Glycoconjugate biosynthesis requires activation of monosaccharides to nucleotide sugars. Monosaccharides are imported into the cell, salvaged from degraded glycans, or derived from other sugars within the cell. Although most glycosylation reactions occur in the Golgi, precursor activation and interconversions occur mostly in the cytoplasm. nucleotide sugar–specific transporters carry the activated donors into the Golgi. This chapter describes how eukaryotic cells accomplish these tasks.

GENERAL PRINCIPLES

Glucose and fructose are the major carbon and energy sources for organisms as diverse as yeast and humans. Both organisms can derive the other monosaccharides needed for glycan biosynthesis from these sources. Not all of these biosynthetic pathways are equally active in all types of cells. However, there are some general principles. Glycan synthesis requires the monosaccharides to be activated to a high-energy donor form. This process requires nucleoside triphosphates (such as UTP or GTP) and a glycosyl-1-P (monosaccharide with a phosphate at the anomeric carbon). Several variations are used, but regardless of the monosaccharide, all must be either activated by a kinase (reaction 1) or generated from a previously synthesized activated nucleotide sugar (reactions 2 and 3):

The most common nucleotide sugar donors in animal cells are shown in Table 4.1. Sialic acids are the only monosaccharides in animals activated as a mononucleotide, CMP-Sia. Iduronic acid does not have a nucleotide sugar parent because it is formed by epimerization of glucuronic acid after its incorporation into glycosaminoglycan (GAG) chains. In some instances, one nucleotide sugar can be formed from another by a nucleotide exchange reaction (reaction 3 above). For example, UDP-Gal is made from UDP-Glc by exchange of Gal-1-P for Glc-1-P.

TABLE 4.1

Activated sugar donors

EXTERNAL SOURCES AND TRANSPORTERS

Three types of sugar transporters carry sugars across the plasma membrane into cells. First are energy-independent facilitated diffusion transporters such as the glucose transporter (GLUT) family of hexose transporters found in yeast and mammalian cells. The genes encoding these proteins are named SLC2A (solute carriers 2A). Second are energy-dependent transporters, for example, the sodium-dependent glucose transporters (SGLT; gene names SLC5A) in intestinal and kidney epithelial cells. The third type are transporters that couple ATP-dependent phosphorylation with sugar import. These are found in bacteria and are not covered in this chapter.

GLUT family transporters were first described in yeast, where at least 18 genes are now known. Humans have 14 GLUT homologs. All of the yeast GLUT transporters are about the same size (40–55 kD) and have similar structures containing 12 membrane-spanning domains, which is typical of many eukaryotic transporters. The transmembrane domains form a barrel with a small pore for the sugar to pass through. Compared to GLUT1, the other family members have a modest 28–65% amino acid identity. The only “sugar transporter signatures” are a few widely scattered glycine and tryptophan residues and one PET tripeptide sequence. That is, there are no obvious major transporter motifs.

Typically, the GLUTs have K_m values for glucose uptake in the 2–20-mM range. In yeast, many transport glucose, but others are specific for galactose, fructose, or disaccharides. Most mammalian GLUT proteins transport glucose or fructose with variable efficiency, but detailed specificity studies for this family have not been done. However, GLUT5 transports fructose, and the GLUT called HMIT is a proton-coupled myo-inositol transporter. GLUT2 also efficiently transports glucosamine.

Glucose is transported from the gut by an energy-requiring Na⁺-dependent glucose transporter (SGLT-1) and is recovered from the kidney filtrates by a related transporter (SGLT-2). The SGLT-type transporters have K_m values of less than 1 mM.

GLUT1–5 have different distributions in mammalian cells and different K_m values that enable them to respond to the availability of glucose (Table 4.2). Although most of the human GLUT members are located on the cell surface, a portion of GLUT4 is associated with intracellular vesicles and recruited to the cell surface in response to insulin. Following carbohydrate-rich meals, glucose transported by SGLT1 in the intestine is thought to promote the recruitment of GLUT2 to the apical surface for enhanced glucose uptake.

TABLE 4.2

Monosaccharide transporters

INTRACELLULAR SOURCES OF MONOSACCHARIDES

Activation and Interconversion of Monosaccharides

Glycogen

Glycogen is an immense molecule that contains up to 100,000 glucose units, arranged in Glcα1–4Glc repeating disaccharides with periodic α1–6Glc branches. It is synthesized on a cytoplasmic protein called glycogenin (see Chapter 19). Glycogen is the major storage polysaccharide in animal cells, and its synthesis and degradation (glycogenolysis) are highly regulated for energy use. Glycogen is synthesized by the addition of single glucose units from UDP-Glc, and it is degraded by glycogen phosphorylase. The non-ATP-dependent reaction forms glucose-1-P by phosphorolysis of glycogen, which can be used directly to form UDP-Glc or converted to glucose-6-P for further catabolism via glycolysis or direct oxidation via glucose-6-phosphate dehydrogenase.

Glucose

Glucose is the central monosaccharide in carbohydrate metabolism, and it can be converted into all other sugars (Figure 4.1). Glucose is first converted to glucose-6-P by hexokinase. In the glycolytic pathway, glucose-6-P is converted to fructose-6-P by phosphoglucose isomerase or into glucose-1-P by phosphoglucomutase. Reaction of glucose-1-P with UTP forms the high-energy donor UDP-Glc. The UDP-Glc pool is quite large, and it is used to synthesize glycogen and other glucose-containing molecules such as glucosylceramide (see Chapter 10) and dolichol-P-glucose, which is used in the N-linked glycan biosynthetic pathway (see Chapter 8).

FIGURE 4.1

Biosynthesis and interconversion of monosaccharides. The relative contributions of each pathway under physiological conditions are unknown. (Rectangles) Donors; (ovals) monosaccharides; (asterisks) control points; (6PG) 6-phosphogluconate; (PEP) phosphoenolpyruvate; (more...)

Some glucose-6-phosphate is acted on by glucose-6-P dehydrogenase, the entry point for the oxidation via the pentose phosphate pathway, to form 6-phosphogluconate and ribose-5-phosphate. These reactions generate NADPH⁺, which is needed to maintain proper redox status.

Glucuronic Acid

UDP-GlcA is synthesized directly from UDP-Glc by a two-stage reaction requiring two NAD⁺-dependent oxidations at C-6. UDP-GlcA is used primarily for GAG biosynthesis (see Chapters 15 and 16), but some N- and O-linked glycans and glycosphingolipids contain glucuronic acid as well. The addition of glucuronic acid to bile acids and xenobiotic compounds increases their solubility, and a large class of glucuronosyl transferases is devoted to such reactions.

Iduronic Acid

Iduronic acid is the C-5 epimer of glucuronic acid, and it is found in the GAGs, dermatan sulfate, heparan sulfate, and heparin. Unlike all other monosaccharides found in glycans, iduronic acid is not directly synthesized from a nucleotide sugar donor. Instead, it is created by epimerization of glucuronic acid after it has been incorporated into the growing GAG chain (see Chapter 16).

Xylose

Decarboxylation of UDP-GlcA gives UDP-Xyl, which is used primarily to initiate GAG synthesis (see Chapter 16). Xylose is also found on proteins that have O-glucose modifications in EGF modules (see Chapter 12). A type II membrane protein performs the decarboxylation reaction using UDP-GlcA transported into the ER or Golgi. In C. elegans, the decarboxylase is called SQV-1, and it colocalizes with the UDP-GlcA transporter (see Chapter 23). In Arabidopsis, another UDP-GlcA decarboxylase also occurs in the cytoplasm, but no ortholog has been identified in animals.

Mannose

Mannose is used for multiple types of glycans (see Chapters 8, 9, 10, 11, and 12). GDP-Man is the primary activated donor. Its production requires prior synthesis of mannose-6-P and its conversion to mannose-1-P. There are two ways to produce mannose-6-P. The first is by direct phosphorylation via hexokinase. The second is through conversion of fructose-6-P to mannose-6-P using the enzyme phosphomannose isomerase. In yeast, loss of this enzyme is lethal. In humans, loss of this enzyme produces a potentially fatal disease called congenital disorder of glycosylation (type Ib) (see Chapter 42). The importance of phosphomannose isomerase is easy to understand because free exogenous mannose is not common in the diet, and this enzyme is the key link between mannose and glucose. Both yeast and human phosphomannose isomerase deficiencies can be rescued by providing exogenous mannose.

Although mannose can be used directly for glycan synthesis, in the “honeybee syndrome” mannose can be lethal! This curious phenomenon occurs when honeybees are fed mannose instead of sucrose or glucose. The bees behave normally for several minutes, but then they suddenly die. The reason is that mannose enters the cells and is phosphorylated by abundant hexokinase using ATP. Because mannose-6-P is now the sole energy source for the bees, it must be converted into fructose-6-P to enter glycolysis. The problem is that honeybees have relatively low phosphomannose isomerase activity compared to hexokinase. This creates a bottleneck and excess mannose-6-P accumulates. It is quickly degraded by phosphatase to free mannose, which is again phosphorylated using the diminishing supply of ATP. Multiple futile cycles deplete the ATP pool and death results. For similar reasons, high mannose is teratogenic in rats, because during early development the embryo relies more on glycolysis rather than oxidative phosphorylation for ATP production. Mice totally lacking phosphomannose isomerase activity die in utero because of a similar honeybee effect.

In mammals, mannose-6-P is converted to mannose-1-P using phosphomannomutase. For this conversion, two isozymes are known in humans, and the loss of one of them produces another congenital disorder of glycosylation (type Ia) that results from underglycosylation of proteins (see Chapter 42). Because mannose-6-P or mannose-1-P are both obligate precursors of GDP-Man, failure to make sufficient amounts of either one reduces the formation of GDP-Man. GDP-Man can be used directly for the formation of the lipid-linked oligosaccharide on the cytosolic face of the ER. GDP-Man can also transfer mannose to dolichol phosphate to form dolichol-P-mannose in the ER membrane.

Mannose-6-P can also condense with phosphoenolpyruvate to form 2-keto-3-deoxy-D-glycero-D galactonononic acid (KDN). This molecule is activated with CTP to produce CMP-KDN. KDN is abundant in trout testis and on their sperm, where it is thought to be important for sperm-egg adhesion.

Fucose

GDP-Fuc can be derived from GDP-Man by the sequential action of two enzymes involving three steps. In the first step, the C-4 hydroxyl group of GDP-Man is oxidized to a ketone (GDP-4-keto-6-deoxy-mannose) by the enzyme GDP-Man 4,6-dehydratase along with the reduction of NADP to NADPH. The next two reactions are catalyzed by a single polypeptide that has epimerase and reductase activity and is well conserved from bacteria to mammals. GDP-4-keto-6-deoxymannose is epimerized at C-3 and C-5 to form GDP-4-keto-6-deoxyglucose, which is then reduced with NADPH at C-4 to form GDP-Fuc (Figure 4.2). The first dehydration step is feedback inhibited by GDP-Fuc. GDP-Fuc can also be synthesized directly from fucose. The first step uses a kinase to make fucose-1-P, which is then converted to GDP-Fuc. Mutant CHO cells that cannot convert GDP-Man to GDP-Fuc form hypofucosylated proteins, but this can be corrected by providing exogenous fucose in the medium. Also, mice genetically deficient in the GDP-Man to GDP-Fuc conversion can be rescued by providing fucose in their food or drinking water. This finding suggests that some transporters may carry other sugars besides glucose, but the quantitative contribution of this pathway is not known. The free fucose concentration in the blood is very low, a few micromolar at most.

FIGURE 4.2

Conversion of activated sugar donors. Steps in the synthesis of GDP-Fuc from GDP-Man (A) and UDP-Gal from UDP-Glc (B). Details of the various enzymes are given in the text. GDP-Fuc synthesis by this route is not reversible, whereas the interconversion (more...)

Galactose

Activated UDP-Gal can be made in several ways. The first is by direct phosphorylation at C-1 to give galactose-1-P, which reacts with UTP to form UDP-Gal. Alternatively, galactose-1-P can be converted to UDP-Gal via a uridyl transferase exchange reaction with UDP-Glc that displaces glucose-1-P. A deficiency in this activity results in a severe human disease called galactosemia, which leads to mental retardation, liver damage, and eventual death if galactose intake is not controlled (see Chapter 42). Finally, UDP-Gal can be formed from UDP-Glc by the NAD-dependent reaction catalyzed by UDP-Gal 4-epimerase. The enzyme first converts the C-4 hydroxyl group to a keto derivative forming NADH from bound NAD⁺. In the next step, the keto group is converted back to a hydroxyl group with opposite orientation and NAD⁺ reforms (Figure 4.2). The same enzyme interconverts UDP-GalNAc and UDP-GlcNAc.

Galactose usually occurs as a pyranose (p) ring in higher animals, but bacteria and pathogenic eukaryotes such as Leishmania and Aspergillus incorporate galactofuranose (f) into their glycans (see Chapter 20). The donor is formed by conversion of UDP-Gal(p)ØUDP-Gal(f) using a flavinadenine-dinucleotide-dependent mutase.

N-Acetylglucosamine

Synthesis of UDP-GlcNAc begins with the formation of glucosamine-6-P from fructose-6-P by transamination using glutamine as the –NH₂ donor. Glucosamine-6-P is then N-acetylated via an acetyl-CoA-mediated reaction to form N-acetylglucosamine-6-P and then isomerized to N-acetylglucosamine-1-P via a 1,6-bis-phosphate intermediate. Similar to the other activation reactions, N-acetylglucosamine-1-P then reacts with UTP to form UDP-GlcNAc and pyrophosphate. Alternatively, GlcNAc can be directly phosphorylated to form N-acetyl-glucosamine-6-P via a kinase that can use either GlcNAc or N-acetylmannosamine. Phospho-N-acetylglucosamine mutase then converts this to N-acetylglucosamine-1-P. This route may account for the efficient salvage of GlcNAc from lysosomal degradation of glycans. Glucosamine can also be used following sequential phosphorylation and acetylation.

N-Acetylgalactosamine

UDP-GalNAc can arise from two routes. One is the direct reaction of N-acetylgalactosamine-1-P with UTP. N-Acetylgalactosamine-1-P is formed by a specific kinase that is distinct from galactose-1-kinase. UDP-GalNAc can also be formed by epimerization of UDP-GlcNAc using the same NAD-dependent epimerase that converts UDP-Glc to UDP-Gal.

Sialic Acids

The term sialic acid is the name given to a group of more than 50 different variations of the two parent compounds, N-acetylneuraminic acid (Neu5Ac), and N-glycolylneuraminic acid (Neu5Gc) as discussed more fully in Chapter 14. Except for the formation of the N-glycolyl derivative as an activated nucleotide sugar, most other modifications of sialic acid probably occur in the Golgi after transfer of the sialic acid to the oligosaccharide acceptor. Formation of CMP-N-acetyl/N-glycolylneuraminic acid is more complicated than formation of the other activated sugars. First, UDP-GlcNAc is converted to N-acetylmannosamine by a single enzyme with two catalytic activities located in different domains of UDP-GlcNAc epimerase/kinase. The first activity involves epimerization at C-2 and cleavage of the UDP to yield N-acetylmannosamine. In the next reaction, this enzyme acts as a kinase using ATP to form N-acetylmannosamine-6-P. Mutations in this enzyme cause two completely distinct metabolic disorders: sialuria and inclusion body myopathy type 2 (see Chapter 42). Knocking out this gene in mice causes early embryonic lethality. In the next step, N-acetylmannosamine-6-P is condensed with phosphoenolpyruvate to form N-acetylneuraminic acid-9-P. Phosphate is then removed by a phosphatase. Activation with CTP yields CMP-N-acetylneuraminic acid. The last step occurs in the nucleus with subsequent export of the activated precursor to the cytoplasm.

There also appear to be alternate pathways for the synthesis of activated sialic acid donors. UDP-GlcNAc epimerase/kinase activity is prominent in relatively few tissues including the liver, salivary gland, and intestinal mucosa, but clearly many other tissues contain sialylated glycoproteins and glycolipids. This finding suggests that other pathways probably exist. Sialic acids can be salvaged from glycoprotein turnover or from plasma and activated by phosphorylation and addition of CMP from CDP. In addition, N-acetylglucosamine 2′-epimerase can generate N-acetylmannosamine, which can be converted to N-acetylmannosamine-6-P by N-acetylglucosamine kinase.

Unusual Sugars in Bacteria and Plants

Fucose is the only deoxyhexose found in animal cell glycans. In contrast, bacterial and plant polysaccharides and glycoproteins frequently contain a variety of deoxysugars, deoxyaminosugars, and branched-chain sugars. These unusual sugars often have potent biological properties. For example, the glycan moiety of streptomycin binds to the minor groove of DNA to form stable antibiotic-DNA complexes. Deoxyhexoses are often immunological determinants of lipopolysaccharides or O-antigens of the Salmonella species. Five of the eight possible 3,6-dideoxyhexoses have been found in these organisms at the nonreducing end of the gram-negative cell wall lipopolysaccharide. Other deoxyhexoses, such as a 4,6-dideoxy-hexose and a 2,3,6-trideoxyhexose, are also biologically significant but uncommon in nature.

Biosynthesis of both deoxysugars and dideoxysugars begins with the oxidation of C-4, which is catalyzed by a 4,6-dehydratase to produce an NDP-4-keto-6-deoxyhexose. This is similar to the first step of the conversion of GDP-Man to GDP-Fuc. The nucleotide (N) differs for the various sugars, and the individual pathways use different dehydratases. For example, biosynthesis of most 3,6-dideoxyhexoses (except colitose) begins with conversion of CDP-glucose to CDP-4-keto-6-deoxyhexose by NAD⁺-dependent CDP-glucose dehydratase. In the biosynthesis of abequose (3,6-dideoxy-D-xylohexose), the product, CDP-6-deoxy-L-threo-D-glycero-hexulose, is then converted in two additional steps to CDP-3,6-dideoxy-D-glycero-D-glycero-4-hexulose by a second dehydratase followed by a reductase.

Amino sugars, such as glucosamine, arise from keto sugars by the addition of an amino group from glutamine (Figure 4.1). In addition, bacteria and plants have many 6-deoxy-hexoses with amino groups in the 2, 3, or 4 positions. For example, duanosamine is a 3-amino-6-deoxyhexose that is found in the antibiotic duanomycin. Here, TDP-glucose is dehydrated to 3-keto-6-deoxyglucose and the amino group is added via a transamination reaction probably involving a vitamin B6-dependent reaction.

Plants and bacteria also contain a number of branched-chain sugars. For instance, apiose is a component of the polysaccharide apiogalacturonan of Lemna minor, and strepose is a component of the antibiotic streptomycin produced by Streptomyces griseus. Apiose (see Figure 4.3A) is synthesized from UDP-GlcA via a 4-keto intermediate that can yield UDP-Xyl or UDP-apiose. Apiose synthesis removes carbon 3 from the chain to give the branched sugar by an unknown mechanism. Streptose (5-deoxy-3-C-formyl-L-lyxose) originates by an intramolecular rearrangement of the hexose of TDP-4-keto-6-deoxy-L-lyxohexose (Figure 4.3B). Although the synthesis of other branched-chain sugars has not been delineated, they probably follow similar reaction pathways.

FIGURE 4.3

Steps in the biosynthesis of (A) the branched sugar donor UDP-apiose from UDP-GlcA and (B) dTDP-dihydrostreptose from dTDP-glucose. A shows the conversion of UDP-GlcA into UDP-apiose, which is used in plant polysaccharides such as apiogalacturonan in (more...)

NUCLEOTIDE SUGAR TRANSPORTERS

In eukaryotes, the nucleotide sugars synthesized in the cytosol or the nucleus are on the “wrong” side of the membrane and must be transported into the ER and Golgi. Negative charge prevents these donors from simply diffusing into these compartments. To overcome this problem, eukaryotic cells have a set of energy-independent nucleotide sugar antiporters that deliver nucleotide sugars into the lumen of these organelles with the simultaneous exiting of nucleoside monophosphates; the latter must first be generated from the nucleoside diphosphates by a nucleoside diphosphatase (Figure 4.4). This transport mechanism was established biochemically in isolated vesicles and genetically in various mutant cell lines. The K_m of the transporters ranges from 1 to 10 μM. Using in vitro systems, the transporters have been shown to increase the concentration of the nucleotide sugars within the Golgi lumen by 10- to 50-fold. This is usually sufficient to reach or exceed the calculated K_m of glycosyltransferases that use these donors.

FIGURE 4.4

Transporters for nucleotide sugars, PAPS (3′ phosphoadenosine-5′-phosphosulfate), and ATP are located in the Golgi membranes of mammals, yeast, protozoa, and plants. These proteins are antiporters and return the corresponding nucleoside (more...)

Most of the antiporters are found in the Golgi, but some are also found in the ER. Thus, they are organelle specific and their location usually corresponds to the location of the downstream glycosyltransferases (Table 4.3 and Figure 4.4). Nucleotide sugar import into the Golgi is not energy dependent or affected by ionophores. However, the import is competitively inhibited by the corresponding nucleoside monophosphates and diphosphates, but not by the monosaccharides. ATP and PAPS (3′ phosphoadenosine-5′-phosphosulfate) transporters are also known; the latter is used for carbohydrate and protein sulfation.

TABLE 4.3

Nucleotide transport in Golgi and endoplasmic reticulum (ER)

Glucuronidation of bile and xenobiotic compounds in the ER is consistent with the presence of the UDP-GlcA transporter in the ER. The presence of a Golgi transporter is consistent with the location of the polymerases that use UDP-GlcA for the formation of GAG chains. The observation that reglycosylation of misfolded glycoproteins occurs in the ER (see Chapter 36) explains the need for an ER UDP-Glc transporter. Under stressful conditions that activate the unfolded protein response, synthesis of lumenal uridine diphosphatase increases to accommodate increased transport of UDP-Glc needed for reglucosylation of misfolded glycoproteins. The existence of UDP-GlcNAc, UDP-GalNAc, and UDP-Xyl transporters in the ER may mean that some reactions thought to occur exclusively in the Golgi may also occur in the ER. A good example is the synthesis of O-fucosylated proteins such as Notch in the ER versus fucosylation of N- and O-linked chains in the Golgi (see Chapter 12). Other, as-yet-undiscovered, glycosylation reactions may also occur in the ER.

For most glycosylation reactions, the nucleotide sugar donates the sugar, resulting in the formation of nucleoside diphosphate, which must be converted into a monophosphate by the nucleoside diphosphatase in the Golgi lumen. Exchange through the antiporters is electroneutral, because the nucleotide sugar with two negative charges (one on each phosphate) enters the Golgi and a nucleotide with a doubly charged single phosphomonoester exits. The antiporter system has the advantage of coupling the rate of nucleotide sugar use with its import. However, the Golgi can accumulate a pool of unused nucleotide sugars for glycosylation. Isolated Golgi preparations are capable of glycosylating partially completed endogenous glycoproteins as well as freely diffusible glycoside acceptors. Another advantage of using the antiporter system is that the nucleotide monophosphate is returned to the cytosol where it will be available for another round of activation. This creates a highly efficient recycling system for the precursors.

Several mutant mammalian cell lines lack specific nucleotide sugar transporters (e.g., UDP-Gal and CMP-Sia), and as a result, they make incomplete sugar chains (see Chapter 46). However, there is some “leakiness” in such mutants. For instance, loss of the UDP-Gal in the Golgi of mutant MDCK cells decreases the synthesis of keratan sulfate and galactosylated glycoproteins and glycolipids, but leaves heparan and chondroitin sulfate unaffected. This is probably because the galactosyltransferases that synthesize the core region tetrasaccharide common to GAG chains have lower K_m values for their donors (see Chapter 16).

Many putative transporters were identified by homology in the genomes of mammals, D. melanogaster, C. elegans, plants, and yeast. Like the GLUT transporters discussed above, all are multi-membrane-spanning (type III) proteins, but the level of amino acid identity does not give any clue to the substrate specificity. The UDP-GlcNAc transporters from mammalian cells and yeast are 22% identical, whereas mammalian CMP-Sia, UDP-Gal, and UDP-GlcNAc transporters have 40–50% identity. Clever domain-swapping experiments show that distinct regions are responsible for functional transport, and engineered chimeric transporters can carry both CMP-Sia and UDP-Gal.

Heterologous expression or rescue of transporter-deficient cell lines can be used to analyze the function of the putative transporters. For example, expressing the C. elegans gene SQV-7 in yeast showed that this one protein transports UDP-GlcA, UDP-GalNAc, and UDP-Gal, whereas mutant alleles cannot transport any of these donors. The human gene SLC35B4 encodes a bifunctional transporter that recognizes UDP-Xyl and UDP-GlcNAc. Also, the GDP-Man transporter of Leishmania can also transport GDP-Fuc and GDP-arabinose. Even though multisubstrate transport is somewhat unusual, it illustrates that functional, biochemical analysis is essential; homology is an insufficient criterion to infer functional specificity. Moreover, not all of the potential transport-like genes have been assigned a specific substrate.

Theoretically, glycosylation may be controlled in part by regulating availability of nucleotide sugars within the Golgi, presumably by regulating the transporters. The subcompartmental location (cis, medial, trans) of the transporters in the Golgi is not known nor are their physical relationships to the various glycosyltransferases they service. Clearly, a functional Golgi compartment requires both the nucleotide sugar donor and the acceptor with a colocalized transferase. There have been few studies on how the actual glycosylation reactions occur within the Golgi. Is it more like solution chemistry or like solid-state transfers? Are there really “soluble pools” of nucleotide sugars? Dramatic time-lapse videos of green fluorescent protein (GFP)-tagged glycosyltransferases show that the proteins are highly mobile within the Golgi, but there is also physical evidence for multiglycosyltransferase complexes involved in the biosynthesis of N-linked glycans, glycosphingolipids, and heparan sulfate. Many transporters appear to function as homodimers, and the GDP-Man transporter in Saccharomyces cerevisiae (VRG4) oligomerizes in the ER and appears to be transported to the Golgi by an active process. Also, synthesis of galactosylceramide occurs in the ER and a portion of the UDP-Gal transporter binds specifically to galactosylceramide transferase and is retained in the ER to provide donor substrate (see Chapter 10).

CONTROL OF NUCLEOTIDE SUGAR LEVELS

How nucleotide sugar levels are controlled and maintained has not been studied extensively, but this topic is likely very important. Table 4.4 shows several key biosynthetic enzymes that are inhibited by their final products in vitro. A human genetic disorder called sialuria validates the importance of this feedback process. In this condition, massive amounts of sialic acid (5–7 gm/day) are secreted into the urine along with various intermediates in the CMP-Sia biosynthetic pathway. This was found to be due to a defective feedback inhibition of the N-acetylglucosamine-2-epimerase/N-acetylmannosamine kinase, the first step in the pathway (see Chapter 42).

TABLE 4.4

Some control points for nucleotide sugar synthesis

Most of the precursor pools turn over within a matter of a few minutes, and the size of various nucleotide sugar pools have been determined by methods with different reliability. But even with the best methods, it is hard to interpret the measured numbers and translate them into a clear picture, because the relative distribution of the precursors in the cytosol and Golgi is not known. The average cellular concentrations of nucleotide sugar precursors may not be very meaningful, because “cytosol” is an operational definition (100,000g supernatant) that may not reflect compartments of cytoplasmic organization.

In whole-animal studies, the GDP-Fuc pool and fucosylated glycans in the intestine can be regulated by the diet and time of weaning. Considering that resident bacteria in the small intestine participate in the induction of fucosylation pathways in the enterocytes, dietary manipulation of glycosylation introduces another level of complexity that has barely been explored. The relationship of amino acid and nucleotide metabolism to nucleotide sugar metabolism is also potentially important, but remains largely unexplored (see Chapter 34).

DONORS FOR GLYCAN MODIFICATION

Glycans can be modified, imparting additional complexity and biological information. Sulfation, phosphorylation, methylation, pyruvylation, acetylation, and acylation have been found and their donors are listed in Table 4.5. In “lower” eukaryotes and bacteria, pyruvic acid is often found as a 1-carboxyethylidene bridge between two hydroxyl groups on a sugar such as galactose. Because all of these reactions occur in the Golgi, clearly there must be carriers or transporters that deliver and orient activated donors for efficient synthesis. As additional modifications of sugar chains made in the ER-Golgi pathway are uncovered, they will likely require specific transporters to carry the activated donors into the lumen of these compartments.

TABLE 4.5

Donors for glycan modifications

SYNTHESIS OF CARRIER LIPIDS

Multiple glycosylation pathways in prokaryotes and eukaryotes require lipid carriers to present monosaccharides and oligosaccharides at the proper location. Undecaprenyl-P (bactoprenol) is the glycosyl carrier for O-antigen, peptidoglycan, capsular polysaccharides, teichoic acid, and mannans in bacteria (see Chapter 20). Dolichol-P serves the same function in eukaryotic cells (see Chapter 8). Dolichol-P-mannose provides all of the mannose for glycophospholipid anchors, C-mannosylated proteins, O-mannose-based chains, and four of the mannose residues of the precursor oligosaccharide used for N-glycan biosynthesis. Dolichol-P-glucose provides glucose for the mature N-linked glycan precursor Glc₃Man₉GlcNAc₂, which itself is built on dolichol pyrophosphate (dolichol-PP).

The formation of dolichol-P involves elongation of farnesyl pyrophosphate with multiple cis-isopentenyl pyrophosphate units. The total number of isoprene units can vary from typically 11 in bacteria (making a C₅₅ bactoprenol chain) up to 21 in mammals. In eukaryotes, the double bond nearest the pyrophosphate must be reduced for the carrier to be functional, but it is unclear whether the phosphates are removed before or after this step. The evolutionary significance of the different chain lengths and reduction of the double bond is not known. Dolichol is phosphorylated by an ATP-dependent dolichol kinase to generate dolichol-P as needed. Because dolichol, dolichol-P, and dolichol-PP are all generated from a common metabolically stable pool, they must be recycled and interconverted as needed. Dolichol occurs in the ER and Golgi and turns over very slowly.

FURTHER READING

1. Neufeld EF, Ginsburg V. Carbohydrate metabolism. Annu. Rev. Biochem. 1965;34:97–312. [PubMed: 14321172]
2. Krieger M, Reddy P, Kozarsky K, Kingsley D, Hobbie L, Penman M. Analysis of the synthesis, intracellular sorting, and function of glycoproteins using a mammalian cell mutant with reversible glycosylation defects. Methods Cell Biol. 1989;32:57–84. [PubMed: 2691861]
3. Lloyd JB. Metabolite efflux and influx across the lysosome membrane. Subcell. Biochem. 1996;27:361–386. [PubMed: 8993166]
4. Hirschberg CB, Robbins PW, Abeijon C. Transporters of nucleotide sugars, ATP, and nucleotide sulfate in the endoplasmic reticulum and Golgi apparatus. Annu. Rev. Biochem. 1998;67:49–69. [PubMed: 9759482]
5. Varki A. Factors controlling the glycosylation potential of the Golgi apparatus. Trends Cell Biol. 1998;8:34–40. [PubMed: 9695806]
6. Schenk B, Fernandez F, Waechter CJ. The ins(ide) and out(side) of dolichyl phosphate biosynthesis and recycling in the endoplasmic reticulum. Glycobiology. 2001;11:61R–70R. [PubMed: 11425794]
7. He XM, Liu HW. Formation of unusual sugars: Mechanistic studies and biosynthetic applications. Annu. Rev. Biochem. 2002;71:701–754. [PubMed: 12045109]
8. Ishida N, Kawakita M. Molecular physiology and pathology of the nucleotide sugar transporter family (SLC35). Pflugers Arch. 2004;447:768–775. [PubMed: 12759756]
9. Tettamanti G. Ganglioside/glycosphingolipid turnover: New concepts. Glycoconj. J. 2004;20:301–317. [PubMed: 15229395]
10. Uldry M, Thorens B. The SLC2 family of facilitated hexose and polyol transporters. Pflugers Arch. 2004;447:480–489. [PubMed: 12750891]
11. Eklund E, Bode L, Freeze HH. Comprehensive Glycoscience. Vol. 4. Elsevier; London: 2007. Diseases associated with carbohydrates/glycoconjugates; pp. 339–372. Chap. 4.19.