Systematic Identification of Microtubule Posttranslational Modification “Readers” by Quantitative Proteomics

Abstract

Microtubules, dynamic polymers assembled from α, β-tubulin dimers, contribute to myriad cellular processes. This is largely attributed to microtubule-associated proteins (MAPs). How MAPs selectively bind microtubules to carry out various functions is not known. The “Tubulin Code” theory proposes that posttranslational modifications (PTMs) of microtubules serve as signs that can be read by specific MAPs, thereby conferring specific functional properties to the microtubules. In support of this hypothesis, “reader” MAPs have been identified for various tubulin PTMs, but, until recently, no systematic screening had been performed to identify readers in an unbiased manner. We addressed this by developing a reader identification pipeline that uses quantitative mass spectrometry to interrogate the microtubule proteome of cells programmed to express specific PTMs. This pipeline can be used to identify readers for any tubulin PTM from various cell types as long as the writer enzymes are known. We also provide an alternative, complementary approach to obtain modified microtubules using a generic writer enzyme in vitro.

1. Introduction

Microtubules (MTs) play an essential role in a variety of cellular processes, including intracellular trafficking, cell motility, and cell division. To accomplish these tasks, MTs bind to various MT-associated proteins (MAPs), including stabilizers, crosslinkers, nucleators, severing factors, and motor proteins. Cellular MTs are not thought to be equal, and mechanisms that allow cells to differentiate between MT sub-populations within the cell have been sought. One idea, referred to as the “Tubulin Code,” proposes that specific posttranslational modifications (PTMs) on MTs cause them to recruit specific MAPs. Therefore, cataloging MAPs that “read” specific tubulin PTMs (code “readers”) is crucial to fully decipher the Tubulin Code. A handful of reader MAPs is known, but prior to our work, systematic screens for readers of any tubulin PTM had not been performed. In our studies, we focused on the tyrosination/detyrosination (Y/ΔY) cycle, the oldest and most studied tubulin PTM. During detyrosination, the C-terminal tyrosine residue of α-tubulin is enzymatically removed by the code writers (e.g., vasohibin (VASH) or MATCAP) and added back by the code eraser (tubulin tyrosine ligase or TTL) [1–5]. To identify readers of the Y/ΔY system, we established an unbiased proteomic screening pipeline to identify code readers by quantitative mass spectrometry [6].

The workflow of the reader identification pipeline is summarized in Fig. 1. This pipeline employs the widely used MT-MAPs co-pelleting assay, in which MAPs are co-pelleted with MTs assembled in cell lysates supplemented with Taxol and GTP. By applying this assay to both wild-type HeLa cells and HeLa cells overexpressing VASH1 detyrosinase, along with its obligate binding partner Small Vasohibin Binding Protein (SVBP), we generate two distinct MT samples—unmodified (tyrosinated) and detyrosinated. Theoretically, these two MT pellet fractions should preferentially contain Y- and ΔY-readers, respectively. After repeating this process three times, we obtain six MT-MAPs fractions (three each for Y and ΔY). These six samples are combined and analyzed by tandem mass tag (TMT) quantitative mass spectrometry (MS) [7]. Since a TMT-sixplex analysis includes three replicates of the MT-MAPs co-pelleting assay, the relative abundance of MAPs in either Y- or ΔY-conditions can be statistically evaluated.

Fig. 1.

Experimental workflow of the reader screening pipeline optimized for α-tubulin Y/ΔY cycle. A total of six independent MT pellet fractions are labeled with six distinct isobaric mass tags. These tags are color-coded in the figure and are denoted by their respective mass-to-charge (m/z) ratios: 126, 127, 128, 129, 130, and 131, which correspond to the different isotopic compositions allowing for multiplexed quantitation in MS analysis

The difference in the affinity of MAPs for unmodified versus modified MTs is often subtle. In our screen, even though our samples were nearly completely unmodified or modified, the difference in MT binding was minimal for many MAPs, including known Y-readers. This is thought to be because many MAPs exhibit multivalent binding to MTs, i.e., the C-terminal tails are not the only binding sites on MTs for these MAPs. Therefore, it is critical to maximize the difference in modification levels between unmodified and modified samples, ideally in binary conditions. In our Y/ΔY reader screen, we use the HeLa Kyoto cell line as the material to isolate MTs and MAPs. Since tubulin in HeLa cells is known to be largely unmodified [8], wild-type HeLa cells serve as an excellent source of unmodified tubulin/microtubules (see Note 1). To obtain a highly modified tubulin/microtubule sample, several methods should be carefully considered and optimized. In this chapter, we will first address this optimization step, then describe the co-pelleting assay, quantitative proteomics, and finally hit validation.

2. Materials

2.1. Generation of Fully Modified MTs

1. pmEGFP-N1-VASH1-IRES-SVBP plasmid [9]

2. Wild-type HeLa Kyoto cells

3. Cell culture medium: DMEM supplemented with 10% fetal bovine serum (FBS) and penicillin-streptomycin.

4. Lipofectamine 2000 transfection reagent

5. Knock-in HeLa cell line expressing VASH1-SVBP [9]

6. 2 μg/mL doxycycline

7. Tetracycline-negative FBS

8. 1 μg/mL puromycin

9. TTLΔ HeLa cell line [10]

10. Tissue culture CO₂ incubator.

11. Benchtop ultracentrifuge, e.g. Optima TM MAX-TL (Beckman Coulter).

12. Rotors TLA45 and TLA110 (Beckman Coulter).

13. Ultracentrifuge tubes.

14. Water bath (37 °C).

15. NP-40 lysis buffer: 6 mM Na₂HPO₄, 4 mM NaH₂PO₄, 2 mM EDTA, 150 mM NaCl, 1% NP40, 1 mM benzamidine, 1 mM PMSF, and 10 μg/mL each of leupeptin, pepstatin, and chymostatin.

16. Sonicator

17. Anti-Y-, ΔY-, Δ2- and a total α-tubulin antibodies

18. Carboxypeptidase A (CPA)

2.2. MT-MAPs Co-pelleting Assay

1. BRB80 buffer: 80 mM PIPES, 1 mM EGTA, 1 mM MgCl₂ pH 6.8.

2. 10× cOmplete mini EDTA free: Dissolve 1 tablet in 1 mL of BRB80 (freshly prepared).

3. 1 M DTT: Dissolve in water and store at −20 °C.

4. 10 mM Taxol: Dissolve in DMSO and store at −20 °C.

5. 100 mM MgGTP: Dissolve in water with equimolar MgSO₄, neutralize and store at −80 °C.

6. Sucrose cushion: 5% sucrose in BRB80 with 10 μM Taxol and 1 mM GTP. Warm up at 37 °C.

7. MT pellet wash buffer: BRB80 with 10 μM Taxol and 1 mM GTP. Warm up at 37 °C.

8. 8 M urea solution: Dissolve 480 mg of urea in 500 μL BRB80 and adjust the volume to 1 mL. Freshly prepared.

2.3. Quantitative Proteomics (TMT)

1. TMTsixplex Isobaric Label Reagent Set, 1 × 0.8 mg (Thermo Fisher).

2. Sep-Pak C18 cartridges (Waters).

3. 15 mM 2-chloracetamide

4. Sequencing grade modified trypsin

5. Vacufuge.

6. 0.1 M TEAB

7. Anhydrous acetonitrile

8. 5% hydroxylamine

9. 0.1% formic acid/2% acetonitrile

10. LC-MS/MS. We use a RSLC Ultimate 3000 nano-UPLC (Dionex) with a 50 cm, 75 μm i.d. C18 column (Thermo Fisher) and an Orbitrap Fusion (Thermo Fisher).

11. Proteome Discoverer (v2.3; Thermo Fisher) or equivalent software

2.4. Hit Validation

1. GFP-tagged mammalian expression plasmid for each hit.

3. Methods

3.1. Strategies to Generate Fully Modified MTs

For successful reader identification, it is important to generate MTs that are post-translationally modified to a high stoichiometry. Various approaches can be used to prepare unmodified and fully modified MTs to use as bait to isolate specific reader MAPs. For example, tubulin PTMs can be increased either by overexpressing the PTM writer or by knocking out the eraser in cells (see Note 2). Alternatively, tubulin PTMs can be promoted by incubating cell lysates with purified writer enzyme prior to (or during) MT polymerization. These approaches can be combined if one approach is not sufficient to generate high levels of tubulin PTM of interest. For the Y/ΔY reader screen, we have effectively used the writer overexpression and in vitro methods. Depending on the PTM of interest, this step requires optimization. Below, we detail our optimization step in the Y/ΔY reader screening.

3.2. PTM Upregulation in Cells

We compared the overexpression of the detyrosination writer (VASH1-SVBP) and the knockout of the eraser (TTL). For writer overexpression, we first transiently expressed a bi-cistronic VASH1-SVBP construct [9] using transfection. We confirmed the expression of VASH1-SVBP proteins and their activity in HeLa cells by detecting ΔY-tubulin in the cell lysate using Western blot analysis. Subsequently, a knock-in HeLa cell line that expresses VASH1-SVBP in a doxycycline-dependent manner was generated [9]. Western blotting of lysates prepared from cells cultured in the presence of doxycycline for 1–4 days confirmed the presence of ΔY-tubulin, similar to transient expression. Next, to test the eraser knockout approach, we knocked out TTL using the CRISPR-Cas9 system, generating a TTLΔ HeLa cell line [10]. When comparing the reduction of Y-tubulin and the increase of ΔY-tubulin, both methods of overexpressing the writer and knocking out the eraser proved to be equally effective, with an approximate 50% reduction of Y-tubulin (Fig. 2a). However, improved levels of detyrosination were observed after in vitro MT assembly, which was induced by the addition of Taxol and GTP to the lysates. Although Y-tubulin was still present in the MT fraction obtained from the TTL knockout lysate, it was undetectable in the lysate prepared from VASH1-SVBP overexpression cells (Fig. 2b), suggesting that detyrosination of Taxol-stabilized MTs occurred in the presence of VASH1-SVBP in cell lysate. The combination of in-cells and in vitro detyrosination successfully generated a fully detyrosinated MT sample. Therefore, we chose the writer overexpression method rather than the eraser knockout condition for further reader screening (see Notes 3–5).

Fig. 2.

Comparison in Y-, ΔY- and Δ2-α-tubulin levels between wild-type HeLa cells (WT), VASH1-SVBP overexpressing (OE) cells cultured with or without 2 μg/mL doxycycline for 3 days, or TTL knockout (TTLΔ) cells. Lysates (a) and the MT pellet fractions (b)

3.2.1. Writer (VASH1-SVBP) Overexpression Method

For the transfection approach, refer to steps 1 through 3. For the use of doxycycline-inducible knock-in cells, refer to steps 4 through 7.

1. Seed wild-type HeLa Kyoto cells in a well of a 6-well plate and culture in DMEM supplemented with 10% fetal bovine serum (FBS) and penicillin-streptomycin.

2. The next day, transfect cells with a bicistronic construct pmEGFP-N1-VASH1-IRES-SVBP that harbors human VASH1-mEGFP and SVBP-FLAG using Lipofectamine 2000 transfection reagent.

3. Cells should be nearly confluent and ready for lysate preparation in the next 1–2 days.

4. Generate a knock-in HeLa cell line expressing VASH1-SVBP [9]. We use a recombination-mediated cassette exchange method to generate a doxycycline-inducible cell line [11, 12] (see Note 2).

5. Seed VASH1-SVBP knock-in HeLa cells in a well of a 6-well plate and culture in DMEM supplemented with 10% tetracycline-negative FBS, penicillin-streptomycin, 1 μg/mL puromycin (see Note 6).

6. Initiate VASH1-SVBP expression by adding 2 μg/mL doxycycline in the culture medium.

7. Cells should be ready for lysate preparation 1–4 days after the addition of doxycycline (see Note 7).

8. Aspirate the culture medium, rinse cells with 1 mL of DPBS and add 300 μL of trypsin to detach cells (volume is given per well). Incubate at 37 °C for a few minutes.

9. Once cells are detached, add 1 mL of culture media to resuspend cells and transfer the whole content to a 1.5 mL Eppendorf tube.

10. Centrifuge at 170× g for 3 min at room temperature.

11. Rinse the pellet with 1 mL of DPBS and centrifuge at 500× g for 1 min at room temperature.

12. Remove the supernatant and keep the tube on ice.

13. Add 100 μL of ice-cold NP-40 lysis buffer, resuspend cells and incubate on ice for 3 min.

14. Sonicate the cell mixture.

15. Centrifuge at 21,000× g for 15 min at 4 °C.

16. Collect the supernatant as cleared whole cell lysate.

17. Measure the protein concentration using the Bradford protein assay.

18. Load 15 μg of lysate proteins per well of a 15-well mini protein gel and run an SDS-PAGE. Along with lysate proteins, load 5 μg of MT-MAPs fraction per lane. For the preparation of MT-MAPs fractions, refer to Subheading 3.3 but omit CPA in step 10.

19. Transfer to a PVDF membrane and perform western blotting using anti-Y-, ΔY-, Δ2- and total α-tubulin antibodies as well as anti-GAPDH antibody as a loading control.

3.2.2. Eraser (TTL) Knockout Method

1. Establish a TTL knockout HeLa cell line using the CRISPR-Cas9 system [10, 13]. We detected three distinct INDELs (note that HeLa cells are hypertriploid) and also confirmed the complete loss of TTL protein in cell lysate by western blotting [10].

2. Seed cells on a 6-well plate and follow steps 8–19 shown above.

3.3. Enzyme Addition In Vitro

An alternative approach to generating highly modified MTs is adding a modification enzyme to the lysate prepared from wild-type cells. Notably, tubulin modification only occurs post-lysate preparation. The advantage of this in vitro method over in-cell modification methods is that this method requires only wild-type cells, negating any requirement for genetically modified cell lines. Crucially, the preparation of two MT samples, both unmodified and modified, from the same cell lysate with an identical proteomic profile eliminates the need to check the expression level of each MAP in lysates prepared from independent cell lines.

To induce the modification of interest in vitro, it is ideal to use recombinantly prepared tubulin-specific writers. However, if applicable, generic enzymes can also promote tubulin PTMs in vitro. When generating ΔY-tubulin/MTs, we utilize Carboxypeptidase A (CPA), a potent enzyme commonly used in generating ΔY-tubulin in vitro [8]. CPA is capable of removing tubulin’s C-terminal tyrosine even on ice for brief periods, making it advantageous as it does not require warm incubation and, subsequently, minimizes undesired MT polymerization prior to full detyrosination. Since the modification process should continue after MT polymerization, we skipped monitoring the modification status in lysate and instead proceeded to the MT assembly to check the Y/ΔY state in the MT-MAPs pellet fractions. In the case of CPA-treated lysate, we detect nearly fully detyrosinated MTs in the MT-MAPs fraction (see Note 8). Again, depending on the PTMs of interest, this process needs careful optimization.

1. Culture wild-type HeLa Kyoto cells on a 2 × 150 mm dish.

2. For each dish, aspirate media, rinse cells with 4 mL DPBS, and add 2 mL trypsin to detach cells. Once cells are detached, add 8 mL DMEM to collect cells. Do the same for all dishes, combine them all, and transfer them to a 50 mL conical tube.

3. Centrifuge, 170× g for 3 min at room temperature.

4. Rinse cell pellet with 35 mL DPBS and centrifuge.

5. Resuspend the pellet in 15 mL DPBS and transfer it to a 15 mL tube.

6. Resuspend the pellet with 1 mL DPBS and transfer it to a 1.5 mL Eppendorf tube.

7. Centrifuge 500× g for 1 min at room temperature.

8. Remove supernatant and approximately 150 μL of cell pellet is recovered. Keep the tubes on ice.

9. Add 1.5× volume of ice-cold 1xBRB80 containing 1.67× cOmplete mini EDTA free protease inhibitor cocktail (final 1×) and 1.67 mM DTT (final 1 mM). For example, if the cell pellet is 150 μL, add 187.5 μL of 1xBRB80, 37.5 μL of 1× cOmplete mini EDTA free protease inhibitor cocktail and 0.375 μL of 1 M DTT.

10. Add 1/100 volume of CPA (see Note 9), Taxol (final 10 μM) and GTP (final 1 mM) to lysate on ice.

11. After 3 min on ice, incubate at 37 °C for 25 min.

12. Put the reaction mix on 100 μL of pre-warmed sucrose cushion in a 1.5 mL tube.

13. Centrifuge with TLA45 rotor at 80,000× g for 30 min at 37 °C.

14. Wash the pellet with 100 μL of warm MT wash buffer twice and resuspend the pellet in 40 μL of cold 8 M urea solution.

15. Resuspend the solution well, incubate on ice for 10 min and transfer to a new tube.

16. Measure protein concentration with Bradford protein assay.

17. Load 250 ng of protein per lane and run an SDS-PAGE.

18. Evaluate the Y/ΔY state by western blotting with Y-, ΔY-, Δ2- and total α-tubulin antibodies.

3.4. MT-MAPs Co-pelleting Assay for MS Analysis

The MT-MAPs co-pelleting assay facilitates the isolation of MTs and MAPs via centrifugation following the in vitro assembly of MTs in the lysate, based on the affinity of MAPs to MTs. Initially, whole protein lysates are prepared from cells, and subsequent incubation of the lysates in the presence of Taxol and GTP promotes MT polymerization. To minimize the contamination of unbound proteins to the MT-MAPs pellet, the assembled MT mix is placed on a sucrose cushion for ultracentrifugation. Eventually, MAPs that co-pellet with MTs are subjected to identification via MS analysis. Below, we provide the procedure for the isolation of MT-MAPs fractions using VASH1-SVBP overexpressing method (Subheading 3.2.1). If the in vitro modification method is used, scale up the procedure described above (Subheading 3.3).

1. Culture cells on 5 × 150 mm dish per condition. Use wild-type HeLa cells and HeLa cells overexpressing VASH1-SVBP (in the presence of 2 mg/mL doxycycline for 3–4 days).

2. For each dish, aspirate media, rinse cells with 4 mL DPBS, and add 2 mL trypsin to detach cells. Once cells are detached, add 8 mL DMEM to collect cells. Do the same to all dishes, combine all and transfer it to a 50 mL conical tube.

3. Centrifuge, 1000× g for 3 min at room temperature.

4. Rinse cell pellets with 35–50 mL DPBS and centrifuge.

5. Rinse with DPBS one more time.

6. Resuspend the pellet with 1 mL DPBS and transfer it to a 1.5 mL Eppendorf tube.

7. Centrifuge 500× g for 1 min at room temperature.

8. Remove supernatant and approximately 500–600 μL of cell pellet is recovered (for each condition). Keep the tubes on ice.

9. Add 1.5× volume of ice-cold BRB80 containing 1.67× cOmplete mini EDTA free protease inhibitor cocktail (final 1×) and 1.67 mM DTT (final 1 mM). For example, if the cell pellet is 600 μL, add 750 μL of BRB80, 150 μL of 1× cOmplete mini EDTA free protease inhibitor cocktail and 1.5 μL of 1 M DTT.

10. Sonicate cells for 10 s, repeat 4 times.

11. Clear lysates by centrifugation at 100,000× g for 1 h at 4 °C using the TLA110 rotor.

12. Transfer the supernatant to a new 1.5 mL tube. The volume is typically ~750 μL. Save ~50 μL for protein quantification and SDS-PAGE analysis.

13. Transfer 650 μL of lysate in a tube and add 0.65 μL of 10 mM Taxol (final 10 μM) and 6.5 μL of 100 mM GTP (final 1 mM) to initiate MT assembly.

14. Incubate at 37 °C for 25 min on a water bath (see Note 10).

15. Layer assembly mix on 500 μL of pre-warmed sucrose cushion. Centrifuge at 80,000× g for 30 min at 37°C.

16. Rinse MT pellets twice with 100 μL of warm MT wash buffer.

17. Resuspend pellets in BRB80 supplemented with 85 μL of cold 8 M urea and incubate on ice for 10 min.

18. Transfer disassembled tubulin/MAP mix to new tubes.

19. Measure protein concentration using Bradford Protein Assay.

20. Transfer MT-MAPs solution that contains a total of 25 μg of protein and adjust the sample volume to 50 μL by adding BRB80 (final protein concentration is 0.5 μg/μL).

21. Snap-freeze the samples with liquid nitrogen and store them at −80 °C until MS analysis.

22. Snap freeze and store the rest of proteins at −80 °C for the future validation step (see Subheading 3.4).

23. Repeat this assay two more times. A total of 6 samples (each 25 μg of tubulin + MAPs) should be subjected to a TMT analysis.

3.5. TMT Quantitative Proteomics

A critical part of the reader identification pipeline is the use of quantitative proteomics for identifying reader MAPs. In a label-free MS analysis, in which we simply compare the peptide spectra counts between different samples analyzed by independent MS runs, it should be noted that comparisons of protein abundance are not quantitative due to the lack of internal control. Since peptide detection can be significantly affected by various factors unique to each sample, it is crucial to combine multiple samples and analyze them in a single run.

In quantitative MS methods such as TMT or Stable Isotope Labeling by Amino acids in Cell culture (SILAC), proteins/peptides are labeled with mass tags or stable isotopes, respectively, to distinguish the peptide spectra belonging to each sample. This enables the evaluation of the relative abundance of specific peptides across different samples. Since TMT-multiplex allows the simultaneous analysis of multiple samples, implementing this approach with multiple biological replicates provides statistical interpretation of the results. The obtained hits should be evaluated based on their enrichment in either unmodified or modified samples, which can be visualized using a volcano plot (Fig. 3). To facilitate the evaluation of the results, it is important to first identify positive controls. In the case of the Y/ΔY reader screen, we first evaluated the MS result by checking known Y-readers, CAP-Gly proteins such as CLIP-170, CLIP-115 and p150^glued (Fig. 3, blue dots). MS analysis shows that these proteins pellet more efficiently with wild-type MTs than with those from VASH-SVBP overexpressing cells. This serves as a proof of principle and suggests that this pipeline may reflect the actual difference in the abundance of readers depending on the PTM states of MTs.

Fig. 3.

An example of a volcano plot showing proteins co-sedimented with Y- or ΔY-MTs in magenta or green background, respectively. X-axis represents log₂ relative abundance of each protein in VASH1 OE over wild-type samples (VASH/WT). Y-axis shows log₁₀ p value. Blue dots highlight CAP-Gly proteins

For the TMT procedure, we simply follow the instructions provided by the manufacturer. After results are obtained, it is essential to properly normalize the MS result to adjust the slightly uneven amount of total proteins in each sample. Alternatively, each score can be normalized against the total amount of tubulin in each sample. We have compared normalization methods against several factors, such as total β-tubulin (average of all β-tubulin isotypes), median β-tubulin isotype, or total protein. However, we saw the same overall trends in the MAP distribution between the unmodified and modified samples. The optimal normalization method should be determined based on specific needs. Note that we did not use the total α-tubulin abundance for normalization because the identification of α-tubulin isoforms relies on the C-terminal peptides, but due to the loss of terminal tyrosine in the modified sample, it appears as if the α-tubulin abundance has reduced significantly in the ΔY condition (VASH1-SVBP overexpression) (Fig. 3, the leftmost dot).

1. For the 6 samples generated in the MT-MAPs co-sedimentation assay, each 25 μg protein sample was digested and labeled with TMT 6-plex isobaric tags following the manufacturer’s protocol.

2. Dilute samples to 1 M urea, reduce with 5 mM DTT for 30 min at 45 °C, then alkylate with 15 mM 2-chloracetamide for 30 min at room temperature in the dark.

3. Add sequencing grade modified trypsin at a 1:25 enzyme:protein for overnight (~16 h) digestion at 37 °C with constant mixing.

4. Stop digestion by acidification, and desalt peptides using Sep-Pak C18 cartridges according to the manufacturer’s protocol.

5. Dry the samples completely using a vacufuge and reconstituted in 100 μL of 0.1 M TEAB.

6. Dissolve TMT 6-plex reagents in 41 μL anhydrous acetonitrile; each digest was added there and incubated at room temperature for 1 h.

7. Quench reactions by adding 8 μL of 5% hydroxylamine for 15 min.

8. Mix labeled samples together and dry via vacufuge.

9. Perform offline high pH reversed-phase fractionation of the combined sample into 8 fractions according to the manufacturer’s protocol.

10. Dry and reconstitute fractions in 9 μL of 0.1% formic acid/2% acetonitrile in preparation for LC-MS/MS analysis.

11. For each fraction, analyze a 2 μL aliquot by LC-MS/MS using a RSLC Ultimate 3000 nano-UPLC (Dionex) with a 50 cm, 75 μm i.d. C18 column (Thermo Fisher Cat # ES903) and an Orbitrap Fusion (Thermo Fisher). A 3-h gradient at 300 nL/min using 0.1% formic acid/acetonitrile (2–22% acetonitrile in 150 min; 22–32% acetonitrile in 40 min; 20 min wash at 90% followed by 50 min re-equilibration) was used.

12. Introduce eluent in the mass spectrometer via an EasySpray source (Thermo Fisher).

13. Set the mass spectrometer to collect MS1 scans (Orbitrap; 120 K resolution; AGC target 2×105; max IT 100 ms) followed by Top Speed MS2 scans (1 m/z isolation width, collision-induced dissociation; ion trap; NCE 35; AGC 5×103; max IT 100 ms).

14. For multinotch-MS3, the top 10 precursors from each MS2 were fragmented by HCD followed by Orbitrap analysis (NCE 55; 60 K resolution; AGC 5×104; max IT 120 ms, 100–500 m/z scan range).

15. For data analysis, use Proteome Discoverer (v2.3; Thermo Fisher) or equivalent software.

16. Search spectral files against the SwissProt human protein database using the following search parameters: Set MS1 and MS2 tolerance to 10 ppm and 0.6 Da, respectively; carbamidomethylation of cysteines and TMT labeling of lysine and N-termini of peptides were considered static modifications; oxidation of methionine and deamidation of asparagine and glutamine were considered variable.

17. Filter identified proteins and peptides to retain only those passing a 1% FDR threshold.

18. Perform quantification only using high-quality MS3 spectra (average signal-to-noise ratio >6 and <30% isolation interference).

3.6. Hit Validation

Hits are prioritized based on preferential enrichment in either Y- or ΔY- samples. Although the TMT-based quantitation reliably reflects the relative abundance of each MAP in two opposite PTM states (see Fig. 1 in [6]), a validation step using Western blot analysis is useful to check the differential binding of each candidate reader to either Y- or ΔY-MT fractions. This is particularly important when multiple isoforms are present for a reader and only specific isoform(s) show preferential binding to the PTM of MTs, as in the case of EML2 (Figs. 1 and 2 in [6]).

It is noteworthy that the expression level of some MAPs can be affected by the overexpression of VASH1-SVBP (or knockout of TTL) (see Note 5). Therefore, it is important to confirm that the protein abundance is comparable in both the wild-type and VASH1-SVBP overexpression lysates. Western blot analysis of the six lysates and MT fractions used for the MS analysis should provide a comprehensive assessment of the reader.

As a final validation, we make constructs of GFP-tagged readers for transient expression in wild-type HeLa cells, followed by the MT-MAPs co-pelleting assay using CPA to confirm each reader’s preferential binding to either Y- or ΔY-MTs. As with any other experiments, it is important to check the functionality of the GFP-tagged reader by live cell imaging or immunofluorescence. If necessary, the position (N- or C-terminal) or choice of tags should be considered. GFP-tagging interferes with MT binding of some MAPs such as EML2-S (Figure S4A in [6]). If this is the case, small affinity tags such as FLAG, c-myc, HA or PA should be considered.

3.6.1. Western Blotting Validation for Endogenous Reader Candidates (Lysate and MTs)

1. Use the 6 lysates and MT pellet fractions used for the TMT MS analysis.

2. Load 15 μg of lysate proteins and 5 μg (for MAP blots) or 200 ng (for tubulin blots) of MT pellet fractions per lane for both unmodified (WT) and detyrosinated (VASH1-SVBP) conditions and run SDS-PAGE.

3. Perform western blot and analyze the Y/ΔY state of tubulin and enrichment of reader candidates.

3.6.2. Western Blotting Validation Using GFP-Tagged Reader Candidates

1. Transfect wild-type HeLa cells with constructs of GFP-tagged reader candidates and perform MT-MAPs co-pelleting assay using CPA for the ΔY-condition.

2. Analyze the differential binding of GFP-tagged reader candidates in Y- and ΔY-MT fractions by SDS-PAGE and western blotting using anti-GFP antibody.

4. Notes

1. As tubulin PTM levels vary among different cell lines, the specific PTM profile in the cell line of choice should be assessed in each study.

2. Alternatively a stable cell line can be generated using lentiviral transduction.

3. Overexpression of writer or knockout of eraser may impact cellular activities and proliferation, or expression and PTMs of a wider variety of proteins, including tubulin. It is important to monitor cell proliferation and morphology and MT configurations (interphase and mitosis) by flow cytometry, live cell imaging or immunofluorescence.

4. It is critical to monitor tubulin PTMs in addition to the PTM of interest. In the case of ΔY upregulation, Δ2-α-tubulin was also generated (Fig. 2b), although it was negligible compared to brain tubulin, which contains abundant Δ2-α-tubulin. If other tubulin PTM levels are significantly altered, this must be taken into account when interpreting the MS result.

5. Overexpression of writer or knockout of eraser may also alter the expression level of other proteins, especially MAPs. Such a differential expression of MAPs between unmodified and modified samples will lead to biased identification of the MAPs in downstream MS analysis. Therefore, it is important to monitor the protein expression level in the lysate for each reader candidate (see Subheading 3.6).

6. Our knock-in system uses a tetracycline promoter. To prevent leaky protein expression, it is important to maintain the cell line using tetracycline-negative FBS.

7. When an inducible system is used for overexpression, it is important to optimize the induction time to maximize the tubulin PTMs. In the case of VASH1-SVBP, we compared 1–4 days after doxycycline addition and decided to use a 3 or 4-day induction period.

8. CPA also generates Δ2- and perhaps Δ3-α-tubulin. Also keep in mind that it may cause proteolytic digestion of other proteins.

9. Before use, mix CPA solution well. Do not centrifuge and remove white matter in the vial.

10. For optimum results regarding the efficient binding of specific MAP, the conditions of the MT assembly step, such as incubation time and salt concentration, may need to be customized. We compared the MT assembly incubation time ranging from 25 to 90 min and found that 25 min was sufficient for various MAPs. After longer incubation, binding efficacy decreased for some MAPs, likely due to protein degradation or gradual changes in tubulin PTM states over time.