Effects of exogenous thyroid hormones on visual pigment composition in coho salmon (Oncorhynchus kisutch)

The spectral quality of light in aquatic environments is spatially and temporally more variable than that in terrestrial environments. Aquatic organisms that move between different habitats, or those that inhabit seasonally variable habitats, are faced with the challenge of tuning their spectral sensitivity to maximize detection and identification of targets(predators, prey and conspecifics). In fishes, spectral sensitivity can be adjusted by the addition or loss of a photoreceptor class or by changes to the visual pigments within the photoreceptors themselves(Beaudet and Hawryshyn, 1999; Bowmaker, 1995).

Visual pigments (VPs) comprise two components: an opsin protein and a chromophore. Specific amino acid sites throughout the opsin protein play key roles in spectral tuning of the resultant VP(Yokoyama, 2000; Yokoyama et al., 2007). There are five classes of vertebrate opsins that are categorized based on amino acid sequence and on spectral absorbance (reviewed in Bowmaker, 1995; Yokoyama, 2000). In addition to expressing a representative of one or more of each of these opsin classes,some fishes have recently been found to express different subtypes of the various opsin classes (Chinen et al.,2003; Matsumoto et al.,2006; Wood and Partridge,1993). Changes in expression levels of opsin subtypes have been associated with ontogenetic changes and metamorphic transitions, some of which have been induced artificially with hormones or the light environment(Beatty, 1975; Carlisle and Denton, 1959; Fuller et al., 2005; Hope et al., 1998; Mader and Cameron, 2004; Shand et al., 2008; Shand et al., 2002; Takechi and Kawamura,2005).

The other component of the VP, the chromophore, can also be varied in some species. Many freshwater and euryhaline fishes have the ability to change which chromophore is incorporated into their VPs, shifting between retinal(aldehyde of vitamin A₁) and 3,4-dehydroretinal (aldehyde of vitamin A₂), or using mixtures of both(Beatty, 1984). The wavelength of maximum absorbance (λ_max) of a VP based on vitamin A₂ is long-wavelength shifted relative to the same opsin combined with vitamin A₁. The long-wavelength shifted vitamin A₂-based VPs are also less thermally stable than vitamin A₁-based VPs, which could have implications for species that inhabit temperate waters where ambient temperatures vary seasonally.

Pacific salmonids are anadromous fishes restricted to temperate climates that are equipped with a remarkably dynamic visual system that varies temporally at different time scales throughout life history(Allison et al., 2006a; Allison et al., 2003; Beatty, 1966; Browman and Hawryshyn, 1994; Hawryshyn et al., 1989; Temple et al., 2006), making this group ideal for investigating adaptive changes in visual pigment composition.

Pacific salmon start life in freshwater as alevin. They become parr once their yolk sac is absorbed. They may spend anywhere from a few days to a few years in fresh water (depending on species) before migrating to sea. A metamorphic event called smoltification precedes seaward migration, after which they are referred to as smolts. Following a period of rapid growth at sea, the length of which differs among species, they return to their natal streams to spawn and die (Groot and Margolis, 1991). During smoltification, they lose most of their ultraviolet-sensitive (UVS) cones through programmed cell death(Allison et al., 2006a). On their return migration back to fresh water, some of these UVS cones are regenerated (Allison et al.,2006a; Beaudet et al.,1997). It has been proposed that the VP vitamin A₁/A₂ ratio in salmon might follow a similar pattern,changing at the time of the metamorphic–migration event, with vitamin A₂ dominating in fresh water and vitamin A₁ dominating at sea (Alexander et al., 1994; Novales Flamarique, 2005). However, recent observations show a seasonal pattern in vitamin A₁/A₂ VP ratio in all ages of coho salmon(Oncorhynchus kisutch) (Walbaum 1792), evidence that this ratio is correlated with seasonal changes in environmental variables and not with the pattern of migration (Temple et al.,2006). With regard to the timing of changes in vitamin A₁/A₂ ratio, we refer to these two models as the migration–metamorphosis and the seasonal hypotheses.

In salmonids, thyroid hormones (THs) are responsible for many of the structural, physiological and behavioral changes associated with smoltification, e.g. silvering, changes in body shape, increased saltwater tolerance, change in rheotaxis, cell proliferation in olfactory epithelium,loss of UVS cones and changes in visual pigment gene expression(Allison et al., 2003; Folmar and Dickhoff, 1980; Grau et al., 1982; Higgs et al., 1982; Hoar, 1988; Lema and Nevitt, 2004; McBride et al., 1982; Specker et al., 2000; Staley and Ewing, 1992; Veldhoen et al., 2006). Earlier research established that TH treatment increased the proportion of vitamin A₂ in the VPs of coho and other salmonids(Beatty, 1972). However, a recent study (Alexander et al.,1998) suggested that the direction of shift in vitamin A₁/A₂ ratio may vary with temperature. Alexander et al.(Alexander et al., 1998) found that coho held in warm water increased the proportion of vitamin A₂in their VPs when treated with exogenous TH, but that coho held in cold water decreased the proportion of vitamin A₂ in their VPs when treated with exogenous TH.

In the present study, we investigated the effect of TH treatment on VP compositions in coho salmon to determine whether there is variability in the direction of vitamin A₁/A₂ ratio shifts at different times of year and under different environmental conditions, including warm and cold temperatures. We report that TH treatment increased the proportion of vitamin A₂ under all conditions tested. We also provide evidence for a change in opsin expression in middle-wavelength-sensitive (MWS)cones.

A series of five experiments was used to compare the effects of TH on vitamin A₁/A₂ VP ratios in coho salmon of various ages,tested at different times of year and held under different environmental conditions (for details, see Table 1).

Coho were provided by two local hatcheries (Robertson Creek hatchery,Department of Fisheries and Oceans, Canada, Port Alberni, British Columbia and Target Marine, commercial hatchery, Sechelt, British Columbia). Both facilities rear coho from eggs to smolts, under natural environmental conditions, in outdoor ponds and tanks. Fish for use in these experiments were transported to the University of Victoria aquatic facilities within a few days of commencing each experiment. Care and treatment of fish was in accordance with University of Victoria's Animal Care Committee, under the auspices of Canadian Council for Animal Care.

Two different TH delivery systems were used for TH treatment: TH-treated food and an exogenous TH bath. Both techniques are effective in stimulating smoltification-like transitions in coho(Alexander, 1998; Alexander et al., 1998; Munz and Beatty, 1965) and vitamin A₁/A₂ ratio shifts in this and other species(Allen, 1977; Allison et al., 2004; Beatty, 1969a; Beatty, 1972; Cristy, 1974; Jacquest and Beatty, 1972). The TH-food treatment was used in Experiments I and II to deliver TH to 150 coho housed in a 750 l cylindrical fiberglass tank with flow-through water replacement. However, it was not possible to control the dose of TH delivered to each fish using TH-treated food. High variation in vitamin A₁/A₂ ratios among the food-treated fish and incomplete transition to vitamin A₂ VP dominance (see Results) led us to use a bath treatment for the remaining experiments in which fewer fish were used. Although an investigation into the differences between these two treatments might be fruitful, it was beyond the scope of this study. Our goal was to determine the direction of shift in vitamin A₁/A₂ ratio and both TH treatments were consistent in this regard.

For Experiments I and II, fish were sampled weekly to track temporal changes in vitamin A₁/A₂ ratio. TH-treated fish were fed commercial salmon pellets sprayed with ethanol containing dissolved l-thyroxine and 3,5,3′-triiodo-l-thyronine (Sigma,St Louis, MO, USA). A mixture of 120 p.p.m. by weight l-thyroxine(T₄) and 12 p.p.m. 3,5,3′-triiodo-l-thyonine(T₃) was used to approximate plasma TH levels found in salmonids prior to smoltification. This dose stimulated smoltification-like transitions in other Oncorhynchus spp.(Ebbesson et al., 2000; Plate, 2001). TH-treated fish were fed the T₃–T₄ diet for 1 month, and then given the control diet for an additional 2 weeks. The control group was fed the same commercial salmon pellets, sprayed only with ethanol, for the entire experiment. Both control and treatment fish were fed to satiation every other day. Weekly, three to ten fish were sampled from both control and treatment groups (for details, see sample sizes in results).

For Experiments III–V, T₄ was dissolved in 1.5 ml of 0.1 moles l^–1 NaOH and added to tank water for a final concentration of 300 μgl^–1 T₄. Control fish received the vehicle only (1.5 ml of 0.1 moles l^–1 NaOH). Water was changed three times per week. Both control and treatment groups were fed to satiation every other day with commercial salmon pellets. Five to 10 fish were sampled from both control and treatment groups after 4–6 weeks of treatment.

Fish were dark adapted for at least 1 hour prior to sacrifice by an overdose of Euganol 100 mg l^–1 (ICN Biomedicals, Irvine, CA,USA), followed by cervical transection. The right eye was enucleated and hemisected along an anterior–posterior axis. A piece of retina 1–2 mm² was cut out of the dorsalmost section of the dorsal hemisphere. This retinal sample was teased apart on a glass coverslip and a drop of minimum essential medium (Sigma, Oakville, ON, Canada; pH adjusted to 7.4–7.6) was applied to the sample. A second cover slip was placed over the sample and sealed with paraffin. All procedures were performed under deep-red illumination (>650 nm) or using a dissecting scope equipped with infrared LED (800 nm) illumination and monitored with a charge-coupled device(CCD) camera.

A CCD–microspectrophotometer (MSP), which has been described previously (Hawryshyn et al.,2001), was used to measure spectral absorbance of individual rod and cone photoreceptors. The CCD–MSP device delivered a short flash[0.05–0.5 s; duration was dependent on intensity and was set to deliver an optimum number of photons per exposure time=total counts (500 000 counts)]of full-spectrum light [300–800 nm; 150 W xenon light source –intensity regulated (Oriel, Stratford, Connecticut, USA)] to the photoreceptor outer segment. Beam size was ∼2×3 μm. After passing through the sample, the transmitted beam was directed through a spectrometer [300 nm blazed grating (Acton Research Corporation, Acton, MA, USA)] and onto a 1340×400 pixel, Peltier cooled (–45°C), back-illuminated CCD detector (Princeton Instruments, Roper Scientific, Trenton, NJ, USA). Photoreceptor absorbance [log₁₀(1/T)] was calculated by comparing the transmitted intensity through the photoreceptor(I_M) to the transmitted intensity through an area clear of debris adjacent to the photoreceptor (I_R) thus, T=I_M/I_R.

Retinal samples were examined under infrared illumination (Schott RG850 filter, Ealing Optics, London, UK) and monitored by an infrared camera(Canadian Photonics Laboratory, Minnedosa, Manitoba, Canada). The search image and infrared filtered beam (Schott RG850 filter) were displayed on a computer monitor. A motorized X–Y stage (Marhauser-Wetzlar GmbH, Germany) was used to position of the photoreceptor outer segment (OS)relative to the measurement beam. The path of the motorized stage was recorded to prevent repeated measurements of photoreceptor OSs. Difference spectra were used to verify that the α-absorption band was due to the presence of a photolabile pigment and were calculated by subtracting the bleached absorbance curve (full spectrum bleach 2–5 s) from the initial absorbance curve.

Criteria for acceptance of absorbance spectra were: (1) presence of a baseline on the long-wavelength limb(Harosi and MacNichol, 1974);(2) λ_max near the expected wavelength for known Oncorhynchus spp. photoreceptors [UVS≈350–380 nm;short-wavelength-sensitive (SWS)≈420–450 nm; MWS≈490–550 nm;and long-wavelength-sensitive (LWS)≈540–630 nm; rod≈500–530 nm (Hawryshyn et al., 2001; Hawryshyn and Harosi, 1994)];(3) minimal absorbance by photoproduct; and (4) signal-to-noise ratio of the main absorption band (α-band) greater than 5:1. Determinations ofλ _max, and percent vitamin A₂ from acceptable absorbance records were performed offline subsequent to initial sampling.

A custom-designed analysis program was used to determineλ _max from absorbance records using existing templates. Each MSP record consisted of over 1000 points collected between 300 and 750 nm. Each record was linear de-trended if necessary(Harosi, 1987). A nine-point adjacent averaging function was used for line smoothing, and the smoothed curve was normalized to zero at baseline on the long-wavelength arm and to one at the center of the α-band. The fit of the normalized curve was compared with a nonlinear least-squares routine to the upper 20% of the weighted vitamin A₁/A₂ ratio averaged Govardovskii et al. (Govardovskii et al.,2000) template (based on the center of the α-peak ±40 nm).

For some rods, we also obtained a second estimate of λ_maxbased on a template created by Munz and Beatty(Munz and Beatty, 1965) for coho rod pigments. Rod absorbance curves were compared (minimum variance fit)to the Munz and Beatty (Munz and Beatty,1965) template, which extends from the λ_max to a point at 20% of the maximum on the long-wavelength arm. This template(Munz and Beatty, 1965)assumes that λ_max values of coho rods vary from 503–527 nm. Their model is in close agreement with Harosi's(Harosi, 1994), which predicts that a vitamin A₁-based VP with a λ_max of 503 nm shifts to 529 nm if the A₁-based chromophore is replaced with a vitamin A₂-based chromophore in the same opsin (see Results). However, many of the rods we measured had λ_max values that exceeded 527 nm and therefore would not fit the Munz and Beatty(Munz and Beatty, 1965)template. In these cases, we used the estimate obtained by the fit to the Govardovskii et al. (Govardovskii et al.,2000) template.

Coho salmon parr, obtained from Robertson Creek hatchery in 2004 and maintained in outdoor aquatic facilities, were dark adapted for 1 hour and then killed by immersion in 300 mg l^–1 tricaine methanesulfonate (Crescent Research Chemical, Phoenix, AZ, USA). Neural retina was dissected free of pigmented epithelium under infrared illumination. Immediately after dissection, total RNA was isolated from the tissue using TRIzol reagent (Invitrogen Canada, Burlington, ON, Canada) as per the manufacturer's recommended protocol. Retinal tissue was placed in a 1.5 ml microcentrifuge tube containing 700 μl TRIzol and was homogenized using a disposable Kontes Pellet Pestle with cordless motor tissue grinder (Kimble Kontes, NJ, USA). Isolated RNA was re-suspended in 50 μl RNase-free water. RNA concentration was determined by measuring absorbance using spectrophotometry at a standard wavelength of 260 nm. Total cDNA was synthesized using 1 μg of total RNA. Each RNA sample was annealed with 500 ng random hexamer oligonucleotide (Amersham Biosciences, Baie d'Urfe,Québec, Canada) and cDNA was prepared using Superscript II RNase H reverse transcriptase (Invitrogen) as described by the manufacturer's protocol.

A degenerate forward primer (5′-GCTATTGAGAGGTACATNGT-3′) was designed based on an alignment of consensus RH2 opsin open reading frame (ORF)sequences and was synthesized by Operon Biotechnologies (Huntsville, AL, USA). A degenerate reverse primer (Johnson et al., 1993) was also used (5′-RAANATNACNGGRTTRAA-3′). All primer pairs used in this study were diluted and combined in an equimolar ratio to a final concentration of 10 μM. Primers were used to amplify cDNA synthesized from 1 μg parr retinal total RNA. The 20 μl reaction contained 20 mmol l^–1 Tris-HCl, 50 mmol l^–1KCl, 1.5 mmol l^–1 MgCl₂, 200 μmol l^–1 dNTPs, 1 μmol l^–1 of each primer, 2μl cDNA diluted 1:20 and 1.0 U Platinum Taq DNA polymerase (Invitrogen). The thermocycle program was 94°C for 9 min, followed by 30 cycles of 94°C for 30 s, 50°C for 1 min and 72°C for 1.5 min, and a final extension at 72°C for 10 min. Amplicons were separated in a 1.5% agarose gel and visualized by ethidium bromide staining. The DNA band was excised from the gel and extracted by freeze–squeeze centrifugation(Smith, 1980). Extracted DNA was cloned into PCR2.1-TOPO vector using the TOPO TA Cloning Kit (Invitrogen). Plasmid DNA was purified using QIAprep Spin Miniprep Kit (Qiagen, Mississauga,ON, Canada) and sequenced (Centre for Biomedical Research DNA Sequencing Facility, University of Victoria, Victoria, BC, Canada). A partial RH2 opsin sequence was obtained using the degenerate primers that differed from the previously reported RH2 sequence for coho(Dann et al., 2004).

The full-length coding sequence of the alternate RH2 was isolated from control fish used in Experiment III using the BD SMART RACE cDNA Amplification Kit (BD Biosciences Clontech, Mississauga, ON, Canada) according to the manufacturer's protocol; with the exception that Platinum Taq (Invitrogen) was used in the PCR reactions. Primers used in the RACE reactions were synthesized by Invitrogen and Operon Biotechnologies, respectively, as follows:3′-RACE primer (5′CTATGCCAGCTTTGCTGCCTGGATT-3′) and 5′-RACE primer (5′-GGCAGCACAGGCCATTGCCATGAC-3′). The 5′-RACE reaction used the following thermal profile: initial denaturation at 94°C for 9 min followed by 5 cycles at 94°C for 30 s,72°C for 3 min followed by 5 cycles at 94°C for 30 s, 70°C for 30 s, 72°C for 3 min, followed by 35 cycles at 94°C for 30 s, 62°C for 30 s and 72°C for 3 min. The 3′-RACE reaction thermal profile,was 94°C for 9 min, followed by 35 cycles of 94° for 30 s, 72°C for 90 s. Amplicons were gel purified, cloned and sequenced as described above.

Sequences obtained from 5′ and 3′ RACE were assembled and compared with sequences in GenBank using Blastn(http://www.ncbi.nlm.nih.gov/BLAST/). Sequence alignments were performed using ClustalW(Chenna et al., 2003).

Statistical analyses were performed using the mean λ_maxvalue obtained for each receptor class from individual fish (see Allison et al., 2004; Jokela et al., 2003). This approach is not typical of MSP studies, which classically report the meanλ _max for each class of receptor from all fish sampled (e.g. Cummings and Partridge, 2001; Harosi and Kleinschmidt, 1993; Hawryshyn et al., 2001; Nawrocki, 1985; Novales Flamarique, 2005). However, it is statistically correct to treat the fish as the sample unit,rather than the individual photoreceptors, because photoreceptors from the same fish lack independence (pseudoreplication). When comparisons were made between fish for a particular receptor class, we used the meanλ _max of all receptors collected from each fish (fish meanλ _max). When comparisons were made between control and treatment groups for a particular receptor class, we used the meanλ _max of all fish, i.e. the mean of the fish meanλ _max values for all fish in each group (group meanλ _max).

One-way analysis of variance was used to detect differences in means among groups over time. Independent sample t-tests were used for comparisons between the control and treatment groups at specific time points(α=0.05).

Exogenous TH treatments (both TH diet and TH bath) resulted in typical parr–smolt-like transitions in all five experiments. TH-treated fish exhibited physical characteristics associated with post smoltification: loss of parr marks, increased silvering, blue–green dorsal coloration, and decreased condition index (body weight to length ratio). None of these changes was observed in control fish.

In Experiments I and II, performed in winter and summer, respectively(Table 1), we made weekly comparisons of the mean λ_max of rods from both control and TH-treated groups. In both experiments, rod λ_max gradually increased in TH-treated groups relative to controls, and after 4–5 weeks TH-treated groups had shifted to significantly (P<0.05) longer wavelengths than controls (Table 2; Fig. 1A,B). At the end of the treatment period, when TH-treated fish were put on the control diet, there was a rapid decrease in λ_max, and after 2 weeks mean λ_max values of TH-treated and control groups were not significantly (P>0.4) different(Fig. 1A,B).

In Experiments III–V, exogenous TH significantly(P<0.001) increased group mean λ_max of rods, MWS and LWS cones. For rods, the long-wavelength shift in group meanλ _max matched that anticipated for a shift from vitamin A₁- to A₂-based VP dominance(Table 2, Fig. 2). The distribution ofλ _max values from all rods from all fish sampled in Experiments III–V extended from ∼500 to 540 nm(Fig. 3A).

For LWS cones, like rods, the mean λ_max values of TH-treated groups were shifted to significantly (P<0.001) longer wavelengths than controls in all three experiments(Fig. 2). Among the three experiments, the control groups did not differ significantly (ANOVA;F₁₆=0.861, P=0.444) nor did the TH-treated groups (ANOVA;F₁₇=0.209, P=0.746). The range of λ_maxvalues for LWS cones extended from 563 nm to 633 nm(Fig. 3B).

For MWS cones, like the rods and LWS cones, the λ_maxvalues of TH-treated groups were shifted to longer wavelengths than the controls in all three experiments. However, we did not perform ANOVA tests because we had evidence that more than one variable was changing (see below). The group mean λ_max values for MWS cones ranged from 501.5 to 547.7 nm (Fig. 2). The range of λ_max values for all individual MWS cones (from all fish,from Experiments III–V) spanned a spectral range from below 490 nm to above 550 nm (Fig. 3C). Both measures of change in λ_max suggest a shift that was greater than predicted for a shift from vitamin A₁- to A₂-based chromophores in a single opsin. There are six empirical models that predict the change in λ_max(Δλ_max=A₂λ_max–A₁λ_max)that occurs when vitamin A₁ and A₂ chromophores are exchanged in a single VP opsin (Bridges,1965; Dartnall and Lythgoe,1965; Harosi,1994; Parry and Bowmaker,2000; Tsin et al.,1981; Whitmore and Bowmaker,1989). For vitamin A₁-based VPs withλ _max values between 495 nm and 512 nm(Fig. 4),Δλ _max is predicted to be between 16.6 nm(Parry and Bowmaker, 2000) and 36.0 nm (Whitmore and Bowmaker,1989). The Δλ_max value we observed for MWS cones was between 46.2 nm and 60 nm (13.8 nm range forΔλ _max is the difference between using the group meanλ _max value or the mean of all individual cones,respectively). Regardless of how it was calculated, the magnitude ofΔλ _max was greater than predicted by any of the existing models. We therefore hypothesized that more than one subtype of the RH2 opsin was being expressed in coho MWS cones. Given this proposed explanation, it was not appropriate to treat all MWS cones as if originating from a single normally distributed population. Therefore, MWS coneλ _max values are hereafter described by the range ofλ _max values recorded from all individual MWS cones from each group of fish.

Estimating the λ_max of the hypothesized opsin subtypes was complicated by the variation in λ_max that resulted from the variable chromophore ratio. To address this, we plottedλ _max values from measurements made on ∼100 MWS–LWS double cones from which both OSs had been recorded(Fig. 5). The expectation for such a plot, for a species with only one opsin in each of its double cone OSs and using only one chromophore type (i.e. not coho salmon), is a single tightly clumped distribution of λ_max values. For a species with only one copy of each opsin but the ability to use both vitamins A₁ and A₂ (as was thought to be the case for coho), the prediction is a distribution of points tightly grouped along a straight line that extends diagonally away from the origin(Loew and Dartnall, 1976). Our observation did not match either of these predictions and, instead, looked like a diagonal line with a positive slope stretched out along the horizontal axis (Fig. 5). This distribution indicated the presence of more than one opsin being expressed in MWS cone OSs.

To estimate λ_max values of the two hypothesized opsin subtypes, we defined the limits of the distribution with lines that took into account a measurement error of ±3 nm on all sides(Fig. 5), except the lower side(see below). The distribution of λ_max values for LWS cones extended from 563 to 633 nm. The distribution for MWS cones extended from 495 to 548 nm. To calculate the λ_max values of the corresponding vitamin A₁- or A₂-based VP pairs, we assigned the lower limit of the λ_max values as the λ_max for the vitamin A₁-based VP and the upper limit as the vitamin A₂-based VP. These values were then used in equations provided in the six published models predicting Δλ_max(Bridges, 1965; Dartnall and Lythgoe, 1965; Harosi, 1994; Parry and Bowmaker, 2000; Tsin et al., 1981; Whitmore and Bowmaker, 1989)to estimate the λ_max of predicted VP pair. As a conservative measure, to reduce the probability of Type I error, we compared our observed data set with the model that predicted the largestΔλ _max for the given vitamin A₁–A₂ VP pair [Harosi's(Harosi, 1994) model for the long-wavelength range and Whitmore and Bowmaker's(Whitmore and Bowmaker, 1989)for the middle-wavelength range].

The LWS cone λ_max distribution was explained by a single LWS opsin combining with vitamins A₁ and A₂ to give a range of 563₁–633₂ nm (subscripts denote vitamins A₁ and A₂ chromophore types, respectively). The most parsimonious model to account for the observed range in MWS cones was to assign the longest value (548 nm) as the vitamin A₂ state of one VP pair and the shortest value (495 nm) as the vitamin A₁ state of a second VP pair. We proposed that one VP pair had a range of 495₁–523₂ nm and the other VP pair had a range of 512₁–548₂ nm.

To test the hypothesis that coho express a second RH2 opsin subtype,retinal material was collected from fish that showed a large range in MWS coneλ _max values. Two subtypes of the RH2 opsin were isolated,cloned and sequenced. One subtype had been previously identified in coho [RH2(Dann et al., 2004)], here renamed RH2A, and the second was a novel RH2 sequence that we have named RH2B(GenBank Accession Number DQ309027) in accordance with nomenclature recently used for other fish opsin subtypes(Carleton and Kocher, 2001; Collin et al., 2003; Matsumoto et al., 2006).

The possession of several amino acids that are conserved among vertebrate opsins suggests that RH2B is a functional opsin. The predicted RH2B amino acid sequence has a lysine at position 296 that forms a Schiff's base linkage with the chromophore (Wang et al.,1980). RH2B has glutamic acid at position 113, which is the counter ion of the protonated Schiff's base(Sakmar et al., 1989; Zhukovsky and Oprian, 1989),and cysteine residues at positions 110 and 187, which form a disulphide bond within the opsin (Karnik et al.,1988). RH2B also has three amino acids, S240, T243 and V250, which are located in cytoplasmic loop 3 and are conserved in all opsins(Archer, 1995).

Alignment of RH2A and RH2B sequences(Fig. 6) reveals 48 amino acid differences (86.1% amino acid sequence identity). The substitution of glutamate (E) in RH2A for glutamine (Q) in RH2B at position 123 (analogous to position 122 in bovine rod opsin) could play an important role in tuning theλ _max of these pigments (see Discussion).

The main findings of these experiments were: (1) exogenous TH induced smoltification-like transitions in pre-smolt coho under all environmental conditions tested and throughout the year; (2) exogenous TH resulted in a long-wavelength shift in rod λ_max in all five experiments;(3) the observed range of group mean λ_max values for rods(503–533 nm) was consistent with a shift in vitamin A₁/A₂ ratio; (4) the range of λ_maxvalues observed for MWS cones extended from 495 to 548 nm, which was greater than predicted for a shift in vitamin A₁/A₂ ratio within a single opsin; and (5) a second RH2 opsin subtype was isolated, cloned and sequenced.

Exogenous TH increased the λ_max of all photoreceptors regardless of differences in age, time of year, rearing conditions or TH delivery method. In rods, the 30 nm increase in group meanλ _max was consistent with a change in vitamin A₁/A₂ VP ratio based on predictedΔλ _max values(Bridges, 1965; Dartnall and Lythgoe, 1965; Harosi, 1994; Parry and Bowmaker, 2000; Tsin et al., 1981; Whitmore and Bowmaker, 1989). This TH-induced increase in vitamin A₂ is consistent with nearly all previous studies with teleosts(Alexander et al., 1998; Allen, 1971; Allen, 1977; Allison et al., 2004; Beatty, 1969a; Cristy, 1974; Jacquest and Beatty, 1972; McFarland and Allen, 1977; Munz and Swanson, 1965; Tsin and Beatty, 1979), except one trial in a study by Alexander et al.(Alexander et al., 1998), which showed a decrease in vitamin A₂ in coho reared at cold temperatures(5°C). Further investigation is required to understand the factors that account for the difference between our results in cold water (3–5°C)and those of Alexander et al. (Alexander et al., 1998).

The λ_max values recorded from individual rods, rather than group mean values, had a range that was greater than theΔλ _max models predicted, and the meanλ _max of rods from the TH-bath treated fish was 532.7 nm,which is long-wavelength shifted by nearly 5 nm relative to previous reports for coho salmon (Alexander,1998; Alexander et al.,1994; Beatty,1972; Munz and Beatty,1965; Temple et al.,2006). One explanation is that our TH-bath treatment may have resulted in a complete transition to vitamin A₂-based VPs, whereas previous studies used fish that may have had intermediary vitamin A₁/A₂ ratios owing to the source of fish or treatments given [e.g. wild versus untreated fish(Alexander et al., 1994; Temple et al., 2006), TH in diet (Alexander et al., 1998)(present study), or intraperitoneal injections of TH for less than 14 days(Beatty, 1972)]. The large shifts in λ_max, resulting from TH treatment may rarely occur naturally, but they demonstrate the extent of adaptability of this species.

Alternatively, it is possible that a second opsin subtype is expressed in coho rods. European eels (Anguilla anguilla L.) express two RH1 opsin subtypes. Under natural conditions, the European eel alters expression of its RH1 opsin subtypes upon migrating from fresh water to the sea(Carlisle and Denton, 1959),and they have been induced to shift between the two subtypes using exogenous hormone treatment (Hope et al.,1998; Wood and Partridge,1993). Recent evidence suggests that rainbow trout (O. mykiss) could also have two RH1 opsin genes(Allison et al., 2006b). The expression of these two genes was shown to vary between TH-treated and control fish, as proposed here for the RH2 opsin subtypes. It is possible that the wide range of values observed for rods was the result of a TH-induced shift to a second RH1 opsin subtype in coho.

In our plot of LWS verses MWS cone λ_max values, we observed three outliers below the lower horizontal line at 563 nm, which delineated the vitamin A₁ state of the proposed LWS opsin(Fig. 5). The data points came from three separate control fish. We are collecting measurements from additional fish as part of a subsequent study to define more accurately the lower limit of λ_max values for LWS cones in coho salmon. Again, we would like to suggest this is preliminary evidence for the presence of multiple subtypes of the LWS opsin.

MSP and molecular approaches have provided independent evidence for the presence of two RH2 opsin subtypes in coho salmon. The λ_maxvalues recorded from MWS cones ranged from 495 nm to 548 nm(Δλ_max=53 nm). This observation argues strongly for two opsin subtypes, especially in the context of several models predicting aΔλ _max no greater than 36 nm(Bridges, 1965; Dartnall and Lythgoe, 1965; Harosi, 1994; Parry and Bowmaker, 2000; Tsin et al., 1981; Whitmore and Bowmaker, 1989). Our molecular analysis led to the isolation and sequencing of a novel subtype of the RH2 opsin gene, thus confirming our hypothesis. When compared with the existing sequence [RH2A (Dann et al.,2004)], there were 48 amino acid differences(Fig. 6). Key among these is likely to be the substitution of glutamate for glutamine at position 123,which is analogous to 122 in bovine RH1. This particular site plays an important role in spectral tuning of both RH1 and RH2 opsins(Yokoyama et al., 1999). The E122Q substitution results in a short-wavelength shift in bovine RH1 by 20–25 nm (Sakmar et al.,1989). In zebrafish, three out of the four RH2 opsin subtypes have glutamate at position 122 and all three absorb maximally at shorter wavelengths (17–38 nm) than the fourth, which has glutamine at position 122 (Chinen et al., 2003; Chinen et al., 2005). Based on our estimates from Fig. 5, the predicted values for the RH2A and RH2B opsin subtypes, when combined with vitamin A₁, were 512₁ nm and 495₁ nm,respectively. This observed shift in λ_max was within the range previously reported for other opsins differing by an E122Q substitution. Based on its sequence and on our MSP results, RH2B was short-wavelength shifted by ∼17 nm relative to RH2A.

The magnitude of shift in λ_max between control and TH-treated groups suggested that TH played a role in regulating expression levels of RH2A and RH2B opsin subtypes. This is consistent with the function of TH as a signaling mechanism in vertebrate metamorphosis(Power et al., 2001), and in retinal development and opsin expression(Allison et al., 2003; Browman and Hawryshyn, 1992; Harpavat and Cepko, 2003; Roberts et al., 2006). Current efforts in our laboratory are aimed at determining the spatiotemporal expression patterns of RH2A and B opsins in coho and other Pacific salmonids.

The ecological significance of the second RH2 opsin subtype was not immediately evident. The λ_max of MWS cones in TH-treated fish were long-wavelength shifted relative to control fish, suggesting that the natural state for post smoltification coho was to express the long-wavelength shifted RH2A opsin. This would seem counterintuitive as after smoltification, coho migrate to sea where the spectral distribution is short-wavelength shifted relative to most freshwater habitats(Tyler and Smith, 1970).

A similar shift in λ_max of MWS cones was reported to occur naturally in coho by Novales Flamarique(Novales Flamarique, 2005). He described the change in MWS cone λ_max as being a`compensatory' shift in response to the simultaneous loss of UVS cones from the retina. We suggest that MWS cones are shifted to longer wavelengths so that they are offset from the main spectral distribution of light along the sidewelling line of sight. Several marine fish have been shown to employ this tactic to enhance the detection of bright reflective targets against the nearly monochromatic blue–green background(McFarland and Munz, 1975). When coho move from fresh water to sea they shift their diet from predominantly terrestrial and aquatic insects to small fish and crustaceans. Crypsis in freshwater streams is accomplished by earthy and dark coloration patterns that match the substrate, therefore prey detection would be optimized by having VPs matched to the spectral background. In the open ocean, crypsis is accomplished by being transparent, or silvery, to reflect the monochromatic background light (Denton and Nicol,1965; Johnsen,2002; McFall-Ngai,1990). The long-wavelength shift in MWS coneλ _max would offset the VP improving the detection of brightly colored prey. The change in MWS cone λ_max might be equally important for conspecific recognition. Coho are territorial while in freshwater streams but join schools when they enter the estuaries, and they alter their appearance at smoltification by changing their dark reddish-brown parr marks and dorsal pigmentation to silvery sides and blue–green dorsal pigmentation (Groot and Margolis,1991).

Our observation, of an increase in vitamin A₂ with the application of TH lends support to the seasonal hypothesis for explaining the timing of shifts in A₁/A₂ ratio. Given that metamorphic transitions are driven by TH signaling mechanisms in vertebrates(Mader and Cameron, 2004; Power et al., 2001), and that TH induces smoltification-like transitions in salmon (see Introduction), then under the migration–metamorphosis hypothesis, it would be expected that TH induce a decrease in A₂ as naturally occurs when salmon undergo smoltification prior to seaward migration. In fact the opposite was observed. The λ_max of rods increased in a manner consistent with an increase in vitamin A₂ under all conditions tested, leading us to conclude that the vitamin A₁/A₂ ratio is independent of smoltification in coho salmon. These findings support our previous observations (Temple et al.,2006) that the A₁/A₂ ratio in coho salmon is linked to seasonal environmental changes, a pattern also observed in other teleosts (Allen et al., 1982; Beatty, 1969b; Dartnall et al., 1961; Ueno et al., 2005), amphibians(Makino et al., 1983) and invertebrates (Suzuki et al.,1984).

The selective advantage of shifting from vitamin A₂ in winter to vitamin A₁ in summer could be that it improves the signal-to-noise ratio in the face of seasonal variation in temperature and light conditions. Vitamin A₁-based VPs are more thermally stable than are vitamin A₂-based VPs (Ala-Laurila et al., 2003; Ala-Laurila et al.,2004; Barlow,1957). And because photoreceptors are unable to distinguish between photo- and thermal-isomerization events(Barlow, 1957) vitamin A₁-based VPs will provide a higher signal-to-noise ratio(Aho et al., 1988; Barlow, 1988). The eyes of exothermic organisms such as fish, amphibians and invertebrates are subject to environmental changes and as temperature increases, the signal-to-noise ratio worsens (Baylor et al., 1980). The possible advantages of long-wavelength-shifted vitamin A₂-based VPs must be countered by the poorer signal-to-noise ratio at higher temperatures (Allen and McFarland,1973). We interpret the variable vitamin A₁/A₂ ratio in coho as a trade-off between increasing spectral breadth of sensitivity and minimizing noise when temperatures rise(Temple et al., 2006).

In summary, exogenous TH produced long-wavelength shifts in photoreceptorλ _max that were explicable by changes in vitamin A₁/A₂ VP chromophore ratio combined with changes in opsin subtype expression. In addition to expressing representatives of all five vertebrate opsin classes, coho were found to express at least two subtypes of RH2 opsin. Our MSP evidence also indicates the possibility of two subtypes of the RH1 and LWS opsin genes. The potential to vary spectral sensitivity, through variable opsin expression together with alterations in the vitamin A₁/A₂ VP ratio, provides coho with a dynamic visual system that operates over spatially and temporally diverse spectral environments encountered throughout their complex life history.

LIST OF SYMBOLS AND ABBREVIATIONS

 
• λmax

  wavelength of maximum absorbance

 
• CCD

  charge couple device

 
• LED

  light emitting diode

 
• MSP

  microspectrophotometer or microspectrophotometry

 
• OS

  outer segment

 
• PCR

  polymerase chain reaction

 
• RPE

  retinal pigmented epithelium

 
• TH

  thyroid hormone

 
• VP

  visual pigment

RH1, RH2, SWS1, SWS2, MWS, LWS, UVS, RH2A, RH2B are all short forms for opsins and cone types as described in the text

We thank managers and staff at the Robertson Creek Hatchery and Target Marine Products for providing us with coho salmon. We also thank Dr Don Allen and Ms Nicola Temple for comments on earlier versions of this manuscript. This research was funded by a NSERC/SSHRC Major Collaborative Research Initiative,Coasts Under Stress grant (P.I. Rosemary Ommer, grant participant C.W.H.), and by a NSERC Equipment Grant to C.W.H. Partial support for S.E.T. came from a King-Platt Memorial Award. C.W.H. is supported by the Canada Research Chair program.