Figure S1.

Phase separating domains are essential for most systems and several combinations of boxB-λ_N22 systems fulfill the independent assembly criterion, related to Figure 3

(A) To test if phase-separating domains are required for the different membrane systems, we used constructs that target PylRS or MCP with or without FUS/EWSR1 individually to the four membranes. Here we show FFC analysis of separate OT organelles with and without phase-separation domains (i.e., PylRS with and without FUS as well as MCP with and without EWSR1). HEK293T cells expressing the dual color ms2 reporter (EGFP^39TAG, mCherry^190TAG::ms2), tRNA^Pyl, a respective PylRS construct without (upper rows) or with phase-separation domain (lower rows), as well as either no MCP protein (first columns) or a membrane-targeted MCP protein without (second columns) or with a phase-separation domain (third columns) in presence of SCO-K. For the OMM, ERM and GM system the variants without FUS loose functionality. Only the PM-targeted PylRS variant is efficient with and without a phase-separation domain. For the case of MCP, the GM and OMM the systems without EWSR1 perform substantially worse. Shown is the sum of at least three independent experiments. (Note that the black line in the FFC dot plots indicates the EGFP signal cut-off used to identify transfected cells). Together, three out of four systems require a protein condensation to be functional, while the PM and PMP systems perform similarly, potentially owing to the anyway good spatial separation of the plasma membrane from the cytoplasm.

(B) To test the independent assembly criterion of the different membrane systems for the λ_N22–boxB system we cloned constructs that target PylRS or λ_N22 peptides (with FUS or EWSR1 respectively) to the four different membranes. Here we show FFC analysis of separate OT organelle constructs. HEK293T cells expressing the dual color boxB reporter (EGFP^39TAG, mCherry^190TAG::boxB), tRNA^Pyl, a respective PylRS OT organelle construct as well as either no λ_N22 peptides (fifth column) or the λ_N22 peptides targeted to each respective membrane. The ncAA SCO-K was present in all experiments. Similar to the experiments for the ms2–MCP systems (Figure 3) we always observed the expected highly selective mCherry^190TAG::boxB production if λ_N22 and PylRS were targeted to the same membrane (vertical population). We also observed enhanced mCherry^190TAG::boxB expression if λ_N22 and PylRS were targeted to the GM and either the ERM or PM. Only the combination of the ERMP and PMP system, and that of the OMMP system with any other system showed no mixing, i.e., a similar expression profile as in the absence of λ_N22 (background expression), shown in the fifth row (-λ_N22) above the cytoplasmic PylRS control, and thus only these fulfilled the independent assembly criterion. Shown is the sum of at least three independent experiments. [Note that the black line in the FFC dot plots indicates the EGFP signal cut-off used to identify transfected cells, the red dashed line highlights background expression (in absence of λ_N22) for each targeted membrane].

(C) FFC gating scheme to identify single HEK293T cells. HEK293T cells were first identified based on SSC-A and FSC-A values. Subsequently, single cells are identified using SSC-A and SSC-W parameters. One representative example is shown. Here, HEK293T cells were transfected with a cytoplasmic PylRS system, the indicated ms2 reporter (EGFP^39TAG, mCherry^190TAG::ms2) and tRNA^Pyl in presence of the ncAA SCO-K. Cells that passed the first gate (top) are subsequently gated for single cells (bottom).

(D) FFC analysis of cells expressing an iRFP::EGFP^39TAG reporter, tRNA^Pyl and PylRS^AF or PylRS^AA. iRFP served as a transfection control, whereas EGFP fluorescence reports on successful amber codon suppression. Cells were incubated either with no ncAA, with SCO-K (250 μM), TCO^∗A-K (250 μM), Cbz-K (250 μM), with 3-IF (1 mM) or SCO-Y (250 μM). For PylRS^AF, robust EGFP expression was observed in the presence of SCO-K, TCO^∗A-K and Cbz-K. For PylRS^AA, EGFP expression was only enhanced in the presence of 3-IF or SCO-Y. Shown is the sum of at least three independent experiments. (Note that the black line in the FFC dot plots indicates the iRFP signal cut-off used to identify transfected cells).