1382
Biochemical Society Transactions (2005) Volume 33, part 6
Defining the protein–protein interactions
of the mammalian endoplasmic reticulum
oxidoreductases (EROs)
S. Dias-Gunasekara and A.M. Benham1
Department of Biological and Biomedical Sciences, University of Durham, South Road, Durham DH1 3LE, U.K.
Abstract
The ER (endoplasmic reticulum) is the site of protein folding for all eukaryotic secreted and plasma
membrane proteins. Disulphide bonds are formed in many of these proteins through a dithiol–disulphide
exchange chain comprising two types of protein catalysts: PDI (protein disulphide-isomerase) and ERO (ER
oxidoreductase) proteins. This review will examine what we know about ERO function, and will then consider
ERO interactions and their implications for mammalian oxidative protein folding.
Identification of EROs (endoplasmic
reticulum oxidoreductases)
In eukaryotic cells, the ER (endoplasmic reticulum) provides
an environment for secretory and outer membrane proteins
to fold into their stable conformations with the assistance of
various chaperone proteins and enzymes. One of the posttranslational modifications that confers structural stability
and enzymatic function is the formation of disulphide
bonds at the correct position during protein folding. With
the discovery of yeast Ero1p [1,2], a redox pathway for
transferring disulphide bonds to a newly synthesized protein
was uncovered. This led to the identification of two hERO
(human ERO) proteins, Ero1α and Ero1β [3,4]. Both hEROs
are capable of rescuing the Saccharomyces cerevisiae ero1-1
thermosensitive mutant and catalysing the oxidation of model
substrates, indicating that these proteins are involved in
disulphide bond formation.
A schematic diagram of the hEROs is shown in Figure 1.
Both hEROs have an N-terminal signal sequence that targets
these proteins to the ER, where they are resident. Ero1α has
two N-glycosylation sites, whereas Ero1β has four, although
whether glycosylation affects protein function is yet unknown. Ero1α and Ero1β both have two conserved redoxactive motifs, a CXXXXC motif near the N-terminus and
a CXXCXXC motif towards the C-terminus. Mutational
and biochemical analysis of the yeast ERO1 gene has shown
that the N-terminal active site cysteines are involved in
a dithiol–disulphide exchange reaction with PDI (protein
disulphide-isomerase; Figure 1C). Accepted electrons are
then transferred on to the C-terminal active site cysteines
that then pass on electrons to molecular oxygen via FAD
[5,6]. The crystal structure of the yeast Ero1p core shows that
Key words: endoplasmic reticulum oxidoreductase, flavoprotein, oxidoreductase, protein
disulphide-isomerase, protein folding, redox.
Abbreviations used: ER, endoplasmic reticulum; ERO, ER oxidoreductase; hERO, human ERO; PDI,
protein disulphide-isomerase; UPR, unfolded protein response.
1
To whom correspondence should be addressed (email adam.benham@durham.ac.uk).
C 2005
Biochemical Society
the first (most N-terminal) cysteine in the CXXCXXC motif
is involved in a long-range intramolecular disulphide bond,
whereas the two more C-terminal cysteines (CXXCXXC)
are redox active and are positioned for electron transfer to
bound FAD [6a].
ER retention of hEROs
S. cerevisiae Ero1p has a C-terminal tail of 127 amino acid
residues that enables this protein to be localized in the ER
[7]. hEROs lack this C-terminal tail and do not have a
conventional KDEL ER retention/retrieval sequence. So how
are the soluble hEROs retained in the ER lumen? Biochemical
analysis has shown that hEROs are closely associated with ER
membranes [4] and more recently Ero1α has been found in
mixed disulphide complexes with ERp44. ERp44 is a relative
of PDI, and interactions with both ERp44 and PDI are likely
to contribute to ER retention of hEROs indirectly, by virtue
of their partner protein’s KDEL (or RDEL) motif [8,9].
Ero1α and Ero1β regulation
Although Ero1α and Ero1β have many similar biochemical
properties, there are some important differences between
them. Ero1β transcript expression can be induced chemically by preventing N-glycosylation and disulphide bond
formation, both of which trigger the UPR (unfolded protein
response) [4]. Ero1α, on the other hand, is regulated by
cellular oxygen tension [10]. Recent evidence examining the
secreted proangiogenesis factor VEGF (vascular endothelial
growth factor) showed that Ero1α is involved in a HIF-l
(hypoxia inducible factor-1)-mediated pathway used to
induce protein secretion under hypoxic conditions. Targeting
Ero1α expression in tumour cells might therefore be a
potential therapeutic target for future cancer therapies [11].
At the mRNA level, Ero1α and Ero1β transcripts also
display differential tissue expression, with Ero1α transcripts
abundant in the oesophagus, compared with the pancreas,
Cellular Information Processing
Figure 1 Schematic representation of hERO proteins and
dithiol–disulphide exchange
(A) Ero1α. (B) Ero1β. (C) Model of the flow of electrons donated from
PDI to molecular oxygen via Ero and FAD. The N-terminal CXXXXC and
C-terminal CXXCXXC active site motifs are shown. Ball and stick represents
potential N-glycosylation sites. s.s, signal sequence.
bond that alters structural elements of the Ero protein.
CXXAXXC and CXXCXXA mutants remain capable of
binding to PDI, but are enzymatically crippled by the loss
of their redox-active cysteines residues. However, Ero1β is
also able to form alkylation-independent complexes and these
interactions were investigated using differentially tagged
Ero1β constructs [12]. A notable proportion of Ero1β exists
as a homodimeric pool at steady state and mutating Cys396
(CXXCXXA) prevents homodimer formation. In addition,
Ero1α and Ero1β are able to form mixed dimers. Using
the crystal contact details from the Ero1p monomer [14] a
symmetrical dimer was modelled in which FAD contributes
up to 20% of the dimer interface [12]. Although the Cys396
residue is buried too deeply within the Ero monomer to form
an intermolecular disulphide bond, its close proximity to
FAD suggests that the mutation might disrupt FAD binding.
This probably leads to collapse of the FAD–FAD interface
and prevention of dimerization.
hERO dimers and their possible function
stomach, testis and pituitary glands, where Ero1β is
strongly expressed [4]. However, the expression of Ero1β
at the protein level and the requirement for induction by
UPR in vivo has remained unknown. We have therefore
investigated the expression of Ero1β proteins in secretory
tissues, including the stomach and pancreas.
Tissue-specific Ero1β expression
Using immunohistochemistry, we have shown that in human
stomach, Ero1β and PDI, along with other ER chaperones,
are highly expressed in enzyme-secreting chief cells. In contrast, in the human pancreas, Ero1β and PDI expression are
not strictly correlated. Ero1β is expressed in both exocrine
(acinar) and endocrine (islet) cells, compared with the mostly
exocrine expression of PDI and its pancreas-specific sister
PDIp [12]. Clearly the control of disulphide bond formation
in vivo is more complicated than we first imagined in
mammals, with Ero1β expression constitutively and strongly
switched on in at least some secretory tissues. Not only is
Ero1β expression tissue-specific, there are also cell-specific
differences in expression within a tissue. These results
prompted us to look in more detail at the types of interactions
that hEROs could facilitate in the ER.
hERO interactions
We and others have shown that both Ero1α and Ero1β
interact with PDI in a disulphide- and alkylation-dependent
manner [12,13]. Using mutational studies of the hERO C-terminal active site, it has been shown that an AXXCXXC
hERO mutant interacts poorly with PDI. This is probably
due to disruption of the long-range intramolecular disulphide
What could be the significance of dimer formation for hERO
function? A dimer might protect non-specific proteins from
oxidative damage by spent electrons, particularly since Eroβ
is highly expressed in secretory tissues. If left unregulated,
electron transfer to EROs could result in the generation of
reactive oxygen species and oxidative stress. Alternatively,
hERO dimers could also be used for the storage of oxidizing
power in tissues with intermittent secretory demands, thus
allowing for rapid mobilization of disulphide bond donors.
However, it would be unexpected for a tissue to switch on
protein expression via the UPR and then store that protein
in an inactive form. A third possibility is that hERO dimers
form part of a larger oxidation complex. The dimer interface
probably does not preclude PDI binding, so it is possible
that PDI–ERO–ERO–PDI tetramers are brought together
to facilitate efficient oxidative protein folding in the ER.
Addressing the true physiological function of hERO dimers
is not trivial, and will require a combination of in vitro assays
and site-directed mutagenesis, as well as the generation of
tissue-specific knockout and knockin mice. The ER still has
much more to tell us about how disulphide bond formation
and protein secretion are organized and regulated in complex
animals.
References
1 Frand, A.R. and Kaiser, C.A. (1998) Mol. Cell 1, 161–170
2 Pollard, M.G., Travers, K.J. and Weissman, J.S. (1998) Mol. Cell 1, 171–182
3 Cabibbo, A., Pagani, M., Fabbri, M., Rocchi, M., Farmery, M.R., Bulleid,
N.J. and Sitia, R. (2000) J. Biol. Chem. 275, 4827–4833
4 Pagani, M., Fabbri, M., Benedetti, C., Fassio, A., Pilati, S., Bulleid, N.J.,
Cabibbo, A. and Sitia, R. (2000) J. Biol. Chem. 275, 23685–23692
5 Tu, B.P. and Weissman, J.S. (2002) Mol. Cell 10, 983–994
6 Frand, A.R. and Kaiser, C.A. (2000) Mol. Biol. Cell 11, 2833–2843
6a Gross, E., Kastner, D.B., Kaiser, C.A. and Fass, D. (2004)
Cell (Cambridge, Mass.) 117, 601–610
7 Pagani, M., Pilati, S., Bertoli, G., Valsasina, B. and Sitia, R. (2001)
FEBS Lett. 508, 117–120
C 2005
Biochemical Society
1383
1384
Biochemical Society Transactions (2005) Volume 33, part 6
8 Anelli, T., Alessio, M., Bachi, A., Bergamelli, L., Bertoli, G., Camerini, S.,
Mezghrani, A., Ruffato, E., Simmen, T. and Sitia, R. (2003) EMBO J. 22,
5015–5022
9 Anelli, T., Alessio, M., Mezghrani, A., Simmen, T., Talamo, F., Bachi, A.
and Sitia, R. (2002) EMBO J. 21, 835–844
10 Gess, B., Hofbauer, K.H., Wenger, R.H., Lohaus, C., Meyer, H.E. and
Kurtz, A. (2003) Eur. J. Biochem. 270, 2228–2235
11 May, D., Itin, A., Gal, O., Kalinski, H., Feinstein, E. and Keshet, E. (2005)
Oncogene 24, 1011–1020
C 2005
Biochemical Society
12 Dias-Gunasekara, S., Gubbens, J., van Lith, M., Dunne, C., Williams, J.A.,
Kataky, R., Scoones, D., Lapthorn, A., Bulleid, N. and Benham, A.M.
(2005) J. Biol. Chem. 280, 33066–33075
13 Benham, A.M., Cabibbo, A., Fassio, A., Bulleid, N., Sitia, R. and
Braakman, I. (2000) EMBO J. 19, 4493–4502
Received 26 July 2005