Proper regulation of a sperm-specific cis-nat-siRNA is essential for double fertilization in Arabidopsis

Abstract

Natural cis-antisense siRNAs (cis-nat-siRNAs) are a recently characterized class of small regulatory RNAs that are widespread in eukaryotes. Despite their abundance, the importance of their regulatory activity is largely unknown. The only functional role for eukaryotic cis-nat-siRNAs that has been described to date is in environmental stress responses in plants. Here we demonstrate that cis-nat-siRNA-based regulation plays key roles in Arabidopsis reproductive function, as it facilitates gametophyte formation and double fertilization, a developmental process of enormous agricultural value. We show that male gametophytic kokopelli (kpl) mutants display frequent single-fertilization events, and that KPL and a inversely transcribed gene, ARIADNE14 (ARI14), which encodes a putative ubiquitin E3 ligase, generate a sperm-specific nat-siRNA pair. In the absence of KPL, ARI14 RNA levels in sperm are increased and fertilization is impaired. Furthermore, ARI14 transcripts accumulate in several siRNA biogenesis pathway mutants, and overexpression of ARI14 in sperm phenocopies the reduced seed set of the kokopelli mutants. These results extend the regulatory capacity of cis-nat-siRNAs to development by identifying a role for cis-nat-siRNAs in controlling sperm function during double fertilization.

Figure 1.

kokopelli is a male gametophytic mutant that shows reduced seed set. (A) Schematic representation of the KPL and ARI14 genomic region. Thick and thin solid lines represent exons and introns, respectively. Dotted lines represent 3′ UTRs. Dashed lines represent the complementary DNA strand. Positions of T-DNA insertions within the kpl-1 and kpl-2 alleles are marked with triangles. (B) Representative silique from a self-pollinated kpl-1/kpl-1 plant showing a reduced seed set. The arrow points to an undeveloped ovule, and the arrowhead points to an aborted seed. (C) Percentage of normal seeds, aborted seeds, and undeveloped ovules from self-pollinated heterozygous (Ht, gray bars, n = 1061) or homozygous (Ho, black bars, n = 1547) kpl-1 plants. Error bars represent standard deviation from the mean. (D) Siliques of reciprocal crosses of kpl-1/kpl-1 plants with wild type (WS4). The arrow points to an undeveloped ovule, and the arrowhead points to an aborted seed. (E) Percentage of normal seeds, aborted seeds, or undeveloped ovules in reciprocal crosses with wild type, using the kpl-1/kpl-1 plants as either male (gray bars, n = 1385) or female (black bars, n = 293). Error bars represent standard deviation from the mean. (F) RT–PCR profile of KPL expression in various tissues. (Rt) Roots; (Sdl) 14-d-old seedlings; (Le) rosette leaves; (CF) closed flowers; (OF) open flowers; (Pis) unpollinated pistils; (Pol) pollen; (Slq) siliques. (G) KPL and ARI14 expression profiles throughout male gametophyte development. Microrraray data was adapted from Honys and Twell (2004). (UNM) Microspores; (BCP) bicellular pollen; (TCP) tricellular pollen; (MPG) mature pollen.

Figure 2.

Defects in ovules receiving kpl-1pollen. (A–D) Siliques from a kpl-1/kpl-1 plant were dissected 1–2 DAP, and then the ovules were cleared and examined using differential interference contrast (DIC) microscopy. The frequency of each phenotype is given in parentheses (n = 764 ovules; an additional 7% had been double-fertilized, but either the embryo or endosperm had arrested). Bars, 50 μm. (A) Unfertilized ovule with central cell nucleus (CCN; false-colored blue) and egg cell nucleus (ECN; false-colored green). (B) Fertilized ovule with embryo (EMB; false-colored red) and developing endosperm (endosperm nuclei; ENDN, false-colored blue). (C) Representative ovule with a two-celled embryo and unfertilized central cell. (D) Representative ovule with endosperm nuclei and an unfertilized egg cell. (E–G) Undeveloped ovules from a self-fertilized kpl-1/kpl-1 plant harboring the HTR10pro:HTR10-mRFP transgene. Sperm nuclei (SN) are bright red dots. Embryo sacs are red due to autofluorescence. Bars, 50 μm. (E) Fluorescent image of an ovule with one sperm (SN). (F) Overlay of fluorescent and bright-field images of the ovule in E, showing a developing embryo (EMB, false-colored green) and the sperm nucleus (SN). (G) Fluorescent image of an ovule with two sperm nuclei (SN). (H–J) Patterns of GUS expression observed after wild-type plants were pollinated using kpl plants carrying FAC1pro:GUS as pollen donor. (H) Doubly fertilized ovule. (I) Embryo (EMB) only. (J) Endosperm (END) only.

Figure 3.

Gene expression of KPL and ARI14 in wild-type and kpl pollen. (A,B) Relative expression levels of KPL and ARI14. Total RNA was extracted from 1-d-old open flowers of kpl homozygous mutants and their corresponding wild-type ecotypes. (A) RT–PCR analyses, with ACT2 used as a control. (B) qPCR analysis. The RNA level of KPL in wild type was set at 100%. Each expression level was normalized to that of IPP2. Standard deviations were calculated from two biological replicates. (C) Fluorescence microscopy images of pollen grains from transgenic plants harboring promoter fusion constructs of KPL and ARI14 to tdTomato. (D–F) Fluorescence microscopy images of pollen from transgenic plants harboring the ARI14pro:mCherry-ARI14 translational fusion construct. Mature pollen grains are visualized with epifluorescence (D) or CLSM (F), and a germinated pollen tube is visualized with epifluorescence (E). Arrows point to the shadows where the sperm are located. Bars: C–E, 20 μm; F, 10 μm.

Figure 4.

ARI14 is regulated post-transcriptionally by KPL. (A,B) GFP expression in leaf epidermal cells of transgenic plants expressing the 35Spro:GFP-ARI14 construct (A) or both the 35Spro:KPL and 35Spro:GFP-ARI14 constructs (B). (C,D) qPCR analyses of KPL (C) and GFP (D) levels in plants expressing 35Spro:GFP-ARI14 or both 35Spro:KPL and 35Spro:GFP-ARI14 (Red plants). Total RNA was extracted from seedlings of F₂ plants selected for the presence of either one or both constructs. The RNA level of the relevant parent was set at 1. Each expression level was normalized to that of IPP2. (E) Northern blot analysis of small RNAs from F₂ progeny of plants carrying both the 35Spro:KPL and 35Spro:GFP-ARI14 constructs, the 35Spro:GFP-ARI14 construct alone, or wild-type (Col) plants. The hybridization probes were a mixture of ARI14-specific oligomers; a U6 probe was used as a loading control for small RNAs. (F) Schematic of the overlapping region of KPL and ARI14. Thick and thin solid lines represent exons and introns, respectively. Dashed lines represent the complementary DNA strand. In the expanded area, double lines represent 3′ UTR, vertical lines represent polyA addition sites, and dotted lines represent putative pre-mRNA downstream from the polyA addition site. Pinhead vertical lines represent the position of the ARI14 cleavage site detected by RLM-5′ RACE. The numbers above each point indicate the position of RLM-5′ RACE products along the ARI14 cDNA sequence, and the numbers in parentheses indicate the frequency of such products at each site. The short line above the KPL exon represents a KPL sequence signature obtained from the sperm siRNA of wild type (Slotkin et al. 2009).

Figure 5.

Accumulation of ARI14 mRNA in different siRNA biogenesis mutants. (A–D) qPCR analysis of ARI14 expression in 1-d-old open flowers of various small RNA biosynthesis mutants and their corresponding wild-type controls. The RNA level of the wild-type control in each group was set at 1. Each expression level was normalized to that of IPP2. Standard deviations were calculated from two to three biological replicas.

Figure 6.

Overexpression of ARI14 in sperm phenocopies the reduced seed set observed in kpl mutants. (A) Total RNA was extracted from 1-d-old open flowers of T₂ progenies from six independent T₁ lines (nos. 28, 39, 45, 85, 92, and 93) that ectopically expressed ARI14 driven by the KPL promoter, and was then used for qPCR analysis. The expression level of ARI14 in wild type was set to 1. Standard deviations were plotted from three replicates. The expression level was normalized to that of IPP2. (B) A representative silique of a self-pollinated T₂ plant overexpressing ARI14. The arrow indicates an undeveloped ovule. (C) Seed set analysis of the plants shown in A. Standard deviations were calculated from seed counts of 10 siliques for each plant.

Figure 7.

Model for biogenesis of KPL-ARI14 nat-siRNAs and their proposed role in fertilization. In wild-type sperm, the KPL-ARI14 nat-siRNAs lead to down-regulation of ARI14. Reduced KPL transcripts in kpl mutants result in accumulation of ARI14 in the sperm. By binding to an unknown substrate in sperm, ARI14, which is likely an inactive ubiquitin E3 ligase, might prevent active ARI family E3 ligases (e.g., ARI13) in sperm from tagging that substrate for degradation, resulting in impaired fertilization.