| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Somatic sequence variation at the Friedreich ataxia locus includes complete contraction of the expanded GAA triplet repeat, significant length variation in serially passaged lymphoblasts and enhanced mutagenesis in the flanking sequence
Introduction
Results
The expanded GAA triplet repeat undergoes significant length variation in serially passaged lymphoblastoid cells
Reversion of the expanded GAA triplet repeat to normal size in circulating peripheral blood cells
The poly(A) tract upstream of the GAA triplet repeat is polymorphic and displays enhanced instability in expanded alleles
Enhanced somatic mutagenesis upstream of the expanded GAA triplet repeat
Discussion
Serially passaged lymphoblastoid cell lines as a model for studying GAA triplet repeat instability
Reversion of the expanded GAA triplet repeat to the normal size in somatic cells
The expanded GAA triplet repeat as a source of genetic instability
Conclusions
Materials And Methods
Serial passages of lymphoblastoid cell lines
Molecular analyses of the GAA triplet repeat expansion
Genotype analyses
Analysis of the poly(A) tract
Analysis of somatic hypermutability
Acknowledgements
References
Somatic sequence variation at the Friedreich ataxia locus includes complete contraction of the expanded GAA triplet repeat, significant length variation in serially passaged lymphoblasts and enhanced mutagenesis in the flanking sequence
Received July 14, 1999; Revised and Accepted September 13, 1999
The vast majority of Friedreich ataxia patients are homozygous for large GAA triplet repeat expansions in intron 1 of the X25 gene. Instability of the expanded GAA repeat was examined in 23 chromosomes bearing 97-1250 triplets in lymphoblastoid cell lines passaged 20-39 times. Southern analyses revealed 18 events of significant changes in length ranging from 69 to 633 triplets, wherein the de novo allele gradually replaced the original over 1-6 passages. Contractions and expansions occurred with equal frequency and magnitude. This behavior is unique in comparison with other large, non-coding triplet repeat expansions [(CGG)n and (CTG)n] which remain relatively stable under similar conditions. We also report a rare patient who, having inherited two expanded alleles, showed evidence of contracted GAA repeats ranging from nine to 29 triplets in DNA from two independent peripheral blood samples. The GAA triplet repeat is known to adopt a triplex structure, and triplexes in transcribed templates cause enhanced mutagenesis. The poly(A) tract and a 135 bp sequence, both situated immediately upstream of the GAA triplet repeat, were therefore examined for somatic mutations. The poly(A) tract showed enhanced instability when in cis with the GAA expansion. The 135 bp upstream sequence was found to harbor a 3-fold excess of point mutations in DNA derived from individuals homozygous for the GAA triplet repeat expansion compared with normal controls. These data are likely to have important mechanistic and clinical implications.
INTRODUCTION
Friedreich ataxia (FRDA), the most prevalent inherited ataxia, is characterized clinically by progressive gait and limb ataxia, dysarthria, areflexia, pyramidal signs, loss of position and vibration senses, cardiomyopathy, diabetes mellitus and secondary skeletal abnormalities (1,2). It is an autosomal recessive disease with onset in early childhood followed by an unremitting course leading to death in the fourth or fifth decade. Over 95% of affected individuals are homozygous for large expansions of a GAA triplet repeat tract situated within an Alu repeat in the first intron of the X25 gene on chromosome 9 (3). Normal and FRDA chromosomes contain 7-38 and 66-1700 triplet repeats, respectively (3-5). Expanded GAA tracts are associated with transcriptional interference (6,7), resulting in greatly reduced levels of X25 mRNA (6,8,9) and thereby the encoded mitochondrial protein, frataxin (10). This deficiency is directly proportional to the size of the GAA expansion (6,7,10), resulting in the observed correlation between disease severity and length of the shorter of the two expanded alleles (4,11,12). The GAA triplet repeat expansion is thought to mediate its inhibitory influence on transcription via the formation of a non-B-DNA structure (6,7), possibly a higher order, intra-molecular, purine·purine·pyrimidine triplex structure termed `sticky DNA' (13).
Most triplet repeats associated with disease demonstrate varying degrees of instability in somatic cells and following intergenerational transmission (14-17). The significance of triplet repeat instability stems from the finding that shorter repeats correlate with a mild phenotype, and vice versa. The nature and degree of instability may depend on several factors, including repeat length, sequence of the repeat and its pre-disposition to form secondary structures, location of the repeat within the gene, cell/tissue type, parent of origin, sequences within and around the repeat, stage of embryonic development, state of epigenetic modification and, possibly, other cis- and trans-acting factors. In general, the long, untranslated triplet repeats [(CGG)n in the 5[prime]-untranslated region (5[prime]-UTR) of the FMR1 gene, (CTG)n in the 3[prime]-UTR of the DMPK gene, and (GAA)n in intron 1 of the X25 gene] show variation of greater magnitude than the relatively short (CAG)n triplet repeats translated into polyglutamine tracts.
In common with the other triplet repeat diseases, the GAA repeat shows obvious intergenerational and somatic length variation. Expanded alleles associated with disease arise from a small pool of uninterrupted `large normal' alleles termed premutations (18,19). Sequence interruptions in the purely repeated GAA triplet motif interfere with the ability of these large normal alleles expanding into the disease-causing range (18,19). Premutation alleles have been found to undergo large expansions in a single generation (5,18-20). Expanded GAA triplet repeats undergo both expansions and contractions when passed through the female germline but predominantly contract following paternal transmission (21,22). The latter was shown to be due to the shorter repeat length in sperm DNA compared with blood DNA, i.e. by a pre-zygotic mechanism (21). In the same study, a single patient was reported with evidence of reversion of the expanded GAA allele to the normal range in his sperm sample. Additionally, a post-zygotic mechanism for GAA triplet repeat instability is suggested by the findings that the degree of contraction in sperm was more pronounced than that actually observed in intergenerational transmission (20,21) and the overall shorter length of expanded alleles in homozygous versus heterozygous carriers (22).
Triplet repeat instability has so far been studied by assessing repeat length variation following parental transmission, by comparing repeat lengths in DNA from various sampled human tissues (somatic and germ cells), by serial passaging of cells in vitro (23-26), and following the introduction of repeats into model organisms, such as Escherichia coli (27-31), yeast (32-34) and mice (35-47). The limited availability of suitable autopsied material from otherwise inaccessible organs has led to the wide use of alternate experimental models. Even though studies in model organisms can allow the dissection of mechanisms potentially involved in triplet repeat instability, it is increasingly becoming clear that the behavior of triplet repeats in human cells may be fundamentally different.
We have used the paradigm of serially passaged lymphoblastoid cell lines from individuals heterozygous or homozygous for expanded GAA triplet repeats to develop a suitable, human-specific, in vitro model for the study of triplet repeat instability. Contrary to what was found for the other untranslated, long triplet repeats in similar experiments (23-26), the expanded GAA triplet repeat displayed significant length variation. The variation seen during somatic propagation indicates that this may indeed represent a suitable model system.
Intermolecular triplex formation in transfected mammalian cells is associated with spontaneous hypermutagenesis in the flanking sequence (48). Given that the GAA triplet repeat has the propensity to form triplexes in vitro (6,13,49,50), the flanking intron 1 sequence was examined for enhanced somatic mutagenesis. The poly(A) tract and a 135 bp sequence situated immediately upstream of the expanded GAA triplet repeat demonstrated enhanced somatic mutagenesis. This indicates that the expanded GAA triplet repeat sequence is not only mutable but is also mutagenic, and suggests the possibility that similar sequences in the genome may serve as sources of genetic instability.
RESULTS
The GAA triplet repeat is situated in intron 1 of the X25 gene in between the right and left arms of an Alu repeat. Figure 1 shows the sequence context around the GAA triplet repeat along with the regions analyzed for somatic variability in this study.
Figure 1. The GAA triplet repeat and its sequence context within intron 1 of the X25 gene. The GAA triplet repeat sequence [(GAA)n, hatched box] is shown flanked on either side by the left and right arms of an Alu repeat (black boxes). Exons 1 and 2 are shown as numbered boxes flanking intron 1. A close-up of the sequence surrounding the GAA triplet repeat depicts the poly(A) tract (pA) immediately upstream and an A/T tail at the 3[prime]-end of the Alu element (boxes filled with horizontal lines). Arrows depict all primers used in this study. The positions of relevant restriction enzyme recognition sequences are indicated. The 135 bp sequence upstream of the GAA triplet repeat sequence analyzed for somatic point mutations is indicated by a black horizontal bar below the left arm of the Alu repeat, flanked by the primers used to amplify it (GAA-F and GAAF-Rev). The probe used for Southern analysis of the GAA triplet repeat expansion (4) is indicated by a horizontal gray bar below exon 1.
The expanded GAA triplet repeat undergoes significant length variation in serially passaged lymphoblastoid cells
To investigate somatic variation of the expanded GAA triplet repeat a total of 16 Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (seven homozygous and nine heterozygous for the GAA expansion) comprising a cohort of 23 expanded alleles were serially passaged 20-39 times in vitro. The size of the expansions contained in these alleles at the start of the experiment ranged from 97 to 1250 (744 ± 66) GAA triplets. The sizes of the expanded alleles were monitored by genomic Southern analysis, initially for every fourth passage. If an alteration was observed, the analysis was extended to include every single passage in between those that showed the alteration. Length variations (contraction or expansion) were scored as such only if the appearance of a new band was detected; disappearance of bands was not scored. In addition, the identity of all cell lines was verified by genotyping DNA from passages preceding and following the observed instability with at least three polymorphic markers.
A total of four out of seven homozygous and five out of nine heterozygous cell lines showed length instability. With six alleles showing more than one length variation, a total of 18 events were scored which included 11 contractions (range 69-633 triplets, mean 204 ± 50) and seven expansions (range 69-351 triplets, mean 225 ± 42). Interestingly, the de novo allele replaced the original allele over 1-6 passages (Fig. 2). The cell lines were therefore only temporarily `mosaic', with the newly derived allele usually remaining stable over the ensuing passages (Fig. 2). Some of the changes were particularly large, resulting in contractions of up to 39% (Fig. 2F) and expansions of up to 149% of the original allele size seen immediately following lymphoblastoid transformation. These length variations were seen throughout the experiment, with no preference for early or late passages (Fig. 3A and B). Of the 18 events detected, nine were observed in each half of the study, i.e. in 240 individual passages over cycles 1-15 and 233 passages comprising cycles 16-39. The expanded GAA triplet repeat seemed to behave differently depending on whether it was in a homozygous or heterozygous state (Fig. 3C). Six of the 10 changes seen in homozygous cell lines involved three pairs of homologous alleles undergoing simultaneous changes of approximately equal magnitude suggesting an inter-allelic mechanism (Fig. 2C and D). While equal numbers of contractions and expansions were seen in homozygous cell lines, six of the eight changes seen in heterozygous cell lines involved contractions.
A - D
 |
E - G
|
Figure 2. Length variation of expanded GAA triplet repeat sequences in serially passaged lymphoblastoid cell lines. Representative genomic Southern blots of cell lines that are homozygous (A-D) and heterozygous (E-G) for the expanded GAA triplet sequence are shown. Passage numbers are indicated at the top of the respective lanes and the sizes of the expanded alleles are indicated in triplet repeats. N, normal allele in heterozygous cell lines; B, lane containing DNA isolated from peripheral blood leukocytes. Examples of homozygous cell lines with alterations involving one (A and B) or both (C and D) expanded alleles are shown. Examples of small (G), medium (E) and large (F) contractions of the single expanded allele in heterozygous cell lines are shown. A gradual replacement of the original by the de novo allele is seen on serial passaging (C-G).
A
 |
B
|
C
|
Figure 3. Analysis of GAA triplet length variation in serially passaged lymphoblastoid cell lines. (A and B) The temporal progression of all expanded GAA triplet repeat alleles in homozygous cell lines that displayed length variation is graphically represented with passage numbers and expansion sizes (in triplet repeats) plotted along the x- and y-axes, respectively. Alleles in homozygous (A) and heterozygous (B) cell lines are shown separately. Passage 0 indicates the `constitutional' allele sizes observed in blood DNA. A wide range of allele sizes showed length variation, with expansions and contractions occurring throughout the passaging experiment. In one of the homozygous cell lines that showed simultaneous alterations of both alleles [between passages 1 and 4 for the alleles represented by unfilled squares and unfilled triangles in (A)], one of two possible combinations of allelic alterations was selected. (C) Graph showing alterations of GAA triplet repeat length in homozygous versus heterozygous cell lines. The single expanded alleles in heterozygous cell lines frequently contracted (negative reading on the y-axis). In homozygous cell lines three examples of simultaneous changes of approximately equal magnitude were seen involving both homologous expanded alleles (double asterisks).
Overall, contractions and expansions occurred with similar frequency (11 and seven, respectively) and magnitude (204 ± 50 and 225 ± 42 triplets, respectively; P = 0.75). As a whole, there was also no significant difference in the sizes of stable or unstable alleles (793 ± 77 versus 700 ± 106; P = 0.49). Alleles involved in contractions (841 ± 81) were somewhat longer than those that showed expansions (602 ± 98; P = 0.08). Among the unstable alleles in heterozygous cell lines, those longer than the median value (>724 triplets) were significantly more likely to contract (P = 0.025). The magnitude of contractions and expansions correlated significantly with the repeat length of the unstable alleles in homozygous cell lines (Pearson coefficient of correlation r = 0.69 and 0.54, respectively).
As another measure of somatic variability of the expanded GAA triplet repeat, the allele sizes observed in peripheral blood DNA were compared with those in lymphoblast DNA immediately following transformation. Expansion sizes in blood were determined for 20 alleles (from eight heterozygotes and six homozygotes) which ranged from 111 to 1105 triplet repeats (750 ± 69). Following transformation, 12 alleles (60%) showed contractions that ranged from 67 to 236 (134 ± 19) triplet repeats, eight alleles remained unaltered, and none of the alleles showed expansions. The alleles that contracted tended to be longer than those that remained stable (870 ± 45 versus 570 ± 141; P = 0.077) but there was no obvious correlation between the size of the expanded alleles and the magnitude of the contraction. There was no difference between the contractions seen in heterozygous and homozygous cell lines (153 ± 46 versus 125 ± 18; P = 0.6). The alterations following lymphoblastoid transformation are graphically depicted for those alleles that displayed length variation on subsequent serial passaging (Fig. 3A and B, compare passages 0 and 1). It should be noted that the size reduction on lymphoblastoid transformation might simply represent selection of B lymphocytes with shorter repeats and not actual contractions.
The normal allele in each of the nine serially passaged heterozygous cell lines was also investigated for somatic instability. PCR followed by high resolution agarose gel electrophoresis (see Materials and Methods) revealed that the normal alleles contained nine (n = 5), eight (n = 3) and less than eight (n = 1) GAA triplets. Analysis of every fourth passage did not reveal any appreciable instability in each of these nine normal alleles (data not shown).
Reversion of the expanded GAA triplet repeat to normal size in circulating peripheral blood cells
In an FRDA patient (patient 71) shown to be homozygous for the GAA triplet repeat expansion by Southern and long-PCR analyses, a faint PCR product of normal size was detected using primers that preferentially amplify normal alleles, suggesting partial reversion of the expanded allele(s) (Fig. 4A and B). Given the possibility of low level, pre-PCR contamination (despite the absence of a similar amplification product in other patient DNA samples), the patient's blood was resampled 28 months after the first collection. Identical results were obtained, indicating the presence of a subpopulation of peripheral blood cells carrying alleles of normal length in vivo. No such PCR product was obtained on analyzing DNA isolated from a sural nerve biopsy (Fig. 4B) or from early passages of two independently transformed lymphoblastoid cell lines. Both the parents of the patient were found to be carriers of the expansion and haplotype analysis was consistent with the patient having inherited both expanded alleles (Fig. 4C). Direct cloning and sequencing of the PCR product suspected to be the product of reversion revealed a heterogeneous mix of alleles, all within the normal size range containing 9-29 uninterrupted GAA triplets. Consistent with the electrophoretic migration of the PCR product with known small normal alleles (Fig. 4B), the majority of clones were found to contain nine GAA triplets (Fig. 4C). This heterogeneous mix of normal alleles is not likely due to PCR or the cloning procedure, since cloning and sequencing of multiple PCR-amplified normal alleles only showed a maximum of ±1 triplet slippage, and never an expansion into the large normal or premutation size range.
A - C
 |
D - F
|
Figure 4. Reversion of the expanded GAA triplet repeat to normal size in somatic cells. (A) Genomic Southern analysis showed that patient 71 (71B) was homozygous for the GAA triplet repeat expansion. A control individual (642) homozygous for normal alleles (N) and another FRDA patient (93) with homozygous expansions are also shown. The size of the expanded allele is indicated in triplet repeats. (B) PCR analysis using primers 104F and 629R showed the presence of a faint normal sized allele (NA) in blood DNA from patient 71 (71B) but not in the corresponding sural nerve sample (71N). A smear is seen above the faint normal allele amplified in 71B. Normal controls with known repeat lengths are also shown; 570 is homozygous for the (GAA)8 allele and 111 is heterozygous for (GAA)9 (small normal) and (GAA)18 (large normal) alleles. The DNA size marker (L) is labeled in kb and a water blank (W) is shown in the last lane. (C) Haplotype analysis of patient 71 and her parents (82 and 159) using polymorphisms CS2 (C or T alleles), ITR3 (C or T alleles) and F5225 (one to four arbitrarily numbered microsatellite alleles) is shown. The GAA triplet repeat is indicated by the number of repeats in normal alleles [(GAA)8 or (GAA)9] and E denotes an expanded allele. In addition to the two expanded parental chromosomes inherited by patient 71, a chromosome whose parental origin cannot be determined (indicated by the N) represents the presence of a sub-population of somatically reverted alleles. (D) PCR analysis using primers 104F and 629R showed the presence of a normal sized allele (NA) in lymphoblastoid DNA from passage 31 of patient 71, but not in earlier or later passages. The lane marked C refers to a control individual heterozygous for (GAA)8 and (GAA)20 alleles. The DNA size marker (L) is labeled in bp and a water blank (W) is shown in the last lane. (E) PCR analysis using primers Bam and 2500F showed the presence of a normal sized allele (NA) in addition to the expanded alleles (EA) in lymphoblastoid DNA from passage 31 of patient 71 (71-31). The same PCR performed using a blood DNA sample (71B) showed only expanded alleles (see Results). N and C refer to lanes containing similar reactions performed using normal controls and a heterozygous carrier, respectively. The DNA size marker is labeled in kb and a water blank (W) is shown in the last lane. (F) Southern blot analysis of DNA isolated from lymphoblastoid cells from patient 71 showed homozygous expanded alleles (sizes indicated in triplet repeats) that were stable on serial passaging. No evidence of reversion to the normal size (N) was seen in passage 31, indicating that the somatic event seen in 4D is likely to represent a small proportion of cells. C and B refer to lanes containing blood DNA samples from normal control and patient 71, respectively.
PCR analysis of DNA from this patient's serially passaged lymphoblasts revealed the presence of a contracted allele at passage 31, resembling a heterozygous carrier (Fig. 4D and E). While PCR with primers 104F and 629R detected the reverted allele in both blood and passage 31 lymphoblastoid DNA (Fig. 4B and D), the long PCR assay detected this only in lymphoblastoid DNA (Fig. 4E), suggesting that the proportion of cells carrying the reversion event in blood may be lower. The contracted allele was again not detectable by Southern analysis of lymphoblastoid DNA (Fig. 4F), indicating that this involved a small subpopulation of cells. The contracted product amplified from passage 31 was cloned, and sequencing of 10 clones revealed eight containing (GAA)8 and two with (GAA)9 inserts. To investigate possible low level contamination of passage 31 with another cell line, genotypes were determined at nine polymorphic loci, seven of which were microsatellites. The genotype of passage 31 was shown to be identical to that of the blood DNA sample of patient 71 (even following prolonged exposure of autoradiographs) making this unlikely. Taken together, these data indicate that the expanded allele(s) in patient 71 is capable of reverting to the normal size range in somatic cells.
It is interesting to note that the contracted product was detected in lymphoblastoid passage 31 but not in earlier or subsequent passages of this patient, or in a total of 32 passages from seven other homozygously expanded patients. The fleeting nature of the appearance of this subpopulation of cells is nevertheless similar to the behavior encountered in some Southern blots where distinct bands disappeared following a single passage in culture (passages 20-21 in Fig. 2D).
The poly(A) tract upstream of the GAA triplet repeat is polymorphic and displays enhanced instability in expanded alleles
The GAA triplet repeat is situated in between the left and right arms of an Alu element, immediately 3[prime] of a mononucleotide tract of adenines (19) (Fig. 1). This poly(A) tract was analyzed for length variation by sequencing cloned PCR products amplified from blood genomic DNA. The poly(A) tracts in individual alleles were analyzed by determining their phase with respect to a (TAA)2/(TAA)4 dimorphism at the 3[prime]-end of the Alu element (6,19) which was conveniently typed in the same sequence reaction. The poly(A) tract in individual alleles from four informative normal individuals was found to be unstable. Among the various clones sequenced, individual alleles were found to vary by as much as 3-6 nucleotides (nt) in length (intra-allelic variation). This instability made it impossible to determine the constitutive length of the poly(A) tracts. However, in the four informative individuals available the various alleles ranged from 12 to 19 (14.4 ± 0.26) nt in length (inter-allelic variation).
Given the limited number of informative chromosomes, the above method could not be used to examine the poly(A) tract upstream of most normal and all expanded GAA triplet repeats in our cohort. A different assay was used that involved restriction digestion of individually purified poly(A) tract-containing fragments from normal and expanded PCR-amplified alleles followed by high resolution polyacrylamide gel electrophoresis. In order to avoid admixture and allow the examination of individual alleles, normal alleles were isolated from heterozygous GAA expansion carriers and expanded alleles were isolated from patients with homozygous GAA expansions of widely varying sizes. This assay confirmed the polymorphic nature of the poly(A) tract (Fig. 5A). In addition, a tail of shorter products was detected, indicative of the intra-allelic variability that we had previously detected by sequencing. This intra-allelic variability was clearly enhanced in chromosomes carrying expanded triplet repeats, and displayed a distinct bias towards deletion (Fig. 5B).
A
 |
B
|
Figure 5. Analysis of the poly(A) tract upstream of the GAA triplet repeat sequence. End-labeled MboII-StuI fragments containing the poly(A) tract in individual alleles are shown resolved by denaturing polyacrylamide gel electrophoresis [(A)n]. (A) Arrows indicate that the poly(A) tract is polymorphic in length. Note the multiple individual bands seen in each lane. (B) Comparison of the poly(A) tracts upstream of normal versus expanded GAA triplet repeats. Poly(A) tracts in expanded chromosomes displayed significantly more variability in the length of the product resolved in each lane (denoted by the brackets on either side of the gel), indicating enhanced somatic instability compared with those in normal chromosomes.
Enhanced somatic mutagenesis upstream of the expanded GAA triplet repeat
To investigate whether the expanded GAA triplet sequence is associated with mutagenesis in its flanking sequence, a 135 bp sequence situated 20 nt upstream of the poly(A) tract flanking the 5[prime]-end of the GAA triplet repeat was analyzed for somatic mutations (Figs 1 and 6). Two FRDA patients (patients 32 and 62) homozygous for large GAA triplet repeat expansions and two controls (patients 570 and 625) with normal GAA repeat lengths were analyzed for this purpose. This region was amplified by PCR, subcloned, and multiple individual clones were sequenced (see Materials and Methods). Analysis of ~100 clones in each group revealed a 1.5-fold enhanced mutational frequency among patient- versus control-derived clones (Table 1). The predominant source of background in this assay is likely to include Taq DNA polymerase-induced mutations, and it is known that T·A->C·G transitions account for up to 70% of all such changes (51). Indeed, 72% of mutations derived from the normal chromosomes were T·A->C·G transitions (Table 1). On the other hand, only 45% of the mutations seen in the patient-derived clones belonged to this class, differing significantly from the experimentally derived expected frequency seen in the normal clones ([chi]2 = 12.26, P = 0.00046). Subtracting all T·A->C·G mutations from both groups revealed a 3-fold enhanced mutagenesis in expanded versus normal chromosomes ([chi]2 = 28.14, P = 1.1 × 10-7) (Table 1 and Fig. 6). Note that the primers used to amplify this sequence did not span the GAA triplet repeat (Fig. 1), and therefore it is unlikely that the observed increase in mutational frequency is caused by a template-dependent PCR mutation bias.
Figure 6. Somatic hypermutation in the sequence immediately upstream of the GAA triplet repeat expansion. A 135 bp sequence immediately upstream of the poly(A) sequence flanking the 5[prime]-end of the GAA repeat was analyzed for mutations. Primers used to amplify the region are indicated by arrows on either end (primer sequences are not included in the 135 bp analyzed). Position numbers are calculated as nucleotides upstream of the 5[prime]-end of the poly(A) tract. Mutations indicated above and below the sequence represent those identified in DNA from homozygous expansion and homozygous normal cases, respectively. Only mutations other than T·A->C·G are indicated (see text). Mutations encountered more than once are indicated above the same position. A [Delta] sign indicates single base deletions. The repeated deletion of the G at -71 only in DNA derived from expanded chromosomes may represent a hotspot. The deletion of T at -91 (indicated by a large arrow) was identified as a constitutive mutation in a normal chromosome from a heterozygous carrier. The vertical line at -128 indicates the 5[prime]-end of the Alu element within which the GAA triplet repeat is contained.
Table 1. Somatic hypermutation in the sequence immediately upstream of the GAA triplet repeat in expanded chromosomes
| GAA alleles | No. of bp | Clones | All mutations | Class of mutationa | Minus T·A->C·G | ||||
| analyzed | Mutations | Frequency (%) | Mutant clones (%) | T·A->C·G (%) | Others (%) | Frequency (%) | Mutant clones (%) | ||
| Expansions | 13 635 | 101 | 42 | 0.308 | 33 (32.7) | 19 (45.2) | 23 (54.8) | 0.169 | 20 (19.8) |
| Normal | 12 420 | 92 | 25 | 0.201 | 21 (22.8) | 18 (72) | 7 (28) | 0.056 | 6 (6.5) |
| Fold difference | 1.53 | 1.57 | 3.02b | 3.05c | |||||
b[chi]2 = 28.14, P = l.1 × 10-7.
c[chi]2 = 29.19, P = 6.5 × 10-8.
The repeated deletion of G at position -71 in patient-derived clones may represent a mutational hotspot (Fig. 6). This mutation was seen independently in sequences derived from both patients analyzed, and its association with other substitutions in individual clones indicated that there might have been at least four independent occurrences. Moreover, this deletion was never encountered in a comparably large number of clones derived from normal controls. Additionally, there was a slight enhancement of T·A->A·T and T·A->G·C transversions in patient versus control clones [19 (8/42) versus 8% (2/25), and 7 (3/42) versus 0% (0/25), respectively]. The significance of this latter finding is unclear, especially since the absolute numbers of these mutations were low and their relative proportions may alter on identification of additional mutants.
The 135 bp sequence analyzed for somatic point mutations partially overlapped with the right arm of the Alu repeat within which the GAA triplet repeat is situated (Figs 1 and 6). The observed mutational spectrum was also analyzed with respect to the status of the Alu repeat following mutagenesis. A search using RepeatMasker (52) showed that 109 bp of this sequence (positions -20 to -128) (Fig. 6) showed 87.2% identity (96.3% similarity, transversion/transition = 0.4) with the AluSq subfamily of Alu repeats. After incorporating all of the mutations observed in patient-derived clones (except T·A->C·G mutations) the same analysis again detected the AluSq subfamily as the best match, but with 75.2% identity and a significantly higher proportion of transversions (88.1% similarity, transversion/transition = 0.86).
To determine whether this flanking mutagenesis was also seen in the germline, a search was performed for constitutive mutations involving this 135 bp sequence. A total of 33 normal and 44 expanded chromosomes were analyzed using chemical cleavage analysis, sequencing and high resolution gel electrophoresis. A novel single base thymine deletion was identified at position -91 in a normal chromosome (Fig. 6; note that it is not possible to determine which one of the three consecutive T residues is deleted and the middle one was selected arbitrarily). This is the third germline mutation to be identified in the Alu element flanking the GAA triplet repeat (19).
DISCUSSION
The GAA triplet repeat expansion in FRDA is unique among repeat expansions associated with human disease for many reasons. It is the only autosomal recessive expansion mutation, the only one to be situated in an intron and the only disease-causing triplet repeat to have the `GAA' sequence (3). This sequence produces a striking asymmetry with purines on one strand and pyrimidines on the other. Such polypurine·poly-pyrimidine sequences have a propensity for unusual DNA structure formation (49). The GAA triplet repeat within the context of limited intron 1 flanking sequence has been shown to adopt an unusual DNA structure under physiological conditions (6), specifically a bimolecular triplex composed of two intramolecular, purine·purine·pyrimidine triplexes (sticky DNA) (13). This is thought to be the underlying basis for the associated length-dependent deficiency of X25 mRNA and frataxin protein in individuals with the expanded GAA triplet repeat mutation (6,7,13). Thus, the length of the expanded repeat shows a direct correlation with parameters of disease severity, offering some scope for genotype-phenotype correlation. Since the length of the GAA triplet repeat expansion is known to vary between somatic cells, this entails that the relative severity of disease pathology is determined at the level of the cell or tissue. Theoretically, therefore, it is possible that a phenotypically relevant cell type may be relatively spared due to a stochastically determined predominance of shorter alleles. Furthermore, the stability of triplexes is known to be affected in vitro by alterations within and flanking polypurine·polypyrimidine sequences (53,54). The propensity of the GAA triplet repeat expansion to form a triplex-based structure may be similarly compromised by such alterations. Somatic mutations flanking the GAA triplet repeat is therefore yet another level at which the deleterious effects of the expansion mutation may be modulated at the cellular level. It is for these reasons that we decided to analyze the GAA triplet repeat and its flanking sequence for genetic instability in somatic cells and to model these in a convenient and versatile in vitro system.
Serially passaged lymphoblastoid cell lines as a model for studying GAA triplet repeat instability
As a model for the analysis of GAA instability in somatic cells, we have investigated the behavior of expanded GAA triplet repeat alleles in serially passaged lymphoblastoid cell lines. These cells are convenient to establish and maintain in culture and allow the study of a wide range of repeat alleles in heterozygous and homozygous states within the context of relevant cis- and possibly trans-acting factors. This model has previously been used to successfully study the instability of the CTG triplet repeat in myotonic dystrophy, which revealed a gradual increase in size of the expanded alleles, consistent with what has been found in peripheral blood DNA of patients (23,25,26,55,56). This system also identified subpopulations of cells bearing altered CTG repeat alleles by small pool PCR, which revealed the dual nature of somatic variability of this repeat: frequent short alterations and rarer, large changes with a distinct contraction bias (23,30). While these facts validate the suitability of this model system for the study of somatic instability of triplet repeats, potential caveats need to be borne in mind. These cell lines are derived from circulating B lymphocytes that have been transformed following EBV infection. They are grown and stored in vitro and do not represent a pathologically relevant cell type.
Compared with the pattern of instability observed with long CTG and CGG triplet repeats in cultured human cells in vitro, the GAA triplet repeat was found to be remarkably prone to significant length variation. Overall, contractions and expansions were equally frequent and of similar magnitude. Half the cell lines showed alterations, with some alleles mutating more than once. Some changes were large, amounting to contractions and expansions of up to 60 and 49% of the original allele sizes, respectively. Expanded alleles in heterozygous cell lines tended to contract, and half the events seen in homozygous cell lines involved both homologous alleles showing identical changes. Perhaps the most striking feature is the complete replacement of the original alleles by the de novo alleles. Even though this took place over 1-6 passages, the apparent lack of selection given the overall equal frequencies of contractions and expansions makes it difficult to explain this phenomenon mechanistically. It should be noted that variation in the proportions of clonal subpopulations within the individual lymphoblastoid cell lines could account for these changes. This is likely to be the explanation for the behavior observed in at least one heterozygous cell line where the expanded allele initially reduced in size, only to apparently return to its original size by re-expanding in a latter passage (Fig. 3B). However, the difference in polarity of the alterations in heterozygous versus homozygous cell lines and the paucity of such changes in cells carrying CTG repeat lengths under similar experimental conditions indicate that other mechanisms may also be involved. Some of these issues are presently being addressed in experiments where cell lines are deliberately mixed prior to serial passage and by clonal purification of lymphoblastoid cell lines.
An important implication of these observations is that significant alterations in GAA triplet repeat length may be possible during post-fertilization somatic proliferation, with tissue-specific differences in repeat length and likely pathological consequence. This is a potential mechanism for the occasional discrepancy observed between the clinical phenotype and the expansion size estimated merely from blood leukocytes. The GAA repeat has previously been shown to have significantly differing sizes within different parts of the brain (57) and amongst various tissues, including lymphoblasts (9), fibroblasts (12), peripheral nerves (58) and sperm (21). We have previously reported (59) a very mildly affected patient, with an age of onset of 43 years, who is presently ambulatory and in gainful employment at the age of 58. In this patient, large expansions with limited somatic variability were detected in DNA isolated from leukocytes, lymphoblasts, fibroblasts, buccal cells and sural nerve, with estimated mean lengths of the expansions being 854 (± 69) and 1283 (± 72) triplets. Such patients could have short expansions in their spinocerebellar tissues despite the large expansions detected in accessible tissues commonly used for genetic diagnosis and counseling.
Several lines of evidence indicate that somatic cells are an important site for instability of triplet repeats. The expanded GAA triplet repeat shows significant length variation among different regions of the brain, but they are similar within developmentally related subdivisions (57). Also, contraction following paternal transmission is generally less pronounced than the size of the GAA triplet repeat in sperm DNA (20,21), indicating that the repeat must expand in post-fertilization somatic divisions. Recent data on the human Huntington's disease locus suggest that the instability of the CAG repeat occurs due to a mitotic mechanism (60). The continuous proliferation of lymphoblastoid cells in our model system differs from the post-mitotic or end cells in vivo, but may mimic the rapidly proliferative stages during embryogenesis, in stem cells and in pre-meiotic spermatogenic cells, i.e. important stages for the generation of phenotypically significant repeat length variation in somatic cells.
The studies on genotype-phenotype correlation have utilized mainly peripheral blood DNA to determine the `constitutional' repeat length in patients. The shortening of the majority of expanded GAA triplet repeats following lymphoblastoid transformation and the unpredictable but significant length variations seen following passaging in vitro should be borne in mind when using lymphoblasts to determine genotypic or phenotypic parameters.
Reversion of the expanded GAA triplet repeat to the normal size in somatic cells
In our study of a patient (patient 71) who had inherited two expanded alleles, a small proportion of peripheral blood cells bearing a heterogeneous mix of normal GAA triplet repeat alleles was identified. This result is similar to the observation of a fully contracted allele in the sperm sample from an FRDA patient in another study (21). Mechanistically, our finding indicates that the expanded GAA triplet repeat can completely revert to the normal size in somatic cells. This reverted allele was identified in two blood samples collected 28 months apart, suggesting that this is either a recurrent phenomenon or that the contraction is present in a hematopoietic precursor cell. It is not known whether the contraction previously reported in sperm originated mitotically in early spermatogenesis or in meiosis. The finding in our patient indicates that the former mechanism is at least possible.
Patient 71 presented with an atypically mild FRDA phenotype and we had previously detected a shorter expansion (but no normal allele) in a peripheral nerve biopsy (58). The contracted allele reported here was only apparent on subsequent PCR analysis using primers that preferentially amplify the normal allele. Although leukocytes are not pathophysiologically relevant to the phenotype, somatic mosaicism for the fully contracted allele may be causally related to the milder phenotype if in fact the same state exists in the patient's nervous system.
Our findings have diagnostic implications in that if the proportion of cells carrying the contracted allele(s) were higher the long-PCR test performed routinely in most DNA diagnostic laboratories would have indicated a heterozygous carrier status. Fortunately, this is likely to be an uncommon finding, but could be suspected in certain rare situations. In the absence of contamination, even a low level reversion event could yield a PCR product since in the long-PCR test the normal allele is preferentially amplified. Southern analysis and possibly PCR analysis of DNA from other tissues could be used to resolve the diagnostic dilemma. Several symptomatic individuals who are heterozygous for the expanded GAA triplet repeat alleles do not have mutations involving their entire coding sequence and all splice junctions (61; unpublished data). It is tempting to speculate that at least in some of these individuals similar contractions may be the underlying mechanism.
The expanded GAA triplet repeat as a source of genetic instability
The expanded GAA triplet repeat sequence similar to triplex DNA is known to interfere with transcription (6,7) and replication (7,50), creating a climate conducive to template mutagenesis (62). Intermolecular triplex-forming oligonucleotides targeted to a polypurine·polypyrimidine tract in a reporter gene resulted in hypermutagenesis flanking the triplex site in transfected mammalian cells (48). The degree of mutagenesis correlated with the inhibition of transcription mediated by triplex formation and was not seen in cells deficient in transcription-coupled repair. The authors speculated that such sequences, if present in the genome, may constitute endogenous sources of genomic instability, presumably by triggering `gratuitous repair' by the stalled transcription complex resulting in hypermutagenesis (48). The enhanced mutagenesis that we report in the sequence immediately flanking the 5[prime]-end of the expanded GAA triplet repeat supports this notion. The difference in the class of mutations affecting the poly(A) sequence (length variation) and the upstream 135 bp intron 1 sequence (nucleotide substitutions) flanking the GAA triplet repeat is likely to be a function of the template per se.
These mutations are reminiscent of the somatic hypermutability induced by a polypurine·polypyrimidine sequence within the APC gene, enhancing colorectal cancer susceptibility in several Ashkenazi Jewish families (63). Within the coding sequence, a T->A transversion results in an uninterrupted tract of eight A residues which, rather than altering the function of the encoded protein, creates a short hypermutable region, indirectly predisposing to neoplastic transformation. In this regard, the role of the poly(A) tract upstream of the GAA triplet repeat remains unknown. However, its presence in both normal and expanded chromosomes suggests that by itself it is unlikely to be sufficient for the etiology of the somatic instability we have observed. A rather large polypurine·polypyrimidine sequence, situated in intron 21 of the PKD1 gene and shown to form multiple non-B-DNA structures, has been suspected to play a similar role (64,65). It was speculated that this sequence could be somatically hypermutable, resulting in the monoclonal nature of the multiple cysts in this form of polycystic kidney disease.
It is not clear whether these somatic mutations will be reflected in the germline, i.e. result in constitutional allelic mutations. The fortuitous detection of multiple germline mutations in the flanking intron 1 sequence suggests that the latter is likely (19; present study). However, the absence of germline exonic mutations in a comprehensive analysis of >360 chromosomes with GAA triplet repeat expansions during the cloning of the X25 gene (3) indicates that this effect, if real, is likely to be limited to the immediate vicinity of the GAA repeat within intron 1. The consequences of these mutations, if any, are unclear. However, polypurine·polypyrimidine sequences such as the GAA triplet repeat located in other parts of the genome may play a role in local genomic instability, with possible phenotypic or evolutionary consequences.
Conclusions
The expanded GAA triplet repeat at the FRDA locus displays significant length variation in patient-derived cells in vitro. Complete reversion of expanded GAA triplet repeats to the normal non-disease size range can occur in circulating peripheral blood leukocytes in vivo. The expanded GAA triplet repeat sequence is associated with enhanced somatic mutagenesis in the sequence located immediately upstream. This includes length variation in a poly(A) tract and a 3-fold excess of point mutations in a 135 bp sequence. These novel findings are likely to further our knowledge of the molecular biology of the most common mutational mechanism in FRDA, and also increase our understanding of the behavior of polypurine· polypyrimidine sequences in the human genome.
MATERIALS AND METHODS
Serial passages of lymphoblastoid cell lines
Peripheral blood samples were collected following written informed consent, and lymphocytes were isolated by differential centrifugation through Ficoll (Pharmacia, Piscataway, NJ). Lymphoblastoid transformation using EBV was carried out as previously described (66). Cells were maintained in suspension in 25 cm2 flasks in RPMI-1640 (Bio-Whittaker, Walkersville, MD) supplemented with 10% fetal bovine serum. At each passage, after thorough resuspension, 1.5 × 106 cells were transferred to 8 ml of fresh medium. To each flask 2 ml of fresh medium was added every other day. After 1 week in culture the next passage was set up similarly and the remaining cells were used for isolation of genomic DNA. The 20-39 passages in our study represented 37-92 population doublings. The latter is likely to be an overestimate given that we measured total cell counts that would not account for the presence of non-viable cells. A total of 16 cell lines were analyzed. Of the nine heterozygous cell lines used in this study, six were from asymptomatic relatives. The remaining three lines were from individuals with ataxia of unknown etiology who were heterozygous for the GAA expansion and were shown not to have mutations in the entire coding sequence and all splice sites of the X25 gene lacking the expansion.
Molecular analyses of the GAA triplet repeat expansion
Long-PCR amplification of expanded alleles was performed using the Bam and 2500F primer pair (3) and normal alleles were additionally analyzed using the GAA-104F and GAA-629R primer pair (11). Both PCRs were performed using the GeneAmp XL PCR kit (Perkin Elmer, Branchburg, NJ) using the cycling parameters previously described (3,11). Southern blot analysis of BsiHKAI-digested genomic DNA was performed as described elsewhere (4). Sizes of all expanded alleles were determined from Southern blots by plotting band sizes against an exponential curve determined by a co-electrophoresed DNA size marker. The exact number of repeats in normal alleles was determined by sequencing, either directly after gel purification or following cloning into the pCR2.1 or pCR2.1-TOPO TA cloning vectors (Invitrogen, Carlsbad, CA). Additionally, normal alleles in serially passaged lymphoblastoid cell lines were analyzed by PCR (using primers GAA-104F and GAA-629R) followed by 3% agarose gel electrophoresis. A range of normal alleles containing known GAA triplet lengths (by direct sequencing) was used as molecular weight standards during electrophoresis. These gels were able to differentiate between the (GAA)8 and (GAA)9 alleles (contained in all heterozygous cell lines in this study) but were unsuitable for resolution of large normal alleles such as (GAA)19 from (GAA)20.
Genotype analyses
Three previously described polymorphisms at the FRDA locus, F5225, CS2 and ITR3, were used for haplotype analysis (18). F5225 and ITR3 are intragenic and CS2 is located 40 kb centromeric to the GAA repeat. F5225 and CS2 were analyzed as described elsewhere (18). A convenient PCR-based assay was designed for analysis of the ITR3 polymorphism by introducing a mutation (underlined) in the reverse primer (3mutR, 5[prime]-TGT GAA CTA AAA TTC TTA GAG GGG GAA ACA GAA GA-3[prime]) to create a novel EarI site if a C is present instead of a T. PCR was performed along with the previously described primer 3F (3). An additional six microsatellite markers from chromosome 19 (D19S220, D19S412, D19S413, D19S571, D19S926 and D19S1150; Research Genetics, Huntsville, AL) were analyzed to rule out DNA and cell line contamination.
Analysis of the poly(A) tract
The poly(A) tract immediately upstream of the GAA triplet repeat was analyzed by direct sequencing and by restriction digestion followed by high resolution electrophoresis. The poly(A) tracts of individual normal alleles were identified by sequencing clones generated from individuals informative for the (TAA)2/(TAA)4 polymorphism situated 150 bp 3[prime] to the GAA triplet repeat (6,19). PCR was performed using primers A/T-FWD (5[prime]-GGT TGC ATT AAG CCA AGA TCG-3[prime]) and GAA-R (3) to generate a 155 or 161 bp product depending on the presence of two or four TAA repeats. The forward primer was end-labeled using polynucleotide kinase (Boehringer Mannheim, Indianapolis, IN) and [[gamma]-32P]ATP (Amersham, Piscataway, NJ), and the polymorphism was typed following denaturing polyacrylamide gel electrophoresis. Genomic DNA from blood samples of heterozygous individuals was used for PCR with primers GAA-104F and GAA-629R (11), cloned into the pCR2.1 TA cloning vector (Invitrogen) and multiple clones were sequenced. The number of adenines in the poly(A) tract was estimated and could be correlated with the presence of (TAA)2 or (TAA)4 within the same sequence ladder, thus allowing the analysis of intra-allelic variability. The poly(A) tract was also analyzed by digesting gel-purified long-PCR products with StuI and MboII. These products were end-labeled and resolved by denaturing polyacrylamide gel electrophoresis. The poly(A) tract was thus effectively separated from the GAA triplet repeat and contained within an ~115 bp StuI-MboII fragment that was polymorphic and unstable. Comparison of the relative instabilities of poly(A) tracts in expanded and normal alleles was performed by electrophoresis after normalizing gel loading amounts for the intensity of the constant 230 bp MboII end fragment.
Analysis of somatic hypermutability
A 135 bp sequence immediately upstream of the poly(A) sequence flanking the 5[prime]-end of the GAA triplet repeat sequence was amplified by PCR using primers GAA-F (3) and GAAF-Rev (5[prime]-TTT TTT GTA TTT TTT AGT AGA TAC T-3[prime]) which generated a 185 bp product. The 135 bp sequence thus analyzed spanned positions -20 to -154 upstream of the 5[prime]-end of the poly(A) tract. Reaction conditions were as follows: 250 ng of genomic DNA from peripheral leukocytes, 20 pM each primer, 125 µM each dNTP, 60 mM Tris-HCl, pH 8.5, 15 mM (NH4)2SO4, 2.5 mM MgCl2 (Buffer C, PCR Optimizer kit; Invitrogen), 10% DMSO and 1 U Taq DNA polymerase. A thirty cycle PCR was performed on a Perkin Elmer thermal cycler at 55°C annealing temperature. PCR products were extracted with 1 vol of chloroform:isoamyl alcohol (24:1) and following ligation into the pMOSBlue T-vector (Amersham) were transformed into competent MOSBlue E.coli cells (Amersham). Recombinant clones were identified following X-Gal selection, sequenced and analyzed for mutations. The strategy of sequencing multiple clones following amplification by PCR has previously been used with success to identify evidence of somatic hypermutation (67).
ACKNOWLEDGEMENTS
We are grateful to the patients and their families for participating in this study. We thank Darren G. Monckton, José M. Barral and Mehreen Hai for critically reviewing the manuscript and for their many helpful suggestions. This work was supported by grants from the Muscular Dystrophy Association (P.I.P. and S.I.B.), the March of Dimes Foundation (P.I.P. and S.I.B.), the Methodist Hospital Foundation (S.I.B.), the National Ataxia Foundation (S.I.B.) and the Oxnard Foundation through the National Ataxia Foundation (T.A.). Part of this work was carried out when S.I.B. was supported by a Post-doctoral Fellowship from the Muscular Dystrophy Association.
REFERENCES
+Present address: Department of Pathology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA
§To whom correspondence should be addressed. Tel: +1 713 798 5823; Fax: +1 713 798 8526; Email: pragna{at}bcm.tmc.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Soragni, D. Herman, S. Y. R. Dent, J. M. Gottesfeld, R. D. Wells, and M. Napierala Long intronic GAA*TTC repeats induce epigenetic changes and reporter gene silencing in a molecular model of Friedreich ataxia Nucleic Acids Res., November 1, 2008; 36(19): 6056 - 6065. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Karthikeyan, J. H. Santos, M. A. Graziewicz, W. C. Copeland, G. Isaya, B. V. Houten, and M. A. Resnick Reduction in frataxin causes progressive accumulation of mitochondrial damage Hum. Mol. Genet., December 15, 2003; 12(24): 3331 - 3342. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Sharma, S. Bhatti, M. Gomez, R. M. Clark, C. Murray, T. Ashizawa, and S. I. Bidichandani The GAA triplet-repeat sequence in Friedreich ataxia shows a high level of somatic instability in vivo, with a significant predilection for large contractions Hum. Mol. Genet., September 1, 2002; 11(18): 2175 - 2187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Karthikeyan, L. K. Lewis, and M. A. Resnick The mitochondrial protein frataxin prevents nuclear damage Hum. Mol. Genet., May 16, 2002; 11(11): 1351 - 1362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Shoichet, A. T. Baumer, D. Stamenkovic, H. Sauer, A. F. H. Pfeiffer, C. R. Kahn, D. Muller-Wieland, C. Richter, and M. Ristow Frataxin promotes antioxidant defense in a thiol-dependent manner resulting in diminished malignant transformation in vitro Hum. Mol. Genet., April 1, 2002; 11(7): 815 - 821. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Gomes-Pereira, M. T. Fortune, and D. G. Monckton Mouse tissue culture models of unstable triplet repeats: in vitro selection for larger alleles, mutational expansion bias and tissue specificity, but no association with cell division rates Hum. Mol. Genet., April 1, 2001; 10(8): 845 - 854. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. I. Bidichandani, C. A. Garcia, P. I. Patel, and M. M. Dimachkie Very Late-Onset Friedreich Ataxia Despite Large GAA Triplet Repeat Expansions Arch Neurol, February 1, 2000; 57(2): 246 - 251. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||