Distinguishing the glycan isomers 2,3-sialyllactose and 2,6-sialyllactose by voltammetry after modification with osmium(VI) complexes

Highlights

• Glycan isomers differing by a single bond can be distinguished by voltammetry.

• Electro-inactive sialyllactose (SL) produces redox peaks after Os(VI)-modification.

• Os(VI)temed complexes react specifically with vicinal diols in SL glycan.

• α(2,6) SL isomer can be modified by three Os(VI) while α(2,3) isomers by two Os(VI).

• Glycan isomer discrimination is almost independent on glycan concentration.

Abstract

Altered glycosylation is a universal feature of cancer cells and certain glycans are well-known markers of tumor progression. In this work we studied two glycan isomers, 2,3-sialyllactose (3-SL) and 2,6-sialyllactose (6-SL), frequently appearing in glycoproteins connected with cancer. A combination of square wave voltammetry and glycan modification with osmium(VI) N,N,N′,N′-tetramethylethylenediamine (Os(VI)tem) allowed to distinguish between these regioisomers, since the 6-SL molecule can bind three Os(VI), while the 3-SL only two Os(VI) moieties, as experiments using capillary electrophoresis, inductively coupled plasma mass spectrometry and thin layer chromatography showed. A similar pattern of Os(VI)-modification was found for isomers of sialyl-N-acetyllactosamine and sialylgalactose. Covalent adducts of Os(VI)tem with glycans yielded three reduction voltammetric peaks. The ratio of peak I/peak II heights depends on the content of individual regioisomer in the sample. Our proposed approach allows the determination of isomer percentage representation in the mixture after one voltammogram recording. These results show a new appropriate method for the discrimination of glycan isomers containing terminal sialic acid important for distinguishing between cancerous and non-cancerous origin of biomarkers.

Introduction

Most of the processes in living cells are in one way or another associated with certain forms of carbohydrate interactions. Recent discoveries show many critical roles for glycans and new glycan functions are constantly being discovered due to their vast structural diversity [1,2]. Any alterations in a glycan's composition or structure could potentially be either a cause or consequential attribute of a biochemical imbalance that we recognize as a „disease“ [3]. Glycan inherent structural diversity creates a major analytical challenge in glycoscience as well as in cancer biomarker research [4]. The sialic acid (N-acetylneuraminic acid, Neu5Ac) is frequently terminal residue of extracellular glycan chains and is typically connected either by α(2,3) or α(2,6) glycosidic bond to galactose (Gal) [5]. Tumor-relating changes are often related with direct structural changes of α(2,3)/α(2,6)-sialic acid glycosidic linkages on glycoproteins [[6], [7], [8]]. New strategies including the detection of protein glycosylation can significantly enhance diagnostic specificity (e.g., in prostate serum antigen (PSA) as the biomarker of prostate cancer [6,9]). Terminal Neu5Ac linked to Gal predominantly via α(2,6)-glycosidic linkage is present in the PSA glycans of healthy individuals. Contrary to this, the presence of α(2,3)-glycosidic linkage of Neu5Ac to Gal is significantly increased in prostate cancer [10]. PSA samples from patients with benign prostate hyperplasia show no increase of α(2,3) linkage of terminal Neu5Ac to Gal. Therefore the monitoring of altered sialylation pattern is a promising tool for distinguishing between PSA of cancerous and non-cancerous origin [9]. Recent progress in understanding of the biological functions of glycans has stimulated an improvement of relevant analytical methods [11,12]. Simple, fast and inexpensive methods suitable for miniaturizing are still required. Electrochemical methods could be appropriate candidates for such glycan analysis, however, most of glycans are electrochemically inactive [13,14]. One of the promising strategies for the glycan study using an electrochemical analysis, is their chemical modification [15,16]. Osmium oxocomplexes with suitable nitrogenous ligands (OsL) were successfully used for labeling DNA [17,18], and proteins [19] as well as saccharides [[20], [21], [22]] for their highly sensitive and selective electrochemical analysis [23]. Oligosaccharides and polysaccharides react with Os(VI)tem (where tem stands for N,N,N′,N′-tetramethylethylenediamine) to yield corresponding stable ligand osmate esters in aqueous media [13,20,24]. Os(VI)L modification products are electrochemically active and produce redox signals on carbon or mercury electrodes [20,24]. Redox peaks occur usually in three reversible pairs and are described as peaks I – III on the hanging mercury drop electrode (HMDE) and peaks α, β and γ at carbon electrodes [13,14,24]. The nature of these signals was not investigated more closely, however it can be expected that similar electrode processes occur when OsO₄ is step-by-step electroreduced as was described by J. Heyrovský [25] and L. Meites [26]. Peak I on the HMDE could correspond to reduction of Os(VI) to Os(IV), peak II of Os(IV) to Os(III) [13] and peak III of Os(III) to Os(II) or to the metal.

In our previous works we showed that two glycan isomers containing sialic acid (different only in a single bond, i.e., 3′-sialyllactose (3-SL) and 6′-sialyllactose (6-SL), can be distinguished using electrochemical approaches based on modification with Os(VI)tem [21] or using lectins [27]. The analysis of glycan isomers is based on a combination of simple chemical modification of the glycans by Os(VI)tem complexes and square wave voltammetry at HMDE or at pyrolytic graphite electrode (PGE) [21]. Os(VI)tem itself usually gives only a small voltammetric response at HMDE and thus its modification products with saccharides can be detected directly in the reaction mixture without any purification [13] in contrast to PGE [21], where unreacted Os(VI)tem interferes with glycan analysis. It was necessary to wash PGE modified by SL-Os(VI)tem before the analysis. Higher isomer concentration was required at PGE, the reproducibility was much worse (RSD 17%) as compared to that obtained with HMDE (RSD 5%). Additionally, the differences between responses of 3-SL and 6-SL isomers were smaller than those at HMDE [24]. The mercury electrodes have extremely broad potential window in the cathodic region, which makes mercury the best available material for the determination of electrochemically reducible analytes. The easily renewable surface eliminates or minimizes problems with electrode passivation and fouling. The atomically smooth surface ensures extremely high reproducibility and excellent sensitivity of the analysis [28,29].

In this work we proposed a method for distinguishing two sialyllactose isomers in their mutual mixture. Both Os-modified trisaccharides (3-SL and 6-SL), yielded three reduction peaks at HMDE [21] higher for 6-SL isomer than for 3-SL. The ratio of peak I/II heights of SL-Os(VI)tem differed for both isomers and linearly increased with increasing content of 3-SL in the mixture. Hence, we were able to determine the percentage representation of individual isomers in the mixture from one voltammogram. Additionally, we found that higher peaks of 6-SL were due to higher content of Os moieties as we showed using complementary methods such as capillary electrophoresis, inductively coupled plasma mass spectrometry and thin layer chromatography. Three Os moieties can be bound to 6-SL, while only two Os moieties bind to 3-SL. We observed similar differences in Os(VI)tem content also for other oligosaccharides with Neu5Ac bound to Gal via α(2,6) or α(2,3) linkage. Percentage content of these glycan isomers was also possible to determine from one measurement.