2.1. Bacterial strains

The P. plecoglossicida wild-type (WT) strain PQLYC4 was isolated from diseased Larimichthys crocea with visceral white nodules disease [21], and cultured in trypticase soy broth (TSB, Hopebio, China) or trypticase soy agar (TSA) at 28°C. E. coli strains DH5α, BL21DE3 (Invitrogen), and S17-1 (AngYuBio) were cultured in Luria-Bertani broth (LB, Hopebio, China) or on LB agar at 37°C.

2.2. Construction of ΔPpClpP1 and ΔPpClpP2 P. plecoglossicida

The ΔPpClpP1 and ΔPpClpP2 mutants were generated through homologous recombination using a suicide vector strategy [22]. The 500 bp upstream (N-terminal) and downstream flanking regions of the PpClpP1 and PpClpP2 open reading frame were amplified using primer sets ΔPpClpP1-U F/R, ΔPpClpP2-U F/R, ΔPpClpP1-D F/R, and ΔPpClpP2-D F/R, respectively. These fragments were subsequently fused by overlap extension PCR using external primers ΔPpClpP1-U F, ΔPpClpP2-U F, ΔPpClpP1-D R, and ΔPpClpP2-D R. The resulting fusion product was cloned into the suicide vector pEX18Tc at KpnI/HindIII restriction sites, generating the recombinant plasmid pEX18Tc-PpClpP1 or pEX18Tc-PpClpP2. After transformation into E. coli S17-1, positive clones were selected on LB agar supplemented with 10 μg/ml tetracycline. Conjugation was performed to transfer pEX18Tc-PpClpP1 or pEX18Tc-PpClpP2 from E. coli S17-1 to wild-type P. plecoglossicida. Primary recombinants were selected on Tryptic Soy Agar (TSA) plates containing both tetracycline (10 μg/ml) and ampicillin (100 μg/ml), taking advantage of the native ampicillin resistance of wild-type P. plecoglossicida. Secondary recombinants were then isolated on TSA plates containing 12% (w/v) sucrose to select for vector loss. Mutant validation was performed using two PCR strategies: (1) amplification across the deletion junction using primers ΔPpClpP1-U F/ΔPpClpP1-D R or ΔPpClpP2-U F/ΔPpClpP2-D R, and (2) internal verification using primer set PpClpP1-F/R or PpClpP2-F/R. All PCR products were confirmed by Sanger sequencing. Primer sequences were designed based on the complete genome sequence of P. plecoglossicida strain PQLYC4 (NCBI accession PRJNA612395), with all oligonucleotides listed in S1 Table.

2.3. Bacterial growth curve and plate titration assays

To evaluate the effect of bortezomib (BTZ) on bacterial growth, P. plecoglossicida strains (PQLYC4-WT and knockout PQLYC4-ΔPpClpP1/ΔPpClpP2) were cultured to an initial optical density at 600 nm (OD₆₀₀) of 0.1, followed by treatment with BTZ at two-fold serial dilution concentrations. Growth was monitored spectrophotometrically (OD₆₀₀) hourly for 15 h to generate growth curves. For plate titration assays, bacterial cultures were adjusted to OD₆₀₀ ≈ 0.4 with double distilled water, and 10-fold serial dilution was performed on TSB containing 1.25 μM BTZ, then plated on TSA plates to determine the number of surviving colony forming bacteria.

2.4. Bacterial motility assays

Wild-type, ΔPpClpP1 and ΔPpClpP2 knockout strains were revived from glycerol stocks stored at –80°C by streaking onto TSA plates. The plates were incubated at 28°C for 24–48 h. A single colony was then inoculated into TSB medium and cultured with shaking at 200 rpm and 28°C until the OD₆₀₀ reached approximately 0.6. Swimming and swarming motilities were assessed using freshly prepared semi-solid agar plates. For swimming motility, 0.3% agar was used, while 0.6% agar was employed for swarming motility. Bacterial cultures were spot-inoculated at the center of the plates. The inoculated plates were incubated upright at 28°C. Phenotypes were observed and photographed after 24 h for swimming motility and 48 h for swarming motility.

2.5. Plasmid construction and protein expression

PpClpP1 (1–183, AXM94812) and PpClpP2 (1–206, WP_003259400.1) and PpClpX (QLB56231.1) DNA were directly amplified with a primer set (S1 Table) from P. plecoglossicida PQLYC4 strain cDNA using RT-PCR, and cloned into pET28a (N-terminal with 6 × His) expression vector using BamHI and XhoI restriction enzymes. Full-length sequences PpClpP1 and PpClpP2 were constructed on the pET-Duet vector (for coexpression of PpClpP1-His₆ and PpClpP2-StrepII). Site-directed mutagenesis of PpClpP1 and PpClpP2 were performed using the DpnI method. The active site mutants of PpClpP1 included S73A, H96A, P151A, and P151D; the BTZ-binding site mutants of PpClpP1 included G44A, E45A, C46A, S47A, F98A, H99A, W100A, T101A, S117A, and D121A. The active site mutants of PpClpP2 included S114A, H139A, D188A, and D188P. All the expression plasmids were expressed in E. coli BL21 cells following overnight expression at 20°C after induction with 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG). Cells were pelleted by centrifugation at 8000 g for 30 min and stored at -20°C until purification.

2.6. Protein purification

The cell pellet was thawed and resuspended in lysis buffer (50 mM Tris pH 7.5, 150 mM NaCl, and 2 mM β-mercaptoethanol with a PMSF (Phenylmethylsulfonyl fluoride). Cells were lysed mechanically by three passes through a high-pressure instrument. Cell lysate was clarified via centrifugation at 17,000 g for 30 min and the supernatant filtered using a 0.45 μm filter (EMD Millipore). Protein were purified using native NiNTA or Strep affinity chromatography followed by Size-Exclusion Chromatography (SEC) using a Superdex 200 Increase columns (Cytiva Bio-technology) and a final SEC buffer containing 50 mM Tris pH 7.5, 150 mM NaCl, and 5% glycerol.

2.7. Real-time quantitative PCR

Culture P. plecoglossicida PQLYC4 in TSB medium and collect cells at different growth stages with OD values (0.1, 0.8, 1.5, and 2.0). According to the manufacturer’s instructions, the cultured cells were then sampled for total RNA extraction using the Eastep super total RNA extraction kit (Promega). The cDNA was synthesized using GoScript reverse transcription mix (Promega). The real-time PCR analysis was carried out using TB Green mix (Takara) and specific primers (S1 Table) in a QuantStudio 5 real-time PCR system (Thermo Fisher Scientific). Gene expression levels were normalized against the reference gene PpGyrB using the 2^-ΔΔCt method. All data were obtained from three independent experiments, and each analysis was performed in triplicate.

2.8. Sequence alignment and phylogenetic tree analysis

Protein sequences were aligned using the CLUSTALW-Multiple Sequence Alignment website (https://www.genome.jp/tools-bin/clustalw), as well as the ESPript 3.0 server Aligned Sequences website (https://espript.ibcp.fr/ESPript/ESPript/). Amino acids have been colored with similarity coloring scheme % MultAlin, Global score 0.7. Phylogenetic trees were constructed using MEGA 11 software for data processing and visualization.

2.9. Size-exclusion chromatography

Size-exclusion chromatography were performed with a HiLoad 16/60 Superdex 200 column (Cytiva Bio-technology) on an ÄKTApurifier System with UV detector (UPC 900, P900, Box900, Frac950, Cytiva Bio-technology). We injected 2 mg of protein via a 1 mL sample loop with the respective buffer at a flow rate of 0.5 mL/min.

2.10. Analytical ultracentrifugation

Sedimentation velocity experiments were conducted using a Beckman Proteomlab XL-I analytical ultracentrifuge equipped with scanning UV/visible optics. An-60 Ti four-hole rotor and cells with 12 mm charcoal centerpieces and quartz windows were used. The assay volume was 400 μL with a protein concentration set to 0.8 mg/mL and samples were analyzed at 30,000 rpm and 16°C scanning continuously until complete sedimentation was achieved. The data were then analyzed using a continuous c(s) distribution and SEDFIT version 13.0b software [23].

2.11. Peptidase activity

All assays of peptidase activity were performed at 37°C for 2 h, in black 96-well plates using a Plate Reader Varioskan LUX (Thermo Scientific). Each well contained 200 μM Suc-LLVY-AMC (MedChemExpress, HY-P1002) fluorogenic peptide, 1 μM recombinant protein in 100 μL of buffer (20 mM HEPES, pH 7.5, 100 mM KCl, 5% glycerin). DMSO as the experimental control group, and DMSO concentration never exceeded 2%. The reaction was initiated by the addition of the enzyme, and peptidase activity was followed in the linear range by monitoring the rate of production of fluorescent 7-amino-4-methylcoumarin-amc from peptide-AMC substrates at 460 nm (excitation at 380 nm) [24]. The deviation of fluorescence value in three independent measurements was not ＞ 5%.

2.12. Cryo-EM sample preparation and data collection

Frozen-hydrated specimens were prepared using a Vitrobot Mark IV plunger (Thermo Fisher Scientific). PpClpP1 protein (as shown in S1C Fig) was loaded at the concentration of 0.5 mg/mL onto a freshly glow-discharged (30 s at 15 mA) holey carbon grid (Quantifoil Cu R1.2/1.3). The excess solution from the grid was blotted for 4 s at 100% humidity and 4°C before the grid was plunged into liquid ethane. For structure determination, the frozen grids were loaded into a 200 kV Glacios electron microscope at School of life sciences, Anhui university for automated image acquisition with EPU [25]. Movies were recorded on a Falcon 4 camera equipped with Slectris energy filter (±5 eV) at 165K nominal magnification (calibrated pixel size of 0.698 Å) and defocus values ranging from -0.4 to -2.6 μm. During data collection, the total dose was 50 e⁻/A². The detailed collection statistics are shown in S2 Table.

2.13. Structure analysis and model refinement

Cryo-EM analysis was performed using CryoSPARC [26]. All frames in each collected movie were aligned and summed using Patch Motion Correction, and CTF estimation were made using Patch CTF Estimation. Blob Picker and Template Picker were used for particle picking, and particles were extracted using a box size of 300 * 300 pixels [27]. 2D classifications and 3D classifications were used to remove junk particles and select the most homogeneous particles for in-depth 3D structural analyses. The final 3D reconstruction for each class was done using Non-Uniform Refinement, and the resulting map was post-processed using DeepEMhancer.

The reported resolution is based on the “gold standard” refinement procedure and the 0.143 Fourier Shell Correlation (FSC) criterion [28]. Local resolution was estimated using Local Resolution Estimation. For model building, the AlphaFold-predicted tetradecameric structure of PpClpP1 was used as initial models to fit into the maps using Chimera, and the resulting models were manually adjusted and rebuilt according to the cryo-EM map in COOT [29]. Phenix real-space refinement was used to refine the models [30]. The refinement statistics are shown in S2 Table. The detailed classifications and map qualities of PpClpP1 are shown in the S2 Fig.

2.14. Nano differential scanning fluorimetry

Protein stability was determined using the nano differential scanning fluorimetry (nanoDSF) method based on intrinsic tryptophan or tyrosine fluorescence [31]. All nanoDSF assays were performed using the Prometheus NT.48 instrument (NanoTemper Technologies). PpClpP1 protein samples (10 μl at 1 mg/mL) and PpClpP1-bortezomib (1 mM) complex were loaded into standard-grade nanoDSF capillaries, placed on the prometheus capillary holder, and subjected to a temperature ramping of 1°C/min from 20 to 94.8°C. The melting point (Tm) onset (°C) and Tm (°C) values, which indicate the structural stability of the samples, were obtained by monitoring the intrinsic tryptophan and tyrosine fluorescence at the emission wavelengths of 330 and 350 nm. The ratio of the fluorescence intensities (F350 nm/F330 nm) was plotted versus temperature or time to generate an unfolding curve. The thermal stability of a sample was described by the thermal unfolding transition midpoint Tm (°C), at which half of the protein population is unfolded. The Tm value corresponded to the inflection point of the unfolding curve and was determined via the derivative of the curve.

2.15. Isothermal titration calorimetry

The affinity experiments were performed in a Nano ITC (low volume) (TA Instruments) [32]. Protein samples were prepared as above. All samples for the assays were prepared in the buffer containing 50 mM Tris (pH 7.5) and 150 mM NaCl. To measure the affinity between bortezomib and PpClpP1 or PpClpP2, 300 μM bortezomib in the syringe was titrated into a sample cell containing 30 μM recombinant protein PpClpP1 or PpClpP2. All experiments were carried out at 25°C. Data correction and analyses were performed in NanoAnalyze software (TA Instruments).

2.16. Molecular simulation of interaction between bortezomib and PpClpP1

Molecular docking was performed to analyze the mechanism of the binding that occurs between bortezomib and PpClpP1 monomer by Autodock Vina (Scripps Research Institute, La Jolla, CA, USA) [33,34]. Energy minimization of the protein and ligand were performed by Molecular Operating Environment (MOE) software. It shows ligand-binding flexibility with the binding pocket residues. The lowest energy conformations were used for analysis and a picture was generated by Pymol software.

2.17. Pull-down analysis of PpClpX with PpClpP proteases

Protein interactions between the unfoldase PpClpX and different PpClpP proteases were analyzed by pull-down assays. His-tagged PpClpX (65–414 aa) and the PpClpP1P2 complex were expressed as described in Materials and methods section 2.5. To obtain Strep-tagged PpClpP1 and PpClpP2, the corresponding genes were cloned into pET-28a (with the His-tag removed and an N-terminal Strep-tag inserted). For His pull-down assays, protein complexes were incubated with Ni-NTA agarose purification resin (Cytiva). Resins were washed with buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.2% Triton X-100, and bound proteins were eluted with elution buffer (300 mM imidazole, 50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.2% Triton X-100). For Strep pull-down assays, protein complexes were incubated with Strep agarose purification resin (Yeasen Biotechnology Co., Ltd.). After washing with the same buffer as above, proteins were eluted with buffer containing 20 mM D-biotin, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.2% Triton X-100. All samples were separated by SDS-PAGE and visualized by Coomassie blue staining.

2.18. Degradation assay of GFP-ssrA and FITC-labeled casein

Degradation assays were performed using GFP-ssrA or FITC-labeled casein as substrates. Each 50 μL reaction contained 1 μM of the specified protein components (BSA, PpClpP1, PpClpP2, PpClpP1P2, PpClpX, PpClpX + PpClpP1, PpClpX + PpClpP2, or PpClpX + PpClpP1P2 complex) were mixed with 50 μL of 1 μM GFP-ssrA or FITC-casein substrate. Reactions were carried out in a buffer containing 25 mM HEPES-KOH (pH 7.5), 150 mM KCl, 20 mM MgCl₂, and 2 mM ATP, and incubated at 37°C for 2 h. Fluorescence was measured using a Molecular Devices FlexStation 3 microplate reader, with an excitation wavelength of 470 nm and an emission wavelength of 510 nm for GFP-ssrA, an excitation wavelength of 498 nm and an emission wavelength of 517 nm for FITC-casein. The GFP-ssrA fusion protein was constructed by fusing the E. coli ssrA sequence to the GFP sequence, cloned into a pET-28a vector, and purified using an E. coli prokaryotic expression system. FITC-labeled casein (C275894) was commercially obtained from Aladdin.

2.19. Statistical analysis

Statistical analysis was performed using the IBM SPSS Statistics 2.0 software platform with one-way ANOVA. A p value <0.05 was considered statistically significant.

3.1. Characterization and enzymatic activity of ClpP proteases from P. plecoglossicida

Many bacterial clpP1 and clpP2 genes are located at distinct genomic loci and the paralogs appear to serve varying functions in vivo [13,16,35]. Similarly, P. plecoglossicida PQLYC4 strain encodes two distinct ClpP protease proteolytic subunits, designated as PpClpP1 and PpClpP2 (Fig 1A). Phylogenetic analysis revealed that PpClpP2 clusters closely with canonical ClpP homologs, showing particularly high similarity to P. aeruginosa ClpP1 (Fig 1B). In contrast, PpClpP1 forms an evolutionarily divergent clade, suggesting either functional specialization or acquisition of unique regulatory mechanisms during evolution. Sequence alignment demonstrated that PpClpP1 contains an N-terminal truncation (>15 amino acid truncation) compared to typical ClpP proteases (S1A Fig). Most notably, PpClpP1 possesses a non-canonical catalytic triad configuration (Ser-His-Pro), differing from both the conserved Ser-His-Asp triad (Fig 1C) or the rare Ser-His-Asn triad (S1A Fig) variant found in some bacterial species. Transcript analysis showed significantly higher basal expression levels of PpClpP2 compared to PpClpP1 (S1B Fig), indicating potential differential roles in cellular metabolism.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Characterization and enzymatic activity of ClpP proteases from P. plecoglossicida.

(A) Schematic representation of a P. plecoglossicida gene cluster encoding Clp protein family components, numbers correspond to genomic loci of the proteins. (B) Phylogenetic tree of ClpP proteins constructed using neighbor-joining method based on multiple sequence alignment (S1A Fig). Bootstrap values (>50%) indicate branch support, with species and protein accessions labeled accordingly. (C) Representative multiple sequence alignment of selected ClpP homologs highlighting catalytic residues: Serine-S and Histidine-H marked with black asterisks, while Aspartic-D and Proline-P are indicated by red asterisks. Size-exclusion chromatography (SEC) elution profiles of recombinant PpClpP1 (D) and PpClpP2 (E). (F) Assessment of peptidase activity in P. plecoglossicida ClpP. (a) Comparative analysis of peptidase activity in PpClpP1 versus PpClpP2. (b) Peptidase activity of PpClpP1 catalytic triad mutants. (c) Peptidase activity of PpClpP2 catalytic triad mutants. Values indicated at the base of each bar denote the fold-change in fluorescence intensity relative to the DMSO-treated control. Bars represent the means of three independent experiments ± SD. All experiments were performed in triplicate. *p < 0.05, **p < 0.01.

https://doi.org/10.1371/journal.ppat.1013909.g001

To investigate the functional consequences of these structural variations, we performed comparative biochemical analyses. We conducted in vitro analytical ultracentrifugation (AUC) assays using recombinant proteins expressed and purified in E. coli (Fig 1D and 1E). AUC confirmed that both proteases form canonical tetradecamers (S1C and S1D Fig). Enzymatic characterization using the fluorogenic substrate Suc-LLVY-AMC revealed that PpClpP2 exhibits substantially higher proteolytic activity, demonstrating a 13-fold greater catalytic efficiency compared to PpClpP1 (Fig 1F-a). To investigate whether the unique proline residue in the catalytic triad contributes to the low peptidase activity of PpClpP1, we generated amino acid substitutions of the catalytic triad residues in both PpClpP1 and PpClpP2 and assessed their enzymatic activity. Substitution of the catalytic serine-S or histidine-H with alanine-A completely abolished peptidase activity in both proteases, confirming their essential roles. In contrast, replacing the proline-P in PpClpP1 with either alanine-A or aspartate-D only partially reduced activity, retaining approximately 60% of wild-type levels (Fig 1F-b). Similarly, mutating the aspartate-D in PpClpP2 to alanine-A or proline-P did not eliminate activity, with residual levels of approximately 16% and 13%, respectively (Fig 1F-c). These results establish that while the catalytic triad is essential for proteolytic function in both proteases, the serine and histidine residues are absolutely indispensable, whereas the third position (proline in PpClpP1, aspartate in PpClpP2) modulates but is not strictly required for activity.

Our findings reveal fundamental functional differences between the two ClpP paralogs in P. plecoglossicida. Whereas PpClpP2 maintains the conserved structural and functional characteristics typical of bacterial ClpP proteases, PpClpP1 exhibits multiple atypical features, including phylogenetic divergence, N-terminal truncation, and a novel catalytic triad configuration. These distinctive attributes position PpClpP1 as an intriguing subject for future studies on protease evolution and non-canonical mechanisms of proteolytic regulation.

3.2. Overall structure of P. plecoglossicida ClpP1

To elucidate the structural features of PpClpP1 and PpClpP2, we attempted to determine their three-dimensional structures using both X-ray crystallography and cryo-electron microscopy (Cryo-EM). However, we only succeeded in solving the structure of PpClpP1. We determined its Cryo-EM structure at an average resolution of 3.07 Å using recombinant protein (S2A–S2F Fig). The asymmetric unit consists of two opposing heptameric rings forming a barrel-shaped homotetradecamer with dimensions of 93 Å in height and 99 Å in diameter (Figs 2A and S3A). Each monomer adopts the canonical ClpP fold, featuring five parallel β-strands (β1-β2-β4-β6-β9) flanked by four α-helices (αA-αD) arranged perpendicular to the β-sheet plane (Fig 2B) [36–38]. Notably, the αD helix forms a distinctive handle domain together with β7 and β8. As predicted by sequence alignment (Fig 1C), structural analysis confirmed the presence of an unprecedented catalytic triad (S73-H96-P151) within the proteolytic chamber. This geometric perturbation at the catalytic center likely underlies the observed attenuation of proteolytic activity. Specifically, a Pro in this position is unable to provide electrostatic stabilization to the positive charge of the doubly protonated form of the His, which acts as the proton acceptor of the catalytic Ser during its activation. The tetradecameric interface is stabilized by interlocking handle domains from opposing heptamers (S3B-a Fig), involving: (1) interactions between two disordered domains (S3B-b Fig), and (2) contacts between αD-helices (S3B-c Fig). The heptameric interfaces feature front-to-back interactions where α-helices of one subunit align with β-sheets of another (S3C-a Fig), with each α-helix and corresponding β-sheet forming three distinct interaction regions (S3C-b–S3C-d Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Structural architecture of PpClpP1.

(A) Surface representation of the tetradecameric PpClpP1 complex (PDB ID 9UXT) showing top and side views, with each monomer in the heptameric rings colored distinctly for clarity. (B) Ribbon diagram of an individual subunit highlighting the secondary structure elements, including six α-helices (denoted with α-letters) and ten β-strands (denoted with β-numbers), with the catalytic triad (S-H-P) prominently marked by yellow circles.

https://doi.org/10.1371/journal.ppat.1013909.g002

3.3. Distinguishing features of P. plecoglossicida ClpP1

Comparative structural analysis with ClpP homologs reveals two defining characteristics that distinguish PpClpP1. First, PpClpP1 possesses the shortest polypeptide chain among characterized ClpP proteases (Fig 3A), lacking the N-terminal α-helical domain (Fig 3B). This truncation results in a dramatically expanded axial pore (42.7 Å diameter) formed by the heptameric ring assembly, significantly larger than those reported for other ClpPs (S4A Fig). Since the N-terminal domain normally constrains pore size and mediates interactions with pore-2 loops of ClpX/C/A unfoldases, its absence likely affects substrate processing and unfoldases recognition [39,40]. Second, while maintaining the conserved spatial arrangement of catalytic residues (Fig 3B and 3C), PpClpP1 uniquely substitutes the conventional aspartate/asparagine with proline (S-H-P triad) - a configuration previously undocumented in bacterial ClpP proteases (Fig 3C). Structural alignment shows that Ser73 and His96 maintain positions similar to those in other ClpPs (Fig 3C-m), while Pro151 occupies a distinct spatial orientation compared to catalytic aspartates in P. aeruginosa ClpP1/ClpP2 (S4B Fig). Although this proline substitution represents a moderate structural change, it causes significant divergence in the orientation of the catalytic triad’s charge-relay system. This altered geometry likely explains PpClpP1’s reduced activity, as the S-H-P configuration may maintain the protease in a suboptimal activation state.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Contrasting structural features of PpClpP1 compared to homologs.

(A) Domain architecture comparison showing sequence length and catalytic active sites between PpClpP1 (blue) and other structurally characterized ClpP homologs (gray). (B) a, Structural superposition of PpClpP1 monomer with representative ClpP structures from E. coli (EcClpP1, 3HLN), P. aeruginosa (PaClpP1/2, 7M1M/7M1L), S. aureus (SaClpP, 3V5E), M. Tuberculosis (MtbClpP1/P2, 6VGK), L. monocytogenes (LmClpP1/2, 4JCT/4JCQ), H. pylori (HpClpP2, 2ZL0), B. subtilis (BsClpP2, 7FEQ), and T. thermophilus (TtClpP2, 6HWM). b-c, Dashed boxes highlight the N-terminal α-helical regions (absent in PpClpP1) and catalytic triad sites, with each homolog colored distinctly. (C) Detailed views of catalytic triad configurations in various ClpP homologs (a-l) and their structural superposition (m), demonstrating the unique Ser-His-Pro arrangement in PpClpP1 compared to conventional Ser-His-Asp/Asn triads in other species, (n) Cryo-EM density for the catalytic triad of PpClpP1. The comparative analysis reveals PpClpP1’s distinctive structural adaptations at both N-terminal and active site regions.

https://doi.org/10.1371/journal.ppat.1013909.g003

3.4. Structural comparison of P. plecoglossicida ClpP1 and ClpP2

Complementing our structural characterization of PpClpP1, we employed computational approaches to investigate its paralog PpClpP2. Using AlphaFold prediction [41], we generated a structural model of PpClpP2 based on fragments from homologous PDB structures (S5A Fig), the model showed high confidence (pLDDT > 90, ipTM = 0.87, pTM = 0.88), supporting its reliability for structural analysis. Three-dimensional comparisons revealed significant architectural differences between the two paralogs. The predicted PpClpP2 structure exhibits distinct dimensional characteristics, with increased longitudinal spacing (116 Å) and reduced transverse dimensions (95 Å) compared to PpClpP1 (Fig 4A). Notably, the axial pore diameter of PpClpP2 heptamers (25.2 Å) is substantially narrower than that of PpClpP1 (42.7 Å) (Fig 4B). This constrained geometry likely enhances substrate stabilization and proteolytic efficiency through a more confined channel architecture. Key structural differences extend to functional domains. Unlike PpClpP1, which lacks N-terminal helices and has an atypical catalytic triad (Fig 3B), PpClpP2 maintains a flexible N-terminal architecture featuring: a mini-helix, dual β-strands, and axial loop (Fig 4C-left). PpClpP2 adopts a compact S-H-D configuration resembling the activated state of P. aeruginosa ClpP1 (PaClpP1), this contrasts with PpClpP1’s atypical S-H-P arrangement (Fig 4C-right). The canonical triad architecture of PaClpP1 correlates with high peptidase activity [42], explaining PpClpP2’s significantly greater enzymatic efficiency compared to PpClpP1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Distinct structural features and functional interplay between PpClpP1 and PpClpP2.

(A) Side views of electrostatic potential surfaces for the PpClpP1 tetradecamer (left) and the predicted PpClpP2 tetradecamer (right) generated by AlphaFold. (B) Top views of electrostatic potential surfaces for the PpClpP1 tetradecamer (left) and the predicted PpClpP2 tetradecamer (right). Inset images illustrate the axial pore diameter sizes of PpClpP1 (42.7 Å) and PpClpP2 (25.2 Å), respectively. (C) Structural alignment of PpClpP1 (Cryo-EM structure) with PpClpP2 (predicted model) highlighting domain organization (left) and catalytic triad configurations (right), including superposition with P. aeruginosa ClpP1 (PaClpP1). (D) Purification of the co-expressed PpClpP1-PpClpP2 complex. (a) Schematic of the initial Strep-tag affinity purification. Supernatant: supernatant after E. colicell lysis; Flow-through: flow-through fraction from the Streptactin resin; Wash: washing with buffer (150 mM NaCl, 50 mM Tris-HCl, pH 8.0); Elution: elution with buffer containing 2.5 mM D-desthiobiotin, 150 mM NaCl, 100 mM Tris-HCl, pH 8.0. (b) Subsequent Ni-NTA affinity purification of the eluate from panel a. Flow-through: flow-through fraction from the Ni-NTA resin; Elution: elution with buffer containing 300 mM imidazole, 150 mM NaCl, 50 mM Tris-HCl, pH 8.0. (E) Comparative analysis of the peptidase activities among the PpClpP1P2 heterotetradecamer, the PpClpP1 homotetradecamer, and the PpClpP2 homotetradecamer. (F) The structure model of PpClpP1P2 heterotetradecamer predicted using AlphaFold. (G) Cartoon model depicting the interaction between a single PpClpP1 subunit and a single PpClpP2 subunit. (H) Analysis of the electrostatic potential at the homomeric (PpClpP1-PpClpP1) and heteromeric (PpClpP1-PpClpP2) interaction interfaces.

https://doi.org/10.1371/journal.ppat.1013909.g004

3.5. Assembly and synergistic activation of the PpClpP1P2 heterotetradecameric complex

The potential formation of heterotetradecameric complexes between PpClpP1 and PpClpP2 in vivo represents an important biological question. Our co-purification experiments revealed stable complex formation when both subunits were co-expressed in E. coli (Fig 4D), AUC analysis further indicated that the PpClpP1P2 complex has a molecular mass consistent with a heterotetradecameric assembly (S5B Fig). Notably, the heterotetradecamer exhibited significantly higher peptidase activity than either PpClpP1 or PpClpP2 alone (Fig 4E), suggesting that complex formation may induce conformational changes in PpClpP1 that enhance its proteolytic efficiency.

To elucidate the structural basis of heterocomplex assembly, we predicted the PpClpP1P2 heterotetradecamer structure using AlphaFold (Fig 4F). The model showed high confidence (pLDDT > 90, ipTM = 0.81, pTM = 0.82), supporting its reliability for structural analysis. Structural alignment demonstrated significant overlap between the handle domains of PpClpP2 and PpClpP1 (Fig 4G). The binding interface between PpClpP2 (binding site 3) and PpClpP1 subunit 2 (binding site 2) is predominantly hydrophobic (Fig 4H), resembling the hydrophobic character of the PpClpP1 homomeric interface (binding site 1). This conserved hydrophobic nature likely facilitates stable hetero-oligomerization through similar interaction mechanisms as those maintaining homo-oligomeric assemblies.

3.6. PpClpX recognizes PpClpP2 to drive proteolysis enhanced by PpClpP1 heterocomplex formation

Based on structural analysis, we hypothesized that PpClpP1, unlike PpClpP2, lacks the N-terminal domain necessary for interaction with unfoldases such as ClpX or ClpC. To validate this hypothesis biochemically, we expressed and purified a truncated variant of PpClpX (residues 65–414, lacking the unstructured N-terminal region, with a molecular weight of approximately 43 kDa). AUC analysis confirmed that the truncated PpClpX protein assembles into a Homohexamer (Fig 5A). Subsequently, pull-down assays demonstrated that whereas PpClpP2 interacts with PpClpX, no interaction was detected between PpClpP1 and PpClpX (Fig 5B and 5C), thus confirming our structural prediction. Furthermore, we verified that the heteromeric PpClpP1P2 complex retains the ability to interact with PpClpX (Fig 5D). The interaction is a prerequisite for forming the active Clp proteasome, in which the unfoldase recognizes specific substrates (such as ssrA-tagged proteins) and translocates them into the degradation chamber of ClpPs for hydrolysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Biochemical and functional characterization of PpClpX interactions with homomeric PpClpP1, PpClpP2, and the heteromeric PpClpP1P2 Complex.

(A) Purification and analytical ultracentrifugation (AUC) analysis of the truncated PpClpX (65-414 aa). The sedimentation profile indicates that PpClpX elutes as a stable hexamer. An SDS-PAGE analysis of the purified protein is shown on the right. (B-D) Pull-down assays assessing the interaction between PpClpX and different PpClpP isoforms. (B) His pull-down assay with PpClpX-His, PpClpP1-Strep can not co-elute with PpClpX-His. (C) Strep pull-down assay with PpClpP2-Strep, PpClpX-His was specifically pulled down by PpClpP2-Strep. (D) Strep pull-down assay with the PpClpP1P2 heterocomplex, PpClpX-His was pulled down by the PpClpP1P2 heterocomplex. (E-F) Degradation assays of fluorescent substrates by the respective active protease complexes. Degradation of GFP-ssrA (E) or FITC-casein (F). Significant protease activity is observed only in the presence of both PpClpX (full-length) and PpClpP2 or the PpClpP1P2 heterocomplex. Values indicated at the base of each bar denote the fold-change in fluorescence intensity relative to the BSA-treated control. Bars represent the means of three independent experiments ± SD. All experiments were performed in triplicate. **p < 0.01.

https://doi.org/10.1371/journal.ppat.1013909.g005

We further evaluated the proteolytic activity of the PpClpX (full-length) in complex with different PpClpP assemblies (ClpP1, ClpP2, or the ClpP1P2 heterocomplex) using two model substrates: GFP-ssrA and FITC-casein. In the absence of PpClpX, neither PpClpP1, PpClpP2, PpClpX, nor PpClpP1P2 alone degraded GFP-ssrA or FITC-casein (Fig 5E and 5F). In contrast, both the PpClpX + PpClpP2 and PpClpX + PpClpP1P2 complexes efficiently degraded these substrates, indicating that PpClpX functions specifically through recognition of PpClpP2 to activate proteolysis. Notably, the proteolytic activity of PpClpX + PpClpP1P2 was slightly higher than that of PpClpX + PpClpP2, suggesting that the heterocomplex PpClpP1P2 possess enhanced hydrolytic capability compared to PpClpP2 alone.

3.7. Bortezomib inhibits PpClpP1 peptidase activity

Our biochemical assays confirmed that both PpClpP1 and PpClpP2 possess peptidase activity, indicating their significant roles in bacterial physiology. We constructed knockout mutants of P. plecoglossicida PQLYC4 by deleting the clpP1 and clpP2 genes individually (S5C and S5D Fig). Deletion of either gene resulted in a significant growth defect compared to the wild-type strain. Notably, the growth impairment observed in the ΔPpClpP2 mutant was more severe than that in the ΔPpClpP1 mutant (Fig 6A). And a significant deficiency in both swimming and swarming motilities were observed in the ΔPpClpP1 and ΔPpClpP2 strains (S5E and S5F Fig). These findings demonstrate that both PpClpP1 and PpClpP2 are indispensable for the normal growth and bacterial motility of P. plecoglossicida.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Characterization of BTZ-mediated inhibition of PpClpP1 and its antibacterial effects.

(A) Determination of the growth curves of the wild-type P. plecoglossicida PQLYC4, ΔPpClpP1, ΔPpClpP2, and complemented strains (designated as cΔPpClpP1and cΔPpClpP2). (B) Differential scanning fluorimetry (DSF) analysis of PpClpP1 in the presence of small molecules (ADEP and BTZ). (C) Isothermal titration calorimetry (ITC) binding isotherm for BTZ (300 μM) titrated into PpClpP1 (30 μM), with derived binding parameters (Kd = 98 ± 3.41 μM). (D) Effect of BTZ on the peptidase activity of PpClpP1. (E) Effect of BTZ on the peptidase activity of PpClpP2. BTZ sensitivity assays of wild-type P. plecoglossicida PQLYC4 (F), ΔPpClpP1 (G), and ΔPpClpP2 mutant strains (H). Left, the bacterial cultures were adjusted to OD₆₀₀ ≈ 0.4 with double distilled water, and 10-fold serial dilution was performed on TSB (with 1.25 μM BTZ), and plated on TSB agar to determine the number of surviving colony forming bacteria. Right, when freshly cultured bacteria grew to an OD₆₀₀ value of 0.1, add BTZ small molecules at different concentrations (μM), spectrophotometry was used to measure the OD₆₀₀ value at appropriate intervals. Bars represent the means of three independent experiments ± SD. All experiments were performed in triplicate.

https://doi.org/10.1371/journal.ppat.1013909.g006

Numerous studies have demonstrated that various small molecules can modulate ClpP protease activity, either activating or inhibiting its function [18–20,43]. In our systematic screening of PpClpP modulators for structural studies, we evaluated two chemically distinct compounds: the natural product acyldepsipeptide (ADEP) [44,45], and the proteasome inhibitor bortezomib (BTZ) [46], both previously identified as ClpP complex modulators and proposed as potential antimicrobial candidates. Differential scanning fluorimetry (DSF) revealed that BTZ induced significant thermal stabilization of PpClpP1 (ΔTm = +10.6°C, Fig 6B), while ADEP showed no stabilizing effect. Isothermal titration calorimetry (ITC) confirmed direct BTZ-PpClpP1 binding with moderate affinity (Kd = 98 ± 3.41 μM, Fig 6C). BTZ, an N-protected dipeptide boronic acid that forms covalent adducts with catalytic serines or threonines, demonstrated potent inhibition of PpClpP1 peptidase activity. At 1 μM concentration, the BTZ completely abolished the Suc-LLVY-AMC hydrolysis (Fig 6D). In contrast, both DSF (S5G Fig) and ITC (S5H Fig) analyses revealed that BTZ does not interact with PpClpP2, and its addition did not affect the peptidase activity of PpClpP2 (Fig 6E). Notably, BTZ exhibited opposing effects on the peptidase activity of different ClpP homologs: it functioned as an activator for TtClpP [20] but as an inhibitor for PpClpP1.

Given that BTZ inhibits the enzymatic activity of PpClpP1, we investigated its direct impact on bacterial growth. Bacterial plate titration assays revealed that BTZ (at 1 μM) significantly suppressed the growth of P. plecoglossicida (Fig 6F-left). Consistent with this, bacterial growth curve analysis demonstrated that BTZ inhibits bacterial growth in a concentration-dependent manner (Fig 6F-right). To ascertain the specificity of BTZ, we tested its effect on isogenic knockout mutants. BTZ failed to inhibit the growth of the ΔPpClpP1 mutant (Fig 6G), whereas it retained a potent bacteriostatic effect on the ΔPpClpP2 mutant (Fig 6H). Collectively, these results indicated that the antibacterial activity of BTZ against P. plecoglossicida is specifically mediated through targeting PpClpP1, and that BTZ inhibits bacterial growth by inhibiting the peptidase activity of PpClpP1.

3.8. Binding mode of bortezomib to PpClpP1

To elucidate the structural mechanism of BTZ-mediated inhibition, we performed molecular docking simulations with the PpClpP1 monomer, which identified a cluster of key potential interacting residues (G44, E45, C46, S47, F98, H99, W100, T101, S117, and D121, Fig 7A-left). Initial simulations positioned BTZ within a hydrophobic pocket formed by loop 3 and helices B/D of individual subunits (Fig 7A-right). However, complete occupancy of the tetradecameric lumen prevented clear visualization of ligand-binding conformations. Selective removal of two opposing monomers revealed BTZ distributed along the proteolytic chamber axis, with predominant binding at the loop 3 and helix B/D interface (Fig 7B). High-resolution docking models identified three key interaction networks (Fig 7C): (1) Hydrogen bonds between BTZ’s boronic acid group and Cys46/Ser47; (2) Hydrogen bonds between BTZ’s pyrazine ring and Thr101; (3) Hydrophobic interactions between BTZ’s phenyl group and Glu45/Phe98. Notably, BTZ’s phenyl ring occupies a position 4.1 Å from catalytic Ser73 and 5.7 Å from His96－ sufficiently proximal to sterically hinder substrate access yet distant from direct catalytic triad engagement (Fig 7D). To delineate the precise contribution of specific amino acids to BTZ binding, we generated a series of alanine-scanning mutants: PpClpP1 (G44A, E45A, C46A, S47A, F98A, H99A, W100A, T101A, S117A, and D121A). Subsequent analysis by DSF (Fig 7E) and ITC (Figs 7F and S6A–S6F) assays confirmed that several of these residues are critical for BTZ binding. Specifically, mutations of E45, C46, S47, F98, W100, and T101 to alanine significantly diminished binding affinity, highlighting their importance for BTZ binding.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Elucidation of the BTZ binding mode within PpClpP1.

(A) 2D-molecular docking (left) and 3D-molecular docking (right) between BTZ and the PpClpP1 monomer. (B) The surface structure diagram of the PpClpP1 tetradecamer, with orange and cyan representing the PpClpP1 monomers in the two heptameric rings, respectively. The black dashed lines outline the removed two PpClpP1 monomers (left). Depiction of the hydrophobic binding pocket that BTZ binds in the context of the PpClpP1 monomer. Red sticks are the small molecule BTZ. (C) Substrate-binding pocket of PpClpP1, the residues involved in the binding to the BTZ are shown as sticks. (D) Enlarged view of the catalytic triad and BTZ in the PpClpP1 monomer. The black measurement lines represent the average distances between residues S73/H96 and the benzene ring on BTZ, which are 4.1 Å and 5.7 Å respectively. (E) Differential scanning fluorimetry (DSF) analysis of various PpClpP1 mutant proteins with BTZ. The ΔTm value represents the temperature difference between the melting temperature (Tm) of the PpClpP1 mutant + BTZ and that of the wild-type PpClpP1 + BTZ. (F) The binding affinity of BTZ to wild-type PpClpP1 and its mutants was quantified by isothermal titration calorimetry (ITC). The derived Kd values are shown within. Results are representative of those from three independent experiments. Data represent mean values ± SD.

https://doi.org/10.1371/journal.ppat.1013909.g007

3.9. Comparative analysis of ligand-ClpP binding modes across species

The structural uniqueness of PpClpP1 prompted a detailed comparison of ligand-binding mechanisms across species. Structural alignment of our BTZ-bound PpClpP1 model (Fig 8A) with the TtClpP-BTZ complex (Fig 8B) revealed both conserved and divergent features. While BTZ occupies similar positions within the hydrophobic vestibule of both proteases, its orientation differs significantly - the boronic acid moiety faces loop3 in PpClpP1 but projects outward in TtClpP (Fig 8C). More importantly, the key residues mediating BTZ binding are completely distinct between PpClpP1 and TtClpP (Figs 8D, 8E and S7A). In TtClpP, BTZ directly engages catalytic triad residues S97 (through hydrogen bonding) and H122 (via π-cation interaction), thereby stabilizing the charge-relay system and enhancing proteolytic activity (Fig 8E) [20]. The distinct binding mode of BTZ to PpClpP1, compared to TtClpP, arises from structural differences in the active site. TtClpP possesses a cavity (S7B Fig) that enables the boronic acid group of BTZ to extend toward and interact with the catalytic residues S97 and H122 (S7C Fig). Conversely, the absence of this cavity in PpClpP1 creates steric hindrance (S7D Fig), preventing the boronic acid group from positioning itself near the catalytic triad (S7E Fig). This binding mode of BTZ fails to stabilize the catalytic triad of PpClpP1. Instead, interactions with residues surrounding the triad may alter its conformation, resulting in reduced or complete loss of enzymatic activity (S7F Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Comparative analysis of ligand binding modes across ClpP homologs.

(A) Cartoon representation of the PpClpP1 monomer in BTZ-bound (marked red) and (B) TtClpP BTZ-bound (marked gray) states. (C) Superposition of BTZ-bound PpClpP (cyan) and TtClpP (light blue) reveals distinct binding geometries (left), with electrostatic potential maps illustrating differential interaction patterns (right). Detailed views of binding interfaces demonstrate residue-specific contacts (red spheres) in PpClpP1 (D) versus TtClpP (E). (F) Comparative analysis includes ADEP-bound (orange) EcClpP dimer, showing inter-subunit binding (violet/light pink subunits) with surface and electrostatic potential representations (right) that contrast sharply with BTZ’s monomer-restricted binding mode.

https://doi.org/10.1371/journal.ppat.1013909.g008

ADEPs represent another important class of ClpP modulators with fundamentally different binding topologies. Structural studies of E. coli ClpP (EcClpP) [18] show that ADEPs occupy an inter-subunit hydrophobic cleft formed by α-helices B-C of one monomer and α-helices A, β-strands 2–4/9 of an adjacent monomer (Figs 8F and S7G), a binding mode conserved in BsClpP and MtbClpP1 (S7H Fig and S7I). This dual-subunit engagement contrasts sharply with BTZ’s monomer-restricted binding and drives distinct antimicrobial mechanisms. ADEP binding induces allosteric pore dilation through N-terminal helix displacement, simultaneously activating uncontrolled proteolysis while blocking Clp-ATPase docking. In contrast, BTZ directly modulates catalytic triad conformation, bypassing ATPase-mediated regulation. These differential binding modes define two distinct antibacterial strategies: ADEP-mediated proteostasis dysregulation versus BTZ-driven direct enzymatic modulation.