Abstract
Some, but not all, plants emit isoprene. Emission of the related monoterpenes is more universal among plants, but the amount of isoprene emitted from plants dominates the biosphere–atmosphere hydrocarbon exchange.
The emission of isoprene from plants affects atmospheric chemistry. Isoprene reacts very rapidly with hydroxyl radicals in the atmosphere making hydroperoxides that can enhance ozone formation. Aerosol formation in the atmosphere may also be influenced by biogenic isoprene. Plants that emit isoprene are better able to tolerate sunlight-induced rapid heating of leaves (heat flecks). They also tolerate ozone and other reactive oxygen species better than non-emitting plants. Expression of the isoprene synthase gene can account for control of isoprene emission capacity as leaves expand. The emission capacity of fully expanded leaves varies through the season but the biochemical control of capacity of mature leaves appears to be at several different points in isoprene metabolism.
The capacity for isoprene emission evolved many times in plants, probably as a mechanism for coping with heat flecks. It also confers tolerance of reactive oxygen species. It is an example of isoprenoids enhancing membrane function, although the mechanism is likely to be different from that of sterols. Understanding the regulation of isoprene emission is advancing rapidly now that the pathway that provides the substrate is known.
INTRODUCTION
It surprises most people to learn that plants emit much more hydrocarbon into the atmosphere than that coming from human activities, especially during extended warm weather (Purves et al., 2004), when hydrocarbon inputs into the atmosphere can be especially deleterious (Monson and Holland, 2001; Purves et al., 2004). This fact is behind the famous quote of Ronald Reagan that ‘approximately 80 % of our air pollution stems from hydrocarbons released by vegetation’ (Pope, 1980). The large amount of hydrocarbon coming from plants was used to suggest that air pollution control was not needed, quoting further: ‘so let's not go overboard in setting and enforcing tough emission standards from man-made sources’. Thus, hydrocarbon emission from vegetation is of immediate societal significance. Most of the hydrocarbon flux from the biosphere to the atmosphere is just one compound, isoprene. Despite the more obvious emissions of pleasant smells such as pine scent and lemon scent (resulting from monoterpenes), isoprene emission is the predominant biogenic source of hydrocarbon in the atmosphere, roughly equal to global emission of methane from all sources (Guenther et al., 2006; Kesselmeier and Staudt, 1999). This surprising finding of such a large flux of isoprene from plants to the atmosphere raises a number of questions, including what happens to the isoprene in the atmosphere and why plants emit isoprene. What is known and new information on why isoprene emission matters and why plants emit isoprene will be discussed and then new information on the biochemical regulation of emission rate will be presented. The history of the discovery of isoprene emission has been described elsewhere (Sanadze, 1991, 2004; Sharkey and Yeh, 2001).
WHY ISOPRENE EMISSIONS MATTER
A second effect of isoprene in the atmosphere is the growth of aerosols. Aerosols are particles in the atmosphere which give rise to Frits Went's famous natural blue hazes (Went, 1960) but also to significant health problems. The yield of aerosol per molecule in the atmosphere is much lower for isoprene than for monoterpenes and larger molecules, but because there is so much more isoprene entering the atmosphere than other molecules, isoprene may be a significant factor in aerosol formation (Claeys et al., 2004; Edney et al., 2005; Kroll et al., 2005, 2006; Ng et al., 2006; Olcese et al., 2007).
WHICH PLANTS EMIT ISOPRENE?
The taxonomic distribution of isoprene emission is broad. Mosses (Hanson et al., 1999), ferns (Tingey et al., 1987), gymnosperms and angiosperms (see http://www.es.lancs.ac.uk/cnhgroup/iso-emissions.pdf for a comprehensive list) all have members that make isoprene but also have members that do not. Isoprene synthase has been sequenced from several Populus species (Miller et al., 2001; Sasaki et al., 2005; Sharkey et al., 2005) and from kudzu (Pureria lobata) (Sharkey et al., 2005). The sequences and gene structures indicate they are part of the TPS-b family (Bohlmann et al., 1998; Trapp and Croteau, 2001) of terpene synthases. Members of this gene family also code for monoterpene and sesquiterpenes synthases in angiosperms but are not found in gymnosperms. Because of this, the evolution of angiosperm isoprene synthases (IspSs) must have occurred after the split between angiosperms and gymnosperms, and so isoprene emission capacity in gymnosperms and angiosperms must have evolved independently. Less is known about the sequences of genes coding for IspS enzymes in mosses and ferns, so no conclusions can be reached at present. As fern and moss sequences become available, comparative genomics will help make the evolutionary origins of isoprene emission and terpene synthases more clear.
Even among angiosperms, isoprene emission capacity may have evolved multiple times. The sequences for IspSs among poplars are very similar (sequences are known for P. alba, P. × canescens, P. tremuloides and P. trichocarpa) (Miller et al., 2001; Sasaki et al., 2005; Sharkey et al., 2005), but this group is very different from the IspS of kudzu (Sharkey et al., 2005). Antibodies against poplar IspS do not cross react with kudzu IspS and vice versa, and neither antibody recognizes oak IspS (Schnitzler et al., 2005; T. D. Sharkey, A. E. Wiberley and A. R. Donohue, unpub. res.). It is likely that isoprene synthesis capacity has evolved multiple times (Harley et al., 1999), possibly from a reservoir of monoterpene synthase genes (Sharkey et al., 2005). Small changes in gene sequence can easily alter both substrate and product specificity of IspS genes (El Tamer et al., 2003; Tholl, 2006; Kampranis et al., 2007). Gene sequences do not support the idea of a single origin of all IspS genes as had been proposed by Hanson et al. (1999). Instead, it appears that isoprene emission is more like the evolution of C4 metabolism, which arose numerous times in response to an environmental constraint (Sage, 2001; Sage and Pearcy, 2000).
Within any particular group of plants, there are some traits that loosely correlate with isoprene emission but there is significant variability. Indeed, there are some peculiar disjunctions. North American oaks all emit isoprene, but many European oaks do not. Instead, among European oaks a variety of behaviours is found. Some clades emit isoprene, some emit monoterpenes in a light-dependent manner, and some emit very little terpene (Loreto et al., 1998; Csiky and Seufert, 1999; Kesselmeier and Staudt, 1999).
WHY PLANTS EMIT ISOPRENE
To ask ‘why’ plants emit isoprene is really asking what advantage isoprene emission provides to the plant that makes it. The energy cost of isoprene emission is quite significant (starting from CO2, 20 ATP and 14 NADPH per isoprene molecule; Sharkey and Yeh, 2001). The balance between cost and benefit likely will vary such that isoprene emission is favoured in some species but not others. This could be a significant influence on the distribution of the capacity for isoprene emission among plants. The nitrogen cost of isoprene emission is small; data for experiments reported in Wiberley et al. (2005) are about 5 mg m−2 IspS of a total of 2·2 g m−2 soluble protein or about 0·2 % of soluble protein is IspS.
Thermotolerance
Thermotolerance has been most often discussed as the advantage plants gain by synthesizing isoprene. The first evidence for thermotolerance was based on a photosystem II chlorophyll fluorescence assay (Sharkey and Singsaas, 1995). This experiment indicated that isoprene had some relationship to temperature effects. Leaf discs that do not show damage in this assay below 45 °C do not respond to isoprene (Logan and Monson, 1999). There are many heat-tolerance mechanisms in plants (Sharkey and Schrader, 2006) and many of them are found in all species. The obvious example is heat shock proteins and factors (Vierling, 1991; Nover et al., 2001). These considerations indicate that isoprene may protect against a specific type of heat stress. Isoprene is emitted from leaves, is light dependent (Sanadze and Kalandaze, 1966; Sanadze, 1969; Rasmussen and Jones, 1973; Tingey et al., 1979; Monson and Fall, 1989; Loreto and Sharkey, 1990) and uses carbon directly from the Calvin cycle of photosynthesis (Delwiche and Sharkey, 1993; Affek and Yakir, 2003; Schnitzler et al., 2004; Ferrieri et al., 2005). While these aspects of isoprene emission pointed toward thermoprotection of leaves and specifically photosynthesis, the temperature environment of leaves was not well known.
It is difficult to measure the temperature of leaves because they are typically very thin and have very little heat capacity. The low heat capacity of leaves and high radiant energy fluxes of sunlight make large, rapid changes in leaf temperature a possibility. A system for measuring leaf temperature under natural conditions was devised (Fig. 1) that is similar to one reported by Drake et al. (1970). A very fine-wire thermocouple (0·079 mm diameter) was threaded through two veins of the abaxial surface of leaves. One joint of the thermocouple was pressed against the leaf surface while the other hung about 3 cm below the leaf. The small diameter of the thermocouple wire reduced heating from sunlight, and in any case, both joints were in the same radiation environment (below the leaf), so any radiation errors would be small. Also, the wire was in contact with the leaf for some distance on either side of the measuring joint, reducing conductivity errors.
Method for measuring leaf temperature. Copper and constantan wires of 0·079 mm diameter were made into a thermocouple that was threaded through adjacent veins so that the thermocouple measuring joint was pressed against the leaf. The use of very small diameter wire increased the response time and decreased radiation errors. This system is similar to one used by Drake et al. (1970) but different from the suggestion of Ehleringer (1991) who recommended that the thermocouples be inserted into the part of the leaf to be measured.
Using the system shown in Fig. 1, leaf temperature in natural conditions has been measured in oaks (Singsaas et al., 1999), aspen (R. R. Wise, Univ. Wisc.-Oshkosh, Oshkosh Wisc, unpubl. res.), cotton (Wise et al., 2004), and mosses (Hanson et al., 1999). In all cases, sunlight caused very large and rapid changes in leaf temperature. In the example shown in Fig. 2, a range of >10 °C is seen to occur throughout the day, except for three periods which correspond with clouds passing overhead. This temperature range was confirmed several times using a hand-held infrared thermometer. The finding of these very large heat flecks allowed a refinement of the thermotolerance hypothesis for isoprene emission. Specifically, isoprene synthesis (and consequent emission) protects against heat flecks. This hypothesis is consistent with the distribution of isoprene emission capacity among plant species. However, like the distribution of C4 metabolism, both environmental and phylogenetic influences can be seen and strict correlations are not the rule.
Temperature of a white oak leaf. Measurement was made at the top of a 30-m Quercus alba tree in Duke Forest, North Carolina.
The thermotolerance hypothesis is also consistent with the distribution of isoprene emission capacity through a canopy. Leaves at the top of a canopy are much more likely to suffer heat flecks and, when measured under identical conditions, leaves at the top of a canopy emit as much as four times more isoprene relative to leaves at the bottom of the canopy (Harley et al., 1996; Sharkey et al., 1996; Singsaas et al., 1999). Differences in IspS activity can account for the canopy position effect (Lehning et al., 2001).
Focusing the thermotolerance hypothesis on heat fleck damage protection made it possible to devise more targeted experimental tests. Additional improvements in experimental design were made possible by the discovery that isoprene is made by the methylerythritol 4-phosphate (MEP) pathway (Schwender et al., 1997), which is sensitive to a specific inhibitor, fosmidomycin (Kuzuyama et al., 1998; Zeidler et al., 1998). As a result, it was possible to test whether plants that had the ability to make isoprene could withstand repeated, short high temperature episodes better than plants that did not. The answer is yes, the capacity for isoprene emission confers tolerance to short high temperature episodes (Sharkey et al., 2001; Velikova and Loreto, 2005). For example, the data in Fig. 3 show that a leaf in which isoprene emission was inhibited by feeding fosmidomycin suffered more heat damage and recovered less than leaves not fed fosmidomycin (endogenous isoprene) or fed fosmidomycin but supplied with isoprene in the gas phase (exogenous isoprene). The control experiment of showing that adding back isoprene in the gas stream restores the thermotolerance of a fosmidomycin-poisoned leaf proves that nonspecific effects of fosmidomycin are not responsible for the results and makes the use of fosmidomycin a very strong experimental system.
Thermoprotection of photosynthetic capacity by isoprene. Photosynthesis of detached kudzu leaves was measured at the indicated temperatures. One leaf was fed water and so made isoprene from endogenous sources. Two other leaves were fed 4 µm fosmidomycin and isoprene emission was monitored until >90% of the isoprene emission capacity was lost. One of these leaves was then provided with 2 µL L−1 isoprene in the air stream (exogenous isoprene treatment).
In other experiments, exogenous isoprene treatment could restore all of the thermoprotection found in leaves emitting isoprene (table 1 of Sharkey et al., 2001). It was shown that this is a general effect of compounds with double bonds (alkenes) and that alkanes enhance thermal damage. Some monoterpenes can also provide thermoprotection (Delfine et al., 2000; Peñuelas and Llusià, 2002; Copolovici et al., 2005). Methyl butenol, which is related to isoprene and emitted by some pine trees (Harley et al., 1998; Gray et al., 2003), has not been tested for its ability to provide thermotolerance.
Feeding isoprene in the gas stream to leaves in which isoprene emission has been poisoned by feeding fosmidomycin, or leaves that do not normally emit isoprene, confers thermotolerance (Sharkey et al., 2001). The protection provided by isoprene can be seen even 24 h after heat stress. For example, Sharkey et al. (2001) reported that photosynthesis of Phaseolus vulgaris was reduced to 80 ± 11 % (mean ± s.e., n = 3) 24 h after a 2 min 46 °C heat spike, but if 2 µL L−1 isoprene (a physiologically relevant level for isoprene inside leaves; Singsaas et al., 1997) was provided in the gas stream during the heat spike, photosynthesis was 96 ± 1 % of the pre-stress value after 24 h. This 16 % difference in photosynthetic capacity can offset the cost of isoprene production in plants that experience such heating episodes.
Genetic engineering has allowed creation of poplar trees that lack the capacity for isoprene emission and these trees show increased damage to photosynthesis by heat spikes relative to control trees (Behnke et al., 2007). Arabidopsis plants transformed with an IspS gene from kudzu (Sharkey et al., 2005) can tolerate heat stress that kills untransformed plants (C. Barta and F. Loreto, Consiglio Nazionale delle Ricerche (CNR), Rome, unpubl. res.). Sasaki et al. (2007) report that arabidopsis expressing poplar IspS are much better able to tolerate heat stress. One the other hand, arabidopsis plants expressing an IspS gene from Populus × canescens did not show enhanced tolerance to heat spikes (Loivamäki et al., 2007a), although the assay did not result in heat spike-induced damage to wild-type plants, and the isoprene emission rate was not very much higher than the background emission from wild-type plants so it is difficult to interpret this experiment. Thus, (a) providing isoprene to plants that do not normally make isoprene, (b) using genetic approaches to induce non-emitting species to make isoprene, (c) using genetic approaches to suppress isoprene synthesis, or (d) using an inhibitor to reduce isoprene emission, all confirm that isoprene provides leaves with the ability to tolerate brief high temperature episodes.
Tolerance of heat flecks can help explain the distribution of the capacity to emit isoprene among plants. Crop plants are selected for rapid growth and this requires open stomata. High stomatal conductance allows high rates of latent heat loss, buffering against heat flecks. Therefore, crop plants should not, and generally do not, emit isoprene. Sustained high temperature presents a type of stress that isoprene emission may not help plants to tolerate. Plants from hot deserts do not emit significant amounts of isoprene. On the other hand, leaves at the tops of trees are subject to intense sunlight and the light (and associated heat gain) can vary over very short periods. Trees are generally the biggest isoprene emitters, especially oak and aspen trees. In the tropics, plant leaves can grow very large, and this creates a large boundary layer insulating the leaf from air temperature, allowing the leaf temperature to exceed air temperature by 10 °C and more. Also, in humid air, heat loss by latent heat of evaporation is reduced. The humid tropics are known to have many isoprene-emitting species (Sharkey and Yeh, 2001). Thus, there is a correspondence between the distribution of isoprene emission capacity among plant species and its presumed function in increasing tolerance of heat flecks suffered by leaves.
Reactive oxygen
A second role for isoprene is in tolerance of ozone and other reactive oxygen species (ROS). Isoprene can prevent visible damage caused by ozone exposure (Loreto and Velikova, 2001; Loreto et al., 2001) and can prevent measurable loss in photosynthetic capacity by ROS (Affek and Yakir, 2002; Peñuelas and Llusià, 2002; Velikova et al., 2004; Peñuelas et al., 2005). However, there is also a report that isoprene emission can exacerbate ozone damage (Hewitt et al., 1990). While IspS gene expression and protein amount are stimulated by high temperature, they are decreased in elevated ozone (Fares et al., 2006; Calfapietra et al., 2007). Therefore, while isoprene protects against both ROS and heat flecks, the physiology of isoprene emission appears related to the protection against heat flecks. It is unlikely that ozone levels were significant over evolutionary time (Jacob, 1999), so it may be that physiological responses to the ROS protection afforded by isoprene has not yet had time to evolve. Lerdau has pointed out that if isoprene emission can increase ozone production when NOx is present, and simultaneously help plants tolerate ozone, ecosystem composition could change as isoprene-emitting species lead to high levels of ozone that they are better able to tolerate (Lerdau, 2007).
Mechanism of isoprene action
The mechanism by which isoprene protects against heat flecks and ROS is unknown. It is tempting to speculate that the same mechanism accounts for both effects of isoprene. Velikova and Loreto (2005) showed that heat flecks caused leaves to accumulate more H2O2 and malondialdehyde (a membrane oxidation product) when isoprene emission was inhibited by fosmidomycin. It is speculated that heat damage to photosynthesis is mediated by ROS, and isoprene protects against ROS by protecting both against experimentally induced ROS and heat-induced ROS. Death of yeast cells by heat shock involves ROS (Davidson et al., 1996). However, H2O2 accumulation can be a signal for inducing gene expression that leads to stress tolerance (Kovtun et al., 2000). It can be difficult to tease apart signals of heat stress from the damage caused by heat stress.
There are significant effects of heat stress on photosynthesis not easily explained by ROS. For example, heat flecks cause more damage when given to leaves in the dark than in the light (Weis, 1982; Schrader et al., 2004). Heat can cause thylakoid membranes to become leaky and stimulate cyclic electron flow (Pastenes and Horton, 1996; Bukhov et al., 1999; Schrader et al., 2004). The cyclic electron flow maintains the proton motive force needed for ATP synthesis (Schrader et al., 2007). The loss of membrane integrity could lead to enhanced levels of malondialdehyde. Thus, it is possible that the mechanism of action of isoprene is to protect membrane integrity, and this protects against heat fleck damage and the effects of ROS. In other words, isoprene could reduce ROS by reducing heat damage directly rather than acting only through quenching of ROS generated by heating. Lui and Huang (2002) reported that cytokinin given to heat-stressed Agrostis palustris (creeping bentgrass) reduced the heat stress and ROS but it was not suggested that the mechanism was quenching the ROS. Isoprene may work through one mechanism that helps leaves tolerate heat and ROS stress or the two mechanisms may be unrelated.
The only direct study of the mechanism by which isoprene might function is that of Siwko et al. (2007). They showed that a moderate amount of isoprene dissolved in a model membrane caused an increase in membrane order equivalent to a 10 °C decrease in temperature. Siwko et al. (2007) conclude that isoprene stabilizes lipid membranes and that their experiments provide a mechanistic basis for the suitability of isoprene for protection against heat spike damage. We agree.
Other hypothesized effects of isoprene
At high concentration, isoprene was shown in one study to speed flowering in arabidopsis (Terry et al., 1995). Isoprene emission also consumes certain metabolites, and it has been proposed that this may be the function of isoprene emission, a ‘safety valve’ to get rid of unwanted metabolites (Rosenstiel et al., 2004) or energy (Sanadze, 2004). However, in both cases these functions have no predictive power; they do not explain why some plants do and some do not emit isoprene. They do not explain why isoprene emission is greater at the tops of trees than lower in the canopy. The metabolite ‘safety valve’ hypothesis, that isoprene emission allows phosphate intermediates that get ‘stuck’ in the MEP pathway, is a futile cycle; futile cycles are normally avoided in metabolism. The first enzyme in the isoprene synthesis pathway is very sensitive to feedback from metabolites further in the pathway (Wolfertz et al., 2004) which is the classic regulatory method of preventing futile cycles in a pathway of this type. The problem of insufficient phosphate turnover is common at low temperature but uncommon at high temperature (Sage and Sharkey, 1987). Therefore, there are mechanisms to prevent the accumulation of dimethylallyl diphosphate (DMADP) and the problem would be expected to be worse at low temperature, when isoprene emission is very low.
It has been hypothesized that isoprene emission can dissipate excess energy when leaves receive more light than they can use (Magel et al., 2007). However, the well-known dissipation mechanisms that give rise to energy-dependent quenching of chlorophyll fluorescence and even photorespiration consume many more times the energy consumed by isoprene emission, making this function for isoprene emission quantitatively insignificant.
Another hypothesis concerning the role of isoprene emission is the ‘opportunist hypothesis’ of Owen and Peñuelas (2005). Clearly, isoprene emission capitalizes on the opportunity to use DMADP, presented by the fact that this metabolite is needed to synthesize many other compounds. The Km of IspS for DMADP is in the millimolar range while the Km of geranyl diphosphate synthase for DMADP is in the micromolar range (Tholl et al., 2001). This difference effectively separates these two metabolic fates of DMADP. This rules out the second component of the opportunist hypothesis as put forward by Owen and Peñuelas (2005), that longer chain isoprenoids will determine the rate of isoprene emission. The opportunist hypothesis has been criticized by Pichersky et al. (2006).
REGULATION OF ISOPRENE EMISSION CAPACITY
Given the importance of isoprene in atmospheric chemistry, it is essential to understand how plants regulate their isoprene emission. Isoprene is synthesized by the action of IspS on DMADP (Silver and Fall, 1991) produced by the MEP pathway (Fig. 4) (Schwender et al., 1997).
The methylerythritol 4-phosphate pathway. G3P = glyceraldehyde 3-phosphate; DXS = deoxyxylulose 5-phosphate (DXP) synthase; DXR = DXP reductoisomerase; MEP = methylerythritol 4-phosphate; CMS = diphosphocytidylyl methylerythritol (CDP-ME) synthase; CMK = CDP-ME kinase; CDP-MEP = CDP-ME 2-phosphate; MCS = methylerythritol 2,4-cyclodiphosphate (ME-cPP) synthase; HDS = hydroxymethylbutenyl diphosphate (HMBPP) synthase; HDR = HMBPP reductase; IDP = isopentenyl diphosphate; DMADP = dimethylallyl diphosphate; IDI = IDP isomerase; IspS = isoprene synthase.
Isoprene emission is modelled as a base (or basal) rate corrected for differences between the conditions of the measurement and the conditions used to determine the basal rate (normally 30 °C and 1000 µmol m−2 s−1). The basal rate was considered to be fairly constant once leaves were fully developed although species dependent (Guenther et al., 1993, 1995). The instantaneous response to temperature was found to be similar among species and in different environments, but the basal rate has turned out to vary considerably. It is well known now that the capacity for isoprene emission is delayed developmentally, with leaves becoming photosynthetically competent as much as weeks before isoprene emission begins (Sharkey and Loreto, 1993; Monson et al., 1994; Goldstein et al., 1998; Kuhn et al., 2004; Mayrhofer et al., 2005; Wiberley et al., 2005). The delay is significantly affected by growth temperature (Monson et al., 1994; Wiberley et al., 2005). Even after leaves are fully developed, air temperature of the previous few hours to weeks affects the base rate of isoprene emission (Goldstein et al., 1998; Fuentes and Wang, 1999; Fuentes et al., 1999; Sharkey et al., 1999; Pétron et al., 2001). Changes in the activity of IspS can be seen in response to temperature of the previous few days (Lehning et al., 2001).
Isoprene emission is remarkably resistant to water stress (Tingey et al., 1981). Water stress that causes nearly complete loss of photosynthetic capacity has only a minor effect on isoprene emission (Tingey et al., 1981; Sharkey and Loreto, 1993; Fang et al., 1996; Pegoraro et al., 2004b; Funk et al., 2005; Monson et al., 2007). Following re-watering, isoprene emission capacity sometimes exceeds the capacity before the stress (Sharkey and Loreto, 1993; Brilli et al., 2007). Isoprene synthase activity is quite robust in response to water stress (Brüggemann and Schnitzler, 2002a; Brilli et al., 2007). The maintenance of isoprene emission and stimulation by water stress can be interpreted as adaptive in light of the thermotolerance hypothesis, since water stress is likely to lead to more frequent heat stress as latent heat loss is reduced with reduced water availability.
Isoprene emission is reduced when plants are grown under elevated ozone and the expression of IspS can be shown to be reduced (Fares et al., 2006; Calfapietra et al., 2007). It is not clear why isoprene emission would be reduced in plants grown in elevated ozone if the adaptive significance of isoprene emission were quenching ozone. Evolution may have resulted in isoprene emission as a mechanism of thermoprotection with the happy consequence that leaves that emit isoprene are also protected against the very recent stress of ozone.
The control of isoprene emission by temperature, water stress, and elevated CO2 or ozone will rest with regulation of IspS and regulation of the supply of DMADP. It was originally assumed that the mevalonic acid pathway was the source of substrate (Sharkey et al., 1991; Sanadze, 2004), but it was demonstrated that a newly discovered pathway for making isoprenoids was the real source (Zeidler et al., 1997). The MEP pathway supplies plastids with DMADP (Schwender et al., 2001). There is evidence for crosstalk between the mevalonic acid pathway in the cytosol and the MEP pathway in chloroplasts (Laule et al., 2003; Dudareva et al., 2005) but movement of substrates from the chloroplast to the cytosol has been demonstrated more often than movement in the other direction. Plants that lack the first enzyme of the MEP pathway are not viable (Estévez et al., 2001) and fosmidomycin, the inhibitor of the second enzyme in the pathway completely (Sharkey et al., 2001) or nearly completely (Loreto et al., 2004) eliminates isoprene emission. The elimination of isoprene emission by fosmidomycin is not consistent with the hypothesis of Sanadze (2004) that a second carboxylation system exists. Therefore, understanding the regulation of the rate of isoprene emission requires understanding the regulation of IspS and the MEP pathway. Molecular tools have become available for dissecting the control of isoprene emission, and these studies plus other biochemical studies of isoprene synthesis regulation are discussed below.
Isoprene synthase regulation
Since the initial discovery of IspS, the link between IspS and isoprene emission has been studied extensively. Several studies have indicated that extractable IspS activity correlates with isoprene emission (Monson et al., 1992; Kuzma and Fall, 1993; Schnitzler et al., 1996, 1997, 2005; Lehning et al., 1999; Brüggemann and Schnitzler, 2002a; Mayrhofer et al., 2005). Recent molecular studies have shown that introduction of an IspS gene into arabidopsis is sufficient to cause the plant to emit isoprene (Sharkey et al., 2005; Loivamäki et al., 2007a; Sasaki et al., 2007).
Thus far, studies of the regulation of IspS have shown that, during leaf development, the onset of isoprene emission is controlled by IspS transcription or mRNA turnover (Mayrhofer et al., 2005; Wiberley et al., 2005). More recently, the same has been shown in developing Populus trichocarpa leaves. The study of emissions from developing and mature leaves is simplified in P. trichocarpa because young stems flush continuously, so an entire series of leaves, from just-emerged to weeks past full expansion, can be studied on one stem. When grown at high temperature, leaves begin to emit isoprene at least 1 week after acquisition of photosynthetic competence, and IspS mRNA and protein begin to accumulate at the same developmental stage. The same transcriptional regulation is observed in leaves growing at low temperature, but such leaves do not begin to emit or accumulate IspS mRNA or protein until several days later than their high-temperature counterparts (Fig. 5) (experimental procedures were as described in Wiberley et al., 2005). This is useful in the creation of mechanistic models of isoprene emission: the amount of isoprene that leaves will emit early in their lives, and how soon they begin to emit, are functions of the temperature at which they develop.
Isoprene emission and photosynthesis rates, and IspS mRNA and protein levels for developing P. trichocarpa leaves. Emission rates were measured at 30 °C and 1000 µmol m−2 s−1 light. Experimental methods were similar to those reported in Wiberley et al. (2005) for kudzu.
Isoprene synthase has a high Km for DMADP and in some cases exhibits sigmoidal kinetics (Sharkey et al., 2005; Schnitzler et al., 2005). As a result, it is easy for isoprene emission to be co-regulated by both the enzyme amount and the substrate amount. There is no direct evidence for post-translational regulation of the activity of IspS but no evidence ruling it out either. In some species, IspS can be found in the soluble fraction and in membrane fractions and the amount in each fraction can vary (Wildermuth et al., 1996; Wiberley et al., 2005). The mechanisms for this and its possible role in regulating isoprene emission are not yet known.
MEP pathway regulation
The MEP pathway provides substrate for the synthesis of numerous terpenoids in addition to isoprene, and has been implicated in regulation of their synthesis. Deoxyxylulose 5-phosphate synthase (DXS), DXP reductoisomerase (DXR) and hydroxymethylbutenyl diphosphate reductase (HDR) have had regulatory roles suggested in production of terpenoids such as carotenoids (Albrecht and Sandmann, 1994; Sun et al., 1998; Lois et al., 2000; Estévez et al., 2001; Carretero-Paulet et al., 2002, 2006; Guevara-Garcia et al., 2005; Muñoz-Bertomeu et al., 2006). Some evidence has linked deoxyxylulose 5-phosphate synthase and isopentenyl diphosphate isomerase to regulation of isoprene emission as well (Brüggemann and Schnitzler, 2002b; Wolfertz et al., 2003, 2004). Given these results and observations of isoprene emission capacities of mature leaves subjected to temperature changes, the role of the MEP pathway in regulation of isoprene emission bears further investigation.
Molecular regulation
Populus trichocarpa is well suited to studies of the molecular regulation of isoprene emission: in addition to producing a continual supply of new leaf tissue and being easy to propagate, its genome has been completely sequenced (Tuskan et al., 2006) making it ideal for molecular studies.
Another advantage of having a sequenced genome available is the ability to do preliminary in silico studies that may indicate regulatory factors testable in vitro. For example, the promoter regions of the MEP pathway genes and IspS have been analysed to identify potential testable regulatory elements. The sequences of these regions (2000 nucleotides upstream of the start codon) were found on the P. trichocarpa genome website (http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home.html) and analysed with PLACE (Higo et al., 1999) and PlantCARE (Lescot et al., 2002), which search DNA sequences for transcription factor binding sites.
Isoprene emission is regulated by heat and exhibits circadian regulation (Wilkinson et al., 2006; Loivamäki et al., 2007b), so a search of the promoters of MEP pathway genes for transcription factor binding sites related to these responses was performed. In Chlamydomonas it has been found that the promoters of heat-shock genes often contain heat-responsive elements within about 300 nucleotides of their start codons (von Gromoff et al., 2006); such genes in arabidopsis and soybean also frequently have a pair of similar elements repeated at least twice, about 20 nucleotides upstream of their transcription start site, and the response of these elements may be increased if they are within 10–30 nucleotides downstream of a CCAAT box (Gurley and Key, 1991; Haralampidis et al., 2002). According to PLACE and PlantCARE analyses, the promoters of most of the MEP pathway genes contain at least one of these elements, but only DXS, methylerythritol 2,4-cyclodiphosphate synthase (MCS) and IspS have two heat shock-response sites between a CCAAT box and putative transcription start site. Only diphosphocytidylyl methylerythritol synthase (CMS), isopentenyl diphosphate isomerase (IDI) and IspS have Chlamydomonas-like heat-responsive elements within a few hundred nucleotides of their start codons. Based on these analyses and consistent with studies described above, IspS is an especially strong candidate for heat-induced expression.
The promoters of some circadian-regulated genes in tomato have a ‘CAA(N)4ATC’ motif within about 300 nucleotides of their start codons (Piechulla et al., 2001); so do poplar DXS, CMS, MCS, HDS and IspS. IspS also has the CCA1/LHY-binding motif (A)5TCT, which controls dawn-phased expression (reviewed by Hotta et al., 2007). A conserved ‘TATTCT’ ten nucleotides upstream of the transcription start site in barley light-responsive genes has been shown to be important in circadian regulation (Thum et al., 2001). This sequence is also found in the proper position in the promoters of poplar DXS, CMS, MCS, HDS, HDR and IspS. In addition, the poplar DXS promoter contains a series of repeated GATA boxes with spacing and position similar to those required for circadian regulation of some arabidopsis genes (Anderson et al., 1994), and the IspS promoter may, also. These findings are consistent with observations to date on the circadian regulation of isoprene emission: isoprene emission and IspS transcript levels show strong circadian rhythms (Wilkinson et al., 2006; Loivamäki et al., 2007b), while DXR transcript levels do not (Mayrhofer et al., 2005). The role of the LHY motif for the IspS promoter was confirmed by an electrophoretic mobility shift assay (Loivamäki et al., 2007b). In both of the cases examined here, DXS and IspS show strong potential for important regulation. These in silico analyses help pinpoint genes that may be key in regulation of isoprene emission under varying conditions, identifying suitable targets for further in vivo and in vitro work.
Regulation through energetics
The MEP pathway requires a significant amount of reducing power and ATP, and this could link regulation of the pathway with photosynthesis. There usually is no emission of isoprene from leaves in darkness. Correlations have been found between leaf ATP content and isoprene emission rate (Loreto and Sharkey, 1993). Models of isoprene emission rate often rely on predictions of photosynthetic electron transport rates to predict isoprene emission rate (Niinemets et al., 1999; Martin et al., 2000; Zimmer et al., 2000).
There are three redox reactions in the MEP pathway. DXR is known to use NADPH, readily available during photosynthetic electron transport. In plants, the HDS enzyme can use electrons transferred directly from the electron transport chain through ferredoxin (Seemann et al., 2006) further connecting MEP pathway activity to photosynthesis. The HDR reductant has not been identified (Seemann et al., 2002). One ATP and one CTP are required in the MEP pathway. The CTP loses two phosphates, and so, presuming the CTP is regenerated by ATP, the total ATP cost is three. If isoprene emission is typically 2 % of photosynthesis on a carbon basis (Sharkey and Yeh, 2001), and there are five carbons per isoprene, the ATP used in the MEP pathway for isoprene synthesis is 0·4 % of that being used for carbon fixation; for reducing power, 0·6 % of that used for carbon fixation is used for MEP pathway reactions. If all of the energy needed to reduce CO2 to sugars is included in the cost of isoprene the totals rise to 2·7 % of ATP and 3·4 % of NADPH is required for isoprene emission at a rate of 2 % of photosynthesis on a carbon basis, not including photorespiration. While these amounts of energy use could lead to loss of the capacity for isoprene emission through evolution if isoprene emission had no value to the plant, these are trivial amounts of energy compared with, for example, photorespiration, where 20–40 % of the total ATP and NADPH usage can be used (Sharkey, 1988). This is why the suggestion that one function of isoprene emission is to dissipate unused energy (Magel et al., 2007) does not hold up under quantitative scrutiny. Because of the very small proportion of energy used by the MEP pathway under normal conditions, the control of the MEP pathway by energetics of the chloroplast is likely to be regulatory rather than by mass action effects. Thus, competition between carbon fixation and the MEP pathway for energy is less likely to be a useful predictor of isoprene emission than is energy status of the chloroplast, which can be unrelated to electron transport rates. Of course, when isoprene emission increases to a large proportion of the carbon fixed the energy cost increases.
Regulation by carbon supply
The MEP pathway draws on the Calvin cycle for carbon. Feeding 13CO2 to leaves results in a rapid appearance of 13C in isoprene (Sanadze et al., 1972; Delwiche and Sharkey, 1993; Loreto et al., 1996; Karl et al., 2002; Affek and Yakir, 2003; Loreto et al., 2004). Similar results have been obtained using 11CO2 (Funk et al., 2004). One of the puzzling findings has been that isoprene does not become completely labelled. In oak and poplar trees it was shown that sugar arriving in the transpiration stream can contribute carbon to isoprene (Kreuzwieser et al., 2002). During water stress, as the availability of carbon in the Calvin cycle becomes limited, more carbon comes from other sources (Brilli et al., 2007). However, Delwiche and Sharkey (1993) pointed out that the first carbon product of photosynthesis, phosphoglyceric acid, shows similar incomplete labelling. Thus, the incomplete labelling may be a general phenomenon related to the availability of carbon within the chloroplast and may not have special significance to isoprene synthesis. The incomplete labelling need not indicate a decoupling between plastid carbon metabolism and isoprene synthesis. It is tempting to assume that the incomplete labelling of isoprene reflects the cytosolic source of pyruvate [through phosphoenolpyruvate (PEP) import into chloroplasts]. However, this has not been directly demonstrated, and given incomplete labelling of PGA, the explanation that incomplete labelling of isoprene is caused by the pyruvate source in the cytosol should be viewed with caution.
Isoprene emission capacity is reduced at high CO2 (Rosenstiel et al., 2003; Centritto et al., 2004; Pegoraro et al., 2004a, b, 2005a; Scholefield et al., 2004). It is not easy to see an adaptive explanation for this response based on the thermotolerance hypothesis, since high CO2 should lead to stomatal closure and increased leaf temperature. Rosenstiel et al. (2003) proposed and presented evidence for a mechanistic explanation. They showed that PEP carboxylase competed for substrate with isoprene emission. Inhibiting PEP carboxylase stimulated isoprene emission (Rosenstiel et al., 2003, 2004). Loreto et al. (2007) found a positive relationship between dark respiration and isoprene emission. This is inconsistent with mitochondrial activity competing with isoprene emission for PEP during the day, and this could indicate that mitochondrial respiration during the day is low or uses substrates other than pyruvate derived from PEP. In their experiments there was a negative correlation between isoprene emission and PEP carboxylase activity, consistent with the hypothesis of competition between PEP carboxylation and isoprene synthesis (Loreto et al., 2007).
On the other hand, Wolfertz et al. (2004) showed that isoprene emission was strongly controlled by the activity of DXS. Aspen grown in elevated CO2 have reduced amounts of IspS, which partially accounts for a long-term reduction in isoprene emission capacity (Calfapietra et al., 2007). This and other data make it clear that there can be several factors controlling the rate of isoprene emission simultaneously. This makes modelling isoprene emission potentially more complex than modelling photosynthesis, where either Rubisco activity or ribulose bisphosphate regeneration dominate the control of the rate of photosynthesis at any given instant (Farquhar et al., 1980). The effect of elevated CO2 on isoprene emission may provide some insights into the control of isoprene emission and has implications for predicting global change effects on the atmosphere (Monson et al., 2007; Possell et al., 2005).
SYNTHESIS
Isoprene emission from plants is an unseen but highly significant component of atmosphere–biosphere interaction. Because it is possible for plants to survive without significant isoprene production and emission, we presume that those plants that do emit derive some benefit that outweighs the cost. Thermotolerance has significant explanatory power and experimental support. Inhibitor studies and genetic approaches have confirmed that thermotolerance is provided by isoprene. Isoprene-emitting plants are also protected against ozone but, given that significant ozone stress is a recent phenomenon, it may be that this is simply a happy coincidence. The regulation of the rate of isoprene emission should reflect the benefits derived from isoprene. It is not surprising, then, that temperature regulates isoprene emission at many different levels. Evidence for regulation at the level of gene transcription has been found but, more generally, the regulation has been difficult to understand. This is in part because there is a significant amount of DMADP in leaves in a compartment that is not accessible to IspS. While isoprene emission is significant at approx. 2 % of photosynthesis, it is hard to do detailed analyses of carbon flux regulation on a background activity 50 times greater than the process under study. Molecular tools are being developed and the use of stable isotopes has led to significant insights into isoprene emission rate regulation. Substantial progress is likely in the coming years. We should be able to answer with increasing depth ‘how and why plants emit isoprene’.
ACKOWLEDGEMENTS
Original research described here was supported by US National Science Foundation grants IBN-0212204 and IOB-0640853. We thank Jon Steichen for preparing Fig. 4.
Comments
We welcome Young’s clarification of the atmospheric chemistry of isoprene. Although we spent some time trying to put together a scheme that was both correct and accessible to the readers of Ann. Bot., it may well be that we got some of the atmospheric chemistry wrong. One area of agreement is that isoprene emission from plants has a significant effect on atmospheric chemistry and it will be very useful to know the biology of this phenomenon as one part of a comprehensive understanding of isoprene.
Thomas D. Sharkey, Amy E. Wiberley, and Autumn R. Donohue, Dept. Biochem. Mol. Biol., Michigan State University, East Lansing, MI 48824 USA (tsharkey@msu.edu) (note new contact information)
Conflict of Interest:
None declared
Sir,
As an atmospheric chemist working on the impact of biogenic emissions, I very much welcomed this invited review on isoprene emission from plants, particularly the opportunity to read a summary of the advances in the understanding of the biological aspects of isorpene emission (phylogeny, biosynthesis etc.) that are often absent from literature in my field. However, I feel it is necessary to clarify the description of tropospheric chemistry and isoprene oxidation presented by the authors.
In their "Why Isoprene Emissions Matter" section, I am not clear whether the authors are stating that the source of OH (the most important tropospheric oxidizing agent) is from conversion of HO2 via reaction with NO. In fact, the major source of OH is from a photolysis channel of ozone that produces high energy oxygen atoms (O1D), followed by reaction with water vapour, such:
1. O3 + hv(lamda < 320nm) -> O2 + O1D
2. O1D + H2O -> OH + OH
Globally, isoprene tends to reduce OH concentrations, compared to no emissions (e.g. Spivakovsky et al., 2000), as it provides a direct sink. Thus, as the OH concentration largely determines the methane lifetime, isoprene can be assigned a global warming potential (Collins et al., 2002); i.e. its emission is relevant to climate as well as air quality.
Furthermore, the authors state that isoprene oxidation proceeds via H -atom abstraction by an OH radical (forming H2O). As isoprene is a (di- )alkene (2 sets of C=C bonds), oxidation by OH proceeds predominantly via an addition mechanism (see Wayne, 2000, pp333-334). Illustrating with the most simple alkene, ethene:
1. CH2=CH2 + OH (+M) -> .CH2-CH2OH (+M)
2. .CH2=CH2OH + O2 -> CH2(OO)-CH2OH
where M is N2 or O2 and CH2(OO)-CH2OH is equivalent to the RO2 peroxy radical described by Sharkey et al.
It is the presence of 2 C=C bonds that make isoprene so reactive in the atmosphere, enabling reactions with ozone and NO3 (produced at nighttime from NO2 + O3) as well as OH. The asymmetry of isoprene results in a great array of oxidation products, some of which retain a C=C bond and are thus still very reactive (e.g. methacrolein, methyl vinyl ketone); see Atkinson & Arey (2003) (and refs. therein) for review of the gas- phase chemistry.
Those further interested in the atmospheric chemistry of isoprene and its impacts should also see the following modelling articles: von Kuhlmann et al. (2004), Fiore et al. (2005), Horowitz et al. (2007), Zeng et al., (2008).
Whilst these details may not be of interest to all this Journal's readership, I feel it is important to correct the mistake in an article that addresses an interdisciplinary area.
Yours faithfully,
Paul Young
Centre for Atmospheric Science, University of Cambridge, UK
References:
Atkinson, R. & J. Arey (2003), Gas-phase tropospheric chemistry of biogenic volatile organic compounds: a review, Atmos. Environ., 37, S197-S219.
Collins, W. J. et al. (2002), The oxidation of organic compounds in the troposphere and their global warming potentials, Clim. Change, 52, 453 -479.
Fiore, A. M. et al. (2005), Evaluating the contribution of changes in isoprene emissions to surface ozone trends over the eastern United States, J. Geophys. Res., 110, D12303, doi:10.1029/2004JD005485.
Horowitz, L. W. et al. (2007), Observational constraints on the chemistry of isoprene nitrates over the eastern United States, J. Geophys. Res., 112, D12S08, doi:10.1029/2006JD007747.
Spivakovsky, C. et al. (2000), Three-dimensional climatological distribution of tropospheric OH: Update and evaluation, J. Geophys. Res., 105, 8931--8980.
von Kuhlmann, R. et al. (2004), Sensitivities in the global scale modelling of isoprene, Atmos. Chem. Phys., 4, 1-17.
Wayne, R. P. (2000), Chemistry of atmospheres, 3rd ed., OUP, Oxford, UK.
Zeng, G. et al. (2008), Impact of climate change on tropospheric ozone and its global budgets, Atmos. Chem. Phys., 8, 369-387.
Conflict of Interest:
None declared