Regulation of circadian behavioral output via a MicroRNA-JAK/STAT circuit

Abstract

Although molecular components of the circadian clock are known, mechanisms that transmit signals from the clock and produce rhythmic behavior are poorly understood. We find that the microRNA miR-279 regulates the JAK/STAT pathway to drive rest:activity rhythms in Drosophila. Overexpression of microRNA miR-279 or miR-279 deletion attenuates rest:activity rhythms. Oscillations of the clock protein PERIOD are normal in pacemaker neurons lacking miR-279, suggesting that miR-279 acts downstream of the clock. We identify the JAK/STAT ligand, Upd, as a target of miR-279 and show that knockdown of Upd rescues the behavioral phenotype of miR-279 mutants. Manipulations of the JAK/STAT pathway also disrupt circadian rhythms. In addition, central clock neurons project in the vicinity of Upd-expressing neurons, providing a possible physical connection by which the central clock could regulate JAK/STAT signaling to control rest:activity rhythms.

Figure 1

Over-expression of miR-279 disrupts locomotor activity rhythms. (A) Map of the miR-279 locus and transgenic constructs. An EP element (NE95-11-24, black triangle) is mapped 1.3kb upstream of miR-279. Two UAS-miR-279 transgenic lines were generated with a 1kb (L) or a 150bp (S) genomic region of miR-279 fused to UAS. A miR-279-GAL4 reporter (Cayirlioglu et al., 2008) comprises the entire promoter of miR-279 fused to GAL4. (B) Over-expression of miR-279 by driving expression of NE95-11-24 or a UAS-miR-279 (L) transgene. q-PCR analysis of total RNA prepared from adult heads. The ratio of mature miR-279/2s rRNA was plotted as mean ± SD (** P<0.001, * P<0.01, by Student’s t-test). Levels were normalized relative to the levels seen in timG4/+ control flies (set as 1). (C) Over-expression of miR-279 leads to behavioral arrhythmia in DD. The genotypes are indicated on top of the panels. The grey and black bars indicate subjective day and night respectively. Average periods (τ) of rhythmic flies are shown at the bottom of the panels. Representative activity records are shown. (D) Panneuronal induction of miR-279 in adulthood leads to a long period, which eventually degenerates into arrhythmia. Flies were reared and then aged for 3 days following eclosion on regular food. They were fed either 500μM RU486 or ethanol (EtOH, vehicle control) from the time of entrainment. Average periods of the first 5 days in DD (τ) are shown beneath the panels. See also Figure S1.

Figure 2

Deletion of miR-279 reduces behavioral rhythmicity without changing PER cycling in small LNvs. (A) The miR-279 excision allele and genomic rescue construct (Cayirlioglu et al., 2008). The null allele ex117-1 (dashed line, Δ1.2kb) completely deletes the miR-279 gene. The miR-279 genomic construct (filled line, 3kb) comprises the entire promoter and coding region of miR-279. (B) The ex117-1 allele completely abolishes production of mature miR-279. q-PCR analysis of total RNA prepared from adult brains. The ratio of mature miR-279/2s rRNA was plotted as mean ± SD (** P<0.001, by Student’s t-test). Levels were normalized relative to those seen in w¹¹¹⁸ control (set as 1). (C) Deletion of miR-279 disrupts activity rhythms, which can be rescued by introducing a miR-279 genomic transgene. After being outcrossed 7 times into a w¹¹¹⁸ background, flies homozygous for ex117-1 live up to two weeks as adults. w¹¹¹⁸ flies derived from siblings during the last outcross were used as wild-type control (indicated as “sibling”). (D) The null allele of miR-279 shows normal PER oscillations and PDF expression in central clock cells. Brains were dissected and stained with PER (red) and PDF (green) antibodies on the third day of DD at the indicated circadian times (CT). See also Figure S2.

Figure 3

Unpaired (upd) is a direct target of miR-279. (A) RNAi knock-down of the JAK/STAT ligand upd abolishes behavioral rhythms. Expression of upd-RNAi in either miR-279 or tim-expressing neurons, but not in Pdf neurons, causes arrhythmia. (B) A highly conserved 8 nucleotide miR-279 target site in the 3′UTR of upd mRNA. A predicted miR-279 binding site (shaded boxes) is conserved within the 3′UTR sequences of upd mRNAs from different Drosophila species (TargetScanFly 5.1). Base-pairing between the consensus target site and the mature miR-279 is shown. A miR-279 mutant containing mismatched nucleotides (green) is shown at the bottom of the panel. (C) miR-279 inhibits expression of an upd 3′UTR-luciferase reporter in cultured Drosophila S2 cells. The entire 750bp sequence of upd 3′UTR, containing one conserved 8mer (red box, by TargetScanFly 5.1) and 4 other putative miR-279 target sites (pink boxes, by RNAhybrid), was fused to a firefly luciferase (F luc, yellow bar) reporter construct (shown beneath the panel). An empty vector, or a wild-type or a mutant miR-279 (as shown in panel B) expression plasmid was cotransfected with the firefly luciferase reporter plasmid. Renilla luciferase was used as transfection control. The ratio of firefly/Renilla luciferase is plotted as mean ± SD of triplicate data points (** P<0.001, by Student’s t-test). The control with the empty vector is set as 1. (D) miR-279 is a negative regulator of upd mRNA in vivo. q-PCR analysis of total RNA prepared from fly brains (red bars) or fly heads (blue bars). The level of upd in ex117-1 was normalized to that in w¹¹¹⁸ control (set as 1, red bars). The levels of upd in flies over-expressing miR-279 were normalized to those in timG4/+ control (set as 1, blue bars). The ratios of upd/actin mRNA are plotted as mean ± SD (** P<0.001, *P < 0.01, by Student’s t-test). (E) RNAi knock-down of upd rescues the arrhythmia of miR-279 nulls. TUG is a comparatively weak tim-GAL4 driver (see text). The ex117-1 mutant and flies expressing upd-RNAi with TUG alone served as controls. See Also Figure S3 and Supplemental Table.

Figure 4

The JAK/STAT signaling pathway is required for locomotor activity rhythms. (A) Panneuronal induction of upd in adulthood leads to arrhythmia. Flies were reared on regular food. 500μM RU486 or ethanol (control) was administered in the food from the time of entrainment. (B) Blocking JAK/STAT signaling with a dominant-negative form of the receptor domeless (dome^ΔCYT) leads to arrhythmia. The dome-GAL4 driver (Ghiglione et al., 2002) causes lethality in males, and thus only females were tested. (C) A hypomorphic mutation of JAK kinase hopscotch (hop²⁵) causes loss of activity rhythms. The hop²⁵ allele is a Q246K point mutation of the hop gene (Luo et al., 1999). (D) Over-expression of a constitutively active form of hop (hop^Tuml) in the adult nervous system disrupts activity rhythms. The gain-of-function mutation, hop^Tuml, is a G341E single amino acid substitution in the hop gene (Harrison et al., 1995; Luo et al., 1995). (E) The hop²⁵ mutants show normal PER cycling and PDF expression in central pacemaker neurons. Brains were dissected and stained with PER (red) and PDF (green) antibodies on the third day of DD at the indicated circadian times (CT). See Also Figure S4.

Figure 5

JAK/STAT signaling is downstream of the central clock and the PDF receptor (PDFR). Flies of the indicated genotypes were collected at different circadian times (CT) on the first (DD1) or second (DD2) day of DD. Protein extracts of brains were subjected to western blot analysis using antibodies specific for STAT92E and loading controls (MAPK or HSP70). Multiple bands are detected for STAT92E, of which the fastest mobility forms are expressed cyclically (indicated with asterisks). Each experiment was conducted three times and quantified STAT92E and loading control levels were normalized to those at CT2 of wild-type controls in each gel. The quantification curves in each panel were plotted as the average ± standard error of the mean (SEM) of three independent western blots. (A and B) STAT92E protein levels cycle in fly brains and are regulated by the circadian clock and PDFR. STAT92E levels at the trough (CT2 and CT20) are significantly lower than at the peak (CT8 and CT14) in w¹¹¹⁸ control flies (P<0.01), but not so in Clk^Jrk mutants (P=0.39) by Student’s t-test (panel A). Although random fluctuations were observed in Clk^Jrk mutants, a 24 h oscillation was not detected in any of multiple experiments. One-way ANOVA detects a significant cycle in both w¹¹¹⁸ and pdfr^han5304 mutants (panel B), and two-way ANOVA indicates that the oscillation is altered in pdfr^han5304 (P<0.05). The Clk^Jrk and pdfr^han5304 mutations were confirmed by probing with CLK and PDFR antibodies using fly head extracts. (C) The cycling of STAT92E is disrupted when hop^Tuml is over-expressed in the adult nervous system. 500μM RU486 or ethanol (control) was administered in the food to induce elavGS from the time of entrainment. One-way ANOVA detects a cycle in the ethanol control (P<0.05), but not in RU486 treated flies (P=0.93). (D) STAT92E protein levels continue to cycle on DD2 in wild type flies but the cycling is dampened in miR-279 mutants. STAT92E levels at the peak (CT8) are significantly higher than at the trough (CT2, CT14 and CT20) in wild-type flies (P<0.01), but not so in ex117-1 flies (P=0.46) by Student’s t-test. The daily expression of STAT92E in ex117-1 flies is significantly different from that in w¹¹¹⁸ flies (P<0.05, by two-way ANOVA). See Also Figure S5.

Figure 6

JAK/STAT signaling and miR-279 are in a circadian output circuit. (A) The expression pattern of upd includes dorsal neurons (DNs) and lateral located neurons (LLNs), which show shifted-phase expression of PER. Brains of upd-GAL4: nGFP flies were dissected and stained with GFP (green) and PER (red) antibodies at indicated Zeitgeber times (ZT) on the 4^th day in LD. PER protein levels are high in the early night in a subset of DNs and LLNs, but high in the early morning in other clock neurons. (B) PDF-containing dorsal projections from the central clock are in the vicinity of upd-expressing neurons. A membrane-targeted mCD8GFP was expressed under the control of upd-GAL4. Brains were dissected and stained with GFP (green) and PDF (red) antibodies. An enlarged image is shown on the right of the panel. upd-expressing neurons are indicated with white arrows. (C) Dome expresses in the mushroom bodies (MBs), pars intercerebralis (PI), and dorsal giant interneurons (DGIs). A mCD8GFP was expressed under the control of dome-GAL4. Brains were dissected at ZT2 on the 4^th day in LD and stained with GFP (green), PER (red) and PDF (blue) antibodies. (D) Model depicting the role of miR-279 and JAK/STAT signaling in a circadian behavioral output. We propose that the central clock affects cyclic secretion of the UPD protein from cells that act downstream of PDF signaling. The mRNA levels of upd in these neurons are negatively modulated by miR-279. UPD may rhythmically activate the DOME receptor in dome-expressing cells, which would lead to daily oscillations of JAK/STAT activity and therefore of STAT92E levels (see Discussion). Rhythmic activity of this pathway is likely required for rest:activity rhythm. See Also Figure S6.