Activity-based proteomics reveals nine target proteases for the recombinant protein-stabilizing inhibitor SlCYS8 in Nicotiana benthamiana

Abstract

Co-expression of protease inhibitors like the tomato cystatin SlCYS8 is useful to increase recombinant protein production in plants, but key proteases involved in protein proteolysis are still unknown. Here, we performed activity-based protein profiling to identify proteases that are inhibited by SlCYS8 in agroinfiltrated Nicotiana benthamiana. We discovered that SlCYS8 selectively suppresses papain-like cysteine protease (PLCP) activity in both apoplastic fluids and total leaf extracts, while not affecting vacuolar-processing enzyme and serine hydrolase activity. A robust concentration-dependent inhibition of PLCPs occurred in vitro when purified SlCYS8 was added to leaf extracts, indicating direct cystatin-PLCP interactions. Activity-based proteomics revealed that nine different Cathepsin-L/-F-like PLCPs are strongly inhibited by SlCYS8 in leaves. By contrast, the activity of five other Cathepsin-B/-H-like PLCPs, as well as 87 Ser hydrolases, was unaffected by SlCYS8. SlCYS8 expression prevented protein degradation by inhibiting intermediate and mature isoforms of granulin-containing proteases from the Resistant-to-Desiccation-21 (RD21) PLCP subfamily. Our data underline the key role of endogenous PLCPs on recombinant protein degradation and reveal candidate proteases for depletion strategies.

Keywords: Nicotiana benthamiana; SlCYS8; activity-based protein profiling; cystatin; papain-like cysteine proteases; protease inhibitor; proteomics.

Figure 1

Sl CYS8 inhibits papain‐like cysteine proteases (PLCPs) in the apoplast of N. benthamiana leaves. Apoplastic fluids from plants agroinfiltrated with an empty vector control (EV) or transiently expressing Sl CYS8 or its inactive mutant ^{Q47 P} S l CYS8 were isolated and incubated with MV201 or fluorophosphonate (FP)‐TAMRA probes to target (a) papain‐like cysteine proteases (PLCPs) and (b) Ser hydrolases (SHs), respectively. Labelled proteins were detected by in‐gel fluorescence scanning. Proteins were electrotransferred for immunodetection of Sl CYS8 in the apoplast. Arrows show signals with different fluorescence intensities in plants expressing Sl CYS8. As controls, apoplastic fluids were mixed and pre‐incubated with specific chemical inhibitors (E‐64 for PLCPs and DCI for SHs) before incubation with or without (Ctrl) the probes. Coomassie blue‐stained gels are shown as loading controls.

Figure 2

PLCPs are suppressed upon Sl CYS8 expression in total leaf extracts. Activity profiles of PLCPs (a), vacuolar‐processing enzymes (b) and Ser hydrolases (c) displayed by labelling with MV201, JOPD1 and FP‐TAMRA, respectively. Plants were agroinfiltrated with an empty vector control (EV) or Sl CYS8 or its inactive mutant ^{Q47 P} S l CYS8. Total leaf extracts isolated at 6 dpi were labelled with respective probes and fluorescent‐labelled proteins were detected by in‐gel fluorescence scanning. Arrows show signals with different fluorescence intensities in plants expressing Sl CYS8. The large subunit of ribulose‐1,5‐bisphosphate carboxylase/oxygenase (RbcL) is stabilized during labelling in Sl CYS8‐expressing leaves. Proteins were electrotransferred for immunodetection of Sl CYS8. As controls, leaf extracts were mixed and pre‐incubated with specific chemical inhibitors (E‐64 for PLCPs, YVAD for VPEs and DCI for SHs) before incubation with or without (Ctrl) the probes. Coomassie blue‐stained gels are shown as loading controls.

Figure 3

Purified Sl CYS8 suppresses labelling of PLCPs in total leaf extracts and apoplastic fluids. (a) Total extracts (TE) and (c) apoplastic fluids (AF) from plants agroinfiltrated with an empty vector control were isolated and pre‐incubated with various concentrations of Sl CYS8 purified from E. coli, prior to PLCP labelling with MV201. Quantification of fluorescent band intensity (b and d) shows that PLCP activity is dependent on Sl CYS8 concentration. As a control, plant extracts were mixed and pre‐incubated with E‐64 inhibitor before MV201 labelling. A Coomassie blue‐stained gel shows Sl CYS8 in the samples.

Figure 4

Activity‐based proteomics reveals nine endogenous PLCPs that are inhibited by Sl CYS8. Total leaf extracts were isolated from plants expressing Sl CYS8 (in vivo) or expressing an empty vector (EV), pre‐incubated with or without 280 nm of purified Sl CYS8 (in vitro). Proteins were labelled with DCG04 and FP‐biotin to label PLCPs and Ser hydrolases, respectively. Labelled proteins were purified and identified by MS. (a) Heatmap showing active enzymes that were significantly enriched (FDR < 0.05) compared to a no‐probe control, fold inhibition serves as a proxy for activity. Most PLCP activities were depleted upon Sl CYS8 expression, while no Ser hydrolase was inhibited. (b) Identified PLCPs were grouped by subfamilies and the activity was normalized to the EV control (100%). Bars are the means of three biological replicates ± SE. Star indicates significant differences from the EV control (Student's t test, P < 0.05). PLCP subfamilies: Xylem and bark Cys proteases (XBCPs), Xylem‐specific Cys proteases (XCPs), Resistant‐to‐Desiccation‐21 proteases (RD21s), Resistant‐to‐Desiccation‐19 proteases (RD19s), Aleurain‐Like proteases (ALPs) and Cathepsin‐B‐like proteases (CTBs). n.d.; not detected. MS dataset is presented in Table S1.

Figure 5

Sl CYS8 inhibits different isoforms of granulin‐containing proteases. (a) Leaf extracts from agroinfiltrated plants overexpressing C14, the tomato RD21‐like protease, with or without Sl CYS8 were labelled with MV201 to monitor PLCP activity. In‐gel fluorescence confirmed inhibition of C14 upon Sl CYS8 co‐expression, while no effect was observed when co‐expressed with the inactive ^{Q47 P} S l CYS8 mutant. As a control, plant extracts were mixed and pre‐incubated with E‐64 inhibitor before incubation with or without (Ctrl) the probe. (b) Immunoblot shows equal amount of the intermediate protease (iC14) and the mature (mC14) form of C14 protease. (c) Coomassie gel shows that protein degradation occurred upon C14 expression and is prevented by co‐expressing Sl CYS8. The large subunit of ribulose‐1,5‐bisphosphate carboxylase/oxygenase (RbcL) is stabilized in Sl CYS8‐expressing leaves.