Translational regulation of the DOUBLETIME/CKIδ/ε kinase by LARK contributes to circadian period modulation

Abstract

The Drosophila homolog of Casein Kinase I δ/ε, DOUBLETIME (DBT), is required for Wnt, Hedgehog, Fat and Hippo signaling as well as circadian clock function. Extensive studies have established a critical role of DBT in circadian period determination. However, how DBT expression is regulated remains largely unexplored. In this study, we show that translation of dbt transcripts are directly regulated by a rhythmic RNA-binding protein (RBP) called LARK (known as RBM4 in mammals). LARK promotes translation of specific alternative dbt transcripts in clock cells, in particular the dbt-RC transcript. Translation of dbt-RC exhibits circadian changes under free-running conditions, indicative of clock regulation. Translation of a newly identified transcript, dbt-RE, is induced by light in a LARK-dependent manner and oscillates under light/dark conditions. Altered LARK abundance affects circadian period length, and this phenotype can be modified by different dbt alleles. Increased LARK delays nuclear degradation of the PERIOD (PER) clock protein at the beginning of subjective day, consistent with the known role of DBT in PER dynamics. Taken together, these data support the idea that LARK influences circadian period and perhaps responses of the clock to light via the regulated translation of DBT. Our study is the first to investigate translational control of the DBT kinase, revealing its regulation by LARK and a novel role of this RBP in Drosophila circadian period modulation.

Figure 1. LARK associates with dbt transcripts.

A, dbt transcripts specifically co-immunoprecipitate with LARK from head tissue lysates. Relative amounts of isolated RNAs using anti-LARK (specific) or anti-EGFP (unspecific) antibodies are shown for the four alternative transcripts of dbt. For each transcript, the mean value of at least 6 Q-RT-PCR experiments is shown. Error bars represent SEM. *** p<10⁻⁶ (Student's t-test). B and C, Purified recombinant LARK protein binds to in vitro-transcribed dbt RNA in UV cross-linking assays. RNA transcribed from GluR2, an unrelated gene, was used as a negative control. Two independent cross-linking assays, utilizing different LARK concentrations, are shown.

Figure 2. LARK regulates the translation of DBT transcripts.

A, Western Blot showing the effect of pan-neuronal LARK knock-down (KD) and overexpression (OE) on DBT abundance assayed at three different zeitgeber times (ZTs). B, Overexpression of DBT alone renders detection of the shorter isoform at ZT2 but not ZT14. A non-specific band was used as loading control in A and B. C, Analyses of dbt mRNA translation using the TRAP method. All samples were collected at ZT2. Left, effect of altered LARK expression in all neurons. Right, effect of altered LARK expression in clock cells. Note that values for KD and OE are respectively normalized to KC and OC; thus the values of both controls were designated as “1” and plotted as one control denoted “C”. Fold changes were calculated from Ct values obtained from Q-RT-PCR that had been normalized to an internal Rp49 control. Average fold change from at least 6 Q-RT-PCR experiments are shown. Error bar represents SEM. *** p<0.001, ** p<0.01, * P<0.05 based on Student's t test comparing the Rp49-normalized Ct values of KD versus KC and OE versus OC.

Figure 3. Translation of dbt transcripts in LD 12∶12 or constant dark (DD) conditions.

Ribosome-bound RNAs were captured by TRAP at indicated Zeitgeber or circadian times and quantified by real-time Q-RT-PCR. Left panel, samples were collected over the course of one day. Right panel, samples were collected during the first and second day of DD. Mean and SEM (error bar) values for at least 6 Q-RT-PCR experiments are shown. One way ANOVA shows a time-dependent change of RC translation in both DD1 (p = 0.036) and DD2 (p = 0.00028), as well as RE translation in LD (p = 0.026).

Figure 4. Altered LARK expression affects light-induced translation of dbt-RE.

A. Light-induced translation of dbt-RE in wild-type flies. Relative translational levels were analyzed by quantifying ribosome-associated transcripts using TRAP and Q-RT-PCR. n≥5 for all data points. Error bars represent SEM. p = 2.91×10⁻⁵ analyzing the effect of light exposure by a two-way ANOVA of light condition and time. B. Altered LARK expression affects light-induced translation of dbt-RE. Light-induced translation of dbt-RE in flies with different LARK levels (KD, Control and OE) were analyzed by TRAP and Q-RT-PCR immediately after light exposure (0 hour) and 4 hours after light exposure. Amounts of ribosome-associated dbt-RE in flies exposed to light were normalized to those in flies kept in darkness (no light). n = 6 for all groups, representing 3 biological replicates, each with 2 technical replicates. Error bars show the possible range of fold change calculated based of the SEM of the QPCR data. * p<0.031 (Student's t test).

Figure 5. Altered LARK expression in PDF neurons affect circadian period.

A, Average period length for various genotypes (n = 192, 144, 64, 130 and 166 for OE, OC, OE^RRM, KD, and KC, respectively). Error bars represent SEM. *** p<10⁻⁵⁶ between OE and OC; P<10⁻⁵⁷ between KD and KC (Student's t-test). B–D, Representative activity plots for various genotypes. E, Immunohistochemistry showing LARK in the PDF neurons of various genotypes. KD, knockdown of LARK in PDF neurons, genotype: w¹¹¹⁸; pdf-gal4 uas-dicer2/+; lark¹ uas-lark^RNAi/+. KC, control line for the knock-down, genotype: w¹¹¹⁸; pdf-gal4 uas-dicer2/+. OE, moderate overexpression of LARK in PDF neurons, genotype: w¹¹¹⁸; pdf-gal4/+; Tub-gal80^ts uas-lark/+. OC, control for the overexpression, genotype: w¹¹¹⁸; pdf-gal4/+; Tub-gal80^ts/+. OE^RRM, Overexpression of a mutant form of LARK with defective RRM domains, genotype: w¹¹¹⁸; pdf-gal4/uas-lark^RRM.

Figure 6. Interactions between lark and dbt modulate circadian period.

A and B, Quantification of average period lengths showing effects of LARK OE or KD in flies heterozygous for various dbt mutations. Numbers shown at the base of the bar chart represent samples sizes for each genotype. Error bars represent SEM. p<0.0001 for all comparisons. C and D, Representative activity plots. For interactions involving LARK OE, genotypes are: without LARK OE, w¹¹¹⁸; pdf-gal4/+; Tub-gal80^ts/dbt. With LARK OE, w¹¹¹⁸; pdf-gal4/+; Tub-gal80^ts uas-lark/dbt. For interactions involving LARK KD, genotypes are: without LARK KD, w¹¹¹⁸; pdf-gal4 uas-dicer2/+; +/dbt, with LARK KD, w¹¹¹⁸; pdf-gal4 uas-dicer2/+; lark¹ uas-lark^RNAi/dbt (dbt here refers to dbt^S, dbt^L, dbt^P, or dbt^AR).

Figure 7. DBT kinase activity is required for the LARK OE phenotype.

A: Representative actograms showing that overexpression of wild-type DBT protein (pdf>dbt) enhances the LARK OE phenotype (producing arrhythmicity) whereas overexpression of a mutant DBT protein lacking kinase activity (pdf>dbt^D132N) suppresses the period-lengthening effect of LARK OE. B: Quantification of percentage rhythmicity in flies overexpressing DBT proteins with or without LARK OE. Genotypes are pdf>dbt alone: pdf-gal4/+; uas-dbt/+ (n = 31). pdf>dbt with LARK OE: pdf-gal4/+; uas-dbt/Tub-gal80^ts uas-lark (n = 31). pdf>dbt^D132N alone: pdf-gal4/+; uas-dbt^D132N/+ (n = 42). pdf>dbt^D132N with LARK OE: pdf-gal4/+; uas-dbt^D132N/Tub-gal80^ts uas-lark (n = 14). *** p<0.0001 by Chi-square test.

Figure 8. Increased LARK expression delays degradation of the PER protein.

The average pixel intensity in confocal images of anti-PER immunoreactivity in large PDF neurons (l-LNvs) are quantified for control and LARK overexpression flies. Samples were taken every 0.5 hours starting at ZT1 till ZT4.5. For each data point, brain hemispheres from 4–6 different animals were analyzed. Error bar represent SEM. A two-way ANOVA of genotype and time find a significant difference between control and LARK OE (p = 0.006289).