Molecular breeding of a novel herbicide-tolerant rice by gene targeting

Authors

  • Masaki Endo,

    1. Plant Genetic Engineering Research Unit, Division of Plant Sciences, National Institute of Agrobiological Sciences, 2-1-2, Kannondai, Tsukuba, Ibaraki 305-8602, Japan,
    Search for more papers by this author
  • Keishi Osakabe,

    1. Plant Genetic Engineering Research Unit, Division of Plant Sciences, National Institute of Agrobiological Sciences, 2-1-2, Kannondai, Tsukuba, Ibaraki 305-8602, Japan,
    Search for more papers by this author
  • Kazuko Ono,

    1. Plant Genetic Engineering Research Unit, Division of Plant Sciences, National Institute of Agrobiological Sciences, 2-1-2, Kannondai, Tsukuba, Ibaraki 305-8602, Japan,
    Search for more papers by this author
  • Hirokazu Handa,

    1. Plant Genome Research Unit, Division of Genome and Biodiversity Research, National Institute of Agrobiological Sciences, 2-1-2, Kannondai, Tsukuba, Ibaraki 305–8602, Japan, and
    Search for more papers by this author
  • Tsutomu Shimizu,

    1. Life Science Research Institute, Kumiai Chemical Industry Co. Ltd, 3360 Kamo, Kikugawa, Shizuoka 439-0031, Japan
    Search for more papers by this author
  • Seiichi Toki

    Corresponding author
    1. Plant Genetic Engineering Research Unit, Division of Plant Sciences, National Institute of Agrobiological Sciences, 2-1-2, Kannondai, Tsukuba, Ibaraki 305-8602, Japan,
    Search for more papers by this author

(fax +81 29 838 8450; e-mail stoki@affrc.go.jp).

Summary

We have previously reported the production of a rice cell line tolerant to the acetolactate synthase (ALS)-inhibiting herbicide bispyribac (BS), and demonstrated that the BS-tolerant phenotype was due to a double mutation in the rice ALS gene. We further indicated that while changing either of the two amino acids (W548 L or S627I) individually resulted in a BS-tolerant phenotype, conversion of both amino acids simultaneously conferred increased tolerance to BS. As the BS-tolerant cell line had lost the ability to regenerate during two years of tissue culture selection, we attempted to introduce these two point mutations into the rice ALS gene via gene targeting (GT). Using our highly efficient Agrobacterium-mediated transformation system in rice, we were able to regenerate 66 independent GT rice plants from 1500 calli. Furthermore, two-thirds of these plants harbored the two point mutations exclusively, without any insertion of foreign DNA such as border sequences of T-DNA. The GT plants obtained in the present study are therefore equivalent to non-GM herbicide-tolerant rice plants generated by conventional breeding approaches that depend on spontaneous mutations. Surprisingly, GT rice homozygous for the modified ALS locus showed hyper-tolerance to BS when compared to BS-tolerant plants produced by a conventional transgenic system; ALS enzymatic activity in plants homozygous for the mutated ALS gene was inhibited only by extremely high concentrations of BS. These results indicate that our GT method has successfully created novel herbicide-tolerant rice plants that are superior to those produced by conventional mutation breeding protocols or transgenic technology.

Introduction

Acetolactate synthase (ALS) catalyzes the initial step common to the biosynthesis of the branched-chain amino acids leucine, isoleucine and valine (Chipman et al., 1998). ALS is the primary target site of action for at least four structurally distinct classes of herbicides (sulfonylureas, imidazolinones, triazolopyrimidine sulfonamides and pyrimidinyl carboxy herbicides) (for review, see Corbett and Tardif, 2006). ALS-inhibiting herbicides control a broad spectrum of grass and broadleaf weeds in crops, including weeds that are closely related to the crop itself and some key parasitic weeds. A major advantage of these compounds is that they are non-toxic to animals, highly selective, and very potent, thereby requiring only low application rates. Thus, ALS-inhibiting herbicides are an essential part of the multi-billion dollar weed control market.

Rotating the use of different combinations of several herbicides and corresponding herbicide-tolerant crops in the same field is one approach to overcoming the problem of herbicide-tolerant weeds. Several variant ALS genes conferring tolerance to ALS-inhibiting herbicides have been discovered in various crops (maize, wheat, rice, oilseed rape and sunflower) through conventional breeding methods such as mutagenesis and selection. However, only a single combination of an imidazolinone herbicide and the corresponding mutant, which was developed by a conventional breeding method, has been developed for commercialization in rice. Hence the goal of this study was to produce a novel rice plant tolerant to the pyrimidinyl carboxy herbicide bispyribac (BS).

In a previous study, as part of a two-year screen of cultured rice cells, we isolated BS-tolerant rice cells that showed hyper-tolerance to BS (Shimizu et al., 2005). BS tolerance was found to be linked to two point mutations in the ALS gene: a tryptophan (TGG) to leucine (TTG) change at amino acid 548 (W548 L), and a serine (AGT) to isoleucine (ATT) change at amino acid 627 (S627I). This double mutation in the rice ALS gene represents a novel combination of spontaneous mutations, with the substitution at the serine position not having been observed previously. Recombinant ALS proteins with each mutation singly, or the combined double mutation, were expressed in Escherichia coli, and colonies were examined for their sensitivity to BS. Although each individual amino acid change in ALS resulted in a BS-tolerant phenotype, conversion of both amino acids conferred increased tolerance to BS (Shimizu et al., 2005). As no plants could be recovered from the BS-tolerant rice cells because of prolonged tissue culture, we decided to directly produce rice plants containing these double mutations. Although chimeric RNA/DNA oligonucleotides have been used to induce site-specific base changes in several plant species (rice, Okuzaki and Toriyama, 2004; tobacco, Beethman et al., 1999; maize, Zhu et al., 1999), this technique can introduce only a single point mutation at a time. In this study, we used a highly efficient T-DNA-mediated gene targeting (GT) system to introduce W548 L and S627I mutations into the ALS gene. The procedure described here may facilitate the site-specific modification of genes in crop plants via GT.

Results and discussion

Experimental design of gene targeting

As shown in Figure 1(a), the GT vector Δ5′ ALS contains a partial ALS coding sequence (BAC clone OSJNBa0052 M16, 137 116–139 050 bp) with two point mutations: tryptophan (TGG) to leucine (TTG) at amino acid 548 (W548 L), and serine (AGT) to isoleucine (ATT) at amino acid 627 (S627I). Both mutations generate recognition sites for the restriction enzyme MfeI (CAATTG), which allows rapid detection of GT events. The ALS sequence on Δ5′ ALS lacks the 5′ coding region corresponding to the first 55 amino acids of ALS, which include the chloroplast-targeting signal (Pang et al., 2004), but contains 6.3 kb of the region downstream region of the ALS gene (BAC clone OSJNBa0052 M16, 139 051–145 372 bp) to allow efficient homologous recombination (HR). This region also contains unknown expressed genes of 2.6 and 0.7 kb (OSJNBa0052 M16.39 and OSJNBa0052 M16.40). We found a silent mutation in the coding region of the ALS gene (C → T at +612) and another silent mutation in the downstream region of the ALS gene (T → C at +2658) in Δ5′ ALS. These silent mutations were generated by PCR amplification during the construction of Δ5′ ALS. To facilitate molecular analysis of the recombination events, we retained these mutations. The ALS gene on Δ5′ ALS is expected to be non-functional, thus recovery of BS-tolerant plants is expected only after HR between Δ5′ ALS and the chromosomal ALS target locus.

Figure 1.

 Strategy for T-DNA-mediated gene targeting of the rice acetolactate synthase (ALS) locus and analysis of homologous recombination events in T0 plants.
(a) Schematic representation of gene targeting (GT) events. The blue boxes represent the coding region of the ALS gene. The thick lines represent flanking plant genomic DNA. A sequence encoding 55 amino acids, including the chloroplast-targeting signal, is deleted in the GT construct Δ5′ ALS, rendering the gene non-functional. The two mutations (W548 L and S627I) that confer herbicide bispyribac (BS) tolerance are marked by solid red vertical lines. Two silent mutations (C → T at +612, T → C at +2658) are marked by solid or dashed red vertical lines. The W548 L and S627I mutations create novel MfeI restriction sites (CAATTG). The positions of primers (F1 and R1) used for polymerase chain reaction (PCR), and the expected sizes of PCR-amplified fragments and their MfeI endonuclease digestion products are shown. The position of probe A used for Southern analysis is indicated on the targeted ALS locus. LB, left border; RB, right border; S, SphI restriction site.
(b, c) Selection of BS-tolerant calli on callus induction medium containing 0.75 μm of BS after infection of rice scutellum-derived calli with Agrobacterium carrying Δ5′ ALS (b) or with Agrobacterium carrying a binary vector harboring a full-length rice ALS gene containing the W548 L and S627I mutations (c).
(d) Example of an initial DNA analysis of BS-tolerant plants by PCR–MfeI digestion analysis. PCR was performed using primers F1 and R1. A band corresponding to 536 bp indicates introduction of only the W548 L mutation but not the S627I mutation into the ALS locus. If true GT has occurred, three fragments (1751, 299 and 237 bp) appear in the T0 generation in addition to the non-digested wild-type band of 2287 bp.
(e) Example of Southern blot analysis with probe A of SphI-digested DNAs from the individual wild-type (WT) and BS-tolerant T0 plants analyzed in (d).

After co-cultivation of Agrobacterium carrying Δ5′ ALS with approximately 1500 rice scutellum-derived calli (2–5 mm in diameter), and subsequent selection of BS-tolerant calli (Figure 1b), 72 independent T0 BS-tolerant plants (BSR 1 to BSR 72) were selected. As a positive control for BS selection, a second batch of rice calli was transformed with a full-length rice ALS gene comprising the 2.3 kb ALS gene coding region harboring the W548 L and S627I mutations flanked 5’ and 3’ by 2.2 and 0.5 kb, respectively, of the surrounding genomic sequence (Osakabe et al., 2005). In this latter transformation, almost all transformed calli produced considerable numbers of BS-tolerant colonies (Figure 1c).

Molecular characterization of the ALS locus in BS-tolerant T0 plants

A total of 72 BS-tolerant T0 plants obtained as above were analyzed to confirm the GT event. As shown in Figure 1(a), PCR amplification using primer F1, which anneals to the ALS promoter region (not present on the targeting construct), and primer R1, which anneals downstream of the ALS gene, yields a PCR product of 2287 bp. PCR products were further analyzed by digestion with MfeI. The presence of an uncut DNA fragment (2287 bp) as well as the three subfragments generated by cutting at the two restriction sites (1751, 299 and 237 bp) may be explained by HR at the target locus creating a heterozygous condition in primary T0 transformed plants. Figure 1(d) shows an example of this PCR–MfeI digestion analysis. Patterns indicating HR (non-digested 2287 bp, digested 1751, 299 and 237 bp) were found in DNA products derived from 66 individual BS-tolerant transformants. In six cases, an unexpected band pattern (non-digested 2287 bp, corresponding to the wild-type ALS locus, 1751 and 536 bp; Figure 1d, BSR 70 and 72) was seen. This result suggests that only one (W548 L) of the two mutations was present in the endogenous ALS locus of these plants. This result could be explained by spontaneous mutation at W548 L. Indeed, rice plants tolerant to BS owing to a W548 L mutation have been recovered by somaclonal variation (Okuzaki and Toriyama, 2004). Alternatively, HR might have occurred at a point between the W548 L and S627I mutations.

Although BS-tolerant mutations can occur spontaneously, the combination of the double mutation W548 L/S627I in the rice ALS gene was discovered only after two years of selection of BS-tolerant rice calli in liquid culture medium (Shimizu et al., 2005). The possibility that these two point mutations were acquired spontaneously during only two weeks of cell culture before BS selection is minimal. Furthermore, the presence of the non-selectable silent mutations, C → T at +612 and C → T at +2658, in the BS-tolerant rice plants demonstrated that the W548 L and S627I mutations were integrated into the rice genome via GT and not by the occurrence of spontaneous mutation (Table 2 and Figure 3c). In the case of true GT, no modification of genomic DNA other than point mutations should have occurred. To verify that this was the case, we performed Southern blot analysis on DNA from 21 randomly selected plants from the 66 individual T0 plants found to possess two MfeI sites in the ALS locus. Southern blot analysis of SphI-digested DNA with probe A (Figure 1e) and DraI-digested DNA with probe C (Figure 2b) revealed that 14 of the 21 T0 plants had the same band pattern as wild-type rice cv. Nipponbare, confirming that no rearrangements had occurred other than HR between the targeting construct (Δ5′ ALS) and the chromosomal copy of the ALS gene.

Table 2.   Segregation of the acetolactate synthase (ALS) locus in the progeny of true gene targeting candidates
T0Polymerase chain reaction (PCR)–MfeI digestionaC → T (+612)bT → C (+2658)c
HomoHeteroWild type (WT)
  1. aHomo, homozygous for the modified ALS locus; hetero, heterozygous for the modified ALS locus; wild type, segregated-out WT ALS locus as determined by PCR–MfeI digestion analysis.

  2. bNucleotide at position +612 of the modified ALS locus in homozygous plants.

  3. cNucleotide at position +2658 of the modified ALS locus in homozygous plants.

BSR 9262CC
BSR 13282TC
BSR 15452CC
BSR 16244CC
BSR 17534CC
BSR 18217CC
BSR 26131TC
BSR 27272CC
BSR 38532CC
BSR 49164CC
BSR 54434TC
BSR 59174CC
Figure 3.

 Molecular analysis of T1 progeny of a gene targeting (GT) candidate.
(a) Diagram showing the T-DNA region of the targeting vector Δ5′ ALS and the targeted ALS locus. The expected band sizes after Southern blot analysis using probe B are shown. Pink lines indicate the polymerase chain reaction (PCR) products used for direct sequencing. Triangles indicate the primers used for amplification of these PCR products. LB, left border; RB, right border; M, MfeI restriction site.
(b) PCR–MfeI digestion analysis of wild-type (WT) and eight randomly selected progeny of BSR 9 (BSR 9-1 to BSR 9-8). PCR was performed using primers F1 and R1. The genotypes of each T1 progeny suggested by this experiment are: homozygous for the modified ALS locus, BSR 9-3; heterozygous for modified ALS locus, BSR 9-1, -2, -4, -6, -7 and -8; homozygous for the wild-type ALS locus, BSR 9-5.
(c) Sequencing chromatograms of the mutation sites. F1 and R1 were used as primers for PCR. Reverse primer R1 was used for direct sequencing analysis of W548 L and S627I mutations, thus the complementary sequences of each mutation points are shown in the chromatograms. Direct sequencing indicated converted nucleotides (TTG at amino acid 548 and ATT at amino acid 627) in BSR 9-3 (homo) or unconverted nucleotides (TGG at amino acid 548 and AGT at amino acid 627, i.e. identical to wild-type) in BSR 9-5 (WT). Two overlapping peaks (TGG and TTG at amino acid 548, and AGT and ATT at amino acid 627) were observed in the heterozygous plant BSR 9-4 (hetero). The reverse primer R4 was used for direct sequencing analysis of the silent mutation C → T at +612. All three plants contained the non-converted nucleotide C at +612. For sequencing analysis of the silent mutation T → C at +2658, F3 and R3 were used as primers for PCR, and forward primer F4 was used for sequencing. Direct sequencing indicated the presence of the converted nucleotide C in BSR 9-3 (homo) and the unconverted nucleotide T in BSR 9-5 (WT). Two overlapping peaks, C and T, were observed in BSR 9-4 (Hetero).
(d) Southern blot analysis of MfeI-digested wild-type (WT) and BSR 9-3, -4 and -5 using probe B.

Figure 2.

 Southern blot analysis of the ALS locus in T0 plants.
(a) Diagram showing the T-DNA region of the targeting vector Δ5′ ALS and the targeted ALS locus. The expected band size after Southern blot analysis using probe C is shown. LB, left border; RB, right border; D, DraI restriction site.
(b) Southern blot analysis with probe C of DraI-digested individual wild-type (WT) and BS tolerant T0 plants. Bands other than 3.4 kb indicate the occurrence of random integration of T-DNA, ectopic GT, or unexpected rearrangements of the ALS locus.

Seven of the plants analysed by Southern blotting showed extra bands. As PCR–MfeI digestion analysis indicated that GT events had occurred in these seven plants, at least in the 5’ flanking region of Δ5′ ALS, additional bands in Southern blot analysis might result from random integration of T-DNA. However, these additional bands could also be explained by the occurrence of HR at one end of the T-DNA and non-homologous end-joining at the other end, or by ectopic GT (Endo et al., 2006a; Hanin et al., 2001; Hohn and Puchta, 2003; Offringa et al., 1993). We eliminated these plants from candidates of the ‘clean GT plants’.

Confirmation of GT modification at the ALS locus in true GT candidates

Analysis of the progeny of primary BS-tolerant recombinants (T1) allows distinction between true GT and ectopic GT events, as T1 generation individuals homozygous for the modified copy of the chromosomal ALS locus would segregate only in the case of true GT. Detailed analysis of true GT candidates was thus performed in T1 progeny. Figures 3 and 4 show examples of such analyses. We analyzed eight randomly chosen T1 progeny plants derived from one true GT T0 plant (BSR 9). Genotyping analysis by PCR–MfeI digestion as described above showed segregation of the wild-type and mutated ALS genes in the T1 generation; one plant homozygous for the modified ALS locus (BSR 9-3), six heterozygous plants (BSR 9-1, -2, -4, -6, -7, -8), and one plant with a wild-type ALS locus (BSR 9-5) were identified (Figure 3b). These results confirm the occurrence of precise gene replacement at the ALS locus of BSR 9. Genomic DNAs from the homozygote (BSR 9-3), one of the heterozygotes (BSR 9-4), and the segregated wild-type plant (BSR 9-5) were isolated and sequenced. Direct sequencing of PCR fragments F1–R1 (see Figure 3a) revealed that the W548 L (TGG to TTG) and S627I (AGT to ATT) mutations co-segregated in these plants (Figure 3c), confirming the results of PCR–MfeI digestion analysis (Figure 3b). As mentioned previously, a silent mutation in the ALS gene (C → T at +612) and another silent mutation in the downstream region of the ALS gene (T → C at +2658) exist in Δ5′ ALS. To confirm the GT event in true GT candidates, we analyzed the existence of these mutations in T1 progeny plants. The T → C mutation at +2658 also co-segregated with W548 L and S627I mutations in BSR 9-3, -4 and -5 (Figure 3c). The C → T mutation at +612 was not detected in these plants (Figure 3c). This means that the C → T mutation at +612 was not integrated into the genome of the BSR 9 plant during GT. The C → T mutation at +612 is located in the vicinity of the left border of Δ5′ ALS, and thus might be excluded during the process of HR. These results indicate that a crossover occurred between +612 and +1643 (W548 L), and that another crossover occurred between +2658 and +8257 in BSR 9. Neither insertion of T-DNA border sequence nor deletion of rice genomic sequences at the positions of the right and left borders were detected in PCR fragments F1–R1 and F2–R2 of the three plants analyzed (data not shown). Southern blot analysis of MfeI-digested genomic DNA also showed segregation of the S627I mutation (Figure 3d). Furthermore, in Southern blot analysis using probe D, which covers Δ5′ ALS, we detected the same band pattern in the wild-type and the plant homozygous for the modified ALS locus (Figure 4b). The results of Southern blot analysis shown in Figure 4(c,d) also confirmed that no modification other than the introduction of the W548 L and S627I mutations had occurred in the ALS locus of these T1 progeny. From these T1 analyses, we concluded that true GT had indeed occurred in BSR 9. As expected, there was no change in ALS gene transcription levels between plants homozygous (GT homo) or heterozygous (GT hetero) for the modified ALS locus and wild-type plants (Figure 4e). For three true GT lines, T2 seeds from GT homo/GT hetero/segregated wild-type T1 plants were germinated on medium containing 1 μm BS. These plants showed the expected segregation ratio of the BS-tolerant trait (Table 1). We also analyzed the progeny of 12 true GT candidates by PCR–MfeI digestion. All 12 lines showed segregation of homozygous plants, plants heterozygous for the modified ALS locus, and plants with the wild-type ALS locus (Table 2). These results confirm the occurrence of precise gene replacement at the ALS locus (true GT) in these plants. All these true GT plants possessed the silent mutation downstream of the ALS gene on the GT vector Δ5′ ALS (T → C at +2658), and the C → T mutation at +612 was detected in three lines (BSR 13, 26 and 54).

Figure 4.

 Southern and Northern blot analysis of the acetolactate synthase (ALS) locus in T1 plants.
(a) Diagram showing the T-DNA region of the targeting vector Δ5′ ALS and the targeted ALS locus. The expected band sizes after Southern blot analysis using probes A, C and D are shown. LB, left border; RB, right border; D, DraI restriction site; C, ClaI restriction site.
(b) Southern blot analysis of ClaI-digested wild-type (WT) and BSR 9-3 (homozygous for the modified ALS locus) with probe D.
(c,d) Southern blot analysis of DraI-digested wild-type (WT) and BSR 9-1 to BSR 9-8 using probe A (c) or probe C (d). Several additional bands other than 4.3 kb in (c) or 3.4 kb in (d) were due to non-specific hybridization, as these additional bands were also seen in the WT lane.
(e) Northern blot analysis with probe C of wild-type (WT) plants, and plants homozygous (homo) and heterozygous (hetero) for the modified ALS locus. The upper panel shows the position of the Oryza sativa ALS transcript (OsALS). The lower panel shows ethidium bromide-stained rRNA as a loading control.

Table 1.   Segregation of bispyribac (BS) sensitivity in the T2 generation of true gene targeting plants
T0 T1 genotypingaNumber of T2 seeds tested BSR BSSExpected ratio (BSR:BSS)
  1. aGenotypes of the T1 seeds were determined by polymerase chain reaction–MfeI digestion analysis. Homo, homozygous for the modified acetolactate synthase (ALS) locus; hetero, heterozygous for the modified ALS locus; WT, segregated-out WT ALS locus.

BSR 9Homo262601:0
Hetero282173:1
Wild-type (WT)300300:1
BSR 54Homo242401:0
Hetero262063:1
WT240240:1
BSR 59Homo242401:0
Hetero302373:1
WT240240:1

Summarizing the results of T0 and T1 analysis, we obtained 72 independent BS-tolerant plants from approximately 1500 transformed rice calli. Of these 72 BS-tolerant plants, 66 plants possessed (functional ALS with W548 L and S627I mutations), via either true or ectopic GT in the genome. Because 14 of 21 randomly chosen T0 plants showed the same band pattern as the wild-type in Southern blot analysis (Figures 1e and 2b), we estimated that true or ectopic GT without random integration of T-DNA occurred in two-thirds of BS-tolerant rice plants. PCR–MfeI digestion analysis of 12 T1 progeny of true ‘clean’ GT candidates confirmed that true GT had occurred in all 12 lines analyzed (Figure 3b and Table 2). Thus, the estimated number of true GT plants without any modification of the genome is two-thirds of the 66 GT plants obtained in this experiment. It is difficult to estimate GT efficiency precisely because a considerable number of transformation events occurred in each callus, judging from both the positive control experiments (Figure 1c) and an Agrobacterium-mediated transformation study with a GFP reporter gene construct (Toki et al., 2006). Based on the transformation efficiency of the GT vector (Δ5′ ALS) and that of positive control experiments, the estimated GT efficiency in our study was roughly 1 event in 2000–3000 potential transformants.

Plants homozygous for the modified ALS gene exhibit hyper-tolerance to BS

We confirmed the acquisition of BS tolerance in GT rice plants by spraying segregated T1 plants with BS (1 kg active ingredient (a.i.) ha−1) (Figure 5a). Rice plants homozygous and heterozygous for the modified ALS locus exhibited tolerance to BS and grew normally.

Figure 5.

 BS sensitivity of segregated T1 plants.
(a) Plants at the 5–6-leaf stage were sprayed with bispyribac (BS) solutions (1 kg a.i. ha−1). T1 plants with a modified ALS locus (either homozygous or heterozygous) showed BS tolerance. T1 plants carrying the wild-type ALS gene showed BS sensitivity, i.e. withered leaves.
(b) Sensitivity of ALS enzyme to BS. The inhibition rates of ALS enzymatic activities at various BS concentrations are shown. ‘Over-expression’ refers to plants over-expressing the mutated ALS gene containing the same two mutations as GT plants.

Next, we analyzed the enzymatic activity of ALS. Protein extracts from wild-type plants, plants homozygous or heterozygous for the modified ALS locus, and plants over-expressing the ALS gene containing the same two mutations (W548 L and S627I), were examined for their sensitivities to BS. In plants over-expressing the ALS gene containing the two point mutations, the mutated ALS gene was driven by a strong 35S promoter, thus the transcriptional level of this mutated ALS gene was much higher than that in wild-type or GT plants (approximately 20-fold higher than wild-type, data not shown). In this experiment, extracts of plants homozygous for the modified ALS locus were extremely tolerant to BS; BS had almost no effect on the enzyme even at a concentration of 100 μm (Figure 5b). Interestingly, the BS tolerance level of GT homozygous plants exceeded that of plants over-expressing the mutated ALS gene. The BS tolerance levels of plants heterozygous for the modified ALS locus and that of plants over-expressing the ALS gene containing the same two mutations were almost the same.

The crystal structure of the catalytic subunit of Arabidopsis ALS in complex with ALS-inhibiting herbicides was recently published (McCourt et al., 2006). The Arabidopsis ALS tetramer consists of four identical subunits with an overall folding similar to that of ALS from Saccharomyces cerevisiae. In plants heterozygous for the modified ALS gene, or in plants over-expressing the mutated ALS gene, the putative ALS complexes would be composed of a mixture of BS-tolerant and BS-sensitive subunits. Presumably chimeric ALS complexes containing both types of subunit would be inhibited by BS. Thus, exclusion of the wild-type, BS-sensitive, ALS allele is important to produce novel crops, which are hypertolerant to BS. In this regard, GT can directly replace a wild-type gene with its mutated counterpart, making it a powerful tool for this kind of gene modification.

Concluding remarks

We were able to obtain clean GT plants with high efficiency (approximately two-thirds of 66 GT plants from 1500 transformed rice calli weighing approximately 30 g). This success can be ascribed to several factors. Firstly, the efficient GT system presented here is highly dependent on our efficient transformation system in rice (Toki et al., 2006). In our transformation system, several to hundreds of independent transformation events occur in each small piece of callus (Toki et al., 2006). Secondly, in contrast to a negative–positive selection GT system (Terada et al., 2002), our GT vector did not contain a negative selection marker, so random integration and/or transient expression of the GT vector was not thought to inhibit growth of potential targeted transformants. Thirdly, we speculate that the condition of the chromatin structure of rice callus cells might enhance HR. Recent reports indicate that an open chromatin structure stimulates HR in higher plants (Endo et al., 2006b; Shaked et al., 2005), and patterns of DNA methylation and histone H3, H4 acetylation have been shown to vary during growth of asynchronous potato cell suspensions (Law and Suttle, 2005).

As the first report of the targeted reconstruction of a transgene in the tobacco genome (Paszkowski et al., 1988), GT events have been described at several loci in dicotyledonous plants (Endo et al., 2006a; Hanin et al., 2001; Kempin et al., 1997; Lee et al., 1990; Miao and Lam, 1995; Shaked et al., 2005). However, in monocots, which represent many important crop plants, only one successful GT event has been reported – a targeted disruption in rice of the Waxy gene, a key enzyme in amylose synthesis, by an exogenous selection marker (hygromycin phosphotransferase) using a negative–positive selection GT system (Terada et al., 2002).

In this study, we succeeded in introducing several point mutations (W548 L and S627I as well as silent mutations) by a T-DNA-mediated GT system. GT plants obtained in this study are equivalent to non-GM herbicide-tolerant rice plants generated by conventional breeding approaches that depend on spontaneous mutations. Furthermore, BS tolerance of GT plants homozygous for the mutated ALS gene exceeded that of conventional transgenic plants over-expressing an ALS gene containing W548 L and S627I mutations, indicating that the GT approach might be superior to the conventional transgenic system in this respect.

Many agronomically valuable phenotypes and natural variations are caused by single, or a small number of, point mutations. In addition to herbicide tolerance, salt tolerance and pigmentation are also potential selectable phenotypes. Beneficial mutations are often discovered under long periods of cell culture, but few plants can be recovered from these cells because of the prolonged cell culture. From this point of view, GT is a powerful tool for producing plants possessing these mutations in situ. Our GT strategy could be applied directly to the improvement of these traits. On the other hand, other agronomically valuable phenotypes caused by single, or a small number of, point mutations are non-selectable at the stage of transformation using current methods (photoperiod sensitivity, Takahashi et al., 2001; spotted leaf phenotype, Yamanouchi et al., 2002; short-day promotion of flowering, Doi et al., 2004; seed shattering, Konishi et al., 2006). If the GT frequency can be improved substantially, co-transformation of a selectable marker gene and a non-selectable GT construct and subsequent identification of desirable targeted events using suitable techniques such as a TILLING (Till et al., 2003) will cope with this problem. In fact, we observed that one-third of GT plants possessed randomly integrated T-DNA (Figures 1e and 2b). Previous attempts to improve GT or HR efficiency have included the induction of double-strand breaks at the target site (Lloyd et al., 2005; Puchta et al., 1996; Urnov et al., 2005), modification of proteins involved in HR (e.g. Rad54, Shaked et al., 2005), and remodeling of chromatin structure (INO80, Fritsch et al., 2004; CAF-1, Endo et al., 2006b). Establishment of a GT system generally applicable to any gene represents the next step in the practical molecular breeding of crops.

Experimental procedures

Vector construction

The residues to be mutated were selected based on analysis of a BS-tolerant rice callus (Oryza sativa L. cv. Kinmaze) produced by somaclonal mutation during tissue culture (Shimizu et al., 2005). In the ALS gene of this BS-tolerant callus, tryptophan 548 was mutated to leucine (W548 L) and serine 627 was mutated to isoleucine (S627I). Both mutations produce novel MfeI sites (Figure 1a). To construct the targeting vector Δ5′ ALS harboring the 5′ truncated ALS gene carrying these two point mutations, genomic DNA from BS-tolerant rice callus was used as a template for PCR with primers Asc-ALS (5′-GGCGCGCCGCCGGCCACGCCGCTCCGGCCGT-3′) and Pac-ALS (5′-TTAATTAAAAACTTGATATTATTCACACAGTGCCCCA-3′). An 8 kb fragment of the PCR product containing the W548 L and S627I mutations was cloned into the AscI/Pac I sites of the binary vector pPZP2028 (M. Kuroda, National Agricultural Research Center, Hokuriku Research Center, Niigata, Japan, and S. Toki, unpublished data), which is a derivative of pPZP202 (Hajdukiewicz et al., 1994), with the addition of rare restriction (AscI/PacI) sites to both ends of the multi-cloning site of pPZP202.

Rice transformation and BS selection

Agrobacterium-mediated transformation of rice (O. sativa L. cv. Nipponbare) was performed as described previously (Toki, 1997; Toki et al., 2006). After co-cultivation of Agrobacterium carrying the binary vector Δ5′ ALS with approximately 1500 rice scutellum-derived calli (approximately 30 g; pre-cultured for 25 days, 2–5 mm in diameter) for 3 days, infected calli were transferred to fresh callus induction medium (Toki, 1997) containing 0.75 μm BS (Kumiai Chemical Industry Co.; http://www.kumiai-chem.co.jp/english/index.html), and BS-tolerant cells were selected over 25 days. Calli that grew vigorously on the selection medium were transferred to regeneration medium containing the same concentration of BS. The regenerated plants were further grown on hormone-free medium to facilitate root growth, and then planted in soil and grown to maturity. These transgenic plants (T0 generation) were self-fertilized, and T1 seeds were collected for detailed analysis.

PCR– MfeI digestion analysis

To confirm the occurrence of GT in BS-tolerant rice plants, we performed PCR analysis coupled with MfeI digestion. We amplified a 2287 bp PCR fragment corresponding to the N-terminal region of ALS using primers F1 (5′- CGTCACCGCGCGCGGACAAAACACCCAC-3′), which anneals to the promoter region of the ALS gene (absent in Δ5′ ALS), and R1 (5′-ACATGATATCTTGTGATGCATATGCCTAC-3′), present in Δ5′ ALS (Figure 1a). The PCR products were digested with MfeI.

Southern blot analysis

Genomic DNA was extracted from leaves of seedlings using the Nucleon Phytopure extraction kit (Amersham Pharmacia Biotech; http://www5.amershambiosciences.com/) according to the manufacturer’s instructions. After SphI, MfeI, DraI or ClaI endonuclease digestion and electrophoresis on a 1% agarose gel, DNA fragments were transferred onto a positively charged nylon membrane (Roche; http://www.roche.com). Probes were prepared using a PCR DIG probe synthesis kit (Roche) and hybridization was performed according to the DIG Application Manual (Roche). Hybridization was performed at 42°C, and washing was performed under high-stringency conditions at 68°C.

Sequence analysis

Amplification using primers F1 and R1 yielded a PCR product of 2287 bp. Primer F1 was used to check the integration of left border sequence or the deletion of any rice genomic sequence, and primer R1 was used to check the presence of the W548 L and S627I mutations. Primer R4 (5′-AGGCGTGCGATGTACCCTGGTAGATT-3′) was used to check the presence of the C → T mutation at +612. Amplification using primers F2 (5′-AGCCTGTCTGCTGGGAGACCACT-3′) (present in Δ5′ ALS) and R2 (5′-AGGATGCTTCTCTCTTCCACCGATCCA -3′) (absent in Δ5′ ALS) yielded a PCR product of 887 bp. Primer F2 was used to check the integration of right border sequence or the deletion of rice genomic sequence. Amplification using primers F3 (5′-TGTCTTCGGCTGGTCTGGGCGCA-3′) and R3 (5′-TCCATTGGCCTAGGTAGTACACACTTACATCA-3′) yielded a PCR product of 1859 bp. Primer F4 (5′-TGGGTACTACTATAGAGAGAGGCTGCATGAAGT-3′) was used to check the presence of the T → C mutation at +2658. In all sequencing analyses, PCR products were used directly as templates for sequencing.

Northern blot analysis

Total RNA was prepared from 6-week-old seedlings using an RNeasy Plant Mini Kit (Qiagen; http://www.qiagen.com/). Total RNA (10 μg) was loaded and separated on a 1% agarose gel, and transferred onto a positively charged nylon membrane (Roche). Probe C was prepared using a PCR DIG probe synthesis kit (Roche), and hybridization was performed according to the DIG Application Manual (Roche). Hybridization was performed at 50°C, and washing was performed under high-stringency conditions at 50°C.

BS sensitivity test of T1 progeny

Plants at the 5–6-leaf stage were sprayed with BS solutions (1 kg a.i. ha−1) containing 5000-fold diluted Kumiten (Kumiai Chemical Industry Co.) as an adhesive. Photographs were taken 9 days after BS treatment.

Enzymatic assay of ALS

Proteins including ALS were extracted from rice shoots. The shoots were homogenized in five volumes of 0.1 m potassium phosphate buffer (pH 7.5) containing 0.5 mm MgCl2, 10% v/v glycerol. The homogenates were filtered through one layer of nylon gauze, and centrifuged at 15 000 g for 20 min. ALS was precipitated from the supernatant with ammonium sulfate at 50% saturation. After centrifugation, at 15 000 g for 20 min, the pellets were dissolved in 0.1 m potassium phosphate buffer (pH 7.5) containing 10% glycerol and 0.5 mm MgCl2. The solution was centrifuged at 15 000 g for 20 min to remove insoluble substances. The supernatant fluids were then desalted on a Sephadex G-25 column (GE Healthcare Bio-Sciences; http://www.gehealthcare.com) equilibrated with the same buffer and stored at −80°C. All operations were carried out at 0–4°C.

Enzyme activity of ALS was assayed at 37°C in 0.5 ml of assay mixture containing 20 mm potassium phosphate buffer (pH 7.5), 20 mm sodium pyruvate, 0.5 mm thiamine pyrophosphate, 0.5 mm MgCl2, 10 μm FAD, and various concentrations of BS (10−9 to 10−4 M). Assays were initiated by adding 100 μl of crude enzyme solution, and terminated after 30 min by the addition of 50 μl of 6 N H2SO4. The amount of acetolactate produced by the enzyme reaction was determined as described previously (Ray, 1984) with the following modifications. The acidified reaction mixtures were heated for 10 min at 60°C, after which 0.5 ml of 0.5% w/v creatine and 0.5 ml of 2-naphtol solution (5% in 2.5 N NaOH), both freshly prepared, were added to the solution, which was then heated for 10 min at 60°C. The absorbance at 525 nm was then determined using a double beam spectrophotometer.

Acknowledgements

We thank R. Aoto, E. Ozawa, A. Nagashii, Y. Nomura and F. Suzuki for their technical help. We also thank P. Maliga for providing pPZP202, H. Rothnie for editing, and B. Hohn for critical reading of the manuscript. This work was supported by the Ministry of Agriculture, Forestry and Fisheries of Japan. This work was also supported by a Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN) grant to S.T. A part of this study was also financially supported by the Budget for Nuclear Research of the Ministry of Education, Culture, Sports, Science and Technology, based on screening and counseling by the Atomic Energy Commission. M.E. was supported by a Japan Society for the Promotion of Science (JSPS) grant-in-aid fellowship.

Ancillary