Antagonistic Gcn5-Hda1 interactions revealed by mutations to the Anaphase Promoting Complex in yeast

Abstract

Background: Histone post-translational modifications are critical for gene expression and cell viability. A broad spectrum of histone lysine residues have been identified in yeast that are targeted by a variety of modifying enzymes. However, the regulation and interaction of these enzymes remains relatively uncharacterized. Previously we demonstrated that deletion of either the histone acetyltransferase (HAT) GCN5 or the histone deacetylase (HDAC) HDA1 exacerbated the temperature sensitive (ts) mutant phenotype of the Anaphase Promoting Complex (APC) apc5CA allele. Here, the apc5CA mutant background is used to study a previously uncharacterized functional antagonistic genetic interaction between Gcn5 and Hda1 that is not detected in APC5 cells.

Results: Using Northerns, Westerns, reverse transcriptase PCR (rtPCR), chromatin immunoprecipitation (ChIP), and mutant phenotype suppression analysis, we observed that Hda1 and Gcn5 appear to compete for recruitment to promoters. We observed that the presence of Hda1 can partially occlude the binding of Gcn5 to the same promoter. Occlusion of Gcn5 recruitment to these promoters involved Hda1 and Tup1. Using sequential ChIP we show that Hda1 and Tup1 likely form complexes at these promoters, and that complex formation can be increased by deleting GCN5.

Conclusions: Our data suggests large Gcn5 and Hda1 containing complexes may compete for space on promoters that utilize the Ssn6/Tup1 repressor complex. We predict that in apc5CA cells the accumulation of an APC target may compensate for the loss of both GCN5 and HDA1.

Figure 1

Mutation to the APC subunit Apc5 reveals antagonistic interactions between Gcn5 and Hda1. (A) Serial 10-fold dilutions of each strain were spotted onto YPD plates from left to right and incubated at the temperatures shown. (B) Serial dilutions using strains expressing the indicated plasmids were spotted onto SD-ura plates containing either 2% glucose or 2% galactose, and grown at 30°C for 2 and 3 days, respectively. (C) Protein lysates were prepared from the mutants shown and characterized by Westerns using the antibodies indicated. Antibodies against GAPDH were used as load controls.

Figure 2

Expression of PDS1, an APC antagonist, is specifically elevated in apc5^CA hda1Δ cells at 37°C. (A) Northern analyses were conducted on total RNAs extracted from the various mutants and probed with sequences derived from the coding regions of the genes indicated. RDN37-2 (RDN1) was used as a loading control. All bands from the 30°C experiments (B) and the 37°C experiments (C) were quantified by ImageJ and normalized to the RDN1 signal. Densitometry was performed on two Northern experiments and two reverse transcriptase experiments. The data was combined and the means and standard errors were plotted.

Figure 3

Histone H3 acetylation at promoter regions is elevated specifically in apc5^CA hda1Δ cells at 37°C. (A) ChIP was performed using lysates derived from the mutants shown and antibodies against total H3 or H3 acetylated at both K9 and K14. A mock treatment lacking antibody was used as a control. Once the crosslinks were reversed and DNA recovered, "end point" PCR was performed using primers against the genes shown that amplified 200 bp regions of the promoter. 10% of the reaction was used as input. (B) Two independent experiments were quantified, with the means and standard errors represented graphically, as previously described [66,70].

Figure 4

Deletion of HDA1 results in increased Gcn5 at promoters. (A) Steady-state Gcn5-TAP in different mutant backgrounds was determined in early log phase asynchronous cells grown at 30°C by Western blotting. Westerns were performed using antibody against TAP and the membrane was stained with Ponceau S to confirm equal protein load. (B) Plasmid borne HA-tagged GCN5 and ELP3, driven by the galactose inducible promoter, were expressed in cells lacking the proteasome ubiquitin receptor Rpn10. Cells were grown overnight in 2% glucose to early log phase. The glucose-supplemented media was washed away and the cells were resuspended in 2% galactose-supplemented media. The cells were then split with one half incubated at 37°C and the other half left at 30°C. The cells were incubated for an additional 4 hours, afterwhich proteins were harvested and analyzed with antibodies against HA or GAPDH as a load control. Controls for the experiment included endogenous APC5-TAP in rpn10Δ cells, as well as the detection of endogenous Clb2 and Ub in WT and rpn10Δ cells using commercially available antibodies. (C) Protein/DNA complexes were recovered from the mutants shown following GAL-induction using antibodies against either the HA epitope, total H3, or H3K9/14^Ac. A mock treatment was conducted where antibody was omitted. Recovered DNA was used as template in "end point" PCR reactions using primers that amplified the promoter regions indicated. 10 μl of each reaction was separated by agarose gel electrophoresis and scanned. (D) Two independent experiments were performed. The gels were scanned and quantified using ImageJ. The means and standard errors were plotted.

Figure 5

Gcn5 promoter occupancy is kept at an equilibrium in WT cells, but increases over time in hda1 cells. (A) Western analyses of Gcn5-HA expression in gcn5Δ and gcn5Δ hda1Δ cells following a 5 hour 4% galactose-induction. Antibodies against GAPDH were used as a load control. (B) A galactose-induction time-course was performed in gcn5Δ and gcn5Δ hda1Δ cells expressing GAL_pro-GCN5-HA. Protein samples were removed at the times shown for Western analyses with antibodies against HA and GAPDH. (C) From the time-course described above, samples were also removed for ChIP. Recovered DNA was used as a template in "end point" PCR reactions. S, sample with antibody; C, control without antibody; I, 10% lysate input. (D) The gel in (C) was scanned, analyzed using ImageJ and plotted.

Figure 6

Tup1 occludes Gcn5 recruitment. A) ChIP was performed using the cells shown expressing GAL_pro-GCN5-HA following a 5 hour galactose induction, as described above. (B) Two independent experiments were scanned and processed using ImageJ, with the means and standard errors shown. (C) Strains lacking TUP1 were constructed in WT and apc5^CA backgrounds. Growth phenotypes were assessed by spot-dilutions, followed by incubation at 34°C and 37°C.

Figure 7

Gcn5 can inhibit Hda1-Tup1 associations at some promoters. (A) Sequential ChIP was used to observe Hda1-Tup1 physical interactions at specific promoters. WT, hda1Δ and gcn5Δ hda1Δ cells expressing combinations of GAL_pro-HDA1-HA and CUP1_pro-TUP1-GST were induced using 4% galactose for 5 hours and 0.4 mM CuSO₄ for 3 hours. ChIP reactions were first performed with antibodies against HA. Bound proteins were eluted from beads using 10 mM DTT for 30 minutes at 37°C. The eluted proteins were then incubated with anti-GST antibodies. The immune complexes were isolated again, cross links were reversed, and "end point" PCR was performed using the recovered DNA as template. (B) Two independent experiments were performed and processed using ImageJ. The means and standard errors are shown. (C) Westerns showing expression of the proteins used in the sequential ChIP experiment.

Figure 8

A model depicting potential interactions between Gcn5 and Hda1. (A) The HAT Gcn5 and the HDAC Hda1 have opposing functions that individually benefit APC function. (B) and (C) Gcn5 and Hda1 appear to compete for Tup1 binding. (B) If Hda1 first gains access to the promoter, recruitment of Gcn5 is partially blocked. (C) Under conditions where gene transcription must be derepressed, Tup1 may recruit Gcn5 to the promoter to prime transcriptional initiation, thus displacing Hda1. The protein labelled × represents a DNA binding factor that recruits the Tup1/Ssn6 corepressor complex to silent genes.