Genome-wide identification, characterization, and expression analysis unveil the roles of pseudouridine synthase (PUS) family proteins in rice development and stress response

Abstract

Pseudouridylation, the conversion of uridine (U) to pseudouridine (Ѱ), is one of the most prevalent and evolutionary conserved RNA modifications, which is catalyzed by pseudouridine synthase (PUS) enzymes. Ѱs play a crucial epitranscriptomic role by regulating attributes of cellular RNAs across diverse organisms. However, the precise biological functions of PUSs in plants remain largely elusive. In this study, we identified and characterized 21 members in the rice PUS family which were categorized into six distinct subfamilies, with RluA and TruA emerging as the most extensive. A comprehensive analysis of domain structures, motifs, and homology modeling revealed that OsPUSs possess all canonical features of true PUS proteins, essential for substrate recognition and catalysis. The exploration of OsPUS promoters revealed presence of cis-acting regulatory elements associated with hormone and abiotic stress responses. Expression analysis of OsPUS genes showed differential expression at developmental stages and under stress conditions. Notably, OsTruB3 displayed high expression in salt, heat, and drought stresses. Several OsRluA members showed induction in heat stress, while a significant decline in expression was observed for various OsTruA members in drought and salinity. Furthermore, miRNAs predicted to target OsPUSs were themselves responsive to variable stresses, adding an additional layer of regulatory complexity of OsPUSs. Study of protein–protein interaction networks provided substantial support for the potential regulatory role of OsPUSs in numerous cellular and stress response pathways. Conclusively, our study provides functional insights into the OsPUS family, contributing to a better understanding of their crucial roles in shaping the development and stress adaptation in rice.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12298-023-01396-4.

Introduction

RNA editing and modification are crucial processes that embellish nearly all types of cellular RNAs, encompassing tRNAs, rRNAs, snRNAs, snoRNAs, lncRNAs, and mRNAs. More than 170 distinct RNA modifications have been identified in different organisms (Machnicka et al. 2013; The RNA modification database-http://medlib.med.utah.edu/RNAmods/). Pseudouridine (Ψ), an isomer of uridine (U), is one of the most abundant and structurally unique modifications in both non-coding (Ofengand 2002) and coding RNAs (Carlile et al. 2014). It is often referred to as the ‘fifth nucleoside’ and it features a C–C bond between the nitrogen base and ribose sugar, providing enhanced flexibility and unique hydrogen bond donor capabilities. Ψ plays essential roles in structure, stability, processing, regulation of mRNA splicing events, interaction with other RNAs and ribosomal proteins, translational efficiency, and fidelity (Schwartz et al. 2014; Spenkuch et al. 2014; Elyer et al. 2019; Charette and Gray 2000; Zlotorynski 2022; Martinez et al. 2022). Pseudouridine synthesis is catalyzed by Pseudouridine Synthase (PUS) enzymes, employing two distinct mechanisms: guide RNA-dependent Ψ catalysis by H/ACA ribonucleoproteins (RNPs) and RNAindependent catalysis by stand-alone PUS enzymes (Karijolich et al. 2015). These stand-alone PUS enzymes are categorized into 6 subfamilies: TruA, TruB, TruD, RluA, RsuA, and PUS10 (McCleverty et al. 2007). Despite limited sequence similarity, almost all stand-alone PUS enzymes share a structurally similar catalytic motif characterized by a highly conserved aspartate (D) residue (Hamma and Ferré-D'Amaré 2006; Fitzek et al. 2018). PUS enzymes are ubiquitous in nature and the genes encoding these enzymes have been identified in all the genomes sequenced to date. Disruptions in Ψ sites due to mutations or modification of PUS enzymes result in various adverse effects in bacteria, yeast, and humans, including cell viability, growth defects, diseases, and stress-related aberrant phenotypes (Rintala-Dempsey and Kothe 2017; Lin et al. 2021).

Our understanding of pseudouridylation in plants remains somewhat limited in comparison to what is known in bacteria, yeast, and humans. In a pioneering study conducted by Sun et al. (2019), Ψ sites were identified in non-coding RNAs as well as mRNAs. The Psi-seq revealed the presence of 187 Ψ sites in rRNA, 282 in tRNA, and 13 in snRNAs of Arabidopsis. A total of 451 Ψ sites were detected within 332 distinct mRNAs which were distributed along the mRNAs, spanning the 5’ untranslated region (UTR), the coding region, and the 3’ UTR, with a discernible positional bias towards the 5’ UTR and coding regions (Sun et al. 2019). Genome-wide studies have identified numerous members of the PUS family across various plant species which underscores their pivotal role in plant evolution, development, and adaptation to the environment (Xie et al. 2022; Niu and Liu 2023). Loss-of-function mutants of SVR1 (Suppressor of Variegation 1 or AtRsuA1) exhibit defects in chloroplast rRNA processing and impairments in chloroplast translation, consequently impacting leaf variegation (Yu et al. 2008). A study by Dong et al. (2020) proposed the involvement of SVR1 in ABA-regulated root development. The PUS proteins are known to possess RNA chaperone activity in addition to their catalytic activity which was demonstrated when disruption of PUS activity in Maa2 did not significantly affect trans-splicing of chloroplast psaA RNA in Chlamydomonas reinhardtii (Perron et al. 1999). Another PUS member, encoded by FCS1 (Leaf Curly and Small 1 or AtRluA1), is localized to mitochondria and is responsible for pseudouridylation of mitochondrial 26S rRNA, which is crucial for mitochondrial translation, thereby regulating seed, root, and leaf development (Niu et al. 2022). AtCBF5 is essential for the development of nodule-like structure and plant growth regulated by thermospermine (Lermontova et al. 2007; Li et al. 2023). While the role of PUS is well-documented in plant development, their involvement in regulating stress responses is just beginning to be understood. Recent research involving integration of genome-wide identification with expression profiling of PUS family in Arabidopsis and maize has shed light on the potential roles of these proteins in fine-tuning stress responses (Xie et al. 2022; Dhingra et al. 2023; Niu and Liu 2023). For instance, the svr1 mutant exhibits reduced sensitivity to phosphate starvation (Lu et al. 2017). In rice, TCD3 plays an essential role in chloroplast rRNA processing and tRNA metabolism, thereby influencing chloroplast development during cold stress (Lin et al. 2020). A similar investigation by Wang et al. (2022) further implicated OsPUS1 (also known as TCD3) in controlling the response to cold stress by inhibiting the excessive accumulation of reactive oxygen species (ROS) in rice. Moreover, a genome-wide association study (GWAS) has identified the OsRluA5 gene (Os06g0717400) as a potential regulator of salinity tolerance in rice (Nayyeripasand et al. 2021).

Advancements in understanding the significance of PUS proteins in plants have primarily originated from research conducted on Arabidopsis. However, our knowledge regarding the functions of PUS proteins in agronomically-important crops such as rice remains relatively limited. Therefore, the present study aimed at identification, classification, and expression profiling of the PUS gene family in rice, uncovering 21 members belonging to 6 subfamilies. We conducted in-depth analyses, including genomic organization, gene structure and phylogenetic analysis, motif and domain composition, homology modeling, and subcellular localization of rice PUS proteins. We further examined non-synonymous SNPs among 3024 rice accessions to gain insights into allelic variations of 21 OsPUSs. To comprehend the regulation of various PUS family members we scrutinized the cis-regulatory elements within their putative promoter regions and identified the miRNAs that potentially target the rice PUS genes. We compiled an expression atlas of 21 OsPUS genes in different tissues and stress conditions utilizing publicly available datasets and validated their expression pattern using quantitative RT-PCR under different abiotic stress conditions which highlighted the differential expression of several OsPUS genes. These findings provide a robust foundation for elucidating the roles of OsPUS genes in developmental processes and under abiotic stress conditions, thereby opening up the possibility of utilizing these genes to bolster stress resilience in rice.

Materials and methods

Identification, classification, phylogenetic analysis, chromosomal distribution, duplication, and structure of OsPUS genes

The BLASTp search was performed with an e value ≤ 1e-10 in the Rice Genome Annotation project (RGAP) using Arabidopsis PUS family members. The sequences matched with those reported by Xie et al. (2022), wherein 22 rice PUS sequences were also identified using Hidden Markov model (HMM)-based search against rice proteome sequences via HMMER with e-value 1*10^–5. We retrieved 22 nucleotide and protein sequences from RGAP (Rice Genome Annotation project) (http://rice.uga.edu). Protein sequences were then submitted to NCBI CDD (Conserved domain database) (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) for domain analysis. Information regarding splice variants, ORFs length, amino acids length, molecular weight, and pIs were retrieved from RGAP. Active site consensus sequence and the position of aspartate (D) residue were manually identified. For phylogenetic analysis, multiple sequence alignment of amino acid sequences of PUS proteins was executed by the ClustalX2.1 and an unrooted neighbor-joining phylogenetic tree with 500 bootstrap value was generated using MEGA X (https://www.megasoftware.net).

The chromosomal locations of all the 21 OsPUS genes were determined and mapped using RGAP and TB tools (v1.115). The MCScanX feature and advanced Circos tool from TBtools was employed to analyze, annotate, and represent the duplication events within the rice genome. For gene structure analysis, full-length genomic and cDNA sequences of all the 21 OsPUSs were retrieved from RGAP by providing the TIGR IDs of all genes. Gene structures, including intron and exon positions, were identified, analyzed, and schematically represented using Genome Structure Display Server (GSDS v2.0; http://gsds.gao-lab.org/).

Prediction of protein features, domain analysis, subcellular localization, and homology-based modeling of OsPUSs

For sequence identity and similarity comparisons, SIAS tool (http://imed.med.ucm.es/Tools/sias.html) was employed. Multiple sequence alignment of PUS proteins was performed using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). The conserved domains of OsPUSs were determined by SMART-EMBL (http://smart.embl-heidelberg.de/) and NCBI-CDD. The schematic diagram was then visualized using TB Tools (v1.115). Subcellular localization of the OsPUSs was predicted using Plant-mPLoc (https://www.csbio.sjtu.edu.cn/bioinf/plant-multi/).

The homology-based modeling of representative members of the 6 PUS subfamilies was performed by submitting their amino acid sequences as the query in the SWISS-MODEL homology-modeling server (https://swissmodel.expasy.org/). The model with the highest QMEAN score was selected and the template amino acid sequence was aligned with the query protein sequence using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). Finally, the homology-based modeling outputs were visualized and represented using UCSF Chimera v1.17.1 software (https://www.cgl.ucsf.edu/chimera/).

Identification of putative cis-regulatory elements (CREs) in the promoters of OsPUS genes

The 2000 bp upstream sequences from the translation start site of OsPUS genes were retrieved from the RGAP. The putative cis-acting regulatory elements in these sequences were predicted using PlantCARE web server (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) and a schematic representation of CREs in the promoters of OsPUS genes was generated using TB Tools (v1.115).

Identification of single nucleotide (SNPs) and amino acid (SAPs) polymorphisms in OsPUS genes and proteins

The TIGR IDs of OsPUS genes were used as query to fetch data for non-synonymous mutations amongst the available 3024 accessions in Rice SNP-Seek database (https://snp-seek.irri.org/_snp.zul) against the reference Nipponbare sequence. Non-synonymous SNPs that could particularly translate to cause any change in the protein sequence were identified. Domain-based localization of SAPs was done on the basis of secondary structure of OsPUS proteins.

In silico investigation of expression patterns of OsPUS genes in development and stress conditions

Genevestigator plant database (https://plants.genevestigator.com/local_plants/index.jsp) was utilized to compile the expression patterns of OsPUS genes in development and stress conditions. For development, Oryza sativa organism, mRNAseq data platform and only experiments with wildtype genetic background parameters were included and the resulting heat map was constructed. For stress conditions, accession numbers GSE14275, GSE41647, GSE6901, E-MEXP-3718, GSE25206, GSE53564, and E-MTAB-4427 were used to extract the expression data of OsPUS genes, and graphically represented using the gplots package v3.1.3 on the open-source softwares R v4.3.1 and R studio v2023.06.1. The details of stress treatments included to prepare expression profile atlas of OsPUSs under different abiotic stress conditions, including heat, drought, salinity, low temperature, chromium treatment, arsenic treatment, and phosphate starvation are mentioned in STable1.

Quantitative RT-PCR-based validation of expression profile of OsPUS genes during abiotic stress

Oryza sativa L. ssp. indica var. Pusa Basmati 1 was employed for expression analysis of OsPUS genes during abiotic stress. Seeds were surface-sterilized with 70% ethanol for 1 min followed by treatment with 2% sodium hypochlorite for 20 min, subsequently washed with water 5 times and soaked overnight. Seeds were then grown hydroponically in rice growth medium (RGM; Yoshida et al. 1976) at 28 ± 2 °C under 16 h light/8 h dark for 7 days. These seedlings were transferred to pots containing potting mix (Soilrite mix) and grown for an additional 3 days. Finally, 10-days-old potted seedlings were subjected to different abiotic stresses such as heat stress (42 °C for 4 h and 6 h) and drought stress (imposed by withholding water for up to 9 and 11 days). For salt stress, hydroponically-grown 10-days-old seedlings were treated with 150 mM NaCl for 4 h and 6 h. Untreated seedlings were kept as control sets for comparison. The tissues were harvested, snap-frozen in liquid nitrogen, and stored at -80ºC for RNA extraction. Total RNA was isolated from samples using TriZOL™ reagent (Thermo Fisher Scientific, USA). 1 μg of total RNA was reverse transcribed using the Verso cDNA synthesis kit as per manufacturer instructions (Thermo Fisher Scientific, USA). Gene expression was determined by Quantitative real-time PCR (qRT-PCR) using PowerUp SYBR Green Master Mix (Applied Biosystems, USA) and with gene-specific primers in CFX Connect Real-Time System (BioRad, USA). The rice eukaryotic initiation factor 1 alpha (eIF1α) gene was employed as an internal loading control. The relative expression was calculated using the 2^−ΔΔCT method (Livak and Schmittgen, 2001). Three independent biological replicates were analyzed with three technical replicates for each cDNA sample. The values were presented as mean ± SEM (Standard Error of the Mean). Student’s t-test was used to compare the significance between the means of different samples (p < 0.05 and p < 0.005). The sequences of primers employed in this study are provided in the Additional file 6.

Prediction of known miRNAs potentially targeting OsPUS genes and their expression analysis

miRNAs targeting OsPUS genes in rice were predicted using psRNATarget (https://www.zhaolab.org/psRNATarget/) against all the rice mature miRNAs as per miRBase Release 21 at default settings (Schema V2 2017 release) and the network was visualized using Cytoscape v3.9.1. Expression profiling of identified miRNAs was carried out using miRid as a query against available rice datasets at PmiRExAt (https://pmirexat.nabi.res.in/searchdb.html).

Generation of protein–protein interaction (PPI) network of OsPUSs

Protein–Protein interactors for 21 OsPUSs were identified using the PPI network generation feature in RiceNETDB (https://bis.zju.edu.cn/ricenetdb/). The obtained interaction network was then visualized using Cytoscape v3.9.1 and was functionally enriched with the STRINGEnrichment App in Cytoscape.

Results

Genome-wide identification, phylogenetic analysis, classification, chromosomal localization, synteny analysis, duplication pattern, and structural organization of PUS genes in rice

To identify the members of the rice PUS family, the BLASTp search was performed in the Rice Genome Annotation project (RGAP) using 20 Arabidopsis PUS family members as our search query. This led to the identification of 22 PUS members in rice. Our findings aligned with the results reported by Xie et al. (2022). For further classification into PUS subfamilies, we employed the NCBI-CDD tool. All the PUS members were manually screened for the presence of active site motifs characteristic of 6 PUS subfamilies. Our analysis revealed that LOC_Os05g43374 was truncated (186 amino acids in length) and lacked the active site PUS motif due to deletions in the core catalytic domain and was therefore eliminated. Another member (LOC_Os03g17920) possessed the PUS motif with a single modification (HRLD to HRLG) in their active site sequence. It was retained as a PUS family member since some PUS enzymes also possess RNA chaperoning activity in addition to pseudouridylation activity (Kurimoto et al. 2020). In conclusion, the rice genome encodes for at least 21 true PUS family members. The nomenclature of these 21 OsPUS genes was performed based on the subfamily names and the phylogenetic relationships among these genes (Fig. 1a). Multiple splice variants were identified for several members as indicated in Table 1, however, we considered only one variant which exhibited the longest coding and corresponding protein sequence for further analyses (Additional files 1, and 2).

Fig. 1.

Phylogenetic analysis, gene and protein architecture of OsPUS genes. a The phylogenetic tree was constructed using the neighbor-joining method and was clustered into paralogous clades (I to XIII) on the basis of ≥ 50% bootstrap value. The gene subfamilies were represented by distinct colored boxes. b The gene architectures displayed the distribution of exons (yellow boxes), introns (lines), and untranslated regions (UTRs, blue boxes). The scale at the bottom can be used to deduce approximate lengths of different gene elements. c The protein architecture of OsPUSs shows distribution of various domains represented as colored boxes, such as S4 (yellow), DKCLD (pink), PUS (blue), and PUA (red) domains (color figure online)

Table 1.

Details of the 21 PUS genes and their predicted proteins in rice

Gene name	TIGR locus ID	CDS length (bp)	Predicted protein
			Size (amino acids)	MW (kDa)	pI	Position of aspartate	Catalytic motif sequence	Subcellular localization
OsTruD1	LOC_Os01g56620.1	2160	720	79.66	7.24	D272	GTKD	Chl, Mt
OsRluA1	LOC_Os02g30840.1	1182	394	44.06	6.66	D162	HRLD	Cy
	LOC_Os02g30840.2	939	313	34.48	6.18	D81		Chl, Nc, CM
OsTruA2	LOC_Os02g44810.1	1746	582	63.77	5.06	D108	ARTD	Cy
OsRsuA1	LOC_Os03g05806.1	1248	416	45.11	10.25	D285	GRLD	Chl
OsRluA4	LOC_Os03g17920.1	945	315	34.82	9.14	G195	HRLG	Chl, Mt
	LOC_Os03g17920.2	894	298	32.81	6.31	G178		–
	LOC_Os03g17920.3	402	134	15.13	9.35	G89		Chl, Nc
OsTruA3	LOC_Os03g21980.1	1551	517	57.7	8.12	D134	SRTD	Chl, Mt
	LOC_Os03g21980.2	1125	375	42.28	5.7	–	–	Cy
OsTruB1	LOC_Os03g25450.1	1776	592	64.81	9.85	D143	GTLD	Nc
OsTruA1	LOC_Os03g26440.1	1236	412	46.15	8.39	D148	GRTD	Chl, Cy
OsPUS10	LOC_Os04g47380.1	1143	381	43.64	8.56	D197	GRED	Cy
OsTruB3	LOC_Os04g48700.1	867	289	32.46	10.79	D120	GTLD	Cy
OsTruB5	LOC_Os05g01750.1	1467	489	55.15	9.57	D305	GTLD	Chl
	LOC_Os05g01750.2	1212	404	45.63	9.87	D305		Chl
OsTruA5	LOC_Os05g19954.1	1029	343	37.95	9.6	D114	GRTD	Chl
	LOC_Os05g19954.2	672	224	25.39	9.83	–	–	Chl
	LOC_Os05g19954.3	666	222	23.91	9.16	D114	GRTD	Chl
	LOC_Os05g19954.4	618	206	23.36	10.08	–	–	Chl, Nc, CM
OsTruA4	LOC_Os06g45250.1	1278	426	46.94	7.5	D192	GRTD	Chl, Mt
OsRluA5	LOC_Os06g50360.1	1338	446	47.84	8.5	D235	HRLD	Chl
	LOC_Os06g50360.2	945	315	33.5	8.52	D235		Chl
OsTruB2	LOC_Os07g44190.1	1803	601	65.87	9.7	D139	GTLD	Nc
OsRluA2	LOC_Os07g46600.1	1203	401	44.48	8.69	D181	HQID	Chl
OsRluA6	LOC_Os08g40860.1	1428	476	53.32	10.23	D230	HRLD	Mt
OsRluA7	LOC_Os09g01620.1	1398	466	51.08	10.15	D228	HRLD	Chl
	LOC_Os09g01620.2	1347	449	49.03	10.27	D228		Chl
OsTruB4	LOC_Os11g38600.1	429	143	15.49	12.19	D79	GTLD	Cy
OsTruA6	LOC_Os12g09290.1	1176	392	42.59	7.19	D121	SRTD	Chl, Cy
OsRluA3	LOC_Os12g37380.1	1038	346	37.22	8.36	D134	NRLD	Chl

Gene IDs of the 21 PUSs were retrieved from RGAP (Rice Genome Annotation Project) and TIGR (The Institute for Genomic Research) databases. Information on splice variants, CDS (coding sequence) length, predicted protein amino acid size, molecular weight, and pIs of rice PUS genes was retrieved from RGAP. The active site consensus sequence and conserved aspartate (D) catalytic residue number were manually screened and its position was calculated. Absence of catalytic D and motif sequence is listed as ‘-’ whereas any change in the catalytic motif sequence is underlined. Subcellular localization of OsPUS proteins was predicted using Plant-mPLoc

MW molecular weight, kDA kiloDalton, bp base pairs, pI isoelectric point, Nc nucleus, Cy cytoplasm, CM cell membrane, Chl chloroplast, Mt mitochondria

An unrooted phylogenetic tree was constructed using the neighbor-joining method for OsPUSs, clustering proteins with a significant bootstrap value ≥ 50%. The tree grouped OsPUS members into 13 paralogous clades: one with 4 members, one with 3 members, three with 2 members each, and eight singletons. Notably, segmentally duplicated members, OsTruB1 and OsTruB2, displayed 100% bootstrap value, signifying high conservation in their protein sequences (Fig. 1a and SFig. 1, 2). We also categorized OsPUS members into six PUS subfamilies wherein we identified 6 members in OsTruA, 5 in OsTruB, 1 in OsTruD, 7 in OsRluA, 1 in OsRsuA, and 1 in OsPus10 family. OsRluA was found to be the largest subfamily, followed by OsTruA and OsTruB. Table 1 provides detailed information about nucleotides and predicted protein sequences of rice PUS family members. The gene structure of OsPUSs was analyzed using Gene Structure Display Server (GSDS v2.0) and it was found that OsPUS genes varied greatly in length, ranging from 429 bp (OsTruB4) to 12 kb (OsTruA6). The intron numbers spanned from 0 to 18 among OsPUS members, with an average of 5.80 introns per PUS gene. Remarkably, OsTruB1, OsTruB2, and OsTruB4 were intronless, while OsTruD1 contained the highest intron count (18). OsRluA3 and OsTruA2 contained one intron each, with the latter having a smaller intron. Intron lengths ranged from 66 bp in OsTruA5 to 6465 bp in OsTruA6, highlighting structural diversity within the rice PUS family (Fig. 1b).

In order to gain insights into the genomic organization of OsPUS genes, their locations were mapped onto the rice chromosomes. Interestingly, these genes were not uniformly distributed across the 12 rice chromosomes. The maximum number of OsPUS genes were found on chromosome 3 (5 members; OsRsuA1, OsRluA4, OsTruA3, OsTruB1, and OsTruA1), while chromosome 10 contained none. Chromosome 2 (OsRluA1 and OsTruA2), 4 (OsPUS10 and OsTruB3), 5 (OsTruB5 and OsTruA5), 6 (OsTruA4 and OsRluA5), 7 (OsTruB2 and OsRluA2), and 12 (OsTruA6 and OsRluA3) each housed only two PUS genes. On the other hand, chromosome 1 (OsTruD1), 8 (OsRluA6), 9 (OsRluA7), and 11 (OsTruB4), harbored only a single PUS member (Fig. 2a). Additionally, gene duplication and synteny analysis revealed a single segmental duplication event between two OsPUS genes, OsTruB1 and OsTruB2, situated on chromosomes 3 and 7, respectively. Notably, no tandem duplication events among OsPUS genes were detected in the rice genome (Fig. 2b).

Fig. 2.

Chromosomal distribution, duplication, and synteny analysis of OsPUS genes. a Distribution of OsPUS genes on 12 rice chromosomes is illustrated wherein the chromosome numbers are indicated on the top and the gene names are annotated on the right side. b Gene duplication and synteny analysis of OsPUSs in the rice genome. The segmentally duplicated gene pair (OsTruB1 and OsTruB2) is linked by the red line between chromosomes, while gray lines in the background represent collinear blocks within the rice genome. All the 12 chromosomes are represented to scale in accordance with their respective actual lengths

Architecture, subcellular localization, and homology modeling of predicted proteins of OsPUS family

To determine the diversification among OsPUS family members, we conducted structural analysis of their predicted proteins (Additional file 2). OsPUS protein lengths ranged from 143 (OsTruB4) to 720 (OsTruD1) amino acids, with molecular weights spanning from 15.49 kDa to 79.66 kDa, respectively. The predicted pI values varied from 5.06 (OsTruA2) to 12.19 (OsTruB4) (Table 1). Pairwise comparisons revealed an average similarity of 19.38% and an average identity of 10.01% among OsPUS proteins (SFig. 1), with high identity (84.77%) and similarity (88.15%) observed between OsTruB1 and OsTruB2 (SFig. 1). Multiple sequence alignment of catalytic domains of OsPUSs revealed conservation of motif II and aspartic acid residue in the OsPUS proteins (Table 1, SFig. 2). Domain analysis using the NCBI-CDD and EMBL-SMART tools indicated that all OsPUSs shared a common catalytic domain, with some variations based on their subfamily (Fig. 1c). For instance, OsTruA members contained XXXRTD motif in the catalytic domain without any additional domains. Out of five members, two members of the OsTruB family (OsTruB1 and OsTruB2) possessed XXGTLD, DKCLD, and PUA domains, while the other three lacked the N-terminus DKCLD and C-terminus PUA domains. (Table 1, Fig. 1c, SFig. 2). The members of the RluA subfamily predominantly had the XXHRLD motif, with a few variants such as XXHQID, XXNRLD, and XXHRLG in some members (Table 1). The S4 domain was present in three (OsRluA1, OsRluA4, and OsRluA5) out of seven OsRluA members. OsRsuA1 contained both S4 domain and the canonical XXGRLD motif (Table 1, Fig. 1c, SFig. 2). Similarly, OsTruD1 and OsPUS10 had catalytic domains with XAGXKD and XXGRED motifs, respectively (Table 1, Fig. 1c). Detailed motif distribution and their amino acid sequences are presented in SFig. 2 and 3.

Localization of PUS proteins significantly impacts RNA substrate diversity, specificity, and their cellular mechanism (Niu and Liu 2023). We employed Plant-mPLoc to predict subcellular localizations of OsPUS proteins, revealing diverse compartments, including nucleus (OsTruB1 and OsTruB2), cytoplasm (OsTruB3, OsTruB4, OsTruA2, OsRluA1, and OsPUS10), chloroplast (OsRluA2, OsRluA3, OsRluA5, OsTruA5, and OsTruB5) and mitochondria (OsRluA6 and OsRluA7). Some OsPUS proteins exhibited dual localization, such as chloroplast and mitochondria (OsTruA3, OsTruA4, OsTruA6, OsTruD1, and OsRluA4) or chloroplast and cytoplasm (OsTruA1 and OsTruA6). It would be worthwhile to experimentally validate the subcellular compartmentalization of these OsPUS proteins in rice.

To understand the biochemical and structural properties of OsPUS proteins, homology-based modeling of a representative member of six subfamilies namely, OsTruA1, OsTruB5, OsTruD1, OsRluA1, OsRsuA1, and OsPUS10 was performed and we found that despite of very low amino acid identity, each OsPUS model shared high structural similarity to its homolog along with the conservation of catalytic domain and asp residue (Fig. 3, SFig. 3). In addition, accessory domains at N- or C-terminal, thumb or/and forefinger loops and 5 PUS signature motifs were also found to be conserved among these proteins. For instance, OsPUS10 adopted a crescent shape-like HsPUS10 (McCleverty et al. 2007) which has THUMP domain at N-terminal and thumb loop, forefinger loop, active site, and catalytic domain at C-terminal (Fig. 3f). In OsTruA1, thumb loop and a protruding forefinger loop, facing away from the active site was observed when aligned with HsPUS1 (Fig. 3a). We performed sequence alignment among TruA homologs and found lack of lysine and proline residues in Motif I of OsTruA1, which is a characteristic of the TruA family (SFig. 2, SFig. 3b). OsRluA1 exhibited similar structural features like EcRluD, consisting of a N-terminal S4 domain which is joined by a flexible linker to the catalytic domain (Fig. 4d). Moreover, except Motif IIIa, conservation of all other PUS motifs was observed in OsRluA1 (SFig. 3e). Similarly, sequence alignments and structural comparisons between OsRsuA1 and EcRsuA delineated that, like EcRsuA, the N-terminal region of OsRsuA1 adopts a fold-like S4 domain and comprises of a central catalytic domain with semi conservation of 5 PUS motifs (Fig. 4e, SFig. 3d). For the TruB family, OsTruB5 homology-based modeling with EcTruB1 exhibited presence of all 5 PUS motifs, thumb loop, and catalytic domain (Fig. 4b, SFig. 3a). The active site was located at the edge of the catalytic domain while no accessory domain, such as the PUA domain at C-terminal, was identified in the OsTruB5. Lastly, in the TruD family, we found two characteristic insertions between the catalytic domain towards the N- and C-terminal, namely R3H and Helix-turn-Helix, respectively, in OsTruD1, similar to other TruD homologs. There is no canonical forefinger or thumb loop identified in this family, but it is suggested that the C-terminal insertion in TruD members facilitates a similar function (Fig. 4c). Remarkably, TruD members have diverged from other PUS subfamilies and carry variations in the 5 PUS motifs (Kaya et al. 2004). All the variations of 5 PUS motifs were also found in OsTruD1 (SFig. 3c).

Fig. 3.

Homology-based modeling of representative members of different OsPUS subfamilies. The homology-based models of OsTruA1 (a), OsTruB5 (b), OsTruD1 (c), OsRluA1 (d), OsRsuA1 (e), and OsPUS10 (f) were generated with evolutionary-related template proteins such as HsPUS1, EcTruB1, HsPUS7, EcRluD, EcRsuA, and HsPUS10, respectively. In these models, the predicted RNA binding loops and accessory domains are depicted as: catalytic domains (yellow), putative active sites (red), forefinger loops (purple), thumb loops (green), insertion domain R3H (blue), helix-turn helix-domain (grey), S4-like domain (cyan), and thump domain (pink) (color figure online)

Fig. 4.

Distribution of cis-regulatory elements (CREs) in putative promoter regions of OsPUS genes. Different CREs identified in OsPUSs promoter sequences (2.0 kb) were represented with TB tools (v.1.115). The position of the potential CREs in the promoter sequences can be inferred from the scale at the bottom

Analysis of polymorphisms in gene and protein sequences (SNPs and SAPs) of OsPUS family members

With an aim to assess allelic variations in OsPUS members across 3024 rice accessions, Rice SNP-Seek database was analyzed, with Nipponbare as the reference genome. Among 15 OsPUS members, a total of 61 non-synonymous SNPs were identified. Interestingly, six genes (OsTruB4, OsTruB5, OsRsuA1, OsTruA2, OsTruA4, and OsTruA5), showed no SNPs, indicating high conservation among them (Table 2). We also analyzed the single amino acid polymorphisms (SAPs) in OsPUS protein sequences, all of which resulted in missense mutations, with no nonsense mutations detected (Table 2). Remarkably, all seven RluA members displayed SAPs, with OsRluA4 having the highest count. Multiple sequence alignments of proteins and analysis of domains revealed that critical amino acids in five PUS signature motifs were conserved in OsPUS proteins, suggesting their favorable selection during evolution (Table 2). Nevertheless, some SAPs were found in other regions of catalytic and accessory domains. Approximately 30% of the SAPs occurred in the catalytic domains of ten OsPUSs (OsTruD1, OsRluA1, OsRluA2, OsRluA6, OsRluA7, OsTruA1, OsTruA3, OsTruA6, OsTruB3, and OsPUS10) and around 13% were in the S4 accessory domain of the OsRluA (OsRluA1, OsRluA4, and OsRluA5). Most SAPs (67.7%) resulted in amino acid group substitutions (polar to non-polar and vice-versa; Table 2). While the precise significance of these variations is difficult to ascertain, these alterations may fine-tune the functionality of PUSs in rice.

Table 2.

Single nucleotide and single amino acid polymorphisms (SNPs and SAPs) in OsPUS genes and proteins, respectively

S. No	Locus ID	Protein name	SNP position on CDS	Original codon	New codon	SAP position on protein	Original AA	New AA	Protein domain exhibiting the polymorphism	Type of AA substitution
1	LOC_Os01g56620_1	OsTruD1	164	CTC	CGC	55	Leu	Arg	–	Non polar to polar basic
			796	GGG	AGG	266	Gly	Arg	Catalytic	Non polar to polar basic
			1112	GTT	GCT	371	Val	Ala	Catalytic	+
			1129	GAG	AAG	377	Glu	Lys	Catalytic	Polar acidic to polar basic
2	LOC_Os02g30840_1	OsRluA1	259	TAT	CAT	87	Tyr	His	S4	Polar to polar basic
			649	GCA	TCA	218	Ala	Ser	Catalytic	Non polar to polar
			979	GAA	AAA	327	Glu	Lys	–	Polar acidic to polar basic
3	LOC_Os02g30840_2	OsRluA1	16	TAT	CAT	6	Tyr	His	S4	Polar to polar basic
			406	GCA	TCA	136	Ala	Ser	Catalytic	Non polar to polar
			736	GAA	AAA	246	Glu	Lys	–	Polar acidic to polar basic
4	LOC_Os03g17920_1	OsRluA4	211	GAG	CAG	71	Glu	Gln	–	Polar acidic to polar
			292; 293	GAT	TGT	98	Asp	Cys	S4	Polar acidic to polar
			321	ATG	ATA	107	Met	Ile	S4	+
			863	ATA	AAA	298	Ile	Lys	–	Non polar to polar basic
5	LOC_Os03g17920_2	OsRluA4	7	GCC	CCC	3	Ala	Pro	–	+
			160	GAG	CAG	54	Glu	Gln	–	Polar acidic to polar
			241; 242	GAT	TGT	81	Asp	Cys	S4	Polar acidic to polar
			270	ATG	ATA	90	Met	Ile	S4	+
			812	ATA	AAA	272	Ile	Lys	–	Non polar to basic
6	LOC_Os03g21980_1	OsTruA3	17	GCG	GTG	6	Ala	Val	–	No change in group charge
			82	TTC	ATC	28	Phe	Ile	–	Aromatic to non polar
			178	TGG	GGG	60	Trp	Gly	Catalytic	Aromatic to non polar
			1502	GCG	GTG	501	Ala	Val	–	+
7	LOC_Os03g21980_2	OsTruA3	1076	GCG	GTG	359	Ala	Val	–	+
8	LOC_Os03g25450_1	OsTruB1	49	GGC	CGC	17	Gly	Arg	–	Non polar to basic polar
9	LOC_Os03g26440_1	OsTruA1	189	TTG	TTT	63	Leu	Phe	Catalytic	Non polar to aromatic
			275	CGA	CAA	92	Arg	Gln	Catalytic	Polar basic to polar
			1201	AGC	GGC	401	Ser	Gly	–	Polar to non polar
10	LOC_Os04g47380_1	OsPUS10	725	GTT	GCC	242	Val	Ala	Catalytic	+
11	LOC_Os04g48700_1	OsTruB3	161	GGT	GAT	54	Gly	Asp	–	Non polar to polar acidic
			239	CAT	CCT	80	His	Pro	–	Polar basic to non polar
			257	TAC	TTC	86	Tyr	Phe	–	+
			293	CGC	CCC	98	Arg	Pro	–	Polar basic to non polar
			508	ATG	GTG	170	Met	Val	Catalytic	+
12	LOC_Os06g50360_1	OsRluA5	377	TTG	TCG	126	Leu	Ser	S4	Non polar to polar
			1315	GGT	AGT	439	Gly	Ser	–	Non polar to polar
13	LOC_Os06g50360_2	OsRluA5	377	TTG	TCG	126	Leu	Ser	S4	Non polar to polar
14	LOC_Os07g44190_1	OsTruB2	20	GCC	GTC	7	Ala	Val	–	+
			45	GAG	GAT	15	Glu	Asp	–	+
			1477	GCG	ACG	493	Ala	Thr	–	Non polar to polar
			1588	GGA	AGA	530	Gly	Arg	–	Non polar to polar basic
15	LOC_Os07g46600_1	OsRluA2	469	GAG	AAG	157	Glu	Lys	Catalytic	Polar acidic to polar basic
			667	ACA	GCA	223	Thr	Ala	Catalytic	Polar to non polar
16	LOC_Os08g40860_1	OsRluA6	1037	ACA	ATA	346	Thr	Ile	Catalytic	Polar to non polar
17	LOC_Os09g01620_1	OsRluA7	65	AAG	AGG	22	Lys	Arg	–	+
			107	GTT	GGT	36	Val	Gly	–	+
			236	GAT	GGT	79	Asp	Gly	–	Polar acidic to non polar
			470	TCG	TTG	160	Ser	Leu	–	Polar to non polar
			913	ATT	GTT	305	Ile	Val	Catalytic	+
18	LOC_Os09g01620_2	OsRluA7	65	AAG	AGG	21	Lys	Arg	–	+
			107	GTT	GGT	35	Val	Gly	–	+
			236	GAT	GGT	79	Asp	Gly	–	Polar acidic to non polar
			470	TCG	TTG	158	Ser	Leu	–	Polar to non polar
			913	ATT	GTT	305	Ile	Val	Catalytic	+
			1295	CAG	CTG	432	Gln	Leu	–	Polar to non polar
19	LOC_Os12g09290_1	OsTruA6	139	GCC	ACC	47	Ala	Thr	–	Non polar to polar
			145	GAC	TAC	49	Asp	Tyr	–	Polar acidic to non polar
			925	GTG	ATG	259	Val	Met	Catalytic	+
20	LOC_Os12g37380_1	OsRluA3	280	GCC	TCC	94	Ala	Ser	–	Non polar to polar

Detailed genomic and protein sequence analysis for assessing genetic variations in OsPUSs across 3024 rice accessions was performed using the Rice SNP-Seek database against the reference Nipponbare sequence. Information about SNPs, codon change, SAPs, amino acid changes, polymorphism in PUS domains, and amino acid substitutions are listed

CDS coding sequence, AA amino acid, +  no change in amino acid substitution, – SAP is not present in any domain of the enzyme

Identification of putative cis-acting regulatory elements (CREs) in the promoter regions of OsPUS genes

Transcriptional regulation in gene expression primarily relies on CREs within the promoter. In this study, a 2.0 kb nucleotide sequence upstream of the individual PUS gene translational start site was analysed using PlantCARE tool, identifying 390 CREs categorized into four groups: stress-related (STRE, DRE, MBS, and W-box) elements (37.4%), hormone-related (ABRE, as-1, TATC-box, and P-box) elements (31.3%), light-responsive (GT1-motif, Sp1, G-box, GATA-motif, TCT-motif, GAP-box, AE-box, I-box, GA-motif, TATC-motif, and GTGGC-motif) elements (25.4%), and development-related elements (CAT-box, CARE, HD-Zip 1, and GCN4_motif; 5.9%). It was found that 95 STREs were present in all OsPUS promoters, except OsPUS10, representing the largest proportion. Additionally, twenty-three DRE core elements were found in 11 OsPUS promoters, with OsTruB1 (5) and OsPUS10 (4) displaying the highest numbers. Fifteen MBS elements were detected in 12 OsPUS genes, with varying frequencies, and 13 W-box elements were identified in 7 OsPUSs, with OsTruA4 (4) possessing the most (Fig. 4, Additional file 3). These findings suggest the likely involvement of OsPUSs in stress responses, hormone regulation, photoperiod sensitivity, and plant development.

Transcription factors (TFs)-binding sites for hormone responses were found in all OsPUS promoters, ranging from 2 (OsTruB5) to 15 (OsTruA3) sites. Notably, the as-1 or activation sequence-1 motif was abundant, except in OsRluA3, OsRluA4, OsTruB3, and OsPUS10 (Fig. 4, Additional file 3). Similarly, the Abscisic acid-responsive elements (ABREs; 64 in total) were prevalent, with OsPUS10 and OsTruA3 containing the most (9 and 8, respectively), while OsTruB5 lacked ABREs. TATC-box and P-box CREs, responsive to gibberellic acid (GA), were also identified in some OsPUS promoter regions. For instance, OsRluA4 and OsRluA6 contained both TATC-box and P-box, whereas OsTruA2, OsRluA2, OsTruB2, and OsTruB3 contained one TATC-box each and OsTruB5 harbored one P-box element. Additionally, light-responsive elements (LREs) and tissue-specific elements were found, including GA motif, GT1 motif, ATC motif, Sp1, G-box, GATA motif, TCT motif, GAP-box, AE-box, I-box, and GTGGC motif. Tissue-specific elements, like 18 CAT-boxes, were associated with meristem expression, with their maximum number identified in both OsTruB1 and OsRluA4. A homeo-domain leucine zipper (HD-Zip) TF-binding site associated with provascular cell fate and induced by wounding (Scarpella et al. 2000) was found in OsRsuA1, and two GCN4 motifs (endosperm-specific) were located in OsTruA3. CARE (CAACT) motifs (seed germination-specific and GA-responsive), were found in OsRluA3 and OsRluA5 (Fig. 4, Additional file 3). Collectively, these findings suggest that OsPUS genes are likely responsive to phytohormones and may play direct or indirect roles in regulating tissue-specific and stress-related responses in rice.

Expression analysis of OsPUS genes during development and abiotic stress

To explore the biological roles of OsPUS genes in rice development and stress responses, we created an expression atlas using Genevestigator data for 19 out of 21 genes, as expression data for OsTruB4 and OsRluA1 were unavailable. The expression patterns revealed varying expression across tissues and developmental stages (Fig. 5a, b). Notably, during the stem elongation stage the TruA gene family exhibited the highest expression, except OsTruA3. Several PUS genes (OsPUS10, OsRluA3, OsRluA4, OsRluA7, and OsTruB2) displayed high expression, with OsTruA6 and OsRluA6 being the highest (Fig. 5a). OsTruB2 and OsTruA6 had significantly high expression during germination and milking stages, respectively. OsRsuA1 showed downregulation during germination, flowering, milking, and doughing stages, but intermediate levels in other stages like seedling, tillering, stem elongation, booting, and heading stages were observed (Fig. 5a). OsTruD1 generally displayed low expression across most developmental stages, with intermediate expression observed at germination, stem elongation, and booting stage (Fig. 5a). Among different anatomical tissues, several OsPUS genes (OsTruA1, OsTruA2, OsTruA6, and OsRluA7) were highly expressed in the shoot apex, with OsTruA6 having the highest levels (Fig. 5b). OsTruA5 showed the highest expression in the coleoptile, while OsTruB2 was most highly expressed in roots. OsTruA6 was upregulated in inflorescence, panicle, pistil, and embryo, similar to OsTruB2, which showed slight upregulation in these tissues. OsPUS10 displayed the highest expression in the leaf collar (Fig. 5b). We conducted a detailed analysis of how OsPUS genes respond to various abiotic stresses and found that most OsPUS genes were either unresponsive or only slightly affected. For example, when subjected to stresses like salt, drought, cold, heavy metals, or phosphate starvation, many OsPUS genes did not show significant changes in their expression levels (Fig. 5c). Drought stress primarily led to downregulation, followed by cold stress, salt stress, and chromium exposure (Fig. 5c). Some OsPUS genes showed increased expression levels under heat, arsenic, or phosphate deficiency. For instance, OsTruD1 had its highest expression during phosphate starvation, while OsRluA5 and OsRsuA1 showed high expression levels on exposure to high temperatures. Arsenic treatment led to appreciable increases in the expression of OsTruB1, OsTruB2, OsTruB3, OsPUS10, and OsTruA2 (Fig. 5c). These findings underscore the diverse roles of OsPUS genes in rice development, encompassing different tissues and growth stages, as well as their involvement in responding to challenging environmental conditions. Details of experiments employed for preparing the expression heat map are presented in STable 1.

Fig. 5.

Expression atlas of OsPUS genes in different developmental stages, tissues, and various abiotic stress conditions. Expression analysis of OsPUS genes at different developmental stages (a), tissues, (b), and under various abiotic stresses (heat, drought, salt, cold, heavy metal treatment, and phosphate deprivation) (c). The color scale represents the percent expression potential (linear scale) for development stages and tissues, while log₂FC expression level for various stress conditions (color figure online)

To further confirm the responsiveness of OsPUS genes and elucidate their expression patterns under various abiotic stresses (specifically heat, salt, and drought stress), we conducted quantitative RT-PCR analysis. Out of the 21 OsPUS genes examined, we were unable to detect any signal for OsTruB4 in qRT-PCR analysis, consequently, it was excluded from further investigation. Our analysis revealed that 20 OsPUS genes exhibited differential responsiveness, characterized in comparison to control conditions (Fig. 6a–c). During heat stress experiments at 42 °C, several OsPUS members, including OsRluA3, OsRluA5, OsRluA6, and OsTruA5 consistently exhibited upregulation at both the 4 h and 6 h time points (Fig. 6a). OsTruB3 and OsTruB5 showed increased expression levels at 4 h, while OsRluA4 displayed a significant increase in expression at 6 h only (Fig. 6a). Under prolonged drought stress, imposed by withholding watering for 9 d and 11 d, OsTruB3 exhibited the highest expression levels, while OsTruB2 demonstrated upregulation at 9 d time-point only (Fig. 6b). In response to high salt concentration (150 mM NaCl), OsRluA4 showed the highest upregulation at both time-points (4 h and 6 h), whereas the expression levels of OsTruB3 were elevated at the 6 h time-point (Fig. 6c).

Fig. 6.

Validation of expression profile of OsPUS genes in response to different abiotic stresses. The expression profiles of 20 OsPUS genes were validated using quantitative PCR in 10-days-old rice seedlings exposed to heat (42 °C for 4 h or 6 h) (a), drought (watering withheld for 9 or 11 days) (b), and salt stress (150 mM NaCl for 4 h or 6 h) (c). qRT-PCR was performed and the expression level in terms of normalized fold change was calculated relative to that in control seedlings by employing the 2^−ΔΔCT method, and the expression of each gene under control conditions was set as 1. Rice eukaryotic Initiation Factor 1 (eIF1A) gene was used as internal control. Values represent the mean ± SEM of three independent biological replicates. Error bars represent the standard errors from three independent replicates. Expression differences in OsPUSs following abiotic stress treatments was assessed by the Student’s t-test (*p < 0.05, **p < 0.005)

Identification of miRNAs targeting OsPUS genes in rice

To uncover the possible miRNAs-mediated regulation of OsPUS gene expression, an in silico analysis utilizing the psRNATarget server was conducted and miRNAs were identified which could potentially regulate the members of OsPUS family (Fig. 7a, Additional File 4). Out of 21 OsPUS genes, 16 were found to be targeted by known plant miRNAs belonging to 36 distinct families, resulting in 14 clusters. The number of miRNAs targeting OsPUSs ranged from 1 (OsRsuA1) to 34 (OsRluA5) (Fig. 7a, Additional File 4). Since miRNAs act based on complementarity, one miRNA can target multiple mRNAs, and multiple miRNAs may act on one mRNA (Ying et al. 2008). Additionally, different members of the same MIR family can target distinct mRNAs (Brennecke et al. 2005). For example, MIR2118 targeted OsTruB1, OsTruD1, and OsRluA1, while MIR444 were predicted to target OsRluA1 and OsTruA6. Similarly, MIR11339 could target OsRluA5 and OsTruA6, and three members of MIR1846 targeted both OsTruA2 and OsRluA3 (Fig. 7a, Additional File 4). One MIR5801 member targeted OsTruD1, while another targeted OsTruA6. Among the 14 clusters, 8 were independent, while two clusters targeting OsRluA5 and OsTruA6 were interconnected through MIR11339 members (Fig. 7a). Notably, 6 clusters with targets OsTruA4, OsRluA2, OsTruA3, OsRluA3, OsRluA6, and OsRluA4 showed extensive interconnections, sharing several miRNAs that could target multiple PUS members from different families. For instance, osa-miR1846f targeted OsTruA4, OsRluA3, and OsTruA2, while osa-miR2873c was found to target OsTruA2 and OsRluA4. Similarly, osa-miR5143 could target both OsRluA4, OsRluA6, and OsTruA3 (Fig. 7a). Most identified miRNAs were predicted to regulate the expression of PUS family members through mRNA cleavage (Additional File 4).

Fig. 7.

Identification of potential miRNAs targeting OsPUS genes. a Schematic representation of targeted regulatory relations between miRNAs and their target PUS genes in rice. Black lines represent the interaction, while the blue and yellow boxes represent the miRNAs and their target OsPUS genes, respectively. b Expression profile of putative miRNAs targeting OsPUSs. The color key represents expression values in log TPM (transcript per million). miRNA expression was determined in various plant tissues and abiotic stress conditions and the OsPUS gene target for corresponding miRNA has been depicted with coloured dots (color figure online)

To uncover the potential functions of PUS family members, we analyzed miRNA expression profiles in rice datasets from PmiRExAt. Notably, most miRNAs showed consistent downregulation during development and under various stress conditions (Fig. 7b). Osa-miR397a, targeting OsRluA7, displayed highest expression in roots and shoots, with appreciable upregulation in drought-stressed seedlings. Another miRNA, Osa-miR408-5p, which targets OsRluA6, displayed intermediate expression levels across roots, shoots, and leaf flowering stages. Osa-miR1871, targeting OsRluA5, exhibited a slight increase in shoots, while Osa-miR1866-5p, predicted to target OsTruA6, was upregulated in endosperm. Intriguingly, Osa-miR820a specifically targeted OsTruD1 and it showed significant levels in leaf flowering stage, endosperm, and drought stress (Fig. 7b). Moreover, miRNAs like Osa-miR160a-5p and Osa-miR1871, which target RluA and TruA members, displayed mild induction under drought stress (Fig. 7b). These findings suggest the involvement of multiple PUS members in shaping stress responses and development in rice.

Construction of protein–protein interaction network for OsPUS proteins in rice

Studying protein–protein interactions (PPIs) is of paramount importance for gaining insights into cellular protein functions. We constructed a PPI network using the RiceNETDB database to elucidate the interactions involving OsPUSs. We found interacting partners for 10 OsPUSs (OsTruB1, OsTruB2, OsTruB3, OsTruD1, OsRluA1, OsRluA5, OsRsuA1, OsTruA1, OsTruA2, and OsTruA6), resulting in 577 interactions among 450 proteins. OsTruB1 had the most interactions, while OsTruB2 and OsTruA6 had only one interacting partner each (Fig. 8, Additional file 5). Some PUSs, such as OsTruB1, OsRluA5, OsRsuA1, OsTruD1, and OsTruA2, could interact with themselves, indicating their ability to form homomers. We also observed a few interactions between members of different PUS families, such as OsRluA1 with OsRsuA1 and OsTruD1 with OsTruB1 (Fig. 8, Additional file 5). The PPI network exhibited a modular structure, comprising 5 major clusters, with nodes predominantly represented by OsTruB1, OsTruA2, OsTruD1, OsTruA1, OsRluA1 and OsRsuA1, and two minor clusters containing OsRluA5 and OsTruB3. Interestingly, we noted several interactions bridging these clusters, but OsTruA6 interacted with only one protein and did not display any connections with other clusters (Fig. 8). Furthermore, some partners were found to interact with more than one OsPUS member, serving as connectors between various clusters. For example, LOC_Os05g42300 (NAF1 domain containing protein) interacted with both OsTruB1 and OsTruB2, while LOC_Os10g32550 (HSP60-3B) exhibited interactions with OsTruB1 and OsRluA5 (Fig. 8, Additional file 5).

Fig. 8.

Protein–protein interaction (PPI) network analysis of rice PUS family members. The PPI network of OsPUSs was generated using the STRINGEnrichment App. The large-sized black nodes depict the OsPUSs (labeled with upper case letters), and the small-sized colorful nodes depict the potential interacting partners. The details of the interacting partner proteins are presented in Additional file 5

The gene ontology (GO) analysis revealed that OsPUSs interactors were involved in critical cellular processes, including translation, ribosome biogenesis, assembly, chromatin organization, protein folding, and gene silencing (Additional file 5). Among these processes, the largest groups of OsPUS interactors were associated with cellular processes and gene expression, comprising 165 and 82 proteins, respectively. The interactors in Cluster 1 (corresponding to OsTruB1) were found to be associated with DNA coiling, RNA processing, protein synthesis, and trafficking along with numerous cellular catabolic and anabolic processes. In contrast, Cluster 5 (related to OsTruD1) consisted of proteins of heat shock response (HSP70), RNA export, heterochromatization, and protein components of ribosomes in both cytoplasm and chloroplast. Cluster 2 (OsRluA5) was primarily comprised of enzymatic proteins, like oxidoreductase, dikinase, thioesterase, glucose-6-phosphate dehydrogenase, and phosphokinase, while Cluster 3 (OsTruB3) consisted of proteins participating in ribosome biogenesis. Similarly, Cluster 4 (OsRluA1 and OsRsuA1) had proteins that are involved in enzymatic processes, such as transcription, and ribosome biogenesis. Notably, Cluster 6 (OsTruA1) encompassed diverse pathways, including aromatic amino acid biosynthesis, reactive oxygen species (ROS) signaling, DNA coiling and replication, transcription, organelle (ribosome and peroxisome), chaperones, organelle and protein trafficking, ubiquitination, cell division (CDKs), and ubiquitinylation. Cluster 7 (OsTruA2) contained ribosomal proteins and a few proteins involved in translation. In summary, the PPI analysis strongly supports the potential regulatory role of OsPUSs in diverse cellular processes and stress response pathways.

Discussion

With an aim to identify and characterize the Pseudouridine synthases family members in rice we conducted a comprehensive analysis and identified 21 members grouped into six distinct subfamilies. Notably, PUS subfamily sizes vary across species (Hamma and Ferré-DꞌAmaré 2006; Penzo et al. 2017), but in plants like rice, they are relatively similar (Xie et al. 2022). For instance, Escherichia coli has four members in both RsuA and RluA subfamilies (Hamma and Ferré-DꞌAmaré 2006; Xie et al. 2022), while in humans there are four members in the RluA, but RsuA subfamily is entirely absent (Penzo et al. 2017). In plants, the RsuA subfamily has undergone a reduction, whereas the RluA and TruA subfamilies have experienced expansion (Xie et al. 2022). PUS10 is absent in E. coli and yeast, but has one copy in archaea, plants, and humans, possibly due to gene dosage effects (McCleverty et al. 2007; Fitzek et al. 2018). Our study aligns with findings in other plant species, revealing RluA and TruA as the largest subfamilies in rice, while one member each in RsuA and PUS10 was retained in rice.

Following the identification of OsPUS members, comprehensive investigations were conducted to analyze gene structure, phylogenetic relationships, chromosomal distribution, and gene duplication events. It was concluded that rice PUS members within each subfamily exhibited similarities as well as differences in their characteristics when compared to PUS genes in other plant species. Gene architecture studies revealed that the average number of introns per PUS gene was 5.80, which was notably lower than that in Arabidopsis (7.65 introns per gene) and maize (8.00 per gene). The CBF5 orthologs in various plant species, such as AtTRUB1 in Arabidopsis and ZmTRUB1A, ZmTRUB1B, and ZmTRUB1C in maize were largely found to be intronless (Xie et al. 2022). Similarly, in rice, OsTruB1 and OsTruB2, which belong to the CBF5/TruB type of PUSs, exhibited intronless organization. In contrast, TruD orthologs displayed the highest number of introns among all plant species (Xie et al. 2022), with OsTruD1 containing the maximum number (18) of introns among all OsPUS genes in rice. It is worth emphasizing that variations in intron lengths among gene family members contribute significantly to functional diversification (He et al. 2012; Mani et al. 2015). Moreover, it was observed that introns in ZmPUS genes were longer than those in AtPUS genes (Xie et al. 2022). Remarkably, OsPUS genes were found to have significantly longer introns compared to that in maize. It was noted that OsTruA6 possessed the longest intron length in rice, whereas ZmRluA6 carried the largest intron in maize, suggesting that members of the PUS family exhibit substantial structural diversity within both monocots and dicots, which highlights the importance of intron length variations in shaping the functional divergence of these genes. Furthermore, the 21 OsPUS loci were mapped onto 12 rice chromosomes, revealing an uneven distribution of OsPUS genes. Duplication plays a crucial role in expanding gene families, thereby resulting in functional diversity and complexity (Cannon et al. 2004). Gene synteny analysis in maize (ZmTRUB1A and ZmTRUB1B) and Arabidopsis (AtTRUA1A and AtTRUA1B) has unveiled a single segmental duplication event for PUS genes in each of these species, with no instances of tandem duplication event (Xie et al. 2022). Interestingly, this study identified no tandem duplications, and only one segmental gene pair, OsTruB1 and OsTruB2, which exhibited a high degree of sequence similarity. Further exploration into the conservation of sequence and structural features, in conjunction with the spatio-temporal expression patterns of these duplicated gene pairs, would provide intriguing insights into their functional evolution.

Extensive analyses of OsPUS proteins revealed structural similarity to PUS enzymes found across diverse organisms, with minor variations by subfamily. True PUS protein feature a catalytic domain with an active site motif II and an aspartic acid residue essential for Ψ synthase activity. In the rice PUS family, despite low sequence similarity and identity, catalytic domains resemble canonical PUS proteins with active site consensus motif II. OsTruA subfamily members contain XXXRTD, OsTruB members possess XXGTLD, OsTruD1 carries XAGXKD, OsRsuA1 harbors XXGRLD, and OsPUS10 has XXGRED consensus signature motifs. OsRluA subfamily predominantly carries the XXHRLD motif, but variants such as XXHQID, XXNRLD, and XXHRLG are present in rice. OsRluA2 and OsRluA3 contain the XXHQID and XXNRLD motif, respectively, while OsRluA4 carries XXHRLG. As XXHQID and XXNRLD variants are absent in algae but widely present in mosses, ferns, and spermatophytes, it is suggested that these variants might have diverged post-vascular plant evolution and may share conserved functions (Xie et al. 2022). Arginine (R) residue, before the catalytic D, is conserved in TruA, RluA, and RsuA PUS subfamilies and is crucial for substrate base flipping and stabilization. XXHQID and XXNRLD variants might undertake similar mechanisms for Ψ synthesis (Xie et al. 2022). Intriguingly, XXHRLG motif, with catalytic D replaced by glycine (G), is universally conserved but its functional significance in the RluA subfamily needs investigation. Mutation of D to a non-polar amino acid alanine (A) results in loss of isomerisation ability in Ψ writers of yeast (Purchal et al. 2022). PUS enzymes have roles beyond isomerization as a catalytically inactive AtRsuA1 was able to complement the leaf variegation phenotype of svr1 mutant (Yu et al. 2008) in Arabidopsis, while in algae, trans-splicing of chloroplast psaA RNA was not substantially impacted by loss of Ψ activity in Maa2 (Perron et al. 1999). In human, TRUB1 promoted let-7 miRNA maturation and OsPUS10 processed miRNA biogenesis, independent of their Ψ activity (Kurimoto et al. 2020; Song et al. 2020). All these studies established that PUS enzymes have biological roles beyond their isomerisation ability and XXHRLG type of RluA members might have other roles than Ψ writers or might function by association with other catalytically active PUSs. Indeed, our protein–protein interaction network analysis predicted that RluA members might interact among themselves or with RsuA PUSs. Furthermore, apart from motif II, 4 other motifs (motif I, IIa, III, and IIIa) are identified in PUS proteins which are involved in Ψ synthesis (McCleverty et al. 2007). All 6 PUS subfamilies carry these 5 PUS motifs wherein TruD members have some variations in the 5 motifs (Kaya et al. 2004). Multiple sequence alignment and sequence comparisons revealed conservation of all these 5 motifs in OsPUS proteins. Interestingly, our analysis of polymorphism in gene and protein sequences unveils no SNPs and SAPs in the 5 signature motifs indicating the favorable selection of these amino acids during evolution.

Structurally, PUS proteins, including OsPUSs, feature 5 motifs and various RNA-binding loops like forefinger and thumb loops, along with accessory domains such as THUMP and S4 at N-terminal and PUA at the C-terminal for substrate recognition and binding (Hamma and Ferré-D'Amaré 2006). Homology modeling of six representative members show structural similarities with their homologs. TruA members lack auxiliary domains, but include forefinger and thumb loops attached to the catalytic domain (Foster et al. 2000). OsTruA1 lacks auxiliary domains, but possesses RNA binding loops similar to EcTruA1. Only two OsTruB subfamily members (OsTruB1 and OsTruB2) have a C-terminal PUA domain and a characteristic DKCLD motif at the N-terminal. In AtTruB1 and ZmTruB1A, 1B and, 1C, both DKCLD and PUA domains are present, similar to bacterial and yeast homologs. Homology modeling of AtPUS10 and OsPUS10 with HsPUS10 recognises the catalytic domain, active site, forefinger, thumb loop, and THUMP domain. Multiple sequence alignment shows conservation in the forefinger loop, indicating similar binding, catalytic mechanisms, and RNA targets among PUS10 homologs (Niu and Liu 2023). In bacterial RsuA and RluA subfamilies, an N-terminal S4 domain recognises RNA substrates and in rice, S4-like domains are identified in OsRsuA1 and 3 members of OsRluA subfamily (OsRluA1, OsRluA4 and OsRluA5). OsRluA subfamily exhibits the highest number of polymorphisms, primarily in OsRluA4, with 13% SAPs identified in S4 domains, potentially impacting plant-specific substrate recognition. In eukaryotes, TruD homologs feature variations in PUS signature motifs and large insertion domains within the catalytic domain, facilitating pseudouridylation of diverse RNA substrates. Modeling studies of OsTruD1 reveal similar structural features, suggesting comparable substrate specificities among TruD homologs. In conclusion, while PUSs differ in amino acid sequences, their conserved structures imply similar mechanisms and RNA substrate recognition.

The precise subcellular localization of a protein is a critical determinant that delineates its fundamental biological function. PUS enzymes exhibit a diverse range of localizations within subcellular compartments. Studies have established that a single PUS enzyme can be found in multiple locations within a cell under normal conditions (Rintala-Dempsey and Kothe 2017). Furthermore, their localizations and activity can be influenced by external stimuli (Schwartz et al. 2014). In the context of plants, PUS enzymes are reported to be found in the nucleus (AtTruD1, AtPUS10, AtTruB1/NAP57 and their maize orthologs ZmTruD1, ZmPUS10, ZmTruB1A), mitochondria (AtRluA1/FCS1), and chloroplast (AtRsuA1/SVR1, AtRluA4, and OsPUS1/TCD3; Lin et al. 2020; Niu et al. 2022; Xie et al. 2022). A subset of plant PUSs is also observed to localize in both the nucleus and cytoplasm (AtTruA5, ZmRluA4, and ZmTruA5). However, in rice, the localization of PUS enzymes, except for OsPUS1/TCD3, confirmed to reside in the chloroplast, is yet to be determined. Predictive models for the subcellular localization of OsPUSs suggest that these proteins may localize to various cellular compartments, including the nucleus, cytoplasm, and cell organelles. To gain a more comprehensive understanding of the functions of these PUS enzymes, it would be beneficial to employ transient gene expression techniques to experimentally confirm their subcellular localization.

Recent findings have increasingly highlighted the impact of PUSs on plant stress and development responses. Our research, including promoter analysis, miRNA predictions, and expression analysis, has provided compelling evidence for rice PUSs role in shaping plants’ response to both internal and external cues. Most OsPUS promoters contain stress- and hormone-related CREs, including DRE (dehydration responsive element), MBS (Myb-binding site), and STRE (stress-response element) which have been well-documented for their responsiveness to drought, salinity, and heat. Notably, the promoters of OsTruA4, OsTruA5, and OsTruB2 harbor the highest number of STREs and show significant modulation in response to various abiotic stresses. Another prevalent CRE in OsPUS promoters is ABRE which is involved in ABA-mediated regulatory network and abiotic stress signaling such as osmotic, salinity, and drought stresses (Kim et al. 2011; Wang et al. 2018). Additionally, as-1 promoter element, known for its responsiveness to phytohormones like auxin, salicylic acid (SA), and methyl jasmonate (MeJA; Xiang et al. 1996) as well as its function in oxidative stress, is also present (Garretón et al. 2002). OsPUS10 and OsTruA3 have the highest number of ABREs, while OsTruA2 and OsTruA3 carry the most as-1 elements. In OsRsuA1, both as-1 and ABREs are present. Mutations in its homolog have been reported to alter ABA responses in Arabidopsis (Dong et al. 2020).

Exploring the expression and regulation of a gene is a fundamental approach to elucidating the significance of the gene within a biological system. We examined the OsPUS gene expression profile and predicted their potential miRNAs regulators. Genevestigator-based expression analysis revealed differential OsPUS expression across development stages, tissues, and abiotic stresses. We further validated these expression patterns under three abiotic stresses (heat, salinity and drought) using qRT-PCR assay. Heat stress induces the maximum number (8) of OsPUS genes, while drought and salinity upregulate only two genes each. In contrast, fewer AtPUS genes (AtTruA1A, AtTruA2, AtTruA3, and AtRluA6) and only ZmTruB1b are significantly upregulated upon heat stress treatment in Arabidopsis and maize, respectively (Xie et al. 2022). Maximum number of OsPUS genes exhibit downregulation in salt stress wherein the members of OsTruA family are most prominent, while only OsRluA4 and OsTruB3 are significantly upregulated. In contrast, higher numbers of PUS genes display upregulation during salt stress in Arabidopsis and maize (Xie et al. 2022). Notably, OsTruB3 is significantly upregulated under all three stresses, making it an intriguing candidate for pseudouridylation research in rice. We further predicted miRNAs belonging to 36 families and found that the majority of these miRNAs are linked to development and stress response. OsTruB1 and OsTruB2, despite high sequence similarity, exhibit distinct expression profiles. Notably, miR2118h and miR2118k targeting OsTruB1, show flower-specific expression (Mittal et al. 2013), and miR2118b–n deletions result in severe male sterility (Araki et al. 2020). OsTruB1 displays low expression in the reproductive tissues, while OsTruB2 is somewhat highly expressed in the same tissues. Although OsTruD1 levels remain unchanged during drought, its levels exhibit downregulation in salinity and heat stress. osa-miR820 which specifically targets OsTruD1, is highly induced under salt stress (Sharma et al. 2015). Members of the MIR160 family (osamiR-160a, b, c, d, e) specifically target OsTruA5, involved in plant immunity (Feng et al. 2022) and drought response (Ding et al. 2013). Remarkably, differential expression levels of members of osamiR-160 are observed in rice varieties exhibiting contrasting responses to drought stress (Cheah et al. 2015). Indeed, in our study, levels of OsTruA5 are significantly reduced under drought stress. It is noteworthy that only one miRNA, osa-miR2275d is predicted to target OsRsuA1, which is responsive to salt and drought stress. OsRsuA1-mediated chloroplastic rRNA pseudouridylation is essential for proper functioning of chloroplast under cold stress (Wang et al. 2022). Levels of osa-miR2275d decrease in rice under short heat duration (Mangrauthia et al. 2017) and its predicted target OsRsuA1 shows moderate induction under heat stress, indicating a probable link of OsRsuA1-mediated pseudouridylation in regulating temperature response in rice. For OsPUS10, six members of the MIR1882 family are predicted to target its transcript and this miRNA family is known to be responsive to Magnaporthe oryzae, a causal organism for rice blast fungus (Li et al. 2020). Moreover, miR166, miR169, miR396, and miR444 which target the members of OsRluA and OsTruA subfamily, miR408 targeting OsRluA6, and miR397 found to targeting OsRluA7, exhibit stress-responsive differential expression (Chandran et al. 2019; Huang et al. 2021; Gao et al. 2022; Feng et al. 2023; Nguyen et al. 2023). It would be worthwhile to elucidate the miRNA-PUS mediated regulation of stress responses in rice.

Protein–protein interactions studies are crucial for gaining a comprehensive understanding of how a particular protein functions and its association with biological pathways in plants. Our analysis for identification of putative protein interactors shows that these proteins are possibly involved in diverse cellular processes, including stress pathways. Notably, OsTruB1, which exhibits mild upregulation by heat stress, is predicted to interact with heat shock proteins, such as HSP17.4 and HSP18.1. It would be interesting to unravel the function of OsTruB1 in regulating heat stress responses in rice. Based on the results obtained it is plausible to perform comprehensive functional studies of members of the PUS family in rice and ascertain whether they are potential candidates for modulating stress responses in rice. Furthermore, understanding the role of pseudouridylation in regulating stress responses in rice is a vital area of research, ultimately benefiting both agriculture and global food security.

Conclusion

In this comprehensive study, we have significantly expanded the current knowledge of PUS enzymes in plants by identifying and characterizing 21 members within the PUS gene family of rice. Our analysis, which included the examination of core catalytic motifs and phylogenetic classification, grouped these 21 members into six distinct subfamilies, with RluA and TruA emerging as the most prominent subfamilies. Our exploration of genomic locations and gene architecture revealed structural diversity among OsPUSs. Intriguingly, gene synteny analysis uncovered only one segmentally duplicated pair (OsTruB1 and OsTruB2), which, despite high sequence similarities, displayed disparate spatio-temporal expression patterns. Detailed investigations into protein structures, domains, and motifs indicated that OsPUSs possess structural characteristics typical of true PUS proteins. However, variations in catalytic core motifs along with occurrence of highest number SAPs in OsRluA members, especially OsRluA4, hinted at structural and functional diversification in plant PUSs. The survey of OsPUS promoters and gene expression patterns under abiotic stress conditions provided compelling evidence for their role in stress adaptation, with OsTruB3 emerging as a key player. Additionally, we identified miRNAs belonging to 36 families that potentially target OsPUSs, which showed dynamic responsiveness to stress stimuli, forming an intricate regulatory web for modulating PUSs expression. Furthermore, through protein–protein interaction analysis, we identified interactors such as heat shock proteins, fortifying the pivotal role of OsPUSs in complex stress response pathways. In conclusion, our study represents a definitive contribution to the broader understanding of PUSs within the plant kingdom, offering profound insights into their structural diversity, protein architectures, and their roles in shaping stress responses in rice. Unraveling the role of pseudouridylation by studying PUSs opens new avenues for enhancing crop production and stress resilience, crucial in the context of global climate change.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

The entire study and experimental scheme was conceptualized, designed and supervised by SK-A. The computational analyses were performed by ML, NB, IK, YD, and SG. YD, IK and NB conducted stress treatments and qRT-PCR analysis. YD and SK-A wrote the manuscript. MA helped with the analysis of results, provided useful suggestions, and critically revised the manuscript. All authors read and approved the final manuscript.

Funding

The financial support from grant no. 38(1525)/23 funded by Council of Scientific and Industrial Research (CSIR), Government of India and Faculty Research Programme (Institute of Eminence), University of Delhi (Sanction order # IoE/2023-24/12/FRP) are acknowledged. YD, IK, and SG are thankful to CSIR and NB is grateful to University Grant Commission (UGC) for research fellowships.

Data availability

All data generated or analyzed during this study are included in the main article as well as in the supplementary and additional files provided.

Declarations

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ethical approval

This is an observational study. No ethical approval is required for this work.

Consent to participation

Informed consent was obtained from all individual participants included in this study.