Circadian control of isoprene emissions from oil palm (Elaeis guineensis)

Authors


For correspondence (fax +44 1524 593985; e-mail n.hewitt@lancaster.ac.uk).

Summary

The emission of isoprene from the biosphere to the atmosphere has a profound effect on the Earth's atmospheric system. Until now, it has been assumed that the primary short-term controls on isoprene emission are photosynthetically active radiation and temperature. Here we show that isoprene emissions from a tropical tree (oil palm, Elaeis guineensis) are under strong circadian control, and that the circadian clock is potentially able to gate light-induced isoprene emissions. These rhythms are robustly temperature compensated with isoprene emissions still under circadian control at 38°C. This is well beyond the acknowledged temperature range of all previously described circadian phenomena in plants. Furthermore, rhythmic expression of LHY/CCA1, a genetic component of the central clock in Arabidopsis thaliana, is still maintained at these elevated temperatures in oil palm. Maintenance of the CCA1/LHY-TOC1 molecular oscillator at these temperatures in oil palm allows for the possibility that this system is involved in the control of isoprene emission rhythms. This study contradicts the accepted theory that isoprene emissions are primarily light-induced.

Introduction

Emissions of the volatile organic compound isoprene (2-methyl-1,3-butadiene, C5H8) from the biosphere affect the oxidative capacity of the atmosphere (Shallcross and Monks, 2000) and hence have the potential to modify regional ozone pollution (Pierce et al., 1998); lead to secondary aerosol formation (Claeys et al., 2004); and indirectly control the lifetime of the greenhouse gas methane (Poisson et al., 2000). Isoprene is synthesized in the chloroplast, mainly from recently assimilated carbon (Karl et al., 2002), and although synthesis is loosely coupled to photosynthetic processes through substrate availability and requirements for adenosine triphosphate, its emission may become uncoupled from photosynthesis, particularly at elevated temperatures (Harley et al., 1999). On time scales of minutes to hours, the primary controls that are known to affect the isoprene emission rate (IE) are photosynthetically active radiation (PAR) and leaf temperature (Harley et al., 1999), which regulate substrate availability and enzyme kinetics (Sharkey et al., 2001). Isoprene is emitted through stomata, but changes in stomatal aperture have negligible effects on its emission due to its high volatility and low solubility in water (Niinemets and Reichstein, 2003a). In the longer term, carbon dioxide concentration (Possell et al., 2004, 2005), leaf phenology (Kuzma and Fall, 1993), leaf nitrogen content (Harley et al., 1994; Rosenstiel et al., 2004), and environmental stresses such as soil water potential (Pegoraro et al., 2004) and herbivory (Funk et al., 1999) also affect the magnitude of IE.

To compile emission inventories for this important compound, branch- and leaf-enclosure techniques are commonly employed to determine IE for individual plant species at 1000 μmol m−2 sec−1 (PAR) and 30°C leaf temperature. Here we define this parameter as IE1000/30. An estimate of IE at ambient PAR and temperature throughout the day for each species is obtained by scaling IE1000/30 using light- and temperature-correction factors (Guenther et al., 1995). Species composition and density are then used to estimate a regional average emission flux. However, there is high variability in IE1000/30 measurements with considerable day-to-day, seasonal, and inter- and intra-species variability (http://www.acd.ucar.edu:8080/voc/vocIndex.jsp; http://www.es.lancs.ac.uk/cnhgroup/iso-emissions.pdf). Significant diurnal variability in the magnitude of IE1000/30 has also been reported for several tree species. Isoprene emission rates from cottonwood and oak, standardized to IE1000/30, were observed to be significantly higher in the afternoon than in the morning (Funk et al., 2003). Yet in the same study, no variability in IE1000/30 for eucalyptus was observed. In other studies, IE1000/30 for oak has been shown to increase over a day by as much as 50%; this increase could be predicted using changes in weather conditions over the previous 24 h (Geron et al., 2000; Sharkey et al., 1999). However, some of the variability in IE1000/30 cannot be explained by present understanding of the effects of short-term environmental and biological controls on isoprene synthesis. As well as changes in IE1000/30 over the course of a day, significant changes have been observed in the concentration of the isoprene precursor dimethylallyl diphosphate (Geron et al., 2000; Rosenstiel et al., 2004; Wolfertz et al., 2003) but without a link being made to possible circadian control.

A wide range of fundamental processes in plant biology are optimized and coordinated each day due to regulation by an endogenous circadian clock. Examples include leaf movements, hypocotyl elongation, gene transcript and protein-abundance rhythms, and the photoperiodic control of flowering time (McClung, 2001). A number of the genes that constitute the molecular components of the positive–negative autoregulatory feedback loops that comprise the plant circadian clock have been identified in Arabidopsis thaliana. CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY), and the pseudo-response regulator/CONSTANS-domain protein TIMING OF CAB EXPRESSION1 (TOC1), form at least one of the interlocking loops at the core of the plant clock (Alabadi et al., 2001). CCA1 and LHY transcript abundance is under circadian control and peaks at dawn (Schaffer et al., 1998; Wang and Tobin, 1998). Orthologous clock genes have been characterized in plant species as diverse as rice (Oryza sativa; Izawa et al., 2003), common ice plant (Mesembryanthemum crystallinum; Boxall et al., 2005), bean (Phaseolus vulgaris; Kaldis et al., 2003) and chestnut (Castanea sativa; Ramos et al., 2005), suggesting that the genetic components of the clock are broadly conserved in higher plants. In Arabidopsis, all known clock-controlled processes have been shown to be ultimately responding to the central CCA1/LHY-TOC1 oscillator, including gas exchange, leaf movements, gene-expression rhythms and flowering time (Dodd et al., 2005). The CCA1 and LHY proteins belong to the Myb family of transcription factors (Schaffer et al., 1998; Wang and Tobin, 1998) and thus act at the transcriptional level to control gene expression.

Here we report that isoprene emission from the fronds of oil palm trees is under the control of a circadian clock. Circadian rhythms of isoprene emission persist for several days at constant light and temperature, are robust and temperature-compensated between 25 and 38°C, and are gated in a circadian fashion. Using an oil palm CCA1/LHY orthologue as a marker, we establish that the molecular clock also operates in oil palm trees up to 38°C, and thus that this clock is potentially capable of coordinating isoprene emission rhythms.

Results

Isoprene emission in oil palm is under circadian control

As noted above, diurnal variability in IE, standardized to IE1000/30, has been observed previously (Funk et al., 2003; Geron et al., 2000; Sharkey et al., 1999). We have confirmed this using field measurements of IE1000/30 from oil palm growing on a plantation in Malaysia, which showed a strong diurnal trend (Figure 1a). This trend was also observed in laboratory-grown plants under controlled conditions (Figure 1b). To determine whether this short-term variability in IE1000/30 is caused by circadian regulation, we monitored IE1000/30 from a frondlet under constant illumination (LL) for 70 h at 1000 μmol m−2 sec−1 and constant 30°C leaf temperature. If isoprene synthesis is under circadian control, we would expect to observe free-running rhythms in IE1000/30 over a period of several days. The frondlet was illuminated from dark at 06:00 (normal ‘lights on’). The IE1000/30 increased for 4 h, reaching a maximum (IEmax) of approximately 40 nmol C m−2 sec−1, before declining over the rest of the day (Figure 2a). IE1000/30 fell during subjective night (when the frondlet is not normally illuminated) to approximately 10% of IEmax before rising again the following subjective morning. Over the following 48 h, oscillations in IE1000/30 continued, but with a dampening of IEmax. The period of oscillation in this free-running cycle was 23.1 h. The persistence of rhythms in the absence of a dark–light cycle or other exogenous signals clearly indicates that isoprene synthesis is regulated by an internal biological clock.

Figure 1.

 Variability in isoprene IE at 1000 μmol m−2 sec−1 and 30°C leaf temperature (IE1000/30).
(a) Field measurements of IE1000/30 from oil palm taken in an oil palm plantation in Sabah, Malaysia. PAR and temperature were held constant at 1000 μmol m−2 sec−1 and 30°C from 10:00 to 16:00 h.
(b) IE1000/30 measured in the laboratory. PAR and temperature were held constant at 1000 μmol m−2 sec−1 and 30°C from 10:00 to 18:00 h.

Figure 2.

 Whole frond- and frondlet-level gas exchange from oil palm. Leaf-level free-running isoprene emission rates (IE) (a–c) and assimilation rates (A) (top trace) and stomatal conductance (gs) (bottom trace) (d–f) from oil palm measured in constant light (LL; 1000 μmol m−2 sec−1) for 70 h at leaf temperature of 30 (a, d); 25 (b, e); and 38°C (c, f). Full-frond IE measured in LL (120 μmol m−2 sec−1) at 25 and 38°C shown in (g, h) respectively. (b) IE values have been multiplied by 10.

To investigate the effects of different leaf temperatures on these free-running rhythms in IE, we repeated the experiment on two frondlets held at leaf temperatures of 25°C (IE1000/25) and 38°C (IE1000/38; Figure 2b,c, respectively). At these temperatures isoprene emissions oscillated over several days, with the magnitude and amplitude of oscillations dependent on leaf temperature and correlated positively with these parameters. The increase in absolute emission rates with increasing temperature is consistent with present understanding of the effects of temperature on isoprene synthase activity (Sharkey et al., 2001). As well as IE increasing with temperature, Figure 2(a–c) shows that the amplitude of oscillations in IE increased at elevated temperatures when the central components of the plant circadian clock are believed to become arrhythmic. The period of the IE rhythm was 25.5 h at 25°C and 21.4 h at 38°C, revealing that the rhythm shows significant temperature compensation. The temperature quotient (Q10) for IE across the temperature range 25–38°C was 0.89, while that for stomatal conductance (gs) was 1.0. This compares well with the Q10 values for leaf movement rhythms across a number of accessions of Arabidopsis, which varied from 0.95 to 1.12 (Edwards et al., 2005).

The robust circadian rhythm in IE at all three temperatures is not reflected in the observed assimilation rates (A): using a mathematical analysis of the data (fast Fourier transform–non-linear least squares analysis, FFT–NLLS), no circadian rhythms were identified in A (Figure 2d–f). This suggests a total uncoupling of IE from photosynthesis. Rhythmicity was observed in gs (Figure 2d–f), but the period length differed from that of IE such that the two parameters were not in phase with one another (Table 1). These results confirm previous findings showing that IE is independent of gs (Fall, 1999; Niinemets and Reichstein, 2003a,b).

Table 1.   Length of period in IE and gs at varying temperatures
MeasurementLeaf temperature (°C)Period (h) CI × 0.5RAE
  1. Period length estimates were generated using a fast Fourier transform–non-linear least squares analysis programme.

  2. CI, confidence interval; RAE, relative amplitude error, a measure of rhythm robustness varying from 0 (a perfect fit to cosine wave) to 1 (not statistically significant).

  3. IE was measured every 45 sec, giving around 20 000 data points for each period calculation. gs was measured every 2 min.

IE2525.50.40.14
3023.10.50.21
3821.40.50.19
gs2521.20.20.10
3026.70.40.11
3822.10.60.16

To eliminate the possibility of cross-talk signalling between neighbouring cells outside the gas-exchange cuvette, which were at a different temperature from the measured cells inside the cuvette, whole-frond measurements of IE were made using excised fronds, with the whole frond kept at a constant temperature. Isoprene emissions from these fronds were measured at leaf temperatures of 25 and 38°C at light intensity of 120 μmol m−2 sec−1 (IE120/25 and IE120/38; Figure 2g,h, respectively). At 25°C (Figure 2g), whole-frond isoprene IE120/25 oscillated with a period similar to that seen in the leaf cuvette (Figure 2b). At 38°C (Figure 2h), the periodicity in IE120/38 was seen for 48 h, after which wilting of the excised frond occurred and IE became arrhythmic. From these whole-frond measurements at both 25 and 38°C, we conclude that no cross-talk signalling occurred between neighbouring cells during the cuvette experiments.

The oil palm molecular clock maintains function at high temperatures

The robust circadian rhythm of isoprene emissions at 38°C (Figure 2c,h) indicates that a circadian clock must be operating in oil palm at a significantly higher temperature than any previously recorded plant circadian rhythm. For example, the circadian rhythm of CO2 fixation in the tropical crassulacean acid metabolism (CAM) species Kalanchoë daigremontiana becomes arrhythmic above 32°C, even in plants that were pre-adapted to a daytime temperature of 34°C for 1 month prior to analysis (Grams et al., 1995). It was therefore important to examine the operation of the molecular clock in oil palm at 38°C to determine whether the molecular oscillator was operating and therefore capable of controlling the observed IE rhythm. We identified an oil palm expressed sequence tag (EST), which we called EgCCA1/LHY, which is an orthologue of the Arabidopsis central clock components CCA1 and LHY (see Experimental procedures). We collected frondlet samples every 6 h from excised oil palm fronds maintained in LL at both 25 and 38°C using a whole-frond chamber. We isolated RNA from each frondlet sample and used a real-time polymerase chain reaction (RT–PCR) technique to examine changes in the steady-state transcript abundance of EgCCA1/LHY (Figure 3). This analysis revealed robust rhythms in EgCCA1/LHY transcript abundance at both 25 and 38°C, supporting the hypothesis that IE could be controlled by the CCA1/LHY-TOC1 central molecular clock. It is noteworthy that the early phase of the EgCCA1/LHY rhythm at 38°C compared with that at 25°C (Figure 3) would be consistent with a shortening of the period of EgCCA1/LHY expression at 38°C. This would correlate with the shortening of the period of the IE rhythm at 38°C relative to 25°C (Figure 2a–c).

Figure 3.

 Steady-state changes in mRNA transcript abundance of EgCCA1/LHY at leaf temperatures of 25°C (bsl00043) and 38°C (•) under constant illumination.

Isoprene emissions are circadian-gated

Circadian clock outputs are often gated such that they can only be activated at a specific phase in the 24-h cycle. For example, the circadian rhythm of nocturnal CO2 fixation in the CAM species Kalanchoë fedtschenkoi is gated in response to temperature (Hartwell et al., 2002). When the temperature was reduced from the non-permissive 30°C to the permissive 15°C during the subjective light period, no CO2 fixation was observed in this species, but when the temperature was reduced to 15°C during the subjective dark period, CO2 fixation was initiated. The IE has previously been assumed to respond directly to light, with zero emission in the dark, that is, emissions will occur only if the frondlet is illuminated. However, if the induction of isoprene is gated by the circadian clock, we would predict that emissions should occur only when a frondlet is illuminated during the subjective light period. Illumination of a frondlet at the beginning of the subjective dark period should not give rise to isoprene emissions until the subsequent subjective light period.

We investigated this by keeping two leaves in the dark at the end of a subjective day: one for 24 h, the other for 36 h. The frondlet kept in the dark for 24 h was illuminated with 1000 μmol m−2 sec−1 at 18:00, at the onset of subjective dark, at a constant leaf temperature of 30°C. Following illumination, IE1000/30 was essentially zero until the start of the next subjective day, at which time it increased significantly (Figure 4, black line). This demonstrates an uncoupling of IE from the light stimulus, and is in contrast to all previous assumptions about the light dependency of IE. Continued illumination of the frondlet, for a further 24 h, resulted in diurnal oscillations in IE1000/30 with a period of approximately 24 h. During the second subjective night, IE1000/30 did not fall to zero but decreased to approximately 20% of the preceding IEmax before increasing during the following subjective day. The frondlet that was kept in the dark for 36 h (Figure 4, open circles) was illuminated at the start of the subjective day. IE1000/30 was zero for the first hour after the frondlet was illuminated, then increased significantly, reaching IEmax at 09:40 before starting to decline. This circadian gating of isoprene emission is similar to the gating described for the acute induction of the chlorophyll a/b-binding protein (CAB) by light in A. thaliana (Millar and Kay, 1996). Figure 4 also shows that the time of day when IEmax was achieved for the frondlet illuminated at 18:00 (black line) was delayed by 9 h when compared with the frondlet illuminated at 06:00 (open circles). This suggests there was a shift in phase of the circadian rhythm due to the frondlet being illuminated at the beginning of a subjective dark period, 12 h out of phase with the prior light entrainment of the frondlet. These results demonstrate that the isoprene emissions are circadian-gated.

Figure 4.

 Circadian gating of isoprene emission rates (IE) from oil palm at 1000 μmol m−2 sec−1 and 30°C.
When a frondlet was kept in the dark for 24 h and illuminated at the start of a subjective day, initiation of IE was delayed by 12 h (black line); when a frondlet kept in the dark for 36 h was illuminated at the start of a subjective day (bsl00043), no delay in the onset of isoprene emission was observed. Comparison of emission rates between the two leaves shows that illuminating the frondlet at the start of subjective night resulted in a shift of phase of approximately 9 h.

Diurnal variability in IE at low light intensities

To determine the effect of different light intensities on the observed diurnal variability in IE, we measured IE at PAR of 100, 200 and 300 μmol m−2 sec−1 on successive days at a constant leaf temperature (30°C). For each light treatment IEmax occurred 2–3 h after illumination (Figure 5). Following IEmax there was a steady decline in IE during the afternoon. This is in contrast to the square wave function previously assumed, in which IE rapidly stabilizes at a given value, irrespective of the time of day, as long as PAR and temperature remain constant. However, the magnitude of IEmax was PAR-dependent, which is consistent with previous reports (Guenther et al., 1993; Lerdau and Gray, 2003).

Figure 5.

 Diurnal variability in isoprene emission rates (IE) at a constant leaf temperature of 30°C and PAR of 100 μmol m−2 sec−1 (•), 200 μmol m−2 sec−1 (bsl00043) and 300 μmol m−2 sec−1 (×), respectively.

Discussion

Despite considerable efforts to elucidate the physiological role of isoprene, its biological function remains unclear. Several hypotheses have been proposed to explain its production by plants, including a role as a mechanism to alleviate oxidative stress (Affek and Yakir, 2002; Loreto et al., 2001); as a protectant of photosynthetic apparatus against thermal damage (Sharkey and Yeh, 2001); or as a mechanism to protect against excessive light (Penuelas and Munne-Bosch, 2005). Therefore a circadian mechanism that allows plants to adjust to the diurnal variability of these different stresses and more efficiently to protect the photosynthetic apparatus in high light and temperature environments would help to increase productivity, and would be of evolutionary and ecological benefit. Circadian clocks represent an evolutionary adaptation allowing plants to optimize their biochemical and physiological activities in anticipation of the varying demands of the day/night cycle, and to react to environmental signals (Millar and Kay, 1996). From an evolutionary perspective, this ability to synchronize with both external and internal environments may be critical to the plant's survival (Dodd et al., 2005).

The circadian rhythm in IE can be measured across a broad range of light and temperature conditions, with the magnitude and amplitude dependent on these two parameters. Furthermore, by studying the transcriptional regulation of the central clock component CCA1/LHY, we have determined that the oil palm molecular clock maintains rhythmicity at a higher temperature (38°C) than any previously recorded plant circadian rhythm. This suggests that a tropical species such as oil palm, which has evolved under temperatures that commonly reach a daytime maximum of 38°C or more, has a clock that is pre-adapted to function at elevated temperatures. In contrast, circadian rhythms in many other plants, including species native to hot climates, have been reported to become arrhythmic at temperatures in the range 30–32°C (Grams et al., 1995). Together, these observations provide circumstantial evidence for a CCA1/LHY-TOC1-type molecular oscillator underlying isoprene emission rhythms in oil palm trees.

At present it is not known at which level(s) circadian regulation is applied to isoprene emission. Investigation of available promoter regions (1000–2000 bp) from poplar isoprene synthase genes (GenBank accession numbers AM084344 and AY341431; sequences in the Populus trichocarpa genome, http://genome.jgi-psf.org/Poptr1/Poptr1.home.html) reveals the presence of multiple copies of circadian-associated cis elements. These include the binding site for CCA1 (AAa/cAATCT), which is conserved in light-harvesting complex (Lhc) gene promoters (Wang and Tobin, 1998) and is found in a number of circadian-regulated gene promoters (reviewed by McClung, 2000; Nozue and Maloof, 2006); the core conserved sequence (CANNNNATC) of a region required for circadian regulation of the tomato Lhc genes (Piechulla et al., 1998); and the ‘evening element’ (AAATATCT), which was identified by in silico analyses of circadian-regulated gene promoters (Harmer et al., 2000). These elements are all closely related and may represent the same conserved sequence. Their presence in the isoprene synthase gene promoters suggests that expression of isoprene synthase genes may be controlled in a circadian fashion at the transcriptional level. Arimura et al. (2004) showed significant variability in isoprene synthase transcript levels between day and night in poplar leaves; this is consistent with the possibility that transcript abundance is under circadian control. Post-transcriptional regulation of circadian rhythms is also known to occur at several levels (reviewed by Nozue and Maloof, 2006; Somers, 1999).

Circadian regulation may also be imposed on genes upstream in the methylerythritol 4-phosphate pathway for isoprene synthesis, resulting in circadian fluctuation of IE. It has been suggested that isoprene emissions may act as a biological ‘safety valve’ for excess chloroplastic energy and carbon (Rosenstiel et al., 2004). Competition for phosphoenylpyruvate (PEP) between cytosolic and chloroplastic processes may result in changes in isoprene emission by preferentially diverting PEP away from the chloroplast (and hence isoprene synthesis) via the anapleurotic pathway for nitrate assimilation. It has long been known that nitrate assimilation is itself under circadian control at the level of nitrate reductase. As a result, when nitrate levels begin to increase during the afternoon to a late evening/early morning maximum (Yang and Midmore, 2005), isoprene emissions may decrease because of increased cytosolic competition for PEP. It is also known that several other components of the PEP metabolic pathway are under circadian control at levels of transcript abundance and protein activity (Hartwell, 2005; Hartwell et al., 1999; Streatfield et al., 1999; Sullivan et al., 2004), although circadian regulation of all of these components does not occur in all species. Clearly, the circadian control of PEP metabolism and chloroplast–cytosol partitioning, and its relation to the circadian rhythms of isoprene emissions reported here, merit further detailed investigation in a species such as oil palm. We are currently investigating this and other molecular levels at which circadian control may be imposed on IE.

Several attempts have been made to scale-up isoprene emissions to produce regional- and global-scale emission inventories (Guenther et al., 1995; Simpson et al., 1999; Stewart et al., 2003). There are already substantial uncertainties in these estimates, and our discovery of strong circadian control of isoprene emissions is an additional complication that now must be taken into consideration when modelling isoprene fluxes from the biosphere to the atmosphere. While it is not known how widespread circadian regulation of isoprene emissions is, it is possible that other plant species as well as oil palm, especially in the tropics, exhibit this characteristic.

Experimental procedures

Field measurements of IE1000/30

Isoprene emission rates (IE1000/30) were measured from three frondlets of the same frond of a 5-year-old oil palm tree growing in an oil palm plantation (Sabah, Malaysia, February 2004). Samples were collected using a leaf cuvette system (LCpro; ADC Biosciences, Hoddesdon, UK) which controlled PAR and leaf temperature (1000 μmol m−2 sec−1; 30°C). Inflow air (350 ml min−1) was filtered through charcoal to remove ozone and other ambient trace gases. Gas samples from the cuvette were taken into Perkin Elmer stainless steel sample tubes (89 × 6 mm OD) at a rate of 100 ml min−1 for 15 min. The sample tubes contained 100 mg Carbotrap (Supelco, Bellefonte, PA, USA) to trap isoprene. After sampling, tubes were sealed with quarter-inch Swagelock caps and polytetrafluoroethylene ferrules. Previous work has shown that samples can be stored for up to 8 weeks at 4°C without significant loss of compound, provided samples are collected from an ozone-free cuvette system (Owen et al., 1997). Isoprene analysis was carried out by thermal desorption/capillary gas chromatography with mass selective detection (Possell et al., 2005).

Plant material and growth conditions

Three oil palm (Elaeis guineensis J. Gaertn.) trees were grown from seed for 2 years in Levingtons M3 compost (Fisons PLC, Ipswich, UK) in 10-l containers under 12 h light (250 μmol m−2 sec−1)/12 h dark with the tree illuminated between the hours of 06:00 and 18:00 at ambient temperatures of 30°C day and 20°C night, reflecting typical understorey conditions in a tropical plantation.

Laboratory measurements of isoprene, CO2 assimilation and stomatal conductance

Isoprene emission rates were measured using two systems. Firstly we used a controlled environment broadleaf cuvette (PP-Systems, Hitchin, UK), modified for isoprene analysis, which allowed real-time measurement of IE with simultaneous measurement of plant physiology (carbon assimilation rate and stomatal conductance) from a 2.5-cm2 section of frondlet held at precise conditions of temperature and PAR. Second, a custom-built chamber for whole-frond measurements, containing approximately 1200 cm2 frond, was used. The latter system allowed real-time measurements of IE but not of physiology.

Leaf physiological responses for the cuvette system were obtained using a CIRAS-1 differential CO2/H2O infrared gas analyser (PP-Systems) to control humidity, CO2 concentration, leaf temperature, PAR and air-flow rate through the cuvette. Water vapour concentration and CO2 concentration were measured with a precision of 0.7 mbar and 0.2 μl l−1, respectively. PAR was measured using a filtered silicon detector with a response range of 400–700 nm with a controllable flux precision of ±10 μmol m−2 sec−1. Inlet CO2 concentration was maintained at 360 μl l−1. Leaf temperature was measured with an infrared radiation sensor. The air-flow rate through the cuvette was 350 ml min−1, giving a residence time of approximately 1 sec.

For whole-frond branch measurements, excised palm fronds were cut under water. The cut end of the frond was kept under water and the whole frond placed in the chamber. Illumination was from cold fluorescent lighting tubes (120 μmol m−2 sec−1). Leaf temperature was controlled using thermocouples connected to a heat mat via a temperature controller [PXZ 4; Fuji Electric (UK Ltd), London], giving a frond surface temperature of 25 or 38°C (±0.3°C). Air temperature, relative humidity and PAR were measured simultaneously (RHT2nl and QS2 sensors; Delta-T Devices Ltd, Cambridge, UK). All environmental parameters were recorded every 5 min using a DL2e data logger (Delta-T Devices). The air-flow rate through the chamber was 3 l min−1, giving a residence time of approximately 10 min. The rate at which air was introduced into both cuvette and frond chamber was mass flow-controlled, and the air was scrubbed of isoprene using activated charcoal.

Air from both leaf and frond chambers was introduced at a rate of 100 ml min−1 into a proton transfer-reaction mass spectrometer (PTR–MS; Ionicon GmbH, Innsbruck, Austria) for isoprene analysis. The PTR–MS was calibrated by dilution of a 700 ppb isoprene gas standard (Linde UK Ltd, Aberdeen, UK) and operated with an E:N ratio of 125 Td (Hayward et al., 2002). Air samples from the cuvette were also collected on 6-mm OD stainless steel sample tubes containing absorbent resin (Carbotrap) for isoprene identification using GC–MS (Turbomatrix Gold; Perkin Elmer, Boston, MA, USA). The analytical procedure followed that of Possell et al. (2005).

Mathematical analysis of rhythms

Individual period estimates for isoprene emission, CO2 assimilation and stomatal conductance were generated by importing data into brass (available from http://www.amillar.org) and using brass to run fast (FFT–NLLS) analysis programmes (Plautz et al., 1997) on each data trace to generate period estimates and relative amplitude errors.

Identification of an oil palm CCA1/LHY orthologue

We identified an oil palm CCA1/LHY EST in the GenBank EST database using the tblastn search algorithm. The translated EgCCA1/LHY EST (GenBank accession number CN600590) aligns with the Arabidopsis AtLHY protein from amino-acid residue 235–350 and shares 35.4% identity with AtLHY across this region. Phylogenetic tree analysis confirmed that this EST was a good orthologue of AtCCA1/LHY and was most closely related to AtLHY (data not shown). We also identified an oil palm EST that is an orthologue of the Arabidopsis polyubiquitin 10 gene (UBQ10), which we used as a loading control in our real-time PCR expression analysis. This gene is routinely used in Arabidopsis as a loading control for real-time PCR analysis (Gould et al., 2006). The EgUBQ10 EST corresponds to GenBank accession number CN600235. Gene-specific primers for real-time PCR of EgCCA1/LHY and EgUBQ10 were designed as follows: EgCCA1/LHYF 5′-ACCCTTTACCCAATTCCACA-3′, EgCCA1/LHYR 5′-CAGGCTTGGATAAGGGTTCA-3′, EgUBQ10F 5′-GACAACGTCAAGGCCAAGAT-3′ and EgUBQ10R 5′-CTCGAGGGTGATCGTCTTTC-3′.

RNA isolation

Frondlet samples were collected and frozen in liquid N every 6 h from an excised oil palm frond maintained in constant light (120 μmol m−2 sec−1) and at a constant temperature of either 25 or 38°C in the branch chamber. Frozen frondlet samples were ground in liquid N using a mortar and pestle, and total RNA was isolated using the acetyl-trimethyl ammonium bromide-based RNA extraction procedure (Hartwell et al., 1996). The quantity and purity of the RNA were determined spectrophotometrically as described by Sambrook et al. (1989).

Reverse transcription and real-time PCR

Total RNA samples (1 μg) were reverse-transcribed using random hexamers and the QuantiTect Reverse Transcription kit with integrated genomic gDNA Wipeout buffer (Qiagen, Crawley, UK) according to the manufacturer's instructions. EgCCA1/LHY transcript abundance was measured in each sample relative to EgUBQ10 using real-time PCR in a Rotor-Gene 3000 (Corbett Research Ltd, Cambridge, UK). Each reaction contained 2 μl cDNA and 6 μl QuantiTect SYBR Green PCR mix (Qiagen) together with the gene-specific primers for EgCCA1/LHY or EgUBQ10. The efficiency value of amplification for each set of primers was determined by measuring the abundance of transcripts from a cDNA dilution series. Efficiency values were computed for each primer set using REST (Pfaffl et al., 2002). Each RNA sample was assayed in triplicate. The EgCCA1/LHY transcript abundance levels were normalized to EgUBQ10 using q-gene software.

Acknowledgements

This work was funded by the Natural Environment Research Council. Additional support was provided by the European Science Foundation VOCBAS programme and the European Commission's Marie Curie Research and Training Network, ISONET.

Ancillary