Autoreceptor control of peptide/neurotransmitter corelease from PDF neurons determines allocation of circadian activity in drosophila

Abstract

Drosophila melanogaster flies concentrate behavioral activity around dawn and dusk. This organization of daily activity is controlled by central circadian clock neurons, including the lateral-ventral pacemaker neurons (LN(v)s) that secrete the neuropeptide PDF (pigment dispersing factor). Previous studies have demonstrated the requirement for PDF signaling to PDF receptor (PDFR)-expressing dorsal clock neurons in organizing circadian activity. Although LN(v)s also express functional PDFR, the role of these autoreceptors has remained enigmatic. Here, we show that (1) PDFR activation in LN(v)s shifts the balance of circadian activity from evening to morning, similar to behavioral responses to summer-like environmental conditions, and (2) this shift is mediated by stimulation of the Gα,s-cAMP pathway and a consequent change in PDF/neurotransmitter corelease from the LN(v)s. These results suggest another mechanism for environmental control of the allocation of circadian activity and provide new general insight into the role of neuropeptide autoreceptors in behavioral control circuits.

Figure 1. Activation of PDFR autoreceptors in LN_vs increases morningness and accelerates free-running rhythms

(A) Average LD locomotor histograms. Filled bars and unfilled bars represent 30 minute bin activity during light phase (ZT0-ZT12) and dark phase (ZT12-ZT0) respectively. Flies expressing t-PDF show greater morning anticipation with increasing numbers of UAS-t-PDF transgenes. Negative controls are flies expressing the inert t-PDF isoform, t-SCR, from six UAS-t-SCR transgenes (pdf > 6x t-SCR). See Supplemental Table 1 for Ns. (B) Average locomotor histograms on the first day of DD (DD1, ZT15-CT18) for the same flies as in (A). The size of morning peak relative to the evening peak increases with increasing dose of t-PDF. See Supplemental Table 1 for Ns. (C) Averaged actograms for 11 days in constant darkness (DD1 – DD11) for the same flies as in (A). Black bars indicate subjective night and grey bars indicate subjective day. See Supplemental Table 1 for Ns. (D) LN_v PDFR activation increases morning anticipatory activity during the 3 hours prior to lights-on (ZT21-ZT0) in LD in a dose-dependent manner. One-Way ANOVA indicates significant differences between genotypes (P < 0.0001), and Bonferroni All-Pairwise Multiple-Comparison test (α = 0.05) reveals significantly greater morning anticipatory activity in pdf > 4x t-PDF than in controls and pdf > 2x t-PDF flies, and even greater morning anticipatory activity in pdf > 6x t-PDF flies. (E) LN_v PDFR activation shifts circadian activity to the subjective morning, as revealed by comparisons of morning ratios, which is defined as the ratio of total activity during the morning peak to the total activity peak during the evening peak (see Methods). One-Way ANOVA indicate significant differences between genotypes (P < 0.0001), and Bonferroni All-Pairwise Multiple-Comparison Test (α = 0.05) reveal significantly greater morningness in pdf > 6x t-PDF flies. (F) Average free-running periods during the first 11 days in DD. Free-running rhythms accelerate with increasing t-PDF copy number. One-way ANOVA indicate significant difference among genotypes (P < 0.0001), and Bonferroni All-Pairwise Multiple-Comparison Test (α = 0.05) indicate significantly shorter free-running periods in pdf > 4x t-PDF and pdf > 6x t-PDF compared to the control. (D–G) Genotype labels at the top of each bar indicate significant differences from, a: pdf > 6x t-SCR, b: pdf > 2x t-PDF, c: pdf > 4x t-PDF, d: pdf > t-PDF, as detected by Bonferroni All-Pairwise Multiple Comparison Test (α = 0.05). Bars and error bars indicate mean and standard errors.

Figure 2. PDFR autoreceptor activation increases morningness through Gα,s-cAMP-pathway

(A) Inhibition of Gα,s or phosphodiesterase-8 overexpression in LN_vs suppress LN_v PDFR activation induced morningness increase. Line graphs represent average LD and DD1 activity of flies expressing Gα,s peptide inhibitor (Gα,s-11) or phosphodiesterase 8 (PDE8) with either 4x t-PDF or 4x t-SCR. One-way ANOVA indicate significant differences in morning anticipatory activity (P < 0.0001) and morning ratio (P < 0.0001) among genotypes. Increase in morning anticipatory activity induced by t-PDF expression is partially suppressed by Gα,s-11 and fully suppressed by PDE8 co-expression. t-PDF expression induced increase in morning ratio is fully suppressed by Gα,s-11 or PDE8 co-expression. See Supplemental Table 1 for Ns. (B) RNAi knockdown of Gα,s suppress morningness increase induced by PDFR activation in LN_vs. Line graphs represent average LD and DD1 activity of flies expressing Dicer, transgenes for RNAi knockdown of Gα,s (Gαs RNAi VDRC#24958 and Gαs RNAi DGRC) and either 4x t-PDF or 4x t-SCR. One-way ANOVA indicate significant differences in morning anticipatory activity (P < 0.0001) and morning ratio (P < 0.0001) among genotypes. Increase in morning anticipatory activity and morning ratio induced by t-PDF expression is fully suppressed by Gαs RNAi VDRC#24958 co-expression. Co-expression of Ga,s RNAi DGRC also fully suppresses the t-PDF induced morning ratio increase, but only partially suppresses morning anticipatory activity increase, likely reflecting the less complete Gα,s knockdown by Gαs RNAi DGRC than by Gαs RNAi VDRC#24958. See Supplemental Table 1 for Ns. (C) Activation of Gα,s pathway by expression of constitutively active Gα,s mutant (Gα,s[GTP]) increases morningness with or without PDFR activation. Line graphs represent average LD and DD1 activity of flies expressing Gα,s[GTP] with either 4x t-PDF or 4x t-SCR. One-way ANOVA indicate significant differences in morning anticipatory activity (P < 0.01) and morning ratio (P < 0.0001) among genotypes. Either t-PDF or Gα,s[GTP] expression significantly increased morning anticipatory activity or morning ratio, without no additive effects of t-PDF and Gα,s[GTP] co-expression on morningness. See Supplemental Table 1 for Ns. (A–C) Yellow indicate day, blue indicate night and subjective night and light blue indicates subjective day. n.s., not significant; *, • = 0.05; **, • = 0.01; ***, • = 0.001; significant difference detected by Bonferroni All-Pairwise Multiple-Comparison Test. Bars and error bars indicate means and standard errors.

Figure 3. Activation of DH31 receptors in LN_vs does not increase morningness

(A–C) Average LD (A) and DD1 (B) locomotor histograms and averaged actograms (C) of flies expressing t-DH31 in LN_vs from either of two different combinations of six independent UAS-t-DH31 transgenes (6x t-DH31(a) or 6x t-DH31(b)). See Supplemental Table 1 for Ns. (D) Mild decrease of morning anticipatory activity by LN_v DH31R activation. One-way ANOVA reveals significant differences (P < 0.05), with only the pdf > 6x t-DH31(b) combination exhibiting significantly less morning anticipatory activity. (E) Mild decrease in morningness by LN_v DH31R activation. One-way ANOVA reveals significant differences (P < 0.001) with only the pdf > 6x t-DH31(a) combination exhibiting significantly less morning ratio. (F) Free-running rhythms in DD are slightly shortened with LN_v DH31R activation. One-way ANOVA, p < 0.0001. (D–F) n.s., not significant; *, • = 0.05; **, • = 0.01; ***, • = 0.001; significant difference detected by Bonferroni All-Pairwise Multiple-Comparison Test. Bars and error bars indicate means and standard errors.

Figure 4. Activation of PDFR autoreceptors in LN_vs specifically depolarizes sLN_vs

(A) Representative 5 second traces of whole-cell current-clamp recordings of sLN_v and lLN_v clock neurons in pdf > 6x t-SCR and pdf > 6x t-PDF flies. Recordings were conducted at ZT22–23. (B) Average resting membrane potential (RMP) of sLN_v and lLN_vs in pdf > 6x t-SCR and pdf > 6x t-PDF flies. t-PDF expression specifically depolarizes sLN_vs, with no effect on lLN_vs. There are no effects on lLN_v RMP, but sLN_vs are significantly depolarized with t-PDF expression. *, p < 0.03, n.s., not significant, unpaired t-test. (C) Autocorrelation analysis of RMP oscillation of sLN_v and lLN_v cells in pdf > 6x t-SCR and pdf > 6x t-PDF flies. sLN_v RMP oscillation is completely suppressed by t-PDF expression, while lLN_vs are not affected. *, p < 0.02, n.s., not significant, Mann-Whitney U Test, See Methods. (D) Representative 5 second traces of whole-cell current-clamp recordings of sLN_v and lLN_v clock neurons in pdf > 6x t-SCR and pdf > 6x t-PDF flies in the presence of bath TTX. Recordings were conducted at ZT22–23. Note the absence of regular RMP oscillations and action potentials. (E) Average resting membrane potentials (RMP) of sLN_vs and lLN_vs in pdf > 6x t-SCR and pdf > 6x t-PDF flies in the presence of bath TTX. sLN_vs are significantly depolarized by t-PDF expression, while lLN_vs are not affected. *, p < 0.01, n.s., not significant, unpaired t-test. (B, C, E) Bars and error bars indicate mean and standard errors. Number of recorded cells is indicated in the bars.

Figure 5. PDFR activation in PDF-negative clock neurons is necessary and sufficient for morning activity

(A) PDFR activation in LN_vs without PDF secretion from LN_vs has no effect on circadian locomotor activity. Average LD and DD1 activity histograms of pdf⁰¹ null-mutant flies expressing 4x t-SCR, two transgene combinations of 2x t-PDF or 4x t-PDF. In LD, stereotypical lack of morning anticipation and phase-advanced evening anticipation of pdf⁰¹ null-mutant flies is seen for all genotypes with no differences in their circadian locomotor profiles. Morning anticipation phase scores indicate absent morning anticipation (phase score = 0.5) in all genotypes (One-way ANOVA, P > 0.85). In DD1, circadian activity peaks are severely dampened or absent in all genotypes. Bars and error bars indicate means and standard errors. (B) PDFR activation in PDF-negative clock neurons is sufficient for circadian activity peaks. Average LD and DD1 activity histograms of pdf⁰¹ null-mutant flies expressing t-SCR or t-PDF from one or two transgenes in PDF-negative clock neurons, using the combination of tim-GAL4 driver active in all clock neurons and pdf-GAL80 to suppress GAL4 activity in PDF-positive LN_vs. t-PDF expression in PDF-negative neurons of pdf⁰¹ null-mutant flies rescues robust morning anticipatory activity in LD as revealed by calculating morning anticipation phase scores. One-way ANOVA indicate significant differences among t-PDF expressing pdf⁰¹ null-mutants and their t-SCR expressing pdf⁰¹ null-mutant controls (P > 0.0001), and Bonferroni All-Pairwise Multiple-Comparison Test (*, α = 0.05) reveals significantly greater morning anticipation in t-PDF expressing pdf⁰¹ null-mutants compared to t-SCR expressing pdf⁰¹ null-mutant controls. In DD1, t-PDF expression in PDF-negative clock neurons of pdf⁰¹ null-mutant also rescues robust morning and evening circadian activity peaks that are severely damped in t-SCR-expressing controls. Bars and error bars indicate means and standard errors. (C) Representative images of anti-PDF immunohistochemistry and average normalized PDF levels in sLN_v dorsal terminals of pdf > 6x t-PDF and pdf > 6x t-SCR flies at ZT22. Bar graph shows pooled results from four independent experiments. PDF accumulation in sLN_v dorsal terminals is greater in pdf > 6x t-PDF flies than in pdf > 6x t-SCR controls. ***, P < 0.0001, un-paired t-test. Bars and error bars indicate means and standard errors. Ns are indicated in the bars.

Figure 6. Blocking classical synaptic transmission suppresses increased morningness induced by LN_v PDFR autoreceptor activation

(A–C) Average LD (A) and DD1 (B) locomotor histograms and averaged actograms (C) of flies co-expressing 4x t-PDF or 4x t-SCR with tetanus toxin light chain (TnTLC). Increased morningness induced by PDFR activation in LN_vs is prevented when classical synaptic transmission is blocked by TnTLC co-expression. See Supplemental Table 1 for Ns. (D) TnTLC co-expression strongly suppresses the increase in morning anticipatory activity induced by LN_v PDFR activation. One-way ANOVA indicates significant differences in morning ratio between genotypes (P < 0.0001), and Bonferroni All-Pairwise Multiple-Comparison Test reveals significantly less morning anticipatory activity in pdf > 4x t-PDF + TnTLC flies than in pdf > 4x t-PDF flies. (E) TnTLC co-expression completely inhibits the morningness increase induced by LN_v PDFR activation. One-way ANOVA indicates significant differences in morning ratio between genotypes (P < 0.001), and Bonferroni All-Pairwise Multiple-Comparison Test reveals significant increase in morning ratio by t-PDF expression and no significant change in morning ratio by t-PDF expression when TnTLC is co-expressed. (F) TnTLC co-expression partially suppresses the period shortening effect of LN_v PDFR activation without affecting free-running period on its own. One-way ANOVA indicates significant differences in morning ratio between genotypes (p < 0.0001). (D–F) **, p<0.01, ***, p < 0.001, significant differences detected by Bonferroni All-Pairwise Multiple-Comparison Test. n.s., not significant. Bars and error bars indicate means and standard errors.

Figure 7. Model of PDFR autoreceptor signaling in LN_vs

PDFR in sLN_vs activates Gα,s and stimulates cAMP synthesis by adenylyl cyclase. The subsequent increase in intracellular cAMP leads to membrane depolarization, increased PDF secretion and altered secretion of an unknown neurotransmitter (NT). Increased PDF and neurotransmitter output from sLN_vs induced by PDFR activation acts on secretion acts on dorsal clock neurons to increase circadian morning activity. PDFR in sLN_vs is activated by increased PDF secretion by lLN_vs in response to light.