Enhancing Maize Transformation and Targeted Mutagenesis through the Assistance of Non-Integrating Wus2 Vector

Efficient genetic transformation is a prerequisite for rapid gene functional analyses and crop trait improvements. We recently demonstrated that new T-DNA binary vectors with NptII/G418 selection and a compatible helper plasmid can efficiently transform maize inbred B104 using our rapid Agrobacterium-mediated transformation method. In this work, we implemented the non-integrating Wuschel2 (Wus2) T-DNA vector method for Agrobacterium-mediated B104 transformation and tested its potential for recalcitrant inbred B73 transformation and gene editing. The non-integrating Wus2 (NIW) T-DNA vector-assisted transformation method uses two Agrobacterium strains: one carrying a gene-of-interest (GOI) construct and the other providing an NIW construct. To monitor Wus2 co-integration into the maize genome, we combined the maize Wus2 expression cassette driven by a strong constitutive promoter with a new visible marker RUBY, which produces the purple pigment betalain. As a GOI construct, we used a previously tested CRISPR-Cas9 construct pKL2359 for Glossy2 gene mutagenesis. When both GOI and NIW constructs were delivered by LBA4404Thy- strain, B104 transformation frequency was significantly enhanced by about two-fold (10% vs. 18.8%). Importantly, we were able to transform a recalcitrant inbred B73 using the NIW-assisted transformation method and obtained three transgene-free edited plants by omitting the selection agent G418. These results suggest that NIW-assisted transformation can improve maize B104 transformation frequency and provide a novel option for CRISPR technology for transgene-free genome editing.


Introduction
Recent advances in genome-editing tools, i.e., CRISPR technology, have demonstrated a great potential for modern crop breeding through de novo domestication [1][2][3] as well as precise trait engineering [4]. In particular, CRISPR-Cas9 has been widely used for different applications including, but not limited to, targeted gene mutagenesis via indel mutations, targeted DNA insertions, precise base changes, RNA editing, transcriptional reprogramming, and chromosomal restructuring [5][6][7][8]. CRISPR technology or any other genomeediting tools, however, cannot be fully utilized without establishing efficient methods for reagent delivery and the regeneration of fertile plants, i.e., genetic transformation [9].
Maize (Zea mays L.) is one of the most important cereal crops worldwide, but efficient genetic transformation methods are only available for a limited number of genotypes or inbreds in the public sector [10]. We recently established a rapid maize inbred B104 transformation method, which only takes about two months to produce rooted transgenic plants [11]. Our ternary vector system, which consists of a T-DNA binary vector and a compatible helper plasmid, demonstrated a robust B104 transformation efficiency with a different reporter and CRISPR-Cas9 vectors [11,12]. In particular, the adoption of the neomycin phosphotransferase II (NptII)/G418 selection system drastically reduced the escape rate from 36-67% to less than 10%, greatly saving time and resources compared to the bar/bialaphos selection system [12]. In this study, we implemented a non-integrating ZmWus2 vector-assisted transformation method [13] to further enhance B104 transformation frequency and tested its potential for recalcitrant genotype B73 transformation and gene editing.
Morphogenic transcription factors (MTF), such as Wuschel2 (Wus2) and Baby Boom (Bbm), are widely applicable to enhance transformation frequency and expand amenable genotypes [14][15][16]. However, constitutive expression of Wus2 or Bbm prevents normal plant regeneration and production of fertile plants [14], thus requiring the removal of the Wus2/Bbm expression cassettes by Cre-mediated inducible excision [17]. To circumvent this issue, Corteva Agriscience's scientists demonstrated an alternative approach using an additional Agrobacterium strain harboring a T-DNA binary vector that carries a Wus2 expression cassette with a strong constitutive promoter [13] (Figure 1a). This approach is based on the observation that Wus2 proteins are diffusible to neighboring cells and can stimulate cell proliferation without the integration of Wus2 T-DNA [14,18]. Using an NIW vector, we tested three thymidine-auxotrophic Agrobacterium strains for maize B104 immature embryo transformation. In addition, we tested if transgene-free gene editing can be achievable for a recalcitrant inbred B73 using the NIW-assisted transformation method. We observed significantly improved B104 transformation frequency and obtained three transgene-free gene-edited B73 plants, suggesting that the NIW-assisted transformation method can be a useful option for maize genome engineering.
plants [11]. Our ternary vector system, which consists of a T-DNA binary vector and a compatible helper plasmid, demonstrated a robust B104 transformation efficiency with a different reporter and CRISPR-Cas9 vectors [11,12]. In particular, the adoption of the neomycin phosphotransferase II (NptII)/G418 selection system drastically reduced the escape rate from 36-67% to less than 10%, greatly saving time and resources compared to the bar/bialaphos selection system [12]. In this study, we implemented a non-integrating ZmWus2 vector-assisted transformation method [13] to further enhance B104 transformation frequency and tested its potential for recalcitrant genotype B73 transformation and gene editing.
Morphogenic transcription factors (MTF), such as Wuschel2 (Wus2) and Baby Boom (Bbm), are widely applicable to enhance transformation frequency and expand amenable genotypes [14][15][16]. However, constitutive expression of Wus2 or Bbm prevents normal plant regeneration and production of fertile plants [14], thus requiring the removal of the Wus2/Bbm expression cassettes by Cre-mediated inducible excision [17]. To circumvent this issue, Corteva Agriscience's scientists demonstrated an alternative approach using an additional Agrobacterium strain harboring a T-DNA binary vector that carries a Wus2 expression cassette with a strong constitutive promoter [13] (Figure 1a). This approach is based on the observation that Wus2 proteins are diffusible to neighboring cells and can stimulate cell proliferation without the integration of Wus2 T-DNA [14,18]. Using an NIW vector, we tested three thymidine-auxotrophic Agrobacterium strains for maize B104 immature embryo transformation. In addition, we tested if transgene-free gene editing can be achievable for a recalcitrant inbred B73 using the NIW-assisted transformation method. We observed significantly improved B104 transformation frequency and obtained three transgene-free gene-edited B73 plants, suggesting that the NIW-assisted transformation method can be a useful option for maize genome engineering.

NIW-Assisted Transformation of Maize B104
We made an NIW vector pKL2391 ( Figure 1b) carrying a strong constitutive promoter for maize Wus2 expression and a visible reporter RUBY, which converts tyrosine to purple pigment betalain [19], to monitor the integration of the NIW construct T-DNA. A CRISPR-Cas9 vector pKL2359 (Figure 1c) [12] was used as a GOI construct and delivered by LBA4404Thy-strain in all experiments. The NIW construct pKL2391 was delivered by one of the three thymidine-auxotrophic Agrobacterium strains for B104 immature embryo transformation. Agrobacterium cells were prepared as described by Hoerster et al. [13]. The GOI and NIW strains were mixed at a 9:1 ratio immediately before the immature embryo infection [13].
As summarized in Table 1, NIW-assisted transformation using Agrobacterium strain LBA4404Thy-(Thy-, thymidine synthase knockout mutant) [20] significantly enhanced B104 transformation frequency compared to the control (18.8% vs. 10.0%; p < 0.05, two proportion z-test). NIW strain EHA105TR (TR, thymidine synthase, and recA knockout) [21] also enhanced transformation frequency by 55% (15.5% vs. 10.0%; p = 0.055, two proportion z-test) but it was not statistically significant. One strain, EHA105Thy- [22], did not have a noticeable impact on the average transformation frequency compared to the control (6.8% vs. 10.0%; p = 0.241, two proportion z-test). Our previous study using LBA4404Thy-strain harboring pKL2359 had a similar result with 7.7% of transformation frequency, which is comparable to the control in this study [12], suggesting that NIW-assisted transformation can be utilized to further enhance maize transformation efficiencies. RUBY reporter was very useful to identify Wus2-integrated callus or regenerated plantlets. Betalain accumulation could be easily identified, and callus pieces or regenerating shoots with purple color were excluded during the subculture, resulting in a low Wus2 co-integration frequency ( Table 1; 2/176 = 1.1%). As previously demonstrated [12], the CRISPR-Cas9 showed efficient targeted mutagenesis, and most of the analyzed T0 plants (80.8-89.5%) carried various indel mutations at the Glossy2 (Gl2) target site (Table S1).

Transgene-Free Gene Editing in Maize B73
Maize inbred B73 is an important public genotype that serves as a reference genome and has been widely utilized in many research programs [23]. However, genetic transformation or gene editing of B73 remains challenging due to its recalcitrance to regeneration. We wanted to explore the potential of the NIW-assisted transformation approach for transgenefree gene editing in B73. Transgene-free CRISPR-Cas9-mediated gene editing has been demonstrated before using a transient expression system in wheat [24]. Recalcitrant genotypes such as B73, however, pose an additional challenge: i.e., poor regeneration efficiency. We tested if the NIW-assisted transformation approach can be combined with the transient expression of CRISPR-Cas9 to generate gene-edited B73 plants without transgene integra-tion ( Figure 2a). We omitted selection agent G418 from all tissue culture media to recover both transgenic and transgene-free regenerants (Figure 2b).  (Figure 2a). We omitted selection agent G418 from all tissue culture media to recover both transgenic and transgene-free regenerants (Figure 2b).
. As summarized in Table 2, we did not recover any regenerated plants from the 136 infected immature embryos with our rapid transformation method (control), confirming the "recalcitrant" nature of the B73 inbred. On the other hand, the NIW-assisted transformation approach was efficient in stimulating B73 regeneration as we recovered 79 independent plants from 395 infected immature embryos (Table 2). Betalain pigmentation was not observed from the 79 regenerants, and PCR results confirmed that Wus2 T-DNA cointegration was not detected from the regenerated plants. We used PCR screening to identify transgenic plants and nine were positive for the CRISPR-Cas9 T-DNA (9/79 = 11.4%), indicating that the vast majority were transgene-free regenerants (70/79 = 88.6%; Table 2). We then PCR-amplified the CRISPR-Cas9 target, Gl2 exon2 region [11,12], and used Sanger sequencing and TIDE/ICE analyses [25,26] to identify gene-edited plants. We found a total of 10 plants that carried indel mutations at the target site: seven T0 plants (Table S3) and three transgene-free plants (Table 3). Seven T0 plants showed biallelic (BI), heterozygous (HT), or mosaic (MO) mutation at the target site (1 HT, 2 BI, and 4 MO plants; Table S3). A loss-of-function mutant was identified among the regenerated T0 plants and showed the expected phenotype (water droplet adherence to the surface of young leaves) when misted (Figure 2c). Importantly, three gl2-edited plants were transgene-free ( Figure S1), indicating that the NIW-assisted transformation approach can be adopted for transgene-free gene editing applications for recalcitrant genotypes such as B73 (Table 3). In our case, one plant had a biallelic mutation (−4/−27), and the other two had heterozygous mutations (+1/0; −1/0), demonstrating an efficient targeted mutagenesis without CRISPR-Cas9 T-DNA integration (Table 3). As summarized in Table 2, we did not recover any regenerated plants from the 136 infected immature embryos with our rapid transformation method (control), confirming the "recalcitrant" nature of the B73 inbred. On the other hand, the NIW-assisted transformation approach was efficient in stimulating B73 regeneration as we recovered 79 independent plants from 395 infected immature embryos (Table 2). Betalain pigmentation was not observed from the 79 regenerants, and PCR results confirmed that Wus2 T-DNA co-integration was not detected from the regenerated plants. We used PCR screening to identify transgenic plants and nine were positive for the CRISPR-Cas9 T-DNA (9/79 = 11.4%), indicating that the vast majority were transgene-free regenerants (70/79 = 88.6%; Table 2). We then PCR-amplified the CRISPR-Cas9 target, Gl2 exon2 region [11,12], and used Sanger sequencing and TIDE/ICE analyses [25,26] to identify gene-edited plants. We found a total of 10 plants that carried indel mutations at the target site: seven T0 plants (Table S3) and three transgene-free plants (Table 3). Seven T0 plants showed biallelic (BI), heterozygous (HT), or mosaic (MO) mutation at the target site (1 HT, 2 BI, and 4 MO plants; Table S3). A loss-of-function mutant was identified among the regenerated T0 plants and showed the expected phenotype (water droplet adherence to the surface of young leaves) when misted (Figure 2c). Importantly, three gl2-edited plants were transgene-free ( Figure S1), indicating that the NIW-assisted transformation approach can be adopted for transgene-free gene editing applications for recalcitrant genotypes such as B73 (Table 3). In our case, one plant had a biallelic mutation (−4/−27), and the other two had heterozygous mutations (+1/0; −1/0), demonstrating an efficient targeted mutagenesis without CRISPR-Cas9 T-DNA integration (Table 3).

Discussion
In this study, we implemented a non-integrating Wus2 approach [13,27] for NIWassisted transformation of maize inbred lines B104 and B73. When combined with our rapid B104 transformation method [11], overall transformation frequency was enhanced by about two-fold (10% vs. 18.8%) by simply mixing two Agrobacterium strains that carry GOI and NIW constructs, respectively, with a 9:1 ratio before immature embryo infection. This simple extra step can significantly save time and resources for researchers to produce transgenic or gene-edited plants. Because the NIW Agrobacterium strain is separately maintained, only GOI Agrobacterium strains need to be modified for different applications, making the NIW-assisted transformation method a versatile option for enhancing maize transformation efficiency.
We tested three different Agrobacterium strains for their capacity to deliver the NIW T-DNA. One strain LBA4404Thy-significantly enhanced B104 transformation frequency, whereas the other two strains, EHA105Thy-and EHA105TR, did not have a marked impact (Table 1). Because Wus2 co-integration inhibits normal plant regeneration, each NIW Agrobacterium strain might have a different optimal mixing ratio to effectively deliver NIW T-DNA without co-integration with the GOI T-DNAs. In addition, because EHA105Thy-and EHA105TR have a different chromosomal background (A. tumefaciens C58) to LBA4404Thy-(A. tumefaciens Ach5), there might be different cell-to-cell interactions, i.e., interbacterial competition, when mixed for co-infection. Future studies might provide further insights into how different Agrobacterium strains can perform in combination with other strains to deliver NIW and GOI T-DNAs.
As a proof of concept, we applied the NIW-assisted transformation approach for a recalcitrant genotype B73 to generate transgene-free gene-edited plants. Importantly, NIW-assisted transformation was efficient in stimulating B73 regeneration without Wus2 integration, as we recovered 79 regenerants from 395 infected embryos (Table 2). In contrast, none were regenerated from the 136 infected embryos of control. Molecular analyses identified nine transgenic plants and three transgene-free edited plants. Based on the number of infected immature embryos, these numbers can be translated to 2.3% of transformation frequency for the transgenic plants (9/395) and 0.8% of the transgene-free targeted mutagenesis rate (3/395). In addition, based on the number of regenerated plants, 11.4% were transgenic and 3.8% were transgene-free edited plants. It needs to be noted that among the three transgene-free edited plants, one was a biallelic mutant and the others were heterozygous mutants; therefore, loss-of-function edited plants without transgenes can be readily obtained in the first generations as biallelic or homozygous mutants, or they can be obtained in the second generation as homozygous mutants after segregation. Given that our rapid transformation method only takes two months to produce first-generation edited plants, transgene-free edited plants can be obtained in about 2-6 months for phenotyping analyses. Interestingly, previous studies reported that an Agrobacterium strain carrying a mutant virD2 gene exhibited significantly impaired T-DNA integration frequency while the transient T-DNA expression was only slightly diminished [28,29]. It would be worthwhile to develop engineered Agrobacterium strains that can efficiently deliver T-DNAs into plant cells without integration. MTFs and CRISPR reagents might be delivered as a single T-DNA to achieve transgene-free gene editing.

NIW Vector Construction
The NIW binary vector was constructed using pTF101.1 [30] as a backbone. First, pTF101.1 DNA was digested with restriction enzymes HindIII and AatII and the 7.4 kb fragment was purified after agarose gel electrophoresis. Next, a 2.6 kb fragment containing 3× viral enhancers and maize ubiquitin promoter and a 1.5 kb fragment containing maize Wus2 coding sequence and In2-1 gene terminator were PCR-amplified from PHP97334 [31], which was kindly provided by Dr William Gordon-Kamm at Corteva Agriscience, using primers listed in Table S2 (ZmWUS2-F1 and TIN2-1-R1; 3×ENH-F1 and PZmUbi-R1). All PCR reactions were carried out using high-fidelity Q5 DNA polymerase (NEB, Ipswich, MA, USA). HiFi DNA assembly mix (NEB) was used for the Gibson assembly, and the resulting plasmid was named pKL2387. This intermediate vector was digested with KpnI and dephosphorylated using alkaline phosphatase to prevent self-ligation. The RUBY reporter with CaMV 35S promoter [19] was PCR-amplified from pCBL101-RUBY [12] using two primer pairs (P35-RUBY-F1 and DODA-R1; DODA-F1 and P35-RUBY-R1) listed in Table S2. Gibson assembly using the NEB Hifi DNA assembly mix resulted in the NIW binary vector pKL2391 (Figure 1b).

NIW-Assisted Transformation of Maize B104 and B73
Seeds of the maize inbred lines were obtained from Dr Erik Vollbrecht (Iowa State University), and immature embryos were produced in the Crop Bioengineering Laboratory greenhouse, Iowa State University (Ames, IA, USA). Maize plants were grown as previously described [11]. Briefly, the greenhouse conditions were as follows: photoperiod, 14 h/10 h (day/night); temperatures, 28 • C/22 • C (day/night). All silks were covered with shoot bags before the emergence to prevent cross-contamination. Each ear was self-pollinated or crossed with sibling plants. B104 and B73 ears containing embryos of appropriate size (1.8-2.0 mm) were harvested 12-14 days after pollination.
Agrobacterium suspension for infection was prepared as described by Kang et al. (2022) [11]. Briefly, a mother plate was prepared by streaking Agrobacterium cells from a glycerol stock on an AB solid media containing 50 mg/L of thymidine and appropriate antibiotics and incubating at 28 • C for 2 days. A working plate was streaked on a YEP medium amended with appropriate antibiotics and thymidine using the cells from the mother plate a day before the infection. Mother plates can be stored at 4 • C for up to 10 days for streaking working plates. Agrobacterium suspension for infection was prepared after embryo isolation. Freshly harvested cells from a working plate were thoroughly resuspended by vortexing in 700A medium supplemented with thymidine (50 mg/L) and acetosyringone (100 µM). The final cell density was adjusted to 0.45-0.55 at OD 550nm . GOI and NIW Agrobacterium suspensions were prepared separately, and two Agrobacterium strains were mixed with a 9:1 ratio (GOI: NIW) immediately before embryo infection.
Dehusked maize ears were disinfected using 20% bleach solution (5.25% sodium hypochlorite) for 20 min and rinsed three times with sterile water. The top of the kernels was removed using a sharp scalpel and embryos were carefully isolated using a sterile micro-spatula. Isolated embryos were collected in 2 mL tubes containing 700A medium and washed once with 700A medium before infection. To infect the embryos, the 700A medium in the 2 mL tubes was carefully removed by pipetting and 1 mL of Agrobacterium suspension was added. After 5 min, the embryos were poured onto a co-cultivation medium, and the excessive amount of infection medium was removed by gentle pipetting. Infected embryos were carefully re-orientated with the scutellum side up and incubated at 20 • C overnight (16-20 h) in the dark. After co-cultivation, all embryos were transferred to a resting medium (scutellum side up) and cultured for one week at 28 • C in the dark. Transient expression of mCherry could be checked 72 h after infection. Developing tissue with somatic embryos was transferred to a maturation medium containing 75 mg/L of G418 and cultured for 4-6 weeks at 28 • C in the dark. To regenerate transgene-free B73 plants, the selection agent G418 was omitted in all tissue culture media. Tissues on the maturation medium were sub-cultured every 2 weeks. Actively regenerating tissues with immature shoots were further transferred to a rooting medium amended with G418 and kept in a light chamber for 2-3 weeks at 28 • C under a 16h/8h photoperiod (light/dark). Rooted plantlets were carefully transferred to 32-cell seed-growing trays filled with autoclaved potting mix Metromix 360 (Sungro, Agawam, MA, USA).

Genotyping of Regenerated Plants
Total genomic DNA was isolated from regenerated plants using a modified version of the protocol described by Edwards et al. (1991) [33]. PCR screening was conducted to identify transgenic T0 plants using primers specific to the zCas9 gene (zCas9-F and zCas9-R; Table S2). To identify transgene-free B73 regenerants, a total of seven primer pairs were used to check the presence of zCas9, mCherry, NptII, and RUBY genes (Table S2). Targeted mutagenesis of the Glossy2 gene was analyzed by Sanger sequencing and TIDE/ICE analyses [25,26] as previously described [11,34]. When the genotyping results by TIDE and ICE analyses were not clear, PCR products were cloned into pJET1.2 vector as instructed by the manufacturer (Thermo Fisher Scientific, Waltham, MA, USA), and 8-12 clones were sequenced per sample using the sequencing primers provided by the kit (pJET1.2 Forward and Reverse sequencing primers).

Conclusions
We adopted the non-integrating maize Wus2-assisted Agrobacterium-mediated transformation method to enhance B104 transformation efficiency and to produce transgene-free gene-edited plants for the recalcitrant genotype B73. When LBA4404Thy-strain was used for both NIW and GOI constructs, B104 transformation frequency was almost doubled by simply mixing a GOI Agrobacterium strain with an NIW strain at a 9:1 ratio, suggesting that this method can be widely applied for maize transformation and gene editing. In conclusion, our results suggest that when combined with the rapid Agrobacterium-mediated transformation method, NIW-assisted transformation can boost B104 transformation effi-