Before working with more complex neurons, we needed a simpler model to establish our methodology. For this, we selected the immortalised rodent neuronal cell line, ND7/23. When undifferentiated, these cells divide rapidly with high protein turnover, making them suitable to optimise the methodology. Additionally, they can be differentiated to provide a good neuronal model. Using undifferentiated ND7/23 cells that we transiently equipped with the translational machinery, we demonstrated that ncAAs could be incorporated across the proteome in place of canonical AAs. We also optimised what sense codon will be targeted and what ncAA was most suitable for follow-up proteomics. Another aim of screening different sense codons was to compare our technique to an established technique that replaces methionine with an ncAA.
Following this AA screen, we optimised conditions for high ncAA incorporation efficiency and a high-yielding click reaction. Additionally, we tested different clickable handles, including fluorophore handles for downstream microscopy and handles for downstream enrichment of labelled proteins. However, long-term studies were not possible when the translational machinery was only transiently in the cells.
To overcome this, ND7/23 cells were generated that stably expressed the translational machinery. Using these cell lines, we optimised our oxidative injury model. We used different compounds to model the acute or chronic injury phases of neurodegeneration. Changes to cell morphology and protein turnover following injury were validated, before we demonstrated that these changes could be followed using proteomics. Proteome changes were captured in ND7/27 cells in response to acute and chronic stress. Using our technique, we also demonstrated that we could achieve greater proteome coverage targeting other canonical AAs relative to methionine.
One major advantage of this GCE-based methodology is the ability to selectively target different subsets of the proteome. All proteins secreted by cells make up the secretome. They are very important in cell-to-cell communication and are typically lost during sample processing. Similarly, proteins located on the outer membrane of cells are essential to cell function but are lost in these steps. By using different clickable handles and labelling approaches, we were able to selectively target these subsets, enrich them and capture changes in expression.
Following the establishment in ND7/23 cells, we wanted to expand this technology into more complex human-derived neurons. Cells stably expressing the translational machinery were generated. However, following rounds of optimisation, we did not observe incorporation of the ncAA using downstream click chemistry with a fluorophore. Different strategies were tested to try to overcome these challenges, while in parallel we also tested a complementary method of proteome labelling. Using these technologies, we optimised our system for the future generation of stable cells. This will allow more biologically relevant studies, including in cells derived from human patients.
