The glycosylation of myeloperoxidase

Biochimica et Biophysica Acta 1804 (2010) 2046–2053

Contents lists available at ScienceDirect

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

The glycosylation of myeloperoxidase Tina Ravnsborg a, Gunnar Houen b, Peter Højrup a,⁎ a b

Institute of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark Statens Serum Institut, Department of Clinical Biochemistry, Artillerivej 5, DK-2300 Copenhagen, Denmark

a r t i c l e

i n f o

Article history: Received 22 March 2010 Received in revised form 29 June 2010 Accepted 1 July 2010 Available online 16 July 2010 Keywords: Myeloperoxidase Glycosylation Cancer Epitope mapping Mass spectrometry

a b s t r a c t The enzyme myeloperoxidase (MPO) is an important part of the neutrophil immune reaction and can be found in alfa granula. The presence of MPO can be used to distinguish acute myelogenous leukemia from acute lymphocytic leukemia. However, the methods employed to do so, such as flow cytometry and immunohistochemistry rely on antibody recognition, and therefore the characterization of the mature MPO, including post-translational modifications, must be considered as important as epitope mapping. MPO has 5 N-linked glycosylation sites, occupied by both high mannose and complex glycan structures. In this study we utilize intact glycopeptide MSMS analysis for site specific characterization of the glycan structures of MPO from a cancer patient. The identified glycan structures are compared to those of MPO from healthy donors, in order to probe for any potential differences that may have diagnostic use. © 2010 Elsevier B.V. All rights reserved.

1. Introduction MPO is an enzyme residing in the azurophilic granules of neutrophils and its main function is to produce hypochlorous acid (HOCl) that serves to kill pathogens as part of the immune response [1–3]. The mature MPO protein consists of 2 light and 2 heavy polypeptide chains with each heavy chain binding a prosthetic heme group [4,5]. Myeloperoxidase (MPO) has for some years received attention based on its application in the distinction between acute myelogenous leukemia (AML) and acute lymphocytic leukemia (ALL) [6–10]. For classification of leukemia cases the presence of MPO may be verified by flow cytometry or immunohistochemistry [11–15]. For these applications, it is important to verify that the antibodies used reliably detect myeloperoxidase in all samples where it is present, whether or not changes in MPO structure or post-translational modifications may have occurred. In this respect, it is necessary to know the structure of MPO as well as all possible post-translational modifications. In addition, it is important to obtain knowledge of the epitopes of the antibodies used, including their potential dependency on posttranslational modifications. Furthermore, MPO has been

Abbreviations: ALL, Acute lymphocytic leukemia; AML, Acute myelogenous leukemia; DHB, 2,5-dihydroxybenzoic acid; FA, Formic acid; Fuc, Fucose; GlcNAc, NAcetylglucosamine; Hex, Hexose; HexNAc, N-Acetylhexosamine; MALDI, Matrix assisted laser desorption ionization; Man, Mannose; MeCN, Acetonitrile; MeOH, Methanol; MPO, Myeloperoxidase; QTOF, Quadropole time of flight; Sia, Sialic acid; TFA, Trifluoroacetic acid ⁎ Corresponding author. Tel.: +45 65502371; fax: +45 65502467. E-mail address: [email protected] (P. Højrup). 1570-9639/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2010.07.001

shown to be the target of autoantibodies in several autoimmune conditions, including several forms of vasculitis [16,17] making epitope mapping and characterization of the mature protein equally important tasks. MPO has been shown to be highly glycosylated [18–24] and the extent and identity of the glycosylations may be of importance in terms of antibody recognition and epitope mapping. Moreover, it has been shown that protein glycosylation may change in cases of cancer [25–27], providing further means of distinction between cancer and non cancer cases. Aside from one consensus site in the propeptide, MPO has 5 potential N-linked glycosylation sites. Previously X-ray crystallography and biochemical studies had indicated that only 4 (N323, N355, N391, N483) of these sites were glycosylated and only N483 had been assigned a full glycan structure (GlcNAc2Man3Fuc1) [18–21,23]. Also, high mannose and complex structures had been suggested based on glycosidase treatment studies, with secreted precursor MPO containing more of the complex type structures than intracellular mature and precursor MPO [22,28]. However; a recent study by Van Antwerpen et al. [24], employing mass spectrometry, have demonstrated that all 5 N-linked glycosylation sites of both recombinant and human MPO are occupied by either complex or high mannose structures. In the same study it was found that the glycosylations were important for the enzyme activity of MPO. Here we address the glycan structure of the 5 N-linked glycosylation sites of MPO by intact glycopeptide MSMS analysis. Furthermore, we compare the glycosylation of MPO between a sample from healthy donors and that of a cancer patient, in order to determine whether any potential differences might be of use for diagnostic purposes.

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

2. Materials and methods 2.1. Purification of MPO A pool of MPO from six healthy donors was purified from a fraction obtained during purification of human neutrophil granulocyte proteinase 3 (PR3) [29]. Briefly, 6 buffy coats were used as a source to obtain neutrophil granulocytes by density gradient centrifugation. The granulocytes were disrupted by nitrogen cavitation and azurophilic granula isolated by density gradient centrifugation with Percoll (GE Healthcare, WI, USA). The granula were mixed 1:1 with 50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-114 and sonicated on ice for 2 × 20 s. The extract was incubated at 37 °C and the water phase (containing MPO) was separated from the detergent phase (containing PR3). The water phase was dialysed against 50 mM Tris, pH 8 and loaded on a Mono S Sepharose column (GE Healthcare, WI, USA), which was eluted with a linear gradient to 1 M NaCl in 50 mM Tris, pH 8.0. MPO containing fractions were located by ELISA using rabbit anti human MPO A-0398 1:1000 (DAKO, Denmark). The fractions were pooled and concentrated using a centriprep cartridge, cut-off 10.000 Da (Millipore, MA, USA). Yield: 0.18 mg, (determined at A430, with ε430 = 1.25). MPO from a patient with chronic myeloid leukaemia was obtained from Dr. Inge Olsson (University Hospital of Lund) and had been purified as described previously [30]. 2.2. Identification of MPO glycopeptides by MALDI MS One microgram of each MPO sample in 10 μl 50 mM NH4HCO3 was incubated with 10 mM dithiothreitol for 30 min at 56 °C, followed by alkylation with 30 mM iodacetamide for 30 min in the dark at room temperature. An additional 5 mM dithiothreitol was added to quench the reaction and the volume was adjusted to 40 μl with 50 mM NH4HCO3. In solution trypsinization was performed by incubating each of the reduced and alkylated samples with 2% trypsin (Promega, WI, USA) (W/W) overnight. at 37 °C. Glycopeptides were enriched on a graphite micro column custom made in a GELoader tip (Eppendorf, Germany), using graphite powder obtained from a prepacked 1 ml graphite column (Alltech, Australia) [31]. The column was equilibrated with H2O, the sample added, and washed twice with H2O. Glycopeptides were eluted on target with 0.6 μl 20% MeCN, 0.1% TFA and 0.6 μl 20 mg/ml DHB in 70% MeCN, 0.1% TFA was added. MS and MSMS were performed on a MALDI QTOF (Micromass, Waters, MA, USA). 2.3. SDS PAGE Aliquots of 1 μg MPO were subjected to SDS PAGE using a precast 10% Tris-glycine polyacrylamide gel (Lonza, Switzerland) and 25 mM Tris, 192 mM glycine, 0.1% SDS as running buffer. Prior to loading on the gel, the samples were mixed in a 1:1 ratio with sample buffer containing 2-mercaptoethanol and bromophenol blue, and boiled for 5 min. Page RulerTM Prestained Protein Ladder (Fermentas, Canada) was used as marker. The gel was run for 1 h at 160 V, stained in 0.25% Coomassie Brilliant Blue R250 in 45% MeOH, 10% acetic acid for 45 min and destained overnight in H2O. 2.4. In-gel digestion and PNGaseF treatment The bands containing MPO were cut from each lane in the gel and washed twice in 50% MeCN and once in 100% MeCN. The gel plugs were first incubated for 30 min at 56 °C with 10 mM dithiothreitol and then for 30 min with 55 mM iodacetamide in the dark at room temperature. The reduction and alkylation was followed by another washing step as described above. Protein digestion was achieved by adding 0.125 μg trypsin (Promega, WI, USA) in 15 μl 50 mM NH4HCO3

2047

to each sample. The gel plugs were left on ice for 10 min, excess liquid was removed, 20 μl of NH4HCO3 was added and the samples left at 37 °C overnight. In one experiment the samples were redigested for 3 h at 37 °C with 0.02 μg AspN (Calbiochem, Merck, UK). One in-gel digested sample from a cancer patient was subjected to PNGaseF treatment by adding 0.5 U PNGase F (Roche, Germany) to the sample and incubating overnight at 37 °C. 2.5. LC-Orbitrap MS and ETD/CID MSMS The digested MPO samples were prepared for LC-Orbitrap MS by purification on custom made micro columns. The columns were made in GELoader tips (Eppendorf, Germany) using a small piece of Empore C18 disk (3 M, MN, USA) as a plug with 1–2 mm oligo R3 resin (Applied Biosystems, CA, USA) on top. The columns were washed in 100% MeCN and equilibrated in 0.1% TFA before samples were loaded. The columns were then washed twice in 0.1% TFA and eluted with 60% MeCN, 0.1% TFA. The eluted samples were dried down and redissolved in 0.3 μl 100% FA followed by 5 μl 0.1% FA. The samples were run on an Easy LC system (Proxeon, Denmark) coupled in-line to an LTQOrbitrap instrument equipped with ETD (Thermo, Germany). Chromatographic separation was obtained using a C18 column, 3 μm Reprosil (Dr. Maisch Gmbh, Germany), of 20 cm in length with an inner diameter of 100 μm and a buffer system consisting of buffer A: 0.1% FA and B: 0.1% FA, 95% MeCN. The samples were loaded with a flow of 550 nl/min and a gradient of 0–34% buffer B in 30 min with a flow of 250 nl/min was applied, followed by 34–100% buffer B in 4 min. For each pre-scan in the Orbitrap the 5 most intense peaks were subjected to MSMS by either ETD or CID in the LTQ, while a high precision full-scan was obtained in the Orbitrap. All together the sample obtained from a cancer patient and the pooled sample obtained from donors were each analyzed five times by in gel digestion and LC-Orbitrap MS and MSMS. Each of the following four experimental combinations were applied once; in gel trypsinization with or without AspN redigestion and MSMS performed by either CID or ETD, with the in gel trypsinization alone combined with CID for fragmentation applied twice. In addition to these five runs, the sample obtained from a cancer patient was analyzed once by in gel trypsinization followed by PNGase F treatment and CID MSMS, and once by in solution trypsinization and CID MSMS. 2.5.1. Antibody binding assays For enzyme-linked immunosorbent assay (ELISA) 10 μg of MPO from a cancer patient was subjected to deglycosylation by incubation with 10 U PNGaseF (Roche, Germany) in 50 μl 50 mM NH4HCO3 at 37 °C overnight. The deglycosylation was confirmed by MALDI QTOF MS after trypsinization and enrichment on graphite as described above. Polystyrene microtitre plates (Nunc, Denmark) were coated for 90 min at room temperature with 1 μg/ml untreated or deglycosylated MPO, 1 U/ml PNGaseF or 5 μl/ml NH4HCO3 in 100 μl 50 mM NaHCO3 pH 9.6 pr well. The plates were washed 3 times and blocked for 15 min in TTN buffer; 50 mM Tris, 0,3 M NaCl 1% tween 20, pH 7.5. The monoclonal antibodies; ab25989 (Abcam, England), LS-C41696 (LifeSpan, WA, USA) and HYB194-01 (SSI, Denmark) were added for 1 h at room temperature in TTN buffer using the ratios 1:1000, 1:1000 and 1:10 respectively. The plates were washed 3 times in TTN buffer and incubated with alkaline phosphatase conjugated secondary antibody A-3688 (Sigma-Aldrich, MO, USA) at a concentration of 1:1000 in TTN buffer for 1 h at room temperature. After a final washing step, 1 mg/ml p-nitrophenylphosphate (Sigma-Aldrich, MO, USA) in 1 M diethanolamine, 0.5 mM MgCl2 was added and the absorbance read at 405 nm, with background subtraction at 650 nm, after 5, 10 and 20 min on a VERSAmax microplate reader (Molecular Devices, CA, USA).

2048

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

2.6. Bioinformatics Peptide masses and fragment ions were calculated using the GPMAW software package (Lighthouse data, Denmark).

3. Results For the initial experiments purified MPO, from either a cancer patient or pooled from healthy donors, was subjected to trypsinization and enriched for glycopeptides on a graphite micro column prior to MALDI MS. Glycopeptides were identified by their theoretical mass and by calculated mass differences corresponding to monosaccharides (Fig. 1, Table 1). All the potential N-linked glycosylation sites were represented in the spectrum from the cancer patient sample, however, only a few could be confirmed by MSMS (marked in Table 1) due to large m/z values and low abundance. In this initial experiment only a few differences in the glycosylation pattern were found between the sample from the cancer patient and the one from the healthy donors. The glycopeptide containing N729 was only seen in the cancer patient sample and not in the donor sample, and the high mannose structure on N323 was seen only in the donor sample. In order to further investigate these differences and also obtain MSMS verifications for all glycopeptides, both samples were subjected to in-gel trypsinization and run on an LC- LTQ-Orbitrap instrument. Each sample was run twice, first using CID for fragmentation and secondly using ETD for fragmentation. Not regarding the glycopeptides, a coverage of 74.7% was obtained for the CID run of the cancer patient sample (Fig. 2). This coverage included the N-linked glycosylation sites N355 and N729, indicating that not all of the MPO molecules contain glycosylations on these two sites. The coverage for the donor sample was similar (data not shown). When treating the trypsinized sample from the cancer patient with PNGaseF prior to LC MS an additional 5.3% coverage was found including the three N-linked glycosylation sites N323, N391 and N483. On the basis of the potential glycopeptides identified in Fig. 1 we verified 38 glycopeptides on the MSMS level from the LC-LTQ-

Orbitrap runs (Table 2). As suspected from the initial MALDI experiments, all 5 N-linked glycosylation sites were glycosylated. N323 was found to be occupied by either a high mannose or a complex structure (Fig. 3). An addition of 16 Da was found on some of these glycopeptides, although, there were too few peptide fragments to assign the precise position of the added mass. However, the deglycosylated peptide containing N323 from the sample treated with PNGaseF was also found in two versions with an added mass of 16 Da that could be assigned to either C316 or C319. We assume that this oxidation has happened after alkylation and is an artifact. The glycosylation on N355 was identified as a high mannose structure (see supplementary data, Fig. 6). Due to a missed cleavage two different versions of the peptide was found. The missed cleavage seems to be affected by the number of mannose residues in the glycan structure, the higher the number of mannose residues the higher the chance of a missed cleavage after position 354 (Table 2). This observation is supported by the fact that only the short version of the peptide shows up in the coverage for the non PNGaseF-treated sample (Fig. 2). Another high mannose structure was found on N391 (see supplementary data, Fig. 6), though with less mannose residues than N355. In the sample from the cancer patient N483 was found to be occupied by either a high mannose or a complex glycan structure (Fig. 4). The high mannose structure was not found in the sample from the healthy donors. The glycopeptide containing N483 with the mass 2949.14 Da, was primarily identified based on glycan fragmentation and precise mass, as only a few or no peptide fragments were found in the spectrum. N729 was found with or without (Fig. 2) a complex glycan structure attached (supplementary data, Fig. 7) and the sialylated version of the glycopeptide (3657.56 Da) was only seen and identified in a single run of the sample subjected to in-solution trypsinization. The glycosylation pattern for all the glycosylation sites is identical between the cancer patient and the donor sample except for the high mannose structure of N483, which was only found in the sample from

Table 1 MPO glycopeptides assigned in the MALDI spectra from Fig. 1. Assignment is based on the theoretical masses and mass differences corresponding to monosaccharide residues. *: assignment confirmed by MSMS on MALDI QTOF. Glycopeptide

N323 SCPACPGSNITIR, HexNAc2Hex2Fuc1 SCPACPGSNITIR, HexNAc2Hex3Fuc1 SCPACPGSNITIR + 16, HexNAc2Hex3Fuc1 SCPACPGSNITIR, HexNAc2Hex5 SCPACPGSNITIR, HexNAc2Hex6 N355 NMSNQLGLLAVNQR + 16, HexNAc2Hex4 NMSNQLGLLAVNQR + 16, HexNAc2Hex5 NMSNQLGLLAVNQR + 16, HexNAc2Hex6 NMSNQLGLLAVNQR + 16, HexNAc2Hex7 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex5 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex6 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex7 N391 ALLPFDNLHDDPCLLTNR, HexNAc2Hex3 ALLPFDNLHDDPCLLTNR, HexNAc2Hex4 ALLPFDNLHDDPCLLTNR, HexNAc2Hex5 ALLPFDNLHDDPCLLTNR, HexNAc2Hex6 N483 SYNDSVDPR, HexNAc2Hex3Fuc1 SYNDSVDPR, HexNAc3Hex3Fuc1 SYNDSVDPR, HexNAc3Hex4Fuc1 SYNDSVDPR, HexNAc3Hex4Fuc1Sia1 N729 DFVNCSTLPALNLASWR, HexNAc2Hex2Fuc1

Theoretical MH+

Observed MH+ Cancer (A)

Donor (B)

2308.98 2471.04 2487.04 2649.08 2811.14

– 2471.13 2487.17 – –

2309.08 2471.17 2487.11 2649.20 2811.26

2628.18 2790.23 2952.29 3114.34 3173.46 3335.52 3497.57

2628.28 2790.33* 2952.38* 3114.45 3173.59 3335.63 3497.66

2628.26 2790.35 2952.41 3114.48 3173.59 3335.68 3497.74

3016.37 3178.42 3340.47 3502.53

3016.48* 3178.54 3340.61* 3502.67

3016.48 – – 3502.69

2090.84 2293.91 2455.97 2747.06

2090.92 2294.03 2456.12 –

2090.92 2294.05 2456.10 2747.15

2840.28

2840.38

–

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

2049

Fig. 1. MALDI MS of MPO glycopeptides. The spectra were obtained on a MALDI Q TOF instrument. MPO from (A) a cancer patient or (B) healthy donors was trypsinated and the resulting glycopeptides were enriched on a micro graphite column prior to MS. All mass values indicated in the spectrum correspond to glycopeptides and have been assigned in Table 1.

the cancer patient. Another difference was found for the N729 site where only the smallest version of the glycopeptide was found in the sample from the donor. However, since this site is not glycosylated on all of the MPO molecules the glycopeptide may be of rather low abundance and could therefore have been lost in the background noise. All glycopeptides were primarily identified from CID spectra. Only one glycopeptide was verified based on ETD alone (marked in Table 2) due to the lack of a CID spectrum. Good ETD spectra were only obtained for those glycopeptides that also yielded good CID spectra with both glycosylation and peptide fragmentation. In the case of the glycopeptides 315–327 and 481–489 where CID gave rise to only a few peptide fragments, ETD performed very poorly due to the inability of these glycopeptides to obtain more than 2 charges under the experimental conditions used. The experiments were repeated with an additional digestion step after trypsination, using AspN to generate alternative glycopeptides containing residue N391 and N483, however, no further glycan structures were found (data not shown). A higher than usual number of amino acid residues was found to be oxidized in both the sample from the cancer patient and in the sample from the donors. Altogether, 13 methionine and 3 tryptophan

residues were observed to be either singly (+ 16 Da) or doubly (+32 Da) oxidized (data not shown). However since these types of oxidations often occur during sample preparation, we cannot rule out that they are merely an artifact. Triple oxidations (+48 Da) were observed on the cysteine residues at positions 309, 398, 663 and 704 corresponding to their conversion into cysteic acid residues (data not shown). This, we believe, is unlikely to be an artifact of sample preparation. Preliminary studies were carried out to address the dependency of glycosylation for antibody binding to MPO. The results showed that for the monoclonal antibodies; ab25989, LS-C41696 and HYB194-01 tested by ELISA, the binding between antibody and MPO was respectively 2.4, 2.9 and 2.0 times higher for untreated MPO as compared to deglycosylated MPO (data not shown). The incubation buffer and PNGaseF alone did not give rise to any antibody binding. 4. Discussion In this study we have performed site specific characterization of the glycan structures of all 5 N-linked glycosylation sites of human MPO. The characterization was carried out using intact glycopeptide MSMS analysis, which revealed high mannose structures on N355 and

Fig. 2. Peptide coverage of MPO from a cancer patient after tryptic digestion. A coverage of 74.7% was obtained and is highlighted in dark gray on the protein sequence (Swiss-Prot accession nr. P05164). Residue 1–48 constitutes the signal peptide and 49–164 the propeptide, therefore these residues were not taken into account when calculating the coverage. The 6 potential N-glycosylation sites are underlined. Following PNGaseF treatment additional peptides were found: 352–368, 355–368, 375–392, 315–327 and 481–489 giving rise to a further coverage of 5.3% (highlighted in light gray). Residue 165–278 constitutes the light chain and 279–745 the heavy chain of MPO.

2050

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

Table 2 Identified MPO glycopeptides from cancer patient and healthy donor. All glycopeptides were verified by high precision mass (10 ppm) and CID MSMS data obtained on an LTQ Orbitrap instrument. In some cases the identifications were backed up by ETD MSMS data. (1) Glycopeptide identified based on precise mass and ETD MSMS alone, fragmentation was seen in both glycan and peptide moiety. (2) Glycopeptide identified based on precise mass and fragmentation of the glycan moiety alone by CID, little or no fragmentation of the peptide moiety was seen. (3) Glycopeptide only determined from an in-solution tryptic digest of MPO, identification based on precise mass and fragmentation of the glycan moiety by CID, no peptide fragmentation was observed. Theoretical M

ID

Cancer

Donor

1634.74 2161.92 2323.98 2486.03 2648.08 2664.08 2810.14 2826.14 2988.19

SCPACPGSNITIR, HexNAc1 SCPACPGSNITIR, HexNAc2Hex2 SCPACPGSNITIR, HexNAc2Hex3 SCPACPGSNITIR, HexNAc2Hex4 SCPACPGSNITIR, HexNAc2Hex5 SCPACPGSNITIR + 16, HexNAc2Hex5 SCPACPGSNITIR, HexNAc2Hex6 SCPACPGSNITIR + 16, HexNAc2Hex6 SCPACPGSNITIR + 16, HexNAc2Hex7

+ + + + + + + + +

+ + + + + + + + +

1780.8 2307.98 2323.98 2470.04 2486.04 3126.26

SCPACPGSNITIR, HexNAc1Fuc1 SCPACPGSNITIR, HexNAc2Hex2Fuc1 SCPACPGSNITIR + 16, HexNAc2Hex2Fuc1 SCPACPGSNITIR, HexNAc2Hex3Fuc1 SCPACPGSNITIR + 16, HexNAc2Hex3Fuc1 SCPACPGSNITIR, HexNAc3Hex4Fuc1Sia1

+ + + + + +

+ + + + + +

2627.18 2789.23 2951.29 3113.34 3172.46 3334.52 3496.57 3658.62

NMSNQLGLLAVNQR + 16, HexNAc2Hex4 NMSNQLGLLAVNQR + 16, HexNAc2Hex5 NMSNQLGLLAVNQR + 16, HexNAc2Hex6 NMSNQLGLLAVNQR + 16, HexNAc2Hex7 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex5 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex6 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex7 NLRNMSNQLGLLAVNQR + 16, HexNAc2Hex8

+ + + + + + + +1

+ + + + + + + +

3015.37 3177.42 3339.47 3501.53

ALLPFDNLHDDPCLLTNR, ALLPFDNLHDDPCLLTNR, ALLPFDNLHDDPCLLTNR, ALLPFDNLHDDPCLLTNR,

+ + + +

+ + + +

2267.88 2429.94

SYNDSVDPR, HexNAc2Hex5 SYNDSVDPR, HexNAc2Hex6

+ +

− −

2089.84 2292.91 2454.97 2746.06 2949.14

SYNDSVDPR, SYNDSVDPR, SYNDSVDPR, SYNDSVDPR, SYNDSVDPR,

+ + + + +2

+ + − + +2

2693.22 2839.28 3001.34 3657.56

DFVNCSTLPALNLASWR, HexNAc2Hex2 DFVNCSTLPALNLASWR, HexNAc2Hex2Fuc1 DFVNCSTLPALNLASWR, HexNAc2Hex3Fuc1 DFVNCSTLPALNLASWR, HexNAc3Hex4Fuc1Sia1

+ + + +3

+ + − −

N323 High mannose

Complex

N355

N391 HexNAc2Hex3 HexNAc2Hex4 HexNAc2Hex5 HexNAc2Hex6

N483 High mannose

Complex HexNAc2Hex3Fuc1 HexNAc3Hex3Fuc1 HexNAc3Hex4Fuc1 HexNAc3Hex4Fuc1Sia1 HexNAc4Hex4Fuc1Sia1

N729

N391, a complex glycan on N729 and either high mannose structures or complex glycans on N323 and N483. This is in concordance with both previous X-ray crystallography, biochemical and glycosidase treatment studies [18–23,28]. We also found evidence suggesting that 3 of the 5 sites are fully glycosylated, whereas N355 and N729 are not glycosylated on all MPO molecules. The findings of this study are in overall agreement with the results presented in a recent paper by Van Antwerpen et al. [24] concerning the glycosylation of human native and recombinant MPO. Only the reported high mannose structures on N729 of human MPO are in contrast to the complex structures observed here. Another less distinct difference between the two studies is the low abundant complex structures on N391 not observed here. These small discrepancies may however be attributed to differences in the experimental setups. The deglycosylation based setup used by Van Antwerpen requires a relatively larger amount of sample, but is likely to provide a better coverage at the expense of the small but always present risk of

contaminating free glycans. In our setup we used much smaller sample amounts and may have missed some of the very low abundant glycan species. However, MSMS analysis of intact glycopeptides has the advantage of making contaminating free glycans irrelevant. According to our results 3 of the 5 glycosylation sites are fully occupied by glycan structures and 2 are occupied in at least a subset of the MPO molecules. As can be seen from Fig. 5, the 5 glycosylation sites are spread fairly even across the surface of MPO. This implies that, with all five sites occupied, a large part of the surface of MPO will be covered by the glycan structures. It is therefore reasonable to speculate whether the glycosylation status of MPO could be of any consequence to the antibody recognition used for detection of MPO in leukaemia classification, and perhaps also to the recognition by autoantibodies in autoimmune conditions such as vasculitis [16,17]. We have here done some preliminary studies on the importance of the glycosylations of MPO in relation to antibody binding. These studies showed a minor decrease in binding of the 3 tested antibodies

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

2051

Fig. 3. MSMS of high mannose and complex glycan structure on residue N323 of MPO. The spectra were obtained in the LTQ of an LTQ Orbitrap system using CID. (A) MSMS of peptides 315–327 with a high mannose structure attached to N323 (M: 2988.21 Da, m/z: 1495.103). An oxidation (+ 16 Da) was assigned to SCPA (315–318), since b4 was found with a mass addition of 16 Da. (B) Peptides 315–327 with a complex glycan structure attached to N323 (M: 3126.28 Da, m/z: 1564.141). In this spectrum the base peak has been reduced by 2/3 in size to enable viewing of the remaining peaks. Both glycopeptides arise from tryptic digestion of MPO from a cancer patient. □: HexNAc, о: Hex, Δ: Fuc, ◊: Sia. For additional peaks * that have been assigned see supplementary data, Table 3.

to deglycosylated MPO as compared to the untreated MPO. This observation suggests that while the glycosylation is not paramount to antibody recognition it may still have some influence on the binding efficiency and thus should be investigated further. Previously a number of studies have been done in effort to map the epitopes of MPO recognized by antineutrophil cytoplasmic autoanti-

bodies (ANCA) found in patients suffering from various forms of vasculitis. However, to our knowledge, none of these studies broach the potential significance of the protein glycosylation in this antibody–antigen interaction [32–35]. It is our belief that taking the glycosylation of MPO into account could provide important information regarding ANCA recognition.

Fig. 4. MSMS of high mannose and complex glycan structure on residue N483 of MPO. The spectra were obtained in the LTQ of an LTQ Orbitrap system using CID. (A) MSMS of peptides 481–489 with a high mannose structure attached to N483 (M: 2429.95 Da, m/z: 1215.98). (B) Peptides 481–489 with a complex glycan structure attached to N483 (M: 2949.16 Da, m/z: 1475.58). In this spectrum the base peak has been reduced by 2/3 in size to enable viewing of the remaining peaks. Both glycopeptides arise from tryptic digestion of MPO from a cancer patient. □: HexNAc, о: Hex, Δ: Fuc, ◊: Sia. For additional peaks * that have been assigned see supplementary data, Table 3.

2052

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

Acknowledgements The Danish Cancer Society is gratefully acknowledged for financial support. Dorthe Tange Olsen is thanked for purification of MPO. Morten Zoega Hansen is thanked for preparation of alfa granula. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbapap.2010.07.001. References

Fig. 5. 3D structure of MPO. The N-linked glycosylation sites (black) are mapped onto the 3D structure of MPO. The structure (3f9p) was obtained from the PDB database and is based on X-ray studies by Carpena et al. [23].

The type of study performed here reveals no information as to the biological function of the profound glycosylations of MPO. Yet, the simple observation that a higher number of mannose residues on the glycan at N355, correlates with a higher risk of a missed cleavage after position 354, could indicate that one purpose served by the glycans is protease protection. From the experiments done here we saw no major differences in the glycan structures of MPO from the cancer patient as compared to that from the healthy donors. Only one small discrepancy was observed between the two samples, namely the high mannose structures containing 5 and 6 mannose residues found on N483 in the cancer sample but not in the donor sample. These high mannose structures were also not reported by Van Antwerpen et al., who used human MPO of non-specified medical origin. However, any conclusions as to the statistical importance of this single variation must await further analysis of additional samples. It has been reported that glycoproteins from cancer samples exhibit altered sialylation both in terms of linkage type and extend of sialylation [25,26]. In our study we do not see such differences, in fact we see only a few sialylated glycopeptides altogether. Methods more specialized for sialic acid enrichment, such as titaniumdioxide [36] and HILIC [37] enrichment, were also employed, but no additional sialylated glycopeptides were found. Extensive sialylation is often present on cell surface and extracellular glycoproteins, and the low degree of sialylation found on MPO may be attributed to it being a stored and not secreted granule protein. It could also be speculated that the observed lack of sialic acid residues, which are rather labile, may be due to the harsh conditions of the granules, partly inflicted by hypochlorous acid produced by MPO itself, and that these conditions are also likely to explain the amino acid oxidations observed in this study.

[1] S.J. Klebanoff, Myeloperoxidase: contribution to the microbicidal activity of intact leukocytes, Science 169 (1970) 1095–1097. [2] S.J. Klebanoff, Myeloperoxidase, Proc. Assoc. Am. Physicians 111 (1999) 383–389. [3] J.K. Hurst, W.C. Barrette, Leukocytic oxygen activation and microbicidal oxidative toxins, Crit. Rev. Biochem. Mol. Biol. 24 (1989) 271–328. [4] M. Yamada, S. Hur, K. Hashinaka, K. Tsuneoka, T. Saeki, C. Nishio, F. Sakiyama, S. Tsunasawa, Isolation and characterization of a cDNA coding for human myeloperoxidase, Arch. Biochem. Biophys. 255 (1987) 147–155. [5] P.G. Furtmüller, M. Zederbauer, W. Jantschko, J. Helm, M. Bogner, C. Jakopitsch, C. Obinger, Active site structure and catalytic mechanisms of human peroxidases, Arch. Biochem. Biophys. 445 (2006) 199–213. [6] J. Storr, G. Dolan, E. Coustan-Smith, D. Barnett, J.T. Reilly, Value of monoclonal anti-myeloperoxidase (MPO7) for diagnosing acute leukaemia, J. Clin. Pathol. 10 (1990) 847–849. [7] D. Drach, J. Drach, H. Glassl, C. Gattringer, H. Huber, Flow cytometric detection of cytoplasmic antigens in acute leukemias: implications for linage assignment, Leuk. Res. 17 (1993) 455–461. [8] M.P. Karnik, C.N. Nair, S.M. Zingde, B.P. Gothoskar, L. Zachariah, S. Barbhaya, S.H. Advani, Use of polyclonal anti-myeloperoxidase antibody in myeloid lineage determination, Indian J. Med. Res. 100 (1994) 272–276. [9] R. Stasi, G. Del Poeta, A. Venditti, A. Bruno, G. Suppo, G. Aronica, G. Di Carlo, G. Papa, Lineage identification of acute leukemias: relevance of immunologic and ultrastructural techniques, Hematol. Pathol. 9 (1995) 79–94. [10] N. Imamura, Sensitive detection technique of myeloperoxidase precursor protein by flow cytometry with monoclonal antibodies, Am. J. Hematol. 58 (1998) 241–243. [11] K. Nakase, M. Sartor, K. Bradstock, Detection of myeloperoxidase by flow cytometry in acute leukemia, Cytometry 34 (1998) 198–202. [12] E. J. Manaloor, R. S. Neiman, D. K. Heilman, M. Albitar, T. Casey, T. Vattuone, P. Kotylo, A. Orazi, Immunohistochemistry can be used to subtype acute myeloid leukemia in routinely processed bone marrow biopsy specimens, Am. J. Clin. Pathol. 113 (2000) 814-822. [13] C.H. Dunphy, J.M. Polski, H.L. Evans, L.J. Gardner, Evaluation of bone marroe specimens with acute myelogenous leukemia for CD34, CD15, CD117, and myeloperoxidase, Arch. Pathol. Lab. Med. 125 (2001) 1063–1069. [14] L. Saravanan, S. Juneja, Immunohistochemistry is a more sensitive marker for the detection of myeloperoxidase in acute myeloid leukemia compared with flow cytometry and cytochemistry, Int. J. Lab. Hematol. 32 (2010) 132–136. [15] H. Tong, C. Lu, J. Zhang, Z. Liu, Y. Ma, Immunophenotypic, cytogenetic and clinical features of 192 AML patients in China, Clin. Exp. Med. 9 (2009) 149–155. [16] A. Wiik, Clinical and pathophysiological significance of anti-neutrophil cytoplasmic autoantibodies in vasculitis syndromes, Mod. Rheumatol. 19 (2009) 590–599. [17] P. Guilpain, A. Servettaz, F. Batteux, L. Guillevin, L. Mouthon, Natural and disease associated anti-myeloperoxidase (MPO) autoantibodies, Autoimmun. Rev. 7 (2008) 421–425. [18] T.J. Fiedler, C.A. Davey, R.E. Fenna, X-ray crystal structure and characterization of halid-binding sites of human Myeloperoxidase at 1.8 Å resolution, J. Biol. Chem. 275 (2000) 11964–11971. [19] T. Liu, W. Qian, M.A. Gritsenko, D.G. Camp, M.E. Monroe, R.J. Moore, R.D. Smith, Human plasma N-glycoproteome analysis by immunoaffinity subtraction, hydrazide chemistry, and mass spectrometry, J. Proteome Res. 4 (2005) 2070–2080. [20] M. Hansson, I. Olsson, W.M. Nauseef, Biosynthesis, processing, and sorting of human myeloperoxidase, Arch. Biochem. Biophys. 445 (2006) 214–224. [21] P. Ramachandran, P. Boontheung, Y. Xie, M. Sondej, D.T. Wong, J.A. Loo, Identification of N-linked glycoproteins in human saliva by glycoprotein capture and mass spectrometry, J. Proteome Res. 5 (2006) 1493–1503. [22] M. Goedken, S. McCormick, K.G. Leidal, K. Suzuki, Y. Kameoka, J.M. Astern, M. Huang, A. Cherkasov, W.M. Nauseef, Impact of two novel mutations on the structure and function of human myeloperoxidase, J. Biol. Chem. 282 (2007) 27994–28003. [23] X. Carpena, P. Vidossich, K. Schroettner, B.M. Calisto, S. Banerjee, J. Stampler, M. Soudi, P.G. Furtmüller, C. Rovira, I. Fita, C. Obinger, Essential role of proximal histidine–asparagine interaction in mammalian peroxidases, J. Biol. Chem. 284 (2009) 25929–25937. [24] P. Van Antwerpen, M. Slomianny, K. Z. Boudjeltia, C. Delporte, V. Faid, D. Calay, A. Rousseau, N. Moguilevsky, M. Raes, L. Vanhamme, P. G. Furtmuller, C. Obinger, M. Vanhaeverbeek, J. Néve, J. Michalski, Glycosylation pattern of mature dimeric

T. Ravnsborg et al. / Biochimica et Biophysica Acta 1804 (2010) 2046–2053

[25]

[26]

[27]

[28] [29]

[30] [31]

leucocyte and recombinant monomeric myeloperoxidase: glycosylation is required for optimal enzymatic activity, J. Biol. Chem. 285 (2010) 16351–16359. U. Kuzmanov, N. Jiang, C.R. Smith, A. Soosaipillai, E.P. Diamandis, Differential Nglycosylation of kallikrein 6 derived from ovarian cancer cells or the central nervous system, Mol. Cell. Proteomics 8.4 (2008) 791–798. G. Tabarés, C.M. Radcliffe, S. Barrabés, M. Ramírez, R.N. Aleixandre, W. Hoesel, R.A. Dwek, P.M. Rudd, R. Peracaula, R. de Llorens, Different glycan structures in prostate-specific antigene from prostate cancer sera in relation to seminal plasma PSA, Glycobiology 16 (2006) 132–145. G.S. Leiserowitz, C. Lebrilla, S. Miyamoto, H.J. An, H. Duong, C. Kirmiz, B. Li, H. Liu, K.S. Lam, Glycomics analysis of serum: a potential new biomarker for ovarian cancer? Int. J. Gynecol. Cancer 18 (2008) 470–475. E. Andersson, L. Hellman, U. Gullberg, I. Olsson, The role of the propeptide for processing and sorting of human myeloperoxidase, J. Biol. Chem. 273 (1998) 4747–4753. N.H. Heegaard, D.R. Jakobsen, D. Klattschou, Purification of Wegener's granulomatosis autoantigen, proteinase 3, from neutrophils by Triton X-114 extraction of azurophilic granules, Anal. Biochem. 253 (1997) 259–262. I. Olsson, P. Venge, Cationic proteins of human granulocytes, Scand. J. Haematol. 9 (1972) 204–214. M.R. Larsen, S.J. Cordwell, P. Roepstorff, Graphite powder as an alternative or supplement to reversed-phase material for desalting and concentration of peptide

[32]

[33]

[34] [35]

[36]

[37]

2053

mixtures prior to matrix-assisted laser desorption/ionization-mass spectrometry, Proteomics 2 (2002) 1277–1287. M.A.P. Audrain, T.A.R. Baranger, N. Moguilevski, S.J. Martin, A. Devys, C.M. Lockwood, J.Y. Muller, V.L.M. Esnault, Anti-native and recombinant myeloperoxidase monoclonals and human autoantibodies, Clin. Exp. Immunol. 107 (1997) 127–134. A.K. Short, C.M. Lockwood, Studies of epitope restriction on myeloperoxidase (MPO), an important antigen in systemic vasculitis, Clin. Exp. Immunol. 110 (1997) 270–276. L. Chang, S. Binos, J. Savige, Epitope mapping of anti-proteinase 3 and antimyeloperoxidase antibodies, Clin. Exp. Immunol. 102 (1995) 112–119. U. Erdbrügger, T. Hellmark, D. O. Bunch, D. A. Alcorta, J. C. Jennette, R. J. Falk, P. H. Nachman, Mapping of myeloperoxidase epitopes recognized by MPO-ANCA using human-mouse MPO chimers, Kidney Int. 69 (2006) 1799-1805. M.R. Larsen, S.S. Jensen, L.A. Jakobsen, N.H.H. Heegaard, Exploring the sialiome using titanium dioxide chromatography and mass spectrometry, Mol. Cell. Proteomics 6.10 (2007) 1778–1787. M. Thaysen-Andersen, I.B. Thøgersen, H.J. Nielsen, U. Lademann, N. Brünner, J. Enghild, P. Højrup, Rapid and individual-specific glycoprofiling of the low abundance N-glycosylated protein tissue inhibitor of metalloproteinases-1, Mol. Cell. Proteomics 6.4 (2007) 638–647.