S100A8 and S100A9—oxidant scavengers in inflammation

Free Radical Biology and Medicine 58 (2013) 170–186

Contents lists available at SciVerse ScienceDirect

Free Radical Biology and Medicine journal homepage: www.elsevier.com/locate/freeradbiomed

Original Contribution

S100A8 and S100A9—oxidant scavengers in inflammation Lincoln H. Gomes a,1, Mark J. Raftery b,1, Wei Xing Yan a,2, Jesse D. Goyette a,3, Paul S. Thomas a, Carolyn L. Geczy a,n a b

Inflammation and Infection Research Centre, University of New South Wales, Sydney, NSW 2052, Australia Bioanalytical Mass Spectrometry Facility, University of New South Wales, Sydney, NSW 2052, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 June 2012 Received in revised form 24 November 2012 Accepted 17 December 2012 Available online 28 December 2012

S100A8 and S100A9 are generally considered proinflammatory. Hypohalous acids generated by activated phagocytes promote novel modifications in murine S100A8 but modifications to human S100A8 are undefined and there is no evidence that these proteins scavenge oxidants in human disease. Recombinant S100A8 was exquisitely sensitive to equimolar ratios of HOCl, which generated sulfinic and sulfonic acid intermediates and novel oxathiazolidine oxide/dioxide forms (mass additions, m/z þ 30 and þ 46) on the single Cys42 residue. Met78(O) and Trp54( þ 16) were also present. HOBr generated sulfonic acid intermediates and oxidized Trp54( þ16). Evidence for oxidation of the single Cys3 residue in recS100A9 HOCl was weak; Met63, Met81, Met83, and Met94 were converted to Met(O) in vitro. Oxidized S100A8 was prominent in lungs from patients with asthma and significantly elevated in sputum compared to controls, whereas S100A8 and S100A9 were not significantly increased. Oxidized monomeric S100A8 was the major component in asthmatic sputum, and modifications, including the oxathiazolidine adducts, were similar to those generated by HOCl in vitro. Oxidized Met63, Met81, and Met94 were variously present in S100A9 from asthmatic sputum. Results have broad implications for conditions under which hypohalous acid oxidants are generated by activated phagocytes. Identification in human disease of the novel S100A8 Cys derivatives typical of those generated in vitro strongly supports the notion that S100A8 contributes to antioxidant defense during oxidative stress. Crown Copyright & 2013 Published by Elsevier Inc. All rights reserved.

Keywords: S100A8 S100A9 Protein oxidation Hypohalous acids Anti-inflammatory Free radicals

S100A8 and S100A9 (also known as MRP8 and MRP14) may act as damage-associated molecular pattern molecules that trigger inflammatory responses [1,2], possibly via NF-kB activation [3]. These comprise 40% of neutrophil cytosolic proteins and some 10fold less in monocytes and are upregulated in numerous cell types by particular agonists or by stress responses. They are released from phagocytes undergoing necrosis [4] or secreted after activation [5], and high plasma levels of S100A8, S100A9, or the heterocomplex

Abbreviations: BALF, Bronchoalveolar lavage fluid; CF, Cystic Fibrosis; CS, Corticosteroid; DTT, Dithiothreitol; GST, Glutathione S-Transferase; oxS100A8, HOCl-oxidized S100A8; ESI–MS, Electrospray Ionization–mass Spectrometry; LC–MS, Liquid Chromatography–Mass Spectrometry; LPS, Lipopolysaccharide; MC, Mast Cell; Met(O), Methionine Sulfoxide; mS100A8, Murine S100A8; MS/MS, Tandem Mass Spectrometry; NO, Nitric oxide; ox, Oxidized; rec, Recombinant; ROS, Reactive Oxygen Species; TFA, Trifluoroacetic Acid; TLR, Toll-Like Receptor; TOF, Time of Flight; Trp( þ16), Oxidation of Trp by addition of one oxygen n Corresponding author. Fax: þ612 9385 1389. E-mail address: [email protected] (C.L. Geczy). 1 These authors contributed equally to this work. 2 Present address: Nerve Research Foundation, University of Sydney, Sydney, NSW 2006, Australia. 3 Present address: Sir William Dunn School of Pathology, South Parks Road, Oxford OX1 3RE, UK.

(S100A8/S100A9, also known as calprotectin) are suggested to be markers of inflammatory disease activity. Numerous extracellular functions are proposed [6] and may depend on their immediate environment, concentration, and structural forms; some require heterocomplex formation [7,8]. Extracellularly, S100A8 and S100A9 are implicated in neutrophil adhesion and transmigration, activation, and apoptosis of some cell types [9,10]. Calprotectin is antimicrobial [11,12] and anti-invasive [13]. S100A8 may propagate inflammation by interacting with Toll-like receptor 4 (TLR-4); S100A9 negates this effect, although the heterocomplex potentiates lipopolysaccharide (LPS) activation of macrophages [10]. Others report that S100A9 is the active component in TLR-4 activation [14]. The receptor for advanced glycation end products is another putative receptor [15]. S100A8 and S100A9 are considered ‘‘adaptive’’ genes [16], but in marked contrast to proinflammatory mediators, S100A8 induction in murine and human macrophages stimulated with TLR-3 or -4 agonists is a secondary event that is IL-10 and protein kinase R dependent [17]. They are also modulated by oxidative stress in some cells [18]. Importantly, corticosteroids (CSs) upregulate these proteins in macrophages in vitro and in vivo [19], and S100A8 is prominent in lungs of mice treated with LPS and CSs [20], suggesting anti-inflammatory roles. In keeping with this,

0891-5849/$ - see front matter Crown Copyright & 2013 Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.freeradbiomed.2012.12.012

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

murine S100A8 (mS100A8) inhibited IgE-mediated mast cell (MC) activation by scavenging reactive oxygen species (ROS) essential for signaling, and intranasal administration suppressed acute murine asthma by reducing key cytokines affecting eosinophil migration and mucus production [21]. We propose that oxidative modifications in S100A8 and S100A9 may regulate key inflammatory processes [22,23]. Murine S100A8 is a two-electron scavenger [8,24,25] of ROS and also scavenges nitric oxide (NO). S-nitrosylated S100A8 can transport NO and may modulate vessel homeostasis and reduces MC activation and leukocyte transmigration in the inflamed microcirculation [22]. mS100A8 also scavenges reactive halides, forming covalent intra- and interprotein disulfide and sulfinamide bonds [25], and is more susceptible to oxidation by HOCl than low-density lipoproteins, and may be atheroprotective [26]. ROS also promote structural changes in S100A9 that alter fugetactic activity [27,28] and S100A9 is S-glutathionylated by S-nitrosoglutathione, a modification that reduced neutrophil adhesion to fibronectin [23]. Thus, particular oxidative modifications of these proteins may limit tissue damage. Epithelial cells and tracheal goblet cells from cystic fibrosis (CF) patients express S100A8/S100A9 [29] and they are induced in LPS-activated bronchial epithelial cells [29]. Proteomic studies identified these S100s in sputum from patients with bronchiectasis, CF [30], or asthma [31] and in bronchoalveolar lavage fluid (BALF) from patients with acute respiratory distress syndrome, CF, or chronic obstructive lung disease, and their differential expression is proposed to regulate distinct mechanisms in respiratory inflammation [32]. Oxidative stress and an imbalance of antioxidant defense mechanisms contribute to asthma progression, exacerbating inflammation and decreasing lung function [33]. ROS are principally generated by activated granulocytes and macrophages via NADPH oxidase, myeloperoxidase, and eosinophil peroxidase. HOCl and HOBr promote common modifications such as 3-chloro/bromotyrosine, which are associated with asthma severity [34]. Until now, the ROS-scavenging capacity of human S100A8 and S100A9 has been unproven in human disease although crosslinked forms implicating this were found in diseased carotid arteries [26]. Here we show that the main products of HOCl oxidation of recombinant (rec) human (h) S100A8 and S100A8 from asthmatic sputum contained sulfinic and sulfonic acid intermediates and novel oxathiazolidine oxide/dioxide adducts on the single Cys42 residue. Met1 and Met78 and Trp54 were also oxidized. HOBr generated sulfonic acid intermediates and Trp54( þ16). S100A8, oxS100A8, and S100A9 were present in asthmatic, but not normal, human lung, and asthmatic sputum had significantly more oxS100A8 than controls. In contrast, HOCl oxidation of the single Cys3 residue in S100A9 was weak; Met63, Met81, Met83, and Met94 were converted to Met(O) in vitro; Met63, Met81, and Met94 were variously present in asthmatic sputum.

Materials and methods Preparation of oxidized S100s Rec-hS100A8, S100A9, and the corresponding Cys42 to Ala42 mutant of hS100A8 were expressed as GST-fusion proteins in Escherichia coli and purified as described [35]. Preparations were stored at  80 1C under argon to prevent oxidation. To investigate differences in reactivity of HOBr and HOCl to rec-hS100A8, protein (2.5 mg) was oxidized (on ice) with 1:1 molar ratios of these, for 15, 30, and 60 min. OBr  solution was generated within 30 min of use, by mixing equimolar amounts of OCl  with Br  at pH 9 in deionized water as described [36].

171

Because of HOBr/OBr  instability, solutions were adjusted to pH 7.4 with HCl immediately before use; the pK of HOBr is 8.7, and at pH 7.4 it is mostly HOBr [36]. Oxidation was quenched with a 10fold molar excess of L-Met in phosphate-buffered saline (PBS; Sigma, St. Louis, MO, USA). Oxidized protein was then separated by SDS–PAGE under reducing (100 mM DTT) or nonreducing conditions and silver-stained. Unoxidized protein incubated for the same times served as control. For HOCl oxidation and generation of oxidized monomer, 10 nmol rec-hS100A8 (100 mg) or rec-hS100A9 (140 mg) in 100 ml PBS was treated with HOCl (molar ratio 1:1, 1:2, or 1:4) on ice for 15 min, as described [26]. For HOBr oxidation 1:1 or 1:10 molar ratios of HOBr:S100A8 were used. Incubations were for 15 min on ice. Products were eluted from C4-RP-HPLC (25/70% acetonitrile/ water); oxS100A8 monomer and dimer eluted at 14.4 and 14.8 min, oxS100A9 monomer and dimer at 14.8 and 15.7 min. Silver-stained SDS–PAGE gels confirmed separation of oxidized monomers and dimers. Peptide generation by in-solution digestion Because conventional analysis of in-gel digests with trypsin was unsuccessful, the following protocol was developed to detect putative modifications generated in S100A8 and S100A9 in vitro. Rec-oxS100A8 and rec-oxS100A9 monomers and dimers were eluted from C4-RP-HPLC (25/70% acetonitrile/water); oxS100A8 monomer and dimer eluted at 14.4 and 14.8 min, respectively, and oxS100A9 monomer and dimer at 14.8 and 15.7 min, respectively. Protein peaks (5 mg) were lyophilized, suspended in 20 mM NH4HCO3 (100 ml), digested for 4 h at 37 1C with 50 ng AspN or trypsin, and analyzed by mass spectrometry (MS; see below). AspN was chosen as a second enzyme to compare its effectiveness with trypsin in in-solution cleavage. MS analysis Protein mass measurements (non-protease-treated recS100 protein) were acquired using an API QStar Pulsar i hybrid tandem mass spectrometer (Applied Biosystems, VIC, Australia). Lyophilized samples (1–5 pmol) were dissolved in water:acetonitrile (50:50 v/v, 0.1% acetic acid) and loaded into nanospray needles (Proxeon, Denmark) and the tip was positioned 10 mm from the orifice. Nitrogen was used as curtain gas and a potential of 900 V applied to the needle. A time-of-flight (TOF)–MS scan was acquired (m/z 250–2000, 1 s) and accumulated for  30 s into a single file. Spectra were deconvoluted using the Bayesian reconstruct algorithm. In some cases, spectra were recorded using liquid chromatography–mass spectrometry (LC–MS) with an 1100 MSD as described [23]. Peptides generated by in-solution digest were separated by nano-LC using an Ultimate HPLC and Famos autosampler system (LC Packings/Dionex, NSW, Australia). Samples (  200 fmol) were concentrated and desalted onto a micro-C18 precolumn (500 mm  2 mm; Michrom Bioresources) with H2O:CH3CN (98:2, 0.05% HFBA) at 20 ml/min. After a 4 min wash the precolumn was switched (Switchos, LC Packings) into line with a fritless nanocolumn manufactured according to Gatlin et al. [37]. Peptides were eluted using a linear gradient of H2O:CH3CN (98:2, 0.1% formic acid, to 36:64, 0.1% formic acid) at  300 ml/min over 30 min. High voltage (2300 V) was applied to low-volume tee (Upchurch Scientific) and a column tip positioned  1 cm from the orifice of an API QStar Pulsar i hybrid tandem mass spectrometer (Applied Biosystems). Positive ions were generated by electrospray and the QStar was operated in informationdependent acquisition mode. A TOF–MS survey scan was acquired

172

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

(m/z 350–1700, 1 s). The two largest multiply charged ions (counts 415) were sequentially selected by Q1 for tandem MS (MS/MS) analysis. Nitrogen was used as collision gas and an optimum collision energy chosen (based on charge state and mass). Tandem mass spectra were accumulated for 2.5 s (m/z 65–2000) with up to two repeat spectra before dynamic exclusion for 45 s. Peak lists were generated using Mascot Distiller (Matrix Science) using the default parameters and submitted to the database search program Mascot (version 2.3; Matrix Science). Manual inspection identified Cys-containing peptides with additions of m/z 30 and 46 Da and modifications containing these mass additions were created and added to the Mascot server and searches repeated. Search parameters were precursor and product ion tolerances of 70.25 and 0.2 Da, respectively; Met(O); Cyscarboxyamidomethylation; sulfenic acid; sulfinic acid; oxathiazolidine oxide þ30; oxathiazolidine dioxide þ46; Met-sulfone; followed by dityrosine; chlorotyrosine derivatives; carbonylation of Lys, Pro, Arg, and Thr and bromination of Lys, His, Arg, Asn, Gln, and Tyr; and oxidation (addition of oxygen, þ16 Da) of His and Trp. Spectra were analyzed with these variable modifications to assist in characterizing other oxidation products, some of which could potentially form cross-linked structures. Some samples were run using an LTQ FT Ultra as described [38]. Weak or poor-quality data were obtained from less abundant samples, but these spectra were consistent with the formation of the unusual adducts. Nonspecified enzyme searches were used to characterize peptide fragments that may have formed independent of enzyme-specific cleavage.

allergic exacerbations were excluded. Subject details are given in Supplementary Table 1. Assessment of S100A8 and S100A9 in sputum Sputum from asthmatic, bronchiectatic, and control subjects was diluted in PBS/0.1% DTT (1/4 v/v) and centrifuged (10 min, 400g) and supernatants were stored at 80 1C. DTT was required to disperse mucus [39]. Protein levels were determined by the Bradford assay (Bio-Rad, Hercules, CA, USA). S100A8 and S100A9 levels in duplicate sputum samples were quantified by ELISA [26]; recombinant standards contained 0.1% DTT. Amounts were expressed as ratios with respect to total protein (expressed as units). Sputum proteins (4 mg for silver stain or 1.5 mg for Western blot) were resolved by 10% SDS–PAGE. As oxS100A8 ELISAs could not be performed because DTT interfered with antibody binding on solid phase, semiquantitative analysis of anti-oxS100A8 reactivity on Western blots was performed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Levels of S100A8 (10 kDa) or oxS100A8 monomer (10 kDa) were quantitated as arbitrary net intensity units, compared to intensities of 250, 100, 50, and 25 ng recS100A8 or 500, 250, 100, and 50 ng recoxS100A8 monomer used to construct standard curves for more quantitative measurement. Standards (250 ng) were routinely separated on the same gel. Reactions were detected by chemiluminescence as described [26]. Density of standard bands was given a value of 1; densities of samples were expressed as a ratio; bands of density greater than the standard were 41, and those less than the standard, o1.

Immunohistochemistry Preparation and analysis of S100A8 and S100A9 in sputum Apparently normal portions of lungs from patients undergoing surgery for lung carcinoma (Central Area Health Service, Sydney, Australia) were obtained with consent under ethical guidelines. Asthmatic lung specimens were from Professor Judy Black (University of Sydney, Australia), obtained with consent. Formalin-fixed, paraffin-embedded tissues (2-mm serial sections) or cytospin preparations of normal human neutrophils were processed and immunostained as detailed [26]. Polyclonal rabbit anti-S100A8, anti-oxS100A8, and anti-S100A9 (as described [35]) and normal rabbit IgG (Dako) (all 5 mg/ml) were used after blocking with normal goat serum (Sigma; 1:5 v/v in PBS). AntioxS100A8 IgG, raised against HOCl-oxS100A8 (4:1 molar ratio HOCl:S100A8), recognizes the oxidized monomer and higher complexes more strongly than does anti-S100A8, but the epitope is uncharacterized [26]. Secondary antibody was biotinylated goat anti-rabbit IgG (Chemicon, VIC, Australia), identified by red alkaline phosphatase staining (Vector Laboratories, Burlingame, CA, USA); counterstaining was with Mayer’s hematoxylin (Sigma). Patient recruitment and sample preparation Stable asthmatic, bronchiectatic, and control subjects were recruited from the Respiratory Medicine Department, Prince of Wales Hospital Sydney, under protocols approved by the Human Research Ethics Committee of the South Eastern Sydney Area Health Service. Spirometry was performed on each subject to assess airway function, according to American Thoracic Society/ European Respiratory Society guidelines, and confirmed that asthmatic and bronchiectatic patients had reduced lung function compared to control subjects. Using a Micro Direct Microlab spirometer (Micro Direct, Lewiston, ME, USA) the best value of three maneuvers was chosen for FVC, FEV1, and FEV1/FVC. Subjects were assessed for smoking history, age, previous/existing medical conditions, current medications, and atopy; those with

Proteins in sputum (600 mg; n ¼ 3 control, n ¼5 asthmatic) were eluted from C8-RP-HPLC in 33–59% acetonitrile/0.1% TFA. For practical reasons, only those samples with a sufficient volume of sputum were chosen for analysis as repeated sample collections were not possible. Control samples eluting as two major peaks at 12–15 and 15–18 min were collected and lyophilized, suspended in PBS (10 ml) and separated by SDS–PAGE under reducing conditions (0.1% DTT), and then Western blotted with anti-S100A8, anti-oxS100A8, or anti-S100A9 IgG. The peak eluting between 12 and 15 min contained predominantly S100A8, oxS100A8, and minor amounts of S100A9. S100A9 was the major protein in the peak at 15–18 min. Both peaks were further resolved using C4-RP-HPLC (25–70% acetonitrile/0.1% TFA). Western blotting of 500 ml lyophilized C4-RP-HPLC eluate (15–18 and 18–21 min) confirmed the presence/absence of oxS100A8, S100A8, and S100A9. Proteins in the lyophilized peaks (5 mg) were suspended in 20 mm NH4HCO3 (100 ml) and digested (4 h at 37 1C) with AspN. Proteins in asthmatic sputum eluted from C8-RP-HPLC in a major peak at 12.3–13.8 min and contained abundant S100A8, oxS100A8, and S100A9. This was further resolved by C4-RP-HPLC; S100A8 and S100A9 eluted in peaks at 14.3 and 15.3 min, confirmed by Western blotting of lyophilized eluates. Because of the varying strengths of the antibodies, and to minimize cross-reactivity, membranes were probed sequentially for anti-oxS100A8, anti-S100A9, and then anti-S100A8 IgG. After reaction with one antibody, membranes were stripped (0.2 M glycine, 0.05% Tween 20, pH 2.5) for 40 min at 80 1C, washed with TBS for 10 min, and then probed. Proteins in the lyophilized peaks (5 mg) were suspended in 20 mm NH4HCO3 (100 ml) and digested (50 ng, 4 h at 37 1C) with AspN or trypsin. Protein mass measurements were acquired as described [23].

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

173

Oxidation of preformed S100A8/S100A9 heterocomplex To investigate effects of oxidation on heterocomplex formation, equimolar amounts (0.4 mM) of rec-hS100A8, S100A9, oxS100A8, or oxS100A9 were suspended in PBS (with 1 mm CaCl2) and various combinations incubated at 37 1C for 30 min. Mixtures were separated by 10% SDS–PAGE under nonreducing conditions, and then silver-stained. To assess the effects of oxidation on the native preformed heterocomplex, rec-hS100A8 and S100A9 (0.4 mM) suspended in Hanks’ balanced salt solution with Ca2 þ and Mg2 þ (GIBCO Invitrogen Corp., Grand Island, NY, USA) were incubated at 37 1C for 30 min and then oxidized with a 1:1 molar ratio of HOCl for 15 min, on ice. The reaction was quenched with 10-fold molar excess L-Met and samples were separated by 10% SDS–PAGE under nonreducing conditions and silver-stained.

Statistical analysis Comparisons of ELISA results (specific levels) between groups and densitometry were analyzed using a one-way ANOVA in conjunction with Bonferroni’s multiple comparison test, which allowed for comparisons between all the groups. Values are reported as means 7 SEM; statistical significance was set at Po0.05.

Results Characterization of oxidation products of S100A8 HOCl generates novel inter- and intramolecular sulfinamide bonds in rec-mS100A8 [25]. Rec-hS100A8 oxidized with 0.5- to 20-fold molar excess HOCl contained DTT-resistant dimers and trimers with as little as 0.5-fold molar excess; multimeric complexes were more evident with higher concentrations (1- to 10-fold) [26]. The susceptibility of rec-hS100A8 to oxidation with HOBr or HOCl was compared. Untreated S100A8 contained predominantly monomers and dimer (Fig. 1A). After DTT reduction, untreated S100A8 migrated predominantly as monomers, with only minor amounts of dimer, indicating reducible disulfide bonds (Fig. 1B). Oxidation with a 1:1 molar ratio of HOCl generated dimers and higher-order products, including trimers, tetramers, pentamers, and hexamers (Fig. 1A). HOBr generated fewer products; the oxidized tetramer was obvious in nonreducing gels (Fig. 1A). Several products were stable after DTT reduction including the HOCl-oxidized tetramer (Fig. 1B), whereas only dimers and trimers were obvious in HOBr-oxidized preparations (Fig. 1B). Thus HOCl was a more effective oxidant of rec-hS100A8. No differences were seen in any preparation oxidized over 60 min, indicating that oxidation was complete within 15 min on ice. Because the oxidized monomer was potentially the most prominent form in asthmatic sputum, conditions were established to generate and characterize this product from rechS100A8. The exquisite sensitivity of S100A8 with a 1:1 molar ratio of HOCl was indicated by changes in elution times from C4RP-HPLC; oxS100A8 eluted as two broad peaks at elution maxima 14.4 and 14.8 min (Fig. 2A, profile 2), whereas S100A8 (Fig. 2A, profile 1) eluted as a single peak after 13.5 min. Although elution times were consistent, quantities of oxidized monomer and dimer varied somewhat from experiment to experiment. SDS–PAGE separation indicated that monomeric oxS100A8 eluted in peak 1 (14.4 min) and monomeric and dimeric oxS100A8 (resistant to reduction by DTT) eluted in peak 2 (elution time 14.8 min;

Fig. 1. Hypohalous acids generate multimeric oxidation products in rec-hS100A8. (A) 2.5 mg S100A8 was oxidized with HOBr or HOCl for 15, 30, or 60 min; separated under nonreducing conditions by 10% SDS–PAGE; and silver-stained. Untreated S100A8 migrated predominantly as monomer and dimer. HOBr-oxidized S100A8 migrated as four distinct components, indicative of monomer, dimer, trimer, and tetramer. HOCl-oxidized S100A8 was similar, with two additional high-mass components (pentamer and hexamer). (B) 2.5 mg S100A8 was oxidized with HOBr or HOCl for 15, 30, and 60 min and separated under reducing conditions (100 mM DTT). Untreated S100A8 migrated predominantly as monomer; some DTT-resistant dimer was seen. HOBr-oxidized S100A8 was predominantly monomer and dimer. HOCl produced a similar pattern although oxidized tetramer was obvious. No apparent differences in intensity or banding pattern were obvious over the three time points in (A) or (B).

Fig. 2B), but monomeric S100A8 in this fraction could not be separated further. Oxidation of the Cys42-Ala42S100A8 mutant with a 1:1 molar ratio of HOCl did not alter its elution time from C4-RP-HPLC (not shown) and no dimer was obvious after SDS–PAGE and silver staining (Fig. 2C), indicating that dimerization of native S100A8 may be dependent on Cys42 but not disulfide in nature. A molar ratio of 2:1 HOCl:S100A8 (Fig. 2A, profile 3) generated aggregates that could not be resolved by C4-RP-HPLC. Consistent with data shown in Fig. 1, HOBr appeared less reactive than HOCl; 1:1 molar ratios generated principally monomeric oxS100A8 with elution maximum at 15.1 min (Fig. 2D, profile 2), later than native S100A8 (13.5 min; Fig. 2D, profile 1), and this migrated as a single 10-kDa band in SDS–PAGE (Fig. 2E). The small peak with elution maximum at 15.6 min was not characterized because of insufficient yield (Fig. 2D, profile 2); 10-fold more HOBr promoted aggregation (Fig. 2D, profile 3). Novel oxathiazolidine dioxide modifications in oxS100A8 Protein fractions were digested with trypsin or AspN to more precisely identify the sites and types of modifications present in S100A8 after HOCl treatment (Fig. 2A, peak 1, profile 2; for deconvoluted masses and ESI–MS see Supplementary Table 2 and Supplementary Figs. 1A and 1B). AspN was used because this had facilitated characterization of the sulfinamide bond in mS100A8 [25]. Monomeric ox-hS100A8 (C4-RP-HPLC, Fig. 2A, profile 2, elution time 14.1–14.6 min) digested with AspN or

174

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

Fig. 2. Hypohalous acids generate structural changes in rec-hS100A8 that alter RP-HPLC retention profiles. (A) 100 mg S100A8 eluted from C4-RP-HPLC at 13.5 min (profile 1). Profile 2 shows separation of S100A8 oxidized with an equimolar amount of HOCl. Two peaks eluted at 14.4 and 14.8 min. Profile 3 highlights the susceptibility of S100A8 to form aggregates using a 2:1 molar ratio of HOCl:S100A8 that could not be separated by C4 RP-HPLC. (B) SDS–PAGE separation (under reducing conditions) and silver staining of oxidized peaks confirmed that mostly monomeric S100A8 (Ox A81) eluted at 14.4 min. Dimeric S100A8 (Ox A82) also contained monomeric protein; this could not be separated further by C4-RP-HPLC. Unmodified S100A8 separated by SDS–PAGE was silver stained for comparative analysis. (C) Nonreducing SDS–PAGE indicated that a 1:1 ratio of HOCl:Cys42-Ala42S100A8 (oxAla A8) did not generate dimers, suggesting that Cys42 was involved in dimer formation after exposure to HOCl. (D) Profile 1 shows 100 mg S100A8 eluting from C4-RP-HPLC at 13.5 min. S100A8 oxidized with HOBr:S100A8 (equimolar ratio) had a major peak with retention time 15.1 min and a second minor peak at 15.6 min (profile 2). The minor peak eluting at 15.6 min contained insufficient protein and could not be characterized. Profile 3 shows that 10-fold more HOBr caused aggregation. (E) Silver staining of PAGE gel (under reducing conditions) of HOBr-oxidized peak 1 (profile 2, 15.1 min) indicated mostly monomeric oxS100A8 (Ox A81).

trypsin yielded sufficient amounts of oxidized peptides for more complete analysis. Spectra were searched using Mascot, with mass tolerances of 0.25 for MS and 0.2 for MS/MS. Met(O), Cys42-sulfenic acid (m/z þ16), Cys42-sulfinic acid (m/z þ32), and Cys42-sulfonic acid (m/z þ 48) were specified as variable modifications. Three additional modifications, þO-H2 (m/z þ14) þO2-H2 (m/z þ30), and O3-H2 (m/z þ46), were added to the Mascot search program and included the possibility of neutral losses of SO2 or SO3. Trypsin digestion of monomeric HOCl-oxS100A8 generated peptides with spectra containing the same Cys42-sulfinic acid and -sulfonic acid modifications identified in the oxidized protein (Table 1). No evidence for intermolecular sulfinamide bonds (m/z þ14) was found. On the other hand, many spectra contained the unusual mass addition of m/z þ30 on Cys-containing peptides, originally observed in deconvoluted spectra (Supplementary Figs. 1B). A second novel adduct, of m/z þ 46 Da, not apparent in HOCl-oxidized mS100A8 was also identified. Figs. 3A–F show extracted ion chromatograms and masses of modified forms of LLETECPQYIR in tryptic digests and the time when MS/MS spectra were acquired (see Supplementary Table 3 for amino acid mass comparisons of the major products). Prominent neutral losses of m/z  64 (corresponding to neutral loss of SO2, from peptides with m/z þ30 additions) and m/z 80 (corresponding to neutral loss of SO3, from peptides with m/z þ46 additions) were observed in MS/MS spectra of these ions.

Based on the fragmentation pattern, the novel m/z þ30 was denoted oxathiazolidine oxide and the m/z þ46 addition defined as oxathiazolidine dioxide. MS/MS of oxS100A8 trypsin-derived peptide LLETECPQYIR was performed as shown in the extracted ion chromatograms (Figs. 3A, B, and D). Although a definitive assignment of Cys modifications was not possible, the oxidation products observed most probably correspond to Cys42-sulfinic and Cys42-sulfonic acid modifications. When Pro is present in a peptide sequence, adjacent fragmentations predominate. N-terminal peptide cleavage results in preferential formation of an abundant y-ion on this residue [40], demonstrated by the intense fragment abundance on Pro43 (labeled as ion y5; Figs. 4A and B). The same approach was used to confirm the presence of the novel þ30 and þ46 mass additions, present only in ions containing the Cys residue (Figs. 4C and 4D), and apparent in Figs. 3A, 3C, 3E, and 3F. MS/MS of the modified S100A8 AspN-derived peptide DDLKKLLETECPQYIRKKGA indicated putative Cys42-sulfinic acid and -sulfonic acid modification (Supplementary Figs. 2A and 2B) and the novel þ46 mass addition (Supplementary Figs. 2C and 2D). The latter was also found in the peptide ETECPQYIRKKGA (Supplementary Figs. 3). The novel þ30 mass addition was not obvious in AspN digests, possibly because of its lower abundance. Post-translational modifications observed in dimeric oxS100A8 samples were difficult to identify precisely. There was no indication of sulfinamide bonds, suggesting that sulfinamide-dependent

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

175

Table 1 Summary of major oxidative modifications in rec-oxS100A8 and oxS100A8 from sputum identified by MS. Sample

Mass addition (Da)

Modification

HOCl-oxidized S100A8 (1:1) AspN digest DDLKKLLETECPQYIRKKGA and ETECPQYIRKKGA DDLKKLLETECPQYIRKKGA and ETECPQYIRKKGA DDLKKLLETECPQYIRKKGA and ETECPQYIRKKGA EFLILVIKMGVAAHKKSHE DVWKFEL

þ 32 þ 48 þ 46 þ 16 þ 16

(10,876 (10,892 (10,890 (10,860 (10,860

Da) Da) Da) Da) Da)

Cys42-sulfinic acid Cys42-sulfonic acid Cys42-oxathiazolidine dioxide Met78(O) Trp54(þ16)

þ 32 þ 30 þ 46 þ 48

(10,876 (10,874 (10,890 (10,892

Da) Da) Da) Da)

Cys42-sulfinic acid Cys42-oxathiazolidine oxide Cys42-oxathiazolidine dioxide Cys42-sulfonic acid

Trypsin digest LLETECPQYIR

HOBr-oxidized S100A8 (1:1), AspN digest ETECPQYIRKKGA DVWFKEL

þ 48 (10,892 Da) þ 16 (10,860 Da)

Cys42-sulfonic acid Trp54(þ16)

Control sputum (n ¼ 2) (15–18 min C4-RP-HPLC), AspN digest MLTELEKALNSII MLTELEKALNSIIa ETECPQYIRKKGA and DDLKKLLETECPQYIRKKGA DVWFKEL

þ 16 þ 16 þ 48 þ 16

(10,850 (10,850 (10,882 (10,850

Da) Da) Da) Da)

Met1(O) Met1(O) Cys48-sulfonic acid Trp54(þ16)

þ 48 þ 46 þ 16 þ 16

(10,882 (10,880 (10,850 (10,850

Da) Da) Da) Da)

Cys42-sulfonic acid Cys42-oxathiazolidine dioxide Met1(O) Trp54(þ16)

þ 32 þ 48 þ 30 þ 46 þ 16 þ 16

(10,866 (10,882 (10,864 (10,882 (10,850 (10,850

Da) Da) Da) Da) Da) Da)

Cys42-sulfinic acid Cys42-sulfonic acid Cys42-oxathiazolidine oxide Cys42-oxathiazolidine dioxide Met1(O) Met78(O)

Asthmatic sputum (n ¼ 5) (peak 1, 14.3 min C4-RP-HPLC) AspN digest ETECPQYIRKKGA and DDLKKLLETECa DDLKKLLETECPQYIRKKGA and ETECPQYIRKKGA MLTELEKALNSII DVWFKEL Trypsin digest KLLETECPQYIR LLETECPQYIR MLTELEK and MLTELEKALNSIIa MGVAAHKK

Oxidative modifications identified in peptides derived from rec-oxS100A8 and control and asthmatic sputum are shown. Modifications were identified in peptides derived from recS100A8 oxidized with HOCl or HOBr at the molar ratios indicated. Oxidation products on Cys42 in S100A8 were identified in in-solution digests by peptide mapping and MS/MS. Oxathiazolidine oxide ( þ30 m/z) and dioxide (þ 46 m/z) adducts were obvious, the former was seen only in tryptic peptides. The oxathiazolidine dioxide derivative was plentiful in peptides from all HOCl-oxidized samples. In sputum, oxidation of Met1 to Met1(O) was seen in 3/3 control samples and of Trp54 to Trp54(þ 16) in 2/3 control samples. Cys42-sulfonic acid was present in 2/3 control samples. Masses of modifications are indicated in parentheses, expressed in Daltons (Da). Oxidation products on Cys42 of S100A8 were identified by peptide mapping and MS/MS. Oxathiazolidine oxide (þ 30 m/z) and dioxide (þ 46 m/z) adducts were obvious, the former only in tryptic peptides. The oxathiazolidine dioxide adduct was observed in S100A8-derived peptides separated from asthmatic sputum (n ¼ 3/5; peptides from 1/5 subjects had no compelling evidence for modified Cys-containing S100A8 residues). a Sequence identified in non-enzyme-specified sample. Modifications observed in these peptides were consistent with modifications identified in enzyme-specified searches.

dimers were unlikely, and these samples were not characterized further. An attempt was made to identify other predicted modifications generated by HOCl, using Mascot searches selecting the following modifications and performing variable modification searches: Met(O)/sulfone; dityrosine; chlorotyrosine derivatives; carbonylation of Lys, Pro, Arg, and Thr and (for HOBr) bromination of Tyr, Lys, His, Arg, Asn, and Gln residues; and oxidation of His or Trp (addition of one oxygen, þ16). Modifications were selected based on their presence in oxidized proteins [41–46] and because no prior examination of these products had been performed. Oxidation of Met78 to Met78(O) was obvious in peaks eluting at 14.4 and 14.8 min by C4-RP-HPLC after AspN digest. Trypsin digests identified only one peptide (S100A878–85), suggesting some oxidation of Met78 to Met78-sulfone, an irreversible modification. Despite identification of several Tyr-containing peptides, only one (S100A87–18) indicated chlorination of Tyr16 to 3-chloro-Tyr. These modifications were found only in monomeric samples, suggesting that they did not contribute to dimer formation. Identification of Trp oxidation products in these samples was complicated by a lack of Trp-containing peptides identified in the trypsin digests; there was

no evidence for His oxidation. Several Tyr-containing peptides were identified in AspN digests of monomeric and dimeric oxS100A8. Oxidation of Tyr19 to 3-chlorotyrosine was apparent in one peptide (S100A814–31, Mowse score 20), indicating a weak spectrum. Taken together, the relative abundance of these modifications was much less than the amounts of the oxathiazolidine adducts identified on Cys42, suggesting that even under favorable conditions, oxidative modifications of Tyr residues, or of Met78(O) to Met78-sulfone, are less compared to oxidation of Cys or Met78. No carbonyl adducts were observed. In AspN digests, addition of one oxygen was obvious on the single Trp54 residue. Weak evidence for oxidation of His17/27 was apparent on the peptide DVYHKYSLIKGNFHAVYR. Similar products were observed in the oxidized dimer. Additional Mascot searches were performed using the same data because some nonspecific peptide-bond cleavages were identified. Nonenzyme-specified searches indicated addition of one oxygen to His83 on the peptide GVAAHKKSHEESHKE derived from monomeric and dimeric samples. Because several residues may be susceptible to oxidation in these peptides, modifications require further characterization.

176

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

Fig. 3. Extracted ion chromatograms and masses of modified forms of LLETECPQYIR. (A) Extraction ion chromatograms from nano-LC–MS of [Mþ 2H]2 þ ions of LLETECPQYIR modified by O2-H2 (þ30), O2 ( þ32), O3-H2 (þ46), or O3 (þ 48). Partial TOF–MS of individual precursor ions, (B) m/z þ 32, þ O2; (C) m/z þ 46, þ O3-H2; (D) m/z þ48, þO3; (E) m/z þ 46, þO3-H2; (F) m/z þ 30, þ O2-H2; at retention times indicated by arrows in (A). Two forms of m/z þ 46, O3-H2 were separated (TOF–MS shown in (C) and (E)) and these were attributed to isomeric oxathiazolidine dioxides. Each modified form of S100A8 separated sufficiently well by nano-LC–MS to ensure MS/MS selection of the desired precursor only (selection windows are shown as shaded color boxes). The switch to MS/MS occurred early in the elution profile with a threshold of 10 counts. This also helped to ensure isolation of a single precursor ion m/z.

MS/MS analysis of HOBr-derived oxidation products identified sulfonic acid intermediates in AspN digests of oxS100A8 monomer (peak 1, profile 2, Fig. 2D; for deconvoluted masses and ESI–MS see Supplementary Table 2 and Supplementary Figs. 4). Apart from oxidation of Trp54 to Trp54( þ16), no other modifications, for example to Tyr [47], were apparent. A summary of the major products identified in HOCl- and HOBr-oxS100A8 peptides is given in Table 1. Characterization of HOCl-oxS100A9 S100A9 had a retention time of 14.1 min (Supplementary Figs. 5A, profile 1) from C4-RP-HPLC; two peaks were generated with S100A9:HOCl (1:1 molar ratio; Supplementary Fig. 5A, profile 2), with longer retention times of 14.8 and 15.7 min, respectively. SDS–PAGE separation (Supplementary Fig. 5B) indicated monomeric oxS100A9 and disulfide-linked dimer (reduced by DTT) in peaks 1 and 2. Deconvoluted masses of unmodified (Supplementary Figs. 6A) and monomeric oxS100A9 products (Supplementary Fig. 6B) were determined by ESI–MS; deconvoluted mass additions are summarized in Supplementary Table 4. In-solution AspN digests of monomeric oxS100A9 (elution maximum 14.8 min, Supplementary Figs. 5A, peak 1, profile 2) and MS analysis of non-enzyme-specified searches indicated that oxidation of the single Cys3 residue to Cys3-sulfinic acid or sulfonic acid, and of Met5 to Met5(O), was weak. Met63/81/83/94

were variably converted to Met(O) by HOCl (Supplementary Figs. 7–10), suggesting low abundance; spectra were weak compared to modifications seen in rec-oxS100A8, consistent with previous indications that S100A9 is less susceptible to oxidation by HOCl [26]. Conversion of Met81 to Met81-sulfone by HOCl was apparent in one peptide (S100A971–91; Mowse score 32) but its very low abundance suggests that this was a relatively unfavorable reaction. Oxidation of Trp and His residues was difficult to characterize accurately. Mascot searches showed no compelling evidence for Trp oxidation. MS/MS spectra for His oxidation was also poor; only weak spectra were available for His91/94 in non-enzymespecified searches and His104 in AspN-specified searches (not shown). The major modifications of isolated peptides are summarized in Table 2. S100 expression in asthmatic lung Immunostaining showed more intense oxS100A8 reactivity (Fig. 5, middle) than S100A8 (Fig. 5, left) in asthmatic lung compared to apparently normal lung (bottom). Positive eosinophils, some neutrophils, and macrophages were localized beneath the basement membrane of the airways and within the lung parenchyma. In contrast, immunoreactivity of oxS100A8 in normal human neutrophils (Fig. 5, inset, middle) was negligible, whereas, as expected, reactivity with anti-S100A8 was strong (Fig. 5, inset, left). S100A9 was localized in a similar pattern (Fig. 5, right).

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

177

Fig. 4. Annotated MS/MS of modified S100A8 tryptic peptides of LLETECPQYIR. (A) MS/MS of S100A837–47 most probably containing a m/z þ 32 addition (Cys42-sulfinic acid) at retention time 25.08 (calculated by nano-LC–MS). (B) MS/MS of S100A837–47 containing a probable m/z þ 48 addition (Cys42-sulfonic acid). Sequence-specific ions are labeled and, as expected, an intense fragment ion y5 was observed on Pro43 (highlighted in box). (C) MS/MS of A837–47 containing a m/z þ 30 addition (retention time 25.808 min, as calculated by nano-LC–MS). (D) MS/MS of A837–47 containing a m/z þ 46 addition (retention time 26.74 min, as calculated by nano-LC–MS) and corresponding to the second eluting peak (a very similar MS/MS spectrum was recorded for [Mþ 2H]2 þ ions eluting at 25.67 min, as calculated by nano-LC–MS). Sequence-specific ions are labeled. Intense fragment ions y5 (Pro residue; shown in box) are observed, as expected. Only ions containing Cys show loss of m/z  64 or  80 (shown in box).

178

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

Analysis of S100A8 and S100A9 in sputum S100A8 levels in asthmatic, bronchiectatic, and control sputum were similar (not shown). S100A9 tended to be higher in asthmatic and bronchiectatic sputum, but was not significantly more than in controls (not shown).

Table 2 Summary of major oxidative modifications in HOCl-oxS100A9 and oxS100A9 from sputum identified by MS. Sample

Mass addition (Da)

Modification

HOCl-oxidized S100A9 (1:1) (14.8 min C4-RP-HPLC), AspN digest EKVIEHIM þ 16 (13,268 Da) Met63(O) þ 16 (13,268 Da) Met81(O) EEFIMLMARLTWASH EEFIMLMARLTWASH þ 16 (13,268 Da) Met83(O) EEFIMLMARLTWASHEKMHEG þ 16 (13,268 Da) Met94(O) Control sputum (n ¼ 3) (18–21 min C4-RP-HPLC) þ 16 (13,258 Da) EKVIEHIME

Met63(O)

Asthmatic sputum (n ¼ 3) (15.3 min C4-RP-HPLC), AspN digest þ 16 (13,258 Da) Met63(O) EKVIEHIME and DLQNFLKKENKNEKVIEHIMEa a þ 16 (13,258 Da) Met81(O) EEFIMLMARLT WASHEKMHEGa þ 16 (13,258 Da) Met94(O) S100A9 peptides from rec-oxS100A9 and oxS100A9 enriched from control and asthmatic sputum are shown. Modifications were identified in peptides derived from recS100A9 oxidized with HOCl at the molar ratio indicated. Oxidation products on Met residues in S100A9 were identified in in-solution digests by peptide mapping and MS/MS. Sputum from 3/3 control and 3/3 asthmatic samples contained Met63(O); of these, one asthmatic sample also contained Met81 and Met94 (O).

Because of the complex nature of sputum samples, Western blotting was conducted to identify S100A8 and S100A9. AntiS100A8 IgG reacted weakly with 5/12 control samples and strongly with 9/14 asthmatic samples and 7/10 bronchiectatic samples (Fig. 6A). Strong anti-oxS100A8 IgG reactivity was seen in 12/14 asthmatic, in 7/10 bronchiectatic, and, more weakly, in 4/12 control sputum samples (Fig. 6B). In all samples, only monomeric S100A8 was obvious, with no evidence of DTTresistant complexes. Western blotting of control sputum with anti-S100A9 IgG (Fig. 6C) indicated strong reactivity in 5/12 samples; a component with a mass typical of the S100A8/S100A9 heterodimer was seen in sputum from two subjects. Reactivity with the monomer was obvious in 9/14 asthmatic samples; only one contained obvious homodimer; 3/14 contained S100A8/S100A9 heterocomplex. Anti-S100A9 IgG reactivity was strong in sputum from 7/10 bronchiectatic subjects; 3/10 contained a component typical of S100A8/S100A9 heterocomplex (24 kDa) (Fig. 6C). OxS100A8 in sputum could not be quantitated by ELISA as samples contained DTT, which interfered with antibody binding in solid phase. Reactivity of Western blots using anti-oxS100A8 and S100A8 IgG were semiquantitated by densitometry. Some 2.3-fold more S100A8 was found in asthmatic sputum and  2.4 times more S100A8 was present in bronchiectatic sputum relative to controls (Fig. 6D). In marked contrast, there was significantly more oxS100A8 (14.7 times) in asthmatic sputum compared to control sputum (Po0.05; Fig. 6E). Because of the relative abundance of oxidized S100A8 in asthmatic subjects, this subgroup was chosen as the potential source for analysis. Novel oxathiazolidine dioxide modifications in S100A8

a

Sequence identified in non-enzyme-specified sample. Modifications observed in these peptides were consistent with those identified in enzymespecified searches.

We were limited to a single sputum sample from each donor. Larger volumes from more donors may have facilitated better MS

Fig. 5. S100A8, oxS100A8, and S100A9 in asthmatic lung and sputum. Positive immunoreactivity (indicated by red staining) of S100A8 (left) and oxS100A8 (middle) was seen in the airway lining in serial sections from an asthmatic lung, representative of five different specimens. Neutrophils, eosinophils, and some macrophages were immunoreactive, mostly accumulated beneath the basement membrane of bronchioles, and spread within the lung parenchyma; the few immunoreactive cells seen in sections from nonasthmatic lung (bottom) were leukocytes within blood vessels. Human neutrophils reacted very weakly to anti-oxS100A8 (inset, middle), whereas S100A8 reactivity (inset, left) was strong. S100A9 (right) was expressed in a similar pattern. IgG control of asthmatic lung is shown in the inset, right.

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

179

Fig. 6. S100A8 and S100A9 in sputum. (A) Western blotting of sputum (1.5 mg) from asthmatic subjects indicated anti-S100A8 immunoreactivity with a 10-kDa component in 9/14 samples; Western blotting showed that 7/10 sputum samples from bronchiectatic samples were reactive, whereas immunoreactivity was observed in only 5/12 control subjects. (B) Positive reactivity with anti-oxS100A8 was obvious in 12/14 asthmatic samples; samples from 4/12 control subjects reacted relatively strongly with anti-oxS100A8 IgG. Western blotting of sputum from bronchiectatic subjects with anti-oxS100A8 indicated immunoreactivity with a 10-kDa component in 7/10 samples. (C) Positive reactivity with anti-S100A9 IgG was apparent in 5/12 control samples. S100A8/A9 heterodimer was observed in two samples. Anti-S100A9 IgG reactivity was seen in 9/14 asthmatic samples; 3/14 contained the heterocomplex. Anti-S100A9 IgG reactivity was strong in sputum from 7/10 bronchiectatic subjects; 3/10 contained heterocomplex (24 kDa). (D) Intensities of components migrating at 10 kDa were quantified relative to a 250 ng protein standard of S100A8 or oxS100A8. Asthmatic and bronchiectatic sputum contained more S100A8 compared to control sputum, but levels were not significantly different (results are means 7 SEM). (E) Intensities of bands migrating at 10 kDa; nP o 0.05 between control and asthmatic sputum levels of oxS100A8. Results are means 7 SEM. Control, n ¼ 12; asthma, n ¼ 14; bronchiectasis, n ¼ 10.

analysis and correlation analysis between oxidation products within and between each subgroup. However, our aim was to identify the particular modifications generated in vitro, to

validate scavenging activity in a human inflammatory condition. Sputum from three control subjects (samples 2, 7, 12; see Figs. 6A and 6B) (sample 7 was from an atopic donor) was separated by

180

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

RP-HPLC. Sputum eluted from C8-RP-HPLC as three broad peaks (12–15 min aliquot comprised two peaks, peak 1 at 13.8 min and peak 2 at 14.5 min, which were pooled; and another peak eluted at 15–18 min; Fig. 7A). Western blotting confirmed that oxS100A8 and S100A8 were prominent in the 12–15 min eluate (Fig. 7B, left and middle). DTT-resistant S100A8 complexes were obvious in both peaks; low levels of dimer (20 kDa) were in the 12–15 min aliquot and a complex migrating at 34 kDa was in the 15–18 min aliquot. Although the latter could indicate S100A82/ S100A9, this did not react with anti-S100A9 IgG, which reacts with S100A8/S100A9 heterocomplexes, whereas anti-S100A8 does not, and so may represent S100A8 complexed with another binding partner in vivo (Fig. 7B, middle). Monomeric S100A9 eluted predominantly at 15–18 min from C8-RP-HPLC; low amounts of S100A92 (28 kDa; Fig. 7B, right) were present. Pooled peaks 1 and 2 from C8-RP-HPLC (Fig. 7A, 12–15 min) were further separated by C4-RP-HPLC, which resolved a major peak (Fig. 7C, profile 1, peak 1, 15–18 min) and a minor peak (Fig. 7C, profile 1, peak 2, 18–21 min). Peak 3 from C8-RP-HPLC (Fig. 7A, 15–18 min pool) eluted as two minor peaks at 16.5 and 18–18.5 min (Fig. 7C, profile 2). Pools were made of proteins eluting between

15–18 and 18–21 min. Western blotting of C4-RP-HPLC eluates from profile 1 (Fig. 7C) confirmed monomeric oxS100A8 and S100A8 and possibly dimeric (20 kDa), trimeric (30 kDa), and tetrameric (40 kDa) complexes (Fig. 7D, left and middle, 15–18 min). Monomeric S100A9 (Fig. 7D, right, 18–21 min) was predominantly in peak 2. Western blotting of C4-RP-HPLC eluate from profile 2 (Fig. 7C) indicated low amounts of oxS100A8 monomer and dimeric complexes (Fig. 7E, left, 15–18 min) and possibly dimeric, trimeric, and tetrameric complexes of native S100A8 in peaks 1 and 2 (Fig. 7E, middle, 15–18 min). Monomeric S100A9 was enriched from C4-RP-HPLC and appeared more abundant than levels eluting in profile 1 (Figs. 7C and D, right); minor reactivity with a component of 28 kDa, the mass of the S100A9 dimer, was also apparent (Fig. 7E, right, 18–21 min). Asthmatic sputum from five donors was analyzed (samples 7, 8, 11, 13, and 14; see Figs. 6A and 6B); samples 7 and 14 were from patients with atopy. These eluted as a broad peak (elution maximum 12.3 min, Fig. 8A) from C8-RP-HPLC and contained a 10-kDa anti-S100A8-reactive component and some homodimer (20 kDa), confirmed by Western blotting (not shown). C4-RP-HPLC resolved three peaks (elution maxima 14.3, 15.3, and 15.6 min; Fig. 8B).

Fig. 7. S100A8 and S100A9 in control sputum. (A) Control sputum (600 mg) eluted from C8-RP-HPLC as three peaks at 13.8, 14.5 (pooled), and 15–18 min. (B) Western blotting of peaks 1 and 2 (12–15 min; 500 ml lyophilized from C8-RP-HPLC) indicated immunoreactivity with anti-oxS100A8, anti-S100A8, and anti-S100A9 antibodies; peak 3 (15–18 min) indicated monomeric S100A9 and slight reactivity with the dimer (  28 kDa). A complex migrating at 34 kDa was identified by anti-S100A8. Data presented are representative of three samples of enriched control sputum. (C) Pooled peaks 1 and 2 and the single peak 3 from C8-RP-HPLC were purified by C4-RP-HPLC, indicated as profiles 1 and 2, respectively. Briefly, profile 1 separated into a major peak (elution time 15–18 min) and a minor component between 18 and 21 min. Profile 2 eluted as small peaks at 15–18 min and 18–21 min. (D) Western blotting of profile 1 from (C) (500 ml lyophilized from C4-RP-HPLC) showed weak reactivity with antioxS100A8; anti-S100A8 reacted with higher-order complexes. Monomeric S100A9 was identified in fractions eluting between 18 and 21 min. (E) Western blotting of profile 2 ((C), 15–18 min) indicated weak reactivity with anti-oxS100A8 and anti-S100A8 IgG; the latter reacted weakly with complexes of the same mass as in (D). Monomeric S100A9 principally eluted between 18 and 21 min. Western blots included 500 ng rec-oxS100A8, S100A8, or S100A9 as positive controls. Data presented are representative of three samples of enriched control sputum.

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

181

Fig. 8. Purification of S100A8 and S100A9 from asthmatic sputum. (A) Asthmatic sputum (600 mg) eluted from C8-RP-HPLC as a broad peak between 11.7 and 12.6 min. (B) Peak 1 from C8-RP-HPLC was purified by C4-RP-HPLC. Peaks eluting at 14.3 (peak 1), 15.3 (peak 2), and 15.6 (peak 3) min were collected. (C) Peaks 1, 2, and 3 were immunoblotted (500 ml, lyophilized to 10 ml) with anti-S100A8 or anti-oxS100A8 IgG under reducing conditions. Peak 1 contained high amounts of 10- and 20 kDa components corresponding to S100A8 monomer and dimer, respectively (native and oxidized). Components in peak 2 had little reactivity with the antibodies; no reactivity was seen in peak 3. (D) The single C8-RP-HPLC peak and the three peaks from C4-RP-HPLC were Western blotted with anti-S100A9 IgG. The C8-RP-HPLC eluate contained components migrating at  14, 24, 28, 38, and 42 kDa and minor amounts of large complexes. Purification by C4-RP-HPLC indicated lower amounts of most components, principally in peak 2, which contained 14-, 28-, and 42-kDa components, corresponding to S100A9, and higher-order complexes. Western blots included 500 ng of each S100 protein as positive control. Data presented are representative of enriched asthmatic sputum from five donors.

Peak 1 (14.3 min) contained strong anti-S100A8 and antioxS100A8 IgG-reactive components (Fig. 8C) migrating at 10 and 20 kDa; there was little S100A8 in peaks 2 and 3. Peak 1 from C8-RP-HPLC contained components with relatively strong reactivity with anti-S100A9, migrating at 14, 24, and 28 kDa, possibly indicating S100A9 monomer and dimer and the S100A8/S100A9 heterodimer. Minor DTT-resistant components separated at 38, 42, and 48 kDa and higher masses. Peaks 1 and 3 from C4-RP-HPLC contained little S100A9; peak 2 was predominantly S100A9 monomer (14 kDa) and dimer (28 kDa) and higher-order complexes (Fig. 8D). S100 proteins in asthmatic sputum separated by C8-RPHPLC and C4-RP-HPLC eluted earlier, with more clearly defined peaks corresponding to S100A8 and S100A9 than seen with control sputum. The delayed elution times for S100A8 and S100A9 in control sputum may be due to their low abundance (15-fold less for oxS100A8) and higher amounts of other proteins than in asthmatic sputum. DTT-resistant complexes were evident in Western blots using anti-oxS100A8 and anti-S100A8; the dimer was most obvious, but only after enrichment by C4-RP-HPLC, and as minor products, these were not characterized further. Because the DTT used to disperse the mucus in sputum could have reduced disulfidelinked dimers, assessment of relative abundance of S100A8 monomers/covalent dimers in sputum was not possible. Control sputum fractions separated by C4-RP-HPLC (5 mg) were analyzed by MS after AspN digestion. Peptides from profile 1, peak

1 (pool 15–18 min; Fig. 7C), indicated oxidation of Met1 to Met1(O) in samples from three of three control subjects; there was weak evidence for Cys42-sulfonic acid in two of three samples; Trp54(þ16) was obvious in two of three samples. Peptides of products separated from asthmatic sputum (peak 1, C4-RP-HPLC, 5 mg; Fig. 8B) contained sulfonic acid intermediates in only one of four AspN-specified Mascot searches. Additional Mascot searches were performed using the same data because some nonspecific peptide bond cleavage was observed previously using similar biological samples (M. Raftery, unpublished observations). Nonenzyme-specified searches identified sulfonic acid in peptides from one additional asthmatic subject, and a trypsin-specified digest identified sulfonic acid in a third (not shown); no sulfinic acid intermediates were obvious in AspN digests, but a single peptide was identified in tryptic digests; three of five peptides contained Met1(O). Met78(O) was identified only in the trypsin-digested sample; two of five contained Trp54(þ16), only evident in AspNspecified digests (see Table 1). The peptide DDLKKLLETECPQYIRKKGA contained oxathiazolidine dioxide (m/z þ 46 addition) in one sample digested with AspN (Fig. 9), and it was in the peptide ETECPQYIRKKGA in another sample. The tryptic peptide LLETECPQYIR contained oxathiazolidine oxide (Fig. 10) and dioxide (m/z þ30 and þ46 addition, respectively) in the single sample digested with trypsin. MS/MS spectra were similar to those generated for digests of recombinant oxS100A8, although associated Mascot scores for

182

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

Fig. 9. (A) Annotated MS/MS of human asthmatic sputum samples eluting at 14.3 min by C4-RP-HPLC digested with AspN. S100A832–51 sequence-specific ions (either as y or b ions) are labeled. Only fragment ions y10 and b11 (Cys42) showed loss of m/z  80, with concomitant gain of m/z þ 46 (shown in box), confirming oxidation of Cys42 to Cys42-oxathiazolidine dioxide. (B) Masses of individual amino acids are listed. Amino acids in bold italics are those contributing to Mowse scores of isolated peptides.

peptides were lower, probably because of the sample complexity and low amounts of S100A8 in the samples. Table 1 summarizes oxidation products identified in S100A8 in enzyme- and non-enzyme-specified searches, in peptides generated from sputum products enriched from asthmatic and control donors. Attempts to minimize irrelevant proteins by RP-HPLC enrichment assisted characterization, but contaminating peptides caused high backgrounds in peptide searches, making identification of residues containing novel oxidative modifications more challenging.

identified in non-enzyme-specified searches, although samples had high background interference, complicating analysis. The single Cys3 residue of S100A9 was not identified in control subjects or from peak 2 derived from asthmatic sputum. There was limited evidence suggesting oxidation of His91/95 in the peptide WASHEKMHEG in one of three asthmatic samples in non-enzyme-specified searches. Analysis was complicated because of the close proximity of these residues to Met94, and more precise identification of modified residues is required. Effects of HOCl oxidation on S100A8/A9 heterocomplex formation

Met residues are oxidized in S100A9 in vivo A summary of HOCl-generated modifications found in S100A9 in vitro and in vivo is presented in Table 2. S100A9 enriched from three of three control samples (profile 2, peak 2, 18–21 min, C4RP-HPLC, Fig. 7C, 5 mg) contained only Met63(O) (Supplementary Fig. 11). S100A9 from one of three asthmatic samples contained Met63(O) on the peptide EKVIEHIME (Supplementary Fig. 12); two of three contained Met63(O) on the peptide DLQNFLKKENKNEKVIEHIME (not shown). Of these, one also contained Met81(O) (Supplementary Fig. 13) and Met94(O) (Supplementary Fig. 14)

When native recS100 proteins were separated by SDS–PAGE, unmodified S100A8 and S100A9 migrated as monomers, dimers, and trimers (Supplementary Fig. 15A). The unmodified S100A8/A9 complex separated into components with masses equivalent to S100A8 monomer and dimer and S100A9 monomer and dimer; the heterocomplex migrated at 24 kDa. After HOCl oxidation, oxS100A8 comprised oxidized monomer and dimer; oxS100A9 contained oxidized monomer, dimer, and trimer (42 kDa). When oxS100A8 or oxS100A9 was combined with its native counterpart (S100A9 or S100A8, respectively), the same complexes were

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

183

Fig. 10. (A) Annotated MS/MS of human asthmatic sputum samples eluting at 14.3 min by C4-RP-HPLC digested with trypsin. S100A837–47 sequence-specific ions (either as y or b ions) are labeled. Only fragment ions b6 and y6 (Cys42) showed loss of m/z  64 with concomitant gain of m/z þ 30 (y6 ion shown in box), confirming oxidation of Cys42 to Cys42-oxathiazolidine oxide. (B) Masses of individual amino acids are listed. Amino acids in bold italics are those contributing to Mowse scores of isolated peptides.

evident, although heterodimers were less obvious. However, when oxS100A8 and oxS100A9 were combined, no heterodimer was obvious. On the other hand, when the preformed S100A8/A9 heterocomplex was oxidized with a 1:1 molar ratio of HOCl, this complex was still resolved (Supplementary Fig. 15B) and no differences in banding patterns from those observed with the untreated control were obvious, suggesting that the preformed heterocomplex may be relatively resistant to HOCl oxidation (Supplementary Fig. 15B).

Discussion Oxidative stress contributes to progression and exacerbation of inflammation, and excessive ROS/RNS reduce lung function [33]. There is increasing evidence that S100A8 and S100A9 may play protective roles, principally through oxidant scavenging [8,21,22,25,26]. S100A8 suppressed MC activation by scavenging ROS required for signaling and reduced symptoms of acute murine asthma when applied intranasally at antigen challenge [21]. Here we confirm that S100A8, and possibly S100A9, may contribute to redox homeostasis in human asthma. Equimolar ratios of S100A8:HOCl generated novel oxathiazolidine adducts on the single Cys residue of rec-hS100A8, a unique product. Next, we used sputum from asthmatic patients as a source of oxidized S100A8 because an antibody that predominantly recognized oxS100A8 indicated strong reactivity in lung and sputum. This is the first report of oxidized S100A8 and S100A9 in a human inflammatory condition, supporting the notion of their oxidant scavenging capacity in vivo. First we characterized oxidation products of the recombinant S100s by hypohalous acids. HOBr was less efficient than HOCl,

although both generated complexes characteristic of oxidized protein [26]. Molar ratios of S100A8:HOCl of 1:2 caused aggregation, indicating exquisite sensitivity, compared to the high levels required to oxidize low-density lipoprotein [48]. The major product in asthmatic sputum was S100A8 monomer, but our initial investigations indicated that this was post-translationally modified. In this study we focused on the structural changes generated in this adduct. A 1:1 molar ratio of HOCl:S100A8 increased its elution time from C4-RP-HPLC, suggesting increased hydrophobicity. Compared to the tight elution profile of recS100A8 (13.5 min, Fig. 2A, profile 1), a more complex profile comprising two peaks that contained monomer and dimer was generated (Fig. 2B). Treatment of the Cys42–Ala42 S100A8 mutant with equimolar HOCl did not promote complex formation or alter its elution profile, suggesting that Cys42 modification contributed to the structural changes seen with the native protein, although modifications in the mutant require validation. Deconvoluted mass analysis of HOCl:oxS100A8 monomer revealed several mass additions, the most abundant probably corresponding to Cys42-sulfinic acid, Cys42-sulfonic acid, and unusual mass additions of m/z þ30 and þ46, indicating as yet unidentified Cys-oxathiazolidine oxide and dioxide oxidation products and oxidation of Met78 to Met78(O). Deconvoluted mass analysis identifies mass increases, and specific modifications to Cys42 and Met78 were confirmed by MS/MS analysis of peptide digests. Cys42 in S100A8 was a critical target of HOCl; Met78 and Trp54 were also oxidized. Oxidation of His17/27/83 was less clearly defined. Kinetic analysis defined Cys and Met residues as the most susceptible to HOCl oxidation, although His residues can also be oxidized [49]. His oxidation can also generate 2-oxo-His, a carbonyl product that may contribute to protein aggregation (reviewed in [50]), and as we found no evidence for intermolecular

184

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

Fig. 11. The likely products of S100 oxidation by hypohalous acids are summarized. The Cys42 adduct, oxathiazolidine dioxide (þ46 m/z mass addition), may have formed by oxidation of Cys42 to sulfenic acid and then further oxidation to sulfinic or sulfonic acids or to the oxathiazolidine oxide ( þ30 m/z mass addition). Upon further oxidation, this forms a dioxide adduct with a þ46 m/z mass addition caused by the neutral loss of SO2 and SO3 to form cyclized products between oxidized Cys42 and nitrogen within the peptide backbone.

sulfinamide cross-linking [25], this may contribute to aggregation of S100A8 seen with higher molar ratios of HOCl. Oxidation by HOBr is largely substrate-dependent [47,51] and Cys and Met residues are generally 10-fold less reactive than with HOCl [51] and Trp is also a likely target [47]. HOBr oxidized Trp54 but Cys-oxathiazolidine adducts were not identified in S100A8. However, studies with other oxidants, such as HOSCN, are required before suggesting that this product is unique to HOCl oxidation. Here we propose that HOCl promotes alternate oxidation of sulfenic acid intermediates in S100A8. These may form between cyclized oxidized Cys42 and nitrogen within the peptide backbone to generate oxathiazolidine dioxide (m/z þ46 Da addition), which was abundant. The less abundant product, oxathiazolidine oxide (m/z þ30 addition), only obvious in tryptic digests, may be a precursor of the final oxathiazolidine dioxide (m/z þ46 Da addition). A scheme detailing putative modifications in S100A8 is presented in Fig. 11. The same modifications were observed in peptides derived from dimeric oxS100A8. Unlike sulfenic acid and sulfhydryl derivatives that can be reduced by thiol reductases and have high reducing potential, sulfinic and sulfonic acid modifications are irreversible [33] and the oxathiazolidine dioxide adduct in S100A8 is also likely to be an irreversible modification. The presence of novel oxathiazolidine adducts in hS100A8 and the apparent absence of sulfinamide modifications characteristic of the HOCl-oxidized mS100A8 highlight how structural differences affect the nature of the modifications. Oxidation of mS100A8 with a 5:1 molar ratio of HOCl generated inter- and intramolecular sulfinamide bonds between Cys and Lys residues (mass addition, m/z þ14) [25]. Comparative analysis of murine and human S100A8 shows that apart from the conserved amino acids TECPQ flanking Cys42, these have low amino acid sequence identity in this region. This encompasses the hinge domain that contributes to functional diversity in S100 proteins, and we reported functional differences between murine and human S100A8 [52]. In particular, substitution of Gly50 in hS100A8 to Asn50 in mS100A8 may contribute to differences in sensitivity. Structural analyses of S100A8 homodimers [53] and S100A8/A9 heterodimers [7] indicate that Gly50 adopts j–c angles that may

reduce backbone clashes likely to occur with bulkier residues such as Asn50 (Supplementary Figs. 16A and 16B). Gly50 resides within the functional hinge domain, and this region may be more flexible in the human protein, enabling Cys42, which is buried in the core of the protein, to become transiently solvent-exposed and thus more prone to oxidation, whereas Asn50 in mS100A8 may not afford the same flexibility. HOCl-treated recS100A9 separated into two peaks from C4-RPHPLC, eluting later than native S100A9 and monomeric oxS100A9, and the disulfide-linked dimer was obvious. Deconvoluted masses of monomeric oxS100A9 indicated possible modification of Cys3, Met residues, and possibly His. Evidence for Cys3 and Met5 modifications in AspN digests of rec-oxS100A9 was weak; Met63/81/83/94 were all oxidized by HOCl to Met(O). Met-sulfone, an irreversible and relatively rare modification, was identified in S100A9 by ESI–MS from PMA-treated neutrophils [23]. We found Met81-sulfone, but this was not abundant. Precise identification of Trp and His residues was difficult and requires further characterization. His oxidation may be important because the His91-X-X-XHis95 motif in S100A9 is implicated in Zn2 þ binding [54], and their oxidation may modify this. S100A8 and S100A9 were expressed in human asthmatic lung and in infiltrating granulocytes; reactivity with anti-oxS100A8 was particularly strong, whereas this reacted only weakly with native S100A8 in neutrophils (Fig. 5, middle, inset). Importantly, oxS100A8, but not S100A8 or S100A9, was significantly elevated in asthmatic sputum compared to controls. Approximately 2  10  7 mol HOCl is generated by 106 activated neutrophils over 2 h [55], although other oxidizing species are also likely to be present in high amounts. In patients with CF, calprotectin reaches  2.6  10  8 mol [56], but free S100A8 levels in asthmatic lung have not been quantitated. However, oxS100A8 was plentiful, and a likely target, particularly in a milieu in which other antioxidant defense systems are compromised [33]. The oxidation products that we describe represent a snapshot of what had occurred, rather than a quantitative measure. These proteins bind heparan sulfate [57] and matrix-associated oxidized S100A8, and S100A9 was apparent in asthmatic but not normal lung (Fig. 5). Binding may concentrate their effects at the extracellular matrix, and the

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

proteins are also likely in BALF [32]. Asthmatic participants were all on inhaled corticosteroids and because these drugs promote S100A8 expression [19,20], levels in asthmatic airways may have been boosted and comparisons of a subgroup of steroid-naı¨ve asthmatics with those on inhaled corticosteroids would be informative. Control sputum contained only weak evidence for S100A8 Cys42-sulfonic acid, and the weak immunoreactivity of enriched fractions with anti-oxS100A8 suggested low abundance, which is not unexpected, given that the airways are under constant immune surveillance against air-borne pathogens. In contrast, Cys42-containing S100A8 peptides from asthmatic sputum were similar to those generated by HOCl-treated recS100A8 (Table 1) and strongly support a role in antioxidant defense. Importantly, we found oxathiazolidine derivatives in three of five samples, confirming likely oxidation by HOCl in vivo. Noncovalent S100A8/S100A9 complexes (calprotectin) are the preferred structural form in neutrophils [10] but we found little of the predicted 24- or 48-kDa forms in unfractionated sputum or in HPLC-enriched fractions. Western blotting of sputum indicated calprotectin in only 2 of 12 controls, 3 of 14 asthmatics, and 3 of 10 bronchiectatic samples. Although we did not analyze the cell types present in sputum, it is possible that neutrophils and/or eosinophils may have contributed to the calprotectin found in these samples. Preliminary data indicated that HOCl had little apparent effect on the preformed calprotectin complex, whereas when S100A8 and S100A9 were oxidized individually, complex formation was reduced. Thus, preformed calprotectin complexes may be more resistant to oxidation although more detailed analysis is required to assess the extent of oxidation of each component and the stability of the noncovalent complex to increasing amounts of oxidant. It is unlikely that all of the S100A8 and S100A9 in asthmatic lung were present as heterocomplexes, as the antiS100A8 antibody used does not recognize the complex. Oxidation in situ may compromise complex formation and the possibility that this alters its functional capacity is worthy of more detailed investigation. S100A9 homodimer [58] (PDB Accession Database ID IIRJ) and S100A8/S100A9 heterodimer (PDB Accession Database ID IXK4) structures show Met63, Met83, and Met94 as relatively solventexposed and probably prone to oxidation, whereas Met81 in S100A9 (oxidized in one sputum sample) is buried in the interior and contributes to the dimer interface. Although somewhat inaccessible, oxidation of Met81 may disrupt the tight packing required for dimerization and/or destabilize the structure. Met1 and Met78 were oxidized to Met(O) in S100A8 from asthmatic sputum, and in S100A9, Met residues were variously modified. Oxidation of critical Met residues can be functionally important [59,60] and HOCl oxidation of S100A9 alters its effects on neutrophil migration, and mutation of Met63 and Met83 to Ala generated an oxidation-resistant protein [28]. However, if Met(O) is regenerated by reductases such as methionine-Rsulfoxide reductase B2 present in lung [61], functions altered by this modification may be reversed and these adducts may not be plentiful. The extracellular effects of S100A8, S100A9, and their heterocomplex on redox regulation are not limited to oxidant scavenging. They inhibit neutrophil oxidative metabolism, and adenosine metabolites are involved in the antioxidative effect [62]. Intracellularly, the heterocomplex promotes NADPH oxidase activation by acting as a scaffold for p67phox, p47phox, and Rac-2 assembly [63–65]. Together with the anti-inflammatory functions reported for S-nitrosylated S100A8 and S-glutathionylated S100A9 [8], the oxidant-scavenging activities of these proteins may represent a small segment of a larger, more tightly controlled redox system that requires further analysis.

185

The mechanisms described here have broad implications in inflammatory conditions in which hypohalous acid oxidants are generated by activated phagocytes and S100A8 and S100A9 are likely to be plentiful. We confirm that S100A8 was a more effective scavenger of low amounts of these oxidants than S100A9. The identification in human disease of the novel S100A8 Cys derivatives typical of those generated in vitro strongly supports the notion that S100A8 contributes to antioxidant defense during oxidative stress and promotes resolution of inflammation.

Acknowledgments We thank Professor Judy Black for asthmatic lung samples. This research was funded by Grant 455307 from the National Health and Medical Research Council of Australia. L.G. held an Australian Postgraduate Award.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.free radbiomed.2012.12.012.

References [1] Foell, D.; Wittkowski, H.; Vogl, T.; Roth, J. S100 proteins expressed in phagocytes: a novel group of damage-associated molecular pattern molecules. J. Leukocyte Biol. 81:28–37; 2007. [2] Bianchi, M. E. DAMPs PAMPs and alarmins: all we need to know about danger. J. Leukocyte Biol. 81:1–5; 2007. [3] Vogl, T.; Tenbrock, K.; Ludwig, S.; Leukert, N.; Ehrhardt, C.; van Zoelen, M. A.; Nacken, W.; Foell, D.; van der Poll, T.; Sorg, C.; Roth, J. Mrp8 and Mrp14 are endogenous activators of Toll-like receptor 4, promoting lethal, endotoxininduced shock. Nat. Med 13:1042–1049; 2007. [4] Voganatsi, A.; Panyutich, A.; Miyasaki, K. T.; Murthy, R. K. Mechanism of extracellular release of human neutrophil calprotectin complex. J. Leukocyte Biol. 70:130–134; 2001. [5] Rammes, A.; Roth, J.; Goebeler, M.; Klempt, M.; Hartmann, M.; Sorg, C. Myeloid-related protein (MRP) 8 and MRP14, calcium-binding proteins of the S100 family, are secreted by activated monocytes via a novel, tubulindependent pathway. J. Biol. Chem. 272:9496–9502; 1997. [6] Donato, R.; Geczy, C. L.; Weber, D. J. S100 proteins. In Encyclopedia of Metalloproteins Chap. 48:1–15; 2012. [7] Korndorfer, I. P.; Brueckner, F.; Skerra, A. The crystal structure of the human (S100A8/S100A9)2 heterotetramer, calprotectin, illustrates how conformational changes of interacting alpha-helices can determine specific association of two EF-hand proteins. J. Mol. Biol 370:887–898; 2007. [8] Lim, S. Y.; Raftery, M. J.; Goyette, J.; Hsu, K.; Geczy, C. L. Oxidative modifications of S100 proteins: functional regulation by redox. J. Leukocyte Biol. 86:577–587; 2009. [9] Donato, R.; Cannon, B. R.; Sorci, G.; Riuzzi, F.; Hsu, K.; Weber, D. J.; Geczy, C. L. Functions of S100 proteins. Curr. Mol. Med. 13:24–57; 2013. [10] Ehrchen, J. M.; Sunderkotter, C.; Foell, D.; Vogl, T.; Roth, J. The endogenous Toll-like receptor 4 agonist S100A8/S100A9 (calprotectin) as innate amplifier of infection, autoimmunity, and cancer. J. Leukocyte Biol. 86:557–566; 2009. [11] Steinbakk, M.; Naess-Andresen, C. F.; Lingaas, E.; Dale, I.; Brandtzaeg, P.; Fagerhol, M. K. Antimicrobial actions of calcium binding leucocyte L1 protein, calprotectin. Lancet 336:763–765; 1990. [12] Corbin, B. D.; Seeley, E. H.; Raab, A.; Feldmann, J.; Miller, M. R.; Torres, V. J.; Anderson, K. L.; Dattilo, B. M.; Dunman, P. M.; Gerads, R.; Caprioli, R. M.; Nacken, W.; Chazin, W. J.; Skaar, E. P. Metal chelation and inhibition of bacterial growth in tissue abscesses. Science 319:962–965; 2008. [13] Hsu, K.; Champaiboon, C.; Guenther, B. D.; Sorenson, B. S.; Khammanivong, A.; Ross, K. F.; Geczy, C. L.; Herzberg, M. C. Anti-infective protective properties of S100 calgranulins. Antiinflammatory Antiallergy Agents Med. Chem 8:290–305; 2009. [14] Loser, K.; Vogl, T.; Voskort, M.; Lueken, A.; Kupas, V.; Nacken, W.; Klenner, L.; Kuhn, A.; Foell, D.; Sorokin, L.; Luger, T. A.; Roth, J.; Beissert, S. The Toll-like receptor 4 ligands Mrp8 and Mrp14 are crucial in the development of autoreactive CD8þ T cells. Nat. Med 16:713–717; 2010. [15] Ramasamy, R.; Yan, S. F.; Schmidt, A. M. RAGE: therapeutic target and biomarker of the inflammatory response—the evidence mounts. J. Leukocyte Biol. 86:505–512; 2009.

186

L.H. Gomes et al. / Free Radical Biology and Medicine 58 (2013) 170–186

[16] Ravasi, T.; Hsu, K.; Goyette, J.; Schroder, K.; Yang, Z.; Rahimi, F.; Miranda, L. P.; Alewood, P. F.; Hume, D. A.; Geczy, C. Probing the S100 protein family through genomic and functional analysis. Genomics 84:10–22; 2004. [17] Endoh, Y.; Chung, Y. M.; Clark, I. A.; Geczy, C. L.; Hsu, K. IL-10-dependent S100A8 gene induction in monocytes/macrophages by double-stranded RNA. J. Immunol. 182:2258–2268; 2009. [18] Grimbaldeston, M. A.; Geczy, C. L.; Tedla, N.; Finlay-Jones, J. J.; Hart, P. H. S100A8 induction in keratinocytes by ultraviolet A irradiation is dependent on reactive oxygen intermediates. J. Invest. Dermatol. 121:1168–1174; 2003. [19] Hsu, K.; Passey, R. J.; Endoh, Y.; Rahimi, F.; Youssef, P.; Yen, T.; Geczy, C. L. Regulation of S100A8 by glucocorticoids. J. Immunol. 174:2318–2326; 2005. [20] Bozinovski, S.; Cross, M.; Vlahos, R.; Jones, J. E.; Hsuu, K.; Tessier, P. A.; Reynolds, E. C.; Hume, D. A.; Hamilton, J. A.; Geczy, C. L.; Anderson, G. P. S100A8 chemotactic protein is abundantly increased, but only a minor contributor to LPS-induced, steroid resistant neutrophilic lung inflammation in vivo. J. Proteome Res. 4:136–145; 2005. [21] Zhao, J.; Endoh, I.; Hsu, K.; Tedla, N.; Endoh, Y.; Geczy, C. L. S100A8 modulates mast cell function and suppresses eosinophil migration in acute asthma. Antioxid. Redox Signaling 14:1589–1600; 2011. [22] Lim, S. Y.; Raftery, M.; Cai, H.; Hsu, K.; Yan, W. X.; Hseih, H. L.; Watts, R. N.; Richardson, D.; Thomas, S.; Perry, M.; Geczy, C. L. S-nitrosylated S100A8: novel anti-inflammatory properties. J. Immunol. 181:5627–5636; 2008. [23] Lim, S. Y.; Raftery, M. J.; Goyette, J.; Geczy, C. L. S-glutathionylation regulates inflammatory activities of S100A9. J. Biol. Chem. 285:14377–14388; 2010. [24] Harrison, C. A.; Raftery, M. J.; Walsh, J.; Alewood, P.; Iismaa, S. E.; Thliveris, S.; Geczy, C. L. Oxidation regulates the inflammatory properties of the murine S100 protein S100A8. J. Biol. Chem. 274:8561–8569; 1999. [25] Raftery, M. J.; Yang, Z.; Valenzuela, S. M.; Geczy, C. L. Novel intra- and intermolecular sulfinamide bonds in S100A8 produced by hypochlorite oxidation. J. Biol. Chem. 276:33393–33401; 2001. [26] McCormick, M. M.; Rahimi, F.; Bobryshev, Y. V.; Gaus, K.; Zreiqat, H.; Cai, H.; Lord, R. S.; Geczy, C. L. S100A8 and S100A9 in human arterial wall: implications for atherogenesis. J. Biol. Chem. 280:41521–41529; 2005. [27] Sroussi, H. Y.; Berline, J.; Dazin, P.; Green, P.; Palefsky, J. M. S100A8 triggers oxidation-sensitive repulsion of neutrophils. J. Dent. Res. 85:829–833; 2006. [28] Sroussi, H. Y.; Berline, J.; Palefsky, J. M. Oxidation of methionine 63 and 83 regulates the effect of S100A9 on the migration of neutrophils in vitro. J. Leukocyte Biol 81:818–824; 2007. [29] Henke, M. O.; Renner, A.; Rubin, B. K.; Gyves, J. I.; Lorenz, E.; Koo, J. S. Upregulation of S100A8 and S100A9 protein in bronchial epithelial cells by lipopolysaccharide. Exp. Lung Res. 32:331–347; 2006. [30] Gray, R. D.; MacGregor, G.; Noble, D.; Imrie, M.; Dewar, M.; Boyd, A. C.; Innes, J. A.; Porteous, D. J.; Greening, A. P. Sputum proteomics in inflammatory and suppurative respiratory diseases. Am. J. Respir. Crit. Care Med. 178:444–452; 2008. [31] Gharib, S. A.; Nguyen, E. V.; Lai, Y.; Plampin, J. D.; Goodlett, D. R.; Hallstrand, T. S. Induced sputum proteome in healthy subjects and asthmatic patients. J. Allergy Clin. Immunol. 128(1176-1184):e1176; 2011. [32] Lorenz, E.; Muhlebach, M. S.; Tessier, P. A.; Alexis, N. E.; Duncan Hite, R.; Seeds, M. C.; Peden, D. B.; Meredith, W. Different expression ratio of S100A8/ A9 and S100A12 in acute and chronic lung diseases. Respir. Med. 102: 567–573; 2008. [33] Comhair, S. A.; Erzurum, S. C. Redox control of asthma: molecular mechanisms and therapeutic opportunities. Antioxid. Redox Signaling 12:93–124; 2010. [34] Xu, K.; Geczy, C. L. IFN-gamma and TNF regulate macrophage expression of the chemotactic S100 protein S100A8. J. Immunol. 164:4916–4923; 2000. [35] Iismaa, S. E.; Hu, S.; Kocher, M.; Lackmann, M.; Harrison, C. A.; Thliveris, S.; Geczy, C. L. Recombinant and cellular expression of the murine chemotactic protein, CP-10. DNA Cell Biol 13:183–192; 1994. [36] Vissers, M. C.; Carr, A. C.; Chapman, A. L. Comparison of human red cell lysis by hypochlorous and hypobromous acids: insights into the mechanism of lysis. Biochem. J. 330:131–138; 1998. [37] Gatlin, C. L.; Kleemann, G. R.; Hays, L. G.; Link, A. J.; Yates 3rd J. R. Protein identification at the low femtomole level from silver-stained gels using a new fritless electrospray interface for liquid chromatography–microspray and nanospray mass spectrometry. Anal. Biochem. 263:93–101; 1998. [38] Raftery, M. J. Enrichment by organomercurial agarose and identification of Cys-containing peptides from yeast cell lysates. Anal. Chem 80:3334–3341; 2008. [39] Efthimiadis, A.; Pizzichini, M. M.; Pizzichini, E.; Dolovich, J.; Hargreave, F. E. Induced sputum cell and fluid-phase indices of inflammation: comparison of treatment with dithiothreitol vs phosphate-buffered saline. Eur. Respir. J. 10:1336–1340; 1997. [40] Boersema, P. J.; Taouatas, N.; Altelaar, A. F.; Gouw, J. W.; Ross, P. L.; Pappin, D. J.; Heck, A. J. Mohammed, S. Straightforward and de novo peptide sequencing by MALDI-MS/MS using a Lys-N metalloendopeptidase. Mol. Cell Proteomics 8:650–660; 2009.

[41] Dalle-Donne, I.; Aldini, G.; Carini, M.; Colombo, R.; Rossi, R.; Milzani, A. Protein carbonylation, cellular dysfunction, and disease progression. J. Cell. Mol. Med. 10:389–406; 2006. [42] Elfering, S. L.; Haynes, V. L.; Traaseth, N. J.; Ettl, A.; Giulivi, C. Aspects, mechanism, and biological relevance of mitochondrial protein nitration sustained by mitochondrial nitric oxide synthase. Am. J. Physiol. Heart Circ. Physiol 286:H22–29; 2004. [43] Kuo, W. N.; Kocis, J. M. Nitration/S-nitrosation of proteins by peroxynitritetreatment and subsequent modification by glutathione S-transferase and glutathione peroxidase. Mol. Cell. Biochem 233:57–63; 2002. [44] Levine, R. L. Carbonyl modified proteins in cellular regulation, aging, and disease. Free Radic. Biol. Med. 32:790–796; 2002. [45] Moller, I. M.; Rogowska-Wrzesinska, A.; Rao, R. S. Protein carbonylation and metal-catalyzed protein oxidation in a cellular perspective. J. Proteomics 74:2228–2242; 2011. [46] Radi, R. Nitric oxide, oxidants, and protein tyrosine nitration. Proc. Natl. Acad. Sci. USA 101:4003–4008; 2004. [47] Pattison, D. I.; Davies, M. J. Kinetic analysis of the reactions of hypobromous acid with protein components: implications for cellular damage and use of 3bromotyrosine as a marker of oxidative stress. Biochemistry 43:4799–4809; 2004. [48] Fu, X.; Mueller, D. M.; Heinecke, J. W. Generation of intramolecular and intermolecular sulfenamides, sulfinamides, and sulfonamides by hypochlorous acid: a potential pathway for oxidative cross-linking of low-density lipoprotein by myeloperoxidase. Biochemistry 41:1293–1301; 2002. [49] Winterbourn, C. C. Comparative reactivities of various biological compounds with myeloperoxidase–hydrogen peroxide–chloride, and similarity of the oxidant to hypochlorite. Biochim. Biophys. Acta 840:204–210; 1985. [50] Hawkins, C. L.; Pattison, D. I.; Davies, M. J. Hypochlorite-induced oxidation of amino acids, peptides and proteins. Amino Acids 25:259–274; 2003. [51] Skaff, O.; Pattison, D. I.; Davies, M. J. Kinetics of hypobromous acid-mediated oxidation of lipid components and antioxidants. Chem. Res. Toxicol. 20:1980–1988; 2007. [52] Lackmann, M.; Rajasekariah, P.; Iismaa, S. E.; Jones, G.; Cornish, C. J.; Hu, S.; Simpson, R. J.; Moritz, R. L.; Geczy, C. L. Identification of a chemotactic domain of the pro-inflammatory S100 protein CP-10. J. Immunol. 150:2981–2991; 1993. [53] Ishikawa, K.; Nakagawa, A.; Tanaka, I.; Suzuki, M.; Nishihira, J. The structure of human MRP8, a member of the S100 calcium-binding protein family, by MAD phasing at 1.9 A resolution. Acta Crystallogr. D Biol. Crystallogr 56:559–566; 2000. [54] Sohnle, P. G.; Hunter, M. J.; Hahn, B.; Chazin, W. J. Zinc-reversible antimicrobial activity of recombinant calprotectin (migration inhibitory factorrelated proteins 8 and 14). J. Infect. Dis. 182:1272–1275; 2000. [55] Weiss, S. J. Tissue destruction by neutrophils. N. Engl. J. Med. 320:365–376; 1989. [56] Gray, R. D.; Imrie, M.; Boyd, A. C.; Porteous, D.; Innes, J. A.; Greening, A. P. Sputum and serum calprotectin are useful biomarkers during CF exacerbation. J. Cystic Fibrosis 9:193–198; 2010. [57] Robinson, M. J.; Tessier, P.; Poulsom, R.; Hogg, N. The S100 family heterodimer, MRP-8/14, binds with high affinity to heparin and heparan sulfate glycosaminoglycans on endothelial cells. J. Biol. Chem. 277:3658–3665; 2002. [58] Itou, H.; Yao, M.; Fujita, I.; Watanabe, N.; Suzuki, M.; Nishihira, J.; Tanaka, I. The crystal structure of human MRP14 (S100A9), a Ca2 þ -dependent regulator protein in inflammatory process. J. Mol. Biol. 316:265–276; 2002. [59] Li, C.; Chen, H.; Ding, F.; Zhang, Y.; Luo, A.; Wang, M.; Liu, Z. A novel p53 target gene, S100A9, induces p53-dependent cellular apoptosis and mediates the p53 apoptosis pathway. Biochem. J. 422:363–372; 2009. [60] Manzanares, D.; Rodriguez-Capote, K.; Liu, S.; Haines, T.; Ramos, Y.; Zhao, L.; Doherty-Kirby, A.; Lajoie, G.; Possmayer, F. Modification of tryptophan and methionine residues is implicated in the oxidative inactivation of surfactant protein B. Biochemistry 46:5604–5615; 2007. [61] Hansel, A.; Jung, S.; Hoshi, T.; Heinemann, S. H. A second human methionine sulfoxide reductase (hMSRB2) reducing methionine-R-sulfoxide displays a tissue expression pattern distinct from hMSRB1. Redox Rep 8:384–388; 2003. [62] Sroussi, H. Y.; Lu, Y.; Zhang, Q. L.; Villines, D.; Marucha, P. T. S100A8 and S100A9 inhibit neutrophil oxidative metabolism in-vitro: involvement of adenosine metabolites. Free Radic. Res. 44:389–396; 2010. [63] Kerkhoff, C.; Nacken, W.; Benedyk, M.; Dagher, M. C.; Sopalla, C.; Doussiere, J. The arachidonic acid-binding protein S100A8/A9 promotes NADPH oxidase activation by interaction with p67phox and Rac-2. FASEB J. 19:467–469; 2005. [64] Doussiere, J.; Bouzidi, F.; Vignais, P. V. The S100A8/A9 protein as a partner for the cytosolic factors of NADPH oxidase activation in neutrophils. Eur. J. Biochem 269:3246–3255; 2002. [65] Doussiere, J.; Bouzidi, F.; Vignais, P. V. A phenylarsine oxide-binding protein of neutrophil cytosol, which belongs to the S100 family, potentiates NADPH oxidase activation. Biochem. Biophys. Res. Commun. 285:1317–1320; 2001.