-
Previous studies have indicated that transformation is much easier and has higher efficiency in japonica than indica rice cultivars[27,28]. Our group previously bred an ideal temperature-sensitive conventional indica cultivar, NGZ[21]. Here we used ZH11, the most popular and widely used japonica cultivar in transgenic rice research, for comparison[29]. ZH11 is a typical japonica cultivar with high native efficiency in Agrobacterium-mediated transformation[29−31]. We selected another five indica rice cultivars for tissue culture evaluation. These cultivars had a range of phenotypes and originated from various regions in South China, Central China, and India. All were mature cultivars that have been consumed for a relatively long time (ranging from the 1960s to 2010s). We compared grain size between these six samples (Fig. 2), and found that they were highly diverse in critical grain-related agronomic traits (Supplemental Table S4). NGZ contained higher alkali spreading value and brown rice rate, lower milled width, chalky grain rate, chalkiness area, and amylose content among these five indica cultivars. Therefore, NGZ is a typical high grain quality rice cultivar that is suitable for functional studies and improvement experiments related to grain size and shape. It has a very low chalky grain rate (1.5%) and chalkiness area (0.5%), and these attributes, along with the shape and milled performance, are desirable to consumers. NGZ also has relatively low amylose content and protein content, giving it superior flavor (Supplemental Table S4)[32,33]. Based on the grain performance (Fig. 2), NGZ appears to be an excellent novel elite rice cultivar for studying grain size and panicles.
Plasmid constructs and transgenic plant development
-
We previously reported that Grain Size and Number 1 (GSN1), which encodes the MAPK phosphatase OsMKP1, negatively regulates the MAPK cascade and controls trade-offs between grain size and number[19]. We here used two plasmids to generate transgenic plants and evaluate tissue culture performance (Supplemental Fig. S1): one for GSN1 overexpression and the other for GSN1 knockout via CRISPR-Cas9 gene editing (Supplemental Fig. S1). A functional gene editing system together with an overexpression system would enable most molecular mechanistic studies in rice. The components of the callus induction medium, resuspension medium, regeneration medium, selection medium, differentiation medium, and root induction medium have been optimized for japonica and indica rice cultivars (Supplemental Table S1).
Callus induction evaluation
-
We put all indica rice cultivars into some aspect of callus induction medium (CIM) to perform callus induction assays. The CIM used for ZH11 was slightly modified with respect to the ratio of sucrose, glucose, and phytohormone (Supplemental Table S1). One hundred healthy, intact grains were selected from each rice cultivar for each transformation event, and there were five biological replicates (500 seeds in total per cultivar). When seeds were cultured on CIM, more than 65% of total seeds were induced into embryogenic calli (Tables 1 & 2). We performed statistical analysis to display the survival rates of all plant tissue culture stages among six cultivars (Fig. 3). After growing on CIM for 14 d, we removed the endosperm and roots to obtain pure callus tissue for the next stage of tissue culture. The japonica cultivar ZH11 had the highest callus induction rate at 88.7%. In comparison, NGZ and ZS97 had similarly high calls induction rates (82%, NGZ; 88.3%, ZS97) (Tables 1 & 2). The Kasalath rice cultivar had the lowest callus induction rate (64.2%). We observed yellowish, compressed callus shapes in most indica cultivars and ZH11. Only very small calli were generated from Kasalath (Supplemental Fig. S2). In general, most indica cultivars formed calli earlier than ZH11 (Supplemental Table S5). Based on the phenotypical and statistical analysis as well as the callus induction rate, we admitted that the six cultivars ranked as follows: ZH11 (88.7%) > ZS97 (83.3%) > NGZ (82%) > YD 6 (77.6%) > HHZ (72.7%) > Kasalath (64.2%) (Fig. 3, Supplemental Fig. S2, Tables 1 & 2). These results confirmed that the indica cultivar NGZ had excellent callus induction capacity under commercial CIM conditions.
Table 1. Number of samples and survival rate percentage at each transformation step with the pYLCRISPR/Cas9Pubi-GSN1-H (GSN1-KO) plasmid in six rice cultivars.
Sample name Sample quantity Mature seeds Callus induction Resistant callus Green spot Shoot induction Transformation NGZ 1000 (100%) 820 (82%)** 111 (26.7%)** 34 (30.6%)n.s 24 (70.5%)n.s 24 (4.8%)n.s ZS97 1000 (100%) 833 (83.3%)* 271 (62.0%)* 41 (15.1%)n.s 26 (63.4%)n.s 26 (5.2%)n.s YD 6 1000 (100%) 776 (77.6%)*** 54 (14.5%)*** 12 (22.2%)* 6 (50%)* 6 (1.2%)* HHZ 1000 (100%) 727 (72.7%)*** 261 (71.1%)* 36 (13.7%)n.s 22 (61.1%)n.s 22 (4.4%)n.s Kasalath 1000 (100%) 642 (64.2%)*** 196 (62.2%)n.s 140 (71.4%)* 17 (12.1%)n.s 17 (3.4%)n.s ZH11 1000 (100%) 887 (88.7%) 189 (41.6%) 55 (29.1%) 24 (43.6%) 24 (4.8%) Data are given as the mean. n.s, no significance, * P < 0.05; ** P < 0.01; *** P < 0.001 compared with the corresponding ZH11 using Student’s t-test. The callus induction rate was caculated from the overall induced callus which was used for the whole studies. Table 2. Number of samples and survival rate percentage at each transformation step with the pCAMBIA1300-GSN1-FLAG (GSN1-OX) plasmid in six rice cultivars.
Sample name Sample quantity Mature seeds Callus induction Resistant callus Green spot Shoot induction Transformation NGZ 1000 (100%) 820 (82%)** 217 (53.5%)* 69 (31.8%)* 39 (56.5%)n.s 39 (7.8%)n.s ZS97 1000 (100%) 833 (83.3%)* 283 (71.5%)*** 56 (19.8%)* 52 (92.9%)n.s 52 (10.4%)n.s YD 6 1000 (100%) 776 (77.6%)*** 243 (60.14%)*** 60 (24.7%)* 17 (28.3%)n.s 17 (3.4%)n.s HHZ 1000 (100%) 727 (72.7%)*** 146 (40.6%)n.s 62 (42.5%)* 23 (37.1%)n.s 23 (4.6%)n.s Kasalath 1000 (100%) 642 (64.2%)*** 229 (45.8%)* 64 (27.9%)* 13 (20.3%)* 13 (2.6%)* ZH11 1000 (100%) 887 (88.7%) 160 (36.9%) 206 (161.5%) 37 (17.9%) 37 (7.4%) Data are given as the mean. n.s, no significance, * P < 0.05; ** P < 0.01; *** P < 0.001 compared with the corresponding ZH11 using Student’s t-test. The callus induction rate was caculated from the overall induced callus which was used for the whole studies. Figure 3.
Statistical analysis of plant tissue culture among six cultivars for the pYLCRISPR/Cas9Pubi-GSN1-H (GSN1-KO) plasmid and the pCAMBIA1300-GSN1-FLAG (GSN1-OX) plasmid. (a) The callus induction rate of six rice cultivars; (b) The resistant callus induction rate of six cultivars; (c) The green spot emerging rate of six cultivars; (d) The shoot induction rate of six cultivars. For callus induction process, ten independent biological repeats were performed (n = 10). For specific plasmid transformation process, five independent biological repeats were carried (n = 5), each performance utilized 100 healthy intact seeds from every cultivar for plant tissue culture process. Data are presented as the mean ± standard deviation. * < 0.05, ** < 0.01, *** < 0.001 (Student's t-test).
Survival rates of transformed resistant calli
-
We next transformed Agrobacterium tumefaciens strain EHA105 with either the overexpression plasmid pCAMBIA1300-GSN1-FLAG (GSN1-OX) or the gene editing knockout plasmid pYLCRISPR/Cas9Pubi-GSN1-H (GSN1-KO). All calli from each rice cultivar were incubated with transformed Agrobacterium in a co-cultivation medium under dark conditions at 28°C for one day. Calli were then transferred to resistant calli selection medium and grown for two weeks under dark conditions for the first round of selection. We observed that fresh calli of different cultivars differed in appearance and behavior. ZH11 calli were more yellow and compressed, had longer induction times, and generated multiple small fresh calli. Calli of the five indica cultivars were generally a lighter yellow (although some were white) and larger. The indica cultivars calli displayed severe browning or even death over the course of the cultivation time. We observed that in comparison to the other rice cultivars, NGZ was more sensitive to Agrobacterium infection (Tables 1 & 2). Statistical analysis showed that except HHZ, the overall resistant calli induction rate was lower for those transformed with the CRISPR plasmid than the overexpression plasmid (Fig. 3). NGZ and YD 6 had the two lowest resistant calli induction rates in GSN1-KO transformation event (Fig. 3). Many NGZ calli had Agrobacterium contamination; kanamycin was insufficient to inhibit agrobacterial growth. Successfully induced NGZ resistant calli were dry and clean (Supplemental Fig. S3). A similar phenomenon was observed in YD 6, but the percentage of contaminated calli was lower than the percentage of brown calli. ZS97 had the best performance, with a higher resistant calli induction rate than even ZH11 (Tables 1 & 2). Kasalath had similar responses to transformation with the two plasmids, whereas NGZ, HHZ and YD 6 had different responses to transformation with the gene editing and overexpression plasmids (Fig. 3). With respect to regeneration rate, the six cultivars ranked as follows: ZS97 (62%, 71.5%) > Kasalath (62.2%, 45.8%) > HHZ (71.1%, 40.6%) > ZH11 (41.6%, 36.9%) > NGZ (26.7%, 53.5%) > YD 6 (14.5%, 60.14%) (Tables 1&2). Overall, these results suggested that transformation with the gene editing plasmid led to a lower resistant calli induction rate, and that NGZ calli were sensitive to Agrobacterium infection.
Differentiation evaluation
-
After 2−3 weeks of regeneration growth, calli were transferred to differentiation medium and incubated at 28 °C under full-light conditions. Interestingly, indica cultivar Kasalath calli transformed with the GSN1-KO construct and japonica cultivar ZH11 calli transformed with the GSN1-OX construct showed large number of green spots (Supplemental Fig. S4). Consistent with previous reports, indica cultivars had lower greening numbers than japonica cultivars[34]. Although a considerable number of regenerated NGZ calli transformed with the GSN1-KO construct died, more than 30% of the calli exhibited their greening area. Kasalath had the highest greening rate at 71.4%. ZS97 and HHZ had the lowest percentage of greening (15.1% and 13.7%). With respect to greening rate, the six cultivars were ranked as follows: Kasalath (71.4%) > NGZ (30.6%) > ZH11 (29.1%) > YD 6 (22.2%) > ZS97 (15.1%) > HHZ (13.7%) (Table 1).
Calli transformed with the GSN1-OX construct had higher greening rates compared to those transformed with GSN1-KO. Surprisingly, ZH11 had a higher greening rate than calli regeneration rate. This phenomenon may be due to the high regeneration capacity of calli and the induction rate in japonica cultivars (Table 2). Furthermore, NGZ had the highest greening amount among all indica cultivars tested, indicating that NGZ had robust viability in plants and calli. Although the resistant callus induction rate was relatively low in HHZ, the greening rate was 42.5% for GSN1-OX transformants, suggesting that HHZ also had a strong capacity for greening (Table 2). We further investigated the size and shape of calli for all six cultivars. NGZ and ZS97 had larger calli and greening areas, whereas YD 6, HHZ, and Kasalath had green spots on smaller calli. Moreover, in comparison to ZH11, all indica cultivars showed darker calli and few were newly generated (Supplemental Fig. S4). With respect to differentiation, the six cultivars ranked as follows: ZH11 (161.5%) > HHZ (42.5%) > NGZ (31.8%) > Kasalath (27.9%) > YD 6 (24.9%) > ZS97 (19.8%) (Table 2).
Shoot and root growth evaluation
-
After 2−3 weeks of growth in differentiation medium, young plants were transferred into shoot induction medium (SIM). During this stage, there was a difference in shoot formation between GSN1-KO and GSN1-OX transformants (Tables 1 & 2). There were fewer GSN1-KO than GSN1-OX lines, and the shoot formation time was longer for GSN1-KO transformants. Although Kasalath plants showed more than 140 greening spots in total, only 17 transformed plants formed shoots (12.1%). Aside from Kasalath, the indica cultivars had higher shoot formation rates than the japonica cultivar ZH11 (Table 2). Similarly, in the GSN1-OX transformation assay, Kasalath had the lowest shoot induction rate (20.3%). Surprisingly, the ZH11 shoot induction rate in GSN1-OX (17.9%) was lower than the shoot induction rate with the GSN1-KO plasmids (43.6%) (Tables 1 & 2). ZS97 had the highest shoot formation rate (92.9%). Moreover, shoot and root lengths were longer in GSN1-OX than GSN1-KO transformants across all cultivars (Supplemental Fig. S5). Hyperhidrosis was not observed at any stage of the plant tissue transformation process. Plants that formed shoots and roots were selected with Hygromycin B. All shoot-forming plants counted in Tables 1 & 2 were confirmed as transformants. Notably, quite amount of transformed YD 6 calli and plants were killed by Hygromycin B in the shoot formation medium, consistent with previous research[35].
T1 transgenic plant phenotyping
-
The T0 seedlings of the six cultivars bearing the GSN1-OX and GSN1-KO plasmids were grown in Hainan (China). T1 seeds were harvested from T0 plants. Leaf pieces (2 cm) from each transformed plant were collected for Hygromycin B verification and PCR product amplification (Supplemental Fig. S6a and S6b). The results showed that all samples had been successfully transformed and revealed alterations at the molecular level. Sanger sequencing was conducted to verify the GSN1-KO constructs in all six cultivars (Supplemental Fig. S7); the results showed that all transformed lines had been successfully modified, functionally disabling GSN1 in those lines. Seed phenotypes were measured and analyzed in all wild-type and transgenic lines in the six cultivars (Fig. 4). The positive transgenic plants' phenotypes are well associated with genotypical alteration. We also performed qRT-PCR and confirmed that the GSN1-OX transformed plants in six cultivars have successfully increased the expression level of GSN1 (Supplemental Fig. S8). Two independent lines derived from each cultivar were selected for each construct to further investigate GSN1 function. Grain length was similar in all rice cultivars; the GSN1-OX and GSN1-KO lines had reduced and increased grain length, respectively. However, Kasalath and HHZ did not show significant grain elongation because the transgenic lines were heterozygous. ZS97, YD 6, and Kasalath showed decreased grain width in GSN1-OX lines, whereas GSN1-KO lines had wider grains. Interestingly, two high grain quality rice cultivars, NGZ and HHZ, showed different trends in seed width in GSN1-OX lines compared to the other cultivars. However, the GSN1-KO lines had wider grains in these cultivars, similar to the other cultivars. Grain thickness was similar between cultivars, except HHZ, for each type of transformant. In general, the GSN1-OX lines had thinner grains than the wild-type lines, whereas GSN1-KO lines had thicker grains. In contrast, HHZ showed the opposite effects in the OX and KO lines. In summary, transforming rice plants with a GSN1 gene editing knockout or overexpression construct led to changes in seed shape across all six rice cultivars, as previously reported[19].
Figure 4.
Phenotypes of T2 seeds of the six cultivars used in this study. For each cultivar, the average grain length (at left in pink, a1−f1), grain width (at middle left in blue, a2−f2), and grain thickness (at middle right in green, a3−f3) is shown for the wild-type, two GSN1-overexpression lines (OX-1 and OX-2), and two GSN1-knockout lines (KO-1 and KO-2). On the right (a4−f4) are grain phenotypes in the wild-type and mutant lines for each cultivar. Scale bar = 1 cm, n = 10.
Expression patterns and single nucleotide polymorphisms (SNPs) in potentially essential genes contributed to rice callus transformation efficiency
-
To further explore the molecular basis of differences in transformation efficiency between the selected indica and japonica cultivars, we performed quantitative reverse transcription PCR (qRT-PCR) to verify expression of callus formation-related genes and to establish the associations between gene expression patterns and cultivars (Fig. 5). The expression of our callus-related genes, including OsCRL, OsBOC1, OsBBM3, and OsSET1, were evaluated[14,36−38] (Fig. 5). The expression of OsBBM3 was significantly higher in callus samples from Kasalath compared to the other cultivars. Also, Kasalath owned the highest expression among all six cultivars of these four genes. The rest callus samples showed relatively lower expression levels (Fig. 5). We also found that HHZ has different expression patterns in OsBOC1, OsCRL and OsSET1 (Higher expression in OsCRL, but lower expression in OsBBM3, OsBOC1 and OsSET1). Except the OsBOC1, the expression of other three genes in YD 6 exhibited significantly lower than that in other cultivars. To further investigate the associations of haplotypes, expression levels and phenotypes, we performed Sanger sequencing to identify insertion/deletion mutations (InDels) and SNPs in these genes (Fig. 6). We cloned the promoter regions, untranslated regions (UTRs), coding regions, and introns of these genes. The sequencing results showed that two OsCRL SNPs were non-synonymous. N143D seemed to exist in japonica and indica, position 143 Asparagine (D) seems to be existed in japonica, and position 143 Aspartic acid (D) are shown in indica cultivars. In this gene, H114N is also another SNP that appeared in different cultivars. Position 114 Aspartic acid appeared only in NGZ and Kasalath (Fig. 6a). OsBOC1 encodes a SIMILAR TO RADICAL-INDUCED CELL DEATH ONE (SRO) protein that contributes to callus browning[14]. We identified two SNPs in coding regions, which is R32S and P377L. In our sequencing results, the R32S and P337L only existed in NGZ and Kasalath (Fig. 6b), the other rice cultivars did not show significant SNPs or InDels at that locus. We did not identify nonsynonymous SNPs or InDels in the coding regions of OsBBM3 or OsSET1 among the tested cultivars (Fig. 6c & d). In summary, our results revealed that the callus induction-related genes might contribute to the callus browning and induction rate variation in japonica and indica cultivars.
Figure 5.
Quantitative reverse transcription PCR (qRT-PCR) analysis of the callus related genes. Samples are harvested from calli after 7 d in CIM. The Actin gene was used as internal control to normalize gene expression data. Values are given as the mean ± SD. * P < 0.05; ** P < 0.01 and *** P < 0.001 compared with the ZH11 control using Student’s t-test. Standard deviations were calculated from three biological replicates.
Figure 6.
Single nucleotide polymorphisms (SNPs) and insertion/deletion mutations (InDels) in four genes related to plant tissue culture in seven cultivars. The key and main SNPs and InDels are shown in the promoter regions, untranslated regions (UTRs), coding regions, and introns. Genes were analyzed for the six cultivars included in the experimental section of this study using Nipponbare as the reference genome. Results are shown for (a) OsCRL, (b) OsBOC1, (c) OsBBM3, and (d) OsSET1. Bases shown in red in (a) and (b) indicate potential functional SNPs that may contribute to callus induction and browning.
-
In summary, we here used GSN1 as a case study to compare the regeneration efficiency of one japonica and five indica rice cultivars. Each cultivar was transformed with a GSN1-overexpression plasmid and a CRISPR-Cas9 GSN1-knockout plasmid. Plants transformed with the gene editing plasmid had lower overall regeneration rates than those with the overexpression plasmid. NGZ and ZS97 had the highest transformation efficiency of the indica cultivars, but both required longer culture times than the japonica cultivar. We also verified expression patterns, SNPs and InDels of selected genes related to callus regeneration among these cultivars and provided possible explanations for differences in transformation efficiency between indica cultivars. Finally, we discussed the disadvantages of NGZ callus browning. We also discussed the advantages and possibility to apply NGZ in the future gene editing and functional studies. Our findings related to NGZ callus regeneration and reducing callus browning are expected to increase the application of NGZ in rice transformation assays.
-
About this article
Cite this article
Chen K, Ye C, Guo J, Chen D, Guo T, et al. 2023. Agrobacterium-mediated transformation efficiency and grain phenotypes in six indica and japonica rice cultivars. Seed Biology 2:4 doi: 10.48130/SeedBio-2023-0004
Agrobacterium-mediated transformation efficiency and grain phenotypes in six indica and japonica rice cultivars
- Received: 20 December 2022
- Accepted: 30 January 2023
- Published online: 30 March 2023
Abstract: Plant transformation and regeneration have been continuously developed over the past four decades. In rice (Oryza sativa L.), Agrobacterium-mediated transformation has high efficiency in japonica and some indica cultivars using mature seeds and immature embryos. However, these protocols have low transformation efficiency in the latest indica cultivar that has been developed in South China since 2010. Here, we explored plant culture regeneration of the high-quality and high-yield indica cultivar Nanguizhan (NGZ) through CRISPR/Cas-based genome editing and traditional overexpression transformations. We compared transformation efficiency in this cultivar to four other widely-used indica cultivars and one japonica cultivar using mature seeds and the gene Grain Size and Number 1 (GSN1) as a case study. We observed universally smaller grain size in overexpression lines and bigger grain size in gene-editing lines among different cultivars. NGZ phenotypes make it an excellent model in which to investigate gene functions. We also examined the single nucleotide polymorphism (SNP) distribution and differences in expression levels of regeneration-related genes in calli, possibly revealing the source of NGZ's advantages in Agrobacterium-mediated transformation. These results shed light on the advanced application of NGZ in gene editing and overexpression transformation related to grain improvement, contributing to the 'rice breeding 4.0 era'.
-
Key words:
- Rice /
- Plant tissue culture /
- Agrobacterium-mediated transformation /
- Grain size /
- CRISPR-Cas9