Thermophilic methane production and oxidation in compost

Abstract

Methane cycling within compost heaps has not yet been investigated in detail. We show that thermophilic methane oxidation occurred after a lag phase of up to one day in 4-week old, 8-week old and mature (>10-week old) compost material. The potential rate of methane oxidation was between 2.6 and 4.1 μmol CH₄ (gdw)⁻¹ h⁻¹. Profiles of methane concentrations within heaps of different ages indicated that 46–98% of the methane produced was oxidised by methanotrophic bacteria. The population size of thermophilic methanotrophs was estimated at 10⁹ cells (g dw)⁻¹, based on methane oxidation rates. A methanotroph (strain KTM-1) was isolated from the highest positive step of a serial dilution series. This strain belonged to the genus Methylocaldum, which contains thermotolerant and thermophilic methanotrophs. The closest relative organism on the basis of 16S rRNA gene sequence identity was M. szegediense (>99%), a species originally isolated from hot springs. The temperature optimum (45–55 °C) for methane oxidation within the compost material was identical to that of strain KTM-1, suggesting that this strain was well adapted to the conditions in the compost material. The temperatures measured in the upper layer (0–40 cm) of the compost heaps were also in this range, so we assume that these organisms are capable of effectively reducing the potential methane emissions from compost.

Thermophilic, Methane oxidation, Methane production, Compost, Methanotrophs, Methylocaldum szegediense

1 Introduction

The most important anthropogenic methane sources are rice paddies, coal mining, enteric fermentation and landfills. The source strength of landfills is estimated to be 22–36 Tg CH₄ a⁻¹[1–3]. In the last decades, methods of mitigation or controlling the methane emission from these sources have been investigated. Composting has been proposed as one method to reduce methane emission from landfills [4,5]. This method of processing organic waste has become an integral part of modern waste management, with the objective of reducing the mass and volume of the biodegradable portion of solid wastes. The resulting compost is mainly used as organic fertilizer. At present, more than 600 full-scale composting plants produce nearly 6–7 million tons of biological waste in Germany [6] and measurements show considerable methane emissions from compost piles [7–10]. These emissions are influenced by the dimensions [10] and density [11] of the heaps, and by the input material [12]. Methane emissions from composting facilities were estimated to be 7.4 × 10⁶ t CH₄a⁻¹[13], which amounts to 0.31–0.44% of the total methane emissions in Germany.

Little is known about the thermophilic microorganisms involved in methane production and oxidation in these environments. In one study, it was shown that mushroom compost piles contain 2 × 10⁸ thermophilic methanogens per gram dry matter, which were mainly identified as Methanobacterium thermoautotrophicum[14]. However, both methane production and oxidation determine the net methane emission to the atmosphere. The oxidation is performed by methane-oxidising bacteria (MB). Phylogenetically MB belong to the Gammaproteobacteria (Methylococcaceae) and Alphaproteobacteria (Methylocystaceae and genera Methylocapsa and Methylocella) [15–19].

Many studies have addressed the balance between production and oxidation of CH₄, for example in flooded rice fields [20,21], and have examined the CH₄ oxidation process in the oxic soil surface layer and the rhizosphere [22–24]. These studies showed that up to 90% of the CH₄ produced is oxidised before it can reach the atmosphere. Under mesophilic conditions, composted material used as landfill cover material showed a methane-consuming potential, or increased the methane oxidation potential of the landfill cover soil [25,26]. Due to self-heating, the temperature in compost is >55 °C, and because the piles are regularly turned, these high temperatures act on the entire material. Mesophilic methanotrophs are therefore unlikely to mitigate methane emission from these piles. The existence of thermophilic methanotrophs in composting material, however, was indicated in one study [27]. In this, it was supposed that thermophilic methanotrophs were responsible for nitrification activity. Molecular analyses also indicated that Methylocaldum was present in landfill cover material that originated from compost material [28].

So far only few investigations have been carried out on methane oxidation under thermophilic conditions. On the basis of accepted definitions [29,30], the methanotrophs preferring hot conditions can be differentiated into thermotolerant (<45°C optimum, growth >45 °C possible), moderately thermophilic (45–60 °C) and strictly thermophilic (>60–70 °C). Four strains of methane-oxidising bacteria, Methylococcus capsulatus, “Mc. Ucrainicus”, Methylocaldum gracile and Md. tepidum, can be classified as thermotolerant; two strains, Mc. thermophilus and Md. szegediense, are moderately thermophilic; and one strain, “Methylothermus” sp. HB, is strictly thermophilic [31].

This paper reports about methane production and consumption in developing compost piles and the impact of thermophilic methanotrophs on the methane emission from these compost piles.

2 Materials and methods

2.1 Composting plant

The composting plant examined is an open windrow composting system, with the modification that the composting process is carried out under a foil (Gore-tex^®) during the initial three weeks. The optimal oxygen supply for microorganisms was ensured via ventilation during this phase. The input material of the plant consisted of about 70% organic material from communal bio-waste collection and 30% loppings (from tree-cutting) and grass. After storage under the gore-tex^® blanking-cover, the piles were set up as trapezoid-shaped piles again (dimension: 30 m length, 3 m width and 1.8 m height). These piles were turned once or twice per week. The mature compost was sieved after 9–13 weeks of processing. Rough-textured material was recycled and transferred to the input material of a new compost process.

2.2 Vertical profiles of in situ methane concentration and temperature in compost piles

The methane mixing ratio in different depths was measured by sampling gas from 20, 50, 80, 110 and 140 cm depth in 4-, 6- and 8-week old as well as in mature compost. Steel capillaries of different lengths (with an inner diameter of 1 mm) were fixed on a wooden post (Fig. 1), which was driven into the pile. Gas samples (30 ml) were then taken with plastic syringes connected to the steel capillaries via butyl tubing and a three-way valve. Syringes were closed with valves and gas samples were analysed as described below. Two profiles were analysed for each pile. Temperatures were measured using a NiCrNi-temperature sensor (1 m) connected to a digital thermometer (J. Dittrich Elektronic, Baden-Baden, Germany).

1

Technical drawing of the gas sampler used for measuring the in situ profile of the methane concentration in compost piles.

2.3 Compost sampling

Three sampling points were selected, one in the centre and one at each end of the pile. About 300 g of compost material was collected at each sampling point from a depth of 20 cm below the surface (at temperatures between 45 and 55 °C). The material was mixed well and sieved (4-mm mesh) in the lab.

2.4 Potential methane oxidation and methane production

To measure the potential methane-oxidising activity, 2 g of mixed and sieved compost samples from heaps aged 4, 8 and 9–12 weeks were incubated under ambient air in 120-ml serum bottles closed with butyl stopper. Methane was added to a final mixing ratio of about 18,000–20,000 ppmv. Incubation was carried out at 25 or 50 °C in the dark. Gas samples (0.3 ml) were taken repeatedly (5 times within 4 days) after the methane addition to determine the rate of potential methane oxidation. Apparent first-order oxidation rate constants were calculated from the exponential decrease of CH₄ in time (Microsoft^®Excel2000) and converted to CH₄ oxidation rates by multiplication with the CH₄ mixing ratio (20,000 ppmv).

To measure the potential methane production, 2 g of compost material from heaps aged 3, 4, 5, 6, 7, 8 and 11 weeks were incubated at 50 °C after flushing the head space with N₂. Accumulated methane was quantified after a 5-day incubation period. Each experiment was carried out in triplicate.

2.5 Temperature effect on methane oxidation and methane production by composted material

To examine the temperature effect on the potential methane oxidation of composting material, 2-g amounts of mature compost were incubated under ambient air in 120-ml serum bottles closed with butyl stoppers. Methane was added to a mixing ratio of about 20,000 ppmv. Samples were incubated at 25, 36, 45, 50, 55, 60 and 70 °C, respectively, in the dark. Each temperature assay was carried out in triplicate. The amount of oxidised methane was quantified by gas chromatography after 28 h of incubation.

To measure the temperature effects on the potential methane production, 2 g of a mixed compost sample from a six-week-old heap was incubated as described above under a dinitrogen atmosphere. Methane production rates were calculated by linear regression analysis using the software package Microsoft^®Excel 2000. Rates are expressed per gram dry weight (gdw) of compost material.

2.6 Gas sampling

The gas samples were taken with gas-tight pressure-lock syringes (A-2 series; Dynatech, Baton Rouge, LA, USA). The amount of gas withdrawn for sampling was replaced by dinitrogen. Methane was analyzed by gas chromatography (Perkin Elmer Autosystem XL B5902), using a flame ionization detector according to [32]. Detection signals were analysed with the integration program Peak Simple (SRI Instruments, CA, USA). Standard gases containing 10, 99 and 1000 ppmv methane (Messer, Germany) were used as reference. The FID-signal was linear up to 4 × 10⁵ ppmv.

2.7 Methane emission estimation

Methane mixing ratios at 20-cm depth and atmospheric mixing ratios (2 ppmv) at the surface were used to estimate methane emission rates from compost piles using Fick’ s first law.

 

where F_methane is the emission density (gCH₄ m⁻²h⁻¹), D_{methane/compost} is the gas diffusion coefficient for CH₄ within the compost (m² h⁻¹), and d(CH₄]/dz is the concentration gradient (g CH₄ m⁻³ compost m⁻¹). The diffusion coefficient was obtained by

 

where D_CH4/air is the diffusion coefficient for CH₄ in air (0.07056 m² h⁻¹ at 9.0 °C [33], φ_g is the air-filled porosity (air m⁻³ of compost), and a and b are dimensionless coefficients accounting for the compost pore tortuosity and size distribution, respectively.

The air-filled porosity φ_g is given by

 

where ɛ is the total porosity (m³m⁻³) and WFPS the water-filled pore space (m³H₂Om⁻³) [34]. D_methane/air is directly dependent on temperature (T) and inversely dependent on pressure (P) [35]. However, pressure effects do not need to be accounted for, since all of the gas samples were taken at the same site. The temperature correction was:

 

where D_CH4/air (T) is the diffusion coefficient of CH₄ at a particular temperature T (in K). To evaluate D_{methane/compost}, approximate values were required for the dimensionless coefficients a and b. The most common values for these empirical coefficients used in soil CH₄ oxidation studies are: a= 0.9 and b= 2.3[35,36].

Methane emission rates were also calculated using CH₄ mixing ratios measured at 60 or 80 cm depth and atmospheric mixing ratios (2 ppmv) at the surface (F_60/80). The ratio of these two fluxes (F₂₀/F_60//0) was used to estimate the percentage of the total potential methane emission that was reduced by methanotrophic activity.

2.8 Extraction of microorganisms from compost material

Ten grams of the freshly collected mixed and sieved compost material were diluted in 90 ml of a sodium pyrophosphate solution (0.18% w/v). The compost dilution was homogenized for 30 s in a sterile Waring-blender (Breda Scientific, Cenco, Meerbusch, Germany) at low power. Fifty millilitre of the homogeneous dilution was transferred to sterile polyethylene sampling tubes (Greiner, Nürtingen, Germany). MPN-analysis or cell fixation was carried out using the supernatant after sedimentation of coarse material for 15 min according to [37].

2.9 MPN, cultivation and isolation of methanotrophs

A serial decimal dilution up to 10⁻⁸ was made in glass tubes containing 4.5 ml of sodium chloride solution (0.18% w/v) for Most-Probable-Number counting according to [38] in quintuplicate. Mineral salts media used for cultivation were prepared according to [39] (Medium A) and [40] (Medium B). Tubes were incubated at 50 °C in closed glass desiccators containing a gas mixture of 20% CH₄, (v/v), 10% CO₂ (v/v) and 70% (v/v) air. A tube from the highest positive dilution step (visible turbidity) was used as the inoculum for further cultivation and isolation on plates on the described media, which were solidified with 1.5% agar. Strain KTM-1 was isolated on medium A after repeated striking.

2.10 Temperature effect on growth and methane oxidation by strain KTM-1

To examine the temperature effect on the potential methane oxidation of the isolated strain KTM-1, sterile serum (120 ml) bottles were filled with agar-medium (B) up to a height of 0.5 cm. With an inoculating loop, one drop of pure liquid culture was spread over the surface. The bottles were then closed with butyl stoppers and filter-sterilised methane (80–100 μmol) was added. Duplicate samples were incubated at 25, 36, 45, 50, 55, 60, and 70 °C, respectively, in the darkness. Growth was analysed visually and the amount of methane oxidised was measured by gas chromatography after eight days of incubation.

2.11 Phylogeny

The 16S rRNA gene of strain KTM-1 was amplified by PCR using different “universal primers”[41] and sequenced. On the nucleic acid level, evolutionary distances between pairs of sequences were calculated by using the Kimura 2-parameter model provided in the MEGA 2.1 [42] software package. Phylogenetic trees were constructed by using the neighbour-joining method supplied by the MEGA 2.1 package.

3 Results and discussion

3.1 Methane production and its regulation

The in situ temperatures of developing compost are shown in Fig. 2(a). Corresponding to the pile age, the in situ temperature at a depth of 20 cm increased from 53 °C in a 4-week old pile up to 65 °C in a mature compost pile. With exception of the 4-week old pile, the temperatures increased also with increasing profile depth. The highest temperature was reached in the 6-week old pile, with almost 78 °C at a depth of 1 m.

2

Vertical profiles of the temperature (a) and the in situ CH₄ mixing ratio (b) in 4-week (♦), 6-week (△), 8-week (▪) old compost piles, and in a pile of mature compost (◻) from a municipal composting plant.

In comparison to the temperature profiles of the various compost piles, Fig. 2(b) shows the corresponding methane mixing ratios for piles of different ages. The lowest methane mixing ratio of about 10 ppmv was observed in the 4-week old pile, while the highest one of >10⁵ ppmv has been detected in the mature compost pile. Thereby the methane mixing ratios gradually increased from 8 ppmv in a 4-week old pile up to 10⁴–10⁵ ppmv in a mature pile, respectively (Fig. 2(b)). Corresponding to the temperature profiles, the methane mixing ratios increased exponentially with the profile depth, starting from 20 cm, in all piles examined.

Due to the initially high microbial activity the easily available organic compounds are converted into biomass and heat. The temperature rises dramatically, as generally observed in compost heaps [43]. This high temperature is maintained due to periodic turning and/or to the use of controlled air flow. The heat dissipation at the heap surface explains the decreasing temperature from the heap centre to the heap surface. After the easily degradable compounds are metabolized, the temperature is decreasing because heat dissipation surpasses the heat production.

Composts are known to host thermotolerant fungi and a variety of thermotolerant and thermophilic bacteria. So far, these bacteria have been mainly investigated by cultivation studies and identified as aerobic bacteria like Thermus thermophilus, Bacillus sp., Hydrogenobacter sp., Thermoactinomyces sp. [43].

Due to the lower oxygen solubility in water at these high temperatures, the compost piles also seem to favour the development of anoxic niches during thermophilic phases, so that a basic requirement for thermophilic methanogenic activity is given as well.

The incubation of compost material of a 6-week old pile at different temperatures under dinitrogen atmosphere showed the highest methane production rate at 60 °C (0.9 μmol(gdw)⁻¹h⁻¹), whereas at 36 and at 70 °C the methane production potential was only about 18% of this value (Fig. 3(a)).

3

Temperature effect on the potential methane production rates of 6-weekold compost material incubated under a dinitrogen headspace (a), and temperature effect on the methane oxidation potential of a mature compost material incubated under oxic conditions with 60–80 μmolCH₄(gdw)⁻¹compost⁻¹ (b) Values represent mean of triplicate incubations ± SD.

These results indicate the presence of a highly adapted methanogenic population to this hot environment, as was also shown by [14], where 2 × 10⁸ thermophilic methanogens per gram dry matter were detected in mushroom compost and identified mainly as M. thermoautotrophicum. Further studies showed a simultaneous increase of methane emission from heaps and ether lipid contents in thermophilic compost material. It has been found [8,44] that methane is produced in compost only under high temperatures. These results indicate that the environmental conditions in compost induced the enrichment of anaerobic microorganisms, like thermophilic methanogens. The high in situ methane concentrations measured in this study facilitate earlier studies that showed considerable methane production in compost piles [6–9].

3.2 Methane oxidation

While at 25 °C low or even no potential methane oxidation was observed (Fig. 4(a)), at 50 °C all examined compost samples showed a methane-consuming activity, either immediately or after a lag-phase of up to 20 h (Fig. 4(b)). According to compost age and the increasing methane mixing ratio, repectively, the consumption rate slowly increased from 2.6 μmolCH₄(gdw)⁻¹h⁻¹ to 4.1 μmolCH₄(gdw)⁻¹h⁻¹(Figs. 2(b) and 4(b)). The measured methane oxidation rate was more than threefold higher compared to the highest detected methane production potential, which was found at 60 °C in 6-week old material (Fig. 3(a)). The temperature optimum of the methanotrophic community in the studied compost material was in the range of 45–55 °C (Fig. 3(b)).

4

Time course of the methane concentration in compost incubation experiments under oxic conditions at 25 °C (a) or 50 °C (b). The compost material was sampled from 4-week old (♦), 8-week old (▪) and mature compost piles (▲). Values are means of n= 3, error bars show one standard deviation.

The results show that, according to the high substrate concentrations and under the given environmental conditions, a moderate thermophilic methanotrophic community has been developed. This is astonishing, since the water solubility of methane and oxygen decreases with increasing temperatures, and thus limits methanotrophic growth [31]. However, the oxygen solubility in water at 60 °C is still about 51% of the solubility at 20 °C [45], and the existence of methanotrophs in thermophilic environments was shown earlier [46,47].

The first results of methanotrophs in compost material were presented by [48]. The potential for methane oxidation of compost material was shown and additional compost increased the methane oxidation potential of landfill cover soils to which it was added [25,26]. However, all these experiments were carried out under mesophilic conditions, in which we found comparatively low or even no potential for methane oxidation (Fig. 4(a)). So, the activity of mesophilic methanotrophs in this study can be excluded.

The in situ activity of methanotrophs is indicated by the exponential decrease of methane mixing ratios from deeper layers to the pile surfaces in all compost heaps examined (Fig. 2(b)).

This result was supported by the quantification of thermophilic methanotrophic bacteria via the MPN-method.

In the upper layer of the mature compost material, 1.8 × 10⁶ to > 3.2 × 10⁷ cultivable methanotrophic cells per g dw were detected. Methanotrophs contributed thus to 0.1% to >3% of the total DAPI-counted cells in cell extracts from mature compost (1 × 10⁹ cells (gdw)⁻¹). However, quantification of methanotrophs using the methane oxidation rates (Fig. 5) resulted in higher values of about 2 × 10⁹ methanotrophs per g dw. This calculation assumes that the V_max cell⁻¹ of the methanotrophs is approximately 1.5 × 10⁻⁹μmolh⁻¹cell⁻¹[49]. From these and the earlier results of [50], who detected 10¹⁰–10¹¹ total cells g⁻¹, we assume that the cell extraction by pyrophosphate yielded only <10% of the total cell numbers.

5

Methane production potential of compost material with different ages. Values represent the methane accumulation after an incubation of five days, given as a mean per gram dry weight in triplicate incubations.

To our knowledge, these are the first results on quantification of moderately thermophilic methanotrophs in an anthropogenic environment. The abundance of methanotrophic bacteria is comparatively high in the examined compost. Population sizes of 10⁵–10⁶ methanotrophs per g, for example, were found in paddy soils [39,51], sizes of 1 × 10⁶ to 9 × 10⁷ methanotrophs per g were found in sediments of a storage lake [39] and >10⁶ methantrophic cells per g were found in acidic Sphagnum peat [17]. A similar high population compared to that in compost was detected in freshwater lake sediment (up to 10⁹ cells (gdw)⁻¹) [49]. Due to the only weak increase in potential oxidation rates with the compost age (Fig. 4(b)), it seems that the methanotrophic population is quite stable during the whole compost maturation. The organic material which was not completely degraded in one composting process was added to a new composting process, so that inoculation of fresh compost with thermophilic methanotrophic bacteria took place on a regular basis. Based on this knowledge, it is not surprising that a stable methanotorophic community, adapted to the given environmental conditions, can be developed.

3.3 Methanotrophic isolate

From the last positive tube of the MPN dilution series we isolated one methanotrophic strain (KTM-1). The temperature optimum for methane oxidation by strain KTM-1 compared to the temperature optimum measured in the compost material was almost identical (Fig. 3 and Table 1). This isolate was identified by 16S rDNA sequence analysis as a strain of Md. szegediense (Fig. 7).

1

Growth and methane oxidation of strain KTM-1 incubated at different temperatures

Temperature (▪C)	Growth	Methane oxidised (μmold−1120 ml−1)
25	−	0
36	+	0.5 ± 1.5
45	+++	3 ± 2.1
50	+++	11.5 ± 0.7
53	+++	10.1 ± 0.1
55	+	8.25 ± 1.3
60	−	0

−no growth; + weak growth; +++ strong growth.

7

Dendrogram showing the phylogenetic affiliation of strain KTM-1 to selected species of methanotrophs within the group of Gammaproteobacteria (gamma) and Alphaproteobacteria (alpha). The bracket encompasses thermotolerant to thermophilic methanotrophs. The dendrogram was constructed based on 16S rRNA gene sequences available from the European Molecular Biology Laboratory databank (accession numbers are given in brackets) and was constructed after multiple alignment of data by ClustalX [53]. Distances (distance options according to the Kimura-2 model) and clustering were calculated with the neighbour-joining method using the software package Mega (Molecular Evolutionary Genetics Analysis) version 2.1 [17]. Bootstrap values based on 1000 replications are given at the branching points.

From these results we suggest that strain KTM-1 was the abundant and active methanotroph in the compost environment. We conclude that KTM-1 is responsible for mitigating methane from the compost piles studied, especially because the temperatures in the upper surface layer (20–40 cm from the surface) ranged from 52 to 65 °C, the optimal range for methane oxidation by this strain. The resistance of KTM-1 to a temperature of 80 °C (data not shown), which is similar to Md. szegediense[52], shows that the highest observed temperature (78 °C) within the compost piles (Fig. 2(a)) should not lead to a strong reduction of viable cells of this species and favour the development of a stable community.

The increasing Cu⁺⁺ concentration per mass of compost, following the microbial degradation of organic material, could be another factor enhancing the development of a Methylocaldum sp. population in compost piles. Growth stimulation for Md. szegediense was observed when additional Cu⁺⁺ was added to the culture media [52].

Md. szegediense was previously isolated from a greenhouse heating system, fed directly from a hot spring connected to a natural gas field [52]. However, compost also seems to be a favourable environment for this species, since in another study [28] hints were found as well for the presence of Methylocaldum sp. in compost material. The habitat description for members of the genus Methylocaldum[52] therefore should be expanded to compost material. However, it will be no surprise to detect this genus in soils that have been fertilized with compost.

3.4 Methane emission

Calculated methane emissions are shown in Fig. 6. The methane emission continuously increased with the pile age, from 2.4 mgCH₄m⁻²d⁻¹ from a 4-week old pile to 2500 mgCH₄m⁻²d⁻¹ from a mature pile, respectively. These calculated rates are comparable to the emission of 6000 mgCH₄m⁻²d⁻¹, measured via a closed chamber method (data not shown) and to earlier studies [8], or to the emission from small heaps [10].

6

Methane emission rates (●) from compost piles of 4-, 6-, and 8-week old compost heaps and a mature compost pile, and % of methane produced (♦) which was oxidized before it reached the atmosphere.

Up to now, only few studies have been reported on methane emissions from composting systems, which makes extrapolation to national or international scales very difficult. Emission rates between 0 and 119 gCH₄m⁻²d⁻¹ from different heaps were shown in a study which documents the high variability due to different practical management of composting facilities [10]. So far, factors influencing the methane emissions from compost piles are pile dimension [10], pile density [11] and the consistence of input material [12]. From this study, methanotrophs seem to be an obvious biological factor regulating the emission from compost piles as well.

Using the methane profile data, calculations showed that 46% of the methane produced in the 4-week old pile, and 84–98% of the methane produced in the 6- and 8-week old and mature compost piles was oxidised by the present methanotrophic community before it reached the atmosphere (Fig. 6).

The relevant methanotrophic organism, which we suggest to be species of Md. szegediense, has a high impact on the methane emission from the piles. This mitigation potential by methanotrophs is comparable to other biological methane sources which are also regulated by methanotrophic activity, such as paddy fields or landfills [22–24]. Our results indicate once more the importance of methanotrophic activity on the methane emission from anthropogenic methane sources.

4 Conclusion

Composting has been proposed as a method to reduce methane emission from landfills [4,5]. Nevertheless, compost piles themselves were shown as sources for atmospheric methane. The extrapolation to national or international scales is very difficult due to the limited number of data. Our results show the activity of thermophilic methane producers and, as a first study, the activity of thermophilic methane oxidisers, which regulates the methane emission from compost heaps. These results should stimulate further studies on composting systems as an interesting environment for thermophilic methane cycles.

Acknowledgements

We thank Katja Grebing for excellent technical assistance and Dr. Peter Dunfield for proofreading the manuscript.