Eukaryotic cells use membrane bound organelles to create specialized biochemical micro-environments. This allows eukaryotes to segregate incompatible biochemical reactions, improving efficiency. While membrane bound organelles are undoubtedly advantageous, their isolation requires mechanisms for communication with the rest of the cell, delivery of proteins and lipids, and selective removal of unnecessary or damaged proteins. By studying how cells communicate between organelles and selectively remove proteins, I found that the ESCRTs function in discrete pathways to maintain lipid and protein homeostasis. ER-PM contact sites are thought to be important for communication between the ER and PM. However, mutants lacking ER-PM contact sites cells grow well, indicating that additional mechanisms facilitate communication between the two membranes. To identify these mechanisms, I used saturating transposon mutagenesis coupled with next generation sequencing to identify synthetic lethal mutants in a strain that lacks ER-PM contact sites. The best hits were components of the ESCRT-III complex. I showed that the synthetic lethality was not due to the ESCRTs canonical function in the multivesicular body pathway, but instead, that ESCRT-III proteins were required to maintain normal lipid synthesis in cells lacking ER-PM contact sites. My data support a model in which ESCRT-III proteins act in a novel pathway to regulate lipid synthesis, and that this pathway is required when cells lose the regulatory feedback provided at ER-PM contact sites. To maintain protein homeostasis, cells use ubiquitin to mark unnecessary or damaged proteins in organelles of the secretory pathway. Ubiquitinated proteins are either degraded by the ERAD pathway, or are trafficked into the MVB pathway, and eventually delivered to the lysosome/vacuole lumen where they are degraded. The vacuole is downstream of the MVB pathway, and there are no known mechanisms to traffic ubiquitinated vacuolar proteins to endosomes. To understand how vacuole membrane proteins are degraded, I used a gene-fusion based genetic selection to identify genes required for vacuole membrane protein degradation. I then used an artificial in vivo ubiquitination assay to show that instead of functioning on endosomes, the ESCRTs can be recruited to the vacuole surface, where they directly internalize ubiquitinated membrane proteins into the vacuole lumen.