Light induced changes in starry flounder (Platichthys stellatus) opsin expression and its influence on vision estimated from a camouflage-based behavioural assay

Correlations between variation in opsin expression and variation in vision are often assumed but rarely tested. We exposed starry flounder (Platichthys stellatus) to either broad spectrum sunlight or green-filtered light in outdoor aquaria for seven weeks and then combined digital-PCR and camouflage experiments to test two hypotheses: i) short-wavelength sensitive opsin expression decreases in a green light environment, and ii) if observed, this change in opsin expression influences colour vision as estimated using a camouflage-based behavioural assay. Of the eight visual opsins measured, Sws1 (UV sensitive) and Sws2B (blue sensitive) expression was significantly lower in fish exposed to green light. However, opsin expression in fish transferred to an arena illuminated with white LED light for three hours after the green light treatment did not differ from broad spectrum controls. Changes in opsin expression in response to artificial light environments have been reported before, but rapid changes over three hours rather than days or weeks is unprecedented. We did not observe a significant difference in a flounder’s camouflage response based on light environment, although broad spectrum fish increased and green-filter fish decreased the pattern contrast when on the blue-green substrate, and this difference approached significance. This pattern is intriguing considering green-filter fish expressed fewer UV and blue opsins and we recommend increased statistical power for future experiments. Together, our results show that starry flounder opsin expression changes rapidly in response to changes in light environment, however, there is no apparent effect on their visually mediated camouflage.


48
The ability to detect and discriminate among different wavelengths of light depends on the 49 diversity of opsins in the photoreceptors of the retina. Humans are considered to be trichromatic, 50 expressing short-wavelength sensitive (OPNSW1), middle-wavelength sensitive (OPN1MW), and 51 long-wavelength sensitive (OPN1LW) opsin genes in retinal cone cells, and rhodopsin (RHO) in 52 rod cells, which are used for scotopic (dim light) vision (1). Teleost fish typically have many 53 more visual opsins than other vertebrates (2,3), largely as a result of lineage-specific tandem 54 duplications (2,4). The advantage of large opsin repertoires, however, is not clear; humans can 55 discriminate between colours (wavelengths) that differ by less than one nm over much of the 56 visible spectrum (5).

58
It may be that a large visual opsin repertoire is a 'toolkit', with subsets of opsins used at different for about 20% of the observed variation in optomotor response (10). In addition Sakai et al. 67 (2016) found that Lws-3 expression increased in guppies grown in orange light and that fish with 68 higher levels of Lws-3 expression had higher visual sensitivity to 600 nm light (11). Conversely, 69 Wright et al. (2020) found colour perception plays an important role in female cichlid mate 70 preference but opsin expression was only weakly correlated and a direct causal link between 71 expression and behaviour was lacking (12).

73
Adaptive camouflage in Pleuronectiformes was first described by Sumner (1911): turbot 74 (Rhomboidichthys podas and Lophopsetta maculata), summer flounder (Paralichthys dentatus), 75 and winter flounder (Pseudopleuronectes americanus) changed patterns over a period of days in 76 response to various mottled sandy substrates and checkerboards (13). Juvenile plaice 77 (Pleuronectes platessa) change colour more quickly (14) and many bothids (e.g., left-eye 78 flounder) can camouflage to environment cues in seconds (15). Rapid changes in camouflage are 79 based on visual cues and can be used to infer visual performance. Flounder camouflage match 80 natural substrates well when modeled to mono-, di-, and tri-chromatic visual systems (16), 81 however, here we elected to use checkboards (as in Mäthger et al. 2006 (17)) to elicit an 82 exaggerated pattern response. We held starry flounder (Platichthys stellatus), a flatfish 83 possessing eight visual opsin genes (9), in outdoor aquaria exposed to either sunlight or green-  96 Fish were exposed to either green-filtered or broad spectrum light for seven weeks. Light 97 transmission (%) for each filter (broad spectrum: Roscolux #3410; green: Roscolux #90) was 98 measured using Ocean Optics QE Pro at -10˚C (sensor), integration time 100 µs, average of 3 99 for each spectrum, boxcar width (2), and electric dark current correction. Over the course of 100 seven weeks they were fed krill at 1% body weight per day, adjusted weekly at an estimated 2% 101 specific growth rate. Feeding occurred once daily through a 2-cm hole in the lid that was 102 otherwise sealed by a black rubber stopper to inhibit non-filtered light from entering the tanks.

103
All procedures were approved by the University of Victoria Animal Care Committee, which 104 abides by regulations set by the Canadian Council for Animal Care. tanks (lwh = 49"×19"×25") were maintained outdoors on a 12˚C closed, recirculating sea water 108 system. Each tank was partitioned with an opaque plastic sheet with perforations to allow sea 109 water to cycle through, but limit light transmission between treatments. Half of each tank was 110 wrapped in Roscosun 1/8 CTO cinematic gel filter (Roscolux #3410) and the other half in Dark 111 Yellow Green cinematic gel filter (Roscolux #90). Two fish were held in each enclosure, one 112 fish was immediately euthanized after 7 weeks exposure, and the other fish proceeded to the 113 camouflage assay for three hours prior to being euthanized. Light transmission (%) for each filter     Granularity analysis similar to that used to quantify cuttlefish camouflage (22) and avian egg 174 pattern (23) was used to get a single measurement for camouflage pattern; cropped images were 175 filtered using each of seven spatial frequency bands, or bandpass filters (i.e., 2, 4, 8, 16, 32, 64, 176 128 pixels). The pattern of individual fish was estimated using the standard deviation of 177 luminance, which measures the overall contrast within an image modelled to human vision.

178
Higher standard deviation of luminance equates to more light-and-dark contrasting patterns (i.e., 179 disruptive or mottle camouflage), whereas low values equate to low pattern contrast (i.e., 180 uniform camouflage).

182
Two-way repeated measures ANOVA was run in R version 3.2.4 using the "nlme" package and 183 Tukey multiple comparisons was run using the "multcomp" package. Analyses were based on 184 standard deviation of luminance from a total of eight fish, held for seven weeks in either broad  190 Eyes were removed and a razor blade was used to cut the cornea exposing the lens and retina.

191
The lens was removed and the retina extracted. Retinas were frozen in liquid Nitrogen and stored  (Table 1). Opsins were multiplexed using FAM and VIC reporter dyes. . Patterns in expression were tested using a paired student's 212 t-test. All statistical tests were evaluated at α = 0.05 level of significance.

217
Experimental animals 218 Fish varied in size but this variation was distributed among treatment and control aquaria. Light 3.110, p = 0.0536) (Fig 2). opsins compared to those exposed to broad spectrum light (student's t-test, t = 3.9414, p = 255 0.01121 and t = 1.1458, p = 0.004792, respectively) (Fig 3). Opsin gene expression levels were 256 the same in fish from the broad spectrum and green-filtered light exposure that were transferred 257 to the behavioural arena and exposed to white LED light for three hours (i.e., the duration of the 258 behavioural assay) (Fig 3).  Microspectrophotometry data indicate that all are translated and that just one type of 271 chromophore is used (9). We predicted opsin expression would be modified by a seven-week 272 exposure to distinct light environments and that changes in opsin expression over that length of 273 time would influence vision. We used a camouflage-based assay to assess visual performance.

298
The rapid plasticity of opsins on the order of hours, rather than days, has implications for the introducing novel ambient light should be used (e.g., a closing cod-end on a trawl net).

316
Varying light environments, driven by water depth or season, affect opsin expression in several 317 species of damselfish, whereas other species appear to have more stable expression patterns (28).

318
In stickleback, opsin expression is shifted toward longer wavelengths in freshwater populations 319 relative to marine populations, and these shifts correlate with differences in the light available decreased pattern contrast and the difference approached significance (Fig 2). The experiment did not control for the difference in overall light intensity between the two 408 environments. The green-filtered environment allowed approximately 12% sunlight through 409 compared to 88% in the broad spectrum environment. Although it is possible the differences in 410 opsin expression could be due to light intensity, the evidence contrary to that point is two-fold.

411
One, the opsins that were expressed significantly lower correspond to the wavelengths of light  Concluding remarks 419 We found significantly greater UV and blue opsin expression after seven weeks in starry