A tobacco etch virus protease with increased substrate tolerance at the P1' position

Abstract

Site-specific proteases are important tools for in vitro and in vivo cleavage of proteins. They are widely used for diverse applications, like protein purification, assessment of protein-protein interactions or regulation of protein localization, abundance or activity. Here, we report the development of a procedure to select protease variants with altered specificity based on the well-established Saccharomyces cerevisiae adenine auxotrophy-dependent red/white colony assay. We applied this method on the tobacco etch virus (TEV) protease to obtain a protease variant with altered substrate specificity at the P1' Position. In vivo experiments with tester substrates showed that the mutated TEV protease still efficiently recognizes the sequence ENLYFQ, but has almost lost all bias for the amino acid at the P1' Position. Thus, we generated a site-specific protease for synthetic approaches requiring in vivo generation of proteins or peptides with a specific N-terminal amino acid.

Figure 1. Activation of the N-degron is rate limiting during substrate depletion by the TIPI system.

A) TIPI efficiency is influenced by three factors, synthesis of the TEV protease by the galactose-inducible GAL1 promoter, proteolysis of the recognition sequence, and degradation of the target protein by the ubiquitin-proteasome system. A reporter protein consisting of two fluorescent proteins (cyan and red) fused together by the TDegX sequence containing the TEV protease recognition sequence (X= amino acid at position P1') and the N-degron sequence. Please note that we follow the original classification of stabilizing and destabilizing residues without considering N-degrons that are produced by N-acetylation. In our constructs, histidine follows X, which prevents acetylation of X in case of Met, Ala, Ser, Cys, Thr, and Val. B) In vivo analysis of the P1' specificity of the pTEV⁺ protease. Processing of the tester constructs CFP-TDegX-RFP (plasmid based) was observed after induction of pTEV⁺ protease production (P_GAL1 -pTEV ⁺ in strain YCT1169) by addition of galactose (2% final concentration). Total cell extracts were fractionated by SDS-polyacrylamide electrophoresis, followed by immunoblotting with antibodies directed against GFP, tRFP and Tub1 (loading control). C) Quantification of X-RFP depletion. RFP fluorescence (same constructs as in B) was measured by a fluorimeter after induction of pTEV⁺ protease synthesis (left graph) and the depletion efficiency of the different substrates was calculated (right graph). Curves are mean values of at least four measurements, normalized to initial RFP fluorescence. Depletion efficiency is represented by the area above each curve (error bars: SEM). D) C-terminal truncation of the TEV protease at position 224 does not influence its activity. The abundance of the tester substrate CFP-TDegF-RFP was followed over time after expression of different pTEV protease variants by fluorimeter measurements (conditions as in C). The plasmid pDS7 was used to express the substrate in yeast strains YCT1243 and YCT1244; error bars represent standard deviation; each construct was measured at least five times.

Figure 2. Generation of a TEV protease that cleaves efficiently the recognition sequence ENLYFQ-R.

A) Scheme of the construct used for the screening procedure: The bidirectional degron module GFP-cODC1-TDegX-RFP (X = F or R) was fused to the phosphoribosylaminoimidazole carboxylase Ade2. Cleavage by the TEV protease leads to activation of the C-degron cODC1 and the N-degron TDegX resulting in proteasomal degradation of Ade2-GFP-cODC1 as well as TDegX-RFP. B) Test for adenine biosynthesis in cells bearing different degron constructs fused chromosomally to ADE2. The yeast strains (ESM356-1, YCT1266, and YCR8) were grown in patches on solid media (YPD, YP+galactose, yeast nitrogen base + 2% glucose, and yeast nitrogen base + 2% galactose; from left to right). C) Scheme illustrating the mutagenesis and selection procedure to obtain a TEV protease which efficiently processes the recognition sequence ENLYFQ-R (left side). The plate is an example to show the difference in color of clones with efficient proteolysis of ENLYFQ-R (red colonies) and clones with insufficient proteolysis (white colonies). Please note that the high degree of red colonies was obtained because the R345G mutant was generated already in the first round of mutagenesis and enriched in subsequent rounds. D) Expression of pTEV2 protease (plasmid-based, R345G mutant) using the GAL1 promoter induces the adenine auxotrophy phenotype in ade2-GFP-cODC1-TDegR-RFP cells (YCR6). Serial dilutions (1:10) were grown on solid media as in B.

Figure 3. In vivo analysis of the P1' Specificity of the pTEV2 protease.

A) Processing of the tester constructs CFP-TDegX-RFP (plasmid encoded) was observed after induction of pTEV2 protease production (P_GAL1 -pTEV2 in yeast strain YCR56). Conditions as in Figure 1B. B) Quantification of the P1' Specificity of the pTEV2 protease. Decrease of full length tester construct after two hours was normalized to initial values and relative efficiency normalized to proline was calculated (cleavage efficiency = ([X]_2h/[Pro]_2h×100-100) ×(−1)), assuming that the recognition sequence with proline at the P1’ Position is not cleaved at all. For each construct two immunoblotting experiments were quantified. Values for constructs with Arg and Phe at the P1’ Position cleaved by the pTEV⁺ protease obtained at the same time are shown as reference. Yeast strains YCR56 (pTEV2 protease production) or YCT1169 (pTEV⁺ protease production) harboring plasmid-based constructs were used for the measurements. C) Quantification of X-RFP depletion. The RFP fluorescence was analyzed by fluorimeter measurements after induction of pTEV2 protease synthesis (upper graph, conditions as in Figure 1C) and the depletion efficiency was calculated (error bars: SEM of at least three experiments). Same constructs as in B. The difference between the arginine construct cleaved by pTEV2 and pTEV⁺ protease is very significant (unpaired t test; p = 0.007).

Figure 4. Influence of the P2' residue on substrate degradation.

A) Analysis of tester construct proteolysis and depletion as well as TEV protease production by immunoblotting. Tester constructs (plasmid based): CFP-TDegXY-RFP, XY=RH, RL, X corresponds to the P1' position, Y to the P2' position; proteases: pTEV⁺ (yeast strain YCT1169), pTEV2 (YCR56). Conditions as in Figure 1B; antibodies directed against tRFP, GFP, TEV, and Tub1 (loading control) were used to obtain the immunoblot. B) The RFP fluorescence of the tester constructs CFP-TDegXY-RFP was followed over time after induction of TEV protease synthesis by fluorimeter measurements (three measurements for each construct; error bars indicate the standard error of the mean; same constructs as in A).

Figure 5. Structural comparison of the TEV protease with the R203G mutant.

A) Ribbon structure of the TEV protease (green) was overlaid with the mutant (dark cyan). The structure of the R203G mutant, which corresponds to R345G in the pTEV⁺ protease, was obtained by homology modeling using an x-ray structure of the TEV protease as template. Views from three different sides are shown. The residues of the catalytic triad H46, D81, and C151 are indicated (TEV protease: blue; mutant: magenta). The two arginine residues close to the catalytic center (R49, R50) are shown in yellow (TEV protease) and light magenta (G203 mutant). The R203 residue is shown in orange, the G203 in red. The two β-sheets, which are mentioned in the text that close the catalytic center are marked by asterisks. B) Surface charge distribution of the TEV protease compared to the R203G mutant. Surface charges were calculated using the software package MolMol. Positive charge is represented by blue color, negative charge by red color.

Figure 6. Comparison of the apparent N-degron strength generated and measured by different methods.

Apparent N-degron strength ordered from high to low destabilizing activity. Pulse chase data were obtained by Bachmaier et al., 1989, fluorescent timer-based measurements by Khmaelinski et al., 2012, fluorescence-based measurements during this study. A color code indicates whether an amino acid is a primary (dark cyan), secondary (blue) or tertiary (red) destabilizing residue at the amino-terminus of a protein, or if it is stabilizing (gold) in the absence of N-acetylation.