We report the development of prox-SILAC method that combines proximity-dependent protein labeling (APEX/HRP) with metabolic incorporation of stable isotopes (pulse-SILAC) to map newly synthesized proteins with subcellular spatial resolution. We apply prox-SILAC to investigate proteome dynamics in the mitochondrial matrix and the ER lumen. Our analysis reveals a highly heterogeneous distribution in protein turnover dynamics within macromolecular machineries such as the mitochondrial ribosome and respiratory complexes I-V, thus shedding light on their mechanism of hierarchical assembly. Furthermore, we investigate the dynamic changes of ER proteome when cells are challenged with stress or when cells are undergoing stimulated differentiation, identifying subsets of proteins with unique patterns of turnover dynamics, which may play key regulatory roles in alleviating stress or promoting differentiation.