Figures
Abstract
A major physiological barrier for aquatic organisms adapting to terrestrial life is dessication in the aerial environment. This barrier was nevertheless overcome by the Devonian ancestors of extant Tetrapoda, but the origin of specific molecular mechanisms that solved this water problem remains largely unknown. Here we show that an ancient aquaporin gene cluster evolved specifically in the sarcopterygian lineage, and subsequently diverged into paralogous forms of AQP2, -5, or -6 to mediate water conservation in extant Tetrapoda. To determine the origin of these apomorphic genomic traits, we combined aquaporin sequencing from jawless and jawed vertebrates with broad taxon assembly of >2,000 transcripts amongst 131 deuterostome genomes and developed a model based upon Bayesian inference that traces their convergent roots to stem subfamilies in basal Metazoa and Prokaryota. This approach uncovered an unexpected diversity of aquaporins in every lineage investigated, and revealed that the vertebrate superfamily consists of 17 classes of aquaporins (Aqp0 - Aqp16). The oldest orthologs associated with water conservation in modern Tetrapoda are traced to a cluster of three aqp2-like genes in Actinistia that likely arose >500 Ma through duplication of an aqp0-like gene present in a jawless ancestor. In sea lamprey, we show that aqp0 first arose in a protocluster comprised of a novel aqp14 paralog and a fused aqp01 gene. To corroborate these findings, we conducted phylogenetic analyses of five syntenic nuclear receptor subfamilies, which, together with observations of extensive genome rearrangements, support the coincident loss of ancestral aqp2-like orthologs in Actinopterygii. We thus conclude that the divergence of sarcopterygian-specific aquaporin gene clusters was permissive for the evolution of water conservation mechanisms that facilitated tetrapod terrestrial adaptation.
Citation: Finn RN, Chauvigné F, Hlidberg JB, Cutler CP, Cerdà J (2014) The Lineage-Specific Evolution of Aquaporin Gene Clusters Facilitated Tetrapod Terrestrial Adaptation. PLoS ONE 9(11): e113686. https://doi.org/10.1371/journal.pone.0113686
Editor: Marc Robinson-Rechavi, University of Lausanne, Switzerland
Received: August 15, 2014; Accepted: October 27, 2014; Published: November 26, 2014
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by the Research Council of Norway (RCN) project #178837/40 and 224816/E40 to (RNF), the Spanish Ministry of Science and Innovation (MICINN) AGL2010-15597 to (JC) and by the National Science Foundation, USA (NSF) IOS 0844818, 1041885 and 1241312 to (CPC). FC was supported by a postdoctoral (Juan de la Cierva Programme) fellowship from MICINN and the RCN (224816/E40). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Jón Baldur Hlidberg is associated with www.fauna.is. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors.
Introduction
Adaptive transitions of organisms between aquatic and terrestrial environments are surprisingly rare during evolution. Available data on biodiversity in water and land reveal that only a restricted subset of the major clades of plants, animals, fungi and microbes have radiated on land [1] The emergence of Tetrapoda from their ancestral aquatic habitats during the Devonian was thus a key transitionary Period that ultimately led to the successful radiation of vertebrates in terrestrial environments [2]. To achieve this success, the ancient ancestors of modern Tetrapoda had to overcome major physiological challenges that included locomotion, breathing and dessication in an aerial environment.
The former problem of locomotion was solved through the evolution of limbs from fins, and considerable research has thus focused on the genes and signalling pathways that alternatively specify the chiridian limbs of tetrapods or the rayed fins of actinopterygian fishes [3]–[8]. These investigations revealed, however, that the same transcription factors, including members of the homeobox (hox) clusters, T-box (tbox) clusters and retinoic acid receptors (rar), are involved in the embryonic formation of the limbs and fins of each lineage. A solution to the second problem of aerial respiration was realised through the early evolution of an air-filled organ that differentiates from the anterior gut as a ventral lung in Sarcopterygii and the most basal actinopterygian fishes (Polypteriformes), or as a dorsal swimbladder in all other Actinopterygii [9]–[11]. The evolution of air-breathing was thus not restricted to the prototetrapod lineage, but has existed since the Paleozoic origin of the osteichthyan fishes (Euteleostomi), with 46 actinopterygian families also having extant species that display this trait [12]. As for the specification of sarcopterygian limbs and actinopterygian fins, recent evidence has shown that conserved transcription factors, including homeobox (nkx2.1), forkhead box (foxa2), wingless (wnt7b), and GATA binding proteins (gata6), direct the development of both sarcopterygian lungs and actinopterygian swimbladders [11]. Consequently, although chiridian limbs are specific to Tetrapoda, and lungs are specific to all Sarcopterygii, the genes involved in their formation are not, they are ubiquitous among all jawed vertebrates (Gnathostomata).
A solution to the adaptive problem of water conservation was first postulated >150 years ago when Claude Bernard introduced the concept of the “milieu interieur” [13]. Chief among the physiological mechanisms that control water homeostasis in Mammalia was the evolution of the kidney and the counter-current mechanism of water conservation (Smith 1953). Extant Mammalia have evolved a tertiary kidney (metanephros) and an impressive counter-current mechanism of water conservation (osmolar urine/plasma ratio: U/P≤25) that in certain groups, such as members of the Rodentia, obviates the need for drinking [14]. The process is mediated through neurohypophysial secretion of vasopressin and type-2 receptor (AVPR2)-mediated trafficking of aquaporin-2 (AQP2) water channels to the apical membrane of the principal cells of renal collecting ducts (antidiuresis) (Deen et al. 1994; Boone and Deen 2008). Although birds also evolved a metanephros and a functional AQP2 mechanism, they show a lesser ability to concentrate urine (U/P≤2.5) [15]. The reduced U/P ratios are due primarily to a paucity of long medullary loops of Henle and the lack of a urea habitus [16]. Reptiles also lack a urea habitus and did not evolve long medullary loops of Henle and thus do not have urine concentrating abilities. As members of the Sauropsida, the reptiles and birds evolved a uric acid habitus [17], [18], a defining metabolic feature of this clade. The almost insoluble nature of this end-product of protein metabolism substantially reduced the need for water excretion and likely alleviated the selection pressure on the evolution of their kidneys. Possibly as an additional adaptation to compensate for the poor counter-current urine concentration ability of the metanephros, sauropsids also evolved a nasal or lingual gland, a homologous organ to the rectal gland of Chondrichthyes (chimaeras, sharks and rays), that secretes concentrated saline (NaCl) [19]–[21]. In the avian lineage, it has been shown that a different water channel (AQP5) is expressed in this organ, and that under osmotic stress avian AQP5 is transcriptionally downregulated in the nasal gland in order to promote hyperosmotic salt secretion and water conservation [22].
In contrast to Mammalia and Sauropsida, extant Amphibia have a secondary kidney (mesonephros), which like the mesonephri of Teleostei, is unable to concentrate their urine above that of their blood plasma osmolarity (U/P≤1). Studies of aquaporins in Amphibia have, however, revealed that AQP2-like (AQP-h2k, HC-2) and AQP5-like (AQP-x5) channels are involved in water conservation [23], [24]. AQP-h2k is localised in the kidney tubules in response to arginine vasotocin (AVT) [25], [26], the vasopressin ortholog that co-evolved in the vertebrate lineage [27], and AQP-x5 mediates fluid secretions in the mucous and small granular glands [28]. The function of AQP-x5 is suggested to aid in the maintenance of moist skin, cutaneous gas exchange and thermoregulation [24], and is thus partly reminiscent of mammalian exocrine sweat glands, which also express AQP5 [29]. In Amphibia, such as hylid treefrogs and bufonid desert toads, additional anuran-specific water channels (AQP-h3 and AQP-h2; AQPa2-type) have been identified that respectively respond to AVT in the ventral skin and urinary bladder to mediate water uptake and water conservation [24], [30]–[35].
From the above, it is clear that AQP2, -5, and the anuran AQPa2-type water channels are intricately involved in the mechanistic basis of water conservation in the three extant clades of terrestrial vertebrates, the Amphibia, Sauropsida and Mammalia. Amongst fish, a recent functional analysis of zebrafish aquaporins revealed that the tertiary structures including the six transmembrane domains, two NPA motifs and aromatic/arginine (ar/R) selectivity filters as well as the permeation properties are similar to the mammalian counterparts [36]. Consequently, both piscine and mammalian aquaporins can be phylogenetically and functionally classified as water-selective classical aquaporins (AQP0, -1, -2, -4, -5 and -6), an ammoniaporin (AQP8), which transports water, ammonia and urea, unorthodox aquaporins (AQP11, and -12) for which cell permation properties have yet to be identified, and classical aquaglyceroporins (Glp: AQP3, -7, -9 and -10), which facilitate the transport of water, arsenic, urea and polyols such as glycerol [36]–[41]. However, although the genomes of the five species of Teleostei studied were found to retain a larger repertoire of aquaporins due to an independent genomic duplication event at the root of the crown clade, they appeared to lack orthologs of AQP2, -5, or -6 [36], [42], [43]. By contrast, a study of lungfishes (Dipnoi) revealed that this ancient lineage of sarcopterygian fishes possess an AVT-AVPR2-AQP2-like system similar to Mammalia [44]. Under drought conditions, the lungfishes estivate and activate this system to regulate an Aqp0-related channel (Aqp0p) to promote antidiuresis [44]. Based upon the preliminary evidence that AQP2, -5, or -a2-related orthologs are absent in actinopterygian fishes [36], [42], [43], we hypothesised that the evolution of these water channels may represent a genomic apomorphy that was positively selected in basal Sarcopterygii as a prelude to terrestrial adaptation. To investigate this hypothesis, we combined selected aquaporin sequencing from agnathan (jawless) and gnathostome (jawed) vertebrates with the assembly of >5200 peptide fragments and their corresponding exon-like sequences and used Bayesian inference to reconstruct the aquaporin superfamilies encoded in 131 deuterostome genomes (see Figure S1 for species interrelationships). The systematic approach of assigning contiguous exon-like fragments to specific subclasses uncovered an unexpected diversity of aquaporins in basal Deuterostomia, and several novel subfamilies of water channel in Vertebrata (Aqp14, -15 and -16). To explain this diversity and the origin of the deuterostome aquaporins, we used Bayesian inference to develope a convergent model that traces the deep evolutionary origins to highly diversified stem subfamilies in Radiata, Porifera, Bacteria and Archaea.
Materials and Methods
Biological samples
Specimens of Smaller spotted catshark were obtained from L'Aquàrium de Barcelona, (Barcelona, Spain) and immediately sacrificed following the procedures related to the care and use of the fish approved by the Ethics Committee of the Institut de Recerca i Tecnologia Agroalimentàries (IRTA, Spain) in accordance with the “Guiding Principles for the Care and Use of Laboratory Animals”. Frozen samples of eyes from US and Iberian Peninsula sea lampreys were obtained from Hammond Bay Biological Station (Michigan, USA), Center for Molecular and Comparative Endocrinology, University of New Hampshire (New Hampshire, USA), and Centre of Environmental Biology, University of Lisbon (Lisbon, Portugal). Specimens of spiny dogfish and Atlantic hagfish were provided by the Mount Desert Island Biological Laboratory (MDIBL; Salisbury Cove, ME, USA) and were treated in accordance with IACUC regulations.
Isolation of aquaporin transcripts
To obtain sequences related to the ancestors of AQP2, -5 or -6, aquaporin encoding cDNAs were isolated by RT-PCR following RNA extraction from visual, metabolic and osmoregulatory tissues, including eye, liver, rectal gland and kidney, using the RNeasy Minikit (Qiagen GmbH, Hilden, Germany), and treated with DNAse using the RNase-Free DNase kit (Qiagen) following the manufacturer's instructions. Total RNA (5 µg) was reverse transcribed using 0.5 µg oligo(dT)17, 1 mM dNTPs, 40 IU RNAse inhibitor (Roche Applied Science, Mannheim, Germany), and 10 IU SuperScript II Reverse Transcriptase enzyme (Life Technologies Corp. Carlsbad, CA), for 1.5 h at 42°C. The PCR was carried out with 0.5-2 µl of the RT reaction in a final volume of 50 µl containing 1 x PCR buffer plus Mg2+, 0.2 mM dNTPs, 1 IU of Taq polymerase (Roche), and 1 µM of primers. Degenerate or specific primers based on available genomic sequences were used to amplify partial cDNAs (Table S1). The PCR products were cloned into the pGEM-T Easy Vector (Promega) and sequenced by BigDye Terminator v3.1 cycle sequencing on ABI PRISM 377 DNA analyzer (Applied Biosystems). The 5′end bearing the C-terminus of sea lamprey aqp01 was cloned by 5′-RACE (Life Technologies Corp.) (Table S1). The spiny dogfish aqp15 and Atlantic hagfish glp cDNAs were also amplified by RT-PCR using total RNA extracted (as previosuly described; [45]) from kidney and esophagus respectively, and using inosine containing degenerate primers (Table S1). Total RNA (4.5 µl) was reverse transcribed using 100 pmole oligo(dT)26, 1 mM dNTPs, 10 IU SUPERase.InTM thermostable RNAse inhibitor, and 100 IU SuperScript III Reverse Transcriptase enzyme (Life Technologies Corp. Carlsbad, CA), for >1 h at 50°C. The degenerate PCR was carried out with 0.5–1 µl of the RT reaction in a final volume of 20 µl containing 1 x standard PCR buffer, 0.2 mM dNTPs, 1.25 IU of Taq polymerase (New England Biolabs, Ipswich, MA), and 5 µM primers. Degenerate or specific primers based on available sequences were used to amplify partial cDNAs. 5′ or 3′ RACE cDNA sequences were amplified using a Marathon cDNA synthesis kit (Clontech, Mountain View, CA) and using Phusion DNA Polymerase (New England Biolabs, Ipswich, MA). cDNA fragments were cloned using a TOPO TA Cloning Kit for Sequencing (Life Technologies Corp. Carlsbad, CA) and Sanger sequenced by MDIBL (Salisbury Cove, ME) or the Clemson University Genomics Institute (CUGI; Clemson, SC).
The nucleotide sequences of smaller-spotted catshark aqp0, -1, -4, spiny dogfish aqp15, sea lamprey aqp01, -3L1, -3L2, and Atlantic hagfish glp are deposited in GenBank under accession numbers KJ784515, KJ784516, KJ784517, KJ815007, KJ784520, KJ784518, KJ784519 and KJ815008, respectively.
Phylogenetic and Syntenic Analyses
Aquaporin orthologs were initially retrieved from public databases (ensembl and GenBank), and the deduced amino acid sequences aligned with default t-coffee v9.01 [46] or L-INS-I MAFFT v7.058b [47] algorithms. These data were converted to codon alignments using Pal2Nal [48] and analysed using Bayesian (Mr Bayes v3.2.2; [49]) and maximum likelihood (PAUP v4b10-x86-macosx; [50]) protocols as described previously [51], [52]). Phylogenetic sorting was then achieved by arranging the sequences in accordance with the resulting tree topologies, and errors in the automated alignment were identified and corrected manually using MacVector (MacVector Inc, Cambridge, UK). Phylogenetic analyses of the deuterostome superfamily were performed on the conserved transmembrane regions between human AQP6 Lys21-Pro237, following removal of the N- and C-termini, while phylogenetic analyes of the separate classes of aquaporin were performed on full-length sequences. For Bayesian analyses the following models were tested for the codon alignments: nucmodel = 4by4 with nst = 2 or codon with nst = 6; rates = gamma and invgamma, respectively; and for amino acid alignments: aamodel = mixed. Based upon the resulting tree topologies and posterior probabilities, no significant differences were noted between nst = 2 or 6, and subsequent analyses utilised the more tractable setting of nst = 2. Markov chain Monte Carlo (MCMC) algorithms were run with 3 heated chains and 1 cold chain with 30 million MCMC generations for the superfamily alignment. Computation time thus varied from several days to 3.5 months for a given alignment. Each run was examined for convergence using Tracer version 1.5 (tree.bio.ed.ac.uk/software/tracer/), and majority rule consensus trees summarized with a burnin of 25%. All trees generated were processed with Archaeopteryx [53] and rendered with Geneious (Biomatters Ltd, New Zealand).
To establish the presence or absence of orthologs in specific taxonomic groups, and to complete partial or poorly predicted sequences, the tblastn algorithm was used to identify 5394 non-redundant aquaporin proteins or protein fragments from whole genome-shotgun (WGS), transcriptome shotgun assemblies (TSA) and expressed sequence tag (EST) databases (www.ncbi.nlm.nih.gov). We initially used full-length aquaporins as reference for the tblastn algorithm, and subsequently used aquaporin regions corresponding to translated exons to retrieve novel sequences. Contiguous nucleotide sequences were then retrieved and trimmed to match each protein fragment, and subsequently concatenated to construct a putative cDNA for each gene. Deduced amino acid sequences from the putative cDNAs were manually incorporated into the alignments previously established using t-coffee or MAFFT, which was then converted to a codon alignment using Pal2Nal. Separate analyses of 10 – 30 million MCMC generations were performed on full-length data sets comprised of classical aquaporins (Aqp0, -1, -2, -4, -5, -5L, -6, -14 and -15), Aqp8 and -16, unorthodox aquaporins (Aqp11 and -12), and aquaglyceroporins (Aqp3, -7, -9, 10 and -13). A full list of accession numbers is provided in Table S2.
To validate the topology of the resulting trees, aquaporins were localised to chromosomes in assembled genomes, and syntenic analyses conducted for conserved flanking genes using the Genomicus (www.dyogen.ens.fr) and ensembl genome browsers or manually assembled via in silico chromosomal and contig walking. This approach identified several conserved nuclear receptors (nr1b, -1d, -1f, -1i and nr4a), which we then examined in a phylogenetic framework as described above based upon 2235 assembled sequence fragments (see Table S3 for accession numbers). These latter analyses were used to evaluate the historical genome duplication events and gene losses associated with aquaporin evolution. Divergence times of the different lineages are based upon previously published timetree data [54].
Results
Four Major Grades of Deuterostome Aquaporin have Parazoan-Radiata Origins
Initial experiments using degenerate primers to identify aquaporins in Chondrichthyes, Hyperoartia (lampreys) and Hyperotreti (hagfishes) resulted in the isolation of four putative classical aquaporins from the two species of shark studied (smaller-spotted catshark, Scyliorhinus canicula; and spiny dogfish, Squalus acanthias), two partial aquaglyceroporins from the sea lamprey (Petromyzon marinus) and a full-length putative aquaglyceroporin from the Atlantic hagfish (Myxine glutinosa). Each sequence was incorporated into the superfamily alignment composed of 852 non-redundant aquaporin transcripts assembled from basal Deuterostomia (Echinodermata), Protochordata (Cephalochordata and Tunicata), Agnatha (Hyperotreti and Hyperoartia), and Gnathostomata (Chondrichthyes, Teleostei, Actinistia, Amphibia, Sauropsida and Mammalia). Incremental Bayesian analyses of the deduced amino acid and codon alignments revealed that the thirteen subfamilies (AQP0 - AQP12) established for Mammalia [55], [56] and the duplicated counterparts amongst the 10 corresponding subfamilies identified in Teleostei [36], [42], [43] are well resolved and can be traced to four major grades of water channel in Echinodermata and Protochordata (Figure 1). The term grade is used to reflect the polyphyletic nature of the four major aquaporin subdivisions. We thus initially phylogenetically classified the four major grades of deuterostome water channels based on their functional properties as classical, water-specific aquaporins (Aqp4, -1, -0, -2, -5, -6), aquaglyceroporins (Glp: including Aqp3, -7, -9, and 10), aquaammoniaporins (Aqp8-type), or unorthodox aquaporins (Aqp11, -12) [37]–[40].
The Bayesian majority rule consensus tree of the codon alignment is rooted with aqpM. The aquaporin subfamilies annotated in the central coil are separated into four major grades: aquaglyceroporins (G), unorthodox aquaporins (U), aquaporin 8-type (8) and classical aquaporins (C). Posterior probabilities are shown at selected nodes. * indicates sequencing of taxon. Evolutionary older nodes associated with Cyclostomata, Chondrichthyes and Actinisia are respectively shaded in yellow, grey and magenta. Teleost and tetrapod subclusters are shaded according to the aquaporin grade, except for sarcopterygian Aqp2, -5, and -6 paralogs, which are shaded in pink.
Surprisingly, however, the genomes of Echinodermata, Hemichordata and Cephalochordata are found to encode multiple copies of the different grades of aquaporin. To validate these observations, we researched the basal deuterostome WGS, TSA and EST databases, and separately analysed 81 assembled orthologs. These latter analyses revealed the existence of five aqp4-like, two to three aqp8-like, two aqp12-like and three glps in Echinodermata, and between one to five paralogs of each grade in the Protochordata (Figure S2). Based upon the loci and clustering of the encoded transcripts, and the low sequence conservation (23–36% deduced amino acid identities between paralogs), it is apparent that duplication of the echinoderm aquaporins occurred prior to the divergence of the echinozoan and asterozoan lineages, with at least two gene couplets, including aqp4l2-aqp4l3 and aqp4l4-aqp4l5, which are respectively colocated on purple sea urchin Scaffold664 and Scaffold916, evolving via tandem duplication. Tandem duplication is also the likely origin of the two aqp12-like paralogs in Cephalochordata, which are juxtaposed in scaffold 207_Cont30099 of the Florida lancelet (Branchiostoma floridae).
To investigate the putative origin and establish a potential root for the four major aquaporin subdivisions, we searched for ancestral eukaryotic sequences within the genomes of Cnidaria (anemones and corals), Porifera (sponges), Fungi and Protists, and to deeper prokaryotic origins in Bacteria and Archaea. The results revealed that expansion of the superfamily into four major grades occurred prior to the evolution of Bilateria (see Text S1, Figure S3–S4), and that Prokaryota encode a surprising diversity of water channels, including a novel clade (AqpN) in Bacteria and Archaea (see Text S1, Figure S5 and Table S2 for the details). The latter data further indicate that there is no clear phylogenetic root for aquaporins, however, AqpM may be considered a good approximation for metazoan trees due to its central position in the mid-rooted prokaryotic tree (Figure S5).
The systematic approach of assembling contiguous exon-like fragments to putative transcripts and establishing their orthology via Bayesian analyses identified an extensive array of non-redundant members of the aquaporin superfamily, including 590 Glps, 286 unorthodox, 221 Aqp8-type and 1004 classical aquaporins in deuterostome organisms (see Text S1). These data revealed that Vertebrata encode a broader repertoire of Glps than previously documented (Figure S6), including several tandem duplicates and a novel subfamily (Aqp13) (Figure S7), that unorthodox aquaporins (Aqp11, and -12) are ubiquitous in Gnathostomata (Figure S8), and that tetraploid Teleostei such as the Atlantic salmon encode up to eight Aqp8-type channels, while diploid Tetrapoda encode only one but retain an additional subfamily (Aqp16) (see Text S1, Figure S9–S10 for the details). In conjunction with the selected sequencing, the assembled data also revealed that two novel classical aquaporin subfamilies (Aqp14 and -15) exist in diverse lineages of vertebrate, which appear to be degraded in Eutheria (see next section).
Diversification and Lineage-Level Loss of Novel Aquaporin Subfamilies in Vertebrata
Phylogenetic analyses of the classical aquaporin-like sequences retrieved from the sea lamprey genome reveal that one can be classified as aqp4, while three more partial sequences cluster as aqp4-like, aqp1-like and aqp0-like subfamily members (Figure 1). The chondrichthyan classical aquaporin sequences isolated from the kidney, rectal gland and eye in the present study, respectively clustered as members of the aqp4, -1, and -0 subfamilies, while the fourth full-length sequence isolated from the kidney of the spiny dogfish clustered on a sister branch to the aqp1 subfamily together with the sea lamprey aqp1-like sequence, the previously analysed zebrafish aqp5/1b duplicate [36], [57] and a novel coelacanth (Latimeria chalumnae) aqp1-related ortholog (Figure 1). Confirmation that the newly isolated spiny dogfish aqp1-related sequence and the novel coelacanth aqp1-related ortholog are not derived isoforms of a gnathostome aqp1 channel was achieved by assembling the full compliment of aquaporins from the ghost shark (Callorhinchus milii), little skate (Leucoraja erinacea) and coelacanth genomes. The Bayesian analyses shown in Figure 1 indicated that the genome of each species encodes an orthologous aqp4, -1 and -0 channel, suggesting that the aqp4-like and aqp1-like channels might represent novel aquaporins not yet identified in Mammalia or any other organism. Equally striking is the observation that three basal aqp2-like paralogs retrieved from the actinistian coelacanth genome (2.6 Mb locus, Scaffold JH126563) co-cluster with the two dipnoan aqp0p orthologs previously identified in lungfish [44], just below the AQP2, -5 and -6 orthologs assembled from tetrapod genomes. The cluster pattern of the tetrapod AQP2 and -5 orthologs was sometimes reversed with equal levels of posterior probability. The preliminary phylogenetic data nevertheless seemed to support our earlier analyses [36], [42], [43] that Teleostei lack functional orthologs of AQP2, -5 and -6. In addition the present observation that two novel classical aquaporins exist in basal gnathostome and agnathan animals suggested that these orthologs might represent ancestral forms of AQP2, -5 or -6 found in Tetrapoda.
To validate the intial findings outlined in Figure 1, we extended the assembly of aquaporin repertoires to include 28 available teleost genomes, together with that of the holostean spotted gar (Lepisosteus oculatus). This resulted in the identification of 2104 peptides, which, together with the corresponding nucleotides, were assembled into 673 actinopterygian-related aquaporins. In addition, to discern whether the novel aquaporins isolated or assembled from lampreys, sharks and the coelacanth are specific to these animals, or represent ancestral orthologs that are lost in modern taxonomic lineages, we expanded our search of the vertebrate genomes to include 89 tetrapod genomes, and examined the syntenic relationships of the vertebrate aqp0, -2, -5 and -6 loci (see below). This allowed two separate data sets consisting of 485 full-length non-redundant classical aquaporin proteins and transcripts to be aligned and analysed using Bayesian protocols. The resulting trees (Figure 2 and Figure S11) were rooted with cnidarian aqp4L1, which we identified as a basal metazoan classical aquaporin (see Text S1), and revealed that extant gnathostome genomes encode eight subfamilies of classical aquaporin rather than six as previously reported [55], [56], [58]. The most basal subfamilies in Vertebrata include canonical aqp4 and the novel aqp4-like orthologs in Cyclostomata and all major clades of Gnathostomata. By localising the sea lamprey aqp4-like sequence to scaffold GL478425, we noted that it is encoded upstream of the aqp0-like sequence, and downstream of the aqp1-like sequence (Figure 3A, Figure S14). This prompted us to search for exons of an aquaporin homolog upstream of aqp0 in each gnathostome lineage, and in doing so we identified 68 orthologs of the sea lamprey aqp4-like gene, including near complete forms in the genomes of the oldest mammalian lineages, Prototheria and Metatheria. In Eutheria, however, we have currently only found exon fragments in members of the Xenarthra, Cetartiodactyla, Chiroptera and Carnivora. The C-terminal fragment identified in the genome of the bottlenosed dolphin (Tursiops truncatus) also contains a premature stop codon, supporting the notion that this branch of aquaporins is degraded to pseudogenes and is thus functionally extinct in Eutheria. Considering that the novel aqp4-like orthologs cluster on a sister branch to the canonical aqp4 transcripts, and that both are encoded in the genomes of Cyclostomata and Gnathostomata, we named the novel subfamily Aqp14 as a putative product of early chordate WGD. Although members of Aqp14 subfamily retain the cononical NPA and NPAR motifs, the low conservation of the primary structures (40–70% amino acid identity compared to methatherian Aqp14; 25–36% amino acid identity compared to human AQP0, -1, -2, -4, -5 or -6) indicates that Aqp14 channels did not evolve under strong purifying selection. Indeed, the divergence of Aqp14 channels from other classical aquaporins is clearly reflected in the longer branch lengths, but is also evident in the absence of aromatic residues in the ar/R selectivity filter [59], [60]. In lieu of the aromatic F on the α-helical transmembrane domain 2 (TMD2) and H on TMD5, the Aqp14 selectivity filter presents residues with non polar aliphatic side chains including A, V or L (TMD2), and A, V, L, I (TMD5), with some avian and metatherian orthologs encoding polar, uncharged T in the latter position (Table 1). Despite these differences, comparison of the five residues (P1–P5) that putatively distinguish between the molecular selectivity of aquaporins and aquaglycroporins [61] confirms that vertebrate Aqp14 channels retain the signature P2–P5 residues of classical water-specific aquaporins (Table S4).
Bayesian majority rule consensus tree of the codon alignment is rooted with cnidarian aqp4L1. Posterior probabilities are shown at each node, with the number of taxa analysed given in square brackets. The scale bar represents the rate of nucleotide substitution per site. Chondrichthyan, actinopteryian and sarcopteygian subclusters are respectively shaded in light grey, cyan and magenta, while evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in yellow and dark grey. The tetrapod AQP2, -5, -5-like (5L), and -6 paralogs are shaded in pink. See Figure S11 for the fully annotated tree.
(A) Synteny is shown in relation to conserved nuclear receptors, and the keratin (KRT), olfactory receptor (OR), vomeronasal receptor (VRE), homeobox C (HOXC) superclusters. Circular arrows indicate that the linkage group is flipped, and coding direction is indicated by the pointed end of the gene symbols. (B) Genomic structure of the Arctic lamprey aqp01 and -14 paralogs.
We next focused on establishing the orthology of the novel aqp1-related transcript isolated from the kidney of the spiny dogfish. The deduced amino acid sequence encodes the six canonical α-helical TMDs, the two NPA motifs, the signature ar/R residues of a classical aquaporin and the P1–P5 residues conserved in tetrapod AQP2, -5 or -6 channels. BLASTp returned hits that were equally related (∼50% amino acid identity) to AQP0, -1 or -5 indicating that this chondrichthyan ortholog could represent an ancestral form of the tetrapod channels. However, we successfully identified orthologous chondricthyan, actinopterygian and sarcopterygian sequences, which when assembled and submitted to Bayesian inference, revealed that the spiny dogfish water channel is an ancestral form of a novel aqp1-related subfamily not found in Mammalia. Based upon the topological position of the subcluster, and syntenic relationship in the genomes of Eutelostomi (Figure S12), we named this subfamily aqp15, a duplicate of aqp1 that putatively evolved from an ancient WGD event.
The origins of vertebrate Aqp0-like and Aqp1-like Integral Membrane Proteins
Having ruled out aqp14 and -15 genes as direct ancestors of tetrapod AQP2, -5 or -6, we re-examined the partial sequences identified in lampreys. In the current version of the sea lamprey genome, the aqp0-like exons downstream of aqp14 encode the N-terminal hemipore, while the aqp1-like exons upstream of aqp14 encode the C-terminal hemipore (Figure 3A, Figure S14), indicating that the remaining hemipores of each integral membrane protein should be respectively located in the 3′ prime and 5′ prime regions of the DNA. However, although we were able to assemble full-length transcripts encoding both hemipores for aqp4 (320 deduced amino acids) and aqp14 (299 deduced amino acids) from the Arctic lamprey genome, only two and three exons are respectively retrieved for the partial aqp0-like and aqp1-like fragments in this latter species (Figure 3B). As for the sea lamprey, the aqp0-like exons are located downstream of aqp14. However, in contrast to the sea lamprey, the Arctic lamprey aqp14, aqp0-like and aqp1-like fragments are encoded in-frame on the reverse strand of two consecutive DNA fragments (APJL01031881, APJL01031882). This alternative arrangement suggested that the five exons might code for a single gene, rather than two as initially surmised. To test this possibilty, we designed forward and reverse primers for the two aqp0 NPA motifs of the smaller-spotted catshark to be used for RT-PCR experiments on total RNA isolated from the eyes of the shark and the eye and kidney tissues of the two species of Hyperoartia, the sea lamprey and the European brook lamprey (Lampetra planeri). This experiment revealed a clear band of the expected size (∼570 bp) for the smaller-spotted catshark, but none for the lampreys (Figure S13). Subsequent sequencing of the 570 bp band confirmed its identity as smaller-spotted catshark aqp0. We therefore used the more complete exon data for the Arctic lamprey to design gene-specific primers for the aqp0-like and aqp1-like fragments and confirmed that each is expressed in a mixture of eye and kidney tissues obtained from the sea lamprey (data not shown). We cloned the aqp1-like C-terminal hemipore by RT-PCR and the use of 5′-prime RACE to isolate the N-terminal hemipore resulted in the amplification of a product, which when sequenced, is identical to the aqp0-like fragment, revealing that despite the non-contiguous arrangement of the exons in the sea lamprey genome, they are encoded by a single gene. Bayesian analyses of the concatenated lamprey nucleotides in a codon alignment consisting of 148 aqp0 and 210 aqp1 transcripts assembled from Gnathostomata, place the hyperoartian transcripts at the base of aqp0 rather than aqp1 with 80–86% prosterior probability (data not shown). However, as shown in Figure 2, when submitted to Bayesian inference in the broader context of all gnathostome classical aquaporins, the hyperoartian transcripts form a monophyletic root (81% posterior probability) to the gnathostome classical aquaporins comprised of aqp1, -15, aqp0 and -2-like subfamilies. This surprising result indicated that seven classical aquaporins (aqp0, -1, -2, -5, -5L, -6, and -15), encoded by four exons in Gnathostomata, may have evolved from a single ancestral form encoded by five exons as noted for extant Agnatha. Closer inspection of the aqp0-like DNA of the Arctic lamprey, however, revealed an RNA-directed DNA polymerase from a jockey-like mobile element encoded in the second intron. We therfore cannot rule out the possibility that two independent classical aquaporins (aqp0 and aqp1) evolved in Agnatha but have subsequently fused due to genomic rearrangements in lampreys. The most ancestral form identified in extant Hyperoartia is therefore annotated as aqp01. To explain this diversity, an overview of the classical aquaporins identified in Deuterostomia is shown in Figure 4.
(A) Phylogenetic relationships of the major deuterostome groups studied. (B) Prevalence of classical aquaporin genes identified in each lineage showing postscript nomenclature for duplicated forms. Coloured dots are specific for the species shown, white dots are found in the same lineage, but a different species, while grey dots indicate pseudogenes.
Genomic Rearrangements Unlink Hox Clusters as Conserved Markers of Aquaporin Evolution
To corroborate the preliminary phylogenetic analyses, we compared the karyotypic loci of the vertebrate aquaporin superfamily and examined the syntenic relationships of genes and superclusters flanking the Primate AQP0, -2, -5, and -6 paralogs (Figure 3A and Figure S14). The HOX clusters are conserved amongst different metazoan Phyla, and have thus been used as markers of genome evolution [62]–[65]. In the present context the teleost HoxAa and HoxAb clusters are respectively linked with aqp10a and -10b, and the HoxBb and HoxBa clusters are respectively linked with the aqp8aa, -8ab and -8bb paralogs, providing support for both tandem duplication and an R3 WGD origin of these aquaporins. However, for the majority of vertebrate aquaporins, the lack of linkage conservation to the HOX clusters limits comparison of their duplication history. For example, with the exception of the fused chromosomes in the metatherian opossum (Monodelphis domestica), the linkage of the HOXB cluster to AQP10 is not conserved in Tetrapoda. In the green anole (Anolis carolensis), the HOXB cluster is distally linked to AQP1, but in Eutheria, it is the HOXA cluster that is proximally linked to AQP1 and AQP4, and while the latter linkage between AQP1 and AQP4 is not conserved in Glires, it may represent an ancestral condition since these paralogs are also linked in some avian and actinopterygian genomes. A similar state appears to have evolved in ancestral Euteleostomi for aqp3 and -7, which are closely linked in the genomes of Actinopterygii and Sarcopterygii. A major exception is the HOXC cluster, which has remained proximally linked to gnathostome aqp0 and -14 at least since the Ordovician (>455 Ma), while the data for Hyperoartia reveal that the linkage between aqp01 and -14 has existed since the Cambrian.
A more holistic view of genomic rearrangements becomes apparent when mapping distantly related gnathostome linkage groups (LGs) encoding aqp0 to the human karyotype (Figure S15). This analysis revealed that much of the gene content of Hsap4 is conserved from the aqp0-bearing orthologons of holostean fishes and galliform birds, such as the turkey (Meleagris gallopavo). Broader rearrangements occurred in the genomes of Teleostei after the lineage diverged from Holostei. Thereafter, teleost gene distribution of the aqp0a and -0b LGs remained relatively stable for >200 million years of evolution. Major rearrangements are also apparent from the aqp0-linkage map of the iguanian green anole compared to the turkey, while in Mammalia, approximately 70% of the gene content of Hsap12 was established >105 Ma, prior to the divergence of Afrotheria and Boreoeutheria. In the central regions of Hsap12 (48.2–57.7 Mb) we found that, despite the extensive genomic rearrangements, 33 genes in addition to the HOXC, keratin (KRT) and olfactory receptor (OR) clusters have remained syntenic to AQP0 for >455 Ma.
Convergent Loss of Syntenic Nuclear Receptors Reflects Aquaporin Evolution
Amongst the syntenic flanking genes, are four nuclear receptors, retinoic acid receptor G (RARG), Rev-erb (NR1D4), vitamin D receptor (VDR), and NR4A1. As a relatively large superfamily with broad distribution, nuclear receptors are regarded as good markers of animal genome evolution [66], so to compare their duplication history to the aquaporins in Deuterostomia, we examined the phylogenetic relationships of the syntenic nuclear receptors and included a fifth subfamily, the rar-related orphan receptors (ROR), since they are syntenic with the aquaglyeroporins. The intriguing results revealed remarkable parallels to the lineage-level loss of deuterostome aquaporins (Table 2, Figures S16–S20 and Table S3 for accession numbers). Beyond the loss of duplicates in the aftermath of the fish-specific WGD, the syntenic loss of NR1D4 in avian and mammalian lineages and the related loss of RORD in Eutheria is highly reminiscent of the loss of AQP6 in Archosauria and AQP14 in Eutheria. Similarly the absence of the constitutive androstane receptor (CAR) in the genomes of Chondrichthyes and Actinopterygii mirrors the absence of AQP2, -5 or -6 in these same lineages, while Sauropsida have alternatively lost the paralogous pregnane-X receptor (PXR). In contrast the presence of tetraparalogous RORs that maintain highly conserved syntenies to the aqualgyeroporins of Euteleostomi, including the duplicated aqp9a, and -9b in Teleostei, suggests that positive and negative selection forces have converged to shape the genomic repertoires of nuclear receptors and aquaporins in Vertebrata.
Apomorphic Aquaporin Gene Clusters in Sarcopterygii
Integration of the phylogenetic and syntenic analyses thus demonstrates that the tetrapod AQP2, -5 and -6 paralogs represent a genomic apomorphy specific to the Sarcopterygian lineage (Table 3, Figure 5). The most ancestral forms are traced to a cluster of three aqp2-like genes encoded in the coelacanth genome between Fas apoptotic inhibitory molecule 2 (faim2) and Rac GTPase activating protein 1 (racgap1) (Figure 3A and Figure S14) with phylogenetically related aqp2-like orthologs (formerly aqp0p, [44]) in the ancient lineage of dipnoan lungfishes (Figure 2). The genomic arrangement of the coelacanth aquaporin cluster is highly reminiscent of the mammalian clusters first noted in humans [67]. However, although the primary structures of the three coelacanth paralogs have diverged significantly (33–40% amino acid substitution), the Bayesian analyses do not provide statistical support for a preferential orthology to either AQP2, -5, or -6. Nevertheless, comparison of the coelacanth and lungfish orthologs to the human primary structures revealed that the basal sarcopterygian channels are collectively more akin to AQP2 (56±3.7%), rather than AQP5 (54±1.5%), or AQP6 (47±1.3%), and we therefore annotated the paralogs as Aqp2a, -2b and -2c. To probe the basis for the reduced evolutionary rate of the coelacanth genes in relation to the amphibian cluster, we used Geneconv [68] to search for replicated nucleotide regions potentially resulting from gene conversion. No outer fragments are detected in either group, however, a significant 86 nucleotide (nt) inner fragment (nt 157–242: aqp2a; Bonferroni-corrected Karlin-Altschul Pvalue = 0.00012) is identified in the coelacanth aqp2a and -2c transcripts. This segment codes for the C-terminal region of TMD2 and α-helix 3, a peptide that includes the first NPA motif.
Schematic illustration showing aquaporin subfamilies identified in deuterstome phyla in relation to geological time.
Amphibia Utilise Ancestral Forms of AQP6 for Water Conservation
Closer examination of the AQP2, -5 and -6 gene loci in Tetrapoda (Figure 3A and Figure S14), reveals that they comprise diverse clusters of four tightly linked genes in the Western-clawed frog (Xenopus tropicalis: AQP6ub-5L-5-6vs), split sets of binary clusters (AQP2-5 and AQP5L-6) in the green anole or a split binary cluster and singleton (AQP2-5 and AQP6) in the Chinese softshell turtle (Pelodiscus sinensis), binary clusters (AQP2-5) in the archosaurian alligators and birds, and a ternary cluster (AQP2-5-6) in Mammalia. These observations also confirmed that canonical AQP2 is not encoded in the genome of the Western-clawed frog [24] or its congener the African clawed frog. The Bayesian analyes show that the wholly aquatic Western-clawed frog has two paralogs of AQP5, a canonical form (AQP5, formerly annotated as AQP-x5, [28]) and a related paralog (AQP5L, formerly annotated as AQP2) encoded between two AQP6 genes, the urinary bladder-type (AQP6ub) and the ventral skin type (AQP6vs) formerly identified as (h2-like and h3-like) type-a2 anuran specific aquaporins [23], [24], [32]. Amongst other Anura the Bayesian data provide statistical support for three paralogs of AQP6 (see Table S2), the urinary bladder type (AQP6ub) and two paralogs (AQP6vs1, -6vs2) that mediate cutaneous hydration through the ventral skin of semi-aquatic and semi-terrestrial anurans [33], [35]. Canonical AQP2 (formerly annotated as AQP-h2k or HC-2, [25], [26]) is found in the genomes of Anura, including the arboreal Pacific treefrog (Pseudacris regilla) and the aquatic green frog (Rana clamitans), but the absence of an assembled linkage map currently precludes comparison of the loci.
Loss and Functional Divergence of AQP6 in Archosauria and Mammalia
The heterogeneous lengths of the branches shown in Figure 2 illustrate the rapid evolutionary rate of the eutherian AQP6 transcripts, which, based upon the human values, have respectively undergone 1.56- and 1.39-fold higher substitution rates compared to the amphibian AQP6ub and AQP6vs transcripts. The rate of divergence is approximately twice that of the human AQP2 and -5 channels, and 3.17-fold higher compared to the coelacanth aqp2 orthologs. Part of this diversification includes the evolution of an extended N-terminal internalising domain (ENID: 10 amino acids in Laurasiatheria and Glires, 13 amino acids in Scandentia and Primates), which determines the intracellular localisation of AQP6 [69], [70], but also the divergence of the C-terminal regions including non-synonymous codon substitutions of putative phosphorylation sites (Table S5), which in AQP2 have been shown to be important for AVPR2-mediated intracellular trafficking (Moeller and Fenton 2012). The high levels of nucleotide exchange combined with genomic rearrangements may have resulted in the loss of AQP6 in the Triassic ancestors of the Archosauria. Amongst fourteen avian and two crocodylian genomes, we located the AQP2-5 binary clusters proximally linked to a second glucagon gene (GCG2), but we have not found AQP6 in these lineages. By contrast we identified novel AQP5-like genes (AQP5L) in the genomes of Iguania (green anole) and Serpentes (Burmese python, Python molurus bivittatus, and king cobra, Ophiophagus hannah). Unlike the linkage map in the green anole, the novel king cobra gene is arranged in a ternary cluster (AQP2-5-5L) flanked by FAIM2 and GCG2, while AQP6 is found in a separate DNA fragment flanked by RACGAP1 (Figure S21).
Discussion
In the present study we examine the origin and diversification of deuterostome aquaporins, and show that the selective evolution of the major channels mediating water conservation in Amphibia, Sauropsida and Mammalia contrast the ubiquity of pattern forming genes that specify the limbs, fins, lungs or swimbladders of Eutelostomi. The gene clusters harbouring paralogous forms of AQP2, -5 or -6 are specific to Sarcopterygii, the lineage that gave rise to the terrestrial vertebrates. This facet implies that aquaporins played a pivotal role for terrestrial radiation.
To understand the origin of the apomorphic aquaporin gene clusters in Sarcopterygii, we constructed a model that incorporates the putative duplication history of the superfamily (Figure 6). Due to the absence of phylogenetic convergence to a single common stem for the metazoan subfamilies, the model traces each grade back to the earliest forms. The primordial channel is thought to have arisen through intragenic duplication of a hemipore to form an integral membrane protein consisting of six TMDs, which when folded created a central pore partially restricted by the two opposing NPA domains [71]. This prototypical structure is now found in all domains of life and forms the basis of the analysed alignments. As in previous studies, the present data support an early split of the superfamily into two major branches: 1. aquaporins in which the central pore evolved a stringent selectivity for water transport, and 2. aquaglyceroporins facilitating the passage of both water and small uncharged solutes such as glycerol [55], [56], [58], [72], [73]. In contrast to previous reports, however, we find an unexpected diversity of integral membrane proteins in Archaea and Bacteria, with robust support for four major clades (see Text S1, Figure S5). Although we do not find more than one paralog in any given species of Archaea, some orders of Bacteria (Lactobacillales) encode multiple copies of GlpF in addition to AqpZ, while others (Bacillales) encode AqpN and AqpZ, and archaeal and bacterial genes are represented in each of the four clades. These novel observations are potentially consistent with the “ring of life” hypothesis involving horizontal gene transfer (HGT) and genome fusions prior to the evolution of Eukaryota [74], [75]. The subsequent endosymbiosis and operational flow of genes to the chromosomes of Eukaryota [76] may in part explain the broad diversity of aquaporins recently documented in single-celled organisms [56]. To date, however, the phylogenetic interrelations of aquaporins between single-celled and metazoan organisms remains mostly obscure, which, as noted in the present study, may be masked by the presence of endosymbiont genes. The current evidence indicates that Protists and Fungi encode both multiple copies of aquaglyeroporins, and a separate set of aquaporins more related to Aqp8 rather than Aqp4 or the unorthodox aquaporins.
Archaea and Bacteria are encapsulated in a putative “ring of life” scenario. Dotted lines indicate uncertain relationships. White nodes indicate tandem duplication, coloured nodes are associated with serial rounds (R1–R3) of whole genome duplication. The nomenclature and model are explained further in the text. The tetrapodomorph shown above node 4 is Tiktaalik roseae, which likely harboured the aqp2-like genes as shown for Actinistia and Dipnoi in Figure 2.
Our previous analyses of the evolution of piscine aquaporins provided evidence for four major grades of integral membrane proteins in Vertebrata [36], [42], [43]. More recently this concept has been extended to plants [77], [78], however the origin of plant glycerol transporters, including NIPs and GIPs, may have occurred via HGT [55], [56], indicating that the gene flows are not coalescent. For basal metazoan organisms, we find putative evidence of HGT in Cnidaria, which lack a phylogenetically resolved branch for Aqp8. In this instance the functions of Aqp8 may have been co-opted by endosymbiotic zooanthellae or bacteria. Amongst Porifera the phylogenetic evidence supports the presence of an ancestral Aqp8-like channel in addition to an aquaglyceroporin, while the latter function may also have expanded via HGT or bacterial symbiosis. Thus, with the exception of an orthologous aqp8 gene in Cnidaria, we successfully traced the four major grades of deuterostome aquaporin to the parazoan-eumetazoan divide, which is estimated to have occurred >1000 Ma [79].
It is well established that serial rounds of WGD played a major role in determining the genomic architecture of chordate animals [65], [80]–[83]. A surprising finding is thus the prevalence of aquaporin clusters encoded in nearly every lineage examined. We not only identify novel gene clusters in Radiata, Echinodermata and Cephalochordata, but find evidence of independent intrachromosomal duplications in Hyperoartia (aqp01, -14, -10L1, -10L2), Chondrichthyes (aqp3, -3L), Actinopterygii (aqp1aa, -1ab1, -1ab2, -3aa, -3ab, -8aa, -8ab, -10aa, -10ab) and Sarcopterygii (aqp2a, -2b, -2c, -5, -5L, -6ub, -6vs). Similar gene clusters have also been reported for Panarthropoda [84], [85]. This recurrent feature of aquaporin evolution indicates that the superfamily is not only prone to errors during DNA replication, but that positive selection pressure favours retention of the duplicates. For example, the recent analysis of the rainbow trout (Oncorhynchus mykiss) genome [86] revealed that approximately half of the protein-coding genes arising from an 88 - 103 Ma WGD in the salmonid lineage [87] are now fractionated, yet our anlysis of the aquaporin repertoire in tetraploid salmonids revealed that they retain twice the gene copy number compared to diploid Teleostei.
In an effort to separate the mechanism of gene expansion between tandem duplication and WGD, we combined broad taxon sampling with synteny. The model thus favours tandem duplication when the gene clusters are proximally linked in the DNA of a restricted subset of organisms, and WGD when the phylogenetic signal supports major branches separated at the level of Hyperoatia, Gnathostomata or Teleostei. The timing of WGD, however, remains an open topic, with recent studies suggesting that two rounds likely occurred before the divergence of ancestral lamprey and gnathostome lineages [88], [89]. If this is true then at least seven classes of aquaporin should have been lost from the genomes of the two species of lamprey studied. Such a scenario cannot be ruled out since genome reduction is a dominant mode of evolution [90], which in Vertebrata has been associated with multiple chromosomal fusions and rearrangements [91]–[93]. A further confounding feature is the discovery of genome remodelling in Hyperoartia, whereby ∼20% of germ line DNA is eliminated from many somatic cell lineages during embryonic development [94] Thus, although we identify seven classes of aquaporin in Hyperoartia, and show that at least seventeen classes exist in extant Gnathostomata, the timing of R2 remains an open question.
The application of the WGD theory is consistent with the evolution of at least three of the novel classes of vertebrate aquaporin reported here, including aqp14, -15 and -16. The modular combination of WGD and tandem duplication also explains why there are eight aqp8 genes in the tetraploid salmon, but only single orthologs in diploid Tetrapoda. The situation for the classical aquaporins is more complex. The model shown in Figure 6 predicts that aqp4 and aqp8 genes likely arose by duplication at node 1, and while the former aqp4 grade of aquaporins expanded independently in Cnidaria, Echinodermata, and Protochordata, diversification into six subfamilies encompassing aqp4, -14, -1, -15, -0 and -2 did not occur until WGD began to shape the vertebrate genomes in the Cambrian. Assuming that R2 occurred after the divergence of ancestral lamprey and gnathostome lineages, we expected to find two classical aquaporins in the hyperoartian genomes that could explain the WGD origin of aqp1, -15, -0 and -2. We initially found evidence that supported this scenario, but we subsequently experimentaly demonstrated that the aqp1-like and aqp0-like fragments are encoded by a single gene proximally linked to aqp14. Nevertheless the unorthodox exon structure coupled with the presence of a mobile element encoded between the two hemipores indicates that two ancestral genes arose at node 2, and subsequently fused in lampreys leaving the aqp01 chimaera and a ghost of a duplication at node 3. Fusion genes are well documented in Drosphila [95] and are recognised as a major source of oncogenes in humans [96]. In lampreys we speculate that the generation of such gene chimaeras may also be a consequence of the programmed genome rearrangements during embryogenesis [97]. While studies of hagfishes (Hyperotreti) may shed light on the origin of the chimaeric gene, the novel observation that an aqp0-like water channel is expressed in the sea lamprey eye is consistant with the evolution of multifocal lenses in Hyperoartia after the lineage diverged from Hyperotreti [98].
It has previously been suggested that duplication of aqp0 in Teleostei and subsequent expansion via tandem duplication could explain the origin of the amphibian AQP2, -5 and -a2 genes [24]. The underlying rationale for this suggestion was the location of aqp0 between faim2 and racgap1 in the medaka (Oryzias latipes), green-spotted pufferfish (Tetraodon nigroviridis) and human genomes. While certainly a plausible scenario, it does not account for the theory of WGD and rediploidisation associated with chromosomal fusions and rearrangements [90]–[93]. Nevertheless, in favour of the tandem duplication theory is the ancient linkage of the three coelacanth aqp2-like genes with faim2, racgap1 and aqp0, and the absence of either aqp2, -5 or -6 in the available genomes of Chondrichthyes and Actinopterygii. On the contrary, in favour of a WGD origin of the apomorphic genes in Sarcopterygii is the new evidence that Gnathostomata encode the novel aqp15 subfamily in addition to the aqp1 and aqp0 subfamilies, while the phylogenetic analyses of syntenic nuclear receptors revealed that lineage-level loss of gene subfamilies is common during vertebrate evolution. Indeed an earlier study examining the phylogenetic relationships of the amphibian CAR genes proposed that this xenosensor arose through WGD, but was subsequently lost in the fish lineage [99]. In the present study we further show parallels between the lineage-level loss the NR1D4 and RORD subfamilies and AQP14 in Eutheria. The model thus posits that the basal sarcopterygian aqp2-like genes may equally well have arisen via WGD at node 4, but were subsequently lost in Actinopterygii. In any event the new data presented here reveal that the genomic apomorphy existed prior to the divergence of the actinistian and lissamphibian lineages >430 Ma [100], and may have arisen as early as the Cambrian. This timing predates the oldest Devonian tetrapodomorph fossil Tiktaalik rosiae [101], [102] and the earliest evidence of tetrapod-like trackways [103] by ∼35–135 million years.
The reduced rate of nucleotide substitution of the coleacanth aqp2-like transcripts is consistant with the significantly slower rate of protein evolution in this organism [104]. The present findings indicate that gene conversion may have been a contributing factor to such morphological retardation. By contrast we find that the AQP2, -5 and particularly -6 transcripts show an accelerated rate of divergence in Tetrapoda compared to the actinistian forms. The Bayesian and syntenic analyses show that the previously identified AQPa2 anuran-specific paralogs (AQP-h2 and -h3, [23], [24], [30], [34], [35]) are ancestral forms of AQP6 encoded at the outer edges of the genomic retro-cluster. These observations suggest that the ancestral function and regulation of AQP6 was closely related to that of amphibian AQP2, whereby neurohypophysial control of vasotocin-like receptors facilitated the transepithelial uptake of water through the ventral skin of Amphibia and the recycling of water from the urinary bladder [25], [30]–[35]. The conservation of the amphibian AQP2, -5, -5L and -6 C-terminal ‘S256’ P1 residues that are recognised phosphorylation sites for AQP2-AVPR2-mediated trafficking [105] further suggests that the proteins have retained ancestral properties. However, this changed in the Sauropsida and Mammalia. In Squamata, the gene cluster is differentially split in Iguania and Serpentes, although both groups retain the novel AQP5L gene. The Archosauria appear to have lost AQP6 leaving only AQP2-5 linked to GCG2 and the conserved flanking genes, a feature that may be related to the evolution of a uric acid habitus and loss of the urinary bladder [17], [18], [106]. In Mammalia only one AQP6 is retained, which unlike AQP14, survived extinction associated with the rapid rate of nucleotide substitution, and neofunctionalised. Mammalian AQP6 is localized in the intracellular vesicles of the glomerular podocyte cell bodies and foot processes, and the intercalating cells in the outer and inner medullary collecting ducts [69]. It has diminished water permeability properties that are only activated at low pH, and it is thus considered a vacuolar-type water channel that contributes to acid-base balance, tubular endocytosis, glomerular filtration, and ion (NO3−, Cl−) transport [29], [107], [108]. The novel functions of mammalian AQP6 and loss in the avian lineage are consistent with the absence of water uptake via the skin or recycling through urinary bladders, and potentially reflects the selection pressure leading to more efficient AQP2-mediated antidiuresis in these clades.
The adaptive function of AQP5 appears to be associated with its glandular localization. In Anura it is expressed in mucus cells and small granular glands of the skin, contributing to the maintenance of moist skin, cutaneous gas exchange and thermoregulation [24], [28]. In Mammalia it is expressed in the lung and sweat glands where, in the latter instance, it also contributes to evaporative heat-loss associated with endothermy [29]. In Sauropsida, AQP5 is expressed in the venom gland of snakes [109], and the nasal salt-secreting glands of birds [22], a related organ to the lingual glands of salt-water crocodiles [19], [21], [110]. Although freshwater alligators lack the lingual salt glands, all Crocodylia shed tears through their lacrimal and tongue-like Harderian glands [106]. Both of these latter glands are specific to the Tetrapoda with the lacrimal appearing secondarily in the Amniota and eventually superceding the porphyin and lipid-secreting mammalian Harderian gland in adult Primates [111]. The expression of AQP5 in lacrimal glands is well documented [29], [112], and the early role of the two types of gland in terrestrial adaptation is associated with hydration of the cornea and the nictitating membrane [113].
The water-conserving role of AQP2 is now well documented in the renal collecting ducts of each of the three extant clades of Tetrapoda, but its evolutionary significance becomes manifest in humans with nephrogenic diabetes insipidus. This hereditary or acquired disease is caused by mutations in AQP2 or the AVPR2 receptor and results in the excretion of large volumes of dilute urine [114]–[116], a rapidly dehydrating syndrome that reflects the osmoregulatory modus operandi of freshwater fish. Indeed polyuria is the ancestral mechanism by which freshwater fish including Hyperoartia, Chondrichthyes and Actinopterygii, maintain their hyperosomotic condition in the face of substantial osmotic gradients (∼240–340: 1 mOsm, [14], [117]–[119]. We therefore suggest that it is the lack of equivalent homeostatic mechanisms involving the AVPR2-AQP2, -5 or -6 systems that has precluded actinopterygian fishes from making the permanent transition from water to land. By contrast, evidence showing that the most basal ichthyan Sarcopterygii (Actinistia and Dipnoi) encode Aqp2-like channels is consistent with observations that estivating lungfishes activate this channel via a similar system to the AVT-AVPR2-AQP2 axis of the mammalian kidney [44] and that this axis is mediated by the sarcopterygian-specific steroid aldosterone [120] to facilitate water conservation in the terrestrial environment.
Although the functional divergence of the apomorphic aquaporin genes took many millions of years before water conservation systems were sufficiently evolved in Tetrapoda to allow permanent habitation of the terrestial environment, it was nevertheless possible due to the fact that they existed in their genomes. In other words, the ability to adapt to novel environments is rooted in the genetic makeup of an organism. This can therefore be considered as genomic competence, a permissive condition that may be modulated through Darwinian selection. Unlike the ubiquitous transcription factors involved in limb and lung development, we propose that the selective evolution of water conserving aquaporins in the sarcopterygian lineage represents a permissive adaptation that facilitated tetrapod terrestrial adaptation.
Supporting Information
Figure S1.
Phylogenetic interrelationships of deuterostome animals for which genomic complements of aquaporins were studied. The number of taxa within each group are shown to the right. Square nodes indicate whole genome duplication. Circular dots on terminal branches indicate an assembled linkage map.
https://doi.org/10.1371/journal.pone.0113686.s001
(PDF)
Figure S2.
Annotated Bayesian majority rule consensus tree of the basal deuterostome aquaporin superfamily. The tree is mid-point rooted. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. Four major grades of aquaporin, Aqp4-like, Aqp8-like, Glp and Aqp12-like are respectively shaded in blue, cyan, green and orange with paralogues of a given subfamily highlighted and numerically labeled.
https://doi.org/10.1371/journal.pone.0113686.s002
(PDF)
Figure S3.
Annotated Bayesian majority rule consensus tree of Parazoa-Radiata aquaporins. The tree is rooted with aqpM. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. Four major grades of aquaporin Aqp4-like, Aqp8-like, Aqp12-like and Glp are respectively shaded in blue, cyan, orange and green with paralogues of a given subfamily highlighted and numerically labeled. Putative endosymbiont aquaporin and GlpF orthologs are respectively shaded in grey and magenta. Tandemly arranged paralogs are shown as numerically linked circles.
https://doi.org/10.1371/journal.pone.0113686.s003
(PDF)
Figure S4.
Summarised Bayesian tree of the Parazoa-Cnidaria-Basal deuterostome aquaporins. The tree is rooted with aqpM. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site.
https://doi.org/10.1371/journal.pone.0113686.s004
(PDF)
Figure S5.
Annotated Bayesian majority rule consensus tree of archaeal and bacterial aquaporins. The tree is mid-point rooted. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. Four major grades of Bacterial and Archaean aquaporins are shaded in colour, with Archaean orthologs highlighted in pink to show the polyphyletic clustering of Archaean AqpZ orthologs.
https://doi.org/10.1371/journal.pone.0113686.s005
(PDF)
Figure S6.
Annotated Bayesian majority rule consensus tree of deuetrostome aquaglyceroporins. The tree is rooted with Hydra vulgaris Glp2. Posterior probabilities resulting from analyses of the codon alignments are shown at each node. Evolutionary older nodes associated with Basal Deuterostomia, Cyclostomata, Chondrichthyes, Holostei and Actinisia are respectively shaded in peach, yellow, grey, cyan and magenta. Teleost and tetrapod subclusters are shaded in green.
https://doi.org/10.1371/journal.pone.0113686.s006
(PDF)
Figure S7.
Syntenic alignment of AQP13. Genomic regions (Mb) are given in parentheses. The coding direction is indicated by the pointed end of the gene symbol.
https://doi.org/10.1371/journal.pone.0113686.s007
(PDF)
Figure S8.
Annotated Bayesian majority rule consensus tree of deuterostome unorthodox aquaporin proteins. The tree is rooted with starlet sea anemone Aqp12. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. Gnathostome Aqp12 and Aqp11 paralogs are respectively shaded cyan, and light orange. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Nodes consistent with whole genome duplications (R2, R3) are labelled.
https://doi.org/10.1371/journal.pone.0113686.s008
(PDF)
Figure S9.
Annotated Bayesian majority rule consensus tree of the deuterostome aquaporin 8 grade. The tree is rooted with bacterial aqpZ. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. The teleost aqp8aa and -8ab tandem duplicates are shaded in green and the teleost aqp8ba and -8bb genomic duplicates are shaded in yellow. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey, while tetrapod AQP16 orthologs are shaded in pink. Nodes consistent with whole genome duplications (R2, R3) are labelled. A schematic of the proposed evolution of actinopterygian aqp8 tandem duplicates and genome duplicates is shown to the right.
https://doi.org/10.1371/journal.pone.0113686.s009
(PDF)
Figure S10.
Syntenic alignment of aqp8 genes in Actinopterygii. Genomic regions (Mb) are given in parentheses with circular arrows indicating that the region is flipped in relation to orthologous regions. Coding direction is indicated by the pointed end of the gene symbol. Genes that are previously isolated and cloned [36], [121] are shown as blue symbols (A) Data for actinopterygian aqp8aa and -8ab paralogs. (B) Data for actinopterygian aqp8ba and -8bb paralogs. The zebrafish segmental duplication is boxed in green.
https://doi.org/10.1371/journal.pone.0113686.s010
(PDF)
Figure S11.
Annotated Bayesian majority rule consensus tree of deuterostome classical aquaporins. The tree is rooted with Nematostella vectensis aqp4L1. Posterior probabilities resulting from analyses of the codon alignments are shown at each node. Nodes consistent with whole genome duplications are labelled with black circles, and nodes consistent with tandem duplications are labeled with black diamonds. Evolutionary older nodes associated with Cyclostomata, Chondrichthyes, Holostei and Actinisia are respectively shaded in yellow, grey, cyan and magenta. Teleost and tetrapod subclusters are shaded in light blue for AQP0, -1, -15, -4 and -14, and pink for AQP2, -5-, 5-like (5L) and -6.
https://doi.org/10.1371/journal.pone.0113686.s011
(PDF)
Figure S12.
Syntenic alignment of AQP15. Genomic regions (Mb) are given in parentheses. The coding direction is indicated by the pointed end of the gene symbol.
https://doi.org/10.1371/journal.pone.0113686.s012
(PDF)
Figure S13.
PCR products amplified from Chondrichthyes and Hyperoartia using primers for aqp0. Representative gel image of RT-PCR analysis using primers flanking the NPA motifs of aqp0 cDNA for the smaller-spotted catshark (Scyliorhinus canicula). Lanes 1-4 are the DNA ladder, and PCR products amplified from the eye mRNA of the smaller-spotted catshark, sea lamprey (Petromyzon marinus) and brook lamprey (Lampetra planeri), respectively.
https://doi.org/10.1371/journal.pone.0113686.s013
(PDF)
Figure S14.
Chromosomal loci and synteny of chordate aquaporins. (A) Linkage groups are drawn to scale within species showing approximate locations of the aquaporin superfamily in each organism. HOX-cluster-bearing chromosomes are coloured according to the key including teleost “a” and “b” duplicated clusters. For organisms without an assembled genome, the presence of a given paralog is annotated with a tick. (B) Syntenic arrangement of vertebrate aqp0, -2, -5, -5L, -6 and -14 genes. Genes are drawn in accordance with their contiguous coding in each genome, with dashed linker lines illustrating rearrangements between the lineages. Nuclear receptors analysed via Bayesian protocols are shown in red, with † indicating extinction of the gene. The AQP2, -5, -5L and -6 gene clusters annotated with pink gene symbols are outlined in light grey to illustrate conservation in Mammalia, but split clusters in Sauropsida. Proximally linked FAIM2 and RACGAP1 are annoted in blue. Circular arrows indicate that the genomic region is flipped.
https://doi.org/10.1371/journal.pone.0113686.s014
(PDF)
Figure S15.
aqp0 linkage group maps to the human karyotype. The linkage group of each organism is mapped using the karyotype view available in Genomicus. The human karyotype is scaled according to the gene copy number (GCN). Boxed numbers represent estimates of the lineage divergence times in millions of years before present (Ma) after Hedges and Kumar [100].
https://doi.org/10.1371/journal.pone.0113686.s015
(PDF)
Figure S16.
Annotated Bayesian majority rule consensus tree of deuterostome retinoic acid receptors. The tree is mid-point rooted. Posterior probabilities resulting from analyses of the codon/amino acid alignments are shown at each node, with the scale bar indicating the rate of substitutions per site. Gnathostome Rara, Rarb and Rarg paralogs are respectively shaded cyan, green and magenta. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Nodes consistent with whole genome duplications (R1, R2, R3) are labelled with black circles. Receptors that are syntenic with aquaporins are indicated with coloured squares.
https://doi.org/10.1371/journal.pone.0113686.s016
(PDF)
Figure S17.
Annotated Bayesian majority rule consensus tree of the deuterostome Nr1d family of nuclear receptors. The tree is rooted with yellow-fever mosquito sevenup (svp). Gnathostome Nr1d1, Nr1d2 and Euteleostomi Nr1d4 paralogs are respectively shaded cyan, green and magenta. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Labels and annotations are as for Supplementary Figure S16.
https://doi.org/10.1371/journal.pone.0113686.s017
(PDF)
Figure S18.
Annotated Bayesian majority rule consensus tree of deuterostome constitutive androgen (CAR), pregnane-X (PXR) and vitamin D (VDR) receptors. The tree is rooted with yellow-fever mosquito svp. Sarcopterygian Car, and gnathostome Pxr and Vdr paralogs are respectively shaded cyan, green and magenta. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Labels and annotations are as for Supplementary Figure S16.
https://doi.org/10.1371/journal.pone.0113686.s018
(PDF)
Figure S19.
Annotated Bayesian majority rule consensus tree of the deuterostome Nr4A family of nuclear receptors. The tree is rooted with yellow-fever mosquito svp. Gnathostome Nr4a1, Nr4a2 and Nr4a3 paralogs are respectively shaded cyan, green and magenta. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Labels and annotations are as for Supplementary Figure S16.
https://doi.org/10.1371/journal.pone.0113686.s019
(PDF)
Figure S20.
Annotated Bayesian majority rule consensus tree of deuterostome retinoic-related orphan receptors (Ror). The tetraparalogous topology of gnathostome Rors is characterised here for the first time, with Rord identified in Actinopteryii and all extant sarcopterygian lineages except Eutheria. The tree is rooted with yellow-fever mosquito svp. Gnathostome Rora, Rorc, Rorb and Rord paralogs are respectively shaded cyan, red, green and magenta. Evolutionary older nodes associated with Cyclostomata and basal Deuterostomia are respectively shaded in orange and grey. Labels and annotations are as for Supplementary Figure S16.
https://doi.org/10.1371/journal.pone.0113686.s020
(PDF)
Figure S21.
Genomic arrangements of the king cobra AQP2, -5, -5L and -6 genes. Exons are labeled in accordance with the coding direction of each gene. Circular arrow indicates that the genomic region is flipped.
https://doi.org/10.1371/journal.pone.0113686.s021
(PDF)
Table S1.
Primer sequences and PCR conditions for cloning of aquaporin cDNAs.
https://doi.org/10.1371/journal.pone.0113686.s022
(PDF)
Table S2.
List of aquaporin accession numbers used in the study.
https://doi.org/10.1371/journal.pone.0113686.s023
(PDF)
Table S3.
List of nuclear receptor accession numbers used in the study.
https://doi.org/10.1371/journal.pone.0113686.s024
(PDF)
Table S4.
Conservation of residues that putatively determine channel transport selectivities of aquaporins and aquaglyceroporins in Deuterostomia.
https://doi.org/10.1371/journal.pone.0113686.s025
(PDF)
Table S5.
Conservation of AQP2-S256 (P1), S261 (P2) in Sarcopterygii.
https://doi.org/10.1371/journal.pone.0113686.s026
(PDF)
Text S1.
Additional text concerning the unexpected diversity of prokaryotic and basal metazoan aquaporins, the broader repertoire of aquaglyceroporins and the ubiquity of unorthodox aquaporins in Vertebrata, and the observation of eight Aqp8s in tetraploid teleostei vs one and a novel subfamily in diploid Tetrapoda.
https://doi.org/10.1371/journal.pone.0113686.s027
(DOC)
Acknowledgments
We thank Nick Johnson and Roger A Bergstedt of the Hammond Bay Biological Station, Millersburg, Michigan, and Stacia A. Sower of the Center for Molecular and Comparative Endocrinology, University of New Hampshire, Durham, New Hampshire, for the US lamprey samples, and Catarina Pereira of the Centre of Environmental Biology, University of Lisbon, Lisbon (Portugal) for samples from lampreys of the Iberian Peninsula. We also thank Patrici Bultó from L'Aquàrium de Barcelona, Barcelona (Spain) for supplying live specimens of smaller-spotted catshark, and Michelle Bailey and Charles Wray from the Mount Desert Island Laboratory for the procurement of experimental animals and other support, and Gordon Cramb and Lara Meischke for their initial work on elasmobranch aqp15. We further thank S. Halifu, A. Oltu and J. Solsvik (University of Bergen, Norway), C. Notredame (Centre de Regulació Genòmica, Barcelona, Spain), and H. Roest-Crollius (Ecole Normale Supérieure, Paris, France) for their assistance. We are also thankful to Christian Zmasek, the developer of Archeopteryx, and to the developers of Geneious Pro and MacVector for incorporating our requested software features.
Author Contributions
Conceived and designed the experiments: RNF JC CPC. Performed the experiments: RNF FC CPC. Analyzed the data: RNF FC JC. Contributed reagents/materials/analysis tools: RNF JBH CPC JC. Wrote the paper: RNF CPC JC. Contributed original artwork: JBH.
References
- 1. Grosberg RK, Vermeij GJ, Wainwright PC. Biodiversity in water and on land. Curr Biol. 2012;22:R900–R903.
- 2. Carroll RL. The origin and early radiation of terrestrial vertebrates. J Paleontol. 2001;75:1202–1213.
- 3. Shubin N, Tabin C, Carroll S. Fossils, genes and the evolution of animal limbs. Nature. 1997;388:639–648.
- 4. Ruvinsky I, Gibson-Brown JJ. Genetic and developmental bases of serial homology in vertebrate limb evolution. Development. 2000;127:5233–5244.
- 5. Zhao XL, Sirbu IO, Mlc FA, Molotkova N, Molotkov A, et al. Retinoic acid promotes limb induction through effects on body axis extension but is unnecessary for limb patterning. Curr Biol. 2009;19:1050–1057.
- 6. Zhang J, Wagh P, Guay D, Sanchez-Pulido L, Padhi BK, et al. Loss of fish actinotrichia proteins and the fin-to-limb transition. Nature. 2010;466:234–U106.
- 7. Freitas R, Gomez-Marin C, Mark WJ, Casares F, Gómez-Skarmeta JL. Hoxd13 Contribution to the Evolution of Vertebrate Appendages. Dev Cell. 2012;23:1219–1229.
- 8. Schneider I, Shubin NH. Making Limbs from Fins. Dev Cell. 2012;23:1121–1122.
- 9. Morris C. The origin orf lungs, a chapter in evolution. American Naturalist. 1892;26:975–986.
- 10. Daniels CB, Orgeig S, Sullivan LC, Ling N, Bennett MB, et al. The origin and evolution of the surfactant system in fish: Insights into the evolution of lungs and swim bladders. Physiol Biochem Zool. 2004;77:732–749.
- 11. Cass AN, Servetnick MD, McCune AR. Expression of a lung developmental cassette in the adult and developing zebrafish swimbladder. Evol Dev. 2013;15:119–132.
- 12.
Graham JB. Air-Breathing Fishes. Evolution Diversity and Adaptation. New York: Academic Press. 1997;299 p.
- 13.
Bernard C. Leçons sur les propriétés physiologique et les alterations pathologique des liquids de l'organsime. France: Baillère. 1859.
- 14.
Schmidt-Nielsen K. Animal Physiology: Adaptation and Environment. Cambridge: Cambridge University Press. 1997.
- 15. Lau KK, Yang Y, Cook GA, Wyatt RJ, Nishimura H. Control of aquaporin 2 expression in collecting ducts of quail kidneys. Gen Comp Endocr. 2009;160:288–294.
- 16. Nishimura H, Fan Z. Sodium and water transport and urine concentration in avian kidney. Symp Soc Exp Biol. 2002;54:129–151.
- 17.
Smith HW. From fish to philosopher. The story of our internal environment. USA: CIBA Pharmaceutical Products Ltd. 1953.
- 18. Nishimura H. Urine concentration and avian aquaporin water channels. Pflug Arch Eur J Phy. 2008;456:755–768.
- 19. Schmidt-Nielsen K, Fänge R. Salt glands in marine reptiles. Nature. 1958;182:783–785.
- 20. Shuttleworth TJ, Hildebrandt JP. Vertebrate salt glands: Short- and long-term regulation of function. J Exp Zool. 1999;283:689–701.
- 21. Cramp RL, Meyer EA, Sparks N, Franklin CE. Functional and morphological plasticity of crocodile (Crocodylus porosus) salt glands. J Exp Biol. 2008;211:1482–1489.
- 22. Müller C, Sendler M, Hildebrandt JP. Downregulation of aquaporins 1 and 5 in nasal gland by osmotic stress in ducklings, Anas platyrhynchos: implications for the production of hypertonic fluid. J Exp Biol. 2006;209:4067–4076.
- 23. Suzuki M, Ogushi THY, Tanaka S. Amphibian aquaporins and adaptation to terrestrial environments: A review. Comp Biochem Physiol A-Mol Integr Physiol. 2007;148:72–81.
- 24. Suzuki M, Tanaka S. Molecular and cellular regulation of water homeostasis in anuran amphibians by aquaporins. Comp Biochem Phys. 2009;153A:231–241.
- 25. Ogushi Y, Mochida H, Nakakura T, Suzuki M, Tanaka S. Immunocytochemical and phylogenetic analyses of an arginine vasotocin-dependent aquaporin, AQP-h2K, specifically expressed in the kidney of the tree frog, Hyla japonica. Endocrinology. 2007;148:5891–5901.
- 26. Pandey RN, Yaganti S, Coffey S, Frisbie J, Alnajjar K, et al. Expression and immunolocalization of aquaporins HC-1,-2, and-3 in Cope's gray treefrog, Hyla chrysoscelis. Comp Biochem Physiol A-Mol Integr Physiol. 2010;157:86–94.
- 27.
Gwee PC, Tay BH, Brenner S, Venkatesh B. Characterization of the neurohypophysial hormone gene loci in elephant shark and the Japanese lamprey: Origin of the vertebrate neurohypophysial hormone genes. 2008; BMC Evol Biol 9: ARTN 47.
- 28. Kubota M, Hasegawa T, Nakakura T, Tanii H, Suzuki M, et al. Molecular and cellular characterization of a new aquaporin, AQP-x5, specifically expressed in the small granular glands of Xenopus skin. J Exp Biol. 2006;209:3199–3208.
- 29. Takata K, Matsuzaki T, Tajika Y. Aquaporins: water channel proteins of the cell membrane. Prog Histochem Cyto. 2004;39:1–83.
- 30. Hasegawa T, Tanii H, Suzuki M, Tanaka S. Regulation of water absorption in the frog skins by two vasotocin-dependent water-channel aquaporins, AQP-h2 and AQP-h3. Endocrinology. 2003;144:4087–4096.
- 31. Hasegawa T, Suzuki M, Tanaka S. Immunocytochemical studies on translocation of phosphorylated aquaporin-h2 protein in granular cells of the frog urinary bladder before and after stimulation with vasotocin. Cell Tissue Res. 2005;322:407–415.
- 32. Ogushi Y, Tsuzuki A, Sato M, Mochida H, Okada R, et al. The water-absorption region of ventral skin of several semiterrestrial and aquatic anuran amphibians identified by aquaporins. Am J Physiol-Regul Integr Comp Physiol. 2010;299:R1150–R1162.
- 33. Ogushi Y, Kitagawa D, Hasegawa T, Suzuki M, Tanaka S. Correlation between aquaporin and water permeability in response to vasotocin, hydrin and beta-adrenergic effectors in the ventral pelvic skin of the tree frog Hyla japonica. J Exp Biol. 2010;213:288–294.
- 34. Shibata Y, Takeuchi H-, Hasegawa T, Suzuki M, Tanaka S, et al. Localization of Water Channels in the Skin of Two Species of Desert Toads, Anaxyrus (Bufo) punctatus and Incilius (Bufo) alvarius. Zool Sci. 2011;28:664–670.
- 35. Saitoh Y, Ogushi Y, Okada, Tanaka S, Suzuki MR, et al. Novel vasotocin-regulated aquaporins expressed in the ventral skin of semiaquatic anuran amphibians: Evolution of cutaneous water-absorbing mechanisms. General Endocrinology. 2014;155:2166–2177.
- 36. Tingaud-Sequeira A, Calusinska M, Chauvigné F, Lozano J, Finn RN, et al. The zebrafish genome encodes the largest vertebrate repertoire of functional aquaporins with dual parology and substrate specificities similar to tetrapods. BMC Evol Biol. 2010;10:38.
- 37. Sweet G, Gandor C, Voegele R, Wittekindt N, Beuerle J, et al. Glycerol facilitator of Escherichia coli: cloning of glpF and identification of the glpF product. J Bacteriol. 1990;172:424–430.
- 38. Preston GM, Carroll TP, Guggino WB, Agre P. Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science. 1992;256:385–387.
- 39. Jahn TP, Moller ALB, Zeuthen T, Holm LM, Klaerke DA, et al. Aquaporin homologues in plants and mammals transport ammonia. FEBS Lett. 2004;574:31–36.
- 40. Wu B, Beitz E. Aquaporins with selectivity for unconventional permeants. Cell Mol Life Sci. 2007;64:2413–2421.
- 41. Hamdi M, Sanchez MA, Beene LC, Liu Q, Landfear SM, et al. Arsenic transport by zebrafish aquaglyceroporins. BMC Mol Biol. 2009;10:104.
- 42. Cerdà J, Finn RN. Piscine Aquaporins: An Overview of Recent Advances. J Exp Zool. 2010;313A:623–650.
- 43. Finn RN, Cerdà J. Aquaporin evolution in fishes. Front Physiol. 2011;2:44.
- 44. Konno N, Hyodo S, Yamaguchi Y, Matsuda K, Uchiyama M. Vasotocin/V2-Type Receptor/Aquaporin Axis Exists in African Lungfish Kidney but Is Functional Only in Terrestrial Condition. Endocrinology. 2010;151:1089–1096.
- 45. Cutler CP, Cramb G. Differential expression of absorptive intestinal and renal tissues of the cation-chloride-cotransporters in the European eel (Anguilla anguilla). Comp Biochem Physiol B-Biochem Mol Biol. 2008;149:63–73.
- 46. Notredame C, Higgins DG, Heringa J. T-Coffee: A novel method for fast and accurate multiple sequence alignment. J Mol Biol. 2000;302:205–217.
- 47. Katoh K, Toh H. Recent developments in the MAFFT multiple sequence alignment program. Brief Bioinform. 2008;9:286–298.
- 48. Suyama M, Torrents D, Bork P. PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res. 2006;34:W609–W612.
- 49. Ronquist F, Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics. 2003;19:1572–1574.
- 50.
Swafford DL. PAUP*. Phylogenetic analysis using parsimony (* and other models). Version 4.0b10 for macintosh.2002; (www.sinauer.com/detail.php?id=8060). Sunderland, MASS: Sinauer Associates Inc.
- 51. Finn RN, Kristoffersen BA. Vertebrate vitellogenin gene duplication in relation to the “3R hypothesis”: Correlation to the pelagic egg and the oceanic radiation of teleosts. PLoS ONE. 2007;2:e169.
- 52. Zapater C, Chauvigné F, Fernandez-Gomez B, Finn RN, Cerdà J. Alternative splicing of the nuclear progestin receptor in a perciform teleost generates novel mechanisms of dominant-negative transcriptional regulation. Gen Comp Endocrinol. 2013;182:24–40.
- 53.
Han MV, Zmasek CM. phyloXML: XML for evolutionary biology and comparative genomics. BMC Bioinformatics 2009;10.
- 54.
Hedges SB, Kumar S. The timetree of life. 2009; New York: Oxford University Press.
- 55. Zardoya R. Phylogeny and evolution of the major intrinsic protein family. Biol Cell. 2005;97:397–414.
- 56. Abascal F, Irisarri I, Zardoya R. Diversity and evolution of membrane intrinsic proteins. Biochim Biophys Acta-Gen Subj. 2014;1840:1468–1481.
- 57. Zapater C, Chauvigné F, Norberg B, Finn RN, Cerdà J. Dual Neofunctionalization of a Rapidly Evolving Aquaporin-1 Paralog Resulted in Constrained and Relaxed Traits Controlling Channel Function during Meiosis Resumption in Teleosts. Mol Biol Evol. 2011;28:3151–3169.
- 58. King LS, Kozono D, Agre P. From structure to disease: The evolving tale of aquaporin biology. Nat Rev Mol Cell Bio. 2004;5:687–698.
- 59. Fu DX, Libson A, Miercke LJW, Weitzman C, Nollert P, et al. Structure of a glycerol-conducting channel and the basis for its selectivity. Science. 2000;290:481–486.
- 60. de Groot BL, Grubmuller H. The dynamics and energetics of water permeation and proton exclusion in aquaporins. Curr Opin Struc Biol. 2005;15:176–183.
- 61. Froger A, Tallur B, Thomas D, Delamarche C. Prediction of functional residues in water channels and related proteins. Protein Sci. 1998;7:1458–1468.
- 62. Amores A, Force A, Yan YL, Joly L, Amemiya C, et al. Zebrafish hox clusters and vertebrate genome evolution. Science. 1998;282:1711–1714.
- 63. Jaillon O, Aury JM, Brunet F, Petit JL, Stange-Thomann N, et al. Genome duplication in the teleost fish Tetraodon nigroviridis reveals the early vertebrate proto-karyotype. Nature. 2004;431:946–957.
- 64. Amemiya CT, Powers TP, Prohaska SJ, Grimwood J, Schmutz J, et al. Complete HOX cluster characterization of the coelacanth provides further evidence for slow evolution of its genome. Proc Natl Acad Sci USA. 2010;107:3622–3627.
- 65.
Braasch I, Postlethwaite JH. Polyploidy in Fish and the Teleost Genome Duplication. In: Soltis PS, Soltis DEeditors. Polyploidy and Genome Evolution. Berlin: Springer-Verlag. 2012;pp. 341–383.
- 66. Garcia HE, Laudet V. Nuclear receptors are markers of animal genome evolution. J Struct Funct Genomics. 2003;3:177–184.
- 67. Ma TH, Yang BX, Umenishi F, Verkman AS. Closely spaced tandem arrangement of AQP2, AQP5, and AQP6 genes in a 27-kilobase segment at chromosome locus 12q13. Genomics. 1997;43:387–389.
- 68.
Sawyer SA. A computer package for the statistical detection of gene conversion. Distributed by the author Department of Mathematics, Washington University in St Louis Sawyer website. 1999. Available: http://www.math.wustl.edu/~sawyer.
- 69. Yasui M, Kwon TH, Knepper MA, Nielsen S, Agre P. Aquaporin-6: An intracellular vesicle water channel protein in renal epithelia. Proc Natl Acad Sci USA. 1999;96:5808–5813.
- 70. Beitz E, Liu K, Ikeda M, Guggino WB, Agre P, et al. Determinants of AQP6 trafficking to intracellular sites versus the plasma membrane in transfected mammalian cells. Biol Cell. 2006;98:101–109.
- 71. Pao GM, Wu LF, Johnson KD, Hofte H, Crispeels MJ, et al. Evolution of the mip family of integral membrane-transport proteins. Mol Microbiol. 1991;5:33–37.
- 72. Park JH, Saier MH. Phylogenetic characterization of the MIP family of transmembrane channel proteins. J Membr Biol. 1996;153:171–180.
- 73. Heymann JB, Engel A. Aquaporins: Phylogeny, structure, and physiology of water channels. News Physiol Sci. 1999;14:187–193.
- 74. Rivera MC, Lake JA. The ring of life provides evidence for a genome fusion origin of eukaryotes. Nature. 2004;431:152–155.
- 75. Lake JA, Sinsheimer JS. The Deep Roots of the Rings of Life. Genome Biol Evol. 2013;5:2440–2448.
- 76. Timmis JN, Ayliffe MA, Huang CY, Martin W. Endosymbiotic gene transfer: Organelle genomes forge eukaryotic chromosomes. Nat Rev Genet. 2004;5:123–U16.
- 77. Soto G, Alleva K, Amodeo G, Muschietti J, Ayub ND. New insight into the evolution of aquaporins from flowering plants and vertebrates: Orthologous identification and functional transfer is possible. Gene. 2012;503:165–176.
- 78. Perez DG, Juliana, Soto G, Alleva K, Jozefkowicz C, et al. Prediction of Aquaporin Function by Integrating Evolutionary and Functional Analyses. J Membr Biol. 2014;247:107–125.
- 79.
Blair JE. Animals (Metazoa). In: Hedges SB, Kumar Seditors. The Timetree of Life. Oxford: Oxford University Press. 2009;pp.220–230.
- 80. McLysaght A, Hokamp K, Wolfe KH. Extensive genomic duplication during early chordate evolution. Nature Genet. 2002;31:200–204.
- 81. Dehal P, Boore JL. Two rounds of whole genome duplication in the ancestral vertebrate. PLoS Biol. 2005;3:1700–1708.
- 82. Putnam NH, Butts T, Ferrier DEK, Furlong RF, Hellsten U, et al. The amphioxus genome and the evolution of the chordate karyotype. Nature. 2008;453:1064–10U3.
- 83. Amores A, Catchen J, Ferrara A, Fontenot Q, Postlethwait JH. Genome Evolution and Meiotic Maps by Massively Parallel DNA Sequencing: Spotted Gar, an Outgroup for the Teleost Genome Duplication. Genetics. 2011;188:799–U79.
- 84.
Drake LL, Boudko DY, Marinotti O, Carpenter VK, Dawe AL, et al.. The Aquaporin Gene Family of the Yellow Fever Mosquito, Aedes aegypti. PLoS One 5.2010.
- 85. Grohme MA, Mali B, Wełnicz W, Michel S. The aquaporin channel repertoire of the tardigrade Milnesium tardigradum. Bioinformatics Biol Insights. 2013;7:153–165.
- 86. Berthelot C, Brunet F, Chalopin D, Juanchich A, Bernard M, et al. The rainbow trout genome provides novel insights into evolution after whole-genome duplication in vertebrates. Nat Commun. 2014;5:3657.
- 87. Macqueen DJ, Johnston IA. A well-constrained estimate for the timing of the salmonid whole genome duplication reveals major decoupling from species diversification. Proc Biol Sci. 2014;281:20132881.
- 88. Kuraku S, Meyer A, Kuratani S. Timing of genome duplications relative to the origin of the vertebrates: Did Cyclostomes diverge before or after? Mol Biol Evol. 2009;26:47–59.
- 89. Smith JJ, Kuraku S, Holt C, Sauka-Spengler T, Jiang N, et al. Sequencing of the sea lamprey (Petromyzon marinus) genome provides insights into vertebrate evolution. Nature Genet. 2013;45:415–421.
- 90. Wolf YI, Koonin EV. Genome reduction as the dominant mode of evolution. Bioessays. 2013;35:829–837.
- 91. Kasahara M, Naruse K, Sasaki S, Nakatani Y, Qu W, et al. The medaka draft genome and insights into vertebrate genome evolution. Nature. 2007;447:714–719.
- 92. Muffato M, Crollius HR. Paleogenomics in vertebrates, or the recovery of lost genomes from the mist of time. Bioessays. 2008;30:122–134.
- 93. Nakatani Y, Takeda H, Kohara Y, Morishita S. Reconstruction of the vertebrate ancestral genome reveals dynamic genome reorganization in early vertebrates. Genome Res. 2007;17:1254–1265.
- 94. Smith JJ, Antonacci F, Eichler EE, Amemiya CT. Programmed loss of millions of base pairs from a vertebrate genome. Proc Natl Acad Sci USA. 2009;106:11212–11217.
- 95. Jones CD, Begun DJ. Parallel evolution of chimeric fusion genes. Proc Natl Acad Sci USA. 2005;102:11373–11378.
- 96. Maher CA, Kumar-Sinha C, Cao X, Kalyana-Sundaram S, Han B, et al. Transcriptome sequencing to detect gene fusions in cancer. Nature. 2009;458:97–U9.
- 97. Smith JJ, Baker C, Eichler EE, Amemiya CT. Genetic Consequences of Programmed Genome Rearrangement. Curr Biol. 2012;22:1524–1529.
- 98. Gustafsson OSE, Collin SP, Kroeger RHH. Early evolution of multifocal optics for well-focused colour vision in vertebrates. J Exp Biol. 2008;211:1559–1564.
- 99. Mathaes M, Burk O, Qiu H, Nusshag C, Goedtel-Armbrust U, et al. Evolutionary History and Functional Characterization of the Amphibian Xenosensor CAR. Mol Endocrinol. 2012;26:14–26.
- 100.
Hedges SB. Vertebrates (Vertebrata). In: Hedges SB, Kumar Seditors. The Timetree of Life. Oxford: Oxford University Press. 2009;pp.309–314.
- 101. Daeschler EB, Shubin NH, Jenkins FA. A Devonian tetrapod-like fish and the evolution of the tetrapod body plan. Nature. 2006;440:757–763.
- 102. Shubin NH, Daeschler EB, Jenkins FA. The pectoral fin of Tiktaalik roseae and the origin of the tetrapod limb. Nature. 2006;440:764–771.
- 103. Niedzwiedzki G, Szrek P, Narkiewicz K, Narkiewicz M, Ahlberg PE. Tetrapod trackways from the early Middle Devonian period of Poland. Nature. 2010;463:43–48.
- 104. Amemiya CT, Alfoeldi J, Lee AP, Fan S, Philippe H, et al. The African coelacanth genome provides insights into tetrapod evolution. Nature. 2013;496:311–316.
- 105. van Balkom BWM, Savelkoul PJM, Markovich D, Hofman E, Nielsen S, et al. The role of putative phosphorylation sites in the targeting and shuttling of the aquaporin-2 water channel. J Biol Chem. 2002;277:41473–41479.
- 106.
Grigg G, Gans C. Morphology & Physiology of the Crocodylia. editors. ‘Fauna of Australia, 2A: Amphibia and Reptilia’. Canberra: Australian Government Publishing Service. 1993;pp. 326–336.
- 107. Yasui M, Hazama A, Kwon TH, Nielsen S, Guggino WB, et al. Rapid gating and anion permeability of an intracellular aquaporin. Nature. 1999;402:184–187.
- 108. Ikeda M, Beitz E, Kozono D, Guggino WB, Agre P, et al. Characterization of aquaporin-6 as a nitrate channel in mammalian cells - Requirement of pore-lining residue threonine 63. J Biol Chem. 2002;277:39873–39879.
- 109. Cardoso KC, Da Silva MJ, Costa GGL, Torres TT, Del B, et al. A transcriptomic analysis of gene expression in the venom gland of the snake Bothrops alternatus (urutu). BMC Genomics. 2010;11:605.
- 110. Taplin LE, Grigg GC. Salt glands in the tongue of the estuarine crocodile Crocodylus porosus. Science. 1981;212:1045–1047.
- 111. Rehorek SJ, Smith TD. The primate Harderian gland: Does it really exist? Ann Anat. 2006;188:319–327.
- 112. Delporte C. Aquaporins in salivary glands and pancreas. Biochim Biophys Acta-Gen Subj. 2014;1840:1524–1532.
- 113. Payne AP. The Harderian gland - A Tercentennial review. J Anat. 1994;185:1–49.
- 114. Oksche A, Moller A, Dickson J, Rosendahl W, Rascher W, et al. Two novel mutations in the aquaporin-2 and the vasopressin V2 receptor genes in patients with congenital nephrogenic diabetes insipidus. Hum Genet. 1996;98:587–589.
- 115. VargasPoussou R, Forestier L, Dautzenberg MD, Niaudet P, Dechaux M, et al. Mutations in the vasopressin V2 receptor and aquaporin-2 genes in 12 families with congenital nephrogenic diabetes insipidus. J Am Soc Nephrol. 1997;8:1855–1862.
- 116. Boone M, Deen PMT. Physiology and pathophysiology of the vasopressin-regulated renal water reabsorption. Pflug Arch Eur J Phy. 2008;456:1005–1024.
- 117. Pang PKT, Griffith RW, Atz JW. Osmoregulation in Elasmobranchs. Am Zool. 1977;17:365–377.
- 118. Beamish FWH. Osmoregulation in Juvenile and Adult Lampreys. Can J Fish Aquat Sci. 1980;37:1739–1750.
- 119.
Evans DH. Osmotic and Ionic Regulation. In: Evans DHeditor. The Physiology of Fishes. Boca Raton: CRC Press. 1993;pp.315–341.
- 120. Colombo L, Dalla Valle L, Fiore C, Armanini D, Belvedere P. Aldosterone and the conquest of land. J Endocrinol Invest. 2006;29:373–379.
- 121. Engelund MB, Chauvigné F, Christensen BM, Finn RN, Cerdà J, et al. Differential expression and novel permeability properties of three aquaporin 8 paralogs from seawater-challenged Atlantic salmon smolts. J Exp Biol. 2013;216:3873–3885.