The myeloperoxidase product hypochlorous acid generates irreversible high-density lipoprotein receptor inhibitors

Abstract

Objective

Elevated levels of advanced oxidation protein products (AOPPs) have been described in several chronic inflammatory diseases, like chronic renal insufficiency, rheumatoid arthritis and atherosclerosis. Recent findings revealed that AOPPs are inhibitors of the major high-density lipoprotein (HDL) receptor, scavenger receptor class B, type 1 (SR-BI). Here we investigated what oxidation induced structural alterations convert plasma albumin into an HDL-receptor inhibitor.

Approach and Results

Exposure of albumin to the physiological oxidant, hypochlorous acid, generated high affinity SR-BI ligands. Protection of albumin lysine-residues prior exposure to hypochlorous acid as well as regeneration of N-chloramines after oxidation of albumin completely prevented binding of oxidized albumin to SR-BI, indicating that modification of albumin lysine-residues is required to generate SR-BI ligands. Of particular interest, N-chloramines within oxidized albumin promoted irreversible binding to SR-BI, resulting in permanent receptor blockade. We observed that the SR-BI inhibitory activity of albumin isolated from chronic kidney disease patients correlated with the content of the myeloperoxidase-specific oxidation product 3-chlorotyrosine and was associated with alterations in the composition of HDL.

Conclusion

Given that several potential atheroprotective activities of HDL are mediated by SR-BI, the present results raise the possibility that oxidized plasma albumin, through permanent SR-BI blockade, contributes to the pathophysiology of cardiovascular disease.

Neutrophils are the first line of defense against invading microorganisms and thus accumulate in high numbers at sites of tissue injury. Neutrophils contain myeloperoxidase (MPO), which is released from activated neutrophils by intraluminal degranulation. MPO catalyses the production of reactive chlorinating species like the powerful oxidant hypochlorous acid (HOCl) (1, 2). Reaction of chlorinated oxidants with plasma proteins, leads to formation of advanced oxidation protein products (AOPPs) (3, 4). Previous studies have shown that plasma albumin may consume the majority of chlorinated oxidants with limited damage to other materials (5). Elevated levels of AOPPs have been described in several chronic inflammatory diseases, like end stage renal disease (ESRD), atherosclerotic coronary artery disease, acute-on-chronic liver failure and rheumatoid arthritis (reviewed in (6, 7)). Moreover, massive oxidation of plasma albumin was demonstrated by mass spectrometry in primary nephritic syndrome involving almost complete sulphonation of the free sulfhydryl group of Cys34 (8, 9). HOCl modifies proteins by various ways, including conversion of cysteine residues to disulphides and higher oxidation products, conversion of methionine residues to methionine sulphoxides, oxidation of tryptophan and chlorination of amino groups and tyrosine (10, 11). Given that plasma contains approximately a 100-fold excess of free amines over thiols, protein-associated N-chloramines are expected to be a major product of HOCl. N-chloramines can subsequently decompose to carbonyls via loss of hydrochlorous acid and hydrolysis of the resultant imine (12, 13).

Albumin isolated from end stage renal disease (ESRD) patients on hemodialysis, but not albumin isolated from healthy controls, markedly inhibits HDL association to scavenger receptor class B type I (SR-BI) (14), indicating that an oxidized fraction of albumin present in uremic blood is recognized by the receptor. SR-BI is expressed in hepatocytes and Kupffer cells of the liver and steroidogenic glands, as well as in endothelial cells, macrophages and dendritic cells (15-17). Several studies have shown that SR-BI is protective against atherosclerotic cardiovascular disease by acting as the primary pathway for disposal of HDL-associated cholesteryl-esters (18, 19). In addition, studies in SR-BI-deficient mice demonstrate the importance of SR-BI expression for α-tocopherol and nitric oxide metabolism, normal red blood cell maturation and female fertility (18). More recently, it has been discovered that the HDL-SR-BI tandem serves other cardioprotective mechanisms, such as promoting endothelial repair and platelet function (20, 21).

In this study, we demonstrate that HOCl-induced N-chloramine formation generates SR-BI inhibitors that can irreversibly block the receptor. The present results provide further evidence that MPO-catalyzed protein damage may contribute to the pathophysiology of cardiovascular disease.

Material and Methods

Blood collection

Blood was taken from ESRD-patients on hemodialysis (HD-patients) prior dialysis sessions and from age- and sex-matched control subjects (14) in agreement with the Institutional Review Board of the Medical University of Graz. The clinical chemistry of control subjects and ESRD-patients is given in Supplementary Table I.

Albumin Preparation

Albumin from ESRD-patients was separated from other plasma proteins by affinity chromatography using HiTrap Blue HP, 1 mL columns (GE Healthcare) according to the instructions of the manufacturer. The purity of the albumin preparations was assessed by SDS-PAGE and subsequent Coomassie staining (Supplementary Fig. I).

AOPP Assay

To 20 μL isolated albumin from patients and controls (10 mg/mL) 2 μL of glacial acetic acid was added and samples were incubated for 10 minutes at room temperature. Subsequently absorbance at 340 nm was determined on a Nano Drop 1000 spectrophotometer. Absorbance was converted into the respective AOPP concentrations by means of a standard curve ranging from 1 to 100 μM chloramine-T as described (3). AOPP concentrations were normalized to the protein concentrations in the individual samples and expressed as nmol AOPP/ mg protein.

Isolation of HDL

HDL (density range 1.125 to 1.21 g/mL) was prepared by discontinuous density ultracentrifugation of plasma obtained from ESRD-patients and control subjects as described (22). Isolated HDL was desalted using PD-10 columns and stored at 4°C for further use.

Homocitrulline (HCit), 3-chlorotyrosine (3-CT) and carboxymethyllysine (CML) quantification

Isolated albumin from HD-patients and control subjects was hydrolyzed with a low-volume hydrolysis method as described (23). Hydrolysates were resuspended in 100 μL 0.2 M Li-citrate buffer (pH 2.2) and derivatized with the EZ:faast Kit (Phenomenex, Aschaffenburg, Germany) according to the manufacturer’s instructions.

Electrospray ionization liquid chromatography tandem mass spectrometry (LC-MS/MS) was used for HCit, 3-CT and CML quantification as described previously (24-26). Results are shown in Supplementary Table II.

Amino Acid Analysis

Albumin preparations were hydrolyzed, and total amino acid analysis was performed as described (23).

Cholesteryl-ester (CE) and phospholipid (PL) quantification

Cholesteryl-ester (CE) and phospholipid content of HDL isolated from plasma of ESRD-patients and control subjects were determined enzymatically (DiaSys, Holzheim, Germany).

Modification of albumin

HOCl-modification of albumin was performed by incubation of albumin with HOCl (HOCl: albumin molar ratios 12.5:1 and 50:1) according to a previous study, showing that an HOCl:albumin ratio of ≈ 35:1 resembles oxidized-albumin isolated from end stage renal patients (29). HOCl-oxidation was performed in PBS (pH 7.4) for 30 minutes at 4°C in absence of free amino acids/carbohydrates/lipids to exclude formation of AGE-like structures (27). Thereafter the extent of modification of individual amino acids was determined by amino acid analysis (Supplementary Table III). As shown by SDS-PAGE analysis, HOCl-modified albumin was only minimally aggregated and appeared mainly as a monomer (Supplementary Fig. I). For carbamylation, albumin (in PBS; pH 7.4) was incubated with potassium cyanate (KOCN) for 4 hours at 37°C as described (26). For carboxymethyllysine (CML) modification of albumin, 10 mg albumin was dissolved in sodium phosphate buffer (0.2 M; pH 7.8) and incubated with 25 μmol sodium cyanoborohydride and 10.8 μmol glyoxylic acid for up to 45 minutes at 37°C. The modified albumin preparations were passed over PD MiniTrap G-25 columns to remove excess reactants and stored at −20°C until further use. Extent of lysine modification of the individual preparations was determined by LC-MS/MS.

Reductive methylation/regeneration of lysine-chloramines

Reductive methylation using formaldehyde as an alkylating reagent and dimethylamine borane complex for reduction was carried out using the JBS Methylation Kit (Jena Biosience, Jena, Germany) according to the manufacturer’s instructions. To reconvert lysine-chloramines to the parent amines, a 5-fold molar excess of L-methionine over HOCl was added 30 minutes after oxidation. After 1 hour incubation at 4°C, samples were passed over PD MiniTrap G-25 columns to remove excess free L-methionine.

Cysteine Protection

Albumin was incubated with a 20-fold molar excess of iodoacetamide in PBS for 1 hour at room temperature in the dark. Cysteine-protected albumin was then passed over PD MiniTrap G-25 columns to remove unreacted iodoacetamide. Cysteine protection was verified using the Thiol Detection Assay Kit from Cayman according to the manufacturer’s instructions (Supplementary Fig. II).

Fluorescamine Assay

200 μL fluorescamine (0.3 mg/mL in dioxane) was added to 600 μL of protein yielding a final protein concentration of 200 μg/mL. Upon reaction with primary amines, fluorescamine yields fluorescent conjugates. Fluorescence measurements were performed on a Flex Station II (Molecular Devices) using the excitation and emission wavelengths of 390 and 460 nm, respectively.

Labeling of HDL and albumin

AF 488 carboxylic acid-2,3,5,6-tetrafluorophenyl ester (AF488) labeling of HDL and albumin was performed according to the manufacturer’s protocol. Labeled (lipo)protein preparations were filtered using PD SpinTrap or PD MiniTrap G-25 columns to remove excess dye and stored at 4°C (HDL) or at −20°C (albumin) until further use.

For 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) labeling, 300 μg DiI were added to 1 mg HDL in PBS. The resulting mixture was incubated for 18 hours at 37°C. Labeled HDL preparations were filtered over PD MiniTrap G-25 columns to remove low molecular weight compounds, sterile filtered and stored at 4°C until further use. HDL labeling with ¹²⁵I-Na iodination was performed as described (28) using N-bromosuccinimide as the coupling agent.

HDL was labeled with [cholesteryl-1,2,6,7-³H]palmitate by CE transfer protein-catalyzed transfer from donor liposomes as described (28).

Cell Culture

SR-BI overexpressing Chinese hamster ovary (CHO) cells (LdlA(SR-BI)) and control CHO cells (LdlA7) were cultured in Ham’s F12 medium as described (14). Both cell lines were kindly provided by Dr. Monty Krieger (Massachusetts Institute of Technology, Boston, USA).

Competition experiments

AF488-HDL or DiI-labeled HDL (50 μg/mL) was added to LdlA(SR-BI) and LdlA7 cells in the presence or absence of the SR-BI inhibitors (modified albumin preparations, 1 mg/mL) for up to 4 hours. Subsequently, cells were rinsed with PBS and detached by addition of accutase for 10 minutes at 37°C. Resuspended cells were transferred into tubes containing 200 μL 1% paraformaldehyde and left for 15 minutes on ice. Finally, cells were washed with PBS and resuspended in fixative solution to assess AF488-HDL association or DiI-HDL uptake by flow cytometry. SR-BI specific AF488-HDL association or DiI-uptake was calculated by subtracting average fluorescence values measured in LdlA7 cells from those measured in LdlA(SR-BI) cells (28).

To test for irreversible binding, LdlA(SR-BI) and LdlA7 cells were incubated with AF488-labeled modified albumin preparations (100 μg/mL) for 2 hours at 4°C. To block unspecific binding, all incubations were performed in the presence of native albumin (5 mg/mL). After incubation, cells were harvested to measure AF488-albumin association, or incubated for additional 3 to 24 hours in the absence or presence of oxidized/carbamylated albumin (3 mg/mL) to demonstrate irreversible binding of oxidized AF488-albumin to SR-BI. Subsequently, cells were harvested and SR-BI specific AF488-albumin association was assessed as described above.

Western Blot

Confluent LdlA(SR-BI) and control LdlA7 cells were incubated with native albumin or HOCl-albumin for 2 hours at 37°C. Subsequently, cells were rinsed and detached with lysis buffer (150 mM NaCl, 25 mM KCl, 10 mM TRIS, 1 mM CaCl₂ and 0.1% Triton × 100) containing a protease inhibitor cocktail.

Aliquots of the individual samples were loaded onto 4-20% tris-glycine gradient gels under reducing (15% mercaptoethanol) and not-reducing conditions. Proteins were separated by SDS gel electrophoresis and transferred to polyvinylidene difluoride membranes (200 mA for 2 hours at 4°C). Membranes were washed, blocked for 1 hour in 5% fat-free dry milk and incubated with a polyclonal anti-SR-BI primary antibody (1:2000) overnight at 4°C. Membranes were washed and HRP-conjugated goat anti-rabbit IgG (1:10000) was applied as a secondary antibody for 2 hours at room temperature. Kodak BioMax light films were exposed for different time periods to visualize SR-BI protein bands.

Immunofluorescence

LdlA(SR-BI) and control LdlA7 cells were seeded in glass bottom dishes. Cells were rinsed with HBSS and incubated with 100 μg/mL AF488-labeled native or HOCl-albumin in the presence of 5 mg/mL native albumin for 2 hours at 37°C. Cells were rinsed with HBSS, fixed with 2% paraformaldehyde for 30 minutes, washed again and stored in PBS at 4°C.

For antibody staining, cells were permeabilized with 0.5% Triton × 100 in PBS for 15 minutes, rinsed and unspecific binding was prevented by incubation with 5% BSA in PBS for 1 hour. Anti-SR-BI antibody was applied in a 1:300 dilution in PBS/5% BSA without prior washing. After 3 hours, cells were washed and incubated with Texas Red-coupled goat anti-rabbit IgG (1:500 in PBS/5% BSA) for 1 hour at room temperature. Cells were rinsed and stored in PBS until microscopic analysis.

AF488-labeled native and HOCl-modified albumin and Texas Red-labeled SR-BI were visualized using a confocal imaging system (Zeiss LSM 510 Meta) with a Plan Neofluar 40×/1.3 N.A. oil-immersion objective. The excitation and emission wavelengths for AF488 were 488 nm and 510-530 nm, respectively. For Texas Red, excitation at 543 nm and emission at >560 nm were used. The optical slice thickness was set to 1.0 μm. 2D images at a central depth of the cell were collected.

³H-CE-HDL/¹²⁵I-HOCl-albumin Turnover in vivo

Following anesthesia, BALB/c mice (n = 6) were injected via the tail vein with [³H]CE-HDL (5×10⁵ cpm) in combination with either 5 mg native albumin (3 mice) or 5 mg HOCl-albumin (3 mice) in 100 μL of PBS. Blood samples were collected from anesthesized mice after 2 and 30 minutes by retroorbital puncture and analyzed by liquid scintillation counting.

Following anesthesia, BALB/c mice (n = 6) were injected via the tail vein with 100 μg ¹²⁵I-labeled native or HOCl-albumin (1×10⁷ cpm) in 100 μL PBS. To determine the amount of ¹²⁵I-albumin remaining in plasma, blood samples were collected by retroorbital puncture from anesthesized mice after 2, 10, 30 and 60 minutes and radioactivity was measured. Experimental protocols were approved by the Animal Care Committee of the Austrian State Department of Science and Research

Statistical analyses

Results are represented as mean ± SEM for the number of performed experiments (n). One-way analysis of variance (ANOVA) was used for multiple comparison and unpaired t-tests for comparison between two groups. Correlations were determined using Pearson product-moment estimates. Significance was accepted at P < 0.05. All statistical analyses were performed using GraphPad Prism Version 4.03.

Results

Modification of albumin-lysine residues transforms albumin into an SR-BI inhibitor

Oxidized plasma albumin is formed by MPO-derived HOCl in vivo (3, 27, 29, 30). Conversion of positively charged ε-amino groups of lysine residues by HOCl results in loss of positive charge through formation of chloramines. The resulting increase in negative charge might affect the interaction of albumin with SR-BI. To specifically elucidate the role of lysine-residues, lysine amine groups were either methylated prior to oxidation to prevent lysine chlorination or free methionine was added to HOCl-treated albumin (chlorinated albumin) to reverse N-chloramine formation (31). HOCl-modification of albumin generated SR-BI ligands, which effectively inhibited Alexa-Fluor 488 labeled HDL (AF488-HDL) binding to SR-BI in competition experiments (Figure 1A). However, when lysine residues of albumin were methylated prior to oxidation, the oxidation product did not interfere with HDL binding to SR-BI. Methionine treatment efficiently regenerated albumin lysine-residues (Supplemental Fig. III) and completely averted HDL displacement by oxidized albumin in competition experiments (Figure 1A). This clearly indicates that N-chloramine formation transforms albumin into an SR-BI inhibitor.

Figure 1. HOCl-induced lysine-modifications transform albumin into an SR-BI inhibitor.

(A) SR-BI overexpressing CHO cells (LdlA(SR-BI)) and control CHO cells (LdlA7) were incubated at 37°C for 2 hours with AF488-labeled HDL (50 μg/mL) in the presence of either 1 mg/mL native albumin or HOCl-albumin. To protect lysine residues from oxidation, albumin was methylated prior to oxidation (Methylated Albumin). HOCl-induced lysine-chloramine formation was reversed by addition of excess free L-methionine 30 minutes after HOCl treatment of albumin (Albumin + L-Methionine) ***P < 0.001 versus native albumin.

(B) LdlA(SR-BI) and LdlA7 cells were incubated at 37°C for 4 hours with DiI-HDL (50 μg/mL) in the presence of 1 mg/mL native albumin or HOCl-albumin. Albumin was oxidized for either 30 minutes at 4°C or for 5 days at 37°C to allow N-chloramine decomposition (aged albumin). To reverse HOCl-induced lysine-chloramine formation, L-methionine was added in excess 30 minutes after the incubation period (Albumin + L-Methionine). ***P < 0.001 versus native albumin of the respective set of samples.

(C) LdlA(SR-BI) and LdlA7 cells were incubated at 37°C for 2 hours with AF488-HDL (50 μg/mL) in the presence of 1 mg/mL native albumin, HOCl-albumin, carbamylated albumin (HCit-albumin) or carboxymethylated albumin (CML-albumin). The extent of modified lysine residues was comparable throughout the different albumin preparations (≈ 6 modified lysine residues of 59). ***P < 0.001 versus native albumin.

(D) LdlA(SR-BI) and LdlA7 cells were incubated at 37°C for 2 hours with AF488-HDL (50 μg/mL) in the presence of 1 mg/mL HOCl-albumin or HCit-albumin preparations containing indicated modified lysine residues per albumin.* P < 0.5; *** P < 0.001 versus 100% AF488-HDL association.

(E) LdlA(SR-BI) and LdlA7 cells were incubated at 37°C for 4 hours with DiI-HDL (50 μg/mL) in presence or absence of plasma which was spiked with indicated concentrations of HOCl-albumin. ***P < 0.001 versus plasma in the absence of oxidized albumin.

(A-E) AF488-HDL association to SR-BI and DiI-HDL uptake by SR-BI were assessed by flow cytometry. AF488-HDL association to SR-BI and DiI-HDL uptake by SR-BI, in absence of any competitor, were set to 100%. Average fluorescence values measured in LdlA7 cells were subtracted from those measured in LdlA(SR-BI) cells to calculate SR-BI specific HDL association and uptake. Data are represented as mean ± SEM of at least two independent experiments performed in triplicates.

(F) BALB/c mice (n=6) were injected via the tail vein with a mixture of 50 μg [³H]CE-HDL (5×10⁵ cpm) and 5 mg of either native albumin or HOCl-oxidized albumin in 100 μL of PBS. The amount of [³H]-CE-HDL remaining in plasma was assessed 2 and 30 minutes after injection by scintillation counting in plasma samples obtained by retroorbital puncture. *P < 0.05; **P < 0.01 versus native albumin.

Next, we investigated whether albumin containing N-chloramines affects SR-BI mediated HDL-lipid uptake. To address this question, we performed competition experiments with DiI-labeled HDL (lipid moiety label) in the presence of native and oxidized albumin (Figure 1B). In line with results shown in Figure 1A, HOCl-albumin effectively attenuated SR-BI-mediated HDL-lipid (DiI) uptake, which was not observed when methionine was added to reverse N-chloramine formation (Figure 1B). Chloramines gradually decompose to carbonyls via the loss of hydrochlorous acid and hydrolysis of the imine (12, 13). To study the impact of N-choramines on HDL-lipid delivery, HOCl-albumin was incubated for 5 days at 37°C to allow complete decomposition of N-chloramines and N-chloramine-derived carbonyls. The resulting “aged” HOCl-albumin preparations were used for further competition experiments. Both, freshly oxidized albumin and aged HOCl-albumin clearly reduced SR-BI-mediated DiI-HDL uptake (Figure 1B). However only the inhibitory effect mediated by freshly oxidized albumin was reversed upon addition of methionine. Based on these results we assume that both, chloramines and their decomposition products, transform albumin into an SR-BI inhibitor. To test whether positive charge elimination through lysine modification by principle renders albumin an SR-BI ligand, albumin-lysine residues were carbamylated with cyanate yielding homocitrulline (HCit) and carboxymethylated to form carboxymethyllysine (CML). To confirm that albumin-lysine residues were modified to about the same extent throughout the three different albumin modifications, lysine modification was determined using amino acid analysis and LC-MS/MS. In competition experiments, HCit-albumin displaced HDL from SR-BI (Figure 1C), providing further evidence that lysine modification is required to generate SR-BI ligands. Surprisingly, CML-albumin failed to interfere with HDL-binding to SR-BI, suggesting that the negative charge of the carboxyl-group of CML and/or structural differences of CML prevent binding to SR-BI (Supplementary Fig. IV). We observed that already a modification less than two of 59 lysine residues in HOCl-albumin was sufficient to generate SR-BI inhibitors, whereas HCit-albumin was less effective with regard to the number of modified lysine residues (Figure 1D).

Based on the assumption that only a fraction of plasma albumin is more extensively modified, we intended to determine the proportion of oxidized albumin necessary to block SR-BI. Notably, already the presence of less than 1% oxidized albumin in plasma effectively inhibited SR-BI mediated DiI-HDL uptake (Figure 1E). To demonstrate in vivo relevance of our data, we examined the plasma clearance of ³H-CE-HDL co-injected with either native albumin or oxidized-albumin in mice. We found that HOCl-albumin significantly impaired ³H-CE-HDL clearance in mice (Figure 1F). As expected, HOCl-albumin was more rapidly cleared than native albumin (Supplementary Fig. V).

HOCl-albumin irreversibly binds to SR-BI

Prompted by the observation that HOCl-albumin was a more effective SR-BI inhibitor than HCit-albumin, we assessed binding affinities of modified AF488-labeled albumin preparations to SR-BI. To inhibit receptor internalization, binding experiments were performed at 4°C. Surprisingly, binding affinities of HCit-albumin and HOCl-albumin to SR-BI were similar (Kd-values: HOCl-albumin 47.1 ± 6.3 μg/mL; HCit-albumin 34.2 ± 5.2 μg/mL) although their dissociation characteristics clearly differed (Figure 2A and 2B). About 90% of chlorinated albumin remained associated to SR-BI even after more than 20 hours, whereas the major portion of bound HCit-albumin dissociated (Figure 2B).

Figure 2. Concentration-dependent association of modified albumin to SR-BI.

(A) LdlA(SR-BI) and LdlA7 cells were incubated with the indicated concentrations of oxidized or carbamylated AF488-albumin for 2 hours at 4°C and AF488-albumin association was assessed by flow cytometry. SR-BI specific AF488-albumin association was calculated as described above. Data are represented as mean ± SEM of at least two independent experiments performed in triplicates. Native AF488-labeled albumin showed no binding to SR-BI (data not shown).

(B) Dissociation of modified albumin. LdlA(SR-BI) and LdlA7 cells were incubated with 100 μg/mL oxidized or carbamylated AF488-albumin for 2 hours at 4°C. Subsequently, cells were washed and incubated with medium containing native albumin (3 mg/mL) for up to 22 hours. The remaining cell-associated AF488-albumin was assessed by flow cytometry and SR-BI specific AF488-albumin binding was assessed as described.

Data are shown as mean ± SEM of 2 independent experiments performed in triplicates. **P = 0.0085 versus carbamylated albumin.

Moreover, a 30-fold molar excess of HOCl-albumin was not able to displace AF488 HOCl-albumin from SR-BI (Figure 3A). In contrast, labeled HCit-albumin was effectively displaced by either HCit-albumin or HOCl-albumin (Figure 3B). Taken together, these results point to an irreversible binding of HOCl-albumin to SR-BI. Similar results were obtained when plasma was spiked with AF488-albumin and then treated with HOCl. (Supplementary Fig. VI).

Figure 3. HOCl-oxidation converts albumin into a permanent SR-BI ligand.

(A) LdlA(SR-BI) and LdlA7 cells were incubated with 100 μg/mL AF488 HOCl-albumin for 2 hours at 4°C. After washing, cells were incubated for another 2 hours with a 30-fold molar excess of native or HOCl-albumin (B) LdlA(SR-BI) and LdlA7 cells were incubated with 100 μg/mL AF488 HCit-albumin for 2 hours at 4°C. After washing, cells were incubated for another 2 hours with a 30-fold molar excess of native, HCit-albumin or HOCl-albumin. *P < 0.05; **P < 0.01 versus native albumin; (C) Albumin was oxidized and incubated for 5 days at 37°C to allow N-chloramine decomposition (aged HOCl-albumin). Subsequently, LdlA(SR-BI) and LdlA7 cells were incubated with 100 μg/mL of the labeled albumin preparation for 2 hours at 4°C. After washing, the cells were incubated for another 2 hours with a 30-fold molar excess of native or HOCl-albumin. ***P < 0.001 versus native albumin; Association of AF488-labeled albumin was measured by flow cytometry and SR-BI specific binding was assessed as described. Data are shown as mean ± SEM of 2 independent experiments performed in triplicates.

To explore the relevance of N-chloramines for irreversible binding, HOCl-albumin was incubated for 5 days at 37°C (aged albumin) to allow complete decomposition of chloramines and N-chloramine-derived carbonyls as described above. Notably, aged HOCl-albumin was displaced from SR-BI by excess HOCl-albumin, indicating that albumin-chloramines mediate irreversible receptor association (Figure 3C).

HOCl-modified albumin cross-links with SR-BI

To test whether oxidized albumin covalently cross-links with SR-BI, HOCl-albumin-treated SR-BI overexpressing cells and control cells were lysed and Western blot analysis was performed. Immunochemical detection of SR-BI on blotted membranes revealed a broad and diffuse high-molecular weight band, presumably indicating SR-BI/HOCl-albumin cross-links (Figure 4). Importantly, the high-molecular weight band was not detected when control LdlA7 cells were treated with HOCl-albumin or when SR-BI overexpressing cells were treated with native albumin. Given that the high-molecular weight band of SR-BI was much less intense when SDS-PAGE was performed under reducing conditions, disulphide bonds might, at least in part, mediate the irreversible binding of HOCl-albumin to SR-BI.

Figure 4. HOCl-treatment of albumin leads to cross-link formation.

LdlA(SR-BI) and LdlA7 cells were incubated with 1 mg/mL native or HOCl-albumin for 2 hours at 37°C. Cells were rinsed, lysed and cell lysates were subjected to SDS PAGE (4-20%) under reducing and not-reducing conditions. Separated proteins were transferred onto polyvinylidene difluoride membranes for SR-BI detection. The arrow indicates monomeric SR-BI.

Due to its susceptibility to oxidation, it is conceivable that oxidative modification of the free cysteine-residue (Cys34) of albumin affects intermolecular cross-link formation with SR-BI. To address this notion, albumin cysteine-residues were alkylated using iodoacetamide prior to HOCl treatment (32). However, competition experiments revealed that alkylating the free cysteine-residue of albumin did not inhibit irreversible binding, suggesting that Cys34 oxidation is not required for cross-link formation (Supplementary Fig. VII).

HOCl-albumin colocalizes with SR-BI in LdlA(SR-BI) cells

To assess whether cells internalize SR-BI/HOCl-albumin complexes, LdlA(SR-BI) and control LdlA7 cells were incubated with AF488-labeled HOCl-albumin for immunohistochemical analysis. Immunohistochemical staining showed that AF488-labeled HOCl-albumin exhibited intracellular colocalization with SR-BI in LdlA(SR-BI) cells, suggesting receptor internalization and efficient uptake of HOCl-albumin (Figure 5). SR-BI association of native AF488-albumin was insufficient to be detected under these experimental conditions. Control LdlA7 cells showed very low uptake of HOCl-albumin (Figure 5).

Figure 5. AF488-labeled HOCl-albumin colocalizes with SR-BI in LdlA(SR-BI) cells.

LdlA(SR-BI) and LdlA7 cells were incubated with 100 μg/mL AF488-labeled native or HOCl-albumin for 2 hours at 37°C. Cells were rinsed, fixed with 2% paraformaldehyde for 30 minutes and permeabilized with Triton × 100 in PBS for 15 minutes. Anti-SR-BI antibody (1:300 in PBS/5% BSA) was applied for 3 hours at room temperature under light exclusion and without prior washing. Texas Red-coupled goat anti-rabbit IgG (1:500 in PBS/5% BSA) was used as secondary antibody.

The SR-BI inhibitory activity of albumin isolated from end stage renal disease (ESRD) patients correlates with the content of 3-chlorotyrosine and the composition of HDL

MPO may not only oxidize proteins locally in inflamed tissue sections, as evidenced by increased levels of circulating AOPPs described in chronic inflammatory diseases, like chronic kidney disease (3). Accordingly we isolated albumin from ESRD-patients on hemodialysis (HD) and determined levels of the MPO fingerprint 3-CT. A more detailed characterization of isolated albumin is given in Supplementary Table II. In line with previous results (14), we observed that HD-albumin interferes with ¹²⁵I-HDL binding to SR-BI (Figure 6A). Herein we demonstrate that the SR-BI inhibitory activity of HD-albumin, measured as the efficacy to attenuate ¹²⁵I-HDL association to SR-BI, significantly correlated with the MPO specific oxidation product 3-chlorotyrosine (Figure 6B). A previous study reported that SR-BI deficiency in mice resulted in the accumulation of cholesteryl-ester-enriched and phospholipid-depleted HDL particles (33). We wondered whether blockade of SR-BI through HD-albumin might be associated with an altered HDL-lipid composition in ESRD patients. To address this question, we isolated HDL from plasma of controls and ESRD-patients, assessed the HDL-CE and HDL-PL content and examined the relationship between the SR-BI inhibitory activity of HD-albumin and the HDL-CE/PL ratio. Of particular interest, we observed that the SR-BI inhibitory activity of HD-albumin significantly correlated with the CE/PL composition of isolated HDL (Figure 6C).

Figure 6. Albumin isolated from ESRD patients interferes with HDL binding to SR-BI.

(A) LdlA(SR-BI) cells and control LdlA7 cells were incubated in DMEM for 2 hours at 37°C with ¹²⁵I-HDL (50 μg of protein/mL) in the presence of 1 mg/mL albumin isolated from plasma of ESRD-patients on hemodialysis (HD) (n=12) and control subjects (n=10). Values measured in LdlA7 cells were subtracted from those obtained in LdlA(SR-BI) cells to calculate SR-BI specific binding. Data are represented as mean ± SEM of two independent experiments performed in triplicates. (B) 3-CT content of the individual albumin-samples was plotted against SR-BI inhibitory activity of individual albumin samples (expressed as SR-BI specific ¹²⁵I-HDL-association). (C) CE/PL ratio of HDL isolated from HD-patients and control subjects was plotted against SR-BI inhibitory activity of individual albumin samples (expressed as SR-BI specific ¹²⁵I-HDL-association).

Discussion

MPO released by degranulation of activated human neutrophils avidly binds to cell membrane glycosaminoglycans, therefore plasma protein oxidation in close vicinity to cells may have important implications for the regulation of cellular function during inflammation (34). Of particular interest, MPO may not only locally oxidize proteins, as shown by increased levels of circulating AOPPs described in several chronic inflammatory diseases (6, 7).

Our study provides the novel evidence that HOCl-induced oxidation of plasma albumin generates irreversible HDL-receptor inhibitors. Previous studies referred to cysteine oxidation as a trigger of HOCl-induced intramolecular and intermolecular sulfenamide-, sulfinamide-, and sulphonamide-generation (35, 36). Our results suggest that oxidation of Cys34 in albumin is unlikely to be involved in cross-link formation with SR-BI, since methylation of Cys34 prior to oxidation did not inhibit irreversible binding to the receptor. However, given that two of the six exoplasmic cysteines of SR-BI are free thiol groups (Cys251 and Cys384) (37), it is tempting to speculate that albumin-chloramines may oxidize SR-BI Cys251 and/or Cys384 thereby forming labile intermolecular sulfenamides or more oxidized sulfinamides / sulfonamides. In line with that assumption is the finding that N-chloramines preferentially react with cysteine residues in plasma and isolated proteins (38). It is of interest to note that protein thiols are decreased in a number of inflammatory diseases (6, 7), which could be due to generation of long-lived, thiol-specific N-chloramine derivatives of HOCl (38). When SR-BI overexpressing cells were incubated with HOCl-albumin, we detected high-molecular weight complexes containing SR-BI by Western blot analysis. Addition of mercaptoethanol markedly reduced the high-molecular weight bands, favouring the assumption that free Cys251 and Cys384 residues of SR-BI may contribute to the HOCl-albumin/SR-BI cross-link formation. However, further studies are required to fully clarify the underlying mechanism(s). A previous study reported that HOCl may induce protein aggregation due to strong, non-covalent interactions between oxidized protein chains (39). Other studies have proposed cross-linking via dityrosine formation or Schiff base formation between proteins through chloramine-derived aldehydes and lysine residues (40). In the present study, we observed that HOCl-albumin, which was incubated at 37°C to induce break-down of chloramines, reversibly binds to SR-BI. Given that N-chloramines decompose into the carbonyl aminoadipic semialdehyde, Schiff base formation may not significantly contribute to albumin-SR-BI cross-link formation.

HOCl reacts with amines at a lower rate than with thiols (41). On the other hand, plasma contains approximately a 100-fold excess of free amines over thiols, therefore a considerable number of protein-associated N-chloramines is expected to be formed. HOCl-mediated oxidation of the major plasma protein, albumin, was previously investigated, both experimentally and computationally, predicting that plasma proteins consume the majority of HOCl with limited damage to other materials (5). Ascorbate or α-tocopherol, even at levels achieved in human supplementation studies, do not attenuate these reactions. On the contrary, elevated thiocyanate levels, as detected in heavy smokers, can modulate HOCl-mediated reactions resulting in formation of hypothiocyanous acid and cyanate, which causes protein carbamylation and oxidation of thiol groups (5, 25, 42). In addition, MPO was shown to promote formation of glycolaldehyde and other reactive aldehydes and may therefore play a pathogenic role by generating advanced glycation end products like CML (12, 43).

Modification/oxidation of lysine-residues alters structural properties of proteins through elimination of positive charge, thereby affecting protein-water or protein-protein interaction. While oxidation and/or carbamylation neutralize the positive charge on the amino group of lysine, reductive methylation selectively modifies lysine residues of albumin without altering the positive charge. Interestingly, when reductively-methylated albumin was oxidized with HOCl, it did not bind to SR-BI, despite oxidation of cysteine, methionine, tyrosine, histidine, and arginine residues (11). This clearly supports the idea that alterations in the particular distribution of charged lysine-residues through post-translational modifications may be crucial for the transformation of albumin into an SR-BI ligand. Given that HOCl-albumin is recognized by the receptor for advanced glycation end-products and CD36 (27, 44), our data raise the possibility that HOCl-modified albumin also irreversibly cross-links with these receptors. Of particular interest, intravenous injections of HOCl-albumin in unilateral nephrectomy rats activated the intrarenal renin-angiotensin system via a CD36-mediated pathway and induced podocyte apoptosis in mice via the receptor of advanced glycation end products (44, 45).

The observation that SR-BI can be irreversibly blocked by oxidized albumin may be of particular relevance, since SR-BI expressed on hepatocytes functions as a positive regulator of cholesterol efflux from macrophages (46). Therefore, oxidized/modified albumin might contribute to cholesterol retention in the arterial wall and alter composition of HDL, as observed in the present study. This is in accordance with a recent studies reporting that compositional alterations in HDL may significantly impair HDL functionality in ESRD (22, 47). In addition, SR-BI is highly expressed in lipid-laden macrophages in atherosclerotic tissue (48); hence blockade of SR-BI-mediated cholesterol efflux may contribute to foam cell formation in the vessel wall. Besides the importance of SR-BI in reverse cholesterol transport, the HDL-SR-BI tandem confers additional atheroprotective actions. For instance, HDL binding to SR-BI protects endothelial cells from apoptosis (20). A previous study reported that treatment of hypercholesterolemic rabbits with repeated intravenous injections of HOCl-albumin significantly increased macrophage infiltration and smooth muscle cell proliferation in atherosclerotic plaques. Interestingly, this was associated with significantly increased plasma total cholesterol, which is mainly carried by HDL in rabbits (49).

In conclusion, the present study demonstrates that modified lysine-residues mediate albumin-SR-BI interaction and that N-chloramines transform plasma albumin into an irreversible SR-BI inhibitor. Blockade of SR-BI through oxidized plasma albumin may alter HDL composition and probably HDL functionality. Given that several potential atheroprotective activities of HDL are mediated by SR-BI, a permanent SR-BI blockade may therefore contribute to the pathophysiology of cardiovascular disease.

Significance.

Scavenger receptor class B, type 1 (SR-BI) mediates cellular uptake of HDL-lipids, is critically involved in reverse cholesterol transport and maintains endothelial function. We describe the novel observation that exposure of plasma albumin to the leukocyte product hypochlorous acid, generates irreversible SR-BI inhibitors. We demonstrate that the SR-BI inhibitory activity of albumin isolated from end stage renal disease patients correlates with the content of 3-chlorotyrosine, a myeloperoxidase specific oxidation product and is associated with alterations in the composition of HDL. Given that several atheroprotective activities of HDL are mediated by SR-BI, receptor blockade through an oxidized fraction of plasma albumin may contribute to the pathogenesis of cardiovascular disease.

List of abbreviations

AF488

Alexa Fluor488 carboxylic acid-2,3,5,6-tetrafluorophenyl ester

AGE

advanced glycation end-product

AOPP

advanced oxidation protein product

CE

cholesteryl-ester

CHO

chinese hamster ovary

CML

carboxymethyllysine

3-CT

3-chlorotyrosine

DiI

1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate

ESRD

end stage renal disease

FBS

fetal bovine serum

HCit

homocitrulline

HD

hemodialysis

HDL

high-density lipoprotein

HOCl

hypochlorous acid

KOCN

potassium cyanate

LC-MS/MS

liquid chromatography tandem mass spectrometry

MPO

myeloperoxidase

PL

phospholipid

SR-BI

scavenger receptor class B type I