Water Ice is a Soft Matrix for the Structural ... - ACS Publications

Article pubs.acs.org/ac

Water Ice is a Soft Matrix for the Structural Characterization of Glycosaminoglycans by Infrared Matrix-Assisted Laser Desorption/ Ionization Lukas Witt,† Alexander Pirkl,† Felix Draude,†,∥ Jasna Peter-Katalinić,‡,§ Klaus Dreisewerd,† and Michael Mormann*,† †

Institute for Hygiene, University of Münster, 48149 Münster, Germany Department of Biotechnology, University of Rijeka, 51000 Rijeka, Croatia § Institute of Medical Physics and Biophysics, University of Münster, 48149 Münster, Germany ‡

S Supporting Information *

ABSTRACT: Glycosaminoglycans (GAGs) are a class of heterogeneous, often highly sulfated glycans that form linear chains consisting of up to 100 monosaccharide building blocks and more. GAGs are ubiquitous constituents of connective tissue, cartilage, and the extracellular matrix, where they have key functions in many important biological processes. For their characterization by mass spectrometry (MS) and tandem MS, the high molecular weight polymers are usually enzymatically digested to oligomers with a low degree of polymerization (dp), typically disaccharides. However, owing to their lability elimination of sulfate groups upon desorption/ionization is often encountered leading to a loss of information on the analyte. Here, we demonstrate that, in particular, water ice constitutes an extremely mild matrix for the analysis of highly sulfated GAG disaccharides by infrared matrix-assisted laser desorption/ionization (IR-MALDI) mass spectrometry. Depending on the degree of sulfation, next to the singly charged ionic species doubly- and even triply charged ions are formed. An unambiguous assignment of the sulfation sites becomes possible by subjecting sodium adducts of the GAGs to low-energy collision-induced dissociation tandem MS. These ionic species exhibit a remarkable stability of the sulfate substituents, allowing the formation of fragment ions retaining their sulfation that arise from either cross-ring cleavages or rupture of the glycosidic bonds, thereby allowing an unambiguous assignment of the sulfation sites.

P

GAGs poses a great challenge not only because of their structural diversity and high negative charge density but also as a consequence of their molecular size. For comprehensive characterization most experimental protocols, therefore, include the depolymerization of the polysaccharide chains into smaller subunits, especially disaccharides. This can be achieved either enzymatically or chemically. The latter approach involves exposition to a solution of nitrous acid, which can be prepared either from sulfuric acid and barium nitrite (pH 1.5) for cleavage of N-sulfated GAGs or from sulfuric acid and sodium nitrite (pH 4.0) for the depolymerization of N-acetylated GAGs.7 The enzymatic approach involves different polysaccharide lyases, depending on the structure of the substrate involved. These enzymes cleave specifically glycosidic linkages by elimination. This gives rise to depolymerization and yields unsaturated oligomers that harbor a double bond between C4 and C5 at the nonreducing end (i.e., the hexuronic acid).5 The resulting disaccharides or higher oligomers can, for example, be

roteoglycans (PGs) are ubiquitous constituents of connective tissue, the surface of mammalian cells, and the extracellular matrix. They play key roles in many biological processes such as morphogenesis, cell growth, and differentiation.1−4 PGs are composed of a core protein and covalently linked glycosaminoglycan (GAG) chains. GAGs are linear polysaccharides that consist of repeating disaccharide building blocks. A distinction is drawn between the following major types: chondroitin sulfate, dermatan sulfate, heparin(Hep)/heparan sulfate(HS), hyaluronan, and keratan sulfate. Except for hyaluronan, all members of the GAG family are sulfated to some extent. Heparin and heparan sulfate represent the most highly sulfated GAG species. Their disaccharide backbone is composed of iduronic acid (IdoA) or glucuronic acid (GlcA) and glucosamine (GlcN), which are linked 4HexAβ1−4GlcNα1. Typically, sulfation positions are found at C2 and/or C6 of the GlcN-residue and at C2 of the uronic acid residue. Furthermore, an N-acetylation instead of Nsulfation is possible.5,6 Due to their tremendous importance in many biological systems, a detailed characterization of GAGs is a prerequisite to understand their structure−function relationships. Analysis of © 2014 American Chemical Society

Received: March 7, 2014 Accepted: May 26, 2014 Published: May 26, 2014 6439

dx.doi.org/10.1021/ac5008706 | Anal. Chem. 2014, 86, 6439−6446

Analytical Chemistry

Article

characterized by application of electrophoretic8,9 or chromatographic techniques.10,11 During the past two decades, mass spectrometry (MS)-based methods have emerged as suitable tools for comprehensive GAG analysis.12−14 First approaches to analyze both underivatized as well as derivatized GAGs by use of mass spectrometry (MS) were based on formation of gas-phase ions by fast atom bombardment (FAB).15,16 In particular, elimination of the labile sulfate groups is a problem frequently encountered, hampering the distinction of intact analyte ions from fragmentation products. The advent of the soft ionization techniques matrix-assisted laser desorption/ionization (MALDI) and electrospray ionization (ESI) allowed for the MS analysis of intact GAGs as unintended loss of the labile sulfate groups during the desorption/ionization process is significantly reduced. Particularly, ESI is employed for the formation of intact gaseous GAG-derived analyte ions, often leading to the formation of multiply charged species which are less prone to sulfate elimination. There has been a number of different approaches, employing this technique including chipbased methodologies and desorption electrospray ionization (DESI) for the analysis of sulfated saccharides.12,14,17−20 Furthermore, ESI allows for a facile hyphenation with chromatographic separation systems enabling online analysis of separated compounds from complex mixtures.13,21−23 For structural characterization of the heterogeneous GAG family members, typically low-energy collision-induced dissociation (CID) is employed.24,25 Cleavages of glycosidic linkages yield fragment ions that allow sequence analysis while cross-ring cleavages deliver structural information, especially on the linkage site between two monosaccharide building blocks. As the energy required to induce formation of cross-ring fragments usually exceeds the activation barrier for SO3 elimination, stabilization of the labile sulfate groups is a prerequisite for assignment of their position. An enhanced stability of the sulfate groups is either observed in multiply charged precursor ions or in ionic species complexed by mono- or divalent cations.25,26 For example, a combination of these techniques has been applied for a complete structural elucidation of Arixtra, the active pentasaccharide epitope involved in the anticoagulative action of low molecular heparin.27 Recent findings show that also electron detachment dissociation (EDD) as well as negative electron transfer dissociation (NETD) are viable techniques to obtain structural information on GAG analytes.28,29 A benefit of these methods is the possible differentiation between the GlcA and IdoA epimers. Several attempts have been made to use UV-MALDI as a method for formation of gaseous GAG-derived ions. It was shown that codesorption of the highly acidic analytes with basic peptides or ion pairing reagents leads to stable ionic noncovalent complexes.30−32 Also, the use of crystalline UV matrices, such as 2,5-dihydroxybenzoic acid,32−35 norharmane,32,36 or α-cyano-4-hydroxycinnamic acid,32 was reported for mass spectrometric studies on sulfated GAG analytes. However, compared to ESI-MS, elimination of the labile sulfate groups is generally occurring, as a result of the relatively large amount of internal excess energy deposited into the analyte ions upon desorption and/or ionization. As an alternative, the applicability of matrices providing softer desorption/ionization under UV-MALDI conditions, viz. liquid matrices (e.g., mixtures of 1,1,3,3-tetramethylguanidine and p-coumaric acid), has been evaluated in several studies. In fact, these demonstrated a considerable improvement with respect to

formation of stable GAG-derived ions inside the ion source.36−39 Compared to UV-MALDI, the use of IR lasers generally provides softer desorption/ionization conditions.40−42 The main difference between the two MALDI regimes lies in the differing energy deposition into the matrix (i.e., electronic excitation by UV photons in a shallow matrix volume and vibrational excitation of a larger volume by IR photons). As a consequence, the material ejection processes vary significantly between the two excitation modes.40−42 The typical emission wavelength used in IR-MALDI of λ = 2.94 μm, provided by either erbium-doped yttrium aluminum garnet (Er:YAG) lasers or by optical parametric oscillator (OPO), systems matches well with the O−H and N−H stretch vibrations of suitable matrices. For example, this allows employing glycerol as IRMALDI matrix. This matrix is vacuum stable at room temperature (RT), provides a near physiological pH and, importantly, a homogeneous sample morphology. IR-MALDIMS with a glycerol matrix used at RT was, for example, applied for the MS analysis of labile analytes such as oligosaccharides,43 glycolipids, 44 double-stranded DNA, 45 and noncovalent protein−carbohydrate complexes.46 In a recent study, Tajiri and Wada have shown that also solid urea can be used for the characterization of labile glycoconjugates. In this case, desorption/ionization was achieved by activation of the carbonyl stretching mode of the matrix with IR photons of a wavelength of λ = 6 μm.47 Gaseous ganglioside- and GAGderived ions were formed in the negative ion mode and loss of the labile sialic acids or sulfate groups, respectively, was observed only to a minor extent. Particularly interesting is the use of water as an IR-MALDI matrix that is providing ideal “near-physiological” conditions. However, due to the high vapor pressure of water, a cooling stage is generally required, unless experiments were performed with an atmospheric pressure (AP) instrument. Recently, the development of a temperature-controlled IR-MALDI sample stage for an orthogonal extracting time-of-flight mass spectrometer (o-TOF-MS) has been described. Notably, desorption/ionization from the water ice matrix at approximately −86 °C and ∼1 mbar of N2 buffer gas pressure resulted in elevated abundances of multiply charged analyte ions.48 Here, we report on the use of glycerol and water ice as IRMALDI matrices for the structural characterization of labile glycosaminoglycans. The IR-MALDI o-TOF mass spectra exhibit the formation of intact deprotonated GAG species in different charge states, depending on the degree of sulfation, which is generally accompanied by only a minor loss of sulfate moieties. Collisional activation of sodium adducts of multiply deprotonated GAG analytes in tandem-MS experiments leads to the formation of fragment ions that arise from either crossring cleavages or rupture of the glycosidic bonds. These diagnostic ions allow for an unambiguous assignment of the sulfation sites.

■

EXPERIMENTAL METHODS Materials. Glycosaminoglycan disaccharides (unsaturated disaccharides obtained from enzymatic cleavage of porcine heparin) were obtained from Iduron (Manchester, U.K.). Heparin was purchased from Acros Organics (Geel, Belgium), and heparinases I, II, and III were from AMS Biotechnology (Abingdon, U.K.). Chondroitin polysaccharide, chondroitinase ABC, mass spectrometry-grade glycerol, all solvents (LC−MS grade) and all reagents used (analytical grade, p.a.) were purchased from Sigma-Aldrich (Seelze, Germany). 6440



Article

In-Solution Digestion of Chondroitin Sulfate and Heparin. Chondroitin sulfate (50 μg) was dissolved in 100 μL digestion buffer composed of 50 mM Tris/HCl buffer, pH 8, 60 mM sodium acetate, and 0.02% bovine serum albumin. To this solution 250 mU of chondroitinase ABC were added, and the reaction mixture was incubated overnight at 37 °C. Heparin (50 μg) was digested by use of 2.5 mU heparinase I, II, and III in a final volume of 100 μL of 0.1 M sodium acetate, pH 7.0, supplemented with calcium acetate (10 mM) overnight at 37 °C. The reactions were terminated by heating for 5 min at 95 °C. Samples were desalted by solid phase extraction using Hypercarb pipet tips (Thermo Scientific, Bellefonte, PA). Briefly, tips were equilibrated five times with 20 μL of releasing solution (40% ACN, 0.1% FA) and three times with 20 μL of binding solution (0.1% FA). Vacuum-dried digests were reconstituted in 20 μL of binding solution and loaded onto the tips by aspirating and expelling 50 times. After washing the samples with 10× 20 μL of binding solution, the disaccharides were eluted by rinsing with 20 μL of releasing solution five times. The eluates were vacuum-dried and stored at −20 °C. Infrared MALDI-o-TOF-MS. IR-MALDI MS experiments were performed by use of an o-TOF mass spectrometer (QSTAR pulsar i, AB SCIEX, Concord, Ontario, Canada), as described previously.48 An Er:YAG laser (Biooptic Laser Systems, Berlin, Germany; λ = 2.94 μm, τ ∼ 150 ns, spot size after focusing: 110 × 220 μm2, pulse repetition rate: 2 Hz) was used for desorption/ionization. Ions were generated at a background pressure in the ion source of 1−2 mbar (N2). For each mass spectrum acquired, 360 laser shots were accumulated. The mass window was set to m/z 100−1000. Matrices used in this study were glycerol and water ice, in both cases without any additional modifiers or supplements. Glycerol preparations were obtained by mixing aqueous analyte solutions 1:1 (v/v) with glycerol and removal of water in vacuo followed by application of 0.3 μL on the sample plate. In the case of RT experiments, a standard stainless-steel sample plate was used that mounted in the default target holder of the oMALDI2 ion source (AB SCIEX). For low-temperature experiments, an inhouse-built cooled sample stage was replacing the standard target holder and maintained at a constant sample temperature of approximately −80 °C during experiments.48 For analysis with water ice as matrix aqueous samples were directly applied onto an infrasil glass slide and immediately dipped into liquid nitrogen to freeze the sample spot. Similarly, the desiccated glycerol samples were prepared for low-temperature experiments.The glass slide was fixed onto the backside of a Peltier element using heat sink paste (WLP 035, Bürklin, Oberhaching, Germany). MS/MS experiments were performed using argon as collision gas. All experiments were performed in the negative ion mode. For data evaluation Analyst Software (AB SCIEX) was used. GAG-derived fragment ions are labeled according to the nomenclature of Domon and Costello.49

Figure 1. IR-MALDI-o-TOF mass spectra of a trisulfated disaccharide standard analyte (D2S6, ΔUA2S−GlcNS6S) employing (a) a glycerol matrix at RT, (b) a frozen glycerol matrix (Tsample: −80 °C), and (c) a water ice matrix (Tsample: −80 °C). The amount applied on the target was 33 pmol in each case, the spectra represent 180 laser shots at a fluence of approximately 4000 J/m2.

ions, formed by deprotonation and complexation with different amounts of sodium cations ([M − 2H + Na]−, [M − 3H + 2Na]−, and [M − 4H + 3Na]−). However, a considerable loss of sulfate groups giving rise to ionic species with m/z 539.90, 517.91, 437.97, and 416.00 (according to the elimination of either one or two sulfate groups) is found for all species. Furthermore, interpretation of the spectra is complicated as a result of the extensive formation of adducts between the GAGs and different matrix-derived ion species. Similar adduct formation has been noted previously in IR-MALDI-MS analyses with glycerol.46 Due to the presence of abundant matrix cluster ions, the spectrum also shows a significant level of chemical background noise. Lowering the sample temperature to −80 °C (cf., Figure 1b) led to a sizable improvement in the quality of the mass spectra. In particular, a significant reduction of sulfate elimination is observed. However, matrix adducts of relatively high abundance remain present in the spectrum, even though the overall chemical noise level decreases. Another significant difference observed in these experiments is the generation of doubly charged ions in the mass range of m/z 285−310, which were not detected when spectra were acquired at RT. The influence of collisional cooling on the overall ion formation was probed by increasing the source pressure from 1 mbar to 2.5 mbar to further suppress thermal fragmentation processes.50 Spectra

■

RESULTS AND DISCUSSION Glycerol and Water Ice As Soft Matrices for the Analysis of Heparan Sulfate-Derived Disaccharides by IR-MALDI Mass Spectrometry. Negative ion mode IRMALDI-o-TOF mass spectra of the triply sulfated heparinderived analyte D2S6 (ΔUA2S−GlcNS6S), obtained by using glycerol at RT, glycerol at −80 °C, and water ice at −80 °C, are shown in Figure 1 (panels a−c), respectively. Evaluation of the mass spectrum obtained by use of glycerol at RT (Figure 1a) reveals the presence of abundant singly charged intact analyte 6441



Article

recorded under these conditions exhibited nearly no loss of sulfate groups, albeit at the expense of a further increased formation of analyte−glycerol adducts (cf., Figure S-1 of the Supporting Information). In contrast, the mass spectrum acquired with the water ice matrix (cf. Figure 1c) does not show any ionic analyte−matrix adducts at both pressure settings. Moreover, it is essentially devoid of matrix-derived chemical background. Analyte ions detected with the highest abundance correspond to the singly and doubly charged sodiated disaccharide ions (e.g., [M − 2H + Na]−, [M − 3H + 2Na]−, and [M − 3H + Na]2−). In addition, even triply charged ions ([M − 3H]3−) are detected. Compared to the spectra obtained with solid glycerol, the loss of sulfate groups is sizably reduced. Similar findings were observed for all other sulfated GAG disaccharides examined (data not shown). In addition, an efficient complexation with sodium, which is observable in all three spectra, the initial sample temperature as well as the choice of the matrix obviously have a notable influence on the stability of the analyte ions. Further stabilization is obtained by improving the efficiency of collisional cooling. While only a relatively low amount of multiply charged ions was produced from the frozen glycerol, the use of water ice as a matrix gave rise to a significantly higher abundance of doubly charged species as compared to singly charged ions. Structural Elucidation of Heparan Sulfate-Derived Disaccharides by IR−MALDI-MS/MS. Recently, Kailemia et al. demonstrated that the use of NaOH as a dopant is advantageous for the structural characterization of GAG precursor ions formed under ESI conditions. Complexation of deprotonated GAGs with sodium ions increases the stability of the analyte ions and facilitates their structural characterization by low-energy CID tandem MS.27 Making use of the highly abundant GAG/Na+-adducts observed in the IR-MALDI mass spectra with the water ice matrix, we next probed structural characterization of GAG-derived analyte ions under the current IR-MALDI o-TOF-MS conditions. In Figure 2, CID spectra of singly charged precursor ions derived from the disaccharide D2S6 (ΔUA2S−GlcNS6S) that are exhibiting a different degree of cationization are compared exemplarily. The CID spectrum of the singly deprotonated species [M − H]− does not provide sufficient information on the analyte structure as only few fragment ion species formed by sulfate loss appear (Figure 2a). With increasing degree of sodium adduct formation, the CID spectra become more information-rich, since unintended sulfate elimination decreases (cf., Figure 2, panels b and c). Best results were obtained for 4-fold deprotonated analytes harboring three sodium ions thereby exhibiting an overall charge of one viz. [M − 4H + 3Na]− (Figure 2d). Under these conditions, only a very low degree of sulfate loss is observed. Instead, a variety of product ions, which were formed by fragmentation of the glycosidic bonds and by cross-ring cleavages under preservation of the sulfate groups, are detected. On the basis of these data, an unambiguous structural assignment of the analytes becomes feasible (vide infra). Thus, best CID results are obtained when the fully sodiated form of the analyte ions, where the number of sulfate groups is equal to the number of sodium ions, is chosen as precursor ion. Similar findings were made when GAG-derived ions of higher charge states were submitted to collisional activation. The MS/MS spectra of doubly charged precursor ions formed from the trisulfated HS disaccharide D2S6 (ΔUA2S−GlcNS6S) (i.e., [M − 2H]2−, [M − 3H + Na]2−,

Figure 2. MS/MS spectra of singly charged precursor ions from trisulfated HS disaccharide D2S6 (ΔUA2S−GlcNS6S, inset) with an increasing degree of Na+-cationization. Precursor ions: (a) [D2S6 − H]−, (b) [D2S6−2H + Na]−, (c) [D2S6−3H + 2Na]−, and (d) [D2S6−4H + 3Na]−. Spectra were recorded by application of 180 laser pulses at a fluence of ∼4000 J/m2, argon served as the collision gas, and Elab was set to 28 eV. The peak labeled with an asterisk corresponds to the [0,2A2 − SO3 − 2 H2O + Na]− ions; ×5 indicates a 5-fold magnification of the absolute intensity scale.

and [M − 4H + 2Na]2−) exhibit an increased stability of the labile sulfate groups with increasing number of associated Na+ ions during CID. Fragmentation of glycosidic bonds as well as cross-ring cleavages can compete with SO3 elimination and yield sulfated fragment ions, which allow for structural assignment (cf. Figure S-2 of the Supporting Information). Evaluation of CID spectra of triply charged GAG ions leads to similar conclusions (cf., Figure S-3 of the Supporting Information). However, since the precursor ion abundance in the latter experiments was relatively low, only a limited number of product ions was detected, restricting the structural information derived from these spectra. Figure 3 depicts MS/MS spectra of trisulfated, disulfated, and monosulfated HS disaccharides. The sulfate groups of D2S6 (ΔUA2S−GlcNS6S) precursor ions ([M − 4H + 3Na]−, m/z 641.90) remained remarkably stable during the CID process and only a minor loss of sulfate was observed. Besides 6442



Article

Figure 3. MS/MS spectra and fragmentation schemes of (a) D2S6, ΔUA2S−GlcNS6S, [M − 4H + 3Na]−, (b) D2S0, ΔUA2S−GlcNS, [M − 3H + 2Na]−, and (c) D0S0, ΔUA−GlcNS, and [M − 2H + Na]− at a collision energy of 28 eV (Elab). The peak labeled with an asterisk in (a) corresponds to the [2,5A2 − SO3 + Na]− ions.

at the C6-position. The signal at m/z 138.97 is ambiguous: it can be assigned either to the formation of 0,4X0 ions, further substantiating the presence of a sulfate group at the C6-position of the reducing end, or it can be annotated as 1,3A1 ions, showing the presence of a sulfation site at the C2-position at the nonreducing end. Additional fragment ions at m/z 480.93 (0,2A2) and m/z 318.97 (2,4A2) verify the sulfation at the C6position of the sugar at the reducing end and an additional third sulfation at the residue of the nonreducing end. This sulfation site is determined by the presence of the product ions with m/z 503.90 (0,2X1) pointing to the C2-position of the uronic acid residue as the only possible position. No evidence for sulfate

glycosidic bond cleavages as shown by the formation of the complementary fragment ion pairs Z1 (m/z 323.94, accompanied by the loss of water) and C1 (m/z 276.96) ions as well as Y1 (m/z 359.97) and B1 ions (m/z 258.95), the majority of product ions is corresponding to cross-ring cleavages. Most of them are found in the pyranose, comprising the reducing end. With the exception of two 0,2A2-ions of minor abundance exhibiting loss of sulfate (m/z 400.92/385.92), all cross-ring cleavage ions indicate an intact sulfation pattern. The 0,2X0-ions (m/z 137.99) imply a sulfation at the amino group corresponding with the observation of the 2,4X0 ions at m/z 299.95, which also corroborates the presence of a sulfate group 6443



Article

patterns possess a structural diversity, which is not easy to decipher. In this context, the successful application of water ice as a matrix in IR−MALDI-MS for sulfated standard disaccharides was expanded to the analysis of samples obtained from heparin and chondroitin sulfate. Prior to mass spectrometric experiments, polymeric samples were enzymatically treated by heparinases I, II, and III and chondroitinase ABC, respectively. Representative mass spectra recorded for Hep and CS digests are shown in Figure 4, together with a

migration processes preceding fragmentation as observed previously in the ESI-MS/MS spectra of singly charged sulfated oligosaccharides was found.51 For the disulfated analyte D2S0 (ΔUA2S−GlcNS), the ions at m/z 539.96 ([M − 3H + 2Na]−) were selected as precursor ions (cf., Figure 3b). Although there is a minor peak corresponding to NaSO4− at m/z 118.94, no fragment ions pointing to a SO3 elimination can be observed. The most prominent ions were observed at the m/z value 318.98 (2,4A2). Together with the ions at m/z 138.97 (1,3A1), the sulfation site of the residue located at the nonreducing end could be determined as C2. The observation of the 0,2X0 ion at m/z 137.99 led to the conclusion that the glucosamine ring is sulfated at the N-position, which is further corroborated by the appearance of 0,4A2-ions at m/z 479.95. Additionally, 0,2A2-ions (m/z 360.99), 0,2X1-ions (m/z 401.98), and 0,2A1-ions (m/z 115.01) were found. Glycosidic bond cleavages could be observed as well, yet the corresponding fragment ions were only of minor abundance (B1, m/z 258.96; C1, m/z 276.96; Z1−H2O, m/z 222.01). Similar results were achieved with a monosulfated analyte (D0S0, ΔUA−GlcNS, m/z 438.04, [M − 2H + Na]−; cf., Figure 3c). The glycosidic bond cleavage again yielded only weak signals; nevertheless, all complementary fragment ions could be detected: B1, m/z 157.02; C1, m/z 175.04; Y1, m/z 258.03; Z1, m/z 240.04, demonstrating the reducing end pyranose as the sulfated residue. While the 2,4X0ions (m/z 219.99) as well as the ions at m/z 334.03 and 378.02 (0,3X1/ 0,4A2) do not allow an exact determination of the sulfate position, the ions at m/z 137.99 (0,2X0), however, indicate a sulfation of the glucosamine ring at the N-position. The analysis of other disaccharide molecules available revealed similar results; a synopsis of observed fragment ions of all investigated analytes is given in Table S-1 of the Supporting Information. The favored complexation of deprotonated GAG analytes with alkali cations during MALDI analysis with the water ice matrix represents a benefit over the widespread electrospray ionization as a complexation with alkali ions is a known prerequisite for successful structural elucidation by CID experiments, as recent publications on this topic reveal.27,52 Though this was already noted in the past,12,25,53 for a long time, multiply charged precursor ions were in the focus as optimal requirement for structural elucidation of GAG analytes. This assumption is obvious as a higher charge state stabilizes the labile sulfate groups in the negative ion mode.26 However, recent developments show that, given a sufficient amount of alkali cations is present in the analyte solution, a full complexation of the sulfate positions with, for example, sodium is easily achieved, which also leads to a more stable sulfation. As described by Wolff et al.,54 for electron detachment and IR multiphoton dissociation, a combination of high charge state and complexation by alkali ions promises best results in fragmentation experiments. Also some of the most recent studies on structural characterization of GAGs describe a twoway approach, in which multiply charged precursor ions are complexed with sodium ions to yield the most informative fragmentation patterns.27,55 However, this technique, in contrast to IR-MALDI with water ice, demands the supplement of relatively large amounts of NaOH to accomplish a successful adduct formation. Application to Samples Obtained from Enzymatic Cleavage of GAG Polymers. The large structural heterogeneity of GAG analytes poses a challenging analytical problem. Polymeric carbohydrate chains exhibiting different sulfation

Figure 4. Comparison of IR-MALDI-o-TOF mass spectra of (a) a chondroitin sulfate digest, (b) a heparan sulfate digest, and (c) an artificial mixture of 12 differently substituted HS disaccharides (6 pmol per analyte species). For the digests, the amount applied to the MALDI target correspond to 1.6 μg of raw polysaccharide. Labeled peaks in (c) correspond to 1[ΔGlcA-GlcN(2S) − 2H]2− and 2 [ΔGlcA,2S-GlcNS,6S − 3H + Na]2−. The number of sulfate groups is given in parentheses for ions with an unknown sulfation pattern.

spectrum obtained from a mixture of disaccharide standards. On the basis of these data, information on the disaccharide composition of the polymers could be extracted. For CS, singly as well as doubly charged ions are detected for monosulfated and acetylated species (m/z 228.55/458.06) and disulfated/ acetylated analytes (m/z 268.52/538.01). Additionally, less intense signals are observed for multiply deprotonated ions being partially neutralized by sodium as well as potassium cations. Signals corresponding to glycosidic bond cleavages also occur to some extent, viz., Z1-ions (m/z 282.04). In general, it has to be noted that monosulfated ionic species account for the major signals, while only weak signals correspond to disulfated analytes. Although this might partly be due to the elimination of sulfate groups, first information on the main constituents is achieved (Figure 4a). 6444



Article

controlled sample stage allowed the use of this matrix compound for the analysis of labile GAG analytes at −80 °C. The negative ion mode mass spectra revealed the presence of intact deprotonated GAG species with only minor extent of sulfate elimination. Depending on the degree of sulfation, doubly- and even triply charged disaccharide ions were observed. Furthermore, the application of water ice as a matrix provided spectra of very low chemical background noise facilitating the interpretation of data. Fragment ions giving rise to improved structural information were obtained from sodium adducts of GAG analytes. These ionic species exhibit a remarkable stabilization of the labile sulfate groups, allowing the formation of fragment ions that arise from either cross-ring cleavages or rupture of the glycosidic bonds and allowed an unambiguous assignment of the sulfation sites. However, so far this technique has only been adopted to the analysis of disaccharides. Thus, it is possible to predict structural information on exhaustively enzymatically treated GAG samples leading to dimers. In the future, emphasis should be placed on further optimization to enable structural characterization of higher oligomers. To conclude, the methodology presented in this study offers novel approaches for the compositional as well as structural analysis of highly complex glycosaminoglycans and possesses the potential to be embedded (e.g., for screening purposes of biological samples).

The heparin-derived sample exhibited an overall higher heterogeneity (Figure 4b). Doubly charged ions were observed for disulfated as well as for one trisulfated ionic species (m/z 247.52 and 287.50). The presence of disulfated and trisulfated analytes was also highlighted by the detection of singly charged deprotonated as well as potassiated ions. Other singly charged ions were generated by ionization of unsulfated/acetylated species (m/z 378.12), monosulfated ions (m/z 416.06), monosulfated/acetylated analytes (m/z 458.07) as well as disulfated ions (m/z 496.01). Analysis of an artificial mixture of 12 different disaccharide analytes produced comparable results, apart from exhibiting some additional signals in the IR-MALDI mass spectra, that are mainly derived from sodium adducts as well as a signal for a disulfated/acetylated species (D2A6, m/z 559.93). The overall signal intensity is comparable for the heparin digest and the artificial mixture while the CS digest results in an approximately doubled signal intensity, most probably due to a less heterogeneous composition of the sample. Since an exact prediction on the structure of the analytes is not possible by simple MS analysis, tandem MS experiments were performed on selected precursor ions. Exemplarily, this approach is shown in Figure 5 for precursor ions derived from a

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as described in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel. +49-251-8358440. Fax: +49-251-8359956. Present Address ∥

Physikalisches Institut, University of Münster, 48149 Münster, Germany.

Figure 5. MS/MS spectrum of monosulfated and acetylated disaccharide ions ([M − 2H + Na]−) derived from a chondroitin sulfate digestion at m/z 480.04 and fragmentation schemes of deduced structures.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support by the German Research Foundation DFG (International Research Training Group 1549 Molecular and Cellular Glyco-Sciences MCGS, Münster/Hyderabad to M.M. and Grant DR 416/10-1 to K.D.) is acknowledged.

digestion of polymeric chondroitin sulfate with chondroitinase ABC at m/z 480.04, representing the sodiated form of a monosulfated/acetylated molecule. By CID experiments, structural motifs can be elucidated. Besides glycosidic bond cleavages expressed by the occurrence of Z1-ions and the respective sodiated congeners (m/z 282.04 and 304.03), also fragment ions arising from cross-ring cleavages are detected. While the 0,4X0 ions (m/z 138.99) indicate the sulfation at the C6-position of the GlcNAc residue, which is further underlined by the formation of 2,4X0-ions (m/z 244.01), sodiated 0,4A2-ions (m/z 420.06) indicate a sulfation at C4 of the same ring. These results unambiguously demonstrate the presence of at least two isomeric monosulfated/acetylated species [i.e., ΔUA-GalN(6S) and ΔUA-GalN(4S) in the digestion mixture].

■

REFERENCES

(1) Taylor, K. R.; Gallo, R. L. FASEB J. 2006, 20, 9−22. (2) Kleinman, H. K.; Philp, D.; Hoffman, M. P. Curr. Opin. Biotechnol. 2003, 5, 526−532. (3) Raman, R.; Sasisekharan, V.; Sasisekharan, R. Chem. Biol. 2005, 3, 267−277. (4) Morgan, M. R.; Humphries, M. J.; Bass, M. D. Nat. Rev. Mol. Cell Biol. 2007, 8, 957−969. (5) Esko, J. D.; Kimata, K.; Lindahl, U. Proteoglycans and Sulfated Glycosaminoglycans. In Essentials of Glycobiology, 2nd ed.; Varki, A., Cummings, R. D., Esko, J. D., Freeze, H. H., Stanley, P., Bertozzi, C. R., Hart, G. W., Etzler, M. E., Eds.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 2009; pp 229−248. (6) Rabenstein, D. L. Nat. Prod. Rev. 2002, 19, 312−331. (7) Conrad, H. E. Curr. Protoc. Mol. Biol. 2001, 32, 17.22.1−17.22.5.

■

CONCLUSION We have shown that water ice is a versatile matrix for the desorption/ionization of sulfated glycosaminoglycans under IRMALDI o-TOF-MS conditions. The use of a temperature6445



Article

(8) Shaklee, P. N.; Conrad, H. E. Biochem. J. 1984, 217, 187−197. (9) Karousou, E. G.; Viola, M.; Genasetti, A.; Vigetti, D.; Luca, G. D.; Karamanos, N. K.; Passi, A. Biomed. Chromatogr. 2005, 19, 761−765. (10) Maccarana, M.; Sakura, Y.; Tawada, A.; Yoshida, K.; Lindahl, U. J. Biol. Chem. 1996, 271, 17804−17810. (11) Takegawa, Y.; Araki, K.; Fujitani, N.; Furukawa, J.; Sugiyama, H.; Sakai, H.; Shinohara, Y. Anal. Chem. 2011, 83, 9443−9449. (12) Zaia, J. Mass Spectrom. Rev. 2004, 23, 161−227. (13) Zaia, J. Mass Spectrom. Rev. 2009, 28, 254−272. (14) Sisu, E.; Flangea, C.; Serb, A.; Zamfir, A. D. Amino Acids 2011, 41, 235−256. (15) Carr, S. A.; Reinhold, V. N. J. Carbohydr. Chem. 1984, 3, 381− 401. (16) Dell, A.; Rogers, M. E.; Thomas-Oates, J. E.; Huckerby, T. N.; Sanderson, P. N.; Nieduszynski, I. A. J. Carbohydr. Res. 1988, 179, 7− 19. (17) Flangea, C.; Sisu, E.; Seidler, D. G.; Zamfir, A. D. Anal. Biochem. 2012, 420, 155−162. (18) Przybylski, C.; Gonnet, F.; Hersant, Y.; Bonnaffé, D.; LortatJacob, H.; Daniel, R. Anal. Chem. 2010, 82, 9225−9233. (19) Camara, J. E.; Satterfield, M. B.; Nelson, B. C. J. Pharm. Biomed. Anal. 2007, 43, 1706−1714. (20) Mormann, M.; Zamfir, A. D.; Seidler, D. G.; Kresse, H.; PeterKatalinić, J. J. Am. Soc. Mass Spectrom. 2007, 18, 179−187. (21) Lawrence, R.; Olson, S. K.; Steele, R. E.; Wang, L.; Warrior, R.; Cummings, R. D.; Esko, J. D. J. Biol. Chem. 2008, 283, 33674−33684. (22) Ridlova, G.; Mortimer, J. C.; Maslen, S. L.; Dupree, P.; Stephens, E. Rapid Commun. Mass Spectrom. 2008, 22, 2723−2730. (23) Gill, V. L.; Aich, U.; Rao, S.; Pohl, C.; Zaia, J. Anal. Chem. 2013, 85, 1138−1145. (24) Zaia, J.; McClellan, J. E.; Costello, C. E. Anal. Chem. 2001, 73, 6030−6039. (25) Zaia, J.; Costello, C. E. Anal. Chem. 2003, 75, 2445−2455. (26) McClellan, J. E.; Costello, C. E.; O’Connor, P. B.; Zaia, J. Anal. Chem. 2002, 74, 3760−3771. (27) Kailemia, M. J.; Li, L.; Ly, M.; Linhardt, R. J.; Amster, I. J. Anal. Chem. 2012, 84, 5475−5478. (28) Leach, F. E., III; Ly, M.; Laremore, T. N.; Wolff, J. J.; Perlow, J.; Linhardt, R. J.; Amster, I. J. J. Am. Soc. Mass Spectrom. 2012, 23, 1488− 1497. (29) Wolff, J. J.; Leach, F. E., III; Laremore, T. N.; Kaplan, D. A.; Easterling, M. L.; Linhardt, R. J.; Amster, I. J. Anal. Chem. 2010, 82, 3460−3466. (30) Juhasz, P.; Biemann, K. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4333−4337. (31) Venkataraman, G.; Shriver, Z.; Raman, R.; Sasisekharan, R. Science 1999, 286, 537−542. (32) Bultel, L.; Landoni, M.; Grand, E.; Couto, A. S.; Kovensky, J. J. Am. Soc. Mass Spectrom. 2010, 21, 178−190. (33) Schiller, J.; Arnhold, J.; Benard, S.; Reichl, S.; Arnold, K. Carbohydr. Res. 1999, 318, 116−122. (34) Nimptsch, A.; Schibur, S.; Schnabelrauch, M.; Fuchs, B.; Huster, D.; Schiller, J. Anal. Chim. Acta 2009, 635, 175−182. (35) Böhme, J.; Anderegg, U.; Nimptsch, A.; Nimptsch, K.; Hacker, M.; Schulz-Siegmund, M.; Huster, D.; Schiller, J. Anal. Biochem. 2012, 421, 791−793. (36) Tissot, B.; Gasiunas, N.; Powell, A. K.; Ahmed, Y.; Zhi, Z.-L.; Haslam, S. M.; Morris, H. R.; Turnbull, J. E.; Gallagher, J. T.; Dell, A. Glycobiology 2007, 17, 972−982. (37) Laremore, T. N.; Murugesan, S.; Park, T. J.; Avci, F. Y.; Zagorevski, D. V.; Linhardt, R. J. Anal. Chem. 2006, 78, 1774−1779. (38) Laremore, T. N.; Zhang, F.; Linhardt, R. J. Anal. Chem. 2007, 79, 1604−1610. (39) Przybylski, C.; Gonnet, F.; Bonnaffe, D.; Hersant, Y.; LortatJacob, H.; Daniel, R. Glycobiology 2010, 20, 224−234. (40) Von Seggern, C. E.; Cotter, R. J. J. Am. Soc. Mass Spectrom. 2003, 14, 1158−1165. (41) Menzel, C.; Dreisewerd, K.; Berkenkamp, S.; Hillenkamp, F. Int. J. Mass Spectrom. 2001, 207, 73−96.

(42) Dreisewerd, K.; Berkenkamp, S.; Leisner, A.; Rohlfing, A.; Menzel, C. Int. J. Mass Spectrom. 2003, 226, 189−209. (43) Dreisewerd, K.; Kölbl, S.; Peter-Katalinić, J.; Berkenkamp, S.; Pohlentz, G. J. Am. Soc. Mass Spectrom. 2006, 17, 139−150. (44) Meisen, I.; Mormann, M.; Müthing, J. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2011, 1811, 875−896. (45) Kirpekar, F.; Berkenkamp, S.; Hillenkamp, F. Anal. Chem. 1999, 71, 2334−2339. (46) Dybvik, A. I.; Norberg, A. L.; Schute, V.; Soltwisch, J.; PeterKatalinić, J.; Vårum, K. M.; Eijsink, V. G. H.; Dreisewerd, K.; Mormann, M.; Sørlie, M. Anal. Chem. 2011, 83, 4030−4036. (47) Tajiri, M.; Wada, Y. Int. J. Mass Spectrom. 2011, 305, 164−169. (48) Pirkl, A.; Soltwisch, J.; Draude, F.; Dreisewerd, K. Anal. Chem. 2012, 84, 5669−5676. (49) Domon, B.; Costello, C. E. Glycoconjugate J. 1988, 5, 397−409. (50) Soltwisch, J.; Souady, J.; Berkenkamp, S.; Dreisewerd, K. Anal. Chem. 2009, 81, 2921−2934. (51) Kenny, D. T.; Issa, S. M. A.; Karlsson, N. G. Rapid Commun. Mass Spectrom. 2011, 25, 2611−2618. (52) Shi, X.; Huang, Y.; Mao, Y.; Naimy, H.; Zaia, J. J. Am. Soc. Mass Spectrom. 2012, 23, 1498−1511. (53) Naggar, E. F.; Costello, C. E.; Zaia, J. J. Am. Soc. Mass Spectrom. 2004, 15, 1534−1544. (54) Wolff, J. J.; Laremore, T. N.; Busch, A. M.; Linhardt, R. J.; Amster, I. J. J. Am. Soc. Mass Spectrom. 2008, 19, 790−798. (55) Kailemia, M. J.; Li, L.; Xu, Y.; Liu, J.; Linhardt, R. J.; Amster, I. J. Mol. Cell. Proteomics 2013, 12, 979−990.

6446


Water Ice is a Soft Matrix for the Structural ... - ACS Publications

Recommend Documents