The Mitotic Cell Atlas is the next phase after a large study of the human genome components required for cell division which was part of the EU MitoCheck Project. The Mitotic Cell Atlas is supported by the EU project MitoSys, which is a systems approach to studying the dynamic protein networks in living human cells. Users of the Mitotic Cell Atlas can choose combinations of mitotic proteins and view in real time their location and which other proteins they associate with during cell division.
In Autumn 2018, the Mitotic Cell Atlas was described in Nature. Lead researcher is Jan Ellenberg of the European Molecular Biology Laboratory (EMBL) in Germany and collaborators are at the Research Institute of Molecular Pathology in Austria.
3D movies of fluorescently labelled proteins in HeLa cells were converted to time-resolved distribution maps of protein concentrations. Proteins were fluorescently labelled by creating cells lines with genome edits that inserted enhanced green fluorescent protein (eGFP) fused to the genes encoding the proteins of interest. Zinc finger nucleases or the CRISPR-Cas9 system were used to introduce fluorescent tags onto the endogenous proten, but for some genes, the gene with eGFP tag was stably integrated as a transgene. An automated pipeline was used to acquire datasets for 28 proteins. The microscope used to collect data is sensitive enough to count and detect differences between 100, 1000 and 10,000 proteins in certain locations.
This pilot set of 28 proteins, already known to be relevant for mitosis, were used to validate the system. The dataset can be mined for new information about dynamic multimolecular events living cells. Commonly researchers study single proteins as multi-color live-cell imaging over the duration of mitosis is very difficult to achieve. The Mitotic Cell Atlas allows the quantitative information from thousands of single-cell experiments to be integrated into the model. The goal of the project is to include all of approximately 600 different mitotic proteins in the atlas. The technology can be applied to study other cell functions such as cell death, cell migration and cancer cell metastasis.
A supervised machine-learning approach was used to define subcellular structures using known resident proteins of compartments and organelles relevant to mitosis which include chromosomes, nuclear envelope, kinetochores, spindle, centrosomes and midbody. A multivariate linear regression model was trained to assign the amounts of proteins of interest to the six compartments and quantitatively compare subcellular fluxes for all proteins.
