(Figure 1, Click figure to enlarge)Background, Name, History, and Biological Aspects:
Role for
Asn-linked oligosaccharides and the Asn-linked biosynthetic pathway:
Protein-linked
carbohydrate structures are among the most complex and diverse set of posttranslational
modifications on intracellular and secreted proteins (8). Large numbers of proteins
contain covalently-bound Asn-linked oligosaccharides, including enzymes, cell
surface receptors, secreted proteins, hormones, immunoglobulins, and viral antigens.
Although there are many examples of the contribution of N-glycans to the bioactivity,
folding, localization, and immunogenicity of the attached polypeptide, the functional
roles of individual oligosaccharide structures on a given glycoprotein are difficult
to predict (9). At the cellular level, Asn- and Ser/Thr-linked glycan structures
have been shown to contribute to several aspects of biological recognition,
including cell adhesion events during development, immune surveillance, inflammatory
reactions, hormone action, viral infection, arthritis, and metastasis of oncogenically-transformed
cells (8).
ER a-mannosidase
I (also known as ER class 1 a1,2-mannosidase and
Man9GlcNAc2-specific processing a-mannosidase,
EC 3.2.1.113) is a key enzyme in the maturation of N-linked oligosaccharides
in mammalian cells (2). The synthesis of Asn-linked oligosaccharides begins
with the transfer of a preformed oligosaccharide precursor, Glc3Man9GlcNAc2,
to the nascent polypeptide chain. a-Glucosidases
and a-mannosidases in the ER trim this oligosaccharide
precursor to primarily Man8GlcNAc2 while subsequent transport
of the glycoprotein into the Golgi apparatus allows further trimming of the
remaining a1,2-linked mannose residues. The resulting
Man5GlcNAc2 oligosaccharide is the substrate for GlcNAc
transferase I, the first enzyme in a cascade that ultimately results in the
formation of complex and hybrid oligosaccharide structures (1)
(Figure 1, Click figure to enlarge)
Asn-linked
glycoproteins in folding and quality control:
A role for glucosylated oligosaccharides containing the Glc1Man9GlcNAc2
structure has been proposed as a part of the quality control process of glycoprotein
folding in the ER. Proteins that are glycosylated as they are extruded into
the lumen of the ER undergo a concerted folding process that is facilitated
by the interaction with a variety of molecular chaperones including BiP, GRP94,
calnexin, and calreticulin (10, 11). The latter two chaperonins have been shown
to contain a lectin activity and initiate their interaction with unfolded glycoproteins
by specifically binding to Glc1Man9GlcNAc2
oligosaccharides (12). Subsequent interactions between the chaperonin and the
polypeptide have been proposed to promote polypeptide folding. Correctly folded
glycoproteins are released from calnexin or calreticulin through the action
of glucosidase I, which removes the last glucose residue, and the glycoconjugate
will then proceed through the secretory pathway to the Golgi complex and other
parts of the cell.
(Figure
2, Click figure to enlarge)
Proteins that have not folded correctly will remain transiently associated with the chaperonins until glucosidase I removes the last glucose residue. Unfolded glycoproteins in the ER containing Man9GlcNAc2 structures can then be acted upon by UDP-Glc:glycoprotein glucosyltransferase to replace the Glc residue to the same position and linkage where it was removed by glucosidase I, thus allowing the protein to re-bind to the chaperonins (13). Since the glucosyltransferase recognizes only unfolded glycoprotein intermediates, it has been proposed that the enzyme acts as the "sensor" of the folding status of the newly synthesized glycoproteins. The cycle of binding to the chaperonins, sugar cleavage by glucosidase I, release from the chaperonins, sugar addition by the glucosyltransferase, and re-binding to the chaperonins can proceed through multiple rounds in an effort to effectively fold the newly synthesized glycoproteins. Under normal conditions glycoproteins that fail to fold after a defined period of time are targeted for translocation into the cytosol via the Sec61p translocon complex followed by degradation by the cytosolic proteosome in a process termed ER-associated degradation (ERAD).
While many of the components involved in glycoprotein folding have been identified, the components that recognize glycoproteins which have failed the folding process and target them for degradation are not well understood. It was originally proposed that mannose trimming could act as a timing step for the quality control cycle by interrupting one or more steps in the glucosylation/de-glucosylation cycle and targeting glycoproteins with trimmed oligosaccharides for degradation (14). Several lines of evidence now indicate that mannose trimming, especially to a specific Man8GlcNAc2 structure (isomer B), plays an important role in targeting unfolded proteins for degradation. The role of a mannose cleavage step in targeting the degradation of a mutant form of carboxypeptidase Y in S. cerevisiae has been shown by blocking the formation of the Man8GlcNAc2 isomer B structure. When glucose or mannose trimming was abrogated through gene disruptions in the processing a-glucosidases or ER mannosidase I, or when the synthesis of the lipid-linked precursor was blocked by gene disruption of the oligosaccharide biosynthetic enzymes, misfolded carboxypeptidase Y accumulated in the ER (15, 16). In mammalian cells, the degradation of a misfolded variant form of a1-antitrypsin or the T cell receptor subunit CD3-d was blocked by treatment of the cells with dMNJ, a potent inhibitor of ER mannosidases (17), indicating the involvement of ER mannosidases. Treatment of cells with kifunensine, a selective inhibitor of ER mannosidase I but not ER mannosidase II, also blocked the degradation of a1-antitrypsin (18) implicating a requirement for the formation of the Man8GlcNAc2 isomer B structure by ER mannosidase I in targeting glycoproteins for degradation.
Two models have been proposed for
the recognition and degradation of unfolded glycoproteins containing the Man8GlcNAc2
isomer B structure. In one model, the Man8GlcNAc2 isomer
B structure can be acted upon by the glucosyltransferase to make Glc1Man8GlcNAc2,
but the product may be cleaved less efficiently by glucosidase II. This would
lead to a prolonged interaction with calnexin/calreticulin and potential targeting
to the Sec61p translocon (4). An alternative model proposes that an, as of yet
unidentified, carbohydrate binding protein, or lectin, may recognize the Man8GlcNAc2
isomer B structure in the context of an unfolded protein to target the glycoconjugate
for translocation (15). In each model discrimination between folded and unfolded
substrates must be made and a low efficiency of recognition of the oligosaccharide
on unfolded proteins would be predicted since transiently unfolded nascent polypeptides
and correctly folded glycoproteins should not be targets for translocation and
degradation. Thus, the conservation of early processing events in oligosaccharide
maturation appears to contribute to the quality control recognition processes
in the ER, and these processes are independent of the roles that the oligosaccharides
may play in later compartments of the secretory pathway.
(Figure
3, Click figure to enlarge).
Diseases
related to ERAD:
Genetic alterations in protein coding regions can commonly cause
alterations in the corresponding translated sequence that can interfere with
the correct folding of the polypeptide. Incorrectly folded translation products
that are not functionally active are the most common cause of human genetic
disease, including inborn errors of metabolism. For mutant unfolded proteins
that enter the secretory pathway, ERAD appears to be a common mechanism for
their degradation. The table below contains a partial list of medically relevant
variant forms of glycoproteins that enter the ER and are degraded by ERAD. Several
members of the list are variant forms of proteins that are rapidly degraded
in the ER. In contrast, others are expressed as functional proteins, even though
they are retained within the ER by an association with chaperonins. Examples
of the latter group include the D-F506 variant form
of CFTR, a protein that normally resides at the cell surface to function as
a regulated chloride channel (19), and the Z variant of a1-antitrypsin,
a serpin that is normally secreted from the liver and circulates to the lung
where it protects the tissue from elastolytic attack (20). Recent data have
indicated that chemical chaperones may aid in inducing the secretion of some
proteins that accumulate in the ER and present a possibility for therapeutic
treatment of affected patients (21), however, this approach is unlikely to be
an effective long term treatment. In order to understand the potential for therapeutic
intervention a better understanding of the mechanisms involved in the recognition
and targeting of these mutant glycoproteins for degradation will clearly be
required.
| Table 1. Examples of ERAD substrates of medical relevance | ||
| ERAD substrate | Associated disease | Ref. |
| a1antitrypsin Z variant (AAT-Z) | Emphysema and liver disease | (22,23) |
| DF508 CFTR | Cystic fibrosis | (19,24) |
| Insulin receptor mutants | Type A insulin resistance | (25) |
| HMG co-A reductase mutants | Cancer, heart disease, cholesterolemia | (26) |
| LDL receptor mutants | Hypercholesteroemia | (27,28) |
| MPO Y173C | Myeloperoidase deficiency | (29) |
| Pro-PTHrP | Hypercalcemia associated with malignancy | (30) |
| Tyrosinase | Amelanotic melanomas | (31) |
| b-Hexosaminidase | Tay-Sachs disease | (32) |
| Pro-collagen IA1 | Osteogenesis imperfecta | (33) |
| a2-plasmin inhibitor | Severe hemorragic disease | (34) |
| Thyroglobulin | Congenital goiter with hypothyroidism | (35) |
| MHC class I | HCMV | (36) |
| CD4 | AIDS | (37) |
| Copper-transporting ATPase | Wilson disease | (38) |
Adapted from Brodsky and
McCracken (39)
Classification
of mannosidases: Class 1 vs. Class 2 mannosidases:
Mannosidases involved in glycoprotein maturation and catabolism
have been divided into three broad classes based on their distinctive substrate
specificities, responses to inhibitors, cation requirements, protein molecular
weights, subcellular localizations, and enzyme mechanisms (2). This classification,
based on biochemical characteristics, has been reinforced by the sequence comparison
of a growing number of family members (2). Two of the mannosidase classes, termed
Class 1 and Class 2 mannosidases, are exo-glycosidases, since they excise mannose
residues from the non-reducing terminus of the oligosaccharide. The Class 1
mannosidases are distinguished by several characteristics including: (a) similarities
in sequence including a conserved 440-510 amino acid catalytic domain, (b) specificity
for cleaving a1,2-mannose linkages, (c) a requirement
for Ca2+ for catalytic activity, (d) sensitivity
to inhibition by the pyranose substrate mimics deoxymannojirimycin (dMNJ) and
kifunensine (KIF) , and (e) cleavage of the glycosidic linkage by inversion
of configuration of the released mannose residue (2). This classification contrasts
them with the more heterogeneous collection of processing and catabolic mannosidases,
termed Class 2 mannosidases, that are present in the ER, Golgi, lysosomes, and
cytosol of mammalian cells, but have additional homologs in a widely diverse
array of organisms including eubacteria and archea. Class 2 mannosidases are
larger (110-135 kDa), do not require cations for catalytic activity, are sensitive
to inhibition by the furanose transition state analogs swainsonine (SW) and
1,4-dideoxy-1,4-imino-D-mannitol (DIM) , and cleave glycosidic linkages by retention
of anomeric configuration of the released monosaccharide (2). A web-based database
(URL: afmb.cnrs-mrs.fr/~pedro/CAZY/ghf.html)
has also collected data on the characteristics of classes of glycosylhydrolases
based primarily on sequence similarity and has also separated the two enzyme
classes into different families. (Class
1 = family 47; Class
2 = family 38).
(Figure
4, Click figure to enlarge)
Class 1
mannosidases: enzymes of the ER and Golgi:
Comparison of the sequences of Class 1 mannosidases makes it
possible to define sub-groups within the members of the family (see fig above).
When the catalytic core sequence of these proteins were aligned and used to
generate a dendrogram of the related sequences, the Class 1 proteins were found
to be partitioned into four discrete subfamilies, three of which contained mammalian
representatives . The first subfamily contains the collection of Golgi processing
a-mannosidases that cleave Man9-8GlcNAc2 to Man5GlcNAc2.
The second cluster of sequences contains ER mannosidase I from yeast, human,
C. elegans, and plant sources. A third cluster of two sequences is unique
to fungi, where they are believed to represent secreted enzymes of unknown function.
The fourth cluster of sequences is a collection of newly identified proteins
that are referred to HTM proteins, for "homologous to mannosidases."
These proteins are conserved from yeast to plants, insects, nematodes and mammals
and have several unique sequence characteristics that will not be discussed
here.
ER a-Mannosidase
I Characteristics:
The Class 1 mannosidases were first identified in Saccharomyces
cerevisiae and mammalian systems, but significant differences in substrate
specificities were identified between the enzymes from the two sources. The
yeast enzyme was shown to cleave only a single mannose residue from Man9GlcNAc2
to form the Man8GlcNAc2 isomer B structure, without any
further cleavage (40, 41). This organism does not process their oligosaccharides
beyond Man8GlcNAc2 before extension into mannan structures.
The enzyme from S. cerevisiae has been the most completely studied of
the Class 1 mannosidases, both as an enzyme purified from natural sources and
by characterization of the recombinant enzyme (40, 41). ER mannosidase I does
not appear to be essential for growth in S. cerevisiae or extension of
mannan structures, but evidence described above indicates that it does appear
to be involved in the timing of glycoprotein folding. Yeast ER mannosidase I
is an ER resident transmembrane protein containing a lumenally oriented catalytic
domain. Like the other Class 1 mannosidases the yeast enzyme can not cleave
aryl-a-mannosides, is inhibited by dMNJ and KIF, but not SW, and requires Ca+2
for catalytic activity. Mutagenesis studies have also identified several
acidic residues that are required for Ca+2 binding
and catalytic activity(42) . The S. cerevisiae enzyme has been overexpressed
in Pichia pastoris and the recombinant enzyme was the first of the Class
1 enzymes in which the mannosidase reaction was shown to proceed with an inversion
of the anomeric configuration of the released monosaccharide (43). The recombinant
yeast ER mannosidase I has been purified in milligram quantities, crystallized,
and was the first glycoprotein processing hydrolase to have its structure determined
(5).
The initial characterization of the early processing mannosidases in mammalian systems was far more complicated than in yeast as a result of the multiple enzyme activities present in different subcellular compartments in the secretory pathway. Early studies on the processing mannosidases took advantage of the ability to identify glycoprotein processing intermediates by pulse labeling cultured cells with radioactive monosaccharide precursors, either in the presence of processing mannosidase inhibitors or following treatment with ionophores, such as CCCP, that blocked the transport of glycoproteins between the ER and Golgi. The early literature on the ER processing mannosidases is full of confusing and contradictory data indicating the presence of dMNJ-sensitive and dMNJ-resistant mannosidase activities producing Man8GlcNAc2 and smaller structures (2). In addition, mannosidases were detected, purified, or cloned that had substrate specificities that did not match the substrate specificities predicted by the metabolic labeling studies in cultured cells. It is now clear that at least some of the prior confusion was the result of the presence of multiple unanticipated mannosidase activities in the ER. Recent data now indicate that there are at least three principal mannosidase activities in the ER of mammalian cells (2). The first mannosidase activity is an endo-a-mannosidase that cleaves Glc3-1Man9GlcNAc2 to Man8GlcNAc2 isomer A. The second activity, termed ER mannosidase I, cleaves Man9GlcNAc2 to Man8GlcNAc2 isomer B, analogous to the processing mannosidase in S. cerevisiae. A third activity, termed ER mannosidase II, cleaves Man9GlcNAc2 to Man8GlcNAc2 isomer C . The latter enzyme is a Class 2 mannosidase, whereas ER mannosidase I is a Class 1 enzyme.
In vitro enzyme assays using ER membrane preparations have been used to examine ER mannosidase I activity in rat liver. This activity was shown to be sensitive to inhibition by dMNJ, KIF, and EDTA, but not SW or DIM (44). Human ER mannosidase I has recently been cloned, expressed, and characterized and found to have a high degree of sequence similarity to yeast ER mannosidase I (45) (GenBank nucleotide entry) (GenBank protein sequence entry) (OMIM entry) (Unigene cluster) (Locus Link) (Sequence family). In addition, two putative open reading frames were identified in the C. elegans genome that have a high degree of sequence similarity to the yeast and human enzymes indicating that they may be nematode orthologs . Sequence analysis indicates that, like yeast ER mannosidase I, the human and C. elegans proteins have a single transmembrane domain near their NH2-terminus. The proposed cytoplasmic tails and the "stem regions" of the enzymes are quite variable in length and in sequence, but the COOH-terminal 440-504 amino acid region that contains the catalytic domain of the yeast enzyme is highly conserved. These four sequences constitute the presently known members of the ER mannosidase I subgroup of Class 1 mannosidases, although all organisms between yeast and mammals would be predicted to contain an equivalent of this Class 1 a-mannosidase sub-group.
Recombinant human ER mannosidase I has been expressed and purified and the enzyme has many catalytic characteristics in common with the yeast enzyme, including substrate specificity, the requirement for Ca+2 for catalytic activity, and inhibition by 1-deoxymannojirimycin (dMNJ) and kifunensine (KIF), but not swainsonine . A recombinant epitope-tagged form of the human enzyme is also localized to the ER of mammalian cells indicating a conservation of localization as well as function of this ER enzyme. Recently the structure of the human enzyme has been determined in the presence and absence of KIF and dMNJ (6). This structure is further described inthe Structural Biology section.
The position of ER mannosidase I as the last fully conserved step in glycoprotein processing between yeast and mammals argues strongly in favor of a critical role for this enzyme in order to maintain selective pressure during eukaryotic evolution. The proposal that the enzyme acts as the timing step in determining the failure of glycoprotein folding in ER and targeting the substrates for degradation is a reasonable hypothesis for this essential role. The mechanism and components involved in this timing step remain to be elucidated.