Nucleotide repeat instability is associated with an increasing number of cancers and neurological disorders. The mechanisms that govern repeat instability in these biological disorders are not well understood. To examine genetic aspects of repeat instability we have introduced an expanded CAG trinucleotide repeat into transgenic mice. We have detected intergenerational CAG repeat instability in transgenic mice only when the transgene was maternally transmitted. These intergenerational instabilities increased in frequency and magnitude as the transgenic mother aged. Furthermore, triplet repeat variations were detected in unfertilized oocytes and were comparable with those in the offspring. These data show that maternal repeat instability in the transgenic mice occurs after meiotic DNA replication and prior to oocyte fertilization. Thus, these findings demonstrate that advanced maternal age is an important factor for instability of nucleotide repeats in mammalian DNA.
Spinocerebellar ataxia type 1 (SCA1) belongs to a group of neurological disorders caused by the expansion of a CAG trinucleotide repeat in the coding region of the associated gene ( 1 ). Other neurodegenerative disorders belonging to this group include spinocerebellar ataxia type 2 (SCA2) ( 2–4 ), Machado-Joseph disease (MJD/SCA3) ( 5 ), spinocerebellar ataxia type 6 (SCA6) ( 6 ), Huntington's disease (HD) ( 7 ), spinal and bulbar muscular atrophy (SBMA) ( 8 ) and dentatorubral-pallidoluysian atrophy (DRPLA) ( 9 , 10 ), which is allelic to Haw River Syndrome ( 11 ). An intriguing aspect of SCA1 genetics is anticipation, a worsening of clinical symptoms and earlier age of onset in successive generations ( 12 , 13 ). The molecular basis of anticipation is the expansion of the CAG repeat upon transmission from parent to offspring.
Three recent transgenic studies have found trinucleotide repeat instability when the transgene is transmitted from parent to offspring. Two of these studies used independent genomic fragments containing the myotonic dystrophy (DM) CTG repeat ( 14 , 15 ). DM is a trinucleotide repeat expansion disorder in which the CTG repeat is located in the 3′-untranslated region of the DMPK gene. In the third study, a 1.9 kb genomic fragment encompassing the 5′-end of the HD gene, including the CAG repeat, was used to generate transgenic mice ( 16 ). These studies found both maternal and paternal intergenerational repeat instability. Both repeat tract expansions and contractions, generally of the order of one to four repeats, were observed. In the HD transgenic mice a paternal age effect was also observed. As the founder animal aged, the size of the repeat tract expansions in his offspring increased. Thus, in the aged founder, transmissions with tract expansions of the order of 10 repeat units were observed.
Although repeat instability has been observed in transgenic mice, the mechanisms that govern repeat instability in these biological disorders are not well understood. In particular, the temporal relationship between repeat instability and mammalian development is unclear. To examine genetic aspects of repeat instability we have introduced a SCA1 cDNA, containing a CAG trinucleotide repeat tract, into transgenic mice and analyzed both maternal and paternal transmissions of the repeat.
Maternal intergenerational CAG repeat instability
The analysis of CAG repeat intergenerational instability was carried out using the single copy SCA1 cDNA transgenic lines designated D02 (expanded allele, 82 CAGs) and C01 (normal interrupted allele, 30 repeats) ( Fig. 1 ). Male transmissions of the transgene carrying the expanded CAG repeat within line D02 were found to be very stable (three unstable in 86 transmissions). In contrast, 67% of female transmissions of this transgene were unstable (143 unstable in 213 transmissions). Interestingly, all of the maternal intergenerational instabilities were contractions in CAG repeat number. These contractions ranged from the loss of a single CAG to a reduction of nine CAG triplets. Maternal instability was observed in transmissions from the founder animal as well as in females from all subsequent generations. D02 females homozygous for the SCA1 transgene demonstrated intergenerational CAG repeat instability identical to that seen in heterozygous females (unpublished data). A similar analysis of CAG intergenerational instability in a transgenic line carrying an unexpanded (30 repeats) allele with an interrupted repeat tract, C01, failed to reveal any instability in either male or female transmissions (58 and 60 transmissions, respectively). This result is in agreement with data from studies of the human disorder that suggests that the presence of a CAT interruption stabilizes the SCA1 CAG repeat tract ( 17 , 18 ). In addition to germline repeat instability, D02 animals were analyzed for somatic repeat instability. In numerous tissues analyzed from these animals no somatic repeat variation was observed (unpublished data).
Effect of maternal age on CAG repeat instability
An interesting aspect of the CAG tract length instability in offspring within transgenic line D02 was enhanced instability of repeats with advancing age of the transmitting mother. For example, in one case it was possible to analyze the SCA1 CAG repeat in four litters from a single D02 female at 21, 24, 30 and 38 weeks of age ( Fig. 2 ). From the first to the fourth litter there was an increase in the magnitude of the CAG repeat contraction. Two of the four transgenic offspring from the fourth litter have CAG tract reductions of nine repeats. Comparing the change in repeat size found in all D02 offspring versus the age of the transmitting mother ( Fig. 3 ) demonstrated an increase in the magnitude of CAG tract intergenerational instability with increasing age of the transgenic mother ( P < 1×10 −5 ). Moreover, in transmissions from females 20 weeks of age or older virtually all of the transmissions were unstable.
Expanded CAG tracts become unstable in oogenesis
To determine the stage during female gametogenesis when the CAG repeat becomes unstable we isolated superovulated unfertilized mature oocytes and performed single cell PCR to examine the CAG repeat size ( Fig. 4 ). Unfertilized oocytes from a 7-week-old transgenic female showed complete stability of the CAG tract ( Fig. 4 A). Of the 10 oocytes examined, none demonstrated a change in length of the SCA1 CAG repeat tract compared with that in the tail DNA of the mother. In contrast, oocytes from a 20-week-old transgenic female displayed repeat tract instability ( Fig. 4 B). In the 10 oocytes examined from this animal, seven had repeat tracts shorter than in the tail DNA of the mother. The size of the contractions were 1, 1, 1, 2, 3, 3 and 4 repeat units. Thus, the frequency and magnitude of the contractions is comparable with CAG repeat tract changes observed in offspring from similar aged mothers ( Fig. 3 ).
We have found that an expanded (mutant) CAG repeat tract exhibits intergenerational instability only when maternally transmitted in SCA1 cDNA transgenic mice. In addition, females that have inherited a CAG tract with a fewer number of repeats transmit alleles with further contractions in repeat number (unpublished data). Maternal contractions were also observed in two other SCA1 transgenic lines: B02 (10 copies with 82 repeats) and E04 (a single copy with 55 repeats) (unpublished data). This suggests that repeat instability in the transgenic mice is integration site independent.
The inability to observe instability in our earlier studies involving SCA1 cDNA transgenic mice ( 19 ) is likely explained by a failure to analyze sufficient numbers of maternal transmissions and offspring of older mothers. The relevance of the repeat instability in the SCA1 transgenic mice to the intergenerational instability observed in the human disorder is not clear. However, the instability observed in the SCA1 transgenic mice begins to address general aspects of DNA integrity during oogenesis.
The finding that maternal instability in the SCA1 transgenic mice was affected by the age of the transmitting mother indicates that there is a window during oocyte development in which CAG instability occurs ( Fig. 5 ). Detection of CAG repeat instability in unfertilized oocytes indicates that the induction of CAG instability occurs prior to fertilization. Moreover, the frequency of repeat instability in unfertilized oocytes isolated from an older female was comparable with CAG instability observed in offspring of similar aged mothers. The inability to detect repeat instability in oocytes collected from a young female suggests that instability occurs while the oocytes are arrested in meiosis I after meiotic DNA replication.
In mice, all oocytes have reached the diplotene stage of the first meiotic prophase by 5 days after birth ( 20 ). Most oocytes die, with only a fraction maturing over the entire lifespan of the female. Previously postulated mechanisms of trinucleotide repeat instability involve DNA replication or recombination ( 21–23 ). Because meiotic DNA replication during female gametogenesis is completed by 5 days after birth, our observations cannot be explained by a DNA replication-based mechanism of CAG instability. A recombination-based mechanism of CAG repeat instability, such as unequal sister chromatid exchange, would also give rise to both CAG expansions and contractions. However, most recombination has initiated by the pachytene stage of meiotic I prophase, which occurs prior to meiotic arrest ( 24 , 25 ). Furthermore, only contractions were observed. Therefore, none of these mechanisms of CAG instability provide a plausible explanation of the maternally derived instability in our SCA1 cDNA transgenic mice.
A mechanism of recombinational repair consistent with our observations is single strand annealing, or end-joining ( 26–28 ). In this type of recombinational DNA damage repair, damaged DNA is excised and the DNA ends, facilitated by homologous pairing, are joined back together. This results in an overall loss of some of the repeating units from which the DNA damage was removed. DNA damage, e.g. oxidative damage, may accumulate over time within the arrested oocytes ( 29 ). The repair of this damage could explain the maternal age relationship and the increased size of CAG contractions. Older oocytes may have accumulated more DNA damage that when repaired gives rise to larger contractions. Our data is consistent with this hypothesis in that we observed CAG instability in mature oocytes isolated from older females, but failed to observe repeat instability in oocytes isolated from young females. This also suggests that the CAG tract may be a hotspot for the accumulation of DNA damage and that repair of this damage leads to repeat tract instability. Moreover, the fact that we have observed only maternal instabilities directly supports the concept that the mechanisms governing instability in female and male gametogenesis are different.
Our transgenic mouse data point to maternal age as an important factor influencing nucleotide repeat tract instability. Moreover, there appears to be a specific period during oogenesis in which the capacity to maintain the integrity of a trinucleotide repeat tract is diminished with increasing age. This temporal relationship between repeat tract instability and oocyte maturation suggests that the efficiency of the DNA repair and recombination machinery varies during oogenesis.
Materials and Methods
The cDNA transgenes used to generate lines D02 and C01 are similar to those previously described ( 19 ) except that they contain the SV40 small t intron within the SV40 polyadenylation sequence. Expression of these cDNA transgenes was driven by the Purkinje cell-specific promoter Pcp-2 ( 19 , 30 ). The D02 and C01 transgenic constructs were assembled by subcloning the 891 bp Pst I- Bam HI fragment containing the SV40 small t intron and polyadenylation signal from pRSVneo ( 31 ) into pBluescript KS+ (Stratagene). This fragment was then released from pBS with Eco RV and Bam HI and subcloned directionally into pGEM11 digested with ApaI (blunted with T4 DNA polymerase) and Bam HI. The resultant clone, pGEM11/polyA was linearized with Xba I and then ligated to the Xba I fragment containing the Pcp-2/SCA1 coding sequence recovered from PS-30 or PS-82 respectively ( 19 ). Orientation of the Pcp-2-SCA1 fragment of the transgene was confirmed by restriction and sequence analyses. The transgenes were recovered from pGEM11 by Not I /Sal I digestion, gel purified and prepared for injection as described ( 20 ). All transgenic mice were generated using the inbred FVB/N strain.
DNA isolation and identification of transgenic animals
Tail DNA was isolated from postnatal animals by methods previously described ( 32 ). Transgenic animals were identified using a transgene-specific PCR assay ( 19 ). The 5′ primer, 5EX2B (5′-AGGTTCACCGGACCAGGAAGG-3′), located within Pcp-2 exon 2, and the ataxin-1-specific 3′ primer RUBY (5′-AATGAACTGGAAGGTGTGCGGC-3′), were used to amplify a 555 bp fragment. Approximately 100 ng genomic DNA was used in a 25 µl PCR reaction that contained 200 µM dNTPs, 1 µM each primer, 2% formamide, 1× standard PCR buffer (Boehringer Mannheim) and 0.5 U Taq DNA polymerase (Boehringer Mannheim). PCR conditions were 94°C, 1 min, 62°C, 1 min, 72°C,1 min for 30 cycles.
CAG repeat PCR analysis
PCR conditions were performed as previously described ( 19 ). In brief, ∼100 ng mouse genomic DNA was used in a 15 µl reaction that contained 1 µM Rep-2 and [γ- 33 P]ATP end-labeled Rep-1 primers, 250 µM each dNTP, 2% formamide, 1.25 mM MgCl 2 , 50 mM KCl, 10 mM Tris-HCl, pH 8.3, and 0.5 U Taq DNA polymerase (Boehringer Mannheim). Cycling conditions were as follows: 94°C for 4 min followed by 30 cycles of denaturing at 94°C for 1 min, annealing at 57°C for 1 min, extension at 72°C for 1 min and final extension at 72°C for 8 min. A 3 µl aliquot of the PCR reaction was mixed with 2 µl formamide containing loading dye, denatured, run on a 5.5% polyacrylamide sequencing gel and visualized by autoradiography. The size of the CAG repeat tract was determined by comparison with a M13 DNA sequencing ladder.
Oocyte collection, preparation and analysis
SCA1 transgenic females were superovulated by administration of pregnant mare serum (PMS) followed by human chorionic gonadotrophin (hCG) 48 h later. The following morning, ovulated unfertilized oocytes were collected. Individual oocytes were lysed for 10 min at 65°C in 5 µl 200 mM KOH, 50 mM DTT and neutralized with 5.2 µl 200 mM Tricine, pH 4.93, and prepared for PCR as previously described ( 33 ). A hemi-nested PCR strategy was used to amplify the SCA1 CAG repeat using the previously described SCA1 -specific primers GCT-435, Rep-1 and GCT-214 ( 17 ).
The first round of PCR was carried out as previously described ( 33 ), using SCA1 primers GCT-435 and Rep-1. Briefly, each 50 µl reaction contained 0.5 µM GCT-435 and Rep-1 primers, 100 µM each dNTP (7-deaza dGTP was used), 3% (v/v) DMSO, 10% (v/v) glycerol, 1.2 mM Mg(OAc) 2 and 1 U rTth DNA polymerase (Perkin-Elmer). The final K + concentration was adjusted to 100 mM by adding 80 mM KOAc. Cycling conditions were as follows: initial denaturation at 94°C for 5 min, followed by denaturation at 94°C for 30 s, extension at 60°C for 5 min for 8 cycles, then denaturation at 94°C for 30 s, extension at 60°C for 3 min for 18 cycles and a final extension at 72°C for 8 min. One microliter of the product from this reaction was used as template in a second round of PCR. Each 20 µl reaction contained 0.175 µM GCT-435 and 0.2 µM [γ- 33 P]-end-labeled GCT-214 primers, 250 µM each dNTP, 2% formamide, 1.25 mM MgCl 2 , 50 mM KCl, 10 mM Tris-HCl, pH 8.3, and 0.5 U Taq DNA polymerase (Boehringer Mannheim). Cycling conditions were as follows: 94°C for 4 min, followed by 30 cycles of denaturing at 94°C for 1 min, annealing at 65°C for 1 min, extension at 72°C for 1 min and a final extension at 72°C for 8 min. A 3 µl aliquot of the PCR was run on a polyacrylamide gel and visualized as described above.
We thank Dennis Livingston for critical reading of the manuscript and thoughtful discussions. The authors would also like to thank Sandra Horn and Barbara Pinch for technical assistance. This work was supported by grants from the National Institute of Neurological Disorders and Stroke of the NIH to H.T.O. (NS22920) and to H.Y.Z. (NS27699) and in part by The Wills Foundation through a grant to M.D.K. H.Y.Z. is a Howard Hughes Medical Institute Investigator.