Phenotypic plasticity, life cycles and the evolutionary transition to multicellularity

SUMMARYUnderstanding the evolutionary transition to multicellularity is a key problem in evolutionary biology (1-4). While around 25 independent instances of the evolution of multicellular existence are known across the tree of life (5), the ecological conditions that drive such transformations are not well understood. The first known transition to multicellularity occurred approximately 2.5 billion years ago in cyanobacteria (5-7), and todays cyanobacteria are characterized by an enormous morphological diversity, based upon which they have been classified into five sections. They range from single-celled species (section I), unicellular cyanobacteria with packet-like phenotypes, e.g., tetrads (section II) and simple filamentous species (section III) to highly differentiated filamentous ones (sections IV and V) (8-10). The unicellular cyanobacterium Cyanothece sp. ATCC 51142, an isolate from the intertidal zone of the U.S. Gulf Coast (11), has been classified as a section I species, and it phylogenetically clusters with the other N2-fixing unicellular cyanobacteria (12). Here we report a facultative multicellular life cycle for a unicellular cyanobacterium, where multicellular filaments and unicellular stages alternate. In a series of experiments we identify the environmental factors underlying the phenotypic switch between the two morphologies. Then we experimentally confirm that the dissolution of filaments into solitary cells is triggered by changes in the external environment, which in turn is modified by the Cyanothece cells. Finally, using numerical models, we test a number of hypotheses regarding the nature of the environmental cues and the physical mechanisms underlying filament dissolution. While results predict that the observed response can be caused by an excreted compound in the medium, we cannot fully exclude changes in nutrient availability (as in (13,14)). The best-fit modeling results demonstrate a nonlinear effect of the compound, which is characteristic for density-dependent sensing systems (15,16). Further, filament fragmentation is predicted to occur by means of connection cleavage rather than by cell death of every alternate cell, which is corroborated by results from fluorescent and scanning electron microscopy. The phenotypic switch between the single-celled and multicellular morphology constitutes an environmentally dependent life cycle, which likely represents an important step en route to permanent multicellularity.