Fungal pathogen isolates and host plants

Maintenance of M. oryzae and medium composition are as described previously³⁸. The M. oryzae wild-type strain Guy11 isolate was grown on complete medium in a growth chamber at 25 °C in a 12-h light–12-h dark photoperiod for 10 days before conidia were collected for inoculation of rice plants. Mycelial growth of M. oryzae strains in different stress conditions was measured on complete agar medium supplemented with or without chemicals for 10 days. The chemicals included 5 mM H₂O₂, which was supplemented to induce oxidative stress; 200 μg ml⁻¹ congo red or calcofluor white, which was used to induce cell wall stress; and 0.01% SDS, which was used to induce cell membrane stress.

Rice infection with M. oryzae was conducted as previously described³⁷. Conidial suspensions (1 × 10⁵ spores ml⁻¹) of M. oryzae strains were used to inoculate rice and incubated for 5–7 days before being photographed. The relative fungal DNA amount was calculated using the threshold cycle value (C_T) of M. oryzae POT2 DNA (MoPOT2) against C_T of rice ubiquitin genomic DNA³⁹. The infection-related development of M. oryzae was examined following previous reports⁴⁰ at indicated hpi with M. oryzae strains. The frequency of plant penetration was assessed by counting the formation of appressoria and/or penetration pegs on leaf sheaths of TP309 rice japonica cultivar or its transgenic lines from at least 50 spores in triplicate on an Axio Imager A2 microscope (Zeiss). For infection-related time-course analyses, hpi denotes hours after inoculation with M. oryzae strains.

R. solani AG1 strain was grown on potato dextrose agar plates at 28 °C before mycelia were collected for inoculation according to previous reports⁴¹. The wild-type F. graminearum PH-1 strain was routinely cultured on potato dextrose agar plates at 25 °C. Construction of F. graminearum deletion mutants and infection assays with wheat seedlings and spikes were conducted following previous reports⁴². The rice TP309 cultivar was inoculated with M. oryzae and R. solani, whereas wheat Nanmai 660 cultivar was inoculated with F. graminearum.

DNA and RNA manipulation

DNA was extracted from samples of fungus or plant using standard cetyltrimethylammonium bromide method according to a previous report⁴³. RNA isolation was conducted as described previously³⁹. In brief, samples were ground into powder with liquid nitrogen and transferred to 2-ml RNase-free tubes (approximately 100 mg per tube). TRIzol (Invitrogen; 15596026CN) reagent was then used for total RNA extraction according to the manufacturer’s instructions. Total RNA of M. oryzae was extracted from samples of vegetative mycelia grown in liquid complete medium or minimal medium, conidiophore on complete medium, appressorium and invasive hypha colonizing leaves. Reverse transcription was conducted with the extracted total RNA using the HiScript III RT SuperMix for qPCR (plus gDNA wiper) kit (Vazyme; R323-01), followed by RT–qPCR amplification with the ChamQ SYBR Color qPCR Master Mix kit (Vazyme; Q411) on a NextGene QL96 instrument (Jie Lai Mei Technology). Specific procedures were conducted according to the manufacturer’s instructions.

RNA-seq analysis of M. oryzae

The M. oryzae wild-type Guy11 strain was allowed to grow under four different conditions before extraction of fungal total RNA using an RNeasy Plant Mini Kit (QIAGEN; 74904). Vegetative mycelia of M. oryzae were sampled from Guy11 grown in liquid complete medium for 48 h; conidia were sampled from Guy11 grown on complete agar medium for 10 days; the appressoria were sampled from Guy11 undergoing infection structure development on hydrophobic borosilicate glass coverslips (Thermo Fisher Scientific; 12-541-B); invasive hyphae were sampled from Guy11 that has infected rice leaf sheaths for 4 days. Two biological replicate samples were prepared for each growth condition. RNA-seq library construction and sequencing analysis were conducted as previously described³⁹. Differentially expressed genes were identified on the basis of an adjusted P value 44.

Generation of M. oryzae mutants

A PCR-based split-marker deletion method was used for targeted gene deletion of lnc117761 as previously described³⁹. A 600-bp DNA region of lnc117761 was targeted for replacement by the selection marker gene HPT, which confers resistance to hygromycin B (Roche; 10843555001). For genetic complementation analysis, the lnc117761 full-length gene expression vector containing its native promoter was constructed in a plasmid containing bialaphos resistance selection marker gene BAR, using a ClonExpress II One Step Cloning Kit (Vazyme; C112). To examine the non-coding property of lnc117761, a complementation plasmid was constructed that contains an lnc117761 variant with the potential short open reading frame in lnc117761 disrupted by shift mutation using the PCR-based point mutation method. Using a similar PCR-based mutation method, another variant of lnc117761 complementation plasmid was generated to disrupt the 16 nt of lnc117761 that base-pair with rice miR5827. PCR-based point mutation was also performed to introduce single-nucleotide replacement in MBS of lnc117761 to produce different lnc117761 variants. The resulting lnc117761 constructs carrying nucleotide mutations were separately transformed into Δlnc117761 for functional complementation analysis (Supplementary Fig. 1). Supplementary Table 8 details the information for all primers, PCR templates and amplicons.

Generation of transgenic rice lines

A CRISPR–Cas9 construction method was used to produce rice lines with miR5827 disrupted (Supplementary Fig. 2). Two 23-nt oligonucleotides (including a protospacer-adjacent motif) targeting the OsmiR5827 precursor near its 5′ and 3′ ends were chosen to delete most of the precursor, and the specificity of the targeting sequences was confirmed by a BLAST search against the rice genome (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Construction of the plasmid and Agrobacterium tumefaciens-mediated transformation of rice were performed according to a previous report³⁷. Genomic DNA samples were isolated from transgenic rice lines following the standard cetyltrimethylammonium bromide method⁴³ and then were used for PCR amplification to confirm the targeted deletion of miR5827-encoding DNA region by sequencing analysis with primer pairs flanking the designed target site. The rice transgenic lines were analysed by PCR to detect the CRISPR–Cas9 construct, and those harbouring edited sequence in the miR5827-encoding DNA region but without the CRISPR–Cas9 insertion in genome were used for phenotypic analyses. The oligonucleotide sequences chosen as guide RNA were listed in Supplementary Table 8.

Rice lines with PKR1 disrupted were also generated through CRISPR–Cas9 method using a 23-nt oligonucleotide as guide RNA. To produce rice transgenic lines PKR1KO–compl.PKR1^mMBS, a PKR1 mutant construct was generated by site-directed mutagenesis in vector expressing full-length wild-type PKR1. MBS (GGTAGTGCTAAATTGTAACGCA) in wild-type PKR1 was changed to a mutant sequence (TTCCACGTTCCCCCTTCCTGCC) to abolish the miR5827-targeting site in PKR1. The PKR1 mutant construct was subsequently introduced into a PKR1KO rice line through A. tumefaciens-mediated transformation.

To overexpress M. oryzae lnc117761 in rice, lnc117761 complementary DNA (cDNA) under control of the 35S promoter of the cauliflower mosaic virus (CaMV 35S) was cloned to binary vector pCAMBIA1300 (Cambia), which contains the HPT selection marker gene. A variant of lnc117761 overexpression plasmid was similarly generated in pCAMBIA1300 to disrupt the 16 nt that base-pair with rice miR5827. The resulting plasmids were separately introduced into rice by A. tumefaciens-mediated transformation. Rice grain length, grain width and 1,000-grain weight were measured using a Mini 1600 automatic analysis system (Jie Lai Mei Technology).

Analysis of plant miRNA and gene expression levels

In the analysis of miRNA level, total RNA was extracted from different plant samples, including leaf and stem tissues from barley (H. vulgare), A. thaliana, potato (S. tuberosum), B. distachyon and wheat (T. aestivum). In brief, the plants were grown in a growth chamber at 25 °C for approximately 30 days before collecting their respective tissues. For each sample, 100 mg of fresh tissue was ground into powder with liquid nitrogen, transferred to a 2-ml RNase-free tube and then processed for total RNA extraction using TRIzol (Invitrogen; 15596026CN). Total RNA was reverse transcribed using miRNA-specific stem-loop reverse transcription primers (Supplementary Table 8) with the miRNA 1st Strand cDNA Synthesis Kit (Vazyme; MR101). The reverse transcription products were subsequently used as a template for qPCR by using miRNA-specific forward primers and the universal reverse primer in miRNA Universal SYBR qPCR Master Mix (Vazyme; MQ101). The U6 RNA served as an internal reference for the detection of miRNAs.

For measuring gene expression, RT–qPCR was performed with the HiScript III RT SuperMix for qPCR (+gDNA wiper) kit (Vazyme; R323-01) and ChamQ SYBR Color qPCR Master Mix kit (Vazyme; Q411). The UBIQUITIN gene (Os03g13170) was used as an internal reference for normalization of gene expression in rice. The potential target genes of miR5827 in rice were determined by psRNATarget tool, and the target genes predicted with high fidelity were selected for confirmation by RT–qPCR analysis with rice total RNA. Following the Minimum Information for Publication of Quantitative Real-Time PCR Experiments standards reported previously⁴⁵, the specificity of primers was examined by BLAST-searching against the plant genome to ensure primer matching to a specific genome locus and by monitoring the melting curve of the RT–qPCR product to ensure a single amplification fragment through single-peak dissociation profiles. Each RT–qPCR assay was repeated with at least two more biological replicates independently with three technical replicates per sample. The 2^−∆∆CT method was used to calculate the relative expression levels with three technical replicates.

In situ hybridization assay

In situ hybridization of lnc117761 distribution pattern was conducted following standard protocols as described previously^46,47 with minor modifications. Rice leaf sheaths were collected from TP309 cultivar grown for 4 weeks in a growth chamber and then used for inoculation with or without M. oryzae Guy11 at a conidial suspension of 2 × 10⁵ spores ml⁻¹. Thirty hours after infection in a high-humidity environment at 25 °C, infected rice leaf sheaths were processed for microscopic analysis as described previously⁴⁸. After successful infection with M. oryzae was confirmed and appressoria and invasive hyphae had developed, the infected rice sheaths were collected and sectioned for fixation with FAA (50% ethyl alcohol, 5% glacial acetic acid and 3.7% formaldehyde) and embedded in wax. Paraffin sectioning was conducted on a microtome (Thermo Fisher Scientific; HM325) to produce 8-μm paraffin sections, which were processed with standard in situ hybridization protocols⁴⁹.

For better comparison, sectioned samples from the same rice sheath were used for detection with probes targeting lnc117761 and control RNA separately using an ISH Detection Kit (Boster Biological Technology; MK1030). lnc117761-targeting probes (5′-biotin-UAUCUAAAUCGCCGACUGUCAGGUUCCUUGCCGU and 5′-biotin-CAAGUAGUUUUGAUUGCUGUGAGGACACUGCUGC) were synthesized by Sangon Biotech (Supplementary Table 8). As a negative control, a neighbouring section derived from the same rice sheath tissue was hybridized to control probes (5′-biotin-UUGUCCCAAGCAGCAGUGUCCUCACAGCAAUCA and 5′-biotin-AUGUUACCCGCCGUUUGAGCAGCACGACCUUU) that target non-transported mRNAs of two control genes (MGG_14955 and MGG_09330, respectively). Meanwhile, an uninfected control sample was included by using rice leaf sheaths without M. oryzae inoculation for hybridization with the same probes. After overnight hybridization at 42 °C, washing and blocking, the section slides were incubated with HRP-labelled streptavidin (Beyotime Biotechnology; A0305) at 37 °C for 30 min. Finally, DAB Horseradish Peroxidase Color Development Kit (Servicebio; G1212) was used to develop brown colour showing the hybridized RNA; meanwhile, blue colouration of cell walls was conducted by haematoxylin counterstaining. Images were captured on Pannoramic scanner (3DHISTECH Pannoramic MIDI II).

Detection of lnc117761 secretion

We used different approaches to investigate lnc117761 trafficking from M. oryzae to rice. In the first approach, rice leaves at and surrounding the infection site were separately sampled for RT–qPCR analysis to assess lnc117761 distribution within rice tissues following a previous report⁵⁰ with minor revisions. We reasoned that if lnc117761 is delivered to infected host cells, it could be transported to surrounding cells, and its presence could be detected by RT–qPCR in neighbouring rice cells before the appearance of fungal cells. In brief, a Guy11 strain constitutively expressing GFP was used to infect detached rice leaves by punch inoculation as previously described⁵¹. Three different concentrations (4 × 10⁴ spores ml⁻¹, 2 × 10⁵ spores ml⁻¹ and 1 × 10⁶ spores ml⁻¹) of M. oryzae spore suspension were used separately to inoculate rice cells. After infection, the infected rice leaves were observed at different time points under a fluorescence microscope to separately collect infected samples (sample 1) from infection sites that contained M. oryzae invasive hyphae and from surrounding tissues (sample 2) that contained no M. oryzae. Thus, sample 1 contains infected rice cells and close neighbouring uninfected cells, whereas sample 2 contains relatively distal surrounding rice cells to avoid M. oryzae contamination. The collected rice leaf segments were then used for extraction of total RNA using TRIzol reagent. RT–qPCR was subsequently performed as described above with the rice UBIQUITIN gene as the endogenous control.

In the second approach, extracellular vesicles of M. oryzae were prepared to assess lnc117761 distribution. In brief, spore suspension of wild-type M. oryzae Guy11 and Δlnc117761 mutant (2 × 10⁵ spores ml⁻¹) was inoculated into liquid complete medium to allow growth for 48 h. Mycelia were collected, washed with distilled water and transferred to minimal liquid medium supplemented with ground powder of rice leaves. Following 24 h of growth at 25 °C and 120 revolutions per minute, the culture was centrifuged at 10,000g for 10 min at 4 °C to collect the supernatant after removing fungal mycelia and rice leaf debris. Extracellular vesicles were then separately isolated from the supernatant of mycelial culture following a previously reported protocol⁵². Supernatant was filtered through a 0.45-μm membrane filter. The filtrate was centrifuged again at 10,000g for 10 min at 4 °C. The supernatant was then subjected to ultracentrifugation at 100,000g for 1 h at 4 °C. The supernatant was discarded, and the vesicle pellet was resuspended in 200 μl of diethyl pyrocarbonate (DEPC)-treated water. Total RNA was extracted from the prepared vesicles and analysed by RT–qPCR as described above.

In the third approach, GFP-expressing Δlnc117761 strain was transformed with an lnc117761–4×Pepper fusion construct, in which 4×Pepper encodes an RNA motif capable of binding to and activating fluorescent dye Pepper620 (ref. ²⁷). This strain expressing lnc117761–4×Pepper was used to inoculate rice leaf sheaths, and then the sheaths were stained with fluorescent dye Pepper620 (Fluorescence Diagnosis Shanghai Biotech Company; HBC620) before microscopic analysis. The epifluorescent analysis was conducted with Olympus BX53, UPlanSApo 100 ×/1.40 oil objective (Olympus) and laser intensity controlled by a VSAOTF100 System and coupled into the light path using a VS-20 Laser-Lens-System (Visitron Systems). Excitation and emission wavelengths were 488 nm and 520 nm for GFP, and 575 nm and 625 nm for Pepper620. Images were recorded using a charged-coupled device camera (Photometrics CoolSNAP HQ2; Roper Scientific).

Analyses of extracellular vesicles

According to the Minimal Information for Studies of Extracellular Vesicles guidelines⁵³, isolated extracellular vesicles of M. oryzae were examined and validated by transmission electron microscopy, lipophilic dye staining and nanoparticle tracking analysis. A 20-μl aliquot of the sample was applied onto a specimen holder. For transmission electron microscopy analysis, a copper grid with its formvar–carbon-coated side facing down was placed onto the surface of the droplet containing isolated extracellular vesicles for 3 min. The grid was then removed and allowed to air-dry at room temperature. Subsequently, the grid was placed (coated side down) onto a droplet of 2% phosphotungstic acid solution for negative staining for 2 min. The grid was removed, and excess stain was carefully blotted away using filter paper. The prepared grid was finally examined on a transmission electron microscope (JEOL; JEM-1400FLASH) to observe the vesicles.

For lipophilic dye staining, a sample containing extracellular vesicles was supplemented with 5 μM of PKH26 red fluorescent dye (Umibio; UR52302), followed by vortex mixing for 1 min and incubation for 10 min. An appropriate volume of 1× phosphate-buffered saline (PBS) was then added to the mixture of extracellular vesicle and dye. Labelled extracellular vesicles were reisolated following the standard extraction protocol to remove unbound dye. The pellet was resuspended in 200 μl of 1× PBS, yielding the fluorescently labelled extracellular vesicles that were observed under a Leica confocal microscope using 551-nm excitation for fluorescence imaging. For nanoparticle tracking analysis, a sample containing isolated extracellular vesicles was sonicated for 15 s (frequency of 20 kHz; 30% of 150-W power; 5-s pulse on–5-s pulse off; repeated three times) to disperse aggregated particles and then diluted to an appropriate concentration with PBS before loading for measurement using a NanoSight NS300 nanoparticle tracking analyser.

GFP reporter assay for miR5827 binding with lnc117761 or PKR1

A previously described GFP reporter assay^54,55 was used to examine the binding of miR5827 with lnc117761 or PKR1. The expression plasmids were constructed in the pCAMBIA1300 binary vector and introduced into Agrobacterium strain GV3101, with its culture resuspended with the infiltration buffer (1 M MgCl₂, 200 mM acetosyringone and 0.2 M 2-morpholinoethanesulfonic acid) to a final concentration of OD₆₀₀ at 0.5. The Agrobacterium suspension was mixed with two combinations as indicated and then placed at room temperature under dark conditions for 3 h. For analysis of miR5827 binding with lnc117761, the 35S-MBS^lnc117761–GFP and 35S-mMBS^lnc117761–GFP reporter constructs were separately transiently expressed alone or co-expressed with 35S-miR5827 in N. benthamiana leaves through Agrobacterium at the indicated concentrations (0.1–0.6  optical density). For analysis of miR5827 binding with PKR1, the 35S-MBS^PKR1–GFP and 35S-mMBS^PKR1–GFP reporter constructs were separately transiently expressed alone or co-expressed with 35S-miR5827 in N. benthamiana leaves as described above. The two combined Agrobacterium suspensions were subsequently co-injected into fresh leaves of tobacco N. benthamiana grown for 45 days. Leaves were examined between 36 hpi and 48 hpi for image acquisition using a Leica confocal laser scanning microscope. Immunoblot analyses were performed following standard protocols. In brief, 15 μg of the total protein was electrophoresed in a 10% SDS–PAGE gel, and the protein blot was probed with anti-GFP (BBI Life Sciences; D191040) diluted 5,000-fold, followed by HRP-conjugated mouse secondary antibody (Invitrogen; 31430) diluted 10,000-fold, to detect and determine GFP accumulation.

ITC assay

ITC was performed at 25 °C with a NANO ITC instrument (TA Instruments) following standard procedures. Single-stranded miR5827, sense-strand RNA transcripts of lnc117761 and PKR1, as well as their different mutants at the miR5827-binding site (mMBS^lnc117761 and mMBS^PKR1), were biochemically synthesized by Sangon Biotech. All biochemically synthesized RNA samples were diluted in DEPC-treated water. Subsequently, 75 µM miR5827 was titrated against 5 µM of lnc117761 RNA (MBS^lnc117761) or 5 µM of PKR1 RNA (MBS^PKR1). Similar experiments were performed with 5 µM mMBS^lnc117761 and 5 µM mMBS^PKR1, respectively, under identical conditions. Binding dissociation constants (K_d) were determined from the integrated injection heats using NanoAnalyze software (TA Instruments) with an independent binding model.

Ligation of interacting RNA and RT–PCR

To assess the interaction between lnc117761 and miR5827, LIGR coupled to RT–PCR experiment was performed following established protocols²⁶ with minor modifications. Rice TP309 seedings grown in a growth chamber for 2 weeks were inoculated with Guy11 spore suspension prepared in 0.1% Tween-20 at a concentration of 5 × 10⁵ spores ml⁻¹. After 30 h of inoculation, infected rice leaves were collected for total RNA extraction with TRIzol reagent. Meanwhile, total RNA samples extracted from mock-inoculated (0.1% Tween-20) TP309 leaves and from M. oryzae hyphae were separately used as control RNA. For crosslinking, 50 μg of each total RNA was supplemented with 20 mg ml⁻¹ of 4′-aminomethyl-4,5′,8-trimethylpsoralen (AMT) (MCE; HY-138287) together with ice-cold Tris–saline buffer (0.15 M NaCl and 0.01 M Tris–HCl; pH 7.2) to a final volume of 40 μl containing 20 μg ml⁻¹ of AMT. The mixture was then irradiated with 365-nm light in an Ultraviolet Crosslinker equipment (Analytik Jena; CL-1000M) for 30 min on ice to crosslink the RNA hybrids containing miR5827 and lnc117761. The crosslinked RNA mixture was next processed for circularization using T4 RNA Ligase 1 (New England Biolabs; M0204) according to the manufacturer’s instructions, followed by lithium chloride (Invitrogen; AM9480) precipitation to purify RNA.

The crosslinked, circularized RNA was enriched by digestion with RNase R (Beyotime; R7092S), which specifically hydrolyses linear RNA. Following this, circular RNA was isolated through lithium chloride precipitation and resuspended in 20 µl of Tris–saline buffer. For crosslink reversal, the precipitated RNA was irradiated for 5-min intervals for 10 min at 254 nm on ice. RNA samples were processed by ethanol precipitation to remove AMT and used as reverse transcription template to synthesize first-strand cDNA with miR5827-specific primer by SuperScript III First-Strand Synthesis System (Invitrogen; 18080-051). The cDNA product was amplified with primer pair LIGR-F1 and LIGR-R (Supplementary Table 8) in the first round PCR, and then with inner primer pair LIGR-F2 and LIGR-R in the second round nested PCR. The PCR product was sequenced to detect the hybrid RNA formed from miR5827 and lnc117761 as a result of their binding.

RNA pull-down analysis for binding between lnc117761 and miR5827

The single-stranded RNA transcripts of full-length lnc117761, mut-lnc117761 carrying mutation in the miR5827 binding site and GFP control were obtained by in vitro transcription with HiScribe T7 Quick High Yield RNA Synthesis Kit (New England Biolabs; E2050S) using 1 μg of each cDNA transcription template produced by PCR amplification with primer pairs listed in Supplementary Table 8. RNA products were purified with Monarch RNA Cleanup Kits (New England Biolabs; T2040), and purified RNA was then labelled with biotin using the Pierce RNA 3′ End Desthiobiotinylation Kit (Thermo Fisher Scientific; 20163) according to the manufacturer’s instructions. An amount of 50 pmol of biotin-labelled RNA transcripts of lnc117761, mut-lnc117761 and GFP was individually added into 50 µl of streptavidin magnetic beads from the Pierce Magnetic RNA-Protein Pull-Down Kit (Thermo Fisher Scientific; 20164) and incubated at 25 °C for 30 min, followed by magnetic purification and washing on a DynaMag-Spin (Invitrogen; 12320D). In the subsequent RNA pull-down procedure, biochemically synthesized 21-nt miR5827 RNA (Sangon Biotech), dispersed in a mixture of RNA binding buffer, was added to magnetic beads bound with biotin-labelled RNA. After incubation at 4 °C for 60 min with rotation, the mixture in magnetic beads was washed three times on DynaMag-Spin. RNA binding products were finally eluted in 50 μl of elution buffer at 37 °C for 30 min.

To detect miR5827 enriched by lnc117761 from the above RNA pull-down assay, first-strand cDNA of miR5827 was reverse transcribed using the SuperScript III First-Strand Synthesis System with a primer specific for miR5827. PCR was then performed to amplify the miR5827 target fragment, which was separated by 1.5% agarose gel electrophoresis and stained. Meanwhile, digital PCR was conducted to quantify the amounts of miR5827 obtained from the RNA pull-down assay using a nanoplate-based QIAcuity Digital PCR System (QIAcuity One 5-Plex). In a 12-μl digital PCR mixture, 1 μl of diluted cDNA was used as template and mixed with 0.48-μl primer pairs and 4 μl of QIAcuity EvaGreen Master Mix (QIAcuity EG PCR; 2500113). The mixture was loaded into a QIAcuity Nanoplate 8.5k 24-well PCR nanoplate (QIAGEN; 250012) and subsequently subjected to reaction under the program of the QIAcuity Digital PCR System. Finally, partitioning and data analysis were performed through QIAcuity software v.2.2.0.26.

Dual-luciferase reporter assay

For the dual-luciferase reporter assay, the DNA sequences of MBS^lnc117761, mMBS^lnc117761, MBS^PKR1 and mMBS^PKR1 were separately cloned upstream of the firefly luciferase gene in the pGreenII 62-SK vector. The recombinant pGreenII 62-SK and pCAMBIA1300-35S–miR5827 constructs were co-transfected into rice protoplasts. At 24 h post-transfection, luminescence of rice protoplasts was measured using the Dual Luciferase Reporter Assay Kit (Beyotime; RG027). The relative LUC/REN ratio was calculated from luminescence readings. Detailed procedures for luminescence measurement followed the manufacturer’s protocol.

EMSA

EMSA was conducted following a previous report⁵⁶ using the Chemiluminescent EMSA Kit (Beyotime; GS606). RNA oligonucleotides (MBS^lnc117761, mMBS^lnc117761, MBS^PKR1 and mMBS^PKR1) and biotin-labelled miR5827 (with biotin conjugated to the 5′ end of the oligo) were synthesized by Sangon Biotech with the following sequences: 5′GGUAACCAUCCUUGGAGACCUUCGUCGCAGUUGCAACAAACAUCUAUGGGAUGCUUAUCU3′ for MBS^lnc117761, 5′GGUAACCAUCCUUGGAGACCUGCGCCGCAGCUCACUACCUCAUCUAUGGGAUGCUUAUCU3′ for mMBS^lnc117761, 5′AGAAAAGGAAUGUAAGCUGGGUAGUGCUAAAUUGUAACGCAUUUUUUCAAUAACCAAUUG3′ for MBS^PKR1, 5′AGAAAAGGAAUGUAAGCUGUUCCACGUUCCCCCUUCCUGCCUUUUUUCAAUAACCAAUUG3′ for mMBS^PKR1 and biotin-5′UUUGUUGCAAUUUGGACUACC3′ for miR5827. RNA oligonucleotide sequences of point mutant MBS^PKR1 (M1-MBS^PKR1 to M20-MBS^PKR1) and point mutant MBS^lnc117761 (M1-MBS^lnc117761 to M16-MBS^lnc117761) carrying targeted replacement at single nucleotide in their MBS are listed in Supplementary Table 8. Detailed EMSA procedures were performed according to the manufacturer’s instructions. Images were captured using a charge-coupled device camera on Touch Imager (e-Blot Life Science). ImageJ was used to quantify the intensity of each band obtained by EMSA.

Uptake of external small RNA by T. aestivum root and coleoptile

Fluorescein-labelled TamiR5827 (FAM-TamiR5827) was synthesized by Sangon Biotech. Coleoptiles and roots were collected from wheat seedlings grown on water agar medium for 5 days. The tissues were immersed in a solution of 100 ng µl⁻¹ of FAM-TamiR5827, with DEPC-treated water serving as a negative control. After 6 h of incubation, half of coleoptiles and roots were transferred to the 50 U µl⁻¹ of Micrococcal Nuclease enzyme buffer (Thermo Fisher Scientific; 88216), whereas the remaining samples were moved to DEPC-treated water at 37 °C for 30 min. Temporary sections were then prepared using freehand sectioning for fluorescence microscopic analysis. In parallel, this assay was performed in T. aestivum protoplasts isolated using the Wheat Protoplast Preparation and Transformation Kit (Coolaber; PPT121).

Sequence analysis of DNA region encoding RNAs similar to miR5827 from different species

The 21-nt mature rice miR5827 sequence derived from the DNA region in rice and a 31-nt sequence where miR5827 is located were used as BLAST search queries against the Ensembl database to identify homologous sequences in the genomes of representative microorganisms and plants. The 31-nt length was selected to include both the 21-nt core-binding site of miR5827 and 5 nt of the flanking sequence on each side, allowing assessment of both the essential binding region and its immediate structural context. The query sequence was aligned to the downloaded genomes using BLAST v.2.12. BLAST search strategy was specifically designed to avoid sampling bias. For each of the two groups (microorganisms and plants), we retained only the single best-matched sequence per species, ensuring that the resulting multiple sequence alignments shown in Supplementary Table 6 reflect the diversity of the sequence motif across kingdoms rather than the abundance within any single genome. The sequence logo was constructed on the basis of the alignment of relevant sequences representing the level of nucleotide occurrence in each position using WebLogo 3 analysis³⁰, which reflects the presence of a similar sequence motif across diverse species. Despite the short sequence, the resulting sequence logos reflect the sequence similarity across diverse species.

Software and algorithms

The software and algorithms used are all publicly accessible. ZEN 2012 imaging software blue edition (Carl Zeiss Microscopy) was used to track and collect the micrographs. Image Lab v.3.0 build 11 (Bio-Rad) was used for DNA electrophoresis and immunoblot data collection. Long non-coding RNA was predicted using Coding Potential Calculator 2, available at https://cpc2.gao-lab.org/. GraphPad Prism is available at https://www.graphpad.com/. ImageJ is available at https://imagej.net/ij/. The miRNA database used is miRbase at https://www.mirbase.org.

Quantification and statistical analysis

To quantify the statistical significance of differences, the probability value (P value) was determined with two-sided Student’s t-test or analysis of variance (ANOVA) using the Microsoft Excel software. No statistical method was used to predetermine the sample size. The sample size was chosen according to common practice in the research field. The data collection was not blinded because when conducting the experiments, we had to be aware of controls and treated groups. The samples were randomly selected for statistic analysis. All values are presented as mean ± s.d. All experiments were repeated at least twice, except where otherwise indicated, to confirm reliability.

Availability of materials

Constructs, transgenic seeds and strains generated in this study are available upon request, but a completed material transfer agreement may be required.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.