A pair of orthologs of a leucine-rich repeat receptor kinase-like disease resistance gene family regulates rice response to raised temperature

Background Rice Xa3/Xa26 disease-resistance gene encodes a leucine-rich repeat (LRR) receptor kinase-type protein against Xanthomonas oryzae pv. oryzae (Xoo) and belongs to a multigene family. However, the functions of most genes in this family are unknown. Results Here we report that two orthologs of this family, the NRKe from rice variety Nipponbare and 9RKe from variety 93-11 at the RKe locus, have similar functions although they encode different proteins. This pair of orthologs could not mediate resistance to Xoo, but they were transcriptionally induced by raised temperature. Transcriptional activation of NRKe or 9RKe resulted in the formation of temperature-sensitive lesion mimics, which were spots of dead cells associated with accumulation of superoxides, in different organs of the transgenic plants. These plants were more sensitive to high temperature shock than wild-type controls. Transgenic plants carrying a chimeric protein consisting of the LRR domain of NRKe and the kinase domain of Xa3/Xa26 developed the same lesion mimics as the NRKe-transgenic plants, whereas transgenic plants carrying another chimeric protein consisting of the LRR domain of Xa3/Xa26 and the kinase domain of NRKe were free of lesion mimic. All the transgenic plants carrying a chimeric protein were susceptible to Xoo. Conclusion These results suggest that the RKe locus is involved in rice response to raised temperature. The LRR domain of RKe protein appears to be important to sense increased temperature. The RKe-involved temperature-related pathway and Xa3/Xa26-mediated disease-resistance pathway may partially overlap.


Background
Disease resistance (R) genes that mediate race-specific resistance are important for plants in defending themselves from various pathogen attacks. Currently, many R genes have been characterized and most of them encode proteins sharing common features. The most prevalent domain of characterized R proteins is the leucine-rich repeat (LRR), which is the major determinant of pathogen recognition [1]. A relatively large number of LRRcontaining R proteins belong to the nucleotide-binding site-LRR class [2]; most of the characterized R genes against Magnaporthe oryzae, which causes rice fungal blast, belong to this class [3]. The second class is the LRR receptor-like R proteins that consist of an extracellular LRR domain and a transmembrane motif [4]; no rice R gene of this class has been characterized so far. The last class includes the LRR receptor kinase-like R proteins, which are only identified in rice so far [5,6]. Both LRR and kinase domain are common structures in many signaling pathways. Plants have a large number of LRR receptor kinase-like proteins [7]. All the LRR receptor kinase-like proteins contain a transmembrane motif, which departs the LRR domain to stay outside of the plasma [7]. Thus, as cell surface receptors, the LRR receptor kinase-like proteins are thought to recognize special extracellular ligands via the LRR domain and initiate the downstream signaling through the intracellular kinase domain.
Rice R gene Xa3/Xa26 confers race-specific resistance to Xanthomonas oryzae pv. oryzae (Xoo) that causes bacterial blight, one of the most devastating diseases of rice worldwide and it encodes a LRR receptor kinase-like protein [6]. Xa3/Xa26 belongs to a tandem clustered multiple gene family in the long arm of rice chromosome 11 [8,9]. Point mutations with positive selection were a major force of the evolution of this family [9]. Both paralogs (in the same rice variety) and orthologs (in different rice varieties) of this family have similar tissue-specific expression patterns, suggesting that they may have similar functions [10,11]. In addition, at least some paralogs of this family have dosage effects, in which their disease resistance functions are associated with their transcript amounts [12,13]. The expression of Xa3/Xa26 is developmentally regulated. It has a lower level of expression at the early developmental stage, but a higher level of expression at the late developmental stage, which results in rice plants carrying Xa3/Xa26 being susceptible to some Xoo strains at the seedling stage but showing enhanced resistance to the same Xoo strains at the adult stages [12]. Rice plants constitutively overexpressing Xa3/Xa26 have an enlarged resistance spectrum, an increased level of resistance, and a whole growth-stage resistance without influencing their morphologies and agronomic performance [12,14]. In addition, the ortholog alleles at Xa3/Xa26 locus confer a durable resistance to Xoo [15]. Another paralog of the Xa3/Xa26 family, the MRKa, could not mediate resistance to Xoo when regulated by its native promoter, but MRKa could confer partial resistance to Xoo when regulated by a strong constitutive promoter [13].
To ascertain the functions of other paralogs of the Xa3/ Xa26 family, we constitutively expressed two orthologs of this family, NRKe and 9RKe, from two rice varieties, at the RKe locus. This pair of orthologs cannot mediate resistance to Xoo as do their paralogs Xa3/Xa26 and MRKa in the present experimental condition. However, they are involved in rice response to raised temperature.

Plasmid construction and rice transformation
To overexpress 9RKe and NRKe in rice, these two genes were amplified using polymerase chain reaction (PCR) primers NRKe-F and NRKe-R (Additional file 1, Table  S1) and genomic DNA from indica rice variety 93-11 (Oryza sativa L. ssp. indica) and bacterial artificial clone (BAC) OSJNBa0004O15 from japonica variety Nipponbare (O. sativa L. ssp. japonica) as template, respectively. The PCR products were digested with BamHI and ligated into the transformation vector pU1301, which contained a maize ubiquitin gene promoter to drive the inserted gene [12].
To study the function of different domains of NRKe, two mutations of NRKe were generated. The kinase domain-deleted NRKe, the NRKe-ΔK, was created by PCR amplification using primers NRKe-F and B+E-3 and the LRR-transmembrane motif-deleted NRKe, the NRKe-K, was created by PCR amplification using primers RKe-6F and NRKe-R (Additional file 1, Table S1). Both truncated genes were amplified using Nipponbare BAC OSJNBa0004O15 as template. The PCR products then were ligated into pU1301 under the regulation of maize ubiquitin gene promoter. Two chimeric genes, Be1 and Be2, which consist of different fragments of NRKe and Xa3/Xa26, were amplified by overlap extension PCR [16]. PCR primers MKb-F, B+E-1, B+E-2, and NRKe-R were used for construction of Be1 and primers NRKe-F, B+E-3, B+E-4, and MKb-R were used for construction of Be2 (Additional file 1, Table S1). Both chimeric genes were amplified using BAC 3H8, which harbors Xa3/Xa26, from rice variety Minghui 63 (O. sativa L. ssp. indica) and Nipponbare BAC OSJNBa0004O15 as templates. The PCR products were ligated into pU1301 under the regulation of maize ubiquitin gene promoter.
All the constructs were respectively transferred into Agrobacterium tumefaciens strain EHA105 by electroporation. Rice transformation was performed by the Agrobacterium-mediated method using calli generated from mature embryos of rice variety Mudanjiang 8 (O. sativa L. ssp. japonica) [17].

Pathogen inoculation
Plants were inoculated with Philippine Xoo strains PXO61, PXO86, PXO71, and PXO99 by the leaf-clipping method at the booting (panicle development) stage [18]. Disease was scored by measuring the percentage disease area (lesion length/leaf length) at 2 weeks after inoculation.

Temperature treatment
Rice plants grown in a greenhouse at approximately 24°C were transferred to growth chambers for temperature experiments at the tillering stage. The growth chamber conditions were controlled as 70% humidity, 14-h light, and 10-h dark at either 24°C or 35°C. For heat shock, transgenic and wild-type plants were grown in the same pots and transferred to a 42°C growth chamber for 32 h (until almost all the leaves of one group in the pot became completely rolled and some leaves died). The plants were then recovered by maintaining them at room temperature for 3 d and phenotypes were recorded.

Gene expression analyses
RNA gel blot analysis was performed as described previously [19]. A 343-bp probe for both NRKe and 9RKe, digested using BamHI and SacI from the PCR product amplified using primers NRKe-F and NRKe-R (Additional file 1, Table S1), was used for hybridization. Quantitative reverse transcription-PCR (qRT-PCR) was performed using gene-specific primers (Additional file 1, Table S1), as described previously [20]. The expression of actin gene was first used to standardize the RNA sample for each qRT-PCR. For each gene, qRT-PCR assays were repeated at least twice, with each repetition having three replicates. Only the results in one repetition were presented when similar results were obtained in repeated experiments. The level of expression relative to control was presented.

Overexpression of RKe resulted in the formation of lesion mimics in rice plants
The paralog RKe of the R gene Xa3/Xa26 family only exists in some rice varieties [9]. The RKe genes in rice varieties 93-11 and Nipponbare, named 9RKe [GenBank accession number: JN176871] and NRKe [JN176870], respectively, had 99.6% sequence identity. Both genes putatively encode proteins consisting of 1097 amino acids but with a five-residue difference (Figure 1a).
At least some paralogs of the Xa3/Xa26 family have dosage effects in rice-bacterium interactions [12,13]. To ascertain whether NRKe and 9RKe also functioned in rice disease resistance, we constitutively overexpressed them in susceptible rice variety Mudanjiang 8, in which RKe was not detected. Twenty-two and 17 independent transformants transformed with NRKe and 9RKe and named NRKe-oe and 9RKe-oe were obtained, respectively. Some of the T 0 plants showed overexpression of NRKe or 9RKe (Additional file 1, Figure S1). All the T 0 plants were inoculated with Xoo strain PXO61 at the booting stage and all the plants were susceptible as the wild-type Mudanjiang 8. After all the T 0 plants were cut at about 12 cm above ground level after seeds had been harvested, the regenerated plants from the stubs were inoculated with other Xoo strains PXO86, PXO71, or PXO99. The transgenic plants were susceptible to these Xoo strains as the wild-type plants. Two independent T 1 families generated from two T 0 plants that overexpressed NRKe were inoculated with Xoo strain PXO61 for further analysis, but both families showed no difference from wild-type Mudanjiang 8. These results suggest that NRKe and 9RKe may not have roles in rice-Xoo interaction.
Interestingly, 11 of the 22 NRKe-oe T 0 plants and 12 of the 17 9RKe-oe T 0 plants developed similar brown lesion mimics (i.e., spots that resemble infection) on leaf blades, leaf sheaths, and peduncles of panicles at the adult stage ( Figure 1b). The lesion mimics could appear at any growth stage as small ones and developed to large ones during plant growth, but were restricted to an area around the initiation point. The lesion also could be formed spontaneously when the transgenic plants were grown in a sterilized tube or flask (Additional file 1, Figure S2). These results suggest that the formation of lesions is not related to pathogen infection.
Staining the leaves of the transgenic plants with visible lesion mimics using trypan blue, an indicator of dead cells [21], showed dark blue-stained spots at the sites around lesion mimics; such dark blue spots were not observed in the leaves of wild-type plants (Figure 1c). Autofluorescence, an indicator of phenolic compounds produced during cell death of hypersensitive response [22], was also detected at the sites of lesion mimics in the leaves of transgenic plants under the microscope after ultraviolet stimulation; no autofluorescence was detected in the leaves of wild-type plants (Figure 1c). The leaves with visible lesion mimics were also stained with 3,3-diaminobenzidin and nitro blue tetrazolium, the indicators of H 2 O 2 and superoxides, respectively [23]. The leaves of the transgenic plants showed markedly increased dark staining at the sites of lesion mimics as compared to the leaves of wildtype plants after both 3,3-diaminobenzidin and nitro blue tetrazolium staining, indicating the accumulation of H 2 O 2 and superoxides in the transgenic plants ( Figure 1c). These results suggest that the lesions of the NRKe-oe and 9RKe-oe plants were related to cell death, which may be associated with the accumulation of superoxides including H 2 O 2 . Because the NRKe-oe and 9RKe-oe plants developed similar lesion mimics, only the NRKe-oe plants were used for some of the further analyses.
To determine whether the development of lesion mimics was due to overexpressed NRKe, two T 1 families (NRKe-oe6 and NRKe-oe15) derived from two of the T 0 plants that developed lesion mimics were further examined for their phenotypes and the existence of NRKe-oe construct. The development of lesion mimics on the leaves cosegregated with the existence of NRKe in the two families ( Figure 1d). These results suggest that overexpressing NRKe caused the formation of spontaneous lesion mimics in rice.

Formation of lesion mimics in NRKe-oe plants was temperature sensitive
Lesion mimic formation is a complex physiologic reaction in plants and may be caused by many biotic or abiotic factors. Because the NRKe-oe plants appeared not to be involved in rice-Xoo interaction, the formation of lesion mimics in these plants may be influenced by other environmental factors. Actually, two batches of NRKe-oe plants were generated in our experiments. The first batch of transgenic plants was planted in the field in early summer and marked lesion mimics were observed on their leaves when the plants grew to the booting stage during the summer. The second batch of transgenic plants was generated about 50 d later and was planted in the field in late summer. When the second batch of plants grew to the booting stage in the autumn, no obvious lesion mimic was observed in these plants. But after the second batch of plants was transferred to a greenhouse that had a higher temperature than the field condition, marked lesion mimics were formed on their leaves. These observations led us to infer that temperature may influence the formation of lesion mimics.  grown at 35°C, the NRKe-oe and 9RKe-oe plants formed marked lesion mimics on their leaves at 10 and 18 d after being at 35°C, respectively, whereas the control Mudanjiang 8, Nipponbare, and 93-11 were free of lesions (Figure 2a). When the 24°C-pretreated NRKe-oe and control plants were transferred to 35°C, the transgenic plants but not control plants developed lesion mimics on their leaves even at 3 d after growing in 35°C (Figure 2b). These results suggest that higher temperature may induce the formation of lesion mimics in NRKe-oe and 9RKe-oe plants.

Temperature influenced the expression of NRKe and 9RKe
To ascertain whether temperature transcriptionally influenced NRKe and 9RKe, we comparatively analyzed the expression of NRKe and 9RKe in wild-type plants at different temperatures. At 35°C, the expression of NRKe and 9RKe was markedly induced at 12 h after temperature treatment and then returned to the basal level of expression at 1 d after temperature treatment; the transcripts increased approximately 2-to 2.5-fold at 12 h after 35°C treatment compared to untreated control plants (Figure 3a). Raising the temperature to 42°C further increased the expression of NRKe and 9RKe; the transcripts of NRKe and 9RKe increased approximately 30-and 7-fold at 12 h after treatment, respectively. However, NRKe and 9RKe expression was only slightly induced when the plants were kept at 24°C.
The promoter regions (approximately 2 kb upstream of translation start codon ATG) of NRKe and 9RKe had 98% sequence identity. The two sequences were analyzed for putative cis-acting elements involved in heat stress response by searching different databases. The most canonical heat shock element (HSE) consists of a repeated core sequence of nGAAn ('n' indicating any nucleotide) in alternate orientation [24]. In the promoter regions of NRKe and 9RKe, two imperfect nGAAn repeats were identified. The first nGAAn repeat was 'acTCaaTTCagGAAt' for NKRe and 'acTCaaTTCagGAta' for 9RKe, and the second was 'aGAAtgGAgaacTCca-tAAatcTCa' for both genes (Figure 3b; Additional file 1, Figure S3). The promoter regions of the two genes also harbored another putative HSE, which was complementary to the HSE 'AAAAAATTTC' of Brassica oleracea  Figure S3). Four CCAAT boxes were also identified in the promoter regions (Figure 3b; Additional file 1, Figure S3). The CCAAT box has been reported to act cooperatively with HSEs to increase promoter activity [25]. The occurrence of putative heat-responsive cis-elements and heat-induced expression suggest that NRKe and 9RKe may function in early response to increased temperature.

Activation of NRKe and 9RKe expression influenced rice response to increased temperature
To examine the inference that NRKe and 9RKe are involved in rice response to increased temperature, NRKeoe or 9RKe-oe plants and their wild-type plants were treated at 42°C. After treatment, the NRKe-oe and 9RKe-oe plants appeared to be more sensitive to the heat shock than their corresponding wild-type plants (Figure 4). The transgenic plants showed lower survival rates compared to wild-type plants. These results further suggest that RKe orthologs influence rice response to temperature.
The kinase domain of Xa3/Xa26 could replace the kinase domain of NRKe for promoting lesion mimics NRKe putatively encodes a receptor kinase-type protein consisting of two domains, a LRR domain and a kinase domain, which are connected by a transmembrane motif. Usually, the two domains function together in a complicated way in signal transduction. One exception is rice XA21D, which is a truncated form of the LRR receptor kinase XA21 conferring resistance to Xoo. XA21D has only the LRR domain of XA21, but functions similarly as XA21 in rice-Xoo interactions [26]. To analyze the relationship of different domains of NRKe, two truncated NRKe fragments were, respectively, overexpressed in rice variety Mudanjiang 8. One truncated NRKe was NRKe-ΔK, which encoded protein did not have the kinase domain of NRKe; another truncated NRKe was NRKe-K, which only encoded the kinase domain of NRKe ( Figure  5a). Eighteen and 18 independent transformants transformed with NRKe-ΔK and NRKe-K, and named NRKe-ΔK-oe and NRKe-K-oe, were obtained, respectively. In the same environment, NRKe-oe plants developed marked lesion mimics on their leaves, but NRKe-ΔK-oe and NRKe-K-oe plants were free of lesions (Figure 5b). These results suggest that a complete NRKe is important for the development of lesion mimics.
NRKe is the paralog of the R gene Xa3/Xa26 that confers race-specific resistance against Xoo in rice [6]. The predicted LRR and kinase domains of NRKe and Xa3/ Xa26 proteins share 68% and 85% sequence identity, respectively (Additional file 1, Figure S4). To determine whether the LRR or kinase domain was critical for the development of lesion mimics, we constructed two chimeric genes, the Be1 and Be2, using the different fragments of NRKe and Xa3/Xa26, and constitutively overexpressed them in rice variety Mudanjiang 8. The Be1 consists of the sequences encoding the region in front of LRR, LRR, juxtamembrane region, and transmembrane region of Xa3/Xa26 and the sequence encoding the kinase domain of NRKe; the Be2 consists of the sequence encoding the region in front of LRR, LRR, juxtamembrane region, and transmembrane region of NRKe and the sequence encoding the kinase domain of Xa3/Xa26 (Figure 6a). Eleven and 13 independent transformants transformed with Be1 and Be2, and named Be1-oe and Be2-oe, were obtained, respectively. All the Be1-oe plants were free of lesions as the wild-type control plants, but 8 of the 13 Be2-oe plants developed lesion mimics on their leaves at the booting stage (Figure 6b). To determine whether the development of lesion mimics was due to overexpressed Be2, two T 1 families (Be2-oe2 and Be2-oe9) derived from two of the T 0 plants that developed lesion mimics were further examined for their phenotypes and the existence of Be2 construct. The development of lesion mimics cosegregated with existence of Be2 in the two families ( Figure  6c). These results suggest that the kinase domain of Xa3/ Xa26 can replace the kinase domain of NRKe to induce formation of lesion mimics.
Because Xa3/Xa26 mediates rice resistance against Xoo, the T 0 Be1-oe and Be2-oe plants were analyzed for their response to Xoo. Most of the T 0 Be1-oe and Be2-oe plants were susceptible to Xoo strain PXO61 as the wild type. Two Be1-oe T 1 families and two Be2-oe T 1 families were further inoculated with PXO61. No increased resistance or susceptibility was associated with the existence of Be1-oe or Be2-oe construct in these T 1 families (Additional file 1, Figure S5). These results suggest that the kinase domain of NRKe cannot replace the functions of the Xa3/Xa26 kinase domain in disease resistance.  Figure 4 Activation of NRKe or 9RKe caused rice to be more sensitive to increased temperature. NRKe-oe and 9RKe-oe were T 2 plants at 4-to 5-leaf stage. WT, wild-type Mudanjiang 8. Bars represent mean (2 technical replicates with each replicate contained 16-20 plants) ± standard deviation.

Discussion
The LRR receptor kinase-like plasma membrane proteins are the largest class of receptor-like kinases in plants and this class of proteins are well-known components in signal transduction during different physiologic activities [27]. For example, the CLAVATA1 regulates stem cell maintenance and differentiation in Arabidopsis [28]. BRI1 participates in the regulation of Arabidopsis growth and development via brassinosteroid signaling [29]. Rice Xa3/Xa26 and XA21 mediate race-specific and major disease resistance to Xoo [5,6]. The RPK1 functions in the abscisic acid signaling pathway related to germination, growth, and stomatal closure in Arabidopsis [30]. ERECTA specifies the size and shape of mature organs in Arabidopsis [31]. A LRR receptor kinase-like gene may be associated with the resistance to Heterodera glycines in soybean and the resistance to H. glycines is temperature sensitive [32,33]. The results presented here add another example that this type of protein, rice NRKe and 9RKe, appears to be involved in rice response to increased temperature. This inference is supported by the following evidences. First, both NRKe and 9RKe were transcriptionally induced by higher temperature compared to control. Second, activation of this pair of orthologs resulted in rice more sensitive to high temperature.
The LRR domain of RKe is critical to sense increased temperature As receptor, the extracellular localized LRR domain of LRR receptor kinase-like protein functions to recognize and bind ligand, which includes proteins, peptides, or nonprotein components, in signal transduction [7]. For example, the LRR domain of BRI1 mediates brassinosteroid signaling by interacting with brassinolide, the active form of brassinosteroids [34]. Arabidopsis FLS2-mediated immunity is initiated by binding a bacterial flagellinderived peptide Flg22 to its LRR domain [35]. Rice Xa21regulated race-specific disease resistance is triggered by binding a sulfated peptide secreted by Xoo to its LRR domain [36]. Constitutive expression of the orthologs in the RKe locus resulted in the formation of lesion mimics, which was related to increased temperature ( Figure 2). Thus, the appearance of lesion mimics in the NRKeor 9RKe-transgenic plants is a marker of rice response to raised temperature through NRKeor 9RKe-initiated signaling, although the appearance of lesion mimics did not  which suggests that the ligand bound to the LRR domains of NRKe and 9RKe to induce the formation of temperature-sensitive lesion mimics was not from the changed environment. Thus, we argue that the ligand may be originated in rice. When the temperature rises, the ligand appears or its concentration is increased and RKe interacts with the ligand to trigger downstream reaction. However, further study is required to examine this hypothesis. The above inference is also supported by the fact that NRKe and Xa3/Xa26 belong to a tandem clustered R gene family; some amino acid sites of LRR domains encoded by the genes in this family are subject to positive selection, whereas the kinase domains encoded by these genes are evolutionarily conserved, suggesting the functional constraint of the kinase domains in this family [9]. However, not all the paralogs in this family possess the ability to confer disease resistance, although they have a similar tissue-specific expression pattern [11,13]. The present results suggest that some paralogs of the R gene family may be involved in other biological processes, like NRKe and 9RKe.

Conclusion
The tandem repeated paralogs of a haplotype and their orthologs in different haplotypes of R gene family provide different resistance specificities and a sequence reservoir for evolutionary forces to rapidly generate new R genes [37]. A previous report revealed that one paralog of tomato R gene Pto family, the Fen, confers sensitivity to fenthion, an organophosphorous insecticide [38]. Our results further suggest that the "defeated" R genes, like NRKe and 9RKe, can also be involved in responses to raised temperature in rice, in addition to serving as the structural reservoir for creating new genes.

Additional material
Additional file 1: Supplemental table and figures. Table S1: PCR primers used for plasmid construction and gene expression analysis. Figure  S1: Expression of NRKe and 9RKe in transgenic plants (T 0 generation) analyzed by RNA gel blot. NRKe-oe, NRKe-overexpressing plants; 9KRe-oe, 9RKe-overexpressing plants; WT, wild-type Mudanjiang 8. Figure S2: NRKe-overexpressing and Be2-overexpressing plants formed lesion mimics spontaneously in sterilized container during tissue culture. Figure S3: Alignment of promoter regions of NRKe and 9RKe. The nucleotides immediately upstream of the translation start codon ATG are numbered as "-1". The putative heat-responsive cis-elements are underlined. HSE (nGAAn core), heat shock element consisting of repeated core nGAAn in alternate orientation; HSE (PlantCARE), heat shock element found in PlantCARE database; CCAAT box (Place), CCAAT box found in PLACE database. Figure S4 Figure S5: Overexpression of Be1 or Be2 could not influence rice response to Xoo strain PXO61. Positive transgenic plants were determined by PCR amplification of Be1 or Be2 using gene-specific primers (Additional file 1, Table S1). Wild type (WT) is Mudanjiang 8. Rb49 is a transgenic line carrying Xa3/Xa26 driven by its native promoter in Mudanjiang 8 background.