Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

unit; Section 11-1C). About 80 different kinds of naturally

occurring glycosidic linkages are known, most of which in-

volve mannose, N-acetylglucosamine, N-acetylmuramic

acid, glucose, galactose, fucose (6-deoxygalactose), N-acetyl-

neuraminic acid (sialic acid), and N-acetylgalactosamine

(Section 11-1C). Glycosidic linkages also occur to lipids

(e.g., glycosphingolipids; Section 12-1D) and proteins

(glycoproteins; Section 11-3C).

Glycosidic bond formation requires free energy input un-

der physiological conditions (G°¿  16 kJ ⴢ mol

1

). This

free energy, as we have seen in the case of glycogen synthe-

sis (Section 18-2B), is acquired through the conversion of

monosaccharide units to nucleotide sugars.A nucleotide at a

sugar’s anomeric carbon atom is a good leaving group and

thereby facilitates formation of a glycosidic bond to a second

sugar unit via reactions catalyzed by glycosyltransferases

Section 23-3. Biosynthesis of Oligosaccharides and Glycoproteins 881

Figure 23-11 The glyoxylate cycle. The cycle results in the net

conversion of two acetyl-CoA to succinate in the glyoxysome,

which can be converted to malate in the mitochondrion for use

in gluconeogenesis. Isocitrate lyase and malate synthase, enzymes

unique to glyoxysomes (which occur only in plants), are boxed in

blue. (1) Glyoxysomal citrate synthase catalyzes the

condensation of oxaloacetate with acetyl-CoA to form citrate.

(2) Cytosolic aconitase catalyzes the conversion of citrate to

isocitrate. (3) Isocitrate lyase catalyzes the cleavage of isocitrate

to succinate and glyoxylate. (4) Malate synthase catalyzes the

COO

–

OOC

COO

–

Citrate

COO

–

OOC

COO

–

Isocitrate

–

OOC

COO

–

Succinate

isocitrate

lyase

H COO

–

Glyoxylate

SCoA

Acetyl-CoA

CoA

Malate

–

OOC

COO

–

malate

dehydrogenase

malate

synthase

–

OOC

COO

–

Oxaloacetate

citrate

synthase

SCoA

Acetyl-CoA

–

OOC

COO

–

Succinate

–

OOC

COO

–

Fumarate

–

OOC

COO

–

Malate

fumarase

succinate

dehydrogenase

–

OOC

COO

–

Malate

–

OOC

COO

–

Oxaloacetate

Gluconeogenesis

NADH+H

NAD

Glyoxysome MitochondrionCytosol

aconitase

citric

acid

cycle

FAD

FADH

NADH + H

NAD

condensation of glyoxylate with acetyl-CoA to form malate.

(5) Glyoxysomal malate dehydrogenase catalyzes the oxidation

of malate to oxaloacetate, completing the cycle. (6) Succinate is

transported to the mitochondrion, where it is converted to

malate via the citric acid cycle. (7) Malate is transported to the

cytosol, where malate dehydrogenase catalyzes its oxidation to

oxaloacetate, which can then be used in gluconeogenesis.

(8) Alternatively, malate can continue in the citric acid cycle,

making the glyoxylate cycle anaplerotic.

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 881

(Fig. 23-12).The nucleotides that participate in monosaccha-

ride transfers are UDP, GDP, and CMP; a given sugar is as-

sociated with only one of these nucleotides (Table 23-2).

A. Lactose Synthesis

Several disaccharides are synthesized for future use as

metabolic fuels. In plants, the major fuel disaccharide is

sucrose (Section 11-2B), whose synthesis is discussed in

Section 24-3Ad. Typical of mammalian disaccharides is

lactose [-galactosyl-(1 S 4)-glucose; milk sugar], which is

synthesized in the mammary gland by lactose synthase

(Fig. 23-13). The donor sugar is UDP–galactose, which is

formed by epimerization of UDP–glucose (Section 17-5B).

The acceptor sugar is glucose.

Lactose synthase consists of two subunits:

1. Galactosyltransferase, the catalytic subunit, which

occurs in many tissues, where it catalyzes the reaction of

UDP–galactose and N-acetylglucosamine to yield N-acetyl-

lactosamine, a constituent of many complex oligosaccha-

rides (see, e.g., Fig. 23-20, Reaction 6).

2. ␣-Lactalbumin, a mammary gland protein with no

catalytic activity, which alters the specificity of galactosyl-

transferase such that it utilizes glucose as an acceptor,

rather than N-acetylglucosamine, to form lactose instead of

N-acetyllactosamine.

B. Glycoprotein Synthesis

Eukaryotic proteins destined for secretion, incorporation into

membranes, or localization inside membranous organelles

contain carbohydrates and are therefore classified as glyco-

proteins. Glycosylation and oligosaccharide processing play an

indispensable role in the sorting and the distribution of these

proteins to their proper cellular destinations. Their polypeptide

components are ribosomally synthesized and processed by

addition and modification of oligosaccharides.

The oligosaccharide portions of glycoproteins, as we

have seen in Sections 11-3C and 12-3Bc, are classified into

three groups:

1. N-Linked oligosaccharides, which are attached to

their polypeptide chain by a -N-glycosidic bond to the

side chain N of an Asn residue in the sequence Asn-X-Ser

or Asn-X-Thr, where X is any amino acid residue except

Pro (Fig. 23-14a).

2. O-Linked oligosaccharides, which are attached to

their polypeptide chain through an -O-glycosidic bond to

the side chain O of a Ser or Thr residue (Fig. 23-14b) or,

only in collagens (Section 8-2Bb), to that of a 5-hydroxyly-

sine (Hyl) residue (Fig. 23-14c).

3. Glycosylphosphatidylinositol (GPI) membrane an-

chors, which are attached to their polypeptide chain through

an amide bond between mannose-6-phosphoethanolamine

and the C-terminal carboxyl group (Fig. 23-14d).

We shall consider the synthesis of these three types of

oligosaccharides in turn.

a. N-Linked Glycoproteins Are Synthesized

in Four Stages

N-Linked glycoproteins are formed in the endoplasmic

reticulum and further processed in the Golgi apparatus.

Synthesis of their carbohydrate moieties occurs in four

stages:

1. Synthesis of a lipid-linked oligosaccharide precursor.

2. Transfer of this precursor to the side chain N of an

Asn residue on a growing polypeptide.

3. Removal of some of the precursor’s sugar units.

4. Addition of sugar residues to the remaining core

oligosaccharide.

882 Chapter 23. Other Pathways of Carbohydrate Metabolism

Figure 23-12 Role of nucleotide sugars. These compounds are the glycosyl

donors in oligosaccharide biosynthesis catalyzed by glycosyltransferases.

Table 23-2 Sugar Nucleotides and Their Corresponding

Monosaccharides in Glycosyltransferase

Reactions

UDP GDP CMP

N-Acetylgalactosamine Fucose Sialic acid

N-Acetylglucosamine Mannose

N-Acetylmuramic acid

Galactose

Glucose

Glucuronic acid

Xylose

glycosyl-

transferase

ROH

OH OH

1

Base

Donor

sugar

1

Nucleotide sugar

Oligosaccharide

–

O O

–

Nucleoside

diphosphate

Acceptor

sugar

Nucleoside

diphosphate

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 882

We shall discuss these stages in order.

b. N-Linked Oligosaccharides Are Constructed on

Dolichol Carriers

N-Linked oligosaccharides are initially synthesized as

lipid-linked precursors. The lipid component in this process

is dolichol, a long-chain polyisoprenol of 14 to 24 isoprene

units (17–21 units in animals and 14–24 units in fungi and

plants; isoprene units are C

units with the carbon skeleton

of isoprene; Section 25-6A), which is linked to the oligosac-

charide precursor via a pyrophosphate bridge (Fig. 23-15).

Dolichol apparently anchors the growing oligosaccharide to

the endoplasmic reticulum membrane. Involvement of lipid-

linked oligosaccharides in N-linked glycoprotein synthesis

was first demonstrated in 1972 by Armando Parodi and Luis

Leloir,who showed that,when a lipid-linked oligosaccharide

containing [

C]glucose is incubated with rat liver micro-

somes (vesicular fragments of isolated endoplasmic reticu-

lum), the radioactivity becomes associated with protein.

c. N-Linked Glycoproteins Have a Common

Oligosaccharide Core

The pathway of dolichol-PP-oligosaccharide synthesis

involves stepwise addition of monosaccharide units to the

Section 23-3. Biosynthesis of Oligosaccharides and Glycoproteins 883

H OH

Glucose

H OH

UDP–galactose

UDP

lactose synthase

H OH

Lactose

H OH

+ UDP

4)-glucose]

[␤-galactosyl-(1

Figure 23-13 Lactose synthase. This enzyme catalyzes the

formation of lactose from UDP–galactose and glucose.

Figure 23-15 Dolichol

pyrophosphate glycoside. The

carbohydrate precursors of N-

linked glycosides are synthesized

as dolichol pyrophosphate

glycosides. Dolichols are long-

chain polyisoprenols (n  14–24)

in which the -isoprene unit is

saturated.

Figure 23-14 Types of saccharide–polypeptide linkages in

glycoproteins. (a) An N-linked glycosidic bond to an Asn residue

in the sequence Asn-X-Ser/Thr. (b) An O-linked glycosidic bond

to a Ser (or Thr) residue. (c) An O-linked glycosidic bond to

a 5-hydroxylysine residue in collagen. (d) An amide bond

between the C-terminal amino acid of a protein and the

phosphoethanolamine bridge to the 6 position of mannose in the

glycophosphatidylinositol (GPI) anchor.The X group (green)

denotes the rest of the GPI anchor (Fig. 12-30).

Asn

Ser (Thr)

Phosphoethanolamine C-terminal

residue

O C

(CH

)

(a)

(b)

(c)

OP O

–

(d)

OHHO

CHR

Mannose

5-Hydroxylysine

O CH

CHCH

CHCH CH

O carbohydrate

–

Isoprene unit

Dolichol

Saturated

-isoprene

unit

JWCL281_c23_871-900.qxd 6/8/10 9:43 AM Page 883

884 Chapter 23. Other Pathways of Carbohydrate Metabolism

growing glycolipid by specific glycosyltransferases to form a

common “core” structure. Each monosaccharide unit is

added by a unique glycosyltransferase (Fig. 23-16). For ex-

ample, in Reaction 2 of Fig. 23-16, five mannosyl units are

added through the action of five different mannosyltrans-

ferases, each with a different oligosaccharide-acceptor

specificity. The oligosaccharide core, the product of Reac-

tion 9 in Fig. 23-16, has the composition (N-acetylglu-

cosamine)

(mannose)

(glucose)

Although nucleotide sugars are the most common

monosaccharide donors in glycosyltransferase reactions,

several mannosyl and glucosyl residues are transferred to

the growing dolichol-PP-oligosaccharide from their cor-

responding dolichol-P derivatives. This requirement for

dolichol-P-mannose was discovered by Stuart Kornfeld,

who found that mutant mouse lymphoma cells (lym-

phoma is a type of cancer) that are unable to synthesize

normal lipid-linked oligosaccharides formed a defective,

smaller glycolipid. These cells contain all the requisite

glycosyltransferases but are unable to synthesize

dolichol-P-mannose (Reaction 4 in Fig. 23-16 is blocked).

When this substance is supplied to the mutant cells, man-

nosyl units are added to the defective dolichol-PP-

oligosaccharide.

Figure 23-16 Pathway of dolichol-PP-oligosaccharide synthesis.

(1) Addition of N-acetylglucosamine-1-P and a second

N-acetylglucosamine to dolichol-P. (2) Addition of five mannosyl

residues from GDP–mannose in reactions catalyzed by five

different mannosyltransferases. (3) Membrane translocation of

dolichol-PP-(N-acetylglucosamine)

(mannose)

to the lumen of

the endoplasmic reticulum (ER). (4) Cytosolic synthesis of

dolichol-P-mannose from GDP–mannose and dolichol-P.

(5) Membrane translocation of dolichol-P-mannose to the lumen

of the ER. (6) Addition of four mannosyl residues from

dolichol-P-mannose in reactions catalyzed by four different

mannosyltransferases. (7) Cytosolic synthesis of dolichol-P-

2 UDP

Dolichol

5 GDP

1 UMP

1 UDP

CDP

CTP

4 GDP

3 UDP3 UDP

4 GDP

Lumen

Cytosol

Endoplasmic

reticulum

membrane

Endoplasmic Reticulum

mRNA

Ribosome

Nascent

polypeptide

= N-Acetylglucosamine

= Mannose

= Glucose

= Dolichol phosphate

glucose from UDP–glucose and dolichol-P. (8) Membrane

translocation of dolichol-P-glucose to the lumen of the ER.

(9) Addition of three glucosyl residues from dolichol-P-glucose.

(10) Transfer of the oligosaccharide from dolichol-PP to the

polypeptide chain at an Asn residue in the sequence Asn-X-

Ser/Thr, releasing dolichol-PP. (11) Translocation of dolichol-PP

to the cytoplasmic surface of the ER membrane. (12) Hydrolysis

of dolichol-PP to dolichol-P. (13) Dolichol-P can also be formed

by phosphorylation of dolichol by CTP. [Modified from Abeijon,

C. and Hirschberg, C.B., Trends Biochem. Sci. 17, 34 (1992).]

See the Animated Figures

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 884

d. Dolichol-PP-Oligosaccharide Synthesis Involves

Topological Changes of the Intermediates

Reactions 1, 2, 4, and 7 of Fig. 23-16 all occur on the cyto-

plasmic side of the endoplasmic reticulum (ER) membrane.

This was determined by using “right-side-out” rough ER

vesicles and showing that various membrane-impermeant

reagents can disrupt one or another of these reactions. Re-

actions 6, 9, and 10 occur in the lumen of the ER as judged

by the inability of concanavalin A, a lectin (carbohydrate-

binding protein), to bind to the products of these reactions

until the membrane is permeabilized. The (mannose)

(N-acetylglucosamine)

-PP-dolichol product of Reaction

2, the dolichol-P-mannose product of Reaction 4, and the

dolichol-P-glucose product of Reaction 7 must therefore

be translocated across the ER membrane (Reactions 3, 5,

and 8) such that they extend from its luminal surface in

order for the synthesis of N-linked oligosaccharides to

continue.The translocations are mediated by specific ATP-

independent flippases.

e. N-Linked Oligosaccharides Are Cotranslationally

Added to Proteins

Vesicular stomatitis virus (VSV), which infects cattle, pro-

ducing influenza-like symptoms, provides an excellent model

system for studying N-linked glycoprotein processing. The

VSV coat consists of host-cell membrane in which a single

viral glycoprotein, the VSV G-protein (not to be confused

with the GTPases involved in signal transduction; Chap-

ter 19), is embedded. Since a viral infection almost totally

usurps an infected cell’s protein synthesizing machinery, a

VSV-infected cell’s Golgi apparatus, which normally con-

tains hundreds of different types of glycoproteins, contains

virtually no other glycoprotein but G-protein. Consequently,

the maturation of the G-protein is relatively easy to follow.

Of the Asn-X-Ser/Thr sites in mature eukaryotic pro-

teins, 70 to 90% are N-glycosylated. Studies of VSV-

infected cells indicate that the transfer of the lipid-linked

oligosaccharide to a polypeptide chain occurs while the

polypeptide chain is still being synthesized. Structural pre-

dictions (Section 9-3A), together with glycosylation studies

of model polypeptides, suggest that the amino acid sequences

flanking known N-glycosylation sites occur at  turns or

loops in which Asn’s backbone N¬H group is hydrogen

bonded to the Ser/Thr hydroxyl O atom (Fig. 23-17a). This

explains why Pro cannot occupy the X position; it would

prevent Asn-X-Ser/Thr from assuming the putative re-

quired hydrogen bonded conformation.

VSV G-protein is N-glycosylated by oligosaccharyltrans-

ferase (OST), a membrane-bound, ⬃300-kD, 8-subunit

enzyme that recognizes the amino acid sequence Asn-X-

Ser/Thr (Fig. 23-16, Reaction 10). Ernst Bause has proposed

a catalytic mechanism for OST in which an enzyme base

Section 23-3. Biosynthesis of Oligosaccharides and Glycoproteins 885

(b)

–

Enzyme

Dol-PP–OH

Inactive, doubly

labeled, enzyme

Epoxyethyl-Gly

Asn

Dol-PP

(a)

–

N-linked

glycoprotein

Enzyme

ⴙ

Asn

Dol-PP

Dol-PP–OH

Thr

Figure 23-17 The

oligosaccharyltransferase

(OST) reaction. (a) The

Asn-X-Thr component of a

hexapeptide model substrate

forms a ring that is closed by

a hydrogen bond from the

Asn amide group to the Thr

hydroxyl group. A base on the

enzyme facilitates the

nucleophilic displacement of

dolichol pyrophosphate from

the oligosaccharide (Sac) by

the amide nitrogen. (b) The

inactivation of the OST by

reacting it with a hexapeptide

containing Asn-Gly-

epoxyethylGly in the presence

of dolichol-PP-oligosaccharide.

This chemically labels the

base with the oligopeptide to

which the oligosaccharide has

become covalently linked.

JWCL281_c23_871-900.qxd 6/8/10 9:43 AM Page 885

abstracts a proton from the Ser/Thr hydroxyl group, which in

turn abstracts a proton from the Asn NH

group, thereby

promoting its nucleophilic attack on the oligosaccharide

(Sac), which then displaces the dolichol pyrophosphate (Fig.

23-17a). This mechanism is supported by the observation

that reacting the OST with dolichol-PP-oligosaccharide and

a hexapeptide model substrate containing the sequence

Asn-X-epoxyethylGly (rather than Asn-X-Ser/Thr) irre-

versibly inactivates the enzyme by covalently linking it to

the now glycosylated hexapeptide (Fig. 23-17b).

f. The Calnexin/Calreticulin Cycle Facilitates

Glycoprotein Folding

The processing of an N-linked core oligosaccharide

begins in the endoplasmic reticulum by the enzymatic trim-

ming (removal) of its three glucose residues (Fig. 23-18, Re-

actions 2 and 3) and one of its mannose residues (Fig. 23-18,

Reaction 4) before the protein has folded to its native

conformation. This is not a straightforward process, how-

ever, because UDP–glucose:glycoprotein glucosyltrans-

ferase (GT), a 1513-residue soluble protein, reglucosylates

the oligosaccharides of partially folded glycoproteins, a reac-

tion that reverses the removal of the last of the three glucose

residues by glucosidase II (Fig. 23-18, Reaction 3).This futile

cycle (most glycoproteins undergo reglucosylation at least

once) is part of a chaperone-mediated glycoprotein folding

process called the calnexin/calreticulin cycle. Calnexin

(CNX; ⬃570 residues), which is membrane bound, and cal-

reticulin (CRT; ⬃400 residues), its soluble homolog, are ER-

resident lectins that bind partially folded glycoproteins bear-

ing a monoglucosylated oligosaccharide in a way that

protects the glycoprotein from degradation and premature

transfer to the Golgi apparatus. If the glycoprotein is re-

leased and deglucosylated before it has correctly folded, GT,

which recognizes only non-native glycoproteins, reglucosy-

lates it so that the CNX/CRT cycle can repeat. CNX and

CRT both also bind ERp57, a 481-residue thiol oxidoreduc-

tase homologous to protein disulfide isomerase (PDI; Sec-

tion 9-2A). While the partially folded glycoprotein is bound

to the complex, ERp57 catalyzes disulfide interchange reac-

tions to facilitate the formation of the correctly paired disul-

fide bonds. The CNX/ ERp57 and CRT/ERp57 complexes

are therefore responsible for the correct folding and disul-

fide bond formation of the glycoproteins in the ER.The im-

portance of this process is demonstrated by the observation

that knockout mice lacking the gene for CRT die in utero.

The X-ray structure of the luminal domain of calnexin

(residues 61–458), determined by Miroslaw Cygler, reveals

a most unusual structure (Fig. 23-19): a compact globular

domain (residues 61–262 and 415–458) from which extends

886 Chapter 23. Other Pathways of Carbohydrate Metabolism

Figure 23-18 The calnexin/calreticulin cycle for glycoprotein

folding in the endoplasmic reticulum. The reactions are catalyzed

by: (1) oligosaccharyltransferase (OST); (2) -glucosidase I; (3)

-glucosidase II, UDP–glucose:glycoprotein glucosyltransferase

= Mannose

= N-Acetylglucosamine

= Glucose

UDP

ERp57

CNX

Transfer via vesicles

to the cis Golgi network

CRT

(GT), calreticulin (CRT), calnexin (CNX), and the thiol

oxidoreductase ERp57; and (4) ER -1,2-mannosidase. [After

Helenius,A. and Aebi, M., Science 291, 2367 (2001).]

JWCL281_c23_871-900.qxd 6/8/10 9:43 AM Page 886

a 145-Å-long arm (residues 270–414).The globular domain

forms a sandwich of a 6-stranded and a 7-stranded antipar-

allel  sheet that binds a Ca

2

ion and which resembles

legume lectins such as concanavalin A (Fig. 8-40). This do-

main binds glucose on its concave (blue) surface, which is

lined by hydrogen bonding groups that model building

suggests binds the (glucose)

(mannose)

portion of cal-

nexin’s natural (glucose)

(mannose)

substrate. The long

arm, which consists of an extended hairpin, is known as the

P domain because it has four copies each of two different

Pro-rich motifs arranged in the sequence 11112222, with

each ⬃18-residue motif 1 in antiparallel association with

an ⬃14-residue motif 2 on the opposite strand of the hair-

pin. Each of these motif pairs has a similar structure, with

its conserved residues maintaining identical interactions in

each pair.The P domain has been shown to form the bind-

ing site for ERp57 in both calnexin and calreticulin.

g. Glycoprotein Processing is Completed

in the Golgi Apparatus

Once a glycoprotein has folded to its native conforma-

tion and ER -1,2-mannosidase has removed one of its

mannosyl residues (Fig. 23-18, Step 4), the glycoprotein is

transported, in membranous vesicles, to the Golgi appara-

tus, where it is further processed (Fig. 23-20).The Golgi ap-

paratus (Fig. 12-58), as we discussed in Section 12-4C, con-

sists of, from opposite the ER outward, the cis Golgi

Section 23-3. Biosynthesis of Oligosaccharides and Glycoproteins 887

Figure 23-19 X-ray structure of the luminal portion of canine

calnexin. The protein is drawn in ribbon form embedded in its

semitransparent molecular surface. The 6- and 7-stranded

antiparallel  sheets of its globular domain are colored orange

and blue, with its remaining portions gray and its bound Ca

2

ion

represented by a blue-green sphere. In the P domain, motifs 1 are

alternately colored green and yellow and motifs 2 are alternately

colored magenta and cyan. [Based on an X-ray structure by

Miroslaw Cygler, Biotechnology Research Institute, NRC,

Montreal, Quebec, Canada. PDBid 1JHN.]

Figure 23-20 Oligosaccharide processing of VSV G-protein in

the Golgi network. The reactions are catalyzed by: (1) Golgi

-mannosidase I, (2) N-acetylglucosaminyltransferase I, (3)

Golgi -mannosidase II, (4) N-acetylglucosaminyltransferase II,

(5) fucosyltransferase, (6) galactosyltransferase, and

(7) sialyltransferase. Lysosomal proteins are modified by:

(I) N-acetylglucosaminyl phosphotransferase and

(II) N-acetylglucosamine-1-phosphodiester

-N-acetylglucosaminidase. [Modified from Kornfeld, R. and

Kornfeld, S., Annu. Rev. Biochem. 54, 640 (1985).]

II 1I

P P P P

UDP

Via vesicles from the

Endoplasmic Reticulum

Exit

cis Golgi

UDP

2UDP

GDP

3 45

Via vesicles

medial Golgi

3CMP

Via vesicles

trans Golgi

Polypeptide

3UDP

= Galactose

= N-Acetylglucosamine

= Mannose

= Sialic acid

= L-Fucose

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 887

network, through which glycoproteins enter the Golgi ap-

paratus; a stack of at least three different types of sacs,

the cis, medial, and trans cisternae; and the trans Golgi net-

work, through which proteins exit the Golgi apparatus.

Glycoproteins traverse the Golgi stack, from the cis to the

medial to the trans cisternae, each of which, as shown by

James Rothman and Kornfeld, contains different sets of

glycoprotein processing enzymes. As this occurs, mannose

residues are trimmed from each oligosaccharide group and

N-acetylglucosamine, galactose, fucose, and/or sialic acid

residues are added to complete the processing of the glyco-

protein (Fig. 23-20; Reactions 1–7). The glycoproteins are

then sorted in the trans Golgi network for transport to

their respective cellular destinations via membranous vesi-

cles (Sections 12-4C and 12-4D).

There is enormous diversity among the different

oligosaccharides of N-linked glycoproteins, as is indicated,

for example, in Fig. 11-32c. Indeed, even glycoproteins with

a given polypeptide chain exhibit considerable microhetero-

geneity (Section 11-3C), presumably as a consequence of

incomplete glycosylation and lack of absolute specificity

on the part of glycosyltransferases and glycosylases.

The processing of all N-linked oligosaccharides is iden-

tical through Reaction 4 of Fig. 23-18, so that all of them

have a common (N-acetylglucosamine)

(mannose)

core

(five “noncore” mannose residues are subsequently trimmed

from VSV G-protein; Fig. 23-20, Reactions 1 and 3). The

diversity of the N-linked oligosaccharides therefore arises

through divergence from this sequence after Fig. 23-20,

Reaction 3. The resulting oligosaccharides are classified

into three groups:

1. High-mannose oligosaccharides (Fig. 23-21a), which

contain 2 to 9 mannose residues appended to the common

pentasaccharide core (red residues in Fig. 23-21).

2. Complex oligosaccharides (Fig. 23-21b), which con-

tain variable numbers of N-acetyllactosamine units as well

as sialic acid and/or fucose residues linked to the core.

3. Hybrid oligosaccharides (Fig. 23-21c), which contain

elements of both high-mannose and complex chains.

It is unclear how different types of oligosaccharides are

related to the functions and/or final cellular locations of

their glycoproteins. Lysosomal glycoproteins, however, ap-

pear to be of the high-mannose variety.

h. Inhibitors Have Aided the Study

of N-Linked Glycosylation

Elucidation of the events in the glycosylation process

has been greatly facilitated through the use of inhibitors

that block specific glycosylation enzymes. Two of the most

useful are the antibiotics tunicamycin (Fig. 23-22a), a hy-

drophobic analog of UDP–N-acetylglucosamine, and baci-

tracin (Fig. 23-23), a cyclic polypeptide. Both were discov-

ered because of their ability to inhibit bacterial cell wall

biosynthesis, a process that also involves the participation

of lipid-linked oligosaccharides. Tunicamycin blocks the

formation of dolichol-PP-oligosaccharides by inhibiting

the synthesis of dolichol-PP-N-acetylglucosamine from

dolichol-P and UDP–N-acetylglucosamine (Fig. 23-16, Re-

action 1). Tunicamycin resembles an adduct of these reac-

tants (Fig. 23-22b) and, in fact, binds to the enzyme with a

dissociation constant of 7 ⫻ 10

⫺9

888 Chapter 23. Other Pathways of Carbohydrate Metabolism

Man Man Man

Man

ManMan

Man Man

Man

α 1,2

α 1,2 α 1,3

α 1,3 α 1,6

α 1,6

α 1,2

α 2,3, or 6 α 2,3, or 6

β 1,4

β 1,2

β 1,4

β 1,2

β 1,4

α 1,6

α 1,3

α 1,6

β 1,4

α 1,2

SA SA

GalGal

GlcNAc GlcNAc

Man

Man Man

Man

β 1,2

β 1,4

Gal

GlcNAc

β 1,4

GlcNAc

Asn Asn

GlcNAc

Asn

–

GlcNAc

β1,4

GlcNAc

–

Fuc

High mannose Complex Hybrid

(a) (b)

(c)

Figure 23-21 Types of N-linked oligosaccharides. Typical

primary structures of (a) high-mannose, (b) complex, and

common to all N-linked oligosaccharides is indicated in red.

[After Kornfeld, R. and Kornfeld, S., Annu. Rev. Biochem. 54,

633 (1985).]

JWCL281_c23_871-900.qxd 7/2/10 12:03 PM Page 888

Bacitracin forms a complex with dolichol-PP that in-

hibits its dephosphorylation (Fig. 23-16, Reaction 12),

thereby preventing glycoprotein synthesis from lipid-

linked oligosaccharide precursors. Bacitracin is clinically

useful because it destroys bacterial cell walls but does not

affect animal cells because it cannot cross cell membranes

(bacterial cell wall biosynthesis is an extracellular

process).

Section 23-3. Biosynthesis of Oligosaccharides and Glycoproteins 889

(a)

(b)

CCC

OHH

CH (CH

)

CHOH

Tunicamycin

n = 8,9,10, or 11

–

O O

–

CCHCH

O P O

–

CHCH

Dolichol phosphate

UDP-N-Acetylglucosamine

Bacitracin

OCHis

(CH

)

AsnD-Asp

D-PheIleD-OrnIle

CysIle

D-Glu LysLeuC

CHCHH

Figure 23-22 Chemical structure of tunicamycin. The

structure of (a) the glycosylation inhibitor tunicamycin is

compared to that of (b) dolichol-P  UDP–N-

acetylglucosamine.

Figure 23-23 Chemical structure of bacitracin. Note that this

dodecapeptide has four

D-amino acid residues and two unusual

intrachain linkages.“Orn” represents the nonstandard amino

acid residue ornithine (Fig. 26-7).

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 889

i. O-Linked Oligosaccharides Are Post-

Translationally Formed

The study of the biosynthesis of mucin, an O-linked gly-

coprotein secreted by the submaxillary salivary gland, indi-

cates that O-linked oligosaccharides are synthesized in the

Golgi apparatus by serial addition of monosaccharide units

to a completed polypeptide chain (Fig.23-24).Synthesis starts

with the transfer of N-acetylgalactosamine (GalNAc) from

UDP–GalNAc to a Ser or Thr residue on the polypeptide by

GalNAc transferase. In contrast to N-linked oligosaccha-

rides, which are transferred to an Asn in a specific amino

acid sequence, the O-glycosylated Ser and Thr residues are

not members of any common sequence. Rather, it appears

that the location of glycosylation sites is specified only by

the secondary or tertiary structure of the polypeptide. Gly-

cosylation continues with stepwise addition of galactose,

sialic acid, N-acetylglucosamine, and/or fucose by the corre-

sponding glycosyltransferases.

j. Oligosaccharides on Glycoproteins Act as

Recognition Sites

Glycoproteins that are synthesized in the endoplasmic

reticulum and processed in the Golgi apparatus are targeted

for secretion, insertion into cell membranes, or incorpora-

tion into cellular organelles such as lysosomes. This suggests

that oligosaccharides serve as recognition markers for this

sorting process. For example, the study of I-cell disease (Sec-

tion 12-4Cg) demonstrated that in glycoprotein enzymes

destined for the lysosome, a mannose residue is converted to

mannose-6-phosphate (M6P) in the cis cisternae of the

Golgi. The process involves two enzymes (Fig. 23-20, Reac-

tions I and II), which are thought to recognize lysosomal

protein precursors by certain structural features on these

proteins rather than a specific amino acid sequence. In the

trans Golgi network, M6P-bearing glycoproteins are sorted

into lysosome-bound coated vesicles through their specific

binding to one of two M6P receptors, one of which is a 275-

kD membrane glycoprotein called the M6P/IGF-II receptor

(because it has been found that this M6P receptor and the

insulinlike growth factor II receptor are the same protein).

Individuals with I-cell disease lack the enzyme catalyzing

mannose phosphorylation (Fig. 23-20, Reaction I), resulting

in the secretion of the normally lysosome-resident enzymes.

ABO blood group antigens (Section 12-3E) are O-linked

glycoproteins.Their characteristic oligosaccharides are com-

ponents of both cell-surface lipids and of proteins that occur

in various secretions such as saliva. These oligosaccharides

form antibody recognition sites.

Glycoproteins are believed to mediate cell–cell recogni-

tion. For example, an O-linked oligosaccharide on a glyco-

protein that coats the mouse ovum surface (zona pellu-

cida) acts as the sperm receptor. Even when this

oligosaccharide is separated from its protein, it retains the

ability to bind mouse sperm.

k. GPI-Linked Proteins

Glycosylphosphatidylinositol (GPI) groups function to

anchor a wide variety of proteins to the exterior surface of

the eukaryotic plasma membrane, thus providing an alter-

native to transmembrane polypeptide domains (Section

12-3Bc; Fig. 12-30). This anchoring results from transami-

dation of a preformed GPI glycolipid within 1 min of the

synthesis and transfer of a target protein to the ER.

Biosynthesis of the GPI core structure (Fig. 23-25a) be-

gins on the cytoplasmic side of the ER with the transfer of

890 Chapter 23. Other Pathways of Carbohydrate Metabolism

Figure 23-24 Proposed synthesis pathway for the carbohydrate

moiety of an O-linked oligosaccharide chain of canine

submaxillary mucin. SA and Fuc represent sialic acid and fucose.

Figure 23-25 GPI anchors. (Opposite) (a) The pathway of

synthesis of the tetrasaccharide core of glycophosphatidylinositol

(GPI).The following enzymes and steps are involved: (1) UDP–

GlcNAc:PI 1 S 6 N-acetylglucosaminyltransferase complex,

(2) GlcNAc–PI de-N-acetylase, (3) inositol acyltransferase,

(4) Dol-P-Man:GlcN–PI/GlcN–(acyl)PI 1 S 4 mannosyltrans-

ferase (MT-I), (5) an ethanolamine phosphotransferase, (6) Dol-

P-Man:Man

GlcN–(acyl)PI 1 S 6 mannosyltransferase (MT-II),

(7) Dol-P–Man:Man

GlcN–(acyl)PI 1 S 2 mannosyltransferase

(MT-III), (8) lipid remodeling (replacement of the fatty acyl

groups on PI), and (9) transfer of phosphoethanolamine from

phosphatidylethanolamine to the 6-hydroxyl group of the

terminal mannose residue of the core tetrasaccharide by an

ethanolamine phosphotransferase. (b) Transamidation of the target

protein, resulting in a C-terminal amide link to the GPI anchor.

UDP

Ser

GalNAc

transferase

UDP

Ser

CMP

Ser

GDP

Ser

GalNAc



GalGalNAc

Gal

β 1,3

GalGalNAc

α 2,6

Fuc

Ser

GalGalNAc

 1,2

Fuc

JWCL281_c23_871-900.qxd 3/24/10 11:33 AM Page 890