Profiling of the viable bacterial and fungal microbiota in fermented feeds using single-molecule real-time sequencing

Abstract

Fermented concentrated feed has been widely recognized as an ideal feed in the animal industry. In this study, we used a powerful method, coupling propidium monoazide (PMA) pretreatment with single-molecule real-time (SMRT) sequencing technology to compare the bacterial and fungal composition of feeds before and after fermentation with four added lactic acid bacteria (LAB) inoculants (one Lactobacillus casei strain and three L. plantarum strains). Five feed samples consisting of corn, soybean meal, and wheat bran were fermented with LAB additives for 3 d. Following anaerobic fermentation, the pH rapidly decreased, and the mean numbers of LAB increased from 106 to 109 colony-forming units (cfu)/g fresh matter. SMRT sequencing results showed that the abundance and diversity of bacteria and fungi in the feed were significantly higher before fermentation than after fermentation. Fifteen bacterial species and eight fungal genera were significantly altered following fermentation, and L. plantarum was the dominant species (relative abundance 88.94%) in the post-fermentation group. PMA treatment revealed that the bacteria Bacillus cereus, B. circulans, Alkaliphilus oremlandii, Cronobacter sakazakii, Paenibacillus barcinonensis, and P. amylolyticus (relative abundance >1%) were viable in the raw feed. After fermentation, their relative abundances decreased sharply to <0.2%; however, viable L. plantarum was still the dominant species post fermentation. We inferred that our LAB additives grew rapidly and inhibited harmful microorganisms and further improved feed quality. In addition, coupling PMA treatment with the Pacific Biosciences SMRT sequencing technology was a powerful tool for providing accurate live microbiota profiling data in this study.

Keywords: bacterial and fungal community; fermented feeds; lactic acid bacteria; propidium monoazide treatment; single-molecule real-time sequencing.

Figure 1.

Boxplots of the α-diversity indexes of bacteria (A) and fungi (B) in feed samples.

Figure 2.

Heat map of the relative abundance of 15 bacteria species detected in feed samples with significant differences.

Figure 3.

UniFrac principal coordinate analysis (PCoA) of the bacterial (A, B) and fungal (C, D) communities in feed samples. Each symbol represents the feed microbiota of one sample; sample group is represented by the respective color.

Figure 4.

Linear discriminant analysis effect size (LEfSe) and linear discriminant analysis (LDA) based on operational taxonomic units characterize the bacteria of the four groups. (A) Taxonomic cladogram produced from LEfSe analysis. Significant discriminant taxon nodes of the CA, CB, PA, and PB groups are represented by green, red, purple, and blue, respectively, while non-discriminant taxon nodes are represented by yellow. Branch areas are shaded according to the highest-ranked variety for that taxon. (B) Horizontal bar chart showing discriminant taxa. Significant discriminant taxa of the CA, CB, PA, and PB groups are represented by red, green, blue, and purple, respectively. The LDA score indicates the level of differentiation among the four groups of feed samples.