Use of cysteine as a spectroscopic probe for determination of heme-scavenging capacity of serum proteins and whole human serum

HAL Id: hal-02127294

https://hal.archives-ouvertes.fr/hal-02127294

Submitted on 13 May 2019

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Use of cysteine as a spectroscopic probe for

determination of heme-scavenging capacity of serum proteins and whole human serum

Remi Noe, Nina Bozinovic, Maxime Lecerf, Sébastien Lacroix-Desmazes, Jordan Dimitrov

To cite this version:

Remi Noe, Nina Bozinovic, Maxime Lecerf, Sébastien Lacroix-Desmazes, Jordan Dimitrov. Use of

cysteine as a spectroscopic probe for determination of heme-scavenging capacity of serum proteins

and whole human serum. Journal of Pharmaceutical and Biomedical Analysis, Elsevier, 2019, 172,

pp.311-319. �10.1016/j.jpba.2019.05.013�. �hal-02127294�

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j p b a

Use of cysteine as a spectroscopic probe for determination of

heme-scavenging capacity of serum proteins and whole human serum

Rémi Noé ¹ , Nina Bozinovic ¹ , Maxime Lecerf, Sébastien Lacroix-Desmazes, Jordan D. Dimitrov ^∗

Centre de Recherche des Cordeliers, INSERM, Sorbonne Université, USPC, Université Paris Descartes, Université Paris Diderot, F-75006 Paris, France

a r t i c l e i n f o

Article history:

Received 19 December 2018 Received in revised form 3 May 2019 Accepted 5 May 2019

Available online 7 May 2019

Keywords:

Heme Hemolysis

Heme-binding proteins Human serum

Absorbance spectroscopy

a b s t r a c t

Heme serves as a prosthetic group of numerous proteins involved in the oxidative metabolism. As result of various pathological conditions associated with hemolysis or tissue damage, large quantities of hemopro- teins and heme can be released extracellularly. Extracellular heme has pronounced pathogenic effects in hemolytic diseases, mediated by its pro-oxidative and pro-inflammatory activities. The pathogenic potential of heme is mostly expressed when the molecule is in protein unbound form. The pathological relevance of free heme deems it necessary to develop reliable approaches for its assessment. Here we developed a technique based on UV–vis absorbance spectroscopy, where cysteine was used as a spec- troscopy probe to distinguish between heme-bound to plasma proteins or hemoglobin from free heme.

This technique allowed estimation of the heme-binding capacity of human serum, of particular heme scavenging proteins (albumin, hemopexin) or of immunoglobulins. The main advantage of the proposed approach is that it can distinguish free heme from heme associated with proteins with a wide range of affinities. The described strategy can be used for evaluation of heme-binding capacity of human plasma or serum following intravascular hemolysis or for estimation of stoichiometry of interaction of heme with a given protein.

© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Heme (Fe-protoporphyrin IX) is a macrocyclic compound that serves as a prosthetic group of many proteins involved in the aerobic metabolism. By transiently interacting with a number of intracellular or membrane-associated proteins, heme also participates in the cell signalling and regulation of cellu- lar functions [1]. Under physiological conditions most of heme is intracellularly sequestered. However, release of large quantities of heme-containing proteins (hemoproteins), such as hemoglobin and myoglobin can occur as consequence of diverse patholo- gies. Thus, damage of erythrocytes due to genetic abnormalities of hemoglobin, infections, trauma, or autoimmunity can result in intravascular hemolysis and liberation of massive quantities of extracellular hemoglobin [2]. In extracellular compartment and upon oxidation, heme relatively easily dissociates from

∗ Corresponding author at: INSERM UMRS 1138, Centre de Recherche des Corde- liers, 75006 Paris, France.

E-mail addresses: jordan.dimitrov@crc.jussieu.fr, jordan.dimitrov@inserm.fr (J.D. Dimitrov).

1

These authors contributed equally to the work.

hemoglobin. As a result of its oxidative potential and prominent hydrophobicity, free heme is inherently toxic molecule. Likewise, heme is considered as endogenous danger signal (alarmin) that activates different types of immune cells and endothelia, increases vascular permeability, triggers complement cascade activation and dysregulates coagulation [3,4]. Accordingly, extracellular hemo- proteins and heme contribute to the pathogenesis of diseases such as malaria, sickle cell disease, sepsis, rhabdomyolysis and other.

In the clinical practice the assessment of hemolysis is performed by measurement of the plasma scavenger of hemoglobin, i.e. hap- toglobin [5]. Thus, a decrease in the plasma concentration of this protein signifies recent hemolytic events. Extensive hemolysis may also result in overwhelming of scavenging capacity of hemopexin, the plasma protein that binds heme with high affinity [6]. In this case, the extracellular heme can associate with lower affinity to other plasma constituents, including albumin, and lipoproteins [6,7]. However, as albumin has a slow exchange rate (half-life of ca. 20 days), recurrent and extensive hemolysis may also result in saturation of the heme-binding capacity of this abundant plasma protein. Since the pathologically-relevant form of heme is the one that is loosely bound to proteins (referred to as free heme), its esti- mation is of utmost clinical importance [3]. However, to the best

https://doi.org/10.1016/j.jpba.2019.05.013

0731-7085/© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.

0/).

312 R. Noé et al. / Journal of Pharmaceutical and Biomedical Analysis 172 (2019) 311–319

of knowledge, there is no estimation of the total heme-scavenging capacity of human serum. There are commercialized techniques for measurement of concentration of total heme in plasma that have been applied in different studies [8–10]. These approaches relay on colorimetric detection of heme or on the pseudo-peroxidase activ- ity of heme in the presence of hydrogen peroxide. Nonetheless, these techniques fail to differentiate the hemoprotein-bound from free heme.

Here we describe a simple method that relays on characteris- tic spectral changes of protein-free heme upon interaction with a thiol-containing compound (cysteine). We demonstrated that this technique could be reliably used for estimation of the free extra- cellular heme. The method can also be applied for evaluation of the heme-binding capacity of different proteins and human serum as well as for assessment of quality of heme-binding proteins used in therapy or as blood substituents.

2. Material and methods

2.1. Materials

Hemin was obtained from Frontier Scientific, Inc. (Logan, UT).

Cysteine, reduced glutathione, potassium cyanide, DMSO, phenyl- hydrazine hydrochloride, 5,5

-dithiobis(2-nitrobenzoic acid), 2,2

- azino-bis(3-ethylbenzothiazoline-6-sulphonic acid (ABTS), human hemopexin and human hemoglobin were obtained from Sigma- Aldrich (St. Louis, MO). All chemicals were with the highest available purity. Human serum albumin (LFB, Les Ulis, France) and human pooled immunoglobulin G (IVIg, Endobulin, Baxter USA) were thoroughly dialyzed against PBS and stored before use at

− 20

C at concentrations of 200 mg/ml and 80 mg/ml, respectively.

Human AB serum obtained from healthy donor was purchased from Etablissement franc¸ ais du sang, Paris, France, (ethical authorization N

12/EFS/079). All stock solutions were freshly prepared and used within 24 h. Hemin was dissolved in DMSO to final concentration of 2 mM. The stock solutions of the other chemicals and proteins were always prepared in PBS.

2.2. Methods

2.2.1. Settings for spectroscopic measurements

If not stated otherwise all measurements of heme interac- tions were performed by using following experimental setting:

the UV–vis absorbance spectra were recorded by Cary-300 spec- trophotometer (Agilent Technologies, Santa-Clara, CA) using 1 ml quartz optical cells (Hellma, Jena, Germany) with 1 cm optical path.

The spectra were recorded in the wavelength range 300–700 nm with spectral resolution of 1 nm and bandwidth set at 2 nm. The absorbance background derived by the buffer only was subtracted from each reading. All measurements were performed at 25

C.

2.2.2. Initial analyses of interaction of heme with low molecular weight ligands

To identify an appropriate probe that interacts only with protein-free heme, we tested binding of cyanide, glutathione and cysteine to free heme and protein-bound heme. Hemin was first diluted to 20 ␮ M in PBS. The UV–vis spectra in absence and pres- ence of 10 mM final concentration of potassium cyanide, reduced glutathione or cysteine were recorded. The heme’s ligands were added immediately before measurement and samples vigorously homogenized.

For examination of the effects of heme ligation protein bound state 20 ␮ M solutions of human hemopexin and human serum albumin in PBS were first treated with 20 ␮ M final concentration of heme. Human hemoglobin at 5 ␮ M (containing 20 ␮ M of heme)

was used directly. After recording the spectra of the protein solu- tions, potassium cyanide, glutathione or cysteine were added to the protein solutions at final concentration of 10 mM. Following, vigorous homogenization, the spectra were measured as described above.

2.2.3. Determination of cysteine concentration for heme-cysteine spectral measurements

Cysteine was diluted in PBS to concentrations of 0.156, 0.312, 0.625, 1.25, 2.5, 5, 10 and 20 mM. The UV–vis spectra of each concentration of cysteine in the presence of 20 ␮ M hemin were measured as described above. To obtain differential spectra, the absorbance spectrum in the range 300–700 nm of hemin at 20 ␮ M was subtracted from the spectra of hemin in the presence of each concentration of cysteine.

2.2.4. Building of a calibration curve for heme-cysteine

Hemin was diluted in PBS to concentrations of 0.312, 0.625, 1.25, 2.5, 5, 10, 20, and 40 ␮ M. The UV–vis spectra of each concentration in the presence of excess of cysteine (10 mM final concentration) were measured as described above. To obtain differential spec- tra, the absorbance in the range 300–700 nm of heme at a given concentration was subtracted from the spectrum of hemin in the presence of cysteine. The obtained differential spectra have promi- nent absorbance maximum at 364 nm. The values of the differential absorbance at this maximum were plotted versus the concentration of hemin.

2.2.5. Evaluation of heme-binding capacity of human plasma proteins

Human serum albumin, and human pooled IgG were diluted in PBS at concentrations ranging from 1.25–320 ␮ M (dilution by fac- tor of 2). Alternatively, human hemopexin and human monoclonal IgG1 (identified in our previous studies to interact with heme) were diluted in PBS in concentrations ranging 0.312–40 ␮ M. Next, 20

␮ M final concentration of hemin was added to each protein dilu- tion and UV–vis spectra were recorded as described above. After to the same samples was added cysteine at 10 mM final concentra- tion and spectra were recorded immediately. Differential spectra were obtained after subtraction of spectrum of protein-heme from the spectrum of protein-heme / cysteine at given protein concen- tration. The plotting of the maximum differential absorbance at 364 nm versus concentration of protein was used for calculation of the heme-binding capacity of albumin or IgG.

2.2.6. Evaluation of heme-binding capacity of human serum Serum obtained from healthy blood donor was assayed at dilu- tions (in PBS) in range from 20 to 5120 × (dilution factor of 2). To each serum dilution was added 20 ␮ M hemin and spectra recorded following intensive homogenization. After the measurement of the spectrum of heme in presence of a given dilution of serum, to the same sample was added 10 mM cysteine, intensively homogenized, and the UV–vis spectrum recorded. The assessment of serum heme- binding capacity was performed by plotting the value of differential absorbance at 364 nm versus the dilution of serum.

2.2.7. Detection of free heme under hemolytic condition

We used leftover serum samples kindly provided by Dr. Lubka

Roumenina (Centre de Recherche des Cordeliers, INSERM U1138,

Paris). As a part of this study wild-type C57BL/6 mice were

obtained from Charles River Laboratories. Eight-week-old female

C57BL/6 WT mice were injected intraperitoneally with 200 ␮ l PBS

or same volume of phenylhydrazine (900 ␮ mol/kg, corresponding

to 0.125 mg/g body weight). The blood was collected from the sub-

mandibular vein 24 h after the induction of hemolysis. Serum was

obtained by centrifugation of blood at 2000 × g for 10 min. These

Fig. 1. Interaction of low-molecular weight heme’s ligands with protein-bound and free heme. (A) Absorbance spectra of 20 ␮M hemin in the absence or presence of 10 mM final concentration of KCN (blue line), of reduced glutathione (green line) or of cysteine (red line). (B) Absorbance spectra of 5 ␮M human hemoglobin (equivalent of 20 ␮M heme) in the absence or presence of 10 mM excess of KCN (blue line), reduced glutathione (green line) or cysteine (red line). (C) Absorbance spectra of 20 ␮M human hemopexin in absence of hemin (gray line) or in presence of 20 ␮M hemin without (dashed line) or with 10 mM excess of KCN (blue line), reduced glutathione (green line) or cysteine (red line). (D) Absorbance spectra of 20 ␮M human serum albumin in absence of hemin (gray line) or in presence of 20 ␮M hemin without (dashed line) or with 10 mM excess of KCN (blue line), reduced glutathione (green line) or cysteine (red line). All dilutions were performed in PBS pH 7.4. In all cases UV–vis absorbance spectra of hemin were recorded in the wavelength range 300–700 nm in quartz optical cell with 1 cm path length. The measurements were performed immediately after addition of the ligands (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

experiments were conducted in accordance with the recommen- dations for the care and use of laboratory animals provided by Charles Darwin ethical committee (Paris, France) and with the eth- ical authorization from the French Ministry of Agriculture, Paris, France (APAFIS 3764-201601121739330v3).

Normal mice serum and serum obtained from mice with acute intravascular hemolysis were first diluted 20 × in PBS. Then, the UV–vis spectra of serum samples were recorded in the absence and the presence of 10 mM of cysteine as described above.

2.2.8. Detection of free heme by apo-peroxidase assay

Horseradish peroxidase (HRP) was obtained from Sigma- Aldrich. To prepare apo-HRP, we used the procedure of extraction by acidified acetone as described in [11]. After dialyses against PBS apo-HRP was stored at − 20

C until use.

Normal human serum was serially diluted in PBS containing 0.1% Tween 20 in range 10–2560 × (dilution factor of 2). Next, to two identical series of serum dilutions samples we added hemin resulting in a final concentration of 10 ␮ M. Following, an incuba- tion of 10 min at room temperature apo-HRP was added resulting in 10 ␮ g/ml. The enzyme was added to a serum dilutions contain- ing hemin as well as to a set of identical serum dilutions in absence of heme. The samples were incubated for 10 min at room temper- ature. After the incubation 50 ␮ l of each sample was transferred to NUNC 96 well polystyrene plates (Thremo Fisher Scientific). The samples were then mixed with 150 ␮ l of reaction buffer - solu- tion of 0.5 mg/ml ABTS, in 0.1 M Citrate-phosphate buffer pH 5.0, containing 6 mM H

O

. The oxidation of ABTS was followed by mea- suring the absorbance at 414 nm with a microplate reader (Infinite 200 Pro, Tecan). The reconstitution of apo-HRP enzyme activity was evaluated after subtraction of the absorbance at 414 nm of

serum samples containing only apo-HRP and only hemin from cor- responding serum dilutions containing both hemin and apo-HRP.

3. Results and discussion

In the present study we aimed to develop a technique for detection and quantification of protein unbound extracellular heme. Such method should be able to distinguish heme-associated to proteins from free heme in complex systems such as blood plasma or serum. Interaction of low molecular weight ligands with heme’s iron leads to considerable changes in the absorbance spectra of the porphyrin molecule. Hence, low molecular weight heme-binding substances can be used as spectroscopy probes for assessing unbound heme. To discriminate protein-bound from unbound heme, the probe should change the absorbance spectrum of unligated heme only. Moreover, the probe ideally should induce distinct spectral changes as compared to those induced by binding of heme to the abundantly present hemoproteins (hemoglobin and myoglobin) and plasma heme-scavenging proteins (hemopexin and albumin).

3.1. Selection of spectroscopy probes for assessment of free heme

The interaction of heme with globins, hemopexin and albu-

min involves coordination of heme’s iron by imidazole group of

histidine residues [12]. This type of coordination results in charac-

teristic bathochromic (red) shifts in the Soret region of the spectra

of the oxidized heme. On the contrary, the coordination of heme

by thiol-containing compounds such as cysteine residues leads to

hypsochromic (blue) shift in the Soret region of UV–vis absorbance

spectrum [13]. Based on this evidence, we reasoned that thiol-

containing substances could serve as spectroscopy probes that

allow discrimination of free heme from protein-bound heme. To

314 R. Noé et al. / Journal of Pharmaceutical and Biomedical Analysis 172 (2019) 311–319

Fig. 2. Quantification of the spectral changes of heme following exposure to cysteine. (A) UV–vis absorbance spectra of hemin at concentrations of 20 ␮M were recorded in absence and presence of 0.156, 0.312, 0.625, 1.25, 2.5, 5, 10 and 20 mM cysteine. The differential spectra were obtained after subtraction of the spectrum of hemin in the absence of cysteine from the hemin spectrum at given cysteine concentration. On the left panel are presented differential spectra. The right panel depicts the cysteine- concentration dependent increase in the absorbance at 364 nm, a characteristic absorbance for complex of hemin with cysteine. (B) UV–vis absorbance spectra of hemin at concentrations of 0.625, 1.25, 2.5, 5, 10, 20 and 40 ␮M were recorded in absence and presence of 10 mM cysteine. The differential spectra were obtained after subtraction of spectrum at given hemin concentration from the spectrum at same hemin concentration in the presence of 10 mM cysteine. On the left panel are presented differential spectra. The right panel depicts the hemin-concentration dependent increase in the absorbance at 364 nm, a characteristic absorbance for complex of hemin with cysteine.

All dilutions were performed in PBS pH 7.4. In all cases UV–vis absorbance spectra of hemin were recorded in the wavelength range 300–700 nm in quartz optical cell with 1 cm path length. The measurements were performed immediately after addition of the ligands.

test this hypothesis, we first measured the UV–vis absorbance spec- tra of oxidized form of heme i.e. hemin in the presence of low molecular weight heme ligands (Fig. 1A). The free hemin in phos- phate buffer demonstrated a characteristic absorbance spectrum with Soret maximum at 388 nm. The addition of excess of high affinity ligand of hemin – cyanide resulted in a prominent alter- ation of the absorbance spectrum with the typical red shift of the absorbance maximum to 423 nm in Soret region and an aug- mentation of the absorbance intensity (Fig. 1A). On the contrary, the incubation of hemin with excess of thiol-containing amino acid, cysteine, resulted in a distinctive blue shift in the absorbance maximum to 364 nm, accompanied by an increase in the overall absorbance intensity (Fig. 1A). Hemin-cysteine spectral changes were independent of the solvent used to solubilize hemin before final 100 × dilution in PBS, albeit a lower overall intensity in case of hemin dissolved in NaOH compared with DMSO was detected (Fig. S1). The exposure to another thiol-containing substance – glu- tathione caused less pronounced spectral changes in the UV–vis spectrum of free hemin (Fig. 1A). To determine whether thiol- containing substances (cysteine and glutathione) are chemically stable in PBS solution at pH 7.4, Elman’s assay was applied (see Supplementary information). We concluded that within the typical

time of titration experiment (40–60 min) the cysteine and glu- tathione stock solutions did no undergo detectable oxidation and can serve as reliable spectroscopic probes (Supple. Table 1 and Fig.

S2).

Further, we examined the effects of low molecular heme ligands on spectral characteristics of heme bound to various proteins. The exposure of human hemoglobin to cyanide resulted in a typical red shift in the Soret region (Fig. 1B). The treatment of the hemoprotein with reduced glutathione had also a profound effect on the UV–vis absorbance spectrum of hemoglobin. Thus, there was a consider- able reduction of the prototypic absorbance peak at 405 nm and appearance of a novel peak at 364 nm (Fig. 1B). Notably, the incu- bation of hemoglobin with an excess of cysteine did not induce any changes in the UV–vis absorbance spectrum (Fig. 1B). Further, we studied the interaction of heme with the high affinity heme scav- enger hemopexin (Fig. 1C). As demonstrated previously, interaction of heme with hemopexin resulted in considerable spectral changes characterized by a blue shift and a pronounced increase in the absorbance intensity around 410 nm [14]. The addition of excess of cyanide or cysteine, however, did not induce considerable modifi- cation of the spectral characteristics of heme-bound to hemopexin.

Nevertheless, reduced glutathione had a profound effect on the

Fig. 3. Evaluation of heme-binding capacity of human plasma proteins. Human serum albumin, human hemopexin, human pooled IgG and human monoclonal IgG1 at concentrations ranging from 0 to 40 ␮M (320 ␮M in the case of pIgG) were incubated in the presence of 20 ␮M hemin. UV-absorbance spectra were recorded in the absence and in the presence of 10 mM cysteine. (A) Differential spectra obtained after subtraction of heme-protein spectrum from the corresponding spectrum in the presence of excess of cysteine. The black lines represent the differential spectra of hemin in the absence of protein. (B) Protein concentration-dependent changes in absorbance maxima at 364 nm derived from the differential spectra. (C) Quantification of available free hemin (gray area of the graphs) in the presence of increasing concentrations of plasma proteins. The quantification analyses were based on the curves displayed on panel B. All measurements were performed in PBS pH 7.4. In all cases UV–vis absorbance spectra of hemin were recorded in the wavelength range 300–700 nm in quartz optical cell with 1 cm path length. The figure shows representative results from two independent experiments.

UV–vis absorbance of heme-hemopexin complex (Fig. 1C). Finally we studied the most abounded heme binding protein in plasma concentration, which is albumin. The exposure of heme-albumin complex to any of the studied low molecular weight heme ligands did not induce any considerable changes of the UV–vis absorbance spectrum (Fig. 1D). Collectively, these data suggested that among studied ligands of heme only cysteine did not affect the spectral properties of heme when in complex with the most pathologically relevant heme-binding proteins. However, when cysteine inter- acts with free hemin there is an appearance of particular spectral changes, which are distinguishable from the changes occurring upon heme binding to the proteins themselves. Consequently, we considered cysteine as the most appropriate probe for developing a strategy for detection and quantification of free heme in complex systems such as blood plasma or serum.

3.2. Titration experiments

To determine the appropriate concentration of cysteine to be used as a probe for heme quantification, the UV–vis absorbance spectra of 20 ␮ M hemin with the increasing concentrations of

cysteine (0.156, 0.313, 0.625, 1.25, 2.5, 5, 10 and 20 mM) were mea- sured. Fig. 2A shows differential absorbance spectra obtained after subtraction of the absorbance spectrum of hemin alone from the absorbance of hemin in the presence of different concentrations of cysteine. The spectroscopic change in differential spectra reached maximum at cysteine concentration of 2.5 mM. We decided to use somewhat higher concentration of 10 mM cysteine to ensure a bet- ter sensitivity of the assay in a complex milieu of the human serum.

Next, we measured the UV–vis absorbance spectra of increasing

concentrations of hemin (0, 0.625, 1.25, 2.5, 5, 10 and 40 ␮ M)

before or after exposure to 10 mM of cysteine. Fig. 2B displays

differential absorbance spectra obtained after subtraction of the

absorbance of hemin alone from the absorbance spectrum of hemin

in the presence of cysteine. The obtained data clearly demonstrated

that the presence of cysteine resulted in a marked blue shift in

the absorbance at Soret region with maximum of 364 nm. The

absorbance intensity at 364 nm increased linearly with augmenta-

tion of the hemin concentration (Fig. 2A). This allowed construction

of plot of hemin concentration versus absorbance increase at

364 nm, which could be used for quantification of protein-unbound

heme.

316 R. Noé et al. / Journal of Pharmaceutical and Biomedical Analysis 172 (2019) 311–319 3.3. Validation of the spectroscopy assay by use of different

heme-binding proteins

Further, we assessed the potential to use cysteine as a probe for detection and quantification of free heme in the presence of typical plasma heme-binding proteins. Human hemopexin binds heme with extremely high affinity (K

value of < 10

M) [6].

This protein circulates in plasma at approximate concentration of 1 mg/ml (17 ␮ M). On the other hand human albumin binds heme with significantly lower affinity (K

value of 10

M) [6,15]. Despite characterized by a relatively low affinity, the interaction of heme with albumin is physiologically relevant as this protein is present in human plasma at high concentration - 35–50 mg/ml (520–750 ␮ M).

Indeed, complexes of heme with albumin have been detected in case of severe hemolytic conditions.

Besides, prototypic heme-binding plasma proteins, previous studies have demonstrated that a substantial fraction of IgG in healthy individuals is able to interact with heme [16]. Therefore, in our analyses we also included pooled human IgG (pIgG) obtained from blood of > 3000 healthy donors and as control a monoclonal human IgG1 that was formerly identified as heme-binding anti- body.

To evaluate heme-binding capacity of albumin, hemopexin and IgG, a fixed concentration of hemin (20 ␮ M) was added to increas- ing concentrations of the plasma proteins in absence or presence of a high molar excess of cysteine (10 mM). The spectra of absorbance of heme-protein complex at given concentration of protein were subtracted from the corresponding spectra of heme-protein com- plex in the presence of cysteine. The obtained differential spectra are displayed on Fig. 3A. As can be observed a decrease of protein concentrations resulted in an appearance and a progressive aug- mentation of absorbance peak at 364 nm in the differential spectra, indicating presence of free hemin. Notably, at high protein con- centrations both differential absorbance spectra of albumin and hemopexin displayed an additional peak of absorbance at 410 nm in the presence of cysteine. This peak, however, is easily distin- guishable and not interfering with the absorbance peak at 364 nm.

In order to quantify free hemin, the absorbance maxima at 364 nm typical for interaction of cysteine with free hemin were plotted versus molar concentration of proteins (Fig. 3B). These plots were further used to estimate the percentage of uncomplexed hemin (Fig. 3C). One strategy for quantification of free hemin is to use calibration curve as this displayed on Fig. 2. Another strat- egy consist in quantifying unbound hemin as proportion change in absorbance at 364 nm compared to the maximal change of the hemin in absence of protein (black lines in Fig. 3A).

The obtained data revealed that at 10 ␮ M human albumin was able to scavenge completely 20 ␮ M of hemin. Although hemopexin binds hemin with considerably higher affinity, complete scaveng- ing of detectable free heme was reached in the presence of 20 ␮ M (i.e. equimolar concentration) of hemopexin (Fig. 3C). This result can be explained by the fact that albumin possesses more than one binding site for heme. Indeed, early spectroscopy studies identi- fied the presence of second affinity binding site for heme on the human albumin [17]. In contrast, structural studies have indicated that hemopexin has only one binding site for heme [6,18]. Cysteine is able efficiently to distinguish protein-free from protein-bound heme. The quantification of free heme in the presence of pooled human IgG, indicated that polyclonal antibodies had markedly lower heme-binding capacity. Indeed, K

value of ca. 5 ␮ M for heme binding to pIgG was recently documented [19]. Consequently, we used very high concentration of pIgG (320 ␮ M). However, even this high protein concentration was not able to scavenge completely the available (20 ␮ M) of hemin (Fig. 3C). This result is consistent with the observation that only a fraction of antibodies in healthy humans are able to interact with heme [16]. Indeed, a monoclonal IgG1

Fig. 4. Evaluation of heme-binding capacity of human serum. Human serum from healthy donor was diluted in the range 20–5120 fold and exposed to 20 ␮M of hemin.

UV-absorbance spectra were recorded in the absence and in the presence of 10 mM

cysteine. The spectra of different dilutions of the human serum were also recorded

and used for correction of the turbidance effect. (A) Differential spectra obtained

after subtraction of heme-serum spectrum at given dilution from the correspond-

ing spectrum in the presence of excess of cysteine. The black line represents the

differential spectra of hemin in the absence of serum. (B) Serum dilution-dependent

changes in absorbance maxima at 364 nm derived from the differential spectra. (C)

Quantification of available free hemin (gray area of the graphs) as a function of

increasing serum dilutions. The quantification analyses were based on the curves

displayed on panel B. All measurements were performed in PBS pH 7.4. The UV–vis

absorbance spectra were recorded in the wavelength range 300–700 nm in quartz

optical cell with 1 cm path length. The figure displays representative results from

two independent experiments.

Fig. 5. Measurement of heme in serum from mice with acute intravascular hemolysis. Left panel depicts UV–vis absorbance spectra of 40 × diluted mouse serum in the absence (blue line) or presence of 10 mM cysteine (red line). The serum was obtained 24 h after induction of intravascular hemolysis. Right panel shows differential spectra after the subtraction of the UV–vis absorbance spectrum of hemolytic serum from hemolytic serum in the presence of cysteine. The UV–vis absorbance spectra were recorded in the wavelength range 300–700 nm using quartz optical cells with 1 cm path length. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

identified in our laboratory as heme-binding antibody, efficiently scavenged available 20 ␮ M of hemin at protein concentration of 10 ␮ M. The capacity of the monoclonal antibody to bind to two- molar excess of heme can be readily explained by the fact that IgG is bivalent with two identical antigen-binding sites.

Cysteine and GSH are reducing agents that can reduce bisul- fide bonds in proteins. To assess whether this process occurs in our setting, we assessed the effect of cysteine and GSH on integrity of human IgG molecule. Surface immobilized anti-human light chain antibodies (anti- ␬ chain and anti- ␭ chain) were used to capture native or the treated IgG, followed by detection of constant portion of heavy chain of immunoglobulins with another antibody spe- cific for human Fc. As can be observed on Supplementary Figure S3 exposure of human IgG to GSH or cysteine did not result in loss of structural integrity of IgG molecule i.e. there is not dissociation of heavy and light immunoglobulin chains that usually occurs upon reduction of IgG molecule.

3.4. Evaluation of heme-binding capacity of normal serum

In order to test the cysteine-based strategy for determination of protein-unbound heme in more complex and physiologically relevant system, we estimated the heme-binding capacity of nor- mal human serum. The ability of serum diluted in the range from 20 – 5120 -folds to scavenge a fixed concentration of hemin (20 ␮ M) was evaluated as described above for the individual pro- teins. The obtained data indicated that the protein-unbound hemin is detectable only following minimum 90-folds dilution of the human serum (Fig. 4). Further dilution of the serum resulted in progressively increased amounts of unbound hemin. Since the concentration of externally added hemin used in these experi- ments is fixed at 20 ␮ M, it can be estimated that undiluted normal human serum has potential to bind approximately 1.8 mM of heme. It can be deduced that albumin is responsible for the pre- dominant fraction of the heme-binding capacity of serum. Thus, the available 520–750 ␮ M of albumin would be able to buffer 1–1.5 mM of free heme. The contribution of the hemopexin to the heme-binding capacity of human serum is significantly lower – 17 ␮ M. Nevertheless, this can be sufficient to ensure transfer of extracellular heme to the liver for ultimate degradation by heme oxygenase-1 and recycling of apo-hemopexin. Other con- stituents of serum such as immunoglobulins, lipoproteins (HDL, LDL) and alpha-1-microglobulin together may also accommodate

significant quantities of heme and contribute to the remaining

> 200 ␮ M heme-binding capacity. However, it should be noted that heme can oxidize LDL and oxidized LDL is regarded as pro-atherogenic [20]. Moreover, when bound to certain hemo- proteins oxidized heme can also retain pathogenic potential [21].

The question of which form(s) of heme are pathologically rel- evant in vivo remains still open in the literature. If only heme bound to hemopexin is functionally quenched and cannot exert pro-inflammatory and cytotoxic effects, then heme-scavenging capacity of serum is rather low. However, much higher protective capacity of serum constituents can be anticipated if only protein unbound heme is pathogenic. In these considerations one should take into account that heme bound to albumin may not be recycled rapidly since the albumin’s circulatory half-life is >20 days. This implies that in cases of chronic intermittent hemolysis the heme- binding capacity of albumin might be decreased or completely saturated, regardless of transfer of part of heme from albumin to hemopexin and further transfer to liver. In this respect, the heme detection strategy proposed in the present article can be useful for evaluation of the total heme-binding capacity of serum of patients with different hemolytic diseases. We expect that as lower is the heme-binding capacity of patients’ sera as the risk for pathogenic effects of protein free or loosely bound heme is elevated.

Finally, to assess whether in condition of acute intravascular hemolysis there is free extracellular heme, we analysed serum from phenylhydrazine treated mice. Phenylhydrazine is a redox- active substance that upon administration in vivo causes abrupt and extensive hemolysis. The use of cysteine as a spectroscopy probe demonstrated that there is no any detectable free heme in mouse serum following acute intravascular hemolysis. This result is in accordance with the estimated considerable heme-binding capacity for human serum. Nonetheless, one can speculate that in conditions of chronic intermittent hemolysis, when heme-binding capacity of plasma is saturated, the free heme can be detected (Fig. 5).

3.5. Evaluation of heme-binding capacity of normal serum by apo-HRP assay

Finally, we compared the proposed approach with a strategy

that has been used for detection of free intracellular heme. Recovery

of the enzymatic activity of apo-horseradish peroxidase (apo-HRP)

provides convenient way for assessment of protein-unbound heme

318 R. Noé et al. / Journal of Pharmaceutical and Biomedical Analysis 172 (2019) 311–319

Fig. 6. Evaluation of heme-binding capacity of human serum by apo-HRP assay.

Human serum from healthy donor was diluted in the range 10–2560 fold and exposed to 10 ␮M of hemin in presence or absence of apo-HRP. (A) Peroxidase enzyme activity was assessed by oxidation of ABTS (absorbance readings at 414 nm).

The curve was generated by subtraction of the peroxidase activities of control sam- ples with separately added hemin and apo-HRP from identical serum dilutions incubated in the presence of both hemin and apo-HRP (B) Quantification of available free hemin (grey area of the graphs) as a function of increasing serum dilutions. The quantification analyses were based on the curves displayed on panel A. (For inter- pretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

[11]. We applied this technique for measurement of free heme in human serum. The obtained data demonstrated that reconsti- tution of peroxidase activity was achieved at considerably lower serum dilutions as compared with heme detected by cysteine. Thus, available heme was detected at serum dilution of 20 × and at dilu- tion of 100 × more than 90% of added heme (10 ␮ M) was available for reconstitution of the enzymatic activity of apo-HRP (Fig. 6).

This result could be explain by the fact that apo-HRP is character- ized by high heme binding affinity [11] and hence it can compete with heme bound with low affinity to serum proteins. This result once again demonstrated the feasibility of using cysteine as probe for detection of free heme with higher capacity to discriminate between protein-bund (irrespectively of affinity) from free heme.

3.6. General reflections and conclusion

The important functions of heme that is unbound to hemo- proteins in processes such as cell signalling, regulation of gene expression and cell metabolism as well as its pathophysiologi- cal relevance, have demanded development of novel assays and

approaches for heme detection and quantification. In recent years a considerable success has been achieved in the design of intracel- lular fluorescent or enzymatic probes for detection of free heme.

These probes are recombinantly expressed, have high sensitivity and selectivity for heme and have been applied successfully for the quantification of free heme and study of heme trafficking in pathogens and model organisms such as malaria parasites [22], Caenorhabditis elegans [23], and Saccharomyces cerevisiae [24]. A strategy has also been developed for measurement of hemoglobin- unbound heme in healthy human erythrocytes. Some label-free approaches for identification of heme have been also proposed [25,26].

Despite the efficient employment of techniques for detection of intracellular heme, the strategies for quantification of heme in plasma or serum are less advanced and they yield more conflict- ing results. The first-generation techniques relayed on colorimetric assay or reconstruction of peroxidase activity by heme [27]. These methods, however, cannot distinguish heme that is bound to hemoglobin from heme-bound to other proteins or free heme and are therefore appropriate only for assessment of total heme in plasma. More specific are attempts that relay on use of immunoas- says for detection of plasma hemoglobin. Again these assays are not capable to estimate heme that is in complex with plasma proteins or free heme. Other techniques relayed on fractionation of plasma and estimation of heme concentrations in protein-free fraction [28].

Recently a methodology for estimation of heme in plasma has been proposed that relay on deconvolution of data from UV–vis absorbance spectroscopy [29]. The advantage of this assay is that it can well distinguish oxidized and reduced forms of hemoglobin from rest of plasma cell-free heme. A high concentration of non- hemoglobin extracellular heme (up to 50 ␮ M in some patients) was detected by applying this technique to plasma samples from sickle cell patients. However, this approach cannot specifically estimate heme species that are not in complex with the plasma proteins.

Another recently proposed technique specifically detects free heme by using heme-binding antibody fragments (scFv) [30]. The advantage of this technique is that it determines protein free heme or heme that binds to scFvs with higher affinities than to plasma proteins. The authors utilized the technique for evalua- tion of heme-binding capacity of mouse plasma. Moreover, the same study revealed that the induction of hemolysis in C57BL/6 mice by phenylhydrazine did not result in saturation of the heme- binding capacity of mouse plasma. It was observed only a transient reduction of the heme-binding capacity of the plasma at 6 h fol- lowing induction of the hemolysis. This study also demonstrated that the heme-binding capacity of plasma was not saturated follow- ing hemolysis in mouse models of sickle cell disease and malaria [30]. Nonetheless, by applying a cellular system with peroxidase reporter authors demonstrated that in the three hemolytic models there was a release of free heme in range 2–5 ␮ M. It remains, how- ever, unclear whether reporter cells acquire directly free heme or they first interact with heme-protein complexes and then heme is transferred intracellularly.

Here we proposed alternative strategy that specifically deter- mines protein free heme. The main advantage of this technique over previous approaches is that heme-detection does not depend on the heme binding affinity for hemoprotein or plasma proteins.

Other advantages of the proposed method are that it is rapid and it requires only simple equipment and reagents.

A shortcoming of the presented method is that it has relatively low sensitivity i.e. minimal reliable detection of heme is ca. 0.5 ␮ M.

The low sensitivity may hamper the application of the strategy

for direct and precise estimation of low free heme concentrations,

as for example of those normally found intracellularly. However,

as previously reported quantities of total extracellular heme are

often in micromolar range in cases of intravascular hemolysis, the

approach can be still useful. We anticipate that the use of cysteine as a spectroscopy probe would be the most appropriate for assess- ment of the changes in the heme-binding capacity of patients’ sera.

Besides this use, our data suggest that the cysteine-based spec- troscopy approach can be utilized for estimation of the number of heme binding sites of a given protein. Likewise, the technique can be suitable for assessment of the functional quality or purity of known heme-binding proteins. We also anticipate that the pre- sented method can be valuable for study of impact of other protein ligands on interaction of heme with the proteins.

In conclusion, here we presented an approach that enables detection and quantification of free heme in presence of heme- binding proteins or blood serum. This approach is based on the peculiar spectral effect of cysteine on protein unbound free heme.

We demonstrated that the technique has a high capacity to dis- criminate between protein-bound from free heme. It can be used in clinical practice for assessment of heme-buffering capacity of serum or plasma or in the research on heme and heme-binding proteins.

Acknowledgment

This work was supported by European Research Council (grant ERC-StG-678905 CoBABATI to J.D.D.) and by INSERM, France.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jpba.2019.05.

013.

References

[1] T. Kuhl, D. Imhof, Regulatory Fe(II/III) heme: the reconstruction of a molecule’s biography, Chembiochemistry 15 (14) (2014) 2024–2035.

[2] D.A. Sears, Disposal of plasma heme in normal man and patients with intravascular hemolysis, J. Clin. Invest. 49 (1) (1970) 5–14.

[3] M.P. Soares, M.T. Bozza, Red alert: labile heme is an alarmin, Curr. Opin.

Immunol. 38 (2016) 94–100.

[4] L.T. Roumenina, J. Rayes, S. Lacroix-Desmazes, J.D. Dimitrov, Heme:

modulator of plasma systems in hemolytic diseases, Trends Mol. Med. 22 (3) (2016) 200–213.

[5] A.W. Shih, A. McFarlane, M. Verhovsek, Haptoglobin testing in hemolysis:

measurement and interpretation, Am. J. Hematol. 89 (4) (2014) 443–447.

[6] D.J. Schaer, P.W. Buehler, A.I. Alayash, J.D. Belcher, G.M. Vercellotti, Hemolysis and free hemoglobin revisited: exploring hemoglobin and hemin scavengers as a novel class of therapeutic proteins, Blood 121 (8) (2013) 1276–1284.

[7] P. Ascenzi, M. Fasano, Serum heme-albumin: an allosteric protein, IUBMB life 61 (12) (2009) 1118–1122.

[8] C.J. Mingone, S.A. Gupte, J.L. Chow, M. Ahmad, N.G. Abraham, M.S. Wolin, Protoporphyrin IX generation from delta-aminolevulinic acid elicits pulmonary artery relaxation and soluble guanylate cyclase activation, American journal of physiology, Lung Cell. Mol. Physiol. 291 (3) (2006) L337–344.

[9] S. Raghuram, K.R. Stayrook, P. Huang, P.M. Rogers, A.K. Nosie, D.B. McClure, L.L. Burris, S. Khorasanizadeh, T.P. Burris, F. Rastinejad, Identification of heme as the ligand for the orphan nuclear receptors REV-ERBalpha and

REV-ERBbeta, Nat. Struct. Mol. Biol. 14 (12) (2007) 1207–1213.

[10] A. Flint, Y.Q. Sun, A. Stintzi, Cj1386 is an ankyrin-containing protein involved in heme trafficking to catalase in Campylobacter jejuni, J. Bacteriol. 194 (2) (2012) 334–435.

[11] H. Atamna, M. Brahmbhatt, W. Atamna, G.A. Shanower, J.M. Dhahbi, ApoHRP-based assay to measure intracellular regulatory heme, Metallomics 7 (2) (2015) 309–321.

[12] A. Margoliash, Heme Schejter, Proteins, Cambridge University Press, 2001.

[13] T. Kuhl, A. Wissbrock, N. Goradia, N. Sahoo, K. Galler, U. Neugebauer, J. Popp, S.H. Heinemann, O. Ohlenschlager, D. Imhof, Analysis of Fe(III) heme binding to cysteine-containing heme-regulatory motifs in proteins, Chem. Biol. 8 (8) (2013) 1785–1793.

[14] G.M. Vercellotti, P. Zhang, J. Nguyen, F. Abdulla, C. Chen, P. Nguyen, C.

Nowotny, C.J. Steer, A. Smith, J.D. Belcher, Hepatic overexpression of hemopexin inhibits inflammation and vascular stasis in murine models of sickle cell disease, Mol. Med. 22 (2016) 437–451.

[15] P.A. Adams, M.C. Berman, Kinetics and mechanism of the interaction between human serum albumin and monomeric haemin, Biochem. J. 191 (1) (1980) 95–102.

[16] J.D. Dimitrov, L.T. Roumenina, V.R. Doltchinkova, N.M. Mihaylova, S.

Lacroix-Desmazes, S.V. Kaveri, T.L. Vassilev, Antibodies use heme as a cofactor to extend their pathogen elimination activity and to acquire new effector functions, J. Biol. Chem. 282 (37) (2007) 26696–26706.

[17] M. Rosenfeld, D. Surgenor, Methemalbumin. Interaction between human serum albumin and ferriprotoporphyrin IX, J. Biol. Chem. 183 (1950) 663–677.

[18] M. Paoli, B.F. Anderson, H.M. Baker, W.T. Morgan, A. Smith, E.N. Baker, Crystal structure of hemopexin reveals a novel high-affinity heme site formed between two beta-propeller domains, Nat. Struct. Biol. 6 (10) (1999) 926–931.

[19] M. Wiatr, N.S. Merle, I. Boudhabhay, V. Poillerat, S. Rossini, M. Lecerf, S.V.

Kaveri, S. Lacroix-Desmazes, L.T. Roumenina, J.D. Dimitrov, Anti-inflammatory activity of intravenous immunoglobulin through scavenging of heme, Mol.

Immunol. (2019), http://dx.doi.org/10.1016/j.molimm.2019.04.020 (in press).

[20] F. Vinchi, M.U. Muckenthaler, M.C. Da Silva, G. Balla, J. Balla, V. Jeney, Atherogenesis and iron: from epidemiology to cellular level, Front.

Pharmacol. 5 (2014) 94.

[21] J. Balla, H.S. Jacob, G. Balla, K. Nath, J.W. Eaton, G.M. Vercellotti,

Endothelial-cell heme uptake from heme proteins: induction of sensitization and desensitization to oxidant damage, Proc. Natl. Acad. Sci. U. S. A. 90 (20) (1993) 9285–9289.

[22] J.R. Abshire, C.J. Rowlands, S.M. Ganesan, P.T. So, J.C. Niles, Quantification of labile heme in live malaria parasites using a genetically encoded biosensor, Proc. Natl. Acad. Sci. U. S. A. 114 (11) (2017) E2068–E2076.

[23] X. Yuan, N. Rietzschel, H. Kwon, A.B. Walter Nuno, D.A. Hanna, J.D. Phillips, E.L.

Raven, A.R. Reddi, I. Hamza, Regulation of intracellular heme trafficking revealed by subcellular reporters, Proc. Natl. Acad. Sci. U. S. A. 113 (35) (2016) E5144–5152.

[24] D.A. Hanna, R.M. Harvey, O. Martinez-Guzman, X. Yuan, B. Chandrasekharan, G. Raju, F.W. Outten, I. Hamza, A.R. Reddi, Heme dynamics and trafficking factors revealed by genetically encoded fluorescent heme sensors, Proc. Natl.

Acad. Sci. U. S. A. 113 (27) (2016) 7539–7544.

[25] U. Neugebauer, A. Marz, T. Henkel, M. Schmitt, J. Popp, Spectroscopic detection and quantification of heme and heme degradation products, Anal.

Bioanal. Chem. 404 (10) (2012) 2819–2829.

[26] A.J. Chen, X. Yuan, J. Li, P. Dong, I. Hamza, J.X. Cheng, Label-free imaging of heme dynamics in living organisms by transient absorption microscopy, Anal.

Chem. 90 (5) (2018) 3395–3401.

[27] S.A. Kuross, B.H. Rank, R.P. Hebbel, Excess heme in sickle erythrocyte inside-out membranes: possible role in thiol oxidation, Blood 71 (4) (1988) 876–882.

[28] O.A. Adisa, Y. Hu, S. Ghosh, D. Aryee, I. Osunkwo, S.F. Ofori-Acquah, Association between plasma free haem and incidence of vaso-occlusive episodes and acute chest syndrome in children with sickle cell disease, Br. J.

Haematol. 162 (5) (2013) 702–705.

[29] J.Y. Oh, J. Hamm, X. Xu, K. Genschmer, M. Zhong, J. Lebensburger, M.B.

Marques, J.D. Kerby, J.F. Pittet, A. Gaggar, R.P. Patel, Absorbance and redox based approaches for measuring free heme and free hemoglobin in biological matrices, Redox Biol. 9 (2016) 167–177.

[30] Z. Gouveia, A.R. Carlos, X. Yuan, F. Aires-da-Silva, R. Stocker, G.J. Maghzal, S.S.

Leal, C.M. Gomes, S. Todorovic, O. Iranzo, S. Ramos, A.C. Santos, I. Hamza, J.

Goncalves, M.P. Soares, Characterization of plasma labile heme in hemolytic

conditions, FEBS J. 284 (19) (2017) 3278–3301.