Abstract

Living things invariably consist of some kind of compartmentalized redox chemistry. Signaling pathways mediated by oxidation and reduction thus derive from the nature of life itself. The role of such redox or metabolic signaling broadened with major transitions in the history of life. Prokaryotes often use redox signals to deploy one or more variant electron carriers and associated enzymes to better utilize environmental energy sources. Eukaryotes transcend the strong surface-to-volume constraints inherent in prokaryotic cells by moving chemiosmotic membranes internally. As a consequence, eukaryotic redox signaling is frequently between these organelle membranes and the nucleus, thus potentially involving levels-of-selection synergies and antagonisms. Gradients of oxygen and substrate in simple multicellular organisms similarly associated metabolic signaling with levels of selection, now at the level of the cell and the organism. By allowing sequestration of large amounts of food, the evolution of the animal mouth was a pivotal event in metabolic signaling, leading to “multicellular” redox regulation. Because concentrated food resources may be patchy in time and space, long-lived sedentary animals with mouths employ such metabolic signaling and phenotypic plasticity in ways that adapt them to the changing availability of food. Alternatively, if the mouth is coupled to a battery of sensory equipment, the organism can actively seek out and sequester patches of food. In these early bilaterians, competition for food resources may have favored rapid development with little subsequent plasticity and metabolic signaling. With rapid dispersal and colonization, such “assembly-line” animals could effectively compete for patchy resources. Limiting metabolic signaling, however, resulted in a cascade of seemingly unrelated changes. These changes derive from the effectiveness of metabolic signaling in policing variation at the cellular level. If the signals an organism uses to control cellular replication are the same as the signals a cell uses to control its own metabolism, then cells that ignore these signals and carry out selfish replication will pay a fitness cost in terms of inefficient metabolism. Bilaterians with limited metabolic signaling thus require other mechanisms to police cell-level variation. Bilaterian features such as restricted somatic cell potency, a sequestered germ line, and determinate growth should be viewed in this context. Bilaterian senescence evolved as a by-product of restricted potency of somatic cells, itself a mechanism of cell policing required by limited metabolic signaling.

Introduction

At its most basic level, life consists of compartmentalized redox chemistry (Martin and Russell 2003). Typically in modern cells, this redox chemistry is based on some form of chemiosmosis, that is, linking oxidation of substrate and electron flow to a membrane-based electrochemical gradient that can be harnessed to do work (e.g., synthesizing ATP) (Fig. 1). While it may be that the first living things employed chemiosmosis (Martin and Russell 2003), it is perhaps more plausible that chemiosmosis was a later invention. In all likelihood, chemiosmosis nevertheless preceded the latest or last universal common ancestor of life (see Valentine 2006 for a discussion of terminology). This nearly universal mechanism of energy conversion has had great effects on the history of life and on the biosphere, e.g., the oxygen-rich atmosphere of modern Earth is a by-product of chemiosmosis (Lane 2002).

Fig. 1

Schemata of the mitochondrial electron-transport chain, showing complexes I–V, coenzyme Q, and cytochrome c. Small arrows trace the flow of electrons from NADH and FADH2 to oxygen. Large arrows show the extrusion of protons (H+) by complexes I, III, and IV and the return of protons to the matrix via complex V, triggering the assembly of ATP (dashed arrow). “A” and “B” indicate the two major sites of ROS formation. Blocking the electron-transport chain “downstream” of these sites causes the electron carriers to become highly reduced and as a consequence to donate more electrons to oxygen, thus increasing ROS. On the other hand, high metabolic demand diminishes ATP and accelerates electron flow. Electron carriers become less reduced and as a consequence donate fewer electrons to oxygen, thus decreasing ROS (Blackstone and Kirkwood, 2003).

Fig. 1

Schemata of the mitochondrial electron-transport chain, showing complexes I–V, coenzyme Q, and cytochrome c. Small arrows trace the flow of electrons from NADH and FADH2 to oxygen. Large arrows show the extrusion of protons (H+) by complexes I, III, and IV and the return of protons to the matrix via complex V, triggering the assembly of ATP (dashed arrow). “A” and “B” indicate the two major sites of ROS formation. Blocking the electron-transport chain “downstream” of these sites causes the electron carriers to become highly reduced and as a consequence to donate more electrons to oxygen, thus increasing ROS. On the other hand, high metabolic demand diminishes ATP and accelerates electron flow. Electron carriers become less reduced and as a consequence donate fewer electrons to oxygen, thus decreasing ROS (Blackstone and Kirkwood, 2003).

By its very nature, chemiosmosis entails signaling pathways mediated by the chemistry of oxidation and reduction, i.e., “redox” signaling. Linking what are ultimately environmental sources and sinks of electrons to an electrochemical gradient and ATP synthesis produces predictable consequences in the event of environmental perturbations of those sources and sinks (Allen 1993). For instance, if the environment becomes stressful and thus unfavorable for cellular replication, yet plenty of substrate remains available, cellular metabolic demand will diminish, and the ATP/ADP ratio will approach one. Oxidation of substrate, however, will continue until the electrochemical gradient is maximal and the membrane-bound electron carriers are highly reduced (Fig. 1). In such circumstances, these electron carriers will typically donate electrons to molecules (e.g., diatomic oxygen) whose reduced products can serve as messengers (e.g., reactive oxygen species, ROS). Such messengers can then trigger the appropriate adaptive response at the cellular level. On the other hand, during starvation (shortage of substrate) the ATP/ADP ratio will approach zero, and the electrochemical gradient becomes minimal. The electron carriers now are relatively oxidized and formation of, for instance, ROS is also minimal. Cellular responses that are diametrically opposite to the first case can then ensue.

Redox or metabolic signaling thus provides a fast and effective link between changes in the environment on one hand and internal cues for adaptive cellular behavior on the other. It is therefore not surprising that redox signaling is an ancient and universal mechanism for environmental signaling pathways (i.e., pathways affected by factors that ultimately originate outside the cell). With each of the major transitions in the history of life, the role of metabolic signaling was nevertheless altered and often broadened. Concomitantly with some of these transitions (e.g., from prokaryotes to eukaryotes and from unicellular to multicellular organisms), redox signaling mechanisms became inexorably linked to levels of selection synergies and antagonisms (Blackstone 1995, 2000). In this context, several important milestones for the role of redox signaling in the history of life are outlined and discussed subsequently. These examples provide useful background for the central question addressed here—the role of metabolic signaling in the evolution of animals.

In animal history, the evolution of the mouth and gut were pivotal events. For a heterotroph, the capacity to sequester large amounts of food may have been a decisive advantage. While competition for food no doubt dates from the earliest living things, the animal mouth likely accelerated resource-related competition. Because concentrated food resources may be patchy in time and space, long-lived sedentary animals with mouths employ metabolic signaling (now at a multicellular level) and phenotypic plasticity to adapt to the changing availability of food. Early-evolving metazoans may thus typify organisms that are highly responsive to redox-related environmental signals in pattern formation (Blackstone and Bridge 2005). This sensitivity to metabolic signals may allow basal metazoans (and by inference, the stem-lineage metazoans) to avoid the trade-off that dominates bilaterian longevity. For instance, life span in mammals requires a balance between the risks of too little and too much self-renewal. Too little self-renewal results in degenerative senescence; too much renewal produces cancer (Tyner et al. 2002; Radtke and Clevers 2005; Beausejour and Campisi 2006). Many cnidarians and sponges have nearly unlimited capacity for self-renewal (Holstein et al. 2003) as well as potential immortality (Martinez 2002). Why do not long-lived cnidarians and sponges die of cancer? As described below, the answer in part may be metabolic signaling: if the signals a cell uses to control its own metabolism are the same as the ones used by the organism to control cells that are selfish in terms of replication, then cancer will entail a heavy fitness cost at the level of the cell.

Because of its flexibility, however, metabolic signaling can be a liability in organisms that develop quickly and precisely with relatively little environmental input. Such patterns of development may be favored in animals in which the mouth is coupled to a battery of sensory equipment. These organisms can actively seek out and sequester patches of food. In the stem lineage leading to bilaterians, selection acting on a life history built about directed movement may have favored rapid completion of pattern formation during embryogenesis with little subsequent plasticity—“one time, permanent” assembly followed by dispersal, colonization, and rapid exploitation of favorable habitats. Rapid and efficient mechanisms—e.g., the canonical Hox system (Kamm et al. 2006)—govern the pattern-forming processes of embryogenesis. These rapid and efficient mechanisms of assembly, however, may have had associated costs. Limiting the use of redox signaling in pattern formation also limited the extent that “honest” metabolic signaling was available to police cancers. As a consequence, adults of the stem bilaterian may have exhibited diminished somatic cell potency, determinate growth, and a sequestered germ line (Blackstone and Ellison 2000; Blackstone and Jasker 2003). Some of the crown-group bilaterians, nevertheless, evolved larger sizes but remained subject to the relentless trade-off between the perils of cancer on one hand and the risks of too little self-renewal on the other. This trade-off has formed the basis for bilaterian senescence. By this view, those bilaterians that do not senesce have returned to the basal condition via convergence, parallelism, or reversal. These themes are further discussed and conclusions are presented as several testable predictions.

A food's-eye view of life

Foremost, bioenergetic metabolism involves linking external electron sources and sinks through a series of living redox couples. Ultimately, via the process of oxidation and reduction, the organism obtains energy from the environment. This has been a common theme of life since its origin. Living things, consisting of compartmentalized redox chemistry, also employ redox signaling as a rapid and effective mechanism to adjust bioenergetic metabolism. The “payoff ” has been a larger dividend in energy. Since this dividend can be spent on a higher rate of replication, the Darwinian imperative is clear. As life became more complex, additional inputs into the web of redox signaling have been added. Given the possibility for multicellular growth, for instance, the organism could spread from an area depleted in electron sources or sinks to one that was rich in this regard. Linking mechanisms of growth to redox signaling would therefore be favored by selection, and such linkage should be expected. Redox signaling is thus likely woven very deeply into the fabric of life. Due to constraints of space, no attempt will be made here to view the entire tapestry. Rather, only a few brief scenes will be described to provide an adequate background for the central comparison of basal metazoans and bilaterians.

Bacteria: masters of metabolic signaling

While bacteria may constitute the simplest extant cells, they are nevertheless far from simple. Redox signaling is among the many sophisticated mechanisms that bacteria can deploy. Control of gene expression commonly occurs in response to changes in redox potential, which is ultimately linked to environmental factors (e.g., light, oxygen, or particular substrates). Two-component signal-transduction systems are often used in adapting bacterial metabolism to environmental conditions (Allen 1993; Georgellis et al. 2001). The Arc system of Escherichia coli provides a well-studied example. Two proteins, ArcA and ArcB, are involved. ArcB is a transmembrane sensor kinase with a loop exposed to the cytoplasm. The cytoplasmic loop contains a conserved histidine residue, which is the site of autophosphorylation. This occurs in response to the redox state of quinone electron carriers of the electron-transport chain (comparable to CoQ in Fig. 1; note that the electron-transport chain of E. coli is similar to, but not entirely the same as the one illustrated). Oxidized forms of quinone inhibit autophosphorylation (Georgellis et al. 2001); if quinones are reduced this inhibition is removed. Subsequent to autophosphorylation, ArcB transphosphorylates the second component, ArcA, which is a global transcriptional regulator. When phosphorylated, ArcA represses the expression of many genes whose products are involved in aerobic respiration and activates many of the genes whose products are involved in fermentation. The logic of the system is thus readily apparent (Allen 1993). The quinone electron carriers remain relatively oxidized and autophosphorylation is inhibited as long as electron transport to the terminal electron acceptor (diatomic oxygen) is possible. If oxygen is not available, electrons “back up” on the electron carriers of the electron-transport chain and these carriers become reduced. Autophosphorylation and transphosphorylation ensue and the components of anaerobic metabolism are activated. Many aspects of bacterial responses to environmental signals are mediated in comparable ways. No doubt such mechanisms are in part responsible for the success that bacteria have enjoyed throughout the history of life.

Eukaryotes: transcending surface-to-volume constraints while simultaneously incorporating a levels-of-selection conflict

For all their sophistication, bacteria are inexorably limited by surface-to-volume constraints (Lane 2005). Because the external membrane system of bacteria is used in chemiosmosis, larger size results in a smaller membrane area and a larger volume. Thus, there is less membrane to convert environmental energy into useable forms and more cytoplasm demanding such conversion. Not surprisingly, bacteria have typically remained small, sacrificing complexity for exquisite environmental tuning coupled to rapid replication. A major innovation of eukaryotes was moving chemiosmotic membranes internally during the endosymbiosis of mitochondria and chloroplasts (Lane 2005). Large eukaryotic cells thus are not necessarily limited in their capacity for energy conversion. As is perhaps typical in the history of life, however, the solution to one problem resulted in the creation of new ones. Because mitochondria and chloroplasts are endosymbionts, and hence evolutionary units capable of heritable variation, redox signaling in eukaryotes is inextricably bound to levels-of-selection synergies and antagonisms (Blackstone 1995). In the stem-lineage eukaryote, metabolic signaling not only served to provide cues for environmental signaling pathways, but also provided the mechanisms by which the symbionts could manipulate the host. Even in modern, crown-group eukaryotes metabolic signaling pathways may remain as vestiges of ancient levels-of-selection conflicts (Blackstone and Green 1999; Blackstone 2000).

An example can illustrate this general point. In a number of animals and perhaps other eukaryotes, release of cytochrome c is a key step in the initiation of programmed cell death via the mitochondrial pathway (Kluck et al. 1997; Bredesen et al. 2006). Cytochrome c is a well-known protein component of the mitochondrial electron transport chain (Fig. 1) and the discovery of its role in programmed cell death was greeted with amazement. Why would an electron carrier of the mitochondrial electron-transport chain signal cell death? The answer likely involves both redox signaling and levels-of-selection conflicts that occurred in the stem-lineage eukaryote (Blackstone and Green 1999; Lane 2005). Descendents of proteobacteria, protomitocondria no doubt were well equipped with redox signaling mechanisms. Sophisticated mechanisms such as the two-component signal-transduction mechanism described earlier cannot be expected to function in this “new” chimeric organism because the host was a very different kind of bacterium. “Generic” signaling mechanisms (Newman 1994) would have a greater likelihood of functioning under these circumstances.

In this context, consider the example described in the introduction. If the environment became stressful for the host and thus unfavorable for replication, yet plenty of substrate remains available, metabolic demand of the host would diminish, and the ATP/ADP ratio of the protomitochondria would approach one. Oxidation of substrate, however, would continue until the electrochemical gradient of the inner protomitochondrial membrane was maximal and the electron carriers were highly reduced. In such circumstances, these electron carriers would donate electrons to diatomic oxygen. In sufficient quantities, ROS thus formed can trigger genetic mutations in the host, leading ultimately to sexual recombination (Blackstone and Kirkwood 2003). As a result, the protomitochondria could thus “engineer” genetically novel host cells that were products of sexual recombination and potentially better able to cope with the original stress. Subsequent to selection, rapid growth and high metabolic demand of the successful host and symbiont population could once again ensue. The electron carriers of the inner protomitochondrial membrane would then become relatively oxidized and only low levels of ROS would be produced.

To enhance this redox signal, the symbiont could manipulate ROS formation by completely blocking the electron-transport chain. Cytochrome c release should be viewed in this context, since its release from the electron-transport chain effectively blocks electron flow (Fig. 1). Cytochrome c release would have the same effect as low metabolic demand of the host—electron carriers “upstream” of cytochrome c would become highly reduced and ROS formation would become maximal. Note that the two major sites of ROS formation in the electron-transport chain are both upstream of cytochrome c (Fig. 1).

As the relationship between the host and the mitochondria became better established, conflicts between the units of selection would require mediation (Michod 2003). To accomplish this, there would need to be selection for more refined signaling mechanisms. “Generic” signaling mechanisms can be effective, but the lack of precision can have costs in fitness. In particular, signaling with high levels of ROS will have a certain amount of collateral damage as otherwise valuable molecules are damaged and require repair or replacement. In this case, selection could favor making cytochrome c itself the signal, rather than the potentially dangerous ROS. ROS formation could thus be diminished without impairing the transduction of the signal (e.g., in addition to deploying antioxidants and uncoupler proteins, by releasing cytochrome c molecules from the mitochondrial intermembrane space—such molecules are not actively participating in electron transfer and thus their release does not block electron flow and upregulate ROS).

This mechanism for triggering sex may have been co-opted into a mechanism for cellular death later in the history of life. In particular, the dynamics of adding a third level of selection (groups of eukaryotic cells) could favor such co-option (Blackstone and Kirkwood 2003). In such a group, selection on the higher-level unit (i.e., the group of cells) may favor eliminating, rather than repairing, damaged cells. Linking a pathway that once led to a repair mechanism (i.e., sex) in a single cell to a pathway leading to elimination of damaged cells (i.e., cell death) in a group would seem to be a rather straightforward co-option.

The history of eukaryotic cells may be littered with similar vestiges of redox signaling. Initially, these mechanisms were used by symbionts to exploit and manipulate host cells. Subsequently, these mechanisms evolved to mediate levels-of-selection conflicts and stabilize the symbiosis on which the eukaryotic cell was built.

Early metazoans: the mouth and its implications

In addition to alleviating surface-to-volume constraints, shifting chemiosmostic membranes internally may have allowed the plasma membrane to develop greater specialization for between-cell interactions. Driven by selection for larger size (Bonner 1998), these interactions may have facilitated the tendency for eukaryotes to form multicellular groups. Such simple groups typically exhibit metabolic gradients between layers of cells, due, for instance, to differences in the availability of substrate and oxygen on the surface versus the interior of the group. These “generic” metabolic gradients may have been used as signals to differentiate somatic and germ cells. Concomitantly, these gradients and their role in signaling would inevitably link levels-of-selection conflicts (now between individual cells and groups of cells) to redox signaling (see subsequently and Blackstone 2000).

The stem-lineage metazoans, perhaps in some ways comparable to modern sponges and placozoans, may have had few if any advantages when compared to other multicellular protists. In this context, one of the most important features derived in animal evolution was the mouth. For a heterotroph, the capacity to sequester and monopolize large amounts of substrate may have been a decisive advantage. At the same time, a large quantity of substrate sequestered in a particular part of the organism would inevitably exaggerate and intensify existing redox gradients. Multicellular redox regulators—collections of metabolically active cells emitting a disproportionate share of redox signals—may have evolved in this context (Fig. 2, Blackstone 2006; Doolen et al. 2007). For instance, if consuming a large amount of substrate triggers contractions of muscle cells involved in circulation, the metabolic demand of these muscle cells can serve as the locus for food-related redox signals. Similarly, in the case of tissues with specialized, energy-intensive functions (e.g., osmoregulation), redox changes related to metabolic demand can signal more general environmental changes. With innovations such as the mouth and multicellular redox regulation, metabolic signaling continued to be of central importance in early evolution of animals.

Fig. 2

Photomicrograph of the base of a polyp of Hydractinia symbiolongicarpus, a colonial cnidarian, stained with H2DCFDA. Once acetate groups are removed by intracellular esterases, H2DCF can be oxidized, typically by peroxides. The resulting fluorescence indicates that this oxidation occurs primarily in the mitochondrion-rich epitheliomuscular cells (the bright, circular areas; each cell is ∼5 μm in diameter). Larger, irregularly shaped, chitin-rich granules autofluoresce and appear toward the outside of the polyp base. The clusters of mitochondrion-rich cells are only found at the base of polyps and are hypothesized to carry out colony-wide multicellular redox regulation (Blackstone 2006; Doolen et al. 2007). For instance, consumption of food by the polyp triggers active contractions of these cells, and this metabolic demand decreases the formation of ROS. Nearby polyps are no longer inhibited by these ROS, and growth of these polyps allows the colony to take advantage of local increases in food as in Fig. 3.

Fig. 2

Photomicrograph of the base of a polyp of Hydractinia symbiolongicarpus, a colonial cnidarian, stained with H2DCFDA. Once acetate groups are removed by intracellular esterases, H2DCF can be oxidized, typically by peroxides. The resulting fluorescence indicates that this oxidation occurs primarily in the mitochondrion-rich epitheliomuscular cells (the bright, circular areas; each cell is ∼5 μm in diameter). Larger, irregularly shaped, chitin-rich granules autofluoresce and appear toward the outside of the polyp base. The clusters of mitochondrion-rich cells are only found at the base of polyps and are hypothesized to carry out colony-wide multicellular redox regulation (Blackstone 2006; Doolen et al. 2007). For instance, consumption of food by the polyp triggers active contractions of these cells, and this metabolic demand decreases the formation of ROS. Nearby polyps are no longer inhibited by these ROS, and growth of these polyps allows the colony to take advantage of local increases in food as in Fig. 3.

Fig. 3

Processed images (Doolen et al. 2007) of genetically identical colonies of P. carnea growing on 18 mm diameter glass cover slips after roughly 30 d of differential feeding (A, control; B, differentially fed). Polyps are bright and circular; stolons are darker and web-like; unencrusted, "inner" areas of cover slip appear dark. The triangular patch of very dense polyp and stolon growth on the lower right edge of colony B is the “fed” area; the remainder of this colony was not directly fed (images courtesy of Kim Cherry Vogt).

Fig. 3

Processed images (Doolen et al. 2007) of genetically identical colonies of P. carnea growing on 18 mm diameter glass cover slips after roughly 30 d of differential feeding (A, control; B, differentially fed). Polyps are bright and circular; stolons are darker and web-like; unencrusted, "inner" areas of cover slip appear dark. The triangular patch of very dense polyp and stolon growth on the lower right edge of colony B is the “fed” area; the remainder of this colony was not directly fed (images courtesy of Kim Cherry Vogt).

The mouth, however, had further implications for animal evolution. Once the capacity to sequester large amounts of substrate evolved, obtaining such large amounts of substrate would become a driving force in evolution. “Super-consumers”—animals with mouths—likely would have perceived the supply of concentrated substrate as patchy in space and time. Long-lived sedentary animals, perhaps using specialized cells (comparable to modern cnidocytes or colloblasts) to entangle and subdue prey, could adapt to such environmental patchiness by morphological plasticity mediated by multicellular metabolic signaling (see next section). Alternatively, in mobile animals coupling the mouth to a battery of sensory equipment would allow seeking out areas that contained concentrated substrate. The mouth, by facilitating consumption, placed additional selective benefits on further sensory direction for that consumption. Once the mouth evolved, heads (and bilaterians) followed.

Early metazoans: environmentally plastic creatures

The tree of life remains poorly known, with many of the branches and their relationships just beginning to be explored (Embley and Martin 2006). The foregoing brief history is intended to suggest what a rigorous comparative analysis cannot yet do: that extensive use of redox and metabolic signaling is a shared primitive feature of all extant organisms. As phylogenies develop and as signaling pathways are elucidated in all of life's branches, a rigorous analysis will become possible. The foregoing is meant to suggest that this assumption remains a plausible one. Further, experimental data does suggest that early-branching metazoans such as cnidarians and sponges employ environmental signaling extensively (Blackstone and Bridge 2005). In particular, long-lived cnidarian colonies are able to adapt to a variable food supply by using metabolic signaling and plasticity.

Podocoryna carnea, a well-studied colonial hydroid, serves as a useful example. Colonies of P. carnea consist of feeding polyps and connecting tube-like stolons. Colonies are grown on glass surfaces in the laboratory and fed brine shrimp nauplii. Typically, such colonies encrust the surface with a dense array of polyps and stolons in the central, older part of the colony and a less dense array in peripheral, newer parts of the colony (Fig. 3A). However, if the colony is differentially fed, so that only some polyps on the periphery receive food, then the fed area of the colony will exhibit denser growth of polyps and stolons as compared to the rest of the colony (Fig. 3B). That metabolic signals are responsible for this differential patterning is suggested by perturbations of the mitochondrial electron-transport chain mimicking greater or lesser amounts of food. The resulting patterns of colony growth are similar to those involving actual perturbations of food supply. Blackstone (2006) further described the basis for these and other results that support the hypothesis of metabolic signaling in colonial hydroids.

Other aspects of metabolic signaling in these organisms may be less well appreciated. By way of comparison, consider relatively derived bilaterians such as mammals. Senescence in mammals is increasingly regarded as the result of a trade-off between the perils of cancer on one hand and the risks of too little self-renewal on the other (Tyner et al. 2002; Radtke and Clevers 2005; Beausejour and Campisi 2006). In these bilaterians, longevity requires tissue maintenance and a capacity for regeneration based on proliferation of unspecialized stem cells. Yet a capacity for too much cellular proliferation puts the organism at risk for hyperproliferative diseases such as cancer. The roots of this trade-off may extend back to the origins of multicellularity. As pointed out by Szathmáry and Wolpert (2003), the fundamental problem in the origin of multicellularity is “… the appropriate down-regulation of cell division ….” Some organisms, however, are able to cope with the risk of hyperproliferation very effectively. For instance, in plants rigid cell walls prevent cell movement and thus proliferative cells cannot become systemic (Buss 1987). Animals, on the other hand, lack any such structural limitations to proliferative cells. Particularly those animals with an extensive capacity for self-renewal are thus expected to be highly vulnerable to cancers. Cnidarians, for instance, seem to have almost unlimited powers of regeneration and self-renewal (Buss 1987; Holstein et al. 2003), e.g., an entire hydra can be reconstituted from a small number of cells. At the same time, many cnidarians seem to be potentially immortal, at least on human timescales (Martinez 2002). A single hydroid clone can be maintained in the laboratory for decades; individual anemones can outlive their human observers; coral colonies in nature may be hundreds or even thousands of years old. A paradox is thus apparent: cnidarians are capable of extreme longevity but at the same time exhibit a seemingly unlimited capacity for self-renewal. Why do not long-lived cnidarians die of cancer?

The answer in part may be metabolic signaling. From a levels-of-selection perspective, the multicellular organism can use metabolic signals to effectively control the replication of cells. Since these signals mimic the ones that the cell uses to control its own metabolism, ignoring these signals inflicts a heavy fitness cost on the cell and its descendents. When confronted with external metabolic signals that direct it to differentiate or die, a cell that continues to replicate may be unable to effectively manage its own metabolism. The well-known Warburg effect (Garber 2006; Lane 2006; Matoba et al. 2006; Trachootham et al. 2006)—the tendency for mammalian cancers to exhibit a glycolytic metabolism even in an aerobic environment—can be understood in this context. Such cells have shut down their mitochondrial metabolism and proceed with highly inefficient glycolysis. Mitochondria are the locus of metabolism and metabolic signaling in a eukaryotic cell (Lane, 2005, 2006), so shutting down mitochondria effectively turns off metabolic signaling. In mammals, such cells are now impossible to regulate without medical intervention. Tellingly, therapeutics that activate mitochondria and metabolic signaling trigger pathways of cell death in cancer cells, while leaving normal cells unaffected (Haridas et al. 2001; Blackstone et al. 2005). In contrast to mammals, cnidarians and other potentially long-lived basal metazoans (e.g., sponges) may be able to keep hyperproliferation in check because they have a more extensive repertoire of metabolic signaling.

Redox signaling with metabolic by-products—ROS—serves as a specific example of this general pattern. Eukaryotic cells use ROS in a variety of intracellular and intercellular signaling pathways (Finkel 2003; Becker 2004; Filomeni et al. 2005). These ROS cannot be downregulated (e.g., by chemical or enzymatic antioxidants) without interfering with the control of normal cellular metabolism. Hence, ROS can be a very effective intercellular signal (Fig. 4; Blackstone 2000). ROS such as peroxide need no receptor and enter cells by diffusing through the plasma membrane. Peroxides can activate signaling pathways involving proteins rich in cysteine or histidine by oxidizing these residues (van Montfort et al. 2003; Filomeni et al. 2005; Lee and Helmann 2006). The activated pathway can then lead to cellular differentiation; at the same time the peroxide is converted to water. On the other hand, if the signaling pathway is mutationally deactivated, peroxides will not be converted to water and may damage components of the cell or trigger active or passive pathways to cellular death. In either case, by “resisting” the differentiation pathway the cell lowers its own fitness.

Fig. 4

Schemata depicting how ROS in general and peroxide in particular can lead to “honest” metabolic signaling. Peroxide enters the cell by diffusing through the plasma membrane; no receptor is necessary. By oxidizing cysteine residues and triggering the formation of disulfide bonds, peroxide activates proteins involved in a signaling pathway that leads to cellular differentiation (A). In the process of activating the signaling pathway, peroxides are converted to water. If, on the other hand, this pathway is deactivated by mutations, peroxides build up in the cell (B). Ultimately, these peroxides trigger cell death by apoptosis or necrosis. Peroxide, a metabolic by-product, can therefore be used effectively by a multicellular organism to suppress cell-level variation (represented by the mutationally deactivated signaling pathway) and favor cellular differentiation. Additional cell-level variation (e.g., upregulation of antioxidant enzymes) could control the excess peroxide, but this would also interfere with normal cellular metabolism, thus again exacting a cost in fitness to the variant cell.

Fig. 4

Schemata depicting how ROS in general and peroxide in particular can lead to “honest” metabolic signaling. Peroxide enters the cell by diffusing through the plasma membrane; no receptor is necessary. By oxidizing cysteine residues and triggering the formation of disulfide bonds, peroxide activates proteins involved in a signaling pathway that leads to cellular differentiation (A). In the process of activating the signaling pathway, peroxides are converted to water. If, on the other hand, this pathway is deactivated by mutations, peroxides build up in the cell (B). Ultimately, these peroxides trigger cell death by apoptosis or necrosis. Peroxide, a metabolic by-product, can therefore be used effectively by a multicellular organism to suppress cell-level variation (represented by the mutationally deactivated signaling pathway) and favor cellular differentiation. Additional cell-level variation (e.g., upregulation of antioxidant enzymes) could control the excess peroxide, but this would also interfere with normal cellular metabolism, thus again exacting a cost in fitness to the variant cell.

The role of the Wnt signaling pathway can be viewed in this context. In bilaterians, Wnt ligands activate cell-surface receptors and stabilize the cytoplasmic protein β-catenin (Bowerman 2005). In turn, β-catenin interacts with FOXO transcriptional factors, which are regulated by insulin and oxidative-stress pathways (Essers et al. 2005). Using metabolic signals, these pathways balance cellular proliferation and quiescence. The Wnt signaling pathway is found in cnidarians (Hobmayer et al. 2000; Kusserow et al. 2005) and may fulfill signaling roles that are more commonly filled by Hox genes in bilaterians (Kamm et al. 2006).

Bilaterians: assembly-line animals

In competition for resources, animals with a mouth likely had a significant advantage compared to sponges, placozoans, and other multicellular protists. One way to employ this advantage is to use specialized cells (e.g., cnidocytes, colloblasts) to entangle and subdue prey that is then engulfed by the mouth. Early in the history of this mouth-bearing clade, however, another kind of animal evolved. Locating a battery of sensory equipment near the mouth and employing active movement, these first bilaterians were super consumers. Developing rapidly, dispersing widely, colonizing readily, these animals could efficiently locate areas of concentrated resources and rapidly exploit them. Slower-developing organisms that lacked the capacity for directed movement were at a competitive disadvantage. Placozoans may have been strong competitors of these first bilaterians, since they too seem to have the capacity to locate and exploit concentrations of resources (personal observation). Placozoans lack features crucial to general success in this competition, in particular a head and mouth, although their small size and simplicity may have allowed them to continue to succeed in some instances. The greater complexity of bilaterians, however, would impose a cost. Rapid development of a complex body plan could only be achieved by precise and unambiguous signaling mechanisms.

In this context, the limitations of metabolic signaling must be recognized. As perhaps is generally found with “generic” mechanisms (Newman 1994), metabolic signaling may be ambiguous under some circumstances. Consider the lack of oxygen and the lack of metabolic demand: both will shift the redox state of electron carriers in the direction of reduction. The metabolic signals will thus be similar even though the appropriate adaptive responses are very different (e.g., suppress aerobic and activate anaerobic metabolism versus dormancy and/or sexual recombination). Because very different environmental stimuli may produce similar redox signals, the complexity of a signaling system based on metabolism may be limited.

The canonical Hox system may be considered one possible solution to the requirement for precise, unambiguous signaling mechanisms that govern rapid, canalized development. The organization of Hox-gene clusters characteristically reflects domains of expression along the anterior–posterior body axis. Through their combined action, these genes are primarily responsible for patterning tissues along that axis (Kamm et al. 2006). As pointed out above, however, limiting metabolic signaling may have had costs. Because bioenergetic metabolism necessarily links external sources and sinks of electrons, monitoring these sources and sinks provides “honest” signals for controlling cell-level variation. Any mutation-based alteration of metabolism will perturb these signals. Cancers are notorious for their altered metabolism (e.g., the Warburg effect and overproduction of ROS) (Garber 2006; Lane 2006; Matoba et al. 2006; Trachootham et al. 2006). Because metabolic signaling is “honest” signaling, it provides effective policing of cancers and other mutation-based events that affect metabolism (Blackstone et al. 2005). Limiting metabolic signaling may thus have required other mechanisms to police cell-level variation. As a consequence, adults of the stem bilaterian may have exhibited diminished somatic cell potency, determinate growth, and a sequestered germ line (Blackstone and Ellison 2000; Blackstone and Jasker 2003).

Bilaterian senescence evolved as a by-product of restricted somatic cell potency, itself a mechanism of cell policing required when metabolic signaling is limited. Senescence became a trade-off between the perils of cancer on one hand and the risks of too little self-renewal on the other (Tyner et al. 2002; Radtke and Clevers 2005; Beausejour and Campisi 2006). Note that these statements are not in contradiction with evolutionary theories of aging (Rose 1991). Rather, these provide a mechanism for an evolutionary view of bilaterian senesence. By this view, the stem-lineage bilaterian (in contrast to basal metazoans) evolved a life history (rapid development, dispersal, colonization, and reproduction) that in turn allowed senescence to evolve. The mechanism that actually led to senescence was restricted somatic cell potency, a mechanism to police cell-level variation in the absence of metabolic signaling.

Conclusions and predictions

The foregoing suggests a research program and at the same time predicts various outcomes from such a program. Important aspects of this program include:

  1. Reconstruction of the tree of life, particularly the tree of basal bilaterians, basal animals, and related taxa. The view outlined here is based on our present understanding of this tree, as for instance described in other papers from the symposium.

  2. Elucidation of signaling pathways in a much greater array of basal bilaterians, basal animals, and related taxa than the current model systems. Only with such a data set and with well-supported phylogenies can the evolution of animal signaling be explicitly analyzed. While such a goal is a crucial requirement of comparative animal biology, admittedly in the current climate for funding its realization will not occur suddenly but rather gradually over a long period of time.

  3. Characterization of homoplasy in the animal tree of life. For instance, the view outlined here suggests that putative reversals of bilaterian traits (e.g., loss of a germ line, loss of senescence) should be accompanied by reversals in signaling (e.g., rederiving a greater role for metabolic signaling).

  4. Illumination of the cellular and molecular mechanisms that allow cnidarians to simultaneously escape both cancer and senescence. The potential biomedical utility of such information should make this a readily achievable goal. Instances of hyperproliferation are known in cnidarians (e.g., hyperplastic stolons of colonial hydroids). How is this hyperproliferation controlled? How do cnidarians respond to perturbations (e.g., radiation) that typically initiate mammalian cancers?

Acknowledgments

Many thanks to the organizers of the symposium, “Key transitions in animal evolution,” for all their efforts. Thanks to the American Microscopical Society, the Society for Integrative and Comparative Biology, and the National Science Foundation for supporting the symposium. Helpful comments on the article were provided by J. Gutterman, R. Meganathan, R. Michod, an anonymous reviewer, and the editor, Harold Heatwole. The National Science Foundation (IBN-00-90580 and EF-05-31654) provided support for the research.

References

Allen
JF
Control of gene expression by redox potential and the requirement for chloroplast and mitochondrial genomes
J Theor Biol
 , 
1993
, vol. 
165
 (pg. 
609
-
31
)
Beausejour
CM
Campisi
J
Balancing regeneration and cancer
Nature
 , 
2006
, vol. 
443
 (pg. 
404
-
5
)
Becker
LB
New concepts in reactive oxygen species and cardiovascular reperfusion physiology
Cardiovascul Res
 , 
2004
, vol. 
61
 (pg. 
461
-
70
)
Blackstone
NW
A units-of-evolution perspective on the endosymbiont theory of the origin of the mitochondrion
Evolution
 , 
1995
, vol. 
49
 (pg. 
785
-
96
)
Blackstone
NW
Redox control and the evolution of multicellularity
BioEssays
 , 
2000
, vol. 
22
 (pg. 
947
-
53
)
Blackstone
NW
Multicellular redox regulation: integrating organismal biology and redox chemistry
BioEssays
 , 
2006
, vol. 
28
 (pg. 
72
-
7
)
Blackstone
NW
Bridge
DM
Model systems for environmental signaling
Integr Comp Biol
 , 
2005
, vol. 
45
 (pg. 
605
-
14
)
Blackstone
NW
Ellison
AM
Maximal indirect development, set-aside cells, and levels of selection
J Exp Zool (MDE)
 , 
2000
, vol. 
288
 (pg. 
99
-
104
)
Blackstone
NW
Green
DR
The evolution of a mechanism of cell suicide
BioEssays
 , 
1999
, vol. 
21
 (pg. 
84
-
8
)
Blackstone
NW
Jasker
BD
Phylogenetic considerations of clonality, coloniality, and mode of germline development in animals
J Exp Zool (MDE)
 , 
2003
, vol. 
297B
 (pg. 
35
-
47
)
Blackstone
NW
Kirkwood
TBL
Hammerstein
P
Mitochondria and programmed cell death
Genetic and cultural evolution of cooperation
 , 
2003
Cambridge, MA
MIT Press
(pg. 
309
-
25
)
Blackstone
NW
Kelly
MM
Haridas
V
Gutterman
JU
Mitochondria as integrators of information in an early-evolving animal: insights from a triterpenoid metabolite
P Roy Soc Lond B
 , 
2005
, vol. 
272
 (pg. 
527
-
31
)
Bonner
JT
The orgins of multicellularity
Integrat Biol
 , 
1998
, vol. 
1
 (pg. 
27
-
36
)
Bowerman
B
Oxidative stress and cancer: a β-catenin convergence
Science
 , 
2005
, vol. 
308
 (pg. 
1119
-
20
)
Bredesen
DE
Rao
RV
Mehlen
P
Cell death in the nervous system
Nature
 , 
2006
, vol. 
443
 (pg. 
796
-
802
)
Buss
LW
The evolution of individuality
 , 
1987
Princeton, NJ
Princeton Press
Doolen
JF
Geddes
GC
Blackstone
NW
Multicellular redox regulation in an early-evolving animal treated with glutathione
Physiol Biochem Zool
 , 
2007
, vol. 
80
 
3
(pg. 
317
-
25
)
Embley
TM
Martin
W
Eukaryotic evolution, changes and challenges
Nature
 , 
2006
, vol. 
440
 (pg. 
623
-
30
)
Essers
MAG
de Vries-Smits
LMM
Barker
N
Polderman
PE
Burgering
BMT
Korswagen
HC
Functional interaction between β-catenin and FOXO in oxidative stress signaling
Science
 , 
2005
, vol. 
308
 (pg. 
1181
-
4
)
Filomeni
G
Rotillio
G
Ciriolo
MR
Disulfide relays and phosphorylative cascades: partners in redox-mediated signaling pathways
Cell Death Diff
 , 
2005
, vol. 
12
 (pg. 
1555
-
63
)
Finkel
T
Oxidant signals and oxidative stress
Curr Opin Cell Biol
 , 
2003
, vol. 
15
 (pg. 
247
-
54
)
Garber
K
Energy deregulation: licensing tumors to grow
Science
 , 
2006
, vol. 
312
 (pg. 
1158
-
9
)
Georgellis
D
Kwon
O
Lin
ECC
Quinones as the redox signal for the Arc two-component system of bacteria
Science
 , 
2001
, vol. 
292
 (pg. 
2314
-
6
)
Haridas
V
Arntzen
CJ
Gutterman
JU
Avicins, a family of triterpenoid saponins from Acacia victoriae (Bentham), inhibit activation of nuclear factor-κB by inhibiting both its nuclear localization and its ability to bind DNA
Proc Natl Acad Sci USA
 , 
2001
, vol. 
98
 (pg. 
11557
-
62
)
Holstein
TW
Hobmayer
E
Technau
U
Cnidarians: an evolutionarily conserved model system for regeneration
Devel Dyn
 , 
2003
, vol. 
226
 (pg. 
257
-
67
)
Hobmayer
B
Rentzsch
F
Kuhn
K
Happel
CM
von Laue
CC
Snyder
P
Rothbächer
U
Holstein
TW
WNT signaling molecules act in axis formation in the diploblastic metazoan Hydra
Nature
 , 
2000
, vol. 
407
 (pg. 
186
-
9
)
Kamm
K
Schierwater
B
Jakob
W
Dellaporta
SL
Miller
DJ
Axial patterning and diversification in the Cnidaria predate the Hox system
Curr Biol
 , 
2006
, vol. 
16
 (pg. 
920
-
6
)
Kluck
RM
Bossy-Wetzel
E
Green
DR
Newmeyer
DD
The release of cytochrome c from mitochondria: a primary site for Bcl-2 regulation of apoptosis
Science
 , 
1997
, vol. 
275
 (pg. 
1132
-
6
)
Kusserow
A
, et al.  . 
Unexpected complexity of the Wnt gene family in a sea anemone
Nature
 , 
2005
, vol. 
433
 (pg. 
156
-
60
)
Lane
N
Oxygen: the molecule that made the world
 , 
2002
Oxford
Oxford University Press
Lane
N
Power, sex, suicide: mitochodria and the meaning of life
 , 
2005
Oxford
Oxford University Press
Lane
N
Power games
Nature
 , 
2006
, vol. 
443
 (pg. 
901
-
3
)
Lee
J-W
Helmann
JD
The PerR transcription factor senses H2O2 by metal-catalysed histidine oxidation
Nature
 , 
2006
, vol. 
440
 (pg. 
363
-
7
)
Martin
W
Russell
MJ
On the origins of cells: a hypothesis for the evolutionary transitions from abiotic geochemistry to chemoautotrophic prokaryotes, and from prokaryotes to nucleated cells
Philos T Roy Soc B
 , 
2003
, vol. 
358
 (pg. 
59
-
85
)
Martínez
DE
Hughes
RN
Senescence and rejuvenation in asexual metazoans
Reproductive biology of the invertebrates XI progress in asexual reproduction
 , 
2002
New York
John Wiley
(pg. 
115
-
40
)
Matoba
S
Kang
J-G
Patino
WD
Wragg
A
Boehm
M
Gavrilova
O
Hurley
PJ
Bunz
F
Hwang
PM
p53 regulates mitochondrial respiration
Science
 , 
2006
, vol. 
312
 (pg. 
1650
-
3
)
Michod
RE
Hammerstein
P
Cooperation and conflict mediation during the origin of multicellularity
Genetic and cultural evolution of cooperation
 , 
2003
Cambridge, MA
MIT Press
(pg. 
291
-
307
)
Newman
S
Generic physical mechanisms of tissue morphogenesis: a common basis for development and evolution
J Evol Biol
 , 
1994
, vol. 
7
 (pg. 
467
-
88
)
Radtke
F
Clevers
H
Self-renewal and cancer of the gut: two sides of a coin
Science
 , 
2005
, vol. 
307
 (pg. 
1904
-
9
)
Rose
MR
Evolutionary biology of aging
 , 
1991
Oxford
Oxford University Press
Szathmáry
E
Wolpert
L
Hammerstein
P
The transition from single cells to multicellularity
Genetic and cultural evolution of cooperation
 , 
2003
Cambridge, MA
MIT Press
(pg. 
271
-
90
)
Trachootham
D
, et al.  . 
Selective killing of oncogenically transformed cells through a ROS-mediated mechanism by β-phenylethyl isothiocyanate
Cancer Cell
 , 
2006
, vol. 
10
 (pg. 
241
-
52
)
Tyner
SD
, et al.  . 
p53 mutant mice that display early aging-associated phenotypes
Nature
 , 
2002
, vol. 
415
 (pg. 
45
-
53
)
Valentine
JW
Ancestors and urbilateria
Evol Dev
 , 
2006
, vol. 
8
 (pg. 
391
-
3
)
van Montfort
RLM
Congreve
M
Tisi
D
Carr
R
Jhoti
H
Oxidation state of the active-site cysteine in protein tyrosine phosphatase 1B
Nature
 , 
2003
, vol. 
423
 (pg. 
773
-
7
)

Author notes

From the symposium “Key Transisions in Animal Evolution” presented at the annual meeting of the Society for Integrative and Comparative Biology, January 3–7, 2007, at Phoenix, Arizona