Expression of ion transport genes in ionocytes isolated from larval zebrafish (Danio rerio) exposed to acidic or Na⁺-deficient water

Abstract

In zebrafish (Danio rerio), a specific ionocyte subtype, the H⁺-ATPase-rich (HR) cell, is presumed to be a significant site of transepithelial Na⁺ uptake/acid secretion. During acclimation to environments differing in ionic composition or pH, ionic and acid–base regulations are achieved by adjustments to the activity level of HR cell ion transport proteins. In previous studies, the quantitative assessment of mRNA levels for genes involved in ionic and acid–base regulations relied on measurements using homogenates derived from the whole body (larvae) or the gill (adult). Such studies cannot distinguish whether any differences in gene expression arise from adjustments of ionocyte subtype numbers or transcriptional regulation specifically within individual ionocytes. The goal of the present study was to use fluorescence-activated cell sorting to separate the HR cells from other cellular subpopulations to facilitate the measurement of gene expression of HR cell-specific transporters and enzymes from larvae exposed to low pH (pH 4.0) or low Na⁺ (5 μM) conditions. The data demonstrate that treatment of larvae with acidic water for 4 days postfertilization caused cell-specific increases in H⁺-ATPase (atp6v1aa), ca17a, ca15a, nhe3b, and rhcgb mRNA in addition to increases in mRNA linked to cell proliferation. In fish exposed to low Na⁺, expression of nhe3b and rhcgb was increased owing to HR cell-specific regulation and elevated numbers of HR cells. Thus, the results of this study demonstrate that acclimation to low pH or low Na⁺ environmental conditions is facilitated by HR cell-specific transcriptional control and by HR cell proliferation.

INTRODUCTION

In freshwater (FW) fish, a large outwardly directed ionic gradient leads to the continuous diffusive loss of salts to the environment largely across the gills of adults or the skin of larvae. FW teleosts achieve ionic homeostasis by minimizing passive salt efflux and by actively absorbing ions from the environment. Beginning with the pioneering work of Krogh (25), the molecular mechanisms underlying ionic regulation in FW fish have been investigated intensely for the past 80 years; the results of which are summarized in several comprehensive reviews (7, 9, 17, 19–21, 28, 30).

Ion uptake in zebrafish larvae is dependent on salt-transporting mitochondrion-rich cells termed ionocytes, which express specific ion transporters responsible for the movement of ions (9, 18, 19). Five types of ionocytes have been identified in zebrafish larvae thus far [for review see (14)], namely, H⁺-ATPase-rich (HR) cells that are involved in H⁺ extrusion and Na⁺ uptake (48), Na⁺-K⁺-ATPase-rich (NaR) cells that are enriched with Na⁺-K⁺-ATPase and involved in transepithelial Ca²⁺ transport (33), Na⁺-Cl⁻ cotransporting (NCC) cells that express the Na⁺-Cl⁻ cotransporter and are involved in both Na⁺ and Cl⁻ uptake (44), SLC26 cells that are involved in Cl⁻ uptake (2), and K⁺-secreting (KS) cells that express inwardly rectifying K⁺ channels (Kir1.1) and are thought to be involved in K⁺ homeostasis (1).

The HR cell, first identified by Lin et al. (35), arguably is the most extensively studied and well-characterized zebrafish ionocyte, for which detailed models have been developed to explain the cellular pathways of Na⁺ uptake and acid extrusion (14). Based on data from immunohistochemistry (IHC) or in situ hybridization, HR cells are known to express H⁺-ATPase (HA), sodium proton exchanger isoform Nhe3b (48), ammonia conducting rhesus C glycoprotein b (Rhcgb, previously known as Rhcg1) (39), carbonic anhydrase (CA) 15a [Ca15a (36)], and cytosolic Ca17a [previously termed CAc (11) or Ca2-like a (36)]. Current models propose that the HR cell is an important site of ammonia secretion via apical membrane Rhcgb (39, 43) and Na⁺ uptake via Nhe3b (10, 42). Additionally, Na⁺ uptake may occur via an unidentified Na⁺ channel [potentially an acid-sensing ion channel (ASIC) coupled to HA (8)]. In either case, Na⁺ uptake is intricately linked to H⁺ secretion. The results of these studies, when taken together, clearly demonstrate redundancy in the functions of the HR cell with respect to Na⁺ uptake and H⁺ excretion, which may contribute to the ability of zebrafish to tolerate a wide range of environmental conditions including acidic water and low environmental Na⁺ levels.

Several studies have evaluated the role of the HR cell in regulating body fluid composition in zebrafish exposed to acidic or low Na⁺ water, with most of them relying on whole-body mRNA measurements of ion transport-related genes and/or numbers of HR cells present on the skin or gills [see Supplemental Table S1 (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.12249284) for a summary of studies presenting mRNA changes in zebrafish exposed to acidic or low Na⁺ water]. These types of studies, while informative, are limited because they cannot distinguish whether the differences in gene expression are the result of an adjustment of ionocyte subtype numbers or transcriptional regulation within individual ionocytes. Additionally, important biological events may be masked by these traditional analytical methods owing to low signal-to-noise ratios [i.e., genes may be expressed (or regulated) in relatively few numbers of a specific ionocyte subtype compared with the vast numbers of cells from other populations], which may or may not contain the transcript of interest. Although the HR cells comprise a small percentage of cells harvested from whole larvae, they play a major role (as described earlier) in whole-body homeostasis. Therefore, studies performed on unique populations of HR cells might allow the detection of gene transcript changes that otherwise might be obscured by high background levels in other cell types.

We are unaware of any data on ionocyte-specific gene expression responses in zebrafish exposed to varying environments including acidic or Na⁺-deficient water. Thus, this study describes a technique to separate the HR cells from other cellular subpopulations. The technique, once validated, was used to measure the gene expression of several HR cell-specific transporters and enzymes in larvae exposed to low pH or low Na⁺ conditions.

MATERIALS AND METHODS

Zebrafish

Adult zebrafish (Danio rerio, Hamilton-Buchanan 1822) were bred and raised in the University of Ottawa Aquatic Care Facility, where they were maintained in plastic tanks supplied with aerated, dechloraminated City of Ottawa tap water at 28°C. Fish were subjected to a constant 14:10-h (light–dark) photoperiod and fed daily until satiation with No. 1 crumble Zeigler (Aquatic Habitats, Apopka, FL). Embryos were collected following standard protocols (46). Briefly, male and females (1 male and 2 females) were placed in breeding tanks, and embryos collected the next day were reared in 50-mL petri dishes supplemented with dechloraminated City of Ottawa tap water with 0.05% ethylene blue, unless otherwise stated. The petri dishes were kept in incubators set at 28.5°C. Dead embryos were removed, and water was changed daily. As all experiments were performed on fish at 4 days postfertilization (dpf), they were not fed. The experiments were conducted in compliance with guidelines of the Canadian Council of Animal Care and after the approval of the University of Ottawa Animal Care Committee (Protocols BL-226 and BL-1700).

Staining Cells with Concanavalin A and MitoTracker CMXRos

The lectin Concanavalin A (ConA) was used previously as a specific marker of the apical membranes of HR cells on the skin of zebrafish larvae (48). The 4 dpf larvae were doubly stained with ConA (50 µg/L)–Alexa Fluor 488 conjugate (Invitrogen, Waltham, MA) for 30 min and with MitoTracker Red CMXRos (MitoRos; Invitrogen) for an additional 6 min at room temperature. The stained fish were washed and euthanized using dechloraminated Ottawa tap water containing 4 mg/mL tricaine methanesulfonate (MS-222; Syndel Laboratories, Nanaimo, BC, Canada) in 1.05 mM Tris base, adjusted to pH 7.0 with NaOH.

Embryo/Larva Digestion

To form a cell suspension, 1,000 larvae (stained with ConA and MitoRos) were transferred to a 50-mL Falcon tube (FroggaBio, North York, ON, Canada) and the liquid was removed. Five milliliters of 5 mM calcium- and magnesium-free EDTA (Bioshop, Burlington, ON, Canada) diluted in phosphate-buffered saline (PBS) (Sigma, St. Louis, MO) was added, and the fish were incubated for 30 min at room temperature.

The cell suspension was agitated by repeated pipetting (1 ml pipette tip volume) of the suspension for 2 min at room temperature to obtain a relatively homogeneous solution. Cell suspensions were then filtered through a 40-µm nylon mesh (Fisher Scientific, IL) into a 50-mL Falcon tube, and the digestion was subsequently stopped by adding 1 ml of Leibovitz’s L-15 medium (Gibco Life Technologies, NY). The isolated cells were washed 3× with L-15 medium, followed by their filtration through another 40-µm nylon mesh. The cells were pelleted by centrifugation at 1,000 rpm for 5 min, and the pellets were resuspended in 5 mL of L-15 medium. The cell suspensions were then processed by the technical staff on a fee-for-use basis at the University of Ottawa Cellular Imaging and Cytometry Facility using fluorescence-activated cell sorting (FACS). Every experimental treatment was replicated six times using 1,000 different larvae from separate breeding events. The cells derived from each sorting were subjected to FACS; samples were never pooled.

Isolation of HR Cells with FACS

The cell suspension was sorted using FACS (MoFlo Astrios EQ) at room temperature with lasers set at 488 nm and 579 nm to detect fluorescence. To minimize RNA degradation and loss of material, the sorted cells were collected directly into lysis buffer (obtained from the RNA extraction kit) and, when necessary, stored at −80°C. A typical FACS run of ~1,000 larvae would generate ~300,000 ConA-positive cells, which in turn yielded ~250 ng of total RNA. Usually, a 3-mL sample (more than 20 million total cells) was sorted within 1–2 h, therefore limiting the time during which transcriptional changes in the isolated cells could occur. The optimal settings for cell sorting were determined empirically in preliminary experiments and reproduced without modification in all subsequent experiments. It is important to note that the final number of sorted HR cells was kept constant for each sort.

Flow Cytometry Gating Strategy

Flow cytometry data analysis is built upon the principle of “gating”. Gates and regions are placed around populations of cells with common characteristics, usually forward scatter, side scatter, and marker expression, to investigate and quantify these populations of interest.

The first step in gating is distinguishing populations of cells based on their forward and side scatter properties. Forward scatter and side scatter provide estimates of the size and granularity of the cells, respectively. Furthermore, it is crucial to gate for single cells and exclude doublets or clumps of cells because flow cytometry is based on single-cell analysis. An illustrated step-by-step gating strategy for FACS analysis is shown in Supplemental Fig. S1, and each step is summarized below.

Step 1: gating based on side and forward scatter.

To distinguish population of cells according to their forward and side scatter properties, we determined the final gate by manually manipulating the settings (iterative gating) to select a region of interest that excludes cell debris and clumps of cells present in total populations of cells (Supplemental Fig. S1).

Step 2: gating based on the side scatter of area and height.

Clumps and doublets were identified using pulse shape analysis (45). The effectiveness of applying pulse shape analysis to gate for single cells is shown in Supplemental Fig. S1C using doubly stained cells to better represent the targeted population of cells. In this plot, each data point represents an “event,” which can represent single cells, doublets, clumps of cells, or debris. The doublets, clumps, and debris exhibit a high degree of autofluorescence, and thus, ideally, they must be excluded as captured events during FACS. Replotting the data from Supplemental Fig. S1A according to cell area and height (Supplemental Fig. S1B) yields a plot in which the single cells fall along a diagonal, while the doublets exhibit increased area relative to their height. Here, we have applied the cell gate (step 1) and pulse geometry gate (step 2) to the total population of cells dissociated from 4 dpf larvae to yield the gated data in Supplemental Fig. S1C, which should represent predominantly single events or cells.

Step 3: gating based on fluorescence (ConA and MitoRos).

An initial experiment used a dissociated cell suspension that was divided into four aliquots, one of which was unlabeled (negative control; Supplemental Fig. S1D) to localize the negative (unstained cell) population and to define areas of background fluorescence for further refinement of gate settings. As shown in Supplemental Fig. S1D, there were 0.2% of events exhibiting background fluorescence for the MitoRos dye; there was no detectable fluorescence for the ConA dye. Two other aliquots were stained separately with each of the dyes to determine whether there was spectral overlap between different fluorophores (ConA and MitoRos). There was 0.02%–0.3% spectral overlap observed in singly stained cells, possibly a result of autofluorescence from debris (Supplemental Fig. S1, E and F). One aliquot was doubly labeled for sorting (Supplemental Fig. S1G) and subjected to a final gating step (step 4).

Step 4: gating based on manual inspection.

The final areas for inclusion were drawn to include events that clearly lie above the background level of fluorescence, apparent from the inspection of the ConA⁺ events (red = HR cells), ConA⁻/MitoRos⁺ events (blue = non-HR cell ionocytes), and ConA⁻/MitoRos⁻ events (purple = negatives or nonionocytes). These thresholds were set by manual inspection and varied little among runs.

Whole Larva mRNA Extraction and cDNA Synthesis

Total RNA was extracted from whole larvae at 4 dpf. Ten larvae were pooled to yield N = 1, and total RNA was extracted with the RNeasy Mini Kit (Qiagen, MD) according to the manufacturer’s instructions. The cDNA was synthesized by treating 1 μg of extracted RNA with DNase (Invitrogen, Waltham, MA) and subsequently reverse transcribed with the iScript cDNA Synthesis Kit (Bio-Rad, CA) according to the manufacturer’s instructions.

RNA Extraction and cDNA Synthesis from Sorted Cells

Total RNA was isolated from the sorted cell populations using the RNeasy Mini Kit (Qiagen, MD), according to the manufacturer’s instructions. The RNA concentration and purity were determined using a NanoDrop spectrophotometer (Bio-Rad, CA). A minimum of 50 ng of total RNA was required for further processing (see below), and only samples with a 260/280 nm ratio of >1.85 were processed further. Extracted RNA was treated with DNase (Invitrogen, Waltham, MA), and 50 ng of total RNA (after DNase treatment) was used to synthesize and amplify cDNA using the QuantiTect Whole Transcriptome Kit (Qiagen, MD), with yields ranging from 10 to 40 µg of total cDNA.

Real-Time PCR

Real-time PCR (RT-PCR) was performed using a Bio-Rad CFX96 qPCR system with SsoFast EvaGreen Supermixes (Bio-Rad, CA) on all samples (cell and whole larva samples were analyzed simultaneously). To allow an accurate interpretation of the data, the reaction efficiencies for each gene were determined from standard curves using pooled samples from each treatment (Supplemental Table S2). PCR conditions were fine-tuned until the efficiencies for each primer pair were within ~90%–110% to ensure that the PCR product of interest was effectively doubling with each cycle (Supplemental Table S2). The optimum annealing temperatures (T_a) of the primers were determined by temperature gradient experiments, with identical reactions containing a fixed primer concentration, across a range of annealing temperatures. The final PCR conditions used for all the primer pairs were as follows: 95°C for 3 min, 40 cycles of 95°C for 20 s, and 58°C for 20 s, with a 5-min final extension at 72°C. All samples were assayed in triplicate. The PCR products were run on a polyacrylamide gel, and single bands were extracted and sequenced to confirm the specificity of the primers. Primers for the genes of interest and the housekeeping genes were identified from the literature (Supplemental Table S2). For each gene, the mRNA abundance specific to HR cells (ConA⁺), non-HR cell ionocytes (MitoRos⁺/ConA⁻), and nonionocytes (ConA⁻/MitoRos⁻) in each treatment group was calculated relative to the mRNA abundance of the negative cells from control fish. Relative mRNA expression was calculated using the delta-delta Ct method (37) and normalized to the mRNA abundance of the several housekeeping genes that remained constant among same treatment groups and samples used in the experiment. For the control versus low Na⁺ treatment experiments, the geometric means of two housekeeping genes, 18s (NM_001105126.2) and efi1-α (NM_ FJ915061.1), were used as an internal control. For control versus low pH experiments, the geometric means of three housekeeping genes, efi-1α, 18s, and α-tubulin (tuba1c; NM_ 001105126.2), were used for the isolated cell samples, while the geometric means of two housekeeping genes, β-actin (actb2: NM_181601.4) and ef1α, were used for the whole larvae. Relative gene expression levels were normalized to the respective levels of the controls for each treatment (pH 7.6 and 800 µM Na⁺ for low pH and low Na⁺ treated fish, respectively).

Immunohistochemistry on Larvae and Sorted Cells

To verify the specificity of ConA for HR cells, larvae were exposed to the ConA–Alexa Fluor 488 conjugate (50 µg/mL; Invitrogen, Waltham, MA) for 30 min at room temperature and fixed using 4% paraformaldehyde (PFA) in 0.1% Tween-20 in PBS (PBS-T), followed by a stepwise dehydration (50% to 100% MeOH in 3× 5-min washing steps). After a stepwise rehydration in 3× 5-min washing steps, larvae were incubated at room temperature with a solution of 3% bovine serum albumin (BSA) containing 0.8% Triton X-100 in PBS-T for 1 h. Larvae were incubated overnight at 4°C with a rabbit primary antibody specific to a region of the A subunit of bovine HA with 100% sequence identity to the zebrafish ortholog (AEMPADSGYPAYLGAR; NM_201135.2), 1:4000 dilution (24); antibody kindly provided by Professor Minoru Uchiyama] in 0.3% BSA containing 0.8% Triton X-100 in PBS-T (31). Larvae were rinsed in 3× 5-min steps in PBS-T and incubated in donkey anti-rabbit secondary antibody conjugated with Alexa Fluor 568 (1:500 dilution; ThermoFisher, Burlington, ON, Canada) in 0.3% BSA containing 0.8% Triton X-100 in PBS-T. After five washes for 5 min each with PBS-T, larvae were mounted on concave glass slides (Fisher Scientific, PA), and images were captured using an A1R+ confocal microscope (Nikon Instruments, Tokyo, Japan). Solid-state lasers emitting at 561 nm and 488 nm were used for the excitation of fluorophore; images were composed using the maximum-intensity projection of Z-stacks of 30 µm total thickness composed of 10 optical sections of 3 µm each.

Dissociated cells from larvae previously stained with ConA and sorted ConA⁺ cells were fixed for 5 min using 0.4% PFA in PBS-T. Subsequently, both samples were incubated at room temperature in a 0.03% BSA containing 0.08% Triton X-100 solution in PBS-T for 1 h. Samples were incubated overnight at 4°C with a primary antibody specific to a region of the A subunit of bovine HA (see above; 1:4000 dilution) in 0.03% BSA containing 0.08% Triton X-100 in PBS-T. Samples were rinsed 3× and incubated in donkey anti-rabbit secondary antibody conjugated with Alexa Fluor 568 (1:5000 dilution; ThermoFisher, Burlington, ON) in 0.03% BSA containing 0.08% Triton X-100 in PBS-T for 30 min. After five washes for 5 min each with PBS-T, both samples (total and sorted cells) were analyzed by flow cytometry (Beckman Coulter “Gallios”), with laser emissions set at 561 nm and 488 nm.

To quantify the percentage of sorted ConA-positive HR cells that also express HA, isolated cells were adhered to glass slides using poly-l-lysine (Sigma-Aldrich, St. Louis, MO). Poly-l-lysine was thawed at room temperature and diluted into a 0.001% solution, which was then used to fully coat Superfrost⁺⁺ microscope slides (Fisher Scientific, PA). The microscope slides were left at room temperature overnight. The poly-l-lysine solution was aspirated the following day, and the slides were rinsed with sterile water and subsequently coated with the cells that had been fixed in PFA. The coated slides were incubated at room temperature in 0.03% BSA containing 0.08% Triton X-100 solution in PBS-T for 1 h, then left overnight at 4°C with primary HA antibody (1:4000 dilution) in 0.03% BSA containing 0.08% Trion X-100 in 0.01% PBS-T. The slides were rinsed 3× and incubated in donkey anti-rabbit secondary antibody conjugated with Alexa Fluor 568 (1:5000 dilution; ThermoFisher, Burlington, ON, Canada) in 0.03% BSA containing 0.08% Triton X-100 in PBS-T for 30 min. After five washes for 5 min each with PBS-T, the stained cells were visualized using a Nikon Eclipse Ni-U upright microscope (Nikon Instruments, Tokyo, Japan), and images were captured using an Andor iXon Ultra EMCCD camera (Andor Technology, Belfast, UK).

Experimental Treatments

Fertilized embryos were transferred immediately into the appropriate treatment media [“normal water” (NW), low Na⁺ water (L-Na⁺), or acidic water] after collection and maintained until 4 dpf. Water containing 500 μmol/L Cl⁻, 250 μmol/L Ca²⁺, 40 μmol/L K⁺, and either 800 μmol/L Na⁺ (NW) or 5 μmol/L Na⁺ (L-Na⁺) (pH = 7.6) was prepared by adding CaCl₂, MgSO₄·7H₂O, KH₂PO₄, K₂HPO₄, and Na₂SO₄ salts (Sigma-Aldrich, St. Louis, MO) to double-deionized water. Acidic water (pH ~4.0) was prepared by adding H₂SO₄ (Sigma-Aldrich, St. Louis, MO) to dechloraminated Ottawa tap water (pH 7.6). For these experiments, dechloraminated Ottawa tap water was used in the control (neutral pH) experiments; the concentrations of Na⁺, Cl⁻, Ca²⁺, Mg²⁺, and K⁺ were identical to the reconstituted NW.

Statistical Analysis

All statistical analyses were performed with SigmaPlot (v. 11, Systat Inc., Chicago, IL). Student’s t-test was used to analyze flow cytometry data (i.e., cell counts between the different treatments). One-way analysis of variance (ANOVA) followed by a Holm–Šídák post hoc test was used to analyze whole-fish RT-PCR data. Two-way ANOVA followed by Holm–Šídák post hoc test was used to analyze data from RT-PCR of sorted cells. When assumptions of normality or equal variance were not satisfied, the corresponding data were transformed using a natural log or square-root transformation. For all analyses, the level of statistical significance was determined using P < 0.05.

RESULTS

Proof-of-Principle Experiments

A first goal of this study was to develop, validate, and optimize cell isolation and sorting techniques to obtain enriched populations of HR cells from zebrafish larvae at 4 dpf. The initial steps involved in vivo confirmation of cell-specific markers that could later be used to separate HR cells from other cell types using FACS. A representative confocal microscopy image of a whole-mount preparation of a 4 dpf larva doubly stained with ConA and MitoRos conjugate dyes revealed, as expected, three distinct populations of cells on the cutaneous surfaces of the yolk sac, trunk, and yolk sac extension (Fig. 1). One population of cells was positive for MitoRos only, whereas another population exhibited colocalization of MitoRos and ConA with variable intensities. A third population of cells did not stain with either MitoRos or ConA (nonionocytes; Fig. 1). No staining of cells below the superficial cutaneous epithelium or within the intestine was ever observed even when the entire thickness of specimens was examined using multiphoton microscopy (data not shown). To verify that ConA can be used as a specific marker for HR cells, specimens that were stained with ConA and MitoRos were subjected to IHC using a homologous HA antibody. Approximately 97% (n = 10 individual larvae) of cells exhibited colocalization of ConA and HA staining (Fig. 2 is a representative image), thereby confirming that ConA can indeed be used as a marker for HR cells in subsequent FACS experiments. A representative microscopy image of unsorted cells dissociated from a pool of 4 dpf larvae (Fig. 3) clearly shows the three subpopulations [HR cells (ConA⁺), other ionocyte subtypes (ConA⁻/MitoRos⁺) only, and nonionocytes (ConA⁻/MitoRos⁻) that were subsequently sorted by FACS. Note that the sorting process altered the morphology of HR cells to cause a reorientation of ConA staining.

Fig. 1.

Representative confocal microscopy images of Concanavalin A (ConA; green) and MitoTracker CMXRos (MitoRos; red) stained cells on the skin of a wild-type 4 days postfertilization zebrafish (Danio rerio) larva. A: the merged image (photographed area is indicated by the dashed red lines on the transmitted light photo in the lower right corner of A) revealed cells that either coexpressed ConA and MitoRos or cells that were positive only for MitoRos. B: an enlarged merged image of a selected trunk region showing colocalization of ConA and MitoRos. The white arrows indicate cells positive for ConA and MitoRos, and the green arrows indicate cells positive for MitoRos only. Scale bars represent 100 μm and 10 μm in A and B, respectively.

Fig. 2.

Representative confocal microscopy images of Concanavalin A (ConA; green) and H⁺-ATPase (HA; red) stained cells on the skin of a wild-type 4 days postfertilization zebrafish (Danio rerio) larva. A: double immunofluorescence staining for ConA and HA indicates coexpression of ConA and HA in many of the stained cells. B: an enlarged image of a selected yolk sac region showing ConA staining. C: an enlarged image of a selected yolk sac region showing HA staining. D: an enlarged image of a selected yolk sac region showing colocalization of ConA and HA. The white arrows indicate cells coexpressing ConA and HA, and the blue arrows indicate cells positive for only HA. Scale bars represent 100 μm in A and 10 μm in B–D.

Fig. 3.

Representative fluorescence microscopy images of unsorted dissociated cells from a pool of zebrafish (Danio rerio) larvae at 4 days postfertilization. A and E: DAPI (4′,6-diamidino-2-phenylindole) staining (blue) showing cell nuclei. B and F: MitoTracker Red CMXRos (MitoRos) staining (red) of mitochondrion rich cells. C and G: concanavalin A (ConA) staining (green) of apical membranes of H⁺-ATPase-rich (HR) cells. D and H: colocalization of DAPI, MitoRos, and ConA. The blue arrows indicate cells positive for DAPI and ConA and low intensity of MitoRos staining, the green arrows indicate cells positive for DAPI and high intensity of MitoRos staining, and the white arrows indicate cells positive for DAPI and high intensity of MitoRos and ConA. Scale bars represent 20 μm.

Postsort Analyses

FACS carried out on a cell suspension that was isolated from pre-stained 4 dpf larvae revealed that signals associated with fluorescence at the expected wavelengths of 488 nm and 579 nm were detected in the ConA and MitoRos stained cells, respectively. FACS sorted the suspension of total cells into three subpopulations based on fluorescence labeling of ConA and MitoRos. A representative FACS experiment (Supplemental Fig. S2A) demonstrated that 19.9% of the total yield of cells were ConA positive (HR cells) with cells exhibiting a range of MitoRos intensity, 32.9% were other ionocyte subtypes (MitoRos⁺/ConA^–), and 37.1% were other cell types (MitoRos^–/ConA^–).

The quality of the sorted cell fractions was assessed for purity (enrichment) using flow cytometry (Supplemental Fig. S2, B–D) or confocal microscopy images (Fig. 4; positive and negative cells were counted based on the fluorescence images using at least 26 images per sort) post-FACS. A representative example of a postsort analysis (Supplemental Fig. S2, B–D) revealed a purity of 96.3% for the sorted HR cells (ConA⁺), 97.4% for the sorted non-HR ionocytes (ConA^–/MitoRos⁺), and 99.4% for the nonionocyte population (ConA^–/MitoRos^–). Quantification of confocal microscopy images of cells post sorting (Fig. 4) demonstrated that ~95% of sorted cells were positive for ConA (N = 8 unique sorts) and thus were deemed to be HR cells. Furthermore, sorted HR cells as well as the unsorted total population of isolated cells (control) were costained with ConA and HA antibody to provide an independent measure of HR cell purity after FACS. The results of both flow cytometry (Supplemental Fig. S3) and fluorescence microscopy (Fig. 5) indicated that ~95% (N = 6 unique sorts) of the sorted HR cells were doubly positive for ConA and HA (i.e., they were confirmed as HR cells).

Fig. 4.

Representative fluorescence microscopy images of an enriched population of putative H⁺-ATPase-rich (HR) cells obtained after sorting (FACS) of dissociated cells from a pool of zebrafish (Danio rerio) larvae at 4 days postfertilization. Sorted cells displayed positive staining for (A) concanavalin A (ConA; green) and (B) MitoRos (red). Coexpression of ConA and MitoRos was demonstrated by overlaying the red and green fluorescence channels. C: magnified images of two cells are included to demonstrate the two patterns of ConA/MitoRos colocalization in the putative HR cells isolated by FACS. D–F: the white arrows indicate cells positive for ConA and exhibiting a high intensity of MitoRos staining, and the blue arrows indicate cells positive for ConA with a low intensity of MitoRos staining. Scale bars represent 20 μm.

Fig. 5.

Representative microscopy images of H⁺-ATPase-rich (HR) cells isolated from a suspension of dissociated cells of 4 days postfertilization zebrafish larvae (Danio rerio) obtained by fluorescence-activated cell sorting (FACS). A–C: transmitted light (A) or fluorescence microscopy images illustrating the colocalization of ConA⁺ (B) and H⁺-ATPase⁺ (C) cells isolated by FACS. Stained and unstained cells were counted and the percentage of HR cells in both the unsorted and postsort cell populations was determined. D: data are presented as means ± SE. Student’s t test, P < 0.05; N = 9 individual sorts; scale bars represent 50 μm.

HR Cell Numbers in Larvae Reared under Low Na⁺ or Low pH Conditions

Flow cytometry (using ConA as the marker) was performed to determine whether the numbers of HR cells increased as a potential compensatory response when 4 dpf larvae were exposed to Na⁺-deficient water (Fig. 6A) or low environmental pH (Fig. 6B). Larvae reared under low Na⁺ conditions exhibited a significantly higher percentage of HR cells (27.9% of total isolated cells) than in the cell populations isolated from fish reared in control water (19.2% of total isolated cells; Fig. 6A). Similarly, the percentage of HR cells was significantly higher (Fig. 6B) in total cell populations that were isolated from larvae reared at pH 4 (~29.9%) compared with larvae maintained at pH 7.6 (~20.8%).

Fig. 6.

The percentage of H⁺-ATPase-rich (HR) cells as determined by flow cytometry in dissociated cell suspensions derived from zebrafish (Danio rerio) larvae reared under varying environmental conditions. The percentage of HR cells was increased in fish reared in low Na⁺ (5 μm) (A) or acidic (pH 4.0) (B) water. Data are presented as means ± SE. *Significant difference between groups (Student’s t test, P < 0.05, N = 8).

Effects of Low Na⁺ or Low pH Exposure on H⁺-ATPase (atpv6v1aa) mRNA Expression in Whole Larvae and Isolated Cells

Relative transcript abundances of atpv6v1aa were assessed by RT PCR using cDNA obtained from whole-fish homogenates or cDNA obtained from three distinct populations of sorted cells, namely, HR cells, other MitoRos-positive ionocytes (i.e., non-HR cell ionocytes), and nonionocyte cells, in control zebrafish larvae and larvae that were exposed to low Na⁺ water (5 µm Na⁺) or acidic water (pH 4).

The low Na⁺ treatment did not affect the expression levels of atpv6v1aa in whole larvae (P = 0.21; Fig. 7A). The relative transcript abundance of atpv6v1aa was significantly elevated in HR cells relative to the non-HR cell and nonionocyte fractions (P < 0.001).

Fig. 7.

The effects of low Na⁺ or low pH exposure on relative mRNA levels of H⁺-ATPase (atpv61aa: NM_201135.2) in whole zebrafish (Danio rerio) larvae at 4 days postfertilization and in three populations of sorted cells obtained by FACS from the same pool of larvae. The h⁺-atpase expression levels determined from mRNA obtained using whole fish (A, C) or isolated cell populations [H⁺-ATPase-rich (HR) cells, non-HR ionocytes, and nonionocytes] (B, D) were compared in fish reared in normal Na⁺ water (800 μm) vs. low Na⁺ (5 μm) water (A, B) and control pH (7.6) vs. low pH (4.0) treatments (C, D). Data are presented as means ± SE. *Significant difference between control and experimental groups. Whole larvae samples were analyzed by one-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Isolated cells were analyzed by two-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Within each group, means that do not share letters are statistically different and P values of <0.05 are considered significant.

There was no significant effect of exposure to low Na⁺ in any of the three sorted cell fractions (Fig. 7B; P = 0.10 for HR cells, P = 0.63 for non-HR cell ionocytes, and P = 0.79 for nonionocytes). However, the low pH treatment elevated the mRNA levels of atpv6v1aa by fivefold in whole larvae (P < 0.01; Fig. 7C), eightfold in HR cells (P < 0.001), and twofold in the non-HR cell ionocytes (P < 0.05; Fig. 7D). The nonionocyte population was unaffected by either treatment.

Effects of Low Na⁺ or Low pH Exposure on nhe3b (slc9a3.2) Expression in Whole Larvae and Isolated Cells

The relative transcript abundance of nhe3b was significantly elevated in HR cells relative to the non-HR cell ionocyte and nonionocyte fractions but only in the low Na⁺ trial (Fig. 8B; P < 0.005). In the low pH trial, there was an overall effect of increased nhe3b levels in HR cells (P < 0.05). The levels of nhe3b were elevated by low Na⁺ exposure when assessed in whole larvae (twofold, P < 0.05), but the increase in mRNA was greater in isolated HR cells (sevenfold, P < 0.001; Fig. 8B). Unlike for low Na⁺, the low pH treatment did not affect the expression levels of nhe3b in whole larvae (P = 0.82; Fig. 8C). The non-HR cell ionocytes (P = 0.23) and the nonionocyte (P = 0.14) populations were unaffected by exposure to acidic water. However, the transcript abundance of nhe3b was significantly elevated by fourfold in HR cell ionocytes (P < 0.001; Fig. 8D).

Fig. 8.

The effects of low Na⁺ or low pH exposure on relative mRNA levels of NHE3b (slc9a3.2: XM_021468124.1) in whole zebrafish (Danio rerio) larvae at 4 days postfertilization and in three populations of sorted cells obtained by FACS from the same pool of larvae. The nhe3b expression levels determined from mRNA obtained using whole fish (A, C) or isolated cell populations [H⁺-ATPase-rich (HR) cells, non-HR ionocytes, and nonionocytes] (B, D) were compared in fish reared in normal Na⁺ water (800 μm) vs. low Na⁺ (5 μm) water (A, B) and control pH (7.6) vs. low pH (4.0) treatments (C, D). Data are presented as means ± SE. *Significant difference between control and experimental groups. Whole larvae samples were analyzed by one-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Isolated cells were analyzed by two-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Within each group, means that do not share letters are statistically different and P values of <0.05 are considered significant.

Effects of Low Na⁺ or Low pH Exposure on Rh Family C Glycoprotein b (rhcgb) Expression in Whole Larvae and Isolated Cells

The relative expression of rhcgb was higher in ionocytes (HR cells and non-HR cell ionocytes) relative to the nonionocyte fraction for the low Na⁺ treatment but not for the low pH treatment (Fig. 9B). The mRNA expression of rhcgb was elevated 10- and 4.2-fold by low Na⁺ exposure in whole larvae (Fig. 9A; P < 0.001) and isolated HR cells (Fig. 9B; P < 0.001), respectively. Interestingly, the low Na⁺ treatment also significantly elevated the expression of rhcgb in the nonionocyte fraction by sevenfold (Fig. 9B; P < 0.001). The mRNA expression of rhcgb was increased 5-fold after exposure to low pH when assessed in whole larvae, 15-fold in the isolated HR cell population (Fig. 9C, D), and 3-fold in the nonionocyte fraction (P < 0.001).

Fig. 9.

The effects of low Na⁺ or low pH exposure on relative mRNA levels of Rh family, C glycoprotein b (rhcgb: NM_017354103.2) in whole zebrafish (Danio rerio) larvae at 4 days postfertilization and in three populations of sorted cells obtained by FACS from the same pool of larvae. The rhcgb expression levels determined from mRNA obtained using whole fish (A, C) or isolated cell populations [H⁺-ATPase-rich (HR) cells, non-HR cell ionocytes, and nonionocytes] (B, D) were compared in fish reared in normal Na⁺ water (800 μm) vs. low Na⁺ (5 μm) water (A, B) and control pH (7.6) vs. low pH (4.0) treatments (C, D). Data are presented as means ± SE. *Significant difference between control and experimental groups. Whole larvae samples were analyzed by one-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Isolated cells were analyzed by two-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Within each group, means that do not share letters are statistically different and P values of < 0.05 are considered significant.

Effects of Low Na⁺ or Low pH Exposure on ca15a and ca17a Expression in Whole Larvae and Isolated Cells

The expression of ca15a was significantly elevated in the sorted HR cells relative to the non-HR cell ionocytes and nonionocytes (Fig. 10B, D) for both low pH and low sodium treatments (P < 0.05). Although trends were apparent, there were no significant increases in ca15a expression in whole larvae or isolated cells (including HR cells) derived from larvae after exposure to low Na⁺ (Fig. 10A, B; P = 0.32). However, the expression of ca15a was elevated after exposure of fish to low pH when determined for whole larvae (5-fold; P < 0.05; Fig. 10C) and to a greater extent in isolated HR cells (10-fold, P < 0.001; Fig. 10D); expression of ca15a was unaffected by acidic water in the other cell types (P = 0.84).

Fig. 10.

The effects of low Na⁺ or low pH exposure on relative mRNA levels of carbonic anhydrase 15a (ca15a: XM_017358460) in whole zebrafish (Danio rerio) larvae at 4 days postfertillization and in three populations of sorted cells obtained by FACS from the same pool of larvae. The ca15a expression levels determined from mRNA obtained using whole fish (A, C) or isolated cell populations [H⁺-ATPase-rich (HR) cells, non-HR cell ionocytes and nonionocytes] (B, D) were compared in fish reared in normal Na⁺ water (800 μm) vs. low Na⁺ (5 μm) water (A, B) and control pH (7.6) vs. low pH (4.0) treatments (C, D). Data are presented as means ± SE. *Significant difference between control and experimental groups. Whole larvae samples were analyzed by one-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Isolated cells were analyzed by two-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Within each group, means that do not share letters are statistically different and P values of < 0.05 are considered significant.

Similar to ca15a, the expression of ca17a was significantly elevated in the sorted HR cells relative to the non-HR cell ionocytes and nonionocytes (Fig. 11B, D; P < 0.05). Similar to the ca15a data, no significant changes in ca17a expression levels were induced by low Na⁺ treatment in either whole larvae or isolated HR cells (Fig. 11A; P = 0.40; Fig. 11B; P = 0.31). However, the ca17a expression levels increased after low pH treatment in samples obtained from whole larvae (5-fold, P < 0.05) and to a greater extent (23-fold, P < 0.001) in isolated HR cells (Fig. 11, C and D) with no significant changes in the non-HR cell fractions (P = 0.80 for non-HR cell ionocytes and P = 0.95 for nonionocytes).

Fig. 11.

The effects of low Na⁺ or low pH exposure on relative mRNA levels of carbonic anhydrase 17a (ca17a: NM_ BC057412) in whole zebrafish (Danio rerio) larvae at 4 days postfertilization and in three populations of sorted cells obtained by FACS from the same pool of larvae. The ca17a expression levels determined from mRNA obtained using whole fish (A, C) or isolated cell populations [H⁺-ATPase-rich (HR) cells, non-HR cell ionocytes and nonionocytes] (B, D) were compared in fish reared in normal Na⁺ water (800 mm) vs. low Na⁺ (5 mm) water (A, B) and control pH (7.6) vs. low pH (4.0) treatments (C, D). Data are presented as means ± 1 standard error of the mean. *Significant difference between control and experimental groups. Whole larvae samples were analyzed by one-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Isolated cells were analyzed by two-way ANOVA followed by the Holm–Šídák post hoc test (N = 6). Within each group, means that do not share letters are statistically different and P values of < 0.05 are considered significant.

DISCUSSION

In this study, we developed and validated a method based on FACS to isolate specific populations of living ionocytes from zebrafish larvae. The technique enabled subsequent analyses of gene (mRNA) expression in three groups of cells—an enriched pool of HR cells, a mixed population of all other (non-HR cell) mitochondrion-rich ionocytes, and an assortment of negative (nonionocyte) cells. The following three major findings emerged:

1) Consistent with current models of HR cell function in zebrafish larvae (e.g., [14]), the mRNA levels of five genes thought to be involved in HR cell Na⁺ uptake and H⁺ secretion (h⁺-atpase, nhe3b, rhcgb, ca15a, and ca17a) generally were enriched in the HR cells relative to the other ionocytes and nonionocytes.

2) Environmental acclimation of zebrafish larvae to water of low Na⁺ content or low pH elicited changes in gene expression in HR cells that tended to be of a larger magnitude than in whole larvae. These results demonstrate an advantage of cell-specific versus whole-larva mRNA analyses.

3) The increases in gene expression in larvae exposed to low Na⁺ or acidic water were the combined result of increased numbers of HR cells and HR cell-specific increases in mRNA expression.

Overall, the results of this study demonstrate that the isolation and analyses of living ionocytes, although labor- and time-intensive, allow for a more accurate assessment of ionocyte-specific mRNA levels of genes of interest compared with the traditional analysis of whole larva gene expression.

Isolation of Ionocyte Subtypes: the Pros and Cons

The use of FACS requires that the cells of interest be labeled and present within a suspension of dissociated cells. In our study, ConA was used to distinguish HR cells from the other cell populations in larvae. Horng et al. (16) first reported that ConA bound to a specific population of cells in the skin of zebrafish larvae that were enriched with HA and thus deemed to be HR cell ionocytes. Therefore, an important initial experiment in the current study was to use an HA antibody in combination with ConA in whole fish and isolated cells to determine the specificity of ConA as a marker for HR cells to be used for FACS. An advantage of labeling cells with ConA is that it permits the isolation of live cells, which facilitates RNA extraction (albeit RNA degradation can be a problem), compared with cells fixed with PFA, whereby crosslinking of protein and RNA hinders RNA extraction (13). Using ConA in combination with the mitochondrial stain MitoRos also allowed the isolation of a non-HR cell pool of ionocytes, a mixed population of ConA^–/MitoRos⁺ NaR and NCC cells. Finally, a third population of remaining cells (nonionocytes) could be sorted simultaneously as a ConA^–/MitoRos^– fraction. Importantly, the staining and sorting procedure excluded the contribution of potential ConA- or MitoRos-positive cells from other tissues (e.g., kidney and intestine) to the final sorted population presumably owing to the brief periods of exposure to these vital dyes and low rates of drinking.

In the current study, not all of the sorted HR cells were MitoRos⁺. Indeed, there was a significant fraction of ConA⁺ cells that typically were categorized as MitoRos^– during FACS. However, although MitoRos staining of these cells was below the threshold set for FACS detection, they nevertheless did exhibit a low level of MitoRos staining when examined by IHC. Thus, the final population of HR cells that was examined for gene expression was a mixture of cells exhibiting a gradient of MitoRos staining. In all cases, however, the cells coexpressed ConA and HA. The variability in the intensity of MitoRos staining of HR cells suggests that these cells may represent different developmental stages or indeed distinct subtypes of HR cells.

A major concern related to analysis of mRNA obtained from whole larvae in studies evaluating ionic or acid–base regulation is that ionocytes are massively underrepresented in comparison with total cells (e.g., ~400 HR cells per fish in comparison with millions of other cells), making it challenging to detect subtle changes in gene expression that may be specific to HR cells or other ionocytes. In the present study, the fraction of ConA⁺ HR cells sorted by FACS (~17%–25%) clearly was higher than the proportion of epithelial HR cells in vivo. The increase in the relative proportion of HR cells was the result of several factors including removal of cellular debris from total cells by centrifugation before sorting and by exclusion of debris, dead cells, and clumps of cells or doublets by sorting from a preselected region that excluded those events during FACS (Supplemental Fig. S1). Furthermore, the method of isolating cells incorporated a relatively brief (30 min) enzyme-free digestion, which may have prevented dissociation of deeper layers of cells. Thus, the cells that were sorted likely originated predominantly from epithelial and subepithelial layers, with cells from deeper tissues being excluded. Regardless, the percentage of ConA⁺/HA⁺ cells recovered by sorting certainly was higher than the expected percentage of HR cells even if only the epithelial cell layer was being sorted. Thus, the technique developed in this study, while capable of yielding highly purified fractions of ConA⁺/HA⁺ cells, may be overestimating the numbers of HR cells originating from the cutaneous epithelium.

Although cell sorting occurred in multiple episodes, numbers of HR cells were kept constant (300,000 HR cells) for each sort, and the total mRNA level for each sort also was kept constant by using 50 ng of total mRNA from all sorted samples. In summary, FACS using ConA and MitoRos is a viable method for isolating highly enriched populations of HR cells and non-HR ionocytes from zebrafish larvae. The main advantage of this approach is that it increases the probability of detecting changes in cell-specific gene expression that otherwise might be undetectable using whole-larva homogenates. However, there are some shortcomings associated with cell sorting before gene expression analysis. Obviously, the procedure is laborious, is time-consuming, and can be expensive (user fees) depending on the type of sorting facility being used. These disadvantages must be weighed against the potential benefits of cell sorting. Another possible negative outcome of cell sorting is that changes in mRNA levels incurred in vivo may be reversed upon removal of the environmental stimulus. In addition to natural mRNA turnover tending to lower levels during sorting, there is a greater potential for RNA degradation by nonendogenous RNase and contamination from exogenous templates.

Responses of HR Cells to Low Environmental Na⁺

It has long been known that exposure of fish to waters of low Na⁺ content elicits compensatory responses aimed at maintaining normal rates of Na⁺ uptake in the face of reduced Na⁺ availability (3, 12, 32, 38, 41). In zebrafish, arguably the most significant compensatory response to low ambient Na⁺ is an enhanced ability to absorb Na⁺ owing to increases in both Na⁺ uptake affinity and capacity (3, 29). It has been suggested that the predominant pathway enabling Na⁺ uptake in zebrafish larvae exposed to low Na⁺ water is via Nhe3b linked to ammonia excretion through Rhcgb (42) that may be assisted by NKA isoform switching (12). In agreement with the proposed metabolon model of Wright and Wood (47), the close association of Nhe3b, Rhcgb and Ca17a (23) facilitates Na⁺ uptake and ammonia excretion by promoting chemical gradients favoring Na⁺/H⁺ exchange and NH₃ diffusion. Accordingly, one would predict HR cell-specific increases in mRNA expression of nhe3b, rhcgb, and ca17a in fish exposed to low environmental Na⁺. In the current study, however, only nhe3b and rhcgb were increased statistically; ca17a levels were unchanged. Statistically significant increases in nhe3b and rhcgb also were detected using whole larvae, although for nhe3b, the magnitude of the mRNA increase was smaller in comparison with that of the HR cell-specific increases. Only two other studies have reported mRNA or protein levels for one or both of these genes in zebrafish larvae acclimated to low Na⁺. Shih et al. (42) measured a small (1.5-fold) increase in nhe3b levels in whole larvae, whereas rhcgb expression was unaltered. In the study of Nakada et al. (39), a single representative Western blot image depicts an apparent increase in Rhcgb protein in 6 dpf larvae acclimated to 1/20× FW. The increases in gene expression of nhe3b and rhcgb reported here are consistent with the metabolon model of Na⁺ uptake that is known to operate in zebrafish larvae under conditions of low ambient Na⁺ (42). However, despite the involvement of Nhe3b in Na⁺ uptake from low Na⁺ water or its capacity to facilitate Na⁺/NH₄⁺ exchange (22), its operation apparently is not essential to maintain normal rates of Na⁺ uptake based on the results of recent nhe3b knockout experiments (50). The continued uptake of Na⁺ in the nhe3b knockouts may reflect the presence of redundant pathways that often are revealed in such knockout studies (49). The lack of an increase in HR cell-specific expression of ca17a was unexpected, especially considering that Ito et al. (23) reported that the expression of ca17a in 7 dpf larvae exposed to low Na⁺ water was increased, in keeping with its role to facilitate an increase in H⁺ production from the catalyzed hydration of CO₂. It is possible that intracellular CA activity was increased despite constant mRNA levels owing to posttranslational modifications (4). In addition to the HR cell-specific increases in gene expression reported in this study, any changes in whole-larva mRNA levels will also reflect the increased numbers of HR cells, especially for genes such as nhe3b that are enriched within the HR cell. The flow cytometry data presented in the current study indicated a significantly higher percentage of HR cells isolated from zebrafish larvae exposed to low Na⁺ (27.9%) compared with those exposed to control water (19.2%).

Previous studies have documented that atpv6v1aa mRNA levels are lowered in whole larvae (42) or adult gill (42, 48) during acclimation of fish to low Na⁺ water. It was argued that a decrease in HR cell HA activity would benefit Nhe3b function in low Na⁺ water by the presumed resultant increase in cytosolic H⁺ activity. In this study, there was no effect of acclimation to low Na⁺ on HR cell-specific atpv6v1aa amounts, although there was an apparent trend for lowered levels (P = 0.10). Similarly, atpv6v1aa levels assessed in whole larvae were unaffected by low Na⁺ acclimation, despite the greater numbers of HR cells known to be enriched in atpv6v1aa. The constancy of atpv6v1aa in the face of HR cell proliferation is in itself an indirect evidence for a decline in HR cell atpv6v1aa levels.

The role of Ca15a in HR cell-mediated Na⁺ uptake, ammonia excretion, and acid–base balance is not well understood. Although it is agreed that Ca15a is positioned to catalyze boundary layer reactions involving CO₂/HCO₃^–owing to its anchorage to the apical membrane of HR cells (36) or other acid-secreting ionocytes (47), there is some debate as to whether it catalyzes dehydration or hydration reactions [for conflicting models, see Guh et al. (14) and Wright and Wood (47)]. Regardless which model is correct, ca15a expression was unchanged by low Na⁺ treatment in the present study, although there was an apparent trend for increased mRNA levels in whole larvae and HR cells. Thus, unlike in the adult zebrafish gill, which exhibits an increase in ca15a during low Na⁺ exposure (23, 36), Na⁺ uptake by the larval HR cell under similar conditions may not be regulated in the same fashion.

Responses of HR Cells to Low Environmental pH

As in several other species, exposure of zebrafish to acidic water causes a transient reduction in whole body Na⁺ levels (26). In zebrafish, the recovery of whole body Na⁺ content results from the stimulation of Na⁺ uptake in the face of persistently elevated Na⁺ efflux (26). The ability of zebrafish to elevate Na⁺ uptake appears to be a unique strategy among FW teleosts to restore Na⁺ balance during exposure to acidic water [reviewed by Kwong et al. (30)].

Although it is generally assumed that exposure of zebrafish to acidic water elicits metabolic acidosis (35) in addition to the well-documented disruption of Na⁺ homeostasis, there are no data to support this view. In other species, the effects of water acidification on blood acid–base status are dependent on water hardness (Ca²⁺ levels) [see review by Kwong et al. (30)]. For the purposes of this study (water Ca²⁺ = 0.25 mM in neutral and acidified water), it was assumed that exposure of larvae to acidic water disrupted both Na⁺ and H⁺ balance, and thus, the experimental design focused on quantifying the numbers of HR cells and the HR cell-specific genes thought to be involved in Na⁺ uptake and H⁺ secretion.

An increase in HR cell density was documented previously as a key physiological response to acid exposure in zebrafish larvae and adult gill (6, 15). The flow cytometry data reported here are consistent with these previous studies, with a significantly larger percentage of HR cells isolated from larvae exposed to low pH (~29%) compared with control larvae (~20%). These results demonstrate that when relying on tissue or whole-body analyses, any changes in mRNA expression of genes localized to HR cells will be the result of increased HR cell numbers coupled with any HR cell-specific changes in gene expression.

Several studies have investigated the effects of low ambient pH on gene expression in zebrafish larvae and adult gill; the results of these experiments are summarized in Supplemental Table S1. Two important findings emerge from these data; first, depending on the gene being evaluated, the results can vary markedly between studies, and second, the effects of external acidification on mRNA levels can differ when comparing adult gill and whole larvae. As pointed out by Lin et al. (34), a possible explanation for the differences is that whole larvae likely express ion transport genes at numerous locations other than the ionocytes of the yolk sac epithelium, including the kidney and intestine. Differences in results between studies, even when assessing equivalent tissues (i.e., gill or whole larvae), might also arise from variability in the proportion of target cells (e.g., HR cells) in the mixed population of cells extracted from the target tissues. The results of this study provide the first data on HR cell-specific changes in gene expression (mRNA levels) in zebrafish larvae exposed to acidic water, which are compared with data obtained from whole larvae. Because we did not measure protein levels, we cannot confirm that the levels of mRNA were correlated with the levels of their corresponding gene products.

Kumai et al. (26) presented evidence supporting a more important role for Nhe3b in Na⁺ uptake in zebrafish larvae reared in acidic water compared with neutral water. However, despite their prediction of increased nhe3b expression, mRNA levels were unchanged in acidic larvae; a result that was attributed to the possible “masking” of HR cell-specific changes by a lack of any changes (or opposing changes) in nhe3b expression in the non-HR cells (26). The results of the current study support their interpretation; nhe3b levels were unchanged in whole larvae but were increased markedly in HR cells. Thus, the finding of reduced nhe3b levels in gill from adults exposed to acidic water (5, 48) should be interpreted with caution until comparable data are obtained from HR cells sorted from adult gill.

The HR Cell Metabolon Model

The increased contribution of Nhe3b to Na⁺ uptake in zebrafish exposed to acidic water may reflect its reliance on the HR cell metabolon (23) whereby Na⁺ uptake is facilitated by the concerted effects of increased ammonia excretion via Rhcgb and increasing intracellular H⁺ availability via CA. The results of this study demonstrating higher levels of rhcgb and ca17a specifically in HR cells are consistent with the model of Na⁺ uptake in acidic water (27), whereby increasing ammonia excretion (as NH₃) serves to increase boundary layer pH, whereas an increase in cytosolic CA activity causes intracellular acidification. Ultimately, both of these factors contribute to increasing Na⁺ uptake via Nhe3b by increasing the H⁺ gradient across the HR cell apical membrane. In addition, because zebrafish Nhe3b also functions as a Na⁺/NH₄⁺ exchanger (22), Na⁺ uptake driven by the gradient provided by internal NH₄⁺ can also maintain Nhe3b activity during external acidification. Kumai and Perry (27) also suggested that the metabolon model of Na⁺ uptake in acidic water was being assisted by an increasing contribution of HA-mediated H⁺ excretion, which serves to enhance NH₃ diffusion. The increased mRNA expression of atpv6v1aa in HR cells in larvae exposed to water of low pH supports its purported role in Na⁺ uptake. Overall, the HR cell-specific increases in transcript levels of nhe3b, ca17a, rhcgb, and atpv6v1aa in fish exposed to acidic water support the notion of the HR cell as the main acid–base regulating cell in zebrafish larvae (15, 16, 19). Interestingly, the nonionocyte population exhibited a significant increase in rhcgb when larvae were exposed to acidic water. It is important to recognize that the nonionocyte fraction presumably included cells from other tissues that could also be important for acid–base regulation and Na⁺ uptake (e.g., [40]).

Similar to previous studies on larvae or gill (34, 36), ca15a expression was increased by acidic water treatment in larvae or HR cells. Depending on whether Ca15a is catalyzing boundary layer hydration of dehydration reactions, its increased activity would promote Nhe3b function either by lowering H⁺ in the boundary layer (via dehydration of HCO₃⁻) or providing a source of H⁺ for NH₃ protonation (via hydration of CO₂); in either case, Na⁺ uptake by Nhe3b is expected to increase.

Perspectives and Significance

Previous research efforts on the gill/skin of zebrafish revealed a complex network of regulatory responses accompanying low Na⁺ and low pH, which have improved our understanding of fundamental ion transport mechanisms in HR cells (see Supplemental Table S1). The results of the current study extend this knowledge by providing the first data on HR cell-specific changes in gene expression as compensatory mechanisms to offset the deleterious effects of low Na⁺ and low pH. Specifically, this study demonstrated that the expression of several key target genes (h⁺-atpase, nhe3b, rhcgb, ca15a, and ca17a) is regulated in HR cells and that this regulation occurred independently of any changes in the number of HR cells.

The primary benefit of FACS is that cell-type-specific mRNA data are generated, which can eliminate the confounding effects of mRNA changes in other cell types. Also, because multiple embryos (1,000 zebrafish larvae) for each sort are usually processed to isolate cells, the effects of embryo-to-embryo variability are reduced. In the current study, the changes in mRNA in the sorted cells were higher (except in one instance) than in the unsorted cells, and the sorting procedure revealed changes in mRNA levels of Nhe3b after low pH exposure that were undetected using unsorted cells. On the other hand, because HR cells are a small population of total cells in zebrafish, sorting this population to high purity requires an immense amount of time and labor.

Presumably, the advantage of FACs will be cell-type-specific, and the decision on whether to pursue FACS as a technique to estimate cellular gene expression should evaluate the benefits against the higher costs.

GRANTS

This research was supported by Natural Sciences and Engineering Research Council (NSERC) Discovery and Research Tools and Instruments grants to S. F. Perry.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.F.P. conceived and designed research; K.S.-M. performed experiments; K.S.-M. analyzed data; S.F.P. interpreted results of experiments; K.S.-M. and S.F.P. prepared figures; K.S.-M. drafted manuscript; S.F.P. edited and revised manuscript; K.S.-M. and S.F.P. approved final version of manuscript.