Metazoan diversification occurred during a time when atmospheric oxygen levels fluctuated between 15 and 30%. The hypoxia-inducible factor (HIF) is a primary regulator of the adaptive transcriptional response to hypoxia. Although the HIF pathway is highly conserved, its complexity increased during periods when atmospheric oxygen concentrations were increasing. Thus atmospheric oxygen levels may have provided a selection force on the development of cellular oxygen-sensing pathways.
Changing levels of atmospheric gasses such as oxygen and carbon dioxide, whether due to human industrial activity or to the processes of nature, have significant implications for the Earth's biota (1, 2, 9, 23, 62). During the 4,500 million year history of the planet, the composition of the thin and delicate lower terrestrial atmosphere has varied greatly and played a central role in driving the evolution of both plants and animals (1, 3, 10, 40, 41). Changes in the composition of the atmosphere are associated with key events in the evolution of life on Earth, including the origination and radiation of metazoans, selective and mass extinctions, and periods of plant and animal gigantism (8, 10, 21, 23, 40). Although the composition of the atmosphere impacted greatly on plant and animal evolution, the converse is also the case with the evolution of living organisms strongly influencing the composition of the atmosphere (for example, the development of photosynthesis was a primary event in the oxygenation of the atmosphere) (1, 30, 45). Thus the intimate and bidirectional relationship that exists between the composition of the atmosphere and the evolution of living organisms has been key to the development of life on Earth.
Following the formation of the planet ∼4,500 million years ago, the first primitive cellular life arose from a single source likely within 500 million years (28). Over the following 1,000 million years, the evolution of simple single-cell prokaryotic life included the growth of masses of cyanobacteria in the Earth's oceans, which evolved the ability to harness the energy inherent in sunlight to fuel metabolism through the process of photosynthesis (53). This process generated molecular oxygen (O2) as a highly reactive and toxic by-product. Initially, the oxygen produced was removed from the atmosphere during the oxidation of inorganic minerals such as iron. After the incorporation of oxygen into these minerals was saturated, the slow and inexorable rise in atmospheric O2 that occurred over the following 3,000 million years (FIGURE 1). initially proved lethal to the vast majority of unicellular anaerobic life on Earth, which could not survive the reactive (oxidative) chemical properties of oxygen (20, 59). However, a group of more complex organisms, the eukaryotes, which formed as a result of a merger (symbiosis) between different families of bacterial cells (28), which had discreet internal membrane-enclosed compartments, could not only withstand the toxic effects of oxygen but utilize its chemical energy to generate cellular ATP in a manner many times more efficient than anaerobic metabolism through the process of oxidative phosphorylation (50). The consequences of this ancient endosymbiotic event, mitochondria, are integral organelles in all eukaryotic cells today and are responsible for the generation of the majority of cellular ATP in metazoans (28, 38). Once armed with this new and highly efficient ability to harness the chemical energy of oxygen matched with a plentiful supply of the critical fuel in the atmosphere, the evolution of complex multicellular organisms, which utilized oxidative metabolism, progressed at a breathtaking rate initiated during the Cambrian era 542–488 million years ago and continued throughout the Phanerozoic eon (from 542 million years ago until the present) (15, 18).
Although central to the evolution of life on Earth, oxygen is not the only atmospheric gas to exert a potent impact on the biota. For example, dramatic changes in atmospheric CO2 levels over the history of the planet also likely impacted on plant and animal evolution (1–3, 40, 41, 49). Indeed, the ratio of CO2 to O2 may have been an important determinant of evolutionary events. However, the role of CO2 in directing evolutionary processes is beyond the scope of the current review.
Matching Atmospheric Oxygen Levels to Evolution
Although photosynthesis in the oceans likely began ∼3,000 million years ago (43), it was not until ∼1,500 million years ago, when photosynthesis-derived oxygen began to accumulate to significant levels in the atmosphere. While atmospheric oxygen levels continued to increase over the following 1,000 million years, the existing life on Earth during this time remained as simple singe-cell organisms (FIGURE 1). However, when atmospheric oxygen levels reached between 15 and 20% between 550 and 600 million years ago, the first animal body plans developed, and the evolution of metazoans started in earnest (9). During the following ∼550 million years of metazoan evolution, oxygen levels in the atmosphere have been estimated to have fluctuated between lows of ∼15% and highs of >30% (4, 6, 9, 10, 12, 23). During this period, metazoan evolution progressed from simple aquatic species such as sea sponges and corals to a dramatic array of marine and terrestrial species including Homo sapiens (H. sapiens) (5).
Evidence for fluctuating atmospheric oxygen concentrations over the last ∼550 million years comes mainly from studying the geochemical cycles of carbon and sulfur (6, 9, 12, 23). Independent models indicate a significant rise in atmospheric oxygen 300–260 million years ago to levels as high as 30–31%, which has been largely attributed to the evolution of large land plants and which is evidenced by the burial of woody plants in coal swamps along with the burial of organic carbon in the oceans (11, 23). In contrast, oxygen levels dropped to levels as low as 15% ∼190–140 million years ago (12). How the fluctuating levels of atmospheric oxygen over the last ∼550 million years impacted on evolution of metazoans is an area in need of further investigation (9, 23). However, existing physiological studies indicate that fluctuations in atmospheric oxygen may have had a significant role to play in helping to shape evolution. For example, studies in fruit flies (Drosophila) have indicated that oxygen is a key determinant of body size in generational studies of D. Melanogaster bred in atmospheric hypoxia (8, 9, 25, 67). Fluctuating oxygen concentrations have also been linked to key stages in evolution over the last ∼550 million years, including the origin of the first animal body plans, the conquest of land by animals, gigantism of various species, and mass extinctions (9).
Because Metazoans evolved to utilize molecular oxygen as the main electron acceptor in oxidative metabolism and energy production, the ability to adapt to and thrive in changing (sometimes excessive and sometimes limiting) oxygen conditions is a key component of evolutionary selection. Indeed, it is likely that the evolution of such pathways was key to the success of the winners of metazoan evolution. Developing our understanding of how atmospheric oxygen levels contributed to the evolution of such pathways may give insight into the impact of atmospheric oxygen on metazoan development.
Hypoxia in Human Health and Disease
In healthy humans, there is a range of physiological oxygen levels within the tissues of the body, ranging from Po2 values of ∼100 Torr in the alveoli of the lungs to less than 10 Torr in tissues such as the medulla of the kidney and the retina (61). Within all of these tissues, the chemical reduction of molecular oxygen in the mitochondria of individual cells during the process of oxidative phosphorylation is central to oxidative metabolism and bioenergetic homeostasis. Because of this, any insufficiency in the availability of molecular oxygen represents a severe threat to continued physiological function and indeed survival. Hypoxia, which occurs when oxygen levels in the microenvironment of a cell, tissue, or organism are reduced relative to the normal physiological state, is associated with a range of physiological and pathophysiological processes (53).
Physiological hypoxia is an important microenvironmental signal in a range of processes including new blood vessel formation (angiogenesis) during development and wound healing, the regulation of vascular tone, and the response to exercise (48). However, tissue hypoxia is also associated with a diverse and wide range of pathophysiological processes including (but not limited to) vascular disease, chronic inflammation, and cancer (33, 48, 61). In vascular diseases such as atherosclerosis and stroke, vascular occlusion leads to acute or chronic tissue ischemia with resultant hypoxia. In chronic inflammatory disease, the greatly increased metabolism of inflamed tissue due to immune cell infiltration matched with vascular dysfunction leads to tissue hypoxia (17, 24, 44, 60). In cancer, the growth of a tumor away from the local blood supply eventually leads to tumor hypoxia. In all of these cases, the induction of a genetic response to hypoxia leads to the expression of genes that are essentially adaptive (or maladaptive in the case of cancer). Seminal discoveries in the last 20 years have greatly enhanced our understanding of the molecular mechanisms underpinning this critical response (33, 53).
Mechanisms of Adaptation to Hypoxia
Molecular oxygen was central to the evolution of multicellular life on Earth through the provision of the fuel to efficiently generate high levels of cellular ATP during oxidative phosphorylation (53). Consequently, during the course of evolution, many organisms have become dependent on a constant supply of oxygen to effectively function. Because of this, it is not entirely surprising that most multicellular organisms also evolved a molecular mechanism by which to respond to conditions where oxygen demand exceeds supply (hypoxia) with the induction of genes that encode proteins that increase oxygen supply and modulate metabolic activity in a hypoxic tissue/organism (33, 53, 61).
A central pathway that coordinates this response in all metazoans involves a transcriptional regulator known as the hypoxia-inducible factor (HIF; Refs. 53, 54). The HIF pathway is highly conserved across species and exists in mammals such as humans as well as in more ancient lineages such as nematodes and corals (27). In humans, HIF consists of one of three oxygen-sensitive alpha subunits (HIFα) and a constitutively expressed beta subunit (HIFβ). Under conditions when sufficient oxygen is available to satisfy bioenergetic requirements, the oxygen not consumed by mitochondrial oxidative phosphorylation is available to permit the activity of a family of four HIF hydroxylase enzymes that hydroxylate defined proline and asparagine residues on the HIFα subunit (26, 33). Three HIF proline hydroxylases (termed PHD1/2/3) have been described that hydroxylate proline residues in the NH2-terminal oxygen degradation domain (NODD; P402) and the COOH-terminal ODD (CODD: P564) on human HIF-1α. This results in targeting of the HIFα subunit as a substrate for an E3 ubiquitin ligase known as the von Hipple Lindau protein (pVHL), which forms a ubiquitin ligase complex leading to the ubiquitination and immediate degradation of HIFα through the ubiquitin-proteasome pathway (FIGURE 2). A further hydroxylation at asparagine 803 [in the COOH-terminal transcriptional activation domain (CTAD)] by factor inhibiting HIF (FIH) prevents transcriptional activity of any HIF that has not been degraded. Thus, under conditions of normoxia, HIF is kept in a state of repression by the activity of the HIF hydroxylases (FIGURE 2).
When oxygen demand exceeds supply, the high affinity of the mitochondrial electron transport chain for oxygen means that virtually all available oxygen is consumed by the process of oxidative phosphorylation leading to a lack of available oxygen for the HIF hydroxylases and their subsequent inhibition (26). Under these conditions, HIFα subunits accumulate within the cytoplasm of the cell, translocate into the nucleus, and bind to HIFβ subunits to form transcriptionally active HIF heterodimers, which coordinate the expression of a family of HIF-dependent adaptive genes. HIF can directly activate the expression of hundreds of direct (primary) target genes and likely impacts on the expression of thousands of secondary genes (53). Genes directly under the control of HIF in mammalian cells include those that contribute to the control of angiogenesis, vascular tone, metabolism, and erythropoesis. The net effect of the expression of these genes is to increase blood and oxygen supply to the hypoxic tissue and also induce a metabolic switch favoring the anaerobic generation of ATP through glycolysis (61). This system is conserved throughout the metazoans in various levels of complexity, as will be discussed below. The central importance of the HIF pathway to physiological processes is exemplified by the fact that homozygous knockout of either pVHL, PHD2, HIF-1α, or HIF-2α leads to embryonic lethality in mice (33). Furthermore, various conditional knockout studies have revealed important tissue-specific roles for HIF in processes as divergent as immune cell survival, kidney fibrosis, and skeletal muscle function during endurance training (19, 29, 39, 57, 63). In subsequent sections of this review, we map the development of components of the HIF pathway to the radiation of metazoans and to changes that occurred in atmospheric oxygen levels during the course of the Earth's history. Our intent is to explore whether the evolving complexity of the HIF pathway is correlated with the prevailing atmospheric oxygen levels of the planet during metazoan evolution.
The Evolution of Components of the HIF Pathway
Orthologs of various components of the human PHD/FIH-HIF-VHL pathway are conserved in organisms as simple as the nematode C. elegans where they also serve hypoxia-sensing functions (22, 47). Notably, however, the components of the HIF pathway have developed a higher degree of complexity possibly because oxygen-sensing requirements became more demanding in larger animals with respiratory/circulatory systems and atmospheric oxygen levels changed. This complexity is reflected by increasing numbers of isoforms of the various components of the pathway (see below). Here, we have examined the conservation of genes with homology to components of the human HIF pathway (HIF-1α/2α/3α, HIF-1β, FIH, PHD1/2/3, pVHL, and CBP/p300) in representative species from the main metazoan Phyla (mammals, birds/reptiles, amphibians, fish, arthropods, and nematodes). Gene homologies were determined using the Homologene search engine (http://www.ncbi.nlm.nih.gov/homologene) and individual papers, which are referenced where appropriate.
Homologs of the gene encoding human HIF-1α (HIF1A) are conserved in the genomes of organisms as primitive as nematodes, coral, and sea anemones and thus appear to be expressed in all metazoans (27). This degree of conservation is striking and underpins the importance of effective oxygen-sensing pathways for the development of all metazoans, irrespective of their natural habitat. Functions of HIF in lower organisms include metabolic switches (e.g., in C. elegans, the HIF pathway is a regulator of metabolism) and tracheal branching in Drosophila (27).
Metazoans, which evolved relatively early, such as coral, worms, and fruit flies have a single HIFα isoform that is most similar to human HIF1α. However, some time after the division of arthropods and fish, which occurred at least 460 million years ago, a second HIF isoform evolved (termed HIF-2α, encoded by the EPAS gene). Although HIF2α is absent in arthropods, it is present in fish species such as zebrafish (42, 51). Following the division of mammals and birds/reptiles that occurred at least 312 million yeas ago, a third isoform (HIF-3α) evolved that is exclusively found in mammals. This indicates the development of increasing numbers of HIFα isoforms with increasing complexity of oxygen-delivery systems during metazoan evolution (27, 33).
As well as increases in the number of HIFα isoforms during metazoan evolution, the rate and mode of molecular evolution has recently been documented in fishes and mammals and has been linked to variable environmental oxygen levels (42, 52). Interestingly, the rate and mode of molecular evolution of HIF-1α in fish differs somewhat from that in mammals, indicating that the different aquatic oxygen tensions experienced by fish may have had a differential impact on the evolution of HIF-1α (16, 42). The molecular evolution of HIF in fish species has been discussed in more detail elsewhere (16, 42, 52).
Thus the early appearance of the HIF1A gene and its constant expression across species independent of fluctuations, which have occurred in atmospheric oxygen levels during the evolution of metazoans, indicates that the HIF pathway appeared early in the evolution of metazoans and is a primary regulator of oxygen-/hypoxia-dependent responses. Although a HIFα isoform has been present since the radiation of the metazoans, the molecular complexity of the molecule has increased over the course of evolution, with primitive forms of HIFα expressing single oxygen-sensitive regulatory domains and more evolved forms of HIFα containing multiple levels of oxygen dependence (27).
The HIF-1β subunit, which is constitutively expressed in mammalian cells and is synonymous with the ARNT protein, is also highly conserved, and orthologs for its gene (ARNT) have been identified in fish (52), fruit flies (65), and nematodes (46), indicating that, like HIF-1α, this gene evolved before the divergence of nematodes and mammals 580–520 million years ago.
Prolyl hydroxylases are a large family of proteins with diverse functions including intracellular signaling and extracellular matrix remodeling. In mammals, three isoforms of HIF prolyl hydroxylases with distinct tissue-specific expression profiles, which are key oxygen sensors regulating the stability of HIF, have been described (33, 36). The three isoforms of HIF prolyl hydroxylases expressed in humans are called PHD1, PHD2, and PHD3 (encoded by the EGLN2, EGLN1, and EGLN3 genes, respectively). PHD2 is the isoform mostly responsible for the regulation of HIF and is thought to be the ancestral form (35). This is supported by evidence that the PHD isoform expressed in the parasitic eukaryotic protist Perkinsus olseni (P. olseni) shares a higher degree of homology with PHD2 than the other isoforms (35). Like HIF, the PHDs are a very well conserved family of genes, and orthologs can be found in flies, worms, beetles, and coral (27). Interestingly, PHD-like enzymes described in P. olseni have been implicated in virulence and pathogenicity, possibly giving a clue to the early evolutionary history of this gene (35). However, the number of PHD isoforms is decreased in lower species. For example, Nematodes and flies have a single PHD isoform known as egl9. Although beetles have two conserved putative sites for proline hydroxylation, only a single PHD has so far been found, leading to the conclusion that the evolution of multiple PHD isoforms occurred after the division of the arthropods and the fish. PHD3 is present in amphibians and fish but not in arthropods and thus represents the second member of the PHD family to evolve (31). A third PHD isoform (PHD1) is present only in mammals and thus appears to have evolved after the division of the mammals and the birds/reptiles (FIGURE 3). The conservation of PHD enzymes is matched by the conservation of target proline residues for hydroxylation in the HIFα proteins expressed (27). Interestingly, recent work has also identified a positively selected haplotype of PHD2 in human populations resident in the high-altitude Tibetan highlands (56). Thus PHDs, like HIF, are highly conserved; however, they show an increased complexity over time, as demonstrated by the appearance of multiple PHD isoforms in vertebrates vs. a single isoform in invertebrates.
Factor Inhibiting HIF
FIH is an asparagine hydroxylase that is critical for the fine tuning of the HIF response (33, 37, 66). The FIH gene demonstrates a variable conservation throughout evolution. Functional FIH homologs are present in mammals but are lacking in some arthropods and nematodes such as D. melanogaster and C. elegans, respectively. However, further analyses revealed the presence of FIH in species that evolved earlier, such as corals, and in species that evolved between the flies and the worms, such as beetles (27). Interestingly, the lack of FIH in D. melanogaster and C. elegans is coupled with a loss of conservation of the COOH-terminal activation domain in HIFα, which harbors the site of hydroxylation by FIH. This variable conservation makes FIH unique among the proteins involved in oxygen sensing to the HIF pathway from an evolutionary perspective.
It has been hypothesized that the variable conservation of the FIH gene is due to its loss from a number of “mono TAD” invertebrate species including D. melanogaster and C. elegans, whereas the “dual TAD” regulatory mechanisms are present in species that evolved before and after (27). In this work, the authors have have hypothesized a number of possible reasons for the loss of these genes in some invertebrates, including the possibility that increased control mechanisms are required for increased complexity in higher species and the possibility that FIH is superfluous in some organisms such as flying insects (27).
The von Hipple Lindau Protein
The von Hipple Lindau gene (VHL), which encodes the von Hipple Lindau protein (pVHL), was first cloned in 1993 (34). Genes with a high degree of similarity to VHL have been described in humans, primates, rodents, and also nematodes (32, 55). Indeed, Bishop et al. (13) demonstrated conservation of the HIF-1/VHL-1/EGL-9 pathway in C. elegans (13). Thus, like HIF-1α and PHD2, VHL is a highly conserved gene that is linked to the regulation of HIF in all metazoans.
Transcriptional Co-activators CBP and p300
The transcriptional co-activators CBP and p300, which interact with HIF in the nucleus to allow the initiation of transcription of HIF-dependent genes, are conserved across species, and a CBP/p300 homolog encoded by the cbp-1 gene can be found in nematodes and fruit flies as well as in mammals (14). Functional homologs of p300/CBP have also been identified in plants such as Arabidopsis thaliana, giving strength to the likelihood for a conserved role for CBP/p300 in all metazoans (14).
Mapping Evolutionary Expression of Components of the HIF Pathway to Ancient Atmospheric Oxygen Levels
In (FIGURE 3), we have utilized a number of databases to correlate the evolution of components of the PHD/FIH-HIF-VHL pathway with metazoan evolution and the estimated atmospheric oxygen concentrations over the last 550 million years. The estimations for periods during which evolutionary divisions between phyla occurred were obtained from The Timetree of Life (5). The upper time limit represents the minimum time since the division of the phyla occurred and the lower limit, which has been termed the “soft maximum,” represents the maximum time since the divergence occurred (for details on how these figures are calculated, see Ref. 5). The atmospheric oxygen levels used were obtained from the updated GEOCARBSULF model (7, 12) and personal communication from Robert Berner (Yale University). It should be noted that an alternative model of ancient atmospheric oxygen levels termed COPSE has also been proposed, which predicts some common and some differential ancient atmospheric Po2 values (6). The presence of homologs of human genes of the PHD/FIH-HIF-VHL pathway were obtained from extensive review of the relevant literature and the use of the Homologene search engine (http://www.ncbi.nlm.nih.gov/homologene).
From these correlations, it appears that the earliest components of the pathway to evolve were HIF-1α, HIF-1β, PHD2, and VHL, homologs for all of which can be found in species as early as nematodes. This indicates that these genes evolved earlier than the radiation of the metazoans, were present in the earliest animal species, and possibly evolved in response to increasing oxygen concentrations in the pre-Cambrian period (from 4,600 to 542 million years ago), during which time prokaryotic life evolved and atmospheric oxygen rose from 0% to over 20% oxygen (12, 28). The next genes in this pathway to evolve were those for HIF-2α and PHD3, both of which are present in fish but not in arthropods, indicating that these genes evolved after the arthropod/fish division but before the fish/amphibian division between 460 and 421 million years ago. The most recent additions to the family are PHD1 and HIF-3α, which appear only in mammals and thus must have evolved <312 million years ago.
Using the above approach, we estimate the minimum times at which various components of the PHD/FIH-HIF-VHL pathway evolved, although the absence of accurate information regarding the actual time of their evolution makes correlation of these events with atmospheric oxygen levels speculative. Nevertheless, some intriguing possibilities can be postulated. First, the evolution of HIF-2α and PHD3 occurred after the division of the Fish and arthropods (460 million years ago) and before the division of the fish and the amphibians (421.5 million years ago). This period coincides with a time during which atmospheric oxygen increased from 17 to 23%. Similarly, HIF-3α and PHD1 evolved after the division of the mammals and the birds/reptiles, which occurred 312 million years ago. During this period, atmospheric oxygen levels were also on the increase from 23% 310 million years ago to their highest point in the course of history of 32% 280 million years ago.
As outlined above, FIH1 demonstrates selective conservation and is expressed in primitive organisms including coral and sea anemone (27) and also fish, amphibians, reptiles, birds, and mammals. However, although the gene is retained in all species since the fish, it is missing in fruit flies and nematodes but not in intermediate species such as beetles. A more in-depth analysis of the phylogenetic digression between beetles and flies reveals that this division occurred between 414 and 307 million years ago (5), which was a time of relative stability in terms of atmospheric oxygen concentrations, indicating that this selective conservatism was not driven by fluctuations in atmospheric oxygen but may have involved other selective forces.
Summary and Conclusions
Metazoans have evolved to use the chemical reduction of molecular oxygen as the primary source of metabolic energy, making a constant supply of oxygen key to organism survival. Because of this, at an early point metazoans began to evolve molecular mechanisms by which to deal with hypoxic insults that involved the increased expression of genes that support adaptation to hypoxia. A master and ancient regulator of this response is the HIF. Recent work has unveiled the molecular mechanisms linking cellular oxygen sensing to the activation of HIF. Here, we propose that the prevailing atmospheric oxygen concentrations may have played a role in shaping the evolution of this pathway. However, some key points need to be considered. First, oxygen is likely not the only gas to drive evolution, and carbon dioxide levels in the atmosphere were also likely important. Second, HIF is not the only hypoxia responsive transcription factor; in fact, over 16 other transcriptional regulators have been demonstrated to possess at least some degree of hypoxia sensitivity (58). Furthermore, the absence of complete analysis of the genomes of species during the relatively long periods between the absolute minimum (as determined by fossil evidence) and the soft maximum (as determined by the molecular clock determinations) for the division time of arthropods and fish and the division of the arthropods and nematodes means we cannot exclude the possibility of selective conservation of genes involved in the PHD/FIH-HIF-VHL pathway during these periods. Future studies should be directed toward taking these important points into account in forming a global view of the role of the Earth's atmosphere on shaping metazoan evolution.
Work from the author's laboratories is funded by Science Foundation of Ireland and the European Union.
No conflicts of interest, financial or otherwise, are declared by the author(s).
- ©2010 Int. Union Physiol. Sci./Am. Physiol. Soc.