Overexpression of the leucine‐rich receptor‐like kinase gene LRK2 increases drought tolerance and tiller number in rice

Summary Drought represents a key limiting factor of global crop distribution. Receptor‐like kinases play major roles in plant development and defence responses against stresses such as drought. In this study, LRK2, which encodes a leucine‐rich receptor‐like kinase, was cloned and characterized and found to be localized on the plasma membrane in rice. Promoter–GUS analysis revealed strong expression in tiller buds, roots, nodes and anthers. Transgenic plants overexpressing LRK2 exhibited enhanced tolerance to drought stress due to an increased number of lateral roots compared with the wild type at the vegetative stage. Moreover, ectopic expression of LRK2 seedlings resulted in increased tiller development. Yeast two‐hybrid screening and bimolecular fluorescence complementation (BiFC) indicated a possible interaction between LRK2 and elongation factor 1 alpha (OsEF1A) in vitro. These results suggest that LRK2 functions as a positive regulator of the drought stress response and tiller development via increased branch development in rice. These findings will aid our understanding of branch regulation in other grasses and support improvements in rice genetics.


Introduction
Rice (Oryza sativa) is one of the most important crops, feeding more than one-third of the world's population. This major staple requires large amounts of water during growth and is therefore susceptible to drought stress. At each stage of growth, drought hinders crop development and decreases yield. Molecular genetic tools aimed at improving rice drought tolerance and maintaining production, while expanding development into regions with limited water resources is therefore important (Fernie et al., 2006;Pennisi, 2008).
Eukaryotic elongation factor (eEF) proteins can be divided into eEF1 and eEF2. eEF1 proteins, which can further be divided into eEF1A, eEF1Ba, eEF1Bb and eEF1Bc subunits, are highly conserved in a number of species (Browning, 1996). eEF1-mediated ammonia acyl-tRNA has also been shown to bind to ribosomes (Riis et al., 1990). Moreover, recent studies have shown that eEF1A is not only important for translation, but is also an important multifunctional protein (Ejiri, 2002;Sasikumar et al., 2012), playing a role in processes such as cell proliferation (Pecorari et al., 2009;Sanders et al., 1992), cell apoptosis (Byun et al., 2009;Shepherd et al., 1989;Zhang et al., 2015), cell morphogenesis (Gross and Kinzy, 2005) and signal transduction (Numata et al., 2000). However, most of these studies have focused on animals and humans, with few findings in plants. In Arabidopsis, eEF1B is associated with cell wall biosynthesis and plant development (Hossain et al., 2012). Moreover, the AtEF2 gene has been shown to be involved in low temperature signalling (Guo et al., 2002). Since 1998, four EF1A genes have also been cloned in rice, but their functions have yet to be reported (Kidou and Ejiri, 1998).
Previously, an eight leucine-rich LRK gene (LRK1-LRK8) cluster, which has been shown to increase grain yield, was cloned from the rice quantitative trait locus (QTL) qGY2-1 (Li et al., 2002a). Haplotype divergence of the LRK locus was subsequently found to be associated with the origin and differentiation of cultivated rice. Moreover, LRK2 was found to be highly expressed in Oryza sativa L. ssp. indica var. 9311, but was undetectable by RT-PCR in Oryza sativa L. ssp. japonica cv. Nipponbare (He et al., 2006). In the present study, we cloned and characterized LRK2 and revealed an increase in branch number and drought tolerance in LRK2-overexpressing transgenic lines. pLRK2::GUS was largely expressed in tiller buds, nodes, roots and anthers. Furthermore, yeast two-hybrid screening and bimolecular fluorescence complementation (BiFC) analysis suggested an interaction between LRK2 and OsEF1A. The molecular mechanisms underlying the response of rice to drought was also discussed with the aim of improving crop growth under potentially adverse conditions.

Features of LRK2 in rice
To confirm the subcellular localization of LRK2, LRK2 with an enhanced GFP was constructed under control of the cauliflower mosaic virus 35S promoter (35S::LRK2:eGFP), then infiltrated into tobacco (Nicotiana benthamiana) leaves by Agrobacteriummediated transient transformation. The resulting construct, 35S::LRK2:eGFP, exhibited eGFP expression in the plasma membrane compared with the control, 35S::eGFP, which showed expression throughout the cell (Figure 2A). In addition, to determine expression patterns, the 2-kb promoter region of the LRK2 gene was cloned into the expression vector pBIN121 with the b-glucuronidase (GUS) reporter gene ( Figure S1A). The construct pLRK2:GUS was subsequently transformed into rice ( Figure S1B). A GUS staining assay of T2 transgenic lines revealed expression in the tiller buds, nodes, roots and anthers ( Figure 2B).
To examine whether LRK2 expression was regulated by drought stress, 7-day-old rice seedlings were subjected to 20% PEG6000 then harvested for RNA extraction at different time points and the transcripts of whole plants were quantified by realtime PCR. As a result, transcript levels were found to be rapidly and strongly induced by drought ( Figure 2C).

Expression of LRK2 in transgenic rice lines
To obtain further insight into LRK2, 2X35S::LRK2 and 2X35S:: antiLRK2 (Figure 3a), plasmids containing the entire LRK2 gene were introduced into Nipponbare, and transformants selected on medium containing hygromycin. The transgenic plants were simultaneously examined by PCR using genomic DNA as a template with specific primers. Eleven and eight independent transformants (T 0 ) were regenerated from hygromycin-resistant calli of the 2X35S::LRK2 and 2X35S::antiLRK2 lines, respectively ( Figure S2). Four T 2 transgenic lines plus the wild type were subsequently selected for expression analysis (M2 and M6 to represent 2X35S::LRK2, and AM5 and AM8 for 2X35S::antiLRK2). Expression levels during the three-leaf stage were determined by semiquantitative and RT-PCR. LRK2 was strongly expressed in M2 and M6, but reduced in AM5 and AM8 (Figure 3b,c), suggesting that the 2X35S::LRK2 and 2X35S::antiLRK2 plasmids were genetically transmitted to the next generation.

Overexpression of LRK2 increases drought tolerance in rice
LRKs reportedly regulate a number of stress responses, including drought, salinity and low temperature (Ouyang et al., 2010;Shen et al., 2015;Yang et al., 2014). The performance of LRK2overexpressing, LRK2-antisense and wild-type plants under drought stress conditions was therefore examined (Figure 4). Under normal conditions, all plants grew well. After 12 days, plants were cultured in 20% PEG6000 solution and then 5 days later, phenotypic changes in wild-type and transgenic plants were observed. Leaf rolling and wilting was delayed in the M2 and M6 transgenic lines compared with the wild type, while in AM5 and AM8 symptoms of drought stress were severe. After 8-day treatment, the M2 and M6 lines showed increased tolerance to drought stress compared with the wild type, while AM5 and AM8 remained sensitive.
Following drought treatment, the plants were watered to induce recovery, and then, the growth status was examined. M2 and M6 recovered and grew more vigorously than the wild type, with survival rates of 73.5% and 65.5%, respectively, compared with 49.2%. In contrast, only 12.2% of the AM5 and 10.2% of the AM8 plants recovered, compared with 28.6% of the wild type. All values were significantly different (P < 0.05, t-test; Figure 5A), suggesting that LRK2 positively regulates the drought stress response in rice.
Large root systems are known to be more conducive to extraction of water from deep soil layers compared with small root systems. The root structures of transgenic and wild-type plants after drought treatment were therefore examined under a stereoscopic microscope. The transgenic plants had significantly more lateral roots and a larger overall root system than the wild type ( Figure 5B). Root activity in rice at different stages directly affects plant growth. In this study, root activity appeared higher in LRK2-overexpressing lines compared with the wild type (Figure 5C). These results suggest that increased root activity in the LRK2-overexpressing lines may be one reason for the increase in lateral root development. To analyse the expression of genes involved in drought tolerance, expression levels of 9-cis-epoxycarotenoid dioxygenase 1 (NCED1), NCED2, plasma membrane intrinsic protein 2;3 (PIP2;3) and ABA-responsive element binding factor (ABF) were detected using real-time PCR in wild-type and transgenic seedlings under drought (20% PEG6000) for 12 h (Hatmi et al., 2015;Redillas et al., 2012;Shi et al., 2015;Wu et al., 2015;Yoshida et al., 2010). The data showed that PIP2;3 was strongly expressed in LRK2 overexpression lines ( Figure S3).

Effect of LRK2 on rice tiller development
To investigate the effect of LRK2 on yield traits, 14 transgenic lines were analysed (M1 to M6, AM1 to AM8; 20 individual plants per line). After cultivating transgenic lines and Nipponbare wildtype plants under identical conditions, the following yield components were examined: numbers of tillers per plant, grains on the main panicle, grains per panicle and grains per plant ( Table 1). The tiller is a specialized grain-bearing branch that forms on unelongated basal internodes. At the tillering stage, the LRK2-overexpressing lines exhibited increased tiller development ( Figure S4A). At maturity, M2 and M6 produced 36% and 32% more panicles than the wild type, respectively, while AM5 and AM8 showed a decrease in panicles of 29.7% and 44.9% compared with the wild type, respectively (  Figure S4C). The number of grains per panicle is affected by the number of primary, secondary and sometimes higher-order panicle branches. The number of primary and secondary panicle branches was therefore counted. The overexpressing lines showed a slight increase in the number of secondary branches on the main panicle compared with the wild type (M2: 7.86%; M6: 6.46%); however, the number of primary branches did not significantly differ (Table 2, Figure S4C). These results suggest that LRK2 improves rice tiller number and the number of grains per plant.
Interaction between LRK2 and eukaryotic elongation factor 1A To identify interactors of LRK2, LRK2 kinase domain (LRK2D) was used as bait in yeast two-hybrid analysis. As a result, six potential interacting molecules were identified. Nucleotide sequence analysis further revealed one cDNA fragment encoding the eukaryotic elongation factor OsEF1A (GenBank Accession no. GQ848073.1), which encodes a protein involved in protein synthesis and cell proliferation. To confirm the direct interaction between LRK2 and OsEF1A, a yeast two-hybrid assay was performed via cotransformation of the pGBKT7-LRK2D bait construct and full-length pGADT7-OsEF1A. As expected, only yeast cells harbouring both LRK2D and OsEF1A grew vigorously on both SD/Leu-Trp-and SD/Leu-Trp-His-/ 50 mM 3-amino-1,2,4-triazole (3-AT) plus X-gal media. In contrast, the wild type grew well on SD/Leu-Trp-but not SD/ Leu-Trp-His-/50 mM 3-AT/ X-gal medium ( Figure 7A). pGBKT7-LRK2D and a truncated version of OsEF1A were also cotransformed with pGADT7, revealing that the C-terminal region alone is sufficient, and indeed necessary, for binding with the LRK2 kinase domain ( Figure 7B). To further verify the interaction between LRK2D and OsEF1A, BiFC analysis of Nicotiana benthamiana leaves was performed ( Figure 7C). Under 513 nm illumination, cells harbouring LRK2D:YEP CE and OsEF1A:YEP NE emitted yellow fluorescence, confirming the interaction between OsEF1A and LRK2.

Discussion
It is predicted that global climate change will increase temperatures, alter geographical patterns of rainfall and increase the frequency of extreme climatic events (Harrison et al., 2014). Drought is a major constraint of crop development and production worldwide. Accordingly, a number of studies have suggested that overexpression of stress-related genes may help improve drought tolerance in cereal crops (Cheng et al., 2015;Uga et al., 2013). In this study, a new rice leucine-rich repeat receptor-like kinase, LRK2, was found to have the ability to increase drought stress by promoting root growth and significantly increasing tiller number, while reducing plant height (Figures 6 and S4; Table S1). The data indicated that the LRK2 gene encodes a protein localized to the plasma membrane, and expressed in tiller buds, nodes, roots and anthers (Figure 2). These results suggest that the LRK2 gene is essential for stable and adequate crop production in drought-prone areas. The mature rice fibrous root system is composed of adventitious and lateral roots. Adventitious root branching results in both large and small lateral roots (Coudert et al., 2010), which are essential for water and nutrient uptake and critical for increased yield under stress (Atkinson et al., 2014;Coudert et al., 2010). In rice, DRO1 was previously found to be negatively regulated by auxin and involved in cell elongation, resulting in an enlarged root system and increased drought avoidance (Uga et al., 2013). Moreover, rice OsAHP1 and OsAHP2 knockdown plants were previously found to exhibit phenotypes representative of a deficiency in cytokinin signalling, including enhanced lateral root growth and resistant to osmotic stress compared with wild-type plants (Sun et al., 2014). Moreover, ectopic OCI expression was found to increase the lateral root density and drought tolerance in Arabidopsis and soya bean (Quain et al., 2014). In the present study, LRK2 overexpression resulted in an improved root system with an increased number of both large and small lateral roots compared with the wild type. This is one reason for the increased drought tolerance in LRK2-overexpressing lines. The 2,3,5triphenyltetrazolium chloride (TTC) test was used here to determine root activity in the LRK2-overexpressing plants (Hu et al., 2016;Steponkus and Lanphear, 1967). Roots of the transgenic plants were significantly more active than those of the wild-type plants ( Figure 5C), further contributing to increased drought resistance in the LRK2-expressing lines. PIP2;3, one of the aquaporins, plays a crucial role in response to drought stress (Yu et al., 2006). SIRK1, a member of the LRK family in Arabidopsis, was shown to interact with and activate PIP2;3 by phosphorylation (Wu et al., 2013). Experiments should be undertaken to evaluate whether LRK2 can specifically phosphorylate PIP2;3.
Plant architecture, such as the structure of the roots, shoots and inflorescences, is affected by branching. Shoot branches in rice are referred to as tillers (Tanaka et al., 2015). Root branches are located underground, while tiller and inflorescence structures in rice undergo lateral branching above ground during the vegetative and reproductive stages, respectively. Tiller number is generally regarded as the determining factor of yield as tillers are specialized panicle-bearing branches (Grillo et al., 2009). Several genes related to tiller development have been characterized; for example, MOC1 is important in both tiller bud formation and outgrowth, while OsTB1 negatively regulates axillary bud outgrowth (Li et al., 2003;Takeda et al., 2003). Moreover, both MOC1 and OsTB1 expression was found in the axillary meristem and tiller buds. MOC1 was also Figure 4 Performance of LRK2 transgenic plants under drought stress. 0 day: Wild-type, M2 and M6 seedlings grown for 12 days under normal conditions. 5, 8 days: Performance of wild-type, M2 and M6 seedlings treated with 20% PEG6000 for 5 and 8 days, respectively. 1, 3, 11 days: Performance of wild-type, AM5 and AM8 seedlings grown under drought stress for 1, 3 and 11 days, respectively. Recovery 20 days: Recovery of the treated wild-type, M2 and M6 seedlings for 20 days. Recovery 8 days: Recovery of the treated wild-type, AM5 and AM8 seedlings for 8 days. detected at the leaf axils. The LRK2 gene studied here was expressed in tiller buds and nodes, suggesting a role in tiller formation and outgrowth. In line with this, LRK1 was previously found to increase both tiller and grain number (Zha et al., 2009). In the current study, overexpression of LRK2 resulted in an increase in tiller number, but no significant change in the number of grains per panicle. Similarly, both LRK1-and LRK2-overexpressing plants exhibited an increase in tiller number. These results suggest that LRK2 and LRK1 are expressed in different tissues, despite belonging to the same gene cluster. However, it is also possible that the diverse phenotypes of the transgenic lines were partly due to the different genetic backgrounds of the rice cultivars used for transformation (i.e. 9311 and Nipponbare).
Plants have evolved a number of morphophysiological and biochemical strategies at both the cellular and molecular levels to allow them to adapt to biotic and abiotic stresses. When plants encounter abiotic stresses, membrane-localized receptors rapidly sense environmental signals and transmit them downstream, thereby activating stress-related responses. During these processes, LRKs play important roles as both sensors and transducers (Lease et al., 1998;Shiu et al., 2004;Torii, 2004). Meanwhile, RPK1 (receptor-like protein kinase 1) is required for embryonic pattern formation and enhances both water and oxidative stress tolerance in Arabidopsis (Mandel et al., 2014;Masle et al., 2005;Meng et al., 2012;Shen et al., 2015;Shpak et al., 2004;Van Zanten et al., 2009). Similarly, BAK1 (BRI1-associated receptor kinase 1) was found to play a role in a number of diverse processes, including brassinosteroid signalling, the phytosulfokine (PSK) signal pathway, light responses, cell death and plant innate immunity (Chinchilla et al., 2007(Chinchilla et al., , 2009Ingram, 2007;Ladwig et al., 2015;Li et al., 2002b;Nam and Li, 2002;Sun et al., 2013). In this study, the novel LRK gene LRK2 was found to function in drought tolerance and tiller development. Identification and further elucidation of the molecular mechanisms of such genes would be valuable in rice production management and genetic improvement studies.
EF1A was first shown to function in protein synthesis, and since then has been implicated in a number of biochemical processes such as interactions with the cytoskeleton (Gross et al., 2005)   proliferation (Sanders et al., 1992). EF1A also interacts with phospho-Akt in breast cancer cells and regulates their proliferation, survival and motility (Pecorari et al., 2009), and is expressed ubiquitously in humans. Although a highly abundant cellular protein associated with the cytoskeleton, the function of EF1A in rice remains unknown. Phylogenetic analysis revealed a close genetic relationship between LRK2 and PSKR1 from A. thaliana. PSKR1 is a PSK receptor that stimulates plant growth and differentiation (Igarashi et al., 2012). In this study, the LRK2 intracellular domain was also found to directly interact with OsEF1A in double molecule fluorescence analysis and a yeast twohybrid assay. The findings further suggest that LRK2 interacts with OsEF1A to regulate plant developmental processes such as cell proliferation, thereby increasing plant branching. Based on the results of this study, a model was proposed to describe the functions of rice LRK2 in regulating tiller size and the drought stress response (Figure 8). By interacting with other molecules such as the eukaryotic translation elongation factor, LRK2 is thought to regulate cellular proliferation, promote branch development and subsequently increase tiller number. The larger root system subsequently contributes to an increase in drought tolerance. As rice is an important food crop and drought one of the main factors affecting crop growth and yield, methods aimed at increasing yield and creating drought-resistant varieties are crucial. The findings of this study demonstrate the potential of LRK2 as a useful tool for crop improvement, particularly with regard to drought tolerance, helping enhance agronomically useful traits such as tiller number, yield and the number of grains per plant. Future research will facilitate further improvements in abiotic stress tolerance in crops through genetic manipulation, thereby paving the way for a new green revolution.

Phylogenetic analysis
Phylogenetic analysis of LRK2 and other LRK proteins was carried out using MEGA 6.0 and a phylogenetic tree constructed using ClustalX 1.83 and UltraEdit21 software.

Subcellular localization of rice LRK2
To investigate the subcellular localization of LRK2, the LRK2 coding sequence was amplified with the cDNA clone as the template, using primers 5 0 -CGGGGTACCATG CAG CCACCTCAT TCTTCATGCA-3 0 and 5 0 -TCCCCCGGGGTCGGA GCCTACACT GTCCAGGCAG-3 0 . The PCR product was used to construct a 35S::LRK2-eGFP fusion plasmid, which was transformed into tobacco leaves by Agrobacterium-mediated transformation, as described previously (Yang et al., 2000). The transformed plants were cultured at 22°C under 16-h light for 24-48 h. The GFP image was subsequently obtained using a Leica TCS SP5 AOBS confocal laser microscope.

Promoter-GUS analysis
The LRK2 promoter, an approximately 2000-bp DNA fragment upstream of the translation start site, was amplified using primers 5 0 -TTGAAGCTTCCTCCCACCTCCAAGTGTTCAAC-3 0 and 5 0 -CCA GGATCCGGTTTTCTGGTGATACTAGCATGGAAG-3 0 from 9311. The DNA fragment was then cloned into the pBI121 expression vector between the HindIII and BamHI sites using the above transformation method. The primers 5 0 -CTGGATCCGTAGATCTG AGGAACCGACGA-3 0 and 5 0 -GAGGACGTCTCACACGTGGTGGT GGTGGT-3 0 were used to confirm the transgenic plants. A GUS assay was performed at various developmental stages as described previously (Jefferson et al., 1987).  Translation of RNA into cDNA was performed using a ReverTra Ace qPCR-RT Kit following the manufacturer's instructions (TOYOBO, Japan). Amplification of genes from the cDNA template was performed using specific primers with high-fidelity primeSTAR HS DNA Polymerase according the user's manual  (TaKaRa, Japan). Diluted reaction products were used as templates for RT-PCR and real-time quantitative PCR analysis. The following LRK2-specific primers were used for RT-PCR: 5 0 -GTCGGTACCATGCAGCCACCTCATTCTTCATGCAAC-3 0 and 5 0 -CAGGTCGACTCAGTCGGAGCCTACACTGTCCAG-3 0 , and the following specific primers for qRT-PCR analysis: 5 0 -TCAGCATC-CAAAAAACAGTTGAAC-3 0 and 5 0 -CTCTGGATCAGAGGTGAAC-GAAC-3 0 . Each data point represents three replicates, and each experiment was repeated twice.

Growth conditions
Rice seeds (O. sativa ssp. japonica cv. Nipponbare) and transgenic plants were sown in pots after 3 day of germination at 37°C. All plants were grown under 28°C/16-h light and 25°C/8-h dark conditions at 75% relative humidity in a glasshouse. For gene expression analysis, the roots of Nipponbare rice seedlings at the three-leaf stage were immersed in PEG6000 (20%) or water for 3, 6, 12 and 24 h. After treatment, the seedlings were harvested and used for total RNA isolation.

Drought stress treatment
For drought stress treatment, 12-day-old 2 9 35S::LRK2, 2 9 35S::antiLRK2 and wild-type plants were grown in pots under the indicated conditions then treated with 20% PEG6000, respectively, until the leaves of the wild type were rolled as a result of drought stress. Plants were then rewatered. The phenotypes of the plants were subsequently observed and photographed at various time points. After recovery, survival rates were calculated by counting plants with green healthy young leaves.

Measurements of root activity
Root activity in terms of TTC reduction was measured as described previously (Hu et al., 2016;Steponkus and Lanphear, 1967). Fresh roots from wild-type and transgenic plants were placed in six test tubes, each containing 2 mg sodium thiosulfate and 5 mL of various concentrations of TTC or distilled water, followed by the addition of 5 mL phosphate buffer (pH 7.5). The roots were incubated at 37°C for 2 h after which 2 mL 1 M sulphuric acid was added to terminate the reaction. The roots were removed and ground in a mortar containing 3 mL ethyl acetate to extract the triphenylformazan (TTCH). TTCH was measured based on the absorbance of the supernatant at 485 nm. A standard curve was constructed with TTCH on the x-axis and OD on the y-axis. The root activity was determined by TTCH concentration for each fresh root as follows: root activity (mg g À1 h À1 ) = TTCH reduction (TTCH mg)/fresh root weight (FW g)/time (h). For each root activity measurement, data points represent the average of three replicates.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: Figure S1 LRK2 promoter expression in transgenic rice. Figure S2 PCR analysis to confirm the transgenic lines. Figure S3 Gene expression levels in rice seedlings in response to drought treatment. Figure S4 Overexpression of LRK2 increased the tiller number in rice.