Phytochrome B integrates light and temperature signals in Arabidopsis

See allHide authors and affiliations

Science  18 Nov 2016:
Vol. 354, Issue 6314, pp. 897-900
DOI: 10.1126/science.aaf5656

Combining heat and light responses

Plants integrate a variety of environmental signals to regulate growth patterns. Legris et al. and Jung et al. analyzed how the quality of light is interpreted through ambient temperature to regulate transcription and growth (see the Perspective by Halliday and Davis). The phytochromes responsible for reading the ratio of red to far-red light were also responsive to the small shifts in temperature that occur when dusk falls or when shade from neighboring plants cools the soil.

Science, this issue p. 897, p. 886; see also p. 832


Ambient temperature regulates many aspects of plant growth and development, but its sensors are unknown. Here, we demonstrate that the phytochrome B (phyB) photoreceptor participates in temperature perception through its temperature-dependent reversion from the active Pfr state to the inactive Pr state. Increased rates of thermal reversion upon exposing Arabidopsis seedlings to warm environments reduce both the abundance of the biologically active Pfr-Pfr dimer pool of phyB and the size of the associated nuclear bodies, even in daylight. Mathematical analysis of stem growth for seedlings expressing wild-type phyB or thermally stable variants under various combinations of light and temperature revealed that phyB is physiologically responsive to both signals. We therefore propose that in addition to its photoreceptor functions, phyB is a temperature sensor in plants.

Plants have the capacity to adjust their growth and development in response to light and temperature cues (1). Temperature-sensing helps plants determine when to germinate, adjust their body plan to protect themselves from adverse temperatures, and flower. Warm temperatures as well as reduced light resulting from vegetative shade promote stem growth, enabling seedlings to avoid heat stress and canopy shade from neighboring plants. Whereas light perception is driven by a collection of identified photoreceptors—including the red/far-red light-absorbing phytochromes; the blue/ultraviolet-A (UV-A) light–absorbing cryptochromes, phototropins, and members of the Zeitlupe family; and the UV-B–absorbing UVR8 (2)—temperature sensors remain to be established (3). Finding the identity (or identities) of temperature sensors would be of particular relevance in the context of climate change (4).

Phytochrome B (phyB) is the main photoreceptor controlling growth in Arabidopsis seedlings exposed to different shade conditions (5). Like others in the phytochrome family, phyB is a homodimeric chromoprotein, with each subunit harboring a covalently bound phytochromobilin chromophore. phyB exists in two photo-interconvertible forms: a red light–absorbing Pr state that is biologically inactive and a far-red light–absorbing Pfr state that is biologically active (6, 7). Whereas Pr arises upon assembly with the bilin, formation of Pfr requires light, and its levels are strongly influenced by the red/far-red light ratio. Consequently, because red light is absorbed by photosynthetic pigments, shade light from neighboring vegetation has a strong impact on Pfr levels by reducing this ratio (8). phyB Pfr also spontaneously reverts back to Pr in a light-independent reaction called thermal reversion (911). Traditionally, thermal reversion was assumed to be too slow relative to the light reactions to affect the Pfr status of phyB, even under moderate irradiances found in natural environments, but two observations contradict this view. First, the formation of phyB nuclear bodies, which reflects the status of Pfr, is affected by light up to irradiances much higher than expected if thermal reversion were slow (12). Second, it is now clear that thermal reversion occurs in two steps. Although the first step, from the Pfr:Pfr homodimer (D2) to the Pfr:Pr heterodimer (D1), is slow (kr2), the second step, from the Pfr:Pr heterodimer to the Pr:Pr homodimer (D0), is almost two orders of magnitude faster (kr1) (Fig. 1A) (11).

Fig. 1 The status of phyB responds to light and temperature.

(A) Three-stage model of phyB (11). Our working hypothesis is that D2 integrates light cues (via k1 and k2) and temperature cues (via kr2 and mainly kr1). (B to E) Warm temperatures reduce Pfr levels of full-length recombinant phyB exposed in vitro to 1 [(B) and (D)] or 5.1 [(C) and (E)] μmol m−2 s−1 of continuous red light. [(B) and (C)] Absorbance kinetics (maximal absorption decreased with temperature, P < 0.05). [(D) and (E)] ∆ absorbance in samples incubated in darkness or exposed to continuous red light to reach a steady state. The difference between Δ absorbance at 665 and 725 nm decreased with temperature (P < 0.01). (F) Warm temperatures reduce the levels of Pfr and D2 in vivo measured in phyA mutant seedlings overexpressing phyB (9) exposed to 1 μmol m−2 s−1 red light. Means ± SE of three biological replicates. (G) Warm temperatures increase kr1 [calculated from (F), P < 0.001].

Physiologically relevant temperatures could change the magnitude of kr1 and consequently affect Pfr and D2 levels, even under illumination (Fig. 1A). To test this hypothesis, we used in vitro and in vivo spectroscopy and analysis of phyB nuclear bodies by means of confocal microscopy. For the first of these approaches, we produced recombinant full-length phyB bearing its phytochromobilin chromophore. When irradiated under continuous red light, the in vitro absorbance at 725 nm reached lower values at higher temperatures, which is indicative of reduced steady-state levels of Pfr (Fig. 1, B and C). We calculated the differences between the steady-state absorbance spectra in darkness and continuous red light (∆ absorbance). The amplitude between the maximum and minimum peaks of ∆ absorbance, which represents the amount of Pfr, strongly decreased between 10 and 30°C (Fig. 1, D and E). This characteristic of phyB differs from the typical behavior of enzymes, which exhibit increased activity over the same temperature range (13).

We also measured with in vivo spectroscopy the steady-state levels of phyB Pfr in seedlings irradiated with continuous red or white light at different temperatures (applied only during the irradiation). Increasing temperatures reduced both the total pool of Pfr and that of D2 (Fig. 1F and fig. S1), which is considered to be the physiologically relevant species for phyB (11). Using these data, we determined kr1, which increased with temperature (Fig. 1G).

phyB nuclear body formation increases with irradiance and red/far-red light ratio (12, 14) because it depends on D2 (11). As a proxy for temperature impact on D2, we used the difference in nuclear body formation in lines of the phyB-9–null mutant rescued with unmodified phyB [phyB–yellow fluorescent protein (YFP)] or either of two chromophore pocket mutants that suppress Pfr thermal reversion in vitro with little to no effect on photoconversion (phyBY361F-YFP and phyBR582A-YFP) (15, 16). De-etiolated (green) seedlings were transferred to the different light conditions (irradiances and red/far-red light ratios) representative of unfiltered sunlight, canopy shade, or cloudy days, in combination with different temperatures applied only during the light treatments (fig. S2). The nuclear body size of phyBY361F-YFP and phyBR582A-YFP was not significantly affected by irradiance (fig. S3) and strongly affected by the red/far-red ratio (fig. S4). This is consistent with the notion that irradiance responses depend on kr1 and kr2 (11), which are affected in the mutants. The size of phyB nuclear bodies varied quadratically with temperature and was largest at ~20°C (Fig. 2A and fig. S5). We tested the hypothesis that the negative phase of this response to temperature is the manifestation of enhanced thermal reversion reducing D2. Toward this aim, we modeled the average size of the phyBY361F-YFP and phyBR582A-YFP nuclear bodies (tables S1 and S2) as a function of both D2 (11) and temperature effects not mediated by changes in D2 (fig. S6). Then, we used this restricted model to predict D2 levels from phyB nuclear body sizes in wild-type lines (Fig. 2B). The difference between the apparent log D2 in wild-type and the log D2 of phyBY361F and phyBR582A in the same light condition is shown in Fig. 2C (difference averaged for all light conditions). The results indicate that high temperatures decrease the apparent D2 for the wild-type phyB under a wide range of light conditions.

Fig. 2 phyB nuclear bodies respond to light and temperature.

(A) Dual response of phyB-YFP nuclear bodies to temperature (white light, 10 μmol m−2 s−1). Scale bar, 5 μm. (B) Estimation of D2 in the wild type by using its average phyB nuclear body size (NB) as input in the model relating NB to D2 in lines expressing stabilized phyB (phyBY361F-YFP and phyBR582A-YFP). (C) Impact of temperature on D2. Difference in log-transformed D2 averaged for 5 to 11 conditions (±SE) covering a wide range of irradiances and red/far-red ratios (temperature effect, P < 0.05).

By using the three approaches above, we showed that the activity of phyB decreases with increasing temperature (Figs. 1 and 2), suggesting two possible biological outcomes. One is that downstream changes in phyB signaling compensate for the temperature effect. The circadian clock provides an example of temperature compensation (17). The other is that phyB perception of temperature cues controls the physiological output. A prediction of the latter hypothesis is that phyB activity (D2) should similarly affect growth independently of whether it is altered by light, temperature, or mutations that stabilize phyB. To test this prediction, we cultivated Arabidopsis seedlings (including phyB genetic variants) at the same irradiance and temperature, sorted them to the different light and temperature environments (fig. S2), and modeled growth under these conditions (table S3) as a function of D2.

The growth responses to temperature (fig. S7) and light (18) are not exclusively mediated by phyB (D2). Thus, we built the model in two steps: first, fitting univariate submodels describing the relationship between growth and the individual factors (D2, temperature effects not mediated by changes in D2, and activity of other photo-sensory receptors), and then combining those components in the final model. To quantify the contribution of D2 (fig. S8), we used growth at 30°C (no low-temperature inhibition of growth) (fig. S9) of all genotypes, including the stabilized phyB variants and the phyB-null mutant (D2 = 0). To quantify the effects of temperature not mediated by changes in D2 (fig. S9B), we used the phyB mutant (no phyB-mediated inhibition) at 1 μmol m−2 s−1 (at this irradiance, and at 30°C, growth is maximal, indicating that other photoreceptors do not make a strong contribution). To quantify the contribution of other photoreceptors (fig. S10), we used the phyB mutant (no phyB-mediated inhibition) at a range of irradiances at 30°C (no low-temperature growth inhibition). The only statistically significant interaction among these terms was between D2 and temperature effects not mediated by changes in D2 (table S4). Therefore, in the final model, growth was inversely related to terms representing the actions of D2, low temperatures (not mediated by changes in D2), other photo-sensory receptors, and the synergistic interaction between D2 and low temperature (not mediated by changes in D2).

We then fitted to the model growth for all 200 light-temperature-genotype combinations. The relationship between observed and predicted data showed no systematic deviation from the 1:1 correlation for the different light (Fig. 3A), temperature (Fig. 3B), or genetic variants with altered Pfr stability (Fig. 3C). Predicted data were obtained with D2 values affected by light, temperature, and genotype. To test the significance of temperature effects mediated by changes in the status of phyB, we recalculated growth by using D2 modified by light and genotype but not by temperature (constant 10°C). This adjustment reduced the growth model goodness of fit (Fig. 3B, inset), indicating that the contribution of phyB-mediated temperature effects on growth is statistically significant and should not be neglected. Because we estimated the effect of D2 using data from a single temperature (fig. S8), our growth model is not based on the assumption that D2 changes with temperature, thus providing confidence that the latter conclusion is genuine.

Fig. 3 phyB mediates growth responses to light and temperature.

(A to C) Observed values of hypocotyl growth (G) in white light–grown seedlings of eight genotypes exposed to 25 combinations of irradiance and temperature versus the values predicted by the growth model. The different irradiances (A), temperatures (B), and genotypes (C) are color-coded to show that the relationship between observed and predicted values is not biased for any of these factors (within the range tested here). Col, Columbia wild type; phyB, phyB null mutant; phyB, phyBY361F, and phyBR582A, transgenic lines expressing wild-type or mutated phyB in the phyB null mutant background. [(B), inset] The goodness of fit of the model (Pearson’s χ2 test) is greatly deteriorated when temperature effects on D2 are not incorporated (both versions of the model have the same number of parameters). (D) Contribution to the inhibition of growth of each one of the three temperature-dependent terms of the growth model. Topmost line is the horizontal base line of no low-temperature effects (G incorporating only light effects at 30°C). Downward, the lines indicate G calculations successively incorporating the phyB-dependent temperature effects, the phyB-temperature interaction, and the phyB-independent temperature effect. The colored areas highlight the contribution of each additional term incorporated in the calculations.

We used the growth model to compare the contribution of each of the three temperature-dependent terms to the inhibition of growth by low temperatures. phyB-mediated effects of temperature contribute to the overall temperature response (Fig. 3D). The effects were large at low irradiances, decreased with intermediate irradiances (light reactions become increasingly important), and increased again at higher irradiances because now D2 affects growth more strongly.

Phytochromes were discovered and have been studied on the basis of their roles as light receptors in plants (6, 7). However, our observations that temperature alters the amount of D2 for phyB (Figs. 1 and 2) and its physiological output in a manner similar to light (Fig. 3) indicate that phyB should also be defined as a temperature cue receptor. phyB requires light to comply with this temperature function by needing light to initially generate the unstable but bioactive Pfr state. Temperature affects the Pfr status of phyB mainly via kr1 in the light (Fig. 1) and via kr2 during the night (19). Receptors are often activated by their ligands; although phyB is activated by red light, it is inactivated by far-red light and high temperatures. This combination of light and temperature perception would serve to integrate the signals controlling photo- and thermo-morphogenesis in ways that optimize the growth of plants exposed to a wide range of environments.

Correction (17 November 2016): Report: "Phytochrome B integrates light and temperature signals in Arabidopsis" by M. Legris et al. (18 November 2016, p. 897). This paper was originally published online as First Release on 27 October 2016. This information has been restored at the end of the article.

Supplementary Materials

Materials and Methods

Figs. S1 to S11

Tables S1 to S4

References (2026)

References and Notes

Acknowledgments: We thank F. Venezia (University of Freiburg) for his valuable help with the analytical equations. This work was supported by Agencia Nacional de Promoción Científica y Tecnológica (grants PICT 2012-1396 and PICT 2013-1444 to J.J.C.), Alexander von Humboldt-Foundation (J.J.C.), Universidad de Buenos Aires (grant 20020100100437 to J.J.C.), Fundación Rene Baron (J.J.C.), the Excellence Initiative of the German Federal and State Governments (EXC 294, project C 20 to E.S. and A.H.), the German Research Foundation (DFG; SCHA 303/16-1 and HI 1369/5-1 to E.S. and A.H.), and the U.S. National Science Foundation (MCB-1329956 to R.D.V.). R.D.V and E.S.B. and the University of Wisconsin–Madison have filed U.S. patent application P140220US02, which relates to the use of mutants phyBY361F and phyBR582A in agriculture. The supplementary materials contain additional data.

Stay Connected to Science

Navigate This Article