https://orcid.org/0000-0002-3644-3028
Figures
Abstract
The multifunctional autoprocessing repeat-in-toxin (MARTX) toxin is the primary virulence factor of Vibrio vulnificus displaying cytotoxic and hemolytic properties. The cysteine protease domain (CPD) is responsible for activating the MARTX toxin by cleaving the toxin precursor and releasing the mature toxin fragments. To investigate the structural determinants for inositol hexakisphosphate (InsP6)-mediated activation of the CPD, we determined the crystal structures of unprocessed and β-flap truncated MARTX CPDs of Vibrio vulnificus strain MO6-24/O in complex with InsP6 at 1.3 and 2.2Å resolution, respectively. The CPD displays a conserved domain with a central seven-stranded β-sheet flanked by three α-helices. The scissile bond Leu3587-Ala3588 is bound in the catalytic site of the InsP6-loaded form of the Cys3727Ala mutant. InsP6 interacts with the conserved basic cleft and the β-flap inducing the active conformation of catalytic residues. The β-flap of the post-CPD is flexible in the InsP6-unbound state. The structure of the CPD Δβ-flap showed an inactive conformation of the catalytic residues due to the absence of interaction between the active site and the β-flap. This study confirms the InsP6-mediated activation of the MARTX CPDs in which InsP6-binding induces conformational changes of the catalytic residues and the β-flap that holds the N terminus of the CPD in the active site, facilitating hydrolysis of the scissile bond.
Citation: Chen L, Khan H, Tan L, Li X, Zhang G, Im YJ (2024) Structural basis of the activation of MARTX cysteine protease domain from Vibrio vulnificus. PLoS ONE 19(8): e0307512. https://doi.org/10.1371/journal.pone.0307512
Editor: Israel Silman, Weizmann Institute of Science, ISRAEL
Received: March 1, 2024; Accepted: July 7, 2024; Published: August 2, 2024
Copyright: © 2024 Chen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The coordinates and structure factors of the pre-CPD C3727A and the post-CPD Δβ-flap have been deposited in the Protein Data Bank with the accession codes, 8YJC and 8YJA, respectively.
Funding: This work was supported by National Research Foundation of Korea (NRF) grants to Y.J.I funded by the Ministry of Education, Science and Technology (Grant No. RS-2023-00241410 and RS-202300221356) and a grant from the Ministry of Oceans and Fisheries R&D project, Korea (Grant No. RS-2021-KS211513).
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Vibrio vulnificus is an opportunistic pathogen that can cause gastroenteritis, wound infection, and septicemia by consumption of contaminated seafood or exposure of skin wounds to contaminated water [1]. Three biotypes of V. vulnificus have been identified based on biochemical characteristics and phylogeny. Biotype 1 including MO6-24/O strain is responsible for the majority of ingestion cases and wound infections. Biotype 2 is a pathogen to farmed eels. Biotype 3 strains such as BAA87 cause wound infections and show hybrid characteristics of biotypes 1 and 2 [1, 2].
V. vulnificus secretes a large multifunctional-autoprocessing repeats-in-toxin (MARTX) as a main virulence factor via an atypical type I secretion system [3]. MARTX is a single polypeptide toxin consisting of central multiple effector domains and repeats-containing arms at the N- and C-termini [4]. The central region of the MARTX toxin of V. vulnificus MO6-24/O contains an autoprocessing cysteine protease domain (CPD) and four effector domains including the domain of unknown function (DUF1), the Rho-inactivating domain (RID), an αβ hydrolase (ABH), and makes caterpillars floppy-like domain (MCF) [4]. The MARTX toxins of the different V. vulnificus subtypes display a variation in the effector domains. CMCP6 and YJ016 strains have an additional Ras/Rap1-specific endopeptidase (RRSP) domain that dysregulates host cell signaling. The BAA87 strain does not contain MCF and RRSP but possesses ExoY-like adenylate cyclase domain (ExoY) and domain X (DmX) [4]. Upon binding to the host cell plasma membrane, the MARTX toxin forms a pore using repeat-containing arms and translocates the effector domains into the cytosol. The CPD directs the proteolytic processing of effector modules after activation by binding to the eukaryotic cell-specific inositol hexakisphosphate (InsP6) [5, 6]. The processing of the first cleavage site of the CPD exposes the active site to other cleavage sites within the toxin, subsequently generating active toxin fragments. DUF1 interacts with prohibitin 1 in the host membrane serving as an initial receptor for the binding of MARTX toxins. RID causes host cell rounding by inactivating Rho-family GTPases that regulate cell cytoskeleton. ABH is a phospholipase A1 intefering with the autophagic pathway and endosomal trafficking. MCF causes apoptotic cell death and Golgi disruption. Consequently, activated toxin effectors exhibit cytotoxic and hemolytic activities and induce host cell death [7–9].
The CPD displays a canonical caspase-like fold with a central β-sheet surrounded by α-helices. InsP6 binds to the CPD with the Kd of 1–2 μM and activates toxin autoprocessing [5]. InsP6 is abundant in the cytosol of mammalian cells at concentrations between 5–100 μM with a long half-life [10, 11]. InsP6 acts as an activator by binding to the site distinct from the active site of CPD. In addition, InsP6-binding stabilizes the CPD structure, facilitating the formation of the enzyme-substrate complex [6, 12]. The MARTX CPD of MO6-24/O strain shows a sequence similarity of 77% to the CPD of the BAA87 strain. The structure of MARTX CPD from V. vulnificus BAA87 biotype 3 was reported revealing attenuated activity of the CPD due to the distinct C-terminal β-flap region [13]. Currently, the MARTX CPD of the MO6-24/O strain is not known and all the reported structures of the MARTX CPDs have active conformations with InsP6-bound forms. The structure of apo-CPD with the inactive conformation is unknown.
Here, we present the high-resolution structures of the MARTX CPD–InsP6 complex of V. vulnificus MO6-24/O. Our results provide a structural basis for the InsP6-activated autoprocessing of CPD by rearrangement of the catalytic residues and the β-flap that holds the N-terminus of the CPD in the active site. After autoprocessing, the CPD adopts a postprocessing form that has a reduced affinity for InsP6 due to the loss of interaction between the β-flap and the N-terminal residues. The structure of the CPD Δβ-flap reveals an inactive conformation of the catalytic residues due to the absence of the interaction with the β-flap, providing a structural insight into the conformational switching of the CPD by InsP6-binding.
2. Materials and methods
2.1. Cloning of cysteine protease domain (CPD) of V. vulnificus MARTX
The DNA encoding the MARTX CPD (residues 3578–3796) was amplified by a polymerase chain reaction from the full clone of the MARTX toxin (NCBI code: WP_058645630.1) isolated from the MO6-24/O strain of Vibrio vulnificus. The MARTX CPD was subcloned into the BamHI/XhoI site of a modified pHIS2 vector. The MARTX CPD was tagged with the N-terminal hexahistidines followed by a thrombin protease cleavage site (LVPR/GS). The inactive mutant of the MARTX CPD was prepared by the point mutagenesis of the active site residue (Cys3727Ala). Other truncation constructs of the CPD were subcloned to the pHIS2 vector with the same procedure.
2.2. Protein expression and purification
Escherichia coli strain BL21(DE3) cells transformed with the plasmids encoding the MARTX CPD were grown to an OD600 of 0.8 at 37°C in LB medium. Cells were induced by the addition of isopropyl β-D-1-thiogalactopyranoside to a final concentration of 0.5 mM and were incubated for 12 h overnight at 20°C prior to harvesting. The cells expressing the CPD were resuspended in 2X PBS buffer containing 20 mM imidazole (lysis buffer) and lysed by sonication. The supernatant containing the His-tagged MARTX CPD was loaded to a Ni-NTA affinity column. The Ni-NTA column was thoroughly washed with the lysis buffer. The target protein was eluted from the column using a buffer containing 100 mM Tris-HCl pH 8.0, 300 mM imidazole, and 300 mM NaCl. The eluate was concentrated to 10 mg ml-1 using Amicon Ultra-15 centrifugal filter. The His-tag was cleaved by 10 international unit (IU) of thrombin protease (Reyon Pharmaceutical) per 10 mg of recombinant protein. The cleaved sample was subjected to size-exclusion chromatography (SEC) on a HiLoad Superdex 200 column equilibrated with 20 mM Tris-HCl pH 8.0 and 150 mM NaCl. The InsP6-loaded MARTX CPD complex for crystallization studies was prepared by supplementing 5 mM of InsP6 (myo-inositol hexakisphosphate, Merck) to the Ni-NTA affinity elutes and incubating for 1 h at room temperature before SEC. The fractions containing the MARTX CPD were concentrated by the centrifugal filter to 20 mg ml-1 for crystallization.
2.3. Crystallization and crystallographic analysis
For the crystallization of the InsP6-loaded MARTX CPD complex, the purified MARTX CPD (residues 3578–3796) was mixed with the additional five-time molar ratio of InsP6 and incubated at room temperature for 1 h. Preliminary crystallization experiments were carried out at 22°C in 96-well using customized crystallization screening solutions by dispensing 0.8 μl protein solution and 0.8 μl precipitant solution by hanging drop vapor diffusion method. The crystals of the pre-CPD-InsP6 complex were grown in 0.1 M HEPES pH 9.0, 30% PEG 8000 in 5 days.
The crystals of the post-CPD Δβ-flap were obtained using the construct (residue 3592–3796) lacking the N-terminal cleavage sequence. The C-terminal β-flap region of the post-CPD was susceptible to proteolytic degradation during protein purification. The purified post-CPD was mixed with the final 5 mM of InsP6 and incubated for 1 hour. The initial crystal appeared in a week in a solution consisting of 0.1 M Na-Acetate pH 4.6, 30% polyethylene glycol (PEG) 4000, and 0.1 M ammonium acetate. The crystallization condition was further optimized to 0.1 M sodium acetate pH 4.0, 25% PEG 3350, 0.2 M ammonium acetate, and 2.5% ethylene glycol via microseeding in 15-well screw-cap plates. A drop consisting of a 1.5 μl protein solution was mixed with a 1.5 μl precipitant solution and equilibrated against a 1 ml reservoir solution. The microcrystal seeds were introduced to the protein-precipitant mixture in an hour and the high-quality crystals with dimensions of 0.1 X 0.1 X 0.15 mm appeared in a week. The crystals of the MARTX CPDs were cryoprotected in a reservoir solution supplemented with 10% glycerol and flash-cooled by immersion in liquid nitrogen. Crystals were preserved in a cryogenic N2-gas stream (~100K) during diffraction experiments. Native diffraction data for the MARTX CPD were collected at a fixed wavelength of 0.97949 Å using an ADSC Q270 CCD detector on the 7A beamline at Pohang Light Source (PLS), Pohang Accelerator Laboratory. All data were processed and scaled using HKL-2000 [14]. The structure of the pre-CPD loaded with InsP6 was determined by molecular replacement using the structure of MARTX CPD (PDB code: 3EEB) as a search model. The single molecule of the MARTX CPD was found in the asymmetric unit using the program Phaser [15], and the density-modified map showed clear electron densities of the CPD and the bound InsP6. The final model was refined to Rwork and Rfree values of 17.6% and 19.3%, respectively using Phenix [16]. The structure of the post-CPD Δβ-flap was determined by molecular replacement using a truncated pre-CPD model. The crystal of the post-CPD did not contain the β-flap due to the proteolytic degradation during protein purification and crystallization. Two molecules of the CPD coordinating a single molecule of InsP6 in the dimer interface were found in the asymmetric unit. The models were built using the software Coot, and the final models were refined to Rwork and Rfree values of 20.7% and 27.7%, respectively (Table 1). The figures for all PDB structures were drawn using the software PyMOL (https://pymol.org).
2.4. In vitro autocleavage assay of the CPD
To monitor the InsP6-dependent autocleavage activities of CPDs, we used the N-terminal 6xHis-tagged CPD construct with a long N-terminal loop (residues 3570–3796) to observe the size difference clearly after cleavage by SDS-PAGE. The catalytically inactive mutant Cys3727Ala was used as a negative control. The 200 μl of 0.1 mM wildtype or mutant MARTX CPD was mixed with 0.5 mM InsP6 and incubated at room temperature with time intervals of 1 and 2 hours. To check the effect of InsP6 concentration on autoprocessing, 20 μM of the CPD was incubated with InsP6 concentrations ranging from 1 μM to 200 μM for one hour. The autocleavage was examined by SDS-PAGE.
2.5. Isothermal titration calorimetry
The InsP6-binding properties of the MARTX CPDs were quantitatively analyzed by ITC using an Affinity ITC calorimeter (low volume cell 190 μl; TA instruments). All proteins were prepared in the identical buffer containing 20 mM Tris-HCl pH 8.0 and 150 mM NaCl. The syringe was loaded with 1 mM of InsP6, and the cell was filled with 300 μl of 0.1 mM MARTX CPD. The titration curve was obtained by injecting 2 μl × 25 aliquots of the InsP6 into the cell at the time interval of 180 sec at 20°C. The enthalpy of the reaction, ΔH0, the binding constant, Kd, and the stoichiometry value, n, were calculated from the measured heat changes, δHi, upon the association of InsP6 and the MARTX CPD. The affinity Kd was estimated with a confidence level of 95% with a single-site binding model. The titration data were analyzed using the NanoAnalyze program (TA instruments) and fitted into an independent binding model.
3. Results
3.1. Structure of the pre-processed form of the MARTX CPD
V. vulnificus strain MO6-24/O contains a MARTX of 4703 amino acids with four effector domains including DUF1, RID, ABH, and MCF (Fig 1A). For the structural studies of the MARTX CPD, we purified the recombinant CPD (residues 3578–3796) containing the N-terminal ten residues upstream of the autocleavage site. The protein was expressed with an N-terminal His-tag and a thrombin protease recognition site for Ni2+-NTA affinity chromatography. The purified V. vulnificus MARTX CPD was a monomer in size exclusion chromatography (SEC). The recombinant CPD displayed InsP6-dependent autoprocessing activity. The incubation of wild-type CPD with 0.5 mM InsP6 resulted in autocleavage of the Leu3587-Ala3588 bond, while the active site mutant C3727A had no proteolytic activity (Fig 1B). The autocleavage was efficient in a low InsP6 concentration of 2 μM, when 20 μM of CPD was incubated for an hour at room temperature. To examine the activated conformation of the unprocessed CPD upon InsP6-binding, we crystallized the pre-processing form of CPD (here referred to as pre-CPD) bound to InsP6 using the catalytically inactive mutant C3727A. We determined the 1.3Å crystal structure of the CPD-InsP6 complex capturing the scissile bond harbored in the active site. The electron densities of InsP6 and the residues coordinating the ligand were well visible (Fig 1C). There was no residue in the disallowed region of the Ramachandran plot except Asn3686. The outlier residue, Asn3686 of the β5-β6 loop, showed well-defined electron density maps by interacting with Asp3732 of the β7-α3 loop (Fig 1C).
(A) Schematic representation of the domain structures of the MARTX toxin of Vibrio vulnificus strain MO6-24/O. The MARTX toxin contains a CPD (residue 3578–3796) and four effector domains including DUF1, RID, ABH, and MCF. (B) The purified CPD wild type (residue 3570–3796) shows autocleavage activity upon the addition of 0.5 mM InsP6. The active site mutant C3727A has no autocleavage activity. To examine the effect of InsP6 concentration on autocleavage, 20 μM of CPD was incubated with a series of InsP6 concentrations ranging from 1 μM to 200 μM for one hour. (C) The 2Fo-Fc electron density maps of the pre-CPD C3727A crystal with the final models superimposed. The three panels of density maps show the InsP6-binding site, N3686 of the β5-β6 loop, and the N-terminal five residues, respectively. (D) The overall structure of the pre-CPD C3727A bound with InsP6. The N-terminal leader, protease core, and β-flap are colored blue, green, and magenta, respectively, according to the annotation of Prochazkova et al [12]. The scissile residue, Leu3587 is shown as sticks with transparent spheres. (E) The N-terminal autocleavage sequence in the substrate-binding cleft. Leu3587 and the catalytic residues, His3678 and Cys3727Ala are shown with dotted sticks. Hydrogen bonds are shown in dotted lines.
The V. vulnificus MARTX CPD displays a canonical CPD structure with the central seven-stranded β-sheet flanked by three α helices (Fig 1D). The core β-sheet of CPD is composed of 12 β-strands. The N-terminal five residues (TKEAL, 3583–3587) upstream of the scissile bond were ordered in the crystal structure (Fig 1C). The N-terminal leader of 34 residues (Thr3583-Arg3616) upstream of the β1-strand wraps around the CPD surface. The C-terminal β-flap composed of the three anti-parallel β-strands (β9-β11) is positioned between the N-terminal autocleavage sequence and the InsP6–binding site. The side chain of the scissile residue Leu3587 is inserted into the catalytic pocket formed between catalytic residues His3678 and Cys3727. The Cβ atom of the mutated Cys3727Ala points toward the amide bond of the scissile bond, which represents the conformation ready for the catalytic reaction (Fig 1E).
The MARTX CPD of V. vulnificus MO6-24/O has sequence similarities of 87% and 77% to the CPDs of the V. cholerae N16961 and V. vulnificus BAA87, respectively (Fig 2A). The overall structure of V. vulnificus MARTX CPD is highly conserved to the structure of V. cholerae CPD with the Cα r.m.s.d. of 0.25 Å for 164 equivalent Cα atoms (Fig 2B). The scissile residues, Leu3587-Ala3588, and the active site residues of the V. vulnificus CPD have almost identical conformations compared to the residues of the V. cholerae CPD. The sequence upstream of the scissile bond is an extended loop in the V. vulnificus MARTX CPD. However, a TEV protease recognition sequence (ENLYFQS) in the corresponding region of the recombinant V. cholerae CPD forms an α-helix, suggesting that sequence variation in the N-terminal region can be accommodated around the substrate-binding cleft. (Fig 2C).
(A) Multiple sequence alignment of cysteine protease domains from MARTX toxins of various Vibrio strains including V. vulnificus MO6-24/O (WP_015728045.1), V. cholerae N16961 (AAD21057.1), V. vulnificus FORC_009 (WP_060534095.1), and V. vulnificus BAA87 (WP_039507922.1). The autocleavage sites in the N-termini of CPDs were indicated by yellow shades with red arrows. The dotted lines indicate the residues invisible in the electron density maps. (B) Structural comparison of V. vulnificus pre-CPD C3727A (this study) and V. cholerae pre-CPD (PDB code: 3FZY) [12]. The catalytic residues (Cys3727 and His3678) and the scissile residue (Leu3587) were shown for each CPD with stick models. (C) Structural comparison of the N-termini of V. vulnificus pre-CPD C3727A (this study) and the recombinant V. cholerae pre-CPD (PDB code: 3FZY). (D) The surface representation of the N-terminal autocleavage site. The N-terminal cleavage sequence is shown in green stick models. The active site residues, Cys3727Ala and His3678, are colored yellow and cyan, respectively.
3.2. The binding of the N-terminal autoprocessing residues to the pocket
The crystal structure shows how the N-terminal substrate sequence is accommodated in the substrate binding cleft of the InsP6-bound CPD. The electron densities of the N-terminal four residues (3583–3586) upstream of Leu3587 were well visible (Fig 1C). The five residues of the autocleavage sequence at the P4, P3, P2, P1, and P1’ positions (P1 and P1’ refer to the N-terminal residue and C-terminal residue to the scissile bond, respectively), Lys3584 –Ala3588, are positioned between the helix ɑ1 and β9 of the β-flap in the pre-CPD. The scissile residue Leu3587 at P1 is inserted into the hydrophobic substrate-binding pocket (Fig 2D). The hydrophobic pocket is composed of five Val residues and one Leu residue. Val3764 and Leu3762 of β9 in the β-flap compose one side of the pocket wall. The backbone of Leu3762 in β9 makes a hydrogen bond with the amide nitrogen of the scissile residue Leu3587. The MARTX CPDs show a preference for small amino acids such as Ala or Ser in the P2 and P1’ positions to avoid steric clashes with Glu3761 of the β-flap (Fig 1E). The conserved residues at P4 and P3 positions also contribute to the substrate recognition in the pocket. Lys3584 and Glu3585 interact with the β-flap and the helix α1, respectively. The side chain stalk of Lys3584 at P4 makes hydrophobic interaction with Trp3779 of the β-flap. The side chain of Glu3585 at P3 interacts with the side chains of Lys3633 and Val3630 of α1.
The scissile bond is inserted between the catalytic dyad residues, His3678 and Ala-substituted Cys3727. The conformations of the catalytic and scissile-bond residues are almost identical to the structure observed for the V. cholerae MARTX CPD-InsP6 complex [5, 12]. The configuration of catalytic residues and the scissile bond suggests that the catalytic Cys is activated by the close alignment of the scissile bond followed by protonation of the leaving group by the catalytic His [17]. In conclusion, in the InsP6-bound form, the β-flap has a conformation that locks the N-terminus of the CPD in the active site by the interaction between the β-flap and the N-terminus, subsequently allowing hydrolysis of the Leu-Ala scissile bond.
3.3. The interaction of InsP6 with the β-flap activates the MARTX CPD for autoprocessing
InsP6 binds to the conserved basic pocket located distal to the substrate binding pocket. The three β-strands (β4, β7, and β8) of the core β-sheet form the floor of the binding pocket (Fig 3A). The β-flap and the N-terminal loop (residue 3612–3616) compose the wall of the binding pocket. InsP6 makes extensive electrostatic interactions with the binding pocket involving 331 Å2 of interacting surface area (Fig 3B). Totally 10 positively charged residues coordinate the phosphate groups of InsP6 (Fig 3A). InsP6-binding stabilizes the β-flap conformation required for the processing of the scissile bond. Three positively charged residues, Arg3769, Lys3770, and Lys3782 of the β-flap make ionic interactions with InsP6. The InsP6-binding to the basic pocket allosterically rearranges the conformation of active site residues by forming hydrogen-bonding networks of the β-flap with the active site and the substrate (Fig 3C). The β9β10 hairpin directly contacts both InsP6 and P1 Leu3587. The β9-strand of the β-flap forms one side of the pocket wall that accommodates Leu3587. The basic residues of the β10-strand interact with InsP6. The β8-β9 loop of the β-flap interacts with the β7-α3 loop locking the catalytic residue Cys3727 in an active conformation.
(A) The InsP6-binding to the pre-CPD C3727A. The electrostatic interactions between InsP6 and the basic residues of the CPD are shown in dotted lines. The N-terminal cleavage sequence is colored in pale blue. The scissile residue Leu3587 is shown by dot representation. (B) Electrostatic surface representation of the basic InsP6-binding site of pre-CPD C3727A. The active conformation of the β-flap induced by InsP6-binding. The dotted lines indicate the hydrogen-bond networks of β-flap with the substrate, β7- α3 loop, and InsP6. (C) Measurement of binding affinities of InsP6 to the pre-CPD C3727A (3570–3796) and CPD ΔNt wt (3592–3796) by ITC. The CPDs of 0.1 mM were titrated with 1 mM of InsP6. The removal of the β-flap abolishes InsP6-binding and autocleavage activities.
The binding affinity of InsP6 to the pre-CPD C3727A was 1.4 ± 0.5 μM measured by isothermal titration calorimetry (Fig 3D). The truncated CPD (residue 3592–3796) representing the post-processed form showed a reduced affinity to InsP6 with a Kd of 21.8 ± 5.0 μM. The data suggest that the pre-CPD containing the N-terminal cleavage sequence is configured for high affinity binding to InsP6. This is consistent with the reported InsP6-binding properties of the V. cholerae MARTX CPD. The pre-CPD of V. cholerae MARTX has an affinity to InsP6 with a Kd value of 0.6 μM [17]. However, the processed CPD of V. cholerae MARTX had a 500-fold reduced affinity to InsP6 and was reactivated for high affinity binding to InsP6 by cooperative binding of a new substrate [12, 13].
The mutation of the positively charged residues in the InsP6-binding pocket, K3787E or H3642E completely abolished InsP6 binding and autoprocessing activity (Fig 3D). The truncated construct lacking the β-flap (Δβ9-β12) was soluble, indicating that the β-flap is not essential for protein stability. However, the Δβ-flap construct completely lost InsP6-binding, indicating that the β-flap is essential for the InsP6-binding and catalytic activity (Fig 3D).
3.4. The CPD Δβ-flap displays an inactive conformation of the catalytic site
After the autoprocessing of the scissile bond, the CPD is reported to adopt a conformation with reduced affinity to InsP6 [12]. To examine the structure of the processed CPD, we purified the wild-type CPD lacking the N-terminal autocleavage sequence (post-CPD, residue 3592–3796). The post-CPD construct in the absence of InsP6 was susceptible to proteolytic degradation of the C-terminal β-flap during protein expression and purification (Fig 4A). In contrast, the recombinant construct of the pre-CPD C3727A did not show significant proteolysis of the β-flap region. This observation indicates that the C-terminal β-flap region is flexible in an apo-form of the post-CPD due to the repulsion of positive residues in the InsP6-binding pocket. This is consistent with the previous report that the β-flap is resistant to limited proteolysis in the presence of InsP6 and the InsP6-binding induces stabilization of the β-flap in the activated CPD [6, 12].
(A) SDS-PAGE analysis of the crystallized CPD constructs. Lane 1, molecular weight marker; Lane 2, the purified CPD (residue 3592–3796) incubated with InsP6 was used for crystallization; Lane 3, the purified CPD protein (residue 3592–3796); Lane 4, purified pre-CPD C3727A (3578–3796). The purified proteins were incubated at room temperature for 48 h before SDS-PAGE analysis. The lane 1 and the other lanes were cut from the identical gel. (B) The 2Fo-Fc electron density maps of the InsP6 binding site in the post-CPD structure. (C) The structure of a dimeric CPD lacking the β-flap. InsP6 binds to the CPD in a 1:2 ratio. One of the protomers shown in gray is related by a noncrystallographic 2-fold axis. The disordered N-terminal loop (3611–3613) is shown in dotted lines. (D) Structural comparison of the InsP6-binding modes in the post-CPD Δβ-flap dimer. The identical InsP6 molecules bound to each promotor are overlayed to show the orientation of InsP6 to each protomer. The ionic interactions of InsP6 with the four basic residues were indicated by yellow dotted lines. (E) Structural comparison of the InsP6-binding to pre-CPD C3727A and the post-CPD wt Δβ-flap. (F) Structural comparison of pre-CPD C3727A and the proteolyzed CPD Δβ-flap. The equivalent catalytic residues, Cys3727, are indicated by red arrows. (G) Structure of the post-CPD Δβ-flap with the inactive conformation. The catalytic dyad residues, His3678 and Cys3727, are shown in dotted stick models.
Approximately 30 C-terminal residues were partially proteolyzed during protein purification, resulting in heterogeneity of the crystallized protein. The purified post-CPD construct was incubated with the final 5 mM of InsP6 for an hour before crystallization setup. InsP6-binding to the post-CPD was expected to stabilize the protein by forming electrostatic interactions between the β-flap and the core of the CPD domain. InsP6 displayed a binding affinity of 21.8 μM to post-CPD which is 16 times weaker than the affinity to pre-CPD. We obtained the crystals in a week at room temperature and determined the structure of the post-CPD at 2.2 Å resolution. The electron density maps were well-defined for the core of two CPDs and an InsP6 molecule (Fig 4B). However, the C-terminal β-flaps (residues 3762–2796) were not visible. In the asymmetric unit of crystals, there were two CPD molecules coordinating a single molecule of InsP6 in the center of the non-crystallographic two-fold axis (Fig 4C). The molecular packing in the crystal lattice does not allow the space for the folded β-flaps in the CPD dimer. Considering the proteolytic susceptibility of the β-flap, it seems that the crystals were formed by proteolyzed fragments lacking the β-flap (post-CPD Δβ-flap) or that the β-flaps were completely disordered. Two protomers have almost identical conformations except for the N-terminal disordered loops (residue 3611–3613). The dimer interface with 810 Å2 of a buried surface area is formed by the antiparallel interaction of β8 strands and α3 helices from each protomer. A single molecule of InsP6 in the crevice of the dimer interface interacts with the four basic residues of the InsP6-binding pockets from each protomer (Fig 4D). InsP6 does not bind tightly to the binding pocket due to the absence of the β-flap. Compared to the structure of the pre-CPD-InsP6 complex, the position of InsP6 in the post-CPD Δβ-flap was shifted 3 Å away from the binding pocket to share the coordination of a single InsP6 molecule by two protomers (Fig 4E). The recombinant construct (residue 3590–3762) of the CPD Δβ-flap corresponding to the crystallized fragment was monomer in SEC and had no InsP6-binding property, suggesting that the observed CPD dimer–InsP6 complex is a crystallographic artifact that was formed during the crystallization process.
Though the dimeric structure of the post-CPD Δβ-flap is physiologically not relevant, the monomeric structure seems to represent the inactive conformation of the core CPD when InsP6 is not loaded to the post-CPD. Compared to the pre-CPD-InsP6 complex, the post-CPD Δβ-flap shows large conformational differences in the α2-β7 and β7-α3 loops (Fig 4F). Due to the lack of interaction between the β-flap and the β7-α3 loop, the catalytic residue Cys3727 is displaced 5.4 Å toward the helix α1, occluding the substrate binding pocket (Figs 1E and 4G). However, the catalytic residue His3678 in the β4-β5 loop has no conformational difference. InsP6-binding contributes to enzyme activation by properly ordering the pocket and active site residues via interaction with the β-flap. Since the flexible β-flap directly contacts both InsP6 and the substrate, cooperative binding of InsP6 and a substrate to the post-CPD is required for catalytic activation.
In conclusion, InsP6-binding to the pre-CPD activates the cleavage of the N-terminal scissile bond, generating the post-CPD form. The post-CPD has a reduced affinity to InsP6 compared to the pre-CPD. When not loaded with InsP6, the post-CPD has an inactive conformation of catalytic residues and flexibility in the β-flap region due to the lack of interaction with a substrate and InsP6.
4. Discussion
Vibrio vulnificus can cause rapid and life-threatening sepsis and wound infections in humans. The virulence of V. vulnificus is governed by many factors including its acid resistance, capsular polysaccharide production, lipopolysaccharide, iron acquisition, cytotoxic factors, and expression of motility and adherence/adhesion molecules [18]. The MARTX toxin of V. vulnificus as a key virulence factor displays cytotoxicity and hemolytic activity [8, 9]. The inactivation of the rtxA1 gene encoding MARTX significantly attenuates the virulence of V. vulnificus [9, 19]. Since the activation of the CPD is essential for the production of active MARTX effectors, the CPD was considered as a potential drug target for inhibition of MARTX activation. Small-molecule inhibitors against the CPD were developed to block the activation of V. cholerae MARTX toxin [12, 20]. However, the inhibition of CPD did not fully abolish the RTX activity and was not effective in blocking the action of the toxin in cells. Since the N terminus is bound within the active site prior to InsP6 binding in the pre-processed CPD, it occludes access of the catalytic Cys to protease inhibitors [12]. Upon autoprocessing of CPD, the exposure of the catalytic Cys to the CPD inhibitors facilitates the inhibition of subsequent processing of effector release. Alternatively, the activation of the CPD using InsP6 mimetics was considered a druggable approach by triggering the toxin auto-proteolysis prior to cell uptake and disrupting the translocation and activation process [6, 21]. Therefore, understanding the precise activation mechanism of the CPD might facilitate the development of new anti-toxin drugs targeting MARTX CPDs.
This study provides a structural basis for how InsP6-binding is communicated to the active site for activation of the V. vulnificus MARTX CPD by structure determination of the CPD–InsP6. The rearrangement of the β-flap by InsP6-binding locks the N terminus of the CPD in the active site and facilitates the hydrolysis of the Leu3587–Ala3588 amide bond. After autoprocessing, the CPD adopts a postprocessing form that has a reduced affinity for InsP6 due to the loss of interaction between the β-flap and the N-terminal substrate residues. The structure of the post-CPD Δβ-flap suggests that the substrate-binding pocket of the InsP6-free form is occluded due to the lack of interaction between the β-flap and the β7-α3 loop. The inactive structure of the post-CPD Δβ-flap correlates with the previous proposal that InsP6-binding regulates exposure of the active site [5, 6]. After the autoprocessing of the first scissile bond, the cooperative binding of both InsP6 and a new substrate to the post-CPD is required for catalytic activation by inducing a conformational change of the active site [12]. Then, the reactivated CPD cleaves the other sites between the effector domains of the MARTX toxin.
Supporting information
S1 Raw image. The original uncropped gel images of Figs 1B and 4B.
https://doi.org/10.1371/journal.pone.0307512.s001
(PDF)
References
- 1. Baker-Austin C, Oliver JD, Alam M, Ali A, Waldor MK, Qadri F, et al. Vibrio spp. infections. Nat Rev Dis Primers. 2018;4(1):8. Epub 2018/07/14. pmid:30002421.
- 2. Baker-Austin C, Oliver JD. Vibrio vulnificus: new insights into a deadly opportunistic pathogen. Environ Microbiol. 2018;20(2):423–30. Epub 2017/10/14. pmid:29027375.
- 3. Boardman BK, Satchell KJ. Vibrio cholerae strains with mutations in an atypical type I secretion system accumulate RTX toxin intracellularly. J Bacteriol. 2004;186(23):8137–43. Epub 2004/11/18. pmid:15547287.
- 4. Kim BS. The Modes of Action of MARTX Toxin Effector Domains. Toxins (Basel). 2018;10(12). Epub 2018/12/06. pmid:30513802.
- 5. Lupardus PJ, Shen A, Bogyo M, Garcia KC. Small molecule-induced allosteric activation of the Vibrio cholerae RTX cysteine protease domain. Science. 2008;322(5899):265–8. Epub 2008/10/11. pmid:18845756.
- 6. Egerer M, Satchell KJ. Inositol hexakisphosphate-induced autoprocessing of large bacterial protein toxins. PLoS Pathog. 2010;6(7):e1000942. Epub 20100708. pmid:20628577.
- 7. Chung HY, Bian Y, Lim KM, Kim BS, Choi SH. MARTX toxin of Vibrio vulnificus induces RBC phosphatidylserine exposure that can contribute to thrombosis. Nat Commun. 2022;13(1):4846. Epub 2022/08/18. pmid:35978022.
- 8. Kim YR, Lee SE, Kook H, Yeom JA, Na HS, Kim SY, et al. Vibrio vulnificus RTX toxin kills host cells only after contact of the bacteria with host cells. Cell Microbiol. 2008;10(4):848–62. Epub 2007/11/17. pmid:18005241.
- 9. Lee JH, Kim MW, Kim BS, Kim SM, Lee BC, Kim TS, et al. Identification and characterization of the Vibrio vulnificus rtxA essential for cytotoxicity in vitro and virulence in mice. J Microbiol. 2007;45(2):146–52. Epub 2007/05/08. pmid:17483800.
- 10. Irvine RF, Schell MJ. Back in the water: the return of the inositol phosphates. Nat Rev Mol Cell Biol. 2001;2(5):327–38. Epub 2001/05/02. pmid:11331907.
- 11. Shen A. Autoproteolytic activation of bacterial toxins. Toxins (Basel). 2010;2(5):963–77. Epub 2010/05/01. pmid:22069620.
- 12. Prochazkova K, Shuvalova LA, Minasov G, Voburka Z, Anderson WF, Satchell KJ. Structural and molecular mechanism for autoprocessing of MARTX toxin of Vibrio cholerae at multiple sites. J Biol Chem. 2009;284(39):26557–68. Epub 2009/07/22. pmid:19620709.
- 13. Choi S, Kim BS, Hwang J, Kim MH. Reduced virulence of the MARTX toxin increases the persistence of outbreak-associated Vibrio vulnificus in host reservoirs. J Biol Chem. 2021;296:100777. Epub 2021/05/17. pmid:33992647.
- 14. Otwinowski Z, Minor W. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 1997;276:307–26. pmid:27754618.
- 15. McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC, Read RJ. Phaser crystallographic software. J Appl Crystallogr. 2007;40(Pt 4):658–74. Epub 20070713. pmid:19461840.
- 16. Liebschner D, Afonine PV, Baker ML, Bunkoczi G, Chen VB, Croll TI, et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr D Struct Biol. 2019;75(Pt 10):861–77. Epub 20191002. pmid:31588918.
- 17. Prochazkova K, Satchell KJ. Structure-function analysis of inositol hexakisphosphate-induced autoprocessing of the Vibrio cholerae multifunctional autoprocessing RTX toxin. J Biol Chem. 2008;283(35):23656–64. Epub 2008/07/02. pmid:18591243.
- 18. Jones MK, Oliver JD. Vibrio vulnificus: disease and pathogenesis. Infect Immun. 2009;77(5):1723–33. Epub 2009/03/04. pmid:19255188.
- 19. Liu M, Alice AF, Naka H, Crosa JH. The HlyU protein is a positive regulator of rtxA1, a gene responsible for cytotoxicity and virulence in the human pathogen Vibrio vulnificus. Infect Immun. 2007;75(7):3282–9. Epub 2007/04/18. pmid:17438022.
- 20. Shen A, Lupardus PJ, Albrow VE, Guzzetta A, Powers JC, Garcia KC, et al. Mechanistic and structural insights into the proteolytic activation of Vibrio cholerae MARTX toxin. Nat Chem Biol. 2009;5(7):469–78. Epub 2009/05/26. pmid:19465933.
- 21. Ivarsson ME, Durantie E, Huberli C, Huwiler S, Hegde C, Friedman J, et al. Small-Molecule Allosteric Triggers of Clostridium difficile Toxin B Auto-proteolysis as a Therapeutic Strategy. Cell Chem Biol. 2019;26(1):17–26.e13. Epub 20181025. pmid:30482680.