Despite large changes in salt intake, the mammalian kidney is able to maintain the extracellular sodium concentration and osmolarity within very narrow margins, thereby controlling blood volume and blood pressure. In the aldosterone-sensitive distal nephron (ASDN), aldosterone tightly controls the activities of epithelial sodium channel (ENaC) and Na,K-ATPase, the two limiting factors in establishing transepithelial sodium transport. It has been proposed that the ENaC/degenerin gene family is restricted to Metazoans, whereas the α- and β-subunits of Na,K-ATPase have homologous genes in prokaryotes. This raises the question of the emergence of osmolarity control. By exploring recent genomic data of diverse organisms, we found that: 1) ENaC/degenerin exists in all of the Metazoans screened, including nonbilaterians and, by extension, was already present in ancestors of Metazoa; 2) ENaC/degenerin is also present in Naegleria gruberi, an eukaryotic microbe, consistent with either a vertical inheritance from the last common ancestor of Eukaryotes or a lateral transfer between Naegleria and Metazoan ancestors; and 3) The Na,K-ATPase β-subunit is restricted to Holozoa, the taxon that includes animals and their closest single-cell relatives. Since the β-subunit of Na,K-ATPase plays a key role in targeting the α-subunit to the plasma membrane and has an additional function in the formation of cell junctions, we propose that the emergence of Na,K-ATPase, together with ENaC/degenerin, is linked to the development of multicellularity in the Metazoan kingdom. The establishment of multicellularity and the associated extracellular compartment (“internal milieu”) precedes the emergence of other key elements of the aldosterone signaling pathway.
- aldosterone pathway
- Metazoan evolution
the kidney of vertebrates plays a major role in the homeostasis of the extracellular fluid. Despite large changes in water and salt intake, the kidney is able to maintain the extracellular osmolarity and volume within very narrow margins. Such a fine control requires specific factors or hormones (66). Aldosterone is the most potent sodium-retaining factor in mammals. During the evolution of vertebrates, aldosterone played a critical physiological role about 300 million years ago following the emergence of amphibia, the first vertebrates to adapt to a terrestrial environment. Comparative studies of physiology, biochemistry, and molecular biology have helped in delineating the most significant steps involved (66). The mechanisms involved in the control of sodium retention appear to be highly conserved, reflecting their biological importance. The most important and conserved function of aldosterone is to promote sodium reabsorption across tight epithelia that display a high transepithelial electrical resistance and an amiloride-sensitive, electrogenic sodium transport. In mammals, such epithelia are found in the distal colon and in the aldosterone-sensitive distal nephron (ASDN) (66). In ASDN, the sodium response (promoting potassium secretion) is localized in the principal cells (Fig. 1), the main target for aldosterone, mediating its effect through a signaling cascade that involves: 1) the mineralocorticoid and glucocorticoid receptors (MR and GR); 2) 11βHSD2, which protects MR from illicit occupation by cortisol or corticosterone; and 3) the transcription of Sgk1, a kinase that controls Nedd4-2, an E3 ubiquitin-protein ligase that, in turn, controls the activity of epithelial sodium channel (ENaC) at the apical membrane. Moreover, aldosterone increases the transcription of the ENaC α-subunit and of α1/β1 Na,K-ATPase subunits. ENaC and Na,K-ATPase are thus the main effectors of aldosterone-dependent sodium reabsorption, which controls renal sodium reabsorption, hence blood volume and blood pressure. Gene inactivation of the Na,K-ATPase α1-subunit is embryonic lethal (34), and gene inactivation of ENaC is lethal shortly after birth (51), demonstrating their limiting role in the aldosterone signaling pathway.
In his famous assay of comparative anatomy and physiology of the kidney during evolution “From Fish to Philosopher” published in 1953 (59), Homer Smith was the first to emphasize the critical role of fluid filtration by the glomerulus and fluid reabsorption by the renal tubule to control the homeostasis of our internal environment during evolution. In contrast to the situation in 1953, we have access to genomic data covering the most important branches of the tree of life. One application of these data is to study in detail the phylogenetic tree of any gene family of interest. In the present study we have focused on the evolution of two sodium transporters, which are the key effectors and limiting factors of the aldosterone response in the mammalian distal nephron.
By searching for homologs in various eukaryotes, from unicellular eukaryotes (“protists”) to multicellular Metazoans, we provide here a novel view of how mammalian aldosterone-dependent control of sodium homeostasis, blood volume, and blood pressure might have evolved.
MATERIALS AND METHODS
Identification of homologous genes.
We first created position-specific score matrices (PSSMs) for each protein. We used the following sequences as starting points: sp:Q16515 (ENaC/degenerin), sp:P05023 (Na,K-ATPase α-subunit), sp:P05026 (Na,K-ATPase β-subunit), and sp:P54710 (Na,K-ATPase γ-subunit). We did PSI-BLAST searches (release 2.2.24+) (2, 11) by applying two runs on the National Center for Biotechnology Information (NCBI) nonredundant protein database “nr” (release 06–10-2010). We used standard parameters (BLOSSUM62, -word_size 3 -gapopen 11 -gapextend 1), except for the short γ-subunit (BLOSUM80, -word_size 2 -gapopen 10 -gapextend 1).
Once the matrices were created, we scanned the Swissprot database with PSI-BLAST searches by doing three runs with the previously built PSSM. We enriched this dataset by scanning the following eukaryotic genomes: the vertebrates Xenopus tropicalis (33) and Takifugu rubripes (5), the chordate Ciona intestinalis (20), the insect Tribolium castaneum (65), the cnidarian sea anemone Nematostella vectensis (49), the placozoan Trichoplax adhaerens (60), the sponge Amphimedon queenslandica (61), the marine choanoflagellate Monosiga brevicollis (43), the soil-living amoeba Dictyostelium purpureum (DoE Joint Genome Institute, unpublished), the Paramecium tetraurelia (6), and finally three excavates: the flagellated protozoan parasites Giardia lamblia (45) and Trypanosoma brucei (8) and the free-living amiboeus bikont Naegleria gruberi (28). An alternative scan of the GenBank NR database provided a very large number of sequences, which was very difficult to visualize but did not provide any additional biological insight. Thus we present only the results of Swissprot + key genomes.
The four gene families were aligned using the GUIDANCE web server (47). We used the MAFFT algorithm, with max-iterate=1000 and refinement strategy=localpair, which corresponds to the L-INS-i option. GUIDANCE estimates a global score for the alignment, and we selected only columns that pass this threshold (GUIDANCE scores: NaKα = 0.79, NaKβ = 0.85, NaKγ = 0.71, ENaC = 0.56). Alignments were visualized with Jalview 2 (70). Using the reduced alignments, we estimated the best evolutionary models using ProtTest 2.4 (1). Phylogenetic trees were estimated with PhyML release 3.0.1 (32), with the appropriate evolutionary model (Na,K-α = LG + G, Na,K-β = LG + I + G, Na,K-γ = JTT + G, ENaC = LG + I + G) and the BEST search method. Branch support is indicated by the aLRT statistic (Chi2-based parametric, option -b2) (4). We verified the consistency of results using Minimum Evolution. We used Protdist (25) to construct the distance matrix and FastME release 2.07 (21) to build the trees (100 bootstraps). Trees were rooted based on the taxonomy (i.e., using outgroup species). The outgroup chosen for NaKα was all other ATPases. The outgroup chosen for NaKβ was the sponge Amphimedon, the most distant animal from Vertebrates. The outgroup chosen for NaKγ was a sequence from a shark, which is the sister group of bony vertebrates. Finally, we rooted the tree of ENaC using the Naegleria sequences, according to known metazoan phylogeny. Trees were drawn using FigTree.
The α-subunit of the Na,K-ATPase pump is found in all three domains of life, consistent with previous observations (18, 55). A rapid survey of the NR dataset shows that it is detected in many bacterial lineages (Tenericutes, Firmicutes, Proteobacteria, and Actinobacteria), in seven Archaea, and most eukaryotic lineages, with the notable exception of plants (data not shown). In animals (Fig. 2), the α-subunit is present in all species sampled, with six copies in Vertebrates: four paralogs of Na,K-ATPase α-subunit and the two paralogs of H,K-ATPase, the gastric and the nongastric α-subunit. The two rounds of whole genome duplication (2R WGD) at the origin of vertebrates (48) are partially responsible for these duplicated genes. A recent study (55) suggested that the frog X. tropicalis (33) has no nongastric H,K-ATPase, but two gastric H,K-ATPases. From our analysis, one gene (Ensembl id: ENSXETP00000014923) belongs to the gastric class, whereas the other (ENSXETP00000002677) belongs to the nongastric class.
Outside vertebrates, the insect genes mostly form a monophyletic group, plus the inclusion of a misplaced sponge sequence. At the basis of the group, we found as expected homologs in Trichoplax and Monosiga. We also detected homologs in D. purpureum and in Naegleria one of the most distant eukaryotic species from Metazoans.
Whereas the α-subunit is found in all the three domains of life, the β-subunit appears restricted to Holozoa, a branch of the opisthokonts that includes all animals and their closest single-celled relatives (such as M. brevicollis). We used the sponge sequence to root the phylogenetic tree (Fig. 3). In Vertebrates, the β-subunit has a similar evolutionary history to the α-subunit. Two rounds of whole genome duplications plus probably a local duplication generated the five groups of paralogs: four paralogs for the Na,K-ATPase β-subunits and a fifth paralog for the H,K-ATPase β-subunit.
Inside the Insecta lineage, which is monophyletic, two independent duplications occurred at the basis, which lead to the three subfamilies Nervana-1, -2, and -3 (7). Another duplication occurred specifically in Drosophila, which led to two others subgroups.
Outside Bilaterians, we found homologs of the β-subunit in the cnidarian N. vectensis, in the placozoan Trichoplax and in the poriferan sea sponge A. queenslandica. Outside Metazoa, only one β-subunit homolog (NCBI number: XP_001748170) has been detected in the genome of M. brevicollis, a choanoflagellate (the sister group of Metazoa). This sequence is placed outside Bilaterians in the minimum evolution tree but misplaced in the arthropod group in the maximum likelihood tree. Thus we can postulate that the β-subunit appeared in the Holozoan ancestor, and was already present at the origin of Metazoans.
Our result is in contradiction with the hypothesis that the β-subunit might share a common origin with the bacterial potassium-transporting ATPase C chain (EC 220.127.116.11) (KdpC) (24, 58). Forcing a multiple alignment of the Na,K-ATPase β-subunit and KdpC, no homology is apparent (Supplemental Fig. S5).1 To have an objective criterion to assign homology, we used the FFAS03 server (35), which is dedicated to the detection of very distant homology; a score −9.5 is considered significant. We compared the same sequences as used in the original paper (58): the β-subunit of the sheep (sp:P05028) and KdpC of Escherichia coli (sp:P03961). Their comparison gives a fold and function assignment system (FFAS) score of −0.84 and only 9% identity, which is not sufficient to assign homology. In contrast, the comparison of the sheep sequence with the placozoan T. adhaerens (XP_002116786) gives a score of −96.7, and 23% of identity; the comparison with the choanoflagellate M. brevicollis (XP_001748170) gives a weaker but still significant score of −16.2 and 16% of identify. Finally, FFAS03 was unable to detect any similarity between the Trichoplax or Monosiga β-subunit and E. coli KdpC. We can conclude that the origins of KdpC and of the β-subunit are probably distinct. Any similarities are probably due to the common constraints of membrane proteins (71).
The family of FXYD proteins, to which the Na,K-ATPase γ-subunit belongs, appears to be quite recent. Apparently, one FXYD protein was already present in the ancestor of Gnatostomes, as we found a PLMS-like exemplar in the shark Squalus acanthias (sp:P82542), which was used to root the tree (Fig. 4). We did not detect any homolog in the sea lamprey Petromyzon marinus (BlastP/tBlastN from the preEnsembl server). But as the sea lamprey genome is not yet completely sequenced, it is difficult to conclude whether FXYD emerged before or after the evolution of Gnatostomes.
The inference of the precise phylogeny is difficult, because the proteins are relatively short (<50 positions analyzed) and provide limited phylogenetic signal. Nevertheless, some clusters emerge.
FXYD1 (phospholemman, PLMS) is present in all Tetrapods but does not form a clear clade in the ML tree. The FXYD2 family (the Na,K-ATPase γ-subunit) appeared in the Tetrapoda ancestor, as we found it in the batracian X. tropicalis, and in all mammals. The FXYD3 (chloride conductance inducer protein Mat-8) and FXYD4 (channel-inducing factor, CHIF) genes are both present in nearly all mammals. They duplicated after the split between Amphibia and Amniota. FXYD5 (Dysadherin) seems to emerge in the Eutheria (placentals) ancestor or was lost in other lineages. The FXYD6 (Phosphohippolin) subfamily is detected in all bony vertebrates. Interestingly, a duplication happened at the base of the Simiiformes primates, including human, which gave birth to the FXYD8 subfamily. In human, FXYD6 is located on chromosome 11, whereas FXYD8 is on the X chromosome. We know nothing about the function of the FXYD8 subfamily. It has been postulated to be a pseudogene (according to the Ensembl annotation). But the multiple sequence alignment and the phylogenetic tree indicate that after an acceleration in evolutionary rate immediately after the duplication, FXYD8 evolved under purifying selection in Simiiformes. This conservation between orthologs challenges the hypothesis of a pseudogene, inviting further functional studies. Finally, another group of FXYD genes is present in the fishes, just after the split between cartilaginous fishes and bony vertebrates, but it is not clearly related to any other cluster. It was probably lost in the Tetrapods' ancestor.
The ENaC/degenerin superfamily contains many subfamilies, which, until now, were only found in Metazoa (42). In vertebrates, this superfamily includes the amiloride-sensitive ENaC, involved in the control of extracellular ionic composition, and the amiloride-sensitive acid-sensing ion channel (ASIC), involved in pain sensation (42). Other ENaC/degenerin members are present in nonvertebrate organisms and have various functions, including osmoregulation, mechanosensation, and water tasting in Drosophila (12, 14); synaptic transmission in mollusks (29); or mecanotransduction in nematodes (63).
The secondary structure of the ENaC/degenerin proteins predicts a short cytoplasmic NH2 terminus, followed by the α-helix of the first transmembrane domain (TM1), a large extracellular domain comprising many β-sheets, two to three cysteine-rich domains (CRDs), followed by the α-helix of the second transmembrane domain (TM2), and finally a short cytoplasmic COOH-terminal domain (36, 42).
ENaC/degenerin was already present in the ancestor of Metazoa, as we found orthologs in Cnidaria, in T. adherens, and in the sponge A. queenslandica. No homolog is found in Monosiga, P. tetraurelia, or D. purpureum (Fig. 5). The tree shows an ancient duplication that occurred in the early Metazoa. More recently, ASIC-1 to -4 were generated by the two rounds of whole genome duplications (2R) at the origin of vertebrates (48), as were the four ENaC genes (α, β, γ, and δ). Only the ASIC-5 gene (also called hINaC) seems to be much older than the origin of vertebrates, as it clusters with a Ciona gene and several Trichoplax and Sponge genes. A specific expansion occurred in the Nematode lineage.
One intriguing result is that we found homologs in Naegleria, an Excavata. All Naegleria sequences cluster together and outside the Metazoan sequences, with strong statistical support. Different scenarios for the divergence of early eukaryotes have been proposed (28, 38). In all of these scenarios, the common ancestor between the Metazoa (supergroup Unikonts) and Excavata (supergroup Bikonts) is the common ancestor of all eukaryotes (50). The specific expansion of the Naegleria lineage is described below.
Otherwise, the overall phylogeny is consistent with the known tree of Metazoans, with specific expansions in nematodes and vertebrates.
Outside Metazoa: the case of N. gruberi.
The detection of homologs in the N. gruberi genome was surprising, and we decided to investigate it further. While the Naegleria sequences are very divergent from the Metazoan sequences, there are three well-defined segments that are conserved between the two taxa (alignment and summary in Supplemental Figs. S6 and S7, respectively): 1) the first conserved segment (position 382–445 on the alignment) includes the α-helix of transmembrane segment TM1 (380–412), followed by the two first β-sheets β1 (417–23) and β2 (436–440). This β-sheet β2 contains the ultraconserved motif FPxxTxC and lies in the core of the extracellular domain; 2) the second conserved segment, from the ultraconserved cysteine C892 to position 1063, encompasses seven β-sheets (β3–β9), with histidine H1028 (H354 in α-hENaC) well conserved in Naegleria; and 3) the COOH-terminal part of the extracellular domain (position 1334 to 1392), which includes the α-helix of transmembrane segment TM2 preceded by the two conserved β-sheets (β11 and β12).
In the NH2-terminal cytoplasmic domain, there are several residues important for the ion selectivity in ASICs, the gating in ENaC, and the function of MEC-4 (42), which are not conserved in the Naegleria sequences.
The animal extracellular domain is rich in cysteine residues and contains seven disulfide bridges, which are widely conserved among homologs [ASIC, ENaC, FaNaC, and degenerin (36)]. One disulfide bridge, C133–C305 in α-hENaC (position 440 and 937 in the alignment), is important for proper protein folding, as a thermosensitive mutation of the first cysteine is linked to the pseudohypoaldosteronism type 1 syndrome (PHA-1) (26). This cysteine belongs to the FPxxTxC motif (127–133 in α-hENaC). In Naegleria, a homologous cysteine is found at position 440 in 11 of the 12 sequences, and all of these 12 possess a cysteine at position 937. This conserved disulfide bridge is the first to form after TM1 bridging of the β2- and β5-domains, suggesting a critical role in protein folding during the translation of the peptide chain. Two other cysteines are conserved among Naegleria sequences only. A third is well conserved in four sequences. They might form a Naegleria-specific CRD, as found in the degenerin of Caenorhabditis elegans (containing three CRDs) and of vertebrates (two CRDs) (63). However, it is difficult to infer if they form disulfide bridges, as we have no direct biochemical or structural information for the C. elegans or Naegleria CRDs.
In addition to the cysteine bridges, in the ENaC family, the loop between β2 and β5 is known to be cleaved by a serine protease (CAP-1 and CAP-2) that is required for channel trafficking and activation (52). In α-rENaC, two sites, 205 and 233 (between C158 and C332), and a third one (between C458 and C472) have been predicted to be cleaved by furin or serine protease (like CAP-1). However, these three sites correspond to large truncated segments in Naegleria (deletion from 451 to 890 and 1060 to 1220 in the alignment).
At the beginning of the TM2 segment, the degenerin site (position 549 in α-hENaC, 1363 in the alignment) has been shown to be important when mutated in the Caenorhabditis MEC-4, causing a permanent opening of the channel (23). Mutation in the human ASIC2a G430 (ACCN1_HUMAN, sp:Q16515) to a large amino acid leads to a constitutively activated channel (69). In several ENaC proteins, including human, a serine occupies this position. In all vertebrate ASICs, and in all Naegleria sequences without exception, there is a glycine. This suggests that the function of this position is conserved in Naegleria. Another motif, the G/SxS selectivity filter (position 1374–1376 in the alignment) is strictly identical in all ASICs (GAS motif) but differs widely among ENaC members. We detect this motif in the Naegleria sequences, but the second residue is a large aliphatic acid (Leu or Ile), and the last serine is replaced by an aspartic acid in most sequences. The motif in Naegleria is thus G-I/L-D. Interestingly, it has been shown that the larger the residues are in this motif, the larger the cations which could pass through the channel (40). In particular, the α-ENaC mutant S589D presents facilitated access for K+ and NH4+, which is not possible in the wild type (41). A large residue such as aspartatic acid will open the pore by moving subunits.
In the COOH-terminal cytoplasmic domain of ENaC, there is an important proline-rich sequence (PPXY) that is involved in ubiquitylation and internalization of the protein; when mutated it leads to a gain of function mutation of ENaC and a severe salt-sensitive hypertensive phenotype in humans (Liddle's syndrome) (56). This domain is not conserved in Naegleria either.
Transition from uni- to multicellularity: creation of the internal milieu.
Unicellular prokaryotes or eukaryotes expose 100% of their surface to the environment, allowing rapid and efficient exchange of substrates and respiration. To face important changes in substrate availability and in osmolarity, unicellulars have developed sophisticated mechanisms to keep a high intracellular potassium and a low intracellular sodium required for replication, transcription, and translation. The passage from single cells to multicellular organisms was a major step in history of life (64). This transition appeared in several lineages of eukaryotes, including plants and animals (53). Multicellularity brought new problems for the cells, for instance substrate or oxygen availability, but offered the distinct advantage of controlling the ionic and osmolarity of the internal milieu. Not surprisingly, genes involved in membrane transport and osmolarity regulation have been reported to have a link to multicellularity (13, 15). In Metazoans, the formation of epithelia with its junctional apparatus and the differentiation of the plasma membrane into an apical and basolateral domain was the key to creating asymmetric distribution of ions and substrates between the environment and the internal milieu (extracellular fluid). During early development, just after fertilization, the egg undergoes several rapid cleavages (2 cells, 4 cells, 8 cells, and so on until the blastula stage). This is an example of a developmental transition from uni- to multicellularity. The first differentiation is the appearance of tight junction as early as the two-cell stage, and the formation of a primitive epithelium at blastula that creates the blastocoel cavity, the first extracellular fluid compartment with a high sodium and low potassium concentration mirroring the intracellular high potassium and low sodium concentration (31).
In this context we would like to discuss the following points: 1) the presence of the ENAC/degenerin family in an unicellular organism (N. gruberi), 2) the emergence of the Na,K-ATPase and ENaC is simultaneous with the emergence of multicellularity of Metazoans, 3) the role of the duplications in vertebrates for the diversification of pumps and channels, and the evolution of the aldosterone signaling pathway.
Presence of the ENAC/degenerin family in the N. gruberi.
The finding of ENaC/degenerin homologs in Naegleria, a eukaryotic microbe from the Excavates supergroup (19), is consistent with two different scenarios. The first one is a vertical inheritance from the last common ancestor of eukaryotes (Fig. 6A). The ENaC/degenerin would have originated in the common ancestor of eukaryotes and then been successively lost in all domains except Metazoans and Excavates. These apparent gene losses could also be due to extreme sequence divergence, making homolog detection impossible. Of note, a similar evolutionary history has recently been reported for molecules involved in integrin-mediated adhesion (57). This vertical inheritance scenario is favored by two observations: 1) horizontal gene transfers are rare in eukaryotes (39), and 2) the phylogeny is consistent with the known tree of eukaryotes (i.e., the Naegleria sequences do not group in the middle of an animal clade). But because horizontal gene transfers are rare does not mean that we should exclude this scenario (Fig. 6B), especially as Naegleria is one of the most represented eukaryote microbes in the number of lateral gene transfers (3). A lateral gene transfer could have happened between the Naegleria and the Metazoa ancestors. In that case, we could not predict the direction, whether the ENaC/degenerin originated in the ancestor of Metazoa and was transferred to the Naegleria ancestor, or if conversely, it originated in the Naegleria and was transferred to Metazoa. Given the absence of ENaC/degenerin in all other lineages of eukaryotes, the lateral gene transfer scenario should be favored for the time being.
More generally, the sequencing of the Naegleria genome has revealed interesting features of eukaryotic evolution (28). Around 40% of eukaryotic gene families have no ortholog in Archae and Bacteria and thus are innovations in the Eukaryotic ancestor. The transition from prokaryote to eukaryote was an important step, notably concerning mobility and sensing of the environment. Some Naegleria species, such as N. gruberi, possess the surprising ability in fresh water to shift rapidly from an amoebaean to a flagellate form, involving generation of de novo organelles (22, 27). It has been demonstrated that this transformation can be prevented by increasing the electrolyte or osmolyte concentration in the water (for instance sodium or potassium chloride) (37). It is tempting to speculate that the mechanism responsible for the sensitivity to ionic changes might be related to the expansion of the ENaC/degenerin gene. As ENaC/degenerin are ion channels highly selective for sodium, they could be candidate sodium or osmolar sensors in N. gruberi.
Emergence of the Na,K-ATPase and ENaC is simultaneous with multicellularity of Metazoans.
A classical question in molecular evolution is to ask when a gene appeared for the first time. When a gene is present in two different organisms (orthologous genes), the most parsimonious hypothesis is that it was already present in their common ancestor.
With our results, we have now a more precise idea of how the mammalian aldosterone pathway was established (Fig. 7). The first gene to appear was the P-ATPase IIC α-subunit. This α-subunit, like many other P-ATPases, is present in all the three domains of life: bacteria, archaea, and eukaryotes. It probably originated at the same time of the origin of the cellular life, within the last universal common ancestor. However, the full function of the Na,K-ATPase requires also the β-subunit; without it the P-ATPase only functions as a unidirectional pump in most organisms. The finding of a distant homolog in the choanoflagellate M. brevicollis suggests that the origin of the β-subunit could be older than previously thought (13, 55), in the Holozoan ancestor (Metazoa and their closest single-celled relatives). Several genes suspected to be important for multicellularity have been found in single-celled relatives (54). The fact that the β-subunit appeared slightly before the emergence of Metazoans leads us to postulate that the full functional IIC-ATPase could have been a factor contributing to multicellularity. It would be interesting to test experimentally whether the Na,K-ATPase α/β in Monosiga is a bona fide sodium/potassium pump or not. Interestingly, the chaperone role of the β-subunit is essential for the formation of tight junctions and epithelial polarity (30).
Slightly after the β-subunit, the ENaC/degenerin superfamily appeared in the Metazoan ancestor. The finding of homologs in the Naegleria genome could suggest an older emergence, but it seems more reasonable, as we did not find this gene in other eukaryotes, to postulate a lateral gene transfer between the Naegleria ancestor and the Metazoan ancestor. In any case, the ENaC/degenerin family was already present in the Metazoan ancestor, along with the full Na,K-ATPase α/β. The fact that these two components are already present in the common ancestor of Metazoans suggests functions similar to those in present day mammals. In particular, the ENaC/degenerin might have functions such as osmo-regulation or water sensing/taste and thus be of great help in a multicellular organism. The emergence of the ion transporters ENaC/degenerin superfamily and the complete form of the active pump Na,K-ATPase (α+β) added another level for strict control of the environment. Like other genes involved in membranes and osmolarity regulation (15), the hypothesis that these genes help the transition to multicellularity is appealing. The advantage of a stricter control allowed cells to become more specialized. For example, cells in the mammalian kidney adopt a polarized form, where ENaC acts to let the sodium enters into the cell from the lumen (apical side), whereas the Na,K-ATPase acts to push out the sodium into the blood and let the potassium enter the cell (basolateral side).
Role of duplication in vertebrates for the diversification of pumps and channels and the aldosterone pathway.
While shifts in function can affect both orthologs and paralogs (62), duplication allows genes to explore different functional space inside the same organism (16). The 2R WGD (48) allowed a diversity of paralogs among the IIC-ATPases (46) and the ENaC/degenerin. While the 2R WGD can result in up to four copies from one single gene, there were additional specific local duplications in vertebrates leading to produce the six α-subunits (4 Na+/K+/2 H+/K+) and the five β-subunits (4 Na+/K+/1 H+/K+). At this time, we cannot define which of the paralogs are from 2R WGD and which are from single duplications. After duplication, these genes were then able to obtain new biochemical functions (e.g., the H,K-ATPases) and/or to evolve new expression patterns [i.e., gastric vs. nongastric for the H,K-ATPase, or the expression in the brain for hENaCδ (68)]. The six α-subunits and the five β-subunits are also able to assemble in various patterns (46), which can lead to subtle changes in function. The expansion of paralogs was important for the adaptation to new tissues in vertebrates. This hypothesis fits also well for the FXYD family, as it appeared in the Gnatostomes ancestor and was subject to different duplications (the 2 WGD and lineage-specific duplications) to produce at least seven paralogs. The many FXYD genes in Vertebrates allow a fine control of the transport properties of the type IIC P-ATPases by mediating tissue specific modulation.
The 2R WGD not only affected the number of pumps and channels in vertebrates but also affected the members of the regulatory pathway and of the aldosterone synthesis. The CYP11A and CYP11B paralogs emerged at the origin of vertebrates. CYP11B1 and CYP11B2 have a complex history. These genes appear in different species by many independent duplications, as seen in primates and mouse (44) or in the guinea pig (10). Apparently, the ancestral CYP11B already possessed the ability to synthesize aldosterone (10).
The evolution of the corticoid receptors has been widely explored (9). Again, these transcription factors have diversified in the vertebrates during the 2R WGD (48). The first round of WGD produced the corticoid receptor and the sex steroids receptor. Then, the second round produced the two paralogous corticoid receptors, the GR (NR3C1) and the MR (NR3C2), and the two paralogous sex steroids receptors, the progesterone receptor (PR, NR3C3) and the androgen receptor (AR, NR3C4).
Finally, two other important enzymes in the aldosterone pathway, Nedd4-2 and Sgk1 (Fig. 1) have been produced by the 2R WGD. The Nedd4-2 ubiquitin-protein ligase is a paralog of Nedd4, as indicated by the C. intestinalis and Ciona savignii outgroups [gene ENSG00000049759 on EnsemblCompara gene trees (67)]. Nedd4-2 possesses various residues that might be phosphorylated, especially the Thr367 and Ser448 (according to the human sequence NEDD4-like, sp:Q96PU5). These two residues are present in two insertions in Nedd4-2, which are absent in Nedd4. In the same manner, the Sgk family contains three members (Sgk-1, -2, and -3) [gene ENSG00000118515 on EnsemblCompara gene trees (67)]. The first round of WGD generated Sgk1/2 and Sgk3. The second round generated the paralogs Sgk1 and Sgk2, while the duplicate of Sgk3 was lost.
Our data together with other recent data (Fig. 7) support the notion that the effectors (channels and pumps) appeared first, then the receptors (MR and then GR), followed by the MR protective enzyme 11βHSD2, conferring a high degree of mineralocorticoid specificity, and finally the ligand aldosterone.
This work was supported by a Leducq Foundation grant.
No conflicts of interest, financial or otherwise, are declared by the author(s).
We acknowledge our dear friend and colleague Jean-Daniel Horisberger for his major contribution in the initiation of this study. We thank Laurent Schild, Kaethi Geering, and Christine Orengo for careful reading of the manuscript and providing helpful comments and suggestions. We also thank three anonymous reviewers for insightful remarks. Genomics sequence data were produced by the US Department of Energy Joint Genome Institute http://www.jgi.doe.gov/ in collaboration with the user community.
Current addresses: R. A. Studer, Dept. of Structural and Molecular Biology, Darwin Bldg., Univ. College London, Gower St., WC1E 6BT London, UK; E. Person, Eawag, Swiss Federal Inst. of Aquatic Science and Technology, Dept. of Fish Ecology and Evolution, CH-6047 Kastanienbaum, Switzerland.
↵1 The online version of this article contains supplemental material.
- Copyright © 2011 the American Physiological Society