Composting procedure and sampling
The HPC and TC were constructed according to Huang et al. (2019a). Briefly, each composting pile was conducted as a mixture of 200 kg of pig manure and rice straw with an initial C/N ratio of 25 and moisture content of 60%. For HPC, the mixture was stirred inside the hyperthermophilic pre-treatment reactor (400 L) and the temperature of the mixed material in the reactor was controlled artificially via an oil bath heated up to 90°C and sustained for 4 h. Then, the mixture was transferred to the TC reactor after the temperature decreased to the ambient temp. For TC, each reactor had a volume of 374 L with a height of 92 cm and an inner diameter of 72 cm. Triplicate samples from each reactor were gathered on day 0, 20, 40 and 60 and used for biophysicochemical analyses and DNA extraction.
Ammonia emission was sampled daily for the first two weeks and once two days for the remaining weeks. An air blower was blown for 1 min to remove some accumulated gas from the headspace before gas sampling. Then a 10 cm plastic tube of 2.5 mm diameter was fixed to the lid with screws and a rubber stopper to sample the gas from the headspace. Gas of 50 ml was sampled at 0, 30 and 60 min after closure.
Analyses of compost
The detection of the composting temperature, pH, moisture, electrical conductivity (EC), total nitrogen (TN), total carbon (TC), inorganic nitrogen (NH4+-N, NO3−-N), protease and ammonification rate were stated previously (Huang et al. 2019a). Therefore, no more repetitive description is made herein. N2O was determined using an Agilent 7890A gas chromatograph (Agilent, Beijing, China) (Huang et al. 2021). The concentration of NH3 was detecd by absorbing NH3 with 2% boric acid and applying the titrimetric method based on Cui et al. (2019). The detailed detection of water soluble organic nitrogen (WSON), the water-soluble nitrogen (WSN) and the calculation of the WSON were stated in the supplemental information.
Analysis of proteolytic bacterial communities
DNA was extracted using the FastDNA SPIN Kit for Soil (MP Biomedicals, Solon, OH, USA) according to the manufacturer’s instruction. For Illumina amplicon sequencing, primers targeting the peptidase genes (npr (npr I/npr II) and sub (sub I/sub II) genes) (Bach et al. 2001) in proteolytic bacteria were modified by adding Illumina adapter consensus sequences to the 5’-ends of the forward and reverse primers to generate amplicons. The detailed PCR reactions, purification of amplicons, multiplexing, library pooling and sequencing process have previously been described (Huang et al. 2018) and are shown in the supplemental information. Sequences were deposited in the NCBI database with the accession numbers SRR13638639-SRR13638685, SRR13674937, SRR13675020 for proteolytic bacterial npr and sub genes.
Statistical analyses
One-way analysis of variance (ANOVA) followed by Newman-Keuls multiple comparison test was performed to evaluate significant differences (p < 0.05) between samples in the physicochemical properties using SPSS 20.0 (SPSS Inc., Chicago, III,U.S.A.) and Origin Pro8G (OriginLab corporation, Northampton, USA). To explore the relationships between proteolytic bacterial community structure and environment parameters, Redundancy analysis (RDA) was performed using the Canoco5 program (Ter Braak et al. 2002). Partial least squares path modeling (PLS-PM) was used to evaluate the complex relationships among diverse variables on NH3 emissions using the R package plspm (v 0.4.7). The biophysicochemical properties (pH, EC, DOC, TOC, TN, NH4+-N, NO3−-N, WSON, WSIN, Protease activity), composting temperature, abundance of dominant proteolytic bacteria (relative abundances of Bacillus megaterium, Staphylococcus cohnii, Bacillus subtilis and Novibacillus thermophilus) and proteolytic bacterial community composition (based on OTU abundances) were included in the model.