The Plant Journal (2015) 82, 93–104

doi: 10.1111/tpj.12793

Comprehensive identification of mutations induced by heavy-ion beam irradiation in Arabidopsis thaliana Tomonari Hirano1,2,†,‡, Yusuke Kazama1,2,†, Kotaro Ishii2, Sumie Ohbu2, Yuki Shirakawa2 and Tomoko Abe1,2,* 1 Innovation Center, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan, and 2 Nishina Center for Accelerator-Based Science, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan Received 30 October 2014; revised 25 January 2015; accepted 5 February 2015; published online 17 February 2015. *For correspondence: (e-mail [email protected]). † These authors contributed equally to this study. Accession numbers: Next-generation sequencing data used in this study are available in the DDBJ Sequenced Read Archive (http://trace.ddbj.nig.ac.jp/dra/ index_e.html) under accession numbers DRA002449, DRA002450, and DRA002451. ‡ Present address: Faculty of Agriculture, University of Miyazaki, 1-1 Gakuen Kibanadai-nishi, Miyazaki 889-2192, Japan.

SUMMARY Heavy-ion beams are widely used for mutation breeding and molecular biology. Although the mutagenic effects of heavy-ion beam irradiation have been characterized by sequence analysis of some restricted chromosomal regions or loci, there have been no evaluations at the whole-genome level or of the detailed genomic rearrangements in the mutant genomes. In this study, using array comparative genomic hybridization (array-CGH) and resequencing, we comprehensively characterized the mutations in Arabidopsis thaliana genomes irradiated with Ar or Fe ions. We subsequently used this information to investigate the mutagenic effects of the heavy-ion beams. Array-CGH demonstrated that the average number of deleted areas per genome were 1.9 and 3.7 following Ar-ion and Fe-ion irradiation, respectively, with deletion sizes ranging from 149 to 602 180 bp; 81% of the deletions were accompanied by genomic rearrangements. To provide a further detailed analysis, the genomes of the mutants induced by Ar-ion beam irradiation were resequenced, and total mutations, including base substitutions, duplications, in/dels, inversions, and translocations, were detected using three algorithms. All three resequenced mutants had genomic rearrangements. Of the 22 DNA fragments that contributed to the rearrangements, 19 fragments were responsible for the intrachromosomal rearrangements, and multiple rearrangements were formed in the localized regions of the chromosomes. The interchromosomal rearrangements were detected in the multiply rearranged regions. These results indicate that the heavy-ion beams led to clustered DNA damage in the chromosome, and that they have great potential to induce complicated intrachromosomal rearrangements. Heavy-ion beams will prove useful as unique mutagens for plant breeding and the establishment of mutant lines. Keywords: heavy-ion beam, mutation, genomic rearrangement, resequencing, Arabidopsis thaliana.

INTRODUCTION Ionizing radiation has long been used as a mutagen for genetic analysis and the investigation of gene functions as well as for plant breeding. Heavy-ion beams, a form of ionizing radiation, have been accepted as an effective mutagen (Tanaka et al., 2010; Abe et al., 2012b). They can induce mutations with high frequency at relatively low doses of radiation and the mutants arising in the irradiated generation often become new cultivars directly. Thus, a small number of genes are expected to be disrupted in the mutant genomes. Such characteristics of heavy-ion irradiation are advantageous for mutation breeding. More than 40 plant cultivars, including torenia (Miyazaki et al., 2006) © 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd

and verbena (Kanaya et al., 2008), have been bred with the aid of heavy-ion beams in Japan. The effectiveness of heavy-ion irradiation may be due to the its linear energy transfer (LET; keV lm 1) value. The LET represents the amount of energy deposited locally by radiation. The LETs of c-rays and X-rays are low and constant (0.2 keV lm 1 for c-rays and 2.0–5.0 keV lm 1 for X-rays). In contrast, the LET of a heavy-ion beam is variable and higher than those of c-rays and X-rays. For example, in the RIKEN RI-beam factory, LETs for use in biological research range from 22.5 to 4000 keV lm 1 (Ryuto et al., 2008). As the energy from heavy-ion beams with 93

94 Tomonari Hirano et al. high LETs is deposited more densely on the target than is the energy from c-rays and X-rays, the beams are efficient in causing double-strand breaks on DNA molecules. We have demonstrated the effect of LET values on both mutation frequency and the molecular nature of induced mutations through studies using Arabidopsis thaliana. A heavyion beam with a LET of 30 keV lm 1 showed the highest mutation frequency in the M2 generation (Kazama et al., 2008). Polymerase chain reaction and sequence analyses of the responsible genes in the heavy-ion beam-induced mutants revealed that 30.0 keV lm 1 irradiation primarily induced small deletions (1–51 bp) (Kazama et al., 2011); 30.0 keV lm 1 irradiation also resulted in a high mutation frequency in the M1 generation (Kazama et al., 2012). This highly efficient LET was designated as LETmax. By contrast, a LET of 290 keV lm 1 resulted in a lower mutation frequency; instead, the proportion of large deletions (>1 kbp) and chromosomal rearrangements was increased (Hirano et al., 2012). The molecular nature of induced mutations is an important factor in selecting the appropriate mutation induction techniques for analysis of gene function. Widely used mutation induction techniques include: ethyl methanesulfonate (EMS) mutagenesis; T-DNA insertion; transposon tagging; and ionizing radiation. Ethyl methanesulfonate mutagenesis predominantly induces point mutations. The insertion of T-DNA (transposon) elements disrupts the coding or intronic sequence of a gene, resulting in the production of ‘leaky’ or knock-out alleles. However, there is an integration bias for the localization of T-DNA insertion sites (Alonso et al., 2003). Therefore, it may be difficult to generate knock-out alleles covering all loci by construction of T-DNA insertion or EMS-mutagenized lines. By contrast, ionizing radiation has been used for the induction of physical deletions in the genome, resulting in a high proportion of knock-out mutations (Sato et al., 2006). Moreover, for heavy-ion irradiation, the size of the induced deletion can be pre-selected as required using LET levels as described above. We have also previously reported that heavy-ion irradiation with LETmax can, in addition, be utilized for reverse genetics, by combining it with single-nucleotide polymorphism (SNP) detection methodologies such as a high-resolution melting curve (Kazama et al., 2011). The large deletions induced by high-LET heavy-ion irradiation would be useful for the disruption of whole genes as well as of multiple, tandemly arrayed genes. An effective technique for the rapid determination of DNA copy number variation (CNV) in plants is arraycomparative genomic hybridization (array-CGH) (Nagano et al., 2008; Yu et al., 2011; Belfield et al., 2014), by which we identified seven deletions and one candidate duplication in an Fe-ion-induced mutant (Kazama et al., 2013). To characterize the molecular nature of induced mutations in greater detail, researchers have applied next-generation

sequencing (NGS) technology to plants (Uchida et al., 2011; Belfield et al., 2012; O’Rourke et al., 2013). Mutations induced by EMS were found to consist mainly of G/C–A/T transitions even upon characterization by NGS (Uchida et al., 2011). By contrast, NGS analysis of six independent Arabidopsis mutants induced by fast-neutron irradiation (another form of ionizing radiation) revealed that single base substitutions were induced more frequently than deletion mutations, and that small deletions (100 kbp.

themselves spanned >100 kbp and had common characteristics (Figure 1, Table S1), containing a significantly small number of protein-coding genes but a large number of transposable elements (v2 test, P < 0.01) in comparison with the average numbers of these on the respective chromosomes of A. thaliana, in which the densities (loci/ 100 kbp) of protein-coding genes and transposable elements are 21.3 and 3.4 (chromosome 3), and 20.2 and 2.6 (chromosome 5), respectively (The Arabidopsis Genome Initiative, 2000). In addition to deletions, candidates for regional duplication were detected on Ar-365-as1 chromosome 4 (1 516 305–2 087 883) and on Ar-443-as1 chromosome 2 (6 966 364–9 222 822 and 9 822 729–10 140 806) by array-CGH (Figure 1). The results obtained by arrayCGH indicate that Ar- or Fe-ion beam irradiation can induce large deletions, consistent with findings from previous, PCR-based, analyses (Hirano et al., 2012; Kazama et al., 2013). We selected the three mutants derived from Ar-ion irradiation for detailed analysis of the induced mutations by resequencing. The selected mutants showed differing

mutation characteristics as detected by array-CGH: the mutant lines Ar-57-al1, Ar-365-as1, and Ar-443-as1 showed no large deletion; deletions between several hundred bp and several kbp with genomic rearrangements; and a 600-kbp deletion with genomic rearrangement, respectively (Table 1). The results of resequencing the mutant genomes are shown in Table S2. More than 99% of the genomes were covered at a depth of 109. From the resequencing data, candidate mutations were detected using the bioinformatics algorithms provided by SAMTOOLS (Li et al., 2009), PINDEL (Ye et al., 2009), and BREAKDANCER (Chen et al., 2009) (Table S3). Although these results also included candidate mutations detected in the mitochondrial and chloroplast genomes, we focused on the candidates detected in the nuclear genome as the targets for analysis in this study. Background mutations harbored in the A. thaliana ecotype Col-0 kept in our laboratory, as well as known sequencing or mapping errors, were also excluded (see Experimental Procedures). Using SAMTOOLS, we were able to detect small mutations including base substitutions, deletions ranging from 1 to 33 bp, and

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

96 Tomonari Hirano et al. Figure 1. The mutated regions comprising multiple copy number variation peaks as detected by array comparative genomic hybridization. The plots of all mutants were merged into one graph. The vertical axis shows the log2 ratio, and the large positive and negative values indicate deletion (Del.) and duplication (Dupli.), respectively. The horizontal axis indicates the positions of the probes in each chromosome (Chr.). The deleted and duplicated regions are represented by the color indicated in the legend above the figure. These regions are made up of more than 10 peaks, and their overall sizes were >100 kbp.

themselves spanned >100 kbp and had common characteristics (Figure 1, Table S1), containing a significantly small number of protein-coding genes but a large number of transposable elements (v2 test, P < 0.01) in comparison with the average numbers of these on the respective chromosomes of A. thaliana, in which the densities (loci/ 100 kbp) of protein-coding genes and transposable elements are 21.3 and 3.4 (chromosome 3), and 20.2 and 2.6 (chromosome 5), respectively (The Arabidopsis Genome Initiative, 2000). In addition to deletions, candidates for regional duplication were detected on Ar-365-as1 chromosome 4 (1 516 305–2 087 883) and on Ar-443-as1 chromosome 2 (6 966 364–9 222 822 and 9 822 729–10 140 806) by array-CGH (Figure 1). The results obtained by arrayCGH indicate that Ar- or Fe-ion beam irradiation can induce large deletions, consistent with findings from previous, PCR-based, analyses (Hirano et al., 2012; Kazama et al., 2013). We selected the three mutants derived from Ar-ion irradiation for detailed analysis of the induced mutations by resequencing. The selected mutants showed differing

mutation characteristics as detected by array-CGH: the mutant lines Ar-57-al1, Ar-365-as1, and Ar-443-as1 showed no large deletion; deletions between several hundred bp and several kbp with genomic rearrangements; and a 600-kbp deletion with genomic rearrangement, respectively (Table 1). The results of resequencing the mutant genomes are shown in Table S2. More than 99% of the genomes were covered at a depth of 109. From the resequencing data, candidate mutations were detected using the bioinformatics algorithms provided by SAMTOOLS (Li et al., 2009), PINDEL (Ye et al., 2009), and BREAKDANCER (Chen et al., 2009) (Table S3). Although these results also included candidate mutations detected in the mitochondrial and chloroplast genomes, we focused on the candidates detected in the nuclear genome as the targets for analysis in this study. Background mutations harbored in the A. thaliana ecotype Col-0 kept in our laboratory, as well as known sequencing or mapping errors, were also excluded (see Experimental Procedures). Using SAMTOOLS, we were able to detect small mutations including base substitutions, deletions ranging from 1 to 33 bp, and

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

98 Tomonari Hirano et al.

(a)

(b)

(c)

Figure 2. Schematic representation of genomic rearrangements caused by Ar-ion beam irradiation. The genomic rearrangements found in Ar-57-al1 (a), Ar-365-as1 (b), and Ar-443-as1 (c) are shown. The zygosity of each mutation in the M2 generation is estimated from the genetic homogeneity of the mutations in the resequencing results from pooled DNA from 40 M3 plants in each line. The rearranged fragments are shown as boxes, with sequence direction indicated. The boxes for the duplicated fragments are marked with a diagonal line pattern. Genomic positions, based on TAIR10, are given for each breakpoint, and the assigned numbers of the rearrangement junctions are in red. For Ar-443-as1 chromosome (Chr.) 3, heterozygous deletions in the M2 generation are illustrated with paired chromosomes, and colored dashed lines indicate areas of homozygous (black) or heterozygous (grey) deletion.

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

Whole-genome analysis of heavy-ion beam mutations 99 Table 3 Numbers of mutated genes and mutations expected to affect gene function in the three mutant lines Ar-57-al1

Missense Nonsense Frameshift Truncationa Amino acid insertion Loss of whole gene Mutated genes/genomeb

Ar-365-as1

Ar-443-as1

Homozygous

Heterozygous

Homozygous

Heterozygous

Homozygous

Heterozygous

6 0 0 3 0 0 7

0 1 1 2 0 0 4

1 0 0 5 0 0 5

3 1 4 13 0 1 19

0 0 0 0 0 11 11

5 2 7 7 1 74 94

The zygosity in each line is estimated from the genetic homogeneity in the M3 population. Truncations included amino acid deletions without a frameshift in the genes; and genes split by genomic rearrangements such as inversions and translocations. b When multiple mutations occurred in one gene, we counted those as a single gene mutation event. a

The size distributions of the simply rejoined deletions, which exclude the deletions at breakpoints of the rearrangement, and the rearranged fragments (≥100 bp), are shown in Figure 3. The maximum fragment size in the Ar-365-as1 mutant was 2.7 Mbp; over half of the fragments were within the several kbp to 1 Mbp size range. Although the simple deletions were fewer than the rearranged fragments (Figure 3), the extremely large deletions (1.2 Mbp being the maximum size in Ar-443-as1) were detected by the resequencing (Table S4). The simple deletions, ranging in size from 3 to 354 kbp, were also detected by array-CGH in the mutants with Ar-ion beam-induced mutations (Table 1). Seventy-six per cent of the breakpoints at the original sites of the rearranged fragments were coincident with deletions (Table 4). On the other hand, at the junctions involved in the rearrangements, regions of microhomology, ranging from 1 to 7 bp, were often used for rejoining breakpoints (62%; Table 4).

Figure 3. Size distribution of the simple deletions and rearranged fragments detected by genome resequencing of Ar-ion-induced mutants. The simple deletions did not include the deletions at breakpoints of the rearrangements. The fragments contributing to duplication, inversion, and translocation events were grouped according to size.

DISCUSSION The LET values of the Ar-ion (290 keV lm 1) and Fe-ion (640 keV lm 1) beams used in this study are much higher than those of c-rays (0.2 keV lm 1) or X-rays (2.0–5.0 keV lm 1). It has been suggested that the complexity and severity of DNA damage increases with the LET value (Hada and Georgakilas, 2008; Sage and Harrison, 2011). Correspondingly, in the Arabidopsis mutants, deletion sizes in the mutated genes increase with increasing LET value (Shikazono et al., 2005; Kazama et al., 2011; Hirano et al., 2012), and the Ar- and Fe-ion beams often induce large deletions (>several kbp) and genomic rearrangements (Hirano et al., 2012; Kazama et al., 2013). Array-CGH analysis determined that the average numbers of deletions (≥149 bp) in the mutant genomes derived from Ar- and Fe-ion irradiation were 1.9 and 3.7, respectively (Table 1; Kazama et al., 2013), and that 81% of the deletions were accompanied by genomic rearrangements (Table 1). These results indicated that irradiation with Arand Fe-ion beams has a large mutational effect on the plant genome, consistent with conclusions obtained from previous PCR-based research (Hirano et al., 2012; Kazama et al., 2013). It has been reported that in A. thaliana the majority of large deletions induced by radiation, such as those of several hundred kbp to several Mbp, were not transmitted to progeny; such non-transmissibility was considered to be due to the deletion of a particular region containing a gene or genes essential for gamete development or viability (Naito et al., 2005). The large deleted areas (>100 kbp) detected as transmissible mutations in this study were commonly characterized as containing few protein-coding genes (Table S1). This suggests that the possibility of the loss of genes essential for viability is greater for large deleted areas containing many protein-coding genes; thus, these areas tend to become non-transmissible mutations. A locational bias based on the chromosomal environment of the deletion, which might influence transmissibility, may also exist.

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

100 Tomonari Hirano et al. Table 4 Characteristics of breakpoints and junctions of rearranged fragments induced by Ar-ion irradiation No. of breakpoints (%)

No. of junctions (%)

Mutant line

+Deletion

–Deletion

+Microhomology

+Insertion of filler DNA

–Microhomology, –insertion

Ar-57-al1 Ar-365-as1 Ar-443-as1 Total

4 14 8 26 (76%)

1 6 1 8 (24%)

4 10 9 23 (62%)

1 3 0 4 (11%)

0 6 4 10 (27%)

Deletions located at the upstream and downstream of original sites of the rearranged fragments were tallied.

Mutant lines, whether induced by chemical mutagens, radiation, T-DNA insertion, or transposon tagging, are important genetic resources and have contributed to the progression of fundamental functional genomics in plants (Krishnan et al., 2009; Bolle et al., 2011). Recently, GABIDUPLO lines were established as mutant resources in A. thaliana, comprising double knock-out lines for two unlinked paralogous genes (Bolle et al., 2013). Genetic resources derived using such strategies will allow further progress in plant science via in-depth gene functional analysis. However, tandemly arrayed genes were not included for development as GABI-DUPLO lines because of the difficulty inherent in creating double mutants for tightly linked genes using crosses between single mutant lines. In A. thaliana, tandemly arrayed genes comprise over 10% of the genes in the genome (The Arabidopsis Genome Initiative, 2000; Rizzon et al., 2006). One method for knocking out genes in a tandem array is by the induction of a large deletion spanning the region of the gene loci (Morita et al., 2007). The appropriate deletion size for disruption of two tandemly arrayed genes is estimated to be around 5–10 kbp, according to the average gene density in A. thaliana (The Arabidopsis Genome Initiative, 2000). Therefore, the sizes of the large deletions induced by high-LET beams were deemed appropriate for the disruption of tandemly arrayed genes. Moreover, our array-CGH assay has previously been shown to be capable of detecting deletions with lengths of 200 bp and more, and deletions within this size range were indeed detected in the Ar- and Fe-ioninduced mutants by the array (Table 1; Kazama et al., 2013). It is thought that array-CGH is well-suited to efficient screening for mutants harboring large deletions. Therefore, lines with disruptions of tandemly arrayed genes could be established by mutation induction using high-LET beams, and mutants with large deletions screened using array-CGH. In this study we resequenced the mutant genomes derived from irradiation with Ar ions and attempted to achieve total mutation profiling of the genomes using three algorithms: SAMTOOLS, PINDEL, and BREAKDANCER. The sizes and types of mutations detected by each program complemented each other as described above, and all of

the detected break ends rejoined reasonably. All of the deleted and duplicated regions flagged by array-CGH were also detected by the resequencing-based method (Tables 1 and S4, Figure S4). These results indicate that our mutation detection platform using NGS data has the appropriate performance standards for detecting total mutations induced by radiation treatment. Although total mutation detection has also been reported in fast neutron-induced mutants, these mutants did not harbor genomic rearrangements (Belfield et al., 2012). In contrast, the mutants used for resequencing in this study provided useful test cases for the comprehensive detection of total mutations; they harbored SNPs and in/dels, as well as genomic rearrangements. For plant mutants induced by EMS, efficient programs or pipelines based on resequencing of bulked segregants in the F2 population have been developed for the identification of genetic mutations responsible for a mutant phenotype (Schneeberger et al., 2009; Austin et al., 2011; Uchida et al., 2011; Abe et al., 2012a). However, another method is required for radiation-induced mutants because of possibility of harboring genomic rearrangements. In this study, we aimed to detect total mutations in the mutants without backcrossing. Despite this restriction, candidate mutations for those responsible for our focused mutant phenotypes (Figure S1) were found in the mutated genes NARA5 (Ogawa et al., 2009) in Ar-57-al1; ASYMMETRIC LEAVES1 (Byrne et al., 2000) in Ar-365-as1, and CYP85A2 (Kim et al., 2005) in Ar-443-as1. For forward genetics strategies in radiation-induced mutants identification of responsible genes will be made even more rapidly and efficiently by resequencing bulked segregants from a backcrossed population and detecting homozygous mutations in the genetic regions using the three algorithms. From the results of practical heavy-ion mutagenesis for mutation breeding to date it was expected that only a small number of genes would be disrupted in the mutant genomes because heavy-ion irradiation can modify a trait of interest without affecting other traits, allowing direct subsequent use of the mutants as new cultivars. In the resequenced mutants, from just five to eleven genes were found to be homozygously mutated (Table 3), supporting this expectation. The number of mutated genes reported in

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

Whole-genome analysis of heavy-ion beam mutations 99 Table 3 Numbers of mutated genes and mutations expected to affect gene function in the three mutant lines Ar-57-al1

Missense Nonsense Frameshift Truncationa Amino acid insertion Loss of whole gene Mutated genes/genomeb

Ar-365-as1

Ar-443-as1

Homozygous

Heterozygous

Homozygous

Heterozygous

Homozygous

Heterozygous

6 0 0 3 0 0 7

0 1 1 2 0 0 4

1 0 0 5 0 0 5

3 1 4 13 0 1 19

0 0 0 0 0 11 11

5 2 7 7 1 74 94

The zygosity in each line is estimated from the genetic homogeneity in the M3 population. Truncations included amino acid deletions without a frameshift in the genes; and genes split by genomic rearrangements such as inversions and translocations. b When multiple mutations occurred in one gene, we counted those as a single gene mutation event. a

The size distributions of the simply rejoined deletions, which exclude the deletions at breakpoints of the rearrangement, and the rearranged fragments (≥100 bp), are shown in Figure 3. The maximum fragment size in the Ar-365-as1 mutant was 2.7 Mbp; over half of the fragments were within the several kbp to 1 Mbp size range. Although the simple deletions were fewer than the rearranged fragments (Figure 3), the extremely large deletions (1.2 Mbp being the maximum size in Ar-443-as1) were detected by the resequencing (Table S4). The simple deletions, ranging in size from 3 to 354 kbp, were also detected by array-CGH in the mutants with Ar-ion beam-induced mutations (Table 1). Seventy-six per cent of the breakpoints at the original sites of the rearranged fragments were coincident with deletions (Table 4). On the other hand, at the junctions involved in the rearrangements, regions of microhomology, ranging from 1 to 7 bp, were often used for rejoining breakpoints (62%; Table 4).

Figure 3. Size distribution of the simple deletions and rearranged fragments detected by genome resequencing of Ar-ion-induced mutants. The simple deletions did not include the deletions at breakpoints of the rearrangements. The fragments contributing to duplication, inversion, and translocation events were grouped according to size.

DISCUSSION The LET values of the Ar-ion (290 keV lm 1) and Fe-ion (640 keV lm 1) beams used in this study are much higher than those of c-rays (0.2 keV lm 1) or X-rays (2.0–5.0 keV lm 1). It has been suggested that the complexity and severity of DNA damage increases with the LET value (Hada and Georgakilas, 2008; Sage and Harrison, 2011). Correspondingly, in the Arabidopsis mutants, deletion sizes in the mutated genes increase with increasing LET value (Shikazono et al., 2005; Kazama et al., 2011; Hirano et al., 2012), and the Ar- and Fe-ion beams often induce large deletions (>several kbp) and genomic rearrangements (Hirano et al., 2012; Kazama et al., 2013). Array-CGH analysis determined that the average numbers of deletions (≥149 bp) in the mutant genomes derived from Ar- and Fe-ion irradiation were 1.9 and 3.7, respectively (Table 1; Kazama et al., 2013), and that 81% of the deletions were accompanied by genomic rearrangements (Table 1). These results indicated that irradiation with Arand Fe-ion beams has a large mutational effect on the plant genome, consistent with conclusions obtained from previous PCR-based research (Hirano et al., 2012; Kazama et al., 2013). It has been reported that in A. thaliana the majority of large deletions induced by radiation, such as those of several hundred kbp to several Mbp, were not transmitted to progeny; such non-transmissibility was considered to be due to the deletion of a particular region containing a gene or genes essential for gamete development or viability (Naito et al., 2005). The large deleted areas (>100 kbp) detected as transmissible mutations in this study were commonly characterized as containing few protein-coding genes (Table S1). This suggests that the possibility of the loss of genes essential for viability is greater for large deleted areas containing many protein-coding genes; thus, these areas tend to become non-transmissible mutations. A locational bias based on the chromosomal environment of the deletion, which might influence transmissibility, may also exist.

© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 93–104

102 Tomonari Hirano et al. may be useful for the elucidation of plant DSB repair pathways and in particular the A-EJ pathway. Mutation breeding remains an important strategy for plant improvement, and there has been high demand for heavy-ion mutagenesis. In the future, genome sequences will be available for breeding programs in greater numbers of plant species. Concomitantly, an efficient system for mutation detection is essential for progress using forward and reverse genetics. We have demonstrated the comprehensive detection of the total mutation profile of the mutant genomes derived from heavy-ion irradiation, and these results are expected to contribute to progress in mutation breeding. High-LET heavy-ion beams have distinctive mutational effects on the plant genome, and thus are useful in plant science as a unique mutagen. EXPERIMENTAL PROCEDURES

(Roche NimbleGen Inc.). For each spot on the array, signal values were calculated as log2 ratios of the Cy3-labeled sample (mutant) versus the Cy5-reference (wild type). Data analysis and visualization of the results were performed using SIGNALMAP software, version 1.9 (Roche NimbleGen Inc.). For peak detection, using the Find Peaks feature in SIGNALMAP, peak window size was 400 or 500 bp and the percentage of peak threshold was changed from 10 to 35%. The parameters and the detected peaks with signal values of 1.2 or greater are shown in Table S5. The candidate deletions detected as positive peaks were confirmed by PCR using genomic DNA from seven individual M3 plants (Figure S2). Following amplification of the regions of the candidate deletions, the fragments were sequenced using the BIG DYE TERMINATOR v. 3.1 Cycle Sequencing Kit and a 3730XL DNA Analyzer (Applied Biosystems, http://www.appliedbiosystems.com/). In cases in which we were unable to amplify the regions in the mutants, we defined that deletion with genomic rearrangements that had occurred in the region (Figure S2b). The primers used for peak confirmation are listed in Table S6.

Irradiation of heavy-ion beams and mutant screening

Resequencing of the mutant genomes

40 17+ Ar and 56Fe24+ ions were accelerated up to 95 and 90 MeV nucleon 1, respectively, at the RIKEN RI-beam factory (RIBF). Dry seeds of the A. thaliana ecotype Col-0 were placed into a plastic bag (Hybri-Bag Hard; Cosmo Bio, http://cosmobiousa.com/) and arranged in a single layer. The seeds were irradiated with each ion species at a dose of 50 Gy by using the E5 beam line in the RIBF. The LET values, which were calculated behind the seeds, were 290 keV lm 1 for the Ar ions and 640 keV lm 1 for the Fe ions. Methods for seed sowing and plant growth conditions were described previously (Kazama et al., 2011). Some phenotypic mutants were obtained in the M2 plants derived from the Ar- and Fe-ion beam irradiations (Hirano et al., 2012; Kazama et al., 2013): pale green leaves (Figure S1a); abnormal leaf shape (Figure S1b,c); or variegated leaves. In self-crossed progenies (M3 generation) of the M2 mutants, we confirmed that the phenotypes of the progenies were identical to the phenotypes observed in the M2 generation. The leaves for genomic DNA isolation were collected from 40 mutants in each M3 line and from wild-type plants. Genomic DNA for array-CGH and NGS was isolated from the leaves using a Nucleon PhytoPure kit according to the manufacturer’s instructions (GE Healthcare UK Ltd, http://www3.gehealthcare.co.uk/).

Array-CGH analysis Custom oligonucleotide array design and array-CGH analysis were performed according to the methods described by Kazama et al. (2013). In brief, the whole-genome sequence of A. thaliana, which was tiled with oligonucleotides, was based on The Arabidopsis Information Resource 8 (TAIR8) (http://www.arabidopsis.org/) for the mutants derived using the Fe-ion beam, and on TAIR10 for the mutants derived using the Ar-ion beam. The sequence of the oligonucleotide probes was initiated every 50 bp across the genome sequence, excluding repetitive sequences, and probe lengths ranged from 50 to 75 bp. The DNA of wild-type and mutant plants was labelled with Cy5 and Cy3, respectively. Hybridization, washing, and scanning of the CGH-array were conducted by Roche NimbleGen Inc. (http:// www.nimblegen.com/). Raw fluorescence intensity data were obtained from scanned images of the oligonucleotide tiling arrays by using NIMBLESCAN version 2.4 extraction software

Resequencing of the mutant lines Ar-57-al1, Ar-365-as1, and Ar-443-as1 was performed at the Takara Dragon Genomics Center (TaKaRa Bio Inc., http://www.takara-bio.com/) using one lane of the HiSeq 2000 sequencing system (Illumina Inc., http://www.illumina.com/). Indexed paired-end DNA libraries were prepared with the TruSeq DNA Sample Preparation Kit (Illumina Inc.) according to the TruSeq DNA Sample Preparation Guide rev. C manual. Cluster generation was performed using a TruSeq PE Cluster Kit v3-cBotHS (Illumina Inc.), and pair-end sequencing (100 bp) was performed using a TruSeq SBS Kit v3-HS. The reads obtained from resequencing were mapped to a reference genome of the A. thaliana ecotype Col-0 (TAIR10) using the Burrows–Wheeler aligner tool (Li and Durbin, 2009). The mapping results are shown in Table S2.

Mutation detection and screening After removal of potential PCR duplicates using PICARD TOOLS (version 1.55; http://picard.sourceforge.net/), candidate mutations were detected by use of SAMTOOLS (version 0.1.16), with default settings, and the candidates which were covered by fewer than three reads were filtered out. For the detection of genomic rearrangements, PINDEL (version 0.2.4.d) and BREAKDANCER (version 1.1) algorithms were also used after removal of the PCR duplicates with PICARD. Standard processes for screening heavy-ion beam-induced mutations from the candidates were performed as follows: (i) Candidate mutations identified by each algorithm that were located in regions of the nuclear genome with read depth ranging from 5 to 1000 were marked for further analysis. (ii) Candidate mutations observed in at least in two independent mutants were excluded as likely background mutations harbored in the ecotype Col-0 kept in our laboratory (Figure S5). (iii) When the proportion of resequencing reads indicating the mutation was

Comprehensive identification of mutations induced by heavy-ion beam irradiation in Arabidopsis thaliana.

Heavy-ion beams are widely used for mutation breeding and molecular biology. Although the mutagenic effects of heavy-ion beam irradiation have been ch...
701KB Sizes 0 Downloads 10 Views