Integrated microbiology and metabolomic analysis reveal the improvement of rice straw silage quality by inoculation of Lactobacillus brevis

Abstract

Background: Ensiling technology holds promise for preserving and providing high-quality forage. However, the preservation of rice straw poses challenges due to its high lignocellulosic content and low water-soluble carbohydrate levels. Developing highly effective lactic acid bacteria (LAB) for rice straw silage remains a priority.

Results: This study evaluated the impact of three LAB strains, Lactobacillus brevis R33 (Lac33), L. buchneri R17 (Lac17), and Leuconostoc pseudomesenteroides (Leu), on the fermentation quality of rice straw silage. Rice straw silage inoculated with Lac33 alone or in combination with other strains exhibited significantly lower neutral detergent fiber (NDF) (66.5% vs. 72.3%) and acid detergent fiber (ADF) (42.1% vs. 47%) contents, along with higher lactic acid (19.4 g/kg vs. not detected) and propionic acid (2.09 g/kg vs. 1.54 g/kg) contents compared to control silage. Bacterial community analysis revealed Lactobacillus dominance (> 80%) and suppression of unwanted Enterobacter and Clostridium. Metabolomic analysis highlighted increased carbohydrates and essential amino acids, indicating improved nutrient values in Lac33-inoculated rice straw silage and a potential explanation for Lac33 dominance.

Conclusions: This research identified a highly efficient LAB candidate for rice straw silage, advancing our comprehension of fermentation from integrated microbiology and metabolomic perspectives.

Keywords: Bacterial community; Fermentation quality; Lactic acid bacteria; Metabolite profiles; Rice straw silage.

Fig. 1

Chemical characteristics of rice straw silage with different strain inoculations. The symbols and error bars represent the mean value and standard deviation of six biological replicates, respectively. Distinct letters on top indicate significant differences across all combinations (Tukey’s honestly significant difference test, P < 0.05). Dashed rectangles in red and blue depict groups inoculated without Lac33 and with Lac33, respectively. Abbreviations: CK, no inoculation; LB, inoculated with LB; Leu, inoculated with L. pseudomesenteroides; Lac17, inoculated with L. buchneri R17; Leu + Lac17, inoculated with L. pseudomesenteroides and L. buchneri R17; Lac33, inoculated with L. brevis R33; Leu + Lac33, inoculated with L. pseudomesenteroides and L. brevis R33; Lac17 + 33, inoculated with L. buchneri R17 and L. brevis R33; Leu + Lac17 + 33, inoculated with L. pseudomesenteroides, L. buchneri R17, and L. brevis R33. DML, dry matter loss; NDF, neutral detergent fiber; ADF, acid detergent fiber; ADL, acid detergent lignin; IVDMD, in vitro dry matter digestibility

Fig. 2

Analysis of bacterial communities in rice straw silage with various strain inoculations. a Principal coordinate analysis (PCoA) of bacterial communities in rice straw silage inoculated with different strains. The 95% confidence ellipses are depicted in respective colors for diverse communities. Dashed circles in blue and red signify groups inoculated without Lac33 and with Lac33, respectively. b Shannon indices of different samples. Boxplots indicate median, 25th and 75th percentiles, while error bars represent standard deviations of three replicates. c Relative abundance of phylum and d genus in bacterial communities of distinct silage samples. CK, no inoculation; LB, inoculated with LB; Leu, inoculated with L. pseudomesenteroides; Lac17, inoculated with L. buchneri R17; Leu + Lac17, inoculated with L. pseudomesenteroides and L. buchneri R17; Lac33, inoculated with L. brevis R33; Leu + Lac33, inoculated with L. pseudomesenteroides and L. brevis R33; Lac17 + 33, inoculated with L. buchneri R17 and L. brevis R33; Leu + Lac17 + 33, inoculated with L. pseudomesenteroides, L. buchneri R17, and L. brevis R33

Fig. 3

Taxonomic characterization of differential bacterial taxa between silage groups with and without L. brevis R33 inoculation. a Manhattan plots display ASVs enriched or depleted between bacterial communities of silage groups with and without L. brevis R33 inoculation. Specific ASVs are labeled with corresponding species, genus, or family. Filled triangles represent enriched ASVs, empty triangles signify depleted ASVs, and filled circles denote non-significant comparisons (FDR-adjusted P < 0.05, Wilcoxon rank sum test). Triangle and circle sizes are proportionate to ASV relative abundances. b Analysis of enriched and depleted ASVs summarized and displayed at the genus level. Abbreviations: CK, no inoculation; LB, inoculated with LB; Leu, inoculated with L. pseudomesenteroides; Lac17, inoculated with L. buchneri R17; Leu + Lac17, inoculated with L. pseudomesenteroides and L. buchneri R17; Lac33, inoculated with L. brevis R33; Leu + Lac33, inoculated with L. pseudomesenteroides and L. brevis R33; Lac17 + 33, inoculated with L. buchneri R17 and L. brevis R33; Leu + Lac17 + 33, inoculated with L. pseudomesenteroides, L. buchneri R17, and L. brevis R33

Fig. 4

Metabolite analysis of rice straw silage with various strain inoculations. a Principal Component Analysis (PCA) of metabolite profiles in rice straw silage with different strain inoculations. The 95% confidence ellipses are shown in respective colors for distinct communities. Dashed circles in blue and red signify groups inoculated without Lac33 and with Lac33, respectively. b Relative abundance of metabolite categories in silage groups with and without L. brevis R33 inoculation. Metabolites classified into different groups and labeled accordingly. c Volcano plots illustrate metabolites up or down-regulated between silage groups without and with L. brevis R33 inoculation. Dashed lines represent fold change (FC > 2 and < -2) and P-value (P < 0.05) thresholds. d Analysis of up and downregulated metabolites summarized and displayed at the subclass level. Abbreviations: CK, no inoculation; LB, inoculated with LB; Leu, inoculated with L. pseudomesenteroides; Lac17, inoculated with L. buchneri R17; Leu + Lac17, inoculated with L. pseudomesenteroides and L. buchneri R17; Lac33, inoculated with L. brevis R33; Leu + Lac33, inoculated with L. pseudomesenteroides and L. brevis R33; Lac17 + 33, inoculated with L. buchneri R17 and L. brevis R33; Leu + Lac17 + 33, inoculated with L. pseudomesenteroides, L. buchneri R17, and L. brevis R33

Fig. 5

KEGG pathway enrichment analysis of differential metabolites between silage groups with and without L. brevis R33 inoculation. a Differential abundance scores of KEGG pathways between silage groups with and without L. brevis R33 inoculation. Pathway size corresponds to differential metabolite counts. Differential abundance scores: 1 signifies upregulated differential metabolites; -1 represents downregulated differential metabolites. b Abundance of specific amino acids in silage groups with and without L. brevis R33 inoculation. Columns and error bars denote mean value and standard deviation of biological replicates (Without R33, n = 30; With R33, n = 24), respectively. Different letters on top indicate significant differences (Tukey’s honestly significant difference test, P < 0.05)

Fig. 6

Correlations between differential bacterial genera and metabolites between silage groups with and without L. brevis R33 inoculation. Heatmap constructed based on Spearman’s rank correlation analysis. Asterisks denote significant correlations (Spearman’s rho > 0.6 and < -0.6, P < 0.05)