SEARCH

SEARCH BY CITATION

Keywords:

  • biomass;
  • ecological impact;
  • genetic modification;
  • growth;
  • population level;
  • simulated stream;
  • transgenesis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. The potential risks of accidental or intentional introduction of transgenic organisms to nature are unknown. We have previously shown that, after being reared in the hatchery, growth hormone transgenic coho salmon Oncorhynchus kisutch can exert a stronger predation effect on natural prey under simulated natural conditions compared with hatchery-reared genetically wild conspecifics. However, when reared in a simulated natural environment, the difference between the two genotypes was greatly reduced.

2. Here, we tested if the difference in predation rate between the two genotypes after being reared in the hatchery diminishes with time spent in a simulated natural environment. Genetically wild coho salmon reared in the hatchery were size matched to younger satiation-fed (fast-growing) and same age restricted-fed (growth rates matched to wild-type) transgenic fish. These three types of predators were released into simulated natural habitats and their effects on prey survival and growth were monitored over a 2-month period.

3. Restricted-fed, but not satiation-fed, transgenic predators consumed significantly more prey (×2·7) relative to wild predators during the first month, with an increase to 3.8 times during the second period. Prey biomass decreased more in the presence of restricted-fed predators during the first period, but a faster growth of these prey during the second period compensated for the higher predation rate resulting in no significant difference in prey biomass among predators after 2 months. Transgenic predators grew more in length than wild type, but all three types converged in weight and condition over time.

4.Synthesis and applications. Behavioural phenotypes developed in the hatchery showed little plasticity after release. Hence, development of policy on the use of transgenic organisms must incorporate knowledge of the effect of environmental conditions, experienced by the organism prior to escape or release, on phenotype. The risk posed by culture-reared organisms may depend on the frequency of escapes, number of individuals escaping and age of escapees. Important to note is that wild-reared organisms may be biologically different from culture-reared and thus require separate evaluation. Our results will be important for science policy makers and regulators to consider when deciding whether to allow commercial application of transgenic species in aquaculture.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Ecological risks associated with transgenic organisms have been hypothesized for many years including the potential negative ecological consequences of fast-growing transgenic fish should they escape or be released into the natural environment (e.g. Tiedje et al. 1989; Kapuscinski & Hallerman 1991; Devlin & Donaldson 1992; Knibb 1997; Snow et al. 2005). Whereas the focus of previous work has been mainly on the fitness of the transgenic fish and the effects of transgene introgression into wild populations (e.g. Dunham et al. 1999; Muir & Howard 1999; Jhingan et al. 2003; Bessey et al. 2004; Sundström et al. 2005), less attention has been devoted to the potential consequences to other species or to the ecosystem as a whole.

Presently, studies of transgenic fish cannot be performed in nature due to the difficulty in extirpating the transgene should negative consequences arise. As an alternative, assessments of the ecological effects of transgenic fish are undertaken in contained laboratory facilities where natural environments are simulated (Devlin et al. 2006). In one such study, we compared predation effects of transgenic and wild fish that had been reared both under simulated natural conditions and in the hatchery (Sundström et al. 2007b). Results showed that transgenic and genetically wild fish reared and tested under simulated natural conditions consumed similar numbers of prey, whereas hatchery-reared transgenic fish consumed many more prey than wild fish reared in the hatchery when tested in the same simulated natural environments. Although the fish were reared for over 21 months, the predation study occurred over a 20-day period and thus we were unable to assess whether the higher predation rate of hatchery-reared transgenic fish compared with that of wild genotypes reared in the hatchery was due to differences in response to the novel conditions in the simulated natural streams or permanent phenotypic differences arising in the hatchery environment.

Transfer from the hatchery to a natural environment is a drastic change for fish and even just a few days of acclimatization can improve subsequent feeding behaviour (Jonsson et al. 1999). Hatchery-reared fish of wild genotypes often require experience before identifying live prey as potential food and they need to learn to pursue and handle the evasive prey as opposed to the formulated pellets they have been fed previously and are accustomed to (Olla et al. 1998; Sundström & Johnsson 2001; Warburton 2006). In comparison, hatchery-reared transgenic fish learn to feed on novel prey faster (Sundström et al. 2004), take less time to regain feeding position after disturbance (Sundström et al. 2003) and more quickly initiate feeding on artificial feed after transfer between hatchery environments (Tymchuk et al. 2005) when compared with hatchery-reared wild-type fish. Thus, it is important to assess the time required by hatchery-reared fish to acclimatize to new environments and determine if this process occurs more quickly in transgenic fish compared with wild fish. As such, hatchery-reared transgenic and wild fish may converge in their predation effects over time and induce similar ecological effects to their environment (Sundström et al. 2007b).

As environmental conditions influence animals’ behaviour and two genotypes may respond differently to the same conditions (genotype × environment interaction, Mackay & Anholt 2007), a single measurement cannot discern if differential responses to a novel environment between two genotypes are solely due to genotype or differential plasticity (genotype × environment interaction). If a permanent difference exists, the average predation rates and differences among predator types would remain the same indefinitely. If differences are due to transgenic fish adapting to the novel environment faster, wild fish will initially lag behind but exert an increasing predation rate over time as they acclimatize to the new environment. Eventually, the predation rates of the wild and transgenic genotypes could then converge to that of natural phenotypes.

Determining whether there is convergence in the predation effects of hatchery-reared transgenic and wild fish released into nature has important implications for the ecological impact of such transgenic fish (Miner et al. 2005). We therefore addressed this question by studying hatchery-reared transgenic and wild genotype coho salmon Oncorhynchus kisutch (Walbaum) predators and by assessing their respective predation effects within a naturalized environment after 1 and 2 months. If differential responses to a novel environment are mainly responsible for the previously observed differences in predation effects of hatchery-reared fish, we would expect to see a reduced difference between the two genotypes over time. Alternatively, the effect of the transgene on phenotypic development in the hatchery environment may produce irreversible effects on feeding behaviour and/or related phenotypic traits that persist even after transfer to a more complex environment. In this case, we would expect that differences in predation rate among predatory types (at least in rank order) would persist throughout the entire experiment.

Material and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The experiments were conducted in 2005 at the DFO/UBC Centre for Aquaculture and Environmental Research (CAER), West Vancouver, Canada, and met requirements for animal rearing as established by the Canadian Council for Animal Care (no. 05-016). This non-commercial research facility was designed to prevent the escape of genetically modified fish to the natural environment.

The present experiment was carried out using three types of size-matched (length) predators that had been reared only under hatchery conditions and fed only artificial salmon feed (Skretting Inc., Bayshore, NB, Canada) (Table 1). The objective was to examine predation effects over time among three predator types: (1) genetically wild salmon fed to satiation (W), (2) fast-growing transgenic salmon fed to satiation (Ts) and (3) transgenic salmon with wild-type growth rates due to a restricted ration diet (Tr). W and Tr were 1 year older (20 months of age) than Ts (9 months of age). Ts and Tr/W salmon must necessarily be derived from different brood years (due to the difference in growth rate). However, they are both derived from crosses involving the same transgenic strain and a random sample of wild-type salmon from the founder wild population. Further, this size matching can only be performed during a period of a few weeks when the size of Ts overlaps with W and Tr from the previous brood year.

Table 1.   Initial length, weight and condition of predator salmon and initial average weight of prey fish
 WTrTsInitial test (d.f. = 2, 20)
  1. W are hatchery-reared fish of wild genetic background fed to satiation during rearing, Tr are transgenic fish fed a diet that restricted their growth rates to match those of the W fish and Ts are transgenic fish fed to satiation resulting in a more rapid growth rate relative to the other two types and hence were 1 year younger.

  2. 1Ts had significantly higher condition than W and Tr (P < 0.001).

Predators
Length (mm)185 ± 4·3192 ± 3·5185 ± 3·5F = 1·09, = 0·36
Weight (g)77·8 ± 4·783·1 ± 3·693·1 ± 5·3F = 2·01, = 0·16
Condition11·23 ± 0·041·18 ± 0·021·46 ± 0·02F = 28·0, < 0·001
Prey
Average weight (g)1·40 ± 0·031·43 ± 0·041·43 ± 0·04F = 0·21, = 0·82

W predators were obtained by mixing gametes from four wild females and four wild males caught in the Chehalis River in south-western British Columbia, Canada. Transgenic genotypes of coho salmon were initially produced by microinjecting the OnMTGH1 growth hormone gene construct into eggs from wild parents from the Chehalis River (Devlin et al. 1994). Stable and characterized transgenic strains were subsequently developed through crosses with wild fish from the Chehalis River at each generation and thus transgenic fish contain on average the same genetic background as the wild fish except for the presence of the OnMTGH1 transgene. Details of the production and subsequent performance of transgenic fish can be found in Devlin et al. (2004). Experimental transgenic fish in the present study were of the F6 (Tr) and F7 (Ts) generation progeny derived from a single founder line (M77). Tr were produced by crossing four fathers homozygous for the transgene insert with the same four wild genotype females as was used to produce wild fish. Ts were produced by crossing five fathers homozygous for the transgene with five wild females from the Chehalis River. It is important to note that the difference in daily specific growth rate (SGR) between transgenic and wild salmon is approximately two- to threefold and can rapidly result in large differences in size. This difference in growth rate is many times greater than interfamily variation observed among wild salmon.

Fish were studied in indoor oval 200-L experimental tanks (80 l × 50 w × 50 h cm) containing coarse gravel, numerous large rocks, and a hollow block that provided hiding places for both prey and predators. Water diverted from the nearby Cypress Creek was used in a flow-through system at a rate of 2–4 L min−1 creating water current similar to that of a slow-moving section in a stream. Water temperature was approximately 12 °C at the beginning of the experiment and dropped to 5 °C at termination due to normal seasonal climate changes. Artificial lighting with simulated natural wavelengths followed the natural photoperiod for the season.

On the first experimental day (17 September), 20 rainbow trout Oncorhynchus mykiss (Walbaum) fry were introduced into each of the 24 landscaped experimental tanks (Table 1). The next day, one coho salmon predator (Ts, Tr or W) was added to each tank which resulted in eight replicates for each predator type. Every other day for the remaining study period, trout prey were fed commercial salmon feed of a size too small for the large predators to profitably feed on. The number of visible prey fish was counted when possible, but, because rainfall periodically made the water murky, fish could not reliably be counted some days. Counting visible prey allowed for the monitoring of behavioural responses of prey to the presence of the three predator types.

Except for the prey trout in the experiment, predators were not provided any additional food, although uncontrolled amounts of various natural prey (mainly mayfly and caddis fly larvae) entered with the creek water. A fine-meshed screen prevented the trout from being washed out with the outflow. After 1 month (14 October), surviving prey were counted and weighed as a group before being returned to their tank and predators were measured in length and weighed. The same procedure was repeated after another month (12 November) after which the experiment was terminated.

Statistical analysis

Differences between predator types in the number of prey consumed were tested with generalized linear models. A negative binomial distribution with log-link function was used allowing spss 16.02 (http://www.spss.com) to adjust for overdispersion by using the dispersion parameter k. Manual iteration was used to determine the k that provided a model deviance closest to 1 (k = 0.8 for period 1 and k = 0.7 for period 2). In addition, the robust estimator was used for the covariance matrix (Norušis 2007). Prey body weight at the beginning of each period was used as a covariate. Because predation rate for the second period would depend on the number of prey consumed during the first period, an offset variable was used as the logarithm of the number of prey still alive at the beginning of this period.

Effects of predator type on SGR of prey were tested with one-way analysis of covariance (ancova) with prey body weight at the start of the period as covariate. SGR was calculated as SGR = 100 × (ln W1 − ln W0)/t, where W0 and W1 are the weights in September and October for the first period and October and November for the second period and t is time in days elapsed between the two dates.

The change in prey biomass for each period was log-transformed to improve normality and then tested with a one-way anova. The change was calculated as the biomass (number of prey × prey weight) at the end of a period minus the biomass at the beginning of the period.

One-way ancovas were used to test for differences between predator types in the specific growth in weight (SGRW) and specific growth in length (SGRL) calculated in the same way as for fry weight above, and change in condition factor (CF) between initial and subsequent sampling points (calculated as the difference between the two CF = W × L−3). Length, weight and CF at the beginning of each period were used as a covariate in the analysis of SGR in length and weight and change in CF respectively. For all analyses, interactions between factors and the covariate were first assessed, but none were significant and therefore excluded from further analysis to increase power.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Predation effects

During the first month, restricted-fed transgenic fish (Tr) consumed 2.7 times more prey (B = 0·99 ± 0·32 SE, = 0·002) than wild predators (Tr, 39 ± 6%; W, 16 ± 3%) with neither of the two differing from satiation-fed transgenic predators (Ts, 24 ± 8%) (Table 2, Fig. 1a). During the second period, absolute proportion of prey consumed was similar to that of the first period for each of the three predators with 35 ± 9% for Tr, 15 ± 6% for W and 18 ± 9% for Ts (Fig. 1a). When taking into consideration the differential predation during the first period and the size of prey at the onset of the second period, Tr predators consumed 3.8 times more prey (B = 1.33 ± 0.67 SE, = 0.047) than W. Tr also consumed more prey than Ts, with no difference between W and Ts (Table 2). During the second period, predators with larger prey at the onset of the period also consumed fewer prey (Table 2). It is important to note that variability increased from period one to period two and a large variation between individual predators was detected with W consuming between 0 and 9, Tr between 5 and 16 and Ts between 0 and 15 of the available 20 fry during the 2-month experimental period. Nevertheless, despite this stochasticity among replicates, differential predation rates (inline image, = 0.029) resulted in fewer prey being alive at the end of the experiment in the presence of restricted-fed transgenic predators compared with both satiation-fed transgenic (= 0.038) and wild (= 0.017) predators with no difference between the latter two predators (Fig. 1a).

Table 2.   Statistical tests of differences in predation rate (prey consumption), specific growth rate (SGR) of prey, change in prey biomass and SGR for weight and length and change in body condition of wild and transgenic coho salmon predators during periods 1 and 2
  Genotype (d.f. = 2, 19)Covariate (d.f. = 1, 19)Pair-wise post hoc test (P)
PeriodW–TrW–TsTr–Ts
  1. Predators were hatchery-reared genetically wild fish (W), age-matched transgenic fish with restricted growth (Tr) and younger transgenic fish fed to satiation (Ts). Covariates for predation rate and prey growth rate were prey size and initial body measurements (length, weight or CFs) at the beginning of the period for predator analyses.

Prey
Predation rate1χ= 9·3; = 0·010χ= 0·8; = 0·380·0030·260·14
2χ= 9·1; = 0·011χ= 12·3; < 0·0010·0490·970·031
Growth rate1F = 0·9; = 0·42F = 1·6; = 0·23   
2F = 4·0; = 0·037F = 4·0; = 0·0600·0460·700·016
Biomass1F = 2·5; = 0·10n/a0·0400·460·16
 2F = 0·8; = 0·27n/a   
Predator
Length1F = 10·3; < 0·001F = 0·1; = 0·800·028<0·0010·053
 2F = 3·5; = 0·052F = 0·8; = 0·400·730·0660·025
Weight1F = 0·57; = 0·58F = 1·5; = 0·23   
 2F = 1·7; = 0·21F = 3·1; = 0·096   
Condition1F = 4·0; = 0·036F = 4·1; = 0·0580·0170·190·76
 2F = 2·1; = 0·16F = 4·1; = 0·056   
image

Figure 1.  Rainbow trout fry survival (a), growth (b) and biomass (c) at the beginning of the study (September) and after 1 (October) and 2 months (November) of exposure to wild (W) and transgenic coho salmon with restricted growth (Tr) or fed satiation ration (Ts). Based on raw data. Error bars denote ±1 SEM.

Download figure to PowerPoint

Prey were initially of the same size (Table 1) and predators did not cause growth rates of fry to differ during the first period (Table 2, Fig. 1b). However, during the second period, prey in the presence of Tr predators grew faster than in the presence of both W and Ts predators, with no difference between the latter two groups (Table 2). Larger prey at the onset of the second period also tended to grow slower than small prey during that period (Table 2). These effects resulted in prey in the presence of Tr being larger (F2,20 = 4.6, = 0.023) after the 2 months compared with prey in the presence of W (= 0.024) and Ts (= 0.013). Overall growth of prey was too large to be explained only by selective predation on the smallest prey as average prey weight increased from 1.42 ± 0.02 (SE) to 2.74 ± 0.11 g (Fig. 1b). Hence, a final average weight much larger than the largest fry at the beginning of the study shows that prey were capable of feeding and growing under the experimental conditions.

Biomass in the presence of each predator type was initially the same (Fig. 1c). After the first period, however, Tr exerted stronger predation than could be compensated for by surviving prey growth so that biomass decreased (Fig. 1c). An increase in biomass in the presence of W predators resulted in these two predators types having significantly different effects on prey biomass (Table 2). Biomass in the presence of Ts predators also increased, but it did not differ significantly from either of the other two predator types. The trend observed for biomass during the first period continued during the second period, although these changes were no longer significantly different (Table 2) due to increased variation. Overall, predation effects resulted in a net increase in prey biomass of 10·3 ± 3·7 g in the presence of W and 3·3 ± 5·5 g in the presence of Ts predators, but a net decrease of 5·5 ± 4·2 g of prey biomass in the presence of Tr.

More fry were visible in the presence of W predators compared with either of the two transgenic types throughout the experiment (Fig. 2). Based on the last 10 days of the second period, a higher proportion of available prey (based on the number of surviving prey at termination) was also observed in the presence of W predators (43 ± 10%; F2,20 = 3·6, = 0·045) compared with both Tr (17 ± 7%; = 0·031) and Ts (16 ± 6%; = 0·027).

image

Figure 2.  Average (±SE) number of prey (rainbow trout) observed per time period in the presence of predatory wild (W) and transgenic coho salmon with restricted growth (Tr) or fed to satiation (Ts). Symbols displaced horizontally for clarity only.

Download figure to PowerPoint

Predator growth

At the start of the experiment, the three types of predators had the same mean length and weight and Ts had higher CF compared with that of W and the restricted-fed (Tr) predators (Table 1).

During the first period, all three predator types differed in length with Ts growing the fastest, followed by Tr and W being the slowest growing type (Table 2, Fig. 3a). During the second period, Ts continued to outgrow the other two types, but W increased in growth and matched the growth of Tr. Over the two periods, Ts grew the most (13·6 ± 1·5 mm) followed by Tr (7·8 ± 1·6 mm) and W (5·6 ± 1·2 mm). Hence, both transgenic types maintained a length growth advantage over the wild type even in the semi-natural environment.

image

Figure 3.  Body length, weight and condition of wild (W) and transgenic coho salmon with restricted growth (Tr) or fed satiation ration (Ts) at the beginning of the study (September) and after 1 (October) and 2 months (November) in simulated natural environments. Based on raw data. Error bars denote ±1 SEM.

Download figure to PowerPoint

All three types of predators lost weight at a similar rate during the first period. During the second period, Ts continued to lose weight, whereas Tr and W started gaining weight (Fig. 3b), although none of these effects were significant. Nevertheless, overall weight losses of 1·7 ± 2·6 g for W, 2·5 ± 2·5 g for Tr and 12·7 ± 3·5 g for Ts resulted in weights being more similar at the end of the experiment than at the start (F2,20 = 0·4, = 0·68; Fig. 3b). It is also important to note that predators probably fed on natural invertebrate prey entering the habitats with the creek water because one W predator gained 3·6 g in weight despite not preying upon any of the fry.

The general increase in length and decrease in weight resulted in reductions in CF for all three types of predators (Fig. 3c). Tr lost relatively more in condition than W during the first period, but otherwise there were no significant differences (Table 2). During both periods, high CF at the start of the period tended to be associated with a greater loss in condition during that period. These changes resulted in CFs converging among the three types at the end of the experiment (F2,20 = 1.42, = 0.27).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The main objective of this study was to evaluate whether differences in predation effects between hatchery-reared wild and transgenic predators persist over time after release to nature or whether differences diminish as predators become acclimatized to the natural environment. Whereas our results show that predators become more similar in body weight and CF over time, we found little evidence of convergence in feeding behaviour between wild-type and restricted-fed transgenic predators reared in the hatchery. Environmental conditions during early life can lead to different behaviours several years later, despite experiencing the same environment in the interim. In Atlantic salmon Salmo salar (L.), different rearing conditions during the first half of their lives resulted in behavioural differences during breeding as adults (Fleming et al. 1997). Hence, differential phenotypes developed early in life in the hatchery may produce permanent effects on behaviour that are not reversed by the naturalization process. This may include differential responses to decreasing photoperiod which reduces feeding motivation in wild-type but not so in hatchery-reared restricted-fed transgenic fish (Lõhmus et al. 2008). However, because this latter study involved fish reared in the hatchery and studied under relatively simple environmental conditions, it cannot be ruled out that under more complex experimental conditions, transgenic fish would respond more similar to wild-type fish across seasons.

Whereas behavioural effects mediated by hormones, metabolites and cell physiology may be reversible, phenotypic changes resulting in permanent differences (e.g. brain structure) may not be so easily changed (Kihslinger & Nevitt 2006; Kihslinger et al. 2006). Although wild and transgenic fish were reared in identical physical structures, the sensory inputs are unlikely to be the same due to differences in the behaviour of the surrounding fish, originating from effects of the transgene and the fact that the two genotypes probably experience similar environments in different ways (Sundström et al. 2007b). Further, a number of structural and physiological differences between wild and the transgenic coho salmon arising in the hatchery environment have been reported. These differences include body shape (Devlin et al. 1995; Ostenfeld et al. 1998), muscle structure (Hill et al. 2000), respiratory system (Stevens & Devlin 2000), swimming capacity (Farrell et al. 1997), digestive system (Stevens & Devlin 2005) and pituitary gland size (Mori & Devlin 1999). Some of these alterations are probably due to the rapid growth per se and other changes may be due to the transgene itself (Stevens & Devlin 2005) causing a difference in phenotype between the two types of transgenic predators and the wild type that may be permanent and last throughout their lives after entry into a natural environment. It is also possible that rapid growth in satiation-fed transgenic predators has phenotypic effects not experienced by the slower growing restricted-fed predators. Such effects could alter their respective consequences in nature and indicate that previous rearing experience in culture or in nature would be likely to be critical determinants of the ecological consequences of specific populations of transgenic fish.

Satiation-fed transgenic fish did not differ from wild type in their predation rate as was found previously (Sundström et al. 2007b). However, it is important to note that in the present study fish were size matched, whereas in the previous study satiation-fed transgenic fish were age matched and thus were almost three times as long as the wild type. Many predatory-related traits depend on body size. One that has been investigated in transgenic fish is swimming speed, where it was found that age is a better predictor of swimming capacity than size (Farrell et al. 1997). Such an effect could be due to the costs associated with rapid growth in the satiation-fed transgenic fish that resulted in a higher CF that makes them less streamlined and inferior swimmers (Billerbeck et al. 2001). Further, reduced swimming capacity of satiation-fed transgenic fish may also explain the increased hiding behaviour of prey in their presence, despite not having the same predation effects as restricted-fed transgenic predators. If satiation-fed predators were as motivated to forage as restricted-fed, particularly during the second period when their condition had decreased, both types may induce the same avoidance behaviour in prey, but if swimming capacity is reduced in the former their actual capture rate may be lower.

Predation on early stages of salmonid fry may have limited effects on later population size as a majority of fry die young due to density-dependent mortality (Chapman 1962; Einum & Fleming 2000; Milner et al. 2003; Einum & Nislow 2005). However, at later life-history stages (from weeks up to months), growth becomes a density-dependant factor (Jenkins et al. 1999; Bohlin et al. 2002). Removal of prey by predation reduces competition and allows survivors to grow faster. This effect is apparent during the second period of the present experiment when prey in the presence of restricted-fed transgenic predators grew faster than prey in the presence of the other two predator types. Although this increase in prey growth partly compensated for the higher mortality of prey, resulting in similar changes in prey biomass among all three predator types during the second period (Fig. 1c), it may be possible that an inability to detect a statistical difference depends more on an increase in variation than a lack of different effects.

Our experimental set-up mimicked a part of a stream in which no emigration or immigration was possible for both prey and predators. Restricted-fed transgenic predators that consumed more prey during the first period compared with the other two types consequently had relatively fewer prey to find and consume during the second period. At some point, this may induce dispersal on the part of the predator in an attempt to find better feeding grounds. Greater tendency to disperse has been documented in this transgenic fish strain at a younger age (Sundström et al. 2007a). Alternatively, as prey populations diminish significantly, prey could also begin to immigrate from surrounding habitats (Chapman 1962; Bujold et al. 2004). Hence, because of these dynamic features associated with the mobility of predator and prey populations, sustained effects of escaped hatchery-reared transgenic fish at the population level will be difficult to predict using closed semi-natural experimental systems.

In our study species, the coho salmon, significant genotype–environment interactions can limit consideration of the present data to that of hatchery-reared individuals only and impacts in nature would depend mainly on the phenotype of the escaped hatchery fish and the frequency and magnitude of escape. Any subsequent generation of transgenic fish resulting from successful breeding of an escaped transgenic fish, or very young transgenic fish escaping, would require separate evaluation of fish reared in nature. Notwithstanding the stability of the phenotype observed in the present study, it is important to examine fitness and consequences of transgenic animals under the range of environments potentially encountered in nature. Policy development on the use and management of transgenic organisms must therefore incorporate not only the complex interaction between genotype and environment on phenotypic development but also the differential response of prey populations within the environment (such as changes to the growth rates of the prey depending on intensity of predation). Our results will be important to consider in assessing the potential harm of captive-reared transgenic organisms and for science policy makers and regulators in deciding whether to allow commercial application of transgenic species in aquaculture.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Larry Kahl for providing wild parental fish from the Chehalis River and Morgan Williams for assisting with sampling. We appreciate the constructive comments made by two anonymous referees. The work was carried out with financial support from the Canadian Regulatory System for Biotechnology (RHD) and the Adlerbert Research Foundation (LFS). LFS was funded by a post-doctoral grant from the Swedish Research Council FORMAS and as a Marie Curie Outgoing International Fellowship under contract MOIF-CT-2005-8141 from the European Community’s Sixth Framework Programme. The present work does not necessarily reflect the Community’s views and in no way anticipates its future policy in this area.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Bessey, C., Devlin, R.H., Liley, N.R. & Biagi, C.A. (2004) Reproductive performance of growth-enhanced transgenic coho salmon (Oncorhynchus kisutch). Transactions of the American Fisheries Society, 133, 12051220.
  • Billerbeck, J.M., Lankford, T.E. & Conover, D.O. (2001) Evolution of intrinsic growth and energy acquisition rates. I. Trade-offs with swimming performance in Menidia menidia. Evolution, 55, 18631872.
  • Bohlin, T., Sundström, L.F., Johnsson, J.I., Höjesjö, J. & Pettersson, J. (2002) Density-dependent growth in brown trout: effects of introducing wild and hatchery fish. Journal of Animal Ecology, 71, 683692.
  • Bujold, V., Cunjak, R.A., Dietrich, J.P. & Courtemanche, D.A. (2004) Drifters versus residents: assessing size and age differences in Atlantic salmon (Salmo salar) fry. Canadian Journal of Fisheries and Aquatic Sciences, 61, 273282.
  • Chapman, D.W. (1962) Aggressive behavior in juvenile coho salmon as a cause of emigration. Journal of the Fisheries Research Board of Canada, 19, 10471080.
  • Devlin, R.H. & Donaldson, E.M. (1992) Containment of genetically altered fish with emphasis on salmonids. Transgenic Fish (eds C.L.Hew & G.L.Fletcher), pp. 229265. World Scientific Publishers, Singapore.
  • Devlin, R.H., Yesaki, T.Y., Biagi, C.A., Donaldson, E.M., Swanson, P. & Chan, W.K. (1994) Extraordinary salmon growth. Nature, 371, 209210.
  • Devlin, R.H., Yesaki, T.Y., Donaldson, E.M. & Hew, C.L. (1995) Transmission and phenotypic effects of an antifreeze GH gene construct in coho salmon (Oncorhynchus kisutch). Aquaculture, 137, 161169.
  • Devlin, R.H., Biagi, C.A. & Yesaki, T.Y. (2004) Growth, viability and genetic characteristics of GH transgenic coho salmon strains. Aquaculture, 236, 607632.
  • Devlin, R.H., Sundström, L.F. & Muir, W.M. (2006) Interface of biotechnology and ecology for environmental risk assessments of transgenic fish. Trends in Biotechnology, 24, 8997.
  • Dunham, R.A., Chitmanat, C., Nichols, A., Argue, B., Powers, D.A. & Chen, T.T. (1999) Predator avoidance of transgenic Channel catfish containing salmonid growth hormone genes. Marine Biotechnology, 1, 545551.
  • Einum, S. & Fleming, I.A. (2000) Selection against late emergence and small offspring in Atlantic salmon (Salmo salar). Evolution, 54, 628639.
  • Einum, S. & Nislow, K.H. (2005) Local-scale density-dependent survival of mobile organisms in continuous habitats: an experimental test using Atlantic salmon. Oecologia, 143, 203210.
  • Farrell, A.P., Bennett, W. & Devlin, R.H. (1997) Growth-enhanced transgenic salmon can be inferior swimmers. Canadian Journal of Zoology, 75, 335337.
  • Fleming, I.A., Lamberg, A. & Jonsson, B. (1997) Effects of early experience on the reproductive performance of Atlantic salmon. Behavioral Ecology, 8, 470480.
  • Hill, J.A., Kiessling, A. & Devlin, R.H. (2000) Coho salmon (Oncorhynchus kisutch) transgenic for a growth hormone gene construct exhibit increased rates of muscle hyperplasia and detectable levels of differential gene expression. Canadian Journal of Fisheries and Aquatic Sciences, 57, 939950.
  • Jenkins, T.M., Diehl, S., Kratz, K.W. & Cooper, S.D. (1999) Effects of population density on individual growth of brown trout in streams. Ecology, 80, 941956.
  • Jhingan, E., Devlin, R.H. & Iwama, G.K. (2003) Disease resistance, stress response and effects of triploidy in growth hormone transgenic coho salmon. Journal of Fish Biology, 63, 806823.
  • Jonsson, S., Brännäs, E. & Lundqvist, H. (1999) Stocking of brown trout, Salmo trutta L.: effects of acclimatization. Fisheries Management and Ecology, 6, 459473.
  • Kapuscinski, A.R. & Hallerman, E.M. (1991) Implications of introduction of transgenic fish into natural ecosystems. Canadian Journal of Fisheries and Aquatic Sciences, 48, 99107.
  • Kihslinger, R.L. & Nevitt, G.A. (2006) Early rearing environment impacts cerebellar growth in juvenile salmon. Journal of Experimental Biology, 209, 504509.
  • Kihslinger, R.L., Lema, S.C. & Nevitt, G.A. (2006) Environmental rearing conditions produce forebrain differences in wild Chinook salmon Oncorhynchus tshawytscha. Comparative Biochemistry and Physiology – Part A: Molecular & Integrative Physiology, 145, 145151.
  • Knibb, W. (1997) Risk from genetically engineered and modified marine fish. Transgenic Research, 6, 5967.
  • Lõhmus, M., Raven, P.A., Sundström, L.F. & Devlin, R.H. (2008) Disruption of seasonality in growth hormone-transgenic coho salmon (Oncorhynchus kisutch) and the role of cholecystokinin in seasonal feeding behavior. Hormones and Behavior, 54, 506513.
  • Mackay, T.F.C. & Anholt, R.R.H. (2007) Ain’t misbehavin’? Genotype–environment interactions and the genetics of behavior. Trends in Genetics, 23, 311314.
  • Milner, N.J., Elliott, J.M., Armstrong, J.D., Gardiner, R., Welton, J.S. & Ladle, M. (2003) The natural control of salmon and trout populations in streams. Fisheries Research, 62, 111125.
  • Miner, B.G., Sultan, S.E., Morgan, S.G., Padilla, D.K. & Relyea, R.A. (2005) Ecological consequences of phenotypic plasticity. Trends in Ecology & Evolution, 20, 685692.
  • Mori, T. & Devlin, R.H. (1999) Transgene and host growth hormone gene expression in pituitary and nonpituitary tissues of normal and growth hormone transgenic salmon. Molecular and Cellular Endocrinology, 149, 129139.
  • Muir, W.M. & Howard, R.D. (1999) Possible ecological risks of transgenic organism release when transgenes affect mating success: sexual selection and the trojan gene hypothesis. Proceedings of the National Academy of Sciences of the United States of America, 96, 1385313856.
  • Norušis, M.J. (2007) spss 15.0 Advanced Statistical Procedures Companion. Prentice Hall, Englewood Cliffs, NJ.
  • Olla, B.L., Davis, M.W. & Ryer, C.H. (1998) Understanding how the hatchery environment represses or promotes the development of behavioral survival skills. Bulletin of Marine Science, 62, 531550.
  • Ostenfeld, T.H., Mclean, E. & Devlin, R.H. (1998) Transgenesis changes body and head shape in Pacific salmon. Journal of Fish Biology, 52, 850854.
  • Snow, A.A., Andow, D.A., Gepts, P., Hallerman, E.M., Power, A., Tiedje, J.M. & Wolfenbargerh, L.L. (2005) Genetically engineered organisms and the environment: current status and recommendations. Ecological Applications, 15, 377404.
  • Stevens, E.D. & Devlin, R.H. (2000) Gill morphometry in growth hormone transgenic Pacific coho salmon, Onchorhynchus kisutch, differs markedly from that in GH transgenic Atlantic salmon. Environmental Biology of Fishes, 58, 113117.
  • Stevens, E.D. & Devlin, R.H. (2005) Gut size in GH-transgenic coho salmon is enhanced by both the GH-transgene and increased food intake. Journal of Fish Biology, 66, 16331648.
  • Sundström, L.F. & Johnsson, J.I. (2001) Experience and social environment influence the ability of young brown trout to forage on live novel prey. Animal Behaviour, 61, 249255.
  • Sundström, L.F., Devlin, R.H., Johnsson, J.I. & Biagi, C.A. (2003) Vertical position reflects increased feeding motivation in growth hormone transgenic coho salmon (Oncorhynchus kisutch). Ethology, 109, 701712.
  • Sundström, L.F., Lõhmus, M., Devlin, R.H., Johnsson, J.I., Biagi, C.A. & Bohlin, T. (2004) Feeding on profitable and unprofitable prey: Comparing behaviour of growth-enhanced transgenic and normal coho salmon (Oncorhynchus kisutch). Ethology, 110, 381396.
  • Sundström, L.F., Lõhmus, M. & Devlin, R.H. (2005) Selection on increased intrinsic growth rates in coho salmon Oncorhynchus kisutch. Evolution, 59, 15601569.
  • Sundström, L.F., Lõhmus, M., Johnsson, J.I. & Devlin, R.H. (2007a) Dispersal potential is affected by growth-hormone transgenesis in coho salmon (Oncorhynchus kisutch). Ethology, 113, 403410.
  • Sundström, L.F., Lõhmus, M., Tymchuk, W.E. & Devlin, R.H. (2007b) Gene-environment interactions influence ecological consequences of transgenic animals. Proceedings of the National Academy of Sciences of the United States of America, 104, 38893894.
  • Tiedje, J.M., Colwell, R.K., Grossman, Y.L., Hodson, R.E., Lenski, R.E., Mack, R.N. & Regal, P.J. (1989) The planned introduction of genetically engineered organisms: ecological considerations and recommendations. Ecology, 70, 298315.
  • Tymchuk, W.E.V., Abrahams, M.V. & Devlin, R.H. (2005) Competitive ability and mortality of growth-enhanced transgenic coho salmon fry and presmolts when foraging for food. Transactions of the American Fisheries Society, 134, 381389.
  • Warburton, K. (2006) Learning of foraging skills by fishes. Fish Cognition and Behavior (eds C.Brown, K.Laland & J.Krause), pp. 927. Blackwell Publishing, Oxford.