Invasion genetics of vendace (Coregonus albula (L.)) in the Inari-Pasvik watercourse : revealing the origin and expansion pattern of a rapid colonization event

In asion gene ics o endace (Co egonus albula (L.)) in he Ina i-Pas ik wa e cou se: e ealing he o igin and expansion pa e n o a apid coloniza ion e en Kim P æbel 1 , Ka l Øys ein Gjelland 1 , E no Salonen 2 & Pe -A ne Amundsen 1 1 Depa men o A c ic and Ma ine Biology, Facul y o Biosciences, Fishe ies and Economics, Uni e si y o T omsø, N-9037 T omsø, No way 2 Finnish Game and Fishe ies Resea ch Ins i u e, Ina i Fishe ies Resea ch and Aquacul u e, Saa ikosken ie 8, FI-99870 Ina i, Finland Keywo ds Biological in asions, e ec i e popula ion size, exo ic species, gene ic bo leneck, apid di e gence. Co espondence Kim P æbel, Depa men o A c ic and Ma ine Biology, Facul y o Biosciences, Fishe ies and Economics, Uni e si y o T omsø, N-9037 T omsø, No way. Tel: +47 776 46107; Fax: +47 776 46020; E-mail: [email p o ec ed] P esen add ess Kim P æbel, Cen e o Ecological and E olu iona y Syn hesis (CEES), Depa men o Biology, Uni e si y o Oslo, P. O. Box 1066, Blinde n, N-0315 Oslo, No way Funding In o ma ion The s udy was suppo ed by he No wegian Resea ch Council (NFR 183984/S30). Recei ed: 28 No embe 2012; Re ised: 1 Ma ch 2013; Accep ed: 9 Ma ch 2013 Ecology and E olu ion 2013; 3(5): 1400– 1412 doi: 10.1002/ece3.552 Abs ac Species in asions can ha e wide- anging biological and socio-economic e ec s and a e gene ally unwan ed by legisla ion. Iden i ica ion o he sou ce popula- ion as well as he ecology and gene ics o bo h he in ade popula ion and he ecei ing communi y is o c ucial impo ance. The apid in asion o a small co egonid ish endace (Co egonus albula) in a majo no he n Eu opean sub- a c ic wa e cou se has esul ed in a labile ecological si ua ion in he ecei ing communi y. The ecological impac o he in asion has been ho oughly docu- men ed, bu he gene ics o he in asion emains o be explo ed. We analyzed he gene ic di e si y and di e gence pa e ns among he wo possible sou ce popula ions om sou he n Finnish Lapland and h ee colonis s popula ions wi hin he Ina i-Pas ik wa e cou se using en mic osa elli e loci in o de o (i) iden i y he mos likely sou ce o he in asion, (ii) e eal he dispe sal pa e n and gene ic s uc u e o he seconda y expansion, and (iii) o in es iga e whe he he ini ial in oduc ion and he seconda y expansion we e associa ed wi h ounde e ec s. We e ealed ha epea ed ansloca ion o endace om Lake Sine € aj€ a i in o a ibu a y lake o L. Ina i in 1964–1966 is he mos plau- sible sou ce o he in asion. Bo h he ini ial in oduc ion and he seconda y expansion we e ound no o be associa ed wi h signi ican ounde e ec s. The seconda y expansion ollowed a s epping s one pa e n and he sou ce and colo- nis popula ions o his expansion ha e unde gone apid gene ic di e gence wi hin a pe iod o 15–35 yea s (ca. 8–17 gene a ions). The apid di e gence may be con ibu ed o lack o gene low among he sou ce and colonis popula- ions due o he ex ensi e hyd oelec ic damming in he wa e cou se. Mul iple in oduc ions and subs an ial gene ic a ia ion in combina ion wi h he boom- and-bus popula ion de elopmen o he species hus likely coun e ac ed he ounde e ec s as well as ueled he apid es ablishmen and expansion o his species wi hin he Ina i-Pas ik wa e cou se. In oduc ion In oduc ions and in asions o exo ic species ep esen a p oblem o global ex en (Vi ousek e al. 1996; William- son 1996; Olden and Rooney 2006). Se e e ecological e ec s may a ise om in asions o possible in e ac ions be ween clima e change and in asions (Sandlund e al. 1999; Mooney and Hobbs 2000; D iscoll e al. 2012), and he consequences may also be associa ed wi h huge economical cos s (Moyle 1986; Mack e al. 2000; Pimen el e al. 2000). Du ing ecen decades, nume ous s udies ha e add essed he ecological e ec s o biological in a- sions, pa icula ly wi h espec o he consequences o he ecei ing communi ies. Nega i e impac s include impe ilmen o na i e species (Allan and Flecke 1993; Pe e son e al. 2004; Bøhn e al. 2008), al e a ions o communi y s uc u e and na u al biodi e si y (William- son 1996; Pa ke e al. 1999), and a wo ldwide bio ic homogeniza ion (Rahel 2000, 2002; Olden and Po 2004; Olden and Rooney 2006). Howe e , species in oduc ions may also be seen as la ge-scale ecological expe imen s, and p o ide unique insigh in o ecological in e ac ions as 1400 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. well as popula ion biology and gene ics (Sakai e al. 2001; Shea and Chesson 2002; Sax e al. 2007; Da is 2009). The gene ic consequences and ou come o an in asion in en o he in ade s and he colonis popula ions, ha is, he co ela ion be ween p opagule p essu e, gene ic a iabili y o he in ade s and he esul ing gene ic a i- abili y in he colonis popula ions, a e no i ial. Colo- nis popula ions, ha is, he popula ions descending om in oduc ion e en s, may apidly de elop di e ences in mo phology and pheno ypes (Facon e al. 2008; Wa d e al. 2012) and in ecology (Sandlund 1992; Sakai e al. 2001; Amundsen e al. 2012) as compa ed o he sou ce popula ion, bu he gene ic componen is o en unknown (e.g., Blanche 2012). Colonis popula ions can gi e ise o seconda y expansions, ha may co e la ge geog aphi- cal scales (Facon e al. 2008; B own and S epien 2009; Reusch e al. 2010). Such demog aphic expansions p opel di e gen adap i e p essu es due o explo a ion o new en i onmen s (Ghalambo e al. 2007), which again may p omo e apid changes in li e his o y o he in ade (Dlu- gosch and Pa ke 2008a; Amundsen e al. 2012; Gu owsky and Fox 2012). While mos s udies ha e ound lowe gene ic a ia ion in he colonis compa ed o he sou ce popula ions (Hen y e al. 2009; Ay es e al. 2010; bu see Dlugosch and Pa ke 2008b o e iew), o he s show no loss o gene ic a ia ion (No ak and Mack 2005; S epien e al. 2005; Wa es e al. 2005). Some s udies e en ind inc eased gene ic a ia ion in he colonis popula ions when hey o igina e om mul iple in oduc ions p omo - ing genomic admix u e (Golub so e al. 1993; Kolbe e al. 2004; La e gne and Molo sky 2007). The ac ion o gene ic a ia ion ans e ed om he sou ce popula ions o he in ade s depends on p opagule p essu e (Lock- wood e al. 2005; Colau i e al. 2006; Simbe lo 2009), dispe sal mode (Hewi 1996; Ib ahim e al. 1996; Wilson e al. 2009), and/o ou e o in asion (Es oup and Guille- maud 2010). The al e ed selec ion p essu e applied when indi iduals explo e new geog aphical anges also ep e- sen s challenges o he in ade (Lee 2002; Sua ez and Tsu sui 2008). Thus, i is he combined e ec o hese ac o s ha de e mine whe he he in ade s a e success ul in hei coloniza ion and p oli e a ion. The p esen s udy conce ns he in oduc ion and subse- quen in asion o an exo ic ish species ( endace, Co egonus albula, Fig. 1) in o he suba c ic Ina i-Pas ik wa e shed (No way, Finland, and Russia). The i s in oduc ion occu ed in 1956, when some endace y we e ansloca ed o s ocking pu poses om Lake Keluj€ a i in cen al Fin- land o he Ina i Ha che y, om whe e hey subsequen ly escaped in o Lake Ina i (Mu enia and Salonen 1992; see also Fig. 2). The second in oduc ion occu ed in 1964– 1966 when endace y we e ansloca ed om Lake Sine € aj€ a i o a small lake in he ca chmen a ea o Lake Ina i. The i s endace we e obse ed in Lake Ina i in 1973. By he ea ly 1980s, a endace popula ion had es ab- lished h oughou he lake (Mu enia and Salonen 1992), inc easing o a peak abundance in 1989 (Salonen 1998, 2004). Du ing his pe iod a downs eam in asion o end- ace appa en ly occu ed o lakes in he Pas ik wa e cou se, he ou le i e om Lake Ina i, whe e he species was obse ed o he i s ime in 1989 (Amundsen e al. 1999). The endace we e obse ed in L. Vagga em o he i s ime in 1991 and in L. Sk ukkebuk a in 1993, and wi hin ew yea s he in ade became an impo an pelagic ish spe- cies in lakes in he Pas ik wa e cou se (Bøhn e al. 2004, 2008). Se e e ecological consequences ha e been docu- men ed o he ecei ing sys em, including changes in biodi e si y and species composi ion, ood web dynamics and ecosys em unc ioning (e.g., Bøhn and Amundsen 1998, 2001; Amundsen e al. 2003). The endace ha e du - ing he in asion en e ed a ypical luc ua ing ‘boom-and- bus ’ de elopmen (Salonen e al. 2007), esul ing in a des abilized aqua ic ecosys em. Al hough documen a ion o he ecological impac o he endace in asion is accumula ing, he gene ics o he in a- sion emains o be add essed. A ecen s udy e ealed ep oduc i e isola ion be ween endace in L. Ina i and a downs eam lake (Amundsen e al. 2012), which ini ia ed his mo e de ailed gene ic s udy o he sys em. Because we hold de ailed empo al and ecological in o ma ion abou he in asion p ocess (see e.g., Mu enia and Salonen 1992; Gjelland e al. 2007; Bøhn e al. 2008) and ha e he special oppo uni y o p ecisely iden i y he sou ce o he in asion, he sys em o e s an excellen model sys em o assess ongoing gene ic and e olu iona y changes. We also expec he downs eam in asion o be highly in luenced by he se en hyd opowe dams ha exis in he wa e cou se (Amundsen e al. 2012). The dams only allow unidi ec- Figu e 1. In ading endace (Co egonus albula) om he Pas ik-Ina i wa e cou se, No way (pho o: K. Ø. Gjelland). K. P æbel e al. In asion Gene ics o Vendace ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 1401 ional downs eam gene low ha would be limi ed o absen du ing no mal i e low condi ions, as he ish la ae o adul s would need o pass h ough dam u bines (see also Amundsen e al. 1999, 2012). Hence, po en ial ac o s ha may limi he success ul coloniza ion in he sys- em may ela e bo h o ounde e ec s and o unidi ec- ional gene low. He e, we used en mic osa elli es o explo e he gene ics o he in oduc ion and subsequen downs eam mig a ion o endace in he wa e cou se by compa ing he wo po en ial sou ce popula ions, he in o- duced popula ion, and wo downs eam colonis popula- ions o igina ing om he seconda y expansion. The main objec i es o he s udy we e o (i) iden i y he mos likely sou ce o he in asion, (ii) e eal he dispe sal pa e n and gene ic s uc u e o he seconda y expansion, and (iii) o in es iga e whe he he ini ial in oduc ion and he second- a y expansion we e associa ed wi h ounde e ec s. Ma e ials and Me hods S udy lakes and sample collec ion To iden i y he mos likely o igin o he in asion, endace we e sampled om he in oduced popula ion in L. Ina i, no he n Finland, and he wo po en ial sou ce popula- ions, L. Sine € aj€ a i and L. Keluj€ a i, cen al Finland (Table 1; Fig. 2). Vendace we e u he mo e sampled om he colonis popula ions in L. Vagga em and Sb (1993) 100 km A c ic Ocean Si Ke A c ic Ci cle In (1973-1982) 1956 1964-66 Vg (1991) (b) (c) Ke Si In Sb Vg Ke Si In Sb Vg 218 222 226 230 234 238 242 246 250 254 258 262 266 270 274 278 282 286 290 166 170 174 176 178 180 182 186 188 190 194 196 198 200 204 216 218 222 224 226 230 181 185 189 193 197 201 213 Ou e ing Inne ing Cla-Te 13 Cla-Te 06 Cocl-la 49 (a) Figu e 2. O igin and sample loca ions (a) and molecula gene ic a ia ion (b, c) o endace in no he n Fennoscandia. In (a) ligh g ay a ows show he di ec ion o d ainage. Also, no e ha hyd opowe dams (indica ed by black lines in he map) a e loca ed be ween he in oduced L. Ina i popula ion (In) and he popula ions wi hin he seconda y expansion (Vg and Sb) limi ing downs eam gene low and uling ou any ups eam gene low. The schema ic illus a ion o he in asion p ocess ollowing a s epping s one (b) and no mal dispe sal pa e n (c) a e illus a ed wi h he co esponding pie cha s o he ela i e allelic equencies o h ee ep esen a i e mic osa elli e loci (ClaTe 13, Cocl-la 49, and ClaTe 06). Dashed lines indica e man-made ansloca ions (L. Sine € aj€ a i and L. Keluj€ a i o L. Ina i) and ull black lines delinea e he na u al seconda y expansion o he in asion. Below a e he colo codes o he alleles pe locus. Sample loca ions a e abb e ia ed acco ding o Table 1. In asion Gene ics o Vendace K. P æbel e al. 1402 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. L. Skukkebuk a in he No wegian pa o he Ina i-Pas ik wa e cou se o in es iga e he gene ics o he seconda y expansion. All samples we e collec ed in he au umn 2008 nea o a he spawning g ounds and consis ed o >92% ipe indi iduals. The e a e se en hyd opowe dam con- s uc ions loca ed below L. Ina i (Fig. 2). Mo e de ailed desc ip ions o L. Ina i and he Pas ik wa e cou se a e gi en by Mu enia and Salonen (1992) and Amundsen e al. (1999), espec i ely. Mic osa elli e DNA ampli ica ion and geno yping Genomic DNA was ex ac ed om gill ilamen s by E-Z96 Tissue DNA Ki (OMEGA Bio- ek, No c oss, GA) ollow- ing he manu ac u e ins uc ions. The indi iduals we e geno yped a 10 mic osa elli e loci (Table S1) a anged in h ee mul iplex polyme ase chain eac ions (PCR) ollow- ing a p e iously desc ibed p o ocol (P æbel e al. in p ess). The PCR p oduc s we e sepa a ed on an ABI 3130 XL Au oma ed Gene ic Analyze (Applied Biosys ems, Fos e Ci y, CA) and alleles sco ed in he GeneMappe 3.7 so - wa e (Applied Biosys ems). A e he i s alida ion o he geno ypes, 3–4% o he indi iduals wi hin each popula ion we e e-ex ac ed and e un a all en loci. The geno ypes esul ing om he ini ial un and he e un we e manually compa ed o all indi iduals o ule ou miss-sco ing o alleles. I any doub occu ed in his compa ison he sam- ples we e e-ex ac ed and e un a all loci o ob ain a con- sensus geno ype. The samples we e inally sc eened o abno mali ies (null alleles, sco ing e o s, e c.) in he so - wa e MICRO-CHECKER 2.2.3 (Van Oos e hou e al. 2004), using 1000 boo s ap eplica ions o gene a e he expec ed homozygo e and he e ozygo e allele size di e - ence equencies. S a is ics The wi hin popula ion gene ic a ia ion indices; numbe o alleles (N A ), expec ed (H e ) and obse ed (H o ) he e ozygos- i y, and he ixa ion index (F IS ) we e es ima ed in GenAlEx 6.41 (Peakall and Smouse 2006; Table S1). De ia ions om Ha dy–Weinbe g equilib ium (HWE) o each popula ion and locus and linkage disequilib ium (LD) among loci and among loci o e all popula ions we e es ed by exac es s (Guo and Thompson 1992) using GENEPOP 4.0 (Rousse 2007). The ables o pai -wise P- alues om he LD and HWE es s we e co ec ed o mul iple compa isons by sequen ial Bon e oni co ec ions (BFCs) ollowing Rice (1989). Allelic (N RA ) and p i a e allelic ichness (N RPA ) pe popula ion we e de e mined, accoun ing o di e ences in sample-sizes, using he a e ac ion p ocedu e o he small- es sample size (100 genes) as implemen ed in he so wa e HP-RARE 1.0 (Kalinowski 2005). To in e which o he wo ini ial endace s ockings (L. Sine € aj€ a i o L. Keluj€ a i) in L. Ina i ha led o suc- cess ul coloniza ion o he Ina i-Pas ik wa e cou se, a neighbo -joining ee was build using Nei e al. (1983) gene ic dis ance (D a ) and nodes we e es ed o obus - ness by 1000 boo s aps using Popula ions 1.2.32 (Langel- la 2005) and iewed in TREEVIEW (Page 1996). A sample o na i e endace (N=20) ob ained om no he n Ge many was used o assis he clus e ing o he neighbo -joining ee. Gene ic di e gence be ween he possible sou ce popula ions and he in oduced L. Ina i popula ion as well as among he colonis popula ions we e es ima ed by pai -wise F ST (Wei and Cocke ham 1984) alues and es ed o s a is ical signi icance (10,000 pe mu a ions) using ARLEQUIN 3.5.1.2 (Exco ie and Lische 2010). The able o P- alues o he pai -wise F ST alues was co ec ed o mul iple compa isons by BFCs ollowing Rice (1989). Bayesian clus e ing as implemen ed in STRUCTURE 2.3.2 (P i cha d e al. 2000; Hubisz e al. 2009), was used o p o ide ano he es ima e o o igin and popula ion s uc u e o he da a. We used a model assuming admix u e and co ela ed allele equencies be ween K popula ions (Bu n-ins o 100,000 eplica ions and 300,000 Ma ko chain Mon e Ca lo (MCMC) eplica es). Sampling loca ions we e used as a p io i in o - ma ion o assis he s uc u ing ( he LOCPRIOR model) Table 1. Loca ions and codes samples included in he gene ic analysis o he no he n Eu opean endace in asion (Fig. 2), wi h la i ude and lon- gi ude (posi ion), sample da e, wa e d ainage, lake size and heigh o loca ion in wa e cou se, da e o ansloca ion ( o L. Keluj€ a i and L. Sine € aj€ a i o L. Ina i) o i s obse a ion on locali y ( o L. Ina i, L. Vagga em and L. Sk ukkebuk a), secchi dep h, and sample size (N). Lake Code Posi ion Sample da e Wa e cou se (Ocean)* A ea (km 2 ) Heigh (m) Yea (Obse ed) Secchi dep h (Max dep h, m) N L. Keluj€ a i Ke 67°28′N, 27°4′E 14.11.2008 Kemijoki (BS) 9 187 1956 2 (10) 52 L. Sine € aj€ a i Si 66°35′N, 25°25′E 25.10.2008 Kemijoki (BS) 9 97 1964–1966 3 (40) 58 L. Ina i In 69°10′N, 27°55′E 30.10.2008 Ina i-Pas ik (AO) 1102 118 1973 5 (95) 55 L. Vagga em Vg 69°13′N, 29°14′E 09.09.2008 Ina i-Pas ik (AO) 34 52 1991 3 (30) 57 L. Skukkebuk a Sb 69°33′N, 30°7′E 16.09.2008 Ina i-Pas ik (AO) 7 21 1993 6 (38) 54 *BS, Bal ic Sea; AO, A c ic Ocean. K. P æbel e al. In asion Gene ics o Vendace ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 1403 as ecommended o weak signals o s uc u ing (Hubisz e al. 2009). All uns we e eplica ed 10 imes a each K=1–5 o con i m consis ency o log-likelihood p oba- bili ies. The mos likely (highes ln P (Χ|Κ)) g ouping was isualized using STRUCTURE HARVESTER (Ea l and onHold 2012). S uc u ing o he endace popula- ions we e also es ima ed by p incipal componen analysis (PCA). The o dina ion o he indi iduals was pe o med in he p og am GenAlEx 6.41 (Peakall and Smouse 2006). Finally, we in es iga ed whe he sha ed p i a e alleles could be iden i ied in pai -wise compa isons o he sou ce L. Keluj€ a i and/o L. Sine € aj€ a i and he in oduced popula ion (L. Ina i), as well as among popula ions wi hin he wa e cou se which would suppo any link ound in he popula ion s uc u ing app oach. The a e- ac ion p ocedu e implemen ed in ADZE 1.0 (Szpiech e al. 2008) we e used o es ima e he pai -wise sha ed p i a e allelic ichness ac oss loci o he possible sou ce popula ions and he in oduced popula ion (L. Ina i) and o he seconda y expansion wi hin he wa e cou se (i.e., be ween L. Ina i-L. Vagga em, L. Ina i-L. Sk ukkebuk a, and L. Vagga em-L. Sk ukkebuk a). All es ima es we e pe o med using a s anda dized sample size co espond- ing o he smalles sample (100 genes). We es ed whe he he ini ial s ocking and he subse- quen in asion was associa ed wi h ounde e ec s by es ing o gene ic bo lenecks among he popula ions using he so wa e BOTTLENECK 1.2.02 (Co nue and Luika 1996). Popula ion bo lenecks will cause a empo- a ily imbalance in he mu a ion-d i equilib ium, whe e addi ions o new alleles ia mu a ion a e balanced by he loss o alleles ia d i (Luika and Co nue 1998). We u ilized his assump ion o iden i y si ua ions whe e he ac ual sample he e ozygosi y exceeds a pe mu ed equilib- ium he e ozygosi y as expec ed unde a mu a ion-d i equilib ium. I he ac ual he e ozygosi y exceeds he equi- lib ium he e ozygosi y i is indica i e o a ecen popula- ion bo leneck (Co nue and Luika 1996). We used 1000 coalescen simula ions and assumed a wo-phased model o mu a ion (TPM) and he mo e conse a i e s ep-wise mu a ion model (SMM). The s a is ical signi i- cance o he de ia ions a equilib ium and obse ed he - e ozygosi ies we e es ed wi h Wilcoxon signed- ank es s. We also es ed o gene ic signa u es o ounde e ec s ollowing he in asion and he seconda y expansion, by es ima es o changes in he e ec i e popula ion size (N e ). Changes in he e ec i e popula ion size will indica e he ela i e con ibu ion o gene ic d i (Hed ick 2000). N e was es ima ed o all popula ions using OneSamp 1.1 (Tallmon e al. 2008). This so wa e uses app oxima e Bayesian compu a ion o es ima e a iance N e om sum- ma y s a is ics ha a e ela ed o N e . We used p io uppe and lowe bounds o N e o 2–1000 and 10,000 eplica ions o gene a e he 95% c edible in e als. Finally, we co ela ed he pai -wise F ST ’s and he di e - ence in expec ed he e ozygosi y be ween he sou ce popu- la ion and Ina i-Pas ik popula ions (DH e ) wi h geog aphic dis ance and yea s om ansloca ion o occu ence. This was pe o med o e eal spa ial and em- po al pa e ns o he mechanism unde lying he ounde e en and o es ima e he ime o he possible expansion om L. Ina i downs eam he Pas ik wa e cou se. We expec ed inc eased di e en ia ion (pai -wise F ST ’s) and di e ence in gene ic di e si y (DH e ) he longe he dis- ance and ime since di e gence. The co ela ion o geno- ype da a and geog aphical dis ance (physical di e gence) om he sou ce popula ion o each o he popula ions in he Ina i-Pas ik wa e cou se we e compa ed wi h he co - esponding co ela ions o he h ee possible expansion imes. The ime om in oduc ion o he ounding o he L. Ina i popula ion was in all es ima es se o 8 yea s, whe eas he expansion ime om L. Ina i downs eam he wa e cou se was es ed o ini ial sigh ing o endace la ae (1973), o al co e o lake (1980), and peak (1989) (Table S2). We used pa ial Man el es s as implemen ed in GenAlEx 6.41 (Peakall and Smouse 2006), o es hese scena ios, using 9999 pe mu a ions o ob ain signi icance. Resul s Geno yping, alida ion, and quali y con ol o geno ypic da a We did no iden i y any misma ch be ween he o iginal indi idual mul ilocus geno ypes and he e-ex ac ed 3–4% eplica es wi hin he p esen da ase . He e ozygo e de ici s we e indica ed by MICRO-CHECKER a BWF1 (L. Vagga em, L. Sk ukkebuk a), C2-157 (L. Sk ukke- buk a), Cocl-la 06 (L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a) Cocl-la 10 (L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a), Cocl-la 49 (L. Sk uk- kebuk a), ClaTe 06 (L. Keluj€ a i), ClaTe 13 (L. Sine € aj€ a i and L. Sk ukkebuk a) all indica ed as caused by he p es- ence o null alleles. To es whe he he loci wi h po en ial null alleles may a ec he esul s he STRUCTURE analy- ses we e pe o med as desc ibed in he Ma e ial and Me h- ods bu wi hou he loci showing he mos he e ozygo e de ici s ac oss popula ions (BWF1, Cocl-la 06, and Cocl- la 10). Nei he he numbe o in e ed clus e s no he popula ion s uc u e was ound o di e when compa ing he esul s wi h and wi hou hese loci (Figs. S1 s. 4). Gi en he ela i ely ew loci used in he s udy we he e o e main ained all loci in he ull analysis o ensu e s a is ical powe . In asion Gene ics o Vendace K. P æbel e al. 1404 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Gene ic a ia ion The s anda d indices o wi hin popula ion gene ic a ia ion a e gi en in Tables 2 and S1. We disce ned 124 alleles among he en mic osa elli e loci assayed in he i e s udied endace popula ions, wi h a wi hin popula ion a ia ion o 47 o 89 alleles. Mean numbe o alleles (N A ) pe locus pe popula ion a ied om 2 o 25 and expec ed (H e ) and obse ed he e ozygosi y (H o ) pe locus pe popula ion a ied om 0.019 o 0.916 and 0.019 o 0.895, espec i ely. All popula ions, excep L. Keluj€ a i, showed signi ican depa u es om HWE associa ed wi h he e ozygo e de i- ci s. Linkage disequilib ium was iden i ied o 4 ou o 45 pai -wise locus es s ac oss popula ions bu all e u ned non-signi ican a e sequen ial BFC. The indi idual locus es s displayed 18 ou o 50 and 7 ou o 50 signi ican de i- a ions om HWE be o e and a e BFCs, espec i ely (Table S1), and 2 o 9 signi ican LD’s wi hin each popula- ion (only h ee signi ican a e BFC, all in L. Sine € aj€ a i). Theallelic ichnesswe ealmos doubleinL.Sine - €aj€a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a (N RA =7.9–8.8) compa ed o L. Keluj€ a i (N RA =4.6) (Table 2). P i a e alle- lic ichness a ied be ween N RPA =0.11–0.97 wi h he lowes N RPA ound in he lowe Pas ik i e lakes. H e and H o we e also highe in L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a (H e =0.596–0.614; H o =0.506–0.546) com- pa ed o L. Keluj€ a i (H e =0.400; H o =0.369) (Table 2). Iden i ying he sou ce o he endace in asion and popula ion s uc u e wi hin he seconda y expansion The pai -wise F ST es ima es o gene ic di e en ia ion be ween he wo possible sou ce popula ions (L. Keluj€ a i/L. Sine € aj€ a i) and L. Ina i sugges ha L. Sine €aj€a i was he sou ce popula ion as his popula ion pai display he lowes , al hough signi ican , di e en ia- ion (L. Keluj€ a i s. L. Ina i, F ST =0.158, P<0.0001; L. Sine € aj€ a i s. L. Ina i, F ST =0.011, P=0.0027; Table S3). L. Keluj€ a i mo eo e appea ed o be he mos isola ed popula ion o he i e popula ions s udied (F ST =0.141–0.158; P<0.0001). Wi hin he Ina i-Pas ik wa e cou se, he in oduced popula ion (L. Ina i) was signi ican ly di e en om he popula ions om he seconda y expansion (L. Vagga em and L. Sk ukkebuk a), whe eas he popula ions om L. Vagga em and L. Sk uk- kebuk a could no be signi ican ly disc imina ed (Table S3). The sequen ial BFCs did no change he signi icance le el o any o P- alues om he pai -wise F ST es ima es. The PCA plo e ealed g oupings o (i) L. Keluj€ a i and (ii) L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukke- buk a (Fig. 3). Each axis, PC1 and PC2, explained 26.6% and 24.4% o he o al a ia ion, espec i ely. The Bayesian clus e ing e ealed ha he indi iduals could be pa i ioned in o wo gene ic clus e s (K=2; ln P (Χ|Κ)SD =7209 1; Fig. S2), sepa a ing L. Keluj€ a i in o one and L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a in o he o he clus e (Fig. 4). The combined app oach o a neighbo -joining ee wi h Nei e al. (1983) gene ic dis ance (D a ) p o ided suppo o L. Sine € aj€ a i being he mos likely sou ce o he endace in he Ina i- Pas ik wa e cou se. Mo eo e , he neighbo -joining ee also p o ided suppo o di e gence be ween he ounde popula ion o L. Sine - € aj€ a i and he in oduced popula- ion in L. Ina i (boo s ap suppo o 100%) as well as be ween L. Ina i and he popula ions om he seconda y expansion (L. Vagga em and L. Sk ukkebuk a, 98%), which is also suppo ed by he low, bu signi ican F ST alues. Finally, he Bayesian clus e ing and he neighbo -joining ee g ouped L. Vagga em and L. Sk ukkebuk a oge he suppo ing he non-signi ican F ST alue es ima ed be ween hese popula ions. A highe mean numbe o sha ed p i a e alleles we e ound wi hin he L. Sine € aj€ a i-L. Ina i pai compa ed o he L. Keluj€ a i-L. Ina i pai (Fig. 5a). The mean numbe o p i a e alleles sha ed by he L. Keluj€ a i-L. Sine € aj€ a i pai esembled ha o he L. Keluj€ a i-L. Ina i pai , sup- po ing he conclusion ha L. Sine € aj€ a i is he sou ce popula ion o he in asion. Fo he seconda y expansion he pai -wise compa ison o mean numbe o sha ed p i- Table 2. Summa y gene ic s a is ics o he endace popula ions included in he s udy. Sample F IS H e H o N A N RA N RPA TPM SMM Ne Ke 0.0772 0.400 0.369 47 4.6 0.79 0.084 0.003 (d, 0.002) 55 (38–153) Si 0.1133*0.614 0.544 89 8.6 0.97 0.322 0.019 (d, 0.009) 234 (130–988) In 0.0870*0.597 0.546 89 8.8 0.73 0.193 0.007 (d, 0.003) 250 (150–746) Vg 0.0849*0.596 0.546 81 7.9 0.11 0.625 0.019 (d, 0.009) 150 (93–422) Sb 0.1664*0.607 0.506 79 8.4 0.58 0.275 0.003 (d, 0.002) 200 (112–703) The coe icien o inb eeding (F IS ), gene ic di e si y es ima es (H e /H o ), o al numbe o alleles (N A ), mean allelic (N RA ), and p i a e allelic ichness (N RPA ). The P- alues o he Bo leneck es a e gi en (TPM and SMM) and signi ican alues (SMM) we e only ound signi ican o he e ozygo e de ici (deno ed wi h d and he co esponding P- alue o he exac es ). The e ec i e popula ion sizes (Ne) wi h c edible in e als a e gi en. Sample codes as in Table 1. *Signi ican he e ozygo e de ici (posi i e F IS ) es ima ed by Ha dy–Weinbe g exac es s. K. P æbel e al. In asion Gene ics o Vendace ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 1405 a e alleles show ha he L. Ina i-L. Vagga em pai sha e ewe p i a e alleles han he L. Vagga em-L. Sk ukkebuk a pai , wi h he L. Ina i-L. Sk ukkebuk a pai sha ing he lowes numbe o p i a e alleles (Fig. 5b). This sugges s a s epping-s one dispe sal whe e L. Ina i endace colonized L. Vagga em and he popula ion in L. Vagga em subse- quen ly ounded he popula ion in L. Sk ukkebuk a (Fig. 2b). Some speci ic alleles (e.g., allele 270 a ClaTe 13, allele 190 a Cocl-la 49, bu see Fig. 2b) we e also obse ed only in he L. Sine € aj€ a i, L. Ina i, L. Vagga em, and L. Sk ukkebuk a popula ions and display an expanding pa - e n, while o he s (e.g., allele 189 a ClaTe 06) we e p esen in all popula ions bu in much highe /lowe equencies in L. Keluj€ a i compa ed o he o he ou lakes (Fig. 2b). Founde e ec s and pa e ns o spa ial and empo al gene ic a ia ion None o he popula ions showed signi ican ly highe expec ed he e ozygosi y (H e ) han equilib ium he e ozy- gosi y (H eq ) in he analysis o bo lenecks assuming a TPM. Simila esul s we e e ealed when assuming a SMM. All es ima es o e ec i e popula ion sizes we e asso- cia ed wi h o e lapping CIs (Table 2). Howe e , he es i- ma e o L. Sine €aj€a i and L. Ina i did no di e in magni ude, suppo ing ha he in oduc ion o endace in o he sys em was associa ed wi h limi ed ounde e ec s (d i ). A dec easing end in e ec i e popula ion size (al hough all CIs o e lapped) was e ealed om L. Ina i o L. Sk ukkebuk a sugges ing an inc eased in luence o d i and es ic ed gene low du ing he expansion. All spa ial and empo al co ela ions o geno ypes we e ound o be signi ican (Table 3). The pe iod o peak abundance o endace (1989) in L. Ina i also co ela e he bes wi h pa e ns o gene ic expansion (DH e ), sugges ing ha he seconda y expansion downs eam he wa e cou se occu ed la e in he in asion p ocess and ha i was associ- a ed wi h gene ic cons ains. This is also in acco dance wi h he obse a ion o endace in ca ches in Pas ik; endace we e i s obse ed in he uppe Pas ik in 1989, and hen successi ely downs eam he wa e cou se (Vagga em in 1991 and Sk ukkebuk a in 1993; Amundsen e al. 1999). Discussion P ima y sou ce o he Ina i-Pas ik endace in asion The e we e wo possible sou ces o he in oduc ion and subsequen downs eam in asion o endace in he Ina i- PC 1 (26.6%) PC 2 (24.4%) Ke Si In Vg Sb Figu e 3. P incipal componen analysis (PCA) plo o he gene ic s uc u ing among he i e endace popula ions. PC1 and PC2 explain 26.6% and 24.4% o he o al a ia ion, espec i ely. Sample abb e ia ions as in Table 1. Ke Si In Vg Sb Ke Si In Vg Sb 100 0.1 Da Ge Ca 98 50 Figu e 4. Gene ic s uc u e among L. Keluj€ a i (Ke), L. Sine € aj€ a i (Si), L. Ina i (In), L. Vagga em (Vg), and L. Sk ukkebuk a (Sb) as e ealed by a neighbo -joining ees using Nei e al. (1983) gene ic dis ance (D a ) and Bayesian clus e ing o all indi iduals using STRUCTURE (Hubisz e al. 2009) assuming wo gene ic clus e s o indi iduals (K=2). Only boo s ap esampling pe cen ages abo e 50 a e shown o he neighbo -joining ee. In he STRUCTURE analysis black lines sepa a e indi iduals om di e en sampling si es (labeled igh ) and each indi idual is ep esen ed by a hin ho izon al line, which is pa i ioned in o K-colo ed segmen s ep esen ing indi idual’s es ima ed membe ship ac ions in Kclus e s. The absolu e mean alues o ln P (X|K) a e plo ed o K=1–5 in supplemen a y Fig. S2. In asion Gene ics o Vendace K. P æbel e al. 1406 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. Pas ik wa e cou se (Mu enia and Salonen 1992; Salonen 1998). Mul ilocus mic osa elli e da a and se e al s a is ical app oaches clea ly iden i ied L. Sine € aj€ a i endace as he mos likely ounde popula ion o Ina i-Pas ik endace. The low, bu signi ican , pai -wise F ST alue ound be ween he endace popula ions o L. Sine € aj€ a i and L. Ina i indica es ha hese popula ions ha e di e ged ecen ly. In compa ison, he F ST alue be ween L. Keluj€ a i and L. Ina i was high and ep esen s he ypi- cal le el o gene ic di e gence epo ed be ween o he na i e pos glacial co egonid popula ions in no he n Fen- noscandia (Øs bye e al. 2006; Saisa e al. 2008; P æbel e al. in p ess). Mo eo e , suppo o L. Sine € aj€ a i being he sou ce popula ion o he in asion was ound in he combined app oach o a neighbo -joining ee and Bayesian clus e ing, whe e gene ic homogenei y was shown o L. Sine € aj€ a i and he Ina i-Pas ik wa e cou se popula ions whe eas L. Keluj€ a i o med i s own clus e . Bayesian clus e ing and neighbo -joining ees ha e suc- cess ully been used in o he s udies o in e ence o ou es and sou ces o biological in asions (see e iew by Es oup and Guillemaud 2010), and suppo i e esul s, as seen he ein, p o ide s ong e idence o he e ealed pa e n. L. Sine € aj€ a i and L. Ina i also sha ed a highe numbe o p i a e alleles han L. Keluj€ a i and L. Ina i, which p o ide addi ional suppo o L. Sine € aj€ a i being he ounde s, as ime since in oduc ion is oo sho o many new p i a e alleles o accumula e. Add essing he numbe o sha ed p i a e alleles be ween pai s o popula ions has been used in se e al o he s udies o iden i y he o igin o in asions and mig a ions (e.g., Szpiech e al. 2008; Bell and Ma ocq 2011). I is also wo h conside ing ha he simila i y in gene ic a ia ion o L. Sine € aj€ a i and L. Ina i p o ide addi ional suppo o his o igin, as biological in asions e y seldom a e associa ed wi h an inc ease in gene ic a ia ion in he in oduced popula ion compa ed o he sou ce popula ion (Nei e al. 1975; Dlugosch and Pa ke 2008b). Thus, aken oge he , he e is compelling e idence ha he second in oduc ion o endace in 1964–1966 om L. Sine € aj€ a i ep esen s he sou ce o he endace in asion and es ablishmen in he Ina i-Pas ik wa e cou se. Founde e ec s in he ini ial Ina i-Pas ik endace in asion Mul iple in oduc ions ha e been shown o e ain gene ic di e si y (Kolbe e al. 2004; Facon e al. 2008; Gelle e al. 2010) and hus inc ease he likelihood o success ul coloniza ion and demog aphic expansion o he in ade (e.g., Dlugosch and Pa ke 2008b; Shine 2012). In he p esen s udy he success ul sou ce popula ion o he in asion was iden i ied as he one ha was in oduced se e al imes in 1964–1966 o a small lake ups eam L. Ina i. The popula ion ha appa en ly did no succeed in es ablishing was in oduced o L. Ina i by escapees om a local ha che y (Mu enia and Salonen 1992). This implies ha he p opagule p essu e o his sou ce likely has been smalle and/o ha he popula ion was p esum- ably less adap ed o he wa e cou se compa ed o he epea ed in oduc ions om L. Sine € aj€ a i. In ac , he esul s show ha he L. Ina i popula ion has simila alle- lic ichness and he e ozygosi y o he sou ce popula ion. Table 3. Spa ial and empo al co ela ions o pai -wise di e ences (F ST ) and di e ences in expec ed he e ozygosi y (DH e ) using pa ial Man el es s. F ST DH e R xy PR xy P Geog aphical dis ance 0.548 0.041 0.987 0.040 Time Ini ial_obse a ion 0.778 0.045 0.623 0.040 Time To al_co e age 0.682 0.045 0.817 0.042 Time Peak_abundance 0.483 0.042 0.912 0.041 The coe icien o eg ession (R xy ) and he co esponding P- alue (9999 pe mu a ions) a e gi en. Geog aphical dis ances and expansion imes used o co ela ions a e gi en in Table S2. DH e , Pai -wise di e - ence in he e ozygosi y be ween L. Sine € aj€ a i and he popula ions wi hin he Ina i-Pas ik wa e cou se. (a) (b) 0.00 0.10 0.20 0.30 0.40 0102030405060708090100 Sample size (genes) Mean numbe o sha ed p i a e alleles SiIn KeIn KeSi 0.00 0.10 0.20 0.30 0.40 010203040 5060708090100 Sample size (genes) InVg InSb VgSb Mean numbe o sha ed p i a e alleles Figu e 5. The mean numbe o sha ed p i a e alleles p i a e o loca ion pai s o L. Keluj€ a i-L. Ina i (KeIn), L. Sine € aj€ a i-L. Ina i (SiIn), and L. Keluj€ a i-L. Sine € aj€ a i (KeSi) (a) and wi hin he Ina i-Pas ik wa e cou se (L. Ina i-L. Vagga em, InVg; L. Ina i-L. Sk ukkebuk a, InSb; and L. Vagga em-L. Sk ukkebuk a, VgSb) (b) as a unc ion o s anda dized sample sizes (in genes). E o ba s ep esen SEM ac oss he 10 mic osa elli e loci. K. P æbel e al. In asion Gene ics o Vendace ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 1407 We did no iden i y loss o gene ic a ia ion ia gene ic bo lenecks o he in oduced L. Ina i popula ion, which ha e been shown o be he case o o he biological in a- sions (e.g., Tsu sui e al. 2000; Hen y e al. 2009; Ay es e al. 2010). Simila e ec i e popula ion size es ima es in he L. Sine € aj€ a i and Ina i-Pas ik popula ions also sug- ges ed li le gene ic d i . Thus, he ini ial ounding e en o endace in he Ina i-Pas ik wa e cou se may be asso- cia ed wi h no o limi ed ounde e ec s. Howe e , we show ha he sou ce and he in asi e popula ions a e gene ically di e en , which sugges ha gene ic d i ac ually has played a ole in he es ablishmen o he col- onis s, since ime om in oduc ion is oo sho o mu a ions o ha e accumula ed ac oss he mic osa elli e loci in high enough equencies o ha e e ec on he esul . Dispe sal pa e n and gene ic s uc u e o he seconda y expansion The esul s e ealed ha he popula ions esul ing om he seconda y expansion downs eam om L. Ina i ha e di e ged gene ically om he in oduced popula ion in a s epping s one pa e n. Such a dispe sal pa e n main ains gene ic a ia ion and limi s di e gence compa ed o o he expansion pa e ns (e.g., Ib ahim e al. 1996; Reusch e al. 2010; Tonione e al. 2011), due o he allelic pa chiness and likely subsequen gene low among demes. The popu- la ions wi hin he seconda y expansion ha e become gene ically dis inguishable om he L. Ina i popula ion in abou 18 yea s/9 gene a ions (maximum 35 yea s/17 gen- e a ions i he Pas ik popula ions we e ounded al eady in 1973 a he ime o i s obse a ion in L. Ina i, bu his scena io appea s highly unlikely om he obse ed occu ences o endace in he wa e cou se; see Amundsen e al. 1999). This is ema kable, especially because he high gene ic a ia ion and la ge consensus and e ec i e popula ion size o he colonis popula ion e ec i ely should coun e ac gene ic di e gence. In ma ine ish spe- cies, such as he ing (Clupea ha engus), capelin (Mallo us illosus), and A lan ic cod (Gadus mo hua), i is well known ha la ge popula ion sizes wi h high gene ic a ia- ion and no ob ious ba ie s o gene low e ec i ely ham- pe he build up o ep oduc i e isola ion (Knu sen e al. 2003; Ma iani e al. 2005; P æbel e al. 2008; bu see e iew by DeWoody and A ise 2000). The majo di e - ence and he mos likely explana ion o he apid di e - gence obse ed among he endace popula ions wi hin he Ina i-Pas ik wa e cou se he e o e appea s o be he p esence o se e al hyd oelec ic dam cons uc ions be ween he L. Ina i popula ion and he wo downs eam popula ions om he seconda y expansion (i.e., L. Sk uk- kebuk a). These dams may e ec i ely hampe downs eam gene low and p omo e he build up o ep oduc i e isola ion du ing he expansion. Adap i e and e olu iona y changes in he in ade –pe spec i es Iden i ying he sou ce popula ion and econs uc ing he ou es o biological in asions a e c ucial o handling and managing in asi e species, as well as o gaining knowledge o he ecology behind success ul coloniza ion e en s. How- e e , in asions also o e he possibili y o s udy e olu ion- a y p ocesses such as na u al selec ion and li e his o y changes as hey un old. Fo example, in ade s may e ol e in esponse o an al e ed selec ion egime compa ed o hei na i e ange (Facon e al. 2006; Sax e al. 2007; Bacig- alupe 2009; Shine 2012). The Ina i-Pas ik wa e cou se is 300–380 km no h o L. Sine € aj€ a i, bu a simila al i ude. Thus, he summe is sho e and colde , and he ice-co - e ed pe iod longe . Especially, empe a u e is an impo an d i e o adap i e changes as i will a ec , o example, egg incuba ion ime, g ow h, physiological p ocesses (Q10), and gene al indi idual i ness (e.g., Mooney and Hobbs 2000). In addi ion, -s a egis s a e o en a o ed du ing biological in asions (Lodge 1993; Facon e al. 2006), and ou ea lie s udy sugges s ha hese in asi e popula ions in he seconda y expansion a eas ha e speeded up hei li e his o y ia g ow h and age a sexual ma u i y (Amundsen e al. 2012). These changes happened wi hin a decade and he sys em he e o e appa en ly ep esen s an excellen example o apid adap i e e olu ion. Fu u e s ud- ies o his sys em may he e o e gain aluable insigh s by iden i ying adap i e pheno ypic ai s ( ansc ip ome) and he gene ic basis o local adap a ion (e.g., using pheno- ypic QTLs) (see e.g., Be na chez e al. 2010). Mo eo e , insigh s would be gained om in es iga ing a ia ion a genes and genomic blocks ela ed o impo an li e his o y pa ame e s such as g ow h, me abolism, and disease esis- ance, o in e adap i e changes o pa e ns o plas ici y among he Sine € aj€ a i-Ina i-Pas ik popula ions. Acknowledgmen s We hank Laina Dalsbø, Tanja L. Haneb ekke, and Cesilie Lien, and o he s a o he Depa men o A c ic and Ma ine Biology, Uni e si y o T omsø, and he Finnish Game and Fishe ies Resea ch Ins i u e, Ina i, o excellen assis ance in ield wo k and in he lab. Jan Die king is hanked o p o iding he C. albula om no he n Ge - many. We hank E ic Taylo , h ee anonymous e e ees, Sh ipa hi Bha , Ma jo ie Cou on, S ein-E ik Fe olden, and Kja an Øs bye o discussions and many aluable commen s on an ea lie d a . The s udy was inanced by he No wegian Resea ch Council (NFR 183984/S30). In asion Gene ics o Vendace K. P æbel e al. 1408 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.