Microbial biomass and community structure were investigated in two composts using the phospholipid fatty acid technique. The composts consisted of shredded straw of Miscanthus with the addition of pig slurry to give an initial C:N ratio of about 25. Samples were taken following changes in the compost temperature (at 5°C intervals) during the first month of composting and additionally after 2 and 3 months. The total microbial biomass, measured as total amount of phospholipid fatty acid, peaked after 1 day with about six times the initial values. The temperature also peaked after 1 day, being above 60°C, and then slowly declined to around 25°C over 3 months. Microbial biomass was approximately halved during this time. When the total amount of phospholipid fatty acid was separated into indicator phospholipid fatty acids for different groups of microorganisms, these groups showed different patterns during the composting process. Gram-positive bacteria increased rapidly with increasing temperature and decreased with decreasing temperature. Gram-negative bacteria and fungi increased initially up to a temperature around 50°C, but decreased during the extreme heating phase. When the temperature declined to about 50°C, the amounts of phospholipid fatty acids indicative of these two groups increased again. The phospholipid fatty acid indicative of actinomycetes, 10Me18:0, was at a low level during the whole experiment, but increased slightly during the last month of composting. The development of the microbial community in the two composting systems was similar during the initial thermophilic phase of the composting process, but the communities after 3 months differed.
Composting is not only a way to reduce human waste and recycle nutrients, but can also produce useful components, e.g. compost for mushroom production [1, 2], soil conditioners [3, 4], or growth media for horticultural plants [5–8]. Since bacteria and fungi are the main organisms responsible for the decomposition, much effort has been put into understanding the changes in the microbial biomass, community structure and activity during the composting process.
A variety of methods have so far been used to investigate the microorganisms during composting. These include traditional plating and identification of the culturable microorganisms [9–14], and more recent techniques measuring ATP content , microbial biomass [16, 17] and potential metabolic abilities determined by the BIOLOG sole carbon utilization test .
One technique, that gives an indication of the microbial community composition without the culturing of organisms on agar media, is the direct analysis of the phospholipid fatty acid (PLFA) pattern of a sample. The PLFA analysis is based on the fact that different subsets of a microbial community differ in their fatty acid composition. Although few PLFAs can be considered to be absolute signature substances for a single species or even a specific group of organisms, it is possible to get indications of changes in major groups, like fungi, actinomycetes, and Gram-positive and Gram-negative bacteria [19, 20]. The PLFA pattern of a compost sample can therefore be regarded as an integrated measurement of all organisms in that sample. The PLFA analysis has been used to study changes in the microbial community in a variety of natural terrestrial environments [21–25], but to our knowledge, only two studies of changes in the microbial community in composts have been carried out using PLFA analysis [26, 27]. Both these studies concentrated on the period after peak temperatures had been reached, and thus did not test if the PLFA technique is sensitive enough to detect rapid changes during the initial temperature increase. Furthermore, neither of these two studies investigated composts consisting of straw material and a nitrogen source.
We have investigated the changes in the microbial biomass and community structure by analysing the composition of the PLFAs during 3 months of composting with special emphasis on the early, thermophilic phase. This was achieved by taking samples following the changes in temperature (at about 5°C intervals) until the temperature reached the level of the surroundings. In particular, we wanted to follow the development of different taxonomic groups of microorganisms (actinomycetes, fungi, Gram-positive and Gram-negative bacteria). Furthermore, since it is well known that there would be rapid changes, first to a thermophilic and then to a mesophilic community, with presumably different PLFA patterns, this would enable us to estimate turnover rates of PLFAs from dead/inactive groups of organisms under natural conditions.
2Materials and methods
2.1Compost set-up and sampling
The material consisted of shredded straw (2–5 cm pieces) of Miscanthus×ogiformis Honda ‘Giganteus’ (Andropogoninae: Poaceae) to which pig slurry was added to give an initial C:N ratio of about 25. The moisture content was adjusted to about 70%. The compost was packed loosely in an open 800 l wooden box (48 kg dry weight of straw) and in a closed 500 l reactor (32 kg dry weight) with forced areation (5 l h−1). Both types of containers were insulated. Water was added twice a week to keep the moisture content stable and the compost in the box was turned every 14 days to insure aerobic conditions. The compost containers were placed in a greenhouse where the temperature was kept between 15 and 20°C during the experiment.
The temperature in the center of the containers was monitored continuously and sampling was done with temperature intervals of about 5°C during the heating phase, beginning just after mixing of the compost. After the heating phase samples were taken after 48, 51, 53, and 93 days of composting in the reactor, and after 51, 55, 78, and 93 days in the box.
In the reactor, the samples were taken through a hole in the middle, close to the center. In the box, the samples were taken in the center at a depth of about 30 cm. At each sampling four subsamples were taken at randomized positions, where the measured temperature corresponded with that of the center. The subsamples were pooled, mixed, and stored at minus 20°C until the extraction took place.
Dry weight was measured after drying for 24 h at 70°C. Organic matter content (OM), based on the dried samples, was calculated as loss on ignition at 550°C for 5 h.
Prior to the extraction procedure 0.5–1 g wet weight of the samples were ball-milled.
The lipid extraction, fractionation, mild alkaline methanolysis and GC analysis used here were described in detail by Frostegård et al. [23, 28]. To summarize shortly: lipids were extracted from the compost samples in a one-phase mixture of chloroform/methanol/citrate buffer and the polar lipids were separated using silicic acid columns, followed by a mild alkaline methanolysis to form fatty acid methyl esters before GC analysis. Fatty acids were designated in terms of total number of carbon atoms:number of double bonds, followed by the position of the double bond from the methyl end of the molecule. The prefixes a and i indicate anteiso and iso branching, cy indicates a cyclopropane fatty acid and methyl branching (Me) is indicated as the position of the methyl group from the carboxyl end of the chain.
The sum of the following fatty acids was considered to represent Gram-positive bacteria: i15:0, a15:0, i16:0, i17:0, a17:0 and 10Me17:0 . The Gram-negative bacteria were represented by the sum of: 16:1ω7t, 16:1ω5, cy17:0, 18:1ω7 and cy19:0. 16:1ω5 and 18:1ω7 are also found in arbuscular mycorrhizal fungi [30, 31], which, however, are very unlikely to appear in compost. As markers for actinomycetes and fungi 10Me18:0 and 18:2ω6,9 were used, respectively [19, 20]. The fungal to bacterial PLFA index ratio was calculated by dividing the amount of 18:2ω6,9 with the sum of bacterial PLFAs. The PLFA pattern of dried straw was analyzed initially. 16:1ω7 and 16:0 constituted over half of the PLFAs found and all other PLFAs were below 10 mol%.
The mole percents of the PLFA values from both composting systems were subjected to a principal component analyses after scaling each variable by dividing with its standard deviation.
The temperature increased quickly after mixing of the compost material, reaching peak values within 20 and 26 h for the box and the reactor, respectively. Then the temperature gradually declined to ambient level over about 3 months (Fig. 1A,B). The organic matter content decreased from 92 to 72% of dry weight, and pH decreased from 8.5 to 7.4 in both systems during this time. Peak temperature in the reactor system was 69°C, and in the box system 64°C. The total amount of phospholipid fatty acids (totPLFA) covaried with the temperature, increasing very rapidly from a starting value around 500 nmol g−1 OM to approximately 6 times higher amounts at or shortly after peak temperature. Then totPLFA decreased gradually, resulting in around 1500 nmol g−1 OM after 3 months of composting. However, after about 30 days the totPLFA increased, without any simultaneous increase in temperature. This was most notable in the reactor system (Fig. 1A). At this time, a large number of fruit bodies of Coprinus cinereus (Schaeffer: Fries) s.f. Gray were present indicating that a major fungal biomass must have been present in the compost.
In total, 32 different PLFAs were detected during the composting process, and 28 of these were identified. The development of three representative PLFAs during composting in the reactor is shown in Fig. 2. The amounts of PLFAs 18:1ω9 and cy19:0 increased until the temperature reached about 50°C, but then they decreased during the period of highest temperature. However, when the temperature had decreased to about 40–50°C (after about 10–20 days of composting), both these PLFAs started to increase again, indicating growth of Gram-negative bacteria and fungi. The PLFA i17:0 only appeared in greater amount when the temperature was above 50°C. The peak value of this PLFA was almost 20 times the initial value. The temperature almost decreased it to initial values at the end of the experiment.
Most PLFAs could be grouped into the same patterns as these illustrated in Fig. 2, both in the reactor and the box system. Thus, PLFAs with a pattern similar to i17:0 (that is with peak abundance during the peak temperature) were i15:0, i16:0, and a17:0, while a range of PLFAs had a pattern with increasing levels when the temperature was below 50°C, including PLFAs indicating eukaryotic organisms (18:2ω6,9, 18:1ω9, 20:4), Gram-negative bacteria (cy17:0, cy19:0, 16:1ω5) and actinomycetes (10Me18:0). Three unidentified PLFAs were also found in this group.
The major shifts in the microbial community during the composting process could be ascertained using marker PLFAs (expressed as mol%) for main taxonomic groups of organism (see Section 2.2 for specific marker PLFAs) (Fig. 3A,B). In total, these marker PLFAs amounted to between 72 and 89% of the totPLFA. The PLFAs representing Gram-positive bacteria increased very rapidly with increasing temperature, especially i16:0. As the temperature started to decrease the amount of Gram-positive PLFA also decreased. During the whole period the amount of actinomycetes (as judged from the amount of 10Me18:0) was at a low level, with only a small increase towards the end of the experiment. The fungi (represented by 18:2ω6,9) increased until temperature reached 50°C, then decreased during the extreme heat phase, and finally started to increase again when the temperature decreased. This was especially evident in the box system (Fig. 3B). The PLFAs representing Gram-negative bacteria showed the same pattern as 18:2ω6,9.
In total, the PLFAs considered to be mainly of bacterial origin accounted for between 34 and 73% of the total amount of PLFAs during the composting period. The changes in the relative amount of fungal and bacterial biomass during composting could be shown by estimating a fungal to bacterial PLFA index ratio over time. In the reactor system, this ratio was 0.37 from start. During the heating phase, the ratio decreased very quickly, reaching a minimum value of 0.007 after 6 days. Then the ratio started to increase again, but never exceeded 0.1. The box system showed the same pattern as the reactor system until day 30, but then a much more pronounced invasion of fungi was evident (Fig. 3B), resulting in a maximum ratio of 0.34 after about 50 days of composting.
To compare the development of the microbial communities of the two composting systems, a principal component analysis of the whole data set (using mol% PLFA) was made (Fig. 4). The first and last sampling (the latter after 3 months), as well as the samplings at peak temperatures are shown in the graph. The lines are graphical representations of the timing of the changes of the whole microbial community. The two first principal components (PC) appeared to represent different communities. The second PC (characterized by high amounts of i15:0, i16:0, i17:0, a17:0, and 17:0) reflected the thermophilic community. PC1 represented the mesophilic community that developed after temperatures had fallen below 50°C. Samples with high values for this PC were characterized by relatively high amounts of especially 10:Me16:0, 10:Me17:0, 10:Me18:0, cy17:0, cy19:0, 16:1ω5, 18:1ω9 and 20:4.
The degree of fatty acid unsaturation in both composting systems decreased during the heating phase and increased during the following phase with lower temperature, although the degree of unsaturation was not completely restored to initial values in the reactor system even after 3 months of composting (Fig. 5A). In addition, the ratio between PLFAs with 16 and 18 carbon atoms showed the expected changes, with almost equal amounts of C16 and C18 in the mesophilic phases, but relatively high amounts of C16 during and just after the thermophilic phase (Fig. 5B). In Fig. 5C, the changes in the ratio between iso and anteiso forms of 15:0 and 17:0 are shown. The ratio for 15:0 increased until the temperature reached maximum levels, in both the reactor and the box system. When the temperature started to decrease, the iso/anteiso ratio also decreased slightly, but during the last 2 months, the ratio increased again, ending with a clear dominance of the iso form. The ratio for 17:0 in the box system showed the same pattern as 15:0, but this was not the case in the reactor system. Until the temperature reached around 50°C the iso/anteiso ratio was 0.8:1.0, but then the anteiso form dominated during the rest of the experiment.
TotPLFA has been suggested as a measure of the total microbial biomass , and has been shown to correlate well with the biomass measurements using the fumigation extraction technique during the period after peak temperature in a municipal solid waste compost . In the present experiment, the gradual decrease in biomass, estimated as totPLFA, after peak temperatures (Fig. 1A,B) is consistent with results obtained by others [18, 26], although we did not find the decrease to be as drastic as in these studies.
TotPLFA increased rapidly during the first day of composting, indicating a doubling time of the microbial biomass in both systems of 6–7 h. However, this is certainly an underestimation of the growth rate for the thermophilic part of the community during this period, since totPLFA in the initial samplings included PLFAs that might partly be of plant origin (e.g. 16:0) or PLFAs from microorganisms not growing at higher temperatures (e.g. 18:1ω7, Fig. 2). Using only PLFAs indicative of Gram-positive bacteria (assuming these to be the main part of the thermophilic community, see below), a doubling time of approximately 4 h was found during the first day of composting.
The pattern of a fast increase in microbial biomass immediately after the composting process started, followed by a gradual decrease might be explained by the availability of nutrients. Initially fresh substrate was colonized rapidly, but as easily degradable nutrients became exhausted a gradual decrease in biomass took place. Although the availability of nutrients thus determined the development of the total microbial biomass over time, different subsets of microorganisms were clearly selected for by the temperature regime during the composting process (Fig. 3A,B). The thermophilic community developing above 50°C appeared to consist of Gram-positive bacteria, as judged from the increase in iso- and anteiso-branched PLFAs (Figs. 2 and 3A,B). These PLFAs are common in the genus Bacillus, a genus well known to be dominant in composts at high temperatures [13, 14, 32]. Strom , for example, identified 87% of isolated bacteria at 40–69°C as Bacillus spp. Iso- and anteiso-branched PLFAs were also prevalent during the thermophilic phase of municipal waste composts , and branched-chain fatty acids increased much during the initial phase with high temperatures in an open windrow compost .
Gram-negative bacteria and fungi appeared to grow only below 50°C. An increase in PLFAs, indicative of these groups of microorganisms, was observed before the temperature had reached peak values, and especially during the cooling phase, where they partly outcompeted the thermophilic, Gram-positive bacteria (Figs. 2 and 3A,B). The increase in totPLFA after about 30 days was assumed to be caused by the invasion of Coprinus cinereus. The increase was partly caused by an increase in 18:2ω6,9 and 18:1ω9, PLFAs common in fungi. Species of Coprinus are often observed on composts .
Actinomycetes (as judged from the proportion of the PLFA 10Me18:0) only increased in the later phase, when the compost temperature was relatively low. This differed from the municipal waste compost studied by Herrmann and Shann , where high relative amounts of 10Me18:0 were found already during the phase with the highest temperature regime.
One must bear in mind that the development of different groups of microorganisms during composting inferred from the changes in the PLFA pattern does not give absolute amounts of biomass of different groups, since conversion factors from the amounts of PLFAs characteristic for the different microorganism groups to actual biomass are lacking. Another problem could be the contribution of PLFA from the straw material, but since the main fraction of PLFA in Miscanthus straw are 16:1ω7c and 16:0, which are not used as marker PLFAs, and the other PLFAs present are divided on all the groups of microorganisms except actinomycetes, the error made should be minor. Furthermore, this error will decrease with time as the straw material is decomposted. Thus, the patterns shown in Fig. 3 must be interpreted with caution, although the main dynamics of the different groups are well-supported by data obtained with traditional isolation techniques [9, 34–37].
Some conclusions on dominant groups can, however, be drawn from the changes in the fungal to bacterial PLFA index ratio. This ratio varied from 0.02 to 0.04 in different agricultural soils low in organic matter to between 0.3 and 0.5 in forest soils . The latter are believed to be dominated by fungal biomass. Since these high ratios are similar to that found in the box system at the end of the composting process, it appears that fungi were the dominant microorganism group at the end of the composting process. On the other hand, at peak temperatures values for the fungal to bacterial PLFA index ratio was below 0.01. This was lower than found in any soil, indicating that, at this point, it was a bacterial-dominated system. Overall, the variation in the fungal to bacterial PLFA index ratio during the composting was as large or even larger than that found in a range of very different soils , emphasizing the drastic changes in microbial community structure that occurred during composting.
An altered PLFA pattern during composting may be due to changes in species composition or due to adaptation to the different temperature regimes by one and the same community. Although changes in e.g. chain lengths or degree of unsaturation of the PLFAs were consistent with changes expected from the different temperature regimes ([38–41], Fig. 5A,B), we believe that the altered PLFA pattern was mostly due to changes in specific groups of organisms. The changes found during the first day of composting, for example, were changes from a mesophilic to a completely different thermophilic community (Fig. 3A,B and Fig. 4), and the major changes in the PLFA pattern during the decrease in temperature was also found between PLFAs from different groups of organisms. A similar conclusion regarding interpretation of the PLFA data was drawn by Hellman et al. .
An increase in a specific PLFA must be related to an increased growth of a species or a group of organisms having that specific PLFA. However, a decrease in a certain PLFA can be caused by at least two mechanisms. First, the organisms are outcompeted and die, and second, the phospholipids have to be broken down in order not to be included in the analyses. Thus, a decrease in a PLFA characteristic of a certain group of microorganisms will always underestimate the real death rate of that organism group. We lack information about the rate at which phospholipids are broken down both in soils and in composts, although studies in marine sediments indicated a fast turnover rate . However, our data suggests a fast turnover rate in the compost. The PLFA 18:1ω9 (common in fungi) that initially increased to around 200 nmol g−1 OM, decreased to less than 50 nmol g−1 OM in 5 days time when temperatures were to high for fungal growth (Fig. 2). This suggests a half-life of phospholipids with this fatty acid of about 1 day during this period. Likewise, i17:0 decreased from a maximum of 600 nmol g−1 OM to around 100 nmol g−1 OM in about 2 weeks, indicating a half life of less than a week. Thus, the decrease in PLFAs during the composting process probably would indicate the death rate of the different groups of organisms rather well.
One objective of the present work was to study if the PLFA technique was sensitive and reproducible enough to be used to monitor the rapid development of the microbial community during the initial phase of the composting process. The very similar development in the two composting systems before, during and shortly after the thermophilic phase (e.g. Fig. 4) indicated that this was the case. Furthermore, the PLFA analysis can also be used to indicate differences during the later stages, since differences in the microbial community in the two composting systems could be detected after three months of composting using PCA (Fig. 4). These differences were caused by differences in the two systems. First, the reactor, as a closed system with forced aeration, remained moist and well aerated, and the material was therefore not disturbed. The open box system without aeration, on the other hand, could become anaerobic and it also tended to dry out due to evaporation, and therefore the material had to be turned and watered regularly. This procedure is known to speed up the composting process . The microbial community in the box system was more altered than in the reactor system (more to the right at the end of composting in Fig. 4), showing that the PLFA pattern can be used to indicate the maturity of the compost. This will be investigated further in the future.
We thank Thomas Læessoe for identification of the Coprinus species, Erling Hyldig for sampling, and Gosha Afshar and Else M. Andersen for technical assistance. We also thank Ulrik Soechting for valuable comments, and Andrea Gargas for revising the English in this paper. This study was supported by a grant from The Danish Research Councils (M.K.).