Chapitre II : Use of a β-glucuronidase (GUS) reporter gene to study mononucleotide instability in Arabidopsis thaliana

Table des matières

Use of a β-glucuronidase (GUS) reporter gene to study mononucleotide instability in Arabidopsis thaliana

Aïda Azaiez, Martine Jean, and François Belzile*

Département de phytologie, Université Laval, Québec, Qc, Canada, G1K 7P4

*Corresponding author :

F. J. Belzile

1243 Pavillon C. E. Marchand, Université Laval, Québec, Canada G1K 7P4

tel : +1 418 656-2131 ext. 5763 ; fax : +1 418 656-7176

e-mail : fbelzile@rsvs.ulaval.ca

Key words : Arabidopsis thaliana , microsatellite instability, frameshift mutations.

Le maintien de la stabilité du génome est la clé de la survie des espèces. Les microsatellites sont des séquences d’ADN répétées en tandem qui représentent des régions instables du génome à cause des variations de leur longueur dues au dérapage de l’ADN polymérase durant la réplication. Afin d’étudier l’instabilité des microsatellites chez Arabidopsis thaliana , un nouveau système rapporteur a été développé. Plusieurs microsatellites mononucléotidiques (G7, G10, G13, G16 et C16) ont été introduits hors cadre de lecture dans le gène rapporteur GUS codant pour la β-glucuronidase. Certaines mutations dans le microsatellite rétablissent l’activité du gène GUS et se traduisent par des secteurs bleus au niveau des cellules de la plante. La fréquence des secteurs bleus a été utilisée pour mesurer l’instabilité des divers microsatellites dans les tissus somatiques, ainsi qu’à la suite de la transmission germinale de nouveaux allèles. Nos résultats ont révélé que l’instabilité des microsatellites augmente avec la longueur de la répétition et qu’elle est également fonction de l’orientation de l’unité répétée. Nous en concluons que le système rapporteur GUS est un outil puissant permettant d’étudier l’instabilité des microsatellites dans le génome des plantes.

Maintenance of genome stability is key to the survival of species. Microsatellites are simple, tandem DNA repeats that represent unstable regions of the genome, undergoing frequent changes in tract length by base additions or deletions due to the DNA polymerase slippage during replication. In order to study microsatellite instability in Arabidopsis thaliana , a novel reporter system was developed. Several constructs containing out-of-frame mononucleotide repeats (G7, G10, G13, G16 and C16) were introduced in the β-glucuronidase ( GUS ) reporter gene. Frameshift mutations occurring in these DNA repeats restore GUS function and produce blue sectors on an otherwise white background. The frequency of blue sectors was used to measure the degree of instability of various microsatellites in the soma, and to follow the germinal transmission of novel alleles. Our results revealed that microsatellite instability increases with the length of repeats and depends on repeat unit orientation. We conclude that our GUS reporter system is a powerful tool to study microsatellite instability in plant genomes.

Maintenance of genetic information, structural integrity and stability of genomes requires precise duplication of DNA to prevent mutations. DNA replication is a complex mechanism, whose fidelity was estimated to be in the range of one error per 1010 nucleotides synthesized (Drake et al ., 1998). Such a low mutation rate is achieved by the sequential operation of three mechanisms: the high selectivity of DNA polymerase, exonucleolytic proofreading, and post-replicative mismatch repair (Kunkel, 1992; Schaaper, 1993; Kroutil and Kunkel, 1998). Several studies have measured mutation frequencies in model systems ( E. coli , S. cerevisiae , cultures of mammalian cells or transgenic mice). A better knowledge of genome-maintenance functions in a plant model system, Arabidopsis thaliana , is of great importance for two major reasons. First, plants are unable to move and are continuously exposed to environmental mutagens, hence they need an efficient system to maintain genome stability. Second, plants lack a reserved germ line. Their gametes are produced from meristematic cells that undergo many somatic divisions. Thus mutations have opportunities to be fixed and can be passed on to subsequent generations (for review, see Hays, 2002).

Mutation rates are much greater in DNA repeats (such as microsatellites) than in non-repetitive DNA sequences (Sia et al., 1997a). Microsatellites are short, tandemly repeated DNA sequences comprised of 1-6 bp per repeating unit. They represent unstable regions of the genome, undergoing mutational changes, usually additions or deletions of integral numbers of repeats. Microsatellites are interspersed throughout eukaryotic genomes and are highly polymorphic in length (Reviewed in Weber, 1990). Their polymorphic nature has made microsatellites useful for genetic mapping (Dib et al ., 1996), studying genomic instability in human diseases (Sutherland and Richards, 1995), and in population genetics (Djian, 1998). Alterations in DNA repeats, particularly in microsatellites, are attributed to either unequal recombination or DNA polymerase slippage (Jinks-Robertson at al ., 1998). Unequal recombination involves the pairing of repetitive tracts on homologous chromosomes in a misaligned configuration. In the DNA polymerase slippage model, a transient dissociation of the nascent and template strands during replication is followed by a misaligned reassociation of the strands leading to unpaired repeats. If the distortion caused by the unpaired repeat is not repaired, this mechanism will lead to either a gain (if the loop is in the nascent strand) or loss (if the loop is on the template strand) of repeat units (reviewed in Umar and Kunkel, 1996). These types of alterations are called frameshift mutations since they create a shift in reading frame when the microsatellite is located within a coding region. In fact, frameshift mutations within the coding region of a gene disturb the reading frame so that the entire set of triplets downstream of the modification is altered. Several studies have demonstrated that tandemly repeated sequences such as mono- and dinucleotide repeats are hotspots for frameshift mutations (see Jinks-Robertson et al ., 1998 for a review of yeast studies).

There are a number of properties of microsatellites that may affect their mutation rate: (i) the length of the repeat unit (mononucleotide vs dinucleotide), (ii) the base composition of the repeat unit, (iii) the number of repeats per tract, (iv) the composition of flanking sequences (v) the degree of perfection of the tract (i.e. presence or absence of interruptions in the tract), (vi) the level of transcription of tracts, and (vii) the orientation of the microsatellite with respect to the replication origin. Several of these parameters have been examined. For instance, mononucleotides are more unstable than dinucleotides which are, in turn, more unstable than trinucleotides, etc. Studies in E. coli (Bichara et al ., 2000), yeast (Henderson and Petes, 1992; Sia et al ., 1997b) and mammalian cells (Boyer et al., 2002) have extended and confirmed these observations. Base composition can also influence frameshift mutation; G or C runs were revealed more unstable than A or T runs (Harfe and Jinks-Robertson, 2000; Boyer et al ., 2002). The length of a repetitive tract affects microsatellite instability in many organisms (Wierdl et al., 1997; Yamada et al ., 2002). The frequency of DNA polymerase slippage in repeat sequences is positively correlated with the length of the tracts (Kroutil et al ., 1996; Tran et al ., 1997; Leonard et al ., 2003). Moreover, the composition of flanking sequences has been reported to affect frameshift mutations in mononucleotide runs in yeast (Harfe and Jinks-Robertson, 2000). Furthermore, the purity of the tract can alter microsatellite instability; variant bases within repetitive tracts lead to increased stability in mammalian cells (Chong et al ., 1995). In addition, elevated levels of transcription result in microsatellite destabilization by increasing DNA polymerase slippage (Wierdl et al ., 1996). Finally, the stability of some triplet repeat microsatellites was shown to be affected by the orientation of the microsatellite with respect to the replication origin (Kang et al ., 1995; Maurer et al ., 1996), although some dinucleotide microsatellites are not affected by sequence orientation (Henderson and Petes, 1992).

In much of the work described above, powerful reporter systems were used to monitor mutation rates. In plants, Kovalchuk et al. (2000) were the first to exploit such a strategy. They used the GUS gene to measure somatic point mutation frequency in transgenic Arabidopsis. The GUS reporter gene is a powerful tool for the assessment of gene activity in transgenic plants. The stability of the protein encoded, the simplicity and the sensitivity of the assays, and the variety of substrates available make GUS an attractive reporter system for plant studies (Jefferson, 1989). Recently, to measure frameshift mutations in microsatellites, a novel system was developed by Leonard et al . (2003) in the course of their work aimed at the phenotypic characterization of an Arabidopsis mismatch repair gene. In the work reported here, we have developed and validated a similar reporter system based on the GUS gene in order to measure frameshift mutations in Arabidopsis plants carrying simple repeats that vary either in length or in orientation of tracts. We show that our reporter system is a powerful tool to study microsatellite instability in plant genomes.

Modified GUS genes containing one of five synthetic microsatellites (G7, G10, G13, G16 and C16) were created. The G16 and C16 tracts were obtained by annealing oligonucleotides G16F (5′-CCCTAGGGGGGGGGGGGGGGGTACCTG-3′) with its complementary strand G16R (5′-CAGGTACCCCCCCCCCCCCCCCTAGGG-3′). The annealed product was cloned into the Msc I restriction site of the GUS gene in pBI221 (Clontech, BD Biosciences, Palo Alto, CA, USA), 620 bp downstream of the ATG (Figure 2.1A). Depending on the orientation of insertion of the synthetic microsatellite, either a G16 or a C16 tract was introduced in the coding strand. Each run contained an Avr II restriction site at one end and a Kpn I site at the other end. In both cases, filling in the Avr II restriction site using DNA polymerase (Klenow fragment) generated a stop codon a few codons downstream of the introduced microsatellite. These derivative constructs were named G16/stop and C16/stop. The G16/stop construct was then used to generate a G7/stop construct by PCR mutagenesis. PCR amplification of the G16/stop sequence was done using forward primer 35S (5′-GGAAACCTCCTCGGATTCCATTGC-3′) and reverse primer G7R (5′-CAGGTACCCCCCCTAGGG-3′). The resulting amplicon (998 pb) carrying a G7 tract was digested with Kpn I and Xba I and substituted for the corresponding region of the G16/stop construct. We generated G10/stop and G13/stop constructs in the same way, using specific mutagenic primers in each case. A 2.4 Kb Hin dIII -Eco RI fragment containing the modified GUS gene, the cauliflower mosaic virus (CaMV) 35S promoter and the NOS terminator were then transferred into either pBI121 (Clontech, BD Biosciences, Palo Alto, CA, USA) or pCAMBIA 1300 (CAMBIA, Canberra, Australia) binary vectors.

Figure 2.1: GUS reporter system to study frameshift mutations . ( A ) Structure of the T-DNA containing a microsatellite-GUS construct. A Hind III /EcoR I fragment containing a modified GUS gene (containing one of five different synthetic microsatellites) was introduced into the pBI121 binary vector. GUS: ß-glucuronidase gene, 35S: cauliflower mosaic virus 35S promoter, NPT II: Neomycinephosphotransferase II, NOS P: nopaline synthase promoter, NOS T: nopaline synthase terminator, RB: right border, LB: left border. ( B ) Detailed view of the synthetic microsatellite inserted into the GUS coding region. The stretch of guanines (G7, G10, G13 or G16) or cytosines (C16; not illustrated) and flanking restriction sites ( Avr II and Kpn I; underlined) adds up to nine codons but does not disrupt the reading frame of the GUS gene. ( C ) After digestion and fill-in at the AvrII site, the coding sequence is out-of-frame and a stop codon interrupts the GUS protein (underlined). (D) Arabidopsis plants carrying the in-frame construct (G16) express GUS protein in all cells and appear as completely blue. ( E ) Plants transformed with an out-of-frame construct (G16/stop) were mostly white except for some blue sectors resulting from frameshift mutations restoring the activity of the reporter gene.

Histochemical staining was done as described by Jefferson et al . (1987). In brief, plants were vacuum infiltrated twice for 10 minutes in sterile staining buffer containing 0.2 mM 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) (Rose Scientific Ltd, Edmonton, Alberta, Canada), 100 mM phosphate buffer (pH 7.0), 0.2% Triton X-100, and 50 mM kanamycin. Afterwards, plants were incubated at 37°C for 48h. Following incubation, plants were cleared by several changes of 70% ethanol to completely eliminate chlorophyll. For plants carrying G7, G10 or G13 runs, the entire rosette of 20-30 plants (4 to 5 weeks-old) was stained. The number of leaves and the total number of blue sectors were counted for each plant using a binocular microscope. For plants containing C16 runs, we stained 24 plants grown in vitro (15-20 days after germination) and sectors were counted on the first three true leaves. For plants carrying G16 runs we stained both plants with a full-rosette grown in soil and in vitro plants. Typically, the scoring was done on T3 or T4 progeny homozygous for the reporter constructs at a single locus. The reversion frequency was expressed as the average number of spots per leaf. To measure the germinal reversion frequency we stained large numbers of in vitro grown plants. The parents of this population were previously tested for the absence of germinal revertant allele to ensure that the germinal reversion events detected had occurred in the generation immediately preceding that under study.

All five out-of-frame constructs (harbouring either a G7, G10, G13, G16 or C16 tract) were transformed into Arabidopsis thaliana ecotype Columbia. Independent T1 transformants were selected and T2 progeny were plated on selective media to identify single locus insertions (lines segregating 3R:1S) for further analysis. T3 lines homozygous for the insertion locus were identified subsequently by progeny testing. A total of 29 lines were thus obtained for the set of out-of-frame constructs: G7/stop (7 lines), G10/stop (4 lines), G13/stop (8 lines), G16/stop (6 lines) and C16/stop (4 lines). All lines were validated by PCR amplification as harbouring the expected microsatellite using primers that border the repetitive tract (data not shown).

In a first set of analyses, we wished to examine the effect of repeat length on microsatellite instability. For this, we measured the reversion rate (frequency of blue sectors) in homozygous T3 progeny of four to eight independent transformants. As can be seen in Figure 2.2A, an increase in the average number of blue sectors per leaf was observed as the length of the repetitive tract increased: 0.5 (G7), 6.5 (G10), 12.2 (G13) and 40.5 (G16). The increase was particularly sharp between the G7 and G10 tracts, as there was thirteen-fold increase between the two. Further incremental additions of three Gs led to more modest increases, on the order of 2- to 3-fold. As shown in Figure 2.2A, G7 and G10 tracts were not significantly different according to Student’s T test, and the same was true of the G10 and G13 tracts. Due to the large interline variability in the G10 construct, a very broad confidence interval was obtained. However, G16 lines exhibited much less interline variation and these differed significantly from other lines.

Second, to study the effect of tract orientation on the reversion rate, we compared G16 and C16 runs (Figure 2.2B). These two microsatellites are identical in sequence as they result from the insertion of the same oligonucleotide duplex in each of the two possible orientations. It is with reference to the coding strand of the GUS gene that we define the G16 and C16 microsatellites. Six G16 and four C16 independent transformants grown in vitro were analysed for the number of blue spots per leaf in the T3 generation. The average number of sectors per leaf in C16 lines was 13.2, more than three-fold lower than in G16 lines (41.2). A statistical analysis indicated that the two sets of lines were significantly different (P < 0.01).

Interestingly, the amount of interline variation differed considerably from one construct to the other. G7 and G16 transformants exhibited very little interline variation whereas G10 and G13 were much more variable. Figure 2.3 shows the results for all eight G13 and six G16 lines, the two constructs that showed the sharpest contrast in this regard. Among G16 lines, the number of blue sectors per leaf was quite uniform as it ranged from 21 to 60 (C.V. = 37%). In contrast, among G13 transformants, a tremendous variation (C.V. = 140%) was seen as the average number of sectors per leaf ranged from as low as 0.3 all the way up to 44.6 sectors per leaf, almost 150 times more.

Figure 2.2 : Effect of repeat length and composition on microsatellite instability (A) Effect of repeat length on the instability of guanine runs. The reversion frequency (average number of blue sectors per leaf) was measured in homozygous T3 progeny of four to eight independent transformants carrying G7, G10, G13 or G16 tracts. Scoring of blue sectors was done on the entire rosette of 20-30 plants grown in soil for 4 to 5 weeks. Numbers on top of each column represent the average number of spots per leaf. Vertical bars show the limits of the 95% confidence interval around the mean. Means followed by the same letter are not significantly different at α = 0.05. (B) Effect of repeat composition on microsatellite instability. Scoring of blue sectors was done on in vitro grown plants from four independent T3 transformants (24 plants for each independent transformant 15-20 days after germination). Numbers on top of each column represent the average number of spots per leaf. Vertical bars show the limits of the 95% confidence interval around the mean. Means followed by the same letter are not significantly different at α = 0.05.

As shown in Figure 2.3, among the eight G13 lines only two (G13-C and G13-F) showed a single T-DNA insertion. These two lines had a low number of sectors per leaf (0.57 and 0.45 respectively) compared to the others. However the line with the lowest number of sectors per leaf (G13-D) showed two copies of the transgene. All other G13 lines were multicopy integrations (three copies at least). In G16 lines, among five independent transformants tested, four had a single-copy insertion (G16-A, G16-D, G16-F and G16-G) and one line (G16-C) had more than three copies (Figure 2.3). Lines that had inserted one transgene show more mutations than the line with a multicopy T-DNA.

Microsatellites are hot spots for frameshift mutations and they represent a sensitive indicator of genomic stability. To measure microsatellite instability in Arabidopsis plants we have developed a powerful and versatile reporter system that can be used to measure the frequency of both somatic and germinally-transmitted mutations. The results obtained led us to conclude that microsatellite instability is positively correlated with the length of the repeated tract, and that there is also a form of strand bias in Arabidopsis as previously reported in other systems.

We have used a reversion assay to measure mutation frequencies for four mononucleotide tracts differing only in the number of repeats. Our results showed a positive correlation between the length of the microsatellite and its instability. This finding is in agreement with earlier observations in other organisms where long repetitive tracts are less stable than short tracts. This relation between length and instability of microsatellites has been demonstrated in many organisms such as Escherichia coli (Bichara et al ., 2000), Saccharomyces cerevisiae (Sia et al ., 1997b; Wierdl et al ., 1997), Caenorhabditis elegans (Degtyareva et al ., 2002), mammalian cells (Yamada et al ., 2002) and more recently in plants (Leonard et al ., 2003).

When we compared the relative mutation frequencies of various microsatellites, we found that a G16 tract was 82-fold more unstable than a G7, a G13 microsatellite 28 times more unstable than a G7, and a G10 approximately 13-fold less stable than a G7 (Figure 2.2A). Rather than being linear, the increases in microsatellite instability was initially very steep (a first 3-bp increment resulting in a 13-fold increase) and then continued at a more moderate pace (the second and third 3-bp increments resulting in further two- and three-fold increases, respectively). We suggest the existence of a surveillance system, presumably the MMR system, which maintains a relative stability in long runs exceeding 10 repeats. The increased instability observed from one construct to the next (with increasing microsatellite length) could be due to the lack of fidelity of DNA polymerase. In vitro studies indicate that DNA polymerase slippage increases as repetitive tracts in the substrate become longer. This is due to the probability of reassociation of DNA strands in a misaligned configuration and a decrease in the efficiency of the proofreading exonuclease (Bebenek et al. , 1990; Kroutil et al ., 1996). Indeed, the effect of proofreading exonuclease was found to be much greater in short tracts compared to long tracts (Tran et al ., 1997). In a recent report, Leonard et al . (2003) reported similar data in Arabidopsis on microsatellites of the same length and base composition as those reported here. The average number of mutations per plant was, however, significantly lower than reported here (0.6, 1.2 and 11.2 mutations per plant for G7, G10 and G13 tracts, respectively). Much of this may be due to the fact that they counted sectors on 2-week-old plants, whereas we counted sectors on plants with fully-developed (12 leaves at least) rosettes on 4- or 5-week old plants grown on soil. In contrast, however, they noted only a modest (two-fold) increase in the instability of a G10 tract relative to a G7 tract whereas the next 3-bp step (G13) resulted in a 10-fold increase relative to a G10 tract. Tran et al . (1997) observed a progression in instability very similar to the one we observed on microsatellites of the same or nearly the same length. Indeed, they reported that an A10 microsatellite was five-fold more unstable than an A7 and that an A14 tract was three-fold more unstable than A10.

In the present report, lines carrying C16 runs were three-fold more stable than those harbouring G16 runs. This is quite perplexing as both microsatellites are essentially the same except for the orientation in which insertion occurred. A previous study in yeast revealed almost the same difference in instability between C10 and G10 tracts; the mutation rate was 3.4 x 10-6 in C10 runs and 10.5 x 10-6 in G10 runs (Gragg et al ., 2002). It has been proposed that such differences in mutation frequency could be attributed either to transcription-coupled repair (TCR) mechanisms or to the unequal fidelity of leading- versus lagging-strand DNA replication (Gragg et al ., 2002). Data from E. coli have indicated that the slippage rate in mononucleotide runs may differ during leading- versus lagging-strand synthesis (Gawel et al ., 2002). Due to their antiparallel nature the two strands are replicated in different fashion and do not produce mutations at the same rate. Since lagging-strand synthesis is less processive than leading-strand synthesis, more frameshift errors might be generated during lagging-strand synthesis (Fijalkowska et al ., 1998; Kunkel and Bebenek, 2000). It does not seem likely, however, that such a hypothesis can explain our results. For this to be the case, one would have to assume that, in many independent transformants, a T-DNA carrying a G16 construct would have systematically inserted in such a fashion that the Gs were found on the lagging strand and conversely for the C16 construct. Although we cannot exclude this possibility absolutely, it would seem quite remote. Similarly, in the case of transcription-coupled repair, as both the G16 and C16 stretches are located on the coding strand in the two different constructs, this should not result in a systematic bias. As neither of the previously proposed hypotheses seems likely, more work will need to be done to resolve this apparent «mystery»!

Unexpectedly, the mutation frequency occasionally varied dramatically between different lines harbouring the same construct, as exemplified by the G13 lines (Figure 2.3). In other cases, a more limited variation was seen, as in the case of lines carrying G16 or C16 microsatellites. Mutation frequency was expected to depend on the copy number of the transgene because the higher the transgene copy number, the more possible targets exist for reversion to the original phenotype (Kovalchuk et al ., 2000). As shown in Figure 2.3, G13 lines exhibiting fewer sectors per leaf (G13-C, G13-D and G13-F) were carrying one or two copies of T-DNA. However, an insertion of three copies of the transgene or more occurred in lines with higher number of sectors per leaf. For the G13 construct, it seems that the number of sectors per leaf was correlated to the number of T-DNAs integrated. Therefore, the heterogeneity in the copy number of the transgene could explain the high variability observed in G13 lines. In G16 lines, however, the correlation between the copy number of the transgene and the mutation frequency is less evident. Indeed, the multicopy line (G16-C) showed the lowest mutation rate, which suggests that a higher copy number could have influenced the frequency of mutation negatively, perhaps by triggering gene silencing. Kovalchuk et al . (2000) reported similar results when studying the relationship between the copy number of the transgene and the reversion frequency in a different reporter construct.

Other factors, apart from copy number of the transgene, such as location of the integrated gene or the level of transgene transcription can also influence the mutation rate. Since transgenes generally integrate in different chromosomal sites, the mutation frequency may be a function of the chromosomal position of the transgene and may reflect different structural or dynamic properties of the chromatin at the locus of the transgene (Puchta et al ., 1995). Thus, each line obtained in this study is the product of an independent T-DNA-insertion event. The transcription level of transgene can also influence frameshift mutations within repeats (transcription-coupled repair). Highly transcribed regions may be repaired more efficiently than those with a low transcription level, leading to lower mutation rates. Turker et al. (1993) found a four-fold reduction in mutation rate in the transcribed mouse adenine phosphoribosyltransferase gene compared to a non-transcribed downstream region.

Unlike animals, plants do not have a specialized cell line predetermined to produce the gametes. In a plant like Arabidopsis, with numerous floral meristems, there is ample opportunity for a mutation arising during somatic development to be transmitted germinally. One would expect that a higher mutation rate in somatic tissues would result in a higher frequency of germinally-transmitted mutations. This would appear to be the case as germinal revertants were only observed in the progeny of lines carrying the G16 construct, those in which the somatic mutation rate was the highest. No germinal revertants were detected in lines carrying the G7, G10 and G13 constructs, presumably because the greater stability of these microsatellites lowered the rate of germinally-transmitted mutations below the detection limit.

In conclusion, the data reported in the present study describe a powerful tool to measure microsatellite instability in Arabidopsis. It is straightforward to score hundreds or thousands of plants at the same time to determine mutation frequency by scoring blue sectors following histochemical staining. The insertion of almost 30 bp (the longest insert in this study) in GUS gene did not disturb its expression hence the robustness of the GUS gene. Our GUS reporter system allowed us to establish not only a relationship between tract length and frameshift mutation rate, but also a correlation between tract orientation and mutation frequency. This GUS system will now serve as a useful tool for the study of repair mechanisms in plants.

Acknowledgements

This work was supported by a research grant from RhoBio to F. J. Belzile.

Bebenek K, Kunkel TA (1990) Frameshift errors initiated by nucleotide misincorporation. Proc Natl Acad Sci USA 87: 4946-4950

Bichara M, Pinet I, Schumacher S and Fuchs R, (2000) Mechanism of dinucleotide repeat instability in Escherichia coli . Genetics 154: 533-542

Boyer JC, Yamada N, Roques N, Hatch SB, Riess K, Farber RA, (2002) Sequence dependent instability of mononucleotide microsatellites in cultures mismatch repair proficient and deficient mammalian cells. Hum Mol Genet 11: 707-713

Chong SS, McCall A, Cota J, Subramony SH, Orr HAT, Hughes MR, Zoghbi HY (1995) Gametic and somatic tissue-specific heterogeneity of the expanded SCA1CAG repeat in spinocerebellar ataxia type 1. Nat Genet 10: 344-350

Clough SJ, Bent AF (1998) Floral dip : a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana . Plant J 16: 735-743

Degtyareva NP, Greenwell P, Hofmann ER, Hengartner MO, Zhang L, Culotti JG, Petes TD, (2002) Caenorhabditis elegans DNA mismatch repair gene msh-2 is required for microsatellite stability and maintenance of genome integrity. Proc Natl Acad Sci USA 99: 2158-2163

Dib C, Fauré S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc S, Hazan J, Seboun E, Lathrop M, Gyapay G, Morissette J, Weissenbach J (1996) A comprehensive genetic map of the human genome based on 5,264 microsatellites. Nature 380: 152-154

Djian P (1998) Evolution of simple repeats in DNA and their relation to human disease. Cell 94: 155-160

Drake JW, Charlesworth B, Charlesworth D, Crow JF (1998) Rates of spontaneous mutation. Genetics

Dubois P, Cutler S, Belzile FJ (1998) Regional insertional mutagenesis on chromosome III of Arabidopsis thaliana using the maize Ac element. Plant J 13: 141-151

Feng Z, Hu W, Komissarova E, Pao A, Hung M-C, Adair GM, Tang M-S (2002) Transcription-coupled DNA repair is genomic context dependent. J Biol Chem 277: 12777-12783

Fijalkowska IJ, Jonczyk P, Maliszewska Tkaczyk M, Bialoskorska M, Schaaper RM (1998) Unequal fidelity of leading strand and lagging strand DNA replication on the Escherichia coli chromosome. Proc Natl Acad Sci USA 95: 10020-10025

Gawel D, Jonczyk P, Bialoskorska M, Schaaper R, Fijalkowska IJ (2002) Asymmetry of frameshift mutagenesis during leading and lagging-strand replication in Escherichia coli . Mutation Res 501: 129-136

Gragg H, Harfe B, Jinks-Robertson S (2002) Base composition of mononucleotide runs affects DNA polymerase slippage and removal of frameshift intermediates by mismatch repair in Saccharomyces cerevisiae . Mol Cell Biol 22: 8756-8762

Harfe BD, Jinks-Robertson S (2000a) Sequence composition and context effects on the generation and repair of frameshift intermediates in mononucleotide runs in Saccharomyces cerevisae . Genetics 156: 571-578

Hays JB (2002) Arabidopsis thaliana , a versatile model system for study of eukaryotic genome-maintenance functions. DNA repair 1: 579-600

Henderson ST, Petes T (1992) Instability of simple sequence DNA in Saccharomyces cerevisiae . Mol Cell Biol 12: 2749-2757

Jefferson RA, Kavanagh T, Bevan MW, (1987) GUS fusions: β-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J 6: 3901-3907

Jefferson RA (1989) The GUS reporter gene system. Nature 342: 837-838

Jinks-Robertson S, Greene C, Chen W, (1998) Genetic instabilities in yeast. In Genetic Instabilities and Hereditary Neurological Diseases. WRaWSA Press, ed, San Diego, pp 485-507

Kang S, Jaworski A, Ohshima K, Wells RD, (1995) Expansion and deletion of CTG repeats from human disease genes are determined by the direction of replication in E. coli . Nature Genet 10: 213-218

Kovalchuk I, Kovalchuk O, Hohn B (2000) Genome-wide variation of the somatic mutation frequency in transgenic plants. EMBO J 19: 4431-4438

Kroutil LC, Kunkel T (1998) DNA replication errors involving strand misalignments. In WRaW ST, ed, Genetic instabilities and hereditary neurological diseases. Academic Press, san Diego, pp 699-716

Kroutil LC, Register K, Bebenek T, Kunkel TA (1996) Exonucleolytic proofreading during replication of repetitive DNA. Biochemistry 35: 1046-1053

Kunkel TA (1992) DNA replication fidelity. J Biol Chem 267: 18251-18254

Kunkel TA, Bebenek K (2000) DNA replication fidelity. Annu Rev Biochem 69: 497-529

Leonard JM, Bollmann SR, Hays JB (2003) Reduction of stability of Arabidopsis genomic and transgenic DNA-repeat sequences (microsatellites) by inactivation of AtMSH2 mismatch-repair function. Plant Physiol 133: 328-338

Maurer DJ, O'Callaghan BL, Livingston DM (1996) Orientation dependence of trinucleotide CAG repeat instability in yeast. Mol Cell Biol 16: 6617-6622

Puchta H, Swoboda P, Gal S, Blot M, Hohn B (1995) Somatic intrachromosomal homologous recombination events in populations of plant siblings. Plant Mol Biol 28: 281-292

Schaaper RM (1993) Base selection, proofreading, and mismatch repair during DNA replication in Escherichia coli . J Biol Chem 268: 23762-23765

Sia EA, Jinks-Robertson S, Petes TD (1997a) Genetic control of microsatellite stability. Mutation Res 383: 61-70

Sia EA, Kokoska R, Dominska M, Greenwell P, Petes TD (1997b) Microsatellite instability in yeast: dependance on repeat unit size and DNA mismatch repair genes. Mol Cell Biol 17: 2851-2858

Sutherland GR, Richards R (1995) Simple tandem DNA repeats and human genetic disease. Proc Natl Acad Sci USA 92: 3636-3641

Tran HT, Keen J, Kricker M, Resnick MA, Gordenin DA (1997) Hypermutability of homonucleotide runs in mismatch repair DNA polymerase proofreading yeast mutants. Mol Cell Biol 17: 2859-2865

Turker MS, Cooper G, Bishop PL (1993) Region-specific rates of molecular evolution: a fourflod reduction in the rate of accumulation of silent mutations in transcribed versus nontranscribed regions of homologous DNA fragments derived from two closely related mouse species. J Mol Evol 36: 31-40

Umar A, Kunkel T (1996) DNA-replication fidelity, mismatch repair and genome instability in cancer cells. Eur J Biochem 192: 563-676

Velvekens D, Van Montagu M, Van Lijsebettens M (1998) Agrobacterium tumefaciens -mediated transformation of Arabidopsis thaliana root explants by using kanamycin selection. Proc Natl Acad Sci USA 85: 5536-5540

Weber JL (1990) Human DNA polymorphisms and methods of analysis. Curr Opin Biotechnol 1: 166-171

Wierdl M, Greene C, Datta A, Jinks-Robertson S, Petes TD (1996) Destabilization of simple repetitive DNA sequences by transcription in yeast. Genetics 143: 713-721

Wierdl M, Dominska M, Petes TD (1997) Microsatellite instability in yeast: Dependence on the length of the microsatellite. Genetics 146: 769-779

Yamada NA, Smith G, Castro A, Roques CN, Boyer JC, Farber RA (2002) Relative rates of insertion and deletion mutations in dinucleotide repeats of various length in mismatch repair proficient mouse and mismatch repair deficient human cells. Mutation Res 499: 213-225