Behav Genet (2015) 45:236–244 DOI 10.1007/s10519-014-9701-7
ORIGINAL RESEARCH
Rhodopsin Gene Polymorphism Associated with Divergent Light Environments in Atlantic Cod Christophe Pampoulie • Sigurlaug Skirnisdottir • Bastiaan Star • Sissel Jentoft • Ingibjo¨rg G. Jo´nsdo´ttir • Einar Hjo¨rleifsson • Vilhja´lmur Thorsteinsson • ´ lafur K. Pa´lsson • Paul R. Berg • Øivind Andersen • Steinunn Magnusdottir • O Sarah J. Helyar • Anna K. Danı´elsdo´ttir
Received: 16 September 2014 / Accepted: 15 December 2014 / Published online: 11 January 2015 Ó Springer Science+Business Media New York 2015
Abstract The spectral sensitivity of visual pigments in vertebrate eyes is optimized for specific light conditions. One of such pigments, rhodopsin (RH1), mediates dimlight vision. Amino acid replacements at tuning sites may alter spectral sensitivity, providing a mechanism to adapt to ambient light conditions and depth of habitat in fish. Here we present a first investigation of RH1 gene polymorphism among two ecotypes of Atlantic cod in Icelandic waters, which experience divergent light environments throughout the year due to alternative foraging behaviour. We identified one synonymous single nucleotide polymorphism (SNP) in the RH1 protein coding region and one in the 30 untranslated region (30 -UTR) that are strongly divergent between these two ecotypes. Moreover, these polymorphisms coincided with the well-known panthophysin (Pan I) polymorphism that differentiates coastal and frontal Electronic supplementary material The online version of this article (doi:10.1007/s10519-014-9701-7) contains supplementary material, which is available to authorized users. Edited by Stephen Maxson. C. Pampoulie (&) I. G. Jo´nsdo´ttir E. Hjo¨rleifsson ´ . K. Pa´lsson V. Thorsteinsson O Marine Research Institute, Sku´lagata 4, 101 Reykjavı´k, Iceland e-mail:
[email protected] S. Skirnisdottir S. Magnusdottir S. J. Helyar A. K. Danı´elsdo´ttir Matis Ltd., Vinlandsleid 12, 113 Reykjavı´k, Iceland B. Star S. Jentoft P. R. Berg Department of Biosciences, Centre for Ecological and Evolutionary Synthesis (CEES), University of Oslo, Oslo, Norway Ø. Andersen ˚ s, Norway NOFIMA, Osloveien 1, 1340 A
123
(migratory) populations of Atlantic cod. While the RH1 SNPs do not provide direct inference for a specific molecular mechanism, their association with this dimsensitive pigment indicates the involvement of the visual system in local adaptation of Atlantic cod. Keywords Atlantic cod Rhodopsin Pantophysin Behaviour type Ecotype Divergence
Introduction Water depth and turbidity modify ambient light spectra through various processes (Jerlov 1976; Thurman and Trujillo 2004). Light is refracted or reflected at the surface and is scattered or absorbed by solid particles within the water column. Particular wavelengths are more effectively scattered and absorbed by water, and a large proportion of light is absorbed within the first few meters. Thus, only one percent of the visible light (400 \ k \ 700 nm; Michiels et al. 2008) reaches depths below 150 m (Jerlov 1976; Michiels et al. 2008; Tyler 2003) of which most of the long wavelengths (red light; 600–700 nm) are absorbed within the first 10 m of the water column. Only short wave length light spectra (blue–green light; 400–500 nm) can be perceived as deep as 1,000 m depth (Jerlov 1976; Michiels et al. 2008; Tyler 2003; Warrant and Locket 2004). Importantly, these processes provide a highly stratified light environment within marine and freshwater ecosystems. A substantial body of literature provides evidence that fish have adapted to such stratified environments by modifying the protein structure of visual pigments responsible for the perception of colours and bright light (cone receptors) as well as of dim-light rod photoreceptors (Ebert and Andrew 2009; Larmuseau et al. 2010; Nakamura et al.
Behav Genet (2015) 45:236–244
2013; Sivasundar and Palumbi 2010; Spady et al. 2005; Yokoyama et al. 2008). In the rhodopsin pigment, multiple amino acid (AA) substitutions affecting the peak spectral sensitivity are found among different teleost species, so that the range of absorbed spectra (kmax of 480–525 nm) is broader in the derived forms compared to the ancestral forms of rhodopsin (kmax of 500 nm) (Yokoyama et al. 2008). In the Pacific bluefin tuna (Thunnus orientalis), a blue-shift caused by a E122Q substitution seems to provide a better detection of blue–green contrast and measure of prey distance in the blue-pelagic ocean (Nakamura et al. 2013). Light conditions may also be an important factor in driving speciation (Shum et al. 2014; Sivasundar and Palumbi 2010). In the rockfish family (Sebastes spp.) of the Pacific Ocean, several species exhibit an AA replacement at the rhodopsin gene attributable to their preference for deep habitats (Sivasundar and Palumbi 2010). Finally, even within a single species, polymorphisms in rhodopsin can be strongly correlated with the photic conditions of the environment. In the North Atlantic beaked redfish Sebastes mentella, the widely-distributed shallow-pelagic form found between 250 and 550 m depth exhibits a GTC codon at position 119, which codes for Valine, while the deeppelagic type found below 500 m displays an ATC, coding for Isoleucine (Shum et al. 2014). Sand goby populations inhabiting the turbid Baltic Sea display the AA substitution F261Y responsible for a red-shift, while populations from the Irish Sea, Bay of Biscay and Iberian peninsula display the A299S substitution responsible for a red-shift too (Larmuseau et al. 2010). Overall, these studies provide decisive evidence for the importance of protein modifications of the rhodopsin in marine vertebrates, reflecting a basic mechanism for ecological processes such as speciation and local adaptation (Ebert and Andrew 2009; Sivasundar and Palumbi 2010). The studies investigating intraspecific rhodopsin polymorphisms have focused on species that occupy relatively confined and shallow depth ranges (but see Shum et al. 2014). Therefore, the ubiquity of such polymorphisms within the visual system of species that exhibit dial vertical migration and seasonal vertical migrations on a larger spatial scale remain relatively unexplored. Based on data storage tags (DSTs), that have been used to investigate seasonal thermo-bathymetric migration patterns in fish (Grabowski et al. 2011; Pa´lsson and Thorsteinsson 2003; Pampoulie et al. 2008; Thorsteinsson et al. 2012), we have previously reported that different individuals of Atlantic cod (Gadus morhua) display divergent feeding behaviour in Icelandic waters. Specifically, individuals tagged at the same spawning ground, exhibit markedly different feeding migration behaviour. The coastal types stay in shallow waters throughout the year, whereas the frontal (migratory) types breed in shallow waters but migrate to deeper waters
237
to feed near thermal fronts (Pa´lsson and Thorsteinsson 2003; Thorsteinsson et al. 2012). These behaviour types, which were recently defined as ecotypes (Grabowski et al. 2011), display genotypic differences at the polymorphic pantophysin locus (Pan I) comprising the Pan IA and Pan IB alleles predominantly observed in the coastal and frontal type, respectively (Pampoulie et al. 2008). Genetic studies of the stationary Norwegian coastal cod and the migratory Northeast Arctic cod have demonstrated similar associations between behaviour and Pan I pattern (Godo and Michalsen 2000; Nordeide 1998; Sarvas and Fevolden 2005; Skarstein et al. 2007). In addition, recent studies investigating a wider range of SNP data have shown that this association is not limited to Pan I, but affects multiple genomic regions in linkage group 1 (Hemmer-Hansen et al. 2013; Karlsen et al. 2013). Interestingly, this linkage group 1 (LG1) also contains genes that, apart from the Pan I locus, may contribute to local adaptation of marine fishes to depth, such as the RH1 opsin gene (rhodopsin) (Hemmer-Hansen et al. 2013; Therkildsen et al. 2013), both known to be under positive selection (Hemmer-Hansen et al. 2013; Pogson and Mesa 2004; Therkildsen et al. 2013). No study, however, has yet investigated variation of the RH1 opsin gene in Atlantic cod. Importantly, extrapolating from the available data and the light absorbance in oceans, the behaviour types that are identified in Icelandic waters are clearly exposed to divergent light environments. Therefore, we investigate the relationship between the polymorphism in the RH1 opsin gene of Atlantic cod, the Pan I locus and behaviour types defined as coastal and frontal using DSTs profiles.
Materials and methods Data storage tags (DSTs) and behavioural classification All surgery was performed under pacific measures aiming at minimizing the suffering of individuals, i.e. tonic immobility or animal hypnoses caused by placing individual cod dorsal side down in tagging saddle, cooling by keeping ample sea water running over gills and side, and for protection of eyes blindfolding with a cloth soaked in seawater wrapped around head. After surgery, antibiotics (engemycin-oxytetracycli, Intervet-Vnr 39 78 85) and vitamin complex (Becoplex Vet.) were given to tagged individuals. The description of the protocols is available at the following website: http://www.hafro.is/skrar/flokkar/ merkingar_thorskur.pdf. Classification of individuals to behaviour types is based on well-described criteria (Grabowski et al. 2011; Pa´lsson and Thorsteinsson 2003; Pampoulie et al. 2008). In short, coastal and frontal behaviours are defined according to the
123
238
tagged individual annual temperature and depth history (Fig. 1). Individuals that spend at least 70 % of their time in shallow waters are classified as the coastal type. Their temperature profiles show an annual pattern; temperature rises towards a maximum in September–October and declines towards a minimum in February–March. Individuals that spend at least 70 % of their time in deeper waters are classified as the frontal type. These fish share the depth range of the coastal type during spawning migrations but migrate to deeper waters (250–600 m) during feeding. Their temperature and depth profiles show evidence for repeated vertical migrations and migration to thermal fronts through extreme temperature recordings found at such locations (\0 °C and [7 °C; see (Thorsteinsson et al. 2012)). A total of 148 individuals were available with DST information. Amplification of microsatellite loci and the Pan I locus Of the DST panel of 148 individuals, 46 were identified as frontal whereas 102 as coastal (see Fig. 1 for typical DSTs profiles), and thus selected for further genotyping at 26 microsatellite loci, namely PGmo38 and PGmo49 (Jakobsdo´ttir et al. 2006), PGmo61, PGmo64, PGmo69, PGmo71, PGmo74, PGmo87, PGmo94, PGmo95, PGmo97, PGmo100, PGmo104, PGmo105, PGmo124, PGmo127 and PGmo134 (Skı´rnisdo´ttir et al. 2008), Gmo2 (Brooker et al. 1994), Gmo8, Gmo19, Gmo34, and Gmo37 (Miller et al. 2000), and Tch5, Tch11, Tch14, and Tch22 Fig. 1 Typical coastal and frontal behaviour DSTs profiles of Atlantic cod. Depth is depicted in black, temperature in red
123
Behav Genet (2015) 45:236–244
(O’Reilly et al. 2000), and the Pan I locus (initially known as cDNA clone GM798, see (Fevolden and Pogson 1997; Pogson et al. 1995)). Primers were modified for four of these microsatellite loci, Gmo8 (forward primer: GAG GCATCTGTCATTCATTTAG), Gmo37 (forward primer TCGGCCTCAGAACATTTAGC and reverse primer TGGCACCGTGGGATACATGG), Tch5 (forward primer TTAATATCACGCACAAATTGCCC and reverse primer TCGCATTGAGCCTAGTTTAC), and Tch22 (forward primer CTCTCTCTGAATCCCTCTGTCTG and reverse primer CTGGCCAAGTTCAGCGG). DNA extraction, polymerase chain reactions (PCR) and genotyping processes for the microsatellite loci are described in (Jakobsdo´ttir et al. 2006; Skı´rnisdo´ttir et al. 2008) and (Pampoulie et al. 2006) with the modification of using the ´ lafsson et al. Teg DNA polymerase (Matis Ltd.) in PCR (O 2010). For the Pan I locus, PCR and genotyping analysis were performed according to (Stenvik et al. 2006). Samples were analysed on an ABI PRISM 3730 sequencer using the GeneScan-500 LIZ internal standard and genotyped with GeneMapper v4.1 (Applied Biosystems). Of the same panel of individuals, the RH1 opsin gene was successfully amplified in 142 specimens (99 coastal and 43 frontal). The RH1 opsin gene was amplified by using primers upstream and downstream of the gene. Primers used for amplification and sequencing of different sections of the RH1 opsin gene are listed in Supplementary Table S1. Primer Rh1039r is based on (Chen et al. 2003) but new forward and reverse primers were designed by
Behav Genet (2015) 45:236–244
using the Clone Manager Suite 7 program (Scientific & Educational Software) on conserved regions of alignment of the RH1 opsin gene. DNA amplifications were carried out in a total volume of 25 ll containing 10–50 ng DNA, 0.75 U of Taq DNA polymerase (New England BioLabs), 1 9 Standard buffer, 200 lM of dNTP, 0.8 lM of the forward primer and 0.8 lM of the reverse primer. The PCR thermal profile was as follows: 4 min at 94 °C followed by 30 cycles of 30 s at 94 °C, 30 s at 58 °C, 90 s at 68 °C, with a final elongation step of 7 min at 68 °C. PCR product cleanup was done by using the ExoSAP-IT kit according to the producer (Affymetrix, Inc; USB product), and the sequencing reactions were performed in both directions by using the BigDye Terminator 3.1 Cycle Sequencing Kit according to the manufacture’s protocol (Applied BioSystems). Sequencing of the fragments was done by using ABI 3730 sequencer (Applied BioSystems). The software Sequencher v5.1 (Gene Codes Corporation) was used to detect point mutations (SNPs) by alignments of all sequences and by using GenBank sequence AF385832 as a reference sequence. Statistical analyses of the microsatellite loci Genetic diversity of coastal and frontal behaviour type was evaluated using allele frequencies, observed (Ho) and unbiased expected heterozygosity (He) as well as Allele number (Ar) calculated in the GENEPOP’007 software (Rousset 2008). Deviations from Hardy–Weinberg expectation (HWE) were tested using Fishers methods (Weir and Cockerham 1984) implemented in GENEPOP and significance was determined with exact tests. We applied the coalescent-based simulation methods of Beaumont and Nichols (Beaumont and Nichols 1996) to detect potential outlier microsatellite loci (loci under selection). Coalescent simulations were performed with the software LOSITAN (Antao et al. 2008) with samples of the same size as the observed samples. A total of 100,000 independent loci were generated with the infinite allele mutation model and the ‘‘neutral’’ mean FST function (outliers loci were excluded to calculate the initial mean FST). Simulated distribution of FST values conditional to heterozygosity under a neutral model were obtained and thus compared to observed FST values to identify potential outlier loci. Genetic differentiation was estimated using theta estimates (h) (Weir and Cockerham 1984) implemented in GENEPOP, and significance was assessed using allelic and genotypic frequency homogeneity tests (5,000 permutations). STRUCTURE 2.3.2 (Pritchard et al. 2000) was used to assess the potential number of genetically distinguishable groups in our samples. The software was run using an admixture model without any a priori structure assumption with a burn-in period of 300,000 iterations followed by
239
600,000 MCMC repetitions. Ten independent trials were run for each predefined K value, with K = 1–4. The results were analysed using STRUCTURE HARVESTER (Earl and vonHoldt 2012) and the optimal number of K was also estimated using the method of Evanno et al. (2005). Statistical analyses of the RH1 opsin gene and the Pan I locus Descriptive analysis of the RH1 opsin gene SNPs and the Pan I locus were performed using SNPStats (Sole´ et al. 2006). The analysis of association was conducted using the behaviour types as the response variable (coastal type vs. frontal type). In our case, the response variable (behaviour) was binary, and thus, an unmatched case–control design and unconditional logistic regression models were used. Individuals with missing values in the response (behaviour), SNPs or additional covariates were excluded from the analysis. For the association study, SNPStats calculates genotype frequencies, proportions, odds ratio (OR) and 95 % confidence intervals (CI), as well as p values of the association. In addition, SNPStats determines Akaike’s Information Criterion (AIC) and Bayesian Information Criterion (BIC) to facilitate selection of the best model for a specific marker. The model considers different scenarios based on the inheritance of the studied markers: codominant (the two alleles of a gene pair are both fully expressed), dominant (one of the allele produces the same phenotypic effect if inherited with a homozygous or heterozygous state), recessive (one of the allele produces its typical phenotypic effect only when paired with the same allele) and overdominant (one of the allele produces the same phenotypic effect when paired with the other allele than when paired with the same allele). The best model fitting to the data is reported as the lowest AIC and BIC values.
Results Genetic analyses at microsatellite loci We genotyped 26 loci in 148 specimens. Genetic diversity measured as He, Ho and Ar is presented in Supplementary Table S2. None of the studied microsatellite loci exhibited HWE disequilibrium (Supplementary Table S2). In addition, for all microsatellite loci used during the present study, a divergence from the neutral expectation could not be observed (Supplementary Fig. S1 and Table S3). Pairwise FST comparisons among behaviour types did not reveal significant differences at microsatellite loci (FST = 0.002, p = 0.338, 95 % CI -0.001 to 0.005) and/or significant inbreeding coefficient (FIS = 0.009, p [ 0.05,
123
240
Behav Genet (2015) 45:236–244
95 % CI -0.008 to 0.030). The Bayesian cluster analysis performed on behaviour types, using STRUCTURE, showed that the most likely number of K was 1 (mean Ln P(D) ± S.D.: K = 1, -16836 ± 2; K = 2, -16912 ± 94; K = 3, -16843 ± 7; K = 4, -16966 ± 19).
segregate at markedly different frequencies among the coastal and frontal types (Table 1). The frequencies of these SNPs in the different populations are comparable to that of Pan I polymorphism (Table 1) and occur in strong linkage disequilibrium (Table 2). Except the SNPs in the 30 -UTR at position 1295, all SNPs were synonymous.
RH1 opsin gene (SNPs) and the Pan I locus Association analysis to behaviour types Sequencing of the RH1 opsin gene resulted in the detection of 18 variable sites in 142 individuals within the gene sequence at position 59, 102, 105, 318, 336, 351, 381, 417, 459, 495, 588, 735, 738, 942, 1041, 1065, 1101, 1220 and two in the 30 -untranslated region (30 -UTR) at position 1295 and 1316. Eleven of these SNPs were observed in a single individual (59, 318, 336, 351, 381, 417, 495, 942, 1041, 1065, 1101) while seven (102, 105, 588, 735, 738, 1220, 1316) exhibited minor allele frequencies (MAF) between 2 and 6 %, i.e. these occurred in 3–8 individuals out of 142 individuals. Nevertheless, two polymorphic sites had higher MAF’s (one synonymous SNP at site 459 [AA153] and one SNP at site 1295 in the 30 -UTR), and these alleles
Table 1 Allele frequencies of the variable RH1 opsin gene SNPs at position 459 (AA153) and in the 30 -UTR (1295) and of the Pan I locus of 99 coastal and 43 frontal Atlantic cod 459 [AA153]
30 -UTR 1295
Pan I
A1
A2
A1
A2
A
B
Coastal
0.76
0.24
0.79
0.21
0.79
0.21
Frontal
0.42
0.58
0.45
0.55
0.44
0.56
Mean
0.65
0.35
0.69
0.31
0.69
0.31
The two different alleles for each polymorphism are designated A1 and A2
Table 2 Linkage disequilibrium statistics among the RH1 gene synonymous SNPs located at position 459 [AA153], at the 30 -UTR (1295) and the Pan I locus 459 [AA153] 481 [AA153]
0
3 -UTR
–
0
30 -UTR 1295
Pan I
0.205
0.205
0.999
0.999
0.930 –
0.931 0.210 0.975 0.974
Pan I
0
0
–
Three commonly used statistics were calculated with their associated p values (below diagonal): D (Linkage disequilibrium, see Lewontin and Kojima 1960), D’ (the standardised value of D; see Lewontin 1964), and the Pearson correlation coefficient (r). Associated p values are presented below diagonal . In each column, the first line indicates D value, the second D’ and the last one r
123
An association analysis of the RH1 opsin gene and the Pan I locus with the behaviour types (response variable) supports highly significant relationships between the SNPs at position 459 (AA153) and 1295 (30 -UTR), and the Pan I locus to the behaviour types (Table 3). The other polymorphic SNPs (102, 105, 588, 735, 738, 1220) within the RH1 opsin gene and the 30 -UTR (1316) were not significantly associated with the behaviour (data not shown). All logistic regression models performed showed an OR higher than 1 (associated with 95 % CI), and resulted in p values lower than 0.001. However, both AIC and BIC suggested that the best models were the codominant and dominant ones for all polymorphic SNPs and the Pan I locus (Table 3). In addition, frontal type displayed more often the codon GCC at the Ala153 while the coastal type displayed the synonymous GCT (See Table 3). Likewise, the Pan IB allele is strongly associated with the frontal type. Further, the mean depth profiles per week of the different RH1 opsin gene haplotypes (based on position 459-AA153 and 30 UTR-1295) of the different behaviour type shows that the most common haplotype of RH1 opsin gene displays strongly differentiated weekly pattern of depth profiles, while intermediate haplotypes exhibit intermediate depth profiles, and by such, suggest that the differentiation is likely to be an on-going process (Fig. 2). The most common RH1 opsin gene haplotype in the frontal behaviour type exhibited average depth of 350 m during winter time while the average depth of the most common haplotype in the coastal behaviour was around 100 m (Fig. 2). Although the weekly pattern of depth profiles was quite different between both behaviour types, RH1 opsin gene haplotypes were not behaviour specific (Fig. 2).
Discussion Here we present a significant association of intraspecific polymorphism at the RH1 opsin gene with the coastal and frontal behaviour types of Atlantic cod. Four major spectral tuning sites have previously been identified through an extensive comparison of sequences from several species, including Atlantic cod, with an aim to understand the evolutionary changes of the visual pigment genes in the Pacific bluefin tuna (Nakamura et al. 2013). Three of these
Behav Genet (2015) 45:236–244 Table 3 Significant associations between the RH1 opsin gene SNPs at position 459 (AA153) and 1295 (30 -UTR), the Pan I locus and coastal and frontal types of Atlantic cod assessed in SNPStats
241
Model
Genotype
Coastal
Frontal
OR (95 % CI)
p value
AIC
BIC
C/C
8
11
20.28 (5.20–79.17)
\0.0001
144.6
153.4
T/C
32
28
12.91 (4.16–40.06)
T/T
59
4
1.00
T/T
59
4
1.00
\0.0001
143.3
149.2
T/C–C/C
40
39
14.38 (4.77–43.40)
T/T-T/C
91
32
1.00
0.007
170.9
176.8
C/C
8
11
3.91 (1.44–10.58)
T/T–C/C
67
15
1.00
\0.0001
164.9
170.9
T/C
33
28
3.91 (1.84–8.32)
–
–
–
4.72 (2.51–8.87)
\0.0001
148.7
154.6
G/G
60
5
1.00
\0.0001
141.5
150.3
G/A
33
28
10.18 (3.59–28.87)
A/A G/G
4 60
9 5
27.00 (6.09–119.78) 1.00
\0.0001
141.8
147.7
G/A–A/A
37
37
12.00 (4.33–33.27)
G/G–G/A
93
33
1.00
0.0022
165.0
170.8
A/A
4
9
6.34 (1.83–21.98)
G/G–A/A
64
14
1.00
\0.0001
161.6
167.4
G/A
33
28
3.88 (1.80–8.35)
–
–
–
6.18 (2.98–12.83)
\0.0001
141.6
147.4
AA
61
4
1.00
\0.0001
139.3
148.1
AB
35
30
13.07 (4.25–40.18)
BB
3
9
45.75 (8.76–238.81)
AA
61
4
1.00
\0.0001
140.8
146.7
AB–BB
38
39
15.65 (5.18–47.30)
AA–AB
96
34
1.00
\0.0001
166.9
172.8
BB
3
9
8.47 (2.17–33.13)
Overdominant
AA–BB AB
64 35
13 30
1.00 4.22 (1.95–9.12)
\0.0001
163.6
169.6
Log-additive
–
–
–
8.24 (3.65–18.58)
\0.0001
139.0
144.9
459 [AA153] Codominant
Dominant Recessive Overdominant Log-additive 30 -UTR 1295 Codominant
Dominant Recessive Overdominant Log-additive Pan I Codominant
SNPStats returns genotype frequencies, odds ratio (OR) and 95 % confidence intervals (CI), and p values for the association, as well as Akaike’s Information Criterion (AIC) and Bayesian Information Criterion (BIC). For each criteria, the two best models are emboldened
Dominant Recessive
sites are conserved in Atlantic cod (E122, F261, A292) apart from the D83N replacement identified in some deepwater species that causes a blue-shift in absorption spectrum (Nakamura et al. 2013). We found that all the representative tuning sites involved in the light sensitivity of the RH1 opsin gene were identical (E122, F261, A292 and D83N) compared to the previous study (Nakamura et al. 2013). Moreover, we did not find any reference in the literature regarding the polymorphic SNPs associated with behaviour at position 459 (AA153) and position 1,295 in the 30 -UTR that are identified in this study. Therefore, based on our results, a plausible molecular mechanism for this SNP variation towards any functional phenotype remains uncertain.
RH1 opsin gene polymorphisms and visual pigments have been studied in great detail in African cichlids with a focus on AA replacements at spectral tuning sites that might be adaptive to light conditions and depth (O’Quin et al. 2011; Spady et al. 2005; Sugawara et al. 2005; Terai et al. 2006). Since the majority of these studies have reported structural AA replacements for the RH1 opsin gene, it is therefore difficult to envision how the polymorphisms identified in the current study contribute to variation in phenotype. However, both AIC and BIC values support the codominant and dominant models which suggests that phenotypic changes related to vision might occur. Nevertheless, if the strong observed association between the RH1 opsin polymorphism and the behaviour types of
123
242
Behav Genet (2015) 45:236–244
Fig. 2 Weekly depth profiles (mean and standard deviation) of the different RH1 opsin gene haplotypes for coastal (left) and frontal (right) behaviour types of Atlantic cod and proportion of each haplotype (right column) based on the two polymorphic SNPs (459 [AA153] and 30 -UTR (1295). C indicates proportions of the haplotype
observed in coastal behaviour; F proportions observed for the frontal behaviour. RH1 variable SNPs are presented in the order of their respective position in the sequences, i.e. 459 [AA153] and 30 -UTR (1295). (–) indicates haplotypes of low frequencies
the Atlantic cod is not spurious and related to choice of habitat depth (Grabowski et al. 2011) through light adaptation, plausible mechanisms might be regulatory rather than structurally. The effect of synonymous mutations has been debated for nearly a decade in human research, and evidence has been mounting concerning their effect on phenotype and fitness via several possible mechanisms (Parmley and Hurst 2007; Venetianer 2012). The efficiency, regulation and kinetics of translation, alternative splicing, changes in the secondary and tertiary messenger structure, and in the half-life of the messenger are all possible mechanisms by which synonymous mutations can lead to altered gene expression (Parmley and Hurst 2007; Venetianer 2012). Interestingly, a recent study of cis-regulatory sequences around opsin genes in African cichlids revealed two mutations within miRNAs target sites that might contribute to differential opsin expression (O’Quin et al. 2011). While alternative splicing is unlikely for the
single exon RH1 opsin gene, translation modification and altered messenger structure or messenger half-life might provide plausible mechanisms explaining the strong association identified between the observed synonymous mutation within the reading frame of the gene, the SNPs in the 30 -UTR and the behaviour types. For instance, the 30 UTR mutation might influence the binding efficiency of specific miRNAs influencing RH1 gene transcript abundance (see Parmley and Hurst 2007), putatively leading to altered visual capacities. In a previous genetic study of the behaviour types the Pan I genotype were found to be associated with the coastal and frontal behaviour types (Pampoulie et al. 2008). Pan I codes for an integral membrane protein expressed in cytoplasmic transport vesicles (Brooks et al. 2000), a function that is difficult to relate to variation in migratory behaviour. The strong linkage of the Pan I locus to other genes located at the LG1, including the RH1 opsin gene
123
Behav Genet (2015) 45:236–244
under investigation here, might be responsible for the relationship detected previously (Pampoulie et al. 2008). The observed association of both Pan I and RH1 polymorphisms to the behaviour types might therefore be due to other genes located on LG1. In the latest unpublished version of the Atlantic cod genome Pan I and RH1 are located on separate scaffolds and at least 3.5 Mbp apart (Tørresen, Nederbragt pers. comm). At present, the different mechanisms potentially responsible for the observed migration pattern of the behaviour types has not been elucidated, but natural selection (Hemmer-Hansen et al. 2013) as well as differential-habitat selection (Grabowski et al. 2011) have been suggested as potential forces maintaining this fine-scale population structure. The detection of highly differentiated genomic regions associated with ecological divergence, such as that identified on LG1, suggests a scenario whereby specific regions are affected by selection due to local adaption in the face of high levels of gene flow (Grabowski et al. 2011; HemmerHansen et al. 2013; Yokoyama and Takenaka 2004). The absence of strong population genetic structure among neutral loci, compared to the pattern of RH1 opsin gene and the Pan I locus, supports such a scenario.
Conclusion We demonstrate a significant association of RH1 genotypes between two ecologically divergent Atlantic cod ecotypes, which exhibit alternative behaviour that exposes them to altered light environments. While the functional consequences of the polymorphisms of the RH1 opsin gene in Atlantic cod remain to be elucidated, our study highlights the potential involvement of the visual system in promoting alternative migratory behaviour of Atlantic cod. Such a finding supports a notion that intraspecific visual adaptation is not limited to those species with a relatively confined and shallow depth range, but reflects a wide spread phenomenon. Acknowledgments We acknowledge funding from the EU-project CODYSSEY (Q5RS-2002-00813) for the tagging experiment and from the Icelandic Ministry of Innovation and Fisheries (Verkefnasjo´ður Sja´varu´tvegsins grant, 2011–2014) for the genetic work. Conflict of Interest Christophe Pampoulie, Sigurlaug Skirnisdottir, Bastiaan Star, Sissel Jentoft, Ingibjo¨rg G. Jo´nsdo´ttir, Einar Hjo¨rle´ lafur K. Pa´lsson, Paul R. Berg, ifsson, Vilhja´lmur Thorsteinsson, O Øivind Andersen, Steinunn Magnusdottir, Sarah J. Helyar, and Anna K. Danı´elsdo´ttir declare that they have no conflict of interest. Human and Animal Rights and Informed Consent The tagging was carried out in strict accordance with the recommendations by the Icelandic Committee for Welfare of Experimental Animals, Chief Veterinary Office at the Ministry of Agriculture, Reykjavik Iceland, under a surgery permit license (No. 0304-1901) issued to
243 V. Thorsteinsson. Informed consent was obtained from all individual participants included in the present study.
References Antao L, Lopes A, Lopes RJ, Beja-Pereira A, Luikart G (2008) LOSITAN: a workbench to detect molecular adaptation based on a FST-outlier method. BMC Bioinform 9:323–327 Beaumont M, Nichols RA (1996) Evaluating loci for use in the genetic analysis of population structure. Proc Roy Soc Lond B 263:1619–1626 Brooker AL, Cook AM, Bentzen P, Wright JM, Doyle RW (1994) Organisation of microsatellites differs between mammals and cold-water teleost fishes. Can J Fish Aquat Sci 51:1959–1966 Brooks CC, Scherer PE, Cleveland K, Whittemore JL, Lodish HF, Cheatham B (2000) Pantophysin is a phosphoprotein component of adipocyte transport vesicles and associates with GLUT4containing vesicles. J Biol Chem 275:2029–2036 Chen WJ, Bonillo C, Lecointre G (2003) Repeatability of clades as criterion of reliability: a case study for molecular phylogeny of Acantomorpha (Teleostei) with large number of taxa. Mol Phylogenetics Evol 26:262–288 Earl DA, vonHoldt BM (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour 4: 359–361 Ebert D, Andrew RL (2009) Rhodopsin population genetics and local adaptation: variable dim-light vision in sand gobies is illuminated. Mol Ecol 18:4140–4142 Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14:2611–2620 Fevolden SE, Pogson GH (1997) Genetic divergence at the synaptophysin (Syp I) locus among Norwegian coastal and North-east Arctic populations of Atlantic cod. J Fish Biol 51:895–908 Godo OR, Michalsen K (2000) Migratory behaviour of north-east Arctic cod, studied by use of data storage tags. Fish Res 48:127– 140 Grabowski TB, Thorsteinsson V, McAdam BJ, Marteinsdottir G (2011) Evidence of segregated spawning in a single marine fish stock: sympatric divergence of ecotypes in Icelandic cod? PLoS ONE 6:e17528 Hemmer-Hansen J, Nielsen EE, Therkildsen NO, Taylor MI, Ogden R, Geffen A, Bekkevold D, Helyar S, Pampoulie C, Johansen T, Carvalho GR, FishPopTraceConsortium (2013) A genomic island linked to ecotype divergence in Atlantic cod. Mol Ecol 22: 2653–2667 Jakobsdo´ttir KB, Jo¨rundsdo´ttir D, Skı´rnisdo´ttir S, Hjo¨rleifsdo´ttir S, ´ , Danı´elsdo´ttir AK, Pampoulie C (2006) Nine Hreggviðsson GO new polymorphic microsatellite loci for the amplification of archived otolith DNA of Atlantic cod, Gadus morhua L. Mol Ecol Notes 6:336–339 Jerlov N (1976) Marine optics. Elsevier, Amsterdam, pp 232. ISBN 0-444-41 490-8 ˚ , Jørgensen TE, Jueterbock A, Karlsen BO, Klingan K, Emblem A Furmanek T, Hoarau G, Johansen SD, Nordeide JT, Moum T (2013) Genomic divergence between the migratory and stationary ecotypes of Atlantic cod. Mol Ecol 22:5098–5111 Larmuseau MHD, Vancampenhout KIM, Raeymaekers JAM, Van Houdt JKJ, Volckaert FAM (2010) Differential modes of selection on the rhodopsin gene in coastal Baltic and North Sea populations of the sand goby, Pomatoschistus minutus. Mol Ecol 19:2256–2268
123
244 Lewontin RC (1964) The interaction of selection and linkage. 1. General considerations; Heterotic models. Genetics 49:49–67 Lewontin RC, Kojima K (1960) The evolutionary dynamics of complex polymorphism. Evolution 14:458–472 Michiels N, Anthes N, Hart N, Herler J, Meixner A, Schleifenbaum F, Schulte G, Siebeck UE, Sprenger D, Wucherer M (2008) Red fluorescence in reef fish: a novel signalling mechanism? BMC Ecol 8:16 Miller KM, Le KD, Beacham TD (2000) Development of tri- and tetranucleotide repeat microsatellite loci in Atlantic cod (Gadus morhua). Mol Ecol 9:238–239 Nakamura Y, Mori K, Saitoh K, Oshima K, Mekuchi M, Sugaya T, Shigenobu Y, Ojima N, Muta S, Fujiwara A, Yasuike M, Oohara I, Hirakawa H, Chowdhury VS, Kobayashi T, Nakajima K, Sano M, Wada T, Tashiro K, Ikeo K, Hattori M, Kuhara S, Gojobori T, Inouye K (2013) Evolutionary changes of multiple visual pigment genes in the complete genome of Pacific bluefin tuna. Proc Natl Acad Sci 110:11061–11066 Nordeide JT (1998) Coastal cod and north-east Arctic cod—do they mingle at the spawning grounds in Lofoten? Sarsia 83:373–379 ´ lafsson K, Hjo¨rleifsdo´ttir S, Pampoulie C, Hreggviðsson GO ´, O Guðjo´nsson S (2010) Novel set of multiplex assays (SalPrint15) for efficient analysis of 15 microsatellite loci of contemporary samples of the Atlantic salmon (Salmo salar). Mol Ecol Resour 10:533–537 O’Quin KE, Smith DA, Naseer Z, Schulte J, Engel SD, Loh Y-HE, Streelman JT, Boore JL, Carleton KL (2011) Divergence in cisregulatory sequences surrounding the opsin gene arrays of African cichlid fishes. BMC Evol Biol 11:120 O’Reilly PT, Canino MF, Bailey KM, Bentzen P (2000) Isolation of twenty low stutter di- and tetranucleotide microsatellites for population analyses of walleye pollock and other gadoids. J Fish Biol 56:1074–1086 ´ K, Thorsteinsson V (2003) Migration patterns, ambient Pa´lsson O temperature, and growth of Icelandic cod (Gadus morhua): evidence from storage tag data. Can J Fish Aquat Sci 60:1409–1423 Pampoulie C, Ruzzante DE, Chosson V, Jo¨rundsdo´ttir TD, Taylor L, Thorsteinsson V, Danı´elsdo´ttir AK, Marteinsdo´ttir G (2006) The genetic structure of Atlantic cod (Gadus morhua) around Iceland: insight from microsatellites, the Pan I locus, and tagging experiments. Can J Fish Aquat Sci 63:2660–2674 Pampoulie C, Jakobsdo´ttir KB, Marteinsdo´ttir G, Thorsteinsson V (2008) Are vertical behaviour patterns related to the pantophysin locus in the Atlantic cod (Gadus morhua L.)? Behav Genet 38:76–81 Parmley JL, Hurst LD (2007) How do synonymous mutations affect fitness? BioEssays 29:515–519 Pogson GH, Mesa KA (2004) Positive Darwinian selection at the pantophysin (Pan I) locus in marine gadid fishes. Mol Biol Evol 2165–2175 Pogson GH, Mesa KA, Boutilier RG (1995) Genetic population structure and gene flow in the Atlantic cod Gadus morhua: a comparison of allozyme and nuclear RFLP loci. Genetics 139:375–385 Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945–959 Rousset F (2008) Genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour 8:103–106 Sarvas TH, Fevolden SE (2005) Pantophysin (Pan I) locus divergence between inshore v. offshore and northern v. southern populations
123
Behav Genet (2015) 45:236–244 of Atlantic cod in the North-east Atlantic. J Fish Biol 67:444–469 Shum P, Pampoulie C, Sacchi C, Mariani S (2014) Divergence by depth in an oceanic fish. PeerJ 2:e525 Sivasundar A, Palumbi SR (2010) Paralle amino acid replacements in the rhodopsins of the rockfishes (Sebastes spp.) associated with shifts in habitat depth. J Evol Biol 23:1159–1169 Skarstein TH, Westgaard JI, Fevolden SE (2007) Comparing microsatellite variation in north-east Atlantic cod (Gadus morhua L.) to genetic structuring as revealed by the pantophysin (Pan I) locus. J Fish Biol 70:271–290 ´ lafsson K, Skı´rnisdo´ttir S, Pampoulie C, Hauksdo´ttir S, Schulte I, O ´ , Hjo¨rleifsdo´ttir S (2008) Characterisation of Hreggviðsson GO 18 new polymorphic microsatellite loci in Atlantic cod (Gadus morhua L.). Mol Ecol Resour 8:1503–1505 Sole´ X, Guino´ E, Valls J, Iniesta R, Moreno V (2006) SNPStats: a web tool for the analysis of association studies. Bioinformatics 22:1928–1929 Spady TC, Seehausen O, Loew ER, Jordan RC, Kocher TD, Carleton KL (2005) Adaptive molecular evolution in the opsin genes of rapidly speciating cichlid species. Mol Biol Evol 22:1412–1422 Stenvik J, Wesmajervi MS, Damsgard B, Delghandi M (2006) Genotyping of pantophysin I (Pan I) of Atlantic cod (Gadus morhua L.) by allele-specific PCR. Mol Ecol Notes 6:272–275 Sugawara T, Terai Y, Imai H, Turner GF, Koblmuller S, Sturmbauer C, Shichida Y, Okada N (2005) Parallelism of amino acid changes at the RH1 affecting spectral sensitivity among deepwater cichlids from lakes Tanganyika and Malawi. Proc Natl Acad Sci 102:5448–5453 Terai Y, Seehausen O, Sasaki T, Takahashi K, Mizoiri S, Sugawara T, Sato T, Watanabe M, Konijnendijk N, Mrosso HDJ, Tachida H, Imai H, Shichida Y, Okada N (2006) Divergent selection on opsins drives incipient speciation in Lake Victoria cichlids. PLoS Biol 4:e433 Therkildsen NO, Hemmer-Hansen J, Hedeholm RB, Wisz MS, Pampoulie C, Meldrup D, Bonanomi S, Retzel A, Olsen SM, Nielsen EE (2013) Spatiotemporal SNP analysis reveals pronounced biocomplexity at the northern range margin of the Atlantic cod Gadus morhua. Evol Appl 6:690–705 ´ K, Jo´nsdo´ttir IG, Pampoulie C (2012) Thorsteinsson V, Pa´lsson O Consistency in the behaviour types of the Atlantic cod: repeatability, timing of migration and geo-location. Mar Ecol Prog Ser 462:251–260 Thurman HV, Trujillo AP (2004) Introductory Oceanography, 10th edn. Prentice Hall, Upper Saddle River Tyler PA (2003) Ecosystems of the deep oceans, 1st edn, Elsevier p 532 Venetianer P (2012) Are synonymous codons indeed synonymous? Biol Mol Concepts 3:21–28 Warrant EJ, Locket NA (2004) Vision in the deep sea. Biol Rev 79:671–712 Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution 38:1358–1370 Yokoyama S, Takenaka N (2004) The molecular basis of adaptive evolution of Squirrelfish rhodopsins. Mol Biol Evol 21:2071–2078 Yokoyama S, Tada T, Zhang H, Britt L (2008) Elucidation of phenotypic adaptations: molecular analyses of dim-light vision proteins in vertebrates. Proc Natl Acad Sci 105:13480–13485
