| Edited by Yoshio Sano. Ryouhei Morita: Corresponding author. E-mail: ryo622@affrc.go.jp. Abbreviation: DSB, Double strand break; LET, Linear energy transfer; NHEJ, Non-homologous end-joining; ROS, Reactive oxygen species; TILLING, Targeting Induced Local Lesions IN Genomes; aCGH, array comparative genomic hybridization. |
Mutants are indispensable and powerful tools used in the analysis of plant gene function. Researchers can link the gene with the phenotype by clarifying the mutant phenotype and the causative gene responsible for that phenotype. Higher plant mutants are usually produced using T-DNA or transposons, chemical mutagens, or ionizing radiation. The insertion of T-DNA or transposons destroys gene structure. Chemical mutagens such as ethyl methanesulfonate (EMS) or N-methyl-N-nitrosourea (MNU) mainly cause base substitution by DNA base alkylation (Greene et al., 2003; Slade et al., 2005; Cooper et al., 2008). Ionizing radiation produces reactive oxygen species (ROS), which interact with DNA and cause oxidative damage such as base modifications and single/double strand breaks (Belli et al., 2002; Roldan-Arjona and Ariza, 2009). Base modifications can cause base substitutions (Shibutani et al., 1991; Tajiri et al., 1995). Double strand breaks (DSBs) are repaired by two major pathways, namely homologous recombination (HR) and non-homologous end-joining (NHEJ)(Britt, 1999; Sankaranarayanan and Wassom, 2005; Kimura and Sakaguchi, 2006). HR is an error-free repair pathway, whereas NHEJ is an error-prone repair pathway, and often causes mutations such as deletions, insertions, and inversions at repair sites (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Kirik et al., 2000). As a result, ionizing radiation has been thought to have the potential to cause various types of mutation (Vizir et al., 1996).
Ionizing radiation can be divided into two classes according to differences in linear energy transfer (LET) (the energy transferred per unit length of the track). Alpha particles, neutrons, and heavy ion beams have high LET, while gamma rays, X-rays, and electron beams have low LET. Neutrons, classified as high LET irradiation, cause deletions of about 300 bp to 12 kbp in length in plant genomes (Sun et al., 1992; Bruggemann et al., 1996; Salmeron et al., 1996; Li et al., 2001; Nagano et al., 2008). Recent research has shown that heavy ion beams, also classified as high LET radiation, tend to induce structural changes in the chromosomes, such as inversions and translocations (Shikazono et al., 2005). However, there is little research analyzing the mutations induced in plant genomes by low LET radiation such as gamma rays. Cecchini et al. (1998) obtained eight mutants by gamma irradiation of Arabidopsis and reported that all mutants had deletions of 5 kbp or more. A number of researchers have reported on gamma irradiation-induced mutations in higher plants (1 bp deletion in Finkelstein et al., 1998; 2 single-base substitutions in Ashikari et al., 2002; 8, 10, 11 and 33 bp deletions in Takano et al., 2005; 2 bp deletion, 4 bp deletion and 4 single-base substitutions in Sato et al., 2006; 15 bp deletion in Ma, 2007; 7 bp deletion in Yan et al., 2007; 5 bp and 14 kbp deletions in Sato et al., 2009), but there is almost no comprehensive information. We therefore irradiated rice with gamma ray, screened for mutants, and determined the mutations. We analyzed 24 mutants and identified the type and size of the mutations. Based on these results, we also discuss a reverse genetics method with gamma irradiation-induced mutations.
A total of 24 mutants were used in this study. All mutants have been obtained from rice (Oryza sativa L.) by irradiating with gamma ray. The irradiated parts, doses, and dose rates are shown in Supporting Data 1. Two lines, namely cao-g1 (YM-15) and wx-g5 (02r200Gy) were derived from japonica rice cv. Hitomebore. Fifteen lines, namely cps-g1 (G010-1D), ga3ox-g1 (G041-9A), gid1-g1 (GAE038-3G), wx-g1 (G041-11H), ga3ox-g2 (GAE010-9H), kao-g1 (G004-11F), pla1-g1 (CNB287), pla2-g1 (G070-11A), wx-g3 (G040-10H), kao-g2 (GAE038-5F), wx-g4 (G038-5F), gid2-g1 (G068-3H), pla1-g2 (HIN2008-154), wx-g6 (G032-3H) and pla2-g2 (CNB337) were derived from japonica rice cv. Nipponbare. Four lines, namely gluA1-g1 (PCM-14), gluA2-g1 (PCM-8, 9, 10, and 11), glb1 (PCM-19) and glu1 (PCM-7) were derived from japonica rice cv. Koshihikari. Lines cao-g2 (G-52), wx-g2 (Odoroki-Mochi) and gluA2-g2 (M539) were derived from japonica rice cultivars Reimei, Takanari and Norin 8, respectively.
We screened six waxy mutants, four short plastochron mutants, and seven gibberellin-related dwarf mutants. The waxy mutants were screened by observing the external appearance of the hulled rice grain, and I2/KI staining. Short plastochron mutants or gibberellin-related dwarf mutants were screened by observing morphological changes at 1 month after planting. Gibberellin-related dwarf mutants were sprayed with 10 mM gibberellin (GA3, Wako Pure Chemical Industries, Osaka, Japan) and were categorized as mutations in gibberellin-deficient if elongation was observed after 10 days or mutations in gibberellin-insensitive if no elongation was observed. Two mutants with chlorophyll b deficiency were screened by Morita et al. (2005). Four mutants with glutelin deficiency were screened by Iida et al. (1997). One mutant with α-globulin deficiency was obtained by Iida et al. (1998).
Mutations were determined using PCR. DNA extraction was performed using the DNeasy Plant Mini Kit (Qiagen, Tokyo, Japan). PCR was performed on the causative genes using primers amplifying approximately 1 kbp. The mutation phenotypes and causative genes were shown in Table 1. When the entire coding region was amplified, we assumed that small deletions or base substitutions had occurred. PCR products were purified with a PCR purification kit (Qiagen), and sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Tokyo, Japan). When the whole or a part of the coding region could not be amplified, we assumed a large deletion, inversion, or translocation had occurred. We designed other primers, and conducted a detailed analysis including TAIL-PCR (Liu et al., 1995) for amplifying flanking sequences. We used Sequencher (Hitachi Software, Tokyo, Japan) for comparisons of the mutant and wild type nucleotide sequences. Results of cao-g1, cao-g2, glu1, and glb1 (Morita et al., 2005, 2007, 2009a) were included in the present analysis.
 View Details | Table 1 Mutation phenotypes, causative genes and the references for the nucleotide sequences of the causative genes |
To characterize DNA lesions induced by gamma-ray irradiation, candidates for causative genes of isolated 24 mutants were analyzed by PCR.
We isolated 7 dwarf mutants with dark green leaves, the characteristics of gibberellin (GA)-related mutants. Therefore, we tried to classify them into GA-deficient mutants and GA-insensitive mutants by assessing restoration of plant height with spraying GA. Among these mutants, two showed GA insensitivity, suggesting that they are categorized into GA-insensitive dwarf (gid) mutants. PCR amplification of GID1 revealed that one mutant harbors 1 bp deletion causing a frame shift in GID1, indicating that this is an allele of gid1 (gid1-g1)(Table 2). GID2 was not amplified in the other GA-insensitive mutant, suggesting that this is an allele of gid2. Further detailed analysis including TAIL-PCR revealed that this mutant (gid2-g1) harbors 42,184 bp (42.2 kbp) deletion involving the entire GID2 gene (Table 2).
 View Details | Table 2 Gamma irradiation-induced deletions and base substitutions |
The remaining 5 GA-sensitive mutants are thought to have lesions in the GA biosynthetic pathway. So, we analyzed the GA biosynthesis enzyme genes in the mutants. Two of them possessed 1 and 3 bp deletions in GA 3β-hydroxylase gene (GA3ox2), respectively, suggesting that they are alleles of ga3ox2 (ga3ox2-g1 and ga3ox2-g2)(Table 2). Another two had 4 and 16 bp deletions in ent-kaurenoic acid oxidase gene (KAO), respectively, suggesting that they are alleles of kao (kao-g1 and kao-g2)(Table 2). The remaining one has 1 bp deletion in ent-copalyl diphosphate synthase gene (CPS), suggesting that it is an allele of cps (cps-g1)(Table 2). The frame shifts occurred in the 5 mutants were thought to disrupt the function of the GA biosynthesis enzymes, resulting in reduced GA content in these mutants.
We newly isolated 5 wx mutants. The sequence analysis of the WX gene revealed that two of them have 2 and 6 bp deletions in the 2nd and 5th exons, respectively (wx-g1, wx-g3)(Table 2). The deletions in wx-g1 result in frame-shifts causing premature stop codon. On the other hand, the deletion in wx-g3 from codon 200 position 2 to codon 202 position 1 has resulted in the amino acid change from Leu-Leu-Cys to Arg. In wx-g5, single-base substitution was observed at the splice donor site of 10th intron (GT → GA), probably greatly reducing splicing efficiency of 10th intron (Table 2). In wx-g4, no amplification of WX gene was observed, suggesting deletion of WX gene (Data not shown). Detailed PCR analysis revealed that this deletion is 9,430 bp (9.4 kbp) long including the entire WX gene (Table 2).
The remaining wx mutant, wx-g6, harbors a structural change involving the WX gene. In this mutant, no PCR amplification occurred when primers designed for amplification of entire 2nd, 3rd and 4th exon were used while other part of WX gene can be amplified (Data not shown). Combination with TAIL-PCR revealed that this mutant has an inversion involving a very long region with one break point located in the 2nd exon and the other 1284.8 kbp upstream, disrupting the WX gene (Fig. 1). 1 and 4 bp deletions were associated with this inversion in the two rejoining sites (Fig. 1).
 View Details | Fig. 1 Gamma irradiation-induced inversions. The positions and nucleotide sequences of break points and rejoined sites of inversion in wx-g6 (A) and pla2-g2 (B) are shown. Schematic representation of gene structure of WX and PLA2 is shown in upper part, respectively. Asterisks indicate the positions of break points of the inversion. Boxes indicate exons. Arrow-shaped boxes indicate direction of the transcription. Underlining in the nucleotide sequences indicates microhomology. Positions of the break points are shown under nucleotide sequences as the position in the genomic clone. |
We further analyzed uncharacterized wx cultivar Odoroki-Mochi, which is induced by gamma rays from cv. Takanari. We named this allele as wx-g2. wx-g2 have 5 bp deletion in 9th exon, resulting in a frame-shift (Table 2).
Three short plastochron mutants, pla1 (Itoh et al., 1998), pla2 (Kawakatsu et al., 2006), and pla3 (Kawakatsu et al., 2009) have been reported. We newly isolated four pla mutants. Sequence analysis of PLA1 in these mutants revealed that two of them (pla1-g1 and pla1-g2) have a 5 bp deletion and a 1 bp substitution, respectively (Table 2). The 5 bp deletion caused a frame-shift of the PLA1 gene. The single-base substitution in pla1-g2 (GAG → GTG) resulted in an amino acid substitution of 407th glutamic acid to valine. The other two short plastochron mutants had DNA lesions in PLA2. pla2-g1 harbors 5 bp deletion in the 1st exon, resulting in a frame-shift (Table 2). Similar to the wx-g6 allele, PCR analysis of PLA2 revealed that pla2-g2 have an inversion of a very long DNA region of 3208.5 kbp (Fig. 1). One break point was located in the 4th intron, disrupting the PLA2 gene. This inversion accompanies 2 bp and 75 bp deletions at the break points (Fig. 1).
We newly analyzed three glutelin subunit mutants isolated by Iida et al. (1997). The sequence analysis of gluA1-g1 (namely 89WPKG30-16 in Iida et al., 1997) revealed that this mutant has 1 bp deletion in the 2nd exon of GluA1 gene (Table 2). The other two mutants had DNA lesions in GluA2 gene. gluA2-g1 (namely 88KG30-913, -958 and -993 in Iida et al., 1997) have 1 bp deletion in 1st exon of GluA2 (Table 2). gluA2-g2 (namely M539 in Iida et al., 1997) have a single-base substitution in the 3rd exon of GluA2, producing a premature stop codon [GAA (Glu) → TAA (Stop)] (Table 2).
The DNA lesions induced by gamma rays in rice observed in this study (24 mutations in total) were summarized in Fig. 1 and Table 2. These data clearly indicated that the most frequent DNA lesion induced by gamma rays is deletion (79.2%). Focused on deletions, most frequent deletion size is 1 bp (31.6%). Small deletions up to 16 bp frequently were observed: 1–16 bp deletions occupied 78.9% of total deletion mutations. The remaining deletions were relatively large deletions of 9.4–129.7 kbp. Interestingly, deletions between 100 bp and 8 kbp were not observed.
In EMS mutagenesis, G/C-to-A/T transition mutations are major one (Greene et al., 2003; Slade et al., 2005; Cooper et al., 2008), but in gamma ray mutagenesis in this study, all three base substitutions were transversion (Table 2). Ashikari et al. (2002) reported that both transition (G/C-to-A/T) and transversion (G/C-to-C/G) were induced by gamma rays. In addition, Sato et al. (2006) reported that three of four base substitutions induced by gamma rays were transversion, and the remaining one was transition. These results suggest that transversion may be the major base substitution induced by gamma irradiation.
Two inversions are observed. The inverted fragments in the inversions are very long (1284.8 kbp and 3208.5 kbp respectively)(Fig. 1), suggesting that gamma rays induce chromosomal level mutation as well as small DNA lesions.
Base insertions were quite rare in the present data. Base insertions are thought to be the consequence of insertion of ‘filler DNA’ in the DSB repair (Gorbunova and Levy, 1997). Kirik et al. (2000) reported that in the repair of DSBs, base insertions are common in tobacco but do not occur at all in Arabidopsis. In our survey of gamma-ray induced mutations in rice, only glu1 involves filler DNA associating with large deletion of 129.7 kbp (Table 2), suggesting that rice, like Arabidopsis, does not use repair mechanisms involving filler DNA insertions very frequently.
Microhomology is frequently present in rejoining sites (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Kirik et al., 2000). We found microhomology in 14 of 19 (73.7%) deletion events (Table 2). Moreover, in wx-6 and pla2-2 where inversions had occurred, microhomology existed in three of the four rejoining sites (Fig. 1), suggesting that the repair of gamma irradiation-induced DSBs frequently uses microhomology.
Analysis of the gamma irradiation-induced mutations showed that 62.5% were small deletions (1–16 bp), 16.7% were large deletions (9.4–129.7 kbp), 12.5% were base substitutions, and 8.3% were inversions (Table 3). Deletions and inversions tend to produce null mutations, while base substitutions may produce silent mutations and are therefore less likely to produce null mutations. We therefore consider that the proportion of base substitutions would be higher than the results seen in this study if all mutations including silent mutations were taken into account.
 View Details | Table 3 Classification of mutations induced by gamma irradiation |
Ionizing radiation produces both DNA strand breaks and oxidative DNA lesions by generating ROS. One of the oxidation product induced by ROS is 8-oxo-7-hydrodeoxyguanosine (8-oxo-dG)(Shibutani et al., 1991). When present in DNA, 8-oxo-dG can induce a G/C-to-T/A transversion. Moreover, 8-oxo-dGTP, major oxidation product of dGTP, is induced in the nucleotide pool by ROS. 8-oxo-dGTP can be incorporated into DNA during DNA replication, then induce both G/C-to-T/A and A/T-to-C/G transversions (Tajiri et al., 1995). The base substitution caused in gluA2-g2 was G/C-to-T/A transversion (Table 2), suggesting that this transversion may have occurred in the processes mentioned above. Remaining two base substitutions were T/A-to-A/T transversions (Table 2). Although the mechanism causing T/A-to-A/T transversions is not clear, it may be possible that gamma irradiation induce this type of transversion frequently.
The gamma irradiation-induced deletions were either 1–16 bp or 9.4–129.7 kbp in length and there were no deletions between 100 bp and 8 kbp in length. The deletion sizes obtained in this study were consistent with the sizes of gamma irradiation-induced deletions in higher plants from other studies (1 bp in Finkelstein et al., 1998; 8, 10, 11 and 33 bp in Takano et al., 2005; 2 and 4 bp in Sato et al., 2006; 15 bp in Ma, 2007; 7 bp in Yan et al., 2007; 5 bp and 14 kbp in Sato et al., 2009). According to Shikazono et al. (2005), deletions of 1–8 bp in length were induced in Arabidopsis by using an electron beam, which is classified as low LET radiation like gamma rays, suggesting that low LET radiation induces deletions of a similar length. In contrast, neutrons, having high LET, induce deletions ranging between about 300 bp and 12 kbp in length (Sun et al., 1992; Salmeron et al., 1996; Bruggemann et al., 1996; Li et al., 2001; Nagano et al., 2008). There may be a substantial difference in the size of deletion caused by low LET and high LET radiation.
Ionizing radiation produces DSBs in DNA. There have been reports that deletions ranging 1 bp and 2294 bp in length, occurring when plants repair a DSB produced in a single location using endonucleases (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Kirik et al., 2000). The sizes of the small deletions (1–16 bp) obtained in this study were included in the sizes of deletions generated by a single DSB repair, suggesting that the small deletions 1–16 bp in length may have occurred during the repair of a single DSB induced by gamma irradiation.
The large deletions 9.4–129.7 kbp in length have not been produced when plants repair a single DSB (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Kirik et al., 2000). Therefore, it is possible that the small deletions (1–16 bp) and the large deletions (9.4–129.7 kbp) have been generated by different processes. Naito et al. (2005) suggested that the deletions of around 500 kbp or more in length may have occurred when DSBs at two locations on the same chromosome caused the loss of a central DNA fragment and the DSBs were subsequently rejoined by NHEJ (Fig. 2). The large deletions 9.4–129.7 kbp in length may also have been produced in this process. Inversions occur when a DNA fragment produced by DSBs at two locations on the same chromosome binds inversely (Fig. 2). The size of the deletion or inversion reflects the distance between the two DSBs on the chromosome. In the gamma ray irradiation, deletions of between 100 bp and 8 kbp in length rarely occur, but deletions of around 10 kbp or more do occur; therefore, gamma ray irradiation may tend to induce DSBs at intervals of around 10 kbp or more. In contrast, neutron radiation causes deletions between about 300 bp and 12 kbp in length (Sun et al., 1992; Salmeron et al., 1996; Bruggemann et al., 1996; Li et al., 2001; Nagano et al., 2008), and may induce DSBs at closer intervals than gamma rays.
 View Details | Fig. 2 The process large deletions and inversions produce. Large deletions occur when two double strand breaks (DSBs) occur simultaneously on the same chromosome, DNA fragment B is lost, and A and C join (Left). Inversions occur when DNA fragment B inverts and binds back in (Right). The distance between the two DSBs determines the size of the deletion or inversion. |
As the distance between two DSBs increases, the size of the deletion increases when the resulting central DNA fragment is lost. Very large deletions of over Mbp occur frequently in irradiated generation (M1), but these are not inherited to the next generation (Naito et al., 2005). This may be because very large deletions frequently involves loss of the genes located in the deletion, which are necessary for survival or formation of reproductive cells (pollen or egg cells) and whose elimination has a lethal effect at the reproductive cell stage (Naito et al., 2005). Large deletions of 400–500 kbp were found in nyc3 alleles (Morita et al., 2009b). NYC3 locates adjacent to the centromere, a region with low gene density (Morita et al., 2009b). We think that this may reflect the situation that only a few genes are lost even with deletions of around 500 kbp when they occur in regions of low gene density.
Taken together, we propose a model describing the inheritance of gamma ray-induced deletions in seed-producing plants. In the gamma ray irradiated generation (M1), deletions occur that are either small in size (several bp) originating from a single DSB or are various sizes (around 10 kbp–over Mbp) originating from two DSBs. Deletions of between 100 bp and 8 kbp are rare to occur probably due to rare occurrence of close DSBs. Since very large deletions of over Mbp are not inherited to the next generation (M2) because of their gametophytic lethal effects, we may only observe small deletions of several bp or deletions of around 10 kbp to around 500 kbp in length in M2 and successive generations. When vegetatively reproducing or apomictic plants are irradiated by gamma rays, cells with deletions of over Mbp may not be eliminated through generations, because of no sexual reproduction. Therefore, very large deletions as well as small deletions, may be inherited in these plants. Because size-classification of deletion events mentioned above was a hypothesis, further experimental proof would be required to support it.
Li et al. (2001) developed a screening method for deletions from 0.8 kbp to 12 kbp using degenerate PCR. Majority of gamma ray-induced deletions are small (1–16 bp), or around 10 kbp or more in length, so this screening method may not be suitable for gamma ray-induced mutations. Alternatively, for small deletions and base substitutions, the TILLING method (McCallum et al., 2000) can be applied. Sato et al. (2006) have succeeded in detection of small deletions and base substitutions using a gamma ray-induced M2 population in rice. For large deletions, it may also be possible to use array comparative genomic hybridization (aCGH) (Nagano et al., 2008), a method detecting large deletions using microarray. These reverse genetic approaches will be useful in both mutation breeding and functional genomics, because they will enable us to obtain mutants efficiently from gamma ray-induced population.
This work was supported by a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Genomic for Agricultural Innovation, AMR-0003) and, in part, by a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Green Technology Project GR-1003).
 View Details | Supporting Data 1 The irradiation parts, total doses and dose rate of the mutants |
|
Index