Environmental factors shaping rumen bacterial and fungal community

Fig. 5 shows the correlation heatmap of important environmental factors (pH, acetic acid, propionic acid, butyric acid, cellulose, and hemicellulose) and bacterial and fungal community composition of the top 20 genera. As shown in Fig. 5(a), pH showed significantly positive correlations with Prevotella, Succiniclasticum, and Oscillibacter, while the other five factors had significantly negative correlations with these bacteria. Furthermore, pH showed significantly negative correlations with Bacteroides, Sphaerochaeta, Paraprevotella, and Parabacteroides, while the other five factors had significantly positive correlations with these bacteria. As shown in Fig. 5(b), pH showed positive correlations with Spizellomyces, Rhizophagus, Basidiobolus, Gonapodya, Mitosporidium, Allomyces, Catenaria, and Conidiobolus, while the other five factors had negative correlations with these fungi. Besides, pH showed negative correlations with Piromyces, Neocallimastix, and Anaeromyces, while the other five factors had positive correlations with these fungi. Therefore, pH was the main factor affecting hydrolysis and acidogenesis in rumen anaerobic fermentation (Darwin and Blignaut, 2019). A neutral pH is suitable for rumen microorganisms for hydrolysis and acidogenesis of lignocellulose (Meng et al., 2016). In this experiment, pH significantly decreased to 5.95 at 72 h due to the accumulation of VFAs. Some studies showed that low pH could inhibit the degradation efficiency of some hydrolytic bacteria and anaerobic fungi, even lead to rumen microbial imbalance affecting the gene expression of CAZymes and VFA metabolic pathway (Faniyi et al., 2019; Li et al., 2021a). The canonical correlation analysis (CCA) also indicated that pH was the main factor affecting bacterial and fungal community composition Fig 4. Therefore, low pH might change the rumen bacterial and fungal communities, and further affect the hydrolysis and acidogenesis of corn straw.

Fig 4. Canonical correlation analysis (CCA) of environmental factors and rumen bacterial (a) and fungal (b) community composition at genus level.

In general, keeping a pH within 5.8-7.2 was suitable for hydrolysis and acidogenesis of rumen microorganisms (Darwin and Blignaut, 2019). In this study, the obvious decrease of pH was caused by the VFA accumulation. Correspondingly, the correlations of VFAs and bacteria and fungi were completely opposite to those of pH. The results of CCA indicated that the VFAs were also the main factor affecting the bacterial and fungal community composition (in Supplementary material). However, some core bacteria and fungi with high relative abundance still existed at low pH, such as Prevotella, Bacteroides, Piromyces, and Neocallimastix etc. In particular, Piromyces and Neocallimastix showed positive correlations with acetic acid, propionic acid, butyric acid, cellulose, and hemicellulose. Therefore, the stop of VFA production might be due to the inhibition of acidic products, rather than the acidosis (Faniyi et al., 2019). Separating the VFAs might be a feasible strategy to maintain the hydrolysis and acidogenesis efficiency of rumen bacteria and fungi (Liang et al., 2020). Membrane separation of VFAs is a potential method to relieve the VFA accumulation, keeping the high activity of rumen microorganisms (Nguyen et al., 2019).

Fig 5. Correlation heatmap of environmental factors and bacterial (a) and fungal (b) communities in top 20 genera (Intensity of color represents correlation co-efficient; 0.01＜p＜0.05 and p(0.01 are denoted as * and **, respectively).

1.3 Co-occurrence network of rumen bacteria and fungi

Co-occurrence network analysis can better explain complex interactions and ecological processes beyond microbial community composition (Chen and Wen, 2021). Fig. 6 shows a co-occurrence networks of bacterial and fungal communities at genus level. Total of 99 nodes (genera) and 1026 edges constituted the bacterial correlation network, and positive correlations accounted for 79.3% in bacterial networks, whereas negative correlations were 20.7% Fig. 6(a). Total of 97 nodes (genera) and 837 edges constituted the fungal correlation network, and positive correlations accounted for 65.0% in fungal networks, whereas negative correlations were 35.0% Fig. 6(b). Positive links were ascribed to cross-feeding and mutualism, while negative links were attributed to amensalism and competition (Wang et al., 2021). In this study, the correlations between bacteria were more complex than those between fungi, showing that bacterial community was more stable due to stronger potential interactions in bacterial network (Chen and Wen, 2021).

Fig 6. Co-occurrence network of bacterial (a) and fungal (b) communities based on Spearman’s correlations at genus level (A connection stands for a strong correlation (R>0.8 or R<-0.8). The red and blue lines show positive and negative interactions between two individual nodes, respectively. The size of each node is proportional to the number of connections).

For the bacterial network, nodes (genera) mainly belonged to Firmicutes (55.6%), Bacteroidetes (20.2%), Actinobacteria (8.08%), and Proteobacteria (5.05%). Firmicutes and Bacteroidetes were mainly phylum reported in most studies, which have the ability to hydrolyze and acidify lignocellulose biomass (Liang et al., 2021b). The genera Eubacterium, Clostridium, Prevotella, Olsenella, Ruminococcus, Treponema, Papillibacter, Bacteroides, Alistipes, Pseudobutyrivibrio, and Fibrobacter were found in the network. Alistipes, Clostridium, Eubacterium, Olsenella, Papillibacter, and Pseudobutyrivibrio showed the positive correlations with other bacteria, while Prevotella showed a negative correlation with other bacteria. Alistipes, Clostridium, and Eubacterium have the ability to degrade lignocellulose to produce VFAs (Won et al., 2020). Olsenella can ferment carbohydrates to produce acetic acid (Mizrahi et al., 2021). Papillibacter and Pseudobutyrivibrio were widely considered to be butyric acid producer (Xue et al., 2020). Although Prevotella is also lignocellulose degrader, Prevotella had more competitors in rumen fermentation system. These genera were also reported to be the core genera of rumen bacteria and played an important role in hydrolysis and acidogenesis of corn straw (Mizrahi et al., 2021). In addition, the low relative abundance of some bacteria, such as Mogibacterium, Marvinbryantia, and Lachnoclostridium etc. showed strong correlations in network, which might play an important role in the whole bacterial network.

For the fungal network, nodes (genera) mainly belonged to Ascomycota (43.3%), Basidiomycota (18.6), Mucoromycota (13.4%), Chytridiomycota (8.25%), and Zoopagomycota (6.19%). The genera Piromyces, Neocallimastix, Anaeromyces, and Orpinomyces showed strong correlations in fungal network. These four genera are mainly cellulose and hemicellulose degraders (Morais and Mizrahi, 2019a), which showed the positive correlations among them and might play a key role in fungal ecosystem. Spizellomyces, Batrachochytrium, and Gonapodya belonging to the phylum Chytridiomycota also showed the strong correlations in network. Similar to bacterial network, some genera with low relative abundance also showed the strong correlations in fungal network, and need to be further explored. In addition, there might exist the synergy and competition relationships between rumen bacteria and fungi in the whole microbial ecosystem (Azad et al., 2020), which need to be further explored.

Carbohydrate-active enzymes composition

In most previous researches, a high relative abundance of GHs was reported, followed by carbohydrate binding modules (CBMs), glycosyl transferases (GTs), carbohydrate esterases (CEs), and other CAZymes in rumen fermentation (Hinsu et al., 2021; Huws et al., 2021). In this study, GHs, GTs, CBMs, and CEs were the main CAZymes at class level involved in carbohydrates metabolism (Fig. 7). The relative abundance of GHs, CBMs, and CEs slightly increased from 3.88%, 0.54%, and 0.47% at 0 h to 4.05%, 0.94%, and 0.50% at 24 h, and significantly decreased to 3.22%, 0.48%, and 0.38% at 72 h, respectively; while the relative abundance of GTs changed insignificantly. GHs include glycosidases and transglycosidases, which are the most abundant and diverse group of enzymes responsible for hydrolysis or transglycosylation of glycosidic bonds, accounting for 50% of the enzymes classified in CAZymes database (Luo et al., 2020). CBMs have no catalytic activity of their own, but help in binding CAZymes (GHs and CEs) to carbohydrates, thus promoting their activities. CEs can remove ester-based modifications existing in monosaccharides, oligosaccharides, and polysaccharides, thus promoting the action of GHs on complex polysaccharides. GTs can catalyze the transfer of specific glycosyl donors to receptors to form glycosidic bonds (Bohra et al., 2019)

Fig 7. Dynamic change of carbohydrate-active enzymes (CAZymes) at class level in rumen fermentation (a), and heatmap of carbohydrate-active enzymes at family level in top 30 (b).

To further explore the change of CAZymes composition, top 30 of CAZymes at family level are shown in Table 1. Cellulases (GH5, GH9, GH51), hemicellulases (GH10, GH28), xyloglucanase (GH94), debranching enzymes (GH77, GH78), oligosaccharide degrading enzymes (GH2, GH3, GH13, GH20, GH29, GH31, GH32, GH36, GH43, GH92, GH95, GH97), and a few CBMs, CEs, and CTs were identified in this study. Among them, GH2 (β-Galactosidases), GH3 (β-Glucosidases), GH13 (α-Amylases), GH31 (α-Glucosidase), GH43 (β-Xylosidase), GT2 (Cellulose synthase), and CE1 (Acetyl xylan esterase) were the most dominant families, which also were the most dominant CAZymes in other studies (Morais and Mizrahi, 2019a). Meanwhile, top 30 heatmap of CAZymes families showed that most CAZymes families significantly decreased after 72 h fermentation (in Supplementary material).

Table 1 Variation of CAZymes relative abundance (%) in top 30 at family level.

The relative abundance of GH5, GH9, and GH51 decreased in 72 h, which were classified as cellulases and showed endoglucanase activity. GH10 and GH28 were classified as hemicellulases, and their relative abundance decreased in 72 h. GH2, GH3, GH13, GH20, GH29, GH31, GH32, GH36, GH43, GH92, GH95, and GH97 were classified as oligosaccharide-degrading enzymes, which showed the activity of galactosidase, amylase, xylosidase, mannosidase, or glucosidase. Apart from that the relative abundance of GH20, GH29, and GH92 increased, the relative abundance of other GHs families decreased in 72 h in rumen fermentation. The relative abundance of CBM6 (cellulose-binding function) and CBM48 (glycogen-binding function) decreased in 72h. In addition, CE1, CE6, and CE8 show the activity of acetyl xylan esterase and pectin methylesterase (Tomazetto et al., 2020), and their relative abundance decreased in 72 h. All of the above showed that enzyme profiles significantly changed during rumen fermentation, and the relative abundance of most CAZymes families significantly decreased, indicating that the hydrolysis efficiency significantly decreased in 72 h. This might be due to the acid accumulation and depletion of biodegradable substrates (Liang et al., 2020). Zhang et al. (2020) proved that the relative abundance of CAZymes was positively corrected with the richness in cellulose and hemicellulose. Meanwhile, the decrease of Prevotella relative abundance might lead to the decrease of CAZymes families, because Prevotella was a major contributor to CAZymes families (Shen et al., 2020). In addition, low pH could affect the expression of CAZymes genes to reduce the hydrolysis efficiency of corn straw (Li et al., 2021a).

Metabolic pathway of VFA production in rumen fermentation

Rumen microorganisms convert lignocellulose to monosaccharides, and further to pyruvate, which is a crucial intermediate to subsequent VFAs (mainly acetic acid, propionic acid, and butyric acid) through different metabolic pathways (Liang et al., 2020). Fig. 8 shows the main metabolic pathways of VFA generation during rumen fermentation of corn straw based on the KEGG mapper. For glucose to pyruvate pathway, the relative abundance of glucokinase (GLK) and pyruvate kinase (PK) decreased from 0.059% and 0.023% at 0 h to 0.048% and 0.016% at 72 h, respectively (in Supplementary material). The decrease of GLK and PK relative abundance led to the decrease of pyruvate production, and then inhibited further VFA generation (Zhang et al., 2020). For pyruvate to acetic acid pathway, the relative abundance of phosphate acetyltransferase (PTA) and acetate kinase (ackA) increased from 0.034% and 0.052% at 0 h to 0.037% and 0.059% at 72 h, respectively. High relative abundance of PTA and ackA promoted the acetic acid formation, which was also observed by Luo et al. (2020).

The pathways of succinic acid and acrylic acid were to synthesize propionic acid in anaerobic fermentation (Liang et al., 2020). In this study, succinic acid pathway mainly produced propionic acid, while the relative abundance of most genes was lower and decreased or unchanged with fermentation time in acrylic acid pathway. After 27 h fermentation, the relative abundance of malate dehydrogenase (MDH), tetrahydrofolate ligase (FHs), fumarate reductase (frdA), and propionate CoA-transferase (PCT) decreased. The decrease of FHs, frdA, and PCT relative abundance were not conducive to the production of propionic acid (Zhang et al., 2020). For pyruvate to butyric acid pathway, the relative abundance of acetyl-CoA C-acetyltransferase (atoB) and phosphate butyryltransferase (PTB) decreased from 0.011% and 0.0070% at 0 h to 0.0089% and 0.0020% at 72 h, respectively. The decrease of atoB and PTB relative abundance greatly reduced the butyric acid production, resulting in more production of acetic acid from acetyl-CoA.

Fig 8. Change of functional genes in principle metabolic pathways for acetic acid, propionic acid, and butyric acid production (The red and green letter represented the increase and decrease of gene relative abundance after 72 h fermentation, respectively).

Meanwhile, the genes related to acetic acid generation (POR, PTA, and ackA) and propionic acid generation (pycB, MDH, FHs, frdA, and MUT) showed higher relative abundance than that to butyric acid generation (atoB, PTB, and HBD) during all fermentation time (in Supplementary material), corresponding to more acetic acid and propionic acid accumulation and less butyric acid accumulation . The decrease of hydrolysis and acidogenesis efficiency with the fermentation time might be due to the down regulation of some key genes, such as GLK, PK, and MDH etc., which was caused by the change of rumen microbial community (Li et al., 2021a). Some studies reported that low pH changed the rumen microbial community composition, and inhibited the microbial activity, further affecting the lignocellulose degradation (Meng et al., 2016; Sung et al., 2007). In this study, pH dropped from 7.45 to 5.95 , which might change metabolic pathways through decreasing the expression of key genes participating in metabolic pathways of glucose and pyruvate into acetic acid, propionic acid, and butyric acid to influence corn straw degradation and VFA generation (Li et al., 2021a).

2. Microbial analysis in rumen semi-continuous reactors

Diversity and abundance

Fig. 9 demonstrates the variation of Shannon and Chao 1 index at different corn straw loads, representing the diversity and abundance of rumen bacteria, respectively. A significant decrease in the bacterial diversity and richness was observed in the reactor at 8% corn straw loads. In a study by Liang et al. (2021), the diversity and abundance of rumen bacteria significantly decreased with increasing rice straw load, and the lowest diversity was observed at 10% load. In most studies of rumen microbial fermentation of biomass waste for VFA production, a significant decrease in bacterial diversity and abundance with high load was also observed (Nguyen et al., 2020; Xing et al., 2020a). This was mainly due to the disappearance of some rumen microorganisms that were not adapted to the in vitro environment. Meanwhile, the increase in substrate load resulted in higher VFA concentration, which further led to a decrease in pH. Low pH was detrimental to rumen microbial growth and might also be one of the reasons for reducing the diversity of rumen microorganisms (Meng et al., 2016).

Fig 9. Shannon and Chao 1 index of rumen bacteria in a reactor at different corn straw loads (Significant differences between corn straw loads were indicated by different letters).

Community composition

Fig. 10 demonstrates the Venn diagram and community of bacteria at phylum and genus level in the reactor at different corn straw loads. Venn diagrams showed that 1665, 1489, and 1164 OTUs were detected at 2.5%, 5.0%, and 8.0% corn straw loads, respectively (Fig. 10a). The shared OTUs accounted 40.3% in 2.5%, 5.0%, and 8.0% corn straw loads, indicating that this semi-continuous reactor shared some microorganisms at different corn straw loads. Meanwhile, the percentage of shared OTUs gradually decreased with the elevation of corn straw loads. Thus, the number and percentage of OTUs gradually decreased with increasing corn straw loads, which also corresponded to a decrease in bacterial diversity in this semi-continuous reactor.

Fig 10. Microbial community composition at different corn straw loads in a rumen semi-continuous reactor. (a) Venn diagram; (b) phylum level; (c) genus level.

Bacteroidetes and Firmicutes were the main phyla in this rumen semi-continuous reactors at different corn straw loads (Fig. 10b), and were also found in most studies on rumen microbial fermentation of biomass waste (Mizrahi et al., 2021). Bacteroidetes and Firmicutes are important phyla in rumen ecosystem and contain various hydrolyzing and acid-producing genera (Morais and Mizrahi; 2019b). As the corn straw loads increased, Bacteroidetes and Firmicutes relative abundance slightly decreased, while Proteobacteria and Actinobacteriota relative abundance significantly increased despite their low relative abundances.

To further explore the variations in rumen bacterial community composition, a total of 17 genera with relative abundance of at least more than 1% were identified in Fig. 10c. In the following study, Rikenellaceae_RC9_gut_group, Prevotella, Saccharofermentans, Ruminococcus, Treponema, and Rhodococcus were the main genera in the reactor at different corn straw loads. The principal co-ordinates analysis results indicated significant changes in bacterial community composition at genus level in the reactor at different corn straw loads (Fig. 11).

Fig 11. The PCoA of microbial community composition at different corn straw loads in a rumen semi-continuous reactor.

As corn straw loads increased, the relative abundance of Rikenellaceae_RC9_gut_group, Christensenellaceae_R-7_group, Ruminococcus, and Treponema, and Prevotellaceae_UCG-004 firstly increased at 5.0% corn straw load, and then significantly decreased at 8.0% load. Rikenellaceae_RC9_gut_group, Ruminococcus, and Treponema are the core hydrolyzing and acid-producing bacteria in rumen fermentation (Huws et al., 2021; Ahmad et al., 2020). Christensenellaceae_R-7_group is widely found in the rumen of cattle, sheep, and deer (Wei et al., 2023). Liang et al. (2021) also reported that an increase in rice straw loads resulted in a lower relative abundance of Rikenellaceae_RC9_gut_group, Christensenellaceae_R-7_group, Ruminococcus, and Treponema. In addition, a significant decrease in the relative abundance of Papillibacter was observed with increasing corn straw loads. Thus, the changes in these genera also reflect changes in the ecosystem in this rumen semi-continuous reactor.

The relative abundance of Prevotella, Saccharofermentans, Succiniclasticum, Lachnospiraceae_NK3A20_group, Ruminococcus_gauvreauii_group, Marvinbryantia, Lachnospiraceae_FCS020_group, Rhodococcus, and Eubacterium_nodatum_group increased with increasing corn straw loads. Especially the relative abundance of Prevotella and Saccharofermentans increased from 4.3% and 0.7% at 2.5% corn straw load to 7.7% and 5.1% at 5.0% corn straw load, respectively, and finally significantly increased to 22.1% and 9.2% at 8.0% corn straw load, respectively. Prevotella is a well-recognized hemicellulose-degrading bacterium in the rumen ecosystem, is widely found in lignocellulose-rich environments, and converts soluble sugars into acetate and propionate (Betancur-Murillo et al., 2023). Saccharofermentans is a cellulose-degrading bacterium in the rumen ecosystem and can produce acetate and propionate (Hu et al., 2021).

Some low relative abundance bacterial genera at 2.5% corn straw load, namely Ruminococcus_gauvreauii_group, Marvinbryantia, Rhodococcus, and Eubacterium_nodatum_group significantly increased to 2.3%, 1.0%, 1.2%, and 1.1%, at 8% corn straw load, respectively. Ruminococcus_gauvreauii_group belonging to the family Lachnospiraceae can convert glucose into acetate (Cheng et al., 2023). Marvinbryantia has been reported as a cellulose-degrading bacterium in the rumen ecosystem (Baniel et al., 2021). Rhodococcus can secrete efficient enzymes that can degrade waste biological resources (e.g., lignin) into lipids (Zhao et al., 2024). The increase of these microorganisms may ensure efficient VFA production in the reactor at 8% corn straw load.

LEfSe analysis

Significant changes in bacterial community composition occurred in the reactor at different corn straw loads. Fig. 12(a) shows the biomarkers of bacteria at genus level at different corn straw loads in this reactor. The 21, 9, and 27 biomarkers were identified at genus level for 2.5%, 5.0%, and 8.0% corn straw loads, respectively. Christensenellaceae_R-7_group, Papillibacter, Lachnospiraceae_XPB1012_group, Ruminiclostriduim, Eubacterium_hallii_group, and Monoglobus were the main biomarkers at 2.5% corn straw load. Treponema, Rikenellaceae_RC9_gut_group, Lachnoclostridium, Lachnospiraceae_AC2044_group, Pseudobutyrivibrio, and Anaerovorax were the main biomarkers at 5.0% corn straw load. Prevotella, Saccharofermentans, Ruminococcus_gauvreauii_group, Pyramidobacter, Acetobacter, Butyrivibrio, Brevundimonas, Pseudoclavibacter, Streptomyces, and Curtobacterium were the main biomarkers at 8.0% corn straw load. Thus, these rumen bacterial biomarkers at different corn straw loads may be important for hydrolysis and acidogenesis.

Fig 12. (a) Histogram of LDA scores (LDA scores >3.0) of microorganisms at the genus level and (b) heatmap of microorganisms and environmental factors based on the Spearman’s rank correlation in the top 30 genera in a reactor at different corn straw loads (Intensity of color represents correlation co-efficient; 0.01＜ p＜ 0.05, 0.001 ＜p＜ 0.01, and p＜ 0.001 are denoted as *, **, and ***, respectively.).

The 8% corn straw load had the most biomarkers (Fig. 6a). Prevotella and Saccharofermentans were the highest abundance genera at 8% corn straw load and were reported to be key hydrolyzing and acid-producing bacteria in the rumen ecosystem (Betancur-Murillo et al., 2023). These two genera are bound to occupy important ecological niches in the microecosystem, which are also guaranteed for corn straw hydrolysis and VFA production. Meanwhile, some low abundance genera such as Pyramidobacter, Acetobacter, Streptomyces, Pseudoclavibacter, and Butyrivibrio were found in 8% corn straw load. Pyramidium is a cellulolytic bacterium in the rumen that produces acetate as the main product (Guo et al., 2021). Acetobacter is often found in higher levels of acetate in rumen fluid (Gu et al., 2021). Butyrivibrio is a recognized fiber-degrading bacterium in the rumen and produces butyrate as its main product (Won et al., 2020). Compared to 2.5% corn straw load, these low abundance genera significantly increased at 8.0% corn straw load . Subtle differences in the microbial community may cause significant performances in hydrolysis and acidogenesis of anaerobic fermentation (Liang et al., 2023). In this study, these differential genera abundances might be the main drivers for corn straw degradation and VFA production at 8.0% corn straw load.

Environmental factors shaping the microbial community

Fig. 6(b) illustrates the correlation between bacteria and environmental factors at the genus level in this reactor. Prevotella, Saccharofermentans, Ruminococcus, Ruminococcus_gauvreauii_group, Prevotellaceae_UCG-001, Succiniclasticum, Marvinbryantia, Rhodococcus, and Moryella showed positive correlations with the VFA, acetate, propionate, and butyrate concentrations, and showed negative correlations with pH and VS removal. Prevotella, Saccharofermentans, Ruminococcus, etc. are the core acid-producing bacteria in the rumen (Morais and Mizrahi; 2019a), and maybe the main VFA producers in this study. Christensenellaceae_R-7_group and Papillibacter showed positive correlations with pH and VS removal, and showed negative correlations with the VFA concentration. In this study, a significant decrease in pH and VS removal occurred with increasing corn straw load. Meanwhile, the decrease in relative abundance of core genera (such as Papillibacter, Christensenellaceae_R-7_group) was observed. Some studies have shown that low pH severely affected the activity of hydrolyzing bacteria in rumen (Liang et al., 2023; Meng et al., 2016). Thus, as the corn straw load increased, accumulation of VFAs caused the decrease in pH, and low pH caused a decrease in the activity of hydrolyzing and acid-producing bacteria, and further led to the lower VS removal.

In this study, the highest concentration of VFAs was observed at 8% corn straw load, but the VS removal was the lowest, and much lower than the efficiency of biomass degradation in the rumen of ruminants. In the rumen of ruminants, VFAs can be absorbed, which may be a strategy to maintain efficient biomass degradation (Nishihara et al., 2023; Weimer, 2022). Therefore, timely separation of VFAs from the rumen semi-continuous reactor may further enhance the degradation efficiency of corn straw, drawing on the digestive strategy of ruminants.

Metabolic pathways

Fig. 13 shows the expression of key functional genes involved in hydrolysis and acidogenesis in the rumen semi-continuous reactor at different corn straw loads based on the KEGG information. Firstly, the main components of cellulose and hemicellulose in corn straw are degraded into hexose and pentose, respectively, and then converted into pyruvate. Finally, pyruvate is fermented into acetate, propionate, and butyrate through different metabolic pathways. As shown in Fig. 13(a), cellulose and hemicellulose were hydrolyzed to glucose and pentose by endoglucanase and beta-glucosidase and endoxylanase and oligosaccharide reducing-end xylanase, respectively, which were further converted to pyruvate by glucokinase and pyruvate kinase. As the corn straw loads in the rumen semi-continuous reactor increased, the expression of hydrolysis related to functional genes gradually increased, indicating that more corn straw was converted to soluble organic matter. The decline of Prevotella relative abundance was accompanied by a significant decrease in the expression of hydrolysis-related functional genes during rumen fermentation of corn straw as reported by Liang et al. (2023). Thus, the increase of core genera relative abundance such as Prevotella, Saccharofermentans, and Ruminococcus might contribute to the high expression of hydrolysis-related functional genes.

Fig 13. Expression of functional genes based on the PICRUSt 2 in a semi-continuous reactor at different corn straw concentrations. (a) Hydrolysis and (b) acidogenesis.

As shown in Fig. 7(b), pyruvate was catalyzed by acetyl CoA synthetase to synthesize acetyl CoA, which was further catalyzed by acetate kinase and butyrate kinase to form acetate and butyrate, respectively. The abundance of acetyl CoA synthetase at 8.0% corn straw load was significantly higher than that of 2.5% and 5.0% corn straw loads. Meanwhile, acetate kinase, phosphate acetyltransferase, butyrate kinase, and phosphate butyryltransferase were significantly expressed at 8.0% corn straw load compared to 2.5% and 5.0% loads. High expression of these functional genes involved in acetate and butyrate synthesis corresponds to a significant increase in acetate and butyrate concentrations at 8% corn straw load.

Generally, propionate is synthesized by the acrylate and succinate pathway (Yin et al., 2023). In the succinate pathway, the abundance of malate dehydrogenase, tetrahydrofolate ligase, and methylmalonyl-CoA mutase at 8.0% corn straw load was significantly higher than that of 2.5% and 5.0% corn straw loads. While the abundance of L-lactate dehydrogenase and butyryl-CoA dehydrogenase at 8.0% corn straw load was significantly lower than that of 2.5% and 5.0% corn straw loads. The abundance of functional genes in the succinate pathway was much higher than that in the acrylic pathway. Thus, the succinate pathway was the major pathway for propionate synthesis with increasing corn straw loads, which might correspond to an increase in the proportion of propionate at 8% corn straw load.

The above results indicated that corn straw loads significantly altered the expression of functional genes involved in hydrolysis and acidogenesis, which was mainly due to changes in the bacterial community composition.

References

1. Morais, S., Mizrahi, I., 2019b. The road not taken: The rumen microbiome, functional groups, and community states. Trends Microbiol. 27(6), 538-549. https://doi.org/10.1016/j.tim.2018.12.011
2. Won, M.Y., Oyama, L.B., Courtney, S.J., Creevey, C.J., Huws, S.A., 2020. Can rumen bacteria communicate to each other? Microbiome 8, 23. https://doi.org/10.1186/s40168-020-00796-y
3. Ahmad, A.A., Yang, C., Zhang, J., Kalwar, Q., Liang, Z., Li, C., Du, M., Yan, P., Long, R., Han, J., Ding, X., 2020. Effects of dietary energy levels on rumen fermentation, microbial diversity, and feed efficiency of yaks (Bos grunniens). Front. Microbiol. 11, 625. https://doi.org/10.3389/fmicb.2020.00625
4. Cheng, Y.F., Wang, Y., Li, Y.F., Zhang, Y.P., Liu, T.Y., Wang, Y., Sharpton, T.J., Zhu, W.Y., 2017. Progressive colonization of bacteria and degradation of rice straw in the rumen by Illumina sequencing. Front. Microbiol. 8, 2165. https://doi.org/10.3389/fmicb.2017.02165
5. Liu, J.H., Zhang, M.L., Xue, C.X., Zhu, W.Y., Mao, S.Y., 2016. Characterization and comparison of the temporal dynamics of ruminal bacterial microbiota colonizing rice straw and alfalfa hay within ruminants. J. Dairy Sci. 99, 9668-9681. https://doi.org/10.3168/jds.2016-11398
6. Jin, W., Wang, Y., Li, Y.F., Cheng, Y.F., Zhu, W.Y., 2018. Temporal changes of the bacterial community colonizing wheat straw in the cow rumen. Anaerobe 50, 1-8. https://doi.org/10.1016/j.anaerobe.2018.01.004
7. Holman, D.B., Gzyl, K.E., 2019. A meta-analysis of the bovine gastrointestinal tract microbiota. FEMS Microbiol. Ecol. 95(6), fiz072. https://doi.org/10.1093/femsec/fiz072
8. Liang, J.S., Zhang, H.B., Zhang, P.Y., Zhang, G.M., Cai, Y.Y., Wang, Q.Y., Zhou, Z.Y., Ding, Y.R., Zubair, M., 2021a. Effect of substrate load on anaerobic fermentation of rice straw with rumen liquid as inoculum: Hydrolysis and acidogenesis efficiency, enzymatic activities and rumen bacterial community structure. Waste Manag. 124, 235-243. https://doi.org/10.1016/j.wasman.2021.02.017
9. Mizrahi, I., Wallace, R.J., Moraïs, S., 2021. The rumen microbiome: balancing food security and environmental impacts. Nat. Rev. Microbiol. 19, 553-566. https://doi.org/10.1038/s41579-021-00543-6
10. Li, Y., Meng, Z., Xu, Y., Shi, Q., Ma, Y., Aung, M., Cheng, Y., Zhu, W.Y., 2021b. Interactions between anaerobic fungi and methanogens in the rumen and their biotechnological potential in biogas production from lignocellulosic materials. Microorganisms 9, 190. https://doi.org/10.3390/microorganisms9010190
11. Rezaeian, M., Beakes, G.W., Parker, D.S., 2004. Distribution and estimation of anaerobic zoosporic fungi along the digestive tracts of sheep. Mycol. Res. 108(10), 1227-1233. https://doi.org/10.1017/S0953756204000929
12. Aydin, S., Yildırım, E., Ince, O., Ince, B., 2017. Rumen anaerobic fungi create new opportunities for enhanced methane production from microalgae biomass. Algal Res. 23, 150-160. https://doi.org/10.1016/j.algal.2016.12.016
13. Faniyi, T.O., Adegbeye, M.J., Elghandour, M.M.M.Y., Pilego, A.B., Salem, A.Z.M., Olaniyi, T.A., Adediran, O., Adewumi, M.K., 2019. Role of diverse fermentative factors towards microbial community shift in ruminants. J. Appl. Microbiol. 127(1), 2-11. https://doi.org/10.1111/jam.14212
14. Darwin, Blignaut, D., 2019. Alkaline treatment for preventing acidosis in the rumen culture fermenting carbohydrates: An experimental study in vitro. J. Adv. Vet. Anim. Res. 6(1), 100-107. https://doi.org/10.5455/javar.2019.f319
15. Meng, Y., Mumme, J., Xu, H., Wang, K.J., 2016. A biologically inspired variable-pH strategy for enhancing short-chain fatty acids (SCFAs) accumulation in maize straw fermentation. Bioresour. Technol. 201, 329-336. https://doi.org/10.1016/j.biortech.2015.11.064
16. Chen, W.D., Wen, D.H., 2021. Archaeal and bacterial communities assembly and co-occurrence networks in subtropical mangrove sediments under Spartina alterniflora invasion. Environ. Microbiome 16(1), 10. https://doi.org/10.1186/s40793-021-00377-y
17. Wang, X.J., Wu, H., Wang, X.Y., Wang, H.X., Zhao, K.K., Ma, B., Lu, Z.M., 2021. Network-directed isolation of the cooperator Pseudomonas aeruginosa ZM03 enhanced the dibutyl phthalate degradation capacity of Arthrobacter nicotianae ZM05 under pH stress. J. Hazard. Mater. 410, 124667. https://doi.org/10.1016/j.jhazmat.2020.124667
18. Xue, Y., Lin, L., Hu, F., Zhu, W., Mao, S.Y., 2020. Disruption of ruminal homeostasis by malnutrition involved in systemic ruminal microbiota-host interactions in a pregnant sheep model. Microbiome 8(1), 138. https://doi.org/10.1186/s40168-020-00916-8
19. Azad, E., Fehr, K.B., Derakhshani, H., Forster, R., Acharya, S., Khafipour, E., McGeough, E., McAllister, T.A., 2020. Interrelationships of fiber-associated anaerobic fungi and bacterial communities in the rumen of bloated cattle grazing alfalfa. Microorganisms 8(10), 1543. https://doi.org/10.3390/microorganisms8101543
20. Hinsu, A.T., Tulsani, N.J., Panchal, K.J., Pandit, R.J., Jyotsana, B., Dafale, N.A., Patil, N.V., Purohit, H.J., Joshi, C.G., Jakhesara, S.J., 2021. Characterizing rumen microbiota and CAZyme profile of Indian dromedary camel (Camelus dromedarius) in response to different roughages. Sci. Rep. 11(1), 9400. https://doi.org/10.1038/s41598-021-88943-9
21. Huws, S.A., Edwards, J.E., Lin, W., Rubino, F., Alston, M., Swarbreck, D., Caim, S., Stevens, P.R., Pachebat, J., Won, M., Oyama, L.B., Creevey, C.J., Kingston-Smith, A.H., 2021. Microbiomes attached to fresh perennial ryegrass are temporally resilient and adapt to changing ecological niches. Microbiome 9, 143. https://doi.org/10.1186/s40168-021-01087-w
22. Luo, J.Y., Huang, W.X., Guo, W., Ge, R., Zhang, Q., Fang, F., Feng, Q., Cao, J.S., Wu, Y., 2020. Novel strategy to stimulate the food wastes anaerobic fermentation performance by eggshell wastes conditioning and the underlying mechanisms. Chem. Eng. J. 398, 125560. https://doi.org/10.1016/j.cej.2020.125560
23. Moraïs, S., Mizrahi, I., 2019a. Islands in the stream: from individual to communal fiber degradation in the rumen ecosystem. FEMS Microbiol. Rev. 43, 362-379. https://doi.org/10.1093/femsre/fuz007
24. Tomazetto, G., Pimentel, A.C., Wibberg, D., Dixon, N., Squina, F.M., 2020. Multi-omic directed discovery of cellulosomes, polysaccharide utilization loci, and lignocellulases from an enriched rumen anaerobic consortium. Appl. Environ. Microbiol. 86, e00199-20. https://doi.org/10.1128/AEM.00199-20
25. Shen, J., Zheng, L.X., Chen, X.D., Han, X.Y., Cao, Y.C., Yao, J.H., 2020. Metagenomic analyses of microbial and carbohydrate-active enzymes in the rumen of dairy goats fed different rumen degradable starch. Front. Microbiol. 11, 1003. https://doi.org/10.3389/fmicb.2020.01003
26. Sung, H.G., Kobayashi, Y., Chang, J., Ha, A., Hwang, I.H., Ha, J., 2007. Low ruminal pH reduces dietary fiber digestion via reduced microbial attachment. Asian Austral. J. Anim. 20, 200. https://doi.org/10.5713/ajas.2007.200
27. Wei, X., Wu, H., Wang, Z., Zhu, J., Wang, W., Wang, J., Wang, Y., Wang, C., 2023. Rumen-protected lysine supplementation improved amino acid balance, nitrogen utilization and altered hindgut microbiota of dairy cows. Anim. Nutri. 15, 320-331. https://doi.org/10.1016/j.aninu.2023.08.001
28. Betancur-Murillo, C.L., Aguilar-Marín, S.B., Jovel, J., 2023. Prevotella: A key player in ruminal metabolism. Microorganisms 11, 1. https://doi.org/10.3390/microorganisms11010001
29. Hu, W., Wu, Y., Bian, Y., Zheng, X., Chen, Y., Dong, L., Chen, Y., 2021. Joint effects of carbon nanotubes and copper oxide nanoparticles on fermentation metabolism towards Saccharofermentans acetigenes: Enhancing environmental adaptability and transcriptional expression. Bioresource Technol. 336, 125318. https://doi.org/10.1016/j.biortech.2021.125318
30. Baniel, A., Amato, K.R., Beehner, J.C., Bergman, T.J., Mercer, A., Perlman, R.F., Petrullo, L., Reitsema, L.J., Sams, S.N., Lu, A., Snyder‐Mackler, N., 2020. Seasonal shifts in the gut microbiome indicate plastic responses to diet in wild geladas. Microbiome 9, 26. https://doi.org/10.1186/s40168-020-00977-9
31. Gu, F., Liang, S., Zhu, S., Liu, J., Sun, H.Z., 2021. Multi-omics revealed the effects of rumen-protected methionine on the nutrient profile of milk in dairy cows. Food Res. Int. 149, 110682. https://doi.org/10.1016/j.foodres.2021.110682
32. Weimer, P.J., 2022. Degradation of cellulose and hemicellulose by ruminal microorganisms. Microorganisms 10(12), 2345. https://doi.org/10.3390/microorganisms10122345
33. Yin, Y., Zhang, T., He, S., Wang, J., 2023. Volatile fatty acids recovery and antibiotic degradation from erythromycin fermentation residues by combined thermal pretreatment and anaerobic fermentation: Insights into microbial communities and metabolic pathways. Bioresour Technol. 387, 129691. https://doi.org/10.1016/j.biortech.2023.129691