3.1. Changes in physicochemical characteristics during co-composting
The main physicochemical parameters of the co-composting piles are shown in Fig. 1. Compost temperature is used as an indicator of microbial activity throughout the composting process, and also an indicator of the maturity and stability of the compost product (Zhang et al., 2020). Among the four treatments, CM + B group arrived the highest temperature very fast (Fig. 1a). There were four phases in CM + B composting, mesophilic phase (< 55°C, day 0–7), thermophilic phase (55–62°C, day 7–32), cooling phase (40–55°C, day 32–41), mature phase (< 40°C, day 41–60). The temperature rose rapidly to 55°C on day 7 and reached 62°C after 20 days, reflecting active microbial degradation during the mesophilic and thermophilic phase. Diverse labile substances (e.g., carbohydrates and proteins) were decomposed rapidly, which released a large amount of heat (Zhou et al., 2018). The thermophilic phase (> 55°C) lasted for approximately 25 days, followed by a cooling phase and a maturation phase during which the temperature gradually declined to approximately 36°C upon the end of the composting. The temperature of the other groups arrived the highest on day 30–45, prolonging the composting period. Water content in all the treatments decreased because of the increasing temperature and ventilation in the composting process (Fig. 1b). The water content of CM + B group dropped the fastest from 70–80–55% during the thermophilic phase because of the sharp temperature change, and gradually stabilized to about 50% in the cooling and mature stage. Compared with CM + B group, the variation of moisture content in other treatments was less dramatic because the temperature increase was slow. Similar to temperature, the pH showed an increasing trend possibly due to the release of NH3 from microbial metabolism (Reyes-Torres et al., 2018), and then fell gradually (Fig. 1c). The highest pH of CM + B, CM + B + N, CM + B + N + S was around 8.99–9.20, which reflected a balance of acid production and ammonia accumulation in the composting piles. The final EC value was observed to be higher than the original value in all treatments (Fig. 1d), which was consistent with the previous studies (Liu et al., 2018). They also fell in the desired range for a mature compost product and unlikely to be phytotoxic (Wang et al., 2020a). Treatment of CM + B also exhibited the highest EC value, increasing from initially 0.55 mS cm-1 to 2.8 mS cm-1. As Fig. 1e and 1f, the contents of TOC and TKN in all treatments declined in the mesophilic phase because of the aerobic degradation of macromolecules and the rapid loss of CO2 and NH3. Then an increase of TOC and TKN was observed in CM and CM + B during thermophilic phase, which may be caused by decomposition of organic matter and N2 fixation (Zhou et al., 2018). Consequently, C/N ratio declined rapidly in the mesophilic stage of composting and stabilized in the cooling and maturation phase (Fig. 1g). GI value greater than 80% is generally considered an indicator that compost is mature and phytotoxicity-free (Yu et al., 2011; Zeng et al., 2009; Zhang et al., 2018). During composting process, GI in CM and CM + B groups kept increasing and met the standard of mature compost product, while GI in CM + B + N and CM + B + N + S was too low to be satisfactory maybe because of the addition of different nitrogen source compared to the other two groups (Fig. 1h).
3.2. Analysis of humic substances
The fluorescence of organic matter would be influenced by the presence of condensed aromatic rings and/or unsaturated aliphatic carbon chains (Yu et al., 2019). Previous studies revealed that the redshift in the maximum fluorescence intensity could be attributed to an increase in aromatic group condensation in these molecules (Lv et al., 2013; Said-Pullicino et al., 2007). The contours of DOM from the four composting piles (CM, CM + B, CM + B + N, CM + B + N + S) showed five peaks marked as A, B, C, D and E (Fig. 2). Peak D and E (Em < 380 nm, Ex < 250 nm) in the early composting process represents aromatic proteins. The peak C (Em > 380 nm, Ex < 250 nm), peak B (Em < 380 nm, Ex > 250 nm) and peak A (Em > 380 nm, Ex > 250 nm) are attributed to fulvic acid, water-soluble microbial metabolites and humic acid, respectively. The fluorescence intensity of peak B decreased from thermophilic stage followed by an increase. The fluorescence intensity of peak D and peak E showed a downtrend while the fluorescence intensity of peak A and peak C showed an uptrend. That suggested during the composting process, the primary reactions were the transformation of protein and water-soluble microbial metabolites to fulvic acid and humic acid, especially during mesophilic phase. After that, the tendency gradually shifted to the stabilization of the newly formed humic acid-like and fulvic acid-like organic materials during the curing and mature phases. Compared with the other treatments, CM + B treatment exhibited an extremely higher fluorescence intensity of peak A. That was an indicator of high DOM conversion efficiency during composting process, possibly due to the appropriate nutrient ratio and fast temperature rise in CM + B. The efficient transformation of DOM was also consistent with GI data because humic substances benefit the seed germination.
3.3. Similarity and diversity of microbial communities
A total of 31, 081 bacterial sequences and 36, 761 fungal sequences from all samples were analyzed after quality filtering. Venn diagrams was used to analyze the similarity of the microbial communities in the samples of the four treatments (Fig. 3a&b). Specially, in CM + B there were 1826 bacterial OTUs and 434 fungal OTUs identified respectively, of which 346 and 168 were unique. That indicated the highest diversity of microbial community in CM + B and significant differences from other treatments. A dramatic variation of bacterial community was also in CM + B observed suggested by the sobs index (Fig. S1A), while change of fungal community was expected to be similar in CM + B, CM + B + N and CM + B + N + S (Fig. S1B). Based on the above results, the treatment CM + B was selected for dynamic analysis of microbial community in compost samples.
The similarity of samples at different time points of CM + B composting process was also analyzed by Venn diagram and PCoA. Bacterial diversity increased significantly in the mesophilic phase and gradually decreased in the cooling phase, while fungal diversity decreased continuously as composting progressed (Fig. 3c&d). Samples of day 7 and day 20 exhibited the most unique bacterial OTUs of 544 and 269 respectively. The fungal taxonomic richness was found to be much lower than the bacteria. Only 6 fungal OTUs were identified from fungal OTU libraries, including Pseudeurotium, Mycothermus, and some other unclassified_k_Fungi. According to PCoA, the close distance between day 40 and day 60 for both bacteria and fungi indicated the stable status of the microbial community after the cooling phase. The short distance between day 7 and day 20 in the bacterial analysis indicated that the bacterial community had a significant change in the mesophilic phase (Fig. 3e), while there was an obvious change in the fungal community on day 20 according to the PCoA (Fig. 3f), indicating that the fungal community was significantly influenced by thermophilic phase in the co-composting process.
3.4. Bacterial community succession and predicted bacterial functions
At phylum level, a closer observation of the bacterial community revealed eight dominant groups in the four composting systems (Fig. 4a). A similar profile of bacterial communities was observed in CM + B + N and CM + B + N + S group, indicating the insignificant effects of sheeting. During mesophilic phase, CM showed no obvious change of bacterial community, while a dramatic change was observed in CM + B possibly due to the fast elevation of temperature. Firmicutes was the most richness phyla in all these compost samples, which plays a major role in the degradation to lignocellulose similar to previous report (Li et al., 2019). Further observation of microbial community analysis at genus level showed the bacterial community structure in CM + B changed dramatically during co-composting process. Acinetobacter (42.5%), Pseudomonas (25.57%), and Psychrobacillus (8.86%) were the most abundant genera in the original co-compost mixture (Fig. 4b). Acinetobacter are ubiquitous in nature and have been found to exist in various environments, including activated sludge, sewage treatment plants and raw wastewater, by different research groups in Australia, Portugal, Korea and Pakistan (Al Atrouni et al., 2016). Pseudomonas aeruginosa is the third most common nosocomial pathogen that is associated with chronic, eventually fatal lung disease in cystic fibrosis (CF) patients, while Pseudomonas syringae and pathovars are prominent plant pathogens (Goldberg et al., 2008). After 7 days of co-composting, those genera disappeared and Bacteroides (19.09%), Paeniclostridium (5.71%), Corynebacterium (5.15%), and Romboutsia (4.70%) became the main genera. The relative abundances of Symbiobacterium, Bacillus, and Thermoflavimicrobium increased by 98.7%, 92.1%, and 94.9% from day 7 to day 20. In contrast, Bacteroides and Corynebacterium declined by 98.7% and 99.2%, respectively, in the thermophilic phase. From day 40 to day 60, Lysinibacillus (13.97–16.61%), Solibacillus (10.42–13.90%), and Acinetobacter (12.12–13.75%) were dominant genera. Lysinibacillus can be used as a potential compost inoculant as it can mediate the variation of C/N ratio, change the activities of carbon- and nitrogen-cycling enzymes, and accelerate the composting process (Ganguly & Chakraborty, 2018). It can also act as an indicator of mature product as its dominant abundance only appeared in the maturation phase. Solibacillus has a high capacity to degrade lignocelluloses and lignin (Huang et al., 2019). Acinetobacter harbinensis, a heterotrophic nitrifying bacterium, has the ability to remove ammonium (Qin et al., 2017). The bacterial community structure of other composting groups in genera level was showed in Fig. S2. A-C.
The metabolic potential of bacterial communities during CM + B composting was evaluated by PICRUSt based on the Clusters of Orthologous Groups (COG) database. As predicted, there were three main functional groups, including metabolism, information storage and processing, and cellular processes and signaling (Fig. 4c). The metabolism group (49.87–52.76%) was the largest of the three groups during co-composting. We noticed that the proportion of the lipid and carbohydrate metabolism group generally increased from day 0 to day 60, which may be related to the high biodegradability of those substrates. However, amino acid metabolism experienced a significant decline during the first week of composting and then increased in the cooling phase and mature phase. Amino acid metabolism is known to enhance microbial growth and activity since amino acids can serve as sources of both carbon and energy for microbes during composting (Bello et al., 2020). Therefore, the energy production and conversion subsystem were further analysed by PICRUSt based on the KEGG database (Fig. S3A). The oxidative phosphorylation metabolism and carbon fixation pathway were two main categories during the composting of CM + B. The relative abundances of genes involved in carbon fixation exhibited an increase during the thermophilic phase. As for nitrogen metabolism, the functional enzymes that showed significant change were associated with nitrification, denitrification and nitrogen fixation (Fig. S3B). The relative abundance of nitrogenase (nif) increased during composting and became the dominant functional enzyme in the thermophilic phase.
3.5. Fungal community succession and predicted fungal functions
For fungal community, six main phyla were detected during co-composting (Fig. 5a). The similarities of fungal community between CM + B and CM + B + N groups indicated that the ratio of cow manure and bedding material possibly have a minor impact at least at phylum level. During the composting of CM + B, Ascomycota made up the greatest proportion of the classified OTUs, and its relative abundance increased from initial value of 47.04–88.05% on day 40 and 90.07% on day 60. At the genus level, the relative abundances of the 20 most abundant classified fungal genera showed obvious variation over the 60-day composting period (Fig. 5b). Orpinomyces, which was detected only in the original cow manure, includes anaerobic fungi that inhabit the gastrointestinal tract of mammalian herbivores and cannot survive in the aerobic compost matrix (Zavrel et al., 2013). The thermophilic genus Mycothermus was dominant during the composting process, especially on day 40 (97.32%) and day 60 (98.77%). The microbial consortium consisting of the thermophilic fungus Mycothermus thermophilus (Scytalidium thermophilum) and a range of thermophilic Proteobacteria and Actinobacteria was primarily responsible to biodegradation of feedstock and release of ammonia (Kertesz & Thai, 2018a). Candida and Aspergillus were detected in all samples, indicating their high adaptability toward diverse environmental conditions (e.g. temperature, moisture and pH). Both of their relative abundance reached the highest value during thermophilic phase, consistent with their thermophilic features. Under relatively high temperature, they were able to degrade various components of lignocelluloses, promote the formation of precursor substances and thus accelerate the synthesis of humic substances (Huang et al., 2019). The fungal community structure of other composting groups was showed in Fig. S2D-F. The fungi in the co-composting of CM + B system, including the 35 most abundant OTUs, were classified by ecological guild and by trophic mode (Fig. 5c). A neighbor-joining phylogenetic tree was constructed to demonstrate the phylogenetic relationships among the main fungal communities in the CM + B co-composting system which showed that pathotrophs and saprotrophs were the dominant fungal trophic modes in this co-composting system and saprotrophic fungi were the most commonly detected taxa, with 13 OTUs belonging to 11 known genera of Ascomycota (e.g., Aspergillus and Candia) and Basidiomycota (e.g., Wallemia). The wood saprotrophs and many undefined saprotrophs were the groups that showed the most significant changes during the co-composting. Notably, the relative abundances of the OTUs of wood saprotrophs started at only 3.5% on day 0 but increased quickly, and they became the dominant fungal group (89.1%) in the mature phase. These data confirmed the efficient conversion of cellulose and hemicellulose during the co-composting of CM + B. Animal and plant pathogens were not detected after 60 days of co-composting in this experiment, indicating that the CM + B co-composting of CM + B improved the safety of raw materials as an agricultural fertilizer. High temperature was reported to be the most important factor in the elimination of pathogens during aerobic composting (Duan et al., 2019). Compared to bacteria, fungi was more advantageous on lignocellulose degradation because of their mycelial structure (Wang et al., 2018c). Ascomycota and Basidiomycota are the dominant fungal phyla for lignocellulose degradation, and high abundances of these phyla promote the degradation of organic waste during composting (Jiang et al., 2019). Mycothermus, Penicillium, and Aspergillus were the core functional genera to produce lignocellulose-degrading enzymes during composting (Kertesz & Thai, 2018b). Specifically, Mycothermus has been reported to produce thermostable cellulases, hemicellulases and xylanase (Basotra et al., 2016; EFSA Panel on Food Contact Materials et al., 2019; Ma et al., 2017), and act as a key organism in the decomposition of plant materials and plant-derived compounds (Cannon & Kirk, 2007). The thermophilic microbe Aspergillus has a high capacity to degrade lignocelluloses and lignin (Huang et al., 2019), and can use malodorous sulfur compounds to reduce odour pollution in composting systems (Liu et al., 2013).
3.6. Relationships of bacterial/fungal communities to physicochemical characteristics and microbial metabolism
CCA, RDA and heatmaps were used to describe the relationship between microbial community during CM + B composting process and environmental factors, including temperature, moisture, pH, and C/N (Fig. 6 &S4). All the p_value data was showed in tables in supplementary file (Table. S1-4). CCA showed the dynamics of the genera data for the top 20 bacterial genera in the different phases of co-composting process (Fig. 6a). Romboutsia, Paeniclostridium and Symbiobacterium were observed mainly on days 7–20 and were positively correlated with temperature (Fig. S4A). Entering thermophilic phase, Ruminiclostridium turned to be the dominant genus and also had a significantly positive relationship with temperature (p_value < 0.05). Bacteroides, Solibacillus and bacillus showed significantly positive relationships with moisture (p_value < 0.05). In contrast, Lysinibacillus, Solibacillus, and Thermobacillus showed negative correlations with moisture and C/N. Ruminiclostridium and Symbiobacterium had a significantly positive relationship with pH (p_value < 0.05). In addition, the five most abundant genera, including Acinetobacter and Bacillus, were all positively correlated with pH at 40 and 60 days. Interestingly, the RDA revealed that only four fungal genera showed positive correlations with temperature not significantly (Fig. 6b and Fig. S4B), including Candida which was dominant on day 20, and Pyrenochaetopsis which was detected only on day 20 (Fig. S4B). Both genera Penicillium and Mortierella detected mainly in the early phase of co-composting, were significantly related to moisture (p_value < 0.05). Unclassified_f__Neocallimastigaceae and unclassified_p__Ascomycota were significantly related to the C/N. In addition, Aspergillus, and Ascochyta were negatively correlated with pH and temperature.
These results indicated that temperature had less influence on the succession of fungal communities than on the succession of bacterial communities. CCA and RDA revealed the dominant bacterial/fungal genera in the original cow manure co-compost mixture was affected mainly by moisture and C/N. Thermophilic genera became dominant producers with increasing temperature. PH was a key factor affecting the bacterial community structure during the curing and mature phases of cow manure co-composting.