H2O2-Based Method for Rapid Detection of Transgene-Free Rice Plants from Segregating CRISPR/Cas9 Genome-Edited Progenies

Genome-editing techniques such as CRISPR/Cas9 have been widely used in crop functional genomics and improvement. To efficiently deliver the guide RNA and Cas9, most studies still rely on Agrobacterium-mediated transformation, which involves a selection marker gene. However, several limiting factors may impede the efficiency of screening transgene-free genome-edited plants, including the time needed to produce each life cycle, the response to selection reagents, and the labor costs of PCR-based genotyping. To overcome these disadvantages, we developed a simple and high-throughput method based on visual detection of antibiotics-derived H2O2 to verify transgene-free genome-edited plants. In transgenic rice containing hygromycin phosphotransferase (HPT), H2O2 content did not change in the presence of hygromycin B (HyB). In contrast, in transgenic-free rice plants with 10-h HyB treatment, levels of H2O2 and malondialdehyde, indicators of oxidative stress, were elevated. Detection of H2O2 by 3,3′-diaminobenzidine (DAB) staining suggested that H2O2 could be a marker to efficiently distinguish transgenic and non-transgenic plants. Analysis of 24 segregating progenies of an HPT-containing rice plant by RT-PCR and DAB staining verified that DAB staining is a feasible method for detecting transformants and non-transformants. Transgene-free genome-edited plants were faithfully validated by both PCR and the H2O2-based method. Moreover, HyB induced overproduction of H2O2 in leaves of Arabidopsis, maize, tobacco, and tomato, which suggests the potential application of the DAB method for detecting transgenic events containing HPT in a wide range of plant species. Thus, visual detection of DAB provides a simple, cheap, and reliable way to efficiently identify transgene-free genome-edited and HPT-containing transgenic rice.


Introduction
The hygromycin phosphotransferase gene (HPT) from E. coli is a positive selection marker for plant transformation [1]. HPT confers resistance to hygromycin B (HyB) in transgenic plants by adding a phosphate to position seven of the destomic acid ring in HyB [1,2]. More importantly, HyB does not affect regeneration and fertility of transgenic plants [3]. Therefore, HPT is broadly used in the transformation of monocot plants, which usually show high levels of natural resistance to another antibiotic, kanamycin [4,5]. Plants effectively selected by using HPT include rice [6], maize [1], wheat [7,8], brachypodium [9,10], and barley [11][12][13].
HyB is an aminoglycoside antibiotic isolated from Streptomyces hygroscopicus [14] that has broad-spectrum activity against both prokaryotic and eukaryotic cells by interfering with translocation of the 70S ribosome and causing misreading of the mRNA template [15,16]. HyB significantly inhibits plant growth and development. Different plant genotypes exhibit variable HyB sensitivity. Culture media containing 50 mg L −1 HyB could fully inhibit the growth of rice callus [17] and cotyledon and leaves of Arabidopsis seedlings [18]; as low as 2.5 mg L −1 HyB was able to restrict the growth of maize cells [19].
Rice is a staple food crop feeding more than half the world's population [20]. To accelerate biotechnological applications, the rice transformation system was established more than two decades ago [6,21]. During rice transformation, HPT is used for screening putative transgenic events and screening massive transgenic progenies or homozygous transgenic lines, thereby replacing the costly and labor-intensive PCR-based or Southern blot analysis [22][23][24]. For efficient HyB screening, seeds are germinated on medium containing HyB [24,25]; however, the germination rate may be confusing in seeds with dormant or low germination vigor, which affects the determination of homozygosity.
CRISPR/Cas technology has emerged as a powerful and promising method to precisely modify the plant genome and efficiently generate transgene-free crops [26,27]. The introduction of a CRISPR/Cas-mediated genome-editing cassette into the plant genome allows for integrating the transgene into one locus and performing the editing at another locus. Therefore, traits can be segregated by sexual reproduction, generating progenies free of the transgene [26][27][28]. The transgene-free crops thus contain biallelic/homozygous mutations and are free of selection markers. A number of useful methods have been developed for screening biallelic/homozygous mutations. However, quick screening of marker-free transgenic plants from a genome-edited (GE) plant population remains a challenge because PCR-based screening is labor-intensive; furthermore, plants cannot survive on selection medium without selection markers. Thus, a reliable, inexpensive, and non-lethal selection method is needed to efficiently distinguish GE plants with or without selection markers.
Leaf painting assay has been used to facilitate the selection of transgenic plants tolerant to antibiotics or herbicides. In cotton, leaves of transgenic cotton treated with 750 mg/L kanamycin exhibited chlorosis after five to seven days and then necrotic patches after 10 days [29]. Transgenic rice [30] and maize [31] expressing the bar gene were able to tolerate the herbicide phosphinothricin. In transgenic maize, more than 95% of transgenic events can be verified by leaf painting assay, with results agreeing with PCR results [31]. Leaf painting assay is simple and efficient, but it takes almost one week to observe the wilt symptoms. Therefore, developing a simple, efficient, and rapid leaf painting assay is needed for high throughput screening of transgenic progenies.
3,3-Diaminobenzidine (DAB) staining is one of the most commonly used methods for H 2 O 2 detection. After being taken up by plants, DAB reacts with H 2 O 2 to form a dark-brown reaction product in the presence of peroxidase [32]. Our recent research indicated that HyB significantly and rapidly enhanced the accumulation of H 2 O 2 in rice leaves [33]. Taking advantage of the high production of H 2 O 2 in plants induced by HyB, we aimed to develop a simple and quick, selection-independent, H 2 O 2 -based assay system for identifying transgenic rice.
In the present study, transgenic and non-transgenic rice could be easily distinguished by the H 2 O 2 -based assay system. The visual selection method provides a quick and reliable way for screening transgene-free GE plants after genome editing in rice. Moreover, we found HyB-induced overproduction of H 2 O 2 in a wide range of plant species, so the H 2 O 2 DAB method may be applicable for efficiently distinguishing a wide range of transgenic and non-transgenic plants.

Hygromycin (HyB) Significantly Increased the Production of H 2 O 2 in Leaves of Rice Seedlings
To examine the effect of HyB on the accumulation of H 2 O 2 , leaf segments of rice seedlings were treated with or without 50 mg mL −1 HyB for various times. H 2 O 2 level increased with increasing treatment time. H 2 O 2 level slightly increased after 6-h treatment and significantly increased after 10-h treatment ( Figure 1).

Hygromycin (HyB) Significantly Increased the Production of H2O2 in Leaves of Rice Seedlings
To examine the effect of HyB on the accumulation of H2O2, leaf segments of rice seedlings were treated with or without 50 mg mL −1 HyB for various times. H2O2 level increased with increasing treatment time. H2O2 level slightly increased after 6-h treatment and significantly increased after 10h treatment (Figure 1).

Figure 1.
Temporal accumulation of hygromycin B (HyB)-induced hydrogen peroxide (H2O2) in rice leaves. Leaf segments of 2-week-old rice seedlings were incubated with or without 50 mg L -1 HyB for the indicated time. Data are mean ± SE (n = 4). * P < 0.05 compared with -HyB.

Accumulation of H2O2 in Transgenic Rice Overexpressing HPT
To examine levels of H2O2 in rice containing a hygromycin-detoxifying gene, we generated transgenic rice harboring HPT. HPT was overexpressed in three transgenic lines (OE-HPT-1, OE-HPT-2, OE-HPT-3), with no transcripts detected in non-transgenic wild-type (WT) plants ( Figure  2A,B). Cultivation of seeds on a hydroponic solution containing 50 mg L −1 HyB showed that the three transgenic lines germinated and grew normally, but germination was inhibited in seeds from WT plants ( Figure 2C). Numbers of germinated plants of WT and three transgenic lines were almost 100% without HyB treatment. In contrast, the germination rate of WT and the three transgenic lines was 0%, 72%, 68%, and 76%, respectively, when seeds were germinated on a hydroponic solution containing 50 mg L −1 HyB ( Figure 2D). Upon treatment with HyB, H2O2 level significantly increased only in non-transgenic rice seedlings and not in the three transgenic lines containing HPT ( Figure  3A). Then, we used DAB treatment to distinguish transgenic and non-transgenic rice. Without HyB treatment, leaf segments of both non-transgenic and transgenic rice showed very little brown color. Conversely, after DAB treatment, leaves of WT but not transgenic rice showed a dark brown color ( Figure 3B). Moreover, HyB treatment significantly increased the production of malondialdehyde in WT but not OE-HPT-1 plants ( Figure 3C), so HyB caused severe oxidative stress, thereby enhancing lipid peroxidation.

Accumulation of H 2 O 2 in Transgenic Rice Overexpressing HPT
To examine levels of H 2 O 2 in rice containing a hygromycin-detoxifying gene, we generated transgenic rice harboring HPT. HPT was overexpressed in three transgenic lines (OE-HPT-1, OE-HPT-2, OE-HPT-3), with no transcripts detected in non-transgenic wild-type (WT) plants (Figure 2A,B). Cultivation of seeds on a hydroponic solution containing 50 mg L −1 HyB showed that the three transgenic lines germinated and grew normally, but germination was inhibited in seeds from WT plants ( Figure 2C). Numbers of germinated plants of WT and three transgenic lines were almost 100% without HyB treatment. In contrast, the germination rate of WT and the three transgenic lines was 0%, 72%, 68%, and 76%, respectively, when seeds were germinated on a hydroponic solution containing 50 mg L −1 HyB ( Figure 2D). Upon treatment with HyB, H 2 O 2 level significantly increased only in non-transgenic rice seedlings and not in the three transgenic lines containing HPT ( Figure 3A). Then, we used DAB treatment to distinguish transgenic and non-transgenic rice. Without HyB treatment, leaf segments of both non-transgenic and transgenic rice showed very little brown color. Conversely, after DAB treatment, leaves of WT but not transgenic rice showed a dark brown color ( Figure 3B). Moreover, HyB treatment significantly increased the production of malondialdehyde in WT but not OE-HPT-1 plants ( Figure 3C), so HyB caused severe oxidative stress, thereby enhancing lipid peroxidation.

DAB-H 2 O 2 Method is Efficient for Detecting HPT-Containing Transgenic Rice
To test the accuracy of the DAB-H 2 O 2 method for distinguishing transgenic plants, we used RT-PCR and the DAB method to examine 24 T1 segregants from one transgenic plant carrying the HPT transgene (OE-HPT-1). On RT-PCR, HPT was detected in transgenic progenies but not null segregants ( Figure 4A). T1 segregants without HPT transcripts showed significant DAB-H 2 O 2 accumulation. Conversely, plants expressing HPT showed no significant DAB-H 2 O 2 accumulation ( Figure 4B). Thus, the DAB-H 2 O 2 method was as reliable as RT-PCR for detecting transgenic plants containing an HyB-resistant gene.

DAB-H2O2 Method is Efficient for Detecting HPT-Containing Transgenic Rice
To test the accuracy of the DAB-H2O2 method for distinguishing transgenic plants, we used RT-PCR and the DAB method to examine 24 T1 segregants from one transgenic plant carrying the HPT transgene (OE-HPT-1). On RT-PCR, HPT was detected in transgenic progenies but not null segregants ( Figure 4A). T1 segregants without HPT transcripts showed significant DAB-H2O2 accumulation. Conversely, plants expressing HPT showed no significant DAB-H2O2 accumulation ( Figure 4B). Thus, the DAB-H2O2 method was as reliable as RT-PCR for detecting transgenic plants containing an HyB-resistant gene.

Detection of Transgene-Free Genome-Edited Plants by DAB Method
To screen transgene-free genome-edited rice, the T1 population of Agrobacterium-mediated CRISPR/Cas9 rice was evaluated with the DAB method. The CRISPR/Cas9 binary vector encoded an HPT gene for transformation selection in the T0 generation [34,35]. After the CRISPR/Cas9 vector is used for gene editing at the target regions, the vector can be removed in the T1 segregation population. Four T1 genome-edited mutants (osrr6/osrr11#2-3-1, osrr6/osrr11#5-1-11, osrr9/osrr10#9-2-4, and osrr9/osrr10#10-3-6) from two transgenic events were used to examine the elimination of the transgene vector. The two transgenic events targeted multiple cytokinin two-component signaling type-A response regulators, OsRR6, OsRR11, OsRR9, and OsRR10. The mutations at the target region were verified by Sanger sequencing (Supplementary Table S1 [34,36]). The presence of the HPT selectable marker was visually selected by the DAB method and RT-PCR.
In osrr6/osrr11 genome-edited lines, single guide RNAs (sgRNAs) specific to editing conserved sequences of OsRR6 and OsRR11 were designed. In osrr9/osrr10 genome-edited lines, sgRNAs specific to editing conserved sequences of OsRR9 and OsRR10 were designed. No significant DAB-H2O2 accumulation was detected in leaves without HyB treatment ( Figure 5A). After HyB treatment,

Detection of Transgene-Free Genome-Edited Plants by DAB Method
To screen transgene-free genome-edited rice, the T1 population of Agrobacterium-mediated CRISPR/Cas9 rice was evaluated with the DAB method. The CRISPR/Cas9 binary vector encoded an HPT gene for transformation selection in the T0 generation [34,35]. After the CRISPR/Cas9 vector is used for gene editing at the target regions, the vector can be removed in the T1 segregation population. Four T1 genome-edited mutants (osrr6/osrr11#2-3-1, osrr6/osrr11#5-1-11, osrr9/osrr10#9-2-4, and osrr9/osrr10#10-3-6) from two transgenic events were used to examine the elimination of the transgene vector. The two transgenic events targeted multiple cytokinin two-component signaling type-A response regulators, OsRR6, OsRR11, OsRR9, and OsRR10. The mutations at the target region were verified by Sanger sequencing (Supplementary Table S1 [34,36]). The presence of the HPT selectable marker was visually selected by the DAB method and RT-PCR.

HyB-Induced Overproduction of H 2 O 2 Observed in Monocot and Dicot Plants
To examine whether H 2 O 2 can be significantly induced by HyB in other plants, we examined plants such as Arabidopsis, tobacco, tomato, and maize. DAB-H 2 O 2 staining showed that H 2 O 2 was highly induced in all tested plants after HyB treatment ( Figure 6). Tobacco showed the strongest accumulation of H 2 O 2 , followed by Arabidopsis and tomato. The monocot plant maize showed a slight enhancement of H 2 O 2 after HyB treatment. accumulation of H2O2, followed by Arabidopsis and tomato. The monocot plant maize showed a slight enhancement of H2O2 after HyB treatment.

Discussion
Genome-editing is a powerful tool for precision breeding in crops. Several countries will not regulate the use of genome-editing techniques in plants, which indicates that there will be more GE crops in the future. Therefore, developing a high-throughput screening platform for rapidly distinguishing transgenic or non-transgenic plants is required to efficiently generate transgene-free genome-edited plants. Our previous study showed that rice seedlings treated with HyB for 12 h markedly increased H2O2 level in non-transformants but not in transgenic rice containing the HPT gene [33]. In this study, we further verified that genome-edited rice progenies containing HPT can also be faithfully reflected by visual detection of the DAB-H2O2 complex in leaves and validated by RT-PCR, which indicates that the H2O2-based leaf painting assay is a rapid and reliable method. As compared with previous studies of cotton [29], rice [30], and maize [31], our H2O2-based method in rice largely decreased the time from five to seven days to 12 h, so the H2O2-based leaf painting assay is a high-throughput system for screening transgene-free genome-edited crops from segregating progenies.

Generation of Hygromycinb (HyB)-Resistant Transgenic Rice
The pPZP/HPH binary vector [21] was used for transformation of HyB-resistant rice plants. The plasmid pPZP/HPH was introduced into Agrobacterium tumefaciens strain EHA105, and embryogenic

Discussion
Genome-editing is a powerful tool for precision breeding in crops. Several countries will not regulate the use of genome-editing techniques in plants, which indicates that there will be more GE crops in the future. Therefore, developing a high-throughput screening platform for rapidly distinguishing transgenic or non-transgenic plants is required to efficiently generate transgene-free genome-edited plants. Our previous study showed that rice seedlings treated with HyB for 12 h markedly increased H 2 O 2 level in non-transformants but not in transgenic rice containing the HPT gene [33]. In this study, we further verified that genome-edited rice progenies containing HPT can also be faithfully reflected by visual detection of the DAB-H 2 O 2 complex in leaves and validated by RT-PCR, which indicates that the H 2 O 2 -based leaf painting assay is a rapid and reliable method. As compared with previous studies of cotton [29], rice [30], and maize [31], our H 2 O 2 -based method in rice largely decreased the time from five to seven days to 12 h, so the H 2 O 2 -based leaf painting assay is a high-throughput system for screening transgene-free genome-edited crops from segregating progenies.

Generation of Hygromycinb (HyB)-Resistant Transgenic Rice
The pPZP/HPH binary vector [21] was used for transformation of HyB-resistant rice plants. The plasmid pPZP/HPH was introduced into Agrobacterium tumefaciens strain EHA105, and embryogenic calli derived from immature seeds of Tainung 67 were transfected as described [21]. Putative transformed calli were selected on HyB (Invitrogen, Carlsbad, CA, USA). Regenerated transgenic plants were grown and self-pollinated for two generations. For generating CRISPR/Cas9-mediated mutants, 20-bp sgRNAs complementary to rice cytokinin signaling type-A response regulators, osrr6, osrr9, osrr10, and osrr11, were designed and cloned into a pAS3 binary vector harboring an sgRNA cassette, a maize UBQ10-drived Cas9, and a hygromycin selection marker for Agrobacterium-mediated transformation [34][35][36]. The transformed calli were first selected on HyB media and the surviving calli were regenerated to T0 plantlets. At the T1 generation, the GE target regions may segregate with CRISRP/Cas9 vectors. To identify the mutants at the T1 generation, the sgRNA target sequences were amplified by PCR and subjected to one-step PAGE [37], high-resolution melt analysis (HRM) [38], or Sanger sequencing [39].

HyB Treatment
Leaf segments of 1 cm from 2-week-old rice seedlings were cut and cultured in sterile distilled water containing HyB. Leaf segments without any treatment were controls. For Arabidopsis, leaves of 2-week-old seedlings were treated with or without HyB. For tomato, tobacco, and maize, leaf discs were used for HyB treatment. All leaf segments/discs were treated with 50 mg L −1 HyB, and plants were incubated in a growth chamber at 27 • C under light for 12 h. We used 10 leaf segments or leaf discs per sample, with 4 biological replicates for each treatment. Representative results are shown.

Visual Detection of H 2 O 2
H 2 O 2 was visually detected in leaves by using DAB as a substrate [40]. Leaf segments or leaf discs were treated with or without 50 mg L −1 HyB. After 12 h, leaves were first rinsed with distilled water and then supplied with DAB solution (1 mg mL −1 ) through the cut ends for 12 [41] with absorbance measured at 410 nm. The amount of H 2 O 2 was calculated by using a standard curve prepared with known concentrations of H 2 O 2 in some experiments, H 2 O 2 content was measured spectrophotometrically after a reaction with DAB solution [42]. The reaction mixture consisted of 2 mL of 50 mM phosphate buffer (pH 6.8) leaf extract supernatant and 1 mg mL −1 DAB solution. The blank reaction consisted of 50 mM phosphate buffer without leaf extract. The absorbance was measured at 465 nm. To quantify DAB staining results, brown areas of the DAB-H 2 O 2 reaction were scanned and measured with ImageJ (https://imagej.nih.gov/ij/). The ImageJ was used to record the grayscale values of all pixels within the brown areas of eight-bit images. An identical noise threshold was used for all analyses and was visually inspected for accurate representation of the particles.
Malondialdehyde, routinely used as an indicator of lipid peroxidation, was extracted with 5% (w/v) trichloroacetic acid, and the content was determined by the thiobarbituric acid reaction as described [43] and expressed on the basis of fresh weight. For protein determination, leaves were homogenized in 50 mM sodium phosphate buffer (pH 6.8). Extracts were centrifuged at 17,600× g for 20 min, and the supernatant was used for determining protein level by the Bradford method [44].

Statistical Analysis
Data are expressed as mean ± SE. Each experiment was carried out in 4 technical replicates and 3 biological replicates. One representative experiment is shown. Differences between measurements were analyzed by Student t test or Duncan's multiple range test. p < 0.05 was considered statistically significant.

Conclusions
We show that the notable difference in H 2 O 2 levels induced by HyB between non-transformant and transgenic rice harboring HPT can be a convenient system for distinguishing transgenic plants. Extraction of DNA or RNA is not necessary in this protocol. Instead, we used a simple DAB histochemical staining method, which was convenient, cheap, quick, and reliable for analysis of transgene-free genome-edited plants. The procedure was feasible for determining transgenic rice containing HPT.