Dolosigranulum pigrum Cooperation and Competition in Human Nasal Microbiota

Abstract

Multiple epidemiological studies identify Dolosigranulum pigrum as a candidate beneficial bacterium based on its positive association with health, including negative associations with nasal/nasopharyngeal colonization by the pathogenic species Staphylococcus aureus and Streptococcus pneumoniae Using a multipronged approach to gain new insights into D. pigrum function, we observed phenotypic interactions and predictions of genomic capacity that support the idea of a role for microbe-microbe interactions involving D. pigrum in shaping the composition of human nasal microbiota. We identified in vivo community-level and in vitro phenotypic cooperation by specific nasal Corynebacterium species. Also, D. pigrum inhibited S. aureus growth in vitro, whereas robust inhibition of S. pneumoniae required both D. pigrum and a nasal Corynebacterium together. D. pigrum l-lactic acid production was insufficient to account for these inhibitions. Genomic analysis of 11 strains revealed that D. pigrum has a small genome (average 1.86 Mb) and multiple predicted auxotrophies consistent with D. pigrum relying on its human host and on cocolonizing bacteria for key nutrients. Further, the accessory genome of D. pigrum harbored a diverse repertoire of biosynthetic gene clusters, some of which may have a role in microbe-microbe interactions. These new insights into D. pigrum's functions advance the field from compositional analysis to genomic and phenotypic experimentation on a potentially beneficial bacterial resident of the human upper respiratory tract and lay the foundation for future animal and clinical experiments.IMPORTANCEStaphylococcus aureus and Streptococcus pneumoniae infections cause significant morbidity and mortality in humans. For both, nasal colonization is a risk factor for infection. Studies of nasal microbiota identify Dolosigranulum pigrum as a benign bacterium present when adults are free of S. aureus or when children are free of S. pneumoniae Here, we validated these in vivo associations with functional assays. We found that D. pigrum inhibited S. aureusin vitro and, together with a specific nasal Corynebacterium species, also inhibited S. pneumoniae Furthermore, genomic analysis of D. pigrum indicated that it must obtain key nutrients from other nasal bacteria or from humans. These phenotypic interactions support the idea of a role for microbe-microbe interactions in shaping the composition of human nasal microbiota and implicate D. pigrum as a mutualist of humans. These findings support the feasibility of future development of microbe-targeted interventions to reshape nasal microbiota composition to exclude S. aureus and/or S. pneumoniae.

Keywords: Corynebacterium; Dolosigranulum pigrum; Staphylococcus aureus; Streptococcus pneumoniae; comparative genomics; interspecies interactions; microbe-microbe interactions; microbiota; nasal; upper respiratory tract.

FIG 1

Individual nasal Corynebacterium species exhibit increased differential relative abundances in the presence of D. pigrum in human nostril microbiota. We used ANCOM to compare the species/supraspecies-level compositions of 16S rRNA gene nostril data sets from (A) 99 children ages 6 to 78 months and (B) 210 adults where D. pigrum was either absent (Dpi−) or present (Dpi+) on the basis of 16S rRNA gene sequencing data. Plots show only the taxa identified as statistically significant (sig = 0.05) after correction for multiple testing within ANCOM. The dark bar represents the median; lower and upper hinges correspond to the first and third quartiles. Each gray dot represents the value for a sample, and multiple overlapping dots appear black. Dpi = Dolosigranulum pigrum, Cac = Corynebacterium accolens, Caa/Cma/Ctu = supraspecies Corynebacterium accolens_macginleyi_tuberculostearicum, Cpr = Corynebacterium propinquum, Cps = Corynebacterium pseudodiphtheriticum, Mno = Moraxella nonliquefaciens. Only three species and one supraspecies of Corynebacterium from among the larger number of Corynebacterium supraspecies/species present in each data set met the significance threshold. Specifically, in the adult nostril data set, there were 21 species and 5 supraspecies groupings of Corynebacterium in addition to the reads of Corynebacterium that were nonassigned (NA) at the species level. These data were previously published (see Table S7 in reference 41). In the pediatric data set, there were 16 species of Corynebacterium in addition to those that were nonassigned among the species-level Corynebacterium reads (see Table S2). The Log relative abundance numerical data represented in this figure are available in Table S1.

FIG 2

D. pigrum growth yields increase on cell-free conditioned agar medium (CFCAM) from nasal Corynebacterium species but not in reverse. (A and B) Growth yield of D. pigrum strains CDC2949-98, CDC4709-98, and KPL1914 was quantified as the number of CFU grown on a polycarbonate membrane placed on (A) cell-free conditioned BHI agar from C. propinquum (aqua green) or C. pseudodiphtheriticum (dark and light green) or (B) cell-free conditioned BHI-triolein (BHIT) agar from C. accolens (blue) and compared to growth on unconditioned BHI agar (dark gray) or unconditioned BHIT agar (light gray), respectively. (C) Growth yield of C. pseudodiphtheriticum KPL1989 on CFCAM from D. pigrum strains (orange) compared to unconditioned medium (white) was assessed similarly. BHIT was used for growth of C. accolens since it is a fatty acid auxotroph and releases needed oleic acid from triolein. Preconditioning strains were grown on a 0.2-μm-pore-size, 47-mm-diameter polycarbonate membrane for 2 days to generate CFCAM. After removal, we then placed a new membrane on the CFCAM onto which we spread 100 μl of target bacterial cells that had been resuspended to an OD₆₀₀ of 0.50 in 1× PBS. After 2 days of growth, CFU were enumerated as described in Materials and Methods. CFU counts were compared independently for each individual strain (A and B, n = 5) or medium (C, n = 4) using a Wilcoxon rank sum test with Bonferroni correction for multiple comparisons to the unconditioned medium. Dark bars represent medians, lower and upper hinges correspond to the first and third quartiles, and outlier points are displayed individually. *, P ≤ 0.05; **, P ≤ 0.001. CSBA, citrated sheep blood agar.

FIG 3

Ten different strains of D. pigrum inhibit methicillin-resistant S. aureus USA300 strain JE2 whereas S. aureus does not inhibit D. pigrum. (A) Ten pregrown D. pigrum isolates produced a diffusible activity that inhibited the growth of S. aureus strain JE2. (B) When S. aureus was pregrown in this assay, there was no visible inhibition of subsequently inoculated D. pigrum. (C) Similarly, we did not observe any inhibition in a pairwise comparison of three representative strains of D. pigrum in this assay. All growth was on BHI agar in the independent experiments (n ≥ 3) represented in panels A, B, and C. Representative images are shown for each strain. The respective pregrown strain (D. pigrum or S. aureus) was resuspended in PBS, and then a 5-μl drop was placed on BHI agar and pregrown for 48 h (D. pigrum) or 24 h (S. aureus). After that, the indicator strain was inoculated at a location adjacent to the pregrown strain. Inhibition was assessed after 24 and 48 h (48-h results are shown here). For panels A and B, similar results were observed using S. aureus Newman.

FIG 4

Lactate production by D. pigrum is insufficient to inhibit pathobiont growth. Strains of S. pneumoniae and S. aureus grew in the presence of higher levels of l-lactic acid than those produced by D. pigrum in vitro. (A) The concentration of l-lactic acid (mM) produced by three D. pigrum strains was measured after 24 h of gentle shaken aerobic growth in BHI broth at 37°C (n = 5) compared to the basal concentration of l-lactic acid in BHI medium alone (none). (B) The average growth (OD₆₀₀) of 4 S. pneumoniae strains in D. pigrum KPL1914 CFCM or in unconditioned BHI broth supplemented with different concentrations of l-lactic acid measured after 19 to 20 h of static aerobic growth at 37°C (n = 4). (C) The average growth (OD₆₀₀) of 2 S. aureus strains in D. pigrum KPL1914 CFCM or in unconditioned BHI broth supplemented with different concentrations of l-lactic acid measured after 19 to 20 h of shaken aerobic growth at 37°C (n = 4). In panels A and B, the average pH of the D. pigrum CFCM was 6.40 (±0.06). The pH of BHI without lactic acid (0 mM) was adjusted to match the pH of the CFCM, to control for any effect of pH alone. Average levels of growth of S. pneumoniae in CFCM and 11 mM l-lactic acid were analyzed independently for each individual strain using a Wilcoxon rank sum test. Dark bars represent medians, lower and upper hinges correspond to the first and third quartiles, and outlier points are displayed individually, except in panel A, where dots for all individual sample values are represented. *, none of the S. pneumoniae or S. aureus strains displayed growth in 55 mM l-lactate.

FIG 5

D. pigrum and C. pseudodiphtheriticum grown together but not D. pigrum alone inhibited S. pneumoniae in an in vitro agar medium-based assay. Representative images are shown of S. pneumoniae 603 growth on (A) BHI medium alone or on CFCAM from (B) C. pseudodiphtheriticum KPL1989, (C) D. pigrum KPL1914, or (D) both D. pigrum and C. pseudodiphtheriticum grown in a mixed inoculum (n = 4). To condition the medium, we cultivated D. pigrum and/or C. pseudodiphtheriticum on a membrane, which was then removed prior to spreading a lawn of S. pneumoniae. For monoculture, 100 μl of either D. pigrum or C. pseudodiphtheriticum, resuspended to an OD₆₀₀ = 0.50, were inoculated onto the membrane. For mixed coculture, 50 μl of D. pigrum (OD₆₀₀ = 0.50) were mixed with 50 μl of C. pseudodiphtheriticum (OD₆₀₀ = 0.50) to yield a final volume of 100 μl for the inoculum, such that each bacterial species was present in the coculture inoculum at half the amount used for the corresponding monoculture inoculum. Images were cropped. Black marks indicate edges where the membrane had been.