Landscape of Fluid Sets of Hairpin-Derived 21-/24-nt-Long Small RNAs at Seed Set Uncovers Special Epigenetic Features in Picea glauca

Abstract

Conifers' exceptionally large genome (20-30 Gb) is scattered with 60% retrotransposon (RT) components and we have little knowledge on their origin and evolutionary implications. RTs may impede the expression of flanking genes and provide sources of the formation of novel small RNA (sRNAs) populations to constrain events of transposon (TE) proliferation/transposition. Here we show a declining expression of 24-nt-long sRNAs and low expression levels of their key processing gene, pgRTL2 (RNASE THREE LIKE 2) at seed set in Picea glauca. The sRNAs in 24-nt size class are significantly less enriched in type and read number than 21-nt sRNAs and have not been documented in other species. The architecture of MIR loci generating highly expressed 24-/21-nt sRNAs is featured by long terminal repeat-retrotransposons (LTR-RTs) in families of Ty3/Gypsy and Ty1/Copia elements. This implies that the production of sRNAs may be predominantly originated from TE fragments on chromosomes. Furthermore, a large proportion of highly expressed 24-nt sRNAs does not have predictable targets against unique genes in Picea, suggestive of their potential pathway in DNA methylation modifications on, for instance, TEs. Additionally, the classification of computationally predicted sRNAs suggests that 24-nt sRNA targets may bear particular functions in metabolic processes while 21-nt sRNAs target genes involved in many different biological processes. This study, therefore, directs our attention to a possible extrapolation that lacking of 24-nt sRNAs at the late conifer seed developmental phase may result in less constraints in TE activities, thus contributing to the massive expansion of genome size.

Keywords: 21- and 24-nt-long small RNAs; MIR loci; Picea glauca; conifers; long terminal repeat-retrotransposons (LTR-RTs); seed set.

Fig. 1.—

Distribution of sRNAs by size (A) and by populations (B), and absolute read numbers of 24- and 21-nt sRNAs (C and D). Note: The long thick dashed lines in ellipses represent the relative change on read numbers of 24nt (in blue) and 21nt (in grey); changes on absolute number of unique sRNA reads (C) and on that of reads mapped to miRBase hairpins (D) in each population over time during seed set of P. glauca; sequencing statistics was presented in supplementary table S1 (Supplementary Material online).

Fig. 2.—

Comparisons of DCL3 homologs (A and B), pgRTL2 relative expression (C) and RPM scaled by the number of unique reads (D). Note: gene trees for DCL3 homologs in gymnosperms and several model angiosperms (A); components of DCL3 domains in Arabidopsis (general), Phoenix canariensis and Picea glauca (B); PAZ and dsRB represent Piwi–Argonaute–Zwille and double stranded RNA-binding domains, respectively; thick dashed line in (B) means the upstream of incomplete pgRTL2 mRNA may contain PAZ domain after mapped to its genome (PG29-v.4); RPM represents reads per million and if the absolute expression of the 24-nt sRNA exceeds 2,000 copies in a single library, it is not used for the calculation of RPM and RPM per unique read number. We excluded around 4-5 sRNA reads per library (see supplementary fig. S4 C, Supplementary Material online).

Fig. 3.—

Numbers of sRNA reads of 24- (A) and 21-nt (B) by seed set phases in populations. Note: the number of reads detected throughout phases and populations respectively provided in black and detailed information listed in supplementary table S3, Supplementary Material online.

Fig. 4.—

Analyses of abundant sRNA populations in 24-/21-nt size classes via their biogenesis sequences (A) and target genes (B and C). Note: the highly expressed sRNA reads provided in supplementary table S4 (Supplementary Material online) with details on computational predictions; one gene may be classified in more than one gene category or none of the categories; gene category and GO code: apoptotic process (GO:0006915), developmental process (GO:0032502), cellular process (GO:0009987), metabolic process (GO:0008152), biological regulation (GO:0065007), localization (GO:0051179), and cellular component organization or biogenesis (GO:0071840).

Fig. 5.—

Summary of distinctive features of 24-nt sRNAs and outstanding questions regarding their production in conifers. Note: Pol II, DCL3, and AGO4/6 are abbreviations of polymerase II, Dicerlike 3, and argonaute 1/4/7, respectively, which are key effectors in 24-nt sRNA biogenesis; RdDM represents the RNA-directed DNA methylation pathway.