Light, power, action! Interaction of respiratory energy and blue light induced stomatal movements

1. Although the signalling pathway of blue light dependent stomatal opening is well characterized, little is known about the interspecific diversity, the role it plays in the regulation of gas exchange and the source of energy used to drive the commonly observed increase in pore aperture. 2. Using a combination of red and blue light under ambient and low [O 2 ] (to inhibit respiration), the interaction between blue light, photosynthesis and respiration in determining stomatal conductance was investigated. These findings were used to develop a novel model to predict the feedback between photosynthesis and stomatal conductance under these conditions. 3. Here we demonstrate that blue light induced stomatal responses are far from universal, and that significant species-specific differences exist both in term of rapidity and magnitude. Increased stomatal conductance under blue light reduced photosynthetic limitation, at the expense of water loss . Moreover, we stress the importance of the synergistic effect of blue light and respiration in driving rapid stomatal movements especially when photosynthesis is limited. 4. These observations will help re-shape our understanding of diurnal gas exchange in order to exploit the dynamic coordination between A and g s , as a target for enhancing crop performance and water use efficiency. theoretical framework for the interplay between photosynthesis and stomatal behaviour following a step increase in red or blue light and highlights the importance of mitochondrial respiration in initiating and driving rapid stomatal movements. These findings provide a unifying theory for the observed “RL” (photosynthesis dependent) and “BL” (photosynthesis independent) induced stomatal opening that has been discussed extensively in the literature. presented here demonstrate that the BL response is far from universal, and that there is a significant inter-specific diversity in stomatal response (both in rapidity and magnitude). Our results give a broader context to the importance of BL for gas exchange and show how BL induced stomatal responses interact with photosynthetic and respiratory processes over the course of the day. At the beginning of the day, BL was shown to uncouple g sw variation from A and enables rapid stomatal opening, removing diffusional limitations and facilitating lower g sw during the night. Respiratory and photosynthetic processes were both required for rapid stomatal movements and to maintain a high g sw . We predict here that the energy produced by respiratory processes could drive up to 60% of the observed diurnal variations in g sw , with photosynthesis providing we have shown that for stomatal closing) maintenance of the steady state g sw in the light. in the absence of energy (both photosynthetic and respiratory), g sw remained unchanged (for in wheat suggesting there is minimal energy requirement to maintain g sw at the current level. These results emphasize the unexploited potential of temporal optimisation of A and g sw over the diurnal period, based on tuning the sensitivity of the stomatal response to BL and the ratio of photosynthetic and driving the the H + -ATPase inhibiting S-type anion resulting in a stronger membrane hyperpolarisation, and the activation of ion channels and K + uptake 2007; et 2020). Overall, BL stimulates energy release and exchange between the mesophyll and guard cells independently of photosynthesis. (7) The rapid removal of photosynthetic limitation and increased stimulation of energetic mechanisms involved in changes in stomatal aperture, induce rapid stomatal kinetics under blue light and enable larger apertures compared to those observed under RL alone. Width proportionality


Introduction
Photosynthesis, the primary determinant of plant biomass depends on light intensity (McCree, 1971) and CO 2 availability at the sites of carboxylation (Farquhar et al., 1980). Gas exchange (CO 2 and H 2 O) in and out of the leaf is controlled by stomata, microscopic pores surrounded by a pair of guard cells that open and close in response to environmental cues and internal signals. In addition to light which is one of the main environmental cues driving variation of gas exchange, endogenous signals such as hormones and the circadian clock can further influence the diurnal behaviour of stomata (Gorton et al, 1993). Stomatal conductance (g s ), a measure of the ease at which gas diffuses through stomata over the leaf surface is closely correlated with the rate of carbon assimilation (A) under steady state conditions, although the mechanisms co-ordinating the two are not clear (Wong et al., 1979;Ball et al., 1987). Variations in g s balance CO 2 uptake and evaporative demands, which happen in opposite directions, resulting in a tradeoff between biomass production and water loss at the plant level (Condon et al., 2002;McAusland et al., 2016). A temporal decoupling of A and g s can appear under dynamic environmental conditions due to the stomatal response being an order of magnitude slower than A responses (McAusland et al., 2016;Taylor & Long, 2017;Adachi et al., 2019). Understanding and optimizing the mechanisms controlling the dynamic coordination between A and g s is an unexploited avenue to increase plant productivity and contribute to achieving food security (Lawson et al., 2010(Lawson et al., , 2012Wu et al., 2019;Leakey et al., 2019).
Photosynthesis and stomatal movements are thought to be part of a positive feedback loop where the products of photosynthesis are used to power changes in guard cell shape, which in turn alters pore dimensions and regulates CO 2 diffusion and A (Farquhar et al., 1978;Buckley et al., 2003).Alternatively, it has been proposed that these products or their intermediates could act as signalling molecules coordinating A and g s (Lee & Bowling, 1993;Mott et al., 2008;Fujita et al., 2013;Mott & Peak, 2018), although the exact signal has yet to be identified. In a situation where stomata are closed (e.g. start of a dark to light transition) such a regulation loop could be expected to result in an initial slow increase in g s due to the diffusive limitation of A, followed by an exponential phase triggered by the rise in A. However, several studies have reported rapid stomatal responses even in dark acclimated plants (Yamori et al., 2020;Flütsch et al., 2020), which suggests that stored energy (in the guard cells or adjacent mesophyll) is used during the initial stomatal opening response (Outlaw & Manchester, 1979;Schnabl, 1980). The energy required for stomatal movements can originate from photosynthesis either from the guard cell chloroplasts or be imported from the surrounding mesophyll cells, although the exact contributions of each are not known and debated (Lawson et al., 2002;Lawson, 2003). The large ratio of mitochondria to

Accepted Article
This article is protected by copyright. All rights reserved chloroplast in guard cells suggests a greater contribution from respiratory processes than photosynthesis (Shimazaki et al., 2007) in maintaining energy supply for stomatal movements. In the guard cell, energy in the form of adenosine triphosphate (ATP) is produced by electron transport within chloroplasts and has been reported to be 80% of that observed in the mesophyll (Lawson et al., 2002;2003). Functional chloroplasts are essential for guard cell energetics and turgor (Azoulay-Shemer et al., 2016) with electron transport potentially providing energy to drive ion exchange and/or used to produce or transform organic compounds (e.g. sugar). These compounds can be subsequently utilized either as osmotica to drive changes in turgor (Horrer et al., 2016) or as substrates to release energy by mitochondrial respiration (Medeiros et al., 2018). It is important to note that previous work quantifying guard cell photosynthesis has reported limited Calvin cycle activity and suggested that sugars present in guard cells are mostly imported from the surrounding mesophyll (Outlaw, 2003).
Stomatal responses to irradiance depend not only on the intensity but also the wavelength, that triggers two distinct light transduction pathways: the red light (RL) and blue light (BL) responses (Shimazaki et al., 2007;Matthews et al., 2020). The RL induced stomatal response is generally described as dependent on photosynthesis and is used to explain the close relationship between A and g s . The BL induced stomatal response occurs at low light intensities and is often considered independent of photosynthesis (as is that the low light levels are not enough to drive photosynthesis). Several studies have suggested that the intensity of the background RL influences the magnitude of the stomatal response to BL (Ogawa, 1981;Assmann, 1988;Shimazaki et al., 2007). BL has been reported to be more effective at opening stomata than RL which involves the release of stored energy and osmotica from starch degradation or lipid metabolism (Horrer et al., 2016;McLachlan et al., 2016). BL is sensed in guard cells by phototropins (Kinoshita et al., 2001) that stimulates stomatal opening by an activation cascade of serine/threonine kinases such as BLUS1 and BHP leading ultimately to the activation of plasma membrane H + -ATPases (Takemiya et al., 2013;Takemiya & Shimazaki, 2016;Hayashi et al., 2017). By activating the H + -ATPase proton pumps on the plasmalemma, whilst simultaneously inhibiting S-type anion channels, BL stimulates membrane hyperpolarisation, the activation of ion channels and K + uptake (Marten et al., 2007;Inoue et al., 2020). However, it is still not clear how blue light-activated phototropins transmit the signal to the H + -ATPases and inhibit plasma membrane anion channels (Marten et al., 2007;Hiyama et al., 2017;Hosotani et al., 2021). The fact that BL induced stomatal opening does not necessarily rely on photosynthesis (Karlsson et al., 1986;Roelfsema et al., 2006) suggests that the guard cell mitochondria play a key role in powering the BL response. Little is known on the interaction between respiratory processes and the BL Accepted Article induced stomatal response, and the impact on the A and g s relationship during a diurnal period. BL could therefore play an important role in the regulatory feedback loop described above between A and g s .
Rapidity and magnitude of changes in g s in response to changing irradiance influence crop photosynthesis (McAusland et al., 2016;Taylor & Long, 2017) and yield under natural environments (Adachi et al., 2019;Yamori et al., 2020). Despite having been identified in several species, the diversity of the blue light dependent stomatal response and its role during a diurnal period are still not well characterized. Recent studies have suggested that stomatal blue-light response is present in seed plants, ferns from early diverged clades, and lycophytes (Doi et al., 2015;Sussmilch et al., 2019) and may have provided a competitive advantage (Doi et al., 2015;Westbrook & McAdam, 2020) for example helping the diversification of modern ferns during the Cretaceous (Cai et al., 2021). The nature of this advantage is unclear and we suggest that it could provide an advantage under dynamic light conditions favouring increased carbon fixation during the diurnal period. Comparing temporal kinetics of g s in response to changes in red light intensity with or without the addition of blue light can reveal the contribution of this signalling pathway to leaf gas exchange in different species. Previous work has suggested that BL stomatal response depends on the level of photosynthesis and/or respiration, however the majority of these studies have only considered short term responses and/or have used epidermal peels and guard cell protoplasts to prevent mesophyll interactions (Mawson, 1993;Suetsugu et al., 2014;Wang et al., 2014).
Here, we measured the impact of different light intensities and spectral quality (RL and BL) on gas exchange in intact leaves and used low [O 2 ] to inhibit respiratory processes to determine the relative contributions of respiration and photosynthesis to the rapidity and magnitude of the stomatal response.

Accepted Article
This article is protected by copyright. All rights reserved and Zea mays (Z.M, maize) were germinated in 5L pots containing peat-based compost (Levington F2S).
Following germination, plants were grown under greenhouse conditions and well-watered during the experiment. Solar radiation was complemented with sodium vapour lamps (∼200 to 400 µmol m −2 s −1 , Hortilux Schreder 600W, Monster, Netherlands) to maintain a 12h photoperiod.

Leaf gas exchange measurements
Net CO 2 assimilation (A) and stomatal conductance to water vapour (g sw ) were measured every 10s on the youngest fully expanded leaf using an infrared gas analyser (Li-Cor 6400 and 6800, Lincoln, NB, USA).
Leaves were first equilibrated at a PPFD of 100 μmol m −2 s −1 until both A and g sw reached 'steady state'.
Once steady state was reached, PPFD was increased to 1000 μmol m −2 s −1 for 30 min before returning to 100 μmol m −2 s −1 for 30 min. The light spectrum was set initially to 'Red only' (RL, peak wavelength: 625 nm) or 'Red+Blue' (90% Red/10% Blue), and once complete, the same protocol was repeated with the light spectrum inversed, by adding or removing 10 μmol m −2 s −1 (at 100 PPFD) and 40 μmol m −2 s −1 (at 1000 PPFD) of blue light (BL, peak wavelength: 475 nm). The leaf cuvette was maintained at 400 μmol mol −1 CO 2 concentration (C a ), a leaf temperature of 22°C (±0.2°C) and a leaf VPD of 1.1±0.1 kPa. All measurements were performed before 2pm to minimize any diurnal or circadian effects on gas exchange.
Measurements under low O 2 concentration (<1%) were performed using an infrared gas analyser (Li-Cor 6800, Lincoln, NB, USA) with the inlet connected to an Oxygen-free Nitrogen cylinder (British Oxygen Company-Industrial Gases, Ipswich, UK). A T-fitting was used to avoid excess flow coming from the pressurized cylinder that could damage the pump. A flow meter monitored the incoming flow and made sure that the excess was vented off and no outside air was pumped in. CO 2 and H 2 O were added to the mix by the Li-Cor 6800 and the infrared signal was corrected for a 1% [O 2 ]. The leaf cuvette was maintained at 400 μmol mol −1 CO 2 concentration (C a ), a leaf temperature of 22°C (±0.2°C) and a leaf VPD of 1.1±0.1 kPa. It is important to note that using [O 2 ] > 1% in wheat led to different results from using [O 2 ] < 1%, as the inhibition of mitochondrial respiration is highly sensitive to [O 2 ] (Forrester et al, 1966;Zabalza et al, 2009) and that values >1% did not produce complete inhibition.
Over a diurnal period, gas exchange measurements were performed simultaneously on flag leaves of two different tillers. Dark acclimated leaves were placed in the leaf cuvette and left to acclimate to the new conditions for 10min in the dark. The Licor 6800 was programmed using a custom python script that Accepted Article recorded gas exchange every 2min and changed the light intensity on average every 4 min to follow a predetermined pattern. The light pattern was described in a table (CSV file) containing the dwelling time and the intensity of the light (Fig. S1). A match of the two IRGA was automatically performed every 30 min to correct for any potential drift during the diurnal period. The leaf cuvette was maintained at 400 μmol mol −1 CO 2 concentration (C a ), a leaf temperature of 22°C (±0.4°C) and a leaf VPD of 1.1±0.1 kPa. Each measurement was started at 8 am to avoid differences due to circadian effects.

Time integrated leaf gas exchange
The responses of A and g sw under 'Red' and 'Red+Blue' lights were integrated starting from the increase in light intensity and for the following 30min. The spline function 'splinefun' was used to produce a continuous output from discrete observations and was used by the 'integrate' function from R to calculate the area under the curve. The percentage difference between values obtained for both light spectra were calculated and used to compare the blue light induced effects on gas exchange. To test if BL induced a significant increase in A and g s , the percentage increase in A and g s under BL was tested using a one sample t-test comparing the percentage increase to 0 for each species.

Modelling the g sw response to a step increase in light intensity
The temporal kinetics of g sw in response to a step change in light intensity was modelled using two sets of equations describing the shape and the magnitude of the response. The shape was modelled using an exponential response: = with s the current value and τ time constant. The steady state target (S) was modelled as an exponential response changing between 0 and 1 and included a feedback loop (S is dependent of the current s) producing a slow initial increase until s reach a threshold value triggering an exponential response of s.
This behaviour resulted in an equation capable of reproducing the exponential and sigmoidal response curve generally observed for g sw : = 1 -with λ the value corresponding to 63% of S. Increasing λ result in an increased initial lag time.
The results were then scaled using:

Accepted Article
This article is protected by copyright. All rights reserved = • ( -) + with and representing the initial and final steady state values for g sw .

Modelling circadian driven g sw response
Temporal kinetics of g sw under weak light intensity can be modelled by two sinusoidal functions describing the variation of the steady state target (S g ) through time (t): with P x the magnitude of response, T m the time at which the maximum response is reached and T s the period.
The rapidity at which g sw followed S g was described by an exponential differential equation: = where represented a time constant (i.e. the time for g sw to reach c.a. 63% of S g ). Different values of were used to describe an increase ( ) and a decrease in . ( )

Modelling the coupled g sw and A response to step changes in light intensity and quality under different [O 2 ]
Modelling the dynamic of g sw was performed using the assumption that the red light induced stomatal response in steady state (G red ) was linearly related with A and R d (Wong et al., 1979;Ball et al., 1987) and was activated in presence of light: The steady state blue light induced stomatal response (G blue ) was modelled as an increase in g sw activated by the presence of blue light:

Accepted Article
with β the increase in g sw induced by blue light.
Both responses were added together to model the steady state g sw (G sw ) response to variations in light intensity and quality: with G min the value of G sw under darkness representing the incomplete stomatal closure.
The steady state target G sw was then used to model the temporal response of g sw with an exponential response: = with τ the time constant representing the time to reach 63% of the total g sw variation. Different values of were used to describe an increase ( ) and a decrease in .

( )
A modified version of the Farquhar, von Caemmerer, Berry model (FvCB, 1980) was used to predict A in steady state (A s ) as a function of g sw . The photosynthetic rate limited by Rubisco activity (A c ) was calculated as: The photosynthetic rate limited by RuBP regeneration (A j ) was calculated as: with g tc the total conductance to CO 2 ( , g bw the boundary layer conductance = 1 ( 1.6 + 1.37 ) to water vapour), g m the mesophyll conductance to CO 2 , Vc max the in vivo maximum rate of RuBP carboxylation, R d the mitochondrial respiration, K m the Michaelis-Menten constant ( , = with τ A the time constant representing the time to reach 63% of the total A variation.

Modelling the diurnal response of g sw and A under fluctuating light intensity
The diurnal response of A and g sw differs from the response to step changes in light intensity by the fact that circadian processes, such as those observed under weak light intensity drive part of the responses.
Therefore, the diurnal model for A and g sw combined the findings from the previously described model.

Bayesian inference
Parameter values from the previously described models were adjusted using 'CmdStan' a software for statistical inference. Data were prepared in R and the models were written in the 'Stan' language. For each model, 4 Monte-Carlo Markov chains were produced that converged to the same optimum values.
There was no divergent transition during the process and the effective sample sizes were all above 100.
The Bayesian inference results in the estimation of 95% credible intervals for each parameter are considered significantly different with p < 0.05 if they are not overlapping (or if the difference between two intervals does not contains 0).

Accepted Article
Using red light (RL, peak wavelength: 625 nm) with or without the addition of blue light (BL, peak wavelength: 475 nm), the rapidity and magnitude of g sw responses were examined in response to a step change in light intensity (mimicking a sun-fleck) in species of scientific or agronomical importance (Fig. 1).
The addition of BL to a RL background induced species-specific stomatal responses to the change in light intensity with increases in g sw (when present) that were not constant over time. Therefore, to compare the effect of BL in different species, the differences between the RL and RL+BL induced responses of g sw and A were expressed as a time integrated difference in percent relative to the RL treatment ( Fig. 2A,B).
The presence of weak BL resulted in most cases in an increase in both g sw and A integrated over time, ranging from a few percent to ca. 100% for g sw and ca. 30% for A. The presence of weak BL significantly enhanced the magnitude of the g sw response in most species except for Solanum tuberosum (S.T) and Zea mays (Z.M). In general, increases in g sw with the addition of BL were accompanied by increases in A, although the differences were not always comparable or significant (Fig. 2B). Interestingly, major crops such as Glycine max (G.M), Triticum aestivum (T.A) and Oriza sativa (O.S) showed substantial increases in time integrated A suggesting that without the BL response, stomata strongly restricted CO 2 diffusion for A ( Fig. 2B). Part of the observed diffusional limitation under RL was due to slow stomatal kinetics, with plants generally displaying a sigmoidal response for g sw with an initial time lag characterized by a quasiabsence of response followed by an exponential increase (Fig. 1). A model describing the kinetic of g sw was used to quantify the importance of these two phases and how the addition of BL altered the kinetics ( Fig. 2C,D). In most species, the initial time lag (λ) was significantly reduced with the addition of BL except for P.G, O.S and A.T, although the differences were relatively small (Fig. S2A, 1C). The time constant (τ) describing the time required to reach 63% of the observed g sw variation showed that the addition of BL did not necessarily resulted in faster stomatal responses (Fig. S2B, 1D). In P.V, A.S and C.C, τ values were significantly higher with the addition of BL, and significantly lower in H.V, T.A, O.S, A.T (Fig. 2D). Overall, stomatal kinetics were significantly altered by the addition of BL with for main consequence an increase in A and a reduction in intrinsic water use efficiency (W i = A / g sw ), although the effects were species specific.
The biological significance of BL on gas exchange was tested over a diurnal period in two species with low and high g sw sensitivity to BL, Nicotiana tobaccum (N.T) and T.A respectively (Fig. 2E,F). A diurnal light regime mimicking natural fluctuations in light intensity was used with or without the addition of BL to a RL background (Fig. S1) to assess the effect of BL on diurnal gas exchange. During the diurnal period, N.T did not display any significant differences in g sw with the addition of weak BL on a RL background (Fig. 2E).

Accepted Article
This article is protected by copyright. All rights reserved Under the same conditions, T.A showed a large increase in g sw through the diurnal period, confirming the large effect of BL on stomatal behaviour described in Fig. 1. In T.A, BL induced a significant increase in A of ca. 9% (2.3 µmol m -2 s -1 ; p<0.05), whilst no significant difference in A was observed for N.T (Fig. 2F). The increase in A in T.A was not sufficient to compensate for the significant decrease of ca. 22% in W i (Fig. S3).
It is interesting to note that over the course of the diurnal period, the difference in g sw with the addition of BL increased with time, possibly driven by an endogenous signal, with differences in g sw values up to 0.2 mol m -2 s -1 in the later part of the diurnal period (Fig. 2E).

Stomatal responses to weak blue light are driven by respiratory processes and depend on endogenous signals
The diurnal BL and RL stomatal responses observed in wheat leaves subjected to a constant weak RL or weak BL (≤ 10 umol m -2 s -1 ) revealed variations in g sw that can be interpreted as a response to an internal signal, here called "endogenous" signal ( Fig. 3A). After an initial increase in g sw , a bi-modal response was observed with peaks at ca. 1h 40 and 5h into the photoperiod that was not reliant on photosynthesis.
Under RL, this endogenous response was shown to contribute up to ca. 20% of the diurnal g sw variation observed in Fig. 2E. In comparison, BL resulted in both a faster initial g sw increase and a 50% higher g sw over the diurnal period.  S4) demonstrating that any lack of response in Fig. 3B was not due to damage or impaired function. Using a pattern of alternating dark and low light (without driving photosynthesis) to maintain a high energy demand for stomatal movement (opening and closing, Fig. 3C), we observed that the rapidity of the g sw response decreased after each cycle and displayed a longer initial lag time (Fig. 3D). The highest g sw achieved during each light period followed a similar trend to those observed in Fig. 3A. Despite the low light intensity used, that was not sufficient to drive photosynthesis above the light compensation point, g sw displayed rapid increases similar to those observed when high light intensity was used (Fig. 1).

Accepted Article
This article is protected by copyright. All rights reserved

Blue light enhances the rapidity of stomatal movements in interaction with respiratory processes
To evaluate the contribution of photosynthetic and respiratory processes on g s responses, wheat leaves were subjected to a step change from dark to high red-light intensity at both ambient and low [O 2 ] (inhibiting respiratory processes) (Fig. 4A,B). Surprisingly, under low [O 2 ] g sw increased to only 38 mmol m -2 s -1 after 60 min (Fig. 4A) and was still slowly increasing after 4h (Fig. 4C) Fig. 4A,B) and therefore most likely captures the key mechanistic responses. The g sw response was 48% faster and g sw increased by 29% under BL. Furthermore, when respiratory processes were inhibited, the model estimated a two-fold decrease in the parameter value controlling the coupling between A and g sw , a doubling in the time required for a full induction of photosynthesis (time constant, K ai ) and a 64% lower maximum carboxylation of Rubisco (Vc max ) (Fig. S6). These results highlighted that BL

Accepted Article
This article is protected by copyright. All rights reserved impacts not only the kinetic of g s but also directly influences the induction of photosynthesis and confirmed the role of respiratory processes in powering stomatal opening even under high irradiance.

Respiration is required for stomatal opening and closing
In light (RL+BL) acclimated leaves, inhibiting respiration (switching from 21% to 1% [O 2 ]) resulted in a slow decrease in g sw , which was independent of the light intensity (Fig. 5A) or the photosynthesis level (Fig. 5B).
When [O 2 ] was restored to 21%, g s slowly returned to its initial level. These data also revealed that the g sw response to intercellular [CO 2 ] (C i , Fig. 5B), which usually induces stomatal opening was overridden under these conditions. Furthermore, stomata in T.A and N.T were unable to fully close for more than 60 min when placed simultaneously under low [O 2 ] and darkness ( Fig. 5B-D) stressing the importance of respiratory energy for both stomatal opening and closing at any time of the diurnal period.

Parallel contribution of respiratory and photosynthetic processes to diurnal gas exchange
Using a leaf gas exchange model including the previous findings, the contribution of respiratory and photosynthetic processes to diurnal gas exchange (Fig. 6A,B) was quantified in response to external and internal cues. The model was able to describe with high accuracy (RMSE for g sw : 0.013 mol m -2 s -1 and for A: 0.6 µmol m -2 s -1 ) the diurnal kinetics of g sw under RL and RL+BL, which provided insights into the contribution of each process driving g sw . Under RL, the feedback loop describing the coordination of A and g sw explained c.a. 80% of the diurnal variation in g sw , whilst the remaining c.a. 20% were due to endogenous signals mostly driven by respiratory processes. Adding weak BL during the diurnal period resulted in a faster increase (K i ) and slower decrease (K d ) in g sw in response to variation in light intensity, promoting higher levels of g sw and lower A limitation (Fig. 6C,D). The g sw response to an endogenous signal(s) was doubled under BL (Fig. 6E,F) and showed a faster increase (Fig. 6G) and a slower decrease ( Fig. 6H) resulting in a higher g sw through the diurnal period. Greater g s resulted in a faster A induction and greater A reached during the diurnal period.
Simulations using the leaf gas exchange model used in Fig. 6 showed that the initial g s value under darkness is an important determinant of the temporal kinetic of A and g s in response to an increase in

Accepted Article
This article is protected by copyright. All rights reserved light intensity (Fig. 7). Under RL, a low g s value induced a strong limitation of A resulting in a slow increase in g s . The presence of weak blue light greatly improved the rapidity and magnitude of the g s response resulting in a faster A induction. Simulations showed that maintaining high g sw values under dark conditions like those observed in N.T (Fig. 5D) can compensate for the lack of blue light induced stomatal opening during induction of A. The model also highlighted the interdependence of A and g s by illustrating how the induction speed of A and the coupling with an increase in g s influences their respective temporal responses. Overall, gas exchange simulations revealed that the coordination between A and g s is not necessarily linear and is greatly improved in favour of A when the blue light pathway is activated.

Discussion
Little is known about the role of BL in the regulation of diurnal gas exchange and the reason it has been evolutionary conserved in many species (Li et al., 2015;Doi et al., 2015). Previous studies hypothesized that BL remove stomatal limitation of photosynthesis early in the morning (Assmann & Shimazaki, 1999), which was supported here by the large and rapid increase in g sw observed in dark adapted plants subjected to weak BL. Our analysis went further and suggested that BL stomatal opening reduces diffusional limitations on A over the diurnal period in most species although at the cost of decreased W i , and is species specific. Traditionally, variations in g s have been predicted from variations in A (Ball et al., 1987), however our findings revealed that this relationship only determines a fraction of the g s achieved during a diurnal period. This is due to the fact that both photosynthesis and respiration are required for rapid stomatal opening and that respiratory processes specifically contribute to BL induced stomatal opening independent of A, which is overlooked in current views on stomatal function. These observations will help re-shape our understanding of the diurnal dynamic coordination between A and g s , an unexploited target for enhanced crop performance and water use efficiency.

Blue light reduces diffusive limitation and improves dynamic photosynthesis
In the species studied here, leaves subjected to BL showed diverse magnitude and rapidity of g sw responses to BL that often resulted in a greater time integrated A (due to a faster induction and increased magnitude of g sw ) and a reduction in W i . In most species, the increase in g s in response to the addition of BL was proportionally greater than the increase in A and is therefore a potential target to enhance W i whilst maintaining A. Some of the major cereal grain and legume crop species used here (rice, wheat, barley and oat, soybean, pea) displayed large and significant g s responses with the addition of BL

Accepted Article
suggesting that breeding programs may have already inadvertently selected for this trait to enhance A, at the detriment of W i . In contrast, species of the Solanaceae Family (tomato, potato, tobacco) displayed only a relatively small or no g s increase with the addition of BL, suggesting a potential evolutionary aspect to the blue light signalling pathway. In wheat, higher g s has been shown to be positively correlated with yield (Reynolds et al., 1999;Fischer & Rebetzke, 2018), due to both reduced diffusional limitation as well as enhanced evaporative cooling for leaf temperature regulation. Despite the fact that most species seem to possess the major known genes involved in the BL signalling pathway (e.g. PHOT1, PHOT2 and BLUS1; Takemiya et al., 2013;Li et al., 2015), some species such as S.T (potato) and Z.M (maize) did not display a significant stomatal response to the addition of BL. The cause of this absence of response is currently unknown but could be a result of constant activation or inactivation of genes involved in BL signal transduction pathway, and further research would be required to understand these inter-specific differences in the BL induced response.
To date, the majority of publications examining BL induced stomatal responses have been performed using short term protocols (sec to min; Matthews et al., 2020) and overlook the long term variations in g s with the addition of BL. Our results revealed that the increase in g sw observed with the addition of BL increased continuously during the diurnal period, independently of the external conditions. This was further supported by the observed impact of the endogenous signal on g sw , which was 50% greater under BL and accounted for ca. 25% of the maximum g sw observed under fluctuating high light intensity. The self-entrained variation in g sw was triggered by weak light intensity (below the compensation point) and resulted in a pattern similar to those reported for the circadian clock (Gorton et al., 1993). These data agree with the concept of a diurnal endogenous signal or circadian system influencing stomatal behaviour over the course of the day, contributing to variations in W i (Gorton et al., 1993;Matthews et al., 2018;Simon et al., 2020) but also highlighting increased sensitivity to BL.

Respiratory processes are essential for rapid variations and maintenance of stomatal conductance
We hypothesized that the coordination observed between A and g s (McAusland et al., 2016) is due in part to the energy requirements of stomatal movements that are fulfilled at different times of the response by mitochondria and/or chloroplasts. Indeed, Mawson (1993) highlighted that the translocation of protons across the guard-cell plasmalemma, the first step of stomatal aperture is an energy-requiring activity. The study suggested that both guard-cell chloroplasts and mitochondria contribute in synergy to supply energy for BL-induced proton pumping in guard-cell protoplasts, as both were inhibited by low [O 2 ].
Earlier work using plants grown with the herbicide norflurazon or using the white areas of variegated plants, showed that the stomata in such plants were still able to respond to BL but the RL response was Accepted Article greatly impaired, suggesting that alternative pathways to photosynthesis can provide the energy for stomatal responses (Karlsson et al., 1983;Roelfsema et al., 2006). The contribution of energy derived from both chloroplasts and mitochondria to drive the stomatal response was confirmed by our results in intact leaves under both red and blue light. During the diurnal period, our results demonstrate the contribution of energy and osmotica originating from the mesophyll on stomatal behaviour, which is not possible to study in guard cell protoplasts. In addition to the activation of the plasma membrane H + -ATPase, BL has been shown to influence sugar/lipid degradation pathways (Horrer et al., 2016;McLachlan et al., 2016) releasing the energy required (e.g. via mitochondria, Medeiros et al., 2018) for the activation/deactivation of ion channels and pumps (Marten et al., 2007;Inoue et al., 2020), which promotes stomatal opening independently of A. By inhibiting respiratory processes during a transition from dark to high light intensity, stomata can rely only on photosynthetic processes (osmoregulation and energy) for opening, which are initially limited by CO 2 diffusion that induces the observed strong coupling between A and g s . This supported the long-standing idea of a positive feedback loop controlling stomatal aperture based on mesophyll photosynthesis, with C i potentially coordinating A and g s (Farquhar et al., 1978). In parallel with this feedback loop, our results suggest that respiratory processes and BL played a key role by initiating and promoting fast stomatal opening independently of A. Our model describing the temporal response of A and g s using a feedback loop estimated a two-fold decrease in the rapidity of photosynthesis induction and maximum rate of carboxylation under low [O 2 ]. In absence of respiratory energy, the initial low and slow g s response limited CO 2 diffusion, prevented the CO 2 and energy (i.e., ATP) requirement of Rubisco activation to be met, and resulted in a lowered maximum carboxylation rate. Figure 8 illustrates the theoretical framework for the interplay between photosynthesis and stomatal behaviour following a step increase in red or blue light and highlights the importance of mitochondrial respiration in initiating and driving rapid stomatal movements. These findings provide a unifying theory for the observed "RL" (photosynthesis dependent) and "BL" (photosynthesis independent) induced stomatal opening that has been discussed extensively in the literature.
It is important to acknowledge that using low [O 2 ] to inhibit respiration can also inhibit other processes such as the production of reactive oxygen species (ROS), which are important signalling molecules in stomatal closure (Enhonen et al., 2017). However, the absence of ROS production under low [O 2 ] would be expected to increase or maintain stomatal aperture relative to ambient [O 2 ] conditions, which was not the case in our experiments (Fig. 4). When [O 2 ] was returned to ambient conditions g sw was restored, suggesting no damage. Although these findings do not totally exclude ROS as a signal, a lack of ROS Accepted Article production could not explain the observed g sw decrease under high light and low [O 2 ] (Fig. 5A). It has been suggested that stomatal closure under darkness (Fig. 5C-D) requires the accumulation of ROS (Desikan et al., 2004;Ma et al., 2018), which may explain why stomatal closure was impaired. In general, it is therefore unlikely that ROS production or lack of it can explain the observed differences in the rapidity of stomatal response shown here.
Over a diurnal period, our results suggest a significant role of BL along with the respiratory processes in powering and maintaining stomatal aperture independently of A. Under a repeated dark light cycle applied over 8 hours, the delay between the application of weak blue light and the g sw response increased after each cycle towards the end of the day, suggesting that stored energy used by respiratory processes had been exhausted. However, it is noteworthy there was sufficient energy to power stomatal movements for the majority of the day. These results suggest that the rapidity and magnitude of the diurnal g sw responses are determined in part by the status of the energy pool, and that photosynthates previously accumulated may influence stomatal behaviour.
Following a light to dark transition, stomatal closure in both wheat and tobacco subjected to low [O 2 ] was greatly impaired as previously observed in wheat and barley (Akita and Moss, 1973) and in Commelina communis L. (Karlsson & Schwartz, 1988). In wheat, an unexpected small increase in g s was observed, and in tobacco g s showed a slow decrease. The g s increase in wheat could be either due to an increase in guard cell turgor and/or a decrease in subsidiary cell turgor (Franks & Farquhar, 2007) (Karlsson & Schwartz, 1988;Willmer & Fricker, 1996) and our results stress the importance of respiratory processes in supplying energy for guard cells movements. Moreover, our results in wheat showed that stomata can maintain aperture for >1h without energy and that energy is only consumed to drive changes in guard cell turgor.
Surprisingly, light acclimated leaves subjected to low [O 2 ] displayed a strong decrease in g s independently Accepted Article of light intensity, highlighting that respiratory processes also play a key role in the maintenance of high g sw throughout the diurnal period. A gas exchange model describing the role of respiratory processes in stomatal behaviour estimated up to 50% of the diurnal g s under a 'natural' light regime was attributed to these processes. Overall these results suggest that the large number of mitochondria in guard cells (Shimazaki et al., 2007) are key for stomatal movements, especially when A is limited (e.g. early in the morning or during sun-flecks), as well as for the maintenance of high g s and stomatal closure (e.g. end of the day or during shade-flecks).

Blue light acts as a dark/light switch optimizing stomatal behaviour for water saving
So far, our results have shown the influence of BL on diurnal stomatal behaviour and the resulting impact on photosynthesis. However, they do not explain why BL induced stomatal opening during the diurnal period was essential for removing stomatal limitation of A in wheat but not necessary in tobacco. One major difference between these two species was the large nocturnal g s observed in tobacco compared to the tight stomatal closure in wheat. The large nocturnal g s in tobacco did not limit photosynthetic induction in the morning, whilst in wheat the rapid stomatal opening induced by BL was essential to remove the early morning diffusional limitation of A. This therefore suggests that a role for BL induced stomatal opening is to enable a low nocturnal g s in wheat without compromising photosynthetic induction, and to increase water savings by closing stomata during the night (Caird et al., 2006). The interspecific differences observed in response to BL suggest that different strategies for the regulation of diurnal gas exchange exist and requires further research to determine species specific advantages.

Conclusions
The BL induced stomatal response is generally assumed to occur in the majority of species, although there is evidence that it is lacking in some ferns (Doi et al., 2015;Westbrook & McAdam, 2020) and facultative CAM plants (when in CAM mode, Gotoh et al., 2019). The findings presented here demonstrate that the BL response is far from universal, and that there is a significant inter-specific diversity in stomatal response (both in rapidity and magnitude). Our results give a broader context to the importance of BL for gas exchange and show how BL induced stomatal responses interact with photosynthetic and respiratory processes over the course of the day. At the beginning of the day, BL was shown to uncouple g sw variation from A and enables rapid stomatal opening, removing diffusional limitations and facilitating lower g sw during the night. Respiratory and photosynthetic processes were both required for rapid stomatal movements and to maintain a high g sw . We predict here that the energy produced by respiratory processes could drive up to 60% of the observed diurnal variations in g sw , with photosynthesis providing

Accepted Article
This article is protected by copyright. All rights reserved the remaining 40% in wheat. Furthermore, we have shown that respiratory processes are essential for stomatal movements (opening and closing) and maintenance of the steady state g sw in the light.
Surprisingly, in the absence of energy (both photosynthetic and respiratory), g sw remained unchanged (for >1h) in wheat suggesting there is minimal energy requirement to maintain g sw at the current level. These results emphasize the unexploited potential of temporal optimisation of A and g sw over the diurnal period, based on tuning the sensitivity of the stomatal response to BL and the ratio of photosynthetic and respiratory energy driving the response.

Accepted Article
This article is protected by copyright. All rights reserved

Accepted Article
This article is protected by copyright. All rights reserved Triticum aestivum (T.A), under fluctuating light intensity. Ribbons represent standard error around the mean. F, Impact of the addition of blue light on diurnal kinetics of A under fluctuating light intensity. Error bars represent the standard error around the mean, n = 5 to 16 biologically independent samples, and * represents a significant difference between light treatments (p<0.05).

Fig. 3: Contribution of respiratory processes to light induced stomatal responses in Triticum aestivum.
A, Observed and modelled (black line) g sw response to constant 10 µmol m -2 s -1 of red or blue light. Shaded area represents the standard error around the mean, n = 5 biologically independent samples. The plants were maintained under dark conditions (grey area) before measurements. B, Response of stomatal conductance (g sw ) to a step change from 0 (dark shaded area) to 5 µmol m -2 s -1 of blue light under 1% and 21% [O 2 ]. Shaded areas represent the standard error around the mean, n = 5 and 6 biologically independent samples. C, Response of stomatal conductance (g sw ) to a cycle of step changes in light intensity from 0 to 10 µmol m -2 s -1 of blue light. The light intensity was maintained for 20 min between steps. D, Successive g sw kinetics in response to step increase from darkness to weak blue light observed in C. All responses were rescaled to start at t=0 highlighting the change in the rapidity of the g sw response over the diurnal period.

Accepted Article
This article is protected by copyright. All rights reserved represents the dark period. Ribbon around the curves represent the standard error around the mean, n = 5 biologically independent samples. D, the same protocol described in (C) was applied to N. tabacum, n = 4 biologically independent samples.