A Koshihikari X Oryza rufipogon Introgression Line with a High Capacity to Take up Nitrogen to Maintain Growth and Panicle Development under Low Nitrogen Conditions

Abstract

Nitrogen (N) is an important macronutrient for plant growth and development. Currently, N fertilizers are required for the efficient production of modern crops such as rice due to their limited capacity to take up N when present at low concentrations. Wild rice represents a useful genetic resource for improving crop responses to low nutrient stress. Here, we describe the isolation and characterization of an introgression line, KRIL37, that carries a small region of the Oryza rufipogon genome in the Oryza sativa L. cv Koshihikari (KH) background. This line was found to grow better under low N conditions and have similar or lower C/N ratios in aerial portions compared to those in the parental KH cultivar, suggesting that KRIL37 has a higher capacity to take up and assimilate N when present at low concentrations. KRIL37 performance in the field was also better than that of KH cultivated without N and fertilizer (−F). Transcriptome analyses of 3-week-old seedlings based on RNA-sequencing revealed that KH induced a wider suite of genes than the tolerant line KRIL37 in response to low N conditions. Some ammonium transporters and N assimilation genes were found to be induced under low N in KRIL37, but not in KH. Our findings suggest that the superior growth performance of KRIL37 under limited N conditions could be due to the expression of wild alleles influencing N uptake and assimilation. Our study demonstrates the potential to use wild rice genomes to improve modern crops for low nutrient tolerance.

Introduction

As the major limiting factor in most agricultural systems, nitrogen (N) is an essential component of several compounds, such as proteins, nucleic acids and some phytohormones. Ammonium (NH₄⁺) and nitrate (NO₃⁻) are the two main forms of inorganic N available for plant growth (Chen et al. 1998). Over the past decades, the need for an increase in food production has been associated with the excessive use of chemical fertilizers in cropping systems. Despite the ecological unsustainability of heavy N fertilization in agriculture due to the loss of applied N to the ecosystem, this practice continues unabated (Boyle 2017).

A reduction in the use of N fertilizers can be achieved by developing sustainable management practices and crop cultivars with improved nutrient uptake and N use efficiency to minimize the impact on the global environment (Boyle 2017). Nitrogen use efficiency (NUE, relates to shoot biomass or grain produced per the nitrogen absorbed or the nitrogen which was available to the plant) which often increases under lower N supply is usually partitioned into components, such as the absorption of nitrogen by the roots (N uptake efficiency, NUpE), as well as the biomass/grains produced per N absorbed into the shoot (N utilization efficiency, NUtE) (Moll et al. 1982, Xu et al. 2012). However, the breeding of varieties with improved nutrient use efficiency and accumulation and with high yield potential depends on large genetic variation for root uptake, root-to-shoot transport, mineral remobilization and grain accumulation, and this variation is lacking in cultivated types due to intensive breeding and selection. Therefore, the genetic diversity present in wild rice could be a source of natural variations that could be tapped to accelerate crop improvement (Ahn et al. 2002, Tester and Langridge 2010). Several introgression lines (ILs) have been developed using wild rice genome species as donors (Ahn et al. 2002, Tian et al. 2006). Studies showed that extensive genomic changes in the genome of rice-Zizania through introgressive hybridization resulted in superior phenotypes (Wang et al. 2013). Also, using O. rufipogon as donor in the background of japonica rice cultivars Koshihikari and Itadaki resulted in partial resistance to rice blast (Hirabayashi et al. 2010).

Many genes involved in the tolerance of plants to low N have been identified. The uptake of N as NO₃⁻ or NH₄⁺ is accomplished by nitrate transporters (NRTs) and ammonium transporters (AMTs), respectively (Masclaux-Daubresse et al. 2010). In rice paddy fields, ammonium is the preferred N source; it is incorporated into amino acids through the glutamine synthetase (GS) and glutamate synthase (GOGAT) N assimilation cycle (Crawford 1995, Lam et al. 1996). NRT1 or NPF (Nitrate Transporter 1/Peptide Transporter) and NRT2 (Nitrate Transporter 2) and AMTs are the most prominent gene families responsible, respectively, for nitrate and ammonium uptake and transport in rice. In general, NRT1 and OsAMT2-4 are low-affinity transporters while NRT2 and OsAMT1 sub-family members are high-affinity transporters (Sonoda et al. 2003, Li et al. 2009). The overexpression of OsNRT2.1 and co-overexpression of OsNRT2.3a and OsNAR2.1 improved nitrate uptake and yield in rice under low NO₃⁻ conditions. Moreover, grain weight and grain N content increased by 26% and 15%, respectively, in OsNRT2.1 overexpressing plants compared to those in WT rice plants under alternating wet and dry conditions (Luo et al. 2018, Chen et al. 2020). NRT1.1B containing Oryza indica allele was associated with enhanced nitrate uptake and root-to-shoot transport and upregulated expression of nitrate-responsive genes resulting in improved grain yield and NUE of japonica variety in field tests (Hu et al. 2015). Similarly, while the expression of OsAMT1.1 and OsAMT1.2 is promoted by the presence of NH₄⁺ in rice roots and plants, overexpression of OsAMT1.1 increased NH₄⁺ uptake, plant growth, grain filling and total number of grains per plant, specifically under low-NH₄⁺ conditions (Ranathunge et al. 2014). These transporters are also regulated by transcription factors (TFs). Over 40 TFs, including WRKY proteins NIN LIKE PROTEIN (NLPs) DNA binding with one finger protein (Dof) and Growth-regulating factor (GRF) are involved in the transcriptional regulation of N transport and assimilation in plants (Vidal et al. 2020). In rice, OsNLP1 positively regulates the expression of several nitrate uptake and assimilation genes (OsNRT1.1a and OsNRT1.1b), AMTs (OsAMT1.1) and NH₄⁺ assimilation genes (OsGS1.1, OsGS2 and OsFd-GOGAT) (Alfatih et al. 2020). Similarly, OsNLP4 affects nitrate reductase (NR) activity in NO₃⁻ assimilation in rice by influencing Fe and Mo accumulation necessary for NR activity (Wang et al. 2021).

Transcriptome analysis through RNA sequencing (RNA-seq) is a high-throughput gene expression profiling and an effective method for studying regulatory networks in crops due to its precise measurement of transcript levels (Wang et al. 2009, Varala et al. 2018). Through transcriptome analysis, several studies have identified the regulators of molecular and physiological responses in crops as well as a large number of differentially expressed genes (DEGs) in response to N levels (Gelli et al. 2014, Sinha et al. 2018). Using lowland (IR64) and upland (Nagina22) rice varieties under different N conditions, transcriptome analysis showed that most DEGs were associated with starch and chloroplast metabolism and signal transduction (Sinha et al. 2018). In sorghum, N-sensitive genotypes increased the abundance of DEG transcripts associated with stress responses including oxidative stress and stimuli, whereas N-tolerant genotypes accumulated transcripts related to high-affinity nitrate transporters (NRT2.2, NRT2.3, NRT2.5 and NRT2.6) and lysine histidine transporter 1 (LHT1), suggesting improved uptake efficiency of inorganic and organic forms of N (Gelli et al. 2014). The calcium-mediated plant–pathogen interaction, mitogen activated protein kinase (MAPK) signaling and phosphatidylinositol signaling pathways were enriched in wheat under low N, indicating that these might be adaptation-related pathways involved in the response to low-N stress (Yan et al. 2021).

In this study, a rice IL, KRIL37 and its parental cultivar Koshihikari (KH) were exposed to low N (LN) and normal N (NN) conditions, and the phenotypic responses of plants as well as their changes at the transcriptome level were determined. The DEGs identified revealed that some NH₄⁺ transporters and N assimilation genes were induced in the IL KRIL37 in response to low N. In contrast, stress response self-degradative pathways, such as autophagy, phagosome formation and diterpenoid biosynthesis pathways, were enriched in the sensitive line KH.

Results

Screening and selection of ILs

The initial selection consisted of the screening of rice ILs developed from parental cultivars KH and genome fragments from wild rice O. rufipogon, respectively, under non-fertilized (−F) paddy field conditions in Fukushima. Using the average panicle weight, two ILs KRIL8 and KRIL37 with high panicle weight (Supplementary Fig. S1) were selected for further screening in Yayoi paddy field (a rice paddy field in the Yayoi campus of the University of Tokyo) with KH in 2017. KRIL37 showed significantly high panicle weight under −F and no N (−N) in Yayoi (Fig. 1) as in Fukushima. Based on the result of this initial screening, the IL KRIL37 and its parental cultivar KH were selected for hydroponic, pot and field N stress experiments and analyzed by RNA-seq.

Fig. 1

Average panicle and straw weight obtained from KRIL37 and KH cultivated at Yayoi paddy field in 2017. (A) Panicle weight. (B) Straw weight. −F, −N and +N indicate no fertilization, no N and fertilizer supply, respectively. n = 10, and asterisks indicate significant differences (t-test, *P < 0.05, **P < 0.01, ns = nonsignificant).

LN stress tolerance experiments

We then tested the physiological characteristics of the selected lines in more detail. Both lines were grown hydroponically under different N conditions, and their growth and C/N ratio were measured; KH and KRIL37 showed different phenotypic responses (Fig. 2A–F) to the different N treatments (LN: 0.4 mM NH₄⁺; NN: 1.6 mM NH₄⁺). At the early growth stage, a reduction in N levels (i.e. the LN treatment) led to an increase in root length, whereas the shoot length remained apparently insensitive to this stimulus in 3-week-old rice seedlings. Although both the root and shoot were generally longer in KRIL37 than in KH plants, only the roots showed over 40% increase in length compared with that of KH under LN conditions (Fig. 2A). In addition, the root and shoot dry weight of KRIL37 was nearly double that of KH after 3 weeks of growth in LN solution (Fig. 2C, D). Similarly, KRIL37 had a lower shoot C/N ratio in comparison to that of KH at 0.4 mM NH₄⁺ (Fig. 2F), suggesting that KRIL37 plants might have a relatively higher N uptake capacity. Interestingly, the variation in biomass (dry weight) highly correlated with the NUE assessed among the two lines. NUpE representing the capture of N from the nutrient solution was 90% higher in KRIL37 compared to KH (Supplementary Fig. S2A), suggesting that N uptake is enhanced in KRIL37. We also calculated NUE as shoot dry biomass/available N. NUE was higher under LN in KRIL37 than KH, but this was not the case under NN (Supplementary Fig. S2D).

Fig. 2

Morphological and physiological analyses of KRIL37 and KH plants cultivated under LN (0.4 mM NH₄⁺) and NN (1.6 mM NH₄⁺) during the seedling stage. The root length (A), shoot length (B), root dry weight (C), shoot dry weight (D), root C/N (E) and shoot C/N ratio (F). n = 5–6, error bars represent the standard deviation and asterisks indicate significant differences (t-test, *P < 0.05, **P < 0.01, ***P < 0.001, ns = nonsignificant). (G) Plants grown hydroponically in modified Kimura’s B solution for 21 d with 0.4 mM NH₄⁺. Scale bars equal 5 cm for roots and 10 cm for the shoot.

Evaluation in pots and paddy fields was also done to assess yield-related traits of both the parental cultivated line KH and the IL KRIL37. As observed for plants in hydroponic cultures, panicle and straw dry weight were significantly higher in KRIL37 than in KH plants grown in pots (Fig. 3A, B) and paddy fields (Fig. 3C–H) under nutrient stress conditions. When grown in pots under LN conditions, KRIL37 produced average panicle and straw weights of 4.8 and 9.04 g/plant, respectively. These results were over 200% higher than those observed in KH plants grown under the same conditions, with average panicle and straw weights of 0.94 and 3.67 g/plant, respectively. A similar trend was observed in plants cultivated under NN conditions in pots.

Fig. 3

Average panicle and straw weight of KH and KRIL37 rice plants grown in pot and field conditions. (A, B) Pot cultivation (n = 6). Paddy field cultivation in Yayoi in 2018 (n = 8) (C, D), paddy field in Shiga in 2018 (n = 24) (E, F) and paddy field in Shiga, 2020 (n = 66–72) (G, H). (I) Straw of KH and KRIL37 following pot cultivation. (LN, low N soil; NN, normal N soil; +N, −N, −F and +F indicate N supply, no N supplied, no fertilization and normal fertilization). Scale bars equal 10 cm. Asterisks indicate significant differences (t-test, *P < 0.05, **P < 0.01, ***P < 0.001, ns = nonsignificant).

Significant higher panicle weight values were consistently obtained from KRIL37 plants compared with those from KH plants cultivated in Yayoi and Shiga fields. However, N and/or nutrient stress caused a reduction in panicle weight in both genotypes; this reduction was more prominent in KH plants relative to KRIL37 plants. At Yayoi paddy field, KRIL37 gave significantly high panicle and straw weight (Fig. 3C, D) under −N and N supply (+N) while at Shiga paddy field, an increase of 42% was observed in KRIL37 panicle yield when compared to the average panicle yield of 13.5 g/plant for KH under −F field conditions in 2018 (Fig. 3E, F). Similar results were obtained from experiments run in 2020 (Fig. 3G, H) and 2021 (Supplementary Fig. S3A) under −F and fertilized (+F) fields in Shiga. We acknowledge that the weight of brown rice is a common measure of rice yield. To examine the relationship between panicle weight and brown rice in our study, we determined the brown rice weight of 2,021 samples. Assessment of brown rice yield (Supplementary Fig. S3B) revealed an average of 176.9 and 339.2 g/m² for KH and KRIL37, respectively, under −F field in Shiga in 2021, indicating the superior performance of the IL under field conditions. This also suggests that our panicle weight data correlate with yield very well.

Transcriptome analysis

To understand gene expression differences between the rice lines under study, shoot and root tissues from KH and KRIL37 grown under LN and NN conditions were subjected to RNA-seq. All treatments had three biological replicates. About 118.9 and 141.9 million raw reads were obtained for KH and KRIL37, respectively. More than 95% of the reads had Phred quality scores at the Q30 level. The high-quality reads were mapped to the reference Nipponbare genome using HISAT2. On average, 91,180 transcripts were selected for biologically meaningful comparisons to identify DEGs in root and shoot tissues of the lines grown under hydroponic conditions. Using normalized counts, a high correlation was observed among replicates of KH and KRIL37 under both LN and NN conditions in various tissues (Supplementary Fig. S4). This indicates a high level of reproducibility among replicates of the same genotype and condition as well as a high level of variability between varieties and N conditions. To determine the effect of N treatments on each genotype, the samples grown under LN and NN conditions were analyzed for DEGs in each tissue (Fig. 4A, B), including the comparisons KRIL37 LN vs. KRIL37 NN and KH LN vs. KH NN [|log2 fold change (FC)| > 1, false detection ratio (FDR) < 0.05]. The number of DEGs in the root (3,824) was twice that in the shoot (1,952) across the two genotypes. Consistently across the two genotypes, DEGs were generally upregulated in roots (2,656 up- vs. 1,826 downregulated) and downregulated in the shoot (1,061 up- vs. 1,454 downregulated). In the root, the number of DEGs between KRIL37 treatments was three times lower (1,065) than that observed among KH treatments (3,617). Interestingly, 858 DEGs in the roots were shared by the two genotypes, including those encoding nitrate transporter (OsNRT1.2), peptide transporter (OsNPF5.5), phosphate (Pi) transporter (OsPHO1;2) and other TFs. Given that KH and KRIL37 have a highly similar genetic background (that of KH), such similarities may have led to similar transcriptome responses to N enrichment, visualized as abundant DEG overlapping (Fig. 4A). Similarly, the number of DEGs among KH treatments (1,572) was significantly higher than that between KRIL37 treatments (943) in shoot tissues. Thus, the N sensitive line KH showed a greater number of genes (5,189 DEGs) responding to LN stress, relative to the tolerant KRIL37 line (2,008 DEGs) in both tissues. To identify the genes specific to KRIL37 and KH, the DEGs obtained in KH comparison (LN vs. NN) were subtracted from those obtained from KRIL37 comparison (LN vs. NN) in both tissues. This led to the identification of 570 KRIL37-specific DEGs, from which 207 and 380 were specific to the root and shoot, respectively, and 3,768 DEGs were specific to KH for both tissues.

Fig. 4

DEGs identified in root (RT) and shoot (ST) tissues of KRIL37 and KH under LN (0.4 mM NH₄⁺) and NN (1.6 mM NH₄⁺) hydroponic growth conditions. Venn diagrams were used to present the upregulated and downregulated genes for different N conditions. Genotype-specific DEGs found in the RT (A) and ST (B). N-response DEGs found in the RT (C) and ST (D).

Differential expression was also tested between the two genotypes to capture their differences across a specific N treatment condition (KRIL37 LN vs. KH LN and KRIL37 NN vs. KH NN). The analyses identified 756 and 233 DEGs in roots under LN (KRIL37 LN vs. KH LN) and NN (KRIL37 NN vs. KH NN) conditions, respectively (Fig. 4C). Likewise, in the shoot, 515 genes were differentially expressed between KRIL37 and KH under LN conditions, whereas 213 were differentially expressed under NN conditions, indicating that LN stress conditions induced a wider dynamic transcriptional response than the NN condition (Fig. 4D). To identify genes specific to the LN and NN conditions (N-responsive), DEGs found under NN conditions (KRIL37 NN vs. KH NN) were subtracted from DEGs from LN and vice versa in both tissues. This led to the identification of 893 LN- and 236 NN-specific genes (Fig. 4C, D). Nitrogen response genes, such as ammonium transporter 1 (OsAMT1) and cytosolic glutamine synthetase 1 (OsGS1), were significantly upregulated in roots relative to shoots under LN conditions, suggesting the importance of roots in nutrient uptake and assimilation.

Gene ontology and pathway analysis

To understand the functions of the DEGs and the biological processes they are involved in, Gene Ontology (GO) enrichment and pathway analyses were done for each genotype-specific and N-responsive group of DEGs. Four sets of DEGs, identified as KRIL37-specific, KH-specific, LN-specific and NN-specific DEGs were used for GO analyses based on a singular enrichment analysis implemented through agriGO v.2 (Tian et al. 2017). In addition, pathway analyses were done using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotation database. The DEGs were classified under ‘molecular function’, ‘cellular component’ and ‘biological process’, involving several GO terms each.

Under ‘biological process’ (Fig. 5A), KRIL37-specific DEGs were grouped under several GO terms; they included the response to chemical stimulus, response to oxidative stress, electron transport chain and protein folding. KH-specific DEGs were grouped under the above-mentioned GO terms and several others, such as carbohydrate catabolic process and response to hormone stimulus. However, cell death and phenylpropanoid and isoprenoid metabolic processes were induced only in KH. Generally, the LN-sensitive genotype KH induced a wider suite of genes associated with response to stress and chemical stimulus than the tolerant line KRIL37. Hence, KRIL37 was not affected by N stress very much compared to KH. Similarly, regulation of gene expression was highly enriched in KH-specific DEGs (177) compared with that in KRIL37 (24).

Under the ‘molecular function’ category (Fig. 5B), both DEG sets induced calcium ion binding and oxidoreductase activity genes. Peroxidase activity was highest in KH-specific DEGs compared to that of KRIL37-specific DEGs. Among the cellular component GO terms (Fig. 5C), cell part, cell, intracellular organelle, cell wall and extracellular region were found to be enriched both for KH-specific and KRIL37-specific DEG sets. Furthermore, the apoplast and external encapsulating structure were enriched in KH-specific DEGs only. The differences in GO enrichment terms showed how the different genotypes responded to LN conditions.

In total, 59 and 15 significant GO terms (Supplementary Fig. S5) were enriched among LN- and NN-specific DEGs, respectively. As expected, several enrichment processes, in addition to a high number of genes involved in stress response and oxidative stress, were found under LN (67) compared to those found for NN (25) conditions in the biological domain (Supplementary Fig. S5A). In addition, no metabolic component was found for the NN condition DEG set; however, peroxidase activity was enriched in LN condition DEG set (Supplementary Fig. S5B). Cellular components such as the apoplast (GO:0048046), extracellular region and chloroplast were only found among LN-specific DEGs, most probably due to their relevance in mineral transport and N foraging under LN conditions.

The pathway analysis of DEGs revealed major plant pathways connected with metabolism, cellular processes and response to abiotic and biotic stresses. A large proportion of the DEGs in both comparisons were involved in the metabolism and biosynthesis of secondary metabolite pathways, with KRIL37-specific DEGs involved in only 10 significant pathways compared with 23 in KH-specific DEGs (Fig. 5D). Some pathways, such as glutathione metabolism, starch and sucrose metabolism, N metabolism and linoleic acid metabolism, were common to both KH- and KRIL37-specific genes. However, specific pathways were also observed for each group. Photosynthesis-antenna proteins, biosynthesis of amino acids and monoterpenoid biosynthesis were pathways unique to KRIL37-specific DEGs. Moreover, only KRIL37-specific DEGs were involved in the circadian rhythm pathway, including circadian-associated rice pseudo response regulator, control of flowering time (OsPRR1, OS02G0618200), regulation of circadian rhythm and flowering time, photoperiodic control of flowering, and osmotic stress response (OsGIGANTEA, Os01g0182600). MAPK signaling pathway, plant hormone signal transduction, brassinosteroid biosynthesis, DNA replication, galactose metabolism and Advanced Glycation Endproducts-Receptor for Advanced Glycation Endproducts (AGE-RAGE) signaling were enriched pathways among KH-specific DEGs. Interestingly, stress-induced cell response pathways, such as autophagy and phagosome, were enriched only in KH DEGs, revealing the extent of LN stress on KH plants. Autophagy is needed for remobilizing amino acids from cytosolic proteins complexes, including Rubisco from chloroplast, into the vacuole by exo- and endopeptidases localized in the vacuole (Ishida et al. 2008). Interestingly, diterpenoid and carotenoid biosynthesis, instead of monoterpenoid biosynthesis, was enriched in KH DEGs. These results showed that there were basic common and unique pathways for both genotypes to cope with environmental N conditions. Similar to the observations on the number of GO terms activated, more pathways were enriched among LN DEGs compared to those enriched among NN DEGs (Supplementary Fig. S5D). Terpenes and secondary metabolite pathways; carotenoid, diterpenoid, zeatin and phenylpropanoid biosynthesis; and other pathways, such as plant hormone signal transduction, MAPK signaling pathway and alanine, aspartate and glutamate metabolism, were enriched under LN conditions, indicating their importance in the response of rice plants to N stress. MAPK-pathway-related proteins, such as OsCHI11 and OsCatA, were significantly upregulated under LN, suggesting that the MAPK pathway is important for LN stress response. Similar findings were recently reported for wheat (Yan et al. 2021).

Fig. 5

GO and KEGG pathway enrichment analysis of KRIL37- and KH-specific DEGs (genotype-specific). GO, biological process (A), metabolic function (B) and cellular component (C). (D) KEGG enrichment analysis of KRIL37-specific and KH-specific DEGs. Values indicate the enrichment ratio, NA = pathway not enriched.

Differential regulation of gene expression by TFs

TFs regulate the expression of several genes; they play an important role in regulating plant growth and in plant adaptation to biotic and abiotic stresses. Comparatively (according to an analysis on PlantTFDB v.5.0), a wider repertoire of TFs (among DEGs) was expressed in KH plants responding to N stress than in KRIL37 plants. There were 222 and 47 TFs encompassing 37 families identified within the KH- and KRIL37-specific DEGs, respectively (Table 1). The WRKY, Ethylene Response Factor (ERF) NO APICAL MERISTEM/Arabidopsis thaliana ACTIVATING FACTOR1/ Arabidopsis thaliana ACTIVATING FACTOR2/CUP-SHAPE COTYLEDON2 (NAC) basic helix-loop-helix (bHLH) and MYB families of TFs were the most abundant, covering over 45% of the total TFs expressed; in addition, SQUAMOSA promoter-binding protein (SBP), GATA family protein, TEOSINTE BRANCHED/CYCLOIDEA/PROLIFERATING CELL FACTOR (TCP), ethylene insensitive-like (EIL), GLABROUS1 enhancer-binding protein11 (GeBP), Nin-like and three-amino-acid-loop-extension (TALE) encoding TF genes were specific to KH DEGs. Surprisingly, NAC, a well-known stress-responsive TF, was not expressed in KRIL37 plants; this was indicative of its sturdiness against N stress. Also, a total of 67 and 20 TFs covering 20 families were identified among LN and NN DEG sets, respectively. Several TF families, such as Related to ABSCISSIC ACID INSENSITIVE3/VIVIPAROUS-1 (RAV), LBD, GeBP, CO-like, TALE, and MYB-related, were significantly induced only under LN stress conditions. Members of the NAC, ERF, WRKY and MYB TF families were the most abundant TFs among LN responsive DEGs. These results clearly show that there are some differences in the response of TFs to N availability.

Table 1

Number of DEGs encoding TFs in the genotype-specific (KRIL37-specific and KH-specific) and N-responsive (LN-specific and NN-specific) DEGs

TFs (description)	KRIL37-specific	KH-specific	LN-specific	NN-specific
RAV (RAV family protein)	–	2	1	–
WRKY (WRKY family protein)	4	24	6	2
SBP (SQUAMOSA promoter binding proteins)	–	4	–
GATA (GATA family protein)	–	2	–	–
Trihelix (Trihelix family protein)	2	6	–	–
bHLH (basic/helix-loop-helix family proteins)	1	22	3	1
Basic-leucine zipper (bZIP family protein)	–	20	3	2
TCP (TCP family protein)	–	3	–
NAC (NO APICAL MERISTEM, Arabidopsis thaliana ACTIVATING FACTOR and CUP-SHAPE COTYLEDON TFs)	–	18	12	1
HSF (Heat stress TFs)	4	3	2	2
Dof (DNA binding with one finger protein)	1	3	–	–
LATERAL ORGAN BOUNDARIES DOMAIN (LBD family protein)	2	2	1
EIL [Ethylene-insensitive-like (EIL) family proteins]	–	2	–	–
NUCLEAR-FACTOR-Y, subunit A (NF-YA family protein)	2	5	1	–
GeBP (GeBP family protein)	–	1		–
NUCLEAR-FACTOR-Y, subunit B (NF-YB family protein)	–	2	1	–
Nin-like [Nin (for nodule inception)-like protein]	–	1	–	–
B3 (B3 family protein)	–	3	–	1
ERF (ERF family protein)	9	18	10	4
DBB (DOUBLE B-BOX family protein)	2	2		–
CO-like [CO (CONSTANS) family protein]	5	2	4	–
ARR-B (ARR-B family protein)	–	1	–	–
G2-like (G2-like family protein)	2	5	1	1
MIKC_MCM1, AGAMOUS, DEFICIENS, and SRF (MIKC-MADS family protein)	–	6	–	1
HB-other (Homeobox-other family protein)	–	1	–	–
WOX (WOX family protein)	–	1	1	–
TALE	–	2	1	–
HD-ZIP (HD-ZIP family protein)	1	13	2	1
C2H2 (C2H2 zinc finger domain)	–	14	4	1
C3H (Cys3His-containing zinc finger domain)	–	3	–	1
LESION SIMULATING DISEASE (LSD family protein)	–	1	–	–
Myeloblastosis (MYB family protein)	3	18	9	2
MYB_related (MYB-related family protein)	7	7	5	–
GIBBERELLIC-ACID INSENSITIVE, REPRESSOR of ga1-3 and SCARECROW (GRAS family protein)	–	2	–	–
“NF-YC”, NUCLEAR-FACTOR-Y, subunit C (NF-YC family protein)	–	1	–	–
Zinc finger-homeodomain (ZF-HD family protein)	–	1	–	–
APETALA2 (AP2 family protein)	2	–	–	–

Nitrogen uptake, assimilation and utilization pattern

The N content in the growth medium significantly affected the expression of N-utilization-related genes in plant tissues; consequently, it influenced N uptake, translocation and assimilation in plants from both genotypes (Fig. 6). The transcription levels of the ammonium transporter OsAMT1;2 (OS02G0620600) were significantly downregulated in about 3-folds in KH plants grown under LN conditions, whereas the expression levels in KRIL37 plants grown under normal and deficient N remained the same, which could have resulted in enhanced ammonium uptake and transport mechanisms in KRIL37 under LN stress. The ammonium transporter OsAMT1;3 (OS02G0620500) expression was found to be similar in KH plants irrespective of the N concentration in the growing medium; however, the levels OsAMT1;3 in KRIL37 LN were significantly higher than those in KH LN in roots. Similarly, although the ammonium transporter OsAMT2:1 (OS05G0468700) was upregulated in KH LN plant tissues, its expression level was no different from that in KRIL37 LN tissues. These high-affinity NH₄⁺ transporters could have positively influenced N uptake in KRIL37 under LN.

Fig. 6

Differential expression of N-utilization-related genes in the ST and RT of rice genotypes KH and KRIL37 subjected to N treatments such as LN (0.4 mM NH4 +) and NN (1.6 mM NH4 +) in hydroponic growth conditions (e.g. ST KH LN vs. NN). Cells with * indicate DEGs (|log₂ FC> 1 or log₂ FC| < −1, FDR < 0.05). Gene codes for each gene can be found at The Rice Annotation Project Database (https://rapdb.dna.affrc.go.jp/).

One striking difference between KRIL37 LN and KH LN involved N-assimilation-related genes, especially in the roots. The genes encoding glutamate synthetase OsGOGAT1 (OS01G0681900), glutamate synthase OsGOGAT2 (OS05G0555600), glutamine synthetase OsGLN1;2 (OS03G0223400) and glutamate-receptor-related protein OsGLR1.2 (OS02G0787600) were significantly suppressed in KH LN (when comparing KH LN vs. KH NN), whereas they were either upregulated or unchanged in KRIL37. For example, mRNA accumulation levels of OS01G0681900 and OS03G0223400 in KRIL37 were not different between LN and NN conditions, but significantly lower levels were recorded in KH. Glutamine synthetases are known to be induced by ammonium, and the relatively lower expression levels observed in KH but not in KRIL37 under LN conditions could be a result of enhanced NH4⁺ uptake in KRIL37 similar to what was reported by Ranathunge et al. (2014). The expression levels of N-assimilation-related genes in KRIL37 correlated with high N uptake and use efficiency observed in KRIL37 compared with KH under N-deficient conditions.

Plant hormones

Phytohormones play a significant role in the regulation of development and environmental responses in plants. Several DEGs related to plant hormone biosynthesis and signaling functions, such as those involving salicylic acid, jasmonic acid, strigolactone and auxin, play key roles in responding to both biotic and abiotic stresses. Much of these were induced in LN-sensitive variety (KH LN vs. KH NN) compared with KRIL37, including gibberellin signaling and cytokinin-O-glucoside biosynthesis, which were induced only in KH but not in KRIL37 (Fig. 7).

Fig. 7

Genotype-specific DEGs related to phytohormone biosynthesis and signaling. The number of DEGs enriched between LN and NN in KRIL37 and KH are shown for each category.

Discussion

In paddy fields, ammonium is preferred as the inorganic N source because poor aeration suppresses the biological process of nitrification (Sasakawa and Yamamoto 1978). However, crop plants use less than half of the applied N, and the excess is lost to the environment, raising concerns. Consequently, reducing fertilizer use is necessary for sustainable agriculture (Socolow 1999, Guo et al. 2010), and crop plants must be enhanced to efficiently absorb and utilize N to maintain yields at reduced fertilizer levels. Therefore, an improved understanding of crop responses to N stress conditions at phenotypic and molecular levels is required to make such genetic improvements. Considering the large genetic variability available in rice for traits including response to N application, we applied a comparative transcriptomic approach to differentiate transcriptional responses in the tissues of two rice genotypes, i.e. KRIL37 and KH; they showed contrasting responses to changes in external ammonium concentrations.

In general, LN conditions affected the overall growth of both rice genotypes in hydroponic, pot and field cultivation; however, the IL KRIL37 was more tolerant or, in some cases, non-responsive to N stress compared to the cultivated type KH. Under LN, root length increased compared to that observed in plants grown under NN conditions. Root growth and development are induced under N-deficient conditions to enhance nutrient foraging to meet the requirement of the plant (Ogawa et al. 2014). Earlier reports indicated that variable N conditions have drastic and contrasting effects in the development of root architecture, including seminal root length, in different species or varieties (Obara et al. 2010). In Arabidopsis thaliana, root elongation has been observed under reduced N concentration (Sánchez-Calderón et al. 2005), whereas a 52% increase in root length was observed in indica rice variety Pokkali grown under similar conditions (Subudhi et al. 2020).

The parameters measured on 3-week-old seedlings showed the superiority of IL KRIL37 over KH plants under LN conditions. The long roots and high root and shoot dry weights observed in KRIL37 plants grown under LN conditions indicated enhanced tolerance by the IL (Fig. 2). Similarly, the high N concentration in shoot tissues of plants grown under LN conditions proved KRIL37 to be more efficient in N uptake than KH. NUpE is influenced by several factors including the mass flow of soil water to the root, root morphology, transporter activity and others (Garnett et al. 2015). It is already reported that plants with relatively high rates of N uptake under limited conditions can use N more efficiently (Barber 1995). The genotypic differences in NUpE translated into highly significant NUE in KRIL37. High root and shoot parameters observed in KRIL37 could be associated with an efficient uptake system when N supply was low.

Straw, panicle and brown rice weights under various nutrient conditions (LN, −N and −F) in pots and paddy fields revealed the sensitivity of KH plants to N deficiency compared to KRIL37 (Fig. 3A–H). Therefore, the morphological and physiological superiority of KRIL37 under LN conditions could be attributed to de novo genomic changes induced by the introgressive hybridization process with its wild relative O. rufipogon. Wild species germplasms are of particular interest as donors of valuable alleles that may have been eliminated or weakened from the cultivated gene pools during domestication (Zhang et al. 2006, Tester and Langridge 2010). Rapid emergence, followed by vigorous, competitive vegetative growth and reproductive development are some features that make wild-type crops very adaptive to stressful conditions compared with cultivated species (Delouche et al. 2007). In rice and other cereals, wild species, including wild relatives, have been used to improve the response of cultivated types to biotic and abiotic conditions (Maxted and Kell 2009). Using chromosome segment substitution, O. rufipogon was used to develop lines with improved agronomic performance under N-deficient conditions (Ogawa et al. 2016). ILs developed from O. sativa cultivar 93-11 and O. rufipogon were found to be more tolerant to salt stress (Wang et al. 2017). Pathogen inoculation studies demonstrated that rice IL RZ35 developed from wild species Zizania latifolia showed enhanced resistance to rice blast compared to that of the recurrent cultivar Matsumae (Wang et al. 2013).

RNA-seq analysis was used to understand the responses to N availability among the IL KRIL37 and the cultivated line KH at the transcriptome level. RNA-seq analysis results showed that 5,189 and 2,008 genes were differentially expressed in KH and KRIL37, respectively, when grown under LN stress and NN conditions. This represents a wider transcriptional response in the N-sensitive line KH compared to that in KRIL37. This suggests that, when facing N-deficient growth conditions, the tolerant line KRIL37 was more transcriptionally stable than KH; these results are in agreement to what was reported for tolerant and sensitive indica genotypes and Corchorus capsularis under stress conditions (Ereful et al. 2016, Yang et al. 2017). Nonetheless, a range of DEGs (1,346) overlapped between the two genotypes (Fig. 4A, B); they might be involved in inherent mechanisms associated with nitrogen metabolism and regulation, regardless of differences in genotype or cultivar. Furthermore, this large number was expected for genotypes with similar genetic backgrounds (as that of KH in the lines under study). Subudhi et al. (2020) reported 415 overlapping genes between indica and japonica rice varieties in response to N stress.

TFs are important for controlling the expression of genes involved in various signaling pathways, plant growth and stress responses. Among the DEGs encoding TFs, WRKY, ERF, NAC, bHLH and MYB TF family members were the most abundant. ERF, WRKY, bHLH, MYB, C₂H₂ and NAC TFs are known to play a multifaceted and overlapping role in the regulation of cell growth, development and proliferation; plant hormone signaling; and biotic and abiotic stress responses, regulation and metabolism (Toledo-Ortiz et al. 2003, Nakano et al. 2006, Rushton et al. 2010). Herein, several ERF, WRKY, bHLH and MYB TF members were induced under N stress. Noteworthily, WRKY family proteins are among the major plant stress regulators through the jasmonic and salicylic acid hormone-mediated defense responsive pathways (Ryu et al. 2006). ETHYLENE-RESPONSIVE ELEMENT-BINDING FACTOR 101 (OsERF101, OS04G0398000), which is involved in salt stress (GO:0009651), water deprivation responses and inflorescence development (GO:0010229), was significantly induced under N stress conditions. Also, the HEAT STRESS TRANSCRIPTION FACTOR (OsHsfC1b, OS01G0733200) and basic helix-loop-helix protein 37 (OsbHLH037, OS01G0218100), affecting abiotic stress and root development, were significantly induced in the roots of KRIL37 when grown under N-deficient conditions, further suggesting a role in tolerance to LN conditions through root development and physiology. The overexpression of RDD1 (Rice Dof daily fluctuations1) in rice increased the uptake and assimilation of ammonium ions under LN conditions due to an enhanced expression of the ammonium transporter OsAMT1.3 in the roots of overexpressing plants. This further resulted in more spikelets/panicle and grain number/plant in the RDD1 overexpressing plants than in the wild type (Iwamoto and Tagiri 2016). From RNA-seq analysis, Yang et al. (2015) reported 85 TFs that, under N deficiency, played an important role in N deficiency response and plant growth. In Cyanidioschyzon merolae alga, R2R3-type MYB TF (CmMYB1) was reported to regulate N assimilation and enhance the expression of CmNRT, CmNAR, CmNIR, CmAMT and CmGS under N starvation (Imamura et al. 2009). In this study, some other TFs were also modulated by N deficiency. GATA TRANSCRIPTION FACTOR7 (Os10g0557600), a brassinosteroid-mediated growth regulator involved in panicle and grain development, was significantly downregulated in KH plants grown under LN, whereas it was upregulated by 2-folds in KRIL37 plants under similar conditions. In A. thaliana, GATA21, a member of the GATA factor family of zinc finger TFs that modulate chlorophyll biosynthesis and glutamate synthase (GLU1/Fd-GOGAT) expression, is induced by nitrate, and loss-of-function mutants cause reduced chlorophyll levels and downregulation of the genes involved in carbon metabolism (Bi et al. 2005). Similarly, the expression of ARABIDOPSIS NITRATE REGULATED 1 (AtANR1), a TF bearing a MADS box, was suppressed under N-deprived conditions, resulting in alterations in lateral root proliferation due to a reduction in sensitivity to NO₃⁻ (Zhang 1998). These data show that genotypic differences in rice influence transcriptional regulation mechanisms, depending on N availability and uptake capacity.

Phytohormones such as auxins and cytokinins have been reported to play crucial roles during plant adaptation to limited N (Krapp et al. 2011). Gibberellin, cytokinins and auxins have been found to be involved in nitrate-signaling pathway networks and in plant developmental processes, including root architecture for stress adaptation (Fredes et al. 2019). For example, crown rootless1 (crl1), a gene involved in crown root formation in rice, is a target of AUXIN RESPONSE FACTOR proteins. Inukai et al. (2005) reported, in their crl1 mutant studies, that auxin promoted the initiation of crown roots in rice. In this study, we found that an auxin efflux transporter (OsPIN2, OS06G0660200) and an auxin efflux carrier domain-containing protein (OsPILS6a) were significantly repressed in the sensitive line KH during LN stress.

Earlier reports have shown that the inhibition of auxin transport leads to blockage of PIN-FORMED1 cycling and results in reduced root development in sensitive genotypes under N stress (Geldner et al. 2001, Gelli et al. 2014). In addition, the BIG GRAIN1 (BG1) gene encodes a positive regulator of auxin response and transport whose action leads to increased indole-3-acetic acid levels in the panicle, resulting in the production of large grains upon activation (Liu et al. 2015). Although cytokinins act as auxin antagonists, they play an important role in the production of lateral root primordia by affecting auxin distribution (Sabatini et al. 1999, Dello Ioio et al. 2007). In rice, overexpression of the cytokinin-induced type-A RESPONSE REGULATOR6 (OsRR6) led to decreased growth of lateral roots (Hirose et al. 2007). We found that OsRR6 (OS04G0673300) was downregulated in both KH and KRIL37 under nutrient stress as a countermeasure to promote root development; however, the expression of OsRR6 was only slightly reduced in KRIL37 compared with that in KH plants. Taken together, these data suggest that phytohormones may have been involved in the reduced root growth observed in KH plants cultivated hydroponically under N stress conditions.

Nitrogen is absorbed and transported by proteins, such as ammonium transporter (AMTs) and NRTs (Xu et al. 2012). AMTs involved in N uptake, such as OsAMT1;2 and OsAMT1;3, are characterized as high-affinity ammonium transporters, suggesting that they should facilitate ammonium uptake under deficient conditions. OsAMT1;2 and OsAMT1;3 showed significantly high expression levels under LN in KRIL37 compared with those in KH under similar N-deficient condition in the root. The variation in expression of such transporters under LN conditions probably arising from the introgressed fragments could play a major role in enhancing N uptake, particularly under limited N conditions in KRIL37, resulting in high N accumulation in shoot and comparatively high NUpE and NUE. Lee et al. (2020) indicated that, the increase in expression levels of OsAMT1;2, OsAMT1;1 and OsAMT1;3 in the rice roots were in good accordance with the higher N uptake measured in the roots using ¹⁵NH₄⁺. In rice, improvement in NUE is influenced by NUpE, and this was corroborated by the increased expression of N transporters (Hu et al. 2015, Chen et al. 2016).

Following uptake, ammonium is incorporated into amino acids, mainly through the GS/GOGAT cycle in plants (Kant et al. 2011). Our results indicated that the expression of several N assimilation genes, including Glutamate synthetase 1 (OsGOGAT1), Glutamate synthase NADH (OsGOGAT2), Glutamine synthetase 1;2 (OsGLN1;2) and Glutamate receptor-related protein 1.2 (OsGLR1.2), was inhibited under LN in KH plants, but upregulated or unchanged in KRIL37 (Fig. 6). Under optimal NH₄⁺ conditions, glutamine synthetase activity is enhanced (Ishiyama et al. 2004, Masclaux-Daubresse et al. 2005). The downregulation of these genes in KH under LN could be due to the absence or reduced levels of the corresponding substrate at each enzyme-catalyzing step of the N assimilation process when facing LN stress conditions. Several studies have reported relatively reduced transcript levels of N assimilation genes under chronic N stress (Sinha et al. 2018, Subudhi et al. 2020).

In contrast, the simultaneous overexpression of OsAMT1.2 and OsGOGAT1 in rice resulted in increased plant height and biomass in the overexpression line compared to the control under low NH₄⁺ conditions; this was due to an increase in ammonium uptake and assimilation resulting in enhanced expression of OsGS1;1 and OsGS1;2 and higher GS activity in the overexpression line (Lee et al. 2020). Here, the allelic differences resulting from O. rufipogon introgression might be responsible for the qualities observed in the IL KRIL37, resulting in the increased expression of OsAMTs for enhanced uptake of N under LN to satisfy the NH₄⁺ substrate requirement of GS/GOGAT (Lea and Miflin 2003), further promoting a relatively better growth than that of KH under N-deficient conditions.

Our study showed the response of IL KRIL37 and its parental cultivar KH to LN and NN supplies and provided clues to further understand the mechanisms, pathways or processes involved in the contrasting adaptation responses of different genotypes to changes in the availability of N. An RNA-seq comparative analysis revealed that KRIL37 induced N-assimilation-related genes for enhanced N uptake and utilization; this could positively influence its adaptation to LN conditions. We speculate that the introgressed genome fragments from O. rufipogon may activate some of the N transport/assimilation genes encoded in the KH genome to enable a better growth of KRIL37 plants under LN conditions. These results provide useful information for future research on molecular breeding of LN-tolerant rice lines or cultivars.

Materials and Methods

Plant materials

The rice IL KRIL37 (cultivated rice KH with an introgressed genomic region from the wild species O. rufipogon) and its recurrent parent KH were used in this study. KH was part of our laboratory stock, while KRIL37 was kindly provided by Dr. Hideyuki Hirabayashi (Hirabayashi et al. 2010).

Hydroponics cultivation

Seeds were incubated at 45°C for 4 days to break dormancy before using for germination. Seeds were surface sterilized with 10× diluted bleach (6% hypochlorite solution) for 30 min and washed with distilled water. Then, seeds were incubated on filter papers with distilled water until germination at 29–30°C inside a controlled growth chamber (for ∼5 d). Germinated seeds were transferred to dark-colored 3-l polyethylene containers with liquid media of different conditions. Modified Kimura’s B solution (Tanaka et al. 2018) was used as the base of the N-free culture media (NH₄Cl, K₂SO4 and CaSO₄·2H₂O were used in place of (NH₄)₂SO₄, KNO₃ and Ca(NO₃)₂·4H₂O, respectively). To create different N growth conditions, 0.4 mM and 1.6 mM NH₄Cl were used in the experiment as LN and NN, respectively. Hydroponic cultivation was done in a growth chamber (NK System Biotron, Nippon Medical and Chemical Instruments Co. Ltd, Osaka, Japan) at 29–30°C with ∼69% Relative humidity under fluorescent lamps and 14.5 h:9.5 h light:dark cycles for 3 weeks. Culture medium was renewed every week. After 3 weeks, seedlings were sampled for measuring shoot and root length, shoot and root dry weight and shoot and root C/N ratio. For dry weight and C/N ratio determination, samples were dried in an oven at 70°C for at least 4 d. To determine the C/N ratio, dried samples were weighed (≤100 mg) in crucible containers for combustion using a CN elemental analyzer (vario MAX cube, Elementar, Kanagawa, Japan).

Estimation of NUpE, NUtE and NUE (Moll et al. 1982): NUpE was estimated as shoot N/N in nutrient solution and NUtE as dry weight of shoot/weight of N in shoot while NUE was estimated as dry shoot weight/N in the nutrient solution.

Pot cultivation

Rice lines were also grown using 2.5-l pots. To create different N soil conditions, 10.5 g (NH₄)₂SO₄/pot or 2.63 g of (NH₄)₂SO₄/pot for NN or LN soil conditions, respectively, were thoroughly mixed with ∼4.4 kg Akadama soil (a low-nutrient soil) (SOWA Recycle Corporation, Tokyo, Japan), watered and let stand for at least 1 week before cultivation. The soil was then used to fill the pots to about 80% of their volume, and 3-week-old seedlings of nearly uniform size were transplanted into the pots; two seedlings per pot. Plants were kept inside a greenhouse under natural light supplemented with artificial light. After maturity, rice samples were harvested and dried for about 2 weeks to determine the panicle and straw dry weights.

Field cultivation

Three- to four-week-old rice seedlings were transplanted for cultivation in the paddy fields at three different locations. At Yayoi campus of the University of Tokyo, physically isolated blocks measuring 1 × 5 m were used. Blocks were subjected to three different fertilization treatments as follows: (5 g N + 20 g P + 5 g K)/m², (0 g N + 20 g P + 5 g K)/m² and (0 g N + 0 g P + 0 g K)/m², for N supply (+N), no N (−N) and no fertilizer supply (−F) soil conditions, respectively. Plants were evenly distributed in 30 rows (1 m long, 10 plants per row) with 20 cm between rows. A paddy field of local farmers at Iinomachi, Fukushima, was also used. This paddy field has long been used for rice production; at the time of the trial described here, rice was grown without additional fertilization. Plant spacing was roughly similar at Iinomachi, Fukushima, to that used at Yayoi fields. Finally, a paddy field at the University of Shiga Prefecture (Shiga, Japan) with physically isolated blocks for fertilizer supply (+F: 5 g N + 5 g P + 5 g K/m²) and −F soil conditions were also used. At the paddy field in the University of Shiga Prefecture, KRIL37 and KH seedlings were included in each experimental soil condition in triplicate, 24 plants per replicate in 2018 and 2021 while 48 plants per replicate in 2020. At maturity, the plants at the center of each plot (n = 8 and n = 24) were manually harvested (including all replicates) at every location and dried for at least 2 weeks before the determination of yield components.

RNA extraction and transcriptome analysis

KRIL37 and KH germinated seedlings from the hydroponic setup described above, grown under LN and NN conditions, were sown in 20-l containers with modified Kimura B solution at various concentrations. After 3 weeks of growth in the nutrient medium, five plants were collected and pooled for each replicate of the LN and NN treatments per cultivar. Root and shoot tissues from each genotype were collected separately, immediately frozen in liquid nitrogen and stored at −80°C. Total RNA was extracted from the tissues using the RNeasy Plant Mini Kit (Qiagen, Tokyo, Japan). Separate libraries for each treatment (LN and NN), tissue (root and shoot) and genotype (KRIL37 and KH) were constructed according to the manufacturer’s protocol. A total of 24 libraries were constructed: eight types of samples with three biological replicates for each. The libraries were prepared using the Lasy-Seq v1.1 protocol (Kamitani et al. 2019) and sequenced in a single-read 50-bp mode using a HiSeq 2,500 platform (Clockmics Inc., Osaka, Japan). Generated reads were first subjected to quality checks to trim ambiguity reads and adapters as well as a specified number of bases at either 3ʹ or 5ʹ end of the reads. fastp (https://github.com/OpenGene/fastp) was then used to remove low-quality reads with Phred Scores < 30 and read lengths < 36 bp. Reads were aligned to the rice reference sequence (from cv. Nipponbare) from the RAP-DB (https://rapdb.dna.affrc.go.jp) using Hisat2 (v2.0.1) (Kim et al. 2015) with default parameters. To obtain expression data, mapped reads were counted for each gene by featureCounts (Liao et al. 2014) using Oryza_sativa.IRGSP-1.0.23.gtf from EnsemblPlants (ftp://ftp.ensemblgenomes.org/pub/release-23) as a gene annotation file. Then, differential expression analysis across samples was conducted using the edgeR package (http://www.r-project.org/) to obtain DEGs at FDR, P value ≤ 0.05 and |log₂ FC| ≥ 1 (for upregulation) and ≤ −1 (for downregulation). For biologically meaningful comparisons, DEGs between plants grown under LN and NN conditions were identified separately for root and shoot tissues within each genotype. GO enrichment of the DEGs was performed using a singular enrichment analysis with agriGO v.2 (Tian et al. 2017), while KEGG pathway analysis was conducted using the KEGG (Kyoto, Japan) Orthology program (http://kobas.cbi.pku.edu.cn/). We used the P values calculated by the hypergeometric test and corrected them by FDR, taking FDR ≤ 0.05 as the threshold to identify the significant functional pathways.

Supplementary Data

Supplementary Data are available at PCP online.

Data Availability

The data underlying this article are available in DNA Data Bank of Japan Sequence Read Archive with accession nos. DRR351110–DRR351133.

Funding

Japan Society for the Promotion of Science, KAKENHI (17K17683, 19H05637 and 18H05490); Cabinet Office, Government of Japan, Moonshot R&D Program for Agriculture, Forestry and Fisheries (founding agency: Bio-oriented Technology Research Advancement Institution).

Acknowledgements

We thank Ms. Kyoko Y-Mogami for technical assistance. The ILs were kindly provided by Dr. Hideyuki Hirabayashi from the National Agriculture and Food Research Organization.

Disclosures

The authors have no conflicts of interest to declare.