CO2 diffusion in tobacco: a link between mesophyll conductance and leaf anatomy

The partial pressure of CO2 at the sites of carboxylation within chloroplasts depends on the conductance to CO2 diffusion from intercellular airspace to the sites of carboxylation, termed mesophyll conductance (gm). We investigated how gm varies with leaf age and through a tobacco (Nicotiana tabacum) canopy by combining gas exchange and carbon isotope measurements using tunable diode laser spectroscopy. We combined these measurements with the anatomical characterization of leaves. CO2 assimilation rate, A, and gm decreased as leaves aged and moved lower in the canopy and were linearly correlated. This was accompanied by large anatomical changes including an increase in leaf thickness. Chloroplast surface area exposed to the intercellular airspace per unit leaf area (Sc) also decreased lower in the canopy. Older leaves had thicker mesophyll cell walls and gm was inversely proportional to cell wall thickness. We conclude that reduced gm of older leaves lower in the canopy was associated with a reduction in Sc and a thickening of mesophyll cell walls.


Introduction
In plants with the C 3 photosynthetic pathway, mesophyll conductance, g m , quantifies the ease with which CO 2 diffuses from intercellular airspace within a leaf to the sites of Rubisco carboxylation within chloroplasts [1,2]. It is one of the three main physiological processes limiting CO 2 uptake and fixation, the others being CO 2 diffusion from the atmosphere to the sub-stomatal cavity (stomatal conductance, g s ) and the biochemical activity of Rubisco and RuBP regeneration. Studies have shown that global crop production needs to double by 2050 to meet the projected demands from a rising population, diet shifts and increasing biofuels consumption [3]. The need to understand and maximize g m is part of the research efforts being made to enhance photosynthesis to improve crop yield. For example, enhancements of photosynthesis through manipulation of chloroplast function will be diminished through the reduction in chloroplast CO 2 partial pressure unless it is combined with improved g m [4][5][6][7][8][9]. Increasingly, crop models are incorporating leaf and canopy level parameters to better predict where photosynthesis improvements can be made (e.g. [10]), and an understanding of mesophyll conductance variation across leaf positions is crucial for this.
At present, there is an incomplete mechanistic understanding of g m (see Cousins et al. [11] for a review of recent developments in g m in C 3 and C 4 species). To enhance CO 2 diffusion, chloroplasts are spread thinly along cell wall surfaces, and their surface area appressing intercellular airspace is up to 20 times leaf surface area and good correlation between g m and S c (chloroplast surface area exposed to the intercellular airspace per unit leaf area) has been observed [12][13][14]. However, across species, this correlation is not unique and mesophyll cell wall thickness and its porosity, as well as membrane permeability to CO 2 and liquid diffusion, are also considered important parameters of g m [12,14,15]. Figure 1 outlines the CO 2 diffusion path within the leaf.
There have been a number of studies that have examined variation in g m with leaf age in tree species [14,[16][17][18] but fewer studies have looked at variation in g m with leaf age in crop species [19][20][21]. Here, we have examined the changes in CO 2 assimilation rate and g m with leaf age and canopy position in tobacco to assess the linkage between g m and leaf anatomy, and provide information on how best to model variation in g m in a C 3 crop canopy.

Plant growth
Tobacco (Nicotiana tabacum, L. cv Samsun) was grown in a naturally lit glasshouse with day/night temperatures set at 28/18°C. Seeds were sown in the glasshouse in commercial seed raising mix, then transferred after two weeks to 5 l pots filled with commercial potting mix supplemented with slow-release fertilizer (Osmocote Exact, Scotts, NSW, Australia). Plants were grown in Canberra, Australia between June and August 2019, with an average day length of 10 h. Average light intensity at midday during the growing period was 700 µmol m −2 s −1 . Plants were watered daily.

Concurrent measurements of gas exchange and
carbon isotope discrimination to quantify mesophyll conductance Gas exchange and carbon isotope discrimination measurements were made as described by Tazoe Figure 1. A Star Wars-inspired cartoon depicting the CO 2 diffusion journey inside the leaf with Princess Leia representing a CO 2 molecule (a) and the corresponding electron micrographs illustrating the actual CO 2 diffusion pathway within a tobacco leaf (b and c). For photosynthetic CO 2 assimilation to occur, CO 2 has to diffuse through stomata (entrance), intercellular airspace, cell wall, membranes ( plasma membrane and chloroplast envelope) and the liquid phase in the cytosol and chloroplast stroma. Once inside the chloroplast, CO 2 is then fixed by Rubisco into energy-rich sugars as part of the photosynthesis process. Bar in b = 10 µm; bar in c = 1 µm.

Anatomical measurements
Leaf anatomy was determined from light and scanning electron micrographs of transverse sections of resin-embedded leaf tissue collected from leaf positions 1 to 10 of three 9-week-old tobacco plants ( figure 3). Leaf tissue was collected from the same area used for the concurrent gas exchange and carbon isotope discrimination measurement and immediately processed for light and electron microscopy as previously described [27]. Light micrographs were obtained using a Leica DM5500 compound  Figure 2. Variation in CO 2 assimilation rate, mesophyll conductance, C i − C c , stomatal conductance and C a − C i over time with increasing leaf age. Gas exchange measurements were made at an irradiance of 1500 µmol m −2 s −1 , ambient CO 2 of 380 µbar, 2% O 2 and a leaf temperature of 25°C. Gas exchange measurements were made concurrently with measurements of carbon isotope discrimination using tunable diode laser spectroscopy for the calculation of mesophyll conductance (see Materials and methods), n = 4. Timepoints significantly different from 6-week-old plants are indicated with an asterisk. The same leaves were followed over time, and measured leaf is denoted with a yellow circle on plant images. Raw data are given in the electronic supplementary material, Data File S1.  royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200040 Table 1. Means comparison using Tukey test of the physiological and anatomical parameters measured between different leaf positions of tobacco plant. The number 1 denotes significant difference at p < 0.05 and 0 denotes no significant difference at p < 0.05.
leaves compared royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200040 Table 1. microscope (Leica Microsystems) and used to measure leaf thickness and mesophyll layer thickness. For scanning electron microscopy, ultrathin sections were mounted onto pieces of silicon wafer and post-stained with aqueous uranyl acetate followed by lead citrate for 10 min each prior to imaging under a Zeiss UltraPlus field emission scanning electron microscope at 2 kV. Scanning electron micrographs were used to measure mesophyll cell wall thickness, chloroplast length, chloroplast thickness, the surface area of mesophyll cells exposed to intercellular airspace per unit leaf area (S mes ) and surface area of chloroplasts exposed to intercellular airspace per unit leaf area (S c ). Electron micrograph measurements were performed according to Evans et al. [13] using at least 600 µm leaf surface length for each leaf position per plant. Stomatal density was measured from positives made with nail polish from hydrophilic vinyl polysiloxane impressions of the abaxial surface of the leaf area measured by gas exchange, and viewed under a Leica confocal microscope and Leica DC500 camera. All measurements were made using ImageJ software (National Institutes of Health) and a Wacom Cintiq graphics tablet (Wacom Technology).

Statistical analysis
Statistical analyses were performed using Student's t-test for figure 2, and all other data using one-way analysis of variance. Means comparison was made at 0.05 significance level using the Tukey test (OriginPro 2020, OriginLab Corporation).

Changes in leaf physiology over time
To understand how mesophyll conductance, g m , is influenced by leaf ageing, leaf physiology traits were repeatedly measured in a tobacco leaf over four weeks of growth. CO 2 assimilation rate, A, stomatal conductance, g s , and g m all decreased in the ageing tobacco leaf over time ( figure 2). The drawdown of CO 2 into the chloroplast (C i − C c ), however, was not significantly changed over the four weeks of measurements. The drawdown of CO 2 into the sub-stomatal cavity (C a − C i ) increased between 6 and 8-week-old leaves, but was not significantly different overall between the 6-week and 9-week measurements (figure 2). Raw data for figure 2 are given in the electronic supplementary material, Data File S1.

The effect of canopy position on leaf anatomy and physiology
The effect of leaf canopy position was further investigated in 9-week-old tobacco plants with 10 leaves through measurement of physiological and anatomical traits. To determine whether there is a direct correlation between the leaf anatomy and g m , structural and ultrastructural analyses were performed using the leaf tissue from the same area as used for physiological measurement. Light micrographs of transverse leaf sections showed that leaf thickness and mesophyll cell layer thickness increase as the leaf matures (figures 3 and 4b,c). Surprisingly this was not accompanied by a significant increase in leaf mass per area (LMA, figure 4a), which varied from 24.8 ± 0.6 g m −2 in leaf 1 to 32.1 ± 2.1 g m −2 in leaf 9. Measurements performed using electron micrographs of the same leaf sections revealed that chloroplast surface area exposed to the intercellular airspace per unit leaf area (S c , figure 4d) and mesophyll surface area exposed to the  table 1). Mesophyll conductance decreased strongly from leaf 6 onwards. Stomatal conductance also showed a similar decrease in the canopy (figure 5d). The variation in the drawdown of CO 2 from the atmosphere into the sub-stomatal cavity (C a − C i , figure 5e) and from the sub-stomatal cavity into the chloroplasts (C i − C c , figure 5c) was less dynamic, but significant differences were still apparent between leaves at the top and bottom of the canopy (table 1). The number of stomata per mm 2 leaf area (stomatal density, figure 5f ) decreased moving down the canopy due to the overall expansion of the pavement cells on the leaf surface as the leaves aged (electronic supplementary material, figure S1). There was a strong correlation between stomatal conductance and stomatal density (y = 0.0034x + 0.092, R 2 = 0.95).
Leaf nitrogen (N) content was highest in the youngest leaves at the top of the canopy, and steadily declined through the lower leaf positions (figure 6a). Rubisco content also decreased down the canopy (figure 6b) and was found to be very closely correlated to leaf nitrogen content (figure 6c, R 2 = 0.99). There was a strong linear relationship between g m and A (y = 0.0126x + 0.142, R 2 = 0.71, electronic supplementary material, figure S2). However, the relationships between g m and Rubisco content or leaf N were best fitted with a second-order polynomial (electronic supplementary material, figure S2) as were the relationships between A and Rubisco content and leaf N (electronic supplementary material, figure S3). This shows that A per Rubisco or leaf N was less in leaves at the top of the canopy than in older leaves further down the canopy.
Electron micrographs also allowed measurements of mesophyll chloroplast thickness, mesophyll chloroplast length and mesophyll cell wall thickness in leaf positions 1-10 ( figure 7). Results showed a decrease in mesophyll chloroplast thickness (figure 8a), an increase in mesophyll chloroplast length (figure 8b) and thickening of mesophyll cell walls (figure 8c) as the leaf matures. Measurements and calculations also revealed that the thicker mesophyll cell wall in older leaves (figure 8c) was directly proportional to the mesophyll resistance (r m = 1/g m , figure 9a). This resistance ranged from 1.5 in young leaves to 4 m 2 s bar mol −1 in lower canopy leaves and the inverse g m ranged from 0.7 to 0.2 mol m −2 s −1 bar −1 . There was only a weak correlation between mesophyll conductance and the chloroplast surface area exposed to intercellular airspace (S c , figure 9b).
4. Discussion 4.1. Mesophyll conductance correlates with CO 2 assimilation rate CO 2 assimilation rate, A, and mesophyll conductance, g m , both decreased as leaves moved lower in the tobacco canopy, with a strong linear correlation between A and g m reflected in a constant difference in CO 2 partial pressure between intercellular airspace and the chloroplasts (C i − C c ) across leaf positions (figure 5, and electronic supplementary material, figure S2). Strong correlations between A and g m have previously been observed with leaf development or ageing [14,16,19,20,28,29]. In a study with transgenic tobacco plants where Rubisco content was reduced, the correlation did not hold and A declined more than g m [13]. This indicates that it is not a mechanistic relationship between A and g m . Both A and g m showed curvilinear responses to Rubisco and leaf nitrogen (electronic supplementary material, figures S2 and S3) suggesting that the link between A and g m is the most useful for the incorporation of g m in canopy photosynthesis and crop models [10,[30][31][32].

Relationship between CO 2 assimilation rate and Rubisco and leaf N content
Rubisco accounts for around 40% of soluble protein in a leaf and around 20% of leaf nitrogen (N) investment in C 3 species [33]. We saw a very close correlation between leaf N and Rubisco, suggesting that the proportional investment in Rubisco across leaves remained constant across the canopy. We also observed curvilinear relationships between A and Rubisco content or leaf N (electronic supplementary material, figure S3) with young leaves having a lower Rubisco and leaf N use efficiency. Curvilinear responses have been observed in the past and two explanations have been put forward [34]. royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200040 The first one suggests it could be a CO 2 diffusion limitation, but since we know that there has been no increase in either C a − C i or C i − C c we can rule out a CO 2 diffusion limitation. The second reason given by Evans [34] is that with increasing leaf N the chlorophyll and electron transport capacity increase and to reach light saturation requires progressively higher irradiances. Since we measured at the same irradiance the maximal rate of the high nitrogen leaf is underestimated. We only made measurements of A at one irradiance and this could be further investigated.

Linking mesophyll conductance and leaf anatomy
Figures 3, 4 and 7-9 highlight the large anatomical changes that occur during leaf development. There was a doubling of leaf thickness but this did not result in significant changes in leaf mass per area (LMA) despite the increase in cell wall thickness. In an across species comparison higher LMA has been associated with decreased g m and an increase in the percentage that cell wall mass contributes to LMA [35]. However, tobacco, as a herbaceous crop species, has a low LMA compared to species tested in that study, and cell wall mass is expected to contribute no more than 10-15% of LMA (figure 3; [35]). It is therefore not surprising that we saw little change in LMA despite an increase in cell wall thickness and reduction in g m . We conclude that increased airspace and cell volume account for these changes.
To facilitate CO 2 diffusion, chloroplasts line mesophyll cell walls adjacent to intercellular airspace. Anatomical parameters such as S c and S mes are similar to those measured in glasshouse grown tobacco in previous studies [13] providing a chloroplast surface area 15 times greater than the projected leaf area. As chloroplasts cover 80% or more of mesophyll cell walls adjacent to intercellular airspace, there is room for only small improvements of g m by increasing S c /S mes . However, S c /S mes has been shown to be less in low light conditions and under stress in Populus tremula, and can vary with growth conditions [14]. Compared to the rapid decline in CO 2 assimilation rate with leaf position, S c remained unchanged up to leaf position 7 and is therefore not the driver for the reduction in CO 2 assimilation rates or g m (figure 9b). Other studies that have also compared changes in A, g m and S c with leaf age variation also reported the poor correlation between g m and S c under these conditions [16,19,20].
As depicted in figure 1 CO 2 has to diffuse from intercellular airspace across the cell wall, the plasma membrane, cytosol, chloroplast envelope and the chloroplast stroma. Of these obstacles in the diffusion path, cell wall thickness and chloroplast shape are measurable components. We observed a continuous increase in cell wall thickness from the top to the bottom of the canopy and observed a strong inverse correlation between g m and cell wall thickness (figure 9a). This highlights the important contribution of cell wall thickness in determining g m and is in line with previous studies [12,15,36,37]. Tobacco like many crop species has relatively thin cell walls [36]; however, it is not just the physical dimension but also cell wall composition that matters and we know little about this at present (see [38] for a review of current knowledge). If we extrapolate to zero cell wall thickness we see a resistance of 0.6 m 2 s bar mol −1 (figure 9a) equating to a mesophyll conductance of 1.6 mol m −2 s −1 bar −1 . While we cannot accurately measure the resistance derived from plant membranes, we know that membrane composition (such as the presence of channels for facilitating CO 2 transfer) together with CO 2 diffusion through the liquid phases also contributes to mesophyll resistance [39]. Nevertheless, our results highlight that thinner cell walls, if structurally possible, could be of benefit to improve mesophyll conductance. royalsocietypublishing.org/journal/rsfs Interface Focus 11: 20200040

Conclusion
We combined gas exchange and anatomical measurements to assess how CO 2 assimilation rate, A, and mesophyll conductance, g m , decreased as leaves aged and whether we could link decreases in g m with leaf anatomy. Surprisingly we observed a decrease in the chloroplast surface area exposed to the intercellular airspace per unit leaf area, S c , only low in the canopy whereas there was a gradual increase in cell wall thickness and an inverse correlation between g m and cell wall thickness. We conclude that reduced g m of older leaves lower in the canopy was associated with a reduction in S c and a thickening of mesophyll cell walls. The relationship between A and g m , however, is the most useful for the incorporation of g m in canopy photosynthesis and crop models. In crop species where mesophyll cell chloroplast cover is high, increasing the conductance across the chloroplast interface is the major challenge for improvements in mesophyll conductance, which will enhance photosynthetic capacity and ultimately increase crop yields.   Figure 9. The relationship between mesophyll resistance (r m = 1/g m ) and mesophyll cell wall thickness (a) and between mesophyll conductance and S c , the chloroplast surface area exposed to intercellular airspace per unit leaf area (b). Each data point corresponds to an average value per leaf from three plants. Raw data are given in the electronic supplementary material, Data File S2.