Introduction

Rice (Oryza sativa L.) is one of the major cereal crops of the world. Despite focused efforts to improve major crops including rice for better yield through traditional breeding, success has been stagnant since last two decades (Siddiq et al. 2021). Fortunately, it is now possible to use biotechnological tools and transgenic approaches to improve yield in agriculturally important crops with far fewer target traits than had been anticipated (Zhang et al. 2001; Li et al. 2020a). Functional genomics, also known as reverse genetics approach, lays down a strong foundation for this transgenic outlook. With the completion of whole genome sequencing of rice, functional annotation of a large number of genes has helped significantly in taking a step forward in this field. Identification and functional characterization of transcription factors are of immense importance as they are master regulators for differential gene expression in myriad physiological processes essential for tissue development, growth and senescence of organs in eukaryotic organisms. Transcription factors (TFs) regulate a plethora of downstream genes and thereby often contribute directly or indirectly in the development of a desirable trait in a multi-pronged approach. Although genome sequences of several crop plants are available, only a small proportion of TFs have been functionally characterized till date. Hence, characterization of unknown TFs related to economically important traits poses an interesting field of study in the context of crop yield improvement.

The NAC designation is derived from No Apical Meristem (NAM), Arabidopsis transcription activation factor (ATAF) and cup-shaped cotyledon (CUC). The proteins belonging to NAC family are unique to plant kingdom, and they constitute one of the largest TF families. The NAC members have concurrently evolved with the transition of plants from aquatic-to-terrestrial habitat as well as from vegetative-to-reproductive growth (Xu et al. 2014; Liu et al. 2023). The TFs of NAC family have been shown to regulate a wide range of developmental processes, including cell division (Kim et al. 2006), shoot apical meristem formation (Kim et al. 2007), vegetative or floral organ development (Liu et al. 2023), fibre development (Ko et al. 2007), embryo development (Duval et al. 2002), seed development (Sperotto et al. 2009) and leaf senescence (Guo et al. 2005; Breeze et al. 2011). TFs in coordination with phytohormones, form a complex regulatory network that controls an array of downstream differential gene expressions. In both rice and millet, 151 genes of NAC TFs have been identified (Nuruzzaman et al. 2010; Dudhate et al. 2021). In barley, 167 genes are annotated to express NAC TFs (Murozuka et al. 2018). The large number of NAC TFs maintained in each genus clearly indicates the importance and diversified role played by them in various vital processes involved in plant life.

Root development along with the maintenance of suitable root architecture is a major event in plant life cycle. Root system is of great significance as it helps the sessile plant to anchor to the substrate, mainly soil, and is responsible for nutrient and water uptake and hence the vitality of the plant. Root system architecture (RSA) is determined by growth rate of primary as well as lateral roots and their angles. The branching of secondary roots ensures radial exploration of soil for water and nutrient absorption by the plant and thus contributes to plant health directly. RSA is highly adaptable to environmental stimuli and often plays first line of defense against various root-associated stresses such as salinity, drought, soil borne pathogens etc. The plant resists these adverse conditions in the root and thereby sustains itself (Seo et al. 2020). In rice, the root system comprises of the seminal root (also known as primary root), the adventitious (or crown) roots and root hairs (Rebouillat et al. 2009; Meng et al. 2019). Development of roots are known to be regulated by multiple phytohormones specially auxin, cytokinin, ethylene and abscisic acid (ABA) (Inukai et al. 2005; Zhang et al. 2018; Li et al. 2020b).

Economically significant agronomic traits of rice are often involved with quantitative trait loci (QTLs). Thus, controlling a transcription factor that can regulate a multitude of downstream targets to modify QTL-related traits is important from agricultural perspective. Tiller number and tiller height of cereal plants are such QTL-associated factors that determine grain yield. Tiller number is directly related to panicle number which in turn is correlated with grain yield (Liang et al. 2014; Wang et al. 2018). Moreover, tiller height is equally important as taller plants are more susceptible to lodging and thereby yield loss. Additionally, it is reported that often increase in tiller number is linked with dwarfism in tiller height in several mutants/ transgenic plants (Liao et al. 2019). However, an excessive increase in tiller number is detrimental for grain yield as often the tillers developed are unproductive. Thus, an optimization of tiller number is required for maximal yield of grains (Zhao et al. 2020).

Genes or QTLs control panicle morphology, which is another agronomic character that is directly related to yield. Panicle length, number of primary and secondary branches, number of grains per panicle, number of chaffy grains are characteristic features of a rice cultivar and they cumulatively determine the net grain yield at the end of cropping season (Sun et al. 2013; Parida et al. 2022). Apart from panicle morphology, grain morphology also contributes to yield as it directly effects grain size and 1000 grain weight. Grain filling is crucial as well from yield perspective. A number of QTLs involved in panicle and grain morphology in rice have been identified over the decades and their roles have been vividly documented by Parida et al. (2022). Studies indicate that often low auxin and cytokinin concentrations are responsible for poor grain filling in rice plants leading to yield loss.

Auxin and cytokinin are major phytohormones implicated in various developmental processes in the plant life cycle. Sometimes auxin and cytokinin play antagonistic roles in different cellular mechanisms. These phytohormones are also known to regulate biosynthesis of each other and consequently, altering tillering pattern in rice (Liu et al. 2011). Thus, auxin and cytokinin forms an integrated system regulating various aspects of plant architecture by modulating each other as well as downstream phytohormone-responsive genes. In root and shoot apical meristems, the auxin: cytokinin ratio dictates the destiny of the meristematic cells (Ruzicka et al. 2009; Su et al. 2011; Azizi et al. 2015). Apart from auxin and cytokinin, other phytohormones, like jasmonic acid (JA), ethylene, brassinosteroids and gibberellic acid (GA) are associated with panicle and grain morphology (Kang et al. 2019; Deveshwar et al. 2020; Lu et al. 2022). Often TFs are integrated in these hormone-regulated pathways to modulate the expression of different downstream genes and thereby regulate various plant processes (Kang et al. 2019).

In this work, we aimed to functionally characterize the OsNAC121 gene (MSU Locus ID: LOC_Os10g42130) of rice (Oryza sativa L.) through reverse genetics approach by developing transgenic rice lines with RNA interference (RNAi)-mediated silencing and overexpression (Ox) of the gene. We speculated that this rice gene could be the putative orthologue of Arabidopsis ANAC089. Through BLASTP analysis using ANAC089 protein sequence, we identified the rice protein sequence A2ZAF9 as the “reciprocal best hit” (having 59.9% similarity and 46.6% identity at the amino acid sequence level) from the indica rice genome. Following the standard and earlier nomenclature (Oryzabase, Fang et al. 2008), this rice gene encoding for the A2ZAF9 protein was designated as OsNAC121. We were curious to study the OsNAC121 gene through functional genomics as ANAC089 has been reported to be involved in diverse physiological processes. ANAC089 is a membrane-associated NAC TF that controls endoplasmic reticulum (ER) stress induced programmed cell death in Arabidopsis and downregulation of ANAC089 imparts ER stress tolerance in Arabidopsis (Yang et al. 2014). ANAC089 is known to negatively modulate floral initiation in Arabidopsis (Jian et al. 2010), and serves as redox dependent suppressor of stomatal ascorbate peroxidase gene expression (Klein et al. 2012). ANAC089 is reported to be regulating seed germination and stress response mediated through hormonal pathways (Albertos et al. 2021). Thus, we hypothesised that the rice gene OsNAC121 could be a potential candidate for functional studies that might be regulating various agronomically significant traits. To understand whether OsNAC121 is a true orthologue of ANAC089 and like other NAC TFs if it can control multiple traits or whether it coordinates with any phytohormone to regulate its downstream genes, we developed OsNAC121 RNAi and Ox lines to decipher the role of OsNAC121 in rice plants. In this work, through detailed analyses of the transgenic rice lines, we have documented the physiological function of OsNAC121 in regulating root development, tillering, panicle morphology, and grain filling in rice plant by articulating the auxin-cytokinin homeostasis. The study demystifies that OsNAC121 modulates phytohormone-mediated processes through determination of the fate of undifferentiated meristematic cells or tissues.

Materials and methods

Plant materials

For this study, indica rice cultivar IR64 was procured from Chinsurah Rice Research Station, Directorate of Agriculture, West Bengal, India. The untransformed control and transgenic lines were grown in a plant growth chamber (Environmental system, Labtech, Daihan) maintained at 16 h light/8 h dark period, 60% humidity and 750–840 lx light intensity. These plants were grown inside a net house for seed development and thereby perpetuation of the rice lines.

Phytohormone treatments

IR64 seeds were surface sterilized and germinated, and then grown aseptically in MS liquid medium (Duchefa Biochemie). Fourteen-day old rice seedlings were treated with different phytohormones (10 μm NAA, 10 μm BAP, 10 μm kinetin, 100 μm ABA, 100 μm JA, 100 μm SA) for different time period to study the gene expression of OsNAC121 in treated seedlings. To analyse the effect of phytohormones on OsNAC121 silencing (RNAi) and overexpression (Ox) in transgenic plants, surface sterilized seeds were germinated, and then grown aseptically for 14 days on the MS medium (Duchefa Biochemie) containing IAA (0, 5 and 10 µM) and BAP (0, 1 µM).

Isolation and cloning of OsNAC121 gene

Total RNA was extracted from IR64 rice shoots using the RNeasy plant mini kit (Qiagen). After DNase I enzyme (Sigma Aldrich) treatment, the isolated total RNA was quantified with a Nanodrop spectrophotometer (NanoDrop, 2000 C, Thermo Scientific). First-strand cDNA synthesis was carried out using 2 µg of RNA and random primers using cDNA synthesis kit (Applied Biosystems). The full-length coding DNA sequence (CDS) of the OsNAC121 gene was PCR amplified from this cDNA using the primer pair (Table S1), which was designed based on the available sequence in the rice genome annotation project (LOC_Os10g42130). The CDS (∼981 bp) was cloned into pUC18 plasmid and introduced into Escherichia coli DH10B (GIBCO BRL) cells. The positive clones were selected by restriction digestion analysis and were confirmed by sequencing. The newly obtained sequence from indica rice cultivar IR64 was deposited in GenBank with the accession no. OP856635.

Preparation of the OsNAC121 gene construct for developing silencing (RNAi) and overexpression (Ox) lines in autologous host rice

For RNAi-mediated downregulation of OsNAC121, a hairpin (hp) RNA forming NAC121 DNA segment was prepared in Agrobacterium-based binary plasmid pCAMBIA1300 (CAMBIA, Canberra, Australia) by using an inverted repeat cloned from the intron1 and exon2 fragment of OsNAC121. The transgene was fused with the rice polyubiquitin (RUBQ) gene promoter and the nopaline synthase (NOS) transcriptional terminator for constitutive expression, as reported previously (Das et al. 2018). The plasmid construct contained the hygromycin phosphotransferase II (hptII) as plant selection marker gene. The recombinant plasmid was designated as pCAM::RUBQ-ihpNAC121-NOS (abbreviated as RNAi) (Fig. S1A). Similarly, the gene construct for constitutive overexpression of OsNAC121 in rice was prepared in Agrobacterium-based binary plasmid pCAMBIA1300. In this case, the transgene was fused with the doubly enhanced CaMV35S promoter and the nopaline synthase (NOS) transcriptional terminator, as used before (Cheng et al. 1998). The recombinant plasmid was designated as pCAM::2 × 35 S-NAC121-NOS (abbreviated as Ox) (Fig. S1B). Both the plasmid constructs were separately introduced into E. coli DH10B strain, and the proper orientation of the genetic elements was verified by plasmid DNA isolation followed by restriction enzyme digestion. Next, the desired RNAi plasmid or Ox plasmid was introduced into A. tumefaciens strain EHA105 by freeze-thaw method. The positive Agrobacterium clone was selected by plasmid DNA isolation, followed by verification of the chimeric plasmid by restriction enzyme digestion.

Transformation, selection and regeneration of transformed rice plants

Independent transgenic rice lines were developed following Agrobacterium-mediated embryogenic callus tissue transformation using the EHA105 strains harbouring RNAi construct or Ox construct as per protocol developed by Das et al. 2018. Putative rice transformants were selected on MS medium containing 50 mg L− 1 hygromycin (Duchefa Biochemie). The selected rice lines of RNAi and Ox events along with untransformed control IR64 plants were grown in net house for further studies and seed collection.

PCR screening of the transformed plants

Genomic DNA was extracted from fresh leaves of the putative rice transformants (selected on the basis of hygromycin resistance) and also from the untransformed control plants by CTAB method (Doyle and Doyle 1990). PCR screening of the putative transgenic lines was performed using the construct-specific primers (Table S1). During PCR, the recombinant plasmid for RNAi construct or Ox construct and the genomic DNA from the untransformed control IR64 plant were used as the positive and negative controls, respectively.

Quantitative reverse transcription PCR (qRT-PCR)

Total RNA was isolated from the whole seedlings, root or shoot tissues of untransformed control or phytohormone treated or transgenic (RNAi or Ox lines) rice plants using the RNeasy plant mini kit (Qiagen), and first strand cDNA was synthesized following the procedure already described in Sect. 2.3. qRT-PCR was performed in the Step One Plus Real-Time PCR System (Life Technologies) with Power SYBR Green PCR Master Mix (Life Technologies). At the end of the PCR cycles, the products were subjected to melt curve analysis to verify the specificity of PCR amplification. Fold changes were calculated using the comparative CT method (Livak and Schmittgen 2001), and CT values for individual variants were compared to those of an endogenous reference gene (HistoneH3 or actin1). The OsNAC121-specific primer pair was used for the expression study of OsNAC121 gene. Similarly, specific primer pairs for several phytohormone-responsive genes were used for qRT-PCR analysis. Primers used for qRT-PCR of different genes are noted in Table S2.

Measurement of root length, panicle and grain size

Images of roots or whole plants were captured and subsequently root length was measured using Image J software.

Microscopy of amyloplasts in root tips

From 14-day old plants (untransformed control and transgenic), about one cm long root tips were excised and soaked in 4% I2-KI solution briefly (Takahashi et al. 2003) and washed thoroughly in distilled water, and immediately observed under light microscope (Olympus).

Gravitropism experiments

Control (i.e., untransformed) and transgenic rice seedlings were grown vertically on solid MS medium for 3 days in the plant growth chamber and then immediately reoriented with a 90° rotation. After 36 h, the root curvature of transgenic and control seedlings was quantified. Similarly for measurement of shoot curvature, control (i.e. untransformed) and transgenic rice seedlings were grown vertically on solid MS medium for 5 days in the plant growth chamber and then immediately reoriented with a 90° rotation. After 48 h, the shoot curvature of transgenic and control seedlings was quantified.

Bioinformatics

For in silico studies on gene expression and protein-protein interaction, web based tools Genevestigator (https://genevestigator.com) and STRING database (https://string-db.org) were used, respectively. In order to have a high confidence level on the predicted data, the parameter for interaction score was set at 0.7 in STRING database.

Statistical analysis

All the statistical analyses of the data obtained during the experiments were carried out using the unpaired t-test with a two-tailed distribution. Each experiment was replicated minimum 2–3 times and the sample size used in each replicate is mentioned in respective figure legend. Changes in measurements were considered statistically significant at P < 0.05 (95% confidence level).

Results

Ubiquitous expression of OsNAC121 is highly induced by phytohormones auxin, cytokinin and abscisic acid

In silico analysis of the expression pattern of OsNAC121 in rice japonica cultivar through Genevestigator, a public database for analysis of transcriptome data, revealed a basal level of constitutive expression throughout the life cycle of rice plant with an increased expression in the early seedling stage (Fig. S2A). The abundance of OsNAC121 transcript was experimentally studied in seedlings of indica rice cultivar IR64. qRT-PCR results of whole seedling confirmed that the gene is expressed in early seedling stage (Fig. S2BI and II). The analysis also clearly showed that at the three-leaf stage (14-day old seedling), the expression is significantly higher in roots compared to shoot tissues (Fig. S2BIII).

Based on microarray data curated from Genevestigator, the gene expression analysis of OsNAC121 predicts its expression upon induction with phytohormones - auxin and cytokinin. To validate experimentally, we treated 14-day old IR64 seedlings with auxin and cytokinin, and performed OsNAC121-specific qRT-PCR. The results confirmed that OsNAC121 is highly induced upon exogenous application of either cytokinin (10 µM BAP or 10 µM kinetin) or auxin (10 µM NAA) in rice root tissues (Fig. 1A). Upon induction for 8 h, there is nearly 17-fold,12-fold and 6-fold increase in OsNAC121 transcript level in presence of BAP, kinetin and NAA, respectively. However, the expression of OsNAC121 is unperturbed by ethephon (ethylene releasing compound) and gibberellic acid (GA) even after 24 h (Fig. 1B). We further examined the inducible nature of OsNAC121 expression by other stress-responsive phytohormones such as ABA, JA and SA. The findings showed that the expression of OsNAC121 is significantly upregulated with ABA but downregulated by both JA and SA, as depicted at three different time points after phytohormone treatment (Fig. 1C-E).

Fig. 1
figure 1

Expression profile of OsNAC121 gene in roots of 14-day old IR64 seedlings upon induction with various phytohormones. (A) Relative expression of OsNAC121 after 8 h in response to treatment with cytokinin (10 µM BAP, 10 µM kinetin) and auxin (10 µM NAA). (B) Relative expression of OsNAC121 after 24 h upon treatment with gibberellic acid (GA, 50µM) and ethylene (ethephon (4 mg/L)). (C, D, E) Relative expression pattern of OsNAC121 at different time points in presence of (C) abscisic acid (ABA, 100µM), (D) jasmonic acid (JA, 100 µM) and (E) salicylic acid (SA, 100 µM). Gene expression data were normalized against the endogenous housekeeping control Histone H3 gene and calculated relative to the expression level in root tissue of 14-day old seedling which was considered as unity. In all the bar graphs, data are expressed as mean ± SEM (n = 3)

Full size image

OsNAC121 modulates crown root formation in rice seedlings

In order to study the loss-of-function and gain-of-function phenotypes of OsNAC121 in autologous host, rice calli were transformed with Agrobacterium tumifaciens strain EHA105 harbouring either silencing construct pCAM::RUBQ-ihpNAC121-NOS (abbreviated as RNAi) or overexpression construct pCAM::2 × 35 S-NAC121-NOS (abbreviated as Ox), respectively (Fig. S1A and S1B). The putative transgenic lines were screened through hygromycin selection and transgene-specific PCR (Fig. S1C-S1E). The expression of endogenous OsNAC121 gene in the RNAi and Ox lines was checked by qRT-PCR. Based on qRT-PCR results, two transgenic lines (RNAi#3 and #14) with relatively lower expression of OsNAC121 for silencing event and two transgenic lines (Ox#1 and Ox#9) with relatively higher expression of OsNAC121 for overexpression event (Fig. S1F and S1G), were selected for further experiments.

Seeds of RNAi silencing lines were germinated on solid MS media along with the untransformed control IR64 plants. The 14-day old plantlets exhibited a stark difference in root architecture. There was a significant decrease in the number of crown roots (Fig. 2A and B) and in the length of the primary roots (Fig. 2A and C) in RNAi lines. Thus, phenotypic characterization of loss-of-function (RNAi) lines indicates the role of OsNAC121 in development of root system in the rice plants. Exogenous application of auxin (10 µM IAA) to the transgenic RNAi lines failed to increase their number of crown roots (Fig. 2D) as well as the length of primary roots (Fig. 2E).

Fig. 2
figure 2

Reduction in number of crown roots and length of primary roots in OsNAC121 RNAi lines did not exhibit any improvement upon exogenous auxin application. (A) Image of 14-day old seedlings revealing differences in number of crown roots (blue arrowheads) and primary root length (purple arrowheads). (B, C) The number of crown roots and length of primary roots in 10-day old seedlings of untransformed IR64, two RNAi lines. (D, E) The number of crown roots and length of primary roots in 10-day old seedlings of untransformed IR64 and two RNAi lines) in the presence of 10 µM IAA. In all the bar graphs, data are expressed as mean ± SEM (n = 10) as per two-tailed t-test at 95% confidence level

Full size image

OsNAC121 regulates development of lateral roots and their growth

To study the lateral root phenotypes of OsNAC121 transgenic lines (RNAi and Ox), seeds were germinated on solid MS medium along with untransformed control IR64 plants. The one month old plantlets exhibited different pattern of lateral root formation (Fig. 3A). The roots of RNAi lines were more fragile (Fig. 3A I and II) compared to the untransformed control IR64 plant (Fig. 3A III), while the roots of Ox lines were stout (Fig. 3A IV and V). The lateral roots of RNAi lines had irregular gap between two lateral roots as well as are of various lengths. In contrast, those of Ox lines showed a regular pattern and uniform increased length (Fig. 3A). The lateral root density is significantly lower in RNAi lines, whereas it is higher in Ox lines compared to the untransformed control IR64 (Fig. 3A and B).

Fig. 3
figure 3

Variation in lateral root structure between untransformed control and transgenic plants. (A) Rice plants exhibiting difference in lateral root pattern in untransformed control, RNAi and Ox lines. (I-II) Fragile and irregular lateral roots in RNAi lines. (III) Lateral root structure of control IR64 plants. (IV-V) Stout and regular pattern of lateral roots in Ox lines. (B) Lateral root density measurement demonstrated significant decrease in RNAi lines and increment in the Ox lines compared to untransformed IR64 plant. In all the bar graphs, data are expressed as mean ± SEM (n = 10) as per two-tailed t-test at 95% confidence level

Full size image

OsNAC121 regulates tiller morphology

The RNAi lines exhibited significantly shorter height accompanied by nearly double the number of tillers with respect to untransformed control IR64 plants (Fig. 4). In contrast, the Ox lines did not show any considerable difference in tiller height as compared to control IR64, but a marginal increase in tiller number was observed (Fig. 4).

Fig. 4
figure 4

Comparison of tiller morphology between untransformed control and transgenic plants. (A) Tiller height of RNAi lines is significantly shorter with respect to untransformed control IR64, while the Ox lines exhibit marginally taller tillers which were statistically non-significant. (B) RNAi lines develop nearly double number of tillers while Ox lines exhibit similar number of tillers as the untransformed control IR64. (C) Images depicting the variation in tiller height among RNAi, Ox and control IR64 plants. (D) Images showing the variation in number of tillers among RNAi, Ox and the control IR64 plants. In all the bar graphs, data are expressed as mean ± SEM (n = 20, per cropping season for three seasons) as per two-tailed t-test at 95% confidence level

Full size image

OsNAC121 plays a role in panicle development

The differences in panicle morphology of the transgenic rice plants with respect to the untransformed control IR64 plants were noted. The panicles of RNAi lines were smaller, bearing lesser grains (one third of that in control), while nearly 47% of these grains were chaffy (Fig. 5). The Ox lines with relatively larger panicles bore fully formed grains and the number of filled grains per panicle (~ 86%) was similar to that in control plants (Fig. 5). Apart from the panicle length and grain number per panicle, the other characteristic features of a panicle structure were also evaluated and a stark difference in RNAi lines were observed in all the cases, as documented in Table S3.

Fig. 5
figure 5

Difference in panicle morphology of OsNAC121 RNAi, Ox and untransformed control IR64. (A) The panicles of transgenic lines are distinctly different from the control IR64 plant. The RNAi panicles are more erect with more chaffy grains and lesser number of total grains, while the Ox panicles are curved and compact compared to control IR64 panicle. The number of grains and percentage of chaffy grains in Ox lines are comparable with that of control IR64. The green arrows indicate the properly filled grains and the black ones show the chaffy grains. (B) Diagrammatic structure of a rice panicle. The morphological features of a panicle which often dictates yield of a rice plant are labeled

Full size image

OsNAC121 plays a role in grain morphology

The grains of RNAi plants were morphologically quite different than the untransformed control IR64 grains. We measured the typical agronomic parameters (viz. the grain length, grain width and 1000 grain weight) which are considered significant from yield perspective. Interestingly, RNAi grains exhibited a significant difference in all the tested parameters with respect to untransformed control IR64 grains (Fig. 6A-D). The Ox grains appeared slightly greater in width but not much variation is observed in grain length or 1000 grain weight when compared with the untransformed IR64 grains (Fig. 6A-D). We also calculated the percentage of chaffy grains in the transgenic lines. The RNAi lines had nearly thrice amount of chaffy grains compared to the untransformed control IR64 (Fig. 6E).

Fig. 6
figure 6

Role of OsNAC121 in grain morphology and grain filling. (A) Images of grains of the RNAi, Ox and untransformed control IR64 plants reveal that the RNAi grains are visibly deformed. (B-C) The RNAi grains are smaller, thinner compared to the control IR64. The grains of Ox lines are significantly wider than the control ones. (D) 1000 grain weight of the RNAi, Ox and control IR64 plants. The RNAi grains are significantly lighter compared to control ones. The Ox grains weigh similar to that of control IR64. (E) Percentage of chaffy grains per panicle in RNAi, Ox and control IR64 plants. The RNAi lines had nearly three times more undeveloped grains compared to the control IR64 or Ox lines. In all the bar graphs (B-E), data are expressed as mean ± SEM (n = 20, per cropping season for three seasons) as per two-tailed t-test at 95% confidence level. (F) Images showing (I) fourteen-day old control IR64 seedling and (II) the different stages of panicle development (viz., unemerged panicle, half emerged panicle, emerged panicle and milky stage) used for isolation of total RNA for the purpose of qRT-PCR analysis. Scale = 2.54 cm. (G) Relative expression of OsNAC121 gene in different reproductive tissues of control IR64 showed about 47-fold higher transcript abundance in the milky stage. Gene expression data were normalized against the endogenous housekeeping control actin1 gene and calculated relative to the expression level in shoot tissue of 14-day old seedling which was considered as unity. In all the bar graphs, data are expressed as mean ± SEM (n = 3)

Full size image

To check the activity of OsNAC121 in grain filling, we performed qRT-PCR of RNA samples isolated from the untransformed IR64 panicles harvested at different stages of reproductive cycle. Interestingly, there was a spike in OsNAC121 expression that is ~ 47-fold higher during the milky stage as compared to the 14d-old shoot tissues (Fig. 6F-G), confirming the role of OsNAC121 in grain filling.

Role of phytohormones auxin and cytokinin in altering plant architecture in transgenic lines

Exogenous application of cytokinin (1µM BAP) prevents lateral root formation in untransformed IR64 as well as all the transgenic lines (both RNAi and Ox) (Fig. S3). On the other hand, exogenous supplementation of auxin (5µM IAA) had no effect on the RNAi lines but exhibited longer lateral roots on the Ox lines (Fig. S4). The application of auxin transport inhibitor TIBA (2,3,5-triiodobenzoic acid) showed lack of lateral roots formation in IR64 and Ox lines, similar to the phenotype of the untreated RNAi lines (Fig. 7A-E). The finding confirms that OsNAC121 does play a crucial role in modifying auxin signalling pathway by disturbing the auxin transport mechanism in lateral root formation.

Fig. 7
figure 7

Phenotypes of roots after exogenous application of auxin transport inhibitor TIBA on untransformed control IR64 and Ox lines, compared to RNAi lines without treatment. (A) Control IR64 seeds were germinated on solid MS medium supplemented with 10 µM TIBA. (B) Control IR64 seeds were germinated on solid MS medium. (C) RNAi seeds were germinated on solid MS medium. (D) Ox seeds were germinated on solid MS medium supplemented with 10 µM TIBA. (E) Ox seeds were germinated on solid MS medium. Roots were imaged after 7 days. Seedlings of IR64 and Ox clearly showed lack of lateral root formation in presence of TIBA, similar to the root phenotype of RNAi lines. Scale = 10 mm

Full size image

To further confirm that OsNAC121 regulates auxin-mediated signalling network, the expression level of a couple of genes involved in auxin transport and signalling was checked through qRT-PCR. Since, the OsNAC121 transcript is expressed at higher level in root tissues of IR64, we have focussed on the root tissues for studying the transcript abundance of downstream genes in both RNAi and Ox lines.

The expressions of genes related to auxin biosynthesis (OsYUCCA5), transport (OsPIN1b), signalling (OsARF16) and auxin degradation (OsGH3.8) were downregulated in root tissues of RNAi lines, while they were upregulated in the Ox lines (Fig. 8A-D). However, the transcript level of an auxin signalling-related gene OsIAA10 was higher in roots of both types of transgenic lines (Fig. 8E). Moreover, we checked the transcript abundance of auxin-responsive cell cycle regulating gene Cyclin Dependent Kinase (OsCDKB1;1) in RNAi lines, which showed its down regulation in root tissues (Fig. 8F). Furthermore, the transcript level of genes involved in initiation of crown roots like Crownless Root 4 (OsCRL4) was also estimated by qRT-PCR. It was found to be downregulated in the root tissues of RNAi lines while upregulated in the Ox lines (Fig. 8G). Since shoot tissue of IR64 exhibited lower expression of OsNAC121 and only RNAi lines showed significant phenotypic variations in shoot architecture, but not the Ox lines, we analysed transcript abundance of a few auxin-responsive genes in the shoot tissues of RNAi lines. qRT-PCR results showed that the expression of OsIAA10 is increased, while OsCDKB1;1 expression is decreased in RNAi shoot tissues (Fig. S5). Thus, cumulatively, auxin-mediated signalling network is severely disturbed in the RNAi lines, suggesting OsNAC121 plays crucial role in auxin-mediated plant architecture, particularly root system architecture in rice.

Fig. 8
figure 8

Expression profile of a few genes involved in auxin-mediated signalling network and a few auxin-responsive genes studied by qRT-PCR in root tissues of RNAi, Ox and untransformed control IR64 plants. Auxin signaling pathway has three main components- auxin biosynthesis, transport and degradation. (A)YUCCA gene products in YUCCA pathway of auxin biosynthesis are the most common source of auxin. Transcript level of OsYUCCA 5 was found to be significantly variable among control and transgenic lines. (B) Auxin transport is mediated by PIN proteins. Transcript abundance of OsPIN1b was noted to be significantly variable between Ox lines and control IR64. (C) Expression of an auxin signaling protein, i.e., auxin responsive factor 16 (OsARF16) was found to be variable in different plants. (D) Transcript abundance of auxin degradation pathway related gene OsGH3.8 was noted to be significantly variable between Ox lines and control IR64. (E) Gene expression level of OsIAA10 was found to be increased in both RNAi and Ox transgenic lines compared to control IR64. (F) Transcript profile of cyclin dependent kinase gene OsCDKB1;1 which is known to be auxin modulated showed drastic reduction in RNAi lines. (G) Expression level of OsCRL4 gene modulating crown root formation was noted to be increased in Ox lines. In all these qRT-PCR studies, gene expression data were normalized against the endogenous housekeeping control actin1 gene and calculated relative to the expression level in root issue of 14-day old seedling which was considered as unity. In all the bar graphs, data are expressed as mean ± SEM (n = 3)

Full size image

To confirm that OsNAC121 regulates cytokinin-dependent and auxin-mediated signalling network for lateral root development, we performed qRT-PCR of a few genes involved in cytokinin signalling pathway. qRT-PCR data revealed that transcript levels of cytokinin synthesis (OsIPT5) and signalling (OsRR1 and OsRR3) associated genes are drastically downregulated but its degradation related gene (OsCKX4) is comparatively less downregulated in the RNAi lines (Fig. 9A-D). While in Ox lines, the transcript abundance of OsIPT5 increased significantly, accompanied by overexpression of OsRR genes but that of OsCKX4 was maintained in normal or slightly higher concentration (Fig. 9A-D). Similar qRT-PCR experiments were performed in the shoot tissues of RNAi plants, which revealed mostly similar results (Fig. S6). Thus, the lower cytokinin level coupled with reduced auxin concentration as well as impaired auxin transport regulates the irregular and frail lateral root formation in the RNAi lines.

Fig. 9
figure 9

Expression profile of a few genes in the cytokinin signalling pathway in roots of untransformed control IR64, RNAi and Ox lines checked by qRT-PCR. (A) Transcript abundance of cytokinin biosynthesis gene OsIPT5 was drastically increased in Ox lines. (B) Transcript expression of OsRR1 involved in cytokinin signalling was significantly decreased in RNAi lines. (C) Expression level of OsRR3, another homologue of OsRR1, was also noticeably low in the RNAi lines. (D) Transcript level of cytokinin oxidase gene OsCKX4 involved in degradation of cytokinin via its oxidation was significantly decreased in RNAi lines. In all these qRT-PCR studies, gene expression data were normalized against the endogenous housekeeping control actin1 gene and calculated relative to the expression level in root issue of 14-day old seedling which was considered as unity. In all the bar graphs, data are expressed as mean ± SEM (n = 3)

Full size image

OsNAC121 RNAi lines showed reduced gravitropism in roots

Phytohormone signalling by auxin, cytokinin and ethylene play crucial role in root gravitropism (Nziengui et al. 2018; Vandenbussche et al. 2012). Our findings on variation of OsNAC121 gene expression, induced by auxin and cytokinin, but unaffected by ethylene (Fig. 1A and B). However, the presence of ethylene insensitive (EIN) cis-elements in the promoter region of OsNAC121 (Fig. S7) and auxin and cytokinin-induced OsNAC121 gene expression pattern (Fig. 1A), intrigued us to study the effect of gravity on the RNAi and Ox lines. The gravity-induced root curvature of transgenic lines was examined after stimulation at 90° to the vertical for 24 h. Results showed that the average root tip angles in RNAi lines varies between 76.5° to 96°, which is significantly lower than the untransformed control IR64 plants (Fig. 10A and B). In contrast, the root tip angles of Ox lines ranged between 101° and 132° which is not significantly different from that of the control IR64 plants (Fig. 10A and B). We also checked the accumulation of statoliths in the root tips of transgenic lines. It was noticed that the RNAi lines had very low to no pattern of amyloplast accumulation in the root tip compared to control roots; while the Ox lines had higher number of amyloplasts accumulated at the root tip (Fig. 10C). Similar experiments to observe auxin-mediated alteration in gravitropic response of shoot tissues were carried out, which exhibited similar negative gravitropic response in RNAi and control IR64 plants but significantly higher negative gravitropism in Ox lines (Fig. S8).

Fig. 10
figure 10

Gravitropic response is compromised in root tissues of OsNAC121 RNAi lines. (A, B) Root tip angle of untransformed control IR64, RNAi and Ox transgenic lines. (C) Amyloplasts accumulation at the root tip of control IR64 and RNAi lines clearly revealed lack of statoliths (black arrows) might be the reason for compromised gravitropism in RNAi lines, while Ox lines showed proper arrangement of statoliths in the root tip. Scale = 50 μm. In all the bar graphs, data are expressed as mean ± SEM (n = 3) as per two-tailed t-test at 95% confidence level

Full size image

Discussion

Auxin, cytokinin and ABA are major phytohormones playing crucial roles in plant growth and development (Blakesley et al. 1991; Laplaze et al. 2007; Liu et al. 2017; Saini et al. 2013). Moreover, it is important to maintain the proper ratio of auxin and cytokinin in meristematic zones in the lateral root meristem (LRM) and shoot apical meristem (SAM), which critically controls organogenesis in plants (Neogy et al. 2021; Su et al. 2011). The expression of OsNAC121 is significantly induced in non-transgenic rice plants in presence of auxin, cytokinin and ABA (Fig. 1A and C), the phytohormones whose roles in root and shoot development are well known. Phenotypic alterations observed in transgenic RNAi and Ox lines with respect to untransformed control IR64 plants indicated OsNAC121 is involved in various aspects of plant architecture, starting from root development to tiller, panicle and grain morphologies (Figs. 2, 3, 4, 5 and 6). However, the entire mechanism of OsNAC121-mediated phytohormone signalling in organ development and structure architecture is yet to be elucidated.

The ease of experimenting supplementation of phytohormones and observe their effect upon root system structure prompted us to supplement the transgenic lines with auxin and cytokinin separately and observe any noticeable changes in the crown root and lateral root structures. The RNAi lines failed to increase number of crown roots and primary root length upon exogenous application of auxin (Fig. 2). Hence, we concluded that low auxin biosynthesis in the roots is not the only factor stimulating these phenotypes in OsNAC121 downregulated plants. Neither exogenous application of cytokinin (BAP), reverses the root structure in RNAi lines (Fig. S3). Thus, the syntheses of these two phytohormones are not the only factors leading to the phenotypes observed in RNAi lines. In auxin signalling pathway, auxin transport from the site of synthesis to the meristematic zone is essential to maintain the auxin gradient at the target site required for the development of roots or shoots (Su et al. 2011). Moreover, the experimental validation provided by application of TIBA (auxin transport inhibitor) to the roots of the Ox lines, exhibiting root phenotypes similar to RNAi lines (Fig. 7) clearly indicates the role of OsNAC121 in auxin transport. To reassert the interplay of OsNAC121 and auxin, we checked the gravitropic response of the untransformed control, RNAi and Ox lines. Auxin distribution is crucial for phototropic and gravitropic responses in plants (Muday 2001). The impaired gravitropic responses due to lack of statoliths in the RNAi line (Fig. 10A-C), indicated a perturbed auxin signalling in these transgenic roots, as observed earlier for pin2 mutants which also exhibited agravitropic response (Nziengui et al. 2018). Further, the Ox lines exhibited altered gravitropic response in shoots, while the RNAi lines had similar changes in shoot angles as the untransformed control (Fig. S8). This phenomenon proved the changed auxin redistribution in the transgenic lines in presence of excess OsNAC121 protein in the shoot tissues while similar effect in the root tissues in absence of OsNAC121. Thus, OsNAC121 mediates gravitropic responses in rice via the auxin signalling pathway. Often, altercation in auxin distribution in the shoot tissues are associated with impaired root development, tiller angle and panicle angle (Su et al. 2011; Huang et al. 2021; Wang et al. 2022).

The literature knowledge gained so far, indicates auxin signalling pathway is mediated by receptors Transport Inhibitor Response1 (TIR1), repressor proteins Aux/IAA, and transcription factor-auxin responsive factors (ARFs) to control the expression of a multitude of auxin-responsive genes required for root formation. Aux/IAAs act as repressors of ARFs that modulate auxin responsive genes (Luo et al. 2018). On the basis of our findings through experimental data analyses (Figs. 2, 3, 4, 5, 6, 8 and 9, S4 and S5) we hypothesize that OsNAC121, an uncharacterized transcription factor (TF) might be involved in this pathway and modulates those genes involved in organ development at the meristematic zones. Bioinformatics-based data mining analysis indicated the interaction of OsNAC121 with two TOPLESS (TPL) proteins - B8AKA4 and A2YRH5 (Fig. S9). TPLs are a class of co-repressor proteins which interact with the repressor motifs found in the activation domains of diverse transcription factors (Plant et al. 2021). TPLs lack any DNA binding domain. Hence, they interact with other TFs to modulate the expression of downstream genes (Xu et al. 2013). As mentioned earlier, auxin signalling pathway is mediated by Aux/IAA proteins and ARFs; and TPLs are an important component of this Aux/IAA-ARF complex (Lavy et al., 2016). It is anticipated that OsNAC121 interacts with TPLs (B8AKA4 and A2YRH5), which in turn interact with OsIAA10 (Fig. S9), an Aux/IAA repressor protein that is known to be involved in auxin signalling cascade (Jin et al. 2016), as depicted in Fig. S10. The interactions predicted in silico are experimentally proved by yeast two-hybrid assay between putative orthologues of Arabidopsis thaliana (Causier et al., 2011). Jin et al. (2016) showed that the P2 capsid protein of rice dwarf virus (RDV) binds to OsIAA10 and blocks its interaction with OsTIR1 and thereby inhibits its ubiquitination leading to phenotypes like shorter primary root length, lesser crown roots, stunted tiller height and reduced seed fertility in infected plants. Similar phenotypes (Figs. 2 and 4A and C A) were observed in our RNAi lines as well. They also documented that OsIAA10 overexpression and degradation resistant lines exhibit same phenocopy as RDV infected ones. Based on our in silico analyses and phenotypes observed in the transgenic RNAi and Ox lines, it seems that OsNAC121 interacts with OsIAA10 physically in association with the TPLs in the presence of auxin, thereby facilitating its ubiquitination and hence promoting the release of ARFs (Fig. S10). Thus, the expressions of downstream auxin-responsive genes are modulated. On the other hand, the OsIAA10 ubiquitination is prevented in absence of OsNAC121, as a result the effects of reprogrammed auxin signalling are observed in RNAi lines. In RNAi lines, with lower concentration of OsNAC121, the transcript level of OsIAA10 is higher in comparison to untransformed IR64 lines (Figs. 8E and S5A). However, the expression patterns of downstream genes which have auxin-responsive cis-elements (AuxREs) in their promoters are highly impacted due to the modulation of OsIAA10 proteins leading to disrupted auxin signalling in transgenic rice plants (Jin et al. 2016). OsIAA10 interacts with various ARFs- OsARF11, 12 16 and 25, and modulates expression of various genes in different manner. osarf12 and osarf16 mutants exhibit susceptibility while osarf11 mutant shows tolerance against RDV infection (Qin et al. 2020). OsARF16, an important protein involved in auxin signalling pathway, is known to be involved in cytokinin-mediated inhibition of phosphate transport and Pi signalling in rice (Shen et al. 2014). Hence, we have investigated the transcript level of OsARF16 in our transgenic plants to observe any effect of OsNAC121 on the expression level of the former. We found that the transcript abundance of OsARF16 in RNAi lines is reduced (Fig. 8C), indicating that the lack of OsNAC121 does interfere with the normal auxin signalling cascade. Moreover, despite the marginal reduction in transcript level of an auxin transporter OsPIN1b in RNAi lines and 1.5 to 2-fold increase in Ox lines (Fig. 8B), the RNAi plants exhibit clear disruption of auxin transport. This implies that lack of OsNAC121 proteins in the downregulated lines lead to such root and shoot phenotypes (Figs. 2, 3, 4, 5, 6 and 7).

Additionally, the lower transcript level of cell cycle dependent kinase gene OsCDKB1;1(Fig. 8F) might result in lower cell division in the meristem of RNAi lines which could lead to their reduction in primary root length as well as stunted tiller height. Moreover, a reduced expression level of OsCRL4 in RNAi lines might be the reason for reduced number of crown roots (Fig. 8G). It is worth mentioning that OsCRL4 (a Lateral Organ Boundaries LOB TF) is known to act upstream of OsCRL1 known for its role in crown root formation (Jung Janelle and McCouch Susan 2013). It is also involved in polar localization of PIN1 auxin transporters (Gao et al. 2014).

In silico-based protein-protein interaction study also indicated the interaction of OsNAC121 with ABA receptor (ABAR8/OsPYL4). The data is inferred from the yeast two-hybrid screening revealing interactions between their putative orthologues in Arabidopsis, i.e. the OsNAC121 orthologue ANAC096 with ABA receptor orthologues PYL 12 and PYL6 (Altmann et al. 2020). An interesting study by Li et al. (2020c) reveals the presence of PYR/PYL/RCAR (abbreviated as PYL) receptor-protein phosphatase 2 A (PP2A) complex, which functions in parallel with the PYL-PP2C-SnRK pathway. PYL-PP2A signaling cascade is involved in lateral root formation as well as in root gravitropism (Li et al. 2020c). Phosphorylation and de-phosphorylation of PIN proteins are very critical for proper auxin transport. Thus, ABA receptor proteins PYLs articulate the directional auxin transport leading to modified root architecture. PYL-PP2A-SnRK signaling cascade is modulated in presence of ABA, which induces OsNAC121. Hence, PYLs bridge the ABA and auxin signaling pathways in root development (Li et al. 2020c). Our qRT-PCR data showed very high level of OsNAC121 expression upon auxin, cytokinin and ABA induction (Fig. 1). In RNAi lines, the expression of genes for both auxin and cytokinin biosynthesis is low as revealed by qRT-PCR data (Figs. 8A and 9A). Thus, the possible alteration in phosphorylation or dephosphorylation of PIN proteins in RNAi might lead to the disruption of auxin signaling, and subsequently resulting in impaired lateral root formation as well as insensitivity towards gravitropism.

Du et al. 2018 has vividly summarized the role of auxin in lateral root formation. Auxin has been reported to articulate the initiation of lateral root primordia (LRM), its spatial distribution, outgrowth as well as emergence. Cytokinin antagonizes the lateral root formation but induces auxin synthesis which promotes the lateral root structures of a plant. In RNAi lines, the expression of genes involved in cytokinin biosynthesis (OsIPT5) and signalling (OsRR1) are downregulated, whereas the OsCKX4 gene expression required for its degradation via oxidation is comparatively higher (Fig. 9), indicating a lower level of cytokinin at root tissues. However, in Ox lines, the transcript abundance of OsIPT5 gene is markedly upregulated (Fig. 9A), implying a higher cytokinin concentration that can lead to formation of stouter and regular lateral roots and a lower transcript level of auxin biosynthesis gene OsYUCCA5 is observed in RNAi lines (Fig. 8A).

Interestingly, the cytokinin oxidase gene OsCKX4 is involved in degradation of cytokinin via its oxidation. Detailed literature study showed that OsCKX4 is known to be involved in various aspects of rice plant architecture. It positively regulates grain size, various panicle morphologies, flag leaf size and internode diameter but negatively regulates tiller number (Gao et al. 2014; Rong et al. 2022). Moreover, OsCKX4 regulation via cytokinin signalling molecule OsRR3 has been reported (Gao et al. 2014). OsRR3 is reported to reduce sensitivity of the plant towards cytokinin (Cheng et al. 2010). OsRR3 is a type A response regulator that functions as a negative regulator of OsCKX4 (Gao et al. 2014). Yeast one-hybrid assay has revealed OsCKX4 interacts with OsARF25 (Gao et al. 2014), and OsARF25 is known to positively regulate grain size (Zhang et al. 2018b). Yeast-two hybrid assay has proved interaction between OsARF25 and OsIAA10 proteins (Qin et al. 2020). Hence, the auxin and cytokinin signalling pathways are interconnected by the OsCKX4 gene product (Fig. S11). Intriguingly, in OsNAC121 RNAi lines, the OsRR3 gene expression is upregulated (Fig. S6B) while OsCKX4 is downregulated (Fig. S6C) in shoot tissues, which explain smaller grains and altered panicle morphology observed in OsNAC121 knockdown lines (Figs. 5 and 6A-C). While in the root tissues, OsCKX4 along with OsARF25 and cytokinin response regulators (OsRR2 and OsRR3), promotes crown root formation (Gao et al. 2014). It is also known that overexpression of OsRR3 induces more lateral roots and longer roots by reducing cytokinin sensitivity (Cheng et al. 2010). The expression of both OsRR genes (OsRR1 and OsRR3) and OsCKX4 are lower (Fig. 9B-D) in the OsNAC121 silencing lines than the untransformed control IR64 plants, thereby explaining the lower number of crown roots observed in the RNAi lines. Thus, a balanced auxin: cytokinin ratio, maintained by auxin-cytokinin feedback regulation is very crucial in the LRM local tissues to decide the fate of these cells (Su et al. 2011; Ruzicka et al. 2009). The activity of transcription factor in co-ordination with phytohormones auxin and cytokinin dictates the undifferentiated cells of LRM and SAM in determining their fates and thereby developments of different organs like root, leaf, panicle, inflorescence etc. (Barton 2010).

In conclusion, our molecular and phenotypic analyses of the transgenic rice lines document that OsNAC121 integrates auxin and cytokinin pathways to modulate a plethora of genes involved in various aspects of plant architecture viz., crown root number, lateral root formation, tiller number, tiller height, panicle and grain morphologies. The altered auxin concentration alone or in conjugation with impaired cytokinin signaling impacts the fate of meristematic cells in the transgenic rice lines. Consequently, different phenotypic variations were observed in the OsNAC121 gene silencing and overexpression lines. The exact mechanism of crosstalk between OsNAC121 and phytohormones is yet to be elucidated. However, OsNAC121 is documented to be an important regulator controlling an array of economically important agronomic traits in rice plant.