Increases in Absolute Temperature Stimulate Free Calcium Concentration Elevations in the Chloroplast

Abstract

Plants need to sense increases in temperature to be able to adapt their physiology and development to survive; however, the mechanisms of heat perception are currently relatively poorly understood. Here we demonstrate that in response to elevated temperature, the free calcium concentration of the stroma of chloroplasts increases. This response is specific to the chloroplast, as no corresponding increase in calcium is seen in the cytosol. The chloroplast calcium response is dose dependent above a threshold. The magnitude of this calcium response is dependent upon absolute temperature, not the rate of heating. This response is dynamic: repeated stimulation leads to rapid attenuation of the response, which can be overcome by sensitization at a higher temperature. More long-term acclimation to different temperatures resets the basal sensitivity of the system, such that plants acclimated to lower temperatures are more sensitive than those acclimated to higher temperatures. The heat-induced chloroplast calcium response was partially dependent upon the calcium-sensing receptor CAS which has been shown previously to regulate other chloroplast calcium signaling responses. Taken together, our data demonstrate the ability of chloroplasts to sense absolute high temperature and produce commensurately quantitative stromal calcium response, the magnitude of which is a function of both current temperature and stress history.

Introduction

Temperature is one of the key environmental parameters affecting all living organisms. Fluctuations in temperature occur seasonally, daily as well as more rapidly and unexpectedly, such as when clouds shield the sun’s heat. Plants have evolved to be able to sense these events, anticipate them when possible and adjust their physiology accordingly (Ruelland and Zachowski 2010, Saidi et al. 2011, Knight and Knight 2012, Mittler et al. 2012). The ability to discriminate a cooling from a heating event, as well as the magnitude of it (e.g. chilling and freezing), is essential for survival (Penfield 2008, Hua 2009, Thomashow 2010, Knight and Knight 2012). Whilst cellular events downstream of temperature changes are well described, the mechanisms for temperature sensing, specifically the early events, are still an open research topic. Indeed, in plants, the specific thermometers for heat and cold have not yet been identified. Amongst the putative temperature-sensing mechanisms, several classes of biological processes have been shortlisted as possible primary sensors. These candidates are not only able to respond to temperature changes directly, but they also activate downstream response pathways (Ruelland and Zachowski 2010). These processes include protein unfolding, changes in the catalytic activity of enzymes, cytoskeleton disassembly, changes in membrane fluidity and chromatin remodeling (explained in detail in several reviews, e.g. Ruelland and Zachowski 2010, Saidi et al. 2011, Knight and Knight 2012, Mittler et al. 2012). Membrane rigidification/fluidization occur nearly concomitantly with the temperature variation, hence these events are likely to be upstream of the others. In Synechocystis, cold sensing is dependent on a histidine kinase (Hik33) whose activation relies on the cold-induced physical rigidification of the membrane (Mikami et al. 2002), while altering membrane fluidity by chemical means to mimic heat-caused de novo synthesis of heat shock proteins (Horvath et al. 1998). Furthermore, in plants, opposite changes in membrane fluidity are responsible for the activation of HAMPK (heat) and SAMK (cold) mitogen-activated protein (MAP) kinases (Sangwan et al. 2002), and Orvar et al. (2000) showed that changes in membrane rigidification act upstream of cytoskeleton remodeling in response to cold. Long-term membrane fluidity modification, where the membrane composition is altered, is also used by plants to acclimate to different temperatures (Murata and Los 1997, Falcone et al. 2004).

A widely studied plant second messenger is calcium (Ca²⁺), which is involved in nearly every aspect of cell physiology and development (Kudla et al. 2010, Batistic and Kudla 2012, Kudla et al. 2018). Alterations in the cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt) have been reported in response to a variety of environmental stimuli (e.g. cold, pathogens, etc.) and they are considered crucial early events in stress response pathways (Sanders et al. 1999, Batistic and Kudla 2012, Kudla et al. 2018). Specific information regarding the nature and the magnitude of the stress is acquired by using different spatio-temporal Ca²⁺ elevations, called ‘Ca²⁺ signatures’ (McAinsh and Hetherington 1998), which differ in parameters such as amplitude, duration, frequency and sublocation of the Ca²⁺ increase (Allen et al. 2001, Miwa et al. 2006, Whalley and Knight 2013).

Relationships between temperature sensing, specifically membrane fluidity, and Ca²⁺ signaling have already been reported. The cold response in plants is strongly dependent on a fast and transient cytosolic Ca²⁺ increase (Knight et al. 1991, Knight et al. 1996), and membrane fluidity changes affect the magnitude of these Ca²⁺ elevations (Orvar et al. 2000). In Physcomitrella patens, heat is responsible for an increase in cytosolic Ca²⁺ levels, leading to activation of the heat shock response (HSR; Saidi et al. 2009), and the extent of the Ca²⁺ heat response (and, consequently, of the HSR) is strongly dependent on membrane fluidity (Saidi et al. 2009, Saidi et al. 2010, Finka and Goloubinoff 2014).

Recently, attention has been focused on understanding Ca²⁺ signaling in the chloroplast. This organelle not only functions as a Ca²⁺ store (Roh et al. 1998, Stael et al. 2012b, Nomura and Shiina 2014, Costa et al. 2018), but also has the ability to generate its own specific Ca²⁺ signals in response to stresses, and hence to contribute to downstream signaling responses (Johnson et al. 1995, Sai and Johnson 2002, Manzoor et al. 2012, Nomura et al. 2012, Kmiecik et al. 2016, Loro et al. 2016, Sello et al. 2016, Sello et al. 2018). Ca²⁺ plays both a regulatory and a structural role in the chloroplast (Stael et al. 2012b, Sello et al. 2016). Importantly, Ca²⁺ is required for PSII assembly, photoprotection and recovery after photoinhibition (Mattoo et al. 1989, Miller and Brudvig 1989, Grove and Brudvig 1998, Yang et al. 2015), but high Ca²⁺ levels are able to inhibit photosynthesis, by acting on the Calvin–Benson cycle (Charles and Halliwell 1980, Kreimer et al. 1988). Furthermore, PSII has been recently identified as the site of action of the small chloroplast-localized heat shock protein 21 (Chen et al. 2017). In 1995 (Johnson et al. 1995), a chloroplast-specific Ca²⁺ increase was measured in response to the light to dark transition, initiating the field of chloroplast Ca²⁺ signaling. More than 20 years later, putative Ca²⁺ channels and transporters have been identified both in the inner envelope and on the thylakoid membranes (Stael et al. 2012b, Nomura and Shiina 2014). Chloroplast Ca²⁺ increases have been reported in response to cold, salt and hyperosmotic stresses (Nomura et al. 2012, Sello et al. 2016) as well as pathogen elicitor molecules (Manzoor et al. 2012, Nomura et al. 2012, Sello et al. 2016), with different kinetics compared with the cytosolic Ca²⁺ counterparts. In the case of response to elicitors and the light–dark transition, the chloroplast-localized Ca²⁺-sensing receptor CAS (Han et al. 2003, Nomura et al. 2008, Vainonen et al. 2008, Stael et al. 2012a, Wang et al. 2012) has been shown to be necessary for the full chloroplast Ca²⁺ response (Nomura et al. 2012). In the case of response to elicitors, the attenuation of chloroplast Ca²⁺ response led to reduced pathogen-related gene expression (Nomura et al. 2012). All the primary stimuli tested to date, apart from the light–dark transition, increase both cytosolic and chloroplast Ca²⁺. To date, no other chloroplast-specific (i.e. not cytosolic Ca²⁺-inducing) stimuli have been identified. Here we report a second instance of a chloroplast-specific Ca²⁺ increase, which occurs in response to heat. We describe its characteristics and discuss the significance of this ability of chloroplasts to sense increases in temperature.

Results

Heat increases free Ca²⁺ concentration in the chloroplast, but not in the cytosol

To examine the role of Ca²⁺ in chloroplast signaling, we treated Arabidopsis thaliana seedlings expressing aequorin targeted to the cytosol (pMAQ2) or stromal compartment (pMAQ6) with a range of stimuli known to induce abiotic stress responses in plants. A chloroplast-specific Ca²⁺ increase was observed in response to heat (Fig. 1). Arabidopsis seedlings were heated on a Peltier element positioned under a photon counting camera, and Ca²⁺-dependent luminescence was collected before and during the heating event. As shown in Fig. 1, Arabidopsis seedlings were kept at 20°C for 2 min and then heated at 40°C for 7 min before dropping the temperature back to 20°C. The 40°C pulse caused a transient increase in the stromal Ca²⁺ levels, up to concentrations of around 0.4–0.5 μM. In contrast, the cytosol did not display any Ca²⁺ increase during the same heat stimulus. However, as can be seen in Fig. 1, the temperature drop from 40 to 20°C was sensed by the plants as a cold shock, which is known to cause a rapid Ca²⁺ peak in the cytosol (Knight et al. 1996, Larkindale and Knight 2002). This cold response was also detected in the chloroplast, leading to the modest increase in stromal Ca²⁺ previously reported (Nomura et al. 2012).

Fig. 1

Plants respond to heat with a chloroplast-specific Ca²⁺ increase. Ca²⁺ elevations in response to heating (20–40°C) and cooling (40–20°C) events in the cytosol (cyt) and chloroplast (chl) are represented through time. Each trace was obtained by averaging the signal recorded from n = 6 for cyt and n = 5 for chl 8-day-old Arabidopsis seedlings. Error bars represent the SD. To mark where the chloroplast Ca²⁺ concentration is significantly different from that of the cytosol, the P-value (gray line) was calculated through time with an unpaired t-test.

In order to test the dose dependency of the Ca²⁺ heat response, a series of temperatures was applied to Arabidopsis seedlings, ranging from mild heat (30°C) to just sublethal temperatures (45°C), with an interval of 2.5°C. Each temperature above 30°C caused a stromal Ca²⁺ increase (Fig. 2A), and the kinetics of the Ca²⁺ were dependent on the temperature sensed, in a dose-dependent manner. For instance, the peak height increased linearly with increasing temperature (Fig. 2B). Conversely, peak time decreased with increasing temperature following a logarithmic relationship (Fig. 2C). Statistical significance of each temperature compared with another is represented in Fig. 2D which shows peak height, and Fig. 2E which shows the time at which the peak occurs. Interestingly, giving plants a 30°C heat stimulus did not cause a stromal Ca²⁺ increase, defining this temperature as the threshold for the chloroplast Ca²⁺ heat response under these conditions. Cytosolic Ca²⁺ increases were monitored for each of the temperatures tested in Fig. 2A, and results are reported in Supplementary Fig. S1, as showing little or no increase.

Fig. 2

The kinetics of the heat-induced chloroplast Ca²⁺ increase is temperature dependent. Chloroplast-targeted aequorin seedlings were exposed to a series of temperatures ranging from 30 to 45°C, at intervals of 2.5°C. (A) Kinetics of the Ca²⁺ increase upon heating. (B) Average relative chloroplastic Ca²⁺ concentration peak height and a linear regression line (R² = 0.9799) interpolating the peaks are represented. (C) Average chloroplastic Ca²⁺ concentration peak times and a logaritmic regression line (R² = 0.9416) interpolating the peaks are represented. (D) Statistical significance of the Ca²⁺ concentration peak height at different temperatures and (E) of peak times was calculated with one-way ANOVA followed by a Tukey’s multiple comparisons test, *P ≤ 0.1, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns = not significant. Data were obtained by averaging four 8-day-old Arabidopsis seedlings, and for each temperature a different set of plants was used. Error bars = SD.

We then tested whether the heat-induced chloroplast Ca²⁺ response was specific to Arabidopsis, or might be conserved amongst plant species. In order to test this, stromal and cytosolic aequorins were transiently expressed in Nicotiana benthamiana, and Ca²⁺ was measured 48 h after infiltration. Supplementary Fig. S2 shows that tobacco is also capable of responding to heat with a transient stromal Ca²⁺ increase; however, the magnitude of the response is lower for the equivalent temperature compared with Arabidopsis (compare Fig. 2A with Supplementary Fig. S2). The Ca²⁺ heat response is also conserved amongst the Arabidopsis ecotypes Col-0 and Ws-0, whose traces are almost identical (Supplementary Fig. S3).

Characteristics of the chloroplast heat response: attenuation and sensitization

Attenuation is a property observed when an organism is repeatedly exposed to a stimulus of the same magnitude within a relatively short time, with the size of the response decreasing each time as a consequence of the previous experience. This was found to be the case for the chloroplast heat response; when seedlings were consecutively exposed to 4 min 40°C heat pulses every 5 min, they showed a reduced Ca²⁺ response upon each subsequent stimulation (Fig. 3A). A stimulation at 45°C following three such 40°C heat pulses was able to re-establish the stromal Ca²⁺ increase, and the magnitude of this elevation was significantly greater than that recorded upon the first 40°C heating pulse (Fig. 3A, B). This property is known as sensitization, and it was able to overcome attenuation.

Fig. 3

The chloroplast Ca²⁺ heat response displays attenuation and sensitization. (A) [Ca²⁺]_chl response to three consecutive heat pulses of the same magnitude (40°C for 4 min) followed by a fourth pulse at a higher absolute temperature (45°C for 4 min). Each heat pulse was separated by a 5 min resting period at 20°C. (B) Relative [Ca²⁺]_chl peak heights of the four individual peaks. Data represent an average of n = 7 Arabidopsis seedlings. Error bars = SD, ****P ≤ 0.0001 calculated by one-way ANOVA followed by Dunnett’s multiple comparison test using the first peak as a control reference.

Heat sensing is mainly dependent upon absolute temperature

In order to investigate whether the rate at which the temperature increase is given is a key parameter of the chloroplast heat response, plants were subjected to an increase from 20 to 40°C at rates of either 0.4, 0.2, 0.15 or 0.1°C s^–1. In Fig. 4A, the chloroplast Ca²⁺ concentration at the peak was plotted against the rate of temperature increase, and all the data fit a horizontal line (R² = 0.0133), which indicated that there is no correlation between the rate of heating and the magnitude of the Ca²⁺ peak. These data indicated that the magnitude of response could be fundamentally dependent upon either absolute temperature or the absolute change in temperature (ΔT) that the plant experienced. To discriminate between these two cases, Arabidopsis plants were heated up to 40°C starting from different initial temperatures, namely 15, 20 and 25°C (Fig. 4B). In this case, the ΔT varies (25, 20 and 15°C, respectively), but the final absolute temperature (40°C) remains the same. Fig. 4B shows that the Ca²⁺ peak values were similar in response to the different ΔT (the horizontal regression line indicates no correlation, R² = 0.0921). On the other hand, when plants were exposed to different absolute temperatures, but the same ΔT of 20°C (heat regimes applied were from 15 to 35°C, from 20 to 40°C and from 25 to 45°C) was maintained, a different pattern emerged. In this case (Fig. 4C) Ca²⁺ peak values were significantly different from each other and linearly proportional to the magnitude of absolute temperature (Fig. 4C). These data clearly demonstrate, therefore, that absolute temperature is the primary parameter regulating the magnitude of the chloroplast heat response.

Fig. 4

The peak level of the heat-induced [Ca²⁺]_chl response is regulated by absolute temperature, not by the heating rate. Each data point represents an average of the value reached at the [Ca²⁺]_chl peak obtained from (A) n = 4 Arabidopsis seedlings exposed to a temperature shift from 20 to 40°C at different rates. For each rate, a different set of plants was used. Rates tested were 0.4, 0.2, 0.15 and 0.1°C s^–1; (B) n = 7 Arabidopsis seedlings exposed to a temperature shift from 15, 20 and 25°C to 40°C at the same rate and (C) n = 7 Arabidopsis seedlings exposed to the temperature shift of 20°C (from 15 to 35°C, from 20 to 40°C and from 25 to 45°C) at the same rate. Data points represent experimental data, interpolated by a regression line; error bars = SD. Statistical significance was calculated with one-way ANOVA followed by a Tukey’s multiple comparisons test, ns = not significant ****P ≤ 0.0001.

Acclimation to different temperatures regimes alters the heat-induced chloroplast Ca²⁺ response

To test the effect of growth history on the heat-induced chloroplast Ca²⁺ response, plants were treated overnight at either 15, 20 or 30°C. The Ca²⁺ response of these three sets of plants to the same heat stimulus (40°C for 7 min) was compared (Fig. 5A). Plants acclimated at different temperatures produced a larger (15°C pre-treatment) or smaller (30°C pre-treatment) stromal Ca²⁺ response to heat compared with the control (20°C pre-treatment). The concentration of Ca²⁺ at the peak is inversely proportional to the acclimation temperature, with 15°C pre-treatment showing the biggest Ca²⁺ response (Fig. 5B). As a control, plants were treated for 30 min at the same acclimation temperatures (15, 20 and 30°C) before the 40°C heat treatment. At this time scale (30 min), acclimation would not be expected to occur. Indeed, 30 min acclimation was not sufficient to affect the heat response, as can be seen by comparing Fig. 5C and D. Differences in baseline Ca²⁺ levels were not detected at the different acclimation temperatures, suggesting that the steady-state levels of stromal Ca²⁺ are kept at the same level in response to the different acclimation temperatures.

Fig. 5

Acclimation to high or low temperature affects the subsequent Ca²⁺ response to heat. Chloroplast-targeted Arabidopsis aequorin lines were acclimated overnight (A and B) or for 30 min (C and D) at 15, 20 or 30°C, then stimulated at 40°C for 7 min. (A) Chloroplast Ca²⁺ kinetics upon heating of the different overnight pre-acclimated lines and (B) average chloroplastic Ca²⁺ concentration peak heights. (C) Chloroplast Ca²⁺ kinetics upon heating of the lines pre-acclimated for 30 min and (D) respective average chloroplastic Ca²⁺ concentration peak heights. Data were obtained by averaging traces of n = 5 Arabidopsis seedlings for overnight acclimation and n = 8 seedlings for 30 min acclimation. Error bars = SD; P-values are represented (**P ≤ 0.01, ****P ≤ 0.0001, ns = not significant) and were calculated with one-way ANOVA followed by a Turkey multiple comparison test.

The heat-induced chloroplast Ca²⁺ response is partially dependent on CAS

The Ca²⁺-sensing receptor CAS has previously been identified as a thylakoid membrane-resident protein postulated to be a Ca²⁺ sensor (Han et al. 2003, Nomura et al. 2008, Vainonen et al. 2008). It has been shown previously to be necessary for full chloroplast Ca²⁺ responses to pathogen elicitors and the light to dark transition (Nomura et al. 2012). Therefore, the effect of heating on chloroplast Ca²⁺ was tested in two independent mutant alleles of the CAS protein (At5g23060) (Vainonen et al. 2008). As can be seen in Fig. 6, whilst the mutants were qualitatively responsive to the heat stimulus, the magnitude of response was significantly reduced to around 50% of the wild-type level.

Fig. 6

Chloroplast-specific Ca²⁺ increases are partially CAS dependent. (A) Representative Ca²⁺ traces of Arabidopsis wild-type Col-0, cas SALK and cas GABI lines in response to a 40°C heat pulse, and (B) average chloroplastic Ca²⁺ concentration peak heights. Data were obtained by averaging four 8-day-old Arabidopsis seedlings, and for each temperature a different set of plants was used. Error bars = SD. Asterisks represent statistical significance compared with the Col-0 control, **P ≤ 0.01, ***P ≤ 0.001 analyzed with one-way ANOVA followed by Dunnett’s multiple comparisons test.

Discussion

In this study, a chloroplast-specific Ca²⁺ signal was identified in response to heat. We demonstrated that this response occurs uniquely in the chloroplastic compartment and that it is dependent upon the magnitude of the temperature applied, not the rate.

Evidence of Ca²⁺ signaling in the chloroplasts has been previously reported in response to pathogens (Manzoor et al. 2012, Nomura et al. 2012, Sello et al. 2016) and abiotic stress (Nomura et al. 2012, Sello et al. 2016), and these stimuli are able to cause both a cytosolic and a stromal Ca²⁺ increase. However, the only other reported case of a chloroplast-specific Ca²⁺ increase was discovered by Johnson et al. (1995) in response to a light to dark transition (Sai and Johnson 2002).

Heat and Ca²⁺ have previously been linked in the literature. It has been shown that Ca²⁺ is able to confer protection against heat stress, specifically preventing oxidative damage, and that it is involved in the acquisition of long-term thermotolerance (Gong et al. 1997, Gong et al. 1998, Larkindale and Knight 2002). Moreover, in moss, specific Ca²⁺ cyclic nucleotide-gated channels (CNGCs) located in the plasma membrane have been shown to regulate the thermosensory response (Saidi et al. 2009, Finka and Goloubinoff 2014). Further evidence of a possible role for Ca²⁺ in heat response pathways comes from the study of unicellular prokaryotic cyanobacteria, where a Ca²⁺ increase analogous to the one presented in this study (Fig. 1) was reported in response to heat shock (Torrecilla et al. 2000). The presence of a similar mechanism in prokaryotes might suggest that such responses were developed before the endosymbiotic event leading to chloroplasts in eukaryotes, and then was conserved in the chloroplast throughout subsequent evolution.

The heat-induced Ca²⁺ response was consistent between different Arabidopsis ecotypes (Col-0 and Ws-0) and it could be observed in different plant species (tobacco and Arabidopsis), suggesting that there may be a common signaling mechanism in higher plants. However, differences in the magnitude of the Ca²⁺increase were observed in Arabidopsis itself and between different species. Each of the two Arabidopsis ecotypes tested, when stimulated at 40°C, responded with a Ca²⁺ increase whose magnitude ranged from 0.3 to 0.7 µM on different days, most probably depending on slight uncontrollable differences in the growth conditions. For this reason, only experiments conducted on the same day were directly compared with each other, and each of them was replicated at least twice to confirm the results. When different species were stimulated by heating, differences were observed in terms of sensitivity. Indeed, stromal Ca²⁺ concentrations were detected in Arabidopsis when stimulated at 40°C comparable with those in tobacco at 45°C (compare Fig. 1 and Supplementary Fig. S2), while at 40°C there is no distinguishable Ca²⁺ peak in tobacco (only a slight Ca²⁺ increase was observed). These differences can be attributed either to a genetic factor distinguishing the thermometer between the two species, or to the different growth temperature regimes applied before the heat treatment (consistent with data shown in Fig. 5). In both species, the relationship between higher temperatures causing a larger Ca²⁺ increase was observed. This relationship is clearly demonstrated in Fig. 2A and B, where the kinetics of the Ca²⁺ curves, as well as peak heights, change progressively with increasing temperature, following a dose–response relationship. Such differences in the Ca²⁺ kinetics are able to be detected by plant cells as unique ‘Ca²⁺ signatures’ (McAinsh and Hetherington 1998), which are crucial to encode different cellular messages. Therefore, the different Ca²⁺ signatures seen at different temperatures might be used by plants to discern one temperature from another, acting as a cellular ‘thermometer’.

One interesting property of the chloroplast Ca²⁺ heat response is that its amplitude attenuates when plants are exposed to consecutive heat stimulation of the same magnitude (Fig. 3). This characteristic is termed attenuation, and it has been previously demonstrated for the cytosolic Ca²⁺ cold response (Plieth et al. 1999). Attenuation is most probably attributable to the activity of channels, which are desensitized by the consecutive stimulations. The possibility that the reduction in the signal may be due to lack of Ca²⁺ available in the stores was excluded by the data shown in Fig. 3, where a higher absolute temperature stimulation restored the Ca²⁺ increase. This property (overcoming attenuation) is known as sensitization and has also been observed for the cold response (Plieth et al. 1999).

Another very important feature observed in the cold-induced cytosolic Ca²⁺ increase is its dependence upon the cooling rate (dT/dt; Plieth et al. 1999), rather than absolute temperature. Hence, we tested the effect of the rate upon the chloroplast heat response, and it emerged (Fig. 4A) that high temperature sensing in plants is mostly dependent upon absolute temperature, rather than the rate. Indeed, for the range of rates tested, the value obtained for the Ca²⁺ peak height was highly similar. This lack of correlation between rate and peak height is an indication that the absolute temperature reached at the end of the heating regime (40°C for all the samples) is the major parameter controlling the Ca²⁺ increase, in stark contrast to the cytosolic Ca²⁺ response to cold.

Fig. 4A suggested the importance of absolute temperature, but it did not formally distinguish whether the response to heat is mainly dependent on absolute temperature or relative temperature change (ΔT). To discriminate between these two options, two experiments were performed, one in which ΔT was varied, whilst absolute T was not (Fig. 4B), and, conversely, in the second, ΔT was fixed to 20°C, but the final absolute T reached was varied (Fig. 4C). While in the first experiment there was no observed correlation between the magnitude of [Ca²⁺]_chl at the peak at different ΔT applied (Fig. 4B), a strong linear dependency was observed for the second case, where the change was in absolute final temperature (Fig. 4C). Notably, results in Fig. 4C are comparable with those shown in Fig. 2B. This evidence strongly suggests that the absolute temperature reached at the end of the heating episode, and not the rate, or the relative temperature change, is the major parameter controlling the heat-induced chloroplast-specific Ca²⁺ increase. Additionally, the major difference in the behavior of the cold response (dependent on the rate of cooling) compared with the heat response reported here is indicative of the fact that two distinct thermometers must be present in plants for sensing increases and decreases in temperatures, respectively. It is interesting that in plants, the cold receptor leading to Ca²⁺ elevation appears to be in the plasma membrane (Plieth et al. 1999), whereas the heat receptor leading to Ca²⁺ elevation is in the chloroplast. In the case of mammals, both cold and heat receptors (themselves Ca²⁺ channels) are located in the plasma membrane (Caterina et al. 1997, McKemy et al. 2002).

When plants are exposed to any temperature changes compatible with plant survival, they are able to adjust the fluidity of their membranes to the new conditions through acclimation, which is a long-term process that involves modifications of the level of saturation of fatty acids (Wilson and Crawford 1974, Pearcy 1978, Graham and Patterson 1982, Murata and Los 1997). This is to maintain the functioning of membrane-resident processes in the face of long-term changes in temperature. As well as these biologically derived changes in membrane fluidity used by plants to perform long-term acclimation to the new temperature regime, rapid changes in membrane fluidity occur as a basic biophysical property of the membranes themselves when the temperature is suddenly modified (Dynlacht and Fox 1992, Mejia et al. 1995, Horvath et al. 1998, Saidi et al. 2009). It is thought that these rapid changes in membrane fluidity are used for temperature sensing by plant cells (Orvar et al. 2000, Sangwan et al. 2002). In this study, we show that acclimated plants respond differently to a heat stimulus according to the temperature they have experienced previously. Indeed plants pre-treated overnight at 15°C, whose membrane will be more fluid due to a higher level of desaturation of the fatty acids, were responding to the same heat stimulus by producing a bigger Ca²⁺ response compared with the control pre-treated at 20°C (Fig. 5A, B). Conversely saturating the membrane fatty acids by pre-acclimating plants at 30°C overnight caused a decreased stromal Ca²⁺ response to heat (Fig. 5A, B). Additionally, if the pre-treatments at 15, 20 and 30°C were reduced to 30 min only, these differences were abolished (Fig. 5C, D), confirming the results obtained in Fig. 4B. These results are consistent with the idea that acclimation leading to changes in membrane fluidity, which is a long-term process, may be responsible for the differences observed in Fig. 5A and B. Additionally, these data indicate that the cellular thermometer involved is able to reset according to the temperature plants have been experiencing before the experimental heating event. Therefore, it might be that rapid changes in membrane fluidity are the primary temperature-sensing event leading to elevations in chloroplast free Ca²⁺ concentration, a theory we will test in the future.

The Ca²⁺-sensing thylakoid protein CAS has been shown in previous studies to be necessary for the chloroplast Ca²⁺ responses to elicitors and the light–dark transition (Nomura et al. 2012). We show that cas mutants displayed a similarly significant reduction in response to heat (Fig. 6). In the case of elicitors, a reduced chloroplast Ca²⁺ response was correlated to reduced expression of salicylic acid-dependent pathogen gene expression and the production of salicylic acid itself (Nomura et al. 2012). This demonstrates that changes in the chloroplast free Ca²⁺ concentration can act as signals regulating downstream processes. Therefore, it is quite possible that the heat-induced chloroplast free Ca²⁺ increases we report here regulate an as yet unidentified downstream response to heating. It will be interesting to identify what these responses are in future work.

In conclusion, we discovered a chloroplast-specific absolute temperature-dependent Ca²⁺ response to heat. This suggest that a plant heat thermometer may be located in the chloroplast. This thermometer is dependent upon CAS protein function and stress history. Determining the nature of this thermometer would be an important target for future work.

Materials and Methods

Plant material and growth conditions

The majority of the experiments were conducted on A. thaliana lines constitutively expressing 35S::apoaequorin either in the cytosol [pMAQ2, Col-0 ecotype (Knight et al. 1991)] or in the chloroplast [pMAQ6, both Col-0 and Ws-0 ecotypes (Ws-0 was a kind gift from Dr. William F. Ettinger, Gonzaga University, Spokane, WA, USA)], and for plant transformation, wild-type Col-0 seeds were used. Two homozygous cas (At5g23060) mutants lines 665G12 (from the GABI-KAT collection) and SALK 070416 (from the Salk collection) were a kind gift from Professor Eva-Mari Aro (Turku University, Finland). Seeds were ethanol-sterilized, sown on 1× Murashige and Skoog (MS; Duchefa Biochemie) medium (Murashige and Skoog 1962) with 0.8% (w/v) agar (Sigma-Aldrich) on Petri dishes, vernalized for a minimum of 48 h at 4°C before growing them at 20°C with a 16/8 h photoperiod at a light intensity of 150 μmol m^–2 s^–1. Imaging experiments were performed on 8-day-old seedlings; aequorin reconstitution was performed on 7-day-old seedlings. For Agrobacterium tumefaciencs-mediated transformation, seedlings were transferred onto 44 mm peat plugs (Jiffy Products International) and grown at 20°C with a photoperiod of 12/12 h until bolting, and 16/8 h after Agrobacterium-mediated transformation (light intensity 150 μmol m^–2 s^–1); for seed collection, individual seedlings were sown on 41 mm peat plugs (Jiffy Products International) and grown at 20°C in a 16/8 h photoperiod (light intensity 150 μmol m^–2 s^–1). Nicotiana benthamiana plants were grown on soil at 27°C for 4 weeks with a 16/8 h photoperiod at 250–300 μmol m^–2 s^–1.

Plant transformation

Plant genetic transformation was performed on A. thaliana Col-0 wild-type plants with the binary construct pMAQ6 (Johnson et al. 1995) using the floral dip method (Clough and Bent 1998). Selection of the primary transformants was performed on MS medium containing kanamycin (50 mg l^–1), and successfully transformed plants were grown to seed as described above. Similarly, the cas mutants were transformed using the pMAQ6 construct in the binary vector pB7WG2 (Karimi et al. 2002) and selected on BASTA. Aequorin-based selection was performed by using a photon-counting camera (for details, see ‘Temperature and chemical treatments of plants’ and ‘In vivo reconstitution of aequorin and Ca²⁺-dependent luminescence measurements’ below), and the total amount of aequorin was measured by changing the temperature to –15°C for 5 min, followed by 2 min at 20°C. Lines with levels of aequorin closest to the average expression were chosen for further experiments.

Temperature and chemical treatments of plants

Fast changes in temperature were performed on a Pelter cooling element (Photek 5.0 and TCS1.0; Photek). Arabidopsis seedlings were laid down on the cooling element on wet filter paper, and covered with cling film, while tobacco detached leaves were flattened with a thin transparent glass plate. Acclimation temperature treatments were performed as follows: 48 h before performing the measurements, plants were reconstituted with coelenterazine at 20°C overnight in the dark. The following day, plants were left at 20°C in the light for 8 h, then transferred for 12 h in darkness at either 15, 20 or 30°C until Ca²⁺ was measured. When the pre-acclimation treatment was reduced to 30 min, plants were reconstituted with colelenterazine overnight at 20°C in darkness, and the next day treated for 30 min at either 15, 20 or 30°C for 30 min in darkness and used for Ca²⁺ measurements.

In vivo reconstitution of aequorin and Ca²⁺-dependent luminescence measurements

Aequorin reconstitution was performed by floating Arabidopsis seedlings on water containing 10 μM coelenterazine in 1% (v/v) methanol (Biosynth). Reconstitution of tobacco plants was performed by infiltrating with a syringe the aequorin-expressing area with a 50 μM coelenterazine solution, in 1% (v/v) methanol 24 h after infiltration. All plants were left in the dark from 12–24 h at 20°C before Ca²⁺ measurements. For Ca²⁺ imaging during temperature treatments, aequorin luminescence was recorded under a plate-intensified charge-coupled camera (Photek 216; Photek). Total aequorin for calibration was measured by decreasing the temperature to –15°C for 5 min, and then back to room temperature to discharge the remaining aequorin. This freezing treatment ruptures all cellular membranes including the chloroplast and allows excess Ca²⁺ from the cell to saturate the aequorin and fully discharge it. Subsequent to this treatment, there is no remaining reconstituted stromal aequorin as previously described (Mehlmer et al. 2012). Calibration was performed as previously described (Knight et al. 1996). Statistical analysis of data involved an unpaired t-test for the comparison of two conditions and one-way analysis of variance (ANOVA) for the comparison of three or more conditions. Subsequently, ANOVAs were followed by post-hoc tests, either by a Tukey’s multiple comparisons test for comparison of each mean with every other mean or by Dunnett’s multiple comparison test for comparing every mean with a control mean. All statistical tests were performed with GraphPad Prism (GraphPad Software, Inc.).

Funding

This study was supported by the EU Marie Curie project CALIPSO [GA 2013-607607 to G.L.].

Acknowledgments

The authors acknowledge Dr. Willian F. Ettinger (Gonzaga University, Spokane, WA, USA) for the pMAQ6 Ws-0 seeds, Eva-Mari Aro (Department of Biology, Plant Physiology and Molecular Biology, University of Turku, Finland) for the cas mutant seeds, and Heather Knight (Durham University, UK), Véronique Larosa (University of Liège, Belgium) and Ute Vothknecht (University of Bonn, Germany) for critical suggestions. We thank Beccy Manning for excellent technical assistance and support.

Disclosures

The authors have no conflicts of interest to declare.