the journal of vol. 268, no. 8, 15, pp. 5783-5789 …sodium salt in 0.1 m carbonate buffer, ph 9.6,...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 8, Issue of March 15, pp. 5783-5789 1993 Printed in C? S. A.

Sugar Chains of Serum Transferrin from Patients with Carbohydrate Deficient Glycoprotein Syndrome EVIDENCE OF ASPARAGINE-N-LINKED OLIGOSACCHARIDE TRANSFER DEFICIENCY*

(Received for publication, September 14, 1992)

Katsuko YamashitaS, Hiroko Ideo, Takashi Ohkura, Keiko Fukushima, Isao YuasaB, Kousaku Ohnoll, and Kenzo Takeshitall From the Department of Biochemistry, Sasaki Institute, Kanda-Surugadai, Chiyoda-ku, Tokyo 101 and the §Department of Legal Medicine and 7lDivision of Child Neurology, Institute of Neurological Science, Tottori University School of Medicine, Nishi-machi, Yonago 683, Japan

The structure of over 93% of the sugar chains of serum transferrin purified from three patients with carbohydrate-deficient glycoprotein (CDG) syndrome was Neu5Ac(~2+6Ga1~1+4GlcNAc/31.-r 2Mana1+6 (Neu5Ac(r2+6Ga1/3l+4GlcNAc/31~ 2Mancrl+3)Man~1+4G1cNAc~l+4GlcNAc, similar to that in a healthy control. On chromatofocusing, CDG syndrome transferrin was separated into three major isoforms, S4, Sz , and So, containing 4, 2, and 0 sialic acids/molecule at pH 5.12 (5.16), 5.42, and 5.80, respectively. On 7.5% SDS-polyacrylamide gel electrophoresis, the molecular masses of transferrin isoforms S4, S2, and SO were 80, 77, and 74 kDa, respectively. Transferrin isoforms S4 and S2 were linked to 2 and 1 mol of sialylated biantennary sugar chainttransferrin molecule, on the other hand, isoform So was not linked to any asparagine-N-linked oligosaccharide. Accord- ingly, CDG syndrome can be concluded to be an asparagine-N-linked oligosaccharide transfer deficiency, although the primary deficient enzyme has not yet been determined.

The carbohydrate-deficient glycoprotein (CDG)’ syndrome, which was first reported by Jaeken et al. in 1984 (l), is an autosomal recessive congenital disorder with severe nervous system involvement, growth retardation, and hepatopathy during infancy (2). A decrease in tetrasialo-transferrin com- pensating for the increases in the di- and asialo-transferrin isoforms is known as a biochemical character (3). Because

* A part of this work was supported by a grant-in-aid for Cooper- ative Research from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. The abbreviations used are: CDG, carbohydrate-deficient glyco-

protein; Neu5Ac, N-acetylneuraminic acid Gal, galactose; GlcNAc, N-acetylglucosamine; Man, mannose; BSA, bovine serum albumin; ConA, concanavalin A; RCA-I, Ricinus communis agglutinin-I; TJA- I, Trichosanthes japonica agglutinin I; AAL, Aleuria aurantia lectin; DSA, Datura stramonium agglutinin; subscript OT, NaB3H4-reduced oligosaccharide; PBS, phosphate-buffered saline; BisTris, bis(2-hy- droxyethyl)iminotris(hydroxymethyl)methane; ELISA, enzyme- linked immunosolvent assay; PAGE, polyacrylamide gel electrophoresis. All sugars mentioned in this paper have the D-configuration except for fucose, which has the L-configuration.

similar phenomena were observed for other serum glycoproteins, this syndrome is suspected to result from a defect in the processing mechanism for asparagine-N-linked sugar chains. However, the mechanism has not yet been determined. We have recently detected three patients with CDG syndrome in two unrelated families, based on the observation of clinical homology, and increases in carbohydrate-deficient transferrin and a,-antitrypsin (4).

In order to elucidate the biochemical mechanism underlying CDG syndrome in this study, serum transferrin molecules from the three patients with CDG syndrome were purified, their characteristics were investigated by means of isoelectric focusing, SDS-polyacrylamide gel electrophoresis, chromatofocusing, and lectin affinity chromatography, and structural studies of sugar chains linked to these transferrin molecules were performed.

MATERIALS AND METHODS

Chemicals, Lectins, and Enzymes-NaB3H4 (490 mCi/mmol) was purchased from Du Pont-New England. a-Methyl- manno no side, lac- btose, and Arthrobacter ureafaciens sialidase (5) were purchased from Nacalai Tesque Inc., Kyoto, Japan. Bio-Gel P-4 (minus 400 mesh) was purchased from Bio-Rad. Concanavalin A (ConA)-Sepharose, cyanogen bromide (CNBr)-activated Sepharose 4B, Sephadex G-100, Mono P HR 5/20 column (0.5 cm inner diameter X 19.6 cm long), and polybuffer 74 were obtained from Pharmacia LKB Biotechnology Inc., Uppsala, Sweden. Diplococcal &galactosidase and p-N-acetylhexosaminidase were purified from culture fluid of Diplococcus pneu- monia according to the method of Glasgow et al. (6). Ricinus communis agglutinin-I (RCA-I)-agarose and Macckia amurensis lectin (MAL) were purchased from Hohnen Oil Co., Ltd., Tokyo. Aleuria aurantia lectin (AAL)-Sepharose (7 mg/ml gel) was kindly provided by Dr. Kochibe of Gunma University. Datura stramonium agglutinin (DSA)- Sepharose (3 mg/ml gel), MAL-Sepharose (10 mg/ml gel), and Tri- chosanthes japonica agglutinin-I (TJA-I)-Sepharose (3 mg/ml gel) were prepared according to the methods described in the previous papers (7 and 8).

Oligosaccharides-Neu5Aca2-+6Galpl-t4GlcNAc~l+2Manal+ 6(Neu5Aca2+3Gal~l+4GlcNAc~l~4(Neu5Aca2+6Gal@l+ 4GlcNAc~1-*2)Manal -3~Man~l+4GlcNAcpl -*4GlcNAco~ (Neu5Acs. Gals. GlcNAcs. Mans. GlcNAc . GIcNAcoT), Neu5Aca2+ 6Gal~1~4GlcNAc/31-+2Mancu1+6(Neu5Aca2+6Ga1~1+ 4GlcNAc/31+2Manal+3)Man~l+4GlcNAc/31+4GlcNAcoT (Neu5Acz. Gal,. GlcNAc,. Man3. GlcNAc. GICNAC~T), Neu5Aca2+ 6Gal@1-+4GlcNAc~l-+2Mancu1”+6(Gal~1-+4GlcNAc~1+ 2 M a n ~ u l - + 3 ) M a n p l + 4 G 1 c N A c ~ l ~ 4 G l c N A c ~ ~ (Neu5Ac. Gal,. GlcNAc,. Man3. GlcNAc. GlcNAco~), and Gal@1+ 4GlcNAc~l+2Mancul+6(Galpl -*4GlcNAc~l+2Mana1+3) Man~l+4GlcNAcpl+4G1cNAcoT (Galp. GlcNAc2. Man3. GlcNAc. GlcNAco~) were obtained from ceruloplasmin by hydrazinolysis followed by reduction with NaB3H4 (9). Mana14(Mancul+3)Manpl+ 4GlcNAc/31+4GlcNAc0~(Man~. GlcNAc . GlcNAco~) was prepared

5783

N-Linked Oligosaccharide Transfer Deficiency 5784

A kDa

B

I -205 - -116 - -97.4

pl - -3.50

- -455

1 2 3 4 5 6 7 1 2 3 4 5 6 7

FIG. 1. SDS-7.5% homogenous polyacrylamide gel electrophoresis ( A ) and isoelectric focusing ( B ) of the purified transferrin samples. Lanes 1-6 in A and B healthy control ( I ) , father of patients 1 and 2 (2), mother of patients 1 and 2 (3) , patient 1 ( 4 ) , patient 2 (5), and patient 3 (6) ; lane 7 in A, molecular markers (Pharmacia): myosin (205 kDa), 8-galactosidase (116 kDa), phospho- rylase b (97.4 kDa), bovine serum albumin (66 kDa) and ovalbumin (43 kDa). Lane 7 in B PI marker proteins (Pharmacia): amylogluco- sidase (3.5), trypsin inhibitor (4.551, &lactoglobulin A (5.20), bovine carbonic anhydrase B (5.85), human carbonic anhydrase B (6.55), and lentil lectins (8.15,8.45,8.65). The electrophoretograms in A and B were Coomassie Brilliant Blue and silver-stained, respectively.

from Galz. GlcNAcz Man3. GlcNAc GlcNAcm by digestion with diplococcal &galactosidase and 8-N-acetylhexosaminidase.

Synthesis of Immobilized Transferrin and Anti-human Transferrin Antibody-42 mg of transferrin (Sigma) in 0.1 M carbonate buffer containing 0.5 M NaCI, pH 8.3, was reacted with 4.6 ml of CNBr- activated Sepharose 4B at 4 "C overnight. The remaining active sites of the Sepharose 4B were blocked by incubation with 0.3 M glycine, pH 8.0.8.9 mg of transferrin was bound per 1 ml of the gel. Sequen- tially, the 4.6 ml of transferrin-Sepharose 4B was packed into a column and 7 ml of anti-human transferrin goat antiserum (Guildhay Antisera Ltd., Guildford, United Kingdom) was applied to the column after equilibration with 10 mM Tris buffer, pH 7.0, containing 0.15 M NaCl (TBS). After washing the column extensively, the anti-transferrin antibody was eluted with 0.3 M glycine-HC1, pH 2.5, containing 0.15 M NaCl, followed by neutralization with 1 M Tris-HCI, pH 8.0. Yield 12 mg. 10 mg of affinity-purified anti-transferrin antibody was immobilized on 5.5 ml of the gel.

Enzyme-linked Immunosoluent Assay (ELISA) for Transferrin- The affinity-purified serum transferrin antibody was dissolved in 20 mM sodium carbonate, pH 9.6, at the concentration of 5 pg/ml. A 100-pl aliquot was applied to each well of microtiter plates (Corning Inc., New York) and then allowed to be adsorbed at 4 "C for 16 h. The plates were then rinsed three times with 400 pl of 10 mM phosphate buffer, pH 7.2, containing 0.15 M NaCl and 0.05% Tween 20 (PBS-Tween) per well, and then blocked with 200 pl of 1% bovine serum albumin (BSA) in 20 mM sodium carbonate, pH 9.6, a t 37 "C for 1 h. Each sample was diluted with 0.1% BSA in PBS-Tween and then added to the transferrin antibody-coated plates. After incubation for 40 min at 37 "C, the plates were washed four times with 400 pl of PBS-Tween/well. Goat anti-human transferrin IgG conjugated with alkaline phosphatase (The Binding Site Ltd., Birmingham, United Kingdom) was added as the second antibody. After 30 min at 37 "C, the enzyme substrate, 6.7 pmol of p-nitrophenylphosphoric acid di- sodium salt in 0.1 M carbonate buffer, pH 9.6, was added. The reaction mixture was incubated at room temperature for 30 min. After stopping the reaction with 100 pl of 1 M NaOH, the released chromogen was measured spectrophotometrically with a spectrometer (EIA Reader, Bio-Rad model 3550).

Sera-Sera were collected from the three patients as described in the previous paper (4) and stored frozen until use. Patients 1 and 2 were a pair of siblings. The three patients were diagnosed as having carbohydrate-deficient glycoprotein syndrome based on the distinc- tive clinical features and multiple carbohydrate-deficient glycoproteins.

Purification of Serum Transferrin-Each serum sample (1.8 ml) was applied to a column of Sephadex G-100 (1.5 inner diameter X 70 cm long) that had been equilibrated with 10 mM Tris-buffered saline, pH 7.0, containing0.02% NaN3 (TBS-NaN3). The transferrin positive fractions obtained on ELISA were pooled and then sequentially applied to the anti-transferrin Sepharose 4B column (4.3 ml). After

A 1 2 3 t t t

a 0

0 m

.- a

(4 0 10 20 (+I D i s t a n c e f r o m O r i g i n ( c m )

FIG. 2. Paper electrophoresis of tritium-labeled oligosaccharides released from the serum transferrin samples. 1 , Neu5Ac. Gal,. GlcNAcz. Man3. GlcNAc . GlcNAcoT; 2, Neu5Acz. Gall- GlcNAc~.Man~.GlcNAc.GlcNAcm; 3, Neu5Ac3.Ga13.GlcNAc3. Man,.GlcNAc.GlcNAcm. A, healthy control; B, father of patients 1 and 2; C, mother of patients 1 and 2; D, patient 1; E, patient 2; F, patient 3.

the column had been extensively washed with TBS-NaN3, elution was carried out with 0.3 M glycine HCI, pH 2.5, containing 0.15 M NaCl, followed immediately by neutralization with 1 M Tris-HC1, pH 8.0. Fractions positive for transferrin eluted from each column were pooled and concentrated with Immersible CX-30 (Japan Millipore Ltd., Tokyo) and then the buffer was changed to the starting buffer for chromatofocusing.

Separation of Transferrin Isoforms-The transferrin isoforms were separated on a Mono P HP5/20 column (0.5 cm inner diameter X 9.6 cm long) connected to a Pharmacia fast protein liquid chromatograph (FPLC delivery system) at 4 "C. After the column had been equilibrated with 0.025 M BisTris, pH 6.65, containing 25 pM FeCI3, about 1.0 mg of each sample in 0.7 ml of the same buffer was injected onto the column. Elution was carried out isocratically with 10 times-diluted polybuffer 74, pH 5.0, containing 25 p~ FeCI3 at the flow rate of 0.5 ml/min. The effluent was monitored as to the absorbance at 280 nm, and transferrin was assayed by ELISA.

SDS-Polyacrylamide Gel Electrophoresis and Isoelectric Focusing- Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- PAGE) was performed under reducing conditions using 7.5% homogenous gel. The electrophoresis buffer strips contained 0.2 M Tris, 0.55% SDS, pH 8.1. The gels contained 0.112 M acetate, 0.112 M Tris, pH 6.4. Isoelectric focusing was performed using PhastGel isoelectric focusing 4-6.5 (Pharmacia LKB Biotechnology, Inc.). The PhastGel isoelectric focusing media were homogeneous polyacrylamide gels (5% T, 3% C) containing Pharmalyte carrier ampholytes.

Release of Asparagine-linked Sugar Chains from the Transferrin Samples-Approximately 100-300 pg of each sample was suspended in 0.3 ml of anhydrous hydrazine and then heated at 100 "C for 8 h as reported previously (10). After N-acetylation, the oligosaccharides released from 100 pg of each sample were reduced with NaB3H4. Lactose (5 nmol) was added to each fraction as an internal standard before the reduction. ["HILactitol and tritium-labeled oligosaccharides were separated by paper chromatography. Based on the specific activity of the NaB3H4 determined from the radioactivity incorporated into lactitol and the molecular mass (80 kDa) of transferrin, the number of sugar chains released from 1 mol of each sample was calculated.


TABLE I Fractions obtained on serial lectin column chromotography, their structures, and the percent molar ratios of desialylated oligosaccharides

released from the six transferrin samples

Healthy Patients and Patient 1 Patient 2 control Mother Father (sister) (brother) Patient Structure (fractions")

% molar ratio

ConA+ AAL- DSA- Gal~l+4GlcNAc~1+2Mana1

I 6

3 Manpl4Rlb

7 Gal@1+4GlcNAc~l+2Manal

Gal~l+4GlcNAc/3l+2Manal ConA+ AAL+ DSA-

I 6

3 M a n p l 4 W

7 Gal~1+4GlcNAc/~'l+2Manal

Galpl+4GlcNAcDl ConA- AAL- DSA+

I 6 M a n a l I

2 6

Galpl+4GlcNAc@l 3 Galpl+4GlcNAcfll+ ZManal/1

Galjfl4GlcNAcpl

7 Man@l+4R1

ConA- AAL+ DSA+

I 6 M a n a l I

2 6

Gal@1+4GlcNAc,91 3 GalB1+4GlcNAcBl+ 2Manal7

0.3 0.2 0.2 0.1 0.2 0.2

7 Manpl+4Rp

96.0 96.1 96.9 97.1 97.5 96.3

2.6 2.8 1.9 1.7 1.5 2.8

1.1 0.9 1.0 1.1 0.8 0.7

Behavior on lectin columns. * R1, GlcNAc@14GlcNAcoT. Rl, GlcNAcpl4(Fucal~)GlcNAcoT.

Affinity Chromatography on Immobilized Lectin Columns-Trit- ium-labeled oligosaccharides derived from transferrin, dissolved in 100 pl of 10 mM Tris-HC1 buffer, pH 7.4, containing 0.02% NaN3 (TB-NaN3), were applied to a ConA-Sepharose (12 mg/ml) column (1 ml), MAL-Sepharose (10 mg/ml) column (1 ml), TJA-I-Sepharose (3 mg/ml) column (1 ml), RCA-I-agarose (4.8 mg/ml) column (1 ml), DSA-Sepharose (3 mg/ml) column, or AAL-Sepharose (7 mg/ml) column. After standing at 4 "C for 15 min, the columns were eluted with 5 ml of the same buffer a t 4 "C, followed by 5 ml of the buffer containing the respective haptens at room temperature. The samples that bound to the ConA-Sepharose, MAL-Sepharose, TJA-I-Sepha- rose, RCA-I-agarose, DSA-Sepharose, and AAL-Sepharose columns were eluted with TB-NaN3 containing 5 mM a-methyl-D-mannoside, 0.4 M lactose, 0.1 M lactose, 10 mM lactose, 1% N-acetylglucosamine oligomers, and 5 mM L-fucose, respectively (11).

Glycosidase Digestion-Radioactive oligosaccharides (0.1-2 nmol) were digested in one of the following reaction mixtures at 37 "C for 17 h: sialidase digestion, 50 milliunits of enzyme in sodium acetate buffer, pH 5.0 (20 pl); digestion with a mixture of diplococcal /3- galactosidase and diplococcal P-N-acetylhexosaminidase, 2 milliunits of @-galactosidase and 5 milliunits of @-N-acetylhexosaminidase in 0.2 M citrate phosphate buffer, pH 5.5 (20 pl).

Other Analytical Methods-Descending paper chromatography was performed with 1-butanol/ethanol/water (4:l:l) as the solvent. High- voltage paper electrophoresis was performed with pyridine/acetate buffer, pH 5.4 (pyridine/acetic acid/water, 3:1:387) a t a potential of

73 V/cm for 90 min. Radiochromatoscanning was performed with a Raytest Radiochromatogram Scanner, model RITA-90. Bio-Gel P-4 (under 400 mesh) column chromatography (1.5 cm inner diameter X 50 cm long) was performed as reported previously (12). Radioactivity was determined with a Beckman liquid scintillation spectrometer, model LS-6OOOLL.

RESULTS

Purification of Serum Transferrin-Serum transferrin samples from the three patients with carbohydrate-deficient glycoprotein syndrome (4), the mother and father of patients 1 and 2, and a healthy control were purified from 1.8 ml of the respective sera by Sephadex G-100 column chromatography followed by affinity chromatography on anti-transferrin- Sepharose (data not shown). The yields of transferrin from the respective sera were 1.8 mg (patient l ) , 2.9 mg (patient 2), 3.3 mg (patient 3), 2.5 mg (father), 2.9 mg (mother), and 3.1 mg (healthy control).

SDS-PAGE and Isoelectric Focusing of Serum Transfer- rin-Fig. 1A shows the results of SDS-polyacrylamide gel electrophoresis of the six transferrin samples saturated with iron. The transferrins derived from the healthy control and the parents of patients 1 and 2 gave single bands correspond-

5786 N-Linked Oligosaccharide Transfer Deficiency

IA A

0 m 0 '0 m K

.-

Elution Volume (ml)

FIG. 3. Bio-Gel P-4 column chromatography of the ConA+ AAL- DSA- fractions of oligosaccharides in Fig. 2 digested with sialidase. A-F, same as in Fig. 2. Peak a, Gal,. GlcNAc,. Man,. GlcNAc .GlcNAco~; b, ManB. GlcNAc. GlcNAcoT. The arrows at the top of the figure indicate the elution positions of glucose oligomers (the numbers indicate the glucose units).

ing to an apparent molecular mass of 80 kDa on reduction, respectively (Fig. lA, lanes 1 3 ) , on the other hand, those from patients 1-3 were resolved into three bands corresponding to different molecular weights (Fig. lA, lanes 4-6). The two additional bands were smaller by about 3 and 6 kDa than that in the healthy control. Fig. 1B shows the isoelectric points of the iron-saturated transferrins (lanes 1-6) after isoelectric focusing in a pH gradient of 4.0-6.5. The iron- saturated transferrin samples from the healthy control and the parents of patients 1 and 2 gave a prominent band a t PI 5.29 (Fig. lB, lanes 1-3), whereas those of the three patients showed not only the band at PI 5.29 but also unusual bands at PI 5.54 and 5.82.

Because transferrins a t PI 5.20, 5.29, 5.43, 5.54, 5.66, and 5.82 are isoforms containing 5, 4, 3, 2, 1, and 0 sialic acid residues/molecule, respectively, as described in the previous papers (3 and 4), the isoform at PI 5.54 in these patients contains 2 mol of sialic acids/molecule and that at PI 5.82 does not contain sialic acid. In order to elucidate the biochemical background of the multiplicity of the transferrin samples purified from the three patients, the sugar chain structure of serum transferrin, which has two potential sites for N-linked oligosaccharides at Asn413 and Asn611 (13), was investigated.

Structural Studies on Sugar Chains Released from Serum Transferrin Molecules of Patients with Carbohydrate-deficient Glycoprotein Syndrome-When tritium-labeled oligosaccharides released from 300 fig of each of the six transferrin samples, i.e. from the healthy control, the parents of patients 1 and 2 and the three patients, were subjected to paper electrophoresis at pH 5.4, they were separated into three acidic components, Al, A2 and A3, as shown in Fig. 2, A-F, respectively. The percent molar ratios of Al, A2, and A3 in the respective samples were calculated, as shown in Fig. 2, A-

I .... I I

m n -

o m 2 0 3 0 4 0 Elution Volume ( ml )

FIG. 4. Chromatofocusing of the serum transferrin samples on a Mono P column. A-F, same as in Fig. 2. The transferrin contents were determined by ELISA (see "Materials and Methods"). Black arrows, the elution positions of partial desialylated transferrin isoform S4. 0, 1, 2, 3, and 4 indicate desialylated, mono-, di-, tri- and tetrasialylated transferrin isoform S4, respectively.

TABLE I1 Percent molar ratios of transferrin isoforms S,, S,, and So separated

from the six transferrin samDles as shown in Fia. 4 Transferrin isoforms

S I S* SO 76 molar ratio

Samples

Healthy control 95 0 0 Parents of patients 1 and 2

Father 93 0 0 Mother 89 0 0

Patient 1 52 32 11 Patient 2 42 37 13 Patient 3 50 31 15

F, on the basis of their radioactivities. Because all the acidic oligosaccharides in fractions AI, A2, and A3 were converted to neutral oligosaccharides on exhaustive Arthrobacter sialidase digestion, the acidic nature of these oligosaccharides can be ascribed to their sialic acid residues. The oligosaccharides in fractions Al, A2, and A3 were confirmed to be monosialyl, disialyl and trisialyl derivatives, respectively, by determining the number of acidic fractions produced from each fraction on partial desialylation according to the method reported previously (14) (data not shown). The respective desialylated oligosaccharides derived from the six transferrin samples were separated into four fractions by serial lectin affinity chroma- tographies, i.e. on ConA, AAL, and DSA-Sepharose, as described in the previous paper (15), and the percent molar ratios of the respective fractions are summarized in Table I. Over 96% of the oligosaccharides from each sample was recovered in the ConA' AAL- DSA- fraction. Sequentially, the respective ConA+ AAL- DSA- fractions from the six transferrin samples were subjected to Bio-Gel P-4 (under 400 mesh) column chromatography. The respective elution pat-


terns showed the same single symmetrical peak at 13.5 glucose units, as shown by the solid lines in Fig. 3, A-F. On digestion with a mixture of diplococcal ,&galactosidase and P-N-acetylhexosaminidase, which specifically hydrolyzes the GalPl+ 4GlcNAcPl-2 group, each radioactive oligosaccharide was converted to a component with the same mobility as authentic Mana. GlcNAc . GlcNAcoT, 2 mol of Gal@l+4GlcNAc@l+B group being released, as shown by the dotted lines in Fig. 3, A-F. The structure of the pentaitol was confirmed to be Manal+6(Manal+3) Manpl+4GlcNAcfi1+4GlcNAcoT, as described in the previous paper (9). These results indicate that over 96% of the desialylated oligosaccharides from the six transferrin samples should equally have non-fucosylated biantennary complex type sugar chains, as shown in Table 1. The remaining minor components were also determined, according to the same methods as reported previously (15) (data not shown). Their structures are summarized in Table I. Acidic oligosaccharide fractions Al, A2, and A3 of the six samples were applied to a Trichosanthes japonica agglutinin- I (TJA-I)-Sepharose column, which specifically interacts with the Siaa2+6GalPl+4GlcNAc group (8), and to a MAL- Sepharose column, which specifically interacts with the Siaa2+3Gal@l+4GlcNAc group (11 and 16), respectively. All the acidic fractions were adsorbed to the TJA-I-Sepharose column and eluted with TB-NaN3 containing 0.1 M lactose, whereas they passed through the MAL-Sepharose column, indicating that they must exclusively have Neu5Aca2- GGal@l+4GlcNAc groups in their molecules.

Because all fractions A1 and A2 of the six samples bound to the ConA column and all fractions A3 passed through the column, the sugar chains of the six samples were equal to each other, irrespective of whether they were from healthy individuals or patients, and the structures of Al, A2, and A3 can be summarized as shown.

which possesses two biantennary oligosaccharides terminat- ing in sialic acid groups, because it was separated into five components on partial sialidase digestion, as shown by the black arrows at the top of Fig. 4A. 0, 1, 2, 3, and 4 indicate the numbers of sialic acids linked to 1 mol of transferrin. How- ever, the transferrin isoforms of the three patients were eluted from the chromatofocusing column at pH 5.12 (5.16), 5.42, and 5.80, as shown in Fig. 4, D-F, respectively. The isoforms eluted at pH 5.12 (5.16), 5.42, and 5.80 contained 4, 2, and 0 sialic acid residues/molecule and were named isoforms S4, Sp, and So, respectively. The percent molar ratios of transferrin isoforms S4, Sz , and So in the six samples were calculated, as summarized in Table 11.

Numbers of Sugar Chains Linked to Transferrin Isoforms S4, SP, and So-The numbers of sugar chains linked to transferrin isoforms S4, S p , and So were determined as follows. Approximately 100 pg of each of the six S4 samples, the S P samples from the three patients, and the So sample from patient 2 were hydrazinolysed and N-acetylated, respectively. After 5 nmol of lactose had been added to the respective sugars released from isoforms S4, Sz, and So, they were reduced with NaB3H4. Based on the specific activity of NaB3H4 determined from the radioactivity incorporated into lactitol and the molecular mass (80 kDa) of transferrin, the numbers of sugar chains released from 1 mol of isoforms S4, Sz, and So were determined. As summarized in Table 111, the numbers of sugar chains per isoforms S4, S p , and So were calculated to be 2.23-1.80, 1.27-0.90, and nearly 0, respectively, and no 3H- labeled N-acetylglucosaminitol could be detected in the sugars released from transferrin isoforms S4, S p , and So. On SDS- PAGE, the molecular masses of transferrin isoforms S4, Sz, and So were 80, 77, and 74 kDa, respectively. Furthermore, the average molecular masses of S4, Sz, and So were determined to be 79,571.7 * 5.3 Da, 77,366.5 * 4.5 Da, and 75,160.7

fFucal Galpl4GlcNAc~1+2Mancul .1

1 6 6

7 3

Al: Neu5Aca24 Man~l+4GlcNAc~l4G1cNAc

Galp14GlcNAc~1-+2Manal

cFucal Neu5Accu2+6Gal~14GlcNAc~1+2Manal 1

1 6 6

7 3

A2: Man/31+4GlcNAc~l+4GlcNAc

Neu5Aca2+6Gal~1~4G1cNAc/31-+2Manal

Neu5Aca2+6Galp14GlcNAc(31 1 6 fFucal

Manal I A3: 7 2 16 6

Neu5Aca2+6Galp14GlcNAcpl Man~l-rlGlcNAc~l+4GlcNAc

Neu5Aca2+6Ga1~14GlcNAc~1+2Manal 7 3

Chromatofocusing of Serum Transferrin Samples-Because a structural difference in the sugar chains linked to the transferrins between healthy individuals and the CDG syndrome cases could not be found, the respective transferrin isoforms showing different isoelectric points were separated by chromatofocusing. When the transferrin samples were applied to a Mono P chromatofocusing column, the transferrin samples of the healthy control, and the parents of patients 1 and 2, were mainly eluted at pH 5.12 (5.16), as shown in Fig. 4, A-C, respectively, and were named isoform S4. The major transferrin isoform, S4, is a tetrasialyl component,

* 3.6 Da by electrospray ionization-mass spectrometry,' and the molecular masses of S p and So were equivalent to those deleting one and 2 mol of sialylated biantennary oligosaccharides from S4, respectively. These results show that transferrin isoforms S p and So are not post-translational products of endo- P-N-acetylglucosaminidase digestion but that the So isoform is a non-glycosylated transferrin and the Sz isoform is linked to one asparagine-N-linked sugar chain.

* H. Ideo, M. Kanai, T. Ohkura, K. Fukushima, and K. Yamashita, manuscript in preparation.

5788 N-Linked Oligosaccharide Transfer Deficiency

TABLE 111 Incorporated radioactivity in 1 nmol of lactose and oligosaccharides released from 1 nmol(80 gg) each of transferrin isoforms S4, S , and SO and

the numbers of asparagine-N-linked oligosaccharides per transferrin isoform Transferrin

[3H]Sugar chain [3H]Lactitol Sugar chains isoforms per isoform

X 10' dpm rnollrnol Healthy control s 4 6.94 3.13 2.22 Parents of patients 1 and 2

Father s4 7.20 3.50 2.06 Mother s4 7.38 3.42 2.16

Patient 1 s 4 7.18 3.98 1 .80 sz 3.63 3.29 1.10

Patient 2 S S 7.43 3.86 1.92 S2 4.85 3.82 1.27 SO 0.20 3.36 0.06

Patient 3 S4 7.88 3.53 2.23 sz 3.87 4.29 0.90

h ConA TJA-I RCA-I : OA

2 0 2

3 OA

5 02

.- C t 0.0

$ u- 0.0 0

0.4 C d 02 E 0 0.0

0 4 0 0 4 8 0 4 8

SO

S2

S.

Elution Volume ( ml ) FIG. 5. Affinity chromatography of transferrin isoforms Sa,

SZ, and SO of patient 2 on ConA, TJA-I, and RCA-I columns. Each sample (1 gg) was dissolved in TB-NaN3 containing 0.1% BSA and then applied on an immobilized lectin column (1 ml, 7.5 mm inner diameter). After being kept at 4 "C for 15 min, each column was irrigated with 5 ml of TB-NaN3 containing 0.1% BSA and then with 5 ml of the same buffer containing 5 mM a-methyl-D-mannoside (ConA), 0.1 M lactose (TJA-I) , or 10 mM lactose (RCA-I) from the positions indicated by the black arrows. A-C, elution profiles on a ConA column; D-F, those on a TJA-I column; G-I, those on a RCA- I column. A, D, and G, So; B, E, and H , Sz; C, F, and I, S4. The same elution patterns were obtained for the transferrin isoforms of patients 1 and 3.

Lectin Affinity Chromatography of Transferrin Isoforms S4, Sz, and So-In order to confirm whether or not the So transferrin isoform is non-glycosylated, and the sugar chains of the Sa and S4 transferrin isoforms are mostly sialylated, the So, Sz, and S4 transferrin isoforms were applied to ConA, TJA-I, and RCA-I columns, respectively. As shown in Fig. 5, transferrin isoforms S4 and Sz bound to the ConA- and TJA-I- Sepharose columns, but passed through the RCA-I-agarose column, indicating that two and one biantennary sugar chains with Neu5Aca24Gal~l+4G1cNAc(31+2 groups at their nonreducing termini are linked to the S4 and Sz isoforms, respectively. On the other hand, the So isoform passed through all three columns, supporting that So is non-glycosylated.

From the results so far described, the structures of transferrin isoforms S4, Sz, and So of carbohydrate deficient syndrome patients can be concluded to be as shown in Fig. 6.

In order to confirm that this syndrome is an N-linked oligosaccharide transfer deficiency, several serum glycoproteins from patients with CDG syndrome were examined by

s4 HZN-Val v v Asn - Asn - Pro-COOH

1 413 61 1 079

sz H2N-Val Asn - Asn - Pro-COOH 1 413 61 1 079

:Hz v H2N-Val Asn - Asn - Pro-COOH

1 413 61 1 679

so H2N-Val Asn - Asn - Pro-COOH 1 413 61 1 679

FIG. 6. The proposed structures of transferrin isoforms S d , Sz, and So derived from carbohydrate deficient glycoprotein syndrome patients.

means of SDS-PAGE,3 isoelectric focusing: and electrospray ionization-mass spectrometry.' Isoforms with higher isoelectric points (PI values) than those in healthy controls were broadly observed for serum glycoproteins, including not only transferrin but also al-antitrypsin, orosomucoid, a'-HS-glycoprotein, antithrombin 111, plasminogen, thyroxine-binding globulin, and Zn-a'-glycoprotein (data not ~ h o w n ) . ~ When antithrombin I11 and orosomucoid were subjected to SDS- PAGE using 7.5% (for antithrombin 111) and 10% (for orosomucoid) homogeneous gels, and immunostained, they showed extra bands corresponding to smaller molecular masses, by about 3 or 6 kDa, than in the normal control, respectively (data not shown).' Asialo-transferrin, al-antitrypsin, and orosomucoid from patients with CDG syndrome showed isoforms with smaller molecular masses, by about 2000-5000 Da, than the normal ones on electrospray ionization-mass spectrometry, respectively (data not shown).' These

3Yuasa, I., Hashimoto, K., Ohno, K., Iijima, K., Yamashita, K., Takeshita, K., and Okada, K. (1993) Clin. Chim. Acta, in press.


results indicate that a defect in glycoprotein biosynthesis caused by impairment of N-linked oligosaccharide transfer is a general phenomenon, which is observed for not only transferrin but also other serum glycoproteins from patients with CDG syndrome.

DISCUSSION

This study clearly demonstrated that the sugar chain structures of serum transferrin purified from patients with CDG syndrome are mature and that one or two potential sites for asparagine-N-linked oligosaccharides of transferrin are non- glycosylated. Accordingly, this carbohydrate-deficient glycoprotein syndrome may be due not to unusual post-translational processing but to a deficiency of a biosynthetic enzyme for dolichol-oligosaccharide intermediates or an N-linked oligosaccharide transferase, although the primary deficient enzyme has not yet been determined. Because similar phenomena were observed for other serum glycoproteins including a,-antitrypsin, orosomucoid, az-HS-glycoprotein, thyroxine-binding globulin, Zn-az-glycoprotein, and several clotting factors, CDG syndrome can be concluded to be a congenital asparagine-N-linked oligosaccharide transfer deficiency. Al- though the results of monosaccharide composition analysis of CDG syndrome transferrin by Stibler and Jaeken (3) were contradictory as to our conclusion, the discrepancy may be due to their using a mixture of transferrin isoforms.

We found previously that asparagine-N-linked oligosaccharides of glycoproteins are degraded by combinations of exo- glycosidases and endo-P-N-acetylglucosaminidase in a series of studies on a congenital exoglycosidase deficiency in relation to the degradation mechanism for glycoproteins (17-20). No [3H]N-acetylglucosaminitol was released from transferrin isoforms So and S p on hydrazinolysis and following reduction with NaB3H4, showing that isoforms So and S p are not produced through the action of an endogenous endo-p-N-acetylglucosaminidase. Furthermore, the possibility remained that transferrin isoforms So and S z were post-translationally produced through modification by an endogenous N-glycanase. But this was ruled out by the following phenomena. If isoform So was N-glycanase-digested transferrin, it should have a more acidic nature than the desialylated S4 and S z transferrin isoforms, because 2 asparagines, 413 and 611, are converted to aspartic acid on N-glycanase digestion (21). Because the isoelectric points (PI values) of desialylated isoforms S4, S z , and So were the same, 5.82 (data not shown), transferrin

isoforms S p and So are not post-translational products of an endogenous N-glycanase.

There have been no reports of congenital disorders in relation to the biosynthetic mechanism for glycoproteins pos- sessing N-linked oligosaccharides, except for ones on congenital dyserythropoietic anemia type I1 (HEMPAS), in which Mana:N-acetylglucosaminyltransferase I1 is deficient (21), and a deficiency of phospho-N-acetylglucosaminyltransferase (I-cell disease) (22 and 23). This report is the first of a congenital asparagine-N-linked oligosaccharide transfer deficiency.

REFERENCES 1. Jaeken, J., van Eijck, H. G., van der Heul, C., Corbeal, L., Eeckels, R., and

2. Jaeken, J., Stibler, H., and Hagberg, B. (1991) Acta Pediatr. S c a d . Suppl.

3. Stibler, H., and Jaeken, J. (1990) Arch. Dis. Child. 6 6 , 107-111 4. Ohno, K., Yuasa, I., Akaboshi, S., Itoh, M., Yoshida, K., Ehara, H., Ochiai,

5. Uchida, Y., Tsukada, Y., and Sugimori, T. (1974) Blochim. Biophys. Acta

6. Glasgow, L. R., Paulson, J. C., and Hill, R. L. (1979) J. Biol. Chem. 2 6 2 ,

Eggermont, E. (1984) Clin. Chim. Acta 144,245-247

376 , 1-71

Y., and Takeshita, K. (1992) Brain Deu. 14,30-35.

360,425-431

Rfil.5-8fi2.1 - - -. - - - - 7. Yamashita, K., Totani, K., Ohkura, T., Takasaki, S., Goldstein, I. J., and

8. Yamashita, K., Umetsu, K., Suzuki, T., and Ohkura, T. (1992) Biochemistry

9. Yamashita, K., Liang, C.-J., Funakoshi, S., and Kobata, A, (1981) J. Biol.

10. Takasaki, S., Mizuochi, T., and Kobata, A. (1982) Methods Enzymol. 8 3 ,

Kobata, A. (1987) J. Biol. Chem. 262,1602-1607

31,11647-11650

Chem. 2 6 6 , 1283-1289

11. Kobata, A,, and Yamashita, K. (1993) in Practical Approach-Glycoprotein 263-268

Analysis (Fukuda, M., and Kobata, A., eds) Oxford University Press, Oxford, in press

12. Yamashita, K., Mizuochi, T., and Kobata, A. (1982) Methods Enzyml. 8 3 , 105-126

13. MacCillivray, R. T. A., Mendez, E., Shewale, J. G., Sinha, S. K., Lineback- Zins, J., and Brew, K. (1983) J. Biol. Chem. 268,3543-3553

14. Yamashita, K., Hitoi, A., Matsuda, Y., Tsuji, A., Katunuma, N., and Kobata,

15. Yamashita. K.. Koide. N.. Endo. T.. Iwakl. Y.. and Kohata. A. (1989) J. A. (1983) J. Biol. Chem. 268,1098-1107,

l%ol. Chhm. 264,2415-2423 I , , , , . ,

16. Knibhs, R. N., Goldstein, I. J., Ratcliffe, R. M., and Shibuya, N. (1991) J.

17. Yamashita, K., Tachibana, Y., Takada, S., Matsuda, I., Arashima, S., and

18. Yamashita, K., Tachibana, Y., Mihara, K., Okada, S., Yabuuchi, H., and

19. Yamashita, K., Ohkura, T., Okada, S., Yabuuchi, H., and Kobata, A. (1981)

20. Tachibana. Y.. Yamashita. K.. Kawaeuchi. M.. Arashima. S.. and Kobata.

Biol. Chem. 2 6 6 , 8 3 4 8



J. Biol. Chem. 266,4789-4798

A. (1981j J.'Biochem. (f'okyo) 90,i29111296 , .

21. Tarentino, A. L., and Plummer, Jr., T. H. (1987) Methods Enzyml. 1 3 8 , 770-77%

22. Fukuda, M. N., Dell, A,, and Scartezzii, P. (1987) J. Biol. Chem. 262,7195-

23. Reitman, A. L., Varki, A., and Kornfeld, S. (1981) J. Clin. Inuest. 67,1574-

24. Hasilik, A., Waheed, A., and von Figura, K. (1981) Biochem. Biophys. Res.

.." . . V

7206

1579

Commun. 9 8 , 761-767

the journal of vol. 268, no. 8, 15, pp. 5783-5789 …sodium salt in 0.1 m carbonate buffer, ph 9.6,...

Documents