Dicer-like 5 deficiency confers temperature-sensitive male sterility in maize

The study investigates the role of the monocot-specific Dicer-like 5 (DCL5) protein in the biogenesis and function of 24-nt phased secondary small interfering RNAs (phasiRNAs) during maize anther development. In plants, small RNAs include miRNAs, heterochromatic siRNAs, and phasiRNAs; in grasses, phasiRNAs are enriched in male reproductive organs. 21-nt phasiRNAs depend on DCL4 and are abundant premeiotically, whereas 24-nt phasiRNAs peak during meiosis and persist afterward, suggesting a potential role in meiosis or associated processes. While DCL1–DCL4 functions are relatively conserved, DCL5 has been proposed to generate 24-nt phasiRNAs in monocots, but its biological function has remained unclear. The authors generated and characterized maize dcl5 mutants to test whether DCL5 is required for 24-nt phasiRNA production and to determine its impact on anther development and fertility, particularly under differing temperature regimes.

Prior work established that 21- and 24-nt reproductive phasiRNAs are abundant in grass anthers, with 21-nt phasiRNAs linked to premeiotic stages and male fertility in rice. DCL1–DCL4 roles span miRNA biogenesis, viral/transgene siRNA processing, and 24-nt hc-siRNA production for RdDM, with some functional redundancy. DCL5 was identified as a monocot-specific Dicer implicated in producing 24-nt phasiRNAs. In maize, 24-nt phasiRNAs coincide with meiotic onset and are enriched in the tapetum. In rice, perturbation of reproductive phasiRNA pathways underlies photoperiod- and thermo-sensitive genic male sterility, implicating small RNAs and lncRNAs (e.g., PMS1, PMS3) in environmentally sensitive fertility. Tapetal defects are a common feature in abiotic stress-induced male sterility across plant species. These studies motivate testing whether DCL5-mediated 24-nt phasiRNAs support tapetal function and male fertility, especially under temperature variation.

• Mutant generation: Used an Agrobacterium-delivered CRISPR/Cas9 system targeting two sites 76 bp apart within Dcl5 to create four alleles (dcl5-1, dcl5-2, dcl5-3, dcl5-4). Guide RNAs were expressed from rice U6 promoters; constructs included maize Ubiquitin1-driven Cas9 and 35S-driven bar. Mutations were identified via T7E1 assay and Sanger sequencing, and stable segregating families established. An independent Mutator transposon insertion allele disrupting exon 3 (dcl5-mu03) was also identified and validated by PCR/Sanger sequencing.
• Growth conditions and fertility assays: Plants grown in greenhouse and field. Controlled day/night temperature regimes: restrictive 28/22 °C, permissive 26/20 °C and 23/20 °C (14 h day/10 h night). A temperature swap experiment moved plants between regimes for defined 3–21 day windows around meiosis to identify phenocritical periods. Male fertility quantified by image-based analysis (PlantCV v2 naive Bayes segmentation) of anther exertion area relative to tassel area; pollen viability assessed with Alexander staining; seed set tested by pollination.
• Microscopy: Confocal microscopy of propidium iodide–stained fixed anthers to assess wall layer architecture; DAPI-stained meiocytes for meiotic progression. Transmission electron microscopy for ultrastructure of tapetum and other layers at meiosis. ScaleP clearing with Calcofluor white and Syto13, followed by multiphoton confocal imaging and 3D rendering (Amira) to quantify tapetal binucleation status.
• Molecular profiling: sRNA-seq from anthers/spikelets and RNA-seq from anthers. sRNA libraries prepared via TruSeq Small RNA or RealSeq-AC; reads trimmed (Trimmomatic) and mapped to B73 v4 (Bowtie). 24-nt phasiRNAs quantified at 176 previously defined 24-PHAS loci using hits-normalized counts, normalized to counts per 20 million (CP20M). RNA-seq libraries prepared with TruSeq Stranded Total RNA with RiboZero-Plant or NEBNext Ultra II Directional kits; mapped with TopHat2; transcripts assembled (Cufflinks) and counted (FeatureCounts); DE analysis with DESeq2. Expression and proteomics data from public resources were integrated to assess Dcl5 spatiotemporal expression.
• Off-target assessment: Checked the most likely off-target, Dcl3, for genome editing; no edits detected.

• Dcl5 null and hypomorphic mutants are male sterile in standard greenhouse and field conditions; wild type and heterozygotes are fertile and morphologically normal.
• Tapetal defects in dcl5-1: despite normal anther wall layer patterning and normal meiosis, tapetal cells at mid-meiosis are pale, distended, largely mononucleate with poorly resolved organelles (TEM), versus densely packed, binucleate cells in fertile siblings. Post-meiotic tapetal programmed cell death is delayed, and the middle layer persists abnormally. Quantitatively, in 1.5 mm anthers (prophase I) tapetal binucleation is reduced ~5-fold in dcl5-1; significant differences persist at 2.5 mm (meiosis II) and 3.0 mm (post-meiosis). Reported p-values: 1.5 mm p=0.0010; 2.5 mm p=0.0007; 3.0 mm p=0.0320.
• Expression localization: Dcl5 transcripts are enriched in tassels and anthers, especially in tapetum (laser microdissection, in situ), with lower levels in meiocytes; DCL5 protein peaks in 2.0 mm meiotic anthers and is low/undetectable in non-anther tissues, indicating tapetal localization of the 24-nt phasiRNA pathway.
• 24-nt phasiRNA biogenesis requires DCL5: Across all four CRISPR alleles and the transposon allele, 24-nt phasiRNAs from 176 24-PHAS loci are dramatically reduced: >99% reduction in dcl5-1, dcl5-2, dcl5-4 and >90% in dcl5-3. Other sRNAs (24-nt hc-siRNAs, 21-nt phasiRNAs) are retained. 24-PHAS precursors accumulate at similar or higher levels in mutants, indicating a processing defect rather than precursor loss.
• Minimal transcriptome changes: Only 23 transcripts show ≥2-fold differential expression in dcl5-1 anthers, mainly genes expressed in later tapetal stages, suggesting changes are indirect and due to developmental arrest.
• Temperature-sensitive male sterility: dcl5-1 and dcl5-mu03 are completely sterile at 28/22 °C (restrictive). At 26/20 °C and 23/20 °C (permissive), dcl5-1 ranges from partially to fully fertile; viable pollen is produced and sets seed. Under permissive temperatures, tapetal binucleation is restored to near-normal levels.
• Crucially, fertility restoration at permissive temperatures occurs without restoration of 24-nt phasiRNAs; dcl5-1 plants remain depleted of 24-nt phasiRNAs under both temperature regimes, and Dcl5 transcript levels and 24-PHAS precursor levels are similar across temperatures.
• Phenocritical window: Temperature swap experiments indicate a ~9-day window spanning meiosis to early post-meiosis (initiation/peak of 24-nt phasiRNAs in normals) during which permissive temperature is required to restore fertility in dcl5-1. Less than nine days in permissive conditions reduces fertility. This suggests DCL5 buffers tapetal development against warm temperatures during this critical period.
• Off-target editing not detected at Dcl3; all four CRISPR alleles and the Mu allele share the male-sterile phenotype, supporting causality of Dcl5 loss.

The findings demonstrate that maize DCL5 is essential for robust male fertility at typical warm growth temperatures by enabling proper tapetal redifferentiation during meiosis. Loss of DCL5 abolishes 24-nt phasiRNA production without broadly altering the mRNA transcriptome, yet profoundly impacts tapetal development (reduced binucleation, delayed degeneration) and prevents anther exertion and viable pollen formation at higher temperatures. The restoration of fertility at cooler temperatures, despite the continued absence of 24-nt phasiRNAs, suggests that 24-nt phasiRNAs are dispensable when development is slower and alternative pathways can compensate, but are required to buffer developmental transitions under faster, warm-temperature growth. The data support a model in which 24-nt phasiRNAs act primarily in the tapetum, possibly by rapidly resetting gene regulatory states during meiosis through interactions with Argonautes or by broadly modulating small RNA occupancy, even though they lack obvious mRNA complementarity. Comparisons with rice highlight that small RNA pathways (21-nt phasiRNAs and lncRNA-derived sRNAs) can underlie environmentally sensitive male sterility, reinforcing the importance of small RNAs in reproductive development under environmental fluctuations. Agronomically, because the restrictive temperature approximates optimal U.S. Corn Belt conditions, DCL5-mediated 24-nt phasiRNAs likely contribute to maintaining high yield potential under warm seasons and may be increasingly critical as climate warming and heat waves intensify.

This work establishes that DCL5 is the dedicated Dicer for 24-nt reproductive phasiRNA biogenesis in maize anthers and that its function is crucial for tapetal development and male fertility under warm, yield-optimal temperatures. dcl5 mutants exhibit tapetal defects, near-complete loss of 24-nt phasiRNAs, and temperature-sensitive male sterility, with a phenocritical window spanning meiosis to early post-meiosis. Fertility can be restored at cooler temperatures without restoring 24-nt phasiRNAs, indicating conditional necessity tied to developmental rate. Future research should elucidate the molecular mechanisms of 24-nt phasiRNA action (targets, AGO interactions), define genetic variation across maize inbreds affecting this pathway, assess roles in other crops, and explore biotechnological strategies to enhance 24-nt phasiRNA functions to buffer reproductive development under heat and other abiotic stresses.

• The mechanistic targets of 24-nt phasiRNAs remain unidentified; they show little complementarity to mRNAs, and transcriptome changes are minimal, limiting insight into direct regulatory effects.
• One RNA-seq library (dcl5-1 restrictive) lacked a biological replicate; dispersion estimates were derived from mean libraries, potentially reducing statistical power.
• Environmental testing focused on temperature; other stresses were not assessed.
• The study was conducted primarily in specific genetic backgrounds; the extent of phenotypic variation across diverse maize inbreds is unknown.
• While meiosis appeared normal cytologically, potential subtle effects in meiocytes or other cell types cannot be fully excluded.