Surface cysteine to serine substitutions in IL-18 reduce aggregation and enhance activity

Plasmid construction and protein expression

All plasmids were synthesized and obtained from GenScript (New Jersey, USA). The open reading frame (ORF) of the WT IL-18, codon-optimized for E. coli Rosetta2 (DE3) pLysS expression system, was inserted between Nde l and Eco RI restriction sites of the pET28a expression vector. The IL-18 double mutant E6K+T63A (IL-18 DM) and IL-18 with multiple mutations E6K+T63A+C38S+C68S+C76S+C127S (IL-18 DM1234) were also obtained from the same source, cloned using similar restriction sites in the pET28a vector. All recombinant plasmids (pET28a-IL-18WT, pET28a-IL-18DM, and pET28a-IL-18DM1234) were transformed into E. coli Rosetta2 (DE3) pLysS cells and selected on Luria Bertani (LB) agar (Sigma-Aldrich, St. Louis, MO, USA) containing 34 mg/ml chloramphenicol (Invitrogen, Waltham, MA, USA) and 50 mg/ml kanamycin (Invitrogen, Waltham, MA, USA). Further, the selected clones were used for IL-18 production. Briefly, the recombinant clones bearing WT and mutated IL-18 were cultivated in LB broth supplemented with 34 mg/ml chloramphenicol and 50 mg/ml kanamycin and incubated in a shaker incubator (Wiggens, Straubenhardt, Germany) at 37 °C, 200 rpm until OD600 reached 0.6. Then, 1 M IPTG (isopropyl β-D-1-thiogalactopyranoside; Sigma-Aldrich, St. Louis, MO, USA) was added to the culture at a final concentration of 0.1 M and incubated at 25 °C, 200 rpm in a shaker incubator for 5 h. Post-production, cells were harvested by centrifugation at 10,000 rpm for 10 min at 4 °C and resuspended in 10 ml PBS for sonication (on ice) with 60% amplitude, pulsed at 10 s on/15 s off for 5 min with breaks between cycles to prevent warming of the mixture. The cell lysate was then centrifuged for 10 min at 10,000 rpm at 4 °C. The supernatant and cell pellet were separated and analyzed on 12% SDS-PAGE.

Step-wise refolding

The sonicated cell pellet was washed with 25 ml PBS thrice and centrifuged at 10,000 rpm for 10 min. The pellet was then collected, weighed, and resuspended in solubilization buffer (base buffer (50 mM Tris-HCl, 150 mM sodium chloride (NaCl), 100 mM glycine, 20% glycerol, pH 8.0) containing 8 M urea) at a final concentration of 2–8 mg/ml. The suspension was mixed in an orbital shaker for 30 min, and after centrifugation at 15,000 rpm for 15 min, the supernatant was filtered through a 0.45 µm syringe filter. The supernatant was dialyzed using slide-A-Lyzer G2 dialysis cassettes (Thermo Scientific, Waltham, MA, USA) in 1 L dialysis buffer (base buffer containing 6 M urea) for 4 h at 4 °C. During this step, 500 mL of base buffer was added to the dialysis buffer every 6 or 12 h 4 times. Finally, the dialysis cassette with the solubilized IL-18 inclusion bodies was placed in 2 ml base buffer for 6 h. The refolded IL-18 proteins (WT and mutants) were harvested and stored at 4 °C until further use.

Western blotting

The proteins obtained after running on a 12% SDS-PAGE gel transferred onto a polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA, USA) by electroblotting at 35 volts for 16 h in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, pH 8.3). Non-specific proteins were blocked by incubating the PVDF membrane in Tris-buffered saline plus 0.1% Tween 20 (TBST) containing 3% (w/v) bovine serum albumin (BSA) (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at 25 °C, followed by washing with TBST thrice. The membrane was then stained with a specific mouse anti-human IL-18 antibody (R&D Systems, Minneapolis, MN, USA; catalog number D043-3) at 1:3,000 dilution for 1 h at 25 °C and washed three times for 5 min each in TBST with gentle agitation. In the next step, horseradish peroxidase-conjugated goat anti-mouse IgG (R&D Systems, Minneapolis, MN, USA; catalog number HAF007) was added at a dilution of 1:10,000 in TBST containing 3% (w/v) BSA, and the membrane was incubated for 1 h at 25 °C. After washing three times in TBST, IL-18 was detected using the Luminata Forte Western HRP substrate (Millipore, Burlington, MA, USA) with the UVITEC Cambridge Gel Documentation System (UVITEC, Cambridge, UK).

Liquid Chromatography with tandem mass spectrometry (LC-MS/MS)

LC-MS/MS analysis was performed using the commercial service of the Proteomics International Company, Australia. Protein samples were trypsin-digested, and peptides were extracted according to standard techniques (Bringans et al., 2008). The peptides were analyzed by electrospray ionization mass spectrometry using a Shimadzu Prominence nano HPLC system (Shimadzu, Kyoto, Japan) coupled to a 5600 TripleTOF mass spectrometer (Sciex, Redwood City, CA, USA). Tryptic peptides were loaded onto an Agilent Zorbax 300SB-C18, 3.5 µm (Agilent Technologies, Santa Clara, CA, USA) and separated with a linear gradient of water/acetonitrile/0.1% formic acid (v/v). Spectra were analyzed to identify the proteins of interest using Mascot sequence matching software (Matrix Science, UK) with the UniProt database.

Protein aggregation assay

The WT and mutated IL-18 proteins were subjected to buffer exchange using a Vivaspin ultrafiltration unit (Sartorius, Göttingen, Germany) against PBS. To generate IL-18 aggregates, 40 µg/ml of each protein was incubated at 37 °C, 220 rpm for 16–18 h in a shaker incubator. The protein aggregation was monitored by ProteoStat Protein Aggregation Assay (Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer’s protocol. Briefly, the proteins were loaded in triplicate in 96 well fluorescent microplates, followed by the addition of ProteoStat detection dye. The reaction mixture was then incubated for 15 min in the dark at room temperature. The fluorescence signal was determined by excitation at 485 nm and emission at 620 nm using a Varioskan™ LUX multimode microplate reader (Thermo Scientific, USA).

IFN-γ induction assay

For the IFN-γ induction assay, the NK-92MI cell line was selected because it is a highly cytotoxic IL-2 independent NK cell line that expresses various types of cytokines (Yan et al., 1998; Tam et al., 1999). The cell line was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in complete α-minimum essential medium (MEM) supplemented with 12.5% fetal bovine serum (FBS) and 12.5% horse serum at 37 °C in a CO₂ incubator. The cell line was used to evaluate the activity of WT and mutated IL-18, and the experiment was performed as previously described (Saetang et al., 2016). Briefly, the NK-92MI cells were suspended in 0.2 ml complete α-MEM at 0. 5 × 10⁶ cells/ml. Then, IFN- γ was induced using each recombinant IL-18 protein at concentrations of 75, 150, and 300 ng/ml, respectively, with 0.5 ng/ml of human IL-12. After 16–18 h of incubation at 37 °C in a CO₂incubator, the supernatants were collected, and IFN-γ levels were estimated using the Human IFN-γ Quantikine ELISA Kit (R&D system, Minneapolis, MN, USA).

MD simulations

The crystal structure of human IL-18 (accession code 3WO2) was taken from the RCSB Protein Data Bank (PDB) (Tsutsumi et al., 2014). The co-crystallized solvents were removed from the structure. The cysteine-serine replacement was performed using the FoldX package (Schymkowitz et al., 2005). All ionizable amino acids were set to their default state at pH 7.0. Neutral histidine was set as a single ϵ-protonated histidine (HIE in AMBER name). Double protonated histidine or disulfide linkages were not observed. The IL-18 protein was then solvated using the TIP3P water rectangular box at a distance of 14 Å from the protein surface. The system was neutralized using sodium ions, and NaCl was added to create a 0.15 mol L⁻¹ NaCl solution (Saetang et al., 2016) using the AMBER (Assisted Model Building with Energy Refinement) 20 force field (Case et al., 2021) via the Leap module in the AMBER20 package. In total, the system consisted of monomeric IL-18, 4 sodium ions, 33 NaCl pairs, and 11437 TIP3P waters.

WT and mutated IL-18 structures were energy-minimized using steeping descent and conjugate gradient methods for 1,000 steps each under periodic boundary conditions. The minimized structures were subjected to canonical (NVT) simulation at 37 °C (310 K), where all protein positions were restrained by a harmonic potential. Temperature was controlled using a Langevin thermostat. All non-bonded and electrostatic interactions were calculated using a cutoff of 16 Å. Each NVT restrained simulation was executed for 200 picoseconds (ps) with a time step of 1 femtosecond (fs) and force constants of 200, 100, 50, 20, and 10 kcal/mol Å-2 in a total of 1 nanosecond (ns). To apply a pressure of 1.013 bar (1 atm), the system was changed to an isobaric-isothermal (NPT) simulation. Temperature and pressure were regulated using a weak-coupling algorithm (Berendsen et al., 1984). The NPT simulation lasted 120 ns, with a time step of 2 fs. The first 60 ns simulation was omitted, and the last 60 ns simulation was taken into the analyzed trajectory. All energy minimizations and molecular dynamics processes were performed using the Particle Mesh Ewald Molecular Dynamics (PMEMD) module implemented in the AMBER20 package. In addition, the trajectory in AMBER coordinate format (.mdcrd) file was converted into a PDB file using cpptraj in the AMBER20 package. Structural visualization and interaction analysis of IL-18 was performed using the Visual Molecular Dynamics (VMD) package (Humphrey, Dalke & Schulten, 1996).

Statistical analysis

All the experiments were performed in triplicate. Statistical analyses were performed using SPSS version 26 (IBM, Armonk, NY, USA), and one-way analysis of variance (ANOVA) followed by Bonferroni or Least Significant Difference (LSD) multiple-comparison tests were applied to compare multiple groups. All comparisons were made with respect to the untreated groups. The P-value < 0.05 was considered statistically significant.

Structural analysis of surface cysteines of IL-18

Cysteine residues present on the surface of IL-18 were analyzed in a previous study (Yamamoto et al., 2004); however, the source was NMR-derived PDB structure (PDB ID: 1J0S) (Kato et al., 2003). Currently, human IL-18 is crystallized together with its receptors (Tsutsumi et al., 2014), allowing researchers to study the role of each amino acid residue of IL-18 in its interaction with the receptor. In this study, we used the crystal structure of human IL-18 in complex with IL-18 receptor α (PDB ID: 3WO3) to re-analyze the role of surface cysteines in their interaction with the receptor. As shown in Figs. 1A and 1B, the surface cysteines, including C38, C68, C76, and C127, were independent and did not interact with each other. The thiol side chain of most cysteines was directed toward the surface of the IL-18 molecule. No intramolecular interactions were observed, suggesting the role of these surface cysteines in intermolecular interactions. Additionally, none of the cysteine residues were observed to be important for binding with IL-18 receptor α (Fig. 1B). This is in accordance with the previous report on NMR-derived IL-18 structure, which showed that amino acid replacement did not disrupt the interaction between IL-18 and its receptors (Yamamoto et al., 2004). Therefore, we mutated the surface cysteines to improve our previously modified human IL-18 double mutant.

Expression and refolding of recombinant IL-18 proteins

Expression of WT and mutated IL-18 proteins was conducted using E. coli Rosetta2 (DE3) pLysS expression system. After cell lysis, the soluble fractions and the insoluble inclusion bodies of crude proteins were analyzed by SDS-PAGE. Figure 2 demonstrates that a protein band of approximately 20 kDa corresponding to the mature form of human IL-18 was detected in both the soluble fractions and inclusion bodies; the same protein band was not observed in control transformants with an empty vector (Fig. S1). However, more than 80% of the proteins were observed in inclusion bodies (Fig. 2A, lanes 4, 5, and 6), suggesting that human IL-18 was expressed primarily in insoluble form as inclusion bodies in E. coli. Therefore, stepwise refolding was performed to refold the IL-18 protein to its correct structure. SDS-PAGE analysis showed that the protein purity was greater than 95% after the refolding step; however, the yield was less than that of the starting source (Fig. 2B). Western blotting confirmed IL-18 specific protein bands using an anti-human IL-18 antibody (Fig. 2C). Further, we validated recombinant IL-18 by excising the protein band from the gel and subjecting it to LC-MS/MS analysis. Tryptic digestion of the protein provided 9 peptide fragments which covered about 48.0% of the pro-form of IL-18 or 58.8% of the mature IL-18 (Figs. 3A and 3B). These data suggest that the refolded IL-18 is the mature form of IL-18.

IL-18 DM1234 showed lower aggregation capability and higher activity

To evaluate the effect of surface cysteine replacement with serine (C-to-S substitution) on IL-18 aggregation, oxidative stress was induced by shaking the refolded protein at 220 rpm in a shaker incubator. Heat (50 °C) was also used to increase the reaction energy. After 16–18 h of induction, the aggregated protein was measured using the ProteoStat protein aggregation assay. This method employs a fluorescent ProteoStat dye for protein aggregate detection, which is more sensitive than the oligomerization assay. Moreover, it is the only method that can detect protein aggregates at visible to sub-visible levels (Oshinbolu et al., 2018). The result demonstrated that protein aggregation was similar in IL-18 WT and IL-18 DM (Fig. 4A). However, IL-18 DM1234 showed a significant decrease in protein aggregation (Fig. 4A). C-to-S substitution lowered the aggregation ability of IL-18 DM1234 by approximately 1.79 and 1.63 folds compared to that of IL-18 WT and IL-18 DM, respectively (Data S1). Interestingly, the IFN-γ induction assay results demonstrated that the mutant IL-18 DM1234 showed a statistically increasing capability to stimulate IFN-γ production from NK-92MI cells (Fig. 4B). The IFN-γ levels induced by IL-18 DM1234 indicated higher efficacy of approximately 10.0 and 2.8 folds compared to that by IL-18 WT and IL-18 DM, respectively (Data S2).

C-to-S substitution may improve the stabilization of IL-18 DM1234 for receptor binding

Based on the results of the IFN-γ induction assay, MD simulations were performed to investigate the role of C-to-S substitution on the structure of IL-18, which may affect the activity of this protein. Figure 5A shows the backbone root-mean-square deviation (RMSD) of each IL-18 protein with different patterns. The results demonstrated that 120 ns of MD simulation with physiological parameters was a sufficient time point for equilibration. Interestingly, IL-18 WT and IL-18 DM revealed similar RMSD values and patterns, whereas IL-18 DM1234 showed lower RMSD deviations (Fig. 5A). In addition to RMSD, MD simulation was used to track the flexibility of IL-18 as a root-mean-square fluctuation (RMSF) value. Figure 5B illustrates the degree of flexibility of amino acid sequences of the three IL-18 proteins. The results showed a similar RMSF value pattern for most amino acid sequences of the IL-18 proteins. However, there were two regions between 30–50 and 100–110 amino acid residues where IL-18 DM showed noticeably higher RMSF values compared with the other regions, indicating that these amino acids displayed greater flexibility.

WT and mutated IL-18 protein structures were visualized and aligned using the VMD program to determine differences in their conformation. The results showed an apparent conformational change in IL-18 DM1234 at the α-helix part of binding site I when compared to IL-18 DM (Figs. 6A and 6B). The binding site I comprise methionine 33 (M33) and aspartate 35 (D35), identified as essential residues for binding of IL-18 with IL-18 receptor α (Kato et al., 2003). Figure 6B demonstrates that the conformational alteration of IL-18 DM1234 resulted in a higher similarity of its binding site I region with IL-18 WT, which may improve the stabilization of IL-18 DM1234 for IL-18 receptors, similar to IL-18 WT structure. Additionally, looking deeper into the direction of amino acids in this area, we observed that D35 of IL-18 DM1234 changed its direction closer to valine 125 (V125) and serine 127 (S127) of IL-18 receptor α; this change was not observed in IL-18 WT (Fig. 6C). Structural alignment revealed that the distance between D35 of IL-18 DM1234 and V125 and S127 of the IL-18 receptor was in the range of 2.88–4.72 Å (Fig. 6B). On the other hand, D35 of IL-18 WT showed a higher distance of 5.28 and 5.95 Å  between V125 and S127, respectively (Fig. 6C). Moreover, D35 of IL-18 DM also showed increased length of these molecular pairs (4.35 and 4.95 Å  for V125 and S127, respectively) when compared to IL-18 DM1234 (Fig. 6D).