Ultraviolet radiation drives methane emissions from terrestrial plant pectins


  • Andy R. McLeod,

    1. Institute of Atmospheric and Environmental Sciences, School of GeoSciences, The University of Edinburgh, Crew Building, West Mains Road, Edinburgh EH9 3JN, UK;
    Search for more papers by this author
  • Stephen C. Fry,

    1. Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Daniel Rutherford Building, Mayfield Road, Edinburgh EH9 3JH, UK
    Search for more papers by this author
  • Gary J. Loake,

    1. Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Daniel Rutherford Building, Mayfield Road, Edinburgh EH9 3JH, UK
    Search for more papers by this author
  • David J. Messenger,

    1. Institute of Atmospheric and Environmental Sciences, School of GeoSciences, The University of Edinburgh, Crew Building, West Mains Road, Edinburgh EH9 3JN, UK;
    2. Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Daniel Rutherford Building, Mayfield Road, Edinburgh EH9 3JH, UK
    Search for more papers by this author
  • David S. Reay,

    1. Institute of Atmospheric and Environmental Sciences, School of GeoSciences, The University of Edinburgh, Crew Building, West Mains Road, Edinburgh EH9 3JN, UK;
    Search for more papers by this author
  • Keith A. Smith,

    1. Institute of Atmospheric and Environmental Sciences, School of GeoSciences, The University of Edinburgh, Crew Building, West Mains Road, Edinburgh EH9 3JN, UK;
    Search for more papers by this author
  • Byung-Wook Yun

    1. Institute of Molecular Plant Sciences, School of Biological Sciences, The University of Edinburgh, Daniel Rutherford Building, Mayfield Road, Edinburgh EH9 3JH, UK
    Search for more papers by this author

Author for correspondence:
Andy McLeod
Tel: +44 131 650 5434
Fax:+44 131 650 0478
Email: andy.mcleod@ed.ac.uk


  • • Recent studies demonstrating an in situ formation of methane (CH4) within foliage and separate observations that soil-derived CH4 can be released from the stems of trees have continued the debate about the role of vegetation in CH4 emissions to the atmosphere. Here, a study of the role of ultraviolet (UV) radiation in the formation of CH4 and other trace gases from plant pectins in vitro and from leaves of tobacco (Nicotiana tabacum) in planta is reported.
  • • Plant pectins were investigated for CH4 production under UV irradiation before and after de-methylesterification and with and without the singlet oxygen scavenger 1,4-diazabicyclo[2.2.2]octane (DABCO). Leaves of tobacco were also investigated under UV irradiation and following leaf infiltration with the singlet oxygen generator rose bengal or the bacterial pathogen Pseudomonas syringae.
  • • Results demonstrated production of CH4, ethane and ethylene from pectins and from tobacco leaves following all treatments, that methyl-ester groups of pectin are a source of CH4, and that reactive oxygen species (ROS) arising from environmental stresses have a potential role in mechanisms of CH4 formation.
  • • Rates of CH4 production were lower than those previously reported for intact plants in sunlight but the results clearly show that foliage can emit CH4 under aerobic conditions.


There remains considerable discussion among plant physiologists about recent observations of the production of CH4 by vegetation foliage under aerobic conditions (Dueck & van der Werf, 2008). Observations by Keppler et al. (2006) first demonstrated that dry and live leaf material could produce CH4, with rates influenced by sunlight and temperature, and this was suggested as a possible explanation for unexpectedly high atmospheric concentrations of CH4 detected over tropical rainforests (Bergamaschi et al., 2007). Biospheric CH4 emissions to the atmosphere are thought to originate mostly from anaerobic microbial processes but experimental observations (Keppler et al., 2006; Sanhueza & Donoso, 2006; Wang et al., 2008), atmospheric measurements (Crutzen et al., 2006; do Carmo et al., 2006; Miller et al., 2007) and subsequent analyses (Bousquet et al., 2006; Houweling et al., 2006; Ferretti et al., 2007) do suggest the possibility of a modest CH4 flux from terrestrial vegetation. Two subsequent investigations were unable to demonstrate CH4 emissions (Dueck et al., 2007; Beerling et al., 2008) but a recent study, using measurements in the dark, has suggested that some woody plants, but not grasses, may emit small amounts of CH4 (Wang et al., 2008). The CH4 was hypothesized to originate from methoxyl groups (Keppler et al., 2006), and recently Keppler et al. (2008) have used isotope analysis to demonstrate an influence of temperature and UV radiation on CH4 emissions from pectin and polygalacturonic acid in vitro. The roles of UV radiation and temperature in CH4 emission from leaves have also been demonstrated by Vigano et al. (2008), while observations that CH4 emissions from vegetation may arise from dissolution of soil CH4 and its transport above ground in the transpiration stream (Rusch & Rennenberg, 1998; Terazawa et al., 2007) have widened the debate about possible CH4 sources in vegetation. At present, the mechanisms by which terrestrial vegetation may emit CH4 from foliage under aerobic conditions remain to be identified (Schiermeier, 2006, 2007; Dueck & van der Werf, 2008).

The role of temperature and sunlight in the observations of Keppler et al. (2006) and the well-known role of UV radiation in photodegradation of plant material (Austin & Vivanco, 2006) led us to investigate the role of UV radiation in CH4 production from foliage. In earlier experiments, we found that leaf litter from oak (Quercus robur L.) grown under elevated UV radiation had an accelerated decomposition rate and reduced extractability of carbohydrates which implied changes to cell wall components (McLeod et al., 2007). Over half of plant primary cell wall dry mass may be comprised of pectins, which are cell wall polysaccharides rich in α-d-galacturonate residues with variable proportions of methylesterification (Seymour & Knox, 2002). We therefore investigated the release of CH4 and other gases from pectin-impregnated glass fibre sheets, dried and exposed to UV radiation (from experimental lamps and from sunlight) inside gas-tight UV-transmitting bags filled with ambient air at 30°C. Care was taken in the experimental design to eliminate possible confounding effects from any contamination of the glass fibre or the release of low-molecular-weight hydrocarbons from the gas bag material. We also examined trace gas production from UV-irradiated leaves of the C3 subtropical herb Nicotiana tabacum (tobacco) that were detached in order to limit transpiration as a possible CH4 source. Sharpatyi (2007) recently suggested that a free-radical process would produce CH4 from plant polysaccharides under the influence of UV radiation, which is well known to produce reactive oxygen species (ROS) in plant tissue (Björn, 2002). We therefore investigated CH4 production from glass fibre impregnated with pectin and the ROS-scavenger 1,4-diazabicyclo[2.2.2]octane (DABCO) (Wang et al., 2006) and also from de-methylesterified pectin. ROS are generated in foliage by a range of environmental stresses, including drought, nutrient deficiency, attempted pathogen infection, high temperature, acidic precipitation and ozone exposure (Apel & Hirt, 2004; Wang et al., 2006). We therefore also compared the effects of UV irradiation on CH4 production from tobacco leaves with effects on CH4 production from leaves infiltrated with water, the singlet oxygen generator rose bengal (Filkowski et al., 2004) or the bacterial pathogen Pseudomonas syringae in order to demonstrate the potential role of ROS in CH4 formation.

Materials and Methods

Pectin sheets

Pectin or pectate-impregnated sheets (20.3 × 25.4 cm) were prepared from glass microfibre filters (Whatman GF/A) and commercial pectin derived from citrus fruits (Sigma-Aldrich Chemical Co., Poole, UK). Before use, glass fibre sheets were baked overnight in a furnace at 300°C in order to remove any organic contaminants. Citrus pectin (10 g; Sigma P9135; galacturonic acid content 84%; methoxy content 9.4%; loss on drying 4.1%) was wetted with 20 ml of ethanol to form a slurry, which was dispersed in c. 950 ml of deionized water with vigorous shaking. After several hours of stirring until the slurry dissolved to form a slightly hazy solution, the volume was adjusted to 1 l with water. Pectate (de-methylesterified pectin) was prepared by addition of 400 ml of 1% pectin solution to 40 ml of 1.0 M NaOH, and incubation at 20°C for 30 min. Then, with vigorous shaking, sufficient 1 M H3PO4 was added to bring the pH to 7.4–7.6. To each 20.3 × 25.4-cm sheet of Whatman GF/A glass microfibre filter, 25 ml of the pectin or pectate solution was then applied and allowed to dry in air. With pectin, this resulted in 240 mg of polysaccharide (210 mg of galacturonic acid residues) per sheet (equivalent to 23.5 mg (= 760 µmol) methyl-ester groups per sheet, giving a theoretical maximum yield of 12.1 mg CH4 (c. 17 ml at STP)). With pectate, there was 175 mg of galacturonic acid residue per sheet. Control sheets of glass fibre were prepared in the same way with 25 ml of 2% ethanol. Pectin sheets containing DABCO (Sigma-Aldrich Chemical Co.) were prepared as above except that the pectin solution contained 8 mM DABCO. The absorbance spectrum of a 2.5 g l−1 solution of pectin in distilled water was measured in a scanning spectrometer (Lambda 900 UV/VIS/NIR; Perkin Elmer Inc., Waltham, MA, USA) for evaluation of the UV absorbance of pectin.

Plant growth and leaf infiltration

Tobacco plants (Nicotiana tabacum L. cv. Xanthi) were grown from seed in a temperature-regulated glasshouse in 7.5-l pots containing a mixture of 75% peat and 25% sand with 750 g of ground limestone, 1200 g of ‘Osmocote Exact High K’ slow-release fertilizer (N:P:K 10 : 11 : 18 + 2 MgO + trace elements; The Scotts (UK) Company, Ipswich, UK) and 6 g of ‘Intercept 60WP’ insecticide (a.i. imidacloprid; Bayer CropScience, Cambridge, UK) per 250 l of soil mixture. The day:night temperature was 21 : 18°C, the light intensity was c.150 µmol m−2 s−1 and daylength was extended to 18 h with supplemental sodium lighting. After 8–10 wk, the youngest fully expanded leaves were infiltrated on the abaxial surface, using a 1-ml syringe, with 2 ml of distilled water, 2 ml of 10 µM rose bengal in water (Sigma-Aldrich Chemical Co.) or 2 ml of a water suspension (optical density (OD600) 0.2) of Pseudomonas syringae pv. tomato DC3000 (PstDC3000) carrying the avirulence gene avrB. The bacterium was grown in King's broth (KB) liquid medium (King et al., 1954) supplemented with 50 mg l−1 rifampicin and 50 mg l−1 kanamycin at 30°C overnight. Bacterial cells were harvested by centrifugation and re-suspended to 0.2 at OD600 (the equivalent of 108 colony-forming units (cfu) ml−1) in 10 mM MgCl2.

Ultraviolet radiation sources

UV radiation was provided by three types of lamp (UV313, UV340 and UB351; The Q-Panel Company, Cleveland, OH, USA) filtered with closely wrapped 125-µm cellulose diacetate, which had a shortwave cut-off at c. 290 nm (cellulose diacetate (CA) lamp filter; Fig. 1b) and so removed UV-C wavelengths (< 280 nm). We also used a filter of 0.036-mm UV-opaque polyester (‘Courtgard’ (CG) lamp filter; CPFilms Inc., Martinsville, VA, USA) to remove UV-B and most UV-A wavelengths (< 380 nm). Examples of the spectral irradiance of lamp/filter combinations, including the gas sampling bag, are shown in Fig. 1(b). Lamp irradiation was adjusted using a phase-angle dimming system. Experiments performed in natural sunlight took place in the horticultural gardens of the University of Edinburgh at UK National Grid Reference NT 270705 (55°55′N, 3°10′W), between 6 and 21 September 2006.

Figure 1.

Methane production from citrus pectin. (a) Linear regression of CH4 production with ultraviolet irradiance weighted with the inset spectral weighting function, using cellulose diacetate (CA)-filtered UV313, UV340 and UV351 lamps, UV-opaque ‘Courtgard’ (CG) polyester-filtered UV313 lamps and sunlight. Open symbols correspond to spectra in (b). For lamp sources, values are means of three replicates and standard errors are less than symbol size except where visible. For sunlight, values are individual measurements. Other symbols (close to the horizontal axis) show CH4 production from pectin plus the reactive oxygen species (ROS)-scavenger 1,4-diazabicyclo[2.2.2]octane (DABCO) and from de-methylesterified pectin. The temperature was 30°C. (b) Spectra of sunlight (SUN) and CA-filtered lamps (UV313, UV340 and UV351) and CG-filtered UV313 lamps (CG313) used in experiments. (c) Transmission spectra (solid lines) of filters used in experiments (CA, cellulose diacetate; CG, ‘Courtgard’ UV-opaque polyester), of materials commonly used to construct experimental plant chambers (GL, 4-mm window glass; PE, 3.8-mm ‘Perspex’; PO, 2.7-mm polycarbonate; SA, 4-mm ‘Sanalux’ glass) and the absorbance spectrum (right-hand axis) of 2.5 g l−1 pectin (PC) in water (hatched line).

Ultraviolet radiation measurements

UV spectral irradiance was measured with a double monochromator spectroradiometer (SR991-PC; Macam Photometrics, Livingston, UK) which was calibrated against tungsten and deuterium lamps traceable to National Physical Laboratory Standards (SR903; Macam Photometrics). During outdoor experiments the solar spectrum was scanned at c. 15-min intervals and monitored continuously with a broad-band UV sensor (Model PMA2102; Solar Light Inc., Glenside, PA, USA) that was used to calculate changes in spectral irradiance between scans.

Preparation of sample bags

New 5-l gas sampling bags of 25-µm UV-transparent polyvinylfluoride film (SKC Inc., Eighty Four, PA, USA) were cut open on one side to allow insertion of glass fibre sheets and re-sealed using 40-µm aluminium (Al) adhesive tape. Bags were flushed five times before filling with 250 ml of stock external ambient air. Each pair of sample bags was used for three replicate experiments (which were determined not to have modified UV transmission of the bag material). Experiments using rose bengal, P. syringae and UV with tobacco were performed with 200 ml of air and a 20 cm × 20 cm window of 4-mm ‘Sanalux’ glass (Deutsche Spezialglas AG, Delligsen, Germany) attached with Al tape and the remainder of the bag shaded with Al foil. The spectral transmissions of polyvinylfluoride film, Sanalux glass and a range of typical chamber and cuvette construction materials were measured using a scanning spectrometer (Perkin Elmer Lambda 900 UV/VIS/NIR) and are shown in Fig. 1(c).

Experimental exposures of pectin and leaves

Experiments on pectin with UV lamps used one sample bag containing one pectin-impregnated glass fibre sheet and another sample bag containing a control glass fibre sheet. Bags were supported on a black butyl rubber sheet (pond liner) on the surface of a thermostatically controlled water bath at 30°C. After 2 h, the CH4 production was determined and then the pectin-impregnated and control sheets were reversed between bags for a further 2-h exposure. This allowed any difference in CH4 production from the sample bags themselves to be eliminated from the estimate of CH4 production from pectins. The CH4 production inside control bags containing control glass fibre sheets was up to 6.6 ng CH4 h−1 in sunlight and a maximum of 24.6 ng CH4 h−1 in lamp experiments at the highest irradiance of UV313 lamps. Outdoor experiments were performed with bags clipped to a temperature-controlled brass plate also covered with a black butyl rubber sheet. Temperature was measured inside a sample bag with thermocouples connected to a PC-based control system that adjusted the temperature of water from a re-circulating water bath. As outdoor UV levels were variable, experiments were conducted for one period of 2 h without reversing the treatment and control bags. However, the bags were reversed before the next experiment.

Experiments with leaves of N. tabacum used one leaf per gas sample bag and were performed inside a growth room at 25–30°C with 18 W m−2 photosynthetically active radiation (PAR: 400–700 nm) provided by fluorescent lamps (Philips Master TLD 36W/830, Philips Lighting, Guildford, UK) and 3.1 W m−2 unweighted total UV provided by CA-filtered UV313 lamps (described in ‘Ultraviolet radiation sources’ above). Experiments with PAR and with PAR plus UV radiation were also performed using empty bags as a control and the control gas production was subtracted from treatment values.

Gas concentration measurement

Methane, ethane and ethylene concentrations were determined with a gas chromatograph (Hewlett Packard Series II 5890; Hewlett Packard, Altrincham, UK) equipped with a flame ionization detector and a column packed with HayeSep Q (80–100 mesh) at 70°C, with N2 as the carrier gas. Peak integration and autosampling were controlled using a chromatography data system (PeakSimple Model 202, SRI Instruments, Torrance, CA, USA). CO2 was measured with a gas chromatograph (Perkin Elmer 8310) equipped with a thermal conductivity detector, using manual injection. Before and after sample analysis, the gas chromatographs were calibrated using standards of known mixing ratios that spanned the sample values.

Statistical analysis

Rates of CH4 production from pectin are reported as the mean of three replicates ± SE and were analysed by linear regression against UV irradiance values. Trace gas concentrations produced by UV, P. syringae and rose bengal treatments of tobacco are also means of three replicates ± SE and were analysed by individual treatment comparison with the water control using ANOVA with replicate gas bags nested within treatment to achieve a repeated measures analysis (Neter et al., 1996). All statistical tests were performed with the MiniTab statistical package (version 14.1, Minitab Ltd, Coventry, UK).

Results and Discussion

The experiments on pectin used fluorescent lamps (filtered to exclude wavelengths < 290 nm) in order to provide both UV-B (280–315 nm) and UV-A (315–400 nm) radiation (Fig. 1b) up to the highest intensity of global erythemal UV irradiance (Liley & McKenzie, 2006) (Table 1) and experiments were also conducted in the field with sunlight (Fig. 1b). The absorbance of a 2.5 g l−1 solution of pectin (Fig 1c) demonstrated the presence of UV-absorbing components. Methane production from pectin (Table 1) had an approximately linear relationship to total UV irradiance for each lamp type alone or for sunlight, but considerable scatter when all sources were plotted together. We therefore calculated UV irradiance using a range of idealized (Micheletti et al., 2003) and common spectral weighting functions (Table 1). Straight-line logarithmic weighting functions (e.g. Micheletti et al., 2003) provided significant linear relationships with CH4 emissions. The best fit was achieved by an idealized function that decayed one decade in 80-nm wavelength (Fig. 1a, inset) giving a significant linear regression between weighted-UV and CH4 production (Fig. 1a).

Table 1.  Ultraviolet irradiance and methane (CH4) production from citrus pectin
Lamp source/locationUltraviolet irradiance (W m−2)Methane (ng g−1 h−1)
Total UV (280–400 nm)UV-A (315–400 nm)UV-B (280–315 nm)ERYaPASbPGRcQUTdDNAeSTNfMEThMeanSE
  • Values of CH4 production are the mean of three replicate measurements for lamp sources and individual measurements in sunlight (see Materials and Methods).

  • a

    ERY: UV (280–400 nm) weighted with the Commission Internationale de l’Eclairage (CIE) erythemal action spectrum (McKinlay & Diffey, 1987).

  • b

    PAS: UV (280–400 nm) weighted with a mathematical formulation (Green et al., 1974) of the general plant action spectrum (Caldwell, 1971).

  • c

    PGR: UV (280–400 nm) weighted with a new plant growth function (Flint & Caldwell, 2003).

  • d

    QUT: UV (280–400 nm) weighted with the pyridine dimer action spectrum (Quaite et al., 1992).

  • e

    DNA: UV (280–400 nm) weighted with the DNA damage action spectrum (Setlow, 1974).

  • f

    STN: UV (280–400 nm) weighted with the plant growth inhibition function (Steinmuller, 1986).

  • g

    MET: Idealized spectral weighting function for CH4 production that decays one decade in 80 nm.

  • h

    Lamps filtered with 125-µm cellulose diacetate which filters UV wavelengths < 290 nm (see Fig. 1b).

  • i

    UV313 lamps filtered with ‘Courtgard’ polyester which filters UV wavelengths < 380 nm (see Fig. 1b).

  • j

    Highest estimated global erythemal irradiation at Cuzco, Peru calculated from the UV index (Liley & McKenzie, 2006).

UV351-CAh5.<< 0.010.611.26146.36.6
UV313-CGi0.050.05< 0.01< 0.01< 0.01< 0.01< 0.01<
7.987.880.< 0.010.851.71180.1
21.3121.< 0.011.874.01448.5
19.0718.910.< 0.011.673.58391.4
Cuzco, Peruj0.63

These rates of CH4 production exceed values previously reported for pectin with CH4-free air inside glass vials (Keppler et al., 2006). Methane production in the dark was undetectable and was reduced to low rates by a UV-opaque filter (Table 1). De-methylesterification of the pectin also reduced CH4 production to low rates (Fig. 1a).

As free-radical mechanisms could also produce carboxylate radicals and potential dimerization of methyl radicals, we also examined the production of ethane, ethylene and CO2, as well as CH4, resulting from UV irradiation of pectin and found production of all four gases with a molecular ratio of CH4:C2H4: C2H6:CO2 of 1.0 : 0.1 : 0.2 : 240 at a total unweighted UV irradiance of 9.48 W m−2 (Table 2). UV-induced production of ROS (Björn, 2002) is also a potential free-radical mechanism leading to CH4 production from the methyl groups of pectins. We therefore examined CH4 production from pectin-impregnated glass fibre containing the ROS-scavenger DABCO (Wang et al., 2006) and found CH4 production reduced to low rates (Fig. 1a), thus demonstrating a role of ROS in CH4 generation from pectins. ROS include hydrogen peroxide (H2O2), the superoxide ion (inline image) and its nonionized equivalent the hydroperoxyl radical (inline image), the hydroxyl radical (OH), and singlet oxygen (1O2). Of these five, only OH has been reported to cause extensive oxidative scission of plant polysaccharides (Fry, 1998; Fry et al., 2001). However, DABCO is usually reported as a singlet oxygen scavenger (Wang et al., 2006), thus implicating singlet oxygen in CH4 generation. More detailed studies using a range of ROS scavengers are needed to evaluate the precise molecular mechanisms of the UV-induced CH4 formation found in this study and those of Keppler et al. (2008) and Vigano et al. (2008), which reported some CH4 production from nonmethylesterified organic material.

Table 2.  Production of trace gases from ultraviolet irradiation of citrus pectin
Methane (ng g−1 h−1)Ethane (ng g−1 h−1)Ethylene (ng g−1 h−1)CO2 (µg g−1 h−1)
  1. Values are the mean of three replicate experiments. Total UV irradiance was 9.48 W m−2 (280–400 nm) with corresponding weighted irradiance values using a range of common spectral weighting functions shown in Table 1.


UV-photosensitizing compounds (e.g. furocoumarins) are abundant in some plants; therefore ROS formation may be much greater in the foliage of certain plant species than in extracted pectins, although cuticular reflectance of UV and UV-screening compounds (McLeod & Newsham, 1997) will reduce the effective exposure of underlying structures. ROS are generated in foliage by a range of environmental stresses and we therefore also investigated the production of CH4 from leaves of the C3 subtropical herb N. tabacum. Infiltration of leaves of N. tabacum cv. Xanthi with the singlet oxygen generator rose bengal (Filkowski et al., 2004) produced leaf necrosis while infiltration with the bacterial pathogen PstDC3000 carrying the avirulence gene avrB (Whalen et al., 1991) resulted in a hypersensitive response (a genetically programmed cell death mechanism; Apel & Hirt, 2004). All treatments resulted in some CH4 formation but even greater amounts of ethylene and ethane over the subsequent 45 h (Fig. 2). Ethane production was only detected after 5 h in all treatments. The amount of CH4 production was significantly different from that of the water control for each of the three treatments (repeated measures ANOVA (UV: F1,4 = 18.51, P = 0.013; P. syringae: F1,4 = 68.48, P = 0.001; rose bengal: F1,4 = 11.20, P = 0.029) but the mean rate caused by UV irradiation was much greater than that of other treatments. Methane production caused by the UV treatment, which appeared linear with time over 45 h, was 12.3 ± 3.2 ng g−1 leaf dry weight h−1, which is similar to rates previously reported for detached leaves from a range of species (Keppler et al., 2006) but much lower than their reported rates for intact plants.

Figure 2.

Trace gas production from tobacco (Nicotiana tabacum cv. Xanthi). (a) Methane, (b) ethane, and (c) ethylene. Leaves were infiltrated with 2 ml of water (open circles), 2 ml of a suspension of Pseudomonas syringae pv. tomato DC3000 carrying the avirulence gene avrB (open squares), 2 ml of a 10 µM rose bengal solution (closed circles) or 2 ml of water plus UV irradiation (closed squares). All treatments were under 18 W m−2 (87 µmol m−2 s−1) photosynthetically active radiation (PAR; 400–700 nm). The UV irradiation treatment was 3.14 W m−2 (280–399 nm) (equivalent to 1.47 W m−2 inline image; see Fig. 1a) and temperature was 25–30°C. Values are means of three replicates and standard errors are smaller than the symbol except where visible. DW, dry weight.

There is a potential for much higher UV-driven emissions at lower latitudes, where all sites between 50°S and 40°N experience peak erythemally weighted (McKinlay & Diffey, 1987) UV irradiances that exceed 0.25 W m−2 (Liley & McKenzie, 2006). Sunlight-driven CH4 emissions from vegetation were suggested by Keppler et al. (2006) as a possible explanation for unexpectedly high atmospheric concentrations of CH4 detected over tropical regions by satellite remote sensing (Bergamaschi et al., 2007; Schneising et al., 2008). More recent studies have shown that the CH4 data retrieval was positively biased in tropical regions as a result of spectroscopic interference by water vapour, but source inversions based upon an updated data retrieval method still point to substantial tropical CH4 emissions (Frankenberg et al., 2008). Conversion of low-latitude erythemal irradiances > 0.25 W m−2 using typical spectra derived from a spectral radiation transfer model (Gueymard, 1995, 2001) and use of the weighting function for CH4 production (Fig. 1a, inset) suggest equivalent values > 11 W m−2 on the CH4-weighted irradiance scale of Fig. 1(a) and > 30 W m−2 CH4-weighted UV for the highest global irradiance measured at Cuzco, Peru (Liley & McKenzie, 2006). However, the relationship between appropriately weighted UV radiation and CH4 production from plants in vivo (and also from dead plant material) should be determined over the full range of global irradiance values before up-scaling to estimates of global emissions as the relationship may not be linear at higher irradiances and may not persist through time as the substrate becomes modified. The steep response of the process to shorter wavelengths (Fig. 1a, inset) makes it essential to filter experimental lamps to remove wavelengths < 290 nm (McLeod, 1997) in order to ensure realistic exposures in experimental studies.

The use of detached tobacco leaves limited any potential contribution of CH4 dissolved in the transpiration stream, but raised a question about the effect of leaf detachment on observed gas production. The leaves infiltrated with water served as a control and indicated a negligible effect of leaf detachment on gas production, including stress ethylene, during the 45-h experiment. Ethylene is a well-known signal molecule produced by environmental stress in plants (Apel & Hirt, 2004; Wang et al., 2006) and the effect of UV-B irradiation on ethylene production from aminocyclopropane-1-carboxylic acid in tobacco and consequent leaf damage has been reported by Nara & Takeuchi (2002). However, ethylene production directly resulting from UV irradiation of pectins (Table 2) suggests a novel mechanism distinct from the classical ethylene biosynthesis pathways (Wang et al., 2006), which may have implications for ethylene signalling.

These observations in which environmental stresses, particularly from UV irradiation, result in CH4 emissions from foliage under aerobic conditions may contribute to resolving the mystery of CH4 sources from terrestrial vegetation in daylight. However, the night-time observations of CH4 (Crutzen et al., 2006) and experimental detection in darkness (Wang et al., 2008) suggest that potential CH4 sources from the transpiration stream (Terazawa et al., 2007) should also be evaluated further. The rates of CH4 production induced by rose bengal and P. syringae were much lower than those induced by UV irradiation, even though the general oxidative stress caused by these chemical and biological treatments were widespread in the leaf and caused cell death. The trace gas emissions caused by these treatments may originate from different processes (such as lipid oxidation) from those caused by UV irradiation. Nevertheless, the potential effects of ROS generated by other environmental stresses should be considered. Experimental studies are generally undertaken using plants that have not been subjected to environmental stress or disease and without UV irradiation (e.g. Dueck et al., 2007; Beerling et al., 2008). Another likely reason why studies have failed to report a role of UV radiation in CH4 production is that materials used in the construction of experimental leaf chambers and cuvettes, such as glass, polymethyl methacrylate (PMMA, ‘Perspex’ or ‘Plexiglas’) and polycarbonate do not transmit all UV wavelengths (Fig. 1c), and plastic polymers may themselves release hydrocarbons under UV irradiation (Lonneman et al., 1981). The suggestion (Schiermeier, 2006) that some plant species emit up to 4000 times more CH4 than others may reflect the variability of leaf structures, UV-screening pigments (McLeod & Newsham, 1997), UV-photosensitizers (Björn, 2002) and ROS-scavenging mechanisms (Apel & Hirt, 2004) found in plants. Previous measurements of CH4 production from vegetation inside glass and ‘Plexiglas’ chambers (Keppler et al., 2006; Dueck et al., 2007) would have excluded some of the more energetic shorter wavelength UV in solar radiation so that rates of CH4 emission in the field may be larger than suggested from past experiments. However, reported experimental CH4 emissions (Keppler et al., 2006; Sanhueza & Donoso, 2006) may also reflect a combination of experimental stress factors that could include heat and desiccation (and UV irradiance) as well as possible artefacts suggested by Kirschbaum et al. (2007). Sharpatyi (2007) also implied that gamma radiation is a driver of free radical mechanisms that may lead to CH4 production, and consequently the effect of gamma sterilization of vegetation samples to eliminate microbial effects (e.g. Keppler et al., 2006) should also be tested carefully to ensure that it does not cause biochemical changes that influence subsequent experimental results.


These results provide further evidence that plant pectins can act as a source of CH4 under aerobic conditions when exposed to UV radiation within the ambient range (280–400 nm) from experimental lamps and sunlight. They also provide a first step in understanding the potential mechanisms by demonstrating a role of ROS in CH4 production and the additional production of ethane, ethylene and CO2.

The rate of CH4 production (c. 13 ng g−1 h−1) from tobacco leaves was similar to values reported for detached leaves by Keppler et al. (2006), but much lower than their values obtained using intact plants. Nevertheless, vegetation stress factors, especially UV radiation, may have wider implications for biosphere–atmosphere interactions, as any ROS-forming mechanism may have the potential to produce not only CH4 but also other atmospheric trace gases such as C2 hydrocarbons and the methyl halides from terrestrial plant sources. A deeper understanding of the biochemical mechanisms for CH4 production may provide the potential to minimize CH4 generation from large-scale planting of crops and trees by selection for effective ROS-scavenging and/or UV-screening properties. Consequently, the potential of high UV irradiance in the tropics and a range of environmental stress factors to cause ROS formation and trace gas emissions requires examination in a range of species, using realistic spectral irradiance and experimental treatments, in order to fully understand the role of these processes in global emissions.


We thank Julia Drewer, Caroline Nichol, Frank Keppler, Ann Webb, Nigel Paul, Ruth Doherty, Geoff Holmes, Robert Vreeburg, Éva Hedig, Paul Palmer and Lesley Yellowlees for helpful discussions and Colin Kaye, Robert Howard, Alex Hart, Bill Adams, Sophie Haupt and Chris McLellan for assistance. This work was supported by research awards from the Natural Environment Research Council (to ARM), the Moray Endowment Fund (to ARM), the Biotechnology and Biological Sciences Research Council (to SCF, GJL and B-WY), a NERC Fellowship (to DSR), the University of Edinburgh Development Trust (to DJM) and the University of Edinburgh Donald Mackenzie Scholarship (to DJM).