Competition among Nasal Bacteria Suggests a Role for Siderophore-Mediated Interactions in Shaping the Human Nasal Microbiota

Abstract

Resources available in the human nasal cavity are limited. Therefore, to successfully colonize the nasal cavity, bacteria must compete for scarce nutrients. Competition may occur directly through interference (e.g., antibiotics) or indirectly by nutrient sequestration. To investigate the nature of nasal bacterial competition, we performed coculture inhibition assays between nasal Actinobacteria and Staphylococcus spp. We found that isolates of coagulase-negative staphylococci (CoNS) were sensitive to growth inhibition by Actinobacteria but that Staphylococcus aureus isolates were resistant to inhibition. Among Actinobacteria, we observed that Corynebacterium spp. were variable in their ability to inhibit CoNS. We sequenced the genomes of 10 Corynebacterium species isolates, including 3 Corynebacterium propinquum isolates that strongly inhibited CoNS and 7 other Corynebacterium species isolates that only weakly inhibited CoNS. Using a comparative genomics approach, we found that the C. propinquum genomes were enriched in genes for iron acquisition and harbored a biosynthetic gene cluster (BGC) for siderophore production, absent in the noninhibitory Corynebacterium species genomes. Using a chrome azurol S assay, we confirmed that C. propinquum produced siderophores. We demonstrated that iron supplementation rescued CoNS from inhibition by C. propinquum, suggesting that inhibition was due to iron restriction through siderophore production. Through comparative metabolomics and molecular networking, we identified the siderophore produced by C. propinquum as dehydroxynocardamine. Finally, we confirmed that the dehydroxynocardamine BGC is expressed in vivo by analyzing human nasal metatranscriptomes from the NIH Human Microbiome Project. Together, our results suggest that bacteria produce siderophores to compete for limited available iron in the nasal cavity and improve their fitness.IMPORTANCE Within the nasal cavity, interference competition through antimicrobial production is prevalent. For instance, nasal Staphylococcus species strains can inhibit the growth of other bacteria through the production of nonribosomal peptides and ribosomally synthesized and posttranslationally modified peptides. In contrast, bacteria engaging in exploitation competition modify the external environment to prevent competitors from growing, usually by hindering access to or depleting essential nutrients. As the nasal cavity is a nutrient-limited environment, we hypothesized that exploitation competition occurs in this system. We determined that Corynebacterium propinquum produces an iron-chelating siderophore, and this iron-sequestering molecule correlates with the ability to inhibit the growth of coagulase-negative staphylococci. Furthermore, we found that the genes required for siderophore production are expressed in vivo Thus, although siderophore production by bacteria is often considered a virulence trait, our work indicates that bacteria may produce siderophores to compete for limited iron in the human nasal cavity.

Keywords: Actinobacteria; Corynebacterium; Staphylococcus; competition; dehydroxynocardamine; iron; nasal microbiome; siderophore.

FIG 1

Inhibition of nasal Staphylococcus spp. by nasal Actinobacteria. (A) Representative images of monocultures of Staphylococcus and cocultures of Actinobacteria (left) with Staphylococcus species isolates (right). The inhibition scoring system is depicted below the coculture images. (B) Twenty-one nasal Actinobacteria isolates (horizontal) were monocultured on BHI agar wells for 1 week before 39 Staphylococcus species isolates (vertical) were spotted adjacent to the Actinobacteria colony. The colonies were cultured together for 1 week before inhibition of the Staphylococcus species colony was scored. The heat map displays the inhibition scores of each Staphylococcus species isolate when paired with the corresponding Actinobacteria isolate. Each interaction was technically replicated at least twice. The gray cells indicate interactions where replicates were in disagreement or the Actinobacteria colony overgrew the well before Staphylococcus inoculation. (Left) Phylogenetic tree of Staphylococcus species isolates built from 685 bp of the 16S rRNA gene amplified with the universal primers 27F and 1492R. (Top) Core-genome phylogenetic tree of Actinobacteria built from 93 conserved, single-copy genes. Both phylogenies are rooted on B. subtilis 168, and nodes with ≥50% bootstrap support are indicated. The dashed yellow lines highlight strong inhibition of CoNS by one clade of Corynebacterium spp. Strains in boldface type are siderophore producers as measured by the CAS assay. Actinobacteria taxa marked with ● harbor a siderophore BGC within their genome.

FIG 2

Comparative genomics of nasal Corynebacterium spp. (A, left) Core-genome phylogeny of nasal Corynebacterium from Fig. 1. The other Actinobacteria node is collapsed, as represented by a triangle. The C. propinquum clade is shaded in orange. (Right) Clustered presence-absence matrix of 1,504 orthologs predicted across the 10 Corynebacterium species genomes. The 1,198 orthologs that were conserved across all 10 genomes were excluded. Orthologs encoding functions pertaining to iron metabolism, iron transport, or siderophore biosynthesis are depicted in red. (B) KO term enrichment analysis of the orthologs uniquely shared among C. propinquum strains relative to all orthologs encoded within all 10 Corynebacterium species genomes. The odds ratios of all 12 significantly enriched (P < 0.05) KO terms are plotted. Note that odds ratios of infinity (Inf) correspond to KO terms where all orthologs that were annotated with the KO term were present in the enrichment set. KO terms in boldface type are related to iron transport or siderophore biosynthesis. The color of each point indicates the significance level, and the size of the point indicates how many of the orthologs were annotated with the KO term.

FIG 3

Siderophore production by C. propinquum inhibits CoNS. (A) CAS assay results from C. propinquum HSID18034 (left) and C. genitalium HSID17239 (right), which strongly and weakly inhibit CoNS, respectively. A shift in the color of the CAS overlay from blue to yellow indicates siderophore production. (B) Coculture plate inhibition between the strains in panel A and CoNS on BHI medium (−Iron) and BHI medium supplemented with 200 μM FeCl₃ (+Iron). (C) Coculture plate inhibition assays with C. propinquum (HSID18034 and HSID18036) and 5 Corynebacterium species non-siderophore producers (HSID17231, HSID17239, HSID17260, HSID17564, and HSID17575) against 22 CoNS strains on BHI medium (−Iron) and BHI medium supplemented with 200 μM FeCl₃ (+Iron). Each pairing was duplicated with consistent results. The error bars represent the standard errors of the means. The photographs in panels A and B are representative of results for duplicate samples.

FIG 4

Corynebacterium propinquum produces the siderophore dehydroxynocardamine. (A) The dehydroxynocardamine BGC from C. propinquum. ORFs encoding biosynthetic and transport-related functions are filled with gray and white, respectively. See Table 1 for specific ORF annotation. The DtxR sequence motif is shown upstream of dnoB. (B) Subset of the molecular network of C. propinquum HSID18034 and C. genitalium HSID17239 agar core extracts. Gray nodes are metabolites shared by both strains, and black nodes are metabolites unique to C. propinquum. There were no detected metabolites that were unique to C. genitalium. The edges are weighted to the cosine score between the two features. A single cluster unique to C. propinquum is outlined in a dashed box. See Fig. S2 in the supplemental material for the full molecular network. (C) Zoomed-in view of the single cluster unique to C. propinquum with the nodes labeled by their corresponding m/z. This cluster contains single- and double-charged states of dehydroxynocardamine, dehydroxynocardamine B, and dehydroxynocardamine C. (D) Structures of dehydroxynocardamine, dehydroxynocardamine B, and dehydroxynocardamine C.

FIG 5

The dehydroxynocardamine BGC is expressed in vivo. Ninety-five nasal metatranscriptomes from 16 subjects in the NIH HMP were pseudoaligned to the genome of C. propinquum HSID18034 to detect the expression of the dehydroxynocardamine BGC (dnoA-G) (white) and the housekeeping genes gyrB, rpoB, sigA, and rpsL (gray). The upper and lower bounds of the notched box plot indicate the 75th and 25th percentiles, respectively. The bars indicate the medians, and the notches represent the 95% confidence intervals of the medians. Each point indicates the expression of a single gene from a metatranscriptome in transcripts per kilobase per million (TPM).