Vernalization, the acquisition of flowering competence after prolonged winter cold in Arabidopsis, requires progressive cold-induced upregulation of VERNALIZATION INSENSITIVE3 (VIN3) to epigenetically silence the floral repressor FLOWERING LOCUS C (FLC). A quantitative model of this network (Antoniou-Kourounioti et al., Cell Systems, 2018) inferred that temperature sensing is distributed across the network rather than localized to one thermosensor, and decomposed the temperature inputs to VIN3 into three phenomenological elements: a long-term integrator L (weeks), a short-term memory of warm spikes S (about a day, reset each evening), and a current-temperature sensor C (1-2 hours). At the time, none had a known molecular identity. Subsequent work identified the NTL8 transcription factor, which accumulates slowly in the cold because slow growth reduces protein dilution, as the molecular basis of the long-term element L. The surviving open core is the molecular and biophysical identification of the two fast thermosensory inputs, C and S: which genes, proteins, RNA-based mechanisms, or chromatin processes implement the rapid temperature response of VIN3 transcription and the circadian-coupled memory of warm spikes, such that their perturbation abolishes the corresponding kinetic signatures under defined fluctuating temperature regimes.

Background

Vernalization is the process by which prolonged exposure to winter cold renders plants competent to flower in spring, and in Arabidopsis thaliana it is centered on the floral repressor gene FLOWERING LOCUS C (FLC). FLC is strongly expressed in autumn; extended cold progressively silences it through Polycomb Repressive Complex 2 (PRC2), a chromatin-modifying complex that deposits the repressive histone mark H3K27me3, thereby locking FLC in an epigenetically silenced state that is stable after cold ends. Cold induction of VERNALIZATION INSENSITIVE3 (VIN3), a PHD-finger protein that recruits PRC2 to FLC chromatin (Sung and Amasino, 2004), is graded, increasing slowly over weeks of cold, and is required for the epigenetic switching phase. Field experiments showed that plants respond not simply to mean temperature: transient daily warm spikes in autumn strongly delay VIN3 upregulation and FLC silencing, so the system effectively measures the 'absence of warmth' over multiple timescales (Hepworth et al., 2018). To formalize this, Antoniou-Kourounioti et al. (2018) fitted a mathematical model (the LSCD model) in which the VIN3 transcription rate is proportional to the product of four inputs: L, a long-term element produced only in the cold (below roughly 15 degrees Celsius) that degrades on a week timescale and behaves like the 'day-degree' counters used in crop modeling (it records time spent within a temperature interval rather than the temperature itself); C, a fast element (responding within 1-2 hours) that takes low values above an intermediate temperature range, maximal values below it, and graded values in between, producing the rapid reduction of VIN3 at high temperature; S, a short-term memory element that is lowered when temperature has spiked high since the previous afternoon and is reset every evening, implying coupling to the circadian clock; and D, a temperature-independent circadian input producing the afternoon peak of VIN3 transcription. The model, also including temperature-sensitive steps downstream of VIN3 in FLC silencing and a VIN3-independent cold input to FLC, accurately predicted FLC dynamics in field data. Its central conclusion is that temperature sensing is broadly distributed across the network, with each element extracting a different feature of the plant's warm/cold exposure history, rather than being localized to a single dedicated thermosensor molecule. The paper stated explicitly that there were no clear molecular candidates for L, S, and C, and that focused genetic screens in defined temperature regimes would be needed to identify them. Subsequently, Zhao et al. (2020) identified the NTL8 transcription factor through a forward genetic screen: NTL8 is produced at a roughly constant rate and is long-lived, so it accumulates slowly in the cold because slow growth at low temperature reduces protein dilution, and it directly upregulates VIN3. This growth-dilution mechanism accounts for the slow, fluctuation-averaging kinetics of the long-term element L, and the original authors' own 2021 review accordingly describes the Long-term pathway as having known components (NTL8). In contrast, the two fast elements remain phenomenological constructs: no gene, protein, RNA mechanism, or chromatin process has been shown to implement the current-temperature response (C) or the evening-reset memory of warm spikes (S), although adjacent molecular findings exist — the clock proteins CCA1 and LHY directly bind a vernalization-responsive cis-element in the VIN3 promoter (molecular basis of the D input), the heat-shock transcription factor HSFB2b represses VIN3 as a warm-conditioned fine-tuning mechanism, and the PRC2 component VRN2, which is stabilized in the cold, is a candidate for the temperature sensitivity of FLC silencing beyond VIN3 induction. Reviews as late as 2024-2026 continue to describe plant temperature sensation as only beginning to be understood, with no identification of S or C.

Problem Statement

The problem, as posed by Antoniou-Kourounioti et al. (2018) and narrowed by subsequent work, is to map the phenomenological thermosensory elements of the vernalization model onto molecular mechanisms in Arabidopsis. The long-term element LL is now substantially accounted for by NTL8 and its growth-dilution-based accumulation (Zhao et al., 2020). The surviving open core is: what are the biophysical and molecular identities of the two fast thermosensory inputs to VIN3 transcription — the current-temperature element CC, which reduces VIN3 within 1-2 hours at high temperature and gives a graded response across an intermediate range, and the short-term memory element SS, which is lowered when temperature spikes above a threshold since the previous afternoon and is reset every evening in a circadian-coupled manner? A complete answer should identify the gene(s), protein(s), RNA-based mechanism(s), or chromatin process(es) responsible, demonstrate by specific perturbation (for example, mutants or inducible alleles) that the corresponding kinetic signature is selectively abolished or altered under defined fluctuating temperature regimes, and show quantitatively that the identified mechanism reproduces the timescales, thresholds, and evening reset that define the model elements. Within the same objective, an answer may also establish related secondary questions raised by the source paper and by subsequent work: whether NTL8 fully accounts quantitatively for LL; the molecular basis of the remaining temperature-sensitive steps of the network (the VIN3-independent cold input to FLC and the temperature sensitivity of VIN3-dependent FLC silencing, for which VRN2 is a candidate); and whether the phenomenological SS/CC decomposition itself requires revision in favor of an alternative network structure that explains the same data. The question is not restricted to a single answer form: genetic, biochemical, or reconstitution-based demonstrations are all admissible, provided the fast temperature dependences of VIN3 regulation are explained mechanistically rather than remaining fitted parameters.

References

  1. Antoniou-Kourounioti RL, Hepworth J, Heckmann A, Duncan S, Qüesta J, Rosa S, Säll T, Holm S, Dean C, Howard M. Temperature Sensing Is Distributed throughout the Regulatory Network that Controls FLC Epigenetic Silencing in Vernalization. Cell Systems 7(6):643-655.e9 (2018). DOI: 10.1016/j.cels.2018.10.011. https://doi.org/10.1016/j.cels.2018.10.011
  2. Zhao Y, Antoniou-Kourounioti RL, Calder G, Dean C, Howard M. Temperature-dependent growth contributes to long-term cold sensing. Nature 583:825-829 (2020). DOI: 10.1038/s41586-020-2485-4. https://doi.org/10.1038/s41586-020-2485-4
  3. Antoniou-Kourounioti RL, Zhao Y, Dean C, Howard M. Feeling Every Bit of Winter - Distributed Temperature Sensitivity in Vernalization. Frontiers in Plant Science 12:628726 (2021). DOI: 10.3389/fpls.2021.628726. https://doi.org/10.3389/fpls.2021.628726
  4. Hepworth J, Antoniou-Kourounioti RL, Bloomer RH, Selga C, Berggren K, Cox D, Collier Harris BR, Irwin JA, Holm S, Säll T, Howard M, Dean C. Absence of warmth permits epigenetic memory of winter in Arabidopsis. Nature Communications 9:639 (2018). DOI: 10.1038/s41467-018-03065-7. https://doi.org/10.1038/s41467-018-03065-7
  5. Sung S, Amasino RM. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 427:159-164 (2004). DOI: 10.1038/nature02195. https://doi.org/10.1038/nature02195
  6. Kyung J, Jeon M, Jeong G, Shin Y, Seo E, Yu J, Kim H, Park CM, Hwang D, Lee I. The two clock proteins CCA1 and LHY activate VIN3 transcription during vernalization through the vernalization-responsive cis-element. The Plant Cell 34(3):1020-1037 (2022). DOI: 10.1093/plcell/koab304. https://doi.org/10.1093/plcell/koab304
  7. Jeong G, Jeon M, Shin J, Lee I. HEAT SHOCK TRANSCRIPTION FACTOR B2b acts as a transcriptional repressor of VIN3, a gene induced by long-term cold for flowering. Scientific Reports 12:10963 (2022). DOI: 10.1038/s41598-022-15052-6. https://doi.org/10.1038/s41598-022-15052-6
  8. 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 & Development 35(11):888-898 (2021). DOI: 10.1101/gad.348362.121. https://doi.org/10.1101/gad.348362.121
  9. Nielsen M, Menon G, Zhao Y, Mateo-Bonmati 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(4):e2311474121 (2024). DOI: 10.1073/pnas.2311474121. https://doi.org/10.1073/pnas.2311474121
  10. Kennedy A, Li M, Vandeperre A, Hameed MU, Van Dyck M, Engelen S, Preston JC, Geuten K. Transcription factor VRT2 reinitiates vernalization when interrupted by warm temperatures in a temperate grass model. Plant Physiology 196(4):2614-2624 (2024). DOI: 10.1093/plphys/kiae498. https://doi.org/10.1093/plphys/kiae498
  11. James AB, Sharples C, Laird J, Armstrong EM, Guo W, Tzioutziou N, Zhang R, Brown JWS, Nimmo HG, Jones MA. REVEILLE2 thermosensitive splicing: a molecular basis for the integration of nocturnal temperature information by the Arabidopsis circadian clock. New Phytologist 241(1):283-297 (2024). DOI: 10.1111/nph.19339. https://doi.org/10.1111/nph.19339
  12. Casal JJ, Murcia G, Bianchimano L. Plant Thermosensors. Annual Review of Genetics 58:135-158 (2024). DOI: 10.1146/annurev-genet-111523-102327. https://doi.org/10.1146/annurev-genet-111523-102327
  13. Yadav AS, Sureshkumar S, Sinha AK, Balasubramanian S. Dispersed components drive temperature sensing and response in plants. Science 388(6752):1161-1166 (2025). DOI: 10.1126/science.adv5407. https://doi.org/10.1126/science.adv5407
  14. Montez M, Schulten A, Menon G, Dean C. Molecular integration of seasonal temperature signals in flowering time control. Nature Reviews Genetics 27(9):695-707 (2026). DOI: 10.1038/s41576-026-00979-7. https://doi.org/10.1038/s41576-026-00979-7