BioID based proteomic screen identifies VDAC2 as a coordinator of Bid dependent apoptotic priming during mitosis

Apoptotic priming refers to the proximity of cells to outer mitochondrial membrane permeabilisation (MOMP), the point of apoptosis commitment. Variations in priming are important for how both normal and cancer cells respond to chemotherapeutic agents. Individual cells adjust their level of priming in response to changing circumstances, such as genomic damage. How changes in apoptotic priming are dynamically coordinated by Bcl-2 proteins remains unclear. We have previously demonstrated that the Bcl-2 protein, Bid, becomes phosphorylated as cell enter mitosis, increasing priming. To understand how cells coordinate dynamic changes in priming, we have performed an unbiased proximity biotinylation (BioID) screen using Bid as bait. This identified several potential regulatory proteins outside the core Bcl-2 family. In particular we found VDAC2 to be essential for Bid phosphorylation dependent changes in apoptotic priming during mitosis. These results show the importance of dissecting the wider Bcl-2 interactome, and identify a key scaffold protein coordinating dynamic changes in single cell priming.


Introduction
Most cells respond to an apoptotic stimulus through mitochondrial outer membrane permeabilisation (MOMP) 1 . MOMP releases cytochrome c into the cytosol to initiate caspase activation. Following MOMP, cell death proceeds with largely invariant kinetics 2 . However, whether or not cells undergo MOMP following an apoptotic stimulus shows significant heterogeneity, both within a cell population as well as between different cells and tissues [3][4][5] . Apoptotic priming refers to how close a cell lies to MOMP, and therefore its subsequent response to damage 6,7 . Indeed, apoptotic priming of cancer cells can predict how a patient might respond to chemotherapy 8,9 . However, how variations in apoptotic priming are established, and how it changes dynamically, remain poorly understood.
MOMP is regulated through interactions between Bcl-2 family proteins 10,11 . The Bcl-2 family comprise: pro-apoptotic multidomain proteins Bax and Bak, which form the pores that drive MOMP; anti-apoptotic proteins, such as Bcl-XL, Bcl-2, and Mcl-1, that suppress MOMP by sequestering pro-apoptotic proteins; and the BH3-only proteins, which regulate the activity of the other two groups. BH3-only proteins may be activators or sensitizers. Activators, including Bim and Bid, bind Bax and Bak directly to initiate MOMP. Activator BH3-proteins can also be sequestered by antiapoptotic Bcl-2 proteins, but are released by the sensitizer subgroup of BH3-only proteins, such as Bad, which compete for binding 12 . Thus, variation in the expression and interaction of the different Bcl-2 family subgroups can set the level of priming, subtle changes in which have profound implications for processes such as embryonic development 13 .
Apoptotic priming can be highly dynamic, adjusting within minutes in response to changes in survival signaling 7 . Mitosis represents part of the cell cycle where dynamic changes in apoptotic priming are precisely controlled 14 . Upon entry to mitosis, cells become primed so that those that fail to divide correctly are eliminated.
However, apoptosis is initially suppressed for a period of time to allow spindle assembly to occur. Once cells satisfy the Spindle Assembly Checkpoint (SAC), they enter anaphase and reduce priming to allow cell division to occur without dying first.
If cells fail to satisfy the SAC sufficiently quickly, they will be eliminated through apoptosis. Thus, mitosis provides an excellent and unique model for understanding how cells dynamically adjust apoptotic priming within precise temporal boundaries.
The shift in apoptotic priming seen upon mitotic entry is linked to both posttranslational modifications to several Bcl-2 proteins [15][16][17][18][19][20] and changes in their expression, driven in part by Myc 21,22 . In particular, we have shown that the BH3only protein Bid is phosphorylated as cells enter mitosis to increase priming 20 , whilst others have shown that the slow degradation of Mcl-1 through mitosis determines how long a cell has to satisfy the SAC to avid apoptosis 16,23 .
How mitosis coordinates temporal adjustments to apoptotic priming is not clear, but has important implications for how cells respond to anti-mitotic drugs. Anti-mitotic drugs exploit the induction of apoptosis in cells arrested during mitosis and they have become an essential component of cancer therapy since the introduction of Taxol in the early 1990's 24 . However it is clear that cancers can become resistant to antimitotics, and cells can exit mitosis without accurately dividing, termed slippage 3,25 .
Slippage occurs through the gradual degradation of cyclin B in the absence of correct spindle attachment. For a cell to die before slippage, it must reach the threshold for MOMP before cyclin B levels decline far enough to result in slippage 25 .
Determining the mechanisms coordinating dynamic changes in apoptotic priming at mitochondria may allow us to improve the response to anti-mitotics.
To this end, we have performed an unbiased proteomic screen based upon proximity-dependent biotin identification (BioID) 26 , using Bid as bait. Using this approach, we identify a number of possible regulatory molecules controlling Bid function. In particular we find that the outer membrane porin, VDAC2, is critical for cells to adjust priming in response to Bid phosphorylation. Thus, this approach identifies mitochondrial proteins involved in co-ordinating Bcl-2 dependent changes in apoptotic priming.

Breast cancer cells show wide variations in apoptotic priming during a prolonged mitosis.
We wanted to use the increased sensitivity of cells to apoptosis during a prolonged mitosis to identify mitochondrial proteins involved in dynamically adjusting priming.
Previous studies have shown that there is considerable inter-and intra-line variability in the response of cells to anti-mitotic drugs 3 . Initially we followed five different cell lines, four cancer lines (MCF7, BT-474, SK-BT-3, MDA-MB-231) and one normal (MCF10A). We used live imaging on unsynchronised, sub-confluent cells in the presence or absence of taxol to determine single cell fate profiles over 65 hours, quantifying entry into mitosis, normal division, apoptosis and slippage (Fig. S1).
There was very little apoptosis in untreated cells. However, the five breast lines show wide variations in their response to prolonged mitosis (Fig. 1A, B). In the presence of taxol, MCF7, BT474, SK-BR-3 and MCF10A cells predominantly underwent apoptosis during mitosis (Fig. 1B, C). Conversely, triple negative MDA-MD-231 cells were prone to slippage. Interestingly, in the presence of taxol, little apoptosis was seen prior to mitosis. In all the lines, taxol significantly increased the time in mitosis compared with untreated cells, although this showed wide variations both within and between lines (Fig.1D).
We have previously shown that the BH3-only protein, Bid, was phosphorylated in mitosis 20 , which was linked to the sensitivity of colon carcinoma cells to taxolinduced apoptosis 20 . We asked if Bid phosphorylation was linked to sensitivity of the breast cell lines to death in mitosis. Mouse and human Bid are phosphorylated on functionally equivalent amino acids, serine 66 in mouse and 67 in human 20 .
However, whereas phosphorylated mouse Bid (mBid) shows a mobility shift by SDS-PAGE, human Bid does not. We therefore expressed an mBid-GFP fusion in the human breast lines in order to observe Bid phosphorylation. Mitotic cells for each line were collected by culturing the cells overnight in Nocodazole, and shaking off those in mitosis. mBid-GFP from mitotic cells was compared with that from unsynchronised cells by SDS-PAGE and immunoblotting (Fig.1E). Despite the variation in response to taxol, all five lines showed a slower migrating band indicative of mBid-GFP phosphorylation in mitosis.

Bid phosphorylation adjusts the set point of apoptotic priming during mitosis in breast cancer cells.
To determine whether Bid phosphorylation sets priming in breast cancer cells, we used a lentiviral approach to simultaneously knock down endogenous human Bid (hBid) whilst expressing mBid-GFP, using an ubiquitin promoter to express at levels similar that of endogenous Bid 20 . We used MCF7 cells, which predominantly undergo apoptosis if mitotic exit is delayed. Stable MCF7 lines were generated that expressed the hBid shRNA, alone or in conjunction with wildtype mBid-GFP (mBidWT-GFP), or mBid-GFP variants where S66 had been substituted to alanine (mBidS66A-GFP), or where the BH3-domain required for interacting with other Bcl-2 proteins had been made non-functional (mBidG94E-GFP). The lines were selected for similar expression levels by FACS ( Fig. 2A). Both mBidWT-GFP and mBidG93E-GFP showed the slower migrating phosphorylated form in mitotic cells as expected ( Fig. 2A). Endogenous hBid expression was knocked down more than 70% (Fig. 2B).
We next performed single cell fate profiling to compare parental MCF7 (WT MCF7) cells with those expressing shBid alone or with the mBid-GFP variants. In the absence of taxol there was no difference between the lines, either for apoptosis or time in mitosis (Fig2C, Fig. S2A ABT-737 alone did not impact on normal mitosis but did have a significant impact on the outcome of a prolonged mitosis. In MCF-7 cells, ABT-737 negated the impact of Bid phosphorylation on priming (Fig. 2E). MDA-MB-231 cell fate was dramatically shifted toward apoptosis by the BH3-mimetic (Fig. S3F). In both lines, the data indicate that dynamic adjustment of priming in mitosis is dependent upon antiapoptotic Bcl-2 proteins. Furthermore, shifting the cell fate away from slippage resulted in cells spending less time in mitosis (Fig.2F, Fig.S2E, Fig.S3G).
Together, these data highlighted Bid phosphorylation as a common mechanism by which mitotic cells dynamically adjust apoptotic priming, even when the overall outcomes appear quite different. We therefore decided to use this observation to identify regulatory complexes controlling temporal changes in priming using an unbiased proteomic strategy.

BioID based proximity labelling allows identification of Bid interaction partners in live cells.
BioID is a powerful approach to identify protein interactions in live cells and has been successfully applied to interrogate protein complexes within the centrosome, the nuclear pore, and cell-ECM adhesions [26][27][28][29] . The strategy involves expressing a bait protein as a fusion with the E.coli biotin ligase BirA containing a R118G amino acid substitution along with a myc epitope tag (from here on referred to as BirA*). In the presence of excess Biotin in the culture media, BirA* fusions generate reactive biotinyl-5'-AMP that can label primary amines. The half-life of the reactive biotinyl-5'-AMP limits the effective labelling radius to ~20 nm 26 . Labelled proteins are isolated by streptavidin affinity purification (Fig.4A). As labelling occurs in situ, it is particularly useful for interrogating complexes that might be disrupted by detergents used in cell extraction, a known issue with Bcl-2 protein interactions 30,31 . We therefore undertook a BioID strategy to identify Bid interactions in live cells.
To validate the functionality of mBid-BirA* fusions, we compared the pro-apoptotic

Proximity biotin labelling identifies Bid interacting partners in mitotic cells.
In order to identify mitosis specific Bid interacting partners, we utilised label free mass-spectroscopy using HeLa cells stably expressing mBid-BirA* ( BirA* tagged bait proteins were expressed from a lentivirus co-expressing tagBFP downstream of a T2A cleavage sequence. As tagBFP was expressed at a 1:1 ratio with each BirA* fusion, we could select stable lines with equivalent expression of each. We generated stable HeLa lines expressing mBidWT-BirA*, mBidS66A-BirA*, mBidG94E-BirA* or venus-BirA* (Fig.4B, Fig.S4A,B). Each line, either arrested in mitosis or left unsynchronised, was maintained in biotin containing media for 16 hours, a time frame that allowed significant enrichment for labelling in mitosis (Fig.   S4C). Labelled proteins were isolated by Streptavidin affinity purification, analysed by LC-MS/MS, and quantified using Progenesis QI. We performed three independent experiments, typically identifying >400 hundred proteins within each sample, with the bait being one of the most abundant (Fig.4C). Enrichment of proteins within each mBid-BirA* sample was determined by comparison with venus-BirA*.
For proteins enriched in mBid-BirA* expressing cells, we applied multiple criteria to determine which to prioritise for further analysis. We initially compared the sequence coverage of each and their relative representation in the contaminant repository for affinity purification (CRAPome), which collates proteins most often found within negative control AP-MS data (Fig.4D) 32 . Several proteins of interest only appeared in two out of the three repeats. Of these, Bax, was identified in mBidWT-BirA* cells and enriched in mitosis. Other candidates within this category, including kinases and signalling enzymes, are of interest with regard to Bid regulation, and will be the subject of future studies. However, for the present study we restricted selection to proteins detected in all three repeats, resulting in a significantly shorter list of candidates for validation (Table 1A and B).
Falling within these strict criteria was the mitochondrial porin, VDAC2, which has been previously linked to Bax and Bak function [33][34][35][36] . We compared enrichment of VDAC2 in unsynchronised and mitotic cells expressing either mBidWT-BirA*, mBidS66A-BirA* or mBidG94E-BirA* (Fig.4E). VDAC2 was enriched in mitotic cells expressing the mBidWT-BirA* bait compared to unsynchronised cells. Furthermore, VDAC2 enrichment was significantly less in mitotic cells expressing either BidS66A-BirA* or BidG94E-BirA*. Other proteins identified in all three repeats did not show mitosis or Bid variant specific enrichment (Fig.4E). For example, mtHsp70 showed no differences between lines expressing mBidWT-BirA* or the S66A and G94E variants. Notably, VDAC1, which is ~10-fold more abundant than VDAC2 in HeLa cells 37 , was not enriched. As quantification was performed using only non-conflicting peptide sequences, any sequence similarity between VDAC isoforms was not a confounding issue.
These results show that proximity labelling can identify potential candidate components of cell-cycle stage specific Bcl-2 protein complexes.

VDAC2 is required for Bid phosphorylation dependent priming in mitosis.
We next asked if VDAC2 influenced apoptotic priming in mitotic cells. To do this, we employed a knock out (ko) strategy in the MCF7 cells, where the response to taxol was influenced by Bid phosphorylation (Fig. 2C). MCF7 cells were transfected with a CRISPR/Cas9 targeting vector with a guide sequence to the start of the proteincoding region of the VDAC2 gene, and single cell clones isolated. Three clones, termed D11, D6 and D4, were identified containing insertions or deletions within the VDAC2 gene (Fig.5A). MCF7 cells have three copies of VDAC2. Sequencing indicated that D11 had all three copies containing frame-shifting insertions, D6 had two frame shift insertions and one WT allele, and D4 two frame shift insertions and one allele with a 3 bp deletion within the coding sequence.
We performed cell fate analysis to compare WT MCF7 cells to the three VDAC2 ko lines (Fig.5B). There were no differences in how the three VDAC2 ko lines progressed through normal mitosis compared to WT cells. However, compared to WT cells all three VDAC2 ko lines showed a shift toward slippage vs. apoptosis when mitosis was prolonged with taxol ( Fig.5B). Although many VDAC2 deficient cells still underwent cell death during prolonged mitosis, the proportion undergoing slippage was comparable to that seen in the shBid MCF7 cells, or those expressing mBidS66A or mBidG94E (cf. Fig.2C). To confirm the effect was due to the loss of VDAC2, we stably re-expressed VDAC2 tagged with a V5 epitope in the MCF7 D11 line. We utilised a lentivirus with an ubiquitin promoter to ensure that we did not overexpress VDAC2-V5, and a T2A cleavage peptide with a downstream tagRFP for selection by FACS. VDAC2-V5 localised to mitochondria (Fig.5C). Cell fate profiling showed that expression of VDAC2-V5 rescued D11 cells so that they now showed the same level of apoptosis in mitosis in taxol as parental MCF7 cells (Fig.5E,   Fig.S5).
Interestingly, BidWT-GFP and BidG94E-GFP were still phosphorylated in D11 cells arrested in mitosis, indicated by the mobility shift by SDS-PAGE (Fig.5D). We performed cell fate profiling in the presence or absence of taxol and/or ABT-737, comparing each mBid-GFP line to WT cells, the D11 ko and the D11-VDAC2-V5 (Fig.S5). Knock down of Bid had no further effect on the level of apoptosis in taxol beyond that already brought about by VDAC2 ko (Fig.5E). mBid-GFP, mBidS66A-GFP or mBidG94E-GFP expressing D11 cells all showed the same proportion of death vs. slippage in taxol as the D11 VDAC2 ko alone (Fig.5E). Furthermore, this level of death in mitosis was comparable to WT MCF7 cells infected with the shBid lentivirus (cf. Fig.2, Fig.S2). Apoptosis in all lines was restored to the level of WT MCF7 cells by treating with ABT-737 along with taxol. ABT-737 alone did not induce any significant level of apoptosis.
Together, the data presented here show that VDAC2 couples apoptotic priming in mitosis to Bid phosphorylation. Thus, VDAC2 is a necessary scaffold for coordinating the temporal regulation apoptotic sensitivity seen in mitotic cells.

Discussion
The mechanistic understanding of apoptosis obtained over the past three decades has ultimately led to novel drugs directly targeting Bcl-2 protein interactions to therapeutically manipulate mitochondrial priming. However, it is still unclear how cells dynamically adjust their level of mitochondrial priming in reaction to changing internal and external signals, a key determinant for how cells will respond to chemotherapeutic drugs. Here we employed an unbiased proteomic approach to identify proteins that control dynamic changes in priming during mitosis, a clinically important chemotherapy target, and identify VDAC2 as a coordinator of dynamic mitochondrial priming in mitotic cells.
Our strategy was based on previous findings showing that cells adjust priming as they enter mitosis. Cell death following delayed mitotic exit is Bcl-2 protein dependent, and several Bcl-2 family members undergo mitosis specific post- We have previously reported that full-length Bid is phosphorylated on serine 66 (67 in human) during mitosis 21 . Our data suggest that this phosphorylation does not convert Bid into a direct activator of MOMP, as with caspase 8 cleavage. Rather, phosphorylated full-length Bid appears to function as a sensitiser, reversibly increasing priming in mitosis. We therefore reasoned that we could use BioID for in situ proximity labelling to identify Bid vicinal proteins in mitosis, and use this is a model to elucidate how post-translational modifications of Bcl-2 proteins coordinate changes in mitochondrial priming. Following labelling, biotinylated proteins can be subjected to stringent affinity purification. This allows extraction of integral mitochondrial membrane proteins that might not be isolated under conditions necessary for more traditional protein-protein interactions approaches, such as coimmune precipitation. Furthermore, this approach avoids known issues with detergents on Bcl-2 protein interactions 28 . The compromise for these benefits is that BioID has the potential to label non-specific proximal proteins 26 . However, we mitigated against detecting false positives by comparative analysis of WT, S66A and G94E Bid variants, all of which showed equivalent subcellular localisation, but had distinct effects on mitochondrial priming 24,42 .
Of the proteins enriched in mBid-BirA* cells relative to venus-BirA* controls, VDAC2 was notable because it was enriched specifically in mitotic cells, and then only when Bid was both functional and able be phosphorylated. Thus, VDAC2 met the criteria expected for a regulator of Bid dependent mitotic priming. The three mammalian isoforms of VDAC, show high sequence similarity, but appear to have distinct roles. proteins could bind VDAC2, it was only required for Bax dependent death. However, an earlier study from the same group found that although loss of VDAC2 reduced both Bax and Bak mitochondrial association, they both retained the ability to kill cells in response to an apoptotic signal 34 . Finally, Bax retrotranslocation off mitochondria also requires VDAC2 33 . The details of these studies, often contradictory, probably reflect varying genetic backgrounds and cell types. However, it is clear that VDAC2 is emerging as a nexus of Bcl-2 protein coordination on mitochondria.
Knock out studies in mice support the idea that the function of VDAC2 in apoptosis regulation is influenced by the genetic and cellular context. Deletion of VDAC2 was originally reported to be embryonic lethal. VDAC2 -/-MEFs isolated from these mixed genetic background mice showed increased susceptibility to Bak activation. In support of these findings, a subsequent thymus specific VDAC2 knock out resulted in increased thymocyte apoptosis, which was rescued by deletion of Bak. However, a recent CRISPR generated deletion of VDAC2 in an inbred mouse line did not result in either unrestrained Bak activation or loss of embryonic viability, although the resulting mice failed to thrive for reasons more related to VDAC2's role in metabolite transport.
Our data here show that the absence of VDAC2 dampens mitotic priming to the same extent as loss of Bid or the ability to phosphorylate Bid. Furthermore, the dual loss of Bid function in mitosis and VDAC2 was not additive, showing that the priming effect of Bid phosphorylation was mediated by VDAC2. VDAC2 therefore appears to coordinate Bid in the context of a sensitizer BH3-protein, temporally adjusting mitochondrial priming in mitosis. Bid, along with Bim and Puma, are defined as direct activator BH3-proteins 38 . Bid as an activator is largely based on its function downstream of caspase 8 cleavage 39 . However, full-length Bid undergoes reversible post-translational modifications 40,41 . It is interesting to speculate that a significant role for Bid may be aligned with sensitisation in response mitosis. Notably, perturbing Bid function or S66 phosphorylation did not alter normal cell cycle progression, just the outcome following delayed mitotic exit.
BioID is a powerful approach that has potential for further dissecting the interactions of Bid and other Bcl-2 proteins in live cells. In this regard, we identified other potential candidates for novel interaction partners of Bid, including kinases, phosphatases and their regulatory subunits. One issue was that some of these proteins were identified through a small number of peptides in two out of three repeats. Whereas it is possible that some of these are real interactions, all need further validation beyond the scope of the present study. We were also surprised that we did not detect significant levels of other Bcl-2 proteins, particularly anti-apoptotic partners for Bid. We did isolate biotin labelled venus-Bcl-XL when it was coexpressed with the activator tBid-BirA*, and detected Bax and Bcl-XL in the stable HeLa lines in two out of three repeats. However only a small number of unique Bax peptides were detected, making quantitative comparisons between conditions unreliable. Despite this, we observed that Bax, was enriched in the presence of mBidWT but not mBidG94E, suggesting we were observing a BH3-domain dependent interaction. However, the fact that we didn't detect other Bcl-2 family partners remains apparent. Perhaps intermediate scaffold proteins such as VDAC2 are more readily labelled in situ. It will be interesting to see whether other Bcl-2 proteins also interact with similar scaffold proteins.
Finally, these results add to our understanding of how cells coordinate the dynamic changes in priming as they progress through mitosis. As with studies on VDAC, studies on mitotic cell death have variously implied significant roles for different specific BH3-proteins, Bak or Bax. A genome-wide screen in HeLa cells identified Bax as key for taxol mediated cell death, whereas the same approach in RKO cells did not directly identify any Bcl-2 proteins, but instead myc dependent transcription of Bid, BIM and Noxa. An siRNA approach in HeLa cells specifically targeted at Bcl-2 proteins identified roles for Noxa and Bim in conjunction with Mcl-1 loss. Together, these studies indicate that the outcome of delayed mitotic exit is not determined by a specific BH3-protein or multi-domain effector, but rather a complex interplay of many, the balance of which will vary between cell and tissues types. Indeed, our data show  Cells were described as being in mitosis from the initial point at which a cell began to round, till the first frame of cytokinesis or membrane blebbing. Cells that exhibited morphology associated with apoptosis such as membrane blebbing, shrinking of cell body, and in the later stages the formation of apoptotic bodies or secondary necrosis were described as having undergone apoptosis in mitosis 38 .
Cells that exited mitosis as a single cell, or as two or more non-identical daughter cells were grouped together under the definition of having undergone mitotic slippage 3 .
Immunofluorescent staining of cells was performed as previously described 39 , using Alexafluor 594 and 488 conjugated secondary antibodies (Stratech).
Immunostained cells were visualised using a Zeiss AxioImager M2 florescence microscope with 63x (NA 1.4) PLAN-APOCHROMAT objective. Images were acquired with Micro-Manager software and processed using ImageJ.
For immunoblotting, proteins were separated by SDS-PAGE and transferred to nitrocellulose. Following incubation with primary antibodies, proteins were detected using IrDye 800 and 680 secodary antibodies and an Odyssey CLx imager (LiCor). Quantification of blots was performed using ImageStudio (LiCor).

Mitotic enrichment of cultured cells.
Asynchronous cell populations were treated with nocodazole (Sigma-Aldrich, 200 ng/mL in complete growth media) for 16 hours to arrest cell in mitosis. Mitotic cells were then dislodged from the cell culture dish by tapping the vessel laterally against the bench. Media containing dislodged mitotic cells was removed and centrifuged at 350 RCF for 5 minutes to pellet cells ready for subsequent analysis.

Flow cytometry analysis of cell cycle.
Cell cycle distribution within a population was quantified by DNA content. Cycling cells were fixed with 70% ethanol and stained with DAPI. Stained cell samples were analysed on a BD Biosciences Fortessa flow cytometer, using Diva Version 8 (BD Biosciences) to collect raw data. Raw data was then analysed using Modfit LT (Version 5, Verity Software House) to quantify distribution between cell cycle phases.

BioID mediated proximity labeling.
For small scale assays such as Western blotting or immunofluorescent staining, cells expressing BirA* fusion proteins were seeded in a 6 well plate to approximately 80% confluence on day of labelling. Growth media was replaced with complete growth media supplemented with 50 µ M biotin (Sigma-Aldrich), and cells cultured for a further 16 hours. Following incubation, biotin supplemented media was removed and replaced with regular growth media for 1 hour. Cells were then washed three times in PBS and assayed as appropriate.
For mass spectrometry experiments, cells expressing BioID fusion proteins were seeded into two 100 mm diameter dishes to be approximately 80% confluent next day. On the day of labelling, growth media was replaced with complete growth media supplemented with 50 µ M biotin (Sigma-Aldrich) and cultured for a further for 16 hours, along with any other drug treatments. Following incubation, biotin supplemented media was removed and replaced with regular growth media for 1 hour to allow free biotin to diffuse out of cells. Cells were then washed three times in PBS and lysed in RIPA supplemented with 1x protease inhibitor cocktail (Millipore), 100 mM Na 3 VO 4 and 500 mM NaF. Following lysis, lysates from both dishes were pooled, protein concentration measured and biotinylated proteins isolated via streptavidin-bead purification.

Streptavidin affinity purification.
Biotinylated proteins were isolated using MagReSyn ® streptavidin (MagReSyn ® ). 100 µ L of bead slurry was decanted into a microcentrifuge tube and exposed to a magnetic field to separate the liquid and bead phases. The liquid phase was discarded and beads were washed three times in ice-cold lysis buffer.
Appropriate volumes of cell lysate were added to beads to yield 1 mg of total protein, and the volume made up to a total of 1 mL with lysis buffer. Lysates and beads were incubated overnight at 4°C on an end-over-end tumbler. The next day beads were collected, washed twice in lysis buffer, once in urea wash buffer Carbamidomethylation of cysteine was applied as a fixed modification, while Nacylation and oxidation of methionine, as well as the biotinylation of lysine, were applied as variable modifications during the database searches.
Label free protein quantification and data analysis.