Dynamics of carbohydrate residues of alpha 1-acid glycoprotein (orosomucoid) followed by red-edge excitation spectra and emission anisotropy studies of Calcofluor White

Dynamics studies on Calcofluor White bound to the carbohydrate residues of sialylated and asialylated alpha 1-acid glycoprotein (orosomucoid) have been performed. The interaction between the fluorophore and the protein was found to occur preferentially with the glycan residues with a dependence on their spatial conformation. In the presence of sialylated alpha 1-acid glycoprotein, excitation at the red edge of the absorption spectrum of calcofluor does not lead to a shift in the fluorescence emission maximum (440 nm) of the fluorophore. Thus, the emission of calcofluor occurs from a relaxed state. This is confirmed by anisotropy studies as a function of temperature (Perrin plot). In the presence of asialylated alpha 1-acid glycoprotein, red-edge excitation spectra show an important shift (8 nm) of the fluorescence emission maximum of the probe. This reveals that emission of calcofluor occurs before relaxation of the surrounding carbohydrate residues occurs. Emission from a non-relaxed state means that Calcofluor molecules are bound tightly to the carbohydrate residues, a result confirmed by anisotropy studies.     

Relation between the secondary structure of carbohydrate residues of alpha1-acid glycoprotein (orosomucoid) and the fluorescence of the protein

We studied in this work the relation that exists between the secondary structure of the glycans of alpha(1)-acid glycoprotein and the fluorescence of the Trp residues of the protein. We calculated for that the efficiency of quenching and the radiative and non-radiative constants. Our results indicate that the glycans display a spatial structure that is modified upon asialylation. The asialylated conformation is closer to the protein matrix than the sialylated form, inducing by that a decrease in the fluorescence parameters of the Trp residues. In fact, the mean quantum yield of Trp residues in sialylated and asialylated alpha(1)-acid glycoprotein are 0.0645 and 0.0385, respectively. Analysis of the fluorescence emission of alpha(1)-acid glycoprotein as the result of two contributions (surface and hydrophobic domains) indicates that quantum yields of both classes of Trp residues are lower when the protein is in the asialylated form. Also, the mean fluorescence lifetime of Trp residues decreases from 2.285 ns in the sialylated protein to 1.948 ns in the asialylated one. The radiative rate constant k(r) of the Trp residues in the sialylated alpha(1)-acid glycoprotein is higher than that in the asialylated protein. Thus, the carbohydrate residues are closer to the Trp residues in the absence of sialic acid. The modification of the spatial conformation of the glycans upon asialylation is confirmed by the decrease of the fluorescence lifetimes of Calcofluor, a fluorophore that binds to the carbohydrate residues. Finally, thermal intensity quenching of Calcofluor bound to alpha(1)-acid glycoprotein shows that the carbohydrate residues have slower residual motions in the absence of sialic acid residues.     

Interaction between carbohydrate residues of alpha(1)-acid glycoprotein (orosomucoid) and progesterone. A fluorescence study

Interaction between progesterone and the carbohydrate residues of alpha(1)-acid glycoprotein was followed by fluorescence studies using calcofluor white. The fluorophore interacts with polysaccharides and is commonly used in clinical studies. Binding of progesterone to the protein induces a decrease in the fluorescence intensity of calcofluor white, accompanied by a shift to the short wavelengths of its emission maximum. The dissociation constant of the complex was found equal to 8.62 microM. Interaction between progesterone and free calcofluor in solution induces a low decrease in the fluorescence intensity of the fluorophore without any shift of the emission maximum. These results show that in alpha(1)-acid glycoprotein, the binding site of progesterone is very close to the carbohydrate residues. Fluorescence intensity quenching of free calcofluor in solution with cesium ion gives a bimolecular diffusion constant (k(q)) of 2.23 x 10(9) M(-1) s(-1). This value decreases to 0.19 x 10(9) M(-1) s(-1) when calcofluor white is bound to alpha(1)-acid glycoprotein. Binding of progesterone does not modify the value of k(q) of the cesium. Previous studies have shown that the terminal sialic acid residue is mobile, while the other glycannes are rigid [Albani, J. R.; Sillen, A.; Coddeville, B.; Plancke, Y. D.; Engelborghs, Y. Carbohydr. Res. 1999, 322, 87-94]. Red-edge excitation spectra and Perrin plot experiments performed on sialylated and asialylated alpha(1)-acid glycoprotein show that binding of progesterone to alpha(1)-acid glycoprotein does not modify the local dynamics of the carbohydrate residues of the protein.     

Effect of binding of Calcofluor White on the carbohydrate residues of alpha1-acid glycoprotein (orosomucoid) on the structure and dynamics of the protein moiety. A fluorescence study

Calcofluor White is a fluorescent probe that interacts with polysaccharides and is commonly used in clinical studies. Interaction between Calcofluor White and carbohydrate residues of alpha1-acid glycoprotein (orosomucoid) was previously studied at low and high concentrations of Calcofluor compared to that of the protein. alpha1-Acid glycoprotein contains 40% carbohydrate by weight and has up to 16 sialic acid residues. At equimolar concentrations of Calcofluor and alpha1-acid glycoprotein, the fluorophore displays free motions [Albani, J. R.; Sillen, A.; Coddeville, B.; Plancke, Y. D.; Engelborghs, Y. Carbohydr. Res. 1999, 322, 87-94], while at high concentration of Calcofluor, its surrounding microenvironment is rigid, inducing the rigidity of the fluorophore itself [Albani, J. R.; Sillen, A.; Plancke, Y. D.; Coddeville, B.; Engelborghs, Y. Carbohydr. Res. 2000, 327, 333-340]. In the present work, red-edge excitation spectra and steady-state anisotropy studies performed on Trp residues in the presence of Calcofluor, showed that the apparent dynamics of Trp residues are not modified. However, deconvoluting the emission spectra with two different methods into different components, reveals that the structure of the protein matrix has been disrupted in the presence of high Calcofluor concentrations.     

Effect of the secondary structure of carbohydrate residues of alpha 1-acid glycoprotein (orosomucoid) on the local dynamics of Trp residues

We studied in this work the relation between the secondary structure of the carbohydrate residues of alpha1-acid glycoprotein and the local motions of Trp residues of the protein. We measured for this purpose the fluorescence emission intensity and anisotropy of the Trp residues between -46 and +30 degrees of the sialylated and asialylated protein. Our results indicate that, in both forms, the global profile of the emission intensity with temperature shows that Trp residues display static and collisional interaction with the neighboring amino acids. However, the profile of the asialylated form is more structured than that observed for the sialylated protein. The Y-plot analysis of the emission-anisotropy results indicated that the frictional resistance to rotation of the surface Trp residue is less important in the sialylated protein than in the asialylated form. This result is in good agreement with the fact that, in the asialylated conformation, the carbohydrate residues are closer to the protein surface than in the sialylated form, thereby increasing the contact of the surface Trp residue with the neighboring amino acids. Also, the interaction between the carbohydrate residues and the surface Trp residue contributes to the modification of the frictional resistance to rotation of the fluorophore.     

Förster energy-transfer studies between Trp residues of alpha1-acid glycoprotein (orosomucoid) and the glycosylation site of the protein

Energy-transfer studies between Trp residues of alpha(1)-acid glycoprotein and the fluorescent probe Calcofluor White were performed. Calcofluor White interacts with carbohydrate residues of the protein, while the three Trp residues are located at the surface (Trp-160) and in hydrophobic domains of the protein (Trp-25 and Trp-122). Binding of Calcofluor to the protein induces a decrease in the fluorescence intensity of the Trp residues accompanied by an increase of that of Calcofluor White. Efficiency (E) of Trp fluorescence quenching was determined to be equal to 45%, and the F?rster distance R(o), at which the efficiency of energy transfer is 50%, was calculated to be 18.13 A. This low distance and the value of the efficiency clearly indicate that energy transfer between Trp residues and Calcofluor White is weak.     

Interaction between calcofluor white and carbohydrates of alpha 1-acid glycoprotein.

Interactions between the fluorescent probe, calcofluor white, and human serum albumin (HSA) and alpha 1-acid glycoprotein (orosomucoid) are compared. The two proteins have comparable isoelectric points, but alpha 1-acid glycoprotein is highly glycosylated (40% of glycans by weight), while the serum albumin is not. Binding of calcofluor to the proteins induces an increase in both the fluorescence anisotropy and the fluorescence intensity of the fluorophore. Also, we found that the calcofluor exhibits a fluorescence emission with a maximum located at 432, 415 or 445 nm, respectively, in the absence of proteins, in the presence of HSA, and in the presence of alpha 1-acid glycoprotein. The stoichiometries of the calcofluor-serum albumin and calcofluor-alpha 1-acid glycoprotein complexes are 2:1 and 1:1, respectively. The association constants are 0.04 and 0.15 microM-1, respectively. The calcofluor does not interact with Lens culinaris agglutinin (LCA), although the protein has a hydrophobic site. Nevertheless, one cannot exclude that the binding of the fluorophore to the HSA is nonspecific. Our results, when compared with those obtained with calcofluor dissolved in the hydrophobic solvent isobutanol, and with the fluorescent probe, potassium 6-(p-toluidino)-2-naphthalenesulfonate (TNS), bound to alpha 1-acid glycoprotein, indicate that the emission of calcofluor bound to HSA occurs from a hydrophobic state, while that of calcofluor bound to alpha 1-acid glycoprotein occurs from a hydrophilic state. The fluorescence intensity of calcofluor decreases in the presence of carbohydrates isolated from alpha 1-acid glycoprotein, while it increases in the presence of alpha 1-cellulose. Thus, calcofluor interacts mainly with the glycan moiety of alpha 1-acid glycoprotein, and its fluorescence is sensitive to the secondary structure of the glycans.     

Interaction between calcofluor white and carbohydrates of alpha 1-acid glycoprotein.

Interactions between the fluorescent probe, calcofluor white, and human serum albumin (HSA) and alpha 1-acid glycoprotein (orosomucoid) are compared. The two proteins have comparable isoelectric points, but alpha 1-acid glycoprotein is highly glycosylated (40% of glycans by weight), while the serum albumin is not. Binding of calcofluor to the proteins induces an increase in both the fluorescence anisotropy and the fluorescence intensity of the fluorophore. Also, we found that the calcofluor exhibits a fluorescence emission with a maximum located at 432, 415 or 445 nm, respectively, in the absence of proteins, in the presence of HSA, and in the presence of alpha 1-acid glycoprotein. The stoichiometries of the calcofluor-serum albumin and calcofluor-alpha 1-acid glycoprotein complexes are 2:1 and 1:1, respectively. The association constants are 0.04 and 0.15 microM-1, respectively. The calcofluor does not interact with Lens culinaris agglutinin (LCA), although the protein has a hydrophobic site. Nevertheless, one cannot exclude that the binding of the fluorophore to the HSA is nonspecific. Our results, when compared with those obtained with calcofluor dissolved in the hydrophobic solvent isobutanol, and with the fluorescent probe, potassium 6-(p-toluidino)-2-naphthalenesulfonate (TNS), bound to alpha 1-acid glycoprotein, indicate that the emission of calcofluor bound to HSA occurs from a hydrophobic state, while that of calcofluor bound to alpha 1-acid glycoprotein occurs from a hydrophilic state. The fluorescence intensity of calcofluor decreases in the presence of carbohydrates isolated from alpha 1-acid glycoprotein, while it increases in the presence of alpha 1-cellulose. Thus, calcofluor interacts mainly with the glycan moiety of alpha 1-acid glycoprotein, and its fluorescence is sensitive to the secondary structure of the glycans.     

Motions studies of the human alpha 1-acid glycoprotein (orosomucoid) followed by red-edge excitation spectra and polarization of 2-p-toluidinylnaphthalene-6-sulfonate (TNS) and of tryptophan residues.

Dynamics studies on tryptophan residues of human alpha 1-acid glycoprotein (orosomucoid) and of 2-p-toluidinylnaphthalene-6-sulfonate bound to the protein are performed. Excitation at the red edge of the absorption spectrum of the tryptophan does not lead to a shift of the fluorescence emission maximum of the fluorophore. This reveals that Trp residues present motions with respect to their microenvironment. This is confirmed by polarization studies as a function of temperature. Excitation at the red edge of the absorption spectrum of TNS leads to an important shift (15 nm) of the fluorescence emission maximum of the probe. This reveals that emission of TNS occurs before relaxation of the amino-acids dipole occurs. Emission from a non-relaxed state means that TNS molecules are bound tightly to the protein, a result confirmed by polarization studies.     

Specificity of sialytransferase: structure of alpha1-acid glycoprotein sialylated in vitro. A new methodology for the determination of the structure of the linkage formed by glycosyltransferase action on galactosyl-oligosaccharide-protein acceptors.

The terminal galactosyl units of desialylated alpha1-acid glycoprotein were selectively labeled with tritium by a galactose oxidase/NaB3H4 procedure. The 3H-labeled glycoprotein was effective as an acceptor in sialytransferase reactions catalyzed by rat liver microsomes in vitro with unlabeled CMP-N-acetyl-neuramininic acid as sialic acid donor. Permethylation/hydrolysis of glycopeptides derived from the resialylated 3H-labeled glycoprotein yielded radioactive 2,3,4-trimethylgalactose indicating that rat liver microsomes are capable of transferring sialic acid to position C-6 of the terminal galactosyl units of desialylated alpha1-acid glycoprotein. No indication was obtained for transfer of sialic acid to other positions. This result is discussed in view of the multiplicity of positions of attachment of sialic acid to galactosyl residues in native alpha1-acid glycoprotein.     

Gel filtration of sialoglycoproteins.

The role of sialic acid in the gel-filtration behaviour of sialoglycoproteins was investigated by using the separated isoenzymes of purified human liver alpha-L-fucosidase and several other well-known sialic acid-containing glycoproteins (fetuin, alpha1-acid glycoprotein, thyroglobulin and bovine submaxillary mucin). For each glycoprotein studied, gel filtration of its desialylated derivative gave an apparent molecular weights much less than that expected just from removal of sialic acid. For the lower-molecular-weight glycoproteins (fetuin and alpha1-acid glyocprotein), gel filtration of the sialylated molecules led to apparent molecular weights much larger than the known values. The data indicate that gel filtration cannot be used for accurately determining the molecular weights of at least some sialoglycoproteins.     

Use of radioactive glucosamine in the perfused rat liver to prepare alpha 1-acid glycoprotein (orosomucoid) with 3H- or 14C-labelled sialic acid and N-acetylglucosamine residues.

1. A method was developed whereby [1-14C]glucosamine was used in a perfused rat liver system to prepare over 2 mg of alpha 1-acid glycoprotein with highly radioactive sialic acid and glucosamine residues. 2. The liver secreted radioactive alpha 1-acid glycoprotein over a 4-6 h period, and this glycoprotein was purified from the perfusate by chromatography on DEAE-cellulose at pH 3.6. 3. The sialic acid on the isolated glycoprotein had a specific radioactivity of 3.1 Ci/mol, whereas the glucosamine-specific radioactivity was 4.3 Ci/mole. The latter amino-sugar residues on the isolated protein were only 13-fold less radioactive than the initially added [1-14C]glucosamine. Orosomucoid with a specific radioactivity of 31.3 microCi/mg of protein was obtainable by using [6-3H]glucosamine. 4. The amino acid composition of the purified orosomucoid was comparable with that found by others for the same glycoprotein isolated from rat serum. A partial characterization of the carbohydrate structure was done by sequential digestion with neuraminidase, beta-D-galactosidase and beta-D-hexosaminidase. 5. Many other radioactive glycoproteins were found to be secreted into the perfusate by the liver. Thus this experimental system should prove useful for obtaining other serum glycoprotein with highly radioactive sugar moieties.     

Separation of anionic oligosaccharides by high-performance liquid chromatography

We have developed methods for rapid fractionation of anionic oligosaccharides containing sulfate and/or sialic acid moieties by high-performance liquid chromatography (HPLC). Ion-exchange HPLC on amine-bearing columns (Micropak AX-10 and AX-5) at pH 4.0 is utilized to separate anionic oligosaccharides bearing zero, one, two, three, or four charges, independent of the identity of the amnionic moieties (sulfate and/or sialic acid). Ion-exchange HPLC at pH 1.7 allows separation of neutral, mono-, di-, and tetrasialylated, monosulfated, and disulfated oligosaccharides. Oligosaccharides containing three sialic acid residues and those bearing one each of sulfate and sialic acid, however, coelute at pH 1.7. Since the latter two oligosaccharide species separate at pH 4.0, analysis at pH 4.0 followed by analysis at pH 1.7 can be utilized to completely fractionate complex mixtures of sulfated and sialylated oligosaccharides. Ion-suppression amine adsorption HPLC has previously been shown to separate anionic oligosaccharides on the basis of net carbohydrate content (size). In this study we demonstrate the utility of ion-suppression amine adsorption HPLC for resolving sialylated oligosaccharide isomers which differ only in the linkages of sialic acid residues (alpha 2.3 vs alpha 2.6) and/or location of alpha 2,3- and alpha 2,6-linked sialic acid moieties on the peripheral branches of oligosaccharides. These two methods can be used in tandem to separate oligosaccharides, both analytically and preparatively, based on their number, types, and linkages of anionic moieties.     

The effect of partial deglycosylation on the structure of alpha1-acid glycoprotein

Changes in structure of alpha1-acid glycoprotein were followed after deglycosylation with neuraminidase, peptide N-glycohydrolase F or with a mixture of exoglycosidases. Partially deglycosylated preparations of alpha1-acid glycoprotein free of sialic acids, one complete saccharide component, sialic acids and one saccharide component and sialic acids and some of the external saccharides were obtained. The effect of these changes in saccharide components on the glycoprotein structure was studied by temperature perturbation difference spectroscopy, fluorescence spectroscopy, fourth-derivative of absorption spectra and spectra of CD. Partial deglycosylation resulted in transformation of the molecule to a more compact state in which phenylalanyl residues were even more buried, tyrosyl residues became more uniform and tryptophyl residues were less exposed. The content of ordered secondary structures decreased. The thermal stability of the molecule was not significantly affected. Removal of one of the five saccharide components from the native molecule had apparently deeper effect than total desialyzation of the glycoprotein.     

Tertiary structure of human alpha1-acid glycoprotein (orosomucoid). Straightforward fluorescence experiments revealing the presence of a binding pocket

Binding of hemin to alpha1-acid glycoprotein has been investigated. Hemin binds to the hydrophobic pocket of hemoproteins. The fluorescent probe 2-(p-toluidino)-6-naphthalenesulfonate (TNS) binds to a hydrophobic domain in alpha1-acid glycoprotein with a dissociation constant equal to 60 microM. Addition of hemin to an alpha1-acid glycoprotein-TNS complex induces the displacement of TNS from its binding site. At saturation (1 hemin for 1 protein) all the TNS has been displaced from its binding site. The dissociation constant of hemin-alpha1-acid glycoprotein was found equal to 2 microM. Thus, TNS and hemin bind to the same hydrophobic site: the pocket of alpha1-acid glycoprotein. Energy-transfer studies performed between the Trp residues of alpha1-acid glycoprotein and hemin indicated that efficiency (E) of Trp fluorescence quenching was equal to 80% and the F?rster distance, R0 at which the efficiency of energy transfer is 50% was calculated to be 26 A, revealing a very high energy transfer.     

Studies on incorporation of mannose into rat alpha 1-acid glycoprotein: evidence for precursor forms in rough membranes

Rats were given pulse injections of D-[14C]mannose and were killed at various times up to 60 min after injection. Rough, smooth, and Golgi fractions were prepared from liver, and alpha 1-acid glycoprotein was isolated from Lubrol extracts of the fractions. The kinetics of incorporation of D-[14C]mannose into total protein, Lubrol protein, and alpha 1-acid glycoprotein showed that proteins associated with rough fractions had particularly high specific radioactivities at early times of incorporation. One explanation for the kinetic data is that glycoproteins contain a high mannose content at early times of assembly of oligosaccharide chains. This idea was confirmed in the case of alpha 1-acid glycoprotein by isolation of a high mannose containing precursor species of alpha 1-acid glycoprotein from rough fractions of liver. This species contained 56 residues of hexose (mainly mannose) compared with 35 residues of hexose (roughly equal amounts of mannose and galactose) which are found in the native protein. It is proposed that the high mannose precursor is a form of alpha 1-acid glycoprotein that exists at an early stage in assembly of the glycoprotein and which contains largely unprocessed carbohydrate chains. In addition, evidence is presented from amino acid analyses and gel electrophoresis of the high mannose precursor and another fraction from which it is formed by limited tryptic treatment, that pro-forms of alpha 1-acid glycoprotein with extensions of the polypeptide chain may also exist.     

The specificity of viral and bacterial sialidases for alpha(2-3)- and alpha(2-6)-linked sialic acids in glycoproteins

The anomeric specificity of six sialidases (Vibrio cholerae, Arthrobacter ureafaciens, Clostridium perfringens, Newcastle disease virus, fowl plague virus and influenza A2 virus sialidases) was assessed with sialylated antifreeze glycoprotein, ovine submandibular gland glycoprotein and alpha 1-acid glycoprotein, resialylated specifically in alpha(2-3) or alpha(2-6) linkage with N-acetylneuraminic acid or N-glycolylneuraminic acid using highly purified sialyltransferases. The rate of release of sialic acid from these substrates was found to correlate well with the specificity observed earlier with the same sialidases using small oligosaccharide substrates, i.e., alpha(2-3) glycosidic linkages are hydrolyzed faster than alpha(2-6) linkages, with the exception of the enzyme from A. ureafaciens. Sialidase activity was higher with N-acetylneuraminic acid when compared with N-glycolylneuraminic acid. The studies also showed that the core oligosaccharide and protein structure in glycoproteins may influence the rate of release for different glycosidic linkages.     

Detailed structural features of glycan chains derived from alpha1-acid glycoproteins of several different animals: the presence of hypersialylated, O-acetylated sialic acids but not disialyl residues

We analyzed carbohydrate chains of human, bovine, sheep, and rat alpha1-acid glycoprotein (AGP) and found that carbohydrate chains of AGP of different animals showed quite distinct variations. Human AGP is a highly negatively charged acidic glycoprotein (pKa = 2.6; isoelectic point = 2.7) with a molecular weight of approximately 37,000 when examined by matrix-assisted laser-desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) and contains di-, tri-, and tetraantennary carbohydrate chains. Some of the tri- and tetraantennary carbohydrate chains are substituted with a fucose residue (sialyl Lewis x type structure). In sheep AGP, mono- and disialo-diantennary carbohydrate chains were abundant. Tri- and tetrasialo-triantennary carbohydrate chains were also present as minor oligosaccharides, and some of the sialic acid residues were substituted with N-glycolylneuraminic acid. In rat AGP, very complex mixtures of disialo-carbohydrate chains were observed. Complexity of the disialo-oligosaccharides was due to the presence of N, O-acetylneuraminic acids. Triantennary carbohydrate chains carrying N,O-acetylneuraminic acid were also observed as minor component oligosaccharides. We found some novel carbohydrate chains containing both N-acetylneuraminic acid and N-glycolylneuraminic acid in bovine AGP. Interestingly, triantennary carbohydrate chains were hardly detected in bovine AGP, but diantennary carbohydrate chains with tri- or tetrasialyl residues were abundant. Furthermore the major sialic acid in these carbohydrate chains was N-glycolylneuraminic acid. It should be noted that these sialic acids are attached to multiple sites of the core oligosaccharide and are not present as disialyl groups.     

The effect of neuraminidase on genetic variants of alpha anitrypsin.

Sera of Pi types M, F, S, Z, IM, FM, MS, and MZ were incubated with neuraminidase and the reaction products followed by electrophoresis. The alpha1 antitrypsin components showed a series of changes in mobility as sialic residues were removed. Removal of sialic acid was confirmed by chemical assay. Results of studies with two different electrophoretic systems suggested that the Z type alpha1 antitrypsin has less sialic acid than the M, F, and S types. There was no evidence that other genetic variants have a reduced sialic acid content. The two major bands of alpha1 antitrypsin seen in certain electrophoretic systems may reflect a difference of one sialic acid residue. It is proposed that the Z protein lacks a carbohydrate chain with two terminal sialic acid residues. This carbohydrate deficiency results in lack of secretion of type Z alpha1 antitrypsin from the endoplasmic reticulum, perhaps because of binding to sites specific for the incomplete glycoprotein or because of aggregation of the Z asialo protein. A carbohydrate chain could be prevented from attaching to the Z type either because of a conformational change or because of the replacement of a carbohydrate-binding asparagine residue in the Z protein.     

Studies on the carbohydrate moiety of α1-acid glycoprotein (orosomucoid) by using alkaline hydrolysis and deamination by nitrous acid

Alkaline hydrolysis followed by deamination with nitrous acid was applied for the first time to a glycoprotein, human plasma alpha(1)-acid glycoprotein (orosomucoid). This procedure, which specifically cleaves the glycosaminidic bonds, yielded well-defined oligosaccharides. The trisaccharides, which were obtained from the native protein, consisted of a sialic acid derivative, galactose and 2,5-anhydromannose. The linkage between galactose and 2,5-anhydromannose is most probably a (1-->4)-glycosidic bond. A hitherto unknown linkage between N-acetylneuraminic acid and galactose was also established, namely a (2-->2)-linkage. The three linkages between sialic acid and galactose described in this paper appear to be about equally resistant to mild acid hydrolysis. The disaccharide that was derived from the desialized glycoprotein consisted of galactose and 2,5-anhydromannose. Evidence was obtained for the presence of a new terminal sialyl-->N-acetylglucosamine disaccharide accounting for approximately 1mol/mol of protein. The presence of this disaccharide may explain the relatively severe requirements for the complete acid hydrolysis of the sialyl residues. The present study indicates that alkaline hydrolysis followed by nitrous acid deamination in conjunction with gas-liquid chromatography will afford relatively rapid determination of the partial structure of the complex carbohydrate moiety of glycoproteins.