Intracellular Detection and Evolution of Site-Specific Proteases Using a Genetic Selection System
Kathryn D.Verhoeven, Olvia C. Altstadt, Sergey N. Savinov
Department of Chemistry, Purdue University, 560 Oval Dr., West Lafayette, IN 47907;
e-mail:
Supporting Information
Table of Contents
Supporting Figures and Tables:
Figure S1. Performance of SCR-integrated reporter strains 2
Figure S2. Mock Selection 3
Figure S3. SDS PAGE analysis of proteolytic processing 4
Figure S4. Solution-phase characterization of TEV-Pr activity 5
Figure S5. Solid-phase implementation of fluorogenic assay 6
Figure S6. Comparative kinetic analysis of evolved proteases. 7
Figure S7. SDS-PAGE gel of immobilized protease samples. 8
Figure S8. Representative structure of fluorogenic peptide substrates. 8
Figure S9. Location of amino acid mutations in evolved proteases 9
Figure S10. Ribbon diagram of active TEV-Pr with bound substrate 10
Table S1. Michaelis-Menten parameters for soluble TEV-Pr 11
Table S2. Vmax/KM values and selectivity ratios for immobilized TEV-Pr and mutant proteases 11
Supporting Methods
Construction of plasmids 12
Construction of strains 13
Solution-phase fluorogenic assay 13
Solid-phase fluorogenic assay 14
Supporting Discussion
Mock Selection 15
Solution-phase kinetic assay with fluorogenic peptides 16
Solid-phase assay with fluorogenic peptides 17
Mutational Analysis 18
Supporting References 19
Supporting Figures and Tables
Figure S1. Performance of the SCR-integrated reporter strains. Reporter strains integrated with nil control (1), wtSCR (2), scrSCR (3), and wkSCR (4) repressors were analyzed by (A) ONPG assay under SCR induction (10 μM IPTG) and growth assay on (B) inducing (LB supplemented with 10 μM IPTG), and (C) selecting (LB supplemented with 10 μM IPTG and 25 mg/mL kanamycin) media.
Figure S2. Mock selection experiment. The strain expressing wtSCR (OC85) was transformed with plasmid mixtures containing pAR-wtTEV and pAR-mtTEV in 1:105 and 1:103 ratios. Transformants were plated on selective media at a density of 105 cfu per plate. The photographs of high-dilution and low dilution plates after a 72-hour incubation at 37 ˚C are shown in panels A and B, respectively. The single surviving colony in panel A is indicated by an arrow, with an enlarged view provided in the inset. The surviving colonies were analyzed by PCR (panel C). Lanes 1 and 11 are molecular weight markers, lanes 2–10 are PCR products of pAR-wtTEV (lane 2), pAR-mtTEV (lane 3), the surviving colony from plate A (lane 4) and survivors from plate B (lanes 5–10).
Figure S3. SDS-PAGE analysis of SCR processing (wtSCR, wkSCR, scrSCR) by co-expressed TEV-Pr variants: wt TEV-Pr and related mutants C151A TEV-Pr, M1, M2, and M3. The assay was performed under constant induction of SCRs (100 μM IPTG) and proteases (1.3 mM Ara).
Figure S4. Solution-phase characterization of TEV-Pr activity with fluorogenic substrates. Time-course of fluorescence change exhibited by soluble wt TEV-Pr (1.2 μM) in the presence of native (H-(2-ABz)-ENLYFQG-(3-NT)-D-OH, dashed line), weak (H-(2-ABz)-ENLYFQD-(3-NT)-D-OH, dotted line), and scrambled (H-(2-ABz)-GYFELNQ-(3-NT)-D-OH, solid line, coincides with x-axis) substrates (60 mM).
Figure S5. Solid-phase implementation of the fluorogenic assay. Time-course of fluorescence change exhibited by wt TEV-Pr (A) and C151A TEV-Pr (B), immobilized through polyhistidine tags on metal-affinity resin, in the presence of native H-(2-ABz)-ENLYFQG-(3-NT)-D-OH, (blue), weak H-(2-ABz)-ENLYFQD-(3-NT)-D-OH (red), and scrambled H-(2-ABz)-GYFELNQ-(3-NT)-D-OH (black) substrates (40 mM).
Figure S6. Comparative kinetic analysis of evolved proteases. Time course of processing P1’-glycine (blue) and P1’-aspartate (red) substrates (40 mM each) by wt TEV (A), M1 (B), M2 (C), and M3 (D) proteases.
Figure S7. SDS-PAGE gel of relative immobilized protease sample concentration levels (27 kDa). Corresponding equivalent volumes of immobilized protease samples were visualized by SDS-PAGE, subjected to quantitative image analysis software, and used to normalize protease sample concentrations for the solid-phase fluorescence assays shown in Figure 5 and Figure S6.
Figure S8. Representative structure of fluorogenic peptides, showing substrates corresponding to the P1’-glycine (R = H) and P1’-aspartate (R = CH2COO-) fluorogens.
Figure S9. Positions of mutations in evolved proteases. Sequence alignment of wt TEV-Pr with evolved proteases M1, M2, and M3, emerging from the first, second, and third rounds of directed evolution, respectively. Catalytic triad residues (red) and amino acid residues corresponding to substrate recognition segments (gray highlight) are indicated. Mutations introduced throughout the evolution iterations are highlighted in yellow.
Figure S10. Ribbon diagram (pdb code: 1LVM) of active TEV-Pr (grey) with bound hydrolysis product Ac-ENLYFQ-OH (blue). Substrate recognition regions (yellow) and catalytic triad residue side chains (red) are indicated within the diagram and the locations of amino acid mutations in the selected proteases M1, M2, and M3 (green) are labeled by arrows, with mutations introduced in the second and third rounds of evolution designated by solid and dashed underlines, respectively.
Table S1. Michaelis-Menten parameters for soluble TEV-Pr obtained with fluorogenic substrates.a
Substrate / KM (mM) / kcat (s-1) / kcat/KM (´ 10-3 M-1s-1)2-ABz-ENLYFQG-(3-NT)-D / 0.09 ±0.02 / 0.20 ± 0.05 / 2.2 ± 0.7
2-ABz-ENLYFQD-(3-NT)-D / 0.07 ± 0.01 / 0.0058 ± 0.0008 / 0.083 ± 0.016
a The kinetic parameters were obtained as described in the corresponding supporting discussion section below.
Table S2. Vmax/KM values and selectivity ratios for immobilized TEV-Pr and mutant proteases.a
Vmax/KM (´ 106 s-1) / Selectivity ratioProtease / wt sbsb / wk sbsb / wk sbs to wt sbs
wt TEV / 30.7 ± 0.4 / 0.391 ± 0.004 / 0.013
M1 / 23.7 ± 0.4 / 0.673 ± 0.003 / 0.028
M2 / 0.70 ± 0.01 / 0.888 ± 0.003 / 1.3
M3 / 0.33 ± 0.01 / 1.331 ± 0.005 / 4.0
a Rate constants were obtained as described in the supporting discussion section.
b Abbreviations correspond to P1’-glycine fluorogenic substrate (wt sbs) and P1’-aspartate fluorogenic substrate (wk sbs)
Supporting Methods
Construction of plasmids
pOC1: The gene coding for 434 DBD was amplified by PCR from pTHCP14 [S1], a derivative of pMAL-c2x (New England Biolabs), with primers 5’-GTTGTTGAGCTCTCGCGAGGATCCGTCGA-CCTTGATATCGTAGGGTTCACA-3’ and 5’-GTTGTTGGTACCATGAGTATTTCTTCCAGGGT-AAAAAGC-3’, designed to eliminate a stop codon and the Shine-Delgarno sequence in the original bicistronic construct. The product of PCR, processed by KpnI and SacI, was subcloned into pTHCP14, linearized with the same restriction enzymes.
pOC2: The cognate TEV-Pr substrate containing the GTTENLYFQSG peptide sequence as an SCR linker was constructed by inserting pre-annealed oligonucleotides 5’-TCGAGGGCACCACCGAAAA-TCTGTATTTTCAGAGCGGTGGTAC-3’ and 5’-CACCGCTCTGAAAATACAGATTTTCGGTGG-TGCCC-3’ into KpnI and XhoI sites of pTHCP14.
pOC4: An SCR containing a non-substrate GSTYFETLNQG sequence was constructed as described for pOC2, using 5’-TCGAGGGCAGCACCTATTTTGAAACCCTGAATCAGGGTGGTA-3’ and 5’-CACCCTGATTCAGGGTTTCAAAATAGGTGCTGCCC-3’ oligonucleotides.
pOC21: A weak substrate SCR containing an Asp mutation at the P1’ position (GTTENLYFQDG) was produced as described for pOC2 with the following oligonucleotides: 5’-TCGAGGGCACCACCGAAAATCTGTATTTTCAGGATGGTGGTAC-3’ and 5’-CACCATCCTGA-AAATACAGATTTTCGGTGGTGCCC-3’.
pET28-wtSCR, pET28-scrSCR & pET28-wkSCR: The SCR genes were subcloned from pOC2, pOC4 and pOC21 into pET28a(+) vector (Novagen) using unique NdeI and SacI sites, generating pET28-wtSCR, pET28-scrSCR, and pET28-wkSCR, respectively.
pAR-wtTEV: TEV-Pr gene was amplified from the pRK793 vector (Science Reagents, Inc) using the following primers: 5’-GAAGAACATGGGAAGCGGTCATCATCATCATCATCATCATGGAG-3’ and 5’-GTTGTTTCTAGATTAATTCATGAGTTGAGTCGCTTCC-3’. The amplified gene was subcloned into the Ara-inducible pARCBD-p expression vector [S2] via PCR-installed NcoI and XbaI sites, yielding pAR-wtTEV.
pAR-mtTEV: A gene coding for an inactive TEV-Pr mutant (C151A) was accessed via a ‘megaprimer’ PCR method [S3], using the following site-directed mutagenesis primer 5’-GGATGGGCAGGCTGG-CAGTCCATTAG-3’. The altered gene was then incorporated into the pARCBD-p vector as described for pAR-wtTEV.
Construction of strains
Integration of the repressor fusions into E. coli strain SNS126 [S1] was performed using a conditional-replication, integration, and modular (CRIM) plasmid system [S4]. Strains expressing integrated substrate, non-substrate, and weak substrate SCRs were designated OC85, OC86, and OC108, respectively. Control strain OC83 (nil integrant) was generated by integrating the unmodified CRIM plasmid.
Solution-phase fluorogenic assay
TEV-Pr was expressed from the commercially available BL21(DE3)-pRIL/pRK793 system (Science Reagents, Inc), following the provided protocol [S5]. The induced cells were lysed using BugBuster® reagent, and a soluble fraction was passed through the TALON metal-affinity beads (BD Bioscience). Treatment of the beads with 200 mM imidazole provided eluates containing purified protease, which were dialyzed overnight into 25 mM NaH2PO4 buffer (pH 8.0) with 10 % glycerol, 200 mM NaCl, 2 mM EDTA, and 10 mM DTT. Protease kinetics were evaluated using the following fluorogenic peptides (>97% purity; Chem-Impex International, Inc.): H-(2-ABz)-ENLYFQG-(3-NT)-D-OH (native), H-(2-ABz)-ENLYFQD-(3-NT)-D-OH (weak), and H-(2-ABz)-GYFELNQ-(3-NT)-D-OH (scrambled). The peptides were incubated with the expressed and purified TEV-Pr at 25 ˚C in 50 mM Tris pH 8.0 buffer containing 2 mM EDTA and 10 mM DTT. Upon internal cleavage of both the native and weak substrates, the fluorescence increase of the unquenched anthranilamide fragment was monitored in real time at 420 nm by excitation at 320 nm.
Solid-phase fluorogenic assay
Plasmids coding for the selected protease were transformed into the BL21 (DE3) expression strain. Protease expression was performed in 25 mL of NZCYM Broth (USB), supplemented with Cm, by inoculating with 250 mL of BL21-protease pre-culture and incubated at 37 oC in a shaker until OD600 reached 0.6. Induction with 1.3 mM Ara was performed at 37 oC overnight, and then cultures were chilled on ice prior to centrifugation at 4000 rpm for 15 min at 4 ºC. The cells were resuspended in 1 mL of lysis buffer (50 mM NaH2PO4, 100 mM NaCl, pH 8.0) and lysed with BugBuster® reagent (Novagen). After the samples were centrifuged at 12,000 rpm for 20 min at 4 ºC, the pellets were solubilized in 1 mL of a denaturing buffer (50 mM NaH2PO4, 50 mM Tris-Cl, 100 mM NaCl, 8 M urea, pH 8.0). Samples were centrifuged at 12,000 rpm for 15 min at 4 oC to remove insoluble aggregates, and then the supernatant was incubated with the TALON metal-affinity resin (BD Bioscience). After washing the resin with the denaturing buffer, the samples were incubated (20 min) with a refolding buffer (50 mM Tris-Cl, 10 % glycerol, pH 8.0) and subsequently washed with the same solution. The bead-immobilized protease samples were resuspended in a protease activity buffer (50 mM Tris-Cl, 10 % glycerol, 1 mM EDTA, 5 mM DTT, pH 8.0) and combined with variable concentrations (20–80 mM) of the fluorogenic substrates in the wells of a black, flat-bottom 96-well plate (Greiner Bio-One). Proteolysis was evaluated using the following fluorogenic peptides (>97% purity; Chem-Impex International, Inc.): H-(2-ABz)-ENLYFQG-(3-NT)-D-OH (native), H-(2-ABz)-ENLYFQD-(3-NT)-D-OH (weak), and H-(2-ABz)-GYFELNQ-(3-NT)-D-OH (scrambled). The samples were monitored in real time by exciting the 2-aminobenzamide fluorophore at 320 nm and collecting emission data at 420 nm using a BioTek Synergy HT bioplate reader. To account for resin interference and photobleaching, resin-only samples were prepared, tested in duplicate, and used to calculate adjusted fluorescence for each protease-resin sample. Relative concentrations of the proteases were established by a quantitative SDS-PAGE analysis using image analysis software [S6].
Supporting Discussion
Mock Selection
To test the performance of the system under the conditions of selection stress, biased combinations of pAR-wtTEV and pAR-mtTEV with a large excess of the latter (1:103 and 1:105) were prepared to simulate the highly underrepresented occurrence of active individuals in naive libraries. The resulting plasmid mixtures were transformed into the wtSCR strain (OC85), and the cell suspensions of transformants were plated at densities of approximately 105 cfu per plate on the selective medium (MMA supplemented with 2% glycerol, 1 mM MgSO4, 10 μg/ml thiamine, 10 mM 3AT, 75 μg/ml Kan, 13 μM Ara, and 10 μM IPTG). The colonies that were visibly growing after 72 hours of incubation were analyzed by a PCR-based selective amplification technique [S7] to reveal a highly reliable retrieval of the rare pAR-wtTEV plasmid, even from a 1:105 mixture (Figure S2 A–C). The exclusive recovery of the plasmid encoding the active protease, as confirmed by plasmid sequencing, is indicative of high genetic stability of the integrated reporter system and low propensity for yielding false-positives. The overall success of this experiment demonstrates the power of the selection-based assay to identify rare functional properties from a large collection of candidates.
Solution-phase kinetic assay with fluorogenic peptides
To confirm and quantify substrate preferences of TEV-Pr and evolved protease, fluorogenic versions of the test sequences were developed by installing a fluorescent 2-aminobenzoic (2-ABz) unit at the N-termini quenched by 3-nitrotyrosine (3-NT) at the opposite end along with an aspartic acid residue for improved solubility [S8]. The peptides—H-(2-ABz)-ENLYFQG-(3-NT)-D-OH (native), H-(2-ABz)-ENLYFQD-(3-NT)-D-OH (weak), and H-(2-ABz)-GYFELNQ-(3-NT)-D-OH (scrambled)—were incubated with expressed autocatalysis-resistant (S219V) TEV-Pr [S5] at 25 oC in 50 mM Tris pH 8.0 buffer containing 2 mM EDTA and 10 mM DTT. Upon internal cleavage of both the native and weak substrates, the fluorescence increase of the unquenched anthranilamide fragment was monitored in real time at 420 nm by excitation at 320 nm (Figure S4). The observed preference for processing the native sequence over the weak substrate is consistent with the significant differences obtained in the growth and ONPG assays of the corresponding SCR-expressing strains (Figure 2). No fluorescence changes were detected with the scrambled substrate, as expected.
To obtain Michaelis-Menten kinetic parameters, the initial proteolysis rates for the recombinant TEV-Pr were collected at variable substrate concentrations (5–100 mM) by exciting the fluorophore at 320 nm and collecting emission data at 420 nm using a BioTek Synergy HT bioplate reader. Non-linear regression analysis (KaleidaGraph, Synergy Software) was used to obtain relevant kinetic parameters. Michaelis-Menten analysis (Table S1) of the enzymatically-active pairs confirmed the poor processing of the substrate, due to the presence of the charged aspartate residue at the P1’ position [S5]. The KM, kcat, and kcat,/KM values for the fluorogenic substrates match closely previously reported data for P1’-glycine (or serine) and P1’-aspartate substrates [S9]. The observed relaxation of the substrate scope for the early M1 protease mutant, combined with the relatively small activity enhancements toward the targeted substrate is a fairly common phenomenon seen in the early stages of directed enzyme evolutions [S10]. As evolution progresses, however, the lack of selective pressure to maintain the original substrate preference and associated toxicity of promiscuous mutants appear to act in concert in narrowing the observed substrate scope. The genetic system discussed herein appears to be, therefore, well suited for applying a substantial level of counter pressure in the search for site-selective proteases rather than enzymes with enhanced scope.