A Universal Positive-Negative Selection System for Gene Targeting in Plants Combining an Antibiotic Resistance Gene and Its Antisense RNA^,

A positive-negative selection system based on a combination of an antibiotic resistance gene and its antisense RNA can enrich gene-targeted cells in rice.

Abstract

Gene targeting (GT) is a useful technology for accurate genome engineering in plants. A reproducible approach based on a positive-negative selection system using hygromycin resistance and the diphtheria toxin A subunit gene as positive and negative selection markers, respectively, is now available. However, to date, this selection system has been applied exclusively in rice (Oryza sativa). To establish a universally applicable positive-negative GT system in plants, we designed a selection system using a combination of neomycin phosphotransferaseII (nptII) and an antisense nptII construct. The concomitant transcription of both sense and antisense nptII suppresses significantly the level of expression of the sense nptII gene, and transgenic calli and plants become sensitive to the antibiotic geneticin. In addition, we were able to utilize the sense nptII gene as a positive selection marker and the antisense nptII construct as a negative selection marker for knockout of the endogenous rice genes Waxy and 33-kD globulin through GT, although negative selection with this system is relatively less efficient compared with diphtheria toxin A subunit. The approach developed here, with some additional improvements, could be applied as a universal selection system for the enrichment of GT cells in several plant species.

Gene targeting (GT; that is, modification of a specific DNA sequence in an endogenous gene by replacement of the target gene with a GT vector through homologous recombination [HR]) is a useful tool in both basic and applied studies. In earlier studies in higher plants, most transferred DNA (T-DNA) containing sequences homologous to the endogenous target gene was found to have integrated randomly into the plant genome through the nonhomologous end joining (NHEJ) pathway, leading to decreased GT frequency (0.01%–0.1% compared with random integration; Paszkowski et al., 1988; Offringa et al., 1990; Puchta et al., 1996). More recently, GT has developed to become a reproducible and general approach, at least in rice (Oryza sativa), through the use of a positive-negative selection system with the hygromycin resistance gene as a positive selection marker and the diphtheria toxin A subunit (DT-A) gene as a negative selection marker (Terada et al., 2002, 2007; Yamauchi et al., 2009; Moritoh et al., 2012; Ono et al., 2012; Ozawa et al., 2012; Dang et al., 2013; Tamaki et al., 2015). Toxic to rice plants, expression of DT-A is also toxic in some dicots, such as Arabidopsis (Arabidopsis thaliana; Czakó et al., 1992; Thorsness et al., 1993; Day et al., 1995; Nilsson et al., 1998; Tsugeki and Fedoroff, 1999; Weijers et al., 2003), tobacco (Nicotiana tabacum; Czakó et al., 1992; Day et al., 1995; Twell, 1995; Uk Kim et al., 1998), and Brassica campestris (Lee et al., 2003). However, successful examples of targeted gene modification through GT with a positive-negative selection system using DT-A have not been reported in higher plants other than rice, even in model plants, such as Arabidopsis or tobacco. This raises the possibility that DT-A negatively affects the growth of GT cells or the nontransformed cells surrounding DT-A-transformed cells because of its strong toxicity to higher plant cells other than rice. Alternatively, transient expression of the DT-A gene from the GT vector before T-DNA integration into the plant genome might kill potential GT cells, leading to their loss.

Previous studies have shown that expression of an antisense transcript of the neomycin phosphotransferaseII (nptII) gene in transgenic tobacco-carrying sense nptII expression cassettes led to a significant reduction in sense nptII gene transcripts and reduced kanamycin resistance, suggesting that a combination of nptII and antisense neomycin phosphotransferaseII (antinptII) genes could be utilized as a positive-negative selection system for GT experiments (Xiang and Guerra, 1993). In our previous work, transgenic rice calli expressing nptII under the control of the Cauliflower mosaic virus (CaMV) 35S promoter but not the nopaline synthase promoter could be selected on medium containing 35 mg L⁻¹ of geneticin (G418), suggesting a narrow range of optimal conditions for G418 selection. Thus, we hypothesized that an effective positive-negative selection system could be established by the use of nptII under the control of the CaMV35S and antinptII under the control of the rice elongation factor 1a (Pef) or maize (Zea mays) polyubiquitin1 promoter (Pubi), which confer higher levels of gene expression than the CaMV35S promoter in rice calli as positive and negative selectable markers, respectively. With the goal of establishing a more generally applicable and more publicly acceptable GT approach in plants, this study investigated whether concomitant expression of the antinptII gene in nptII-expressing transgenic rice would function as a negative selection marker in rice. Furthermore, this approach was also applied successfully to knockout of the endogenous waxy gene, encoding a granule-bound starch synthase (Wang et al., 1995), and the 33-kD globulin (Glb33) gene, encoding glyoxalase I (defined as a major allergen of rice; Usui et al., 2001).

RESULTS AND DISCUSSION

An Antisense nptII Transcript Suppresses Transcription of the nptII Gene, Affecting the G418-Resistant Phenotype of Transgenic Rice

One-week-old rice calli were infected with Agrobacterium tumefaciens harboring pZDnptIIGFP containing CaMV35S promoter::nptII and rice actin promoter::GFP expression cassettes or pZDnptIIGFPantinptII containing CaMV35S promoter::nptII, rice actin promoter::GFP, and maize Pubi::antinptII expression cassettes (Fig. 1A). Three-day cocultured calli were transferred and selected on N6D medium containing 35 mg L⁻¹ G418 for 4 weeks. The emergence of G418-resistant calli was decreased in pZDnptIIGFPantinptII transgenic calli compared with pZDnptIIGFP transgenic calli (Fig. 1B; Table I), whereas northern-blot analysis (Fig. 1C) revealed that nptII mRNA levels were greatly reduced in pZDnptIIGFPantinptII transgenic calli, which could eventually proliferate on selection medium containing G418. G418-resistant pZDnptIIGFP and pZDnptIIGFPantinptII-transformed calli were transferred to shoot regeneration medium containing 35 mg L⁻¹ G418. Hardly any G418-resistant regenerated plants were obtained from pZDnptIIGFPantinptII-transformed calli (Fig. 1D; Table I).

Figure 1.

Effect of antinptII expression on transcript levels of the sense nptII gene and the G418-resistant phenotype in rice. A, Structure of pZDnptIIGFP and pZDnptIIGFPantinptII vectors. The pZDnptIIGFP vector carries a CaMV 35S promoter (P_35S)::nptII expression cassette and a Pact::s65tGFP expression cassette. In addition, pZDnptIIGFPantinptII carries a Pubi::antinptII expression cassette. B, G418-resistant calli 14 d after onset of selection. C, Transcript levels of the nptII gene in pZDnptIIGFP and pZDnptIIGFPantinptII transgenic rice calli. Northern-blot analyses were performed with digoxygenin-labeled antinptII RNA probes and 10 μg of total RNA extracted from pZDnptIIGFP and pZDnptIIGFPantinptII transgenic rice calli. D, G418-resistant regenerated plants 1 month after transfer to regeneration medium. LB, Left border; RB, right border.

Table I. Suppressive effect of antinptII expression on G418-resistant phenotypes in transgenic rice.

—, Transgenic calli were not regenerated to plants.

Experiment	pZDnptIIGFP			pZDnptIIGFPantinptII
	A. tumefaciens-Infected Calli	Lines Successful in Clonal Propagation	Lines Successfully Regenerated	A. tumefaciens-Infected Calli	Lines Successful in Clonal Propagation	Lines Successfully Regenerated
	No.	No. (%)		No.	No. (%)
1	62	26 (41.9)	—	61	10 (16.4)	—
2	60	36 (60.0)	—	60	12 (20.0)	—
3	71	39 (54.9)	12 (30.8)	67	11 (16.4)	0 (0)
4	24	15 (62.5)	10 (66.7)	24	12 (50.0)	1 (8.3)

These data showed that expression of antinptII under the control of the maize Pubi had obvious effects on the transcript levels of both the sense nptII gene driven by the CaMV35S promoter and the G418-resistant phenotype in rice calli and plants, suggesting that the combination of CaMV35S promoter::nptII with maize Pubi::antinptII expression cassettes—named the nptII-antinptII system—could be used as a positive-negative selection system for GT experiments.

GT with Positive-Negative Selection Using the nptII-antinptII System

We next attempted to modify endogenous genes through GT with the nptII-antinptII system to apply this unique positive-negative selection system to identify GT cells. The Waxy gene encoding granule-bound starch synthase was chosen as the first target for knockout through GT, because two groups have already been successful in modifying this gene through GT using positive-negative selection (Terada et al., 2002; Ozawa et al., 2012). To evaluate the selection efficiency of the nptII-antinptII system, we generated two types of GT vector: pANwaxy and pKODwaxy. Both vectors carry a negative selection marker encoding either antinptII or the DT-A gene under the control of the maize Pubi and Pef promoter at both sides of the T-DNA region and a 6.0-kb fragment containing a waxy-encoding region with a rice actin terminator (Tact) and CaMV35S promoter::nptII expression cassette as a positive selectable marker in the first intron (Fig. 2A). Four-week-old rice calli were inoculated with A. tumefaciens harboring pANwaxy or pKODwaxy and selected for 4 weeks on N6D medium containing G418. In total, 381 G418-resistant calli were obtained from 2,948 pieces of calli transformed with the pKODwaxy vector. PCR analysis with the primers shown in Figure 2A revealed that nine callus lines (pKODwaxy-A1, pKODwaxy-A2, and pKODwaxy-B1–pKODwaxy-B7) were identified as GT candidate calli from among 381 G418-resistant calli, giving a ratio of GT candidate calli to G418-resistant calli of 2.36% (Table II), in agreement with previous findings (Terada et al., 2007; Shimatani et al., 2014). In contrast to calli transformed with pKODwaxy, a large number of G418-resistant calli were propagated from one piece of callus infected with A. tumefaciens harboring pANwaxy. In the first transformation experiment, an enormous number of G418-resistant calli lines was obtained from 800 pieces of calli transfected with pANwaxy. Among them, 840 calli lines were selected randomly to identify GT candidates by PCR analysis (Table II). In the second transformation experiment, 737 G418-resistant calli were obtained from 37 pieces of calli transfected with pANwaxy and subjected to PCR screening for isolation of GT candidates (Table II). However, we selected only two or one (pANwaxy-A1 and pANwaxy-A2 or pANwaxy-B) callus lines as GT candidates by PCR from 840 and 737 pieces of G418-resistant calli (0.24% and 0.14%), respectively (Table II). GT candidate calli were transferred to shoot regeneration medium, and the regenerated plants were analyzed further.

Figure 2.

Knockout of rice waxy gene through GT with positive-negative selection. A, Schematic representation of the waxy gene knockout strategy through GT with positive-negative selection using the nptII-antinptII system (pANwaxy) or nptII and DT-A (pKODwaxy). Lines 1 and 4 show the genomic structure of the wild-type waxy gene and the knockout waxy gene locus through GT, respectively. The GT vectors pANwaxy and pKODwaxy carry antinptII or DT-A under the control of the Pubi or Pef, respectively, as a negative selection marker and a 6.0-kb waxy coding region (gray boxes) containing a Tact and nptII gene under the control of the CaMV35S promoter (P_35S) as a positive selection marker in the first intron of the waxy gene. Black arrowheads represent primers for PCR screening to identify GT candidates. White arrowhead indicates the primer used to analyze segregation of the modified waxy gene in T1 progeny. The numbers between black arrowheads indicate the lengths of the PCR fragments. LB, Left border; RB, right border. B to D, Southern-blot analysis with probe 1 (B), probe 2 (C) and probe 3 (D) shown in A using EcoRV-digested genomic DNA of wild-type (Wt) plants and the regenerated plants of four independent lines with pANwaxy-A (pANwaxy-A1 and pANwaxy-A2), pANwaxy-B, pKODwaxy-A (pKODwaxy-A1 and pKODwaxy-A2), and pKODwaxy-B (pKODwaxy-B1–pKODwaxy-B4). M, Molecular weight marker.

Table II. GT experiments targeting the waxy locus.

Vector and Experiment	A. tumefaciens-Infected Calli	G418-Resistant Calli	Targeted Callia
	No.		No. (%)
pANwaxy
A	800	>840b	2 (0.24)
B	37	737	1 (0.14)
pKODwaxy
A	1,906	291	2 (0.69)
B	1,042	90	7 (7.78)

^aThe GT frequency was calculated as the ratio of targeted calli to G418-resistant calli.

^bThese calli were subjected to PCR analysis, although more than 840 calli were obtained from A. tumefaciens-infected calli.

To determine whether true GT events had occurred at the waxy locus in GT candidates, Southern-blot analysis was performed with EcoRV-digested genomic DNA isolated from regenerated plants derived from candidate GT calli. Using probe 1, a 13.2-kb band derived from the wild-type waxy locus and a 9.6-kb band expected to be derived from the modified waxy locus through true GT were detected in pANwaxy-A2, pKODwaxy-A1, pKODwaxy-A2, pKODwaxy-B1, pKODwaxy-B2, pKODwaxy-B3, and pKODwaxy-B4 (Fig. 2B). However, a band of 13.2 kb and a band with a molecular weight higher than 9.6 kb were observed in pANwaxy-A1 and pANwaxy-B (Fig. 2B). Southern-blot analysis with probe 2 revealed 7.4-kb fragments derived from the waxy knockin allele in all GT candidate plants (Fig. 2C). In addition to 7.4-kb fragments, one or more additional bands were detected in pANwaxy-A1, pANwaxy-A2, and pANwaxy-B (Fig. 2C), suggesting that one or more T-DNA fragments were integrated simultaneously at other loci through NHEJ. A fragment containing the wild-type waxy locus and a 7.4-kb fragment from the modified waxy locus through true GT were observed in all GT candidate plants (Fig. 2D). These results suggest that true GT events had occurred at one allele of the waxy gene in pANwaxy-A2, pKODwaxy-A1, pKODwaxy-A2, pKODwaxy-B1, pKODwaxy-B2, pKODwaxy-B3, and pKODwaxy-B4. One-sided invasion, which arises from one HR event and another NHEJ event at the target locus, might be generated at one allele of the waxy locus. Alternatively, ectopic GT, resulting from ectopic integration of a recombinant molecule produced by HR between the randomly integrated GT vector and a copy of the target gene, was thought to have occurred elsewhere in the genome in pANwaxy-A1 and pANwaxy-B. Such imprecise GT events have been reported in Arabidopsis and rice (Offringa et al., 1993; Hanin et al., 2001; Hohn and Puchta, 2003; Endo et al., 2006, 2007).

To assess the generality of this approach, the Glb33 gene encoding glyoxalase I (defined as a rice allergen) was disrupted through GT with positive-negative selection using DT-A or nptII-antinptII as the negative selectable marker. We inoculated rice calli with A. tumefaciens-harboring pANglb or pKODglb vector. These constructs carry antinptII (pANglb) or DT-A (pKODglb) gene expression cassettes flanking the 6.3-kb fragment encoding the Glb33 gene containing a replacement of the 451-bp fragment including the third exon and a portion of the fourth exon with a Tact and the nptII expression cassette as a positive selectable marker (Fig. 3A). In total, 1,460 and 93 G418-resistant calli were obtained from 188 and 2,715 pieces of calli for the pANglb and pKODglb constructs, respectively (Table III). Three (pANglb-A1, pANglb-A2, and pANglb-B) and two (pKODglb-A and pKODglb-B) calli lines were identified as GT candidates in G418-resistant calli transformed with pANglb and pKODglb, respectively, by PCR analysis (GT candidate calli:G418-resistant calli: 0.21% and 2.15%, respectively; Table III). Plants could be regenerated from GT candidate calli of pANglb-A1, pANglb-A2, pKODglb-A, and pKODglb-B but not pANglb-B. Southern-blot analysis using genomic DNA digested with SpeI and XbaI revealed that a true GT event had occurred in one allele of the Glb33 locus in all GT candidates (Fig. 3, B–D). However, one or two additional bands were detected by all probes in pANglb-A2, suggesting that at least one copy of T-DNA was integrated at another locus concomitant with true GT at the Glb33 locus (Fig. 3, B–D). Nevertheless, our findings indicate that positive-negative selection using the nptII-antinptII system had been successful in identifying transgenic calli in which true GT had occurred.

Figure 3.

Knockout of the rice Glb33 gene through GT with positive-negative selection. A, Strategy for knockout of the Glb33 gene through GT with positive-negative selection using the nptII-antinptII system (pANglb) or nptII and DT-A (pKODglb). Lines 1 and 4 show the genomic structure of the wild-type Glb33 gene and knockout Glb33 gene locus through GT, respectively. The GT vectors pANglb and pKODglb carry an antinptII (Pubi::antinptII and Pef::antinptII) or DT-A (Pubi::DT-A and Pef::DT-A) expression cassette, respectively, as a negative selection marker and a 6.3-kb Glb33 coding region (gray boxes) containing Tact and P35S::nptII expression cassette as a positive selection marker instead of the third exon of the Glb33 gene. Black arrowheads represent primers for PCR screening to identify GT candidates, and the white arrowhead indicates the primer used to analyze segregation of the modified Glb33 gene in T1 progeny. The numbers between black arrowheads indicate the lengths of the PCR fragments. LB, Left border; RB, right border. B to D, Southern-blot analysis with probe 1 (B), probe 2 (C), and probe 3 (D) shown in A using SpeI/XbaI-digested genomic DNA of wild-type (Wt) plants and the regenerated plants of four independent lines with pANglb-A (pANglb-A1 and pANglb-A2), pKODglb-A, and pKODglb-B. M, Molecular weight marker.

Table III. GT experiments targeting the Osglb33 locus.

Vector and Experiment	A. tumefaciens-Infected Calli	G418-Resistant Calli	Targeted Callia
	No.		No. (%)
pANglb
A	87	820	2 (0.26)
B	101	640	1b (0.15)
pKODglb
A	1,238	44	1 (2.27)
B	1,477	49	1 (2.00)

^aThese calli were subjected to PCR analysis, although more than 840 calli were obtained from A. tumefaciens-infected calli.

^bRegenerated plants could not be obtained.

Segregation of the Targeted Allele in T1 Progeny

To test whether the target gene was modified exactly at its endogenous locus through GT with the nptII-antinptII system, inheritance of the modified target gene to T1 progenies was evaluated by PCR analysis. After GT, both the waxy and Glb33 genes segregated according to a Mendelian ratio in T1 progeny derived from all GT candidate T0 plants, regardless of which negative selectable marker had been used (Supplemental Table S1). These data indicate that the GT vector was integrated into the waxy locus by one-sided invasion in pANwaxy-A1 and pANwaxy-B, because the T1 progenies from these GT candidates could segregate homozygous for the modified waxy gene allele. If the ectopic GT event had occurred in pANwaxy-A1 and pANwaxy-B, the GT-modified waxy gene would not have segregated in the T1 progenies. Furthermore, Southern-blot analysis with genomic DNA extracted from T1 plants of pANwaxy-A2, pKODwaxy-A1, and pKODwaxy-A2 revealed that disruption of the waxy gene through GT was inherited stably to T1 progenies (Fig. 4, A–C; Supplemental Fig. S1). Consistent with these results, we confirmed that the modified Glb33 alleles were able to segregate in pANglb-A1 and pKODglb-A progenies (Supplemental Fig. S2). Among 14 plants of pANwaxy-A2 T1 progenies, an extra band detected by probe 2 was found in plants either heterozygous or homozygous for the modified waxy allele (Fig. 4B). Thus, random integration of T-DNA was thought to have occurred at a region adjacent to the waxy locus. In addition, we confirmed that seeds of plants carrying a homozygous deletion of the waxy gene through GT stained light brown by iodine staining; the representative waxy phenotype of seeds is shown in Figure 4D.

Figure 4.

Southern-blot analysis and phenotype of T1 plants derived from GT candidate pANwaxy-A2. A to C, Southern-blot analysis with probe 1 (A), probe 2 (B), and probe 3 (C) shown in Figure 2A using EcoRV-digested genomic DNA of wild-type (Wt) plants and 14 T1 plants of pANwaxy-A2 (wild-type waxy gene, line nos. 1–4 [+/+]; heterozygous knockout waxy gene, line nos. 5–12 [+/−]; and homozygous knockout waxy gene, line nos. 13 and 14 [−/−]). M, molecular weight marker. D, The waxy phenotype of wild-type (left) and T1 seeds of pANwaxy-A2 (wild-type [+/+] or heterozygous knockout [+/−] waxy gene [center]; and homozygous knockout [−/−] waxy gene [right]) stained with iodine.

Taken together, a one-sided invasion, ectopic GT, or concomitant integration of T-DNA at another locus was found in GT candidate calli introduced by pANwaxy and pANglb, suggesting that the pressure of negative selection by antinptII under the control of the Pef promoter or maize Pubi was not sufficient to kill the cell upon random integration of the GT vector. This notion was supported by the abundance of G418-resistant calli with positive-negative selection using the nptII-antinptII system. It has been reported that expression of a small hairpin RNA rather than the concomitant expression of sense and antisense RNA was able to effectively suppress target gene expression by RNA interference in Arabidopsis (Chuang and Meyerowitz, 2000). Therefore, the utilization of a small hairpin RNA targeting the nptII gene as a negative selection marker might be more effective for the establishment of an efficient positive-negative selection system in rice.

Cytosine deaminase (codA) derived from Escherichia coli converts 5-fluorocytosine into the toxic 5-fluorouracil (Perera et al., 1993) and has been used as a conditional negative selection marker to modify endogenous genes through positive-negative selection-mediated GT in Lotus japonicas and Arabidopsis (Thykjaer et al., 1997; Gallego et al., 1999; Xiaohui Wang et al., 2001). However, ectopic GT plants but not true GT plants could be obtained with GT using the codA gene as a negative selectable marker (Thykjaer et al., 1997; Gallego et al., 1999; Xiaohui Wang et al., 2001). Recently, we reported that a mutant type of codA with one amino acid change, Asp to Ala at amino acid 314 (D314A; Mahan et al., 2004), could act as a more effective negative selection marker than wild-type codA in rice and could be used as a negative selection marker for enriching GT cells (Osakabe et al., 2014).

In conclusion, this study shows the successful use of the nptII-antinptII system to modify an endogenous gene through GT in rice. However, the negative selection pressure of antinptII is lower than that of DT-A; thus, additional improvement will be required for future applications. The properties of this system will be useful for GT experiments in plant species that are recalcitrant to genetic transformation, because unlike with DT-A, rarely emerged GT cells might not be lost upon transient expression of the antinptII negative selection marker gene. In addition, we have been successful in establishing a technique that allows the introduction of desired mutations into target genes through GT and subsequent positive marker excision from the GT locus using the piggyBac transposon without leaving any dispensable sequences (Nishizawa-Yokoi et al., 2015). GT with positive-negative selection and piggyBac-mediated marker excision permits the introduction of a desired modification into any gene, at least in rice plants. It has also been reported recently that the piggyBac transposon is also active in tobacco plants (Johnson and Dowd, 2014), suggesting that marker excision through piggyBac transposition might be generally applicable to several plant species. Because nptII is one of the selection marker genes most widely available in plants, the positive-negative selection with the nptII-antinptII system could well become a universal selection system for the enrichment of GT cells in many plant species. To verify the generality and effectiveness of our positive-negative GT system with nptII-antinptII or codA gene in several plant species, we plan to apply this approach to enrich GT cells in tobacco plants.

MATERIALS AND METHODS

Vector Construction

The coding region of nptII and antinptII was amplified by PCR using the primers 5′-tctagaatggggattgaacaagatgga-3′ (XbaI site in italics) and 5′-actagtgagctctcagaagaactcgtcaagaag-3′ (SacI site in italics) for nptII and 5′-cacctctagagacgtcccgggtcagaagaactcgtcaagaa-3′ (XbaI site in italics) and 5′-gagctccctaggagcgctcatggggattgaacaagatgg-3′ (SacI site in italics) for antinptII and cloned into pCR2.1 vector using TOPO Cloning Methods (Life Technologies). A 0.8-kb nptII or antinptII fragment was cloned into the XbaI/SacI site between the CaMV 35S promoter (P35S) and the nopaline synthase terminator (Tnos) in pENTR(L1–L4) or between the maize (Zea mays) Pubi and the Arabidopsis (Arabidopsis thaliana) ribulose-bisphosphate carboxylase small-subunit gene terminator (TrbcS) in pENTR(L3–L2), respectively. For construction of pZDnptIIGFP and pZDnptIIGFPantinptII, a cassette of P35S::nptII::Tnos in pENTR(L1–L4), rice (Oryza sativa) actin promoter::GFP::Tact in pENTR(R4–R3), Tact in (L3–L2), or Pubi::antinptII::TrbcS in pENTR(L3–L2) was combined and cloned into the pZD202Hyg vector (Nishizawa-Yokoi et al., 2014) using Multisite Gateway (Life Technologies).

The binary vector pAN was constructed as follows. A 0.8-kb antinptII fragment was cloned into the XbaI/SacI site between the Pef and the Arabidopsis polyA-binding protein terminator (Tpab) in pENTR(L3–L2). Cassettes representing Pef::antinptII::Tpab, Pubi::antinptII::TrbcS, and attR1-ccdB-attR2 were cloned sequentially into the HindIII/PacI, AscI/BsrGI, and SpeI/SacII sites in pPZP200MCS (Nishizawa-Yokoi et al., 2014), respectively, yielding pAN. A 3.0-kb waxy gene fragment containing 2.7 kb of upstream 5′ untranslated region, first exon, and first intron was digested with BamHI/PacI and cloned into pENTR(L1–L4), yielding pENTR(L1–L4)waxyBP. A 3.0-kb waxy gene fragment spanning the 1st to the 10th intron was amplified using the primers 5′-caccggcgcgcctctagatcttgtgttcaactctcg-3′ (XbaI site in italics) and 5′-ttaattaaggttccttccctaatattcgag-3′ (PacI site in italics) and cloned into the XbaI/PacI site in pENTR(L3–L2), yielding pENTR(L3–L2)waxyXP. A 3.2-kb Glb33 gene fragment spanning the 2.1-kb upstream region to the second exon was amplified using the primers 5′-taggcgcgccgattgagcaggcaaaggaca-3′ (AscI site in italics) and 5′-gcttaattaaccatacttaatggtgcgatc-3′ (PacI site in italics) and cloned into the AscI/PacI site in pENTR(L1–L4), yielding pENTR(L1–L4)Glbf4-r4. A 3.1-kb Glb33 gene fragment spanning the fourth exon to 0.9 kb downstream of the 3′ untranslated region was amplified using the primers 5′-taggcgcgccatttcgctattgcaactgag-3′ (AscI site in italics) and 5′-gcttaattaatatgactccagcccaactct-3′ (PacI site in italics) and cloned into the AscI/PacI site in pENTR(L3–L2), yielding pENTR(L3–L2)Glbf5-r5. A 4.1-kb fragment containing an nptII expression cassette (Tact::P35S::nptII::rice heat shock 17.3 protein terminator) was cloned into the AscI/PacI site in pENTR(R4–R3), yielding pENTR(R4–R3)TactP35SnptIIThsp. pENTR(L1–L4)waxyBP, pENTR(R4–R3)TactP35SnptIIThsp, and pENTR(L3–L2)waxyXP were combined and cloned into pAN or pKOD4 (Nishizawa-Yokoi et al., 2015) vector, yielding pANwaxy or pKODwaxy, respectively. pENTR(L1–L4)Glbf4-r4, pENTR(R4–R3)TactP35SnptIIThsp, and pENTR(L3–L2)Glbf5-r5 were combined and cloned into the pAN or pKOD4 vector, yielding pANglb or pKODglb, respectively.

Agrobacterium tumefaciens-Mediated Transformation

Agrobacterium tumefaciens-mediated transformation of rice ‘Nipponbare’ calli was performed as described previously (Toki, 1997; Toki et al., 2006). GT transformation was performed according to our previous studies (Nishizawa-Yokoi et al., 2015). Calli transformed with A. tumefaciens harboring pANwaxy, pKODwaxy, pANglb, and pKODglb were selected on callus induction (N6D) medium solidified with 0.8% (w/v) Bactoagar (Becton, Dickinson and Company) containing 35 mg L⁻¹ G418 and 25 mg L⁻¹ meropenem (Wako Pure Chemical Industries). Candidate GT calli were transferred to regeneration medium with 25 mg L⁻¹ meropenem, and shoots arising from callus were transferred to Murashige and Skoog medium (Murashige and Skoog, 1962) without phytohormones.

RNA Extraction and Northern-Blot Analysis

Total RNA was extracted from rice calli using an RNeasy Plant Mini Kit (QIAGEN). Total RNA (10 μg) was fractionated in a 1.2% (w/v) agarose gel containing 2% (v/v) formaldehyde. Northern-blot analysis and the synthesis of specific RNA probes for antinptII were performed with a Digoxigenin Northern Starter Kit (Roche Diagnostics) according to the manufacturer’s protocol.

Screening of GT Candidates by PCR and Southern-Blot Analysis

After a 4-week selection period, genomic DNA was extracted from small pieces of G418-resistant calli transformed with A. tumefaciens harboring a GT vector using Agencourt Chloropure (Bechman Coulter) according to the manufacturer’s protocol. PCR amplifications were performed with KOD FX or KOD FX neo (TOYOBO) using primer sets as follows: for 5′ amplification of waxy, waxy GT-F: 5′-aatccaatccaatccatcacataatgcaag-3′ and Tact-R: 5′-ctgacgatgagaatatatctgatgctgtga-3′; for 3′ amplification of waxy, Thsp17.3-F: 5′-acatacccatccaacaatgttcaatccctt-3′ and waxy GT-R: 5′-atgtggaaaccagtcttgccttcgatga-3′; for 5′ amplification of Glb33, glb GT-F: 5′-gtatggtagtgtactatatccctggtcaaatctgc-3′ and Tact-R; and for 3′ amplification of Glb33, Thsp17.3-F and glb GT-R: 5′-agagggttgcgtatctttggagtcg-3′.

For Southern-blot analysis, genomic DNA was extracted from GT candidate plants derived from PCR-positive calli using the Nucleon Phytopure Extraction Kit (GE Healthcare) according to the manufacturer’s protocol. Genomic DNA (2 μg) was digested with EcoRV (for the waxy locus) or SpeI/XbaI (for the Glb33 locus) and fractionated in a 0.7% (w/v) agarose gel. Southern-blot analysis was performed according to the Digoxigenin Application Manual (Roche Diagnostics). Specific DNA probes for waxy and Glb33 were synthesized with a PCR Digoxigenin Probe Synthesis Kit (Roche Diagnostics) using the primers waxy probe 1 (5′-ggcgaagtaactccagatcg-3′ and 5′-tcggattcgttgacaagttaagg-3′), waxy and glb33 probe 2 (nptII-specific primers: 5′-ttgaacaagatggattgcac-3′ and 5′-ggcatcgccatgtgtcacga-3′), waxy probe 3 (5′-cccatgtggaaaccagtctt-3′ and 5′-cttgtggagttagccggaag-3′), glb33 probe 1 (5′-ggatgggatggtgagatcat-3′ and 5′-ccgcagcttatgaggactct-3′), and glb33 probe 3 (5′-cccatgccttgctattgttt-3′ and 5′-ccatcagggtcaaggaaaga-3′).

Segregation Analysis of the Targeted Allele in T1 Progeny Using PCR

Genomic DNA was extracted from T1 progeny plants derived from self-pollinating pANwaxy-A2, pKODwaxy-A1, pKODwaxy-A2, pANglb-A1, and pKODglb-A T0 plants and subjected to PCR analysis using primer sets as follows: for assessment of the insertion of positive selection marker into the waxy gene through GT, Thsp17.3-F/waxy GT-R; for amplification of a 4.1-kb wild-type waxy gene fragment, waxy-F (5′-agagggggagagagagatcg-3′)/waxy GT-R; for assessment of the insertion of positive selection marker into the glb33 gene through GT, Thsp17.3-F/glb GT-R; and for amplification of the 4.2-kb wild-type glb33 gene fragment, glb-F (5′-ggtccaagaactccagataagagag-3′)/glb GT-R. T1 plants carrying the wild-type waxy gene contained only a 4.1-kb wild-type waxy gene fragment. Heterozygous or homozygous waxy knockout plants through GT contained both the 4.1-kb wild-type waxy gene fragment and the 3.7-kb knockout waxy gene fragment or only the knockout waxy gene fragment, respectively. The 4.2-kb wild-type glb33 gene fragment was detected in T1 plants carrying the wild-type glb33 gene and the heterozygous glb33 knockout and not the homozygous glb33 knockout, and the 3.8-kb knockout glb33 gene fragment was observed in heterozygous and homozygous glb33 knockout plants but not in wild-type plants.

Iodine Staining

The T1 seeds of plants carrying wild-type, heterozygous, or homozygous deletions of the waxy gene through GT were hand cut with a razor blade. The bottom one-half of the seed was stained with 20-fold diluted Lugol’s solution, and the upper one-half of the seed carrying the scutellum was inoculated onto Murashige and Skoog medium. The GT genotype for the waxy gene was identified by PCR using the primers described above.

Sequence data from this article can be found in the Rice Genome Annotation Project (http://rice.plantbiology.msu.edu/) data libraries under the following accession numbers: rice waxy (LOC_Os06g04200) and rice Glb33 (LOC_Os08g09250).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Molecular analysis of T1 plants derived from pKODwaxy-A1 and pKODwaxy-A2.

• Supplemental Figure S2. Molecular analysis of T1 plants derived form pANglb-A1 and pKODglb-A.

• Supplemental Table S1. Segregation analysis of GT plants in T1 progeny.

Glossary

G418

geneticin

GT

gene targeting

HR

homologous recombination

NHEJ

nonhomologous end joining

T-DNA

transferred DNA