Starvation induces phenotypic diversification and convergent evolution in Vibrio vulnificus

Abstract

Starvation is a common stress experienced by bacteria living in natural environments and the ability to adapt to and survive intense stress is of paramount importance for any bacterial population. A series of starvation experiments were conducted using V. vulnificus 93U204 in phosphate-buffered saline and seawater. The starved population entered the death phase during the first week and approximately 1% of cells survived. After that the population entered a long-term stationary phase, and could survive for years. Starvation-induced diversification (SID) of phenotypes was observed in starved populations and phenotypic variants (PVs) appeared in less than 8 days. The cell density, rather than the population size, had a major effect on the extent of SID. SID was also observed in strain YJ016, where it evolved at a faster pace. PVs appeared to emerge in a fixed order: PV with reduced motility, PV with reduced proteolytic activity, and PV with reduced hemolytic activity. All of the tested PVs had growth advantages in the stationary phase phenotypes and increased fitness compared with 93U204 cells in co-culture competition experiments, which indicates that they had adapted to starvation. We also found that SID occurred in natural seawater with a salinity of 1%-3%, so this mechanism may facilitate bacterial adaptation in natural environments.

Figure 1. Survival of V. vulnificus cultures in various media.

(A) Dynamics of V. vulnificus populations in phosphate-buffered saline (PBS, •), natural seawater (NSW, ◊), half-strength natural seawater (HNSW, □), and tryptic soy broth with 1.5% NaCl (TSBS, △) during the first week after inoculation. The results shown were derived from three replicate populations. (B) Dynamics of V. vulnificus populations in PBS (•) and TSBS (△) during an extended period of starvation (214 days). Representative results are shown from one experiment. (C) Electron micrographs of V. vulnificus cells showing the cellular morphology of a log-phase population (Log phase), a population starved in PBS for 4 days (PBS), and a population starved in TSBS for 4 days (TSBS). Most of the cells starved in TSBS were ruptured. All of the bars represent 2 µm. (D) The diameters and lengths of log-phase cells (△) and cells starved in PBS for 4 days (•).

Figure 2. Phenotypic diversification of the V. vulnificus 93U204 population during starvation in phosphate-buffered saline.

(A) Relative abundance of phenotypic variants in starved populations from two replicate trials. The population gradually diversified from the 93U204 M3H2P3 phenotype. (B) Compositions of populations, which were categorized based on their motility (top panel), hemolytic activity (middle), and proteolytic activity (bottom panel). WT, phenotype of the parental 93U204. (C) Association of diversity with mortality in starved populations. Twenty populations were starved in phosphate-buffered saline, and their survival and diversity were measured on days 14 or 15. The figure shows the distinction between low survival/un-diversified populations (left) and high survival/diversified populations. These 20 populations were derived from 12 separate inoculations (indicated by different colors).

Figure 3. Phenotypic variant (PV) compositions of different sets of starvation populations.

(A) The PV compositions of two sets of starvation populations, which started with the same inoculum, were analyzed for 33 days (n = 5). (B) Summaries of the phenotypic composition in terms of the motility (M), hemolysis (H), and proteolysis (P) activity levels are shown in. (C) Cluster analysis of the PV composition of each population. Green rectagulars, populations received inoculum 1; organge, populations received inoculum 2; blue, unstarved control. The analysis was based on the Bray-Curtis similarity and paired group methods.

Figure 4. Competition between various phenotypic variants (PVs) derived from different survivor populations during starvation survival.

Three PVs, i.e., C30-2 (A), D33-4 (B), and B33-12 (C), were competed with 93U204 cells in starvation conditions. The top panels show the population size and mutant phenotype dynamics. The bottom panels show the results of the competition experiments. For each mutant-wild-type (WT) pair, we measured the survival of the mutant-only (□), WT-only (•), or WT-mixed with 0.1% mutant (△) populations. The relative abundance of cells with the mutant phenotype was determined by examining 94 colonies in each trial. All three of the tested PVs increased their relative abundance after 14 days. Only one representative set of results (trial 1) is shown. The motility (Mot), hemolysis (Hem), proteolytic activity (Pro), growth on TCBS (TCBS), and growth on TSAS (growth) phenotype of each mutant was included in the figure and the characteristics used for differentiate mutant from WT colonies were shown in bold.

Figure 5. Effects of population density and population size on starvation-induced diversification.

93U204 cells were starved in 10 mL phosphate-buffered saline at densities of 10⁹, 10⁷, 10⁵, and 10³ CFU/mL (A), or in 100, 10 and 3 mL at a density of 10⁹ CFU/mL (B). The Shannon diversity index (top), percentage of cells with the wild-type (WT) phenotype (middle), and survival rate (bottom) of each population on days 14 (white bar), 28 (light grey bar), and 42 (dark grey bar) are shown. At day 14, phenotypic diversity (indicated by Shannon index) correlated positively with cell density, and percentage of WT and ratio of final to initial populations correlated negatively with cell density. After day 14, the trends in Shannon index and WT cell percentage were less apparent. Increased population size did not lead to more beneficial mutations in the population.

Figure 6. Starvation-induced diversification (SID) of V. vulnificus strains 93U204 and YJ016 in seawater.

We tested the SID in natural seawater samples with salinities of 1%, 2%, or 3%. The results for 93U204 are shown in (A) and (B), while those for YJ016 are shown in (C) and (D). The starting concentration was 10⁹ CFU/mL. The phenotypic variant compositions (A and C) and the summarized population compositions (B and D) for motility (upper), hemolysis (middle), and proteolysis activity (lower) are also shown.