INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

2-Methyl-1,3-butadiene (isoprene) is emitted from leaves of various plant species (Kesselmeier and Staudt, 1999) and influences the trace-gas composition of the troposphere by reacting with OH radicals and NO_x to generate tropospheric ozone (Trainer et al., 1987; Chameides et al., 1988).

It has been demonstrated that isoprene is produced in the chloroplasts from primary photosynthetic products via the 2-methylerythritol-4-phosphate (MEP) pathway (Zeidler et al., 1997; Lichtenthaler, 1999). Fast labeling by 99% (v/v) ¹³CO₂ indicated direct coupling between isoprene and photosynthesis (Sharkey et al., 1991a), although uncertainty remains concerning the potential contribution of cytosolic substrates to chloroplastic isoprene production, such as isopentenyl pyrophosphate (IPP) produced via the mevalonic acid pathway (Lichtenthaler et al., 1997a). Process-based isoprene emission models use photosynthesis as a starting point (Niinemets et al., 1999; Zimmer et al., 2000) assuming direct coupling between the two processes. However, there are indications for incomplete coupling in some cases, such as emission of isoprene in the absence of net assimilation in the dark (Shao et al., 2001) or under CO₂-free air (Monson and Fall, 1989; Loreto and Delfine, 2000; Affek and Yakir, 2002), that requires isoprene synthesis using alternative carbon source(s). Isoprene protection, such as against oxidative damage (Affek and Yakir, 2002), may depend on such decoupling to maintain isoprene production when photosynthesis is at least partially inhibited.

Isotopic discrimination against ¹³C (, where  = R_substrat/R_product  1, and R = ¹³C/¹²C) occurs during diffusion of CO₂ into leaves (4.4) and during CO₂ fixation by Rubisco (29; Roeske and O'Leary, 1984; Guy et al., 1993). The combined, scaled effects of diffusion and carboxylation lead to photosynthetically fixed carbon and organic matter depleted in ¹³C with respect to atmospheric CO₂. Further carbon isotope discrimination occurs during lipid synthesis due to an isotopic effect that may be associated with decarboxylation in the production of acetyl CoA from pyruvate (DeNiro and Epstein, 1977; Melzer and Schmidt, 1987) and/or due to site-specific isotopic heterogeneity in pyruvate in which the carboxyl carbon is relatively enriched (Rinaldi et al., 1974; Gleixner and Schmidt, 1997).

Discrimination against ¹³C was also observed in isoprene production. Isoprene emitted from red oak was depleted in ¹³C by 3 with respect to photosynthetically fixed carbon (Sharkey et al., 1991b). This was suggested to be associated with isotopic discrimination by the pyruvate dehydrogenase complex as part of the mevalonate pathway. However, this interpretation should be revised to fit the currently accepted MEP pathway (Lichtenthaler, 1999), for which the possible discrimination steps have not yet been explicitly identified.

In the present work, we use the carbon isotopic compositions of isoprene and of newly fixed carbon to examine the coupling between isoprene production and photosynthetic carbon flow and to extend the knowledge of ¹³C discrimination in isoprene synthesis in leaves of myrtle (Myrtus communis), reed (Phragmites australis), buckthorn (Rhamnus alaternus), and velvet bean (Mucuna pruriens).

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Source Effect

A change in the isotopic composition of the CO₂ supplied to a leaf enclosed in a gas-exchange cuvette was immediately reflected in the ¹³C content of the fixed carbon, _fixed (where  = [R_sample/R_standard  1] × 10³, R = ¹³C/¹²C and the standard is Vienna Pee Dee Belemnite). It was partially reflected within 5 min in the ¹³C content of isoprene, _isop, emitted from myrtle leaves and reached a constant value within about 30 min (Fig. 1). The relatively rapid response of _isop to labeling implied small isoprene pool size in the leaves, which was confirmed also by direct measurements. Only approximately 100 nmol m² isoprene was obtained by extractions from leaves of myrtle-1.

View larger version (23K): [in this window] [in a new window]   Figure 1.   Changes in the isotopic composition of photosynthetically fixed carbon (from on-line gas exchange and isotopic measurements) and of isoprene emitted from a branch of myrtle-3, in response to a rapid change in the isotopic composition of the source CO2 (at a time marked by the vertical lines). The numbers denote the discrimination in isoprene production relative to fixed carbon (c-isop in per mil). Photosynthesis was brought to steady state before the onset of the experiment and all conditions other than the 13C of the source CO2 were kept constant throughout the experiments.

As expected, _isop values were significantly lower than the ¹³C values of concurrently fixed carbon, _fixed (Fig. 1). In contrast to expectations, however, the apparent discrimination against ¹³C from carbon fixed in photosynthesis to concurrently emitted isoprene, _C-isop (where _C-isop = [_{fixed  isop}]/[_isop/1,000+1]), was sensitive to the ¹³C of the CO₂ supplied (Fig. 1). Similar results were obtained using leaves of myrtle, velvet bean, and buckthorn, as summarized in Figure 2. The variations in the apparent discrimination in response to changes in the ¹³C of the CO₂ supply are reflected in the slopes of _isop versus _fixed, which were smaller than 1 and varied between 0.72 ± 0.02 and 0.91 ± 0.03 (Fig. 2, a-e). In reed, on the other hand, the apparent discrimination did not vary significantly with changes in the ¹³C of the source CO₂, resulting in a slope of 1 (Fig. 2f).

View larger version (40K): [in this window] [in a new window]   Figure 2.   Relationships between the isotopic composition of the photosynthetically fixed carbon (fixed;3b from on-line gas exchange and isotopic measurements) and that of isoprene (isop) emitted by leaves of myrtle-1 through -3 (a-c, ), velvet bean (d), buckthorn (e), and reed (f). The vertical lines indicate the isotopic composition of leaf organic matter (leaf; n between 4 and 10; SE between 0.2 and 0.6). The horizontal lines denote the apparent discrimination (c-isop, calculated from the plot assuming that leaf represents the mean fixed during the growth period of the leaf). The slopes of myrtle, buckthorn, and velvet bean were significantly lower than 1 based on t test for ([slope  1]/slope SE), with n  2 degrees of freedom. Slope SE was 0.03 (n = 106), 0.02 (n = 83), 0.02 (n = 42), 0.04 (n = 33), 0.02 (n = 28), and 0.04 (n = 30) for plots a to f, respectively. The gray triangles and dashed line in plot a denote the influence of ABA on fixed and isop in a branch of myrtle-1.

In myrtle, we also modified _fixed without changing the CO₂ supply, by changing c_i/c_a through stomatal closure induced, in turn, by abscisic acid (ABA) treatments (30-100 µM). Stomatal conductance decreased from 0.24 ± 0.09 to 0.02 ± 0.01 mol m² s¹ (average ± SE, n = 3) and net assimilation decreased from 7 to 1 µmol m² s¹. This led to a decrease in c_i/c_a from 0.84 ± 0.002 to 0.61 ± 0.004 and in photosynthetic discrimination, yielding more positive _fixed (compare with Farquhar et al., 1982) and _isop values (Fig. 2a). The relationships of _isop versus _fixed showed an even lower slope than that observed in the CO₂-labeling experiments (above). Additional ABA treatments on other myrtle branches showed similar results with relatively large variations in slopes among branches (data not shown).

The variations in apparent discrimination indicated that some of the isoprene was not labeled by recently fixed carbon and that carbon from a source independent of current assimilation (termed below alternative carbon source) was incorporated into isoprene (see "Discussion"). The measurements described below were performed to characterize this unlabeled isoprene and to recognize its carbon source.

CO₂-Free Air

Measurements were carried out under CO₂-free air (or in the dark, see below) when there is no photosynthesis, to obtain isoprene that cannot be labeled by concurrent photosynthesis. After switching to CO₂-free air, isoprene emission was sustained for several hours, without significant change in emission rates (Affek and Yakir, 2002). _isop values under CO₂-free air were constant over approximately 2 h of treatment with a mean value of 41.7 ± 1.4 (average ± SE, n = 6, in myrtle-1 and -2 plants). This value was independent of _isop values during the several hours preceding the CO₂-free air treatment that ranged between 35 and 50, for different pretreatment in which ¹³C of the CO₂ supply ranged between 8 and 31. These results indicated that isoprene was produced under CO₂-free conditions from an unlabeled carbon source. Only when _isop preceding the treatment was as low as approximately 60 was some depletion in _isop observed, i.e. from 41.7 to 47.9 ± 1.4 (n = 7). Isoprene emission under CO₂-free air was observed also in reed and buckthorn leaves with _isop values of 44.4 ± 2.2 (n = 2) and 45.8 ± 0.6 (n = 7), respectively, irrespective of the _isop values preceding the CO₂-free air treatment that ranged between 35 and 50.

Emission in the Dark

In the dark, isoprene emission from myrtle decreased rapidly to detection limit levels not sufficient for isotopic analysis. Consequently, _isop was measured in the light until a steady-state value was observed and then during the first few minutes of darkness. _isop values in the dark followed changes in _isop values in the preceding light period, such as due to changes in source CO₂. _isop-dark was slightly more depleted than _isop-light (by 2.7 ± 0.7; n = 4) for CO₂ source during the preceding light period with ¹³C values of either 8 or 31.

Glc Labeling

To examine incorporation of glycolytic carbon into isoprene (Fig. 3), we fed both myrtle and buckthorn leaves with ¹³C-enriched Glc (¹³C_Glc was 10 as compared with ¹³C of leaf organic matter of 29 or 30). No change in isoprene emission rates was observed during the feeding treatments. _isop and _fixed or _respired were measured at ambient or zero CO₂ concentrations, respectively, with and without Glc feeding. Under CO₂-free air, both respired CO₂ and isoprene were clearly labeled by the enriched Glc and both to a similar extent (Table I). Under ambient CO₂ concentration, Glc feeding led to a slight decrease in net assimilation rates, but neither _isop nor _fixed changed significantly.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Schematic representation of isoprene biosynthesis pathway and possible coupling to cytosolic Glc metabolism and IPP. Also noted are the sites of action of the inhibitor fosmidomycin (Fellermeier et al., 1999) and of the isotopic discrimination by DXS. DHAP, dihydroxyacetone phosphate; DMAPP, dimethylallyl pyrophosphate; TPP, thiamine pyrophosphate.

View this table: [in this window] [in a new window]   Table I.   Effects of feeding leaves with 13C-enriched Glc (13C = 10) on the isotopic composition of isoprene and respired CO2, under CO2-free air conditions Each pair of lines denotes one branch. Average and SE refer to three or four measurements for either control of treatment.

Cytosolic IPP

To examine the potential effect of cytosolic IPP and the possibility that it contributes unlabeled carbon to isoprene, we reduced chloroplastic IPP formation using fosmidomycin. In myrtle, buckthorn, and reed, fosmidomycin led to inhibition of isoprene emission, although about 10% of the emission rates persisted even after treatment (Fig. 4; compare with Loreto and Velikova, 2001). Partial inhibition resulted in ¹³C depletion of _isop and an increase of apparent discrimination, _C-isop, from an average of 9.4 ± 0.5 before treatment to 11.4 ± 0.5 during inhibition (average ± SE, n = 8, P < 0.001; Table II). This increase in _C-isop was observed even when _fixed and _isop before the fosmidomycin treatment were highly depleted. Notably, inhibition with fosmidmycin resulted also in a 43% increase in leaf pyruvate content from 23 ± 3 µmol m² in control leaves to 33 ± 2 µmol m² after 4 h of inhibitor treatment (3 µM; P < 0.02, n = 4).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Effects of fosmidomycin inhibition on net assimilation () and isoprene emission rates () in a branch of myrtle-1. The vertical lines indicate the beginning of fosmidomycin feeding (5 µM at 12:30 and 10 µM at 16:20 PM). Leaf temperature was 26°C, light intensity was 250 µmol m2 s1, and ci varied between 220 and 250 µL L1. Due to the decrease in net assimilation observed at the high fosmidomycin used here, we used only up to 5 µM fosmidomycin in the isotopic analysis experiments.

Isotopic Composition of Isoprene Emitted to the Atmosphere

Short-term _C-isop in myrtle-1 leaves was measured under ambient air and high flow rate, to obtain typical physiological c_i and _fixed values. This yielded _C-isop of 7.7 ± 0.7 (_isop = 29.2 ± 0.6, _fixed = 21.7 ± 0.3, n = 8). However, measurements as in Figure 1 cannot be used to estimate the intrinsic isotopic discrimination in isoprene production, as indicated by the results presented in Figure 2, because of the deviations from the 1:1 in the _fixed versus _isop relationships. Alternatively, an estimate for long-term mean _C-isop under natural conditions can be obtained by substituting ¹³C of total leaf organic matter (_leaf) for _fixed. For myrtle-1 this yielded mean _C-isop value of 7.7 (Fig. 2a), consistent with the values obtained in the short-term measurement. Using the same approach, mean _C-isop values in velvet bean, reed, and buckthorn were estimated to be 4.2, 11.0, and 9.6, respectively (Fig. 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Incomplete Coupling between Isoprene and Assimilation

The time course of changes in _isop values after a change in the isotopic composition of source CO₂ (Fig. 1) confirmed the dynamic connection between CO₂ assimilation and isoprene productions (Sharkey et al., 1991a). Quantitatively, however, the coupling between isoprene production and recently fixed carbon was clearly incomplete. If concurrent photosynthetically fixed carbon was the only carbon source for isoprene, it should be reflected in _isop versus _fixed relationships of 1:1. But the observed slopes of these relationships were considerably smaller than 1 (Fig. 2). Such behavior indicates a significant contribution from alternative carbon source(s), with ¹³C content independent of the ¹³C of the CO₂ source. Contribution from such alternative carbon source(s) with constant ¹³C content would be expected to enhance _C-isop when its ¹³C value is more negative than current _fixed, and vice-versa when it is more positive than current _fixed. Such response was clearly observed in the results reported in Figure 1.

As expected, possible contribution from stationary isoprene pool in the leaves could be ruled out based on the direct measurement of only approximately 100 nmol m² isoprene extractable from myrtle leaves. Such stationary pool would support about 100 s of emission (i.e. based on typical emission rate of 10 nmol m² s¹ and 10% contribution from alternative sources), whereas emission from an unlabeled carbon source was sustained for several hours.

Interestingly, in previously reported labeling experiments, partial ¹³C labeling of isoprene has also been observed, even though the question of alternative carbon sources was not directly addressed. In the study of Sharkey et al. (1991b) with red oak, _C-isop was slightly smaller when ¹³C-depleted CO₂ was used (slope of 0.96 ± 0.04, n = 6). In another labeling study (Delwiche and Sharkey, 1993), most of the isoprene emitted from red oak was rapidly labeled by ¹³CO₂. But after 20 min, only 80% of the isoprene was labeled, which was similar in magnitude and time scale to the labeling pattern of phosphoglyceric acid (Canvin, 1979; Delwiche and Sharkey, 1993). Such results also suggest a contribution from an alternative, slow turnover, carbon source for isoprene or for phosphoglyceric acid and subsequently for isoprene (Delwiche and Sharkey, 1993). Such incomplete coupling between isoprene and concurrently assimilated carbon may be important in models using net assimilation to predict isoprene emission. Low isoprene emission was observed in Scots pine (Pinus sylvestris) with only approximately 90% labeling of the isoprene by ¹³CO₂ (Shao et al., 2001). This was explained by the possible existence of two carbon sources, whose turnover rates are different. Similarly, only approximately 90% labeling by newly fixed ¹³CO₂ was observed in non-stored sabinene in Fagus sylvatica (Kahl et al., 1999) and non-stored -pinene and 3-methyl-3-buten-1-ol in Quercus ilex (Loreto et al., 1996a, 1996b). The question of alternative carbon source(s) for isoprene recently received renewed interest, and during the revisions of this paper, two other labeling studies provided evidence for contributions of extrachloroplastic carbon source (Karl et al., 2002b) or xylem-transported Glc (Kreuzwieser et al., 2002).

The results of our labeling experiments were supported by those for the ABA treatments. Modification of _fixed via changes in stomatal conductance and c_i/c_a, rather than CO₂ labeling, resulted in the expected enrichment in _fixed and _isop. But here too, coupling between _fixed and _isop was incomplete (Fig. 2a). The lower slope in the ABA treatments, as compared with the labeling experiments, was possibly due to decreased net assimilation rates with stomatal closure that would increase the relative contribution of the alternative source(s). Sharkey et al. (1991b) observed a more pronounced enrichment in _isop in red oak, under conditions of very low c_i, when _fixed is expected to become more enriched.

Large variations in the slopes of the _isop versus _fixed relationships from 1% to approximately 0.7% was observed (Fig. 2). This reflected species (genetic) effects, but likely also the effects of environmental factors on individual plants. The dynamic nature of the variable coupling between isoprene emission and concurrent photosynthesis was particularly evident in reed. In this case, a slope of 1 (Fig. 2) indicated full coupling to photosynthesis, but the emission under CO₂-free air suggested engagement of an alternative carbon source. Such dynamic response is significant because it may help explain the reduced sensitivity of isoprene emission to stress effect, as compared with photosynthesis (Sharkey and Loreto, 1993; Loreto and Delfine, 2000; Affek and Yakir, 2002). Further, such effect would enhance the potential protection effects by isoprene against, for example, oxidative stress when photosynthesis is partly inhibited (Affek and Yakir, 2002).

Alternative Carbon Source(s) for Isoprene

Characteristics

Glycolytic Sources

Cytosolic IPP

Isotopic Discrimination

Isotopic Discrimination in the MEP Pathway

Isotopic Composition of Isoprene Emitted to the Atmosphere

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Plants of velvet bean (Mucuna pruriens) were grown from seeds (Glendale Enterprises Inc., De Funiak Springs, FL) under ambient light and temperatures. Measurements were conducted on fully expanded, attached leaves. Mature branches of myrtle-1 and -2 (Myrtus communis) and buckthorn (Rhamnus alaternus) and leaves of reed (Phragmites australis) were cut under water from plants grown on the campus of the Weizmann Institute of Science (Rehovot, Israel) and were kept with the stem immersed in deionized water. Some measurements were done using attached myrtle leaves from plants grown in pots in a greenhouse. Potted plants were transferred to the campus at least a week before measurements (myrtle-3). Typical isoprene emission rates were approximately 10 nmol m² s¹ in all plants used.

Gas Exchange

Net assimilation rates, isoprene emission rates, and the isotopic composition of isoprene were measured with a leaf gas exchange system centered on a flow-through leaf cuvette in which the leaves were sealed. Net assimilation rates were measured using an infrared gas analyzer (Li-6262, LI-COR, Lincoln, NE) under light intensity of 1,000 µmol m² s¹ photosynthetically active radiation and leaf temperature of 26°C ± 0.5°C. An aliquot of the air in the leaf cuvette was pumped through a loop on a six-port valve (Valco Instruments Co, Houston, TX), which was cooled by a mixture of ethanol and dry ice (75°C) for trapping the hydrocarbons. After trapping, the valve was switched to a flow of He, the loop was rapidly heated (200°C), and the trapped hydrocarbons passed to a gas chromatography (GC) column (GC-HP 5890, Wilmington, DE). The details of the gas exchange and isoprene sampling system are given in Affek and Yakir (2002).

Isotopic Analysis of Isoprene and Leaf Organic Matter

The isotopic composition of isoprene (_isop) was determined after GC separation using a Q-plot GC column (Supelco, Bellefonte, PA; 30 m long, 0.32 mm inner diameter; temperature program, 50°C for 1 min  10°C min¹  170°C for 2 min) and combustion to CO₂. This GC column was selected because it provided good separation of isoprene and the large CO₂ peak (that interfered with isoprene in other columns). Retention times of CO₂ and isoprene were 300s and 700s, respectively, providing complete separation. A selection valve (see below) enabled us to direct the CO₂ peak to the flame ionization detector (FID), whereas only the isoprene peak was directed to the mass spectrometer. Testing on other columns (with myrtle and buckthorn) confirmed that they emit only isoprene (Affek and Yakir, 2002), and velvet bean and reed are commonly known as such. Furthermore, we tested separation of isoprene from several monoterpenes and 2-methyl-3-buten-2-ol and obtained good results with the Q-plot column.

A selection valve (MOVPT-1/100 pneumatic valve, SGE, Melbourne, Australia) was used to direct the flow eluting the GC column to either a FID or a combustion oven (CuO, 850°C) in which the desired GC peaks were each combusted quantitatively to CO₂. FID was used for isoprene concentration measurements. The CO₂ resulting from the combustion oven was used for isotopic composition measurements. The CO₂ from combustion was dried in a cold trap and fed in a flow of He, through an open split, to an isotope ratio mass spectrometer (IRMS; Optima, Micromass, Manchester, UK), where masses 44, 45, and 46 were measured. The ratio 45 to 44 was normalized to a pulse of CO₂ reference gas injected before each sample. The isotopic results are expressed in the  (per mil) notation versus Vienna Pee Dee Belemnite standard, where  = (R/R_std  1) × 1,000 and R, R_std are the isotopic ratios ¹³C/¹²C of the sample and the standard, respectively.

For isotopic calibration, liquid isoprene was injected into a pre-evacuated glass bulb (10 L, 60 mtorr), and air or N₂ was added to atmospheric pressure. Aliquots were sampled and measured at the end of each experiment in a similar manner to the air in the leaf cuvette. Typical precision for isoprene standard measurements was ±0.3, and the ¹³C of isoprene standard in either air or in N₂ was the same, indicating no influence of CO₂ on _isop. The ¹³C of the liquid isoprene was predetermined by comparison with international and laboratory working standards, which were measured by conventional on-line combustion elemental analyzer (EA1109 CHN-O, Carlo Erba Instruments, Milan) connected to IRMS (Optima, Micromass). Ground whole dry leaves were measured by combustion in the same elemental analyzer to obtain ¹³C values of total leaf organic matter (_leaf).

Isotopic Analysis of CO₂

_isop was compared with the isotopic composition estimated for newly fixed carbon (_fixed), based on isotopic measurements of the CO₂ in the leaf cuvette. Samples of the air entering and leaving the leaf cuvette were dried and collected in glass flasks (100 mL) and ¹³C of the CO₂ was measured as described by Gillon and Yakir (2000). The CO₂ from each flask was trapped in liquid N₂ in a sample loop (1/16 inch outer diameter, 350 µL), which was then heated; and the sample was carried in a flow of He (80 mL min¹), separated on a Porapak QS packed column (Supelco; 2 m, 50-80 mesh, 50°C) or Haysep D (Supelco; 3 m, 80°C), and analyzed for isotopic composition by IRMS (either Optima or a 20-20 [PDZ Europa, Crewe, Cheshire, UK]). _fixed was estimated by on-line discrimination calculations (Evans et al., 1986) using the CO₂ concentrations and ¹³C in the air entering and leaving the leaf cuvette.

CO₂ was calibrated by measuring air from cylinders of 400 or 1,000 µL L¹ CO₂ having different ¹³C values that were, in turn, calibrated by comparing with a cylinder of 400 µL L¹ CO₂ of known ¹³C value. Typical precision of ¹³C analysis in CO₂ was ±0.15.

Air containing CO₂ of various isotopic compositions was supplied to leaves during experiments. Ambient air (¹³C of approximately 8) was pumped through a 50-L external buffering volume and dried using drierite (8 mesh; W.A. Hammond Drierite, Xenia, OH). Cylinder air containing no CO₂ was mixed via mass flow controllers (MKS1179A, MKS instruments, Andover, MA) with cylinder air (Gordon Gas and Chemicals, Tel Aviv) containing 1% or 2.5% (v/v) CO₂ whose ¹³C value was 48, 33, or 27. Cylinder air mixture containing 383 or 366 µL L¹ CO₂ whose ¹³C value was 13 or 42, respectively, were also used.

CO₂-Labeling Experiments

Experiments testing the effect of the _fixed on _isop were performed by measuring both parameters when the leaf cuvette was supplied with air of ambient CO₂ concentration and of a certain ¹³C value. After approximately 2 h of constant _fixed and _isop, the source CO₂ was rapidly switched, changing the ¹³C but not any of the gas-exchange parameters. _fixed and _isop were measured for an additional 2 h under the new source CO₂.

Few experiments were done in the dark or under CO₂-free air. Gas-exchange parameters and the isotopic composition of isoprene and CO₂ were measured in the light and under ambient CO₂ concentration during few hours to obtain control values. For CO₂-free air measurements, the air supply to the leaf cuvette was switched rapidly to CO₂-free air. CO₂ produced by respiration resulted in 10 to 20 µL L¹ CO₂ in the cuvette. Gas-exchange parameters and the isotopic composition of isoprene and CO₂ were measured under these conditions during 2 to 4 h. For measurements in the dark, the leaf cuvette was covered by a dark cloth. The isotopic composition of isoprene was measured within 10 min (after longer times, isoprene emission was too low for isotopic measurements).

Isoprene Pool in Leaves

The amount of isoprene stored in 10 myrtle leaves was measured by freezing leaves in liquid N₂ immediately after cutting and extracting the volatile fraction by heating under a flow of He (Loreto et al., 1998). The sample was dried by magnesium perchlorate (Aldrich Chemical Co., Milwaukee), trapped in a loop cooled by a mixture of ethanol and dry ice, heated, and measured by GC, as described above. Magnesium perchlorate was examined separately and was found to influence neither the concentration nor the isotopic composition of an isoprene standard.

Labeled Glc and Inhibitor Feeding

All chemicals used in our experiments were fed to the leaves as aqueous solutions through the petiole. Fosmidomycin (Molecular Probes, Eugene, OR) was fed at concentrations of 5 to 20 µM for emission rates measurements and 2 to 5 µM for isotopic analysis experiments. Measurements were performed during approximately 2 h before feeding and then during 2 h, beginning approximately 1 h after onset of fosmidomycin feeding. D-Glc (BDH Chemicals, Poole, Dorset, UK; ¹³C = 10) was fed at concentration of 15 mM. Measurements were done during few hours before feeding at both ambient CO₂ concentrations and CO₂-free air. Then, feeding of Glc were performed during approximately 2 h under ambient CO₂ concentration and continued for an additional 2 h under CO₂-free air.

Pyruvate Content

Water soluble fraction of leaves was extracted from leaf-discs (total area of 15 cm²) as described by Duranceau et al. (1999). Pyruvate was separated by ProStar HPLC (Varian, Palo Alto, CA) with an AS11 column (Dionex, Sunnyvale, CA) at 24°C, with NaOH concentration gradient of 0.4 to 22.5 mM as eluent, at 2 mL min¹, and measured by a Dionex ED50 electrochemical detector.

Statistical Analysis

Statistical analysis was done using the t test and regression functions in the data analysis add-in from Microsoft Excel 2001 for Macintosh (Microsoft Corp., Redmond, WA).