INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

A wide array of volatile organic compounds (VOCs) are emitted into the atmosphere by the leaves of many plant species (Graedel, 1979; Lerdau et al., 1997; Kesselmeier and Staudt, 1999; Kreuzwieser et al., 1999). Among the biogenic VOCs studied to date, isoprene (2-methyl-1,3-butadiene) is quantitatively the most important, with as much as 500 Tg year¹ estimated to be emitted globally from vegetation (Guenther et al., 1995), exerting profound effects on atmospheric chemistry (Monson and Holland, 2001). In the presence of nitrogen oxides and sunlight, isoprene oxidation can lead to the production of tropospheric ozone (Trainer et al., 1987; Chameides et al., 1988). Not all plants emit isoprene. Most that do are woody in growth habit and are represented by North American species of oaks (Quercus spp.), willows (Salix spp.), poplars (Populus spp.), and spruce (Abies spp.; Harley et al., 1999).

Methylbutenol (MBO; 2-methyl-3-buten-2-ol) is a C₅ terpenoid, similar to isoprene, that is emitted by several pine species native to the western United States, including ponderosa (Pinus ponderosa), lodgepole (Pinus contorta), and gray pine (Pinus sabiniana; Harley et al., 1998). Emissions of MBO from pine needles are dependent on both photosynthetic photon flux density (PPFD) and temperature (Harley et al., 1998), showing behavior that is similar to leaf emissions of isoprene (Monson and Fall, 1989; Loreto and Sharkey, 1990). Like isoprene, MBO emissions from forests represent a significant source of reactive carbon to the atmosphere (Harley et al., 1998). Despite the important influence of isoprene and MBO emissions on atmospheric chemistry, the biological functions of these VOCs remains open to debate (Logan et al., 2000; Sharkey and Yeh, 2001).

Two lines of evidence suggest that isoprene and MBO arise from similar biochemical pathways in plants. First, both of these C₅ hemiterpenes are formed via the recently discovered non-mevalonate, 1-deoxyxylulose 5-phosphate/2-C-methylerythritol 4-phosphate (DOXP/MEP) pathway of isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) formation (Lichtenthaler et al., 1997; Zeidler and Lichtenthaler, 2001). This pathway utilizes pyruvate and glyceraldehyde-3-phosphate for the synthesis of these C₅ prenyl diphosphates (Lichtenthaler, 1999). All plastidic terpenoids studied thus far, including isoprene, MBO, carotenoids, and the phytol side chain of chlorophyll (Chl), are formed via the DOXP/MEP pathway (Lichtenthaler, 1999). Second, Fisher et al. (2000) detected and partially purified an MBO synthase from gray pine needles that catalyzes the conversion of DMAPP to MBO. This reaction is similar to the formation of isoprene via isoprene synthase (Silver and Fall, 1995), except that a putative allylic carbocation intermediate is hydrated rather than undergoing proton extraction and rearrangement. These reactions are shown in Figure 1.

View larger version (9K): [in this window] [in a new window]   Figure 1.   Schemes showing the reaction mechanism by which DMAPP is catalytically converted to isoprene or MBO by isoprene synthase or MBO synthase.

The regulatory mechanisms controlling isoprene and MBO emission are not completely understood, although both short- and long-term variation in isoprene emission appears to depend on the activity of isoprene synthase (Monson et al., 1992; Kuzma and Fall, 1993; Schnitzler et al., 1997). Diurnal fluctuations in isoprene emission rates are generally well correlated with PPFD (Sharkey and Loreto, 1993) and leaf temperature (Monson et al., 1992). The PPFD dependence of isoprene emission has been attributed to control by photosynthetic electron transport (Niinemets et al., 1999; Zimmer et al., 2000), as well as direct light activation of isoprene synthase (Wildermuth and Fall, 1996; Fall and Wildermuth, 1998). Isoprene synthases from aspen (Populus tremuloides), velvet bean (Mucuna sp.), and oak, when assayed in vitro, have temperature response profiles that are generally similar to that for corresponding leaves (Monson et al., 1992; Schnitzler et al., 1996). Long-term and seasonal variation in isoprene emission has been reported (Monson et al., 1994), and in at least one case, the seasonality of emission was related to seasonal variation in extractable isoprene synthase activity (Schnitzler et al., 1997).

Although much of our current understanding of the regulation of leaf level isoprene emissions has been interpreted with respect to the activity of isoprene synthase, it has been suggested that the regulation of isoprene emission may also, in part, depend on the availability of DMAPP substrate (Wildermuth and Fall, 1996; Fall and Wildermuth, 1998; Sharkey and Yeh, 2001). Presumably, activity of the DOXP/MEP pathway must be regulated to accommodate the large changes in isoprene and MBO emission that are known to occur. Preliminary reports have suggested that transcripts of two enzymes of the DOXP/MEP pathway (deoxyxylulose synthase and deoxyxylulose reductoisomerase) are induced and accumulate in leaves of an isoprene emitter when switched to high temperature (Sharkey and Yeh, 2001).

Recently, we developed a method to quantify DMAPP, an end product of the DOXP/MEP pathway, in biological samples (Fisher et al., 2001). Using this technique, we set out to determine if patterns of accumulation of DMAPP in leaves and needles vary among isoprene- and MBO-emitting and nonemitting plant species. In this study we provide the first evidence for differences in the regulation of DMAPP levels in leaves. Because production of DMAPP is compartmentalized in plant cells (Lichtenthaler et al., 1997), occurring separately in both the cytosol and plastids, we present evidence for the localization of DMAPP in cottonwood (Populus deltoides) leaves.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Diurnal Time Course of DMAPP Content

In a past study, we developed methods for quantifying DMAPP in leaf and needle tissues with recovery greater than 95%. Using this method, we observed large diurnal increases in relative DMAPP content in leaves of the isoprene-emitting species cottonwood when sampled at predawn and midday (Fisher et al., 2001). Prompted by these results, in the current study, we measured DMAPP throughout the day in leaves from several trees of cottonwood exposed to a natural photoperiod (Fig. 2). The relative content of DMAPP in whole-leaf preparations changed through the day, and appeared to correspond with changes in incident PPFD. Quantifiable levels of DMAPP were clearly present in leaves at the beginning and end of the photoperiod. By midday, relative DMAPP content increased nearly 2.5-fold, returning to near predawn levels by dusk.

View larger version (29K): [in this window] [in a new window]   Figure 2.   Diurnal variation in the PPFD (400-700 nm) and whole-leaf DMAPP concentration for leaves of potted cottonwood that were grown in the University of Colorado greenhouse. The DMAPP symbols represent the mean ± SE of age-matched leaves from five replicate plants.

Survey of DMAPP Content in Leaves of Isoprene- and MBO-Emitting and Nonemitting Species

We next quantified DMAPP content, at two times of day (predawn and midday), in leaves from seven species growing on the campus and in the greenhouses of the University of Colorado (Fig. 3A). The species chosen represented a mix of known isoprene-emitting and nonemitting species. We were interested in determining if the large diurnal increases in DMAPP content that were observed in cottonwood were a general phenomenon or a specific characteristic of isoprene-emitting plants. All leaves examined contained quantifiable levels of DMAPP at both predawn and midday. However, the absolute content of DMAPP varied markedly among the species examined, with isoprene emitters generally exhibiting the highest levels of DMAPP. The nonemitter common pole bean (Phaseolus vulgaris) had the lowest DMAPP content at both predawn and midday (5.37 ± 0.18 and 7.64 ± 0.35 nmol g¹ fresh weight, respectively). DMAPP levels at predawn and midday in the weak isoprene emitter cowpea (Vigna unguiculata) were similarly low (5.92 ± 0.46 and 12.43 ± 1.31 nmol g¹ fresh weight, respectively). In contrast, leaves from the strong isoprene emitter red oak (Quercus rubra) contained the highest levels of DMAPP at predawn and midday (119.4 ± 32.65 and 290 ± 29.42 nmol g¹ fresh weight, respectively). All of the isoprene emitters tested, and the nonemitting common pole bean, showed significant increases in whole-leaf DMAPP content from predawn to midday (P < 0.05). The two nonemitters common walnut (Juglans nigra) and red maple (Acer rubrum) exhibited no significant diurnal changes in DMAPP content (P = 0.73 and 0.48, respectively). Although absolute levels of DMAPP varied significantly among the isoprene-emitting and nonemitting species, isoprene-emitting species displayed a 110% to 149% increase in DMAPP content from predawn to midday, whereas nonemitters exhibited smaller increases from 9% to 42% (Table I).

View larger version (20K): [in this window] [in a new window]   Figure 3.   The midday (MD) and predawn (PD) whole-leaf DMAPP concentrations for several isoprene-emitting species (A) and MBO-emitting species (B) growing on the University of Colorado campus or in the greenhouses. Horizontal bars represent the mean ± SE of leaves from three to five replicate plants. A single asterisk by each pair of values represents predawn and midday differences at the significance level of P < 0.01 and a double asterisk represents predawn and midday differences at the significance level of P < 0.05.

To determine if large diurnal changes in DMAPP content were also characteristic of MBO-emitting species, a similar analysis was performed on needles collected from trees of four pine species growing in a greenhouse at the University of Colorado (Fig. 3B). The DMAPP content of needles varied significantly among the species examined. Similar to the results for isoprene emitters, all of the pines tested contained quantifiable levels of DMAPP at predawn, with the nonemitting eastern white pine containing the greatest levels (300.7 ± 22.40 nmol g¹ fresh weight). Needles collected at midday from the MBO emitter, ponderosa, displayed the greatest DMAPP contents of all pine species (654.4 ± 74.44 nmol g¹ fresh weight). All of the MBO-emitting species exhibited significant increases in whole-needle DMAPP content when comparing predawn and midday measurements (P < 0.05). In contrast, needles from the nonemitter eastern white pine displayed no significant diurnal changes in DMAPP content (P = 0.10). Despite the large variation in DMAPP content among the four pine species tested, MBO-emitting species exhibited pronounced increases in DMAPP content predawn to midday (110%-1,063%), whereas the nonemitting species eastern white pine did not (Table I).

In general, we expressed DMAPP content on a fresh weight basis. To determine if the basis for expression alters the interpretation of results, we performed additional analyses of leaf and needle tissue, expressing DMAPP content on several common bases, including nmol g¹ fresh weight, nmol g¹ dry weight, nmol mg¹ Chl and nmol mg¹ protein (Table II). Irrespective of the basis for expression, the midday content of DMAPP varied consistently among the species examined. Among fully expanded leaves and needles, the absolute content of DMAPP was highest in the nonemitting eastern white pine. In contrast, young leaves of cottonwood contained the lowest levels of DMAPP. Similarly, young leaves of cottonwood contained significantly less DMAPP for a given basis of expression than a similarly orientated mature leaf (P < 0.05).

Cellular Localization of DMAPP

Next, we determined the relative intracellular localization of DMAPP in leaves of the isoprene emitter cottonwood harvested at midday. Leaves were fractionated using a nonaqueous fractionation technique (Gerhardt and Heldt, 1984), and DMAPP levels in each fraction were subsequently measured. A typical density gradient with associated distributions of markers and DMAPP is shown in Figure 4. Chl, the chloroplast marker, occurred in the upper, lightest, fractions; the greatest percentage of vacuolar marker (-mannosidase) occurred in the lower, heaviest, fractions; and the cytosolic marker (phosphoenolpyruvate carboxylase) generally partitioned toward the middle of the gradient. This relative distribution of compartment markers is typical for plant leaves after centrifugation in nonaqueous gradients (Gerhardt and Heldt, 1984; Sharkey and Vanderveer, 1989; Stitt et al., 1989; Moore et al., 1995). In comparison with the distribution of markers, DMAPP was predominately associated with Chl. In every case, approximately 70% of the DMAPP recovered from the entire gradient was partitioned into the lightest two fractions, and it was always visually associated with Chl. The correlation coefficient between percent DMAPP and percent Chl in numerous gradients was generally greater than 0.98 (data not shown).

View larger version (29K): [in this window] [in a new window]   Figure 4.   The partitioning of various markers and DMAPP within the nonaqueous fractions that were used. Data are from a single gradient representative of five different experiments.

The activities or amounts of markers and analytes in the gradient fractions were examined by graphical analyses to determine relative distributions between the cytosol and chloroplasts. Using this two-compartment analysis technique, we determined that approximately 65% to 70% of the DMAPP in leaves of cottonwood harvested at midday occurs in the chloroplast (Table III). Attempts were made to perform similar analyses on predawn leaves. Unfortunately, we were unsuccessful in measuring active phosphoenolpyruvate carboxylase from these preparations. However, several attempts did provide anecdotal evidence that the greatest fraction of predawn DMAPP was recovered from fractions lower in the gradient, compared with midday DMAPP. However, the lack of a suitable predawn cytosolic marker precludes any further interpretation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The work presented here represents the first detailed measurements of DMAPP in plant leaves. This is of general interest in plant biology because DMAPP is one of the essential precursors of all plant terpenoids. Two key findings are presented. First, we detected a diurnal increase in DMAPP content that was pronounced in leaves of isoprene and MBO emitters, but there was little or no increase in leaves of nonemitters. Second, in isoprene-emitting cottonwood leaves, DMAPP was largely localized in chloroplasts. Both of these observations may be directly related to the biological rationale for light-dependent hemiterpene emission from leaves, a subject of some debate for several decades (Logan et al., 2000; Sharkey and Yeh, 2001), and are of general interest in understanding the regulation of higher terpenoid biosynthesis in leaves.

We first observed large diurnal changes in DMAPP content in leaves of cottonwood while developing a sensitive nonradioactive method for quantifying DMAPP in biological samples (Fisher et al., 2001). In the current study, a more complete characterization revealed that DMAPP levels vary diurnally in a manner that roughly parallels incident PPFD (Fig. 2). This pattern of DMAPP accumulation is remarkably similar to previously reported patterns of isoprene emission versus PPFD (Monson and Fall, 1989; Loreto and Sharkey, 1990; Monson et al., 1992). When additional leaves of isoprene-emitting and nonemitting species were compared, we found that the capacity to emit isoprene was clearly associated with large diurnal variation in DMAPP content (Fig. 3; Table I). Differences in the magnitude of diurnal DMAPP accumulation were also seen in a comparison of MBO-emitting and nonemitting pine species (Fig. 3; Table I), and the differences in accumulation correlate well with known patterns of emission (for estimates of MBO emission rates in these species, see Harley et al., 1998).

There is clearly a tendency for amplification of the capacity for light-dependent DMAPP production in isoprene- and MBO-emitting species. However, the fact that even some nonemitters exhibit diurnal increases in DMAPP suggests that the light-dependent response may be a more general characteristic of plant metabolism. The light-dependent synthesis of IPP and DMAPP could support the production of higher isoprenoids during periods of high photosynthetic activity, including the replacement and repair of damaged Chls, carotenoid molecules, and pigment-protein complexes that occur at high PPFD (Niyogi, 2000).

When midday DMAPP contents were expressed on different measures, and compared in isoprene- and MBO-emitting and nonemitting species, no significant differences in interpretation were found (Table II). However, it was apparent that young (isoprene-emitting) leaves of cottonwood contained significantly less DMAPP on all measures than did similarly orientated fully mature leaves. This may simply reflect a partially depleted pool of free DMAPP as a result of increased carotenoid and Chl biosynthesis during leaf development. It may also represent an overall decrease in the synthesis of DMAPP as a result of the presumed reduction in photosynthetic metabolism. Although the results are preliminary, they suggest that the availability of DMAPP may, in some way, be associated with the ontogenetic delay in isoprene emission that has been observed in previous studies (Kuzma and Fall, 1993).

Cellular Localization of DMAPP

It is now known that DMAPP is produced by two distinct and independent biosynthetic routes within the plant cell; the classic mevalonic acid pathway in the cytosol and the DOXP/MEP pathway of plastids (Lichtenthaler et al., 1997). To interpret the relevance of our measures of DMAPP content for influencing VOC emission, it is necessary to provide an estimate of the relative distribution of DMAPP between these two cellular compartments. We used nonaqueous fractionation as a method for determining the intracellular distribution of DMAPP in leaves of cottonwood (Table III). There are two ways of calculating subcellular metabolite distributions from nonaqueous fractionation data: a two-compartment analysis (Gerhardt and Heldt, 1984) and a three-compartment analysis (Riens et al., 1991). Because DMAPP is generally thought to only occur in the cytosol and chloroplasts, and the presence of phosphorylated intermediates are generally absent from the vacuole (Farré et al., 2001), the two-compartment model is justifiable (see discussion in "Materials and Methods"). Results from our analysis suggest that a significant percentage (65%-71%) of DMAPP measured at midday in cottonwood leaves is associated with the chloroplast. It seems likely that these two distinct pools of DMAPP derive from the two independent pathways for isoprenoid synthesis (Licthenthaler et al., 1997). However, whether the two pathways exchange intermediates is not well understood, although Heintze et al. (1990) and McCaskill and Croteau (1995) have suggested that IPP could be transported across the plastid envelope.

Using the results from the localization experiment, it is possible to derive an estimate of the concentration of DMAPP in isoprene-emitting chloroplasts. Our measurements reveal that the whole-leaf concentration of DMAPP in isoprene emitters ranges from 2 to 45 µmol m². If it is assumed that 60% of DMAPP occurs in the chloroplast, and that the ratio of chloroplast volume to total leaf area is 0.009 L m² (as determined in Quercus robur by Zimmer et al., 2000), then the concentration of chloroplastic DMAPP would range from 0.13 to 3.0 mM. This estimate is consistent with the relatively high K_m values that have been reported for isoprene synthase (0.5-8 mM; Silver and Fall, 1995; Schnitzler et al., 1996) and MBO synthase (5 mM; Fisher et al., 2000). The concentration of chloroplastic DMAPP is similar to those reported for other chloroplastic phosphorylated intermediates, including 3-phosphoglycerate (2.0-4.3 mM), dihydroxyacetone phosphate (0.21-0.32 mM), and Fru-1,6-bisphosphate (0.55-1.04 mM; Winter et al., 1993, 1994; Leidreiter et al., 1995).

When taken together, the large diurnal variation in DMAPP content associated with isoprene and MBO-emission, as well as the localization of DMAPP to the cottonwood chloroplast, suggests that the production of DMAPP may participate in the regulation of isoprene and MBO emission from plant leaves. Previous studies have primarily emphasized the importance of the synthase enzymes in regulating light-dependent VOC emission (Sharkey and Yeh, 2001). However, in at least one study of isoprene emission in oak, Lehning et al. (1999) reported no diurnal changes in extractable isoprene synthase activity despite marked diurnal variation in emission rate, concluding that diurnal variation in leaf temperature could account for observed diurnal changes in isoprene emission rate. However, this analysis depended on the assumption of constant and saturating DMAPP substrate conditions. Results from our study suggest that a constant supply of DMAPP does not occur in isoprene-emitting species during the diurnal cycle. This raises the possibility that some of the diurnal variation in isoprene emission observed by Lehning et al. (1999), and in general the PPFD dependence of isoprene and MBO emission, is due, in part, to variation in DMAPP substrate supply.

Substrate level regulation of isoprenoid biosynthesis is not without precedent. In animals (Edmond and Popjak, 1974), and in some plants (Nes and Bach, 1985; Bach et al., 1999), excess carbon is diverted away from isoprenoid production by the "mevalonate shunt," a series of reactions converting DMAPP to dimethylacrylyl-CoA, an intermediate of Leu metabolism. This shunting of carbon away from isoprenoid biosynthesis has been shown to regulate isoprenoid synthesis in embryonic Drosophila melanogaster cells (Havel et al., 1986) and in rat (Rattus norvegicus) liver (Marinier et al., 1987). Recently, overexpression of either deoxyxylulose synthase or deoxyxylulose reductoisomerase, two key enzymes in the DOXP/MEP pathway, led to the accumulation of various isoprenoids such as Chls, carotenoids, and abscisic acids in Arabidopsis (Estévez et al., 2001) and an increase in essential oil accumulation in peppermint (Mentha × piperita; Mahmoud and Croteau, 2001), respectively. These results support the role of substrate level regulation of isoprenoid biosynthesis.

Although our understanding of the regulatory mechanisms controlling isoprene and MBO emission are improving, the ultimate reason(s) for these emissions remains unclear. Several experiments have provided support for the hypothesis that volatile isoprenoids (isoprene, monoterpenes, and other short-chain hydrocarbons) protect leaves from short high-temperature bursts (Sharkey and Yeh, 2001). However, evidence for increased thermotolerance due to isoprene was not found in at least two studies (Logan and Monson, 1999; Logan et al., 1999). Other experiments demonstrate that isoprene protects leaves exposed to high-ozone episodes and therefore isoprene may play an important antioxidative role in plant leaves (Loreto et al., 2001). However, previous studies have suggested that the hydroperoxides produced by the reaction between ozone and isoprene may be as damaging to leaves as ozone alone (Hewitt et al., 1990; Sauer et al., 1999).

An alternative hypothesis to explain the possible advantages of isoprene emission has focused on the regulation of chloroplast metabolism through maintenance of appropriate pool sizes for various metabolites (Logan et al., 2000). Although support for this hypothesis remains correlative in nature, results from this study suggest that the total flux of carbon through the DOXP/MEP pathway can be significant. The conversion of DMAPP to isoprene and MBO, and their emission, may help regulate the size of the midday DMAPP pool, and thereby release previously unavailable phosphate.

Results from this study highlight the need for an enhanced understanding of the role of the DOXP/MEP pathway in regulating the biosynthesis of isoprenoid VOCs. The pronounced diurnal variation in DMAPP associated with isoprene and MBO emission suggests an inherent difference in the regulation of this pathway in species that synthesize isoprene and MBO, compared with nonemitting species. Alteration in the activity of this pathway may result from altered activities of DOXP/MEP enzymes, enhanced supply of pathway precursors, or both. A careful reexamination of the sources of carbon entering the DOXP/MEP pathway will be required to further test the hypothesized role of isoprene emission in metabolic regulation, as well as refine the potential significance of substrate level control of isoprene and MBO emission from leaves and needles. At present, we are developing new experimental systems using intact chloroplasts, which will allow us to test the validity of these ideas.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Leaves and needles were collected from trees growing on the campus and in the greenhouses of the University of Colorado. Whole leaves were collected directly into liquid N₂ at two times of the day, immediately before sunrise (predawn) and at the middle of the photoperiod (midday). All leaves were fully exposed sun leaves, and were uniformly orientated with respect to incident light and of similar age. Leaves were stored in liquid N₂ until further processing.

Assay and Quantification of DMAPP

DMAPP was quantified as described previously (Fisher et al., 2001). In brief, the levels of DMAPP were determined by measuring the amount of isoprene released into the headspace of a sealed reaction vial upon acidification. Frozen leaves were ground to a fine powder with added liquid N₂ in a mortar and pestle, and 100 mg of the resulting powder was rapidly transferred to a 4.8-mL glass vial containing 500 µL of distilled water. This suspension was immediately acidified by the addition of 500 µL of 8 M H₂SO₄ and sealed with a Teflon-lined septum. Vials were incubated for 1 h at 30°C, after which 1.5 mL of the vial headspace was withdrawn with a gas-tight syringe, and analyzed for isoprene by gas chromatography, described in detail in Fisher et al. (2001). Calibration of the detection system was performed using authentic isoprene standard (Scott Specialty Gases, Longmont, CO). Isoprene production from DMAPP hydrolysis was calibrated using aqueous solutions of DMAPP synthesized and purified as previously described (Davisson et al., 1985). Although isoprene is one of only several hydrolysis products resulting from the acidification of DMAPP, we have found that the production of isoprene is linear from about 2 pmol to 15 nmol DMAPP with this method. Recovery of DMAPP in leaf and needle samples was greater than 95% in all cases. Although needles of MBO-emitting pines contain small amounts of MBO, about 0.06 nmol g¹ fresh weight in illuminated ponderosa (Pinus ponderosa; Fisher, 2001), and MBO produces isoprene upon acidification (Fisher et al., 2001), these levels are insignificant in the DMAPP determinations.

Nonaqueous Fractionation

The nonaqueous fractionation procedures were similar to those reported elsewhere (Gerhardt and Heldt, 1984; Sharkey and Vanderveer, 1989). Sun-exposed cottonwood (Populus deltoides) leaves were collected directly into liquid N₂ at midday. After removing the major veins, 5 g of leaves was ground in liquid N₂ using a mortar and pestle. Powdered material was transferred to cold lyophilization jars and immediately placed on a freeze-drying apparatus (Labconco, Kansas City, MO), with a collector temperature of 80°C. The jars were kept in liquid N₂ until the air pressure was reduced to 10 mtorr. Samples were dried overnight, and 200 mg of dried sample was added to 20 mL of heptane (at 4°C) in a cutoff 50-mL centrifuge tube housed in a 250-mL beaker containing partially frozen liquid heptane. The sample was sonicated for 30 s (power setting 3, 30% duty cycle), and filtered successively through 200- and 80-µm nylon nets (Tetko Inc., Briarcliff Manor, NY). The sample was precipitated by centrifugation at 100g for 3 min. Leaf pellets were resuspended in 1 mL of heptane (at 4°C), and each sample was loaded onto a 7-mL discontinuous gradient of heptane/tetrachloroethylene, which ranged from 1.35 to 1.60 g mL¹. Samples were centrifuged at 4°C in a swinging bucket rotor (model HB-4, Sorvall Instruments, Newtown, CT) at 25,000g for 3 h. Gradients were fractionated into six portions from the top using a transfer pipette (typically about 1 mL for each). For marker assays, 0.5 mL was diluted with cold heptane in microfuge tubes, and the gradient material was precipitated by centrifugation at 10,000g for 5 min. The remaining gradient material from each fraction was diluted with heptane and collected by centrifugation. Residual heptane was evaporated overnight at 4°C under reduced pressure in a vacuum jar with desiccant and paraffin.

Marker Assays

The dried fractions were resuspended in 5 mL of buffer containing 0.1 M Bicine (pH 7.8), 5 mM MgCl₂, 1 mM EDTA, and 5 mM dithiothreitol. The samples were sonicated for 20 s (power setting 3, 30% duty cycle), and centrifuged at 10,000g for 5 min. Supernatants were used for enzyme assays and the pellet was used for Chl determinations. Chl was used for the chloroplast marker. Chl was determined by adding 1 mL of 80% (v/v) acetone to the pellet, sonicating for 30 s, centrifuging for 2 min in a microcentrifuge, and repeating the extraction twice. The concentration of Chl (a + b) was determined using the extinction coefficients of Porra et al. (1989).

Phosphoenolpyruvate (PEP)-carboxylase (PEPC; E.C. 4.1.1.31) activity was used as a marker for the cytosol. PEPC activity was assayed as described by Krall and Edwards (1993), with 0.3 mM NADH, 5 units mL¹ malate dehydrogenase, and 2.5 mM PEP in a final volume of 1 mL. After a 5-min pre-incubation at 25°C, the reaction was initiated with PEP, and PEPC activity was measured as the reduction in NADH detected as a spectrophotometric decrease in A₃₄₀.

For the vacuolar marker, we used -mannosidase (E.C. 3.2.1.24) activity, assayed in duplicate for 60 min at 37°C as described by Sharkey and Vanderveer (1989). Total protein was determined by the method of Bradford (1976).

DMAPP in Gradient Fractions

Dried gradient fractions were resuspended in 0.5 mL of distilled water, rapidly transferred to a 4.8-mL glass vial, acidified with 8 M H₂SO₄, and assayed as described above. In an initial experiment, the recovery of DMAPP from fractionated material was determined. The recovery of DMAPP added to Chl-containing fractions was 91%.

Data Analysis

The distribution of DMAPP was calculated according to a two-compartment, graphical analysis as described by Gerhardt and Heldt (1984) and Sharkey and Vanderveer (1989). Because DMAPP is thought to be largely associated with the chloroplast and cytosol, we calculated (from the appropriate graphs) the relative distribution of DMAPP between these two compartments. Because this analysis assumes that the DMAPP in each fraction is only from the chloroplast stroma and cytosol, it is necessary to estimate the contribution of the vacuole to determine the error in this analysis. Assuming all of the DMAPP in the heaviest fraction was vacuolar (an overestimate) and applying the ratio of DMAPP to vacuolar marker to the top fractions, less than 9% of the DMAPP in the top fractions came from vacuolar contamination. Because the majority of DMAPP (>70%) in multiple gradients consistently partitioned into the two lightest fractions, and were predominantly associated with the Chl fractions, we feel it is justifiable to apply this two-compartment approach.