Endoplasmic reticulum-associated degradation of glycoproteins
bearing Man5GlcNAc2 and Man9GlcNAc2 species in the MI8-5
CHO cell line
Franc¸ois Foulquier, Sandrine Duvet, Andre
´
Klein, Anne-Marie Mir, Fre
´
de
´
ric Chirat and Rene
´
Cacan
Unite
´
de Glycobiologie Structurale et Fonctionnelle, CNRS-UMR 8576, IFR 118, GDR CNRS 2590, Universite
´
des Sciences
et Technologies de Lille, Villeneuve d’Ascq, France
Endoplasmic reticulum-associated degradation of newly
synthesized glycoproteins has been demonstrated previously
using various mammalian cell lines. Depending on the cell
type, glycoproteins bearing Man9 glycans and glycoproteins
bearing Man5 glycans can be efficiently degraded. A wide
variety of variables can lead to defective synthesis of lipid-
linked oligosaccharides and, therefore, in mammalian cells,
species derived from Man9GlcNAc2 or Man5GlcNAc2 are
often recovered on newly synthesized glycoproteins. The
degradation of glycoproteins bearing these two species has
not been studied. We used a Chinese hamster ovary cell
line lacking Glc-P-Dol-dependent glucosyltransferase I to
generate various proportions of Man5GlcNAc2 and

sidase, probably ER mannosidase I, and this constitutes a
putative degradation signal [5]. Recently, Frenkel et al. [6]
demonstrated that the degradation of unstable N-glycopro-
teins accumulated in a rough ER subcompartment (QC
compartment) involves trimming of the sugar chain to
Man(6–5)GlcNAc2 species. Furthermore, by using a mutant
cell line that synthesizes truncated Man5GlcNAc2, Ermon-
val et al. [7] showed that the degradation of a soluble form
of ribophorin I required the formation of Man4GlcNAc2.
After the trimming of Man residues, these misfolded
N-glycoproteins are then degraded by a specific pathway
named endoplasmic reticulum-associated degradation
(ERAD). This pathway involves a deglycosylation step
located in the cytosol, and the soluble oligomannosides
released by this process were then submitted to the action of a
1,4-dideoxy-1,4-imino-
D
-mannitol (DIM)-sensitive cytosolic
mannosidase [8], leading to the formation of a specific
Man5GlcNAc1 isomer [9] before entering the lysosomal
compartment [10].
Thus, it appears that at least three oligomannoside
structures may be involved as degradation signal for ERAD
(Man8GlcNAc2 and Man6GlcNAc2 for glycoproteins
bearing Man9 species and Man4GlcNAc2 for glycoproteins
bearing Man5 species). However, this has been observed
using different cell lines. So we decided to study the
degradation of N-glycoproteins when these oligomannoside
structures are both transferred to newly synthesized glyco-
proteins in the same cell line.

nosides, we demonstrate that glycoproteins bearing Man5
are preferentially degraded over glycoproteins bearing
Man9.
Experimental procedures
Materials
[2-
3
H]Man (429 GBqÆmmol
)1
) was from Amersham (Little
Chalfont, Bucks, UK). Trypsin and castanospermine were
from Sigma. Kifunensine and DIM were from ICN, Orsay,
France. Peptide N-glycanase F (PNGase) was from Biolabs.
b-Galactosidase and b-hexosaminidase (both isolated from
Jack bean) were from Oxford GlycoSystems (Abingdon,
Oxon, UK). Glucosidase II purified from rat liver was a gift
from T. Butters (Oxford Glycobiology Institute, University
of Oxford, UK).
Cell culture
The Glc-P-Dol-dependent glucosyltransferase I-deficient
mutant Chinese hamster ovary (CHO) cell line (MI8-5)
was a gift from S. S. Krag (Johns Hopkins University,
Baltimore, MD, USA). This cell line was grown in
a-minimal essential medium (Gibco-BRL) supplemented
with 10% (v/v) fetal bovine serum, at 34 °C under 5%
CO
2
.
Metabolic labeling of oligosaccharides and pulse-chase
experiments

Soluble oligomannoside fractions obtained after the
sequential extraction were desalted on Bio-Gel P2 eluted
with 5% (v/v) acetic acid. Glycoprotein fractions obtained
at the end of the sequential extraction were digested with
trypsin (1 mgÆmL
)1
)in0.1
M
ammonium bicarbonate
buffer, pH 7.9, overnight at room temperature.
Glycopeptides were then treated with 0.5 U PNGase in
50 m
M
phosphate buffer, pH 7.2, for 4 h to release
oligosaccharides. Oligosaccharide moieties were released
from LLO by mild acid treatment (0.1
M
HCl in
tetrahydrofuran) for 2 h at 50 °C. The oligosaccharide
fractions were then desalted on Bio-Gel P2 eluted with
5% (v/v) acetic acid. Analysis was performed by HPLC
on an amino-derivatized Asahipak NH2P-50 column
(250 mm · 4.6 mm; Asahi, Kawasaki-ku, Japan) with a
solvent system of acetonitrile/water from 70 : 30 (v/v) to
50 : 50 (v/v) at a flow rate of 1 mLÆmin
)1
over 90 min.
Oligomannosides were identified on the basis of their
retention times compared with well-defined standards [11].
Elution of the radiolabeled oligosaccharides was moni-

LLOs were extended as a consequence of activation of
the unfolded protein response (UPR). To obtain various
proportions of Man5 and Man9 species on newly
synthesized glycoproteins, MI8-5 cells were preincubated
forupto120minatlowGlcconcentration(0.175m
M
)
and incubated with labeled Man for 60 min. Figure 1
shows that, when the preincubation period was increased,
the UPR was activated, because, after 20, 40 and 120 min
of Glc deprivation, the radioactivity shifted from the
truncated species (Man3GlcNAc2-PP-Dol and Man5Glc-
NAc2-PP-Dol) to Man9GlcNAc2-PP-Dol (Fig. 1A–C).
To demonstrate that glycans bound to proteins are not
the result of modification by processing enzymes, the
same experiment was performed in the presence of
kifunensine, an inhibitor of the demannosylation process.
Figure 1D,H shows that incubation in the presence of
20 l
M
kifunensine did not affect the proportion of Man5
and Man9 species bound to LLOs and glycoproteins.
This demonstrates that both species were transferred with
different proportions according to the relative abundance
of each species present on the LLOs (Fig. 1E–G).
Ó FEBS 2003 Endoplasmic reticulum-associated degradation (Eur. J. Biochem. 271) 399
Dual fate of Man5GlcNAc2 bound to glycoproteins
in the MI8-5 cell line
As described previously for MI8-5 cells, the transfer of
Man9GlcNAc2 to glycoproteins led to the formation of

kifunensine, the formation of
Man4GlcNAc2 and Man8GlcNAc2 species was strongly
inhibited (Fig. 1H). This suggests the involvement of class I
mannosidase in these demannosylation processes [14].
Fate of glycoproteins bearing Man5GlcNAc2 and
Man9GlcNAc2
Depending on the preincubation time at low Glc concen-
tration, various proportions of Man5GlcNAc2 and
Man9GlcNAc2 species could be transferred to proteins.
To obtain an equal distribution of the radioactivity between
the Man5 and Man9 populations bound to the glycopro-
teins, cells were preincubated for 40 min in 0.175 m
M
Glc,
labeled for 1 h, and chased in the culture medium (5 m
M
Fig. 1. HPLC analysis of LLO and oligo-
mannoside species bound to newly synthesized
glycoproteins during Glc deprivation of MI8-5
cells. MI8-5 cells were preincubated in
0.175 m
M
Glc for 20, 40 and 120 min, and
then pulsed for 1 h with [2-
3
H]Maninthe
same medium. Cells were then submitted to
the sequential extraction procedure. Oligo-
mannosides bound to LLO (A, B, C and D)
and glycoproteins (E, F, G and H) were ana-

most of the Man5GlcNAc2 was converted into Man4Glc-
NAc2, in contrast with the small amount of Man6GlcNAc2
originating from the Man9GlcNAc2 species.
ERAD of glycoproteins bearing Man5 and Man9
populations in MI8-5 cells
The N-glycosylation process is accompanied by the release
of soluble oligomannosides. At least some of this soluble
material has been shown to originate from glycoprotein
degradation [16,17] after the quality control mechanism.
Figure 4A shows the pattern of oligomannosides bound to
glycoproteins after 20 min of preincubation at 0.175 m
M
Glc followed by 1 h of labeling and a 1 h chase in culture
medium containing 5 m
M
Glc and 5 m
M
Man. Under these
conditions as described previously [17], the profile of soluble
oligomannosides contained three major species: Man4Glc-
NAc1, Man5GlcNAc1 and Glc1Man5GlcNAc1 (Fig. 4B).
The smaller species (from Man3GlcNAc1 to Man1Glc-
NAc1) correspond to lysosomal degradation of the soluble
oligomannosides, as previously demonstrated [18]. As these
species may result from the action of cytosolic mannosidase
[8], the same experiment was performed in the presence of
DIM to avoid cytosolic demannosylation of glycans
released during the degradation process. Figure 4C shows
the profile of soluble oligomannosides when the same
experiment was performed in the presence of 1 m

teins were analyzed by HPLC as described in Experimental
procedures. (C) HPLC profile of oligomannosides obtained after
incubation with purified glucosidase II. G1M9 indicates oligoman-
nosides containing one Glc, nine Man, and two GlcNAc residues. M4,
M5, M8 and M9 indicate oligomannosides containing four, five, eight
and nine Man residues and two GlcNAc residues, respectively. X
indicates an unidentified peak.
Ó FEBS 2003 Endoplasmic reticulum-associated degradation (Eur. J. Biochem. 271) 401
Glc1Man9GlcNAc1 obtained with soluble oligomanno-
sides was twofold higher than the ratio Glc1Man5Glc-
NAc2/Glc1Man9GlcNAc2 obtained with the glycoprotein
pattern. This indicates that, when glycoproteins bearing
Man9 and Man5 are synthesized in the same cell line, those
bearing Man5 are more efficiently degraded.
Furthermore, when various proportions of Man5 were
transferred to glycoproteins as a result of different preincu-
bation times at low Glc concentration, a direct relationship
was observed between the level of soluble oligomannosides
released and the proportion of Man5 species bound to the
glycoprotein fraction (Fig. 4D).
Glycoproteins bearing only Man9 species can be obtained
in two ways: (a) induction of UPR by extensive preincuba-
tion at low Glc concentration (this treatment produces
extension of oligomannosides bound to lipid intermediates)
as in Fig. 4D (0% of Man5 population); (b) incubation of
cells in the culture medium (5 m
M
Glc) which considerably
reduces the labeling. In the two cases, glycoprotein degra-
dation, measured by the ratio radioactivity bound to soluble

Man5 and Man9 populations in the same cell type. To
avoid preferential transfer of glucosylated species from
LLO, we used MI8-5 CHO cells deficient in Glc-P-Dol-
dependent glucosyltransferase I. As observed previously for
other cell lines [23], the stress induced by preincubation at
low Glc concentration leads to a cellular response named
UPR, the first level of which is the extension of LLOs.
Depending on the preincubation time, various proportions
of Man5GlcNAc2 and Man9GlcNAc2 are transferred to
proteins. The ER processing of these two species has been
studied. As expected, Man9 species gave Glc1Man9Glc-
NAc2, which revealed the action of UGGT. For Man5Glc-
NAc2 species, although it has been claimed that UGGT is
not active on this species, we have clearly demonstrated the
formation of Glc1Man5GlcNAc2. The formation of this
species has been observed previously in mutant cell lines
[24,25]. However, this result indicates that this reglucosy-
lation step is not specific to the Man-P-Dol-deficient
mutant cell line, but can be observed when Man9 species
are present on newly synthesized glycoproteins. Similarly,
the Man9 and Man5 species undergo demannosylation
reactions. Man5GlcNAc2 is converted into Man4Glc-
NAc2, which cannot be reglucosylated, and Man9Glc-
NAc2 can be demannosylated first into Man8GlcNAc2
and, after a longer chase period, into Man6GlcNAc2. This
species has been observed on resident glycoproteins, as
demonstrated for ribophorin I [26] and unstable glyco-
proteins [6].
Fig. 3. Evolution of Man9 and Man5 populations bound to proteins during a pulse-chase experiment. MI8-5 CHO cells were pulse-labeled with
[2-

tive degradation signals for glycoproteins bearing Man9 and
Man5, respectively. As we only took into account mono-
glucosylated species, the degradation of the Man5 popula-
tion was probably underestimated.
We have reported previously [27] that the level of soluble
oligomannosides released during N-glycosylation is higher
in Man-P-Dol-deficient cell lines than in wild-type cells,
which are able to elongate their lipid intermediates. We have
also demonstrated that the pattern of these oligomannosides
is related to the pattern of the glycans bound to the newly
Table 1. Radioactivity bound to Glc1Man5 and Glc1Man9 species
during chase experiments in MI8-5 cells in the presence of 1 m
M
DIM.
MI8-5 cells in the presence of 1 m
M
DIM were pulsed for 1 h and
chased in culture medium containing 5 m
M
Glc and 5 m
M
Man for 1 h
after 20 min preincubation with 0.175 m
M
Glc. Cells were then sub-
mitted to the sequential extraction procedure. The radioactivity bound
to Glc1Man5 and Glc1Man9 species was measured after counting of
the glycoprotein and soluble oligomannoside fractions, taking into
account the percentage of these species as determined by HPLC.
Values (d.p.m.) from three different experiments are shown.

absence (B) or presence (C) of DIM. (D) Relationship between the
radioactivity bound to soluble oligomannoside material and the pro-
portion of Man5 species (Man4GlcNAc2, Man5GlcNAc2 and
Glc1Man5GlcNAc2) transferred to glycoproteins. These different
proportions were obtained by using different preincubation times in
0.175 m
M
Glc before the 1-h pulse (from 120 min preincubation for
0% Man5 species to 20 min preincubation for 50% Man5 species).
OS, Soluble oligomannosides. M4, M5, M6, M7, M8, M9 indicate
oligomannosides with four, five, six, seven, eight and nine Man resi-
dues and two GlcNAc residues at the reducing end. G1M5 indicates
oligomannosides with five Man and two GlcNAc residues at the
reducing end and one Glc residue, respectively. G1M9 indicates oligo-
mannosides with one Glc residue and nine Man and two GlcNAc
residuesatthereducingend.M4Gn1,M5Gn1,M6Gn1,M7Gn1,
M8Gn1, and M9Gn1 indicate oligomannosides with four, five, six,
seven, eight and nine Man residues and one GlcNAc residue at the
reducing end. G1M5Gn1 and G1M9Gn1 indicate oligomannosides
with one Glc residue, five or nine Man residues and one GlcNAc
residue at the reducing end.
Ó FEBS 2003 Endoplasmic reticulum-associated degradation (Eur. J. Biochem. 271) 403
synthesized glycoproteins [18]. We have shown here that,
when Man5GlcNAc2 oligosaccharides are transferred to
glycoproteins in cells able to synthesize and transfer
Man9GlcNAc2, the glycoproteins bearing these truncated
glycans are preferentially degraded. It means that the
mechanism of degradation found in mutant cells is also used
by normal cells.
Acknowledgements

5. Jakob, C.A., Burda, P., Roth, J. & Aebi, M. (1998) Degradation
of misfolded endoplasmic reticulum glycoproteins in Saccharo-
myces cerevisiae is determined by a specific oligosaccharide
structure. J. Cell Biol. 142, 1223–1233.
6. Frenkel, Z., Gregory, W., Kornfeld, S. & Lederkremer, G.Z.
(2003) ER-associated degradation of mammalian glycoproteins
involves sugar chain trimming to Man
6)5
GlcNAc
2
. J. Biol. Chem.
278, 34119–34124.
7. Ermonval, M., Kitzmuller, C., Mir, A.M., Cacan, R. & Ivessa,
N.E. (2001) N-glycan structure of a short-lived variant of ribo-
phorin I expressed in the MadIA214 glycosylation-defective cell
line reveals the role of a mannosidase that is not ER mannosidase I
in the process of glycoprotein degradation. Glycobiology 11,
565–576.
8. Weng, S. & Spiro, R.G. (1996) Endoplasmic reticulum kifu-
nensine-resistant alpha-mannosidase is enzymatically and
immunologically related to the cytosolic alpha-mannosidase.
Arch. Biochem. Biophys. 325, 113–123.
9. Kmie
´
cik, D., Herman, V., Stroop, C.J., Michalski, J.C., Mir,
A.M., Verbert, A. & Cacan, R. (1995) Catabolism of glycan
moieties of lipid intermediates leads to a single Man
5
GlcNAc
oligosaccharide isomer: a study with permeabilized CHO cells.

12721.
17. Cacan,R.,Duvet,S.,Labiau,O.,Verbert,A.&Krag,S.S.(2001)
Monoglucosylated oligomannosides are released during the
degradation process of newly synthesized glycoproteins. J. Biol.
Chem. 276, 22307–22312.
18. Duvet, S., Labiau, O., Mir, A.M., Kmiecik, D., Krag, S.S., Ver-
bert, A. & Cacan, R. (1998) Cytosolic deglycosylation process of
newly synthesized glycoproteins generates oligomannosides pos-
sessing one GlcNAc residue at the reducing end. Biochem. J. 335,
389–396.
19. Spiro, M.J. & Spiro, R.G. (1991) Potential regulation of
N-glycosylation precursor through oligosaccharide-lipid hydro-
lase action and glucosyltransferase-glucosidase shuttle. J. Biol.
Chem. 266, 5311–5317.
20. Rearick, J.I., Chapman, A. & Kornfeld, S. (1981) Glucose star-
vation alters lipid-linked oligosaccharide biosynthesis in Chinese
hamster ovary cells. J. Biol. Chem. 256, 6255–6261.
21. Marquardt, T. & Freeze, H. (2001) Congenital disorders of gly-
cosylation: glycosylation defects in man and biological models for
their study. Biol. Chem. 382, 161–177.
22. Freeze, H.H. (2002) Human disorders in N-glycosylation and
animal models. Biochim. Biophys. Acta 1573, 388–393.
23. Doerrler, W.T. & Lehrman, M.A. (1999) Regulation of the
dolichol pathway in human fibroblasts by the endoplasmic
reticulum unfolded protein response. Proc. Natl Acad. Sci. USA
96, 13050–13055.
24. Ermonval, M., Duvet, S., Zonneveld, D., Cacan, R., Buttin, G. &
Braakman, I. (2000) Truncated N-glycans affect protein folding in
the ER of CHO-derived mutant cell lines without preventing
calnexin binding. Glycobiology 10, 77–87.