Solutions to the public goods dilemma in bacterial biofilms

Abstract

Bacteria frequently live in densely populated surface-bound communities, termed biofilms [1-4]. Biofilm-dwelling cells rely on secretion of extracellular substances to construct their communities and to capture nutrients from the environment [5]. Some secreted factors behave as cooperative public goods: they can be exploited by nonproducing cells [6-11]. The means by which public-good-producing bacteria avert exploitation in biofilm environments are largely unknown. Using experiments with Vibrio cholerae, which secretes extracellular enzymes to digest its primary food source, the solid polymer chitin, we show that the public goods dilemma may be solved by two very different mechanisms: cells can produce thick biofilms that confine the goods to producers, or fluid flow can remove soluble products of chitin digestion, denying access to nonproducers. Both processes are unified by limiting the distance over which enzyme-secreting cells provide benefits to neighbors, resulting in preferential benefit to nearby clonemates and allowing kin selection to favor public good production. Our results demonstrate new mechanisms by which the physical conditions of natural habitats can interact with bacterial physiology to promote the evolution of cooperation.

Figure 1

V. cholerae faces a public goods dilemma when digesting chitin

1. (A)

   Transcript levels of chiA-1 and chiA-2 in response to different substrates [27]. The black dashed line represents no change in expression relative to growth in glucose. Error bars are standard deviations.

2. (B)

   Wild type in competition with chitinase non-producers (ΔchiA-1,2) in liquid cultures supplemented with GlcNAc (red), and (GlcNAc)₂ (blue). Δf_producer is the change in frequency of the chitinase producer in the population after 10 cell divisions, and f_0,producer is the initial seeding frequency of the producers.

3. (C)

   Cell divisions in shaking cultures grown on solid chitin. Red and blue bars are data from pure cultures of ΔchiA-1,2 and wt, respectively, and striped bars represent co-cultures at different frequencies of wt. Error bars are the range of means from n = 4 independent replicas.

4. (D)

   Wild type in competition with chitinase non-producers (blue) and mutants that produce inactive chitinase proteins (red) in mixed liquid with chitin flakes. Δf_producer was measured after 36 h.

Figure 2

Solutions to the public goods dilemma

1. (A)

   Wild type in competition with chitinase non-producers in non-shaken liquid with chitin flakes. The two strains had either no additional mutations, or both carried identical additional mutations as indicated in the legend. ΔflaA mutants have no flagellum, ΔvpsL mutants cannot make the biofilm matrix, ΔvpvC^W240R mutants are matrix hyper-producers. The dashed line indicates the maximum Δf_producer for a given f_0,producer, i.e. fixation of the producers. The strains were initially seeded with 10³ CFU/mL. Error bars represent the range of means from n = 4 independent replicas.

2. (B)

   Wild type in competition with chitinase non-producers in microfluidic chambers subjected to a range of flow speeds. Error bars are standard deviations.

Figure 3

Thick biofilms localize public goods close to the producers

1. (A)

   The maximum biofilm thickness that developed over 180 h for different strains. Error bars are standard deviations.

2. (B)

   Model calculation of the flux of GlcNAc molecules that escape from the biofilm (J_out) as a function of biofilm thickness. J_out is normalized by J_in, the flux of GlcNAc molecules that enter the biofilm at the chitin surface due to the action of the chitinases. Evaluating the model with biofilm densities that are 10% and 1% of the confluent density yielded the black and green data points, respectively.

3. (C)

   Lower values of the normalized cross-correlation indicate higher spatial segregation of chitinase producers and non-producers. The cross-correlation was evaluated for competitions of wild type and ΔchiA-1,2 (blue), and for competitions in the vpvC^W240R background (magenta), 108 h after inoculation. Error bars are standard deviations.

4. (D)

   Confocal image of a population of chitinase producers (yellow) competing against ΔchiA-1,2 (red) on chitin (blue) in non-shaken liquid, initially seeded at f_0,producer = 0.5. Figure S3G shows this image at full resolution. There are three regions with significant growth, which are dominated by producers. The inset illustrates that in these regions, the chitinase producers exhibited the matrix hyper-producer phenotype (thick biofilms), due to spontaneous mutations, and they outcompeted the ΔchiA-1,2 mutant.

Figure 4

Flow removes public goods and minimizes access for non-producers

1. (A)

   Microscope image acquired for a competition of wt (yellow) versus ΔchiA-1,2 (red) without flow (Pe = 0) on solid chitin (blue). The competition was seeded with f_0,producer = 0.5.

2. (B)

   Microscope image acquired for competitions of the same strains and f_0,producer as in panel A, but with flow at 2100 μm/s (corresponding to Pe = 590).

3. (C)

   In the absence of flow, mathematical models show that the flux of GlcNAc molecules escaping from the biofilm (J_out) sets up a long-range concentration gradient.

4. (D)

   Fast flow over the biofilm generates a boundary layer, which causes rapid transport of GlcNAc to the bulk fluid, leading to a lower GlcNAc concentration at the biofilm surface for the same flux J_out. The displayed concentration profile was calculated for flow speeds corresponding to Pe = 590.

5. (E)

   Model calculation for the ratio of the GlcNAc concentration at the top of the biofilm with flow and without flow. Blue squares indicate points on the curve for the Péclet numbers at which competitions were performed in Figure 2B. The shading close to Pe = 0 indicates that the model is valid for Pe ≫ 1.