Glycobiology, 2001, Vol. 11, No. 8 663-676
© 2001 Oxford University Press
Diversity of sialic acids revealed using gas chromatography/mass spectrometry of heptafluorobutyrate derivatives
2Laboratoire de Chimie Biologique, CNRS UMR 8576, 59655 Villeneuve d’Ascq Cedex France, and 3Biochemisches Institut, Universität Kiel, Olshausenstrasse 40, D-24098 Kiel, Germany
Received on January 16, 2001; revised on March 20, 2001; accepted on March 21, 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
The fine structural motifs of sialic acids, a frequent terminal monosaccharide of glycans, seem to contain essential biological properties. To identify such subtle structural differences, a reliable method was developed for the qualitative and quantitative identification of sialic acids present in different tissues and fluids. This method involved, after liberation of sialic acids by mild acid hydrolysis, their methyl esterification using diazomethane in the presence of methanol and the formation of volatile derivatives using heptafluorobutyric anhydride. The derivatives were analyzed by gas chromatography coupled to mass spectrometry in the electron impact mode. This technique allowed the separation and identification of a large variety of sialic acids, including different O-acylated forms of N-acetyl and N-glycolyl neuraminic acids and of 3-deoxy-D-glycero-D-galacto-nonulosonic acid (Kdn). This method allowed also identifying 8-O-methylated and 8-O-sulfated derivatives, de-N-acetylated neuraminic acid, and 1,7-sialic acid lactones. Compounds present in very complex mixtures could be identified through their fragmentation patterns. Because of the stability of the heptafluorobutyrate derivatives, this method presents important improvements compared to the previous techniques, because it can be frequently applied on very small amounts of crude samples. This methodology will support progress in the field of the biology of sialic acids.
Key words: Kdn/lactone/lactyl/methyl/sulfate
|
Introduction |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
In recent years, increasing interest was drawn to the diversity of sialic acids, because these monosaccharides could play important biological functions (for reviews, see Varki, 1992
, 1997; Schauer and Kamerling, 1997; Sinha et al., 2000). For example, 9-O-acetyl N-acetyl neuraminic acid (Neu5Ac) is a specific ligand for the agglutinin of influenza virus C (Roy et al., 1992), whereas 4-O-acetylated Neu5Ac is a specific ligand for the agglutinin of the murine hepatitis virus (Regl et al., 1999). Based on previous work (for review, see Schauer and Kamerling, 1997), sialic acids present an extreme diversity, and more than 40 different compounds were identified differing in the presence or absence of an amino group in position 5 (sialic acids or 3-deoxy-D-glycero-D-galacto-nonulosonic acid [Kdn]), different acylations of the NH2 group at position 5 (acetyl, glycolyl), and various substituents of the different hydroxyl groups (acetyl, lactyl, methyl, sulfate, phosphate, etc.). Different complex monosaccharides presenting similarities with sialic acids were identified in bacteria, which may represent important epitopes (Knirel et al., 1987, 1997).
Several methods were developed for the identification and quantitative determination of these compounds. The use of fluorescent labeling specific for 2-keto carboxylic acid (Hara et al., 1989; Klein et al., 1997; Iwersen et al., 1998) followed by the separation of the compounds by high-performance liquid chromatography (HPLC) provided a very sensitive method. However, the separation power of HPLC was still insufficient for resolving very complex mixtures, whereas coupling with mass spectrometry (MS) was difficult for routine analyses. Furthermore, this method could not reveal possible lactones of sialic acids, which might play important biological roles. In contrast, gas chromatography (GC)/MS methods using the derivatization of the methyl esters of sialic acids as trimethyl-silyl (TMS) ethers, allowed easy GC/MS analyses (Kamerling et al., 1975; Schauer et al., 1976; Kamerling and Vliegenthart, 1982). However, these methods also suffered from some difficulties. One was that, when applied to biological samples presenting different impurities, the silylation reaction had a poor yield and the derivatives were unstable. The other was that the TMS derivative of the semi-acetalic group was unstable and was partially lost by pyrolysis in the injector of the gas chromatograph in a poorly reproducible way. Consequently, instead of the two peaks of the anomers of each sialic acid, four peaks were produced, as already observed for glucosamine (Maes et al., 1999). One way to circumvent these problems was the use of strong acylating agents, like heptafluorobutyric anhydride (HFBAA), resulting in high-mass compounds with a very poor adsorption on classical methyl-siloxane liquid phases (Zanetta et al., 1999) and, consequently, that eluted at relatively low temperature. Furthermore, the derivatives present a very strong stability with time, the reaction can take place quantitatively even in the presence of salts, proteins, or other contaminants and the semi-acetalic group is not derivatized. This article reports the methodology of formation of their volatile derivatives and presents the GC/MS data obtained in the electron impact mode allowing the identification of different sialic acid derivatives found in various biological samples.
|
Results and discussion |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
Initial attempts for GC/MS analysis of sialic acid using heptafluorobutyrate derivatives
Preliminary experiments indicated that the carboxyl group had to be blocked to get volatile derivatives. We first tried to methyl esterify the carboxyl groups using iodomethane (ICH3) in dimethyl sulfoxide (Powell and Harvey, 1996
), a procedure allowing a quantitative reaction as evidenced by matrix-assisted laser desorption and ionization time-of-flight analysis of sialylated compounds (Zanetta et al., 2000). However, this procedure was not compatible with a subsequent analysis as heptafluorobutyrate (HFB) derivatives, because HI and I2 liberated from ICH3 during the methyl esterification formed stable complexes with HFBAA and heptafluorobutyric acid in such a way that the acylation mixture could not be evaporated to dryness before injection into the GC/MS apparatus. This resulted in a strong solvent tailing and artifactual peaks. Furthermore, it was observed that pure Neu4,5Ac2 and Neu5,9Ac2, showed a partial de-O-acetylation on this treatment. We therefore used the diazomethane methyl esterification procedure (Figure 1), but the yield of methyl esterification was extremely low when dry sialic acid samples were treated with a solution of diazomethane in ether because of the insolubility of sialic acids in ethyl-oxide. In fact, the yield of the reaction could be increased to completion (see the data for sialyl-lactose, Figure 2c) adding first anhydrous methanol (25–200 µl for 200 µl of the diazomethane reagent; Kamerling and Vliegenthart, 1982; Fontaine et al., 1994). A total methyl esterification could be obtained only after 1 h at room temperature. For security, we used 4 h or overnight at room temperature in routine analysis. In fact, the sialic acids could be left for more than 1 week in closed vials containing the reagent without alteration of the yield. After evaporation of the reagent, high-temperature acylation with HFBAA (150°C for 5 min) resulted into an extremely low tailing of the solvent and a very few artifactual peaks. When working with pure O-acetylated Neu5Ac derivatives (Neu4,5Ac2 and Neu5,9Ac2, Neu5,7,9Ac3), only extremely minor traces of Neu5Ac were detected (Figure 2a). This indicated that the methyl esterification procedure and the high-temperature acylation with HFBAA did not provoke a significant destruction of O(N)-acylated sialic acids.
|
|
To examine if the yield of the reaction (and especially that of the methyl esterification) was quantitative, experiments were performed in the presence of two internal standards. One was Glc and the other the O-methyl glycoside of the methyl ester of sialic acid obtained after methanolysis of Neu5Ac with an almost quantitative yield (Zanetta et al., 1999
). These experiments indicated that for the standard sialic acids mentioned above the yield was quantitative. Because of different retention times (RTs) and differences in the fragmentation patterns, the data obtained for Glc indicated that the methyl esterification with diazomethane did not result in the formation of O-methyl glycosides but in anomers having a free OH group on the C(2) carbon atom. This point was of importance because of the absence of molecular ions in the electron impact (EI) mode and of fragmentations in the CI modes for all sialic acid derivatives, it was not possible to directly determine if the compounds were O-methyl glycosides (loss of 31/32; methanolate/methanol) or not (loss of 17/18; OH/H2O). Because problems of stability with time were encountered for TMS derivatives, the same samples were analyzed at different periods of time after the addition of the acylating mixture (200 µl acetonitrile and 25 µl HFBAA) after evaporation of the diazomethane/ether mixture. Samples were immediately heated at 150°C for 5 min, evaporated under nitrogen, dissolved into dried acetonitrile, and analyzed. The results were compared to samples remaining for 1 month at room temperature in the acylating mixture, heated at 150°C for 5 min, evaporated under nitrogen, and analyzed. The two types of samples were indistinguishable. This indicated that standing at room temperature in the acylating mixture did not introduce degradation of the sialic acid derivatives. This was not the case when already analyzed samples (depleted of the acylating mixture) were standing for such a period of time, even in closed vials. The quantity of the initial sialic derivatives was decreased concomitantly with the appearance of additional peaks (including de-O-acetylated sialic acids). This indicated that O-acyl groups were relatively unstable when standing for a long period of time in a medium especially enriched in heptafluorobutyric acid, but remain stable in the presence of an excess of HFBAA. Nevertheless, already analyzed samples could be stored unaffected for at least 3 months after the addition of 25 µl of HFBAA.
GC separation of the HFB derivatives of sialic acids
Analyses of standard samples indicated that most of the HFB derivatives of the methyl esters of sialic acids could be separated using classical methyl-siloxane columns. As previously reported, HFB derivatives of sugars poorly interacted with these liquid phases (Zanetta et al., 1999). For example, the HFB derivatives of malto-tetraose (mass 3214 Da) is eluted before the C24:0 fatty acid methyl ester (mass 382 Da). Most of the HFB derivatives of the methyl esters of sialic acids were eluted between 140 and 200°C, depending on their mass and on their degree of initial O-substitution and the nature of their N-substitution. For example, the HFB derivative of the methyl ester of Neu5Ac (mass 1107 Da) was eluted (Table I) before that of Neu4,5Ac2 (mass 953 Da). The derivatives of the different mono-O-acetylated Neu5Ac were eluted before those of the di-O-acetylated Neu5Ac. The derivative of N-glycolyl neuraminic acid (Neu5Gc) (mass 1319 Da) was eluted at a temperature higher than that of Neu5Ac, whereas the derivative of Kdn (mass 1262 Da; without N-acyl group) was eluted before that of Neu5Ac (Table I and Figure 2b). The general rules of separation were that compounds having additional acetyl groups were retarded despite their lower mass, whereas compounds with a 9-O-lactyl group were eluted before the 9-O-acetyl compounds despite their higher mass (Table I).
|
The derivatives of sialic acids were completely separated from the derivatives of pentoses, deoxy-hexoses, hexoses, hexitols, and N-acetyl-hexosamines present in some samples (Figure 1b). Indeed, the hydrolysis procedure used for the liberation of sialic acids was able to cleave (although with a much lower yield) the glycosidic bonds involving other peripheral monosaccharides. When these compounds were major nonreducing terminal constituents, they became present at an important level compared to sialic acids. Derivatives of the commonly found monosaccharides other than sialic acids were easily characterized by their A1 ion (Kochetkov and Chizhov, 1966
) at m/z 946/947 for hexoses, 720/721 for pentoses, 734/735 for deoxy-hexoses, 790/791 for N-acetyl-hexosamines, and 778/779 for uronic acid methyl esters. Interferences were actually observed in several samples with the derivatives of di- and trisaccharides made of glucose eluted as couples of anomers in the area of the derivatives of acetylated sialic acids (Figure 2b,c). These compounds could be easily identified through specific fragmentation patterns in the EI mode (intense A1 ion at m/z 947 and the sequence of ions derived from the A series by successive losses of heptafluorobutyric acid at m/z 733 and 519). It should be stressed that for sialyl-lactose, the present technique allows the simultaneous determinations of sialic acids and of lactose (Figure 2c). The integration of the TIC responses of the product from sialyl(Neu5Ac)lactose, 34.9% for Neu5Ac and 65.1% lactose, gave a result in perfect agreement with the mass of these compounds. Such a result could not be obtained for glycosphingolipids because of the absence of cleavage of the sphingosine–monosaccharide bond under the hydrolysis conditions used for the liberation of sialic acid.
Identification of the parent members of the families of sialic acids
All sialic acids derivatives (except those derived from Kdn) gave an extremely predominant ß-anomer (96%) after acid hydrolysis. As shown in Figure 3a,b, Neu5Ac and Neu5Gc presented very different fragmentation patterns allowing their unambiguous identification in complex mixtures. Anomers of Kdn were found in a proportion close to 2/3–1/3. The proportion between the anomers changed slowly if samples were kept in the acylation mixtures for a long time because of mutarotation due to the presence of a free hydroxyl group on the C(2) carbon atom. As for Neu5Ac and Neu5Gc (not shown), the EI spectra of the HFB derivatives of Kdn anomers were significantly different (Figure 3c,d), the difference being concerned mostly with the intensities of the ions.
|
Specificity of the fragmentation patterns of Neu5Ac, Neu5Gc, and Kdn derivatives
Studies of the fragmentation patterns of these parent compounds and of O-acylated derivatives indicated that the fragmentation of Neu5Ac, Neu5Gc, and Kdn derivatives were similar (Figure 3a–d; see also Table I and Figure 4 for reporter ions). Indeed, for all these compounds, basic ions were related to the HFB derivatives (69 for CF3; 119 for CF3CF2; 169 for CF3CF2CF2). The molecular ion was undetectable, as did in most cases the A1 ion (M – H2O), but the higher-mass ion actually detectable were issued from a loss of mass of 76/77 (the resulting ion will be termed L1 in the following to avoid a confusion with the nomenclature proposed by Kochetkov and Chizhov, 1966
). It probably resulted from the loss of the substituents outside the pyranic ring on the C(2) carbon atom (losses of COOCH3 + H2O and of COOCH3 + OH). In fact, ions with the same mass could be formed by another mechanism, the loss of water between C(2) and C(3) being accompanied by the loss of the substituent of the C(5) for Neu5Ac derivatives or acetic acid from O-acetylated carbon atoms for acetylated derivatives of Neu5Ac, Neu5Gc, and Kdn. For Neu5Ac, an additional loss of the acetamido group of the C(5) carbon atom (58/59) gave a very weak ion at m/z 971/972, likely the parent ion of a series obtained by additional losses of HFBA residues giving the characteristic doublet ion at 757/758 and the ions at 543/544 and 329/330. Another series of ions was derived from an additional loss of 15 amu from the L1 ion, likely corresponding to the loss of the methyl of the acetamido group: ions at m/z 801/802, 587/588, and 373/374.
|
Neu5Ac showed a doublet of ions unrelated to the previous families at m/z 704/705. Because the same doublet was found only in Neu5Ac and Neu4,5Ac2, it was suggested that it included the chain outside the pyranic ring (i.e., the sequence C(9)H2OHFB-C(8)OHFB-C(7)OHFB). The nature of this ion remained speculative, but it could correspond to the C(9)–C(5) chain with the loss of the substituents of the C(5) carbon atom, similar to the K1 ions observed for derivatives of hexoses (Kochetkov and Chizhov, 1966
). Neu5Gc was characterized by an intense ion at m/z 227 corresponding to [CF3CF2CF2COOCH2]+ issued from the glycolyl group. This ion at 227 (Figure 4) constituted a reporter ion for most derivatives of Neu5Gc. Other series of ions were derived from M – 18 (loss of H2O) or M – 31 (loss of OCH3) by additional loss of heptafluorobutyric acid groups (m/z 214). It should be stressed that another series of ions were found derived from M – 50/51 (corresponding to a loss of a fluorine/hydrofluoric acid and of a methanolate group [M – F/HF – OCH3]). This loss of mass is generally accompanied by an additional loss corresponding to a mixed anhydride between two neighboring O-acylated carbon atoms. For example, compounds having non-O-acylated C(9) and C(8) gave an ion at M – 50–410 (410 is the mass of [CF3CF2CF2CO]2O). Another loss of mass relative to M – 50 was 256 for compounds having a single acetyl group on C(7)–C(9) and 102 ([CH3CO]2O) for those having two acetyl groups on these carbon atoms. For compounds having an 8-O-sulfate group the same mechanism gave an intense positively charged ion at m/z 295 (Figure 4) corresponding to a charged mixed anhydride between heptafluorobutyric and sulfuric acid ([CF3CF2CF2CO-O-SO3H2]+). This series of ions was extremely important for confirming the position of the substituents outside the pyranic ring.
In the area of low masses, several ions were of fundamental interest for the identification of the different series of compounds. For example, a basic ion at m/z 81 was present, corresponding to the pyranic ring depleted of all substituents with a series of double bounds in resonance. An ion at m/z 238 was characteristic of Neu5Ac derivatives was already found to be characteristic of 
-amino-alcohols like long-chain bases (Pons et al., 2000).
Mono- and multi-O-acylated derivatives of Neu5Ac
All mono-O-acetylated derivatives of Neu5Ac (mass 953 Da) could be identified through the intense doublet of ions at m/z 861/862 (Figure 5a–d) derived from the L1 ion by the loss of an additional methyl group. The derivative of Neu4,5Ac2 (especially abundant in equine material) could be unambiguously identified throughout the different mono-O-acetylated Neu5Ac (Figure 5a) because of the presence of additional intense doublet of ions at m/z 704/705 (see Specificity of the fragmentation patterns of Neu5Ac, Neu5Gc, and Kdn derivatives) and 490/491 derived from the former by a loss a heptafluorobutyric acid group. The two other derivatives lacked this ion. Neu5,7Ac2 (Figure 5b) was identified through two ions at m/z 790. Neu5,9Ac2 (Figure 5c) was characterized by its intense ion at m/z 490, a medium intensity ion at m/z 530 (L1 – [CH3CONH2 + C(9)H2OCOCH3 + CF3CF2CF2COOH]), its weak ion at 745 and overall by the ion at m/z 73 [C(9)H2OCOCH3]+ (Figure 4). Neu5,8Ac2, identified as a very minor constituent in bovine submandibular gland mucin (BSM), showed intense characteristic ions at m/z 492 and 307 (Figure 5d). These compounds showed significantly different RTs (Table I) and are separated from each other. Interferences in the interpretation of the spectra of O-acetylated derivatives of Neu5Gc could be solved because the corresponding ions showed a significantly lower mass (859/860 instead of 861/862). The reverse was observed with the O-acetylated derivatives of Kdn, because the second-order ions showed masses significantly higher (predominant ion at m/z 863). As shown in Table I, the relative retention times (RTr) to Neu5Ac provided an additional unambiguous confirmation of the nature of the compounds.
|
As expected from the data of the literature (for review, see Schauer and Kamerling, 1997
), a single lactylated Neu5Ac was identified, the lactyl group being in position 9: Neu5Ac9Lt (Figure 5e). As other lactylated compounds (see below), Neu5Ac9Lt presented a very intense ion at m/z 112 that allowed the almost unambiguous identification of these compounds (Figure 4). This ion probably resulted from the elimination of a HFBAA group between the C(9) and C(8) substituents followed by the formation of a cyclic ion with double bonds in resonance. Di-O-acetylated derivatives of Neu5Ac (mass 799 Da) gave a common L1 reporter ion at m/z 707/708. Unfortunately, such compounds were rare in our samples and only Neu5,7,9Ac3, Neu5,8,9Ac3, and Neu4,5,9Ac3 were identified (Figure 6a–c). Neu5,7,9Ac3 was characterized by ions at m/z 73 and m/z 153 (ring depleted of substituents except the C(7) carbon atom) and m/z 379 (the latter plus the HFB derivative of the C(8) carbon atom). Neu5,8,9Ac3 was characterized by the presence of intense ions at m/z 145 (substituted C(9) and C(8)) and 103 (the latter minus CH2=C=O; Figure 4). Neu4,5,9Ac3 was characterized by the ion at m/z 73 (acetylated C(9)) and by the absence of the ion at m/z 238 simultaneous with the presence of the ion at m/z 84 (Figure 4).
|
A single 9-O-lactylated-mono-O-acetylated derivative of Neu5Ac was also detected (Figure 6d) corresponding to Neu5,7Ac29Lt. This compound was characterized by a series of ions at m/z 562, 451, and 238 together with the basic ion of lactyl substituent at m/z 112.
Highly acetylated derivatives of Neu5Ac were detected as very minor compounds in the mucins from the human saliva and in horse serum glycoproteins. They included the per-acetylated sialic acid Neu4,5,7,8,9Ac5 (Figure 7b) and Neu4,5,7,9Ac4 (Figure 7a).
|
Mono- and multi-acylated derivatives of Neu5Gc
Surprisingly, no mono-O-acetylated derivative of Neu5Gc (mass 1165 Da) was identified in our analyses, although many derivatives of Neu5Gc were detected through the ion at m/z 227, in contrast with di-O-acylated compounds. The di-O-acetylated derivatives of Neu5Gc (mass 1011 Da) were characterized by ions at m/z 706. Two compounds responding to this criterion were identified as the Neu7,9Ac25Gc in BSM (Figure 7c) and Neu4,7Ac25Gc in human saliva (Figure 7d).
Three compounds possessing both the ion at m/z 227 (characteristic of the Neu5Gc family) and at m/z 112 (characteristic of the lactylated compounds) were detected (Figure 8). They included Neu8Ac5Gc9Lt (Figure 8a), Neu4Ac5Gc9Lt (Figure 8b), and the di-acetylated compounds Neu7,8Ac25Gc9Lt (Figure 8c) and Neu4,7Ac25Gc9Lt (Figure 8d).
|
Mono- and di-acetylated derivatives of Kdn
Only a few reports were concerned with the O-acetylation of Kdn, this substance being abundant only in a few species. A single O-acetylated Kdn was totally identified: Kdn9Ac (Iwasaki and Troy , 1990
; Kanamori et al., 1990; Strecker, 1997). Studies of sialic acids of the mucins of the eggs of Rana temporaria, Rana arvalis, and Pleurodeles waltii allowed the identification of several other Kdn derivatives (Figure 3 and Figure 9). Although the study of Kdn derivatives was very often difficult because of the absence of standard compounds, it was possible because the major Kdn derivatives were previously characterized using nuclear magnetic resonance as Kdn and Kdn9Ac in a molar ratio close of 1:1 (Strecker, 1997). Kdn9Ac was easily identified (Figure 10a,b), because it possessed the ion at m/z 73 characteristic of 9-O-acetylated compounds besides the ion at m/z 863 characteristic of mono-O-acetylated Kdn. Two other minor mono-O-acetylated Kdn derivatives were identified as Kdn5Ac (Figure 9c,d) and Kdn7Ac (Figure 10e).
|
|
Several di-O-acetylated Kdn were also detected as very minor compounds specifically in the eggs of R. temporaria (Figure 1) and of P. waltii, all presenting a characteristic ion at m/z 709 (Figure 10). Most of these compounds (mass 954 Da) showed a series of ions corresponding to an initial loss of mass of 51 Da (simultaneous loss of 31 (OCH3) and 20 (HF)). Kdn8,9Ac2 was easily identified through intense ions at m/z 73, 103 and 145 (Figure 10a). Kdn7,9Ac2 (Figure 10b) was identified by the ions at m/z 73 and 152. Two other di-O-acetylated Kdn derivatives were detected, corresponding to Kdn4,7Ac2 (Figure 10c) and Kdn4,5Ac2 (Figure 10d). The possibility of an acetylation in position 4 of sialic acids from the R. temporaria mucins was supported by the unambiguous detection of the very minor Neu4,5Ac2, Neu5Ac being also detected as a medium intensity constituent (Figure 1b).
Intramolecular lactones of sialic acid derivatives
Substances, reproducibly present at variable intensities depending on the samples, were identified as lactones derived from different sialic acids but not from Kdn. These compounds lacked the molecular ion [M]+°, but had a series of weak ions at M – 20 (M – HF) and M – 37 (M – [F + H2O]). Series of ions were derived from M – 18 by the additional loss of heptafluorobutyrate/and or acetate. These compounds lacked systematically the M – 76/77 and/or M – 91/92 ions characteristic of all sialic acids described above. This suggested that the carboxyl group of the C(1) carbon atom was not in the form of a methyl ester, but was involved in an intramolecular lactone. For the lactone derived from Neu5Ac (mass 879 Da), these ions corresponded to m/z 859, 842, and 814, respectively. An intense ion was derived from the M – 18 ion by the loss of a heptafluorobutyric acid group (m/z 647 for the lactone of Neu5Ac). A weak but fundamental ion for the identification of these compounds was the ion at m/z 136, corresponding to the formation of a double ring (Figure 4). Based on these criteria, 1,7-lactones derived from Neu5Ac (Figure 11a), Neu5Gc (Figure 11b), Neu5,9Ac2 (Figure 11c), and Neu4,5,9Ac3 (Figure 11d) were identified (but not exclusively) in the eggs of Bufo bufo.
|
The question remained of whether these not-yet-identified compounds were natural compounds or artefacts produced during acid hydrolysis. Evidence was provided for the formation of similar lactones using carbodiimides (Yu and Ledeen, 1969
), but not in 2 M acetic acid. Based on parallel analyses of different samples treated under the same conditions, it appeared that the 1,7 lactones were likely endogenous products. For example, the mucin of the egg of B. bufo showed the predominance of the Neu5Ac1,7-lactone, the proportion of Neu5Ac being 0.5% of the corresponding lactone. The other major sialic acids in this sample were Neu5Gc and its 1,7-lactone in equivalent amounts. In BSM, in contrast, Neu5Ac1,7-lactone was present but represented only 1% of the quantity of Neu5Ac. The Neu5Ac1,7-lactone was reproducibly not detectable in equine submandibular gland mucin (ESM) and represented less than 0.5% of Neu5Ac in ovine submandibular gland mucin (OSM), indicating that sample specificity for these compounds.
8-O-methylated sialic acids
Several samples (including the mammalian mucin OSM) showed a weak peak with an RT in between Neu5Ac and Neu5,9Ac2 endowed with a typical fragmentation pattern, most of its intense ions being absent from the other compounds. Although the molecular ion [M]+° was absent for the most frequent compound, all fragment ions could be interpreted starting from a mass of 925 amu. From a low intensity L1 ion at m/z 848, a sequence of ions was observed corresponding to sequential losses of mass of 227 amu (C(9)H2OCOCF2CF2CF3), 44 amu (C(8)HOCH3), and 226 amu (C(7)HOCOCF2CF2CF3) characteristic of a O-methyl group on the C(8) carbon atom, the C(9) and C(7) being initially nonsubstituted. It was concluded that the compound was Neu5Ac8Me (Figure 12a). Such a compound was already described in lower organisms (Kamerling and Vliegenthart, 1982) and its presence in mammals was surprising, but it could not be an artifact due to the methodology. Indeed, treatment of standard Neu5Ac with the diazomethane reagent up to 15 days at room temperature did not produce Neu5Ac8Me. Furthermore, because the O-methyl groups are stable to the most common chemical procedures used for the liberation of monosaccharides from glycoconjugates (in contrast with N- and O-acyl groups), this compound should be distinguished from the bulk of other sialic acids (including Kdn) in classical monosaccharide analyses. This point was actually verified using GC/MS analyses of the HFB derivatives obtained after methanolysis (Zanetta et al., 1999). The HFB derivative of the methyl ester of the O-methyl glycoside of the 8-O-methylated neuraminic acid (Neu) (eluted just before the major anomer of sialic acid) was actually detected, although at a low level (2% of sialic acids) in OSM, the mucin in which it was present at the same level. In the methanolysis products of other samples, this compound was absent or present at an extremely low level (0.01% of the total sialic acids) in agreement with the data obtained in the procedure reported here. Neu5,9Ac28Me (Figure 12b) was identified in the sialic acids of the skin of Anguilla anguilla. The two compounds were characterized by an ion at m/z 578 and 424, respectively, the former corresponding to the pyranic ring depleted from all its substituents except the C(7)–C(9) substituted chain. The reason such a compound was not detected before in higher animals might be its extremely low level.
|
8-O-sulfated sialic acids
Several samples presented variable abundance compounds showing a basic ion at m/z 122 and two other intense ions at m/z 295 and 365 (Figure 13a–b). Two different compounds, some of the ions differing by a mass difference of 154 (difference between HFB and acetate), were detected in BSM. The sequences of ions showed differences of 80 and/or 110, suggesting the presence of a sulfate or a phosphate group in the molecule. Based on the fragmentation patterns, the sulfate/phosphate group was on the C(8) carbon atom (Figure 13a–b). The ion at m/z 122 was a key for understanding the nature of these compounds. Indeed, it corresponded to a cyclic compound, the sulfate/phosphate group being involved in a di-ester bond with the hydroxyl groups of the C(8) and C(9) carbon atoms (Figure 4). Based on the isotopic composition of the peak at m/z 122, it was concluded that the acylating group was a sulfate and not a phosphate group. Two compounds present at a significant level in BSM were identified as Neu5Ac8S (Figure 13a) and Neu4,5Ac28S (Figure 13b). A third compound present at an extremely low level in BSM could correspond to Neu5Gc8S (not shown). Such compounds were previously identified, but always in lower organisms (Kamerling and Vliegenthart, 1982
; Slomiany et al., 1981; Bergwerff et al., 1992).
|
The ion at m/z 295 corresponded to a protonated (on the sulfate part) of a mixed anhydride between heptafluorobutyric and sulfuric acid (Figure 4). The presence of this ion asked the question of the mechanism of formation of the mixed anhydride. Previous studies of per-acetylated and of per-trifluoroacetylated monosaccharides mentioned losses of mass corresponding to the elimination of acetic and trifluoroacetic anhydrides, respectively (for review, see Kochetkov and Chizhov, 1966
). In all the compounds identified in this study, positive ions corresponding to anhydrides were absent except for the previous fragment, a proton being captured by the sulfuric acid part. This capture is not possible for the other putative anhydrides (HFBAA, acetic anhydride, mixed heptafluorobutyric–acetic anhydride, mixed glycolyl–acetic anhydride). However, losses of mass corresponding to the uncharged anhydrides were evident and provided a confirmation of the position of the substituents on the C(8) and C(9) carbon atoms. Indeed, the different compounds could be identified frequently through ions at m/z M – 50 – mass of the anhydride formed between acid substituents of C(8) and C(9) (see above). The reason that sulfated compounds were volatile remained puzzling, but it was unlikely that the methyl ester of the sulfate group formed using diazomethane methyl–esterification would resist to the acylation procedure. The actual reason could be a protective effect of the HFB groups on the C(7) and C(9) carbon atoms against formation of hydrogen bonds between the sulfate group and oxygen atoms of the liquid phase of the capillary column. The fact that only three compounds were detected (Neu5Ac8S, Neu5Gc8S, and Neu4,5Ac28S) that all had HFB groups on the C(7) and C(9) might be indicative of this protective effect, the other unprotected compounds being not volatile or interacting too strongly with the liquid phase. This assumption was sustained by the observation that the use of shorter chain fluorinated acylating agents, such as trifluoroacetic anhydride or pentafluoropropionic anhydride, could not allow the detection the previous 8-sulfated compounds.
Presence of Neu
A compound (representing 5% of Neu5Ac, the major sialic acid of OSM) was eluted at an RT between those of Neu5Ac and Neu5,9Ac2 (Figure 13c). It did not present the ions at 238 and at 227 characteristic of Neu5Ac and Neu5Gc, respectively. The two anomers (the major represented 99.5 %; Figure 13c), showed an intense ion at m/z 505, due to the pyranic ring substituted with two HFB groups. This indicated that the amino group was initially not substituted. Consequently, this compound corresponded to Neu, the amino group on the C(5) carbon atom being initially free and, consequently, transformed into the N-HFB derivative. Such a compound has not been so far identified as a free molecule, although indirect evidence was provided immunologically for its presence in gangliosides associated with cancer (Sjöberg et al., 1995). The possibility that it was a degradation product was unlikely, since this compound co-existed with O-acylated compounds, which are much less stable than N-acylated compounds. The explanation why Neu can be identified using the present method is that the heptafluorobutyric anhydride acylation is able to disrupt the Schiff base spontaneously formed between the amino group on the C(5) and the keto group on the C(2) of Neu liberated on hydrolysis, progressively and definitely blocking the amino group as its HFB derivatives.
Sialic acids with a second amino group
Several extremely minor compounds with characteristic intense ions at m/z 703, 518, 304, and 276 were specifically detected in one of the BSM preparations analyzed in this study. These compounds also had the particularity of having both the ions at 238 and 227 characteristic of derivatives of Neu5Ac and Neu5Gc, respectively. The first series of ions was 1 amu lower than for the other compounds, suggesting the presence of a second amino group in the molecule. The ion at 227 indicated the presence of 5-glycolyl groups, whereas the ion at 238 suggested a N-acetyl group in another position vicinal to a hydroxyl group substituted by a HFB group. Based on the fragmentation patterns, it was deduced that these compounds were derivatives of Neu5Gc having an N-acetyl group on the C(7) carbon atom. These compounds were Neu9Ac7Am5Gc and Neu4,9Ac27Am5Gc (Figure 12c–d), in which Am represented an acetamido group. They were likely due to a bacterial contamination of this particular BSM preparation. Indeed, similar compounds were previously identified in somatic antigens of Pseudomonas aeruginosa (Knirel et al., 1987, 1997), although the previous compound (5,7-di-acetamido-3,5,7,9-tetradeoxy-L-glycero-L-manno-nonulosonic acid; pseudaminic acid) did not contain an N-glycolyl group and O-acetyl groups were not detected.
Analysis on crude samples
This method could be applied successfully to very crude samples, like crude mucins from the eggs of frogs (or a single egg as the starting material and less than 1% injected onto the GC/MS apparatus), total saliva proteins, and so on. Because HFB derivatives are stable, the analysis could be performed without elimination of the protein or lipid part of the sample after hydrolysis. Because no purification of the liberated sialic acids is needed, a complete analysis can be performed starting routinely from nanogram amounts of total sialic acids. Major contaminants already observed were oligosaccharides derived from glucose present in the samples, but these constituents could be easily identified through their intense L1 ion at 947 (Figure 1b). This was true not only for the major constituents but also for compounds representing a few percent of the major compounds. For example, the area between RT 21.50 min and RT 24.00 min in Figure 1b showed extremely weak peaks, some of them being other sialic acids that could be identified together with contaminant oligosaccharides and phthalates. Fatty acid methyl esters or sphingoid bases derived from glycosphingolipids did not interfere with the sialic acid determination because: (1) the acetamido bonds (like the other bonds) were not cleaved significantly on the hydrolysis conditions used for sialic acid determinations; and (2) fatty acid methyl esters, when present, are, in majority, eluted after the area of separation of sialic acid derivatives.
Major contaminants actually observed were derived from phthalates. Consequently, plastic vessels were not recommended in the handling of samples for these analyses. Nevertheless, these compounds could be easily identified by their ions at 149 and the absence of ions at 69, 119, and 169, in such a way that they did not significantly interfere with the identification of the sialic acids. The other major interferences were due to methyl-siloxane compounds derived from the liquid phase but also from residues of previous analyses of TMS derivatives, intense ions at 73, but revealed by the presence of an intense ion at m/z 225, absent from HFB derivatives.
Although the present method allows the identification of most compounds, confirmations of the structures can be obtained through a complete monosaccharide analysis after cleavage of the glycosidic bonds by acid-catalyzed methanolysis followed by GC/MS analysis of the compounds as HFB derivatives (Zanetta et al., 1999). Such analyses could be performed on the same sample after analyses of the sialic acids. This allows detecting the presence of Neu (produced from Neu5Ac and Neu5Gc and of their O-acylated derivatives), Kdn derivatives, and O-alkylated derivatives. Additional confirmations can be obtained using a previous alkaline elimination of O-acyl groups (NaOH 0.1 M for 2 h at 37°C followed by neutralization with hydrochloric acid, evaporation, and formation of volatile derivatives). This allows identifying the disappearing peaks as O-acylated derivatives.
Conclusions and perspectives
This method presents important advantages compared to previously described techniques, essentially due to the stability of the HFB derivatives and the absence of derivatization of the semi-acetalic group with HFBAA. In contrast with other methods, it does not need intermediate steps of purification of sialic acids after the mild hydrolysis step. For glycolipids or oligosaccharides, the samples can be analyzed in the same vial, without elimination of the other derivatives. For glycoprotein analysis, a short-term centrifugation step is needed to eliminate the insoluble material from material with a very low sialic acid content. In fact, in most cases, the centrifugation step can be avoided, but cleaning of the GC injector should be performed frequently. Although soluble peptides and amino acids are still present, they do not interfere with the analysis. This type of contamination cannot be easily solved using shorter fluorinated anhydrides such as trifluoroacetic anhydride (Zanetta et al., 1972) and pentafluoropropionic anhydride (unpublished data).
From the MS analysis, HFB derivatives of sialic acids present the advantage of having specific ions that allow the immediate selection of these compounds in very impure samples by chromatogram reconstitution using the ion at m/z 169, characteristic of HFB derivatives. Furthermore, the ions of HFB derivatives are generally not saturating relative to the higher-mass ions, and it is possible to record spectra for all ions above 45 amu. This appears to be a significant point because important ions for the identification of some compounds have relatively low masses (ion at m/z 73 for 9-O-Ac, 84 for 4-O-Ac [except Neu4,5Ac2], 112 for 9-O-lactyl, 122 for 8-O-sulfate, etc.). Furthermore, the modalities of fragmentation of these derivatives allow a relatively easy structural assignment for most compounds.
This method is by far more sensitive than others. The first important point is that it does not need a purification step of the liberated sialic acids, which may lead to loss of sialic acids and to isomerization of O-acetyl groups. Indeed, classical techniques used a dialysis of the hydrolyzed mixture to recover sialic acids in the dialysate, followed by ion exchange chromatography. This procedure needs relatively high quantities of sialic acids (Schauer and Kamerling, 1997), even if the following analysis steps are very sensitive and specific. Routinely, 0.1–1 ng total sialic acids are injected onto the GC/MS apparatus to avoid saturation of the TIC detector. The second important point is the wide range of detection of specific ions and/or complete mass spectra (1/1000) of the MS detectors. Sialic acids present at a concentration less than 1/1000 relative to the major peaks can be unambiguously identified. Most of the extremely minor TIC peaks (most of them appear as definite peaks only examining areas not containing the major peaks) give spectra allowing the identification of these compounds.
Because purification of the liberated sialic acids (for review, see references in Schauer and Kamerling, 1997) is not needed, the method allows identifying compounds that do not respond to the expected properties of sialic acids, that is, monosaccharides with a single carboxylic acid negative charge. The presence of 1,7 lactones has been probably missed because of the absence of a negative charge in these compounds. Such compounds cannot be revealed using the reagents specific for 2-keto-carboxylic acids but appear as the corresponding sialic acid after alkaline treatment. Furthermore, lactones may have been lost during ion exchange chromatography. Similarly, Neu was not detectable because of its transformation into a Schiff base. Sulfated sialic acids should have been also missed during ion-exchange chromatography because of the strong acidic charges of these molecules. Although we expect that multi-O-acetylated 8-O-sulfated sialic acids will not be volatile, this technique allowed to clearly identifying Neu5Ac8S, Neu5Gc8S, and Neu4,5Ac28S in routine analyses.
Because of its simplicity and sensitivity, this method allows routine analyses on minute quantities of material, for example, a single frog egg. It will be possible to analyze directly human samples to evaluate if qualitative and quantitative modifications of sialic acids occur in different pathologies, especially cancer, and to analyze in different tissues the presence of putative ligands of bacterial or viral agglutinins. But it could also open a wide field of research in the field of biochemistry because it could reveal the presence of unexpected sialic acids in different organisms and tissues. This could be the basis of specific searches of new enzymatic activities, needed for the synthesis and degradation of these compounds, as already discovered in several systems. This could be also the basis for the study of new molecular and biological interactions in the function of fundamental physiological mechanisms, still unexpected because of the absence of knowledge on the existence of strange sialic acid derivatives. One recent example was the identification of the high affinity ligand of human interleukin 4, the Neu5Ac1,7-lactone (Cebo et al., 2001), a compound for the first time detected at the surface of human leukocyte using the present methodology.
The finding of new sialic acid species and the occurrence of rare types like 8-O-methylated or 8-O-sulfated, known so far only in echinoderms (Bergwerff et al., 1992; Schauer and Kamerling, 1997; Schauer, 2000), in animals higher than these, even in mammals, is a challenge for enzymologists and cell biologists, to search for the responsible enzymes and genes on the one hand and for the cell biological significance on the other. This is also valid for the sialic acid lactones identified and for the nonsubstituted Neu molecule itself. In addition to the ongoing research on the metabolism of O-acetylated sialic acids in mammals, the biosynthesis of the various O-acetylated Kdn derivatives found in amphibian needs elucidation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
Standard Neu5Ac, Neu5Gc, fetuin, and horse serum were purchased from Sigma. Neu4,5Ac2 on the one hand and Neu5,9Ac2, Neu5,7,9Ac3, on the other hand, were liberated from ESM and BSM, respectively, using propionic acid hydrolysis (Mawhinney and Chance, 1994
), followed by dialysis and cation and anion exchange chromatographies (Reuter and Schauer, 1994) and, finally, column cellulose chromatography (Schauer, 1970). DiazogenTM was from Acros. HFBAA was from Fluka, Merck, or Acros. CM-Trisacryl was from Pharmacia. ESM, BSM, and OSM were prepared according to Tettamanti and Pigman (1968). Eggs of B. bufo, R. temporaria, P. walti, Bombina bombina, Triturus alpestris, and total extracts from the skin of A. anguilla were kindly provided by Dr. G. Strecker (CNRS UMR 8576, Villeneuve d’Ascq, France).
Gas-liquid chromatography and MS
For GC/MS analysis, the gas-liquid chromatography separation was performed on a Carlo Erba GC 8000 gas chromatograph equipped with a 25 m x 0.32 mm CP-Sil5 CB low bleed/MS capillary column, 0.25 µm film phase (Chrompack France, Les Ullis, France). The temperature of the Ross injector was 260°C, and the samples were analyzed using the following temperature program: 90°C for 3 min then 5°C/min until 260°C. The column was coupled to a Finnigan Automass II mass spectrometer (mass detection limit 1000) or, for mass larger than 1000, to a Riber 10-10H mass spectrometer (mass detection limit 2000). The analyses were performed routinely in the EI mode (ionization energy 70 eV; source temperature 150°C). When necessary, detection was performed in the chemical ionization mode in the presence of ammonia (ionization energy 150 eV, source temperature 100°C). The chemical ionization detection was performed for positive ions or negative ions, in separated experiments, the latter allowing the quite specific detection of HFB derivatives with a higher sensitivity (Zanetta et al., 1999).
Cleavage of sialic acid residues from glycoproteins and glycolipids and preparation of the samples for derivatization
All experiments were performed in heavy walled Pyrex tubes with a Teflon-lined screw cap. In preliminary experiments, sialic acids were liberated from glycolipids and from glycoproteins using two different techniques: diluted formic acid at pH 2.0 for 90 min at 80°C, or 2 M acetic acid for 90 min at 80°C. Because of similar cleavage yields and of the easier elimination under vacuum of acetic acid relative to formic acid, the former was used in all experiments.
The tubes containing the glycoprotein or glycolipid samples were evaporated under vacuum at room temperature then supplemented with 2 M acetic acid (at least 2 ml/mg protein or 2 ml/100 µg glycolipid, with a maximum of 500 µl per reaction vial). The tightly closed vials were gently shaken and placed in an oven at 80°C. After 90 min, the samples were cooled, and after addition of two drops of toluene, were evaporated to dryness using a rotary evaporator (with a Teflon-lined adaptation to the reaction vessels) at room temperature. They were treated differently depending on their nature and on their content in sialic acids:
(a) For most mucins and sialylated glycolipids containing high levels of sialic acids, the dried samples were directly derivatized and analyzed without additional purification.
(b) For protein fractions containing very low levels of sialic acids, the evaporated hydrolyzed samples were dissolved into 1 ml water and left for 1 h at 4°C. The precipitate was eliminated by centrifugation at 4°C for 30 min at 4000 rpm, and the supernatant was evaporated again under vacuum at room temperature. The dry sample was submitted to derivatization.
(c) For total cell homogenates or samples containing extremely low amounts of sialic acids and glycoproteins showing a very weak precipitation, the samples were treated as above (b) and resuspended in a small volume of water (100 µl/mg initial protein). It was passed through a small column of CM-Trisacryl (200 µl of swollen gel in a Pasteur pipette closed with glass wool). The run through of the column and 500 µl additional water eluate were recovered and evaporated with a rotary evaporator as above, then derivatized as below.
Derivatization of sialic acids
The dry samples (0.1–1 µg of total sialic acid) were supplemented with 100–200 µl of anhydrous methanol (Kamerling and Vliegenthart, 1982; Fontaine et al., 1994) at the bottom of the vial (to partially dissolve sialic acids), followed by the addition of 200 µl of a diazomethane solution in ether, and the tubes were tightly closed. Diazomethane was prepared in a WheatonTM apparatus according to the procedure proposed by the manufacturer, and the diazomethane solution was kept as aliquots in tubes identical to the reaction vials at room temperature in a ventilated hood (diazomethane is a strong irritatant and carcinogenic agent; consequently drastic cares have to be taken during handling). The samples were left for 4 h at room temperature without stirring. The reagents were evaporated to dryness under a stream of nitrogen in a ventilated hood, and then supplemented with 200 µl acetonitrile and 25 µl HFBAA. The closed vials were heated for 5 min at 150°C in a sand bath, cooled at room temperature, and the samples were evaporated in a stream of nitrogen in a ventilated hood. The residue was solubilized into the required volume of acetonitrile previously dried on calcinated calcium chloride (Zanetta et al., 1999), and aliquots were injected onto the Ross injector of the GC/MS apparatus. Optimal analyses were obtained injecting 0.1–1 ng of each sialic acid derivatives in the GC/MS apparatus, but compounds could be identified safely at the picogram level. If analyses are performed on higher amounts of material, the quantities of the reagents have to be increased in proportion.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
The authors thank Gérard Strecker and co-workers for providing eggs of amphibians and Marzog el Madani for assistance in the isolation of standard sialic acids. This manuscript is dedicated to our esteemed colleague and friend Professor André Verbert, director of the CNRS UMR 8576, who died suddenly on May 20, 2000.
|
Abbreviations |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
BSM, bovine submandibular gland mucin; CI, chemical ionization; Di-Glc, dimer of Glc; EI, electron impact; ESM, equine submandibular gland mucin; GC, gas chromatography; HFB, heptafluorobutyrate; HFBAA, heptafluorobutyric anhydride; HPLC, high-performance liquid chromatography; ICH3, iodomethane; Kdn, 3-deoxy-D-glycero-D-galacto-nonulosonic acid; MS, mass spectrometry; Neu, neuraminic acid; Neu5Ac, N-acetyl neuraminic acid; Neu5Gc, N-glycolyl neuraminic acid; OSM, ovine submandibular gland mucin; RT, retention time; RTr, relative retention time; TIC, total ion count; TMS, trimethyl-silyl; Tri-Glc, trimer of Glc. The nomenclature of the different sialic acids was according to Schauer and Kamerling (1997)
.
|
Footnotes |
|---|
1 To whom correspondence should be addressed
|
References |
|---|
|
TopAbstractIntroductionResults and discussionMaterials and methodsAcknowledgmentsAbbreviationsReferences |
|---|
Bergwerff, A.A., Hulleman, S.H.D., Kamerling, J.P., Vliegenthart, J.F.G., Shaw, L., Reuter, G., and Schauer, R. (1992) Nature and biosynthesis of sialic acids in the starfish Asterias rubens. Identification of sialo-oligomers and detection of S-adenosyl-L-methionine: N-acylneuraminate 8-O-methyltransferase and CMP-N-acetylneuraminate monooxygenase activities. Biochimie, 74, 25–38.[Medline]
Cebo, C., Dambrouck, T., Maes, E., Laden, C., Strecker, G., Michalski, J.-C., and Zanetta, J.-P. (2001) Recombinant human interleukins, IL-1, IL-1ß, IL-4, IL-6 and IL-7, show different and specific calcium-independent carbohydrate-binding properties. J. Biol. Chem., 276, 5685–5691.
Fontaine, T., Fournet, B., and Karamanos, Y. (1994) A new procedure for the reduction of uronic acids containing polysaccharides. J. Microbiol. Meth., 20, 149–157.
Hara, S., Yamaguchi, M., Takemori, Y., Furuhata, K., Ogura, H., and Nakamura, M. (1989) Determination of mono-O-acetylated N-acetylneuraminic acids in human and rat sera by fluorometric high-performance liquid chromatography. Anal. Biochem., 179, 162–166.[Web of Science][Medline]
Iwasaki, M., Inoue, S., and Troy, F.A. (1990) A new sialic acid analogue, 9-O-acetyl-deaminated neuraminic acid, and alpha-2, 8-linked O-acetylated poly(N-glycolylneuraminyl) chains in a novel polysialoglycoprotein from salmon eggs. J. Biol. Chem., 265, 2596–2602.
Iwersen, M., Vandammme-Feldhaus, V., and Schauer, R. (1998) Enzymatic 4-O-acetylation of N-acetylneuraminic acid in guinea pig liver. Glycoconj. J., 15, 895–904.[Web of Science][Medline]
Kamerling, J.P., and Vliegenthart, J.F.G. (1982) Gas-liquid chromatography and mass spectrometry of sialic acids. In R. Schauer, ed., Sialic acids. Chemistry, metabolism and function. Springer Verlag, New York, pp. 95–125.
Kamerling, J.P., Verluis, C., and Schauer, R. (1975) Identification of O-cetylated N-acylneuraminic acids by mass spectrometry. Carbohydr. Res., 41, 7–17.[Web of Science][Medline]
Kanamori, A., Inoue, S., Iwasaki, M., Kitajima, K., Kawai, G., Yokoyama, S., and Inoue, Y. (1990) Deaminated neuraminic acid-rich glycoprotein of rainbow trout egg vitelline envelope. Occurrence of a novel alpha-2, 8-linked oligo(deaminated neuraminic acid) structure in O-linked glycan chains. J. Biol. Chem., 265, 21811–21819.
Klein, A., Diaz, S., Ferreira, I., Lamblin, G., Roussel, P., and Manzi, A.E. (1997) New sialic acids from biological sources identified by a comprehensive and sensitive approach: liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS) of SIA quinoxalinones. Glycobiology, 7, 421–432.
Knirel, Y.A., Kocharova, N.A., Shaskhov, A.S., Dmitriev, B.A., Kochetkov, N.K., Stanislavsky, E.S., and Mashikova, G.M. (1987) Somatic antigens of Pseudomonas aeruginosa. The structure of O-specific polysaccharide chains of the lipopolysaccharides from P. aeruginosa O5 (Lanyi) and immunotype 6 (Fisher). Eur. J. Biochem., 163, 639–652.[Web of Science][Medline]
Knirel, Y.A., Moll, H., Helbig, J.H., and Zahringer, U. (1997) Chemical characterization of a new 5, 7-diamino-3, 5, 7, 9-tetradeoxynonulosonic acid released by mild acid hydrolysis of the Legionella pneumophila serogroup 1 lipopolysaccharide. Carbohydr. Res., 304, 77–79.[Web of Science][Medline]
Kochetkov, N.K., and Chizhov, O.S. (1966) Mass spectrometry of carbohydrate derivatives. Adv. Carbohydr. Chem., 21, 39–93.
Maes, E., Florea, D., Delplace, F., Lemoine, J., Plancke, Y., and Strecker, G. (1998) Structural analysis of the oligosaccharide-alditols released by reductive beta-elimination from the oviductal mucins of Rana temporaria. Glycoconj. J., 14, 127–146.
Maes, E., Strecker, G., Timmerman, P., Leroy, Y., and Zanetta, J.-P. (1999) Quantitative cleavage of the N-glycosidic bond under the normal conditions of methanolysis used for the analysis of glycoprotein monosaccharides. Anal. Biochem., 267, 300–308.[Web of Science][Medline]
Mawhinney, T.P., and Chance, D.L. (1994) Hydrolysis of sialic acids and O-acetylated sialic acids with propionic acid. Anal. Biochem., 223, 164–167.[Web of Science][Medline]
Pons, A., Popa, J., Portoukalian, J., Bodennec, J., Ardail, D., Kol, O., Martin-Martin, M.-J., Hueso, P., Timmerman, P., Leroy, Y., and Zanetta, J.-P. (2000) Single step GC/MS analysis of glycolipid constituents as heptafluorobutyrate derivatives with a special reference to the lipid portion. Anal. Biochem., 284, 201–216.[Web of Science][Medline]
Powell, A.K., and Harvey, D.J. (1996) Stabilization of sialic acids in N-linked oligosaccharides and gangliosides for analysis by positive ion matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun. Mass. Spectrom., 10, 1027–1032.[Web of Science][Medline]
Regl, G., Kaser, A., Iwersen, M., Schmid, H., Kohla, G., Strobl, B., Vilas, U., Schauer, R., and Vlasak, R. (1999) The hemagglutinin-esterase of mouse hepatitis virus strain S is a sialate-4-O-acetylesterase. J. Virol., 73, 4721–4727.
Reuter, G., and Schauer, R. (1994) Determination of sialic acids. Methods Enzymol., 230, 168–199.[Web of Science][Medline]
Roy, R., Andersson, F.O., Harms, G., Kelm, S., and Schauer, R. (1992) Synthesis of esterase-resistant 9-O-acetylated polysialoside as inhibitor of influenza-C virus hemagglutinin. Angew. Chem. Int. Ed., 31, 1478–1481.
Schauer, R. (1970) Biosynthesis of N-acetyl-O-acetylneuraminic acids. II. Substrate and intracellular localization of bovine acetyl-coenzyme A: N-acetylneuraminate-7- and 8-acetyltransferase. Hoppe Seylers Z. Physiol. Chem., 351, 595–602.[Web of Science][Medline]
Schauer, R. (2000) Structure, function and metabolism of sialic acids. In Carbohydrates in chemistry and biology. Part II: Biology of saccharides vol. 3. Wiley-VCH, Wienheim, pp. 227–243.
Schauer, R., and Kamerling, J.P. (1997) Chemistry, biochemistry and biology of sialic acids. In Montreuil, J., Vliegenthart, J.F.G., and Schachter, H., eds., Glycoproteins II, New comprehensive biochemistry, vol. 29b. Elsevier, Amsterdam, pp. 243–402.
Schauer, R., Haverkamp, J., Wember, M., Vliegenthart, J.F.G., and Kamerling, J.P. (1976) N-acetyl-9-O-L-lactylneuraminic acid, a new acylneuraminic acid from bovine submandibular gland. Eur. J. Biochem., 62, 237–242.[Web of Science][Medline]
Sinha, D., Chatterjee, M., and Mandal, C. (2000) O-acetylated sialic acids—their detection, biological significance and alteration in diseases. Trends Glycosci. Glycotechnol., 12, 17–33.
Sjöberg, E.R., Chammas, R. Ozawa, H., Kawashima, I., Khoo, K.H., Morris, H.R., Dell, A., Tai, T., and Varki, A. (1995) Expression of de-N-acetyl-gangliosides in human melanoma cells is induced by genistein or nocodazole. J. Biol. Chem., 270, 2921–2930.
Slomiany, A., Kojima, K., Banas-Gruszka, Z., and Slomiany, B.L. (1981) Structure of a novel sulfated sialoglycosphingolipid from bovine gastric mucosa. Biochem. Biophys. Res. Commun., 100, 778–784.[Web of Science][Medline]
Strecker, G. (1997) Amphibian glycoproteins. In Montreuil, J., Vliegenthart, J.F.G., and Schachter, H. (eds.), Glycoproteins II, New comprehensive biochemistry, vol. 29b. Elsevier, Amsterdam, pp. 243–402.
Strecker, G., Wieruszeski, J.-M., Michalski, J.-C., Alonso, C., Boilly, B., and Montreuil, J. (1992) Characterization of Le(x), Le(y) and A Le(y) antigen determinants in KDN-containing O-linked glycan chains from Pleurodeles waltlii jelly coat eggs. FEBS Lett., 298, 39–43.[Web of Science][Medline]
Tettamanti, G., and Pigman, W. (1968) Purification and characterization of bovine and ovine submaxillary mucins. Arch. Biochem. Biophys., 124, 41–50.[Web of Science][Medline]
Varki, A. (1992) Diversity in the sialic acids. Glycobiology, 2, 25–40.
Varki, A. (1997) Sialic acids as ligands in recognition phenomena. FASEB J., 11, 248–255.[Abstract]
Yu, R.K., and Ledeen, R. (1969) Configuration of the ketosidic bond of sialic acid. J. Biol. Chem., 244, 1306–1313.
Zanetta, J.-P., Breckenridge, W.C., and Vincendon, G. (1972) Analysis of monosaccharides by gas-liquid chromatography of the O-methyl glycosides as their trifluoroacetate derivatives. Application to glycoproteins and glycolipids. J. Chromatogr., 69, 291–304.[Web of Science][Medline]
Zanetta, J.-P., Gouyer, V., Maes, E., Pons, A., Hémon, B., Zweibaum, A., Delannoy, P., and Huet, G. (2000) Massive in vitro synthesis of tagged oligosaccharides in 1-benzyl-2-acetamido-2-deoxy-alpha-D-galactopyranoside treated HT-29 cells. Glycobiology, 10, 565–575.
Zanetta, J.-P., Timmerman, P., and Leroy, Y. (1999) Gas-liquid chromatography of the heptafluorobutyrate derivatives of the O-methyl-glycosides on capillary columns: a method for the quantitative determination of the monosaccharide composition of glycoproteins and glycolipids. Glycobiology, 9, 255–266.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Ludwig, S. M. Theissen, M. J. Morton, and M. J. Caplan The Cytoplasmic Tail Dileucine Motif LL572 Determines the Glycosylation Pattern of Membrane-type 1 Matrix Metalloproteinase J. Biol. Chem., December 19, 2008; 283(51): 35410 - 35418. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Nelson, A. Chowdhury, J. B. Harris, Y. A. Begum, F. Chowdhury, A. I. Khan, R. C. LaRocque, A. L. Bishop, E. T. Ryan, A. Camilli, et al. Complexity of rice-water stool from patients with Vibrio cholerae plays a role in the transmission of infectious diarrhea PNAS, November 27, 2007; 104(48): 19091 - 19096. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Popa, A. Pons, C. Mariller, T. Tai, J.-P. Zanetta, L. Thomas, and J. Portoukalian Purification and structural characterization of de-N-acetylated form of GD3 ganglioside present in human melanoma tumors Glycobiology, April 1, 2007; 17(4): 367 - 373. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Robbe, C. Capon, E. Maes, M. Rousset, A. Zweibaum, J.-P. Zanetta, and J.-C. Michalski Evidence of Regio-specific Glycosylation in Human Intestinal Mucins: PRESENCE OF AN ACIDIC GRADIENT ALONG THE INTESTINAL TRACT J. Biol. Chem., November 21, 2003; 278(47): 46337 - 46348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Sonnenburg, H. van Halbeek, and A. Varki Characterization of the Acid Stability of Glycosidically Linked Neuraminic Acid. USE IN DETECTING DE-N-ACETYL-GANGLIOSIDES IN HUMAN MELANOMA J. Biol. Chem., May 10, 2002; 277(20): 17502 - 17510. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||