The long form of pVHL is artifactually modified by serine protease inhibitor AEBSF

Abstract

von Hippel–Lindau protein (pVHL) is the tumor suppressor responsible for ubiquitylating the hypoxia‐inducible factor (HIF) family of transcription factors for degradation under normoxic conditions. There are two major pVHL isoforms with the shorter isoform (pVHL₁₉) lacking the acidic domain present in the N‐terminus of the longer isoform (pVHL₃₀). Although both isoforms can degrade HIF and suppress tumor formation in experimental systems, previous research suggests that pVHL₃₀ can undergo posttranslational modifications (PTM) and interact with unique proteins. Indeed, pVHL₃₀ has long been observed to migrate as two species on a reducing polyacrylamide gel, indicating the presence of an uncharacterized PTM on the slower‐migrating pVHL₃₀ without an identifiable biological consequence. Thus, there has been considerable effort to elucidate the exclusive biological activity of pVHL₃₀, if any, by first defining the unique features of the slower‐migrating species. We show here that the migration of pVHL₃₀, but not pVHL₁₉, is retarded by 4‐(2‐aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF), an irreversible serine protease inhibitor commonly found in protease inhibitor cocktails.

1. INTRODUCTION

von Hippel–Lindau protein (pVHL) is the substrate‐conferring F‐box component of an E3 ubiquitin ligase complex. The best characterized target of this complex is the alpha subunit of the hypoxia‐inducible factor (HIF) transcription factor. Under normoxic conditions, HIFα is rapidly hydroxylated by prolyl hydroxylase (PHD) enzymes, which use molecular oxygen as a co‐substrate. ¹ Hydroxylated HIFα is then bound by pVHL and targeted for degradation via the ubiquitin‐proteasome pathway. ² , ³ Under hypoxic conditions, HIFα escapes recognition via pVHL and is consequently stabilized to dimerize with HIFβ to form a functional transcription factor that positively regulates factors governing cellular response and adaptation to hypoxia such as glycolysis, erythropoiesis and vasculogenesis. Mutations in the components of the HIF pathway, including pVHL, PHD2, and HIF2α, underlie numerous cancerous and non‐cancerous pathologies. ⁴ , ⁵ , ⁶ , ⁷

There are two isoforms of pVHL. pVHL₃₀ represents the full‐length amino acid chain (1–213) encoded by the VHL gene. pVHL₁₉ is a shorter isoform that is translated from an internal methionine (54–213). ⁸ , ⁹ It is still unclear whether there is a functional significance to having two isoforms. Both pVHL₃₀ and pVHL₁₉ possess the alpha domain necessary for forming an E3 complex and the beta domain that mediates HIFα binding. ¹⁰ , ¹¹ Further, it has been shown that both isoforms can regulate HIFα levels and suppress tumor growth. ⁸ , ⁹ The first 53 amino acids of pVHL₃₀, absent in pVHL₁₉, encode a repetitive stretch of acidic residues. On the basis of multiple sequence alignments, it appears that the long form of pVHL lacks the acidic domain in non‐mammalian species. ¹² Within kingdom mammalia, different orders possess a different number of acidic repeats; primates (7–8 units), cingulata (5–6 units), artiodactyla (4–5 units), proboscidea (2–3 units), and lagomorpha (1 unit). ¹²

Several unique properties have been attributed to pVHL₃₀ and, by extension, to this acidic domain. First, several studies have suggested that pVHL₃₀ is localized primarily to the cytoplasm and pVHL₁₉ to the nucleus. ⁸ , ¹³ Second, three serine residues (33, 38, 43) present in the acidic domain have been shown to be phosphorylated by casein kinase II (CK2). ¹⁴ Phosphorylation by CK2 is suggested to be important for pVHL₃₀ tumor suppressor function. Third, pVHL₃₀ but not pVHL₁₉ has been shown to interact with p14/ARF, a tumor suppressor that positively regulated p53 by inhibiting E3 ubiquitin ligase MDM2. ¹² , ¹⁵ These bodies of work suggest that the acidic domain of pVHL₃₀ may be biologically active.

In previous publications, it is apparent that pVHL₃₀ can migrate as a doublet on an acrylamide gel. ² , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ The nature of the slower‐migrating band is unknown. We hypothesized that the elucidation of the posttranslational modification (PTM) of pVHL₃₀ may help shed light on a potentially unique function of pVHL₃₀.

2. RESULTS

2.1. pVHL₃₀, but not pVHL₁₉, migrates as a doublet

In previous publications, ¹⁷ , ¹⁸ we employed an in vitro transcription and translation (IVTT) system where rabbit reticulocyte lysate is used to produce HA‐pVHL₃₀. The programmed rabbit reticulocyte is then incubated with biotinylated HIF1αOH peptides immobilized on streptavidin agarose beads. By pulling down on the beads, we can visualize levels of associated pVHL via western blotting. Curiously, we observed that a 10% input of HA‐pVHL₃₀ migrated as a single species on 10% acrylamide gel while HA‐pVHL₃₀ pulled down by HIF1αOH peptide migrated as a doublet (Figure 1a). In other experiments, we noted that HA‐pVHL₁₉ does not migrate as a doublet when pulled down via HIF1αOH (Figure 1b). By conducting immunoprecipitation experiments with 3xFLAG‐pVHL₃₀, we realized that this construct migrated as a singlet (Figure 1c). We considered multiple possibilities. One was that the HA tag promoted PTM of pVHL₃₀. The second possibility was that the N‐terminal 3xFLAG tag, which is larger than an HA tag, prevented the PTM of pVHL₃₀. To test the latter hypothesis directly, we cloned a C‐terminal 3xFLAG pVHL₃₀. Indeed, this construct migrated as a doublet (Figure 1c). Further, when we added hydroxylated and unhydroxylated peptide to the pulldown buffer, we found no modulation in the proportion of the slower‐ and faster‐migrating species. These results suggested that the pVHL modification was independent of an interaction with HIFα (Figure 1c). To further interrogate the role of N‐terminal tags on the modification of the pVHL₃₀ acidic domain, we cloned an untagged pVHL₃₀ construct. Much like the HA‐tagged construct, untagged pVHL₃₀ migrated as a doublet (Figure 1d).

FIGURE 1.

pVHL₃₀ migrates as a doublet. (a, b, and d) Biotinylated HIF1α peptides were immobilized on streptavidin‐agarose beads and incubated with in vitro transcribed and translated (IVTT) pVHL in EBC buffer supplemented with protease inhibitors. Streptavidin beads were pulled down (PD), and levels of pVHL were visualized via immmunoblotting (IB). (c) IVTT pVHL was incubated in EBC buffer supplemented with protease inhibitors with or without HIF1α peptides. 3xFLAG‐pVHL was immunoprecipitated (IP) with an anti‐FLAG antibody, and levels of pVHL were visualized via IB. (e) HEK293a cells were transiently transfected HA‐pVHL constructs for 48 hr before being lysed in EBC buffer supplemented with protease inhibitors. Hundred micrograms of total protein was resolved via SDS‐PAGE, and levels of pVHL were visualized via immunoblotting

We observed a similar phenomenon when HEK293A cells were transiently transfected with either pVHL₃₀ or pVHL₁₉ (Figure 1e). Our first hypothesis was that the shift in apparent molecular weight may represent phosphorylation of serine residues 33, 38, and 43. ¹⁴ However, site‐directed mutagenesis of these three serine residues to alanine did not abolish the slower‐migrating band (Figure 1e). At this point in our investigation, we hypothesized that a novel PTM localized to the acidic domain of pVHL₃₀ was being catalyzed by an active enzyme present in both rabbit reticulocyte lysate and HEK293A cells.

2.2. AEBSF retards pVHL₃₀ migration

We further completed a series of biochemical experiments by omitting and altering certain conditions of the pVHL pulldown experiment. We found that the slower‐migrating band does not appear when the pulldown is conducted in a buffer with neutral pH (Figure S1). The slower‐migrating band becomes incrementally more prominent with increasing alkalinity of the pulldown buffer (Figure S1). Notably, pVHL₃₀ migrated as a singlet when protease inhibitor cocktail was omitted from the pulldown buffer (Figure 2a). The addition of a protease inhibitor cocktail to the rabbit reticulocyte lysate yielded the slower‐migrating pVHL₃₀ species when directly resolved on an acrylamide gel (Figure 2b). We hypothesized that the protease inhibitor cocktail may be inhibiting an enzyme that removes a PTM from pVHL₃₀ or that the modified pVHL₃₀ species is proteolytically degraded. However, further attempts to probe the conditions necessary for the modification of pVHL₃₀ were hampered by the inability to modify the rabbit reticulocyte lysate without also directly affecting pVHL₃₀ translation/stability. To circumvent this issue, we purified recombinant pVHL₃₀ and pVHL₁₉ in complex with elongin B and elongin C from BL21 Escherichia coli. pVHL₁₉ migrated as a singlet even when incubated with protease inhibitor and rabbit reticulocyte lysate (Figure 2c). Conversely, pVHL₃₀ was purified as a singlet but migrated as a doublet when incubated with protease inhibitor and rabbit reticulocyte (Figure 2c). This process was temperature dependent as the proportion of modified pVHL₃₀ was increased when the incubation temperature was increased from 4°C to 30°C (Figure 2c). Unexpectedly, the presence of a protease inhibitor cocktail was both necessary and sufficient for the modification of pVHL₃₀ (Figure 2c).

FIGURE 2.

AEBSF covalently modifies pVHL₃₀. (a) Biotinylated HIF1α peptides were immobilized on streptavidin‐ agarose beads and incubated with in vitro transcribed and translated (IVTT) pVHL in EBC buffer supplemented with or without protease inhibitors. Streptavidin beads were pulled down (PD), and levels of pVHL were visualized via immmunoblotting (IB). (b) Rabbit reticulocyte lystate was programmed with a HA‐pVHL₃₀ construct in the presence or absence of protease inhibitors. Two microliter of rabbit reticulocyte lysate was resolved via SDS‐PAGE. (c–e) 2 μg of purified pVHL complex (pVHL₁₉/pVHL30, elongin B, and elongin C) was incubated at 30°C for 2 hr with or without rabbit reticulocyte lysate and with or without protease inhibitors. One microgram of protein was resolved via SDS‐PAGE, and levels of pVHL were visualized via immunoblotting

We next screened individual protease inhibitors and observed that incubation with the serine protease inhibitor 4‐(2‐aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) alone is sufficient to obtain two migratory species of pVHL₃₀ (Figure 2d). Incubation of pVHL₃₀ with other protease inhibitors, including E64, leupeptin, bestatin, and pepstatin, did not induce a slower‐migratory shift in pVHL₃₀ (Figure 2d). Under increasing concentrations of AEBSF, pVHL₃₀ began to migrate in a laddering pattern (Figure 2e). pVHL₁₉ also began to migrate as multiple species but only under higher concentrations of AEBSF. These results demonstrate that AEBSF covalently modifies pVHL₃₀ at concentrations normally used in experiments. As AEBSF predominantly modifies serine and tyrosine residues, ²⁷ and mutation of the three serine residues found in the acidic domain failed to prevent pVHL₃₀ modification (Figure 1e), we hypothesized that AEBSF may be modifying the lone tyrosine residue found in the acidic domain (Tyr23). However, Y23F mutation did not prevent pVHL₃₀ modification (Figure S2). The first 53 amino acids of pVHL₃₀ lack lysine and histidine residues that are other possible sites of AEBSF modification. ²⁷

We next aimed to identify the modified site by mass spectrometry. AEBSF‐modified pVHL₃₀ was enzymatically digested using trypsin and samples were analyzed by LC–MS. As AEBSF treatment results in a modification with a calculated mass shift of 183.035, we used an open search with this mass shift as a potential modification on any amino acid, in addition to other typical post‐translational modifications (see Section 4). This resulted in the identification of 11 unique peptides, representing sequence coverage of approximately 50%. Based on the in silico tryptic digestion of pVHL₃₀, the identified peptides correspond to the majority of predicted MS‐amenable peptides in the protein sequence. None of the identified peptides, however, mapped to the N‐terminal region of the protein. This was not surprising, considering that no tryptic cleavage site is present within the first 63 amino acid residues. Moreover, digestion with a different protease such as Glu‐C would be of little use considering that over one‐third (21/60) of the first 60 amino acids of the protein are glutamic acid residues. However, we did identify a peptide (IAHQ[+183]R) with a putative AEBSF‐modification near the C‐terminus of the protein (Figure S3). Interestingly, AEBSF appears to modify a glutamine residue at position 209 in this case. To the best of our knowledge, glutamine residues have not been reported to be covalently modified by AEBSF. As the C‐terminus region is common to both pVHL₁₉ and pVHL₃₀, this modification is unlikely to explain the slower migration seen specifically with pVHL₃₀.

3. DISCUSSION

PTMs have long been recognized as an important means of regulating protein activity, localization, stability, and function. Numerous pVHL PTMs have been reported, including phosphorylation, ¹⁴ , ²⁸ , ²⁹ SUMOylation, ¹⁹ ubiquitylation, ³⁰ neddylation, ¹⁶ , ³¹ and N‐terminal cleavage. ³² Of these potential PTMs, phosphorylation of serine residues 33, 38, and 43 and cleavage C‐terminal to Tyr23 are localized to the N‐terminal acidic domain unique to pVHL₃₀. Thus, we were intrigued to study a slower migrating pVHL₃₀ species (Figure 1a), especially once the potential modification was found to be localized to the acidic domain, and the possibility of serine phosphorylation was ruled out (Figure 1b,e). Experiments with both IVTT pVHL₃₀ and purified pVHL₃₀–elongin B–elongin C complex show that the slower migrating pVHL₃₀ species is an AEBSF‐related artifact.

Other proteins, including amyloid beta, ³³ transthyretin, ³⁴ and the ζ of the T cell receptor, ³³ , ³⁴ , ³⁵ have been found to be covalently modified by AEBSF. In the aforementioned cases, the modification by AEBSF was detected by mass spectrometry. In the case of the T cell receptor, the protein modification was not observable by SDS‐PAGE. Two‐dimensional electrophoresis studies of human plasma revealed that AEBSF alters the migration profile of several proteins. ²⁷ AEBSF inhibits serine proteases through covalent modification of serine residues in the active site, thus forming a sulfonate ester. AEBSF has also been reported to modify serine residues in other chemical contexts in addition to tyrosine, histidine, lysine, and N‐terminal amine nucleophiles. ²⁷ Interestingly, we identified a Gln209 as being covalently modified by AEBSF. This modification, however, is unlikely to cause the slower migration associated with pVHL₃₀ as it also found in pVHL₁₉. It is likely that modification of a residue unique to pVHL₃₀ retards protein migration, but no such site was identified in our MS/MS experiment. It is also possible that the acidic domain of pVHL₃₀ interacts with the core of the pVHL protein and activates a nucleophile that otherwise would not react with AEBSF.

AEBSF preferentially modifies active, nucleophilic serine residues, which has led to the development of AEBSF as an activity‐based probe (ABP) for serine proteases. ³⁶ Interestingly, further use of AEBSF as an ABP uncovered the specific labeling of functional tyrosine residues in the active site of glutathione‐s‐transferase enzymes. ³⁷ This work has been extended to designing ABPs that specifically react with specific tyrosine residues in the active site of the mRNA‐decapping scavenger enzyme DcpS. ³⁸ Thus, AEBSF as ABP could be useful for specifically labeling pVHL₃₀ and not pVHL₁₉. However, this might be a challenge given that we found a residue common to both pVHL₁₉ and pVHL₃₀ that can react with AEBSF.

In addition to identifying a putitive and artifactual PTM of pVHL₃₀, we uncovered several conditions that prevent this modification. The presence of modified pVHL₃₀ is more pronounced under alkaline condition. This might explain why the pVHL₃₀ doublet is not observable in experiments where buffer conditions are more neutral. ³⁹ Further, the use of a N‐terminal 3xFLAG tag prevents modification of pVHL₃₀ by AEBSF (Figure 1d). This observation suggests that the lengthy 3xFLAG tag interacts with the acidic domain. This notion is supported by the observation that intracellular truncation of the pVHL₃₀ acidic domain, hypothesized to be catalyzed by chymotrypsin, is inhibited by N‐terminal HA and 3xFLAG tags. ³²

Here, we show that the migration of pVHL₃₀ is artifactually modified by the protease inhibitor AEBSF. This phenomenon is dependent on the first 53 amino acids unique to the longer isoform of pVHL. Since it is unknown whether the artifactual, slower‐migrating species of pVHL₃₀ has or influences any biological activity, any investigation into the potential function of the acidic domain, which is predicted to be disordered, ¹² should avoid the use of any N‐terminal tags.

4. MATERIALS AND METHODS

4.1. Plasmids

Construction of the following plasmids has been described previously; pcDNA3‐HA‐VHL₃₀(WT), ⁴⁰ pcDNA3‐3xFLAG‐VHL₃₀(WT), pACYCDuet‐1 plasmid encoding untagged elongin B and elongin C(17–112), pGEX‐4T‐1‐GST‐VHL₁₉. ⁴¹ pcDNA3‐HA‐VHL₁₉(WT), pcDNA3‐VHL₃₀(WT), and pGEX‐4T‐1‐GST‐VHL₃₀ were subcloned from pcDNA3‐HA‐VHL₃₀(WT). pcDNA3‐HA‐VHL₃₀(S33A/S38A/S43A) was generated by site‐directed mutagenesis of pcDNA‐HA‐VHL₃₀(WT). C‐terminal tagged pcDNA3‐3xFLAG‐VHL₃₀(WT) was generated by amplifying VHL₃₀ from pcDNA3‐HA‐VHL₃₀(WT) with a reverse primer containing 3xFLAG and subcloning into an empty pcDNA3 vector.

4.2. Antibodies

Anti‐HA (C29F4; 1:2,000 dilution) was obtained from Cell Signaling Technology. Anti‐FLAG (F1804; 1:5,000) and Anti‐tubulin (T5168; 1:5,000) were obtained from Sigma‐Aldrich. Anti‐VHL (sc‐135657, 1:1,500) was obtained from Santa Cruz.

4.3. Peptides

HIF‐1α (556–564; DLDLEMLAPYIPMDDDFQL) peptides with N‐terminal biotinylation and C‐terminal amidation modifications were custom synthesized by Genscript. All peptides were reconstituted to 2 mg/mL, as measured by A280, using sterile DMSO, aliquoted, and stored at −80°C.

4.4. Protein expression and purification

BL21(DE3) E. coli cells were co‐transformed with a dual expression construct encoding untagged elongin B (full‐length, 1–118) and elongin C (17–112) and a construct encoding either N‐terminal GST‐tagged pVHL₁₉ or N‐terminal GST‐tagged pVHL₃₀. Bacterial cultures (1 L) were grown to an OD₆₀₀ of approximately 0.6. Expression of the pVHL–elongin B–elongin C complex was induced via addition of IPTG (final concentration of 1 mM). Following induction, cells were grown for an additional 3.5 hr at a temperature of 37°C. Bacterial pellets were resuspended in 20 mM HEPES pH 7.4, 200 mM NaCl freshly supplemented with 10 mM DTT and lysed using a cell disruption unit at a pressure of 30 kPSI. Cell lysate was cleared via centrifugation (34,000g for 40 min). Cleared lysate was applied to a column of glutathione sepharose resin (GE Life Sciences). The column was washed with 20 mM HEPES pH 7.4, 200 mM NaCl and protein was eluted with 20 mM HEPES pH 7.4, 200 mM NaCl supplemented with 10 mM reduced glutathione. Protein solution was concentrated using a 4‐mL centrifugal concentrator with a molecular weight cut‐off of 3,500 Da (Pall Corporation). The concentrated protein solution was purified on a Superdex 200 10/300 size exclusion chromatography column equilibrated with 20 mM HEPES pH 7.4, 200 mM NaCl, and 1 mM DTT. The monomeric VBC complex was diluted to a concentration of 0.7 mg/mL. The GST‐tag was cleaved via incubation (60 hr) with 1 U thrombin per mg of VBC protein at 4°C. Free GST was removed by applying the protein solution to regenerated glutathione sepharose resin. The protein solution was concentrated as discussed above and size exclusion chromatography was employed to purify soluble, monomeric VBC complex. Purity was confirmed via SDS‐PAGE analysis. Aliquots of protein were frozen at −80°C at a concentration between 1 and 2 mg/mL.

4.5. In vitro binding assay

The in vitro pVHL‐HIF‐1α binding assay was performed according to a previously published protocol. ⁴² VHL constructs was expressed in a transcription and translation (TNT) rabbit reticulocyte lysate system (Promega, Cat. No. L1170) and incubated with 2 μg of biotinylated HIF‐1α peptide (hydroxylated or unhydroxylated), immobilized on streptavidin agarose beads, in either 500 μL of EBC buffer (50 mM Tris–HCl pH 8.0, 120 mM NaCl, 0.5% [v/v] NP‐40) supplemented with protease inhibitors for 2 hr at 4°C. Following incubation, beads were washed ×5 with NETN buffer (20 mM Tris–HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% [v/v] NP‐40). Biotinylated peptide was pulled down via streptavidin agarose beads and protein was eluted by boiling beads in sample buffer. Samples were resolved on a 10% acrylamide gel via SDS‐PAGE.

4.6. In vitro VHL modification assay

Two μg of purified pVHL–elongin B–elongin C was incubated in 20 μL of 50 mM Tris–HCl pH 8.0, 120 mM NaCl for 2 hr at 30°C. Two microliter of rabbit reticulocyte was added to indicated reaction conditions. For individual protease inhibitors, the following working concentrations were utilized; leupeptin (100 μM), pepstatin A (10 μM), E‐64 (10 μM), AEBSF (1 mM), bestatin hydrochloride (10 μM). After incubation, the reaction was stopped with sample buffer. Samples were resolved on a 10% acrylamide gel via SDS‐PAGE.

4.7. Cell culture and transfection

HEK293a cells were obtained from American Type Culture Collection (ATCC). Cells were maintained in DMEM (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum (Wisent) and were grown at 37°C in a humidified, 5% CO₂ atmosphere. The indicated plasmids were transiently transfected into HEK293a cells using polyethylenimine.

4.8. Liquid chromatography–mass spectrometry

Purified pVHL was incubated with AEBSF as indicated above. To avoid inhibition of trypsin activity by excess AEBSF, samples were subjected to buffer exchange in Amicon Ultra 3 kDa MWCO columns (Millipore) using 50 mM NH₄HCO₃. Samples were next reduced with DTT (5 mM, 20 min, 60°C) and alkylated with iodoacetamide (10 mM, 30 min, room temperature, light protected). Protein digestion was performed with sequencing‐grade, TPCK‐treated, modified trypsin (Promega) for 16 hr at 37°C. The resulting peptide samples were desalted using C18 chromatography columns and lyophilized. Peptides were re‐suspended in 0.1% formic acid analyzed by LC–MS.

Liquid chromatography was conducted using a C18 pre‐column (Acclaim PepMap 100, 2 cm × 75 μm ID, Thermo Scientific) and a C18 analytical column (Acclaim PepMap RSLC, 50 cm × 75 μm ID, Thermo Scientific), running a 120 min reversed‐phase gradient (0–40% ACN in 0.1% formic acid) at 225 nL/min on an EASY‐nLC1200 pump (Thermo Scientific). Mass spectrometry was performed on a Q‐Exactive HF instrument (Thermo Scientific). An MS scan was performed with a resolution of 60,000 followed by up to 20 MS/MS scans (minimum ion count of 1,000 for activation) using higher energy collision induced dissociation (HCD) fragmentation. Dynamic exclusion was set for 5 s (10 ppm; exclusion list size = 500). Raw data was analyzed for identification of peptides and post‐translational modifications using the FragPipe platform. ⁴³ Spectra were searched against the pVHL₃₀ protein sequence. Search parameters specified a parent ion mass tolerance of 10 ppm, and an MS/MS fragment ion tolerance of 0.4 Da, with up to two missed cleavages allowed for trypsin (excluding K/RP). Variable modifications of 0.98 on N or Q, 15.99 on M, 57.02 on C and 183.04 (corresponding to the mass difference resulting from the AEBSF‐induced modification) on any amino acid were set in the search parameters. Spectra were manually inspected for quality assurance and confirmation of proper assignment of modifications.

AUTHOR CONTRIBUTIONS

Daniel Tarade: Conceptualization; data curation; formal analysis; investigation; methodology; supervision; validation; visualization; writing‐original draft; writing‐review and editing. Shelley He: Formal analysis; investigation; methodology; validation; visualization; writing‐review and editing. Jonathan St‐Germain: Data curation; formal analysis; investigation; methodology; validation; visualization; writing‐review and editing. Avi Petroff: Investigation; validation. Anya Murphy: Investigation. Brian Raught: Resources; supervision; writing‐review and editing. Michael Ohh: Conceptualization; formal analysis; funding acquisition; methodology; resources; supervision; writing‐original draft; writing‐review and editing.

Supporting information

Figure S1 Alkaline conditions promote covalent modification of pVHL₃₀. Biotinylated HIF1α peptides were immobilized on streptavidin‐ agarose beads and incubated with in vitro transcribed and translated (IVTT) pVHL in EBC buffer supplemented with protease inhibitors. Streptavidin beads were pulled down (PD), and levels of pVHL were visualized via immmunoblotting (IB).

Figure S2 AEBSF does not modify pVHL₃₀ Tyr23. HA‐pVHL₃₀ was in vitro transcribed and translated (IVTT) and incubated with the indicated concentration of AEBSF for 2 hr at 30°C. Levels of pVHL were visualized via immmunoblotting (IB).

Figure S3 Covalent modification of pVHL Gln209 by AEBSF. Purified pVHL₃₀ complex was incubated with or without AEBSF for 2 hr at 30°C. AEBSF was eliminated from the sample using an Amicon Ultra 3 kDa MWCO columns to buffer exchange. Protein was reduced (5 mM DTT) and alkylated (10 mM iodoacetamide) prior to incubation with trypsin for 16 hr at 37°C. Resultant peptides were desalted on a C18 column and lyophilized. Peptides were re‐suspended in 0.1% formic acid analyzed by LC–MS. Liquid chromatography was conducted using a C18 pre‐column and a C18 analytical column, running a 120 min reversed‐phase gradient (0–40% ACN in 0.1% formic acid) at 225 nL/min on an EASY‐nLC1200 pump. Mass spectrometry was performed on a Q‐Exactive HF instrument. An MS scan was performed with a resolution of 60,000 followed by up to 20 MS/MS scans (minimum ion count of 1,000 for activation) using higher energy collision induced dissociation (HCD) fragmentation. Dynamic exclusion was set for 5 s (10 ppm; exclusion list size = 500). Raw data was analyzed for identification of peptides and post‐translational modifications using the FragPipe platform. Spectra were searched against the pVHL₃₀ protein sequence. Amino acid sequence of pVHL₃₀ with AEBSF‐modified tryptic peptide highlighted is provided. Search parameters specified a parent ion mass tolerance of 10 ppm, and an MS/MS fragment ion tolerance of 0.4 Da, with up to two missed cleavages allowed for trypsin (excluding K/RP). Variable modifications of 0.98 on N or Q, 15.99 on M, 57.02 on C and 183.04 (corresponding to the mass difference resulting from the AEBSF‐induced modification) on any amino acid were set in the search parameters. Spectra were manually inspected for quality assurance and confirmation of proper assignment of modifications.

Tarade D, He S, St‐Germain J, et al. The long form of pVHL is artifactually modified by serine protease inhibitor AEBSF . Protein Science. 2020;29:1843–1850. 10.1002/pro.3898