Download Redox rescues virus from eR trap - Institute of Molecular Life Sciences

news and views
The authors have investigated the epigenetic memory of an active state and propose
a role for H3.3 in this process. How the epigenetic memory of silent chromatin is handled will be the next item on the agenda,
and is likely to be equally important for
reprogramming efficiency. Does memory of
silent states occur as a default mechanism in
the absence of H3.3 simply because no transcription priming event can occur? It may
thus be important to further analyse the
links between replication and transcription
reprogramming. In this context, it would be
interesting to examine whether the high replication reprogramming promoted by mitotic
extracts12 is paralleled by a higher incorporation of replicative H3. With the cellular
systems now available13,14,15, analysis of the
dynamics of histone H3 variants will surely
be important to fully understand the network
of interactions and find ways to manipulate
the system.
Although the study by Ng and Gurdon raises
several questions, it reinforces the importance
of histones as a possible means of conveying
epigenetic information. These advances, combined with other reprogramming techniques
such as cellular fusions13 or recently developed
cellular models in which a limited set of factors is expressed in somatic cells14,15, should
extend our basic understanding of epigenetics
and open avenues for the treatment of diseases
such as cancer, and for the development of cell
replacement therapies.
1. Ng, R. K. & Gurdon, J. B. Nature Cell Biol. 10, 102–
109 (2008).
2. Tsunoda, Y. & Kato, Y. Differentiation 69, 158–161
(2002).
3. Gurdon, J. B. J Embryol. �������������
Exp. Morphol. 10, 622–640
(1962).
4. Santos, F. et al. Curr Biol.13, 1116–1121 (2003).
5. Ng, R. K. & Gurdon, J. B. Proc. Natl Acad. Sci. USA
102, 1957–1962 (2005).
6. Sarma, K. & Reinberg, D. Nature Rev. Mol. �����������
Cell. Biol.
6, 139–149 (2005).
7. Loyola, A. & Almouzni, G. Trends Biochem. Sci.
���� 32,
425–433 (2007).
8. van der Heijden, G. W. et al. Mech. Dev. 122, 1008–
1022 (2005).
9. Yang, X. et al. Nature Genet. 39, 295–302 (2007).
10.Klose, R. J. & Zhang, Y. Nature Rev. Mol. Cell. Biol. 8,
307–318 (2007).
11.Wysocka, J. et al. Cell 121, 859–872 (2005).
12.Lemaitre, J. M., Danis, E., Pasero, P., Vassetzky, Y. &
Mechali, M. Cell 123, 787–801 (2005).
13.Blau, H. M., Chiu, C. P. & Webster, C. Cell 32, 1171–
1180 (1983).
14.Meissner, A., Wernig, M. & Jaenisch, R. Nature
Biotechnol. 25, 1177–1181 (2007).
15.Takahashi, K. et al. Cell 131, 861–872 (2007).
Redox rescues virus from ER trap
Christoph J. Burckhardt and Urs F. Greber
How viruses manage to resist physical and chemical stress and yet open their protective coats during cell infection has been a
longstanding, fundamental question. A study with the DNA tumour virus SV40 now shows that protein folding and quality-control
factors of the endoplasmic reticulum reshuffle disulfide bonds within the viral capsid, providing a molecular mechanism for the
exit of infectious virions from the endoplasmic reticulum.
The concept of a viral replication cycle is
exceedingly simple, yet its mechanisms hold
many secrets. Upon engulfment of a virus by
the host cell, uncoating of the viral genome
from the protective capsid is required for the
activation of viral genes, using the transcriptional machinery of the host cell to drive synthesis of viral progeny. Newly formed particles
exit the cell, infect new cells, and the cycle starts
again. On their way into a cell, animal viruses
pass through cellular compartments where they
encounter triggers that initiate viral coat destabilization. Examples of destabilization events
include virus–receptor interactions, shedding
of minor capsid proteins, or the low pH in the
digestive tract or in endosomal vesicles activating viral or cellular proteases that catalyse
limited proteolysis1. In the past, therapeutic
Christoph J. Burckhardt & Urs F. Greber are at
the Institute of Zoology, University of Zurich,
Winterthurerstrasse 190, CH-8057 Zürich,
Switzerland.
email: [email protected]
inhibition of uncoating has been a prominent
strategy in the development of anti-viral drugs,
particularly for non-enveloped viruses, such as
picornaviruses2. Although in principle successful, the strategy has been largely abandoned due
to the rapid appearance of viral escape mutants
in patients treated with such antiviral drugs.
Development of therapeutics directed against
the host may be an interesting alternative.
A recent paper by Schelhaas et al.3 reveals
new host functions required for the infectious
entry of the polyoma virus Simian Virus 40
(SV40). Polyoma viruses are non-enveloped
DNA viruses that cause cancer in rodents, and
traces of their genome can be found in human
tumours. Notably, SV40 was accidentally inoculated into millions of humans during vaccinations with chemically inactivated poliovirus
in the 1950s when the vaccine virus was produced in SV40-infected monkey cells. The pas����
sage of SV40, mouse polyomavirus (mPy) and
human BK virus into cells is intriguing, given
that these viruses use glycolipids as their cell
nature cell biology volume 10 | number 1 | JANUARY 2008 © 2008 Nature Publishing Group
attachment sites4. Polyoma viruses are assembled in the nucleus under reducing conditions,
and their capsids become oxidized outside the
cell. The spatial separation of viral assembly
and disassembly explains in part why a virus
can be stably assembled in an infected cell and
disassembled on entry into a naïve cell5.
The work by Schelhaas et al.3 shows how a
virus uses the protein folding and quality-control apparatus for uncoating and membrane
translocation. It provides a molecular explanation for two long-standing observations,
namely that the incoming SV40 is found in
large amounts inside the ER6 and that infectious particles are present in the cytosol7. The
SV40 capsid has icosahedral symmetry and is
composed of 72 homopentamers of the major
capsid protein VP1. Twelve pentamers are
five-coordinated, the others are six-coordinated. It is unusually structured because the
VP1 proteins are linked together by a network
of interchain Cys9–Cys9 and Cys104–Cys104
disulfide bonds (Fig. 1).The carboxy-terminal
news and views
a
ERp57
Caveolar
endocytosis
ERAD
PDI
Derlin-1
Sel1L
Ca2+ high
ER
b
Cys
104
Cys9
Cys9
Cys
104
Cys104
ERp57
Cys104
Cytosol
Ca2+ low
Cys104
Cys9
Cys9
Cys104
Figure 1 ERp57 and low calcium are involved in SV40 pentamer dissociation. (a) Infectious entry
of SV40 into human and monkey cells occurs via caveolin- and lipid raft-dependent endocytosis on
binding to the GM1 ganglioside receptors at the plasma membrane13,14. From the caveosome, the
viruses are transported inside vesicles along microtubules to the ER where they accumulate 15. A
cryo-electron microscopy structure of intact SV40 capsids is shown in blue; reduced and calciumdepleted particles that released the pentavalent pentamers of the major capsid protein VP1 are shown
in yellow3. The ER thiol-disulfide oxidoreductase ERp57 isomerises SV40 interchain disulfide bonds.
PDI, Derlin-1 and Sel1L — three proteins involved in ER-associated degradation (ERAD) — are then
thought to retrotranslocate particles to the cytosol, where the low calcium concentration leads to
pentamer dissociation. Cytosolic particles then enter the nucleus via the nuclear pore complex, where
viral transcription and replication take place. (b) Model of SV40 disulfide isomerisation mediated by
ERp57. Individual monomers in a five-coordinated pentamer (blue) of the major capsid protein VP1
are connected to neighbouring six-coordinated pentamers (orange and green) by interchain Cys9–Cys9
and Cys104–Cys104 disulfide bonds (possible bonds are displayed by dashed lines). During disulfide
bond isomerisation, intrachain Cys9–Cys104 disulfide bonds are formed in the five-coordinated
pentamer of the VP1 molecule whereas the six-coordinated pentamers are connected by Cys104–
Cys104 disulfide bonds. This isomerisation reaction uncouples the five-coordinated pentamer from the
disulfide bond network of the virus capsid, which may provide a signal for recognition of the particle by
the ERAD machinery for translocation to the cytosol.
peptide of VP1 extends to binding sites on VP1
molecules of neighbouring pentamers that are
stabilized by calcium ions8. In the life cycle of
polyoma viruses, progeny capsids are assembled in the nucleus under reducing conditions
and oxidized outside the cell. They do not come
in contact with the enzymatic redox system in
the ER; the site of viral uncoating.
The initial trigger for the research by
Schelhaas et al. was the observation that in
SDS polyacrylamide gels, treatment of nonreduced virus with alkylating agents yields VP1
multimers indicative of intersubunit disulfide
bonds. These are not observed in the absence
of alkylation, where free cysteines are available for reshuffling the interchain disulfides
of VP1. Moreover, alkylation of virus blocked
10
infection and alkylated virus did not release
VP1, indicative of a requirement for disulfide
isomerisation rather than disulfide reduction
for productive infection. Depletion of the two
ER thiol–disulfide oxidoreductases, ERp57 and
the closely related protein disulfide isomerase
(PDI), by small-interfering RNA (siRNA) led
to a significant reduction of infection, but only
the depletion of ERp57 inhibited VP1 release
in vivo, indicating that only ERp57 resolves
the interchain disulfide bonds. ERp57 is an
essential soluble protein of the ER, where it
catalyses disulfide reduction, isomerization
and dithiol oxidation in substrate proteins.
These findings are complemented by an in
vitro assay where both purified ERp57 and PDI
were able to induce the loss of VP1 on isolated
viruses. To determine whether ERp57 and PDI
serve as isomerases or reductases, alkylated
(non-isomerizing) virus was used. Following
treatment with ERp57, alkylated SV40 capsids did not release VP1 monomers although
PDI was able to release VP1, suggesting that
ERp57 functions as an isomerase and PDI as
a reductase. However, native SV40 released
VP1 following ERp57 treatment, supporting
the siRNA knockdown data. The disulfide
isomerization activity of ERp57 is needed to
resolve interchain disulfide bonds of the pentamers, yielding the intrachain disulfide bond
Cys9–Cys104 (Fig. 1). Nevertheless, isomerization of disulfide bonds is not sufficient for
the dissociation of the pentamers; the loss of
the VP1-associated Ca2+ ions is also required.
Cryo-electron microscopy of partially uncoated
viruses revealed that it was the five-coordinated pentamers that were released, suggesting that they are less tightly associated than
the six-coordinated pentamers. Interestingly,
the Cys9–Cys104 disulfide bond is already
in place in native mouse polyoma virus VP1,
which has no interchain disulfide bonds and
hence does not require the Erp57 activity for
infection9. The conformational changes around
the disulfides of SV40 VP1 weaken the capsid
such that the myristoylated amino-terminal
region of VP2 may swing out, and potentially
positions the capsid to the lumenal face of the
ER membrane. Electron micrographs suggest
that SV40 is proximal to the lumenal ER membrane6. A similar principle of conformational
change is used by mPy, where in vitro experiments have shown that the thioredoxin-like
ERp29 is required for membrane association
of the virion9.
But how is a nucleo-protein complex of
50 nm in diameter translocated through the
ER membrane? The observation that proteasome inhibitors block SV40 entry suggests
that ER-associated degradation (ERAD) may
be involved 3. ERAD requires chaperones that
recognize misfolded proteins and the retrotranslocation machinery10. Cholera toxin, for
example, requires reduced PDI for unfolding
and presentation to the ERAD machinery.
Accordingly, knockdown of PDI, Derlin-1 and
Sel1L reduced SV40 infection, suggesting a role
of ERAD in SV40 retrotranslocation. Derlin-1
is a putative pore-forming component, whereas
Sel1L may be required for substrate recognition. The sensor that recognizes the disulfidereshuffled virus in the ER has not been
identified, but may involve PDI or the signal
nature cell biology volume 10 | number 1 | JANUARY 2008
© 2008 Nature Publishing Group
news and views
peptide peptidase, perhaps recognizing VP2
(ref. 10). This sensor may target the virus to
the retrotranslocation complex. The rather low
efficiency of SV40 release from the ER into the
cytosol (estimated to be 1%) may explain why it
remains unclear whether the translocated virus
particles are ubiquitinated, a modification that
normally occurs on proteins that are targeted
to ERAD. Nevertheless, the translocated virus
with isomerized disulfides sheds the VP1 pentamers in the low-calcium conditions of the
cytosol. This poises the particle for the release
of infectious DNA into the nucleus, as shown
for other DNA viruses11.
One can speculate that the co-option of
the ERAD machinery by SV40 occurs in
conjunction with biosynthetic functions of
the ER membranes, such as lipid synthesis or
lipid droplet formation12. Viral escape from
the ER is, however, reminiscent of viral gene
products that co-opt the ERAD machinery
to induce rapid degradation of host receptors
or of immune surveillance molecules, such
as major histocompatibility complex (MHC)
class I proteins10. The ER redox system contains more than fifteen different disulfide isomerases; other viruses, such as the papilloma
viruses or picornaviruses, that visit the ER or
interfere with ER functions, may use a similar
pathway as SV40, perhaps by engaging different isomerases. Future studies will uncover
additional host factors, and these may be useful
targets for anti-viral therapies as they are not
subject to viral resistance mutations.
1. Guglielmi, K. M., Johnson, E. M., Stehle, T. & Dermody,
T. S. Curr. Top.
����������������
Microbiol. Immunol.
�������� 309, 1–38
(2006).
2. Li, C. et al. Curr. Med. Chem. 14, 847–856 (2007).
3. Schelhaas, M. et al. Cell 131, 516–529 (2007).
4. White, M. K. & Khalili, K. Virology 324, 1–16
(2004).
5. Greber, U. F., Singh, I. & Helenius, A. Trends Microbiol.
2, 52–56 (1994).
6. Kartenbeck, J., Stukenbrok, H. & Helenius, A. J. Cell
Biol. 109, 2721–2729 (1989).
7. Greber, U. F. & Kasamatsu, H. Trends Cell Biol. 6,
189–195 (1996).
8. Stehle, T., Gamblin, S. J., Yan, Y. & Harrison, S. C.
Structure 4, 165–182 (1996).
9. Magnuson, B. et al. Mol. Cell 20, 289–300 (2005).
10.Loureiro, J. & Ploegh, H. L. Adv. Immunol. 92, 225–
305 (2006).
11.Greber, U. F. & Way, M. Cell 124, 741–754 (2006).
12.Ploegh, H. L. Nature 448, 435–438 (2007).
13.Tsai, B. et al. Embo J. 22, 4346–4355 (2003).
14.Smith, A. E., Lilie, H. & Helenius, A. FEBS Lett. 555,
199–203 (2003).
15.Pelkmans, L., Kartenbeck, J. & Helenius, A. Nature Cell
Biol. 3, 473–483 (2001).
Cilia put a brake on Wnt signalling
Xi He
Two studies suggest that the primary cilium, a microtubule-based structure protruding from the surface of most vertebrate cells,
has a role in restraining Wnt/β-catenin signalling. These findings have implications for the pathogenesis of a plethora of diseases
associated with abnormal cilia; however, the mechanism linking Wnt signalling and cilia remains a mystery.
Signalling by the Wnt family of secreted growth
factors is important in cell proliferation and
differentiation during development1,2. Wnt proteins engage multiple receptors and activate a
number of distinct signalling pathways3,4. These
include the canonical Wnt/β-catenin pathway,
which controls stability of the transcription coactivator β-catenin (and therefore gene expression)1–3, and the non-canonical Wnt/planar cell
polarity (PCP) pathway, which primarily governs cytoskeletal architectures associated with
cell polarity and movement4. Defects in Wnt
signalling pathways are associated with human
cancers and other diseases1,2.
The primary cilium consists of a microtubule-based axoneme enclosed by a specialized
plasma membrane5,6. The cilium protrudes
into the extracellular space and is tethered to
the cell by the basal body, a centriolar-based
Xi He is at the F. M. Kirby Neurobiology Center,
Children’s Hospital Boston, Harvard Medical School,
Enders 461.2, 300 Longwood Avenue, Boston,
Massachusetts 02115, USA.
e-mail: [email protected]
structure that forms the interface between the
cilium and the cytoplasm5,6 (Fig. 1). The formation and function of the cilium requires
intraflagellar transport (IFT), which shuttles
molecules in and out of the cilium via IFT proteins and microtubule-based motor complexes,
including Kinesin-2 and Dynein5,6. The cilium
proteome consists of hundreds to a thousand
polypeptides5, reflecting diverse cilia functions
that include cell motility, chemo- and mechano-sensing, and signal transduction5,6. Many
human diseases, collectively referred to as
ciliopathies, are associated with defective cilia
formation and/or function, such as polycystic
kidney disease (PKD), nephronophthisis, primary cilia dyskinesia, situs inversus (reversal
of left-right asymmetry), and Bardet-Biedl
Syndrome (BBS)5,6.
A critical role of the cilium in cell signalling
was discovered in the study of neural patterning by Sonic Hedgehog in mice7. The role of
cilia in Hedgehog signalling is unique to vertebrates. Cilia are present in most vertebrate
cells, whereas in Drosophila Melanogaster,
nature cell biology volume 10 | number 1 | JANUARY 2008 © 2008 Nature Publishing Group
only sperm and sensory neurons exhibit cilia.
Hedgehog signalling therefore occurs independently of cilia in flies, supporting the idea
that Hedgehog signalling uses both common
and distinct mechanisms in Drosophila and
vertebrates7. By comparison, the Wnt signalling pathway is highly conserved between flies
and vertebrates1–4. Given that cilia cannot have
a widespread role in Drosophila, the findings
by Katsanis and colleagues and Reiter and colleagues that cilia restrain Wnt/β-catenin signalling in vertebrates is surprising8, 9.
Earlier studies have implicated ciliary proteins
in Wnt and PCP signalling10,11. Inversin (Inv),
mutations of which cause nephronophthisis
characterized by renal cysts and situs inversus,
was found to inhibit Wnt/β-catenin signalling10.
However, Inv protein localization to other cellular compartments in addition to cilia complicates
the interpretation that its ciliary role is directly
implicated in the control of Wnt signalling. Inv is
also required for Xenopus gastrulation10, a process governed by Wnt/PCP signalling4. Inv binds
to and promotes the degradation of cytosolic,
11