Construction of safer hirudin-derived peptides with enhanced anticoagulant properties based on C-terminal active residue adaptation and modification

Graphical abstract

Highlights

• •The inhibitory binding and anticoagulant mechanism of hirudin C-terminal on thrombin were clarified.
• •The role of π-π stacking and π-cation interactions in hirudin-thrombin affinity was revealed.
• •The importance of tyrosine sulfation in enhancing hirudin anticoagulant activity was highlighted.
• •In vitro sulfation catalysis was conducted using the leech homolog TPST.
• •Safer, more effective hirudin-derived anticoagulant lead compounds were designed and validated.

Abstract

Introduction

Hirudin exerts anticoagulant effects by inhibiting the binding and catalytic activity of thrombin to fibrinogen. However, its rigid N-terminal region irreversibly occupies the thrombin catalytic center, raising concerns about potential bleeding.

Objectives

In this study, a novel lead compound, WPHVC_V1, which is based on the competitive binding mechanism of the hirudin variant WPHV_C, was developed and validated for in vitro and in vivo activity and safety.

Methods

Saturation mutagenesis, molecular dynamics simulations and mutant protein activity assays were used to elucidate the competitive anticoagulant mechanism between WPHV_C and thrombin. Next, a recombinantly expressed tyrosylprotein sulfotransferase was used to modify and confirm the sulfation site on the C-terminal tyrosine of hirudin. Finally, a multisite aromatic amino acid mutation strategy was implemented to design and synthesize the lead anticoagulant, WPHVC_V1.

Results

The acidic amino acid cluster in WPHV_C formed strong electrostatic interactions with the positively charged thrombin exosite I, blocking fibrinogen binding. The introduction of aromatic amino acids further stabilized the complex through π-π stacking and π-cation interactions. For example, mutation of 13E to A decreased the free energy of dissociation (ΔG) from 19.27 to 10.93 kcal·mol⁻¹ and shortened the thrombin time (TT) from 42.00 s to 30.94 s, whereas mutation of 26 K to W increased the ΔG to 24.70 kcal·mol⁻¹ and prolonged TT to 51.92 s. In addition, the aromatic effect of 20Y, combined with sulfation, synergistically enhanced binding. Based on these findings, the newly designed WPHVC_V1 showed a ΔG of 37.24 kcal·mol⁻¹ and, at 0.1 mg/ml, increased TT/APTT/PT from 41.72/14.38/15.86 s (WPHV_C) to 62.08/23.38/22.22 s. In in vivo studies, WPHVC_V1 achieved tail thrombus inhibition in the mouse tail by reducing the length of the thrombus from 3.562 cm in CK to 1.853 cm (1.729 cm for sodium heparin and 2.530 cm for WPHV_C), completely inhibited thrombus formation in a carotid artery model and reduced tail bleeding time by 35.2 s compared with heparin sodium. Safety evaluations revealed that WPHVC_V1 did not cause hemolysis, had no significant effect on blood pressure or cause pathological changes in major organs.

Conclusion

These findings provide an initial foundation and sequence reference for the development of safe and effective anticoagulant drugs with potential for clinical translation.

Introduction

Cardiovascular disease is the leading cause of mortality worldwide. Ischemic heart disease and stroke, which are usually induced by abnormal clot formation, are the most prevalent and lethal causes [1,2]. The presence of pathological factors can interfere with normal hemostasis, leading to uncontrolled thrombosis and vessel occlusion [3]. Thrombin plays a central role in thrombosis by converting fibrinogen into fibrin to form clots and activating platelets to accelerate aggregation [4,5]. Therefore, thrombin inhibition based anticoagulant therapy has become a key strategy for the treatment of cardiovascular and thrombotic diseases.

Direct thrombin inhibitors (DTIs), such as cardiovascular surgery and acute coronary syndrome, are widely used in clinical anticoagulant therapy [6]. The most common DTIs include dabigatran [7], argatroban [8], and bivalirudin [9]. By directly binding to thrombin to block its interaction with substrates, these DTIs have the advantages of rapid onset, predictable effects, and a short duration of action [10]. Although synthetic DTIs are widely used in clinical settings, natural anticoagulants have received significant research attention because of their unique mechanisms of action and potential therapeutic advantages.

Hirudin, which is extracted from a medicinal leech (Hirudo medicinalis), is the most potent natural inhibitor of thrombin [11,12]. The anticoagulant mechanism of hirudin involves two parts. First, the C-terminus can bind to exosite I on the H-chain of thrombin via electrostatic interactions. Its ability to target thrombin allows hirudin to compete with fibrinogen for the active site of thrombin [13]. Second, the N-terminus forms a tight structure with three disulfide bonds and blocks the catalytic center of thrombin, thereby inhibiting the hydrolysis of fibrinogen by thrombin to form fibrin (Fig. S1). The specific binding of hirudin to thrombin provides a valuable template for the design of DTIs [14]. For example, hirudin-derived lepirudin and bivalirudin have been successfully used in the clinical treatment of acute coronary syndrome and percutaneous coronary intervention [15,16]. However, hirudin and its derivatives may lead to a potential risk of bleeding due to their irreversible inhibitory effects on thrombin [17], and there are currently no suitable antidotes to reverse bleeding in emergency situations. However, their clinical application is limited, especially in surgical or trauma patients, who require rapid restoration of blood clotting function [10]. Therefore, despite its potent anticoagulant potential, hirudin has not been widely applied to treat cardiovascular disease [18]. The development of safer and more effective new DTIs remains an important and urgent research priority.

To increase the anticoagulant activity of hirudin or reduce the risk of bleeding, various approaches have been undertaken by researchers to promote its clinical application. For example, Sun et al. designed the hirudin mutant HM2-E60D-I62D through mutant screening and molecular dynamics simulations, which significantly improved its antithrombotic efficacy [19]. Additionally, Tian et al. developed an elastin-like polypeptide-hirudin fusion protein, which not only enables long-term antithrombotic therapy but also effectively decreases the risk of bleeding associated with hirudin [20]. Previously, our group discovered a novel variant of hirudin (WPHV1, Whitmania pigra Hirudin Variant 1) with anticoagulant activity derived from the nonhematophagous medicinal leech Whitmania pigra [21]. When the N-terminus of WPHV1 was removed and only its C-terminal chain was expressed in Escherichia coli, the recombinant protein still retained anticoagulant activity. Another study revealed that the C-terminus of hirudin could inhibit thrombin-induced platelet aggregation without affecting its ability to catalyze the hydrolysis of chromogenic substrates [22].

Therefore, it is hypothesized that hirudin can exert anticoagulant effects by competitively binding to thrombin, where fibrinogen binds via the C-terminal structure alone. This finding provides a new perspective for the clinical use of hirudin: removing the N-terminal structure could alleviate the irreversible inhibition of thrombin and reduce potential bleeding risk. However, the anticoagulant activity of hirudin decreased after the N-terminus was deleted. Therefore, enhancing the anticoagulant activity of the C-terminus of hirudin has become an urgent issue. Studies have shown that the binding of hirudin to thrombin is closely related to electrostatic interactions. In hirudin, these interactions originate from a C-terminus rich in acidic amino acids and the sulfation of the tyrosine residue at position 63 [23]. The C-terminus of WPHV1 contains nine acidic amino acids and one tyrosine residue capable of sulfation, demonstrating great potential for binding to thrombin and potentially serving as a natural template for the development of new hirudin-derived DTIs.

This study investigated the binding mechanism between the C-terminus of WPHV1 (WPHV_C) and thrombin, analyzed the effect of sulfation modifications on its anticoagulant activity, and proposed methods to optimize anticoagulant activity through amino acid mutations. The aim of this research was to design a novel C-terminal derived hirudin anticoagulant peptide that could effectively reduce the clinical risk of bleeding while maintaining significant anticoagulant activity. Through comprehensive mechanistic studies, this research has provided new insights into anticoagulant drug development, facilitating the implementation of safer and more effective anticoagulant treatment strategies, thereby contributing to improved outcomes in cardiovascular disease patients.

Materials and methods

Amino acid sequence and protein structure analysis, molecular docking with thrombin, and saturation mutation prediction of WPHV_C

Amino acid sequences and structural information for WPHV1 and other hirudin variants were obtained from our previous study [21] and UniProt [24] (https://www.uniprot.org/). Multiple sequence alignments of their C-terminal regions were conducted via ClustalW in MEGA 7.0 [25] and visualized with Jalview [26]. Tyrosine sulfation sites were predicted with Sulfinator [27] (https://web.ExPASy.org/sulfinator/). The protein structures of hirudin and thrombin (PDB No.: 1HRT) were retrieved from the Protein Data Bank (https://www.pdb.org/) and analyzed for visualization, comparison, and surface potential via ChimeraX 1.4 [28].

Flexible molecular docking between WPHV_C and thrombin was performed via the docking_protocol tool in Rosetta [29]. Thrombin was modeled as a rigid receptor, whereas WPHV_C was treated as a flexible ligand. One thousand relaxed conformations of WPHV_C were generated via an unconstrained relax protocol, and the top 50 scoring models were selected to form a flexible ligand ensemble. Subsequently, 5,000 docking simulations were executed, and the complex with the highest score was chosen as the WPHV_C-thrombin binding model. Saturation mutagenesis of all amino acids in WPHV_C was conducted via flex_ddG [30] in Rosetta to predict the changes in the binding free energy (ΔΔG) values of the mutations. The results were visualized via TBtools [31].

Prediction and analysis of electrostatic and aromatic π-electron interactions in the WPHV_C-thrombin binding model

Electrostatic and π interactions in the WPHV_C-thrombin binding model were analyzed via Python scripts and ChimeraX. Structural data of the WPHV_C-thrombin complex were processed, and distances between side-chain atoms of acidic and basic amino acids were calculated with the Bio.PDB library. Specifically, the distances between OD1 and OD2 of aspartic acid (ASP) and OE1 and OE2 of glutamic acid (GLU) in WPHV_C and between NZ of lysine (LYS), NH1 and NH2 of arginine (ARG), and ND1 and NE2 of histidine (HIS) in thrombin were measured. Additionally, distances between the centers of aromatic rings in aromatic amino acids of both WPHV_C and thrombin, as well as between aromatic ring centers in WPHV_C and side chains of basic amino acids in thrombin, were calculated. The computations were validated by comparing the results with visualization outputs from ChimeraX.

Molecular dynamics simulations and binding energy analysis of WPHV_C and its mutants

Molecular dynamics simulations of WPHV_C and its mutants were conducted via GROMACS (2022.5) [32] on a Windows 10 system with an NVIDIA A100-PCIE-40 GB GPU. Initially, the CHARMM27 all-atom force field together with the SPC/E explicit water model was selected to construct the simulation system. A rectangular water box (dimensions of 7.0 × 6.5 × 19 nm³) was then set up under periodic boundary conditions, and Na⁺/Cl⁻ ions were added to the system to balance the charge [19]. After system construction, energy minimization was performed via the steepest descent method (100000 steps with a force convergence threshold of 100 kJ·mol⁻¹·nm⁻¹), and the system was equilibrated under the NVT (V-rescale) and NPT (V-rescale) ensembles to ensure stability. After the entire simulation system was prepared, the complex was subjected to a molecular dynamics simulation via the Umbrella Sampling (US) method to study the change in binding energy during the dissociation process of WPHV_C from thrombin (Fig. S2). The center of mass distance between WPHV_C and thrombin was defined as the reaction coordinate, a harmonic restraint force constant of 1500 kJ·mol⁻¹·nm⁻² was used, and WPHV_C was pulled away from the thrombin binding site at a rate of 0.005 nm·ps⁻¹ over 1000 ps, with coordinates and data recorded every 1 ps. US was subsequently performed throughout the entire dissociation process, with overlapping windows set at a length of 0.2 nm to ensure that sufficient sampling of the free energy changes was achieved. Each iteration was run for 5,000,000 steps to ensure that the system fully explored the free energy surface. After sampling, the data from each window were integrated via the Weighted Histogram Analysis Method (WHAM) to calculate the Gibbs free energy change (ΔG) and assess the stability of the simulated system through the ΔG convergence curve. To assess the reliability of the results, three independent simulations were performed for each mutant and a representative result is presented in the main text (Fig. S3).

CDS synthesis, expression vector construction, and prokaryotic expression and purification of WPHV_C and its mutants

The coding sequence (CDS) of WPHV_C, its mutants, and lead compounds were optimized via the E. coli codon optimization tool ICOR [33] and synthesized by the Chinese Academy of Agricultural Sciences (sequences are provided in Table S1). The optimized CDS was inserted between the HindIII and NotI restriction sites of the expression vector pET21a. The recombinant plasmids were transformed into E. coli BL21 (DE3) competent cells, and positive clones were selected and verified by sequencing.

The verified recombinant strains were inoculated into 200 ml of LB medium containing ampicillin at a 1:20 ratio and cultured at 37 °C with shaking at 200 rpm until the OD600 reached approximately 0.6. IPTG was then added to a final concentration of 0.5 mM, and induction was conducted at 20 °C with shaking at 140 rpm for 20 h. Following induction, the cells were collected, washed, and resuspended in PBS buffer (pH 7.4), followed by ultrasonic disruption. The cell lysate was centrifuged, and the supernatant was purified via a His-tag protein purification kit (Beyotime Biotech. Inc.).

Protein samples were concentrated and dialyzed via an Amicon Ultra-4 centrifugal filter (Millipore) with a 1 kDa molecular weight cutoff. Protein concentration was determined via the Bradford assay (Biotopped). The expression and purity of the recombinant proteins were analyzed via SDS-PAGE and Coomassie Brilliant Blue staining. Finally, the recombinant protein was dialyzed into the appropriate buffer and adjusted to the desired concentration.

In vitro anticoagulant activity assays of WPHV_C

Thrombin time (TT), activated partial thromboplastin time (APTT), and prothrombin time (PT) were measured via coagulation assay kits from Shanghai Sun Biotech Co., Ltd. The assays were performed on an SC40 (LG-PABER-I) semiautomatic coagulation analyzer. Platelet-poor plasma (PPP) was prepared from rabbit blood and anticoagulated with sodium citrate (Sbjbio). Considering the precision and detection range of the semiautomatic thrombin analyzer, concentration ranges of 0.025, 0.05 and 0.1 mg/ml were chosen to cover the possible effective concentration intervals. To ensure homogeneity, all protein samples were individually diluted with 50 mM Tris-HCl buffer (pH 7.4) prior to the experiment. Heparin sodium (140 U/mg, Solarbio) was used as a positive control. Each sample was measured five times, and the mean values and standard deviations were calculated. Duplicate experimental data were processed via the pandas library in Python, and independent samples Student’s t-tests were performed via the scipy.stats library. Pairwise comparisons were made between different groups, and t-statistics and p-values were output for each pair of groups to assess their significance.

For TT measurements, 160 µl of PPP was placed in a sterile centrifuge tube, followed by the addition of 20 µl of the protein sample. The mixture was incubated at 37 °C for 3 min. Subsequently, 20 µl of TT reagent was added, and the clotting time was recorded. For APTT measurements, 100 µl of PPP was placed in a sterile centrifuge tube. Then, 40 µl of APTT reagent preheated to 37 °C and 20 µl of the protein sample were added. The mixture was incubated at 37 °C for 5 min, followed by the addition of 40 µl of preheated 0.025 mol/l CaCl₂. The APTT value was recorded. For PT measurements, 160 µl of PPP was placed in a sterile centrifuge tube, and 20 µl of the protein sample was added. The mixture was incubated at 37 °C for 3 min. Then, 20 µl of PT reagent was added, and the PT value was recorded.

Identification, characterization, and in vitro sulfation catalysis of homologous TPST from W. pigra

A homologous tyrosylprotein sulfotransferase (TPST) protein (WP_TPST) was identified from the W. pigra genome via the TPST-HMM tool [34] (https://sz.bjfskj.com/), and its annotation was extracted from the GFF3 file. Detailed information on the cDNA of W. pigra was provided in our previous study [21]. Primers targeting the 5′UTR and 3′UTR of WP_TPST were designed for PCR amplification. The PCR products were purified, ligated into a blunt-end vector, transformed into E. coli T1 cells, and positive clones were sequenced to obtain the WP_TPST sequence.

The tertiary structure of WP_TPST was predicted via AlphaFold2 [35]. TM-score and RMSD calculations for protein comparisons were performed with the US-align online tool [36] (https://zhanggroup.org/US-align/). Flexible molecular docking between WPHV_C and WP_TPST was conducted following the methodology described in Method 2.1.

The CDS of WP_TPST was cloned and inserted into the pET43.1a(+) vector via the HindIII and NotI restriction sites and transformed into E. coli for expression. The expression and purification procedures were identical to those used for WPHV_C. The NusA tag was removed via enterokinase.

An in vitro tyrosine sulfation assay was performed with WP_TPST (0.5 mg/ml) and WPHV_C (0.5 mg/ml) in a reaction mixture containing 20  mM MES buffer (pH 6.5), 100 μM PAPS lithium salt hydrate, 0.1 % Triton X-100, and 50 mM NaCl. For the control group, double-distilled water (ddH₂O) was used to replace the missing components in the reaction mixture. After incubation at 25 °C for 1 h, the reaction products were fixed in a 96-well plate, blocked with 5 % skim milk, incubated with an anti-sulfotyrosine primary antibody, and then incubated with an HRP-conjugated goat anti-mouse IgG secondary antibody. The HRP color reaction was subsequently developed using TMB, and the reaction was terminated with H₂SO₄. The absorbance at 450 nm was measured, and the blank sample absorbance value was subtracted for calibration [37]. Reactions were incubated at 25 °C for 24 h, and TT was measured to assess anticoagulant activity.

In vivo anticoagulant activity of WPHV_C and its lead compounds in mouse models

The animal study complied with the ARRIVE guidelines [38]. Five-week-old male Kunming mice (Swiss strain, SPF grade, Beijing Hua Fukang Biotechnology Co., Ltd) were randomly assigned to four groups and housed under standard conditions (22–24 °C, 12-hour light/dark cycle) with free access to food and water. The experimental protocol was approved by the Animal Ethics Committee of the Institute of Medicinal Plant Development, Chinese Academy of Medical Sciences, following National Institutes of Health guidelines (project approval number: SLXD-20241112013). The 3R principles were followed to minimize animal suffering. The mice were anesthetized with 2.5 % isoflurane. Tail incision bleeding time and blood collection were performed after the mice were anesthetized. Euthanasia was performed using an overdose of carbon dioxide.

Preliminary experiments with small samples revealed that a dose of 3 mg/kg protein was effective in demonstrating in vivo potency. All animal experiments were divided into four groups: two experimental groups received 3 mg/kg WPHV_C or WPHVC_V1; a blank control group received saline; and a positive control group received 3 mg/kg heparin sodium (140 U/mg). All the groups had five replicates. The drug was administered by subcutaneous injection.

Mouse tail thrombosis model: In vivo anticoagulant activity was evaluated in a carrageenan-induced mouse tail thrombosis model [39]. Subcutaneous injection of 60 mg/kg of γ-carrageenan (CAS: 9000–07-1) was used to induce thrombus formation 30 min after anticoagulant injection [40]. Photographs of mouse tail thrombi were taken 24 h after injection with a scale bar placed during imaging. The length of the well-demarcated dark thrombus area in the mouse tail was measured via ImageJ software by scale bar calibration. The changes in thrombus length reflect the inhibitory effects of each group on thrombus formation.

Mouse carotid artery thrombosis model: Mice were anesthetized 30 min after anticoagulant injection. The carotid artery was exposed and isolated via sterile surgical scissors. A filter paper soaked in 5 % FeCl₃ was then placed on the carotid artery for 2 min. After removal of the filter paper, carotid blood flow was observed and recorded via a laser Doppler speckle blood flow imaging system.

Tail incision bleeding time in mice: A 3 mm deep incision was made 4 cm from the base of the mouse tail 30 min after anticoagulant injection. Bleeding time was recorded from the onset of bleeding until no blood remained on the filter paper after gentle blotting every 8 s.

In vivo plasma thrombin activity test in mice: Blood was collected from the mice into sodium citrate anticoagulation tubes 30 min, 2 h, and 4 h after anticoagulant injection. The plasma was separated and subjected to TT, APTT, and PT assays to evaluate the effects of the recombinant hirudin protein in vivo. The methods for TT, APTT, and PT followed those described in Method 2.5, without the addition of any extra anticoagulants.

Initial validation of WPHVC_V1 features and security

Mouse red blood cells (RBCs) (2 %) were incubated with saline, WPHVC_V1 (1 mg/ml) or Triton X-100 (0.1 %) at 37 °C for 4 h. After centrifugation, the absorbance of the supernatant was measured at 580 nm [41].

Healthy mice were injected with WPHVC_V1 (3 mg/kg), and saline was used as a control. The blood pressure of the mice was measured after administration via the tail-cuff method. A pulse cuff was attached to the tail of each mouse, and changes in the blood flow signal were detected by inflating and deflating the cuff, with blood pressure values determined from the appearance and disappearance of the pulse wave [42]. The heart, liver, spleen, lung and kidney were then collected. The organs were embedded, sectioned and stained with H&E, and the stained tissue sections were examined under a microscope.

Whole blood clots obtained from mouse clotting were gently washed with saline and transferred, to EP tubes to which 3.0 ml of PBS buffer was added for soaking. WPHVC_V1 (1 mg/ml) was then added, with saline used as a control. After incubation in vitro at 37 °C for 2 h, the supernatant was collected, and its absorbance was measured at 580 nm [43].

Results

Characterization, mechanism of thrombin binding, and predictive analysis of saturated amino acid mutations in WPHV_C

To investigate the interaction mechanisms of WPHV1 and its C-terminal structure with thrombin during anticoagulation, their structures and molecular binding models were predicted. Fig. 1A and 1B show the amino acid sequence and protein structure of WPHV1. Surface potential visualization (Fig. 1C) revealed that the transmembrane domain is neutral, the N-terminal domain is partially positively charged, and the C-terminal chain is negatively charged. Although differences in the C-terminal sequences of hirudin variants were observed, their key features remained consistent (Fig. 1D). For example, the C-terminus of WPHV1 contains a sulfatable tyrosine and nine acidic residues. These abundant acidic side chains enhanced the binding affinity between the hirudin C-terminus and thrombin, as well as its anticoagulant activity.

Fig. 1.

Sequence and structural characteristics of WPHV_1, the predicted molecular model of WPHV_C-thrombin binding, and the saturation mutation results. (A) Amino acid sequence of WPHV_1. (B) WPHV_1 model predicted by AlphaFold2. (C) Surface electrostatic potential of WPHV_1, with red indicating negative potentials and blue indicating positive potentials. (D) C-terminal sequence alignment of hirudin variants. (E) Predicted docking model of WPHV_C binding to thrombin, showing the thrombin H-chain (green), L-chain (brown), and WPHV_C (red). (F) Surface electrostatic potential of the WPHV_C-thrombin complex. (G) Heatmap of ΔΔG for WPHV_C-thrombin binding after saturation mutation of WPHV_C. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Since WPHV_C predominantly presents disordered chains in solution, flexible docking was used to simulate its binding to thrombin (Fig. 1E). The surface potential visualization (Fig. 1F) revealed that the acidic tail of WPHV_C was tightly bound to thrombin basic exosite I, suggesting its potential to inhibit thrombin. To explore the molecular mechanism underlying the interaction between WPHV_C and thrombin, a systematic saturation mutagenesis scan of 26 amino acid residues in WPHV_C was performed via the Flex ddG method, and the effect of each mutation on the ΔΔG of the complex was quantitatively assessed (Fig. 1G). A decrease in ΔΔG indicates an increase in complex stability, whereas an increase in ΔΔG reflects a decrease in binding affinity. The main results were as follows. Nine acidic amino acid residues and tyrosine residues capable of being sulfated were found to contribute to the stability of the complex (with ΔΔG > 0 after mutation). Among these, the contributions of glutamic acid (E) at position 13 and tyrosine (Y) at position 20 were the most significant (with average ΔΔG values of + 1.66 kcal·mol⁻¹ and + 2.69 kcal·mol⁻¹, respectively, after mutation). In addition, the introduction of phenylalanine (F), histidine (H), tryptophan (W) or tyrosine (Y) was observed to significantly reduce ΔΔG (average value of −1.03 kcal·mol⁻¹), suggesting that mutations introducing aromatic amino acids at appropriate positions may universally contribute to the stability of the WPHV_C-thrombin complex. Among these mutations, the K26W mutation had the most significant effect (ΔΔG = −2.09 kcal·mol⁻¹).

Electrostatic interactions between WPHV_C and thrombin and the Impact of aromatic amino acid mutations on the binding stability

Since the electrostatic interactions between WPHV_C and thrombin are key factors affecting anticoagulant activity, the electrostatic interactions were predicted via a binding model. The electrostatic interactions occur between acidic and basic amino acids within 20 Å and form stable ionic bonds within 6 Å. The distances between the acidic side chains of nine acidic residues in WPHV_C and the basic side chains of basic residues in thrombin are shown in Table 1. (1) All acidic residues interact with the 19 basic residues of thrombin within 20 Å, suggesting that they contribute to thrombin binding. (2) Thrombin basic residues 21 K, 68R, 73R and 107 K, located within 6 Å of WPHV_C acidic residues, interact with exosite I. (3) 13E forms close interactions with several thrombin basic residues and is the most active acidic residue, including four within 6 Å. (4) Within 6 Å, residues 16E, 21D, 22D, and 24D interact with each other, whereas residues 12D, 14E, 23D, and 25D do not. These results show that there is a positive correlation between the number and proximity of electrostatic interactions and the predicted ΔΔG from saturation mutagenesis.

Table 1.

Interactions number and distance range between acidic amino acids of WPHV_C and basic amino acid side chains of thrombin.

	< 6 Å	6 ∼ 12 Å	12 ∼ 20 Å
12D	0	0	22
13E	4	7	27
14E	0	3	10
16E	1	6	25
21D	1	5	16
22D	2	10	22
23D	0	7	12
24D	2	7	21
25D	0	4	20

Saturation mutagenesis revealed that mutations in certain aromatic amino acids (F, H, W, Y) significantly decreased ΔΔG, thereby increasing the binding affinity of WPHV_C for thrombin. This effect is likely due to the hydrophobicity and π-electron interactions of the aromatic residues, which increase the binding stability. Aromatic amino acids contribute through π-π stacking (between aromatic rings, 4–7 Å) and π-cation interactions (between aromatic rings and basic amino acids, 4–6 Å). Such interactions between WPHV_C and thrombin were identified (Fig. 2). Specifically, residue 7H of WPHV_C forms π-π stacking with thrombin residues 47Y and 50 W; residue 20Y interacts with thrombin residues 19F and 71Y. Additionally, WPHV_C residues 2Y and 20Y form π-cation interactions with thrombin residues 178R and 62R, respectively.

Fig. 2.

Aromatic amino acid π interactions between WPHV_C and thrombin. (A) π-π stacking between aromatic rings (4–7 Å). (B) π-cation interactions between aromatic rings and basic amino acids (4–6 Å).

To understand the decrease in ΔΔG due to aromatic mutations, further simulations and analyses were performed. Each WPHV_C residue was replaced with F, H, W or Y and energy minimization simulations were performed via GROMACS. The π-π stacking and π-cation interactions at the mutated sites are as follows (Table 2): (1) Mutations at residues 5A, 8Q, 22D, 24D, and 26 K introduce new π-π stacking or π-cation interactions, which are consistent with the significant decreases in ΔΔG. (2) Mutations in residues 6 M, 9S, 13E and 16E also introduce new interactions. These residues are crucial residues that maintain structural stability, but aromatic mutations weaken these effects. The decrease in ΔΔG is smaller than that of nonaromatic mutations and is not sufficient to significantly improve the binding stability. (3) Mutations in other sites do not result in new interactions or significant changes in ΔΔG. These results indicate that the π-π stacking and π-cation interactions introduced by aromatic residues play a key role in enhancing the binding stability between WPHV_C and thrombin.

Table 2.

Number of π-π stacking and π-cation interactions resulting from aromatic amino acid mutations in WPHV_C.

	π-π				π-cation
	F	H	W	Y	F	H	W	Y
2Y					1	1		1
5A	1	1		1			1
6M					2	2	1	2
7H	1	2	1
8Q		1	1			1
9S	1	1	1	1
13E					2	2	2	2
16E					1	1	1	1
20Y	2	2	2	2	1		2
22D	1	1	1	1	1	1	1	1
24D					1	1	2	1
26K					2	1	2	1

Molecular dynamics simulations reveal the role of key amino acids in WPHV_C-thrombin and validation anticoagulation

Molecular dynamics (MD) simulations were performed with GROMACS to verify the predicted role of acidic and aromatic amino acids in the binding of WPHV_C to thrombin. To assess the effects of acidic residues, nine acidic amino acids were mutated to alanine in three groups based on their contribution to binding energy and proximity to the basic residues of thrombin: (1) the most active residue 13E; (2) residues within 6 Å of basic residues of thrombin (16E, 21D, 22D, 24D); and (3) residues between 6 Å and 20 Å away (12D, 14E, 23D, 25D). For aromatic amino acid mutations, residue 5A is altered to phenylalanine and residues 8Q, 22D, 24D, and 26 K are altered to tryptophan based on interaction analyses. Using these mutant models, Gibbs free energy changes (ΔG) were calculated during the pulling of WPHV_C from the WPHV_C-thrombin composite model via umbrella sampling (US) simulations (Fig. 3A, Fig. S3). The ΔG convergence curve revealed that when the center-of-mass distance between WPHV_C and thrombin reached approximately 4 nm (sampling extended to ∼ 6 nm), the dissociation ΔG tended to stabilize, indicating sufficient sampling of the system. In the wild-type (WT) model, the ΔG was 19.27 kcal·mol⁻¹. Mutating acidic residues significantly decreases ΔG, most notably E13A, highlighting the importance of 13E (E13A: 10.93 kcal·mol⁻¹, DE_16/21/22/24_A: 14.32 kcal·mol⁻¹, DE_12/14/23/25_A: 16.86 kcal·mol⁻¹). Aromatic mutations significantly increased the ΔG (A5F: 28.56 kcal·mol⁻¹, Q8W: 27.42 kcal·mol⁻¹, D22W: 23.49 kcal·mol⁻¹, D24W: 22.51 kcal·mol⁻¹, K26W: 24.70 kcal·mol⁻¹), suggesting enhanced binding affinity. These results underscore the critical roles of acidic and aromatic amino acids in WPHV_C binding to thrombin.

Fig. 3.

Prediction of binding energy and validation of the anticoagulant activity of WPHV_C WT and mutants with thrombin. (A) Convergence curve of ΔG when WPHV_C is pulled apart from thrombin. A larger ΔG indicates a more stable WPHV_C-thrombin binding model. (B) The final protein concentration was 0.05 mg/ml. (C) The final protein concentration was 0.1 mg/ml. n = 5; the error bars indicate the standard deviation. The significance of all the experimental groups with CK was < 0.0001. * represents significant differences between the mutants and the WT. *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001.

To validate these findings and assess their effects on anticoagulant activity, the WT and mutant WPHV_C proteins were cloned, expressed, and purified in E. coli BL21(DE3). Two concentrations (0.05 mg/ml and 0.1 mg/ml) of the protein were investigated for TT assays. Compared with the control (CK), the WT significantly prolonged TT at both concentrations, indicating its strong anticoagulant activity (Fig. 3B and 3C). Mutation of acidic residues to alanine decreased TT, especially E13A, indicating reduced activity (e.g., at 0.1 mg/ml, the TT for E13A was 30.94 s versus 42.00 s for the WT). The introduction of aromatic amino acids significantly enhances the anticoagulant activity of WPHV_C, with TT at 0.1 mg/ml prolonged by an additional 3–10 s for all five aromatic mutants. Mutation of 22D and 24D to aromatic amino acids results in a slightly smaller increase, which may be due to the loss of electrostatic interactions that reduce the binding efficiency.

Analysis of the effect of sulfonatable 20Y in WPHV_C on anticoagulant activity

The only sulfonatable residue in WPHV_C is the tyrosine residue at position 20. The sulfation of tyrosine introduces a negatively charged sulfonate group (SO₃⁻), which is similar to the side chains of acidic amino acids aspartic or glutamate. To assess the role of electrostatic interactions, 20Y was mutated to aspartic (Y20D, acidic side chain: –COOH) or glutamate (Y20E, acidic side chain: –CH2-COOH) to simulate the properties of sulfonated tyrosine. These mutations resulted in multiple electrostatic interactions within 6 Å between the mutated residues and the basic amino acids 62R and 65 K on the H-chain of thrombin (Y20D: 4 interactions; Y20E: 6 interactions) (Fig. 4A and 4B). No such interactions were observed between 62R or 65 K and the nine other acidic residues in WPHV_C, suggesting that 62R and 65 K interact primarily with the sulfonate group of 20Y (Fig. 4C). MD simulations (Fig. 4D) revealed that the mutation of 20Y to alanine (Y20A) significantly reduced the ΔG from 19.27 kcal·mol⁻¹ to 12.26 kcal·mol⁻¹, indicating a decrease in affinity. Compared with the WT, Y20A presented reduced anticoagulant activity, with TT activity decreasing to 28.82 s and 34.04 s at 0.05 mg/ml and 0.1 mg/ml, respectively (Fig. 4E). For the mutants in which 20Y is replaced by acidic amino acids, the binding energy and TT of Y20D are slightly lower than those of the WT, whereas those of Y20E are nearly identical to those of the WT. These results suggest that the introduction of acidic residues partially compensates for the loss of aromatic interactions through new electrostatic interactions, but the anticoagulant activity of WPHV_C is not significantly improved (0.05 mg/ml: p = 0.82; 0.1 mg/ml: p = 0.20).

Fig. 4.

Mechanistic resolution and validation of sulfonatable tyrosine residues in WPHV_C. (A) Schematic diagram of the electrostatic interactions between Y20D and the basic amino acids in thrombin. (B) Schematic diagram of the electrostatic interactions between Y20E and the basic amino acids in thrombin. (C) Hypothetical schematic diagram of the electrostatic interactions between sulfonated 20Y and the basic amino acids in thrombin. (D) MD simulation results of the ΔG between WT and its Y20A, Y20D, and Y20E mutants with thrombin. (E) Standard TT coagulation tests conducted for WT and its Y20A, Y20D, and Y20E mutants. n = 5; the error bars indicate the standard deviation. * represents significant differences between the mutants and the WT. *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001, ns: p > 0.05.

Functional assessment of WP_TPST mediated tyrosine sulfation in WPHV_C

Using TPST-HMM, a TPST protein (WP_TPST) and its corresponding gene were identified from the W. pigra genome, and its CDS was confirmed through cloning and sequencing (Table S2). The structure of WP_TPST was predicted by AlphaFold2 and compared with that of human TPST2 (PDB No. 3AP1) (Fig. 5A). The value of the RMSD is 0.96, and the TM-score is 0.98, indicating that their catalytic core structures are highly conserved (a RMSD < 2.0 indicates structural similarity; a TM-score ≥ 0.5 indicates a similar global topology, with a maximum value of 1.0). Based on studies of the catalytic mechanism of human TPST2, the binding pocket of TPST, which is located between the α2-α3 helices and the βe sheet, was identified. The molecular binding model of WPHV_C to WP_TPST was predicted via Rosetta (Fig. 5B). Visualization of the surface electrostatic potential revealed that the interaction is due primarily to electrostatic interactions between acidic and basic amino acids (Fig. 5C).

Fig. 5.

Study of tyrosine sulfation catalyzed by WP_TPST on WPHV_C. (A) Structural comparison of human TPST2 (yellow) and WP_TPST (light blue). (B) Predicted catalytic binding model of WP_TPST with WPHV_C (red). (C) Surface electrostatic potential of the catalytic pocket, with red indicating negative potentials and blue indicating positive potentials. (D) Tyrosine sulfate group structure ELISA specific antibody detection (OD = 450 nm). (E) TT results for the in vitro system comprising WP_TPST, WPHV_C, and PAPS compared with those of the controls (the final protein concentration in all systems was 0.05 mg/ml, n = 5). Significance was graded via abcd, and p < 0.01 between the two groups was classified as different. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

To verify the ability of WP_TPST to catalyze tyrosine sulfation in WPHV_C, its CDS (excluding the transmembrane region) was cloned and inserted into the pET-43.1a vector, and soluble WP_TPST was expressed in E. coli. A complete catalytic system with WP_TPST as the enzyme, WPHV_C as the substrate, and PAPS as the sulfonate donor (WP_TPST + WPHV_C + PAPS) was established. Several control groups were also established: (1) buffer only, (2) individual components (WP_TPST, WPHV_C, or PAPS), and (3) combinations of two components (WP_TPST + WPHV_C, WP_TPST + PAPS, or WPHV_C + PAPS). ELISA was used in this study to detect the tyrosine sulfation modification of WPHV_C. The method is based on the principle that tyrosine residues modified by sulfation are specifically recognized and bound by an anti-sulfotyrosine antibody. The results revealed that an antibody positive signal was detected only in the complete sulfonation catalytic system (WPHV_C + WP_TPST + PAPS) (Fig. 5D), indicating that WP_TPST successfully catalyzed the formation of tyrosine sulfate groups on WPHV_C and that these modifications were specifically recognized by the antibody. Despite the relatively low absorbance values of the ELISA, it was preliminarily confirmed that WP_TPST possessed catalytic activity for the natural tyrosine sulfation modification of the WPHV_C protein. To assess the effect of WP_TPST-mediated sulfation, the anticoagulant activity of WPHV_C before and after sulfation was measured via the TT assay, as the introduction of sulfonate groups may increase its binding affinity to thrombin (Fig. 5E). These results indicate that neither WP_TPST nor PAPS alone, nor their combination, has anticoagulant activity. Combinations of WPHV_C with either WP_TPST or PAPS did not increase their anticoagulant activity. Surprisingly, in the complete reaction system, a 10 s prolongation of the TT of WPHV_C was observed, which also indicates that WP_TPST successfully catalyzed the tyrosine sulfation of WPHV_C.

Design and in vitro and in vivo anticoagulant activity of the lead compound WPHVC_V1

Based on the analysis of the binding mechanism between WPHV_C and thrombin, acidic and aromatic amino acid modifications were introduced into its structure to design a new anticoagulant lead compound, WPHVC_V1. Specifically, residue 2Y was mutated to 2E, a highly conserved glutamate commonly found in native hirudin and its variants. This modification increased the overall acidity of the protein. In addition, aromatic amino acids were introduced at positions 5A, 8Q, and 26 K to strengthen π-π stacking and π-cation interactions with thrombin. The acidic residues at positions 22D and 24D were retained to maintain stable electrostatic interactions with thrombin and to improve solubility. MD simulations indicated that the ΔG required to dissociate WPHVC_V1 from thrombin was 37.24 kcal·mol⁻¹, which was substantially greater than the 19.27 kcal·mol⁻¹ observed for WPHV_C (Fig. 6A).

Fig. 6.

Prediction of thrombin binding and in vitro and in vivo anticoagulant activity of the lead compound WPHVC_V1. (A) Energies of WPHVC_V1 binding with thrombin. (B-D) Standard coagulation assay for TT, APTT, and PT of WPHV_C, WPHVC_V1, and heparin sodium. The light blue dashed line represents the CK reference value. (E) Effects of WPHV_C, WPHVC_V1, and heparin sodium on thrombus formation in the tails of mice 24 h after carrageenan injection. (F) Quantification of thrombus tail length in the mice. (G) In vivo anticoagulant activity of WPHV_C, WPHVC_V1 and heparin sodium was assessed by carotid blood flow after FeCl₃-induced carotid thrombosis. Red represents blood flow, green represents low blood flow, and blue represents no blood flow. (H) The bleeding time of the mouse tail incision was recorded every eight seconds. (I-K) Effects of WPHV_C, WPHVC_V1, and heparin sodium administration for 30 min on TT, APTT, and PT in isolated plasma from mice. n = 5; the error bars indicate the standard deviation. *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001, ns: p > 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In the in vitro experiments, TT, APTT, and PT were measured for WPHV_C, WPHVC_V1, and heparin sodium (140 U/mg) at final concentrations of 0.025, 0.05, and 0.1 mg/ml, respectively (Fig. 6B, 6C, and 6D). Under identical conditions, WPHVC_V1 demonstrated superior anticoagulant effects compared with WPHV_C. At 0.1 mg/ml, the TT, APTT, and PT values for WPHV_C were 41.72 s, 14.38 s, and 15.86 s, respectively. In contrast, the corresponding values for WPHVC_V1 increased to 62.08 s, 23.38 s, and 22.22 s. The CK values were 18.52 s, 8.92 s, and 11.48 s, respectively. Although the anticoagulant effect of WPHVC_V1 was slightly lower than that of heparin sodium in the TT assay, it showed a modest advantage in APTT and PT measurements.

To evaluate the in vivo anticoagulant activity of the lead compounds, a carrageenan-induced mouse thrombosis model was employed. Thrombus formation was assessed by measuring the length of thrombi in the tails of the mice (Fig. 6E and 6F, Fig. S4). The results and statistical analyses demonstrated that the recombinant proteins WPHV_C and WPHVC_V1 significantly inhibited thrombus formation. Compared with the CK treatment (3.562 cm), the WPHV_C and WPHVC_V1 treatments reduced the average thrombus length to 2.530 cm and 1.853 cm, respectively. The lead compound WPHVC_V1 had a significantly greater effect than the WT, while its anticoagulant efficacy was comparable to that of heparin sodium (1.729 cm).

In the FeCl₃-induced mouse carotid artery thrombosis model, the inhibitory effects of WPHV_C and WPHVC_V1 on thrombus formation were evaluated by monitoring carotid blood flow (Fig. 6G). The results revealed that in the CK group injected with saline, carotid blood flow was interrupted after FeCl₃ treatment because of complete thrombus formation. The WPHV_C group showed initial thrombus inhibition, but some animals still showed incomplete thrombus formation with reduced blood flow. In contrast, thrombus formation was completely prevented in both the WPHVC_V1 group and the heparin sodium group, and unobstructed vessel flow was maintained throughout the observation period without any interruption or reduction in blood flow.

The tail bleeding time results in the mice indicated that WPHV_C and WPHVC_V1 significantly prolonged the bleeding time (Fig. 6H), demonstrating their ability to inhibit blood coagulation in vivo. Compared with CK, WPHV_C and WPHVC_V1 extended the bleeding time by 33.6 and 75.2 s, respectively. Although the bleeding time with heparin sodium was 35.2 s longer than that with WPHVC_V1, the difference was not statistically significant (p = 0.11).

Coagulation parameters in mouse plasma were monitored at 30 min, 2 h and 4 h after administration. At 30 min after administration, in vivo thrombin activity parameters revealed that (Fig. 6I, 6 J, and 6 K), under normal physiological conditions (CK group), the average TT/APTT/PT values of mouse plasma were 15.38/10.82/8.16 s. After administration of WPHV_C, these three indicators were significantly prolonged to 23.72/14.50/10.44 s (p < 0.01), indicating that thrombin activity was effectively inhibited. WPHVC_V1 showed even greater thrombin inhibition, with TT/APTT/PT values reaching 31.94/20.06/14.34 s, which was significantly different from that of the WPHV_C group (p < 0.01) and not significantly different from that of the heparin sodium group (32.60/18.32/14.20 s, p > 0.05). These findings demonstrated that the anticoagulant activity of WPHVC_V1 was significantly superior to that of the natural structure. At 2 h after administration, almost all the anticoagulant activity of the WT sequence WPHV_C was lost, and the anticoagulant activity of the designed WPHVC_V1 was also significantly reduced (Fig. S5). At 4 h after administration, the anticoagulant activity of WPHVC_V1 was lost, and the anticoagulant activity of the positive control heparin sodium was also significantly reduced, suggesting that the drug may have been gradually eliminated.

Initial validation of WPHVC_V1 features and security

With the in vitro and in vivo anticoagulant effects of the designed lead compound WPHVC_V1 confirmed, its safety was preliminarily evaluated both in vitro and in vivo. First, mouse RBCs (2 %) were used to test whether WPHVC_V1 would cause hemolysis in vitro. The RBCs were incubated with saline, WPHVC_V1, or Triton X-100 at 37 °C for 4 h. After centrifugation, the supernatant was collected, and its absorbance was measured at 580 nm (Fig. 7A). Triton X-100, a surfactant used for cell lysis that disrupts red blood cells, results in high absorbance in the supernatant due to the release of abundant hemoglobin. In contrast, no significant increase in absorbance was observed in either the saline or WPHVC_V1 groups, indicating that WPHVC_V1 did not cause a hemolytic reaction. The effect of WPHVC_V1 on blood pressure in mice was then assessed. Compared with the saline control, WPHVC_V1 at the concentration studied did not induce significant changes in blood pressure (Fig. 7B). The systolic blood pressure (SBP) ranged from 114 to 121 mmHg, and the diastolic blood pressure (DBP) ranged from 69 to 73 mmHg, which are within the normal range. The organ sections from the WPHVC_V1 group and saline group revealed that at the concentration and duration studied, no obvious pathological changes or side effects were observed in the major organs, including the heart, liver, spleen, lung and kidney (Fig. 7C). In addition, blood clots prepared from mouse blood were incubated in vitro with WPHVC_V1, and compared with saline, WPHVC_V1 did not exhibit significant in vitro thrombolytic activity (Fig. S6).

Fig. 7.

Initial validation of WPHVC_V1 security. (A) In vitro hemolysis evaluation. Absorbance of the supernatant of the mouse RBCs at 580 nm at 4 h postincubation with WPHVC_V1. (B) The tail-cuff method was used to measure blood pressure in mice after the administration of WPHVC_V1. (C) Histological examination. Representative H&E-stained images of normal tissues from mice postinjection of WPHVC_V1. The scale bar is 100 μm. n = 3; the error bars indicate the standard deviation. ****: p < 0.0001, ns: p > 0.05.

Discussion

Although oral anticoagulants have been widely used for long-term therapy, there is an urgent need for effective, rapid and safe parenteral DTIs for the treatment of acute thrombotic events such as deep vein thrombosis and pulmonary embolism [44,45]. Full-length hirudin, a classic natural DTI, exerts its anticoagulant effects via dual binding sites at its N-terminus (irreversible inhibition of the thrombin catalytic center) and C-terminus (competitive binding of thrombin exosite I). The rigid, disulfide-bonded N-terminus irreversibly occupies the thrombin active site and lacks a rapid neutralization method (such as protamine to neutralize heparin [46]), making potential bleeding risk a concern [47,48]. In this study, a novel hirudin variant was identified in W. pigra. By selectively removing the N-terminal domain and retaining and optimizing the C-terminal competitive binding region, we developed the lead compound WPHVC_V1, which avoids irreversible inhibition of the thrombin catalytic center at the molecular structural level.

WPHV_C competitively binds thrombin at the same site as fibrinogen, inhibiting fibrinogen binding and exerting anticoagulant effects. Potentially active amino acid sites were predicted by saturation mutagenesis, and the US method provided accurate MD simulations of WPHV_C-thrombin binding [32]. Mutant recombinant proteins were then produced and validated via in vitro assays. Nine acidic amino acids were found to play critical roles in the WPHV_C-thrombin interaction [13]. In particular, 13E formed the most electrostatic interactions with basic residues of thrombin, and mutation of 13E to alanine significantly reduced both the ΔG and anticoagulant activity.

In addition to electrostatic forces, the introduction of aromatic amino acids (A5F, Q8W, D22W, D24W, and K26W) at specific sites significantly stabilized WPHV_C-thrombin binding, probably via π-π stacking and π-cation interactions [49,50]. These substitutions increased the ΔG values and prolonged TT in vitro by 3–10 s at 0.1 mg/ml. However, replacement of 22D and 24D with aromatic residues offered limited improvements in ΔG and anticoagulant activity, suggesting that electrostatic interactions remain essential. However, discrepancies between ΔG values and in vitro TT assays arose because the MD simulations involved only the WPHV_C-thrombin binary complex, whereas plasma contains fibrinogen and other competing thrombin receptors. Nevertheless, the overall trends of the aromatic amino acid mutants at different concentrations were consistent: an improved ΔG was correlated with increased anticoagulant activity, confirming both the importance of aromatic residues and the reliability of the MD simulation predictions.

Tyrosine sulfation at the C-terminus of hirudin is critical for enhancing anticoagulant activity through strong electrostatic interactions (via robust salt bridges and hydrogen bonds) [51]. In WPHV_C, 20Y is the only residue susceptible to sulfation. In addition to sulfation, the aromatic ring of 20Y contributes additional binding strength to thrombin through π-π stacking and π-cation interactions. MD simulations and TT assays revealed that replacing 20Y with alanine significantly reduced anticoagulant activity, highlighting its importance. Although the introduction of acidic amino acids at position 20 created new electrostatic interactions, they could not compensate for the loss of the aromatic π-electron system. Research has shown that the electrostatic effects and spatial accessibility of sulfate groups introduced by tyrosine sulfation may be able to surpass those provided by naturally acidic side chains [52]. Thus, preserving the aromatic ring of 20Y while adding sulfation results in a synergistic enhancement. Therefore, achieving tyrosine sulfation is considered optimal for improving the anticoagulant activity of WPHV_C.

Artificial tyrosine sulfation of hirudin remains a challenge. New synthetic approaches, such as expanding the genetic code to include sulfate groups, often have low efficiency and cytotoxicity [53,54]. Recently, tyrosylprotein sulfotransferase (TPST) was shown to catalyze protein sulfation in vitro, suggesting a practical method to sulfonate the C-terminal tyrosine of hirudin [37]. WP_TPST from W. pigra was therefore used to catalyze WPHV_C sulfation in vitro. Molecular modeling predicted that electrostatic interactions primarily govern WP_TPST recognition of WPHV_C [34,55]. The abundance of acidic residues and the chain-like structure of WPHV_C facilitate the access of TPST to the sulfation site, thereby increasing the catalytic efficiency. Specific antibody ELISA and TT assays confirmed that WP_TPST transfers the sulfate group from PAPS to 20Y on WPHV_C, thereby enhancing the electrostatic interactions between WPHV_C and thrombin and improving its anticoagulant activity. Although recombinant WP_TPST expressed in E. coli has relatively low enzymatic activity, further optimization could make this a viable strategy for tyrosine sulfation of recombinant hirudin proteins.

By introducing multiple aromatic amino acid mutations, the binding affinity of WPHV_C for thrombin was increased, leading to the design of a novel lead peptide, WPHVC_V1. MD simulations revealed that WPHVC_V1 required a greater ΔG to dissociate from thrombin than WPHV_C, indicating a more stable complex. In vitro coagulation assays (TT, APTT and PT) revealed that both WPHV_C and WPHVC_V1 had good anticoagulant activity, with WPHVC_V1 showing greater potency. TT directly reflects the efficiency of thrombin-catalyzed fibrinogen conversion, whereas APTT and PT measure the initial phases of the intrinsic and extrinsic pathways, both of which converge on thrombin generation. Thus, WPHV_C and WPHVC_V1 prolong TT, APTT and PT by inhibiting the interaction between thrombin and fibrinogen. In contrast, heparin sodium indirectly suppresses thrombin by increasing antithrombin III (ATIII) activity, which also prolongs these parameters [56]. In vivo experiments in mice confirmed that 30 min after administration, the plasma coagulation parameters followed the same trend as the in vitro tests, indicating effective thrombin inhibition. The anticoagulant activity of WPHVC_V1 was significantly greater than that of the wild type, reinforcing the efficacy of the introduction of multiple aromatic amino acids into the C-terminal region of hirudin. However, the anticoagulant activity of both WPHV_C and WPHVC_V1 decreased significantly at 2 h post dose and no effect were observed at 4 h, which is in agreement with the literature reporting the short (1–2 h) in vivo half-lives of hirudin derivatives (e.g., recombinant hirudin and bivalirudin) [47]. Further pharmacokinetic studies in higher animal models (rabbits, dogs or nonhuman primates) are needed to elucidate their absorption, distribution, metabolism and excretion (ADME) and provide essential data for clinical translation.

Two acute thrombosis models with different mechanisms and sites of action were used to evaluate the in vivo anticoagulant activity of WPHV_C and WPHVC_V1. In the carrageenan-induced mouse tail thrombosis model, a classic model of acute venous thrombosis induced by inflammation, the mean thrombus length inhibited in the WPHVC_V1 group compared with the CK group was reduced from 3.562 cm to 1.853 cm, which was close to that of the heparin sodium group (1.729 cm) and significantly lower than that of the WPHV_C group (2.530 cm). In a FeCl₃-induced mouse carotid artery thrombosis model, which simulates acute arterial thrombosis via iron ions, oxidative stress and endothelial injury were induced. WPHV_C reduced clot formation but still left some vessels with reduced blood flow. In contrast, both WPHVC_V1 and heparin sodium completely prevented thrombus formation and maintained normal blood flow. Overall, these results suggest that WPHVC_V1 has good anticoagulant potential in both venous and arterial acute thrombosis scenarios.

Crystal structure studies have shown that the three-dimensional conformations and active site binding patterns of human thrombin and rodent thrombin are similar [57,58]. This study confirmed the anticoagulant activity of WPHV_C and its derivative WPHVC_V1 through in vitro rabbit plasma experiments and in vivo mouse models. However, the mammalian coagulation cascade is complex, and species-specific differences affect the sensitivity of thrombin generation to stimulatory factors such as tissue factors [59]. Human thrombogenesis also involves unique regulatory networks, including platelet-thrombin feedback and stabilization of fibrin cross-linking, which are difficult to replicate in animal models. These factors may lead to discrepancies between effective doses in animal studies and human therapeutic needs. Therefore, future research should establish humanized evaluation systems, including in vitro human plasma assays and transgenic animal models, which are critical for clinical translation and minimizing therapeutic risks.

After confirming the in vitro and in vivo anticoagulant activities of the lead compound WPHVC_V1, a preliminary safety evaluation was conducted. In vitro tests revealed no hemolysis within the observation period, and histopathological analysis of major organs from treated mice revealed no significant pathological changes. However, the balance between anticoagulant efficacy and bleeding risk remains a key challenge. In a mouse tail incision experiment, although the bleeding time of the WPHVC_V1 group was 35.2 s shorter than that of the sodium heparin group (p = 0.1077, not statistically significant), there was still a trend toward prolonged bleeding. Future studies in more advanced animal models are needed to systematically evaluate the therapeutic potential and bleeding risk of WPHVC_V1.

WPHVC_V1 is a novel hirudin derivative that retains classical targeting of thrombin exosite I via acidic residues. By introducing aromatic amino acid mutations, it gains additional π-π stacking and π-cation interactions, improving its complex stability with thrombin. Compared with full-length hirudin, WPHVC_V1 may reduce the risk of bleeding by avoiding irreversible inhibition of the catalytic center of thrombin, and its competitive binding to exosite I offers the possibility of developing a targeted antidote. As a DTI candidate, WPHVC_V1 blocks thrombin-mediated fibrinogen hydrolysis in vitro. Thrombin not only catalyzes fibrin formation, but also promotes coagulation by activating factor XI (FXI), factor V (FV) and platelets, and triggers inflammatory responses and protease-activated receptor (PAR) signaling, which can lead to cardiovascular complications [60,61]. While oral anticoagulants (e.g., rivaroxaban) effectively prevent thrombosis by inhibiting upstream coagulation factors, they cannot neutralize preformed thrombin [62]. In contrast, the direct inhibitory properties of WPHVC_V1 allow it to disrupt existing thrombin, reduce positive feedback loops and attenuate inflammation, endothelial dysfunction and vascular permeability via PAR inhibition [63,64], making it well suited for acute thrombosis or high-thrombin scenarios. As an injectable DTI with rapid onset of action, WPHVC_V1 is suitable for acute anticoagulant needs, such as perioperative use. Oral DTIs such as dabigatran, which require metabolic activation through the gastrointestinal tract, are more suitable for long-term anticoagulation [65]. WPHVC_V1 also offers advantages over heparin, including a lower molecular weight, a well-defined chemical structure and a predictable dose–response relationship. Although the benefits of WPHVC_V1 have been assessed mainly through theoretical analyses of its molecular structure and preliminary analysis of its pharmacodynamic activity, comprehensive pharmacodynamic and pharmacokinetic studies, together with a detailed assessment of anticoagulant efficacy, bleeding risk and safety, are needed to fully explore its clinical potential.

Conclusion

The detailed investigation of the anticoagulant mechanism of WPHV_C confirmed the critical role of acidic amino acids in binding to thrombin exosite I and revealed the importance of the π-electron cloud system of aromatic amino acids in this process. These findings provide valuable guidance for developing novel anticoagulant peptides or optimizing the structure of natural proteins. Additionally, the sulfation modification of the C-terminal tyrosine was investigated, and preliminary in vitro experiments demonstrated that sulfation catalyzed by TPST enzymes significantly enhanced its anticoagulant activity.

Based on the elucidated anticoagulant mechanism, a lead compound, WPHVC_V1, was designed, and its anticoagulant effects were validated both in vitro and in vivo. Preliminary validation of its drug safety is needed. The compound also presented advantages such as a low molecular weight, simple structure, and high solubility. However, the study models had certain limitations, and further translational research is needed to confirm the long-term safety and efficacy of the lead compound.

CRediT authorship contribution statement

Xiaozhe Yi: Data curation, Formal analysis, Software, Visualization, Writing – original draft. Xiaoli Wu: Writing – review & editing. Erhuan Zang: Writing – review & editing. Xiaoli He: Formal analysis, Methodology. Xinyi Chang: Methodology. Jinxin Liu: Conceptualization, Project administration, Supervision. Linchun Shi: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing.

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.

Appendix A. Supplementary data

The following are the Supplementary data to this article: