Zebrafish larvae exhibit rheotaxis and can escape a continuous suction source using their lateral line

Abstract

Zebrafish larvae show a robust behavior called rheotaxis, whereby they use their lateral line system to orient upstream in the presence of a steady current. At 5 days post fertilization, rheotactic larvae can detect and initiate a swimming burst away from a continuous point-source of suction. Burst distance and velocity increase when fish initiate bursts closer to the suction source where flow velocity is higher. We suggest that either the magnitude of the burst reflects the initial flow stimulus, or fish may continually sense flow during the burst to determine where to stop. By removing specific neuromasts of the posterior lateral line along the body, we show how the location and number of flow sensors play a role in detecting a continuous suction source. We show that the burst response critically depends on the presence of neuromasts on the tail. Flow information relayed by neuromasts appears to be involved in the selection of appropriate behavioral responses. We hypothesize that caudally located neuromasts may be preferentially connected to fast swimming spinal motor networks while rostrally located neuromasts are connected to slow swimming motor networks at an early age.

Figure 1. Larvae burst away from a continuous suction source at 5 days post fertilization.

A. Schematic of the experimental suction chamber which is immersed in a large volume bath and connected with tubing to a retention flask and vacuum motor. B. Comparison of two suction avoidance behaviors. Position of a larva (black line) drifting towards and then bursting away from the suction source (origin at 0 cm), and another larva continuously holding station instead of bursting (gray line). C. Time series of a rheotactic larva (gray circles) escaping the suction source (arrow) with a quick burst of swimming. Frames taken every 350 ms.

Figure 2. Larval zebrafish demonstrate rheotaxis by moving their body to orient upstream in the presence of current.

A. Characteristic example of a larva turning its body upstream to align with a current created by a continuous suction source. The initial body angle relative to the flow direction (t₆) decreases (gray arrows) with time such that at the end of a one second video sequence the body is aligned parallel to, and facing away from, the flow (t₀). B. All body angles decreased over time in the presence of a current, indicating a robust rheotactic response in freely swimming larvae with an intact lateral line (n = 8 larvae). The data for larva in A are highlighted by white circles and joined by a solid black line. The solid gray line for the last three data points (highlighted in the black box) represents the average change in body axis between t₃ and t₀. C. Euthanized, and therefore passively drifting, larvae show a tendency to slowly self-orient to the current (n = 7 larvae). The solid gray line represents the average rate of change in body angle between t₃ and t₀. D. Comparison of the change in body angle between timepoints t₃ and t⁰, when fish responded robustly to the flow. Live fish turn faster to align themselves with the flow than what is expected for a passively drifting fish (Student's unpaired, one-tailed T-test, * p<0.01, 7≤N≤8).

Figure 3. Characteristics of bursting behavior in the suction chamber for untreated larvae.

A. Percentage of fish initiating a burst at a given radial distance from the suction source, where the origin of the suction source is at 0 cm. Most larvae initiate a swimming burst between 0.5–1.0 cm (1–2 body lengths) away from the suction source, while far fewer larvae initiate a burst closer or further away (N = 98). B. Relationship between burst distance and location of burst initiation. The closer to the suction source a burst is initiated the farther the distance traveled during the burst (N = 98, R² = 0.09, p = 0.02). C. Velocity of passively drifting euthanized larvae as a function of distance from the suction source. The average drifting velocity of ten bodies was smoothed using a cubic spine (gray shaded area represents the standard error of the mean, N = 10). D. Average burst velocity as a function of the location of burst initiation. The closer to the suction source the burst is initiated, the faster the burst velocity (N = 98, R² = 0.06, p = 0.01).

Figure 4. The effects of selective neuromast ablation on the ability of larvae to avoid a suction source.

A. DASPEI-labeled neuromasts in 5 day post fertilization larvae with sections of the posterior lateral line (PLL) ablated with neomycin. Intact neuromasts labeled with DASPEI are highlighted with white arrowheads, while white boxes indicate regions where neomycin was applied. Note that due to the transparency of the larvae, at times labeled neuromasts from the opposite side of the body are seen. Five different treatments were tested, from top to bottom: (1) larvae with rostral neuromasts of the PLL ablated, (2) middle neuromasts of the PLL ablated, (3) caudal neuromasts of the PLL ablated, (4) complete PLL ablated and (5) sham treated control group. B. Percent larvae that escape (black bars) and are captured by (white bars) the suction source. There is a significant difference between the control (N = 78) and the complete PLL ablated group (N = 36) as well as between the control and the caudal neuromasts ablated group (N = 32). We also found a significant difference between the complete PLL ablated group and middle neuromasts ablated group (N = 27). No significant effects where found for the rostral neuromasts ablated group (N = 18). All groups were tested using a Fisher's exact test (***p<0.001, **p<0.01, *p<0.05). C. Time series showing the position of a larva with caudal neuromasts ablated (white circles) captured by the suction source (located at the origin of the coordinate system), and a larva with middle neuromasts ablated (black circles) bursting away from the suction source. Start and endpoint of each path are indicated.