Galactosaminoglycan Function and Oligosaccharide Structure Determination

This review will discuss the importance of sequencing long chondroitin sulfate and dermatan sulfate chains specifically derived from decorin. Decorin is a member of the small leucine-rich repeat proteoglycans and ubiquitously expressed primarily in the skin. Sequence information and diverse function of glycosaminoglycans is further influenced by variable expression through the core protein indicating the importance to analyse glycosaminoglycans from specific proteoglycans.


INTRODUCTION
Glycosaminoglycans are linear polysaccharides exhibiting a wide range of biological function. The primary structure of several glycosaminoglycans is organized macroscopically by segregation into alternating blocks of specific sulfation patterns and microscopically by formation of oligosaccharide sequences with specific binding functions. The biological importance of the fine structures of dermatan sulfate has been demonstrated by the finding that particular sulfated oligosaccharide sequence within these glycosaminoglycans can confer growth factor or anticoagulant activity.

Structural Analysis of Oligosaccharides Derived from Galactosaminoglycans
For structure investigation of CS/DS glycoforms, development of high performance specific methods was lately required, among which, mass spectrometry (MS) contributed an essential progress to the field [28]. MS with electrospray (ESI) became nowadays the most valuable technique for structural and quantitative CS/DS investigation because it uniquely offers the combination of analytical versatility, sensitivity, and precision.
Prior to ESI introduction difficulties encountered in the ionization of CS/DS mixtures limited for long time the MS potential for their structural elucidation. The most severe problems in this regard were related to the impossibility to obtain high ionization yield for long chains, hamper unspecific in-source decomposition of the labile sulfate group, generate the multiply charged ions of high molecular weight species and finally distinguish the isobaric structures. Under appropriate optimized conditions, and negative ion mode detection settings, intact negative ions can be formed and transferred into the MS without a significant in-source fragmentation, which would primordially lead to desulfation and induction of spectrum/interpretation artifacts. Another advantage of the method is that the majority of modern mass spectrometers working in conjunction with ESI exhibit the tandem MS facility, which gives the possibility to sequence the precursor ion and thus obtain straightforward information upon the repeating disaccharide unit, number of sulfate groups and their precise location along the CS/DS chain and/or within a certain monomer ring [29,30]. In the case of chain depolymerization using lyases complex CS/DS mixtures containing intact chains of variable length are formed. In ESI-MS, because of chain constitution, the monoisotopic peaks corresponding to charge states that equal ½ the number of the repeating units are observed at the same m/z value. This isobaric peak overlapping results in another specific challenges of ESI-MS of GAGs in general and CS/DS in particular since it makes the screening spectra difficult to interpret and jeopardizes the isolation of the precursor ion for tandem MS. For these reasons, the screening of the complex DS mixtures requires the combination of MS with an efficient separation technique properly optimized for this type of analysis. From this point of view, ESI is also advantageous as it allows for on-line coupling of MS with chromatographic separation methods such as size exclusion chromatography (SEC) [31], high performance liquid chromatography [32] and capillary electrophoresis (CE) [33].
Novel insights into the domain structure of the hybrid CS/DS chains obtained from bovine aorta by enzymatic depolymerization using chondroitin B lyase, which is the characteristic of oligosaccharides representing defined hybrid molecules bearing a single DS disaccharide unit at the non-reducing end, linked to a variable number of CS disaccharide units at the reducing terminal, were recently obtained using an analytical approach based on CE in conjunction with negative ESI-quadrupole time-of-flight tandem mass spectrometry (QTOF-MS/MS) [34]. By CE fractionation followed by ESI-MS detection, several oligosaccharides containing one SO 3 per disaccharide were identified. The species were all detected bearing one double bond per molecule, which clearly indicated that contain L-IdoA building block formed by elimination during the enzymatic cleavage of the polysaccharide chain. The ion corresponding to a composition of Δ-unsaturated 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 3( 4S) octasaccharide, has been selected as a precursor for sequencing in a tandem MS experiment. Beside ions indicative of a regularly sulfated pattern, two detected oversulfated fragments showed that in the investigated octasaccharide chain one disaccharide unit might be disulfated. Moreover, it could be postulated that the ions might arise from either the same sequence species, or document even two different locations of the oversulfation. Additional significant structural aspect revealed by this analysis is the determination of the sulfate linkage site. According to the relative intensity of typical cleavages determined by tandem MS, it was clearly established that in the disaccharide units constituting the octasaccharide CS/DS of the bovine aorta, the GalNAc is sulfated in position 4 while the individual disaccharide cleaved-off by the lyase, is sulfated in position 6 of the GalNAc moiety.
Studies dedicated to hybrid CS/DS chains derived from decorin secreted by human skin fibroblasts, showed their interaction with FGF-2, as the oligosaccharides release by chondroitin B lyase digestion [35]. For identification of the biologically active sequence, we reported upon the efficiency of our improved protocol, which targeted the structural analysis of enzyme-resistant oligosaccharides larger than standard trisulfated hexasaccharides [35]. The method was based also on CE for separating oversulfated species in off-line combination with nanoESI-QTOF-MS/MS. The heterogeneity of the oligosaccharide mixture was first demonstrated by CE with UV detection. Further, by nanoESI-QTOF-MS analysis of the CE fractions, up to 12-mer oligosaccharides with different degrees of sulfation were identified. Another interesting information emerging from the MS screening of the separated CE fractions was that the species with high molar sulfate content could be clearly separated from the non-sulfated ones, present in the CS/DS mixture released by β-elimination. For strict determination of the degree of sulfation in single GAG species and delimiting the real under-and nonsulfated species from the possible artifacts induced by the in-source decay of the sulfate groups in the MS mode, this aspect was crucial.
In analogy to the previous reports upon ESI-MS methods for GAG oligosaccharide analysis [28,29] it was observed that in the negative ESI-MS, the in-source desulfation may be reduced by acquiring the spectra under mild values of the MS counterelectrode potential. Using this procedure, it was possible to detect up to fully sulfated dodecasaccharide and a number of intact oversulfated species (Table 1). Detailed structural characterization was achieved by fragmentation of the novel CS/DS-containing hexasaccharide, 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 2 (5S), for determination of the sulfation pattern along the carbohydrate chain. According to the MS/MS data three sulfates were found distributed in the IdoAGalNAcGlcA moiety, offering two structural motifs: one containing the sulfated L-IdoA and the disulfation of GalNAc moiety, and the other with the both UroA moieties and the GalNAc each monosulfated (Fig.1). The structural data gathered for the novel oligosaccharide 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 2 (5S) confirmed the presence of a tetrasulfated tetrasaccharide partial sequence assigned either to the IdoA(S)GalNAc(S)GlcA(S)GalNAc(S) or to the IdoA(S)GalNAc(2S)GlcAGalNAc(S) moiety.
In our recent study, the hybrid CS/DS chains derived from decorin secreted by human skin fibroblasts were also released by β-elimination and further subjected to depolymerization using chondroitin AC lyase. Following the procedure we described previously [35] the resulting oligosaccharide mixture was separated by SEC. The octasaccharide fraction was collected, purified and analyzed by advanced, highly sensitive chip electrospray ionization using fully automated sample infusion into QTOF-MS. The chip nanoESI-MS revealed hexa-, penta and quadruply charged molecular ions corresponding to the unsaturated regularly sulfated DS octasaccharide having the composition of 4,5-Δ-GlcAGalNAc[IdoAGalNAc] 3 (4S), and a triply charged ion assigned to an undersulfated octasaccharide motif 4,5-Δ-GlcAGalNAc[IdoAGalNAc] 3 (3S) (Fig.2). The oligosaccharide 4,5-Δ-GlcAGalNAc[IdoGalNAc] 3 (4S) could be part of the structure for activation heparin cofactor II where   [36].
With neoglycolipid microarrays Yamaguchi and co-workers [37] showed that DS-oligosaccharides larger than octamers bind preferentially FGF-7, HGF and heparin cofactor II. These results show the  (1400-1700). Assignment of fragment ions is according to Domon and Costello [39]. Reprinted with permission from [38]. importance of sequencing longer galactosaminoglycan chains to determine the binding and active structure. Reliable data upon extended galactosaminoglycan chains [38] in mixtures from conditioned culture medium of 293 cells transfected with cDNA for decorin, containg mainly CS structures [11], were obtained using a novel approach based on on-line CE/ESI-QTOF-MS and tandem MS. An unsaturated oversulfated tetradecasaccharide assigned to the composition of 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 6 (9S), a saturated eicosasaccharide IdoAGalNAc[GlcAGalNAc] 9 (11S) an unsaturated octadecasaccharide 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 8 (10S) and an abundant unsaturated eicosasaccharide bearing eleven sulfate groups having the structure of 4,5-Δ-IdoAGalNAc[GlcAGalNAc] 9 (11S) were identified. These data demonstrated the complexity of the sample, which is related to the elevated length of the GAG oligosaccharide chains present, as well as their type and high level of sulfation. The molecular ions obtained in this experiment were shown to carry different numbers of sulfate groups per disaccharide unit and were all expressing oversulfation of the molecules. High percentage of species separated and detected in the mass spectra was observed to carry one double bond, originating from the specific eliminative action of chondroitin B lyase on GalNAc-IdoA linkages. The structural information upon the oversulfated unsaturated eicosasaccharide 4,5-Δ IdoAGalNAc [GlcAGalNAc] 9 (11S) were acquired by the advanced data-dependent MS method. In the MS to MS/MS switching approach applied to the pentacharged 4,5-ΔIdoAGalNAc[GlcAGalNAc] 9 (11S) (Fig. 3) the product ion analysis generated fair fingerprint ion set for identification of the molecular structure in terms of epimerization and sulfation pattern (Fig. 4). According to this diagnostic group of ions and the rest of regularly and oversulfated fragments, the additional sulfate group is indubitably linked to the last GalNAc moiety at the reducing end.
All these structural studies demonstrate in the first place that the determination of molecular characteristics of CS/DS, is an essential prerequisite in understanding their functions. To understand the functional significance of CS/DS one must consider both the structure of the CS/DS chain and the core protein, in this case decorin protein core, which leads the GAG chain into the extracellular matrix. Secondly, these findings show that the structural analysis must start with the identification of all glycoforms in complex mixtures obtained after detachment from core-protein by chemical or enzymatic methods followed by collecting and assembling detailed data upon epimerization, sulfate content and its distribution. Lastly, it appears clear that the presence of regular or irregular domains plays a critical biological role and can be directly recognized only by sequencing single components in tandem MS experiments.