Peeling back the many layers of competitive exclusion

Abstract

Baby chicks administered a fecal transplant from adult chickens are resistant to Salmonella colonization by competitive exclusion. A two-pronged approach was used to investigate the mechanism of this process. First, Salmonella response to an exclusive (Salmonella competitive exclusion product, Aviguard^®) or permissive microbial community (chicken cecal contents from colonized birds containing 7.85 Log₁₀Salmonella genomes/gram) was assessed ex vivo using a S. typhimurium reporter strain with fluorescent YFP and CFP gene fusions to rrn and hilA operon, respectively. Second, cecal transcriptome analysis was used to assess the cecal communities' response to Salmonella in chickens with low (≤5.85 Log₁₀ genomes/g) or high (≥6.00 Log₁₀ genomes/g) Salmonella colonization. The ex vivo experiment revealed a reduction in Salmonella growth and hilA expression following co-culture with the exclusive community. The exclusive community also repressed Salmonella's SPI-1 virulence genes and LPS modification, while the anti-virulence/inflammatory gene avrA was upregulated. Salmonella transcriptome analysis revealed significant metabolic disparities in Salmonella grown with the two different communities. Propanediol utilization and vitamin B12 synthesis were central to Salmonella metabolism co-cultured with either community, and mutations in propanediol and vitamin B12 metabolism altered Salmonella growth in the exclusive community. There were significant differences in the cecal community's stress response to Salmonella colonization. Cecal community transcripts indicated that antimicrobials were central to the type of stress response detected in the low Salmonella abundance community, suggesting antagonism involved in Salmonella exclusion. This study indicates complex community interactions that modulate Salmonella metabolism and pathogenic behavior and reduce growth through antagonism may be key to exclusion.

Keywords: Salmonella; antimicrobials; attenuation; competition; exclusion; pathogen.

Figure 1

Experimental approach to revealing the mechanism of action of competitive exclusion. A two-prong approach involved determining cecal community metabolism relative to Salmonella abundance in vivo (A) and Salmonella expression in response to a permissive or exclusive community ex vivo (B). In the in vivo study, 2-day-old, specific pathogen-free white leghorn chickens were orally administered Salmonella Typhimurium SL1344 (1 × 10⁵ CFU). At different ages, chickens were sacrificed, the ceca were aseptically collected from each bird, and nucleic acids were extracted. Salmonella invA qPCR was used to determine Salmonella abundance in the cecal community samples. The cecal community RNA was sequenced and annotated using the MG-RAST pipeline. The ex vivo study focused on Salmonella response when grown microaerophilically in a simulated cecal medium, Ex Vivo Cecal Contents (EVCC), containing a permissive or exclusive community. A fluorescent S. typhimurium SL1344 rrn promoter::yfp, hilA operon::cfp reporter strain was placed within dialysis tubing, physically separate from the permissive or exclusive community, enabling the collection of Salmonella cells for analysis. Jellyfish green fluorescent protein variants YFP and CFP served as reporters for Salmonella growth and SPI-1 invasion gene reporter, respectively. Fluorescence was measured using fluorescence-activated cell sorting. The cecal community from a 35-day-old chicken with high Salmonella abundance served as the permissive community, while the exclusive community consisted of the cecal community comprising the competitive exclusion product Aviguard^®. Salmonella grown in EVCC alone served as a control. Select Salmonella metabolic genes, upregulated in either or both communities, were deleted in a S. typhimurium SL1344 fluorescent reporter strain, and mutants were compared to the wild-type reporter strain for growth in the exclusive community.

Figure 2

A rrn promoter jellyfish yellow fluorescent protein (YFP) gene fusion as a reporter for Salmonella growth. Lambda red was used to construct rrn tagged with YFP. The rrn promoter was cloned upstream of a “promoter-less” yfp and in tandem with cat. PCR primers were designed to overlap with the target insertion site, an intergenic site between thrW and STM0324. The λ Red system (Datsenko and Wanner, 2000) was used to introduce this reporter into S. enterica Typhimurium SL1344. S. typhimurium SL1344 strain YC1104 with the rrn-yfp reporter was grown aerobically to OD 1.0, λ 600 nm, in LB broth (A) or M9 minimal medium with 0.4% glycerol as a carbon source (B) with aeration and observed with a fluorescence microscope. (C) Fluorescence-activated cell sorting (FACS) analysis of the Salmonella rrn-yfp reporter strain grown aerobically, in LB broth to mid-exponential phase (OD 0.5, λ 600 nm) at 37°C or 42°C to demonstrate fluorescence intensity with a high growth rate.

Figure 3

The Salmonella ex vivo growth in permissive (B) or exclusive (C) communities. FACS analysis was used to measure the rrn-yfp promoter activity in S. typhimurium SL1344 grown in simulated cecal medium (EVCC) (A) within a permissive or exclusive community. Salmonella YFP expression was monitored by FACS analysis at 3, 6, 9, 12, and 24 h using the parental strain as a negative fluorescence control. Cyan (3 h), blue (6 h), green (9 h), and red (24 h) lines mark the peak fluorescence levels of the reporter strain grown alone in EVCC (A) in order to demonstrate changes in intensity when grown in a permissive (B) or exclusive (C) community.

Figure 4

A hilA operon-cfp promoter fusion using jellyfish cyan fluorescent protein (CFP) as a reporter for Salmonella pathogenicity. A “promoter-less” cfp was placed between iag and the transcriptional terminator for the hilA operon using λ Red recombineering (Datsenko and Wanner, 2000). The Salmonella reporter strain was grown in EVCC medium alone or EVCC medium inoculated with the permissive or exclusive cecal community at (A) 3 h or (B) 6-h incubation. FACS analysis was used to measure the hilA operon-cfp promoter fusion in S. typhimurium SL1344 grown in EVCC medium alone (red line) or with a permissive (blue line) or exclusive (green line) community. The S. typhimurium SL1344 parental strain (orange line) served as a negative fluorescence control in FACS analysis. The CFP plasmid vector, pMG34, was used as a fluorescence-positive control (cyan line) (Miyashiro and Goulian, 2007). While a fluorescent population of cells were detected for Salmonella grown in EVCC alone or in a permissive community, an exclusive community repressed the Salmonella hilA locus, which is a global regulator of the pathogenicity island 1 (SPI-1) associated type 3 secretion system (T3SS).

Figure 5

Differences in the Salmonella response to cecal permissive (A) and exclusive (B) communities ex vivo. The Salmonella SL1344 rrn-yfp, hilA operon-cfp reporter strain was grown in EVCC alone or with a permissive or exclusive community. Total RNA was extracted from Salmonella after 6-h growth. Genes reported with elevated expression and significant differences in expression (p < 0.0001) for growth in the permissive or exclusive community versus the negative community control were categorized based on function.

Figure 6

Transcriptional analysis of Salmonella virulence expression after growth in permissive or exclusive communities. FACS analysis was used to measure the hilA operon-cfp promoter fusion expression in S. typhimurium SL1344 grown in EVCC or with permissive or exclusive communities. Because Cfp expression was diminished in the Salmonella reporter strain after 6 h of growth in the exclusive community, the reporter strain was harvested after 6 h for microarray analysis. Significant transcription differences (*p < 0.0001) were observed in the SPI-1 locus, its ancillary T3SS effectors sopE2, pipA-D, and sopB, the type I fimbriae operon, flagellin, and polymyxin resistance expression. A heat map scale depicts the magnitude of microarray signal.

Figure 7

Details of ex vivo Salmonella transcriptional response to an exclusive community. Salmonella SL1344 rrn-yfp, hilA operon-cfp reporter strain was grown in co-culture in EVCC with an exclusive community or media alone. Total RNA was extracted from Salmonella after 6-h growth. Genes reported with significant differences in expression (p < 0.0001) versus the medium control were categorized based on function.

Figure 8

Differences in the Salmonella response to cecal permissive or exclusive communities ex vivo. The Salmonella SL1344 rrn-yfp, hilA operon-cfp reporter strain was grown in EVCC alone or with a permissive or exclusive community. Total RNA was extracted from Salmonella after 6-h growth. Genes reported with significant differences in expression (p < 0.0001) for both communities versus the negative community control were categorized based on function.

Figure 9

A Bayesian network analysis of cecal transcriptome, focused on fermentation, from chickens with high (>5.85 Log₁₀ Salmonella genomes/g cecal contents) (A) or low (B) Salmonella abundance. The network depicted was identified using the score-based learning algorithm Hill-Climbing. The data were obtained from KO metabolism data sets in MG-RAST. Connections were identified among the 45 enzyme transcripts by network analysis, of which 36 were identified as single or multiple connections, producing 20 and 10 connections from the transcriptomes of cecal communities with high or low Salmonella abundance, respectively. Arrows indicate the direction and strength of the connection. Enzymes are color-coded based on their associated propanediol (green), butyrate (blue), or acetate (yellow) fermentation pathways. Enzymes identified in these networks were: methyl-malonyl-CoA decarboxylase (1); methylglyoxal synthase (2); lactaldehyde dehydrogenase (3); hydroxyacylglutathione hydrolase (4); glycerol dehydrogenase (5); glycerol dehydratase (6); propanediol dehydratase (7); propionate kinase (8); propionyl-CoA carboxylase (9); propionate-CoA transferase (10); methyl-malonyl-CoA mutase (11); propionyl-CoA synthase (12); propionaldehyde dehydrogenase (13); butyryl-CoA dehydrogenase (14); glutaconate-CoA transferase (15); hydroxybutyryl-CoA dehydratase (16); hydroxybutyryl-CoA dehydrogenase (17); thiolase (18); ferredoxin hydrogenase (19); pyruvate oxidase (20); acetyl-CoA synthase (21); acetyl-CoA hydrolase (22); acetate kinase (23); pyruvate ferredoxin oxidoreductase (24); acetyl-CoA synthetase (25); aldehyde dehydrogenase (26); phosphotransacetylase (27); lactate dehydrogenase (28); phosphoketolase (29); glutamate synthase (30); methylaspartate ammonia-lyase (31); ethanolamine ammonia-lyase (32); ethanolamine utilization protein (33); 4-aminobutyrate aminotransferase (34); cobalamin (vitamin B12) (35); and cobalamin biosynthesis (36). The following enzymes were not identified in any network analysis: lactaldehyde reductase, butyrate kinase, pyruvate dehydrogenase, citramalate synthase, serine dehydratase, alanine dehydrogenase, formate hydrogenlyase, formate C-acetyltransferase, and glutamate mutase. Numbers in bold are enzymes shared between cecal communities with high and low Salmonella abundance. Circles with thick-colored borders are for enzymatic transcripts with 98–100% nucleotide identity and 99–100% coverage by BLAST scores for intestinal species including Escherichia coli, Enterococcus faecium, and Flavonifractor plautii.

Figure 10

The effect of mutations in one or more metabolic pathways involved in Salmonella catabolism in an ex vivo intestinal environment containing an exclusive community. λ Red was used to create single and double mutations in Salmonella SL1344. pGLOW, a fluorescent reporter that fluoresces under aerobic and anaerobic conditions, was used to monitor Salmonella growth over 48 h. Mutant and wild-type (WT*) S. typhimurium SL1344 strains were grown in EVCC with exclusive community; WT was grown in EVCC alone as a control. Oxyrase (Sigma-Aldrich; St. Louis, MO) was added to the medium and overlaid with mineral oil to create and maintain low oxygen conditions. The selected mutations affect the catabolism of: (A) microbial metabolites (eut, pdu, prp), vitamin B12 synthesis (cbi), vitamin B12 uptake (btuC), or (B) sugar utilization (fucI, mgl) or anaerobic respiration (nar, nir).

Figure 11

A Bayesian network analysis of the cecal transcriptome, focused on stress response, from chickens with high (>5.85 Log₁₀ Salmonella genomes/g) (A) or low (B) Salmonella abundance. The network depicted was constructed using the score-based learning algorithm Hill-Climbing (arrows with solid line). Arrows indicate the direction and strength of the connection. Enzymes are color-coded based on the stress response. Enzymes or genes identified in the high Salmonella abundance community (A) are: (1) organic hydroperoxide resistance protein (ohr); (2) universal stress protein E (uspE); (3) yciT; (4) cspI; (5) pspC; (6) rseA; (7) arcA; (8) glutaredoxin; (9) nrd; (10) yncG; (11) fmrR; (12) betI; (13) yehZ; (14) ybaT; (15) hdeA; (16) propanediol diffusion facilitator; (17) glutamate-1-semialdehyde aminotransferase. Enzymes or genes identified in the low Salmonella abundance community (B) are: (1) yersiniabactin non-ribosomal peptide synthase; (2) putative bacterial hemoglobin; (3) gshF; (4) hemW; (5) tehB; (6) thioesterase domains of type I polyketides. Circles with thick red, green, or blue borders are for enzymatic transcripts with 98–100% nucleotide identity, by BLAST scores, to Escherichia coli, Enterococcus spp. or both, respectively.