Mechanisms and applications of disulfide bond formation
Professor Lloyd Ruddock
Biocenter Oulu and Department of Biochemistry
FIN-90014 University of Oulu
Tel: +358-8-553 1136
Fax: +358-8-553 1141
Background and significance
Protein folding for both soluble and membrane proteins is an essential,
complex multi-factorial process. Around one-third of all human proteins
fold in the endoplasmic reticulum (ER) and have an added complexity during
protein folding - the formation of native disulfide bonds.
The formation of these covalent bonds is often the rate limiting step
of protein folding in vivo and in vitro.
At its simplest native disulifde bond formation can be broken down
into two distinct steps, oxidation of two cysteines to form a disulfide bond and
subsequent isomerization of the disulfide bonds formed to the native functional state.
In practice, the mechanisms of native disulfide bond formation are much more complex
and the complexity is further increased by having multiple parallel pathways for their
formation (figure 1). Currently nearly 40 proteins have been implicated in native
disulfide formation in the ER of mammalian cells.
While many of the participants in the cellular process of native disulfide bond
formation in the ER are known, their individual physiological roles and/or mechanisms
of action are often unclear. Furthermore, virtually nothing is known about their
synergy of action with each other or with other ER resident molecular chaperones
and folding catalysts or with components of the ER-associated degradation pathway (ERAD).
However, protein disulfide isomerase (PDI) is central to most of the known mechanisms
for disulfide bond formation (figure 1) and, as the name suggests, proteins belonging
to this family are the only known, physiologically relevant, route for disulfide bond
isomerization in protein substrates with substantial regular secondary structure.
Figure 1: Pathways for disulfide bond formation in the ER.
Dotted arrows represent hypothetical pathways, red arrows indicate pathways in which our group
has recently generated significant new data, the green arrow indicates a pathway for which the
physiological relevance is unclear but which has recently been shown to be the major route in
the periplasm of some bacteria. ROS = reactive oxygen species
Due to the complexities of their formation the production of proteins that contain
disulfide bonds is difficult. This severely inhibits scientific progress in understanding a myriad
of mechanistic processes and imposes limitations on the biotechnology industry for the production
of therapeutic proteins. While disulfide bond containing proteins can be made in eukaryotic systems,
the most common route for their biopharmaceutical scale production involves heterologous expression
of (mammalian) proteins in the cytoplasm of E.coli. Here there are no mechanisms for disulfide bond
formation and the recombinant proteins form insoluble inclusion bodies.
has two pathways to ensure that its cytoplasm is reducing i) using thioredoxins/thioredoxin
reductase and ii) using glutathione/glutathione reductase (figure 2). When both pathways are
disrupted, for example in the commercial strains origami, rosetta-gami (Novagen) or SHuffle (New England Biolabs),
the cytoplasm is less reducing and disulfide bonds can form. These systems allow the production of disulfide
bond containing proteins, however, disulfide formation in these strains is very inefficient.
Figure 2: Pathways for disulfide bond reduction in the cytoplasm of
E.coli. To disrupt both pathways the most common route is to knock-out the gor and trxB genes.
Three distinct, but interconnected areas of progress are outlined below. These range from the
fundamental understanding of protein disulfide bond formation in the ER to the application
of this understanding for improved technologies for recombinant protein production and back
to fundamental understanding of protein biogenesis.
i) Disulfide bond formation in the ER
For more than a decade the major route for disulfide bond formation in the ER was thought
to be via the oxidation of PDI (or a PDI-family member) by the sulphydryl oxidase Ero1 (figure 1),
with the role of other systems, such as the glutathione redox buffer being downplayed.
These views sometimes overlooked decades of in vitro studies which showed that the optimal
glutathione ratios for native disulfide bond formation matched those found in vivo.
Similarly, once it was reported that Ero1 made one molecule of hydrogen peroxide per
disulfide bond the field shifted towards the view that disulfide bond formation per se
was toxic to cells, but this overlooks both the inefficiency of this model and the
chemistry of peroxide and thiols. Two years ago we reported that in vitro peroxide
could efficiently be used to make disulfide bonds and that native disulfide bond
formation was not only possible, but kinetically faster than the use of a glutathione
buffer, with minimal side reactions. This lead to the subsequent identification by us
and others of multiple catalytic systems for this process and the radical shift in
viewpoint from peroxide being a problem to peroxide being a possible major route for
disulfide bond formation in the ER, a shift reinforced by the discovery that Ero1 is
not only not an essential gene product in mammals, but that its deletion has minimal
Other small molecules also potentially play an important role in disulfide bond
formation in the ER. In the past year we reported studies on dehydroascorbate (DHA),
one of the oxidized forms of vitamin C, is able to extremely rapidly oxidise dithiols to
disulfides in peptides and non-native proteins in the absence of a catalyst. This study also
sparked considerable interest and we and others are trying to distinguish if this is a
relatively minor route for disulfide bond formation in vivo (via the antioxidant role of ascorbate)
or whether there is a catalytic system for making DHA in the ER of some organisms.
Also within the last year we published a long standing and controversial study on the kinetics
of oxidation and reduction of PDI by glutathione, which when combined with some recent in vivo
studies on how rapidly the optimal glutathione redox buffer is restored after treatment of cells
with reductants we hope will allow glutathione to retake its rightful place centre-stage in
oxidative folding in the lumen of the ER.
ii) Identification of human ER-resident proteins involved in protein folding
There are already over 80 human proteins directly implicated in protein folding in the ER.
However, given the recent identification of a significant number of these, including ERp18,
ERp27 and two PDI-peroxidases identified by our group there are likely to be a significant
number yet to be identified or to which the wrong physiological role has been assigned.
Identification of soluble ER-resident proteins is aided by the requirement for an N-terminal
signal sequence and a C-terminal KDEL-like ER retrieval motif. It is known that variants
of KDEL also work to keep proteins ER-resident, with 24 possible variants being listed as
the PROSITE-motif for the ER-localization of soluble proteins. Using novel reporter
constructs we defined 37 new retrieval motifs that work in human cells and determined
the specificity of the three human KDEL-receptors to both existing and novel KDEL-variants.
For this we were awarded Biocenter Oulu discovery of the year. These new ER-retrieval motifs
are found on 29 human proteins which enter the secretory pathway and we have confirmed the
ER localisation of 16 of them, most of which we believe are functionally linked to protein
folding. We have also shown that the C-terminal retention motif is not only limited to the
terminal four amino acids, but must be extended to include positions -5 and -6 as well.
iii) The CyDisCo system
From our studies on understanding the mechanisms for disulfide bond formation we have
developed systems which allow efficient disulfide bond formation in the cytoplasm of
E. coli. These systems are based on the expression of a catalyst of disulfide bond
formation and a catalyst of isomerization. While the latter idea has been widely used,
the former idea has not previously been considered in main due to misunderstanding of
the effects of knocking-out the gor and trxB genes (figure 2). This system, known as
the CyDisCo system (cytoplasmic disulfide bond formation in E.coli), has a variety of
formats, but all share the same basic principles.
Initial studies with the CyDisCo system shows that the system is very successful,
that knockout of the reducing pathways in the cytoplasm is not essential for disulfide
bond formation and that high yields of active, correctly folded, eukaryotic proteins
can be obtained. The potential importance of these papers is demonstrated by the fact
that both were accessed over 700 times within one month of acceptance and the second
is the 3rd most highly accessed paper from the journal from this year.
Current versions of the CyDisCo system, while still not optimal for all proteins,
can easily be transferred between bacterial strains and allow the production of human
proteins with multiple disulfide bonds in E.coli grown in shake flasks with yields of
up to 100 mg/litre culture.
The overall aim of the group is to provide a complete molecular description of the
processes by which protein folding occurs within the ER and the application of this
knowledge for the efficient production of disulfide bond containing proteins of
scientific, medicinal or biotechnological importance.
The research within the group can currently be subdivided into three overlapping and inter-dependent areas:
- Pathways involved in oxidative protein folding
- Structure based molecular enzymology
- Development of the CyDisCo system
In addition there is a smaller subproject arising out of studies connected with CyDisCo connected
with determinants of membrane protein topology. The research work is highly collaborative and
involves collaborations with 11 groups in 6 countries.
In particular the group aims to fully exploit the new opportunities arising from the development
of the CyDisCo system for the molecular characterization of proteins involved in disulfide bond
formation in the ER. New important insights have already been achieved by being able to make
homogenously folded proteins in high yields and we want to extend this to proteins involved in
ERAD and to look at the inter-connectedness and synergy in the folding and ERAD pathways.
Recent publications from the group:
- Nguyen V.D., Hatahet F., Salo K.E., Enlund E, Zhang C. and Ruddock L.W.
Pre-expression of a sulfhydryl oxidase significantly increases the yields of
eukaryotic disulfide bond containing proteins expressed in the cytoplasm of
E.coli. Micro. Cell Fact. (2011) 10:1
- Nguyen V.D., Saaranen M.J., Karala A.R., Lappi A.K., Wang L., Raykhel I.B.,
Alanen H.I., Salo K.E., Wang C.C. and Ruddock L.W. Two endoplasmic reticulum
PDI-peroxidases increase the efficiency of the use of peroxide during disulfide
bond formation. J. Mol. Biol. (2011) 406, 503-515
- Lappi A.K. and Ruddock L.W. Re-examination of the role of interplay between
glutathione and protein disulphide isomerase. J. Mol. Biol.(2011) 409, 238-249
- Alanen H.I., Raykhel I.B., Luukas M.J., Salo K.E.H. and Ruddock L.W. Beyond
KDEL: The role of positions 5 and 6 in determining ER localization. J. Mol. Biol.(2011) 409, 291-297
Karala A.R., Lappi, A.K. and Ruddock, L.W. Modulation of an active-site
cysteine pKa allows PDI to act as a catalyst of both disulfide bond formation and isomerization.
J.Mol.Biol. (2010) 396, 883-892
- Karala A.R. and Ruddock, L.W. Bacitracin is not a specific inhibitor of protein
disulfide isomerase. FEBS J. (2010) 277, 2454-2462
- Hatahet F., Nguyen V.D., Salo K.E.H. and Ruddock L.W. Disruption of reducing
pathways is not essential for efficient disulfide bond formation in the cytoplasm
of E.coli. Micro. Cell Fact. (2010) 9, 67
- Saaranen M.J., Karala A.R., Lappi, A.K. and Ruddock L.W. The role of
dehydroascorbate in disulfide bond formation. Antioxid. Redox Signal. (2010) 12, 15-25
- Karala, A.R., Lappi A.K., Saaranen M. and Ruddock L.W. Efficient peroxide
mediated oxidative refolding of a protein at physiological pH and implications for oxidative
folding in the endoplasmic reticulum. Antioxid. Redox Signal. (2009) 11, 963-970
- Rowe, M., Ruddock L.W., Kelly G., Schmidt J., Williamson, R., and Howard, M.
Solution structure and dynamics of ERp18, a small ER resident oxidoreductase.
Biochemistry (2009) 48, 4596-4606
- Hatahet, F. and Ruddock L.W. Modulating proteostasis: peptidomimetic inhibitors
and activators of protein folding. Curr. Pharm. Des. (2009) 15, 2488-2507
- Saaranen M.J., Salo K.E.H., Latva-Ranta M.K., Kinnula, V.L. and Ruddock, L.W.
The C-terminal active site cysteine of Escherichia coli glutaredoxin 1 determines
the glutathione specificity of the second step of peptide deglutathionylation.
Antioxid. Redox Signal. (2009) 11, 1819-1828
- Karala A., Lappi A.K. and Ruddock L.W. The role of conserved arginine of PDI
in catalysis. FEBS J. (2009) 276, 143-144
- Hatahet F. and Ruddock L.W. Protein disulfide isomerase: A critical evaluation
of its function in disulfide bond formation. Antioxid. Redox Signal. (2009)
- Byrne L.J., Sidhu A., Wallis A.K., Ruddock L.W. Freedman R.B., Howard M.J. and
Williamson R.A. Mapping of the ligand binding site on the b' domain of human PDI:
interaction with peptide ligands. Biochem J. (2009) 423, 209-217
- Wallis A.K., Sidhu A., Byrne L.J., Howard M.J., Ruddock, L.W., Williamson, R.A.
and Freedman, R.B. The ligand-binding b' domain of human protein disulphide-isomerase
mediates homo-dimerization. Protein Sci (2009) 18, 2569-2577