Using Sulfur Stable Isotopes to Understand Feeding Behavior and

May 28, 2015 - *Phone: (418) 654-2640; fax: (418) 654-2600; e-mail: [email protected]. ... selenium (Se) concentrations in yellow perch (Perca flaves...
0 downloads 5 Views 575KB Size
Subscriber access provided by NEW YORK UNIV

Article

Using Sulfur Stable Isotopes to Understand Feeding Behavior and Selenium Concentrations in Yellow Perch (Perca flavescens) Dominic Ponton, and Landis Hare Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b00718 • Publication Date (Web): 28 May 2015 Downloaded from http://pubs.acs.org on June 5, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

Environmental Science & Technology

Using Sulfur Stable Isotopes to Understand Feeding Behavior and Selenium Concentrations in Yellow Perch (Perca flavescens)

DOMINIC E. PONTON AND LANDIS HARE*

Institut national de la recherche scientifique, Centre – Eau Terre Environnement (INRS-ETE), Université du Québec, 490 rue de la Couronne, Quebec City, QC, Canada, G1K 9A9

Word count: 4,935 words + 2,100 words (3 figures at 600 + 1 figure at 300) = 7,035 words.

Keywords: Selenium, Prey, Sulfur isotopes, Bioaccumulation, Biomagnification, Food web, Fish, Yellow perch, Perca flavescens

* Corresponding author phone: (418) 654-2640; fax: (418) 654-2600; e-mail: [email protected]

1 ACS Paragon Plus Environment

Environmental Science & Technology

1

Abstract

2

We measured selenium (Se) concentrations in yellow perch (Perca flavescens) muscle

3

and their prey collected from four Se-contaminated lakes located near metal smelters in

4

the eastern Canadian cities of Sudbury and Rouyn-Noranda. Yellow perch Se

5

concentrations were related to their weight in two of the four lakes. Measurements of

6

sulfur stable isotopes (δ34S) in yellow perch muscle and stomach contents showed that

7

larger fish tended to feed less on zooplankton and more on benthic invertebrates than

8

did smaller fish. Because Se concentrations are lower and δ34S signatures are higher in

9

zooplankton than in sediment-feeding invertebrates, there was an inverse relationship

10

between animal Se concentrations and δ34S signatures in all of our study lakes. δ34S

11

signatures were highly effective in characterizing these food web relationships.

12

Selenium concentrations in yellow perch were 1.6 times those of its prey, which

13

indicates that Se is biomagnified by this fish in our study lakes. Estimated Se

14

concentrations in yellow perch gonads suggest that in two of our study lakes a third of

15

fish are at risk of reproductive toxicity.

16

Introduction

17

The minimum Se concentration required to satisfy physiological needs is only a factor

18

of four below the concentration that can be toxic to fish.1 Thus high concentrations of

19

Se in runoff from dryland irrigation or ash from coal-fired power plants have caused

20

deformities in the embryos of fish and aquatic birds.2,3 Metal mines and smelters can

21

also be important sources of Se contamination to aquatic systems.4 Yet in most such

22

systems, environmental studies have focused on metals such as cadmium (Cd), copper

23

(Cu), lead (Pb), nickel (Ni) and zinc (Zn) rather than on Se.5,6,7 For example, in eastern

24

Canada, toxic effects on aquatic animals from major smelters in Rouyn-Noranda

25

(Quebec) and Sudbury (Ontario) have been attributed to Cd8,9 and Ni,10 respectively.

26

Selenium was not considered in these studies, in spite of the fact that high Se

27

concentrations are found in the biota of lakes downwind from these point sources.11,12,13

28

Rather, Se was studied more as a mercury (Hg) antagonist11,12 than as a potentially toxic

29

element in its own right. Although emissions reductions from these smelters have led to 2 ACS Paragon Plus Environment

Page 2 of 25

Page 3 of 25

Environmental Science & Technology

30

lower trace metal concentrations in the water, surface sediments and animals of some

31

nearby lakes,14-18 trace metal concentrations in surface sediments have not returned to

32

background levels and they remain high at depth, which suggests that they still pose a

33

risk to benthic invertebrates18,19,20 and to the fish that feed on them.21,22

34

We set out to study the influence of fish feeding-behavior on their Se

35

concentrations in four lakes that have been contaminated by the Rouyn-Noranda and

36

Sudbury smelters. As a model organism, we chose the yellow perch, Perca flavescens

37

(Percidae), since it persists in many metal-contaminated boreal-forest lakes and thus

38

has been the subject of studies on the accumulation, trophic transfer and toxicity of

39

Cd, Cu, Ni and Zn.9,23 In contrast, the relationships between Se concentrations in

40

yellow perch, their diet and toxic effects have not been studied.

41

Selenium uptake by fish is likely to differ from that of metals such as Cd and

42

Ni because the latter are reported to be taken up mainly from water via the gills,24,25

43

whereas fish take up Se mainly from their food.5,26 Given this fact, the exposure of

44

yellow perch to Se is likely to vary according to their diet. Thus, early in their life,

45

when yellow perch feed mostly on zooplankton,27,28 they will be exposed to Se from

46

prey in the water-column compartment, whereas as they grow and switch to feeding

47

on benthos27,28 they will be increasingly exposed to Se taken up by invertebrates in

48

the sediment compartment. At even greater size, yellow perch can feed on small fish

49

that could themselves be either planktivorous or benthivorous.27,28 Since Se

50

concentrations are likely to differ among zooplankton, benthos and small fish, Se

51

exposure to yellow perch is likely to change as they grow.29

52

To determine the diet of yellow perch, we measured sulfur (S) stable isotopes

53

in this predator and in its prey. This approach is based on the fact that δ34S signatures

54

tend to be lower in sediment-feeding than in plankton-feeding invertebrates and these

55

differences determine the δ34S signatures of their predators.30 Mechanisms that could

56

explain this difference include fractionation during bacterial dissimilatory sulfate

57

reduction,31 sulfate diffusion across the sediment-water interface,32 S fractionation in

58

the drainage basins of these lakes,33 and historical changes in the isotopic

59

composition of sulfate due to variations in atmospheric SO2 sources.34 Measurements

60

of sulfur stable isotopes in fish provide a longer term assessment of diet than does the 3 ACS Paragon Plus Environment

Environmental Science & Technology

61

study of gut contents since the composition of gut contents varies from day to day

62

and is influenced by the resistance of various prey types to digestion.

63

To determine if Se is biomagnified, we compared Se concentrations in yellow

64

perch and their prey. Lastly, we evaluated whether the yellow perch populations that

65

we studied are likely to be at risk by comparing Se concentrations in their muscle

66

with those of toxicity guidelines, as well as by estimating Se concentrations in their

67

gonads and comparing these to published values for Se reproductive toxicity. Given

68

the paucity of studies on Se in freshwater food webs, especially in lakes,35 the results

69

of our study will be useful for understanding differences in Se concentrations among

70

prey types, how diet influences fish Se exposure, how Se exposure varies with fish

71

age and the possible consequences of such trends for Se risk assessments in lakes.

72

Methods

73

Study Sites and Animal Collection and Handling. We collected yellow perch and

74

their prey from mid-May to mid-June for three years (2010 to 2012) in four lakes

75

located on the Precambrian Shield that were downwind from metal smelters located

76

near Sudbury, Ontario, and Rouyn-Noranda, Quebec, Canada (Supporting

77

Information (SI), Table S1). We chose these lakes because of their high Se

78

concentrations in lakewater (SI, Table S1), sediments (SI, Table S1) and zooplankton

79

(> 10 µg g-1).13 At the time of sampling, the water columns at our sampling sites were

80

of uniform temperature and saturated in oxygen.

81

In each year, we collected 5 to 12 yellow perch in the littoral zone of each

82

lake using a 0.5-cm mesh-size seine net (75 m × 1 m), which gave a total of 12-30

83

fish from each lake. Live fish were held in a cooler with aerated lakewater until they

84

were measured, weighed and sacrificed on shore by decapitation and decerebration.

85

We removed the contents of each fish stomach and a piece of anterior dorsal muscle.

86

Since we used the other organs of these fish for biochemical measurements (not

87

reported), in 2009 we collected yellow perch from another Sudbury-area lake (Lohi

88

Lake; 46o23'N, 81o02'W) to compare Se concentrations in fish muscle with those in

89

gonads (SI, Figure S1) so that we could estimate the toxic potential of Se in the

90

reproductive organs of yellow perch from our study lakes. Fish parts and stomach 4 ACS Paragon Plus Environment

Page 4 of 25

Page 5 of 25

Environmental Science & Technology

91

contents were placed individually in acid-washed, 1.5-mL, polypropylene, centrifuge

92

tubes and stored at -20 oC. In the laboratory, fish stomach contents were thawed,

93

identified and counted in ultrapure water under a dissecting microscope then refrozen

94

at -20 oC for later analysis. Benthic prey were collected using an Ekman grab at the same site in each lake

95 96

at which we collected yellow perch. Mud samples were sieved in lakewater using a

97

0.5-mm mesh-size net and the invertebrates retained were held in a plastic bag in

98

lakewater until their return to the laboratory where they were sorted according to

99

lowest taxonomic level possible (SI, Table S2) and held at 4 oC for 3 days to depurate

100

their gut contents. We retained only prey types that we had collected in yellow perch

101

stomach contents for Se and S analyses. Three individuals of each taxon were placed

102

on a pre-weighed piece of acid-washed Teflon sheeting in an acid-washed micro-

103

centrifuge tube that was frozen at -20 °C until analysis. Pelagic prey were collected at night by hauling a 64 µm mesh-size plankton

104 105

net horizontally in the water-column. We sieved bulk zooplankton to separate larvae

106

of the phantom midge Chaoborus from micro-crustaceans (cladocerans and

107

copepods). We verified under a microscope that plankton fractions were composed of

108

at least 90% micro-crustaceans by volume. Samples for chemical analysis were

109

prepared by placing either 10-20 similar-sized fourth-instar Chaoborus larvae or ~10

110

mg wet weight of crustacean zooplankton on acid-washed, pre-weighed pieces of

111

Teflon sheeting in acid-washed micro-centrifuge tubes and frozen at -20 °C until

112

analysis.

113

Analyses. Frozen samples were freeze-dried (FTS Systems) and crushed into a

114

homogeneous powder in polypropylene centrifuge tubes using a plastic pestle and

115

subsamples were taken for the measurement of Se and stable isotopes of S. Subsamples (3-4 mg; Sartorius M2P PRO 11 balance) for S stable-isotope

116 117

analysis were put into tin capsules and weighed, along with ~8 mg of vanadium

118

pentoxide powder as a catalyst, and placed in a 96-well microplate. Blind duplicates

119

were made for quality assurance (reported as standard deviations in Figures 1, 3 and

120

Table S2). Sulfur isotopic signatures are reported as δ34S (‰) = [(34S/32S sample /

121

34

S/32S standard) – 1] × 103, where the S standard is from the Canyon Diablo Troilite 5 ACS Paragon Plus Environment

Environmental Science & Technology

122

(CDT). Sulfur stable-isotopes were measured by elemental analyzer – isotope ratio

123

mass spectrometry at Iso-Analytical Limited (Crewe, UK) using a Sercon elemental

124

analyzer and 20-20 mass spectrometer (Sercon Ltd, Crewe, UK). The certified

125

reference materials used during S isotopic-analyses were IAEA-SO-5 (barium sulfate,

126

δ34S = + 0.5‰; IAEA) and IA-R027 (whale baleen, δ34S = + 16.30‰; Iso-Analytical

127

working standard). Values for both were within the certified ranges (+0.3 ± 0.2 ‰

128

and +16.2 ± 0.3 ‰, respectively; ± standard deviation (SD)). Although we also

129

measured carbon and nitrogen stable-isotopes, these data are not shown because they

130

varied little among prey types and among yellow perch from a given lake.

131

Subsamples (1 to 15 mg dry weight) for Se analysis were weighed and placed

132

in acid-washed, 15 mL, high-density polyethylene bottles where they were digested at

133

room temperature for 2 days in concentrated nitric acid (Aristar®, ACS grade; 100 µL

134

per mg sample dry weight) followed by 1 day in concentrated hydrogen peroxide

135

(TraceSELECT®, Ultratrace grade; 40 µL per mg sample dry weight); digestate

136

volume was completed to 1 mL per mg sample dry weight using ultrapure water (18

137

MΩ•cm). Certified reference material (lobster hepatopancreas, TORT-2, National

138

Research Council of Canada) was submitted to the same digestion procedure.

139

Selenium of mass 82 g mol-1 was measured using Inductively Coupled Plasma - Mass

140

Spectrometry (ICP-MS; Thermo Elemental X Series) and interferences with bromine

141

(Br) and hydrogen (81Br + 1H) were corrected using a standard curve of several Br

142

concentrations. Selenium in TORT-2 (5.7 ± 0.1 (± SD) µg g-1; n = 4) was within the

143

certified range (5.6 ± 0.7 (95% confidence interval) µg g-1) and the detection limit of

144

the ICP-MS for Se was 0.2 µg L-1. Standard deviations reported for individual values

145

(Figures 1, 3 and S2) are for three analyses of the same sample.

146

Statistical Tests. Year to year differences in yellow perch muscle Se concentrations

147

were assessed using ANOVA followed by either a Student T test or its non-

148

parametric equivalent (Mann-Whitney U test).

149

We compared yellow perch whole weight to their muscle Se concentrations

150

and δ34S signatures. Where relationships were significant, we tested the adequacy of

151

fit of either a humped-shaped, log-normal, curve (Se concentrations in Lake Dufault

152

fish) or exponential curves (Se concentrations and δ34S signatures in Kelly Lake fish) 6 ACS Paragon Plus Environment

Page 6 of 25

Page 7 of 25

Environmental Science & Technology

153

when assumptions of normality (Shapiro-Wilk test) and equality of variances

154

(Levene’s test) were satisfied. We then determined yellow perch weight at maximum

155

Se concentrations using these bivariate scatterplots. In the case of Kelly Lake, we

156

determined fish weight (x) at maximum [Se] by iteration from x = 0 up to a y value

157

that was not significantly different from the plateau (± SD). This standard deviation is

158

the weight of the variance from our measured [Se] from five different fish from Kelly

159

Lake analyzed independently three times. Note that we did not consider the

160

significant linear relationship between yellow perch weight and δ34S signatures for

161

Lake Rouyn because it depended on a single large individual.

162

Slopes and y-intercepts of Se concentrations as a function of δ34S signatures

163

were compared independently for yellow perch muscle and invertebrates from each

164

of the four lakes using an analysis of covariance and a general linear model. If

165

regressions were significantly different, a Tukey multiple comparisons test was

166

conducted to determine which lakes differed from one another (see SI, Table S3, for

167

detailed results). The lack of significant differences between the slopes and y-

168

intercepts of prey data and fish for Lakes Dufault and Osisko, as well as for Lakes

169

Kelly and Rouyn (post hoc Tukey multiple comparisons; SI, Table 3), indicated that

170

prey data for these pairs of lakes could be combined. Combined prey Se

171

concentrations and δ34S signatures were compared using ANOVA followed by a non-

172

parametric Dunn’s test on ranks. These same lake pairings were used to evaluate the

173

risk posed by Se to yellow perch and the maximum consumption values of yellow

174

perch by humans.

175

Statistical analyses were carried out using Systat version 11, SigmaStat

176

version 3.5 and Jump 9.0 (SAS Institute Inc.). A P value of 0.05 was used as the

177

threshold for significance.

178

Results and Discussion

179

For a given lake, there was no significant difference (P > 0.05) among years in Se

180

concentrations in yellow perch muscle (except for a slight difference between 2011

181

and 2012 for Kelly Lake fish) and so we combined data for the three sampling years.

7 ACS Paragon Plus Environment

Environmental Science & Technology

182

Selenium concentrations in yellow perch muscle varied widely among

183

individuals from a given lake (Figure 1, left panels; SI, Table S2). In Lakes Dufault

184

and Kelly, Se concentrations increased significantly (P < 0.05) with increasing fish

185

weight up to ~28 g and ~20 g, respectively (Figure 1, left panels), as determined by

186

regression analysis (see Statistical Tests). Above this weight, Se concentrations in

187

yellow perch muscle (Figure 1, left panels) either declined somewhat (Lake Dufault)

188

or remained stable (Kelly Lake). No significant trends between Se concentrations and

189

fish weight were observed in Lakes Osisko and Rouyn (Figure 1, left panels).

190

In fish species other than yellow perch, size and Se concentrations are

191

reported to show either a positive relationship (several marine predators36 and

192

trout37), a negative relationship (catfish38 and halibut37), or no relationship (eels,39

193

carp,40 common carp, green sunfish, bluegill38 and Arctic char41). The lack of

194

consistent trends between fish size and Se concentrations among fish species, as well

195

as for the same fish species at different sites (our study), suggests that fish weight is

196

not a good predictor of Se concentrations in fish tissues. Indeed, fish feeding habits

197

have been shown to control Se concentrations in two sympatric fish species that

198

depended on either a high-Se (bivalve based) or low-Se (crustacean based) food

199

chain.29 The fact that trophic transfer factors ([Se]predator / [Se]prey) tend to be

200

similar among fish species42 also supports the idea that feeding habits are the major

201

determinant of fish Se concentrations. In general, the high efficiency of Se

202

assimilation by fish and invertebrates (usually > 50%)42-45 coupled with their low Se

203

uptake rates from solution tends to favor food as their main Se source.26,42,46

204

To determine yellow perch feeding habits and their relationship to individual

205

size, we measured S stable isotopes in this fish species and its prey. There was no

206

significant difference (P > 0.05) between the mean (±SD) δ34S signatures of yellow

207

perch muscle (-3.0 ± 1.3 ‰) and those of the prey in their guts (-2.8 ± 2.9 ‰), which

208

is consistent with that fact that there is little fractionation of sulfur isotopes between

209

consumers and their food.32,47

210

The δ34S signatures of zooplankton in Lakes Kelly and Rouyn were

211

significantly more positive than those of sediment-feeding chironomids (grey bars in

212

lower panel of Figure 2). The trend was the same in Lakes Dufault and Osisko, 8 ACS Paragon Plus Environment

Page 8 of 25

Page 9 of 25

Environmental Science & Technology

213

although the difference between these prey groups was not significant, although the

214

δ34S signatures of chironomids and plankton-feeding minnows were significantly

215

different in these lakes (grey bars in upper panel of Figure 2). In both pairs of lakes,

216

the mean δ34S signatures of invertebrates feeding at the sediment-water interface were

217

intermediate between those of zooplankton and burrowing chironomids (grey bars in

218

Figure 2).

219

Selenium concentrations also differed between planktonic and benthic prey,

220

with the former being lower than the latter in both pairs of lakes (solid bars in Figure

221

2). In Lakes Dufault and Osisko, there was also a significant difference in Se

222

concentrations between zooplankton and epibenthic invertebrates (solid bars in the

223

upper panel of Figure 2).

224

Since zooplankton tend to have more positive δ34S signatures and lower Se

225

concentrations than benthic invertebrates (Figure 2), small yellow perch (< 10 g)27

226

and minnows feeding on zooplankton should mirror these trends. This was the case in

227

Lakes Dufault and Kelly where Se concentrations tended to be lower and δ34S

228

signatures higher in small yellow perch than in those of intermediate size up to ~24 g

229

(Figure 1; mean of 28 g and 20 g, as explained above). However, this trend for small

230

fish was not seen in the two other lakes (Osisko and Rouyn) and in all lakes there was

231

considerable variability among small yellow perch. This variability is likely

232

explained by the fact that in our study lakes small (< 10 g) yellow perch fed on both

233

zooplankton (6 fish from 3 lakes) and benthos (9 fish from 4 lakes; SI, Table S2). For

234

example, in Kelly Lake, we collected a 1 g fish that had consumed 22 chironomids

235

whereas an 8 g fish had only zooplankton in its stomach (SI, Table S2).

236

Intermediate-sized yellow perch (10 to 24 g) from Lakes Dufault and Kelly

237

had lower δ34S signatures and higher Se concentrations than did smaller fish (Figure

238

1), which suggests that as fish age they eat less zooplankton likely because this prey

239

type provides less individual biomass than do benthic invertebrates.27 Above a mean

240

mass of ~24 g (see above), the few yellow perch that we collected had Se

241

concentrations (Figure 1, left panels) that tended to either remain stable (Kelly Lake)

242

or decline somewhat (Lake Dufault). This difference is likely explained by the types

243

of prey available to yellow perch in a given lake. If, on the one hand, small fish such 9 ACS Paragon Plus Environment

Environmental Science & Technology

244

as minnows are not abundant (as in Lake Kelly, where no minnows were found in gut

245

contents; SI, Table S2), yellow perch are likely to remain dependent on benthos and

246

thus their δ34S signatures and Se concentrations should change little as they grow

247

(assuming no change in fish growth rates). On the other hand, in lakes where

248

minnows are more readily available as prey (as in Lake Dufault, where 4 yellow

249

perch had minnows in their guts; SI, Table S2), it would be energetically

250

advantageous for yellow perch to shift from feeding on benthic invertebrates to

251

feeding on minnows.27,28 If these prey fish are planktivores, then Se concentrations in

252

large yellow perch are likely to decline (as in Lake Dufault; Figure 1), whereas if

253

prey fish are benthivores then the Se concentrations of large yellow perch are likely

254

to remain stable (assuming no change in fish growth rates). The Se concentrations of

255

channel catfish are also reported to change with fish size, that is, Se concentrations

256

decline as the fish’s diet shifts from benthic to non-benthic animals.38

257

Ignoring differences in yellow perch weight, there was a clear negative

258

relationship between muscle Se concentrations and δ34S signatures. On the one hand,

259

lower Se concentrations and more positive δ34S signatures were associated with

260

yellow perch that depended on zooplankton, either directly as prey, or indirectly

261

when feeding on planktivorous minnows (Figure 3, closed symbols; SI, Table S3).

262

On the other hand, higher Se concentrations and more negative δ34S signatures were

263

associated with yellow perch feeding on benthic chironomids or benthivorous

264

minnows (Figure 3, closed symbols; SI, Table S3). Prey of yellow perch followed

265

similar trends (Figure 3, open symbols; SI, Table S3). Trout are also reported to have

266

higher Se concentrations when feeding on sediment-based food webs than when

267

feeding on periphyton exposed to Se in the water column.48 Likewise, lake chub are

268

reported to have higher Se concentrations when feeding on free-living, unidentified,

269

benthic invertebrates compared to when they feed only on Chironomus larvae raised

270

in the laboratory in the absence of sediments.49 The extensive review of Janz et al.50

271

discusses the importance of the sediment compartment as a source of Se for lake food

272

webs.

273 274

Higher Se concentrations in benthic as opposed to planktonic invertebrates, and thus in the fish that feed on them, could be explained in two ways. First, Se 10 ACS Paragon Plus Environment

Page 10 of 25

Page 11 of 25

Environmental Science & Technology

275

concentrations could be higher in sediments than in particles in the water column

276

(mainly algae). Although total Se concentrations did not differ among these particle

277

types (SI, Table S1), their concentrations of bioavailable Se could be different. This

278

possibility is supported by the fact that sympatric Chironomus species differ in their

279

trace element concentrations depending on whether they feed mainly on deep anoxic

280

sediment or on surface oxic sediment.51,52 In the case of Se, Chironomus species

281

feeding on deep sediments had Se concentrations that were double those of species

282

feeding on surface sediments in spite of the fact that total Se concentrations were

283

similar in sediments from the two zones.53 The similar physiology of these species

284

was purported to rule out physiological differences as the main factor explaining

285

differences in their trace-element concentrations.51,52,53 Higher Se bioavailability in

286

anoxic sediments than in oxic sediments is likely dependent on their relative

287

concentrations of organic Se, which tends to be more bioavailable than inorganic

288

Se.13,54,55 For example, organic selenide produced by diatoms and sedimentary

289

bacteria is reported to be more bioavailable to bivalves than is abiotically-formed

290

elemental Se56 and Chironomus larvae have been shown to take up organic selenide

291

in sediments to a greater extent than inorganic elemental Se.57 Abiotic reactions at the

292

surface of sediments that reduce selenite to elemental Se58 and then biotic processes56

293

that use it to produce organic Se in anoxic sediment (30 to 40% of total Se in

294

sediments)50,59 would favor greater Se bioavailability to Chironomus and other

295

invertebrates that feed directly or indirectly (via predation) on anoxic sediments.

296

Physiological differences among various invertebrate groups could also

297

explain in part differences in their Se concentrations. Thus the crustacean zooplankter

298

Daphnia is reported to lose from 15-50% of its Se daily through the production of

299

offspring,60 whereas there would be no such loss from immature larval insects. This

300

difference could explain in part the lower Se concentrations that we measured in

301

zooplankton crustaceans compared to those in benthic insects. Overall, a combination

302

of physiological, behavioral and geochemical factors is likely to explain differences

303

in the Se concentrations of various invertebrate groups from our study lakes.

304 305

Selenium concentrations in yellow perch muscle were strongly correlated with those of prey in their gut contents (Figure 4), which suggests that prey Se 11 ACS Paragon Plus Environment

Environmental Science & Technology

306

concentrations are the main determinant of those in yellow perch. However, scatter

307

around the regression line in Figure 4 could be explained by physiological factors

308

such as yellow perch growth rates. Since the regression line in Figure 4 is above the 1

309

to 1 line, Se is biomagnified between yellow perch and its prey. The trophic transfer

310

factor for yellow perch (1.6 ± 0.7, SD, n = 51) is at the high end of those reported for

311

various fish species (0.5 to 1.6).42 The equation for the regression in Figure 4, along

312

with values for sediment-feeding invertebrates such as Chironomus or tubificid

313

oligochaetes, could be used to estimate maximum Se concentrations in yellow perch

314

from other lakes. Although the highest Se concentrations that we measured were in

315

Chironomus larvae, Se concentrations in the oligochaete Tubifex tubifex are reported

316

to be higher than those of Chironomus riparius raised in the same sediments,43 which

317

suggests that tubificids (uncommon in our study lakes) could also be used for

318

estimating maximum Se exposure to yellow perch.

319

Our results suggest that benthivorous fish are likely to attain Se concentrations

320

that are from 2 to 5 times higher than those of sympatric planktivorous fish, which

321

would lead to a correspondingly higher risk of Se toxicity for fish feeding on benthos.

322

To assess the likelihood that yellow perch in our study lakes are suffering toxic

323

effects, we compared their Se concentrations to those reported to induce deformities

324

in the taxonomically-related Centrarchidae.1 Since the Se concentrations reported for

325

this sister family of the Percidae are in terms of whole fish, we multiplied our values

326

for Se in muscle by 0.75 (as reported for the pike Esox lucius).61 The highest Se

327

concentration that we measured in yellow perch muscle (42 µg g-1) corresponds to a

328

whole body Se concentration of 32 µg g-1, which is the value at which deformities

329

occur in 10% of Centrarchidae,1 whereas the mean Se concentration in yellow perch

330

from Lakes Kelly and Rouyn (27 ± 8 µg g-1 in muscle; 20 µg g-1 in whole body)

331

corresponds to 6% deformed fish and the low mean Se value for fish from Lakes

332

Dufault and Osisko (14 ± 4 µg g-1 in muscle; 11 µg g-1 in whole fish) suggests that

333

deformities would not be above background level in these lakes.1

334

The likelihood of toxic effects due to Se exposure can also be assessed by

335

comparing our yellow perch Se measurements to the ovarian Se concentration that is

336

reported to produce reproductive toxicity in 10% of fish of a given species (20-25 µg 12 ACS Paragon Plus Environment

Page 12 of 25

Page 13 of 25

Environmental Science & Technology

337

Se g-1 dry weight for more than 10 Canadian fish species).62 To estimate Se

338

concentrations in yellow perch gonads, we assumed that muscle Se concentrations are

339

155% of those in gonads (SI, Figure S1; values for Se in ovaries fell within those for

340

testes) such that yellow perch having a muscle Se concentration > 31 µg g-1 would be

341

at risk of Se reproductive toxicity. Using this threshold, we estimate that about one-

342

third (38% and 33%, respectively) of the yellow perch from Lakes Kelly and Rouyn,

343

are at risk of Se toxicity, whereas no Se concentrations exceeded this threshold value

344

in the other two lakes. The fact that Se concentrations in the biota and water of most

345

lakes in our study areas tend to be lower than those in Lakes Kelly and Rouyn,11,12,13,

346

our study

347

these lakes. Although we did not measure yellow perch population sizes in our four

348

study lakes, we note that the lowest catch per unit effort (data not shown) was in the

349

two lakes in which Se toxicity is most likely to occur (Lakes Kelly and Rouyn).

350

suggests that Se is not a major risk factor for yellow perch in the majority of

To determine the suitability of yellow perch from our study lakes as a food

351

source, we estimated consumption limits for humans. The maximum recommended

352

Se consumption is 400 µg of Se per day63 or 12 mg per month. Given the mean Se

353

values in yellow perch from Lakes Dufault and Osisko (14 ± 4 µg g-1 dry weight or

354

2.8 µg g-1 wet weight; dry to wet weight factor of 564), a person consuming 12 filets

355

per month (filet size of 0.23 kg fresh weight) would reach this limit, whereas in Lakes

356

Kelly and Rouyn (27 ± 8 µg g-1 dry weight) only 6 filets would be needed to reach

357

the maximum recommended Se intake.

358

Overall, our results suggest that freshwater fish species whose food webs are

359

based on planktonic organisms will have lower Se concentrations than fish dependent

360

on sediment-based food webs. Furthermore, since many freshwater fish species feed

361

mainly on zooplankton as larvae and then shift to benthic invertebrates and eventually

362

fish, their Se exposure is likely to change as they grow. We observed this trend for

363

yellow perch, but it is also likely the case for other sports fish such as walleye

364

(Sander vitreus), largemouth bass (Micropterus salmoides), pike (Esox lucius) and

365

lake trout (Salvelinus namaycush) that show similar age-related shifts in diet. Lastly,

366

our results suggest that changes in Se exposure within the life of a given species can

367

be site specific depending on the densities of various prey types in a given lake. Thus 13 ACS Paragon Plus Environment

Environmental Science & Technology

368

the risk of Se toxicity is likely to be highly variable depending on the fish species, its

369

life stage and the site at which it is collected. Given the importance of diet in

370

determining Se exposure and effects on fish, an efficient technique is needed to

371

determine the dependence of various fish species and life stages on planktonic and

372

benthic food chains. Our results suggest that measurements of S stable isotopes in

373

fish fill this need.

374

Acknowledgments

375

We acknowledge John Gunn and the Cooperative Freshwater Ecology Unit at the

376

Vale Living with Lakes Center in Sudbury, Ontario for their help in the field. The

377

assistance of Maikel Rosabal, Julien Lacharité, Nicolas Fabien-Ouellet and Isabelle

378

Lavoie is also greatly appreciated. We thank Marc Amyot, Marie-Noële Croteau and

379

Claude Fortin for their comments on an earlier version of the manuscript. Our study

380

was supported by funding from the Natural Sciences and Engineering Research

381

Council of Canada.

382

Supporting Information Available

383

Table S1 gives information about lake locations, lakewater chemistry and Se

384

concentrations in sediments and in microplankton. Table S2 reports the length, wet

385

weight and stomach contents (when present) of all yellow perch (P. flavescens) that

386

we collected as well as Se concentrations and S stable isotopic signatures in prey and

387

in yellow perch dorsal muscle. Table S3 gives details about the correlations presented

388

in Figure 3. Figure S1 compares Se concentrations in the various tissues and organs

389

of P. flavescens from Lohi Lake, Ontario. This material is available free of charge via

390

the internet at http://pubs.acs.org.

14 ACS Paragon Plus Environment

Page 14 of 25

Page 15 of 25

Environmental Science & Technology

Literature Cited 1. 2. 3. 4. 5.

6. 7. 8.

9.

10.

11.

12.

13. 14.

Lemly, D. Teratogenic effects of selenium in natural populations of freshwater fish. Ecotoxicol. Environ. Saf. 1993, 26, 181–204. Presser, T. S.; Ohlendorf, H. M. Biogeochemical cycling of selenium in the San Joaquin Valley, California, USA. Environ. Manag. 1987, 11, 805–821. Lemly, A. D. Symptoms and implications of selenium toxicity in fish: the Belews Lake case example. Aquat. Toxicol. 2002, 57, 39–49 Nriagu, J. O.; Wong, H. K. Selenium pollution of lakes near the smelters at Sudbury, Ontario. Nature 1983, 301, 55–57. Luoma, S. N.; Rainbow P. S. Metal contamination in aquatic environments: science and lateral management. Cambrige University Press, New York. 2008. 573 pp. Hare, L. Aquatic insects and trace-metals – Bioavailability, bioaccumulation and toxicity. Crit. Rev. Toxicol. 1992, 22,327–369. Gunn, J. M. Restoration and recovery of an industrial region. Springer-Verlag, New York. 1995. 358 pp. Borgmann, U.; Norwood, W. P.; Reynoldson, T. B.; Rosa, F. Identifying cause in sediment assessments: bioavailability and the sediment quality triad. Can. J. Fish. Aquat. Sci. 2001, 58, 950–960. Campbell, P. G. C.; Hontela, A.; Rasmussen, J. B.; Giguère, A.; Gravel, A.; Kraemer, L.; Kovesces, J.; Lacroix, A.; Levesque, H.; Sherwood, G. Differentiating between direct (physiological) and food-chain mediated (bioenergetic) effects on fish in metal-impacted lakes. Human Ecol. Risk Assess. 2003, 9, 847–866. Borgmann, U.; Nowierski, M.; Grapentine, L. C.; Dixon, D. G. Assessing the cause of impacts on benthic organisms near Rouyn-Noranda, Quebec. Environ. Pollut. 2004, 129, 39–48. Pyle, G. G.; Rajotte, J. W.; Couture, P. Effects of industrial metals on wild fish populations along a metal contamination gradient. Ecotox. Environ. Safety 2005, 61, 287–312. Belzile, N.; Chen, Y. W.; Yang, D. Y.; Thi Truong, H. Y.; Zhao, Q. X. Selenium bioaccumulation in freshwater organisms and antagonistic effect against mercury assimilation. Environ. Bioindicat. 2009, 4, 203–221. Ponton, D. E.; Hare, L. Relating selenium concentrations in a planktivore to selenium speciation in lakewater. Environ. Pollut. 2013, 176, 254–260. Croteau, M. N.; Hare, L.; Tessier, A. Increases in food web cadmium following reductions in atmospheric inputs to some lakes. Environ. Sci. Technol. 2002, 36, 3079–3082.

15 ACS Paragon Plus Environment

Environmental Science & Technology

15. Belzile, N.; Morris, J. R. Lake Sediments: Sources or sinks of industrially mobilized elements? In: Restoration and recovery of an industrial region. Gunn, J. M. Ed.; Springer-Verlag, New York. 1995. pp 183-193. 16. Keller, W.; Heneberry, J. H.; Gunn, J. M. Effects of emission reductions from the Sudbury smelters on the recovery of acid- and metal-damaged lakes. J. Aquat. Ecosyst. Stress Recovery 1999, 6, 189–198. 17. Gallon, C.; Gobeil, C.; Tessier, A.; Carignan, R. Stable Pb isotopes and PAHs as indicators of lead contamination sources in a lake of the Rouyn-Noranda area. J. Phys. IV 2006, 107, 505–508. 18. Croisetière, L.; Hare, L.; Tessier, A.; Cabana, G. Sulphur stable isotopes can distinguish trophic dependence on sediments and plankton in boreal lakes. Freshwater Biol. 2009, 54, 1006–1015. 19. Warren, L. A.; Tessier, A.; Hare, L. Modelling cadmium accumulation by benthic invertebrates in situ: The relative contributions of sediment and overlying water reservoirs to organism cadmium concentrations. Limnol. Oceanogr. 1998, 43, 1442–1454. 20. Wesolek, B. E.; Genrich, E. K.; Gunn, J. M. Use of littoral benthic invertebrates to assess factors affecting biological recovery of acid- and metal-damaged lakes. J. N. Amer. Benthol. Soc. 2010, 29, 572–585. 21. Janz, D. M.; Liber, K.; Pickering, I. J.; Wiramanaden, C. I. E.; Weech, S. A.; Gallego-Gallegos, M.; Driessnack, M. K.; Franz, E. D.; Goertzen, M. M.; Phibbs, J.; Tse, J. J.; Himbeault, K. T.; Robertson, E. L.; Burnett-Seidel, C.; England, K.; Gent, A. Integrative assessment of selenium speciation, biogeochemistry, and distribution in a northern coldwater ecosystem. Integr. Environ. Assess. Manage. 2014, 10, 543–554. 22. Farag, A. M.; Woodward, D. F.; Goldstein, J. N.; Brumbaugh, W.; Meyer, J. S. Concentrations of metals associated with mining waste in sediments, biofilm, benthic macroinvertebrates, and fish from the Coeur d’Alene river basin, Idaho. Arch. Environ. Contam. Toxicol. 1998, 34, 119–127. 23. Pierron, F.; Normandeau, E.; Amery Defo, M.; Campbell, P. G. C.; Bernatchez, L.; Couture, P. Effects of chronic metal exposure on wild fish populations revealed by high-throughput cDNA sequencing. Ecotoxicology 2011, 20, 1388– 1399. 24. Kraemer, L. D.; Campbell, P. G. C.; Hare, L.; Auclair, J. C. A field study examining the relative importance of food and water as sources of Cd for juvenile yellow perch (Perca flavescens). Can. J. Fish. Aquat. Sci. 2006, 63, 549–557. 25. Lapointe, D.; Couture, P. Influence of the route of exposure on the accumulation and subcellular distribution of nickel and thallium in juvenile fathead minnows (Pimephales promelas). Arch. Environ. Contam. Toxicol. 2009, 57, 571–580 26. Stewart, A. R.; Grosell, M.; Buchwalter, D.; Fisher, N.; Luoma, S. N.; Mathews, T.; Orr, P.; Wang, W. X. Bioaccumulation and trophic transfer of selenium. In Ecological assessment of selenium in the aquatic environment. Chapman P. M.; 16 ACS Paragon Plus Environment

Page 16 of 25

Page 17 of 25

Environmental Science & Technology

27.

28. 29.

30.

31.

32. 33.

34. 35.

36.

37.

38.

39.

Adams, W. J.; Brooks, M. L.; Delos, C. G.; Luoma, S. N.; Maher, W.A.; Ohlendorf, H. M.; Presser, T. S.; Shaw, D. P., Eds.; CRC Press, New York. 2010. pp 93-139. Rasmussen, J. B.; Gunn, J. M.; Sherwood, G. D.; Iles, A.; Gagnon, A.; Campbell, P. G. C.; Hontela, A. Direct and indirect (foodweb mediated) effects of metal exposure on the growth of yellow perch (Perca flavescens): Implications for ecological risk assessment. Human Ecol. Risk Assess. 2008, 14, 317–350. Boisclair, D; Leggett, W. C. Among-population variability of fish growth. 2. Influence of prey type. Can. J. Fish. Aquat. Sci. 1989, 46, 468–482. Stewart, A. R.; Luoma, S. N.; Schlekat, C.; Doblin, M. A.; Hieb, K. A. Food web pathway determines how selenium affects aquatic ecosystems: A San Francisco Bay case study. Environ. Sci. Technol. 2004, 38, 4519–4526. Croisetière, L.; Hare, L.; Tessier, A.; Cabana, G. Sulphur stable isotopes can distinguish trophic dependence on sediments and plankton in boreal lakes. Freshwater Biol. 2009, 54, 1006–1015. Brunner, B.; Bernasconi, S. M. A revised isotope fractionation model for dissimilatory sulfate reduction in sulfate reducing bacteria. Geochim. Cosmochim. Acta 2005, 69, 4759–4771. Peterson B. J.; Fry, B. Stable isotopes in ecosystem studies. Annu. Rev. Ecol. Syst. 1987, 18, 293–320. Cole, J. J.; Carpenter, S. R.; Pace, M. L.; Van de Bogert, M. C.; Kitchell, J. L.; Hodgson, J. R. Differential support of lake food webs by three types of terrestrial organic carbon. Ecol. Letters 2006, 9, 558–568. Nriagu, J. O.; Soon, Y. K. Distribution and isotopic composition of sulfur in lake sediments of northern Ontario. Geochim. Cosmochim. Acta 1985, 49, 823–834. Chapman P. M.; Adams, W. J.; Brooks, M. L.; Delos, C. G.; Luoma, S. N.; Maher, W.A.; Ohlendorf, H. M.; Presser, T. S.; Shaw, D. P., Eds. Ecological assessment of selenium in the aquatic environment. CRC Press, New York. 2010. 339 pp. Kojadinovic, J.; Potier, M.; Le Corre, M.; Cosson, R. P.; Bustamante, P. Bioaccumulation of trace elements in pelagic fish from the western Indian ocean. Environ. Poll. 2007, 146, 548–566. Burger, J.; Gochfeld, M.; Jeitner, C.; Pittfield, T.; Donio, M. Heavy metals in fish from the Aleutians: Interspecific and locational differences. Environ. Res. 2014, 131, 119–130. May, T. W.; Fairchild, J. F.; Petty, J. D.; Walther, M. J.; Lucero, J.; Delvaux, M.; Manring, J.; Armbruster, M. An evaluation of selenium concentrations in water, sediment, invertebrates, and fish from the Solomon River Basin. Environ. Monit. Assess. 2008, 137, 213–232. Polak-Juszczak, L.; Robak, S. Mercury toxicity and the protective role of selenium in eel, Anguilla anguilla. Environ. Sci. Pollut. Res. 2015, 22, 679–688. 17 ACS Paragon Plus Environment

Environmental Science & Technology

40. Levengood, J. M.; Soucek, D. J.; Sass, G. G.; Dickinson, A.; Epifanio, J. M. Elements of concern in fillets of bighead and silver carp from the Illinois river, Illinois. Chemosphere 2014, 104, 63–68. 41. Gantner, N.; Power, M.; Babaluk, J. A.; Reist, J. D.; Köck, G.; Lockhart, L. W.; Solomon, K. R.; Muir, D. C. G. Temporal trends of mercury, cesium, potassium, selenium, and thallium in Arctic char (Salvelinus alpinus) from lake Hazen, Nunavut, Canada: Effects of trophic position, size, and age. Environ. Toxicol. Chem. 2009, 28, 254–263. 42. Presser, T. S.; Luoma, S. N. A methodology for ecosystem-scale modeling of selenium. Integr. Environ. Assess. Manag. 2010, 6, 685–710. 43. Dubois, M.; Hare, L. Selenium assimilation and loss by an insect predator and its relationship to Se subcellular partitioning in two prey types. Environ. Poll. 2009, 157, 772–777. 44. Schlekat, C. E.; Lee, B. G.; Luoma, S. N. Assimilation of selenium from phytoplankton by three benthic invertebrates: effect of phytoplankton species. Mar. Ecol. Progr. Ser. 2002, 237, 79–85. 45. Reinfelder, J. R.; Fisher, N.S. The assimilation of elements ingested by marine copepods. Science 1991, 251, 794–796. 46. Wang, W. X.; Fisher, N. S. Delineating metal accumulation pathways for marine invertebrates. Sci. Total Environ.1999, 238, 459–472. 47. McCutchan, J. H.; Lewis, W. M.; Kendall, C.; McGrath, C. C. Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulphur. Oikos 2003, 102, 378–390. 48. Orr, P. L.; Guiguer, K. R.; Russel, C. K. Food chain transfer of selenium in lentic and lotic habitats of a western Canadian watershed. Ecotox. Environ. Safe. 2006, 63, 175–188. 49. Phibbs, J.; Franz, E.; Hauck, D.; Gallego-Gallegos, M.; Tse, J. J.; Pickering, I. J.; Liber, K.; Janz, D. M. Evaluating the trophic transfer of selenium in aquatic ecosystems using caged fish, X-ray absorption spectroscopy and stable isotope analysis. Ecotox. Environ. Safe. 2011, 74, 1855–1863. 50. Martin, S.; Proulx, I.; Hare, L. Explaining metal concentrations in sympatric Chironomus species. Limnol. Oceanogr. 2008, 53, 411–419. 51. Proulx, I.; Hare, L. Differences in feeding behaviour among Chironomus species revealed by measurements of sulphur stable isotopes and cadmium in larvae. Freshwater Biol. 2014, 59, 73–86. 52. Proulx, I. Évaluation du potentiel d'utiliser les larves de Chironomus comme biomoniteurs de la biodisponibilité des éléments traces dans les sédiments. Doctorat en Sciences de l'eau, Université du Québec, Institut national de la recherche scientifique. Québec, QC, Canada. 2014. 275 pp. 53. Baines, S. B.; Fisher, N. S.; Doblin, M. A.; Cutter, G. A. Uptake of dissolved organic selenides by marine phytoplankton. Limnol. Oceanogr. 2001, 46, 1936– 1944. 18 ACS Paragon Plus Environment

Page 18 of 25

Page 19 of 25

Environmental Science & Technology

54. Besser, J. M.; Canfield, T. J.; Lapoint, T. W. Bioaccumulation of organic and inorganic selenium in a laboratory food-chain. Environ. Toxicol. Chem. 1993, 12, 57–72. 55. Schlekat, C. E.; Dowdle, P. R.; Lee, B. G.; Luoma, S. N.; Oremland, R. S. Bioavailability of particle-associated Se to the bivalve Potamocorbula amurensis. Environ. Sci. Technol. 2000, 34, 4504–4510. 56. Franz, E. D.; Wiramanaden, C. I. E.; Gallego-Gallegos, M.; Tse, J. J.; Phibbs, J.; Janz, D. M.; Pickering, I. J.; Liber, K. An in situ assessment of selenium bioaccumulation from water-, sediment-, and dietary-exposure pathways using caged Chironomus dilutus larvae. Environ. Toxicol. Chem. 2013, 32, 2836–2848. 57. Chen, Y. W.; Thi Truong, H. Y.; Belzile, N. Abiotic formation of elemental selenium and role of iron oxide surfaces. Chemosphere 2009, 74, 1079–1084. 58. Martin, A. J.; Simpson, S.; Fawcett, S.; Wiramanaden C. I. E.; Pickering, I. J.; Belzile, N.; Chen, Y. W.; London, J.; Wallschläger, D. Biogeochemical mechanisms of selenium exchange between water and sediments in two contrasting lentic environments. Environ. Sci. Technol. 2011, 45, 2605–2612. 59. Yu, R. Q.; Wang, W. X. Kinetic uptake of bioavailable cadmium, selenium, and zinc by Daphnia magna. Environ. Toxicol. Chem. 2002, 21, 2348–2355. 60. Muscatello, J. R.; Bennett, P. M.; Himbeault, K. T.; Belknap, A. M.; Janz, D. M. Larval deformities associated with selenium accumulation in northern pike (Esox lucius) exposed to metal mining effluent. Environ. Sci. Technol. 2006, 40, 6506– 6512. 61. DeForest, D. K.; Gilron, G.; Armstrong, S. A.; Robertson, E. L. Species sensitivity distribution evaluation for selenium in fish eggs: considerations for development of a Canadian tissue-based guideline. Integr. Environ. Assess. Manag. 2012, 8, 6–12. 62. Food and Agriculture Organization. Human vitamin and mineral requirements. FAO/WHO expert consultation on human vitamin and mineral requirements, 2001. http://www.fao.org/docrep/004/Y2809E/y2809e0l.htm. 63. CRESP, Consortium for Risk Evaluation with Stakeholder Participation. Note on: Wet to dry weight conversions for biota from Amchitka and Kiska. CRESP Amchitka Radionuclide Data Set, 2006. http://www.cresp.org/Amchitka/Final_WW_DW_3_13_06.pdf.

19 ACS Paragon Plus Environment

Environmental Science & Technology

Figure Captions Figure 1. Mean (±SD) Se concentrations ([Se], µg g-1 dry weight; left panels) and sulfur stable-isotope signatures (δ34S, ‰; right panels) of yellow perch muscle as compared to whole-fish wet weights (g) in our four study lakes. Fish having wet weights < 28 g and < 20 g (see text) in Lakes Dufault and Kelly, respectively, are represented by solid symbols, whereas heavier fish are represented by open squares.

Figure 2. Selenium concentrations ([Se], µg g-1 dry weight; black bars) and sulfur stable-isotope signatures (δ34S, ‰; gray bars) of prey consumed by yellow perch that were collected in either Lakes Dufault and Osisko (upper panel) or Lakes Kelly and Rouyn (lower panel). Different letters represent a significant difference between prey types. Figure 3. Selenium concentrations ([Se]; µg g-1 dry weight) as a function of sulfur stable-isotope signatures (δ34S; ‰) in yellow perch muscle (closed circles) or in their prey (open squares) from our four study lakes. All linear regressions are significant at P < 0.03. Details of these regressions are presented in SI, Table S3. Figure 4. Selenium concentrations ([Se]; µg g-1 dry weight) in the muscle of yellow perch collected from our four study lakes as a function of the Se concentrations of prey in their stomachs. The equation for the regression line is [Se]Perca flavescens = 0.9 ± 0.1 ∗ [Se]Perca flavescens prey + 7 ± 2 where P < 0.001 and uncertainties are standard errors. The dotted line represents a 1:1 relationship.

20 ACS Paragon Plus Environment

Page 20 of 25

Page 21 of 25

Environmental Science & Technology

25

0

Dufault

-1

15

-2

10

-3

5

-4

0 20

40

60 110 115 0

20

40

60 110

-5 115 6 4

Kelly

40

2

30

0

20

-2

Kelly

10

-4

0

-6 0

15

30

70

80 0

15

30

70

80

25

-2

20

-3

15

-4

10

-5

5

Osisko

Osisko

0 0

5

10 15 60

50

90

120 0

5

10

30 20 10

Rouyn

Rouyn

0 0

15

30

45

60 160180 0

15

30

-7 120 -1 -2 -3 -4 -5 -6 -7 -8 60 160 180

15 60

40

45

-6

90

Perca flavescens weight (g w.w.)

Fig. 1.

21 ACS Paragon Plus Environment

34

50

0

δ S(‰)

[Se] Perca flavescens (µg g-1)

20

Dufault

Environmental Science & Technology

Fig. 2

22 ACS Paragon Plus Environment

Page 22 of 25

Page 23 of 25

Environmental Science & Technology

25

A

Osisko

20

-1

[Se] (µg g )

B

Dufault

20

r2 = 0.49

15 10

25

r2 = 0.22

15 r2 = 0.72

10

5

r2 = 0.40

5

0 0 -12 -10 -8 -6 -4 -2 0 2 4 -12 -10 -8 -6 -4 -2 50 50 C D Rouyn 2 r = 0.39 40 40 30 20

0

2

4

Kelly 2

r = 0.57

30 20

r2 = 0.40

10

10

0 -12 -10 -8 -6 -4 -2

0

2

4

r2 = 0.28

0 -12 -10 -8 -6 -4 -2

34

δ S(‰)

Fig. 3

23 ACS Paragon Plus Environment

0

2

4

-1

[Se]P. flavescens (µg g )

Environmental Science & Technology

Page 24 of 25

2

r = 0.50

40 30 20 10 0 0

10

20

30

40 -1

[Se] P. flavescens prey (µg g ) Fig. 4

24 ACS Paragon Plus Environment

Page 25 of 25

Environmental Science & Technology

25 ACS Paragon Plus Environment