Background

The protein composition of any given cell is intimately linked to its state of differentiation and functionality. Changes in a cell’s proteome may reflect its response to cell-intrinsic cues or to signals originating from elsewhere inside the respective organism or its environment. In turn they are indicative of the significance of those signaling cues. Deciphering proteomes and their dynamics in a cell type-specific fashion has thus become a main focus in current research, reaching a better understanding of molecular events underlying physiological or pathophysiological processes. Any proteomic approach in this direction, however, is challenged by the heterogeneity of cell types that are interconnected within a tissue or organ of interest. In the brain, for instance, different types of neurons and glial cells form the networks required to control animal or human behavior. Moreover, it is well established that information processing within these networks leading to long-term memory is strictly dependent on de novo protein synthesis and degradation. While this has been exemplified for a number of neuronal proteins (e.g., immediate early gene proteins), it is obvious that proteins that are up- or down-regulated in just a limited number of cells (or even are regulated oppositely in different groups of cells) may easily escape conventional modes of detection, where cellular proteomes are averaged across entire brain areas.

A number of labeling methods for cellular proteomes have been published in the last two decades, e.g., using isotope-coded affinity tags (Gygi et al., 1999) or isobaric tags for relative and absolute quantification (Ross et al., 2004), quantitative proteomic analysis using samples from cells grown in ¹⁴N or ¹⁵N media (Washburn et al., 2002; MacCoss et al., 2003), and stable isotope labeling by amino acids in cell culture (Ong et al., 2002; Andersen et al., 2005). Moreover puromycin (Schmidt et al., 2009) and non-canonical amino acids, e.g., azidohomoalanine (AHA) or homopropargylglycine, in combination with click chemistry have been used to decipher cellular proteomes (Link et al., 2003; Link and Tirrell, 2003; Beatty et al., 2006; Dieterich et al., 2006; Link et al., 2006; Dieterich et al., 2010). All of these strategies, however, fail to uncover cell-type specific proteomes within tissue or organ samples. Most recently, novel strategies to resolve this issue have been reported for C. elegans and Drosophila (Elliott et al., 2014; Erdmann et al., 2015; Yuet et al., 2015). They have in common the use of either a mutated aminoacyl-tRNA synthetase or an orthogonal aminoacyl-tRNA synthetase/tRNA for tagging of newly synthesized proteins with food-supplied non-canonical amino acids. Specifically, we could show that upon cell type-specific expression of a mutant Methionyl-tRNA synthetase (MetRS^L262G) as achieved by employing the well-established Gal4/UAS-system, the non-canonical amino acid ANL can be incorporated into proteins of selectable cell types in living Drosophila larvae and adult flies. An accompanying study by Niehues et al. (2015) used this method to show the causal involvement of mutated glycyl-tRNA synthetase in a model for the neurodegenerative Charcot Marie Tooth disease.

ANL-containing proteins can either be analyzed in protein extracts by using biochemistry and mass spectrometry or can be visualized in situ by fluorescence microscopy (Erdmann et al., 2015; Niehues et al., 2015). For more information see ‘Click Chemistry (CuAAC) and detection of tagged de novo synthesized proteins’. The following protocol details the metabolic labeling of proteins in larvae and adult flies with ANL.