Methoxyl groups of plant pectin as a precursor of atmospheric methane: evidence from deuterium labelling studies


Author for correspondence:
Frank Keppler
Tel: +49 6131 305 316
Fax:+49 6131 305 388


  • • The observation that plants produce methane (CH4) under aerobic conditions has caused considerable controversy among the scientific community and the general public. It led to much discussion and debate not only about its contribution to the global CH4 budget but also about the authenticity of the observation itself. Previous results suggested that methoxyl groups of the abundant plant structural component pectin might play a key role in the in situ formation process of CH4. Here, this effect is investigated using an isotope labelling study.
  • • Polysaccharides, pectin and polygalacturonic acid, with varying degrees of trideuterium-labelled methyl groups in the methoxyl moieties, were investigated for CH4 formation under UV irradiation and heating.
  • • A strong deuterium signal in the emitted CH4 was observed from these labelled polysaccharides.
  • • Results clearly demonstrate that ester methyl groups of pectin can serve as a precursor of CH4, supporting the idea of a novel chemical route of CH4 formation in plants under oxic environmental conditions.


It is well known that plants emit a wide range of volatile organic compounds (VOCs), such as isoprenoids and oxygenated compounds (e.g. methanol and acetone), to the atmosphere (Kesselmeier & Staudt, 1999). However, it was only recently found that plants also produce CH4 when we showed that intact living plants, as well as plant litter, produce this important greenhouse gas in an oxygen-rich environment and release it to the atmosphere (Keppler et al., 2006). We also reported that CH4 emissions were sensitive to both temperature and natural sunlight irradiation. The fact that plants produce CH4 under aerobic environmental conditions is a surprising finding because, before our observations, biological formation was only considered to occur by microbial activity under strictly anaerobic conditions, in environments such as wetlands rice paddies, or the intestinal tract of ruminants. Our controversial findings have led to considerable discussion and debate as to their authenticity and their implications for the CH4 global budget and global warming (Kirschbaum et al., 2006; Schiermeier, 2006; Dueck et al., 2007; Evans, 2007). Most importantly, Dueck et al. (2007), utilizing 13C-labelled plants, reported that there were no substantial emissions of CH4 from living plants, which obviously casts serious doubt on our work. However, on the other hand, a very recent study by Wang et al. (2008) reported emissions of CH4 from several shrubs of the Mongolian steppe, confirming the finding of aerobic methane formation in plants.

Interestingly, again very recently, via a personal communication, we became aware of work conducted in the late 1950s at the Academy of Sciences of Georgia (Tbilisi) on emissions of VOCs from leaves of Populus simonii Carr. and P. sosnowskyi Grossh (Sanadze & Dolidze, 1960). In that study, the researchers suggested that plants could emit CH4 as well as ethane, propane, isoprene and several other volatile organic components. However, no follow-up studies on CH4 release were undertaken, as this group very successfully focused on isoprene emissions instead. Even though this early report and the more recent space, aircraft and surface observations, suggesting a strong CH4 source in tropical forest regions (Frankenberg et al., 2005; Crutzen et al., 2006; Bergamaschi et al., 2007; Miller et al., 2007; Sinha et al., 2007), together with the Wang et al. (2008) study provide some support for our contention that vegetation releases CH4, the mechanism of its formation remains unknown. Elucidation of the underlying pathway for aerobic CH4 is crucial to enable a full understanding of its role and hence its global significance.

As a first step, it is important to gain information about precursor compounds in plants that could give rise to CH4. Based on previous results we suggested the possibility of the involvement of the methyl moiety of the esterified carboxyl group (methoxyl group) of pectin (Keppler et al., 2004). Indeed, in experiments with apple pectin we not only observed emission of CH4 but also noted that the emission rate was broadly similar to that measured with detached leaves (Keppler et al., 2006). However, even though those results indicated a role for pectin, they provide no proof for the involvement of the pectin methoxyl group in CH4 formation.

Stable isotope analysis is a powerful tool that we recently employed to demonstrate that plant pectin and lignin methoxyl groups have unique carbon isotope signatures (Keppler et al., 2004), and also to establish the relationship between hydrogen isotopes of wood lignin methoxyl groups and meteoric water (Keppler et al., 2007). In this investigation we again employ this technique, together with pectin and polygalacturonic acid (PGA) (Fig. 1), modified to contain varying degrees of methoxyl moieties with varying degrees of trideuterium-labelled methyl groups, to demonstrate that plant pectin methoxyl groups are a precursor of CH4 and thus could be a source of it in an oxic environment.

Figure 1.

Chemical structure of pectin (a) and polygalacturonic acid (PGA) (b).

Materials and Methods

Isotopic labelling of pectin and polygalacturonic acid

All reagents were purchased from Sigma-Aldrich Company Ltd except for low methoxyl pectin (GENU® pectin LM-101 AS), which was a gift from CP Kelco Ltd (Lille Skensved, Denmark) and methyl D-galactopyranoside which was purchased from CMS Chemicals (Abingdon, UK).

Methyl esterification

Methyl esterification was performed essentially as described by van Alebeek et al. (2000). Briefly, samples of polygalacturonic acid (2 ± 0.01 g) were weighed into 100 ml volumetric flasks and solutions of anhydrous methanolic H2SO4 (0.02 N, 100 ml) containing either 0, 5 or 20% tetradeuterated methanol (v/v) were added. Samples were incubated at 4°C for 47 d with occasional shaking and then centrifuged (770 g, 10 min). The methanolic H2SO4 was decanted and the remaining sample washed with propan-2-ol (3 × 40 ml), with the supernatant being discarded following centrifugation. The samples were left overnight at room temperature to allow evaporation of the remaining propan-2-ol, and then distilled water (20 ml) was added to the samples, which were then thoroughly mixed, frozen and lyophilized over a 6 d period. Dried samples were ground to a powder using a pestle and mortar. Low methoxyl pectin was methylated by the procedure described earlier, except that only 5% tetradeuterated methanol was used for derivatization and the incubation period was 14 d.

Determination of deuterium label in pectin and polygalacturonic acid

Methoxyl content was determined by measuring the release of methyl iodide (CH3I) using the Zeisel technique as described by Keppler et al. (2007). Incorporation of the deuterium label of the ester group in pectin and modified PGAs was also determined using CH3I. For both quantification and measurement of label, gas chromatography-mass spectrometry (GC/MS) was employed. The GC oven was equipped with a PoraPLOT Q column (12.5 m × 0.25 mm × 8 µm) and programmed to hold at 80°C for 1 min and then ramp at 10°C min−1 to 160°C. The injector port was maintained at 250°C and the sample (50 µl) was injected split at a ratio of 100 : 1. The mass spectrometer was operated in the selected ion monitoring (SIM) mode measuring ion currents at m/z 127, 142 and 145. CH3I was quantified by direct comparison of sample peak areas for the sum of ion currents m/z 142 and 145 with a calibration curve obtained with authentic standard. Percentage label was calculated as follows: (integral of ion current at m/z 145/sum of integrals of ion currents at m/z 142 and 145) × 100.

Methoxyl content and degree of labelling for esterified polygalacturonic acids and pectin are presented in Table 1.

Table 1.  Methoxyl content and % trideutered methyl groups in methyl-esterified polygalacturonic acids and pectin
SampleMethoxyl content (%)% Labelled methoxyl groupsTheoretical δD(CH4) (‰ vs VSMOW)*
  • *

    Calculated using the following equation: δ2H (‰) = ((2H/1H methoxyl × 0.75) + (2H/1Hstandard × 0.25) − 2H/1Hstandard))/2H/1Hstandard × 1000‰ and the assumption that for the four hydrogen atoms of the formed CH4, three hydrogen atoms (75%) are derived from the methoxyl group (OCH3) and one (25%) comes from either surrounding water or the organic model compound with a theoretical value of δ2H of 0‰ (2H/1Hstandard).VSMOW, Vienna Standard Mean Ocean Water.

Polygalacturonic acid esterified with unlabelled methanol1.44
Polygalacturonic acid esterified with 5% CD3OD labelled methanol1.486.1  312 000
Polygalacturonic acid esterified with 20% CD3OD labelled methanol2.3022.61 410 000
Pectin treated with 5% CD3OD labelled methanol4.080.05    1 660

Temperature and illumination experiments

Lypholized milled samples (c. 200 mg) in glass vials (fused quartz (Suprasil), 40 ml) sealed with caps containing a PTFE-lined silicone septa were heated for 24 h at 40, 60, and 80°C or illuminated with a 250 W Osram Vitalux lamp (UVA 320–400 nm, UV-B 280–320 nm). For more details on the characteristics of the lamp, we refer readers to Vigano et al. (2008). All experiments were performed in laboratory ambient air, and thus the initial CH4 concentration in the vial was approx. 1800–2000 ppb with δ2H values c. −90‰. δ2H of CH4 was measured by GC-IRMS at the end of each experiment. For blank measurements, empty vials (containing only laboratory ambient air) were prepared at the same time as the samples and treated under identical conditions.

For the light experiments, vials were placed approx. 35 cm below the lamp, and temperature measured during experiments was in the range between 30 and 38°C. The total unweighted UV-B radiation was c. 3.7 W m−2 and UV-A radiation was c. 42 W m−2. The UV content (UV-A and UV-B separately) was determined with a Waldmann UV meter (Waldmann, Schwenningen, Germany). The relative spectral distribution measurements and the calibration of the Waldmann device were performed with a calibrated standard UV-visible spectroradiometer (model 752, Optronic Laboratories Inc, Orlando, FL, USA).

Isotope ratio monitoring mass spectrometry for determination of δ2H values on CH4

Gas samples were transferred from the vials to an evacuated 40 cm3 sample loop. CH4 was trapped on Hayesep D, separated by gas chromatography from interfering compounds and transferred via an open split to the isotope ratio mass spectrometer (ThermoFinnigan Deltaplus XL, Thermo Electron Corporation, Bremen, Germany). Concentration (reproducibility ± 20 ppb at ambient concentration) and δ2H values were determined using a methane standard with known concentration and isotopic composition as internal reference, and a measurement of the inlet pressure in the sample loop. Values of δ2H (‰) relative to that for Vienna Standard Mean Ocean Water (VSMOW) are defined by the equation δ2H (‰) = (2H/1Hsample/2H/1Hstandard) − 1. Throughout the paper we use the form δD(CH4) values instead of δ2Hmethane values.

A GC-FID instrument for grab sample analysis (reproducibility ± 10 ppb) was additionally used for verification of the IRMS concentration measurements.

Results and Discussion

In a first approach, we employed pectin which had a low degree (0.05%) of trideuterated methyl groups in the methoxyl moieties (Table 1) for heating experiments at 80°C. The emission rate was measured to be c. 2.5 ng g−1 DW h−1 and the δ2H values of CH4 (δD(CH4) values) changed from c. −83‰ to 1‰ within 14 h (Table 2). Calculation of the change in δ2H values, together with the change in CH4 concentration in the vials, revealed that at least 80% of the emitted CH4 must have been derived from the methoxyl groups of pectin. These first results provided strong evidence of the involvement of pectin methoxyl groups in CH4 formation. However, because of the complexity of the labelling of the methyl groups in the modified pectin, we decided that it would be more appropriate to utilize another model compound. Therefore, as pectin is a polysaccharide composed primarily of partially esterified α-(1-4)-linked galacturonic acid units, we employed polygalacturonic acid for further experiments. Using polygalacturonic acid as the model compound has a couple of major advantages over pectin itself. Firstly, since it contains no methoxyl groups, it could be used as the control compound to determine if CH4 formation can also occur from the free acid; and secondly, as it is easily methyl-esterified with methanol, methyl ester derivatives with varying amounts of trideuterium-labelled methyl groups could be conveniently prepared.

Table 2.  Methane emission rates and methane δD values under various detailed experimental conditions
ComponentTemperature (°C)Lamp (Vitalux)Duration (h)Blank (ppb)δD(CH4) blank (‰)End (ppb)δD(CH4) end (‰)δD(CH4) theoretical* (‰)Emission rate (ng g−1 DW h−1)
  • *

    Theoretical value was calculated by the assumption that the increase of the mixing ratio in the vial (end – blank) comes from methane that is entirely derived from labelled methoxyl groups (see theoretical methane values in Table 1).

  • Mixing ratios of CH4 measurements are means of two independent analytical methods (SD ± 23 ppb at ambient concentration).

Pectin untreated80141937−832181   −107 3.7
Pectin label 0.05%80 141937−832074      1     182.5
PGA untreated40241937−861962    −85 0.2
PGA methyl-esterified 0% label40241937−861981    −870.3
PGA methyl-esterified 6% label40241937−862029   1 860  2 7400.7
PGA methyl-esterified 23% label40241937−86n.d.   7 310n.d.
PGA untreated60241896−831983    −89 0.7
PGA methyl-esterified 0% label60241896−832012    −89 0.9
PGA methyl-esterified 6% label60241896−832139 12 800 27 5002.0
PGA methyl-esterified 23% label60241896−832051 54 600105 0001.2
PGA untreated80241877−861960    −85 0.6
PGA methyl-esterified 0% label80241877−862225   −124 2.2
PGA methyl-esterified 6% label80241877−862075 20 000 29 4001.3
PGA methyl-esterified 23% label80241877−862365126 000284 0003.1
PGA untreated30–38+ 21918−78n.d.    −84 n.d.
PGA methyl-esterified 0% label30–38+ 21918−782048    −92 9.2
PGA methyl-esterified 6% label30–38+ 21918−782104  7 800 27 60013.8
PGA methyl-esterified 23% label30–38+ 21918−78n.d. 57 500 n.d.
PGA untreated30–38+ 71937−652067    −662.6
PGA methyl-esterified 0% label30–38+ 71937−652893   −12718
PGA methyl-esterified 6% label30–38+ 71937−653510 39 500138 00016
PGA methyl-esterified 23% label30–38+ 71937−653397163 000600 00015
PGA untreated30–38+141943−1012196   −1022.7
PGA methyl-esterified 0% label30–38+141943−1013755   −18919
PGA methyl-esterified 6% label30–38+141943−1013574 59 000142 00033
PGA methyl-esterified 23% label30–38+141943−1014710233 000825 00058

Polygalacturonic acids derivatized with anhydrous methanolic H2SO4 containing either 0, 5 or 20% tetradeuterated methanol (v/v) were found to contain, respectively, 0, 6.1 and 22.6% trideuterated methyl groups in their methoxyl moieties. From this point forward, these methylated PGAs will be referred to as unlabelled, PGA6 and PGA23. Results from experiments where labelled and untreated PGA samples were incubated at temperatures between 40 and 80°C are shown in Fig. 2. As was expected, high δD(CH4) values were measured for the labelled PGA samples (up to 1 400 000‰, see Table 1) and thus it was essential that ambient laboratory air (c. 1900 ppb with δD(CH4) values in the range of −90‰) was used for headspace in the reaction vials so as to avoid massive contamination and memory effects in the analytical system. Headspace from both PGA6 and PGA23 showed a continuous increase in the δD(CH4) values with increasing temperature (Fig. 2a), whereas with unlabelled PGA a slight decrease in δD(CH4) values with increasing temperature (Fig. 2b) was noted. Relative to the labelled samples, the δD(CH4) values of the unlabelled PGA sample only differed marginally from that of laboratory air. The strong increase in the δD(CH4) values for labelled PGA samples is accompanied by an increase in emission rates at higher temperatures. Emission rates for the esterified PGAs were found to range between 0.3 and 0.7 ng g−1 DW h−1 at 40°C, and between 1.3 and 3.1 ng g−1 DW h−1 at 80°C, rates in a similar range to that recently reported for apple pectin (Keppler et al., 2006). Formation of CH4 from untreated PGA was much lower, c. 0.2 and 0.7 ng g−1 DW h−1 at 40 and 80°C, respectively. The calculated ratio of label between PGA23 and PGA6 samples is 3.7 (22.6/6.1), and this ratio was generally reflected in the calculated ratios of the δD(CH4) values of CH4 formed during the incubation periods. Based on the δD(CH4) values and the increase in headspace CH4 concentration in the vials during the incubation period, it is possible to calculate the percentage of CH4 arising directly from the methyl moiety of the methoxyl groups. This calculation shows that, on average, about two-thirds of CH4 is formed from methoxyl groups, with a range of 48–68% observed (estimated error ± 25%). One possible explanation for this observation is that the methyl moiety is not transferred intact during CH4 formation. Using GC/MS, with the instrument employed in the selected ion monitoring mode, headspace from PGA23 which had been incubated at 80°C was shown to have peaks at both m/z 16 and 19 at the expected retention time of an authentic CH4 standard. The presence of the peak at m/z 19, which was absent in CH4 formed from incubation of an unlabelled PGA sample, indicated the presence of CH4 containing three deuterium atoms, an observation which shows that the methyl group from methylated PGA was transferred intact during the heating process. An alternative explanation could be the release of CH4 as a result of desorption processes, as recently suggested by Kirschbaum et al. (2007). As reported earlier, untreated PGA, containing no methoxyl groups, showed small emissions of CH4 during heating experiments (c. 0.2 and 0.7 ng g−1 DW h−1 at 40 and 80°C, respectively). Since there was no measurable, or only marginal, change in isotope values of the CH4 observed during these experiments, this indicates that the released CH4 from the untreated PGA must have a similar value to CH4 of the background ambient air, and thus it is considered likely that this fraction is derived from desorption processes during heating of the investigated organic material. If we assume that this process is occurring similarly in all conducted experiments using PGA (untreated and methyl-esterified), then we can recalculate the percentage of CH4 arising directly from the methyl moiety of the methoxyl groups by correcting for CH4 release of the untreated PGA. In this case, our calculations show values in the range 63–126% (mean value 94 ± 25%), suggesting that methyl-esterified groups can fully explain the in situ formation of methane during heating experiments from pectin.

Figure 2.

Heating experiments with polygalacturonic acid (PGA) and methyl-esterified PGAs. (a) δ2Hmethane values of methyl-esterified PGAs; (b) δ2Hmethane values of blank, PGA and nonlabelled methyl-esterified PGA. VSMOW, Vienna Standard Mean Ocean Water.

In addition to heating, natural sunlight has been shown to have an even more pronounced effect on CH4 emissions from pectin (Keppler et al., 2006). Moreover, with the recent studies of A. R. McLeod (pers. comm.) and Vigano et al. (2008), it has now become more evident that UV light plays an important role in the production of CH4 from plant matter, and for more detailed information about this we refer readers to their work. Thus in a second set of experiments we decided to investigate the influence of UV radiation on isotopically labelled PGA samples. We used a 250 W Osram ‘Vitalux’ lamp that produces UV-A (320–400 nm), UV-B (280–320 nm) and barely detectable traces of UV-C (< 280 nm) with a spectral distribution shown in Fig. 3 (for more details see Vigano et al., 2008). For our experiments we used a total unweighted UV-B irradiance of c. 3.7 W m−2, which is similar to typical UV-B irradiances found in the tropics. Typical ambient (nonweighted) summer UV-B irradiances near the Earth's surface range from 2 W m−2 at mid-latitudes to 4 W m−2 in the tropics (Bernhard et al., 1997).

Figure 3.

Spectral distribution of the Osram Vitalux lamp.

This study aimed to demonstrate mechanisms and the molecular source for aerobic production of CH4 from pectin and the potential role of UV radiation. Realistic environmental UV exposure requires careful attention to spectral distribution and the spectral weighting of experimental lamps (Björn & Teramura, 1993), so further studies are required to ensure accurate simulation of ambient UV exposures and the extent and quantification of these processes in natural sunlight. Similar to the heating experiments, CH4 formed from both PGA6 and PGA23 samples showed an increase in the δD(CH4) values with increasing irradiation time (Fig. 4a), reaching values of up to c. 230 000‰ after 14 h whilst the unlabelled PGA sample showed a significant decrease from −101 to −189‰ over the same time period (Fig. 4b). δD(CH4) values of the untreated PGA did not differ significantly from that of laboratory air. Emission rates from untreated PGA were found to range from 2 to 3 ng g−1 DW h−1, whilst, in contrast, rates for all methylated PGAs were found to range between 9.2 and 36 ng g−1 DW h−1, which is one to two orders of magnitude higher than the rates measured during heating experiments at 40°C. With light, the percentage of CH4 calculated to be directly derived from methoxyl groups ranged from 28 to 41%, with, on average, one-third found to originate from this source. Correcting these values for the observed methane release from untreated PGA, the percentage of CH4 calculated to be directly derived from methoxyl groups increases to c. 50% (mean value 51 ± 25%); values for all samples being in the range 30–86%. This proportion differs from that observed for the temperature experiments, possibly indicating different pathways involved in CH4 formation in the two processes. It should be mentioned that the CH4 fraction not originating directly from the methyl moiety of the methoxyl pool cannot be fully explained by formation from nonesterified PGA, as those emission rates are almost an order of magnitude lower than the rates observed for methoxylated PGAs. Therefore it would appear that esterification of the carboxyl moiety in pectin is also important for the observed secondary CH4 formation process during UV-light experiments, in that once the methyl group is removed, it increases the possibility of CH4 formation from other carbon atoms within the PGA structure.

Figure 4.

Irradiation experiments with polygalacturonic acid (PGA) and methyl-esterified PGAs. (a) δ2Hmethane values of methyl-esterified PGAs; (b) δ2Hmethane values of blank, PGA and nonlabelled methyl-esterified PGA. VSMOW, Vienna Standard Mean Ocean Water.

An example of a free radical process leading to formation of CH4 during photochemical-induced degradation of polysaccharides was recently presented by Sharpatyi (2007). Although free radical processes are known to generate a cascade of reactions, from a chemical viewpoint CH4 formation from carbon moieties other than methoxyl of PGA is difficult to envisage. The significance of this pathway of CH4 for aerobic formation in plant tissue will require further intensive isotope labelling investigations.


Our results provide unambiguous isotope evidence that methoxyl groups of pectin can act as a source of atmospheric CH4 under aerobic conditions. As previously shown (Keppler et al., 2006), emissions of CH4 from pectin are strongly dependent on temperature and exposure to light, in particular in the UV range.

For more detailed studies on the light effect, we would refer readers to the study of Vigano et al. (2008), in which the role of UV light in the formation of methane from dried and fresh detached leaves is described in considerable detail. Although the mechanism is still unknown, this study is an important first step in gaining more knowledge of potential plant precursor components that will enable delineation of the reaction mechanism, an essential requirement to fully understand the environmental significance of aerobic methane formation.


We are grateful to Andy McLeod for his helpful comments on the manuscript. The European Science Foundation (ESF) is acknowledged for a VOCBAS Research Grant awarded to FK. We thank Carina van der Veen for the GC-FID analysis.