Histone deacetylase complexes promote trinucleotide repeat expansions

Abstract

Expansions of DNA trinucleotide repeats cause at least 17 inherited neurodegenerative diseases, such as Huntington's disease. Expansions can occur at frequencies approaching 100% in affected families and in transgenic mice, suggesting that specific cellular proteins actively promote (favor) expansions. The inference is that expansions arise due to the presence of these promoting proteins, not their absence, and that interfering with these proteins can suppress expansions. The goal of this study was to identify novel factors that promote expansions. We discovered that specific histone deacetylase complexes (HDACs) promote CTG•CAG repeat expansions in budding yeast and human cells. Mutation or inhibition of yeast Rpd3L or Hda1 suppressed up to 90% of expansions. In cultured human astrocytes, expansions were suppressed by 75% upon inhibition or knockdown of HDAC3, whereas siRNA against the histone acetyltransferases CBP/p300 stimulated expansions. Genetic and molecular analysis both indicated that HDACs act at a distance from the triplet repeat to promote expansions. Expansion assays with nuclease mutants indicated that Sae2 is one of the relevant factors regulated by Rpd3L and Hda1. The causal relationship between HDACs and expansions indicates that HDACs can promote mutagenesis at some DNA sequences. This relationship further implies that HDAC3 inhibitors being tested for relief of expansion-associated gene silencing may also suppress somatic expansions that contribute to disease progression.

Figure 1. Mutation or chemical inhibition of yeast HDACs suppresses TNR expansions.

(A) Reporter with (CTG)₂₀ permits expression of the reporter gene CAN1, and results in canavanine sensitivity. Expansions of ≥6 repeats alter transcription initiation, incorporating the out-of-frame ATG codon that blocks expression of CAN1 (X). Canavanine resistance ensues. (B) PCR products displayed on a high-resolution polyacrylamide gel. All expansion results reported here include PCR validation. (C) Expansion rates in mutants of Rpd3L (sin3 or pho23), Hda1 (hda3), or both (pho23 hda3 or sin3 hda3). TNR reporter integration sites are indicated in the figure. Error bars, ±SEM; * p<0.05 compared to wild type; + p<0.05 compared to wild type and to each single mutant (details in Table S1). (D) Cells were grown 13–14 generations in liquid culture ±30 µg/ml TSA, followed by expansion analysis. Error bar, ± SEM; * p = 0.02 compared to DMSO-only control, n = 5 independent measurements. (E) Accumulation of acetylated histone H3 in yeast cells with impaired HDAC activity. Immunoblot results of 15 µg protein from whole cell lysates. Top, acetylated H3; bottom, total H3. Values below the blot show the ratio of acetylated H3/total H3. (F) Expansion sizes, derived from PCR analysis. 26 genetically independent expansions for wild type, 17 for sin3, 25 for hda3, and 8 for hda3 sin3.

Figure 2. Chemical inhibition or RNAi knockdown of HDAC3 in human SVG-A cells suppresses expansions.

(A) The genetic assay is essentially as described . Cells were treated with either HDAC inhibitor 4b, compound 3, or DMSO only. Alternatively, siRNA was used with scrambled siRNA as a control. Expansions are scored using yeast as a biosensor, and total plasmid counts are monitored by bacterial transformation for enhanced sensitivity. (B) Expansion frequencies as a function of inhibitor dose, compared to DMSO-treated control cells. Blue, 4b-treated; red, compound 3-treated. Error bar, ±SEM; * p<0.05 compared to DMSO-treated cells. Details in Table S3. (C) Expansion frequency after RNAi. Knockdown efficiency, judged by three independent immunoblots, averaged 76(±8)% for HDAC3 and 76(±2)% for HDAC1. Error bars, ±SEM; * p<0.05 compared to scrambled control. Details in Table S3. (D) Expansion sizes, derived from PCR analysis. 21 genetically independent expansions for DMSO, 16 for 4b (combined data from 10 µM and 20 µM treatments), 28 for scrambled siRNA, and 13 for HDAC3 siRNA. (E) Cell viability measured by nigrosin staining just prior to cell harvest. (F) Representative immunoblot of acetylated histone H4 and total histone H4 upon treatment with 4b; data summary in Figure S6. (G) Expansion frequencies after RNAi against histone acetyltransferases. Error bars, ±SEM; * p<0.05 compared to scrambled control.

Figure 3. Evidence that Rpd3L acts in trans to promote expansions.

(A) sin3 mutants suppress expansion rates when the TNR reporter is integrated at “hot” zone, INO1 on chromosome X and a “cold” zone, SPS2 on chromosome IV. Error bars, ±SEM; * p<0.05 compared to wild type. (B, C) Chromatin immunoprecipitation using antibodies against pan-acetylated histone H4 or total H4. Underline indicates the TNR reporter integration site at INO1 (B) or SPS2 (C). Rand, control reporter with randomized sequence in place of triplet repeat. Error bars, ±SEM. Primer site details are provided in Figure S8. (D) Expansion rates in single or double mutants of sae2, mre11, exo1, and/or sin3. The reporter was (CTG)₂₀-CAN1 integrated on chromosome II. Error bars, ±SEM; * p<0.05 compared to wild type. Details for panels (A–D) are in Table S4. (E) Model for HDAC promotion of expansions in yeast. 1. Acetylated Sae2 (Ac-Sae2) is marked for degradation, but it is stabilized by deacetylation in an Rpd3L- and Hda1-dependent manner . The same HDACs may deacetylate other factors relevant to expansions, thereby stabilizing them or influencing their activities. The action of Rpd3L and Hda1 is counterbalanced by one or more HATs that await identification. 2. Sae2 along with another nuclease, Mre11, cleaves TNR DNA, possibly in a hairpin structure, to initiate the expansion pathway. 3. The cleaved TNR undergoes additional processing steps to complete the expansion.