Massive contractions of myotonic dystrophy type 2-associated CCTG tetranucleotide repeats occur via double-strand break repair with distinct requirements for DNA helicases

Abstract

Myotonic dystrophy type 2 (DM2) is a genetic disease caused by expanded CCTG DNA repeats in the first intron of CNBP. The number of CCTG repeats in DM2 patients ranges from 75 to 11,000, yet little is known about the molecular mechanisms responsible for repeat expansions or contractions. We developed an experimental system in Saccharomyces cerevisiae that enables the selection of large-scale contractions of (CCTG)100 within the intron of a reporter gene and subsequent genetic analysis. Contractions exceeded 80 repeat units, causing the final repetitive tract to be well below the threshold for disease. We found that Rad51 and Rad52 are involved in these massive contractions, indicating a mechanism that uses homologous recombination. Srs2 helicase was shown previously to stabilize CTG, CAG, and CGG repeats. Loss of Srs2 did not significantly affect CCTG contraction rates in unperturbed conditions. In contrast, loss of the RecQ helicase Sgs1 resulted in a 6-fold decrease in contraction rate with specific evidence that helicase activity is required for large-scale contractions. Using a genetic assay to evaluate chromosome arm loss, we determined that CCTG and reverse complementary CAGG repeats elevate the rate of chromosomal fragility compared to a short-track control. Overall, our results demonstrate that the genetic control of CCTG repeat contractions is notably distinct among disease-causing microsatellite repeat sequences.

Keywords: DNA repair; DNA repeats; homologous recombination; microsatellites.

Fig. 1.

Experimental system to study large-scale CCTG repeat contractions in vivo. a) DNA repeats are cloned into the artificial intron, derived from ACT1, of a URA3 reporter gene. The reporter gene is integrated ∼1-kb downstream of the replication origin ARS306. The starting strain with (CCTG)₁₀₀ is Ura−. Repeat contraction renders the cells Ura+. b) Spot assays to evaluate the growth phenotypes of yeast strains with (CCTG)₁₀₀ and (CAGG)₁₀₀ repeats as well as nonrepetitive DNA. c) Distribution of remaining repeat length in Ura+ clones, evaluated by PCR and Sanger sequencing. The median number of repeats was 9. d) Rate of large-scale contraction for (CCTG)₁₀₀ strains, shown with 95% CIs. Rate is calculated using the number of Ura+ clones beginning with 12 independent cultures and the MSS-MLE with a correction for sampling and plating efficiency.

Fig. 2.

Comparison of URA3 reporter gene expression in strains with and without CAGG/CCTG repeats. a) Representation of PCR amplicons to evaluate: i) spliced URA3, ii) 5′ unspliced URA3, and iii) 3′ unspliced URA3. Dashed line denotes that primers will not anneal to the intron sequence. b) Relative expression of spliced and unspliced URA3 transcripts compared to ACT1 endogenous control in 4 yeast strains. Bar graph colors correspond to amplicon colors as in a). The relative expression in the long intron strain was set to 1 for comparison, and the numerical fold difference is in comparison to the long intron strain. At least 3 biological replicates of cDNA for each strain were used for the analysis. The mean relative expression values and SE are plotted. Statistical significance was evaluated by unpaired t-test (*P < 0.05; **P < 0.01) for each primer set in the no intron, (CCTG)₁₀₀, or (CAGG)₁₀₀ cDNA sample compared to the long intron strain.

Fig. 3.

Genetic analysis of large-scale CCTG repeat contractions. Rate of large-scale contraction for (CCTG)₁₀₀ strains, shown with 95% CIs (dashed lines for WT strain). Rate is calculated using the number of Ura+ clones beginning with 12 independent cultures and the MSS-MLE with a correction for sampling and plating efficiency.

Fig. 4.

CCTG repeats elevate chromosomal fragility in an orientation-dependent manner. a) Experimental system to study repeat-induced DNA fragility via chromosomal arm loss. CCTG and CAGG repeats were integrated at the reporter locus, lys2 (blue). DSB formation followed by telomere addition (gray triangle) will result in increased arm loss frequency of the left chromosomal arm. This region (∼40 kb) contains no essential genes and the selectable markers CAN1 and ADE2. Mutations in these genes result in canavanine-resistant (Can^R) and adenine auxotroph clones (Ade−) that appear red when plated on selective media. The closest origin of replication is ARS507. As such, the lagging strand template for DNA replication is noted at the top of the figure. The CCTG and CAGG orientations refer to their placement on the lagging strand template. b) Growth of strains on media containing canavanine and low adenine (5 μg/mL). (GAA)₅ and (GAA)₂₂₀ strains are described in (41). c) CCTG/CAGG repeats elevate arm loss rates. Arm loss rates of (GAA)₅, (CAGG)₁₀₀, (CAGG)₁₃₈, and (CCTG)100 from 3-day incubation on selective media. Arm loss rates were calculated with FluCalc, which uses the MSS-MLE model with a correction for sampling and plating efficiency. Error bars indicate SE from 3 independent experiments. Statistical significance was evaluated by unpaired t-test (*P < 0.05) for each CAGG/CCTG strain compared to the nonfragile (GAA)₅ control.

Fig. 5.

Proposed model of CCTG repeat contraction in budding yeast. During DNA replication, secondary structures such as dumbbells and hairpins form on ssDNA. ssDNA formation could be elevated due to the uncoupling of DNA polymerases and helicase at the replication fork when replicating through the DNA repeats (red/blue tracts). DNA DSBs occur at the CAGG/CCTG repeats, possibly mediated by a structure-specific nuclease(s). Sgs1 helicase promotes DNA strand unwinding, translocating in a 3′ to 5′ direction. At the DSB end close to the start of the repeat tract, the 5′ end will be displaced by Sgs1 (hairpin omitted for clarity). At the 3′ end, Rad52 (black circle) will help load Rad51 (yellow diamond) in a 5′ to 3′ direction. At the other DSB end in the absence of Dna2 endonuclease activity, Sgs1 can unwind the repeat DNA more extensively. Hairpin formation can occur on the newly displaced CAGG strand (^). Rad51 and Rad52 promote DNA invasion of the CCTG 3′ overhang to the homologous template. Because of the repetitive nature of the DNA template, out-of-register alignment toward the distal end of the repeat tract will result in a massive contraction. Importantly, only events where the invading DNA end aligns toward the edge of the repeat tract on the template strand will result in a large enough repeat contraction for Ura+ selection. MutSβ may be involved in stabilizing secondary structures and/or processing flap removal following strand annealing.