Characterization of CP activity against B. thailandensis in vitro.

(A) Phylogenomic tree of CPs (yellow) and model pathogen 8. thailandensis (Bt) (red), interspersed amongst publicly available genomes (no color). (B) Schematic of competition assay used to test in vitro inhibitory capabilities. Percent growth of Bt is calculated as the CFU in co-culture at 24 hours divided by the CFU in the Bt monoculture at 24 hours multiplied by 100. Co-culture inhibition was tested for multiple starting densities of CP (see Figure 1C x-axis). Starting density of Bf and CP at a 1:1 ratio are 3.33×104 CFU/ml per organism. Bt monoculture is inoculated at 3.33×104 CFU/ml. (C) Dose-response curves for CPs against Bt. Relative IC50 is denoted with a vertical dashed-dotted line, and 100% and 50% growth are denoted with horizontal dotted lines. Percent growth at each density is represented as mean± SEM.

Summary of pathogen inhibition by CPs.

Relative IC50 values and 95% confidence intervals (95% CI) are shown for each inhibitory CP. Not determined (n.d.) indicates that a value could not be measured. CPs are listed in order of decreasing NI values.

Production of a specialized metabolite plays a role in the antagonistic activities of CPS.

(A) Schematic representation of supernatant inhibition experiment workflow. Supernatant from a Bf and CP co-culture or from Bf monoculture (supernatant) is mixed 1:1 with 2x LSM. This mixture is used as the growth medium for subsequent Bf growth curves. Fold-change in Bt growth is calculated by dividing the maximum growth of Bt in co-culture supernatant by the maximum growth of Bt grown in supernatant from itself. (B) Left: Heatmap of fold-change in max CFU of Bt grown in supernatant from each inhibitory CP and Bt co-culture relative to Bt grown in supernatant from itself. Co-culture supernatant was collected at 72 hours. A fold-change of 1 indicates no difference in growth in co-culture supernatant compared to Bt growth in supernatant from itself. A fold-change less than 1 indicates a reduction in pathogen growth. Right: Boxplot of growth inhibition by CPS supernatant. Data are represented as mean± SD. (C) Agar diffusion assays testing the CPS wildtype strain (CPS WT) and sfp mutant strain (CP8Δsfp) for inhibition of Bt after 48 hours. (D) Growth curves of CP8 WT (black) and CP8b.sfp (red) (E) Dose-response curve of CP8 WT (black) and CP8b.sfp (red) against Bt after 24 hours of co-culture. Percent growth at each density is represented as mean ± SEM.

Metabolic niche overlap is indicative of pathogen and CP antagonism.

(A) Dose-response curve for CP7 grown in unsupplemented LSM (black) or with additional glucose (red). Percent growth at each density is represented as mean ± SEM. (B) Activity of CP7 against Bt at a 1:100 ratio with or without glucose supplementation (red and gray, respectively). Data are summarized as mean ± SD. ***P<0.001. (C) Carbon utilization heatmap for CPs and Bt. (D) Venn diagram of carbon sources utilized by CP7 and Bt. (E) Plot of rank percent Bt growth at a 1:1 ratio after 24 hours vs rank of Niche Index for each CP. Correlation is determined by Spearman rank correlation (r = −0.6367, P < 0.0001, 95% Cl [−0.7891 −0.4261], N = 49). (F) Total carbon consumed by CPs grown in combination with Bt at 36 hours (mean± SD). Activity indicates relative amount of Bt inhibition at a 1:1 ratio from most(+++) to least(+) inhibitory. *P=0.05; **P<0.01. (G) Total electron flux at 24 hours. Red dashed line indicates the estimated abundance of Bt in each co-culture. (H) Summed electron flux of lactate (black), proline (white), and aspartic acid (gray) at 24 hours in different CP co-culture conditions (mean± SD). *P<0.05.(I) Receiver operator characteristic curve showing the ability of growth on lactate to distinguish inhibitory CPs from non-inhibitory CPs. An AUROC of 1 indicates perfect distinction between inhibitors and non-inhibitors; an AUROC of 0.5 (dashed red line) indicates no distinction between inhibitors and non-inhibitors.

Niche overlap aids in identification of efficacious CP combinations.

(A) Venn diagram representation of the Niche Index Fraction (NIF) calculation. Yellow represents the fraction numerator and red the denominator. (B) Heatmap of NIF values for each colonizing CP. (C) Dose-response curve for the CP8/CP19 combination (mean ± SEM). (D) Dose-response curve for the CP19/CP13 combination (mean ± SEM). (E) Dose-response curve for the CP13/CPB combination (mean± SEM). (F) Lactate consumed (as determined via GC-TOF) for 3 conditions, normalized to CFU of each CP inoculated (mean ± SD). ****P<0.0001. (G) Dose-response curve for the CP8/CP19 (red) and CP8sfp/CP19Δsfp (black) combinations. (H) Schematic of the proposed mechanism for Bt inhibition by the CP8/CP19 combination. CP8 produces a specialized metabolite that inhibits growth of the Bt. Additionally, increased lactate consumption by CP19 in the presence of CP8 further reduces Bt growth. Overall, minimal niche overlap between the CPs, and high overlap with Bt, allows for further inhibition of Bt with minimal inter-CP antagonism.

CPs colonize the mouse airway and protect against respiratory Bt infection.

(A) Schematic of colonization testing for CPs. Each CP (106 CFU) is administered to the lower airway via OPA. After 7 days, the airway tissues (lungs and trachea) are collected, homogenized, and plated in order to enumerate viable CPs (expressed as CFU per tissue homogenate) (B) Colonization results from airway tissues collected at 7 days following CP administration. The dotted line represents the minimum CFU at which the CP is considered able to colonize in a reliably detectable manner. Data are represented as the mean ± SD. (C) Schematic of the method for survival testing. Each CP (106 CFU) is administered to the lower airway via OPA at 3, 5 or 7 days post-GP a normally lethal dose of of Bt (3×104 - 5×105 CFU) is administered to the lower airway via OPA. Survival is measured over the course of 10 days post-Bt. (D) Survival data from mice administered no CP (vehicle control) and challenged with Bt (E) Survival data from mice administered CPs at 3 days (blue), 5 days (orange), or 7 days (red) prior to Bt challenge. The Niche Index value for each CP is listed in the lower-right corner of its associated graph. Comparison of survival rate with versus without CP treatment was accomplished using the Mantel-Cox test ****P<0.0001, ***P<0.0002, **P<0.01. A summary of the relationship between days surviving and Niche Index value for each dosing schedule is also shown (F) Pathogen loads in airway tissues of mice prophylactically treated with CPs. Each CP or PBS (“no CP” negative control) is administered to the lower airway via OPA, 3 days later Bl (3×104 - 5×105 CFU) is similarly administered, and 3 days later (i.e., 3 days post-challenge, 6 days post-treatment) the airway tissues are collected for enumeration of Bl CFU. Mean Bt CFU counts are as follows for each group: CP19 (1.6×103), CP17 (1.7×103), CP20 (2.2×106), CP26 (8.7×106), CP13 (1.3×107), CPS (3.9×108), No CP (4.6×108). Asterisks indicate significant differences in Bt CFU between treatment versus no treatment groups as determined using Dunn’s multiple comparisons test ****P<0.0001, ***P<0.0002, **P<0.01. The relationship between the Niche Index values of CPs and the Bl CFU detected in mice treated with the CPs is indicated in the upper left corner.

Proposed mechanism of CP protection.

Left: During lower airway infection, bacterial growth elicits inflammation from the host, with concurrent endothelial and epithelial injury. Intra-alveolar oedema introduces additional nutrients from the blood into the alveoli and lung lumen, enabling further growth of the pathogen. This positive feedback loop results in uncontrolled growth of the pathogen, ultimately facilitating its instantiation. Right: Probiotics delivered directly to the lower airway may limit infection by reducing nutrient abundance after oedema, which prevents further growth of the pathogen and thereby breaks the infection-promoting feedback loop.

Growth curves in LSM and inhibition of CPs by Bt.

(A) CPs and Bt were individually inoculated into LSM at 106 CFU/ml, and the cultures were incubated at 37C with shaking (278rpm) for 48 h. (B) Co-culture data measuring CFUs of each CP after co-culture with Bt (+) at a 1:1 ratio or each CP grown alone (-) in LSM. ****P<0.0001; ***P=0.0002; **P=0.0021

Plaque assays to detect phage isolated from CPs.

Supernatants from phage-induced CP cultures were serially diluted 1:10 and spotted onto soft agar containing Bt. Supernatant from a phage-induced culture of Bt strain E421, which produces a Bt-targeting phage, generated visible plaques as expected, demonstrating that the phage detection method was effective.

Addition of glucose decreases antagonism between CP19 and Bt.

Dose-response curve of Bt in co-culture with CP19 in standard LSM (5.5mM glucose) (black) or LSM supplemented to 43mM glucose (red). Vertical dotted lines indicate the IC50 in the low or high glucose condition. Data are represented as the mean ± SEM.

Dose-response curves for pairwise CP combinations with low Niche Index Fractions

(A) IC50 curve for the CP20/CP17 combination (red diamonds). Single strain inhibition curves (black) are shown for CP20 (circles) and CP17 (triangles). (B) IC50 curve for the CP26/CP19 combination (red diamonds). Single strain inhibition curves (black) are shown for CP26 (circles) and CP19 (triangles).

qPCR generated growth curves for CP8, CP19 and Bt grown in co-culture.

CP8, CP19 and Bt were co-cultured in LSM for 36 hours at 37C and 278rpm. Samples were periodically taken to quantify their OD using qPCR. Three biological replicates of co-cultures were made and two technical replicates of each co-culture analyzed. All 6 points for each organism at each timepoint are displayed. Bt growth curve is denoted as black open circles, CP8 as blue closed circles and CP19 as purple closed circles.

CP8 colonization of the airway is significantly reduced in Bt-infected mice.

CP8 (106 CFU) was administered to the airway via OPA and after 3 days challenged with either PBS (CP8 alone) or Bt (3×104 - 5×105 CFU) (CP8 + Bt) via OPA. Airway tissues were recovered at 3 days post-challenge (6 days post-CP8 administration) for enumeration of CP8 load (expressed as CFU per tissue homogenate). **P<0.002.

Survival of mice treated with non-viable CPs prior to pathogen challenge.

CPs were rendered non-viable (via UV and heat treatment) and administered to the mouse airway (via OPA) at 3 days (blue), 5 days (orange), or 7 days (red) prior to challenge with Bt (via OPA). Survival was monitored for 10 days following pathogen challenge. For CP20, CP26, CP13, and CP8 treatment at 3 days prior to pathogen challenge was the only dosing regimen investigated because these CPs provided little to no protection when viable bacteria were administered at 5 days or 7 days prior to pathogen challenge (see Fig. 5E). Asterisks indicate significant differences between survival of mice treated with PBS (negative control) versus the non-viable CP, whereas plus signs indicate significant differences between survival of mice treated with the non-viable CP versus the viable CP, as calculated using the Holm-Šidák multiple comparisons test (****or ++++P<0.0001, ***or +++P<0.0002, ** or ++P<0.0021, * or +P<0.0332).

CP19 consumption of lactate and inhibition of pathogen growth in mouse airway tissues ex vivo.

(A) Lactate levels in airway tissue homogenates treated with PBS (-CP19) or CP19 (106 CFU) (+CP19) for 24 hours. ****P<0.0001. (B) Growth of Bt in airway tissue homogenates pre-treated with PBS (-CP19) or CP19 (106 CFU) (+CP19) for 24 hours. Pre-treated tissue homogenates were filter sterilized, the filtrates were inoculated with Bt (106 CFU), and OD600 measurements were taken every 15 minutes to monitor growth of the pathogen over the course of 12 hours.

Representation of CP families in mouse airway microbiome profiles.

Family-level phylogenetic assignments of the CPs were compared to those of mouse airway microbiome constituents previously identified in four independent 16S rRNA profiling efforts21–25. An identical analysis was performed for “background” (B) strains (contaminants recovered from liver and spleen). For each CP and B family, listed are: 1) The number of times that it is represented in a given airway microbiome profile (“# Hits”); 2) Its relative abundance in the profile (“% Hits”); and 3) Its relative rank within the profile, with the most abundant family in the profile designated as “Rank” = 1. For each profile, the total number of 16S rRNA sequences placed at the family level, and the total number of families represented, are listed the headings of the “# Hits” and “Rank” columns, respectively.

Representation of CP genera in mouse airway microbiome profiles.

Genus-level phylogenetic assignments of the CPs, and of the B strains, were compared to those of mouse airway microbiome constituents previously identified in four independent 16S rRNA profiling efforts21–25. For each CP and B genus, listed are: 1) The number of times it is represented in a given airway microbiome profile (“# Hits”); 2) Its relative abundance in the profile (“% Hits”); and 3) Its relative rank within the profile, with the most abundant genus in the profile designated as “Rank” = 1. For each profile, the total number of 16S rRNA sequences placed at the genus level, and the total number of genera represented, are listed the headings of the “# Hits” and “Rank” columns, respectively.

Representation of CP species in mouse airway microbiome profiles.

Species-level phylogenetic assignments of the CPs, and of the B strains, were compared to those of mouse airway microbiome constituents previously identified in four independent 16S rRNA profiling efforts21–25. For each CP and B species, listed are: 1) The number of times it is represented in a given airway microbiome profile (“# Hits”); 2) Its relative abundance in the profile (“% Hits”); and 3) Its relative rank within the profile, with the most abundant species in the profile designated as “Rank” = 1. For each profile, the total number of 16S rRNA sequences placed at the species level, and the total number of species represented, are listed the headings of the “# Hits” and “Rank” columns, respectively.

Relative IC50 and Niche Index Fraction values for each CP combination tested.

Relative IC50 values are reported as IC50 (95% confidence interval). Non-inhibitory combinations have relative IC50 values listed as n.a. (not applicable).