POTENTIAL FOR ECOLOGICAL EFFECTS AND GENE FLOW RESULTING FROM GROWTH HORMONE TRANSGENIC
ATLANTIC SALMON (SALMO SALAH) INTERACTIONS WITH WILD CONSPECIFICS
By o Darck Thomas Rhea] Moreau
A thesis submitted 10 the School of Graduate Studies in pania] fulfillment of the requirements for the degree of Doctor of Philosophy
Ocean Sciences Centre Cognitive and Behavioural Ecology
FacullyofScicnce Memorial University of Newfoundland St. John's. Newfoundland & Labrador
Canada
Octobcr2011
Abslrael
Growtll IU,lnnone (Gl J) tr<.lnsgenic Atl<.lntic S<llmon (SII/IIIO ,m/llrj l'xhibit tremendous growth rateS under h<.ltchcry condition~, This pl1l'notypi(' rl'sponsc has creall'd interest within the ilquacultuTe industry: however. possihk eseapel' evelllS have raised concerns rcg<.lrding their pOl<'ntinl ecologic'al impacts, This thesis appli('d an eeo~
cvolutiorwry approacll tn empiricully aSSeSS the potential ecologicnl cneets of Gil transgenic Allanlic sulrnon on norHransgenic intruspecitlc populutions. Specifically, Illy goal was to explore thc rciativc sUTI'jval and reproductive success of Gil tr;rrrsgenic and norHmnsgenic salmon undcr nenr-rmtural conditions, To nccomplislr tllis. h'y litnl'ss- rcl;rted traits were compared ~lIlcen Glilransgenic and norl-tmnsgl'nic Atlanlic sllimon over periods ofthl'ir life c'yele whcn natural sl'Jcction is typiClllly inlcnsc, Spceilicllily.
this thesis focused on the youllg-of-thl'-ycar stream and the brceding periods, TlI'ostudil's (Chapters 2 and 3) compared litness-related tmits between trJnsgenic ami non-transgcnie Atlmllic salmon during cllrl) lite history, Cimpler tllO e.\plored tire potential dirtercnces in dcvelopmenwl nLle lind respirntory metabolism bCl\\Ci:n transgenic and non-transgenic siblings at three early stages nf IiI<:: tlK' eyed-ernbr)", aJcvin (larvJI) and lirst-fceding fry (juvcnile) stnges, Chaptl'rthrc('c,>.:plorcd the loraging behaviour ami tire grolllh and survival oftmnsgenic and non-transgenic tirst-feeding fry rcarl'd undl'r lOll feed, stream-like conditions. Colledil'Cl), the results ofthcs.t.' ehaptcr~
suggest tlHlI lhere is JIl nntogenl'tic dela) in the phl'nOI) pil' rl'SpOnS(' induced h~ the [rarr~gcne_ ~uch tll;]l biologic'ally signilicanl dilkn.:nces in lilne~~-rclated tr,lit~ h..'[I\l'l'n Gil tran~genic and non-tr;rrrsgl'rric At];rntic s.;rlrnnn arl' minimal durin); this critic,11 e;rrl~
lili: histor) [X'riod,
rile ilnalt\lo studies (Chaptcrs 4 and 5) comparcd fitness-related traits betll'ccn transgenic and non-transgenic Atlantic salmon during the reproductive phase of the life cycle, Tht" fourth chapter compHed tht" breeding performance of growth hormone transgcnic ~nd non-transgenic Atlantic salmon males of hoth alternative reproductiw phenotypes to test for the potelltial of the transgene to introgress into wild populations.
The titih chapter uscd populations ofGH transgenic and non-transgcnic Atlantic salmon sirolings 10 elucidate tht" elTeCls of growth nn precocious parr matumtioll. Collectively.
these data suggcst that trallsgClli!; m,lks nMy \ .. ~pcriCIKC rcdu!;cd rcprodu!;tivc SUlTess rdativc to non-transgcni!; individlHlis. llowcvcr. Ihc pot,·nti'li for thc lr:msgcnc to inlrogress inlO wild populations was demonstrmed. The empirical contributions of this thesis inform dccisions regarding thc potcntial ccological impacts associmcd \Iilh GH transgcnic Atlantic salmon.
iii
Acknowledgements
This thesis has coincided with m)' 1,lte twenties As is COl11mOI1 II ilh this period of life. I haw undergonc a great period of person~1 dev\:lopment and personal understamJing. While Ihc ne:>.:t scnlCIH;C may sound vcry \:Iiche. il is ;11~0 very accur,lte M) years as a gmduate sludenl have h\:lped Ille del1ne who I al11 and th..: type of life I Ilouid like to lead. The graduat..: student experience has been an exe..:ptionalvenue to l!l1denake this phase in the journey of life. There arc many peopk you haw helped m..:
either directly or indirectly and I could not possibly do them all justice here lIollevel".
therenre(l/i:windividualsthm I would like 10 lIlenlion
First and forel11ost. I would like 10 thank Isabel Cosla for her low :md companionship. Wilhout Isabcrs support I may vcry Ilcll h,lve losl my mind somc time ago. A specialthauks also goes out 10 all the friends I h(lve made throughout the )c,lrs The growlh one e;o.;periences participaling in diV("rsc social activities is underrated hy till'
;Ieadem),. I haw no regrets about living lif..:: as full as I hale during my tilll": at MUN I am also vel)' thankful 10 m)' ~upervisors. Ian Fleming and Gunh I'kClher. and m) commitln.· memlwr. Kurt Gall1~rl. II has been a pleasure to work wilh three researchers that hal'e been so sueeesstill within their respective areas of research. My e;o.;posure to Iheir often dissimilar perspn'livcs on biological queslions has providl'd Illl' II ilh a \\ell rounded lense wilh which to design and inll'rprl'l biological rese:lrch. This ha~ 1\lrther be..:n (tided b~ n dynllmic group ofsludel1l colleagues. \Iho. like me. rl'll'l in IIK'ntall'Xl'rcis('s Ihal challenge the statlls ljllO. I'inall). I'd likt'tnthank Ill)' tamily I\ho, de~pitl' being supportive. haH' now quite undl'rslood what it b I hflle b..:en doing nut in NCllf()undl,md ,Iff this time. or why Ilnpefu!l), il Ilillnmh' Sl"nSl" unl" d:I).
TahlcofContcnts Abstract
Acknowledgements
Table ofContents
Listof Tables x
List ofl-igures xiii
ListofAbbreviationsandSymbols Co-AuthorshipStatement
Cha ptc r I:Gcncrallntroduction
1.1Introduction
1.2 Genetic Background
1.3PhenotypicExprression
1.4Domesticat ionSelectionand Divergence 10
1.5 Effects in Non-nativeHabitats 1.6 EffcctswithinNative Habitats
1.6.1 Competition
1.6.2Intcrbr ccding andIntrogrcssion 1.7 Casc Study ofSalmonidGrowthEnhanccmcnt
1.7.1Comparing Growth-selectedandGIl-enhanced Phenotypes IX
1.8 ThesisOutline LL
1.9Referen ces
Chapte r 2: Grow th horm onetransgcn csisdocsnot in fl ue nc etc rritorialdomin anc c or growthandsurviva lof first-fc cd ing AtlanticsalmonSa lmosa la r in food-li mit ed strca m mic rocos ms
Abstrac t 2.1 Introduction
2.2Meth ods
2.2.1ExperimentalAnimals 2.2.2Domin anceTrials
2.2.3Stream Micro cosm
2.2.4 StatisticalAnalyse s
2.3Result s ,'0
2.3.1 DominanccTrials 2.3.2Strea m Microcosm
2.4Discussion 2.5Acknowledge me nts 2.6Ref er en ces
vi
Cha p te r 3:Dela yedphenotypicexpre ssionof growth hormonctransgcne sisduring ea r lyo ntogeny in Atlantic sa lmon(Sa lmo salar)
Abstract
3.1Intro d uct ion
3.2Met ho ds
3.2.1Expe rimentalAn ima ls
3.2 .2Respiro me trySyste ms
3.2.3Respirom etryProtocol 0.'
3.2 .4Development
3.2.5Dat a Ana lyse s
3.3Results 01
3.3.1 Resp irom etr y
3.3.2Development
3.4Discu ssi on
3.5 Acknowledge me nts
3.6:References
vii
Cha pte r 4: Reproductive performanceofaltcrnativc malcphenotypesof growth hormonc transgcnic Atla ntic salmon(Salmosala r)
Abstract ...,10
4.1Introduction
4.2Methods
4.2.1ExperimentalFish
4.2.2Exper imental Design
4.2.3Behavio uralObservations
4.2.4l'arentalAnalyses
4.2.5StatisticalAnalyses
.. ...118
4.3Results .:..:.
4.3.1AnadromousMales
4.3.2PrecociousMale I'arr
4.4 Discussion 4.5Acknowledgments 4.6References
viii
Chapter 5:Enha nce dgrowth redu cesprccoci almalematurati oninAtla nticsa lmo n (Sa lrno sala r)
Abstrac t
5.1 Introd uction
5.2Method s
5.2 .1Dat a Analyses
5.3Results , ,-' L.
5.4Discussion
5.4.1Implications
5.5Ackno wledgme nts
5.6Refer ences
Chapte rIi:Gene r a l Discu ss ion
6.1Gene ralDiscussion
6.2Conclusion
6.3Referen ces
List ofTabl~s
rabie- 1.1. (j~ner31 palterns or fitn .... ss·related trait diwrgem:e b~IIH-~n \\·ild and growth sekct<.:d (rarmed). growth horl11on~ (Gil) transgenic or (ill treated salmonid popnlations compared in artificial laboratory tanks or aquaria. Trait direction rclkets a comparison of the growth enhanced relative to wild·type tishes. Re1i:rences provided arl'not exhaustive
and arc intended merely as exampks .39
Table 1.2. General panerns of fitness-related trait divergence between wild and growth seiccted Olumed). growth hormone (GH) transgenic or GH trcaled salmonid populations compared in natural or near-natural laboratory environments. Trait dire(tion rctkcts a comparison or the growth enhanced relative to wild-type fishes. Each numerical value in the direction column represents ol1e measurement from a singk treatment within (I given study. TillIS. some studies contribute multiple values 10 the direction column. Rl-fer~nces provided arc not exhaustive and ar~ inl~nded mncly as t'xample~. . .42 Table 2.1. The me:m .i S.L initial and final mass (M) and I~)rk length (/.I) mcasun:mcnts of first· feeding growth hormonl: transgenil" and non-tr,lIlsgenie S(l11II1) S(llar rr) rransgenie ,lIld non-transgeni\; high and low d~nsitics inlll'ar- n,!\uralstream
rablc 3.1 C<lndid,lte models (ANOVA) desnibing the l'lTel"ls or ramily "rig ill and growth hormone transgenesis on the routine o.\ygen conSlimption (mg 02 g.1 h(l, response variable) of l\tlal11ie salmon (SII/II/V sa/(//") at three stages of ellrf)' fife histllr"). k represents the number of predil·to~ in eadl model. Mod~1 fit is rcpresentl-d by comparing A,A Ie. and Akaikc wl-ights (I\'i)' A,A Ie, refns tllth~ changc in Ale,. between OI1l' model and thl.' 1I10dl.'III'ith lowest Ale score and 11", refers to the probability that the focal modci provides the best representa1ionoftheda1lt rciutivetotheothl.'rcalldidatc Illodels following repeated analyses (II', SUI11IO 1.0). [3oth fixed ~md mixed Illodeb arc prl.'sentcd beeallse a sil1gle selection criterion wus inapproprimc Ill1crprt-wtioll i~ ba~cd predominately 011 the mixed modefs. \\'ith fixed elleet, models used lor inferences ahout
the !ixed elleet variable (genotype)... 101
Table 3.2. Candidate models (ANOVA) describing tile efleets of fhmily origin and hormone transgenl'sis on the mass (g: respollse variable) ,)f f\tiantie salmon lifehistory.krl'presenlsthenlJlllb('rofpr<,dictorsin l'aeh model. Model fit represented by comparing L\,AIC, and Akaik<' w<'ights t.,AIC, reters to the change in AlC, between one model and the modclwith lowest score and n', refers to tile probability thm the focal model provides the be~t represell1atioll of the data relalive 10 the other candidate models lollowing repealed analyses (II", Sllill 10 1.0). Both l"txed and mixed modcl~ arc presented becau~e a single selection criterion was inappropriale. Interpretation i~ based predominately on the mixed models. with Ii:-.:ed effecls moJeb uSl'd for inferences (Iooulthe fi:-.:ed effect v,lriable (genotype) .. 102
rable 3.3. CandiJate moJels (binomial logistic describing lhe effecl~ of family and growth hormone Iransgenesis on lime of Allanlic ~alm\'ll (0'''/1110 rhe respon~e variable represents lhe proportion of halclwJ individuals carrying the transgene. k repre~ents the number of predicto~ in each model. Model tit is represenleJ by comparing L\,AlC anJ Akaik~ weighl~ (Wi). A,Ale refers 10 lhe change in Ale between pne moJel anJ the rnoJel with lowt"sl Ale scnrt" and '\.! rd"as to thc' probabililythalthl' focal model provides the best represcltlation of the data relative to the other candidaTe models lollowing repealed analyses (It", sum 10 1.0). Both lixed and mixed models arc presented because a single selection criterion was inappropriate.
Interpretation is based predominately on the mixeJ models. with fi:\ed eili:ets models used lor inferences aboUT the fi~ed efleet variable (gt·notype).. . .. 103 rable 3.4. Candidate models (ANOVA) dl'seribing the effcet~ of jiunii)" origin and growth homlOlW lransgenesi~ on iJlevill physi!;al dlaradt"risTi!;s Ire~p"nse variables: yolk-
~<.I!; are<.l (mill"). rna~s (g). iJnd knglh (Illm)1 of Ali<Jnlic salmon (Sa/mo sa/ar). k r~pre~enls lhe number of pn:diuor~ in e<.l!;h model. Model fit is represenlcd A,AlC and Akaike weights !\,AlC rel,:rs To Ill<." dlange in Ale between one
anJ lhe mOlJeI wilh s!;ort" and lhe focalmodcl
proviJes lhe best rep~senlation of The relative TO thl' lollowing repeated nnnlyses (II", ~urn to 1.0). 1101h li .... ed anJ mi~cd h~!;ause a single seledion nilcrion was inappropriatl·. lnlerprclation predominalely on lhc mi .... ed models. with fixed efleets models used for inkrences aboul
the l!xed efleCl I'ariable (genotype).. I O~
Tabk 3.5. Pearson's prodlld-rnomenl correlalions performed llsing family-In'l"f llh."allS K' l'xpINc' assOI:ialions beTween egg chamCTl'risties of lmnsgenic and non-transgenil:
Allanlil: salmon (5(1/mo sa/ar). Egg characterislics included initial dry mass (M,.~!J.
50%. hatch (dd~,.,J and mass·independent oxygen conSllmption (egg~'l(), Ill'
and alevin characteristics near emergence. including yolk are~l (:\,,~d. Ilct
lenglh(L.lnml .. 105
Tablc 4.1. /I.·lean lork length (em: ±S.E.) and mass (g: ±S.E.) of th.: Illatur.: Atlantic salmon used in eompetitive oreeding experiments (JII transg,-ni'·'llld non- tnlnsgenicaiternmivereproductivephenotypes. In trials cOlllpilr.:d age 0;
tran~g"-nic (T) VC"f'iUS 1+ non-tr,msgenie (NT) parr and five tri,ll~ cornpared 0+ transgenic v.:rSLIS 0+ non-transgenic parr: th~-size ofpilrr involved lIre reported separmely for each
;rge !;llmparison below. The N for each fish type is provided in p;rrentheses.. . .. 140 Tahle 4.2. An Clhograrn descrihing the spawning h..-h;rv;ollrs rm:aslired during compctitive trials hetween transgenic and non-tr;rnsgcnic Atlantic salmon mak-s the arwdromolls and parr rcproductive phenotypes ... 141 Table 4.3. Nest lidelity (proportion of time spent with nesting female) of arwdromous gro\\1h hormone transgenic and nOIHransgenie Atlantic salmon males during paired competitive breeding trials. Each breeding trial ineludcd phases ofeompetition alld no competition. Duringeompctition. both the transgenic and non-transgenie malescompetcd directly for breeding opportunities with the female. During no competition. males had sole acecss to a spawning lema Ie. Data Ii·om each trial II·cre analysed 60 min hl'i"ore (pre- spawn) and 30 min aftcr (post-spawn) cach spawning cvent. In trials with no spawning_
were based on observations conducted for 5 min intervals 30 min lor the
Tabk 4.4 Nest lidelity (proportion of tim..-spent with nesting r..-male) during paired eompetitive hreeding trials ofm;rture male p;rfr that were growth hornl()rll" transg<·nil" and non-transgenic. The lirst and second (8) spawns from each trial werc analysed tor the period 52.5-12.5 min the spawn (prc-.I"!)(/H"II). 12.5 min on either side-of the spnwn (.lpall·II). and 12.5-22.5 :Iller the spawn (po.\·/-SPi/\,.,I)... . .... 143 Table 4.5. Tile fertilisation sueee~s (prol>onion of eggs knilized) of wild IlnadromOllS males and growth horllIonetransgenic and non-transgenic llIatlirellIaleparrduring II jJair-wiseeompctitiwbrecdingtrials. Rcprcs.:ntation indirat<-s the nUllltl<"ro1"tri;lls II"lwrc succcssflli fcnilis<llionwasohservcdhyamalrtype ... 144
List of Figures
Figure 1.1. 1'1011' cilurl summuri7inJ,! the potentiul ecological 1111(1 genetic impllets of aqll:leultureeseapces in nativealld 11011 native habitats... . ... .46
2.1. The experimental apparatus lIsed during pair~lI'ise dominance trials with
""" 10). Located immediately downstream of each other lI'ere eomest
enclosures. enelosurc consisted of mesh ends and PVC side purtitions. The mesh partition scparating each cnclosllre could be rai~ed. allowing intruder entry during prior residencc trials. Fn:shwatcr nowed thrQugh a spray bar allQwing u current :ICroSS the width Qfthe apparatus at apPrQ:-.:imutcly 30-50 mm sec". A wut.:rdepth 01'70-100 mill was l11uintuined by the addition of a modified. grey pl~stic bOllom covl:red with a thin layer of a nlltural grllvcl substr;lte. Si/.l' differences between til<'" upstream and d()wnstream contest ureas were al'counted for in the expcrimcntal d..-sign. Specilically.all prior n:sidenl"t.~ trials were conducted in the downstream eonlcst ellclosures and an equal Ilumher of cohabitunt trials were conducted in lIpstr ... arn and downstream
enclosures... . . . .... 70
Figure 2.2. F:-.:pcril11ental stn:um mien)<.'osills (/I '" 8) used to compare the efleets 01 density on the growth und survival of transgenic and non-transgcnic Sa/Illu sa/ar fl)' Inno\\' spray bars were positioned behind a screen partition ill the IIpstrcum end ofc:lch stream. creating a unidirectional. clockwise !low within ea,h trough. Ar/l'lIIill spp. Jrip i(}()d (klivery tubes wer, positioncd just ~bove th, w~tcr surf~[,e at IO()-140 111m b~l()w the upstream screen and again half way down th, microco~m to ,nsure feed w(lIlIJ he a~c,ssibk tl1l; full kngth of each stream. The current speed within each streal11 ranged frplll 120-ISO 111m s" upslr,amt(l}O-SO 111m ~.I downstreal11. The b0110111 of each chnnnel
\\'~~ [,()vereJ with 5-15 111111 gravel and 50-150 111111 rocks to creatc habitat he1,rogelleity ... 71
2.3. Th, p~t-r()rmaIK" Ji~rlu}'ed in p..-rcentages. hormone transgenic Juring pair-wisc dominancc contcsts with fry IInder three scenarios of competition (cohabitant. rcsiJent and intruJer). l'erfonmuKe was nJ<>asur<,d
by wins (blu,k). I"sses (white) and draws {grey)... 72
Figure 2.5. The of growth 10w(LD}
wH.kr The tr' arc
c;ltegori>'ed a~ f(\ll(\w~: II!) lransgenil' (filled circle). HD norHransgenic (open eircic·). LD tr,m~genic triangle) and U) non-transgenic triangle). The M and LI r~btiunship rt:pres~nh .. d by nwan ± S.I resiJu,ris II ith II linenr regre~sil)n 01 initi;llaTld final natural
eonsumption (mg OJ g.1 hr·l) of tran~gel1i .. · anJ al~vin full siblings Transgenic and non-transgenic mean values within repr~st:llted by Illack anJ whitc circles. respectively. The shon and long dashcd linl's represent th~ m(>r;rll lransgcnic and non-transgenic rneans. respeetively.. . .. 106 Figurc 3.2. The tillle ofhmeh (degree days) of full-sibling transg~ni(, anJ non-transg(·nic Atlml1ic salmon .wllll") from eight farnili(>s. Tht'se data an' r<'pr<'sl'nkd as
,em"""',,
r'm~'n'o", of"PI>co""""'y
I 00 individuals per family.. . 107 transg~nic;mdFigure4.1. An illustration of the naturalis(>i.lstre<.llllllleSOCllsm(1.2SlllxIS'(ullxO.25mpa channel). which was dividt'd into two channels and used 10 compar~ th~ rerrodul"ti\"~
performance of growth hormone transgenic and non-transgenic Atl<.lTlti .. · s;rlnwn (Sulll/o saInI") mnles. both asanndromolls fish nnd precoeial parr. Behavioual data l\"erceolieCied n combination of video observation and PIT tng dctection. with the resp(>ctivc cameras nnd antenna moved in response to the location of iCnwle nesting nelivity. Thickarroll"s indicatc the direction ofwateriloll" ... 137 Figure 4.2. Swndard box plol frequencies of (A) overt aggressive behaviours by transgl'nic and non-trallsg~nie anadroillolis and parr males during paired compctitiVl"
brecding trials and (13) quivering by transgenic and non-transgenic mwdwllloliS nwics during the l"Ompl,titive and ph;rses. For graphical purposes. thes~ data wac standardised to a 90 minut~ r~riod. Th~ top <.Ind bullom of ~ach hox represents the lower (25%) quantiles. respectivcly. The horizonlalline within e~ch bo.\ Illcdi~n. The vcrtienl lines (whiskers) c.\tending fn'lll thc upper (Ind lower quanliics 1he maximum (lnd minimum values of 11ll' distribll1ion. excluding the The oUIli .. 'rs arc represented by the dots iocllted
h~yond the maximuill and minimuill whiskers l.ll!
Figure 5.1. The inc idence(%)matur emaletransgen ic andnon-tra nsgenicAtlant ic salmon parr (Salmosalar)duringthe first(0+)andsecond (I+) yearsoflite.High and lowfeedlevels were appliedonly duringthe firstyearof life.Therea fte r ma intenance levelswereused. Theerrorbarsreprese nt the 95%confidence inter va lsaroundthe ... ... .. . . ... .. ... . . .. ... .. ... ... . . .... .. . . ... ... ... ...165
Figure 5.2.The mean (A) wetmass(g) and(B)fo r klength(mrn)of transgen ic andnon- transgenic precociousmale Atlanticsalmon(Salmosalariduringthefirst (0+)andsecond (1+)yearsoflife.High and lowfeedlevels were appliedonly duringthe firstyear oflife.
Therea fter maintenancelevels wereused.Theerror barsreprese nt the95'Yoconfide nce
intervalsaround themean.. . 166
Figure5.3. The length-m ass relatio nshipfortransge nicand non-transgen icprecoc ious maleAtlanticsalmo n(Salmosalariduring the first(A) andsecond(13)yearsoflife.
and low feed levels were app liedonlyduringthe firstyearof life. Therea fte r
Figure5.4.Natural log transfo rmedgonada landsoma ticmass(g) of1+matu re male transgenic and non-transgen ic AtlanticsalmontSalmo salari parr. Thedas hedandsolid linesof best fitreprese nt the non-transgenicand transgenic parr.
List nf Abbrc"i:ltion.~ and Symbol.~
011-Growth Ilormonc
OII-IGF-l -Growth I-Ionnone-Insulin-likc Growth Factor I lJNA -Deoxyribilnue!;'ic Acid
c.-Cirell RNA -Ribonucleic Acid II1RNA -Mtossengcr Ribilllurleie Acid eDNA - Complementnry Deoxyribonucleic Add N -Nonh
\V-West spp.-Spccies(pluml) L-Light D-Dark I'. -Versus II-Sample Size M-Mass llllpubl.-Unpublislwd c.g.-E.\1l1llple i.e.-itl est (in othcr\\ords) 11I-Metcl3 em-Centimetre., mm-Millilllctres /'1 -Fork Length
L -I.englh
G -Instantaneous Growth Ratto GLM - G<.-n<.-ral Linear Moo<.-I
Tuk<.-y liS!) -luk<.-y Ilonc~t Signilicaln Difl(.-rcnl"C P(-valuc)-I'robilbilit) V,lhrc
LR -Logistic Regression
"i-Chi-squarc
""-EqlWI Sign
x _ Multiplication Symbol S.I::. -Standard Error
%-I'<.-rc<.-nt
> -GrcalcrThan
<-LcssThan g-Graflls
± -I'lusor Minus cl al. -etalii (andotlwrs)
USDA -United States Departmcnt of Agriculture OSC Occan Sci~oncc~ Ccnlrc
S-Stoconds liD-High Dcnsil\
L1)-LO\, Density
I.og-I.og:nithm
PCR -Polymerase Chnin Rcnction
°C_ DegrecsCeisills min-Minutes mg-Milligrams
hr-I-Iour
ANOVA -Analysis of Varia nee AIC -Akaike Information Criterion
D,AIC. -Akaike Inforlll~ltion Criterion (corn:c1cd for SIll<l1l ~ampl~ ~i/.d A, -R~i<lliv~ l'erlorm<lllCC (stalistinll IllIKlt"I)
1<',-AkaikeWeighls
k -Rcprc~enlS the number ofrr~diclor~ in ~<ll,h mlKl~1 rv1(h-o.,ygcn Consumption
err-Canadian FOlllldation lor Inno\'ation
NSERC National Science <ll1d Ellgille~ring Ih-S~:Jrch Council Jd -D~gr~e Days
A -AI~vin Ch<lraClerislics
I'VC-l'olyvinylChloridc I'll' -I'ussive Int~gnlled 'iransponun IIDD/DVD -I lard Disk Drive/Digital Video Disc V Volt or Test SI:l1istic lor Wilco)(on Signed Runk Test pl-I\:linnlitn:-
ng-Nanograms 111M Millimolar
dN'I"P -l)eQxynu,'kOlidc "I'riphosphate pM-MicrOlllolar
KCI Potassium Chloride
Tris-HCI-Tris(hydroxymcthyl)aminomethane-Hydrogen Chloride
MgCI,-Magnt"siulll Chloride lJ -Enzyme Unit
t Test Stmislie for Paired t-test UnRII-Gonadotropin Rckasing Ilorlllonc I" Transgenic
1\'T Non-transgenic
F-Test Statistic fOflllodds using the I-'distrihution
Co-authorshipStatellu'lIt
rhe work described in this thesis was 1I0t an individual contribution Co-allihor contribution fore(lCh chnpter is described below
CIWfI/('rl' Mo,t or this introductory ehaptl'r will also servc as a book chapter in ;111
upcoming Wilcy-Blackwdl public<ltion. D<lrek Moreau was responsible lor thc background research and writing. J,1I1 Flcming (co-<llIthor) nnd Garth Flctdicr (oook editor) providcd cditori(11 assistance;lI1d intellectual input
/I/O/Hili. D. fH .. alld I.A. Flemillg. 2UI I.
nit,
1'0lelllia/ (',:"Iogim/lllld g<'lIC'Iic" illlpllC/.> (~(IIIIIIIICIIIIIII"<' bioledmologics; E,;o-""olllliOI/(/I"Y ('OJlSi(/eralioIlS fiJI' 11f(/I/(/gillg lIlt,
""I<'
/"('1'01111/(111. III. AqllllCU/lllre /Jioicc/lllology ('(/. F/cl("h('l·. G.L. lind His('. AU,.). Will,.\"- Illlld,"d/. lIo/>o/;{'II. NJ. USA. P. XX-Xx.
C/Wfi/tT 2: This chaptCf is published in the Journal of Fish Biology (Wiley-Blackwell).
Darek MorC<ll1 W<lS rl'~ponsible for tl1l' concept dl'vl'loplllenL cxperinlt'ntal dcsign.
constnll"tion. data collection and analysis. and manuscript prep;lralion. Ian Fleming (l'O-
author) assisted with concept development and e.~JX'rirncntal dcsigll. and cditorinl nnd intcllccltlal input. Garth I'k1Chcr(co-1l11thor)providcdeditori;lIassistancc:tnd intl'IIL'clual input Joe Ilrown (co-author) assisted II ith l'onci'pt d~·vl.'lopml.'nt nnd provkkd intcllccltlalinputllpl"iori
MOI"<'(IlI. D. TR .. 1.11. nemillg. (iJ Fle/ch"r. alld .I.A. IJI"!!H"II. 2011. (;"')\1"111 /wl"lmm(' //"lII/.,·gclI(,.,·i.< (/(1/'.' 110/ /1'I"I'ilo/"ili/ i/ollliIWII("t' O/" gl"oll"lli 111/(1 .'·lIn·hlll /i.-nlillg Alllllllie .1'1/1111<111 L il/ .filOd-limilnl .I·/r('(/I/I microcosm.,·. o(
Fish liio/!!!!,y 7N: 72o-74IJ.
Chop/"I' J: Darek Morc(lu \\a~ rcsponsiblc for 11ll" nllll"l'pt dl"li'loplllcnt. c,.,pl'rinl~·ntnl design. construction. data collection nnd 1I1Hllysis. (Ilul manuscript prcparmion 1(111 Fleming (co-1Hl1hor) ,ls~isted \\ ith concept dewlopmcnt <Inti l'xpcrim~"lt<l1 (k~igll. and p/Ovidcd editorinl nnd intellectulll input. Kurt Gamperl (co-author) as~istcd \Iith
~xperirnental design. ~diting and inkllectual inpul. (j~r1h Fletcher (co-author) prtlvitletl editorial as~i~larKe and intdkctual input
ClW{JIi'1" 4: This chapter is published in I·:volutionary Applications (Wiley-Illack\\ellj
Darek Moreau was responsihle I(lr the concept devclopmerll. experimenwl design.
construction. data collectioll and llrwlysis. and mlllluseript preparation. Ian Fleming (co-
author) assisted with concept developmcl1l and experimental design. and provided editorial and intelleclllal inpul. Corinne Conway (co-alrthor) assish:d with experimental design and dataeollection
m(/I,.
I)()I
Chap/a 5. This chapter has been accepted for publication in Function;rl Enllogy (l1ritish
Ecologic(1i Society). Darek 1\1oreau was r,·sp,msible for the experimental design.
constnu;tion. data collection. data analysis and manuscript preparation. Ian Fleming (eo-
auth0r) assisted with c(mcept development and experimental design. and provi(kd etlitorialand intellectual inpul.
("fWI'I'"/" n. Some paragraphs of this conclusion chapter will also snVt' as a book chapta in an upcoming Wiley-Hlaekwell publication. Darek MorellU was responsihle tilr the background research and writing. Ian Fleming (nl-author) and Garth fktchcr (hook etlit(lr) pwvitietiedilOrial assistmlccand intellectual inpul.
a/fll(/Cllfllll"t' ""","',,,",',,''''', E",~,"'"""'""",,v
n·I'ofu/ilm. In
",,,/,,, , ,,, ,,, ,,,,,",,',,,1
hlllt"
C hapter 1
General Introduction
1.1: Introduction
Hunwn histor) has hc~n in~:»tric<lbly sh;!p~d by th~ ~:»plnit;J1ion of !;al'tive animals. In conjunction with plant agrieuiturl'. captiv .... -r .... ared anim;!ls provided alKestr;11 humans with an abundant and r .... !atively stabl .... food surplus. r .... suiting in in!;reased popul;!tion d .... nsiti .... s ;lI1d station~ry !;omnHlIliti"s. The advcnt of ugri!;ul1ure pcrmit1ed the diwrsificntion of labour I .... ading to rapid adv;lIlc~s in tedlilology ;md nwrl' sophistiGlteu hierar!;hical political systems that gradually came to dominate traditional hUlller-gmhcn:r
s(l(ieties. Di,n!lond (2002) has suggested tlwt increased human population sizes. a r .... dudion (ovCfe:»ploitationj in large m;!tntnal popul;!tions. ;!nu tedmolngi!;ul adVillK!;S leading to fond storag .... wert~ prim;lry reasons for the initial transit inn to agricullure. "
simil;]r, contemporary transition app .... ars to h .... an:ompanying Ihe global tleL"line in aqu:tli!;
food resources. which has h~en brought about largely by incn:;lsed human ~:»pl()itali"n and damage 10 :Jquali!; !;!;osystcms. In parallel with our Holo,'ene <llKestors. we have eSl'alatcd Ihe dev .... lopllll'ni of new technologies for th;.' eXploilation ofcaptiVl" ;l'lliatic animals(i.c.aquacuiture)tosupplemenicaplUrc-basedtishcries.
"qu:tnliture. Ihe niltivation of aqu<lti!; organisms. is a primary tool llsed to supplement depressed glohallisheries ;]nd. in till' l(ltul";'·. may help conserve hi."avily exploited populations. Over the last flny years. the so called "[3ltle Revolution" has tfallsfortlwd aqu:Jcultllri." from a localized agricultural activity producing less th;]11 one million tonnes oI1(l{-..:1 annually 10 ,I global industry. pn-..:!ucing m~arly Si:»ly million hlnne~
"f«loti annually (FAO. 2(10). COlllnwrcial aquaculture aCC()UTlt~ for lwarly half 01";]11 food !ish production Ilorldll'ide and that proponion is expected to continue to grol\.
The rapid growth of aquaculture has been aided tremendously hy the application nfscience. Indeed. hiotechnology is partly responsihlo: lor thc unpreecdcl11ed growth in the volume and diversity of species involved in modern aquaculture production (Duancct
al. 2(07). Advancing beyond hreeding only thc largcst :md healthiest l1sh. hUI11:mscHn nO\1 ~,,·kl·t individu~1 lish with the aid of observations at the gene level (c.g. gcnc exprcssion proliks, (Jullntitative trait loci) or manipulate the genome dirl'ctly (c.g.
trnnsgenesis). It is expected thm biotcchnologics will C()llIillllC to aid til<.' development of trnitsadvamageotls for production
Whik;!quacuhure isdewloping in tl1l' biotechnology ern. it is also devcloping in ,1Il erd ofcnvironl1lcntal c()n~ciousm:ss. There is now considcrnble pressure to minimize anthnlpog .. 'nil" impacts on the t'llvironment and biodiversity. which is largely in response to thc unprecedcnted t'nvironmental changes we arc seeing globally. COl11l11crcinl aquaculture has not escdped allention and associated environmental concerns have been a centre of dchate (C{)~ta-Pierce 2002: FAO 201 0). From an eco-evohl1i()I1~lry per~pectiVl'.
perhaps the most complc~ concern il1v()lve~ the potenti.11 n'ological and genctic impacts
"fa(Juacultureescapl,cson surroundingceosystcms
There arc numcrous sllidies on thc ccologic;!1 and gcnctic effects ofaquaculturl' cscapees(revielled inlJl1erand 1-.pifanio20U2: l\aylorct al. 2005: Thorstad ct nl. 2001':) While lhe Illlldamrntal questions rcmain similar. it is widely recognized that modern r.iotechnologics. sllch as tmnsgcnesis. inlroduce nell" challenges to this already <,omp!.::.\
Ihi~ dis~l'rlatioJ\ provides a thorollgh empiric;ll ;1~seSSlllell1 or the potential ellvirollmcntal ellccts 01 growth hormone ((ill) tran~genic Allantil' salmon (5"/11111 \"IIIa,-j enlry inlo lhe wild:;1 leading candidatc for conHllercializatiol1 I'rior to presenling the
empirical core of this thcsis. I revil'll the eco-evolutionary contcxt and ~'.~;sting lik'mtllrl' on the potential impacts ofaquacilliure l'scapces. Specilically. tlK' eurrelll rhaptl'r 100(:uses on the similarities and dilTerenees betl\eell Icehnologies applil'd in aid of intentional sell'l'tion (i.e. Illarkl'r-assislcd broodstol'k development) and thOSl> il1lolving dirl'l·t genl'tic ehangl'(i,e.transgcnes;s). I shall providcevidenel' suggesting that the cl'ological and genetic impacts of transgenic animals may be more ditlictllt 10 predict than that of animals produced through mnrker-assistcd brood stock dcvelopmcnt and introduce till' empirical Ilorkeontained in thisdissenation.
1.2: (;cnctil'hackground
Aquacullllre biotechnologics. stich as marker-assisled broodSlo(:k developnK'nl.
Ihm do nOI involve dirccl genetic manipulation remain simil,lr \() traditional intentional sekction (i.e. anificial sekction baSc'd ~tridly on the phenotypil' e:-;pr('ssion uf traits) rhis is heeause till' only procl'ss that dilTers is thaI by which parenls an: sell-etl'd. In terlllS ofevolutionar) processes, th('y all involve Illullipk gencrations ofsdel'tion that re~tlit in eoncllrrclll genomic changes and novel phenotypes under polygenic control (Mignon-Graslcau 2005: knsen 2006). Thus. I "ill not dblingllish bellleen 'Hlim'lls produced using such bioteclmolngil'S and those produced llsing traditional artiril'ial st.'kctionand "ill rctl'rtobolh as farmed strains.
Farmed specics lend 10 IHlve loll' genelic diversily "ilhin and among popubtiollS rel,llile to Ih;11 of"ild populatiolls. Genetic si1llilaritie~ app .. :ar hI he tl1l' re~ult of: I) Ihe I()\\ numb,'r of initial hroodstock smlre .. · populalions and 2) Ih(' selet'lion of similar Inlib over multiple generations !'Of example. Atiantil' salmon (S(I/III(I s"ltlr) hreeding
programs hav(' developed with local populmions in scveral places. including Easlern Canada. Norway. :lnd Scotland (Ferguson el al. 2007). For euch region. howe\'er.
programs have ('ilher hegun hy. orendcd up eonecmratingon IlO morclhan a fell slrnins.
which arc nOI likely rcpresenl<llivc of loc<ll populalion Slructurc (Gjedrcm 1'1 al. 1991: Gjoen :md Bentscn 1997: G1cbe 1998). Moreover. some cultured stocks ,Ire tr,mspl:mted or hybridized with locul culturcd slrains. further homogeni;-:ing strains used in thl' il1dUslry (Ferguson el al. 2007). Thus. Ihe rel,llively low num~r of strains involv('d in broodstock development contributes to genetic simil~rities among artilil'iully seit'etl'd populmions and genelic dissimilurilics relative 10 wild populalions.
The seleclive pressures on fanned POpuhltiollS arc Onl'll very similar: that is selection lor high growth and survival under like culture condilions. !'rOIll an evoluliooary perspeetivc. parallel el'olution among domestic,lted strains might be expecled. whereby different lineag{'s shan' genotypic similarilics due 10 simil,lr evoltltionary pressures (Fosler and Bilker 2004: Schluter el al. 2004). Indeed. Roberge et al. (2006. 2008) demonslmled parallels bctwcen the transcription profiles of two le,tding
"tlalllie s.1lmon :lqlJ;lculturl" strains from East("rn C(mada and one from Norway
"n,llogous p;l1ternS have also b<'<.-n obs.:.'rvcd in the tr:lnseriptoillc of closely relmed C01wn congeners (C{).1"")l'irllll ht".h{/(It'lr~'{'. I., l!in/lllllfl) in response III artilicial seleelioll (Chaudh<ll)' t"l al. 200!\: Ilovav C! al. 2008). Evidence for convergent el'oluti()n als"
('.\ists among IJroSOI'Iti/u ,\'11",,11.,'('/11"11 popul;ninlls \Ihen eX('JI)sed to lalxlratnry environments over multiple generations (Mato~ el :11. 2000: 2002). Thes..: linding~
Sllggl"st that bOlh additive and non-additivc genelic varimion IIW) COllwrgl" amung dislinct poPlil~l1ions experiencing ~imilar pressures from ar!ifll'ial sclection.
The phenotypic enhancement strategy oj transgenesis is very di1rercnt n·om intentioJ1:l1 selection. Speeitic phenotypes arc targeted by gene irlsertion and thus the traditional. polygenic. process of artificial selection can be bypassed, The 1110st C0l111110n method of creating tral1sgenic tish is currently cytoplasmic 111icroinjection. \Ihere llluitiple copies of the transgene arc inserted into the cytoplasm ofa recently fcnilizedegg and the- transgene-(s) are-ineorponlle-d into the de-vcloping zygote-·s genome haphazardly (llu d <II. 1992: Iycng<lre! <II. 1996: TW},1ll<1J12005). Individualsexpre-ssing the-de-sired phenotypic tnlit arc crossed with wild-type fish. The transgenic offspring ,Ire then cross~"d with non-transgenic conspecitics ovcr l11ultipt.: g\:n~rations. rorming a ~tabk tran~g"l1ic line. Suc("e~srul transgencsis can avoid the lime- and re-sourees involve-d in an nnili("ial seiel,tinn program. Moreover. transgcnesis may allow for the development of traits not allainnble through selective breeding by adding genl's that code for proteins not
pfl"sent within the host genomr (c.g. earbohydratl' ml'tabolism or freezl' resistance:
Fle1l:hn d al. I 994). As <I resuit. a transgenic br('lodstock can be developed in the ab~enee or intentional selection. Thl'rcforc. tnlnsgl'nie broodstocks may be more similnr g ... n ... ti!,;ally (i ... with exl"t~pti{)n or the transge-ne(s)} 10 the-ir "ild source populations than arc 1:1rmed broodstocks. that have been selected for prodUdion traits ror g ... rwrations This may <liS!) reJu,·e the negativc litness nlT1sequcnl"t~s for transgenic aninwls in the wild. The iitnl'ssconsequenccs. howevl'f.will depend on 11('1\\ thctransgene(sl inh:ral"ls lI"ith the orgallism·s c~isting genctic I1r'CilitCClUre. with which it has not l'(wv()ived
13: Phcllot),pifcxprcssion
Phenotypic expression among transgenic organisms hlls been shown 10 vary by integrmion position, copy number, construe\. strain. and species (Twyman 2005: Gong et al. 2()07: N<lT11 et al. 20(7). Rcscarelli.'rs dcwloping transgenil' organisms haw no\cd substantial phenotypic dilTerences between individuals successfully integrating the' transgene within the same' populmion. These inner diflercllees are due to epistasis resulting from molecular level \'ariation during the integration process. The genomic location of tmllsgene il1te'grmion during initial inse'rtion is the major eaus\.' of this variation: known as position/integration eflects (Iycngar ct al. 1996: Twyman 2(05) I'.ssentially, epistatic interaclions bclwe ... n th ... transg ... n ... allli n ... ighhouring genes aff ... cttlw aclivityofthe local molecular region. which mily influence th<.' plwnotype. Anothcreause of molecular level varimion isthe l1umberoftransgene eopics that integrate into tile host
genome; known as dosage effects (Twyman 2005). ("O)1y number Illil)' arkct tl1<.' ~nn{)lInt of protein produced by the lransgenc loci and. consequently. thc overall phcnotypc i"lwsc sourccs ofphen01Ypic V(lri(llion arc dillicultto predict (/ priori and may haw Illncss consequence's. Continued efi(,rts to dcvelop or adaptillore predidabl.:. em~icT1t and practical gene transfer methods for fish and shclilish sp,"ck~ ~ould he lIn assct I,'r the aquantllure industry and the riskas~essmrnt process (i'Jam etal. 2007)
Genetic recombination may renrrange transgrnes or their IOGltioll. as lI"ith any endogenous gene seljuelKc ovcr time. In somt:' "ountri~~. I(:da<ll kgislalinn r('quirrs till"
drll\ollstration of genetic sl~bility in a trnnsgenie strain for severn I generations (("EPA 1999: USFDA 2009). Therefore, genclic slability (lilhc lransgeni~ loci <md the t:lrg,"led phenotyp~ will ne~d to he maintain~d for sn'rral g('JI<.'ratiolls to cOI1lI11C"rcialil'e
aljuacultur~ b;okehnologi~s (Yaskow;ak d al. 2006). However. a fl'w st~ble-gelh.'r:ltions does not preclude reeombinmion m the transgene loei in future generations. The eflect 01 recombination on the location and struclllre of the transgene loci cannot be predicted nor can the re-sulting phcnotype or effect on fitness. I'unhermorc. tnlllsgenes arc invariably designed to behave as gcnes of major cfleet: genes that inlluence the phenotype more so than most genes. Thert."fore. recombination m the transgene loci may rc~ult in a greater il1nll~lIee on fitne-ss-related traits than rceombinmion at 1110st other loci. There is some evidence of transgene instability among some populiltion~ of mud loach (Misl!.III"IIIIS mi;o!cpis) and carp (Labw rollilr/: Num et ;11. 1999: Kim l"t al. 2004: V~lIug(lpalct al.
2004). Thus. position eflects caused by gelll:tic rccomhillatioll may ehang~ th~ phenotYPl·
of a transgenic linc between generntiolls. This is complicatnl further \\"h~n II·c" c()n~ider the dleet of the background genotype on the pllellotypic en"ccts indllc~d hy till" tran~gcnc
The phenotypic response to a stable transgene construct <.:all vary nlll~i{krahl) Wi1hill and hetwn~n popUlations and sr~ci~s. Asidt." from dilTcrenccs caused by construct dt:~ign. plci<1lropy induced by a particulartrnnsgene is aflected by the composition nL and interaction Ilnl(lIlg the gl"neS (If the r~ceiving animal. The- dilTcrential resp.)USC among specic"s to a spccifi~d trallsg~nc" con~truCl is well documented (Nanl et al. 2008). There arc also examples of this phenomenon 1Imong dilferenl p"puiatiolls within a sp~<.:i~s. FOI example. Devlin el III. (2001) found th~t th~ growth r~spons~ of wild and i:lrIlled rain!>..-lw trout (OlicorhYllciw., IIn·ki.\·s) populations 10 transg.enc (OnM1"(1I11) intrnJul"lioll din,:r~d Th~ growth respons," ,,1' till" wilJ strain was far grcatcr th~n thm of the lill"llled slnlin whidl had b~~n sekcted lor rapid growth over sevenll g~nenllion~. Th~ non-tran~g~nic limn strllill. howe\'er. Olltgrew 1he trlmsgenic lI"ild strain. These results might have heen
inllueneed by position or dosllge elTects, Illlllever. ~ recent study by ;>.Jen:gard ct ,Ii (2008a) using GH implants to eompure the growth respon~es of two wild and {)Ill" f:lrmed strain of' Atlantic salmon (511//110 Milar) found simihlr results Crable 1.2). 'I'his conlirmed that genetic background can be
, I
key tactor in the degrceofresponsc tosurplement<ltion, whether through transgenesisor impiantationIn ~u!llInary, intentional srlection ill aquaculture broodstocks Icads to genetic homogeneity hrought <lbout hy tht" usc "ffcw strains alld sUhsrqurnt parallrl ('\olutioll.
Moreowr. intt"ntional sclection :llIows gcnc~, ~nd consrquently. phCllOtypic traits (() co- evolve over lime. Conwrsely. phenotypesofdifferenttransgcnic lincscan varydul' to tr:msgenc position and/or dOSllgC effrets or dim~rcnces in the gcnctic background of the p;]rcnt strains. This potcntial lor a highcr dcgrce of dissimilarity suggests thut the evolutionary pressure on wild populmions from intcrbreeding with tr,msgenic animals may inducc a greater array of pleiotropic dkcts thull th~1 nhs,:rwd from in((.orhrecding with thrmed animals, Hc\\\'c"cr. Ol'cr time. tr.l!lsg~llic S1r~ins Ill.,y l°),.periencc intt"lltinnai sclcction such thm litncss rcdllctionsare simil:lrto farrm:d strains. Thus. aticast in the ahscncenfintentionai selection, the ccologi<.:al and gcnctic impacts oftmnsgcnic anillluis may be more difficult to predict than those ciluscd hy anirlwls produced through traditional intentionni selection.
ro illustwte the above. I shall compare what is kno\\1l ahout potclltial ecological and genetic cffccts c<lllscd hy aqual"ulturlocso.:apccs originating n'om trnditional breeding program~ with those originating from trnnsgcnir manipulation. Despite the dil'crsit~ ot spccics uscd in aquaculture. the tbcus will he Oil salmonid fishe~ heG!u~e "f an untllrlunah.o pUlicity of data addrcssing ccological ;]nd gCIlClir effects of nthlT srecies
Moreover. snlrnonids arc ol1e of fell t,IXOIl II here fitness-rel'lted consequl-nee~ of transgenesishavcheeninl'('stigalCdandar('alsothefocusofthisdiSS(-rtntion
1.4: I)omcslkalioll s('lcclioli and di\'crgcncc
Aquaculture strains arc genetically and phenotypically distinct from their source wild popUlations due to the process of domestication s..-iection (e.g .. lJller ;lnd Epifanio 2002: Ferguson ct al. 2(07). iJonll'stication selection rcf('rs to the dim-rcnt forces allh:ting g('ndic change in l:nptive-renred versus wild populations. This genetic chmlgc
may oc,cur for a rwmhl.'r of diR'ct and indir('ct rcasons. Direct genetic ch;mge or intl'ntional dOllll-stil:ation selection. rders to sl'kctive breeding lor desired Imits. such as
those targeted in traditional aquaculturc practices. Gl'ni.'tmnsfer biotechnologies arc also a direct Illethod of gl'ndic change>: howevcr. as previously described. Ihey Illay ditl"cr in
fundal11entalways
Domcslicmion selection C;1Il also innuence bOlh farl11i.>d and Iransgl'nic animals tllrOllgh indirect gl-Ill'lic change. ThaI is. rearing a population in captil'ity II ith 110 guided sckclion can lead 10 divergencc from the source population. Unintcntion;11 selection may l11anilest itsclfthrough 1\10. most ollcn concurrent. foutes, First. ch;mgi.-s may result fmm
gcnctic drill duc 10;111 irmdvcrtent sampling bias in thl' Iliid founder population of a eaptil'e-bred line: known as n founderell"cCI (Frankham d al. 2002: Allendorfand Luil-.art 20(17). Second. changes Illa) result from abiotic and biotic dill"crences hl.'tlleen Iliid ;lIld
eaptill' rearing enl'ironl11l'nl~ (Price 1999: EinUIll and Fkming 2001: Ilunlingl,'rd 200~) The result is the polential for indirect selection of traits that inneaS(' litness in 111l' captin' eT1vimnnK'nt (I:ncomio l'1 al. 2005: Shoemaker el al 20(6) or. COIl\'er~el~. rel;J.\ed
10
s~l~dion on Irait~ Ihat d~l:rcas~ fltn~~s in th~ wild (Fleming and (irnss 19119: 199.,' Wnples 1(99). Unimelllional selcction is hypothcsizcd to eorrelme with thc numOcr of gcnermions in caplivil)' (Araki CI al. 2007: Caroftino ct al. 2008). So long as lhc cnvirOllmCIll is hcid conslnn1. lilrillcd and transgcnic pOPllbtions should eXI>cricncc simiiaruililltcntional SCICClivc prcssurcs: unlcssthcrc is SOIllC sort ofuniquc imemtlion octwccnthctr:lIlsgcnc:lIldthccnvironlllcnt.
Thc cflects of domestication sclcction on Ihe gcnctic and phcnotypic chanldcristics ofaquacu!turc animnls lead to various l)(ltential envirunllwntal impacts upon rl·l~as~ into natur~. Figur~ 1.1 summari/.~s tlw mC<.:hanisms r~-,ponsihk for su<.:h impacts within tour <.:ategori~s: dir~ct ecological ~1Tects. indirect ecologi<.:al em'tlS, dirl'ct geneti<.: dT~cts and indirect genetic efTeets. Most of these potentinl impads dcp.:nd un whdher the cultured animals arc ~nl<'ril1g hahitat occupied hy populatiolls with whil"ll th~y can il1t~rhreed. As such. I shall discuss the efreds in nOll-nmive (cxotic) and Ilative habilatssepamlcly
1.5: Efrn'tsinl\ou-nalinJ-lahit:lts
The p!.ltenti:ll genetic effech of tiJrmed ;md lransgenic :lI1imals ar~ unlikd) t"
diner fundamentally \\h~n invading non-native habitats that lack genetically compatibk l1l'tl'rospeciflc populmiol1s. This is Ix'cause the comple.\ genctie cffects associilled with species inlerhreedil1g and genes intr()gres~ing ar~ ahsent. Interhre~ding and intn>grl·ssi('n
~rc Ilot synonymous. Interbrceding rctCrs only 10 the ;tet of se)(uill reprodu<:tion bellleen t\\O discrct~ pOPlilations, while intr(lgr~ssi()n rl·1~r~ to lh~ suc<.:essful trans1, .. r of g~nl's from OTIC gene pool to another by intcrbreeding (Fral1kh~11ll el ~ll 2002: Allendorf ami
Luikan 2007). We recognize thm e.~olie habilats may contain closely relaled specks II ilh which an invading populmion can introgress (revielled in Allendorf and Le,lr~ 2001: Allendorf and Luikan 2007): however hybridization usually occurs al 10\1 niles and results in i,1lCnile offspring or olfsprin),\ wilh ),\re<ltly reduced reprodm:tin' ~ucn·ss.
Should hybridization resul1 in Ihe introgression ofa Ir,lI1sgcn(' il1lo <I nalive spn,ies. Ihel1 the dTeels arc likcly 10 become similar 10 those Ihal occur in nalive habilals (sec beiOII) I·ll-rl·. hOllcv(·r.1 focus on elTeels inthcabsl'nceofintrogrcssion.
Till' dTeets on ('cology and biodiversilY resulling from species inV:ISil)ns hale b-i.·en reponed for deeadcs (e.g. Levine and I)"Anlonio 1999: Simberlo1l" and Von Holle 1999: Dc Sill'a CI al. 2006). The establishment of an e:l:olic specie;, depends on the frequency <lnd magnitude of the invasion. Ihe relalive lilness of 11K' coloniLer and Ih..:
vulnembililY oflhe ecosysiCl11 (Ruesink 2005; Olden ct al. 2006). An aquatic species Ihal is fanned extensively is likely 10 hal'c ample opponunity (Illd sufticiell\ numbers to invade hosl eeosyslems due 10 the polen1ial for reCllrring esc<lpe ev..:nts <lnd till' scale of f:mning (e.g. Fleming el al. 2000. Bckkevold el al. 2006: Thorswd et al. 200S). Similarl).
mostaqu!lcliliured species arl" reared in ('lwironm('I1IS where IheirabililY 1010icrale 10cll1 abiotic Ihctors is lICit understood. Casal (200b) rqlon('d a lis\ of [Iw lOp dgill..:en invasive finfish spccics repon ... d 10 hal'c ncgalivc elTecls Oil IOC~11 eC(ls~slems. (If\\hich Ihine..:n(72%) hu\'eocellIIscdill ilqlmClll1l1re
Many of the l"cologic;ti cllcds ofeseap<.'d aqu:lclIl1urc aninwls arc COIIIIlWIl \<) bUlh norHwtiv(' and nalive Iwbilals (sec bclo\l\. 110\I..:ver. therl' arc also 1II;1II~
eeOsyslel1l~ic\"C1 abiolie and biolic indireCl efkcls thai arc mme likel) Il' N'Cllf in n"n- naliV<' habilals. Indireel cll-':e\s rc1cr \0 changes Illedialed lhrough 11lhird wlII]lnl1ern
12
(abiotic or biotic) of the (.'eosystem. Sm;h indirect mcchanisms indudl' hnbitat (llodificmion. apparent lacilit:lIion. apparelll competition and the lrophil: cascnde cl"kct (Goldschmidt CI al. 1993: Shurin e\ al. 2002: While el al. 2006). Indirect effects arc more likcly in exolic Iwbitals due 10 Ihe p01elllial tor exotic species [(l illlcrael Ililh Ihe environm(.'nl in a nO\'el manne'rand.thus. inlluencecommunilyslruelure and funclion.
The genelic impaels ofaquaeuilure animals eillcring norHlalil'e Iwhitatsean be mediated indirectly through ecological interactions. Indirect genetic enccts refer In changes in local hcterospecilic populations. Such interaclions may l"Il.mgl' thl' sdeclivc pr(.'ssures eXjK"rienced by wild poputnliol1s. r(.'suiling in g(.'nolypie and phl'flOtypk ,Ldaptations (Waples 1(9). In COnlrasl. such ehang('s may mnnif(,'stlhcmsehes through neg:ltivc inter,lCtions r('sulting in a reduction in population siJ;(.'and geLwti('divcrsity.
1.6: ":ffo.,{·tswithin Native H:lhitals 1,6.1: Competitioll
In terms of risk assessnll'nl. the effects of farmed and transgl'nil' aninw"
inleracting II ith wild conspecitic or IK'tcrosp(.'eifie populmions with which the) can inlerbrced is likely Ihe 1110st critil:al scenario. This is becnuse Ibe con~equenees nfdirect genclic meehnnisms. such as an inl'asion oftarmed thh gene complexes or transgl'IIl'S.
arc eXlccptionall) ditlil"llh In predict (lx'vlin ct al. 200n. 2007: Hindar CI ,II. 2006:
K:Lpuscinski ctul. 2007)
Ecological cffccts call b(' l'ausl,d by dirl.'l·1 111('chanislll~. such a~ dbea~e. prcdali(ltl.
and intcrtCrcnce competition. or indireci llleclwnisllls. such as c~ploilatile competition (Weir and Granl 200S). In sal(llonid~. comp'."lilion b\.'tll('cn farmed and 1I;ld eon~pcdtks
13
is well described. Salmon ids arc intertCrence competitors both as jUI'eniles. when competing for 1()raging territories. ,md <IS adults. when eOl1lpcling for bn:cding opportunitics. I.<lh-b<lst"u stuuit"~ gt"naaliy rt"port innt"ast"djuvt"nile aggrl'ssion and pOOf rt"productive brhaviours in farmed individuals (Eirlllm and Fleming 2001: Fkming and I'ctrrsson 2001: Weir and Grant 2005). Thesc distinct pallerns ofcomprtition indicate substantial resourcr oVl'rlap Iheretore. ecological cne-cts may depend on their rl'spectiVl' drnsitirs. their rrlative competitive abilities and thr carrying capacity 01" the ecosystem (Webcr and Fausch 2003). However. it should be noted that genc Ir:ms1i:r lllodiliC(ltions. such as. :m addition of a gcnc surrorting carhuhydrate mclailolisrn (I'itkarwn cl al. 19l)i)) could allt.-cl prey choice and large scale foraging pallcrns. In laet.
therr is some evidence of snwll scale ditlerrnees bell\een GH transgenic <lnd non·
transgenic coho salmon loraging pallrrns (Sundstrom et al. 2007).
1.6.1: Interbreeding ulld inlmgre.nimr
rile dirrl"rence~ ubserved in the competitivr ability of fanned and wild salmonids suggest diftCrences in survivlllllnd rcprodlKtivc sm:ccss. Despitl' 1011 a survival 01 l~rmed juvenilrs. comptlitivr displacement of wild individuals has been observed in Atlantic salmon. indicating potentially negativc ecological cftCcls (Fleming cl al. 2000;
McGinnity 1997: [l...h:Ginni!y ,:t al. 20(n). Iloweva. adult farllll'J ~trains show poor rates
"I' r~lurn to th~ ~pawning groullJ~ (McGinnit) ,'1:11. 2003) and di1li:rences in spnwning
belwviOllr lhat correlale cxpe~tcdly wilh reduced rerrodllctivc ~Ul"l"e~s (I.'lerning l·t al.
1996.2000: Weir et al. 2004). Reduecd n:produetivc 5ucee~s. hOI\'cI<:r. ilppc,lr~ no! In carryover 10 males thm maturl' prn:nciously in fresh watn a~ parr. Iwving never been to
14
th(.'oec<ln. i>.1alcs of this alternative lifchistory tactic. whieh area fraction oftlw size 01 the anadromous males (i.e. males that have bc~'n to sea). may brn'd successfully by 'snc:lking' fertilizations (reviewed by Fleming and Reynolds 2004). Filrmcd nl<lles e);pr(.'~sing this <llternative reproductive phenotypc can compctc sucecssfull) II ith their wilu eOllllterparts for breeding opportunities. leading to equal or ewn superior rcproductivc success (Garant et al. 2003: Weir et al. 2(05). TIlliS, the avail<lhiL' t"vidc'nc(.' trOIll sall110nid lishes demonstrates Ihatl·OI11I)(.'tition can r(.'sult in till' rcdllcl'd titl1l'SS of wild strains. Mon:owr. farm(.'d individuals hav(.' poor lifetimc reproductive success in the wild. but can contribute to suhsequent g(.'n(.'rations and maytl1l'refore influence tile fitness of wild populations (Fleming o:t ~1. 20UO: McGinnity d al. 2003: Hilldar et al. 2006).
rill' interbrt"eding of farm(.'d and wild poplilations gives rise to concerns abollttllc potential negative rffeets of altering wild gcne pools vin introgres~ion. Such c()nc(.'rn~ arc' b:lsed on the evidence suggesting Ihat genetic ~Illl phenotypic diffrrentiation bCh,cen Sillm()niJ populations has ~{bpti\'e ,ignifieaIICl' (Garcia de Leaniz et al. 2007: Carlsun and Seamons 20()8) <lllli thatthi, would h(.' threatened hy introgression.
Captively reared populations llsually haw IOll'crgeneticdivcrs ity hec<lus(.' they ar(.' ollen closcd and havc reduced c1fec1ivc popubtion si/(.'~ rdati\'(.' tll thos(.' in till' ",ild (i'rankhaI1l2()OX). Thispal1rrn has becll obscrved rcpcalcdly inaquacuitur(.' hrnmlstocks (f';>;adnctylo~ ct al. 1999: Skanla el al. 2005 Frost ct al. 2006). Signiticant one-way gene 110\' due to eseap(.'es could shiti the genetic composiliol1 of the wild populations to'lanls that of the cultured broodstod (Fleming et al. 2000: .\1cGillnity ct al. :W03: Ilind'lr c't al 2006). Subsequent n:dll(:lions in tlwg(.'nctir diwrsity of\\ild populations 'Iould make til(.'m nwr~ vulnnahi<' to (.'lIvirOIlJl1<'l11al change and. in e~trellle cases. could lead to
"
extinction (Frunkh:Ull et :11. 2002: Allendorf und Luikan 2007: Cl1rtson l1nd Seal11on~
200S)
Species where there i~ 10\\ gene flo\\ between populmions and l1 high degree of local adaptmion. such as sall11onids.are panicularly vulncrable 10 olllbrc('ding depression.
Outbreeding depression refers to combining allcles from dill"crent populations adaptcd to difl"crcllt environlllental conditions. re_~ulting in the reduced ntllCs~ of tlK' hyhrid populatioll (Wolfet al. 2000: Fmnkharn ("I al. 2002: Allcndorfand Luikan 20(7). Further harmful dli...:ts llIay occur if the illlerbreeding populations disrupt cll-adapted gl·l1l' complcxcs upon recombination in subseqllent generations. Co-adapled gl'nl' eornpk ... s arc sets of loci that undergo litness-rclated epistatic inll"ractions (Wolfet :11. 2000:
Frnnkhmn et al. 2002: Alicndorfand Luikart 2(07). Conscquelllly. interbreeding lh:1\\i.'en a \\ild und a captivc-rcun.:d popui(ltion lIluy result in outbreeding depn.:ssion in the hybrid progeny and the breakdown of eo-adaptl'd genl' complexes in subsequent gelll'rations.
rhere is fairly consistent empirical {'vidence of out breeding depression cau~ed from the interbreeding of II ild and limned or wild and non-local salmonid populations (reviewed in Fergusonelal.l007:Garciadel.e:mi/("tal.l007: Fruserl0()S)
1.7: Case stud~' ofsalll1ollid j!rowth enhanccment
rhc goal of this thesis is 10 provide empirical inlornwtilll1 Oil the potential environmental and genetic elYects of gro\\th hormone (Gil) tran~genic Atlanlie ,,,lm(\I1
(S"IIII" .w/ar) enlry into the wild. rhe preceding disulssi,'n Sl'i Ihl· en,-e\'''llIti"nar~
l·onle.~1 "ilhin "hich IIll' issui.' of "quaeuitllr(' biolel'lU1ologic, 111:1) lx' aS~i.'s,cd. Till"
16
1011011 ing discussion provides inform(llion on Ihe Sial(' of current ('mpirical knowkdge and a COmn1enl<lry on till" lilerailire ri:l<lled directly loth<: focusoflhisdiss('rtalion.
Till" salmonid growth-enhancem<:111 litcrauU"l.' consists of studies inlc~ligaling lhe
filness-rebted traits of groll"th-scll'cled :lquaculture (f~lrIncd) fish. grolllh hormone (Gil) tr<lnsgenic fish and fish adminislered exogenous GIl. Thi.' laner group has be~'n utilized e:»lensiVl'ly in lhe lasl 15 years as a pro:»)' for lransgenic individuals. TypicallY.lhis melhod relics on a cominuous. sloll-releasc bovine Gil formulalion (posilaCijl1: Mon';lnlO Company: S1. Louis, USA) thm is impl;mted inlO lhc pcrilOlll"al cavily (McLean el al.
19(7). Thc ap~al of lhis subslitutc is that it allows field comparison~ of trealed :1I1d unlre<lled wild fish: an option n01 available lor lransgenic animals. Field experiments
<111011' for the' complc .... il)' of n<llurc. whic.h cannot be' mimicked fully in laboratory cnvironmenls. Furthermore. llll" lise of wild fish climinates lhc poten1iall) cont()IInding effects of the caplive re(lring elll'ironmenl OJ] phenOlypic development. Mokcular-il-wl complil·alions associated wilh lhe dCl'elopmcnt oftrnnsgenic lines. such as position or dOS:lge dlCcts. also need not ocacollccrn.
rhe obvious limitalion ofe:>.ogenous Gil adminislralion is thaI il i~ not a complete ph)siological <:quivail-nt of lh~· i.'udogenous Gil proollelion illduced b) :1 lran~gelll"
Sciemilic lillderst<lnding ol"lhe eflec1s oflhe grow1h hormone-insulin-lih· gro\\lh 1~lclor I ,yslt'm (GII-IGF-I) on St'vo:ral <lS~t·1S of sahnonid physiology TCmains uncertain (l3jorn~~on 1997: BjornSSOIl el (II. 2002). Moreover. 11K' endocrinologic;ll cfl~l"I~ (,f a Gil implanl compared 10 a lr<ll1sg('nC· nre nOI knOI\TI. This 1l1a~' be P:lrticillarl) ~ignit1c;1I11 II11l'1l 11(' t·onsider 11K' inhcrelll eomplexilY ofbiologkal c)cks. Speci1i(ally. Ihe ell~l"IS ofseasol1(l1ily nnd nge Oil <.ill-IOI'-1 indul·ed phl'IlOl)pil' variatioll Wilh lhis C:1Ie:ll in
17
mind. til\: 1:11( .. cls of c:-.:ogenous GH administration ,lppear t<J stimulate similar ph .... nolypic dwngt's as lhm sCt'n wilh tr;lJlsgt'nt'~is (Tab It's 1.1 and 1.2). This mah,s it a uSl,ful tool when studying life history periods where we e:-.:peet animals arc under tlw inllul'n.,;c of sustained GH production. such as thaI ofjllvenile salmonids in the spring and sllmnwr (11j{irnsson 1997: Fleming el al. 2002). Therefore. I shall consider exogenous GH administration as analogous 10 GH lransgellesis for the purposes of this chapter and rcli:r to thelll collectivclyas GH-enhalleed I1shes.
1.7.1: Ctllllpurill!; !;mwth-.\·e/ectl~J ulld Gil-eli/lanced plwllotypes
A range of similar experiml,nts have been performed 011 growth sl'lcl:lcd (fanned) ,md GI-I-enhuneed (GH transgenic and GH treated) salmonid tish .... s. Overall. the difT,'r .... nees obs .... rv.:d lor ti1l1 .... ss-related traits offilrmed and GH-enhaneed tish .... s r.:l:l1iv.:
to wild-typ .... individll;lls ur .... quit .... similar in hutcIH:ry-typ .... environments (Tabk 1.1) rhese d .. ta generally indie .. te in<:r .... as .... d growth p01 .... nlial and
I, ...
ding Ilwtiv'ltion.reductions in anlipredator behaviour and differences in various physiologicall'orrl"iak~
rl"iativt' to wild salillonids
Th .... se sirnilarpatterns may correspond toequivalcnl proc .... ss .... s intllleneing lllnllcd and (;II-.... nlwnct'd animals. A principal pht'notypit changt' [t'suiting from (JlI transgenesis or exogenous treatmem is. unequivocally. incr .... as .... d GH production rhl' principul phenotype tnrg .... t .... d in s(limonid nqll(lcuitlirc is growth (!I1d l~lrmcd s<llilloilids h:1\ ~ hcen as~()(;ia1CJ with all inlTea .... e in l'iruil<lling growth IWrIllUlli.' during til<" ju\enilc' growth phase (jfth~ir lik history (rtc:ming .... t al. 2002: Devlin et al. 200')). rhus. similar
18
pluripotent .:11ects on fitness-related traits ma~ be intima!ely aSSO(i,lh:d with Ill( (hange~
in (ndCK:rine growth regulation inbOlh farmed and Gil-enhanced salmon ids.
Upon (omparison of larmed ,md Glt-enhan,,:d individuals rel3tiw to wild individuals in more complex cnvironments. vnriabilit)' begins \0 emerge Crable 1.2). In natural and ncar-natural cxperimentul streams. cvidcnn~ of increased growth and decre<l~ed <lntipredalqr hehaviour is (OnSislenl. Ilowevcr. Ilw dirl'C1ion offitness-rclatcd Ira its such ns survivn!. rl'productive su<:cess. energy USl' and competitive bclwviollr is inconsistent. Trait divcrgence between limned :md GH-enhiHKed individuals n:I,ltive to wild individuals may relket genetic andlor environmental dit1"crenn·s_ For (<lmparison among studies. it is imponant to identify and. where applicable. wn\rol for stl~h sotlr(·es of trail differentiation. Otherwise it isdiflkultto intcrwhethcrageneti(predisposition is responsible for Irail dit1cn:IKes or. <litern<ltively. if differences rl,lkCl a plastic responsc to unique l'nvironmcl1ts. This is 110t always casy to accomplish beelluse. for e~alllple.
rcsearcll ellnnol be (ondueted with transgenic org(miSIllS inlhe wild. Thereillfe. unlike thrmed strains. (omparing fitness-related traits hc!ween wild-reared tr<lnsgcnil' and 11011- tr<lnsgeni( straillS m<l)' not he possible. In one study. Besscy e\ al. (2004) slwII"ed similar paHcrns of reproductive trait divergence between GH transgenic (olw salmon and wild individuals as has bc·en observed b~t\\'rrn fimnrd and wild individtwls: suggesting that for thesc traits. rellring history Ill(l), h,· a morc (ritiGll !;Ktor Ihan tran~genesis (·Iable 1.2).
When (jll-cnhanced individuals arc compared to wild-type indil'iduals with the S.11llC background genotype and rearing history. under nil1ural or ncar-natural bbor<lh)r) (ondilions, l1l(lllY litness-related tr~ils appear unallc·(ll·d h) treatnwilt (Iahk 1.2). For e.\:1I11pic. natural br<'cding c.\perilllell1sshow no dilferelKcs ill counship ~ha\'i(\lIrand
19
reproductiw SllC[<~SS hell\ecn (jll-enhallc<,d and -uncnhaneed individuals when they share the saillc background genotype and rcilring history (Ile~scy cl al. 2004: Sundt- Hansen 2008). Similarly. studies measuring traits injuvcnilc parr{> 2g) undcrtllc threat of predation consistently find in!;reases in growth. hUI no diffcn:m:cs ill survival bl,tween GH-enhanced and wild individlWls (Johnsson et al. 1999: Julm%llll cl al. 2000: Johnsson and BjOrnsson 2001; Sundstrom ct al. 2007: Sundt-Ilansen 2008. SlIndt-llanscll el ;t!
2009)
In tenllsoffitncss.many slwJicsonjllvcnilelishfail 10 CaplUfl'a signifil'alll
period ofcllrly life hislory: the onset ofexogenolls keding. The transition to exogenous lOod resources is a critical period of survival for young fry. where individuals eonl"ron1 an environmcnt saturated with compctitors and pred:Hors: feSlllting in high rates ofll1<lrl;t!ily (Elliol 1994: EinulII and Fleming 2000: Kennedy el al. 2(08). Few sludi,:s haw nlr<lsured lillless-related traits in lanned and GI·j-enhanced fry « 2 g). In m:ar-natural laboratory conditions. Sundstriim et al. (2004: 20(5) showcd Ihal (jlltransgcnic coho fry experiencc increased predator-induced lllorialilY. rrdueed dispersal and equal growth lillder moderate to high ked levels rdatiw to wild-Iy~ fry. 11,,\\cwr. dcrrcasnl gn\\lth was (lhservcd und<·r 1<11\ Itcd lewis Converscty. a reecnt mark-recapture study. I\here UH-implanted Allanlic snllllon fry were released into ~ natural st~am found llO diftcrellces in ~lIr\'iv(!I or di~pcr~aL and reduccd growth of (jll-implanttd felalil'e t() control fry (Sllndt-Jlan~cll et al. 200S). II is lInelcar I,hether these inconsistencies relk,,1 phcnotypic differences resuhing Ii-,ml intrinsic I'ariati,)(l hctll<'('n sp(" .. ks, .. nil(llleelllent method., (i.c. tfansg"lIesis versus implalllntionj or expcr'illlental ,'l1vinmnwnts (e.g relmive levels of pre dati un and l'Olllpl'litiun). II is. howel"."r, likcly a combination ofslIch
20
factors. Nonetheless. II hen genetk background and I"CHing environment ,Irl' controlled.
differences in tltncss-rel<tted trail~ <lppe<lr Ic~s likel) between (ill-nlhancl,d ;md ulll"nhanced individuals eomp<lrl,d undl'r ntar-natuml conditions. Furthcrmon.::. thi."re is Sl)me evidt'net to suggest that high levels of prcdation and foraging comp-ctition can ri.'duee survival and growth in GII-cnhanced individuals at the critical. lirst-tCedingst;lge
The above discussion highlights a principal reason why tnmsgenic lmimals m;IY hav(' a lower elcmcnt of predictability than farmcd 1lI1imais with respect to risk assessmcnt. Farmcd individuals have c.~pcricnced gcnermionsofintentionli1 selection for production-rdev<tnt traits that are likely controlled by many gi.'IK'S of sillall l"Ili."ct. This results in genetic divergent(' from wild populations. Evidence of eonsistcnt. negative fitness eonsequcnces have been demonstrated bl'tween !'<trmed ;md wild salnHl11 populations. Due to loll' stnlin v;lri<ttion ,md simii;n sel(,ttivc pri.'sslIr('s among aquaculture brood stocks. variability in the conSt4ucnces of recombination is prinfipally caused by the responSl' of the receiving popul<tti01] h) the brc<lkdown of nHldaph.:d polygenic complc.\e~.
In contrast to aquaculturt strains. transgenesis GI11 induce growth in lishi.'~ "I' ull~t'lected and diwrse genetic b:lckgrounds under controlled el1virnnmel11s. As genl's "I major cnccl. transgtlK'S ('an have a gre<t1<'r intlutncl" 011 plwnot) pic ('xprcssion than most other genes. WIll'n growth-enhanced lish :lfC compared to II ild-type individu<tls 01 ,imilar genetic background. trait dillcfCnces become evidentllnder "Ibmat,'r) condition,.
Iltllle\'l"r. under natural conditions. litni.'ss-rl"lated trait dilli.'ri.'nces arc variable. Ileah or none.\istent: indicative of gene b) environment interartions (Devlin et al. 200 .. t: 2()06:
2007: SundstrOm ("\ ill. 2007b). Evidentl:-. it appears the badground gcn"m~ C;1Il
21
