The MAP kinase p38 is part of Drosophila melanogaster's circadian clock

Abstract

All organisms have to adapt to acute as well as to regularly occurring changes in the environment. To deal with these major challenges organisms evolved two fundamental mechanisms: the p38 mitogen-activated protein kinase (MAPK) pathway, a major stress pathway for signaling stressful events, and circadian clocks to prepare for the daily environmental changes. Both systems respond sensitively to light. Recent studies in vertebrates and fungi indicate that p38 is involved in light-signaling to the circadian clock providing an interesting link between stress-induced and regularly rhythmic adaptations of animals to the environment, but the molecular and cellular mechanisms remained largely unknown. Here, we demonstrate by immunocytochemical means that p38 is expressed in Drosophila melanogaster's clock neurons and that it is activated in a clock-dependent manner. Surprisingly, we found that p38 is most active under darkness and, besides its circadian activation, additionally gets inactivated by light. Moreover, locomotor activity recordings revealed that p38 is essential for a wild-type timing of evening activity and for maintaining ∼ 24 h behavioral rhythms under constant darkness: flies with reduced p38 activity in clock neurons, delayed evening activity and lengthened the period of their free-running rhythms. Furthermore, nuclear translocation of the clock protein Period was significantly delayed on the expression of a dominant-negative form of p38b in Drosophila's most important clock neurons. Western Blots revealed that p38 affects the phosphorylation degree of Period, what is likely the reason for its effects on nuclear entry of Period. In vitro kinase assays confirmed our Western Blot results and point to p38 as a potential "clock kinase" phosphorylating Period. Taken together, our findings indicate that the p38 MAP Kinase is an integral component of the core circadian clock of Drosophila in addition to playing a role in stress-input pathways.

Figure 1. p38 MAPK expression pattern in adult male Canton S brains.

p38 MAPK distribution within the circadian clock was investigated immunohistochemically with an antibody directed against Drosophila p38b (A–C) and against phosphorylated human p38 (D–G). A–C: Staining with anti-p38b (green) in Canton S wildtype brains was visible in many cell bodies close to the lateral clock neurons, but co-labeling with anti-VRI (magenta) and anti-PDF (blue) revealed clear p38b expression in the l-LN_vs (white stars in B3) and the s-LN_vs (white stars in C3). B1–B4 represents a close-up of l-LN_vs, C1–C4 a shows close-up of s-LN_vs. Furthermore, we found staining in the entire cortex including the region of the dorsal neurons (see Fig. S1). D–G: Staining with anti-p-p38 (green) was restricted to fewer neurons, but revealed again staining in the entire cortex that was stronger at night (F, ZT21) than during the day (D, ZT9). Double-labeling with anti-VRI (magenta) and anti-p-p38 antibody (green) revealed active p38 only in 2 clock neurons, the DN_1as (white stars in G2). Also in these cells, p-p38 staining intensity depended on the time of day, showing a higher level of active p38 at ZT 21 (G2, white stars) than at ZT9 (E2). E1–E3 and G1–G3 represent a close-up of DN_1as. Scale bar = 10 µm.

Figure 2. Daily p38 mRNA (A) and protein expression (B–D) in Canton S wildtype.

A: Quantitative real-time PCR on head extracts revealed constant mRNA expression throughout the day with allover higher levels of p38b compared to p38a (p<0.001). B: Antibody staining with anti-p-p38 on adult brains displayed rhythmic phosphorylation of p38 in DN_1as in LD with significant higher p-p38 levels occurring during the night than in the day (p<0.05). C: A highly significant reduction of active p38 in DN_1as at CT6 compared to CT18 in DD indicates a clock-controlled activation of p38 (p<0.001) D: Only a 15 minute light pulse (LP) during subjective night (CT18) and not during the subjective day (CT6) leads to a reduction in active p38 in DN_1as, suggesting a clock-dependent photic reduction of active p38. The “C” in D indicates control brains without 15 minute light pulse (LP). Error bars show SEM. Significant differences (p<0.05) are indicated by *, highly significant differences (p<0.001) by **.

Figure 3. Locomotor activity rhythms of p38b knockdown flies with respective controls.

Flies were recorded in LD 12∶12 for 6 days and subsequently in DD for at least 14 days. A daily average activity profile for day 2–7 in LD was calculated for each genotype and is shown above the double-plot of a representative actogram (left panels in A and B). In addition, for each genotype the onset of evening activity in LD (upper right panel in A and B) as well as the average free-running period in DD (lower right panel in A and B) was determined and is depicted as boxplot in the right panel. While average activity profiles of dicer2;UAS-p38b^RNAi/tim(UAS)-Gal4;+ (A) displayed wildtype-like behavior in LD with activity bouts around lights-on and lights-off, evening activity onset of dicer2;UAS-p38b^RNAi/Pdf-Gal4;+ flies (B) was significantly delayed compared to respective controls. When transferred to DD, dicer2;UAS-p38b^RNAi/tim(UAS)-Gal4;+ mainly became arrhythmic (see Table 1 and lower left panel in A), just 7% remained rhythmic displaying a significant longer free-running period than UAS- and Gal4-controls (lower right panel in A). p38b knockdown in PDF-expressing neurons also significantly lengthened the free-running period in DD as compared to respective controls (lower panels in B), and 58% of the flies remained rhythmic (see Table 1). Bars above the daily average activity profiles and actograms depict the light regime of the LD 12∶12 cycle and black arrowhead indicate the shift to constant DD. Black lines in daily average activity profiles represent mean relative activity, gray lines SEM and dotted grey lines the calculated evening activity onset. Gray lines in boxplots illustrate the median, boxes 25–75%, and whiskers 10–90% of the data. UAS refers to respective UAS-control, GAL4 to respective Gal4-control and UAS>GAL4 to the experimental line. Significant differences (p<0.05) are indicated by *, highly significant differences (p<0.001) by **. Numbers in brackets indicate n.

Figure 4. Locomotor activity rhythms of flies expressing a dominant-negative form of p38b (p38b^DN-S) in Drosophila clock neurons and respective controls.

Both, expression of a dominant negative form of p38b in either all clock neurons (UAS-p38b^DN-S;tim(UAS)-Gal4/+;+) or just in a subset of clock cells, the PDF-positive LN_vs (UAS-p38b^DN-S;Pdf-Gal4/+;+), resulted in a diurnal activity profile with a significantly delayed evening activity onset in comparison with respective controls (upper panels in A and B). This delay in evening activity is accompanied by a significantly prolonged free-running period in UAS-p38b^DN-S;tim(UAS)-Gal4/+;+ (lower panels in A) as well as in UAS-p38b^DN-S;Pdf-Gal4/+;+ flies(lower panels in B), when released into constant darkness. For recording and processing of activity data as well as for figure labeling see Figure 3.

Figure 5. Locomotor activity rhythms of p38a knockdown flies and respective controls.

Average activity profiles of dicer2;UAS-p38b^RNAi/tim(UAS)-Gal4;+ (upper panel in A) and dicer2;UAS-p38b^RNAi/Pdf-Gal4;+ (upper panel in B) displayed wildtype-like behavior in LD with activity bouts around lights-on and lights-off and did not differ from those of control flies. Evening activity onset was not delayed in the two mutant strains. However, when released into constant darkness both, p38a knockdown in TIM- (lower panels in A) and PDF-expressing neurons (lower panels in B) resulted in significant prolonged free-running rhythms in comparison to respective controls. For recording and processing of activity data as well as for figure labeling see Figure 3.

Figure 6. Locomotor activity rhythms of p38b and p38a null mutants and hypomorphic double mutant flies.

Both p38 null mutants, p38b^Δ45 (upper panels in A) and p38a^Δ1 (upper panels in B), displayed wildtype-like behavior with activity bouts around lights-on and lights-off when recorded in LD 12∶12. Even if evening activity onset of p38a^Δ1seems to be delayed compared to w¹¹¹⁸, this delay did not result in a longer free-running period under constant darkness (lower panels in B). Similarly, flies, lacking the p38b gene, also showed comparable free-running rhythms as their respective controls (lower panels in A). Activity data in C show two representative single actograms of a double mutant strain with a hypomorphic p38b allele (p38b^Δ25;p38a^Δ1). Since these flies are hardly viable and die within 3–6 days after emergence of the pupa, flies were already entrained to LD12∶12 during pupal stage and subsequently monitored in DD conditions after eclosion. Even if periodogram analysis was not possible due to the short recording period, p38b^Δ25;p38a^Δ1 flies clearly showed a long free-running period when kept in constant darkness (C). For recording and processing of activity data as well as for figure labeling see Figure 3.

Figure 7. Daily oscillations of nuclear PER in s-LN_vs and l-LN_vs of flies expressing a dominant negative form of p38b in these cells.

Flies were entrained in LD 12∶12, dissected every one to two hours and staining intensity of nucleus and cell body was measured as described in Material and Methods. Nuclear PER staining intensities were normalized to total staining and tested for statistically significance. Expression of the dominant negative form of p38b phase delayed nuclear accumulation of PER in the s-LN_vs (A) and l-LN_vs (B). Arrows indicate the maxima of nuclear PER staining that occurred significantly later in UAS-p38b^DN-S;Pdf-Gal4/+;+ flies than in control flies. This delay in nuclear PER accumulation in PDF-positive clock neurons is well consistent with the shifted evening activity in these flies. Bars above the graphs depict the light regime of the LD 12∶12.

Figure 8. p38b promotes PER phosphorylation during the dark phase.

To analyze daily phosphorylation of PER in flies that express the dominant-negative form of p38b in clock neurons and photoreceptor cells, we performed Western blots on head extracts after 4 days entrainment to LD 12∶12 cycles. According to our behavioral data, timing of PER accumulation was not affected in experimental flies (UAS-p38b^DN-S;cry-Gal4/+;+) in comparison with their respective control (A). However, regarding the degree of PER phosphorylation we observed differences at all time points when we compared both genotypes. For better comparison Western blots were repeated and samples of control and UAS-p38b^DN-S;cry-Gal4/+;+ flies were plotted side by side for each ZT (B). Interestingly, flies with impaired p38 signaling indeed had less phosphorylated PER, showing the largest differences to the controls at the end of the night. Western blots were repeated 4 times and always gave similar results. Bars above the blots depict the light regime of the LD 12∶12. The “C” refers to respective control, DNS to UAS-p38b^DN-S;cry-Gal4/+;+.

Figure 9. p38b phosphorylates PER in vitro.

To test whether p38b phosphorylates PER in vitro, either non-radioactive kinase assays followed by urea-PAGE (A–D) or radioactive kinase assays with autoradiography (E–F) were performed. A–D: Non-radioactive kinase assays were conducted with poly-histidine tagged p38b (His₆-p38b) and two truncated GST-tagged PER isoforms, GST-PER^1–700 (A,B) and GST-PER^658–1218 (C,D). Samples were subsequently separated with urea-PAGE and visualized by Coomassie staining (A,C). To further confirm PER's position in the gel two samples of the same gel were additionally blotted onto nitrocellulose membrane and immunolabeled using an anti-PER antibody and a secondary fluorescent antibody (B,D). While GST-PER^1–700 without kinase did not shift within 60 minutes, the addition of His₆-p38b induced a downward shift of GST-PER^1–700 indicating phosphorylation of PER (A; dotted line). Immunoblots with anti-PER further confirmed the size as well as the shift of the GST-PER^1–700 band (B). In addition to GST-PER^1–700, GST-PER^658–1218 also displayed band shifts after incubation with His₆-p38b (C). This was most prominent after 60 minutes, when addition of His₆-p38b resulted in two distinct shifted bands (black arrowheads), which could be additionally confirmed by Western blots (D). Time scale below graphs represents minutes after addition of His₆-p38b, the “C” refers to control and represents substrate samples without kinase and ATP. (E) Radioactive in vitro kinase assays were conducted with the indicated GST-PER fusion proteins and GST-p38b. Control reactions were performed in the absence of GST-p38b or with GST in combination with GST-p38b. Coomassie staining proved loading of the indicated protein combinations. Below, phosphorylation of GST-PER proteins was detected by autoradiography. (F) For quantitative analysis five independent in vitro kinase assay experiments were performed and analyzed. For each reaction within a single experiment, autoradiography signal intensities were normalized to the corresponding Coomassie stained protein band. Values in the graph are shown as percentages of GST-PER^658–1218 phosphorylation (100%; * p<0.05, ** p<0.005).