The methylerythritol phosphate (MEP) pathway in plants produces the prenyl precursors for all plastidic isoprenoids, including carotenoids and quinones. The MEP pathway is also responsible for synthesis of approximately 600 Tg of isoprene per year, the largest non-methane hydrocarbon flux into the atmosphere. There have been few studies of the regulation of the MEP pathway in plants under physiological conditions. In this study, we combined gas exchange techniques and high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS-MS) and measured the profile of MEP pathway metabolites under different conditions. We report that in the MEP pathway, metabolites immediately preceding steps requiring reducing power were in high concentration. Inhibition of the MEP pathway by fosmidomycin caused deoxyxylulose phosphate accumulation in leaves as expected. Evidence is presented that accumulation of MEP pathway intermediates, primarily methylerythritol cyclodiphosphate, is responsible for the post-illumination isoprene burst phenomenon. Pools of intermediate metabolites stayed at approximately the same level 10 min after light was turned off, but declined eventually under prolonged darkness. In contrast, a strong inhibition of the second-to-last step of the MEP pathway caused suppression of isoprene emission in pure N2. Our study suggests that reducing equivalents may be a key regulator of the MEP pathway and therefore isoprene emission from leaves.
The methylerythritol phosphate (MEP) pathway (Fig. 1) is one of two pathways in plants responsible for the biosynthesis of dimethylallyl diphosphate (DMADP) and isopentenyl diphosphate (IDP), the building blocks of all isoprenoids including carotenoids, monoterpenes, and prenyl chains of chlorophylls and quinones. The MEP pathway in plants is essential as it has been shown that mutants lacking this pathway were unable to develop functional chloroplasts (Mandel et al. 1996; Lois et al. 1998). MEP pathway-derived DMADP also leads to the production of approximately 600 Tg of isoprene (C5H8) per year, or about 1/3 of global hydrocarbon emission from all sources (Guenther et al. 2006). This biogenic isoprene contributes to tropospheric ozone production and affects formation of aerosols (Went 1960; Chameides et al. 1988; Kiendler-Scharr et al. 2009). However, relatively little is known about the regulation of the MEP pathway in plants.
A useful probe of the MEP pathway flux in vivo is isoprene emission. In emitting species, isoprene emission accounts for well over 90% of the flux through the MEP pathway (Sharkey, Loreto & Delwiche 1991). When light is turned off, isoprene emission from a leaf quickly declines to almost zero within 10 min, presumably because the MEP pathway requires energetic cofactors from the light reactions of photosynthesis (three ATP- and three NADPH-equivalents per C5 molecule) (Sharkey, Wiberley & Donohue 2008). The integral of post-illumination isoprene emission has been proposed to reflect the pool size of plastidic DMADP (Rasulov et al. 2009). Interestingly, it had been observed in poplars and oaks that isoprene emission rises again in darkness before eventually falling off to zero on a longer time scale (‘post-illumination burst’) (Monson et al. 1991; Rasulov et al. 2010, 2011; Li, Ratliff & Sharkey 2011). In a revisit to this phenomenon, we hypothesized that a pool of intermediate metabolites in the MEP pathway may be trapped upon an almost instantaneous depletion of reducing power during the first few moments of darkness, and that these metabolites were later converted to isoprene as reducing power becomes available (Li et al. 2011). The size of the pool of metabolites giving rise to the post-illumination burst is comparable with the size of the DMADP pool and these two pools responded similarly to environmental variables (Li et al. 2011; Rasulov et al. 2011).
To understand the nature of post-illumination isoprene burst and gain insights into regulation of the MEP pathway, it would be useful to measure levels of leaf MEP pathway metabolites under physiological conditions. Some of the best efforts to date include studies using radioisotopes and 31P-NMR studies. Using 31P-NMR, it has been shown that methylerythritol cyclodiphosphate (MEcDP), a cyclic compound produced by MEcDP synthase (MDS, step 5), accumulates in leaves (Rivasseau et al. 2009; Mongélard et al. 2011). MEcDP was the only MEP pathway metabolite detectable by 31P-NMR (Rivasseau et al. 2009).
In this study, we present a method for profiling of MEP pathway metabolites in leaves based on high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS-MS). With the exception of diphosphocytidylyl methylerythritol phosphate (CDP-MEP), an unstable intermediate, all MEP pathway metabolites were successfully measured. We show that metabolites immediately before steps in the MEP pathway that require reducing power are in high concentration whereas other metabolites are present in very low concentrations. The MEP pathway intermediate metabolites (MEcDP in particular) were the sources of post-illumination isoprene burst. Leaves exposed to pure N2 accumulated very high levels of MEcDP, suggesting that the second-to-last step of the pathway was inhibited in the absence of both CO2 and O2.
MATERIALS AND METHODS
Hybrid aspen (Populus tremula × alba) grown in Michigan State University greenhouses were used for this study. Unless otherwise stated, whole plants were brought into the lab in insulated transporters and attached leaves were used for gas exchange experiments. In the fosmidomycin-feeding experiment where detached leaves were used, the leaf was swiftly cut at the base of the petiole with a fresh razor blade, and then immediately recut underwater and transported to a floral tube under water.
Gas exchange experiments
The experimental set-up for gas exchange was described by Li et al. (2011). Synthetic air of 80% N2 and 20% O2 was mixed using mass flow controllers and supplied to a LI-6400 Portable Photosynthesis System (Li-Cor Biosciences, Lincoln, NE, USA), where CO2 was added to bring the air to 400 µmol mol−1 CO2. A custom-built leaf cuvette was put in line with the LI-6400 chamber to allow real-time monitoring and recording of gas exchange parameters (e.g. carbon assimilation, transpiration and Ci). The exhaust gas of LI-6400 was connected to a Fast Isoprene Sensor (FIS, Hills Scientific, Boulder, CO, USA) where simultaneous isoprene measurements were made. Air was humidified by passing an adjustable portion of N2 through a humidifier. Cuvette temperature was controlled by passing water from a water bath through the leaf chamber. Leaf temperature was monitored using a thermocouple wired from LI-6400 into the leaf cuvette. The light source was an SL3500 white light emitting diode array (Photon Systems Instruments, Drasov, Czech Republic). All leaves were acclimated at standard conditions (30 °C leaf temperature, 1000 µmol m−2 s−1 photosynthetic photon flux density at leaf level) for at least 1 h before the start of an experiment.
Extraction of leaf metabolites
During a gas exchange experiment, the leaf was quickly taken out of the cuvette and immediately frozen in liquid N2. The whole leaf was then ground up in liquid N2 and extracted with 3:1:1 acetonitrile-isopropanol-50 mm ammonium acetate in dH2O (adjusted to pH 10 with ammonium hydroxide). The leaf extract was centrifuged at 14 000 g for 10 min and the supernatant was stored at −80 °C for later analysis.
Metabolites of the MEP pathway were separated and quantified on a ZIC-pHILIC column (Merck SeQuant, Umeå, Sweden) fitted to a 3200 Q-TRAP mass spectrometer (Applied Biosystems, Carlsbad, CA, USA) that was coupled with two LC-20AD HPLC pumps and a SIL-HTc autosampler (Shimadzu, Kyoto, Japan). Standards of MEP pathway metabolites were separated using a binary gradient consisting of 50 mm ammonium acetate adjusted to pH 10.0 with ammonium hydroxide and acetonitrile (Supporting Information Table S1). Electrospray ionization (ESI) in negative ion mode was used. Enhanced product ion scans were first run to assess fragmentation patterns and determine the optimal product ions for monitoring. Compound-dependent MS parameters were then optimized (Supporting Information Table S2). Mass spectra were acquired in multiple-reaction monitoring mode for the optimized precursor/product ion pairs (Fig. 1).
As isotope-labelled standards were not commercially available, metabolite levels in leaf samples were quantified based on linearly fitted curves of unlabelled standards. Standards of MEP pathway metabolites, including deoxyxylulose phosphate (DXP), MEP, diphosphocytidylyl methylerythritol (CDP-ME), MEcDP, hydroxymethylbutenyl diphosphate (HMBDP), DMADP and IDP, were acquired from Echelon Biosciences (Logan, UT, USA). Due to their short shelf life under even deep-freeze conditions, CDP-MEP standards were freshly prepared from CDP-ME, by the CMK-catalyzed reaction using a CDP-MEP Synthesis Kit (Echelon Biosciences). Newly synthesized CDP-MEP was quantified in two ways, by measuring ATP consumption and CDP-ME consumption in the reaction. All other reagents used were analytical or HPLC grade and used without further purification.
Quality control and data analysis
Recovery ratios were determined by spiking of leaf samples with authentic standards. For paired comparisons, leaves frozen in liquid N2 were divided into two similar-sized fractions, one of which was spiked during the grinding step. The unspiked sample was then normalized to the pellet weight of the spiked sample for determination of endogenous metabolite levels. Recovery ratios have been taken into account in the data presented in this paper. Limits of detection (LOD) were defined as the lowest concentrations that gave a signal-to-noise ratio of two. One-way analysis of variance and unpaired t-tests (P < 0.05) were used to detect significant differences between groups. Statistical analyses were carried out in Origin (OriginLab, Northampton, MA, USA).
Separation and quantitation of MEP pathway metabolites by HPLC-MS-MS
To separate MEP pathway metabolites that are highly polar, we employed hydrophilic interaction chromatography (HILIC) using a zwitterionic stationary phase attached to polymeric beads. A binary gradient was used for separation of the metabolites (Supporting Information Table S1). Standards of MEP pathway metabolites eluted between 4.4 and 9.1 min (Fig. 2). DMADP and IDP could not be differentiated based on fragmentation patterns or by chromatography in this study. MEP pathway metabolites in leaves were extracted in a 3:1:1 acetonitrile-isopropanol-aqueous buffer. This extraction buffer was chosen as it approximated the initial mobile phase and ensured rapid arrest of cellular metabolism at the same time. Recovery ratios of standards added to leaf samples were determined for each metabolite (Table 1). With the exception of CDP-MEP, all metabolites were recovered in leaves with good consistency. Low recovery of CDP-MEP is likely due to the innate instability of this compound.
Table 1. Retention time and recovery ratios of MEP pathway metabolites in leaf extract
Each number in recovery ratios denotes the mean ± SE (n = 7).
Levels of MEP pathway metabolites in leaves during a light-to-dark switch
When light was switched off, isoprene emission from an aspen leaf initially declined (phase I) and then rose in the dark before declining again to approximately zero (phase II, Fig. 3a). To determine the source of the post-illumination isoprene burst, MEP pathway metabolites were measured in leaf samples taken at three time points: in light, 10 min into darkness (before the burst) and 40 min into darkness (after the burst) (Fig. 3b). In the light, three intermediate metabolites – DXP, MEcDP and HMBDP – accumulated to significant levels. In particular, MEcDP measured 0.92 µmol m−2. Levels of MEP and CDP-MEP were below LOD. Whole-leaf DMADP and IDP measured 0.79 µmol m−2 in total. After light was switched off for 10 min and isoprene emission declined to its first minimum (Fig. 3a), DMADP/IDP levels were reduced to 13% of previous levels in light. In contrast, levels of MEcDP were unaffected and slightly increased to 0.96 µmol m−2. Levels of other metabolites were reduced to between 18% (CDP-ME) and 64% (HMBDP) of their respective levels under light. At 40 min into darkness, the amount of all metabolites decreased. The amount of isoprene emission during phase II is compared with the amount of decrease in metabolites in phase II, in Fig. 3b, second panel.
Levels of MEP pathway metabolites in fosmidomycin-fed leaves, and leaves acclimated in pure N2
Isoprene emission levels and MEP pathway metabolites were measured in leaves fed with 20 µm fosmidomycin for over 80 min (Fig. 4a). Fosmidomycin inhibits DXR, the enzyme catalyzing step 2 of the MEP pathway (Fig. 1). Isoprene emission of fosmidomycin-fed leaves was reduced by 93% compared with the control emission levels. DXP levels increased 40-fold in fosmidomycin-fed leaves while other downstream metabolites were reduced to between 16% (CDP-ME) and 38% (HMBDP) of their control levels (Fig. 4b).
Acclimation of a leaf in N2 quickly inhibited isoprene emission, in a similar fashion to the suppression of isoprene emission by darkness (Fig. 5a). However, no ‘burst’ was observed in N2. A large transient overshoot in isoprene emission was observed when O2 and CO2 were switched back on, but subsequently isoprene emission failed to fully recover over a 2 h period, remaining ∼20% below pretreatment values. In leaves acclimated in N2, MEcDP increased 32-fold while DMADP/IDP decreased to 6% of control levels (Fig. 5b). Levels of other metabolites decreased to between 26% (CDP-ME) and 85% (DXP) of their respective controls.
MEP pathway metabolites accumulate before steps requiring reducing equivalents
Intermediate metabolites in the MEP pathway can account for post-illumination isoprene burst
Upon darkening of a leaf, isoprene emission dropped rapidly to almost zero within 10 min (phase I, Fig. 3a). This decline in isoprene emission could be, in theory, caused by a decrease in (1) carbon input into the MEP pathway; (2) ATP levels; (3) ferredoxin levels; or (4) levels of NADPH. In this study, we show that in phase I, DMADP/IDP levels were reduced by 87% while MEcDP levels slightly increased. Under prolonged darkness (phase II), MEcDP eventually decreased to 12% of the level found in illuminated leaves. Taken together, this suggests that steps between MEcDP and DMADP/IDP (steps 6 and 7) were blocked in phase I. This blockage was reversed over time, and intermediate metabolites in the MEP pathway were metabolized to isoprene, leading to the post-illumination isoprene burst in phase II.
Step 6 in the MEP pathway, catalyzed by the iron-sulphur protein HMBDP synthase (HDS), is tightly coupled to photosynthesis (Seemann et al. 2002, 2006). In contrast to its homolog in Escherichia coli, the plant HDS cannot use the NADPH/flavodoxin/flavodoxin reductase system, but instead could use isolated thylakoid preparations as the sole electron donor upon illumination (Seemann et al. 2005, 2006). Reducing power required by HDS in light can be directly shuttled via ferredoxin from electron transport of the light reactions and in the absence of NADPH (Okada & Hase 2005; Seemann & Rohmer 2007). Therefore, it is likely that this step is rapidly slowed down (incompletely inhibited) upon darkness, trapping the MEcDP pool. Over time, this inhibition is relieved as HDS switches to NADPH as its electron source in the dark (Seemann & Rohmer 2007). NADPH could be low immediately after switching the light off and then be replenished through the pentose phosphate pathway. The first step in the pentose phosphate pathway catalyzed by plastidic glucose-6-phosphate dehydrogenase is redox sensitive and is activated in the darkness by oxidation (Lendzian & Ziegler 1970; Wenderoth, Scheibe & von Schaewen 1997). The levels of DXP and HMBDP were decreased by less than 50% in phase I, while other metabolites (CDP-ME and DMADP/IDP) showed much larger reductions, suggesting that step 2 and step 7 were also slowed down in phase I, consistent with control by redox status. It is likely that no new carbon came into the MEP pathway in the darkness, otherwise additional isoprene emission would have been observed.
The DMADP/IDP measured in this study includes both plastidic and cytosolic pools. It is generally accepted that the darkness-labile or fosmidomycin-labile part of DMADP/IDP pool is in the chloroplasts, while the remaining portion is of cytosolic origin (Rosenstiel et al. 2002). The integral of isoprene emission during phase I has also been hypothesized to be a measure of the plastidic DMADP pool in vivo (Rasulov et al. 2009). In this study, we show that the size of post-illumination burst (integrated emission in phase II) agrees well with the sum of the metabolic intermediates that are lost during the burst (Fig. 3b). The size of plastidic DMADP pool measured here is in good agreement with previous reports (Behnke et al. 2007; Rasulov et al. 2009). It is interesting to note that the cytosolic pool (measured in the dark) was much smaller than the plastidic pool in the leaf.
Acclimation in nitrogen turns off isoprene emission by inhibiting the second-to-last step of the MEP pathway
It has been reported long ago that isoprene emission is rapidly abolished in N2 (i.e. CO2- and O2-free air) (Fig. 5a) (Loreto & Sharkey 1990, 1993). The suppression of isoprene production in N2 is not due to a limitation on carbon supply as isoprene emission is much less affected in CO2-free air in which photorespiration would cause a significantly more negative carbon balance (Loreto & Sharkey 1990). In leaves acclimated in N2, MEcDP accumulated to very high levels while DMADP decreased by over 90%. Leaves followed for 1 h in N2 did not show a second burst, indicating that inhibited flow through step 6 cannot be reestablished during sustained exposure to N2 the way it can be in sustained darkness (Fig. 5b). The subsequent overshoot in isoprene emission when O2 and CO2 were restored to the air around the leaf presumably came from accumulated MEcDP. In the absence of O2 and CO2, ribulose 1·5-bisphosphate carboxylase/oxygenase (Rubisco) will be inactive, resulting in very little sink for electrons from photosynthetic electron transport. The NADP-malate dehydrogenase activation state can be used as a proxy for plastidic redox status and it has been shown that NADP malate dehydrogenase in Phaseolus vulgaris leaves exposed to a nitrogen atmosphere becomes fully activated (Scheibe 1987; Weise et al. 2006). We suggest that in N2, the redox poise of the cell is disrupted, causing a depletion of reducing equivalents required for step 6; alternatively, the [4Fe-4S] cluster of HDS is oxygen sensitive (Seemann et al. 2005; Rivasseau et al. 2009), and perhaps it suffered from increased turnover rate. This is supported by the observation that, following an initial burst of isoprene emission upon returning to air, emission capacity was irreversibly reduced (Fig. 5a).
In conclusion, we show that darkness and N2 inhibit MEP pathway at the same step (step 6) but in a different fashion. Profiling of the MEP pathway metabolites suggest this step and the two other steps requiring reducing power are likely the points of regulation of isoprene emission under different environment conditions. Indeed, flux through HDS measured by 31P-NMR positively correlates with light intensity and temperature (Mongélard et al. 2011). In addition to enhancing our understanding of factors controlling isoprene emission from leaves, a better understanding of the regulation of MEP pathway has other important practical implications. For example, the MEP pathway in bacteria and pathogens represents an ideal therapeutic target for development of new drugs. In addition, there is a growing interest in metabolic engineering of the MEP pathway to produce compounds of important medical and commercial values, such as taxol, vitamins and biofuels.
We thank Dr Frank Telewski for providing poplar trees; Drs A. Daniel Jones, Chao Li and Lijun Chen for assistance with LC-MS-MS; Echelon Biosciences for the gift of CDP-MEP synthesis kit; and Alex Corrion and Dr Sean E. Weise for assistance with gas exchange experiments. This work was supported by a National Science Foundation Grant IOS-0950574 to T.D.S.