Cold-induced antisense transcripts (COOLAIR) originating from the 3′ end of the Arabidopsis floral repressor FLOWERING LOCUS C (FLC) were shown in 2009 to be early actors in vernalization — the epigenetic silencing of FLC by prolonged winter cold — and the source paper left open six mechanistic questions: how cold induces COOLAIR, how antisense transcription reduces sense transcription, how COOLAIR relates to Polycomb silencing, whether effects are cis or trans, how redundant the pathways are and which RNA features matter, and whether 3′-origin antisense regulation is a general genome-wide mechanism. The 2013–2026 literature establishes that several components are now substantially resolved: cold induction is driven by CBF transcription factors binding CRT/DRE elements at the FLC 3′ end, with an R-loop at the COOLAIR promoter repressing it in warm conditions; COOLAIR and Polycomb Repressive Complex 2 (PRC2) act as parallel pathways with distinct dynamics rather than in a recruitment relationship; COOLAIR acts strictly in cis, with sense and antisense transcription mutually exclusive at individual loci; and COOLAIR is dispensable for completion of vernalization. The surviving open core is the precise molecular mechanism by which antisense transcription in cis quantitatively shuts down sense FLC transcription during cold (transcriptional interference versus RNA-product-mediated chromatin effects), the untested strict necessity of antisense transcription because no genotype lacking all antisense initiation at FLC has been constructed, and the untested genome-wide hypothesis that 3′-origin antisense initiation causally regulates cognate sense genes in a condition-dependent manner.

Background

Vernalization is the process by which prolonged exposure to winter cold renders plants competent to flower in spring. In Arabidopsis thaliana this works through the epigenetic silencing of FLOWERING LOCUS C (FLC), a gene encoding a MADS-box transcription factor (MADS-box proteins constitute a conserved family of DNA-binding transcription regulators) that represses flowering. Cold exposure first causes a progressive shutdown of FLC transcription; later, Polycomb Repressive Complex 2 (PRC2 — the multiprotein complex that trimethylates histone H3 at lysine 27, a repressive mark denoted H3K27me3) deposits H3K27me3 at an intragenic nucleation region of FLC, and the mark then spreads across the locus, providing a mitotically stable memory of winter. Cold-induced plant-homeodomain proteins such as VIN3 and VRN5 associate with PRC2 during this switch, and LHP1 is the Arabidopsis protein that binds H3K27me3 to maintain repression.

In 2009, Swiezewski, Liu, Magusin and Dean reported that a set of long noncoding antisense transcripts, collectively named COOLAIR, initiate just downstream of the FLC sense poly(A) site — that is, at the 3′ end of the gene — and are transcribed in the direction opposite to the FLC mRNA across the whole locus. COOLAIR transcripts are upregulated at the start of a cold treatment, earlier than other vernalization markers and independently of them, and their induction coincides with reduced sense transcription. Alternative splicing and alternative polyadenylation generate proximal COOLAIR isoforms that terminate within the gene and distal isoforms that extend to the FLC promoter. Attaching the antisense promoter sequences to a reporter gene was sufficient to confer cold-induced silencing on the reporter, an effect the authors attributed to promoter interference, meaning collision or occlusion between the transcription machineries of the two oppositely oriented transcription units. The paper proposed that antisense transcription transiently silences FLC transcription while PRC2 subsequently installs epigenetic memory, and it closed with the hypothesis that antisense initiation from 3′ ends of genes might be a general, condition-dependent mechanism regulating the cognate sense gene.

The intervening literature has clarified much of this picture. The cold-induction pathway is now defined: C-repeat/dehydration-responsive element (CRT/DRE) binding factors (CBFs — cold-induced AP2/ERF-family transcription factors classically studied as regulators of freezing tolerance) bind CRT/DRE elements at the 3′ end of FLC and are necessary and sufficient for cold-induced COOLAIR expression, whereas in warm conditions an R-loop (a stable RNA–DNA hybrid) covering the COOLAIR promoter, stabilized by the homeodomain protein AtNDX, keeps COOLAIR repressed. Single-molecule RNA fluorescence in situ hybridization showed that sense and antisense transcription, although able to co-occur in one cell, are mutually exclusive at each individual FLC locus, establishing cis action: the antisense transcript affects only the copy of the gene from which it is transcribed. COOLAIR RNA processing is coupled to chromatin regulation: splicing and proximal polyadenylation of COOLAIR — promoted by the RNA-binding protein FCA and by a CPF-like termination module (the APRF1/TOPP4/LD complex, analogous to the yeast cleavage and polyadenylation factor machinery) — connect antisense transcription to histone changes such as loss of the active mark H3K36me3 (trimethylation of histone H3 lysine 36) and demethylation of H3K4 by the FLD demethylase. COOLAIR RNA itself folds into evolutionarily conserved, elaborate secondary structures that are remodeled by cold, and mutations altering specific structural elements change FLC expression. Finally, quantitative experiments combined with mathematical modeling have resolved the COOLAIR–PRC2 relationship: the two operate as parallel pathways with different dynamics — a fast COOLAIR-mediated transcriptional shutdown and a slow PRC2-mediated epigenetic switch — both influenced by the cold-accumulating transcriptional regulator NTL8, and COOLAIR is dispensable for completion of vernalization.

What remains unresolved is the mechanism of the cis-repression event itself: the physical basis of the mutual exclusivity between opposing transcription units (RNA polymerase II collision, promoter occlusion, or chromatin-mediated feedback), the relative contributions of the act of antisense transcription versus the COOLAIR RNA molecule, and how these couple quantitatively to the observed histone-modification kinetics. In addition, no genotype has yet been constructed in which all antisense initiation at FLC is removed — every existing COOLAIR-defective line retains cryptic antisense promoters that become more active when the endogenous COOLAIR promoter is mutated — so strict necessity of antisense transcription has not been tested, and the genome-wide generality hypothesis remains unquantified.

Problem Statement

The problem, inherited from Swiezewski et al. (2009), has one scientific objective — to establish whether and how transcription initiating at the 3′ end of a gene causally represses transcription of the cognate sense gene in cis — with three linked components, all of which are required to settle the objective:

  1. Molecular mechanism of cis repression at FLC. Determine the physical mechanism by which COOLAIR antisense transcription quantitatively reduces sense FLC nascent transcription during cold: the relative contributions of transcriptional interference between the opposing transcription units (RNA polymerase II collision or promoter occlusion) versus effects of the antisense transcription process or the COOLAIR RNA product on chromatin (coupling of COOLAIR splicing and termination to loss of H3K36me3 and other histone modifications), and the causal chain that renders sense and antisense transcription mutually exclusive at individual loci. A complete answer should account quantitatively for the known kinetics: early COOLAIR induction during cold, progressive sense-transcription shutdown, and the parallel PRC2-dependent H3K27me3 nucleation.

  2. Necessity of antisense transcription. Construct a genotype in which all antisense initiation at FLC is eliminated — the endogenous COOLAIR promoter together with the cryptic antisense promoters that are activated when it is mutated or deleted — and determine from quantitative silencing dynamics (rate and cell-to-cell uniformity of sense transcriptional shutdown, H3K27me3 nucleation kinetics, and vernalization response) whether any antisense transcription at the locus is required. All COOLAIR-defective genotypes produced so far retain cryptic antisense transcription, so this question is open in its strict form even though COOLAIR is known to be dispensable for completion of vernalization. A negative answer — no antisense transcription required at the locus — would refute the causal premise of component 1, so it must be reported as such rather than as a resolution of component 1.

  3. Genome-wide generality. Quantify, by systematic functional tests across the genome, how frequently antisense initiation from 3′ regions of genes causally regulates cognate sense transcription in response to environmental or developmental cues, distinguishing causal regulation from transcriptional coincidence.

References

  1. Swiezewski S, Liu F, Magusin A, Dean C. Cold-induced silencing by long antisense transcripts of an Arabidopsis Polycomb target. Nature 462:799-802 (2009). DOI: 10.1038/nature08618. https://doi.org/10.1038/nature08618
  2. Sun Q, Csorba T, Skourti-Stathaki K, Proudfoot NJ, Dean C. R-loop stabilization represses antisense transcription at the Arabidopsis FLC locus. Science 340:619-621 (2013). DOI: 10.1126/science.1234848. https://doi.org/10.1126/science.1234848
  3. Marquardt S, Raitskin O, Wu Z, Liu F, Sun Q, Dean C. Functional consequences of splicing of the antisense transcript COOLAIR on FLC transcription. Molecular Cell 54:156-165 (2014). DOI: 10.1016/j.molcel.2014.03.026. https://doi.org/10.1016/j.molcel.2014.03.026
  4. Wang ZW, Wu Z, Raitskin O, Sun Q, Dean C. Antisense-mediated FLC transcriptional repression requires the P-TEFb transcription elongation factor. Proceedings of the National Academy of Sciences USA 111:7468-7473 (2014). DOI: 10.1073/pnas.1406635111. https://doi.org/10.1073/pnas.1406635111
  5. Csorba T, Questa JI, Sun Q, Dean C. Antisense COOLAIR mediates the coordinated switching of chromatin states at FLC during vernalization. Proceedings of the National Academy of Sciences USA 111:16160-16165 (2014). DOI: 10.1073/pnas.1419030111. https://doi.org/10.1073/pnas.1419030111
  6. Castaings L, Bergonzi S, Albani MC, Kemi U, Savolainen O, Coupland G. Evolutionary conservation of cold-induced antisense RNAs of FLOWERING LOCUS C in Arabidopsis thaliana perennial relatives. Nature Communications 5:4457 (2014). DOI: 10.1038/ncomms5457. https://doi.org/10.1038/ncomms5457
  7. Li P, Tao Z, Dean C. Phenotypic evolution through variation in splicing of the noncoding RNA COOLAIR. Genes and Development 29:696-701 (2015). DOI: 10.1101/gad.258814.115. https://doi.org/10.1101/gad.258814.115
  8. Wu Z, Ietswaart R, Liu F, Yang H, Howard M, Dean C. Quantitative regulation of FLC via coordinated transcriptional initiation and elongation. Proceedings of the National Academy of Sciences USA 113:218-223 (2016). DOI: 10.1073/pnas.1518369112. https://doi.org/10.1073/pnas.1518369112
  9. Rosa S, Duncan S, Dean C. Mutually exclusive sense-antisense transcription at FLC facilitates environmentally induced gene repression. Nature Communications 7:13031 (2016). DOI: 10.1038/ncomms13031. https://doi.org/10.1038/ncomms13031
  10. Hawkes EJ, Hennelly SP, Novikova IV, Irwin JA, Dean C, Sanbonmatsu KY. COOLAIR antisense RNAs form evolutionarily conserved elaborate secondary structures. Cell Reports 16:3087-3096 (2016). DOI: 10.1016/j.celrep.2016.08.045. https://doi.org/10.1016/j.celrep.2016.08.045
  11. Chen M, Penfield S. Feedback regulation of COOLAIR expression controls seed dormancy and flowering time. Science 360:1014-1017 (2018). DOI: 10.1126/science.aar7361. https://doi.org/10.1126/science.aar7361
  12. Tian Y, Zheng H, Zhang F, Wang S, Ji X, Xu C, He Y, Ding Y. PRC2 recruitment and H3K27me3 deposition at FLC require FCA binding of COOLAIR. Science Advances 5:eaau7246 (2019). DOI: 10.1126/sciadv.aau7246. https://doi.org/10.1126/sciadv.aau7246
  13. Jiao F, Pahwa K, Manning M, Dochy N, Geuten K. Cold induced antisense transcription of FLOWERING LOCUS C in distant grasses. Frontiers in Plant Science 10:72 (2019). DOI: 10.3389/fpls.2019.00072. https://doi.org/10.3389/fpls.2019.00072
  14. Zhao Y, Zhu P, Hepworth J, Bloomer R, Antoniou-Kourounioti RL, Doughty J, Heckmann A, Xu C, Yang H, Dean C. Natural temperature fluctuations promote COOLAIR regulation of FLC. Genes and Development 35:888-898 (2021). DOI: 10.1101/gad.348362.121. https://doi.org/10.1101/gad.348362.121
  15. Jeon M, Jeong G, Yang Y, Luo X, Jeong D, Kyung J, Hyun Y, He Y, Lee I. Vernalization-triggered expression of the antisense transcript COOLAIR is mediated by CBF genes. eLife 12:e84594 (2023). DOI: 10.7554/eLife.84594. https://doi.org/10.7554/eLife.84594
  16. Yang M, Zhu P, Cheema J, Bloomer R, Mikulski P, Liu Q, Zhang Y, Dean C, Ding Y. In vivo single-molecule analysis reveals COOLAIR RNA structural diversity. Nature 609:394-399 (2022). DOI: 10.1038/s41586-022-05135-9. https://doi.org/10.1038/s41586-022-05135-9
  17. Mateo-Bonmatí E, Montez M, Maple R, Fiedler M, Fang X, Saalbach G, Passmore LA, Dean C. A CPF-like phosphatase module links transcription termination to chromatin silencing. Molecular Cell 84:2272-2286 (2024). DOI: 10.1016/j.molcel.2024.05.016. https://doi.org/10.1016/j.molcel.2024.05.016
  18. Nielsen M, Menon G, Zhao Y, Mateo-Bonmatí E, Wolff P, Zhou S, Howard M, Dean C. COOLAIR and PRC2 function in parallel to silence FLC during vernalization. Proceedings of the National Academy of Sciences USA 121:e2311474121 (2024). DOI: 10.1073/pnas.2311474121. https://doi.org/10.1073/pnas.2311474121
  19. Long X, Cai Y, Wang H, Liu Y, Huang X, Xuan H, Li W, Zhang X, Zhang H, Fang X, He H, Xu G, Dean C, Yang H. Cotranscriptional splicing is required in the cold to produce COOLAIR isoforms that repress Arabidopsis FLC. Proceedings of the National Academy of Sciences USA 121:e2407628121 (2024). DOI: 10.1073/pnas.2407628121. https://doi.org/10.1073/pnas.2407628121
  20. Zeng X, Gao Z, Gu J, Qin G, Ou Y, Gao Z, Wang P, He Y. CK2 kinase-PRC2 signalling regulates genome-wide H3K27 trimethylation and transduces prolonged cold exposure into epigenetic cold memory in plants. Nature Plants 11:1572-1590 (2025). DOI: 10.1038/s41477-025-02054-1. https://doi.org/10.1038/s41477-025-02054-1
  21. Schulten A, Jang GJ, Payne-Dwyer A, Fiedler M, Nielsen ML, Mateo-Bonmatí E, Bienz M, Leake MC, Dean C. VEL-dependent polymerization maintains the chromatin association of Polycomb proteins for the switch to epigenetic silencing. Molecular Cell 85:3321-3332 (2025). DOI: 10.1016/j.molcel.2025.08.002. https://doi.org/10.1016/j.molcel.2025.08.002
  22. Maple R, Menon G, Duncan S, Schulten A, Sindalovskaya M, Yang H, Bloomer R, Howard M, Dean C. Co-transcriptional splicing changes combine with reduced productive transcription initiation for cold-induced repression of FLC. Nucleic Acids Research 54:gkaf1462 (2026). DOI: 10.1093/nar/gkaf1462. https://doi.org/10.1093/nar/gkaf1462