Complex N-linked Chains (1,3–5)

Much of what we know of this pathway comes from analysis of products that accumulate in patients missing one of the degradative enzymes (Table 18.1) or from structural analysis of monosaccharide-labeled glycoproteins during degradation in perfused rat liver. By carrying out the latter studies in the presence of inhibitors of different lysosomal enzymes, a picture emerges of simultaneous and independent bidirectional degradation of the protein and the carbohydrate chains (Figure 18.1). The relative degradation rates vary depending on structural and steric factors of the protein and the sugar chains. The accumulation of GlcNAcβ1–4GlcNAcβ-Asn in cells that cannot cleave the GlcNAcβ-Asn linkage clearly shows that degradation of the sugar chain does not require prior cleavage from the protein. Much of the protein is probably degraded before N-glycan catabolism begins. Removal of the core fucose (Fucα1–6GlcNAc) and probably any peripheral fucose residues linked to the outer branches of the chain (Fucα1–3GlcNAc) appears to be the first step in degradation, since patients lacking this enzyme still have intact N-linked chains bound to asparagine. Glycosylasparaginase (aspartyl-N-acetyl-β-d-glucosaminidase) then cleaves the GlcNAcβ-Asn bond, provided the α-amino group is not in peptide linkage. In rodents and primates, chitobiase (an endo-β-N-acetylglucosaminidase) removes the reducing GlcNAc, leaving the oligosaccharide with only one terminal GlcNAc. In many other species, splitting of the chitobiose linkage (GlcNAcβ1–4GlcNAc) uses the β-N-acetylhexosaminidase mentioned below as the last step in degradation. Because either pathway appears to get the job done, the presence of this endo-β-N-acetylglucosaminidase (also called chitobiase) in some species is unexplained. The oligosaccharide chain is then sequentially degraded by sialidases and/or α-galactosidase and then β-galactosidase, β-N-acetylhexosaminidase, and α-mannosidases. The remaining Manβ1–4GlcNAc is split by mannosidase to mannose and GlcNAc or, in those species that do not have chitobiase, to chitobiose, which is then degraded by β-N-acetylhexosaminidase.

Figure 18.1

Degradation of complex N-linked oligosaccharide chains in primates and rodents. The lysosomal degradation pathway of glycoproteins carrying complex-type oligosaccharides proceeds simultaneously on the protein and oligosaccharide moieties. The oligosaccharides (more...)

Lysosomal sialidase (neuraminidase), β-galactosidase, and a serine carboxypeptidase called protective protein/cathepsin A form a complex in the lysosome that is required for efficient degradation of sialylated glycoconjugates. Cathepsin A protects β-galactosidase from rapid degradation and also activates the sialidase precursor. Mutations in this protective protein lead to galactosialidosis in which the simultaneous deficiencies in β-galactosidase and α-sialidase are secondary effects.

Other oligosaccharides having GalNAcβ1–4GlcNAc or GlcAβ1–3Gal or GalαGal on the outer branches must first have these residues removed by β-N-acetyl-hexosaminidase, β-glucuronidase, and α-galactosidase prior to any further digestion of the underlying oligosaccharide chain.

High-mannose-type Oligosaccharides (6–7)

High-mannose-type chains are hydrolyzed in the lysosome by an α-mannosidase to yield Manα1-6Manβ1–4GlcNAc, a common intermediate derived from hybrid and complex chains, in addition to the high-mannose type. A second α1,6-specific mannosidase can cleave this linkage in humans and rats, but only on molecules that have a single core region GlcNAc, i.e., those generated by chitobiase cleavage. Finally, β-mannosidase completes the job.

ER and Cytosolic Degradation of Oligosaccharides Derived from Dolichol Precursors and Misfolded Glycoproteins (6–8)

High-mannose-type oligosaccharides are also partially degraded in the cytosol by a different set of α-mannosidases with higher pH optima. These enzymes degrade free glycans liberated from the dolichol-linked precursors or from newly synthesized but misfolded glycoproteins that are en route to proteasome-mediated degradation. The degradation of the lipid-linked precursor begins with a pyrophosphatase cleave of the nonglucosylated cytosolic-facing precursor, primarily Man₅. Normally, the next step in the biosynthetic pathway involves flipping the lipid-linked oligosaccharide to the lumen of the ER for completion. Oligosaccharide precursors that fail to flip may be degraded in this way. Once the phosphorylated Man₅ chain is liberated from the lipid, a phosphatase and a cytosolic endo-β-N-acetylglucosaminidase remove the phosphate and reducing GlcNAc. Other dolichol-linked oligosaccharides may also be hydrolyzed within the lumen of the ER, and transported into the cytosol, but the details of this pathway and its relation to cleavage of oligosaccharides from proteins in the ER are less clear. Regardless of the mechanism, the cytosolic endo-β-N-acetylglucosaminidase mentioned above (a chitobiase distinct from the lysosomal one) clips a single GlcNAc. Several mannose residues are also removed, producing the same Man₅ structure mentioned above. This molecule is then imported into the lysosome for its final degradation. Up to one third of the total lipid-linked oligosaccharides can be recovered as free glycans. These cleavages may be important for coordinating the level of oligosaccharide intermediates with protein synthetic rates.

Poorly folded glycoproteins were previously thought to be degraded in the ER, but more recently, it has become clear that this degradation occurs in the cytosol as an ATP-dependent process. The emerging view is that misfolded glycoproteins are reverse-translocated into the cytoplasm through the Sec61 translocation channel in the ER. The release of the polypeptide into the cytoplasm depends on ubiquitination, and final degradation is proteasome-dependent. Depending on the protein, the N-glycan may require a small amount of mannose trimming prior to ubiquitination, but complete degradation of the protein requires removal of all N-glycans using a cytosolic peptide-N-glycoamidase (not endoglycosidase) activity. These liberated oligosaccharides can then be cleaved by the cytosolic endo-β-N-acetylglucosaminidase and α-mannosidases similar to the degradation of the chains liberated from lipid-linked oligosaccharide.

Some evidence suggests that the removal of glycan chains from proteins might be physiologically important. Proximal peptide-N-glycoamidases can release the entire oligosaccharide chain, and in doing so, they convert an asparagine to an aspartic acid residue. Amino acid sequencing of some proteins shows aspartic acid where the DNA sequence predicts asparagine. The liberated glycans cannot be simply added to other glycoproteins in the ER, since addition requires the pyrophosphate-linked dolichol donors and not just the sugar chain itself. The liberated chains may need to be exported to prevent them from competing with ER lectins such as calnexin, calreticulin, or ERGIC53, a mannose-binding lectin in the ER.

O-Linked Oligosaccharides (9)

Degradation of typical α-GalNAc-initiated O-glycans has not been systematically studied. Since so many of the structures seen on N-linked chains are also found on O-glycans, they are probably degraded by the same group of exoglycosidases discussed above. The exception to this of course is the linkage region, GalNAcα-O-Ser/Thr. Patients with Schindler disease lack an α-N-acetylgalactosaminidase. This enzyme is specific for α-GalNAc and will not cleave α-GlcNAc. Patients lacking α-N-acetylgalactosaminidase accumulate GalNAc-containing glycopeptides in their urine, but curiously they also accumulate more complex, extended glycopeptides containing GlcNAc, galactose, and sialic acid. The structures are the same as those found on some native glycoconjugates. Their production may result from a general slowdown in degradation of the oligosaccharides, but this seems unlikely. Alternatively, the terminal GalNAcα-O-Ser/Thr glycopeptide could be generated as a normal degradation product using normal glycosidases and then rebuilt once again. How would this happen? There are several possibilities. One is that some of the partially graded glycopeptides enter a compartment (Golgi?) that contains the appropriate glycosyl transferases and sugar nucleotides. Monosaccharides are added sequentially before the products exit the cell. In this way, the glycopeptides would be behaving like oligosaccharide primers (see Chapter 40). However, another possibility is that the concentration of GalNAcα-O-Ser/Thr is high enough in the lysosome that lysosomal glycosidases use them as acceptors in a series of transglycosylation reactions. α-N-acetylgalactosaminidase has also been found to be part of the sialidase/β-galactosidase/cathepsinA complex mentioned above. If lysosomal glycosidases form complexes to degrade glycoconjugates more efficiently, partially degraded substrates may be in a preferred location to be acceptors in transglycosylation reactions. The same α-GalNAcase probably removes terminal α-GalNAc from blood-group-A-containing glycans (GalNAcα1,3Gal) and some glycolipids such as the Forsmann antigen, GalNAcα1,3GalNAcβ1,3Galα1,4Galβ1,4Glcβ-Cer.

Hyaluronan

The most abundant GAG is HA, which is degraded in the lysosome by hyaluronidase (see Figure 18.2). The products are tetrasaccharides and larger fragments. Because HA is not synthesized on a protein, proteolysis is not required, and it is not sulfated; therefore, only β-glucuronidase and β-N-acetylhexosaminidase are needed for degradation, which occurs sequentially from the nonreducing end. Pig liver hyaluronidase has a sequence identical to that of the heme-binding serum protein hemopexin. Although serum hemopexin binds to HA, it does not appear to have hyaluronidase activity. Expression of the pig hyaluronidase enzyme in baculovirus produces a protein with hyaluronidase activity, showing that this primary sequence has the information needed for enzymatic activity. The relationship between lysosomal hyaluronidase and hemopexin is still unclear.

Figure 18.2

Degradation of hyaluronan. Hyaluronidase, an endoglycosidase, cleaves large chains into smaller fragments, each of which is then sequentially degraded from the nonreducing end.

Heparan Sulfate (14–15)

HS chains are first degraded by an endoglucuronidase followed by well-ordered sequential degradation (see Figure 18.3). A terminal IdoA-2-sulfate must be desulfated by a specific IdoA-2-sulfatase to make it a suitable substrate for α-iduronidase. If a GlcA-2-sulfate were at this position, a GlcA-2-sulfatase would first remove the sulfate, followed by β-glucuronidase cleavage. The terminal GlcNSO₄ is the next sugar for cleavage, but this requires two steps. The sulfate is removed by N-sulfatase forming glucosamine, but this residue cannot be cleaved by α-N-acetylglucoaminidase. The amino group must be N-acetylated by an N-acetyl transferase embedded in the lysosomal membrane. In the first step, acetyl-CoA donates the acetyl group to a cytosolic-facing histidine residue at neutral pH. The acetyl group then becomes available on the luminal side of the lysosomal membrane and transferred at low pH to the amino group of glucosamine on the partially degraded HS chain. The terminal α-GlcNAc can now be cleaved. If the glucuronate residue is 2-sulfated, the sulfate is first removed, followed by β-glucuronidase cleavage. If the next α-GlcNAc is 6-O-sulfated, the sulfate is removed by a specific GlcNAc-6-sulfatase. Figure 18.3 and Table 18.2 provide a listing of the enzymatic defects and the disorders they cause.

Figure 18.3

Degradation of heparan sulfate. An endoglucuronidase first cleaves large chains into smaller fragments, and then each monosaccharide is removed from the nonreducing end as described in the text. N- and O-sulfate esters must first be removed before the (more...)

Dermatan Sulfate and Chondroitin Sulfate

A combination of endoglycosidases, sulfatases, and exoglycosidases degrade DS in the lysosome (Figure 18.4). Iduronate-2-sulfatase is followed by α-iduronidase. The terminal GalNAc-4-SO₄ can be removed by either of two pathways. In the first pathway, a GalNAc-4-sulfate sulfatase acts, followed by β-N-acetylhexosaminidase A or B to remove GalNAc. In the second, β-N-acetylhexosaminidase A removes the entire GalNAc-4-SO₄ unit followed by sulfatase cleavage. Absence of this sulfatase (MPS VI) is unique to the DS pathway. β-glucuronidase cleaves the β-GlcA residue and the process is repeated on the rest of the molecule.

Figure 18.4

Degradation of dermatan/chondroitin sulfate. Sequential degradation can proceed by two different routes. One route uses GalNAc-4-SO₄ase followed by cleavage with β-N-acetylhexosaminidase A or B. The other route uses only β-N-acetylhexosaminidase (more...)

To degrade CS, GalNAc-6-SO₄ sulfatase and GalNAc-4-SO₄ sulfatase work in combination with β-N-acetylhexosaminidase A or B and β-glucuronidase. Hyaluronidase may also degrade CS, but no CS-specific endoglycosidases have been found.

Keratan Sulfate

KS is a heavily sulfated polylactosamine chain. Although some organisms have an endo-β-galactosidase that could cleave similar chains, mammalian cells do not have an endoglycosidase to break KS down. Sequential action of sulfatases and exoglycosidases is needed. Galactose-6-SO₄ sulfatase is the same enzyme that desulfates GalNAc-6-sulfatase in CS degradation. This hydrolysis is followed by β-galactosidase digestion, leaving a terminal GlcNAc-6-SO₄. Desulfation by the sulfatase followed by cleavage with β-N-acetylhexosaminidase A or B eliminates GlcNAc-6-SO₄. Alternatively, β-N-acetylhexosaminidase A can directly release GlcNAc-6-SO₄ followed by desulfation of the monosaccharide.

Degradation of the Linkage Regions of Proteoglycans

Degradation of the core region O-linked KS (skeletal type; type II) probably occurs by the same route as the other typical O-linked chains (Figure 18.5). The familiar N-linked corneal-type KS (type I) and the more recently discovered family of N-linked GAG chains are probably also degraded by the same set of enzymes used for N-linked degradation. The more typical xylose-linked GAG chains (DS, CS, HS) all share a common core tetrasaccharide. The finding of GlcAβ1,3Galβ1,3Galβ1,4Xylβ-O-Ser and xylose- and galactose-terminated core fragments suggests that endo-β-xylosidase and endo-β-galactosidase exist. An endo-β-xylosidase has been detected in the rabbit liver. It is possible that similar enzymes cleave different types of GAGs, but the details of how the linkage regions of GAGs are degraded remain to be established.

Figure 18.5

Stepwise lysosomal degradation of keratan sulfate. Sequential degradation of KS occurs from the nonreducing end, and like the degradation of DS and CS, the terminal GlcNAc-6-SO₄ can be cleaved sequentially by a sulfatase and then by β-N-acetyl-hexosaminidase (more...)

Specialized Enzymes and Proteins for Glycosphingolipid Degradation

Some exoglycosidases for glycoprotein and GAG chain degradation also degrade glycolipids, but others are unique to glycolipid degradation. The absence of these unique enzymes causes glycolipid storage diseases listed in Table 18.3. Glucocerebrosidase, also called β-glucoceramidase, is specific for the degradation of the Glcβ-Cer bond, and its loss causes Gaucher's disease.

A specialized β-galactosidase, called β-galactoceramidase, hydrolyzes the bond between galactose and ceramide, and it can also cleave the terminal galactose from lactosyl ceramide. Its loss produces Krabbe disease. Galactosyl-ceramide is often found with a 3-sulfate ester (sulfatide), and thus a specific sulfatase, arylsulfatase A, is needed for its removal prior to β-galactosylceramidase action. Loss of this sulfatase causes metachromatic leukodystrophy and the accumulation of sulfatide. Glycolipids terminated with α-galactosamine residues are degraded by a specific α-galactosidase and its loss causes Fabry disease.

Activator Proteins

Two genes code for all of the known sphingolipid activator proteins. These proteins can also be used to form complexes with more than one degradative enzyme for more efficient hydrolysis.

G_M2 Activator

G_M2 activator protein forms a complex with either G_M2 or G_A2 and presents them to β-N-acetylhexosaminidase A for cleavage of terminal β-GalNAc. Sugars that lie too close to the lipid bilayer apparently have limited access to the soluble hexosaminidase. The activator protein binds a molecule of glycolipid, forming a soluble complex that can be cleaved by the hexosaminidase. The resulting product is now inserted back into the membrane, and the activator presents the next G_M2 molecule, and so on. Genetic loss of this activator protein causes the accumulation of G_M2 and G_A2, resulting in the AB variant of G_M2 gangliosidosis. The activator is probably targeted to the lysosomes via the Man-6-P pathway.

Saposins

Saposins are derived from a 524-amino-acid precursor called prosaposin which is processed into four homologous activator proteins, Sap-A, Sap-B, Sap-C, and Sap-D, each having about 80 amino acids and each having different properties. Sap-A and Sap-C both help β-glucosyl and β-galactosyl ceramidase degradation. Sap-B assists arylsulfatase A, α-galactosidase, α-sialidase, and β-galactosidase. The mechanism is thought to be similar to that of the G_M2 activator by partially lifting a glycolipid molecule out of the membrane to form a complex that is cleaved by the soluble enzyme, followed by reinsertion of the substrate back into the membrane. Sap-D and Sap-B also assist in sphingomyelin degradation by sphingomyelinase. Complete absence of prosaposin is lethal in humans, and deficiencies in Sap-B and Sap-C lead to defects that clinically resemble aryl sulfatase A deficiency (metachromatic leukodystrophy) and Gaucher's disease. These symptoms might be predicted based on the enzyme that these saposins assist.

A very rare human disorder is the multiple sulfatase deficiency. All sulfatases are casualties of this defect, since they all require a posttranslational conversion of an active site cysteine residue to a 2-amino-3-oxopropionic acid for activity. In essence, the S group of cysteine is replaced by a double-bonded oxygen atom that probably acts as an acceptor for cleaved sulfate groups. Loss of the enzyme carrying out that reaction leads to inactive sulfatases.

General Inhibitors

Since nearly all lysosomal enzymes have acid pH optima, degradation can be severely curtailed by increasing the intralysosomal pH from 5–5.5 to 7, using agents such as ammonium chloride and chloroquine. These agents lower the rate of degradation, but they may not completely block it. Some lysosomal enzymes have relatively broad pH optima, and some of these enzymes actually have pH optima higher than the lysosome. It has been suggested that lysosomes may not always remain at low pH. Some “lysosomal” enzymes are actually present and active in early and late endosomes that have a higher pH than lysosomes.

Specific Inhibitors

Glycosylation inhibitors are covered in Chapter 40; however, some of these also inhibit specific lysosomal enzymes. For example, swainsonine blocks lysosomal α-mannosidase in addition to blocking the α-mannosidase II involved in glycoprotein processing. Sheep and cattle become neurologically deranged by eating food rich in swainsonine, which is also called locoweed. Since there are no neurological effects seen in transgenic knockout mice lacking the processing α-mannosidase II, it is likely that these temporary symptoms are induced because lysosomal α-mannosidase is inhibited. Undegraded oligosaccharides probably accumulate in affected animals. Although many inhibitors affect various enzymes when tested in in vitro enzymatic assays, they may not be effective in cells or whole animals because they may not enter the lysosome, or if they do, their concentration may not be sufficient.

Perspective on Synthesis and Degradation and Genetic Disorders

Genetic defects in lysosomal enzymes have provided the best insight into the catabolic pathways and their importance. For example, cocultivation of fibroblasts from patients with different storage diseases leads to the disappearance of stored material in both types of cells. Each supplies the corrective factors (lysosomal enzyme) that the other lacks through uptake via the Man-6-P receptors on the cell surface (for further details, see Chapter 23 on the discovery of Man-6-P targeting of lysosomal enzymes). Cross-correction highlights the fact that only a relatively small amount of enzyme may be needed to prevent accumulation of storage products and the complications they cause. Some estimates suggest that only 10% of normal β-N-acetylhexoaminidase activity may be sufficient to prevent pathological symptoms of Tay-Sachs disease. Since accumulation of stored material depends on the relative rates of glycoconjugate synthesis and degradation, the near absence of a glycosidase can be ameliorated by reducing the synthetic rate of the stored compound. This was cleverly tested in a mouse model of Tay-Sachs disease by using N-butyldeoxynojirimycin, an inhibitor of glucosylceramide synthase, the first step in glycosphingolipid biosynthesis. When normal mice were treated with this compound, the amount of glycosphingolipids fell 50–70% in all tissues without obvious pathological effects. When this compound was given to Tay-Sachs mice, the accumulation of G_M2 in the brain was blocked and the amount of stored ganglioside was reduced. Thus, reducing the synthesis of the primary precursor reduced the load of G_M2 that could be degraded by the mice.

As is discussed in Chapter 6, it is assumed that most of the monosaccharides derived from glycoconjugate breakdown are reutilized. The few studies exploring this issue suggest that reutilization of glycosidase-liberated sugars may be quite substantial.

Glycoprotein Reglycosylation

A variety of different experiments clearly show that the half-life of specific membrane proteins is longer than the half-life of their glycans. The more terminal monosaccharides turn over faster than those closer to the reducing end of the chain, suggesting that the monosaccharides are sequentially removed by exoglycosidases. This may occur at the cell surface or when proteins are endocytosed in the course of normal membrane recycling. Since the mildly acidic late endosomes contain lysosomal enzymes, whose pH optima can be quite broad, terminal sugars such as sialic acid can be cleaved. If the endocytosed proteins are not degraded in lysosomes, they may encounter sialyltransferases in the Golgi, become resialylated, and reappear again on the cell surface. A similar situation may occur if a glycoprotein has lost both sialic acid and galactose residues from its glycans. Some membrane proteins synthesized in the presence of oligosaccharide processing inhibitors can still reach the cell surface in an unprocessed form. Subsequent incubation in the absence of inhibitors leads to normal processing over time. The extent of processing and the kinetics depend on the protein and the cell line. Since Golgi enzymes are not distributed identically in all cells, it is difficult to make a general statement about how much “reprocessing” can occur. Most studies have monitored N-linked glycans, but membrane proteins with O-glycans show a similar response. Newly arrived plasma membrane proteins will acquire additional O-glycans at previously unoccupied glycosylation sites over time. Initiation of O-glycans is thought to occur in the cis-Golgi, but it is difficult to determine whether there is a small amount of O-GalNAc transferase that is active in a late (trans) Golgi compartment or whether the acceptor protein briefly encounters a high concentration of the glycosyltransferase in an early compartment.

Glycosphingolipid Recycling

The majority of the newly synthesized glucosyl ceramide arrives at the cell surface by a Golgi-independent cytosolic pathway. Some of this material, as well as Glc-Cer generated by partial degradation of more complex glycolipids, can recycle to the Golgi. There, like glycoproteins, they can serve as acceptors for synthesis of longer chains. In the lysosome, sphingosine and sphinganine are produced by complete degradation of glycosphingolipids, and reutilized. Together, these pathways, especially the latter one, may actually account for the majority of complex glycolipid synthesis in many cells. Depending on the physiological state of the cell and synthetic demands, the de novo pathway starting from serine and palmitoyl CoA may account for only 20–30% of the total synthesis. Shuttling the recycled components through and among the various organelles appears to involve vimentin intermediate filaments. Mechanisms and details of this process are presently lacking, but these studies offer a physiological function for glycolipid transfer proteins that were purified from brain preparations many years ago. For details of the biosynthetic pathway, see discussion in Chapter 9.