Methyltransferase-like (METTL) homologues participate in Nicotiana benthamiana antiviral responses

ABSTRACT

Methyltransferase (MTase) enzymes catalyze the addition of a methyl group to a variety of biological substrates. MTase-like (METTL) proteins are Class I MTases whose enzymatic activities contribute to the epigenetic and epitranscriptomic regulation of multiple cellular processes. N⁶-adenosine methylation (m⁶A) is a common chemical modification of eukaryotic and viral RNA whose abundance is jointly regulated by MTases and METTLs, demethylases, and m⁶A binding proteins. m⁶A affects various cellular processes including RNA degradation, post-transcriptional processing, and antiviral immunity. Here, we used Nicotiana benthamiana and plum pox virus (PPV), an RNA virus of the Potyviridae family, to investigated the roles of MTases in plant–virus interaction. RNA sequencing analysis identified MTase transcripts that are differentially expressed during PPV infection; among these, accumulation of a METTL gene was significantly downregulated. Two N. benthamiana METTL transcripts (NbMETTL1 and NbMETTL2) were cloned and further characterized. Sequence and structural analyses of the two encoded proteins identified a conserved S-adenosyl methionine (SAM) binding domain, showing they are SAM-dependent MTases phylogenetically related to human METTL16 and Arabidopsis thaliana FIONA1. Overexpression of NbMETTL1 and NbMETTL2 caused a decrease of PPV accumulation. In sum, our results indicate that METTL homologues participate in plant antiviral responses.

KEYWORDS:

• Methyltransferase
• METTL
• N⁶-methyladenosine
• antiviral immunity
• Potyviridae
• RNA virus

Introduction

Methyltransferase (MTase) enzymes catalyze the addition of a methyl group to a variety of biological substrates including small molecules, lipids, proteins, and nucleic acids. S-adenosylmethionine (SAM)-dependent MTases of the Class I fold act in the majority of known methylation reactions; they have emerging roles in multiple biological functions and are key regulators of cellular epitranscriptomic dynamics.^1–5

Over 100 types of chemical RNA modifications have been discovered so far, and N⁶-methyladenosine (m⁶A) is the most ubiquitous one.⁶ It is widely distributed in coding and non-coding RNAs of cellular organisms as well as viruses.^7–11 In plants, m⁶A has conserved roles in diverse RNA metabolic processes including mRNA splicing, nuclear export, stability, translation, and small RNA maturation.^9,12 m⁶A abundance is jointly regulated by MTases and MTase-like (METTL) proteins, demethylases, and m⁶A binding proteins, also known as writers, erasers, and readers, respectively.^7,13

Within metazoans, METTLs constitute a protein family whose members have key roles in the methylation of DNA and RNA molecules.¹⁴ In mammals, the main components of the m⁶A-MTase complex include METTL3, METTL14, Wilms’ tumor 1-associating protein (WTAP) and KIAA0823. METTL3 is the SAM-binding subunit,⁷ which is highly conserved in eukaryotes and whose deletion in mice leads to early embryonic death.¹⁵ Similarly, T-DNA insertion disruption of Arabidopsis thaliana MTA (TAIR: AT4G10760), an METTL3 homologue, causes an embryo lethal phenotype.¹⁶ METTL14 is another component of the m⁶A-MTase complex,⁷ which can heterodimerized with and potentiate the enzymatic activity of METTL3. MTB (TAIR: AT4G09980) is an METTL14 homolog shown to be a component of the A. thaliana complex responsible for m⁶A deposition in mRNAs.¹⁷

MTase domains recur in diverse viral taxa,¹⁸ but their contribution in m⁶A deposition and cellular epitranscriptomic dynamics is currently unknown. Several studies nonetheless showed that cellular METTLs affect viral infection of eukaryotic hosts. Silencing of METTL3 and METTL14 decreases human immunodeficiency virus 1 (HIV-1) replication.¹⁹ Conversely, knockdown of the same genes increases Zika virus (ZIKV) production.²⁰ m⁶A modification of hepatitis C virus (HCV) RNA is mediated by METTL3 and METTL14, and regulated by WTAP.²¹ Depletion of METTL3, METTL14, or WTAP increases production of HCV infectious particles,²² indicating a negative role of m⁶A in HCV infection. Recent research indicates that m⁶A plays a regulatory role in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, too.²³ The host writer complex installs m⁶A in SARS-CoV-2 genomic RNA and replication intermediates, and knockdown of METTL3 and METTL14 promotes SARS-CoV-2 replication. These findings support the involvement of m⁶A in antiviral immunity.

m⁶A has multiple roles in regulation of plant development, biotic stress, and abiotic stress responses.²⁴ More recently, m⁶A has been proposed as a component of novel plant antiviral defense system that was shown to modulate plant RNA virus replication and movement.^10,25,26 Evolutionary and experimental evidence supports the m⁶A implication in the infection of members of Potyviridae, the largest family of plant RNA viruses.^27,28 Proteins of the alkylation B (AlkB) superfamily show RNA demethylase activity, m⁶A removal, and proviral functions. AlkB homologues have been identified in Potyvirus genomes; in turn, downregulation of host AlkB-like genes significantly decreases the accumulation of plum pox virus (PPV; genus Potyvirus).²⁹ Wheat yellow mosaic virus (WYMV; genus Bymovirus) triggers transcriptome-wide m⁶A changes in wheat,³⁰ and the WYMV NIb associates with Triticum aestivum m⁶A methyltransferase B (TaMTB) to promote viral infection.³¹

Despite these advances, plant METTL homologues and their roles in plant-virus interaction are poorly characterized. In this study, we investigate the roles of METTLs from Nicotiana benthamiana during PPV infection. Our transcriptomic analysis of N. benthamiana plants shows that viral infection significantly alters the expression of multiple MTases. Two N. benthamiana METTL transcripts (NbMETTLs), hereinafter referred to as NbMETTL1 and NbMETTL2, were then cloned and characterized by in silico sequence analyses, as well as structural modeling of the encoded proteins. Finally, we found that overexpression of NbMETTLs could promote the N. benthamiana resistance to PPV infection.

Results

Viral infection alters the expression of N. benthamiana methyltransferase (MTase) genes

To understand how viral infection influences MTase expression in N. benthamiana, plants were infected with PPV and upper uninoculated leaves were subjected to RNA sequencing (RNA-seq) alongside uninfected control samples (Figure 1). Three biological replicates were analyzed for each condition. Sequencing reads were filtered to remove poor quality reads and sequencing adapters; transcript read counts were then obtained using an augmented N. benthamiana transcriptome previously assembled from PPV-infected plant samples and a revised proteome annotation.^32,33 Bioinformatic analysis of the obtained RNA-seq data detected expression of over 40,000 transcripts. Hierarchical clustering of the RNA-seq normalized count values highlighted two significantly divergent groups comprising either PPV-inoculated or control samples, thus confirming the consistency and transcriptomic similarities of samples for each of the conditions analyzed (Figure S1). To identify differentially expressed MTases, on the one hand, read counts of the two conditions were compared to compute false discovery rate (FDR) values, and, on the other hand, transcripts were queried against the SUPERFAMILY protein domain database for their functional annotation (Figure 1b; Dataset S1). Our analysis identified 166 transcripts encoding putative MTases with significantly altered expression in PPV-infected samples compared to healthy control plants (FDR <0.05; Figure 1c; Dataset S1).

Figure 1. Expression of S-adenosylmethionine-dependent methyltransferases (SAM-MTases) during virus infection. (a) Phenotypes of plum pox virus (PPV)-inoculated and control Nicotiana benthamiana plants. (b) Transcriptomes of PPV-inoculated and control plants were analyzed by RNA-seq with three biological replicates per condition. The MA plot shows the transcript fold changes (FC) and normalized counts (CPM); red dots show differentially expressed transcripts (FDR <0.05) with annotated MTase domains. (c) Heatmap shows sample profiles of the differentially expressed MTase transcripts (FDR <0.05). (d) Plots show RT-qPCR quantification values (mean ± SD) of two N. benthamiana METTL homologues in PPV-infected or healthy control samples; p values by Student’s t test are shown.

Cloning of MTase transcripts from N. benthamiana

Based on RNA-seq results, we designed primers that were used to amplify two putative MTase transcripts from samples of PPV-infected plants. RT-PCR products spanning two full-length open reading frames were obtained, cloned into binary vectors for plant expression and sequenced (Figure S2A). The obtained sequences of the two N. benthamiana MTases, herein referred to as NbMETTL1 and NbMETTL2, have been deposited at NCBI (NbMETTL1, GenBank: MZ423213; NbMETTL2, GenBank: MT107162). RT-qPCR quantification of two N. benthamiana MTase transcripts in PPV-infected or healthy control samples indicated that their accumulation is generally downregulated during virus infection, albeit to varying extents (Figure 1d).

The cloned N. benthamiana MTases are METTL homologues

Translation of the cloned sequences results in NbMETTL1 with 224 aa (GenBank: QYD01611) and NbMETTL2 with 310 aa (GenBank: QPC97719). The protein physiochemical properties were analyzed by the ProtParam webserver (Table 1) and percentages of amino acid content are shown in Table S1. Among other parameters, NbMETTL1 and NbMETTL2 have the molecular formula C₁₁₄₇H₁₇₉₄N₂₉₆O₃₂₉S₉ and C₁₅₀₉H₂₄₀₆N₄₃₀O₄₇₁S₁₉, average molecular weights of 25 kDa and 34 kDa, and PI of 6.16 and 5.52, respectively.

Table 1. Physicochemical parameters of the two cloned NbMETTL proteins.

Parameter	NbMETTL1	NbMETTL2
Molecular formula	C1147H1794N296O329S9	C1509H2406N430O471S19
Molecular weights	25283.19	34717.44
Theoretical PI	6.16	5.52
Instability index	46.42	42.19
Aliphatic coefficient	95.22	79.55
Total number of negatively charged residues (Asp+Glu)	25	45
Total number of positively charged residues (Arg+Lys)	22	35
Grand average of hydropathicity (GRAVY)	−0.147	−0.462

Protein sequence search against hidden Markov models of the SUPERFAMILY protein profile database showed a significant match to the SAM-dependent MTase domain (profile SSF53335) at positions 63–173 of NbMETTL1 (e-value 1.54e-20) and 61–195 of NbMETTL2 (e-value 1.54e-20) (Figure S2B).

We compared sequences of the cloned MTases and those of reference species to evaluate their evolutionary relationship. Phylogenetic analysis showed that the two cloned sequences from N. benthamiana belong to the METTL family, whose members are involved in a variety of epigenetic and epitranscriptomic processes.¹⁴ N.benthamiana MTases are thus METTL homologues that clustered with the human METTL16 forming a monophyletic sister clade to human METTL21A and METTL21D (Figure S2C). Interestingly, human METTL16 and its Arabidopsis thaliana homologue FIONA1 (FIO1; TAIR: AT2G21070) are m⁶A RNA methyltransferases with emerging roles in mRNA maturation, stress response, and development.^34–37

NbMETTLs are structurally related to A. thaliana FIO1 and human METTL16

SAM-dependent MTases share little sequence identity, but a seven-beta-strand structural core is conserved across METTL proteins alongside the motif E/DxGxGxG involved in SAM binding.^2,4,5 Recent deep learning algorithms can predict high-resolution protein structures from primary sequences.^38,39 The three-dimensional structure of NbMETTLs was predicted by OmegaFold; models obtained showed a conserved core with high prediction confidence scores, which was flanked by less conserved termini with a variable number of disordered residues (Figure 2(a–d)).

Figure 2. Structural modeling and analysis of NbMETTLs. Tertiary structural models of NbMETTL1 (a) and NbMETTL2 (b) are shown; colors by prediction confidence scores. Secondary structures derived from the predicted models of NbMETTL1 (c) and NbMETTL2 (d) are shown. (e) The structural similarity matrix of NbMETTLs and reference MTases; colors by Dali Z scores. (f) The structural dendrogram of NbMETTLs and reference MTases; branch support values are shown. Structural models of NbMETTLs were computed by OmegaFold, A. thaliana MTA and FIO1 models were from the AlphaFold database; human METTL3 and METTL16 crystal structures were reported.

To confirm that NbMETTLs are structurally related to known METTL enzymes, structural comparison and alignment analyses were done with reference structures obtained by X-ray crystallography or high-resolution modeling. Results from a structural similarity matrix and hierarchical clustering indicate that NbMETTLs have spatial topologies close to A. thaliana FIO1 and human METTL16, which were in turn significantly different to those of human METTL3 and its A. thaliana homologue MTA (Figure 2e,f).

Structural superimposition of human METTL16 (PDB: 6GFN) and models of NbMETTLs revealed high conservation of the secondary structures of the MTase core. The aligned primary sequences showed moderate-to-low conservation, but the E/DxGxGxG motif known to be involved in SAM binding could be readily identified in NbMETTL1 and NbMETTL2 (Figure 3). The findings from structural modeling thus corroborated the results from functional annotation and primary sequence analyses, and indicate that the cloned NbMETTLs are Class I MTase enzymes that use SAM as the methyl donor to chemically modify one or more specific substrates.

Figure 3. Structural sequence alignment of NbMETTLs and METTL16. Sequence conservation and consensus secondary structures (PROMALS3D) are shown for the structural superimposition of human METTL16 (PDB: 6GFN) and structural models of NbMETTLs. Inverted triangles highlight the Class I MTase conserved motif E/DxGxGxG involved in SAM binding.⁵.

NbMETTLs promote the N. benthamiana resistance against PPV infection

To verify the influence of NbMETTL proteins on viral infection, overexpression constructs were used in transient expression assays. The expression vectors pEAQ-HT-DEST1-NbMETTL1 and pEAQ-HT-DEST1-NbMETTL2 were transformed into Agrobacterium strains that were infiltrated into N. benthamiana leaves. Compared to the control condition, transcript quantification results showed a > 150-fold accumulation increase of NbMETTLs in infiltrated samples, indicating that the overexpression constructs were functional (Figure 4a).

Figure 4. Effect of NbMETTL overexpression on plant virus accumulation. N. benthamiana leaves were infiltrated with NbMETTL expression constructs and inoculated with PPV; leaves were analyzed at 4 days post infiltration and upper uninoculated leaves were analyzed after 12 days. (a) Plots show RT-qPCR transcript quantification values (mean ± SD) in samples collected from leaves infiltrated with NbMETTL overexpressing constructs. (b) Plots show PPV RNA quantification values (mean ± SD) in samples collected from upper uninoculated leaves and measured by RT-qPCR. (c) Viral protein accumulation in upper uninoculated leaves was measured by immunoblotting with an anti-PPV coat protein (CP) serum; RuBisCO large subunit is shown as a loading control. Plots show signal quantification values (mean ± SD). CTRL, control condition; p values by Student’s t test are shown.

At 2 days post infiltration, PPV was mechanically inoculated on the infiltrated leaves, and upper uninoculated leaves were analyzed after 12 days. Compared to the control condition, plants treated with NbMETTL1 and NbMETTL2 showed a reduction of the PPV RNA levels as detected by RT-qPCR analysis (Figure 4b).

Immunoblotting of the upper uninoculated leaf samples was done with an anti-PPV coat protein (CP) serum to support the findings at a protein level. In agreement with the RNA analysis results, a decrease in CP amount was detected in NbMETTL-overexpressing samples compared to the control condition (Figure 4c). Statistical analysis of the quantification values showed that the changes in viral RNA and protein levels were significant, that is a significant reduction of PPV accumulation was detected in plants infiltrated with NbMETTLs compared to the control condition (Figure 4b,c).

To corroborate the findings, immunoblot detection of the viral reporter GFP was done. Results showed low GFP accumulation in plants overexpressing NbMETTLs and high GFP levels in the control samples (Figure S3). Taken together, our results reveal that overexpression of NbMETTL homologues promotes PPV resistance in N. benthamiana.

Discussion

MTases catalyze the addition of a methyl group to a variety of biological substrates. METTLs are Class I MTases whose enzymatic activities contribute to the epigenetic and epitranscriptomic regulation of multiple cellular processes. m⁶A is a common post-transcriptional modification of eukaryotic mRNAs and non-coding RNAs. It is a dynamic and reversible process regulated by methyltransferases, demethylases, and RNA binding proteins. m⁶A affects a variety of RNA metabolic processes in eukaryotic cells,¹³ and is evolutionary conserved in divergent plant taxa.⁹ Some RNA modifications play a significant role in antiviral responses,^11,18,40,41 among them m⁶A is increasingly recognized as a key regulatory component of plant virus immunity.^10,25,26

Evidence indicates that m⁶A plays a critical role in plant-virus interaction and plant antiviral response. m⁶A has been identified in the RNA genome of alfalfa mosaic virus, whose accumulation and movement were impaired by genetic depletion of a plant m⁶A demethylase.^25,26,42 Compared to controls, expression of the m⁶A “writer” genes OsMAT3 and OsMAT4 was significantly increased in plants infected with rice stripe virus and rice black-streaked dwarf virus (RBSDV).⁴³ In tobacco, m⁶A level was globally reduced after tobacco mosaic virus infection; this may be related to increased or decreased expression of host m⁶A demethylases and methyltransferases, respectively.⁴⁴ In wheat, transcriptome-wide m⁶A profiling of two varieties with different resistance to WYMV indicates that m⁶A may be involved in plant-host interaction by regulating biological pathways related to protein post-translational modification and defense response.³⁰ In watermelon, cucumber green mottle mosaic virus infection caused a significantly decrease in m⁶A amount.⁴⁵ We recently reported the identification of atypical potyviruses encoding putative RNA demethylase domains of the AlkB superfamily and showed that downregulation of genes of the plant-specific ALKBH9 clade promotes plant resistance to potyvirus infection.²⁹ Although these studies revealed that m⁶A is conserved in plants and involved in plant virus infection, the regulatory mechanism of m⁶A remains unknown.

To promote our understanding of methylation dynamics in plant-virus interaction, we investigated the roles of host MTases upon infection of PPV, a member of the Potyvirus genus. Wheat and common bean transcriptomics showed that expression of MTases such as TaFIP37–1 and PvMTC was altered during infection of Potyviridae members.^30,46 Here, we analyzed the transcriptomes of N. benthamiana infected with PPV and identified multiple differentially expressed transcripts with annotated MTase domains. Among these, an MTase of the METTL family showed a significant downregulation compared to controls.

To characterize the function of METTL proteins in N. benthamiana, we successfully amplified and cloned the sequences of two METTL transcripts (NbMETTLs). In silico analyses confirmed the presence of a MTase domain of the Class I fold, which was supported by secondary and tertiary structure predictions and modeling. Although full-length sequence comparisons showed size and composition variability in N and C termini, the core MTase domain was relatively well conserved. Phylogeny analysis showed that the cloned NbMETTLs cluster with human METTL21A, METTL21D, and METTL16. The last of which is an m⁶A MTase with emerging roles in mRNA maturation and stress response;^34,35 the identified NbMETTLs may have a similar activity.

Some m⁶A-related proteins have been directly involved in antiviral immunity.¹¹ Studies in animal systems found that HCV, ZIKV and SARS-CoV2 fitness can be significantly affected by reducing the expression of host methyltransferases or demethylases.^{8,19,20,23,47} Downregulation of the small brown planthopper METTL3 and METTL14, components of the m⁶A MTase complex, was shown to promote replication of the plant virus RBSDV in its natural vector.⁴⁷ In plants, recent work shows that AlkB RNA demethylase homologues positively regulate virus infection indicating that a decrease in m⁶A levels negatively affect plant antiviral immunity.^25,26,29 Proviral functions of AlkB homologues is supported by their identification in diverse plant virus taxa,^27,48,49 whose RNA demethylase activity has been experimentally validated for RNA viruses of the families Potyviridae, Betaflexiviridae and Closteroviridae.⁴¹ Consistently, we found that the accumulation of PPV was significantly reduced in N. benthamiana plants treated with NbMETTL homologues with putative methyltransferase activity. These findings suggest that m⁶A is a positive regulator of plant antiviral immunity.

In summary, we cloned transcripts the N. benthamiana METTL homologues NbMETTL1 and NbMETTL2, then we performed bioinformatic analysis and investigated their roles in PPV infection. Plant virus infection is dependent on host resources and factors.⁵⁰ Our results are consistent with an emerging model in which a balance between cellular methylase and demethylase activities can tune plant antiviral immunity and, in turn, virus fitness. This set basis for novel strategies for virus control based on epitranscriptomic reprogramming of crops.

Materials and methods

Plant materials and virus inoculation

N. benthamiana plants were grown in a greenhouse at ~23ºC, 16 h light and 8 h dark. A full-length cDNA copy of a PPV isolate adapted to Nicotiana spp., tagged with a GFP gene, and inserted into a binary vector was reported.⁵¹ The PPV clone was delivered to N. benthamiana plants by Agrobacterium-mediated inoculation as detailed.⁵² Upper uninoculated leaves were collected, stored at −80°C, and used as the inoculum source; PPV inoculum preparation in phosphate buffer and mechanical inoculation were done as described.³³

RNA sequencing

Total RNA samples were purified from three biological replicates/condition, libraries were prepared and sequenced (2 × 150 nt) in an Illumina platform (BGI, China).²⁹ Raw RNA-seq reads were filtered with fastp⁵³ to remove poor quality reads and adapter contaminations. Transcript read counts were obtained by Salmon⁵⁴ and an augmented N. benthamiana transcriptome including the PPV genome;³³ the complete N. benthamiana draft genome sequence v2.6.1 (https://solgenomics.net/ftp/genomes/Nicotiana_benthamianaV261/) was used as the decoy sequence.⁵⁵ Differential expression analysis was done with edgeR.⁵⁶ Hierarchical clustering analysis of transcriptome profiles was done using pvclust with average linkage clustering and 1000 bootstrap replications.⁵⁷ For transcript functional annotation, translation products were obtained using OrfM⁵⁸ and protein domains were identified using InterProScan and the SUPERFAMILY-1.75 database.^59,60

Reverse transcription quantitative PCR (RT-qPCR)

Total RNA was purified and used in cDNA synthesis reactions; cDNA aliquots were then used in PCR reactions performed with MonAmp^TM SYBR Green qPCR mix (Monad, China) and gene-specific primers (Table S2) in an FTC-3000P Real-Time Quantitative Thermal Cycler (Funglyn Biotech, Canada). Expression was normalized using NbUBI as a reference, and fold changes relative to the control condition were calculated by the ΔΔ^CT method.²⁹

mRNA cloning and transient expression

NbMETTL1 and NbMETTL2 transcripts were amplified by RT-PCR reactions including N. benthamiana total RNA samples and transcript-specific primers (Table S2). The products were inserted into the pMD19-T plasmid and confirmed by Sanger sequencing. Recombinant pMD19-T plasmids carrying the NbMETTL sequences were linearized by the restriction enzyme SmaI and used in Gateway recombination reactions (Thermo Fisher Scientific) alongside the binary vector pEAQ-HT-DEST1⁶¹ to assemble the plant expression vector pEAQ-HT-DEST1-NbMETTL1 and pEAQ-HT-DEST1-NbMETTL2. The resulting constructs were transformed into Agrobacterium strain C58C1 hosting a disarmed pTi by the freeze-thaw method.⁵² Agrobacterium strains carrying the NbMETTL expression vectors were cultured overnight at 28ºC. Cultures were then centrifuged at 4000 rpm at room temperature (5 min), the supernatant was discarded and bacterial pellets were suspended at OD₆₀₀ = 1 with a buffer containing 0.5 M 2-(N-Morpholino)ethanesulfonic acid hydrate (pH 5.6), 1 M MgCl₂ and 0.1 M acetosyringone. Bacterial suspensions were incubated at room temperature (3 h), and then used to infiltrate leaves of three-week-old N. benthamiana plants using a needleless syringe. Samples from leaves infiltrated with the empty pEAQ-HT-DEST1 vector were used as a control.

In silico sequence and structure analysis

Sequences were aligned using ClustalW; the evolutionary history was inferred using MEGA7 and the maximum likelihood method based on the Kimura 2-parameter model.⁶² In silico analyses of physicochemical properties, secondary structures and conserved domains of two NbMETTL proteins were done using the ExPaSy ProtParam and the SUPERFAMILY-1.75 database.⁶⁰ The three-dimensional structural models of NbMETTLs were predicted by OmegaFold;³⁸ models of A. thaliana MTA (AF-O82486-F1-model_v4) and FIO1 (AF-Q5XEU1-F1-model_v4) were downloaded from the AlphaFold protein structure database;³⁹ crystal structures of the human METTL3 (PDB: 5K7W) and METTL16 (PDB: 6GFN) catalytic domains were reported.^63,64 Structures were visualized and imaged using UCSF ChimeraX;⁶⁵ sequence alignments based on structural superimposition were generated using PROMALS3D.⁶⁶ The structural similarity matrix was obtained using Dali,⁶⁷ and the structural dendrogram by US-align⁶⁸ pairwise alignment (Table S3) and pvclust with average linkage clustering and 10,000 bootstrap replications.⁵⁷

Protein immunodetection

Total protein extracts from plant samples were prepared, resolved by SDS-PAGE and electroblotted onto nitrocellulose membranes as described.³³ Membranes were stained with Ponceau S for loading control; immunodetection was conducted using as the primary antibodies rabbit anti-PPV coat protein (CP) serum³³ and mouse anti-GFP monoclonal antibody (AE012, ABclonal). Horseradish peroxidase-conjugated goat anti-rabbit IgG (ab205718, Abcam) or goat anti-mouse IgG (AS003, ABclonal) were used as the secondary antibody; immunostained proteins were visualized by enhanced chemiluminescence detection.

Statistics

Student’s t test was used for two-group comparisons; significance levels of p values are indicated in the figures. Quantification values from immunoblotting assays were normalized by the sum of the replicate approach;⁶⁹ samples were from two independent experiments, each one with 12 plants per treatment. In RNA-seq analyses, the Benjamini–Hochberg method was applied on the p values to control the false discovery rate (FDR).⁷⁰

Acknowledgments

We are grateful to J.A. García (CNB, CSIC, Spain) for materials.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data supporting the findings of this study are available within the article, its supplementary materials or are available from the corresponding authors; the cloned gene sequences are available at NCBI (NbMETTL1, MZ423213; NbMETTL2, GenBank: MT107162). Raw RNA-seq files are available at NCBI under BioProject ID PRJNA953827 (http://www.ncbi.nlm.nih.gov/bioproject/953827).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15592324.2023.2214760.

Additional information

Funding

This study was supported by National Science Foundation of China (grants 31770165 and 31860489 to M.Z.) and Major project of Nature Science Foundation of Inner Mongolia of China (grants 2021ZD06 to M.Z.). F.P. is supported by a “Juan de la Cierva Incorporación” contract (IJC2019-039970-I) from Ministerio de Ciencia e Innovación (Spain), and grants MiniVi (ELIXIR-IIB, Cineca, Italy) and BCV-2023-1-0021 (Red Española de Supercomputación, Spain).