Soils were collected from Italy (Vercelli), China (Fuyang), and Thailand (Suwan). The Italian soil (Vercelli), a silt loam, was originally from a rice field (45.9°N, 171.6°E) at the Italian Rice Research Institute in Vercelli, Italy (Holzapfel-Pschorn et al., 1986), and has subsequently been used for cultivation of rice in our glasshouse. The China soil (Fuyang), a clay loam, was collected in 2007 from a rice field (30.1°N, 119.9°E) at the China National Rice Research Institute in Fuyang, Hangzhou (Peng et al., 2008). The Thailand soil (Suwan), a clay soil, was collected in 2010 from an upland field site (14.6°N, 101.3°E) at the National Corn and Sorghum Research Center of Kasetsart University (Suwanwajokkasikit Field Corps Research Station), Nakornratchasima province, Thailand (Thongsaga et al., 2010). The field had been cropped with maize for several years. The content of C (%) of the soils from Vercelli, Fuyang, and Suwan was 1.3 ± 0.01, 2.4 ± 0.1, and 2.0 ± 0.1, respectively. The N content (%) was 0.13 ± 0.01, 0.25 ± 0.01, and 0.19 ± 0.03, respectively. The soils were air-dried and stored in dry state at room temperature. This treatment ensures that methanogenic activity rapidly recovers after flooding (Mayer and Conrad, 1990). Rice straw originated from rice plants and maize straw from maize plants, both grown in our glasshouse using Italian rice field soil or commercial standard soil, respectively. The straw was air-dried and ground using a blender. The C and N contents of maize straw were 40.2% C and 2.4% N, of rice straw 41.7% C and 0.7% N. The δ¹³C of the maize straw was −12.97 ± 0.05‰, that of rice straw was −30.96 ± 0.10‰.

Soil slurries were prepared by mixing 600 g dry soil with 600 mL of deionized, sterile, anoxic water, and incubating the mixture in 2 L bottles under a headspace of N₂ at 25°C for 4 weeks. The soil slurry was then amended with either rice straw or maize straw at a ratio of 10 mg dry ground straw per gram dry soil, and aliquots (10 g slurry) dispensed into 26 mL pressure tubes. The tubes were closed with black rubber stoppers, flushed with N₂, pressurized to 0.5 bar overpressure, and incubated at 25°C. The tubes with soil slurry were prepared in numerous parallels, of which triplicates were sacrificed for chemical analyses. The gas phase of some incubations was amended with CH₃F, an inhibitor of acetoclastic methanogenesis (Janssen and Frenzel, 1997; Conrad and Klose, 1999b), at a ratio of 1 or 2%. The initial pH values of the soil slurries were between pH 6.1 (Vercelli) and 7.3 (Suwan), but upon anoxic incubation narrowed to pH values between pH 6.7 and 7.0.

Methane was quantified in a gas chromatograph (GC) equipped with methanizer and flame ionization detector, H₂ in a GC with a reduced gas detector (Conrad and Klose, 1999a). The partial pressures of CH₄ measured in the headspace of the incubations were converted into molar quantities using the molar volumes of an ideal gas at the different temperatures. The isotopic composition (δ¹³C) of CH₄ and CO₂ were determined in a Finnigan gas chromatograph combustion isotope ratio mass spectrometer (GC–C–IRMS) system (Thermoquest, Bremen, Germany; Fey et al., 2004). Concentrations of acetate in the liquid phase of the soil incubations were analyzed by high pressure liquid chromatography (HPLC; Krumböck and Conrad, 1991), the δ¹³C of acetate by a HPLC–C–IRMS as described before (Conrad et al., 2007). The δ¹³C of the methyl group of acetate was determined by off-line pyrolysis followed by GC–C–IRMS analysis (Conrad et al., 2007). The C and N content of soil and straw were analyzed on a CHNS-elemental analyzer by the Analytical Chemical Laboratory of the University of Marburg. The δ¹³C_org of straw and the straw-amended soils were analyzed at the Centre for Stable Isotope Research and Analysis (KOSI) at the University of Göttingen using an elemental analyzer coupled to an IRMS. The δ¹³C_org were for Italian, China, and Thailand soil, respectively, −27.1, −27.6, and −22.3‰ for rice straw-amended and −24.6, −25.9, and −18.6‰ for maize straw-amended soil.

DNA was extracted from the soil using the Fast DNA Spin Kit for soil (MP, Heidelberg, Germany) following the manufacturer’s instructions as described in detail by Kolb et al. (2005). DNA extraction was optimized to obtain maximum copy numbers of archaeal 16S rRNA in quantitative PCR (qPCR; see below), which was by combining the DNA from 1, 2, and 3 extractions for soil from Vercelli, Suwan, and Fuyang, respectively. The analysis of terminal restriction fragment length polymorphism (T-RFLP) of archaeal 16S rRNA genes was done as described (Chin et al., 1999) using the primer combination Ar109f/Ar915r (Grosskopf et al., 1998), with the reverse primer labeled with FAM (6-caboxyfluorescein). The 16S rRNA gene amplicons were digested with TaqI (Fermentas, St. Leon Rot, Germany), and the products were size-separated in an ABI 3130 DNA sequencer (Applera, Darmstadt, Germany). The abundance of total archaeal 16S rRNA gene copies was determined by qPCR (Angel et al., 2011, 2012). The abundance of individual groups of archaea was calculated from the relative abundances of their T-RFs multiplied with the copy numbers of total archaeal 16S rRNA genes determined by qPCR (Conrad and Klose, 2006). The abundance of total copies of the mcrA gene, coding for a subunit of the methyl coenzyme M reductase of methanogens was also determined (only in Suwan soil) by qPCR (Angel et al., 2011). The abundance of individual groups of methanogens (only in Suwan soil) was determined by TaqMan qPCR targeting specific 16S rRNA gene sequences (Yu et al., 2005) as described by Angel et al. (2012).

Values of δ¹³C are defined by

with R = ¹³C/¹²C of sample (sa) and standard (st) respectively, using VPDB carbonate as standard. Fractionation factors for a reaction A → B are defined after (Hayes, 1993; Angel et al., 2012):

sometimes expressed as isotopic enrichment factor ε ≡ 1 − α (in units of per mile). The fractionation factor for conversion of CO₂ to CH₄ is given by

where δ13CCH4-mc is the δ¹³C of the CH₄ derived from H₂ + CO₂, i.e., the CH₄ produced in the presence of CH₃F, assuming that acetoclastic methanogenesis was then completely inhibited.

Relative contribution of H₂ + CO₂-derived CH₄ to total CH₄ was determined using the following mass balance equation (Conrad, 2005):

where fCO2,CH4 is the fraction of CH₄ formed from H₂ + CO₂, δ13CCH4 the δ¹³C of total produced methane, δ13CCH4-mc the δ¹³C of CH₄ derived from H₂ + CO₂, and δ13CCH4-ma the δ¹³C of CH₄ derived from acetate determined by:

where δ13Cac-methyl is the δ¹³C of the methyl group of acetate accumulated and εac - methyl,CH4 is the isotopic enrichment factor (equivalent to αac - methyl,CH4) during acetoclastic methanogenesis. It is assumed, that the methyl group of acetate was converted to CH₄ with εac - methyl,CH4=-15‰, which is close to the values determined previously for straw-amended or acetate-amended rice field soil dominated by Methanosarcina species as acetoclastic methanogens (Penning and Conrad, 2007; Goevert and Conrad, 2009).

Methane was produced from each soil amended with either rice straw or maize straw (Figure 1). Vigorous CH₄ production started after about 2 days in soil from Vercelli and Fuyang, and after about 7 days in soil from Suwan. The soil from Suwan was not an authentic rice field soil and had so far not been flooded. Note that the soil from Suwan was preincubated under anoxic conditions without straw amendment for 38 days before the experiment shown in Figure 1C was started. Furthermore, in contrast to soils from Vercelli and Fuyang, production of CH₄ from Suwan soil was only observed when straw was added (data not shown). However, the maximum rates of CH₄ production were in a similar range (140–200 nmol h⁻¹ g⁻¹ dry soil) in all three soils, irrespectively which type of straw was added. Addition of methyl fluoride (1–2%) partially inhibited CH₄ production in all incubations (Figure 1). The pH values during methanogenesis were between pH 6.8 and 7.1.

Acetate transiently accumulated in all incubations indicating that it was produced during degradation of organic matter (straw) and consumed by methanogenesis (Figure 2). The maximum amount of acetate accumulated accounted for about 30–50% of CH₄ produced until the end of the incubation (a concentration of 1 mM acetate corresponds to an amount of 1 μmol g⁻¹ dry soil). Addition of CH₃F resulted in inhibition of acetate consumption so that acetate concentrations stayed high until the end of incubation (Figure 2). Only in Vercelli soil, acetate steadily decreased after 300 h of incubation, probably since Vercelli soil was only treated with 1% CH₃F, whereas the other soils were treated with 2% CH₃F. Since CH₃F concentrations usually decrease slowly during incubation (Conrad and Klose, 1999b; Penning and Conrad, 2006), the CH₃F concentrations in Vercelli soil probably fell below the critical inhibitory threshold of 1% so that acetate could again be consumed by methanogens (Penning et al., 2006b). Increase of acetate (Fuyang) and/or decrease of acetate (Vercelli, Suwan) was generally faster with maize straw than with rice straw (Figure 2), being probably a reason for why CH₄ production was also somewhat faster with maize straw than with rice straw (Figure 1).

Hydrogen, which is next to acetate another important substrate for methanogenesis, also accumulated transiently (Figure 3). As previously reported (Conrad, 2002) the highest H₂ concentrations were reached during the initial lag phase of CH₄ production. During the main phase of CH₄ production H₂ partial pressures were generally lower than 50 Pa. In Vercelli and Fuyang soil, but not in Suwan soil, H₂ partial pressures seemed to be generally about 2 Pa higher with rice straw than with maize straw (Figure 3).

The δ¹³C of the produced CH₄ was generally higher for the soils amended with maize straw than with rice straw and was generally lower in the treatments with CH₃F than in the untreated controls (Figure 4). The δ¹³C of the produced acetate was also generally higher for the soils amended with maize straw than with rice straw, and increased faster with time in the absence than in the presence of CH₃F (Figure 5). This behavior is reasonable, since δ¹³C of the residual acetate should increase because of carbon isotope fractionation during acetate consumption (cf. Penning and Conrad, 2007).

Note that δ¹³C of acetate-methyl is given in Figure 5. The data were directly measured in the incubations of Vercelli and Fuyang soil. The δ¹³C of acetate-methyl was about 10‰ lower than that of total acetate (Figure 6). For Suwan soil, we only analyzed the δ¹³C of total acetate. The corresponding values of acetate-methyl shown in Figure 5 were calculated by subtracting 10‰ from the δ¹³C of total acetate.

The fractionation factors (αCO2,CH4) for hydrogenotrophic methanogenesis (calculated by Eq. 3) were in a range of 1.045–1.060, 1.050–1.060, and 1.040–1.055 for soils from Vercelli, Fuyang, and Suwan, respectively. In general, values were by about 0.005–0.010 units higher in soil amended with rice straw than with maize straw (data not shown).

The percentage contribution of hydrogenotrophic methanogenesis to total CH₄ production was calculated (using Eq. 4) by mass balance of δ¹³C values of CH₄ (Figure 4) and acetate-methyl (Figure 5) and scaled against the amount of CH₄ produced (Figure 1). The results show that this contribution was initially high, decreased during the time when the transiently accumulated acetate was consumed, and then increased again (Figure 7). Interestingly, the pattern was very similar for treatments with maize straw versus rice straw.

The abundance of methanogenic archaea in the authentic rice field soils from Vercelli and Fuyang was quantified before and after incubation with straw using qPCR and T-RFLP targeting archaeal 16S rRNA genes. The T-RFs detected could be affiliated to different methanogenic groups based on previous studies using T-RFLP and cloning/sequencing of archaeal genes (Chin et al., 1999; Lueders and Friedrich, 2000; Peng et al., 2008; Wu et al., 2009). The following T-RFs (in brackets) could be affiliated: Methanomicrobiales (84 bp), Methanobacteriales (92 bp), Methanosarcinaceae (186 bp), Methanosaetaceae (284 bp), and Methanocellales (393 bp). The T-RFs of 74, 381, >738 bp were only detected in low relative abundance (<2%), and thus were not further considered. In Vercelli and Fuyang soil the T-RF = 186 bp represents mainly Methanosarcinaceae. Although this T-RF can also represent non-methanogenic crenarchaeotes (Rice cluster VI or group I.1b), they generally constitute only a minor fraction of T-RF = 186 bp in rice field soil and can be neglected.

Terminal restriction fragment length polymorphism patterns of archaeal 16S rRNA genes slightly changed during incubation with rice and maize straw in both Vercelli and Fuyang soil with the relative abundance of the 186 bp peak increasing during the phase of active acetate consumption (data not shown). Similar patterns have been observed before and were interpreted as growth of acetoclastic Methanosarcina spp. (Conrad and Klose, 2006; Peng et al., 2008). In both Vercelli (Figure 8) and Fuyang (Figure 9) soil the numbers of total archaea significantly increased after incubation with straw, more in Vercelli than in Fuyang, and more with maize straw than with rice straw. The proliferating archaeal community in Vercelli soil consisted mainly of Methanosarcinaceae and Methanobacteriales, while Methanocellales were much less abundant (Figure 8). In Fuyang soil, on the other hand, Methanosarcinaceae and Methanocellales were the most abundant methanogens, but Methanobacteriales were also of relatively large importance (Figure 9). Addition of CH₃F inhibited the proliferation of the methanogenic populations in Vercelli and Fuyang soil (Figures 8 and 9).

The soil from Suwan had previously not been used for cultivating rice. Although archaea were abundant, the T-RFLP pattern was almost exclusively dominated by the 186 bp T-RF. In contrast to Vercelli and Fuyang soil, the abundance of total archaea did not increase upon incubation with straw (Figure 10). Sequencing of six clones of archaeal 16S rRNA genes generally resulted in assignment to crenarchaeotes (accession numbers: JQ286402–JQ286407). We therefore assumed that the archaeal 16S rRNA in this soil was mainly represented by crenarchaeotes, whose abundance was not enhanced by addition of straw. We therefore decided to quantify methanogens directly by applying qPCR assays targeting the mcrA gene that is unique to methanogens. Using this assay, a significant proliferation of methanogens by two orders of magnitude was detected after incubation with straw reaching almost the abundance of total archaeal 16S rRNA gene copies (Figure 10). The methanogenic community consisted almost equally of Methanosarcinaceae and Methanocellales, with Methanomicrobiales as only minor additional component (Figure 10). Proliferation of Methanosarcinaceae and total methanogens was slightly inhibited by the presence of CH₃F (Figure 10).

Hence, the differences were all in details of the methanogenic degradation process and the active methanogenic populations, while the overall methanogenic path was very similar. This demonstrates an amazing constancy in system functioning despite all the differences in substrate quality (mainly based on straw type) and archaeal community composition and abundance (mainly based on soil type). We had previously observed a similar constancy in methanogenic paths in soils vegetated with cultivated rice versus wild rice, although differences in turnover rates existed (Conrad et al., 2009). However, these soils had a similar archaeal community composition. On the other hand, we have observed dramatic differences in CH₄ production and emission when rice was grown on different soils, and roots were colonized with crossly different methanogenic populations (Conrad et al., 2008). Therefore, we do not generally exclude that the archaeal methanogenic community composition and abundance might have a substantial effect on methanogenic functioning. In the present study, however, the methanogenic functioning was generally similar.

In general, CH₄ production was at first dominated by hydrogenotrophic methanogenesis which subsequently decreased as acetoclastic methanogenesis became progressively more important, and finally increased again. This temporal pattern was paralleled by the initial peak in H₂ accumulation, followed by the slower transient accumulation of acetate, which was eventually depleted to rather low concentrations characteristic for steady state between fermentative production and methanogenic consumption of acetate. Similar temporal patterns of the relative contribution of hydrogenotrophic versus acetoclastic methanogenesis have been observed before (Conrad et al., 2002; Fey et al., 2004; Penning and Conrad, 2007). Such temporal pattern has been explained by a conceptual model based on the sequential degradation of organic matter by hydrolytic fermentative, and methanogenic microbes, by the competition of sulfate-reducing and iron-reducing bacteria for H₂ and by the relatively late activation of acetoclastic methanogens (Conrad, 2002; Glissmann and Conrad, 2002).

Incubation of maize straw or rice straw under methanogenic conditions generally resulted in increase of total methanogenic archaea irrespectively of the soil used. In the Vercelli and Fuyang soils, the archaeal 16S rRNA genes almost exclusively belonged to methanogens and copy numbers were relatively abundant (10⁶–10⁸ g⁻¹ dry soil) right from the beginning of incubation, as expected for authentic rice field soils (Krüger et al., 2005; Ma et al., 2010). In the Suwan soil by contrast, copy numbers were in the beginning rather low (10³ g⁻¹ dry soil), which is normal for upland soil (Angel et al., 2012). In all of the soils, however, numbers of methanogens eventually increased during methanogenic degradation of straw. The most abundant methanogens always belonged to the family of Methanosarcinaceae, presumably since acetoclastic Methanosarcina species proliferated by consuming the transiently accumulated acetate. Such proliferation during degradation of straw has been observed before and the final abundance was found to be consistent with theoretical calculations based on the energetic conditions in the soil for acetate conversion to CH₄ (Conrad and Klose, 2006; Angel et al., 2012). The most abundant hydrogenotrophic methanogens were Methanobacteriales in Vercelli soil, and Methanocellales in the other two soils. The results on Vercelli (Weber et al., 2001; Conrad and Klose, 2006) and Fuyang (Peng et al., 2008) soil is consistent with previous observations. Suwan soil had not been flooded before and thus initially contained only a low abundance of methanogens. In the end, however, Methanosarcinaceae and Methanocellales dominated. In other upland soils these two methanogenic groups were also found to be the most prevalent ones, both before first flooding and after an extensive phase of methanogenesis (Angel et al., 2012). These two methanogenic groups are in particular known for potentially being able to deal with toxic O₂ species and surviving oxic soil conditions (Erkel et al., 2006; Angel et al., 2011).

Addition of CH₃F generally resulted in substantial inhibition of CH₄ production, irrespectively of the type of soil or straw. As CH₃F is known as specific inhibitor of acetoclastic methanogens (Janssen and Frenzel, 1997), one would expect that acetate consumption by this group of methanogens is inhibited. Thus, members of the Methanosarcinaceae should be impeded in growth unless they use an alternative energy source such as CH₄ production from H₂/CO₂. A previous study using methanogenic rice roots has indeed shown that growth of acetoclastic Methanosarcinaceae was inhibited by CH₃F (Penning and Conrad, 2006). In the present study using Vercelli soil addition of CH₃F also significantly inhibited the growth of the potentially acetoclastic Methanosarcinaceae, but it also inhibited growth of non-acetoclastic Methanobacteriales. The same was observed in Fuyang soil amended with maize straw, while the methanogenic population in the rice straw treatments were not significantly inhibited at all. In Suwan soil, on the other hand, again only the growth of Methansarcinaceae but not of Methanocellales was inhibited by CH₃F. Although we have no final explanation for the different behavior, we assume that it is due to the fact that Methanosarcina species are not restricted to acetoclastic methanogenic metabolism, and may have used for example H₂/CO₂ for growth in the presence of CH₃F, and that growth of hydrogenotrophic methanogens was unspecifically inhibited (Conrad and Klose, 1999b).

Due to inhibition of acetoclastic methanogenesis addition of CH₃F inhibited or delayed the consumption of the accumulated acetate and resulted in lower δ¹³C of the produced CH₄. The δ¹³C of the CH₄ produced in the presence of CH₃F should represent the δ¹³C of hydrogenotrophically produced CH₄ (δ13CCH4-mc). Since the fractionation factor for hydrogenotrophic methanogenesis is larger than for acetoclastic methanogenesis (Conrad, 2005), δ13CCH4-mc is expected of being lower than the δ¹³C of total CH₄ or of acetoclastically produced CH₄ (δ13CCH4-ma), which was actually observed. The fractionation factors of hydrogenotrophic methanogenesis that were calculated from the values of δ13CCH4-mc (Eq. 3) were in a range of αCO2,CH4=1.04-1.06, which have quite commonly been observed in rice field soils (Sugimoto and Wada, 1993; Chidthaisong et al., 2002; Penning and Conrad, 2007; Conrad et al., 2009). The same range of α_{CO2, CH4} has been used for calculation of (fH2) of total CH₄ produced from CO₂ reduction (Eq. 4) during studies on rice fields in TX, USA (Bilek et al., 1999) and Italy (Krüger et al., 2002). If acetoclastic methanogenesis is not completely inhibited (which might have been the case in Vercelli and Fuyang soil, but not in Suwan soil), δ13CCH4-mc would be less negative than expected and values of fH2 would be overestimated (sensitivity analysis see Conrad et al., 2002). However, the range of fH2 values was similar in soil from Suwan (25–80%) as in soils from Vercelli (15–75%) and Fuyang (25–100%). Nevertheless, it should be noted that the fH2 values that were finally reached were relatively high (50–70%) compared to a value of 33% that is expected when carbohydrates are methanogenically decomposed under steady state conditions (Conrad, 1999).

Computation of fH2 is also affected by the value of δ13CCH4-ma, which was based on measured values of δ¹³C of acetate-methyl and the assumption of a fractionation factor for acetoclastic methanogenesis of αac - methyl,CH4=1.015 (equivalent to εac - methyl,CH4=-15‰). Such value is reasonable because acetoclastic methanogenesis was apparently dominated by Methanosarcina. In soil, Methanosarcina species exhibit εac - methyl,CH4 values that are about 5–10‰ less negative than in pure culture which is probably due to limitation by acetate (Penning and Conrad, 2007; Goevert and Conrad, 2009). In systems dominated by Methanosaeta species, values of εac - methyl,CH4 are even less negative, i.e., on the order of > −10‰ (Valentine et al., 2004; Penning et al., 2006a), and fractionation may become actually zero in systems where acetate supply is so low that it is quantitatively utilized as it is produced. Calculation of fH2 results in increasing values with αac - methyl,CH4 becoming smaller (equivalent to εac - methyl,CH4 becoming less negative). However, we may assume that this was not the case in our experiments, which generally showed acetate concentrations well above 100 μM.

Taking all these uncertainties into account it is striking that amendment with either maize straw or rice straw had a similar effect on the overall value of fH2 and its temporal change. Hence, amendment of soil with maize straw may be used for experiments in which the partitioning of CH₄ production derived from root versus straw decomposition is determined. Such determinations are even feasible under field conditions by measuring the δ¹³C in the total CH₄ produced by comparing fields amended with either rice straw or maize straw. We are presently about doing such studies.