Photoperiodic control of seasonal growth is mediated by ABA acting on cell-cell communication

See allHide authors and affiliations

Science  13 Apr 2018:
Vol. 360, Issue 6385, pp. 212-215
DOI: 10.1126/science.aan8576

Dormancy by communication shutdown

Trees become dormant in winter, with encapsulated buds protected against harsh conditions. Tylewicz et al. found that, as the days get shorter, communication channels between cells in aspen trees shut down. The blocked plasmodesmata sequester the dormant meristems from growth signals. Growth-promoting signals can be turned on and off relatively rapidly, but the closed plasmodesmata are not so nimble. Thus, despite the occasional sunny day, the trees stay dormant until spring.

Science, this issue p. 212


In temperate and boreal ecosystems, seasonal cycles of growth and dormancy allow perennial plants to adapt to winter conditions. We show, in hybrid aspen trees, that photoperiodic regulation of dormancy is mechanistically distinct from autumnal growth cessation. Dormancy sets in when symplastic intercellular communication through plasmodesmata is blocked by a process dependent on the phytohormone abscisic acid. The communication blockage prevents growth-promoting signals from accessing the meristem. Thus, precocious growth is disallowed during dormancy. The dormant period, which supports robust survival of the aspen tree in winter, is due to loss of access to growth-promoting signals.

Dormancy protects meristematic cells of perennial plants in temperate and boreal ecosystems by preventing growth during winter. Release from dormancy enables reinitiation of growth when favorable conditions return in spring (1). Shorter photoperiods as winter approaches (2) induce growth cessation, formation of a bud that encloses the arrested leaf primordia and shoot apical meristem (SAM) (Fig. 1A), and bud dormancy (3, 4). Longer photoperiods alone cannot promote growth in dormant buds; prolonged exposure to low temperatures is required to release dormancy (5, 6). We show that blockage of symplastic communication mediated by the action of abscisic acid (ABA) is part of the photoperiodically controlled dormancy mechanism in hybrid aspen.

Fig. 1 Hybrid aspen plants with attenuated ABA responses fail to establish dormancy.

(A to C) Buds of (A) wild type, (B) abi1-1 line 1, and (C) abi1-1 line 3 after 11 weeks of short photoperiod. (D to F) Unlike in (D) WT, buds burst in (E) abi1-1 line 1 and (F) abi1-1 line 3. (G to I) Transmission electron microscopy (TEM) micrographs of apices of actively growing (G) WT plants, (H) abi1-1 line 1, and (I) abi1-1 line 3, showing plasmodesmata lacking electron-dense dormancy sphincters. (J to O) Sphincters are observed after 5 and 10 weeks of short photoperiod in apices of [(J) and (M)] wild-type plants (indicated with arrowheads), but not [(K) and (N)] abi1-1 line 1 or [(L) and (O)] abi1-1 line 3. Scale bar, 200 nm.

Short photoperiods induce expression of ABA receptors and increase ABA levels in hybrid aspen buds (4, 7). ABA regulates dormancy (8). Therefore, we probed ABA’s role in photoperiodic control of bud dormancy. First, we generated hybrid aspen plants with reduced ABA responses by expressing the dominant-negative abi1-1 allele of ABI1, a key ABA-signaling gene (9). Those hybrid aspens that expressed abi1-1 had reduced ABA responses, manifested by weak induction of the ABA-inducible gene KIN2, compared with that of wild-type (WT) controls (fig. S1). We then assessed bud dormancy by exposing WT and abi1-1 plants to 11 weeks of short photoperiod followed by transfer to long photoperiod without the low-temperature treatment required for dormancy release. Both WT and abi1-1 plants ceased growth and set buds after 4 weeks of short photoperiod (Fig. 1, A to C), but after 11 weeks of short photoperiod followed by long photoperiod, WT buds remained dormant, whereas abi1-1 buds reactivated growth within 11 to 15 days (Fig. 1, D to F). Thus, attenuation of ABA responses compromised photoperiodic control of bud dormancy and not growth cessation.

We investigated transcriptomic responses to short photoperiod in WT and abi1-1 apices in order to understand ABA-mediated control of dormancy. After 6 and 10 weeks of short photoperiod, respectively, we detected 9290 and 3053 differentially expressed genes in WT and 10,514 and 2149 differentially expressed genes in abi1-1 (line 1) apices (table S1). A large number of transcripts for plasmodesmata-associated proteins responded to short photoperiod. Plasmodesmata closure (by callosic dormancy sphincters) correlates with dormancy and their opening with dormancy release in diverse plants, including hybrid aspen and charophycean algae such as Chara (6, 10, 11). Of 187 poplar homologs of Arabidopsis genes encoding proteins enriched in plasmodesmata (12), 62 and 47 were induced after 6 and 10 weeks in WT apices, respectively, and of these, 53.2 and 76.6% were differentially expressed in abi1-1 relative to WT apices at these time points (table S2). Expression of GERMIN-LIKE 10; REMORIN-LIKE 1 and 2, which are implicated in plasmodesmata function (13); and CALLOSE SYNTHASE 1, which is required for callose deposition (6), was progressively up-regulated, whereas that of GH17-39, a glucanase implicated in sphincter removal (6), was down-regulated in WT apices after 6 and 10 weeks of short photoperiod. These genes showed an altered response to short photoperiod in abi1-1 plants (fig. S2). Thus, ABA mediates short-photoperiod response of the plasmodesmata-related transcriptome.

Transcriptomic analysis prompted us to investigate ABA’s role in plasmodesmata closure (Fig. 1, G to O). Under long photoperiod, WT and abi1-1 lines 1 and 3 had similar frequencies of “closed” plasmodesmata with dormancy sphincters (12.5 versus 17.4 and 13.5%, respectively). After 5 weeks of short photoperiod, corresponding frequencies were 78% in WT and 5.5 and 17.4% in abi1-1 apices, respectively, and after 10 weeks, frequencies increased to 83.6% in WT plants but fell to 2.2 and 0.5% in abi1-1 lines 1 and 3, respectively. Thus, ABA mediates plasmodesmata closure in response to short photoperiod. Plasmodesmata closure is not required for growth cessation (because growth cessation occurs in abi1-1 plants) and indicates association of plasmodesmata closure with bud dormancy, both being mediated by the same factor, ABA.

To investigate ABA-mediated plasmodesmata closure’s role in short photoperiod–induced dormancy, we overexpressed PDLP1 (PLASMODESMATA-LOCATED PROTEIN 1), which impairs trafficking via plasmodesmata (14) and phenocopying plasmodesmata blockage by dormancy sphincters, in abi1-1 plants (fig. S3). Both abi1-1/PDLP1 double transformants and parental abi1-1 plants ceased growth and formed buds under short photoperiod (Fig. 2, A to C), but subsequent exposure to long photoperiod only reactivated growth in the latter (Fig. 2, D to F). Thus, PDLP1 expression suppressed abi1-1 plants’ bud dormancy phenotype, although KIN2 expression responses to ABA remained attenuated in abi1-1/PDLP1 (fig. S4). Thus, expression of PDLP1 was sufficient to restore bud dormancy in abi1-1/PDLP1 plants without the restoration of general ABA responses.

Fig. 2 PDLP1 expression restores bud dormancy in abi1-1 plants.

(A to C) Buds of (A) abi1-1, (B) abi1-1/PDLP1 line 1, and (C) abi1-1/PDLP1 line 3 after 11 weeks of short photoperiod. (D to F) Transfer to long photoperiod results in bud burst in (D) abi1-1 plants but not in (E) abi1-1/PDLP1 line 1 or (F) abi1-1/PDLP1 line 3.

PICKLE (PKL) is an antagonist of polycomb repression complex 2, which is implicated in seed dormancy (15, 16). PKL expression was down-regulated in WT plants but up-regulated in abi1-1 plants under short photoperiod (fig. S5). Hence, we investigated whether PKL could be involved in plasmodesmata closure and dormancy regulation mediated by ABA. Thus, we examined plasmodesmata in abi1-1 plants with suppressed PKL activity (abi1-1/PKLRNAi) (RNAi, RNA interference) (fig. S6). Under long photoperiod, frequencies of plasmodesmata with dormancy sphincters were comparable in abi1-1 (13.1%) and abi1-1/PKLRNAi lines 9 (19.4%) and 11 (18.4%) (Fig. 3, A to C). After 5 weeks of short photoperiod, the frequencies increased in the abi1-1/PKLRNAi lines (to 34.4 and 28.5%, respectively), but not abi1-1 plants (16.4%) (Fig. 3, D to F). After 10 weeks of short photoperiod, the frequencies further increased in abi1-1/PKLRNAi lines 9 and 11 to 84.6 and 74.5%, respectively, but fell in abi1-1 plants (5.2%) (Fig. 3, G to I). PKL down-regulation in abi1-1/PKLRNAi also suppressed expression defects of plasmodesmata markers evident in abi1-1 plants (fig. S7). Although both abi1-1 and abi1-1/PKLRNAi plants ceased growth and set buds (Fig. 3, J to L), abi1-1/PKLRNAi buds remained dormant and did not reactivate growth (unlike nondormant abi1-1 buds) after long photoperiod exposure after 11 weeks of short photoperiod (Fig. 3, M to O). Thus, PKL down-regulation restores plasmodesmata closure and bud dormancy defects in abi1-1 plants, suggesting that ABA mediates plasmodesmata closure and bud dormancy by suppressing PKL.

Fig. 3 PKL down-regulation restores dormancy sphincters and bud dormancy in abi1-1 plants.

(A to C) TEM micrographs of apices of actively growing (A) abi1-1 plants, (B) abi1-1/PKLRNAi line 9, and (C) abi1-1/PKLRNAi line 11, showing plasmodesmata lacking electron-dense dormancy sphincters. (D to I) After 5 and 10 weeks of short photoperiod, sphincters were not observed in [(D) and (G)] abi1-1 apices but were present in abi1-1/PKLRNAi apices of [(E) and (H)] lines 9 and [(F) and (I)] 11 (arrows). Scale bar, 500 nm. (J to L) Buds of (J) abi1-1 plants, (K) abi1-1/PKLRNAi line 9, and (L) abi1-1/PKLRNAi line 11 after 11 weeks of short photoperiod. (M to O) After a shift to long photoperiod, buds burst in (M) abi1-1 plants but not in (N) abi1-1/PKLRNAi line 9 or (O) abi1-1/PKLRNAi line 11.

Plasmodesmata closure could mediate dormancy by limiting access of SAM to growth-promotive signals. We investigated this hypothesis by analyzing responses of WT and abi1-1 buds to FLOWERING LOCUS T1 (FT1), a seasonal growth regulator induced during dormancy release and before bud growth resumes (6, 17). We grafted scions of WT and abi1-1 plants exposed to 10 weeks of short photoperiod (in order to induce plasmodesmata closure and dormancy) onto rootstocks of FT1-expressing plants (18). Although buds of WT scions did not reactivate growth, new leaves emerged from buds of abi1-1 scions under a continued short photoperiod (Fig. 4). Thus, plasmodesmata closure, as in WT plants, was associated with buds’ failure to respond to FT1 or FT1-derived growth-promotive signals, corroborating the involvement of plasmodesmata in photoperiodic control of ABA-mediated bud dormancy.

Fig. 4 FT1-expressing stocks can reactivate growth in abi1-1 scions under short photoperiod.

WT and abi1-1 buds after 10 weeks of short photoperiod before grafting, and a further 2 and 7 weeks of short photoperiod after grafting of WT and abi1-1 scions on FT1-expressing stocks. Buds remained dormant in WT scions but burst in abi1-1 scions.

Thus, short photoperiods suppress FT2, which causes growth cessation and amplifies the ABA response by enhancing levels of ABA and ABA receptors (4, 7). ABA suppresses PKL and induces callose synthase to block plasmodesmata and maintains these blockages by repressing antagonistic glucanases (fig. S8). Hence, attenuating ABA responses not only results in a failure to induce plasmodesmata closure at dormancy onset but also in fewer subsequently closed plasmodesmata. Plasmodesmata closure through PKL down-regulation or PDLP1 expression, which both target cell-cell communication, suppresses dormancy defects in abi1-1 plants. PDLP1 expression restores dormancy without suppressing ABA response defects in abi1-1 plants. Thus, plasmodesmata closure is essential to dormancy and occurs downstream of ABA-mediated control of dormancy in response to shorter photoperiods.

With plasmodesmata closed, growth arrest is maintained even in the presence of growth-promoting signals. Reopening of closed plasmodesmata in dormant buds occurs slowly and only after prolonged exposure to low temperature. Hence, dormancy prevents precocious activation of growth. On the other hand, in the absence of dormancy and plasmodesmatal closure, growth cessation induced by short photoperiod can be quickly reversed. Thus, dormancy, unlike growth cessation, adds robustness to the mechanism that is crucial for perennial survival and longevity in the face of changing seasons.

Supplementary Materials

Materials and Methods

Figs. S1 to S8

Tables S1 to S3

References (1935)

References and Notes

Acknowledgments: Funding: Grants from Vetenskapsrådet (VR-2016-04430) and the Knut and Alice Wallenberg Foundation (2014-0032) to R.P.B. are gratefully acknowledged. Author contributions: S.T., A.P., S.M., P.M., A.A., R.K.S., and J.I. performed experiments. N.M. and T.R.H. performed the transcriptomics analysis. R.P.B., D.M.E., J.L.B., and Y.H. designed experiments and provided intellectual input. All authors contributed to writing the manuscript. Competing interests: The authors declare no competing interests. Data and materials availability: Raw RNA sequencing reads are available at the European Nucleotide Archive ( under accession no. PRJEB23073. All other data needed to evaluate the conclusions in the paper are present in the paper or the supplementary materials.
View Abstract

Navigate This Article