Combination of Hairy Root and Whole-Plant Transformation Protocols to Achieve Efficient CRISPR/Cas9 Genome Editing in Soybean

The new gene-editing technology CRISPR/Cas system has been widely used for genome engineering in various organisms. Since the CRISPR/Cas gene-editing system has a certain possibility of low efficiency and the whole plant transformation of soybean is time-consuming and laborious, it is important to evaluate the editing efficiency of designed CRISPR constructs before the stable whole plant transformation process starts. Here, we provide a modified protocol for generating transgenic hairy soybean roots to assess the efficiency of guide RNA (gRNA) sequences of the CRISPR/Cas constructs within 14 days. The cost- and space-effective protocol was first tested in transgenic soybean harboring the GUS reporter gene for the efficiency of different gRNA sequences. Targeted DNA mutations were detected in 71.43–97.62% of the transgenic hairy roots analyzed as evident by GUS staining and DNA sequencing of the target region. Among the four designed gene-editing sites, the highest editing efficiency occurred at the 3′ terminal of the GUS gene. In addition to the reporter gene, the protocol was tested for the gene-editing of 26 soybean genes. Among the gRNAs selected for stable transformation, the editing efficiency of hairy root transformation and stable transformation ranged from 5% to 88.8% and 2.7% to 80%, respectively. The editing efficiencies of stable transformation were positively correlated with those of hairy root transformation with a Pearson correlation coefficient (r) of 0.83. Our results demonstrated that soybean hairy root transformation could rapidly assess the efficiency of designed gRNA sequences on genome editing. This method can not only be directly applied to the functional study of root-specific genes, but more importantly, it can be applied to the pre-screening of gRNA in CRISPR/Cas gene editing.


Introduction
Soybean [Glycine max (L.) Merr.] is an economically important crop for feed, oil, and protein products. It contains about 40% protein and 20% oil in the seed [1,2]. In the past decades, significant progress has been made in soybean functional genomics and its application in molecular breeding [3][4][5]. Genome editing is a tremendous strategy for efficient and targeted genome manipulations, especially for crops that have complex genomes and difficulty being improved through conventional breeding approaches [6,7]. The targeted genome editing methods have undergone three generations of technological development and improvement, from zinc-finger nucleases (ZFNs) to transcription activator-like effector nucleases (TALENs), and recently the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein (CRISPR/Cas) [8]. All three methods induce double-stranded breaks in the target genome DNA, which are subsequently repaired 2.7% to 80.0%, respectively. The editing efficiency of stable transformation was positively correlated with those of hairy root transformation with a Pearson correlation of efficient (r) of 0.83. Meanwhile, we efficiently edited the targeted GUS and soybean endogenous PDS genes for editing in hairy roots. Using the pre-screening of gRNAs, we were able to achieve whole-plant editing events for 26 soybean genes, which provides a potential way to study soybean gene function and genetic improvement.

A Rapid and Efficient Protocol for A. rhizogenes-Mediated Hairy Root Transformation System in Soybean
To use the hairy root transformation system to assess of CRISPR/Cas9 constructs, we first made efforts to optimize the hairy root transformation protocol for its robustness and effectiveness. Previously, hairy root transformation was performed by stabbing the hypocotyl of soybean seedlings as described [33]. By inoculating hypocotyls adjacent to cotyledonary nodes of the one-week-old seedlings with A. rhizogenes stain K599, transgenic hairy roots were obtained in 4 weeks ( Figure 1A-C).
Plants 2023, 12, x FOR PEER REVIEW 3 of 13 different gRNAs in hairy root and whole-plant transformation ranged from 5.0% to 88.8%, and 2.7% to 80.0%, respectively. The editing efficiency of stable transformation was positively correlated with those of hairy root transformation with a Pearson correlation of efficient (r) of 0.83. Meanwhile, we efficiently edited the targeted GUS and soybean endogenous PDS genes for editing in hairy roots. Using the pre-screening of gRNAs, we were able to achieve whole-plant editing events for 26 soybean genes, which provides a potential way to study soybean gene function and genetic improvement.

A Rapid and Efficient Protocol for A. rhizogenes-Mediated Hairy Root Transformation System in Soybean
To use the hairy root transformation system to assess of CRISPR/Cas9 constructs, we first made efforts to optimize the hairy root transformation protocol for its robustness and effectiveness. Previously, hairy root transformation was performed by stabbing the hypocotyl of soybean seedlings as described [33]. By inoculating hypocotyls adjacent to cotyledonary nodes of the one-week-old seedlings with A. rhizogenes stain K599, transgenic hairy roots were obtained in 4 weeks ( Figure 1A-C). To improve the robustness and effectiveness of the hairy root transformation protocol, we tried split-imbibed soybean seeds as explants for the A. rhizogenes-mediated hairy root transformation. The explant preparation method was referred from the soybean cotyledenary nodes whole-plant transformation [22,23]. In brief, the overnight imbibed seeds were excised into two half-seeds, then inoculated with A. rhizogenes suspension culture for 30 min. The infected cotyledons were then transferred to root induction (RI) medium to produce hairy roots. Hairy roots emerged at 6 days post-inoculation (DPI). By 14 DPI, the transgenic roots had grown to 1-2 cm, which is suitable for DNA extraction and further identification ( Figure 1D-H). To improve the robustness and effectiveness of the hairy root transformation protocol, we tried split-imbibed soybean seeds as explants for the A. rhizogenes-mediated hairy root transformation. The explant preparation method was referred from the soybean cotyledenary nodes whole-plant transformation [22,23]. In brief, the overnight imbibed seeds were excised into two half-seeds, then inoculated with A. rhizogenes suspension culture for 30 min. The infected cotyledons were then transferred to root induction (RI) medium to produce hairy roots. Hairy roots emerged at 6 days post-inoculation (DPI). By 14 DPI, the transgenic roots had grown to 1-2 cm, which is suitable for DNA extraction and further identification ( Figure 1D-H).
To verify the effectiveness of the protocol, the binary plasmid pTF102 [23] was introduced into A. rhizogenes strain K599 to overexpress GUS reporter gene in hairy roots. The number of generated hairy roots per explant was calculated at 6, 8, 10, and 14 DPI. Putative transgenic roots started to emerge at 6 DPI and increased to an average of 8.4 at 14 DPI (Figure 2A). GUS (β-glucuronidase) staining and polymerase chain reaction (PCR) verification of each hairy root generated from explants were carried out. Results showed that nearly all regenerated hairy roots are PCR-and GUS-positive ( Figure 2B,C). The transformation frequency reached 94.3% with the RI medium containing 5 mg/L glufosinate as the selecting agent.
Plants 2023, 12, x FOR PEER REVIEW 4 To verify the effectiveness of the protocol, the binary plasmid pTF102 [23] was in duced into A. rhizogenes strain K599 to overexpress GUS reporter gene in hairy roots. number of generated hairy roots per explant was calculated at 6, 8, 10, and 14 DPI. P tive transgenic roots started to emerge at 6 DPI and increased to an average of 8.4 a DPI ( Figure 2A). GUS (β-glucuronidase) staining and polymerase chain reaction (P verification of each hairy root generated from explants were carried out. Results sho that nearly all regenerated hairy roots are PCR-and GUS-positive ( Figure 2B,C). transformation frequency reached 94.3% with the RI medium containing 5 m glufosinate as the selecting agent. In summary, by using the imbibed seeds as explants, we were able to reduce the ration of the hairy root transformation procedure from the classic protocol of 3-4 w to 2 weeks. Since Ri medium contains the selective agent, the vast majority of the h roots obtained are transgenic.

CRISPR/Cas9-Induced Mutagenesis of GUS Transgene in Soybean Hairy Roots
Although whole-plant transformation of soybean has become routine, the trans mation efficiency is often low and the procedure to obtain transformed plantlets take least three months. Thus, using hairy root system to assess the effective target gRNA p to conducting the whole-plant genome editing would be greatly save time and costs evaluate the gene-editing efficiency of different target designs of the GUS reporter g four GUS-CAS9 constructs with different gRNA sequences were made to edit the G gene in the GUS-expressing soybean lines [23]. The GUS-expressing soybean lines are homozygous transgenic soybean containing a single copy of GUS expression cass driven by the CaMV35S promoter. Four gRNAs were designed to target the 5′ end of G (5′Target1 and 5′Target2) or the 3′ end (3′Target1 and 3′Target2) ( Figure 3A). The structs were introduced into A. rhizogenes stain K599 to produce transformed hairy ro in which the previously transferred GUS-coding sequence was edited. The success of G editing was evident by both GUS staining and sequencing of target region. While transgenic roots generated from the control construct containing only the Cas9 expres cassette without gRNA were stained dark blue (the top rows in Figure 3B), GUS stain of the transgenic roots showed significant reduction in blue color ( Figure 3B). When b GUS alleles were edited, no blue color was seen in GUS staining. In summary, by using the imbibed seeds as explants, we were able to reduce the duration of the hairy root transformation procedure from the classic protocol of 3-4 weeks to 2 weeks. Since Ri medium contains the selective agent, the vast majority of the hairy roots obtained are transgenic.

CRISPR/Cas9-Induced Mutagenesis of GUS Transgene in Soybean Hairy Roots
Although whole-plant transformation of soybean has become routine, the transformation efficiency is often low and the procedure to obtain transformed plantlets takes at least three months. Thus, using hairy root system to assess the effective target gRNA prior to conducting the whole-plant genome editing would be greatly save time and costs. To evaluate the gene-editing efficiency of different target designs of the GUS reporter gene, four GUS-CAS9 constructs with different gRNA sequences were made to edit the GUS gene in the GUS-expressing soybean lines [23]. The GUS-expressing soybean lines are the homozygous transgenic soybean containing a single copy of GUS expression cassette driven by the CaMV35S promoter. Four gRNAs were designed to target the 5 end of GUS (5 Target1 and 5 Target2) or the 3 end (3 Target1 and 3 Target2) ( Figure 3A). The constructs were introduced into A. rhizogenes stain K599 to produce transformed hairy roots, in which the previously transferred GUS-coding sequence was edited. The success of GUS editing was evident by both GUS staining and sequencing of target region. While the transgenic roots generated from the control construct containing only the Cas9 expression cassette without gRNA were stained dark blue (the top rows in Figure 3B), GUS staining of the transgenic roots showed significant reduction in blue color ( Figure 3B). When both GUS alleles were edited, no blue color was seen in GUS staining.  The GUS gene-editing efficiency was calculated by the number of edited GU out of the total GUS copies. The generated transgenic hairy roots include three ty not edited (0 of 2 alleles edited), heterozygous mutant (1 of 2 alleles edited) and h gous mutant (2 alleles edited). The editing rates were 71.43%, 77.27%, 79.41%, and for the gRNAs of 5′Target1, 5′Target2, 3′Target1 and 3′Target2, respectively (Fi Table 1).  The GUS gene-editing efficiency was calculated by the number of edited GUS alleles out of the total GUS copies. The generated transgenic hairy roots include three types, i.e., not edited (0 of 2 alleles edited), heterozygous mutant (1 of 2 alleles edited) and homozygous mutant (2 alleles edited). The editing rates were 71.43%, 77.27%, 79.41%, and 97.62% for the gRNAs of 5 Target1, 5 Target2, 3 Target1 and 3 Target2, respectively ( Figure 3B, Table 1). To verify the knockouts and determine the genetic modifications of the transgenic roots, Sanger sequencing was carried out on the amplicons that amplified using the PCR primer pairs covering the target regions. Results showed that the most common mutation were short (1-32 nt) ( Figure S1). In addition, sequence insertion and (up to 35 bp), and base substitution were also commonly seen ( Figure S1).

Gene-Editing Efficiency of the Two Homeologous Genes of Soybean Phytoene Dehydrogenase (PDS)
Soybean has a paleopolyploid genome, in which nearly 75% of genes are present in multiple copies [24]. To evaluate the efficiency of the CRISPR/Cas9 system in editing the duplicated genes simultaneously, we first designed gRNAs targeting the common sequence of the two soybean phytoene dehydrogenase (GmPDS) genes, GmPDS1 (Glyma.11G253000) and GmPDS2 (Glyma.18G003900). Phytoene desaturase plays critical functions in pigment synthesis, and disruption of its functions leads to albino plants [34]. Five gRNAs were designed and evaluated in hairy root transformation system with various gene-editing efficiencies ranging from 35.6 to 100.0%. Then, we selected GmPDS1/2 gRNA-1, GCATTAAT-GATCGGTTACAATGG, which showed 100% PDS gene-editing efficiency in hairy root, to construct CRISPR/Cas9 vector for stable soybean transformation. Twenty-six Cas9-positive T 0 plants were generated. Sequencing of the target region of GmPDS genes showed that among 26 Cas9-positive events, 7 events had mutation on GmPDS2, and 4 events had mutation in both GmPDS1 and GmPDS2, with the gene-editing efficiencies for GmPDS2 only and both genes of 26.9% and 15.4%, respectively. The pds1/2 mutants with both GmPDS1 and GmPDS2 gene edited showed bleached leaf color, which can be observed during shoot elongation, rooting, and the seedling stages ( Figure 4A). Sequencing of the four events with both GmPDS alleles edited showed that the editing patterns included base deletion from −1 to −11, base addition from +1 to +23, and a base substitution of 3 (Table 2, Figure S1). As the GmPDS homozygous mutant seedlings cannot survive, the white-green heterozygous mutants that contain at least one wild-type allele of GmPDS1 or GmPDS2 were used for the segregation assay. Results show that the cotyledon color of the T 1 seedlings of pds heterozygous mutant follows a Mendelian segregation pattern ( Figure 4B,C). base substitution were also commonly seen ( Figure S1).

Gene-Editing Efficiency of the Two Homeologous Genes of Soybean Phytoe (PDS)
Soybean has a paleopolyploid genome, in which nearly 75% of gen multiple copies [24]. To evaluate the efficiency of the CRISPR/Cas9 syst duplicated genes simultaneously, we first designed gRNAs targeting quence of the two soybean phytoene dehydrogenase (GmPDS) (Glyma.11G253000) and GmPDS2 (Glyma.18G003900). Phytoene desatur functions in pigment synthesis, and disruption of its functions leads to a Five gRNAs were designed and evaluated in hairy root transformation s ous gene-editing efficiencies ranging from 35.6 to 100.0%. Then, we sel gRNA-1, GCATTAATGATCGGTTACAATGG, which showed 100% PDS ciency in hairy root, to construct CRISPR/Cas9 vector for stable soybean Twenty-six Cas9-positive T0 plants were generated. Sequencing of the GmPDS genes showed that among 26 Cas9-positive events, 7 events h GmPDS2, and 4 events had mutation in both GmPDS1 and GmPDS2, with efficiencies for GmPDS2 only and both genes of 26.9% and 15.4%, respect mutants with both GmPDS1 and GmPDS2 gene edited showed bleached can be observed during shoot elongation, rooting, and the seedling sta Sequencing of the four events with both GmPDS alleles edited showed patterns included base deletion from −1 to −11, base addition from +1 to substitution of 3 (Table 2, Figure S1). As the GmPDS homozygous mutant survive, the white-green heterozygous mutants that contain at least one of GmPDS1 or GmPDS2 were used for the segregation assay. Results sh ledon color of the T1 seedlings of pds heterozygous mutant follows a Me tion pattern ( Figure 4B,C).

Gene-Editing Efficiency of Different Targets of Soybean Transparent Testa 8a (GmTT8a) and GmTT8b
The described hairy root transformation system was applied and evaluated in editing the pair of soybean function genes, GmTT8a (Glyma.02G147800) and GmTT8b (Glyma.10G026000). In this attempt, gRNAs were designed for targeting to the common region of both GmTT8a and GmTT8b, or targeting to either of GmTT8a and GmTT8b alone. PCR primers were selected in the locus-specific region to distinguish the mutations in the different locus.
A total of 10 gRNAs were designed, including 2 for both of GmTT8a and GmTT8b, 3 for GmTT8a, and 5 of GmTT8b (Table 3). The mutation rate of these targets was calculated as the number of hairy roots containing edited sequence at the target site divided by the total number of hairy roots that contain the selectable marker gene bar and Cas9 gene ( Table 3). Editing of the target region was achieved in 6 out of the 10 gRNAs, including GmTT8a/b-1, GmTT8a-2, -3, and GmTT8b-2, -3, -5, with different editing efficiency ranging from 15% to 100% (Table 3). Based on the data, we selected GmTT8a/b-1, GmTT8a-2, and GmTT8b-2 as the gRNAs for the whole-plant transformation. As showed in Table 3, using GmTT8a-2 and GmTT8b-2 gRNAs, we were able to obtain five and eight whole plant mutants for GmTT8a and GmTT8b, respectively. No double mutant of GmTT8a and GmTT8b were identified among the 30 Cas9 positive transgenic events. It is possible that knocking out both GmTT8a and GmTT8b caused lethality.

Achievement of Efficient CRISPR/Cas9-Induced Targeted Mutations of Other Functional Genes in Soybean
We applied the hairy root transformation system to pre-screening the efficient gRNA for the whole-plant genetic transformation. We collected different experiments in the lab to prove the importance of pre-screening of targets using hairy root system. Table 4 showed the overall CRISPR/Cas9 gene-editing events for 13 gRNAs involving six different pairs of soybean homoeologous genes. Experimental data of the gRNA1, 2, and 4 are collected from published studies from the lab [35][36][37]. The gRNAs 5-1 and 6-3 were failed to achieve any transgenic hairy roots so that they were abandoned for stable transformation ( Table 4). The target regions of the transgenic hairy roots obtained from the gRNAs 5-2 and 6-2 were detailed listed ( Figure S3). Among the gRNAs selected for stable transformation, the editing efficiencies of hairy root transformation and stable transformation ranged from 5.0 to 88.8% and 2.7% to 60.0%, respectively ( Table 4). The editing efficiencies of stable transformation were positively correlated with those of hairy root transformation with a coefficient of determination (R2) of 0.70, i.e., a Pearson correlation coefficient of 0.83 (Supplementary Table S1). By pre-screening the effective gRNAs, we successfully obtained gene-editing mutants from 16 gRNA, which targeted 10 pairs of soybean homeologs and four single-copy genes (Tables 4 and 5). The gene-editing efficiencies ranged from 2.7% of gRNA6-1 (Table 4) to 80.0% of gRNA9-1 (Table 5).

Discussion
In the study, by using the imbibed seeds as explants, the duration of soybean hairy root transformation procedure was reduced to 2 weeks (Figure 1), which is at least 1 week shorter than the classic seedling-stabbing protocol [33]. A significant positive correlation was detected between the editing efficiencies of gRNAs in hairy root transformation and whole-plant transformation (Table S1). Thus, the modified hairy root transformation protocol can be used to assess the editing efficiency of designed gRNAs prior to wholeplant transformation in soybean.

The Modified Hairy Root Transformation System Not only Shortens the Duration, but also Guarantees the Ratio of Transgenic Roots in the Obtained Hairy Roots
In this study, we used split-seeds as explants for hairy root transformation. The explant preparation method was referred from the A. tumefaciens-mediated whole soybean plant transformation procedure [22,23]. Germline cells infected and transformed by A. rhizogenes here are the same as the reported whole-plant transformation protocol, that is, the cotyledons axillary bud primordium. The major differences between this modified protocol and the classic seedling stabbing protocol [33] include the following: (1) the explants used in our protocol have undergone only 16 h of immersion; (2) the selective agent corresponding to the resistance marker gene contained in the binary vector was added to the Ri medium. As a result, the modification not only shortens the duration of obtaining transgenic roots by one week, but also ensures that the hairy roots obtained are produced by T-DNA insertion. Using the modified protocol, the verification of the hairy roots obtained can then be eliminated.

Impartance of Pre-Evaluation of the Editing Efficiency of gRNAs in Hairy Root Transformation
The target sites for CRISPR/Cas9-based genome editing can be designed manually or assisted by website tools such as CRISPR-P [38] and CRISPR-GE [39]. However, the CRISPR/Cas9 system is unable to edit all targetable genomic sites with full efficiency in vivo due to the genome complexity. Thus, it is necessary to pre-screen the valid gRNAs before whole-plant transformation begins. In this study, 26 soybean genes were simultaneously edited in CRISPR/Cas9 based on hairy root and whole-plant transformation. This study, for the first time, compared the gene-editing efficiencies of a set of gRNAs in the two Agrobacterium species-mediated transformation systems. Result showed that the editing efficiency of a certain gRNA in whole-plant transformation is highly correlated with that in hairy root transformation with a Pearson correlation coefficient of 0.83 (Supplementary Table S1). The use of hairy root system to pre-evaluate the editing efficiency of gRNAs can also be applied to other plant species that are recalcitrant to plant transformation to avoid waste caused by poor gRNA design in the whole plant transgenic process.
In this study, we proposed a new use of hairy root transformation system, which is to pre-assess the effectiveness of gRNAs designed in CRISPR/Cas9-mediated gene-editing vectors before the whole-plant transformation begins. The protocol can screen gRNAs for a better editing efficiency in the target within two weeks. The expected gene-editing plants could be successfully obtained through the transformation of whole soybean plants using the pre-selected gRNAs.
Pre-assessment of gRNA editing efficiency can also be achieved more rapidly using simpler systems. For instance, using Agrobacterium-mediated tobacco infiltration [40], the gRNA and CAS9 proteins can be expressed transiently in tobacco leaves, and the editing efficiency of gRNAs in the targeting gene of the tobacco genome can be assessed within three days. A recent paper successfully used the engineered tomato-spotted wilt virus to deliver the CRISPR/Cas components in various plant species [41]. It provides a promising tool for gene editing in the plant species that are recalcitrant to tissue-culture-based plant transformation in the future.

Plant Materials and Growth Conditions
Soybean cultivar Williams 82 was used for gene-editing experiments. In the experiment, to test the editing efficiency of gRNAs targeting the transgene GUS, transgenic soybean (c.v. Williams 82) harboring a single copy of 35S::GUS expression cassette was used. Soybean plants were grown in a greenhouse at 30 • C day/25 • C night and with cycles of 16 h of light/8 h of dark.

Vector Construction
The binary plasmid pTF102 [23] was used for over-expressing the GUS reporter gene. The vector contains the expression cassettes of a phosphinothricin acetyl transferase (bar) gene conferring resistance to herbicide phosphinothricin, an intron-containing GUS gene in its T-DNA region.
The CRISPR/CAS9 construct was based on the pBlu-gRNA vector and CAS9 MDC123 (Addgene plasmid # 59188 and 59184, Watertown, MA, USA). The target sites were designed using the webtool of http://skl.scau.edu.cn/ (accessed on 15 January 2023) [34]. The target sequences were synthesized and cloned into pBlu/gRNA at the BbsI site and under the control of the U6 promoter. The construct was then digested with EcoRI to generate the gRNA cassette and inserted into destination vector CAS9 MDC123. The sequences of the analyzed soybean genes were downloaded from phytozome (https://phytozome-next. jgi.doe.gov/ (accessed on 15 January 2023)). The resulting constructs were named as "genename-CAS9", e.g., gus-CAS9, pds-CAS9, tt8a-CAS9, and so on. To knockout the pair of the homoeologous genes, gRNA was designed in the common region of the genes. All the corresponding primers are listed in Supplementary Table S2.

Hairy Root Transformation
For hairy root transformation, constructs were transformed into A. rhizogenes strain K599 by the heat shock method.
A. rhizogenes-mediated hairy root transformation was performed according to the literature [37]. The modified protocol uses imbibed seedlings as explants. Briefly, soybean seeds were surface-sterilized for 10-12 h using chlorine gas in a sealed desiccator. The sterilized seeds were imbibed in sterile water at 25 • C for 16 h (overnight). The cotyledons and hypocotyls (about 5 mm) from a single seed were cut evenly into two half-seeds. The prepared explants were then inoculated with A. rhizogenes strain of K599 containing the corresponding CRISPR/Cas9 gene-editing vectors for 30 min. After inoculation, the explants were evenly (adaxial side up) placed into root induction medium containing B5 salts and vitamins, 3% sucrose, 0.8% agar, 0.58 mg/L MES (pH 5.8), filter-sterilized 1.67 mg/L BAP, 250 mg/L cefotaxime, and 5 mg/L glufosinate (the selective agent for the vector), and incubated in a growth room at 25 • C. Hairy roots emerged at 6 DPI and grew to 1-2 cm long at 14 DPI.

Whole Plant Transformation Experiment of Soybean
After assessing the editing efficiency in hairy roots, constructs containing the efficient gRNA were transformed into A. tumefaciens strain LBA4404. The stable transgenics soybean plants were generated via A. tumefaciens -mediated cotyledenary node protocol [23]. Soybean "Williams 82" was used as the transformation recipient. Transgenic plants were identified by selectable cas9 gene amplification and leaf painting with glufosinate (135 mg/L).

Verification of CRISPR/Cas9-Induced Mutations in Transgenic Roots and Plants
To verify the knockouts and determine the genetic modifications in the transgenic roots or plants, Sanger sequencing was carried out on the amplicons that amplified using the PCR primer pairs covering the target regions. Briefly, genomic DNA of the transformed hairy roots or stable transgenic soybean plants was extracted using the TPS [100 mM Tris-HCl (pH 8.0), 10 mM EDTA (pH 8.0), and 1 M KCl] method. PCR amplification was performed using the primers listed in Supplementary Table S2. The reaction conditions were as follows: 95 • C for 2 min, 34× (95 • C for 10 s, 58 • C for 15 s, 72 • C for 15 s), 72 • C for 5 min.
Transgenic roots or plants were first verified for the existence of Cas9 gene using the Cas9-F/R primers. Cas9 positive transformants were further amplified to obtain the 400~800 bp fragments covering the gRNA targeting region. PCR products were then purified and sequenced. When gRNA targets a common region of a pair of homologous genes, gene-specific PCR primers should be designed to distinguish the two genes.

Conclusions
In summary, we reported an efficient soybean hairy root transformation protocol. Using the modified protocol, transgenic hairy roots could be obtained within 2 weeks, which is at least 1 week shorter than the classic seedling-stabbing protocol. More importantly, we found that there is a significant positive correlation between the editing efficiency of gRNAs in hairy root transformation and whole-plant transformation. Thus, the modified hairy root transformation protocol can be used to assess the editing efficiency of designed gRNAs prior to whole-plant transformation in soybean.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/plants12051017/s1, Figure S1. Sequencing results of the targeted region in transgenic hairy roots; Figure S2. GmPDS sequencing results of the targeted region in transgenic hairy roots; Figure S3. Sequencing results of target regions of transgenic hairy roots for the selected sgRNA; Table S1. Correlation of editing efficiency of hairy root transformation and whole plant transformation; Table S2. Primers used in this study.
Author Contributions: Q.K., J.L. and H.S. designed the research, analyzed the data, and wrote the manuscript; S.W. and X.F. helped discussion. All authors discussed the data and reviewed and commented on the manuscript. All authors have read and agreed to the published version of the manuscript.