High Stomatal Conductance in the Tomato Flacca Mutant Allows for Faster Photosynthetic Induction

Due to their slow movement and closure upon shade, partially closed stomata can be a substantial limitation to photosynthesis in variable light intensities. The abscisic acid deficient flacca mutant in tomato (Solanum lycopersicum) displays very high stomatal conductance (gs). We aimed to determine to what extent this substantially increased gs affects the rate of photosynthetic induction. Steady-state and dynamic photosynthesis characteristics were measured in flacca and wildtype leaves, by the use of simultaneous gas exchange and chlorophyll fluorometry. The steady-state response of photosynthesis to CO2, maximum quantum efficiency of photosystem II photochemistry (Fv/Fm), as well as mesophyll conductance to CO2 diffusion were not significantly different between genotypes, suggesting similar photosynthetic biochemistry, photoprotective capacity, and internal CO2 permeability. When leaves adapted to shade (50 µmol m−2 s−1) at 400 µbar CO2 partial pressure and high humidity (7 mbar leaf-to-air vapour pressure deficit, VPD) were exposed to high irradiance (1500 µmol m−2 s−1), photosynthetic induction was faster in flacca compared to wildtype leaves, and this was attributable to high initial gs in flacca (~0.6 mol m−2 s−1): in flacca, the times to reach 50 (t50) and 90% (t90) of full photosynthetic induction were 91 and 46% of wildtype values, respectively. Low humidity (15 mbar VPD) reduced gs and slowed down photosynthetic induction in the wildtype, while no change was observed in flacca; under low humidity, t50 was 63% and t90 was 36% of wildtype levels in flacca. Photosynthetic induction in low CO2 partial pressure (200 µbar) increased gs in the wildtype (but not in flacca), and revealed no differences in the rate of photosynthetic induction between genotypes. Effects of higher gs in flacca were also visible in transients of photosystem II operating efficiency and non-photochemical quenching. Our results show that at ambient CO2 partial pressure, wildtype gs is a substantial limitation to the rate of photosynthetic induction, which flacca overcomes by keeping its stomata open at all times, and it does so at the cost of reduced water use efficiency.


INTRODUCTION
In the leaves of higher plants, stomata balance carbon uptake against water loss. They achieve this balance by dynamically regulating stomatal aperture in response to intrinsic and extrinsic factors. Typically, stomatal aperture decreases in low irradiance or darkness, and increases in high irradiance. Stomatal opening after sudden increases in irradiance is slow compared to changes in Calvin cycle metabolism (McAusland et al., 2016), with time constants in the range of 4 to 29 min (Vico et al., 2011). Due to the slow opening of stomata, the increase of stomatal conductance (g s , mol m −2 s −1 ) from a low initial value is assumed to be one of the three main limitations of net photosynthesis rate (A, µmol m −2 s −1 ) in response to increases in irradiance (e.g., Urban et al., 2008;Kaiser et al., 2016;Li et al., 2016;Kaiser et al., 2017a;Zhang et al., 2018;Zhang et al., 2020). Given that solar irradiance incident on a leaf often fluctuates, these dynamic limitations of photosynthesis decrease photosynthetic irradiance use efficiency . Improving g s , including its dynamics, is an attractive means with which to improve both irradiance and water use efficiency (Lawson and Blatt, 2014;Vialet-chabrand et al., 2017). Increases in the limitation imposed upon A by g s can be identified via transient decreases of leaf internal CO 2 partial pressure (C i , µbar). The other two main limitations during photosynthetic induction arise from slow rates of change in the activity of enzymes involved in ribulose-1,5bisphosphate (RuBP) regeneration, and from slow activation of Rubisco (Pearcy et al., 1996;Way and Pearcy, 2012;Kaiser et al., 2015;Kaiser et al., 2018). These limitations occur additionally to the limitations at steady state due to, e.g., the rate of electron transport, Calvin cycle metabolism, sucrose metabolism, or mesophyll conductance (g m , mol m −2 s −1 ).
Disentangling stomatal and other limitations during photosynthetic induction is difficult. Many treatments affecting g s also affect other transient limitations of photosynthetic induction, such as leaf temperature (Kaiser et al., 2017a) and salt stress (Zhang et al., 2018). Similarly, some genetic mutations affecting g s , such as the abscisic acid (ABA) deficient aba2-1 mutant in Arabidopsis thaliana, showed enhanced A/C i responses compared to its wildtype, Col-0 , suggesting pleiotropic effects, which may confound effects of altered g s . Also, models estimating transient stomatal limitation have often been based on linearnot curvilinear -A/C i relationships (Woodrow and Mott, 1989;Tinoco-Ojanguren and Pearcy, 1993), or are based on steady-state A/C i responses (Kaiser et al., 2017a). In fact, the maximum rate of carboxylation (V cmax ), for example, increases strongly during photosynthetic induction (Soleh et al., 2016;Taylor and Long, 2017). In another approach, the limitation during induction was attributed to g s alone up to the point where A reached 95% of the steady-state value (McAusland et al., 2016), entirely ignoring limitations by RuBP regeneration and Rubisco activation kinetics. Tools to better separate stomatal from other limitations are thus warranted, and mutants or transformants with substantially altered stomatal characteristics but similar photosynthetic biochemistry (Raissig et al., 2017;Papanatsiou et al., 2019;Tomimatsu et al., 2019;Kimura et al., 2020;Yamori et al., 2020) can be counted among such tools.
The tomato (Solanum lycopersicum L.) flacca mutant has a 80% to 90% lower ABA content than its wildtype (Tal and Nevo, 1973;Sagi et al., 2002). Flacca leaves exhibit a very high g s , without affecting the A/C i response, suggesting that photosynthetic capacity is independent of ABA (Bradford et al., 1983). Lack of ABA has not been found to affect g m in the aba-1 mutant of Nicotiana plumbaginifolia (Mizokami et al., 2015) but to our knowledge has not been determined in flacca. The aim of this study is to determine the dynamic limitations of photosynthetic induction due to g s in tomato leaves. For this, steady-state and dynamic photosynthesis characteristics were measured in the ABA-deficient flacca mutant and its wildtype, by the use of simultaneous gas exchange and chlorophyll fluorometry.

Plant Material
Seeds of tomato cv. Rheinlands Ruhm wildtype (LA0535) and flacca (LA0673) were obtained from the Tomato Genetics Resource Center (University of California, Davis, USA). Seeds were germinated in stonewool plugs (Grodan, Roermond, NL). A week after sowing, they were transferred to stonewool cubes (10 cm × 10 cm × 7 cm; Grodan). Plants were grown in a climate chamber under a day/night cycle of 16/8 h (day/night), 20/18°C temperature, ambient CO 2 partial pressure, 70% relative air humidity, and 154 µmol m −2 s −1 photosynthetically active radiation (PAR; measured 10 cm above table height), which was provided by fluorescent tubes (Master TL-D 58W/840 Reflex Eco; Philips, Eindhoven, the Netherlands). Stonewool cubes were standing in a layer (height, 1-2 cm) of nutrient solution (Yara Benelux B.V., Vlaardingen, the Netherlands), which was replenished every 1 to 2 days and contained 12.4 mM NO 3 − , 7.2 mM K + , 4.1 mM Ca 2+ , 3.3 mM SO 4 2− , 1.8 mM Mg 2+ , 1.2 mM NH 4 + , 1.1 mM PO 4 3− , 30 mM BO 3 3− , 25 mM Fe 3+ , 10 mM Mn 2+ , 5 mM Zn 2+ , 0.75 mM Cu + , and 0.5 mM MoO 4 2− (EC 2.1 dS m −1 , pH 5.5). Between 1 and 4 weeks after sowing, flacca plants were sprayed daily with a solution containing 10 mM ABA, 0.01% (w/v) Triton-X, and 0.1% (v/v) ethanol (Bradford et al., 1983), using commercially available horticultural hand sprayers. Wildtype plants were sprayed with a control solution containing Abbreviations: A, net photosynthesis rate; A g , gross photosynthesis rate; ABA, abscisic acid; ANOVA, analysis of variance; C a , leaf external CO 2 partial pressure; C c , chloroplast CO 2 partial pressure; C i , substomatal CO 2 partial pressure; F, fluorescence emission from leaves under actinic irradiance; F m , maximal fluorescence in dark-adapted leaves; F m ', maximal fluorescence in light-adapted leaves; F o , minimal fluorescence in dark-adapted leaves; F v /F m , maximum quantum efficiency of PSII photochemistry; F v ′/F m ′, efficiency of open PSII traps; g m , mesophyll conductance; g s , stomatal conductance; J, rate of linear electron transport; MPF, multi-phase flash protocol; PAR, photosynthetically active radiation; qP, coefficient of photochemical quenching; R d , mitochondrial respiration; s, lumped parameter used to convert F PSII to J; TPU, triose phosphate utilization capacity; WUE i , intrinsic water use efficiency; V cmax , maximum rate of carboxylation; VPD, leaf-to-air vapour pressure deficit; G*, CO 2 compensation point in the absence of mitochondrial respiration; F PSII , photosystem II operating efficiency. 0.01% Triton-X and 0.1% ethanol. Untreated flacca plants are smaller and show much higher transpiration rates than the wildtype, together with leaf epinasty and strong root formation along the stem (Tal, 1966). Growing flacca with application of ABA causes plants to grow similarly well as the wildtype (Imber and Tal, 1970). When the application of ABA is stopped, flacca reverts to its mutant phenotype within days, including alwaysopen stomata (Imber and Tal, 1970). All chemicals were purchased from Sigma (St. Louis, MO, USA).

Measurements
When plants were between 5 and 6 weeks old, the fourth leaf, counting from the bottom, was selected for measurements. ABA spraying was stopped seven days before the start of measurements, to allow the high g s phenotype of flacca to reassert itself. Measurements were performed in a lab, using the LI-6400 XT photosynthesis system (LI-COR Biosciences, Lincoln, Nebraska, USA) equipped with a fluorescence chamber (leaf area: 2 cm 2 ). Conditions inside the leaf chamber during measurements were: 25°C chamber temperature, 7 mbar leaf-to-air vapour pressure deficit (VPD; except when stated otherwise) and a flow rate of 500 µmol s −1 . Irradiance was provided by LEDs in a 90/10 red/blue irradiance mixture, with peak intensities at wavelengths of 635 and 465 nm, respectively. For all measurements, five plants per genotype were used (n = 5). All measurements were performed on the same spot of a leaf, to reduce measurement noise caused by spatial variation (Lawson and Weyers, 1999;Matthews et al., 2017): (a) darkadapted F v /F m , (b) A/PAR curves at 2% oxygen, (c) A/C i curves at 2% oxygen, (d) A/C i curves at 21% oxygen, (e-g) photosynthetic induction under three different environmental conditions (described below). While measurements a-d were performed in the same sequence, the order of photosynthetic induction measurements was randomized for each plant. Values of A were corrected for CO 2 leakage based on the manufacturers' suggestions. Measurements were started at 7:30 in the morning and took 8 to 9 h to complete per leaf.

Dark-Adapted F v /F m and Net CO 2 Exchange in Darkness
Leaves were dark-adapted for 20 minutes. Then, net CO 2 exchange in darkness (A dark ) was logged, after which a weak measuring beam was turned on to measure F o . Then, F m was determined by exposing the leaf to a single-pulse saturating flash of~9.000 µmol m −2 s −1 intensity and 1-s duration. Dark-adapted

A/PAR Curves at 2% Oxygen
A gas mixture containing 2% oxygen and 98% nitrogen was fed to the inlet of the LI-6400 XT. Leaf external CO 2 partial pressure (C a ) was set to 2000 µbar, and irradiance was set to 200 µmol m −2 s −1 . After reaching steady-state A, irradiance was decreased in steps of 150, 100, 70, 50, and 30 µmol m −2 s −1 , and A was logged for 30 s after reaching steady-state A, at steps of 5 s. Values were later averaged at each step to reduce measurement noise.

A/C i Curves at 2% Oxygen
A gas mixture containing 2% oxygen and 98% nitrogen was fed to the inlet of the LI-6400 XT. Irradiance was set to 1,500 µmol m −2 s −1 , and C a was set to 150 µbar. After reaching steady-state A, C a was decreased in steps of 130, 100, 70, and 50 µbar. A and C i were logged as described above. At each C a , the infrared gas analysers were matched.

A/C i Curves at 21% Oxygen
Irradiance was set to 1,500 µmol m −2 s −1 , and C a was set to 400 µbar. After reaching steady-state A, C a was decreased in steps of 300, 200, 130, 100, 70 and 50 µbar. Then, C a was raised to 400 µbar and after reaching steady-state A, C a was increased in steps of 600, 750, 900, 1,100, 1,400, 1,700, and 2,000 µbar. A and C i were logged after reaching steady-state (3-5 min per step) as described above. At each C a , the infrared gas analysers were matched. Parameters describing maximum rate of carboxylation (V cmax ), rate of linear electron transport at the measuring irradiance (J 1500 ) and triose phosphate utilization capacity (TPU) were determined using the excel solver tool by Sharkey (2016). Additionally, operating and maximal fluorescence in light-adapted leaves (F and F m ′, respectively) were determined at each C a by using a multi-phase flash protocol (MPF; Loriaux et al., 2013). The maximum intensity of the MPF was~9.000 µmol m −2 s −1 , the durations of the three phases were 0.3, 0.7 and 0.4 s respectively, and the percentage decrease of flash intensity during phase two was 60%. These MPF settings were found to yield the most accurate results in pilot experiments (data not shown). Photosystem II operating efficiency (F PSII ) was calculated as F PSII = (F m ′ − F)/F m ′. Mesophyll conductance (g m ) was determined following the variable J method proposed by Harley et al. (1992); the variables A, C i , and F PSII to calculate g m were determined at a C a of 400 µbar and an irradiance of 1500 µmol m −2 s −1 . Parameters to calculate g m , namely R d , G* and s, were determined from A/C i and A/PAR measurements at 2 and 21% oxygen following Yin et al. (2009).

Photosynthetic Induction
Leaves were adapted to 50 µmol m −2 s −1 until g s was constant (40-60 minutes). Irradiance was then increased to 1500 µmol m −2 s −1 in a step change, and gas exchange values were logged every 2 s for 60 min. These measurements were performed at C a and air humidity close to the plant's growth conditions (400 µbar C a , 7 mbar VPD), termed "control" hereafter. Photosynthetic induction was additionally assessed under two other conditions: "high VPD" (15 mbar) and "low CO 2 " (200 µbar), keeping all other conditions the same. During photosynthetic induction, chlorophyll fluorescence was measured using a saturating MPF (described above) once every minute during the first ten minutes, and once every two minutes thereafter. Photosynthetic induction (PI, %) was calculated as a percentage of the total change between initial A (A i ) and final A (A f ) of each transient: PI = (A-A i )/(A f -A i )*100. Intrinsic water use efficiency (WUE i ) was calculated as WUE i = A/g s . Non-photochemical quenching (NPQ) during photosynthetic induction was calculated as NPQ = (F m − F m ′)/ F m ′. The coefficient of photochemical quenching (qP) and the efficiency of open photosystem II traps (F v ′/F m ′) were calculated after Oxborough and Baker (1997) Oxborough and Baker (1997).

Statistical Analysis
All statistical tests were performed at P=0.05 as threshold for significance. Where appropriate, a two-sided Student's t-test was used to determine significant differences between genotypes. For photosynthetic induction under different environmental conditions, a two-way analysis of variance (ANOVA) was performed, and interaction means were separated based on Fisher's least significant difference test. Residuals were tested for normal distribution (Shapiro-Wilk test) and equal variances were assumed for treatment groups. If the requirement for normal distribution was not fulfilled, the procedure was repeated on log-transformed data. If after log transformation residuals still did not show normality, Kruskal-Wallis one-way ANOVA, considering six treatments (three environmental conditions times two genotypes), was performed using the original data. In case of significant treatment effects, Dunn's test of multiple comparisons was performed to identify differences between the six treatments. All statistical tests were performed in Genstat (VSN international, Hempstead, UK) except for Dunn's test, which was performed in R (R Core Team, 2020) using the dunn.test package (Dinno, 2017).

Steady-State CO 2 and Irradiance Responses of Photosynthesis
Wildtype and flacca leaves showed very similar responses of A and F PSII to C i (Figure 1). In the CO 2 range 50 to 300 µbar, A increased near-linearly, then peaked at~500 µbar and with further increases in C i , A declined in both genotypes. Compared to A, F PSII peaked at lower C i (~300 µbar) and exhibited a stronger decline with further increases in C i . Parameters describing photosynthetic capacity, i.e., V cmax , J 1500 and TPU, were not significantly different between genotypes (Figure 1, insert). Mesophyll conductance and its components were not significantly different between genotypes (except C i , which was significantly greater in flacca, Figure 2C), although flacca tended to show greater values for A, J, R d , and C c ( Figure  2). In dark-adapted leaves, A was −1.2 ± 0.1 µmol m −2 s −1 in wildtype and −1.9 ± 0.1 µmol m −2 s −1 in flacca leaves (p=0.008). At low irradiance (50 µmol m −2 s −1 ), on the other hand, A was similar between genotypes ( Table 1).

Response of Photosynthetic Gas Exchange to a Stepwise Irradiance Increase
Next, we tested how gas exchange in wildtype and flacca leaves that had been adapted to low irradiance (50 µmol m −2 s −1 ) reacted to a stepwise increase to high irradiance (1500 µmol m −2 s −1 ). In wildtype leaves, the rate of photosynthetic induction was slower at high VPD, compared to the other two treatments (low CO 2 or high VPD; Figure 3A), while in flacca, there was no difference between control and high VPD treatments ( Figure 3B). However, while in wildtype leaves the rate of photosynthetic induction was the same in the control and low CO 2 treatments, in flacca, induction at low CO 2 was slower than in the control treatment ( Figures 3A, B). The flacca mutation had significant effects on the times to reach 50 (t 50 ) and 90% (t 90 ) of full photosynthetic induction: under control conditions, t 50 was 91% and t 90 was 46% of wildtype values in flacca, while under high VPD, t 50 was 63% and t 90 was 36% of wildtype values in flacca (Figure 4). Both indices were not significantly different between genotypes under low CO 2 (Figure 4). Transient A was higher in flacca than in wildtype leaves, and in both genotypes was slightly higher in control than in high VPD, as well as substantially reduced at low CO 2 (insets in Figures 3A, B; Table 1).
In flacca, A showed a small decrease between~1.5 and 2.0 min after the stepwise increase in irradiance under control and high VPD conditions ( Figure S1B). These dynamics are very similar to those previously seen in shade-adapted wildtype tomato leaves undergoing photosynthetic induction under high CO 2 partial pressure (Kaiser et al., 2017b), and in the present study were observed neither at low CO 2 ( Figure S1B), nor in wildtype leaves ( Figure S1A). The most likely explanation for this phenomenon is a transient mismatch between the rate of CO 2 fixation in the Calvin cycle and downstream sucrose metabolism, which would transiently limit the availability of free phosphate in the chloroplast (Prinsley and Leegood, 1986;Stitt and Grosse, 1988;Stitt and Quick, 1989).
During the complete trajectory of photosynthetic induction, stomatal conductance (g s ) in wildtype leaves was strongly increased by lowering CO 2 , and substantially reduced by increasing VPD ( Figure 3C; Table 1). In flacca, g s was lower FIGURE 1 | Steady-state responses of net photosynthesis rate (A) and photosystem II operating efficiency (F PSII ) to leaf internal CO 2 partial pressure (C i ), in wildtype (WT) and flacca (flc) leaves, measured at 21% oxygen. Insert: parameters derived from curve fitting (after Sharkey, 2016) to A/C i response (unit for all: mmol m −2 s −1 ), namely maximum rate of carboxylation (V cmax ), rate of linear electron transport at 1500 mmol m −2 s −1 PAR (J 1500 ) and triose phosphate utilisation capacity (TPU); P values for comparisons between WT and flc based on two-sided t-test. Symbols and numbers represent averages ± SEM, n = 5.
in the high VPD treatment compared to both other treatments during photosynthetic induction ( Figure 3D; Table 1). Intriguingly, while the low irradiance adapted (initial) g s in wildtype leaves reacted to the treatment levels in a predictable way, i.e., decreasing under high VPD and increasing under low CO 2 relative to control ( Figure 3C), initial g s in flacca did not ( Figure 3D; Table 1). Also, g s in flacca was substantially higher than that in wildtype leaves in all cases, as would be expected of an ABA mutant.
In control conditions, C i in wildtype leaves showed an initial decrease in the first 10 minutes of photosynthetic induction, after which it gradually increased as stomata opened ( Figure 3E). This decrease was exacerbated in the high VPD treatment. Even after partial recovery of C i due to stomatal opening, C i did not reach control values when approaching steady state ( Figure 3E; Table 1). In flacca leaves, C i decreased less strongly under both control and high VPD conditions, and did not increase much during the remainder of photosynthetic induction ( Figure 3F).  Also, C i time courses in both of these treatments were virtually indistinguishable in flacca, which is explained by the diminished reduction of g s under high VPD ( Figure 3D). Under low CO 2 , C i was similar in wildtype and flacca, displaying only small decreases in the beginning of photosynthetic induction without subsequent recovery ( Figures 3E, F). Intrinsic water use efficiency (WUE i ) was roughly twice as high in wildtype compared to flacca leaves ( Figures 3G, H). In the wildtype, a high VPD resulted in large increases (+31%), and a low CO 2 in large decreases (−56%), of WUE i relative to control conditions ( Figure 3G). In flacca, WUE i was markedly reduced (−50%) under low CO 2 relative to the other two conditions ( Figure 3H). While WUE i showed strong dynamics in the first 30 minutes after exposure to high irradiance under high VPD and control conditions in the wildtype, it plateaued early after an initial increase (<5 min) in all other cases ( Figures 3G, H). Leaves initially adapted to 50 µmol m −2 s −1 PAR were exposed to 1500 µmol m −2 s −1 PAR at time = 0 min. Data were logged at 400 µbar CO 2 and 7 mbar leaf-to-air VPD (Control), 200 instead of 400 µbar (Low CO 2 ) and 15 instead of 7 mbar (High VPD). Lines and symbols represent averages ± SEM, n = 5.

Chlorophyll Fluorescence Dynamics During Photosynthetic Induction
In wildtype leaves, an initial rapid increase in F PSII within the first 8 min was followed by a slower, more gradual increase towards a steady state under control and high VPD conditions until~40 min ( Figure 5A). In flacca leaves in control and high VPD conditions, a gradual decrease was observed after the initial increase in F PSII ( Figure 5B). In both genotypes, F PSII under low CO 2 stabilized quickly at lower values and then plateaued. Dark-adapted F v /F m was 0.82 in both genotypes (n.s., Figure 5B, inset). A rapid increase in NPQ in the first five minutes was followed by a decrease until 10 to 20 min, which was followed by a slower increase until the final measurement after 60 minutes ( Figures 5C, D). Under control and high VPD conditions, NPQ initially rose to much higher values in wildtype (1.6-1.7) compared to flacca leaves (1.4-1.5). The subsequent decrease to a local minimum showed a greater amplitude in wildtype (~0.1) compared to flacca leaves (~0.05). Under low CO 2 , NPQ tended to be greater in both genotypes compared to the other treatments. The coefficient of photochemical quenching (qP) showed dynamics similar to those of F PSII ( Figures  5E, F). qP and F PSII were highly correlated in all treatments (R 2 >0.99). The efficiency of open photosystem II traps (F v ′/F m ′) showed dynamics that were the inverse of those of NPQ ( Figures  5G, H); NPQ and F v ′/F m ′ were highly correlated (R 2 >0.99). These correlations suggest that F PSII dynamics were largely due to changes in qP rather than changes in F v ′/F m ′ or NPQ (Baker et al., 2007).

Stomatal Effects on Rate of Photosynthetic Induction
Next, we explored several ways to visualize the effects of (partially) closed stomata on photosynthetic induction. First, initial, low-irradiance adapted g s was plotted against t 90 ( Figure  6A). Across both genotypes, there was a consistent thresholdtype relationship between initial g s and t 90 : at initial g s <0.4 mol m −2 s −1 , t 90 increased strongly with decreases in initial g s , reaching values of~15 min at an initial g s of 0.11 mol m −2 s −1 . At initial g s >0.4 mol m −2 s −1 , the value of t 90 (~5 min) was unaffected by further increases in initial g s . Roughly, a similar threshold was visible between t 50 and initial g s , as initial g s <0.2 mol m −2 s −1 tended to increase t 50 , while t 50 was unaffected by differences in initial g s in the range 0.2 to 0.8 ( Figure 6A, inset). Initial g s versus t 50 or t 90 did not show a similar relationship at low CO 2 ( Figure S2) and was therefore omitted from Figure 6A. Photosynthesis integrated over the initial five minutes of photosynthetic induction scaled well with C i integrated over the same period ( Figure 6C). This suggests that A was affected by C i , while the change in C i was due to treatment and/or genotype effects on g s . Finally, plotting J vs. gross photosynthesis rates (A g , calculated as A plus R d ) produced very similar results between control and high VPD conditions in flacca leaves ( Figure 6D), while in wildtype leaves A g showed higher values for the same J in control compared to high VPD conditions ( Figure 6B). Under low CO 2 , both genotypes showed decreased A g for a given J. Also, while in flacca leaves the plots of A g vs. J were highly linear ( Figure 6D), in wildtype leaves they showed an upwards curvature at higher A g /J values, i.e., A g increased more strongly than J ( Figure 6B). This upwards curvature is indicative of an increase in the rate of carboxylation relative to the rate of oxygenation, most likely caused by an increase in C i due to increased stomatal opening (Kaiser et al., 2017a;Zhang et al., 2018).

DISCUSSION
In recent years, the dynamic responses of photosynthesis to fluctuating light, and their limitations, have received more attention. It is now widely recognised that increasing photosynthesis is a pathway to increasing crop productivity (Ort et al., 2015), that crops frequently encounter light intensity fluctuations , and that alleviating some of the limitations acting on photosynthesis transients can strongly increase biomass . Indeed, research looking into reductions of limitations acting on dynamic photosynthesis currently receives a lot of attention (Tanaka et al., 2019;Acevedo-Siaca et al., 2020;Kimura et al., 2020;Yamori et al., 2020).

Greater Stomatal Conductance Increases the Rate of Photosynthetic Induction
Responses of A to C i , as well as mesophyll conductance, were similar between the flacca mutant and the wildtype (Figures 1 and 2), while stomatal conductance was strongly enhanced in flacca leaves A B FIGURE 4 | Time to reach 50 (A, t 50 ) and 90% (B, t 90 ) full photosynthetic induction of wildtype and flacca leaves after a step change from low to high irradiance. Experimental conditions identical to those described under Figure 3. Different letters depict significant differences (P<0.05) between ranks, as determined by Dunn's test. Bars and error bars represent averages ± SEM, n = 5.
( Figures 3C, D). This confirms our hypothesis that the flacca mutant is a useful system for strongly reducing stomatal limitations and, via their reduction, better understanding their effects. This is similar to earlier reports using genotypes with "always-open" stomata (Tomimatsu and Tang, 2012;Tomimatsu et al., 2019;Kimura et al., 2020;Yamori et al., 2020). Our results further suggest that the faster photosynthetic induction observed in flacca compared to wildtype leaves under control and high VPD conditions ( Figures 3A, B and 4) was indeed due to much higher stomatal conductance (difference between genotypes in initial g s was~0 .4 mol m −2 s −1 ). Initial g s in leaves adapted to darkness or shade strongly impacts on rates of photosynthetic induction upon illumination with high irradiance. It has been reported previously that parameters such as the time needed to reach 50 or 90% of full photosynthetic induction (t 50 and t 90 , respectively) show a strong bimodal relationship with initial g s (Valladares et al., 1997;Allen and Pearcy, 2000;Kaiser et al., 2016), similar as shown in the present study ( Figure 6A).
The decrease in initial g s in the wildtype upon high VPD ( Figure 3C) translated into a marked decrease in C i during photosynthetic induction ( Figures 3E and 6C). This decrease was less strong in control conditions and was barely visible in flacca under control or high VPD ( Figure 3F), as initial g s in flacca did not react to high VPD ( Figure 3D). The reduced transient availability of C i decreased the rate of photosynthetic induction in the wildtype under high VPD ( Figure 3A). This reduction, in turn, fed back on dynamic F PSII (which was slowed down; first 15 minutes in Figure 5A), NPQ (which initially increased to higher levels and then relaxed less quickly than under control conditions; Figure 5C) and the relationship between gross photosynthesis and electron transport ( Figure 6B). Unlike initial g s , stomatal opening (difference between initial and final g s ) was not different between wildtype and flacca leaves under control (~0.4 mol m −2 s −1 ) and high VPD conditions (~0.3 mol m −2 s −1 ; Table 1). Perhaps surprisingly, photosynthetic induction was not different between genotypes under low CO 2 partial pressure (Figures 3 and 4). This may be explained in two ways: firstly, initial g s in the wildtype increased, from 0.22 mol m −2 s −1 at 400 µbar to 0.34 mol m −2 s −1 at 200 µbar, while that in flacca did not ( Table 1). This g s increase in wildtype leaves almost halved the difference in initial g s between genotypes (0.4 ! 0.24 mol m −2 s −1 ). Secondly, any positive effect that the remaining difference in initial g s may have had on photosynthetic induction in flacca was probably additionally decreased by low CO 2 availability. Initial g s in Leaves Lacking ABA Does Not React to Low CO 2 or High VPD A striking finding of the present study was that while g s in wildtype leaves (as expected) increased upon reductions in CO 2 partial pressure ( Figure 3C, Table 1), g s in flacca was virtually unchanged ( Figure 3D). This confirms that ABA is part of the CO 2 signalling pathway in stomatal regulation (reviewed in Engineer et al., 2016). Under high VPD (15 mbar), g s in flacca leaves adapted to low irradiance was similar to that at low VPD (7 mbar; Figure 3D). In wildtype leaves, stomata again responded as expected, by reducing their aperture under increased VPD ( Figure 3C). At high irradiance, however, g s in flacca did respond to the increase in VPD, as its value was reduced by~0.15 mol m −2 s −1 compared to that at 7 mbar ( Table 1). While there is ongoing controversy about the role of ABA in stomatal sensing of humidity, Merilo et al. (2018) showed that a number of genotypes that are either ABA deficient or ABA insensitive closed their stomata when exposed to an increase in VPD (at 150-500 µmol m −2 s −1 PAR). The authors explained this phenomenon (which is at variance with Mcadam et al., 2015;McAdam et al., 2016) as most likely being a hydropassive response, resulting from much higher initial g s in ABA mutants, and thus a greater drop in humidity in the substomatal cavity (Merilo et al., 2018).

A B D C
FIGURE 6 | Effects of stomatal conductance on photosynthetic induction. (A) relationship between stomatal conductance of leaves adapted to 50 µmol m −2 s −1 PAR (initial g s ) and time to reach 90% of full photosynthetic induction (t 90 ) after a switch to high irradiance, inset: relationship between initial g s and t 50 ; (B) relationship between electron transport rate (J) and gross photosynthesis rate (A g ) in the wildtype, arrows show sequence in which data were logged; (C) relationship between integrated net photosynthesis rate (∫A) and integrated leaf internal CO 2 partial pressure (∫C i ) during the first five minutes of photosynthetic induction, different letters indicate statistically significant (P<0.05) differences between treatments; (D) relationship between J and A g in flacca. Experimental conditions identical to those described under Figure 3. Symbols represent averages ± SEM, n = 5.

Limitations of the Study
In this study, we only used one mutant (flacca, LA0673) to examine the effects of open stomata on the rate of photosynthetic induction. Ideally, a larger number of mutants, each resulting in differential g s relative to its wildtype, should be used; this to make sure that the observed effects on photosynthetic induction were truly caused by g s rather than some other putative effects of the flacca mutation. Secondly, during growth flacca plants were regularly sprayed with ABA, following the recommendations of Imber and Tal (1970). It may be that this ABA application triggered unwanted responses in the plants, although based on all results presented here it seems that photosynthesis was fully functional in flacca.

CONCLUSIONS AND OUTLOOK
The current study suggests that in wildtype leaves, g s exerts a substantial limitation on non-steady state photosynthesis. The flacca mutant can partially overcome this limitation through stomata that remain open in low light, resulting in substantially reduced t 90 in ambient CO 2 partial pressure (t 90 was 36-46% in in flacca relative to wildtype values). Nevertheless, while flacca is a good system for testing (dynamic) stomatal limitation in the laboratory, breeding for a similar stomatal behaviour will not be useful for most crops (except possibly for production in wetland areas), as this improvement of photosynthesis in fluctuating light will come with a significant reduction in water use efficiency and a major fitness disadvantage. A more promising approach may be to improve stomatal responsiveness to light intensity fluctuations, as this can potentially increase both light and water use efficiency under fluctuating light intensities. Recent, promising examples of increased g s responsiveness are the BLINK1 transformant (Papanatsiou et al., 2019) and the PATROL1 overexpressor (Kimura et al., 2020).

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/ Supplementary Material.

AUTHOR CONTRIBUTIONS
EK and AM designed the study with input from all other authors. EK performed measurements and data analysis. EK wrote the manuscript with input from all other authors.