Dynamics of plasmid-mediated niche invasion, immunity to invasion, and pheromone-inducible conjugation in the murine gastrointestinal tract

Abstract

Microbial communities provide protection to their hosts by resisting pathogenic invasion. Microbial residents of a host often exclude subsequent colonizers, but this protection is not well understood. The Enterococcus faecalis plasmid pCF10, whose conjugative transfer functions are induced by a peptide pheromone, efficiently transfers in the intestinal tract of mice. Here we show that an invading donor strain established in the gastrointestinal tract of mice harboring resident recipients, resulting in a stable, mixed population comprised of approximately 10% donors and 90% recipients. We also show that the plasmid-encoded surface protein PrgB (Aggregation Substance), enhanced donor invasion of resident recipients, and resistance of resident donors to invasion by recipients. Imaging of the gastrointestinal mucosa of mice infected with differentially labeled recipients and donors revealed pheromone induction within microcolonies harboring both strains in close proximity, suggesting that adherent microcolonies on the mucosal surface of the intestine comprise an important niche for cell-cell signaling and plasmid transfer.

Fig. 1. Colonization, niche invasion, and plasmid transfer in mice mono-associated with E. faecalis donor/recipient communities.

As described in Methods, germ-free mice were initially gavaged with either wild-type recipients (OG1ES) (A, C) or donors, either OG1Sp:pCF10 (B) or OG1Sp:pCF10ΔprgB (D) to establish a GI resident population. Three days later, the mice were gavaged with the opposite cell type, which comprised the invading strain. The populations of recipients, donors, and transconjugants in feces were enumerated by plating on selective media at various time points from 1–35 days after the addition of the invading strain. Each plotted point represents results from a single mouse. A subset of mice from these experiments was also sacrificed at selected time points and bacterial associated with different sections of the intestine were enumerated, as shown in Supplementary Fig. 2. OG1ES is marked in blue, OG1Sp in red. The presence of plasmid pCF10 is indicated by a triangle (∆), the plasmid pCF10ΔprgB is represented by a diamond shape (◊). Statistical analysis for the data presented in Figs. 1 and 2 were performed using Student’s t test with Welch’s correction using GraphPad (v. 9) and R (v. 3.53) software. The number of input animals for each experiment are as follows: A- 26 (7 cages); B- 14 (4 cages); C- 8 (2 cages); D- 8 (2 cages). Detailed enumeration data are listed in the Source Data File.

Fig. 2. Role of strain background in niche invasion, invasion immunity, and pCF10 transfer.

Experiments analogous to those shown in Fig. 1 were carried out (same enumeration methods and data analysis): A, B show the plasmid-free strain OG1ES and OG1Sp in the resident or invader roles. C, D are analogous to the experiments shown above in Fig. 1A, B, but with OG1ES as pCF10 donor strain and OG1Sp as plasmid-free recipient. OG1ES is marked in blue, OG1Sp in red; the presence of plasmid pCF10 is indicated by a triangle (∆). Numbers of mice (cages): A- 8 (2); B- 8(2): C- 12(3): D- 8(3). Detailed Enumeration data are listed in the Source Data File.

Fig. 3. Success as an invader is dependent on strain background, pCF10, and PrgB.

The invasion index (II) is defined as the cell number of the invading strain divided by the cell number of the resident strain in feces (the same primary data used to generate Figs. 1 and 2), at either day 1 or Day 21. Statistical analysis was performed by ANOVA, followed by tests for normal distribution and Mann–Whitney ranked test two-tailed for pair-wise comparison. P values for the population differences between the time points are as follows (invader/resident): ES:pCF10/Sp- <0.001, ES/SpΔprgB- 0.001, ES/Sp- 0.004, ES/Sp:pCF10- <0.001, Sp:pCF10/ES- <0.0001, SpΔprgB/ES- 0.181, Sp/ES- 0.004, Sp/ES:pCF10- 0.505. Detailed statistical analysis is presented in the Source Data File.

Fig. 4. IFM imaging of Recipient (R) and Donor (D) interactions in the murine GI tract.

The cell wall of the recipient strain (OG1ES) was pre-labeled with the fluorescent d-amino acid HADA (blue). Donors (OG1Sp) constitutively express the tdTomato fluorescent protein (orange-red) and, when driven by the presence of nearby recipients, inductively expressed GFP (green); for convenience, the color scheme for identification of different cell types is also depicted in the lower-left corner of the figure. Gnotobiotic mice were gavaged with either Recipient first (OG1ES; A–E) or Donors first (OG1Sp:pCF10; F–I). After 72 hours, the other strain (D for the R first mice; R for the D first mice) was gavaged and, 6 hours post gavage, the mice were killed. The images shown are representative of nine animals. Donors invade recipients. A Villi and microvilli (top) can be seen in this orthogonal projection of a portion of the proximal ileum. B At higher magnification, numerous ~1 μm structures labeled in blue (arrows; blue circles) can be seen attached and clustered on the epithelial surface (not all cells are labeled to allow for unrestricted viewing). In addition, some orange-red labeled cells are clustered nearby (red circles). C In a matched sample from a different mouse, clusters of blue cells (recipients; HADA) can be seen on the epithelium with a small cluster of yellow cells (red circles). Separating the fluorescent channels into Red+Blue (D) and Green only (E), reveals that the yellow cells in C are composed of cells labeled both red and green, consistent with Donor cells (red) adjacent to Recipient cells (blue) expressing the inducible GFP from the pCF10 plasmid. Notably, there is also at least one Donor cell in this field (D: white single outline) that is not expressing GFP (E, white single outline), suggesting it has not yet been induced by the presence of neighboring Recipients. Recipients invade donors F Numerous donors (red) can be seen in this proximal ileum sample with many fewer recipients (blue; blue arrows) present. G A matched sample from another mouse provides a similar field of view: many donors (red) and a few recipients (blue; blue arrows). H High magnifications show similar clusters of donors and recipients as seen in C–E. Although limited in number, I appears to show at least one donor (red) that has been induced (red circle) and is beginning to express the inducible GFP from pCF10.

Fig. 5. Models for pCF10-mediated conjugation and invasion of resident populations in the mouse intestine.

A Donors invade recipients Donor cells (blue) in the lumen can attach to recipient microcolonies (red), or directly to the mucosal surface. Because of the high pheromone C concentration in the colony, donors that attach are induced to transfer plasmids (black circles) to adjacent recipients. The spread of the plasmid through the microcolony is limited by the gradually increasing levels of I resulting from plasmid dissemination and from the pheromone-induced expression of the prgQ gene encoding the I peptide; this generates a colony comprised of plasmid-containing and plasmid-free cells. B Cells and small aggregates from the colony likely re-enter the planktonic phase by direct exit or when the colony is disrupted by turnover of the mucosal surface, and re-attach to generate new microcolonies. C Recipients invade donors when recipients invade a niche containing resident donors, the high pre-existing concentration of I coupled with the low population density of the C-producing invaders minimizes the expression of transfer functions. The rare transconjugants generated probably result from contact-dependent pheromone induction of donors adjacent to the attached recipient. Unless the invading recipients have a plasmid-independent fitness advantage, the transconjugants have no advantage over the resident donors, and they are eventually eliminated, along with the unmated recipients are eliminated. As in B. cells and small aggregates from the colonies may re-enter the lumen and then re-attach, but this does not prevent the eventual elimination of the invaders.