Mapping protein networks in yeast mitochondria using proximity-dependent biotin identification coupled to proteomics

Summary Proximity-dependent biotin identification (BioID) permits biotinylation of proteins interacting directly, indirectly, or just localized in proximity of a protein of interest (bait). Here, we describe how BioID coupled to proteomics and network biology can be used to map protein proximities in yeast mitochondria, aiding in visualization of complex protein-protein interaction landscapes. For complete information on the use and execution of this protocol, please refer to Singh et al., 2020.


Primer design
Primers must contain at their 5 0 ends homologous sequences to the gene of interest. More specifically, the forward primer must contain an overhang of about 40-51 nucleotides ending at the Figure 1. Construction of a plasmid for C-terminal BirA*-tagging of a protein of interest (A) The sequence encoding BirA* tag was inserted into the plasmid pYM28 carrying an HIS3 auxotrophic marker. (B) A segment of the plasmid including linker sequence, BirA* coding sequence, stop codon, and the auxotrophic marker is amplified by PCR. Then, this DNA segment is inserted by homologous recombination into the genomic DNA, thereby fusing BirA* to the C terminus of the protein of interest.

OPEN ACCESS
penultimate codon of the open reading frame and followed by S3 sequence (5 0 -CGTACGCTG-CAGGTCGAC-3 0 ). The reverse primer features the S2 sequence (5 0 -ATCGATGAATTCGAGCTCG-3 0 ) at the 3 0 end and, at 5 0 end, an overhang downstream (at least 70 bp) of the stop codon of the protein of interest. The size of the overhang should be about 40-51 nucleotides. Hence, following homologous recombination, the native stop codon of the gene of interest is replaced by 5 0 -S3 sequence / linker / BirA* / STOP / terminator/ HIS/TRP selection marker / S2-3 0 sequence.
CRITICAL: Make sure that the region between the stop codon and the reverse primer binding site does not contain promotor/terminator sequences of adjacent genes. Typically, if the region is >200 bp further from the ORF of the adjacent gene, that would be considered appropriate for homologous recombination. For tagging of Mrpl4 we used Forward primer ''Mrpl4_S3'': 5 0 -AAATACTTGAAACAGCTAATTCATGCAAGCTCCGTG-GAGCAAGCAACTGCAcgtacgctgcaggtcgac-3 0 and Reverse primer ''Mrpl4_S2'': 5 0 -GAA-TACATACAGCGTATAGCTATATAAAGGTGTATGTATGTGTGTATGTATatcgatgaattcgagctcg-3 0 . The underlined bases correspond to the S3 and S2 sequences.
2. Set up the PCR reaction as follows.
3. Run the PCR on the following cycling conditions.
Note: This protocol is for Q5 polymerase (NEB).

PCR product purification
Use any of the commercially available PCR clean-up kits for this step. The PCR product is dissolved in sterile deionized H 2 O and ready for transfection. Mass spectrometer Q Exactive HF (Thermo Fisher Scientific). Other mass spectrometer capable of Data-Dependent Acquisition (DDA) with sufficiently high resolution and scan speed, like LTQ-Orbitrap Elite and Velos series, Orbitrap Tribid instruments, or TOF instruments can be used as well. Operate the Q Exactive HF in DDA positive mode with a survey full scan MS spectra (from m/z 375-1,600) acquired with resolution R = 60,000 at m/z 400 (AGC target of 3 3 10 6 , maximum ion time 60 ms). The 10 most intense peptide ions with charge states between 2 and 5 are sequentially isolated to a target value of 1 3 10 5 , R = 60,000 at m/z 400, maximum ion time 60 ms, isolation window 2.0 m/z and fragmented at 27% normalized collision energy. Typical mass spectrometric conditions are: spray voltage, 1.5 kV; no sheath and auxiliary gas flow; heated capillary temperature, 300 C; ion selection threshold, 33.000 counts, S-lens RF level 50.

Stock solution preparation
Note: autoclave the same day after preparation. YP can be stored at room temperature for long term.
Note: autoclave the polyethylene glycol 50% solution, and sterile filter the Lithium acetate 1 M solution before use. All reagents can be stored at room temperature except for SS DNA, which is stored at À20 C.

Reagent Final concentration (g/L) Amount
Bacto tryptone 20 g/L 20 g Note: Used when made fresh.
Note: Can be stored at 4 C for up to 4 months. Add Protease cocktail inhibitor fresh before use.
Note: Can be stored at 4 C for up to 4 months.
Note: Make fresh before use. Can be used up till 24 h.
Caution: TFA solutions are corrosive. Prepare the solutions in a fume hood and handle with gloves. Make fresh before use. Protocol Caution: TFA solutions are corrosive. Prepare the solutions in a fume hood and handle with gloves. Make fresh before use.

Reagent
Caution: TFA solutions are corrosive. Prepare the solutions in a fume hood and handle with gloves. Make fresh before use.
Caution: This buffer must be prepared fresh. Otherwise, the pH will begin to increase due to the high volatility of NH 4 OH, and its elution strength will decrease.
Caution: TFA solutions are corrosive. Prepare the solutions in a fume hood and handle with gloves. Make fresh before use.

STEP-BY-STEP METHOD DETAILS
Tagging the protein of interest via homologous recombination

Timing: 6 days
This step is required to genomically tag BirA* to the C-terminal of the protein of interest, using transfection mediated homologous recombination. Tagging is confirmed by Western blotting using anti-BirA antibody. 1. Grow the yeast cell line of interest for 12-16 h in YP + 2% dextrose. 2. Dilute the culture to 0.2 OD in YP + 2% dextrose and incubate under shaking at 30 C until the culture reaches an OD of $1.0. 3. Harvest 1.5 mL of cells by spinning at 3,000 3 g for 5 min at $21 C. 4. Remove supernatant. 5. Resuspend cells in 1 mL of sterile water and spin at 3,000 3 g for 5 min at $21 C. 6. Remove supernatant. 7. Perform PCR-tagging via homologous recombination. a) Resuspend the cells in 100 mM lithium acetate and shake them for 10 min at 30 C. b) Pellet the cells at 3,000 3 g for 5 min and resuspend them by vortexing in 360 mL of HR buffer that includes the DNA fragment for homologous recombination. c) Incubate the cells at 42 C for 45 min under constant shaking. d) Spin down the cells at 3,000 3 g at $21 C for 5 min, remove supernatant and resuspend them in 1 mL of YP + 2% dextrose. e) Incubate the cells for 1 h at 30 C under constant shaking. f) Spin down the cells at 3,000 3 g at $21 C for 5 min and resuspend them in 100 mL of water, and spread them on an appropriate selective agar plate. 8. After three days ($72 h), single-cell colonies should emerge on the plate.
Pause Point: the plates can be stored at 4 C for up to 2 weeks. 9. pick single colonies and streak them on a fresh selection plate and let them grow at 30 C for 36-48 h.
Pause Point: the plates can be stored at 4 C for up to 2 weeks.
10. Inoculate the clones in YP + 2% dextrose and let them grow for 12-16 h at 30 C under constant shaking. 11. Perform protein extraction. a) Harvest 1 OD of cells and spin them down. b) Resuspend the cells in 0.2 mL of 0.1 M NaOH and incubate at $21 C for 5 min. c) Spin the cells down at 3,000 3 g for 5 min at $21 C and remove supernatant. d) Resuspend the cells in 30 mL of 13 SDS sample buffer. e) Boil the sample at 95 C for 5 min. 12. Load the sample on an SDS-PAGE gel. 13. Run the gel according to the manufacturer instructions. 14. Transfer the protein on a nitrocellulose membrane. 15. Stain the membrane with Ponceau S to confirm transfer. 16. Block the membrane with 5% skim milk powder in TBS for 1 h at $21 C. 17. Decorate the membrane with the anti-BirA antibody (Diluted 1:500 in 5% skim milk powder in TBS) on a shaker for 12-16 h at 4 C. 18. Wash the membrane three times with TBS. 19. Expose the membrane to the anti-mouse secondary antibody conjugated to HRP (Diluted 1:3,000 in 5% skim milk powder in TBS) for 1 h at $21 C. 20. Wash the membrane three times with TBS. 21. Incubate the membrane with ECL mix and check for correct BirA* tagging of the protein of interest.

Labeling cells with biotin and extracting crude mitochondria
Timing: 3-4 days ll

OPEN ACCESS
In this step, biotin labeling of BirA*-tagged cells and subsequently extract crude mitochondria is performed.
Note: Perform three biological replicates for each strain.
22. Grow the cell line containing the BirA*-tagged version of the protein of interest in YP + 2% glycerol or galactose supplemented with 50 mM biotin. a) Start the culture in 20 mL in YP + 2% glycerol or galactose and grow 12-16 h at 30 C under shaking. b) Increase the culture volume to 200 mL by supplementing YP + 2% glycerol or galactose. c) Bring the volume to 2 liters, add biotin powder into the culture to reach a final concentration of 50 mM and let the culture grow for $15 h at 30 C under shaking.
CRITICAL: It is important to harvest the cells for mitochondrial extraction when they are in logarithmic growth phase. The duplication time might be influenced by the strain's genetic background. In our hands, starting a 2 liters culture with an OD 600 $ 0.025-0.05 typically results on an OD $1.5 after 15 h.
23.When the OD 600 is $1.5, harvest the cells by centrifuging at 3,000 3 g for 5 min at $21 C. 24. Remove supernatant. 25. Resuspend the cells with water. 26. Spin down the cells at 2,900 3 g at $21 C for 5 min. Remove supernatant. 27. Weight the wet cell pellet, expected to be around 6 grams for a 2 liters culture harvested at OD 600 $1.5. 28. Resuspend the cells in MP1 buffer and incubate at 30 C for 10 min under shaking. 29. Spin down the cells at 2,900 3 g at 21 C for 5 min and wash them with 1.2 M sorbitol. 30. Spin down the cells at 2,900 3 g at 21 C for 5 min and resuspend them in MP2 buffer for 1 h at 30 C under shaking. 31. Disrupt the cells with a tight-fitting glass Dounce homogenizer. a) Spin down the cells at 3,000 3 g at 4 C for 5 min and resuspend them in half of the required homogenization buffer. b) Dounce 10 times. c) Spin down the homogenate for 5 min at 2,900 3 g at 4 C and save the supernatant at 4 C. d) Resuspend the pellet with the second half of homogenization buffer. e) Dounce 10 times. f) Spin down the homogenate for 5 min at 2,900 3 g at 4 C and add this supernatant to the one saved before. 32. Spin down the total supernatant at 2,900 3 g for 5 min at 4 C. 33. Transfer the supernatant in a fresh bottle. 34. Spin down the supernatant at 17,500 3 g for 15 min at 4 C and discard the supernatant. 35. The pellet contains crude mitochondrial extract. Resuspend them in SH buffer ($10 mg/mL) and fast freeze in liquid nitrogen followed by storage at À80 C.
Pause Point: mitochondria can be stored at À80 C for up to 1 year.

Purification of biotinylated proteins
Timing: 2 days In this step, total biotinylated proteins from the mitochondrial extract are purified using streptavidin magnetic beads and digested on-bead by trypsin. The peptides are then prepared for subsequent analysis by mass spectrometry.

OPEN ACCESS
STAR Protocols 1, 100219, December 18, 2020 36. Thaw mitochondria (3 mg of total protein per biological replicate) on ice or at 4 C. 37. Spin down the mitochondria for 10 min at 10,000 3 g at 4 C. 38. Remove the supernatant and solubilize the mitochondria in 300 mL of 1% SDS, and incubate for 5 min at 50 C. Please see critical point below.
CRITICAL: This step is important as it will extract and denature all the proteins. This will reduce non-specific binding of proteins to streptavidin magnetic beads (step 43).
39. Dilute the mitochondrial lysate by adding 2,700 mL of RIPA buffer supplemented with 500 U of Benzonase and incubate for 30 min on ice (4 C). 40. Clarify the mitochondrial lysate by spinning at 16,000 3 g for 5 min at 4 C. 41. Transfer the supernatant into a fresh 15 mL tube. 42. In the meanwhile, equilibrate streptavidin magnetic beads equivalent to 60 mL of beads slurry (to reach 50 mg protein/mL beads slurry) with 1 mL of RIPA buffer. 43. Spin down the beads, remove the supernatant, and transfer the equilibrated beads into the 15 mL tube with the mitochondrial lysate. 44. Incubate on a rotary shaker for 3 h at 4 C. 45. After the incubation, remove the supernatant (not bound fraction) by either centrifugation at 3,000 3 g for 5 min at $21 C or by using a magnet.
CRITICAL: Make sure all the beads have been spun to the bottom of the tube as a pellet. Presence of beads in the supernatant will give a brown tinge to the solution.
46. Add 1 mL of RIPA buffer to wash the beads. 47. Transfer the beads in a fresh 1.5 mL tube and load onto a magnetic rack.
Note: The magnet will attract the beads to one side of the tube. 57. Resuspend the beads twice in 150 mL of ABC buffer, and harvest the supernatant. 58. Merge the three supernatants together (total volume= 500 mL). 59. Ensure the high purity of the elution by keeping the tubes on the magnetic rack for 5 min, and transferring the elutions in a new tube.
Pause Point: The elutions can be stored at À80 C for weeks before proceeding to the next step.
60. Lyophilize the elutions by solvent removal in a SpeedVac vacuum concentrator. 61. Resuspend the digested proteins in 30 mL of 0.1% TFA and incubate for 10 min at 4 C. 62. Check the pH. It should be between 2 and 3, otherwise adjust with 1% TFA. 63. Prepare StageTips with 2 layers of SDB-RPS per sample by piercing the SDB-RPS extraction disk with a 16-gauge Kel-F Hub Needle (90516 from Hamilton) and pushing the plugs down with a 1 mm OD metal wire into a 200 mL pipette tip, as described previously (Rappsilber et al., 2007). Commercial alternatives are also available (Tips-RPS-M.T1.200 from Affinisep).

OPEN ACCESS
64. Pre-equilibrate the StageTips sequentially with 50 mL ACN, 50 mL of Conditioning Buffer 2 and 50 mL of Conditioning Buffer 3 by centrifugation for 5-8 min at 1,500 3 g at $21 C, or until no liquid remains in the StageTip. 65. Load each sample into a StageTip. 66. Wash the StageTips sequentially with Wash Buffer 1 and Wash Buffer 2 by centrifugation for 5-8 min at 1,500 3 g at $21 C, or until no liquid remains in the StageTip. 67. Elute the peptides with 60 mL of elution buffer and collect the peptides into a clean low protein binding tube by centrifugation for 5 min at 500 3 g at $21 C, or until no liquid remains in the StageTip. 68. Lyophilize the elutions by solvent removal in a vacuum concentrator for 30 min at 45 C, or until $2 mL remains. 69. Add 15 mL of MS loading buffer (2% ACN, 0.3% TFA) and resuspend by sonication.
Pause Point: The samples can be stored at À80 C for weeks before proceeding to the next step.

Mass spectrometry and MaxQuant analysis
Timing: 1-3 days In this step, the samples are processed by mass spectrometry and the raw data is analyzed by MaxQuant.
70. Place the samples in the LC autosampler cooled to 4-8 C and analyze 5 mL of each sample using the LC-MS setup as described in Material and Equipment. 71. Analyze the raw BioID data using MaxQuant (Cox and Mann, 2008;Tyanova et al., 2016) and perform the downstream bioinformatics analysis as described below. 72. Specify the parameters used for searching label-free BioID data using MaxQuant. Typical parameters for MQ 1.6.15.0 are described in the mqpar.xml file (https://github.com/ wasimaftab/BioID_Proteomics/Example/MQ_param). See (Tyanova et al., 2016) for more details on MaxQuant settings.
Note: The raw files of the proteomic results that we present in this manuscript was processed using MQ v1.5.2.8. Therefore, users attempting to reproduce the results with later versions of MQ might get slightly different results.

Statistical quantification and generation of volcano plots
Timing: 30-60 min In this step, two-group statistical quantification analysis is done on the Maxquant output data using a code developed using R programming. The code also generates volcano plots for easy visualization.
To implement this analysis in your workflow, refer to the Quantification and statistical analysis section for a step-by-step description.

Generation of protein-proximity networks
Timing: 30-60 min a) If you run multiple two-group comparisons, you will end up creating multiple "Results_ timestamp" folders. Inside each "Results_timestamp" folder there will be one "time stamp_final_data.tsv" file. b) Create a folder in your preferred location and copy those "timestamp_final_data.tsv" into that folder. c) Run the "automate_nw_tab_gen.R" code in RStudio by clicking Source and select the folder where you copied the .tsv files (see Example/Limma_output/Final_tsvs folder). d) The code will first ask users to specify a log fold-change cutoff. Then, for every .tsv file in that folder, it will print a list of iBAQ/LFQ column names and ask you to enter a bait name from that list. Here assumption is that iBAQ/LFQ columns will contain bait names. e) Using that information, the code will create a virtual list of bait-prey interactions that have fold change R cutoff (e.g., 1.5) and p < 0.05. f) After that, the code will print a message if that file is processed successfully. g) Finally, when all .tsv files are processed successfully, code will concatenate those virtual lists into a single one, print top 10 rows that list and save all the proximity interactions as "Links.xlsx" in the same directory where you copied .tsv files earlier (see Example/Lim ma_output/Final_tsvs/Links.xlsx). h) To visualize a bait-prey interaction network, open Cytoscape and click File / Import / Network from File. and load "Links.xlsx." i) For ''Bait'' select Meaning: ''Source Node'' and Data type: ''String.'' For ''Prey'' select Mean ing: ''Target Node'' and Data type: ''String.'' For ''Fold change'' select Meaning: ''Edge attri bute'' and Data type: ''Floating point.'' Click ''OK.'' j) Click Edit / Remove Self-Loops. to remove self-interactions involving baits. k) The network displayed represents Bait-prey undirected network which can be customized depending on the user preference.
Note: Tutorials for navigating through Cytoscape applications are available on GitHub by following this link: https://github.com/cytoscape/cytoscape-tutorials/wiki .This section of the protocol was tested on Windows 10 (64 bit) using Cytoscape 3.7.2.
Optional: Bait-prey edge table can also be generated manually in Microsoft Excel. The above code automates the process and eliminates the risk of human error.

EXPECTED OUTCOMES
Mitochondria contain proteins of dual genetic sources, and a small genome present in the mitochondrial matrix that is responsible for the production of a few key subunits of the oxidative phosphorylation system. Hence, assembly of respiratory chain complexes I, III, and IV requires expression of nuclear and mitochondrial encoded genes. To avoid accumulation of non-assembled subunits, the activities of both genetic systems need to be coordinated, which occurs by feedback loops that regulate mitochondrial translation. This translational control in mitochondria depends on the action of nuclear encoded factors called translational activators (Ott et al., 2016). The best studied translation feedback loops adjust translation of COB and COX1, two mRNAs coding for the cytochrome b subunit of complex III and subunit Cox1 of complex IV, respectively. The conserved Cbp3-Cbp6 complex interacts with newly synthesized cytochrome b to support its assembly and maturation (Gruschke et al., 2011;Gruschke et al., 2012). At the same time, the occupancy of Cbp3-Cbp6 with cytochrome b sets the levels of cytochrome b production. In cases where assembly proceeds, Cbp3-Cbp6 will dissociate from cytochrome b to stimulate a new round of COB mRNA translation (Gruschke et al., 2012;Hildenbeutel et al., 2014). To this end, free Cbp3-Cbp6 interacts with the tunnel exit of the mitochondrial ribosome to liberate COB mRNA and induce its translation . These results prompted the question whether mitochondrial ribosomes are specifically equipped with factors supporting the biogenesis of select mitochondrial encoded proteins. In order to test this hypothesis, we used the above protocol to perform C-terminal BirA* tagging for Cbp3, Mrpl4, a mitoribosomal protein located at the rim of the polypeptide tunnel exit (Gruschke et al., 2010), and Kgd4, used as control (Figure 2A). Kgd4 is a soluble protein of the alpha ketoglutarate dehydrogenase complex present in the mitochondrial matrix (Heublein et al., 2014) and not directly involved in mitochondrial gene expression. In a pilot experiment, we purified biotinylated proteins from strains containing the BirA*-tag on Mrpl4 and Kgd4 ( Figure 2B). This led to the co-purification of Kgd2, a subunit of the mitochondrial alpha ketoglutarate dehydrogenase, in the case of Kgd4-BirA*, and of Cbp3 in the case of Mrpl4-BirA*. This latter observation confirms a previous finding obtained by chemical crosslinking showing that Cbp3 is found in close proximity to Mrpl4 (Gruschke et al., 2011).
Then, we applied BioID coupled to proteomics to determine Mrpl4 and Cbp3 proximity interactomes. As previously , we used Kgd4 as a contrasting bait to eliminate background, non-specific biotinylation and performed quantitative analysis. The results are presented in the two-side volcano plots, and the hits are considered significant if they have a fold-change enrichment of R1.5 and p <0.05 ( Figures 2C and 2D). The BioID analyses of Kgd4 indeed revealed proximity to Kgd2 and Kgd1, two catalytic subunits of the alpha ketoglutarate dehydrogenase, further confirming that Kgd4 is part of the complex ( Figures 2C and 2D, left side). Moreover, the proximity interactome of Mrpl4 contained Cbp3-Cbp6 together with ribosomal proteins and other respiratory chain assembly factors ( Figure 2C, right side). Likewise, BioID analyses revealed the localization of various respiratory chain assembly factors in vicinity to Cbp3 ( Figure 2D, right side). Finally, we created a protein-proximity network by merging the two BioIDomes into a bait-prey network ( Figure 2E). This includes the significant hits from both Mrpl4 and Cbp3 BioIDs, revealing the common hits between the two baits. By combining multiple BioIDs, we can declutter complex biological processes into a comprehensive and elegant network that can aid in generating novel mechanistic insights, as previously demonstrated .

QUANTIFICATION AND STATISTICAL ANALYSIS
The code performs two-group comparison in proteomics data available in GITHUB repository: https://github.com/wasimaftab/BioID_Proteomics. Figure 3 is implemented in R programming language and the needed libraries are installed automatically. 2. Make sure you have mitochondrial proteins (identified with UniProt ID) in your proteingroups.txt file otherwise the code will terminate with a message. For the list of experimentally determined mitochondrial proteins, See Mitochondrial_Proteins folder: https://github.com/wasimaftab/ BioID_Proteomics/tree/master/Mitochondrial_Proteins.

Note:
The code can be modified for BioID analysis in other organisms, e.g., humans. To do so, replace the list of yeast mitochondrial proteins in https://github.com/wasimaftab/ BioID_Proteomics/blob/master/Mitochondrial_Proteins/ Experimentally_Determined_Mitochondrion_Proteins.txt with the most recent list of experimentally determined mitochondrial proteins for the organism (e.g., MitoCarta3.0 for humans (Rath et al., 2020)).
3. First, we recommend to install RStudio v1.3.959 or higher and R v4.0.2 or higher in Windows OS. Then, run the limma_main.R code in RStudio by clicking Source in the latest version of RStudio  Table 1: Proteins that do not participate in the two groups comparison are saved in the results folder as tsv files with their iBAQ/LFQ intensities.
5. The code is interactive, and it will help users to provide the correct names as input for bait and control as it appears in the proteingroups.txt file columns. Part of the name (case insensitive) is also accepted. It will then ask users if they want to median normalize their data prior to two-group comparison. if the median normalization is chosen, the code will boxplot (in RStudio ''Plots'' panel) the data before and after normalization. We did not choose to normalize as our replicates were highly consistent (see Figure 4). However, in case the replicate medians are not consistent it is recommended to median normalize them before performing two-group comparison. The code will also plot the data distribution (histogram) before and after missing value imputation. 6. After a successful run, the code will create a volcano plot in html format and a tsv file containing final data inside a folder called "Results_timestamp" with the current system "timestamp" in the same directory where the limma_main.R file is present. Two such sample result folders can be found in Example/Limma_output folder: https://github.com/wasimaftab/BioID_Proteomics/ tree/master/Example/Limma_output .  7. The plot can be visualized in any browser and saved as png file. 8. In addition, the code will save ''exclusively enriched'' proteins (if any) in control and bait replicates with corresponding LFQ/iBAQ values in the "Results_timestamp" folder.
Note: Since the missing values are imputed randomly (from normal distribution), users can see minute change in numbers associated with fold change and p values across multiple runs of the pipeline. However, the major patterns in the data do not change.
Note: The statistical data analysis pipeline relies on the user input for bait and control names, which should match the corresponding column names in the proteingroups.txt file. Therefore, name the samples carefully before submitting to mass spec to get a smoother data analysis.  As an example, we named the three replicates of Mrpl4 as ''Mrpl4_1'', ''Mrpl4_2,'' ''Mrpl4_3''. Then, while running ''limma-main.R,'' we entered Mrpl4 as bait name and the R code was able to acquire data for all the 3 replicates of Mrpl4 without any ambiguity.

LIMITATIONS
BioID is a powerful tool to identify proteins localized in proximity of a bait. However, the size of the BirA* tag ($38 kDa) might compromise the localization and the function of the bait protein. For example, the import in mitochondria could be affected. Hence, it is of high importance to verify the subcellular localization and the functionality of the tagged bait. Moreover, BirA* biotinylates primary amines. As consequence, if a nearby protein has no exposed primary amines, it will not be biotinylated and hence will not be detected by mass spectrometry.

TROUBLESHOOTING Problem 1
Tagging BirA* to a bait protein may disrupt its molecular functioning. By utilizing non-functional bait proteins, we may identify false positive proximity interactions and/or miss genuine interactors. Hence, the resultant network would not reflect the true physiological context and the data may be considered unreliable and misleading.

Potential solution 1
It is important to test the molecular functioning of the BirA*-tagged bait by a robust and quick assay.
In the case of Mrpl4 and Cbp3, a loss of their function would result in the loss of cellular respiration. Hence, we assayed for respiratory growth of the cell lines under non-fermentable carbon source, i.e., glycerol, and proceeded ahead with only those cell lines that displayed no growth defects.

Problem 2
While running the statistical data analysis pipeline described in "Quantification and statistical analysis", user might encounter the following error: Error in plot.new() : figure margins too large This happens when the plot pane has very little/no area to plot. Once you hit this error, code does not proceed further and gets killed.

Potential solution 2
To solve this problem, user needs to increase the area in RStudio plot pane. The issue and its potential solution are presented in a video in the following link: https://github.com/wasimaftab/BioID_Proteomics/tree/master/Troubleshooting/plot_issue.webm

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact Martin Ott (martin.ott@gu.se).

Materials availability
Yeast strains generated in this study are available by contacting Martin Ott. Plasmids used in the study are available on Addgene, ID: 160290 and 160291.

Data and code availability
The code required to perform statistical analysis on the BioID LC-MS data and to automate creation of Bait-Prey network from the analyzed datasets is available on GitHub repository: https://github. com/wasimaftab/BioID_Proteomics.