Testing for the fitness benefits of natural transformation during community-embedded evolution

Abstract

Natural transformation is a process where bacteria actively take up DNA from the environment and recombine it into their genome or reconvert it into extra-chromosomal genetic elements. The evolutionary benefits of transformation are still under debate. One main explanation is that foreign allele and gene uptake facilitates natural selection by increasing genetic variation, analogous to meiotic sex. However, previous experimental evolution studies comparing fitness gains of evolved transforming- and isogenic non-transforming strains have yielded mixed support for the 'sex hypothesis.' Previous studies testing the sex hypothesis for natural transformation have largely ignored species interactions, which theory predicts provide conditions favourable to sex. To test for the adaptive benefits of bacterial transformation, the naturally transformable wild-type Acinetobacter baylyi and a transformation-deficient ∆comA mutant were evolved for 5 weeks. To provide strong and potentially fluctuating selection, A. baylyi was embedded in a community of five other bacterial species. DNA from a pool of different Acinetobacter strains was provided as a substrate for transformation. No effect of transformation ability on the fitness of evolved populations was found, with fitness increasing non-significantly in most treatments. Populations showed fitness improvement in their respective environments, with no apparent costs of adaptation to competing species. Despite the absence of fitness effects of transformation, wild-type populations evolved variable transformation frequencies that were slightly greater than their ancestor which potentially could be caused by genetic drift.

Keywords: Acinetobacter baylyi; community evolution; evolution of sex; horizontal gene transfer; natural transformation; species interactions.

Fig. 1.

Illustration of the evolution experiment and direct competition assays. (a) Two focal strains (competent wild-type and isogenic non-competent counterpart) were cultured separately in their respective treatment conditions. (b) Focal strains were subjected to two treatment conditions: monoculture (abiotic) or co-culture with five competitor species. (c) Cultures were propagated by replacing spent broth with fresh media resulting in an approximately 34-fold dilution each passage. After the sixth passage, all species except the focal strain are killed off using LB agar amended with apramycin. The focal strain and freezer stocks of the competitors were allowed to grow to maximum density in LB broth before being inoculated together for another week of passaging. (d) After five full weeks of passaging, the focal strains are selected for again with use of apramycin agar and frozen in 25 % glycerol at −70 °C until tested against a common competitor to measure relative fitness. Figure created with BioRender.com.

Fig. 2.

Selection-rate constants of evolved populations standardised to ancestral populations in respective assay conditions (y=0). Asterisks describe significant differences (emmeans test) between evolved populations within the same assay conditions. Asterisks (bottom) describe significant differences (t-test) between evolved populations and zero. Top, middle, and bottom bands of the boxes denote the 75 % quartile, the median, and the 25 % quartile, respectively. Box whiskers are 1.5× the interquartile range. Fitness differences of each of the six biological replicates per treatment were measured in triplicate. (*, P<0.05; **, P<0.01; ***, P<0.001).

Fig. 3.

Treatment-level transformation frequencies of ancestral and evolved wild-type populations using heat-killed cell lysate. Transformation frequencies were measured in triplicate per biological replicate. The Abiotic Culture, Biotic Culture, and WT ancestor treatments had 6, 5, and one biological replicates, respectively.

Fig. 4.

Variation of transformation frequencies of biological replicates within evolutionary treatment groups for heat-killed cell lysate. Transformation frequencies were measured in triplicate per biological replicate. Points are biological replicates and bars are means±one standard error. Asterisks denote significantly different transformation frequencies between intra-treatment biological replicates (Kruskal-Wallis, P<0.05).