3474
Research Article
The oligosaccharyltransferase subunits OST48, DAD1
and KCP2 function as ubiquitous and selective
modulators of mammalian N-glycosylation
Peristera Roboti and Stephen High*
Faculty of Life Sciences, The University of Manchester, Michael Smith Building, Oxford Road, Manchester M13 9PT, UK
*Author for correspondence ([email protected])
Journal of Cell Science
Accepted 20 February 2012
Journal of Cell Science 125, 3474–3484
ß 2012. Published by The Company of Biologists Ltd
doi: 10.1242/jcs.103952
Summary
Protein N-glycosylation is an essential modification that occurs in all eukaryotes and is catalysed by the oligosaccharyltransferase (OST)
in the endoplasmic reticulum. Comparative studies have clearly shown that eukaryotic STT3 proteins alone can fulfil the enzymatic
requirements for N-glycosylation, yet in many cases STT3 homologues form stable complexes with a variety of non-catalytic OST
subunits. Whereas some of these additional components might play a structural role, others appear to increase or modulate Nglycosylation efficiency for certain precursors. Here, we have analysed the roles of three non-catalytic mammalian OST components by
studying the consequences of subunit-specific knockdowns on the stability and enzymatic activity of the OST complex. Our results
demonstrate that OST48 and DAD1 are required for the assembly of both STT3A- and STT3B-containing OST complexes. The
structural perturbations of these complexes we observe in OST48- and DAD1-depleted cells underlie their pronounced
hypoglycosylation phenotypes. Thus, OST48 and DAD1 are global modulators of OST stability and hence N-glycosylation. We
show that KCP2 also influences protein N-glycosylation, yet in this case, the effect of its depletion is substrate specific, and is
characterised by the accumulation of a novel STT3A-containing OST subcomplex. Our results suggest that KCP2 acts to selectively
enhance the OST-dependent processing of specific protein precursors, most likely co-translational substrates of STT3A-containing
complexes, highlighting the potential for increased complexity of OST subunit composition in higher eukaryotes.
Key words: Endoplasmic reticulum, Glycoprotein synthesis, Membrane protein complex, STT3 proteins
Introduction
The diversity and complexity of the eukaryotic proteome is
greatly enhanced by a variety of covalent modifications, and for
proteins entering the secretory pathway the most prevalent of
these is the addition of branched chain oligosaccharides to
suitable asparagine residues. This N-linked glycosylation acts to
recruit specific chaperones present in the lumen of the
endoplasmic reticulum (ER) that promote protein folding,
assembly and disulphide bond formation, contributes to the
removal of misfolded glycoproteins, and influences protein
trafficking through the secretory pathway (Aebi et al., 2010).
N-glycosylated proteins mediate numerous fundamental cellular
processes, and defects in their production are linked to a variety
of human diseases (Ohtsubo and Marth, 2006; Zelinger et al.,
2011). In eukaryotes, the process of protein N-glycosylation is
catalysed by the oligosaccharyltransferase (OST), a membrane
protein complex of the rough ER that transfers preassembled
high-mannose oligosaccharides to asparagine residues present in
the appropriate context, or sequon (N-x-S/T; x?P), of a nascent
polypeptide chain (Kelleher and Gilmore, 2006). These Nsequons are efficiently modified when present in flexible and
surface-exposed segments of a polypeptide (Kowarik et al.,
2006a; Lizak et al., 2011). Interestingly, the bulk of protein
N-glycosylation appears to occur in a co-translational manner;
with the process coupled to ongoing protein synthesis at, and
translocation into and across, the ER membrane (Harada et al.,
2009; Ruiz-Canada et al., 2009).
Whilst eukaryotes such as trypanosomatids and giardia are able
to N-glycosylate newly synthesised polypeptides using a singlecomponent enzyme, a member of the STT3 family (Castro et al.,
2006; Izquierdo et al., 2009; Kelleher and Gilmore, 2006), many
eukaryotic OSTs consist of multiple subunits. The composition
and subunit organisation of the mammalian enzyme is particularly
complicated, having two homologues of Saccharomyces
cerevisiae Stt3p, STT3A and STT3B. STT3 proteins are the
catalytically active components of the OST (Nilsson et al., 2003;
Yan and Lennarz, 2002) and associate with a common set of noncatalytic subunits: ribophorin I (Ost1p in yeast), ribophorin II
(Swp1p), OST48 (Wbp1p), DAD1 (Ost2p), N33 or IAP (Ost3p or
Ost6p), and OST4 (Ost4p), generating two distinct OST isoforms
in mammals (Kelleher et al., 2003). The less enzymatically active,
more selective STT3A isoform catalyses co-translational
N-glycosylation, while the more active, less selective STT3B
isoform can modify some sites skipped by STT3A and also
mediate post-translational N-glycosylation of defined sequons
(Ruiz-Canada et al., 2009). Hence, as observed in Trypanosoma
brucei (Izquierdo et al., 2009), it appears that different mammalian
STT3 isoforms perform distinct but partially redundant roles,
presumably ensuring optimal protein N-glycosylation of their
respective substrates at the ER (Ruiz-Canada et al., 2009).
Journal of Cell Science
Modulatory subunits of the OST
Many studies of non-catalytic OST subunits have used
Saccharomyces cerevisiae as a model system (Yan and
Lennarz, 2005), and mutations in several of these components,
including Ost1p, Ost2p and Wbp1, result in reductions in Nlinked glycosylation and in some cases destabilisation of the OST
complex (Kelleher and Gilmore, 2006). More recently, the Ost3p
and Ost6p subunits have been shown to have disulphide
oxidoreductase activity that may enhance the N-glycosylation
of defined N-sequons by preventing the nascent polypeptide from
prematurely forming disulphide bonds during co-translational Nglycosylation (Schulz et al., 2009). The contribution(s) of noncatalytic OST subunits to protein N-glycosylation in mammals is
less well defined. Ribophorin I (the homologue of Ost1p)
regulates the delivery of a subset of membrane proteins to the
OST catalytic core (Wilson and High, 2007; Wilson et al., 2008);
whilst ribophorin II is required for the efficient N-glycosylation
of P-glycoprotein, although its precise contribution in this context
is unclear (Honma et al., 2008). Previous studies of DAD1, the
mammalian equivalent of Ost2p, have relied solely on a
temperature sensitive mutant (Kelleher and Gilmore, 1997;
Sanjay et al., 1998), and to date there have been no attempts at
perturbing the function of mammalian OST48, the Wbp1
homologue. The identification of two novel OST subunits DC2
(or OSTC) and keratinocyte-associated protein 2 (KCP2) (Roboti
and High, 2012; Shibatani et al., 2005), present only in higher
eukaryotes, adds to the potential mechanistic complexity of the
enzyme in higher eukaryotes. Whilst mammalian DC2 shows
weak homology to S. cerevisiae Ost3p and Ost6p (Shibatani et al.,
2005), a search for conserved domains in KCP2 found no
homology to known protein families. Hence, the potential
function of KCP2 during protein N-glycosylation is unclear
(Kelleher and Gilmore, 2006; Roboti and High, 2012), and
lacking any experimental support (Wilson et al., 2011).
In this study, we assess the contributions of three poorly
characterised non-catalytic subunits to the structural and
functional integrity of mammalian OST complexes. OST48,
DAD1 or KCP2 were each knocked down using a small
interfering RNA (siRNA) approach and the effects of these
treatments on the stability of a range of individual OST subunits
was investigated. We have also addressed the contribution of
these three subunits to the stability of native OST complexes,
and determined the effect of their knockdown upon the
N-glycosylation of a small set of model substrates using two
different approaches. We show that OST48 and DAD1 have a
major role in stabilising both STT3A- and STT3B-containing
OST complexes; hence their depletion results in a severe and
global hypoglycosylation phenotype that is directly comparable
to that seen upon the depletion of both catalytic STT3 proteins.
Our studies of KCP2 reveal that it is essential for the formation of
the major STT3A-containing OST complex, and we show that
KCP2 acts to facilitate N-glycosylation in a substrate-specific
manner, underlining the additional complexity of the process in
higher eukaryotes.
Results
OST48 and DAD1 are required for the stability of both
catalytic STT3 proteins
We first assessed the effect of knocking down previously defined
OST components on the steady-state levels of the catalytic
STT3A and STT3B subunits and vice versa. Triton X-100
extracts of HeLa cells, prepared 72 h after transfection with
3475
siRNAs specific for a particular OST subunit, were analysed by
quantitative immunoblotting (Fig. 1; Table 1). OST48 siRNA
treatment markedly decreased OST48 levels by ,4-fold,
resulting in ,2-fold reduction of both STT3A and STT3B and
a substantial destabilisation of DAD1 and KCP2 (Fig. 1A, cf.
lanes 1 and 4; for quantification see Fig. 1B; Table 1). Depletion
of DAD1 (,7-fold reduction) led to a 2- to 3-fold decrease in
STT3A, STT3B and OST48, and a ,6-fold reduction of KCP2
(Fig. 1A, cf. lanes 1 and 5; Fig. 1B; Table 1). Various siRNAs
targeting KCP2 were tested and we found one to be extremely
effective at knocking down KCP2 protein levels (.10-fold
reduction; Fig. 1A, aKCP2 panel, cf. lanes 1, 6 and 7). In contrast
to the OST48 and DAD1 knockdowns, this KCP2 siRNA did not
affect the stability of STT3A, STT3B, or OST48, although we did
see a modest reduction in DAD1 levels (Fig. 1A, cf. lanes 1 and
6; Fig. 1B; Table 1).
The STT3A siRNA substantially reduced STT3A protein
levels (,7-fold decrease), also resulting in a ,2-fold loss of
OST48 and DAD1, and an even more striking reduction in KCP2
levels (Fig. 1A, cf. lanes 1 and 2; Fig. 1B; Table 1). Importantly,
pronounced destabilisation of KCP2 was also observed with a
second siRNA targeting STT3A (siSTT3A #2; supplementary
material Fig. S1), and after knocking down STT3A in HepG2
cells (Fig. 1C), thereby ruling out off-target or cell-type-specific
effects. On this basis, the reduction of KCP2 we observe in
OST48- and DAD1-depleted cells may actually reflect the
significant loss of STT3A resulting from these treatments.
Conversely, the efficient knock down of KCP2 had no effect
on STT3A levels, and in fact only DAD1 levels showed a
significant change displaying a modest reduction (Fig. 1B;
Table 1). As previously observed, knockdown of STT3A had
little effect on STT3B (Ruiz-Canada et al., 2009). In contrast, the
siRNA-mediated 3-fold reduction of STT3B led to a marked
increase in STT3A levels (cf. Ruiz-Canada et al., 2009), but
reduced DAD1 by 2-fold with modest effects on OST48 and
KCP2 levels (Fig. 1A, cf. lanes 1 and 3; Fig. 1B; Table 1).
Destabilisation of these OST subunits did not noticeably affect
the steady-state levels of the a-subunit of the ER resident Sec61
translocon complex (Fig. 1A, aSec61a panel) (Ruiz-Canada
et al., 2009; Sanjay et al., 1998). Taken together, the mutual
destabilisation of multiple OST subunits upon the siRNAmediated reduction of individual non-catalytic components is
consistent with the loss of certain subunits perturbing the
complex as a whole.
In order to rule out any loss of OST components resulting from
the Triton X-100 extraction procedure we used to prepare the cell
extracts, we also analysed whole cell lysates resulting from direct
solubilisation in SDS sample buffer (supplementary material
Fig. S2A,B). The effects of the various siRNA treatments upon
the five OST subunits analysed are comparable using both
approaches (cf. Fig. 1B; supplementary material Fig. S2C). The
only exception is that we no longer observe an increase in STT3A
signal following STT3B depletion. This is in contrast to both our
data using the Triton X-100 approach, and an independent study
that analysed digitonin high-salt extracts following siRNAmediated depletions (Ruiz-Canada et al., 2009). It is worth
noting that the levels of STT3A present in HeLa cells appear
much lower than other OST subunits including STT3B (Roboti
and High, 2012), and this may be reflected in the sensitivity of its
detection. We also investigated the possibility of a linkage
between OST subunit expression at the mRNA level rather than
Journal of Cell Science
3476
Journal of Cell Science 125 (14)
Fig. 1. Effect of siRNA-mediated depletion of selected
OST subunits on stability. (A) HeLa cells were
transfected with siRNAs targeting STT3A, STT3B, OST48,
DAD1, KCP2, or a non-targeting control (nt) and
immunoblotted for the proteins indicated after 72 h.
(B) Quantitative representation of the data from several
experiments as presented in A. Fluorescent signals were
quantified, normalised to the corresponding a-tubulin
signals, and expressed as a percentage of the signal in
control cells (see also Table 1). Values are means 6 s.e.m.
for n$7 (*P,0.05, **P,0.01, ***P,0.001). (C) HepG2
whole-cell lysates prepared 72 h after siRNA transfection
were analysed by immunoblotting for the indicated proteins.
via the co-stabilisation of individual protein subunits. The OST48
and DAD1 proteins appear to be the most sensitive to the
depletion of other OST subunits (see Table 1); hence, we used
RT-PCR to analyse the levels of their mRNAs following the
knockdown of each of the five subunits under study. For both
OST48 and DAD1 we see no indication of any substantial
destabilisation of their mRNAs by siRNAs targeting the other
subunits (supplementary material Fig. S3A,B). Thus, within the
limitations of the semi-quantitative approach used, we conclude
the depletion of any one of several OST subunits may result in
the destabilisation of other members of the same protein complex
(see also Discussion).
N-glycosylation enhances protein folding and degradation of
non-native polypeptides (Aebi et al., 2010); hence any
perturbation of this process might increase levels of aberrant
proteins, resulting in an unfolded protein response (UPR) (Ron
and Walter, 2007). We examined the inositol-requiring protein-1
(IRE1) branch of the UPR after OST subunit knockdown, but
found no evidence for any X-box binding protein-1 (XBP1)
mRNA splicing indicative of ER stress (supplementary material
Fig. S4). We conclude that despite efficient knockdowns of
several individual components, either the residual OST activity
is sufficient to cope with the cellular requirement for
N-glycosylation under our experimental conditions and/or there
is substantial functional redundancy between the distinct
mammalian OST complexes containing STT3A or STT3B (cf.
Ruiz-Canada et al., 2009; see also below).
OST48 and DAD1 are required for the oligomeric assembly
of native OST isoforms
To test the hypothesis that the wide-ranging effects of OST48 or
DAD1 depletion (Fig. 1; Table 1) reflect the disruption of native
complexes, we monitored OST complex formation by blue native
(BN)-PAGE analysis (Wittig et al., 2006) of siRNA-treated cells.
In the first instance, we examined the incorporation of the
relevant OST subunits in untreated HeLa cells (Fig. 2A). As
previously revealed in our studies using HepG2 cells (Roboti and
High, 2012), STT3A was primarily detected in a , 500 kDa
Table 1. OST subunit protein level in siRNA-treated HeLa cells as a percentage of non-targeting control
Protein
siSTT3A
siSTT3B
siOST48
siDAD1
siKCP2
STT3A
STT3B
OST48
DAD1
KCP2
14.962.8
87.564.0
65.866.0
42.666.5
10.462.7
142.3612.0
32.564.0
81.066.3
53.866.5
80.5610.9
42.263.7
54.162.2
23.262.7
18.361.9
32.264.5
33.563.6
51.666.5
42.765.1
14.362.3
16.863.3
93.765.9
101.8610.3
119.8613.6
68.369.4
8.762.8
3477
Journal of Cell Science
Modulatory subunits of the OST
Fig. 2. Depletion of non-catalytic subunits disrupts OST complexes. (A) HeLa cell digitonin extracts were resolved by 4–10% gradient BN-PAGE and
analysed by immunoblotting with antibodies against Sec61a, KCP2, STT3A, STT3B, DAD1 and OST48. Protein complexes containing STT3A (OSTC500),
STT3B (OSTC550) and Sec61a (Sec61C) are indicated. (B,C) HeLa cells were transfected with non-targeting siRNA (nt) or siRNAs targeting STT3A, STT3B,
OST48, DAD1 or KCP2, processed as for A, and analysed by SDS-PAGE (B) or BN-PAGE (C) and immunoblotting. In C, discrete OSTC550, OSTC500 (arrow)
and OSTC470 (open circle) species are indicated. For the aSTT3A panel in C, two separate blots were aligned using the OSTC500 species of STT3B-depleted
samples as a reference.
complex (OSTC500; Fig. 2A, lane 3) while STT3B was present
in a larger complex(es) of ,520–580 kDa (OSTC550; Fig. 2A,
lane 4). On the basis of co-migration, both OSTC500 and
OSTC550 appear to contain OST48 and DAD1 (Fig. 2A, lanes 5
and 6). A substantial amount of KCP2 appears to be present in the
STT3A-containing isoform OSTC500, consistent with previous
evidence of a physical association (Roboti and High, 2012).
Some KCP2 may also be present in a larger complex that comigrates with STT3B-containing OSTC550 species (Fig. 2A,
lane 2; see also Roboti and High, 2012).
Digitonin-solubilised homogenates were prepared from
siRNA-treated HeLa cells (cf. Ruiz-Canada et al., 2009) and
standard immunoblot analysis following SDS-PAGE confirmed
the relevant knockdowns (Fig. 2B). The immunoblot analysis of
discrete OST complexes, resolved by BN-PAGE using an
antibody specific for STT3A, revealed an almost complete loss
of the OSTC500 species in cells with siRNA-mediated reductions
in the levels of STT3A, OST48, and DAD1 (Fig. 2C, aSTT3A
panel, cf. lanes 1, 2, 5 and 6). In contrast, the OSTC500 appeared
unaffected by STT3B depletion (Fig. 2C, aSTT3A panel, cf.
lanes 1 and 3, 4). Intriguingly, reduced KCP2 levels led to a loss
of OSTC500 species with the concurrent appearance of a fastermigrating, STT3A-containing species denoted OSTC470
(Fig. 2C, aSTT3A panel, cf. lanes 1 and 7, open circle). Very
low levels of STT3B-containing OSTC550 were observed in cells
knocked down for STT3B, OST48, and DAD1 (Fig. 2C, aSTT3B
panel, cf. lanes 1, 3, 4, and 5). In contrast, depletion of STT3A or
KCP2 had no obvious effect on these species (Fig. 2C, aSTT3B
panel, cf. lanes 1, 2 and 6).
Analysis of both OST48- and DAD1-containing species further
validated the results detailed above, revealing a pronounced
reduction in the levels of both OSTC500 and OSTC550 after
depletion of OST48 or DAD1 (Fig. 2C, aOST48 and aDAD1
panels, cf. lanes 1, 4 and 5) and the appearance of a OSTC470like species in KCP2-knockdown cells (Fig. 2C, aOST48 and
aDAD1 panels, lane 6, open circles). A comparable ,470 kDa
species containing low levels of OST48 and DAD1 is also
apparent in control and STT3B-knockdown cells (Fig. 2C,
3478
Journal of Cell Science 125 (14)
aOST48 and aDAD1 panels, lanes 1 and 3, cf. open circles).
Most likely our antibody against STT3A is unable to detect this
comparatively low-abundance species in the absence of KCP2
knockdown (cf. Fig. 2C, aSTT3A panel, lanes 1, 3, 4 and 7),
and we conclude OSTC470 is an STT3A-containing OST
subcomplex that lacks KCP2 (cf. Roboti and High, 2012).
Reprobing of the immunoblots showed that depletion of the OST
subunits did not affect the steady-state levels of the physically
adjacent Sec61 translocon complex (Sec61C; see supplementary
material Fig. S5), confirming the localised effects of OST subunit
loss. In terms of the discrete OST complexes detectable by BNPAGE analysis, the knockdown of OST48 and DAD1 perturbs
complexes containing both STT3 isoforms. In contrast, the
knockdown of individual STT3 isoforms selectively perturbs
only OST complexes containing the target isoform, whilst the
knockdown of KCP2 results in a mobility shift of the major
STT3A-containing complex and has no visible effect upon the
STT3B-containing species.
Journal of Cell Science
OST48, DAD1 and KCP2 facilitate protein N-glycosylation
in vitro
Having established that the depletion of individual OST subunits
appears to be accompanied by distinct effects on the structural
integrity of the whole complex, an in vitro analysis of the
functional consequences of these various subunit depletions was
performed. HeLa cells were knocked down for the relevant OST
subunit, semi-permeabilised with digitonin, and the Nglycosylation capacity of the ER of these preparations was then
analysed using a well-established cell-free system (Wilson
and High, 2007). Aquaporin 1 (AQP1) is a channel-forming
protein with a complex membrane topology, and the polypeptide
chain is known to be exclusively N-glycosylated at Asn-42
(supplementary material Fig. S6A). When its biogenesis was
analysed using the cell free system, a ,50% reduction in the
level of N-glycosylated AQP1 (AQP1-1) was observed in
STT3A-depleted cells, whilst depletion of STT3B resulted in a
,25% reduction (Fig. 3A). Strikingly, the knockdown of OST48
or DAD1 led to a greater than 70% reduction of N-glycosylation
in a parallel analysis (Fig. 3A). Collectively, these results suggest
that the single N-sequon in AQP1 can be modified by STT3A and
to a lesser extent by STT3B, emphasising their overlapping roles
(Ruiz-Canada et al., 2009). The even more pronounced inhibitory
effects of OST48 and DAD1 knockdowns that we observe are
consistent with the disruption of both OST isoforms following
these treatments. Intriguingly, KCP2 depletion resulted in a
,25% reduction in the levels of N-glycosylated AQP1 (Fig. 3A),
providing the first evidence that KCP2 plays any functional role
during this process. We ascribe this contribution of KCP2 to
substrates utilising STT3A-containing OST complexes for their
N-glycosylation (see below).
Prepro-alpha-factor (ppaF) is a yeast secretory glycoprotein
with a cleavable signal sequence and three sites for Nglycosylation (supplementary material Fig. S6B). A modest
decrease in the levels of fully glycosylated paF (paF-3) was
observed in STT3A-, STT3B- and KCP2-knockdown cells, and
this was accompanied by an increase in non-glycosylated species
consistent with these subunits contributing to OST activity (paF0; Fig. 3B). In contrast, although the levels of paF-3 were not
significantly affected by the OST48 or DAD1 knockdowns, the
levels of paF species lacking one N-glycan (paF-2) were visibly
increased, especially by the OST48 siRNA treatment (Fig. 3B),
indicating hypoglycosylation. Asialoglycoprotein receptor
(ASGPR) is a type II membrane protein with two Nglycosylation sequons (supplementary material Fig. S6C). In
this case, the individual knockdown of the two catalytic
STT3 subunits had no measurable effect on the levels of
fully glycosylated ASGPR (ASGPR-2; Fig. 3C), suggesting
substantial redundancy of the different mammalian OST
complexes for modification of this precursor. Intriguingly,
normal glycosylation of ASGPR was also detected after KCP2
depletion (Fig. 3C), indicating any role of this component in
relation to protein N-glycosylation may be precursor-specific.
Conversely, depletion of OST48 or DAD1 again resulted in the
appearance of intermediates lacking one N-glycan (ASGPR-1;
Fig. 3C), as previously observed with paF (Fig. 3B).
Furthermore, we also observed a similar effect when we
studied a second type II membrane protein, the invariant chain
(Ii) of the major histocompatibility complex (MHC) class II,
using the same approach (supplementary material Fig. S7). Taken
together, the clearly visible accumulation of hypoglycosylated
species observed upon the depletion of OST48 and DAD1
(Fig. 3B,C; supplementary material Fig. S7; cf. Silberstein et al.,
1995; te Heesen et al., 1992) are entirely consistent with the
functional perturbation of OST complexes containing both STT3
isoforms (cf. Figs 1, 2; also see Discussion).
Lastly, opsin, a polytopic membrane protein with two Nsequons (supplementary material Fig. S6D) was analysed. In this
case, evidence for alterations at the qualitative level were far less
obvious than for the other precursors with no hypoglycosylated
forms apparent (Fig. 3D). Quantification showed that any
potential changes in N-glycosylation were not statistically
significant, and we conclude that the effects of OST subunit
depletion detected using our in vitro assay are substrate-specific.
Overall, our in vitro analysis suggests that OST48, DAD1 and
KCP2 can all facilitate N-glycosylation in a substrate-specific
manner. Notably, in every case we found that the effect of
tunicamycin treatment of cells was more pronounced than the
knockdown of any of the OST subunits analysed here, including
either of the two mammalian STT3 isoforms (Fig. 3). We
conclude that although both quantitative immunoblotting of
individual OST subunits and BN gel analysis of OST complexes
indicate substantial siRNA-mediated perturbation, the residual
OST activity is present in relative excess compared to the low
levels of substrates that are synthesised in vitro.
OST48, DAD1 and KCP2 are required for efficient
N-glycosylation in HeLa cells
In order to further explore the effects of OST subunit depletion,
we next adopted a cell-based system to study two previously
characterised glycoproteins (Ruiz-Canada et al., 2009). Firstly,
human coagulation factor VII (FVII) was transiently expressed in
HeLa cells, and its biosynthesis and N-glycosylation analysed by
pulse labelling. After a 10 min pulse of radiolabel, fully
glycosylated (FVII-2) and mono-glycosylated (FVII-1) species
were readily apparent, but only a trace of non-glycosylated
(FVII-0) species was seen (Fig. 4A; 4B, lane 2). The most
striking effects were observed upon depletion of STT3A, OST48
and DAD1, and in each case the levels of FVII-0 were noticeably
increased, whilst FVII-2 levels was decreased and FVII-1 levels
not significantly affected (Fig. 4B, cf. lanes 2, 3 5 and 6). In
contrast, depletion of STT3B or KCP2 had no statistically
significant effect in this assay (Fig. 4B, cf. lanes 2, 4 and 7).
3479
Journal of Cell Science
Modulatory subunits of the OST
Fig. 3. Depletion of non-catalytic OST subunits perturbs N-glycosylation in vitro. HeLa cells treated as for Fig. 1 were semi-permeabilised and used for
the in vitro translation of (A) aquaporin 1 (AQP1), (B) prepro-alpha-factor (ppaF), (C) asialoglycoprotein receptor (ASGPR) and (D) opsin. Tunicamycin (tun)treated cells provided a comparison. Fully glycosylated, non-glycosylated and glycosylated intermediates are indicated for each precursor. Glycoforms were
quantified, normalised to corresponding total protein levels, and expressed relative to the normalised signal in non-targeting siRNA-treated cells. Values show the
mean 6 s.e.m. for n53 (*P,0.05, **P,0.01, ***P,0.001).
Prosaposin (pSAP) is an abundant endogenous glycoprotein of
HeLa cells (Ruiz-Canada et al., 2009) and the early stages of its
biosynthesis were also studied by pulse labelling. In control cells,
the predominant form of pSAP recovered is the fully-glycosylated
form (Fig. 4C; 4D, lane 2, pSAP-5), whilst tunicamycin treatment
effectively prevented any N-glycosylation to yield nonglycosylated pSAP (Fig. 4D, cf. lanes 1 and 2, pSAP-0). As for
exogenously expressed FVII, depletion of STT3A, OST48, or
DAD1 resulted in a substantial perturbation of pSAP Nglycosylation, whilst reduction of STT3B had little, if any, effect
(Fig. 4D, cf. lanes 2–6). In each case, the reductions in the level of
pSAP-5 were accompanied by increases in hypoglycosylated
species (Fig. 4D, pSAP-2, -3 and -4; see also quantification).
Strikingly, depletion of KCP2 also had a clear effect on pSAP Nglycosylation, in this case reducing pSAP-5 levels and increasing
pSAP-4 species (Fig. 4D, cf. lanes 2 and 7; see also
quantification). On this basis, we conclude that OST48, DAD1,
and KCP2 all contribute to the efficient N-glycosylation of pSAP.
Given the effects of STT3A and STT3B depletion are comparable
to those previously reported (cf. Ruiz-Canada et al., 2009), we
believe that these data most likely reflect the contributions of these
non-catalytic subunits to the STT3A-dependent co-translational
processing of pSAP (see below).
Whilst STT3A-containing OSTC500 species appear to be a
major site for co-translational N-glycosylation, STT3Bcontaining OSTC550 species may perform overlapping
functions, resulting in at least partial redundancy (Ruiz-Canada
et al., 2009). Hence, the hypoglycosylation of pSAP we observe
upon depletion of OST48, DAD1 or KCP2 might either be a
direct reflection of reduced OSTC500 function or the result of
pSAP being shunted into an alternative, OSTC550-mediated
route. To further address this issue we looked at combinations of
double knockdowns using pSAP N-glycosylation as a read-out.
As expected, the depletion of both STT3A and STT3B had a
more pronounced effect than a reduction of STT3A alone (Fig. 5,
cf. lanes 1, 2 and 4). Compellingly, this combination mirrored the
Journal of Cell Science
3480
Journal of Cell Science 125 (14)
Fig. 4. Depletion of non-catalytic
OST subunits results in protein
hypoglycosylation in HeLa cells.
(A,C) Diagrams of (A) human
coagulation factor VII (FVII) and
(C) prosaposin (pSAP) precursor
showing the signal sequence (blue) and
the N-glycosylation sequons (red). In C,
disulphide bonds (green lines) and the
saposin domains (four yellow blocks
labelled A–D) are shown.
(B,D) Glycoforms of (B) exogenous
coagulation factor VII (FVII) and (D)
endogenous prosaposin (pSAP) in HeLa
cells treated with the indicated siRNAs
or tunicamycin (tun) were measured by
immunoprecipitation and
phosphorimaging. Different glycoforms
of each precursor are indicated, as is a
non-glycosylated background species
(asterisk, see also supplementary
material Fig. S8). Glycoforms in the
siRNA-treated cells were quantified,
normalised to the total precursor levels
in the same samples, and expressed
relative to the normalised signal in nontargeting siRNA-treated cells. Values
are mean 6 s.e.m. for n53 (*P,0.05,
**P,0.01, ***P,0.001). In D, the
inset shows the data clustered by pSAP
N-glycosylation status.
depletion of OST48 alone (Fig. 5, cf. lanes 4 and 5), supporting
our proposal that reductions in OST48 levels perturb both classes
of OST complex, i.e. those containing STT3A and those
containing STT3B. This is consistent with the destabilisation of
both OSTC500 and OSTC550 species we see upon OST48
knockdown (Fig. 2). The pronounced effect of OST48 reduction
(Fig. 5, cf. lanes 1 and 5) may also reflect a role in supplying
lipid-linked oligosaccharide donors to the active site of OST
complexes (cf. Kelleher et al., 2007). In contrast, co-depletion of
DAD1 and OST48 is comparable to OST48 depletion alone
(Fig. 5, cf. lanes 5 and 9), presumably since the loss of OST48
results in a substantial loss of DAD1 (Table 1).
Although our data suggest that co-depletion of STT3B with
OST48 or DAD1 may have a modest corrective effect (Fig. 5,
lanes 5–8; see also quantification), in practice this effect most
likely reflects the up-regulation of STT3A expression resulting
from STT3B depletion (Fig. 1; Table 1); see also (Ruiz-Canada
et al., 2009). In the light of the distinct combinatorial effects of
STT3B depletion with STT3A, OST48 or DAD1 that we observe,
the absence of any additional effect of STT3B depletion in
combination with the KCP2 knockdown (Fig. 5, cf. lanes 10 and
11) is very informative. Our previous work had shown a physical
link between STT3A and KCP2 (Roboti and High, 2012) that we
have gone on to extend in this study (Figs 1, 2). Furthermore, we
see clear evidence that KCP2 is required for the efficient Nglycosylation of many, but not all, precursors (Figs 3, 4). In
comparison to the co-depletion of STT3A and STT3B, the
absence of any additional defect in pSAP N-glycosylation upon
co-depletion of KCP2 and STT3B strongly suggests that loss of
KCP2 perturbs OSTC500 activity in a way that cannot be
compensated for by the OSTC550 species (Fig. 6).
Discussion
Although the effect of siRNA-mediated depletions of several
mammalian OST subunits have been reported in recent years, our
data provides the first analysis of OST48 and DAD1 function
using this technique. At the level of its individual subunits, we
see clear evidence for the mutual stabilisation of components of
the OST, consistent with their assembly into a discrete
complex(es). Hence, in terms of OST subunit stability, the
consequences of the siRNA-mediated depletion of an individual
non-catalytic component, in this case OST48 or DAD1, are at
least as severe as the knockdown of the catalytic subunits. By
these criteria, the behaviour of KCP2 is distinct. Hence, whilst
KCP2 depletion has a marginal effect on the levels of other OST
subunits, its own stability is strongly dependent on the presence
of each of the other OST subunits tested, with the clear exception
of STT3B. These data are entirely consistent with the original
suggestion that KCP2 is a subunit of the mammalian OST
complex (Shibatani et al., 2005), and strongly support our own
contention that it is primarily associated with OST complexes
containing the STT3A isoform (Roboti and High, 2012).
Furthermore, since KCP2 lacks N-linked glycans (Roboti and
High, 2012), this effect cannot be mediated by its inefficient
glycosylation in HeLa cells depleted for other OST subunits.
Interestingly, this apparently one-way stabilisation displayed
between STT3A and KCP2 is comparable to the relationship
between other ER membrane protein assemblies, such as the
Modulatory subunits of the OST
3481
Journal of Cell Science
Fig. 6. Schematic showing a model for the sequential scanning of nascent
pSAP for N-sequons by STT3A- and STT3B-containing OST complexes.
Co-translational glycosylation of pSAP is preferentially catalysed by
OSTC500 in control and STT3B-depleted cells (route 1). In STT3A-depleted
cells, routes 1 and 2 are perturbed, hence the only alternative for skipped Nsequons is modification by OSTC550, albeit with low efficiency (route 3). In
KCP2-knockdown cells pSAP is selectively N-glycosylated by OSTC470
(route 2) despite the integrity of OSTC550 (route 3). In OST48- and DAD1depleted cells all three routes are perturbed, and remaining pSAP Nglycosylation reflects global residual OST activity.
Fig. 5. Co-depletion studies confirm OST48 function is ubiquitous,
whereas KCP2 is selective. pSAP glycoforms were measured in siRNAtreated HeLa cells as for Fig. 4, and symbols are as previously defined.
Glycoforms were quantified, normalised to total pSAP levels in the same
samples, and expressed relative to the normalised signal in non-targeting
siRNA-treated cells.
HRD1-SEL1L ubiquitin ligase complex (Iida et al., 2011). In the
round, these data highlight the role of multiple non-catalytic
subunits in stabilising the two classes of mammalian OST
complexes, and show that this role is not restricted to DAD1
(Sanjay et al., 1998) and ribophorin I (Ruiz-Canada et al., 2009).
Presumably, this mutual subunit dependence reflects the close
spatial relationship between the various components of
eukaryotic OST complexes (Chavan et al., 2006; Karaoglu
et al., 1997; Yan et al., 2005).
By using BN-PAGE combined with immunoblotting we were
able to consider the effect of depleting individual OST subunits
upon the stability of discrete OST complexes containing one or
other of the two catalytic STT3 subunits that we previously
defined in HepG2 cells (Roboti and High, 2012). Most
significantly, we observe that the depletion of either of the
DAD1 or OST48 has an even more pronounced destabilising
effect on global OST complex stability than the individual
knockdown of either catalytic STT3 isoform alone. Hence, we
conclude that mammalian OST48 and DAD1 are important for
the stability of two distinct classes of OST complexes that
contain either STT3A or STT3B (Fig. 6). In fact, our data also
indicate that the OST48 and DAD1 subunits stabilise each other
(see Table 1), consistent with both the previous study of a
temperature-sensitive version of DAD1 (Sanjay et al., 1998) and
evidence for a direct interaction between these two components
(Kelleher and Gilmore, 1997). However, although this means that
we cannot easily ascribe a more important role in OST complex
stability to one or other of these two subunits, the more
pronounced effect of their depletion upon global OST complex
stability is readily apparent. Furthermore, the yeast homologues
of both OST48 (Wbp1p) and DAD1 (Ost2p) are known to be
essential for cell viability (Silberstein et al., 1995; te Heesen et al.,
1992), underpinning the suggestion that these components play a
central role in maintaining the structural integrity of mammalian
OST complexes. Although we favour a model where the siRNA
depletion of one subunit results in the degradation of other
protein subunits by destabilising the OST complex, additional
levels of regulation, such as feedback loops that influence the
translation of mRNAs encoding other OST subunits (Mick et al.,
2010), may also contribute to their loss.
In contrast to all of the other OST subunit knockdowns that we
analysed by BN-PAGE and immunoblotting, the effect of KCP2
depletion upon detectable OST complexes was comparatively mild.
In the case of the STT3B-containing OST complexes (OSTC550),
we saw no effect, supporting our model that KCP2 is predominantly
incorporated into STT3A-containing complexes (Roboti and High,
2012). In contrast, loss of KCP2 appears to prevent the complete
assembly of OSTC500, resulting in the accumulation of an STT3Acontaining OST subcomplex (cf. Roboti and High, 2012; Shibatani
et al., 2005; Wang and Dobberstein, 1999). Since we do not have
good quality antibodies for all known mammalian OST subunits,
the discrete OSTC470 species observed upon KCP2 depletion may
also lack additional, as yet unidentified, components. Nevertheless,
these results suggest that the assembly of the major OSTC500
species detectable after BN-PAGE could arise via the recruitment
of KCP2 to a smaller subcomplex (OSTC470) that contains at least
STT3A, OST48 and DAD1, and most likely other known subunits
of the mammalian OST.
Overall, our in vitro analysis of OST function indicates that
OST48, DAD1 and KCP2 facilitate N-glycosylation in a
precursor-specific manner. The comparatively mild phenotype
observed in OST subunit-depleted cells (cf. tunicamycin
treatment) most likely reflects residual OST activity after
siRNA treatment. We typically observed more severe Nglycosylation defects in OST48- and DAD1-depleted cells than
in cells treated with the STT3A or STT3B siRNA alone. These
Journal of Cell Science
3482
Journal of Cell Science 125 (14)
data are consistent with a model where the loss of these two noncatalytic subunits destabilises both STT3A- and STT3Bcontaining OST complexes and thereby more effectively
inhibits N-glycosylation (Fig. 6). Interestingly, AQP1 proved to
be the most sensitive substrate for perturbations of OST-mediated
protein N-glycosylation using our in vitro assay. This most likely
reflects the fact that the glycosylation of Asn-42 of AQP1 is
complex, and relies on the initial translocation of TM2 into the
ER lumen prior to the subsequent topological maturation of the
precursor into its final structure (Buck et al., 2007; Foster et al.,
2000). Our studies of protein N-glycosylation using a cell-based
system bear out this model. Thus, the N-glycosylation of both
FVII and pSAP was quantitatively less efficient after knocking
down OST48 than STT3A, whilst knocking down STT3B had no
detectable effect (cf. Ruiz-Canada et al., 2009). Furthermore, the
knockdown of KCP2 resulted in a mild yet specific perturbation
of pSAP N-glycosylation, confirming the effect of KCP2
depletion we observed in vitro and showing that KCP2 is a
bone fide functional component of the mammalian OST.
Although a recent study concluded that KCP2 is dispensable
for mammalian OST activity, only one authentic protein substrate
was analysed in vitro (Wilson et al., 2011). We now provide the
first evidence that KCP2 levels influence protein N-glycosylation
in a substrate-specific manner, with the most striking effect
observed upon the biogenesis of endogenous pSAP in HeLa cells.
KCP2 depletion is the only treatment that results in an altered
OSTC species rather than a loss of detectable OST complexes. On
this basis, we conclude that the majority of pSAP N-glycosylation
occurs via the STT3A-containing OSTC470 species following
KCP2 knockdown, and speculate that although these KCP2depleted OSTs are suboptimal, nascent pSAP polypeptides are still
modified here in preference over the alternative STT3B-containing
OSTC550 species (Fig. 6). Hence, when we co-deplete KCP2 and
STT3B we see no additional defect in pSAP N-glycosylation.
In contrast, when we co-deplete STT3A and STT3B, the
hypoglycosylation phenotype of pSAP is comparable to that of
depleting OST48 alone (cf. Fig. 6). In contrast to the role of other
non-catalytic OST subunits (Schulz et al., 2009; Wilson et al.,
2008), the precise contribution of KCP2 during protein Nglycosylation remains to be established. However, given KCP2
appears to function primarily, if not exclusively, in combination
with the STT3A isoform (Roboti and High, 2012; see also this
study), it seems likely that the role of KCP2 relates to the
co-translational modification of nascent polypeptide chains.
Furthermore, the most significant defect resulting from KCP2
depletion appears to be an increase in forms of pSAP lacking one of
the five N-linked glycans normally attached (i.e. less pSAP-5 and
more pSAP-4; see Fig. 4D); a phenotype that is quite distinct from
that of the other siRNA depletions. On this basis, we speculate that
KCP2 may preferentially facilitate the N-glycosylation of a
particular N-sequon within the nascent pSAP chain (cf. Fig. 4C).
Alternatively, the alteration in pSAP N-glycosylation we observe
upon KCP2 depletion may simply reflect a stochastic reduction in
the occupancy of all five potential sites averaged across the whole
population of polypeptides synthesised during our analysis. At an
organismal level, our study is consistent with the suggestion that
there may be some correlation between increases in the complexity
of OST subunit composition and the diversity of protein substrates
and N-sequons that the resulting complex is capable of modifying
(Kowarik et al., 2006b). Indeed, higher eukaryotes seem to have
acquired additional non-catalytic subunits, as exemplified by
KCP2, that are over and above known and predicted homologues
of the well-defined yeast OST subunits (Kelleher and Gilmore,
2006; Shibatani et al., 2005). Conversely, various kinetoplastid
single-component OSTs display the capacity to N-glycosylate an
array of substrates (Castro et al., 2006; Izquierdo et al., 2009;
Kelleher and Gilmore, 2006; Nasab et al., 2008). Thus, the idea that
OST complexity directly reflects substrate complexity may be over
simplistic, and non-catalytic OST subunits may perform other roles
that remain to be delineated.
Materials and Methods
Cell culture and transfection
All siRNAs used in this study were custom synthesised by QIAGEN and their
sense strand sequences are in Table S1 in the supplementary material. QIAGEN
AllStars non-targeting siRNA was used as a negative control. HeLa and HepG2
cells were grown in DMEM supplemented with 10% FBS and 2 mM L-glutamine
at 37 ˚C under 5% CO2. Approximately 2.66105 HeLa cells were seeded in 10 cm
dishes 24 h before transfection with siRNA duplexes (supplementary material
Table S1) using Oligofectamine (Invitrogen) according to the manufacturer’s
instructions. After 7 h, the transfection medium was replaced by complete growth
medium and the cells were typically assayed 72 h post-transfection. For analysis of
FVII N-glycosylation, cells seeded in 6-well plates were transfected with a human
FVII expression plasmid (pCMV6-XL4-FVII; OriGene Technologies, Rockville,
MD, USA) 72 h after siRNA-treatment and were analysed 24 h later. For each
well, 2 mg of plasmid DNA was transfected with FuGENE HD (Roche) using the
manufacturer’s instructions. HepG2 cells were reverse-transfected with nontargeting or STT3A siRNAs (100 nM final concentration) using siPORT NeoFX
(Ambion) according to the manufacturer’s instructions and harvested after 72 h.
Immunoblotting
Cells were solubilised at 4 ˚C by a 1 h incubation with Triton X-100 lysis buffer
containing 10 mM Tris-HCl pH 7.6, 140 mM NaCl, 1 mM EDTA, 1% Triton X100, and a protease inhibitor cocktail. Cell extracts were clarified by centrifugation
and soluble proteins were quantified using the BCA Protein Assay kit (Pierce).
Proteins (,30 mg) were separated by SDS-PAGE and analysed by infrared
immunoblotting (Roboti and High, 2012). Alternatively, cells were directly
solubilised in SDS-sample buffer (0.1 M Tris-HCl pH 6.8, 5 mM EDTA, 0.5 M
sucrose, 0.5% L-methionine, 0.01% Bromophenol Blue, 40 mM DTT, 3% SDS)
and equal volumes subjected to SDS-PAGE and infrared immunoblotting.
Antibodies against STT3A, STT3B, OST48, DAD1 and KCP2 have been
previously described (Roboti and High, 2012). Antisera recognising a-tubulin
and Sec61a were from Keith Gull (University of Oxford, UK) and Richard
Zimmermann (Saarland University, Germany), respectively. Fluorescent-labelled
secondary antibodies were purchased from LI-COR Biosciences (Cambridge, UK).
The fluorescent bands were visualised and quantified on the Odyssey Infrared
Imager (LI-COR Biosciences) using the software provided by the manufacturer.
Semi-quantitative RT-PCR analyses
After 72 h of siRNA-treatment, HeLa cells were harvested in Trizol (Invitrogen)
and total RNA was extracted according to the manufacturer’s instructions. RNA
(1 mg) was reverse-transcribed into single-stranded cDNA using the AMV-RT
system (Roche). One-eighth and one-sixteenth of the synthesised cDNA was used
as template for PCR amplification of XBP1 and OST48 or DAD1 mRNAs,
respectively. A standard PCR protocol using Taq polymerase and human specific
primers for Xbp1 (forward: 59-ACAGCGCTTGGGGATGGATG-39; reverse:
59-TGACTGGGTCCAAGTTGTCC-39), OST48 (forward: 59-CACCTTAGTGCTGCTGGACA-39; reverse: 59-CGAAGGGGAGAAAATGATGA-39), DAD1 (forward:
59-GTTATGTCGGCGTCGGTAGT-39; reverse: 59-ACAAGAGATGAAGCCCGAGA-39) was followed. Parallel amplifications with primers for GAPDH cDNA
(forward: 59-GACCCCTTCATTGACCTCAACTACATG-39; reverse: 59-GTCCACCACCCTGTTGCTGTAGCC-39) were performed. The numbers of cycles
(between 25 and 31) giving the linear range were adjusted depending on the genes
amplified. PCR products were resolved on 2% agarose gels and visualised on a UV
transilluminator. Band intensities were quantified AIDA v3.52 image-analyser
software (Raytest).
BN-PAGE
BN-PAGE was based on the protocol of Wittig and colleagues (Wittig et al.,
2006). Twenty passes through a ball-bearing homogeniser (10 mm clearance) were
used to homogenise ,25 mg of sedimented siRNA-treated HeLa cells in 500 ml
sucrose buffer (83 mM sucrose, 6.6 mM imidazole-HCl pH 7.0). The pellet
obtained from 80 mg wet weight homogenised cells was solubilised in 35 ml
solubilisation buffer (50 mM NaCl, 2 mM 6-aminohexanoic acid, 1 mM EDTA,
50 mM imidazole-HCl pH 7.0) containing 1% digitonin for 20 min at 4 ˚C.
Insoluble material was removed by centrifugation at 16,000 g for 10 min at 4 ˚C,
Modulatory subunits of the OST
and the supernatant was mixed with 1/8 volume of 50% glycerol and 1/40 volume
of 5% (w/v) Coomassie Blue G250 (Serva) in 500 mM 6-aminohexanoic acid.
Samples were run on a 4–10% polyacrylamide gradient gel as described previously
(Roboti and High, 2012) and then analysed by immunoblotting.
Metabolic labelling and immunoprecipitation
Cells were starved in a methionine- and cysteine-free DMEM (GIBCO)
supplemented with 2 mM L-glutamine for 20 min at 37 ˚C, and then incubated
in fresh starvation medium containing 22 mCi/ml [35S]Met/Cys protein labelling
mix (PerkinElmer; specific activity .1,000 Ci/mmol) for 10 min. Tunicamycin
(Calbiochem) was added at a final concentration of 10 mg/ml ,20 h prior to the
starvation and was also included throughout the starvation and pulse. Cells were
washed twice with PBS and solubilised with Triton X-100 lysis buffer. Clarified
lysates were denatured with 1% SDS at 37 ˚C for 30 min, then diluted 5-fold with
Triton X-100 lysis buffer containing 8 mM cold Met/Cys and 1 mM PMSF, and
pre-cleared by incubation with pansorbin (Calbiochem) at 4 ˚C for 1 h. The precleared lysates were incubated overnight with antibodies specific for either saposin
D (Konrad Sandhoff, University of Bonn, Germany) or FVII (R&D Systems), and
immune complexes were captured using Protein-A-Sepharose beads (GenScript,
USA). Beads were washed three times with Triton X-100 buffer before eluting
proteins with SDS-PAGE sample buffer. Samples were denatured at 37 ˚C for 1 h,
resolved by SDS-PAGE, and analysed by phosphorimaging (FLA-3000; Fuji).
[35S]-labelled proteins were quantified using AIDA v3.52.
Journal of Cell Science
Endo H/PNGase F treatment of immunoprecipitated pSAP
Immunoprecipitated pSAP was digested with Endo H or PNGase F (New England
Biolabs) according to the manufacturer’s instructions. Briefly, the protein ASepharose antibody-antigen complexes were denatured in 20 ml of 16glycoprotein
denaturing buffer (0.5% SDS and 40 mM DTT) at 37 ˚C for 30 min. The denatured
samples were then digested with 500 units of Endo H in 40 ml of reaction buffer
(50 mM sodium citrate pH 5.5, 0.25% SDS, 20 mM DTT) or 500 units of PNGase
F in 40 ml of reaction buffer (50 mM sodium phosphate pH 7.5, 0.25% SDS,
20 mM DTT, 1% NP-40) for 30 min at 37 ˚C. The digestions were stopped by
adding SDS-PAGE sample buffer and the reaction products were separated by
SDS-PAGE and visualised by phosphorimaging.
In vitro glycosylation assay
siRNA-treated cells were semi-permeabilised (Wilson et al., 1995) and used as a
source of ER membranes for in vitro translation. Briefly, the plasma membrane
was permeabilised with KHM buffer (110 mM potassium acetate, 20 mM HEPES
pH 7.2, and 2 mM magnesium acetate) supplemented with 40 mg/ml digitonin
(Calbiochem) on ice for 5 min. Permeabilisation efficiency was confirmed by
staining cells with the Trypan Blue dye. Semi-permeabilised cells were treated
with HEPES buffer (50 mM potassium acetate and 90 mM HEPES pH 7.2) to
remove the cytosol. The membrane pellet fraction was treated with micrococcal
nuclease to remove endogenous RNA, and then resuspended in KHM buffer. Nglycoproteins were translated in a rabbit reticulocyte lysate system (Promega) for 1
h at 30 ˚C in the presence of 0.88 mCi/ml [35S]Met/Cys, an in vitro synthesised
mRNA transcript, and ,0.86105 semi-permeabilised cells. Then, 0.1 M
aurintricarboxylic acid was added to inhibit translation initiation and the
reactions were incubated at 30 ˚C for 10 min. Membrane-associated proteins
were isolated by centrifugation at 16,000 g for 5 min, washed with KHM buffer,
and resuspended in SDS-PAGE sample buffer before denaturation at 37 ˚C for 1 h.
Samples were resolved by SDS-PAGE and analysed by phosphorimaging.
Data and statistical analysis
Data are expressed as means 6 s.e.m., and two-tailed t-tests were performed to
determine P values for paired samples using Prism 4 (GraphPad).
Acknowledgements
We are indebted to Keith Gull, Konrad Sandhoff and Richard
Zimmermann for providing antibodies, and William Skach for
helpful discussions regarding the N-glycosylation of AQP1. We
thank Quentin Roebuck for the affinity purification of antibodies
against several OST subunits, and Pawel Leznicki and Lisa Swanton
for critical feedback during the preparation of this manuscript.
Funding
This work was supported by the Biotechnology and Biological Sciences
Research Council [grant number BB/E01979X/1 to P.R. and S.H.].
Supplementary material available online at
http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.103952/-/DC1
3483
References
Aebi, M., Bernasconi, R., Clerc, S. and Molinari, M. (2010). N-glycan structures:
recognition and processing in the ER. Trends Biochem. Sci. 35, 74-82.
Buck, T. M., Wagner, J., Grund, S. and Skach, W. R. (2007). A novel tripartite motif
involved in aquaporin topogenesis, monomer folding and tetramerization. Nat. Struct.
Mol. Biol. 14, 762-769.
Castro, O., Movsichoff, F. and Parodi, A. J. (2006). Preferential transfer of the
complete glycan is determined by the oligosaccharyltransferase complex and not by
the catalytic subunit. Proc. Natl. Acad. Sci. USA 103, 14756-14760.
Chavan, M., Chen, Z., Li, G., Schindelin, H., Lennarz, W. J. and Li, H. (2006).
Dimeric organization of the yeast oligosaccharyl transferase complex. Proc. Natl.
Acad. Sci. USA 103, 8947-8952.
Foster, W., Helm, A., Turnbull, I., Gulati, H., Yang, B., Verkman, A. S. and Skach,
W. R. (2000). Identification of sequence determinants that direct different
intracellular folding pathways for aquaporin-1 and aquaporin-4. J. Biol. Chem. 275,
34157-34165.
Harada, Y., Li, H., Li, H. and Lennarz, W. J. (2009). Oligosaccharyltransferase
directly binds to ribosome at a location near the translocon-binding site. Proc. Natl.
Acad. Sci. USA 106, 6945-6949.
Honma, K., Iwao-Koizumi, K., Takeshita, F., Yamamoto, Y., Yoshida, T., Nishio,
K., Nagahara, S., Kato, K. and Ochiya, T. (2008). RPN2 gene confers docetaxel
resistance in breast cancer. Nat. Med. 14, 939-948.
Iida, Y., Fujimori, T., Okawa, K., Nagata, K., Wada, I. and Hosokawa, N. (2011).
SEL1L protein critically determines the stability of the HRD1-SEL1L endoplasmic
reticulum-associated degradation (ERAD) complex to optimize the degradation
kinetics of ERAD substrates. J. Biol. Chem. 286, 16929-16939.
Izquierdo, L., Schulz, B. L., Rodrigues, J. A., Güther, M. L., Procter, J. B., Barton,
G. J., Aebi, M. and Ferguson, M. A. (2009). Distinct donor and acceptor
specificities of Trypanosoma brucei oligosaccharyltransferases. EMBO J. 28, 26502661.
Karaoglu, D., Kelleher, D. J. and Gilmore, R. (1997). The highly conserved Stt3
protein is a subunit of the yeast oligosaccharyltransferase and forms a subcomplex
with Ost3p and Ost4p. J. Biol. Chem. 272, 32513-32520.
Kelleher, D. J. and Gilmore, R. (1997). DAD1, the defender against apoptotic cell
death, is a subunit of the mammalian oligosaccharyltransferase. Proc. Natl. Acad. Sci.
USA 94, 4994-4999.
Kelleher, D. J. and Gilmore, R. (2006). An evolving view of the eukaryotic
oligosaccharyltransferase. Glycobiology 16, 47R-62R.
Kelleher, D. J., Karaoglu, D., Mandon, E. C. and Gilmore, R. (2003).
Oligosaccharyltransferase isoforms that contain different catalytic STT3 subunits
have distinct enzymatic properties. Mol. Cell 12, 101-111.
Kelleher, D. J., Banerjee, S., Cura, A. J., Samuelson, J. and Gilmore, R. (2007).
Dolichol-linked oligosaccharide selection by the oligosaccharyltransferase in protist
and fungal organisms. J. Cell Biol. 177, 29-37.
Kowarik, M., Numao, S., Feldman, M. F., Schulz, B. L., Callewaert, N., Kiermaier,
E., Catrein, I. and Aebi, M. (2006a). N-linked glycosylation of folded proteins by
the bacterial oligosaccharyltransferase. Science 314, 1148-1150.
Kowarik, M., Young, N. M., Numao, S., Schulz, B. L., Hug, I., Callewaert, N., Mills,
D. C., Watson, D. C., Hernandez, M., Kelly, J. F. et al. (2006b). Definition of the
bacterial N-glycosylation site consensus sequence. EMBO J. 25, 1957-1966.
Landolt-Marticorena, C. and Reithmeier, R. A. (1994). Asparagine-linked oligosaccharides
are localized to single extracytosolic segments in multi-span membrane glycoproteins.
Biochem. J. 302, 253-260.
Lizak, C., Gerber, S., Numao, S., Aebi, M. and Locher, K. P. (2011). X-ray structure
of a bacterial oligosaccharyltransferase. Nature 474, 350-355.
Mick, D. U., Vukotic, M., Piechura, H., Meyer, H. E., Warscheid, B., Deckers, M.
and Rehling, P. (2010). Coa3 and Cox14 are essential for negative feedback
regulation of COX1 translation in mitochondria. J. Cell Biol. 191, 141-154.
Nasab, F. P., Schulz, B. L., Gamarro, F., Parodi, A. J. and Aebi, M. (2008). All in
one: Leishmania major STT3 proteins substitute for the whole oligosaccharyltransferase complex in Saccharomyces cerevisiae. Mol. Biol. Cell 19, 3758-3768.
Nilsson, I. M. and von Heijne, G. (1993). Determination of the distance between the
oligosaccharyltransferase active site and the endoplasmic reticulum membrane. J.
Biol. Chem. 268, 5798-5801.
Nilsson, I., Kelleher, D. J., Miao, Y., Shao, Y., Kreibich, G., Gilmore, R., von
Heijne, G. and Johnson, A. E. (2003). Photocross-linking of nascent chains to the
STT3 subunit of the oligosaccharyltransferase complex. J. Cell Biol. 161, 715-725.
Ohtsubo, K. and Marth, J. D. (2006). Glycosylation in cellular mechanisms of health
and disease. Cell 126, 855-867.
Popov, M., Tam, L. Y., Li, J. and Reithmeier, R. A. (1997). Mapping the ends of
transmembrane segments in a polytopic membrane protein. Scanning N-glycosylation
mutagenesis of extracytosolic loops in the anion exchanger, band 3. J. Biol. Chem.
272, 18325-18332.
Roboti, P. and High, S. (2012). Keratinocyte-associated protein 2 is a bona fide subunit
of the mammalian oligosaccharyltransferase. J. Cell Sci. 125, 220-232.
Ron, D. and Walter, P. (2007). Signal integration in the endoplasmic reticulum
unfolded protein response. Nat. Rev. Mol. Cell Biol. 8, 519-529.
Ruiz-Canada, C., Kelleher, D. J. and Gilmore, R. (2009). Cotranslational and
posttranslational N-glycosylation of polypeptides by distinct mammalian OST
isoforms. Cell 136, 272-283.
Sanjay, A., Fu, J. and Kreibich, G. (1998). DAD1 is required for the function and the
structural integrity of the oligosaccharyltransferase complex. J. Biol. Chem. 273,
26094-26099.
3484
Journal of Cell Science 125 (14)
Journal of Cell Science
Schulz, B. L., Stirnimann, C. U., Grimshaw, J. P., Brozzo, M. S., Fritsch, F.,
Mohorko, E., Capitani, G., Glockshuber, R., Grütter, M. G. and Aebi, M. (2009).
Oxidoreductase activity of oligosaccharyltransferase subunits Ost3p and Ost6p defines
site-specific glycosylation efficiency. Proc. Natl. Acad. Sci. USA 106, 11061-11066.
Shibatani, T., David, L. L., McCormack, A. L., Frueh, K. and Skach, W. R. (2005).
Proteomic analysis of mammalian oligosaccharyltransferase reveals multiple
subcomplexes that contain Sec61, TRAP, and two potential new subunits.
Biochemistry 44, 5982-5992.
Silberstein, S., Collins, P. G., Kelleher, D. J. and Gilmore, R. (1995). The essential
OST2 gene encodes the 16-kD subunit of the yeast oligosaccharyltransferase, a highly
conserved protein expressed in diverse eukaryotic organisms. J. Cell Biol. 131, 371-383.
Skach, W. R., Shi, L. B., Calayag, M. C., Frigeri, A., Lingappa, V. R. and Verkman,
A. S. (1994). Biogenesis and transmembrane topology of the CHIP28 water channel at
the endoplasmic reticulum. J. Cell Biol. 125, 803-815.
te Heesen, S., Janetzky, B., Lehle, L. and Aebi, M. (1992). The yeast WBP1 is
essential for oligosaccharyl transferase activity in vivo and in vitro. EMBO J. 11,
2071-2075.
Wang, L. and Dobberstein, B. (1999). Oligomeric complexes involved in translocation of
proteins across the membrane of the endoplasmic reticulum. FEBS Lett. 457, 316-322.
Wilson, C. M. and High, S. (2007). Ribophorin I acts as a substrate-specific facilitator
of N-glycosylation. J. Cell Sci. 120, 648-657.
Wilson, C. M., Roebuck, Q. and High, S. (2008). Ribophorin I regulates substrate
delivery to the oligosaccharyltransferase core. Proc. Natl. Acad. Sci. USA 105, 95349539.
Wilson, C. M., Magnaudeix, A., Yardin, C. and Terro, F. (2011). DC2 and
keratinocyte-associated protein 2 (KCP2), subunits of the oligosaccharyltransferase
complex, are regulators of the gamma-secretase-directed processing of amyloid
precursor protein (APP). J. Biol. Chem. 286, 31080-31091.
Wilson, R., Allen, A. J., Oliver, J., Brookman, J. L., High, S. and Bulleid, N. J.
(1995). The translocation, folding, assembly and redox-dependent degradation of
secretory and membrane proteins in semi-permeabilized mammalian cells. Biochem.
J. 307, 679-687.
Wittig, I., Braun, H. P. and Schägger, H. (2006). Blue native PAGE. Nat. Protoc. 1,
418-428.
Yan, A. and Lennarz, W. J. (2005). Unraveling the mechanism of protein Nglycosylation. J. Biol. Chem. 280, 3121-3124.
Yan, A., Wu, E. and Lennarz, W. J. (2005). Studies of yeast oligosaccharyl transferase
subunits using the split-ubiquitin system: topological features and in vivo interactions.
Proc. Natl. Acad. Sci. USA 102, 7121-7126.
Yan, Q. and Lennarz, W. J. (2002). Studies on the function of oligosaccharyl
transferase subunits. Stt3p is directly involved in the glycosylation process. J. Biol.
Chem. 277, 47692-47700.
Zelinger, L., Banin, E., Obolensky, A., Mizrahi-Meissonnier, L., Beryozkin, A.,
Bandah-Rozenfeld, D., Frenkel, S., Ben-Yosef, T., Merin, S., Schwartz, S. B. et al.
(2011). A missense mutation in DHDDS, encoding dehydrodolichyl diphosphate
synthase, is associated with autosomal-recessive retinitis pigmentosa in Ashkenazi
Jews. Am. J. Hum. Genet. 88, 207-215.