Abstract
A rapid environmental change represents stressful conditions for living organisms and may reduce the demographic components of fitness (survival rate, reproduction), leading to a risk of extinction. Phenotypic plasticity, whereby a given genotype expresses different phenotypes in different environments, enables organisms to cope with fluctuating environments and avoid extinction when it is adaptive, i.e. when it increases mean fitness across environments. Conversely, fitness reduction induced by environmental changes is generally considered as maladaptive plasticity, due to passive response to stress. But can stress-induced death, leading to severe population decline, sometimes be interpreted as adaptive plasticity? And how may selection operate on such a trait in fluctuating environments? I investigated these questions with the halotolerant microalga Dunaliella salina, where programmed cell death has been established. We showed that after a salinity rise, one of the strains from our collection displayed a massive population decline (-69% in one hour), but that this decline was followed by a demographic rebound. Another genetically-related strain did not exhibit this initial decline, but grew more slowly in the second phase. Strikingly, the declining strain presented a positive correlation between decline and subsequent growth, which was more pronounced in conditions more favourable to growth (more light, less competition). Moreover, the decline could be diminished by a programmed cell death inhibitor, indicating that is not merely a passive response to stress. We have then grown these two strains in competition for 26 weeks, alternating intermediate and high salinity. The declining strain persisted until the end of the assay at frequencies higher than 50% at the end of each salinity cycle, despite its massive decline at each salinity rise. I showed that this co-existence may be partially explained by the competition near the stationary phase, which attenuates the effect of the initial decline-rebound dynamics. Finally, I modelled 3 mechanisms which may explain both the decline-rebound pattern and the competition with the non-declining strain: an altruistic hypothesis where dying cells release substrates usable by the remaining cells; a trade-off between salinity tolerance and reproduction; and a population heterogeneity in cell condition, where the elimination of damaged cells result in greater mean growth rate. Together, my results suggest that cell death can be advantaged by natural selection, and potentially represents a form of adaptive plasticity in response to an environmental stress.