Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

Contents lists available at SciVerse ScienceDirect

Comparative Biochemistry and Physiology, Part A

j ourna l homepage: www.e lsev ie r .com/ locate /cbpa

Effects of low salinity media on growth, condition, and gill ion transporter expressionin juvenile Gulf killifish, Fundulus grandis

Joshua Patterson a, Charlotte Bodinier b, Christopher Green a,⁎a Louisiana State University Agricultural Center, Aquaculture Research Station, Baton Rouge, LA 70820, USAb Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, USA

⁎ Corresponding author. Tel.: +1 2257652848; fax: +E-mail address: cgreen@agcenter.lsu.edu (C. Green).

1095-6433/$ – see front matter © 2012 Elsevier Inc. Alldoi:10.1016/j.cbpa.2011.12.019

a b s t r a c t

a r t i c l e i n f o

Article history:Received 17 October 2011Received in revised form 21 December 2011Accepted 26 December 2011Available online 8 January 2012

Keywords:Fundulus grandisGrowthMitochondria-rich cellsOsmoregulationSalinityTeleosts

The Gulf killifish, Fundulus grandis, is a euryhaline teleost which has important ecological roles in thebrackish-water marshes of its native range as well as commercial value as live bait for saltwater anglers. Ef-fects of osmoregulation on growth, survival, and body condition at 0.5, 5.0, 8.0 and 12.0‰ salinity were stud-ied in F. grandis juveniles during a 12-week trial. Relative expression of genes encoding the ion transportproteins Na+/K+-ATPase (NKA), Na+/K+/2Cl− cotransporter(NKCC1), and cystic fibrosis transmembraneconductance regulator (CFTR) Cl− channel was analyzed. At 0.5‰, F. grandis showed depressed growth,body condition, and survival relative to higher salinities. NKA relative expression was elevated at 7 dayspost-transfer but decreased at later time points in fish held at 0.5‰ while other salinities produced nosuch increase. NKCC1, the isoform associated with expulsion of ions in saltwater, was downregulated fromweek 1 to week 3 at 0.5‰ while CFTR relative expression produced no significant results across time or sa-linity. Our results suggest that Gulf killifish have physiological difficulties with osmoregulation at a salinityof 0.5‰ and that this leads to reduced growth performance and survival while salinities in the 5.0–12.0‰are adequate for normal function.

© 2012 Elsevier Inc. All rights reserved.

1. Introduction

Gulf killifish Fundulus grandis are abundant in coastal marshesfrom northeastern Florida to Veracruz, Mexico (Rozas and LaSalle,1990; Williams et al., 2008) and constitute an ecologically importantportion of the estuarine nekton in these areas (Kneib, 1997). This spe-cies is a popular live baitfish along the southern Atlantic and Gulf ofMexico coasts where most marketed fish are taken from wild stockswith seasonally inconsistent availability (Green et al., 2010). Egg pro-duction per gram female per day in F. grandis is at least 6-fold lowerthan in fathead minnow Pimephales promelas, a successfully culturedfreshwater (FW) species accounting for 20% of baitfish sales in theUnited States (Clemment and Stone, 2004; Kumaran et al., 2007;Green et al., 2010). While fecundity is low, Fundulus larvae are highlydeveloped at hatch with fully functional eyes and mouthparts(Armstrong and Child, 1965). Larvae also possess functional gills,which take over the task of osmoregulation from the external epithe-lia shortly after hatching (Katoh et al., 2000), although the precise on-togeny of ion regulation and oxygen uptake in larval gills remains asubject of discussion (Rombough, 2007). Despite knowledge of theontogeny of embryonic and larval development, few studies have ex-amined early-life stage growth and survival in F. grandis.

1 2257652877.

rights reserved.

F. grandis is able to survive and reproduce across a broad salinityrange. Embryos will develop and hatch in salinities ranging from0 to 80‰ (Perschbacher et al., 1990) and fish have been collectedfrom waters ranging from 0.05 to 76‰ (Simpson and Gunter, 1956).Salinity has been shown to influence growth rates across all life histo-ry stages in many fish species and improved growth at intermediatesalinities has often been associated with a lower standard metabolicrate (Bœuf and Payan, 2001). Growth of the euryhaline Southern Bra-zilian flounder Paralichthys orbignyanus was reduced in FW, althoughsurvival was unaffected (Sampaio and Bianchini, 2002). Research ex-amining cellular and molecular aspects of osmoregulation and theireffects on and survival in F. grandis is valuable for further develop-ment of culture methods in this species. Work with the Fundulusgenus provided much of the early knowledge base on ion transportin the gills of saltwater fish, with the mummichog Fundulus heterocli-tus as the primary experimental animal (Wood and Marshall, 1994).F. heteroclitus is a closely related congener whose range overlaps F.grandis in northeastern Florida (Gonzalez et al., 2009). This specieshas been the subject of extensive work on osmoregulation(Marshall et al., 2002; Wood and Laurent, 2003; Scott et al., 2004,2005; Wood and Grosell, 2008; Hyndman and Evans, 2009;Flemmer et al., 2010; Whitehead et al., 2011) while considerablyless attention has been devoted to salinity effects on F. grandis. Meta-bolic costs of acclimation to FW in adult F. grandiswere demonstratedby Kolok and Sharkey (1997) who reported significantly lower criticalswimming speeds (Ucrit; Beamish, 1978) in FW acclimated fish versus

416 J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

fish acclimated to 10‰ brackish-water. In addition to reduced Ucrit,the authors also observed 40% mortality after a swimming challengein FW and no mortalities in identical brackish-water trials.

Fish in non-isosmotic media must actively compensate for passivediffusion of salts and water between blood and environment. In tele-osts, the salinity range which generally minimizes the osmotic gradi-ent between blood and water is ≈10–12‰ and it has been proposedthat energy conserved due to reduced osmoregulatory activity nearthese salinities is redirected for growth (Varsamos et al., 2005). Forexample, a series of differential salinity exposures estimated the isos-motic point of P. orbignyanus at 10.9‰ (Sampaio and Bianchini, 2002)and the plasma osmolality of juvenile dusky kob Argyrosomus japoni-cus was consistently around 360 mOsm l−1 or 12‰ (Bernatzeder etal., 2008). Once the gills are fully developed, they are the main siteof osmotic regulation through secretion of ions in seawater and up-take of ions in FW (Perry, 1997; Evans et al., 2005). Within the gillsof young juvenile to adult fish, specialized mitochondria-rich cells(MRC, formerly referred to as chloride cells) perform the ion andwater trafficking tasks of osmoregulation. In F. heteroclitus embryos,MRC are located largely in the yolk-sac membrane, with a gradual mi-gration to the external epithelia beginning several days before hatchand MRC shared between external epithelia, gills, and opercularmembrane until ~25 days post hatch at which point the gills andopercular membrane dominate (Katoh et al., 2000). A number oftransmembrane proteins expressed in MRC execute various osmoreg-ulatory functions and the direction of water and ion flow is depen-dent upon their apical or basolateral location in the cell membrane(Marshall et al., 2002; Hiroi and McCormick, 2007; Bodinier et al.,2009b). Recently described ion transport proteins which are thoughtto facilitate absorption of ions in hyposmotic media include ClC-3, amember of the ClC chloride channel family (Tang et al., 2010), theNa+/H+ exchangers NHE2 and NHE3 (Yan et al., 2007; Ivanis et al.,2008), Na+/Cl− cotransporter (NCC), a cation-chloride cotransporter(Hiroi et al., 2008; Inokuchi et al., 2009), H+-ATPase (Lin et al., 2006),and the SLC26 anion exchange family (Bayaa et al., 2009). The precisefunctions of these newly described ion transport protein are still openfor discussion. Three ion transport proteins found in MRC which co-operate to regulate ionic concentration in teleost blood were chosenfor this study: Na+/K+-ATPase (NKA), Na+/K+/2Cl− cotransporter(NKCC1), and the cystic fibrosis transmembrane conductance regula-tor (CFTR) Cl− channel (Marshall and Singer, 2002; Hiroi andMcCormick, 2007). Despite their general classification as secretoryproteins involved in hyposmoregulation, NKCC1 and CFTR were cho-sen for this experiment in F. grandis because of interest in examiningtheir potential transcriptional regulation and because many of theproteins involved in ion absorption in euryhaline fishes have beenonly recently described and their functions are still a subject of dis-cussion. For example, work on ClC-3, the Cl− transporter which likelyplays an absorptive role in hyposmotic environments (Tang et al.,2010), had not yet been published when the present study was con-ducted. While NKCC1 and CFTR change distribution within MRC de-pendent upon osmotic status of the animal (Marshall et al., 2002),NKA remains embedded in the basolateral membrane of MRC whereit generates an electrochemical gradient through hydrolysis of ATP.This electrochemical gradient is necessary for the activity of othertransmembrane proteins, including NKCC1 and CFTR, whose abun-dance and location determine the rate and direction of ion transfer(Marshall and Singer, 2002; Bodinier et al., 2009b). Reduced growthrates observed in P. orbignyanus from a FW environment may havebeen partially attributable to the high energetic costs of increasedNKA activity (Sampaio and Bianchini, 2002). Reduced NKA activityin isosmotic conditions may provide an energetically favorable envi-ronment for other teleost species (Lin et al., 2003; Saoud et al.,2007; Partridge and Lymberry, 2008). Previous work with mRNA ex-pression of these three genes in euryhaline fishes provides somebackground for the present study. Adults of the F. grandis congener

F. heteroclitus up-regulated NKA mRNA in the gills after transferfrom brackish water to FW (Scott et al., 2004; Scott et al., 2005). Inthe euryhaline Europoean sea-bass, Dicentrarchus labrax, CFTRmRNA was down-regulated in the gills after transfer from seawater(SW) to FW (Bodinier et al., 2009b). Similarly, expression of NKCC1was significantly lower in the gills of FW acclimated D. labrax thanthose acclimated to SW (Lorin-Nebel et al., 2006).

Knowledge of the effects of low salinities on growth of juvenilefish could provide practical information for culture of early-lifestage F. grandis and gene expression analyses may reveal how the ef-fects of physiological compensation at low salinities manifest at themolecular level. The objective of this study was to evaluate the abilityof juvenile F. grandis to survive and grow across a hyposmotic salinitygradient and to examine long term expression patterns of genesencoding important ion transporters in the gills.

2. Materials and methods

2.1. Experimental systems and water quality

This study was conducted at the Aquaculture Research Station(ARS; Louisiana State University Agricultural Center, Baton Rouge,LA, USA) in four identical, adjacent recirculation systems maintainedat salinities of 0.5, 5.0, 8.0 and 12.0‰ using Crystal Sea Marinemix(Marine Enterprises International Inc., Baltimore, MD, USA). Replicatetanks were 100-L glass aquaria containing 75 L of recirculating waterwith aeration provided by a single airstone. Photoperiod was set at13 h light/11 h dark. Water remained at ambient temperature andwas monitored independently in each recirculation system usingtemperature loggers which recorded temperature every 30 min forthe duration of the experiment. Each systemwas serviced by a bubblewashed bead filter and a 25 W ultraviolet sterilizer.

Salinity and dissolved oxygen (DO) were recorded every 48 h.Dechlorinated municipal water was used to compensate for evapora-tion. Total ammonia nitrogen (TAN) and nitrite nitrogen (nitrite)were measured by the salicylate and diazotization methods, respec-tively. Titration kits were used to test alkalinity and hardness(reported as CaCO3) of the treatment water. Sodium bicarbonatewas added to the systems during the experiment to maintain alkalin-ity. An ion meter was used to determine pH. Water quality parame-ters other than salinity and DO were tested once a week for theduration of the experiment.

2.2. Stocking and feeding

Juvenile F. grandiswere obtained from a population cultured at theARS and held in a single recirculating tank with aeration and 7.0‰ sa-linity. Prior to the current study fish were fed a commercial starterdiet containing approximately 52% protein and 14% lipid. Groups ofjuvenile fish were grown in triplicate at four different salinities; 0.5,5.0, 8.0, and 12.0‰. No acclimation occurred during transfer from7.0‰ to other salinities, so the fish were considered directly trans-ferred for purposes of molecular work. The growth trial lasted12 weeks with gill samples taken after weeks 1, 3, and 7. The discrep-ancy in length of gill sampling and growth trials was attributable todifferential effective time scales for the two aspects of the study. Westopped gill sampling for gene expression analysis after week 7 be-cause regulation was expected to be complete, with expression levelsstable by this point. The growth and body condition trial was extend-ed five weeks past the end of gill sampling to allow a sufficient periodfor treatment effects to fully manifest. Salinities of 0.5, 5.0, and 8.0‰were chosen to represent varying degrees of hyposmotic stress.Each of the four systems was randomly stocked in triplicate withfish at a mean mass of 0.50±0.01 g. Seventy-five fish were initiallyused in each replicate for a density of one fish per L. Initially, fishwere fed the commercial starter diet, however necropsy of fish

417J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

sampled after one week revealed extreme hepatic lipidosis and dur-ing weeks 2–12 a crumbled commercial diet containing approximate-ly 32% protein and 4% lipid was fed to reduce caloric intake among alltreatments. Fish were fed 4% of body weight per day divided intomorning and afternoon feedings with the amount fed adjusted fol-lowing biweekly growth sampling.

2.3. Sampling procedures

At stocking,wetmass (nearest 0.0001 g) and total length (TL), or thedistance from the tip of the snout to themost distal portion of the caudalfin (nearest mm) were determined for a random sample of 25 individ-uals per replicate. After initial stocking the same parameters weremea-sured for a random sample of 20 individuals per replicate every 14 days.Length and weight data were used to calculate growth rates as well asrelative condition factor at final sampling. Mean relative condition fac-tor was compared amongst salinities using the calculations of Le Cren(1951). Because all fish were reared in the same conditions prior tobeing stocked at various salinities, length and weight data from initialstocking was used to create a log transformed plot with the equation:

log W¼ log aþb log L ð1Þ

whereW is weight (g), L is TL (cm), b is the slope of the line and log a isits intercept. Subsequently, the values a and nwere used as constants inthe equation:

W ¼ aLb ð2Þ

to calculate the ‘normal’ weight for an individual based on TL, W wascalculated for each fish at final sampling and a relative condition factor(Kn) was determined as a ratio of recorded final weightW to predicted‘normal’ weight for each individual using the formula:

Kn ¼ W

Wð3Þ

Specific growth rate (SGR) was calculated by the formula:

SGR ¼ 100 lnW2− lnW1ð Þ=T ð4Þ

whereW1 andW2 are mean initial and final weight, respectively and Tis the number of days in the growth trial.

2.4. RNA extraction and cDNA synthesis

Following anesthesia with 60 mg/L MS-222, gills were removedfrom three fish per replicate after weeks 1, 3, and 7. Gill samplesfrom all replicates of a single salinity were pooled to provide ninesamples for each time and salinity combination. The entire gill lamel-lae were immediately flash-frozen in liquid nitrogen and stored at−80 °C until RNA isolation.

Five of the nine gill samples from each salinity and date combina-tion were randomly chosen for analysis. Frozen organs were groundto a fine powder. Approximately 100 mg of this powdered tissue wasimmediately transferred into 1 mL of TRIzol reagent (Invitrogen,Carlsbad, CA, USA), vortexed, and passed 10 times through a 20-1/2gauge needle to ensure complete homogenization. Total RNA was iso-lated using the TRIzol reagent according to themanufacturer's instruc-tions with slight modification. To increase RNA purity, the TRIzoltreatment was repeated on each sample as described by Galvez et al.(2007). After the second TRIzol isolation, RNA from each sample wasdissolved in 50 μL of RNAase-free water with concentration subse-quently quantified using a NanoDrop 1000 (Thermo Fisher ScientificInc., Wilmington, DE, USA). Reverse transcriptase polymerase chainreaction (RT-PCR) was used to synthesize cDNA from 2 μg of totalRNA in a reaction volume of 20 μL. RT-PCR was performed using the

Applied Biosystems high capacity cDNA reverse transcription kitwith RNAase inhibitor (Life Technologies Inc., Carlsbad, CA, USA).

2.5. Quantification of gene expression by real-time PCR

The mRNA encoding genes NKA, NKCC1, CFTR, and elongation factor(EF1α) (GenBank accession nos. AY057072, AY533706, AF000271, andAY430091, respectively) was quantified in duplicate from gills at eachof the four experimental salinities and each of three time periods withinsalinity. The reference gene EF1α has been frequently used in studies onexpression of ion transport proteins in fish (Galvez et al., 2007; Bodinieret al., 2009b; Tipsmark et al., 2011) andwas employed here to determinerelative changes in target gene mRNA. First strand cDNA was quantifiedusing a NanoDrop 1000 and diluted 20 times in nuclease-free water.Quantitative real-time RT-PCR (qPCR) analyses were performed usingan iCycler iQ Real-Time PCR Detection system and the iQ SYBR Greensupermix (Bio-Rad Laboratories Inc., Hercules, CA, USA). Primers forEF1α and CFTR were designed using Primer Express software (AppliedBiosystems, Calsbad, CA, USA) and were reported in Scott et al. (2004).Primers for NKCC1 andNKAwere designed byMeng et al. (unpublished)using Primer 3 software v 0.4.0 (National HumanGenomeResearch Insti-tute, USA). Primer sequences were; NKCC1 (NKCC1F,5'-CCACTGG-TATTCTGGCTGGT-3' and NKCC1R,5'-GAAACTGCAACCCCCAAGTA-3')and NKA (NKA1F,5′-ACTGCCAAGGCCATTGCTAA-3′ and NKA1R,5′-AAC-GACGCAAGCTTTGGCAT-3′). Standard curves were run for each primerby amplifying serial dilutions of a random mixture of experimentalcDNA using thermal cycling protocols described below. Efficiencies forall four primers were between 1.8 and 2.0. Samples contained 4 μLnuclease-free water, 5 μM forward primer, 5 μM reverse primer, 10 μLiQ SYBRGreen supermix, and 5 μL of 1:20 diluted template cDNA. The cy-cling conditions were: denaturation and hot-start polymerase activation(95 °C, 4 min); 50 cycles (95 °C, 10 s; 60 °C, 30 s); melt curve (95 °C to65 °C in 0.5 °C increments). Samples not producing expectedmelt curveswere discarded from analysis. Critical threshold (Ct) values were auto-matically calculated for each sample using algorithms in the qPCR soft-ware (Bio-Rad). Ct values were normalized and reaction efficiencieswere calculated using the LinRegPCR data analysis program (version12.5, download: http://LinRegPCR.HFRC.nl, Ruijter et al., 2009). Reactionefficiencies for all reactions were between 85% and 110% with 100%representing a doubling in PCR product after every cycle. A blank controlcontaining water was included with each reaction to test for environ-mental contamination. Each reaction included a blank control withwater. Ct values for EF1αwere minimally variable amongst different sa-linity treatments at 21.95±0.32, 21.67±0.20, 21.78±0.37, and 22.08±0.42 for salinities of 0.5, 5.0, 8.0, and 12.0‰, respectively. This lack of var-iation coupledwith past utilization as a reference gene in similar applica-tions confirms the functionality of EF1α for this experiment. Relativeexpression of target genes was calculated for 100 copies of the referencegene using the formula of Rodet et al. (2005):

N ¼ 100x 2 Ct reference gene – Ct target geneð Þ ð5Þ

2.6. Statistical analysis

Data were reported as mean±standard error of the mean (SEM).Relative condition factor (Kn) and growth rates were comparedamongst salinities using one-way analysis of variance (ANOVA) andRyan-Einot-Gabriel-Welsch (REGWQ) post-hoc tests for pairwisecomparisons of treatment groups. Survival from stocking until termi-nation of the experiment was assessed using a Chi-Square analysiswith the null hypothesis that survival rates should remain constantacross salinities. For gene expression data, normality was testedwith the Kolmogorov–Smirnov-test and homoscedasticity was testedwith the Fisher–Snedecor F test. Then, multiple comparisons amongstsalinity treatments and across time periods at the same salinity were

0.9

1.0

0.5 g/L 5.0 g/L 8.0 g/L 12.0 g/L a

a

418 J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

performed using one-way ANOVA and Tukey post-hoc tests for pair-wise comparisons of treatment groups. All hypotheses were testedat a significance level of α=0.05 and all tests were performed usingStatistical Analysis Software (SAS Institute Inc., Cary, NC, USA).

Mea

n g

/ fis

h

0.4

0.5

0.6

0.7

0.8

b

ab

3. Results

3.1. Water chemistry

Water chemistry parameters are reported in Table 1. Temperaturein the four adjacent systems was allowed to fluctuate with ambientlab conditions so inter-system temperatures would line up to thegreatest extent possible. Amongst the four systems, mean tempera-ture was 26.80±0.12 °C.

Days After Stocking0 14 28 42 56 70 84

Fig. 1. Growth of juvenile F. grandis stocked at one fish per L and fed a 32%-CP, 4% fatdiet for 12 weeks. Letters indicate significance at final sampling (REGWQ; P>0.05).

3.2. Growth, survival and condition

Data from biweekly sampling is detailed in Fig. 1. Growth isreported as mean change in g per fish sampled. Weight gain andSGR were significantly lower than all other treatment levels in the0.5‰ salinity while the 12.0‰ treatment had significantly highervalues in these parameters than the 5.0‰ treatment (Table 2). Surviv-al was lowest in fish reared at 0.5‰ and salinity was found to have asignificant effect on this parameter. Final condition factor (Kn) wassignificantly lower in the 0.5‰ treatment (Table 2), indicating thatfish from this salinity had lower body mass per unit length comparedto the experimental population.

3.3. Relative expression of NKA, NKCC1, and CFTR

Fish reared at a salinity of 0.5‰ had significantly higher NKAmRNA quantity at week 1. After week 1 there were no significant dif-ferences in relative expression of NKAmRNA amongst salinities. Fromweek 1 to week 3, fish from both the 0.5 and 12.0‰ treatmentsshowed a statistically significant decrease of NKA mRNA of ≈68%while fish from the intermediate salinities did not show a significantchange (Fig. 2).

The NKCC1 mRNA quantity was not significantly differentamongst salinities at week 1. In week 3 NKCC1 relative expressionvalues were lower in fish reared at 0.5 and 12.0‰ compared withfish from the 5.0 and 8.0‰ treatments, however only the differenceat 5.0‰ was statistically significant. Significant NKCC1 decrease of≈45% from week 1 to week 3 was observed in fish reared at 0.5‰.Over the same time period, less substantial non-significant decreaseswere observed in the higher salinities (Fig. 3). No significant differ-ences in relative expression of CFTR mRNA were observed amongstsalinities or over time (Fig. 4).

Table 1Mean water chemistry parameters throughout the 12-week growth trial.

Parameter Nominal salinity (‰)

0.5 5.0 8.0 12.0

Salinity (‰) 0.54±0.01 5.06±0.02 7.94±0.02 11.99±0.03DO (ppm) 7.01±0.05 6.93±0.05 6.76±0.06 6.67±0.05pH 8.24±0.04 8.22±0.05 8.33±0.02 8.12±0.04Hardness (mg/L)a 106.0±3.5 336.6±9.4 863.6±12.4 1047.3 ±16.3Alkalinity (mg/L)a 99.7±2.7 101.2±3.4 99.8±3.2 99.3±2.9TAN (mg/L) 0.62±0.04 0.64±0.03 0.71±0.04 0.69±0.03Nitrite (mg/L) 0.14±0.01 0.15±0.02 0.13±0.01 0.16±0.01

a Reported as CaCO3.

4. Discussion

4.1. Growth, survival, and condition

The growth, survival, and body condition data in this study dem-onstrate the indirect metabolic expense to F. grandis juveniles main-taining ion balance in a hyposmotic environment over extendedtime periods. We observed significantly reduced body condition fac-tor in a recirculating culture setting at low salinity and this effect ap-pears to persist for some marsh dwelling species in the wild. Sailfinmolly Poecilia latipinna and western mosquitofish Gambusia affinisshare habitat with F. grandis and individuals of these two species col-lected from coastal Louisiana marshes demonstrated significantly re-duced body condition factors at FW (0.53±0.09‰) compared withbrackish (9.32±0.35‰) sites (Martin et al., 2009). Osmoregulatoryability in F. grandis may vary by life-stage, as other work has shownsignificant effects of ontogeny on salinity tolerance in euryhaline tel-eost species (Watanabe et al., 1985; Varsamos et al., 2005; Bodinier etal., 2009a; Bodinier et al., 2010). In F. grandis the FW critical swim-ming speed experiments of Kolok and Sharkey (1997) and earlylife-stage (hatching to 2.5 weeks) survival and growth at various sa-linities work by Perschbacher et al. (1990) confirm the physiologicaleffects of hyperosmoregulation observed in the present study.Perschbacher et al. (1990) found no significant differences in hatch-ing percentage of F. grandis eggs across a salinity range from 0 to35‰ with the exception of a 10‰ trial where a fungus associatedwith that salinity group may have been the cause of low hatchrates. In the same study, larvae in the 0‰ salinity group experienced40% mortality from hatching until 2.5 weeks after hatching. This ratewas significantly less than larvae reared at 5.0‰, the next incrementalsalinity level. Kolok and Sharkey (1997) observed 40% mortality in

Table 2Growth, survival, and condition data for F. grandis fed 32% protein and 4% lipid over a12-week trial at different salinities. Letters denote statistical significance across param-eters (REGWQ; P>0.05).

Parameter Nominal salinity (‰)

0.5 5.0 8.0 12.0

Initial weight (g) 0.53±0.02a 0.51±0.02a 0.49±0.02a 0.46±0.02a

Final weight (g) 0.62±0.03b 0.79±0.05ab 0.84±0.05a 0.91±0.07a

Weight gain (g) 0.09±0.03c 0.28±0.03b 0.35+0.04ab 0.45±0.03a

SGR (%/day) 0.18±0.06c 0.51 ±0.04b 0.63±0.07ab 0.81±0.02a

Survival (%) 59.3 86.5 96.3 89.7Initial Kn 1.053a 1.010a 0.999a 0.989a

Final Kn 0.976b 1.091a 1.096a 1.104a

Weeks Post-Transfer

731

Rel

ativ

e N

KA

Exp

ress

ion

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.5 ppt5.0 ppt8.0 ppt12.0 ppt

bbb

a*

a*

aa

a

a

a

aa

Fig. 2. Relative expression of NKA in gills after direct transfer from 7.0‰ (n=5 for eachpoint). The reference gene is elongation factor EF1α. Letters denote significance acrosssalinities at the same time point while an asterisk (*) denotes significance within salin-ity across time points (Tukey; P>0.05).

Weeks Post-Transfer731

Rel

ativ

e C

FTR

Exp

ress

ion

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.5 ppt5.0 ppt8.0 ppt12.0 ppt

Fig. 4. Relative expression of CFTR in gills after direct transfer from 7.0‰ (n=5 foreach point). The reference gene is elongation factor EF1α. There was no statistically sig-nificant difference across salinities at a time point or within salinities across timepoints (Tukey; P>0.05).

419J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

FW acclimated F. grandis challenged in a critical swimming speed ex-periment while fish acclimated and exercised in brackish water(10‰) experienced no mortalities. In the current study, mean mortal-ity across the three 0.5‰ replicates was 40.7% over a 12-week period.

4.2. Intraspecific adaptations

Phenotypic and physiological differences may exist among a sub-set of individuals which hinder their ability to successfully maintainion balance in FW. Previous work has identified differential intraspe-cific adaptations to FW acclimation in juvenile D. labrax resulting inmortality for that proportion of individuals unable to properly coordi-nate ion conservation (Nebel et al., 2005). In the kidneys of non-acclimating D. labrax, NKA abundance and activity was reduced andfish were unable to produce urine hypotonic to their blood. D. labraxattempted to compensate for renal inability to conserve ions by in-creasing active absorption of ions through the gills. In this scenario,NKA abundance and activity increased in the gills relative to success-fully acclimating fish and an excessive proliferation of branchial MRCwas observed, possibly to the point of interfering with gas exchange(Nebel et al., 2005). It remains to be determined if specific consistent

Weeks Post-Transfer731

Rel

ativ

e N

KC

C E

xpre

ssio

n

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

0.5 ppt5.0 ppt8.0 ppt12.0 ppt

a

ab

b

b*

a

a

a

a

a

a

a

a

Fig. 3. Relative expression of NKCC1 in gills after direct transfer from 7.0‰ (n=5 foreach point). The reference gene is elongation factor EF1α. Letters denote significanceacross salinities at the same time point while an asterisk (*) denotes significance with-in salinity across time points (Tukey; P>0.05).

cellular level differences in ion regulation among a population subsetlead to reduced osmoregulatory ability in some F. grandis.

4.3. Critical salinity

In F. grandis larvae and juveniles a critical salinity appears to existsomewhere below 5.0‰. Whitehead et al. (2011) reported that F. het-eroclitus experienced an energetically taxing transition in salinitiesb1‰ which caused wild populations to adapt and reside in osmoticniches on either side of a≈0.5‰ salinity barrier. F. grandis in this exper-iment were part of a cultured population at the ARS. Parent fish of thiscultured population were taken from coastal Louisiana brackish watermarshes. Inland FW populations of F. grandis have been documented(Campbell et al., 1980) andfish from these populationsmayhave an im-proved ability to maintain internal ion concentrations in a hyposmoticenvironment as documented in F. heteroclitus (Whitehead et al.,2011). The present study produced satisfactory growth and body condi-tion in juvenile F. grandis at salinities of 5.0‰ or greater but a significantreduction in survival, SGR, and body condition occurred at 0.5‰. Fur-ther workwith F. grandis in the 0.5 to 5.0‰ salinity range could identifyan inflection point where manipulation of gill morphology and energyused on osmoregulation reduces the ability to effectively culture thisspecies.

4.4. Osmoregulatory gene expression

Much of the energy used in osmoregulation is devoted to synthe-sis of ion trafficking proteins and active maintenance of an electro-chemical gradient by NKA (Hwang and Lee, 2007). While cells ofthe kidney, intestines, and opercular epithelia are known to functionin ion transfer between animal and environment (Nebel et al., 2005;Scott et al., 2005, 2006) gene expression was examined in the gills be-cause this organ has been identified as the primary site of osmoregu-lation in juvenile and adult fishes (Katoh et al., 2000). NKA is thoughtto be responsible for the basolateral exit of Na+ from gill MRC duringFW hyperosmoregulation (Hwang and Lee, 2007). Due to its expres-sion in the dense mitochondria of MRC and roles in ion secretion inSW and basolateral exit of absorbed Na+ in FW, NKA can be used asan index or marker of ionoregulation in the gills (Evans et al., 2005;Bodinier et al., 2010; Ostrowski et al., 2011). In the present study rel-ative NKA expression was significantly higher at 0.5‰ relative to thehigher salinities at 7 d after direct transfer from 7.0‰. Scott et al.(2005) reported a statistically significant two-fold increase in activity

420 J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

of NKA in the gills of F. heteroclitus at 12 h, 3 d, and 7 d after transferfrom 10% SW to FW. In a separate study on F. heteroclitus, NKA wassignificantly upregulated by two-fold at 4 d post transfer from brack-ish water to FW and expression levels had dropped by 30 d posttransfer (Scott et al., 2004), a pattern consistent with observationsof F. grandis NKA expression in the present study. NKCC1 is generallythought to be the secretory isoform of the protein while NKCC2 orother forms may be active in Cl− uptake (Hwang and Lee, 2007;Hiroi et al., 2008). At week 1, relative expression of NKCC1 in F.grandis was not elevated or depressed in at any salinity. Marshall etal. (2002) suggested that redistribution of NKCC in F. heteroclitus isrelatively slow following a salinity transfer. Statistically significantdown-regulation of the gene encoding NKCC1 was observed at week3 for the 0.5‰ salinity as could be expected for an ion secretory pro-tein in the gills of a FW acclimating fish. Scott et al. (2004) observeddownregulation of NKCC1 after FW transfer in F. heteroclitus, althoughthe change in expression level occurred quickly and was more tran-sient than observed in the present study. CFTR has been immunoloca-lized to the basolateral membrane of MRC in freshwater killifishsuggesting a potential mechanism of Cl− exit into the blood(Marshall et al., 2002). However, CFTR is associated with hyposmore-gulation in most euryhaline and marine species and down-regulationof the gene or decreases in protein abundance upon FW transfer havebeen observed in D. labrax and Hawaiian goby, Stenogobius hawaiien-sis (McCormick, et al., 2003; Lorin-Nebel et al., 2006; Bodinier et al.,2009a;). In Fundulus, the method of Cl− regulation may be different.Scott et al. (2004) observed relocation of CFTR proteins within MRCbut comparable involvement in Cl− management after transfer frombrackish water (10‰) to both FW and SW. It is also possible thatthe absence of variation in expression of CFTR mRNA is due to the ex-istence of other chloride channels such as ClC-3 or the SLC26 family(Bayaa et al., 2009; Tang et al., 2010). The potential presence ofthese channels in F. grandismay mean that the activity of CFTR is lim-ited in hypotonic media and no changes in transcriptional activity ofthe genes encoding this protein should be expected. Alternatively,the lack of significant differences in CFTR expression across time orat different salinities may suggest that expression levels stabilize by7 d post transfer. For both CFTR and NKCC1, a reduced interval untilfirst sampling may have produced observable gene regulation results.It is possible that sampling fish 24 or 72 h after transfer would haveshown regulation in these genes, and that this regulation is completeat 7 d. Future work could utilize a reduced sampling interval to exam-ine possible expression changes for CFTR and NKCC1 in F. grandis.Across all three genes, the intermediate salinities of 5.0 and 8.0‰did not produce a statistically significant difference in expressionfrom each other or over time. Transfer from 7.0‰ to either of these in-termediate salinities may not represent a biologically significantchange in osmolarity strong enough to produce observable variationsin transcriptional regulation of the ion trafficking proteins examinedin the current study.

When analyzing relative expression of genes from fish in the 0.5‰salinity treatment it is necessary to consider the 40.7%mortality rate. Al-though mortalities were not individually recorded, they were observedthroughout the duration of the experiment. If differential intraspecificacclimation exists in F. grandis at near-FW conditions as discussed inSection 4.2 for juvenileD. labrax (Nebel et al., 2005), samplingwhich oc-curred at later dateswould have had an increased likelihood of choosingan individual with good ability to ionoregulate in FW as those with poorability weremore likely victims of mortality. Relative expression of NKAat 0.5‰ strongly downregulated fromweek 1 toweek 3 and stabilized atweek 7. Statistically significant downregulation was also observed inNKCC1 at 0.5‰ from week 1 to week 3 with stabilization at week 7.The general trend of stabilization in gene expression over time observedat 0.5‰ salinitymay be partially attributable tomortality of over expres-sing individuals attempting to compensate for an inability to maintainosmotic homeostasis for other reasons, such as renal ion excretion.

The present study examined RNA-level gene expression over anextended time period relative to that which has normally been testedin the past. Most researchers have tested short-term effects of salinityacclimation on osmoregulatory gene expression and supposed thatafter 7–14 d expression has stabilized and there are no furtherchanges. This study tested animals at 7 d but also at 21 d and 49 d.Changes which occurred from 7 d to 21 d cannot be definitively clas-sified as outside the 14 d maximum normally tested. However,changes observed between 21 d and 49 d would generate interestingquestions about stabilized regulation of relative gene expression overtime. Some differences in expression were observed in this later timeperiod but high levels of variation generally meant these differenceswere not statistically significant. It may be worthwhile for futurework to examine properties of relative gene expression beyond 14 d.

5. Conclusions

It is clear that F. grandis have the ability to hyperosmoregulate, butthe metabolic challenges of doing so produce adverse effects ongrowth. Elevated expression of NKA at 0.5‰ salinity shows how thismetabolic demand can manifest on a cellular and molecular level.Ecological considerations for this observation should focus on thecritical salinity discussed in Section 4.3. In Louisiana, where saltwaterintrusions are becoming more frequent and persistent (Shaffer et al.,2009), F. grandis populations which reside on the FW side of a≈0.5‰salinity barrier could be replaced by less FW tolerant conspecifics. In aculture situation without ready access to saline water a balance mustbe reached between the expenses of maintaining ideal water ion con-centrations and reduced growth observed as salinity decreases. Thisstudy casts doubt on the ability to successfully grow F. grandis at sa-linities of 0.5‰ or lower, although adaptive populations may existor may be possible to produce through selective breeding. Furtherwork should be able to determine a critical salinity below 5.0‰where culture is no longer feasible.

Acknowledgments

We would like to thank Dr. Jerome LaPeyre and Dr. Fernando Gal-vez for their technical assistance.

References

Armstrong, P.B., Child, J.S., 1965. Stages in the normal development of Fundulus hetero-clitus. Biol. Bull. 128, 143–168.

Bayaa, M., Vulesevic, B., Esbaugh, A., Braun, M., Ekker, M.E., Grosell, M., Perry, S.F., 2009.The involvement of SLC26 anion transporters in chloride uptake in zebrafish(Danio rerio) larvae. J. Exp. Biol. 212, 3283–3295.

Beamish, F.W.H., 1978. Swimming capacity. Fish Physiol. 7, 101–187.Bernatzeder, A.K., Cowley, P.D., Hecht, T., 2008. Effect of short term exposure to the an-

aesthetic 2-phenoxyethanol on plasma osmolality of juvenile dusky cob, Argytoso-mus japonicus (Sciaenidae). J. Appl. Ichthyol. 24, 303–305.

Bodinier, C., Boulo, V., Lorin-Nebel, C., Charmantier, G., 2009a. Influence of salinity onthe localization and expression of the CFTR chloride channel in the ionocytes ofDicentrarchus labrax during ontogeny. J. Anat. 214, 318–329.

Bodinier, C., Lorin-Nebel, C., Charmantier, G., Boulo, V., 2009b. Influence of salinity onthe localization and expression of the CFTR chloride channel in the ionocytes of ju-venile Dicentrarchus labrax exposed to seawater and freshwater. Comp. Biochem.Physiol. A 153, 345–351.

Bodinier, C., Sucré, E., Lecurieux-Belfond, L., Blondeau-Bidet, E., Charmantier, G., 2010.Ontogeny of osmoregulation and salinity tolerance in the gilthead sea breamSparus aurata. Comp. Biochem. Physiol. A 157, 220–228.

Bœuf, G., Payan, P., 2001. How should salinity influence fish growth? Comp. Biochem.Physiol. C 130, 411–423.

Campbell, S.L., Hilles, D.M., Milstead, E., 1980. Inland records for Fundulus grandis(Cyprinodontidae) in Texas. Southwest. Nat. 25, 271–272.

Clemment, T., Stone, N., 2004. Collection, removal, and quantification of eggs producedby rosy red fathead minnows in outdoor pools. N. Am. J. Aquacult. 66, 75–80.

Evans, D.H., Pierarini, P.M., Choe, K.P., 2005. The multifunctional fish gill: dominant siteof gas exchange, osmoregulation, acid–base regulation, and excretion of nitroge-nous waste. Physiol. Rev. 85, 97–177.

Flemmer, A.W., Monette, M.Y., Djurisic, M., Dowd, B., Darman, R., Gimenez, I., Forbush,B., 2010. Phosphorylation state of the Na+–K+–Cl− cotransporter (NKCC1) in the

421J. Patterson et al. / Comparative Biochemistry and Physiology, Part A 161 (2012) 415–421

gills of Atlantic killifish (Fundulus heteroclitus) during acclimation to water of vary-ing salinity. J. Exp. Biol. 213, 1558–1566.

Galvez, F., Franklin, N.M., Tuttle, R.B., Wood, C.M., 2007. Interactions of waterborne anddietary cadmium on the expression of calcium transporters in the gills of rainbowtrout: influence of dietary calcium supplementation. Aquat. Toxicol. 84, 208–214.

Gonzalez, I., Levin, M., Jermanus, S., Watson, B., Gilg, M.R., 2009. Genetic assessment ofspecies ranges in Fundulus heteroclitus and F. grandis in northeastern Florida saltmarshes. Southeast. Nat. 8 (2), 227–243.

Green, C.C., Gothreaux, C.T., Lutz, C.G., 2010. Reproductive output of Gulf killifish at dif-ferent stocking densities in static outdoor tanks. N. Am. J. Aquacult. 72, 321–331.

Hiroi, J., McCormick, S.D., 2007. Variation in salinity tolerance, gill Na+/K+-ATPase,Na+/K+/2Cl− cotransporter and mitochondria-rich cell distribution in three sal-monids Salvelinus namaycush, Salvelinus fontinalis, and Salmo salar. J. Exp. Biol.210, 1015–1024.

Hiroi, J., Yasumsu, S., McCormick, S.D., Hwang, P.-P., Kaneko, T., 2008. Evidence for anapical Na-Cl cotransporter involved in ion uptake in a teleost fish. J. Exp. Biol.211, 2584–2599.

Hwang, P.-P., Lee, T.-H., 2007. New insights into fish ion regulation andmitochondrion-rich cells. Comp. Biochem. Physiol. A 148, 479–497.

Hyndman, K.A., Evans, D.H., 2009. Effects of environmental salinity on gill endothelinreceptor expression in the killifish, Fundulus heteroclitus. Comp. Biochem. Physiol.A 152 (1), 58–65.

Inokuchi, M., Hiroi, J., Watanabe, S., Hwang, P.-P., Kaneko, T., 2009. Morphological andfunctional classification of ion-absorbing mitochondria-rich cells in the gills of Mo-zambique tilapia. J. Exp. Biol. 212, 1003–1010.

Ivanis, G., Esbaugh, A.J., Perry, S.F., 2008. Branchial expression of localization of SLC9A2and SLC9A3 sodium/hydrogen exchangers and their possible role in acid–base reg-ulation in freshwater rainbow trout (Oncorhynchus mykiss). J. Exp. Biol. 211,2467–2477.

Katoh, F., Shimizu, A., Uchida, K., Kaneko, T., 2000. Shift of chloride cell distributionduring early life stages in seawater-adapted killifish, Fundulus heteroclitus. Zool.Sci. 17 (1), 11–18.

Kneib, R.T., 1997. The role of tidal marshes in the ecology of estuarine nekton. Ocea-nogr. Mar. Biol. Annu. Rev. 35, 163–220.

Kolok, A.S., Sharkey, D., 1997. Effect of freshwater acclimation on the swimming perfor-mance and plasma osmolarity of the euryhaline Gulf killifish. Trans. Am. Fish. Soc.126, 866–870.

Kumaran, S., Lochmann, R., Stone, N., Kachowski, A., Lee, Y.-W., 2007. Effects of dietswith or without menhaden fish meal and oil on egg size, hatchability, and frysize for rosy red fathead minnow. N. Am. J. Aquacult. 69, 419–428.

Le Cren, E.D., 1951. The length-weight relationship and seasonal cycle in gonad weightand condition in the perch (Perca fluviatilis). J. Anim. Ecol. 20, 201–239.

Lin, Y.M., Chen, C.N., Lee, T.H., 2003. The expression of gill Na, K-ATPase in milkfish,Chanos chanos, acclimated to seawater, brackish water and fresh water. Comp. Bio-chem. Physiol. A 135, 489–497.

Lin, L.-Y., Horng, J.-L., Kunkel, J.G., Hwang, P.-P., 2006. Proton pump-rich cell secretesacid in skin of zebrafish larvae. Am. J. Physiol. Cell Physiol. 290, 371–378.

Lorin-Nebel, C., Boulo, V., Bodinier, C., Charmantier, G., 2006. The Na+/K+/2Cl− cotran-sporter in the sea bass Dicentrarchus labrax during ontogeny: involvement in os-moregulation. J. Exp. Biol. 209, 4908–4922.

Marshall, W.S., Singer, T.D., 2002. Cystic fibrosis transmembrane conductance regulatorin teleost fish. Biochim. Biophys. Acta 1566, 16–27.

Marshall, W.S., Lynch, E.M., Cozzi, R.R.F., 2002. Redistribution of immunofluorescenceof CFTR anion channel and KCC cotransporter in chloride cells during adaptationof the killifish Fundulus heteroclitus to sea water. J. Exp. Biol. 205, 1265–1273.

Martin, S.B., Hitch, A.T., Purcell, K.M., Klerks, P.L., Leberg, P.L., 2009. Life history varia-tion along a salinity gradient in coastal marshes. Aquat. Biol. 8, 15–28.

McCormick, S.D., Sundell, K., Björnsson, B.J., Brown, C.L., Hiroi, J., 2003. Influence of sa-linity on the localization of Na+/K+-ATPase, Na+/K+/2Cl− cotransporter (NKCC)and CFTR anion channel in chloride cells of the Hawaiian goby (Stenogobiushawaiiensis). J. Exp. Biol. 206, 4575–4583.

Nebel, C., Romestand, B., Nègre-Sadargues, G., Grousset, E., Aujoulat, F., Bacal, J., Bon-homme, F., Charmantier, G., 2005. Differential freshwater adaptation in juvenilesea-bass Dicentrarchus labrax: involvement of gills and urinary system. J. Exp.Biol. 208, 3859–3871.

Ostrowski, A.D., Watanabe, W.O., Montgomery, F.P., Rezek, T.C., Shafer, T.H., Morris Jr.,J.A., 2011. Effects of salinity and temperature on the growth, survival, whole bodyosmolality, and expression of Na+/K+-ATPase mRNA in red porgy (Pagrus pagrus)larvae. Aquaculture 314, 193–201.

Partridge, G.J., Lymberry, A.J., 2008. The effect of salinity on the requirement for potas-sium by barramundi (Lates calcarifer) in saline groundwater. Aquaculture 278,164–170.

Perry, S., 1997. The chloride cell: structure and function in the gills of freshwater fishes.Annu. Rev. Physiol. 59, 325–347.

Perschbacher, P.W., Aldrich, D.V., Strawn, K., 1990. Survival and growth of the earlystages of Gulf killifish in various salinities. Prog. Fish Cult. 52, 109–111.

Rodet, F., Lelong, C., Dubos, M.-P., Costil, K., Favrel, P., 2005. Molecular cloning of a mol-luscan gonadotropin-releasing hormone receptor orthologue specificallyexpressed in the gonad. Biochim. Biophys. Acta 1730, 187–195.

Rombough, P., 2007. The functional ontogeny of the teleost gill: which comes first, gasor ion exchange? Comp. Biochem. Physiol. A 148, 732–742.

Rozas, L.P., LaSalle, M.W., 1990. A comparison of the diets of Gulf killifish, Fundulusgrandis Baird and Girard, entering and leaving a Mississippi brackish marsh. Estu-aries 13, 332–336.

Ruijter, J.M., Ramakers, C., Hoogaars, M.H., Karlen, Y., Bakker, O., van den Hoff, M.J.B.,Moorman, A.F.M., 2009. Amplification efficiency: linking baseline and bias in theanalysis of quantitative PCR data. Nucleic Acids Res. 37 (6), e45.

Sampaio, L.A., Bianchini, A., 2002. Salinity effects on osmoregulation and growth of theeuryhaline flounder Paralichthys orbignyanus. J. Exp. Mar. Biol. Ecol. 269, 187–196.

Saoud, I.P., Kreydiyyeh, S., Chalfoun, A., Fakih, M., 2007. Influence of salinity on survival,growth, plasma osmolality and gill Na+-K+-ATPase activity in the rabbitfish Siga-nus rivulatus. J. Exp. Mar. Biol. Ecol. 348, 183–190.

Scott, G.R., Richards, J.G., Forbush, B., Isenring, P., Schulte, P.M., 2004. Changes in geneexpression in gills of the euryhaline killifish Fundulus heteroclitus after abrupt sa-linity transfer. Am. J. Physiol. Cell Physiol. C300–C309.

Scott, G.R., Claiborne, J.B., Edwards, S.L., Schulte, P.M., Wood, C.M., 2005. Gene expres-sion after freshwater transfer in gills and opercular epithelia of killifish: insightinto divergent mechanisms of ion transport. J. Exp. Biol. 208, 2719–2729.

Scott, G.R., Schulte, P.M., Wood, C.M., 2006. Plasticity of osmoregulatory function in thekillifish intestine: drinking rates, salt and water transport, and gene expressionafter freshwater transfer. J. Exp. Biol. 209, 4040–4050.

Shaffer, G.P., Wood, W.B., Hoeppner, S.S., Perkins, T.E., Zoller, J., Kandalepas, D., 2009.Degradation of baldcypress-water tupelo swamp to marsh and open water insoutheastern Louisiana, U.S.A.: an irreversible trajectory? J. Coastal Res. 54,152–156.

Simpson, D.G., Gunter, G., 1956. Notes on habitats, systematic characters, and life his-tories of Texas salt water cyprinodonts. Tulane Stud. Zool. 4, 114–134.

Tang, C.-H., Hwang, L.-Y., Lee, T.-H., 2010. Chloride channel ClC-3 in gills of the euryha-line teleost, Tetraodon nigroviridis: expression, localization and the possible role ofchloride absorption. J. Exp. Biol. 213, 683–693.

Tipsmark, C.K., Breves, J.P., Seale, A.P., Lerner, D.T., Hirano, T., Grau, E.G., 2011. Switch-ing of Na+, K+-ATPase isoforms by salinity and prolactin in the gill of a cichlid fish.J. Endocrinol. 209, 237–244.

Varsamos, S., Nebel, C., Charmantier, G., 2005. Ontogeny of osmoregulation in postem-bryonic fish: a review. Comp. Biochem. Physiol. A 141, 401–429.

Watanabe, W.O., Kuo, C.-M., Huang, M.-C., 1985. The ontogeny of salinity tolerance inthe tilapias Oreochromis aureus, O. niloticus, and an O. mossambicus x O. niloticus hy-brid, spawned and reared in freshwater. Aquaculture 47, 353–367.

Whitehead, A., Roach, J.L., Zhang, S., Galvez, F., 2011. Genomic mechanisms of evolvedplasticity in killifish distributed along an environmental salinity gradient. Proc.Natl. Acad. Sci. U. S. A. 108, 6193–6198.

Williams, D.A., Brown, S.D., Crawford, D.L., 2008. Contemporary and historical influ-ences on the genetic structure of the estuarine-dependent Gulf killifish Fundulusgrandis. Mar. Ecol. Prog. Ser. 373, 111–121.

Wood, C.M., Grosell, M., 2008. A critical analysis of transepithelial potential in intactkillifish (Fundulus heteroclitus) subjected to acute and chronic changes in salinity.J. Comp. Physiol. B 178, 713–727.

Wood, C.M., Laurent, P., 2003. Na+ versus Cl− transport in the intact killifish after rapidsalinity transfer. Biochim. Biophys. Acta 1618, 106–119.

Wood, C.M., Marshall, W.S., 1994. Ion balance, acid–base regulation, and chloride cellfunction in the common killifish, Fundulus heteroclitus— a euryhaline estuarine tel-eost. Estuaries 17 (1A), 34–52.

Yan, J.-J., Chou, M.-Y., Kaneko, T., Hwang, P.-P., 2007. Gene expression of Na+/H+ ex-changer in zebrafish H+-ATPase-rich cells during acclimation to low-NA+ andacidic environments. Am. J. Physiol. Cell Physiol. 293, 1814–1823.