Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Sep 5.
Published in final edited form as: Eur J Pharmacol. 2023 Jun 21;954:175873. doi: 10.1016/j.ejphar.2023.175873

New Insights on the Differential Interaction of Sulfiredoxin with Members of the Peroxiredoxin Family Revealed by Protein-Protein Docking and Experimental Studies

Murli Mishra 1,#, Hong Jiang 1,#, Qiou Wei 1,2,*
PMCID: PMC10426277  NIHMSID: NIHMS1914618  PMID: 37353187

Abstract

Sulfiredoxin (Srx) is the enzyme that restores the peroxidase activity of peroxiredoxins (Prxs) through catalyzing the reduction of hyperoxidized Prxs back to their active forms. This process involves protein-protein interaction in an enzyme-substrate binding manner. The integrity of the Srx-Prx axis contributes to the pathogenesis of various oxidative stress related human disorders including cancer, inflammation, cardiovascular and neurological diseases. The purpose of this study is to understand the structural and molecular biology of the Srx-Prx interaction, which may be of significance for prediction of target site for the novel drug-discovery. Homology modeling and protein-protein docking approaches were applied to examine the Srx-Prx interaction using online platforms including I-TASSER, Phyre2, Swissmodel, AlphaFold, MZDOCK and ZDOCK. By in-silico studies, A 26-amino acid motif at the C-terminus of Prx1 was predicted to cause a steric hindrance for the kinetics of the Srx-Prx1 interaction. These predictions were tested in-vitro using purified recombinant proteins including Srx, full-length Prxs, and C-terminus deleted Prxs. We confirmed that deletion of the C-terminus of Prxs significantly enhanced its rate of association with Srx (i.e. >1000 fold increase in the ka of the Srx-Prx1 interaction) with minimal effect on the rate of dissociation (kd). Differential interaction of Srx with individual members of the Prx family was further examined in cultured cells. Taken together, these data add novel molecular and structural insights critical for the understanding of the biology of the Srx-Prx interaction that may be of value for the development of targeted therapy for human disorders.

1. Introduction

Sulfiredoxin (Srx) is an exclusive enzyme that reduces the over-oxidized members of typical 2-Cys Peroxiredoxin (Prx) family, including Prx1 to Prx4 (Biteau et al., 2003). Srx utilizes ATP and magnesium (Mg2+) or manganese (Mn2+) as cofactors and forms sulfinic phosphoryl ester followed by thiosulfinate intermediate to reduce over-oxidized Prxs (Biteau et al., 2003; Lowther and Haynes, 2011). Srx is evolutionarily conserved in the majority of eukaryotes, but rare in prokaryotes with few exceptions (e.g. cyanobacteria). Apart from antioxidant function, Srx can also catalyze the deglutathionylation of proteins including actin, Prx2 and protein phosphatase (Findlay et al., 2006; Park et al., 2009). There can be few others, yet to be identified substrates that are deglutathionylated by Srx (Forshaw et al., 2021; Gao et al., 2009). Srx expression is altered in multiple types of human cancer. Our previous work has demonstrated oncogenic association of Srx with skin, colon, and lung tumorigenesis; where it is found to be highly over-expressed in tumors compared to adjacent normal tissue (Jiang et al., 2015; Jiang et al., 2014; Wei et al., 2013; Wei et al., 2011; Wu et al., 2014). Although the physiological as well as pathological significance of Srx has been studied in detail, the biochemistry of the Srx-Prx interaction and how this axis can be targeted for treatment have not been fully investigated and utilized.

The Srx-Prx axis can be explored as therapeutic target as well as therapeutic tools depending on their role in particular pathological condition. For example, individual Prx-isoforms can be considered as good therapeutic targets in lung cancer (Wei et al., 2011), glioblastoma (Kim et al., 2012), colorectal cancer (Wei et al., 2013), prostate cancer (Ummanni et al., 2012) etc. where they protect tumor cells. It is important to evaluate the risk-benefit ratio of targeting the individual members of the Srx-Prx axis as they also have protective role in normal (non-tumor) tissue. The Srx null mice have normal phenotype under laboratory conditions (Wei et al., 2013). Prx3 knockout mice born and mature normally (Li et al., 2007). Prx4 knockout mice have mild prostate atrophy (Iuchi et al., 2009). Prx1 and Prx2 knockout mice are reported to have some issue with erythropoiesis (Lee et al., 2003; Neumann et al., 2003) but they are otherwise normal. Hence, the majority of proteins in the Srx-Prx axis can be knocked-out without any life-threatening issue. Considering the risk associated with cancer, it is worth exploring the targets that can prolong the lives of patients by few extra years. Hence, benefits associated with targeting Srx or individual Prx outweighs the risk associated with it and the Srx-Prx system can be considered a therapeutic target in cancer. On the other hand, individual Prx-isoforms can be explored as therapeutic or diagnostic tools in Parkinson’s disease, Alzheimer’s disease, and diabetic complications (Chen et al., 2008; Findlay et al., 2005; Hu et al., 2011; Yoshida et al., 2009; Zeldich et al., 2014). These differential properties of individual components of the Srx-Prx system draw our attention towards differences in molecular properties of individual Prx-isoforms that gives them ability to play such diverse roles. Improved understanding of these molecular differences will help us in therapeutic intervention of the Srx-Prx system. However, lack of structural insights on the Srx-Prx interaction hinders the development of targeting strategies such as identification of small molecules and natural products to disrupt this axis for treatment of human disorders.

The molecular characteristics of protein-protein interaction must be identified to design a good targeting strategy for inhibition of such interactions. 3-dimensional structure of protein as well as its individual components can play major or minor roles to modulate the protein interaction. The understanding of molecular structure of individual proteins is the first criteria that must be fulfilled to study the effect of 3-dimensional structure of individual proteins on protein-protein interaction. Protein structure can be predicted experimentally using X-ray crystallography and NMR studies. In the absence of experimental data, it can be predicted computationally by homology modeling. Homology modeling is one of the most popular method for prediction of protein structures based on the known structure of homologous proteins with some sequence identity (Ring and Cohen, 1993). It is not trivial to predict the structure covering full length of protein using experimental methods as crystallizing whole protein is a cumbersome task that can be affected by multiple experimental factors leading to lower confidence in predicted structure. The ease of prediction of structure covering full protein sequence by homology modeling leads to popularity of this method. Homology modeling has already established its utility in making hypothesis for molecular studies (Kirubakaran et al., 2013; Trujillo et al., 2015). Protein structures predicted using these methods can be used computationally for protein-protein docking studies.

Protein-protein docking is a unique computational tool to identify the points of contact during protein-protein interaction that can help in designing the targeting strategy to inhibit those interactions (Park et al., 2015). Predictions of docking studies can be further confirmed experimentally using amino acid mutations and deletion studies. Recombinant proteins can be designed with mutations at individual points of contacts or deletion of particular domain. These mutants can be used to study the effect of particular mutations on protein-protein interactions. Once confirmed, amino acids from aforementioned experiments can be defined as the binding sites for protein-small molecule docking or virtual screening to identify the inhibitors of interaction. We carried out protein-protein docking studies followed by deletion mutation to confirm the role of Prx C-terminal arm in the Srx-Prx binding.

2. Materials and Methods

2.1. Homology modeling and protein-protein docking studies

Crystallographic structures of truncated Srx (PDB 1XW3 or 1XW4 with ADP bound) and its complex with Prx1 (PDB 2RII) were available in Protein Data Bank as previously documented. Those entries actually do not cover the full sequence of these proteins. Therefore, homology modeling was used to predict full-length structure of all Prxs and Srx. We used multiple online homology modeling programs in this experiment, which include I-TASSER (iterative threading assembly refinement) (Roy et al., 2010; Zhang, 2008), Phyre2 (Kelley and Sternberg, 2009), Swissmodel (Arnold et al., 2006), and AlphaFold (Jumper et al., 2021; Varadi et al., 2022). After comparison of the predicted structures, we used MZDOCK (Pierce et al., 2005) for prediction of dimeric structure from monomeric structure predicted by I-TASSER and Phyre2. Followed by homology modeling studies, we carried out protein-protein docking studies using ZDOCK (Pierce et al., 2011) online server to identify structural characteristics of interaction. We analyzed the final output of these experiments using PyMOL visualizer and labeled the binding and catalytic site components.

2.2. Purification of recombinant proteins and study of the Srx-Prx1 interaction in vitro by IP

The Srx, Prx1wildtype, Prx1mutant (in which last 22 amino acid from C-terminal are deleted), Prx4wildtype, and Prx4mutant were cloned into pRSET B vector and transformed into E. Coli BL21(DE3) strain for protein expression. The Srx coding region was inserted between BamHI and EcoRI restriction sites of pRSET B. All other coding regions of the Prx wildtype as well as Prx mutants were inserted between BamHI and HindIII restriction sites of pRSET B. All recombinant proteins had a (His)6 tags at N-terminal. The BL21(DE3) bacteria were cultured at 37 °C in LB broth media (Sigma-Aldrich). After the addition of isopropyl-1-thio-D-galactopyranoside (1.0 mM), the cultures were incubated for 4 hours at room temperature. The cells were lysed using lysis buffer [8M Urea (pH 8.0), 100 mM Monosodium phosphate, 10 mM Tris Base]. The protein purification was carried out using Ni2+ charged IMAC Select Affinity Gel column for purification of His-tagged proteins following the manufacturer’s suggested protocol (Qiagen). The column was washed with wash buffer [20 mM Imidazole (pH 8.0), 300 mM NaCl, and 50 mM Monosodium phosphate]. The protein was eluted with elution buffer [300 mM Imidazole (pH 8.0), 300 mM NaCl, 50 mM Monosodium phosphate] and was dialyzed against multiple rounds of 20 mM Tris-HCl buffer for overnight to remove extra salts and re-nature the protein. To study the protein-protein interaction of purified recombinant proteins, Srx (2 µg) was incubated with multiple concentrations of Prx1 wildtype and Prx1mutant (i.e. 500 ng, 1µg, 2 µg, 4µg) in 500 µL of IP buffer for 2 hours at 4 °C. After 2 hours, IP and western-blot was carried out using standard procedures as previously reported (Wei et al., 2011).

2.3. Study of the Srx-Prx interaction kinetics using Surface Plasmon Resonance

The interaction of Srx (analyte) with Prx1wildtype and Prx1mutant (ligand) were measured by Surface Plasmon Resonance (SPR) technique using ProteOnTM XPR36 instrument (Bio-Rad). The ligand was immobilized on the GLC sensor chips (Bio-Rad) using amine coupling method. Ligand capturing on the GLC chip was performed as per the manufacturer’s protocol. The SPR assay was performed with assay buffer containing 20 mM HEPES, pH7.5, 150 mM NaCl, 5 mM DTT, 100 µM TCEP, 0.005% tween20, 0.1% BSA and 5 mg/ml dextran. Several different concentrations of pure recombinant Srx (analyte) were used to evaluate the ligand-analyte binding. The data were acquired and processed by ProteOn manager software and Langmuir 1:1 evaluation model is used for analysis.

2.4. Western Blot and Immunoprecipitation (IP) assay in HEK293T cells

HEK293T (human embryonic kidney) cell line was originally purchased from ATCC and cultured in Dulbecco’s modified Eagle’s medium (Gibco) supplemented with 10% fetal bovine serum (Gibco) and antibiotics including penicillin-streptomycin solution (Thermo scientific) [penicillin 100 U/ml and streptomycin 100 µg/ml] and 5 µg/ml gentamycin (Gibco, Life Technologies). Cells were cultured in CO2 incubator with an atmosphere of 100% humidity and 5% CO2 at 37 °C. Transient transfection for protein expression was performed as previously reported using Fugene 6 (Matthews et al., 2007). Cells were transfected with expression vector for Flag-tagged Srx. Cells were divided into three groups and were treated with vehicle, 500 µM H2O2 and 1000 µM of H2O2 for 10 minutes and then washed three times with base medium before replenished with fresh complete culture medium. Vehicle treated group was used as control. Western blot and IP were carried out using standard procedures as previously reported (Wei et al., 2011). Briefly, cells were gently washed three times with saline before being harvested in radioimmunoprecipitation assay (RIPA) buffer containing 50 mM Tris at pH 7.4, 150 mM NaCl, 1% NP-40, 1 mM EDTA, 0.25% sodium deoxycholate, 0.6 mM PMSF and a mixture of 1% protease cocktail inhibitors (Santa Cruz Biotech, Dallas, TX). Cell lysates were mixed with Laemmli buffer in the presence of β-mercaptoethanol and denatured by heating. Proteins were separated by sodium-dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred onto PVDF membrane. Membranes were blocked for one hour in 5% nonfat dry milk in tris buffer before overnight incubation with diluted primary antibody. The membrane was then washed with TBST and incubated for one hour with HRP-conjugated secondary antibody. After multiple washing steps, signals were detected using western dura chemiluminescence substrate (Pierce) and bands were visualized using Amersham Imager 680 (GE Healthcare Biosciences). Band intensity was quantitated using Image J software. IP procedure using anti-Flag antibody was performed as previously reported and followed by western blotting. Primary antibodies used including anti-Srx, anti-Prx4 (Proteintech), anti-Flag and anti-β-actin (Sigma–Aldrich).

2.5. Statistical analysis

Quantitative data were presented as means ± standard deviation (x̅±SD). Data were analyzed with indicated statistical methods by using SigmaPlot (version 13.0). SPR data was analyzed using Langmuir 1:1 evaluation model. For calculation of the p-value, parameters of two-tailed, 95% confidence interval were used for all analysis. p≤ 0.05 is considered statistically significant.

3. Results

3.1. Complete 3D-structure of full length Srx and Prxs predicted using homology modeling

There are no crystallography studies of full-length Srx or Prxs to reveal their complete structures in detail. Based on the available crystallography study of the Srx-Prx1 complex (Jonsson et al., 2009), we used freely available homology modeling tools to predict structure with Prx1 structure as a quality control. Several homology modeling programs were used to predict the monomeric and dimeric structure of Srx and Prxs. Accuracy of these tools substantially differed from each other. We finally selected the tool based on how closely it could predict the Prx1 structure compared to the published. Once selected, the same tool was used further for homology modeling of other PRX structures. Swissmodel predicted quite close but only partial structure which we could not use for our purpose. Phyre2 and AlphaFold gave good results too. However, I-TASSER proved to be the best for all proteins under this study. Although I-TASSER is time consuming but the ability of predicting best structure from minimal information was appreciable. Predicted structure of Prx1 and Srx monomer from I-TASSER were obtained and used in this study (Fig. 1). Prx1 monomer was further uploaded to MZDOCK to predict the structure of Prx1 dimer. We received multiple structures with different scores for each of our prediction, and best prediction was determined by the priority of individual score and a close comparison of previously reported partial structure of the Srx-Prx1 complex from crystallography studies. An overhanging of the C-terminal arm in Prx1 that covers the peroxidatic (CP) and resolving (CR) cysteine of Prx1 dimer was observed (Fig. 1C). After deletion of the c-terminal 22 aa, the predicted structure of Prx1mutant dimer is lack of such overhang (Fig. 1D). Similar strategy was used to predict the 3-dimensional structures of other 2-Cys containing Prxs including Prx2, 3 and 4. We found that structures of these Prxs largely resemble that of Prx1 except the difference in the orientation of the C-terminal overhang (Fig. 2).

Fig. 1. Predicted example structures of Srx and wildtype/mutant Prx1 homodimers obtained from I-TASSER modeling.

Fig. 1

Open in a new tab

(A, B) Structure of Srx in ribbon (A) and spheres (B). (C, D) Structures of homodimers formed by wildtype Prx1 (C) and c-terminal deleted Prx1mutant (D) in ribbon. Critical cysteines in Srx and Prx1 are marked in red, other amino acids in Srx that forms a groove to bind Prx1 are marked in magenta, the C-terminal 22 aa of Prx1 are marked in orange. Note that the C-terminal arm of Prx1 is hanging out of the dimer.

Fig. 2. Predicted example structures of other typical 2-Cys containing Prx dimers including Prx2–4 obtained from I-TASSER modeling.

Fig. 2

Open in a new tab

Structures of wildtype Prx2 (A), Prx3 (B), and Prx4 (C) obtained by homology modeling. Critical cysteines are marked in red, the C-terminal 22 aa of Prx are marked in orange.

3.2. Protein-protein docking studies reveal that the c-terminus of Prxs creates a steric hindrance for the Srx-Prx interaction

Previous studies of the Srx-Prx1 complex structure using crystallography revealed that two molecules of Srx interacts with the dimer of Prx1 through a hydrophobic binding pocket, this complex structure is available at PDB (Jonsson et al., 2009) (Fig. 3A and 3B). In this study, the N-terminal truncated Srx was used to facilitate the formation of crystals. The hydrophobic pocket of Srx along with CP and CR of Prx were defined as the Srx-Prx interaction interface. From I-TASSER prediction, the N-terminus of Srx swings away from the hydrophobic pocket and thus does not affect the binding of Prxs. Predicted full-length structures of Srx and Prx1 were used in protein-protein docking studies, which were carried out using ZDOCK online server. Docking output was visualized using PyMOL visualizer. Results of docking indicate that the overhanging arm of Prx1 may partially block the binding part of Prx1 that needs to be accessed by Srx, which creates a significant steric hindrance that negatively impacts the binding of Srx to Prx1 (Fig. 3E). Deletion of the C-terminus of Prx1 in the modeling leads to the absence of such hindrance (Fig. 3F). Similar steric hindrance is also observed in the modelling of Srx binding to other typical 2-Cys Prxs. However, the extent of hindrance is not likely to be uniform due to the variation of the orientation of each C-terminal arm in individual Prx. Theoretically, the existence of the C-terminus and their orientation may thus result in affinity differences between the binding of Srx to individual member of the typical 2-Cys containing Prxs. Therefore, it is likely possible that deletion of the C-terminal arm in Prxs will reduce such steric hindrance and results in higher binding affinity to Srx, as such deletion does not affect their major conformation or the motif within the vicinity of the protein-protein interaction interface (Fig. 3F).

Fig. 3. The c-terminus of Prxs creates a steric hindrance for their interaction with Srx in protein-protein docking studies using ZDOCK.

Fig. 3

Open in a new tab

(A,B) Crystallography structure of N-terminal truncated Srx (white) complexes with Prx1 (green). (C,D) The structure of Srx in full length predicted using homology modeling. The N-terminal amino acids 1–37 was marked yellow to highlight the truncated part and its orientation with respect to the hydrophobic Prx-binding pocket. (E,F) Structures of full-length Srx bound to wildtype Prx1 in full length (E) or mutant Prx1 missing the c-terminus (F). The binding site was highlighted in magenta, the catalytic/peroxidatic cysteine of Prx1 in red and the last 22 aa of the C-terminus of Prx in orange.

3.3. Examination of the Srx-Prx interaction in vitro using purified recombinant proteins in full length or truncated mutants.

Above homology modeling predicts that differences in the orientation of C-terminal arm in different Prx may affect their binding to Srx. To test such predictions experimentally, coding sequences of human Srx or Prxs were inserted into plasmid vectors to make constructs that express recombinant full-length or truncated proteins in E. Coli (Fig. 4A). In truncation mutants of Prx1 or Prx4, the c-terminal 22 amino acids were deleted. We successfully purified recombinant full-length Srx, full-length Prx1 and c-terminus deleted Prx1 (Prx1mutant), full-length Prx4 and c-terminus deleted Prx4 (Prx4mutant) (Fig. 4B). These proteins were used to examine their ability and affinity to bind with Srx by immunoprecipitation (IP) with anti-Srx antibody. When equal amounts of Prx1wildtype and Prx1mutant were incubated with a fixed amount of Srx, more of Prx1mutant were pulled-down compared to Prx1wildtype (Fig. 5A and 5B). The differences in binding is not so obvious at lower Prx concentrations as at such concentrations there is plenty of Srx available for each molecule of Prx1wildtype as well as Prx1mutant. Hence, all the Prx in the media is pulled down by anti-Srx antibody. However, the differences become quite obvious at higher Prx1 concentrations. As the Srx concentration becomes a limiting factor for interaction, the greater fraction of the Prx with higher affinity (i.e. Prx1mutant) for Srx is pulled-down with anti-Srx antibody. These results confirm that Prx1mutant has higher steady-state interaction potential for Srx compared to Prx1wildtype. It also emphasizes the possibility that C-terminal arm of Prx1 may be causing some steric hindrance for the Srx-Prx1 interaction. Experiments above indicate that the deletion mutation enhances the steady-state of Srx-Prx interaction. However, it was not clear whether effect on the steady state interaction is due to changes in the rate of association or dissociation or both. Surface plasmon resonance (SPR) is a technique of choice to study kinetics of protein-protein interaction (Huber and Mueller, 2006). This technique was used for studying the effect of C-terminal deletion on the kinetics of the Srx-Prx interaction (Fig. 6). In the case of Srx-Prx1 interaction, the c-terminal deletion mutation resulted in more than 1,000-fold increase in association rate constant (ka) compared with wildtype Prx1 (Fig. 6A and 6B). In other words, in the presence of equivalent molar concentrations of Prx1wildtype and Prx1mutant, the ka for the Srx-Prx1mutant interaction was more than 1,000 fold higher compared to ka for the Srx-Prx1wildtype interaction. In the case of Srx-Prx4 interaction, deletion of the C-terminal arm in Prx4 resulted in more than 100 fold faster association (ka) (Fig. 6C & D). However, in both cases we found that the deletion of c-terminus did not significantly affect the dissociation rate constant (kd) (table 1).

Fig. 4.

Fig. 4

Open in a new tab

Purification of recombinant Srx and Prxs in full length or c-terminal deleted mutants. All recombinant proteins were His and Xpress tagged at the N-terminus. Recombinant proteins including Srx (A), Prx1 and mutant, Prx4 and mutant (B) were expressed in E. Coli, purified using affinity chromatography and examined by coomassie blue staining.

Fig. 5. Deletion of the C-terminal arm in Prxs enhances their interaction with Srx in vitro.

Fig. 5

Open in a new tab

For the same amount of Prx1 wildtype and mutant incubated with fixed amount of Srx, more of Prx1mutant is pulled-down along with Srx compared to Prx1wildtype (A) Western blot showing amount of Prx1wildtype and Prx1mutant pulled down along with Srx, (B) Quantitated values of Prx1widltype and Prx1mutant pulled down at each concentration of Prx1.

Fig. 6. Deletion of the C-terminus of Prxs increases the rate of association to Srx but does not change its dissociation.

Fig. 6

Open in a new tab

The binding kinetic parameters of Srx to different forms of Prxs including full-length Prx1 (A), Prx1mutant (B), full-length Prx4 (C), and Prx4mutant (D) were determined using SPR technique. Ka: rate of association constant; Kd, rate of dissociation constant. KD, equilibrium constant.

Table 1. The C-terminal deletion mutation increases the Srx-Prx affinity:

The kinetic parameters calculated using SPR indicates faster rate of association and higher affinity of Srx for Prx1mutant compared to Prx1wildtype

Parameters (Unit) Ka (1/Ms) Kd (1/s) KD (M) Comments
Prx1 wildtype 6.93 E−01 5.32 E−04 7.69 E−04 Prx1 has slow rate of association but very slow rate of dissociation. It results in longer time required to form the Srx-Prx interaction but highly stable complex.
Prx1mutant
(last 22 amino acids from Prx1 C-terminal are deleted)		 2.54 E+03 4.44 E−04 1.75 E−07 Deletion mutation results in more than 1,000 fold increase in rate of association with minimal effect on rate of dissociation.
Prx4 wildtype 5.09 E−01 1.66 E−03 3.26 E−03 Compared to Prx1 wildtype, the Srx-Prx4 complex dissociates faster
Prx4mutant
(last 22 amino acids from Prx1 C-terminal are deleted)		 9.04 E+01 2.69 E−03 2.97 E−05 Deletion mutation results in more than 100 fold increase in rate of association with minimal effect on rate of dissociation.
Open in a new tab

3.4. Examination of the Srx-Prx interaction in culture cells demonstrate the differences in interaction of Srx with Prxs.

In mammalian cells, it is well documented that Prxs react with hydrogen peroxide to mediate oxidative signaling as well as to keep redox balance through the oxidation and further hyperoxidation of their catalytic cysteine residues. The primary function of Srx is to reactivate hyperoxidized Prxs through the reduction reaction, thus the accessibility of individual Prx to Srx may affect its rate of reduction. Therefore, we asked whether there is any difference in the formation complex between Srx with individual Prx in cultured cells. HEK293 cells in culture were temporarily exposed to hydrogen peroxide to induce oxidative stress, and its effect on the levels of hyperoxidized Prxs as well as the formation of the Srx-Prx complexes were examined. We found that short exposure of these cells to hydrogen peroxide induces significantly increased levels of hyperoxidized Prxs without affecting their protein expression levels (Fig. 4A left panel). Anti-Srx antibody was used to pull down the Srx-Prx complexes, and we found that Srx interacts with various basal levels of individual Prx, mainly with Prx1, 2 and 4. In the case of Prx3, there is very low basal level of interacting with Srx, but exposure to hydrogen peroxide significantly increases the formation of the Srx-Prx3 complexes (Fig. 7A right panel and 7B). This increase may be due to the mitochondria-specific localization of Prx3 whereas other Prxs are mainly localized in the cytosol, and the translocation of Srx from the cytosol to mitochondria under oxidative stress as previously reported. Such differences in the interaction of these Prxs to Srx under oxidative stress conditions may be attributed to molecular rearrangements based on the individual orientation of the C-terminal arm in Prxs as predicted from homology modeling. Another possibility is that the C-terminal arm of Prx may affect conformational change as it is known to re-orient itself after oxidation. Taken together, all these factors lead to the variations on the members of Prx family to interact with Srx and different rate of reactivation.

Fig. 7. Effect of H2O2 treatment on the Srx interaction with various typical 2-Cys Prx-isoforms in cultured HEK293T cells.

Fig. 7

Open in a new tab

(A) Representative western blot results showing the relative amounts of Prx-isoforms and oxidized Prxs in cell lysates, and the amounts of Prxs pulled down by anti-Srx IP under control or oxidative stress conditions. (B) Quantitative plot of band intensity of individual Prx-isoforms pulled down under oxidative stress condition.

4. Discussion

The molecular interaction between proteins can regulate variety of cell signaling processes leading to their unique role in physiological homeostasis as well as pathological conditions. The importance of these macromolecules has led to the development of multiple tools that can help to get insight from their molecular structures. There are various experimental methods available to study the structure of proteins. However, these methods have limitations in maintaining the conformation of native protein in an environment suitable for structural prediction by NMR or X-ray crystallography. It takes years of research by a group of structural chemists to figure out the structure of a simple protein. Often times, these research efforts are not enough to predict the complete structure of proteins. The time and efforts required for structural predictions using experimental methods and the limitations of these methods led to development of computational tools that can help structural chemists to temporarily fill the gap in existing knowledge. Homology modeling is such a computational method of protein structural prediction that got great potential to fill the gap in existing data within acceptable limits of error (Khan et al., 2015).

In our study, we used I-TASSER, Phyre2, Swissmodel, and AlphaFold to predict protein structure. Each of these methods utilizes different algorithms and techniques and has pros and cons for protein structure prediction. Among them, AlphaFold is relatively new and may not be as widely used or tested as some other methods. It uses advanced deep learning algorithms that are able to accurately predict both global and local structural features. It is able to produce high-quality models for a wide range of protein targets, including many previously unsolved structures. However, it requires a significant amount of computational power and resources to run. Phyre2 and Swissmodel are able to predict protein structures for a wide range of protein sequences and can be used to model both single-domain and multi-domain proteins. They are not as accurate as some other methods for predicting local structural features and are difficult to produce high-quality models for proteins with low sequence identity to known structures. I-TASSER uses advanced algorithms for identifying structural templates, which results in higher accuracy predictions. It is fast and produces accurate models for a wide range of protein targets and can be used to model both single-domain and multi-domain proteins. In particular, its accuracy is further increased for proteins that have close homologs with known structures. Predicted structures of Srx and Prxs from these software are largely similar, but predictions from I-TASSER have more details thus were used in our study. Nevertheless, each method has its own strengths and weaknesses, and the choice of method depends on the specific needs of each application.

In addition, I-TASSER is a tool designed based on multiple templates that includes inbuilt fragment guided molecular dynamics simulation. As our homology modeling tool included the molecular dynamics simulation, we did not consider additional molecular dynamics tools. The molecular dynamics simulation was not performed at that time as we were more interested in studying the steric factors affecting the Srx-Prx interactions. Our group is continuing the study the Srx-Prx interactions at molecular levels. In the future, we may study non-steric factors contributing to Srx-Prx interactions and include insights from molecular dynamics simulation if possible along with publication of further in-vitro data.

Protein-protein docking is computational method that helps in understanding of protein functions and molecular characteristics by filling the gap in existing knowledge about protein-protein interactions (Xue et al., 2015). Protein-protein docking software can be used to study the dynamics and flexibility of protein complexes, which can help explain how proteins interact and function. Such software include HADDOCK (Geng et al., 2017), RosettaDock (Lyskov and Gray, 2008), PatchDock (Schneidman-Duhovny et al., 2005), and ZDOCK (Pierce et al., 2011), among others. Each software has its own strengths and weaknesses, and the choice of software depends on the specific needs of the user. For example, ADDOCK is particularly useful for studying large protein complexes, while RosettaDock is known for its accuracy in predicting the side-chain conformations of the interacting proteins. PatchDock is a fast and user-friendly software, but its accuracy is lower than that of other more complex software. We used ZDOCK for the Srx-Prx interaction, as it is a widely used software for protein-protein docking studies and it provides valuable insights to help identify potential drug targets. ZDOCK can also be used to predict the binding affinity and orientation of protein complexes, which can be used to guide experimental studies. Overall, protein-protein docking software can be a valuable tool for studying protein interactions, but the accuracy of the results should be interpreted with caution and experimental validation is often necessary to confirm the results. In our study, results of protein-protein docking software were further tested using purified recombinant proteins and explored in cultured human cells.

The main function of the Srx-Prx system is to protect host cells from oxidative damages. This property of the Srx-Prx system becomes harmful to host organism when it starts protecting the survival of tumor cells. As shown in published literature, the Srx-Prx system is altered in multiple types of cancer and they function as activators or enhancers of oncogenic signaling to promote cancer development. How individual isoform of Prx contributes to different signaling pathways remains elusive. Further understanding of the Srx-Prx interaction can help in designing good targeting strategies against cancer development. Previous structural details of the Srx-Prx interaction provides some insights that the YF motif presents in C-terminal arm of Prx1 occludes the Srx-Prx interaction (Jonsson et al., 2008a). It is possible that the YF motif is responsible to hold C-terminal arm in particular orientation where it causes steric hindrance for the Srx-Prx interaction resulting in reduced rate of association. However, there is also another possibility that the C-terminal arm of Prx1 may help to stabilize the Srx-Prx complex. In this study, we utilized both computational prediction and experimental data to characterize the Srx-Prx interaction. Evidence from existing literature suggests the similarity in interaction of Srx with all members of typical 2-Cys Prx. In particular, the hydrophobic pocket of Srx binds to Prx homodimer in a region where peroxidatic cysteine and resolving cysteine of alternate monomer is located. Crystallography studies revealed that the hydrophobic pocket of Srx, consisting of residues Pro52, Leu82, Phe96, Val118, Val127 and Tyr128, mediates its interaction with Prx1 as well as other 2-Cys containing Prxs (Jonsson et al., 2008a; Jonsson et al., 2008b). This physical interaction interface is the same in our wildtype or c-terminus deletion mutants, but the access to the site is affected by the steric hindrance. Moreover, the convention in the field is that differences between typical 2-Cys Prx interacting with Srx come from their subcellular localization and not molecular characteristics (Noh et al., 2009). However, by virtue of being different proteins of same subfamily, they also have minor differences in their characteristics. Our computational prediction indicated that those minor differences in interaction can be due to different orientation of C-terminal arm of Prx. To confirm our prediction on steric hindrance and role of Prx C-terminal arm in the Srx-Prx interaction, recombinant proteins were purified and used in binding kinetic studies. Our data confirms this steric hindrance as deletion of C-terminal arm results in increased steady state interaction between Srx and Prx, which is further confirmed by kinetic studies that indicate that deletion mutation results in much faster association of Prx with Srx with very low effect on dissociation. In cultured cells, the differential interaction of Srx with individual Prx under normal and oxidative stress conditions were also observed. Taken together, this study adds insights to the molecular characteristics of the Srx-Prx interaction and may help in designing future targeting strategies for the inhibition of the Srx-Prx interaction.

  • *

    Understand the structural and molecular biology of the Srx-Prx interaction is of significance for prediction of molecular target site for the novel drug-discovery process.

  • *

    Homology modeling and protein-protein docking approaches predicts that the C-terminus of Prx1 causes a steric hindrance the interaction.

  • *

    Differential interaction of Srx with individual members of the Prx family was confirmed in cultured cells.

  • *

    Our data add novel molecular and structural insights critical for the understanding of the biology of the Srx-Prx interaction that may be of value for the development of targeted therapy for human disorders.

5. Acknowledgement

This work was supported by the National Institutes of Health (NCI R01CA222596), Department of Defense (W81XWH-16-1-0203), American Cancer Society (RSG-16-213-01-TBE) and the Kentucky Lung Cancer Research Program (KLCRP2016). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or other funding agencies.

Abbreviations

CP

Peroxidatic Cysteine (N-terminal Cysteine)

CR

Resolving Cysteine (C-terminal Cysteine)

GLC

General Ligand coupling chip with Compact capacity

I-TASSER

Iterative Threading ASSEmbly Refinement

Prx(s)

Peroxiredoxin(s)

SPR

Surface Plasmon Resonance

Srx

Sulfiredoxin

ZDOCK

Zhiping (name of principal investigator) DOCK

Footnotes

Conflict of Interests Statement.

None.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  1. Arnold K, Bordoli L, Kopp J, Schwede T, 2006. The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics 22, 195–201. [DOI] [PubMed] [Google Scholar]
  2. Biteau B, Labarre J, Toledano MB, 2003. ATP-dependent reduction of cysteine-sulphinic acid by S. cerevisiae sulphiredoxin. Nature 425, 980–984. [DOI] [PubMed] [Google Scholar]
  3. Chen L, Na R, Gu M, Salmon AB, Liu Y, Liang H, Qi W, Van Remmen H, Richardson A, Ran Q, 2008. Reduction of mitochondrial H2O2 by overexpressing peroxiredoxin 3 improves glucose tolerance in mice. Aging cell 7, 866–878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Findlay VJ, Tapiero H, Townsend DM, 2005. Sulfiredoxin: a potential therapeutic agent? Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie 59, 374–379. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Findlay VJ, Townsend DM, Morris TE, Fraser JP, He L, Tew KD, 2006. A novel role for human sulfiredoxin in the reversal of glutathionylation. Cancer research 66, 6800–6806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Forshaw TE, Reisz JA, Nelson KJ, Gumpena R, Lawson JR, Jonsson TJ, Wu H, Clodfelter JE, Johnson LC, Furdui CM, Lowther WT, 2021. Specificity of Human Sulfiredoxin for Reductant and Peroxiredoxin Oligomeric State. Antioxidants (Basel) 10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Gao XH, Bedhomme M, Veyel D, Zaffagnini M, Lemaire SD, 2009. Methods for analysis of protein glutathionylation and their application to photosynthetic organisms. Mol Plant 2, 218–235. [DOI] [PubMed] [Google Scholar]
  8. Geng C, Narasimhan S, Rodrigues JP, Bonvin AM, 2017. Information-Driven, Ensemble Flexible Peptide Docking Using HADDOCK. Methods Mol Biol 1561, 109–138. [DOI] [PubMed] [Google Scholar]
  9. Hu X, Weng Z, Chu CT, Zhang L, Cao G, Gao Y, Signore A, Zhu J, Hastings T, Greenamyre JT, Chen J, 2011. Peroxiredoxin-2 protects against 6-hydroxydopamine-induced dopaminergic neurodegeneration via attenuation of the apoptosis signal-regulating kinase (ASK1) signaling cascade. The Journal of neuroscience : the official journal of the Society for Neuroscience 31, 247–261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Huber W, Mueller F, 2006. Biomolecular interaction analysis in drug discovery using surface plasmon resonance technology. Curr Pharm Des 12, 3999–4021. [DOI] [PubMed] [Google Scholar]
  11. Iuchi Y, Okada F, Tsunoda S, Kibe N, Shirasawa N, Ikawa M, Okabe M, Ikeda Y, Fujii J, 2009. Peroxiredoxin 4 knockout results in elevated spermatogenic cell death via oxidative stress. Biochem J 419, 149–158. [DOI] [PubMed] [Google Scholar]
  12. Jiang H, Wu L, Chen J, Mishra M, Chawsheen HA, Zhu H, Wei Q, 2015. Sulfiredoxin Promotes Colorectal Cancer Cell Invasion and Metastasis through a Novel Mechanism of Enhancing EGFR Signaling. Mol Cancer Res. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jiang H, Wu L, Mishra M, Chawsheen HA, Wei Q, 2014. Expression of peroxiredoxin 1 and 4 promotes human lung cancer malignancy. American journal of cancer research 4, 445–460. [PMC free article] [PubMed] [Google Scholar]
  14. Jonsson TJ, Johnson LC, Lowther WT, 2008a. Structure of the sulphiredoxin-peroxiredoxin complex reveals an essential repair embrace. Nature 451, 98–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jonsson TJ, Johnson LC, Lowther WT, 2009. Protein engineering of the quaternary sulfiredoxin.peroxiredoxin enzyme.substrate complex reveals the molecular basis for cysteine sulfinic acid phosphorylation. The Journal of biological chemistry 284, 33305–33310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jonsson TJ, Murray MS, Johnson LC, Lowther WT, 2008b. Reduction of cysteine sulfinic acid in peroxiredoxin by sulfiredoxin proceeds directly through a sulfinic phosphoryl ester intermediate. The Journal of biological chemistry 283, 23846–23851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A, Bridgland A, Meyer C, Kohl SAA, Ballard AJ, Cowie A, Romera-Paredes B, Nikolov S, Jain R, Adler J, Back T, Petersen S, Reiman D, Clancy E, Zielinski M, Steinegger M, Pacholska M, Berghammer T, Bodenstein S, Silver D, Vinyals O, Senior AW, Kavukcuoglu K, Kohli P, Hassabis D, 2021. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kelley LA, Sternberg MJ, 2009. Protein structure prediction on the Web: a case study using the Phyre server. Nat Protoc 4, 363–371. [DOI] [PubMed] [Google Scholar]
  19. Khan FI, Wei DQ, Gu KR, Hassan MI, Tabrez S, 2015. Current updates on computer aided protein modeling and designing. Int J Biol Macromol 85, 48–62. [DOI] [PubMed] [Google Scholar]
  20. Kim TH, Song J, Alcantara Llaguno SR, Murnan E, Liyanarachchi S, Palanichamy K, Yi JY, Viapiano MS, Nakano I, Yoon SO, Wu H, Parada LF, Kwon CH, 2012. Suppression of peroxiredoxin 4 in glioblastoma cells increases apoptosis and reduces tumor growth. PloS one 7, e42818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kirubakaran P, Karthikeyan M, Singh Kh D, Nagamani S, Premkumar K, 2013. In silico structural and functional analysis of the human TOPK protein by structure modeling and molecular dynamics studies. J Mol Model 19, 407–419. [DOI] [PubMed] [Google Scholar]
  22. Lee TH, Kim SU, Yu SL, Kim SH, Park DS, Moon HB, Dho SH, Kwon KS, Kwon HJ, Han YH, Jeong S, Kang SW, Shin HS, Lee KK, Rhee SG, Yu DY, 2003. Peroxiredoxin II is essential for sustaining life span of erythrocytes in mice. Blood 101, 5033–5038. [DOI] [PubMed] [Google Scholar]
  23. Li L, Shoji W, Takano H, Nishimura N, Aoki Y, Takahashi R, Goto S, Kaifu T, Takai T, Obinata M, 2007. Increased susceptibility of MER5 (peroxiredoxin III) knockout mice to LPS-induced oxidative stress. Biochem Biophys Res Commun 355, 715–721. [DOI] [PubMed] [Google Scholar]
  24. Lowther WT, Haynes AC, 2011. Reduction of cysteine sulfinic acid in eukaryotic, typical 2-Cys peroxiredoxins by sulfiredoxin. Antioxid Redox Signal 15, 99–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lyskov S, Gray JJ, 2008. The RosettaDock server for local protein-protein docking. Nucleic Acids Res 36, W233–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Matthews CP, Birkholz AM, Baker AR, Perella CM, Beck GR Jr., Young MR, Colburn NH, 2007. Dominant-negative activator protein 1 (TAM67) targets cyclooxygenase-2 and osteopontin under conditions in which it specifically inhibits tumorigenesis. Cancer research 67, 2430–2438. [DOI] [PubMed] [Google Scholar]
  27. Neumann CA, Krause DS, Carman CV, Das S, Dubey DP, Abraham JL, Bronson RT, Fujiwara Y, Orkin SH, Van Etten RA, 2003. Essential role for the peroxiredoxin Prdx1 in erythrocyte antioxidant defence and tumour suppression. Nature 424, 561–565. [DOI] [PubMed] [Google Scholar]
  28. Noh YH, Baek JY, Jeong W, Rhee SG, Chang TS, 2009. Sulfiredoxin Translocation into Mitochondria Plays a Crucial Role in Reducing Hyperoxidized Peroxiredoxin III. The Journal of biological chemistry 284, 8470–8477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Park H, Lee H, Seok C, 2015. High-resolution protein-protein docking by global optimization: recent advances and future challenges. Curr Opin Struct Biol 35, 24–31. [DOI] [PubMed] [Google Scholar]
  30. Park JW, Mieyal JJ, Rhee SG, Chock PB, 2009. Deglutathionylation of 2-Cys peroxiredoxin is specifically catalyzed by sulfiredoxin. The Journal of biological chemistry 284, 23364–23374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pierce B, Tong W, Weng Z, 2005. M-ZDOCK: a grid-based approach for Cn symmetric multimer docking. Bioinformatics 21, 1472–1478. [DOI] [PubMed] [Google Scholar]
  32. Pierce BG, Hourai Y, Weng Z, 2011. Accelerating protein docking in ZDOCK using an advanced 3D convolution library. PloS one 6, e24657. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ring CS, Cohen FE, 1993. Modeling protein structures: construction and their applications. FASEB J 7, 783–790. [DOI] [PubMed] [Google Scholar]
  34. Roy A, Kucukural A, Zhang Y, 2010. I-TASSER: a unified platform for automated protein structure and function prediction. Nat Protoc 5, 725–738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Schneidman-Duhovny D, Inbar Y, Nussinov R, Wolfson HJ, 2005. PatchDock and SymmDock: servers for rigid and symmetric docking. Nucleic Acids Res 33, W363–367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Trujillo K, Paoletta S, Kiselev E, Jacobson KA, 2015. Molecular modeling of the human P2Y14 receptor: A template for structure-based design of selective agonist ligands. Bioorg Med Chem 23, 4056–4064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Ummanni R, Barreto F, Venz S, Scharf C, Barett C, Mannsperger HA, Brase JC, Kuner R, Schlomm T, Sauter G, Sultmann H, Korf U, Bokemeyer C, Walther R, Brummendorf TH, Balabanov S, 2012. Peroxiredoxins 3 and 4 are overexpressed in prostate cancer tissue and affect the proliferation of prostate cancer cells in vitro. Journal of proteome research 11, 2452–2466. [DOI] [PubMed] [Google Scholar]
  38. Varadi M, Anyango S, Deshpande M, Nair S, Natassia C, Yordanova G, Yuan D, Stroe O, Wood G, Laydon A, Zidek A, Green T, Tunyasuvunakool K, Petersen S, Jumper J, Clancy E, Green R, Vora A, Lutfi M, Figurnov M, Cowie A, Hobbs N, Kohli P, Kleywegt G, Birney E, Hassabis D, Velankar S, 2022. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res 50, D439–D444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wei Q, Jiang H, Baker A, Dodge LK, Gerard M, Young MR, Toledano MB, Colburn NH, 2013. Loss of sulfiredoxin renders mice resistant to azoxymethane/dextran sulfate sodium-induced colon carcinogenesis. Carcinogenesis 34, 1403–1410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Wei Q, Jiang H, Xiao Z, Baker A, Young MR, Veenstra TD, Colburn NH, 2011. Sulfiredoxin-Peroxiredoxin IV axis promotes human lung cancer progression through modulation of specific phosphokinase signaling. Proc Natl Acad Sci U S A 108, 7004–7009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wu L, Jiang H, Chawsheen HA, Mishra M, Young MR, Gerard M, Toledano MB, Colburn NH, Wei Q, 2014. Tumor promoter-induced sulfiredoxin is required for mouse skin tumorigenesis. Carcinogenesis 35, 1177–1184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Xue LC, Dobbs D, Bonvin AM, Honavar V, 2015. Computational prediction of protein interfaces: A review of data driven methods. FEBS Lett 589, 3516–3526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Yoshida Y, Yoshikawa A, Kinumi T, Ogawa Y, Saito Y, Ohara K, Yamamoto H, Imai Y, Niki E, 2009. Hydroxyoctadecadienoic acid and oxidatively modified peroxiredoxins in the blood of Alzheimer’s disease patients and their potential as biomarkers. Neurobiology of aging 30, 174–185. [DOI] [PubMed] [Google Scholar]
  44. Zeldich E, Chen CD, Colvin TA, Bove-Fenderson EA, Liang J, Tucker Zhou TB, Harris DA, Abraham CR, 2014. The neuroprotective effect of Klotho is mediated via regulation of members of the redox system. The Journal of biological chemistry 289, 24700–24715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zhang Y, 2008. I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 9, 40. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES