Figure 1.
Identification of aa 100-200 as the transcriptional activation domain for WC-1.
(A) Schematic depiction of the domain architecture of the WC-1 protein. WC-1 contains 1,167 amino acids and black bars with corresponding numbers of amino acids represent different deletions of the N-terminal region of WC-1. AUG(long) and AUG(short) mark the beginnings of long and short forms of WC-1; PolyQ, putative transcription activation domain; LOV, light, oxygen, or voltage domain containing FAD chromophore binding site; PAS, Per-Arnt-Sim domain mediating interaction with WC-2; NLS, putative nuclear localization signal; Zn, zinc finger DNA-binding domain (B) Racetube analyses of wild-type (WT) and WC-1 deletion strains. Period is reported in hours ± one standard deviation, n = number of racetubes. Replicate tubes are shown, and the vertical black lines in racetubes mark daily growth fronts of the strains. 328-4 (ras-1bd A) served as the WT. Strains bearing deletions of both polyQs (Δ16-57 and 1097-1128), Δ1-100, and Δ201-300 showed WT circadian periods. Strains lacking aa 100-200 (Δ1-200, Δ1-300, and Δ101-200) are arrhythmic (C) Results from quantitative RT-PCR analysis of frq and wc-1 mRNA levels in WT and in the wc-1 deletion strains noted. (D) Western blot showing expression levels of WC-1 and FRQ in WT and wc-1 mutants. LL, constant light; DD, hours after the light to dark transition; CT, circadian time. Both WT WC-1 and mutants were C-terminally tagged with V5. Δ101-200 showed a significant decrease of frq expression and further deletion of aa 1-200 or 1-300 further dimished frq expression. Elimination of the two polyQs (Δ16-57 and 1097-1128; four progeny of identical genetype from one cross are shown) had no impact on FRQ level at CT5.
Figure 2.
WCC interacts with SWI/SNF in vivo and in vitro.
(A) Co-IP demonstrating interaction of WC-1 with SWI1 in vivo. SWI1 was C-terminally tagged with a V5 epitope and immunoprecipation was performed using WC-2 antibody. WC-1 was pulled down with WC-2 as well as a reduced amount of SWI1. (B) Expression and purification of GST-WC-1 fusion proteins. GST, GST WC-1 1-300 6xHis, and GST WC-1 100-300 6xHis lacking the N-polyQs were expressed in bacteria and purified. Gel stained with Coomassie Blue. (C) N-terminal fragments of WC-1 extending from 1-300 or 100-300 bind to SWI1 in vitro. Neurospora crude cell lysates were mixed with beads to which the GST-tagged WC-1 fragments were bound, and bound SWI1 was visualized by virtue of a C-terminal V5 tag; see Materials and Methods. GST alone failed to pull down SWI1 while WC-1 aa1-300 or aa 100-300 pulled down SWI1 at a similar level. Likewise negative control proteins actin, and two transcription factors encoded by NCU05051 and NCU07728 were not bound by WC-1 fragments. (D) Affinity purification of the Neursopora SWI/SNF complex showed the presence of different subunits in a 1∶1 stoichiometry except for SWI3. SWI1 was tagged with V5 at the C-terminus, centrifuged lyate was incubated with V5 antibody, and the gel was sliver-stained. Individual bands were excised and each protein identified via mass spectrometry.
Figure 3.
Loss of FRQ expression in SWI/SNF subunit knockouts.
(A) Expression levels of FRQ and WC-1 were followed by Western blotting in WT, Δswp59, Δswi1, and Δsnf5. Two dark time points were chosen to examine FRQ expression, CT5 (DD16) when newly synthesized FRQ is seen and CT17 (DD28) when old FRQ is hyperphosphorylated and begins to be degraded in WT. Non-specific bands were shown for equal loading. (B) Corresponding data for frq and wc-1 mRNA is shown.
Table 1.
Neurospora SWI/SNF subunits and knockouts.
Figure 4.
Reduced frq expression and loss of rhythmicity in SWI/SNF subunit knockouts as assayed by a luciferase reporter.
frq transcription in WT, Δswi1, Δsnf5, and Δswp59 was examined using the frq C box fused to codon-optimized firefly luciferase (transcriptional fusion). Strains were grown on 0.1% glucose racetube medium containing 5 mM luciferin in a 96 well plate and synchronized by growth in constant light for 48 hours followed by transfer to darkness. The luciferase signal was followed for longer than 6 days with sampling every 30 minutes. Each strain was repeated three times. Δswp59 showed a WT oscillation, Δsnf5 also oscillated in a circadian manner despite an extremely low amplitube, while frq trancription was completely abolished in Δswi1.
Figure 5.
The binding of SWI/SNF to the C box relies on aa 1-300 of WC-1 and increases prior to the peak of rq expression.
(A) ChIP experiment performed on chromatin isolated at DD16 when frq expression is maximal. WC-1 and WC-2 had a normal binding to the C box of frq in a strain bearing WC-1Δ1-300, while SWI2 binding was impaired in this strain. The annealing positions of the primer set used to detect the C box corresponds to the middle set shown in Figure 6A. Average values are plotted as a percent of total with error bars representing the standard error of the mean (SEM) (n = 3, ***p<0.0005). Samples were grown for 16 hours in the dark, formaldehyde-crosslinked, and harvested. (B) ChIP was used in a timecourse analysis of the association of WC-2, SWI-1 and histone H3 with the C box. Samples were gorwn for the indicated number of hours in darkness (DD) prior to harvesting and processing for ChIP as described in (A). Error bars represent the standard error of the mean ( = 3).
Figure 6.
SWI1 is required for rhythmic opening of the C box through removal of NucB.
(A) Relative positions of resident nucleosomes, NucB and NucA at the C box (see also reference 31). Arrows indicate primer sets used to detect these regions. (B) Nuclei were isolated by fractionation from samples harvested at indicated hours in the dark and digested with MNase. 20 nanograms of gel-purified mononucleosomal DNA were loaded in each lane. (C) The density of NucB was measured by real-time PCR using a primer set shown in (A, the left primer set) and gel purified mononucleosomal DNA as template (n = 3, error bars represent ± SEM, ***p<0.0005, **p<0.005). NucB occupancy of the C box oscillates between low (CT0 and 21) and high (CT 13-16) in WT as reported by Belden et al. (2007b), whereas NucB is always bound to the C box in Δswi1. NucA and 3.303 (an unregulated control region of genomic DNA distinct from frq) served as controls for equal DNA inputs [31]. (D) Difference in frq promoter chromatin were probed with a limited nuclease digestion assay as described in Experimental Procedures. Cultures were harvested from WT or from Δswi1 cultures at the times in darkness indicated.
Figure 7.
A working model for WC-1-dependent recruitment of SWI/SNF to initiate frq transcription.
In the night before frq transcription starts, NucB is in a repressive state occluding the C box as fostered by chromatin remodelers that may include CHD-1. When active WC-1/WC-2 begin to bind at the C box, an event roughly coincident with CSW binding, SWI/SNF is recruited and it helps to fully remodel the C box to an active state, removing NucB, and also beginning to bring about the DNA looping (curved gray line) that is essential to bring the WCC transcriptional activation domains to the transcription start site (TSS) for stable recruitment of general transcription factors (GTFs). For acute light-induction of frq, an action not requiring SWI/SNF, WCC binds directly to the PLRE rather than the C box and recruits unknown remodelers to drive frq transcription at levels much higher than in the dark.