New T-DNA binary vectors with NptII selection and RUBY reporter for efficient maize transformation and targeted mutagenesis

Dear Editor, Maize (Zea mays L.) is one of the most important cereal crops for food, feed, and biomaterials. Although genetic transformation protocols have been established for maize inbreds, their applications are limited to certain genotypes, and overall transformation frequencies remain low. One of the widely applicable approaches for enhancing plant transformation frequencies is using superbinary vectors or ternary helper plasmids that carry extra copies of essential Agrobacterium virulence genes to improve T-DNA delivery capacities (Komari et al. 1986; Anand et al. 2018; Zhang et al. 2019). Ternary helper plasmids allow T-DNA vector construction procedures simpler than that of superbinary vectors (Anand et al. 2018; Zhang et al. 2019; Kang et al. 2022), but they also require that the T-DNA vectors, which carry the gene of interest (GOI), must have the compatible origin of replications (ORIs) and different antibiotics resistance genes for bacterial selection. We recently demonstrated a rapid maize B104 transformation method (Kang et al. 2022) using a ternary helper plasmid and QuickCorn media recipes (Masters et al. 2020) to produce rooted transgenic plantlets in just 51 d. Compared with the conventional transformation method (Raji et al. 2018), which requires over 150 d to generate transgenic plantlets, the QuickCorn method saves 3 to 4 mo to reach the same. Because of the rapid in vitro tissue culture time, the effective plant selectable marker genes are critical for the success of the transformation. In the work of Kang et al. (2022), we observed as high as two-thirds of nontransgenic plants in bialaphos-resistant and regenerated plants. While the selection system of the bar gene and herbicide bialaphos works well in extended callus subculture and transformation protocol (Raji et al. 2018), it appears to be inefficient in eliminating nontransformed tissue fast enough in the QuickCorn method (Kang et al. 2022), resulting in high escape rates of nontransgenic plants. In this paper, we report new T-DNA binary vectors with neomycin phosphotransferase II (NptII) as a selectable marker gene for plant transformation (Fig. 1) that showed improved maize B104 transformation and CRISPR-Cas9-mediated mutagenesis efficiencies with low escape rates (<6%). A ternary vector system for Agrobacterium-mediated plant transformation. A) Schematic illustration of a ternary vector system consisting of 3 compatible plasmid vectors: a T-DNA binary vector, a helper plasmid, and a disarmed tumor-inducing (Ti) plasmid. The T-DNA binary vector carries GOI that will be transferred to the plant cells. The helper plasmid carries extra copies of essential virulence genes (VIR), whereas the disarmed Ti plasmid encodes all the necessary virulence (VIR) genes for the T-DNA transfer. All 3 plasmids must have compatible origins of replication to stably coexist in the same cells. C1, C2, pAT, Agrobacterium chromosomes C1 and C2, and the pAT plasmid. Empty T-DNA binary vectors with pVS1 ORI for transgene overexpression B) or CRISPR–Cas-mediated gene mutagenesis C). These vectors are compatible with ternary helper plasmids such as pKL2299 (Kang et al. 2022; Addgene #186332) with an RK2 ORI. pCBL101 (Addgene #199722), pCBL101.1 (Addgene #199724), and pCBL101gw (Addgene #199725) are identical, except that pCBL101 and pCBL101.1 have MCS, whereas pCBL101gw has Gateway attR1/R2 cassettes for LR clonase reaction for cloning a GOI into the vector. pCBL101.1 was generated by removing a BsaI restriction site within the pVS1 ORI and inserting new MCS with dual BsaI restriction sites for Golden Gate assembly. pKL2351 (Addgene #199721) is a Gateway destination vector (attR1/R2) and has a red fluorescent protein (mCherry) gene driven by the 35S promoter as a visible marker and a maize ubiquitin gene promoter (PZmUbi; Christensen et al. 1992) for transgene expression. All 3 vectors contain neomycin phosphotransferase II (NptII; Fraley et al. 1983) gene driven by PZmUbi as a plant selectable marker. D) T-DNA binary vectors used for maize B104 transformation. pCBL101-RUBY (left; Addgene #199723) carries the RUBY reporter (He et al. 2020) that contains betalain biosynthesis genes (CYP76AD1, DODA, and Glucosyl transferase). pKL2359 (right) carries a maize codon–optimized Cas9 from Streptococcus pyogenes (SpCas9) and a sgRNA1 targeting maize glossy2 gene. RB, T-DNA right border; LB, T-DNA left border; P35S, CaMV 35S promoter; CmR, chloramphenicol resistance gene; ccdB gene codes for toxin protein CcdB; pVS1, a broad host range ORI from Pseudomonas pVS1 plasmid for Agrobacterium; pBR322, a high copy number ORI for Escherichia coli; KmR, Kanamycin resistance gene; SpR, spectinomycin resistance gene. E) Summary of maize B104 transformation. F) Summary of CRISPR–Cas9-mediated Glossy2 mutagenesis. HO, homozygous; BI, biallelic; HT, heterozygous; MO, mosaic mutations. G) Scutella tissue of immature embryo, 3 d postinfection. H) Embryogenic callus tissue on maturation medium, 3 wk postinfection. I) Immature shoots on maturation medium, 5 wk postinfection. J) Transgenic plantlets, 7 wk postinfection. The immature embryo and embryogenic callus tissues G and H) were infected with Agrobacterium LBA4404Thy- strain carrying a binary vector pCBL102-RUBY (Supplemental Fig. S1) and other images were taken from transgenic plants produced by pCBL101-RUBY D). Our ternary vector system consists of 4 major components as illustrated in Fig. 1A; (i) Agrobacterium C1 and C2 chromosomes plus pAT plasmid (Goodner et al. 2001; Wood et al. 2001); (ii) disarmed Ti plasmid; (iii) the helper plasmid (pKL2299, Kang et al. 2022, Addgene #186332); and (iv) the T-DNA binary vector (this work). The Agrobacterium strain used in the work is LBA4404Thy- (Masters et al. 2020), a thymidine auxotrophic strain (kindly provided by Dr. William Gordon-Kamm of Corteva Agriscience) that can only survive in media supplemented with thymidine. The helper plasmid, pKL2299 (Kang et al. 2022), with RK2 ORI can be readily combined with T-DNA binary vectors with compatible ORIs such as pVS1, pRi, or pSa (Anand et al. 2018). The T-DNA binary vectors described in this work are designed to facilitate gene overexpression and CRISPR–Cas9-mediated targeted gene mutagenesis in plants, especially maize (Fig. 1, B and C; Supplemental Fig. S1). The overexpression vectors pCBL101 (Addgene #199722), pCBL101.1 (Addgene #199724), and pCB101gw (Addgene #199725; Fig. 1B) have spectinomycin-resistant gene, while the CRISPR vector pKL2351 (Addgene #199721; Fig. 1C) has kanamycin-resistant gene for bacterial selection. Both types of vectors have pVS1 ORI and thus are compatible with our helper plasmid pKL2299 (Addgene #186332). We used the NptII/geneticin (G418) in these vectors for plant selection to reduce escape rates observed by Kang et al. (2022) when using the bar/bialaphos for transformation. Bialaphos works as a glutamine synthetase inhibitor, and plant cells treated with bialaphos accumulate ammonia to a higher level, which is toxic under photorespiratory conditions (Donn and Köcher 2002). Under nonphotorespiratory conditions, however, increased ammonia level does not cause phytotoxic effects in the plant cells (Lea and Ridley 1989), making the bar/bialaphos selection system less efficient in the QuickCorn method. On the other hand, geneticin inhibits polypeptide synthesis by blocking the elongation step directly, causing phytotoxicity to cells (Wilmink and Dons 1993). Recently, the NptII/G418 selection system has been reported for maize leaf transformation with <10% escape rates (Wang et al. 2023). GOI can be cloned into these vectors via multiple cloning sites (MCSs) in pCBL101 or pCBL101.1 (Fig. 1B) or Gateway cloning in pCBL101gw (Fig. 1B) and pKL2351 (Fig. 1C). To test these vectors for maize B104 transformation, we cloned a new reporter system called 35S-RUBY, which includes a Cauliflower Mosaic Virus 35S promoter (P35S) driving the expression of 3 enzymes that convert tyrosine into a purple pigment betalain (He et al. 2020), into pCBL101 (pCBL101-RUBY, Fig. 1D; Addgene #199723). To test the efficiency of CRISPR–Cas9-mediated mutagenesis, maize codon–optimized SpCas9 and single-guide RNA (sgRNA) with an engineered scaffold (Dang et al. 2015; Supplemental Fig. S2) targeting maize Glossy2 gene were cloned into pKL2351 via the 3-way Gateway cloning method (pKL2359, Fig. 1D) as previously described (Lee et al. 2019). Detailed information about vector construction is described in the Supplemental data. Maize B104 transformation of immature embryos was performed using the QuickCorn method (Kang et al. 2022; Supplemental data) and also LBA4404Thy- strain harboring the helper plasmid pKL2299 and binary vectors pKL2013 (McCaw et al. 2021), pCBL101-RUBY, or pKL2359 (Fig. 1D). As summarized in Fig. 1E, pCBL101-RUBY and pKL2359 were efficient in B104 transformation with overall transformation frequencies of 14.8% and 7.7%, respectively, comparable to 10.3% of pKL2013. Importantly, using NptII/geneticin drastically reduced the escape rates from 35.6% to 66.6% (pKL2013 in Fig. 1E; Kang et al. 2022) to 0% to 5.7% (pCBL101-RUBY and pKL2359 in Fig. 1E) without compromising the transformation frequencies. Low escape rates can significantly save time and resources when generating and screening transgenic plants. We then analyzed the CRISPR–Cas9-mediated gene-editing efficiencies. An identical spacer sequence was used for Glossy2 gene editing in both pKL2013 (Kang et al. 2022) and pKL2359. The adoption of an engineered sgRNA scaffold (Dang et al. 2015) demonstrated efficient gene editing using a CRISPR/Cas9 system (Fig. 1F). Overall indel mutation frequencies were 90% to 100% for both pKL2013 and pKL2359, indicating a highly efficient CRISPR–Cas9-mediated gene editing. Particularly, the frequencies of loss-function-mutants (i.e. homozygous [HO] and biallelic [BI] mutations) were over 50% (35/61 = 57.4% for pKL2359; 7/14 = 50% for pKL2013; Fig. 1F; Supplemental Table S1). These results suggest that the engineered sgRNA scaffold (Supplemental Fig. S2) can be successfully used for CRISPR–Cas9-mediated gene editing in maize. Appropriate and efficient marker genes are essential for plant transformation, and they can be very useful in monitoring plant transformation procedures and screening transgenic plants or identifying transgene-free segregants. Here, we show that the RUBY reporter (He et al. 2020) can serve as an excellent visible marker in maize transformation. Using the constitutive 35S promoter driving the RUBY expression cassette, betalain accumulation was observed from the very early stages during the transformation procedures in the immature embryos 2 to 3 d after infection (Fig. 1G). Betalain accumulation patterns were highly variable across different tissues and developmental stages in maize (Figs. 1, G to J and 2). As can be seen in Fig. 1, H to J, betalain accumulation was not homogenous in the callus tissues, regenerating shoots, and rooted plantlets. Figure 2A shows 3 plants transplanted from Petri dishes to the soil on the same day but grew drastically different in size. Extensively pigmented plants showed stunted growth, presumably due to tyrosine deficiency (Fig. 2A; Supplemental Fig. S3H). Betalain was observed from almost all tissue types: leaves (Fig. 2, A to C), mature roots (Fig. 2, D and E), anthers (Fig. 2, H and I), ear (Fig. 2, J, K, M to P), and endosperm (Fig. 2, K and L). Some plants exhibited stripes in the leaves (Fig. 2B) and others had dark purple leaves (Fig. 2C). The cross section of the purple stem showed that betalain was highly accumulated along the vascular bundle tissues (Fig. 2, F and G). While betalain was not observed from the pollen (Fig. 2H), when the yellow transgenic pollen was used to pollinate wild-type B104 ears, the resulting ears exhibited interesting betalain pigmentation patterns across the ears (Fig. 2P). It is worth noting that some plants that did not show betalain pigmentation at early stages became pigmented in the later developmental stages, or only a part of certain tissues accumulated betalain (Fig. 2, B, M to O). Usually, T1 seedling colors were consistent with kernel colors: i.e. seedlings germinated from purple kernels had purple shoots and roots, whereas those germinated from yellow kernels had green shoots and white roots (Supplemental Fig. S3). Betalain accumulation pattern in various transgenic maize tissues. A) Transgenic plants with varying levels of betalain accumulation. Three plants of the same age but different betalain accumulation in tissues. Heavily pigmented plants (middle and right) exhibited stunted growth. B) Transgenic plant with betalain strips in the leaves. C) Variation of betalain accumulation in T0 leaf tissues. Wild-type B104 (W); transgenic T0 plant (T). D) Transgenic root tissue. E) Cross section of the T0 root tissue. F) Cross section of the T0 stem. G) Close-up of a stem cross section with vascular bundles showing betalain accumulation. H) Tassel and pollen of the T0 plant. No betalain accumulation was observed in pollen. I) Anthers of the T0 plant. J) Young ear tip of the T0 plant. K) Cross section of a young ear (bottom part of J). L) Cross section of a kernel (23 d after pollination, T0 × WT B104 pollen). Betalain is mainly accumulated in the endosperm and at the tip of the scutellum. M to P) Ears from different transgenic events. All images were taken from transgenic plants produced by pCBL101-RUBY (Fig. 1D) except for O [T0 produced by pCBL102-RUBY (Supplemental Fig. S1) × WT B104 pollen] and P (WT B104 × pCBL102-RUBY transgenic pollen). In summary, we adopted NptII/geneticin for our ternary vector system for plant transformation and greatly reduced the escape rates for the rapid Agrobacterium-mediated B104 transformation. Our empty vectors are compatible with many existing Gateway entry vectors that are available from Addgene, and new GOIs and CRISPR–Cas reagents can be easily cloned into them for plant transformation and gene-editing applications. The engineered sgRNA scaffold (Dang et al. 2015) showed promising results for CRISPR–Cas9-mediated gene editing efficiencies, with about two-thirds of the T0 plants being loss-of-function mutants. Although we tested only single sgRNA for gene editing, we generated Golden Gate entry vectors for up to 8 sgRNAs (Supplemental Fig. S2), and thus, these can be used for a wide range of applications. Moreover, we observed that the betalain biosynthesis genes, RUBY, produce purple pigments in most maize tissues throughout the developmental stages, indicating that RUBY can serve as an excellent visible marker for maize transformation. To reduce betalain accumulation in plant tissue that can lead to growth retardation, strategies of using weaker constitutive or inducible promoters to drive RUBY gene system could be considered. Together with the helper plasmid pKL2299 (Kang et al. 2022), our new T-DNA binary vectors and engineered sgRNA entry vectors can provide highly efficient systems for both gene overexpression and targeted mutagenesis in maize. The authors thank Dr. William Gordon-Kamm at Corteva Agriscience for providing PHP81856 plasmid DNA and auxotrophic Agrobacterium strain LBA4404Thy- strain, and Dr. Erik Vollbrecht at Iowa State University for providing maize B104 seeds. K.W.'s contribution to this work is partially supported by (while serving at) the National Science Foundation. Supplemental Table S1. Glossy2 genotyping results by TIDE and ICE analyses. Supplemental Table S2. List of oligonucleotides used in this study. Supplemental Table S3. List of gBlock gene fragments used in this study. Supplemental Figure S1. Map of the T-DNA binary vector pCBL102-RUBY. Supplemental Figure S2. Maps of guide RNA entry vectors. Supplemental Figure S3. Betalain pigmentation in T1 seedlings. Supplemental Materials and methods. A) Plasmid DNA vector construction. B) Maize B104 transformation. C) Genotyping analysis. This project was partially supported by the National Science Foundation (NSF) Plant Genome Research Program award IOS-1917138 and NSF Established Program to Stimulate Competitive Research's Research Infrastructure Improvement Program award 2121410 to K.W. and K.L., by the seed grant fund from the Crop Bioengineering Center of Iowa State University to K.L., by the National Institute of Food and Agriculture of United States Department of Agriculture Hatch project #IOW04714, and by State of Iowa funds. All study data are included in the article or Supplementary Data.