Experiment 1—larval response to a rapid attenuation of stimulus light
Acropora tenuis (Dana, 1846) is a common reef-building scleractinian coral in shallow water habitats throughout the Indo-Pacific Ocean. Seven adult colonies of A. tenuis were collected at 2–5 m depth from Backnumbers Reef (S18°29.26′, E147°09.18′), a mid-shelf reef in the central Great Barrier Reef (GBR) in November 2018, and transferred in flow-through tanks over 8 h by ship to flow-through aquaria in the National Sea Simulator (SeaSim), Australian Institute of Marine Science (AIMS), Queensland, Australia. This facility uses natural coastal seawater filtered to 1 µm and the range of water quality parameters matched that of mid-shelf reefs including Backnumbers Reef in November: temperature 26.5–27.5 °C, salinity 36.4–36.5 psu and pH 8.13–8.17. Immediately after spawning on November 5, gamete bundles were mixed to fertilize eggs, and cultures of embryos then larvae were maintained in 500 L flow through seawater tanks (0.5 µm filtered), with the motile aposymbiotic planula larvae becoming competent to settle four days after the spawning. Five to nine day old larvae were used in this experiment, with 10–15 larvae transferred into a rectangle polystyrene chamber (6.5 cm × 3.5 cm × 1 cm) filled with 15 mL 0.5 µm-filtered seawater (FSW). To examine whether larvae respond to rapid changes in the photon flux density of stimulus light, the swimming behavior of larvae was observed under the following light scheme. Firstly, a single long side of the test chamber was illuminated for 120 s using a 50 µmol/m2/s white LED light (ISC-201-2; CCS Inc., Kyoto, Japan), and the normal swimming activity of the larvae was recorded. Subsequently, the stimulus light was rapidly (< 1 s) attenuated using neutral density filters (No. 209, 210, 211, 298 and 299; Lee Filters, Andover, UK) such that the same white light illuminated the test chamber but with various photon flux densities. The five neutral density filters produced six different light stimuli each with a different photon flux density (9.1, 4.6, 2.5, 1.1, 0.55, 0.31 µmol/m2/s), all measured at the center of the experimental chamber using a Jaz spectrometer (Jaz-EL200; Ocean Optics, Dunedin, FL, USA). The swimming behavior of the larvae during the light exposure was recorded from above with a NIKON 1 J5 digital camera equipped with a NIKON Ai AF MICRO-Nikko 60 mm macro lens (NIKON, Tokyo, Japan). To avoid any potential effect of diurnal variation in the swimming activity on the results, all observations were recorded between 9:00 and 11:00, and the temperature was maintained at 26 °C.
Experiment 2—larval response to changes in the wavelength of stimulus light
Adult colonies of A. tenuis were collected at 1–3 m depth before the predicted spawning day in November 2017 from Little Pioneer Bay, Orpheus Island, GBR (S18°36.35′, E146°29.14′) and were placed into out-door raceways with constant flow-through seawater at the Orpheus Island Research Station (James Cook University). Spawning occurred on 8 November, 2017 and fertilization and larval culturing followed the methods of Experiment 1, however, the planula were maintained in plastic buckets (15 L) in 0.5 µm filtered seawater which was changed daily. 10–15 larvae aged 7–9 days were transferred into the test chamber as previously described. We first illuminated the larvae with 75 µmol/m2/s white LED light for 2 min and then the white light was replaced with one of each quasi-monochromatic light stimuli (400-, 455-, 500-, 528-, 590-, 625- or 660-nm light) with the same photon flux density as the white light (75 µmol/m2/s) for 3 min each. Different wavelengths of light were produced using eight different colours of HOLOLIGHT (Pi PHOTONICS, Inc., Hamamatsu, Japan) connected to DMX light controller. Larval swimming behaviors were again recorded from above.
To investigate the wavelength sensitivity of the larvae at higher spectral resolution, we observed the swimming behavior under monochromatic light stimuli in increments of 20 nm between 320 and 680 nm using the Okazaki Large Spectrograph (OLS) at the National Institute for Basic Biology (NIBB), Okazaki, Japan34. For this experiment, we collected colonies of A. tenuis from < 3 m depth on the fringing reef on Sesoko Island (N26°37.58′, E127°52.01′) and kept them in flow-through aquaria at Sesoko Station (Tropical Biosphere Research Center, University of the Ryukyus, Okinawa, Japan). Four days after spawning, motile planula larvae were transported to NIBB following the method described previously35. Briefly, larvae were transferred into 1 L plastic bottles at a density of 3000 individuals/L and then shipped by express service or as hand luggage. The transport duration did not exceed 2 days. The survival of coral larvae was > 90% with this transportation procedure. After arrival, larvae were transferred to plastic bowls with transport water, and then gradually acclimated to artificial seawater (LIVESea Salt, DELPHIS, Hyogo, Japan) at 26 °C. These larvae were kept at a density of approximately 1000 individuals/L and seawater was exchanged daily. 10–15 larvae in the test chamber were illuminated by 35 µmol/m2/s white LED light for two minutes and then the white light was replaced with each of monochromatic light stimuli (320–680 nm) with the same photon flux density (35 µmol/m2/s) produced by OLS from a single long side of the test chamber. Swimming tracks were recorded from above. The photon flux densities were measured at the center of the experimental chamber using Light Analyzer LA-105 (Nippon Medical and Chemical Instrument CO., Osaka, Japan). The difference in the photon flux densities between experiments (50, 75 and 35 µmol/m2/s, respectively) resulted from the difference in light apparatus used among experiments. In Experiment 2, to equalize the photon number of each light stimulus with different wavelengths, the photon flux density of all light stimuli was set to match that of a wavelength of light with the weakest photon flux density. The photon flux density values of stimulus light used in this study (< 100 µmol/m2/s) were within the range of those measured in cryptic habitats like shaded areas or vertical/downward-facing horizontal surfaces of substrata in shallow—middle waters (< 10 m depth)15,36, where Acropora tenuis larvae are likely to settle7,16.
Tracking analysis of recorded larval swimming
All movie files from Experiment 1 and 2 were processed with in-house applications written in Objective-C, C and Ruby, and the swimming speeds of the larvae were calculated using R version 3.5.137. First, we extracted movie frames and converted them to a sequence of still images at 20 frames per second; this frame rate value was adopted so that identical larvae mostly overlapped between two successive movie frames. The movie frames were not extracted from full movies because of the large data size, instead, we clipped two-seconds-long movie fragments (40 frames each) every 10 s from each full movie. Next, for each 2 s-long movie fragment, we highlighted the larval silhouettes by the background subtraction and then used global thresholding to extract larval bodies. Trajectories of identical larvae were produced by frame-by-frame tracking based on nearest neighbor algorithm. Subsequently, we selected proper trajectories of larvae in each movie fragment based on their size and roundness. We calculated the mean swimming speed of each larval trajectory in 2 s-long movie fragments by measuring the distance travelled between two successive movie frames divided by 0.05 s. We also calculated the “relative swimming speed”, defined as the swimming speed divided by the mean swimming speed before switching the light stimuli (i.e. divided by the mean swimming speed from 1 to 120 s in each movie).
Mathematical modeling and simulation
We constructed a mathematical model for larvae exhibiting a step-down photophobic response and computed the spatial distribution of the “modeled larvae” in two-dimensional chambers (Supplementary Fig. S4). Each individual larva was modeled as a self-propelled particle moving a constant velocity (2.0 mm/s) as described by Vicsek et al. (1995)38. The larva moves straight ahead and bounces off the outer boundary of the chamber following the law of reflection (angle of incidence = angle of reflection). The larva does not interact with other larvae, i.e. there were no interactions such as collision and attraction/repulsion. Fluid flow was not assumed.
To mimic an experimental chamber with shallow depth having light gradient along x-axis, a two-dimensional rectangular area was assumed which was divided into “light” and “dark” regions at “light–dark boundary” which corresponds to the position (x) at which the light intensity is equal to the larval detection threshold (Supplementary Fig. S4a). We assumed “small” and “large” rectangular areas for the simulation. A two-dimensional “small” rectangular area (7.5 × 2.5 cm), corresponding to experimental chambers used in previous phototaxis studies10, and “large” rectangular areas (1.5, 3, 5, 10, 20, 30, and 100 × 0.1 m) assumed the natural open water habitats of corals.
The assumption of a step-down photophobic response in the simulation was described in Supplementary Fig. S4b. When the modeled larva moves from the light to the dark region, it stops moving for a short time period after passing through the light–dark boundary (“time required to stop swimming”, [1] in Supplementary Fig. S4b), remains in a motionless state for a given period (“duration of no swimming period”, [2] in Supplementary Fig. S4b), and then resumes movement along a randomly determined direction with a constant velocity. The position of light–dark boundary and the time required to stop swimming were defined as variables, and the time required to stop swimming was the time value with the mean τ ± 50% fluctuation given by a uniform distribution. The time series change in the mean swimming speed of modeled larvae after passing through the light–dark boundary is shown in Supplementary Fig. S4c. The larvae which pass through the light–dark boundary from the dark side to light side do not respond and continue to swim at a speed of 2.0 mm/s. The period over which larvae stopped swimming ([2] in Supplementary Fig. S4b) did not affect the distribution of larvae in any of the simulations tested, and thus this period was set at a constant value of 4 min. Parameter values including the swimming velocity (2.0 mm/s) and the duration of the no swimming period (4 min.) were based on experimental measurements. Each simulation contains 1000 modeled larvae in a rectangular field.
Molecular phylogeny of candidate photoreceptor proteins of Acropora tenuis
Putative opsin and photolyase/cryptochrome (PLs/CRYs) homologs in Acropora tenuis were obtained from the genome sequence data of Acropora tenuis (The ReFuGe 2020 Consortium39, https://aten.reefgenomics.org/). Amino acid query sequences of previously defined opsins and PLs/CRYs were subjected to BLAST searches (BLASTP and TBLASTN) with an E-value cutoff of 10–10. Collected homologs were aligned and trimmed using PRANK40 and TrimAl41, and then the ML tree was reconstructed using RAxML-HPC version 842 assuming the LG + F + Γ (for opsins) or LG + Γ (for PLs/CRYs) model of protein evolution, which were the best-fit models for the aligned sequences selected by Aminosan43, respectively. The ML tree was visualized with FigTree software (https://tree.bio.ed.ac.uk/software/figtree/).
Ethics
All samples of Acropora tenuis were collected under Great Barrier Reef Marine Park Authority permits G12/35236.1 and G17/713999.1 and Okinawa prefectural government permission no. 29-73.
Source: Ecology - nature.com
