Natural variation of the MIR168a promoter influences its activity and miR168a abundance in rice

Abstract

miR168a, a plant-specific microRNA (miRNA) derived from the MIR168a gene, plays a pivotal role in modulating rice blast disease resistance and critical agronomic traits such as flowering time and yield. However, the regulatory mechanisms governing the MIR168a promoter remain poorly understood. This study identified a 1661 bp upstream fragment of the mature miR168a as highly active in promoter function. Sequence alignments revealed substantial variation in MIR168 promoters across plant species. Analysis of over 4000 rice accessions showed that while the MIR168a promoter exhibited abundant SNPs and InDels, miR168a itself had no such polymorphisms. Based on promoter polymorphisms, the MIR168a promoter was classified into three haplotypes, with Hap2 and Hap3 showing higher activity than Hap1. Through DNA fragment swapping and site-directed mutagenesis, the T site in Hap2 and the A site in Hap3 were identified as critical determinants of promoter activity. Rice accessions containing these sites exhibited significantly higher miR168a abundance compared to Hap1 accessions. Population genetic and evolutionary analyses revealed that highly active MIR168a promoters in Hap2 and Hap3 predominantly occur in indica accessions and trace their origins to wild rice. Furthermore, the nucleotide diversity of the MIR168a promoter in cultivated rice was markedly lower than in wild rice, likely reflecting artificial selection during domestication and artificial selection. Breeders may have favored rice lines harboring MIR168a promoter variants with reduced activity, such as Hap1 accessions, underscoring its potential for breeding programs. Additionally, miR168a expression was induced in all three haplotypes following infection with Magnaporthe oryzae. These findings illuminate the natural variation in MIR168a promoter sequences and their influence on miR168a expression activity, offering new insights for rice improvement strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12284-025-00842-z.

Background

MicroRNAs (miRNAs) are a class of small RNAs derived from the post-transcriptional processing of miRNA genes (MIR) (Chen 2009; Li et al. 2010). Transcription begins when RNA polymerase II (Pol II) is recruited to the MIR promoter by the transcriptional coactivator mediator, forming primary miRNAs (pri-miRNAs) (Kim et al. 2011; Rogers & Chen 2013). These pri-miRNAs are subsequently processed by DICER-LIKE (DCL) proteins into precursor miRNAs (pre-miRNAs) and mature miRNAs, typically 19–24 nucleotides (Margis et al. 2006; Achkar et al. 2016). Mature miRNAs associate with Argonaute (AGO) proteins to form the RNA-induced silencing complex (RISC), which primarily mediates target gene regulation through mRNA degradation (Liu et al. 2004), translational inhibition (Fabian et al. 2010; Li et al. 2014), and transcriptional silencing via DNA methylation (Khraiwesh et al. 2010; Wu et al. 2010). Accumulating evidence highlights the critical role of miRNAs in modulating growth, development, metabolism, and responses to biotic and abiotic stresses in both plants and animals (Achkar et al. 2016; Gebert & MacRae 2019). The MIR gene promoter contains TATA box and cis-regulatory elements, demonstrating its regulation by various transcriptional factors (Megraw et al. 2006; Zhao et al. 2013). In Arabidopsis thaliana, the SANT-domain-containing protein POWERDRESS (PWR) enhances the expression of MIR172a-c genes, promoting floral determinacy (Yumul et al. 2013). In contrast, MIR172d expression is PWR-independent but suppresses APETALA2 (AP2) gene expression, forming feedback regulatory loops (Yant et al. 2010). Other transcription factors, such as SQUAMOSA PROMOTER BINDING PROTEIN-LIKE7 (SPL7) and AtMYB2, modulate miRNA expression under copper deficiency and phosphate stress (Yamasaki et al. 2009; Baek et al. 2013), respectively. Similarly, abscisic acid (ABA)-responsive elements in the MIR168a promoter interact with ABRE-binding factors to regulate miR168a expression (Li et al. 2012). In mammals, oncogenic transcription factors such as c-Myc and Ras-responsive element-binding protein (RREB1) regulate MIR promoter activity, influencing tumorigenesis (Chang et al. 2008; Kent et al. 2010). These findings underscore the biological significance of MIR promoter regulation.

The diversity among MIR isoform genes suggests differential regulatory mechanisms underlying gene expression (Rogers & Chen 2013). miR168 is a highly conserved, plant-specific small RNA involved in diverse biological processes. It modulates growth and development, including somatic embryogenesis (Siddiqui et al. 2019), phase transitions, and fruit development (Xian et al. 2014). miR168 also modulates abiotic stress responses, such as low-potassium stress in tomato (SlmiR168), where its overexpression enhances root and leaf development (Liu et al. 2020). In biotic stress, pathogens and elicitors induce miR168 to modulate plant resistance. For instance, in Malus hupehensis, Botryosphaeria dothidea activates the MIR168 promoter, increasing miR168 levels and enhancing pathogen resistance (Yu et al. 2017). Conversely, miR168 induction by tombusvirus inhibits AGO1 translation, weakening antiviral responses (Várallyay et al. 2010).

Moreover, miR168 demonstrates cross-kingdom regulation. Plant miR168a has been detected in human and animal tissues, where it binds to low-density lipoprotein receptor adapter protein 1 (LDLRAP1) mRNA, reducing LDLRAP1 levels and elevating LDL in mouse plasma (Zhang et al. 2012). These multifaceted roles highlight the significance of miR168 in both plant and non-plant systems, further emphasizing the importance of understanding its regulation at the promoter level.

The rice miR168 family includes MIR168a and MIR168b, producing two mature isoforms, among which miR168a is predominantly expressed. By repressing AGO1, a core component of the RNA-induced silencing complex (RISC), miR168 serves as a central node in the rice miRNA regulatory network. In a comparative study of the disease-resistant rice accession IRBLkm-Ts and the susceptible accession Lijiangxintuanheigu (LTH), inoculation with M. oryzae followed by miRNA deep sequencing revealed that M. oryzae induces the expression of miR168a. However, no expression of miR168b was detected (Li et al. 2014). Overexpression of miR168a significantly suppresses AGO1 expression, leading to increased susceptibility to M. oryzae, delayed flowering, and reduced yield-related traits. Conversely, inhibiting miR168a expression enhances rice resistance, increases tiller numbers, shortens flowering time, and improves yield (Wang et al. 2021). These findings highlight the potential of miR168a modulation for improvement of agronomic traits. Despite this, the specific cis-elements critical for the activity of the MIR168a promoter remain unclear.

Given miR168a’s pivotal role in the rice miRNA regulatory network and its influence on vital agronomic traits, it is scientifically imperative to identify the cis-elements responsible for MIR168a promoter activity. In this study, we first validated the activity of the MIR168a promoter using transient expression assays. Tissue-specific expression patterns were then examined through GUS reporter assays. To explore MIR168a promoter haplotypes and their expression activities, we analyzed over 4,000 genomic datasets, combining bioinformatics approaches with transient expression assays to identify cis-elements regulating MIR168a promoter activity. Finally, we treated rice haplotypes with M. oryzae to evaluate miR168a expression patterns. The results revealed that the MIR168a promoter exhibits substantial polymorphisms, with haplotypes containing the T or A site displaying high activity. Moreover, the low nucleotide diversity of the MIR168a promoter in cultivated rice compared to wild rice indicates that artificial selection during domestication may have favored variants with reduced activity. These findings offer valuable insights into the regulatory mechanisms of the MIR168a promoter and provide a foundation for manipulating miR168a expression to improve rice agronomic traits.

Materials and methods

Plant materials and growth conditions

The rice (Oryza sativa) accessions used in this study were LTH, Nipponbare (NPB), Yahui2115 (R2115), Basmati1, Yihui3551 (Y3551), Guang8B (G8), NamRoo, Wushansimiao (WSSM), Digu (DG), Yixiang 1B (YX1), CG14, N22, W194, W222, W297, W026, W166, C006, W197, W276, and W211. These accessions were cultivated in a growth chamber maintained at 25 °C, with 70% relative humidity and a 14/10 h (h) day/night cycle. Nicotiana benthamiana was cultivated in a chamber at a constant 25 °C day/night temperature, with a 16 h photoperiod and at 70% relative humidity.

Activity analysis of MIR168a promoter with various lengths

To identify the MIR168a promoter activity, we fused five fragments, ranging from 1 to 2313 bp upstream of the miR168a, with luciferase (LUC), following a previous report with slight modifications (Zhao et al. 2021). Briefly, the MIR168a promoter fragments were cloned using genomic DNA (gDNA) of NPB as a template with specific primers listed in Table S1. The amplified PCR products were recovered via electrophoresis and subsequently cloned into the pCAMBIA1300-LUC vector at the EcoRI and KpnI sites (Zhao et al. 2021). Plasmids were extracted and transformed into Agrobacterium strain GV3101. The optical density (OD) was adjusted to 0.5, and Agrobacterium was infiltrated into N. benthamiana leaves. The infiltrated leaves were imaged after 36 h using a CCD camera with 10 μg/mL D-luciferin potassium salt as the substrate.

Dual LUC reporter assay to verify the activity of different haplotypes of the MIR168a promoter

The dual LUC reporter assay was performed as previously described with slight modifications (Zhao et al. 2021). Briefly, RLUC (renilla luciferase) was amplified using pCAMBIA1300-35S::RLUC-NOS plasmid as a template and cloned into the pCAMBIA1301 vector at the NcoⅠ and PmlⅠ sites, generating pCAMBIA1301-35S::RLUC-NOS. Subsequently, 35S::LUC-NOS was amplified using pCAMBIA1300-35S::LUC-NOS plasmid as a template and cloned into the pCAMBIA1301-35S::RLUC-NOS vector at the HindⅢ and EcoRⅠ sites, generating pCAMBIA1301-35S::LUC-NOS-35S::RLUC-NOS. The MIR168a promoter was cloned from accessions, including N22, NPB, Basmati1, Y3551, LTH, NamRoo, WSSM, DG, YX1, CG14, W222, W297, W026, W166, C006, W197, W211, and W276, using gDNA as template. The amplified and gel electrophoresis-recovered PCR products were cloned into the HindⅢ and XbaⅠ sites of the pCAMBIA1301-35S::LUC-NOS-35S::RLUC-NOS vector, respectively, to obtain pCAMBIA1301-MIR168apro::LUC-NOS-35S::RLUC-NOS plasmids. Plasmids harboring the different haplotypes of the MIR168a promoter were extracted and transformed into Agrobacterium strain GV3101. Subsequently, the OD was adjusted to 0.5, and the solution was infiltrated into N. benthamiana leaves. Leaf samples were collected at 36 h after infiltration. Total protein was extracted following the instructions of the Dual-Luciferase Reporter Gene Assay Kit (Beyotime). LUC and RLUC signals were detected using the Glomax instrument from Promega Company.

Verification of the regulatory element that affects the MIR168a promoter activity

To determine the truncation impacting the activity of the MIR168a promoter, we divided the MIR168a promoter into three truncations: the Key Region (KR), the KR upstream truncation, and the KR downstream truncation. Three truncations of the MIR168a promoter from Hap1 were swapped with the corresponding truncations from Hap2 and Hap3, respectively, using the ClonExpress MultiS One Step Cloning Kit (Vazyme) to clone the fragment-swapped promoter into HindIII and Xba1 sites of the pCAMBIA1301-35S::LUC-NOS-35S::RLUC-NOS vector, generating the following plasmids: pCAMBIA1301-Pro_(LTH-KR)-DG::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_LTH-(DG-KR)::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_(DG-KR)-LTH::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_DG-(LTH-KR)::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_(LTH-KR)-YX1::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_LTH-(YX1-KR)::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_(YX1-KR)-LTH::LUC-NOS-35S::RLUC-NOS, and pCAMBIA1301-Pro_YX1-(LTH-KR)::LUC-NOS-35S::RLUC-NOS.

To verify the effect of the T site on the MIR168a promoter activity of Hap2, we used the gDNA from LTH and WSSM as templates, and either the NEBridge® Golden Gate Assembly kit or overlap extension PCR was used to mutate the “CC” in the MIR168a promoter of these two backgrounds to “TT”. Subsequently, the same method was used to introduce the “CC” mutation into the MIR168a promoter cloned from DG and CG14, replacing “TT” with “CC”, generating the following plasmids: pCAMBIA1301-Pro_lth::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_dg::LUC-NOS-35S::RLUC-NOS, pCAMBIA1301-Pro_wssm::LUC-NOS-35S::RLUC-NOS, and pCAMBIA1301-Pro_cg14::LUC-NOS-35S::RLUC-NOS.

To verify the effect of the A site on the MIR168a promoter activity of Hap3, we employed the same methods to introduce the “A” mutation into the MIR168a promoter cloned from LTH background, replacing “CCCC” with “A”, and the “CCCC” mutation into the MIR168a promoter cloned from YX1, replacing “A” with “CCCC”, generating pCAMBIA1301-Pro_{LTH(CCCC to A)}::LUC-NOS-35S::RLUC-NOS and pCAMBIA1301-Pro_{YX1(A to CCCC)}::LUC-NOS-35S::RLUC-NOS, respectively.

The Agrobacterium strain GV3101, which harbored each plasmid, was adjusted to an OD of 0.5 before being injected into N. benthamiana leaves. The infiltrated leaves were imaged using a CCD camera 36 h after injection, with 10 μg/mL D-luciferin potassium salt as the substrate. Total protein was extracted directly according to the manufacturer’s instructions for the Dual-Luciferase Reporter Gene Assay Kit (Beyotime). LUC and RLUC signals were detected using the Glomax instrument from Promega Company. All the primers are listed in Table S1.

Haplotype and nucleotide diversity analysis

The variation data and genomic sequences of 4736 cultivated rice accessions and 33 wild rice accessions were obtained from the Rice Variation Map V2.0 Database (RiceVarMap, https://ricevarmap.ncpgr.cn/) (Zhao et al. 2014), the Rice Resource Center (http://113.54.11.180/) (Qin et al. 2021), and the Rice Super Pan-genome Information Resource Database (RiceSuperPIRdb, http://www.ricesuperpir.com/) (Shang et al. 2022). Variants were filtered using a minor allele frequency (MAF) threshold of ≥ 0.05, and accessions with heterozygous genotypes or missing data within the genomic region of miR168a were excluded. Haplotype analysis was performed using the geneHapR (v1.2.4) (Zhang et al. 2023) package in R (v4.2.2), and subpopulations were grouped according to the RiceVarMap database with minor modifications. The geographical coordinates of cultivated rice accessions were obtained from the RiceVarMap database and clustered by major regions, and then visualized on a world map using the ggplot2 (v3.4.1) (Wickham 2011) package in R (v4.2.2) to analyze the geographic distribution patterns of the three haplotypes. The phylogenetic tree of the MIR168a promoter was generated using the UPGMA method in MEGA5 (v11.0.13) (Tamura et al. 2021), and the haplotype network was constructed by POPART6 (v1.7) (Leigh & Bryant 2015). The nucleotide diversity for each subpopulation in the MIR168a promoter and flanking regions (~ 100 kb) was analyzed using DnaSP6 (v6.12.03) (Rozas et al. 2017) with 2000 bp window and 100 bp step length.

Inoculation of M. oryzae and treatment of PAMPs

The M. oryzae strain GZ8 used in this study was cultured as previously described (Zhao et al. 2020). Briefly, GZ8 was cultured on oatmeal and tomato agar (OTA) medium at 28 °C under 12 h light and 12 h dark cycles. The hyphae were scratched and incubated for four days under the same conditions to induce spore formation. Three-week-old rice seedlings were prepared for spray inoculation by adjusting the spore concentration to 1 × 10⁵ spores/mL. Inoculated leaves were harvested at the designated time points. For miR168a and pri-miR168a abundance analysis following PAMP treatment, the leaves of 3 week-old seedlings were immersed in water, chitin (20 μg/mL), and flg22 (μM) solutions for 0, 6, and 12 h. Leaf samples were subsequently collected and used for RNA extraction.

RNA extraction and stem-loop reverse transcription quantitative PCR (RT-qPCR)

RNA extraction was conducted as previously described (Zhao et al. 2020). Briefly, total RNA was extracted from the collected leaf samples using TRIzol reagent (Invitrogen). RNA quality and quantity were assessed using a spectrophotometer. Stem-loop RT-qPCR was performed to quantify miR168a amounts following previously established methods (Wang et al. 2021). RT-qPCR was used to quantify pri-miR168a levels. U6 and Ubiquitin (Ubi) were used as internal controls. The QX96B (Sichuan Jielaimei Technology Co., Ltd Chengdu, China) was used for RT-qPCR. All the primers used in this study are detailed in Table S1.

MiR168apro::β-glucuronidase (GUS) material construction, GUS staining, and tissue specificity of MIR168a promoter expression

To construct MIR168apro::GUS lines, NPB gDNA was used as a template to clone the MIR168a promoter. The PCR product was purified and cloned into the pCAMBIA1300-GUS vector at the EcoRⅠ and BamHⅠ sites. Subsequently, transgenic lines were generated via Agrobacterium-mediated genetic transformation. Hygromycin was employed to select positive lines. GUS staining was performed as previously described with minor modifications (Tu et al. 2011). Briefly, rice seedlings were immersed in the GUS staining solution [0.1 M sodium phosphate buffer (pH 7.0), 10 mM EDTA, 0.1% Triton X-100, 1 mM K₃[Fe(CN)₆], 2 mM X-Gluc] at 37 °C for 12 h. The staining solution was removed using 95% ethanol. The tissues were photographed using a DSLR camera (Nikon) and a stereo fluorescence microscope (Olympus).

GUS activity detection

GUS activity driven by the MIR168a promoter in the rice seedlings was assessed as previously described with minor modifications (Jefferson et al. 1987; Lardet et al. 2011). Briefly, fluorometric measurements of GUS activity were conducted using 4-methylumbelliferyl-β-D-glucuronide (MUG) as a substrate. Rice leaves were frozen in liquid nitrogen and subsequently ground. The frozen tissue powder was homogenized in extraction buffer. GUS activity of the total protein was measured using the Varioskan LUX Multimode Microplate Reader (Thermo Fisher Scientific).

Results

MIR168a is responsive to pathogen‐associated molecular patterns and M. oryzae

To assess the responsiveness of MIR168a to pathogen-associated molecular patterns (PAMPs) and M. oryzae infection, we quantified the relative abundance of miR168a in the rice lines NPB, LTH, and R2115, which exhibit different phenotypes of blast disease resistance. Following treatment with chitin and flg22, miR168a expression was significantly upregulated in NPB at 6 and 12 h post-treatment with flg22 and at 6 h post-treatment with chitin, compared to the mock control (Fig. 1a). In LTH, miR168a expression increased significantly at 6 h following flg22 treatment and at both 6 and 12 h after chitin treatment (Fig. 1b). In contrast, R2115 showed significant miR168a induction at 12 h after chitin treatment, whereas flg22 treatment only elicited a weak response at the same time point (Fig. 1c). To rule out the influence of miR168a processing process on its accumulation in different genetic background, we quantified the abundance of pri-miR168a and found patterns similar to those of mature miR168a upon PAMP treatment (Fig. S1a, b, and c).

Fig. 1.

The expression pattern analysis of miR168a upon PAMP and M. oryzae treatment. a–c For PAMP treatment, 3 week-old seedlings of NPB, LTH, and R2115 were treated with H₂O, flg22 (2 μM), and chitin (20 μg/ml). Samples were collected at 0, 6, and 12 h post-treatment. U6 was used as the internal reference for normalization. Error bars indicate the standard deviation (SD) (n = 3). Significant differences (P < 0.05) were determined by one-way ANOVA followed by Tukey’s test. d–f For M. oryzae treatment, seedlings of the same age were spray-inoculated with the GZ8 strain (1 × 10^–5 spores/mL). Samples were collected at 0, 12, 24, 48, and 72 h post-inoculation (hpi). Total RNA was extracted, and RT-qPCR was performed to detect the expression of miR168a. U6 was used as the internal reference for normalization. Error bars indicate the standard deviation (SD) (n = 3). Different letters indicated significant differences (P < 0.05), as determined by one-way ANOVA followed by Tukey’s test

To determine if the change in miR168a accumulation leads to suppression of OsAGO1 expression upon PAMP treatment, we quantified the expression levels of OsAGO1a and OsAGO1b, which are targeted by miR168a (Wu et al. 2009). In NPB, OsAGO1a and OsAGO1b were significantly upregulated at 6 and 12 h post-treatment with flg22 and chitin, except for OsAGO1a, which showed only weak induction at 12 h post-treatment with flg22 (Fig. S2a and d). In LTH, flg22 treatment induced weak upregulation of OsAGO1a at 6 h (Fig. S2b), while OsAGO1b was significantly upregulated at both 6 and 12 h (Fig. S2e). Chitin treatment significantly upregulated OsAGO1a at 12 h (Fig. S2b) and OsAGO1b at both 6 and 12 h (Fig. S2e). In R2115, flg22 and chitin treatments significantly increased the expression of both OsAGO1a and OsAGO1b at 12 h (Fig. S2c and f). These results demonstrate that OsAGO1 is also responsive to PAMPs. Although OsAGO1 is a known target of miR168a, PAMP treatments induced the expression of both miR168a and OsAGO1, consistent with previous findings that elicitors upregulate both (Wu et al. 2015). Therefore, subsequent experiments focused solely on miR168a expression.

Next, we analyzed miR168a and pri-miR168a expression in NPB, LTH, and R2115 following inoculation with M. oryzae strain GZ8. In NPB, miR168a expression was significantly upregulated, peaking at 24 and 72 h post-inoculation (hpi) (Fig. 1d). In LTH, miR168a expression was induced as early as 12 hpi and remained elevated until 72 hpi (Fig. 1e). In contrast, miR168a expression peaked at 24 hpi but gradually declined thereafter in the resistance line R2115 (Fig. 1f). Similarly, pri-miR168a expression was significantly induced at 12, 24, 48, and 72 hpi in NPB; at 12, 48, and 72 hpi in LTH; at 24 and 72 hpi in R2115 (Fig. S1d, e, and f). These findings indicate that miR168a is responsive to M. oryzae infection.

To confirm the responsiveness of the MIR168a promoter to PAMPs, we generated transgenic lines expressing GUS under the control of the MIR168a promoter (MIR168apro::GUS) in the NPB background. GUS staining revealed strong signals in coleoptiles, radicles, leaves, sheaths, and roots (Fig. 2a). RT-qPCR confirmed that miR168a expression was ubiquitous in the leaf, sheath, and root tissues of NPB (Fig. 2b). Then, we treated MIR168apro::GUS transgenic lines and NPB plants with flg22 and chitin, followed by GUS staining of roots and leaves. No GUS signal was detected in NPB plants treated with H₂O or PAMPs (Fig. 2c). However, MIR168apro::GUS lines displayed weak GUS signals in roots and leaves following H₂O treatment, while PAMP treatments produced strong GUS signals in these tissues (Fig. 2c). Quantitative analysis of GUS activity in the roots and leaves of the MIR168apro::GUS-7 line corroborated the histochemical staining results, with significantly enhanced fluorometric activity after PAMP treatment (Fig. 2d and e). These results demonstrate that miR168a is ubiquitously expressed in seedlings and is significantly up-regulated by PAMPs.

Fig. 2.

Histochemical GUS staining and GUS activity analysis after PAMP treatment were performed in MIR168apro::GUS-transformed rice. a GUS staining in different developmental stages and tissues. NPB gDNA was used as a template to clone the MIR168a promoter, and the Agrobacterium-mediated genetic transformation was used to construct MIR168apro::GUS lines. b Stem-loop RT-qPCR was employed to confirm that miR168a expression was ubiquitous in the leaf, sheath, and root tissues of NPB. c GUS staining was performed on three-week-old MIR168apro::GUS-transformed rice lines with and without flg22 or chitin treatment. d, e GUS activity analysis following flg22 and chitin treatments in roots (d) and leaves (e). Error bars indicate the standard deviation (SD) (n = 4). Significant differences (P < 0.05) were determined by one-way ANOVA followed by Tukey’s test

A 1.6-kb promoter is essential for maximal MIR168a expression

In eukaryotic organisms, gene expression is tightly regulated by promoter activity. The cis-regulatory elements in promoter regions, such as transcription factor binding sites and enhancers, govern both basal transcriptional activity and dynamic gene expression in response to diverse biotic and abiotic stresses, thereby enabling plants to be resilient to changing environmental conditions. To explore the cis-regulatory elements in the MIR168a promoter, we exploited the LUC-reporter system and constructed LUC reporter constructs using five fragments of varying lengths (2313 bp, 1661 bp, 661 bp, 448 bp, and 205 bp) upstream of mature miR168a, with NPB genomic DNA serving as the template (Fig. 3a). LUC assays revealed that all fragments were capable of driving LUC expression in N. benthamiana, though the 1661 bp fragment exhibited the strongest activity, while the 661 bp and 448 bp fragments showed the weakest (Fig. 3b and c). Notably, the 205 bp fragment exhibited significantly higher promoter activity than 661 bp and 448 bp fragments. We speculate that this difference may be attributed to the presence of cis-regulatory elements within the 661 bp and 448 bp sequences that repress their activity. These results imply that key cis-regulatory elements in the promoter play a pivotal role in determining MIR168a expression levels.

Fig. 3.

MIR168a promoter activity identification. a Schematic diagram showing different lengths of the MIR168a promoter. b LUC activity in N. benthamiana leaves transformed with reporter construct harboring different lengths of the MIR168a promoter. The 35S::LUC served as a positive control, while 35S::eYFP was used as a negative control. c Quantification of LUC activity from b. Different letters indicate significant differences (P < 0.05), as determined by one-way ANOVA followed by Tukey’s test

The MIR168a promoter displays three haplotypes with global distribution in rice

Previous studies have demonstrated the extreme conservation of mature miR168 sequences and their secondary structures across Brassicaceae species (Gazzani et al. 2009). To determine whether the MIR168 promoter and miR168 sequence exhibit similar conservation across the plant kingdom, we retrieved MIR168 sequences from 107 plant species using the PmiREN database (https://www.pmiren.com/). Sequence analysis revealed remarkable conservation of the mature miR168 sequence, with only minimal base pair differences (Fig. S3 and S4). In contrast, the alignment of MIR168 promoter sequences from 27 plant species indicated significant variability, with limited homology restricted to a few regions (Fig. S5 and S6).

To further explore variability in rice, we analyzed SNPs and InDels in the MIR168a promoter and miR168a stem-loop sequences using the RiceVarMap v2.0 database (http://ricevarmap.ncpgr.cn/) and Rice Resource Center (http://113.54.11.180/). The miR168a stem-loop region was highly conserved, with no SNPs or InDels detected, whereas the MIR168a promoter displayed extensive polymorphisms (Fig. 4a and S9a). These findings indicate that although the mature miR168 sequence is evolutionarily conserved, its promoter region exhibits substantial divergence across rice accessions

Fig. 4.

The MIR168a promoter was classified into three haplotypes with global distribution. a The MIR168a promoter haplotype analysis using a population of over 4000 rice accessions. The numbers at the top panel represent the positions of the base pairs in the promoter relative to the miR168a. The red star denotes the position of a “Key Region (KR)”. b The geographical distribution of rice containing different haplotypes of the MIR168a promoter

.

Haplotype analysis of the MIR168a promoter in over 4000 rice accessions, based on the 2-kb upstream sequence, identified three major haplotypes (Hap1, Hap2, and Hap3) (Fig. 4a). Hap1 was found to be the most widely distributed globally, while Hap2 and Hap3 were predominantly localized in Asia, East Africa, and West Africa. Their prevalence declined with increasing latitude, particularly in regions such as North America, South America, South Africa, and Europe (Fig. 4b). These observation implies that artificial selection prefers Hap1 to Hap2 and Hap3.

MIR168a promoter haplotypes exhibit differential activity

To examine whether different haplotype confers different promoter activity, we cloned MIR168a promoters from rice accessions representing different haplotypes: NPB (Hap1); DG (Hap2); YX1 (Hap3); and CG14 (mixed haplotype, which refers to a rare MIR168a promoter variant that cannot be categorized as Hap1, Hap2, or Hap3, was excluded during haplotype identification due to its low frequency). As expected, dual-luciferase reporter assays revealed distinct promoter activities among haplotypes in N. benthamiana leaves after Agrobacterium infiltration at 36 h (Fig. 5a and b). Intriguingly, the promoters from DG (Hap2) and YX1 (Hap3) exhibited significantly higher LUC activity than that of Hap1. To exclude that the lower activity of Hap1 is the NPB accession-specific, we further tested the promoter activity from the other six accessions and confirmed the observation that Hap1 refers lower promoter activity (Fig. 5b). Notably, the mixed haplotype promoter (CG14) displayed activity comparable to Hap2 and Hap3 (Fig. 5b and c).

Fig. 5.

Promoter activity analysis of different haplotypes of the MIR168a promoter. a Schematic diagram of the dual-luciferase reporter system used to assess the activity of different haplotypes of the MIR168a promoter. b LUC activity in N. benthamiana leaves transformed with the reporter construct harboring different haplotypes of the MIR168a promoter. c Dual-luciferase assay to measure the activity of different MIR168a promoter haplotypes. Error bars indicate the standard deviation (SD) (n = 34). Different letters represent significant differences (P < 0.05), as determined by one-way ANOVA followed by Tukey’s test. d Relative miR168 abundance in different haplotypes. RT-qPCR was performed to assess miR168a expression across different haplotypes. Different letters indicated the significant differences (P < 0.05), as determined by one-way ANOVA followed by Tukey’s test

The data presented above clearly indicate that the MIR168a promoter activity in Hap2 and Hap3 accessions is significantly higher than in Hap1. Several accessions included in Fig. 5, characterized by genetic diversity, were previously used in rice pan-genome analyses, highlighting their rich polymorphic profiles (Qin et al. 2021). To further validate the differential promoter activities among haplotypes, we selected rice accessions strictly classified under Hap2 and Hap3 based on the RiceVarMap V2.0 database. Promoters were cloned from these accessions for functional assays. Accessions W222, W297, W026, and W166 represented Hap2, while C006, W197, W276, and W211 belonged to Hap3. LUC assays revealed that promoter activity in all Hap2 and Hap3 accessions significantly exceeded that of NPB and LTH (Fig. S7a). Subsequent experiments using MIR168a promoters from W222 and W297 (Hap2) and C006 and W197 (Hap3) confirmed that their activity was markedly higher than Hap1 accessions (Fig. S7b).

To test whether promoter activity determines miR168a expression, we conducted stem-loop RT-qPCR to quantify miR168a abundance across accessions representing three haplotypes and a mixed haplotype (Fig. 5d). Consistent with the luciferase assay results (Fig. 5c and Fig. S7), miR168a levels were significantly higher in Hap2, Hap3, and mixed haplotype than in Hap1 (Fig. 5d). Similarly, pri-miR168a expression was also significantly elevated in Hap2, Hap3, and mixed haplotype compared to Hap1 (Fig. S8). These findings suggest that the higher activity of the MIR168a promoter correlates with higher miR168a expression.

The “key region” of the MIR168a promoter is essential for maximal promoter activity

To pinpoint specific regions within the MIR168a promoter responsible for its activity, we performed a detailed segmentation analysis. The promoter was partitioned based on two criteria: proximity of SNPs or InDels to miR168a and sequence differences in these polymorphic sites between Hap1 and Hap2/Hap3. Through this analysis, position 164 (marked with an asterisk in Fig. 4a) was identified as a critical site and designated as the ‘‘key region’’ (KR), based on the criteria described above. Using KR as a boundary, the promoter was divided into three segments (Fig. 6a). Since MIR168a promoter sequences are nearly identical among accessions within the same haplotype, with minimal variations that do not significantly affect promoter activity (Fig. 4 and Fig. 5), we selected accessions LTH (Hap1), DG (Hap2), and YX1 (Hap3) for subsequent functional validation.

Fig. 6.

The ‘‘Key Region’’ in the MIR168a promoter affects promoter activity. a Schematic diagram of the MIR168a promoter truncation swapping, where KR represents the ‘‘Key Region’’ of the promoter. The number above each truncation indicates the length scale. b, d Dual-luciferase reporter system was used to assess the activity of the indicated MIR168a promoter. The upper panel (b) shows the truncation swapping between Hap1 and Hap2, while the lower panel (d) represents the truncation swapping between Hap1 and Hap3 (c, e) Quantitative LUC/RLUC ratio analysis of the results from panels (b) and (d). Error bars indicate the standard deviation (SD) (n = 8 and 14, respectively). Different letters indicated significant differences (P < 0.05), as determined by one-way ANOVA followed by Tukey’s test.

To test the function of KR, we swapped promoter fragments between Hap1 (LTH) and Hap2 (DG) or Hap3 (YX1), generating eight distinct MIR168a promoter constructs (Fig. 6a). Dual-luciferase reporter assays performed in N. benthamiana leaves following Agrobacterium infiltration revealed that the activity of constructs (DG-KR)-LTH and LTH-(DG-KR) was consistent with DG and significantly higher than LTH, (LTH-KR)-DG, and DG-(LTH-KR) (Fig. 6b and c). No significant differences were observed among LTH, (LTH-KR)-DG, and DG-(LTH-KR), confirming that KR plays a pivotal role in the elevated promoter activity of Hap2.

Similarly, for Hap3, the constructs (YX1-KR)-LTH and LTH-(YX1-KR) exhibited promoter activity levels comparable to YX1, both significantly higher than LTH, (LTH-KR)-YX1, and YX1-(LTH-KR) (Fig. 6d and e). Again, no significant differences were detected among LTH, (LTH-KR)-YX1, and YX1-(LTH-KR), reinforcing the essential role of KR in driving Hap3 promoter activity. These findings collectively demonstrate that the KR within the MIR168a promoter is a critical determinant of its high expression activity. This key regulatory region likely underpins the enhanced promoter-driven miR168a expression observed in Hap2, Hap3, and mixed haplotype.

The T site in Hap2 and the A site in Hap3 are critical regulatory elements influencing MIR168a promoter activity

Given the importance of the KR in regulating MIR168a promoter activity, we hypothesized that specific SNPs or InDels within the KR in Hap2 and Hap3 are responsible for differences in expression activity. To identify the key regulatory sites, we conducted a comparative sequence analysis of the KR among three haplotypes. This revealed SNPs and InDels at four distinct locations, designated as site 1, site 2, site 3, and site 4 (Fig. S9b). Since no significant differences were observed in promoter activity among Hap1 accessions, including NPB, N22, Basmati1, Y3551, NamRoo, G8, and LTH (Fig. 5), we ruled out the involvement of site 1 and site 2 in regulating promoter activity. Instead, we focused on the “TT” sequence at site 3 (designated as the T site) and the “A” at site 4 (designated as the A site), hypothesizing that these elements influence promoter activity in Hap2 and Hap3, respectively (Fig. S9b).

To validate the role of the T site in Hap2, we engineered mutations by substituting “CC” with “TT” in the KR of the MIR168a promoters from LTH and WSSM and replacing “TT” with “CC” in DG and CG14. These modified promoters were named Prolth, Prowssm, Prodg, and Procg14, respectively (Fig. 7a and b). Dual-luciferase reporter assays performed in N. benthamiana leaves following Agrobacterium infiltration showed that Pro_lth and Pro_wssm exhibited significantly enhanced activity compared to the original Pro_LTH and Pro_WSSM (Fig. 7c). Conversely, Pro_dg and Pro_cg14, in which the T site was removed, showed substantially reduced activity compared to the native MIR168a promoters from DG and CG14, respectively (Fig. 7c). Quantitative analysis further demonstrated that the LUC/RLUC ratio for Pro_lth and Pro_wssm was significantly higher than their respective wild-type promoters, while the ratio for Pro_dg and Pro_cg14 was markedly lower than their unmutated counterparts (Fig. 7d). These findings establish the T site as a critical determinant of MIR168a promoter activity in Hap2, with its mutation leading to a significant decrease in activity.

Fig. 7.

The T and A sites in the MIR168a promoter impact its activity. a A schematic diagram illustrating the detection of MIR168a promoter activity after mutation of the T and A sites using the dual-luciferase reporter system. b Sequence alignment diagram after mutating “CC” to “TT” in Hap1 and mutating “TT” to “CC” in Hap2. c Dual-luciferase reporter assay detects the expression activity of different types of MIR168a promoters in (b). d Quantitative LUC/RLUC ratio analysis in (c). e Sequence alignment diagram after mutating “CCCC” to “A” in Hap1 and mutating “A” to “CCCC” in Hap2. f Dual-luciferase reporter assay detects the expression activity of different types of MIR168a promoters in (E). g Quantitative analysis of LUC/RLUC ratio in (f). In (d) and (g), error bars indicate the standard deviation (SD) (n = 4 and 16, respectively). Significant differences (P < 0.05) were determined by one-way ANOVA followed by Tukey’s test

Similarly, to investigate the A site’s impact on MIR168a promoter activity in Hap3, we mutated “CCCC” to “A” in the KR of the LTH promoter and reversed the mutation, changing “A” to “CCCC,” in YX1 (Fig. 7e). The resulting constructs were tested using LUC reporter assays. The promoter activity of LTH (CCCC to A) was significantly higher than that of the unmutated LTH, while the activity of YX1 (A to CCCC) was significantly reduced compared to YX1 (Fig. 7f). Quantitative analysis confirmed these results, with LTH (CCCC to A) exhibiting markedly enhanced activity and YX1 (A to CCCC) showing a dramatic reduction in activity (Fig. 7g). These results demonstrate that converting “CCCC” to “A” in LTH significantly boosts promoter activity, while the reverse mutation in YX1 diminishes it. These findings reveal that the T site within the KR is essential for the enhanced promoter activity observed in Hap2, while the A site plays an equally critical role in Hap3. These key regulatory elements underscore the functional significance of KR in modulating the expression activity of the MIR168a promoter across rice haplotypes.

The T and A sites of the MIR168a promoter originate from wild-type rice

To examine the distribution of different MIR168a promoter haplotypes across rice subspecies, including Ind, TeJ, TrJ, Jap, Aus, Aro, and Mix, we conducted a bioinformatic analysis using accession data from various regions (http://ricevarmap.ncpgr.cn/). The results revealed that Hap1, which contains the C site, is present in all the examined subspecies. In contrast, Hap2 (containing the T site) and Hap3 (containing the A site) are predominantly distributed in Ind, with only a few instances observed in Mix (Fig. 8a).

Fig. 8.

The phylogenetic relationship and genetic difference of the MIR168a promoter. a Distribution of MIR168a promoter of different haplotypes in rice subspecies; Ind, indica; TeJ, temperate japonica; TrJ, tropical japonica; Jap, japonica; Aro, aromatic; Mix, admixed. b Nucleotide diversity and selection analyses in the MIR168a promoter and flanking regions (~ 100 kb). Genomic sequences of cultivated rice accessions and 33 wild accessions were obtained from RiceVarMap v2.0 (http://ricevarmap.ncpgr.cn/), RiceSuperPIRdb (http://www.ricesuperpir.com/), and Rice Resource Center (http://113.54.11.180/). c Phylogenetic relationship of the MIR168a promoter generated from cultivated rice accessions and 33 wild rice accessions d Haplotype network of the MIR168a promoter. The size of the circle is proportional to the number of samples for each haplotype. Black dots on the lines represent mutational steps between haplotypes.

To determine whether domestication influenced the MIR168a promoter through artificial selection, we analyzed genome data from cultivated rice varieties and 33 wild rice accessions, focusing on the genetic diversity of the MIR168a promoter and its flanking regions. In wild rice, nucleotide diversity for the MIR168a promoter was significantly higher than in cultivated rice, where nucleotide diversity was notably reduced in subspecies such as TeJ, TrJ, Jap, Aus, Aro, and Mix (Fig. 8b). This reduced diversity in cultivated rice likely reflects artificial selection, which may be linked to the negative regulation of rice blast resistance and agronomic traits by miR168a (Wang et al. 2021).

To further trace the origins of the T and A sites, we analyzed genomic sequences from 20 wild rice accessions and compared them with the KR region of MIR168a promoter haplotypes. Both the T and A sites were identified in the MIR168a promoter of wild rice (Fig. S10). Phylogenetic analysis confirmed that these sites originated from wild rice (Fig. 8c). Additionally, a haplotype network was constructed to describe the evolutionary relationships and mutation steps of MIR168a promoter haplotypes (Fig. 8d). Wild rice haplotypes were categorized into Hap4, Hap5, and Hap6 (Fig. S11), showing that the T and A sites in the MIR168a promoter were derived from distinct wild rice lineages. These findings conclusively indicate that the T and A sites in the MIR168a promoter trace their origins to wild rice.

miR168a is responsive to blast infection across all rice haplotypes

Given the MIR168a promoter's responsiveness to PAMP treatment and its haplotype-specific expression activity (Figs. 2 and 5), we investigated whether miR168a expression shows haplotype-specific responses to M. oryzae infection. To address this, we selected rice accessions representing Hap1, Hap2, and Hap3 and measured miR168a expression using stem-loop RT-qPCR. Across all analyzed accessions, miR168a expression was significantly induced at different time points post-inoculation, with the most pronounced increases observed in YX1 (Fig. 9).

Fig. 9.

Relative abundance of miR168 in different haplotypes upon rice blast infection. a–i Stem-loop RT-qPCR was performed to analyze the expression pattern of miR168a in N22 (a), Y3551 (b), G8 (c), DG (d), W166 (e), W222 (f), YX1 (g), W197 (h), and W211 (i). Among them, N22, Y3551, and G8 belong to Hap1. DG, W166, and W222 belong to Hap2. YX1, W197, and W211 belong to Hap3. Different letters indicated the significant differences (P < 0.05), determined by one-way ANOVA followed by Tukey’s test

In Hap1 accessions, the miR168a expression level was highest at 72 hpi than at other time points in N22 (Fig. 9a). Among other Hap1 accessions, peak expression occurred at 24 hpi in Y3551 and 12 hpi in G8 (Fig. 9b and c). For Hap2, miR168a expression was significantly induced at 24 and 48 hpi in DG (Fig. 9d); at 12, 24, 48, and 72 hpi in W166 (Fig. 9e); at 24 and 72 hpi in W222 (Fig. 9f). Similarly, in Hap3, miR168a expression peaked at 72 hpi in YX1 and W197 (Fig. 9g and h), at 12 and 24 hpi in W211 (Fig. 9i). Similarly, pri-miR168a expression was significantly induced upon M. oryzae infection at the indicated time points across all tested haplotypes (Fig. S12). In addition, we evaluated the resistance of rice accessions carrying different MIR168a promoter haplotypes to the M. oryzae strain GZ8. The results showed that LTH, NPB, N22, and W222 were susceptible to GZ8, whereas all other tested accessions displayed resistance (Fig. S13). Accordingly, GZ8 represents a compatible strain for LTH, NPB, N22, and W222, but an incompatible strain for the remaining accessions. Notably, GZ8 infection induced miR168a expression across all haplotypes under both compatible and incompatible interactions (Fig. 1d–f and Fig. 9). These results indicate that, despite variations in the timing and magnitude of miR168a induction among haplotypes, blast infection universally triggered its upregulation, suggesting that miR168a induction by M. oryzae is not haplotype-specific.

Discussion

Rice blast is one of the most devastating fungal diseases threatening global rice production. The most efficient and cost-effective strategy to combat this disease is deploying resistance mediated by NLR (nucleotide-binding leucine-rich repeat-containing) genes (Li et al. 2019a). In addition, miRNAs play a critical role in regulating rice immunity by modulating PTI and ETI signaling through their diverse target genes. This regulation ensures a balance between disease resistance and growth (Feng et al. 2021; Zhao et al. 2023). Among these, miR168a is a key regulator that negatively affects rice blast resistance. Overexpression of miR168a leads to diminished resistance to rice blast, delayed flowering, and reduced yield. Conversely, suppressing miR168a abundance using artificial target mimics enhances rice blast resistance, shortens flowering time, and improves yield (Wang et al. 2021). Similarly, knocking out MIR168a results in phenotypes resembling MIM168 plants, increased tillering, and earlier flowering (Zhou et al. 2022). These findings highlight miR168a's pivotal role in finely tuning rice immunity and agronomic traits.

Our study showed that MIR168a responds to both PAMP and M. oryzae treatments (Fig. 1), suggesting its involvement in rice immune response to biotic stress. The MIR168a promoter is active in both roots and leaves and is similarly inducible by PAMPs (Fig. 2). Based on nucleotide polymorphisms in the MIR168a promoter, we classified it into three haplotypes (Fig. 4). Promoter activity was significantly higher in Hap2 and Hap3 than in Hap1, correlating with elevated miR168a expression levels in these haplotypes (Fig. 5d). Further analysis revealed that the T site in Hap2 and the A site in Hap3 are key determinants of promoter activity, with mutations at these sites causing a marked reduction in activity (Figs. 6 and 7). These haplotypes, containing the T or A sites, were predominantly identified in indica rice, and our findings suggest these sites originated from wild rice (Fig. 8). Lastly, miR168a expression was consistently induced across all three haplotypes following M. oryzae treatment (Fig. 9).

The promoter of MIR168a is polymorphic, while mature miR168a is highly conserved

miRNAs are loaded into AGO1 to form the RISC complex, which regulates gene expression at transcriptional and post-transcriptional levels, influencing diverse biological processes in eukaryotes (Zhao et al. 2023). miR168, a plant-specific small RNA, targets and cleaves AGO1 mRNA. In turn, AGO1 stabilizes miR168 by incorporating it into the RISC complex (Vaucheret et al. 2006). This intricate self-regulatory circuit between miR168 and AGO1 is essential for the proper functioning of the miRNA pathway and plays a pivotal role in plant growth, development, and disease resistance.

Mature miR168 is highly conserved through evolution. Previous studies have shown that, in Brassicaceae plants, the sequence, secondary structure, and thermodynamic stability of miR168 are remarkably conserved (Gazzani et al. 2009). Consistent with these findings, our results (Fig. S3 and Fig. S4) demonstrate that mature miR168a is completely identical across all rice subspecies, with no observed nucleotide variations (Fig. S9a). This high conservation likely reflects the critical role of miR168 in targeting AGO1 mRNA and regulating the miRNA network. Although AGO1 is present across various organisms (Fang & Qi 2016; Zhao et al. 2023), miR168 remains uniquely plant-specific.

In contrast to the conserved mature miR168, the MIR168 promoter exhibits significant variability (Fig. S5 and S6). Polymorphisms in the promoter region can affect gene expression by altering promoter activity and transcription factor binding (Wu et al. 2014; Li et al. 2017). For example, studies on the ABCB1 gene have shown that promoter haplotypes can influence promoter activity by modifying transcription factor interactions (Speidel et al. 2018). To assess the impact of polymorphism on the MIR168a promoter, we conducted a comprehensive analysis across rice subspecies. Our findings revealed that the MIR168a promoter can be grouped into three haplotypes. Hap2 and Hap3 display greater polymorphism than Hap1 (Fig. 4a), and promoter activity varies significantly among haplotypes (Fig. 5 and Fig. S7).

Together, these results illustrate that while mature miR168a remains highly conserved across rice subspecies, the MIR168a promoter is highly polymorphic. This promoter variability drives differences in miR168a expression among rice accessions with distinct haplotypes (Fig. 5d). Thus, polymorphisms in the MIR168a promoter are closely linked to the regulation of miR168a abundance.

The loss of T or A sites in the MIR168a promoter may be attributed to artificial selection and domestication

In plants, gene expression is a dynamic process that varies across developmental stages, tissues, and environmental conditions, thereby regulating growth, development, and adaptability. Promoters play a central role in controlling gene expression by determining initiation, termination, and transcription levels. A typical promoter consists of a core promoter and upstream cis-regulatory elements (CREs) (Tang and Zhang 2023). The core promoter interacts with RNA polymerase and transcription factors at the transcription start site to initiate gene expression, while upstream CREs bind gene-specific transcription factors, known as trans-acting factors (TREs) (Tang and Zhang 2023). These interactions enable precise gene expression regulation.

Natural variations in CREs can influence transcription factor interactions, altering promoter activity and affecting plant traits. For instance, in rice, an SNP in the Bsr-d1 promoter differentiates LTH and DG accessions. This variation enables MYBS1 to bind to the bsr-d1 promoter from DG, reducing its activity and inhibiting hydrogen peroxide degradation, thereby enhancing blast resistance in DG (Li et al. 2017). Similarly, variations in the Stay-Green (OsSGR) promoter accelerate senescence in indica rice by inducing earlier and stronger OsSGR expression (Shin et al. 2020). Additionally, variations in the GSE5 promoter define haplotypes linked to grain size diversity, with certain haplotypes reducing GSE5 expression and producing wider grains (Duan et al. 2017). These examples underscore how natural promoter variation affects growth, development, and stress responses in plants.

Analysis of MIR168a promoter haplotypes revealed that Hap2 and Hap3 exhibit significantly higher expression activity compared to Hap1 (Fig. 5 and Fig. S7). Notably, the presence of a ‘‘TT’’ or ‘‘A’’ at position -164 in the promoters of Hap2 and Hap3, respectively, enhances promoter activity (Figs. 6 and 7). We hypothesize that these T and A sites enable binding to an unidentified transcriptional activator, thereby increasing expression levels. However, our sequence analysis using the PlantPAN 4.0 database (http://plantpan.itps.ncku.edu.tw/plantpan4/index.html) did not identify known transcription factors that bind to this motif. This suggests that the position -164 sequence may interact with unknown or uncharacterized transcription factors, warranting further investigation.

Additionally, the high-order structure of the promoter might influence its activity. For example, G-quadruplex (G4) secondary structures in promoters are known to regulate gene expression (Robinson et al. 2021). It is possible that the "TT" or "A" at position -164 alters the higher-order structure of the MIR168a promoter, enhancing its expression activity. Further research is needed to elucidate these mechanisms. Exploring the interactions between position -164 and potential transcription factors, or analyzing structural variations in the MIR168a promoter across Hap2 and Hap3, could provide valuable insights into the regulation of MIR168a expression.

Rice accessions corresponding to Hap2 and Hap3 were predominantly from the indica subspecies (Fig. 8a). In contrast, the total number of rice accessions with Hap1 exceeded those with Hap2 and Hap3 (Figs. 4b and 8c, d). Previous studies have shown that reducing miR168a abundance enhances resistance to blast disease while simultaneously improving several yield-related agronomic traits (Wang et al. 2021). This suggests that Hap2 and Hap3, which carry the T or A sites in the MIR168a promoter, were less favored during artificial domestication. Breeders likely prioritized rice lines with reduced MIR168a promoter activity to balance blast resistance and desirable agronomic traits.

This hypothesis is further supported by nucleotide diversity data, as the nucleotide diversity of the MIR168a promoter in cultivated rice is significantly lower than in wild rice (Fig. 8b). The loss of T or A sites in the MIR168a promoter, therefore, appears to be a consequence of artificial selection. These findings highlight how human-mediated selection can shape the genetic architecture of promoter regions, influencing plant traits over evolutionary timescales.

However, elucidating how distinct haplotypes of the MIR168a promoter modulate miR168a expression to influence OsAGO1 transcript levels and rice blast resistance remains a scientifically significant yet technically challenging endeavor. This complexity arises from two principal obstacles. First, robust resistance to M. oryzae in rice is typically governed by major resistance (R) genes (Huang et al. 2023), whereas microRNAs such as miR168a are generally considered to function as fine-tuners of the immune response (Li et al. 2019b). Second, the extensive genetic diversity among rice accessions precludes direct functional comparisons. To overcome these limitations, two experimental strategies may be employed in future work: (1) generating near‑isogenic lines (NILs) that differ only at the MIR168a promoter through repeated backcrossing—a time‑ and labour‑intensive process; and (2) employing precision genome‑editing techniques to create rice lines with identical genetic backgrounds but distinct MIR168a promoter haplotypes. Given that the expression activities of different MIR168a promoter haplotypes were evaluated using transient assays in N. benthamiana, future work will focus on generating transgenic rice lines harboring each promoter haplotype within an identical genetic background, coupled with validation of haplotype-specific promoter activity in rice protoplasts.

Editing the MIR168a promoter may enhance rice resistance and yield

CRISPR/Cas9 gene-editing technology has transformed functional genomics research (Gosavi et al. 2020). Unlike modifications to a gene’s coding region, editing its promoter leaves the gene-encoded product intact, avoiding the risk of extreme phenotypes caused by loss or alteration of gene function. This approach offers a promising strategy for exploring gene functional diversity and is particularly advantageous for breeding programs. Populations with edited promoters exhibiting variable expression activities can generate germplasm with polymorphic quantitative traits or novel phenotypes derived from the same genes. For instance, targeted promoter editing has proven effective in developing disease-resistant crops. In rice, CRISPR/Cas9 has been used to modify the EBE region of the OsSWEET promoter, which is recognized by transcription activator-like effectors of pathogens, thereby conferring broad-spectrum resistance to bacterial blight (Oliva et al. 2019).

Rice yield depends on three core factors: panicle number, grains per panicle, and grain weight (Yin et al. 2021). However, these traits often exhibit trade-offs. For example, Ideal Plant Architecture 1 (IPA1), a gene with pleiotropic effects, enhances grain number per panicle but reduces tiller number, limiting its yield potential. Recent findings by Song et al. (2022) offer a solution to this trade-off. Screening for regulatory region deletions in the IPA1 promoter identified a mutant (IPA1-pro10) with a 54-bp deletion, enabling simultaneous increases in tiller number, grains per panicle, and panicle weight, thereby significantly enhancing grain yield (Song et al. 2022).

In our research, we found that MIR168a plays a key role in regulating disease resistance, flowering time, and yield traits. Suppressing MIR168a expression enhanced disease resistance, improved yield, and shortened flowering time (Wang et al. 2021). Additionally, our analysis revealed that the presence of T or A sites in the MIR168a promoter of Hap2 and Hap3 enhances its activity (Figs. 6 and 7). This finding highlights the potential of molecular breeding techniques to edit the MIR168a promoter in elite rice lines for developing superior germplasm resources. By generating promoter-edited populations with varying MIR168a activity, it is possible to achieve differential miR168 expression, leading to lines with improved disease resistance and yield.

Conclusion

In this study, we identified a 1661-bp upstream fragment of MIR168a with the highest promoter activity. Analysis of over 4,000 rice accessions revealed extensive polymorphisms in the MIR168a promoter, which were classified into three haplotypes, with Hap2 and Hap3 showing significantly higher activity than Hap1. We found that specific nucleotide sites, “TT” in Hap2 and “A” in Hap3, play a critical role in enhancing promoter activity and increasing miR168a abundance. Notably, these active promoters are predominantly found in indica accessions and trace their origins to wild rice, highlighting their potential for breeding applications.

Abbreviations

miRNA

MicroRNA

MIR

MiRNA genes

Pol II

RNA polymerase II

DCL

DICER-LIKE

AGO

Argonaute

RISC

RNA-induced silencing complex

SPL7

SQUAMOSA PROMOTER BINDING PROTEIN-LIKE7

ABA

Abscisic acid

RREB1

Ras-responsive element-binding protein

LDLRAP1

Low-density lipoprotein receptor adapter protein 1

LTH

Lijiangxintuanheigu

NPB

Nipponbare

Y3551

Yihui3551

G8

Guang 8B

WSSM

Wushansimiao

DG

Digu

YX1

Yixiang 1B

LUC

Luciferase

gDNA

Genomic DNA

OD

Optical density

RLUC

Renilla luciferase

OTA

Oatmeal and tomato agar

RT-qPCR

Reverse transcription quantitative PCR

GUS

β-Glucuronidase

MUG

4-Methylumbelliferyl-β-D-glucuronide

Hap

Haplotype

CREs

cis-regulatory elements

TREs

Trans-acting factors

PAMPs

Pathogen‐associated Molecular Patterns

Author contributions

Z.-X. Z. and W.-M. W. conceived and designed the study. Z.-X. Z., X.-L. Y., R. L., S.-J. Y., X.-X. Y., and Y.-P. T. performed the experiments. Z.-X. Z. and D.-M. G. were responsible for bioinformatics analysis. J. F., Y.-Y. H., J.-W. Z., H. W., G.-B. L., and Y. L. analyzed the data. M.-L. G., Y. L., and D.-Q. L. collected the references. Z.-X. Z., B. H., and W.-M. W. wrote the manuscript. All authors discussed the results and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (32101728 to Z.-X. Z., 32350410431 to B. H., U19A2033 to W.-M. W., and 32121003), the Open Project Program (SKL-KF202423) of State Key Laboratory of Crop Gene Exploration and Utilization in Southwest China, and Sichuan Province Breeding Project (2021YFYZ0021).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.