Discovery of novel inhibition site centered on 114-bit tryptophan of Thioredoxin reductase 1 through computer-aided drug design

Haoyi Yang; Hengyu Wang; Jia Feng; Jianmin Liao; Yuanyuan Lu

doi:10.1016/j.arabjc.2023.104642

View/Download PDF

Buy Reprints

PDF

Translate this page into:

Original article

04 2023

:16;

104642

doi:

10.1016/j.arabjc.2023.104642

Discovery of novel inhibition site centered on 114-bit tryptophan of Thioredoxin reductase 1 through computer-aided drug design

Haoyi Yang, Hengyu Wang, Jia Feng, Jianmin Liao^⁎, Yuanyuan Lu^⁎

China Pharmaceutical University, School of Life Sci & Technol, Nanjing 210009, China

⁎Corresponding authors at: Department of Marine Pharmacy, China Pharmaceutical University, 24 Tong Jia Xiang, Nanjing 210009, China. liaojianmin_cn@163.com (Jianmin Liao), luyy@cpu.edu.cn (Yuanyuan Lu)

Received: 2022-9-19, Accepted: 2023-1-29,

Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

Peer review under responsibility of King Saud University.

Abstract

Thioredoxin reductase 1 (TrxR1) is an oxidoreductase playing the important role in the tumor cells. It is a new type of drug therapy target. Most of the existing TrxR1 inhibitors act directly covalently on the active sites. Herein, molecular docking-based virtual screening approach was used to screen inhibitors with new binding site of TrxR1 from the SPECS database. After experimental test, compound 22 was identified as the reversibility inhibitor of TrxR1 U498C mutant (It has similar structure and function to replace the wild-type TrxR1 which is difficult to express) with IC₅₀ value of 15.31 ± 0.57 μM. The molecular docking results showed that the interaction between compound 22 and TrxR1 was centered on inactive site Trp¹¹⁴. Furthermore, phenazine compounds 24–30 with similar structures as 22 were also screened out from our phenazine database. Compounds 24–27 had longer chain structures and better inhibitory activity than compound 22, while compounds 28–30 were the opposite. Compounds 24–27 can be more stably bound in the protein cavity on Trp¹¹⁴ than compounds 28–30. Then we verified amino acids centered on Trp¹¹⁴ can regulate TrxR1 activity by amino acids mutation. Taken together, A new inhibition site are found that can regulate TrxR1 U498C mutant activity by acting on amino acids sequence at inactive sites centered on Trp¹¹⁴ and can provide ideas for the discovery and research of new TrxR1 inhibitors.

Keywords

Thioredoxin reductase 1

Computer-aided drug design

Virtual screening

Molecular docking

Trp114

Show Related Articles from PubMed

1 Introduction

Thioredoxin reductase 1 (TrxR1) is an NADPH-dependent selenocysteine-containing protein with a FAD domain and belongs to the pyridine nucleotide disulfide oxidoreductase family (Mercatelli, 2017). When it works, electrons are transferred from FAD to the N-terminal active site (-Cys⁵⁹-Val-Asn-Val-Gly-Cys⁶⁴-), then through the C-terminal active site (-Gly-Cys⁴⁹⁷-Sec⁴⁹⁸-Gly), finally arrive at Trx1 (Zhang, 2022). In vivo, TrxR1 is the key point of the redox system, relates with various diseases (Dong, 2016). In tumor cells, the highly expressed TrxR1/Trx1 system is one of the basic redox regulation and antioxidant defense systems against reactive oxygen species (ROS) which is obtained by rapid proliferation and vigorous metabolism in a variety of tumor cells (Tamura and Stadtman, 1996).

From the recent study, TrxR1 inhibitors which have been found or under development have many shortcomings. For example, heavy metal-containing TrxR1 inhibitors, such as Cisplatin and Auranofin, have poor targeting due to their irreversible covalent binding, causing many adverse effects such as gastrointestinal adverse reactions and nephrotoxicity (Roder and M.j., 2015). Natural products such as Curcumin also have the disadvantages of rapid degradation, low bioavailability and poor metabolism in vivo (Ste Marie, 2020). As far as we know, there are most of the inhibitors act irreversible and directly at active sites of TrxR1 reported previously. So we hope to find a new binding site and a new binding mode.

Computer-aided drug design (CADD) offers opportunities for the rapid and effective development of innovative drugs, including virtual screening, molecular docking, and molecular dynamics simulation (Yu, 2022). At present, a variety of drugs have been developed using this technology, such as Donepezil for the treatment of early Alzheimer's disease (Pandey and Singh, 2020), anti-HIV infection drug Indinavir, and anti-platelet drug Tirofiban (Dar, 2018). We expected to find new TrxR1 inhibitors by combining CADD techniques with previously discovered areas of action.

In this work, we used CADD technology to gain more binding mode between inhibitor structures and enzyme. The different compounds were found from the SPECS database to do preliminary test. DTNB method was used to measure the enzyme activity after adding inhibitors to evaluate the inhibitory activity. But the expression system we usually used cannot add selenium to the protein correctly. Therefore, in this paper, we used TrxR1 U498C mutant with similar structure and function to the wild-type as a substitute (Cheng, 2010). At the same time, we selected four cells with high expression of TrxR1 to preliminarily test the cytotoxicity. To explore whether inhibitor and enzyme was reversible binding, we used ultrafiltration desalination techniques to test the strength of the compound's binding to the enzyme. The macromolecular protein could be retained in the ultrafiltration membrane, while the loosely bound small molecule inhibitor would be eluted from the enzyme. After knowing that inhibitors could inhibit enzyme activity, we wanted to know the binding amino acids of inhibitors bound to enzyme. Autodock Vina was used for analysis to infer possible sites of action. The binding amino acids we analyzed before was verified by amino acids mutation. TrxR1 E106A/N107A/G110A/S111A/W114A/ G115A/Y116A/V118A/U498C mutant (AAAAAAAAC or E106A/N107A/G110A/ S111A/W114A/G115A/Y116A/V118A/U498C) was designed and produced for follow-up test. Combining with the previous experimental results, we found that the binding mode of the compound to the protein has the following characteristics: 1. Reversible binding, 2. Binding produces an inhibitory effect at the inactive site. We speculate that the binding of the compound to the enzyme is noncompetitive inhibition. In summary, this work attempts to provide a new inhibitor binding site for TrxR1, which can provide ideas for the development of new TrxR1 inhibitors in the future.

2 Results

2.1

2.1 Virtual screening based on molecular docking

Previous research found that the compound 2, 2′: 6′, 2′'-terpyridine platinum (II) chloride can inhibit the activity of TrxR1. Through a series of studies, the binding region between 2, 2′: 6′, 2′'-terpyridine platinum (II) chloride and TrxR1 was obtained (PDB library: 2zzb) (Lo, 2009). But we don’t know the binding amino acid sequence or amino acid pockets with similar positions.

After that, we used this binding region to do the molecular docking-based virtual screening. 306,709 compounds in the SPECS database were docked into the binding region of TrxR1 using Autodock Vina. According to the results of molecular docking, we performed cluster analysis of compounds with an affinity of less than −8.2 kcal/mol to ensure the diversity of skeletons, the top-ranked 23 compounds with low docking energy (Table 1 compounds information, Fig. 1 structures of compounds) were purchased for further in-vitro experiment.

Table 1 Basic information and preliminary docking scores of top-ranked 23 compounds.

NO.	ZINC ID	SPECS ID	Affinity (kcal/mol)
1	ZINC08398361	AP-406/41885709	−9.2
2	ZINC08454559	AK-968/41024260	−9.1
3	ZINC08443595	AK-968/12975147	−8.9
4	ZINC02070190	AG-690/36877026	−8.8
5	ZINC08453935	AH-262/36520008	−8.7
6	ZINC08397961	AN-329/40221887	−8.5
7	ZINC00990140	AK-820/13220140	−8.5
8	ZINC08455074	AK-968/41171962	−8.4
9	ZINC08383903	AB-131/42301036	−8.4
10	ZINC17130694	AK-968/15360546	−8.4
11	ZINC02071782	AG-690/40139005	−8.4
12	ZINC00712263	AG-205/33685064	−8.3
13	ZINC08436819	AN-988/14419107	−8.3
14	ZINC08424380	AG-690/11384631	−8.3
15	ZINC02094416	AN-655/14573038	−8.3
16	ZINC00990152	AK-820/13220165	−8.3
17	ZINC08441699	AF-399/42017518	−8.3
18	ZINC08443716	AK-968/13029089	−8.3
19	ZINC02052499	AG-670/41783010	−8.3
20	ZINC02750193	AK-968/14002216	−8.3
21	ZINC12779070	AK-968/41017609	−8.3
22	ZINC08439364	AQ-390/43238273	−8.2
23	ZINC08441713	AF-399/42017530	−8.2

Structural formulas of top-ranked 23 virtual screened compounds.

2.2

2.2 Inhibitory activity of virtual screening compounds

In order to test the inhibitory activity of the virtual screening compounds against TrxR1, we used DTNB method to determine the inhibitory effects of the top-ranked 23 compounds. Due to the inefficient expression of selenoprotein in E. coli, and the Autodock molecular docking software could not recognize selenium. To match the CADD figures with the experimental figures, cysteine (Cys) with a similar structure and function to selenocysteine (Sec) was used as the substitute. Then we used Sec to Cys mutant TrxR1 (GCCG or U498C) to go on the next research. In previous studies, TrxR1 U498C mutant and wild-type TrxR1 had also been proved to have similar structures and reduction functions (Cheng, 2010). The experiment results we get by TrxR1 U498C mutant could solve the problems in the search for new TrxR1 inhibitors and could be used as a reference for wild-type TrxR1.

First, preliminary screening was performed to estimate the activity of TrxR1 U498C mutant under 5 μM inhibitors (Fig. 2a) and the inhibitors’ cytotoxicity with four TrxR1 high-expressing cancer cell lines (Table S1). As shown in Fig. 2a, compound 22 showed more inhibitory activity than other inhibitors. After that, we mainly studied and discussed compound 22. By setting different inhibitor concentrations, a series of inhibitory rates under the corresponding inhibitory concentrations were obtained. Through nonlinear fitting, the IC₅₀ of compound 22 for TrxR1 U498C mutant was calculated to be 15.31 ± 0.57 μM (Fig. 2b), which had certain inhibitory effect.

Inhibitory activity and molecular docking results of compound 22. (a) Preliminary enzyme inhibition results of top-ranked 23 virtual screening compounds. 5 μM inhibitors are incubated with TrxR1 U498C mutant at 37 °C for 30 min, as determined by the DTNB method. (b) Compound 22 IC50 determination of TrxR1 U498C mutant. After the inhibitor and enzyme (0.5 mg/mL) are incubated at 37 °C for 30 min, the absorbance is measured by the DTNB method. These data represent the mean ± SD (n = 3). (c) 3D conformation of compound 22 interacting with TrxR1 (PDB id: 2zz0). (d) 2D conformation of compound 22 interacting with TrxR1 (PDB id: 2zz0). The purple dotted line indicates Pi-Pi interaction, the yellowish-brown dotted line indicates Pi-Sulfur interaction, the light blue dotted line indicates Carbon Hydrogen interaction, and the light pink dotted line indicates Alkyl interaction. (e) The spatial conformation of TrxR1 (PDB id: 2zz0). (f) The spatial conformation of compound 22 binds to TrxR1 (PDB id: 2zz0). The green conformation is the C-terminal active site dominated of TrxR1 by Cys497 and Cys498, the purple conformation is the Trp114, the blue conformation is the N-terminal active site (-Cys59-Val-Asn-Val-Gly-Cys64-) of TrxR1, the yellow conformation is compound 22, and the white arrow refers to the 52–54 (–Arg52 -Trp53 - Gly54-) amino acids pocket.

Now we knew that the compound 22 had inhibitory activity, then we wanted to know the inhibitory activity was reversible or not. TrxR1 irreversible inhibitors, such as Auranofin or other heavy metal-containing compounds, showed excellent inhibitory activity at the enzyme level (IC₅₀ < 0.4 μM). But they had poor selectivity caused many strong toxic side effects. Although reversible protein inhibitors were not as effective as irreversible inhibitors at the enzyme level, they had stronger targeting and lower toxic side effects after entering the human body (Kiely-Collins, 2021).

We found that when the final concentration of compound 22 was 25 μM, it could effectively inhibit the activity of TrxR1 U498C mutant (inhibition rate > 50 %). To test the binding strength between the compound and the enzyme, we used the ultrafiltration tube to remove the drug from the enzyme drug mixture (Lamport and Várnai, 2013). The enzyme with large molecular weight (55 kDa) could be intercepted in the ultrafiltration tube (the pore of ultrafiltration membrane can intercept the molecular weight of 30 kDa) by the filter membrane, while the inhibitor with small molecular weight could pass through the filter membrane. If the binding ability between enzyme and inhibitor was weak, the enzyme and inhibitor would be separated after ultrafiltration and the inhibition effect would reduce or disappear. In our experiment, the mixture of enzyme and inhibitor was desalted by ultrafiltration tube, and the color of the compound disappeared obviously. When the enzyme activity was detected by DTNB method, there was no significant difference in absorbance between the treatment group and the control group (Fig. S1a).

In order to more visually show the change of inhibitory effect of compound 22 on TrxR1 U498C mutant after ultrafiltration, we converted the change of absorbance into the value of enzyme activity then used histograms to present enzyme activity data. We found that after ultrafiltration, the enzyme activity values of the treatment group and the control group were close and showed no significant difference (Fig. S1b), which was consistent with the previous experimental results of absorbance change. So far, it could be judged that the inhibitor is reversible inhibition, which was also consistent with our previous speculation. The binding position of compound 22 with the enzyme was located on the enzyme surface and did not enter the conformation center. This led external changes can easily affect the binding and separation of compound 22 and enzyme, resulting in the shedding of compounds from the protein.

Next, by analyzing the structure of compound 22 and its interaction amino acids with enzyme, we wanted to obtain more clues for finding new TrxR1 inhibitors with similar structures.

2.3

2.3 Molecular docking found the binding amino acids of compounds and enzyme

In previous studies, we found that compound 22 has inhibitory activity, but the binding amino acids with enzyme were not clear. In order to observe the interaction between compound 22 and the enzyme, we used Autodock Vina software to analyze it. From the results of molecular docking (Fig. 2c and Fig. 2d), it could be observed that the main tetracyclic structure of compound 22 interacted with Trp¹¹⁴ through Pi-Pi bond and could interact with Cys⁴⁹⁷. The tetracyclic structure was on the upper side of the Trp¹¹⁴, and the remaining pentacyclic phenyl structure could extend downward along the Trp¹¹⁴.

Here, we had two conjectures about the inhibition principle of the compound 22. The first was that compound 22 could anchor the inactive sites centered on Trp¹¹⁴ and capture Cys⁴⁹⁷, made the C- terminal active site cannot swing back and forth, thus prevented electrons at the N-terminal active site from transferring to Trx1. Made Trx1 unable to be reduced, caused redox imbalance.

The second was that compound 22 anchored Trp¹¹⁴, and the pentacyclic phenyl structure would bend down to the amino acids pocket of position 52–54 (–Arg⁵²-Trp⁵³-Gly⁵⁴-, the position indicated by the white arrow in Fig. 2e) block the contact between the C-terminal active site (-Gly-Cys⁴⁹⁷-Sec⁴⁹⁸-Gly) and the N-terminal active site (-Cys⁵⁹-Val-Asn-Val-Gly-Cys⁶⁴-) (Fig. 2e and Fig. 2f, the green one is the C-terminal active site which transfer electrons to Trx1, and the blue one is the N-terminal active site which accept electrons from FAD, and the purple one is the Trp¹¹⁴). When compound 22 combined with Trp¹¹⁴, it could block the contact between the C-terminal active site (green conformation) and the N-terminal active site (blue conformation), thus it could block the transmission of electrons between the two active sites, destroyed the redox balance and caused cell death.

The binding mode of compound 22 was obtained by combining Autodock vina, its tetracyclic and pentacyclic phenyl structure had attracted our attention. Then, compound 22 was divided into two parts according to its structure (Fig. 1). Results of reference molecular docking, the different binding amino acids between different parts of inhibitors were analyzed. The tetracyclic structure was called the first part, and the pentacyclic phenyl was called the second part. We found that the first part mainly interacted with Cys⁴⁹⁷ and Trp¹¹⁴, and the second part mainly interacted with Gln¹⁰⁶ and Trp¹¹⁴ and bended down to the amino acids pocket at positions 52–54. We speculated that the first inhibition principle was related to the structure of part 1, and the second inhibition principle was related to the structure of part 2. We found both inhibition principles require the compound could bind to the position of Trp¹¹⁴. Therefore, we thought Trp¹¹⁴ was very important, which was the center of multiple binding amino acids.

Due to the weak activity of compound 22, we wanted to find compounds with better activity. We speculated that phenazine compounds with similar skeletal structures would also have the similar inhibitory effect on TrxR1.

A series of compounds were screened from our phenazine database. Compounds 24–27 (Fig. 3a, Compounds details can be found in (Lu, 2017) had similar structures to compound 22 and had a longer chain of structures. We measured their inhibitory activity (Fig. 3b-e and Table 2). It could be seen that the newly screened compounds had better TrxR1 U498C mutant inhibitory activity. Compound 27 had the best inhibitory effect, which was close to the listed TrxR1 inhibitor Auranofin. Then we wanted to find the binding amino acids of compounds 24–27 to analyze how they combined. After the molecular docking of these four compounds (Fig. 4a for binding mode and Table 3 for docking energy), it could be seen that these four compounds interacted with Trp¹¹⁴, and the pentacyclic phenyl structure could enter the amino acids pocket at positions 52–54. The interactions between the compounds and enzyme showed that Trp¹¹⁴, Gly¹¹⁰, Gln¹⁰⁶, Asn¹⁰⁷, and Ser¹¹¹ were key amino acids and Trp¹¹⁴ was still the main binding center.

Compounds 24–27 structures and compounds 24–27 IC50 curve. (a) Structures of compounds 24–27. (b-e) Compounds 24–27 inhibit TrxR1 activity. 0.25 mg/mL TrxR1 U498C mutant is incubated with inhibitors of different concentrations for 30 min, and the absorbance is measured at 415 nm.

Table 2 IC₅₀ data of compounds 24–27 for inhibiting TrxR1. 0.25 mg/mL TrxR1 U498C mutant is incubated with inhibitors of different concentrations for 30 min, and the absorbance is measured at 415 nm.

NO.	IC₅₀（μM）
24	4.373 ± 0.055
25	1.283 ± 0.211
26	1.714 ± 0.128
27	0.910 ± 0.029

These data represent the mean ± SD (n = 3).

Molecular docking data and amino acids preference of compounds 24–27. (a) Molecular docking data of compounds 24–27. The first column shows the 3D conformation of the compounds in TrxR1 (PDB id: 2zz0), and the second column shows the 2D display of the amino acids of the compounds interacting with TrxR1 (PDB id:2zz0). The green dotted line indicates Hydrogen interaction, and the purple dotted line indicates Pi-Pi interaction. The light green dotted line represents Carbon Hydrogen interaction, the yellowish-brown dotted line represents Pi-Sulfur interaction, and the pink dotted line represents Pi-Alkyl interaction. (b) Amino acids preference of different inhibitors. The probability is represented by a pie chart.

Table 3 Docking energy of compounds 24–30.

NO.	Affinity (kcal/mol)
24	−7.4
25	−7.6
26	−8.2
27 28 29 30	−7.6 −5.7 −5.6 −5.2

This was a very interesting result. It was demonstrated that compounds 24–27 had similar binding amino acids to compound 22 and could block electron transfer between the C-terminal active site and the N-terminal active site centered on Trp¹¹⁴. Compared with compound 22, compounds 24–27 had longer chain structures, so they could go deeper into the amino acids pocket at positions 52–54. It was consistent with the result that the inhibitory effect was stronger in the determination of enzyme activity. It also proved that compounds could inhibit TrxR1 by steric hindrance.

2.4

2.4 Verification of inhibitors binding site

In previous experiments, we analyzed the possible binding amino acids of compounds with TrxR1 through molecular docking software. We found that the binding amino acids were centered on Trp¹¹⁴. In order to verify the binding amino acids of the compounds, combined the molecular docking results, the TrxR1 E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A/U498C mutant (AAAAAAAAC or E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A/U498C) was designed and produced. When the amino acids were mutated into alanine with simpler spatial structure, the interaction between amino acids and the inhibitors would be destroyed. If the inhibitory effect of the inhibitors disappeared or weakened, it was proved that the mutated amino acids were the main binding sites. As shown in Fig. 5a, the enzyme activity of TrxR1 E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A/U498C mutant added compound 22 and compounds 24–27. After mutation of Trp¹¹⁴-centered amino acids, the inhibitory effect of compounds decreased or even disappeared comparing with TrxR1 U498C mutant. The result proved that the amino acids centered on Trp¹¹⁴ were the key sites for the series of compounds, which led to the inhibition of TrxR1.

(a) The TrxR1 U498C mutant and E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A/U498C mutant (0.95 mg/mL) with 20 μM compound 22 and 2 μM compounds 24–27 after incubation for 30 min determination of inhibition effect. Analysis by t-test: compare with the control group, * * means the difference is very significant (p < 0.01), * means the difference is significant (p < 0.05), no label means the difference is not significant. (b) 5 μM compounds 28–30 enzyme activity data. (c) Structures of compounds 28–30. (d) Docking results and binding energy of compounds 28–30.

So far, we had found that the amino acids centered on Trp¹¹⁴ play an important role in the inhibitory activity. However, did the slight difference of the binding modes between compound 22 and compounds 24–27 lead the different inhibitory activity? Then we sorted out the results of molecular docking.

The amino acids preferences were used to answer the subtle differences of binding amino acids between different compounds. In the first three conformations of three times docking results, the more frequently amino acids appeared in the conformation of molecular docking, the more it preferred to which amino acid, and the easier bound to this amino acid.

We found that compound 22 and compounds 24–27 have different amino acids preferences. It could be seen from the results (Fig. 4b) that compound 22 had a stronger preference for Cys⁴⁹⁷ and Gln¹⁰⁶ than compounds 24–27, and compounds 24–27 had a stronger preference for Gly¹¹⁰ than compound 22. All of them had a strong preference for Trp¹¹⁴. It could be seen that the structural difference between compound 22 and compounds 24–27 obviously led to the difference in binding amino acids between compounds and enzyme, which might cause the difference in their inhibitory activities. The mutation of these amino acids would also affect the inhibitory activity of inhibitors in different degrees.

At the same time, we found that when the amino acids were mutated into alanine with smaller spatial structures, the enzyme activity increased significantly (Fig. 5a). This also demonstrated with the previous experimental results that the amino acids at position 106–118 were located between the C-terminal active site and the N-terminal active site, and played a role to regulate enzyme activity. When this sequence was mutated into amino acids with smaller spatial structures, it was equivalent to open the door to the N-terminal active site, making the contact between the C-terminal active site and the N-terminal active site more frequently, so the catalytic efficiency of the enzyme was more efficient. Therefore, we found that the activity of the TrxR1 enzyme can be regulated by interacting with the amino acids centered on Trp¹¹⁴.

Compared with compounds 24–27, we found that when chlorine atoms were added to the end of the compounds, its ability of inhibiting TrxR1 was improved (compared with the inhibitory activity of compounds 24 and 25, 26 and 27 in Table 2). We speculated that the chlorine atom had a larger spatial structure, which could more effectively prevent the approach between the two active sites. This was consistent with our speculation: electrons transfer between the C-terminal and the N-terminal active sites could be hindered by increasing steric hindrance.

5. Explored the effect of chain length of compounds on their activity.

To further verify that the inhibitors could produce inhibitory effect by increasing the steric hindrance centered on Trp¹¹⁴, we selected compounds 28–30 (Compounds details can be found in (Lu, 2017) with shorter chain structures from our phenazine database. By measuring the enzyme activity of compounds 28–30, compounds 28–30 in 5 μM showed no inhibitory activity (Fig. 5b and Table 4). In molecular docking, we found that compounds 28–30 have higher docking energy than compound 22 and compounds 24–27 (Fig. 5d and Table 3), in other words, the binding ability of compounds 28–30 to binding sites was less than compound 22 and compounds 24–27. By analyzing the molecular docking conformation of compounds 28–30, we found that compounds 28–30 could not go deep into the amino acids pocket of 52–54 and could not stably bind to the Cys⁴⁹⁷ because of its shorter chain structures and higher binding energy, so it could not produce a strong steric hindrance or stably capture the active site of Cys⁴⁹⁷, and thus could not show strong activity of inhibiting TrxR1.

Table 4 Enzyme activity data of TrxR1 after adding 5 μM compounds 28–30.

NO.	Enzyme avtivity(U/mg prot)
NC	56.81 ± 2.217
28	68.53 ± 0.308
29	60.54 ± 1.340
30	60.54 ± 1.627

These data represent the mean ± SD (n = 3).

3 Discussion

Tumor cells will accumulate excessive reactive oxygen species (ROS) due to their rapid proliferation and high metabolic rate. In order to resist the large number of ROS produced by metabolism, the antioxidant system will be activated. Therefore, it is feasible to inhibit the growth and development of tumor cells by targeting the antioxidant system (Cheung and Vousden, 2022; Sahoo, 2022; Asahina, 2022 ). The intracellular Trx1/TrxR1 system is an important node and component of the redox system. In previous studies, TrxR1 passes the electrons of NADPH to Trx1 through the approaching of the two active sites (Fritz-Wolf, 2011), and its’ expression increased in tumor cells have been widely reported (Zhang, 2021). TrxR1 is very important to maintain tumor phenotype and is considered to be a promising target for chemotherapy (Jovanović, 2020). In this study, we obtained the compounds which could bind to TrxR1 through virtual screening and found new sites where the inhibitors produced inhibition through molecular docking. By analyzing the structures of the inhibitors, compounds with higher activity were screened out, and the newly discovered active site was verified by amino acids mutation.

Because the efficiency of expressing selenoprotein in the conventional heterologous expression system is too low, we chose the TrxR1 U498C mutant as an alternative and proved in previous studies that it has similar structure and function to the wild type. Through preliminary inhibitory activity detection of 23 compounds, we selected compound 22 as the target molecule with better effect of inhibiting TrxR1. Through the determination of various data, we found that compound 22 could reversibly bind to TrxR1. After that, we used molecular docking technology to simulate the binding mode and binding amino acids between compound 22 and TrxR1. The tetracyclic structure of compound 22 bound to Trp¹¹⁴, while the pentacyclic phenyl structure bound to Trp¹¹⁴ and Gln¹⁰⁶ and then bent downward close to the 52–54 amino acids pocket near the N-terminal active site. We found that Trp¹¹⁴ always interacts with the tetracyclic structure in the multiple docking conformations so it was the central binding amino acids of the compound to the TrxR1. We had two speculations to explain its inhibitory activity. One was that compound 22 bound to amino acids centered on Trp¹¹⁴ then directly bound to Cys⁴⁹⁷, made the C-terminal active site cannot swing to get elements from N-terminal. The second was that after binding to Trp¹¹⁴ and surrounding amino acids, the compound blocked the contact between the two active sites through steric hindrance. All of the speculations are that the compound binds to the inactive site and inhibits the activity of the enzyme. Then the amino acids centered on Trp¹¹⁴ were presumed to be a cumulative role in the inhibitory effect.

In previous studies, TrxR1 inhibitors generally act directly on the N-terminal or C-terminal active sites, such as classical anticancer drug Auranofin (Pickering, 2020); natural active substances Shikonin (Zhang, 2022) and Curcumin. However, they did not indicate that whether the inactive sites close to the N-terminal or C-terminal active sites play the important role in the inhibitory effect. But in our study, the inhibitors we found can produce inhibition by binding to the inactive sites centered on Trp¹¹⁴, which had never been reported before.

Next, by analyzing the structure of the compound 22, we screened compounds 24–30 with similar structures from our phenazine database. All of them have a rigid multi-element ring plane is used as the subject structure of the compound, and the structure of the compound is extended through the linkage of the five membered nitrogen heterocycle or the benzene ring. Compared with compound 22, compounds 24–27 had longer chain structures (have more ring-shaped structures connected together) (structure in Fig. 3a). The compounds 24–27 were verified to have better inhibitory activity than compound 22 through experiments. Using molecular docking software, we found that they mainly bind to Trp¹¹⁴ by their rigid multi-element ring plane and interact with many surrounding amino acids. They could use their longer chain structures to go deep into the 52–54 sites to produce stronger steric hindrance effect than compound 22. Furthermore, compounds 28–30 (structure in Fig. 5c) with shorter chain structures did not show strong inhibition in the enzyme activity test (Table 4). We speculated that they could not go deep into the 52–54 amino acids pocket, resulting in smaller steric hindrance. The molecular docking results also showed that the binding force of compounds 28–30 to the enzyme were less than compounds 22 and 24–27 (Table 3), which may reduce the inhibitory effect.

To prove our speculation, we selected TrxR1 E106A/N107A/G110A/S111A/ W114A/G115A/Y116A/V118A/U498C mutant to do the experiment. These amino acids centered on Trp¹¹⁴ were found in molecular docking where compounds could interact with TrxR1. If the compounds bind to this region to produce inhibitory activity, the inhibitory activity would decrease or disappear when this segment of amino acids were mutated. Subsequent experiments also confirmed our speculation. After mutation of E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A amino acids, the inhibitory effect of compounds decreased or even disappeared comparing with TrxR1 U498C mutant. At the same time, this was the first time we knew the inhibitor could combine with this region to produce inhibitory effect by mutating the amino acids sequence centered on Trp¹¹⁴.

In the previous experiments, we found that compound 22 has two characteristics of reversible inhibition and binding to the inactive site center on Trp¹¹⁴ to prevent electron transfer between two active sites. We speculated that the binding mode of compound 22 to protein was similar like noncompetitive inhibition. Compounds 24–27 have similar chemical structure and binding mode with protein to compound 22, and we speculated that they are also noncompetitive inhibition. We are eager to prove this conclusion. However, due to the limited experimental conditions, the microplate reader fluctuates violently at low substrate concentration and cannot get scientific results. Here, in order not to make unnecessarily mislead for people who read this article, we can only speculate that it is noncompetitive inhibition based on the information obtained before, and we hope it can be confirmed later.

Subsequently, we also found that when the amino acids centered on Trp¹¹⁴ were mutated into the alanine with a smaller space structure, its catalytic activity increased significantly, which further verified the conclusion that our inhibitors could affect the enzyme activity by binding to this sequence and increasing steric hindrance (Fig. 5a). At the same time, we analyzed the amino acids preference of compounds 24–27 and compound 22. Result showed that the different binding amino acids caused by the different structures. It might explain why the inhibitory activity of compounds 24–27 were higher than compound 22.

However, combined with previous studies, we found that the activity of compounds at the cell level were not consistent with the enzyme level. In previous studies, we detected the sensitivity of cells against virtual screening compounds. Among the four cell lines, A549 cells were the most sensitive to compound 22 (Table S1). Then in another our published article we found compounds 24–27 showed similar results (A549 IC₅₀ values were 12.3 ± 0.74 μΜ; 11.8 ± 2.2 μM; 39.5 ± 4.6 μM; > 50 μM, the LO2 IC₅₀ values was > 50 μΜ, respectively) (Lu, 2017); we have the following conjecture: A549 cells belong to lung cancer cells. Compared with other tissues, the lung is in a high oxygen environment and is more vulnerable to oxidative stress damage (Zablocka-Slowinska, 2018). Therefore, in order to cope with the external environment, lung cancer cells have higher expression of antioxidant system and are more susceptible to external regulation. We also found some structures may show strong inhibitory activity at the enzyme level, they would show negative results at the cell level due to the presence of cell membrane or many interfering substances at the cell level, which was expected to be solved in the follow-up research.

Although in this article, we started from the virtual screened compound 22 to the phenazine compounds 24–27, we provided a new binding site of compounds and proteins. We believe that most of the compounds with the structure consistent with the above-mentioned binding mode can bind to this site to inhibit the activity of the enzyme, and are not limited to phenazine compounds.

Taken together, by using CADD technology and combining with experiments, the new binding sites centered on Trp¹¹⁴ were found to inhibit TrxR1. We look forward to providing a reference for the research and development of new TrxR1 inhibitors.

4 Materials and method

4.1

4.1 Virtual screening based on molecular docking

This subject selected SPECS database (306709 molecules) (https://www.specs.net) to do molecular docking-based virtual screening. Human thioredoxin reductase I (PDB id: 2zzb) was obtained from protein data bank (https://www.rcsb.org/). Protein TrxR1 used Autodocktools 1.5.6 (Sanner, 1999; Morris, 2009) to add polar hydrogen and charge, and finally converted to pdbqt format. The coordinates of protein TrxR1 active pocket were set to: center_ x = -20.734, center_ y = 0.046, center_ z = 0.028; size_ x = 15, size_ y = 15, size_ z = 15. In addition, the parameter num_modes were set to 1. Unless otherwise specified, other parameters adopt default values. Finally, after the results were split with openbabel software, the script was used to convert it into pdbqt format in batch, and then Autodock Vina 1.1.2 was used for batch molecular docking (Trott and Olson, 2010).

4.2

4.2 Produced heterologous TrxR1

The recombinant E. coli BL21 (pET28a⁽⁺⁾-TrxR1^U498C) had been constructed in our work. Other recombinant strains constructed in this study were based on this strain. And the inducible expression process of all the recombinant strains in this study was as below. For seed cultivation, 20 μL of glycerol stock strain was inoculated into 20 mL LB medium supplemented with 50 mg/L kanamycin. The seed culture was shaken at 37 °C for 16 h. Then, 2 mL seed culture was inoculated at 200 mL LB medium supplemented with 50 mg/L kanamycin. The resulting culture was shaken at 37 °C for 4 h at 220 rpm, when its optical density at 600 nm reached 0.6–0.8, isopropyl-β-d-Thiogalactoside (IPTG) was added to induce expression of the TrxR1. After that, the culture temperature was reduced to 16 °C and at 220 rpm for another 18 h.

4.3

4.3 TrxR1 activity assay (DTNB method)

The DTNB method was based on the absorbance of TNB at 415 nm to determine the enzyme activity (1.36 × 10⁷ mL/mol/cm). 140 μL 0.1 M PE buffer (77.4 mL of 1 M Na₂HPO₄ and 22.6 mL of 1 M NaH₂PO₄ were mixed to a fixed capacity of 1 L) were incubated at 37 °C, added 20 μL 5,5′-Dithiobis-(2-nitrobenzoic acid) (DTNB, Shanghai yuanye Bio-Technology Co., ltd) solution (12 mM) and 20 μL nicotinamide adenine dinucleotide phosphate (NADPH, Shanghai yuanye Bio-Technology Co., ltd) solution (2 mM) as substrate, added 20 μL heterologous expression of purified TrxR1 started the catalytic reaction. Recorded the absorbance change every minute after the start of the reaction, and set up a blank group without enzyme.

Definition of enzyme activity: at 37 °C, the catalytic consumption of 1 nmol of NADPH or the catalytic production of 2 nmol of TNB per milligram and per minute is a unit of enzyme activity.

TrxR1 (U/mg prot) = (Δ A _measuring-Δ A _blank) / (ε × d) × 10⁹ × V _total / (Cpr × V _sample) / T.

ε: Molar extinction coefficient of DTNB at 415 nm, 1.36 × 10⁷ mL/mol/cm; d: 0.6 cm; V _total: total reaction volume, 0.2 mL; Cpr: protein concentration of supernatant (mg/mL); V _Sample: volume of sample to be tested in the reaction, 20 μL; T: Reaction time, min.

4.4

4.4 Molecular docking verification binding site

TrxR1 (PDB id: 2zz0) added polar hydrogen and removed water molecules using Autodocktools 1.5.6, and finally converted to pdbqt format. The coordinates of protein TrxR1 active pocket were set to: center_ x = -20.734, center_ y = 0.046, center_ z = 0.028; size_ x = 15, size_ y = 15, size_ z = 15. Autodock Vina 1.1.2 was used for molecular docking. Docking results used the Discovery Studio Visualizer to analyze docking results.

4.5

4.5 Inhibitor binding site validation test

Took TrxR1 U498C and E106A/N107A/G110A/S111A/W114A/G115A/Y116A/V118A/U498C mutant, incubating with final concentration of 20 μM compound 22 and 2 μM compounds 24–27 (from our phenazine database) at 37 °C for 30 min, and the enzyme activity was detected by DTNB method. At the same time, the blank group without protein and the control group without inhibitor were established.

4.6

4.6 Statistical method for amino acid preference of inhibitors

The molecular docking was repeated three times, and the conformation with the lowest three binding energy among the molecular docking results were used for statistical analysis, and then the results were presented using the sector diagram.

5 Conclusion

Together, our work focused on the discovery of the new inhibitor binding site for TrxR1. Starting with virtual screening and going through a series of experiments, the compounds had been proved to inhibit TrxR1 activity by binding amino acids centered on Trp¹¹⁴ which had never been reported before. It indicates that this site can regulate the enzyme activity. Our findings provide a new idea for the TrxR1 inhibitors design and suggest they’re good application prospects in the future.

Acknowledgements

The study received financial support from National Nature Science Foundation of China (Grant No. 81872757). National Innovation and Entrepreneurship Training Program for Undergraduate (No. 202210316063Y).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

Asahina Y., . Wakabayashi, The mammalian-type thioredoxin reductase 1 confers a high-light tolerance to the green alga Chlamydomonas reinhardtii. Biochem. Biophys. Res. Commun.. 2022;596:97-103.
[Google Scholar]
Cheng Q., . The selenium-independent inherent pro-oxidant NADPH oxidase activity of mammalian thioredoxin reductase and its selenium-dependent direct peroxidase activities. J. Biol. Chem.. 2010;285(28):21708-21723.
[Google Scholar]
Cheung E.C., Vousden K.H., . The role of ROS in tumour development and progression. Nat. Rev. Cancer. 2022;22(5):280-297.
[Google Scholar]
Dar K.B., . Modern computational strategies for designing drugs to curb human diseases: a prospect. Curr. Top Med. Chem.. 2018;18(31):2702-2719.
[Google Scholar]
Dong C., . Role of thioredoxin reductase 1 in dysplastic transformation of human breast epithelial cells triggered by chronic oxidative stress. Scient. Rep.. 2016;6:36860.
[Google Scholar]
Fritz-Wolf K., . Crystal structure of the human thioredoxin reductase-thioredoxin complex. Nat. Commun.. 2011;2:383.
[Google Scholar]
Jovanović M., . Novel TrxR1 inhibitors show potential for glioma treatment by suppressing the invasion and sensitizing glioma cells to chemotherapy. Front. Mol. Biosci.. 2020;7:586146
[Google Scholar]
Kiely-Collins H., . The role of reversible and irreversible covalent chemistry in targeted protein degradation. Cell Chem. Biol.. 2021;28(7):952-968.
[CrossRef] [Google Scholar]
Lamport D.T.A., Várnai P., . Periplasmic arabinogalactan glycoproteins act as a calcium capacitor that regulates plant growth and development. New Phytol.. 2013;197(1):58-64.
[Google Scholar]
Lo Y.C., . Terpyridine-platinum(II) complexes are effective inhibitors of mammalian topoisomerases and human thioredoxin reductase 1. J. Inorg. Biochem.. 2009;103(7):1082-1092.
[Google Scholar]
Lu Y., . Design, facile synthesis and biological evaluations of novel pyrano[3,2-a]phenazine hybrid molecules as antitumor agents. Eur. J. Med. Chem.. 2017;127:928-943.
[Google Scholar]
Lu Y., . Design, combinatorial synthesis and biological evaluations of novel 3-amino-1'-((1-aryl-1H-1,2,3-triazol-5-yl)methyl)-2'-oxospiro[benzo[a] pyrano[2,3-c]phenazine-1,3'-indoline]-2-carbonitrile antitumor hybrid molecules. Eur. J. Med. Chem.. 2017;135:125-141.
[Google Scholar]
Mercatelli N., . MiR-23-TrxR1 as a novel molecular axis in skeletal muscle differentiation. Scient. Rep.. 2017;7(1):7219.
[Google Scholar]
Morris G.M., . AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J. Comput. Chem.. 2009;30(16):2785-2791.
[Google Scholar]
Pandey, S. Singh, B.K., 2020. De-novo Drug Design, Molecular Docking and In-Silico Molecular Prediction of AChEI Analogues through CADD Approaches as Anti-Alzheimer's Agents, Curr Comput Aided Drug Des 16 (1) 54-72
Pickering I.J., . Direct observation of methylmercury and auranofin binding to selenocysteine in thioredoxin reductase. Inorg. Chem. 2020;59(5):2711-2718.
[Google Scholar]
Roder, C. Thomson M.J., 2015. Auranofin: repurposing an old drug for a golden new age, Drugs R D 15 (1) 13-20
Sahoo B.M., . Reactive oxygen species (ROS): key components in cancer therapies. Anticancer Agents Med. Chem.. 2022;22(2):215-222.
[Google Scholar]
Sanner M.F., . Python: a programming language for software integration and development. J. Mol. Graph Model. 1999;17(1):57-61.
[Google Scholar]
Ste Marie E.J., . Can Selenoenzymes Resist Electrophilic Modification? Evidence from thioredoxin reductase and a mutant containing α-methylselenocysteine. Biochemistry. 2020;59(36):3300-3315.
[Google Scholar]
Tamura T., Stadtman T., . A new selenoprotein from human lung adenocarcinoma cells: purification, properties, and thioredoxin reductase activity. Proc. Nat. Acad. Sci. U. S. A.. 1996;93(3):1006-1011.
[Google Scholar]
Trott O., Olson A.J., . AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem.. 2010;31(2):455-461.
[Google Scholar]
Yu D., . Recent advances in application of computer-aided drug design in anti-influenza A virus drug discovery. Int. J. Mol. Sci.. 2022;23(9):4738.
[Google Scholar]
Zablocka-Slowinska K.Z., . Systemic redox status in lung cancer patients is related to altered glucose metabolism. PLOS ONE. 2018;13(9):e0204173.
[Google Scholar]
Zhang S.P., . Discovery of hydroxytyrosol as thioredoxin reductase 1 inhibitor to induce apoptosis and G. Exp. Ther. Med.. 2021;22(2):829.
[Google Scholar]
Zhang Y., . Thioredoxin reductase 1 inhibitor shikonin promotes cell necroptosis via SecTRAPs generation and oxygen-coupled redox cycling. Free Rad. Biol. Med.. 2022;180:52-62.
[Google Scholar]

Appendix A

Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.arabjc.2023.104642.

Appendix A

Supplementary material

The following are the Supplementary data to this article:

Supplementary data 1

Download

Show Sections

[1] Asahina Y., . Wakabayashi, The mammalian-type thioredoxin reductase 1 confers a high-light tolerance to the green alga Chlamydomonas reinhardtii. Biochem. Biophys. Res. Commun.. 2022;596:97-103.
[Google Scholar]

[2] Cheng Q., . The selenium-independent inherent pro-oxidant NADPH oxidase activity of mammalian thioredoxin reductase and its selenium-dependent direct peroxidase activities. J. Biol. Chem.. 2010;285(28):21708-21723.
[Google Scholar]

[3] Cheung E.C., Vousden K.H., . The role of ROS in tumour development and progression. Nat. Rev. Cancer. 2022;22(5):280-297.
[Google Scholar]

[4] Dar K.B., . Modern computational strategies for designing drugs to curb human diseases: a prospect. Curr. Top Med. Chem.. 2018;18(31):2702-2719.
[Google Scholar]

[5] Dong C., . Role of thioredoxin reductase 1 in dysplastic transformation of human breast epithelial cells triggered by chronic oxidative stress. Scient. Rep.. 2016;6:36860.
[Google Scholar]

[6] Fritz-Wolf K., . Crystal structure of the human thioredoxin reductase-thioredoxin complex. Nat. Commun.. 2011;2:383.
[Google Scholar]

[7] Jovanović M., . Novel TrxR1 inhibitors show potential for glioma treatment by suppressing the invasion and sensitizing glioma cells to chemotherapy. Front. Mol. Biosci.. 2020;7:586146
[Google Scholar]

[8] Kiely-Collins H., . The role of reversible and irreversible covalent chemistry in targeted protein degradation. Cell Chem. Biol.. 2021;28(7):952-968.
[CrossRef] [Google Scholar]

[9] Lamport D.T.A., Várnai P., . Periplasmic arabinogalactan glycoproteins act as a calcium capacitor that regulates plant growth and development. New Phytol.. 2013;197(1):58-64.
[Google Scholar]

[10] Lo Y.C., . Terpyridine-platinum(II) complexes are effective inhibitors of mammalian topoisomerases and human thioredoxin reductase 1. J. Inorg. Biochem.. 2009;103(7):1082-1092.
[Google Scholar]

[11] Lu Y., . Design, facile synthesis and biological evaluations of novel pyrano[3,2-a]phenazine hybrid molecules as antitumor agents. Eur. J. Med. Chem.. 2017;127:928-943.
[Google Scholar]

[12] Lu Y., . Design, combinatorial synthesis and biological evaluations of novel 3-amino-1'-((1-aryl-1H-1,2,3-triazol-5-yl)methyl)-2'-oxospiro[benzo[a] pyrano[2,3-c]phenazine-1,3'-indoline]-2-carbonitrile antitumor hybrid molecules. Eur. J. Med. Chem.. 2017;135:125-141.
[Google Scholar]

[13] Mercatelli N., . MiR-23-TrxR1 as a novel molecular axis in skeletal muscle differentiation. Scient. Rep.. 2017;7(1):7219.
[Google Scholar]

[14] Morris G.M., . AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J. Comput. Chem.. 2009;30(16):2785-2791.
[Google Scholar]

[15] Pandey, S. Singh, B.K., 2020. De-novo Drug Design, Molecular Docking and In-Silico Molecular Prediction of AChEI Analogues through CADD Approaches as Anti-Alzheimer's Agents, Curr Comput Aided Drug Des 16 (1) 54-72

[16] Pickering I.J., . Direct observation of methylmercury and auranofin binding to selenocysteine in thioredoxin reductase. Inorg. Chem. 2020;59(5):2711-2718.
[Google Scholar]

[17] Roder, C. Thomson M.J., 2015. Auranofin: repurposing an old drug for a golden new age, Drugs R D 15 (1) 13-20

[18] Sahoo B.M., . Reactive oxygen species (ROS): key components in cancer therapies. Anticancer Agents Med. Chem.. 2022;22(2):215-222.
[Google Scholar]

[19] Sanner M.F., . Python: a programming language for software integration and development. J. Mol. Graph Model. 1999;17(1):57-61.
[Google Scholar]

[20] Ste Marie E.J., . Can Selenoenzymes Resist Electrophilic Modification? Evidence from thioredoxin reductase and a mutant containing α-methylselenocysteine. Biochemistry. 2020;59(36):3300-3315.
[Google Scholar]

[21] Tamura T., Stadtman T., . A new selenoprotein from human lung adenocarcinoma cells: purification, properties, and thioredoxin reductase activity. Proc. Nat. Acad. Sci. U. S. A.. 1996;93(3):1006-1011.
[Google Scholar]

[22] Trott O., Olson A.J., . AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem.. 2010;31(2):455-461.
[Google Scholar]

[23] Yu D., . Recent advances in application of computer-aided drug design in anti-influenza A virus drug discovery. Int. J. Mol. Sci.. 2022;23(9):4738.
[Google Scholar]

[24] Zablocka-Slowinska K.Z., . Systemic redox status in lung cancer patients is related to altered glucose metabolism. PLOS ONE. 2018;13(9):e0204173.
[Google Scholar]

[25] Zhang S.P., . Discovery of hydroxytyrosol as thioredoxin reductase 1 inhibitor to induce apoptosis and G. Exp. Ther. Med.. 2021;22(2):829.
[Google Scholar]

[26] Zhang Y., . Thioredoxin reductase 1 inhibitor shikonin promotes cell necroptosis via SecTRAPs generation and oxygen-coupled redox cycling. Free Rad. Biol. Med.. 2022;180:52-62.
[Google Scholar]

Discovery of novel inhibition site centered on 114-bit tryptophan of Thioredoxin reductase 1 through computer-aided drug design

Abstract

Keywords

Thioredoxin reductase 1

Computer-aided drug design

Virtual screening

Molecular docking

Trp114

1 Introduction

2 Results

2.1 Virtual screening based on molecular docking

2.2 Inhibitory activity of virtual screening compounds

2.3 Molecular docking found the binding amino acids of compounds and enzyme

2.4 Verification of inhibitors binding site

3 Discussion

4 Materials and method

4.1 Virtual screening based on molecular docking

4.2 Produced heterologous TrxR1

4.3 TrxR1 activity assay (DTNB method)

4.4 Molecular docking verification binding site

4.5 Inhibitor binding site validation test

4.6 Statistical method for amino acid preference of inhibitors

5 Conclusion

Acknowledgements

Declaration of Competing Interest

References

Appendix A

Supplementary material

Appendix A

Supplementary material

Supplementary data 1

Suggested read for related articles: