Triggering and modulation of a complex behavior by a single peptidergic command neuron in Drosophila

Abstract

At the end of their growth phase, Drosophila larvae remodel their bodies, glue themselves to a substrate, and harden their cuticle in preparation for metamorphosis. This process-termed pupariation-is triggered by a surge in the hormone ecdysone. Substrate attachment is achieved by a pupariation subprogram called glue expulsion and spreading behavior (GSB). An epidermis-to-CNS Dilp8-Lgr3 relaxin signaling event that occurs downstream of ecdysone is critical for unlocking progression of the pupariation motor program toward GSB, but the factors and circuits acting downstream of Lgr3 signaling remain unknown. Here, using cell-type-specific RNA interference and behavioral monitoring, we identify Myoinhibiting peptide (Mip) as a neuromodulator of multiple GSB action components, such as tetanic contraction, peristaltic contraction alternation, and head-waving. Mip is required in a pair of brain descending neurons, which act temporally downstream of Dilp8-Lgr3 signaling. Mip modulates GSB via ventral nerve cord neurons expressing its conserved receptor, sex peptide receptor (SPR). Silencing of Mip descending neurons by hyperpolarization completely abrogates GSB, while their optogenetic activation at a restricted competence time window triggers GSB-like behavior. Hence, Mip descending neurons have at least two functions: to act as GSB command neurons and to secrete Mip to modulate GSB action components. Our results provide insight into conserved aspects of Mip-SPR signaling in animals, reveal the complexity of GSB control, and contribute to the understanding of how multistep innate behaviors are coordinated in time and with other developmental processes through command neurons and neuropeptidergic signaling.

Keywords: Drosophila; innate behavior; neuromodulation; neuropeptide; relaxin.

Fig. 1.

Proper glue expulsion and spreading behavior requires Mip in neurons. (A) Representative profile of muscle contractions during the stages of the pupariation motor program detected with muscle-expressed GCaMP (mhc>>GCaMP). (B) Mip RNAi in R48H10> cells does not induce gross defects in puparium shape. (C) RNAi screen to detect neuropeptides required for the PMP by crossing R48H10> with RNAi lines targeting the indicated genes. Percentage of animals performing normal GSB (Top, N ≥ 5, except CCAP and ETH, N = 4) and puparium aspect ratio (Bottom, N ≥ 43). Dots represent one animal. Horizontal orange line, median. Kruskal–Wallis and Dunn’s test. (D) Knockdown of Mip in R48H10> cells suppresses glue expulsion. Sgs3::GFP labels a glue protein (cyan) of representative wandering larva and T0 white prepupa. (E) Percentage of larvae that expel glue. (F) Panneuronal Mip RNAi alters GSB and suppresses glue expulsion. Percentage of larvae that perform a normal GSB (Top) and release glue (Bottom). Blue numbers, N. Pairwise Fisher’s exact test, alpha = 0.05 with Bonferroni correction. Same blue letter, P > 0.008 in E; P > 0.016 in F. Data for E, F, and S2D were collected in parallel as part of the same experiment and represented separately for clarity. For this reason, UAS-Mip-IR/+ values are the same in all cited panels. (Scale bar, 1 mm.)

Fig. 2.

Mip is required in Mip2M descending neurons for proper glue expulsion and spreading behavior. (A) Third instar larval CNS stained with anti-Mip. VNC: Ventral nerve cord, br: brain. (B) Max projection of CNS from R48H10>mCD8::RFP (magenta) larvae stained with anti-Mip antibody (green). Arrows and arrowheads point to double-labeled cells. f: esophageal foramen. (C) Detail of protocerebral double-labeled pair (arrows). (D) Detail of 6VNC neuron cell bodies (arrowheads). T1-T3, thoracic segments 1 to 3. (E) All Mip2M>-positive cells express Mip (arrowheads). The Top and Bottom images are from different confocal stacks. Mip RNAi in Mip2M> cells affects GSB (F) and glue expulsion (G). Blue numbers, N. Asterisk, Fisher’s exact test with Bonferroni correction when required, P = 0.0008 in F, P < 0.00001 against controls in G. (H) Dendritic and axonal compartments of Mip2M neurons labeled with Denmark and syt::GFP respectively. (I) Single-cell clone of a Mip descending neuron (Mip2M>MCFO). CI: central intermediate, CL: central lateral longitudinal tracts. (Scale bar, 50 µm.) Data for G and S3G were collected in parallel as part of the same experiment and represented separately for clarity. For this reason, UAS-Mip-IR/+ values are the same in all cited panels.

Fig. 3.

Mip descending neuron activity and neuropeptide expression are indispensable for GSB. Mip¹ mutant larvae perform an abnormal GSB and retain glue in the salivary glands. (A) Percentage of larvae that perform a normal GSB and expel glue (marked by Sgs3::GFP). (B) Mip2M> expression of UAS-Mip-UTRs rescues glue expulsion to control levels. (C and D) Percentage of larvae that perform a normal GSB and expel glue upon silencing of Mip2M neurons with tetanus toxin (TNT) (C) and hyperpolarizing Kir2.1 channel (D). An inactive version of TNT (TNT(−)) served as control. Kir2.1 expression completely blocked GSB. (E) Representative traces of muscle GCaMP fluctuations of the indicated genotypes. GSB is labeled with a gray horizontal bar. Control genotypes perform a strong tetanic contraction that expels glue (orange dot). Mip RNAi in Mip2M neurons impedes tetanic contraction, while Kir2.1 silencing completely blocks GSB. Blue numbers, N. Fisher’s exact test with Bonferroni correction when required, alpha 0.05. Same blue letter: P > 0.016 in A and D; P > 0.008 in B. Asterisk: P < 0.005 in C, P < 0.00001 in A and D.

Fig. 4.

Optogenetic stimulation of Mip2M neurons induces GSB-like behaviors during a specific developmental window. Wandering larvae fed with all-trans-retinal (ATR) or not (−ATR) were optogenetically stimulated every 5 min. Response to light was evaluated in 1-h time windows beginning 180 and 60 min before post-GSB (“3 to 2 h before post-GSB” and “1 to 0 h before post-GSB” respectively). (A) Schematics of the selected time windows relative to the pupariation motor program. (B) Rectangular boxes represent light stimuli, filled with pink if the stimulation induced a GSB-like response or white, otherwise. The period when larvae performed pre-GSB contractions is shaded in green. 5-min intervals without light stimulation are represented by a line, or in yellow when larvae executed a spontaneous GSB. ND, not detected, larvae reached post-GSB before 2 h of recording. (C) Quantification of the percentage of light pulses that were followed by a GSB-like behavior per larva in −ATR and +ATR conditions at the indicated time periods. Dots: one animal horizontal bar is mean value. Blue numbers: N. Asterisks: Kruskal–Wallis followed by Dunn’s test, P < 0.007 for +ATR comparisons. Mann–Whitney test, P < 0.0001 for 1 to 0 h, −ATR vs. +ATR.

Fig. 5.

SPR is required in VNC neurons for proper GSB execution. (A) The SPR gene is required for proper GSB, while the SP gene is not. Percentage of animals that perform a normal GSB [SPR (−/−): Df(1)Exel6234. SP (−/−): ∆¹³⁰. (+/+, +/+), SP[∆¹³⁰] background controls]. (B) Percentage of larvae that expel glue (marked by Sgs3::GFP). (C) Confocal stack projections of L3 larval CNS from SPR>mCD8::GFP animals. DAPI counterstain, nuclei (blue). Dotted line, CNS contour. (Scale bar, 50 µm.) (D) Percentage of animals that execute a normal GSB. Panneuronal (57C10>) or VNC-specific (tsh>) RNAi knockdown of SPR alters the execution of GSB. Two different SPR RNAi constructs were tested. Blue numbers, N. Pairwise Fisher’s exact test with Bonferroni correction at alpha = 0.05. Same blue letter: P > 0.016 in A, P > 0.008 in B. Asterisks: P < 0.0001 against corresponding control conditions in D.

Fig. 6.

Mip from Mip2M neurons ensures a stereotyped motor pattern during GSB. (A) Categories used to classify behavioral subunits that comprise GSB. (B) Representative ethograms (N = 9 to 10) displaying the sequence of behaviors and duration of each element of GSB in wild-type animals (+/+) and Mip mutants. Each row represents the color-coded behavioral sequence of one animal expressing mhc>>GCaMP. (C) Quantification of the effect of Mip mutation on the occurrence of backward short peristaltic waves. (D) The back-and-forth alternation of peristaltic waves is lost in Mip¹mutants, resulting in a low opposing transition index. (E) Panneuronal (R57C10-GAL4) or Mip2M-neuron-specific Mip RNAi affects GSB behavioral sequence. Depicted are 9 to 10 representative ethograms of indicated genotypes. (F) The consistency of the back-and-forth transitions is reduced upon panneuronal or Mip2M Mip knockdown, resulting in a low opposing transition index. (G) GSB behavioral sequence is restored in Mip mutants by expression of the UAS-Mip-UTRs construct with Mip2M-GAL4. (H) The back-and-forth alternation of peristaltic waves is rescued in Mip¹mutants expressing UAS-Mip-UTRs in Mip descending neurons, increasing the opposing transition index to control values. Blue numbers, total number of larvae. C, D, F, and H: Dots represent one larva; the horizontal line is mean value. Asterisk: Mann–Whitney test, P < 0.0001 in C and D. Kruskal–Wallis and Dunn’s test, P < 0.01 against respective controls.

Fig. 7.

SPR is required in neurons to determine and shape GSB action components. (A) Ethograms displaying the sequence and duration of GSB component acts. Panneuronal (R57C10-GAL4) SPR RNAi phenocopies Mip mutants. (B) The back-and-forth alternation of peristaltic waves is lost upon panneuronal SPR knockdown, resulting in a reduced opposing transition index. (C) Quantification of the effect of SPR knockdown on the occurrence of backward-short peristaltic waves. Blue numbers, N. B and C: Dots represent one larva, horizontal line is mean value. Asterisks: Kruskal–Wallis and Dunn’s test, P < 0.05 against controls in B and C.

Fig. 8.

Model of Mip and SPR activity during GSB. Dilp8 production in the larval cuticle epidermis during the end of the third instar larval growth period signals to Lgr3+ neurons in the VNC and unlocks the progression of the Pupariation Motor Program, of which GSB is one of the stages (17). Lgr3+ neurons in turn may communicate with Mip descending neurons (MipDNs) in a still uncharacterized way. Depolarization of MipDNs triggers GSB, while secretion of Mip modulates the activity of downstream SPR+ neurons in the VNC to orchestrate the characteristic sequence of GSB action components. Depolarization of MipDNs in the absence of Mip (Mip mutant) leads to an abnormal GSB, characterized by an uncoordinated subset of its action components, while MipDN silencing by hyperpolarization completely abrogates the behavior. In the absence of Mip or Mip-DN neuron activation, the rest of the pupariation motor program occurs and is mostly normal.