Plant-specific BLISTER modulates miRNA biogenesis by regulating MIR transcription, HYL1 phosphorylation, and nuclear transport in Arabidopsis

Abstract

MicroRNAs (miRNAs), processed from primary transcripts by the microprocessor complex (MC), serve as crucial post-transcriptional regulators in eukaryotes. The stability and nuclear localization of HYL1, a core MC component, are essential for maintaining complex activity. In this study, we demonstrate that the plant-specific protein BLISTER (BLI) plays a key role in miRNA biogenesis by regulating MIR transcription, HYL1 phosphorylation, and HYL1 transport in Arabidopsis. The bli mutant exhibits increased accumulation of specific miRNAs accompanied by enhanced HYL1-containing D-body formation. Biochemical evidence indicates that BLI negatively regulates MIR transcription. Moreover, BLI promotes HYL1 dephosphorylation, which facilitates its degradation. Furthermore, BLI interacts with KETCH1 to orchestrate HYL1 nuclear import. These findings establish a novel regulatory paradigm where a plant-specific protein integrates transcriptional control and post-translational modification to coordinate miRNA production, advancing our understanding of plant gene regulation mechanisms.

Graphical Abstract

Graphical Abstract.

Introduction

MicroRNAs (miRNAs) are noncoding RNAs ranging from 21 to 24 nucleotides in length that negatively regulate gene expression at the post-transcriptional level through complementary base pairing. These miRNAs play crucial roles in a wide array of biological processes [1, 2]. The biogenesis of miRNAs involves a series of carefully coordinated steps [1]. In plants, the transcription of genes encoding miRNAs (MIR) into primary miRNAs (pri-miRNAs) is primarily carried out by DNA-dependent RNA polymerase II (Pol II), with assistance from the Mediator, Elongator, and transcription factors such as NEGATIVE ON TATA LESS2 (NOT2), DAWDLE (DDL), and CELL DIVISION CYCLE5 (CDC5) [3–6]. The RNase III enzyme DICER-LIKE 1 (DCL1) processes these pri-miRNAs into precursors miRNAs and subsequently into miRNA/miRNA* duplexes within the nucleus [7]. Over time, accumulating evidence has suggested a close interplay between the transcription process of MIR genes and the processing of pri-miRNAs [4, 8].

Efficient processing of miRNAs also necessitates the involvement of HYPONASTIC LEAVES 1 (HYL1) and SERRATE (SE) in forming the microprocessor or Dicing complex along with DCL1 [9–12]. Both SE and HYL1 have been shown to enhance the precision of pri-miRNA processing [13]. Similar to most proteins, the proper functioning of the plant microprocessor components often requires post-translational modifications. In the case of the plant microprocessor, the phosphorylation of HYL1 is vital for efficient miRNA production. MITOGEN-ACTIVATED PROTEIN KINASE3 (MPK3) and SUCROSE NONFERMENTING1-RELATED PROTEIN KINASE2 (SnRK2) are known to phosphorylate HYL1 [14–17]. HYL1 contains seven serine residues that serve as phosphorylation sites, with S42 and S159 being particularly crucial for its function. Substituting either of these residues with aspartate or glutamate codons leads to significant phenotypic alterations [14].

The degradation of HYL1 is regulated by light conditions. During daylight, the RING-finger E3 ligase COP1 (Constitutive Photomorphogenic 1) translocates to the cytoplasm and protects HYL1 from degradation by inhibiting the activity of the cytoplasmic protease HCS1 (HYL1-cleavage subtilase 1). However, in darkness, COP1 returns to the nucleus, allowing HCS1 to cleave the N-terminus of HYL1 [18–21]. Furthermore, the phosphorylation of S159 protects HYL1 from degradation by confining HYL1 in the nucleus, and the hyperphosphorylation mimic HYL1 cannot be recruited to the processing complex [22]. CPL1/2 (C-TERMINAL DOMAIN PHOSPHATASE-LIKE 1 and 2) and PROTEIN PHOSPHATASE 4 (PP4) play essential roles in the dephosphorylation of HYL1, thereby facilitating miRNA biogenesis [14, 23, 24]. Additionally, AAR2, a homolog of a U5 small nuclear RNP assembly factor in yeast and humans, functions in pre-messenger RNA splicing and enhances HYL1 dephosphorylation in Arabidopsis [25]. The import of HYL1 from the cytoplasm into the nucleus is facilitated by KARYOPHERIN ENABLING THE TRANSPORT OF THE CYTOPLASMIC HYL1 (KETCH1) [26]. However, it remains unclear whether the phosphorylation modification of HYL1 influences its transport process.

BLISTER (BLI), a unique protein exclusive to plants, engages in intricate interactions with CURLY LEAF (CLF), a component of the Polycomb repressive complex 2 (PRC2) and a histone methyltransferase. This interaction plays a crucial role in regulating the expression of polycomb-group target genes [27]. Prior investigations have also established BLI’s negative impact on the endoplasmic reticulum stress response by disrupting the functionality of IRE1A [28]. Furthermore, BLI orchestrates chromatin dynamics and fatty acid synthesis in seeds through its interaction with the transcription factor WRINKLED1 (WRI1) [29]. Additionally, it plays a pivotal role in the vacuolar transport of soluble proteins and the recycling of endocytosed plasma membrane proteins [30]. Most recently, research has revealed that BLI participates in skotomorphogenesis through its interaction with the GSK3-like kinase BRASSINOSTEROID INSENSITIVE2 (BIN2) and the transcription factor BRASSINAZOLE RESISTANT1 (BZR1) [31], further expanding our understanding of its diverse and significant functions in plants. Furthermore, BLI contains nuclear localization signal and nuclear export signal motifs, an N-terminal IDR domain, a structural maintenance of chromosomes (SMC)-like domain, and a C-terminal coiled-coil domain. The SMC-like domain of BLI mediates its interaction with CLF or WRI1 [29, 31], while the C-terminal coiled-coil domain mediates its interaction with the core retromer or BIN2 [30, 31].

In this study, we elucidate the multifaceted role of BLI in miRNA biogenesis in Arabidopsis. We demonstrate that BLI suppresses MIR gene transcription through direct binding to their promoters. Conversely, BLI deficiency enhances the formation of HYL1-associated D-bodies, consequently increasing the abundance of specific miRNAs. Biochemical analyses reveal that BLI regulates HYL1 phosphorylation status, with bli-1 mutant showing an elevated ratio of phosphorylated to non-phosphorylated HYL1. This post-translational modification appears to influence HYL1 nuclear retention, as both BLI overexpression and bli-1 mutation enhance HYL1 nuclear accumulation. Furthermore, we establish that BLI mediates HYL1 nuclear translocation through physical interaction with the nuclear transport regulator KETCH1. These findings reveal a sophisticated regulatory network where BLI coordinates transcriptional and post-translational control mechanisms in miRNA production, providing new perspectives on the functional versatility of this evolutionarily conserved plant-specific protein.

Materials and methods

Accession numbers

Sequence data in this article can be obtained in the Arabidopsis Genome initiative data library with the following accession numbers: BLI (AT3G23980), HYL1 (AT1G09700), SE (AT2G27100), DCL1 (AT1G01040), KETCH1 (AT5G19820), IMB1 (AT5G53480), and IMPA6 (AT1G02690).

Plant materials and growth conditions

All Arabidopsis thaliana strains used in this study are in the Columbia-0 (Col-0) accession. Arabidopsis seedlings were grown under long-day condition (16-h-light/8-h-dark) at 20°C–22°C. The T-DNA insertion line bli-1 and transgenic line BLI complementation were obtained from Dr Jiaqiang Sun [31]. To generate HYL1-YFP/bli-1, pMIR167a::GUS/bli-1, and pMIR172b::GUS/bli-1 lines, bli-1 was crossed with HYL1-YFP, pMIR167a::GUS, and pMIR172b::GUS reporter lines, respectively [6, 32, 33].

Constructs

For bimolecular fluorescence complementation (BiFC) assay, full-length coding sequences (CDSs) of respective genes were cloned into the pSAT1-nVenus-C or pSAT4-cCFP-C vector [34]. The DNA fragment containing cCFP- or nVenus-gene was released by I-SceI restriction enzyme and subsequently cloned into the pPZP-RCS2-ocs-bar vector. For co-immunoprecipitation (Co-IP) assay, to obtain 35S::BLI-MYC construct, BLI complementary DNA (cDNA) was PCR (polymerase chain reaction)-amplified and cloned into the pENTR/D-TOPO (Invitrogen) and then introduced into pEarleyGate203 [6]. The full-length CDS of BLI was cloned and inserted downstream of the 35S promoter and fused with GFP in pZP211 vector to generate 35S::BLI-GFP. For luciferase (LUC) complementation imaging (LCI) assay, the full-length CDSs of respective genes were cloned and fused with N-terminal or C-terminal of luciferase protein and ligated into the pEAQ vector [35]. For generation of 35S::HYL1^S>A-GFP (seven serine codons S7, S8, S42, S60, S85, S89, and S159 mutated to alanine) and 35S::HYL1^S>D-GFP (seven serine codons S7, S8, S42, S60, S85, S89, and S159 mutated to aspartate), the full-length CDSs of HYL1^S>A and HYL1^S>D were separately cloned by overlapping PCR and ligated into the pZP211 vector. More details for the constructs and primer sequences used in this study are listed in Supplementary Table S2.

qRT-PCR

Total RNA was extracted from the Arabidopsis seedlings or inflorescence tissues by TRIzol reagent (Takara) and treated with RNase-free DNase, followed by cDNA synthesis using RevertAid Reverse Transcriptase (Vazyme, China) with oligo(dT) primers. Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed using iQ SYBR Green Supermix (Bio-Rad) on the Bio-Rad CFX96 system. Primers used are listed in Supplementary Table S2.

Small RNA sequencing

Total RNA was extracted from the inflorescence tissues of wild type (WT) and bli-1 was harvested as three biological replicates and used for small RNA sequencing (Lianchuan, China). The generated sequencing data were then analyzed for miRNA abundance according to the method described previously [36].

RNA stability measurement

RNA stability was assessed using the method described previously [37]. Briefly, 12-day-old seedlings were first equilibrated by incubating them overnight in 1/2 MS liquid medium. Subsequently, cordycepin was added to achieve a final concentration of 0.6 mM. Seedlings were then collected at various time points (0 min, 30 min, 60 min, 90 min, and 120 min) for RNA extraction. qRT-PCR was conducted to quantify the levels of pri-miRNAs.

Cell-free protein degradation assay

The cell-free protein degradation assay was conducted as outlined previously [23]. Briefly, seedlings of different genotypes were grounded in liquid nitrogen and subsequently dissolved in lysis buffer (comprising 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 2 mM DL-Dithiothreitol (DTT), and 10% glycerol). The supernatants were collected and supplemented with cycloheximide (CHX) to attain a final concentration of 500 mM. Each supernatant, derived from the plant material, was then incubated at 28°C. Samples were taken at specified time points and subjected to protein gel blot analysis to ascertain the protein levels.

In vivo protein degradation assay

The treatment of the seedlings was based on previous study [38]. Briefly, 10-day old seedlings of different genotypes were incubated in dH2O and MG132 (50 μM) with or without CHX (300 μM). Samples were taken at specified time points and subjected to protein gel blot analysis to ascertain the protein levels. The subsequent protein extraction method was consistent with cell-free protein degradation assay.

GUS histochemical staining

GUS histochemical staining assay was performed as described previously [39]. Briefly, tissues from bli-1 or WT plants harboring pMIR167a::GUS or pMIR172b::GUS were incubated overnight in the staining solution at 37°C. Seventy percent ethanol was used for tissue clearing before imaging.

Co-IP and protein detection

The Co-IP analyses performed using the 35S::BLI-GFP transgenic Arabidopsis rosette leaves or Nicotiana benthamiana leaves coexpressed relevant fusion proteins that were directed by the 35S promoter. The total proteins were extracted using the lysis buffer [50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 5 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0, 0.1% Triton X-100, and 0.2% NP-40] with freshly added Phenylmethysulfonyl fluoride (PMSF) (10 mM) and 1× protease inhibitor cocktail. Anti-GFP and anti-Myc-conjugated agarose beads (CWBIO, KTSM1301 and KTSM1306), and anti-HYL1 (New England Biolabs) were used for the immunoprecipitation. For gel blot analysis, primary antibodies (anti-HYL1, 1:5000; anti-GFP, 1:5000, TransGen, HT801; anti-MYC, 1:5000, TransGen, HT101; anti-histone H3, 1:1000, Beyotime, AF0009; anti-Rubisco, 1:2000, Beyotime, AG5359; anti-Actin, 1:5000; CWBIO, CW0264M) and secondary antibodies anti-mouse lgG and anti-rabbit lgG (1:2000 and 1:1000; Beyotime; A0192 and A0208) were used to detect protein levels.

BiFC assay

The relevant nYFP and cYFP constructs were separately transformed into the Agrobacterium tumefaciens strain GV3101. The active strains were collected and equally mixed with each other in infiltration buffer (10 mM MES, pH 5.6, 150 mM acetosyringone, and 10 mM MgCl₂), and then infiltrated into the 3-week-old N. benthamiana leaves. Plants were kept in the low-light condition for 12 h and then transferred to the bright-light condition for another 36 h. After expression, a confocal microscopy (Carl Zeiss, LSM880) was used to detect the BiFC signals (originally yellow fluorescence). In each experiment, five independent N. benthamiana leaves were infiltrated and analyzed, and three independent biological replications were performed with similar results.

ChIP assay

Chromatin immunoprecipitation assay (ChIP) was performed using 3-week-old 35S::BLI-GFP transgenic Arabidopsis rosette leaves. The procedure was performed as described previously [40]. Briefly, ∼2 g of leaves was harvested and crosslinked in 1% (v/v) formaldehyde for 10 min, followed by 5-min neutralization with 0.125 M glycine and then washing five times with distilled water. The leaves were grounded in liquid nitrogen and then dissolved the powder with lysis buffer (50 mM Tris–HCl, pH 8.0, 10 mM EDTA, 0.5% Triton X-100, 200 mM NaCl, and 1 mM phenylmethyl sulphonyl fluoride). After sonication, the chromatin complex was immunoprecipitated by anti-GFP antibody (TransGen, HT801) or anti-RPB2 (ABclonal, A5928), and then by protein G plus agarose (Thermo Fisher, 10003D). Finally, the GFP-specific enrichments were analyzed by qPCR. Primers used are listed in Supplementary Table S2. The percentage of input was calculated by determining 2^−ΔCt {ΔCt = 2^{−[Ct(ChIP)−(Ct(input)−log(input dilution factor)]}, input dilution factor = fraction of the input chromatin saved⁻¹}.

Nuclear-cytoplasmic fractionation

Nuclear-cytoplasmic fractionation was performed using 3-week-old Arabidopsis rosette leaves. The procedure was performed as described previously [41]. Briefly, 1 g leaves was grounded in liquid nitrogen and dissolved the powder with 2 ml lysis buffer (20 mM Tris–HCl, pH 7.5, 25% glycerol, 250 mM sucrose, 20 mM KCl, 2.5 mM MgCl₂, 2 mM EDTA, 5 mM DTT, and 1× protease inhibitor mixture). Then, the lysates were filtered through two layers of Miracloth. The filtered solution was centrifuged at 1500 × g at 4°C for 10 min and the supernatant was collected to centrifuge at 10 000 × g at 4°C for 15 min as the cytoplasmic fraction. For nuclear fractions, the sediment was firstly washed six times with nuclear resuspension buffer (20 mM Tris–HCl, pH 7.5, 2.5 mM MgCl₂, 25% glycerol, and 0.2% Triton X-100) and resuspended with 500 μl NRB2 (20 mM Tris–HCl, pH 7.5, 10 mM MgCl₂, 250 mM sucrose, 0.5% Triton X-100, 5 mM β-mercaptoethanol, and 1× protease inhibitor mixture), and then carefully overlaid on top of 500 μl NRB3 (20 mM Tris–HCl, pH 7.5, 10 mM MgCl₂, 1.7 M sucrose, 0.5% Triton X-100, 5 mM β-mercaptoethanol, and 1× protease inhibitor mixture). Finally, the pellet was centrifuged at 16 000 × g for 45 min at 4°C and suspended in nuclear resuspension buffer to obtain nuclear fractions. Three independent biological replications were performed with similar results.

Detection of phosphorylated protein abundance

The total proteins were extracted using the Co-IP lysis buffer without EDTA and with the addition of a phosphatase inhibitor cocktail. For the nuclear and cytoplasmic proteins, the lysis buffer also does not contain EDTA and the phosphatase inhibitor cocktail is added. The preparation of Phos binding reagent (Phosbind) acrylamide gels followed the APExBIO protocol. Briefly, Phos-tag acrylamide (final concentration 50 μM) and MnCl₂ (final concentration 100 μM) were added to the SDS–PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) gels. HYL1 protein was separated using 7.5% polyacrylamide gel. Before electropotential transmembrane, the gel was incubated in a transfer buffer containing 10 mM EDTA for 10 min (three times) and then washed with pure transfer buffer three times.

Firefly luciferase complementation assay

The firefly luciferase complementation assay was performed as described previously [42, 43]. Briefly, the relevant nLUC and cLUC constructs were separately transformed into the A. tumefaciens strain GV3101. The active strains were collected and equally mixed with each other in infiltration buffer (10 mM MES, pH 5.6, 150 mM acetosyringone, and 10 mM MgCl₂) and then infiltrated into the 3-week-old N. benthamiana leaves. The LUC fluorescence was measured 48 h post-infiltration using a Lumina II instrument (Xenogen, Alameda, CA, USA). In each experiment, five independent N. benthamiana leaves were infiltrated and analyzed, and three independent biological replications were performed with similar results.

Results

BLI is involved in miRNA biogenesis

The bli-1 mutant is characterized by a pleiotropic phenotype that closely resembles the hyl1-2 mutant in terms of miRNA biogenesis, as illustrated in Supplementary Fig. S1. Notably, the BLI gene fully restored the mutant phenotype observed in bli-1, such as reduced organ size and shorter roots (Fig. 1A and B), indicating that BLI is indeed the gene responsible for the pleiotropic phenotype of bli-1. In contrast, the overexpression of BLI (BLI-OX) did not exhibit significant phenotypic deviations compared to the WT (Fig. 1A and B). Based on these observations, we hypothesized that BLI might regulate the accumulation of miRNAs. Through small RNA sequencing analysis, we discovered a substantial global increase in miRNAs in bli-1 (Fig. 1C and Supplementary Data S1). Further validation using qRT-PCR confirmed that all tested miRNAs were significantly upregulated in bli-1 compared to WT (Fig. 1D). Since miRNAs negatively regulate the accumulation of their target messenger RNAs (mRNAs), we measured the levels of target mRNAs for selected miRNAs in both WT and bli-1 mutant using qRT-PCR (Fig. 1E). As anticipated, we observed decreased levels of target transcripts in bli-1 compared to WT. Additionally, we examined the levels of miRNAs and their targets in BLI-OX and found no significant differences compared to WT (Supplementary Fig. S2A and B).

Figure 1.

BLI is required for miRNA biogenesis. (A) Phenotypes of 3-week-old plants of the indicated genotypes. comple: 35s::BLI-GFP/bli-1; OX1: 35S::BLI-GFP/Col-1#; OX17: 35S::BLI-GFP/Col-17# (Scale bar, 1 cm). (B) Phenotypes of 7-day-old plants of the indicated genotypes (Scale bar, 1 cm). (C) The scatter plot shows the miRNA variation distribution of bli-1 compared with WT. (D) The levels of miRNAs detected by RT-qPCR. miRNA levels in bli-1 were normalized to those of U6 RNAs and compared with WT (set as 1). (E) The levels of miRNA target transcripts in WT and bli-1 detected by RT-qPCR. The levels of miRNA target transcripts were normalized to those of UBIQUITIN 5 (UBQ5) and compared with WT (set as 1). Error bars in panels (D) and (E): Standard deviations (SD) of at least three replicates (*: P < .05; **: P< .01 with two-sample t-test).

BLI negatively regulates MIR transcription

To ascertain whether the heightened miRNA levels observed in bli-1 are a consequence of altered expression of miRNA biogenesis genes, we conducted qRT-PCR to quantify the transcript abundance of DCL1 complex genes. The mRNA levels of these genes mirrored those in the WT control (Supplementary Fig. S3A), suggesting that BLI does not play a role in regulating the transcription of key miRNA biogenesis genes.

Next, we investigated whether BLI modulates MIR transcription. Initially, we assessed the accumulation of pri-miRNAs using qRT-PCR and discovered a reduction in pri-miRNA abundance in bli-1 compared to the WT (Fig. 2A). Subsequently, we examined whether the MIR promoter activities were impacted in bli-1 by crossing the pMIR167a::GUS and pMIR172b::GUS reporter lines with bli-1. These reporter lines have been previously employed to monitor the promoter activities of MIR167a and MIR172b [6, 32]. Intriguingly, histochemical staining of inflorescences with GUS revealed elevated GUS protein levels in the bli-1 background compared to the WT (Fig. 2B). Furthermore, qRT-PCR analysis of RNAs isolated from inflorescences confirmed the increased GUS transcript levels in bli-1 (Fig. 2C). We also investigated the splicing of intron-retention pri-miRNAs using reverse transcriptase polymerase chain reaction with primers spanning the intron region. However, no significant differences were observed between bli-1 and WT, indicating that BLI may not affect the splicing of certain pri-miRNAs (Supplementary Fig. S3B and C).

Figure 2.

BLI negatively regulates MIR transcription. (A) qRT-PCR analysis of pri-miRNAs in WT and bli-1. UBQ5 was used as an internal control. pri-miRNAs levels in bli-1 were normalized to those of UBQ5 and compared with WT (set as 1). (B) Histochemical GUS staining of inflorescences in WT and bli-1 harboring the pMIR167a::GUS or pMIR172a::GUS transgene. Ten plants containing GUS were analyzed for each genotype. (C) Levels of the GUS transcripts in WT and bli-1 harboring the pMIR167a::GUS or pMIR172a::GUS transgene as determined by qRT-PCR. GUS levels in bli-1 were normalized to those of UBQ5 and compared with WT (set as 1). (D) The association of BLI-GFP with MIR promoters in transgenic plants containing 35S::BLI-GFP. The occupancy of BLI-GFP at MIR promoters was detected by qPCR following ChIP. The intergenic region between At2g17470 and At2g17460 (POL II C1) was used as a negative control. IP was performed using anti-GFP antibodies. Lowercase letters indicate significant difference (P < .05). (E) The occupancy of Pol II at MIR promoters detected by ChIP assay in bli-1 and WT. Lowercase letters indicate significant difference (P < .05; by Student’s t-test). Error bars in panels (A), (C), and (D): SD of at least three replicates (*: P < .05; **: P< .01 with two-sample t-test).

Moreover, we conducted qRT-PCR to analyze the levels of pri-miRNAs in BLI-OX and found them to be decreased compared to the WT (Supplementary Fig. S4A). By crossing the pMIR167a::GUS reporter line with BLI-OX, we observed decreased GUS protein levels in the inflorescences of BLI-OX compared to the WT, and qRT-PCR analysis confirmed the reduced GUS transcript levels (Supplementary Fig. S4B and C). These findings suggest that BLI negatively regulates MIR promoter activity. One plausible explanation is that BLI is recruited to the MIR promoters to regulate their activity. Therefore, we performed ChIP experiments on BLI-OX transgenic plants using antibodies against GFP to ascertain whether BLI binds to the MIR promoters. Indeed, qPCR detected the presence of the examined MIR promoters in BLI-GFP immunoprecipitates, whereas no such detection was observed in the no-antibody controls (Fig. 2D). To confirm that BLI is a negative transcription cofactor of MIR, we monitored the occupancy of RNA Pol II at the promoters of MIRs in WT and bli-1 by ChIP analysis using a specific antibody raised against the RNA Pol II subunit B2 (anti-RPB2) (Fig. 2E). The results revealed that RNA Pol II enrichment was improved at MIR promoters in bli-1. Collectively, our results indicate that BLI inhibits MIR transcription.

BLI influences the localization of HYL1 in D-bodies and associates with microprocessor components

Considering the observed decrease in pri-miRNA transcripts and the increase in MIR promoter activity in bli-1 mutant, we hypothesized that the processing of pri-miRNAs is accelerated in the mutant. To validate this hypothesis, we analyzed the decay of pri-miRNAs in seedlings treated with the transcription inhibitor cordycepin. Prior research has demonstrated that when transcription is inhibited, the levels of pri-miRNAs are expected to decline due to processing by DCL1 and RNA degradation [33]. Indeed, both WT and bli-1 seedlings exhibited a time-dependent decrease in pri-miRNA levels. However, the decline in pri-miRNA levels was more rapid in bli-1 compared to WT (Supplementary Fig. S5), suggesting accelerated pri-miRNA processing or reduced RNA stability in bli-1. To further explore the potential mechanisms by which BLI inhibits pri-miRNA processing, we introduced HYL1-YFP into bli-1 through genetic crossing and compared the YFP signal in root cells between WT and bli-1 mutant. Notably, we observed a significant increase in HYL1-YFP-labeled D-bodies in bli-1 compared to WT plants (Fig. 3A and B). This finding aligns with the increased accumulation of miRNAs in bli-1 and suggests that the absence of BLI promotes the localization of HYL1 in D-bodies and the assembly of the microprocessor complex, thereby facilitating pri-miRNA processing.

Figure 3.

BLI influences the localization of HYL1 in D-bodies and associates with microprocessor components. (A) Representative images of HYL1-YFP in WT and bli-1. Arrowheads indicate D-bodies (Scale bar, 20 μm). (B) Violin plots showing the number of HYL1-YFP-labeled D-bodies per cell in WT and bli-1. Quantification was performed by counting >500 cells from 15 roots for each genotype. P-value was calculated by the Wilcoxon test. (C) The interaction of BLI with DCL1, SE, and HYL1 detected by BiFC. Paired cCFP- and nVenus-fusion proteins were coexpressed in N. benthamiana leaves. The combination of SE-CFP and VENUS-HYL1 was used as a positive control. Scale bars, 20 μm. (D) LCI assay showing the interaction of BLI with HYL1 and BLI with SE in N. benthamiana leaves. nLUC, the N-terminal fragment of LUC; cLUC, the C-terminal fragment of LUC; EV, empty vector. Scale bar represents the relative luminescence intensity. (E–G) Co-IP between BLI with HYL1 and BLI with SE. IP was performed using anti-GFP, anti-HYL1, or anti-MYC antibodies. Proteins of HYL1, BLI-GFP, or BLI-MYC were detected by immunoblot with anti-HYL1, anti-GFP, or anti-MYC antibodies. Inputs show the total protein before IP.

We speculated that BLI interacts with the microprocessor to play a role in miRNA biogenesis. To test this hypothesis, we first conducted BiFC analysis in N. benthamiana leaves to determine whether BLI interacts with HYL1. In this experiment, BLI was fused with Cyan Fluorescent Protein (CFP) at its C-terminus (BLI-CFP), while HYL1 was fused to VENUS at its N-terminus (VENUS-HYL1). We included VENUS-HYL1 and SE-CFP as a positive control, which have been shown to associate with each other in the nucleus [17], and BLI-CFP and VENUS-GUS as a negative control. Coexpression of VENUS-HYL1 with BLI-CFP resulted in YFP signals in both the nucleus and cytoplasm. Similarly, YFP signals in the nucleus were observed when BLI-CFP was coexpressed with VENUS-DCL1 and VENUS-SE. Furthermore, significant D-bodies were observed in the coexpression of BLI-CFP with VENUS-SE (Fig. 3C). Next, we performed LCI assays in N. benthamiana leaves to examine the physical interaction between BLI and its potential binding partners. BLI and HYL1 or SE were separately fused to the amino-terminal part (nLUC) and the carboxyl-terminal part of LUC (cLUC) to generate nLUC-BLI and cLUC-HYL1 or cLUC-SE, respectively. Strong interaction signals were observed in samples coexpressing nLUC-BLI with cLUC-HYL1 or cLUC-SE, but not in negative controls (Fig. 3D). Finally, we confirmed the interactions between BLI and HYL1, as well as between BLI and SE, using Co-IP assays. In BLI overexpressing lines, we used antibodies against GFP and HYL1 to verify the interaction between HYL1 and BLI (Fig. 3E and F). In the SE-BLI Co-IP experiment conducted in N. benthamiana leaves, BLI and SE proteins were fused with GFP or MYC tags, respectively, and antibodies against GFP and MYC were used to confirm the interaction between SE and BLI (Fig. 3G). Collectively, these findings indicate that BLI associates with microprocessor components and influences the localization of HYL1 in D-bodies, providing insights into the role of BLI in miRNA biogenesis and pri-miRNA processing.

BLI is required for HYL1 degradation

The notable elevation of HYL1-YFP-labeled D-bodies in bli-1 mutant prompted us to formulate a hypothesis regarding BLI’s potential role in regulating the transport of HYL1 between the cytoplasm and the nucleus. To validate this hypothesis, we isolated proteins from nuclear and cytoplasmic fractions of cells and quantitated the abundance of HYL1. Intriguingly, the overall abundance of HYL1 was higher in both bli-1 and BLI-OX plants compared with in WT (Fig. 4A). Moreover, both bli-1 and BLI-OX plants exhibited an increased nucleus/cytoplasm ratio of HYL1 (Fig. 4A). Since similar alterations were noted in nuclear protein abundance, we postulated the existence of distinct regulatory mechanisms governing the nucleus–cytoplasm distribution of HYL1 in response to BLI deficiency and overexpression. Prior research has indicated that HYL1 is a protein with a brief half-life and undergoes degradation in darkness or shaded conditions [18, 19, 25]. We initially investigated whether the stability of the HYL1 protein was altered in bli-1 or BLI-OX plants. To do this, we exposed 10-day-old seedlings to darkness for 2 days and assessed HYL1 protein levels in WT, bli-1, and BLI-OX. As anticipated, HYL1 levels decreased in WT seedlings under dark treatment, aligning with previous findings [18]. Under dark conditions, HYL1 protein levels were more stable in bli-1 compared to WT, whereas they were comparable between BLI-OX and WT (Fig. 4B). This result implies that the absence of BLI enhances the stability of the HYL1 protein in darkness.

Figure 4.

Loss of BLI inhibits HYL1 degradation. (A) Protein gel blot analysis of HYL1 levels in total, nuclear (N-HYL1), and cytoplasmic fractions (C-HYL1) of WT, bli-1, and BLI-OX. The levels of HYL1 were determined with anti-HYL1 antibody HYL1. Actin, histone H3, and Rubisco were used as total, nuclear, and cytoplasmic markers, respectively. HYL1 levels were normalized against actin, histone H3, and Rubisco expressed in values relative to WT. (B) Protein gel blot analysis of HYL1 in 10-day-old WT, bli-1, and BLI-OX seedlings in the light or transferred to the dark and grown for 2 days. (C) The cell-free protein decay assay of HYL1 in WT, bli-1, and BLI-OX seedling lysates treated with CHX or CHX together with phosphatase inhibitors (PI) for 0–2 h. CBB staining was used as loading control. (D) The in vivo protein decay assay of HYL1 in WT, bli-1, and BLI-OX seedling lysates treated with CHX or CHX together with PI for 0–2 h. CBB staining was used as loading control.

To further examine whether the absence of BLI can influence HYL1 degradation, we conducted a cell-free protein decay assay by incubating HYL1 with lysates from 10-day-old seedlings in the presence of CHX to inhibit de novo protein synthesis. To ascertain the significance of HYL1 phosphorylation in its degradation, phosphatase inhibitors were added to the lysates. The result shows that the decay rate of the HYL1 protein was inhibited following treatment with phosphatase inhibitors (Fig. 4C). We analyzed HYL1 protein levels 2 h post-treatment. In the control without phosphatase inhibitors, HYL1 levels exhibited a similar significant decline in both WT and BLI-OX, but the decay rate was considerably slower in bli-1 (Fig. 4C). Similar results were found in in vivo experiments (Fig. 4D). Therefore, BLI is crucial for HYL1 degradation.

BLI is involved in HYL1 phosphorylation

Previous studies have demonstrated that the phosphorylation of HYL1 plays a crucial role in controlling its subcellular localization, with phosphorylated HYL1 forming a stable pool resistant to degradation [22]. To investigate the phosphorylation status of HYL1, we utilized phos-tag gels and conducted independent experiments. The results revealed that, compared to WT plants, the hyperphosphorylated form of HYL1 was increased in bli-1 mutant, while the hypophosphorylated state of HYL1 was comparable to that of both WT and BLI-OX lines (Fig. 5A). Previous research has shown that the phosphorylation of serine 159 (S159D) in HYL1 restricts its subcellular localization to the nucleus [22]. To further examine the distribution of hyper/hypophosphorylated HYL1, we performed nuclear–cytoplasmic fractionation. Our findings indicated that, compared to WT, the ratio of hyper/hypophosphorylated HYL1 was significantly increased in the cytoplasm and slightly increased in the nucleus of bli-1 mutant. This observation aligns with the higher stability of HYL1 in bli-1, as HYL1 phosphorylation occurs in the cytoplasm. Regarding the lack of significant accumulation of hyper/hypophosphorylated HYL1 in the nucleus of bli-1, this may be attributed to the low proportion of S159D in bli-1. In contrast, the ratio of hyper/hypophosphorylated HYL1 in both the nucleus and cytoplasm of BLI-OX lines was similar to that of WT. Notably, the total nuclear HYL1 level was higher in both bli-1 and BLI-OX compared to WT, consistent with previous observations (Fig. 5B). These results suggest that the absence of BLI promotes HYL1 phosphorylation, thereby facilitating its nuclear localization.

Figure 5.

The phosphorylation of HYL1 is enhanced in bli-1. (A) Phosphoprotein mobility shift assay using Phostag in WT, bli-1, and BLI overexpression. Hypo- and hyperphosphorylated HYL1 forms are indicated as P⁻ and P⁺, respectively. Actin was used as the loading control. (B) Phosphoprotein mobility shift assay using Phostag detecting P⁻ and P⁺ HYL1 in the nuclear and cytoplasmic fractions of WT, bli-1, and BLI overexpression. Histone H3 and Rubisco were used as nuclear and cytoplasmic markers, respectively. (C) Phenotypes of 3-week-old plants of the indicated genotypes (Scale bar, 1 cm). (D) The levels of miRNAs in the indicated genotypes detected by RT-qPCR. The levels of miRNA were normalized to those of U6 and compared with WT (set as 1). Error bars in panel (D): SD of at least three replicates (*: P < .05; **: P< .01 with two-sample t-test).

The aforementioned findings prompted us to investigate the genetic interactions between bli-1 and hyl1-2. However, we were unable to obtain the bli-1 hyl1-2 double mutant, leading us to speculate that this double mutant may be lethal or nonviable due to the multiple functions of these genes in various biological processes. The increased levels of hyperphosphorylated HYL1 and miRNAs in bli-1 also prompted us to examine whether BLI influences miRNA biogenesis through HYL1 phosphorylation. We found that expressing a non-phosphorylated form of HYL1, with all seven predicted serine phosphorylation sites mutated to alanine (HYL1^S>A) or aspartic acid (HYL1^S>D), could partially rescue the developmental defects of bli-1. Notably, the rescue effect of HYL1^S>A was more significant than that of HYL1^S>D (Fig. 5C). Similarly, the accumulation of examined miRNAs in bli-1 was also partially rescued by both HYL1^S>A and HYL1^S>D (Fig. 5D). Moreover, we investigated the nucleus and cytoplasm localization of both HYL1^S>A-GFP and HYL1^S>D-GFP, and a strong signal of D-bodies was detected in HYL1^S>A-GFP (Supplementary Fig. S6). These results indicate that there is an optimal phosphorylation state of HYL1 for miRNA processing in bli-1 and that the hypophosphorylated form of HYL1 is more active in Arabidopsis development than the hyperphosphorylated form.

BLI associates with KETCH1 and assists HYL1 translocalization

A previous study has reported that KETCH1, a karyopherin, facilitates the transport of cytoplasmic HYL1 [26]. Based on this finding, we hypothesized that the overexpression of BLI aids in the nuclear localization of HYL1 through karyopherin proteins. To identify potential karyopherin proteins involved in this process, we conducted immunoprecipitation coupled with mass spectrometry (IP–MS) in BLI-OX plants. Notably, we found that KETCH1 could be successfully co-immunoprecipitated with BLI-GFP (Supplementary Table S1). To validate the interaction between BLI and KETCH1, we performed both BiFC and Co-IP analyses. In the BiFC assay, KETCH1 was fused to VENUS at its N-terminus (VENUS-KETCH1), and coexpressed with BLI-CFP in N. benthamiana leaves. We observed YFP signals in both the cytoplasm and nucleus (Fig. 6A). In the Co-IP assay, BLI was fused to the MYC tag, while KETCH1 was fused to the GFP tag. Using antibodies against GFP and MYC, we detected BLI-MYC and KETCH1-GFP in the immunoprecipitates of each other (Fig. 6B and C). Additionally, we fused the RFP tag with HYL1 and BLI proteins, and coexpressed BLI-RFP with KETCH1-GFP and HYL1-RFP with BLI-GFP in N. benthamiana leaves, respectively. Confocal microscopy revealed that BLI-RFP colocalized with KETCH1-GFP in both the nucleus and cytoplasm, similar to the coexpression of HYL1-RFP with BLI-GFP (Supplementary Fig. S7).

Figure 6.

BLI associates with KETCH1 and assists HYL1 translocalization. (A) The interaction of BLI with KETCH1 detected by BiFC. Paired cCFP- and nVenus-fusion proteins were coexpressed in N. benthamiana leaves. Scale bars, 20 μm. Co-IP between BLI and KETCH1. IP was performed using anti-GFP (B) or anti-MYC (C) antibodies. Proteins of BLI-MYC and KETCH1-GFP were detected by immunoblot with anti-MYC or anti-GFP antibodies. Inputs show the total protein before IP. (D,E) LCI assay showing the interaction of KETCH1 with HYL1 in N. benthamiana leaves was improved by BLI. nLUC, the N-terminal fragment of LUC; cLUC, the C-terminal fragment of LUC; EV, empty vector. Scale bar represents the relative luminescence intensity. (F) KETCH1 and HYL1 synergistically interact with BLI, as determined in Co-IP assay. Total protein extracts from N. benthamiana leaves expressing both 35S:KETCHI-GFP and 35S:HYL1 without or with 35S:BLI-MYC were immunoprecipitated with the immobilized anti-GFP and anti-HYL1 antibody.

To further investigate our hypothesis that BLI assists KETCH1 in translocating HYL1, we analyzed the interactions between KETCH1, HYL1, and BLI using LCI and Co-IP assays. In the LCI assay, KETCH1 was fused to nLUC to generate nLUC-KETCH1. A strong interaction signal was observed in the sample coexpressing nLUC-KETCH1 with cLUC-HYL1, but not in the negative controls. Moreover, the interaction signal was enhanced when BLI-MYC was coexpressed (Fig. 6D). Consistent with this, the interaction between KETCH1-GFP and HYL1 was also strengthened with the overexpression of BLI-MYC in the Co-IP assay (Fig. 6E). We also demonstrated that the phosphorylation state of HYL1 has no impact on its transport by KETCH1 through a Co-IP assay (Supplementary Fig. S8). Previous research has demonstrated that BLI is involved in regulating the transport of soluble vacuolar proteins [30]. Therefore, we speculated that BLI may be a typical cofactor of transport proteins. To test this hypothesis, we performed a BiFC assay with IMB1 and IMPA6, which are homologous to KETCH1 and putative importins. We fused IMB1 and IMPA6 to VENUS at their N-termini, respectively, and observed YFP signals in both the nucleus and cytoplasm of the coexpression of VENUS-IMB1 and BLI-CFP, as well as the coexpression of VENUS-IMPA6 and BLI-CFP (Supplementary Fig. S9). Collectively, these data suggest that BLI is a partner of KETCH1 and assists in the transport of cytoplasmic HYL1.

Discussion

In plants, miRNAs regulate the expression of genes involved in diverse biological processes. Accumulated previous studies have demonstrated numerous factors involved in miRNA biogenesis through various mechanisms. However, many factors that affect plant growth and development through miRNA biogenesis remain undiscovered. Herein, we unveil novel roles of the plant-specific BLI protein in miRNA accumulation by regulating MIR transcription and HYL1, including its localization in D-bodies, phosphorylation, and nuclear and cytoplasmic distribution.

This study provides evidence to support the theory that BLI inhibits MIR transcription. The ChIP assay indicates that BLI can be recruited to the promoters of MIR genes to inhibit the binding of Pol II at MIR promoters. However, a prior study has shown that BLI acts as a transcriptional activator, regulating the association of WRI1 with fatty acid biosynthesis genes in vivo by directly influencing WRI1 recruitment [29]. Furthermore, the recruitment of BLI to the promoter regions of BR-responsive genes occurs in a BZR1-dependent manner [31]. For the miRNA biogenesis genes, CDC5 and HYL1 have been implicated in MIR transcription. CDC5 associates with MIR promoters and is essential for Pol II occupancy [6]. HYL1 also interacts with MIR gene promoters and functions in the transcription initiation of RNA Pol II [44]. These findings suggest that BLI may regulate MIR transcription by interacting with other factors to indirectly bind to MIR promoters, further influencing Pol II occupancy or transcription initiation. We observed that BLI and SE formed foci within the nucleus (Fig. 3A), a phenomenon not observed between BLI and DCL1 or HYL1. These foci might be D-bodies or other protein aggregates formed by BLI and SE, such as liquid–liquid droplets [34]. Further research is needed to determine whether BLI participates in SE phase separation to assist scaffold protein SE in recruiting DCL1/HYL1 to RNA substrates, or whether BLI and SE form additional functional aggregates.

HYL1 is a short-lived protein that is not stabilized in dark conditions. The degradation of HYL1 is influenced by its phosphorylation status. We observed that the rate of HYL1 degradation in the dark is slower in bli-1 mutant compared to WT plants. Additionally, the decay rate of HYL1 was much slower in bli-1 than in WT plants in a cell-free protein decay assay, and this decay was inhibited after the treatment with phosphatase inhibitors. These observations indicate that BLI is required for HYL1 degradation and dephosphorylation. We indeed found a significant increase in hyperphosphorylated HYL1 in both the total cellular and cytoplasmic fractions of bli-1 mutant compared to WT plants, as well as a slight increase in hyperphosphorylated HYL1 in the nuclear fraction of bli-1. This is consistent with the higher stability of HYL1 in bli-1 mutant, as the phosphorylation and degradation of HYL1 occur in the cytoplasm. However, the hyperphosphorylation of HYL1 in bli-1 seems to contradict the increased miRNA levels and the increased amount of HYL1-located D-bodies in these mutant. We thus speculate that the more stable HYL1 pool formed by phosphorylated HYL1 is conducive to the formation of D-bodies. Previous studies have shown that SE is a phosphorylated protein [35, 36], and the hypophosphorylated SE can effectively restore the defects of the se-1 mutant [35]. Therefore, whether BLI also affects the phosphorylation modification of SE needs further study.

The phosphorylated sites of HYL1 are specifically regulated by multiple kinases, such as MPK3, SnRK2, and other unknown kinases, suggesting their multiple underlying functions in cellular processes. Therefore, there is a possibility that certain amino acid sites in HYL1 positively influence the formation of D-bodies and the accumulation of miRNAs. The current study showed that mimicked unphosphorylated HYL1 can only partially rescue the increased pri-miRNA levels and decreased miRNA levels in hyl1-2 mutant [25], indicating that there is an optimal state of hyper/hypophosphorylated HYL1 that contributes to pri-miRNA processing. Previous study found that the excessive accumulation of hypophosphorylated SE interferes with the assembly of SE-scaffolded macromolecular complexes, which also shows that the modification state of proteins in vivo is not an extremist pursuit, but may require an optimal ratio of active/non-active to ensure the multifunctional regulation of proteins [35]. Our finding of the increased amount of HYL1-located D-bodies in bli-1 mutant and the ability of both HYL1^S>A and HYL1^S>D to partially rescue the accumulation of examined miRNAs in bli-1 also supports this hypothesis. Although we suspect that there must be hyperphosphorylation of serine 42 or/and serine 159 on HYL1 in bli-1 due to its several phenotypic defects, it is meaningful to explore the hyper/hypophosphorylated serine sites of HYL1 and their ratio in bli-1, which might reveal new positive roles of hyperphosphorylated HYL1 in miRNA accumulation.

Due to the increased nucleus/cytoplasm HYL1 ratio in BLI-OX plants, we focused on the influence of BLI on HYL1 translocation. Based on IP–MS and protein interaction analysis, we confirmed the partnership of BLI with KETCH1. A previous study also demonstrated the positive role of BLI in the trafficking of soluble vacuolar cargoes [30]. Furthermore, we verified that the expression of BLI promotes the interaction between HYL1 and KETCH1. This also explains why BLI negatively regulates MIR transcription but BLI-OX plants have similar miRNA levels as WT plants. Whether the phosphorylation modification of HYL1 in the cytoplasm is related to its transport process requires more systematic research.

Increasing evidence indicates that BLI, a plant-unique protein, has diverse roles in various biological processes. The multiple abilities of BLI are attributed to its multifunctional motifs or domains. Therefore, it is meaningful to explore the interaction network of these domains of BLI with different partners, which will be beneficial for speculating on other roles of BLI in diverse biological processes.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was supported by grants from the National Natural Science Foundation of China (31970602 to S.Z.), State Key Laboratory of Wheat Improvement “811” project (SKL81110) and Taishan Scholars (tsqn20161020). Funding to pay the Open Access publication charges for this article was provided by State Key Laboratory of Wheat Improvement “811.”

Data availability

All study data are included in the article and/or SI Appendix. The RNA-Seq data have been submitted to the National Center for Biotechnology Information Sequence Read Archive (http://www.ncbi.nlm.nih.gov/sra) under Bio Project accession number PRJNA1234704.