MutSβ abundance and Msh3 ATP hydrolysis activity are important drivers of CTG•CAG repeat expansions

Abstract

CTG•CAG repeat expansions cause at least twelve inherited neurological diseases. Expansions require the presence, not the absence, of the mismatch repair protein MutSβ (Msh2-Msh3 heterodimer). To evaluate properties of MutSβ that drive expansions, previous studies have tested under-expression, ATPase function or polymorphic variants of Msh2 and Msh3, but in disparate experimental systems. Additionally, some variants destabilize MutSβ, potentially masking the effects of biochemical alterations of the variations. Here, human Msh3 was mutated to selectively inactivate MutSβ. Msh3-/- cells are severely defective for CTG•CAG repeat expansions but show full activity on contractions. Msh3-/- cells provide a single, isogenic system to add back Msh3 and test key biochemical features of MutSβ on expansions. Msh3 overexpression led to high expansion activity and elevated levels of MutSβ complex, indicating that MutSβ abundance drives expansions. An ATPase-defective Msh3 expressed at normal levels was as defective in expansions as Msh3-/- cells, indicating that Msh3 ATPase function is critical for expansions. Expression of two Msh3 polymorphic variants at normal levels showed no detectable change in expansions, suggesting these polymorphisms primarily affect Msh3 protein stability, not activity. In summary, CTG•CAG expansions are limited by the abundance of MutSβ and rely heavily on Msh3 ATPase function.

Figure 1.

Loss of Msh3 causes a defect in CTG repeat expansions but not contractions. (A) Schematic showing predicted CRISPR/Cas9 cut site in exon 2 of human Msh3 gene. Sequencing results show in-frame nine base pair deletion with corresponding loss of codons for amino acids 116–118. (B) Representative immunoblot with an antibody against Msh3 residues 136–349, showing loss of Msh3 expression in Msh3^−/−cells in comparison to Msh3^+/+ wild type cell line. Expression of Msh2 and Msh6 was retained. (C) Quantitative analysis of Msh protein expression in Msh3^+/+ and Msh3^−/−cell line, normalised to actin and relative to Msh3^+/+. n = 9 for Msh3 levels; n = 7 for Msh3^+/+Msh2 and Msh6 protein levels; and n = 6 for Msh3^−/−Msh2 and Msh6 protein levels. *P = 2.23 × 10⁻²⁰. (D) Growth curve analysis of Msh3^+/+and Msh3^−/−cell lines, n = 3. (E) Expansion frequencies of Msh3^+/+and Msh3^−/− cell lines, n = 3 for Msh3^+/+and n = 4 for Msh3^−/−, *P = 0.0019. (F) Contraction frequencies of Msh3^+/+and Msh3^−/− cell lines, n = 3. For panels C-F, error bars denote ±SEM.

Figure 2.

Msh3 overexpression cell line shows increased expansion frequency and increased MutSβ formation with no effect on contraction frequency. (A) Representative immunoblot showing Msh3 protein expression in Msh3 add-back cell line (Msh3^1.7X) in comparison to Msh3^+/+and Msh3^−/− cell lines. (B) Quantitative analysis of Msh3 protein expression normalised to actin and relative to Msh3^+/+cell line. Msh3 add-back cell line has an average Msh3 expression level of 1.7 times wild type expression. n = 9 for Msh3 protein expression in all cell lines, n = 7 for Msh2 and Msh6 protein levels in Msh3^+/+ cell lines, n = 6 for Msh2 and Msh6 protein levels in Msh3^−/− cell line and n = 5 for Msh2 and Msh6 protein levels in Msh3^1.7xcell line. *P = 2.23 × 10⁻²⁰ and **P = 1.87 × 10⁻⁶, both compared to Msh3 level in Msh3^+/+cells. ***P = 0.032 compared to Msh6 level in Msh3^+/+cells. (C) Expansion frequencies for Msh3^−/−, Msh3^+/+ and Msh3^1.7Xcell lines. n = 4 for Msh3^−/−and n = 3 for Msh3^+/+and Msh3^1.7X. *P = 0.029 compared to wild type. (D) Contraction frequencies of Msh3^−/−, Msh3^+/+and Msh3^1.7Xcell lines, n = 3. (E) Representative immunoblot for co-immunoprecipitation of Msh3 and Msh2 and investigation of MutSβ formation in Msh3^−/−, Msh3^+/+ and Msh3^2.9X cell lines. (F) Quantitative analysis of immunoprecipitated Msh3 and Msh2 protein abundance, n = 4 for all samples. *P = 5.44 × 10⁻¹³ for Msh3 and P = 4.30 × 10⁻¹¹ for Msh2 compared to wild type; and **P = 1.66 × 10⁻⁴ for Msh3 and P = 4.96 × 10⁻³ for Msh2 compared to wild type. For panels B, C, D and F error bars denote ±SEM.

Figure 3.

Msh3^E976A ATPase Walker B mutant shows a decreased expansion frequency compared to Msh3^+/+ wild type. (A) Schematic showing Walker B motif consensus sequence and the E976A mutation, ‘h’ denotes any hydrophobic amino acid. (B) Representative immunoblot showing Msh protein expression in Msh3^+/+ and Msh3^E976Acell lines. Msh3^E976A cells show Msh3 expression at near wild type levels. (C) Representative immunoblot of co-immunoprecipitation of Msh3 and Msh2 and MutSβ formation in Msh3^+/+and Msh3^E976A cell lines. (D) Expansion frequencies for Msh3^E976A cell line in comparison to Msh3^+/+, n = 3, *P = 0.037. (E) Contraction frequencies of Msh3^E976Aand Msh3^+/+ cell lines, n = 5 for Msh3^+/+and n = 4 for Msh3^E976A. (F) Representative ATPase analysis of MutSβ and MutSβ-Msh3E976A (MutSβ EA) in the absence or presence of mismatched DNA containing a two-nucleotide insertion/deletion mispair. Pi, [³²P]phosphate. (G) Average ATP hydrolysis activity, n = 3. *P = 0.028 compared to wild type MutSβ assayed without DNA (unshaded bars); **P = 0.001 compared to wild type MutSβ assayed with DNA (shaded bars). For panels D, E and G, error bars denote ±SEM.

Figure 4.

Msh3 polymorphic cell lines show no significant difference in expansion frequency in comparison to Msh3^+/+ cell line. (A) Schematic showing polymorphic changes to Msh3 sequence. (B) Representative immunoblot showing Msh protein expression in Msh3^T363Iand Msh3^T1045Acell lines at comparable expression levels to Msh3^+/+wild type cells. (C) Representative immunoblot of co-immunoprecipitated Msh3 and Msh2 showing MutSβ complex formation in Msh3^T363Iand Msh3^T1045A cell lines alongside Msh3^+/+cells. (D) Expansion frequencies for polymorphic Msh3 cell lines assayed in parallel to Msh3^+/+. Error bars denote ±SEM, n = 3.