ORIGINAL PAPER

Eco-immunology of fish invasions: the role of MHC variation

C.Monzón-Argüello & C. Garcia de Leaniz &G. Gajardo &

S. Consuegra

Received: 21 November 2013 /Accepted: 31 March 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract The relationship between invaders and the patho-gens encountered in their new environment can have a largeeffect on invasion success. Invaders can become free fromtheir natural pathogens and reallocate costly immune re-sources to growth and reproduction, thereby increasing inva-sion success. Release from enemies and relaxation of selectivepressures could render newly founded populations more var-iable at immune-related genes, such as the major histocom-patibility complex (MHC), particularly when they have dif-ferent origins. Using rainbow and brown trout, two of theworld’s most successful fish invaders, we tested the generalhypothesis that invaders should display high intrapopulationimmunogenetic diversity and interpopulation divergence, dueto the interplay between genetic drift and successive waves ofgenetically divergent introductions. We analysed genetic di-versity and signatures of selection at the MHC class II βimmune-related locus. In both species, MHC diversity (allelicrichness and heterozygosity) for southern hemisphere popula-tions was similar to values reported for populations at theirnative range. However,MHC functional diversity was limited,and population immunogenetic structuring weaker than thatobserved using neutral markers. Depleted MHC functionaldiversity could reflect a decrease in immune response,immune-related assortative mating or selection for resistanceto newly encountered parasites. Given that the role of MHC

diversity in the survival of these populations remains unclear,depleted functional diversity of invasive salmonids couldcompromise their long-term persistence. A better understand-ing of the eco-immunology of invaders may help in managingand preventing the impact of biological invasions, a majorcause of loss of biodiversity worldwide.

Keywords Oncorhynchus mykiss . Salmo trutta . Biologicalinvasions .MHC . Enemy release . Immunogenetics

Introduction

Biological invasions are the subject of recent scientific con-troversies that have important economic and societal implica-tions (Simberloff et al. 2013), costing billions to global econ-omies (Pimentel et al. 2005). Invasive species are importantdrivers of ecological change (Strayer et al. 2006) and biodi-versity decline (Butchart et al. 2010), but predicting theirimpacts has proved elusive. Invasive provide some of themost striking examples of rapid evolution (Buswell et al.2011; Carroll 2007; Hendry et al. 2008; Whitney and Gabler2008), and understanding the basis of establishment success isconsidered key for their management. Pathogens can hamperthe reproduction and development of invaders and constraintheir invasive potential (Torchin and Mitchell 2004), thereforepotentially influencing the outcome of biological invasions. Abetter understanding of the eco-immunity of invaders could,therefore, provide useful insights into the drivers of establish-ment success.

In some cases, invasion success can be explained by theenemy release hypothesis (Keane and Crawley 2002).According to this, invaders are liberated from their naturalenemies (parasites, pathogens and predators) when they colo-nize new environments, enabling them to reallocate costlyimmune defence resources to growth and reproduction that

Electronic supplementary material The online version of this article(doi:10.1007/s00251-014-0771-8) contains supplementary material,which is available to authorized users.

C. Monzón-Argüello :C. Garcia de Leaniz : S. Consuegra (*)Department of Biosciences, Swansea University, Swansea SA2 8PP,UKe-mail: s.consuegra@swansea.ac.uk

G. GajardoLaboratorio de Genética, Acuicultura y Biodiversidad, Universidadde Los Lagos, Osorno, Chile

ImmunogeneticsDOI 10.1007/s00251-014-0771-8

may facilitate invasions (Lee and Klasing 2004). Some sup-port for this comes from the observation that some populationsintroduced into new environments are less likely to be infectedthan native populations (Torchin andMitchell 2004), althoughthe role of novel pathogens on invaders appears to be morecomplex than what the simple release hypothesis would sug-gest (Colautti et al. 2004). Indeed, founder effects and bottle-necks could eventually result in immunogenetic losses, ren-dering invaders more susceptible to novel parasites (Whiteand Perkins 2012). However, relaxation in parasite selectivepressures and successive waves of invaders each bringingdifferent parasites and immunogenetic diversity into recipientecosystems could result in a rapid divergence of populations.In this sense, aquaculture-mediated introductions, the maincause of aquatic invasions together with shipping (Molnaret al. 2008; Naylor et al. 2001), provide good opportunitiesto test the role of immunocompetence on fish invasions, asthey tend to consist of multiple introductions from differentgeographical locations (Consuegra et al. 2011). Invasionsoriginating from aquaculture escapes could result in an im-mune repertoire highly divergent among populations due tothe interplay between genetic drift and successive waves ofgenetically diverse introductions.

The genes of the major histocompatibility complex (MHC)are excellent candidates for this study. Central to the immuneresponse, MHC genes encode for proteins that presentpathogen-derived antigens to T cells, initiating the adaptiveimmune response (Janeway et al. 2004) and are amongst themost polymorphic and best-studied functional genes in verte-brates (Hughes and Yeager 1998). Variation in the residuesthat bind antigens from pathogens is critical for the effective-ness of the immune response (Hedrick and Kim 2000) and isthought to be maintained not only by balancing selectiondriven by pathogens (either through overdominance orfrequency-dependent selection, Doherty and Zinkernagel1975; Slade and McCallum 1992) but also by mate choice(Apanius et al. 1997; Consuegra and Garcia de Leaniz 2008).Parasites seem to be the main agent of selection acting onMHC genes, as suggested by multiples lines of evidenceprovided by heterozygote advantage (Kurtz et al. 2004;Wegner et al. 2003), rare allele advantage (Schwensow et al.2007), the association of individual MHC alleles and/or ge-notypes with susceptibility to specific pathogens (Bonneaudet al. 2006b; Gómez et al. 2010) and changes in allele fre-quencies after parasite exposure (Eizaguirre et al. 2012).

Rainbow trout (Oncorhynchus mykiss) and brown trout(Salmo trutta) are two salmonids that have been introducedworldwide for sport fishing and aquaculture (Froese and Pauly2013; Lowe et al. 2000). The two species have non-overlap-ping native ranges but have converged in many parts ofChilean Patagonia (Correa and Hendry 2012; Young et al.2010), where they have established and tend to have contrast-ing introduction histories and dispersal patterns (Young et al.

2010). In Chile, rainbow and brown trout were originallyintroduced for recreational purposes in 1905, most likely asimported ova from US, Germany and England, although addi-tional sources cannot be completely ruled out (Basulto 2003).Both species were then shipped from Chile to the FalklandIslands between 1936 and 1947, although only brown troutsurvived and founded self-sustained populations (Arrowsmithand Pentelow 1965). Brown trout displays a narrower geo-graphic range than rainbow trout in Chile but has a strongerimpact on native fishes (Young et al. 2010; Correa and Hendry2012), having dispersed mostly through stocking and naturalcolonization (Gajardo and Laikre 2003; Schröder and Garciade Leaniz 2011; Garcia de Leaniz et al. 2010). In contrast,rainbow trout is only present in Chile, where its spread hasbeen facilitated by massive escapes of farmed fish since the1990s, following the rapid expansion of the Chilean salmonindustry (Gajardo and Laikre 2003; Schröder and Garcia deLeaniz 2011; Garcia de Leaniz et al. 2010; Consuegra et al.2011). Here, we examined patterns of neutral (microsatellites)and functional (MHC class II β) genetic diversity in twoecologically similar invasive salmonids with different modesof dispersal in the southern hemisphere to test the generalhypothesis that enemy release would result in high immuno-genetic diversity and population divergence, particularly in thecase of rainbow trout aided by secondary releases.

Material and methods

Fish sampling and laboratory procedures

We analysed a fragment of 254 bp of the exon 2 of the MHCclass II β gene, containing most of the peptide-binding region(PBR) in 151 brown trout from six rivers in Chile and threerivers in the Falkland Islands (Fig. 1; Table 1) using theprimers CL007 (Landry et al. 2001) and AL1002 (Olsenet al. 1998). Approximately 50 ng of extracted DNAwas usedin 20 μL PCR mixes containing 0.2 μM of each primer,0.25 mM dNTPs, 0.5 U of Taq DNA polymerase (Bioline,London, UK), 1x buffer and 2.5 mM MgCl2. Thermal condi-tions consisted of 5 min initial denaturation cycle (94 °C)followed by 35 cycles of 1 min at 94 °C, 1 min at 57 °C,1 min at 72 °C and a final extension cycle of 10 min at 72 °C.Amplified fragments were directly sequenced using the samePCR primers and resolved in a 3130 automated sequencer(Applied Biosystems). Resulting sequences were alignedusing BioEdit 7.0.5.3 (Hall 1999) and compared with previ-ously described brown trout sequences retrieved fromGenBank, as in Consuegra and Garcia de Leaniz (2008), inorder to assign alleles. Sequences that had not been previouslydescribed were cloned using the TOPO TA Cloning® Kit forSequencing (Invitrogen), and six clones per individual wereselected for forward and reverse sequencing. Only alleles

Immunogenetics

identified in at least two independent PCRs were consideredfor subsequent analyses; these are alleles that appeared in two

independent PCRs from the same individual or in the inde-pendent PCRs of at least 2 individuals. We compared MHC

Fig. 1 Sampling locations ofrainbow trout (Oncorhynchusmykiss) and brown trout (Salmotrutta) populations in Chile andthe Falkland Islands. Open andclosed circles represent rainbowtrout and brown trout populations,respectively, while stars representrivers sampled for both

Table 1 Diversity indices for 9 brown trout and 10 rainbow trout populations at neutral microsatellite (Microsat) and functional loci (MHC)

Species Population Marker N K AR Ho He J’

Brown trout Chile Golgol Microsat/MHC 21 5.571/14 4.347/8.306 0.665/0.778 0.662/0.897 –

Butalcura Microsat/MHC 22 4.643/11 3.895/6.241 0.615/0.737 0.631/0.782 –

Blanco-Enco Microsat/MHC 19 5.071/8 4.071/6.632 0.665/0.889 0.655/0.859 –

Pangal Microsat/MHC 23 4.143/8 3.551/6.161 0.625/0.737 0.622/0.825 –

Encanto Microsat/MHC 21 5.214/7 4.134/5.925 0.617/0.947 0.649/0.838 –

Bonito Microsat/MHC 20 5.214/7 4.316/6.221 0.683/0.800 0.673/0.832 –

Falklands Estancia Brook Microsat/MHC 23 7.929/21 5.634/11.137 0.724/0.810 0.765/0.965

Finlay Creek Microsat/MHC 23 2.786/4 2.446/3.084 0.408/0.650 0.412/0.541

Sarnys Creek Microsat/MHC 15 3.143/6 2.664/6.000 0.408/0.714 0.408/0.791

Rainbow trout Chile Encanto Microsat/MHC 23 7.429/11 6.412/9.118 0.737/0.913 0.750/0.865 0.470

Nilque Microsat/MHC 23 6.714/12 6.135/9.255 0.615/0.826 0.760/0.850 0.216

Pescadero Microsat/MHC 24 7.429/16 6.455/12.752 0.643/0.958 0.753/0.928 0.817

Blanco-Correntoso Microsat/MHC 20 6.875/11 6.253/9.422 0.627/0.842 0.744/0.824 0.668

U23 Microsat/MHC 17 7.286/14 7.023/12.410 0.739/0.941 0.798/0.914 0.977

Aitoy Microsat/MHC 16 7.714/12 7.478/11.282 0.767/0.813 0.830/0.913 0.875

Pangal Microsat/MHC 16 6.714/14 6.425/12.69 0.739/0.875 0.764/0.885 0.992

Bonito Microsat/MHC 24 7.714/12 6.592/9.197 0.690/0.727 0.768/0.836 0.363

Golgol Microsat/MHC 24 7.143/9 5.906/7.409 0.678/0.792 0.716/0.822 0.216

Cendoya Microsat/MHC 24 4.857/5 4.283/4.988 0.554/0.792 0.588/0.794 0.004

N sample size, K number of observed alleles, AR allelic richness (based on seven diploid individuals), Ho observed heterozygosity, He expectedheterozygosity, J’ evenness population admixture index

Immunogenetics

variability in brown trout with that of a 237-base pair fragmentof the exon 2 of the MHC class II β previously amplified in208 rainbow trout from 10 populations (Fig. 1; Table 1) asdetailed in Monzón-Argüello et al. (2013). Microsatellite datapreviously published for both species were used as a baselinefor comparisons between neutral (microsatellites) and func-tional (MHC) markers (7 and 14 microsatellites for rainbowtrout and brown trout, respectively, Consuegra et al. 2011;Monzon-Arguello et al., under revision and stored in Figshareunder DOI http://dx.doi.org/10.6084/m9.figshare.953191).Rainbow trout had been classified as farm escapees,naturalized or hybrids based on their microsatellitegenotypes (Consuegra et al. 2011).

MHC class II β variability and tests for selection

Observed heterozygosity (Ho) and unbiased expected hetero-zygosity (He) were assessed using GENETIX 4.05 (Belkhiret al. 2004). Deviations from Hardy-Weinberg (HW) equilib-rium following sequential Bonferroni correction (Rice 1989)were estimated using Arlequin 3.5 (Excoffier and Lischer2010). The Ewens-Watterson homozygosity test of neutrality(Ewens 1972; Watterson 1978) with Slatkin’s exact P values(Slatkin 1994; Slatkin 1996) was used to assess deviationsfrom the hypothesis of neutral selection at the MHC locus.The relationship between population diversity (AR—allelicrichness and Ho) at MHC and neutral microsatellites wasinvestigated by using the Pearson correlation coefficient, aftertesting for normality using SPSS v.19. We analysed the cor-relation of genetic distance between populations, measured aspairwise FST, between microsatellites and MHC using aMantel test implemented in Arlequin. A Mantel test was alsoused to analyse isolation by distance (IBD). In both species,AR and population differentiation (FST) were determinedusing FSTAT 2.9.3 (Goudet 1995). To investigate populationstructure, we used the model-based clustering method imple-mented in STRUCTURE 2.3.3 (Pritchard et al. 2000). Foreach K (1–10), we computed 10 iterations with a burn in of20,000 and 80,000 MCMC replicates using the admixturemodel with allele frequencies correlated. To assess the mostlikely number of clusters, we calculated ΔK followingEvanno et al. (2005).

We estimated the dissimilarity between MHC class II βalleles within individuals using the number of non-synonymous substitutions per non-synonymous site (Ka) inDnaSP 5.10 (Librado and Rozas 2009), and population ob-served Ka was compared against random Ka, calculated as inConsuegra and Garcia de Leaniz (2008). Differences in meanKa between populations were compared using one-wayANOVA.

Evidence for selection was assessed by using three differ-ent codon-based maximum likelihood approaches. CODEMLimplemented in PAML v4.6 (Yang 2007), was used to

estimate ω, which is the ratio of the non-synonymous (dN)to synonymous substitutions (dS) over the entire set of se-quences. We compared models that only consider neutral anddeleterious non-synonymous mutations (M1 and M7) withmodels that allow for selection (M2, M3 and M8). Nestedmodels (M8 and M7; M2, M3 and M1) were compared with alikelihood ratio test (LRT) (Yang and Swanson 2002), and anAkaike information criterion (AIC) was used to compare non-nested models. Maximum likelihood trees for the analysiswere built using DNAML from PHYLIP (Felsenstein 1995).A Bayesian approach implemented in CODEML was used toidentify residues under positive selection in the β domain, andsites with a posterior probability >95 % were considered to bepositively selected under the model that best fitted the data. Inaddition, we employed two recent methods implemented inHyPhy (http://www.datamonkey.org/) (Wayne et al. 2010),the mixed effects model of evolution (MEME) (Murrellet al. 2012) which detects both episodic and pervasive positiveselection at individual sites and the Fast Unbiased BayesianApproximation (FUBAR) that detects positive selection undera model that allows site-to-site rate variation (Murrell et al.2013). We only considered sites that attained a significancelevel of 0.05 in MEME and 0.95 in FURBAR. In order toavoid the potential confounding effect of recombination(Shriner et al. 2003), we first examined the presence of intra-genic recombination using the GARD analysis also imple-mented in DataMonkey.

Changes in the amino acid sequence of the MHC bindingpockets can alter their binding properties and affect the rangeof pathogen peptides an individual can respond to(Schwensow et al. 2007). In order to take this into account,MHC class II β sequences were clustered into supertypesbased on the amino acid sequence properties of all positivelyselected sites (PSS) as detailed in Ellison et al. (2012), basedon the procedure of Doytchinova et al. (2005).

Results

MHC variability

A total of 40 MHC class II β alleles were identified in browntrout (19 in Chile and 26 in the Falklands; Figure S1; Table 1);33 of which represent novel sequences (GenBank AccessionNo. JX646900 – JX646932). After sequential Bonferronicorrection, the MHC class II β locus deviated significantlyfromHWexpectations in only one population, Estancia Brook(Ho=0.810; He=0.965; P=0.002; Table 1). The Ewens-Watterson test following sequential Bonferroni correctionrejected the null hypothesis of neutrality only in EstanciaBrook (Falkland Islands) (Slatkin exact P=0.005). All rain-bow trout populations were in HWequilibrium. There were nosignificant differences in MHC AR or Ho between rainbow

Immunogenetics

and brown trout (P=0.238 and P=0.158, respectively).Non-significant associations between spatial and MHC geneticdistances appear inconsistent with a pattern of isolation bydistance in both species in Chile (brown trout, P=0.406;rainbow trout, P=0.377). In contrast, there was a highlysignificant correlation between genetic distance at microsatel-lite loci and at the MHC class II β locus in Chileanpopulations (brown trout, r=0.676, P=0.023; rainbowtrout, r=0.492; P=0.010).

Population differentiation in Chile (FST) was higher amongbrown trout than among rainbow trout populations for theMHC class IIβ locus (brown trout FST=0.084 versus rainbowtrout FST=0.070; P<0.001). There was a significant degree ofpopulation differentiation among all Chilean and Falklandbrown trout populations (FST=0.181, P<0.001), with browntrout populations within the Falklands displaying the highestlevel of differentiation (FST=0.194, P<0.001).

STRUCTURE analyses indicated that Chilean rainbowtrout were fairly uniform in MHC genotype, with no evidenceof population structuring (Figure S2A). Similarly, there wasno obvious structuring within Chilean brown trout popula-tions (Figure S2B), whereas brown trout from the Falklandsshowed a pronounced differentiation with two of the popula-tions (Finlay Creek and Sarnys Creek) grouping together andthe third one (Estancia Brook) more closely associated to theChilean populations (Figure S3).

MHC allelic dissimilarity (measured as the mean rate ofnon-synonymous substitutions, Ka) was significantly lowerthan the random expectation in four of the Chilean and twoof the Falkland brown trout populations (Chile mean Ka=0.098, n=103, 95 % confidence interval (CI)=0.011;Falkland mean Ka=0.087, n=48, 95 % CI=0.017; P<0.010;Fig. 2 and Figure S3). In the remaining ones, Ka values werehigher than expected by random in one case (Encanto meanKa=0.134; 95 % CI=0.019) and non-significantly differentfrom random expectation in two others (Blanco-Enco meanKa=0.123; 95 % CI=0.026; Estancia Brook mean Ka=0.110;95 % CI=0.027). The mean rate of dissimilarity (Ka) was notsignificantly different between Chilean and Falkland browntrout populations (P=0.461).

Based on a similar fragment size (72 sites in rainbow troutand 81 in brown trout), rainbow trout displayed lower MHCallelic dissimilarity than brown trout populations (P<0.001);however, as inmost of the brown trout populations,MHC allelicdissimilarity was significantly lower than random expectation innine of the ten rainbow trout populations (mean Ka=0.060, n=208, 95 % CI=0.005; Fig. 2 and Figure S4), with the exceptionof the river Encanto (mean Ka=0.074; 95 % CI=0.012).

Cluster analysis revealed 10 consensus brown troutsupertypes (based on amino acid sequences), each possessingbetween 1 and 6 alleles (Fig. 3). Supertypes and theirbootstrapping values were consistent among all methods used(Figure S5). In general, clusters included private alleles from

both Chile and Falkland Islands, but some supertypesconsisted basically of unique alleles from one region (e.g.supertype 1 comprised mostly private alleles from Falklands,while supertype 3 was mainly made of private alleles fromChile), resulting in significant differences in supertype com-position among populations (FST=0.168,P<0.001). All of thebrown trout from two of the populations in the Falklands(Finlay Creek and Sarnys Creek) carried alleles fromsupertypes 2, 5 or both, whereas the third population(Estancia Brook) displayed a higher diversity of supertypesmore similar to the Chilean populations. Thirty percent of theindividuals carried alleles from the same supertype, 50 % ofthem with both alleles belonging to supertypes 5 or 7.

In rainbow trout, cluster analysis revealed six consensussupertypes, each possessing between 2 and 9 alleles (Fig. 4).Supertypes and their bootstrapping values were also consis-tent among all methods used (Figure S6), and the distributionof supertypes in rainbow trout differed significantly among

Falklands Brown Trout

Chilean Rainbow Trout

a

b

c

Fig. 2 MHC class II β dissimilarity (Ka; indicated by arrows) of aChilean brown trout, b Falkland brown trout and c Chilean rainbow troutcompared to random expectations based on 100 permutations of MHCallelic frequencies in each group

Immunogenetics

populations (FST=0.040; P<0.001). As for brown trout, 32 %of the individuals carried alleles from the same supertype,with a clear predominance of supertype 4 among those (65 %).

Signatures of selection

Several codons of the MHC class II β locus were identified asbeing under positive selection using three different methods.GARD identified breaking points in codon 65 of brown troutand 103 of rainbow trout. In Chilean brown trout, maximumlikelihood models that allow for positive selection fitted thedata significantly better than those that assume only neutral orconserved mutations (Table S1A). AIC suggested that modelM2, which allows for positive selection, fitted the data betterthan the rest of the models and identified 21% of sites as beingunder positive selection (ω=7.72). All three methods werecoincident in identifying three sites under selection in browntrout 66, 77 and 80 in Chile (Table 2(A)). In the Falkland

brown trout, M3 model, which assumes three site classes,fitted the data significantly better than the rest (Table S1B).In this case, the three methods were coincident in identifyingfive sites under selection (codons 8, 20, 74, 77, 80). None ofthese codons was identified as a potential recombinationbreakpoint, and two of them were coincident with thoseidentified in Chilean brown trout (Table 2(B)).

In rainbow trout, the M3 model fitted the data significantlybetter than the others (Table S1C). Estimates from M3 iden-tified 16 % of the sites as being under positive selection in thesequences (ω=22.56). Codons 35, 47 and 53 (Table 2(C))were identified by all methods as being under selection.

Discussion

Species introduced into novel environments are often freefrom their natural pathogens fairly rapidly (Mitchell and

Fig. 3 Phenetic tree based on a cluster analysis (Ward’s algorithm) defining MHC brown trout supertypes. Asterisk and double asterisks show privatealleles in Chile and the Falkland Island populations, respectively

Immunogenetics

Power 2003; Torchin et al. 2003), and it usually takes a longerperiod of time for new pathogens to become established (Leeand Klasing 2004). During the initial invasion stages, hence,invaders can benefit from lower pathogen loads compared tothose of conspecifics living within the natural range (Cornelland Hawkins 1993; Kennedy and Pojmanska 1996). We thushypothesized that successful invaders would display highimmunogenetic diversity at genes related to the cell-mediated response (such as the MHC genes) and, due to thecombined founder effects and/or potential new waves of in-vaders, high population differentiation. We tested this hypoth-esis by comparingMHC class IIβ genetic diversity in two co-occurring invasive salmonids from Chile and the Falklands(brown trout and rainbow trout), which are ecologically sim-ilar but have different introduction histories. We found levelsof MHC class II β variation for brown and rainbow troutsimilar or greater than those reported for natural populationsat their native range (e.g. Aguilar and Garza 2006). Multipleintroductions from several sources might have been able toovercome the potential effects of founder events by introduc-ing new genetic diversity (Consuegra et al. 2011; Monzón-Argüello et al. 2013). However, we also found evidence of

lower diversity than expected by random in most populationsof both species at the functional level (measured as amino acidsimilarity between alleles), potentially caused by a decrease indiversity related to the immune response. In theory, successfulinvaders could display reduced immune activity, in particularthat associated with systemic inflammation (including MHC-mediated T cell immunity), and reallocate costly energy re-sources to other processes such as growth and reproduction(Lee and Klasing 2004). Moreover, at least 30 % of individ-uals of both species possessed alleles belonging to the samesupertype, with one supertype being clearly predominant inrainbow trout and two supertypes being predominant in browntrout.

Low allelic dissimilarity could also be indicative ofassortative MHC mating (i.e. reproductive pairing ofindividuals genetically more similar at the MHC classII β locus than would be expected by random mating) orreflect selective pressures of new pathogens encounteredin the new environment. Assortative mating can play arole in sympatric speciation, contributing to pre-matingisolation, and also in the genetic isolation of populationswhen they come into secondary contact (Bolnick and

Fig. 4 Phenetic tree based on a cluster analysis (Ward’s algorithm) defining MHC rainbow trout supertypes

Immunogenetics

Kirkpatrick 2012). MHC-related disassortative matinghas been observed in a number of species, including housemouse, humans and salmonids (Consuegra and Garcia deLeaniz 2008; Mays and Hill 2004; Penn and Potts 1998),although the difference between disassortative mating andmating for heterozygosity does not seem completely clear(Bonneaud et al. 2006a; Roberts et al. 2006). In contrast,examples of MHC-mediated assortative mating are less abun-dant (Roberts et al. 2005) and none of them in salmonids.Testing for assortative mating was outside the remit of this studyand would require experimental evidence, but in any case, lowdissimilarity seems to contrast with most MHC studies whereallele dissimilarity has been commonly identified as a signatureof balancing selection (Bernatchez and Landry 2003).

We found little evidence of population structuring amongChilean brown trout or rainbow trout in relation to MHC.These results contrast with the high level of structuring previ-ously observed within the same populations using neutralmarkers (microsatellites): very admixed rainbow trout popu-lations (Consuegra et al. 2011) and very structured and

differentiated brown trout populations (Monzon-Arguelloet al. in review). Although the difference could be the resultof using a singleMHCmarker, using the samemarker, we stillobserved a clear structuring between brown trout populationsfrom the Falklands and Chile and among the three populationsfrom the Falklands. For both species, genetic differentiation atthe MHC class II β gene was correlated with neutral FST,suggesting a role of neutral evolutionary processes in currentpopulations. However, there was also evidence of selectionacting on the PBR of the MHC class IIβ in both species whenrates of non-synonymous versus synonymous substitutionswere considered, and a number of sites appear to be clearlyunder selection using three different methods. These signa-tures were fairly consistent between geographical regions inthe case of brown trout, suggesting that they may correspondto signatures of selection generated in the original populationsbefore they were introduced in the southern hemisphere. Thiscould be because significant dN-to-dS ratios take a long time toaccumulate, but they may also take an equally long time todisappear in the absence of selection, not necessarilyreflecting the effect of current selective pressures (Garriganand Hedrick 2003).

Finally, we found very uniform patterns of MHC distribu-tion in both species, contrasting with the differences found forneutral markers, where brown trout populations were highlystructured whereas rainbow trout displayed high levels ofadmixture (Consuegra et al. 2011; Monzon-Arguello et al.under review). This pattern could be explained by the purgingof MHC alleles present in fish farms after rainbow troutescape into the wild, as suggested by our previous analyses(Monzón-Argüello et al. 2013). Thus, we did not find evi-dence of an effect of new invasion waves, in the form ofaquaculture releases, in the immunogenetic structure of therainbow trout populations. Instead, we found a similar patternof reduced functional MHC diversity in both trout species.

The eco-immunology of invasions is an emerging field inneed of empirical studies that consider the genetic aspectsunderlying the immune response, particularly consideringthe overarching influence of founder effects for the successfulestablishment of invaders (White and Perkins 2012). Weanalysed the diversity of MHC class II β in two contrastingsalmonid invaders introduced into the southern hemisphereand potentially liberated of their natural enemies. MHC classII β genetic diversity did not appear to be reduced in terms ofallelic richness or heterozygosity in any of the two species.However, we found evidence of low functional MHC diver-sity (measured as amino acid dissimilarity) in most popula-tions of both species, high percentage of individuals with twoalleles from the same functionally similar supertype and lowerpopulation genetic structuring than that observed at neutralmarkers, suggesting a potential reduction in MHC functionaldiversity that could reflect either a decrease in cell immuneresponse, assortative mating or new pathogen pressures, all

Table 2 Results from three different codon-based maximum likelihoodapproaches; mixed effects model of evolution (MEME), Fast UnbiasedBayesian Approximation (FUBAR) and the model in CODEML that bestfit the data to estimate positive selected sites along the MHC class II βdomain in Chilean brown trout (A), Falkland brown trout (B) and Chileanrainbow trout (C)

Model Positively selected sites

(A) Chilean brown trout

MEME 52, 57, 63, 66, 77, 80

FURBAR 8, 12, 31, 33, 66, 74, 77, 80

M2 (Positiveselection)

4Y**, 6R*, 8A**, 22 L*, 31A**, 33Y**, 52 K*,63I**, 66Q**, 74Y**, 77P**, 80D**, 81I**,82D**

(B) Falkland brown trout

MEME 8, 20, 27, 52, 66, 74, 77, 80, 82

FURBAR 8, 20, 31, 34, 52, 74, 77, 80, 81, 82

M3 (Discrete) 4E**, 5Q**, 6 V**, 7 V**, 8R**, 9Q**,11R** ,12 F**, 19G**, 20I**, 22 F**,24D**, 27 V**, 30 K**, 31A**, 33Y**,34 V**, 43Y**, 49H**, 52 K**, 57 W**,61G**, 62P**, 63E**, 66Q**, 67E**,68 L**, 70E**, 74 V**, 77G**, 78 N**,80A**, 81I**, 82D**

(C) Chilean Rainbow Trout

MEME 8, 35, 47, 53

FURBAR 4, 6, 17, 27, 35, 47, 53, 58, 61, 66

M3 (Discrete) 4I**, 6 F**, 7I*, 8D**, 11 V**, 14 K**,15 V**, 17H**, 18I**, 27Y**, 35 V*,41 W**, 47 L**, 53G**, 54E**, 57S**,58Y**, 61H**, 63A**, 64D**, 65I**, 66Y**

In bold are the sites identified as being under positive selection by thethree methods

(*P<0.05, **P<0.01)

Immunogenetics

hypotheses that deserve further experimental studies. Therelationship between MHC diversity and the long-term per-sistence of small populations is unclear (Radwan et al. 2010).While some species seem to thrive even after severe bottle-necks have depleted their MHC diversity, it could be that thesespecies only represent rare examples that survived despite theloss of MHC variation (Radwan et al. 2010). Given thepotential importance of host-parasite relationships for theestablishment and long-term persistence of invasive species,we suggest that a better understanding of the eco-immunologyof invaders may help in managing and preventing the impactof biological invasions.

Acknowledgments We thank Kyle Young, Hector Venegas, PatriciaBeristain, Jose Sanzana, Anita Cerda, Gabriel Orellana, DelphineVanhaecke and several volunteers for collecting the samples in Chile;Amy Ellison, Kirsten Skot and Candida Nibau for the help with labora-tory analyses; and Nuria Varo for the statistical advice. Funding for thisstudy was provided by the DEFRA Darwin Initiative ‘Reducing theImpact of Exotic Aquaculture on Chilean Aquatic Biodiversity’ (GrantNo. 162/15/020) and a post-project award ‘Protecting galaxiids fromsalmonids invasions in Chile and the Falklands’ (Grant No. EIDPOC041; http://www.biodiversity.cl) to CGL, GG and SC with additionalsupport from the University of Los Lagos (Chile). CMAwas funded bythe Fundación Alfonso Martín Escudero (Spain).

References

Aguilar A, Garza JC (2006) A comparison of variability and populationstructure for major histocompatibility complex and microsatelliteloci in California coastal steelhead (Oncorhynchus mykissWalbaum). Mol Ecol 15:923–937

Apanius V, Penn D, Slev PR, Ruff LR, Potts WK (1997) The nature ofselection on the major histocompatibility complex. Crit RevImmunol 17:179–224

Arrowsmith E, Pentelow FTK (1965) The introduction of trout andsalmon to the Falkland Islands. Salmon Trout Mag 174:119–129

Basulto S (2003) El largo viaje de los salmones: una crónica olvidada,propagación y cultivo de especies acuáticas en Chile, p. 299.Editorial Maval, Santiago de Chile

Belkhir K, Borsa P, Chikhi L, Raufaste N, Bonhomme F (2004)GENETIX 4.05, logiciel sous Windows TM pour la genetique despopulations. Laboratoire Genome, Populations, Universite deMontpellier II, Montpellier, France

Bernatchez L, Landry C (2003) MHC studies in nonmodel vertebrates:what have we learned about natural selection in 15 years? J EvolBiol 16:363–377

Bolnick DI, Kirkpatrick M (2012) The relationship between intraspecificassortative mating and reproductive isolation between divergentpopulations. Curr Zool 58:484–492

Bonneaud C, Chastel O, Federici P, Westerdahl H, Sorci G (2006a)Complex MHC-based mate choice in a wild passerine. Proc R SocB Biol Sci 273:1111–1116

Bonneaud C, Pérez-Tris J, Federici P, Chastel O, Sorci G (2006b) Majorhistocompatibility alleles associated with local resistance to malariain a passerine. Evolution 60:383–389

Buswell JM, Moles AT, Hartley S (2011) Is rapid evolution common inintroduced plant species? J Ecol 99:214–224

Butchart SHM, Walpole M, Collen B et al (2010) Global biodiversity:indicators of recent declines. Science 328:1164–1168

Carroll SP (2007) Natives adapting to invasive species: ecology,genes, and the sustainability of conservation. Ecol Res 22:892–901

Colautti RI, Ricciardi A, Grigorovich IA, MacIsaac HJ (2004) Is invasionsuccess explained by the enemy release hypothesis? Ecol Lett 7:721–733

Consuegra S, Garcia de Leaniz C (2008) MHC-mediated mate choiceincreases parasite resistance in salmon. Proc R Soc B Biol Sci 275:1397–1403

Consuegra S, Phillips N, Gajardo G, Garcia de Leaniz C (2011) Winningthe invasion roulette: escapes from fish farms increase admixtureand facilitate establishment of non-native rainbow trout. Evol Appl4:660–671

Cornell HV, Hawkins BA (1993) Accumulation of native parasitoidspecies on introduced herbivores: a comparison of hosts as nativesand hosts as invaders. Am Nat 141:847–865

Correa C, Hendry AP (2012) Invasive salmonids and lake order interactin the decline of puye grande Galaxias platei in western Patagonialakes. Ecol Appl 22:828–842

Doherty PC, Zinkernagel RM (1975) Enhanced immunological surveil-lance in mice heterozygous at the H-2 gene complex. Nature 256:50–52

Doytchinova IA, Walshe V, Borrow P, Flower DR (2005) Towards thechemometric dissection of peptide–HLA-A*0201 binding affinity:comparison of local and global QSARmodels. J Comput AidedMolDes 19:203–212

Eizaguirre C, Lenz TL, Kalbe M, Milinski M (2012) Rapid and adaptiveevolution of MHC genes under parasite selection in experimentalvertebrate populations. Nat Commun 3:621

Ellison A, Allainguillaume J, Girdwood S et al (2012) Maintainingfunctional major histocompatibility complex diversity under in-breeding: the case of a selfing vertebrate. Proc R Soc B Biol Sci279:5004–5013

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clustersof individuals using the software STRUCTURE: a simulation study.Mol Ecol 14:2611–2620

Ewens WJ (1972) The sampling theory of selectively neutral alleles.Theor Popul Biol 3:87–112

Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series ofprograms to perform population genetics analyses under Linux andWindows. Mol Ecol Resour 10:564–567

Felsenstein J (1995) PHYLIP (Phylogeny Inference Package)Department of Genetics. University of Washington, Seattle

Froese R, Pauly D (Eds.) (2013) FishBase. World Wide Web electronicpublication.www.fishbase.org. Accessed Feb 2014

Gajardo G, Laikre L (2003) Chilean aquaculture boom is based on exoticsalmon resources: a conservation paradox. Cons Biol 17:1173–1174

Garcia de Leaniz C, Gajardo G, Consuegra S (2010) From best to pest:changing perspectives on the impact of exotic salmonids in thesouthern hemisphere. Syst Biodivers 8:447–459

Garrigan D, Hedrick PW (2003) Detecting adaptive molecular polymor-phism: lessons from the MHC. Evolution 57:1707–1722

Gómez D, Conejeros P, Marshall S, Consuegra S (2010) MHC evolutionin three salmonid species: a comparison between class II alpha andbeta genes. Immunogenetics 62:531–542

Goudet J (1995) FSTAT (version 1.2): a computer program to calculate F-statistics. J Hered 86:485–486

Hall T (1999) BioEdit: a user-friendly biological sequence alignmenteditor and analysis program for Windows 95/98/NT. NucleicAcids Symp Ser 41:95–98

Hedrick P, Kim T (2000) Genetics of complex polymorphisms: parasitesand maintenance of MHC variation. Harvard University Press,Cambridge

Hendry AP, Farrugia TJ, Kinnison MT (2008) Human influences on ratesof phenotypic change in wild animal populations. Mol Ecol 17:20–29

Immunogenetics

Hughes AL, Yeager M (1998) Natural selection at major histocompati-bility complex loci of vertebrates. Ann Rev Genet 32:415–435

Janeway CA, Travers P, Walport D, Shlomchik MJ (2004)Immunobiology: the immune system in health and disease, vol 1,(6th ed.). Garland Science: New York.

Keane RM, Crawley MJ (2002) Exotic plant invasions and the enemyrelease hypothesis. Trends Ecol Evol 17:164–170

Kennedy CR, Pojmanska T (1996) Richness and diversity of helminthparasite communities in the common carp and in three more recentlyintroduced carp species. J Fish Biol 48:89–100

Kurtz J, Kalbe M, Aeschlimann P et al (2004) Major histocompatibilitycomplex diversity influences parasite resistance and innate immuni-ty in sticklebacks. Proc R Soc B Biol Sci 271:197–204

Landry C, Garant D, Duchesne P, Bernatchez L (2001) “Good genes asheterozygosity”: the major histocompatibility complex and matechoice in Atlantic salmon (Salmo salar). Proc R Soc B Biol Sci268:1279–1285

Lee KA, Klasing KC (2004) A role for immunology in invasion biology.Trends Ecol Evol 19:523–529

Librado P, Rozas J (2009) DnaSP v5: a software for comprehensiveanalysis of DNA polymorphism data. Bioinformatics 25:1451

Lowe S, BrowneM, Boudjelas S (2000) 100 of the world’s worst invasivealien species. In: IUCN/SSC Invasive Species Specialist Group,Auckland

Mays HL Jr, Hill GE (2004) Choosing mates: good genes versus genesthat are a good fit. Trends Ecol Evol 19:554–559

Mitchell CE, Power AG (2003) Release of invasive plants from fungaland viral pathogens. Nature 421:625–627

Molnar JL, Gamboa RL, Revenga C, Spalding MD (2008) Assessing theglobal threat of invasive species to marine biodiversity. Front EcolEnviron 6:485–492

Monzón-Argüello C, Garcia de Leaniz C, Gajardo G, Consuegra S (2013)Less can be more: loss of MHC functional diversity can reflectadaptation to novel conditions during fish invasions. Ecol Evol 3:3359–3368

Murrell B, Wertheim JO, Moola S et al (2012) Detecting individual sitessubject to episodic diversifying selection. PLoS Genet 8:e1002764

Murrell B, Moola S, Mabona A et al (2013) FUBAR : A Fast,Unconstrained Bayesian AppRoximation for inferring selection.Mol Biol Evol 30:1196–1205

Naylor RL, Williams SL, Strong DR (2001) Aquaculture—a gateway forexotic species. Science 294:1655–1656

Olsen KH, Grahn M, Lohm J, Langefors A (1998) MHC and kindiscrimination in juvenile Arctic charr, Salvelinus alpinus (L.).Anim Behav 56:319–327

Penn D, Potts WK (1998) Chemical signals and parasite-mediated sexualselection. Trends Ecol Evol 13:391–396

Pimentel D, Zuniga R, Morrison D (2005) Update on the environmentaland economic costs associated with alien-invasive species in theUnited States. Ecol Econ 52:273–288

Pritchard JK, Stephens M, Donnelly P (2000) Inference of populationstructure using multilocus genotype data. Genetics 155:945–959

Radwan J, Biedrzycka A, Babik W (2010) Does reduced MHC diversitydecrease viability of vertebrate populations? Biol Cons 143:537–544

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43:223–225Roberts SC, Little AC, Gosling LM et al (2005) MHC-assortative facial

preferences in humans. Biol Lett 1:400–403

Roberts SC, Hale ML, Petrie M (2006) Correlations between heterozy-gosity and measures of genetic similarity: implications for under-standing mate choice. J Evol Biol 19:558–569

Schröder V, Garcia de Leaniz C (2011) Discrimination between farmedand free-living invasive salmonids in Chilean Patagonia using stableisotope analysis. Biol Invas 13:203–213

Schwensow N, Fietz J, Dausmann KH, Sommer S (2007) Neutral versusadaptive genetic variation in parasite resistance: importance ofmajorhistocompatibility complex supertypes in a free-ranging primate.Heredity 99:265–277

Shriner D, Nickle DC, Jensen MA, Mullins JI (2003) Potential impact ofrecombination on sitewise approaches for detecting positive naturalselection. Genet Res 81:115–121

Simberloff D, Martin J-L, Genovesi P et al (2013) Impacts of biologicalinvasions: what’s what and the way forward. Trends Ecol Evol 28:58–66

Slade RW, McCallum HI (1992) Overdominant vs. frequency-dependentselection at MHC loci. Genetics 132:861–864

Slatkin M (1994) An exact test for neutrality based on the Ewenssampling distribution. Genet Res 64:71–74

Slatkin M (1996) A correction to the exact test based on the Ewenssampling distribution. Genet Res 68:259–260

Strayer DL, Eviner VT, Jeschke JM, Pace ML (2006) Understanding thelong-term effects of species invasions. Trends Ecol Evol 21:645–651

Torchin ME, Mitchell CE (2004) Parasites, pathogens, and invasions byplants and animals. Front Ecol Environ 2:183–190

Torchin ME, Lafferty KD, Dobson AP, McKenzie VJ, Kuris AM (2003)Introduced species and their missing parasites. Nature 421:628–630

Watterson GA (1978) The homozygosity test of neutrality. Genetics 88:405–417

Wayne D, Poon AF, Frost SDW, Pond SLK (2010) Datamonkey 2010: asuite of phylogenetic analysis tools for evolutionary biology.Bioinformatics 26:2455–2457

Wegner KM, Reusch TBH, Kalbe M (2003) Multiple parasites aredriving major histocompatibility complex polymorphism in thewild. J Evol Biol 16:224–232

White TA, Perkins SE (2012) The ecoimmunology of invasive species.Funct Ecol 26:1313–1323

Whitney KD, Gabler CA (2008) Rapid evolution in introduced species,‘invasive traits’ and recipient communities: challenges for predictinginvasive potential. Divers Distrib 14:569–580

Yang ZH (2007) PAML 4: Phylogenetic analysis by maximum likeli-hood. Mol Biol Evol 24:1586–1591

Yang Z, SwansonW (2002) Codon-substitutionmodels to detect adaptiveevolution that account for heterogeneous selective pressures amongsite classes. Mol Biol Evol 19:49–57

Young KA, Dunham JB, Stephenson JF et al (2010) A trial of two trouts:comparing the impacts of rainbow and brown trout on a nativegalaxiid. Anim Conserv 13:399–41

Data accessibility

Raw data on microsatellites have been stored in Figshare (http://dx.doi.org/10.6084/m9.figshare.953191) and will be made accessible throughFigshare when the paper is accepted.

Immunogenetics