3.1. Effect of endolysin LyJH307 on Ruminal fermentation characteristics and qPCR
Compared to CON, endolysin LyJH307 supplementation did not affect gas production in any of the investigated incubation times (Table 2). The pH was significantly higher in the LyJH307 group than in the CON group (P = 0.0335), whereas IVDMD, NH3-N, and total VFA production were not changed by endolysin LyJH307 supplementation (IVDMD, P = 0.3597; NH3-N, P = 0.1485; and total VFA, P = 0.2592). In each proportion of individual VFAs, the proportion of acetate was significantly higher in the LyJH307 group than in the CON group (P = 0.0362), whereas that of propionate was significantly lower in the LyJH307 group (P = 0.0379), thereby increasing the acetate to propionate ratio (A:P ratio) in LyJH307 (P = 0.0291). However, endolysin LyJH307 supplementation did not affect the proportion of butyrate, iso-butyrate, valerate, and iso-valerate (butyrate, P = 0.2435; iso-butyrate, P = 0.3502; valerate, P = 0.2752; and iso-valerate, P = 0.3502). Endolysin LyJH307 supplementation significantly increased only the D-lactate concentration compared to the CON group (P = 0.0340) and total lactate concentration was high in the LyJH307 group (P = 0.0824).
Table 2. Gas production and rumen fermentation parameters by endolysin LyJH307 supplementation in an in vitro experiment at 12 h of incubation
|
Treatments1)
|
|
|
Items*
|
CON
|
LyJH307
|
SEM
|
P-value
|
Gas 6 h, mL/g DM
|
52.2
|
50.0
|
1.10
|
0.1216
|
Gas 9 h, mL/g DM
|
87.3
|
82.7
|
1.91
|
0.1121
|
Gas 12 h, mL/g DM
|
106.1
|
103.0
|
2.21
|
0.2533
|
IVDMD, %
|
40.06
|
38.73
|
1.1631
|
0.3597
|
NH3-N, mg/100 mL
|
7.87
|
9.21
|
0.553
|
0.1485
|
pH
|
5.57
|
5.69
|
0.033
|
0.0335
|
Total VFA, mM
|
83.12
|
89.11
|
4.388
|
0.2592
|
Acetate, mmol/mol
|
427.50
|
448.30
|
6.359
|
0.0362
|
Propionate, mmol/mol
|
341.50
|
325.20
|
4.422
|
0.0379
|
Butyrate, mmol/mol
|
199.90
|
195.30
|
3.017
|
0.2435
|
Iso-butyrate, mmol/mol
|
1.96
|
2.16
|
0.171
|
0.3502
|
Valerate, mmol/mol
|
22.21
|
20.94
|
0.802
|
0.2752
|
Iso-valerate, mmol/mol
|
6.96
|
8.12
|
1.726
|
0.6329
|
A:P ratio
|
1.25
|
1.38
|
0.033
|
0.0291
|
Total lactate, mM
|
16.76
|
18.89
|
0.872
|
0.0824
|
D-lactate, mM
|
9.74
|
11.26
|
0.404
|
0.0340
|
L-lactate, mM
|
7.02
|
7.63
|
0.642
|
0.4552
|
* SEM, standard error of the mean; DM, dry matter; IVDMD, in vitro dry matter digestibility; NH3-N, ammonia-nitrogen; VFA, volatile fatty acids; A:P ratio, acetate to propionate ratio.
1) CON, corn grain with elution buffer of the same volume of endolysin treatment; LyJH307, corn grain with endolysin LyJH307 (0.2% of dietary DM).
The absolute abundance of total bacteria and ciliate protozoa did not show any significant differences between treatments (Table 3, total bacteria, P = 0.6036; ciliate protozoa, P = 0.8271), whereas the absolute abundance of S. bovis was significantly decreased by endolysin LyJH307 supplementation (P < 0.0289).
Table 3. Microbial absolute abundances by endolysin LyJH307 supplementation in an in vitro experiment at 12 h of incubation
|
Treatments1)
|
|
|
Items*
|
CON
|
LyJH307
|
SEM
|
P-value
|
Total bacteria2)
|
3.82
|
3.68
|
0.231
|
0.6036
|
Ciliate protozoa2)
|
1.37
|
1.43
|
0.2063
|
0.8271
|
Streptococcus bovis3)
|
5.70a
|
1.76b
|
1.161
|
0.0289
|
* SEM, standard error of the mean
1) CON, corn grain with elution buffer of the same volume of endolysin treatment; LyJH307, corn grain with endolysin LyJH307 (0.2% of dietary DM).
2) ×1010 copies/mL of rumen fluid.
3) ×108 copies/mL of rumen fluid.
3.2. Effect of endolysin LyJH307 on the rumen microbiota
In the present study, 260,866 sequences were obtained by the 16S rRNA sequence analysis with an average of 43,477 ± 12,838 sequences per sample. Through quality filtering by QIIME2 (Q score > 20), 148,786 sequences (57% of the raw reads) were generated with an average of 24,797 ± 9,238 sequences per sample. There were no significant differences in the alpha diversity measurements, including Shannon’s index, Simpson’s index, Chao1 estimates, and evenness (Figure 1, Shannon’s index, P = 0.5127; Simpson’s index, P = 0.5127; Chao1 estimates, P = 0.2752; and evenness, P = 0.1266)). For the beta diversity, we presented two PCoA results based on Bray-Curtis dissimilarity and weighted UniFrac distance, followed by PERMANOVA analysis (Figure 2). The PCoA result based on Bray-Curtis dissimilarity showed statistical tendency between the CON and LyJH307 groups (Figure 2A, P= 0.0970), whereas there was no significant change in the PCoA result based on the weighted UniFrac distances (Figure 2B, P = 0.5030).
In the Venn diagram made by the major microbiome data, 16 of 17 phyla were shared, and 57 of 63 families were shared between the CON and LyJH307 groups. At the phylum and family levels, specific taxa were observed only in the LyJH307 group (Figure 3A and 3B). In addition, 114 of 129 genera were shared, and specific taxa in the LyJH307 group were higher than those in the CON group (Figure 3C).
The predominant taxa with relative abundance above 0.5% in at least one group are shown in Figure 4. At the phylum level, eight phyla had a relative abundance > 0.5%, and Bacteroidota (CON, 60.9% vs. LyJH307, 56.7%), Firmicutes (CON, 27.2% vs. LyJH307, 27.2%), and Proteobacteria (CON, 7.3% vs. LyJH307, 11.2%) were the three dominant phyla (Figure 4A), while the other phyla had a relatively minor proportion (both treatments, relative abundance < 1%). A total of 19 families were the dominant taxa that had a relative abundance of > 0.5% (Figure 4B), and there were no significant differences in major families between the CON and LyJH307 groups. At the genus level, the three dominant genera in the CON group were Prevotellaceae (41.1%), Rikenellaceae (11.1%), and Lachnospiraceae (8.3%), whereas the three dominant genera in the LyJH307 group were Prevotellaceae (39.8%), Succinivibrionaceae (10.7%), and Rikenellaceae (9.3%) (Figure 4C).
The differentially abundant taxa at the phylum, family, and genus levels between the CON and LyJH307 groups were identified using LEfSe (Figure 5). At the phylum level, the relative abundance of Elusimicrobiota and Verrucomicrobiota were significantly higher in the LyJH307 group than in the CON group (Figure 5). At the family level, LyJH307 supplementation significantly increased the relative abundance of three genera within Verrucomicrobiota (WCHB1-41, vadinBE97, and Victivallaceae), one genus within Firmicutes (Lactobacillaceae), and Desulfuromonadaceae (Figure 5B). At the genus level, a total of 11 genera including Lachnoclostridium, WCHB1-41, unclassified genus Selenomonadaceae (UG_Selenomonadaceae), Paraprevotella, vadinBE97, Ruminococcusgauvreauii group, Lactobacillus, Anaerorhabdusfurcosa group, Victivallaceae, Desulfuromonadaceae, and Sediminispirochaeta had significantly higher relative abundance in the LyJH307 group than in the CON group; thus, there were no genera enriched in the CON group (Figure 5B). Among the differentially abundant genera, several genera were only detected in the LyJH307 group, including Ruminococcusgauvreauii group, Anaerorhabdusfurcosa group, Victivallaceae, Desulfuromonadaceae, and Sediminispirochaeta.
3.3. Effect of endolysin LyJH307 on the predicted functions of the microbiota
To identify the functional changes of ruminal bacteria metabolism by LyJH307 supplementation, the functional composition profiles were predicted from 16S rRNA gene sequencing data with PICRUSt2, and the various functional features were predicted from the 16S rRNA gene using five different reference databases (KEGG, Pfam, COG, enzyme classification, and MetaCyc) (Table 4). LyJH307 affected the overall increase in the numbers of total functional features (KEGG orthologs, P = 0.0559; KEGG modules, P = 0.0535; Pfam, P = 0.0302; COG, P = 0.0566; and MetaCyc pathways P = 0.0969) at least showing statistical tendency (Table 4). However, the overall predicted functional features in the KEGG pathways were enriched in the CON group, with 14 KEGG pathways (other glycan degradation, ko00511; alanine, aspartate, and glutamate metabolism, ko00250; one carbon pool by folate, ko00670; amino sugar and nucleotide sugar metabolism, ko00520; zeatin biosynthesis, ko00908; biosynthesis of siderophore group nonribosomal peptides, ko01053; galactose metabolism, ko00052; glycosaminoglycan degradation, ko00531; protein digestion and absorption, ko04974; streptomycin biosynthesis, ko00521; fructose and mannose metabolism, ko00051; starch and sucrose metabolism, ko00500; pyrimidine metabolism, ko00240; and inositol phosphate metabolism, ko00562), except for biotin metabolism (ko00780) that was higher in the LyJH307 group (Figure 6A). In the KEGG modules, the predicted functional features were found to be enriched differently between the CON and LyJH307 groups (Figure 6B). In the CON group, keratan sulfate degradation (M00079), lipopolysaccharide transport system (M00250), ATP synthase (M00164), glycolysis (M00001), F-type ATPase (M00157), ascorbate biosynthesis (M00114), and aminoacyl-tRNA biosynthesis (M00359) were higher in the KEGG modules, whereas assimilatory sulfur reduction (M00176), PTS system (M00276), fatty acid biosynthesis (M00082), iron (Ⅲ) transport system (M00190), proline biosynthesis (M00015), biotin biosynthesis (M00123), and ADP-L-glycero-D-manno-heptose biosynthesis (M00064) were higher in the LyJH307 group (Figure 6B). Despite the changes in the low hierarchical level (KEGG pathways and modules), only two features, the biosynthesis of other secondary metabolites (P = 0.0809) and nucleotide metabolism (P = 0.0809), decreased by endolysin LyJH307 supplementation among the top 20 KEGG level 2 (Supplemental Table 1).
Table 4. Number of predicted functional features including KEGG hierarchies (orthologs, modules, and pathways), enzyme classification, MetaCyc pathways, Pfam, and clusters of orthologous genes by LyJH307 supplementation in an in vitro experiment at 12 h of incubation
|
Treatments1)
|
|
|
Items*
|
CON
|
LyJH307
|
SEM
|
P-value
|
KEGG orthologs
|
5137
|
5354
|
60.281
|
0.0559
|
KEGG modules
|
270
|
277
|
2.082
|
0.0535
|
KEGG pathways
|
137
|
142
|
3.215
|
0.2381
|
Pfam
|
5954
|
6254
|
75.508
|
0.0302
|
COG
|
3887
|
4007
|
36.226
|
0.0566
|
EC
|
1603
|
1656
|
21.759
|
0.137
|
MetaCyc pathways
|
328
|
339
|
3.756
|
0.0969
|
* SEM, standard error of the mean; KEGG, Kyoto Encyclopedia of Genes and Genomes; COG, clusters of orthologous genes; EC, enzyme classification.
1) CON, corn grain with elution buffer of the same volume of endolysin treatment; LyJH307, corn grain with recombinant LyJH307 (0.2% of dietary DM).
3.4. Correlations between microbial taxa and ruminal fermentation characteristics
Strong correlations (|r| > 0.8, P < 0.05) between ruminal fermentation characteristics and differently abundant taxa analyzed by LEfSe were detected (Figure 7). Lactobacillaceae, Desulfuromonadaceae (family), Ruminococcusgauvreauii group, Lactobacillus, and Desulfuromonadaceae (genus), all of which were higher in the LyJH307 group than in the CON group, were positively correlated with ruminal pH. In the individual VFA proportions, Elusimicrobiota, Desulfuromonadaceae (family), Lachnoclostridium, unclassified genus (UG)_Selenomonadaceae, Ruminococcusgauvreauii group, and Desulfuromonadaceae (genus) were all increased in the LyJH307 group and were positively correlated with the acetate proportion and negatively correlated with that of propionate. Verrucomicrobiota, WCHB1-41, Lactobacillaceae, Lachnoclostridium, WCHB1-41, UG_Selenomonadaceae, Paraprevotella, and Lactobacillus were positively correlated with D-lactate concentration.