The genome of zebrafish (Danio rerio) encodes two unlinked genes equally closely related to the SLC4A2/AE2 anion exchanger genes of mammals. One of these is the recently reported zebrafish ae2 gene (Shmukler BE, Kurschat CE, Ackermann GE, Jiang L, Zhou Y, Barut B, Stuart-Tilley AK, Zhao J, Zon LI, Drummond IA, Vandorpe DH, Paw BH, Alper SL. Am J Physiol Renal Physiol Renal Physiol 289: F835–F849, 2005), now called ae2.1. We now report the structural and functional characterization of Ae2.2, the product of the second zebrafish Ae2 gene, ae2.2. The ae2.2 gene of zebrafish linkage group 24 encodes a polypeptide of 1,232 aa in length, sharing 70% amino acid identity with zebrafish Ae2.1 and 67% identity with mouse AE2a. Zebrafish Ae2.2 expressed in Xenopus oocytes encodes a 135-kDa polypeptide that mediates bidirectional, DIDS-sensitive Cl−/Cl− exchange and Cl−/HCO3− exchange. Ae2.2-mediated Cl−/Cl− exchange is cation independent, voltage insensitive, and electroneutral. Acute regulation of anion exchange mediated by Ae2.2 includes activation by NH4+ and independent inhibition by acidic intracellular pH and by acidic extracellular pH. In situ hybridization reveals low-level expression of Ae2.2 mRNA in zebrafish embryo, most notably in posterior tectum, eye, pharynx, epidermal cells, and axial vascular structures, without notable expression in the Ae2.1-expressing pronephric duct. Knockdown of Ae2.2 mRNA, of Ae2.1 mRNA, or of both with nontoxic or minimally toxic levels of N-morpholino oligomers produced no grossly detectable morphological phenotype, and preserved normal structure of the head and the pronephric duct at 24 h postfertilization.
- chloride/bicarbonate exchanger
- Xenopus oocyte
- isotopic flux
- in situ hybridization
- two-electrode voltage clamp
- N-morpholino oligomer
chloride bicarbonate exchangers of metazoan organisms contribute to regulation of intracellular pH (pHi), cell volume, and intracellular chloride concentrations ([Cl−]i) in most body cell types (17). In polarized epithelial cells, basolateral Cl−/HCO3− exchangers participate in epithelial secretion of acid or Cl−, whereas apical Cl−/HCO3− exchangers mediate epithelial secretion of HCO3− and/or Cl− absorption. Cl−/HCO3− exchangers are encoded by SLC4 and SLC26 gene families. We have previously described the Slc4a1/Ae1 (10) and Slc4a2/Ae2 gene products of zebrafish (14). Zebrafish slc4a1/ae1 was positionally cloned as the hematopoietic mutation retsina. Retsina/ae1 is expressed in the erythropoietic tissue of the head kidney, and its loss of function is associated with severe hemolytic, dyserythropoietic, spherocytic anemia associated with erythroid-specific defects in cytokinesis (10). Zebrafish Slc4a2/Ae2.1 was the most abundant of the cDNAs homologous to retsina/slc4a1/ae1 cloned by low-stringency hybridization screening from the same kidney cDNA library. Ae2.1 mRNA and protein are expressed predominantly in the proximal pronephric duct but not in the surrounding erythropoietic tissue of the head kidney (14).
The SLC4 gene family in mammals includes among the Na+-independent Cl−/HCO3− exchanger members at least three genes differing in tissue distribution and regulatory properties (11, 17). We therefore sought additional AE-related SLC4 genes in the zebrafish. We report here that the zebrafish genome encodes a second slc4a2 gene, ae2.2, and so rename the originally reported zebrafish ae2 gene as ae2.1. The ae2.2 gene, though located on a different chromosome, shares a common intron-exon genomic organization with the homologous ae2.1 gene . The Ae2.2 polypeptide encodes a Na+- and voltage-independent, electroneutral anion exchanger inhibited by pHi and extracellular pH (pHo), and is activated by NH4+. Ae2.2 mRNA expression appears absent from the pronephric duct or hematopoietic cells, but is expressed at low levels in brain, eye, and additional structures. N-morpholino oligo-mediated individual knockdown of the mRNAs encoding Ae2.2 or Ae2.1, or combined knockdown of both, produced no detectable gross alteration of early embryonic development, including apparent preservation of normal structure of the head and pronephric duct.
Molecular cloning of zebrafish ae2.2 cDNA.
An arrayed zebrafish phagemid cDNA library in pBK-CMV (Stratagene, La Jolla, CA) screened at low stringency with a 32P-labeled zebrafish slc4a1/ae1 cDNA (10) yielded 11 clones encoding ae2.1 (described in Ref. 14 as zAe2), and one clone (CG385) with terminal sequences identical to Zebrafish Information Network (ZFIN) clone CB330 (19). Alignment of the completely sequenced clone CG385 with AE2 sequences revealed three deletions unrelated to consensus intron-exon junctions, as well as an internal translocation of sequence from the 5′-coding region to the middle of the cDNA. Independent PCR amplification products from the same library were sequenced as pools to identify sequences bridging the deletions. 5′- and 3′-RACE products generated from the same cDNA library were sequenced as pools or as individual clones in pCRII (Invitrogen) to extend the cDNA sequence into 5′- and 3′-untranslated regions (utr). These procedures restored most of the ae2.1-like open reading frame, with the exception of a 27-codon region corresponding to aa 406–432 of the predicted Ae2.2 polypeptide.
We therefore generated cDNA (Retroscript; Ambion) from pooled 24-h postfertilization zebrafish embryo total RNA (RNeasy; Qiagen). From this cDNA, we amplified a cDNA encoding a complete Ae2.2 open reading frame using as forward, oligonucleotide 5′-TCTGCAGTGAGCACTTTGAGG-3′ from the 5′-utr and as reverse oligonucleotide, 5′-GCCAATAGTGTGATTGATAG-3′ from the 3′-utr (Hot start with 36 cycles of 94°C denaturing × 45 s/60°C annealing × 2 min/72°C extension × 6 min). No deletion/translocation products resembling library cDNA CG385/CB330 were subsequently identified. Occasional but reproducible nucleotide ambiguities in the full-length sequence from pooled embryo PCR products (see Supplemental Fig. 1A; the online version of this article contains supplemental figures and tables) suggested common polymorphisms in the Ae2.2 cDNA. This impression was confirmed by complete sequencing of individual cDNA subclones in pCRII from multiple, independent PCR amplifications, to exclude artifactual PCR-induced mutations. Two ae2.2 alleles were identified. Allele 1 sequence (see Supplemental Fig. 1) was assembled from plasmid 15-3 (GenBank accession no. DQ286970) and from 3′-utr sequence from the extensively overlapping clone CG385. Intact plasmid 15-3 was used for further study. Plasmid 15–18 (GenBank accession no. DQ287972) encoded the second allele of ae2.2. DNA sequence analyses were carried out with the GCG suite of programs (Univ. of Wisconsin Genetics Computing Group, Madison, WI).
After completion of this work, the ae2.2 gene was annotated in the Danio rerio genome Zv6 release as genomic clone NC_007135. However, the predicted encoded Ae2.2 mRNA XM_694225 and Ae2.2 polypeptide XP_699317 are currently artifactual sequences, each marked by an ectopic internal duplication and an internal deletion.
Radiation hybrid mapping of the zebrafish slc4a2.2/ae2.2 gene.
The genomic location of the slc4a2.2/ae2.2 gene was determined with the Goodfellow T51 radiation hybrid (RH) panel (5, 8). A specific Ae2.2 cDNA probe without similarity to sequences then in the database was amplified from the 3′-utr with forward primer 5′- CCGTGAGAACAACTGAATGT (nt 4102–4121) and reverse primer 5′- TTCATGATTTGATGGCAACA (nt 4356–4337), as previously described (20). Following initial denaturation at 94°C for 4 min, PCR included 35 cycles of 94°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 1 min, with a final 72°C extension for 10 min. Raw RH mapping scores (20) were analyzed with Instant RH mapping software (http://afrhmaps.tch.harvard.edu/ZonRHmapper/instantMapping.htm) based on the T51 RH panel map calculated using SAMapper 1.0 (18, 20).
In situ hybridization.
An Ae2.2 in situ hybridization probe was synthesized using the PCR-amplified 5′-RACE clone 12 in pCRII that contains 183 nt of 5′-utr plus the 5′-most 735 nt of Ae2.2 coding sequence, a region absent from zebrafish Ae1 cDNA and with very low similarity to vertebrate Ae2 and Ae3 genes. The probe template was linearized with SpeI for transcription from the T7 promoter to generate antisense cRNA probe, or linearized with XhoI for transcription from the SP6 promoter to generate a sense cRNA probe. Probes were labeled with digoxigenin-UTP using the DIG RNA Labeling Mix (Roche) and purified with G-50 Sephadex Quick Spin Columns (Roche). In situ hybridization was carried out as described previously (14).
cRNA expression in Xenopus oocytes.
cDNA encoding zebrafish Ae2.2 (plasmid 15-3) was subcloned into the Xenopus oocyte expression vector pXT7, linearized with XbaI, and used as template for synthesis of capped cRNA with T7 polymerase. Plasmid encoding zebrafish ae2.1 (plasmid 254–130) was linearized with XbaI, and transcribed with T3 polymerase as previously described (MEGAscript; Ambion, Austin, TX) (14). cRNAs were purified with the RNeasy kit (Qiagen), and cRNA integrity was verified by formamide agarose gel electrophoresis. Female X. laevis anesthetized with 0.17% tricaine were subjected to partial ovariectomy in accordance with protocols approved by the Institutional Animal Care and Use Committee of Beth Israel Deaconess Medical Center. Excised, minced ovarian segments were incubated 1 h with gentle shaking at room temperature in 2 mg/ml type A collagenase (Roche) in ND-96, containing (in mM) 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, 5 HEPES, and 2.5 Na pyruvate, pH 7.4, supplemented with 5 mg/100 ml gentamicin. Stage V-VI oocytes were washed, manually defolliculated, and maintained at 19°C in ND-96/pyruvate/gentamicin. On the same or next day oocytes were manually microinjected (Drummond) with 50 nl water or solution containing 3 or 10 ng cRNA as indicated. Oocytes were maintained 2–5 days in ND-96-gentamicin at 19°C with daily medium change until use for experiments.
36Cl− flux assays in X. laevis oocytes.
36Cl− influx assays were performed as 30-min uptakes in ND-96 lacking pyruvate and gentamicin and containing 10 μM bumetanide. Total bath [Cl−] was 103 mM. For 36Cl− efflux assays, individual oocytes were injected with 50 nl of 260 mM Na36Cl (10–20,000 cpm; ICN Biochemicals). Following a 5- to 10-min recovery period in Cl−-free ND-96 containing (in mM) 96 Na isethionate, 2 K gluconate, 1.8 Ca gluconate, 1 Mg gluconate, and 5 HEPES, pH 7.4, the efflux assay was initiated by transfer of single oocytes into borosilicate tubes containing efflux solution. At defined intervals, 0.95 ml of efflux solution was removed for scintillation counting and replaced with an equal volume of fresh solution. Following completion of the assay with a final efflux period in Cl−-free medium or in the presence of the inhibitor DIDS to confirm oocyte integrity, each oocyte was lysed in 100 μl 1% SDS. Samples were counted for ≥5 min, such that the magnitude of two SD was <5% of the sample mean. In some efflux experiments, ND-96 Na+ was completely replaced by K+ (KD96) or by N-methyl-d-glucamine, or Cl− was completely replaced by 72 mM gluconate plus 24 mM HCO3− in the presence of 5% CO2. In other experiments, 20 mM NaCl was replaced by equimolar NH4Cl.
Two-electrode voltage clamp of X. laevis oocytes.
Two-electrode voltage clamp was performed as previously described, with modifications (2, 14). Oocytes injected 2–3 days previously with water or with 10 ng Ae2.2 cRNA were mounted in a perfusion chamber on the stage of an Olympus IMT-2 microscope and subjected to 800-ms voltage clamp steps from −100 mV to +40 mV in 20-mV increments in ND-96 bath.
Immunofluorescence detection of zebrafish ae2.2 polypeptide in transiently transfected HEK-293 cells.
Zebrafish Ae2.2 cDNA (plasmid 15-3) subcloned into pcDNA3 was transiently transfected into subconfluent HEK-293 cells on glass coverslips using lipofectamine (Invitrogen). After 48 to 72 h, cells were fixed with 2% paraformaldehyde, quenched, and immunostained with affinity-purified anti-mouse AE2 aa 1224–1237 (1:2,000) in the presence of irrelevant peptide or of specific peptide antigen at 24 μg/ml. After being washed, the coverslips were incubated with fluorophor-conjugated secondary anti-rabbit Ig (Jackson Immunochemicals, Westport, PA) and mounted. Images were recorded with a Bio-Rad MRC-1024 laser-scanning confocal fluorescence microscope.
Antisense N-morpholino-oligo knockdown studies.
N-morpholino-oligos (MO) of a design verified with GeneTools software were synthesized by GeneTools (Philomath, OR; see Supplemental Table 3). Ae2.2 oligos MO-1 (spanning the intron 14-exon 15 junction of the ae2.2 gene) and MO-2 (spanning the exon 15-intron 15 junction of the ae2.2 gene) were designed to interrupt the Ae2.2 open reading frame by deletion of exon 15 from Ae2.2 mRNA. Ae2.1 oligos MO-5 (spanning the intron 5-exon 6 junction) and MO-6 (spanning the exon 6-intron 6 junction) were designed to interrupt the Ae2.1 open reading frame by deletion of exon 6 from Ae2.1 mRNA. One-cell embryos were injected with 1 nl volumes containing pairs of MOs at individual concentrations of 0.2 to 1.5 mM (total injected load 3.3–25 ng). Knockdown efficacy was estimated from the relative abundance of RT-PCR products generated from wild-type mRNA and from the intended knockdown mRNA product, as detected by agarose gel electrophoresis. The oligonucleotide sequences used for these diagnostic RT-PCR amplifications are presented in Supplemental Table 3. MO-3 and MO-4 were found to be ineffective as reagents for Ae2.1 knockdown.
Among the 14 cDNA clones selected from the zebrafish adult kidney cDNA library by low-stringency screen with an Slc4a1/Ae1 probe, only three of the isolated cDNAs did not encode Ae2.1 (14). One of these three (the Ae2.1-related CG385) had terminal sequence identical to then unannotated shotgun clone CB330 (ZDB-GENE-030429-14), represented by GenBank partial cDNA sequences CB417274 (5′-end) and CB417275 (3′-end). Complete sequencing of CG385 revealed three internal deletions and one internal rearrangement. After Ae2.2 sequences were extended at both ends by 5′- and 3′ RACE, cDNAs encoding a putative full-length Ae2.2 open reading frame were amplified from zebrafish whole embryo cDNA prepared from pooled embryos.
Two allelic full-length coding sequences of Ae2.2 cDNA were determined. Supplemental Fig. 1 indicates sites at which sequences of alleles 1 and 2 differ. The two nonsynonymous coding polymorphisms (so named because they alter amino acid sequence) and 19 additional synonymous polymorphisms (leaving amino acid sequence unaltered) that distinguish the two alleles are summarized in Supplemental Table 1. Three additional nonsynonymous coding polymorphisms and eight more synonymous polymorphisms contributing to neither allele 1 nor allele 2 were also found in two or more Ae2.2 cDNA clones isolated in early stages of the study from independent pools of embryos (Supplemental Table 1). Thus, the ae2.2 gene is highly polymorphic.
Mapping of the zebrafish ae2.2 gene.
The ae2.2 gene was assigned to linkage group 24 (LG24) by RH mapping on the Goodfellow TH51 RH panel. The LG24 markers unp191, fj04c06.x1, and zfish44908-1385h06 at 29 cR exhibited linkage to the ae2.2 gene with logarithmic odds (LOD) scores of 7.5. LG24 markers z4921 at 31 cR, fb61h03.x1 at 35 cR, and z24003 were linked to ae2.2 with respective LOD scores of 6.7, 5.9, and 5.1. The RH localization of the ae2.2 gene to LG24 was confirmed by physical linkage of LG24 marker fi03f09.x1 to a portion of the ae2.2 gene in fosmid CH1073-406K10 from the double-haploid library (M. Caccamo, Wellcome Trust Sanger Institute, personal communication from Zv7 genomic assembly). Both ae2.2 on LG24 and ae2.1 on LG2 are syntenic with the human SLC4A2 locus at chromosome 7q36 (Table 1). Although most of the surrounding linked zebrafish genes have a duplicated homolog elsewhere in the genome, the homologous cul1b and cul1a genes of zebrafish have each remained linked to ae2.2 and to ae2.1, respectively (Table 1). None of the human genes adjacent to the human AE3 locus on chromosome 2q35 are syntenic with the zebrafish ae2.1 or ae2.2 genes.
Exon-intron organization of the zebrafish ae2.2 gene.
The exon-intron organization of the zebrafish ae2.2 gene (Table 2) was revealed by nucleotide sequence alignment of the ae2.2 cDNA sequence with ae2.2 gene fragments (Supplemental Fig. 1). Twelve of the 23 ae2.2 exons are identical in length to the corresponding exons of the ae2.1 gene. All 23 introns of the ae2.2 gene are bound by consensus donor and acceptor splice sites, as is also true for zebrafish ae2.1 and mouse Ae2 genes. The 22 exon-exon junctions of the ae2.2 gene include 20 with locations precisely aligned with those of the zebrafish ae2.1 gene. Sixteen of the 21 exon-exon junctions within the ae2.2 gene protein coding region exhibit amino acid sequences across the splice junctions that are conserved in the Ae2.1 polypeptide (14). The exon 13-exon 14 junction in all three genes falls after codon position +1, but the ae2.2 gene has a single codon insertion (E654) at that site (Fig. 1). The exon 10-exon 11 junction is located identically in the zebrafish ae2.2 and 2.1 genes, but is shifted in the mouse gene. The exon 5-exon 6 junction in all three genes falls after coding position +2, despite the low degree of amino acid sequence conservation in a wide region surrounding this splice junction.
Deduced amino acid sequence of the zebrafish ae2.2 polypeptide.
The 1232 aa zebrafish Ae2.2 polypeptide is 70% identical to its homologous zebrafish Ae2.1 polypeptide (Fig. 1). Kyte-Doolittle hydropathy analysis predicts for Ae2.2 a hydrophilic NH2-terminal cytoplasmic domain of ∼701 aa, followed by a hydrophobic polytopic transmembrane (TM) domain of ∼494 aa comprising 12–14 predominantly α-helical TM spans and a COOH-terminal hydrophilic cytoplasmic domain of ∼37 aa. Three consensus N-glycosylation sites are located in the nominally extracellular loop between putative TM5 and TM6. The putative covalent binding site for stilbene disulfonate inhibitors of anion transport is also conserved at the extracellular end of TM5 at K838 of Ae2.2. The putative electrostatic binding site for carbonic anhydrase 2 (aa 1207–11) is identical to that in human AE1 and mouse Ae1. Pairwise alignments of the Ae2.2 aa sequence with those of other SLC4/AE family members (Supplemental Table 3) revealed greater identity with zebrafish Ae2.1 and with AE2 of other species than with AE1 or AE3 polypeptides. This held true also in comparisons of individual subdomains among the orthologous polypeptide sequences.
Zebrafish Ae2.2 is a cation-independent, electroneutral Cl−/anion exchanger.
Ae2.2 polypeptide was abundantly expressed in Xenopus oocytes (Fig. 2A), and its expression was associated with concentration-dependent 36Cl− influx at rates comparable to those exhibited by zebrafish Ae2.1 (Fig. 2B). The Cl− transport was confirmed as Cl−/Cl− exchange by demonstration of trans-Cl−-dependence of 36Cl− efflux (Fig. 2, C and D). Ae2.2 also mediated Cl−/HCO3− exchange (Fig. 3). Both Cl−/Cl− and Cl−/HCO3− exchanges were sensitive to the stilbene disulfonate inhibitor, DIDS (Figs. 2, C and D and 3). zAe2.2-mediated Cl−/Cl− exchange was cation-independent, insofar as bath Na+ substitution with N-methyl-d-glucamine (Fig. 4A) or with K+ (Fig. 4B) diminished neither 36Cl− efflux rate constants nor sensitivity to inhibition by DIDS.
Zebrafish Ae2.2-mediated anion exchange is electroneutral.
The lack of effect of bath K+ substitution on Ae2.2-mediated 36Cl− efflux suggested membrane potential-independence. Therefore, Ae2.2-expressing oocytes were subjected to two-electrode voltage clamp. Fig. 5 shows that resting currents and reversal potential were no different than in water-injected oocytes, that bath gluconate substitution decreased rather than increased inward current, and that subsequent addition of DIDS was without effect. Thus, zAe2.2-mediated anion exchange was electroneutral and was not accompanied by increased anion conductance of the type exhibited by trout Ae1 (3).
Zebrafish Ae2.2 is subject to acute regulation.
Mouse Ae2 and zebrafish Ae2.1 are acutely and independently regulated by intracellular and pHo and by NH4+ (14). We therefore examined zebrafish Ae2.2 for similar types of acute regulation. Fig. 6 shows that the rate of 36Cl−/Cl− exchange by Ae2.2 is nearly doubled by bath addition of NH4+. As shown in Fig. 7, A and B, removal of butyrate following a period of butyrate-induced intracellular acidification accelerates the rate of 36Cl−/Cl− exchange by Ae2.2 nearly threefold. However, the relative inhibition of Ae2.2 by 40 mM butyrate may not be so great as for Ae2.1 (Fig. 7 and Ref. 14). Fig. 7C shows that acidic pHo inhibits Ae2.2 with a pHo(50) value of 7.33 ± 0.16 (n = 10). Although this value is slightly alkaline-shifted compared with that of Ae2.1 (6.94 ± 0.19, n = 9; Ref. 14), the shift is not statistically significant (P = 0.14).
Localization of ae2.2 mRNA expression.
Recombinant Ae2.2 polypeptide and Ae2.1 polypeptide were detected equally well by antibody to the mouse AE2 COOH-terminal peptide (Fig. 7D), such that this antibody could not be used for unambiguous localization of Ae2.2 polypeptide. Therefore, Ae2.2 mRNA was localized in embryos by whole mount in situ hybridization (Fig. 8). Staining was faint at the 5, 10, and 17 somite stages (not shown). At 24 h postfertilization, low-abundance mRNA was evident in eye (likely retina) and tectum, and in a faint axial pattern suggestive of axial vasculature. The observed pattern reproduced that reported for in situ hybridization with a probe generated from the multiple deletion clone CB330 (19). Ae2.2 expression was not detected in pronephric duct, the major site of Ae2.1 expression. Ae2.2 localization was not altered in hematopoiesis-defective cloche mutants at the same developmental stages (not shown).
As shown in Fig. 7D by immunofluorescence in transiently transfected HEK-293 cells, and in Fig. 2A by immunoblot in oocytes, Ae2.2 and Ae2.1 are both detected by antibody to the highly conserved mouse AE2a COOH-terminal aa 1124–1237. Similar cross-reactivity was noted for antibody-to-mouse AE2a aa 330–352 (not shown). Therefore, neither antibody reagent is appropriate for immunolocalization of Ae2.2 polypeptide expression in zebrafish.
N-morpholino oligo knockdown of Ae2.2 and of Ae2.1 mRNAs.
Injection of pairs of MOs at concentrations of 1.5 mM, 0.75 mM, and 0.4 mM each led to dose-dependent hemorrhagic necrosis in brain and eye accompanied by developmental delay, constriction of yolk sac extension, and retarded axial extension with enhanced body curvature (not shown). Injection of individual MO pairs at the concentration of 0.2 mM did not produce grossly evident toxicity (Supplemental Fig. 2A). At these concentrations, knockdown of Ae2.1 mRNA was nearly complete, as evidenced by generation of the intended exon 6 deletion product leading to termination of the open reading frame (Supplemental Fig. 2C). Knockdown of Ae2.2 mRNA was also of high efficiency. However, in addition to the intended exon 15 deletion product terminating after Ae2.2 TM span 1 (Supplemental Fig. 2B), the MO-1/MO-2 pair also generated two alternate deletion products with in-frame deletions of 24 or 48 nt at the 3′ end of exon 15. As documented by DNA sequencing, these deletion products encoded Ae2.2 polypeptides lacking 8 (aa 761–768) or 16 amino acids (aa 761–776) from putative TM span 3 of Ae2.2, both predicted to be nonfunctional and/or unstable.
Knockdown of either Ae2.2 or Ae2.1 using MO concentrations of 0.4 mM or greater led to dose-dependent hemorrhagic necrosis. Individual knockdown of Ae2.2 or Ae2.1 using MO concentrations of 0.2 mM produced little if any necrosis, and produced mild developmental delay without evident specific developmental phenotype (Fig. 9). However, even with MO concentrations > 0.2 mM, those embryos without evidence of necrosis developed normally. Coinjection of both MO pairs at 0.2 mM per MO was accompanied in some embryos by mild toxicity (occasional hemorrhagic necrosis) and by an enhanced, nonspecific developmental delay phenotype (Supplemental Fig. 2A). These surviving morphants also progressed to develop an ultimately normal phenotype (Fig. 9).
Low stringency screening for Ae1 homologs of the zebrafish uncovered two zebrafish Ae2 genes, ae2.1 (14) and ae2.2 (present study). The two genes are each syntenic with human AE2, exhibit highly conserved intron-exon structures and encode multiple polymorphisms. The ae2.1 and ae2.2 genes encode polypeptides that share 70% amino acid identity, and each mediates electroneutral, cation-independent, anion exchange with similar properties of acute regulation. Whereas Ae2.1 is expressed at moderate levels in embryonic development and is predominantly localized at later stages in the anterior pronephric duct, Ae2.2 is expressed at low levels in head structures and in a diffuse pattern suggestive of axial vasculature. MO knockdown of Ae2.2 and Ae2.1 or both resulted in no apparent specific defect in early development. MOs at higher concentrations produced a nonspecific phenotype of developmental retardation, and inhibition of axial extension with curvature.
Ae2 gene duplication in zebrafish.
The highly conserved genomic structure and the conserved sequence encoding a conserved domain arrangement with conserved motifs all suggest that the ae2.2 and ae2.1 genes arose by gene duplication, a remnant of the whole genome duplication in ray-finned fishes believed to have occurred between 250M and 400M years ago (13). The 70% amino acid sequence identity is consistent with this remote gene duplication. Among the three of 21 exon-exon junctions of the ae2.2 gene coding region, which do not share superimposable locations with the zebrafish ae2.1 gene and the mouse Ae2 gene, the exon 10–11 junction is identical in the ae2.2 and ae2.1 genes, but differs in the mouse Ae2 gene, consistent with the occurrence of this gene duplication after the evolutionary divergence of fish from other vertebrates (13). The exon 5–6 junction which differs between the ae2.2 and ae2.1 genes is the site of alternative promoter usage in the mouse gene generating the functionally distinct Ae2c1 polypeptide (7). The ae2.1 exon 13–14 junction is conserved in mouse Ae2 gene but not in zebrafish ae2.2. This junction lies adjacent to a predicted proteolytic cleavage site, which, in the homologous Ae1 genes, defines the topological boundary separating most of the NH2-terminal cytoplasmic domain from the TM domain.
The hypothesis of Ae2 gene duplication is substantiated by shared synteny of both zebrafish Ae2 genes, ae2.1 on LG2 and ae2.2 on LG24, with apparently nonoverlapping regions of human chromosome 7q36 closely adjacent to the human AE2 gene. Such dual LG24 and LG2 synteny with human 7q36 is shared by only one linked zebrafish gene pair shown in Table 1, cul1b and cul1a. Additional genes pairs preserve syntheny for only one of the duplicated genes. Thus abcf2, smarcd3, and eng2b reside on LG2, but their duplicated homologs are found en bloc on LG7, suggestive of a postduplication intrachromosomal recombination. The duplicated homolog of LG2 gene shhb, shha, is found on LG15. Additional examples of preserved gene duplications in the zebrafish genome include insulin-like growth factor 1 receptors (12), voltage-gated Na+ channels (9), and two of the 29 muscular dystrophy gene homologs (15). Genes encoding cKit receptors and cKit ligands are both duplicated, and have evolved paralogous receptor-ligand specificity (6).
However, many neighboring genes of human chromosome 7 immediately adjacent to the AE2 gene have only one known homolog in the zebrafish genome, a situation reflecting most of the zebrafish genome. This is consistent with zebrafish diploidy, and reflects the evolutionary loss of one of the duplicated genes (13). Fish genes preserved in duplicated form are hypothesized to have been maintained under selective pressure. Most fish gene coding regions have evolved faster than their mammalian orthologs (1), as perhaps reflected in the high degree of polymorphism detected in the ae2.2 gene (Supplemental Table 1; note, however, that 10 nonsynonymous and 15 synonymous cSNPs have been reported in the human AE2 gene as of Sept. 2007). Many fish gene pairs also exhibit asymmetrically accelerated evolution of the paralogs (1). However, this generalization appears not to apply to zebrafish ae2.2 and ae2.1 coding regions, which are equidistant in sequence similarity from the most closely related mammalian SLC4 polypeptides. Thus, the preservation of two ae2 genes evolving at apparently equivalent rates supports the hypothesis that both gene products play important functions in the zebrafish, despite the absence of identifiable specific phenotype arising from mRNA knockdown during early development.
Functional characterization of ae2.2 polypeptide.
Maintenance of two related genes through evolution might reflect acquisition of distinct functional activities or distinct patterns of acute regulation. However, the Ae2.2 polypeptide did not differ from the Ae2.1 polypeptide in any tested functional index. Ae2.2 mediated Cl−/Cl− and Cl−/HCO3− exchange that was DIDS-sensitive, cation-independent, and electroneutral, all properties shared with Ae2.1. Ae2.2-mediated Cl−/Cl− exchange was also acutely activated by NH4+, acutely inhibited by acidic pHi, and acutely inhibited by acidic pHo, again properties in common with Ae2.1. The pHo(50) value of Ae2.2 was not significantly different from that of Ae2.1. Thus, as judged by the assays tested to date, distinct anion transport function does not explain preservation of the duplicated Ae2 genes in evolution.
It is curious that three of four cSNPs of the NH2-terminal cytoplasmic domain of Ae2.1 (P147R, I149T, E152D) reside within a region, which, when mutated in mouse Ae2a or when absent from mouse Ae2c1, lead to an alkaline-shifted pHo(50) (7). The single TM domain cSNP (V882M) is located at the COOH-terminal end of the ecto-loop connecting TM5 and TM6 (“Z loop”) immediately adjacent to two residues, which, when mutated in mouse Ae2, acid-shifts or alkali-shifts pHo(50), respectively, (16) (only the latter is conserved in Ae2.1 as K884).
Localization of Ae2.2 mRNA in zebrafish embryo.
Ae2.2 and Ae2.1 mRNAs do differ in localization pattern of expression. In contrast to the moderately high level of Ae2.1 expression in pronephric duct (14), Ae2.2 is absent from pronephric duct and expressed at low level in eye (likely retina) and tectum, and in axial vasculature. Ae2.2 is expressed at lower levels than Ae2.1 during early somite stages of development. The low expression level is evident also in the proportional representation of cDNAs encoding Ae2.1 (51 hits, most from kidney, some from testis, few from heart) and Ae2.2 (4 hits, all from embryo) in the Expressed Sequence Tag (EST) database (Sept 15, 2007).
Role of the two ae2 genes in embryonic development.
Preservation of both ae2 genes through evolution suggests a selective advantage or selective pressure at work. Distinct sites of expression constitute the best evidence for nonoverlapping functions of Ae2.2 and Ae2.1. However, these nonoverlapping expression patterns do not reveal whether in each expression site Ae2 function is essential or redundant.
Combined knockdown of both Ae2.2 and Ae2.1 (at effective MO concentrations, which, as individual MO pairs, were apparently nontoxic) produced minimal generalized developmental delay and minimally impaired axial elongation (Fig. 9 and Supplemental Fig. 2A). In all surviving embryos (the large majority of those injected) this developmental delay did not lead to any abnormality grossly evident at later developmental stages.
Perspectives and Significance
We conclude that neither Ae2.1 nor Ae2.2 are essential for normal early embryonic development. This conclusion is consistent with the developmental role of the single mouse Ae2 gene, whose knockout allows early embryonic development and birth, but is manifest as peri- and postnatal growth retardation and periweaning death (4). Ae2.2 may contribute to regulation of cell pH, cell volume, and cell [Cl−] in embryonic head structures, as Ae2.1 likely does in the anterior segment of the pronephric duct. However, the knockdown results suggest that the zebrafish genome expresses redundant functions that compensate for loss of Ae2.2 and Ae2.1 expression during early development. These redundant functions may be encoded by additional SLC4 genes, by SLC26 genes, or by yet other genes. Knockdown embryos might exhibit a phenotype under imposed stress conditions indicating conditional requirement for one or both Ae2 polypeptides. Ae2.2 and/or Ae2.1 might alternately play more important physiological roles in the more mature or adult fish at times beyond the efficacy of MO-mediated mRNA knockdown.
This work was supported by National Institutes of Health Grants DK-43495 (to S. L. Alper), HL-32262 and DK-70838 (to B. H. Paw), and DK-07199 (Beth Israel Deaconess Renal Training Grant to J. S. Clark), by March of Dimes Foundation Grants 5-FY04-21 and 1-FY06-365, the William Randolph Hearst Foundation (to B. H. Paw), and by fellowships from the National Kidney Foundation (to C. E. Kurschat) and from Mt. Holyoke College (to A. Hsu).
We thank Gabriele E. Ackermann and Alan K. Stuart-Tilley for discussion and assistance.
↵* B. E. Shmukler, J. S. Clark, and A. Hsu contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society