Regulation of microtubule detyrosination by Ca²⁺ and conventional calpains

In eukaryotic cells, microtubules (MTs) are key components of the cytoskeleton. They play a crucial role in cell division, differentiation, growth, migration, polarity and intracellular trafficking. MTs mostly consist of 13 protofilaments formed through head-to-tail polymerization of α- and β-tubulin heterodimers.

The tubulin family has expanded during evolution with expression of several α- and β-tubulin isotypes [9 and 10 in humans, respectively (HGNC database, 2022; Tweedie et al., 2021)]. Both isotypes can undergo different post-translational modifications (PTMs), which label a subset of MTs and thereby act as traffic signals. PTMs regulate the stability and function of MTs, thus defining the so-called ‘tubulin code’ (Garnham and Roll-Mecak, 2012; Janke, 2014).

The removal of the C-terminal tyrosine of α-tubulin, called detyrosination, is one of the first discovered PTMs of MTs (Janke, 2014). Tubulin detyrosination occurs in a great variety of species ranging from invertebrates to humans and affects almost all α-tubulin isotypes (Janke, 2014; Redeker, 2010). In contrast to classical proteolytic cleavage, tubulin detyrosination is reversible. Upon depolymerization of MTs, the tyrosine residue can be re-ligated by tubulin tyrosine ligase (TTL), thereby returning αβ-tubulin dimers to the polymerization-depolymerization cycle. In fact, the presence or absence of the tyrosine residue has dramatic consequences for tumor progression, cardiomyocyte function and neuronal organization (Chen et al., 2018; Dubey et al., 2015; Erck et al., 2005; Heinz et al., 2017). Tyrosination and detyrosination of MTs can act as a binary on/off switch for MT function (Garnham and Roll-Mecak, 2012). Detyrosinated (detyr-) MTs are more stable, protected from kinesin-13-induced depolymerization and serve as tracks for transport, whereas tyrosinated (tyr-) MTs play a structural role (Kaul et al., 2014; Peris et al., 2009; Sirajuddin et al., 2014). The identity of the α-tubulin tyrosine carboxypeptidase (TTCP) was not known until recently when the vasohibin (VASH)/small vasohibin binding protein (SVBP) complexes were discovered as the enzymes responsible for MT detyrosination (Aillaud et al., 2017; Nieuwenhuis et al., 2017). Vasohibins are represented by two similar and functionally redundant proteins: VASH1 and its homolog VASH2. They belong to the group of transglutaminase-like cysteine proteases but contain a non-canonical catalytic triad made of cysteine, histidine and the backbone of a leucine residue (Aillaud et al., 2017). Vasohibins are widely distributed in eukaryotes, have broad tissue expression, and VASH1 is more abundant, especially in the brain (Aillaud et al., 2017).

As detyrosination has such a strong impact on MT functions, there are likely regulatory mechanisms that can control the activity of vasohibins spatially and temporally. The enzymatic activity of both vasohibins is facilitated by association with the small chaperone protein SVPB (Wang et al., 2019). However, if and how the detyrosination activity of the vasohibin proteins is regulated other than by binding to SVBP remains unknown. Furthermore, several human cell lines [e.g. human embryonic kidney cells (HEK293T), human bone osteosarcoma epithelial cells (U2OS)] depleted of vasohibins and SVBP and mice lacking SVBP still exhibit some remaining detyrosinated MTs. Although the presence of non-tyrosinated α4-tubulin could explain this, the existence of other enzyme(s) responsible for this MT modification cannot be excluded (Aillaud et al., 2017; Liao et al., 2019; Nieuwenhuis et al., 2017; Pagnamenta et al., 2019).

In this study, we aim to get new insights on processes regulating this α-tubulin PTM by investigating the effects of intracellular Ca²⁺ and the conventional Ca²⁺-dependent proteases calpain 1 and 2 on MT detyrosination both in in vitro reconstitution assays and in HEK293T cells.

HEK293T cells were previously used as a model system to study TTCPs. We first wanted to analyze the endogenous levels of detyr-α-tubulin in these cells. Basal levels of detyr-α-tubulin were low, as judged by western blotting (WB), whereas overexpression of VASH1 and SVBP resulted in a drastic increase in detyr-α-tubulin in both WB and immunocytochemistry (ICC) (Fig. 1A–C; Fig. S1). Expression of VASH1 or SVBP alone led to a mild or no increase in detyr-α-tubulin levels, respectively (Fig. 1A,B), probably due to lack of detectable endogenous VASH1 in HEK293T cells [PAXdb protein abundance database, https://pax-db.org/ (Feb. 2022) (Geiger et al., 2012; Wang et al., 2015)]. A PAXdb search revealed that SVBP can be found at the protein level in HEK293T cells. Therefore, it is possible that a small increase in detyrosination upon overexpression of VASH1 alone is the result of the reconstitution of the VASH1–SVBP complex. Thus, HEK293T cells display a large dynamic range of detyr-α-tubulin, represent a sensitive system to study changes in MT detyrosination, and can be used as a good cellular model for dissecting MT detyrosination pathways. As many signaling pathways respond to changes in intracellular Ca²⁺, we first tested if removal of Ca²⁺ by using the cell-permeable Ca²⁺ chelator BAPTA-AM affects α-tubulin detyrosination in HEK293T cells. Although the endogenous detyrosination levels were low, a strong reduction of more than 50% of MT detyrosination was observed after 4 h of BAPTA-AM treatment (Fig. 1D,E). Taxol-induced stabilization of MTs did not influence the reduction in detyrosination caused by BAPTA-AM application, indicating that this change occurs on MTs rather than tubulin dimers. As TTL acts on soluble tubulin dimers, the observed decrease in detyrosination likely results from diminished TTCP activity.

Previous work suggested that the Ca²⁺-dependent non-lysosomal cysteine protease calpain 1 can cleave VASH1, thereby altering its angiogenic activity in the extracellular space (Saito et al., 2016). Calpain 1 and 2 (CAPN1 and CAPN2, respectively) are the conventional and ubiquitously expressed calpain family members. They play an important role in cell signaling and cytoskeleton dynamics due to limited proteolysis of their substrates – both in physiological and pathological conditions (Moldoveanu et al., 2002). As they are Ca²⁺-dependent and intracellular Ca²⁺ concentration can rapidly change depending on the physiological status of a cell, we set to test whether calpains might be involved in MT detyrosination. We performed knockdown (KD) experiments using calpain 1 and 2 siRNAs (Fig. 2). The specificity of the KD in HEK293T cells was confirmed by WB analysis after 24 h of transfection (Fig. 2A; Fig. S2). Co-transfection of VASH1–GFP and SVBP with either calpain 1 or 2 siRNAs led to a significant decrease in detyr-MTs compared to detyr-MT levels in scrambled control siRNA (Fig. 2B,C). The combination of both siRNAs turned out to be less efficient, probably due to the reduced amount of the individual siRNAs. These data indicate that the conventional calpains play a role in the regulation of the MT cytoskeleton, and that the presence of calpains is required for efficient MT detyrosination.

Calpains are endopeptidases, and predictions for calpain cleavage sites indicate that calpains require longer stretches within the substrate sequence (Sorimachi et al., 2012). Thus, it is unlikely that calpains detyrosinate α-tubulin themselves, but rather act upstream of a TTCP. Although VASH1 was previously proposed to be a substrate of calpain 1 (Saito et al., 2016), the consequences of VASH cleavage on the detyrosination of MT are completely unknown. Therefore, we performed a set of cell-free assays to investigate whether calpain 1 could indeed cleave VASH1 or VASH2, and if this impacted VASH activity on MTs (Fig. 3).

To isolate VASH1 or VASH2 independent of N- or C-terminal cleavage, we decided to co-purify them using SVBP–Myc as a bait, as the SVBP-binding site lies within the central region of VASH1 and VASH2, shown previously by co-immunoprecipitations of truncated constructs and solved protein structures (Kadonosono et al., 2017; Liao et al., 2019; Wang et al., 2019; Zhou et al., 2019). After on-bead calpain 1 digestion and subsequent separation on an sodium dodecylsulfate (SDS) gel, prominent bands were excised and used to identify potential new N-and C-termini (Fig. 3A) using liquid chromatography-mass spectrometry (LC-MS) analysis (Deng et al., 2015; Schnölzer et al., 1996). While new C-termini could only be identified within the linker region and GFP (data not shown), several new N-termini were found. For VASH2, an N-terminal stretch (aa 36–47) was identified, within which several almost ladder-like distributed cleavage sites were localized (Fig. 3B; Tables S3 and S4). This amino acid stretch lies in the N-terminal disordered region in both human and mouse vasohibins. For VASH1, more distinct cleavage sites between residues 19A and 20A and between 24T and 25A were identified in both the chymotrypsin and trypsin digestion experiments and with a large number of peptide spectral matches (PSMs) (Fig. 3B; Tables S1 and S2). These positions are in agreement with the cleavage specificity of calpain 1 according to the MEROPS peptidase database (https://www.ebi.ac.uk/merops/) (Rawlings et al., 2017), in which calpain prefers alanine in the P1′ site (Fig. S3A). Another less abundant cleavage site, between S14 and A15, was also identified in digestion experiments. This site contains an alanine residue in the P1′ position and a proline residue in the P3′ position, which is in line with calpain's cleavage preference (Fig. S3A). Moreover, recently published structures of vasohibins suggest a disordered N- and C-terminus (Adamopoulos et al., 2019; Ikeda et al., 2020; Li et al., 2019, 2020; Liao et al., 2019; Liu et al., 2019; Wang et al., 2019; Zhou et al., 2019), and therefore, the alanine-rich N-terminus would be accessible to calpain, which prefers interdomain unstructured regions in its substrate binding pocket (Hosfield et al., 1999; Moldoveanu et al., 2002; Sorimachi et al., 2012). Furthermore, we cannot exclude that calpain 1 additionally cleaves SVBP, as the SVBP band was not detectable after calpain 1 digestion on the SDS gel (Fig. 3A,B). However, it was not possible to identify newly generated N- or C-terminal ends of SVBP in the MS analysis. Of note, previously reported extracellular cleavage corresponding to amino acids 86–87 and 328–329 in the mouse protein (Sonoda et al., 2006) could not be confirmed, suggesting a differential cleavage and functional regulation of vasohibins intracellularly and extracellularly. As VASH1 was previously suggested to be a calpain substrate, we decided to further focus on VASH1.

After identifying these new cleavage sites, we investigated the functionality of cleaved VASH1 on MT detyrosination by performing a cell-free MT detyrosination assay (Fig. 3C–F; Fig. S3), as our cellular data suggested an effect of calpain on VASH1. Polymerized MTs were prepared from porcine brain tubulin. Interestingly, although it has generally been believed that tyr-α-tubulin, in contrast to detyr-α-tubulin, can easily incorporate into MTs and therefore two cycles of polymerization and depolymerization should lead to an enrichment of tyr-α-tubulin, we could not find experimental evidence for this (Fig. S3B,C). Next, MTs were incubated with purified Twin-Strep-tag-VASH1–GFP and SVBP, with or without prior calpain 1 digestion, for different time intervals ranging from 5 to 30 min (Fig. 3C–F). WB analysis confirmed that the presence of calpain 1 led to the cleavage of VASH1, as the full-length VASH1 band was no longer detectable (Fig. 3D) and truncation products could be detected by WB (Fig. S3E). As expected, the addition of VASH1 led to a strong reduction in tyr-α-tubulin and an increase in detyr-α-tubulin (Fig. 3E,F). However, this effect was largely independent of prior VASH1 cleavage by calpain 1, indicating that calpain 1 cleavage of VASH1 does not affect its MT-detyrosinating catalytic abilities (Fig. 3E,F). This experiment can also rule out the possibility that calpain 1 itself cleaves the C-terminal tyrosine of α-tubulin. Treatment with a tenth of the concentration of VASH1 (Fig. S3D–G) showed only a subtle decrease of tyr-α-tubulin over time, and again, prior calpain 1 cleavage did not have an effect. Thus, the possibility of oversaturation in the experimental setup could be excluded.

To further confirm these findings in a cellular context, we cloned and overexpressed different VASH1 truncation constructs in HEK293T cells and quantified MT detyrosination after immunostaining: 87–375-VASH1 and 87–328-VASH1, representing the previously described ΔN and ΔNΔC VASH1 upon extracellular cleavage (Sonoda et al., 2006) and the newly identified 20–375- and 25–375-VASH1 truncations (Fig. 4A). Expectedly, full-length VASH1 led to a strong increase in MT detyrosination, whereas 87–375-VASH1 and 87–328-VASH1 were significantly less efficient, indicating that N- and C- terminal truncation renders the protein functionally less active in the context of MT detyrosination (Fig. 4B,C). Importantly, overexpression of VASH1 (20–375 and 25–375) truncation constructs, representing the newly identified calpain 1 cleavage products, led to the same increase in MT detyrosination as the full-length construct (Fig. 4B,C).

Our in vitro results suggest that the N-terminal cleavage of VASH1 mediated by calpain 1 does not change the catalytic activity of VASH1 (Fig. 3). In HEK293T cells, which do not express endogenous VASH1, overexpression of calpain 1 cleavage-mimicking truncated VASH1 constructs versus overexpression of full-length VASH1 did not show direct effects of calpain cleavage of VASH1 on MT detyrosination activity. However, as this pathway is not endogenously available in HEK293T cells, a fine regulating effect of calpain cleavage of VASH1 and/or VASH2 under physiological conditions cannot be ruled out. This regulation could potentially be mediated by changes in recruitment of VASH1 to MTs either directly or via MT-associated proteins (MAPs), as suggested for VASH2 and MAP4 (Yu et al., 2021). HEK293T cells express α4-tubulin [https://pax-db.org/ (Geiger et al., 2012; Wang et al., 2015)], which lacks the C-terminal tyrosine, and this could contribute to the pool of detyrosinated tubulin without the necessity of being cleaved. However, if all detyr-α-tubulin comes from α4-tubulin, then it would likely randomly incorporate into different MTs without labeling a specific subset of MTs, as the lack of C-terminal tyrosine does not seem to affect dimer incorporation into MTs (Chen et al., 2021). HEK293T cells have individual MTs which are continuously decorated with detyr-α-tubulin (Fig. 1C). Therefore, it is probable that another Ca²⁺- and/or calpain-dependent pathway upstream of a yet unknown TTCP exists.

Taken together, we confirmed that SVBP is a potent activator of VASH1 activity, whereas Ca²⁺ and conventional calpains play a role in fine-tuning MT detyrosination. Future work could shed light on how conserved and ubiquitous these regulations are, and whether Ca²⁺ and calpains can dynamically change the cellular MT detyrosination landscape depending on local processes and the activity status of the cell.

All used plasmids and cloned constructs are listed in Table S5. Cloning of SVBP–sfGFP–His construct was performed by amplification of SVBP by PCR from a construct with SVBP–Myc–Flag (which was a gift by M. Moutin, Grenoble Institut des Neurosciences) with primers containing restriction sites for XbaI and NotI. In the plasmid Flag–full-length-VASH1–sfGFP–His (gift by M. Moutin, Grenoble Institut des Neurosciences), VASH1 was replaced by SVBP using XbaI and NotI. For cloning of the mouse VASH1 truncation constructs, the respective nucleotides were amplified via PCR from the full-length construct with primers containing the restriction sites XbaI and AgeI, and cloned back into the full-length backbone using standard restriction digestion and ligation cloning protocols. The constructs containing the internal ribosome entry sites (IRES), VASH1-87–375–sfGFP–His IRES SVBP–Myc and VASH1-87–328–sfGFP–His IRES SVBP–Myc, correspond to the N-terminally cleaved 36 kDa and the N- and C-terminally cleaved 27 kDa form of human VASH1 from Sonoda et al. (2006). The Twin-Strep-tag–VASH1–GFP IRES SVBP construct was obtained by cold fusion cloning of a gBlock containing the Twin-Strep-tag (the Twin-Strep-mCherry empty vector was a gift from Amol Aher and Anna Akhmanova, Utrecht University, The Netherlands) into the Bsu15I digested full-length VASH construct.

HEK293T cells were grown in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, #41966-029) with 10% fetal calf serum (FCS; Thermo Fisher Scientific, #10270106) and 1× penicillin-streptomycin (Thermo Fisher Scientific, #15140122) at 37°C in 5% CO₂ and 95% humidity, and split twice per week. Cells for immunochemistry were plated on poly-L-lysine (Sigma P2636)-coated glass coverslips, for biochemistry directly in six-well plates. HEK293T cells were authenticated and tested for contamination by the supplier and used within 35 passages.

Cells were transfected using maxPEI or Lipofectamine 2000 (for siRNA experiments) using the manufacturers' protocols 1 day after splitting, and were then fixed or harvested after 24 h. For siRNA experiments, 15 pmol siRNA was used in combination with 0.6 µg DNA per well in a 12-well plate (1 ml volume). Only 0.3 µg pmaxGFP as a transfection control was used. Amounts were accordingly adjusted for use in six-well plates. Treatments of HEK293T cells were performed so that all groups were fixed and/or harvested at the same time. Cells were treated with 50 µM BAPTA-AM for 4 h with the addition of 1 nM taxol for the last 30 min. Details of suppliers and catalog numbers are shown in Table S5.

Immunocytochemistry (ICC) on HEK293T cells was performed as described previously (Bär et al., 2016). Briefly, cells were fixed with 4% paraformaldehyde (PFA) with 4% sucrose, washed in PBS, permeabilized in 0.2% Triton X-100 in PBS and blocked in blocking buffer (BB; 10% horse serum, 1% Triton X-100 in PBS). Primary antibodies were diluted in BB (anti-detyr-α-tubulin, 1:400; anti-α-tubulin, 1:600) and coverslips incubated overnight at 4°C. After additional washing with PBS, the corresponding fluorescently labeled secondary antibodies (all diluted 1:500 in BB) were applied for 1–2 h at room temperature, before final washing with PBS, incubation with 1 µg/ml DAPI and mounting of coverslips on objective slides with Mowiol. Details of suppliers and catalog numbers are shown in Table S5.

Cell homogenates/extracts of HEK293T cells were prepared as follows: cells were shortly washed with warm PBS, harvested in Tris-buffered saline (TBS; 20 mM Tris-HCl 150 mM NaCl, pH 7.4) containing 1% Triton X-100 and protease inhibitor cocktail (PI). Hot 4× SDS sample buffer [250 mM Tris-HCl, pH 6.8; 8% (w/v) SDS; 40% (v/v) glycerol; 20% (v/v) β-mercaptoethanol; 0.008% bromophenol blue, pH 6.8] was added directly and samples were boiled for 5 min. Equal amounts of samples were separated on a 10% polyacrylamide gel and transferred to a PVDF membrane. For BAPTA-AM experiments, cells were centrifuged at 1000× g for 3 min at 4°C, the supernatant discarded, cells resuspended in TBS and processed as described above.

The membranes were blocked for 1 h in 5% milk in TBS with 0.1% Tween-20 (TBS-T) at room temperature and incubated overnight at 4°C with primary antibodies in TBS with 0.02% NaN₃ (anti-detyr-α-tubulin, 1:1000; anti-β-actin, 1:5000; anti-tyr-α-tubulin, 1:1000; anti-calpain 1, 1:200; anti-calpain 2, 1:200; anti-VASH1, 1:100; anti-GAPDH, 1:200). After washing in TBS-T and TBS successively, the membranes were incubated with secondary antibodies in TBS-T with 5% milk for 1–2 h at room temperature, washed and processed for signal detection using ECL solution on an INTAS ChemoCam Imager (Intas Science Imaging) or BioRad VersaDoc device. Detyr-α-tubulin, tyr-α-tubulin and total α-tubulin were detected on different membranes, and the loading controls (actin and GAPDH) were detected on identical membranes and used for normalization. The Fiji tool ‘Analyze gels’ (Schindelin et al., 2012) was used to quantify band intensities. Details of suppliers and catalog numbers are shown in Table S5.

Widefield imaging was performed at a Nikon Eclipse Ti-E microscope controlled by VisiView software and equipped with standard GFP, RFP and Cy5 filters. Illumination was achieved by an LED light source. Use of the 100× (Nikon, ApoTIRF 1.49 oil), 60×1.4 NA (Nikon, CFI Plan Apo Lambda Oil) or 40×1.3 NA (Nikon, CFI Plan Fluor Oil) objectives for imaging of transfected HEK293T or neurons yielded pixel sizes of 65 nm, 108 nm or 162.5 nm, respectively. Images were taken at 16-bit depth and 2048×2048 pixel size on an Orca flash 4.0LT CMOS camera (Hamamatsu). In part, several images on different z-positions were taken to cover the entire cells and maximum projections were calculated for representation.

HEK293T cells were transfected with Flag–VASH1–GFP and SVBP–Myc constructs or with Flag–VASH2–GFP and SVBP–Myc constructs using maxPEI. After 24 h, cells were washed with cold TBS (20 mM Tris-HCl, 150 mM NaCl, pH 8.0), and harvested in extraction buffer (20 mM Tris-HCl, 150 mM NaCl, 0.5% Triton X-100, 2× PI, pH 8.0). Extraction was performed for 30 min on ice, followed by centrifugation at 14,000 g at 4°C and the supernatant was collected. Myc-Trap Agarose beads A (Chromotek, yta-2) were prepared by three rounds of washing in extraction buffer and centrifugation at 1000 g for 5 min at 4°C. Beads were incubated with HEK293T cell extracts at 4°C overnight at slow rotation. The unbound fraction was removed by centrifugation at 1000 g for 5 min at 4°C. The beads were washed three times in extraction buffer and equally split into two samples for on-bead calpain digestion and control. The beads were then washed three times with TBS to remove protease inhibitors and washed once with TBS containing 2 mM CaCl₂. 50 µl TBS containing CaCl₂ (all components being 1.2× concentrated) was added to both samples. In vitro calpain cleavage was achieved by adding 10 µl calpain (>20 U, resulting in 1× concentration of TBS) and incubation at 30°C for 5 min. The reaction was stopped by addition of 20 µl hot 4× SDS buffer, and the sample was boiled for 5 min. The control sample was treated equally except that 10 µl H₂O was added instead of calpain. Samples were separated on a commercial 4–12% Bis-Tris Plus SDS gel (Thermo Fisher Scientific, #NW04120BOX).

Gel bands were cut into 2 mm³, destained using 100 mM ammonium bicarbonate (ABC) and 30% acetonitrile (ACN) in 50 mM ABC buffer, followed by dehydration with ACN and then dried using vacuum centrifugation. Samples were reduced with dithiothreitol (10 mM) in triethylammonium bicarbonate (TEAB) buffer (pH 8.5) for 30 min at 60°C and alkylated with iodoacetamide (55 mM) at room temperature for 15 min.

The gel pieces were washed with 30% ACN in HEPES buffer (25 mM) and then dehydrated with ACN. After drying, the amino groups (N-terminus and lysine residues) were reductively dimethylated using 40 mM formaldehyde and 20 mM sodium cyanoborohydride in HEPES buffer (250 mM, pH 7) at 37°C for 16 h. The reaction was quenched using 0.9 M of Tris buffer (pH 6.8) for 3 h followed by multiple washing steps, i.e. 50 mM ABC in 30% ACN (15 min), 100% ACN (15 min) and drying in a vacuum centrifuge (10 min). After quenching, the gel bands were incubated with either trypsin (50 ng) or chymotrypsin (100 ng) in the presence of 50% H₂¹⁸O (heavy water). Trypsin samples were incubated with 25 mM ABC, whereas chymotrypsin samples also included 2 mM CaCl₂ in the digestion buffer. Samples were incubated overnight at 37°C and after digestion, the reaction was quenched using 10% formic acid (FA). The peptides from the gel bands were extracted with 50% ACN in 1% FA and 100% ACN using sonication. Pooled extracts were dried down and the samples resuspended in 3% ACN and 0.1% trifluoroacetic acid (TFA) prior to LC-MS.

In-gel digested samples were analyzed on a Dionex Ultimate 3000 nano-UHPLC coupled to a Q Exactive mass spectrometer (Thermo Fisher Scientific, Bremen). The samples were washed on a trap column (Acclaim Pepmap 100 C18, 5 mm×300 μm, 5 μm, 100 Å, Dionex) for 4 min with 3% CAN and 0.1% TFA at a flow rate of 30 μl/min prior to peptide separation using an Acclaim PepMap 100 C-18 analytical column (50 cm×75 μm, 2 μm, 100 Å, Dionex). A flow rate of 300 nl/min using eluent A (0.05% FA) and eluent B (80% ACN and 0.04% FA) was used for gradient separation. The spray voltage applied on a metal-coated PicoTip emitter (10 μm tip size, New Objective, Woburn, MA, USA) was 1.7 kV, with a source temperature of 250°C. Full-scan MS spectra were acquired between 300 and 2000 m/z at a resolution of 70,000 at m/z 400. The ten most intense precursors with charge states greater than 2+ were selected with an isolation window of 2.1 m/z and fragmented by higher energy collisional dissociation (HCD) with normalized collision energies of 27 at a resolution of 17,500. Lock mass (445.120025) and dynamic exclusion (15 s) were enabled.

The MS raw files were processed by Proteome Discover 2.2 (Thermo Fisher Scientific, version 2.2.0.388) and MS/MS spectra were searched using the Sequest HT algorithm against a database containing common contaminants (45 sequences), the canonical human database and predicted peptides for recombinant Flag–VASH1–GFP–His. For trypsin, the enzyme specificity was set to semi-Arg-C with two missed cleavages allowed. For samples incubated with chymotrypsin, no enzyme specificity was used. An MS1 tolerance of 10 ppm and an MS2 tolerance of 0.02 Da was implemented. Oxidation (15.995 Da) of methionine residues, dimethylation (28.031 Da) on the peptide N-terminus and heavy water (2.004 Da) on the peptide C-terminus were set as a variable modification. Carbamidomethylation (57.021 Da) on cysteine residues, along with dimethylation on lysine residues were set as a static modification. The minimum peptide length was set to six amino acids and the peptide false discovery rate (FDR) was set to 1%. Peptide peak intensities were calculated using the Minora algorithm. Peak intensities were only used if they were identified with high confidence.

Abundances of all the N-termini that were identified with high confidence for bands 2,3,4 and 1 (control) for both chymotrypsin (Table S1) and trypsin (Table S2) digestions of VASH1 are provided. Similar data for VASH2 digestion with chymotrypsin (Table S3) and trypsin (Table S4) are also shown. Samples were analyzed from the lowest molecular mass to highest molecular mass, with the control being measured last. Unless noted otherwise, all peptides shown have their N-terminus and their lysine residues are dimethylated. The large number of N-termini identified in the control sample may be due to degradation of VASH1 or insufficient quenching of labeling reagents prior to enzymatic digestion. N-termini were only considered to be calpain cleavage events if they were identified in both the chymotrypsin and trypsin experiments and were also not observed in the calpain control. The most probable calpain cleavage sites, based on intensity and number of PSMs, are highlighted.

HEK293T cells were transfected at 60–70% confluency using maxPEI with Twin-Strep-tag–VASH1–GFP IRES SVBP construct 24 h before purification. Cells were washed once with ice-cold TBS and harvested in ice-cold TBS subsequently. The cells were centrifuged at 1000 g for 3 min at 4°C and resuspended in extraction buffer [20 mM Tris-HCl, 300 mM NaCl, 1% Triton X-100, 5 mM MgCl₂, 1× cOmplete^TM, EDTA-free Protease Inhibitor Cocktail (Sigma-Aldrich), pH 8.0] and incubated for 30 min on ice. The cell lysates were centrifuged at 14,000 g for 15 min at 4°C. The supernatants were collected and incubated with Strep-Tactin Sepharose High Performance beads (GE Healthcare) for 1 h at 4°C on slow rotation. The beads were centrifuged at 1000 g for 1 min at 4°C and the cell lysate was removed. Beads were washed three times with washing buffer (20 mM Tris-HCl, 150 mM NaCl, 0.5% Triton X-100, 2 mM EGTA, 1× cOmplete^TM, EDTA-free Protease Inhibitor Cocktail, pH 8.0). Elution was performed for 10 min in elution buffer (100 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 50 mM D-Biotin, 1 mM DTT) followed by centrifugation at 1000 g for 1 min.

The purified protein was stored on ice until further use. Purification was controlled by adding 5 µl (Fig. 3) or 20 µl (Fig. S3) of 2× SDS buffer to 5 µl (Fig. 3) or 20 µl (Fig. S3) of the first elution, boiling the sample for 5 min and running SDS-PAGE with a bovine serum albumin (BSA) standard on a 10% SDS-polyacrylamide gel. The gel was stained with Coomassie Brilliant Blue R250 (Roth) for 30 min at room temperature and destained with MilliQ water overnight at room temperature.

1 µM of purified Twin-Strep-tag–VASH1–GFP was incubated with 5 µl calpain 1 (>10 U) in 100 mM TBS with 2 mM CaCl₂ for 5 min at 37°C. Afterwards, the samples were put on ice immediately and used straight away.

The calpain 1 cleavage of Twin-Strep-tag–VASH1–GFP was analyzed using a 12% SDS-polyacrylamide gel and followed by WB. Blocking was done for 30 min with 5% milk powder in TBS-T. The primary antibodies used were anti-calpain 1 (1:100 in TBS with 0.02% NaN₃) and anti-VASH1 (1:100 in TBS with 0.02% NaN₃). The secondary antibody used was anti-mouse-HRP (1:20,000 in TBS-T). The blots were developed using ECL solution on an INTAS ChemoCam or Bio-Rad VersaDoc Imager. Details of suppliers and catalog numbers are shown in Table S5.

Two rounds of MT polymerization and cold-induced depolymerization were performed. A mix of 20 µM of porcine brain tubulin protein (Cytoskeleton/Tebu-Bio) with 1 mM GMPCPP (Jena Bioscience) in PEM80 (80 mM Pipes, 1 mM EGTA, 4 mM MgCl₂) was made on ice and subsequently incubated at 37°C for 30 min for the production of MT seeds. The sample was centrifuged at 120,000 g for 5 min at 25°C to remove polymerized MTs from solution. The supernatant with the unpolymerized tubulin was discarded, and the pellet was resuspended in PEM80 to about 20 µM tubulin assuming 80% recovery. The resuspended tubulin was incubated for 20 min on ice to depolymerize the MTs. Then, GMPCPP was added to a concentration of 1 mM and incubated 5 min more on ice and the process was subsequently repeated. These cycles of polymerization and depolymerization were thought to result in tyr-α-tubulin enrichment. However, WB analysis of cycled and non-cycled tubulin showed no difference in the amount of detyr- or tyr-α-tubulin (Fig. S3A,B). Therefore, for experiments with 10 nM VASH1 concentration, MTs were just polymerized without prior rounds of polymerization and depolymerization.

After the second round of MT polymerization and depolymerization, the sample was incubated for 30 min at 37°C to obtain MTs. Purified Twin-Strep-tag–VASH1–GFP (1 µM) was incubated with calpain 1 (>10 U) in 100 mM TBS with 2 mM CaCl₂ for 5 min at 37°C. 0.1 µM or 10 nM of this in vitro calpain 1-cleaved Twin-Strep-tag–VASH1–GFP, 0.1 µM or 10 nM of purified Twin-Strep-tag–VASH1–GFP or elution buffer without protein as a negative control were added to MTs and incubated for 5, 10, 20 or 30 min (and 45 min for 10 nM VASH1) at 37°C. The reaction was stopped by addition of hot 4× SDS buffer and samples were boiled immediately for 5 min.

Equal amounts of the samples were analyzed by SDS-PAGE with either 10% or 12% SDS-polyacrylamide gel, followed by WB. Blocking was done for 30 min with 5% milk powder in TBS-T. The primary antibodies used were anti-detyr-α-tubulin (1:1000 in TBS with 0.02% NaN₃), anti-tyr-α-tubulin (1:1000 in TBS with 0.02% NaN₃) and anti-α-tubulin (1:2000 in TBS with 0.02% NaN₃). The secondary antibodies used were anti-rabbit-HRP (1:20,000 in TBS-T) and anti-mouse-HRP (1:20,000 in TBS-T). The blots were developed using ECL solution on an INTAS ChemoCam or BioRad VersaDoc Device. Band intensity integration areas were analyzed using the ‘Analyze gels’ tool of Fiji (Schindelin et al., 2012). The ratio of tyr-α-tubulin to anti-α-tubulin was calculated and normalized to the corresponding control. Details of suppliers and catalog numbers are shown in Table S5.

The detyr-α-tubulin/α-tubulin ratios in immunostainings were analyzed as follows: cells were outlined by using the total tubulin channel, that was after strong smoothing and despeckling, semi-automatically thresholded and manually corrected to exclude dividing cells, dense cell clusters and parts of cells on the edge of the field of view. Afterwards, the GFP channel was smoothed and despeckled and used to identify transfected cells in the same way, and cell outlines were corrected with the help of the total tubulin channel. Subtraction of the total cell outline and the transfected cell outline was calculated to obtain the outline of non-transfected cells. Total tubulin and detyr-α-tubulin average intensity values within the transfected and untransfected cells were measured in the original images. For each image, the detyr-α-tubulin/α-tubulin ratio of transfected and untransfected cells was calculated, and afterwards, intensities of transfected cells were normalized to untransfected cells. Data from several independent experiments were normalized to control, before pooling data. An experimenter blinded to the experimental groups performed all analyses.

Statistical analysis was performed using Prism version 6.07 and 8 (GraphPad). Data are represented as mean±s.e.m. Data were tested for normal distribution (D'Agostino and Pearson omnibus normality test) and the statistical test was chosen accordingly. Individual channels of multi-color micrographs were contrasted for better representation, with identical settings within each experiment. All analysis was performed on raw images.

We would like to thank Lisa Mallis for excellent technical support, Daniela Hacker for help with initial antibody characterizations, Marie-Jo Moutin for kindly providing the Flag–full-length-VASH1–sfGFP–His, SVBP–Myc–Flag, Flag–full-length-VASH1–sfGFP–His IRES SVBP–Myc and Flag–full-length-VASH2–sfGFP–His IRES SVBP–Myc constructs, as well as Amol Aher and Anna Akhmanova for kindly providing the Twin-Strep-mCherry cloning vector.