Electroretinography

For the spectral experiment, 16 M. atlanticus were placed in conditions of different light spectra, but near identical irradiance (Fig. 1a, b). Light intensity in each condition was 500 μW s⁻¹ cm⁻² at the water’s surface, provided by sixteen 1.3 m-long, custom-made LED light bars (BML Horticulture, Austin, TX, USA). ‘Red’ and ‘blue’ conditions consisted of 50 nm bandwidth spectra centered on 590 and 420 nm, respectively (Fig. 1). These spectra are detectable on the extreme ends of juvenile M. atlanticus wavelength sensitivity based upon previous microspectrophotometric (MSP) analyses [15 (link)]. Thus, these light conditions should activate long-wavelength sensitive (LWS) and short-wavelength sensitive (SWS) cone types independently. After 60 and 120 days, the spectral sensitivity of four M. atlanticus from each condition was tested in vivo by electroretinography, after which the fish were euthanized, eyecups removed, and their eyes fixed for histological analyses. Of the sixteen fish in the spectral experiment, one died from natural causes during the course of the experiment, so a total of three fish (as opposed to four) were tested as part of the red condition at the first timepoint.Fig. 1

Lighting conditions of spectral and intensity experiments. In the spectral experiment, juvenile Megalops atlanticus were kept in lighting conditions of different spectra but identical irradiance (500 µW s⁻¹ cm⁻²). The ‘blue’ condition (a) and ‘red’ condition (b) were centered on 420 nm light and 590 nm light, respectively. In the intensity experiment, M. atlanticus were kept in lighting conditions of different irradiances but identical spectra. The ‘bright’ condition (c) and ‘dim’ condition (d) consisted of 2.5 mW s⁻¹ cm⁻² and .325 µW s⁻¹ cm⁻² white light, respectively. Insets: emission spectra and irradiance values for each condition, y-axes are normalized irradiance (μW s⁻¹ cm⁻²)

For the intensity experiment, eight M. atlanticus were placed in conditions of different light irradiance, but near-identical spectra (Fig. 1c, d). The light spectrum of each condition was engineered to approximate that of natural sunlight, and were provided by four 1.3 m-long, custom-made ‘Solar-Max’ LED light bars (BML Horticulture Inc., Austin, TX, USA). Preliminary analysis of M. atlanticus retinal sensitivity by electroretinography included differentiation of light activation of rod- and cone-based systems by analysis of a-wave and b-wave amplitudes [38 (link), 39 (link)]. An abrupt change occured at ~ 3 μW s⁻¹ cm⁻² (Additional file 1: Figure S1). Thus, light intensities for the ‘bright’ and ‘dim’ light conditions were set at 2.5 mW and .325 μW s⁻¹ cm⁻², respectively, in order to activate cone-based and rod-based vision. Light intensities were measured and set using a radiant power meter with a silicon photodiode (Ophir Photonics, North Andover, MA). Changing the intensity of light bar output did not affect the light spectrum (measured with a Hyper OCR spectroradiometer, Satlantic LP, Halifax, Nova Scotia) between the two lighting conditions. After 60 days, the retinal sensitivity of the four M. atlanticus from each condition was tested in vivo by electroretinography, after which the fish were euthanized, eyes were removed, and eyecups prepared and fixed for immunofluorescence analyses (see below).