferentially taking up ADP and ejecting ATE these proteins are also coded by nuclear genes.
Interpretation of these data on modern mitochondria begin with the recognition that the higher level unit of evolution (the host cell plus the clonal mitochondrial population) cannot control variant mitochondria as long as the individual mitochondrion can control its own replication and energy allocation (i.e., in the cytoplasmic environment, individual selection on the symbiont will always be stronger than group selection). Because energy production is fundamental to all cellular processes, including replication, the eukaryotic cell will always be vulnerable to the mitochondrion that is energetically selfish. Selection on the higher-level unit thus necessitated the transfer to the nucleus of the key mitochondrial genes mentioned above; this chimeric nucleus could then control the replication, transcription, energy metabolism, and energy allocation of the entire mitochondrial population.
In some cases, it is possible that more extreme measures than these were invoked. Animal mitochondria in particular have certain features (e.g., one or few copies of tRNA and rRNA genes, lack of spacer DNA between coding regions) that leave them very vulnerable to loss of function if a single mutation occurs. Gabriel et al. (1993) suggested that the genomes of these mitochondria are constructed so that a single mutation will lead to "mutational meltdown" and extinction of that lineage. By eliminating a wide range of mitochondrial variants, these uniquely derived features of the mitochondrial genome may represent a "fail-safe" measure to protect the cell and the clonal mitochondrial population from the mutational origin of a selfish mitochondrial variant.
Mitochondrial Control of the Cell
Early in the evolution of the endosymbiosis, a host cell could eliminate mitochondria from some of its descendants by evolving any of several variant life histories, for example, a rapid rate of cell division relative to that of the mitochondria. To counter such evolutionary threats, mitochondria must have the means to assess, and perhaps even signal, major features in the life history of the cell as a whole, such as cell division. To achieve a stable association, mechanisms of intracellular communication must thus be standardized between host and mitochondria, and signaling factors that the mitochondrion perceives must also be perceived by the cell and must be relevant to the cell's life history.
ATP and Phosphorylation Cascades.--The presence of aerobic mitochondria in an originally anaerobic host likely increased the availability of ATP by an order of magnitude, and ATP production has remained the principal mitochondrial function in eukaryotes. Cellular functions clearly require ATP, and the cytosolic ATP/ADP ratio may regulate mitochondrial respiration rate under some circumstances (Brand and Murphy 1987). Soon after the initial host-mitochondria association was established, a crude sort of cell-mitochondrial communication may thus have been possible. Major events in the life history of cells (e.g., cell division) require quantities of ATP. A certain minimum number of well-supplied mitochondria would be necessary before cytosolic energy reserves would be sufficient to initiate the process of cell division. In this way, the clonal mitochondrial population could achieve sufficient numbers to ensure representation in, for instance, both daughter cells. In this context, a new perspective on post-translational regulation of proteins by phosphorylation is suggested. Such phosphorylation occurs when a kinase transfers the terminal phosphate group from ATP to certain residues of a specific target protein, thus increasing that protein's activity. In some circumstances, phosphorylation could signal appropriate levels of energy availability. Phosphorylation cascades are characteristic of the regulation of eukaryotie cell-division cycles (Murray and Kirschner 1989; O'Farrell 1992), and this seems to be a uniquely derived feature (see Newton and Ohta 1990).
Calcium Fluctuations.--In eukaryotic cells, the concentration of cytosolic [Ca.sup.2+] controls many cellular activities ranging from the particular (e.g., excitability, contraction, exocytosis, growth) to the general (metabolism and gene expression; see Carafoli 1987; Berridge and Irvine 1989; Tsien and Tsien 1990; Allbritton et al. 1992; Bandtlow et al. 1993; Galione 1993). Cytosolic and mitochondrial [Ca.sup.2+] are also important regulators of mitochondrial respiration rate (Hansford 1985; Brand and Murphy 1987). Calcium ions thus might constitute a common currency for communication between the cell and the mitochondria. In fact, there has been considerable interest in the possibility of mitochondrial regulation of intracellular calcium and calcium signaling (Nicholls and Akerman 1982). In extant eukaryotes, both the affinity of mitochondria for [Ca.sup.2+] and the rate at which mitochondria import this cation are too low for mitochondria to regulate intracellular calcium concentrations effectively (Chance 1965; Hansford 1985; Carafoli 1987). Nevertheless, isolated mitochondria can take up large quantities of [Ca.sup.2+] (e.g., Lehninger 1964), and this may occur in vivo when calcium ions reach abnormally high concentrations (Alberta et al. 1983; Carafoli 1987).
The situation in primitive eukaryotes was perhaps more complex. The very low concentration of [Ca.sup.2+] in modern eukaryotes may be a derived feature and may have evolved early in the history of the eukaryotes (Ballantyne and Chamberlin 1988). Since mitotic microtubule systems require low and regulated [Ca.sup.2+] concentrations for tubulin polymerization (Borisy et al. 1975; Alberts et al. 1983), attention has focused on how calcium regulation may have affected the evolution of mitosis and meiosis (Margulis et al. 1976; Ballantyne and Chamberlin 1988). For instance, Margulis et al. (1976) suggested that calcium carbonate depositional systems, which produce the shells and skeletons of many marine organisms, may have first evolved to remove intracellular calcium and thus stabilize the mitotic apparatus. Depending on the timing of the endosymbiosis, mitochondria may have also played a role in maintaining low intracellular calcium levels and thus in regulating cell division (Ballantyne and Chamberlin 1988). Early in the evolution of the endosymbiosis, mitochondria may have accumulated stores of calcium. Perturbations of the redox state of mitochondria (for instance, if cells became overcrowded with mitochondria and substrate limited) may have caused the release of these calcium stores with consequent effects on the host cells (see Allen and Balin 1989).
Possible Sequellae Continued: A Closer Look at Multicellular Eukaryotes
The above discussion employed uniquely derived features of modern eukaryotic cells to infer the nature of the hostmitochondria interaction within a single cell. These interactions, near the inception of eukaryotic lineages, may have channeled the further evolution of these lineages. Multicellular eukaryotic organisms, particularly animals with motile cells, are faced with many of the same units-of-evolution challenges that confronted primitive eukaryotic cells (Buss 1987; Maynard Smith 1988). Cells within a multicellular animal compete for access to the germ line, and individual mitochondria would be expected to also compete for such access. Further, the competition between host cell and mitochondrion may have provided a framework for the evolution of competition between cells in a multicellular animal. For instance, competition between cells would require effective mechanisms of intercellular communication. Given such mechanisms, cells could assess, and perhaps even manipulate, other cells' activities. Communication mechanisms that allowed mitochondria to predict and manipulate the life histories of their host cells would seem obvious candidates to be co-opted into such between-cell signaling. For these reasons, mechanisms that became established to permit hostsymbiont communication may have been co-opted into allowing cell-cell communication, particularly in animals.
I will briefly discuss two aspects of the relationship between mitochondria and multicellular eukaryotes: (1) competition among mitochondria for access to the germ line, and (2) putative similarities between cell-cell communication in multicellular organisms and the mechanisms of host-symbiont communication described above.
Access to the Germ Line
Inhabiting a multicellular organism with specialized somatic and germ cells presents mitochondria with new challenges. In a clonal organism, only mitochondria inhabiting the totipotent stem cell lineage have a chance of "colonizing" a new organism; in aclonal organisms only those mitochondria inhabiting the germ line have this chance. If a genetically variant mitochondrion arose such that it could trigger the formation of stem cells or germ cells, or both, this variant would have a tremendous selective advantage over other mitochondria. In a few (organismal) generations, such a genetic variant could predominate and eliminate other mitochondria. This process exactly parallels competition for the germ line among genetically variant cells (Buss 1987).
In this context, the nature of the so-called "germinal plasm" is of interest. These dense, cytoplasmic materials are often present in germ cells of various animals, and a role as a specific determinant of the germ line has been suggested. Typically, mitochondria are closely associated with this germ plasm. In fact, in a recent study of Drosophila, it was found that mitochondrial large ribosomal RNA is a component of the germinal plasm (Kobayashi et al. 1993). Although it is not yet clear how mitochondrial transcripts are exported into the cytoplasm, some data suggest that this mtRNA may participate in the determination of the germ line (Kobayashi et al. 1993). Cytoplasmic substances very similar to the germ plasm are also found in the totipotent stem cells of hydroids and flatworms, and mitochondria are usually in close association with these substances (Hay and Coward 1975; Noda and Kanai 1977). It is not yet known whether mtRNA is present in the cytoplasm of these stem cells.
Effects of Extracellular ATP.--In complex eukaryotes, extracellular ATP has a number of effects including alterations of ion fluxes, phospholipid metabolism, growth, differentiation, and secretion (see Dubyak and Fedan 1990). It is suggested above that ATP levels provided a principal mechanism for communication between the host cell and mitochondria early in the evolution of the endosymbiosis; the use of ATP in intercellular communication may have been derived from this original intracellular signaling function.
Second Messengers.--"Second-messenger" systems are common in eukaryotes, allowing cell-cell communications in a variety of circumstances (e.g., Berridge and Irvine 1989 Berridge 1993). Typically, stimulation of cell-surface receptors and hydrolysis of a membrane-bound inositol lipid produces at least two second messengers, diacylglycerol and inositol 1,4,5-trisphosphate. The former activates protein kinase C, which phosphorylates a variety of enzymes, while the latter triggers the release of nonmitochondrial calcium stores, thus initiating a variety of events, including the further activation of the calcium-sensitive protein kinase C. These ubiquitous signal transduction mechanisms are now known to regulate a large array of cellular processes including metabolism, secretion, contraction, neural activity, and cell proliferation (Berridge and Irvine 1989). These pathways are reminiscent of possible host-mitochondria communications early in the evolution of the endosymbiosis. The postulated existence of mitochondrial signaling to host cells via calcium fluxes may have provided the raw material that led to the evolution of the observed system of second messengers for signaling between the cells of extant eukaryotic animals.
Free-Radical Theory of Development.--Metabolic processes in general and concentrations of mitochondria in particular were once considered key elements in the regulation of metazoan development (e.g., Child 1941; Gustafson and Lenique 1952). The difficulty of distinguishing cause from effect in these theories and the rise of the paradigm of variable gene activity (e.g., Davidson 1986) has led to the abandonment of general theories of metabolic control of development. Recently, there has been a resurgence of interest in this topic, reconceptualized as the "free radical" theory of development (Allen and Balin 1989; Fanburg et al. 1992; Shweiki et al. 1992). A variety of developmental phenomena suggest that oxidative gradients may produce gradients of gene activity. The perspective developed here views such theories in an evolutionary context; the effects of oxidative gradients on development in metazoan animals may represent no more than an extension of the free radical control of the primitive host by the protomitochondria. Again, the system of signaling between mitochondria and host cells may have been co-opted and elaborated into a system of signaling between the cells in multicellular animals.
Free-Radical Theory of Aging.--The free-radical theory of aging (elaborated by Harman 1962, 1981) suggests that senescence in metazoan animals is largely a consequence of the sum of deleterious free radical reactions going on continuously throughout cells and tissues. This theory has been criticized as " nonevolutionary " (Rose 1991). In other words, if organismal senescence were selected against, there would seem to be no limits to the evolution of better defenses against free radicals, for example, heightened SOD and associated enzyme activity. From the viewpoint of evolutionary theories of aging, for example, mutation accumulation and negative pleiotropy (Rose 1991), free radicals are only one of many potential proximate causes of senescence, while the ultimate cause of senescence is the declining force of natural selection with the increasing age of the organism (Medawar 1952; Williams 1957; Hamilton 1966; Charlesworth 1980; Rose 1991). Nevertheless, free radicals are implicated in the aging process to an extent perhaps not predicted by current evolutionary theories (Chance et al. 1979; Richter 1988; Floyd 1991; Halliwell and Aruoma 1991; Stadtman 1992; Wallace 1992; Orr and Sohal 1994). The role of free radicals in metazoan aging follows from the perspective developed above; early in the history of the endosymbiosis, mitochondria used free radicals to manipulate the life history of their host cells, forming DNA adducts and mutations in the host to trigger sexual recombination and opportunities for mitochondria to colonize genetically novel hosts with minimal risk. This ancient system of signaling senescence may have remained intact and in many cases may still provide the proximate mechanism of aging in populations of multicellular organisms with the appropriate age-specific life histories.
Darwin developed and tested his theory of evolution by natural selection using the natural history data available in the nineteenth century, that is, data from multicellular organisms. Since Darwin's time, and particularly in the last 50 yr, whole new classes of natural history data have been observed, i.e., data from molecules, organelles, and cells. Using a units-of-evolution approach, these new classes of data provide an added proving ground for Darwin's theory of evolution (Lewontin 1970). The challenge of modern biology is thus to elucidate not only biological mechanisms, but the possible selective agents that produced these mechanisms (Buss 1987).
Employing a units-of-evolution perspective on the endosymbiont theory, the above discussion suggests that (1) the host-mitochondria interaction at its inception must be conceptualized in terms of synergistic and antagonistic effects between two units of evolution, (2) ecological and physiological differences between aerobic protomitochondria and their anaerobic primitive hosts are central to understanding these costs and benefits, and (3) the host-mitochondria interactions may have channeled the evolution of many uniquely derived features of eukaryotic cells so that an understanding of this interaction can provide insight into some key aspects of the biology of eukaryotes. Although a considerable amount of correlational and associational data support these statements, key questions remain. For instance, is there phenotypic selection on mitochondria in modern eukaryotic cells? If the normal cell can eliminate mitochondria on the basis of, say, their rate of ATP export, many questions about mitochondrial population dynamics would be answered (see Avise 1993). Many questions about the phylogeny of eukaryotes also remain. It is an open question as to whether an extant group of eukaryotes constitutes a sister-group to the mitochondria-bearing eukaryotic clade. Such a group would be invaluable for more thoroughly evaluating the character states discussed here.
Additional intriguing avenues of inquiry are suggested. Basal groups of extant eukaryotes may provide insight into alternative mechanisms for resolving host-mitochondria conflict. For instance, the editing of mitochondrial RNA by nuclear enzymes in the kinetoplastids (e.g., Koslowsky et al. 1992) suggests an additional mechanism by which the nucleus can diminish the expression of mitochondrial variation. Questions about the organization of the eukaryotic genome are also suggested. It would be of interest to identify the nuclear genes that now regulate the transcription and replication of the mitochondrial genome. If, for instance, some of these genes employ the DNA-binding helix-turn-helix motif, this would imply that the population of such genes in the nucleus of eukaryotes is of chimeric origin. Genes that regulate the transcription and replication of the mitochondrial genome would be expected to show affinities to the regulatory genes of the [alpha]-purple eubacteria.
Although there can be no definitive direct tests as to the nature of a biological interaction that began a billion or more years ago, several relevant experiments on extant eukaryotes are suggested. In particular, an exploration of the costs of aerobic respiration should be undertaken, for instance, by growing yeast under aerobic and anaerobic conditions and measuring mutation and recombination rates in both treatments. In addition, mutation rates of cells with mitochondria that are phosphorylating maximally could be compared to rates of cells with mitochondria in the resting state. The growing interest in endogenous oxidants in normal (Allen and Balin 1989; Fanburg et al. 1992) and abnormal development (Ames and Gold 1990a,b), and the increasing focus on how oxidative gradients function mechanistically, particularly in relation to gene activity, will provide additional levels of relevant detail. Plausibly, the role of oxidative gradients in directing development has declined since the early history of the endosymbiosis; in other words, as the host-mitochondria conflict ameliorated, genetic (i.e., nuclear) control of life histories came to predominate. Oxidative mechanisms in development may therefore be overlaid by reinforcing genetic mechanisms (see Newman and Comper 1990; Newman 1992). Ultimately, such predictions can be used to guide reductionist research programs in biology.
[Figures 1 to 3 ILLUSTRATION OMITTED]
Leo Buss' The Evolution of Individuality led me to think about mitochondria in this way. Discussions with D. Bridge L. Buss, and B. Chance helped to formulate these ideas. seminar presented to the Johnson Foundation at the University of Pennsylvania further stimulated this review. Comments and criticism of the manuscript were provided by D Berrigan, D. Bridge, L. Buss, B. Chance, U. Smith, G. Vermeij, G. Wagner, and several anonymous reviewers. The National Science Foundation (BSR-88-05691, IBN-94-07049 and the Graduate School of Northern Illinois University provided support.
Alberts, B., D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson. 1983. The molecular biology of the cell. Garland, New York.
Allbritton, N. L., T. Meyer, and L. Stryer. 1992. Range of messenger action of calcium ion and inositol 1,4,5-trisphosphate. Science 258:1812-1815.
Allen, R. G., and A. K. Balin. 1989. Oxidative influence on development and differentiation: an overview of a free radical theory of development. Free Radicals in Biology and Medicine 6:631-661
Ames, B., and L. S. Gold. 1990a. Chemical carcinogenesis: too many rodent carcinogens. Proceedings of the National Academy of Sciences, USA. 87:7772-7776.
--. 1990b. Too many rodent carcinogens: mitogenesis increases mutagenesis. Science 249: 970-971.
Archibald, J. D. 1994. Metataxon concepts and assessing possible ancestry using phylogenetic systematics. Systematic Biology 43:27-40.
Avise, J. C. 1993. The evolutionary biology of aging, sexual reproduction, and DNA repair. Evolution 47:1293-1301.
Ballantyne, J. S., and M. E. Chamberlin. 1988. Adaptation and evolution of mitochondria: osmotic and ionic considerations. Canadian Journal of Zoology 66:1028-1035.
Baltscheffsky, H., and M. Baltscheffsky. 1981. Mitochondrial ancestor models: Paracocaus, Rhizabium, and Rhodospirillum. Pp. 519-540 in C. P. Lee, G. Schatz, and G. Dallner, eds. Mitochondria and microsomes. Addison-Wesley, New York.
Bandtlow, C. E., M. F. Schmidt, T. D. Hassinger, M. E. Schwab and S. B. Kater. 1993. Role of intracellular calcium in NI-35 evoked collapse of neuronal growth cones. Science 259:80-83.
Behm, C. A. 1991. Fumarate reductase and the evolution of electron transport systems. Pp. 88-108 in C. Bryant, ed. Metazoan life without oxygen. Chapman and Hall, London.
Bell, G. 1988. Sex and death in Protozoa. Cambridge University Press, Cambridge.
Berridge, M. J. 1993. Inositol trisphosphate and calcium signaling. Nature 361:315-325.
Berridge, M. J., and R. F. Irvine. 1989. Inositol phosphates and cell signaling. Nature 341:197-205.
Borisy, G. G., J. M. Marcum, J. B. Olmsted, D. B. Murphy, and K. A. Johnson. 1975. Purification of tubulin and associated high molecular weight proteins from porcine brain and characterization of microtubule assembly in vitro. Annals of the New York Academy of Sciences 253:107-132.
Brand, M. D., and M. P. Murphy. 1987. Control of electron flux through the respiratory chain in mitochondria and cells. Biological Reviews 62:141-193.
Brusca, R. C., and G. J. Brusca. 1990. Invertebrates. Sinauer, Sunderland, Mass.
Bryant, C., ed. 1991. Metazoan life without oxygen. Chapman and Hall, New York.
Buss, L. W. 1987. The evolution of individuality. Princeton Uni versity Press, Princeton, N.J.
Carafoli, E. 1987. Intracellular calcium homeostasis. Annual Re views of Biochemistry 56:395-433.
Cavalier-Smith T. 1987a. The simultaneous symbiotic origin of mitochondria, chloroplasts, and microbes. Annals of the New York Academy of Sciences 503:55-71.
--. 1987b. Eukaryotes with no mitochondria. Nature 326:332-333.
Chance, B. 1965. The energy-linked reaction of calcium with mitochondria. Journal of Biological Chemistry 240:2729-2748.
Chance, B., H. Sies, and A. Boveris. 1979. Hydroperoxide metabolism in mammalian organs. Physiological Reviews 59:527-605.
Charlesworth, B. 1980. Evolution in age-structured populations. Cambridge University Press, Cambridge.
Chapman, D. J., and H. Gest. 1983. Terms used to describe biological energy conversion, electron transport processes, interactions of cellular systems with molecular oxygen, and carbon nutrition. Pp. 459-463 in J. W. Schopf, ed. Earth's earliest biosphere. Princeton University Press, Princeton, N.J.
Child, C. M. 1941. Patterns and problems in development. University of Chicago Press, Chicago.
Clayton, D. A. 1991. Replication and transcription of vertebrate mitochondrial DNA. Annual Review of Cell Biology 7:453-78.
--. 1992. Transcription and replication of animal mitochondrial DNA's. International Review of Cytology 141:217-232.
Cosmides, L. M., and J. Tooby. 1981. Cytoplasmic inheritance and intragenomic conflict. Journal of Theoretical Biology, 89:83-129.
Davidson, E. H. 1986. Gene activity in early development, 3d ed. Academic Press, Orlando, Fla.
Dubyak, G. R., and J. S. Fedan, eds. 1990. Biological actions of extracellular ATP. Annals of the New York Academy of Sciences 603:1-542.
Dutton, P. L., and D. F. Wilson. 1974. Redox potentiometry in mitochondrial and photosynthetic bioenergetics. Biochimica et Biophysica Acta 346:165-212.
Eberhard, W. G. 1980. Evolutionary consequences of intracellular organelle competition. Quarterly Review of Biology 55:231-249.
Erecinska, M., and D. F. Wilson. 1982. Regulation of cellular energy metabolism. Journal of Membrane Biology 70:1-14.
Fanburg, B. L., D. J. Massaro, P. A. Cerutti, D. B. Gail, and M. A. Berberich. 1992. Regulation of gene expression by [0.sub.2] tension. America Journal Physiology 262:L235-L241.
Fenchel, T., and C. Bernard. 1993. A purple protist. Nature 362:300.
Fenchel, T., and B. J. Finlay. 1994. The evolution of life without oxygen. American Scientist 82:22-29.
Floyd, R. A. 1991. Oxidative damage to behavior during aging. Science 254:1597.
Fox, G. E., E. Stackebrandt, R. B. Hespell, et al. 1980. The phylogeny of the prokaryotes. Science 209:457-463.
Gabriel, W., M. Lynch, and R. Burger. 1993. Muller's ratchet and mutational meltdowns. Evolution 47:1744-1757.
Galione, A. 1993. Cyclic ADP-ribose: a new way to control calcium. Science 259:325-326.
Gest, H. 1980. The evolution of biological energy-transducing systems. FEMS Microbiology Letters 7:73-77.
Gogarten, J. P., H. Kibak, P. Dittrich, L. Taiz, E. J. Bowman, B. J. Bowman, M. J. Manolson, R. J. Poole, T. Date, T. Oshima, J. Konishi, K. Denda, and M. Yoshida. 1989. Evolution of the vacuolar [H.sup.+]-ATPase: implications for the origin of eukaryotes. Proceedings of the National Academy of Sciences, USA 86:6661-6665.
Gray, M. W. 1989. Origin and evolution of mitochondrial DNA. Annual Reviews of Cell Biology 5:25-50.
--. 1992. The endosymbiont hypothesis revisited. International Review of Cytology 141:233-335.
Gray, M. W., and W. F. Doolittle. 1982. Has the endosymbiont hypothesis been proven? Microbiology Review 46:1-42.
Gustafson, T., and P. Lenique. 1952. Studies on mitochondria in the developing sea urchin egg. Experimental Cell Research 3:251-274.
Halliwell, B., and O. I. Arnoma. 1991. DNA damage by oxygen-derived species. FEBS Letters 281:9-19.
Hamilton, W. D. 1966. The moulding of senescence by natural selection. Journal of Theoretical Biology 12:12-45.
Hansford, R. G. 1985. Relation between mitochondrial calcium transport and control of energy metabolism. Review of Physiological and Biochemical Pharmacology 102:1-72.
Harman, D. 1962. Role of free radical in mutation, cancer, aging, and the maintenance of life. Radiation Research 16:753-763.
--. 1981. The aging process. Proceedings of the National Academy of Sciences, USA 78:7124-7128.
Hassan, H. M., and J. R. Schiavone. 1991. The role of oxygen free radicals in biology and evolution. Pp. 38-64 in C. Bryant, ed. Metazoan life without oxygen. Chapman and Hall, London.
Hay, E. D., and S. J. Coward. 1975. Fine structure studies on the planarian, Dugesia. Journal of Ultrastructure Research 50:1-21.
Hull, D. L. 1988. Interactors versus vehicles. Pp. 19-50 in H. C. Plotkin, ed. The role of behavior in evolution. MIT Press, Cambridge, Mass.
Hurst, L. D. 1992. Intragenomic conflict as an evolutionary force. Proceedings of the Royal Society of London, B 248:135-140.
Hurst, L. D., and W. D. Hamilton. 1992. Cytoplasmic fusion and the nature of sexes. Proceedings of the Royal Society of London B 247:189-194.
Iwabe, N., K. Kuma, M. Hasegawa, S. Osawa, and T. Miyata. 1989. Evolutionary relationship of archaebacteria, eubacteria, and eukaryotes inferred from phylogenetic trees of duplicated genes. Proceedings of the National Academy of Sciences, USA 86:9355-9359.
Jenkins, R. J. F. 1991. The early environment. Pp. 38-64 in C. Bryant, ed. Metazoan life without oxygen. Chapman and Hall, London.
Kessel, E. L. 1955. Mating activities of balloon flies. Systematic Zoology 4:97-104.
Knoll, A. H. 1992. The early evolution of eukaryotes: a geological perspective. Science 256:622-627.
Kobayashi, S., R. Amikura, M. Okada. 1993. Presence of mitochondrial large ribosomal RNA outside mitochondria in germ plasm of Drosophila melanogaster. Science 260:1521-1524.
Koslowsky, D. J., H. U. Goringer, T. H. Morales, and K. Stuart. 1992. In vitro guide RNA/mRNA chimaera formation in Trypanosoma brucei RNA editing. Nature 356:807-809.
Lehninger, A. L. 1964. The Mitochondrion. W. A. Benjamin, New York.
Lewontin, R. C. 1970. The units of selection. Annual Review of Ecology and Systematics 1:1-18.
Loomis, W. F. 1988. Four billion years. Sinauer, Sunderland, Mass.
Lloyd, E. A. 1988. The structure and confirmation of evolutionary theory. Greenwood Press, Westport, Conn.
--. 1992. Unit of selection. Pp. 334-340 in E. F. Keller and E. A. Lloyd, eds. Keywords in evolutionary biology. Harvard University Press, Cambridge, Mass.
Margulis, L. 1970. Origin of eukaryotic cells. Yale University Press, New Haven, Conn.
--. 1981. Symbiosis in cell evolution. W. H. Freeman, San Francisco.
Margulis, L., and D. Sagan. 1986. Origins of sex. Yale University Press, New Haven, Conn.
Margulis, L., J. C. G. Walker, and M. Rambler. 1976. Reassessment of roles of oxygen and ultraviolet light in Precambrian evolution. Nature 264:620-624.
Maynard Smith, J. 1978. The evolution of sex. Cambridge University Press, London.
--. 1988. Evolutionary progress and levels of selection. Pp. 219-230 in M. H. Nitecki, ed. Evolutionary progress. University of Chicago Press, Chicago.
McCarroll, R., G. J. Olsen, Y. D. Stahl, C. R. Woese, and M. L. Sogin. 1983. Nucleotide sequence of the Dictyostelium discoideum small-subunit ribosomal ribonucleic acid inferred from the gene sequence: evolutionary implications. Biochemistry 22:5858-5868.
--. 1991. A Darwinian view of symbiosis. Pp. 26-39 in L. Margulis and R. Fester, eds. Symbiosis as a source of evolutionary innovation. MIT Press, Cambridge, Mass.
Medawar, P. B. 1952. An unsolved problem of biology. H. K. Lewis, London.
Murray, A. W., and M. W. Kirschner. 1989. Dominoes and clocks: the union of two views of the cell cycle. Science 246:614-621.
Muller, M. 1988. Energy metabolism of protozoa without mitochondria. Annual Review of Microbiology 42:465-488.
Newman, S. A. 1992. Generic physical mechanisms of morphogenesis and pattern formation as determinants in the evolution of multicellular organization. Pp. 241-268 in J. E. Mittenthal and A. B. Baskin, eds. The principles of organization in organisms. Addison-Wesley, Reading, Mass.
Newman, S. A., and W. D. Comper. 1990. "Generic" physical mechanisms of morphogenesis and pattern formation. Development 110:1-18.
Newton, A., and N. Ohta. 1990. Regulation of the cell division cycle and differentiation in bacteria. Annual Review of Microbiology 44:689-719.
Nicholls, D. G., and K. Akerman. 1982. Mitochondrial calcium transport. Biochimica et Biophysica Acta 683:57-88.
Noda, K., and C. Kanai. 1977. An ultrastructural observation on Pelmatobydra robusta at sexual and asexual stages, with special reference to "germinal plasm." Journal of Ultrastructure Research 61:284-294.
O'Farrell, P. H. 1992. Cell cycle control: many ways to skin a cat. Trends in Cell Biology 2:159-163.
Orr, W. C., and R. S. Sohal. 1994. Extension of life-span by overexpression of superoxide dismutase and catalase in Drosophila melanogaster. Science 263:1128-1130.
Richter, C. 1988. Do mitochondrial DNA fragments promote cancer and aging? FEB 241:1-5.
Rivera, M. C., and J. A. Lake. 1992. Evidence that eukaryotes and eocyte prokaryotes are immediate relatives. Science 257:74-76.
Rose, M. R. 1991. Evolutionary biology of aging. Oxford University Press, Oxford.
Rowlands, T., P. Baumann, and S. P. Jackson. 1994. The TATA-binding protein: a general transcription factor in eukaryotes and archacbacteria. Science 264:1326-1329.
Runnegar, B. 1991. Oxygen and the early evolution of the metazoa. Pp. 65-87 in C. Bryant, ed. Metazoan life without oxygen. Chapman and Hall, London.
Schwartz, R. M., and M. O. Dayhoff. 1978. Origins of prokaryotes, eukaryotes, mitochondria, and chloroplasts. Science 199:395-403.
Schopf, J. W., J. M. Hayes, and M. R. Walter. 1983. Evolution of earth's earliest ecosystems. Pp. 361-384 in J. W. Schopf, ed Earth's earliest biosphere. Princeton University Press, Princeton N.J.
Shweiki, D., A. Itin, D. Soffer, and E. Keshet. 1992. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359:843-848.
Sober, E. 1984. The nature of selection. MIT Press, Cambridge, Mass.
Sogin, M. L., J. H. Gunderson, H. J. Elwood, R. A. Alonso, D. A. Peattie. 1989. Phylogenetic meaning of the kingdom concept: an unusual ribosomal RNA from Giardia lamblia. Science 243:75-77.
Spenser, D. F., M. N. Schnare, and M. W. Gray. 1984. Pronounced structural similarities between small subunit ribosomal RNA genes of wheat mitochondria and Escherichia coli. Proceedings of the National Academy of Sciences, USA 81:493-497.
Stadtman, E. R. 1992. Protein oxidation and aging. Science 257:1220-1224.
Tsien, R. W., and R. Y. Tsien. 1990. Calcium channels, stores, and oscillations. Annual Reviews of Cell Biology 6:715-760.
Vossbrinck, C. R., J. V. Maddox, S. Friedman, B. A. Debrunner-Vossbrinck and C. R. Woese. 1987. Ribosomal RNA sequence suggests microsporidia are extremely ancient eukaryotes. Nature 326:411-414.
Wallace, D. C. 1992. Mitochondrial genetics: a paradigm for aging and degenerative diseases? Science 256:628-632.
Williams, G. C. 1957. Pleiotropy, natural selection, and the evolution of senescence. Evolution 11:398-411.
--. 1975. Sex and evolution. Princeton University Press, Princeton, N.J.
--. 1992. Natural selection. Oxford University Press, Oxford. Wilson, E. B. 1925. The cell in development and heredity. Macmillan, New York.
Woese, C. R., O. Kandler, and M. L. Wheelis. 1990. Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proceedings of the National Academy of Sciences, USA 87:4576-4579.
Yang, D., Y. Oyaizu, H. Oyaizu, G. J. Olsen, and C. R. Woese. 1985. Mitochondrial origins. Proceedings of the National Academy of Sciences, USA 82: 4443-4447.
Zillig, W., R. Schnabel, and K. O. Stetter. 1985. Archaebacteria and the origin of the eukaryotic cytoplasm. Current Topics in Microbiology and Immunology 114:1-18.
|Printer friendly Cite/link Email Feedback|
|Author:||Blackstone, Neil W.|
|Date:||Oct 1, 1995|
|Previous Article:||Selection on fecundity and variation in the degree of sexual size dimorphism among spider species (Class Araneae).|
|Next Article:||Genetic constraints on macroevolution: the evolution of host affiliation in the leaf beetle genus Ophraella.|