In vivo monitoring of circadian rhythms depends on reliable and non-invasive detection methods. This is often achieved by expressing reporter genes heterologously under the control of a circadian promoter. The activity or fluorescence of the gene product is then used as a readout. To avoid the generation of such reporter strains, we recently established a reporter-free detection method for cyanobacterial batch cultures. To determine whether these rhythms are driven at the level of individual cells or result from population-based effects, such as gating of cell division, we analysed individual Synechocystis sp. PCC 6803 cells by combining a microfluidic cultivation technique with time-lapse microscopy imaging. Hundreds of time-lapse image sequences were processed using our deep learning cell segmentation workflow. Although the cells had been entrained by a 12-hour light-dark cycle, cell division did not display a circadian rhythm. This indicates the absence of circadian gating of cell division in Synechocystis. Instead, we observed a circadian oscillation in the average brightness of the phase contrast in individual Synechocystis cells. To demonstrate how phase-contrast analysis of single cells can be complemented by backscatter analysis of batch cultures, we investigated the wildtype, a deletion mutant known to affect circadian rhythms (∆kaiC3) and complementation strains. We concluded that phase contrast and backscatter likely measured the same rhythmic changes in the refractive index of the cells. The method presented here will advance circadian research by enabling the analysis of circadian rhythms in individual cells without the need for expression of reporter molecules.
