Reassimilation of Leaf Internal CO 2 Contributes to Isoprene Emission in the Neotropical Species Inga edulis Mart.

: Isoprene (C 5 H 8 ) is a hydrocarbon gas emitted by many tree species and has been shown to protect photosynthesis under abiotic stress. Under optimal conditions for photosynthesis, ~70%–90% of carbon used for isoprene biosynthesis is produced from recently assimilated atmospheric CO 2 . While the contribution of alternative carbon sources that increase with leaf temperature and other stresses have been demonstrated, uncertainties remain regarding the biochemical source(s) of isoprene carbon. In this study, we investigated leaf isoprene emissions ( Is ) from neotropical species Inga edulis Mart. as a function of light and temperature under ambient (450 µ mol m − 2 s − 1 ) and CO 2 -free (0 µ mol m − 2 s − 1 ) atmosphere. Is under CO 2 -free atmosphere showed light-dependent emission patterns similar to those observed under ambient CO 2 , but with lower light saturation point. Leaves treated with the photosynthesis inhibitor DCMU (3-(3,4-dichlorophenyl)-1,1-dimethylurea) failed to produce detectable Is in normal light under a CO 2 -free atmosphere. While strong temperature-dependent Is were observed under CO 2 -free atmosphere in the light, dark conditions failed to produce detectable Is even at the highest temperatures studied (40 ◦ C). Treatment of leaves with 13 C-labeled sodium bicarbonate under CO 2 -free atmosphere resulted in Is with over 50% containing at least one 13 C atom. Is under CO 2 -free atmosphere and standard conditions of light and leaf temperature represented 19% ± 7% of emissions under ambient CO 2 . The results show that the reassimilation of leaf internal CO 2 contributes to Is in the neotropical species I. edulis . Through the consumption of excess photosynthetic energy, our results support a role of isoprene biosynthesis, together with photorespiration, as a key tolerance mechanism against high temperature and high light in the tropics.


Introduction
Isoprene (2-methyl-1,3-butadiene, C 5 H 8 ) is a reactive hydrocarbon gas emitted in large amounts to the atmosphere by many plants [1,2]. Approximately 500 Tg C year −1 is released as isoprene by vegetation [3]. Tropical rainforests are an important source of isoprene to the atmosphere and estimates suggest that they are responsible for 80% of global isoprene emissions (Is) [4]. Isoprene has an tolerance mechanism under plant stress that leads to a decrease in stomatal conductance and CO 2 uptake. Under these conditions, isoprene could still be synthesized using the CO 2 released by photorespiration, offering a protective mechanism by consuming excess photosynthetic energy and reducing equivalents. These results deepen our understanding about Is by tropical species under different environmental conditions commonly experienced in the tropics, such as high irradiance and temperature, and contributes to the modeling of Is from terrestrial ecosystems in the tropics under climate change.

Plant Material and Experimental Design
Six I. edulis trees with a height ranging from 5 to 10 m were used in this study. The experimental trees occur naturally near the Laboratory of Plant Physiology and Biochemistry at the National Institute for Amazonian Research (INPA) campus III in Manaus, Brazil. This tropical species was selected because of its high reported Is and for its ability to maintain high transpiration rates for many hours following branch detachment from the tree [25]. From October 2014 to February 2016 Pn and Is rates were quantified from healthy mature leaves, located in the upper third of the canopy. For each day of the study, one branch near the top of one of the plants was cut and placed in tap water before being recut and transported to the laboratory. The leaf gas exchange measurements occurred between 9:00 AM and 5:00 PM, with the branch exposed to ambient light and temperature conditions.

Isoprene Emissions and Net Photosynthesis
Pn and Is rates were quantified from I. edulis leaves using a portable open gas exchange system (IRGA) (LI-6400XT, LI-COR Inc., Lincoln, NE, USA) with an artificial light source 6400-02B Red Blue. The flow rate entering the LI-6400XT leaf chamber was set to 537 mL min −1 . A fraction of air exiting the leaf chamber was used to determine Is using two methods: quantified in real-time using a high sensitivity quadrupole proton transfer reaction-mass spectrometry (PTR-MS, Ionicon Analytik, Innsbruck, Austria) or thermal desorption (TD) gas chromatography-mass spectrometry (GC-MS, 5975C series, Agilent Technologies, Santa Clara, CA, USA). Using four-way junction fitting, air exiting the leaf chamber was delivered or to the PTR-MS (40 mL min −1 ) or to the TD tube (100 mL min −1 when collecting), with the remainder of the flow diverted to the vent/match valve within the LI6400XT. The excess flow entering the vent/match valve was maintained to at least 200 mL min −1 [25]. Background measurements were performed with an empty leaf chamber, at the beginning and at the end of each experiment.

Light and Temperature Response of Isoprene
For each I. edulis leaf, either a light or temperature response curve was generated. Leaves were evaluated for their response of Pn and Is to changes in PPFD (Photosynthetic Photon Flux Density) at 0, 25, 50, 75, 100, 250, 500, 750, 1000, 1500, and 2000 µmol m −2 s −1 under constant leaf temperature (30 • C) in both a CO 2 -free atmosphere (0 µmol mol −1 , 4 leaves) and ambient CO 2 (450 µmol mol −1 , 7 leaves). To achieve a CO 2 -free atmosphere inside the LI-6400XT leaf chamber, the CO 2 mixer was turned off and the soda lime was put to fully scrub. Leaf temperature response curves (25→ 30→ 35→ 40 • C) were measured under constant PPFD (1000 µmol m −2 s −1 ), again in both CO 2 -free atmosphere (n = 3) and ambient CO 2 (n = 4). Before each series of measurements, the leaves were acclimated for~15 min in the chamber, until the stomatal conductance and photosynthesis were stable.
To quantify the ratio of Is between CO 2 -free atmosphere (0 µmol mol −1 of CO 2 , 8 leaves) and ambient CO 2 concentrations (450 µmol mol −1 of CO 2 , 8 leaves), gas exchange was measured with leaves of I. edulis under standard conditions of light and leaf temperature (1000 µmol m −2 s −1 PPFD, 30 • C leaf temperature).

Isoprene Emission under Limiting Conditions for Net Photosynthesis
To evaluate Is under limiting conditions for Pn we measured gas exchange under two limiting situations: (a) using a specific photosynthesis inhibitor, DCMU (3-(3,4-dichlorophenyl)-1,1-dimethylurea) or (b) in the dark. DCMU blocks photosynthetic electron flow from photosystem II to the plastoquinone pool [36], inhibiting photosynthesis.
For the inhibitor experiment, a 180 µM DCMU solution, was prepared and adjusted to pH 7.5. A leaflet was cut from a leaf and placed in the DCMU solution for two hours, until the inhibitor took effect. This inhibition time was determined during previous experiments to be sufficient to decrease Pn to near 0 µmol m −2 s −1 in leaflets under standard conditions (i.e., with CO 2 , PPFD and leaf temperature at 450 µmol mol −1 , 1000 µmol m −2 s −1 , and 30 • C, respectively). Following uptake of the DMCU solution by the transpiration stream, a light response curve was measured under CO 2 -free atmosphere and constant leaf temperature (30 • C).
For the dark experiment, a temperature response curve was established in the absence of light. During the measurements the leaves were kept under CO 2 -free atmosphere and zero PPFD. The samples were collected in TD tubes, using the LI-6400XT/GC-MS methodology [25].

13 C-labeling of Leaf Isoprene Emissions Using Sodium Bicarbonate 13 C Delivered through the Transpiration Stream
In order to observe possible reassimilation of internal CO 2 by photosynthesis through isoprene labeling molecules, measurements of Is under CO 2 -free atmosphere were carried out on detached fully expanded I. edulis leaflets in a solution of 13 C labeled sodium bicarbonate (NaH 13 CO 3 ) (n = 5). Fresh solutions of sodium bicarbonate 13 C (20, 60, 120 mM) were prepared with the pH adjusted to 7.0. For each concentration five leaflets were cut and recut in the sodium bicarbonate 13 C solution, and leaf gas exchange measurements were initiated. In sequence, the measurements were made for two hours, under CO 2 -free atmosphere at constant PPFD (1000 µmol m −2 s −1 ) and temperature (30 • C) using the LI-6400XT/GC-MS methodology. Isoprene 13 C-labeling is reported as the 13  Under ambient CO 2 and leaf temperature of 30 • C Is and Pn were stimulated together as PPFD increased ( Figure 1a). However, while Pn saturated at roughly 1000 µmol m −2 s −1 PPFD, Is showed no sign of saturation up to the highest fluxes of PPFD applied (2000 µmol m −2 s −1 ) (Figure 1a). When Is was expressed in µmol C m −2 s −1 (IsC) and plotted against Pn ( Figure 1b) the results show that during high PPFD intensities, Is keeps increasing, representing a higher fraction of Pn relative to lower PPFD intensities. The light response curve in a CO2-free atmosphere shows that Pn remained negative, while light stimulated Is in a similar pattern to that observed under ambient conditions of CO2 ( Figure 2a). However, under CO2-free atmosphere, average Is at maximum PPFD reached 59% of the observed under ambient CO2 conditions (21.3 nmol m −2 s −1 at CO2-free atmosphere vs. 36 nmol m −2 s −1 at ambient CO2, respectively). The fact that Pn remained negative while the leaf was in the chamber demonstrated that the enclosure CO2 concentrations were below that of the CO2 compensation point, and that the leaf was a net source of CO2 to the air, likely due in-part to the stimulation of photorespiration. When real-time Is data was collected, Is saturated at lower PPFD intensity (1500 µmol m −2 s −1 PPFD, using average values) under CO2-free atmosphere relative to ambient CO2 conditions (Figure 2a vs. Figure 1a), indicating a limitation of carbon for isoprene production.  The light response curve in a CO 2 -free atmosphere shows that Pn remained negative, while light stimulated Is in a similar pattern to that observed under ambient conditions of CO 2 ( Figure 2a). However, under CO 2 -free atmosphere, average Is at maximum PPFD reached 59% of the observed under ambient CO 2 conditions (21.3 nmol m −2 s −1 at CO 2 -free atmosphere vs. 36 nmol m −2 s −1 at ambient CO 2 , respectively). The fact that Pn remained negative while the leaf was in the chamber demonstrated that the enclosure CO 2 concentrations were below that of the CO 2 compensation point, and that the leaf was a net source of CO 2 to the air, likely due in-part to the stimulation of photorespiration. When real-time Is data was collected, Is saturated at lower PPFD intensity (1500 µmol m −2 s −1 PPFD, using average values) under CO 2 -free atmosphere relative to ambient CO 2 conditions (Figure 2a vs. Figure 1a), indicating a limitation of carbon for isoprene production. The light response curve in a CO2-free atmosphere shows that Pn remained negative, while light stimulated Is in a similar pattern to that observed under ambient conditions of CO2 ( Figure 2a). However, under CO2-free atmosphere, average Is at maximum PPFD reached 59% of the observed under ambient CO2 conditions (21.3 nmol m −2 s −1 at CO2-free atmosphere vs. 36 nmol m −2 s −1 at ambient CO2, respectively). The fact that Pn remained negative while the leaf was in the chamber demonstrated that the enclosure CO2 concentrations were below that of the CO2 compensation point, and that the leaf was a net source of CO2 to the air, likely due in-part to the stimulation of photorespiration. When real-time Is data was collected, Is saturated at lower PPFD intensity (1500 µmol m −2 s −1 PPFD, using average values) under CO2-free atmosphere relative to ambient CO2 conditions ( Figure 2a vs. Figure 1a), indicating a limitation of carbon for isoprene production.  To extend the study of light responses of Is and Pn, a specific inhibitor of photosynthesis (DCMU-3-(3,4-dichlorophenyl)-1,1-dimethylurea) was provided to a I. edulis leaflet under CO 2 -free air. Although light-stimulated Is could be observed from the detached leaf placed in water as a control (up to 15 nmol m −2 s −1 ), Is were small to negligible and did not increase as a function of PPFD when DCMU was provided to the detached leaflet ( Figure 3). As stomatal conductance (gs) in the DCMU fed leaflet was similar or higher than gs in the water control leaf, this effect could not be attributed to a potential stomatal closure induced by DCMU application. To extend the study of light responses of Is and Pn, a specific inhibitor of photosynthesis (DCMU-3-(3,4-dichlorophenyl)-1,1-dimethylurea) was provided to a I. edulis leaflet under CO2-free air. Although light-stimulated Is could be observed from the detached leaf placed in water as a control (up to 15 nmol m −2 s −1 ), Is were small to negligible and did not increase as a function of PPFD when DCMU was provided to the detached leaflet ( Figure 3). As stomatal conductance (gs) in the DCMU fed leaflet was similar or higher than gs in the water control leaf, this effect could not be attributed to a potential stomatal closure induced by DCMU application.

Isoprene Emissions under CO2-Free Atmosphere are Strongly Stimulate by Leaf Temperature Increase but are Eliminated in the Dark
Under temperature response curves with constant light and CO2 conditions (1000 mol m −2 s −1 and 450 mol mol −1 ) Is was strongly stimulated as leaf temperature increased and did not show a clear observable saturation point while reaching an average of 76 nmol m −2 s −1 at the highest leaf temperature (40 °C, Figure 4a). When IsC was plotted against Pn there was a negative relationship between these two parameters (Figure 4b), which represents the uncoupling between the Pn and Is as leaf temperature increases. This is driven by both the decrease in Pn and the increase in Is with increasing leaf temperature.  Figure 4a). When IsC was plotted against Pn there was a negative relationship between these two parameters (Figure 4b), which represents the uncoupling between the Pn and Is as leaf temperature increases. This is driven by both the decrease in Pn and the increase in Is with increasing leaf temperature.  In order to further evaluate the dependence on Is and Pn, a measurement of these variables was made as a function of leaf temperature, but in the dark to eliminate photosynthesis. When a temperature response curve was performed in the dark under a CO2-free atmosphere, Is was not detectable, even at the highest temperatures studied (40 °C) ( Figure 6). In contrast, in the light, increases in leaf temperature stimulated Is, which continued to increase up to the highest leaf temperature studied (40 °C). When the same measurements were made under a CO 2 -free atmosphere, Is continued to increase with leaf temperature up to 14-20 nmol m −2 s −1 at 40 • C while Pn remained near-zero or slightly negative (e.g., −1.0 µmol m −2 s −1 ) ( Figure 5).  In order to further evaluate the dependence on Is and Pn, a measurement of these variables was made as a function of leaf temperature, but in the dark to eliminate photosynthesis. When a temperature response curve was performed in the dark under a CO2-free atmosphere, Is was not detectable, even at the highest temperatures studied (40 °C) ( Figure 6). In contrast, in the light, increases in leaf temperature stimulated Is, which continued to increase up to the highest leaf temperature studied (40 °C). In order to further evaluate the dependence on Is and Pn, a measurement of these variables was made as a function of leaf temperature, but in the dark to eliminate photosynthesis. When a temperature response curve was performed in the dark under a CO 2 -free atmosphere, Is was not detectable, even at the highest temperatures studied (40 • C) ( Figure 6). In contrast, in the light, increases in leaf temperature stimulated Is, which continued to increase up to the highest leaf temperature studied (40 • C). As it was determined that Is are stimulated by increases in light and leaf temperature under CO2-free air, we calculated Is under a CO2-free atmosphere relative to ambient CO2 under standard conditions of light and temperature (1000 mol m −2 s −1 PPFD, 30 °C, n = 8). When Is emissions under a CO2-free atmosphere were quantified from eight I. edulis leaves, it was determined that those emissions amounted to 19% ± 7% of emissions under ambient CO2.

Sodium Bicarbonate 13 C Leaf Feeding
In order to verify the contribution of CO2 reassimilation from the decarboxylation process in isoprene molecules, sodium 13   As it was determined that Is are stimulated by increases in light and leaf temperature under CO 2 -free air, we calculated Is under a CO 2 -free atmosphere relative to ambient CO 2 under standard conditions of light and temperature (1000 µmol m −2 s −1 PPFD, 30 • C, n = 8). When Is emissions under a CO 2 -free atmosphere were quantified from eight I. edulis leaves, it was determined that those emissions amounted to 19% ± 7% of emissions under ambient CO 2 .

Sodium Bicarbonate 13 C Leaf Feeding
In order to verify the contribution of CO 2 reassimilation from the decarboxylation process in isoprene molecules, sodium 13 C-bicarbonate solutions were provided to I. edulis leaflets under CO 2 -free atmosphere with PPFD and temperature constant (1000 µmol m −2 s −1 , 30 • C). Within minutes of placing the leaflet in the solution, emissions of 13 C-labeled isoprene could be observed (Figure 7). When leaflets were placed in the 20 mM solution of sodium bicarbonate 13 C they emitted Is with one labeled carbon, representing 26% of the emissions of unlabeled isoprene molecules. Emissions of completely labeled molecules [ 13 C5]isoprene were nearly undetectable. With the increase of the molarity (20-120 mM), the percentage of Is with 13 C-labeling also increased relative to the unlabeled Is, as did the number of carbon atoms 13 C-labeled per molecule. In the solution with 120 mM of sodium bicarbonate 13 C, the percentage of molecules with one 13 C atom reached 54%, with two When leaflets were placed in the 20 mM solution of sodium bicarbonate 13 C they emitted Is with one labeled carbon, representing 26% of the emissions of unlabeled isoprene molecules. Emissions of completely labeled molecules [ 13 C 5 ]isoprene were nearly undetectable. With the increase of the molarity (20-120 mM), the percentage of Is with 13 C-labeling also increased relative to the unlabeled Is, as did the number of carbon atoms 13 C-labeled per molecule. In the solution with 120 mM of sodium bicarbonate 13 C, the percentage of molecules with one 13 C atom reached 54%, with two labeled carbons 25%, with three labeled carbons 8%, with four labeled carbons 2%, and with all labeled carbons reaching up 0.35% (relative to unlabeled Is).

Under CO 2 -Free Air, Isoprene Emissions Display a Similar Light and Temperature Response Pattern to That under Ambient Conditions, But Is Eliminated when Photosynthesis Is Inhibited
The high correlation between Is and Pn as a function of light (Figure 1) is widely discussed in literature, since isoprene production via the MEP pathway utilizes, as the first precursor, a product of the Calvin-Benson cycle, glyceraldehydes-3-phosphate (G3P), which is responsible for three of the five carbons in the isoprene skeleton and requires 14 mol of NADPH and 20 mol of ATP for each mole of isoprene synthesized [37][38][39][40][41][42] (Figure 8). Thus, it is expected that the relationship between Pn and Is be close under ambient CO 2 conditions (Figure 1b).
The production of isoprene from the tropical species I. edulis under a CO 2 -free atmosphere in this study is consistent with previous research on several temperate plant species including myrtle (Myrtus communis), buckthorn (Rhamnus alaternus), velvet bean (Mucuna pruriens), gray poplar (Populus x canescens), aspen (Populus tremuloides), and tasmanian bluegum (Eucalyptus globulus) [16,27,32,43]. We demonstrate for the first time that Is from I. edulis under CO 2 -free air are stimulated by light and temperature in a similar fashion to emissions under ambient CO 2 , albeit with important differences (Figures 2b and 5b). Relative to emissions under ambient CO 2 , Is under a CO 2 -free atmosphere were reduced and saturated at low light levels (e.g., 100 µmol m −2 s −1 PPFD) versus no clear light saturation under ambient CO 2 up to 2000 µmol m −2 s −1 PPFD. Other studies that observed Is under limited or zero CO 2 have suggested that alternative carbon sources, that do not come directly from photosynthesis, such as stored starch and carbohydrates transported in the xylem stream help to maintain the carbon supply for isoprene when a reduced atmospheric CO 2 source is limiting G3P formation, while NADPH and ATP are still producing by the photochemical phase of photosynthesis [26]. In this study, we provide quantitative evidence that most carbon alternatives for isoprene production are derived from the reassimilation of internal CO 2 sources.
Previous studies have sought to understand how these alternative carbon sources can maintain high rates of Is when Pn is reduced under stress [25,34]. However, the decrease in Pn does not necessarily mean that gross photosynthesis rates are decreasing, but rather that decarboxylation processes are being stimulated (especially photorespiration), increasing the CO 2 released directly into the cell. At increasing temperatures the oxygenating reaction of RuBisCO has a relative advantage over the carboxylation reactions because (1) stomatal closure reduces the uptake of atmospheric CO 2 , (2) the solubility of CO 2 declines stronger than the solubility of O 2 , and (3) the kinetic properties of RuBisCO decrease the affinity for CO 2 more strongly than that for O 2 , causing a proportional increase in photorespiration [44].
Indeed, the Is is stimulated by elevated temperatures (Figure 4), and this dependence is a consequence of high temperature optimum for isoprene synthase and the use of excess energy not used in the Calvin-Benson cycle by the enzymatic reactions of the MEP pathway [45][46][47].
However, when leaves were fed with the photosynthesis inhibitor DCMU under a light response curve it was not possible to observe Is anymore (Figure 3). The same result occurred when the light was turned off during a temperature response curve ( Figure 6). This confirms that Is is dependent on the photochemical products of photosynthesis (NADPH and ATP). Once these energetic cofactors are no longer produced, it is not possible to maintain the isoprene synthesis by the MEP pathway demonstrating that the Is is dependent on these energetic cofactors products but can be maintain without the external carbon supply during several hours.

CO 2 Reassimilation from Decarboxylation Process as an Alternative Carbon Source to Isoprene Synthesis
With respect to the molecular composition, 60% of carbon that comprises the isoprene molecule is assigned to G3P, as one of initial precursors of the MEP pathway [15,48]. When the external CO 2 source was removed from the reference air stream, 19% ± 7% of Is continued relative to Is observed under ambient CO 2 . This percentage is in the range observed by studies that analyzed the contribution of alternative carbon sources to isoprene synthesis by providing atmospheres of 13 CO 2 to leaves. In these studies, it was observed that~10%-30% of the carbon atoms were not from atmospheric CO 2 , but instead originated from other sources [23,26,27]. Although the possibility of reassimilated CO 2 as a source of carbon for isoprene has been discussed in the literature, our results support a possibility that it represents the main alternative carbon source for isoprene and is largely responsible for maintaining high Is rates under abiotic stress conditions like high temperatures, which decrease Pn.
A study carried out in leaves of Populus sp. under 0 µmol mol −1 CO 2 observed that Is rates were little affected relative to the measures carried out under ambient CO 2 [32]. However, the authors observed that, when photorespiration was inhibited by the absence of O 2 , it was no longer possible to observe Is. In our study, we observed that the Is rates are high under photorespiratory conditions by stimulating the photorespiratory process by removing the CO 2 from the reference air.
In agreement with these results, a previous study provided a [2− 13 C]glycine (a photorespiratory intermediate) solution to I. edulis leaves under high temperature and observed Is with labeled carbon atoms [25]. The authors suggested that, in situations where photorespiration is stimulated, reassimilation of photorespiratory CO 2 by photosynthesis acts as a protective mechanism to avoid oxidative stress by consuming excess ATP and NADHP produced by the light reactions. Likewise, our results suggest that increasing the carbon flux through the MEP pathway, which also consumes excess photosynthetic ATP and NADPH, further increases plant tolerance to abiotic stress including high temperatures and drought. Thus, the coupled activity of enhanced photorespiration and the MEP pathway together with reassimilation of internal CO 2 sources may offer plants protection against abiotic stress by helping balance the sources and sinks of photosynthetic ATP and NADPH [44] (Figure 8). When sodium bicarbonate (NaHCO3) was provided to I. edulis leaves under CO2-free air, the bicarbonate could be decarboxylated by carbonic anhydrase enzyme. The CO2 released by the decarboxylation could be fixed by the Calvin-Benson cycle, constituting a G3P molecule, being able to enter into the DOXP/MEP pathway and compose isoprene molecules as observed (Figure 7).
Under conditions of limiting CO2, a potential alternative source to isoprene synthesis is the CO2 Figure 8. Simplified scheme of the main decarboxylation processes that may contribute as alternative sources for isoprene synthesis: respiration, photorespiration, and the first reaction of the DOXP/MEP pathway.
When sodium bicarbonate (NaHCO 3 ) was provided to I. edulis leaves under CO 2 -free air, the bicarbonate could be decarboxylated by carbonic anhydrase enzyme. The CO 2 released by the decarboxylation could be fixed by the Calvin-Benson cycle, constituting a G3P molecule, being able to enter into the DOXP/MEP pathway and compose isoprene molecules as observed (Figure 7).
Under conditions of limiting CO 2 , a potential alternative source to isoprene synthesis is the CO 2 released into the cell by the decarboxylation process. A previous study estimated that the refixation rate of mitochondrial respiration was related to the incomplete carbon labeling of Is, suggesting that respiratory CO 2 can contribute to isoprene formation [24]. Another study in Populus deltoides Barr. leaves measured Is under N 2 and 13 CO 2 atmosphere and observed that the inhibition of photorespiration did not remove the 12 C in isoprene released from leaves under a 13 CO 2 atmosphere and suggested that the unlabeled carbon in isoprene molecules is not derived from photorespiratory carbon [23]. On the other hand, in I. edulis under [2− 13 C]glycine, feeding a strong labeling of isoprene molecules was observed, suggesting that the CO 2 released from photorespiration can be incorporated into isoprene molecules [25]. In addition to (photo)respiration, other decarboxylation processes could be acting as alternative carbon sources to isoprene synthesis, including biosynthetic pathways such as the DOXP/MEP, fatty acid, mevalonate, fermentation, and the oxidative pentose phosphate pathway [48][49][50].
The incorporation of labeled CO 2 released by bicarbonate decarboxylation into isoprene molecules was observed within minutes after placing the leaf in a 13 C sodium bicarbonate solution (Figure 7). This confirms the contribution of CO 2 reassimilation to isoprene synthesis and suggests that other decarboxylation processes can contribute to isoprene synthesis as a primary carbon source under reduced atmospheric CO 2 availability. These results have relevance for the next decades of climate change predictions [51], since higher temperatures increase respiration and photorespiration and consequently the amount of CO 2 released in to the leaf. This CO 2 can contribute to Is in stress situations that could decrease the stomatal conductance and the carbon uptake.

The Reassimilation of CO 2 in the Cell Interior Linked to Isoprene Emission: Its Importance for Plant Physiological Functioning under Changing Environmental Factors
In our experiment, high temperatures and light intensities under a CO 2 -free atmosphere caused a remarkable depression of Pn and substantial increase in Is, suggesting that other sources than recently assimilated carbon are used for isoprene formation. Subsequently, when labeled sodium bicarbonate was provided, the emissions of labeled isoprene indicate that the reassimilation of CO 2 released in the chloroplast can contribute as a carbon source to isoprene synthesis. This is in line with evidence from Loreto et al. [24] who showed refixation of unlabeled CO 2 derived from respiration is an alternative source of carbon for isoprene in nonstressed leaves.
Under normal conditions, the reassimilation of CO 2 in the chloroplast can reach up to 40% [52] and increase to more than 80% under low CO 2 conditions and high temperature [53]. The amount of photorespiratory or respiratory CO 2 emitted or recycled by leaves have been associated with an improvement of the photosynthetic carbon gain [52,53]. Therefore, the reassimilation of CO 2 in the chloroplast can contribute significantly for photosynthetic carboxylation efficiency by minimizing the loss of photosynthetic carbon caused by stimulus of photorespiration and respiration.
Previous report have pointed out that the stimulus of internal recycling of photorespired and respired CO 2 in plants may become more important at conditions with high temperatures, low CO 2 , water stress, or high irradiance [54][55][56][57]. However, in some cases, the respiratory rate is dramatically inhibited by water stress or extreme temperature conditions, while the photorespiration rate increases significantly [58]. Efficient internal recycling of photorespired CO 2 linked to isoprene synthesis might improve tolerance against high irradiance and temperature, commonly experienced in the tropics, besides increasing the overall carbon budget of plants. Nevertheless, the contribution of each of these decarboxylation processes to the CO 2 balance is highly variable, depending on the species, stage of growth, biotic and abiotic factors [48,[50][51][52]. In addition, there is little evidence for widespread uses of an effective mechanism to trap and reassimilate photorespired CO 2 in tropical species, and particularly with relation with Is. Thus, subsequent research is needed to clarify the overall significance of trapping of CO 2 in the cell interior and its flux to isoprene production, particularly in tropical forests, where the bulk of the world's Is occur, and both temperature and drought risk are expected to increase according to most climate scenarios [59][60][61].

Conclusions
Previous studies have shown that plants can use nonatmospheric carbon sources to produce isoprene. Most studies have focused on stored carbon sources including carbohydrates (starch and glucose) and their metabolites (pyruvate and phosphoenolpyruvate) as potentially important alternative carbon sources and few studies are made with tropical species. In this study, we show that the CO 2 released by decarboxylation processes is an important carbon source for I. edulis-a widespread tropical specie in Amazon forest. Once decarboxylation processes release CO 2 within the leaf mesophyll, our observations suggest that this process acts as a quantitatively important carbon source for isoprene syntheses mainly during stress situations when the stomata are closed and the amount of internal CO 2 decreases. This study enhances our knowledge of isoprene emission by tropical species under conditions that plants often experience in the tropics, where the amount of irradiance and temperature, the two main environmental drivers of isoprene fluxes, are high throughout the year. The reassimilation of released CO 2 within the leaf and its use for isoprene synthesis may be a key tolerance mechanism against a specific type of stress, for example high irradiance and high temperature in tropical environments.