Sociality in Escherichia coli: Enterochelin Is a Private Good at Low Cell Density and Can Be Shared at High Cell Density

Abstract

Many bacteria produce secreted iron chelators called siderophores, which can be shared among cells with specific siderophore uptake systems regardless of whether the cell produces siderophores. Sharing secreted products allows freeloading, where individuals use resources without bearing the cost of production. Here we show that the Escherichia coli siderophore enterochelin is not evenly shared between producers and nonproducers. Wild-type Escherichia coli grows well in low-iron minimal medium, and an isogenic enterochelin synthesis mutant (ΔentF) grows very poorly. The enterochelin mutant grows well in low-iron medium supplemented with enterochelin. At high cell densities the ΔentF mutant can compete equally with the wild type in low-iron medium. At low cell densities the ΔentF mutant cannot compete. Furthermore, the growth rate of the wild type is unaffected by cell density. The wild type grows well in low-iron medium even at very low starting densities. Our experiments support a model where at least some enterochelin remains associated with the cells that produce it, and the cell-associated enterochelin enables iron acquisition even at very low cell density. Enterochelin that is not retained by producing cells at low density is lost to dilution. At high cell densities, cell-free enterochelin can accumulate and be shared by all cells in the group. Partial privatization is a solution to the problem of iron acquisition in low-iron, low-cell-density habitats. Cell-free enterochelin allows for iron scavenging at a distance at higher population densities. Our findings shed light on the conditions under which freeloaders might benefit from enterochelin uptake systems.

Importance: Sociality in microbes has become a topic of great interest. One facet of sociality is the sharing of secreted products, such as the iron-scavenging siderophores. We present evidence that the Escherichia coli siderophore enterochelin is relatively inexpensive to produce and is partially privatized such that it can be efficiently shared only at high producer cell densities. At low cell densities, cell-free enterochelin is scarce and only enterochelin producers are able to grow in low-iron medium. Because freely shared products can be exploited by freeloaders, this partial privatization may help explain how enterochelin production is stabilized in E. coli and may provide insight into when enterochelin is available for freeloaders.

FIG 1

Growth curves of wild-type E. coli BW25113 grown with (open squares) or without (closed squares) added Ent and the ΔentF mutant grown with (open triangles) and without (closed triangles) added Ent. We used low-iron medium for these experiments and supplemented with 5 μg/ml of Ent where indicated. Cultures were run in duplicate, and ranges are within the symbols.

FIG 2

The ΔentF mutant can grow in low-iron medium when in coculture with the wild type. E. coli BW25113 arsB::kan (WT1) or ΔentF arsB::kan (ΔentF) was mixed at a 1:1 ratio with E. coli BW25113 arsB::cat (WT2). (A) Competitive index; (B) final cell density. The initial cell density was about 10⁵ cells per ml. Total cells per milliliter and strain frequencies were measured at the start of the experiment and after 14 h. The competitive index was calculated as described in Materials and Methods. Results from three independent experiments are shown, and the bars indicate the means.

FIG 3

Negative frequency-dependent fitness of the ΔentF mutant. Competition experiments were conducted with WT1 or ΔentF against WT2. Initial ratios ranged from 1,000:1 to 1:1,000, and initial cell densities were about 10⁶ cells per ml. Cells were grown in low-iron medium. (A) CI of competitions between the ΔentF mutant and WT2; (B) total yield of the cultures. At high frequencies of the ΔentF mutant, its fitness and total culture yield fall off. Controls with WT1 and WT2 show a CI of about 1 regardless of initial frequency, and total yields also remain constant. Frequencies were measured at the start of the experiment and after 14 to 16 h of growth. Also shown are growth of WT2 in culture with the ΔentF mutant (C) or with WT1 (D). Generations per hour were calculated as generations over the growth period divided by the hours of growth.

FIG 4

Fitness of the ΔentF mutant versus the wild type depends on initial cell density. The E. coli BW25113 arsB::cat strain was grown together with the E. coli BW25113 arsB::kan strain or the ΔentF mutant (E. coli ΔentF arsB::kan) in low-iron medium with or without supplements as indicated. Starting ratios were 1:1, and initial cell densities are shown. Blue bars represent the two Ent wild-type strains in unsupplemented low-iron medium. The CI was about 1 regardless of inoculum size. Black bars represent the ΔentF mutant and the wild type in unsupplemented medium. The ΔentF mutant is competitive only at initial cell densities of 10³ to 10⁴ or greater. Green bars represent the ΔentF mutant and wild type in medium supplemented with Ent (5 μg/ml). The mutant is competitive regardless of inoculum size. Cell numbers and frequencies were determined immediately after inoculation and at 14 h. The bars represent results from two independent experiments, and error bars show ranges.

FIG 5

Growth of wild-type E. coli in the presence of transferrin is dependent on inoculum size. E. coli BW25113 and the ΔentF mutant were grown in low-iron medium with human apotransferrin and sodium bicarbonate. Open circles represent the ΔentF mutant started from low density and closed circles the ΔentF mutant started from high density. Open triangles represent the wild type started from low density and closed triangles the wild type started from high density. Open diamonds represent the wild type started from low density with Ent added, and closed diamonds represent the wild type started from high density with Ent added. Cultures were run in duplicate, and ranges are within the symbols.

FIG 6

Growth rate of wild-type E. coli in low-iron medium is not affected by inoculum size. E. coli BW25113 was grown in low-iron medium starting at various densities as indicated (initial cell density). Cell numbers were determined by plate counting at the time of inoculation and at 18 h. In all cases, growth was in logarithmic phase at 18 h. The number of generations is the number of doublings in 18 h. Cultures were run in triplicate. One of the lowest-density starting cultures failed to grow, presumably because this culture did not receive any cells during inoculation; this culture was excluded from analysis.