Figure 1.
Activity of Recombinant DBT Purified from Sf9 Cells
(A) Recombinant DBT purified from Sf9 cells has protein kinase activity. GST-DBT and GST-DBT (K38A) proteins purified from Sf9 cells were incubated with casein and radioactive P-32 labeled γ-ATP in kinase buffer. Casein is strongly phosphorylated by GST-DBT. Mutant GST-DBT (K38A) substantially reduces casein phosphorylation. The lower panel indicates the amount of GST-DBT and GST-DBT (K38A) used in the kinase assay as detected by western blot with anti-DBT antibody. Molecular weight markers are shown on the left.
(B) DBT mutations affecting behavioral rhythms also affect kinase activity. Either wild-type DBT (DBT), DBT arrhythmic (DBTAR), long- (DBTL) or short- (DBTS) period rhythm mutants were expressed and purified as GST fusion proteins from Sf9 cells, and used to phosphorylate casein. Top: GST-DBT levels measured by anti-GST western blot. Bottom: equal amounts of casein were phosphorylated with wild-type GST-DBT (lane 4) or with mutant GST-DBT proteins: GST-DBTAR (lane 1), GST-DBTL (lane 2), and GST-DBTS (lane 3).
(C) Kinase activity of GST-DBT fusion proteins. Relative kinase activity is shown as a ratio of the casein phosphorylation signal to total GST-DBT level. Relative kinase activity of wild-type GST-DBT was set as 1.
Figure 2.
DBT Preferentially Phosphorylates PER
Fragments or full-length sequences of proteins required for circadian clock function in Drosophila were fused to GST and expressed and purified from E. coli. Purified fusion proteins were treated with either wild-type DBT (odd-numbered lanes) or an inactive DBT mutant (K38A) as a negative control (even-numbered lanes). The top panel shows autoradiograph of incorporated phosphate; the lower panel shows total protein stained with Amido black. Molecular weight markers are indicated to the right of each panel. Sequences fused to GST as substrates for phosphorylation: lanes 1–2, TIM aa 507–1,150; lanes 3–4, TIM aa 222–577; lanes 5–6, CLK (-N terminus) aa 1–600; lanes 7–8, full-length CYC; lanes 9–10, full-length SGG; and lanes 11–12, PER amino acids 1–640. Molecular weight markers are shown on the right.
Figure 3.
DBT Phosphorylates Three Regions in PER
Different fragments of PER were fused to GST and expressed and purified from E. coli. Each fusion protein was phosphorylated with recombinant DBT, resolved in 10% SDS-PAGE, and transferred to nitrocellulose. Top panels in (A, (B), and (C) show signals from incorporated radioactive phosphate. Lower panels in (A), (B), and (C) show total proteins stained with Ponceau S. Numbers on top of each lane indicate the amino acids in PER included in each fusion protein. Molecular weight markers are shown on the right of each panel.
(A) shows mapping of phosphorylation in the middle of PER (between residues 570–645).
(B) shows mapping of phosphorylation near the N-terminus (between residues 149–170).
(C) shows mapping of phosphorylation near the C-terminus (between residues 1,101–1,224).
Table 1.
Phosphopeptides Generated from GST-PER Fusion Proteins by Reaction with DBT
Figure 4.
Distribution of Phosphorylated Regions in PER Relative to Functional Domains or Motifs
Colored blocks show functional domains found in PER. Three regions phosphorylated by DBT are shown as segments of PER protein sequence, with phosphorylated residues established in this study indicated in red. Four serines between aa 143–169, as well as serine 604, are shown in blue, as they are potential targets within DBT-derived phosphopeptides, but could not be directly assessed for phosphorylation in this study; see also the text). PAS indicates position of PAS domain (contains aa ∼220–450) [51]; CLD, cytoplasmic localization domain that promotes PER cytoplasmic localization in the absence of TIM (aa 452–512) [52]; S, per-short domain, causes short-period behavioral rhythms when variably mutated (indicated as a green bar; 585 and 601 indicate the first and last amino acid of the motif, respectively) [24]; SD, per-short downstream domain, contains many DBT phosphorylation sites (indicated by the red bar above the protein sequence; 604 and 629 indicate the first and last amino acids, respectively, of the region examined in this study); CCID, CLK-CYC inhibitory domain, represses CLK-dependent transactivation when overexpressed in cultured cells [39]. (N and C indicate the amino- and carboxy-terminal ends of the schematic PER).
Figure 5.
Repression of CLK-Dependent Activation of per-luciferase Reporter by PER Phosphorylation Site Mutants
Bar graphs indicate relative luciferase activity levels from S2 cells cotransfected with Clk and 50 ng of either wild-type (+) or mutant per expression plasmids along with the per-luc reporter. Normalized luciferase activities were plotted as relative activation compared to expression in the absence of per (–). Numbers below the graph indicate positions of serine or threonine residues in PER that were mutated to alanine (PER mutant N5A has amino acids 149, 151, 153, 164, and 169 mutated to alanine, and PERΔS has a 51-nt deletion in the per-short region as described in Materials and Methods).
(A) Repression activity of PER mutants. Sequence on top shows the DBT-mediated phosphorylation sites in the per-short and perSD domains (in red), and the multiple residues mutated to alanine (in blue) for mutants 4A and 6A. Values are averaged over three experiments; error bars indicate standard error of the mean. Lane 1, no added CLK. Remaining lanes, CLK plus indicated per PER repressor.
(B) Stability of the PER mutants. The assay was performed as described above with S2 cells transfected with equal amounts (50 ng) of mutant or wild-type per DNA. The panel at the top shows the amount of PER protein detected by western blot analysis at the time the luciferase assay was conducted. In this case, equivalent, small samples of the luciferase assays were subjected to SDS-PAGE, and the presence of PER protein was detected using anti-PER antibodies.
(C) Quantification of per-luc repression by PER mutants as a function of the amount of PER protein. Increasing amounts of per+ and per mutant DNA were used to transfect S2 cells. In three independent experiments, the level of repression was determined as described above, and the amount of PER present in each assay quantified by western blot. The levels of repression shown represent activity per unit protein as assessed by the comparison of the quantified western blots. This comparison indicates that PER S589A and PERΔS are both more active repressors than the wild-type PER protein.
Table 2.
Dosage-Dependent Shortening of the Period Length of Locomotor Activity Rhythms by perΔS Transgenes
Figure 6.
Quantification of Clock Gene mRNA Expression
mRNA expression of cycling genes making up the circadian clock in flies was measured by northern blot or RNase protection from two time-course experiments (one of which is shown in Figure S2). An average of the normalized values was obtained and plotted. Genes assayed are indicated to the right of their respective profiles. RNA profiles on the left are from 12 h/12 h, light/dark cycles (LD, indicated by alternating open and closed horizontal bars), right-side profiles are from DD time courses (constant darkness with subjective day indicated by hatched bar). Numbers on the vertical axes indicate relative expression (maximal expression in wild type is 1.0). Numbers on the horizontal axes indicate time (hour within a 24-h cycle) when RNA was collected. Blue, wild-type (wt) time courses; pink, time courses from flies homozygous for one perΔS transgene (per0;perΔSx1); and yellow, time courses from flies homozygous for two perΔS transgenes (per0;perΔSx2). For (K) and (L), the green line represents the endogenous per RNA present in the per0 background, to which the levels were normalized. Notice the change in the scale due to the increased expression of perΔS RNA. Error bars show square root of variance from two independent samplings.
Figure 7.
PERΔS Protein Expression in Flies with Two to Four perΔS Transgenes
Western blot analyses of PER from fly heads collected at the indicated times in LD (indicated by altering open and closed horizontal bars). Newly eclosed flies were entrained in LD for at least 3 d prior to collection. Genotypes assayed are indicated at the bottom of each panel (wt, wild type; [perΔS]x2 indicates presence of two homozygous perΔS transgenes (four copies total). PERΔS protein signals are visible in the upper panels. Only PERΔS proteins were detected in transgenics due to the per null (per0) background. DBT expression (lower panel) was used as a loading control.
Figure 8.
Activity of the per-Short and perSD Domains Regulated by DBT Phosphorylation
A schematic of Drosophila PER protein is shown with the per-short (per-S) and per-short downstream (per-SD) domains highlighted. Residues phosphorylated by DBT in vitro are indicated (P). Three phosphorylation states are indicated, as well as effects on stability and activity as suggested by our studies of cultured cells and transgenic flies. It is proposed that phosphorylation of the per-short domain inhibits DBT-directed phosphorylation of the per-short downstream domain (see text).