Clean recovery and recycling of seasonal surplus forage grass by microbial driven anaerobic fermentation: a case study of napiergrass
Chemical and Biological Technologies in Agriculture volume 9, Article number: 91 (2022)
In this study, the anaerobic fermentation technique was conducted to accomplish the clean recycling of surplus napiergrass. The surplus napiergrass harvested at two harvest dates (early vegetative stage, NI; late vegetative stage, NII) was treated as follows: (i) natural fermentation of NI (NNI); (ii) natural fermentation of NII (NNII) and stored for 1, 3, 7, 15, 30 and 60 days. After 60 days of anaerobic fermentation, NNI had higher lactic acid concentration and ratio of lactic to acetic acid, but lower pH value and ammonia–nitrogen concentration than NNII. Lactobacillus and Enterobacter were, respectively, dominant in both 7-day NNI and NNII, while Lactobacillus was the most abundant genus in 30-day NNI and NNII. Both harvest date and store time altered the bacterial co-occurrence networks of fresh and fermented napiergrass. The complexity of the bacterial networks decreased from NII, NI, NNII to NNI. The correlations were primarily positive in the bacterial networks of NI, NII, NNII-7 and NNII-30 with positive correlative proportion of 53.0%, 64.3%, 53.1% and 55.6%, but negative in those of NNI-7 (47.4%) and NNI-30 (46.2%) with positive correlative proportion of 47.4% and 46.2%, respectively. Overall, the fermentation quality and microbial community structure of napiergrass during anaerobic fermentation were highly influenced by harvest date and store time. Based on the principle of stable fermentation and high quality, anaerobic fermentation of NI for at least 15 days is recommended. The in-depth understanding of microbial community dynamics and co-occurrence networks during anaerobic fermentation of napiergrass is important for revealing the fermentation mechanism and can contribute to resource recycling without increasing cost.
Feed production is closely related to the environment in terms of water consumption, land use and climate change, since it is an essential industry that required resources, such as water, land, and energy. Given the increasing global population growth and demand for the animal product, these relationships might be more tightly in the future. As a key link of the feed industry, forage grass production presents a distinct seasonal characteristic throughout the year . Feedback from forage growers indicated that in the fast-growing season (May–October), the yield of forage grass increases wildly, accounting for more than 70% of the annual growth, especially in July and August, accounting for more than 40% of the annual growth. It frequently happens that the yield of forage grass exceeds the need for utilization . It’s common to leave surplus forage grass in the fields or discard directly without utilization. In resource waste assessment, the waste of forage grass is often not as prominent as the waste of food, and only a few studies have emphasized forage waste. Based on the concept of resource utilization, long-term preservation of surplus forage grass by anaerobic fermentation may increase its add-value and transform the original linear economy into a circular economy .
Anaerobic fermentation of forages (ensilage) is based on the principle that lactic acid bacteria (LAB) ferment available sugars in plants to produce lactic acid-based organic acids, and rapidly reduce pH value so as to effectively inhibit the activities of harmful microorganisms to achieve the purpose of long-term preservation . However, it is worth noting that the fermentation quality varies greatly in the anaerobic fermentation of forage at different harvest dates. Oliveira et al.  and Toruk et al.  have shown that the harvest date of forage is the main factor determining the nutritional value and fermentation quality. In the case of guinea grass, van Niekerk et al.  reported that the anaerobic fermentation of guinea grass prepared at early vegetative and boot stages was lactate-type, while the anaerobic fermentation of guinea grass prepared at the full bloom stage was acetate-type. Similarly, there is also a need to determine the appropriate harvest date of surplus napiergrass for anaerobic fermentation. However, to the authors’ knowledge, no studies regarding the effects of different harvest dates on the fermentation quality of surplus forage grass were reported.
Along with the study of anaerobic fermentation, culture-based methods are no longer sufficient to clearly present the microbial community succession from fresh forages to fermentative products. Recent advances in culture-independent analyses, such as high-throughput sequencing technology, have enabled researchers to explore the microbial population shifts involved in the growing and anaerobic fermentation process, and to mine more crucial biological information . Therefore, in this study, surplus napiergrass (Pennisetum purpureum Schumach.) was used to evaluate the effects of harvest date and store time on its fermentation quality, microbial community and co-occurrence networks during anaerobic fermentation. The obtained results may, therefore, provide a basic reference for contributing to resource recycling without increasing cost and transferring a single economy to a circular economy.
Surplus forage grass collection
Napiergrass was surplus and left in the field of Baima National Agricultural-tech Zone (Jiangsu, China) without harvest. A 30 m2 field was separated into three equal blocks (for replicates) and the obtained each block was further divided into two equal plots (for two harvest dates). After 12 weeks of planting, half of the napiergrass was harvested on August 15, 2018, and after 18 weeks of planting, the remaining napiergrass was harvested on September 26, 2018, to obtain two batches of napiergrass (NI, the early vegetative stage; NII, the late vegetative). The harvest time was in the morning with clear weather, and the stubble height was about 15 cm. Each batch of fresh napiergrass was immediately cut into about 2 cm lengths by a forage cutter, mixed thoroughly and split into two parts for fresh sample analysis and anaerobic fermentation preparation, respectively.
Anaerobic fermentation preparation
A total of 36 bags (2 harvest dates × 6 store time × 3 replicates per treatment) were prepared and the treatments were set as follows: (i) natural fermentation of NI (NNI) and (ii) natural fermentation of NII (NNII). Specifically, approximately 0.45 kg of thorough-mixing material was packed into a UV-sterilized polythene bag (size: 300 × 400 mm), sealed by an automatic vacuum sealer and stored under surrounding temperature (25–30 ℃) for 1, 3, 7, 15, 30 and 60 days of anaerobic fermentation.
Bio-chemical composition analyses
Before analyses, the fresh or fermented sample was blended thoroughly. About 300 g sample was dried at 65 °C for 48 h in an air-forced oven to determine dry matter (DM) content. The oven-dried sample was then milled to pass through a 1-mm sieve. The water-soluble carbohydrates (WSC) content of fresh and fermented samples was analyzed with anthrone-sulfuric acid . The buffering capacity (BC) of the fresh sample was quantified by titration . The neutral and acid detergent fiber (NDF and ADF) content of the fresh sample was quantified by the method of Van Soest et al. . The total nitrogen (TN) content of the fresh sample was quantified by a Kjeltec 8200 Kjeldahl N analyzer (Foss Analytical AB, Höganäs, Sweden). The crude protein (CP) content of fresh and fermented samples was obtained through multiplying TN by 6.25.
After extraction of the fresh or fermented sample with deionized water (1:3 ratio) at 4 °C for 24 h, the above extracts were filtered with 4 layers of sterile cheesecloth and filter paper. The pH of fresh or fermented samples was immediately recorded by a glass electrode pH meter. The ammonia–nitrogen (NH3–N) concentration of the fermented sample was quantified by the phenol–hypochlorite procedure . The organic acid concentrations of fermented sample including lactic acid (LA), acetic acid (AA), propionic acid (PA) and butyric acid (BA) were quantified by the 1260 Infinity HPLC system (Agilent Technology Co., Ltd., Waldbronn, Germany) .
After homogenization of fresh or fermented sample with sterile saline solution (1:9 ratio) at 120 rpm, 37 °C for 2 h, 1 mL homogenized solution was serial-diluted for the enumeration of LAB, aerobic bacteria, yeasts, molds and enterobacteria . The microbial number was recorded in colony-forming units (CFU), transformed to logarithmic form and expressed on a fresh material (FM) basis. After filtrating with 2 layers of sterile cheesecloth, the obtained filtrate was collected for subsequent bacterial DNA extraction.
High-throughput sequencing analysis
Bacterial DNA extraction, PCR amplification and 16 s rRNA paired-end sequencing were conducted as reported in our previous study . Briefly, after quality filtering and chimeric sequences removal, the qualified reads were obtained and further clustered into operational taxonomic units (OTUs). The OTUs were analyzed at phylum and genus levels based on the SILVA database. Bacterial alpha diversities (Shannon, Chao1, Ace, Sobs, Simpson and Coverage indexes) and beta diversity (Bray–Curtis distance metric) were calculated by the QIIME software. Through R software (ver. 4.1.3), the vegan package was run to construct principal coordinates analysis (PCoA) plots for beta diversity analysis, the ggplot2 package was to construct stream graphs showing the bacterial community successions, and the pheatmap package was to construct heatmaps visualizing the Spearman’s correlation relationships between fermentation products and bacterial communities.
Co-occurrence network analysis
The co-occurrence pattern was constructed by calculating multiple abundance correlations based on a genus-level matrix using Networkx (ver. 2.6.3). Only genera of relative abundance > 0.05% were considered. If Spearman correlation coefficient (ρ) > 0.50 and p < 0.05, co-occurrence is considered to be robust. The co-occurrence networks were visualized using Gephi (ver. 0.9.2). Nodes represent individual bacterial genera, and edges represent the pairwise correlation between nodes in the bacterial network. The calculated topological characteristics of bacterial networks include positive (co-occurrence) and negative (mutually exclusive) correlation numbers, network diameter, average shortest path length, average clustering coefficient, average connectivity (degree), closeness centrality, betweenness centrality, modularity, etc.
The effects of harvest date, store time and their interactions on chemical composition, fermentation quality and microbial population were investigated using the GLM of SAS (ver. 9.2; SAS Institute Inc., NC, USA) following the model as follows:
where Yij refers to the dependent variable; μ refers to the overall mean; Gi refers to the effect of harvest date (i = 2, NI vs. NII); Dj refers to the effect of store time (j = 6, 1, 3, 7, 15, 30 and 60); (G × D)ij refers to the interaction effects of harvest date and store time; and eijk refers to the residual error. Comparisons between two harvest date were performed through t test when the fixed effect of harvest date was significant. The differences were considered statistically significant at p < 0.05.
Characteristics of surplus napiergrass before anaerobic fermentation
The harvest date significantly (p < 0.05) affected all measured chemical and microbial parameters except the pH value and mold count of napiergrass (Table 1). The WSC and CP content, the BC and the LAB and enterobacteria number decreased (p < 0.05), while the DM, NDF and ADF contents and the aerobic bacteria and yeasts numbers increased (p < 0.05) as harvest date was delayed.
Fermentation quality of surplus napiergrass after anaerobic fermentation
Harvest date or store time had significant (p < 0.05) effects on the pH, the LA, AA and BA concentrations and the ratio of lactic to acetic acid (LA/AA), while their interaction significantly (p < 0.05) affected the LA concentration and LA/AA (Table 2). The pH of NNI and NNII sharply (p < 0.05) declined during the first 7 days of anaerobic fermentation reaching the lowest value (3.66 and 4.00) on day 15 and day 30 of anaerobic fermentation, respectively, then increased slightly. During the whole anaerobic fermentation, NNI always had a lower pH value than NNII (p < 0.05). The LA concentration presented the opposite trend to the variation of pH value, with the highest concentration of 52.0 and 39.2 g/kg DM on day 30 of anaerobic fermentation in NNI and NNII, respectively. Regardless of harvest date, the AA concentration increased with the store time prolonged. Along with the anaerobic fermentation process, the LA/AA value showed an upward and then downward tendency, with a maximum of 9.58 in NNI on day 15 of anaerobic fermentation and a maximum of 4.89 in NNII on day 30 of anaerobic fermentation. The BA of NNI and NNII was always less than 2 g/kg DM.
The DM content was affected by harvest date, the WSC content was affected by store time, and the NH3–N concentration was affected by their interaction (p < 0.001). As anaerobic fermentation proceeded, the DM content remained relatively stable, but the WSC content decreased and the NH3–N concentration increased (p < 0.05). NNII always had higher DM content and NH3–N concentration than NNI (p < 0.05).
Harvest date and store time significantly (p < 0.001) affected LAB, aerobic bacteria, yeasts, molds and enterobacteria numbers, while their interaction significantly (p < 0.05) affected LAB and aerobic bacteria numbers (Table 3). The LAB number of NNI and NNII showed an upward and then downward tendency, but the aerobic bacteria, yeast, molds and enterobacteria numbers constantly decreased to a low or undetected level. The LAB number of NNI and NNII increased rapidly during the first 3 days of anaerobic fermentation and detected the maximum (8.66 and 7.39 log10 CFU/g FM) on day 15 of anaerobic fermentation. Wherein, NNI had more LAB than NNII (p < 0.05). Although the number of aerobic bacteria, yeasts and enterobacteria in NNI was significantly (p < 0.05) lower than that in NN1I within the first 15 days of anaerobic fermentation, there was no difference in the number of these microorganisms between NNI and NNII at the end of anaerobic fermentation (60 d).
Bacterial community of surplus napiergrass before and after anaerobic fermentation
The alpha diversities of the microbial community in fresh and fermented napiergrass are presented in Fig. 1A. The Shannon, Chao1 and Sobs indexes were highest in fresh NII, followed by fresh NI and finally fermented samples. Among all samples, NNI always had numerically (p > 0.05) or statistically (p < 0.05) lower Shannon, Chao1, Ace and Sobs indexes than NNII. Compared with fresh samples (NI and NII), the Shannon, Chao1, Ace and Sobs indexes of fermented samples (NNI and NNII) decreased after 7 days of anaerobic fermentation, and further decreased in NNI but increased in NNII after 30 days of anaerobic fermentation. The lowest Shannon, Chao1, Ace and Sobs indexes were observed in NNI on day 30 of anaerobic fermentation. The average Coverage index of all sequenced samples was greater than 0.99. The beta diversities of the microbial community in fresh and fermented napiergrass were assessed by PCoA. A well separation was found between the symbols of fresh and fermented sample in the 3D-PCoA plot except for NNII-7 and NNII-30 (Fig. 1B). Among them, NI and NII, NNI-7 and NNII-7 as well as NNI-30 and NNII-30 were well-separated in Fig. 1C–E. In addition, NI, NNI-7 and NNI-30 as well as NII and NNII (NNII-7 and NNII-30) were well-separated in Fig. 1F and G.
As shown in Fig. 2A, Proteobacteria, Firmicutes, Actinobacteriota and Bacteroidota were detected in both NI and NII and Deinococcota was an additional phylum in NII. Proteobacteria was the abundant phyla in NI and NII, with a relative abundance of 64.3% and 69.2%, respectively. With the growth of napiergrass, the relative abundances of Actinobacteriota and Bacteroidota increased from 5.55% and 0.21% to 20.0% and 6.86%, but Proteobacteria and Firmicutes decreased from 74.3% and 19.9% to 59.2% and 10.8%, respectively. After 7 days of anaerobic fermentation, the relative abundance of Firmicutes in NNI and NNII increased in varying degrees, accompanied by a decrease in the relative abundance of Proteobacteria, Actinobacteriota and Bacteroidota. After 30 days of anaerobic fermentation, Firmicutes (> 60%) dominated in the microbiota of all samples, especially in that of NNI.
The number of genera with a relative abundance greater than 1% in NI and NII was 10 and 24, respectively (Fig. 2B). The most abundant genus in NI was Acinetobacter (26.5%), followed by Enterobacter (15.0%) and Pseudomonas (12.4%), while Pseudomonas (17.6%), Enterobacter (9.14%), Sphingomonas (7.87%), Aureimonas (6.31%) and Rhizobium (6.01%) were the 5 genera with high relative abundance in NII. From NI to NII, the relative abundance of Pseudomonas increased from 12.4% to 17.6%, whereas Acinetobacter, Enterobacter and Lactococcus decreased from 26.5%, 15.0% and 8.01% to 2.36%, 9.14% and 2.65%, respectively. With the process of anaerobic fermentation, Acinetobacter and Sphingomonas decreased to an undetectable level. Differently, after 7 days of anaerobic fermentation, the relative abundance of Pediococcus and Lactococcus in NNI increased up to 10.1% and 22.5%, respectively, and the relative abundance of Klebsiella in NNII increased up to 12.9%. After 30 days of anaerobic fermentation, Lactobacillus dominated the bacterial community of both NNI and NNII, with relative abundance accounting for 63.1% and 34.1%, respectively.
The stream graphs showed that both harvest date and storage time had a remarkable impact on the succession of bacterial communities during the anaerobic fermentation of surplus napiergrass (Fig. 2C–H). Although the variation of harvest date affected the bacterial community succession of surplus napiergrass (Fig. 2C, F), the anaerobic fermentation process had a more significant effect on the bacterial community succession (Fig. 2D, E, G, H).
Bacterial co-occurrence networks of surplus napiergrass before and after anaerobic fermentation
The co-occurrence network based on the correlation coefficient matrix, to a certain extent, can reflect the relationships between microbial members. Thus, the bacterial co-occurrence networks of fresh and fermented napiergrass based on Spearman’s rank correlation were separately created at two harvest dates to clearly understand the effects of harvest date on the interrelationships of bacterial members (genera). Based on co-occurrence network analysis (Fig. 3A–D and Table 4), the number of nodes and edges in bacterial networks was ranked as follows: NII > NI > NNII > NNI. The genera with high closeness centrality, high mean degree and low betweenness centrality were Lactococcus, Hafnia-Obesumbacterium, Enterococcus and Curtobacterium in NI, Microbacterium and Roseomonas in NII, Lactobacillus and Pediococcus in NNI-7, Weissella, Lactobacillus, Enterococcus and Pantoea in NNII-7, Lactobacillus and Lactococcus in NNI-30, and Lactobacillus, Pedicoccus and Kosakonia in NNII-30 (Fig. 3E and F). Furthermore, the correlations of nodes were primarily positive (proportion of positive edges) in the bacterial networks of NI, NII, NNII-7 and NNII-30 but negative (proportion of negative edges) in those of NNI-7 and NNI-30 (Fig. 3A–F).
Correlation analysis of chemical composition and epiphytic microbiota as well as fermentation products and bacterial communities
Before anaerobic fermentation (Fig. 4A), Acinetobacter was negatively (p < 0.01) related to DM content, with correlation coefficients of -0.943. Paenibacillus was positively correlated with CP content (R = 0.880, p < 0.05) and WSC (R = 0.941, p < 0.01). Positive correlations were observed between CP content and Rhizobium (R = 0.957, p < 0.01). After anaerobic fermentation (Fig. 4B), Lactobacillus was positively correlated with LA (R = 0.580, p < 0.05) and LA/AA (R = 0.776, p < 0.01) concentrations, but negatively correlated with DM content (R = −0.692, p < 0.05) and pH value (R = −0.873, p < 0.001). Similarly, Leuconostoc was positively correlated with LA/AA (R = 0.713, p < 0.01), but negatively correlated with DM content (R = −0.860, p < 0.001). Enterobacter, Klebsiella and Enterococcus were positively (p < 0.01) correlated with DM content, with correlation coefficients of 0.783, 0.755 and 0.715, whereas negatively (p < 0.05) correlated with LA/AA, with correlation coefficients of -0.853, −0.790 and −0.609, respectively. There was positive correlation between Pseudomonas and LA concentration (R = 0.630, p < 0.05) and negative correlation between Lactococcus and NH3–N concentration (R = −0.566, p < 0.05).
Effects of harvest date on the characteristics of surplus napiergrass
The harvest date is reported to be the most important factor affecting forage quality at harvest . In this work, with the delay of harvest date, the CP content decreased, while the DM, NDF and ADF content increased, which was in line with the study of Silva et al. . The increase in DM content could be explained by the deposition of cell walls (structural carbohydrates) produced by photosynthesis. While the increase in NDF and ADF content could be attributed to the decrease in leaf–stem ratio, since the cell wall components in the stem are higher than that in the leaf [18, 19]. Meanwhile, the increased proportion of NDF and ADF led to a relative decrease in CP and WSC content [18, 20]. Queiroz et al.  indicated that the decrease of CP and WSC content during forage growth might be the result of the ‘dilution effect’ caused by the increasing proportion of cell walls. Moreover, these variations could also be attributed to the plant varieties, geographical location, climate, harvest season and fertilization. The CP content of forage grass is known to influence its BC , and the decrease in CP content may explain the decline in BC.
As Macarisin et al.  reported, the bacterial number and diversity can be sharply impacted by the development stages of leaves and plants. Throughout the growth cycle of forage grass, the external environment (e.g., solar radiation, temperature and rainfall) and internal environment (e.g., plant morphology, moisture content and leaf thickness) have been changing, and they are reported to impact microbial colonization . In this work, the number of LAB and enterobacteria decreased with the increase of maturity, which may be caused by the decrease of water and WSC content on the napiergrass surface at the late harvest date. Recent studies showed that sugar and volatile organic compounds secreted by forage play an important role in determining the microbial population of forage grass [25, 26]. It is known that microbes including LAB are enriched on sugar-rich plants. Thus, the lower WSC content in NII could explain its decreased number of epiphytic LAB. Moreover, the nutrient release from aging tissue and leaves in NII was considered to be beneficial to microbial growth , which could explain the higher aerobic bacteria and yeast numbers in NII.
Effects of harvest date and store time on the fermentation quality of surplus napiergrass after anaerobic fermentation
During the anaerobic fermentation of surplus napiergrass, the pH increased with the advancement of harvest date, which was in line with the result of Faria et al. . Similarly, Abdelhadi and Tricarico  also found that the silage pH increased as the stage of forage maturity increased. The above results were considered to be related to the decrease of WSC and water content during the growth of forage. According to McDonald et al. , as forage matured, the decrease of water activity (aw) and WSC content in fresh material lowed the production of LA and other organic acids during anaerobic fermentation, thereby increasing the pH value. Hence, a higher pH value and lower LA concentration were observed in NNI than that in NNII. The LA/AA of NNI and NNII were always greater than 2 throughout the anaerobic fermentation, indicating that all anaerobic fermentation of napiergrass presented lactic acid-type (homolactic) fermentation. As in most studies, AA concentration increased with the extension of store time, and this could be related to the activities of AA-producing microorganisms. The BA produced by clostridia is easily found in the anaerobic fermentation of material with DM less than 30% . However, negligible BA (< 2 g/kg DM) was detected in NNI and NNII, indicating that no serious clostridial fermentation occurred in this study.
The significantly higher DM content of NNII than that of NNI could ascribe to the higher DM content in NII than that in NI. Based on the fermentation parameters, such as pH, LA/AA and BA, NNII seems to be effectively preserved. However, the high NH3–N concentration (150 g/kg TN) detected in NNII suggested that the anaerobic fermentation of NNII had severe protein degradation and nutrient loss. As an indicator of protein degradation, NH3–N concentration has long been used to evaluate the fermentation quality of anaerobic fermentation. Although the appropriate DM content and low BC of NII could benefit its subsequent LA fermentation, the insufficient WSC (< 50 g/kg DM) content and LAB number (< 5.0 log10 CFU/g FM) finally determined the worse fermentation quality of NNII. The degradation degree of forage protein depends on the decline rate of pH during anaerobic fermentation . The rapid pH decline of NNI in the initial phase of anaerobic fermentation effectively inhibited the degradation and deamination of protein, resulting in the acceptable NH3–N concentration (< 100 g/kg TN) of NNI. Furthermore, the high proteolysis degree in aging and dead tissues of mature forage (NII) might also contribute to the high NH3–N concentration in NNII .
The higher number of LAB in NNI was related to the higher WSC content and epiphytic LAB of NI. As O2 depleted during anaerobic fermentation, the number of aerobic bacteria and molds in both harvest dates rapidly decreased to an undetectable level. In addition, the negligible yeasts and enterobacteria in NNI could be explained by its low pH value and high LA concentration and LAB number. Furthermore, the decline of yeasts in NNI was also associated with the massive proliferation of LAB, which reduced the niche available for yeasts.
Effects of harvest date and store time on the bacterial community of surplus napiergrass before and after anaerobic fermentation
The Coverage index of all sequenced samples was above 99%, indicating that the sequencing depth was sufficient for reliable analysis of microbial community. Bacterial alpha diversity of surplus napiergrass, characterized by Shannon, Chao1, Ace, Sobs and Simpson indexes, increased with the delay of harvest date. The anaerobic fermentation process further decreased its bacterial alpha diversities, which could be due to the deactivation of acid- and anaerobic-intolerant epiphytic bacteria . Mendez-Garcia et al.  found that low pH is the main reason for the decrease of microbial diversity in acidic environments. The adequate WSC content (> 50 g/kg DM) and LAB number (> 5.0 log10 CFU/g FM) of NI ensured the LAB proliferation and rapid acidification during anaerobic fermentation, thereby reducing the bacterial alpha diversity of NNI.
The PCoA plots were plotted to visualize the differences in bacterial community composition between treatments as distances between symbols. The well separation of symbols NI and NII showed great differences in the composition of bacterial community for surplus napiergrass harvested at these two harvest dates, and this discrepancy could be associated with climate, the physio-biochemical characteristics of forage grass, or other factors . Meanwhile, the separated clustering between the fresh and fermented samples was, as abovementioned, due to the deactivation of acid- and anaerobic-intolerant epiphytic bacteria during anaerobic fermentation. In addition, the separation among NI, NNI-7 and NNI-30 suggested that the composition of bacterial community in NNI was distinctly different at different store times. The decrease of the distance from the symbols of NII and NII-7 to NII-7 and NII-30 showed that, as ensiling proceeded, the bacterial community composition of NNII tended to be similar.
This obvious succession of bacterial community from Proteobacteria to Firmicutes before and after anaerobic fermentation can ascribe to the suppression of aerobic genera (Sphingomonas, Acinetobacter, etc.) and the bloom of LAB (mainly Lactobacillus, Pediococcus, Weissella, and Lactococcus). Anaerobic conditions favor the growth of Firmicutes, because this genus is common in anaerobic fermentation . Pediococcus, Weissella and Lactococcus are generally considered early colonizers during anaerobic fermentation [34, 35] due to their weaker tolerance to acid compared with Lactobacillus. However, the initial acid environment established by these genera is particularly suitable for the proliferation of Lactobacillus , which explained the overwhelming dominance of Lactobacillus in both NNI and NNII after 30 days of anaerobic fermentation. As expected, the relative abundance of Enterobacter in NNI decreased continuously during anaerobic fermentation, while it is worth noting that the relative abundance of Enterobacter in NNII increased first and then decreased, and this could be related to the high pH value of 7-day NNII. Enterobacter are commonly considered undesirable microbes because this genus can metabolize LA and WSC into AA and amino acids into ammonia [37, 38]. Although studies have shown that under anaerobic fermentation of material with 25% DM, the pH value below or equal to 4.35 can effectively inhibit the activity of Enterobacter [39, 40], a certain abundance of Enterobacter was still detected in NNI even its pH value was low. This finding might be attributed to the existence of several acid-resistant strains of Enterobacter. This genus has been reported to protect itself and grow well in some adverse environments, i.e., low pH conditions .
Different from Lactobacillus and Enterobacter, Pseudomonas has not been widely studied in anaerobic fermentation. As aerobic and non-fermentative bacteria, Pseudomonas is generally thought to be difficult to survive under anaerobic and acidic conditions. However, after an initial inhibition, a high relative abundance of Pseudomonas was observed in 30-day NNI and NNII. These results are unexpected but consistent with the findings of Dong et al. . Previous studies have reported that, under certain conditions, Pseudomonas can grow anaerobically using acetate and nitrate as electron acceptors . Overall, according to the stream graphs, store time seems to have a greater influence on the bacterial community compositions of napiergrass than harvest date.
Effects of harvest date and store time on the bacterial co-occurrence networks of surplus napiergrass before and after anaerobic fermentation
Co-occurrence networks among bacterial genera for fresh and fermented napiergrass were constructed to clearly understand the effects of harvest date on the correlation and interaction of the epiphytic and anaerobic fermentative microbiome. Based on the number of nodes and edges, the complexity of bacterial networks in fresh napiergrass increased from the early vegetative stage to the late vegetative stage (Table 4, Fig. 3A, andB). Differently, anaerobic fermentation decreased the complexity of bacterial networks, with the simplest bacterial correlation structures in 30-day NNI (Table 4, Fig. 3C–F). High fermentation quality was accompanied by low network complexity, which is consistent with Bai et al. .
As reported by Banerjee et al. , the negative correlation of co-occurrence networks indicates a possible competition for resources and common predators, while the positive correlation indicates symbiotic or cooperative relationships within microbial taxa. The lower proportion of negative correlations in the network of NII (35.7%) than those in the network of NI (47.0%) manifested that, with the growth of napiergrass, the competition among bacterial taxa in its epiphytic microbiota weakened. Co-occurrence network analyses have been increasingly applied to interrogate community stability based on topological properties [45, 46]. According to the findings of previous studies, microbial networks with lower positive correlations (or higher negative correlations) among members are more stable [47, 48]. Thus, the lower proportion of positive correlations in the network of NNI-7, NNII-7, NNI-30 and NNII-30 (47.4%, 53.1%, 46.2% and 55.6%) than those in the networks of NI and NII (53.0% and 64.3%) reflected that the bacterial networks of fermented samples were more stable than those of fresh samples.
According to Berry and Widder , the keystone taxa in the bacterial community can be identified by the combined scores of low betweenness centrality, high closeness centrality and high mean degree. Correspondingly, Lactococcus, Hafnia-Obesumbacterium, Enterococcus and Curtobacterium in NI, Microbacterium and Roseomonas in NII, Lactobacillus and Pediococcus in NNI-7, Weissella, Lactobacillus, Enterococcus and Pantoea in NNII-7, Lactobacillus and Lactococcus in NNI-30, and Lactobacillus, Pedicoccus and Kosakonia in NNII-30 were identified as the keystone taxa (Fig. 3G, H). It should be noted that the keystone taxa in this work were not necessarily the ones with the highest relative abundance. Similarly, previous studies found that although keystone taxa have considerable effects on bacterial communities and functions, their abundance is not proportional to their effects [43, 50, 51].
Relationships between chemical composition and epiphytic microbiota as well as fermentation products and bacterial communities
In fresh napiergrass, the Acinetobacter was negatively correlated with DM content, indicating that this genus prefers moist habitats. Indeed, Acinetobacter can be readily found in water, soil and sewage . In addition, the significant negative correlation between WSC content and Paenibacillus could be due to the fact that this genus can utilize various sugars for growth [52, 53]. While the significant positive correlation observed between CP content and Rhizobium could be explained by their nitrogen-fixing capacity. It is commonly known that rhizobia can convert inorganic nitrogen from the atmosphere into organic nitrogen for use by host plants. In the anaerobic fermentation of napiergrass, there were positive correlations between Lactobacillus and LA concentration or LA/AA but negative correlation between Lactobacillus and pH value or BA concentration, which further confirmed that this genus had strong acid resistance and played a crucial role in the decline of pH in late anaerobic fermentation . The significant negative correlation between NH3–N concentration and Lactococcus suggested that the abundant distribution of Lactococcus in the initial phase of anaerobic fermentation could reduce the production of NH3–N. The positive correlation between LA concentration and Pseudomonas was attributed to the increased relative abundance of Pseudomonas on day 30 of anaerobic fermentation. Higher relative abundances of Pseudomonas were observed in 30-day anaerobic fermentation with high LA concentrations. Both Serratia and Hafnia-Obesumbacterium belong to Enterobacteriaceae and have been reported to have proteolytic activity , which might explain the positive correlation between these two genera and NH3–N concentration in this work.
Although surplus napiergrass is considered as ‘waste’, it is still in good condition with low BC but high WSC and CP content. Harvest date and store time, especially store time had remarkable effects on fermentation quality, microbial community and co-occurrence networks during the anaerobic fermentation of napiergrass. The higher LA concentration and LA/AA in NNI were related to the higher WSC content in NI, the higher abundance of Lactococcus and Pediococcus in initial anaerobic fermentation and the higher abundance of Lactobacillus in late anaerobic fermentation. The NNI and NNII both displayed lactate-type fermentation, but the latter had an unacceptable NH3–N concentration. Therefore, anaerobic fermentation of NI for 15 days or more is recommended for the cleaner production of surplus napiergrass. Anaerobic fermentation improved the quality and add-value of surplus napiergrass in terms of fermentation characteristics, bacterial community and functional profiles, and can be a clean solution for leftover forage grass.
Availability of data and materials
All data generated or analyzed during this study are included in this published article.
- NI :
Napiergrass harvested at the early vegetative stage
- NII :
Napiergrass harvested at the late vegetative stage
- NNI :
Natural fermentation of NI
- NNII :
Natural fermentation of NII
Lactic acid bacteria
Neutral detergent fiber
Acid detergent fiber
Operational taxonomic units
Ribosomal database project
Principal coordinates analysis
- G × D:
The interaction of harvest date and store time
The ratio of lactic to acetic acid
Garcez Neto AF, dos Santos TM, da Silva J, Fernandes SR. Effect of whey permeate and Lactobacillus buchneri on biomass conservation, chemical characteristics and aerobic stability of elephant grass silage. Waste Biomass Valori. 2020;12:879–93.
Snyman LD, Joubert HW. Effect of maturity stage and method of preservation on the yield and quality of forage sorghum. Anim Feed Sci Technol. 1996;57:63–73.
de Mello Santos VH, Campos TLR, Espuny M, de Oliveira OJ. Towards a green industry through cleaner production development. Environ Sci Pollut Res Int. 2022;29:349–70.
Weinberg ZG, Muck RE. New trends and opportunities in the development and use of inoculants for silage. FEMS Microbiol Rev. 1996;19:53–68.
Oliveira SS, Costa KAdP, Souza WFd, Santos CBd, Teixeira DAA, Silva VCe. Production and quality of the silage of sorghum intercropped with Paiaguas palisadegrass in different forage systems and at different maturity stages. Anim Prod Sci. 2020;60:694–704.
Toruk F, Gonulol E, Kayisoglu B, Koc F. Effects of compaction and maturity stages on sunflower silage quality. Afr J Agr Res. 2010;5:55–9.
van Niekerk WA, Hassen A, Bechaz FM. Influence of growth stage at harvest on fermentative characteristics of Panicum maximum silage. S Afr J Anim Sci. 2010;40:334–41.
McAllister TA, Duniere L, Drouin P, Xu S, Wang Y, Munns K, Zaheer R. Silage review: Using molecular approaches to define the microbial ecology of silage. J Dairy Sci. 2018;101:4060–74.
Thomas TA. An automated procedure for the determination of soluble carbohydrates in herbage. J Sci Food Agric. 1977;28:639–42.
Playne MJ, McDonald P. The buffering constituents of herbage and of silage. J Sci Food Agric. 1966;17:264–8.
Van Soest PJ, Robertson JB, Lewis BA. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J Dairy Sci. 1991;74:3583–97.
Broderick GA, Kang JH. Automated simultaneous determination of ammonia and total amino acids in ruminal fluid and in vitro media. J Dairy Sci. 1980;63:64–75.
Zhao J, Dong Z, Li J, Chen L, Bai Y, Jia Y, Shao T. Ensiling as pretreatment of rice straw: The effect of hemicellulase and Lactobacillus plantarum on hemicellulose degradation and cellulose conversion. Bioresour Technol. 2018;266:158–65.
Zhao J, Yin XJ, Wang SR, Li JF, Shao T. Separating the effects of chemical and microbial factors on fermentation quality and bacterial community of Napier grass silage by using gamma-ray irradiation and epiphytic microbiota transplantation. Anim Feed Sci Technol. 2021;280:115082.
Zhao J, Yin X, Dong Z, Wang S, Li J, Dong D, Shao T. Using gamma-ray irradiation and epiphytic microbiota inoculation to separate the effects of chemical and microbial factors on fermentation quality and bacterial community of ensiled Pennisetum giganteum. J Appl Microbiol. 2022;132:1675–86.
Buxton DR. Quality-related characteristics of forages as influenced by plant environment and agronomic factors. Anim Feed Sci Tech. 1996;59:37–49.
Silva LFPE, Cassoli LD, Roma LC, Rodrigues ACD, Machado PF. In situ degradability of corn stover and elephant-grass harvested at four stages of maturity. Sci Agric. 2008;65:595–603.
Ma J, Sun G, Shah AM, Fan X, Li S, Yu X. Effects of different growth stages of amaranth silage on the rumen degradation of dairy cows. Animals. 2019;9:793.
Xue Y, Bai C, Sun J, Sun L, Chang S, Sun Q, Yu Z, Yin G, Zhao H, Ding H. Effects of locations and growth stages on nutritive value and silage fermentation quality of Leymus chinensisin in Eurasian steppe of northern China. Grassl Sci. 2018;64:40–50.
Onyango CM. Preharvest and postharvest factors affecting yield and nutrient contents of vegetable amaranth (var. Amaranthus hypochondriacus). Wageningen: Wageningen University; 2010. p. 129.
Queiroz FEd, Rocha VR, Monção FP, Rigueira JPS, Parrella RAdC, Rufino LDdA, Santos ASd, Cordeiro MWS. Effect of row spacing and maturity at harvest on the fermentative profile, aerobic stability, and nutritional characteristics of biomass sorghum (BRS 716) silage in the semiarid region of Brazil. Rev Bras Zootecn. 2021;50:e20200254.
McDonald P, Henderson AR, Heron SJE. The biochemistry of silage. 2nd ed. Bucks: Chalcombe Publications; 1991.
Macarisin D, Patel J, Bauchan G, Giron JA, Ravishankar S. Effect of spinach cultivar and bacterial adherence factors on survival of Escherichia coli O157:H7 on spinach leaves. J Food Prot. 2013;76:1829–37.
Kinkel LL, Wilson M, Lindow SE. Plant species and plant incubation conditions influence variability in epiphytic bacterial population size. Microb Ecol. 2000;39:1–11.
Farre-Armengol G, Filella I, Llusia J, Penuelas J. Bidirectional interaction between phyllospheric microbiotas and plant volatile emissions. Trends Plant Sci. 2016;21:854–60.
Sharma M, Sudheer S, Usmani Z, Rani R, Gupta P. Deciphering the omics of plant-microbe interaction: perspectives and new insights. Curr Genomics. 2020;21:343–62.
Thompson IP, Bailey MJ, Fenlon JS, Fermor TR, Lilley AK, Lynch JM, Mccormack PJ, Mcquilken MP, Purdy KJ, Rainey PB, Whipps JM. Quantitative and qualitative seasonal changes in the microbial community from the phyllosphere of sugar beet (Beta Vulgaris). Plant Soil. 1993;150:177–91.
Faria WG, Goncalves LC, Ribeiro GO, Carvalho WTV, Mauricio RM, Rodrigues JAS, Faria WG, Saliba EOS, Rodriguez NM, Borges ALCC. Effect of grain maturity stage on the quality of sorghum BRS-610 silages. Arq Bras Med Vet Zoo. 2011;63:1215–23.
Abdelhadi LO, Tricarico JM. Effects of stage of maturity and microbial inoculation at harvest on nutritive quality and degradability of grain sorghum whole-plant and head-chop silages. Anim Feed Sci Technol. 2009;152:175–85.
Dong Z, Shao T, Li J, Yang L, Yuan X. Effect of alfalfa microbiota on fermentation quality and bacterial community succession in fresh or sterile Napier grass silages. J Dairy Sci. 2020;103:4288–301.
Mendez-Garcia C, Pelaez AI, Mesa V, Sanchez J, Golyshina OV, Ferrer M. Microbial diversity and metabolic networks in acid mine drainage habitats. Front Microbiol. 2015;6:475.
Guan H, Yan Y, Li X, Li X, Shuai Y, Feng G, Ran Q, Cai Y, Li Y, Zhang X. Microbial communities and natural fermentation of corn silages prepared with farm bunker-silo in Southwest China. Bioresour Technol. 2018;265:282–90.
Zhao X, Liu J, Liu J, Yang F, Zhu W, Yuan X, Hu Y, Cui Z, Wang X. Effect of ensiling and silage additives on biogas production and microbial community dynamics during anaerobic digestion of switchgrass. Bioresour Technol. 2017;241:349–59.
Dong Z, Li J, Chen L, Wang S, Shao T. Effects of freeze-thaw event on microbial community dynamics during red clover ensiling. Front Microbiol. 2019;10:1559.
Muck RE. Recent advances in silage microbiology. Agr Food Sci. 2013;22:3–15.
Cai Y, Benno Y, Ogawa M, Ohmomo S, Kumai S, Nakase T. Influence of Lactobacillus spp. from an inoculant and of Weissella and Leuconostoc spp. from forage crops on silage fermentation. Appl Environ Microbiol. 1998;64:2982–7.
Chen D, Zheng M, Guo X, Chen X, Zhang Q. Altering bacterial community: a possible way of lactic acid bacteria inoculants reducing CO2 production and nutrient loss during fermentation. Bioresour Technol. 2021;329:124915.
Ostling C, Lindgren S. Influences of enterobacteria on the fermentation and aerobic stability of grass silages. Grass Forage Sci. 1995;50:41–7.
Pedroso AF, Adesogan AT, Queiroz OC, Williams SK. Control of Escherichia coli O157:H7 in corn silage with or without various inoculants: efficacy and mode of action. J Dairy Sci. 2010;93:1098–104.
Weissbach F, Honig H. On the anticipation and control of the run of fermentation in silage making from extensively grown forages. Landbauforsch Volk. 1996;46:10–7.
Pereira OG, Rocha KD, Fortes Ferreira CLDL. Chemical composition, characterization, and population of microorganisms on elephantgrass “Cameroon” (Pennisetum purpureum, Schum.) and its silages. Rev Bras Zootecn. 2007;36:1742–50.
Moore ERB, Tindall BJ, Martins Dos Santos VAP, Pieper DH, Ramos J-L, Palleroni NJ. Nonmedical: Pseudomonas. In: Dworkin M, Falkow S, Rosenberg E, Schleifer K-H, Stackebrandt E, editors. The prokaryotes: a handbook on the biology of bacteria volume 6: proteobacteria: gamma subclass. New York: Springer, New York; 2006. p. 646–703.
Bai J, Ding Z, Ke W, Xu D, Wang M, Huang W, Zhang Y, Liu F, Guo X. Different lactic acid bacteria and their combinations regulated the fermentation process of ensiled alfalfa: ensiling characteristics, dynamics of bacterial community and their functional shifts. Microb Biotechnol. 2021;14:1171–82.
Banerjee S, Kirkby CA, Schmutter D, Bissett A, Kirkegaard JA, Richardson AE. Network analysis reveals functional redundancy and keystone taxa amongst bacterial and fungal communities during organic matter decomposition in an arable soil. Soil Biol Biochem. 2016;97:188–98.
de Vries FT, Griffiths RI, Bailey M, Craig H, Girlanda M, Gweon HS, Hallin S, Kaisermann A, Keith AM, Kretzschmar M, Lemanceau P, Lumini E, Mason KE, Oliver A, Ostle N, Prosser JI, Thion C, Thomson B, Bardgett RD. Soil bacterial networks are less stable under drought than fungal networks. Nat Commun. 2018;9:3033.
Hernandez DJ, David AS, Menges ES, Searcy CA, Afkhami ME. Environmental stress destabilizes microbial networks. ISME J. 2021;15:1722–34.
Coyte KZ, Schluter J, Foster KR. The ecology of the microbiome: Networks, competition, and stability. Science. 2015;350:663–6.
Grilli J, Rogers T, Allesina S. Modularity and stability in ecological communities. Nat Commun. 2016;7:12031.
Berry D, Widder S. Deciphering microbial interactions and detecting keystone species with co-occurrence networks. Front Microbiol. 2014;5:219.
Banerjee S, Schlaeppi K, van der Heijden MGA. Keystone taxa as drivers of microbiome structure and functioning. Nat Rev Microbiol. 2018;16:567–76.
Xu D, Wang N, Rinne M, Ke W, Weinberg ZG, Da M, Bai J, Zhang Y, Li F, Guo X. The bacterial community and metabolome dynamics and their interactions modulate fermentation process of whole crop corn silage prepared with or without inoculants. Microb Biotechnol. 2021;14:561–76.
Ash C, Priest FG, Collins MD. Molecular identification of rRNA group 3 bacilli (Ash, Farrow, Wallbanks and Collins) using a PCR probe test. Anton Leeuw. 1993;64:253–60.
Li M, Zhang L, Zhang Q, Zi X, Lv R, Tang J, Zhou H. Impacts of citric acid and malic acid on fermentation quality and bacterial community of cassava foliage silage. Front Microbiol. 2020;11:595622.
Winters AL, Cockburn JE, Dhanoa MS, Merry RJ. Effects of lactic acid bacteria in inoculants on changes in amino acid composition during ensilage of sterile and non-sterile ryegrass. J Appl Microbiol. 2000;89:442–51.
This work was financially supported by the Joint Fund Project of the National Natural Foundation of China (U20A2003).
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Zhao, J., Yin, XJ., Li, JF. et al. Clean recovery and recycling of seasonal surplus forage grass by microbial driven anaerobic fermentation: a case study of napiergrass. Chem. Biol. Technol. Agric. 9, 91 (2022). https://doi.org/10.1186/s40538-022-00360-w
- Surplus forage grass
- Harvest date
- Anaerobic fermentation
- Microbial community
- Co-occurrence network