PHOTOSYNTHETIC TRAITS OF CANOPY LEAVES OF DINIZIA EXCELSA (FABACEAE)

ABSTRACT The response of leaf traits to irradiance and [CO2] in canopy leaves of several tall trees remains to be determined under natural conditions. Thus, the objective of this work was to determine gas-exchange parameters in sun and shade leaves of Dinizia excelsa Ducke in 35-45 m tall trees of Central Amazonia. We assessed light saturated photosynthesis (Amax), stomatal conductance (gs), mesophyll conductance (gm), transpiration rates (E), water use efficiency (WUE), intrinsic water use efficiency (WUEi), maximum electron transport rate (Jmax), the maximum carboxylation rate of Rubisco (Vcmax), intercellular CO2 concentration (Ci)specific leaf area (SLA) and fresh leaf thickness. We also estimated the CO2 concentration at the chloroplast level (Cc) and determined the light and CO2 saturated photosynthesis (Apot). Amax was obtained at light saturation (1200 µmol m-2 s-1), whereas Apot, Vcmax, Jmax and gm were obtained after constructing A/Ci response curves. There was a significant difference between sun and shade leaves in Ci and Cc, but for other parameters no differences were observed. Amax was positively correlated with gs, gm and E, and there was also a significant correlation between gs and gm (p ≤ 0.05), as well as between Jmax and Vcmax. Thicker leaves had higher values of Amax, gs, Ci, Cc and E. Apot was limited by the electron transport rate and by low gm. The canopy of the tree caused a decrease in irradiance (30-40%), but this reduction was not enough to reduce important photosynthetic parameters. Thus, all resources allocated to leaf production led to maximum use of the solar energy received by the leaves, which allowed this species to grow at fairly rapid rates.


1.INTRODUCTION
In the last decades the eff ect of the increase of atmospheric CO 2 concentration on plant physiology has been intensively investigated due to its presumed consequences on climate changes (Manter and Kerrigan, 2004;Knauer et al., 2019). It is well-known that stomata play an important role on carbon uptake, as they impose the fi rst major diff usional limitation to CO 2 diff usion, from the atmosphere to carboxylation sites (Nascimento and Marenco, 2013;Xiong et al., 2018). The photosynthetic capacity of a leaf depends on diff usive (i.e. stomatal conductance -g s and mesophyll conductance -g m ) and non-diff usive factors (i.e. the maximum carboxylation rate of Rubisco -V cmax and maximum electron transport rate, J max ). V cmax and J max can be obtained from response curves of photosynthesis to intercellular CO 2 concentration -A/C i (Farquhar et al., 1980;Stinziano et al., 2019). Both V cmax and J max are important parameters in modeling studies that aim to predict the impacts of climate change on plant functioning (Knauer et al., 2019).
Among the factors that aff ect photosynthesis, the response of stomatal functioning to environmental and endogenous factors has been the subject of much research (Terashima et al., 2011;Mendes and Marenco, 2014;Marenco et al., 2017). Although g m is also quite relevant to CO 2 diff usion and photosynthesis, it attracted less attention in the past decades, but in recent studies, the importance of g m to carbon assimilation has been highlighted (Flexas et al., 2016;Knauer et al., 2019). Nevertheless, research is still needed to assess the relevance of g m in tropical species, particularly the Amazonian species.
The strata of tropical forests, as well as gap opening and closing lead to the formation of a light gradient in the canopy of a tree. This gradient can lead to changes in the strategy of the leaf for light capture and light use effi ciency, and thereby to variations in leaf structure and physiology (Givnish, 1984;Clark and Clark, 1992;Marenco et al., 2017). In fact, light is probably the most important environmental factor aff ecting plant establishment, growth and survival (Niinemets et al., 2015;Gitelson et al., 2017).
The angelim (Dinizia excelsa Ducke, Fabaceae) is an emerging tree that can reach 50-60 m in height and 1.0 to 1.8 m in diameter (Lorenzi, 1992). It has a dense wood -0.91 g cm -3 (Fearnside, 1997), and trees over 20 cm in diameter have a growth rate of 5.30 mm per year (Schwartz et al., 2016). In the emergent phase, the angelim canopy stands out in the forest landscape; hence it is not shaded by neighboring trees. It has wide distribution in the Brazilian Amazon, can occur in density of one individual per 6 ha, and the trees have good silvicultural performance (Dick, 2001;Ferreira et al., 2004). Angelim is economically important in silvicultural systems and in the timber industry. The wood of angelim is resistant to the attack of fungi and termites, and it is used in the manufacture of decorative plates, construction, shipbuilding, woodworking, and carpentry (Melo and Varela, 2006;Oliveira et al., 2008). In addition, the angelim can also be used in garden and urban aff orestation (Lorenzi, 1992).
Gas exchange studies of Amazonian tree species under natural conditions are relevant due to current trends in global warming and atmospheric CO 2 enrichment (Magrin et al., 2014). Thus, measurements of gas exchange parameters (e.g. g s , A max , V cmax and J max ) in Amazonian species is of paramount importance for the construction of models aimed at predicting the performance of Amazonian ecosystems. Thus, the objective of this work was to determine the eff ect of leaf position in the canopy stratum on leaf traits of angelim trees in the Central Amazon. Leaf trait measurements included light saturated photosynthesis (A max ), g s , g m , transpiration (E), water use effi ciency (WUE), intrinsic water use effi ciency (WUEi), J max , V cmax , intercellular CO 2 concentration (C i ), specifi c leaf area (SLA) and fresh leaf thickness (FLT). It was also estimated the concentration of CO 2 at chloroplast level (C c ) and determined light and CO 2 saturated photosynthesis (A pot ).

2.MATERIALS AND METHODS
The study was conducted at a 10-ha plot of a dense terra-fi rme forest fragment at the Colosso Reserve (02° 24' 13.2"S, 59°51' 54"W). The region has a humid equatorial climate, with annual precipitation of 2240 mm, distributed over a rainy season from November to May (> 180 mm per month) and a mild dry season from June to September (≤ 100 mm per month). October is a transitional month. The average annual air temperature is 26.7°C, and the average relative humidity is 84%.
In this study, six trees (35-45 m in height and 1.30-2.00 m in diameter) of angelim (Dinizia excelsa Ducke, Fabaceae) were used. In the experiment we used a completely randomized design with two treatments and six replications (trees). The treatments were the leaves from two positions in the canopy: upper part of the tree (hereinafter referred to as sun leaves, which were under direct solar radiation) and lower part of the canopy (shade leaves), which received about 60-70% of total solar radiation. From these canopy strata branches were detached for data collection. They were about 7 to 12 cm in diameter and approximately 12 m in length. All branches were above the canopy of neighboring trees and the lower branches of a tree were shaded only by the upper ones. We used detached branches because of the impossibility of accessing the canopy of the tree for data collection.
To assess the eff ect of branch detachment on stomatal conductance (g s ), a previous study was carried out using two gas exchange systems (Li-6400, Li-Cor, Lincoln, USA), collecting data simultaneously from two small branches (about 4 cm in diameter) from the same tree, which was accessible from a 40-m observation tower. After a stabilization period of 15 minutes, one branch was randomly selected and kept intact while the other branch was severed from the tree by a rapid cut, without interrupting gas-exchange measurements. The measurement process on both leaves (i.e. from the undisturbed branch and the severed one) was uninterrupted for 60 min, and during this time stomatal conductance remained similar in both branches. Thus, in the current experiment gas exchange and fl uorescence data were measured within that time interval, and for further precaution data were collected on branches thicker than 4 cm in diameter.
Gas exchange and fl uorescence measurements were made with a gas exchange system (Li-6400, Li-Cor, Lincoln, NE) with a 2-cm 2 integrated fl uorescence chamber head (Li-6400-40, Li-Cor). Just after detachment, the branch was taken to the gas exchange instrument for data collection. The time lag from branch detachment up to the leafl et insertion in the leaf chamber of the gas exchange system was about 10 minutes. After a stabilization period (5-10 min) at ambient CO 2 (380 µmol mol -1 ), photosynthetically active radiation (PAR) of 1200 μmol m -2 s -1 (light saturation) and ambient temperature (27 °C), photosynthetic rates (A) as a function of the intecellular CO 2 concentration -C i (A/ C i response curves) were measured. The A/C i response curves were generated by increasing the reference CO 2 concentration from 50 to 2000 µmol mol -1 in nine steps, i.e. 380, 250, 100, 50, 380, 550, 1000, 1,500 and 2,000 µmol mol -1 (Long and Bernacchi, 2003). Light saturated photosynthesis (A max ) was determined at light saturation (a value determined in a previous experiment) and a CO 2 concentration of 380 μmol mol -1 . Whereas the light and CO 2 saturated photosynthesis (hereinafter termed potential photosynthesis -A pot ) was measured at light saturation and a reference CO 2 concentration of 2000 μmol mol -1 . The maximum carboxylation rate of Rubisco (V cmax ) and maximum electron transport rate (J max ) were calculated according to Farquhar et al. (1980).
Where A c and A j denote the net photosynthetic rates limited by Rubisco activity and electron transport rate (A j ), respectively. * represents the CO 2 compensation point in the absence of respiration in the light; C i is the intercellular CO 2 concentration, O represents the oxygen concentration in the intercellular spaces; K c and K o represent the Michaelis-Menten constant of Rubisco for carboxylation and oxygenation, respectively. V cmax values were standardized to 25 ° C (Medlyn et al., 1999).
The quantum yield of the photosystem II system (Φ PSII ) was calculated as previously described (Nascimento and Marenco, 2013). C c values were determined according to Epron et al. (1995), whereas g m was obtained as follows: g m = A/(C i -C c ). We also calculated, water use effi ciency (WUE) the photosynthesis to transpiration ratio (A/E) and the intrinsic water use effi ciency (WUEi) the A/g s ratio. Data were collected from September to November 2010 from two healthy and physiologically mature leafl ets per stratum, and from six trees and two canopy strata per tree.
Data were subjected to analysis of variance (ANOVA) and the t test was used to determine signifi cant diff erences between sun and shade leafl ets. The relationships between quantitative variables (e.g. A max versus V cmax , A pot versus J max ) were examined by regression analysis. Pearson's correlation coeffi cient was used to evaluate the relationships between gas exchange variables and morphological traits (SLA and leafl et thickness). Statistical analyzes were performed using Statistica 10.0 software (StatSoft, Tulsa, OK, USA).

3.RESULTS
There was no signifi cant diff erence between sun and shade leafl ets in A max , A pot , g s , g m , V max , J max , VPD, T leaf , WUE, WUEi SPAD values, and chlorophyll contents (p ≥ 0.05; Table 1). Only C i and C c showed diff erences between canopy strata, i.e. sun versus shade leafl et (p ≤ 0.05; Table 1). On average, the A max and A pot values were 8.7 and 24.1 µmol m -2 s -1 for the sun leafl ets, whereas shade leafl ets had A max and A pot values of 5.9 and 19.3 µmol m -2 s -1 , respectively (Table 1). In ambient [CO 2 ], the g m values were 46% (sun leafl ets) and 29% (shade leafl ets) lower than those recorded for g s ( Table 1). The maximum quantum efficiency of PS II (F v /F m ratio) did not vary between sun and shade environment. There was a signifi cant eff ect of the ambient condition on FLT (p ≤ 0.05; Table 1). Mean FLT values were 0.20 and 0.16 mm for the sunny and shade conditions, respectively. However, SLA showed no signifi cant diff erences between the sun and shade environment (p ≥ 0.05; Table 1). As the ambient condition had no eff ect on gas exchange parameters (A max , A pot , g s and g m ), data from both environments were pooled for correlation analysis.
A max was positively correlated with g s (r = 0.95, Table 2), g m (r = 0.96), and E (r = 0.92). However, A max negatively correlated with VPD (r = -0.52), T leaf (r = -0.32) and WUE (r = -0.28). There was a positive correlation between g s and g m (r = 0.83, Table 2), and, it was also found that g s decreased with increasing in VPD and T leaf (r = -0.65 and r = -0.46; respectively). However, there was a weak correlation between g m and VPD (r = -0.37) and T leaf (r = -0.20), and between WUE and g s (r = -0.30) and g m (r = -0.155). WUE and WUEi were positively correlated with the g m /g s ratio (r = 0.99; Fig. 1).
At ambient CO 2 concentration, C i and C c were positively correlated with g s (r = 0.78) and g m (r = 0.42, Table 2). There was also a positive correlation between J max and V cmax (r = 0.81), and between V cmax (and J max ) and A max and A pot (Table 2).
V cmax and J max were positively correlated with g s (r  0.6) and g m (r  0.9, Table 2). Gas exchange parameters were correlated with SLA and FLT (Table 2). Thicker leaves had higher values of A max (r = 0.34), g s (r = 0.46), C i (r = 0.52), C c (r = 0.52) and E (r = 0.38). SLA was negatively correlated with VPD, T leaf ,V cmax and J max (p ≤ 0.05). On the other hand, FLT had no effect on g m , V cmax and J max (p ≥ 0.05). Finally, SLA increased with increasing LFT, which was not expected (r = 0.58, p ≤ 0.001). In Figure 2, irrespective of the environment (sun -shade), the initial portion of the curve showed a linear association between photosynthesis (A) and C i and C c indicating a limitation imposed by Rubisco carboxylation rate. The increase of both C i and C c led A to increase to the point of photosynthesis limitation by electron transport rate (dashed line in Figure 2). Below the colimitation point (C i values of 433-544 μmol mol -1 , indicated by the vertical line in Figure  2 A, B), A was limited by Rubisco carboxylation rates. The C c values at the colimitation point were 201-252 μmol mol -1 (indicated by the vertical line in Figure 2 C, D). On the other hand, above the colimitation point (high CO 2 concentration) A was limited by electron transport rates (dashed lines in Figure 2).
The g m values varied as a function of C i or C c (Fig. 3). At low CO 2 concentrations (C i ≤ 400 µmol mol -1 ), g m increased linearly with increasing C i or C c (Fig. 3). On the contrary, g m decreased at high CO 2 concentrations (Fig. 3). In comparison with g m values recorded at ambient [CO 2 ],(C i of 200-300 mol mol -'  ), g m decreased 70% at high CO 2 concentration (Fig.  3).

4.DISCUSSION
The mean F v /F m value recorded in this study (0.78; Table 1) is within the range of values for non--stressed leaves, i.e. F v /F m of about 0.80 (Björkman and Demmig, 1987). Very low F v /F m values, e.g. 0.60 or lower (Magalhães et al., 2009) often indicate the occurrence of photoinhibition. Therefore, it is concluded that in this study there was no photoinhibition of photosynthesis in D. excelsa.
The A max values observed in this study are similar to those found in other studies, such as D. excelsa (Miranda et al., 2005), Minquartia guianensis, Coussapoa orthoneura and Protium opacum (Magalhães et al., 2009;Marenco et al., 2014). The chloroplast CO 2 concentration (C c ) was around 50% lower than the concentration observed in the intercellular spaces (Table 1). Taking C i as the base line, the decline in C c is due to the barriers against CO 2 diff usion, from the intercellular space to the carboxylation site in the chloroplast (Niinemets et al., 2015; Tosens and Laa-  nisto, 2018). The greatest resistance to internal diff usion of CO 2 seems to be related to plasma membranes and chloroplast membranes (Warren, 2009;Peguero--Pina et al., 2017).
At the ambient CO 2 concentration (380 µmol mol -1 ) there was a close relationship between A max and g s and g m ( Table 2). This occurs because both conductances determine the CO 2 fl ux to the carboxylation site in the chloroplast (Flexas et al., 2013). Stomatal and mesophyll resistance to CO 2 fl ux account for about 40% of photosynthesis limitation in well-irrigated plants (Yamori et al., 2006;Tosens et al., 2016). Thus, the strong correlation between g s and g m indi- Os numerais (nos painéis) mostram os valores para C i e C c em μmol mol -1 no ponto de colimitação. Cada símbolo representa a média de seis árvores e dois folhetos (de duas folhas) por árvore.
cates a coupling between these parameters. The high correlation between A max and g s (and g m ) suggests that diff usive factors were of paramount importance for determining carbon assimilation in D. excelsa. The decline of g s under high VPD conditions corroborates that some ambient conditions can lead to a reduction in photosynthetic rates. Park and Furukawa (1999) showed that photosynthesis and stomatal conductance measured in tropical trees decreased due to increased VPD. Stomatal response to air humidity, temperature and VPD has been studied for decades, and it is known that stomata respond to changes in leaf tissue water content or to variations in guard cell water potential (Buckley, 2019).
Transpiration (E) was strongly correlated with g s , which is expected as the stomata play a key role in the control of leaf transpiration. Although g m is expected to have little impact on leaf transpiration (Ouyang et al., 2017), in this study we found a positive correlation between g s and g m , which helps to explain the positive correlation between g m and WUEi, as reported by Jahan et al. (2014). A high g m /g s ratio contributes to increased water use effi ciency, particularly under water stress. Thus, it has been suggested that plants with potential for acclimation to drought have a high g m / g s ratio (Giuliani et al., 2013). The strong relationship observed between the g m /g s ratio and WUE and WUEi found in this study supports the hypothesis of coupling between g s and g m . These results suggest that in Dinizia excelsa water effi ciency can be increased by improving g m .
Besides g s , g m also aff ects the photosynthetic capacity of the leaf, as shown by the close correlation between V cmax , J max , and g m (Table 2). Indeed, V cmax and J max were more infl uenced by g m than by g s . In Figure  2, one can see that Rubisco carboxylation rate was the most limiting factor for photosynthesis (up to the CO 2 concentration indicated by the arrows), which is in agreement with Sage and Kubien (2007) and Mendes et al. (2017). At higher CO 2 concentrations, however, the electron transport rate became the most limiting factor a closer correlation between A pot and J max was observed, which is in agreement with Mendes et al., (2017).
It was observed that g m showed a rapid response (minutes) to changes in CO 2 concentration. However, it is unknown whether this response is a pattern also shared by other Amazonian tree species. At C i values of 0-400 μmol mol -1 (C c < 200 μmol mol -1 ), g m increased linearly with increasing CO 2 until it reached a maximum value (at C c ≈ 300 μmol mol -1 ). On the other hand, at high C i values g m decreased with increasing CO 2 concentration, which suggests that at that CO 2 condition photosynthesis is no longer limited by the availability of CO 2 at the intercellular spaces but by mesophyll resistance and electron transport rates.
Leaf thickness is one of the key leaf traits that aff ect g m (Terashima et al., 2011). However, the signifi cance of this eff ect was not detected in D. excelsa probably because there was little variation in FLT (0.16-0.21 mm). However, it is worth noting that FLT positively aff ected g s , C i and C c . This suggests that in thicker leaves there was an increase in the volume of intercellular space, perhaps as a mechanism to maximize mesophyll CO 2 concentration. A large intercellular space can contribute to reduce the eff ect CO 2 limitation under partial stomatal closure (Shao et al., 2008). SLA values recorded at the lower part of the canopy (shade leafl ets, Table 2) were lower than those reported for understory trees (12 and 22 m 2 kg -1 ) in the central Amazon (Mendes et al., 2013), which indicates that even leaves of the innermost part of the canopy were receiving relatively high levels of solar radiation. An increase in SLA is often related to a decrease in leaf thickness. Thus, the results presented in this study diff er from the classical pattern that shows a negative relationship between SLA and leaf thickness (Niinemets, 1999). This discrepancy can be attributed to the fact that in this study leaf thickness was measured in fresh leaves, and variation in leaf water contents may lead to divergence with the most common SLA-leaf thickness relationship.

5.CONCLUSIONS
The canopy leaves of angelim causes a decrease in irradiance (30-40%), which does not appear to be high enough to negatively aff ect important gas exchange parameters, such as V cmax and J max . This is quite important for the carbon economy of the tree, as it allows a maximum use of the solar energy received by the leaves. This ultimately contributes to enhance carbon uptake. As a result this species has a fairly fast growth rate, even when it produces wood of high density. On the other hand, an open canopy can lead to high transpiration rate, which might have a negative impact on photosynthesis if drought periods become longer as predicted by climate models.
In the study it is shown that the mesophyll resistance plays is an important role in CO 2 diff usion. Therefore, it is recommended whenever possible to include g m values in gas exchange calculations to obtain more accurate values of photosynthetic parameters. At low [CO 2 ] the photosynthetic rates were limited by Rubisco carboxylation rate, but at high [CO 2 ] photosynthesis was limited by both g m and the electron transport rate. Changes in CO 2 concentrations have an eff ect on g m , which has its maximum at C c values of about 300 μmol mol -1 . However, a substantial increase in atmospheric CO 2 concentration may lead to an increase (> 50%) of the mesophyll resistance. The results presented in this study may be useful in the construction of climate models that aim to predict the eff ects of global climate change on the Amazon ecosystem.