SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Rice field soils contain a thermophilic microbial community. Incubation of Italian rice field soil at 50°C resulted in transient accumulation of acetate, but the microorganisms responsible for methane production from acetate are unknown. Without addition of exogenous acetate, the δ13C of CH4 and CO2 indicated that CH4 was exclusively produced by hydrogenotrophic methanogenesis. When exogenous acetate was added, acetoclastic methanogenesis apparently also operated. Nevertheless, addition of [2-13C]acetate (99% 13C) resulted in the production not only of 13C-labelled CH4 but also of CO2, which contained up to 27% 13C, demonstrating that the methyl group of acetate was also oxidized. Part of the 13C-labelled acetate was also converted to propionate which contained up to 14% 13C. The microorganisms capable of assimilating acetate at 50°C were targeted by stable isotope probing (SIP) of ribosomal RNA and rRNA genes using [U-13C] acetate. Using quantitative PCR, 13C-labelled bacterial ribosomal RNA and DNA was detected after 21 and 32 days of incubation with [U-13C]acetate respectively. In the heavy fractions of the 13C treatment, terminal restriction fragments (T-RFs) of 140, 120 and 171 bp length predominated. Cloning and sequencing of 16S rRNA showed that these T-RFs were affiliated with the bacterial genera Thermacetogenium and Symbiobacterium and with members of the Thermoanaerobacteriaceae. Similar experiments targeting archaeal RNA and DNA showed that Methanocellales were the dominant methanogens being consistent with the operation of syntrophic bacterial acetate oxidation coupled to hydrogenotrophic methanogenesis. After 17 days, however, Methanosarcinacea increasingly contributed to the synthesis of rRNA from [U-13C]acetate indicating that acetoclastic methanogens were also active in methanogenic Italian rice field soil under thermal conditions.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Methane emission from rice field soils contributes about 13% to the global budget of atmospheric CH4 which is an important greenhouse gas (Lelieveld et al., 1998). Production and emission of CH4 is a function of soil temperature, which for example changes during the day or over the season (Schütz et al., 1990). Temperature increase was found to affect not only the rate of CH4 production but also the composition of the methanogenic community and its function (Chin et al., 1999; Fey and Conrad, 2000). This is interesting, since little is known how soil microbial communities react upon temperature changes, for example those expected by increase of the global surface temperature (Davidson and Janssens, 2006). Structure and function of the community undergoes a drastic shift when temperature increases above 42–46°C (Fey et al., 2001; Wu et al., 2006; Conrad et al., 2009; Rui et al., 2009). At moderate temperatures CH4 is always produced by a combination of acetoclastic and hydrogenotrophic methanogenesis involving Methanosarcinaceae, Methanosaetaceae, Methanomicrobiales, Methanobacteriales and Methanocellales (Rice Cluster 1). However, under thermal conditions CH4 is formed by hydrogenotrophic methanogensis with Methanocellales prevailing (Conrad et al., 2009). Although air temperature in Chinese rice fields easily reaches 40°C (Rui et al., 2009), the soil microbial community probably rarely, if at all, experiences the temperature (50°C) that is used for the present study. Nevertheless, it is noteworthy that Italian rice field soil in which the temperatures range typically from 15°C to 30°C (Schütz et al., 1990), as well as rice field soils from other geographical regions (Wu et al., 2006), contains a microbial community that can potentially function and degrade organic matter to CO2 and CH4 under thermal conditions.

Acetate is the most abundant intermediate of organic matter decomposition in anoxic rice field soil (Takai, 1970; Fey and Conrad, 2000; Glissmann and Conrad, 2000). Normally acetate is utilized by acetoclastic methanogenesis involving species of Methanosarcina or Methanosaeta (Chin et al., 1999; Fey and Conrad, 2000; Lueders and Friedrich, 2000). In this reaction, acetate is cleaved and the methyl and carboxyl groups are converted to CH4and CO2 respectively. However, acetate can in principle also be converted to CH4 involving syntrophic acetate oxidation followed by hydrogenotrophic methanogenesis. In this process, both the methyl and the carboxyl groups of acetate are oxidized to CO2 plus H2, which are then used by hydrogenotrophic methanogens for production of CH4 (Schink and Stams, 2006).

Syntrophic acetate oxidation is basically the reversal of chemolithotrophic acetogenesis (4 H2 + 2 CO2[RIGHTWARDS ARROW] CH3COOH + 2 H2O; ΔG°′ = −95 kJ mol−1). Thus, the reaction is endergonic under standard conditions and is thermodynamically favoured by increasing temperature. Therefore, many of the syntrophic acetate-oxidizing bacteria so far isolated are thermophiles, e.g. the so-called strain AOR (Lee and Zinder, 1988) and Thermacetogenium phaeum (Hattori et al., 2000). However mesophilic acetate oxidizers also exist, e.g. Clostridium ultunense (Schnürer et al., 1996) and Candidatus‘Contubernalis alkalaceticum’ (Zhilina et al., 2005).

Our previous results showed that acetate accumulated in Italian rice field soil at elevated temperature (> 41°C) (Fey et al., 2001). Under these conditions acetoclastic methanogenesis did not operate and CH4 was produced exclusively from H2/CO2. The methanogenic archaeal community was dominated by hydrogenotrophic Methanocellales (Rice Cluster I), while acetoclastic Methanosaetaceae were absent and Methanosarcinaceae were only a minor population (Fey et al., 2001; Conrad et al., 2009). A similar behaviour was also shown for a Chinese rice field soil at 45°C (Rui et al., 2009). Nevertheless, acetate consumption is possible under these conditions, since acetate only accumulated transiently and was utilized subsequently (Rui et al., 2009; Noll et al., 2010). Rui and colleagues (2009) found that in a Chinese rice field members of Clostridium (clusters I and III) were always dominating, but that members of Acidobacteria were selected at 45°C, while members of Bacteroidetes and Chlorobi were selected at 15°C and 30°C. In Italian rice field soil, the abundance of populations of Clostridium, but also of Acidaminococcaceae and Heliobacteraceae, was found to parallel thermal acetate turnover (Noll et al., 2010). However, a direct identification of the bacterial populations responsible for thermal acetate degradation has so far not been attempted.

Stable isotope probing (SIP) allows identification of microbial populations that are actively incorporating a 13C-labelled substrate into the DNA or rRNA (Radajewski et al., 2000; Lueders et al., 2004a). The sensitivity of rRNA-SIP is significantly higher than of DNA-SIP because of the generation of labelled rRNA even in the absence of replication, whereas the recovery of labelled DNA requires at least one cell division. The objective of our study was to identify the acetate-utilizing bacteria and archaea being active in methanogenic Italian rice field soil under thermal conditions by studying the time course of rRNA-SIP and DNA-SIP using terminal restriction fragment length polymorphism (T-RFLP) and cloning-sequencing analysis.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Catabolism of acetate

Four sets of rice soil slurry incubations (at 50°C) were performed: (i) treatment with [U-13C]acetate, (ii) treatment with [2-13C]acetate, (iii) treatment with unlabelled acetate and (iv) control treatment without addition of acetate. Acetate was consumed after a lag phase of 15 days (Fig. 1A). After the added acetate was depleted, acetate was added another four times giving a total of 140 µmol of exogenous acetate. Acetate consumption during the second, third, fourth and fifth addition of acetate became increasingly faster being completed within 7, 4, 3 and 2 days respectively.

image

Figure 1. Concentrations of (A) acetate and (B) CH4, and (C) H2 partial pressures in soil slurries after addition of acetate (unlabelled, [2-13C]acetate and [U-13C]acetate) and without addition of acetate (control) during incubation at 50°C over 67 days; (D) apparent isotope fractionation factor (αapp) without addition of acetate (control) and after addition of unlabelled acetate; (E) relative abundance of 13CH4 and 13CO2 (atom%) after addition of [2-13C]acetate; means ± SE of triplicates. Acetate addition is indicated by arrows.

Download figure to PowerPoint

Methane was produced in all incubations. However, in those amended with acetate, accumulation was faster and reached higher concentrations till the end of incubation (67 days) (Fig. 1B). The same was found for CO2 accumulation (data not shown). A total of 52 µmol of 13C carbon (37%) of the [2-13C]acetate added and 74 µmol of 13C carbon (27%) of the [U-13C]acetate added were finally recovered as gaseous 13CH4 and 13CO2 at day 32. Similarly low carbon recovery has been observed before in Italian rice field soil (Chidthaisong et al., 1999; Hori et al., 2007).

The H2 partial pressures increased with incubation time and reached the maximum at day 17 (Fig. 1C). H2 partial pressures were higher in the soils amended with exogenous acetate than in the unamended control, in particular during the first 15 days. This difference resulted in a more negative ΔG for hydrogenotrophic methanogenesis in amended samples, which was exergonic with values from −10 to −20 kJ mol−1 (Fig. 2A). However, H2 partial pressures were also generally sufficiently low to allow exergonic oxidation of acetate to H2 + CO2 with ΔG values range within −20 to −50 kJ mol−1 (Fig. 2B).

image

Figure 2. Gibbs free energy (ΔG) of (A) hydrogenotrophic methanogenesis, (B) syntrophic acetate oxidation and (C) syntrophic propionate oxidation for the incubation conditions at 50°C.

Download figure to PowerPoint

When soil was incubated with [2-13C]acetate, the 13C content of CH4 was initially higher than that of CO2 (Fig. 1E). Nevertheless, significant amounts of 13C-labelled CO2 accumulated and eventually reached a similar atom percentage (27–30% 13C) after 20–25 days (Fig. 1E), demonstrating that the methyl group of acetate was not only reduced but also oxidized to a large extent.

The 13C content of CH4 and CO2 was also analysed in the soil slurries incubated with unlabelled acetate (natural abundance of 13C) or without acetate amendment. The resulting δ13C values of CH4 and CO2 were used to calculate the apparent fractionation factor αapp for the conversion of CO2 to CH4 (Fig. 1D). In the unamended soil αapp stayed at a relatively high value of about αapp ≈ 1.08–1.09, which is characteristic for CH4 production by hydrogenotrophic methanogenesis only (Fey et al., 2004). In the acetate-amended soil, however, the apparent fractionation factor decreased to about αapp ≈ 1.05 at day 7. Such a value is characteristic for the operation of both acetoclastic and hydrogenotrophic methanogenesis (Fey et al., 2004). Thereafter, values of αapp increased again eventually reaching similarly high values as the unamended control soil on day 17 (Fig. 1D). Later on, however, αapp tended to decrease again.

Interestingly, the concentration of propionate in the slurries incubated with acetate increased to 0.17 mM at day 15, which was significantly higher than that of control slurry, and then decreased below the detection limit (Fig. 3A). The relative abundance of 13C in propionate from the unlabelled treatment was 1%, which is the natural abundance of 13C. However, from the treatments with [2-13C]acetate and [U-13C]acetate propionate contained an excess of 9% and 14% 13C respectively (Fig. 3B). While H2 partial pressures were sufficiently low to allow syntrophic degradation of propionate to acetate, CO2 and H2, which was exergonic with values from −10 to −30 kJ mol−1, the reverse reaction was thermodynamically not permissive (Fig. 2C).

image

Figure 3. (A) Concentrations of propionate in soil slurries after addition of acetate (unlabelled, [2-13C]acetate and [U-13C]acetate) and without addition of acetate (control) during incubation at 50°C over 67 days; and (B) relative abundance of 13C (atom%) in propionate at day 11 and day 15 (means ± SE of triplicates).

Download figure to PowerPoint

SIP targeting bacterial 16S rRNA and 16S rRNA genes

Using [U-13C]acetate both rRNA-SIP and DNA-SIP were applied to identify acetate-utilizing bacteria in Italian rice field soil at 50°C. After 17 days of incubation the bulk of bacterial 16S rRNA was still found in the ‘light’ fractions of a density CsTFA gradient (Fig. S1B). However, after 21 days one ‘light’ peak at 1.77 g ml−1 and one ‘heavy’ peak at about 1.79 g ml−1 were observed (Fig. S1C). Bacterial DNA was only found in the ‘light’ fractions of a CsCl gradient after 11 days of incubation (Fig. S1D), but after 32 and 67 days of incubation, a substantial peak of ‘heavy’ density at about 1.72−1.73 g ml−1 was obtained (Fig. S1E and F).

The bacterial T-RFLP profiles of rRNA and DNA density gradient fractions after different days of incubation with either 12C- or 13C-labelled acetate are shown in Fig. S2. A more detailed time-course analysis of the relative abundance of individual terminal restriction fragments (T-RFs) across the density gradient is shown in Fig. 4A and B for rRNA and DNA respectively. For rRNA, it was mainly the T-RFs of 140 bp and 171 bp length that increased in abundance at higher density, in particular after day 17, while the T-RF of 150 bp length decreased strongly (Fig. 4A). For DNA, patterns were more complex but showed that the 150 bp T-RF was dominating the light density fractions in the beginning of incubation, whereas the 120 bp and 171 bp T-RFs dominated the heavy density fractions after 32 and 67 days incubation respectively (Fig. 4B).

image

Figure 4. Relative abundance of different bacterial T-RFs across CsTFA (for rRNA) or CsCl (for DNA) buoyant densities of isolated (A) 13C-labelled rRNA and (B) 13C-labelled DNA.

Download figure to PowerPoint

Four clone libraries were generated using samples of different buoyant density as indicated by asterisks in Fig. S1. One clone library used ‘light’ templates of rRNA for control, and three clone libraries used ‘heavy’ templates, two of which were from rRNA and one from DNA. The phylogenetic affiliation of all bacterial clones analysed is summarized in Table 1. The placement of selected representative clones from ‘heavy’ templates is shown in a phylogenetic tree (Fig. 5).

Table 1.  Phylogenetic affiliations and numbers of 16S rRNA sequences retrieved in clone libraries generated from density-resolved nucleic acids.a
Phylogenetic groupDay 17 (‘heavy’)b B4Day 21 (‘light’)b C9Day 21 (‘heavy’)b C4Day 67 (‘heavy’)c F5
No. of clonesT-RFd (bp)No. of clonesT-RFd (bp)No. of clonesT-RFd (bp)No. of clonesT-RFd (bp)
  • a.

    Characteristic T-RFs for different clone groups are given.

  • b.

    13C enrichment of rRNA.

  • c.

    13C enrichment of DNA.

  • d.

    Terminal restriction fragment lengths are shown in base pairs (bp).

  • Terminal restriction fragments with relative abundance of more than 6% are indicated in boldface.

  • Terminal restriction fragments detected in more than one phylogenetic group are marked with an asterisk.

Acidobacteriaceae515020150, 1523150, 152  
Candidate division OP102140*, 3034140*3295, 303, 305  
Chloroflexi  145614532453
Bacillales        
 Thermoactinomyces    1134  
Clostridiales        
 Clostridium  21941280  
 Thermoanaerobacteriales        
 Thermoanaerobacteriaceae3140*, 171*1866140*, 170, 171*5140*, 171*
 Moorella  1132    
 Thermacetogenium17140*, 151, 171*  18140*, 1519140*, 171*
Unknown        
 Symbiobacterium1132116314120, 122, 124, 132, 163, 171*4120, 132, 163
 Total (clones) 12528 30 47 20 
image

Figure 5. Neighbour-joining tree of representative bacterial 16S rRNA and DNA clone sequences generated from density-resolved nucleic acids extracted from 13C-acetate consuming anoxic rice field soil. R17B4 clones: ‘heavy’ fraction (1.801 g ml−1 CsTFA) from rRNA at day 17; R21C4 clones: ‘heavy’ fraction (1.801 g ml−1 CsTFA) from rRNA at day 21; R21C9 clones: ‘light’ fraction (1.769 g ml−1 CsTFA) from rRNA at day 21; D67F5 clones: ‘heavy’ fraction (1.743 g ml−1 CsCl) from DNA at day 67. Numbers in parentheses (base pairs) indicate the lengths of T-RFs determined in silico. The scale bar represents 10% sequence divergence; GenBank accession numbers of reference sequences as indicated. AY608936 and AB362275 were used as outgroup sequences.

Download figure to PowerPoint

Phylogenetic analysis of the ‘light’ clones revealed a community that was clearly dominated by populations of the Acidobacteriaceae (67% of all clones; Table 1). The Acidobacteriaceae-related sequences (e.g. GU984467) have 99% similarity with a clone sequence (FM956805) previously retrieved from Chinese rice field soil (Rui et al., 2009). However, sequences related to the ‘Candidate division OP10’ (13%), and the genus Clostridium (6%) were also detected. To a minor extent (< 6%) also members of Thermoanaerobacteriaceae, Symbiobacterium and Chloroflexi were detected in light-density rRNA.

In contrast, the clones derived from ‘heavy’ rRNA from day 17 and day 21 showed a clear predominance of sequences related to members of novel clades affiliated with Thermoanaerobacteriaceae (Fig. 5). Further clones formed a distinct cluster related to the Acidobacteriaceae (not shown in Fig. 5), the candidate division OP10, and Symbiobacterium (Table 1; Fig. 5). Furthermore, we detected one clone each related to Clostridium, Thermoactinomyces and Chloroflexi (Table 1). The clone library generated from the ‘heavy’ DNA also consisted of similar phylogentic groups as those obtained from ‘heavy’ rRNA (Table 1, Fig. 5).

We used the sequence data obtained from the clone libraries to assign major T-RFs observed in the different bacterial fingerprints to defined phylogenetic lineages. Thus, the predominant 150 bp and 152 bp T-RFs of the ‘light’ nucleic acid fingerprints represented members of the Acidobacteriaceae, which also dominated the ‘light’ rRNA clone library (Table 1). In contrast, the T-RFs predominating the ‘heavy’ fractions clearly represented members of Thermacetogenium (140 and 171 bp) and the Thermoanaerobacteriaceae (mostly 171 bp) (Fig. 5). However, clones clustering with Symbiobacterium exhibited various T-RF lengths, including a T-RF of 120 bp length (Table 1).

SIP targeting archaeal 16S rRNA and 16S rRNA genes

Stable isotope probing targeting rRNA and DNA was also used to identify the archaea that assimilated [U-13C]acetate after different days of incubation at 50°C. Compared with day 1, at which acetate was not yet degraded, the amount of archaeal rRNA detected in the ‘heavy’ fractions increased with time, and finally after 21 days, a ‘light’ and a second ‘heavy’ peak were obtained (Fig. S3A–C). In DNA-SIP ‘light’ and ‘heavy’ DNA became apparent after 32 days of incubation, and after 67 days, the ‘heavy’ peak was larger than the ‘light’ peak (Fig. S3D–F).

Archaeal T-RFLP profiles of rRNA and rRNA genes from different density fractions after different times of incubation in the presence of [U-13C]acetate are shown in Fig. S4. The T-RFLP of the archaeal 16S rRNA templates showed four major T-RFs of 90, 186, 380 and 393 bp. The relative abundance of these T-RFs changed with the buoyant density of the gradient centrifugation and with the time of incubation (Fig. 6A). Although the 393 bp T-RF was the dominant T-RF, the 186 bp T-RF became increasingly more abundant in particular in the heavy fractions. Already after 17 days, the 186 bp T-RF specifically increased in the heavy fractions, indicating that archaea representing this T-RF reacted upon acetate addition by synthesizing ribosomes. Nevertheless, the archaea represented by the 393 bp T-RF usually dominated the ribosomal RNA. This dominance was even more pronounced when analysing the DNA (Fig. 6B). However, a T-RF of 158 bp length also specifically increased in abundance in the heavy fractions (Fig. 6B).

image

Figure 6. Relative abundance of different archaeal T-RFs across CsTFA (for rRNA) or CsCl (for DNA) buoyant densities of isolated (A) 13C-labelled rRNA and (B) 13C-labelled DNA.

Download figure to PowerPoint

Two archaeal 16S rRNA clone libraries were generated from the ‘heavy’ fractions after reverse transcription from ribosomal RNA after 21 (R21C5) and 67 (D67F6) days of incubation. Randomly selected clones were sequenced and used to phylogenetically characterize the acetate-assimilating archaea (Fig. 7). The T-RFs found in the T-RFLP fingerprint analysis were assigned to Methanosarcinaceae (33% in R21C5 library, 186 bp T-RF), Methanocellales (Rice Cluster 1) (67% in R21C5 library and 83% in D67F6 library, 393 bp), and Methanocellales (Rice Cluster 1) (17% in D67F6 library, 158 bp). The T-RFs of 90 bp and 740 bp lengths were not represented in the clone library, but may be assigned to Methanobacteriales and Crenarchaeota, respectively, based on previous studies using Italian rice field soil (Chin et al., 1999).

image

Figure 7. Neighbour-joining tree of representative archaeal 16S rRNA and DNA clone sequences generated from density-resolved nucleic acids extracted from 13C-acetate consuming anoxic rice field soil. R21C5 clones: ‘heavy’ fraction (1.793 g ml−1 CsTFA) from rRNA at day 21; D67F6 clones: ‘heavy’ fraction (1.735 g ml−1 CsCl) from DNA at day 67. Numbers in parentheses (base pairs) indicate the lengths of T-RFs determined in silico. The scale bar represents 10% sequence divergence; GenBank accession numbers of reference sequences as indicated.

Download figure to PowerPoint

Bacteria and Archaea in unamended soil slurries

Ribosomal RNA was extracted from the control soil samples, which had not been amended with acetate and analysed by T-RFLP. The bacterial community in these samples was dominated by the 150 bp T-RF (up to 62% relative abundance) until the end of incubation at day 28 (Fig. 8A). However, the relative abundance of the 150 bp T-RF gradually decreased after day 15, while that of the 140 bp T-RF gradually increased until both exhibited similar abundance of about 31–33% (Fig. 8A). The relative abundance of the 171 bp T-RF ranged within 7–15% relative abundance, and that of the 120 bp T-RF increased at the very end to about 20% relative abundance (Fig. 8A). The data show that the ribosomes of bacteria characterized by T-RFs of 140 bp and 171 bp length were relatively abundant in unamended soil.

image

Figure 8. Relative abundance of different (A) bacterial and (B) archaeal T-RFs in rRNA isolated from soil slurry incubations without addition of acetate (control) (means ± SE of triplicates).

Download figure to PowerPoint

The archaeal community was dominated by the 393 bp T-RF (88% of relative abundance) till day 15. Then the relative abundance of the 393 bp T-RF decreased gradually to 39%, while that of the 186 bp T-RF increased from 0.8% to 37%. These data show that although the Methanocellales (393 bp) were the dominant contributors of archaeal rRNA in unamended soil under thermal conditions, the Methanosarcinaceae (186 bp) increasingly contributed to synthesis of rRNA after 25 days of incubation (Fig. 8B).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Italian rice field soil incubated at 50°C mainly contained Methanocellales (Rice Cluster 1, 393 bp T-RF) and produced CH4 with a 13C-isotopic signature characteristic for hydrogenotrophic methanogenesis (Conrad et al., 2009). Acetate transiently accumulated but was eventually consumed again. Exogenously added acetate was also consumed albeit after a lag phase, but then progressively faster. Our experiments showed that the acetate consumption under thermal conditions was only partially explained by acetoclastic methanogenesis but in addition involved syntrophic bacterial acetate oxidation coupled to hydrogenotrophic methanogenesis. The bacteria and archaea involved in acetate degradation were identified by SIP.

Stable isotope probing technique essentially requires the addition of 13C-labelled substrate in relatively large amounts, thus considerably increasing the in situ availability. In this study, the concentration of 13C-labelled acetate was not higher than 7 mM because the highest concentration of acetate during pre-incubation time was 5 mM. Such acetate concentrations are commonly reached in rice field soil (Glissmann and Conrad, 2002; Penning and Conrad, 2007; Rui et al., 2009). However, evidence for acetoclastic methanogenesis, such as the temporal decrease of αapp (Fig. 1D), was only obtained in soil that was amended with exogenous acetate, but not in unamended soil (Fig. 1D). Furthermore, after 15 days of incubation of the unamended control treatment there was a gradual increase in the relative abundance of ribosomal RNA of putatively acetoclastic Methanosarcinaceae (186 bp T-RF) (Fig. 8B). Hence, it is likely that thermophilic and acetoclastic Methanosarcinaceae were active and synthesized ribosomes (Fig. S3A), although they did not proliferate and form new DNA (Fig. S3B). These methanogens were closely affiliated to Methanosarcina barkeri (AF028692) and Methanosarcina thermophila (M59140). The highest temperatures for the growth of M. barkeri is 42°C (Maestrojuan and Boone, 1991). However, M. thermophila can be active under thermal conditions. Both Methanosarcina species can use besides acetate a range of different substrates including H2/CO2 (Mladenovska and Ahring, 1997), so that they do not necessarily need catabolizing acetate all the time. Nevertheless, it is likely that the Methanosarcina species used acetate for acetoclastic methanogenesis, since part of the methyl group of acetate was converted to CH4.

However, our results also showed that a large part of the methyl group of acetate was oxidized to CO2 demonstrating that syntrophic acetate oxidation was operating. This result was predicted by previous observations that acetate can be degraded in thermally incubated paddy soil while methane is produced exclusively by reduction of CO2 (Fey et al., 2001; Conrad et al., 2009; Rui et al., 2009; Noll et al., 2010). However, it has not been known which bacteria might be responsible for oxidizing acetate under these conditions.

In the present SIP experiments, already after 17 days of incubation, the bacterial community with 13C-labelled ribosomal RNA consisted mainly of members of the Thermacetogenium genus and the Thermoanaerobacteriaceae. Thermacetogenium-related sequences (e.g. GU984434) which were closely affiliated with the clone sequences of FM956805 (91% sequence similarity) retrieved from a hyperthermophilic anaerobic glucose-degrading reactor (Tang et al., 2008) and AB332113 (91%) retrieved from a thermal methanogenic bioreactor utilizing propionate (Sugihara et al., 2007) demonstrated that these detected taxa play a ubiquitously important role in anaerobic digestion under thermal conditions.

The Thermacetogenium genus contains the thermophilic syntrophic acetate oxidizer T. phaeum, which, however, showed only about 89% sequence similarity to sequences retrieved in this study (Fig. 5). The bacterial sequences clustering within Thermacetogenium exhibited three different T-RFs, i.e. with 140, 151 and 171 bp length. The 171 bp T-RF was the closest one to T. phaeum (AB020336). Thermacetogenium phaeum was isolated from a thermal (55°C) anaerobic methanogenic reactor treating kraft-pulp waste water (Hattori et al., 2000). It oxidizes acetate in co-cultures with Methanothermobacter thermautotrophicus and is also able to grow acetogenically using various substrates, such as pyruvate, methanol or H2/CO2. Thermacetogenium phaeum can operate the CO dehydrogenase/acetyl-CoA pathway reversibly both for acetate oxidation and reductive acetogenesis (Hattori et al., 2005). One of the Thermoanaerobacteriaceae-related sequences was identical to a sequence retrieved from the thermophilic methanogenic rice field soil enrichment ‘MRE50’ (Erkel et al., 2005) so that one may speculate that the identified Thermoanaerobacteriaceae are syntrophic partners of RC-I methanogens (Methanocellales).

Our data show that bacteria related to Symbiobacterium also synthesized ribosomal RNA after 21 days of incubation. The most closely related isolate was Symbiobacterium thermophilum (93% sequence similarity) (Ohno et al., 1999) (Fig. 5). Symbiobacterium thermophilum is a thermophilic bacterium, which was isolated from compost (Suzuki et al., 1988). This bacterium exhibits a characteristic growth dependence on microbial commensals. For example, it grows only when co-cultured with a Bacillus sp. (Ohno et al., 1999). Since the T-RFs characteristic for Symbiobacterium sp. (e.g. 120 bp T-RF) were not very abundant, we suggest that these bacteria were probably only of limited importance for the acetate oxidation observed in our incubations.

Acetate also transiently accumulated in unamended control soil, albeit at low concentration. This acetate production is probably due to anaerobic degradation of soil organic matter including degradation of rice straw (Glissmann and Conrad, 2000). T-RFLP results from the control soil showed that bacteria characterized by the 140 bp and 171 bp T-RFs started to synthesize rRNA during the phase (Fig. 6) when the accumulated acetate was consumed again (Fig. 1A). Hence, bacteria affiliated to Thermoactogenium and Thermoanaerobacteriaceae, which are characterized by these T-RFs, were probably involved in syntrophic acetate oxidation also when no exogenous acetate was added.

It is interesting that the ‘heavy’ nucleic acid fractions did not reveal the presence of Geobacter or Anaeromyxobacter species, which had been found as the predominant acetate-assimilating bacteria in methanogenic rice soil at 25°C (Hori et al., 2007). These species are apparently mesophiles and were not active at elevated temperature. Also, populations of Clostridium cluster I and II, and of Acidaminococcaceae and Heliobacteraceae, being potentially involved in acetate consumption at 45°C (Noll et al., 2010), were not dominant in the present study. A possible explanation is that these species were not identified by SIP but by their correlation with acetate concentrations.

Whereas the ‘heavy’ rRNA clone library was dominated by Thermacetogenium-, Thermoanaerobacteriaceae- and Symbiobacterium-related sequences, the ‘light’ clone library from the same rRNA gradient was dominated by Acidobacteriaceae-related sequences (e.g. GU984467) which were most closely affiliated with one clone sequence (FM956805, 99% sequence similarity) retrieved from a Chinese rice field soil (Rui et al., 2009). These authors found that Acidobacteria species became dominant in the late successions of rice field soil incubated at 45°C, when accumulated acetate was consumed again. Their results were based on analysis of 16S rRNA genes. The present data showed that ribosomes of Acidobacteriaceae-related bacteria were apparently of general importance in rice field soil incubated under thermal conditions, but were not specifically involved in acetate consumption.

The products of syntrophic acetate oxidation are CO2 and H2, which can be further converted to CH4 by thermophilic hydrogenotrophic methanogens. Members of the hydrogenotrophic Methanocellales (Rice Cluster 1, 393 bp T-RF) were the main archaea to form detectable ‘heavy’ rRNA and DNA from 13C-labelled acetate. Besides the 393 bp T-RF, Methanocellales with a 158 bp T-RF were also detected, but only by DNA-SIP not by RNA-SIP. The sequence of these methanogens has 100% similarity with clone sequence PLI-11 (AJ556317) that was retrieved from a Philippine rice field soil (Pila) incubated under thermal conditions (Wu et al., 2006). The reason for not detecting them by RNA-SIP might be the dilution with high rRNA copy numbers of other methanogens (Lueders et al., 2004a). The order Methanocellales (formerly Rice Cluster I) has been characterized by two isolates, a mesophilic (Sakai et al., 2008) and a thermophilic one (Sakai et al., 2010). The thermodynamic conditions for hydrogenotrophic methanogenesis were usually permissive but the free energy available was not very large albeit within the range reported (Hoehler, 2008). We nevertheless assume that the methanogens live in juxtaposition to the syntrophic acetate oxidizers, where H2 concentrations are presumably higher and allow for more negative ΔG.

Besides acetate, also propionate accumulated transiently at 50°C, albeit only to much lower concentrations. Interestingly, 14% and 9% of the propionate carbon was labelled with 13C when incubated in the presence of [U-13C]acetate and [2-13C]acetate respectively (Fig. 3B). Hence, some of the propionate was apparently produced from acetate. A similar observation has previously been made on washed rice roots, where either [1-14C]acetate or [2-14C]acetate accounted for 60–76% of total propionate production (Conrad and Klose, 1999). An explanation might be that propionate is produced by reduction of 13C-labelled acetate and CO2 in a reversal of the anaerobic propionate oxidation pathway (Schink and Stams, 2006). This reaction has been shown to occur in Desulfobulbus propionicus (Laanbroek et al., 1982). We then might assume that bacteria related to syntrophic propionate oxidizers catalysed the reductive formation of propionate from acetate plus CO2. Syntrophic propionate oxidizers have been characterized in Italian rice field soil using SIP indicating the involvement of Syntrophobacter, Smithella and Pelotomaculum species (Lueders et al., 2004b). These bacteria were identified in rice soil at moderate temperatures. In our experiments at elevated temperatures, however, these bacteria were not detected indicating that other bacteria were involved in propionate formation from acetate, perhaps species related to Thermoactogenium and Thermoanaerobacteriaceae, which were actually labelled with 13C derived from acetate. Note, however, that propionate synthesis from acetate was thermodynamically not permissive as the H2 partial pressures measured were too low. Hence, this reaction could have occurred only in microsites with elevated H2. A further possibility is that propionate was not formed by reduction of acetate, but became labelled in the carboxyl group by exchange with 13CO2. However, the content of 13C in CO2 produced during incubations with [2-13C]acetate was not much higher than that in the propionate, so that almost all of the propionate must have undergone such an exchange reaction, which is unlikely.

In conclusion, we identified that uncultured Thermoanaerobacteriaceae oxidized the acetate transiently accumulated in Italian rice field soil at 50°C. Our findings underscore the importance of conducting time-course analysis in both rRNA-SIP and DNA-SIP to trace the soil microbial community actively consuming the added 13C-labelled acetate. Using SIP, we were able to identify acetate oxidizers known from literature, but also novel ones, to be active in rice field soil under thermal conditions. It remains to be shown whether the detected acetate oxidizers would also operate in the mode of chemolithotrophic acetogenesis. The identification of acetate oxidizers was facilitated by the use of thermal conditions, under which syntrophic acetate oxidation has a thermodynamic advantage compared with moderate temperatures. Another advantage was the fact that rice field soils exhibit a naturally high acetate concentration so that it was justified to add 13C-acetate at the concentrations required for SIP. Other environments, like lake sediments, where occurrence of syntrophic acetate oxidation has also been suggested (Nüsslein et al., 2001), are more difficult to assess by the SIP technique because of the low steady-state concentrations of acetate. It remains to be shown which microbial taxa are involved in syntrophic acetate oxidation in such environments.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Soil incubation

Soil samples were collected from the rice fields at the Italian Rice Research Institute in Vercelli, Italy, were air dried and stored as dry lumps at room temperature. The soil was a sandy-loamy silt (27% sand, 58% silt, 15% clay) (Holzapfel-Pschorn et al., 1986). The dry lumps were broken and passed through a stainless steel sieve (2 mm). Soil slurries were prepared in 2 l bottles using 700 g of dry soil, 0.35 g of dry ground rice straw and 700 ml of anoxic sterile water. The bottles were closed with red rubber stoppers, flushed with N2, pressurized to 0.5 bar overpressure, and incubated at 25°C for 47 days and then at 50°C for 40 days, during which acetate accumulated and then was degraded. After verifying that the archaeal 16S rRNA gene-based T-RFLP (see Conrad et al., 2009) in the incubations had reached the characteristic pattern, i.e. dominance of Methanocellales (Rice Cluster I), the slurry was dispensed into 26 ml pressure tubes using about 10 g for each tube.

Three sets of slurries were incubated adding unlabelled, [2-13C] and [U-13C]acetate (99 atom%; Campro Scientific GmbH, Berlin, Germany) equivalent to a final concentration of about 5.5 mM, whereas control slurries received the same amount of distilled water. The tubes with soil slurry were prepared in numerous parallels (once more flushed with N2) and then incubated at 50°C, of which triplicates were sacrificed for determination of CH4 accumulation and composition of the methanogenic microbial community after 1, 2, 4, 7, 9, 11, 15, 17, 21, 25, 28, 32 and 67 days of incubation. Aliquot samples from the gas headspace were analysed, and the soil slurries were stored frozen (−75°C) for nucleic acid extraction.

Methane and CO2 were quantified in a gas chromatograph equipped with methanizer and flame ionization detector (Conrad and Klose, 1999). Hydrogen was analysed by a GC equipped with a HgO-to-Hg conversion detector (Schuler and Conrad, 1990). The isotopic composition of CH4 and CO2 was determined in a Finnigan gas chromatograph combustion isotope ratio mass spectrometer system (GC-C-IRMS) (Thermoquest, Bremen, Germany) (Fey et al., 2004). After addition of 13C-labelled acetate, the 13C content of CH4 and CO2 were high and therefore, are given in atom%. However, after addition of unlabelled acetate or without acetate addition the 13C content of CH4 and CO2 was on the order of the natural abundance and therefore, is given as δ13C. The apparent isotopic fractionation factor for conversion of CO2 to CH4 was determined by αapp = (δ13CO2 + 103)/(δ13CH4 + 103).

Liquid samples for analysis of organic acids dissolved in the pore water were recovered as described by Hori and colleagues (2007). Whenever most of the added acetate was consumed, as analysed by high-pressure liquid chromatography (HPLC), acetate was added again (five times during incubation, about 140 µmol in total).

The concentrations of gases and dissolved compounds in the slurry incubations were used to calculate the Gibbs free energy (ΔG) of hydrogenotrophic methanogenesis, syntrophic acetate oxidation and syntrophic propionate oxidation for the incubation conditions at 50°C (Conrad and Wetter, 1990).

Nucleic acid extraction and gradient centrifugation

Simultaneous extraction of DNA and RNA was conducted with ∼0.5 g (wet weight) of soil samples as previously described (Noll et al., 2005) with modified TPM buffer [50 mM Tris-HCl (pH 7.0), 1.7% (w/v) polyvinylpyrrolidon K25, 20 mM MgCl2, 1% (w/v) sodium dodecyl sulfate] and phenol-based lysis buffer [5 mM Tris-HCl (pH 7.0), 5 mM Na2EDTA, 1% (w/v) sodium dodecyl sulfate, 6% (v/v) water-saturated phenol]. The aqueous phase of samples was precipitated with 0.6 vol. of isopropanol and then incubated at −20°C for 1 h or overnight.

RNA for reverse transcription-polymerase chain reaction (RT-PCR) was obtained by removal of co-extracted DNA with RNase-Free DNase Set (Qiagen, Hilden, Germany) at room temperature for 30 min. RNA was further recovered using the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany) and kept at −80°C until used as template for cDNA synthesis. RNA was confirmed to be DNA-free by the absence of PCR products after amplification of 16S rRNA genes with universal primers 27F/1492R (Weisburg et al., 1991) for bacteria and Ar7f /Ar1384r (Lueders et al., 2004a) for archaea. The DNase digestion step was repeated as necessary. The integrity of the received small-subunit ribosomal RNA (16S rRNA) was checked by standard agarose gel electrophoresis using 1% agarose and ethidium bromide staining. Nucleic acid concentrations were estimated using a Nano-Drop spectrophotometer ND-1000 (Thermo Fisher Scientific, Wilmington, DE, USA).

Isopycnic centrifugation of soil DNA and rRNA was performed, as previously described (Lueders et al., 2004a) with minor modifications of the number and volume of the gradients, in caesium chloride (CsCl) and caesium trifluoroacetate (CsTFA) centrifugation gradients respectively. Twelve equal gradients (∼450 µl) of the density-separated DNA and rRNA were fractionated, and the CsCl and CsTFA buoyant density (BD) of each fraction was determined, and nucleic acids were precipitated for subsequent quantitative and qualitative community analyses.

Quantification of rRNA genes and rRNA in density gradient fractions

Bacterial and archaeal 16S rRNA from gradient fractions was quantified by quantitative PCR and quantitative RT-PCR with primer pair Ba519f/ Ba907r (Stubner, 2002) and A364aF/A934b (Kemnitz et al., 2005) respectively. Copy numbers were standardized by using dilution series (102–107 molecules µl−1) of almost full-length 16S rRNA genes of Escherichia coli strain JM109 or a M. barkeri strain (Accession No. AY641448) as described by Lueders and colleagues (2004a).

Community analyses by T-RFLP analysis

Terminal restriction fragment length polymorphism analysis of density-resolved bacterial and archaeal communities from gradient fractions was performed by PCR or RT-PCR using primer pairs Ba27f-FAM/Ba907r and Ar109f/Ar912rt-FAM respectively. For reverse transcription, 5 µl of random hexamer primers (50 ng µl−1, Invitrogen, Karlsruhe, Germany), 5 µl of dNTPs (2 mM, Promega, Mannheim, Germany) and 2.4 µl of water were added to 1 µl of total RNA samples, incubated for 5 min at 65°C, and chilled on ice for at least 1 min. Samples were incubated with 4 µl of 5× First-Strand Buffer (Invitrogen), 1 µl of dithiothreitol (0.1 M, Invitrogen), and 20 units of SUPERase-In™ RNase Inhibitor (20 U µl−1, Ambion, Austin, TX, USA) and 200 units of SuperScript III reverse transcriptase (200 U µl−1, Invitrogen) at 25°C for 5 min and 50°C for 60 min. After inactivation of the reverse transcriptase by heating at 70°C for 15 min, the reaction product was subjected to PCR with DNA template. The thermal profile of the PCR included 30 cycles of primer annealing at 52°C for 45 s, primer extension at 72°C for 1.5 min and denaturing at 94°C for 45 s. The final elongation step was 5 min. Amplicons were digested by MspI (Bacteria) and TaqI (Archaea, Chin et al., 1999) size-separated on an ABI Prism 373 DNA sequencer (Applera, Darmstadt, Germany).

Sequencing and phylogenetic analysis

Selected density fractions of bacterial and archaeal DNA and rRNA were amplified for cloning using the primer set Ba27f/Ba907r and A364aF/A934b, respectively, under the thermal conditions mentioned above. The PCR and RT-PCR products were ligated into the plasmid vector pGEM-T Easy (Promega), and the ligation mixture was used to transform E. coli TOP10 competent cells (Invitrogen) according to the manufacturer's instructions. Clones were randomly selected and sequenced at GATC Biotech AG (Konstanz, Germany). Raw sequence data were assembled and checked with the Lasergene software package DNASTAR (Madison, USA). Chimeric structures were detected by Bellerophon program on the Greengenes website (DeSantis et al., 2006). Phylogenetic analyses were conducted using the ARB software package (http://www.arb-home.de) as described previously (Liu et al., 2009). Phylogenetic trees with reference 16S rRNA sequences (> 1400 nucleotides) from SILVA (Pruesse et al., 2007) using ‘aligner’ tool were calculated using neighbour-joining method. Sequence data have been submitted to the GenBank database under Accession No. GU984420 to GU984568.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Dr Marc Dumont for introduction of the SIP technique and Melanie Klose, Peter Claus for excellent technical assistance. This study was supported by the Fonds der Chemischen Industrie. F.L. received a postdoctoral fellowship of the Max-Planck Society.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1. Quantification of archaeal templates across gradient fractions by (A–C) quantitative reverse transcription-PCR and (D–F) PCR. The gradient fractions, characterized by their buoyant densities, were generated using CsTFA density gradient centrifugation of (left panel) rRNA and using CsCl density gradient centrifugation of (right panel) DNA extracted from anoxic rice field soil after (A) 1 day, (B) 17 days, (C) 21 days, (D) 11 days, (E) 32 days and (F) 67 days of incubation. Fractions from which T-RFLP fingerprints (triangles) or clone libraries (asterisks) were generated are indicated.

Fig. S2. T-RFLP fingerprints of density-resolved bacterial communities retrieved (A–C) from selected rRNA and (D–F) from selected DNA. The T-RFLP assays were performed using gradient fractions from nucleic acids extracted from samples incubated with unlabelled acetate and [U-13C]acetate as indicated in Fig. S1. CsTFA (for rRNA) or CsCl (for DNA) buoyant densities of respective gradient fractions are given in brackets. The fragment lengths (in base pairs) of important T-RFs (as mentioned in the text) are given. Amplicons were generated with Ba27f-FAM/Ba907r primers and digested with MspI.

Fig. S3. Quantification of bacterial templates across gradient fractions by (A–C) quantitative reverse transcription-PCR and (D–F) PCR. The gradient fractions, characterized by their buoyant densities, were generated using CsTFA density gradient centrifugation of (left panel) rRNA and using CsCl density gradient centrifugation of (right panel) DNA extracted from anoxic rice field soil after (A) 1 day, (B) 17 days, (C) 21 days, (D) 11 days, (E) 32 days and (F) 67 days of incubation. Fractions from which T-RFLP fingerprints (triangles) or clone libraries (asterisks) were generated are indicated.

Fig. S4. T-RFLP fingerprints of density-resolved archaeal communities retrieved (A–C) from selected rRNA and (D–F) from selected DNA. The T-RFLP assays were performed using gradient fractions from nucleic acids extracted from samples incubated with unlabelled acetate and [U-13C]acetate as indicated in Fig. S3. CsTFA (for rRNA) or CsCl (for DNA) buoyant densities of respective gradient fractions are given in brackets. The fragment lengths (in base pairs) of important T-RFs (as mentioned in the text) are given. Amplicons were generated with Ar109f/Ar912rt-FAM primers and digested with TaqI.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
EMI_2289_sm_fS1-4.doc1399KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.