Multispecies biofilm architecture determines bacterial exposure to phages

Abstract

Numerous ecological interactions among microbes-for example, competition for space and resources, or interaction among phages and their bacterial hosts-are likely to occur simultaneously in multispecies biofilm communities. While biofilms formed by just a single species occur, multispecies biofilms are thought to be more typical of microbial communities in the natural environment. Previous work has shown that multispecies biofilms can increase, decrease, or have no measurable impact on phage exposure of a host bacterium living alongside another species that the phages cannot target. The reasons underlying this variability are not well understood, and how phage-host encounters change within multispecies biofilms remains mostly unexplored at the cellular spatial scale. Here, we study how the cellular scale architecture of model 2-species biofilms impacts cell-cell and cell-phage interactions controlling larger scale population and community dynamics. Our system consists of dual culture biofilms of Escherichia coli and Vibrio cholerae under exposure to T7 phages, which we study using microfluidic culture, high-resolution confocal microscopy imaging, and detailed image analysis. As shown previously, sufficiently mature biofilms of E. coli can protect themselves from phage exposure via their curli matrix. Before this stage of biofilm structural maturity, E. coli is highly susceptible to phages; however, we show that these bacteria can gain lasting protection against phage exposure if they have become embedded in the bottom layers of highly packed groups of V. cholerae in co-culture. This protection, in turn, is dependent on the cell packing architecture controlled by V. cholerae biofilm matrix secretion. In this manner, E. coli cells that are otherwise susceptible to phage-mediated killing can survive phage exposure in the absence of de novo resistance evolution. While co-culture biofilm formation with V. cholerae can confer phage protection to E. coli, it comes at the cost of competing with V. cholerae and a disruption of normal curli-mediated protection for E. coli even in dual species biofilms grown over long time scales. This work highlights the critical importance of studying multispecies biofilm architecture and its influence on the community dynamics of bacteria and phages.

Fig 1. E. coli embedded within V. cholerae cell groups can evade exposure to phages in the surrounding medium.

(A) Time-lapse series of a dual culture biofilm of E. coli (yellow) and V. cholerae (purple), undergoing T7 phage exposure (infected E. coli cells reporting in cyan/white). The biofilm was grown for 48 h prior to continuous phage introduction thereafter. Time points noted in the upper right of each panel represent time since phage introduction was started. (B) The neighborhood biovolume fraction (biovolume fraction within a 6 μm around each segmented bacterium) of the merged biovolumes of both V. cholerae and E. coli for the first time point in panel A. (C) Mean V. cholerae fluorescence signal found around E. coli cells in biofilms with and without phage exposure (Mann–Whitney U test with n = 9). (D) E. coli biovolume normalized to biovolume prior to the introduction of phage in dual culture with V. cholerae and monoculture controls (Mann–Whitney U test with n = 9, n = 3). (E) Total biovolume of E. coli in dual culture and monoculture control biofilms with and without phage exposure at equivalent time points (Mann–Whitney U tests with n = 8, n = 8, n = 6, n = 7 from left to right). The data underlying this figure can be found in S1 Data.

Fig 2. E. coli evasion of phages within V. cholerae biofilms depends on the high cell–cell packing produced by WT V. cholerae.

(A) E. coli biovolume over time in dual culture conditions with either V. cholerae WT or V. cholerae ΔrbmA (n = 7, n = 8, n = 3–6, n = 3–8 from top to bottom in legend). (B) Average distance between E. coli cells and either V. cholerae WT or ΔrbmA in a triculture condition with or without phage exposure (Wilcoxon paired comparison tests with n = 9). (C, D) Representative images from the triculture condition with E. coli (yellow), V. cholerae WT (purple), and V. cholerae ΔrbmA (cyan) (C) without phage exposure and (D) after phage exposure. (E) PFU recovered after incubation of starting T7 phage inoculum with either no bacteria, V. cholerae, E. coli WT, or E. coli ΔtrxA over a 60-min time course. E. coli ΔtrxA allows for T7 phage attachment and genome ejection, but not for phage replication. Each trajectory shows the data for 1 run of each treatment (n = 3 for each treatment, giving 3 traces per treatment). (F) The neighborhood biovolume fraction of the merged biovolumes of both V. cholerae genotypes and E. coli from panel (D). The data underlying this figure can be found in S1 Data. PFU, plaque-forming unit; WT, wild type.

Fig 3. E. coli biofilms’ normal production of curli matrix protein is interrupted in co-culture with V. cholerae to the extent that phage protection is no longer provided by E. coli biofilm matrix.

(A) E. coli biovolume normalized to biovolume prior to phage introduction in dual culture and monoculture conditions for both E. coli WT and E. coli ΔcsgA (Mann–Whitney U tests with n = 7, n = 12). (B) Total E. coli biovolume with and without phage treatments at equivalent time points (Mann–Whitney U tests with n = 12). In these experiments, in contrast with Fig 1E, biofilms were grown for longer periods before phage addition such that E. coli WT on its own could produce protective curli matrix prior to phage addition. (C) Frequency distribution of csgBAC transcriptional reporter fluorescence around E. coli in monoculture and dual culture conditions. (D) Frequency distribution of curli immunofluorescence intensity in proximity to E. coli in monoculture and dual culture conditions. (E) Dual culture conditions of E. coli (yellow) and V. cholerae (purple) before phage exposure (top) and after 16 h of continuous phage exposure (bottom). (F) Monoculture conditions of E. coli before phage exposure (top) and after phage exposure (bottom). The data underlying this figure can be found in S1 Data.

Fig 4. Population dynamics of E. coli (yellow) and V. cholerae (purple) in monoculture and dual culture conditions, where biofilms grew for 48 h prior to phage exposure, and phage exposure was applied continuously for 96 h thereafter.

(A, B) Representative images from time course imaging of (A) E. coli monoculture and (B) co-culture with V. cholerae. (C, D) Magnification of E. coli phage infection (reporting in cyan/white) within a cluster embedded in a larger colony of V. cholerae at (C) 96 h and (D) 120 h. Expanded fields of view in (C) and (D) are denoted by checked white boxes in panel B. (E) E. coli population dynamics in monoculture and in co-culture with V. cholerae, with and without T7 phage exposure from 48 h onward (n = 7, n = 8, n = 6–7, n = 6–8 from top to bottom in the legend). Note that the data in the gray circles and squares through 120 h are repeated from the gray data in Fig 2A. (F) V. cholerae population dynamics in monoculture and in co-culture with E. coli, with and without T7 phage exposure from 48 h onward (n = 4, n = 4, n = 4–8, n = 4–8 from top to bottom in the legend). The data underlying this figure can be found in S1 Data.