draft...draft 5 84 for this study. we used atlantic salmon (salmo salar linnaeus), a primarily...

Draft

Do prior diel thermal cycles influence the physiological

response of Atlantic salmon (Salmo salar) to subsequent heat stress?

Journal: Canadian Journal of Fisheries and Aquatic Sciences

Manuscript ID cjfas-2016-0157.R1

Manuscript Type: Article

Date Submitted by the Author: 22-Jun-2016

Complete List of Authors: Tunnah, Louise; Mount Allison University,

Currie, Suzanne; Mout Allison University MacCormack, Tyson J.; Mount Allison University, Biology

Keyword: Heat Shock Proteins, METABOLISM < General, Oxygen Consumption, Salmonid, TEMPERATURE < General

https://mc06.manuscriptcentral.com/cjfas-pubs

Canadian Journal of Fisheries and Aquatic Sciences

Draft

1

Do prior diel thermal cycles influence the physiological response of Atlantic salmon (Salmo 1

salar) to subsequent heat stress? 2

3

LOUISE TUNNAH, SUZANNE CURRIE, TYSON J. MACCORMACK 4

5

Affiliations: 6

Louise Tunnah ([email protected]) 7

Department of Biology, Mount Allison University, Sackville, NB, Canada 8

9

Suzanne Currie ([email protected]) 10

Department of Biology, Mount Allison University, Sackville, NB, Canada 11

12

Tyson J. MacCormack ([email protected]) 13

Department of Chemistry and Biochemistry, Mount Allison University, Sackville, NB, Canada 14

15

Author for Correspondence: 16

Louise Tunnah 17

Department of Biology 18

Mount Allison University 19

Sackville, NB, E4L 1G7 20

Canada 21

Phone: (902) 830-4966 22

E-mail: [email protected] 23

Page 1 of 48



Draft

2

ABSTRACT 24

25

We designed two environmentally relevant thermal cycling regimes using monitoring data from 26

an Atlantic salmon river to determine if exposure to prior diel cycles stimulated protective 27

mechanisms (e.g. heat hardening), and/or resulted in physiological and cellular stress. Wild fish 28

were exposed to three days of diel cycling in the lab and then exposed to an acute thermal 29

challenge near their upper reported critical temperature. We measured routine metabolic rate 30

across the time course as well as indicators of physiological status (e.g. plasma glucose and 31

osmolality) and cellular stress (e.g. heat shock protein 70). We observed that thermal cycling 32

altered physiological and cellular responses, compared to an acute heat shock, but saw no 33

differences between cycling regimes. Unique temperature regime and tissue specific responses 34

were observed in heat shock protein induction, metabolites, haematology and osmotic indicators. 35

Routine metabolic rate was not affected by the thermal cycling and increased according to Q10 36

predictions. While we report unique physiological and cellular responses between all treatment 37

groups, we did not observe a clear indication of a heat hardening response.38

Page 2 of 48



Draft

3

INTRODUCTION 39

40

The temperature in natural ecosystems is rarely constant; real-world scenarios involve 41

varying magnitudes of diel temperature fluctuation, coupled with sporadic extreme events (Ma et 42

al. 2015). The dynamic nature of these cycles creates variability in temperature parameters such 43

as the mean, rate of change, maximum temperature achieved, magnitude of change and 44

accumulated heat exposure over time (i.e. thermal load). For aquatic ectotherms like fish, whose 45

body temperature is normally within 1 ºC of the surrounding water (Hazel and Prosser 1974), 46

temperature has critical and pervasive effects on biology. Due to logistical difficulties in 47

execution, and prolonged experimental duration, few studies examine the biological effects of 48

environmentally relevant diel thermal cycles. However, those studies that have been conducted 49

in fish show both the metabolic rate (e.g. Lyytikäinen and Jobling 1998; Beauregard et al. 2013, 50

Oligny-Hébert et al. 2015) and physiological responses (e.g. Mesa et al. 2002; Narum et al. 2013; 51

Eldridge et al. 2015) of thermally cycled fish are altered compared to those maintained at a 52

constant temperature. Whether these diel thermal cycles can enhance a fish’s thermal tolerance 53

and provide protection during a subsequent heat event is presently unclear. This idea has 54

parallels to heat hardening, where a short (minutes to hours) exposure to a sublethal temperature 55

increases survivorship following a subsequent more severe heat stress (Loeschcke and Hoffmann 56

2007) and is reminiscent of ischaemic preconditioning in mammalian (Murry et al. 1986) and 57

fish hearts (Gamperl et al. 2001). If exposure to prior thermal cycling does impact thermal 58

tolerance, then understanding both the nature of the cycle and the induced biological effects is 59

important and may enhance the efficacy of conservation, management, and remediation efforts. 60

Page 3 of 48



Draft

4

The range of temperatures tolerated by a fish is bounded by an upper and lower critical 61

temperature (Tcrit) within which lies an optimum temperature range (Topt) where performance is 62

maximal. Metabolic physiology is highly temperature-dependent due to the inextricable link 63

between temperature and the reaction rate of biological processes. Routine metabolic rate (RMR) 64

and maximal metabolic rate (MMR) increase almost exponentially with temperature; however, 65

MMR plateaus and may even decline at high temperatures (Eliason and Farrell 2016). All life 66

processes (e.g. foraging, growth, predator evasion, and reproduction) must occur within an 67

animal’s aerobic scope, which is defined as the difference between MMR and RMR at a given 68

temperature. Aerobic scope is generally considered to be widest at a fishes Topt and smallest 69

around Tcrit (Steinhausen et al. 2008), although this may not always be the case (Clark et al. 70

2013). A thermal stress response may occur if water temperatures stray outside of a species’ Topt, 71

and particularly if they approach either the upper, or lower Tcrit (for review see Currie and 72

Schulte 2014). Such increases in temperature result in increased levels of circulating 73

catecholamines (LeBlanc et al. 2011, 2012; Templeman et al. 2014) that mobilize energy 74

reserves and enhance oxygen delivery. From a cellular perspective, the induction of heat shock 75

proteins (HSPs) are an important adaptive component of the thermal stress response, maintaining 76

the structure and function of cellular proteins that may otherwise be compromised with exposure 77

to high temperatures. Overall, the heat shock response has been primarily studied following 78

exposure to a single heat shock; however, redband trout (Oncorhynchus mykiss gairdneri) 79

experiencing a diel thermal cycle show acclimation in expression of hsp mRNA over 6 weeks 80

(Narum et al. 2013). It is unclear how different diel thermal cycles will affect the heat shock 81

response, particularly at the biologically active protein level. This paucity of physiological 82

information on diel temperature cycling over multiple days in any fish species was this impetus 83

Page 4 of 48



Draft

5

for this study. We used Atlantic salmon (Salmo salar Linnaeus), a primarily cold-water adapted 84

species, as a model to assess indicators of metabolic, physiological, and cellular stress in 85

response to an experimental design that mimicked the naturally occurring diel temperature cycles 86

in this species’ river habitat during the summer months. 87

Freshwater habitats are home to approximately 40% of all identified fish species 88

(Lundberg et al. 2000) and are particularly vulnerable to climate change (Woodward et al. 2010). 89

The Miramichi River in New Brunswick, Canada, like many other temperate river systems, 90

experiences diel cycles in temperature and therefore provides an excellent model to investigate 91

the biological impacts associated with warming rivers. The Miramichi River is recognized for 92

having the largest Atlantic salmon run in eastern North America (DFO 2013); but population 93

numbers are at record lows (ICES 2015) with warming river temperatures in the summer months 94

thought to be a contributing factor to this overall population decline. The optimum temperature 95

range for Atlantic salmon is thought to span 6-20 ºC, within which maximal growth occurs at 16-96

17 ºC (Jensen et al. 1989). In Atlantic salmon acclimated to 15 ºC, the upper lethal temperature 97

(survival <10 min) is reported as 32 ºC (Elliot 1991). Data from ongoing conservation efforts 98

(e.g. Miramichi River Environmental Assessment Committee) for summer 2014 and 2015 shows 99

the average water temperature in July and August to be 21.0 ± 0.6 ºC, with peak temperatures of 100

27.9 ± 0.9 ºC. The rate of temperature change is often rapid, warming on average 0.43 ± 0.03 101

°C·h-1 (maximum 1.4 °C·h-1), thus allowing little time for acclimation. Diel thermal cycles are 102

not always constant, and the mean temperature can increase over several days if overnight 103

cooling is not sufficient before warming begins again the following day (see Figure 1B). 104

Exposure to these naturally occurring, repeated, daily fluctuations in temperature may provoke a 105

heat hardening response, whereby fish are better able to withstand subsequent, possibly more 106

Page 5 of 48



Draft

6

severe, temperature increases. Alternatively, the accumulated thermal load from these repeated 107

exposures could cause stress, damage, and depletion of energetic stores, as Atlantic salmon 108

juveniles are reported to cease feeding above 23 °C (Breau et al. 2011). 109

Some information on the biological responses of Atlantic salmon exposed to high 110

temperatures is available. Maximum aerobic metabolic rate in wild 2+ individuals occurs at ~ 24 111

ºC (540 mgO2·kg-1·h-1), and above this temperature muscle glycogen stores become depleted and 112

blood and muscle lactate levels increase (Breau et al. 2011). Parr removed from the Miramichi 113

River at 27 ºC had significantly higher whole body HSP70 protein and mRNA levels than fish 114

caught in the same location at 20 ºC indicating that these stress proteins are induced in the wild 115

(Lund et al. 2002). Thermal tolerance does appear to be plastic in Atlantic salmon, as Elliot 116

(1991) determined that increased acclimation temperature increased the upper lethal tolerance of 117

Atlantic salmon. Furthermore, Anttila et al. (2014) showed increases in maximum heart rate and 118

onset temperature of cardiac arrhythmias in Atlantic salmon reared at warmer (+ 8 ºC) 119

temperatures. Notably, the temperatures chosen in these two studies were held constant and 120

maintained for weeks to months. 121

The goal of this study was to assess whether recent exposure to environmentally relevant 122

diel thermal cycles influences the response to a subsequent acute high temperature event in 123

Atlantic salmon. Specifically, we were interested in whether or not two distinct, summer water 124

temperature scenarios involving diel cycling over multiple days stimulated protective 125

mechanisms (e.g. heat hardening), and/or resulted in physiological and cellular stress. One 126

thermal profile reflected a greater magnitude of temperature change and higher daily maximum 127

temperatures (diel cycle A) and one mimicked increasing mean temperatures and a higher 128

thermal load (diel cycle B). We compared these scenarios to an acute heat shock after a constant 129

Page 6 of 48



Draft

7

thermal history of 16 ºC as well as a 16 ºC negative control. While the metabolic rate of Atlantic 130

salmon was not differentially altered by either diel thermal cycle, we report unique physiological 131

and cellular responses between all treatment groups, but did not observe a clear indication of a 132

heat hardening response. 133

134

135

MATERIALS AND METHODS 136

137

Animal collection and holding conditions 138

139

Wild (1+) Atlantic salmon (Salmo salar) parr were collected using a Smith-Root LR-24 140

backpack electrofisher, in June, 2013, from the Cains River (N46°25’59.5; W066°01’29.1- 141

N46°26’05.0; W066°01’10.8 ± 15m), a tributary of the Southwest Miramichi River, New 142

Brunswick, Canada. At capture, body masses and fork lengths (FL) were approximately 5 g and 143

7 cm respectively. Fish were immediately transported to Mount Allison University, Sackville, 144

NB, and held in a 750 L tank of recirculating freshwater maintained at 16 ± 1.0 °C under natural 145

photoperiod conditions. Fish were initially fed a mixture of dehydrated krill and commercial 146

pellet feed (Corey Nutrition Company, Fredericton, NB; 1.0 mm) twice daily until satiation. 147

Subsequently, parr were weaned from krill and fed exclusively pellets to satiation once per day 148

for the remainder of holding. Feeding was ceased 24 h prior to experimentation to limit 149

elevations in metabolic rate due to specific dynamic action. Experiments were conducted 150

between November, 2014 and 2015, prior to which fish had been maintained at 16 ± 1.0 °C for at 151

least 10 months. All experimental fish (n = 28) had undergone smoltification as characterized by 152

Page 7 of 48



Draft

8

the absence of parr spots and were either 3+ or 4+ animals1. The FL and body mass of 153

experimental fish ranged between 26.5 – 41.5 cm and 165.5 – 802.1 g (these ranges are large due 154

to the extended experimental duration with this experimental design). All experimental protocols 155

were performed in accordance with the Canadian Council on Animal Care guidelines and 156

research was approved by the Mount Allison University Animal Care Committee (Protocol # 14-157

03). 158

159

Swim tunnel respirometry 160

161

All temperature manipulations and metabolic rate measures were performed individually 162

within a swim tunnel respirometer. The swim tunnel consisted of a 30 L measurement chamber, 163

within a larger 120 L chamber of aerated water (Loligo Systems, Tjele, Denmark), which was 164

continually passed through a 200 L aquarium bio- and charcoal- filter to prevent accumulation of 165

ammonia and nitrite wastes. Low levels of these waste products were verified periodically using 166

Nutrafin colourimetric kits (Hagen Inc., Canada), and the system was completely drained and 167

refilled between experimental animals. Water velocity, O2 concentration, and temperature were 168

continuously monitored and controlled by a DAQ-M data acquisition instrument and AutoResp 169

software (Loligo Systems). The water velocity within the measurement chamber was calibrated 170

using a vane wheel probe and handheld flow sensor (Höntzch, Waiblingen, Germany) and a solid 171

blocking correction factor was applied to correct water velocity to body lengths per second 172

(BL·s-1) for each experimental fish. Dissolved O2 content was measured using a dipping probe 173

mini-sensor (PreSens, Regensburg, Germany) and was calibrated using a 2-point calibration in 174

1 See supplemental table S1

Page 8 of 48



Draft

9

aerated tank water (100%), and a solution of sodium sulphite (0%). Temperature within the 175

measurement chamber was monitored using a platinum resistance temperature probe. During the 176

flush phase a 57 L·min-1 submersible aquarium pump, controlled by the DAQ-M, was used to 177

exchange the water inside the measurement chamber with aerated water from the reservoir. This 178

pump was turned off during the closed loop measurement phase. Ambient temperature in the 179

swim tunnel was maintained by continuously circulating water through a chiller (Aqua Logic 180

Inc, San Diego, USA). To manipulate temperature for each experimental condition, a 120 L hot 181

water reservoir was created using a rod immersion heater (Cole-Parmer, Montreal, Canada) on a 182

floating platform. A 10 L·min-1 submersible pump also controlled by the DAQ-M pumped hot 183

water into the ambient reservoir of the swim tunnel. By opposing action of the chiller and this 184

hot water pump, the temperature within the swim tunnel could be increased or decreased in a 185

controlled and reproducible manner. 186

187

Experimental procedure 188

189

Fish were removed individually from the holding tank and lightly sedated in buffered 190

anaesthetic solution (125 mg·L-1 tricaine methanesulfonate, 250 mg·L-1 NaHCO3) to measure 191

body mass, FL, body depth, and width (to the nearest 0.5 cm). An initial blood sample (0.5 mL; t 192

= 0 h) was drawn via caudal puncture using a 23 gauge needle and syringe rinsed with 193

heparinised salmonid saline to prevent clotting (in mmol·L-1: 124 NaCl, 5.1 KCl, 1.9 MgSO4, 1.5 194

Na2HPO4, 11.9 NaHCO3, CaCl2, 5.6 glucose, 100 U·mL-1 heparin). Salmon were then transferred 195

to the darkened swim tunnel respirometer and allowed to recover overnight on a constant flush 196

cycle at a water speed of 0.5 BL·s-1. In a subset of five fish, O2 consumption was monitored 197

Page 9 of 48



Draft

10

overnight to ensure sufficient recovery time before commencing experiments2. Following 198

recovery, fish were subjected to one of four experimental thermal regimes (see below and Table 199

1) and O2 consumption was measured over 84 h. Salmon were then removed from the swim 200

tunnel, sedated, and a second blood sample (~2 mL) taken in the same manner described above. 201

White muscle, heart, liver, and gill tissues were sampled following transection of the spinal cord, 202

flash frozen in liquid nitrogen, and stored at -80 ºC for subsequent analysis. 203

204

Thermal profiles 205

206

Two environmentally relevant thermal profiles were developed to mimic summer water 207

temperature scenarios in the Miramichi River based upon monitoring data from the Miramichi 208

River Environmental Assessment Committee (Fig. 1). These profiles are referred to as cycle A 209

and B throughout the text. Cycle A mimics a river water scenario where the temperatures 210

oscillate daily by the same magnitude. In this diel cycle, the change in temperature (∆T) is 8 °C, 211

fluctuating between 16 and 24 °C and ending with a temperature ramp to 27 °C over 12 h (∆T 11 212

°C), with a cumulative thermal load, or the area under the curve, of 349 °C·h (Table 1). Cycle B 213

represents a scenario with a smaller daily magnitude of change (∆T between 3.5 and 6 °C), and a 214

larger cumulative thermal load of 415 °C·h (Table 1), mimicking river conditions shown in Fig. 215

1B. To compare these thermal profiles with other studies on thermal stress, we conducted an 216

acute, single heat shock as a positive control, in which temperature was ramped from 16 to 27 °C 217

over a 12 h period. Finally, a negative control experiment was performed in which temperature 218

was maintained at 16 °C for 84 h (Table 1). 219

2 See supplemental Figure S1

Page 10 of 48



Draft

11

220

Oxygen consumption measurements 221

222

Routine mass specific metabolic rate (ṀO2) was measured through intermittent closed 223

loop respirometry. Oxygen consumption was calculated from the decline in dissolved oxygen 224

content within the measurement chamber of the swim tunnel over 300 s after a 30 s wait period. 225

Between measurement periods the chamber was flushed with aerated water from the reservoir for 226

570 s to return oxygen levels to approximately 100% saturation. One full flush/measure cycle 227

therefore took 15 min, allowing 4 ṀO2 readings per hour across the 84 h experimental period. 228

Autoresp software (Loligo Systems) automatically calculated ṀO2 accounting for fish mass, 229

water temperature, salinity, and barometric pressure. To account for background microbial 230

oxygen consumption, oxygen depletion was monitored in an empty chamber, both before and 231

after a fish was placed in the setup. The slope of background oxygen depletion was always 232

negligible and therefore metabolic rates are reported uncorrected. Any ṀO2 values generated 233

from an oxygen depletion slope with an R2 of less than 0.7 were removed in the interest of data 234

integrity. Additionally, any individual ṀO2 values 25% greater than the determined maximum 235

metabolic rate (see below) were considered outliers and were removed (n = 18 out of 236

approximately 7300 total measurements). 237

238

Maximal metabolic rate 239

240

In a separate group of four fish (679.9 ± 52.1 g), maximum and routine metabolic rates 241

were determined at 16°C to derive aerobic scope at this temperature. Fish were individually 242

Page 11 of 48



Draft

12

placed in the swim tunnel respirometer and allowed to acclimate overnight using the same 243

intermittent-loop respirometry parameters detailed above. Fish were not anesthetized, and no 244

preliminary blood sample was taken from this group of animals. Routine metabolic rate was 245

determined from the average of the final 10 recorded ṀO2 values the following morning. Each 246

fish was then moved to a separate tank and chased to exhaustion (determined as a lack of 247

response to handling) over a period of 5 – 10 min. Each fish was then quickly transferred back to 248

the swim tunnel respirometer and sealed in a closed loop. Respirometry measurements 249

commenced within ~1 min. O2 depletion was continuously recorded over a period of 5 min, and 250

an ṀO2 value was automatically calculated using AutoResp software every 30 s over this 251

measurement period. Maximum metabolic rate was determined by averaging the three highest 252

ṀO2 values immediately following this exhaustive exercise protocol. Aerobic scope was 253

calculated as the difference between MMR and RMR. 254

255

Blood and plasma analyses 256

257

Within 5 min of blood sampling, haematocrit (Hct, %), and haemoglobin concentration 258

(Hb, g·L-1) in whole blood were measured using a SpinCrit Microhaematocrit centrifuge 259

(SpinCrit, Indiana, USA), and HemoCue® Hb 201+ system (Ängelholm, Sweden) according to 260

the manufacturer’s instructions. A correction factor for fish blood was applied to all Hb values as 261

per Clark et al. (2008). Mean corpuscular haemoglobin concentration (MCHC, g·L-1) was 262

subsequently calculated from these two parameters. The remaining blood sample was then spun 263

at 14,000 g for 5 min at 4°C, any buffy coat was discarded, and the plasma supernatant and RBC 264

pellet were flash frozen separately in liquid nitrogen and stored at -80 °C. Plasma was 265

Page 12 of 48



Draft

13

subsequently analysed to determine osmolality (mmol·kg-1) using a Wescor Vapro 5520 Vapour 266

Pressure Osmometer (Logan, Utah, USA), and chloride ion concentration (mmol·L-1) using a 267

Chloride Analyzer 925 (Nelson Jameson Inc., Marshfield, WI, USA). Plasma samples were 268

deproteinized by diluting 1:2 (v:v) with 6% perchloric acid (PCA), and centrifuging at 13,000 269

rpm for 3 min at 4 °C. The glucose concentration in the supernatant of each sample was 270

determined relative to a standard curve of D-glucose in a combined assay solution with final in 271

well concentrations: 100 mmol·L-1 imidazole, 5 mmol·L-1 MgCl2, 4 mmol·L-1 ATP, 0.32 272

mmol·L-1 NADP+, 1 U glucose-6-phosphate dehydrogenase, 1 U hexokinase. The reaction was 273

allowed to run to completion at 37°C and the absorbance of generated NADPH measured at 340 274

nm using a SpectraMax M5 Plate Reader and SoftMax Pro software (Molecular Devices, 275

Sunnyvale, USA). 276

277

Tissue lactate 278

279

Lactate was quantified in plasma, heart, and liver tissues as an indication of anaerobic 280

metabolism. Ground heart, and liver samples were diluted 1:4 (w:v) in 8% PCA containing 1 281

mmol·L-1 EDTA using a PowerGen 125 homogenizer (Fisher Scientific, Ottawa, Canada). 282

Plasma samples were diluted 1:2 (v:v) in 8% PCA and vortexed to mix. All samples were then 283

centrifuged at 13,000 rpm for 4 min at 4 °C to pellet cell debris and protein. A known volume of 284

supernatant was removed and neutralized using a solution of 2 mmol·L-1 KOH and 0.4 mmol·L-1 285

imidazole. Samples were centrifuged again at 13,000 rpm at 4 °C for 1 min and the resulting 286

supernatant removed for lactate quantification. Samples were analysed in triplicate relative to a 287

standard curve of L-lactic acid. Final in-well concentrations of all assay components were: 0.16 288

Page 13 of 48



Draft

14

mol·L-1 glycine, 0.13 mol·L-1 hydrazine, 1.9 mol·L-1 NAD+, 10 U lactate dehydrogenase. The 289

reaction was allowed to run to completion at room temperature, after which absorbance of 290

generated NADH was read at 340 nm using a SpectraMax M5 Plate Reader and SoftMax Pro 291

software. 292

293

Na+/K+-ATPase Activity 294

295

Gill Na+/K+-ATPase activity (NKA; µmol·mg total protein-1·h-1) was measured as an 296

indicator of ion regulation. The protocol was as per McCormick (1993) with the exception that 297

gill tissue used for analysis had been frozen and ground over liquid nitrogen (see below). 298

299

Protein extraction and quantification 300

301

Gill, heart, liver, and white muscle tissues were ground over liquid nitrogen and diluted 302

10-15x (w:v) in protein extraction buffer (50 mmol·L-1 Tris base, 2% SDS; pH 7.5) with 0.05 303

µL·mg tissue-1 protease inhibitor cocktail (PIC004.1; Bioshop, Ontario, Canada). Samples were 304

then homogenized using a Fisher Scientific PowerGen 125, and centrifuged at 14,000 g for 10 305

min at 4 °C, and the supernatant removed and stored at -80°C. Each RBC pellet was diluted 1:1 306

(v:v) in heparinised salmonid saline with a 1x working concentration of protease inhibitor 307

cocktail (PIC003.1; BioShop ) and transferred to a FastPrep MatrixD cell lysing tube (MP 308

Biomedical, California, USA) containing 1.4 mm ceramic beads. Tubes were shaken for 45 s at 309

4.0 m·s-1 to lyse cells, and then cooled on ice. Tubes were centrifuged at 16,000 g for 3 min at 4 310

°C to pellet the beads and the soluble protein supernatant was removed and stored at -80 °C. 311

Page 14 of 48



Draft

15

Soluble protein concentrations in all samples were assayed using a DC Protein Assay kit (Bio-312

Rad, Mississauga, ON, Canada). Samples and standards (bovine serum albumin; Bio-Rad) were 313

diluted in protein extraction buffer or saline as appropriate and their absorbance read in triplicate 314

at 750 nm with a SpectraMax M5 plate reader and SoftMax Pro software. 315

316

Heat shock protein immunoblotting 317

318

As an indicator of cellular stress, we measured HSP70 in salmon RBCs, and tissues. 4 µg 319

of soluble protein sample was loaded per well, and for quantification purposes a four-point 320

standard curve of ‘supersample’ containing a mixture of all tissues from a representative fish was 321

run on each gel. Electrophoresis was performed using Bolt™ 4-12% Bis-Tris mini gels 322

(ThermoFisher Scientific, Ontario, Canada) and MOPS SDS running buffer (ThermoFisher 323

Scientific). Gels were run at 200V and separated proteins were transferred to a polyvinylidene 324

difluoride (PVDF) membrane at 20V using the Bolt™ Mini Blot module and transfer buffer (25 325

mmol·L-1 bicine, 25 mmol·L-1 bis-tris, 1 mmol·L-1 EDTA; pH 7.2). Membranes were then 326

blocked for 1 h at room temperature, or overnight at 4 °C, in 5% milk powder dissolved in Tris-327

buffered saline with Tween 20 (TBS-T; 25 mmol·L-1 Tris, 138 mmol·L-1 NaCl, 0.1% Tween 20, 328

pH 7.6). Membranes were incubated in polyclonal rabbit affinity purified HSP70 primary 329

antibody (AS05 061A, Agrisera, Vännäs, Sweden) at a concentration of 1:10,000 in 1% milk 330

powder TBS-T solution for 1 h at room temperature. This antibody is specific to the inducible 331

isoform of salmonid HSP70, and does not detect the constitutive protein3 and therefore provides 332

a powerful tool to assess induction of HSP70 with thermal stress. Secondary antibody (goat anti-333

3 See supplemental figure S2 for representative western blot images

Page 15 of 48



Draft

16

rabbit IgG-HRP; SAB 300, Enzo Life Sciences, New York, USA) was also used at 1:10,000 in 334

1% milk powder TBS-T solution for 1 h at room temperature. Chemiluminescent detection of 335

protein bands was performed using ECL Select (GE Healthcare, Buckinghamshire, UK). Blots 336

were imaged using a VersaDoc™ MP 4000 Molecular Imager (Bio-Rad) and analysed using 337

Image Lab® software. HSP70 in each sample was interpreted relative to a fit of the 338

‘supersample’ curve. Equal protein loading was visually assessed by Coomassie staining PVDF 339

membranes following immunodetection as per Welinder and Ekblad (2011). 340

341

Protein turnover / damage 342

343

Tissue and RBC ubiquitin was measured as an indirect indicator of protein turnover 344

and/or damage, using dot blots. Sample protein (0.2 µg·µL-1) and a standard curve of 345

‘supersample’ (4x – 0.25x) was loaded onto a nitrocellulose membrane (Bio-Rad) and blocked in 346

5% BSA/TBS-T. Membranes were probed for ubiquitin with a mouse primary antibody (BML-347

PW8805-0500, Enzo Life Sciences; 1:2,500 in 5% BSA/TBS-T), which detects 348

polyubiquitinylated protiens (i.e. those targeted to the proteosome for degradation) but not mono 349

or free ubiquitin. The secondary antibody was a goat anti-mouse IgM (ab97230, AbCam Inc, 350

California, USA; 1: 20,000 in 0.1% BSA/TBS-T). Levels of ubiquitin in each sample were 351

visualized and quantified as detailed for immunoblotting above. Equal protein loading was 352

visually assessed via staining with Ponceau S following immunodetection. 353

354

Statistical analyses 355

356

Page 16 of 48



Draft

17

All statistical analyses were performed using R-studio (version 3.2.1). All experimental 357

data were assessed, and appropriately transformed when required, to meet the parametric 358

assumptions of homogeneity of variance (Levene’s test), and normality of residuals (Shapiro-359

Wilk test). Tissue HSP70 and ubiquitin data were analysed separately using two-way ANOVAs 360

(tissue - 5 levels; treatment – 4 levels). However, due to a statistically significant tissue x 361

treatment group interaction, the data for both parameters were split by tissue to analyse the 362

response between treatments in each tissue separately using a one-way ANOVA (α-critical = 363

0.05). Fish masses, gill NKA, ṀO2 at 27 ºC, and tissue lactate concentrations were also analysed 364

using one-way ANOVAs to compare between treatment groups (α-critical = 0.05). When a 365

significant effect was detected, Tukey post-hoc tests were performed to assess where treatment 366

groups differed from one another. 367

All variables where repeated samples were taken over time (plasma osmolality, chloride, 368

glucose, lactate, ṀO2 at 21 ºC, Hb, Hct, and MCHC) were analysed using a split-plot ANOVA to 369

compare the effects of fish nested within treatment group (between-subjects factor), and 370

sampling time point (within-subjects factor). When an overall significant effect of time was 371

detected, a Tukey’s post-hoc test was used (α-critical = 0.05). When a significant interaction was 372

detected, data were split to compare both factors across treatments (4 levels, randomized block 373

ANOVA with Tukey post-hoc tests), and over time (2 levels, two-tailed paired t-test); the α-374

critical level for both analyses was therefore adjusted to 0.025. 375

376

377

RESULTS 378

379

Page 17 of 48



Draft

18

Rate of O2 consumption 380

381

The rate of O2 consumption was measured using intermittent-loop respirometry to 382

characterize the effects of four different thermal profiles on aerobic metabolic rate (Fig. 2). 383

Metabolic rate data were not scaled to the mean mass of all fish used in the study as we found no 384

relationship between ln fish mass and ln ṀO2 (R2 = 0.003) at a given temperature. Similarly, we 385

found no correlation between fish mass and either: osmolality, chloride, glucose, lactate, or 386

HSP70 in any tissue. Predictably, ṀO2 increased and decreased with increases and decreases in 387

water temperature, respectively. In the negative control group, ṀO2 remained constant across the 388

84 h experimental period at 97.4 ± 0.6 mgO2·kg-1·h-1, a value consistent with those obtained in 389

other studies on salmon (Deitch et al. 2006; Eliason and Farrell 2016). Baseline ṀO2 was similar 390

in the positive control group until 72 h (107.5 ± 0.8 mgO2·kg-1·h-1), and then increased with 391

temperature, peaking at 27 °C (268.3 ± 21.5 mgO2·kg-1·h-1). In diel thermal cycle A, the highest 392

ṀO2 readings coincided with the daily temperature peak. In this group, variability between 393

individuals was low, and ṀO2 peaked at 264.5 ± 16.1 mgO2·kg-1·h-1 when the temperature was 394

27 °C. In diel thermal cycle B, variability between individuals was higher, making the overall 395

ṀO2 response harder to interpret; however, it appears the daily peak in ṀO2 occurs before the 396

temperature peak. At 84 h and 27 °C, ṀO2 was 281.5 ± 23.3 mgO2·kg-1·h-1. The ṀO2 values at 397

the end of the experiment (at 27 °C) in each treatment group were not significantly different from 398

one another (p = 0.83; Fig. 3A). When ṀO2 at a common temperature of 21 °C was examined 399

within the two diel cycling groups over days 1 – 4, no differences were observed (p = 0.73). 400

However, overall, fish in cycle B had significantly higher ṀO2 at 21 °C than those in cycle A (p 401

= 0.02; Fig 3B). This 21 °C temperature was chosen as it was the highest temperature common to 402

Page 18 of 48



Draft

19

all treatment groups across all four experimental days. 403

We also determined aerobic scope in a subset of Atlantic salmon at 16 °C. An aerobic 404

scope of 278.4 ± 3.9 mgO2·kg-1·h-1 was determined from a resting ṀO2 of 81.7 ± 3.5 mgO2·kg-405

1·h-1 and a maximal value following exhaustive exercise of 360.3 ± 7.3 mgO2·kg-1·h-1. 406

407

Plasma and tissue metabolites 408

409

We found a highly significant treatment group x time interaction (p < 0.001) in plasma 410

glucose suggesting that responses were distinct among temperature treatments (Table 2). Plasma 411

glucose significantly decreased over time (from t = 0 to t = 84 h) in the negative control group (p 412

= 0.014), and significantly increased in cycle A (p = 0.005). At t = 0 h plasma glucose was the 413

same across all treatment groups (p = 0.30), but at t = 84 h it was significantly elevated in the 414

positive control (p < 0.001), cycle A (p < 0.001), and cycle B (p = 0.006) compared to the 415

negative control. 416

Neither heart (p = 0.36), nor white muscle (p = 0.26) lactate concentration significantly 417

changed across treatments (Fig. 4). However, as with plasma glucose, there was a significant 418

treatment x time interaction (p < 0.001) in plasma lactate (Table 2). The concentration of lactate 419

in plasma at t = 84 h was significantly elevated above the negative control in both positive 420

control (p < 0.001) and cycle B (p = 0.020) treatments. In fish subjected to thermal cycle B, 421

plasma lactate increased from t = 0 h to t = 84 h (p = 0.005), while approaching significance (p = 422

0.029, α-critical = 0.025) in fish subjected to thermal cycle A. 423

424

Ionoregulation 425

Page 19 of 48



Draft

20

426

To assess whether exposure to elevated temperatures led to osmotic stress, we measured 427

plasma osmolality and plasma chloride concentrations (Table 2). Plasma osmolality changes 428

were dependent on thermal group, as indicated by a significant treatment x time interaction (p = 429

0.003). Plasma osmolality remained unchanged over time (from t = 0 h to t = 84) in the negative 430

control (p = 0.41), but increased significantly in the positive control (p = 0.007), thermal cycle A 431

(p = 0.007), and thermal cycle B (p = 0.005) treatments. Initially, at t = 0 h, plasma osmolality 432

was the same in all fish across thermal groups (p = 0.062). However, after exposure to 27 °C at t 433

= 84 h, plasma osmolality was significantly elevated above negative control levels in both the 434

positive control and cycle A (p < 0.001) fish. The plasma osmolality of fish in cycle B was not 435

significantly different from any other treatment group (p > 0.05). Plasma chloride concentration 436

was the same in all fish, across all treatment groups at both t = 0 h (p = 0.74) and at t = 84 h (p = 437

0.038; α-critical = 0.025). In thermal cycle B, plasma chloride concentration was significantly 438

higher than t = 0 h levels at the end of the experiment (t = 84 h; p = 0.004). 439

440

Haematology 441

442

Neither Hb or Hct was significantly different among treatment groups, but both did 443

decrease significantly over time from t = 0 h to t = 84 h in the negative control (Hb - p < 0.001; 444

Hct - p = 0.01), and thermal cycle A (Hb - p = 0.023; Hct - p = 0.028) groups. Changes in both 445

Hb and Hct approached significance in the positive control group (p = 0.066 and p = 0.055 446

respectively). We calculated MCHC from Hb and Hct (Table 2) and did not observe significant 447

differences among any of the temperature groups (p = 0.92), nor did MCHC change within any 448

Page 20 of 48



Draft

21

treatment group from t = 0 h to t = 84 h (p = 0.95). 449

450

Cellular Stress and Damage 451

452

The expression of indicators of cellular stress (HSP70) and damage (ubiquitin) were 453

highly tissue-specific as shown by a significant treatment group x tissue type interaction (p < 454

0.001) for both variables. Consequently, the data were split to analyse the stress response in each 455

tissue separately (Fig. 5). HSP70 was undetectable in the negative control group (16 ºC) and for 456

all tissues at t = 0 (16 ºC) regardless of treatment group (data not shown). In the positive control, 457

HSP70 was induced in white muscle, heart, and liver (Fig. 5 E, G, I) but not in RBCs or gill (Fig 458

5 A, C). HSP70 was induced in all tissues in both diel thermal cycling groups with the same 459

magnitude of induction (p > 0.05). The induction of HSP70 in gill, white muscle, and RBC was 460

greatest in diel cycles A and B (Fig 5A, E, G). In comparison, the greatest HSP70 increase in 461

heart and liver occurred in the positive control group (Fig. 5G, I). 462

We measured ubiquitin for a gross indication of whether the HSP70 induction correlated 463

with changes in protein damage/turnover. We observed increases in ubiquitin in at least one of 464

the treatment groups in all tissues except liver (p = 0.35, Fig. 5J). In gill, only thermal cycle A 465

resulted in a significant increase in ubiquitin compared to the negative control (p = 0.006, Fig. 466

5B). In the positive control group, the acute heat shock resulted in a significant increase in 467

muscle, RBC, and heart ubiquitin (p < 0.05, Fig. 5D, F, H). As was observed with HSP70, the 468

magnitude of ubiquitin expression in both diel thermal cycling groups was statistically similar (p 469

> 0.05) in all tissues, with the exception of white muscle where cycle B was significantly higher 470

(p = 0.031) than cycle A. 471

Page 21 of 48



Draft

22

472

473

DISCUSSION 474

475

Although the response of fishes to increases in temperature has been extensively studied 476

in a variety of species, little work has focused on environmentally relevant diel thermal cycles, 477

particularly over multiple days. The goal of this study was to use Atlantic salmon (Salmo salar) 478

as a model to investigate ecologically relevant thermal profiles and assess their effects on 479

indicators of metabolic, physiological, and cellular stress. We designed our thermal profiles 480

using data from the Miramichi River and used wild fish caught from the same location to 481

ecologically ground our laboratory experiments. Two likely outcomes from exposure to repeated 482

fluctuations in temperature are possible and not necessarily mutually exclusive: 1) a protective 483

heat hardening response that increases thermal tolerance, or 2) a detrimental accumulation of 484

heat exposure-related stress, cellular damage, and depletion of energetic stores. Salmon in the 485

Miramichi River survive at mean summer water temperatures (21 ºC) which exceed the reported 486

upper optimal temperature for salmon (20 ºC; Jensen et al. 1989), thus enhanced thermal 487

tolerance due to chronic exposure to such sub-optimal thermal conditions seems plausible. 488

Conversely, Atlantic salmon population numbers are currently well below historical levels, 489

suggesting a potential detrimental impact of high temperature exposures. We did not find clear 490

evidence of a heat hardening response, nor did exposure to any thermal treatment group appear 491

detrimental or induce a severe stress response, at least over the time course tested. It is possible, 492

although currently unexamined in the literature, that a longer period of thermal cycling may 493

induce a heat hardening response in this species. As predicted, we did observe clear differences 494

Page 22 of 48



Draft

23

in the physiological and cellular responses of fish exposed to diel thermal cycling in comparison 495

to fish kept at a stable temperature. Given that these two different diel thermal cycles had similar 496

overall mean temperatures (Table 1), we are confident that our findings are a result of 497

temperature cycling and not differences in mean temperature. We observed unique responses in 498

some physiological parameters with thermal cycling, although the changes noted were subtle. 499

This suggests that thermal cycling or temperature variance is more important than the nature of 500

the diel cycling (e.g. accumulated thermal exposure, magnitude, rate of temperature change, or 501

mean), at least with regard to the two different cycling profiles tested here. 502

In general a 10 ºC increase in temperature doubles or triples RMR in fish (i.e. a Q10 of 2-503

3; Reid et al. 1997). Increasing aerobic scope either through decreased RMR and/or increased 504

MMR increases the amount of available energy and thus provides an overall benefit to the fish. 505

We saw no evidence that a prior exposure to environmentally relevant thermal cycles could 506

lower RMR, and thus increase aerobic scope, during a subsequent acute thermal shock. At 27 ºC 507

(84 h), RMR was statistically similar in all groups (Fig. 4A), and was also the same on days 1 - 4 508

at a common temperature of 21 ºC in diel cycles A and B (Fig. 4B). The Q10 of ṀO2 for the 12 h 509

ramp to 27 ºC on day 4 did not vary either, fluctuating only from 2.1 to 2.5 between experimental 510

groups. Therefore, it appears that RMR is not modulated by diel thermal cycling, and that Q10 511

effects are the primary determinant of RMR under the conditions tested. We did notice that the 512

variability in ṀO2 between individuals was higher in diel cycle B; seasonal effects offer a 513

potential explanation. Evans (1984) determined RMR to be maximal in the spring and to shift 514

independently from changes in environmental temperature. Notably, four of six individuals in 515

diel thermal cycle B were tested in the spring. 516

Aerobic scope can be expressed as an absolute value, or as a ratio termed factorial 517

Page 23 of 48



Draft

24

aerobic scope (FAS = MMR/RMR). A FAS value ≤ 1 indicates MMR has been reached or 518

exceeded, and the fish is relying on anaerobic processes to fuel function; the duration of survival 519

under such conditions is limited (Portner and Knust 2007). Limitations in fish availability only 520

allowed routine and maximum metabolic rate to be determined in the same individual (n = 4) at 521

one temperature (16 ºC), and FAS in this group was 4.4 ± 0.1. If we use the MMR at 16 ºC to 522

calculate FAS at 27 ºC in the positive control and both diel thermal cycling groups, FAS drops to 523

~1.4. This is almost certainly an underestimate of actual FAS since MMR (the numerator) also 524

increases with temperature, although to an unknown extent before plateauing. Regardless, it 525

appears that even in this worst-case-scenario at 27 ºC, Atlantic salmon are still not fully reliant 526

upon anaerobic metabolism to fuel function. This finding is supported by the overall lack of 527

lactate accumulation in any tissue, although without data on lactate flux (i.e. production and 528

utilization rates) we cannot conclusively rule out the activation of anaerobic metabolism. The 529

literature value for the upper incipient lethal temperature (at which fish can survive for at least 7 530

days) in young (0 - 1+) Atlantic salmon acclimated to 15 ºC is 27.6 ± 0.27 ºC (Elliot 1991). 531

Collectively, our data suggest that at 27 ºC, fish do retain aerobic metabolic capacity and no 532

mortality was observed under the conditions tested. That said, a preliminary experiment holding 533

fish overnight at 27 ºC resulted in mortality after only ~5 h. Juvenile fish are generally 534

considered to have higher thermal tolerance than adults (Fowler et al. 2009; Breau et al. 2007). It 535

is possible that the older and larger fish used here compared to Elliot (1991; 0 – 1+) may account 536

for differences in upper incipient lethal temperatures. 537

Enhanced oxygen demand due to exercise or elevated water temperatures necessitates 538

increased oxygen uptake at the gill. The mechanisms for increasing oxygen uptake (e.g. 539

increased ventilation, gill surface area, blood perfusion, etc.) are at odds with those required to 540

Page 24 of 48



Draft

25

minimize ion loss. This conflict has been termed the osmo-respiratory compromise (see Sardella 541

and Brauner 2007 for review). Thus, a teleost fish in freshwater exposed to high temperature 542

would be predicted to gain water and lose ions to their hypotonic surroundings. Anticipated 543

physiological ramifications may include decreased plasma osmolality and ions in addition to 544

lowered Hb and Hct due to haemodilution. Contrary to these predictions, we observed a 545

significant increase in plasma osmolality over time in all experimental treatment groups. This 546

increase is largely accounted for by increases in plasma glucose, lactate, and ions (chloride 547

measured here, and presumably its primary counter ion sodium). NKA activity was the same 548

across all experimental groups (p = 0.122; data not shown). Collectively, these data suggest that 549

an osmo-respiratory compromise is not occurring following exposure to 27 ºC in Atlantic 550

salmon. 551

Haematocrit and Hb decreased over time in all experimental groups, although only 552

significantly in the negative control and diel cycle A. We propose that these haematological 553

effects are due to an effect of fasting (Gillis and Ballantyne 1996; Djordjevic et al. 2012) rather 554

than a result of haemodilution; but they may reduce oxygen carrying capacity. Any fasting 555

responses here are an unavoidable by-product of fish not being fed during experimentation to 556

avoid post prandial increases in ṀO2. 557

Heat shock proteins have been implicated in the acquisition and decay of thermotolerance 558

in fish since the in vitro work of Mosser and Bols (1988) in rainbow trout fibroblasts. The 559

induction of HSPs is a plastic response, and one that acclimation temperature can modulate. For 560

example: increasing the acclimation temperature of the eurythermal goby (Gillichthys mirabilis) 561

by 20 degrees (from 10 to 30 ºC) increased the induction temperature for both HSP70 and 562

HSP90 significantly (Dietz 1994; see also Oda et al. 1991; Dietz and Somero 1992; Nakano 563

Page 25 of 48



Draft

26

2000; Nakano and Iwama 2002 for other examples). We show that both the HSP70 and ubiquitin 564

responses in Atlantic salmon are tissue specific and dependent on the thermal profile experienced 565

by the fish. The undetectable levels of HSP70 in any tissue at 16 ºC reinforce that this was an 566

appropriate holding temperature for Atlantic salmon and that confinement in the swim tunnel 567

was not stressful. Not surprisingly, in all tissues studied we saw marked differences in HSP 568

expression between the two diel thermal cycling groups and the positive control group, where the 569

temperature was maintained at a constant 16 ºC. However, in no tissue was there a significant 570

difference in HSP70 between diel thermal cycles A and B. Perhaps the accumulated thermal 571

loads experienced by fish in these two groups (349 and 415 ºC·h-1 respectively) were not 572

sufficiently different to differentially influence the HSP70 response. 573

In addition to different thermal loads, each experimental treatment is distinct in the rate 574

and magnitude of temperature change, and the presence or absence of diel cycling, all of which 575

have the potential to influence the cellular stress response. In RBC, gill, and white muscle, 576

increases in HSP70 appeared to mirror increases in thermal load. HSP70 induction was highest 577

in the diel thermal cycling groups (349 - 415 ºC·h) in these tissues. This pattern of HSP70 578

induction was not consistent in the heart and liver, where HSP70 was ~2.5 times higher in the 579

positive control than in either thermal cycling group. Of the tissues examined here, heart and 580

liver are arguably the most likely to experience oxygen limitation due to the single series 581

circulatory anatomy of fish (Satchell 1991). If the unique HSP70 response in heart and liver is 582

due to their susceptibility and/or decreased tolerance to hypoxia, then the greater rate of 583

temperature change in the positive control compared to the thermal cycling groups, coupled with 584

the lack of prior diel cycling, may explain this HSP70 induction pattern. 585

Page 26 of 48



Draft

27

Dot blotting was successfully used by Hofmann and Somero (1995) to correlate ubiquitin 586

conjugated proteins with HSP70 levels in the intertidal mussel (Mytilus trossulus) and similarly 587

by MacLellan et al. (2015) in the gill of an elasmobranch fish (Squalus acanthias). However, 588

increases in protein degradation can also result from overall increases in temperature, 589

particularly when coupled with starvation (Houlihan 1991; McCarthy and Houlihan 1997). In 590

general, the level of ubiquitination corresponded to the amount of HSP70 in a given tissue (e.g. 591

ubiquitin levels were lowest in RBCs and gill as were HSP70 levels). One notable exception 592

occurred in white muscle, where ubiquitin levels were high while HSP70 was relatively low. We 593

speculate that the high ubiquitin levels in this tissue were due to increases in protein turnover. 594

White muscle constitutes ~70% of salmonid body mass (Eddy and Handy 2012) and thus offers 595

the largest, most accessible source of protein for degradation to fuel energy metabolism. 596

Although this study was designed around the Miramichi River system, temperature 597

fluctuations are a broadly occurring natural phenomenon in the aquatic environment, and thus 598

conclusions drawn here may also be extended to other ecosystems and species. We made several 599

predictions at the outset of this study: 1) that fish exposed to diel thermal cycling would respond 600

differently to a 27 ºC exposure than fish maintained at 16 ºC, 2) that the two different diel cycles 601

tested would result in unique biological effects, and 3) that diel cycling would result in either 602

beneficial heat hardening, and/or detrimental cellular stress and damage. We did indeed observe 603

differential responses between thermal cycled and non-cycled fish, as well as subtle differences 604

between the two different thermal cycling treatments. We did not see clear evidence of a heat 605

hardening response, nor did exposure to 27 ºC appear to induce a severe physiological stress 606

response under the conditions tested. Resting metabolic rate adhered to predicted increases based 607

on Q10 effects, indicating that salmon have little capacity to regulate RMR in response to 608

Page 27 of 48



Draft

28

fluctuations in temperature; metabolic shifts at the cellular and signalling levels remain possible 609

but were not examined here. Collectively, our data suggest that three days of environmentally 610

relevant diel temperature fluctuation does not substantially alter the response of Atlantic salmon 611

to temperatures approaching their Tcrit. It is possible that diel thermal cycling impacts 612

cardiorespiratory function in Atlantic salmon and improves function in a way not assessed in the 613

current study. Atlantic salmon do exhibit cardiorespiratory plasticity which can contribute to 614

increased thermal tolerance and survival at elevated temperatures (Anttila et al. 2014). 615

A direct application of this study is refinement of the regulatory policies in Atlantic 616

salmon river systems. At present the Miramichi River is closed to angling activities when the 617

minimum water temperature for two consecutive days is ≥20 ºC (DFO 2012). Neither diel 618

thermal cycling group tested here meets these closure criteria, yet fish exhibit evidence of 619

cellular, haematological, and osmotic changes that would presumably also occur in the wild. 620

Bringing together our reported physiological changes with the knowledge that post release 621

mortality following angling in warm water (20 ºC) is high in Atlantic salmon (~40%; Wilkie et 622

al., 1996), suggests that re-evaluation of river closure criteria is warranted. Furthermore, the 623

differences shown here between constant and cycling thermal histories suggest that conclusions 624

drawn from acute laboratory studies conducted at stable acclimation temperatures may not be 625

reflective of responses in the wild. Moving forward, researchers should make every effort to 626

incorporate ecological relevance into their experimental designs. 627

628

629

ACKNOWLEDGEMENTS 630

631

Page 28 of 48



Draft

29

The authors would like to acknowledge the Miramichi River Environmental Assessment 632

Committee for access to their monitoring data. Additional and sincere appreciation are extended 633

to Neal Callaghan for respirometry data analysis, Dr. Andrea Morash for aerobic scope 634

determination, Emily Corey and colleagues for fish collection, Courtney Aubé for laboratory 635

assistance, and Wayne Anderson for animal husbandry. The authors have no ethical or financial 636

conflicts of interest to declare. Financial support was provided by Natural Sciences and 637

Engineering Research Council Discovery Grants (SC, TJM), a Canadian Foundation for 638

Innovation John R. Evans Leadership Fund award (TJM), New Brunswick Innovation 639

Foundation (NBIF) STEM Graduate scholarship (LT), a New Brunswick Environmental Trust 640

Fund, and NBIF Emerging Projects awards (SC, TJM). 641

642

Page 29 of 48



Draft

30

REFERENCES 643

644

Anttila, K., Couturier, C.S., Øverli, Ø., Johnsen, A., Marthinsen, G., Nilsson, G.E., and Farrell, 645

A.P. 2014. Atlantic salmon show capability for cardiac acclimation to warm temperatures 646

[online]. Nature Comm. 5. doi:10.1038/ncomms5252. 647

Beauregard, D., Enders, E., Boisclair, D., and Kidd, K. 2013. Consequences of circadian 648

fluctuations in water temperature on the standard metabolic rate of Atlantic salmon parr (Salmo 649

salar). Can. J. Fish. Aquat. Sci. 70(7): 1072–1081. doi:10.1139/cjfas-2012-0342. 650

Breau, C., Cunjak, R.A. and Bremset, G. 2007. Age-specific aggregation of wild juvenile 651

Atlantic salmon Salmo salar at cool water sources during high temperature events. J. Fish Biol. 652

71: 1179–1191. 653

Breau, C., Cunjak, R.A. and Peake, S.J. 2011. Behaviour during elevated water temperatures: 654

can physiology explain movement of juvenile Atlantic salmon to cool water? J. Anim. Ecol: 80: 655

844–853. 656

Clark, T.D., Eliason, E.J., Sandblom, E., Hinch, S.G. and Farrell, A.P. 2008. Calibration of a 657

hand-held haemoglobin analyser for use on fish blood. J. Fish Biol. 73: 2587–2595. 658

Clark, T.D., Sandblom, E. and Jutfelt, F. 2013. Aerobic scope measurements of fishes in an era 659

of climate change: respirometry, relevance and recommendations. J. Exp. Biol. 216: 2771–2782. 660

Currie, S. and Schulte, P.M. 2014. Thermal Stress. In The Physiology of Fishes fourth edition. 661

Edited by D.H. Evans, J.B. Claiborne and S. Currie. CRC Press, Boca Raton. pp. 257–288. 662

Page 30 of 48



Draft

31

Dietz, T.J. and Somero, G.N. 1992. The threshold induction temperature of the 90-kDa heat 663

shock protein is subject to acclimatization in eurythermal goby fishes (genus Gillichthys). Proc 664

Nat Acad Sci. USA 89(8): 3389-3393. 665

Dietz, T.J. 1994. Acclimation of the threshold induction temperatures for 70-kDa and 90-kDa 666

heat shock proteins in the fish Gillichthys mirabilis. J. Exp. Biol. 188: 333–338. 667

Deitch, E.J., Fletcher, G.L., Petersen, L.H., Costa, I.A.S.F., Shears, M.A., Driedzic, W.R. and 668

Gamperl, A.K. 2006. Cardiorespiratory modifications, and limitations, in post-smolt growth 669

hormone transgenic Atlantic salmon Salmo salar. J. Exp. Biol. 209: 1310–1325. 670

DFO. 2012. Temperature threshold to define management strategies for Atlantic salmon (Salmo 671

salar) fisheries under environmentally stressful conditions. DFO Canadian Science Advisory 672

Secretariat Science Advisory Report. 673

DFO. 2013. Atlantic Salmon (Salmo salar) returns to the Miramichi River (NB) for 2012. DFO 674

Canadian Science Advisory Secretariat Science Advisory Report, 2013/009. 675

Djordjevic, B., Kristensen, T., Øverli, Ø., Rosseland, B.O. and Kiessling, A. 2012. Effect of 676

nutritional status and sampling intensity on recovery after dorsal aorta cannulation in free-677

swimming Atlantic salmon (Salmo salar L.). Fish Physiol. Biochem. 38: 259–272. 678

Eddy, B.F. and Handy, R.D. 2012. General principles of fish physiology: living in water. In 679

Ecological and Environmental Physiology of Fishes. Oxford University Press, Oxford. pp. 43-680

101. 681

Page 31 of 48



Draft

32

Eldridge, W.H., Sweeney, B.W. and Law, J.M. 2015. Fish growth, physiological stress, and 682

tissue condition in response to rate of temperature change during cool or warm diel thermal 683

cycles. Can. J. Fish. Aquat. Sci. 72: 1527–1537. 684

Eliason, E.J. and Farrell, A.P. 2016. Oxygen uptake in Pacific salmon Oncorhynchus spp.: when 685

ecology and physiology meet. J. Fish Biol. 88: 359–388. 686

Elliott, J.M. 1991. Tolerance and resistance to thermal stress in juvenile Atlantic salmon, Salmo 687

salar. Freshwater Biol. 25: 61–70. 688

Evans, D.O. 1984. Temperature independence of the annual cycle of standard metabolism in the 689

pumpkinseed. Trans. Am. Fish. Soc. 113: 494–512. 690

Fowler, S.L., Hamilton, D. and Currie, S. 2009. A comparison of the heat shock response in 691

juvenile and adult rainbow trout (Oncorhynchus mykiss) — implications for increased thermal 692

sensitivity with age. Can. J. Fish. Aquat. Sci. 66: 91–100. 693

Gamperl, A.K., Todgham, A.E., Parkhouse, W.S., Dill, R. and Farrell, A.P. 2001. Recovery of 694

trout myocardial function following anoxia: preconditioning in a non-mammalian model. Amer J 695

Physiol: Regul, Integr Comp Physiol. 281: R1755–R1763. 696

Gillis, T.E. and Ballantyne, J.S. 1996. The effects of starvation on plasma free amino acid and 697

glucose concentrations in lake sturgeon. J. Fish Biol. 49: 1306–1316. 698

Hofmann, G. and Somero, G. 1995. Evidence for protein damage at environmental temperatures: 699

seasonal changes in levels of ubiquitin conjugates and hsp70 in the intertidal mussel Mytilus 700

trossulus. J. Exp. Biol. 198: 1509–1518. 701

Page 32 of 48



Draft

33

Houlihan, D.F., Livingstone, D.R. and Lee, R.F. 1991. Protein turnover in ectotherms and its 702

relationships to energetics. In Advances in Comparative and Environmental Physiology. Edited 703

by R. Gilles. Springer-Verlag, Berlin. pp. 1-45. 704

ICES. 2015. Report of the working group on North Atlantic salmon (WGNAS), 17-26 March, 705

Moncton, Canada. ICES CM 2015/ACOM:09. 332 pp. 706

Jensen, A.J., Johnsen, B.O. and Saksgard, L. 1989. Temperature requirements in Atlantic salmon 707

(Salmo salar), Brown trout (Salmo trutta), and Arctic char (Salvelinus alpinus) from hatching to 708

initial feeding compared with geographic distribution. Can. J. Fish. Aquat. Sci. 46: 786–789. 709

LeBlanc, S., Middleton, S., Gilmour, K.M. and Currie, S. 2011. Chronic social stress impairs 710

thermal tolerance in the rainbow trout (Oncorhynchus mykiss). J. Exp. Biol. 214: 1721–1731. 711

LeBlanc, S., Hoglund, E., Gilmour, K.M. and Currie, S. 2012. Hormonal modulation of the heat 712

shock response: insights from fish with divergent cortisol stress responses. Amer J Physiol: 713

Regul, Integr Comp Physiol. 302: R184–R192. 714

Loeschcke, V. and Hoffmann, A.A. 2007. Consequences of heat hardening on a field fitness 715

component in Drosophila depend on environmental temperature. Am Nat. 169: 175–183. 716

Lund, S.G., Caissie, D., Cunjak, R.A., Vijayan, M.M. and Tufts, B.L. 2002. The effects of 717

environmental heat stress on heat-shock mRNA and protein expression in Miramichi Atlantic 718

salmon (Salmo salar) parr. Can. J. Fish. Aquat. Sci. 59: 1553–1562. 719

Lundberg, J.G., Kottelat, M., Smith, G.R., Stiassny, M.L.J. and Gill, A.C. 2000. So many fishes, 720

so little time: an overview of recent ichthyological discovery in continental waters. Annals of the 721

Missouri Botanical Garden. 87: 26. 722

Page 33 of 48



Draft

34

Lyytikäinen, T. and Jobling, M. 1998. The effect of temperature fluctuations on oxygen 723

consumption and ammonia excretion of underyearling Lake Inari Arctic charr. J. Fish Biol. 52: 724

1186–1198. 725

MacLellan, R.J., Tunnah, L., Barnett, D., Wright, P.A., MacCormack, T. and Currie, S. 2015. 726

Chaperone roles for TMAO and HSP70 during hyposmotic stress in the spiny dogfish shark 727

(Squalus acanthias). J. Comp. Physiol., B. 185: 729–740. 728

Ma, G., Hoffmann, A.A. and Ma, C.-S. 2015. Daily temperature extremes play an important role 729

in predicting thermal effects. J. Exp. Biol. 218: 2289–2296. 730

McCarthy, I.D. and Houlihan, D.F. 1997. The effect of temperature on protein metabolism in 731

fish: the possible consequences for wild Atlantic salmon (Salmo salar L.). In Global Warming: 732

Implications for Freshwater and Marine Fish. Edited by C.M. Wood & D.G. McDonald. 733

Cambridge University Press, Cambridge. pp. 51-78. 734

Mesa, M., Weiland, L. and Wagner, P. 2002. Effects of acute thermal stress on the survival, 735

predator avoidance, and physiology of juvenile fall chinook salmon. Northwest Science. 76: 736

118–128. 737

Mosser, D.D. and Bols, N.C. 1988. Relationship between heat-shock protein synthesis and 738

thermotolerance in rainbow trout fibroblasts. J. Comp. Physiol., B. 158: 457–467. 739

Murry, C.E., Jennings, R.B. and Reimer, K.A. 1986. Preconditioning with ischemia: a delay of 740

lethal cell injury in ischemic myocardium. Circulation. 74: 1124–1136. 741

Nakano, K. 2000. The cellular heat shock response and thermotolerance of fishes. PhD thesis, 742

University of British Columbia, Vancouver, BC. 743

Page 34 of 48



Draft

35

Nakano, K. and Iwama, G.K. 2002. The 70-kDa heat shock protein response in two intertidal 744

sculpins, Oligocottus maculosus and O. snyderi: relationship of hsp70 and thermal tolerance. 745

Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 133: 79–94. 746

Narum, S.R., Campbell, N.R., Meyer, K.A., Miller, M.R., and Hardy, R.W. 2013. Thermal 747

adaptation and acclimation of ectotherms from differing aquatic climates. Mol Ecol 22(11): 748

3090–3097. 749

Oda, S., Mitani, H., Naruse, K. and Shima, A. 1991. Synthesis of heat shock proteins in the 750

isolated fin of the Medaka, Oryzias latipes, acclimatized to various temperatures. Comparative 751

Biochemistry and Physiology Part B: Comparative Biochemistry. 98: 587–591. 752

Oligny-Hébert, H., Senay, C., Enders, E.C. and Boisclair, D. 2015. Effects of diel temperature 753

fluctuation on the standard metabolic rate of juvenile Atlantic salmon (Salmo salar): influence of 754

acclimation temperature and provenience. Can. J. Fish. Aquat. Sci. 72: 1306–1315. 755

Portner, H.O. and Knust, R. 2007. Climate change affects marine fishes through the oxygen 756

limitation of thermal tolerance. Science. 315: 95–97. 757

Reid, S.D., Wood, C.M. and McDonald, D.G. 1997. Interactive effects of temperature and 758

pollutant stress. In Global Warming: Implications for Freshwater and Marine Fish. Cambridge 759

University Press, Cambridge. pp. 325-350. 760

Sardella, B.A. and Brauner, C.J. 2007. The osmo-respiratory compromise in fish. In Fish 761

respiration and environment. Edited by M.N. Fernandes. Science Publishers, Enfield. pp. 147–762

165. 763

Page 35 of 48



Draft

36

Satchell, G.H. 1991. The peripheral circulation. In Physiology and Form of Fish Circulation. 764

Cambridge University Press, Cambridge. pp. 41-55. 765

Steinhausen, M.F., Sandblom, E., Eliason, E.J., Verhille, C. and Farrell, A.P. 2008. The effect of 766

acute temperature increases on the cardiorespiratory performance of resting and swimming 767

sockeye salmon (Oncorhynchus nerka). J. Exp. Biol. 211: 3915–3926. 768

Templeman, N.M., LeBlanc, S., Perry, S.F. and Currie, S. 2014. Linking physiological and 769

cellular responses to thermal stress: β-adrenergic blockade reduces the heat shock response in 770

fish. J. Comp. Physiol., B. 184: 719–728. 771

Welinder, C. and Ekblad, L. 2011. Coomassie staining as loading control in western blot 772

analysis. Journal of Proteome Research. 10: 1416–1419. 773

Wilkie, M.P., Davidson, K., Brobbel, M.A., Kieffer, J.D., Booth, R.K., Bielak, A.T., and Tufts, 774

B.L. 1996. Physiology and survival of wild Atlantic salmon following angling in warm summer 775

waters. Trans. Am. Fish. Soc. 125(4): 572–580. 776

Woodward, G., Perkins, D.M. and Brown, L.E. 2010. Climate change and freshwater 777

ecosystems: impacts across multiple levels of organization. Philos. Trans. R. Soc., B, 365: 2093–778

2106 779

780

Page 36 of 48



Draft

37

TABLES 781

Table 1: Detailed time and temperature parameters for each experimental treatment group 782

Thermal Profile Time (h) Range

(°C) ∆ °C

Ramp

Rate

(°C·h-

1)

Overall

Mean

(°C)

Thermal

Load

(°C·h)*

NEGATIVE

CONTROL

0 - 84 16 0 --

15.7 0

POSITIVE

CONTROL

0 – 72 16 0 --

16.7 66

72 - 84 16 - 27 11 + 0.92

DIEL CYCLE A 0 – 11 16 - 24 8 + 0.73

20.2 349

12 – 23 24 - 16 8 - 0.73 24 – 35 16 - 24 8 + 0.73 36 – 47 24 - 16 8 - 0.73 48 – 59 16 - 24 8 + 0.73 60 – 71 24 - 16 8 - 0.73 72 - 84 16 – 27 11 + 0.92

DIEL CYCLE B

0 - 10 16 - 21 5 + 0.50

21.0 415

13.5 – 20.5 21 – 17.5 3.5 - 0.50 24 - 34 17.5 – 22.5 5 + 0.50

37.5 – 44.5 22.5 - 19 3.5 - 0.50 48 – 58 19 - 24 5 + 0.50

61.5 – 68.5 24 - 21 4 - 0.57 72 - 84 21 - 27 6 + 0.50

* Thermal load is calculated as the cumulative area under the curve of each treatment’s thermal 783 profile where the temperature is above 16°C. 784

Page 37 of 48



Draft

38

785

Table 2: Plasma (osmolality, chloride, glucose, and lactate) and RBC (Hb, Hct, and MCHC) 786

parameters in Atlantic salmon among four experimental temperature groups at two sampling 787

time points (t = 0 h and t = 84 h), and the calculated difference (∆) between sampling times. 788

789

Notes: 790 All data presented as mean ± S.E.M. n = 6 in each treatment group. Except t = 0 h lactate (n = 5). 791 Different letters show significant difference between treatment groups at that sampling time. 792 Asterisk (*) indicates significant difference between t = 0 h and t = 84 h within treatment group 793

TIME NEGATIVE

CONTROL

POSITIVE

CONTROL

DIEL

CYCLE A

DIEL

CYCLE B

PLASMA Osmolality (mmol·kg-1)

t = 0 h 298.5 ± 5.2 302.5 ± 2.7 302.6 ± 1.2 291.3 ± 2.4

t = 84 h 295.1 ± 3.0 a 320.7 ± 3.4

b* 318.3 ± 2.1

b* 308.8 ± 4.8

ab*

∆ - 3.4 ± 3.4 + 18.2 ± 3.7 + 15.6 ± 1.8 + 17.5 ± 3.3

Chloride

(mmol·L-1)

t = 0 h 133.0 ± 1.6 136.3 ± 2.5 135.3 ± 3.1 135.5 ± 1.3

t = 84 h 138.3 ± 2.0 140.0 ± 2.0 143.3 ± 2.9 149.0 ± 3.0 *

∆ + 5.3 ± 2.7 + 3.8 ± 3.7 + 8.2 ± 5.7 + 13.5 ± 2.6

Glucose

(mmol·L-1)

t = 0 h 5.3 ± 0.5 4.8 ± 0.1 4.8 ± 0.2 5.3 ± 0.2

t = 84 h 4.0 ± 0.1 a* 5.6 ± 0.5

b 5.6 ± 0.3 b* 5.2 ± 0.2

b

∆ - 1.4 ± 0.4 + 0.9 ± 0.5 + 0.8 ± 0.2 - 0.06 ± 0.3

Lactate

(mmol·L-1)

t = 0 h 5.8 ± 0.5 5.1 ± 0.4 3.8 ± 0.7 4.7 ± 0.2

t = 84 h 3.1 ± 0.6 a 9.6 ± 1.8

b 6.4 ± 0.8

a 7.0 ± 0.7

b*

∆ - 2.8 ± 0.7 + 4.5 ± 1.0 + 2.6 ± 0.9 + 2.2 ± 0.6

RBC Haemoglobin (g·L-1)

t = 0 h 86.6 ± 4.7 94.1 ± 2.9 99.1 ± 2.6 87.6 ± 9.2

t = 84 h 62.6 ± 7.3 * 83.8 ± 2.1 79.3 ± 4.8 * 75.0 ± 5.7

∆ - 24.0 ± 5.3 - 10.3 ± 4.4 - 19.8 ± 6.1 - 12.7 ± 12.0

Haematocrit (%)

t = 0 h 44.8 ± 3.2 45.8 ± 1.5 49.4 ± 2.0 43.0 ± 3.3

t = 84 h 31.3 ± 3.7 * 40.7 ± 1.8 39.7 ± 2.6 * 38.9 ± 3.8

∆ - 13.5 ± 3.5 - 5.1 ± 2.1 - 9.7 ± 3.2 - 4.1 ± 6.4

MCHC (g·L-1)

t = 0 h 197.1 ± 13.3 205.8 ± 4.9 201.9 ± 7.1 201.5 ± 8.7

t = 84 h 202.8 ± 15.7 207.0 ± 4.7 201.1 ± 7.1 197.2 ± 11.8

∆ + 5.7 ± 7.6 + 1.1 ± 3.8 - 0.8 ± 10.2 - 4.2 ± 10.4

Page 38 of 48



Draft

39

For all plasma parameters α-critical = 0.025. For all RBC parameters α-critical = 0.05. 794

FIGURE LEGENDS 795

796

Figure 1: Hourly river water temperatures in the summer of 2015, from Doaktown monitoring 797

station on the Miramichi River, NB. Data generously provided by the Miramichi River 798

Environmental Assessment Committee. Panel A is representative of a consistent daily oscillation 799

in temperature (~6.5 ºC). Panel B is representative of a natural increase in temperature over 800

several days creating a ramp. 801

802

Figure 2: Atlantic salmon mean M� O2 (mgO2·kg-1·h-1 ± SEM) over time (black) in response to 803

temperature fluctuation (grey) in four treatment groups: negative control (A), positive control 804

(B), diel cycle A (C), and diel cycle B(D). n = 6 fish within each treatment group. 805

806

Figure 3: ṀO2 at 27 ± 0.5 ºC in positive control and diel cycles A and B at the end of the 807

experimental time course (A) and at 21 ± 0.5 ºC on days 1 – 4 in both diel cycling groups (B). A. 808

n = 6 in each treatment group. No significant difference between treatment groups (p = 0.83; 1W 809

ANOVA). B. For each time within each treatment group n = 6, with the exception of diel cycle B 810

on day 4 where n = 5. Asterisk indicates that treatment groups are significantly different from 811

one another (p = 0.019; split-plot ANOVA). 812

813

Figure 4: Heart (A), and white muscle (B) lactate concentrations (mmol • L-1; mean ± S.E.M) in 814

Atlantic salmon. n = 6 in each treatment group. A. Heart lactate concentrations do not differ 815

significantly between treatment groups (p = 0.26; 1W ANOVA), B. White muscle lactate 816

concentrations do not differ significantly between treatment groups (p = 0.36; 1W ANOVA) 817

Page 39 of 48



Draft

40

818

Figure 5: HSP70 and ubiquitin in gill (A, B), white muscle (C, D), RBCs (E, F), heart (G, H), 819

and liver (I, J) in Atlantic salmon (Salmo salar) exposed to four temperature treatments: negative 820

control, positive control, and diel cycles A and B. All graphs show relative band/dot density 821

(mean ± SEM). n = 6 fish within each treatment group for each tissue type. RBC data shown in 822

panels E + F were sampled at 84 h prior to terminal sampling. Due to a significant treatment x 823

tissue interaction (p < 0.001; 2W ANOVA) data were split to analyse change between treatment 824

groups in each tissue separately using 1W ANOVAs. Different letters indicate statistically 825

significant (p < 0.05) differences between treatment groups. A. Significant treatment group effect 826

(p = 0.023). B. Significant treatment group effect (p = 0.001). C. Significant treatment group 827

effect (p < 0.001). D. Significant treatment group effect (p < 0.001). E. Significant treatment 828

group effect (p < 0.001). F. Significant treatment group effect (p = 0.003) G. Significant 829

treatment group effect (p < 0.001). H. Significant treatment group effect (p = 0.003). I. 830

Significant treatment group effect (p < 0.001). J. No significance treatment group effect (p < 831

0.35). 832

Page 40 of 48



Draft

Figure 1

248x350mm (300 x 300 DPI)

Page 41 of 48



Draft

Figure 2

190x121mm (300 x 300 DPI)

Page 42 of 48



Draft

Figure 3

164x231mm (300 x 300 DPI)

Page 43 of 48



Draft

Figure 4

151x254mm (300 x 300 DPI)

Page 44 of 48



Draft

Figure 5

158x268mm (300 x 300 DPI)

Page 45 of 48



Draft

Figure S1: Mean MO2 (mgO2 • kg-1

• h-1

± SEM) during overnight acclimation to swim tunnel and

recovery from initial blood sample in five Atlantic salmon. One fish from the negative control and two

diel cycling groups, and two fish from the positive control. For comparison, the shaded region shows the overall

mean (solid line) plus/minus 1 std. deviation (dashed lines) of the negative control group across the 84 h

experimental time course.

Page 46 of 48



Draft

Supplemental Table 1: Sample size (n), experiment start date, length (cm), body mass (g), percent decrease in

body mass at experiment cessation (%), and treatment group (mean ± S.E.M) length, mass, and percent mass

change in all Atlantic salmon used in this study.

Treatment n Start Length Starting ∆ Mass Mean Mean Mass Mean

Group Date (cm) Mass (g) (%) Length (cm) Mass ∆

(cm) (%)

NEGATIVE Nov 2014 26.5 166 -

CONTROL Nov 2014 28 202.5 -

6 Nov 2014 26.5 165.5 - 29.75 ± 2.0 276.2 ± 74.4 a -

Nov 2014 28 214.7 -

Apr 2015 30 268.3 -

Sept 2015 39.5 640.3 -1.3

POSTIVE Jan 2015 29.0 226.1 -

CONTROL Feb 2015 30.0 268.9 -

6 Feb 2015 30.0 271.9 - 33.25 ± 2.1 407.1 ± 94.1 a -2.4 ± 0.7

June 2015 31.5 299.6 -3.8

Sept 2015 37.5 573.7 -1.5

Oct 2015 41.5 802.1 -2.0

DIEL CYCLE A July 2015 32.0 316.2 -1.3

July 2015 34.0 415.7 -6.5

6 Aug 2015 37.5 520.9 -5.8 36.4 ± 1.4 489.2 ± 56.0 b -3.9 ± 1.0

Aug 2015 34.5 412 -5.3

Aug 2015 40.5 567.7 -1.2

Sept 2015 40.0 702.7 -3.0

DIEL CYCLE B Apr 2015 28.0 279.9 -

Apr 2015 28.5 257.0 -

6 Apr 2015 31.0 271.0 -7.3 32.0 ± 1.3 370.8 ± 53.7 a -5.4 ± 1.0

May 2015 35.0 384.3 -

June 2015 34.0 436.0 -5.2

Oct 2015 35.5 596.4 -3.8

AEROBIC Nov 2015 - 578.0 -

SCOPE 4 Nov 2015 - 745.0 - - 679.9 ± 52.1 b -

Nov 2015 - 605.0 -

Nov 2015 - 791.0 -

- No available data indicated by (-) - Different lowercase letters indicate statistically significant (α-critical = 0.05) differences in fish masses between treatment groups.

Page 47 of 48



Draft

A. Gill and Heart HSP70

B. Liver and Muscle HSP70

Figure S2: Representative western blot images for the HSP70 salmonid specific antibody

(polyclonal rabbit affinity purified; AS05 061A, Agrisera) A. Lanes 1 – 4: supersample standard curve (2x

– 0.25x), lane 5: negative control gill, lane 6: positive control gill, lane 7: diel cycle A gill, lane 8: diel cycle B gill,

lane 9: negative control heart, lane 10: positive control heart, lane 11: diel cycle A heart, lane 12: diel cycle B

heart. B. Lanes 1 – 4: supersample standard curve (2x – 0.25x), lane 5: negative control liver, lane 6: positive

control liver, lane 7: diel cycle A liver, lane 8: diel cycle B liver, lane 9: negative control muscle, lane 10: positive

control muscle, lane 11: diel cycle A muscle, lane 12: diel cycle B muscle.

Page 48 of 48



draft...draft 5 84 for this study. we used atlantic salmon (salmo salar linnaeus), a primarily...

Documents