Abstract
Plants have elaborate genetic mechanisms controlling developmental responses to environmental stimuli. A particularly important environmental parameter is temperature. Previous work has identified ELF3 and PIF4 as key players in the Arabidopsis thaliana temperature response, and the ELF3 polyglutamine (polyQ) domain as a source of functional variability in ELF3. We used transgenic analysis to test the hypothesis that ELF3 polyQ variation modulates temperature sensing as an input to temperature-mediated morphological changes. We found little evidence that the polyQ domain has a specific role in temperature sensing beyond mediating overall ELF3 function. Instead, we made the serendipitous discovery that ELF3 plays a role in thermoresponsive flowering at elevated temperatures, a response previously shown to require PIF4. Unexpectedly, ELF3’s role in thermoresponsive flowering was independent of PIF4. Here, we present evidence that ELF3 acts through the photoperiodic pathway, indicating a previously unknown symmetry between low and high ambient temperature responses. These findings tie together disparate observations into a coherent framework in which multiple pathways converge in accelerating flowering in response to temperature, with some such pathways modulated by photoperiod.
Introduction
The responses of plants to temperature variation are of central importance to food security in a changing world (Battisti and Naylor, 2009). The elucidation of genetic pathways underlying these responses is therefore a central mission of plant science (Quint et al., 2016). Many studies have made impressive strides in understanding the phenomena of circadian temperature compensation (Gould et al., 2006; Edwards et al., 2006; Thines and Harmon, 2010), thermoresponsive flowering (Blázquez et al., 2003; Balasubramanian et al., 2006; Strasser et al., 2009; Kumar et al., 2012; Song et al., 2013), and temperature effects on plant morphology (Koini et al., 2009; Johansson et al., 2014; Lee et al., 2014; Box et al., 2014; Mizuno et al., 2014b; Raschke et al., 2015). Various studies have converged on PIF4 as a master regulator of temperature responses, and ELF3 as an input to PIF4 integration, among many other genes and pathways. Given previously-described regulatory interactions between ELF3 and PIF4 (Nusinow et al., 2011; Nieto et al., 2014; Nozue et al., 2007), it is reasonable to predict that these components of thermal response operate in the same pathway for various thermal response phenotypes (Lorenzo et al., 2016). Recent reports support this expectation, focusing largely on hypocotyl elongation as a model for understanding this regulatory relationship (Box et al., 2014; Mizuno et al., 2014b; Raschke et al., 2015).
ELF3 serves to repress hypocotyl elongation by reducing PIF4 levels. This repression of PIF4 occurs at both the transcriptional level, through the role of ELF3 in the Evening Complex (EC) (Nusinow et al., 2011; Nozue et al., 2007), and at the post-translational level, through PIF4 destabilization by the phytochrome phyB in cooperation with ELF3 (Lorrain et al., 2008). Light sensing enforces circadian oscillations of the EC and other components, leading to calibration of the circadian clock (McWatters et al., 2000; Covington et al., 2001), which in turn results in diurnal repression of hypocotyl elongation through repression of PIF4 and its semi-redundant paralog PIF5 (Nusinow et al., 2011; Nozue et al., 2007). ELF3 also plays a crucial role as a flowering repressor (Zagotta et al., 1996). Consequently, elf3 null mutants are skotomorphogenic, with elongated hypocotyls under light conditions, and flower early.
PIF4 is one of a family of basic helix-loop-helix (bHLH) “phytochrome-interacting factors” (PIFs), transcription factors with overlapping functions promoting skotomorphogenesis. Under dark conditions, the PIFs act to target phyB for ubiquitin-mediated degradation by the E3 ubiquitin ligase COP1, thereby repressing photomorphogenesis (Leivar and Quail, 2011). Under light conditions, degradation of PIFs is mediated by direct interactions with photoactivated phyB (Lorrain et al., 2008). PIF4 is distinct from the other PIFs in having specific roles in temperature sensing and flowering (Proveniers and van Zanten, 2013). pif4 null mutants show short hypocotyls with photomorphogenic attributes even in the dark (Bai et al., 2012).
At elevated ambient temperatures (27°-29°), the wiring of these various signaling pathways and the role of these proteins change. Several independent studies have recently found that warm temperatures, specifically during dark periods (Thines et al., 2014), inhibit the activity of the EC by an unknown mechanism (Mizuno et al., 2014b; Box et al., 2014; Raschke et al., 2015), leading to increased expression of PIF4 and its targets (Koini et al., 2009; Proveniers and van Zanten, 2013). This increased PIF4 activity leads to several morphological temperature responses through various signaling pathways. For instance, hypocotyls are elongated following PIF4 induction of auxin signaling via an auxin biosynthesis pathway (Proveniers and van Zanten, 2013; Lee et al., 2014). PIF4 is required for the acceleration of flowering at 27°C under short photoperiods (Kumar et al., 2012; Thines et al., 2014). Chromatin accessibility at the FT promoter may also be a limiting factor for PIF4 binding (Kumar et al., 2012), though these observations have been disputed (Galvão et al., 2015). In contrast, under continuous light, pif4 null mutants have an intact temperature-dependent acceleration of flowering (Koini et al., 2009). pif1, pif3, pif4, and pif5 null mutants show no aberrant flowering phenotypes under standard growth conditions either singly or in combination (Shin et al., 2009). Lastly, pif4 null mutants lose the normal elongation of petioles under high temperatures (Koini et al., 2009). It is unclear why PIF4 does not affect thermoresponsive flowering under continuous light; yet, this phenomenon may reflect low PIF4 levels under these conditions due to inhibition by phyB. Regardless of light conditions, however, PIF4 is always necessary for petiole elongation under increased temperatures. Moreover, under longer photoperiods and higher temperature a flowering acceleration still exists (Balasubramanian et al., 2006; Koini et al., 2009), requiring that some PIF4-independent thermoresponsive flowering pathway must be in effect at higher ambient temperatures. Recent reviews of the literature tend to emphasize the primacy of PIF4 in this response (Song et al., 2013; Wigge, 2013; Liu et al., 2015), although the condition of high ambient temperature with short photoperiods is probably rare in the field.
Modeling has suggested the existence of a “Y” component of temperature sensing upstream of PIF4 (Johansson et al., 2014). Other recent studies have identified ELF3 as a plausible candidate for Y, given its function upstream of PIF4 at room temperature (Nusinow et al., 2011), its role in temperature response (Mizuno et al., 2014b; Box et al., 2014; Raschke et al., 2015), its physical interaction with PIF4, and its post-translational inhibition of PIF4 (Nieto et al., 2014). However, further studies have implicated other candidates, such as FCA (Lee et al., 2014), and mathematical modeling has suggested that ELF3/EC complex regulation alone is insufficient to explain PIF4 thermal regulation (Seaton et al., 2015; Box et al., 2014). The exact mechanism of this response has yet to be unraveled.
Specifically, the mechanism by which EC/ELF3 activity is reduced under elevated temperatures (“temperature sensing”) is not known. We recently used transgenic experiments to demonstrate that ELF3 function is dependent on the unit copy number of its C-terminal polyglutamine (polyQ) tract (Undurraga et al., 2012). This protein region is likely disordered, and disordered proteins can evince structural changes in response to physical parameters such as temperature (Uversky, 2009). Thermal remodeling of this polyQ tract is a plausible mechanism by which ELF3 activity could be modulated through temperature. Moreover, this polyQ tract shows substantial natural variation, with most variants showing strong interactions with the genetic background, such that different backgrounds prefer their endogenous ELF3-polyQ variant (Undurraga et al., 2012). Furthermore, in flies, variable repeats are associated with local temperature compensation adaptations (Sawyer, 1997). In short, the ELF3-polyQ is an attractive candidate for adaptive variation in the ecologically relevant trait of temperature response (Gemayel et al., 2010).
In this study, we used previously described transgenic polyQ variants of ELF3 in two A. thaliana genetic backgrounds to dissect the contribution of the polyQ tract to temperature response, and to better understand ELF3 function in temperature sensing. We show that polyQ repeat copy number modulates temperature sensing by affecting ELF3 function. However, this effect appears to be largely the result of overall effects on ELF3 function rather than being specific to temperature. Surprisingly, we found that ELF3’s role in thermoresponsive flowering appears to be entirely independent of PIF4 in the reference genetic background, Columbia. We postulate that ELF3’s primary role in thermoresponsive flowering is PIF4-independent and occurs through the photoperiodic pathway, and that this role is in turn dependent on the genetic background.
Results
The hypocotyl elongation temperature response is modulated by the ELF3 polyQ tract affecting overall gene function
Many recent studies have noted the involvement of ELF3 in temperature-dependent hypocotyl elongation (Mizuno et al., 2014a, 2014b; Box et al., 2014; Raschke et al., 2015), concluding that ELF3 protein activity is reduced under elevated temperatures, thereby relieving ELF3 repression of PIF4 expression. PIF4 up-regulation then leads to the observed hypocotyl elongation. We examined whether polyQ tract variation in ELF3 in two backgrounds affects hypocotyl elongation at 27° (Figure 1, Table S1). In the Ws background (Figure 1A), the endogenous ELF3 (16Q) partially rescues elf3, and another variant (9Q) fully rescues the hypocotyl temperature response. Other variants with more or fewer Qs rescue inefficiently. In the Col background, we observed a somewhat different pattern, though the same lines, in addition to the 20Q variant, showed partial rescue (Figure 1B, Table S1). However, the endogenous 7Q variant, among other variants, failed to rescue the response, consistent with previous observations that these transgenic lines are hypomorphic in the Col background (Undurraga et al., 2012). Overall, patterns of hypocotyl response to elevated temperature were not linear, though it is unclear in the Col background whether this might be due to poor complementation. Thus, the response to elevated temperature was similar to other, previously tested ELF3-dependent traits (Undurraga et al., 2012). Robust thermal responses were strongly correlated with the overall functionality of each ELF3 variant (Figure 1C), consistent with recent reports suggesting that the mechanism of ELF3 thermal response in hypocotyl elongation is the de-repression of PIF4 (Raschke et al., 2015; Nomoto et al., 2012; Mizuno et al., 2014b, 2014a, 2015; Box et al., 2014). Among Ws lines, which do show robust rescues with the endogenous variant, we cannot reject the hypothesis that the different transgenic lines respond to elevated temperature in proportion to ELF3 functionality. From our data, conclusions about the effects of polyQ variation on thermoresponsive hypocotyl elongation in the Col background would be premature. Nonetheless, our various transgenic lines remain informative as an allelic series of ELF3 function. In summary, one means by which ELF3 polyQ variation mediates thermally-responsive hypocotyl elongation is through affecting overall ELF3 functionality.
Expression of PIF4 and PIF4 targets as a function of temperature and ELF3
To evaluate the hypothesis that the temperature response defects in the transgenic lines is due to up-regulation of PIF4 and PIF4 targets, we measured transcript levels of PIF4 and its target AtHB2 in seedlings of selected lines from both backgrounds at 22°C and 27°C (Figure 1D). Like other groups (Raschke et al., 2015; Mizuno et al., 2014b), we observed an inverse relationship between ELF3 functionality and transcript levels of PIF4 and AtHB2, and also that ELF3 function has a larger effect on expression of PIF4 than of AtHB2. Similar to previous observations, there appeared to be an effect of elevated temperature on PIF4 transcript levels independent of ELF3 genotype. Col transgenic lines complement PIF4 repression poorly, consistent with their poor complementation of hypocotyl phenotypes. We observed a small effect of elevated temperature on AtHB2 expression in WT lines, but AtHB2 responses in transgenic lines (even those with intact thermal responses) were less clear. We conclude that de-repression by ELF3 is one factor potentially controlling temperature-dependent PIF4 activation. Notably, the ELF3 lines with the strongest temperature response (for example 16Q in the Ws background) showed the most robust repression and de-repression of PIF4 expression. However, it was not clear whether de-repression of PIF4 and its targets is sufficient to explain temperature response defects of elf3 null mutants.
ELF3-polyQ variation affects thermoresponsive adult morphology and flowering time
Following the expectation that ELF3’s temperature response acts through PIF4, we reasoned that ELF3 should also play a role in other PIF4-dependent thermal responses. One well-known response to elevated temperature involves adult morphologies, wherein leaf angle from the ground increases and petioles elongate. pif4 mutants do not have these responses when grown at elevated temperatures (Koini et al., 2009). We therefore measured petiole length in our ELF3-polyQ transgenic lines, expecting that, due to general PIF4 de-repression, poorly-functioning ELF3-polyQ lines would show no response (perhaps due to constitutively elongated petioles, similar to hypocotyls; Figure 2). In stark contrast to these expectations, we found that all lines had a robust petiole response to temperature (Figure 2 A, B, Table S2). This effect was apparent in both Ws (Figure 2A) and Col backgrounds (Figure 2B). Moreover, this response is actually accentuated in elf3 null mutants and in poorly-functioning ELF3 polyQ variants (Figures 2A, B).
Further, we measured flowering time in transgenic lines as the number of rosette leaves at flowering (Figures 2C, D, Table S3). Accelerated flowering is a well-known temperature response for which PIF4 is not required under longer photoperiods (Koini et al., 2009). Hence, we expected that loss of ELF3 function should also not affect thermoresponsive flowering. In contrast to our expectations, in the Col background, elf3 mutants had an abrogated flowering response to elevated temperature (Figure 2D). Moreover, most variants in the Col background entirely failed to rescue this phenotype. Previous work indicated that many of these lines fail to rescue other elf3 null phenotypes (Undurraga et al., 2012), but the involvement of ELF3 in thermoresponsive flowering is unexpected. In contrast to Col, Ws is known to lack a robust flowering response to elevated temperature under these conditions (Ibañez et al., 2015), and indeed, variants in the Ws background generally showed no thermoresponsive flowering (Figure 2C). Thus, ELF3 polyQ variation does not suffice to enhance the negligible thermoresponsive flowering in the Ws background under these conditions. In light of this data, the roles of ELF3 and PIF4 in the elevated temperature response appear to be independent of one another under these experimental conditions. This result is intriguing, given that the PIF4 pathway is the best-recognized mechanism for thermoresponsive flowering at high temperatures (Song et al., 2013; Wigge, 2013; Kumar et al., 2012; Liu et al., 2015). Therefore, we suggest that ELF3 acts in a PIF4-independent pathway for thermoresponsive flowering at high temperatures, similar to the acknowledged complexity of thermoresponsive flowering at lower temperatures (Strasser et al., 2009; Blázquez et al., 2003; Song et al., 2013).
ELF3 regulates thermoresponsive flowering under long days, and is not required for PIF4-dependent thermoresponsive adult morphologies
To directly address the relationship of ELF3 and PIF4 in adult thermoresponsive phenotypes, we grew pif4 and elf3 mutants with various thermal treatments. Previous experiments showing pif4 thermoresponsive adult phenotypes were grown in continuous light and transferred to 27°C at 14 days (Koini et al., 2009), whereas we transferred from 22°C to 27°C at day one. Hence, it was possible that the observed inconsistencies between elf3 and pif4 effects on adult thermoresponsive phenotypes were a trivial consequence of experimental conditions. For instance, exposure of early seedlings to elevated temperature may have unexpected consequences in later ontogeny. Consequently, we tested both transfer conditions under long days to ensure that our results were comparable to previous experiments (Figure 3). We found that the effect of different experimental treatments is negligible, though the longer 27°C treatment (transfer at day one) showed a slightly stronger morphological response than the shorter 27°C treatment (transfer at 14 days, Figure 3A, B). Further, our results under long days replicated previous observations under continuous light (Koini et al., 2009), showing that PIF4 is essential for petiole elongation (Figure 3B), but dispensable for thermoresponsive flowering (Figure 3C). Our PIF4 results were in direct contrast to ELF3, which was dispensable for petiole elongation (Figure 3B), but essential for thermoresponsive flowering (Figure 3C). Indeed, petiole elongation is potentially hyper-responsive in elf3 (Figure 3A, Figure 3B). These results confirm the apparent independence of ELF3 and PIF4 in these specific responses, and the independence adult thermal responses from early seedling temperature stimuli.
One open question was whether the dispensability of ELF3 for petiole elongation reflected increased importance of other inputs to PIF4, such as FCA, which is involved in PIF4-dependent thermoresponsive petiole elongation in seedlings (Lee et al., 2014). We therefore measured adult thermoresponsive petiole elongation in fca mutants (Figure 3D), and unexpectedly found no substantial difference between fca mutants and WT Col. We suggest that while PIF4 influences both adult petiole and hypocotyl elongation in response to elevated temperatures, regulatory rewiring across development occurs by removing FCA and ELF3 as inputs to PIF4-dependent thermomorphogenesis.
A second question was whether elf3 still affects thermoresponsive flowering in the Ws strain under other conditions; for instance, the closely related Ws-2 strain evinces a robust thermoresponsive flowering acceleration at 22°C relative to 16°C but not at 28°C relative to 22°C (Ibañez et al., 2015). We therefore assayed flowering in Ws and the Ws-derived null mutant elf3-4 at these temperatures (Figure 3E). Under these conditions, Ws robustly accelerated flowering at 22°C, whereas elf3-4 showed no perceptible difference in flowering between the two temperatures. Thus, ELF3’s role in thermoresponsive flowering is not restricted to the Col strain or a certain temperature, but rather is necessary for whatever thermoresponsive reaction norm a strain may have for flowering.
ELF3 and PIF4 regulate adult thermoresponsive phenotypes independently
If ELF3 and PIF4 were truly independent in controlling thermal responses of adult phenotypes under long days, then elf3 pif4 double mutants would show approximately additive phenotypes. We generated elf3 pif4 double mutants and subjected them to the same experiments as above (omitting again the 14 day transfer treatment). Our results indicated that flowering and petiole elongation appear to constitute independent temperature responses, with PIF4 controlling the former and ELF3 controlling the latter in additive fashions (Figure 4). That is, elf3 pif4 double mutants showed negligible thermoresponsive flowering like elf3, and a negligible petiole response like pif4. Additionally, elf3 pif4 flowered slightly later than elf3 at 22°, while maintaining a negligible thermal response in flowering, indicating that elf3 mutants are not simply restricted by a physiological limit of early flowering. The additivity of these phenotypes establishes that, under these conditions, ELF3 and PIF4 must operate in separate thermal response pathways.
Previous studies have indicated that other members of the PIF family, such as PIF1, PIF3, and PIF5, have minimal roles in these same thermal response phenotypes (Koini et al., 2009; Stavang et al., 2009; Proveniers and van Zanten, 2013), and pif4 pif5 double mutants retain substantial thermoresponsive flowering even under shorter 12 hour light: 12 hour dark photoperiods (Thines et al., 2014). These previous findings suggest that our results are not explained by redundancy between PIFs. However, to exclude this possibility, we evaluated thermoresponsive flowering in pif4 pif5 mutants (Figure 4D), because PIF5 is most often considered to act redundantly with PIF4 (Filo et al., 2015; Thines et al., 2014; Sun et al., 2013). As expected, both pif5 single mutants and pif4 pif5 double mutants have intact thermoresponsive flowering. These observations indicate that redundancy with other PIFs is not responsible for the apparent independence of PIF4 and ELF3.
Overall, the strong photoperiod-dependence of PIF4-related thermoresponsive flowering necessitates the existence of some pathway or pathways independent of PIF4 under long days, given the persistence of the phenomenon under these conditions. Based on our data, ELF3 acts in one such pathway.
Thermoresponsive flowering under long days can operate through the photoperiodic pathway
ELF3’s role in thermoresponsive flowering at low ambient temperatures has previously been shown to operate in the photoperiodic pathway, through repressing GI expression, which in turn represses GI’s direct activation of FT (Sawa and Kay, 2011; Jang et al., 2015). To evaluate whether this pathway might explain our results (Figure 5), we measured transcript levels of GI and CO in wild-type and elf3 mutants under 22°C and 27°C (Figure 5A). We found that GI is strongly up-regulated in elf3 null mutants of Col and Ws backgrounds, confirming previous reports in Col (Jang et al., 2015; Mizuno et al., 2014a). Further, wild-type Ws shows approximately five-fold higher basal GI levels compared to Col, which do not increase at higher temperatures. In contrast, Col shows very low basal GI levels that increase at higher temperatures to approximately the same levels as Ws. CO levels, however, are not substantially affected by either elf3 mutation or increased temperature, consistent with previous reports (Strasser et al., 2009; Jang et al., 2015), and in fact may decrease at 27°C. Thus, robust thermoresponsive flowering is correlated with low basal levels of GI, and with temperature-dependent GI up-regulation, as observed in Col. High basal GI levels in Ws may be associated with this strain’s lack of thermoresponsive flowering and deficient circadian temperature compensation at higher ambient temperatures (Ibañez et al., 2015; Edwards et al., 2005, 2015). These observations support the model under which ELF3 acts in the photoperiodic pathway to engender thermoresponsive flowering, just as it does in response to lower ambient temperatures (Strasser et al., 2009; Jang et al., 2015).
If the photoperiodic pathway contributes to thermoresponsive flowering at elevated ambient temperatures in long days (LD), we would expect mutants in this pathway to show abrogated temperature responses. Under short days (SD), members of both the autonomous and photoperiodic pathways are essential to the high temperature flowering response (Balasubramanian et al., 2006). These two pathways also contribute independently to thermoresponsive flowering at low temperatures (16°C vs. 23°C) (Blázquez et al., 2003; Strasser et al., 2009). Altogether, we would expect that a photoperiodic thermoresponsive flowering pathway would operate independently of both PIF4 and the autonomous pathways in long days. It is not clear whether the autonomous pathway would be independent of PIF4, given known interactions between FCA and PIF4 (Lee et al., 2014).
To evaluate whether these past results under other conditions also apply to LD and elevated temperatures, we measured flowering time at 22°C and 27°C in mutants in the photoperiodic pathway (gi, co, Figure 5B). We also tested mutants of the gibberellin pathway (spy), and a terminal floral integrator (soc1), which are not expected to be necessary for thermoresponsive flowering. We found robust thermal responses in all mutants except elf3 and gi, closely agreeing with previous results under other conditions (Balasubramanian et al., 2006; Strasser et al., 2009; Sawa and Kay, 2011; Jang et al., 2015), and implicating GI (but not CO) as an actor in thermoresponsive flowering at elevated temperatures. Collectively, these experiments suggest that the photoperiod pathway is necessary in promoting thermoresponsive flowering in long days, and expression data in this and other studies suggests that ELF3 is likely to act within this pathway.
Discussion
ELF3 and PIF4 are both crucial integrators of temperature and light signaling in controlling A. thaliana development. Recent literature has emphasized the centrality of PIF4-dependent thermoresponsive regulation in a variety of phenotypes, including in flowering (Kumar et al., 2012; Song et al., 2013; Wigge, 2013). Here, we extend the previous observation that PIF4 is dispensable for thermoresponsive flowering under long photoperiod conditions (Koini et al., 2009), and identify ELF3 as essential for thermoresponsive flowering under conditions in which PIF4 is dispensable. Our results integrate previous knowledge about thermoresponsive flowering under other conditions, and identify at least one pathway for this response that do not involve PIF4. Moreover, we show that while polyQ variation in ELF3 affects ELF3 function, the polyQ tract may not constitute a simple temperature-responsive component in itself (in Ws at least). Our results incorporate current findings into classic models of thermal responses in A. thaliana, allowing a comprehensive view of the genetic underpinnings of this agronomically crucial plant trait.
ELF3 polyglutamine variation affects thermoresponsive traits in part by modulating overall ELF3 activity
In previous work, we demonstrated that polyQ variation in ELF3 is (1) common, (2) affects many known ELF3-dependent phenotypes, and (3) is dependent on the genetic background (Undurraga et al., 2012). Following the recent discoveries that ELF3 is involved with temperature response (Mizuno et al., 2014b; Box et al., 2014; Raschke et al., 2015), we confirmed that ELF3 polyQ variation also affects temperature response phenotypes in a background-dependent fashion. However, we found little support for the hypothesis that the polyQ tract has a special role in temperature sensing. Instead, as was the case for other ELF3-dependent phenotypes, ELF3 polyQ variation appears to affect overall ELF3 functionality, with less functional ELF3 variants lacking robust temperature responses. The molecular mechanism by which polyQ variation affects ELF3 functionality remains unknown, in spite of substantial phenotypic characterization (Undurraga et al., 2012). Targeted mechanistic study will be required to answer this question.
ELF3-PIF4 relationship in thermoresponsive morphologies
One question that remains unanswered is to what extent ELF3 participates in PIF4-dependent thermoresponsive morphologies. While our study and previous work (Mizuno et al., 2014a; Box et al., 2014; Raschke et al., 2015) support a PIF4-ELF3 link in temperature-dependent hypocotyl elongation, this relationship disappears in the analogous case of temperature-dependent petiole elongation. These results can be explained by many hypotheses. First, it is possible that ELF3 regulation of PIF4 is only relevant at the early seedling stage. Alternatively, the model that elevated temperature down-regulates ELF3 could be inadequate, and rather PIF4 up-regulation at elevated temperatures could be due to other inputs. One candidate that modulates PIF4 responsiveness is FCA; however, thermoresponsive hypocotyl elongation is actually intensified in fca mutants (Lee et al., 2014), as is petiole elongation in seedlings. More work will be needed to understand the relationship of ELF3 and PIF4 in thermomorphogenesis.
Natural variation in temperature response
Several studies have indicated that different A. thaliana strains respond to temperatures differently, either shifting or inverting the reaction norm of the phenotype in question (Edwards et al., 2005; Ibañez et al., 2015; Edwards et al., 2015). Ws has a shifted reaction norm with respect to temperature compared to Col for photoperiod-related phenotypes, including flowering. For instance, Ws displays accelerated flowering at 23°C vs. 16°C (Ibañez et al., 2015), but accelerates flowering no further at 27°C. Thus, if thermoresponsive flowering involves ELF3 in Ws, elf3 Ws mutants would lack flowering acceleration at 22°C relative to 16°C, and indeed we show this to be the case. Another example of differential mutational effects among strains is that gi mutants in the Ler background display robust thermoresponsive flowering (Blázquez et al., 2003; Balasubramanian et al., 2006). It is unclear whether this finding is due to altered wiring of pathways between these backgrounds.
Thermoresponsive flowering can require PIF4 or ELF3, depending on photoperiod
Under various conditions, both ELF3 and PIF4 have been found to be crucial for thermoresponsive flowering; ELF3 is involved under long days comparing 16°C/23°C temperature conditions (Strasser et al., 2009; Jang et al., 2015), and PIF4 is involved under short days comparing 22°C/27°C temperature conditions (Kumar et al., 2012). Other members of the autonomous and the photoperiodic pathways have also been implicated in thermoresponsive flowering (Blázquez et al., 2003; Balasubramanian et al., 2006; Strasser et al., 2009) (besides other pathways, (Lee et al., 2007)). Consequently, some combination of these pathways, modulated by experimental conditions, must require ELF3 and/or PIF4. We and others (Koini et al., 2009; Thines et al., 2014) have observed that PIF4 and its paralogs are not required or minimally required for proper thermoresponsive flowering under longer photoperiods and elevated temperatures. Hence, some other pathway or pathways must be responsible for such a temperature response. Furthermore, we and others (Strasser et al., 2009; Jang et al., 2015) have shown that ELF3 and the photoperiod pathway (excluding CO) are essential for proper thermoresponsive flowering under long days. It has been previously shown that PIF4 and the photoperiodic pathway contribute to thermoresponsive flowering via independent pathways (Kumar et al., 2012), suggesting that under longer photoperiods PIF4 activity is inhibited, allowing other mechanisms to dominate thermoresponsive flowering.
These collected observations lead us to propose a model of thermoresponsive flowering, in which PIF4, ELF3, the photoperiodic pathway, and other pathways interact depending upon condition and genetic background (Figure 5D). Under short days or other short photoperiods, phyB activity is down-regulated, leading to up-regulation of PIF4 (Duek and Fankhauser, 2005; Huq and Quail, 2002; Luo et al., 2014; Lorrain et al., 2008), which at high levels occupies the promoter of the flowering integrator FT and induces flowering (Kumar et al., 2012). Constitutive overexpression of PIF4, PIF5, and PIF3 under long day conditions induces early flowering (Galvão et al., 2015), supporting the hypothesis that differences in PIF levels underlie the difference in PIF4 importance between long and short photoperiods. PIF4 activation thus appears to be the dominant mechanism of thermoresponsive flowering under short photoperiods, although pif4 null mutants still show some thermal response. However, under longer photoperiods, phyB up-regulation leads to an attenuation of PIF4 activity, and consequently the role of PIF4 and other PIFs becomes negligible (Koini et al., 2009). This allows canonical ambient temperature responses (such as the photoperiodic pathway, including ELF3, (Strasser et al., 2009; Jang et al., 2015)) to take a dominant role in thermoresponsive flowering. Several reports have indicated that GI and COP1, but not CO, are involved in thermoresponsive flowering (Balasubramanian et al., 2006; Strasser et al., 2009; Jang et al., 2015), with GI directly binding the FT promoter (Jang et al., 2015). Under each of these conditions, FT-induced flowering is activated by a different signaling cascade initiated by elevated temperature. This interpretation leads to a holistic and coherent view of how light and temperature responses are integrated in this important plant trait.
To summarize, at least three independent mechanisms have been described that promote thermoresponsive flowering in any context. These include the photoperiodic pathway (PHYB/ELF3/GI/COP1), the autonomous pathway (PHYA/FCA/FVE/TFL1/FLC), and the PIF4-dependent pathway (PIF4/H2A.Z/gibberellin), all of which converge by regulating FT (although the last pathway may also act through other integrators (Thines et al., 2014; Galvão et al., 2015). The collective results of our experiments and previous work suggest that the first two pathways are necessary but not sufficient for thermoresponsive flowering, and that the third (PIF4) is sufficient but not necessary for thermoresponsive flowering. The truth of this characterization deserves further study in understanding the interdependencies of the three pathways. For instance, it has been suggested that PIF4 binding to the FT promoter is dependent on cooperativity with a second photoperiod-controlled actor (Seaton et al., 2015). Notably, in our interpretation the photoperiodic pathway omits CO, which shows neither a thermoresponsive flowering defect nor temperature-dependent expression responses in any study we are familiar with (Balasubramanian et al., 2006; Sawa and Kay, 2011; Jang et al., 2015).
In conclusion, we observe that ELF3 is involved in the hypocotyl response to elevated temperature as reported previously, and that this response can be abrogated by poorly-functioning ELF3-polyQ variants. We further demonstrate that ELF3 has little effect on the petiole temperature response, and is necessary for the flowering temperature response, suggesting that it functions independently of PIF4, potentially in the photoperiodic pathway. These results reiterate the complexity of these crucial environmental responses in plants, and will serve as a basis for further development of our understanding of how plants respond to elevated temperatures. In the context of climatic changes, this understanding will serve those attempting to secure the global food supply.
Methods
Plant materials and growth conditions
All mutant lines (except pif4-2 elf3-200) were either described previously or obtained as T-DNA insertions from the Arabidopsis Biological Resources Center at Ohio State University (Alonso et al., 2003; Kleinboelting et al., 2012), and are described in Table S11. pif4-2 elf3-200 was obtained via crossing and genotyping. T-DNA insertions were confirmed with primers described in Table S10. For hypocotyl assays, seedlings were grown for 15d in incubators set to SD on vertical plates as described previously (Undurraga et al., 2012). All plates were incubated at 22° for one day, after which one replicate arm was transferred to an incubator set to 27°, with another replicate arm maintained at 22°. For flowering time assays, plants were stratified 3-5d at 4° in 0.1% agarose and seeded into Sunshine #4 soil in 36-pot or 72-pot flats to germinate at 22° under LD. Replicate arms were subsequently transferred to 27° LD conditions as indicated, with others remaining at 22°. Different temperature treatments of the same experiment were identical with respect to randomization, setup, and format. At 25d, petiole length and whole leaf length (including petiole) of the third leaf were measured, and the ratio of these values was further analyzed. Flowering was defined as an inflorescence ≥1cm tall; at this point, date and rosette leaf number were recorded.
Trait data analysis
All data analysis was performed using R v3.2.1 (R Core Team, 2015). Where indicated, temperature responses were modeled using multiple regression in the form Phenotype ∼ μ + βGGenotype + βTTemperature + βGxT(Genotype x Temperature) + βEExperiment + Error. All experiments were included in models for transgenic experiments, and thus the βE term describes systematic variation between experiments, whereas line-specific effects among transgenics should be modeled in the error term. Where temperature responses are reported, they consist of the βT + βGxT terms and associated errors where σr is the standard error for βT and σGxT is the standard error for βGxT), and thus are corrected for systematic experimental variation and temperature-independent genotype effects. Analysis scripts and data are provided at https://figshare.com/s/129525f02ef6e66f7bed and as File S1.
Gene expression analyses
Seedlings were grown for 1d under LD at 22°, after which one replicate arm was transferred to LD at 27°, with another replicate arm remaining at 22°, and all seedlings were harvested 6d later at indicated times. At harvest, ∼30mg aerial tissue of pooled seedlings was frozen immediately in liquid nitrogen and stored at -80°. RNA extraction, cDNA synthesis, and real-time quantitative PCR were performed as described previously (Undurraga et al., 2012), using primers in Table S10. Transcript levels were quantified using the ΔΔCt method (Pfaffl, 2001).
Author contributions. Designed research: MOP AL CQ. Performed experiments: MOP AL. Analyzed data: MOP. Wrote the paper: MOP CQ.
Supplemental Data
Table S1. Regression analysis of hypocotyl elongation temperature response among Col and Ws transgenic lines.
Table S2. Regression analysis of petiole: leaf length ratio temperature response among Col and Ws transgenic lines.
Table S3. Regression analysis of rosette leaf number at flowering temperature response among Col and Ws transgenic lines.
Table S4. Regression analysis of rosette leaf number at flowering and petiole: leaf length ratio temperature responses in elf3 and pif4.
Table S5. Regression analysis of rosette leaf number at flowering temperature response in Ws and elf3-4.
Table S6. Regression analysis of petiole: leaf length ratio temperature response in Col and fca mutants.
Table S7. Regression analysis of rosette leaf number at flowering temperature response in elf3 pif4 double mutants.
Table S8. Regression analysis of rosette leaf number at flowering temperature response in pif4 pif5 double mutants.
Table S9. Regression analysis of rosette leaf number at flowering and petiole: leaf length ratio temperature responses in flowering pathway mutants. Table S10. Primers used in this study. Table S11. Mutant lines used in this study. File S1. Data and analysis code.
Acknowledgments
We thank Philip Wigge and Jaehoon Jung for ideas, helpful conversations, sharing unpublished data, and comments on this manuscript. We thank Evan Eichler for use of the LightCycler instrument. We thank members of the Queitsch lab for helpful discussions. This work was supported by National Institutes of Health New Innovator Award DP2OD008371 to CQ.