Heterotrimeric G protein are involved in the regulation of multiple agronomic traits and stress tolerance in rice

Background The heterotrimeric G protein complex, consisting of Gα, Gβ, and Gγ subunits, are conserved signal transduction mechanism in eukaryotes. Recent molecular researches had demonstrated that G protein signaling participates in the regulation of yield related traits. However, the effects of G protein genes on yield components and stress tolerance are not well characterized. Results In this study, we generated heterotrimeric G protein mutants in rice using CRISPR/Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats) gene-editing technology. The effects of heterotrimeric G proteins on the regulation of yield components and stress tolerance were investigated. The mutants of gs3 and dep1 generated preferable agronomic traits compared to the wild-type, whereas the mutants of rga1 showed an extreme dwarf phenotype, which led to a dramatic decrease in grain production. The mutants showed improved stress tolerance, especially under salinity treatment. We found four putative extra-large G proteins (PXLG)1–4 that also participate in the regulation of yield components and stress tolerance. A yeast two hybrid showed that the RGB1 might interact with PXLG2 but not with PXLG1, PXLG3 or PXLG4. Conclusion These findings will not only improve our understanding of the repertoire of heterotrimeric G proteins in rice but also contribute to the application of heterotrimeric G proteins in rice breeding.

generated diversity in grain sizes [16,17], and a deletion of the cys-rich domain of the dep1-1 allele resulted in a dense and erect panicle architecture, and a significant increase of grain number per panicle [18][19][20]. Taken together, the heterotrimeric XLG proteins and noncanonical Gγ subunits that do not exist in the repertoire of heterotrimeric G proteins in animals might identify additional signal transduction pathways and expand the repertoire of heterotrimeric G proteins.
CRISPR/Cas9 systems have been successfully applied as efficient genome editing tools in a number of plant species [21][22][23][24]. The CRISPR/Cas9 gene-editing technology has been used to verify the function of important yield-related genes, such as DEP1, GS3, IPA1, and Gn1a [22]. In the present study, we generated one Gα mutant, one Gβ mutant, five Gγ mutants, and four extra-large Gα mutants using CRISPR/Cas9 gene-editing technology. We assessed the yield components and stress tolerance of the mutants, revealing the distinctly different effects of the heterotrimeric G protein mutants on yield components and stress tolerance. We aimed to gain insight into the heterotrimeric G proteins in the rice genome, with the goal of providing better information and germplasm resources for rice breeding.

Heterotrimeric G protein-encoding gene investigation of the transformation receipt
The japonica rice variety Sasanishiki was used as wild type (WT) for the transformation in this study. Sasanishiki is a typical Japanese japonica variety that was widely cultivated in Northeast China before the 1980s. Sasanishiki then became a major backbone parent, and many rice varieties in Northeast China share the genome segment from Sasanishiki. Sasanishiki thus represents an ideal receipt for transformation in this study, not only for the functional demonstration of heterotrimeric G protein-encoding genes in rice, but also for the genetic improvement by molecular breeding in Northeast China. Thus, we first assessed the heterotrimeric G proteinencoding genes of Sasanishiki. The Gα, Gβ, and five Gγ genes of Sasanishiki were investigated ( Table 1). The results showed that Sasanishiki shared the G protein genes to Nipponbare. No frame shift was detected between Nipponbare and Sasanishiki. Thus, we assumed that all seven G protein genes in Sasanishiki were functional.
Generation of heterotrimeric G protein mutants using CRISPR/Cas9 In order to evaluate the effects of heterotrimeric G protein-encoding genes on rice agronomic traits, we used CRISPR/Cas9 to specifically induce mutagenesis of the G protein. We sequenced 20 plants of each mutant to examine the mutation efficiency in the T 0 generation. The result showed that mutations occurred in 64.29% of plants, and 10.71% of sequenced plants had a putative homozygous mutation ( Fig. 1 and Table 2). Seven T 1 lines with homozygous mutations and the WT were used in further analyses. At least two independent transgenic lines were obtained for each Heterotrimeric G protein encoding gene except for rgg2 (Fig. 1). Forty T 1 plants for each mutant and WT were sown in the experimental field of Shenyang Agricultural University (N41°, E123°) on 23 April 2018. We recorded the heading time when the first panicle had emerged from the sheath. The results showed that the rgb1, rgg1, rgg2, gs3, and dep1 mutants underwent heading significantly earlier than the WT. The rga1 and ggc2 mutants showed a similar heading time as the WT (Fig. 2).

The yield components of heterotrimeric G protein mutants
The heterotrimeric G protein is involved in the regulation of multiple agronomic traits, including grain size and panicle architecture. Thus, we compared the yield components of the seven G protein mutants and the WT (Fig. 2). The results showed that the independent mutant lines exhibited similar phenotypic changes. The rga1 showed a significant dwarf phenotype, with a dramatic decrease in panicle length, grain length, 1000 grain weight, and setting rate compared to that of the WT. The rgb1 mutant exhibited a semi-dwarf phenotype, and the panicle number was less than that of the WT. Sun  [31,32]. However, the three base and six base deletions in rgb1-1 and rgb1-2 did not cause a frame shift, which might be the reason for the survival of rgb1-1 and rgb1-2. In the Gγ subunits, the rgg1 mutant had longer but fewer panicles than that of the WT, and the rgg2 mutant showed a decrease in plant height compared to that of the WT. The ggc2 and dep1 mutants exhibited a decrease in plant height, and the dep1 mutant had a shorter grain but a greater grain number per panicle compared to that of the WT. The gs3 mutant had a higher panicle number per plant and a longer grain length, which led to an increase in the 1000 grain weight compared to that of the WT.
The stress tolerance of the heterotrimeric G protein mutants We then examined the drought, chilling, and salinity tolerance of the heterotrimeric G protein mutants (Fig. 3).
As the independent mutant lines exhibited similar phenotypic changes, we only used the number 1 transgenic line   for each heterotrimeric G protein for the stress tolerance survey. In the drought tolerance survey, the rga1 and dep1 mutants showed an enhanced drought tolerance compared to the WT, whereas the rgb1, rgg1, rgg2, gs3, and ggc2 mutants exhibited similar deficiencies caused by drought. In the chilling treatment test, the rgb1 and ggc2 mutants were more sensitive than the WT. The rga1 and dep1 mutants showed improved chilling tolerance compared to the WT. In the salinity treatments, all of the heterotrimeric mutants showed enhanced salinity tolerance compared to the WT, especially the gs3 and dep1 mutants. Taken together, these results suggest that the heterotrimeric G protein might participate in multiple stress response mechanisms.

The extra-large Gα protein in rice
The bioinformatics analysis showed that there were four putative extra-large Gα protein (pxlg1-4) encoding Fig. 3 The stress tolerance of the heterotrimeric G protein mutants. a The non-treatment control of WT and heterotrimeric G protein mutants. bd Photos of the plants treated with drought, chilling, and salinity stress. e-g The survival rates after drought, chilling, and salinity treatment. Bar = 10 cm, * indicates significance at the 5% level genes in the genome of rice (Table 3). We then generated mutants of four pxlg genes (Fig. 4). The four pxlg mutants showed an early flowering phenotype compared to the WT (Fig. 5). The plxg1 mutant had a shorter panicle but a longer grain than that of the WT. The panicle number of pxlg1 was significantly increased compared to that of the WT. The settings rate of pxlg2 and pxlg3 were decreased compared to that of the WT. The pxlg1 had a longer grain length than that of the WT, that led to an increase in 1000 grain weight. The two independent transgenic lines of pxlg2 showed slight differences in 1000 grain weight, the pxlg2-1 showed significantly decrease in 1000 grain weight, whereas the pxlg2-2 had a similar 1000 grain weight compared to the WT. We also investigated the stress tolerance of the four pxlg mutants (Fig. 6). In the drought treatment, pxlg4 showed enhanced drought tolerance compared to the WT and the other three pxlgs. In the chilling treatment, pxlg4 exhibited improved chilling tolerance compared to the WT.
In the salinity treatment, all four pxlgs showed enhanced salinity tolerance, particularly pxlg3 and pxlg4, which had greater chilling tolerance than that of WT, pxlg1, and pxlg2. We subsequently surveyed whether the PXLGs interacted with RGB1 in rice using a yeast two hybrid analysis. The results showed that RGB1 interacted with PXLG2, but not with PXLG1, PXLG3, or PXLG4 (Fig. 7).

Discussion
The heterotrimeric G proteins participate in a number of biological processes. Previous studies have shown that heterotrimeric G protein mutants exhibited various phenotypic changes. The null mutant of the RGA1 gene exhibits a severe dwarf phenotype in rice [33,34], whereas the down-regulation of the expression of the RGA1 gene causes a semi-dwarf phenotype [19]. Microarray analysis of RGA1 showed that RGA1 might participate in the regulation of multiple abiotic stresses, such as drought, salinity, heat, and cold tolerance [35]. During drought, rga1 mutant plants exhibit greater stomatal conductance than the WT, but both genotypes exhibit the same transpirational water loss per unit leaf area [25]. A recent study showed that RGA1 could regulate photoprotection and photoavoidance in rice [26]. The COLD1 gene could regulate Gα to confer chilling tolerance in rice [36]. Our studies confirmed that the null mutants of RGA1 showed a severe dwarf phenotype and exhibited stress tolerance under drought, chilling, and salinity treatments. The suppression of RGB1 causes dwarfism and browning of the internodes and lamina joint regions [27], and the null mutant of RGB1 was lethal under the genetic background of ZH11 [31,32].
Our study also showed that the mutation of rgb1 caused a dwarfish plant architecture, with the panicle length being shorter than that of the WT. We concluded that rgb1-1 and rgb1-2 harbor mild alleles of RGB1 as the 3bp and 6-bp deletions did not cause frame shifts. Overexpression of RGG2 in Nipponbare led to a reduced plant height and decreased grain size, whereas the mutants generated via CRISPR/Cas9 in the Zhenshan 97 background exhibited enhanced growth, including elongated internodes, increased 1000-grain weight and plant biomass, and enhanced grain yield per plant [29]. However, the rgg2 CRISPR/Cas9 gene-edited plants under the Sasanishiki genetic background showed a similar phenotype to that of WT (Fig. 2). Interestingly, the mutant of rgg1 showed enhanced growth, which was similar to the rgg2 mutants in the Zhenshan 97 genetic background (Fig. 2). These results demonstrate that both RGG1 and RGG2 might act as regulators of plant growth in rice. Recent molecular studies identified two noncanonical Gγ subunits DEP1 and GS3 as major quantitative trait loci (QTLs) [28]. Several studies identified the GS3 locus as a major QTL for grain length, grain width, and grain weight, and DEP1 corresponded to the dense erect panicle and increased grain number [18,30,37]. Our study showed that the knock-out mutation of gs3 could significantly increase the grain length compared to that of the WT, whereas the knock-out mutant of dep1 showed significantly decreased grain length. The dep1 mutants had a shorter panicle than the WT but exhibited a significant increase in grain number per panicle. The mutants of another atypical Gγ protein gene, GGC2, did not exhibit a preferential phenotype as observed in the gs3 and dep1 mutants. The panicle size and setting rate were penalized in the ggc2 mutant. Our study also indicated that the rgb1, rgg1, rgg2, gs3 and dep1 mutants underwent heading significantly earlier than the WT, results that had not been reported in previous studies. The early heading phenotype observed in the present study indicated that the G protein signaling might participate in the control network of plant heading.
Recently, a "self-inhibition" model was constructed to explain the gain of function mutants of dep1 and gs3. The hypothesis is that the C-terminal domains inhibited the N-terminal domains in non-canonical Gγ proteins [12,30]. Although both C-terminal and N-terminal regions is thought to be involved in protein interaction, the molecular mechanisms of the observed selfinhibition of GS3 and DEP1 remain elusive [19,30]. As DEP1 plays a crucial role in the understanding of G protein in rice and in rice breeding, a number of studies employing CRISPR/Cas9 technology have been conducted to verified the function of DEP1 [22,31,[38][39][40][41]. The DEP1 CRISPR/Cas9 gene-edited plants with an eliminated C-terminal exhibited an increase in grain number per panicle, reductions in plant height, panicle length, and grain length, and an erect panicle architecture, whereas the plants that had lost both the Gγ and cys-rich domains showed reductions in setting rate and grain number per panicle. Additionally, we noticed the truncated dep1 with an eliminated C-terminal also exhibited a reduction in grain number per panicle under the genetic background of Nongken 57 and Wuyunjing 8 [42,43]. These results suggested the truncated dep1 alleles could generate opposite effects to grain number per panicle under different genetic background. The present study showed that the mutants with both the Gγ domain and cys-rich domains eliminated showed an increase in grain number per panicle (Fig. 2). Li et al. generated three lines harboring different length of truncation at Cterminal of DEP1 using CRISPR/Cas9 gene editing technology, and the three lines exhibited similar phenotypic changes of the grain size [40]. Taken together, we hypothesized that both the two types of DEP1 alleles (the allele eliminated Gγ and cys-rich domains, and allele eliminated only cys-rich domains) could increase the  Studies of the extra-large Gα protein in Arabidopsis have demonstrated that the heterotrimeric XLG-Gβγ proteins represent additional signal transduction mechanisms in plant heterotrimeric G protein signaling [9][10][11]44].
However, the function of XLGs in the rice genome is unclear. The bioinformatics analysis indicated that there were four putative extra-large Gα protein encoding genes (PXLG1-4) in the genome of rice. We generated the mutants of four pxlg1-4 mutants using CRISPR/Cas9. All four pxlg mutants showed an earlier heading phenotype than the WT, and various phenotypic changes in yield components and stress tolerance were observed in the (See figure on previous page.) Fig. 5 Genetic effects of the putative extra-large Gα protein on plant architecture, panicle size, grain size, and yield components. a The whole plants of the putative extra-large Gα protein mutants. Bar = 10 cm. b The panicles of the putative extra-large Gα protein mutants. Bar = 1 cm. c The grains of the putative extra-large Gα protein mutants. Bar = 1 cm. d-i The plant heights, heading times, panicle numbers, grain numbers per panicle, 1000-grain weight, and setting rates of the putative extra-large Gα protein mutants. * indicates significance at the 5% level Fig. 6 The stress tolerance of the putative extra-large Gα protein mutants. a The non-treatment control of the WT and putative extra-large Gα protein mutants. b-d Photos of plants treated with drought, chilling, and salinity stress. e-g The survival rates after drought, chilling, and salinity stress. Bar = 10 cm. * indicates significance at the 5% level comparison between the WT and mutants. As the yeast two hybrid analysis showed that RGB1 could only interact with PXLG2, we expected that the mutant of pxlg2 could generated obvious agronomic changes compared to other mutants of PXLGs. However, the mutant of pxlg2 showed mediocre phenotypic change among the four mutants of pxlgs, whereas the mutant of pxlg1 showed more severe changes in panicle length and grain size, and mutant of pxlg4 showed improved stress tolerance compared to other mutants of PXLGs. These results indicated that PXLG2 might participate in the regulation of yield components and tress tolerance through the direct interaction with RGB1. Nevertheless, the interaction among four PXLGs needs further study.
Yield is the latest addition to the growing list of critical traits modulated by heterotrimeric G signaling in plants [12]. By combining different G protein variants, the grain length could be decreased by up to 35% or increased by up to 19%, leading to an over 40% decrease or a 28% increase in grain weight [31]. The present study constructed heterotrimeric G protein mutants in an identical genetic background. The effects of the heterotrimeric G protein in the regulation of yield components and stress tolerance were assessed. We found that the gs3 and dep1 mutants showed an ideal phenotype regarding both yield components and stress tolerance. The pxlg4 mutant exhibited similar yield components but enhanced stress tolerance compared to the WT. The grain yield per plant of the WT was 17.38 ± 0.99 g, and those of pxlg1-1 and pxlg1-2 were 29.96 ± 1.59 g and 28.88 ± 0.95 g, respectively. Thus, pxlg1 could improve the grain yield per plant by 69% under our experimental conditions, indicating that PXLG1 has a potential application in high yield rice breeding. Taken together, the results from the manipulation of G protein signaling might lead to improvements in yield and stress tolerance in rice and other crop species.

Conclusions
The present study generated one Gα, one Gβ, and five Gγ mutants using CRISPR/Cas9 gene-editing technology. The Fig. 7 Interaction of RGB1 with PXLGs. In the yeast two hybrid assay, RGB1 is used as a prey (AD), and the PXLGs are used as bait (BD) (a) The corresponding positions in the self-activation test and yeast two hybrid assay. b The result of self-activation test and yeast two hybrid assay. Two plasmids containing either an AD or BD construct were introduced into a yeast strain and transformants were grown on selective medium lacking Leu and Trp. pGBKT7-p53 + pGADT7-T was used as a positive control, and pGBKT7-p53 + pGADT7-LaminA was used as a negative control bioinformatics analysis showed that there were four putative extra-large Gα protein encoding genes (PXLG1-4) in the genome of rice. The yield components and stress tolerance were surveyed. The results showed that the heterotrimeric G proteins were involved in the regulation of yield components and stress tolerance. The yeast two hybrid assay showed that PXLG2 might interact with RGB1. These findings not only improve our understanding of the repertoire of heterotrimeric G proteins in rice but also contribute to the application of heterotrimeric G proteins in rice breeding.

CRISPR/Cas9 vector construction and plant transformation
The experiment was conducted under the genetic background of the Japanese commercial japonica cultivar Sasanishiki. The seed of Sasanishiki (JP number: 5354) was ordered from the Genebank Project, NARO (Tsukuba, Japan). According to the standard material transfer agreement of Genebank, the seeds from Genebank were available to conduct scientific research and education. In order to perform the CRISPR/Cas9 gene editing, we conducted the vector construction as described in our previous study [45]. We conducted the rice transformation according to previously described method [46]. Genomic DNA was extracted from the mutant plants, and primer pairs flanking the designed target site were used for PCR amplification. The 200-500 bp PCR products were sequenced directly using the Degenerate Sequence Decoding method [21]. The primer used in the present study was listed in additional file 1.

Field experiments
Field experiments were conducted at the experimental farm of Shenyang Agricultural University, Shenyang, China (N41°, E123°) in 2018. Two independent transgenic lines for each heterotrimeric G protein were prepared except for rgg2, which only had single transgenic line. Seeds of mutants and WT were sown on 24 April, and transplanting (one seedling per hill) was conducted on 23 May. A randomized block design with three replicates was used in the present study. The plot was 5.4 m 2 and contained 120 plants with 30 cm × 15 cm intervals. The cultivation method and field management were as described previously [38]. We harvested the aboveground parts of 24 plants for each transgenic line at the mature stage (35 days after the full heading stage). The number of grains per panicle and grain-filling percentage were calculated based on the above data.

Stress tolerance investigation
The fully filled and uniform rice seeds of the WT and mutants were washed with 70% (v/v) ethanol for 30 s and then washed three times with sterile water. The stress treatment was conducted 3 weeks after sowing. To test chilling tolerance, 27 seedlings per genotype were treated at 2-4°C for 72 h. Subsequently, they were moved to a temperature-controlled (30°C day/22°C night) greenhouse for recovery. For the drought treatment, 27 seedlings per genotype were exposed to dehydration stress in a greenhouse (30°C day/22°C night) by withholding water for up to 10 d until variation in the stress response was observed. We then added water for recovery. We compared the salinity tolerance of the WT and homozygous mutant plants in the greenhouse, and the three-week-old plants were treated with fresh groundwater and 0.75% NaCl solution (pH = 7), respectively. After 2 weeks of treatment, the plants were treated with fresh groundwater for recovery. The survival rate was determined after 2 weeks of recovery. Each line was replicated three times.

Statistical analyses
Statistical analyses were performed using SPSS version 13.0 (SPSS, Chicago, IL, USA) and analyzed with Student's t test and two-way analysis of variance (ANOVA). Values represent the mean ± standard deviation (SD) of three replicates, and significant differences at the 0.05 level are indicated by asterisks or different letters in the figures.

Yeast two-hybrid analysis
Yeast two-hybrid experiments were performed using the Matchmaker Two-Hybrid System (Clontech). The coding sequences of RGB1, and four PXLGs were amplified using primers listed in additional file 1. Then we cloned the obtained fragments into PGADT7 and PGBKT7 separately. The detail of yeast two hybrid assay was described in elsewhere [31].

Additional file 1. The primer used in present study
Abbreviations CRISPR: Clustered regularly interspaced short palindromic repeats; PXLG: Putative extra-large G protein; QTL: Quantitative trait loci; SD: Standard deviation; WT: Wild type