Insertion sequence-mediated phage resistance contributes to attenuated colonization of cytolytic Enterococcus faecalis variants in the gut

Abstract

Specific elimination of cytolytic Enterococcus faecalis from the intestinal microbiota by bacteriophages (phages) attenuates ethanol-induced liver disease in pre-clinical studies; however, other clinical phage therapy studies have reported the occurrence of phage-resistant variants. Here, we assessed phage resistance using a cytolytic E. faecalis clinical isolate, EF01. After infecting EF01 with ΦEf2.1 (Myoviridae) or ΦEf2.2 (Podoviridae), four host variants (R-EF01^ΦEf2.1-A and R-EF01^ΦEf2.1-B from infection with ΦEf2.1, and R-EF01^ΦEf2.2-A and R-EF01^ΦEf2.2-B from infection with ΦEf2.2) were isolated. Although isolate R-EF01^ΦEf2.2 exhibited resistance to both phages, isolate R-EF01^ΦEf2.1 demonstrated partial resistance only to ΦEf2.1. Whole-genome sequencing of these four isolates revealed an insertion sequence, IS256, -mediated disruption of xylA in R-EF01^ΦEf2.1-A and R-EF01^ΦEf2.1-B. In addition, a non-synonymous mutation in epaR, essential for the complete Enterococcus polysaccharide antigen (Epa), was identified in the R-EF01^ΦEf2.2-A isolate. Furthermore, R-EF01^ΦEf2.2 isolates exhibited IS256-associated chromosomal deletions and lacked galE, a gene involved in Epa biosynthesis. After gavaging mice with EF01 WT, R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates, colonization of R-EF01^ΦEf2.2 isolates was significantly attenuated. R-EF01^ΦEf2.2 isolates exhibited less resistance to the bile salt sodium deoxycholate and showed reduced adherence to intestinal cell monolayers, suggesting that phage-resistant variants with alterations in bacterial surface molecules, potentially including those involved in Epa biosynthesis, reduced pathogen fitness by attenuating gut colonization. In summary, IS256 is involved in phage resistance of a cytolytic E. faecalis clinical isolate, and certain phage resistance mechanisms could contribute to favorable clinical outcomes by promoting the swift elimination of phage-resistant variants in the treatment of alcohol-associated hepatitis.

Importance: Phage therapy is a promising approach for precise editing of the gut microbiota. Notably, the specific elimination of cytolytic E. faecalis from the intestinal microbiota by phages attenuates ethanol-induced liver disease in pre-clinical studies. Despite the great promise of phage therapy, the occurrence of phage-resistant variants represents a concern for the successful development of phage-based therapies. In this context, we assessed phage resistance using a cytolytic E. faecalis clinical isolate. Isolated phage-resistant variants acquired mutations or deletions of Epa biosynthesis-related genes and exhibited attenuated colonization in the gut. These phage-resistant variants showed less resistance to bile salts and reduced adherence to intestinal cell monolayers. These results suggest that even if phage-resistant variants arise during phage therapy, certain mechanisms of phage resistance may contribute to the rapid elimination of phage-resistant variants promoting favorable clinical outcomes in the treatment of alcohol-associated hepatitis.

Keywords: bacteriophage; cytolysin; fitness cost; gut-liver axis; microbiome editing; phage therapy; trade-off.

Fig 1

Infection of cytolytic E. faecalis strain EF01 with phages and isolation of EF01 variants. (A) Schematic representation of liquid-based infection of EF01 with ΦEf2.1 and ΦEf2.2. Each group was conducted in duplicate. (B) OD₆₀₀ values of each independent replicate every 24 h. (C) Schematic representation of the isolation of EF01 variants, designated as R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.1-B, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates. (D) OD₆₀₀ values of EF01 R-EF01^ΦEf2.1 and R-EF01^ΦEf2.2 cultures 24 h post-inoculation (hpi). (E) Colony-forming activity of EF01 R-EF01^ΦEf2.1 and R-EF01^ΦEf2.2 cultures 24 hpi. Values are presented as mean ± SD (n = 3). Significance against EF01 WT was analyzed by Dunn’s test based on One-way ANOVA analysis: ****P < 0.0001. (F) Killing activity of ΦEf2.1 and ΦEf2.2 to EF01 R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.1-B, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates compared to EF01 WT. Efficiency of plating (EoP) values represent phage infectivity against tested strains compared with the plaque-forming activity of phages against the EF01 WT strain.

Fig 2

Disruption of xylA by IS256 in phage-resistant variants, R-EF01^ΦEf2.1. (A) Short reads from R-EF01^ΦEf2.1-A and R-EF01^ΦEf2.1-B, mapped against the parental strain EF01 WT visualized using CLC Genomics Workbench. The left panel shows the R-EF01^ΦEf2.1-A isolate, and the right panel shows the R-EF01^ΦEf2.1-B isolate. Insertion sequences were detected around the xylA regions. (B) Detection of xylA and xylB regions in EF01 WT, R-EF01^ΦEf2.1-A, and R-EF01^ΦEf2.1-B isolates by PCR using primer pairs recognizing the upstream region of xylA and downstream region of xylB. (C) Genome sequences around the xylA regions of R-EF01^ΦEf2.1-A and R-EF01^ΦEf2.1-B isolates. Insertion sequences were identified by the IS finder.

Fig 3

Characterization of Epa- and Epa-related genes in phage-resistant variants, R-EF01^ΦEf2.2. Schematic explanation of chromosomal deletions in R-EF01^ΦEf2.2-A (A) and R-EF01^ΦEf2.2-B (B) variants. F1 and R1 represent the primer pair for long amplicon sequencing using the R-EF01^ΦEf2.2-A genome, and F2 and R2 represent the primer pair for long amplicon sequencing using the R-EF01^ΦEf2.2-B genome. Red lines indicate deleted sequences in phage-resistant variants. (C) Detection of galE in EF01 WT, R-EF01^ΦEf2.2-A ,and R-EF01^ΦEf2.2-B isolates by PCR. poll was used as a control. (D) Schematic representation of Epa biosynthesis, particularly for rhamnan-decoration. epaR plays a fundamental role in transferring α-glucose to the core. αGlcNAc is converted from αGalNAc by galE and catalyzed by epaX and epaS. αGlu: α-glucose, Rha: rhamnose. (E) Gradient gel electrophoresis of extracted Epa from EF01 WT, R-EF01^ΦEf2.1, and R-EF01^ΦEf2.2 isolates. (F) Adsorption rate of ΦEf2.1 (left panel) and ΦEf2.2 (right panel) against EF01 WT. Data are presented as mean ± SD (n = 4). (G) Adsorption rate of ΦEf2.1 and ΦEf2.2 against R-EF01^ΦEf2.1 and R-EF01^ΦEf2.2 isolates compared to EF01 WT after 10 min of adsorption time. Data are presented as mean ± SD (n = 4). Significance against EF01 WT was analyzed by Dunn’s test based on one-way ANOVA analysis: ****P < 0.0001.

Fig 4

Colonization of phage-resistant variants in the gut. (A) The levels of colony-forming activity in fecal samples harvested at 24, 48, and 72 h after gavaging conventional mice with EF01 WT, R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B transformed with pSL101P_16S Each colony was detected on a BHI-spc agar plate (500 µg/mL). Values are presented as mean ± SD (n = 7) for each group. Significance against EF01 WT was analyzed by Dunn’s test based on one-way ANOVA analysis: *P < 0.05, **P < 0.01. (B) The proportion of mice that tested positive for cylL_S in qPCR analysis using fecal DNA. n = 13 for each group. Significance against EF01 WT was analyzed by Fisher’s exact test: *P < 0.05.

Fig 5

Adherence and bile salt resistance of phage-resistant variants. (A) Representative fluorescent microscopic imaging of adherent EF01 WT, R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates toward Caco2 cells. “Mock” indicates that no bacteria were inoculated into the cells. (B) The levels of adherent EF01 WT, R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates determined by counting colonies on a BHI-agar plate after washing the cells and detaching bacterial strains. Data are presented as mean ± SD (n = 8). Significance against EF01 WT was analyzed by Dunn’s test based on one-way ANOVA analysis: *P < 0.05, **P < 0.01. (C) The colony-forming activity of EF01 WT, R-EF01^ΦEf2.1-A, R-EF01^ΦEf2.2-A, and R-EF01^ΦEf2.2-B isolates on BHI-agar plates containing 0%, 0.04%, 0.06%, and 0.08% sodium deoxycholate. Values are presented as mean ± SD (n = 4). Significance against EF01 WT was analyzed by Dunn’s test based on two-way ANOVA analysis: *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig 6

Schematic explanation of fitness trade-off between phage sensitivity and gut colonization. (Left) Complete Epa in the parental E. faecalis strain provides cell wall integrity, offering protection from bile salts and enabling adhesion to intestinal epithelial cells, thereby contributing to gut colonization. Epa also serves as the primary receptor for Enterococcus phages. (Right) Phage-resistant variants resulting from changes in the bacterial surface structure, potentially including Epa biosynthesis defects, block Enterococcus phage infection; however, these variants simultaneously lose bile salt resistance and adhesion to intestinal epithelial cells as a fitness cost for gaining phage resistance, which would accelerate the elimination of the pathobiont from the gut microbiota even if phage resistance develops. Such alterations represent a key fitness trade-off in phage therapy targeting bacterial microbiota in the gut.