Winter fields antisense RNAs to kick off flowering.

FLOWERING LOCUS C (FLC), a MADS-box transcription factor, plays a major role in determining flowering time in Arabidopsis In this issue of Genes & Development, Zhao and colleagues (pp. 888-898) elucidate the role of COOLAIR antisense noncoding RNAs in FLC regulation through field trials and laboratory experiments. COOLAIR-mediated FLC silencing is induced by the first seasonal frost in the field and thus acts as a key molecular indicator during autumn for winter arrival.

Temperature is a major seasonal cue for plants to align flowering time. High activity of the central floral repressor FLC inhibits untimely flowering in the model plant Arabidopsis and also crop species (Michaels and Amasino 1999; Takada et al. 2019). A thorough dissection of FLC regulation in Arabidopsis identified an impressive collection of contributing mechanisms (Wu et al. 2020). A tantalizing observation is that defects in conserved factors regulating RNA polymerase II (RNAPII) transcription often result in seemingly specific flowering time phenotypes through altered FLC expression. Allelic variation at FLC in Arabidopsis accessions collected around the world highlighted DNA sequence polymorphisms at this locus to the local adaptation of flowering (Hepworth et al. 2020). A fascinating aspect of FLC expression regulation is the responsiveness to long cold periods (winter) that trigger PRC2-based chromatin silencing to ensure low FLC expression when conditions are favorable in spring.

Like many other genes in Arabidopsis, several transcript isoforms can be generated from the FLC locus (Thomas et al. 2020). Perhaps surprisingly, COOLAIR, a group of long noncoding antisense RNAs expressed from the FLC locus, emerged as central coordinator of FLC expression. The discovery of COOLAIR provided a previously missing target for pathways mediating alternative polyadenylation (Liu et al. 2010). Moreover, cold results in COOLAIR induction, offering a molecular rationale for how high FLC expression can be reversed by cold to allow flowering. Mutations in 3′ end formation complexes flower extremely late due to high FLC activity, but cold treatment and resulting COOLAIR induction can substitute for 3′ end formation defects and accelerate flowering (Koornneef et al. 1991). COOLAIR thus provides the FLC locus with a neat cold sensor to initiate repression in preparation for spring. COOLAIR induction in laboratory conditions peaks after 2–3 wk of cold exposure, but induction can be detected within hours after cold treatment by methods detecting nascent RNAPII transcription (Kindgren et al. 2020). These findings raised the question of how COOLAIR activity is triggered by natural environments.

Zhao et al. (2021) clarify FLC regulation in natural temperature regimes. The investigators assayed the expression of FLC and COOLAIR isoforms in samples from key selected accessions grown in three field sites across Europe. Field trial data revealed that COOLAIR is highly induced when plants experience a dip below freezing temperature, likely analogous to the first frost in autumn. The freezing-induced COOLAIR spike coincides with the reduction of FLC transcripts. Importantly, controlled environment chambers that recapitulate natural temperature regimes could validate these findings, and show that a freezing-induced COOLAIR spike represents a key trigger for FLC repression. These findings support the view that the COOLAIR-FLC circuitry of repression is triggered by an exposure to first frost in natural temperature regimes, and that this property likely underpins environmental adaptation.

This study presents an authoritative clarification of the functional role of COOLAIR in FLC repression. Testing the function of lncRNA represents a formidable challenge, since it is often unclear what aspect of the lncRNA needs to be disrupted to reveal the key activity (Bassett et al. 2014). Among lncRNAs, the functional characterization of antisense lncRNA is a particularly tough nut to crack since mRNA and antisense lncRNA are generated from the same DNA template. The opportunities for mutational approaches targeting transcription of the antisense DNA strand specifically are thus severely limited, in particular in experimental systems such as Arabidopsis, where knock-in mutational approaches are not commonly available.

Zhao et al. (2021) offer an impressive collection of molecular approaches to alter COOLAIR expression. These include two strategies to achieve COOLAIR overexpression in cis. They first isolated T-DNA insertion mutants that result in COOLAIR overexpression, and second, dominant mutations (ntl8-D3) of the transcriptional regulator of COOLAIR, NTL8 (Zhao et al. 2020). Quantitative single-cell imaging methods establish that COOLAIR induction anticorrelates with unspliced FLC transcript levels (used as a proxy for transcription), strongly supporting a role in repression. The inverstigators describe five experimental strategies to engineer specific COOLAIR loss-of-function mutants. Four of these rely on a two-step approach based on the flc-2 fast neutron mutation where most FLC sequences have been deleted. The flc-2 mutant is then complemented with genomic clones of FLC where sequences in the 3′ end linked to transcriptional termination of FLC mRNA and the induction of COOLAIR are replaced. The fifth strategy relies on CRISPR/Cas9 deletion of sequences in the FLC terminator region that include the transcription start sites (TSSs) of some COOLAIR isoforms. A comprehensive analysis of FLC and COOLAIR transcription in this collection of mutants aiming to eliminate COOLAIR expression showed that these strategies were most effective for isoforms initiating close to the mutated sequences. However, none of the strategies fully eliminated distal COOLAIR transcription overlapping the TSSs of FLC mRNA. Targeted analysis of TSSs in COOLAIR mutants showed that novel antisense transcript isoforms are generated when initiation of COOLAIR is repressed. A hotspot of TSSs for the novel antisense transcript isoforms is the first intron of FLC, reminiscent of convergent antisense species (CAS), previously identified by genome-wide analyses in Arabidopsis (Kindgren et al. 2020). These analyses provide detailed insight into the fluidity of antisense transcript isoforms that may have contributed to prior confusion about the functional role of COOLAIR in FLC repression when only a selection of mutants was analyzed. Double mutants between ntl8-D3 and COOLAIR mutants showed that FLC reduction by ntl8-D3 requires the generation of COOLAIR, arguing for some functional specification of COOLAIR.

In conclusion, careful and holistic analyses of temperature-sensitive FLC expression from the field to the laboratory illustrates how the interplay between mRNA and lncRNA isoforms results in chromatin-based regulation of plant gene expression. This study offers insight into how temperature-responsive gene expression is achieved by plants, which should offer inspiration for strategies aiming to promote food security in a changing climate.