To identify RWP-RK domain-containing transcription factors in rice (Oryza sativa), the rice protein database was obtained from JGI (https://phytozome-next.jgi.doe.gov; accessed on 5 Jan 2025; Osativa 323 v7.0) (Goodstein et al., 2012). Raw Hidden Markov Model (HMM) data of the conserved RWP-RK (PF02042) domain were downloaded from the Pfam database (http://pfam.xfam.org; accessed on 5 Jan 2025) and used to search for orthologues in the rice genome (Mistry et al., 2021). The RWP-RK domain in candidate proteins was verified using the NCBI Conserved Domains search tool (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi; accessed on 5 Jan 2025) (Marchler-Bauer et al., 2017; Wang et al., 2023). The physicochemical properties of OsRWP-RK proteins were predicted using ProtParam (https://web.expasy.org/protparam/; accessed on 5 Jan 2025) (Wilkins et al., 1999), and their subcellular localizations were predicted using Cell-PLoc (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/; accessed on 5 Jan 2025) (Chou and Shen, 2010).

Arabidopsis RWP-RK protein sequences were obtained from PlantTFDB v5.0 (https://planttfdb.gao-lab.org; accessed on 3 Jan 2025) (Jin et al., 2017). Sequences of 28 RWP-RK proteins from Arabidopsis thaliana and Oryza sativa were aligned using Mafft, and a phylogenetic tree was constructed with the maximum likelihood (ML) method in Fasttree, including bootstrap analysis with 500 replicates. Chromosomal localization and synteny analysis were performed using TBtools (version 2.194) (Chen et al., 2023).

Gene structure data were obtained from the genome annotation file on JGI (https://phytozome-next.jgi.doe.gov; accessed on 5 Jan 2025; Osativa 323 v7.0) and analyzed using TBtools (version 2.194) (Goodstein et al., 2012; Chen et al., 2023). Conserved motifs of OsRWP-RK proteins were identified with MEME (http://meme-suite.org/tools/meme; accessed on 3 Jan 2025) (Bailey et al., 2015), and the top three motifs were extracted. Conserved domains were predicted using NCBI-CDD (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi; accessed on 3 Jan 2025) and visualized in TBtools (version 2.194) (Marchler-Bauer et al., 2017; Chen et al., 2023; Wang et al., 2023).

The 3D structure of the OsRWP-RK proteins was predicted using the AlphaFold Protein Structure Database (https://alphafold.com/; accessed on 14 Jan 2025) (Jumper et al., 2021; Varadi et al., 2023).

To analyze the promoter sequences of OsRWP-RK transcription factors, 2000 bp upstream of each gene’s initiation codon was extracted using TBtools (version 2.194) (Chen et al., 2023). Cis-elements were predicted with PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html; accessed on 6 Jan 2025) and visualized using TBtools (version 2.194) (Lescot et al., 2002; Chen et al., 2023).

Expression data for five rice tissues (leaf, root, stem, inflorescence, and seed) were obtained from the Rice Gene Annotation Project (https://rice.uga.edu/expression.shtml; accessed on 7 Jan 2025). Heatmaps and histograms were generated using GraphPad Prism (version 9.5).

Rice seedlings of the Japonica variety ZH11 were grown for 15 days in a modified Kimura B nutrient solution (1 mM NO₃ ^- + 1 mM NH₄ ⁺) and then transferred to a nitrogen-free solution for 5 days. After nitrogen starvation, plants were returned to the modified solution, and roots were collected at 0, 10, 30, 60, and 120 minutes to obtain RNA. RNA was extracted using RNA isolator Total RNA Extraction Reagent (R401-01, Vazyme) and reverse transcribed with HiScript III All-in-one RT SuperMix (R333, Vazyme). Quantitative PCR was performed using ChamQ Universal SYBR qPCR Master Mix (Q711-02, Vazyme) on a LightCycler^® 480 Instrument II. The primers used for qRT-PCR are listed in Supplementary Data Sheet S1 . The experiment was conducted with three biological replicates.

The Oryza sativa genome contains 13 RWP-RK genes encoding 14 protein isoforms, including two variants of OsNLP5 ( Table 1 ), all harboring the conserved RWP-RK domain ( Figure 1 ). The coding sequence (CDS) lengths of the OsRWP-RK family range from 699 bp (OsRKD6) to 2829 bp (OsNLP1), with corresponding protein lengths spanning 232 (OsRKD6) to 942 amino acids (OsNLP1). Their molecular weights vary from 27.4 kDa (OsRKD6) to 104.7 kDa (OsRKD5), and predicted protein isoelectric points range from 5.17 (OsNLP6) to 8.46 (OsRKD4) ( Table 1 ). Proteins in the NLP subfamily generally possess longer CDS and higher molecular weights compared to those in the RKD subfamily ( Table 1 , Supplementary Figure 1 ), whereas OsRKD5 exhibits similar features to the NLP subfamily ( Table 1 ).

Subcellular localization prediction indicated that the majority (10/14) localized exclusively to the nucleus, consistent with their predicted roles as transcription factors ( Table 1 ). Notably, two exceptions were identified: OsNLP6 and OsRKD5 exhibited cytoplasmic localization ( Table 1 ), suggesting potential non-nuclear functions such as post-translational regulation or signaling. Additionally, OsRKD3 and OsRKD9 showed dual cytoplasm-nucleus distribution ( Table 1 ), implying shuttling dynamics that may regulate their activity.

These 13 OsRWP-RK genes are unevenly distributed across 8 of the 12 chromosomes, with chromosome 1 harboring three genes, chromosomes 2, 4, and 9 each containing two, and chromosomes 3, 6, 11, and 12 each with one gene ( Figure 2A ). Notably, OsNLP1, OsNLP4, and OsNLP6 are positioned near chromosome ends, suggesting potential telomeric regulatory roles ( Figure 2A ).

A phylogenetic tree of RWP-RK proteins from Oryza sativa and Arabidopsis thaliana was constructed to investigate their evolutionary relationships ( Figure 2B ). The analysis revealed that RWP-RK proteins are classified into four major groups, each containing specific genes from both species. Group 1 contains AtNLP1-5 and OsNLP1, OsNLP4. Group 2 includes AtNLP6-7, OsNLP3, and OsNLP6. Group 3 clusters AtNLP8-9, OsNLP2, OsNLP5.1, and OsNLP5.2. Group 4 contains all OsRKDs and AtRKDs. The NLP subfamily is distributed across Groups 1-3, while Group 4 represents the RKD subfamily, with clear differentiation between the two subfamilies, indicating functional divergence during evolution. The presence of multiple members in each group suggests that gene duplication events contributed to the expansion of the gene family. Furthermore, the clustering of genes from both Arabidopsis and rice indicates a shared evolutionary origin, suggesting that the RWP-RK gene family is ancient and conserved across species.

To elucidate structural diversity among OsRWP-RK genes, we analyzed their exon-intron organization, conserved motifs, and protein domain ( Figures 1A, B ). The genes consist of 3-6 introns and 4-7 exons, exhibiting variation in exon/intron length and arrangement. Additionally, structural differences between the two subfamilies may result in distinct regulatory patterns or expression profiles, potentially influencing their biological functions.

Motifs are small, conserved sequence fragments with functional significance in proteins. Motif analysis helps determine protein function. Using the MEME tool, three conserved motifs were identified in OsRWP-RK proteins ( Figure 1C ). Each protein contained 1-5 conserved motifs. The RKD subfamily retained the same functional motifs, while the NLP subfamily acquired two additional motifs, reflecting functional differences between the subfamilies.

Protein conserved domain analysis reveals core functional regions, evolutionary history, and biological functions of proteins. In this study, conserved domains were identified using the NCBI CDD database ( Figure 1D ). The results demonstrated that OsRKD subfamily members contain a single RWP-RK domain. In contrast, OsNLP proteins possess both the RWP-RK and PB1 domains ( Figure 1D ). This suggests that NLP subfamily members may adapt to specific functions by acquiring new domains, while RKD members retain their original domains to maintain basic functions. Additionally, the PB1 domain in the OsNLP subfamily is predominantly located at the C-terminus ( Figure 1D ), and the conserved spatial relationship between the RWP-RK and PB1 domains suggests a shared origin. Notably, OsRKD5 contains both a PLN03077 and a GBP_C domain ( Figure 1D ), indicating that the gene acquired new functions during evolution, enhancing rice’s adaptation to its environment.

The 3D structure of a protein dictates its function, with features such as the catalytic site and ligand binding site determined by its spatial arrangement. We predicted the 3D structures of OsRWP-RK family proteins, including OsNLP and OsRKD subfamily members ( Figure 3 ; Supplementary Figure 2 ). The models reveal differences in domain composition and overall folding. The OsNLP subfamily exhibits a relatively conserved 3D structure. Two large conserved domains (green and orange) are consistently located at the N-terminus ( Figure 3 ), forming the structural core and occupying a substantial volume. Two smaller regions, located near the C-terminus (with OsNLP4 containing only one), are spatially arranged around the large regions ( Figure 3 ). These domains are connected by flexible loops, suggesting potential cooperative function. Based on Figure 1D , the two small C-terminal domains are likely the RWP-RK domain (purple) and the PB1 domain (pink). Their spatial positioning suggests favorable orientation for DNA binding and nitrate signal perception (Konishi and Yanagisawa, 2013, 2019). In contrast, OsRKD proteins display greater structural variability, indicating potential functional diversity. Most OsRKD members exhibit a simplified architecture, often retaining only a single RWP-RK domain. This minimal structure may be sufficient for specific transcriptional regulation, such as the activation of genes involved in reproduction (Purwestri et al., 2023). Notably, OsRKD5 possesses a more complex structure, potentially due to the presence of additional domains—PLN03077 and GBP_C—as shown in Figure 1D . These structural characteristics provide valuable insights into the functional specialization and regulatory mechanisms of the OsRWP-RK protein family.

Synteny analysis of gene families reveals their evolutionary history, functional conservation, and inter-species relationships, and provides insights into adaptation, environmental response, and functional evolution. In this study, synteny analysis of the OsRWP-RK family highlights the conservation of the RWP-RK genes in rice and Arabidopsis, shedding light on their evolutionary patterns ( Figure 4 ). The analysis shows significant collinearity within the rice genome, such as between OsRKD6 (Chromosome 6) and OsRKD5 (Chromosome 2) ( Figure 4A ). Additionally, OsNLP2 (Chromosome 4) exhibits strong collinearity with a region on chromosome 2 ( Figure 4A ), suggesting gene duplication and evolutionary divergence, leading to the distribution and functional diversification of the OsRWP-RK family in rice. Furthermore, synteny between the rice and Arabidopsis RWP-RK gene families reveals four collinear gene pairs ( Figure 4B ), indicating that these families have retained conservation through evolution, likely originating from a common ancestor and undergoing gene recombination and species-specific divergence over time.

Cis-acting elements are crucial for gene regulation and can predict gene expression patterns under different tissue, organ, or environmental conditions. In this study, promoter analysis revealed diverse cis-regulatory elements in the 2 kb upstream regions of OsRWP-RK genes. Among the 80 detected elements, the CAAT-box and TATA-box are the most critical elements ( Figure 5B ; Supplementary Data Sheet S2 ). The CAAT-box enhances basal transcription by binding to CAAT-binding proteins, while the TATA-box facilitates the assembly of transcription initiation complexes, aiding RNA polymerase II in locating the transcription start site. Abundant cis-elements associated with hormone and abiotic stress responses were also identified ( Figure 5A ). Notable hormone-related cis-elements include ABRE (abscisic acid response), TGACG-motif and CGTCA-motif (jasmonic acid response), and as-1 (salicylic acid response). Abiotic stress-related elements include MYB and G-box (light response), MYC and MBS (drought response), MYC and LTR (low-temperature response), ARE (anaerobic response), and STRE (stress-responsive) ( Figures 5A, B ; Supplementary Data Sheet S2 ). All in all, the promoter region of the OsRWP-RK family contains various cis-acting elements that regulate gene expression and are involved in hormonal and stress responses.

To investigate the transcription patterns of OsRWP-RK genes, we obtained a gene expression matrix from the Rice Gene Annotation Project (https://rice.uga.edu/expression.shtml; accessed on 7 Jan 2025) ( Figure 6 ). Most NLPs are expressed across all tissues, suggesting their role in growth regulation in rice. ( Figures 6A, B ). OsNLP1-3, OsNLP5.1, and OsNLP6 are expressed at significantly higher levels in roots (TPM>10; Figures 6A, B ), suggesting shared transcriptional regulation under variational nutrient conditions. Furthermore, the OsRKD subfamily is generally expressed at low levels across tissues (TPM<10; Figure 6A ), while OsRKD1, OsRKD4, OsRKD6, OsRKD8, and OsRKD9 show relatively higher expression in inflorescences, exceeding levels in other tissues by at least two-fold ( Figures 6A, B ), further supporting subfamily-specific functional roles.

To further explore the diversity of transcriptional responses among OsRWP-RK genes upon nitrogen re-supply, we analyzed temporal expression profiles of 10 representative genes using qRT-PCR across five time points (0, 10, 30, 60, and 120 minutes; Figure 7 ). Most genes responded rapidly to nitrogen, with peak expression occurring at 10 minutes. For example, OsNLP1 and OsRKD3 showed strong and transient upregulation, followed by a sharp decline to near-baseline levels by 30 minutes. These early-transient responders may act as immediate nitrogen sensors or signal initiators. OsNLP2, OsNLP4, and OsNLP5.1 were also initially induced at 10 minutes, but their transcript levels subsequently declined to below or near baseline. This pattern suggests a transient and tightly controlled regulatory role, potentially functioning in early signaling checkpoints. OsNLP5.2 and OsRKD1 maintained elevated expression levels throughout the time course, indicating prolonged involvement in transcriptional regulation during the nitrogen response. OsNLP6 exhibited a unique two-phase expression pattern, with an initial increase at 10 minutes, a transient decrease at 30 minutes, and a second rise at 60 minutes. This oscillatory behavior may reflect feedback regulation or involvement in multiple signaling phases. OsNLP3 displayed a gradual increase in expression, reaching peak levels at 60 minutes. This delayed response suggests a role downstream of initial nitrogen signaling, possibly in the regulation of secondary targets (Zhang et al., 2022).