An Integrative Biology Approach to Quantify the Biodistribution of Azidohomoalanine in Vivo

Abstract

Background Identification and quantitation of newly synthesized proteins (NSPs) are critical to understanding protein dynamics in development and disease. Probing the nascent proteome can be achieved using non-canonical amino acids (ncAAs) to selectively label the NSPs utilizing endogenous translation machinery, which can then be quantitated with mass spectrometry. Since its conception, ncAA labeling has been applied to study many in vitro systems and, more recently, the in vivo proteomes of complex organisms such as rodents. We have previously demonstrated that labeling the murine proteome is feasible via injection of azidohomoalanine (Aha), an ncAA and methionine (Met) analog, without the need for Met depletion. With the ability to isolate NSPs without applying stress from dietary changes, Aha labeling can address biological questions wherein temporal protein dynamics are significant. However, accessing this temporal resolution requires a more complete understanding of Aha distribution kinetics in tissues. Furthermore, studies of physiological effects of ncAA administration have been limited to gross observation of animal appearance and behavior. Results To address these gaps, we created a deterministic, compartmental model of the -kinetic transport and incorporation of Aha in mice. Parameters were informed from literature and experimentally. Model results demonstrate the ability to predict Aha distribution and labeling under a variety of dosing paradigms and confirm the use of the model as a tool for design of future studies. To establish the suitability of the method for in vivo studies, we investigated the impact of Aha administration on normal physiology by analyzing plasma and liver metabolomes following various Aha dosing regimens. We show that Aha administration induces metabolic alterations in mice. However, these changes are minimal as reflected by the small percentage of metabolites that are differentially abundant between non-injected controls and Aha treatment groups. Conclusions Our results demonstrate that we can reproducibly predict protein labeling and that the administration of this analog does not significantly alter in vivo physiology over the course of our experimental study. We expect this model to be a useful tool to guide future experiments utilizing this technique to study proteomic responses to stimuli.