Methanogenese und anaerober Acetatumsatz in einem sauren Hochmoor

Zielsetzung:

Erstmalig wurde hier die syntrophe Acetatoxidation in einem sauren Hochmoor gezeigt.

Dieser Prozess ist bisher nur in Bioreaktoren und Seesedimentproben nachgewiesen worden. Gute Voraussetzugen für das Aufdecken dieses Prozesses in diesem Habitat waren die Abwesenheit von alternativen Elektronenakzeptoren wie Fe(III), SO42- und NO3-, der hohe Anteil von hydrogenotroph gebildetem CH4 und die Azidität der Proben, die den Prozess aus thermodynamischen Gründen fördern kann.

Hier wurde der Prozess über den Umsatz von [2-¹³C]- beziehungsweise [2-¹⁴C]-Acetat verfolgt. Die markierte Methylgruppe kann entweder über die acetoclastische Methanogenese ins CH4 oder über die syntrophe Oxidation von Acetat zunächst ins CO2

und über die hydrogenotrophe Methanogenese ins CH4 gelangen. Um auf das aus der syntrophen Acetatoxidation stammende CH4 zu fokussieren, wurde die acetoclastische Methanogenese spezifisch mit Hilfe von CH3F gehemmt. Um Struktur und Funktion der methanogenen Gemeinschaft zu korrelieren, wurde eine phylogenetische Analyse des Gens für die 16S rRNA und des Gens für die α-Untereinheit der Methyl-CoenzymM-Reduktase durchgeführt. Um einen potentiellen Kandidaten für die syntrophe Acetatoxidatio zu finden wurde außerdem das Gen für die Formyltetrahydrofolate-Synthetase analysiert.

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

Methanogenesis and Anaerobic Acetate Turnover in an Acidic Peat Bog (submitted to Environmental Microbiology)

Martina Metje and Peter Frenzel

Max-Planck-Institute for Terrestrial Microbiology, D-35043 Marburg, Germany Abstract

We investigated anaerobic acetate oxidation (AAcO), homoacetogenesis and methanogenesis in an acidic Sphagnum peat by following the turnover of ¹⁴C-bicarbonate and [2-¹³C]- or [2-¹⁴C]-acetate, respectively. AAcO could be detected in samples collected over five years, however to a varying degree. The lack of alternative electron acceptors sugested, that AAcO was based on the syntrophic oxidation of acetate to H2 and CO2. CH3F, which specifically inhibits acetoclastic methanogenesis, did not affect AAcO, as predicted for a H₂-based syntrophy. 60-100% of methanogenesis were based on H2/CO2. The majority of sequences could be affiliated to hydrogenotrophic methanogens (Methanobacteriales, Methano-microbiales, Rice Cluster I and II).

Production of ¹⁴CH4 and ¹⁴C-acetate from ¹⁴C-bicarbonate indicated both hydrogenotrophic methanogenesis and homoacetogenesis. Homoacetogenesis was strongly increased by amending peat slurries with H2. Analysis of the formyltetrahydrofolate synthetase gene (fhs) revealed a new deep branching group of acetogenic sequences, possibly being responsible for homoacetogenesis and AAcO. Incubations with different H2/CO2 ratios showed that CO2 became limiting under high H2 partial pressures, with homoacetogens and methanogens competing for CO2. In addition to ¹⁴CH4 and ¹⁴C-acetate, ¹⁴C-butyrate and

14C-ethanol accumulated as potential by-products of homoacetogenesis, and ¹⁴C-propionate from the reductive carboxylation of acetate and CO2. Residual AAcO in the presence of chloroform (CHCl3) or AQDS (anthraquinone-2, 6-disulfonate) indicated that acetate was not only oxidized via the acetyl-CoA-pathway suggesting the presence of an yet unknown oxidative pathway.

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

Introduction

Peatlands cover 400 million km² worldwide with their distribution centred in the North (Gorham, 1991). Despite low rates of primary production, northern peatlands have accumulated 200-455 Pg carbon since the last glaciation, corresponding to 20-30% of the global soil organic carbon pool (Frolking et al., 2001; Gorham, 1991). Hence, they have been and may still be substantial sinks for atmospheric CO2 (Harden et al., 1992; Smith et al., 2004). In addition, they are sources for atmospheric CH4, contributing approximately 30% to emissions from biogenic sources(Matthews et al., 1991; Mitra et al., 2005).

In the absence of alternative electron acceptors like Fe(III) or SO42-, methanogenesis is the terminal step in anaerobic mineralization. The main precursors of CH4 are H2/CO2 and acetate. When carbohydrates are the primary substrate, theoretically 2/3 of CH4 originate from acetate. This has been proofed for many methanogenic environments (Conrad, 1999).

However, in lake sediments, e.g. Knaack Lake or Lake Constance, acetate has been shown to be the sole precursor of CH4 (Phelps and Zeikus, 1984; Schulz and Conrad, 1996). In other methanogenic environments the contribution of acetate to methanogenesis was very low or completely absent (Galchenko, 1994; Lansdown et al., 1992; Metje and Frenzel, 2005; Namsaraev et al., 1995). For one northern peatland it has even been shown that acetate was the dominant end product of anaerobic degradation, exceeding CH4 production rates by up to 120 times (Duddleston et al., 2002; Hines et al., 2001).

An alternative acetate degrading process in the absence of electron acceptors is the syntrophic oxidation of acetate to H2 plus CO2. In nature this process may account for acetate consumption, when acetoclastic methanogenesis is missing. However, syntrophic acetate oxidation is thermodynamically very unfavourable and highly dependent on syntrophic partners, removing the intermediate H2 (Conrad, 1999; Schink, 1997). To date, the syntrophic oxidation of acetate has been described for mesophilic and thermophilic biogas reactors and for lake sediments (Karakashev et al., 2006; Nüsslein et al., 2001;

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

is considered to be the reverse of homoacetogenesis. In which direction the organism operates, depends on the H2 partial pressure (Hattori et al., 2000; Hattori et al., 2005).

Most studies in microbial ecology of northern wetlands have either dealed with processes or with community structures. Regarding the latter, basically primary succession transects and ecohydrological gradients have been studied (Galand et al., 2002; Høj et al., 2005;

Juottonen et al., 2005; Merilä et al., 2006). It is still unresolved how terminal processes are regulated, and especially the fate of acetate is under discussion.

A pilot study revealed that under methanogenic conditions acetate was oxidized anaerobically, indicating a syntrophic relationship. Therefore, we investigated methanogenesis, syntrophic acetate oxidation and homoacetogenesis in samples from an acidic peat bog from Estonia. Samples were incubated at 25°C, which is close to the temperature optimum of methanogenesis at this site (Frenzel and Karofeld, 2000), and which corresponds to near-surface summer temperatures. First, AAcO was measured by [2-¹³C]- or [2-¹⁴C]-acetate turnover, and homoacetogenesis and hydrogenotrophic methanogenesis by ¹⁴CO2 turnover. In addition, the carbon flow during degradation of organic matter was measured with and without inhibitors (BES, CHCl3, CH3F, and FCH2COO^-). Secondly, to evidence the influence of varying CO2-and H2-concentrations on terminal processes, we incubated peat slurries under varying H2/CO2 ratios. In order to evaluate the pontential processes, we calculated Gibb’s free energies including actual concentrations, pH and temperature. Finally, structure and function were correlated by phylogenetic analysis of the 16S rRNA gene of Archaea, of the methanogens‘ α-subunit of Methyl-Coenzyme M (mcrA) (Hales et al., 1996; Lueders et al., 2001), and the formyltetrahydrofolate synthetase (fhs) of the acetyl-CoA pathway (Leaphart et al., 2003;

Leaphart and Lovell, 2001).

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

Results Pilot studies

A very first experiment with [2-¹⁴C]-acetate in 2000 indicated that part of the methyl group was not only converted to CH4, but also oxidised to CO2 (data not shown). Using the same peat sample, we made a pilot study amending the peat with [2-¹³C]-acetate (final concentration 10 µM; n=4). The amendment did not stimulate methanogenesis (data not shown). However, the heavy methyl-carbon was detected not only in CH4, but also in CO2

(Figure 1A). This peat contained no other electron acceptors like oxidised nitrogen species, microbial reducible Fe(III), or SO42-, suggesting a different oxidation process being responsible for AAcO.

In a second pilot experiment, different inhibitors were used to test which processes might have been involved in anaerobic oxidation of [2-¹³C]-acetate. Utmost care was taken to avoid contaminations with O2. The autoclaved samples and the controls produced on average 0.23 and 230 nmol CH4 g · DW^-1 · d^-1, respectively. The treatments with AQDS (anthraquinone-2, 6-disulfonate), BES (2-bromoethane sulfonate), or a mixture of both, showed only a minor production between 0.15 and 0.47 nmol CH4 g · DW^-1 · d^-1 on average. The ¹³C-signal became visible nearly immediately in CO2 (Figure 1B) and in CH4

(data not shown).

No signal was detectable in the autoclaved and in the BES-treated samples suggesting a biological process and the involvement of methanogens, respectively. Both treatments with AQDS, an analogon to oxidised humics, showed an isotopic signal in CO2 (Figure 1B).

This signal, however, was nearly two orders of magnitude smaller than that in the controls.

The low-molecular weight AQDS is often used as an analogon for humics (Lovley et al., 1999). It may function as an e^- -shuttle for iron-reducing bacteria, but it may also shift product patterns of fermenting bacteria towards more oxidized end products (Benz et al., 1998). The reliability of AQDS as a model may be questionable, but a minor biological

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

Experiment I.

In further experiments peat samples were incubated at 25°C, which is near the temperature optimum of methanogenesis for this bog (Frenzel and Karofeld, 2000). In addition, this comes close to the in situ summer temperatures (Frenzel & Karofeld, unpublished). CH4

was produced at a rate of 0.28 µmol ⋅ gDW^-1 ⋅ day^-1. The specific inhibition of acetoclastic methanogenesis with CH3F verified that around 80% of methanogenesis was based on H2/CO2 (Figure 2A). Corresponding to this, CH3F caused an equimolar accumulation of acetate (Table 2; molar ratio: 1:1). Gross acetate accumulation never exceeded CH4

accumulation neither in control samples nor in samples inhibited with CH3F (Table 2).

CO2-accumulation was not influenced by CH3F (Figure 2A). BES completely inhibited methanogenesis (Figure 2A). Butyrate accumulated (Table 2) suggesting that it served as a precursor for CH4 with acetate as an intermediate of syntrophic butyrate oxidation. H2

concentrations were strongly dependent on the type of inhibitor, reaching its maximum after five days of incubation (Figure 2A).

AAcO was determined following the turnover of [2-¹⁴C]-labeled acetate. After 16 days of incubation 75.8% ± 2.7 (SE; n=3) of the available [2-¹⁴C]-acetate (Σ¹⁴C-productsmax, equation (1)) was recovered as gaseous compounds, (27.9% ± 2.9 as ¹⁴CO2 and 47.8% ± 3.8 as ¹⁴CH4). The remaining was recovered as ¹⁴C-propionate and ¹⁴C-butyrate. With CH3F, 75.8% ± 9.0 was recovered as gaseous compounds (SE; n=2) (38.8% ± 8.5 as ¹⁴CO2

and 37.0% ± 0.5 as ¹⁴CH4), the balance being made by ¹⁴C-butyrate. With BES no accumulation of longer chained fatty acids from [2-¹⁴C]-acetate occurred, indicating that butyrate did not originate from acetate, but rather from upstream fermenting processes being consistent with the results from BES inhibition (Table 2). ¹⁴CO2 reached steady state conditions soon after addition of [2-¹⁴C]-acetate (Figure 3), whereas ¹⁴CH4 increased linearly. CH3F did not affect AAcO (Figure 3). Referring to Σ(¹⁴CH4 + ¹⁴CO2) in samples with CH3F, AAcO accounted for 37.8% ± 3.4 of total acetate turnover, corresponding to 44.3 nmol CO2 · gDW^-1 · day^-1 ± 2.3 and 18.0 nmol CH4 · gDW^-1 · day^-1 ± 0.1, respectively. 73% ± 9 and 65% ± 6 of total radioactivity were recovered in control samples and in samples with CH3F, respectively. Summed up, the results verified the presence of AAcO.

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

Following the turnover of ¹⁴C-bicarbonate revealed that CO2 was not used for fatty acid production at all. However, around 21.5% ± 2.0 of added ¹⁴C-bicarbonate was recovered as

14CH4 after 16 days of incubation. 111% ± 17 of total radioactivity were recovered.

Experiment II.

The biochemical pathway leading to AAcO, and in particular the organisms involved could not be deduced from experiment I alone. Low levels of Sulfate (8µM) and lacking Fe(II)-accumulation led us assume that AAcO happened to be syntrophic, suggesting the utilization of the acetyl-CoA pathway. Therefore, more detailed experiments with peat sampled in August 2004 were carried out. In order to follow the carbon-flow, we applied CHCl3 and FCH2COO^-, respectively. In addition, peat samples were incubated with 1.5 bar H2. Since syntrophic acetate oxidation is highly dependent on low H2 partial pressure, AAcO should be quenched if the process is based on a syntrophic relationship.

Compared to the first experiment, little radioactivity was recovered from gaseous compounds (data not shown). Nearly no ¹⁴CH4 was produced from [2-¹⁴C]-acetate, suggesting that CH4 was completely based on H2/CO2. Similarly, syntrophic acetate oxidation did not constitute an important CO2 source for hydrogenotrophic methanogenesis. Instead, a net accumulation of ¹⁴C-labeled ethanol, acetate, propionate and butyrate from both, ¹⁴C-bicarbonate and [2-¹⁴C]-acetate, occurred with and without inhibitors (supplementary data Figure 1 and 2). During four weeks of incubation 35.1% ± 3.1 of the available [2-¹⁴C]-acetate were recovered as ¹⁴CH4 and ¹⁴CO2. Thereof, 31.5% ± 7.1 were recovered as ¹⁴CO2. Vac calculated by equations (1) to (3) was changing with time: -162.1 nmol · gDW^-1 · day^-1 ± 47.4 (day 1 to 2; ±SE; n=2), -37.8 nmol · gDW^-1 · day^-1

± 10.2 (day 2 to 7), -20.8 nmol · gDW^-1 · day^-1 ± 7.6 (day 7 to 19), -58.5 nmol · gDW^-1 · day^-1 ± 33.6 (day 19 to 28) (±SE; n=4). However, the percentage recovered as ¹⁴CH4 was small. Referring to ¹⁴C-acetate measured at day 1, in control samples 66% ± 9 of total radioactivity were recovered. ¹⁴CO2 reached steady state five days after amendment with

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

However, 47.9% ± 6.0 of the added ¹⁴C-bicarbonate was converted to ¹⁴CH4 during 28 days of incubation. 50.6% ± 4.9 were recovered as Σ (¹⁴C-fatty acids + ¹⁴C-ethanol), thereof 19.5% ± 2.1 as ¹⁴C-acetate, corresponding to up to -50.7 nmol · gDW^-1 · day^-1 ± 8.9. H2 stimulated the production of ¹⁴C-labeled acetate from ¹⁴C-bicarbonate more than 10-fold, starkly arguing for homoacetogenesis (Supplementary data, Figure 2). The stimulation could be followed by higher consumption rates of both ¹⁴C-labeled (data not shown) and unlabeled CO2 (Figure 2B), even resulting in CO2 net consumption. At the same time, the production of ¹⁴C-ethanol, -propionate and –butyrate from both, [2-¹⁴ C]-acetate and ¹⁴C-bicarbonate, was stimulated, but to a lower extent (Supplementary data, Figure 2).

CHCl3, which among others inhibits the acetyl-CoA-pathway (Chidthaisong and Conrad, 2000; Conrad and Klose, 2000), decreased the accumulation of ¹⁴C-labeled acetate from

14C-bicarbonate by around 30% compared to the controls (Supplementary data, Figure 2).

Methanogenesis was completely inhibited, and H2-partial pressures increased steadily (Figure 2B). FCH2COO^-, which inhibits the citric acid cycle (Chidthaisong and Conrad, 2000), decreased [2-¹⁴C]-acetate turnover by >80% (Supplementary data, Figure 1) and completely inhibited ¹⁴C-bicarbonate incorporation into fatty acids or ethanol (Supplementary data, Figure 2). FCH2COO^- completely inhibited methanogenesis and kept H2-partial pressures below 20 Pa during incubation time (Figure 2B). In addition, H2, FCH2COO^- and CHCl3 decreased CO2-accumulation (H2 > FCH2COO^- > CHCl3; Figure 2B).

Acetate accumulated, but did not exceed CH4 production rates (Table 2). FCH2COO^-, CHCl3 and H2 impeded acetate accumulation (FCH2COO^- > H2 > CHCl3; Table 2). At the same time, FCH2COO^- and H2 led to an increase in propionate and butyrate accumulation (FCH2COO^- > H2, Table 2). Considering carbon balance, the stimulating effect of FCH2COO^- on propionate production corresponded to lower CH4 and CO2 accumulation rates (Figure 2B, Table 2). CHCl3 hardly affected propionate and butyrate accumulation (Table 2).

Summed up, the experiment revealed the presence of both AAcO and homoacetogenesis in control samples. Homoacetogenesis was strongly stimulated by H2. Methanogenesis was

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

completely based on H2/CO2. In addition, the production of propionate and butyrate from both, CO2 and acetate was detected. Like homoacetogenesis, these processes were stimulated by H2. In contrast, methanogenesis was inhibited by high H2 partial pressure, indicating competition between methanogens and homoacetogens. In further experiments the effect of shaking the incubation tubes and the inhibiting effect of high H2 partial pressure were investigated. Total methanogenesis and methanogenesis from ¹⁴ C-bicarbonate were not affected by CH3F. However, the minor amount of ¹⁴CH4 originating from [2-¹⁴C]-acetate further decreased when samples were incubated with CH3F, indicating that CH3F might have an unknown side-affect.

Experiment III.

Since shaking during incubation led to a reversible inhibition of methanogenesis (see material and methods), we propose that methanogenesis depends on aggregate formation with other organisms building up syntrophic consortia. In order to prove this hypothesis, peat slurries were incubated for 28 days (i) at 150 rpm until day five, (ii) at 150 rpm from day 5 on, and (iii) without shaking. Indeed, methanogenesis was reversibly inhibited regardless of the sequence of shaking and non-shaking (Figure 4A). In addition, CO2

consumption in control samples exceeded CO2 accumulation from day 20 on (4B). This resulted in a lower CO2/CH4 ratio compared to samples that were shaken during incubation (Figure 4C) thus indicating that CO2 scavenging was more effective in non-shaken samples. Phase contrast microscopy revealed that cell aggregate formation only occurred in samples that were not shaken during incubation.

Experiment IV.

The inhibition of hydrogenotrophic methanogenesis by incubation under a H2 atmosphere (Figure 2B) was reproduced in a further experiment with peat sampled in 2005 (data not shown). In order to check the reason for this effect, we incubated peat slurries under different gas atmospheres and followed the turnover of ¹³CO . Hydrogenotrophic

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

hydrogenotrophic methanogenesis and homoacetogenesis were strongly stimulated compared to controls (0.4 ± 0.02 µmol · gDW^-1 · day^-1)when samples were incubated with 80% H2/20% CO2 (1.2 ± 0.08 µmol · gDW^-1 · day^-1), and to a lower extent when incubated with 20% H2/80% CO2 (0.9 ± 0.17 µmol · gDW^-1 · day^-1; Table 3). Incubation with 100%

CO2 did not alter total methanogenesis but slightly stimulated acetoclastic methanogenesis (0.1 ± 0.02 µmol · gDW^-1 · day^-1) compared to control samples (100% hydrogentrophic methanogenesis), which was shown by inhibition with CH3F (data not shown). None of the treatments led to changes in pH.

The high portion of hydrogenotrophic methanogenesis in control samples was further illustrated by stable isotope signatures, which were measured repeatedly during the experiment. Values for δ¹³C of CH4 ranged between –50 and –68 ‰, being consistent with the dominance of hydrogenotrophic methanogenesis (Whiticar, 1999).

In order to investigate AAcO in peat sampled in 2005, we again followed the turnover of [2-¹⁴C]-acetate without and with CH3F. CH3F did not affect methanogenesis but reduced for unknown reasons ¹⁴CH4 production (data not shown). AAcO measured as ¹⁴CO2

production was affected only slightly. In accordance with former results, shaking during incubation inhibited methanogenesis. Surprisingly, AAcO was not affected by shaking (data not shown).

Structure of the microbial community.

In peat sampled in 2003, 30% of all T-RFs retrieved when the experiment began, exhibited a 92-bp fragment, 30% a 185-bp fragment, 10% a 393-bp fragment, and 30% a 494-bp fragment (Figure 5). After 16 days of incubation 50% of all T-RFs exhibited a 92-bp fragment. The relative frequency of the 185-92-bp and 393-92-bp fragment did not change, whereas the relative T-RF frequency of the 494-bp fragment became less important (10%).

In peat sampled in 2004, 45% of all T-RFs retrieved when the experiment began exhibited a 92-bp fragment, 20% exhibited a 185-bp fragment, 10% a 393-bp fragment and 5% a 494-bp fragment (Figure 5). The T-RF frequencies of the 92-bp fragment showed the highest relative abundance in all treatments (40-70%). The relative frequency of the 185-bp fragment decreased compared to start samples (3-8%), whereas the relative abundance

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

of the 393-bp fragment increased (9-30%). Further fragements (76-bp, 82-bp, 105-bp, 807-bp) were negligible (Figure 5).

Based on the phylogenetic analysis of the Archaeal 16S rRNA gene (52 clones in 2003 and 15 in 2004) and on calculated clone frequencies, samples from both years were dominated by methanogens using H2/CO2 as substrate (2003: 86%; 2004: 60%; Methanobacteriales, Methanomicrobiales, Rice Cluster I, Rice Cluster II; Figure 6). Sequences of Rice Cluster II were present in samples from 2003 with a clone and T-RF frequency of around 20%.

Rice Cluster I occurred in both years (clone frequency: 13%). Crenarchaeota and Rice Cluster V only played a minor role.

An analysis of the clone sequences in silico showed that the 92-bp RF and the 185-bp T-RF could be affiliated with Methanobacteriales and Methanosarcinales, respectively. The 393-bp fragment could be affiliated with Rice Cluster I and Methanomicrobiales, the 105-bp fragment with Rice Cluster I, the 494-105-bp fragment with Rice Cluster II, the 76-105-bp and 807-bp fragments with Rice Cluster V and the 793-bp and 810-bp fragment with Crenarchaeota. Hence, the major archaeal groups detected by T-RFLP analysis were covered by the clone library.

The phylogenetic analysis of the McrA amino acid sequences (45 clones in 2003 and 38 in 2004) showed a lower diversity (Figure 7). However, the percentage of methanogens using H2/CO2 as their sole substrate (2003: 62.2%; 2004: 79%) again corresponded to the high percentage of hydrogenotrophic methanogenesis. The phylogenetic analysis of the FTHFS amino acid sequences from experiment II (2004) showed that all sequences clustered within a completely acetogenic cluster in group A, as defined by Leaphart et al. (Leaphart et al., 2003) (Figure 8).

Discussion

This is the first detailed study focusing on the interaction of anaerobic acetate oxidation,

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

discovered a highly dynamic system, which is sensitive to even slight changes in substrate concentrations.

Due to the refractive character of Sphagnum-peat, turnover rates were low (Figure 2A and B; Table 2). However, since low turnover rates seem to increase the H2 scavenging capabilities of methanogens, the presence of syntrophic processes may become more probable (Eichler and Schink, 1984). Methanogenesis was mainly based on H2/CO2 (60-100%, Figure 2, Table 2), as shown for other acidic peatlands (Duddleston et al., 2002;

Hines et al., 2001; Horn et al., 2003; Lansdown et al., 1992; Metje and Frenzel, 2005;

Williams and Crawford, 1984). The high percentage of hydrogenotrophic methanogenesis was mirrored in the 16S rRNA and mcrA clone frequencies (Figure 6 and 7). In contrast to the well-studied Turnagain Bog in Alaska (Duddleston et al., 2002; Hines et al., 2001), acetate was not a dominating end product: its concentration never exceeded that of CH4

(Table 2).

Which process was responsible for acetate depletion? Our tracer experiments evidenced that AAcO played a considerable role, at least in peat sampled in 2003. It was also detected in samples taken in 2000, 2001, 2004, and 2005, but with a lesser (2004) or unknown importance. Since the specific inhibition of acetoclastic methanogenesis with CH3F (Frenzel and Bosse, 1996; Janssen and Frenzel, 1997; Penning and Conrad, 2006) did not affect AAcO (Figure 3), acetoclastic methanogenesis could not have been involved. A few acetogens are known being capable of syntrophic acetate oxidation in co-culture with methanogens. Dependent on the H2 partial pressure, the acetyl-CoA pathway is used for both syntrophic acetate oxidation and homoacetogenesis (Lee and Zinder, 1988; Schnürer et al., 1996; Schnürer et al., 1999; Zinder and Koch, 1984). Hence, the ongoing production of ¹⁴CH4 from [2-¹⁴C]-acetate with CH3F (60% in Experiment I compared to the controls) was attributed to the reduction of syntrophically produced ¹⁴CO2 (Figure 3). This is in accordance with the lack of alternative electron acceptors, and further supported by thermodynamic calculations. Hence, we conclude that AAcO was the main process preventing acetate accumulation even in the absence of acetoclastic methanogenesis.

Even though the methanogenic population was relatively constant (Figure 5), the system switched between different terminal processes in peat sampled in 2003 and 2004,

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

respectively, depending on even slight changes in H2 partial pressure (Figure 2B, Table 2).

Besides CH4 and acetate, propionate and butyrate occurred as end products (Table 2, Supplementary data Figure 1 and 2) and were favoured by high H2 concentration, too.

Butyrate can be a by-product of homoacetogenesis (Lynd and Zeikus, 1983), and propionate may originate from the reductive carboxylation of acetate as it has been shown for Desulfobulbus propionicus (Laanbroek et al., 1982).

In contrast to peat sampled in 2003, AAcO played only a minor role in peat from 2004, whereas homoacetogenesis was more important. ¹⁴C-tracer experiments even argued for a recycling of acetate between AAcO and homoacetogenesis (Supplementary data Figure 1 and 2). Corresponding to this, Gibb’s free energies indicated thermodynamic equilibrium between syntrophic acetate oxidation and homoacetogenesis (Table 1): during incubation,

∆G values for both processes shifted between +/–12 and +/- 5 kJ ⋅ reaction^-1. Anaerobic bacteria can operate close to the thermodynamic equilibrium, but even slight changes of substrate or product concentrations may be sufficient to shift the metabolic system into the opposite direction (Schink, 1997).

In sediments and sludges, optimal metabolite transfer can be achieved by forming aggregates of different microorganisms (Conrad et al., 1985; Conrad et al., 1986). Indeed, aggregates were observed microscopically, and shaking obviously disrupted aggregates and terminal mineralization (Figure 4). We conclude that methanogens depended on juxtaposition to their syntrophic partners. Previously this has been also shown for rice paddy soils (Dannenberg et al., 1997). Aggregates may have facilitated acetate cycling even if bulk measurements indicated thermodynamically unfavourable conditions. Until today, homoacetogens are the only known microorganisms being capable of syntrophic acetate oxidation via a reverse acetyl-CoA pathway. Indeed, we could show the potential for homoacetogenesis by ¹⁴C-acetate formation from ¹⁴CO2 (Supplementary data, Figure 2). In addition, we analyzed the fhs-gene. All sequences clustered within an exclusively acetogenic group (Leaphart et al., 2003) and apart from SO42- reducers, which are also

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

energies (Table 1). Formate can largely be excluded as an intermediate; it was not detected.

In general, mainly H2 is assumed to be exchanged in aggregates in order to grant a high flux at low H2 partial pressures. However, since CO2 consumption exceeded CO2

production in non-shaken samples (Figure 4B and C) also CO2 scavenging by methanogens seems to be more effective in aggregates. This suggests that aggregate formation facilitates the exchange of both H2 and CO2.

The fractionation of stable carbon isotopes during hydrogenotrophic methanogenesis is related to ∆G, with hydrogenotrophic methanogenesis becoming more favourable with decreasig αCO2-CH4 (Penning et al., 2005). In Figure 9, the empirical α-∆G relations are compared to a semi-Gauss-curve derived from culture experiments (Penning et al., 2005).

Following the argumentation in Penning et al., the H2 partial pressure measured in the headspace probably underestimates the partial pressure in the aggregates. Hence, αCO2-CH4

can be supposed to give a more realistic estimate of the true ∆G experienced by the methanogenes, while the calculated ∆G is merely an apparent value. The stippled lines projecting the endpoints of the measured αCO2-CH4-∆G relations onto the semi-Gauss-curve suggest that hydrogenotrophic methanogenesis became less favourable with time (Figure 9). Since H2 was exchanged within aggregates, we suppose that the competition for H2

became stronger during time course of the experiment.

However, not only H2 is exchanged in aggregates, but CO2, too. Corresponding to this and against all thermodynamic predictions (Table 1), incubating under high H2 partial pressure inhibited hydrogenotrophic methanogenesis by 60 to 80% (Figure 2B, Table 2). According to this, it has been shown previously in rice root incubations that CH4 production by Rice Cluster I was suppressed by high H2 partial pressure (Lu et al., 2005). This was proposed to be due to the inability of Rice Cluster I methanogens to adapt to artificially high H2

concentrations. Since Rice Cluster I was also recovered in our peat (Figures 7 and 8), this may be a reason for inhibition of methanogenesis with high H2 partial pressure. However, in parallel to the inhibition of methanogenesis, homoacetogenesis was stimulated more than 10-fold, considering ¹⁴C-bicarbonate turnover (Supplementary data Figure 2), and Gibb’s free energy for homoacetogenesis became sufficiently exergonic for the formation of 1 mol ATP · reaction^-1 (Table 1). According to current opinions, methanogens and

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

2001; Schulz et al., 1997). However, correlating with homoacetogenesis, CO2 net accumulation changed into CO2 net consumption from day 20 on (Figures 2B). Incubating under 80% H2/20% CO2 released the inhibition of methanogenesis. This suggests that even under high H2 partial pressure homoacetogens seem to be more effective CO2 scavengers than methanogens, leading to CO2 partial pressures below the threshold of methanogenesis.

In principle similar, yet at different gas mixing ratios, stimulation of hydrogenotrophic methanogenesis in samples from the acidic McLean Bog was only successful when homoacetogens were inhibited by rifampicin (Bräuer et al., 2004).

We got also indication for an yet unknown acetate oxidising process that might contribute to AAcO: CHCl3, which is known to inhibit the acetyl-CoA pathway (Chidthaisong and Conrad, 2000), only slightly affected the oxidation of [2-¹⁴C]-acetate (Supplementary data Figure 1). In contrast, the incorporation of ¹⁴C-bicarbonate into ¹⁴C-acetate was partly inhibited (Supplementary data 2). FCH2COO^-, which is an inhibitor of the aconitase in the citric acid cycle (Chidthaisong and Conrad, 2000; de Graaf et al., 1996), completely inhibited [2-¹⁴C]-acetate oxidation (Supplementary data Figure 1), but also homoacetogenesis (data not shown). Acetate may have also been oxidized via humic acid reduction, as indicated by the effect of AQDS (Figure 1B). However, a final conclusion about this alternative pathway is not yet possible.

Concluding remarks

Peat samples collected over five years showed all the potential for syntrophic acetate oxidation. Syntrophic acetate oxidation was largely coupled to methanogenesis. It came along with a methanogenic community dominated by hydrogenotrophic methanogens, and with a high ratio of CH4 originating from H2/CO2. We propose that at least part of AAcO was attributed to homoacetogens: homoacetogenesis was stimulated under elevated H2

partial pressure, and sequences of the fhs gene clustered between two completely homoacetogenic groups indicating that they represented homoacetogens. Even if samples

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

either involving fermenting or humin-reducing bacteria. In summary, our results demonstrate how important it is covering more than just one moment in time to analyze the microbial potential in an ecosystems.

Experiments comparing agitated with non-agitated samples suggested that AAcO depended on aggregate formation. Labelling experiments showed that not only H2 was exchanged between the syntrophic partners, but also CO2. Under elevated H2 partial pressure homoacetogens seem to scavenge CO2 more effectively than methanogens.

Hence, CO2 can become an important regulator for methanogenesis and homoacetogenesis.

Acknowledgements

The authors thank Peter Claus, Alexandra Hahn, and Jochen Scheld for excellent technical help with the pilot experiments. Edgar Karofeld introduced us to mire ecology and helped with peat sampling. This study was supported by a grant from the Deutsche Forschungs Gemeinschaft (SFB 395, ‘‘Interaction, Adaptation and Catalytic Capabilities of Microorganisms in Soil’’).

Material and Methods Sampling site and peat

Peat was sampled from a raised bog (Männikjärve Bog) at the Endla Nature Reserve in central Estonia (58°52’27.9’’N; 26°15’6.8’’E). The mean annual temperature in this region is around 4.8°C, and mean annual precipitation is 589 mm (www.klimadiagramme.de).

The samples were taken from Sphagnum-cuspidatum-hollows located in the central ridge-hollow-complex. The only other macroscopic plant beside Sphagnum cuspidatum was Rhynchospora alba.

For the first pilot study peat was sampled in October 2000 below the water-table between 10 and 30 cm deep. Peat was stuffed below the water-table into glass jars and closed with metal lids while still submersed. The peat was stored at 4°C in the dark till use for the pilot experiments some months later. For the second pilot study peat was sampled from the same site in April 2001. Samples were taken below the water table (20-40 cm) and filled

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

bubble-free into sterile polyethylene terephthalate bottles (Nalgene^®). Samples were stored at 4°C until analysis.

Due to other sampling campaigns disturbing this site, we had to shift later on to another Sphagnum-hollow about 30 m apart, but with the same topological and botanical characteristics. Peat was sampled in October 2003, August 2004, and November 2005, respectively. Loss on ignition was 98%. The pH varied between 4.4 in 2003 and 3.6 in 2004, respectively.

Incubation

For the first pilot experiment, about 10mL peat were filled into 60-mL serum bottles, capped with butyl stoppers and flushed with N2. The bottles were incubated at 16°C in the dark. Headspace concentrations of CH4, CO2, and ¹³C-signatures of CH4 and CO2 were measured regularly. Three bottles served as a control while four were amended with a N2 -bubbled solution of [2-¹³C]-acetate, increasing the pore water acetate concentration by about 10 µM. For the second pilot experiment, 30 g peat were filled into sterile 120-mL serum bottles, capped with butyl stoppers and flushed with N2 in an anaerobic box. Three replicates were amended with (i) distilled water, (ii) AQDS (anthraquinone-2, 6-disulfonate) to a final concentration of 20 mM, (iii) BES (2-bromoethane sulfonate) to a final concentration of 40mM, and (iv) AQDS and BES to a final concentration of 20 and 40 mM, respectively. Another four replicates of treatment (ii) were autoclaved. All solutions had been bubbled with N2 and autoclaved under N2. Finally, a [2-¹³C]-acetate solution was injected to a concentration of 10µM. The bottles were incubated at 25°C in the dark. Headspace concentrations of CH4, CO2, and ¹³C-signatures were measured regularly starting about one hour after adding [2-¹³C]-acetate.

Further experiments were carried out with peat sampled in October 2003 (experiment I), August 2004 (experiments II & III) and November 2005 (experiment IV), respectively.

Peat samples were homogenized as described previously (Metje and Frenzel, 2005), and

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

1997; Penning and Conrad, 2006). An oxygen-free solution of sodium BES (final concentration 40mM) was used to completely inhibit methanogenesis (Metje and Frenzel, 2005) (controls, n=3; CH3F, n=3; BES, n=3). Headspace concentrations of CH4, CO2 and H2 were measured by gas chromatography after equilibrating gas and liquid phases.

Samples were incubated for 17 days. CH4 and CO2 were measured daily, H2 three times during incubation. Pore water samples were taken at day 0, 8 and 16 and analyzed for fatty acids and alcohols. Similarly, slurry samples were taken to measure Fe(II) concentrations.

In order to determine acetate turnover, we followed the turnover of [2-¹⁴C]-acetate sodium salt (57.0 mCi/mmol, Amersham Life Sciences, UK) or ¹⁴C-bicarbonate (53.0 mCi/mmol, Moravek Biochemicals, Brea, CA, USA). The total radioactivity per tube was 60507 Bq for [2-¹⁴C]-acetate or 65891 Bq for ¹⁴C-bicarbonate (controls, n=3; CH3F, n=2, BES, n=2).

In experiment II with peat sampled in 2004 we applied additional inhibitors (CHCl3 and FCH2COO^-) to follow carbon flow. Here, the total radioactivity per tube was 165000 Bq for [2-¹⁴C]-acetate and 55000 Bq for ¹⁴C-bicarbonate (controls, n=4; FCH2COO^-, n=4;

CHCl3, n=4; H2, n=4). Samples were incubated for 29 days. CH4 and CO2 were measured every second day, H2 was measured eight times during incubation. Pore water samples were taken at day 0, 5, 17, and 27. Since a high H2 partial pressure quenches syntrophic acetate oxidation (Hattori et al., 2001; Hattori et al., 2005), we checked if H2 was the intermediate by incubating peat samples with 1.5 bar H2. In order to guarantee H2 -equilibrium between gas and liquid phase, samples were shaken during incubation (150 rpm). However, this led to a reversible inhibition of methanogenesis. Therefore, in experiment III samples were incubated at 25 rpm from day 6 on. To further evaluate this effect peat samples were incubated for 28 days (i) at 150 rpm until day five, (ii) at 150 rpm from day 5 on, and (iii) without shaking.

CHCl3 was used to inhibit methanogenic activity (Chidthaisong and Conrad, 2000; Chin and Conrad, 1995; de Graaf et al., 1996). In addition, CHCl3 has been shown to inhibit the activity of homoacetogenic bacteria and acetate-consuming sulfate-reducing bacteria (Scholten et al., 2000). A dose-response experiment showed that CHCl3 completely inhibited methanogenesis at all tested concentrations (20 to 200 µM), whereas CO2 -accumulation was not affected (data not shown). CHCl3 was added to a final concentration of 70 µM by syringe. FCH2COO^- has been applied to distinguish between acetoclastic and

3. Ergebnisse

__________________________________________________________________________________________________________________________________________________

hydrogenotrophic methanogenesis, or to study the relative importance of acetate as the key intermediate in organic carbon mineralization (de Graaf et al., 1996; Schulz and Conrad, 1996). FCH2COO^- was added to a final concentration of 0.5 mM.

Experiment IV aimed to study the effect of different gas atmospheres with and without inhibitors. Peat sampled in 2005 was incubated under 100% N2 (BES, CHCl3, CH3F), 100% H2 (CHCl3, CH3F), 80% H2 20% CO2 (BES, CHCl3, CH3F), 20% H2 80% CO2

(CH3F) or 100% CO2 (CH3F), respectively. Here, hydrogenotrophic methanogenesis and homoacetogenesis were followed by the addition of ¹³C-bicarbonate. ¹³CO2 was obtained from completely dissolving ¹³C-sodium-bicarbonate (min. 98 atom %; Campro Scientific, Berlin, Germany) in 85% ortho-phosphoric acid (Fluka Chemie AG, Buchs, Switzerland).

The solution was prepared in a 150mL serum bottle, which had been flushed with N2 and evacuated. ¹³CO2 was taken from the headspace with a gas tight Luer-Lok^TM syringe (Becton Dickinson EDC, Temse, Belgium) and injected into sample flasks. Removed

13CO2 was balanced by saturated O2-free NaCl solution. Start headspace concentrations of CO2 in samples with 100% N2 and 100% H2 were 43.8 ppmv ± 0.8 (n=23), corresponding to 70.0 ¹³C-atom% ± 1.67 (n=23). In samples with 80% H2 20% CO2, 20% H2 80% CO2 or 100% CO2, respectively, ¹³C-bicarbonate was added at 8.5 atom% ± 0.34 (n=24). Slurry samples were incubated in 150 mL serum bottles with natural rubber stoppers (Ochs, Göttingen) at 25°C for 14 days. Controls were run in quadruplicates and inhibitor experiments in triplicates.

CH4, CO2 and H2 headspace partial pressures were measured as described above. Pore water samples for fatty acids and alcohol measurements were taken on day six and 14. The pH was measured at start and end of the experiment. In order to investigate AAcO also in peat sampled in 2005, we followed the turnover of [2-¹⁴C]-acetate without and with CH3F.

In addition, we investigated the effect of shaking during incubation on methanogenesis and AAcO (n=3).

Im Dokument Terminale Prozesse des anaeroben Abbaus in sauren Torfmooren der Subarktis und des südlichen Boreals (Seite 89-136)