Survival of the first rather than the fittest in a Shewanella electrode biofilm

Abstract

For natural selection to operate there must exist heritable variation among individuals that affects their survival and reproduction. Among free-living microbes, where differences in growth rates largely define selection intensities, competitive exclusion is common. However, among surface attached communities, these dynamics become less predictable. If extreme circumstances were to dictate that a surface population is immortal and all offspring must emigrate, the offspring would be unable to contribute to the composition of the population. Meanwhile, the immortals, regardless of reproductive capacity, would remain unchanged in relative abundance. The normal cycle of birth, death, and competitive exclusion would be broken. We tested whether conditions required to set up this idealized scenario can be approximated in a microbial biofilm. Using two differentially-reproducing strains of Shewanella oneidensis grown on an anode as the sole terminal electron acceptor - a system in which metabolism is obligately tied to surface attachment - we found that selection against a slow-growing competitor is drastically reduced. This work furthers understanding of natural selection dynamics in sessile microbial communities, and provides a framework for designing stable microbial communities for industrial and experimental applications.

Fig. 1. Experimental platform.

a Anaerobic three-electrode bioreactor. Fresh sterile anaerobic medium is pumped at a constant rate into a growth chamber under positive pressure from a flow of sterile humidified argon. Both spent medium and argon exit through an air-locked outlet. Cells are grown on a graphite working electrode poised at anodic potentials with a platinum counter electrode and a Ag/AgCl reference electrode to control the voltage. A magnetic stir bar mixes the planktonic phase. b Only cells attached to the anode can grow. Hence, the strain ratio in the planktonic phase equals the ratio on the anode times the relative growth rate (see Note in Supplementary Information).

Fig. 2. Planktonic competition.

Anaerobic competition for lactate with fumarate as the terminal electron acceptor. The slopes of the solid lines (selection coefficients per MR-1 + gfp generation estimated by linear regression) reveal strong selection against strain ∆llpR when placed in competition with strain MR-1 + gfp. a Three replicate competitions in planktonic batch growth with an overall selection coefficient of −0.321 ± 0.010 gen⁻¹ (relative growth rate w = 0.536 ± 0.014). b Five replicate competitions across two experiments in bioreactors with electrodes unpoised; overall selection coefficient of −0.324± 0.007 gen⁻¹ (relative growth rate w = 0.533 ± 0.010). Different symbols (white or black diamonds, circles, and squares) represent replicate competitions. Dashed lines indicate 95% confidence bands derived from linear regression.

Fig. 3. Biofilm competition.

Anaerobic competition for lactate in bioreactors with electrodes poised (black circles, squares, and diamond, along with black lines). a Two replicate competitions conducted over 671 h (~165 MR-1 + gfp divisions) with abundances of each strain measured by flow cytometry. b Three replicate competitions conducted for 1488 h (~365 MR-1 + gfp divisions) and measured by dilution plate counting. For comparison, gray lines extrapolate the planktonic selection in bioreactors with unpoised electrodes (white diamonds). c Confocal microscopy volume projection of a mixed-culture biofilm at 929 h confirms that MR-1 + gfp cells (green) still sparsely populate the biofilm which remains dominated by ∆llpR cells (magenta). Symbols indicate experimental replicates (i.e., one symbol for each bioreactor), solid and dashed lines indicate linear regression and associated 95% confidence bands, respectively.

Fig. 4. Alternative hypothesis testing.

a The carbon source determines the metabolic rate of strain ∆llpR (solid line). Adding d,l-lactate provides a second carbon source, d-lactate, that is utilized preferentially and equally by both strains. The immediate increase in current from strain ∆llpR (but not from strain MR-1 + gfp, represented by a dashed line) demonstrates that it is limited by its capacity to metabolize l-lactate during growth on electrodes. Data shown is representative from N = 2 experimental replicates. b Letting air into an anaerobic bioreactor (black squares and solid line) provides planktonic cells with O₂ as a terminal electron acceptor. The ensuing competition causes a decline in the relative abundance of strain ∆llpR. Once anaerobic conditions have been re-established, strain ∆llpR returns to its original frequency, reflecting migration from a biofilm that is resilient to such perturbations. For comparison, strain ∆llpR is at a strong disadvantage during competition in an aerobic planktonic culture (white circles and dashed line). Data shown represent a single experiment.