Melatonin action in a midbrain vocal-acoustic network

Abstract

Melatonin is a well-documented time-keeping hormone that can entrain an individual's physiology and behavior to the day-night cycle, though surprisingly little is known about its influence on the neural basis of social behavior, including vocalization. Male midshipman fish (Porichthys notatus) produce several call types distinguishable by duration and by daily and seasonal cycles in their production. We investigated melatonin's influence on the known nocturnal- and breeding season-dependent increase in excitability of the midshipman's vocal network (VN) that directly patterns natural calls. VN output is readily recorded from the vocal nerve as a 'fictive call'. Five days of constant light significantly increased stimulus threshold levels for calls electrically evoked from vocally active sites in the medial midbrain, supporting previous findings that light suppresses VN excitability, while 2-iodomelatonin (2-IMel; a melatonin analog) implantation decreased threshold. 2-IMel also increased fictive call duration evoked from medial sites as well as lateral midbrain sites that produced several-fold longer calls irrespective of photoregime or drug treatment. When stimulus intensity was incrementally increased, 2-IMel increased duration only at lateral sites, suggesting that melatonin action is stronger in the lateral midbrain. For animals receiving 5 days of constant darkness, known to increase VN excitability, systemic injections of either of two mammalian melatonin receptor antagonists increased threshold and decreased duration for calls evoked from medial sites. Our results demonstrate melatonin modulation of VN excitability and suggest that social context-dependent call types differing in duration may be determined by neuro-hormonal action within specific regions of a midbrain vocal-acoustic network.

Keywords: Melatonin; Midshipman fish; Periaqueductal gray; Vocalization.

Fig. 1.

Midshipman vocalizations and experimental setup. (A) Natural (adapted from Rubow and Bass, 2009) and fictive calls of male midshipman fish. Long-duration advertisement hums are produced in the summer at nighttime. Short-duration agonistic grunt trains can be produced at any time of day or year. (B) A schematic sagittal view of the midshipman brain showing the vocal control network (adapted from Chagnaud et al., 2011). Stimulation in the midbrain vocal-acoustic complex (mVAC) evokes readily recorded fictive calls from the vocal nerve. The mVAC receives input from the forebrain vocal-acoustic complex (fVAC), from which fictive calls can also be evoked. The mVAC drives the hindbrain vocal pattern generator, which provides a precise and synchronous code for sonic muscle contraction and consists of the vocal pre-pacemaker nucleus (VPP), vocal pacemaker nucleus (VPN) and vocal motor nucleus (VMN). (C,D) Schematics of photoperiod, drug treatment and neurophysiology stimulation regimes used in this study. (C) 5LL fish were implanted with 2-indolmelatonin (2-IMel) or vehicle before subjective lights-off and moved to constant light (LL) for 5 days. Light gray boxes represent subjective night. 5DD fish were held in 5 days of constant darkness (DD) and injected daily around subjective lights-off with vehicle, luzindole or 4P-PDOT. Dark gray boxes represent subjective day. Black arrows indicate time of treatment, and the gray arrow indicates time of neurophysiology for both 5LL and 5DD fish. Sessions (120 min) consisted of stimulation at indicated times after fish were acclimated on the rig for 1 h; 40 stimuli were delivered at each time point except for 120 min, when an additional 60 stimuli for a total of 100 were delivered (highlighted by *). Ten minutes later, a stimulus–response curve (SRC) was collected without moving the electrode, where stimulus intensity was increased to the indicated % of baseline threshold, recording 10 fictive calls every 5 min. The stimulus electrode was then immediately moved to a lateral site in the midbrain to collect another SRC. (D) LD fish were tested for medial and lateral midbrain stimulation comparisons and were housed in 15 h:9 h light:dark. The neurophysiology stimulation regime followed that of 5LL and 5DD animals except only 40 stimuli were delivered at the 120 min trial, and only one SRC was collected after 120 min sessions without moving the electrode.

Fig. 2.

Effects of photoregime on vocal excitability inferred from control animals. Fictive call duration (A) and threshold (B) over 120 min sessions, as well as medial (C) and lateral (D) SRCs recorded from control fish of each photoregime/treatment group: LD, 5LL and 5DD. (E) Threshold levels at baseline trials of medial and lateral SRCs. For LD fish that received lateral midbrain stimulation only, lateral SRCs were recorded without prior medial SRCs (Lateral 1st). All other groups received medial stimulation first (Medial 1st). Data are presented as means ± s.e. n.s., non-significant; *P<0.042.

Fig. 3.

2-IMel increases vocal excitability in animals kept under 5LL. (A) 2-IMel treatment increased fictive call duration in a time-dependent manner during the 120 min stimulus trial and decreased stimulus threshold (B). (C,D) 2-IMel increased fictive call duration in stimulus response curves (SRC) only when lateral midbrain was stimulated. (E) 2-IMel decreased stimulus threshold measured at the onset of SRCs. All fish were stimulated at a medial site first for the 120 min session and SRC, followed by recording a second SRC at a lateral site. Data are presented as means ± s.e. n.s., non-significant. Note for the 120 min session duration (A), the means from the first 40 responses were separated from the last 60 fictive call responses (displaced to the right), hence the two data points at 120 min.

Fig. 4.

Effects of melatonin receptor antagonists on vocal excitability in animals kept under 5DD. Melatonin receptor antagonists luzindole and 4P-PDOT have mixed effects on fictive call 120 min session duration (A), 120 min session threshold (B), medial SRC duration (C) and lateral SRC duration (D). (E) Treatments had no effect on SRC threshold levels, which were higher at lateral sites. Data are presented as means ± s.e. n.s., non-significant. Note for 120 min session duration measurements (A), the means from the first 40 responses were separated from the last 60 fictive call responses (displaced to the right), hence the two data points at 120 min.

Fig. 5.

Vocal attributes depend on site of midbrain stimulation. (A,B) Five representative fictive vocal responses evoked by medial (Ai) and lateral (Bi) midbrain stimulation at 200% SRC, taken from a fish that was treated with 2-IMel. Fictive vocal traces outlined in gray boxes were enlarged in Aii and Bii. Gray arrowheads point to stimulus artifacts. From the same fish and stimulation trial, fictive call firing frequency, measured in inter-pulse intervals (IPI), taken from 200% threshold medial SRC trial (Aiii) and lateral SRC trial (Biii) showed no significant site-dependent differences. (C–F) Quantification of fictive call duration (C), stimulus threshold (D), latency (E) and stimulus response curves (F) resulting from either medial or lateral midbrain stimulation in non-treated fish held in normal LD cycles (n=5 per group). Data are presented as means ± s.e. All stimulus-site comparisons shown in C–F are statistically significant (P<0.006).