Báo cáo khoa học: N-Glycosylation is important for the correct intracellular localization of HFE and its ability to decrease cell surface transferrin binding pptx

N-Glycosylation is important for the correct intracellular
localization of HFE and its ability to decrease cell surface
transferrin binding
Lavinia Bhatt
1
, Claire Murphy
2
, Liam S.O’Driscoll
2
, Maria Carmo-Fonseca
3
, Mary W. McCaffrey
1
and John V. Fleming
2,3
1 Department of Biochemistry, Biosciences Institute, University College Cork, Ireland
2 Department of Biochemistry, School of Pharmacy and ABCRF, University College Cork, Ireland
3 Institute of Molecular Medicine, University of Lisbon, Portugal
Keywords
HFE; N-glycosylation; transferrin; transferrin
receptor 1; b2-microglobulin
Correspondence
J. V. Fleming, Department of Biochemistry
and School of Pharmacy, University College
Cork, Cork, Ireland
Fax: +353 21 4901656
Tel: +353 21 4901679
E-mail: j.fl
Note
L. Bhatt and C. Murphy contributed equally
to this work

(Received 8 February 2010, revised 14 May
2010, accepted 2 June 2010)
doi:10.1111/j.1742-4658.2010.07727.x
HFE is a type 1 transmembrane protein that becomes N-glycosylated dur-
ing transport to the cell membrane. It influences cellular iron concentra-
tions through multiple mechanisms, including regulation of transferrin
binding to transferrin receptors. The importance of glycosylation in HFE
localization and function has not yet been studied. Here we employed bio-
informatics to identify putative N-glycosylation sites at residues N110,
N130 and N234 of the human HFE protein, and used site-directed muta-
genesis to create combinations of single, double or triple mutants. Com-
pared with the wild-type protein, which co-localizes with the type 1
transferrin receptor in the endosomal recycling compartment and on dis-
tributed punctae, the triple mutant co-localized with BiP in the endoplas-
mic reticulum. This was similar to the localization pattern described
previously for the misfolding HFE-C282Y mutant that causes type 1 hered-
itary haemachromatosis. We also observed that the triple mutant was func-
tionally deficient in b2-microglobulin interactions and incapable of
regulating transferrin binding, once again, reminiscent of the HFE-C282Y
variant. Single and double mutants that undergo limited glycosylation
appeared to have a mixed phenotype, with characteristics primarily of the
wild-type, but also some from the glycosylation-deficient protein. There-
fore, although they displayed an endosomal recycling compartment/punc-
tate localization like the wild-type protein, many cells simultaneously
displayed additional reticular localization. Furthermore, although the
majority of cells expressing these single and double mutants showed
decreased surface binding of transferrin, a number appeared to have lost
this ability. We conclude that glycosylation is important for the normal
intracellular trafficking and functional activity of HFE.
Structured digital abstract

l
MINT-7896236, MINT-7896218: beta2M (uniprotkb:P61769) physically interacts (MI:0915)
with HFE (uniprotkb:
Q30201)byanti bait coimmunoprecipitation (MI:0006)
l
MINT-7896162: TfR1 (uniprotkb:P02786) and HFE (uniprotkb:Q30201) colocalize (MI:0403)
by fluorescence microscopy (
MI:0416)
Abbreviations
ER, endoplasmic reticulum; ERC, endosomal recycling compartment; HH, hereditary haemachromatosis; b2M, b2 microglobulin; MHC, major
histocompatability complex; PNGase F, N-glycosidase F; Tfn, transferrin; TfR1, transferrin receptor 1; TfR2, transferrin receptor 2.
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3219
Introduction
The hereditary haemochromatosis (HH) protein HFE
(high Fe) is a type 1 transmembrane protein that plays
an important role in controlling physiological iron
homeostasis [1–3]. It is widely expressed throughout
the body with expression highest in cells that are
involved in iron metabolism [4–6]. Mutations in the
HFE protein cause type 1 HH, which is an inherited
disease of iron metabolism that results in iron overload
in several organs [4,7]. The HFE mutation detected in
the majority of HH patients results in the replacement
of cysteine residue 282 with tyrosine (C282Y). The
mutant protein is unable to form a structurally impor-
tant disulfide bridge required for HFE interactions
with b2 microglobulin (b2M) [4,5,8–11]. In the absence
of b2M binding, the protein misfolds and is retained in
the endosplasmic reticulum (ER) where it induces an
unfolded protein stress response that is characterized

by alternative splicing of XBP-1 and increased expres-
sion of CHOP and BiP [12–14]. A second well-
described HFE mutation associated with HH leads to
the replacement of histidine at residue 63 with aspar-
tate. This mutant is capable of b2M interaction and
cell-surface expression but is unable to regulate cellular
iron uptake like the wild-type HFE protein [4,15].
Although much insight into HFE function has been
gained through studying the cellular and biochemical
properties of these different mutant proteins, the exact
mechanism by which HFE regulates intracellular iron
levels is still not completely understood.
The HFE primary sequence exhibits significant
homology to major histocompatability complex
(MHC) class I molecules and the protein is organized
into a1, a2 and a3 structural domains that resemble
those described for MHC class I and related proteins
[4,16]. The N-terminal a1 and a2 domains come
together to form a superstructure composed of two
a helices layered on top of eight anti-parallel b sheets.
In MHC class I proteins this a1/a2 superstructure
forms a peptide-binding groove that mediates antigen
binding and presentation to CD8
+
cytolytic T cells.
In HFE, the proximity of the two a helices and the
presence of amino acid side chains that project into
the groove appear to prevent peptide binding [16]. The
a3 region, like its homologous domain in MHC
class I, is an immunoglobulin-like domain that medi-

ates binding to b2M [16,17]. C-Terminal residues of
HFE mediate its retention in the cell membrane.
Shortly after HFE was discovered it was reported to
co-localize and interact with the type 1 transferrin
receptor (TfR1) [5,18]. TfR1 mediates the endocytosis
of iron-loaded transferrin into acidic endosomes where
the iron is released and transported into the cytoplasm
via the Nramp2-DCT1 iron transporter. Apo-transfer-
rin and TfR1 are recycled to the cell surface where
apo-transferrin is released [3,19]. Crystallography stud-
ies suggest that the a3 stem of HFE lies parallel to the
cell membrane and that the a1/a2 superstructure inter-
acts with helical regions located within TfR1. In this
way, it is possible for two HFE proteins to be posi-
tioned at either side of the TfR1 homodimer and form
a tetrameric complex that exhibits twofold symmetry
[17]. Reports from crystallography experiments have
been supported by mutagenesis studies that identified
residues located at the end of an a-helical region of the
HFE a1 domain (V100 and W103A) as being of par-
ticular importance for TfR1 interactions [16,17,20].
The effect of HFE binding to TfR1 is to lower the
affinity of the receptor for transferrin [15]. This most
likely reflects the existence of overlapping HFE and
transferrin-binding sites on the receptor [21,22]. Suc-
cessive studies indicate that HFE and TfR1 co-localize
during endosomal trafficking, although there are con-
tradictory reports as to whether TfR1 recycling is
affected by HFE [23–29].
Despite these well-described interactions, there is

mounting evidence that HFE regulation of cellular
iron levels may not depend solely on TfR1 binding
[30,31]. Attention has shifted to a second transferrin
receptor, TfR2, whose pattern of expression is more
restricted than that of ubiquitously expressed TfR1
[32]. Levels of TfR2 are highest in hepatocytes, the
predominant site of HFE expression, and recent stud-
ies have confirmed that the two proteins are capable of
interacting [33,34]. The nature of these interactions dif-
fers from those observed between HFE and TfR1 in
that they are mediated by the a3 domain of HFE, as
opposed to the a1/a2 superstructure [33]. An emerging
model, therefore, is that TfR2 competes with TfR1 for
l
MINT-7896258, MINT-7896317, MINT-7896330, MINT-7896348, MINT-7896366: HFE (uni
protkb:
Q30201) and transferrin (uniprotkb:P02787) colocalize (MI:0403)byfluorescence
microscopy (
MI:0416)
l
MINT-7896149: HFE (uniprotkb:Q30201) and BiP (uniprotkb:P11021) colocalize (MI:0403)
by fluorescence microscopy (
MI:0416)
N-Glycosylation of HFE L. Bhatt et al.
3220 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
HFE binding. This occurs maximally at high concen-
trations of transferrin. The resulting HFE–TfR2 com-
plex, which is stabilized at high iron concentrations, is
believed to somehow regulate the expression of other
genes involved in iron metabolism. This includes hepci-

din, a 25 amino acid antimicrobial peptide that is
expressed in liver cells and is now recognized as a key
regulator of iron homeostasis in the body. Hepatocel-
lular hepcidin mRNA levels have been shown to be
regulated by HFE, and are altered in haemachromato-
sis patients with the C282Y mutation [35–37].
The importance of N-glycosylation with respect to
protein expression and function is highly variable.
Roles have been described in the secretion, stability
and oligomerization of proteins [38,39], the bioactivi-
ties of enzymes [40] and the binding affinities of
ligands and receptors [41]. In many instances, specific
functions can be attributed to glycosylation at specific
sites. For example, the human gonadotropin a subunit
has N-glycosylation sites at residues Asn52 and Asn78
that have been shown to differentially regulate receptor
signalling and secretion, respectively [38,39]. Another
example is the type 1 transferrin receptor, which has
N-glycosylation sites at residues Asn251, Asn317 and
Asn727. Mutation of Asn727 decreases cell-surface
expression, whereas mutation at the other two sites
does not [42].
HFE becomes glycosylated during post-translational
processing. Transfection studies have confirmed that
this involves N-glycosylation, and incubation of lysates
from HFE-expressing cells with N-glycosidase F
(PNGase F) leads to the accumulation of lower molec-
ular mass HFE proteins [13,43]. The carbohydrate
moiety undergoes processing and endoglycosi-
dase H-resistant HFE isoforms can be detected by

30 min post translation [10,13,18,23]. Although these
studies demonstrate that HFE is glycosylated, the
specific role, if any, that glycosylation might play in
cellular HFE function has not previously been studied.
In this article, we map the sites of HFE N-glycosyla-
tion and examine the importance of glycosylation on
parameters of protein localization and function.
Results
Tunicamycin treatment results in a reticular
pattern of HFE localization
Previous studies have demonstrated that HFE under-
goes post-translational N-glycosylation. As a first step
towards assessing the importance of N-glycosylation
on HFE expression, we transiently transfected HuTu80
to express HFE-WT–HA and cultured the cells in the
presence or absence of tunicamycin to inhibit glycan
production. Control, untreated cells predominantly
exhibited a punctate pattern of HFE expression with a
tubulovesicular concentration in the pericentrosomal
region (Fig. 1A,D), consistent with previous observa-
tions [11,28]. Immunostaining with anti-TfR1, anti-
Rab11a and Rab11-FIP3 Ig has identified the HFE-
containing pericentrosomal compartment of HuTu80
cells as the endosomal recycling compartment (ERC)
[28]. Treatment of HFE-WT–HA-expressing cells with
tunicamycin altered this pattern of localization and
resulted in a reticular pattern of cell localization
(Fig. 1B,D). Immunostaining showed significant
co-localization with the ER chaperone protein BiP,
demonstrating that the HFE-WT–HA was now locali-

zing primarily to the ER (Fig. 1B,D).
A similar pattern of reticular expression and BiP
co-localization was observed when HuTu80 cells were
transfected to express the HH-causing HFE-C282Y
variant (Fig. 1C,D), which has been shown through
multiple biochemical and microscopy approaches to be
retained in the ER [10,13,18,28,29].
HFE is glycosylated at residues Asn110, Asn130
and Asn234
Although the results in Fig. 1 point towards an impor-
tant role for glycosylation in HFE localization, it
remains possible that the effects of tunicamycin treat-
ment were indirect. To directly examine the impor-
tance of glycosylation on HFE, it was necessary to
generate an N-glycosylation-deficient mutant. To this
end, we used a bioinformatic prediction program
(netnglyc 1.0 Server; Technical University of Den-
mark) to identify putative glycosylation sites in the
protein. Consistent with previous predictions [18], we
identified three high-probability sites: asparagines at
positions 110, 130 and 234. Starting with wild-type
HFE, we generated all possible combinations of single
and double putative N-glycosylation site mutants
using site-directed mutagenesis. The wild-type and
mutant expression constructs were transfected into
HEK293T cells and the lysates analysed by western
blotting.
The results from these experiments, which are shown
in Fig. 2A, indicate that the introduction of single ala-
nine mutations at N110, N130 and N234, respectively,

resulted in the production of HFE proteins that
migrated with increased mobility on SDS/PAGE com-
pared with the wild-type. This suggested that all three
sites in the wild-type proteins are capable of becoming
glycosylated. The decrease in apparent molecular mass
became even more pronounced for proteins containing
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3221
combinations of double mutants, which displayed a
lower apparent molecular mass than either wild-type
or single mutant forms of the protein.
Although immunoblot analysis demonstrated that
the single and double mutants had decreased mass
compared with the wild-type protein, they still
appeared to be of higher molecular mass than the
unglycosylated form of the wild-type HFE protein –
which was produced when wild-type-expressing cells
were treated with tunicamycin (Fig. 2A; WT-Tunica).
This suggested that both the single and double
mutants were still partially glycosylated. To test this,
we transfected HEK293T cells to express either the
wild-type or mutant proteins, and incubated the cells
in the presence or absence of tunicamycin. Drug
treatments resulted in the accumulation of forms of
the mutant proteins that were of lower apparent
molecular mass and of similar size to the unglycosy-
lated form of the wild-type protein (see Fig. 2B for
single mutants and Fig. 2C for double mutants). This
suggested that the mutant proteins do indeed still
undergo limited glycosylation. For the single

mutants, additional supporting evidence for the per-
sistence of N-linked glycans was obtained by PNG-
ase F digestions of immunoprecipitated proteins,
which then migrated with lower apparent molecular
mass compared with the undigested forms (results
not shown).
A
B
C
D
Fig. 1. Inhibition of N-glycosylation influ-
ences patterns of HFE intracellular localiza-
tion. HuTu80 cells were transfected with
constructs expressing the HFE-WT–HA
(A, B) and HFE-C282Y–HA (C) proteins,
HFE-WT-expressing cells were incubated for
1 h in the absence (A) or presence (1B) of
2 lgÆmL
–1
tunicamycin (Tunica) as indicated.
Cells were immunostained with anti-HA and
anti-BiP Ig, and processed for fluorescence
microscopy. Co-localization masks were
created as described in Materials and meth-
ods, and represent areas with overlapping
green and red pixels converted to white.
Scale bar, 10 lm. Identical results were
obtained when cells were transfected with
constructs directed to express amino-tagged
GFP–HFE-WT and GFP–HFE-C282Y, and in

general we found that HA and GFP tags
could be interchanged without altering the
pattern of cell localization (data not shown).
Figures shown are representative of at least
three independent experiments. (D) Graph
showing the relative amounts of transfected
cells exhibiting punctate or reticular localiza-
tion of expressed HFE proteins (n = 3).
N-Glycosylation of HFE L. Bhatt et al.
3222 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
HFE NNN110/130/234/AAA triple mutant is
glycosylation deficient
To investigate whether the three N-glycosylation sites
studied to date are the only sites of HFE N-glycosyla-
tion – and with the aim of producing an HFE mutant
that is completely deficient in N-glycosylation – we
used site-directed mutagenesis to create a NNN110/
130/234AAA triple mutant. To determine the effect of
these combined mutations on HFE, we transfected
HEK293T cells to express wild-type, single, double or
triple mutants.
Western blot analysis of cell lysates shown in Fig.3A
indicated that the triple mutant fractionated with a
lower molecular mass than the wild-type protein, and
either the single or double mutants. This lower molecu-
lar mass form appeared to be the same size as the
unglycosylated form of the wild-type protein produced
in tunicamycin-treated cultures (Fig. 3A; WT-Tunica).
These data suggested that all potential glycosylation
sites had been mutated. To confirm this, we transfected

HEK293T cells to express wild-type or triple mutant
forms of HFE and incubated the cells in the presence
and absence of tunicamycin. Drug treatment resulted
in the accumulation of an unglycosylated lower molec-
ular mass form of the wild-type HFE protein, whereas
it had no effect on the apparent molecular mass of the
triple mutant (Fig. 3B).
In a second approach, HFE was immunoprecipitated
from wild-type or triple-mutant-expressing cells and
incubated with PNGase-F. As is shown in Fig. 3C,
enzyme treatment of wild-type HFE resulted in the
production of a lower molecular mass product. Treat-
ment had no detectable effect on migration of the tri-
ple mutant, which had the same apparent molecular
mass as the PNGase F-treated wild-type protein. We
conclude that HFE is normally glycosylated in vivo at
three sites (N110, N130 and N234), and that mutation
of these sites gives rise to an HFE protein that is
N-glycosylation deficient.
N-Glycosylation of HFE is required for its
appropriate localization to the ERC
Our results from Fig. 1 indicated that tunicamycin
treatment of HFE-WT-expressing cells results in a
reticular localization pattern. To definitively establish
the importance of N-glycosylation on HFE localization
in HuTu80 cells, we transfected cells to express GFP-
tagged forms of wild-type or triple-mutant HFE. As
shown in Fig. 4A, HFE-WT–GFP localized predomi-
nantly to a tubulovesicular structure near the nucleus
with some punctate staining, similar to that observed

in Fig. 1A. The GFP-tagged triple mutant, by contrast,
predominantly displayed a reticular localization pat-
tern (Fig. 4A,B). This was similar to the expression
pattern previously observed for HFE-WT in tunica-
mycin-treated cells (Fig. 1B).
A
B
C
Fig. 2. Characterization of N-glycosylation site single and double
mutants. (A) HEK293T cells were transfected to transiently express
HFE-WT–HA, HFE-N110–HA, HFE-N130A–HA, HFE-N234A–HA, HFE-
NN110/130AA–HA, HFE-NN130/234AA–HA or HFE-NN234/110AA–
HA proteins. HFE-WT–HA expressing cells were incubated for 16 h
before lysis in the presence or absence of 1 m
M tunicamycin (Tunica).
Cleared cell lysates were fractioned by 11% SDS/PAGE for immuno-
blotting with a mouse anti-HA Ig. (B) Transiently transfected HEK293T
cells expressing HFE-WT–HA, HFE-N110–HA, HFE-N130A–HA or
HFE-N234A–HA were incubated for 16 h in the presence or absence
of 1 m
M tunicamycin as indicated (Tunica). HA-tagged proteins in
cleared cells lysates were detected by immunoblotting with a mouse
anti-HA Ig. (C) Transiently transfected HEK293T cells expressing HFE-
WT–HA, HFE-NN110/130AA–HA, HFE-NN130/234AA–HA or HFE-
NN234/110AA–HA protein were incubated for 16 h in the presence or
absence of 1 m
M tunicamycin as indicated (Tunica). HA-tagged pro-
teins in cleared cells lysates were detected by immunoblotting with a
mouse anti-HA Ig.
L. Bhatt et al. N-Glycosylation of HFE

FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3223
HuTu80 cells transfected with HFE-WT or HFE tri-
ple-mutant proteins were subsequently immunostained
with an anti-TfR1 Ig. The results from these
experiments demonstrated that the wild-type protein
co-localizes with TfR1 predominantly in the tubulove-
sicular perinuclear ERC and discrete punctae
(hereafter referred to as ERC/punctate pattern of
localization). The triple mutant, by contrast, shows no
TfR1 co-localization (Fig. 4A).
Although these results attest to the importance of
N-glycosylation for the normal cellular localization of
HFE, we wondered whether this was a cumulative
effect or whether there are specific glycosylation sites
that are more important than others for ensuring
expression and recycling of the protein. To test this,
we transfected HuTu80 cells to express HFE-N110A–
GFP, HFE-N130A–GFP or HFE-N234A–GFP pro-
teins. In all instances, we observed that the majority of
transfected cells displayed an ERC/punctate pattern of
localization, similar to the wild-type protein
(Fig. 4A,B). Interestingly, a significant number of cells
displaying this ERC/punctate pattern simultaneously
displayed reticular localization in the same cells
(50 ± 2% of N110A-expressing cells, 50 ± 2% of
N130A-expressing cells and 58 ± 5% of N234A-
expressing cells, n = 3). This was a feature also of
cells expressing HFE double mutants, in which we like-
wise observed an ERC/punctate localization pattern in
the majority of cells (Fig. 5A,B) and a significant num-

ber of these cells simultaneously displaying reticular
localization (55 ± 7% of NN110/130AA-expressing
cells, 51 ± 4% of NN130/234AA-expressing cells and
67 ± 7% of NN234/110AA-expressing cells, n = 4).
This type of mixed phenotype, where reticular locali-
zation was observed in cells that already had the cor-
rect ERC/punctate pattern, was not observed to any
significant degree in cells expressing the wild-type or
triple-mutant proteins.
N-Glycosylation is important for interactions with
b2M
The wild-type HFE protein interacts with b2M during
transport to the cell surface. It is commonly reported
that misfolding and ER retention of the HFE-C282Y
variant happens specifically because this b2M interac-
tion does not occur [5,9–11]. The HFE triple mutant,
just like the HFE-C282Y mutant, shows a reticular
pattern of localization and in immunostaining studies
was seen to co-localize in the ER with BiP (Fig. 6A).
Accordingly, we wondered whether this pattern of
localization reflected an underlying inability to interact
with b2M or whether, in addition to b2M binding, the
HFE protein needs to be appropriately glycosylated in
order to successfully transit the ER. HEK293T cells
were transfected to express wild-type or mutant forms
of the HFE protein. Cells were lysed and an anti-b2M
Ig used to immunoprecipitate b2M and any interacting
proteins. As shown in Fig. 6B, the triple mutant
showed significantly decreased interactions with b2M.
A

B
C
Fig. 3. Characterization of N-glycosylation site triple mutant. (A)
HEK293T cells were transiently transfected with constructs
expressing HFE-WT–HA, HFE-N110–HA, HFE-N130A–HA,
HFE-N234A–HA, HFE-NN110/130AA–HA, HFE-NN130/234AA–HA,
HFE-NN234/110AA–HA or HFE-NNN110/130/234–HA (Triple) pro-
teins. HFE-WT–HA expressing cells were incubated for 16 h before
lysis in the presence or absence of 1 m
M tunicamycin (Tunica).
HA-tagged proteins were detected by fractionation of cleared cell
lysates on 11% SDS/PAGE for immunoblotting with a mouse anti-
HA Ig. (B) Transiently transfected HEK293T cells expressing
HFE-WT–HA or HFE-NNN110/130/234–HA (Triple) proteins were
incubated for 16 h in the presence or absence of 1 m
M tunicamycin
(Tunica) as indicated. (C) HFE-WT–HA or HFE-NNN110/130/
234AAA–HA proteins were transiently expressed in HEK293T cells.
Forty-eight hours after transfection cells were harvested and HA-
tagged proteins were immunprecipitated with a polyclonal rabbit
anti-HA Ig. Immunoprecipitated proteins were digested with PNG-
ase F and fractionated by 11% SDS/PAGE for immunoblotting with
a mouse anti-HA Ig.
N-Glycosylation of HFE L. Bhatt et al.
3224 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
This deficiency could not be attributed to a specific
glycosylation site, because each of the single mutants
retained the ability to interact with b2M. The HFE-
C282Y mutant was employed in these experiments as a
negative control (Fig. 6B).

Combined with our earlier cell localization results,
the data presented in Fig. 6 point towards an
important role for N-glycosylation in HFE folding.
However, once again, it is only when all three sites
are mutated that we observe a significant loss of
function.
N-Glycosylation is important for HFE regulation
of transferrin binding
Previous reports have established that in certain cell
types HFE acts to regulate intracellular iron levels by
decreasing the binding of transferrin to transferrin
receptors and reducing cellular iron uptake as a conse-
quence. Indeed, it is commonly believed that iron over-
load in HH occurs because the misfolding HFE-
C282Y variant fails to make it to the cell surface and
is unable to exert this control. We wanted to determine
A
B
Fig. 4. Intracellular localization of HFE triple and single N-glycosylation site mutants. (A) HuTu80 cells were transfected to express HFE-WT–
GFP, HFE-N110A–GFP, HFE-N130A–GFP, HFE-N234A–GFP or HFE-NNN110/130/234AAA–GFP triple mutant. Sixteen to eighteen hours post
transfection the cells were immunostained with an anti-TfR1 Ig, and processed for fluorescence microscopy. Scale bar, 10 lm. (B) Graph
showing the relative amounts of transfected cells exhibiting ERC/punctate or reticular localization of expressed HFE proteins (n = 3).
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3225
whether N-glycosylation is important for HFE regula-
tion of transferrin uptake. To this end, transfected
HuTu80 cells expressing HA-tagged forms of wild-type
or mutant HFEs were analysed for their ability to bind
fluorescently labelled transferrin. Consistent with previ-
ous reports, we noted that cells expressing wild-type

HFE displayed a striking decrease in cell-surface bind-
ing of transferrin, and that little or no reduction in
transferrin binding was observed in cells expressing the
HFE-C282Y mutant (Fig. 7A).
A
B
Fig. 5. Intracellular localization of HFE
double N-glycosylation site mutants. (A)
HuTu80 cells were transfected to express
HFE-NN110/130AA–HA, HFE-NN130/
234AA–HA or HFE-NN234/110AA–HA. Six-
teen to eighteen hours post transfection the
cells were immunostained with an anti-TfR1
Ig, and processed for fluorescence micros-
copy. Scale bar, 10 lm. (B) Graph showing
the relative amounts of transfected cells
exhibiting ERC/punctate or reticular localiza-
tion of expressed HFE proteins (n = 4).
N-Glycosylation of HFE L. Bhatt et al.
3226 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
In experiments to compare the status of our glyco-
sylation mutants in this assay we observed that the tri-
ple mutant displayed a phenotype indistinguishable
from the C282Y mutant, with an almost complete loss
of the ability to regulate transferrin binding (Fig. 7B).
By contrast, cells transfected with either the single or
double mutants were all capable of reducing transfer-
rin binding. In each case, however, there tended to be
a decrease in the proportion of cells that retained
this ability compared with the wild-type protein

(Fig. 7B–D). This effect was strongest for the 110/130
double mutant (Fig. 7D).
Discussion
Although N-glycosylation can dramatically alter the
structure and function of many proteins, there are also
cited instances in whch the mutation of glycosylation
sites has little or no effect [42,44]. The importance of
glycosylation is highly variable, therefore, and even in
cases where it is important, the effect may be either
direct or indirect. In this study, we set out to explore
the importance of HFE glycosylation. In doing so, we
aimed to expand on previous studies, which despite ref-
erence to glycan addition and the development of
endo-H resistance [13,18], nevertheless failed to identify
the role, if any, that glycosylation plays in HFE locali-
zation and function. We established for the first time
that the protein becomes N-glycosylated at asparagine
residues 110, 130 and 234, and that mutation of all
three sites results in the production of a protein that is
glycosylation deficient. Glycosylation at each of the
substrate asparagine residues can occur independently
of the glycosylation status at the other two sites.
A
B
Fig. 6. ER localization of the HFE-triple mutant and interaction of N-glycosylation site mutants with b2M. (A) HuTu80 cells transfected to
express HFE-triple mutant–HA were immunostained with anti-HA and anti-BiP Ig, and processed for fluorescence microscopy. Co-localization
masks were created as described in Materials and methods. Scale bar, 10 lm. (B) Constructs expressing HFE-WT–HA, HFE-NNN110/130/
234–HA (triple), HFE-N110A–HA, HFE-N130A–HA, HFE-N234A–HA or HFE-C282Y–HA proteins were transiently transfected into HEK293T
cells. Forty-eight hours after transfection cells were harvested. HA-tagged proteins in cleared cell lysates were fractionated on 11% SDS/
PAGE for immunoblotting with a mouse anti-HA Ig (upper). b2M and b2M-interacting proteins were immunoprecipitated using a rabbit anti-

b2M Ig and precipitated proteins were fractionated by SDS/PAGE (11%) for detection of HA-tagged proteins by immunoblot using a mouse
anti-HA Ig (lower). In a complementary series of experiments NiNTA–agarose was used to precipitate His-tagged versions of the WT, C282Y
and triple-mutant HFE proteins from transfected HEK293T cells. Immunoblots confirmed that only the wild-type HFE protein was capable of
co-precipitating significant quantities of b2M (data not shown).
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3227
A
B
Fig. 7. Regulation of transferrin binding by
N-glycosylation site mutants. HuTu80 were
transfected to express (A) HFE-WT–HA or
HFE-C282Y–HA, (B) triple mutant HFE-
NNN110/130/234–HA (triple) or single
mutants HFE-N110A–HA, HFE-N130A–HA
and HFE-N234A–HA or (C) double mutants
HFE-NN110/130AA–HA, HFE-NN234/110A–
HA or HFE-NN234/110AA–HA, as indicated.
Sixteen hours post transfection the cells
were serum starved for 2 h followed by
incubation with Alexa Fluor 594-bound Tfn
for 1 h at 4 °C. Cells were immunostained
with an anti-HA Ig and processed for fluo-
rescence microscopy. Scale bar, 10 lm. (D)
Graph showing the percentage of transfect-
ed cells with reduced transferrin binding in
response to the expression of various HFE
proteins as indicated. *P < 0.05, **P < 0.01
by Student’s unpaired t-test (n = 6).
N-Glycosylation of HFE L. Bhatt et al.
3228 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS

From the outset, we were interested in looking at
the importance of overall glycosylation patterns on
HFE intracellular trafficking, localization and func-
tion. Initially this was done in the context of inhibiting
cellular glycan production, with tunicamycin treatment
altering HFE localization so that we observed a reticu-
lar pattern of protein localization. Tunicamycin dis-
rupts the glycosylation status of many cellular
proteins, however, and it remained a possibility that
the observed changes were indirect. By mapping the
relevant asparagine residues, and disrupting the glyco-
sylation status specifically of HFE, it allowed us to
C
D
Fig. 7. (Continued).
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3229
establish the importance of HFE glycosylation and to
confirm ER localization in its absence. It is noteworthy
that just a single glycosylation site has been identified
in homologous MHC class I proteins, at a position
comparable with N110. In contrast to our HFE data,
where protein glycosylation is important for function,
studies with MHC class I proteins HLA-A2 and HLA-
B7 suggest that it is fully functional in the absence of
glycosylation [45]. Despite the sequence and structural
similarities between the two proteins, therefore, we
observed a clear divergence from MHC class I in this
regard.
Parameters of protein function can frequently be

attributed to glycosylation at specific sites. A good
example of this is TfR1, in which mutation of one of
the three glycosylation sites (Asn727) results in a retic-
ular rather than ERC pattern of localization [41,42].
HFE does not appear to be regulated in this way, and
despite the overall importance of glycosylation, our
results nevertheless argue against the existence of a sin-
gle glycosylation site that is of such absolute impor-
tance that it is capable in its own right of regulating
localization, b2M interactions or transferrin binding.
Instead, we observed a tendency towards a mixed phe-
notype for the partially glycosylated mutants. There-
fore, although cells expressing either the double or
single mutants showed incorporation into the ERC
and punctate patterns of cellular localization similar to
the wild-type protein, a significant number of transfect-
ed cells demonstrated both ERC/punctate and reticular
localization within the same cell. Furthermore, in cells
expressing either the double or single mutants, we
tended to see a decrease in the number of cells with
the ability to reduce transferrin binding compared with
the wild-type protein. It should be emphasized in all
instances, however, that the majority of single or
double mutant transfected cells continued to exhibit
ERC/punctate localization and regulate transferrin
binding, consistent with production and localization of
a functional protein.
Many studies have demonstrated the importance of
glycosylation in the folding and/or oligomerization of
trafficking proteins, with underglycosylated proteins

frequently becoming misfolded and exhibiting a reticu-
lar pattern of cellular localization [41,46–48]. Evidence
to suggest that N-glycosylation is important also for
HFE folding can be observed in many of our experi-
ments. The pattern of ER localization observed for the
triple mutant, or following tunicamycin treatment of
wild-type expressing cells, is similar to that detected
for the misfolding HFE-C282Y variant. The triple
mutant, like HFE-C282Y, also demonstrated a marked
decrease in interactions with b2M, which is believed to
be important for HFE folding and stabilization [4,5,8–
11]. Our combined biochemical and microscopy studies
reveal a consistent picture linking the cell biology of
the HFE glycosylation-deficient mutant with that of
the HFE-C282Y protein that is known to misfold and
induce a cellular unfolded protein response. We pro-
pose that glycosylation is important for the folding of
HFE and is essential for transport and exit of the pro-
tein from the ER. The importance of glycosylation is
cumulative, however, with all three glycosylation sites
requiring mutagenesis before an absolute effect was
observed. Therefore, although glycan addition at just a
single site – no matter which one – was sufficient to
ensure production of protein that could function in the
regulation of transferrin binding, our results likely
reflect the fact that the efficiency of folding is compro-
mised by underglycosylation, affecting some but not
all proteins as they transit the ER. This would in part
explain the observed intermediate phenotype for par-
tially glycosylated proteins.

As might be expected from the ER-localized and
potentially misfolding triple mutant, we observed little
or no regulation of transferrin binding in cells express-
ing glycosylation-deficient HFE. It remains unclear,
however, whether this deficiency is occurring solely as
a result of protein mislocalization or whether there
might be glycosylation events that are of specific rele-
vance for this parameter of HFE function. The results
for the 110/130 double mutant were particularly inter-
esting, not only was there a decrease in the number of
transfected cells that could regulate transferrin binding
compared with the wild-type, but there was also a sig-
nificant decrease compared with all other single and
double mutants. Current understanding suggests that
the majority of cell types regulate cellular iron levels
by binding of transferrin to the type 1 transferrin
receptor [15]. The crystal structure of HFE complexed
to the TfR1 has been determined and although the
two proteins interact over a relatively large interface, it
has been shown that residues V100 and W103 (V78
and W81 of the mature protein) are of particular
importance [16,17,21,22]. Additional missense muta-
tions at residues I105 and G93 in this region have also
been implicated in the disruption of iron metabolism
[49]. In structural terms, this corresponds to a helical
and loop region in the a1 domain of HFE that incor-
porates both the N110 and N130 residues. Our studies
therefore place these N-glycosylation sites in a region
of the protein that has previously been mapped
as being of functional importance for TfR1 binding.

It remains possible that the 110/130 combination
of mutations may directly disrupt regulation of
TfR1, although specific co-immunoprecipitation and
N-Glycosylation of HFE L. Bhatt et al.
3230 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
functional studies with TfR1 would be required to
further explore this possibility.
The molecular basis for the observed deficiency in
b2M binding of the triple mutant remains unclear. One
possibility is that glycosylation deficiency compromises
HFE folding and the misfolded protein cannot bind
b2M. At a cellular level, however, underglycosylation
of HFE that compromises protein trafficking may mean
that it never accesses b2M. Further experiments will be
needed to clarify the exact sequence with which HFE
binds its interacting partners and progress the work of
Gross et al. [18] who first addressed the sequential
order and glycosylation dependence of events.
In summary, our results demonstrate that N-glyco-
sylation is required for the appropriate localization of
HFE, presumably due to effects on trafficking. Future
studies will address the role that this may play in the
functional regulation of transferrin binding.
Materials and methods
Cell culture
HEK293T and HuTu80 cells were obtained from the Amer-
ican Type Culture Collection and cultured in Dulbecco’s
modified Eagle’s medium containing 10% fetal bovine
serum and 1% penicillin/streptomycin solution. Cells were
maintained in a 5% CO

2
humidified incubator at 37 °C.
HEK293T were transfected with 1–2 lg of HFE-expressing
plasmid DNA using calcium phosphate precipitation.
HuTu80 cells were transfected using Effectene (Qiagen,
Crawley, UK) according to the manufacturer’s protocol.
Unless otherwise stated, HFE-transfected HEK293T and
HuTu80 cells were co-transfected with an equal concentra-
tion of pCDNA3.1–b2M or pSPORT–b2M as described
previously [11,12,28]. Tunicamycin was purchased from
Sigma (Arklow, Ireland).
Expression constructs and site-directed
mutagenesis
Mammalian expression vectors pEP7–HA, pEP7–HFE–
HA, pEP7–HFE-C282Y–HA, pEP7–HFE–GFP, pEP7–
HFE-C282Y–HA, pCDNA3.1–b2M, pSPORT6–b2M,
pEGFP–N1 HFE WT pEGFP-N1-HFE-WT and pEGFP-
N1-HFE-C282Y have been described previously [12,13,28,
50–52]. The pEP7–HFE–HA vector was used as template
to mutate putative N-glycosylation sites in the HFE coding
sequence using the QuikChange Site-Directed Mutagenesis
protocol (Stratagene, Agilent Technologies, Dublin,
Ireland). To generate the pEP7–HFE-N110A–HA vector
we used the following primer set: sense, ATGGAAAATC
ACGCCCACAGCAAGGAG; antisense, CTCCTTGTCG
TGGGCGTGATTTTCCAT. To generate the pEP7–HFE-
N130A–HA mutation we used the following primer set:
sense, ATGCAAGAAGACGCCAGTACCGAGGGC; anti-
sense, GCCCTCGGTACTGGCGTCTTCTTGCAT. To
generate the pEP7–HFE-N234A–HA vector we used the

following primer set: sense, TACTACCCCCAGGCCATC
ACCATGAAG; antisense, CTTCATGGTGATGGCCTG
GGGGTAGTA. Bold type indicates the mutations. For
simplicity, the proteins expressed from these constructs are
frequently described in the text and figures as the single
mutants.
Double glycosylation site mutations were generated by
using expression constructs for the single mutants as tem-
plates for site-directed mutagenesis. The HFE NN110/
130AA mutant was made by introducing the N130A muta-
tion into the pEP7–HFE-N110A–HA plasmid. The HFE
NN130/234AA mutant was made by introducing the
N234A mutation into the pEP7–HFE-N130A–HA plasmid.
The HFE NN234/110AA mutant was made by introducing
the N110A mutation into the pEP7–HFE-N234A–HA
plasmid. For simplicity, the proteins expressed from these
vectors are frequently described in the text and figures as
double mutants.
An expression construct lacking N-glycosylation sites
was generated by introducing the N234A mutation into the
pEP7–HFE-NN110/130AA–HA plasmid and sequenced.
Throughout the text and figures, the protein expressed from
this construct is frequently referred as the triple mutant.
Mutated HFE cDNAs were subcloned into the BglII/
SalI sites of the pEP7–GFP vector [52].
Western blot analysis
Experimental cells were lysed in RIPA buffer containing
complete protease inhibitor (EDTA-free; Boehringer,
Dublin, Ireland) and phosphatase inhibitors (1 mm sodium
vanadate and sodium fluoride). Cleared cell lysates were

fractionated on 11% SDS/polyacrylamide gels and trans-
ferred to nitrocellulose membranes for western blotting with
a monoclonal anti-HA Ig (Covance, Crawley, UK). Fluores-
cently tagged goat anti-(mouse secondary Ig) IRDye 680DX
(LI-COR Biosciences, Cambridge, UK) was used for pri-
mary antibody detection and blots were scanned using an
infrared imager (Odyssey; LI-COR Biosciences). Associated
software provided quantification of band intensity.
PNGase F digestion
For in vitro deglycosylation experiments, HEK293T cells
expressing wild-type and mutant HFE variants were lysed
in RIPA buffer. Cleared cell lysates were precleared with
20 lL of Protein A/G PLUS-Agarose (Santa Cruz Biotech-
nology, Heidelberg, Germany) and HA-tagged proteins
were immunoprecipitated using a 1 : 200 dilution of rabbit
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3231
polyclonal a-HA (Santa Cruz Biotechnology) and 20 lLof
Protein A/G agarose. Immunoprecipitates were recovered
by centrifugation and washed three times in RIPA buffer.
Immunoprecipitated proteins were boiled for 10 min in the
presence of glycoprotein denaturing buffer (New England
Biolabs, Hitchin, UK) and transferred to a new tube.
Samples were then incubated at 37 °C for 1 h with
PNGase F and NP-40 as advised by the manufacturer
(New England Biolabs) to remove N-glycans. Samples were
denatured by boiling in Laemmli buffer for 5 min and frac-
tionated on 11% polyacrylamide gels before transfer and
western blotting as described above.
Co-immunoprecipitation

HEK293T cells were transfected to express wild-type,
C282Y, triple-mutant or single-mutant forms of HFE.
Forty-eight hours later the cells were harvested in 50 mm
Hepes buffer (pH7.4) containing 250 mm NaCl, 0.5%
Tween-20 and 0.5 mm dithiothreitol and sonicated on ice. A
sample of cleared cell lysate was fractionated on 11% SDS/
PAGE for anti-HA immunoblotting as described above.
The remaining cleared cell lysates were precleared with
20 lL of Protein A/G PLUS-Agarose (Santa Cruz Biotech-
nology) and proteins that were interacting with endogenous
b2 were immunoprecipitated using a 1 : 600 dilution of rab-
bit polyclonal a-b2M (Abcam, Cambridge, UK) and 20 lL
of Protein A/G agarose. Immunoprecipitates were recovered
by centrifugation and washed three times in lysis buffer.
Immunoprecipitated proteins were boiled for 10 min in
Laemmli buffer and fractionated on 11% SDS/PAGE gel
for immunoblotting with a mouse anti-HA Ig.
Immunofluorescence microscopy
HuTu80 cells were cultured and transfected, and immuno-
fluorescence microscopy was performed essentially as
described [28]. Primary antibodies used were mouse mono-
clonal anti-HA (Abcam), rabbit anti-BiP (Abcam) and anti-
TfnR (Zymed, Biosciences, Dun Laoighaire, Ireland). The
secondary antibodies used were goat anti-mouse conjugated
to TRITC (Jackson ImmunoResearch, Newmarket, UK) or
Alexa Fluor 488 (Molecular Probes, Biosciences, Dun Lao-
ighaire, Ireland), and donkey anti-rabbit conjugated to
TRITC or Alexa Fluor 488. Coverslips were mounted in
MOWIOL (CalBiochem, Nottingham, UK). For Tfn-bind-
ing experiments, cells were serum starved for 2 h and then

incubated for 1 h, at 4 °C, with 5 lgÆmL
)1
of Alexa -
Fluor 594-labelled iron-saturated holotransferrin (Invitro-
gen, Biosciences, Dun Laoghaire, Ireland). Images were
recorded using a Zeiss LSM 510 META confocal micro-
scope fitted with a 63·/1.4 plan apochromat lens. Co-locali-
zation masks were generated with Zeiss image examiner
Software. Overlapping green and red pixels were extracted
from the merged image and converted to white.
Acknowledgements
This work was supported by grants from Fundac¸ a
˜
o
para Cieˆ ncia e a Technologia (POCI/SAU-MMO/
61129/2004) and the Health Research Board (RP/
2006/294) to J.V.F and Health Research Board Grant
(RP/2005/107) and Science Foundation Ireland (SFI)
Investigator Grant (05/IN.3/13859) to MMC. The
authors wish to thank Maria de Sousa and Sergio de
Almeida for useful discussions.
References
1 Fleming RE, Britton RS, Waheed A, Sly WS & Bacon
BR (2005) Pathophysiology of hereditary hemochroma-
tosis. Semin Liver Dis 25, 411–419.
2 Pietrangelo A (2002) Physiology of iron transport and
the hemochromatosis gene. Am J Physiol Gastrointest
Liver Physiol 282, G403–G414.
3 Hentze MW, Muckenthaler MU & Andrews NC (2004)
Balancing acts: molecular control of mammalian iron

metabolism. Cell 117, 285–297.
4 Feder JN, Gnirke A, Thomas W, Tsuchihashi Z, Ruddy
DA, Basava A, Dormishian F, Domingo R, Ellis MC,
Fullan A et al. (1996) A novel MHC class I-like gene is
mutated in patients with hereditary haemochromatosis.
Nat Genet 13, 399–408.
5 Parkkila S, Waheed A, Britton RS, Bacon BR, Zhou
XY, Tomatsu S, Fleming RE & Sly WS (1997) Associa-
tion of the transferrin receptor in human placenta with
HFE, the protein defective in hereditary hemochroma-
tosis. Proc Natl Acad Sci USA 94, 13198–13202.
6 Parkkila S, Waheed A, Britton RS, Feder JN, Tsuchih-
ashi Z, Schatzman RC, Bacon BR & Sly WS (1997)
Immunohistochemistry of HLA-H, the protein defective
in patients with hereditary hemochromatosis, reveals
unique pattern of expression in gastrointestinal tract.
Proc Natl Acad Sci USA 94, 2534–2539.
7 Pietrangelo A (2006) Hereditary hemochromatosis.
Annu Rev Nutr 26, 251–270.
8 Muckenthaler MU, Rodrigues P, Macedo MG, Minana
B, Brennan K, Cardoso EM, Hentze MW & de Sousa
M (2004) Molecular analysis of iron overload in beta2-
microglobulin-deficient mice. Blood Cells Mol Dis 33,
125–131.
9 Feder JN, Tsuchihashi Z, Irrinki A, Lee VK, Mapa
FA, Morikang E, Prass CE, Starnes SM, Wolff RK,
Parkkila S et al. (1997) The hemochromatosis founder
mutation in HLA-H disrupts beta2-microglobulin inter-
action and cell surface expression. J Biol Chem 272,
14025–14028.

10 Waheed A, Parkkila S, Zhou XY, Tomatsu S, Tsu-
chihashi Z, Feder JN, Schatzman RC, Britton RS,
Bacon BR & Sly WS (1997) Hereditary hemochroma-
tosis: effects of C282Y and H63D mutations on
N-Glycosylation of HFE L. Bhatt et al.
3232 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS
association with beta2-microglobulin, intracellular
processing, and cell surface expression of the HFE
protein in COS-7 cells. Proc Natl Acad Sci USA 94,
12384–12389.
11 Bhatt L, Horgan CP & McCaffrey MW (2009)
Knockdown of beta2-microglobulin perturbs the
subcellular distribution of HFE and hepcidin. Biochem
Biophys Res Commun 378, 727–731.
12 de Almeida SF, Fleming JV, Azevedo JE, Carmo-Fons-
eca M & de Sousa M (2007) Stimulation of an unfolded
protein response impairs MHC class I expression.
J Immunol 178, 3612–3619.
13 de Almeida SF, Picarote G, Fleming JV, Carmo-Fonse-
ca M, Azevedo JE & de Sousa M (2007) Chemical
chaperones reduce endoplasmic reticulum stress and
prevent mutant HFE aggregate formation. J Biol Chem
282, 27905–27912.
14 de Almeida SF & de Sousa M (2009) The unfolded
protein response in hereditary hemochromatosis. J Cell
Mol Med 12, 421–434.
15 Feder JN, Penny DM, Irrinki A, Lee VK, Lebron JA,
Watson N, Tsuchihashi Z, Sigal E, Bjorkman PJ &
Schatzman RC (1998) The hemochromatosis gene
product complexes with the transferrin receptor and

lowers its affinity for ligand binding. Proc Natl Acad
Sci USA 95, 1472–1477.
16 Lebron JA, Bennett MJ, Vaughn DE, Chirino AJ,
Snow PM, Mintier GA, Feder JN & Bjorkman PJ
(1998) Crystal structure of the hemochromatosis protein
HFE and characterization of its interaction with
transferrin receptor. Cell 93, 111–123.
17 Bennett MJ, Lebron JA & Bjorkman PJ (2000) Crystal
structure of the hereditary haemochromatosis protein
HFE complexed with transferrin receptor. Nature 403,
46–53.
18 Gross CN, Irrinki A, Feder JN & Enns CA (1998)
Co-trafficking of HFE, a nonclassical major histocom-
patibility complex class I protein, with the transferrin
receptor implies a role in intracellular iron regulation.
J Biol Chem 273, 22068–22074.
19 Fleming MD, Romano MA, Su MA, Garrick LM,
Garrick MD & Andrews NC (1998) Nramp2 is mutated
in the anemic Belgrade (b) rat: evidence of a role for
Nramp2 in endosomal iron transport. Proc Natl Acad
Sci USA 95, 1148–1153.
20 Lebron JA, West AP Jr & Bjorkman PJ (1999) The
hemochromatosis protein HFE competes with transfer-
rin for binding to the transferrin receptor. J Mol Biol
294, 239–245.
21 West AP Jr, Giannetti AM, Herr AB, Bennett MJ,
Nangiana JS, Pierce JR, Weiner LP, Snow PM &
Bjorkman PJ (2001) Mutational analysis of the
transferrin receptor reveals overlapping HFE
and transferrin binding sites. J Mol Biol 313, 385–

397.
22 Lebron JA & Bjorkman PJ (1999) The transferrin
receptor binding site on HFE, the class I MHC-related
protein mutated in hereditary hemochromatosis. J Mol
Biol 289, 1109–1118.
23 Salter-Cid L, Brunmark A, Peterson PA & Yang Y
(2000) The major histocompatibility complex-encoded
class I-like HFE abrogates endocytosis of transferrin
receptor by inducing receptor phosphorylation. Genes
Immun 1, 409–417.
24 Salter-Cid L, Brunmark A, Li Y, Leturcq D, Peterson
PA, Jackson MR & Yang Y (1999) Transferrin receptor
is negatively modulated by the hemochromatosis pro-
tein HFE: implications for cellular iron homeostasis.
Proc Natl Acad Sci USA 96, 5434–5439.
25 Roy CN, Penny DM, Feder JN & Enns CA (1999) The
hereditary hemochromatosis protein, HFE, specifically
regulates transferrin-mediated iron uptake in HeLa
cells. J Biol Chem 274, 9022–9028.
26 Laham N, Rotem-Yehudar R, Shechter C, Coligan JE
& Ehrlich R (2004) Tranferrins receptor association and
endosomal localization of soluble HFE are not
sufficient for regulation of cellular iron homeostasis.
J Cell Biochem 91, 1130–1145.
27 Ikuta K, Fujimoto Y, Suzuki Y, Tanaka K, Saito H,
Ohhira M, Sasaki K & Kohgo Y (2000) Overexpression
of hemochromatosis protein, HFE, alters transferrin
recycling process in human hepatoma cells. Biochim
Biophys Acta 1496, 221–231.
28 Bhatt L, Horgan CP, Walsh M & McCaffrey MW

(2007) The hereditary hemochromatosis protein HFE
and its chaperone beta2-microglobulin localise
predominantly to the endosomal-recycling
compartment. Biochem Biophys Res Commun 359 ,
277–284.
29 Pinto JP, Ramos P & de Sousa M (2007) Overexpres-
sion of HFE in HepG2 cells reveals differences in
intracellular distribution and co-localization of wt- and
mutated forms. Blood Cells Mol Dis 39, 75–81.
30 Zhang AS, Davies PS, Carlson HL & Enns CA (2003)
Mechanisms of HFE-induced regulation of iron
homeostasis: insights from the W81A HFE mutation.
Proc Natl Acad Sci USA 100, 9500–9505.
31 Carlson H, Zhang AS, Fleming WH & Enns CA
(2005) The hereditary hemochromatosis protein, HFE,
lowers intracellular iron levels independently of
transferrin receptor 1 in TRVb cells. Blood 105,
2564–2570.
32 Kawabata H, Yang R, Hirama T, Vuong PT, Kawano
S, Gombart AF & Koeffler HP (1999) Molecular clon-
ing of transferrin receptor 2. A new member of the
transferrin receptor like family. J Biol Chem 274,
20826–20832.
33 Chen J, Chloupkova M, Gao J, Chapman-Arvedson TL
& Enns CA (2007) HFE modulates transferrin
receptor 2 levels in hepatoma cells via interactions that
L. Bhatt et al. N-Glycosylation of HFE
FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS 3233
differ from transferrin receptor 1–HFE interactions.
J Biol Chem 282, 36862–36870.

34 Goswami T & Andrews NC (2006) Hereditary hemo-
chromatosis protein, HFE, interaction with transferrin
receptor 2 suggests a molecular mechanism for mamma-
lian iron sensing. J Biol Chem 281, 28494–28498.
35 Ludwiczek S, Theurl I, Bahram S, Schu
¨
mann K &
Weiss G (2005) Regulatory networks for the control of
body iron homeostasis and their dysregulation in HFE
mediated hemochromatosis. J Cell Physiol 204, 489–
499.
36 Muckenthaler M, Roy CN, Custodio AO, Minana B,
deGraaf J, Montross LK, Andrews NC & Hentze MW
(2003) Regulatory defects in liver and intestine implicate
abnormal hepcidin and Cybrd1 expression in mouse
hemochromatosis. Nat Genet 34, 102–107.
37 Nicolas G, Viatte L, Lou D-Q, Bennoun M, Beaumont
C, Kahn A, Andrews NC & Vaulont S (2003) Constitu-
tive hepcidin expression prevents iron overload in a
mouse model of hemochromatosis. Nat Genet 34,
97–101.
38 Matzuk M & Boime I (1988) The role of the aspara-
gine-linked oligosaccharides of the alpha subunit in the
secretion and assembly of human chorionic gonadotro-
phin. J Cell Biol 106, 1049–1059.
39 Matzuk M, Keene J & Boime I (1989) Site specificity
of the chorionic gonadotropin N-linked oligosaccha-
rides in signal transduction. J Biol Chem 264,
2409–2414.
40 Makanji Y, Harrison CA, Stanton PG, Krishna R &

Robertson DM (2007) Inhibin A and B in vitro
bioactivities are modified by their degree of glycosyla-
tion and their affinities to betaglycan. Endocrinology
148, 2309–2316.
41 Williams A & Enns C (1991) A mutated transferrin
receptor lacking asparagine-linked glycosylation sites
shows reduced functionality and an association with
binding immunoglobulin protein. J Biol Chem 266,
17648–17654.
42 Williams A & Enns C (1993) A region of the C-terminal
portion of the human transferrin receptor contains an
asparagine-linked glycosylation site critical for receptor
structure and function. J Biol Chem 268, 12780–12786.
43 Schimanski LM, Drakesmith H, Sweetland E, Bastin J,
Rezgui D, Edelmann M, Kessler B, Merryweather-
Clarke AT, Robson KJ & Townsend AR (2009) In vitro
binding of HFE to the cation-independent mannose-6
phosphate receptor. Blood Cells Mol Dis 43, 180–193.
44 Elbein AD (1987) Inhibitors of the biosynthesis and
processing of N-linked oligosaccharide chains. Annu
Rev Biochem 56, 497–534.
45 Parham P, Alpert BN, Orr HT & Strominger JL (1977)
Carbohydrate moiety of HLA antigens. Antigenic prop-
erties and amino acid sequences around the site of gly-
cosylation. J Biol Chem 252, 7555–7567.
46 Zhang Y & Dahms NM (1993) Site-directed removal of
N-glycosylation sites in the bovine cation-dependent
mannose 6-phosphate receptor: effects on ligand
binding, intracellular targetting and association with
binding immunoglobulin protein. Biochem J 295(Pt 3),

841–848.
47 Weitz G & Proia RL (1992) Analysis of the glyco-
sylation and phosphorylation of the alpha-subunit of
the lysosomal enzyme, beta-hexosaminidase A, by
site-directed mutagenesis. J Biol Chem 267, 10039–
10044.
48 Ng DT, Hiebert SW & Lamb RA (1990) Different roles
of individual N-linked oligosaccharide chains in folding,
assembly, and transport of the simian virus 5 hemagglu-
tinin-neuraminidase. Mol Cell Biol 10, 1989–2001.
49 Barton JC, Sawada-Hirai R, Rothenberg BE & Acton
RT (1999) Two novel missense mutations of the HFE
gene (I105T and G93R) and identification of the S65C
mutation in Alabama hemochromatosis probands.
Blood Cells Mol Dis 25, 147–155.
50 Fleming JV, Fajardo I, Langlois MR, Sanchez-Jimenez
F & Wang TC (2004) The C-terminus of rat l-histidine
decarboxylase specifically inhibits enzymic activity and
disrupts pyridoxal phosphate-dependent interactions
with l-histidine substrate analogues. Biochem J 381(Pt
3), 769–778.
51 Fleming JV, Sanchez-Jimenez F, Moya-Garcia AA,
Langlois MR & Wang TC (2004) Mapping of catalyt-
ically important residues in the rat l-histidine decar-
boxylase enzyme using bioinformatic and site-directed
mutagenesis approaches. Biochem J 379 (Pt 2), 253–
261.
52 Fleming JV & Wang TC (2003) The production of
53–55-kDa isoforms is not required for rat l-histidine
decarboxylase activity. J Biol Chem 278 , 686–694.

N-Glycosylation of HFE L. Bhatt et al.
3234 FEBS Journal 277 (2010) 3219–3234 ª 2010 The Authors Journal compilation ª 2010 FEBS