Further studies on a highly purified glycoprotein from the intimal region of porcine aorta

Biochimica et Biophysica Acta, 439 (1976) 26-37 © Elsevier Scientific Publishing Company, Amsterdam - - Printed in The Netherlands BBA 37374

F U R T H E R STUDIES ON A H I G H L Y P U R I F I E D G L Y C O P R O T E I N FROM T H E I N T I M A L R E G I O N OF PORCINE AORTA

B R U C E I. ROBERTS and P R E M A N A N D V. W A G H

Connective Tissue Laboratory, Veterans Administration Hospital, 300 E. Roosevelt Road, Little Rock, Ark. 72206 (U.S.A.) (Received December 17th, 1975)

SUMMARY

Highly purified glycoprotein from the intimal region of porcine aorta was isolated with minor modifications of the procedure described previously. The molecular weight of the glycoprotein as determined by sedimentation equilibrium method either in presence of 0.1 M NaCI or 6 M guanidine.HC1 containing fl-mercaptoethanol was 72 000. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the native glycoprotein and its S-carboxyamidomethyl derivative at different acrylamide concentrations showed no difference in the molecular weight indicating the absence of subunits. Attempts to determine the identity of the amino-terminal acid by a dansylation technique indicated that the amino group is not flee. The carboxy-terminal amino acid was found to be serine after treatment of the glycoprotein with carboxypeptidase A. The glycoprotein did not contain an alkali-labile (O-glycosidic) carbohydrate-protein linkage as tested by the fl-elimination reaction. The release of monosaccharides from the glycoprotein as a function of time was studied employing mild acid hydrolysis (0.5 M HCI, 80 °C) and also by the use of neuraminidase, tt-Dand fl-D-glucosidases and fl-D-N-acetylglucosaminidase. From the observations on the release of monosaccharides and analogy with standard features determined by other investigators on soluble aortic glycoproteins, a prediction has been made as to the general features of the carbohydrate moiety of the glycoprotein.

INTRODUCTION

Mammalian arterial wall contains a variety of glycoproteins, some of which can be readily extracted with neutral buffers. Several soluble glycoproteins, as distinct from those which require more rigorous conditions for extraction, e.g. the use of trichloroacetic acid, high alkaline pH or urea, have been purified from bovine aorta [1, 2]. Although structural studies on the carbohydrate component of these glycoproteins have been attempted [2, 3], very little is as yet known regarding the nature of the carbohydrate moiety of glycoproteins from the aorta. We have previously reported a procedure for the isolation of a novel glycoprotein from the intimal region of porcine aorta [4]. The intimal glycoprotein contained 4 mol of hexosamine, 3 mol each of glucose and galactose, 2 mol of mannose

27 and 1 mol each of fucose and sialic acid per mol of glycoprotein (molecular weight 72 000). We report here some of the molecular properties of the glycoprotein and propose a tentative structure for the carbohydrate component consistent with the experimental results. MATERIALS AND METHODS Chemicals were obtained as follows: DEAE-cellulose (Cellex D, 0.7 mequiv./g), Coomassie brilliant blue R-250 dye and Bio-Gel P-150, Bio-Rad; Sephadex G-25 and G-15, Pharmacia Fine Chemicals; bovine pancreatic carboxypeptidase A, porcine pancreatic carboxypeptidase B, neuraminidase (Clostridium perfringens), and Dnsamino acids, Sigma Chemical Co. ; fl-o-glucosidase (sweet almond) and t~-o-glucosidase (yeast), Boehringer Mannheim Corp. All other chemicals were of reagent grade. fl-o-N-Acetylhexosaminidase (Jack bean) and bovine submaxillary mucin were gifts of Dr. Y. T. Li and Dr. Ward Pigman, respectively. Thoracic aortae from freshly killed pigs (about 6 months of age) were obtained from a local abattoir. The intima powder was prepared by the procedures previously described [4, 5]. Isolation of porcine intimal glycoprotein. The procedure for the isolation of glycoprotein described previously [4] was used with some modifications. In the present study, instead of chromatography on the first DEAE-cellulose column, a batchwise DEAE-cellulose fractionation method was used allowing us to process larger amounts of tissue. The dialyzed (NH4)2SO4 precipitate solution was added dropwise to a 1:1 (v/v) slurry of DEAE-cellulose in 0.005 M Tris.HCl buffer containing 0.001 M EDTA, pH 7.0 (standard buffer) under constant stirring. (A ratio of 2 ml packed resin per mg protein was necessary for maximal adsorption of the glycoprotein). The slurry was stirred for 3 h. The resin collected by centrifugation (1020 × g, 30 rain) was resuspended in an equal volume of standard buffer and stirred as above. After centrifugation, the supernatant fraction was discarded. Thereafter, the resin was stirred in an equal volume of standard buffer containing 0.2 M NaC1 for 3 h and centrifuged. This 0.2 M NaC1 elution was repeated two more times. The supernatant fractions from the three centrifugations were combined, dialyzed against standard buffer and lyophilized. The lyophilized product was dissolved in a small amount of standard buffer and dialyzed exhaustively against the same buffer. The solution was then applied to the second DEAE-cellulose column essentially as described previously [4]. The protein eluted in the region of the gradient corresponding to 0.16-0.18 M NaC1 contained the intimal glycoprotein. Small amounts of contaminating proteins remained due to these modifications. The contaminants were removed by passage of the glycoprotein fraction through a column of Bio-Gel P-150. Elution was performed with the standard buffer. Effluent from the column was monitored for protein at 280 nm employing an automatic ultraviolet analyzer. A major protein peak, eluting soon after the void volume of the column, was pooled, dialyzed against deionized water and lyophilized. This fraction was homogeneous as observed by disc gel electrophoresis and was identical in composition to the intimal glycoprotein reported previously [4]. Results from six different experiments were reproducible. The purity of the glycoprotein was tested by electrophoresis on polyacrytamide

28 gels. Electrophoresis was performed in 7-7.5 ~ gels using Tris/glycine buffer, pH 8.4 [6], and in 7.5~o gels using 0.1 M sodium phosphate buffer, pH 7.2, containing 0.1 sodium dodecyl sulfate [7]. Gels were stained and destained [8] using Coomassie brilliant blue R-250 dye. Analytical methods. Analysis of individual neutral sugars released from the glycoprotein by acid hydrolysis or by the action of glycosidases was accomplished by gas-liquid chromatography of the alditol acetates [9, 10]. Sialic acid was determined by the thiobarbituric acid method [11]. Hexosamine was quantitated by the procedure of Gatt and Berman [12]. Glucosamine was identified as the sole amino sugar by use of a Beckman 120C amino acid analyzer following hydrolysis of the glycoprotein in 2 M HCI for 16 h at 100 °C. The method of Reissig et al. [13] was used for N-acetyl amino sugar analysis. Other analytical procedures have been previously described [4]. Molecular weight determhration. The molecular weight of porcine intimal glycoprotein was determined both by sedimentation equilibrium method [14] using a Beckman Model E ultracentrifuge and by sodium dodecyl sulfate-polyacrylamide gel electrophoresis [7, 15]. In sedimentation equilibrium studies, the glycoprotein was dissolved in 0.1 M NaC1 and in 6 M guanidine.HCl containing 0.1~o flmercaptoethanol. Mixtures containing porcine intimal glycoprotein and standard proteins of known molecular weights were electrophoresed on 5, 7.5 and 10~ polyacrylamide gels in 0.1 M sodium phosphate buffer at pH 7.2 containing 1 ~ sodium dodecyi sulfate. Electrophoresis was performed at a current of 3.3-4 mA (25 oC) per tube for 1.5-5 h. In a separate experiment, the glycoprotein was first reduced with/3mercaptoethanol in 6 M urea and then alkylated with iodoacetamide [16]. The Scarboxyarnidomethylated derivative was electrophoresed in sodium dodecyl sulfate gels in a manner identical to that described for the native molecule. Terminal amino acid determination. Attempts were made to identify the amino-terminal amino acid by reaction of the glycoprotein (6 mg) with Dns-chloride [17]. The reaction mixture was passed through a Sephadex G-25 column. The Dnsglycoprotein derivative contained in the void volume of the column effluent was hydrolyzed in 6 M HC1 under N2 for 19 h at 105 °C. The hydrolysate was examined for Dns-amino acids by electrophoresis [17] at pH 2.0 (8 ~ formic acid in water) and at pH 4.4 (pyridine/acetic acid/water; 10:20:250, v/v). Thin-layer chromatography of the hydrolysate [18] was conducted on silica gel sheets (Eastman Kodak, Type K 301R) in three solvent systems: solvent I (chloroform/tert-amyl alcohol/acetic acid; 70:30: 3, v/v); solvent II (chloroform/tert-amyl alcohol/formic acid; 70: 30:1, v/v); solvent III (benzene/pyridine/acetic acid; 80:20:2, v/v). The fluorescent spots were detected under ultraviolet light. The COOH-terminal amino acid of glycoprotein was determined by the carboxypeptidase methods [19]. Carboxypeptidase A (20 #g, 12 units) and carboxypeptidase B (150/~g, 2 units) were reacted with 13.5 and 15.5 mg glycoprotein, respectively. Reaction tubes with appropriate controls containing enzyme but no glycoprotein were analyzed concurrently. Aliquots were removed at 0, 1 and 3 h. Free amino acids in these aliquots were quantitated using the Beckman 120C amino acid analyzer. Release of monosaccharides from porcine intimal glycoprotein by miM acid hydrolysis. The glycoprotein (12-25 mg) was dissolved in 10 ml of 0.5 M HCI and hydrolyzed at 80 °C over a 24-h period. Suitable aliquots (0.15-0.40 ml) were removed at various times. Free neutral sugars in the aliquots were quantitated by gas-liquid

29 chromatography. Hexosamine and sialic acid were determined by colorimetric procedures. Treatment of porcine intimal glycoprotein with glycosidases. The glycoprotein (1-3 mg) was incubated with neuraminidase (30 munits) in 0.25 ml of 0.1 M sodium acetate buffer, pH 4.5, at 37 °C. Aliquots (0.05 ml) were removed at various times, immersed in boiling water for 1 min and analyzed for free sialic acid. The action of fl-o-N-acetylhexosaminidase on the native glycoprotein and that exposed briefly to mild acid hydrolysis was studied as follows: Native porcine intimal glycoprotein (2-5 mg) was incubated with the enzyme (21 munits) in 1.0 ml of 50 mM sodium citrate buffer, pH 5.2, at 30 °C. Aliquots (0.2 ml) were removed and analyzed for free N-acetylhexosamine. A sample of native glycoprotein (7.6 mg) was hydrolyzed in 0.6 ml of 0.5 M HC1 at 80 °C for 2 h to remove acid-labile monosaccharides. The hydrolysate was dialyzed extensively against several changes of water and the non-dialyzable material was lyophilized. The dried protein was dissolved in 1.0 ml of 0.05 M NaOH and was acetylated by the procedure of Roseman and Daffner [20] to replace acetyl groups that may have been lost from the acetyl amino sugars during the hydrolysis. The acetylated protein was desalted by passage through a Sephadex G-25 column, the protein eluate was lyophilized and dissolved in 0.4 ml of 50 mM sodium citrate, pH 5.2. fl-o-N-Acetylhexosaminidase (20 munits) was added to this solution and the reaction mixture was incubated at 30 °C. Aliquots (0.2 ml) were taken at various times and analyzed for free N-acetylhexosamine. After 6 h of incubation, the remainder of the incubation mixture was adjusted to 2 M HCI, hydrolyzed for 16 h at 105 °C and analyzed for total hexosamine. The native and asialo-glycoproteins were treated with a-D- and fl-Dglucosidases. The various glycosidase activities present in the partially purified fl-Dglucosidase preparation were identified using o- or p-nitrophenyl glycosides. The preparation contained 15~ as much fl-D-galactosidase activity as that of fl-Dglucosidase activity. Very small amounts (less than 1 ~o of the fl-D-glucosidase activity) of fl-D-N-acetylhexosaminidase and a-o-mannosidase were present, a-DGlucosidase and a-D-galactosidase activities were not detectable. No contaminants were found in the purified a-D-glucosidase preparation. Native glycoprotein (5.9 mg) was treated with fl-D-glucosidase (2 I.U.)in 1.2 ml of 0.1 M sodium acetate buffer, pH 5.0, at 37 °C. Aliquots (0.2 ml) were removed at 0, 3 and 9 h, deionized on columns containing Dowex 50 (H +) and Dowex 1 (HCO3-), and analyzed for released neutral sugars by gas-liquid chromatography. After 9 h, a-D-glucosidase (1 I.U.) was added to the reaction mixture and the incubation continued for another 15 h. Two 0.2-ml aliquots at 12 and 24 h were analyzed for sugars. Asialo-glycoprotein (2.4 mg), obtained by treatment with neuraminidase for 24 h, was incubated with fl-o-glucosidase (2 units) in the same manner. Aliquots (0.15 ml) were taken at 0, 6 and 16 h and analyzed for neutral sugars as described above. Test for O-glycosidic linkage in the intimal glycoprotein. Two solutions of intimal glycoprotein (250/~g/ml and 1 mg/ml) were prepared in 0.5 M NaOH. The change in absorbance at 240 nm of the first solution was observed for 6 h at 25 °C and that of the other was observed for 36 h at 50 °C to detect the formation of unsaturated amino acids [21]. A similar solution containing bovine submaxillary mucin was tested to confirm the reliability of the reaction.

30 RESULTS

The molecular weight of porcine intimal glycoprotein The results of all the molecular weight determinations on the glycoprotein are summarized in Table I. Fringe displacement plots from sedimentation equilibrium analysis were linear when the glycoprotein was dissolved in NaC1 or in guanidine. HCI. Only a single protein band was visualized on polyacrylamide sodium dodecyl sulfate electrophoresis both with the native and S-carboxyamidomethylated porcine intimal glycoprotein. The variation of molecular weight as a function of acrylamide concentration was that which would be predicted for a glycoprotein containing 3 - 4 ~ carbohydrate [15]. From these data a molecular weight of 72 000 has been adopted for the intimal glycoprotein.

TABLE I MOLECULAR WEIGHT DETERMINATION OF PORCINE INTIMAL GLYCOPROTEIN Figures in parentheses indicate number of determinations. Method

Molecular weight

(A) Sedimentation equilibrium (1) NaC1, 0.1 M (2) Guanidine' HCI, 6 M

72 110 (4) 72 370 (3)

(B) Sodium dodecyl sulfate gel electrophoresis (1) Native glycoprotein 5.0 ~ acrylamide 7.5 ~ acrylamide

10.0~ acrylamide (2) S-Carboxyamidomethyl glycoprotein 5.0 ~ acrylamide 10.0~ acrylamide

85 400 (1) 77 500 (2) 74 800 (1) 75 400 (2) 70 400 (3)

NH2- and COOH-terminal amino acids Several attempts were made to identify the NHz-terminal residue of intimal glycoprotein. When the Dns-glycoprotein was hydrolyzed and the products in the hydrolysate were examined by either thin-layer chromatography or by high voltage electrophoresis, only two fluorescent spots were detected. These were identified as Dns-e-lysine and Dns-amine. In all experiments we failed to detect a Dns-a-amino acid derivative. We therefore conclude that the amino-terminal amino acid in the glycoprotein does not have an a-amino group available for reaction with Dnsreagent. Treatment of the intimal glycoproteins with carboxypeptidase B and analysis of the aliquots of the reaction mixture revealed that none of the basic amino acids was released from the glycoprotein. Several free amino acids appeared in the digest when the glycoprotein was treated with carboxypeptidase A (Fig. l). The results suggest that the COOH-terminal residue is serine. However, the data do not allow ruling out glycine as the COOH-terminal amino acid.

31 0.5

,

,

,

~

,

,

T~ °0.4 et. o >-

o

0.3

o v

.<, .-L

40.1 0

0.0

1

2 TIME

(HOURS)

Fig. 1. Carboxypeptidase A reaction with intimal glycoprotein. Giycoprotein (0.19/~mol) was incubated at 37 °C in 0.2 M pyridine/acetate, pH 8.8, containing 1.6~ sodium dodecyl sulfate in the presence of 12 units of the enzyme. Final volume was 2 ml. Aliquots (0.5 ml) were removed at 0, 1 and 3 h and analyzed for free amino acids. The release of the amino acids is expressed as a molar ratio (molecular weight of the glycoprotein, 72 000) after substracting the zero time values. O---O, serine; I---I, glycine; ?q---D, alanine; and O---O, aspartic acid.

Release of sugars from the carbohydrate component of porcine intimal glycoprotein Acid hydrolysis. The release of sialic acid from the native glycoprotein was studied by hydrolysis in 0.5 M HC1 at 80 °C. The amount of sialic acid present in the hydrolysate fell progressively between 1 and 6 h (Fig. 2). Extrapolation to zero time indicated that 1 mol of sialic acid per mol of glycoprotein had been released. In other experiments, treatment of the glycoprotein with 0.1 M HCI at 80 °C for 1 h also resulted in the cleavage of one sialyl residue. After dialysis of the hydrolysate, the remaining protein contained less than 0.1 residue of protein-bound sialic acid as observed by further hydrolysis in 0.1 M HC1 at 80 °C. The intimal glycoprotein was hydrolyzed in 0.5 M HC1 at 80 °C for 24 h and the release of hexosamine and neutral sugars was followed as a function of time. One residue each of glucose, galactose and mannose and 0.3 residue of fucose are released within the first 2 h of hydrolysis (Fig. 3). A second residue each of glucose and galactose was released at a slower rate between 2 and 18 h. The third residue of each of these sugars was released rapidly thereafter. The second mannose residue was not released until after 18 h. Only one-half residue of glucosamine was released within the first 2 h and an additional 2.5 residues were released at a steady, linear rate throughout the remaining period of hydrolysis. In a separate experiment, the glycoprotein was hydrolyzed in 0.5 M HC1 at

32 1.5

J

i

i

-~

i

I

I

I

I

I

.~.1

I

o

== o o >/0 ~ O ' - ' / p ~ ~' (D :E

~.0 •

/ /

t~ < t~

~i// / / /

0,5

/ / / (J < c.) --

/

/

/ 0.0

2

4

' 6

..,'

I 14

I 16

I 18

I 20

I $' 2

I 24

.. ""

, 40

i 42

TIME (HOURS)

Fig. 2. Release of sialic acid from intimal glycoprotein by mild acid hydrolysis and by neuraminidase. Glycoprotein was heated in 0.5 M HC1 at 80 °C. Aliquots were analyzed for free sialic acid at 1, 4 and 6 h (Q---O); glycoprotein was treated with neuraminidase (CI. perfringens) in 0.1 M acetate, pH 4.5, at 37 °C. Aliquots were taken at 15, 24 and 42 h and analyzed for sialic acid (O---O).

z

_

3.0

Q

¢D

2.5

o

Q =E v

2.0

,=, .-1

1.5

0

<=: o

o/

1.0

(::) 0

0

0.5

0,0 0

:5

6

9

12

15

18

21

24

TIME (HOURS)

Fig. 3. Release of neutral sugars and glucosamine from intimal glycoprotein by mild acid hydrolysis. The glycoprotein was heated in 0.5 M HCI at 80 °C. Aliquots were removed at various times and analyzed for glucosamine ( A - - - A ) , glucose (Q---O), galactose (I~---C]), mannose ( O - - - O ) and fucose

(m---m).

33 80 °C for 24 h, and the hydrolysate was separated into two fractions, void volume and salt volume, on a precalibrated Sephadex G-15 column. Each fraction was analyzed for total and free glucosamine [12]. The high molecular weight material in the void volume fraction contained 0.93 mol of glucosamine per mol of glycoprotein originally hydrolyzed, while the salt volume contained 2.81 mol of glucosamine per mol of hydrolyzed glycoprotein. From these data we concluded that after 24 h of acid hydrolysis, all covalently bound sugars except one glucosamine residue were released from the intimal glycoprotein. The apparent order of release was: sialic acid and fucose followed by one residue each of glucose, galactose and mannose followed by a second residue each of glucose and galactose. Two residues of glucosamine were released during the period that these second residues of glucose and galactose appear in the hydrolysate. The third residue each of glucose, galactose and glucosamine and the second residue of mannose were released last. The fourth residue of glucosamine apparently remained bound to the protein.

Glycosidases Treatment of the glycoprotein with neuraminidase released one residue of sialic acid within 15 h (Fig. 2). There was no additional release when the reaction was continued up to 42 h. When aliquots of the reaction mixture at 24 h of incubation were dialyzed, the non-dialyzable material was devoid of sialic acid. Therefore, in other structural investigations asialo-glycoprotein was routinely prepared by treatment of the native molecule with neuraminidase for 24 h. The native glycoprotein was treated with fl-o-N-acetyl-glucosaminidase from jack bean meal. The reaction was followed for 44 h. Aliquots of the incubation mixture were analyzed for free N-acetylglucosamine. No free N-acetylglucosamine was released from the native molecule. In another experiment, the glycoprotein was first hydrolyzed in 0.5 M HCI at 80 °C for 2 h. This treatment releases one residue each of glucose, galactose, mannose, fucose and sialic acid but less than one residue ofglucosamine (Fig. 3). The hydrolysate was neutralized, passed through a Sephadex G-25 column, and the high molecular weight material eluting in the void volume was reacetylated and exposed to fl-o-Nacetylglucosaminidase. Examination of aliquots taken over a 6-h incubation period showed that no N-acetylhexosamine was released. These data indicate that the native glycoprotein does not contain N-acetylglucosamine at a non-reducing terminus in fllinkage and that mild acid hydrolysis does not expose any N-acetylglucosaminyl residues which can be released by fl-N-acetylglucosaminidase. The native glycoprotein was treated with a partially purified preparation of fl-o-glucosidase which contained significant activity against fl-galactosides. After 9 h of incubation, a-o-glucosidase was added to the reaction mixture. Aliquots were analyzed at 3, 9, 14 and 24 h for neutral sugars (Fig. 4). One residue of glucose was released within 3 h. One residue of galactose was released at a slower rate over a 24-h period. Addition of a-glucosidase appeared to have no effect on the carbohydrate structure. The action of fl-D-glucosidase on the asialo-glycOprotein is also shown in Fig. 4. The release of glucose from the asialo-glycoprotein was essentially similar to that from the native glycoprotein. However, the enzyme released one residue of galactose

34 1.5

o o

o

1.0

~n 0

0 •

O

0 0

~n

u~

0.5

o

0.0 0

3

6

9

12

15

18

51

24

T I M E (HOURS)

Fig. 4. Treatment of native and asialo intimal glycoprotein with c~-o-glucosidase and fl--D-glucosidase. Native glycoprotein was treated with fl-D-glucosidase. Aliquots were removed at 3 and 9 h At 10 h, a-D-glucosidase was added and the hydrolysis was continued for an additional 14 h. The aliquots were analyzed for glucose (©---O) and galactose ([]---[]) by gas-liquid chromatography. Asialo-glycoprotein was treated with fl-o-glueosidase in a manner identical with the native molecule. a-D-Glucosidase was not used. O---O, glucose and I1---II, galactose.

very rapidly from the asialo-glycoprotein and a second residue of galactose at apparently the same rate as from the native glycoprotein. These results suggest that the carbohydrate component contains a glucosyl fl galactosyl fl (next residue) structure at a non-reducing terminus. The rapid release of a second galactose residue from the asialo-glycoprotein suggests the presence of galactose penultimate to the sialic acid, fl-linked to the next monosaccharide.

Test for alkali-labile linkage between the carbohydrate and protein Solutions of glycoprotein heated in N a O H failed to show the formation of unsaturated amino acids [21], resulting from the fl-elimination of carbohydrate components. The solution of bovine submaxillary mucin showed a linear increase in absorption at 240 nm for nearly 2 h, confirming the presence'of several O-glycosidic linkages between the carbohydrate and protein moieties [22]. There was no increase in absorbance when alkaline solutions of glycoprotein were incubated at 25 °C for 6 h and at 50 °C for 36 h, indicating that the intimal glycoprotein does not contain an alkali-labile linkage.

35 DISCUSSION In a previous study, it was suggested that porcine intima contained a glycoprotein unique in its characteristics [4]. The glycoprotein molecular weight of 72 000 is similar to that of a glycoprotein isolated from bovine aorta [2]. The porcine intimal glycoprotein, however, contains only 3.3~ carbohydrate in contrast to 15.8~ reported for the bovine aortic glycoprotein. The molceular weight of the porcine glycoprotein was essentially the same when determined by sedimentation equilibrium in the presence of either 0.1 M NaC1 or 6 M guanidine. HC1, and by gel electrophoresis of the native glycoprotein and its S-carboxyamidomethyl derivative in the presence of sodium dodecyl sulfate and fl-mercaptoethanol, suggesting that the porcine intimal glycoprotein does not contain subunits. Several attempts to determine the NH2-terminal amino acid employing a dansylation technique were unsuccessful, apparently because the amino group of the terminal amino acid is not free to react with dansyl reagent. It may be possibly blocked by an acetyl group as in ovalbumin [23] or may exist in the cyclic pyroglutamic form as in at-acid glycoprotein [24]. The data obtained using carboxypeptidase enzymes suggest that the COOH-terminal amino acid is serine. The glycoprotein carbohydrate component contains glucosamine, glucose, galactose, mannose, fucose and sialic acid in a molar ratio of 4: 3: 3 : 2:1 : 1, respectively. The data obtained from acid hydrolysis of the porcine intimal glycoprotein together with those obtained from treatment with glycosidases are insufficient to indicate the precise sequence of individual sugars. However, the information presented in this paper combined with other information in the relevant literature, does allow some speculation concerning the structural features of the prosthetic group. A postulated structure is shown in Fig. 5. The monosaccharides are numbered parenthetically for ease of identification. NeuAc (I)

Glu (13)

Gal (3)

Gal (5)

GIu (14)

GIcNAc (4)

GIcNAc (6)

GIc_NAc (7)

(FUC)? (2)

Man 18)

Glu 111)

Man ~'~G

cNAc (10}~'~~

AspNH2

$

Ga, ([2) ( ? ) -

Fig. 5. A predicted structure for the carbohydrate moiety of porcine intima] g]ycoprotein. NeuAc, N-acetylneuraminic acid; GIcNAc, N-acetyl-D-glucosamine.

One residue of sialic acid (NeuAc) (1) is at a terminal non-reducing position. It is quantitatively released by the action of a bacterial neuraminidase which is specific for a,2 sialyl --~ (3 or 6) linkages. The sialic acid is most probably linked to galactose ((1)-(3)) as has been observed in other aortic glycoproteins [2, 25]. Fucose is rapidly released by mild acid hydrolysis and may be located at a non-reducing terminus. The disaccharides (3)-(4) and (5)-(6) may be 4-O-fl-D-galactopyranosyl-fl-D-Nacetylglucosamine similar to disaccharides seen in bovine aorta glycoproteins [3, 26]. The terminal trisaccharide (1)-(3)-(4) has been shown to be present in an aortic

36 glycopeptide [25]. Galactose residues (3) and (5) are presumed to account for the two galactose residues released by acid hydrolysis in the first 18 h. The trisaccharide (7)-(8)-(10) has been commonly found within the inner core of oligosaccharides of several glycoproteins [27, 28]. The complete release of the three glucosamine residues (4), (6) and (7) would require cleavage of two bonds for each amino sugar, in agreement with the similar rates of release of these amino sugars by mild acid hydrolysis. Mannose residue (8) might be rapidly released from the structure due to the lability of the mannose-glucosamine linkage [29]. Mannose residue (9) would be released much more slowly from linkage with three glucosamines, appearing in the hydrolysate only after 18 h. A glycosaminyl linkage between glucosamine (10) and asparagine may be present in the structure. It was observed that after 24 h mild acid hydrolysis of the glycoprotein, one residue of glucosamine remained associated with the protein. No alkali-labile carbohydrate-protein linkage is present. The two glucose residues (13) and (14) shown at the non-reducing terminii account for the glucose released during the first 18 h of acid hydrolysis. One of these can be released by fl-glucosidase. A glycopeptide from bovine aortic glycoprotein has previously been shown to contain glucose in a terminal position [2]. The trisaccharide (2)-(11)-(12) has been placed as a separate carbohydrate structure. A glycopeptide containing glucose, galactose and fucose has been reported from two different aortic glycoproteins [3, 26]. The carbohydrate-peptide linkage for this small side chain is not known. The structure for the carbohydrate moiety presented in Fig. 5 appears to conform with all the data presented. The proposed structure must still withstand results of future investigations including methylation studies, oxidation with periodate and the use of other specific glycosidases. ACKNOWLEDGEMENTS This work was supported by Veterans Administration Research Funds, Project No. 9166-01. We express our gratitude to Dr. Y. T. Li and Dr. W. Pigman for their generous gifts of fl-D-N-acetylglucosaminidase and bovine submaxillary mucin, respectively. We thank Mrs. Diane Butler for secretarial assistance. REFERENCES 1 Radhakrishnamurthy, B., Fishkin, A. F., Hubbell, G. J. and Berenson, G. S. (1964) Arch. Biochem. Biophys. 104, 19-26 2 Maier, V. and Buddecke, E. (1971) Hoppe-Seyler Z. Physiol. Chem. 352, 1338-1346 3 Radhakrishnamurthy, B. and Berenson, G. S. (1973) J. Biol. Chem. 248, 2000-2006 4 Wagh, P. V. and Roberts, B. I. (1972) Biochemistry I 1, 4222-4227 5 Roberts, B. I., White, H. J., Read, R. C. and Wagh, P. V. (1974) Atherosclerosis 20, 533-537 6 Davis, B. J. (1964) Ann. N.Y. Acad. Sci. 121,404-427 7 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412 8 Catterall, W. A. and Pedersen, P. A. (1971) J. Biol. Chem. 246, 4987-4994 9 Lehnhardt, W. F. and Winzler, R. J. (1968) J. Chromatogr. 34, 471-479 10 Neidermeier, W. (1971) Anal. Biochem. 40,'465-475 11 Aminoff, D. (1961) Biochem. J. 81, 384-392 12 Gatt, R. and'Berman, E. R. (1966) Anal. Biochem. 15, 167-171

37 13 14 15 16 17 18 19 20 21

Reissig, J. L., Strominger, J. L. and Leloir, L. F. (1955) J. Biol. Chem. 217, 959-966 Yphantis, D. (1964) Biochemistry 3, 297-317 Segrest, J. P. and Jackson, R. L. (1972) Methods Enzymol. 28, 54. Carlsen, R. B., Bahl, O. P. and Swaminathan, N. (1973) J. Biol. Chem. 248, 6810-6827 Gray, W. R. (1967) Methods Enzymol. 11, 139-151 Morse, D. and Horecker, B. L. (1966) Anal. Biochem. 14, 429-433 Ambler, R. P. (1967) Methods Enzymol. 11, 155-166 Roseman, S. and Daffner, I. (1956) Anal. Chem. 28, 1743-1747 Neiderhiser, D. H., Planter, J. J. and Carlson, D. M. (1971) Arch. Biochem. Biophys. 145, 155163 22 Tanaka, K. and Pigman, W. (1965) J. Biol. Chem. 240, PC 1487-1488 23 Narita, K. and Ishi, J. (1962) J. Biochem. Tokyo 52, 367-373 24 Schmid, K., Kaufmann, H., Isemura, S., Bauer, F., Emura, J., Motoyama, T., Ishiguro, M. and Nanno, S. (1973) Biochemistry 12, 2711-2724 25 Klemer, A. and Nager, C. (1967) Z. Naturforsch. 22, 456 26 Radhakrishnamurthy, B. and Berenson, G. S. (1966) J. Biol. Chem. 241, 2106-2112 27 Spiro, R. G. (1970) Annu. Rev. Biochem. 39, 599-638 28 Wagh, P. V., Bornstein, J. and Winzler, R. J. (1969) J. Biol. Chem. 244, 658-665 29 Marshall, R. D. and Neuberger, A. (1972) in Glycoproteins (Gottschalk, A., ed.), pp. 224-251, Elsevier, Amsterdam