Index INTRODUCTION MATERIALS AND METHODS Plant materials and DNA extraction Examination of deletion by PCR TAIL-PCR and sequencing Gene prediction Genomic sequencing of “Koshihikari” RESULTS Analysis of glu1 mutation Genes included in the deletion Repair of deletion DISCUSSION References

INTRODUCTION

One useful method for analyzing gene functions is to produce knockout mutants (Bouche and Bouchez, 2001). This is because the relevant gene function can be examined if the knocked-out gene can alter the phenotype. In higher plants, the following methods are generally in use for producing knockout mutants: employing chemical mutagens such as ethylmethane sulfonate (EMS); inserting transposons or T-DNA; and utilizing ionizing radiation such as X-rays, gamma rays, neutrons, or ion beams. EMS causes knockout by base substitution and transposons cause knockout by insertion, whereas ionizing radiation causes knockout by chromosomal structural changes including deletion, inversion, and reciprocal translocation. Such chromosomal structural changes are less likely to occur with other mutagens and they can be called radiation-specific mutation.

In sequenced plant genomes, 14% or more of the genes formed tandem array (Arabidopsis Genome Initiative, 2000; International Rice Genome Sequencing Project, 2005). Studying of such a gene function is very difficult, because mutating only one gene in a tandem array often produces no phenotypic change because of a functional redundancy. It is often necessary to knock out more than one gene to obtain a mutant phenotype, although creating such double mutants by crossing each mutant is very difficult because they are very tightly linked. Both RNA interference and antisense method can be down-regulation of tandem array genes and generate phenotypic change. However, both methods may produce variable results because they do not disrupt gene function perfectly.

Glutelin is a major rice seed storage protein and is coded by a multigene family. Depending on the homology, it is classified into two subfamilies, GluA and GluB, which show 60–65% of homology in their amino-acid sequences (Takaiwa et al., 1991). In both subfamilies, the precursor proteins are processed into acidic subunits (37 to 39 kDa) and basic subunits (22 to 23 kDa). Various types of glutelin mutants have been obtained from screening procedures with SDS-PAGE (Iida et al., 1993; 1997; Qu et al., 2002).

glu1 has been obtained as a mutant lacking of band a-1, which is the largest glutelin acidic subunit, by gamma irradiation (Iida et al., 1997). The visual characters of glu1 were not different from original cultivar, “Koshihikari”. Two-dimensional electrophoresis of glu1 endosperm protein revealed disappearance of spot 1a. The causative gene was mapped to the chromosome 2, but the mutation of glu1 has not been identified. During their analysis of the low-glutelin mutant LGC-1, Kusaba et al. (2003) found that GluB5 and GluB4 of the GluB family were located 3.8 kb apart from each other in a tandem array in a tail-to-tail orientation, and that those two genes had a very high level of homology. They further speculated that glu1 was a mutant related to GluB5 and GluB4 as was the case with LGC-1. They came to this conclusion from the following observations: the deletion occurring in GluB5 and GluB4 was the cause of the LGC-1 phenotype; LGC-1 and glu1 exhibited a common phenotype that loses spot 1a; and when genomic DNA was digested with Hind III and examined by Southern blot analysis using GluB1 cDNA as a probe, a band of 8.5 kb was detected in the wild type but not in LGC-1 or glu1.

We analyzed the mutation that occurred in glu1. It was revealed that a large scale deletion of 129.7 kb took place in the glu1 mutant, and that this deletion, in turn, caused a knockout of both GluB5 and GluB4 which exist in a tandem array. The amino acid sequence of the acidic subunit of GluB5 perfectly matched with that of GluB4, suggesting that only the mutation involving both GluB5 and GluB4 results in the lack of the glutelin subunit a-1 or spot 1a in two-dimensional electrophoresis. Our finding suggests that gamma-ray have a great advantage to disrupt tandem repeated and functionally redundant genes.

MATERIALS AND METHODS

Plant materials and DNA extraction

The glu1 homozygous mutant has been obtained from rice (Oryza sativa L.) by irradiating the cultivar “Koshihikari” with gamma rays (see Iida et al., (1997) for details). DNA was extracted from 100 mg of mature leaves using DNeasy Plant Mini Kit (QIAGEN, Tokyo, Japan).

Examination of deletion by PCR

PCR was performed in a total volume of 20 μl, containing 20 ng of DNA, 2 μl of 10 × ExTaq buffer (TaKaRa Bio, Shiga, Japan), 2 μl of 2.5 mM dNTP, 10 pmol of each primer, and 0.5 U of Taq DNA polymerase (ExTaq, TaKaRa Bio). The PCR condition was: 30 cycles of 94°C for 40 s, 58°C for 40 s, and 72°C for 40 s.

TAIL-PCR and sequencing

TAIL-PCR was performed as described by Liu et al. (1995) using Taq DNA polymerase (rTaq, Takara Bio), TAIL-1, TAIL-2, and TAIL-3 as spesific primers (Table 1) and 5’-GTNCGA(G/C)(A/T)CAN(A/T)AGC-3’ as an adapter primer (Miyao et al., 2003). The PCR product was purified with QIAquick Gel Extraction Kit (QIAGEN) and the directly sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Tokyo, Japan). The comparison between the resulting nucleotide sequence and a “Nipponbare” genomic clone was made using Sequencher (Hitachi software, Tokyo, Japan) program.

View Details

Table 1. Nucleotide sequences of the primers used in this study

Gene prediction

For gene prediction, RiceGAAS (http:// ricegaas.dna.affrc.go.jp/) was employed.

Genomic sequencing of “Koshihikari”

The amplification of “Koshihikari” DNA fragment corresponding to nucleotides 44130 in AP005428 was performed in a total volume of 20 μl, containing 20 ng of DNA, 2 μl of 10 × ExTaq buffer (TaKaRa Bio), 2 μl of 2.5 mM dNTP, 10 pmol each of primers 173690-F and 174140-R (Table 1), and 0.5 U of Taq DNA polymerase (ExTaq, TaKaRa Bio). The PCR condition was: 30 cycles of 94°C for 40 s, 58°C for 40 s, and 72°C for 40 s.

The amplification of “Koshihikari” DNA fragment corresponding to nucleotides 18727 in AP005875 was performed in a total volume of 40 μl, containing 20 ng of DNA, 4 μl of 10 × LA Taq buffer (TaKaRa Bio), 4 μl of 2.5 mM dNTP, 4 μl of 25 mM MgCl₂, 20 pmol each of primers 40866-F and R1 (Table 1), and 0.5 U of Taq DNA polymerase (LA Taq, TaKaRa Bio). The PCR condition was: 35 cycles of 94°C for 40 s, 58°C for 40 s, and 72°C for 3 min 30 s.

The PCR product was purified with a QIAquick Gel Extraction Kit (QIAGEN), and the directly sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems).

RESULTS

Analysis of glu1 mutation

It has been suggested that a deletion involving both GluB5 and GluB4 might occur in glu1 (Kusaba et al., 2003). In order to confirm this deletion, we conducted PCR by the primers having the sequences in the first exon of the GluB5 on the chromosome 2 (F1 and R1 of Fig. 1A). GluB5 and GluB4 are reported to have an extremely high level of homology, 99.8%, from the initiating codon to the stop codon and that they match perfectly except for 3 base substitutions in the fourth exon and 2 base substitutions in the third intron (Kusaba et al., 2003). When we compared the sequences of GluB5 and GluB4 of “Nipponbare” (Accession No. AP005428), the sequences from the initiation codon to a 1000-bp upstream point matched perfectly with each other (GluB5: 42842 to 43841, GluB4: 51289 to 52288; data not shown). As a result, when PCR was carried out with primers F1 and R1 using a wild type genomic DNA as a template, the first exon was amplified for both the GluB5 and GluB4 (Fig. 1A). When PCR was carried out using glu1 genomic DNA as a template, no amplification of this DNA fragment was observed (lane 1 of Fig. 1B). Similarly, when the amplification was attempted for the third exon using primers F2 and R2, and for the fourth exon using primers F3 and R3, DNA fragment was not amplified, either (lane 3 and 5 of Fig. 1B). These findings indicate that the deletion took place in a region at least from the first exon of the GluB5 to the first exon of GluB4 in glu1.

View Details

Fig. 1. Identification of deletion in glu1 by PCR. (A) Structure of GluB5 and GluB4 and position of primers. GluB5 and GluB4 are located in line about 3.8 kb apart in a tail-to-tail orientation. In GluB5 and GluB4, the transcriptional orientations are denoted by white arrows, the exons by white squares, and the introns by solid lines. The black arrows indicate positions of annealing for the primers. The characters indicate the names of primers. (B) PCR amplification of the DNA regions around GluB5 and GluB4. In the even-numbered lanes, genomic DNA from glu1 is used as a template. In the odd-numbered lanes, genomic DNA from “Koshihikari” is used as a template. The names of the primers used are displayed above the lane numbers. DNA size markers are shown on the left side.

When we attempted to amplify the region from about 300 bp upstream of the initiation codon to the middle of the first exon of GluB5 and GluB4 using primers F0 and R4, amplification of the DNA fragment was observed (lane 7 of Fig. 1B). In our attempt to amplify the region from the initiation codon to 3 kb upstream for both GluB5 and GluB4, upstream of GluB5 was amplified (lane 9 of Fig. 1B), while upstream of GluB4 was not (lane 11 of Fig. 1B). This indicates that the deletion took place from the last half of the first exon of GluB5 to at least 3 kb upstream of the initiation codon of GluB4 in glu1. An attempt to amplify the 3’ side of the first exon of GluB5 with TAIL-PCR yielded a PCR product of about 1.5 kb (data not shown). In a determined 568-bp sequence of this PCR product, the first 122 bp was the same as the first exon of the GluB5 (1 to 122 bp of Fig. 2A). The remaining 446 bp was then subjected to BLAST alignment, and the entire sequence except for G at position 123 was found to match the downstream sequence of 129.7 kb (Accession No. AP005875). The primer glu1-del-R1 (Fig. 1A) was then constructed using the sequence of AP005875 for PCR with the primer F1, and the band was found when genomic DNA of glu1 was employed as a template (lane 13 of Fig. 1B). This strongly suggested that a deletion of 129.7 kb took place in glu1, and that this deletion caused a loss of a glutelin subunit corresponding to the band a-1 through the loss of GluB5 from the last half of the first exon downstream and the loss of the entire GluB4 (Fig. 2B). Finally, we attempted to amplify the region from about 20 kb, 40 kb and 70 kb downstream of GluB5. DNA fragment was not amplified when we used glu1 genomic DNA as a template (Fig. 2C). This suggested that all the 129.7-kb DNA region was lost completely from the entire genome in glu1.

View Details

Fig. 2. Nucleotide sequences of PCR product obtained by TAIL-PCR and structure of deletion in glu1. (A) Nucleotide sequences of PCR product. The determined 568-bp sequence is shown. The first 122 bp matches with the first exon of the GluB5 gene. The nucleotide sequence from 124 to 568 bp indicated by boldface type matches with AP005875. The inserted G at position 123 is enclosed in a square. (B) Structure of deletion in glu1. The deletion in glu1 took place in the nucleotide sequence from the first exon of the GluB5 gene to a 129.7-kb downstream point. The region lost through the deletion is shaded in gray. The positions of annealing for the primers F1, glu1-del-R1, 20kb-F, 20kb-R, 40kb-F, 40kb-R, 70kb-F and 70kb-R are indicated by black arrows. (C) PCR amplification of the DNA regions from about 20kb, 40kb and 70kb downstream of GluB5 in glu1 and wild-type. DNA size markers are shown on the left side.

Genes included in the deletion

The nucleotide sequence of “Nipponbare” corresponding to the deletion in glu1 consisted of 129,744 bp. In order to study genes included in the deleted region, gene prediction was performed. The existence of 25 genes together with GluB5 and GluB4 were predicted (Table 2). Among these 25 genes, there was one gene which corresponds to a full length cDNA clone (No. 11 of Table 2). There were also three genes which showed high level homologies (E-value = 0.0) with expressed sequence tags (ESTs)(No. 4, 7, and 20 of Table 2). Moreover, there were ten genes which were predicted to encode retrotransposons (No. 12, 14, 15, 18, and 22–27 of Table 2). From the above findings, 17 genes, excluding the retrotransposons, were predicted in the deleted region.

View Details

Table 2. Genes included in the deletion in glu1

Repair of deletion

A large deletion such as the one that occurred in glu1 is created by a double strand break occurring in two positions in the DNA, followed by a loss of the DNA fragment in the center; this is then repaired by non-homologous end-joining (NHEJ)(Sachs et al., 2000). When the nucleotide sequence of the TAIL-PCR product was compared with the genomic sequence of “Nipponbare”, it matched with AP005428 from T at posision one to T at position 122, and with AP005875 from T at position 124 to the end (Fig. 2A). This indicates that a double-strand break occurred after T at position 122 on the 5’ side, and before T at position 124 on the 3’ side, followed by a loss of DNA fragment in the center, then repaired by NHEJ in glu1. G at position 123 did not match with either AP005428 (position 44130) or AP005875 (position 18727). We determined the genomic sequence of “Koshihikari” corresponding to G at position 123 and confirmed that G does not exist in the genomic sequence of “Koshihikari” (data not shown, see Materials and Methods), strongly suggested that this base G had been inserted during the repair.

DISCUSSION

The present study revealed that a deletion of 129.7 kb took place in the gamma-irradiation-induced mutant, glu1, and also revealed that this deletion knocked out two genes, GluB5 and GluB4, which form a tandem array with an extremely high level of homology and caused a loss of band a-1. It is known that a small deletion is also induced by gamma irradiation (Morita et al., 2005; Takano et al., 2005), suggesting that a large deletion, such as the one found in glu1 is not always induced by gamma irradiation. Although the size of the deletion in the genomic sequence of “Nipponbare” is equivalent to 129,744 bp, the actual size of deletion and the number of genes predicted in the deletion are thought to be different due to differences in genomic structures such as insertions or deletions because it originates from the cultivar “Koshihikari”. Assuming that the number of genes in the deleted region is the same for both “Nipponbare” and “Koshihikari”, 17 genes are knocked out in glu1. Knockout mutants are highly effective for analyzing gene functions, and glu1 can be used for a functional analysis of these genes in addition to GluB5 and GluB4. One of the gene, No.11 was annotated as COV1. cov1 mutant, which showed slower growing, twisted leaves and significant defect in the patterning and development of the vascular bundles compared with wild-type, has been isolated in Arabidopsis (Parker et al., 2003). On the other hand, the visual characters of glu1 were not different from wild-type. At least five COV1 homologous genes including the gene No.11 were found in rice genome by using BLAST program (Data not shown). Remaining four genes in glu1 might have a functional redundancy of COV1 in rice.

Band a-1 which is lost in glu1 is an acidic subunit of glutelin (Iida et al., 1997), and the deduced amino acid sequence for the acidic subunit region is perfectly identical for both GluB5 and GluB4. Thus, in order to obtain a band a-1 deficient mutant, it is necessary to create a double mutation of GluB5 and GluB4. Aside from radiation, transposons and EMS are frequently used as mutagens. Transposons induce mutation by insertion and EMS induces point mutations by base substitution. It will be stochastically extremely difficult to induce mutations simultaneously in two genes with transposons or point-mutation-inducing EMS when genes are located in extremely close proximity, which is the case for GluB5 and GluB4. There are reports wherein the function of genes with redundant function was successfully elucidated through creating individual mutants which were then crossed to make a double mutant (for review see Bouche and Bouchez, 2001). However, it will also be extremely difficult to obtain a double mutant by crossing when genes are located in the close proximity. Thus, the band a-1 deficient mutant glu1 is thought to be a mutant which is specific to DNA-deletion-inducing radiation. There are a lot of genes formed tandem-array in plant genomes. For example, 17% of Arabidopsis genes (Arabidopsis Genome Initiative, 2000) and 14% of Rice genes (International Rice Genome Sequencing Project, 2005) are members of tandem-arrayed gene families. Radiation is believed to be an effective way to obtain mutants of genes formed tandem-array with redundant function.

A large deletion is said to be generated by the repair of a double-strand break in DNA with NHEJ. Naito et al. (2005) demonstrated that large deletions with an average size of 2 Mb or more were generated in the M₁ generation obtained by pollination with Arabidopsis pollen that had been irradiated with gamma rays and a carbon ion beams. However, most of the large homozygous deletions were not inherited by M₂ generation, and the authors speculated that the deletion eliminated genes which are involved in formation of pollen and egg cells. While 17 genes including GluB5 and GluB4, which are lost in glu1, are probably not ones essential for survival or germinal differentiation because individuals with homozygous deletion can be obtained. Conversely, if no genes essential for survival exist in the region, large deletions such as those reported by Naito et al. (2005) are likely to be maintained. It has been reported that there are numerous repeat sequences and a low density of functional genes in the proximity of the centromere (Nagaki et al., 2004; Wu et al., 2004). In such a region, gamma-ray-induced large scale deletions are expected to be easily inherited by the M₂ or later generations.

Kusaba et al. (2003) found that when genomic DNA was digested with Hind III and examined by Southern blot analysis using GluB1 cDNA as a probe, a band of 8.5 kb was detected in the wild type but not in glu1. We searched Hind III recognition site around GluB5 and GluB4 of “Nipponbare” (AP005428), and found that Hind III recognition site was in the upstream of GluB5 (43370 to 43375) and GluB4 (43370 to 43375) respectively. The distance from one Hind III recognition site to another was 8.4 kb, suggesting that these two Hind III recognition sites produced a band which was detected in wild-type. Almost all DNA region between two Hind III recognition sites was lost in glu1, suggesting that the disappearance of the 8.5-kb band in glu1 was caused by the 129.7 kb-deletion.

In glu1, an insertion of base G was found to have taken place during the repair. It is well known that insertion of several base pairs to several thousand base pairs takes place with NHEJ (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Shikazono et al., 2001; 2005), and it is thought to be a usual phenomenon which occurs during repair by NHEJ. In addition, although short homology is often found at the terminal ends with repairs by NHEJ (Gorbunova and Levy, 1997; Salomon and Puchta, 1998; Shikazono et al., 2001; 2005), there is no short homology in glu1, suggesting that the repair in glu1 was made without short homology.

glu1 has been screened from progeny of gamma-ray-irradiated plants with a forward genetics approach. On the other hand, if we could screen the target mutant from the mutant group with a reverse genetics approach, it could serve even more uses. For a deletion of several thousand base pairs, the screening method utilizing PCR by Li et al. (2001) can be employed. This technique generates primers which sandwich the target gene and then searches for individuals with a deletion wherein small PCR products are more amplified compared with the wild type. Because of the high detection sensitivity of this technique, it can screen mutant DNAs in bulk. Using this technique, Li et al. (2001) successfully screened deletion mutants of 0.8 to 12 kb from fast-neutron-irradiated Arabidopsis. Currently, we are attempting to detect deletions in the gamma-ray-irradiated mutant group using the same technique. If this technique can work with gamma irradiation, it is anticipated to make screening of target mutants possible not only in rice but also in various other species since gamma irradiation is easier and safer than fast neutron irradiation.

This work was supported by a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan (Green Technology Project GR-1003) and , in part, by the budget for Nuclear Research of the Ministry of Education, Culture, Sports, Science, and Technology.