EAAT5b and EAAT7 mediate UV-blue and red but not green light signal transmission.

Monochromatic ERG was recorded on eaat5b-/- (A-C), eaat7-/- (D-F) and eaat5b-/-; eaat7-/- (G-I). B-wave amplitudes are plotted in box-and-whisker plots with fish-specific data overlaid as a swarmplot. Homozygous mutation in eaat5b or in eaat7 causes a decrease in the b-wave if stimulated red light (C and F) but not if stimulated with green or UV-blue light (A, B, D and E). Double KO animals display a reduced b-wave amplitude for UV-blue (G) and red light (I) but not for green light (H).

EAAT5b distributes anisotropically to EAAT7 and focuses in the Strike Zone in contact with UV cones.

(A) Projection of the 6 dpf whole larval retina stained for EAAT5b (magenta) and EAAT7 (yellow) shows that both proteins appear co-expressed along most of the OPL apart from the temporo-ventral area called SZ where relative intensity of EAAT7 drops while EAAT5b raises, indicating distribution anisotropy. (B) Schematics depicting the process used to measure the relative intensity of fluorescent signals along the OPL. Left: One section (retinal midsection) where most of the photoreceptor layer is perpendicular to the z-axis is selected. Right: on the selected section, the signal intensity for the different channel is measured by tracing a line along the visible OPL starting from the ventral extreme of the SZ and the average grey level under the trace is recorded and normalized to have values between zero and one for all channels. (C) Virtual midsection of a 6 dpf WT retina with nuclei (grey), UV cones (cyan), EAAT5b (magenta) and EAAT7 (yellow) labelled. (D) Top: Relative intensity values of the three fluorescent channels along the whole OPL. Full-colored line represents the traces’ mean and semitransparent area represents 95% CI. Bottom: Pearson’s correlation matrix pooling all the measured traces. While the overall correlation score is relatively high, probably due to partial co-localization, EAAT5b and UV cone pedicles appear to define a cluster on their own when compared to Eaat7 traces. (E) Virtual midsection of a 6 dpf WT retina with nuclei (grey), Red cones (cyan), EAAT5b (magenta) and EAAT7 (yellow) labelled. (F) Top: Relative intensity values of the three fluorescent channels along the whole OPL. Full-colored line represents the traces’ mean and semitransparent area represents 95% CI. Bottom: Pearson’s correlation matrix pooling all the measured traces. Notably, while EAAT7 and EAAT5b still correlate with a high score amongst each other, the former’s traces appear to be more similar to red cone pedicles compared to EAAT5b. This is probably due to the differences at the Strike Zone. For the whole panel, “N”: nasal, “D”: dorsal, “T”: temporal, “V”: ventral, “SZ”: strike zone.

UV stimuli can be used to simulate UV-bright moving paramecia in a LED-based custom setup.

(A) Schematics of the custom setup used to deliver virtual prey-capture stimuli. (B) Schematics representing a section of the UV-light stimulator and approximate light pathway through pinholes and lenses. (C) Diagram representing the stimulus paradigm during one iteration. After two seconds from the start of the round, the LED turns ON along with the servomotor, which turns left to right by 20° (∼90° on the projected screen). This takes three seconds, after which the LED turns OFF and the servomotor resets position. The same operation reiterates in the opposite direction after four seconds. (D) Sample tracking graphs from the setup. Top panel represents eye-specific angle variations over time (blue and green) with rolling averages used to identify events in orange. The bottom panel represents tail total curvature over time (red) with rolling average used to identify events in orange. Dots represent a detected event. (E) Relative intensities of the UV and yellow LEDs used for the validation measured by spectrometer.

Larval zebrafish show hunting response preferentially to UV stimuli in an intensity-dependent trend.

6 dpf WT zebrafish larvae were exposed to yellow (A) and UV (B) full intensity hunting stimuli. While response was extremely low for the yellow light (1 response overall; A), UV stimuli elicited consistent response in the firsts stimulation rounds, gradually waning through each repetition (B). WT larvae exposed with incremental UV light intensity showed increasing response rates, reaching a response plateau between the last two intensity levels (C). For all graphs in the figure: top shows fish specific results through the whole assay (left) and pooled events distribution over single epoch (right); bottom shows summary statistics as response percentages (reactions/fish on the left, reaction/fish/epoch on the right).

eaat5b KO mutants do not show a significant difference in UV hunting response compared to WT.

eaat5b mutants (n=23) begin reacting to the incremental prey stimulus one intensity-step later than WT (n=25) and display overall lower reactions/fish. While WT larvae start showing responses with stimuli delivered already at 30% of intensity (A), eaat5b mutants start showing reactions from 40% (B left). WT larvae reach peak responses at intensity 80% with a markedly higher response (A) compared to eaat5b mutants (B). Pooled responses proportion does not show any significant difference between WT and eaat5b mutant fish (two proportions z-test p>0.05) (C left). Comparing stimulus intensity-specific responses also leads to a non-significant difference (Kolmogorov-Smirnov test p>0.05) (C right). For all graphs in the figure: top shows fish specific results through the whole assay (left) and pooled events distribution over single epoch (right); bottom shows summary statistics as total response proportions (successful hunting responses/ delivered stimuli) and intensity specific percentages (reaction percentage/ stimulus intensity).

eaat7 KO mutants exhibit a higher peak response to an artificial UV prey stimulus compared to WT.

eaat7 mutants begin reacting to the incremental prey stimulus one intensity-step later than WT and display overall higher reactions/fish. While WT larvae start showing responses with stimuli delivered at 30% of intensity (A left), eaat7 mutants start showing reactions from 20% (B left). Both WT and eaat7 mutant larvae reach peak responses at intensity 90% but surprisingly fish lacking the glutamate transporter show a markedly higher response (A left) compared to WT (B left). Pooled responses proportion is significantly increased in eaat7 mutant fish compared to WT (two proportions z-test p=0.0011) (C left). Comparing stimulus intensity-specific responses leads to a non-significant difference instead (Kolmogorov-Smirnov test p>0.05) (C right). For all graphs in the figure: top shows fish specific results through the whole assay (left) and pooled events distribution over single epoch (right); bottom shows summary statistics as total response proportions (successful hunting responses/ delivered stimuli) and intensity specific percentages (reaction percentage/ stimulus intensity).

eaat5b but not eaat7 mutants show lower response to high contrast long wavelength OMR.

(A) To test long-wavelength range sensitivity, we build an OMR setup where a projector sends moving grating at the bottom of an arena organized in rows. Fish swimming in the rows are tracked by a high-speed camera. (B) We provide a long wavelength targeting paradigm divided in four steps, where two centering stimuli composed by black and white gratings moving towards the center of the arena, guide the larvae in a central position. After each centering stimulus, still (during the Habituation phase) and rightward-moving (during the Stimulus phase) gratings composed of black and red (at different contrasts) bands are displayed and the larval rightward movement percentage over total movement is calculated for both the long-wavelength phases. (C) The setup was used for testing sensitivity loss in both eaat mutants. As expected, eaat5b mutants have an overall reduced response to the stimulus (top left), particularly significant at higher contrasts (Repeated measure Anova; p<0.05). Conversely, eaat7 mutants do not show such a defect (top right). In all the tests, fish exposed to a still stimulus (Habituation phase) have no preference independently of the contrast displayed.

mGluR6b OPL distribution peaks in the SZ in a similar pattern to Eaat5b.

(A) Virtual midsection of a 6 dpf WT retina with nuclei (grey), mGluR6b (cyan) and EAAT5b (magenta) labelled.

(B) Relative intensity values of the two fluorescent channels along the whole OPL. Full-colored line represents the traces’ mean and semitransparent area represents 95% CI.

eaat5b and eaat7 mutants show same total activity as their respective control groups during black and red OMR assay.

Average total movement of eaats mutant larvae and their respective control groups exposed to red and black moving gratings at different contrasts. Overall, no major differences in activity can be detected in any of the groups. Both controls and eaat mutants show an increase in overall activity while the gratings are moving (top), while activity remains relatively stable when the stimulus remains still (bottom).