Genetic variation can influence protein expression through transcriptional and post-transcriptional mechanisms, and these effects may be conserved across tissues or specific to one. To characterize the shared and tissue-specific effects of natural genetic diversity on the proteome, we combined RNA-seq and multiplexed, quantitative mass spectrometry with a genetically diverse mouse population, the Diversity Outbred (DO) heterogeneous stock. We measured genome-wide transcript and protein abundance in livers and kidneys from 192 DO mice, and mapped quantitative trait loci that influenced transcript (eQTL) and protein (pQTL) expression. We identified nearly 3,000 pQTL in each of the liver and kidney, divided equally between local and distant variants. Local pQTL generally had larger effects on protein abundance, these effects were conferred primarily through transcriptional mechanisms, and half showed conserved protein responses in both tissues. In contrast, distant pQTL influenced protein abundance nearly exclusively through post-transcriptional mechanisms and most were observed to be specific to the liver or kidney. We applied mediation analysis and identified a second protein or transcript as the causal mediator for half of the significant distant pQTL. Furthermore, we identified groups of proteins within known pathways that shared coincident subthreshold distant pQTL for which we could identify a single causal protein intermediate from the same pathway, demonstrating the power of integrating ontology and mediation analyses to tease out subtle but real genetic effects from mapping populations with modest sample sizes. Overall, our analysis revealed extensive tissue-specific networks of direct protein-to-protein interactions that act to achieve stoichiometric balance of functionally related enzymes and subunits of multimeric complexes.