Molecular Insights and Functional Consequences of the Interaction of Heme with Activated Protein C

Abstract

Aims:

In hemolysis, which is accompanied by increased levels of labile redox-active heme and is often associated with hemostatic abnormalities, a decreased activity of activated protein C (APC) is routinely detected. APC is a versatile enzyme that exerts its anticoagulant function through inactivation of clotting factors Va and VIIIa. APC has not been demonstrated to be affected by heme as described for other clotting factors and, thus, is a subject of investigation.

Results:

We report the interaction of heme with APC and its impact on the protein function by employing spectroscopic and physiologically relevant methods. Binding of heme to APC results in inhibition of its amidolytic and anticoagulant activity, increase of the peroxidase-like activity of heme, and protection of human umbilical vein endothelial cells from heme-induced hyperpermeability. To define the sites that are responsible for heme binding, we mapped the surface of APC for potential heme-binding motifs. T²⁸⁵GWGYHSSR²⁹³ and W³⁸⁷IHGHIRDK³⁹⁵, both located on the basic exosite, turned out as potential heme-binding sites. Molecular docking employing a homology model of full-length APC indicated Tyr²⁸⁹ and His³⁹¹ as the Fe(III)-coordinating amino acids.

Innovation:

The results strongly suggest that hemolysis-derived heme may directly influence the protein C pathway through binding to APC, conceivably explaining the decreased activity of APC under hemolytic conditions. Further, these results extend our understanding of heme as a multifaceted effector molecule within coagulation and may allow for an improved understanding of disease development in hemostasis under hemolytic conditions.

Conclusion:

Our study identifies APC as a heme-binding protein and provides insights into the functional consequences.

Introduction

The redox active molecule heme is known for its essential role as the prosthetic group in a number of hemoproteins, such as hemoglobin and cytochromes. Each of up to 25 trillion circulating red blood cells (RBCs) in the human body is considered to contain 2.5 × 10⁸ hemoglobin molecules with 5 × 10⁹ molecules of heme bound, which roughly corresponds to 80 mM heme (9, 10). Under hemolytic conditions, heme is released as a result of erythrocyte lysis and subsequent degradation of hemoglobin, which, in turn, may lead to repletion of the heme scavenger hemopexin (57). Heme may also adhere to other plasma proteins, such as albumin and α1-microglobulin, resulting in a pool of so-called “labile heme” (63).

Labile heme is well known as a regulator of various proteins in a broad range of biochemical and pathophysiological processes, such as transcription, signal transduction, inflammation, or hemostasis, but it is still unresolved as to what extent it might influence coagulation processes, for example, in hemolytic states (19, 37, 57).

Intravascular hemolysis has been reported to increase the thrombotic risk (1, 44). This can be partially explained by hemolysis-induced coagulation activation through exposure of negatively charged phospholipids from ruptured RBCs, induction of tissue factor expression, and stimulation of von Willebrand factor release from endothelial cells (37, 57). However, the underlying mechanisms of these events have not yet been elucidated (1, 16, 54).

Interference of heme with endogenous inhibiting systems such as the protein C (PC) pathway, might contribute to the hemolysis-induced hypercoagulable state. PC is the zymogen of the serine protease activated PC (APC) that is synthesized in a vitamin K-dependent manner in the liver and circulates at a concentration of 70 nM in human blood (23). PC consists of a heavy and a light chain connected by a disulfide bond (32). The light chain is composed of a Gla domain and two epidermal growth factor (EGF)-like domains, whereas the heavy chain comprises the serine protease domain. Depending on the glycosylation pattern of the heavy chain, different molecular forms (α, β, and γ) of PC exist (20).

PC becomes activated to APC through limited proteolysis and cleavage of the activation peptide by the thrombin-thrombomodulin complex on the luminal surface of endothelial cells. Once formed, APC downregulates further thrombin formation by proteolytic inactivation of the procoagulant cofactors activated factor VIII (FVIIIa) and activated factor V (FVa) (56). In addition, APC can exert pleiotropic cytoprotective effects, including the activation of antiapoptotic and anti inflammatory pathways enhancing, for example, endothelial barrier integrity. These effects are induced by APC-mediated activation of the G-protein coupled protease-activated receptors 1/3 (PAR1/3) and involve binding of APC to the endothelial protein C receptor (EPCR) (21).

Innovation

Several coagulation factors have been earlier described to be affected by heme, which, together with the occurrence of coagulopathies during hemolytic events, inevitably suggests a link between hemolysis and coagulation. For the first time, we demonstrate transient heme binding to activated protein C (APC), an anticoagulant serine protease, and identify two potential binding sites. Further, our findings on the functional consequences for APC on heme binding led us to assume that a heme overload under hemolytic conditions might directly affect the protein C pathway and, consequently, exerts an effect on clotting processes.

The increased thrombotic risk of inherited PC deficiency, inherited protein S (PS) deficiency, and APC resistance emphasize the importance of this anticoagulant pathway (3, 43, 66). In addition, Gabre et al. reported that APC can induce the expression of heme oxygenase 1 (HO-1), an enzyme that catalyzes the degradation of heme (17). Earlier, however, an upregulation of PC and thrombomodulin by HO-1 was shown by Fei et al. (13). Continuing the line, in patients with sickle cell disease (SCD) accompanied by severe hemolysis, decreased levels of APC were partially reported based on the amidolytic activity of the enzyme, however, without describing the underlying mechanism (60, 67).

In previous studies, we identified potential heme-regulated proteins such as interleukin-36α and cystathionine-β-synthase, which were unknown to bind heme (6, 39, 49, 61, 70). Herein, we provide evidence that human APC is a heme-binding protein and thus might represent a heme-regulated protein, as well.

Surface plasmon resonance (SPR) spectroscopy and in silico studies were used to characterize the respective interaction. Potential heme-binding sites on the surface of APC were identified and analyzed by ultraviolet/visible (UV/vis) spectroscopy. To visualize heme binding to APC, we employed molecular modeling and docking studies by using an in-house generated homology model. Moreover, we demonstrated an impact of heme on the amidolytic and anticoagulant function of APC and observed an increased peroxidase-like activity of the heme-APC complex. Apart from that, the cytoprotective functions of APC were not influenced by heme binding as human umbilical vein endothelial cells (HUVECs) were protected from heme-induced cytotoxicity in the presence of APC.

Taken together, our data provide novel valuable insights into the regulation of the anticoagulant protein APC by heme, which may contribute to our understanding of the physiological functions of APC as well as the development of hemolytic diseases associated with hemostatic abnormalities.

Results

Heme binds to human APC

To demonstrate and characterize the binding of heme to APC, SPR measurements of the heme-APC interaction were performed. The analysis of the association of heme with APC by using a global 1:1 kinetic model revealed a equilibrium dissociation constant (K _D) value in the nanomolar range (400 ± 47 nM) according to the best fit (Fig. 1). The binding of heme to APC is transient, which is characterized by a rapid association (association rate constant [k _a] = 2.06 × 10⁴ m⁻¹·s⁻¹) and dissociation process (dissociation rate constant [k _d] = 8.08 × 10⁻³·s⁻¹). A stoichiometry of ∼1:1.3 (protein:heme) was determined, which indicates either overlaying binding events of heme to different binding sites, nonspecific binding, or both.

FIG. 1.

Heme binds to APC. Single-cycle kinetic SPR analysis of the heme interaction (187.5–3000 nM) with APC revealed at least one specific heme-binding site with a binding affinity in the nanomolar range (K _D ∼ 400 ± 47 nM). The K _D value was calculated by using a kinetic global fit (red curve). Moreover, the determined stoichiometry of 1:1.3 indicates that there are overlaying binding events of specific and/or unspecific interactions. The fast association (k _a ∼ 2.06 × 104 s⁻¹) and dissociation rates (k _d ∼ 8.08 × 10⁻³ s⁻¹) refer to a transient interaction of heme and APC. Arrows indicate the five consecutive injections of heme in different concentrations which are noted above. APC, activated protein C; k _a, association rate constant; k _d, dissociation rate constant; K _D, equilibrium dissociation constant; SPR, surface plasmon resonance. Color images are available online.

Heme binds to APC-derived peptides

Previously, it was shown that nonapeptides represent suitable models for analyzing heme-regulatory/binding motifs (HRMs/HBMs) on the surface of proteins (39, 40, 49). Four APC-derived nonapeptides were included in this study (peptides 1–4) (Supplementary Table S1), which were predicted by the in-house developed webserver HeMoQuest (48) based on earlier established criteria for heme binding to proteins (5, 7, 38, 39). These peptides, containing either histidine and/or tyrosine as the coordinating residue, were synthesized and analyzed for their heme-binding capacity by UV/vis spectroscopy.

Heme binding was observed by a Soret band shift to ∼420 nm for all peptides (Fig. 2 and Supplementary Fig. S1). It was previously demonstrated by different spectroscopic methods, including resonance Raman spectroscopy, continuous-wave electron spin resonance spectroscopy, and two-dimensional nuclear magnetic resonance spectroscopy, that a shift to ∼420 nm is typical for a mixture of a hexa- and pentacoordination or hexacoordination of a heme-peptide/protein complex (5, 7, 39, 61). According to the best fit, a stoichiometry of 1:1 was determined for all heme-peptide complexes investigated herein. The highest heme-binding affinities were observed for peptides 1 (TGWGYHSSR) and 2 (WIHGHIRDK) displaying K _D values of 0.46 ± 0.25 μM and 1.65 ± 0.51 μM, respectively.

FIG. 2.

Identification of HRMs on the surface of APC. (A) Homology model of APC depicted as a ribbon structure (left) and a space-filling model (right), with the heavy chain in red, and the light chain in blue with calcium ions (yellow). (B) Sequence of APC divided into light and heavy chain. HRMs are highlighted in green. (C, D) Motifs 1 (TGWGYHSSR) and 2 (WIHGHIRDK) are found within the heavy chain, here shown as a zoom-in into the whole structure of APC. The potential coordination sites are accentuated by stick bonds (top). The affinity for heme binding to these stretches was estimated by UV/vis binding studies of heme to motif-derived nonapeptides. These peptides (each 10 μM) were incubated with heme (0–40 μM). Depicted are the difference spectra. For both peptides, a high affinity for heme was demonstrated (bottom). HRMs, heme-regulatory motifs. Color images are available online.

The sequence stretch of peptide 1 is located close to the active site of APC and found within a coiled structure (Fig. 2A–C). In contrast, the motif of peptide 2 is part of a helix, further away of the active site and C-terminal of the heavy chain, but within the basic exosite of the protein (Fig. 2A, B, D). Two further potential motifs (peptides 3 and 4) (Supplementary Table S1) revealed only moderate affinities for heme and, in contrast to 1 and 2, they are distant from the active site. The motifs 1 and 2, thus, represented promising candidates for potential heme-binding sites in APC.

In silico study of heme-APC complex via molecular docking and molecular dynamics simulations

In silico structural modeling of APC was employed in this study, because an experimentally derived structure of the full-length APC has not been available so far (45). Earlier reported homology models of APC were performed for either the serine protease domain or the zymogen form of PC (14, 51).

In our work, the five top-ranked templates used to build the initial homology models of APC were PDB ID: 1AUT, 3F6U, 2AER, 1W0Y, 3HPT, and they had an alignment target coverage of 83%, 83%, 94%, 93%, and 80%, respectively. The final hybrid model (Fig. 2A) consisted of 405 residues [full coverage of APC(1–405), i.e., without activation peptide] and included contributions from the following initial models: 1AUT-∼(39–393), 1W0Y-∼05(1–52), 1W0Y-∼02(68–76), and 2AER-∼01(394–406) (Supplementary Table S2). To evaluate the quality of the model, Z-scores, which indicate the total deviation in energy against energy distributions derived from entirely random conformations, were used as a metric. In an ideal situation, a value of 0 is expected but in most real-world cases values close to 0 are acceptable (68). Accordingly, the overall Z-score obtained as a weighted average of the individual Z-scores of three distinct components, namely, dihedrals, one-dimensional packing, and three-dimensional (3D) packing, was found to be −1.407. Ramachandran plot analysis of the structure showed that 87.2% of the models' residues were in the most favored regions, 12.2% of residues in the allowed region, and a meagre 0.6% of the residues in the disallowed region, indicating that the final model obtained was of good quality.

As a result of focused molecular docking experiments on the HRMs 1–4, clusters of docking complexes were automatically ranked and sorted by the Autodock Vina scoring function used (65). Each cluster was then manually inspected to select only those complexes that had coordination interactions between potential HRMs and a surface-bound heme molecule (based on a distance cutoff of 3 Å) that suggests transient heme binding. Our analysis revealed that two (motifs 1 and 2) out the four investigated motifs exhibited an excellent heme-binding potential based on favorable heme-motif distances and the relative strength of their association quantified by Poisson-Boltzmann binding energies.

The stability of the heme-motif interaction for the motifs 1 and 2 was evaluated based on the observed residence of heme at the motif in APC through the entire course of the 50 nanoseconds simulation runs, quantified by the distances between the Fe³⁺ ion of heme and the side chain oxygen of the Tyr²⁸⁹ residue (motif 1) and the side chain nitrogen atom of the coordinating His³⁹¹ residues (motif 2), respectively (Supplementary Table S3). For motif 1 this average distance (between Tyr²⁸⁹ and the Fe³⁺ ion of heme) was 2.16 Å and the Poisson-Boltzmann binding energy (see Materials and Methods section) estimated to quantify the heme-motif association was −168.08 kJ/mol. Motif 2 displayed similar positive binding characteristics, as the distance between the His³⁹¹ residue of the motif was at an average of 2.07 Å and a consequent binding energy of −255.49 kJ/mol. These observations serve as corroboration to the experimental results at the peptide level. Based on the distances discussed herein, though it might be the case that both the motifs 1 and 2 stand a similar chance and potential of heme coordination, the binding energy analysis suggests that motif 1 with a more positive value (−168.08 kJ/mol) might have a higher preference over motif 2 (−255.49 kJ/mol), since the binding energy quantifies the strength of the association of heme to its respective motifs. On the other hand, observations from motifs 3 and 4 showed clear indications of poorer heme-binding capabilities (Supplementary Table S3 and Supplementary Fig. S2). Motif 3, despite holding heme at an average distance of 2.18 Å away from the motif, had a much lower binding energy of −326.51 kJ/mol. The average heme-motif distance among all the four investigated motifs was the highest for motif 4 with an associated binding energy of −291.95 kJ/mol at 4.03 Å.

With the individual motifs investigated, a 50 nanoseconds molecular dynamics (MD) simulation of heme bound to both motif 1 and 2 showed that both the heme molecules stayed bound to the surface of APC with heme-motif distances under ∼3 Å over the course of the simulation (Fig. 3A and Supplementary Video S1). This suggests that APC has the capacity to bind more than one heme molecule and in theory can bind up to two heme molecules at a time.

FIG. 3.

Heme coordination by APC via Tyr²⁸⁹ and His³⁹¹ has a structural impact on APC. (A) Two heme molecules docked to APC through interaction with Tyr²⁸⁹ (right) and His³⁹¹ (left). The respective motifs are highlighted in green within the structure. (B) The per RMSF plots were computed from two different 300 nanoseconds long MD simulations run on APC. The RMSF plot from the simulation with no heme molecules bound to APC (black line) has a flexibility characteristic of most proteins, with the terminal ends of the protein showing higher mobility. In comparison, in case of APC with heme molecules bound to both motifs 1 and 2 by molecular docking, the entire protein has a pronounced loss of mobility as indicated by the RMSF plot (red line). MD, molecular dynamics; RMSF, residue root mean square fluctuation. Color images are available online.

The structural consequences of heme binding to APC were assessed via two separate 300 nanoseconds long MD simulations conducted on both the heme-bound and -unbound form of the protein. We observed that the association of the two heme molecules with APC had a marked impact on the dynamics of APC by inducing a significant reduction of the inherent flexibility/mobility of the protein. The per residue root mean square fluctuation (RMSF) of the residues was used as an MD-derived measure to compare the dynamics of the heme-bound and -unbound form of APC. The RMSF values of the two simulations clearly indicate that heme binding to APC has a dampening effect on the natural dynamics of the protein (Fig. 3B). This structural impact can be hypothesized to have a functional consequence on the protein.

Heme inhibits the amidolytic activity of APC

The amidolytic activity of APC was analyzed in the presence of different concentrations of p-Glu-Pro-Arg-MNA as the substrate, revealing a Michaelis-Menten constant (K_M) value of 0.42 ± 0.07 mM, which is in the same range as reported for other chromogenic peptide substrates of APC (Supplementary Fig. S3) (59). In the presence of heme, the inhibition of the APC activity by heme as a competitive inhibitor revealed an inhibitor constant (K_i) value of 12.56 ± 2.31 μM (Fig. 4A). In comparison, the amidolytic activity of APC was also studied in the presence of two other porphyrin analogs, namely zinc protoporphyrin IX (ZnPPIX) and protoporphyrin IX (PPIX) as well as iron(III) chloride at concentrations ranging from 0 to 50 μM. Heme inhibited the activity of plasma-derived APC (pdAPC) and recombinant APC (rAPC) in a dose-dependent manner with a half maximal inhibitory concentration (IC₅₀) value of 10.41 ± 1.24 μM and 3.88 ± 0.87 μM, respectively, according to the best fit (Fig. 4B, C). At the highest concentration of heme (50 μM) applied, an inhibition of 80% of pdAPC's activity was observed. A Hill slope of 2.0 suggests two binding sites for heme with positive cooperativity (n = 2). A similar effect toward pdAPC was observed after incubation with PPIX, revealing an IC₅₀ value of 8.46 ± 1.68 μM, n = 2 (Fig. 4D). ZnPPIX had an even stronger effect by decreasing the pdAPC activity to 1.42% (IC₅₀ = 4.33 ± 0.87 μM, n = 2) (Fig. 4E), which has been shown for zinc(II) ions as well by Zhu et al. (77). Iron (III) ions reduced the pdAPC activity only to 66.83% at a concentration of 50 μM (Fig. 4F). To compare the effect of heme on APC's activity with that on other enzymes of the blood coagulation cascade, the enzymatic activities of the serine protease thrombin and the transglutaminase factor XIIIa (activated factor XIII [FXIIIa]) were observed in the presence of heme. Both thrombin and FXIIIa were not affected by heme in their enzymatic activities (Fig. 4G, H). Indeed, these proteins were not detected as heme-binding proteins with the publicly available webserver HeMoQuest (48), indicating reliability of the program and the underlying search criteria regarding HBMs.

FIG. 4.

Effect of heme on the amidolytic activity of APC. (A) Evaluation of K_i (12.56 ± 2.31 μM) of heme (0–50 μM) for pdAPC (100 nM) with different substrate concentrations (● 0.05 mM, ■ 0.1 mM, ▼ 0.5 mM, ♦ 1 mM, ○ 2.5 mM) by using Dixon plot (reciprocal reaction rate ν plotted against increasing concentrations of heme as the inhibitor). Inhibition of the amidolytic activity of (B) pdAPC (100 nM), and (C) rAPC (50 nM) by heme (0–50 μM). IC₅₀ values were determined to be 10.41 ± 1.24 μM and 3.88 ± 0.87 μM, respectively. (D, E) Inhibition of pdAPC by heme derivatives PPIX and ZnPPIX with IC₅₀ values of 8.46 ± 1.68 μM and 4.33 ± 0.87 μM. (F) The effect of FeCl₃ on pdAPC in the same concentration range seems to be negligible due to the fact that only the highest concentration shows a minimal effect, reducing the activity to ∼66.83%. (G, H) In comparison to APC, ns effect of heme was observed on the enzymatic activity of α-thrombin (G) and rFXIIIa (H), as analyzed by applying the one-way ANOVA test. The black dashed lines represent the normalized activity of the enzymes without addition of heme (100%), respectively. Values are normalized to the enzymatic activity of the respective enzymes, renouncing the addition of porphyrins or iron ions. (I) Peroxidase-like activity of heme is increased in complex with APC ([APC+heme]) up to 512.33% ± 66.50%. The dashed line represents the normalized activity of heme (100%) and the heme-amyloid beta complex (∼500% (71); [Aβ+heme]) for comparison. ANOVA, analysis of variance; IC₅₀, half maximal inhibitory concentration; K_i, inhibitor constant; ns, not statistically significant; pdAPC, plasma-derived activated protein C; PPIX, protoporphyrin IX; rAPC, recombinant activated protein C; ZnPPIX, zinc protoporphyrin IX.

Heme-APC complex shows peroxidase-like activity

The peroxidase-like activity of heme-protein complexes has been reported earlier, with amyloid beta being the most prominent protein representative evoking a significantly increased activity (∼500%) compared with uncomplexed heme (2, 71). The activity of the heme-APC complex was determined by using an experimental set-up established earlier (2, 64, 71). Normalized against the basal heme activity, the heme-APC complex showed a more than fivefold higher activity (512.33 ± 66.50%; Fig. 4I) and, thus, is one of the most efficient molecules regarding this effect compared with the peptides and proteins that have been studied so far (28, 29, 64, 71).

Heme is able to inhibit the anticoagulant activity of APC

To investigate the impact of heme on the anticoagulant function of APC, the conventional activated partial thromboplastin time (aPTT)-based clotting assay was performed with plasma samples. Heme alone had no significant effect on aPPT, whereas APC (5 nM) prolonged the clotting time of normal plasma by approximately twofold (Fig. 5A). An incubation of APC with 1 μM heme did not change this effect. In contrast, the preincubation of APC with heme at 10 μM and higher concentrations caused a marked reduction of the clotting time, counteracting the anticoagulant function of APC. Thus, the clotting time was effectively shortened to that of normal plasma without addition of APC.

FIG. 5.

Heme-mediated inhibition of the anticoagulant function of APC. (A, B) Dose-dependent inhibition of pdAPC by heme, observed in a conventional aPTT-based clotting assay. Preincubated with APC (5 nM), the lowest effective heme concentration was 10 μM [without additional HSA, (A)] and 100 μM [in the presence of 0.1% HSA, (B)], respectively. (C, D) Further, the heme analogs ZnPPIX and PPIX were also capable of inhibiting the anticoagulant function of APC in both settings. Statistically significant differences in comparison to the sample without heme are shown according to one-way ANOVA. *p ≤ 0.05; **p < 0.01. aPTT, activated partial thromboplastin time; HSA, human serum albumin.

On preincubation with heme (up to 50 μM), the addition of exogenous albumin (0.1%) completely restored the clotting time to the level when only APC was applied (Fig. 5B). At 100 μM heme, however, a tendency toward inhibition of the anticoagulant function of APC was observed again (not statistically significant). Due to the solubility and high aggregation potential of heme in correlation with the required assay conditions, the effect of higher concentration could not be investigated despite the pathological relevance. Further studies demonstrated that heme could already induce a slight inhibition of APC's anticoagulant without any preincubation, which is more emphasized after a preincubation time of 15 min or longer (Supplementary Fig. S4).

Since the heme analogs ZnPPIX and PPIX were potent to inhibit the amidolytic activity of APC similar to heme, their impact on the anticoagulant function was investigated as well. Similar to heme, none of the analogs had a direct influence on aPTT. However, after preincubation with APC and the addition of this mixture to plasma, aPTT was shortened (Fig. 5C, D). In the presence of additional human serum albumin (HSA), an effect comparable to heme was observed (Fig. 5D). The lowest concentration of heme still exerts activity in the absence of HSA (i.e., 10 μM) and it seemed to be insufficient for the heme analogs (Fig. 5C). To achieve the same effect as heme in this approach, an increase of their concentration to 100 μM was required (Supplementary Fig. S5).

APC protects HUVECs from heme-mediated hyperpermeability

To study the influence of heme on the cytoprotective function of APC, an endothelial barrier protection assay with HUVECs was performed. Heme exposure of the HUVECs resulted in a concentration-dependent increase of fluorescein isothiocyanate (FITC)-dextran flux through the cell monolayer (Supplementary Fig. S6). In the presence of APC, the HUVECs were protected from heme-induced hyperpermeability. Therefore, the cytoprotective function of APC was not influenced by heme (Fig. 6).

FIG. 6.

APC's barrier protective function is not altered by heme. Heme exposure (120 μM) causes an increase of the endothelial cell monolayer permeability, which is completely prevented by a 30-min preincubation of heme with APC (20 nM). Thrombin (20 nM) alone significantly increases the permeability of the cell monolayer, whereas APC (10 nM) protects from thrombin-induced endothelial barrier permeability, which is determined here as fluorescence intensity as a measure for the flux of FITC-labeled dextran through the endothelial cell monolayer.**p < 0.01. FITC, fluorescein isothiocyanate; RFU, relative fluorescence units.

Discussion

This study provides evidence that heme binds to APC and thereby impacts its function (Fig. 7). We demonstrate a heme-binding affinity for APC in the nanomolar range, which is typically found for heme-regulated proteins, as can be exemplified with the transcription repressor Bach1 (K _D = 140 nM) or heme oxygenase 2 (K _D = 33 nM [dithiol state], K _D = 350 nM [disulfide state]) (47, 76). The transient manner of heme binding to APC is illustrated by both rapid association and dissociation of the heme molecule. A stoichiometry of 1:1.3–1:2 suggested one to two binding sites for heme with either overlaying binding events, different binding affinities of the two motifs, or a mixture of specific and unspecific binding events. pdAPC represents a mixture of different glycosylated forms of APC (α, β, and γ). Different accessibility of potential binding sites through a missing glycosylation may result in different binding properties, which, in turn, may potentially affect the stoichiometry. This is further evidenced by the computational studies using a combination of molecular docking and MD simulations, suggesting heme binding to both motifs 1 and 2 with a comparably high potency.

FIG. 7.

APC is the first clotting inhibitor that is reported to bind heme. Previously, there was no report on APC as a heme-binding protein. In this study, the binding of heme to APC was demonstrated with a K _D value in the nanomolar range. This resulted in a change of its amidolytic activity, whereas FIIa and FXIIIa were not affected in their activity in the presence of heme (dashed boxes). An impact of heme or hematin on the plasma levels or function of other coagulation factors was earlier reported (light red boxes) (19, 31, 57). FVIII(a) and fibrinogen are known to bind heme (dark red boxes, accentuated by the heme symbol) (31, 55). For other factors, there has been no link to heme binding reported so far (gray boxes). FII, prothrombin; FIIa, thrombin; FIX, factor IX; FIXa, activated factor IX; FMx, crosslinked fibrin; FV, factor V; FVa, activated factor V; FVIIa, activated factor VII; FVIII, factor VIII; FVIIIa, activated factor VIII; FX, factor X; FXa, activated factor X; FXI, factor FXI; FXIa, activated factor FXI; FXIII, factor XIII; FXIIIa, activated factor XIII; TM, thrombomodulin. Color images are available online.

We identified T²⁸⁵GWGYHSSR²⁹³ (motif 1) as the most probable heme-binding site on the surface of APC, with Tyr²⁸⁹ as the favorable coordination site and the highest affinity compared with other motifs. This motif is located close to the active site, which might explain the effect of heme on the amidolytic activity of APC, accomplished by either disturbing access to the active site or a conformational change of the protein infringing the integrity of the active site (Fig. 8). Based on experimental and MD simulation results, we suspect that W³⁸⁷IHGHIRDK³⁹⁵ (motif 2) with His³⁹¹ as the coordinating residue is one further high-affinity heme-binding site that could be responsible for overlaying binding effects, which were observed in the SPR studies or represent the second binding site as determined with the amidolytic activity assay in the presence of heme (Fig. 8).

FIG. 8.

Hemolysis-derived labile heme and its interaction with APC. (A) Under hemolytic conditions, the level of labile heme increases as a consequence of RBC rupture and subsequent degradation of hemoglobin (Hb) (57). This leads to simultaneous intravascular occurrence of labile heme and clotting factors. Consequently, heme can exert its regulatory function by affecting different coagulation factors (different gray circles) with respect to their plasma levels and/or their functional activity (19, 31, 57). (B) Regulatory heme binding to proteins occurs to short surface-exposed sequence stretches. In case of APC, the heme-binding site (green) is close to its catalytic site (red). Therefore, heme binding to APC occurs not through a direct binding to the catalytic site. There is, thus, no direct correlation between the heme-binding affinity of APC and the inhibitory effect of heme on APC. In contrast, heme binding, especially if involved in specific binding sites (e.g., FVa- and FVIIIa [PDB: 2r7e]-binding sites; blue), can induce conformational changes that might explain the effect on the activity and/or function of APC. (C) Mapping the interaction sites of APC with its interaction partners in comparison with the sites of the potential heme-binding motifs (green) reveals that motif 1 is very close to the catalytic triad (red) of APC, with high proximity to its active site serine residue (zoom-in, right). In addition, it is part of the important sites for interaction with the clotting factors FVa and FVIIIa (blue), which serves as an explanation for the inhibitory effect of heme on APC's anticoagulant function. Motif 2 (zoom-in, left) is closer to the interaction site of APC with PAR1 (brown), suggesting a potential relevance of heme for the cytoprotective function of APC. Further, interaction sites for phospholipids (yellow), EPCR (gray), PCI, and the glycosaminoglycan heparin (purple) are labeled (69). Interaction sites and respective interaction partners have the same color and are connected by dashed lines. Glycosylation sites are highlighted (magenta) (22). EPCR, endothelial protein C receptor; PAR1, protease-activated receptor 1; PCI, protein C inhibitor; RBC, red blood cell. Color images are available online.

Molecular docking studies allowed for the assumption that heme binding to both motifs is possible, which fits to the aforementioned results. Moreover, two other motifs showed moderate heme-binding properties based on experimental and computational approaches. Therefore, a recruiting binding process could be conceivable. The proposed binding sites (motif 1 and 2) were already described as potential binding sites for zinc ions of APC (77). Moreover, the sequence of motif 2 was earlier reported to be part of a sequence stretch that is crucial for the anticoagulant function of human APC (46). The potential coordinating site of motif 1 is highly conserved in APC from different species, whereas motif 2 is not only missing in some species but also highly conserved, if present (Supplementary Table S4). However, both motifs are part of the basic exosite, which is essential for the interaction of APC with FVa and FVIIIa and proper presentation of both factors to the active center (Fig. 8B, C) (56). Heme binding to APC may, thus, impair the catalytic activity of the protein and its interaction with other coagulation factors. Our MD-based observations indicate that the heme-induced conformational change could be initiated by a reduction of the protein flexibility. However, a clear characterization of the conformation change itself is not within the scope of this study, as this might happen at a much larger simulation time scale than 300 nanoseconds due to the size of the system (>250,000 atoms) and is solely hypothetical.

However, heme is, indeed, able to inhibit the amidolytic activity of rAPC and pdAPC with an IC₅₀ value of 3.88 ± 0.87 μM and 10.41 ± 1.24 μM, respectively. The recombinant form merely differs from the pdAPC in the glycosylation pattern (75). rAPC often shows a higher content of the highly active β-APC (diglycosylated heavy chain) and only 50%–60% of the α-form (triglycosylated heavy chain), whereas the pdAPC contains 70%–80% of α-APC (22). In addition, the oligosaccharides in rAPC are described to be less branched, since they are only diantennary (74). Thus, the higher impact of heme on the recombinant form might be due to less hidden binding sites and an improved access of heme to these sites.

Nevertheless, the fact that heme is able to impair the conversion of a small peptide substrate by APC might be an indication for the lower APC activity measured during hemolytic crisis in SCD patients (60, 67). The APC preincubation with PPIX resulted in a similar effect regarding the amidolytic activity, which demonstrates the importance of the protoporphyrin ring system for the interaction. Interestingly, ZnPPIX, which is naturally formed in iron deficiency (41), had an even stronger effect on APC. The amidolytic activity was almost completely suppressed.

By now, ZnPPIX itself was only reported to be a potent inhibitor of HO-1 reducing APC levels in septic mice (13). However, zinc(II) ions were shown to inhibit the enzymatic activity of APC toward a peptide substrate and FVa with a binding affinity in the micromolar range (K _D = 7.4 μM) (77). In the present experimental set-up, ZnPPIX seems to have an even higher effect, which again would confirm the significance of the porphyrin ring for interaction. Physiologically, increased levels of ZnPPIX accompanying iron deficiency states were earlier associated with malaria, which in the course of disease development results in hemolysis (27). Under these conditions, ZnPPIX could potentially contribute to decreased APC amidolytic activity. Our results demonstrate that the inhibitory effect is provoked not only by heme but also by other PPIX derivatives, which may lead to an even stronger effect under physiological conditions.

In further support of our hypotheses, heme and its analogs were capable of inhibiting APC's anticoagulant function in an aPTT-based clotting assay. Therefore, heme even enabled the complete elimination of the anticoagulant effect of APC. As one of the classical heme scavenger proteins, albumin binds heme and captures it under hemolytic conditions (57, 63). Thus, on addition of exogenous HSA, we could observe that APC is protected from inhibition by heme to a certain extent, fully recovering the anticoagulant function of APC. However, heme applied in the highest possible concentration with respect to the assay conditions, that is, 100 μM, again induced a reduction of APC's anticoagulant function. It can, thus, be assumed that massive local increase of heme, as occurring during hemolytic events, leads to an inhibition of the anticoagulant properties of APC (Fig. 8). This scenario is conceivable at least in case of a complete saturation of the heme scavengers, since already a comparably low heme concentration of 10 μM could induce the inhibition of APC in our approach.

Further, the heme-APC complex was found to exert a much higher peroxidase-like activity (∼512.33%) than heme only (100%), an effect that was also demonstrated earlier for the respective complexes with amyloid beta or fibrinogen (2, 28, 71). Interestingly, we recently reported a peroxidase-like activity of >360% for the complex of heme with the APC-derived peptide WIHGHIRDK (motif 2) (64). It is, therefore, highly probable that heme binding to this specific motif is responsible for generating a suitable environment for the binding of hydrogen peroxide (H₂O₂) and the oxidation of the substrate.

This function might be particularly important during hemolytic events in sepsis, which are associated with the formation of reactive oxygen species such as H₂O₂ (11, 33). Further, heme itself is well known for its cytotoxic effect through induction of oxidative stress, which, in turn, has been suggested to promote coagulation (18, 42). Again, this would support the procoagulant effect of heme. In addition, APC exerts cytoprotective functions through interaction of its acidic exosite with the receptor PAR1 (21, 73). Interestingly, the potential HBM 2 is close to this PAR1 interaction site in APC (Fig. 8C). As a consequence, investigation of the cytoprotective functions of APC in the context of high heme concentrations is of great importance.

As reported earlier (62), heme exposure of endothelial cells resulted in a concentration-dependent barrier dysfunction of the endothelial monolayer. Our results obtained using an HUVEC-based barrier assay support these findings. In the presence of APC, however, the cytotoxic effect of heme was partly neutralized, indicating that the cytoprotective functions of APC are not inhibited by heme binding. Thus, the interaction between heme and APC seems to be of bilateral nature. Although heme partially inhibits the anticoagulant functions of APC, the cytoprotective functions remain intact.

Herein, we report for the first time a direct interaction of heme with the serine protease APC, providing a first evidence for specific binding motifs as well as for functional consequences resulting from heme binding to APC. To what extent heme might influence the PC system through interaction with APC depends on time and spatial hierarchy of the simultaneously occurring reactions as well as on concentration of other heme-binding proteins (apart from heme-scavenging proteins) and their respective heme-binding affinity (Figs. 7 and 8A). However, it is tempting to speculate that heme-induced inactivation of APC contributes to the hypercoagulable state associated with severe hemolysis.

Materials and Methods

Reagents

Hemin [Fe(III) PPIX], later designated as heme, was obtained from Frontier Scientific (Logan, UT). Stock solutions of 1 mM heme were freshly prepared as described earlier (5, 7, 38, 39, 71). The stock solutions were diluted to the desired concentration in the respective assay buffer directly before use. Stock solutions of PPIX and ZnPPIX (Frontier Scientific) were prepared identically to heme. Ferric chloride (FeCl₃, anhydrous; Merck, Darmstadt, Germany) was dissolved 1 mM in water (double-distilled) for stock solution and freshly diluted before use.

Plasma-derived human APC (HCAPC-0080; Hematologic Technologies, Essex Junction, VT), human α-thrombin (T6884; Sigma-Aldrich, St. Louis, MO), recombinant human factor XIIIa (T070; Zedira, Darmstadt, Germany), and HSA (A3782; Sigma-Aldrich) were commercially obtained. The recombinant human APC (Drotrecogin alfa, Xigris^®) was obtained from Eli Lilly and Company (Indianapolis, IN). As a drug, Xigris was in use for the treatment of severe sepsis but was withdrawn from the market due to a lack of efficiency. It differs from pdAPC only in its glycosylation pattern, which was demonstrated to result in considerable differences in activity (74, 75). The purity of all proteins was examined by sodium dodecyl sulfate polyacrylamide gel electrophoresis.

SPR-binding studies

SPR-binding studies were performed on a Biacore T200 instrument (GE Healthcare Europe GmbH, Freiburg, Germany) as described earlier (49, 70). APC was dissolved in 10 mM acetate buffer (pH 4.5) at a concentration of 10 μg/mL and immobilized by amine coupling on a CM5 sensor chip (GE Healthcare) to a response unit (RU) of 2261 RU. Binding kinetics were determined by a titration of five consecutive injections with different heme concentrations (187.5, 375, 750, 1500, 3000 nM), prepared by dilution of the stock solution (1 mM) with running buffer. Measurements were performed at 25°C in running buffer (20 mM 2-Amino-2-(hydroxymethyl)propane-1,3-diol [Tris]-HCl, 150 mM NaCl, 0.05% Tween 20, pH 7.4) with a flow rate of 30 μL/min in the standard single-cycle kinetics method (Biacore T200 control software; GE Healthcare). Regeneration of the sensor chip surface was achieved by injection of 10 mM NaOH. Binding kinetics was evaluated by global fitting with the Biaevaluation software (GE Healthcare).

Screening and selection of potential heme-binding sites in APC

As described earlier, potential HBMs were selected as nonapeptide sequences by using the sequence-based detection of heme binding motifs (SeqD-HBM) search algorithm, which was recently manifested with the webserver HeMoQuest (5, 7, 38, 39, 48, 72). HeMoQuest is a computational approach for the prediction of transient heme binding to proteins (48). After entering the sequence of APC, 10 motifs were shortlisted based on the search criteria (a potential coordinating site [Cys, Tyr, or His], a positive net charge, and surface-exposed). Five cysteine-based motifs were excluded since the cysteine residues are involved in disulfide bridges (45), and one motif included a glycosylation site (Asn²³⁵) (22), hindering heme binding. Finally, four motifs were synthesized as nonapeptides for studying their heme-binding properties (see Synthesis, Purification, and Amino Acid Analysis of APC-Derived Peptides section) (Supplementary Table S1).

Synthesis, purification, and amino acid analysis of APC-derived peptides

Automated solid-phase peptide synthesis of APC-derived peptides 1–4 was carried out by standard 9-fluorenylmethyl-oxycarbonyl/tert-butyl (Fmoc/tBu) strategy on Rink amide MBHA resin, as described earlier (5, 7, 38, 39, 71). Subsequently, the crude peptides were purified by preparative reverse-phase high-performance liquid chromatography. The purity of the peptides was examined by analytical high performance liquid chromatography (LC-10AT; Shimadzu); the molecular identity was confirmed by an electrospray ionization mass spectrometer (ESI-MS, micrOTOF-Q III; Bruker Daltonics GmbH, Bremen, Germany) connected to a Dionex Ultimate 3000 LC 8 (Thermo Fisher Scientific, Dreieich, Germany) (Supplementary Table S1). As described earlier (5, 7, 38, 39, 71), amino acid analysis was performed to determine the peptide content, which was taken into consideration for the following studies.

Analysis of heme binding to APC-derived peptides by UV/vis spectroscopy

Heme binding to peptides 1–4 was studied by UV/vis spectroscopy, as described earlier (39). In brief, different concentrations of heme (0.4–40 μM) were incubated for 30 min with the peptide (10 μM) in 100 mM HEPES (2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid) buffer (pH 7.0). Subsequently, the absorbance spectra (300–600 nm) were recorded on a Multiskan GO microplate spectrophotometer (Thermo Fisher Scientific). Difference spectra were obtained by subtracting spectra of heme and peptide, respectively, from the spectra of the heme-peptide complexes for each heme concentration. K _D values were determined by using GraphPad prism 8.0 software and the equation earlier described (39, 53).

Molecular dynamics and docking simulation studies of the heme-APC complex

First, homology modeling was employed to construct the full-length 3D structure of APC in YASARA structure (version 18.2.7) (34, 35). The program used five initial structures (PDB ID: 1AUT, 3F6U, 2AER, 1W0Y, 3HPT) as templates in the construction of 17 different initial models based on different alignment variants. A hybrid model was built by combining the best parts of the 17 initial models. Finally, the Gla domain of the hybrid model was manually replaced by the Gla domain obtained from an earlier study that contained the calcium ions at their precise locations (51). This model was energy minimized by using the steepest descents and stimulated annealing protocol as implemented by the YASARA macro rm_run.mcr. This minimized structure that converged to a total energy of −45552.48 kJ/mol was subjected to a 300 nanoseconds MD simulation intended to produce an equilibrated conformational ensemble of the protein. The simulated system attained conformational equilibrium with an RMSD change of <2.5 Å after ∼150 nanoseconds of the simulation, and the final 100 nanoseconds of the simulation was considered as the production phase from where snapshots were selected for further analysis.

Using selected structures obtained from the production phase of the MD simulation trajectory of the model, we performed focused molecular docking studies of heme to potential HRMs of APC. The snapshots were selected arbitrarily on visual examination of the trajectory, specifically choosing ones where the coordinating residue of (H or Y) of the potential HRM was found with good surface exposure. The heme molecule was downloaded from the ChemSpider database (ChemSpider ID 16739951) (50). For each potential HRM, a cubic simulation cell (15 × 15 × 15 Å) was built around the expected coordinating residue.

The AutoDock Vina algorithm was used to run docking simulations via YASARA (35, 65). The ensemble docking approach was used by creating a protein receptor ensemble of five high-scoring solutions of the side chain rotamer network at a temperature of 298 K (65). To each rotamer, heme was docked 20 times, resulting in a hundred docking poses per run. The resulting poses were clustered to produce a list of ligand-receptor complex conformations, sorted by predicted binding energies and a 5 Å (default value used by Vina) heavy ligand atom RMSD threshold to separate the clusters.

The MD simulations of selected heme-APC docked complexes were conducted on the YASARA version 19.1.27 platform to assess the stability of heme binding to the specific HRMs, thereby validating the docking hit that was found (8, 36). Heme docking to APC was considered successful in poses where the distance between the Fe³⁺ ion of heme and the non-protonated nitrogen atom of histidine or the hydroxyl oxygen of tyrosine is ≤3 Å. First, selected docked complexes for each HRM were loaded into YASARA as starting structures for the MD simulations. All MD simulations were conducted in the NPT ensemble with a 2 fs time step, at 298 K and a pressure maintained at 1 bar. Periodic boundary conditions were used on the simulation cell, which was constructed to have a minimum distance of 10 Å from all protein atoms. Long-range forces were computed with a cutoff of 8 Å, and the particle-mesh Ewald method was used (12).

Three rounds of simulations were conducted for each complex, the first of which was run for 20 nanoseconds. Further, two independent validation runs for 50 nanoseconds each were conducted by creating mildly altered initial configurations of the system via energy minimizations (Supplementary Fig. S7). The first 20 nanoseconds run was used to verify whether the docked complex was stable in the first place and the subsequent two 50 nanoseconds validation runs were done to be sure that the binding persists even when the initial coordinates of the complex are altered due to the fact that MD simulations are highly sensitive to the initial configuration of the system (4). Analyses of the MD trajectories were done in both YASARA and VMD (version 1.9.3) by first converting the trajectory into the GROMACS XTC format (26, 30). Molecular graphics were created from UCSF Chimera (1.13.1) (52), and plots were generated by using GraphPad prism 8.0 software. To enable a qualitative comparison of the heme-protein interaction between various motifs, the binding energy computed by the Poisson-Boltzmann method as implemented by the md_analyzebindingenergy.mcr macro in YASARA was used on the trajectories of the 50 nanoseconds validation runs (15). The energies computed here are not to be considered in any sense comparable to the real or experimentally obtained binding potentials, but only as arbitrary measures for a relative comparison to quantify the interaction of heme to the different motifs.

Amidolytic activity of human APC, thrombin, and factor XIIIa (FXIIIa) in the presence of heme

The impact of heme on the catalytic activity of APC was spectrophotometrically analyzed on a Multiskan GO microplate spectrophotometer (Thermo Fisher Scientific) based on a diagnostic assay, which is commonly used for the quantification of functionally active PC (Berichrom^® Protein C). APC (100 nM) was preincubated with different concentrations of heme (0–50 μM) for 15 min at room temperature in the dark. Ninety microliters of the heme-APC complex was placed into a 96-well microplate (Greiner Bio-One GmbH, Frickenhausen, Germany), and 10 μL of the chromogenic substrate p-Glu-Pro-Arg-MNA (Berichrom Protein C kit; Siemens Healthcare Diagnostics, Erlangen, Germany) was added in a concentration varying from 0 to 2.5 mM. The assay was performed in HEPES buffer (100 mM, pH 7.4).

K_M was obtained by nonlinear regression fitting using the Michaelis-Menten equation, whereas the inhibition constant K_i was graphically analyzed by using Dixon plot (GraphPad prism 8.0 software). Performing the assay with a constant substrate concentration of 300 μM allowed for the comparison of the effect of heme with that of other porphyrin analogs (ZnPPIX and PPIX) and iron(III) chloride. Further, the impact of heme on rAPC (Xigris) was investigated. The concentration of rAPC needed to be adapted to 50 nM. A range of heme concentrations (0–10 μM) was applied.

The same procedure was used for thrombin, but with a final enzyme concentration of 25 nM and heme in concentrations of 0–50 μM. Again, p-Glu-Pro-Arg-MNA served as the substrate. The activity of FXIIIa was monitored in Tris buffer (0.05 M Tris, 0.154 M NaCl, 0.01 M CaCl₂, pH 7.5) by using H-Tyr-Glu(pNA)-Val-Lys-Val-Ile-Gly-NH₂ (100 μM) as the substrate (25). The final concentration of FXIIIa was 400 nM. Heme was added in final concentrations of 0–50 μM. The increase of absorbance due to substrate conversion was monitored every 10 s for 20 min at 405 nm. The remaining activity of the enzymes as the change of absorbance per second was normalized against the basal activity without any potential ligand. Data were analyzed as the mean of triplicate, and the IC₅₀ value was determined by fitting the curves with nonlinear regression (GraphPad prism 8.0 software).

Monitoring the peroxidase-like activity of the heme-APC complex

As described earlier (64, 71), 42 μM of heme was preincubated with 42 μM of pdAPC for 30 min in phosphate buffered saline (PBS) buffer (pH 7.4). A mixture of 3,3′,5,5′-tetramethylbenzidine (TMB; 1.66 mM) and H₂O₂ (163.2 mM) served as the substrate. After transfer of 10 μL of the heme-APC complex (final concentration 1 μM) to 200 μL of the substrate, the absorbance was detected at 652 nm. For comparison, the same experiment was performed with heme as well as with APC only as controls. Data were normalized against the activity of heme and evaluated as the mean of triplicates.

Anticoagulant activity of APC in the presence of heme

The impact of heme on the anticoagulant activity of pdAPC was analyzed by using an aPTT-based clotting assay, performed with the semi-automatic ball coagulometer KC10 (Amelung, Lemgo, Germany) (24). APC (5 nM) was incubated with different concentrations of heme (0–100 μM) for 1 h in Dulbecco's phosphate buffered saline (DPBS) buffer (with Ca²⁺ and Mg²⁺) or in DPBS with 0.1% HSA (with Ca²⁺ and Mg²⁺). After addition to the 10-fold volume of citrated pool plasma (in-house preparation), 100 μL of the respective mixture was immediately incubated with 100 μL of the contact phase activator Actin FS (Siemens Healthcare Diagnostics) at 37°C for 3 min followed by the addition of 100 μL of 30 mM CaCl₂ solution (Siemens Healthcare Diagnostics), as described earlier (24).

As a control, pool plasma containing the same amount of heme without addition of exogenous APC was subjected to the assay system. Clotting times, given in seconds, were automatically determined by the coagulometer. Data were evaluated as the mean of triplicates (GraphPad prism 8.0 software). In a second approach, the impact of heme with the lowest potent concentration was compared with that of its analogs ZnPPIX and PPIX (10 μM in the absence and 100 μM in the presence of 0.1% HSA).

Evaluation of the cytoprotective function of APC in the presence of heme

The endothelial barrier permeability assay (Merck) was used to determine the cytoprotective function of pdAPC after preincubation with heme. As described earlier (58), HUVECs were seeded (2.5 × 10⁴ cells/well) onto collagen-coated culture inserts of a 96-well plate and subsequently cultured in endothelial cell growth medium (ECGM, C-22010; PromoCell, Heidelberg, Germany) for 4 days to reach maximum confluency. The upper chamber contained 125 μL and the lower chamber contained 250 μL of ECGM. One hour before starting the experiment, ECGM was exchanged with phenol red-free endothelial cell basal medium (ECBM, C-22215; PromoCell). APC (20 nM) was preincubated with heme (120 μM) for 30 min in phenol red-free ECBM. Subsequently, the mixture was applied to the upper chamber and the cells were incubated for 3 h at 37°C, 5% CO₂.

As controls, buffer, APC, and heme were separately incubated with the cell monolayer. After washing the cells three times with PBS, the APC/heme mixture was replaced by either thrombin (20 nM) or ECBM as a control for 30 min at 37°C. Later, the content of the upper and lower chamber was replaced by FITC-labeled dextran (75 μL of 1:40 dilution in phenol red-free ECBM) and 250 μL phenol red-free ECBM, respectively, and incubated for 20 min at room temperature. Finally, the determination of the fluorescence intensity (λ_ex = 485 nm, λ_em = 535 nm) with a Synergy 2 microplate reader allowed for the quantitation of the flux through the endothelial cell monolayer at a rate proportional to the permeability. Data are presented as the mean of quadruplicates.

Statistical analysis

Ordinary one-way analysis of variance with post hoc Tukey's multiple-comparisons test was used to calculate the indicated p values. A p value of 0.05 was assumed to be significant. Calculations were performed by using GraphPad prism 8.0 software.

Footnotes

Authors' Contributions

M.-T.H., N.S.H., B.P., and D.I. conceived and designed the study. M.-T.H. performed the experiments and collected the data. N.S.H. performed the barrier protection assay. N.A., A.A.P.G., and A.B. carried out the computational studies. M.-T.H., N.S.H., B.P., A.A.P.G., A.B., and D.I. analyzed the data. B.P. and J.O. contributed analysis tools. The article was written through the contributions of all authors.

Acknowledgments

The authors like to thank Kornelia Hampel and Bastian Zimmermann (Biaffin GmbH Kassel) for SPR measurements. In addition, the authors thank Amelie Wißbrock (University of Bonn) and Sneha Singh (University Hospital Bonn) for useful scientific discussions. They are also grateful to Charlotte A. Bäuml and Sabrina Linden (University of Bonn) for technical assistance.

Author Disclosure Statement

The authors declare no competing financial interests.

Funding Information

Financial support of M.-T. H. by the University of Bonn is gratefully acknowledged.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

Supplementary Video S1

Abbreviations Used

References

Ataga

. Hypercoagulability and thrombotic complications in hemolytic anemias. Haematologica, 94: 1481–1484, 2009.

Atamna

and Boyle

. Amyloid-beta peptide binds with heme to form a peroxidase: relationship to the cytopathologies of Alzheimer's disease. Proc Natl Acad Sci U S A, 103: 3381–3386, 2006.

Bertina

, Koeleman

BPC

, Koster

, Rosendaal

, Dirven

, de Ronde

, van der Velden

, and Reitsma

. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature, 369: 64–67, 1994.

Braxenthaler

, Unger

, Auerbach

, Given

, and Moult

. Chaos in protein dynamics. Proteins, 29: 417–425, 1997.

Brewitz

, Goradia

, Schubert

, Galler

, Kühl

, Syllwasschy

, Popp

, Neugebauer

, Hagelueken

, Schiemann

, Ohlenschläger

, and Imhof

. Heme interacts with histidine- and tyrosine-based protein motifs and inhibits enzymatic activity of chloramphenicol acetyltransferase from Escherichia coli. Biochim Biophys Acta Gen Subj, 1860: 1343–1353, 2016.

Brewitz

, Hagelueken

, and Imhof

. Structural and functional diversity of transient heme binding to bacterial proteins. Biochim Biophys Acta Gen Subj, 1861: 683–697, 2017.

Brewitz

, Kühl

, Goradia

, Galler

, Popp

, Neugebauer

, Ohlenschläger

, and Imhof

. Role of the chemical environment beyond the coordination site: structural insight into Fe(III)protoporphyrin binding to cysteine-based heme-regulatory protein motifs. ChemBioChem, 16: 2216–2224, 2015.

Chen

YC.

Beware of docking!. Trends Pharmacol Sci, 36: 78–95, 2015.

D'Alessandro

and Zolla

. Proteomic analysis of red blood cells and the potential for the clinic: what have we learned so far?. Expert Rev Proteomics, 14: 243–252, 2017.

10.

Donegan

, Moore

, Hanna

, and Reddi

. Handling heme: the mechanisms underlying the movement of heme within and between cells. Free Radic Biol Med, 133: 88–100, 2019.

11.

Effenberger-Neidnicht

and Hartmann

. Mechanisms of hemolysis during sepsis. Inflammation, 41: 1569–1581, 2018.

12.

Essmann

, Perera

, Berkowitz

, Darden

, Lee

, and Pedersen

. A smooth particle mesh Ewald method. J Chem Phys, 103: 8577–8593, 1995.

13.

Fei

, Meng

, Zhao

, Kang

KAI

, Tan

, Pan

, Luo

, Liu

WEN

, Nan

, Jiang

, Krissansen

, Zhao

, and Sun

. Enhanced induction of heme oxygenase-1 suppresses thrombus formation and affects the protein C system in sepsis. Transl Res, 159: 99–109, 2012.

14.

Fisher

, Greengard

, and Griffin

. Models of the serine protease domain of the human antithrombotic plasma factor activated protein C and its zymogen. Protein Sci, 3: 588–599, 1994.

15.

Fogolari

, Brigo

, and Molinari

. The Poisson-Boltzmann equation for biomolecular electrostatics: a tool for structural biology. J Mol Recognit, 15: 377–392, 2002.

16.

Frimat

, Tabarin

, Dimitrov

, Poitou

, Halbwachs-Mecarelli

, Fremeaux-Bacchi

, and Roumenina

. Complement activation by heme as a secondary hit for atypical hemolytic uremic syndrome. Blood, 122: 282–292, 2013.

17.

Gabre

, Chabasse

, Cao

, Mukhopadhyay

, Siefert

, Bi

, Netzel-Arnett

, Sarkar

, and Zhang

. Activated protein C accelerates venous thrombus resolution through heme oxygenase-1 induction. J Thromb Haemost, 12: 93–102, 2014.

18.

Görlach

. Redox regulation of the coagulation cascade. Antioxid Redox Signal, 7: 1398–1404, 2005.

19.

Green

and Ts'ao

. Hematin: effects on hemostasis. J Lab Clin Med, 115: 144–147, 1990.

20.

Greffe

, Manco-Johnson

, and Marlar

. Molecular forms of human protein C: comparison and distribution in human plasma. Thromb Haemost, 62: 902–905, 1989.

21.

Griffin

, Zlokovic

, and Mosnier

. Activated protein C, protease activated receptor 1, and neuroprotection. Blood, 132: 159–169, 2018.

22.

Grinnell

, Walls

, and Gerlitz

. Glycosylation of human protein C affects its secretion, processing, functional activities, and activation by thrombin. J Biol Chem, 226: 9778–9785, 1991.

23.

Gruber

and Griffin

. Direct detection of activated protein C in blood from human subjects. Blood, 79: 2340–2348, 1992.

24.

Hamedani

, Rühl

, Zimmermann

, Heiseler

, Oldenburg

, Mayer

, Pötzsch

, and Müller

. In vitro evaluation of aptamer-based reversible inhibition of anticoagulant activated protein C as a novel supportive hemostatic approach. Nucleic Acid Ther, 26: 355–362, 2016.

25.

Hardes

, Becker

, Zouhir Hammamy

, and Steinmetzer

. Design, synthesis, and characterization of chromogenic substrates of coagulation factor XIIIa. Anal Biochem, 428: 73–80, 2012.

26.

Hess

, Kutzner

, Van Der Spoel

, and Lindahl

. GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation. J Chem Theory Comput, 4: 435–447, 2008.

27.

Hiremath

, Sullivan

, Tripathi

, Black

, and Sazawal

. Effect of Plasmodium falciparum parsitemia on erythrocyte zinc protoporphyrin. Clin Chem, 52: 778–779, 2006.

28.

Hou

, Zhang

, Wu

, Wang

, Zhang

, Li

, Wang

, Xue

, and Wang

. Label-free detection of fibrinogen based on fibrinogen-enhanced peroxidase activity of fibrinogen-hemin composite. Analyst, 143: 725–730, 2018.

29.

Huang

, Yang

, Xu

, Zhang

, Gao

, and Li

. Insulin enhances the peroxidase activity of heme by forming heme-insulin complex: relevance to type 2 diabetes mellitus. Int J Biol Macromol, 102: 1009–1015, 2017.

30.

Humphrey

, Dalke

, and Schulten

. VMD: visual molecular dynamics. J Mol Graph, 14: 33–38, 1996.

31.

and Huang

. Haem-assisted dityrosine-cross-linking of fibrinogen under non-thermal plasma exposure: one important mechanism of facilitated blood coagulation. Sci Rep, 6: 1–8, 2016.

32.

Kisiel

. Human plasma protein C: isoation, characterization, and mechanism of activation by α-thrombin. J Clin Invest, 64: 761–769, 1979.

33.

Kolls

. Oxidative stress in sepsis: a redox redux. J Clin Invest, 116: 860–863, 2006.

34.

Krieger

, Joo

, Lee

, Raman

, Thompson

, Tyka

, Baker

, and Karplus

. Improving physical realism, stereochemistry, and side-chain accuracy in homology modeling: four approaches that performed well in CASP8. Proteins, 77: 114–122, 2009.

35.

Krieger

and Vriend

. YASARA View—molecular graphics for all devices—from smartphones to workstations. Bioinformatics, 30: 2981–2982, 2014.

36.

Krieger

and Vriend

. New ways to boost molecular dynamics simulations. J Comput Chem, 36: 996–1007, 2015.

37.

Kühl

and Imhof

. Regulatory Fe II/III heme: the reconstruction of a molecule's biography. ChemBioChem, 15: 2024–2035, 2014.

38.

Kühl

, Sahoo

, Nikolajski

, Schlott

, Heinemann

, Imhof

, Kühl

, Heinemann

, Schlott

, Imhof

, and Sahoo

. Determination of hemin-binding characteristics of proteins by a combinatorial peptide library approach. ChemBioChem, 12: 2846–2855, 2011.

39.

Kühl

, Wißbrock

, Goradia

, Sahoo

, Galler

, Neugebauer

, Popp

, Heinemann

, Ohlenschläger

, and Imhof

. Analysis of Fe(III) heme binding to cysteine-containing heme-regulatory motifs in proteins. ACS Chem Biol, 8: 1785–1793, 2013.

40.

Kumar

, Wißbrock

, Goradia

, Bellstedt

, Ramachandran

, Imhof

, and Ohlenschläger

. Heme interaction of the intrinsically disordered N-terminal peptide segment of human cystathionine-β-synthase. Sci Rep, 8: 1–9, 2018.

41.

Labbé

, Vreman

, and Stevenson

. Zinc protoporphyrin: a metabolite with a mission. Clin Chem, 45: 2060–2072, 1999.

42.

Larsen

, Gouveia

, Soares

, and Gozzelino

. Heme cytotoxicity and the pathogenesis of immune-mediated inflammatory diseases. Front Pharmacol, 3: 77, 2012.

43.

Lipe

and Ornstein

. Deficiencies of natural anticoagulants, protein C, protein S, and antithrombin. Circulation, 124: 365–369, 2011.

44.

Mannucci

, Lobina

, Caocci

, and Dioguardi

. Effect on blood coagulation of massive intravascular haemolysis. Blood, 33: 207–213, 1969.

45.

Mather

, Oganessyan

, Hof

, Huber

, Foundling

, Esmon

, and Bode

. The 2.8 Å crystal structure of Gla-domainless activated protein C. EMBO J, 15: 6822–6831, 1996.

46.

Mesters

, Houghten

, and Griffin

. Identification of a sequence of human activated protein C (residues 390-404) essential for its anticoagulant activity. J Biol Chem, 266: 24514–24519, 1991.

47.

Ogawa

, Igarashi

, Nishitani

, Shibahara

, and Fujita

. Heme-regulated transcription factor BACH1. J Heal Sci, 48: 1–6, 2002.

48.

Paul George

, Lacerda

, Syllwasschy

, Hopp

M-T

, Wißbrock

, and Imhof

. HeMoQuest: a webserver for qualitative prediction of transient heme binding to protein motifs. BMC Bioinformatics, 21: 124, 2020.

49.

Peherstorfer

, Brewitz

, Paul George

, Wißbrock

, Adam

, Schmitt

, and Imhof

. Insights into mechanism and functional consequences of heme binding to hemolysin-activating lysine acyltransferase HlyC from Escherichia coli. Biochim Biophys Acta Gen Subj, 1862: 1964–1972, 2018.

50.

Pence

and Williams

. ChemSpider: an online chemical information resource. J Chem Educ, 87: 1123–1124, 2010.

51.

Perera

, Foley

, Darden

, Stafford

, Mather

, Esmon

, and Pedersen

. Modeling zymogen protein C. Biophys J, 79: 2925–2943, 2000.

52.

Pettersen

, Goddard

, Huang

, Couch

, Greenblatt

, Meng

, and Ferrin

. UCSF Chimera—a visualization system for exploratory research and analysis. J Comput Chem, 25: 1605–1612, 2004.

53.

Pîrnău

and Bogdan

. Investigation of the interaction between flavonoids and human serum albumin. Rom J Biophys, 18: 49–55, 2008.

54.

Rehani

, Mathson

, Belcher

, Vercellotti

, and Slungaard

. Heme potently stimulates tissue factor expression by peripheral blood monocytes: a novel mechanism for thrombosis in intravascular hemolytic diseases. Blood, 122: 2215, 2013.

55.

Repessé

, Dimitrov

, Peyron

, Moshai

, Kiger

, Dasgupta

, Delignat

, Marden

, Kaveri

, and Lacroix-Desmazes

. Heme binds to factor VIII and inhibits its interaction with activated factor IX. J Thromb Haemost, 10: 1062–1071, 2012.

56.

Rezaie

. Exosite-dependent regulation of the protein C anticoagulant pathway. Trends Cardiovasc Med, 13: 8–15, 2003.

57.

Roumenina

, Rayes

, Lacroix-Desmazes

, and Dimitrov

. Heme: modulator of plasma systems in hemolytic diseases. Trends Mol Med, 22: 200–213, 2016.

58.

Russo

, Soh

UJK

, Paing

, Arora

, and Trejo

. Caveolae are required for protease-selective signaling by protease-activated receptor-1. Proc Natl Acad Sci U S A, 106: 6393–6397, 2009.

59.

Sala

, Owen

, and Collen

. A functional assay of protein C in human plasma. Blood, 63: 671–675, 1984.

60.

Schnog

, Gillavry MR

Mac

, Zanten AP

Van

, Meijers

JCM

, Rojer

, Duits

, Cate

, and Brandjes

DPM

. Protein C and S and inflammation in sickle cell disease. Am J Hematol, 76: 26–32, 2004.

61.

Schubert

, Florin

, Duthie

, Brewitz

, Kühl

, Imhof

, Hagelueken

, Schiemann

, Henning Brewitz

, Kühl

, Imhof

, Hagelueken

, and Schiemann

. Spectroscopic studies on peptides and proteins with cysteine-containing heme regulatory motifs (HRM). J Inorg Biochem, 148: 49–56, 2015.

62.

Singla

, Sysol

, Dille

, Jones

, Chen

, and Machado

. Hemin causes lung microvascular endothelial barrier dysfunction by necroptotic cell death. Am J Respir Cell Mol Biol, 57: 307–314, 2017.

63.

Soares

and Bozza

. Red alert: labile heme is an alarmin. Curr Opin Immunol, 38: 94–100, 2016.

64.

Syllwasschy

, Beck

, Družeta

, Hopp

M-T

, Ramoji

, Neugebauer

, Nozinovic

, Menche

, Willbold

, Ohlenschläger

, Kühl

, and Imhof

. High-affinity binding and catalytic activity of His/Tyr-based sequences: extending heme-regulatory motifs beyond CP. Biochim Biophys Acta Gen Subj, 1864: 129603, 2020.

65.

Trott

and Olson

. AutoDock Vina. J Comput Chem, 31: 445–461, 2010.

66.

De Visser

MCH

, Rosendaal

, and Bertina

. A reduced sensitivity for activated protein C in the absence of factor V Leiden increases the risk of venous thrombosis. Blood, 93: 1271–1276, 1999.

67.

Westerman

, Green

, Gilman-Sachs

, Beaman

, Freels

, Boggio

, Allen

, Zuckerman

, Schlegel

, and Williamson

. Antiphospholipid antibodies, proteins C and S, and coagulation changes in sickle cell disease. J Lab Clin Med, 134: 352–362, 1999.

68.

Wiederstein

and Sippl

. ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucleic Acids Res, 35: 407–410, 2007.

69.

Wildhagen

KCAA

, Lutgens

, Loubele

STGB

, ten Cate

, and Nicolaes

GAF

. The structure-function relationship of activated protein C: lessons from natural and engineered mutations. Thromb Haemost, 106: 1034–1045, 2011.

70.

Wißbrock

, Goradia

, Kumar

, Paul George

, Kühl

, Bellstedt

, Ramachandran

, Hoffmann

, Galler

, Popp

, Neugebauer

, Hampel

, Zimmermann

, Adam

, Wiendl

, Krönke

, Hamza

, Heinemann

, Frey

, Hueber

, Ohlenschläger

, and Imhof

. Structural insights into heme binding to IL-36α proinflammatory cytokine. Sci Rep, 9: 16893, 2019.

71.

Wißbrock

, Kühl

, Silbermann

, Becker

, Ohlenschläger

, and Imhof

. Synthesis and evaluation of amyloid β derived and amyloid β independent enhancers of the peroxidase-like activity of heme. J Med Chem, 60: 373–385, 2017.

72.

Wißbrock

, Paul George

, Brewitz

, Kühl

, and Imhof

. The molecular basis of transient heme-protein interactions: analysis, concept and implementation. Biosci Rep, 39: BSR20181940, 2019.

73.

Yamaji

, Wang

, Liu

, Abeyama

, Hashiguchi

, Uchimura

, Biswas

, Iwamoto

, and Maruyama

. Activated protein C, a natural anticoagulant protein, has antioxidant properties and inhibits lipid peroxidation and advanced glycation end products formation. Thromb Res, 115: 319–325, 2005.

74.

Yan

, Chao

, and van Halbeek

. Novel Asn-linked oligosaccharides terminating in GalNAcβ (1 → 4)[Fucα (1 → 3)]GlcNAcβ (1→•) are present in recombinant human protein C expressed in human kidney 293 cells. Glycobiology, 3: 597–608, 1993.

75.

Yan

SCB

, Razzano

, Chao

, Walls

, Berg

, McClure

, and Grinnell

. Characterization and novel purification of recombinant human protein C from three mammalian cell lines. Nat Biotechnol, 8: 655–661, 1990.

76.

and Ragsdale

. Evidence that the heme regulatory motifs in heme oxygenase-2 serve as a thiol/disulfide redox switch regulating heme binding. J Biol Chem, 282: 21056–21067, 2007.

77.

Zhu

, Ubhayasekera

, Nickolaus

, Sun

, Tingsborg

, Mowbray

, and Schedin-Weiss

. Zinc ions bind to and inhibit activated protein C. Thromb Haemost, 104: 544–553, 2010.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

122.94 MB

0.60 MB