ENVIRONMENTAL RISKS IN USING GH TRANSGENIC ATLANTIC SALMON AND RAINBOW TROUT FOR COMMERCIAL MARINE PRODUCTION IN CANADA

Arnold Sutterlin1, Garth Fletcher2, Choy Hew3, and Tillmann Benfey4

1Aqua Bounty Farms, Fortune RR#4, Souris, P.E.I. canada, C0A 2B0; 2A/F Protein Canada and Ocean Sciences Centre, Memorial University, St. John's, NF Canada A1C 5S7; 3Hospital for Sick Children, Dept. Clinical Biochemistry and Chemistry Univ. Toronto, Ont., Canada M5G 1L5; and 4Dept. Biology, Univ. New Brunswick, Fredericton, N.B.Canada E3B 6E1

SUMMARY

The general progress and present status toward commercialization of salmon and trout transgenic fish for an "all fish" chimeric growth hormone gene construct are reviewed. We anticipate no risks to human health as the entire gene construct is derived from edible species of fish; however, there remain a number of environmental concerns that need to be addressed before the technology can be applied by the aquaculture industry. Under pending Canadian regulations for the use of transgenic fish in enclosed research facilities or hatcheries, a case is made that compliance with proposed containment measures as shown in the design of the facility should result in little or no environmental risks to natural salmonid populations. Attempts are then made to assess the effectiveness of current fish containment and sterilization techniques in order to reveal factors of risk involved in commercially culturing these fish in land-based systems and in ocean pens. Many of the risk issues confronting the application of transgenic technologies are not particulary new and have been encountered in the past by fisheries managers and aquaculturists in relation to salmonid transfers, introductions and selective breeding. Some contrived examples of impacts (real and perceived) are developed for non-endemic species such as steelhead trout cultured in Newfoundland. Factors considered include probabilities of escape, adult survival, migratory behaviour, spawning success, sex ratios, reproductive sterility (triploidy effectiveness) and offspring survival, inter- and intra specific competition and gene fitness. It is concluded that the magnitude of risk will depend on the nature of the gene construct and its phenotypic effect, and will vary between different salmonid species and ecological regions. Present sterility techniques will probably be adequate for some species in most circumstances, but may not sufficiently reduce risks (or be commercially viable) for other species under other conditions. Considering the lack of present understanding of the fitness (behavioural, physiological and genetic) of such transgenic fish, it may be exceedingly difficult to predict impacts in many situations. If sterile transgenic fish were to be used in aquaculture, risk to native salmonid populations would likely be no greater (and probably less) than under present operating conditions in which non sterile fish are being used world wide.

Key words: Transgenic, fish, Atlantic salmon, rainbow trout, growth hormone, environment, risk, containment, triploidy

INTRODUCTION AND BACKGROUND

Workers at Memorial University began injecting salmon eggs in 1983 with an antifreeze protein gene construct derived from winter flounder in an attempt to provide protection from ice formation in the tissues of salmon being cultured in cold climates particularly in Eastern Canada. Although the gene has been incorporated in subsequent generations in a heritable fashion, the quantities of the protein produced in salmon to date have not been insufficient to provide a safe margin of thermal protection (Fletcher et al., 1988). In 1988 injections were initiated in Atlantic salmon using a transgene consisting of a fish (eel pout) antifreeze protein gene promoter fused to the growth hormone cDNA from chinook salmon. Gene injections through the micropyle of fertilized, but non-water-activated salmon eggs, have routinely led to about 2 to 3% incorporation of the gene in developing juveniles. These fish consisit of a mixture of somatic and germ line mosaics (Fletcher and Davis, 1991). Of these gene carriers, about 25% of the fish actually pass the gene (and the rapid growth characteristic) to a of the F1 generation, resulting in anywhere from 1 to 20% transgenic progeny (Hew et al., 1995). A stable inheritance pattern at the F2 generation has been demonstrated and explained by from 1 to 4 chromosome inserts, with transgenic fish growing up to ten times times faster than their non-transgenic siblings. Male heterozygous brood fish presently on hand are capable of producing 50% fast growing transgenic salmon when mated to wild female salmon and indications to date suggest that size grading at a weight of less than 5 grams is an effective method of separating transgenic fish. In some of the transgenic lines of fish, the transgene, although incorporated (as detected by PCR), is not expressed in the form of enhanced growth. Problems exist in certain transgenic lines with skeletal abnormalities and with higher than normal early mortalities rates. Techniques (reverse PCR) are available to identify the exact number of chromosome inserts possessed by a fish and which fish have homozygous inserts; work is planned to apply these methods to enhance the breeding program. Preliminary results suggest that gross food conversion efficiency of transgenic salmon are likely to be as good and or better than that of normal salmon. A/F Protein Canada has licenced the patented technology to companies in Scotland and New Zealand and is actively working with these companies to assist them in developing transgenic fish using their existing brood stock. Injections of the same construct into rainbow trout eggs were initiated this Spring at our new research facility at Prince Edward Island, and we anticipate that progress will be more rapid with this species as the triploid technology is perfected and readily applicable for the production of reproductively sterile trout. Injections are also planned for Arctic char this fall.

POTENTIAL NEGATIVE IMPACTS OF TRANSGENIC FISH

ON WILD POPULATIONS

Many of the concerns (Kapuscinski and Hollermann, 1990 and 1991) associated with the use of transgenic fish in fish culture are not particularly new and have been encountered in the past in one form or another by the aquaculture industry and fisheries managers (see NASCO CN(93)29, 1993). These relate to the use and possible impact of the escape of exotic (non-endemic) or highly selected species, strains or stocks, fish hybrids and fish with manipulated chromosome sets. If the new or genetically modified fish is reproductively sterile, the fear of genetic introgression, genetic drift or swamping of native gene pools is lessened (Devlin and Donaldson, 1992). Even then, however, a considerable degree of apprehension among fisheries managers and the public is likely to remain even with sterile transgenic fish and must be addressed.

As discussed later, unless the aquaculture operation is entirely land-based with rigid containment methods in place, there is always the possibility of sterile transgenic fish escaping into the wild. This is particularly the case for culture operations using lake or ocean cages where the appearance of aquaculture fish returning to local rivers has been well documented (Ludwig, 1996; Gausen and Moen, 1991; Skaala et al., 1991). Present production of Atlantic salmon is almost inclusively conducted in sea pens and unless the performance of transgenic salmon becomes truly remarkable, it is doubtful that a sizeable portion of the industry would switch to more contained land-based, pumped systems in light the high capital and operating costs of the later. However, sufficient containment of transgenic trout in land-based farms using tanks or raceways would appear to be possible Even in situations where industry has for environmental reasons adopted the use of non-transgenic, all-female, triploid fish (99% sterile), there is still concern over the direct effect of escaped, long-living, free-ranging predators on natural fish populations as well as their potentially competitive interactions for resources such as food and space (Sutterlin and Collier, 1991). However, the use of sterile transgenic fish does allow for a period of assessment with the possibility of reversing negative consequences through discontinuing use of the novel genotype (Hallermann and Kapuscinski, 1992). Although it is the expressed intension of A/F Protein to maintain all experimental non-sterile transgenic broodfish and juveniles in secure confinement while refining techniques for reliably inducing sterility prior to any application for field trials, the two most disconcerting questions that remain (and are discussed later) are:

What is likely to be an accepted level of sterility for transgenic salmonids for use in commercial land-based and cage rearing operations?

How can the issue of direct competition of sterile transgenic fish on the ecology of natural systems be addressed?

CONTAINMENT OF TRANSGENIC FISH FOR RESEARCH PURPOSES

A/F Protein has recently purchased a salmon hatchery on, Prince Edward Island, specifically for the purpose of conducting research and developing transgenic salmonid brood stock (Figure 1). The following set of selection criteria set out in the initial search for a facility have been satisfied:

The plant should be based on a sizeable quantity of ground water of high water hardness.

The plant should have sufficient water and tank size and space to rear salmon and trout to maturity.

The facility should already be disease certified or be capable of FHPR Certification.

All rearing space should be confined within a secured building

All water should leave the plant through a common drain

Effluent water should not discharge into a major system containing salmonid fish

The building should be structurally sound and of sufficient elevation to preclude flooding

Security systems must include automatic telephone call out and remote paging

Living accommodations should be available in the building so as provide 24 hr security

Plant modifications presently underway to meet containment requirements consist of :

The installation of a hall way to allow entry to the upper level office /living complex while preventing access to the rearing area

The replacement of present door locks with heavy duty units

The installation (figure 2) of an in-ground concrete sump equipped with two sets of 1mm mesh screening to filter all effluent and automatic methods to detect screen fouling

The recruitment of an antisocial, vocal canine with an attitude problem

The size of the screening has been set to retain the smallest life history stage of the organisms in use. For salmonids this involves preventing the passage of first feeding fry with the lower size limit set for that of Arctic char which are smaller than the fry of rainbow trout and much smaller than that of salmon. At least three levels of screening are provided for each rearing unit. Incubators are equip. with 1 mm screens that filter the water from each stack of incubators containing transgenic eggs. Water discharge from the fry tanks pass through a rotating drum filter (200 micron screen) which is necessary to remove sediments prior to heat effluent heat recovery using heat exchanger and heat pump. The back wash from this screen enters a domestic septic tank and blind weeping field. All other effluent leaving the hatchery must pass through a double (1 mm) screened concrete sump buried in the ground and accessible through a locked steel hatch. Inexpensive aluminum window screening works fine but needs to be supported by 1/4 inch galvanized hardware cloth. It is important that the water be allowed to fall a distance of at least 12 inches before striking the screen or the screen will quickly foul. Test with submerged screens proved unsatisfactory as then fouled in several hours.

Should the upper most screen become plugged and begin to overflow onto the lower screen, a float switch activates a telephone monitoring system that begins calling hatchery personnel at their homes as well as activating their pagers indicating a screen overflow code. The lower screen provides the filtration until personnel can replace or clean the upper screen.

Presently in Canada, regulations concerning contained research, field trials and commercial production of transgenic fish are evolving under the Federal Department of Fisheries and Oceans in the form of a draft document entitled: Transgenic Aquatic Organisms: Policy and Guidelines for Research with, or for Rearing in Natural Aquatic Ecosystems in Canada. The policy is presently in its 3rd or 4th draft and is still under government and industry review. A/F Protein will likely be the first private company in Canada to apply for a research permit under the new regulations and in consultation with DFO in Ottawa has determined that voluntary compliance with the draft regulations could provide some valuable practical experience for both government and our company prior to any required licencing or legislative implementation of this policy. A/F Protein applied in early April , 1996 for the prototype research permit and has been working with DFO officials locally and in Ottawa to install the mutually acceptable containment measures described above. We anticipate a final inspection of the facility in August, 1996.

REPRODUCTIVE STERILITY OF TRANSGENIC FISH

Presently available and potential future methods of sex reversing and sterilizing salmonid fish have been described and/or reviewed by Devlin and Donaldson ( 1992), Hunter and Donaldson (1983) and Pepper (1991).

The most promising of presently available techniques include the following:

Direct oral androgen treatment of juvenile fish producing a high frequency of sterility in both sexes. However there is some public concern (perhaps of questionable merit) over treating even juvenile food fish with steroids.

Hormonally sex-reverting female fish to male fish and using the resulting milt containing only X chromosomes to create an all-female population of transgenic fish. This approach would only be acceptable in highly specific situations where wild males of the same species were not available to mate with escaped transgenic females.

Inducing triploidy in the above all-female population by the use of heat or pressure shock shortly after fertilization with the sperm bearing only X chromosomes. Triploidization of mixed-sex population of eggs is not recommended as male triploid salmonids do sexually mature and could possibly contribute to spawning activities resulting in lethal aneuploid offspring.

Although there are a number of alternate pathways to achieve the later objective, the final outcome results in a population that is entirely female providing the sexual genotype of milt donor is properly identified and providing the lots of fish or eggs do not become mixed up with diploid groups at the transgenic production facility. Maintaining juvenile diploid brood stock or other diploid production lines in separate rooms or buildings would be a recommended approach. In rainbow trout and Atlantic salmon, identification of first generation sex-reversed females from normal males can quite easily be accomplished by examination of the gross testis and sperm duct morphology, and a molecular sex probe, although perhaps desirable, is not required. In any event, a simple gonad examination of a small sample of juvenile progeny will confirm the mono sex status, and in subsequent sex-reversed brood stock generations this problem becomes inconsequential.

There also are also a few uncertainties relating to the interactions of the triploid and transgenic condition that might influence the final breeding methods (maternal or paternal transgene contribution) to be adopted depending on the relative commercial suitability of heterozygous or homozygous transgenic fish. Because there is a doubling of maternal chromosomes involve in the triploidization process, disruption in transgene control is a possibility although this does not seem to be the case with other genes in triploid fish (Benfey, 1997).

THE NUMBERS GAME

Quite apart from the concern over direct effects of escaped sterile transgenic fish described later, the acceptable frequency of non sterile transgenics fish in a production lot destined for a commercial hatchery and eventually to a sea pen grow-out operation, might logically also depend on the following factors:

The past record (frequency and magnitude) and probability for future escapes of fish at these sites

The survival and reproductive effectiveness of the transgenic fish

The presence, size and importance (environmentally, commercial and recreational) of local native salmonids of the same or hybridizable species

The nature, stability and resulting phenotypic fitness due to the transgene in the initial and subsequent generations

In an attempt to avert any genetic consequences, perhaps one of the more uncertain obstacles for the producer of transgenic fish will relate to the efficacy of the triploid induction process which will determine the frequency of non-sterile fish in the production lot and as such could become an environmental control issue. The effectiveness of heat and pressure treatments, summarized in Table 1, suggest that with experienced operators it should be possible to produce at least 99% sterile fish even if the treatment results in a financially tolerable 20% loss of eggs. There also appears to be a consensus developing by scientists and industry that pressure is a more reliable method than temperature for triploid induction. The cost of a 4-liter pressure vessel ($C 20,000) and time required to treat salmon eggs at a 20,000 eggs per batch (30 minutes) does not appear to be commercially insurmountable.

For example, it may be feasible to examine the blood cells (Coulter sizing or determining their DNA content by flow cytometry) from 100 or possibly even 1000 fry resulting from each pressure treated group of eggs which would detect the presence of a single diploid fish in a sample of X number of fish at the Y percent confidence level when they are present in the population at the Z % incidence. Batches with unacceptable levels of diploid fry could then be excluded from further commercial production. Table 2 (from Ossiander and Wedemeyer, 1973 ) shows the sampling effort required to detect a single diploid fish at the 95% confidence level when their frequency in the population varies between 0.5 and 10%. However, it should be noted in Table 1 that the size of the sample in these studies are all under 100 fish and this low level of screening would (from Table 2) not likely be adequate to detect a low diploid incidence in large populations. Therefore, the question remains: what level of triploidy will be deemed environmentally acceptable and can sufficiently sensitive and affordable screening methods be devised to meet these requirements?

It becomes apparent from Table 2 that treatment lots based on eggs from each female (5,000-10,000 eggs) will not be a cost efficient method of screening if blood samples from the required number of individual fry must be examined. Flow cytometry, however, is sufficiently sensitive to enable pooled blood samples from fry derived each treated female to be used. Screening treated fry from pooled samples of treated eggs from several females may have some disadvantages in that variations in triploidy effectiveness, possibly due to a parental dam effect, could result in the entire lot of fry being later rejected. Although it would be advantageous to remove any unacceptable lots of eggs from production as soon as possible, rejection of large egg lots even in a normal hatchery situation is costly, and with the added effort of including the transgenic component in the development, it could be prohibitive. In any event, it would appear that further work is required to resolve the best sampling and screening methods once some guidelines develop concerning acceptable limits of sterility.

Because a single large commercial hatchery can produce in excess of one million smolts per year, what confidence level should be prescribed for such a hatchery in order to reduce risks to an acceptable level prior to authorization for the stocking of smolts in sea pens? If we look at other regulations that depend on detection probabilities, such as the Canadian Fish Health Regulations, which deal in matters of similar concern being designed to prevent the transfer of infectious diseases to natural population, we find that the required sampling for certification of a facility to practice in the interprovincial (and often international) transfer of gametes or fish requires a maximum sample size of 150 fish per lot. This sampling level is sufficient to detect a pathogen carrying fish at the 95% confidence level if carriers are prevalent at 2% (Gillespie et al. 1974). Is the risk of negtive impacts from gene transmission any more or less serious than those resulting from disease transmission and if so how much?

FITNESS CONSIDERATIONS

Confounding the controversial issue related to the fitness of the transgenic fish, successful application of the above sterility techniques must also take into account the general suitability of the triploid condition and its impact on commercially important production traits. Few consumer problems are anticipated with chromosome set manipulated fish in that polyploidy has been used for decades in plant agriculture, and triploid trout have been on the market in Europe and North America for several years and do not differ from normal fish in nutritional quality or consumer safety. It would be encouraging to report that the triploid condition offers no disadvantages to the fish farmer, and this appears to be the situation with rainbow trout but not with Atlantic salmon.

The Bay D'Espoir salmon growers on the South Coast of Newfoundland have been successfully rearing all-female triploid rainbow trout in sea pens since 1988 with a production of about 1000 mt of 3 kg trout in 1996 (Sutterlin and Collier, 1991). Because this species is not endemic to insular Newfoundland, it was at government insistence that only sterile trout be used to insure that this species would not become established and compete with native salmonids such as Atlantic salmon and brook trout. Aside from the delay and start up cost to initiate the lengthy sex-reversing process, triploid trout have proven to offer several production advantages over normal trout in that they can be reared to a much larger size and the meat quality and marketing schedule is not compromised by the detrimental effects of sexual maturation. Such non-reproducing trout have also been shown to offer some promise in trials in Cape Breton, N.S. (Stewart, 1996) and in the United Kingdom (Johnstone, 1996). A/F Protein has only recently injected trout eggs with the GH construct at its new hatchery in P.E.I., and it is encouraging to know that the sterilization technology is now available for this species.

The situation with producing non-reproductive Atlantic salmon does not appear to be as straightforward despite attempts by government agencies to encourage the industry in Canada, Australia and Europe to begin producing such environmentally friendly salmon. The problems apparently do not reside with either the sex-reversal or triploid induction methods (which are nearly identical to those for rainbow trout), but relate to the general performance and fitness of the triploid condition. Although all-female, triploid Atlantic salmon seem to perform well under optimum conditions they often exhibit a number of undesirable production traits in the commercial situation not observed in their diploid counterparts (Benfey, 1991). These include: reduced survival during egg incubation and during marine grow-out (McGeachy et al, 1996), diminished tolerance of chronic stress (Benfey, 1996), increased incidences of lower jaw abnormalities (Sutterlin et al, 1987; Jungalwalla, 1991; Hughes, 1992; King and Lee, 1993; and Lee and King, 1994 ), greater susceptibility to cataract formation (Wall and Richards, 1992) as well as a number of hematological abnormalities (Benfey, 1997). It remains to be seen if these deficiencies can be resolved through improved nutrition, husbandry or by selecting different genetic stocks. In certain countries with more moderate winter temperatures, a steriodogenic burst in growth during the early stages of sexual maturity are being capitalized upon and this advantage would be lost if fish were sterile. At this time, the prospects of easily getting the transgenes into suitably performing sterile Atlantic salmon are not altogether clear and perhaps methods other than triploidy should also be explored (Devlin and Donaldson, 1992).

Different constructs of the same or different genes can be expected to have a range of phenotypic effects of varying environmental consequence ranging from that of a relatively benign effect of a non-expressed marker gene (perhaps of use in population studies) to genes that alter reproductive success. These factors, particularly the last relating to transgene fitness in an open system, are going to be exceptionally difficult if not impossible to predict reliably, and the possible requirement for lengthy and costly monitoring programs over a considerable migratory range could well exceed the capacity of government or industry, thereby possibly precluding the use of transgenic fish in such extensive culture systems. Theoretical treatments using techniques in population genetics to predict rates of genetic introgression similar to that by Hutchings (1991) and Muir et al. (1996) could provide some valuable insight, however a component of relative fitness is required for such an analysis. In the case of a salmon containing a GH transgene, it would be extremely useful to know something about the physiological energetics of these fish including some estimate of metabolic rates, gross and net food conversion efficiencies, maintenance ration levels and the influence of temperature on these parameters. It is the intention of A/F Protein to undertake such studies which are also extremely important from the commercial perspective. It might develop that the energetic costs of foraging to satisfy appetite and growth potential cannot readily be met resulting in a neutral or negative competitive advantage. For example, Dunham et al.(1995) found that GH transgenic catfish, which exhibit a rather modest (33 %) increase in growth rates in aquarium studies, showed no growth advantage over normal fish when reared in ponds on natural foods without supplemental feeding, and that the enhanced growth rates of transgenic fish are predominately confined (as with salmon) to the juvenile stages and as such do not contribute inordinately to reproductive success. These authors, in their review of the literature, conclude that it is exceptionally difficult to create a new genotype in established natural populations through the infusion of domestic or even wild genes.

There may be a limit to usefulness of such population and genetics models in that the track record in Fisheries Resource Management in Canada in recent years, despite efforts by some very capable fisheries scientists, has resulted in a considerable loss of public confidence in the ability to manage such complex systems; to expect fish biotechnologist to make predictions likely to be more reliable maybe asking a lot!

SOME FISH TALES

The general welfare of Atlantic salmon, often referred to as the "King of the Sportfish," is of considerable concern to a variety of user and special interest groups which include sport/recreational fishermen, commercial fisherman, conservation associations, indigenous peoples, fish farmers and the occasional gourmet chef. As a means of revealing the variety of perceived impacts (both negative and positive), the following hypothetical case history studies ("transgenic scenarios") are provided as contrived examples of two extreme levels of impact as might be anticipated by groups having different perspectives.

Let us assume that a cooperative of salmonid farmers from a small coastal community in Newfoundland that own their own hatchery and sea pens wish to put a sizeable component of their production into transgenic trout fish in order to accelerate growth rates and become more competitive. Because of the excellent performance in the past, they plan to continue using use all-female, sterile triploid trout in a location that does not have native rainbow trout, and they feel confident that there will be little or no negative impact on local Atlantic salmon or brook trout populations. The hatchery manager decides to import GH transgenic fry from a certified (disease and transgenic) source after they had been screened to insure that they are 100% all-female and at least 99% triploid. After six months of fresh water culture in the hatchery, one million smolts are stocked into estuarine pens located thorough the Bay. Two years later, the trials are deemed a success as improved growth rates are observed at low winter temperatures in both fresh and salt water, less energy is required for heating water at the hatchery and the production time and production cost is improved considerably. Although there are the usual number of fish escaping, no negative impact is observed by threse reproductively incapable trout on salmonid populations in local rivers. Because of their greater food and energy requirements, it is discovered that the trout do not survive well outside of the protective aquaculture environment and as an added benefit, a very lucrative estuarine summer and winter ice sport fishery develops to exploit these trophy size trout that have escaped which employs many local people including aboriginal groups that serve as fishing guides. The industry expands providing employment for many of the fisherman displaced as a result of the failure of the traditional cod fishery. It would appear that Canadian biotechnology has given these farmers an edge whereby they can better compete with Chilean trout farmers who are blessed with warmer temperatures, better growth rates and less expensive feed.

In the second scenario, things seem to go wrong at the onset. Ice damage is exceptionally bad the first winter and combined with several seal attacks, 5% of the total stock ( 50,000 fish) inadvertently escaping from the net enclosures. This happens close to harvesting time and the fish are averaging 3 kg in weight. All escaping fish survive and all migrate to the nearest salmon river which has environmental temperatures and spawning habitat behaviourally conducive to mating by entire group of trout. Of the 50,000 fish, only 500 (1%) are non-sterile diploid fish and produce eggs, but there is a paucity of male trout around since rainbow trout are not endemic in this area and hybridization with native species is not possible. Nevertheless, a few stray male rainbow trout from farming activities in Cape Breton (400 km distant) or from established populations on the Avalon Peninsula manage to find their way to this river and 1.5 million eggs (50%) are successfully fertilized and deposited. Incubation temperatures are ideal and 10% of the eggs hatch successfully and the resulting fry emerge from the gravel at an opportune time, to encounter excellent food resources resulting in 10% survival to the smolt stage. This results in 15,000 transgenic trout smolts going to sea each containing 50% of the transgenes possessed by their female parent. After two to three years, 1500 surviving transgenic trout, consisting of both males and females, return to the same river and undertake spawning among themselves and with other transgenic trout that have more recently escaped. Because their spawning time overlaps with that of Atlantic salmon and brook trout, the rainbows compete for spawning habitat and also dig up eggs previously deposited by the two native species. Their absolute fecundity is also great because of their trangenically enhance oceanic growth rates, and many eggs are deposited by each female. The juvenile transgenic trout are always hungry and exceptionally aggressive, displacing salmon from the most productive feeding habitat. Many of the large transgenic trout overwinter in the estuary and wait at the mouth of the rivers in the Spring for a free meal of vulnerable salmon smolts migrating to sea. The trout eventually become self-sustaining and spread to many other rivers. After several years of this abuse, the native salmon and brook trout populations begin to decline. Sport fishing revenues cease and the aboriginal food fishery fails. Fisheries officers are instructed to confiscate and destroy all remaining transgenic fish, and the fish farmers suffer severe economic hardship. There is a profusion of accusations, allegations and litigations accompanied by some bad press and a large migration of lawyers to this Newfoundland outport community. The remaining free-ranging transgenic trout become so large and aggressive that the Canadian Coast Guard has to issue a warning to bathers and small boat operators, and an element of public safety is added to the debate.

The above "Jurassic Jaws" scenario is put forth not as an attempt to be humorous, facetious or to trivialize the seriousness of such issues, but to challenge the government, scientific community and the public to sort out which elements are likely to be possible, which are probable and which might be best considered pretentious or fictitious. While the first scenario might be considered especially optimistic in that no technical or environmental problems arose during this field trial, many of the predictions made in the second scenario ignore many biological facts, are alarmist and extremely unlikely.

The required analysis to use transgenic Atlantic salmon in the farming industry in British Columbia (which as with trout above would be also be a non indigenous species) would likely involve assessing elements of risk similar to that above with a determination being made as to the acceptability of using mono-sex fish over truly sterile fish. In eastern Europe, the State on Maine, USA and in Eastern Canada where Atlantic salmon are endemic, the added impact component of genetic introgression of the transgene within the species would have to be included in the risk analysis and the frequency of non sterile transgenic fish could become the central issue. A proposal to rear transgenic trout or Arctic char in a ground water hatchery in central Canada would involve somewhat different considerations. It is therefore, apparent that such exercises will have to be quite gene, species and site specific, and simple generalizations concerning the risks (or benefits) of transgenic fish are not likely to be particularly useful.

Although perhaps a far-fetched justification for the commercial use of transgenic fish, the follow question may be legitimate. If the previously documented neutral or negative effect that the sterile triploid or monosex condition seems to have on commercial Atlantic salmon operations could be offset by a positive transgenic component, would the industry not be more inclined to use sterile fish? And if such were the case, would the risks (real or perceived) to wild stocks be any greater than that projected or perceived under current practices?

ACKNOWLEDGEMENTS

The authors of this paper would like to thank the Canadian funding agencies (The National Sciences and Engineering Research Council, and the National Research Council of Canada) for the many years of support for this research. We particularly acknowledge the assistance of our colleagues at St. Andrews, N.B., Dick Saunders (DFO) and Greg Goeff (HMSC), who have contributed much to this project. Comments on the manuscript by Dr. Dick Alderson were most helpful.

REFERENCES

Benfey, TJ (1996) Use of all-female and triploid salmonids for aquaculture in Canada. Bull. Aquac. Assoc. Can. 96-2: 6-8.

Benfey, TJ (1997) The physiology and behaviour of triploid fish. Rev. Fish. Sci. (In press).

Benfey, TJ and Sutterlin, AM (1984) Triploidy induction by heat shock and hydrostatic pressure in landlocked Atlantic salmon (Salmo salar L.). Aquaculture 36: 359-367.

Chourrout, D (1984) Pressure-induced retention of the second polar body and suppression of the first cleavage in rainbow trout: production of all-triploids, all-tetraploids, and heterozygous and homozygous diploid gynogenetics. Aquaculture 36: 111-126.

Devlin, RH and Donaldson, EM (1992) Containment of genetically altered fish with emphasis on salmonids. In: Transgenic Fish. Ed. C.L. Hew and G.L. Fletcher, Word Scientific Pub. Co., Singapore. Chap. 13: 229-265.

Dunham, RA; Chen, TT; Powers, DA; Nichols, A; Argue, B and Chitmanat,C (1995) Predator avoidance, spawning and foraging abilility of transgenic channel catfish with rainbow trout growth hormone gene. In: Proceedings of Biotech. Risk Assessment Symp. (EPA and USDA), Pensecola Beach, Florida.

Fletcher, GL and Davies, PL (1991) Transgenic fish for aquaculture In: Genetic Engineering ed: JK. Setlow, Plenum Press, N.Y. Vol. 13: 331-370.

Fletcher, GL Shears, MA King, MJ. Davies, PL and Hew, CL (1988) Evidence for antifreeze protein gene transfer in Atlantic salmon (Salmo Salar). Can. J. Fish. Aquat. Sci. 45(2) 352-357.

Gausen, D and Moen, V (1991) Large-scale escapes of farmed Atlantic salmon (Salmo salar) into Norwegian rivers threaten natural populations. Can J. Fish. Aquat. Sci. 48: 426-428.

Gillespie, DC , Evelyn, TPT , Fransi, C , Mackelvie, RM and Neufeld, N (1974) Methods for the detection of certain pathogens of salmonid fishes. Dept. Envir. Fish. and Mar. Serv. Ottawa, Miscellaneous Special Publication No.23.

Gray, AK, Evan, MA and Thorgaard, GH. (1993) Viability and development of diploid and salmonid hybrids. Aquaculture 112: 125-142.

Hallerman, EM and Kapuscinski, A (1992) Ecological and regulatory uncertainties associated with transgenic fish. In: Transgenic Fish. Ed. C.L. Hew and G.L. Fletcher, Word Scientific Pub. Co., Singapore, Chap. 12: 209-228.

Hew, CL., Fletcher, GL and Davies, PL (1995) Transgenic salmon: tailoring the genome for food production. Jour. Fish. Biol. (Supplement A) 47:1-19.

Hughes, DR (1992) Lower jaw deformity in farmed Tasmanian Atlantic salmon, Salmo salar (Salmoniformes, Teleosti). Final report. In: Proceedings of the Saltas Research Review Seminar, 29th April, 1992. P. 17-64. Hobart: Salmon Enterprises of Tasmania.

Hunter, GA and Donaldson, EM (1983) Hormonal sex control and its application in fish culture. In: Fish Physiology, Vol. IX, Reproduction, Part B, Behaviour and Fertility Control (W.S. Hoar, D.R. Randall and E.M. Donaldson, eds.) Chap. 5. P. 223-303. Academic Press, New York, N.Y.

Hutchings, JA (1991) The threat of extinction to native populations experiencing spawning intrusions by cultured Atlantic salmon. Aquaculture 98:119-132.

Johnstone, R (1996) Experience with salmonid sex-reversal and triploidisation technology in the United Kingdom. Bull. Aquac. Assoc. Can. 96-2: 9-13.

Johnstone, R (1985) Induction of triploidy in Atlantic salmon by heat shock. Aquaculture 30: 375-380.

Jungalwalla, PJ (1991) Production of non-maturing Atlantic salmon in Tasmania. Can. Tech. Rep. Fish. Aquat. Sci., 1789: 47-71.

Kapuscinski, AR and Hollermann, EM (1990) Anticipating environmental impacts of transgenic fish. Fisheries 15(1):2-11

Kapuscinski, AR and Hollermann, EM (1991) Implications of introduction of transgenic fish into natural ecosystems. Can. J. Fish. Aquat. Sci. Can. J. Fish. Aquat. Sci. 48 (Suppl. 1) 99-107.

King, H and Lee, P (1993) Progress report: Jaw deformities and respiratory physiology of triploids. In: Proceedings of the Saltas Research Review Seminar, 29th April, 1993. p. 37-44. Hobart: Salmon Enterprises of Tasmania.

Lee, P and King, H (1994) Effects of reduced dietary energy on the incidence of jaw deformities in Tasmanian Atlantic salmon. In: Reports from the Saltas 1993-94 Research and Development Programme. P. 61-69. Hobart: Salmon Enterprises of Tasmania.

Lincoln, RF and Scott, AP. (1983) Production of all-female rainbow triploid rainbow trout. Aquaculture 30: 375-380.

Ludwig, B (1996) Position of the British Columbia Ministry of Environment, Lands and Parks on the use of non-reproductive Atlantic salmon. Bull. Aquacult. Assoc. Can. 96-2:32-37.

Mc Geachy, SA; Benfey, TJ and Friars, GW. (1995) Freshwater performance of triploid Atlantic salmon (Salmo salar) in New Brunswick aquaculture. Aquaculture 137: 334-341.

Muir, BR; Howard, R; Martins, R and Bidwell,C. (1996 in preperation) Risk assessment in aquaculture: Development of models using results obtained from Medaka transformed with growth hormones. 8th Symposium on Environmental Releases of Biotechnology Products: Risk assessment methods and research progress. "Risk Assessment st the Crossroads" Ottawa, Ontario Canada June 23-26, 1996.

NASCO (1993) Impacts of salmon aquaculture (a bibliography). North American Salmon Conservation Organization Document CNL 9329, 8 pgs.

Ossianger, FJ and Wedemyer, G (1973) Computer program for sample sizes required to determine disease incidence in fish populations. J. Fish. Res. Bd. Can. 30:1383-1384.

Pepper, VA (1991) Proceedings of the Atlantic Canada Workshop on methods for the production of non-maturing salmonids. Can. Tech. Rep. Fish. Aquat. Sci. No. 1789, 152 pp.

Quillet, E and Gaignon, JL. 1990. Thermal induction of gynogenesis and triploidy in Atlantic salmon (Salmo salar) and their potential interest for aquaculture. Aquaculture 89: 351-364.

Skaala, O, Jorstad, KE and Borgstrom, R (1991) Genetic impact on wild populations from fish farming: experiments with genetically tagged brown trout (Salmo trutta L.). In: Abstracts from the International Symposium of Biological Interactions of Enhanced and Wild salmonids. June 17-20, 1991. Nanaimo British Columbia, Canada p. 22.

Stewart, R (1996). Triploidy: a commercial application in Canada. Bull. Aquac. Assoc. Can. 96-2: 29-31.

Sutterlin, AM and Collier C (1991) Some observations on the commercial use of triploid rainbow trout and Atlantic salmon in Newfoundland. Proceedings of the Fish Sterility Workshop, Feb. 20, Halifax, N.S. Can. Tech. Rep. Fish. Aquat. Sci. No. 1789. 15-33.

Sutterlin, AM, Holder, J and Benfey, TJ (1987) Early survival rates and morphological abnormalities in landlocked, anadromous and hybrid (landlocked x anadromous) diploid and triploid Atlantics salmon. Aquaculture: 64: 157-164.

Wall, AE and Richards, RH (1992) Occurrences of cataracts in triploid Atlantic salmon (Salmo salar) on four farms in Scotland. Vet. Rec., 131: 553-557.


Table 1. Best published results for triploidy induction success and survival in Atlantic salmon and rainbow trout*
Species Method n %3n % Survival Reference
Hatch Feedling
Atlantic salmon heat 30 100 93 -- Benfey and Sutterlin, 1984
heat 50 100 67 -- Johnstone, 1985
heat 28 100 84 -- Quillet and Gaigon, 1990
heat 80 100 -- 80 Sutterlin et al., 1987
pressure 18 100 89 -- Benfey and Sutterlin, 1984
pressure 112 100 100 -- Pepper, 1996 (pers. comm.)
pressure 83 100 96 100 McGeachy et al., 1995
Rainbow trout heat 15 100 80 79 Gray et al., 1993
heat 24 100 -- -- Lincoln and Scott, 1983
pressure 15 100 85 -- Chourrout, 1984

* Best performances were selected on the assumption that experienced operators should be able to obtain comparable results. Survival rates are relative to controls.


Table 2. Sample sizes (n) required to detect at least one diploid fish in a population (N) at a given diploid frequency (F as %) at the 95% confidence limit.
Population Percentage diploid incidence (F)
Size (N) 0.5 1.0 2.0 3.0 4.0 5.0 10.0
50 46 46 46 37 37 29 20
100 93 93 76 61 50 43 23
250 192 156 110 75 62 49 25
500 314 223 127 88 67 54 26
1000 448 256 136 92 69 55 27
2500 512 279 142 95 71 56 27
5000 562 288 145 96 71 57 27
10000 579 292 146 96 72 57 27
100000 594 296 147 97 72 57 27
1000000 596 297 147 97 72 57 27
1000000 or more 600 300 150 100 75 60 30


Figure 1. Ground floor layout of A/F Protein's research and brood stock facility (Aqua Bounty Farms) located in Prince Edward Island, Canada. Required containment and security modifications include entry way restrictions, improved door locks, and sump for effluent screening.

Figure 2. Design of effluent containment system at Aqua Bounty Farms, indicating sump, screening and alarm system.