HEK293 Flp-In T-Rex were authenticated by STR analysis with The Center for Applied Genomics Genetic Analysis Facility (Sick Kids Hospital, Toronto). HeLa cells and primary fibroblasts were not independently authenticated Mycoplasma contaminationCell lines were routinely monitored for mycoplasma contamination as assessed by a commercial kit (MycoAlert, Lonza). Commonly misidentified lines (See ICLAC register)No commonly misidentified cell lines were used in this study.

INTRODUCTIONCompartmentalization is an essential characteristic of eukaryotic cells, ensuring that cellular processes are partitioned to defined subcellular locations. High throughput microscopy 1 and biochemical fractionation coupled with mass spectrometry 2-6 have helped to define the proteomes of multiple organelles and macromolecular structures. However, many compartments have remained refractory to such methods, partly due to lysis and purification artefacts and poor subcompartment resolution. Recently developed proximity-dependent biotinylation approaches such as BioID and APEX provide an alternative avenue for defining the composition of cellular compartments in living cells (e.g. 7-10 ). Here we report an extensive BioID-based proximity map of a human cell, comprising 192 markers from 32 different compartments that identifies 35,902 unique high confidence proximity interactions and localizes 4,145 proteins expressed in HEK293 cells. The recall of our localization predictions is on par with or better than previous large-scale mass spectrometry and microscopy approaches, but with higher localization specificity. In addition to assigning compartment and subcompartment localization for many previously unlocalized proteins, our data contain finegrained localization information that, for example, allowed us to identify proteins with novel roles in mitochondrial dynamics. As a community resource, we have created humancellmap.org, a website that allows exploration of our data in detail, and aids with the analysis of BioID experiments. BODYProximity-dependent labelling approaches have rapidly grown in popularity, as they provide a robust way to label the environment in which a protein resides in living cells 7,8 . In the most widely used of these techniques, BioID, a mutant E. coli biotin ligase -BirA* (R118G) -is fused in-frame with the coding sequence of a bait polypeptide of interest, and the resulting fusion protein expressed in cultured cells. While BirA* can activate biotin to biotinoyl-AMP, the abortive mutant enzyme exhibits a reduced affinity for the activated molecule. A reactive intermediate is thus released into the local environment that can react with free epsilon amine groups on nearby lysine residues 7 . This ability for BirA* to label a local environment has led to BioID being employed by multiple laboratories to define the composition, and in some cases the overall organization, of both membrane-bound and membraneless organelles (e.g. 7-10 ).Here, we set out to map a human cell by profiling markers (consisting of full-length proteins or targeting sequences) from 32 cellular compartments. These compartments include the cytosolic face of all membrane-bound organelles, the ER lumen, subcompartments of the nucleus and mitochondria, major membraneless organelles such as the centrosome and the nucleolus, and the main cytoskeletal structures (actin, microtubules and intermediate filaments). Several proteins were also queried throughout the endomembrane system to identify components enriched at locales a...

The polycomb repressive complex 2 (PRC2) histone methyltransferase plays a central role in epigenetic regulation in development and in cancer, and hence to interrogate its role in a specific developmental transition, methods are needed for disrupting function of the complex with high temporal and spatial precision. The catalytic and substrate recognition functions of PRC2 are coupled by binding of the N-terminal helix of the Ezh2 methylase to an extended groove on the EED trimethyl lysine binding subunit. Disrupting PRC2 function can in principle be achieved by blocking this single interaction, but there are few approaches for blocking specific protein-protein interactions in living cells and organisms. Here, we describe the computational design of proteins that bind to the EZH2 interaction site on EED with subnanomolar affinity in vitro and form tight and specific complexes with EED in living cells. Induction of the EED binding proteins abolishes H3K27 methylation in human embryonic stem cells (hESCs) and at all but the earliest stage blocks self-renewal, pinpointing the first critical repressive H3K27me3 marks in development.

α-N-terminal methylation of proteins is an important post-translational modification that is catalyzed by two different N-terminal methyltransferases, namely NTMT1 and NTMT2. Previous studies have suggested that NTMT1 is a tri-methyltransferase, whereas NTMT2 is a mono-methyltransferase. Here, we report the first crystal structures, to our knowledge, of NTMT2 in binary complex with S-adenosyl-L-methionine as well as in ternary complex with S-adenosyl-L-homocysteine and a substrate peptide. Our structural observations combined with biochemical studies reveal that NTMT2 is also able to di-/tri-methylate the GPKRIA peptide and di-methylate the PPKRIA peptide, otherwise it is predominantly a mono-methyltransferase. The residue N89 of NTMT2 serves as a gatekeeper residue that regulates the binding of unmethylated versus monomethylated substrate peptide. Structural comparison of NTMT1 and NTMT2 prompts us to design a N89G mutant of NTMT2 that can profoundly alter its catalytic activities and product specificities.