Mechanism-based traps enable protease and hydrolase substrate discovery

Abstract

Hydrolase enzymes, including proteases, are encoded by 2-3% of the genes in the human genome and 14% of these enzymes are active drug targets¹. However, the activities and substrate specificities of many proteases-especially those embedded in membranes-and other hydrolases remain unknown. Here we report a strategy for creating mechanism-based, light-activated protease and hydrolase substrate traps in complex mixtures and live mammalian cells. The traps capture substrates of hydrolases, which normally use a serine or cysteine nucleophile. Replacing the catalytic nucleophile with genetically encoded 2,3-diaminopropionic acid allows the first step reaction to form an acyl-enzyme intermediate in which a substrate fragment is covalently linked to the enzyme through a stable amide bond²; this enables stringent purification and identification of substrates. We identify new substrates for proteases, including an intramembrane mammalian rhomboid protease RHBDL4 (refs. ^3,4). We demonstrate that RHBDL4 can shed luminal fragments of endoplasmic reticulum-resident type I transmembrane proteins to the extracellular space, as well as promoting non-canonical secretion of endogenous soluble endoplasmic reticulum-resident chaperones. We also discover that the putative serine hydrolase retinoblastoma binding protein 9 (ref. ⁵) is an aminopeptidase with a preference for removing aromatic amino acids in human cells. Our results exemplify a powerful paradigm for identifying the substrates and activities of hydrolase enzymes.

Fig. 1. Dap-mediated HtrA2 substrate identification in mammalian cell lysate.

a, HtrA2(S306Dap)–HA–Strep and its conjugates were enriched from cell lysate with anti-HA beads and detected with an anti-Strep antibody. Control experiments were performed with wild-type (WT) HtrA2 and the catalytically inactive S306A mutant. Input: HtrA2 variants in cell lysates before incubation. b, Venn diagram showing the number of proteins identified in HtrA2(S306Dap) elution compared with controls. Proteins identified in at least two of the three replicates were considered as positively identified. c, Volcano plot based on label-free quantification (LFQ) values for the proteins identified in HtrA2(S306Dap) and wild-type HtrA2 samples. The black line represents the cut-off curve for significance (S₀ = 1, FDR < 0.01). Each data point is calculated in Perseus using n = 4 for each HtrA2 variant. The dot representing ornithine aminotransferase (OAT) is labelled in red. d, Wild-type HtrA2 or HtrA2(S306A) (1 μM) was added into Expi293 cell lysate and incubated for the indicated time at 37 °C. Red arrowhead, full-length OAT; blue arrowhead, wild-type HtrA2-dependent proteolytic fragments. GAPDH was used as a loading control. The experiment in a was performed in biological triplicate, and the experiment in d was performed in two biological replicates, both with similar results. For gel source data, see Supplementary Fig. 1.

Fig. 2. Optical activation of a Protease(Dap) substrate trap in human cells.

a, The mass of TEV(C151pc-Dap) before and after illumination of human cells expressing the protein. TEV containing pc-Dap was produced using the DapRS–tRNA_CUA pair and a TEV gene bearing the amber codon (TAG) at position 151 in the presence of pc-Dap. The grey trace shows proteins purified before illumination. Fully protected TEV(C151pc-Dap) bearing an acetyl group ([Ac-TEV(C151pc-Dap)]: expected 39,245.4 Da, observed 39,244.4 Da). The blue trace shows proteins purified immediately after illumination of cells. The fully deprotected product ([Ac-TEV(C151Dap)]: expected 38,948.2 Da, observed 38,945.2 Da); the deprotection intermediate ([Ac-TEV(C151Dap_inter)]: expected 39,052.2 Da, observed 39,054.0 Da). The brown trace shows proteins purified 6 h after illumination of cells. b, TEV variants and GFP-s were co-expressed in HEK293T cells for 48 h. Total lysate was analysed by anti-Strep (for TEV) and anti-GFP antibodies. Wild-type TEV quantitatively cleaved GFP-s to GFP. β-Tubulin was used as a loading control. c, Detection of TEV(Dap)–GFP conjugate after Strep-tag enrichment. Samples, not illuminated (lane 4) or after UV illumination (lanes 1–3 and 5–10) were collected at indicated time points. Input, cell lysates before immunoprecipitation probed with anti-Strep and anti-GFP antibodies. β-Tubulin was used as a loading control. Experiments in a–c were performed in two biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Fig. 3. Discovery of RHBDL4 substrates.

a, RHBDL4 variants were expressed in Expi293 cells. Immunoblotting analysis of RHBDL4 variants enriched from an equal number of cells after optical activation and substrate capture. Ala, RHBDL4(S144A); Dap, RHBDL4(S144Dap). Red arrows, RHBDL4(S144Dap)-specific higher molecular mass bands. b, Venn diagram showing the number of proteins identified in RHBDL4(S144Dap) elution with respect to controls. Proteins found in at least two of the three replicates were considered to be positively identified. c, d, Volcano plots based on the LFQ values for the identified proteins for RHBDL4(S144Dap) versus wild-type RHBDL4 (c; S₀ = 1, FDR < 0.01) and RHBDL4(S144Dap) versus RHBDL4(S144A) (d; S₀ = 1.5, FDR < 0.05). Black lines represent the cut-off curve for significance. Each point was calculated in Perseus using n = 3 for each RHBDL4 variant. ER-resident candidates are marked in red and select candidates are also labelled. e, RHBDL4 cleaves CCDC47. SP, signal peptide; V5, V5-tag; TMH, transmembrane helix; HA, HA-tag; red arrows, RHBDL4 cleavage sites; black circles, full-length CCDC47; blue triangles, N-terminal proteolytic fragments; red triangles, C-terminal proteolytic fragments; asterisk, bands present without RHBDL4. f, RHBDL4(S144Dap) conjugates to endogenous BiP and the cleavage of endogenous BiP by wild-type RHBDL4. Cell lysates before immunoprecipitation (input) and the conjugates were directly visualized by anti-BiP and anti-Strep antibodies after Strep-tag enrichment. Red arrow, the proteolytic fragment of endogenous BiP cleaved by wild-type RHBDL4. g, Cleaved BiP is secreted into the medium. The medium was separated into supernatant (SN) and microvesicles (MV). Red arrow, the proteolytic fragment. Whereas full-length BiP was detected by both anti-BiP and anti-KDEL antibodies, cleaved BiP can only be detected with an anti-BiP antibody. Asterisk indicates non-specific bands. Revert 700 total protein stain was used as a loading control. Experiments in a, f, were performed in biological triplicate; the experiment in e was repeated in three biological replicates; the experiment in g was repeated in two biological replicates; all with similar results. For gel source data, see Supplementary Fig. 1.

Fig. 4. Identifying RBBP9(Dap75) conjugates.

a, The Dap-containing tryptic peptide sequence (Pept(Dap-X)) of RBBP9 with certain b and y ion masses of Pept(Dap) modified by mx. b, Pipeline to identify mx in live cells. RBBP9(Dap75-X) conjugates were affinity purified and trypsinized. The resulting peptide pool was analysed by LC–MS/MS and peptides with molecular mass no less than MS(Pept(Dap)) were individually selected for mx calculation (for example, the blue peak represents Pept(Dap), mx = 0). The experimental MS2 peaks were compared to theoretical MS2 peaks for scoring as described in Methods. In the example shown, the peaks are colour coded as b, y, b+mx or y+mx ions, using the colour scheme in a. c, mx obtained from the top-scoring spectra were plotted. Each dot represents the mass shift of the observed peptide relative to the parental peptide. +113 (Leu/Ile), +131 (Met), +147 (Phe), +163 (Tyr), +186 (Trp), +297 (pc-Dap), +42 and +77 (consistent with near cognate suppression of this amber codon in mammalian cells with Gln and Tyr, respectively). d, The entire mass of RBBP9 variants and Dap conjugates purified from human cells. Green trace (i): wild-type RBBP9 ([Ac-(RBBP9-Met)]: expected 23,995 Da, observed 23,994.5 Da); black trace (ii): RBBP9(S75A) ([Ac-(RBBP9(S75A)-Met)]: expected 23,979 Da, observed 23,978.5 Da); brown trace (iii): RBBP9(S75pc-Dap) before illumination ([Ac-(RBBP9(S75pc-Dap)-Met)]: expected 24,291 Da, observed 24,290.5 Da); purple dashed trace (iv): RBBP9(S75Dap) deprotected from RBBP9(S75pc-Dap) in vitro ([Ac-(RBBP9(S75Dap)-Met)]: expected 23,994 Da, observed 23,993.5 Da); multicolour trace (v): RBBP9(S75pc-Dap) purified from cells after illumination and substrate trapping for 3 h ([Ac-(RBBP9(S75Dap_inter)-Met)]: expected 24,098 Da, observed 24,097 Da; [Ac-(RBBP9(Dap-L/I)-Met)]: expected 24,107 Da, observed 24,108 Da; [Ac-(RBBP9(Dap-F)-Met)]: expected 24,141 Da, observed 24,140.5 Da; [Ac-(RBBP9(Dap-Y)-Met)]: expected 24,157 Da, observed 24,156 Da; [Ac-(RBBP9(Dap-W)-Met)]: expected 24,180 Da, observed 24,179 Da). The experiment in c was performed in biological triplicate. In d, the entire mass acquisition of trace (v) was performed in two biological replicates with similar results, the other traces were acquired once.

Fig. 5. Characterization of RBBP9 aminopeptidase activity.

a, RBBP9 is specific for aromatic amino acids. The graph shows the catalytic efficiency of RBBP9 on 19 AA–AMCs relative to its catalytic efficiency on Phe–AMC. The bar graph represents the mean of n = 2 independent measurements. b, c, RBBP9 cleaves aromatic amino acids from the N terminus of peptide hormones (nociceptin or MENK). The full-length (FL) peptides and the products (DePhe¹ or DeTyr¹) after incubating with wild-type RBBP9 or RBBP9(S75A) were determined by mass spectrometry. Black solid line shows detection of product after incubating with WT RBBP9; brown solid line shows detection of product after incubating with RBBP9(S75A); black dashed line shows detection of full-length peptide after incubating with wild-type RBBP9; brown dashed line shows detection of full-length peptide after incubating with RBBP9(S75A). d, e, The crystal structure of RBBP9 in complex with Phe. d, surface view of RBBP9 and sphere representation of Phe (purple). The surface of Tyr99, Leu103, Phe140 and Leu141 is shown in yellow. e, Ribbon diagram of RBBP9 with key residues shown as sticks. Side chains of Tyr99, Leu103, Phe140 and Leu141 are shown in yellow; side chains of Glu108 and Ser76 are shown in green. The hydrogen bond between the α-amine group of Phe and the side chain of Glu108 (2.7 Å) is represented by a red dashed line. The side chain of Ser76 or Tyr99 may also form hydrogen bonds with the α-amino group of Phe (dashed brown lines, bond distances of 2.8 Å or 2.9 Å, respectively). Source data

Extended Data Fig. 1. Strategies for Dap-mediated hydrolase substrate discovery.

Genetically encoded pc-Dap in place of the catalytic serine or cysteine in the active site of a hydrolase enables the deprotection of pc-Dap to Dap and the activation of a hydrolase substrate trap that can covalently capture substrates. Purified recombinant and pre-activated Dap-containing hydrolase can be added to lysate or extract and used to capture substrates. Hydrolase containing pc-Dap can also be expressed directly in mammalian cells. The hydrolase substrate trap can then be optically activated in live cells to capture substrates. Covalent conjugates can be enriched by immunoprecipitation with stringent washing to remove non-covalent binders and the conjugates can be visualized and identified by immunoblot and mass spectrometry-based methods. Control experiments use the wild-type (WT) enzyme that does not form a stable covalent acyl-enzyme conjugate and the catalytically inactive mutant (Ala, catalytic serine/cysteine is mutated to alanine) of the hydrolase that does not react with substrates. We note that not all the Dap-containing hydrolase will necessarily be found in conjugates; in general, we expect the fraction of Dap-containing hydrolase in conjugates to be a function of hydrolase abundance, substrate abundance, effects on the rate of acyl-intermediate formation resulting from replacing the catalytic nucleophile with Dap, and the stability of the acyl-intermediate.

Extended Data Fig. 2. TEV(C151Dap) specifically traps its substrate in Expi293 cell lysate.

TEV variants (WT, Ala, Dap, 1 μM) were incubated with control GFP-c (a polypeptide “Gly-Gly-Gly-Ser-Gly-Gly-Gly-His6” was attached at the C-terminus of GFP) or substrate GFP-s (a TEV cleavage sequence “Glu-Asn-Leu-Tyr-Phe-Gln-Gly-His6” was attached at the C-terminus of GFP. The cleavage between Gln and Gly is underlined) in Expi293 cell lysate at 37 °C for 3 h. The concentration of GFP-s or GFP-c was (a) 1 μM or (b) 5 μM. Dap^*: reaction of TEV(C151Dap) with GFP-s in Tris buffer. The input (reaction before IP) and Strep-tag enriched TEV species were analyzed by SDS-PAGE (Coomassie staining) and WB (anti-Strep for TEV and anti-GFP). (a) and (b) were repeated in three biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 3. Dap-containing proteases selectively capture known substrates of their parent enzymes.

(a) The change of AMC fluorescence resulting from cleavage from the C-terminus of Ubiquitin (Ub) or ubiquitin-like molecules (SUMO, NEDD8, ISG15) was followed upon mixing with UL36^USP. WT UL36^USP from human herpesvirus 1 specifically hydrolyses Ub-AMC, but not other ubiquitin-like protein-AMC (UBL-AMC) molecules. The deubiquitination activity of UL36^USP is lost when Cys65 is mutated to Ser. This data is consistent with the previously reported specificity of UL36^USP. (b) UL36^USP(C65Dap) specifically reacts with Ub-AMC to form the UL36^USP(Dap)-Ub conjugate. In contrast, no conjugates were observed between UL36^USP(Dap) and SUMO1, NEDD8 or ISG15. (c) WT SCoV2-PLpro selectively hydrolyses ISG15-AMC in preference to other Ub/UBL-AMC molecules; this data is consistent with the previously reported specificity of SCoV2-PLpro. The hydrolysis of ISG15-AMC is abrogated when the catalytic Cys111 of SCoV2-PLpro is mutated to Ala. (d) SCoV2-PLpro(C111Dap) specifically reacts with ISG15-AMC, generating the SCoV2-PLpro(Dap)-ISG15 conjugate. *: UBL-AMC independent higher MW bands resulting from PLpro(C111Dap) self-reaction. (a) and (c) were generated using n = 3 independent measurements. The line represents the means of three measurements. (b) was repeated twice and (d) was performed in three biological replicates with similar results. For gel source data, see Supplementary Fig. 1. Source data

Extended Data Fig. 4. UL36^USP(C65Dap) and SCoV2-PLpro(C111Dap) form conjugates with endogenous ubiquitin in live cells.

(a) A protease containing pc-Dap in place of its catalytic serine or cysteine is produced in mammalian cells by genetic code expansion, creating a photocaged protease trap. The trap is activated by illuminating cells. The activated trap covalently and specifically captures substrate fragments in acyl-enzyme complexes linked through stable amide bonds. (b) Immunoblotting analysis of UL36^USP variants enriched from an equal number of cells after optical activation and substrate capture. The conjugates formed between UL36^USP(C65Dap) and endogenous proteins were detected by an anti-Strep antibody and an anti-Ub antibody; this demonstrates that UL36^USP(C65Dap) captures Ub and di-ubiquitin (Ub₂) in cells. Catalytically inactive UL36^USP(C65S) non-covalently associated with Ub chains to co-IP them (lanes 2 and 6), while UL36^USP(C65Dap) formed conjugates with endogenous Ub molecules (lanes 7–9). Input: cell lysates before IP probed with an anti-Strep antibody. These data are consistent with previous work demonstrating that UL36^USP is a deubiquitinase in cells. (c) Immunoblotting analysis of SCoV2-PLpro(C111Dap) variants enriched from an equal number of cells after optical activation and substrate capture. The conjugates formed between SCoV2-PLpro(C111Dap) and endogenous proteins were detected by an anti-Strep antibody and an anti-Ub antibody; this demonstrates that SCoV2-PLpro(C111Dap) captures Ub and Ub₂ in cells. Catalytically inactive SCoV2-PLpro(C111A) and SCoV2-PLpro(C111pc-Dap) without illumination non-covalently associated with Ub chains to co-IP them (lanes 2, 4 and 6), while SCoV2-PLpro(C111Dap) also formed conjugates with endogenous Ub molecules (lanes 8 and 9). Input: cell lysates before IP probed with an anti-Strep antibody. Note that SCoV2-PLpro has previously been shown to cleave poly-Ub chains in cells^,. While SCoV2-PLpro will also cleave ISG15, this protein is not expressed in Expi293 cells without stimulation^,. (b) and (c) were repeated in two biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 5. Formation of RHBDL4(Dap)-pTα conjugates in Expi293 cells.

(a) Schematic representation of the model substrate pTα. SP: signal peptide; FLAG: FLAG-tag; TMH: transmembrane helix; V5: V5-tag. Red arrows: RHBDL4 cleavage sites; black arrow: SPase cleavage site. (b) Co-expression of C-terminal Strep-tagged RHBDL4 variants and pTα in Expi293 cells for 40 h. pc-Dap was added to produce full-length RHBDL4(S144pc-Dap). Cell lysates were analysed by anti-Strep (for RHBDL4), anti-FLAG and anti-V5 antibodies. Blue triangles: N-terminal proteolytic fragments of pTα; Red triangles: C-terminal proteolytic fragments of pTα. (c) Detection of RHBDL4(Dap)-pTα conjugates after Strep-tag enrichment. Samples without illumination (lane 4) and with illumination (lanes 1–3 and 5–11) were collected at indicated time points. Input: cell lysate before IP analysed by anti-Strep and anti-FLAG antibodies. (d) The formation of conjugates was monitored for a longer period of time. RHBDL4(Dap)-pTα conjugates were gradually degraded 4 h after UV irradiation. RHBDL4 variants were also gradually degraded over time pTα*, the deglycosylated form of pTα. The experiments were repeated in two biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 6. RHBDL4 cleaves transmembrane substrate candidates.

RHBDL4 cleavage assays for (a) MFN2, (b) EMD, (c) LEMD2 and (d) CCDC47. The putative cleavage sites – the approximate positions of which were estimated based on the MW of the proteolytic bands – are indicated by red arrows in schematic representations. HA: HA-tag; GFP: GFP-tag; SP: signal peptide; V5: V5-tag. The transmembrane helices are labeled in grey. (a) WT RHBDL4 cleaved MFN2, a multiple membrane-spanning protein resident in mitochondria or the ER. Red triangle: the C-terminal proteolytic fragment. (b) WT RHBDL4 cleaved EMD, a type II transmembrane protein in the nuclear inner membrane. Red triangle: the C-terminal proteolytic fragment. (c) WT RHBDL4 cleaved LEMD2, a multiple membrane-spanning protein in the nuclear inner membrane. Two minor cleavages (red triangles) would be in the perinuclear space, but the major cleavage site (blue triangle) in the domain resident in nucleus might be cut by the inverted RHBDL4 (see Supplementary Note 1). (d) RHBDL4 cleaved both full-length (FL) CCDC47 and CCDC47(21–135), generating the same MW N-terminal proteolytic fragment (lanes 2 and 5, blue triangle); this suggested that RHBDL4-mediated cleavage is independent of the transmembrane helix of CCDC47. Blue triangle: the N-terminal fragment; Red triangles: C-terminal proteolytic fragments. (e) Comparison of the proteolytic fragments to truncated CCDC47 standards confirmed that the cleavages are in the luminal domain of CCDC47. WT + FL: FL CCDC47 cleaved by WT RHBDL4. (f) Over-expressed WT RHBDL4 cleaved endogenous CCDC47. Red triangles: proteolytic fragments detected by an anti-CCDC47 antibody. The C-terminal proteolytic fragments of MFN2, LEMD2 and CCDC47 were probed with an anti-HA antibody, while the proteolytic fragment of EMD was detected by an anti-GFP antibody. The N-terminal proteolytic fragments were probed with an anti-V5 antibody. RHBDL4 variants were detected by an anti-Strep antibody. (a-d) were repeated in three biological replicates with similar results. (e) was performed once. (f) was repeated in two biological replicates with similar results.  For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 7. RHBDL4 cleaves endogenous BiP in mammalian cells and facilitates the secretion of N-terminal proteolytic fragments.

(a) BFA inhibitory assay was performed for endogenous BiP. The secretion of BiP proteolytic fragment into extracellular media was inhibited by Brefeldin A (BFA) compared to cells treated with DMSO (lanes 5 vs 2, 11 vs 8, anti-BiP). Especially when the medium was changed, no BiP proteolytic fragment was detected in the medium (lane 11, dashed red arrow). Treatment of BFA greatly increased the endogenous level of BiP (lanes 4–6 vs 1–3, anti-BiP in cell lysate). In comparison, the expression of endogenous BiP was slightly increased when WT RHBDL4 or RHBDL4(S144A) was expressed (lanes 2 and 3 vs 1, anti-BiP in cell lysate). BFA treatment or RHBDL4 variant expression facilitated the secretion of full-length BiP (lane 4 vs 1 or 2 vs 1), but not the proteolytic fragment generated by RHBDL4 cleavage. RHBDL4 expression was detected by an anti-Strep antibody in cell lysate. Revert 700 total protein staining was used as loading control. (b) Endogenous BiP was cleaved by endogenously expressed RHBDL4. WT RHBDL4-dependent BiP proteolytic fragment in the media was detected by an anti-BiP antibody in WT HCT116 cells (red arrow), but not in RHBDL4 knockout HCT116 cells. Transferrin: loading control; additionally, the same samples run in different lanes were stained by Revert 700 for total protein loading control. Endogenous BiP was detected by an anti-BiP antibody and RHBDL4 was detected by an anti-RHBDL4 antibody in the cell lysate. All experiments were repeated in two biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 8. RHBDL4 cleaves other ER-resident proteins before ER retention motifs and facilitates the secretion of N-terminal proteolytic fragments.

RHBDL4 cleavage assays for V5-tagged (a) PDIA6, (b) Calreticulin and (c) ERP44. The putative cleavage sites are indicated by blue arrows in the schematic representations. (a) WT RHBDL4 cleaved PDIA6 at the C-terminus, resulting in secretion of the N-terminal fragment (blue arrow) into the medium. (b) WT RHBDL4 cleaved Calreticulin at the internal region (minor cleavage) and the C-terminus (major cleavage), resulting in secretion of the N-terminal fragments (blue arrows) into the SN. (c) WT RHBDL4 cleaved ERP44 before RDEL sequence. The secreted proteins were deglycosylation mix II sensitive. Because we could not source an antibody to specifically detect the RDEL sequence, we inserted an HA-tag four amino acids before the RDEL sequence for detection. Anti-V5 detected proteolytic fragments 1 and 2 (p1 and p2), and p1 was also detected by the anti-HA antibody, indicating that one cleavage might happen after the HA-tag. The expression of RHBDL4 was detected by an anti-Strep antibody. RHBDL4 cleavage assays for (d) endogenous PDIA6 and (e) endogenous Calreticulin. (d) WT RHBDL4 cleaved endogenous PDIA6. The proteolytic fragment of PDIA6 (red arrow) without the KDEL sequence was secreted into the SN. (e) WT RHBDL4 cleaved endogenous Calreticulin. The proteolytic fragment of Calreticulin (red arrow) without the KDEL sequence was secreted into the SN. Expression of RHBDL4 did not increase the endogenous level of PDIA6 and Calreticulin (lanes 2 and 3 vs 1). The expression of RHBDL4 is shown in Fig. 3g. BFA inhibitory assays for (f) PDIA6 and (g) Calreticulin. The secretion of proteolytic fragments of PDIA6 and Calreticulin was inhibited compared to cells treated with DMSO (lanes 5 vs 2, 11 vs 8). Endogenous PDIA6 and Calreticulin were detected by an anti-PDIA6 and an anti-CALR antibody, respectively. The expression of RHBDL4 is shown in Extended Data Fig. 7a. (h) Endogenous RHBDL4 cleaves endogenous Calreticulin. WT RHBDL4-dependent Calreticulin proteolytic fragments in the extracellular media were detected by an anti-CALR antibody in WT HCT116 cells (red arrows), but not in RHBDL4 KO HCT116 cells. The secretion of the major proteolytic fragment (near full-length Calreticulin, low intensity) and the minor proteolytic fragment (high intensity) were both detected. Transferrin: loading control. The endogenous RHBDL4 is shown in Extended Data Fig. 7b. (a-h) were repeated in two biological replicates with similar results. For gel source data, see Supplementary Fig. 1.

Extended Data Fig. 9. The distribution of cysteine and serine protease clans.

The clan distribution of cysteine proteases from (a) animals and (b) viruses. The clan distribution of serine proteases from (c) animals and (d) viruses. We have demonstrated the Dap-mediated protease substrate capture method for multiple cysteine proteases (TEV (Clan PA), UL36^USP (Clan CA) and SCoV2-PLpro (Clan CA)) and serine proteases (HtrA2 (Clan PA), RHBDL4 (Clan ST) and RBBP9 (Note that RBBP9 has not been included in the MEROPS database by September 30, 2021. RBBP9 possesses the “alpha-beta hydrolase” fold, suggesting that it belongs to Clan SC). The data used to generate this figure were downloaded from the MEROPS database. Source data