Selectable System for Monitoring the Instability of CTG/CAG Triplet Repeats in Mammalian Cells

Despite substantial progress in understanding the mechanism by which expanded CTG/CAG trinucleotide repeats cause neurodegenerative diseases, little is known about the basis for repeat instability itself. By taking advantage of a novel phenomenon, we have developed a selectable assay to detect contractions of CTG/CAG triplets. When inserted into an intron in the APRT gene or the HPRT minigene, long tracts of CTG/CAG repeats (more than about 33 repeat units) are efficiently incorporated into mRNA as a new exon, thereby rendering the encoded protein nonfunctional, whereas short repeat tracts do not affect the phenotype. Therefore, contractions of long repeats can be monitored in large cell populations, by selecting for HPRT+ or APRT+ clones. Using this selectable system, we determined the frequency of spontaneous contractions and showed that treatments with DNA-damaging agents stimulate repeat contractions. The selectable system that we have developed provides a versatile tool for the analysis of CTG/CAG repeat instability in mammalian cells. We also discuss how the effect of long CTG/CAG repeat tracts on splicing may contribute to the progression of polyglutamine diseases. A growing number of neurodegenerative diseases such as myotonic dystrophy, Huntington's disease, and spinocerebellar ataxias have been found to result from expansion of CTG/CAG trinucleotide repeats. Normal individuals generally have fewer than 30 repeats at a disease locus, whereas affected individuals may have from 30 to several thousand repeat units, depending on the disease. The mechanism by which an expanded repeat causes the characteristic features of a disease is complex and depends on the orientation of the repeat within the gene. In myotonic dystrophy, for example, the mRNA from affected individuals carries an expanded CUG repeat in its 3′ untranslated region that may contribute to disease pathogenesis by binding CUG binding proteins, which can alter the splicing of other, unrelated mRNAs (see reference 51 for a review). Additional factors such as nuclear retention of the CUG-containing transcripts (49) and hypermethylation of the adjacent chromatin (48) may also contribute to the disease phenotype. In Huntington's disease and spinocerebellar ataxias 1, 2, 3, 6, and 7, a CAG trinucleotide repeat is expanded within the coding region of the affected gene. The resulting neurodegeneration has been linked to accumulation of aberrant polyglutamine-containing proteins, which induce neuronal death (see reference 55 for a review). The exact mechanism leading to expansion of trinucleotide repeats is unknown. It is likely to be related to the ability of repeat tracts to form unusual DNA secondary structures such as hairpins and slipped-strand DNA duplexes, which can interfere with aspects of DNA metabolism (see reference 52 for a review). Both Escherichia coli and Saccharomyces cerevisiae have been used as model systems to study the instability of CTG/CAG repeats. Virtually every process that exposes single strands of DNA destabilizes triplet repeats, including transcription (2, 45), nucleotide excision repair (35, 38), mismatch repair (21, 42, 44, 46), replication (17, 22, 30, 43), and recombination (9, 18-20, 39, 40). CTG/CAG triplet repeats also cause double-strand DNA breaks in yeast (9, 20). Similarly, studies with mammalian cells have shown that DNA replication (6, 37), mismatch repair (23, 27, 53), and proximity to CpG islands (4) contribute to destabilization of triplet repeats. Based on these studies, several models of repeat expansion have been proposed (reviewed in references 3 and 52). Small changes in repeat length may be caused by the slippage of DNA polymerases, while larger changes may result from errors of DNA repair machinery. For example, if single- or double-strand breaks are formed close to the repeat tract, flaps, hairpins, or other complex DNA structures could form at the ends, leading to errors in DNA repair. Furthermore, alternative DNA structures formed at the repeat locus might by themselves be recognized by DNA repair proteins (such as mismatch repair machinery), which could lead to aberrant processing and promote repeat expansions or contractions (reviewed in reference 47). Understanding the mechanisms that lead to expansion of triplet repeats may allow development of approaches to prevent the lengthening of repeat tracts or even to induce contractions of the expanded repeats in order to stop the progression of neurodegenerative disorders. To search for genetic factors or therapeutic treatments that affect repeat stability, it is critical to use a sensitive assay. Assays based on inactivation of a selectable reporter gene such as URA3 or chloramphenicol acetyltransferase have been developed for yeast (34, 41) and E. coli (14). However, to achieve the ultimate goal of finding a cure for triplet repeat diseases, it is essential to find treatments that are effective in mammalian cells. Presently, trinucleotide repeat instability is assayed in mammalian cells by methods such as small-pool PCR and GeneScan, which can detect frequencies of repeat change in the range from 10−2 to 10−3, and thus lack the sensitivity of a selectable genetic assay. In this report we describe a new genetic assay for trinucleotide repeat contractions in mammalian cells that is selective and quantitative. The assay is based on the novel finding that long CAG repeats cloned into an intron of a reporter gene disrupt correct splicing and become incorporated into mRNA, thereby inactivating the gene. Using this system, we demonstrate that aphidicolin and hydroxyurea, which affect DNA replication, and gamma irradiation, which induces DNA breaks, destabilize long repeat tracts in mammalian cells. The selectable system that we have developed provides a versatile tool for the further analysis of CTG/CAG repeat instability