UCH-L1 aggresome formation in response to proteasome impairment indicates a role in inclusion formation in Parkinson’s disease Helen C. Ardley, Gina B. Scott, Stephen A. Rose, Nancy G. S. Tan and Philip A. Robinson Molecular Medicine Unit, University of Leeds, Clinical Sciences Building, St. James’s University Hospital, Leeds, UK Abstract Aggresomes are associated with many neurodegenerative disorders, including Parkinson’s disease, and polyglutamine disorders such as Huntington’s disease. These inclusions commonly contain ubiquitylated proteins. The stage at which these proteins are ubiquitylated remains unclear. A malfunc- tion of the ubiquitin/proteasome system (UPS) may be asso- ciated with their formation. Conversely, it may reflect an unsuccessful attempt by the cell to remove them. Previously, we demonstrated that overexpression of Parkin, a ubiquitin- protein ligase associated with autosomal recessive juvenile Parkinsonism, generates aggresome-like inclusions in UPS compromised cells. Mutations in the de-ubiquitylating enzyme, UCH-L1, cause a rare form of Parkinsonism. We now dem- onstrate that overexpression of UCH-L1 also forms ribbon-like aggresomes in response to proteasomal inhibition. Disease- associated mutations, which affect enzymatic activities, significantly increased the number of inclusions. UCH-L1 aggresomes co-localized with ubiquitylated proteins, HSP70, c-tubulin and, to a lesser extent, the 20S proteasome and the chaperone BiP. Similar to Parkin inclusions, we found UCH-L1 aggresomes to be surrounded by a tubulin rather than a vimentin cage-like structure. Furthermore, UCH-L1 aggre- gates with Parkin and a-synuclein in some, but not all inclu- sions, suggesting the heterogeneous nature of these inclusion bodies. This study provides additional evidence that aggre- gation-prone proteins are likely to recruit UPS components in an attempt to clear proteins from failing proteasomes. Fur- thermore, UCH-L1 accumulation is likely to play a pathological role in inclusion formation in Parkinson’s disease. Keywords: de-ubiquitylating enzyme, inclusion bodies, neu- rodegeneration, Parkinson’s disease, ubiquitin. J. Neurochem. (2004) 90, 379–391. Neurodegenerative disorders such as Parkinson’s (PD), Alzheimer’s (AD) and Huntington’s (HD) diseases are characterized by the selective and symmetric loss of neurones in specific regions of the brain. This loss affects motor, sensory or cognitive systems, causing progressively increas- ing disability leading to the death of the patient. These disorders are pathologically characterized by neuronal pro- teinaceous inclusions of insoluble, unfolded, ubiquitylated polypeptides that fail to be degraded by the 26S proteasome (Lowe et al. 1988; Lang and Lozano 1998; Kaytor and Warren 1999; Sherman and Goldberg 2001). Their apparent stability may, in part, be due to the decreased levels of 26S proteasomal activity, which is associated with increasing age (Goto et al. 2001; McNaught et al. 2002a). The molecular changes underlying the initiation and progression of sporadic idiopathic disease are poorly under- stood. However, genetic studies have provided strong evidence that implicate abnormal processing of a variety of cellular proteins via the ubiquitin/proteasome system (UPS) in these disorders, particularly in the development of PD. Indeed, of the familial forms of PD, a-synuclein, Parkin, UCH-L1 and DJ-1 are associated with the UPS (Wilkinson et al. 1989; Shimura et al. 2000, 2001; Choi et al. 2001; Miller et al. 2003; Tofaris et al. 2003). Parkin and UCH-L1 are components of the system, a-synuclein and DJ-1 are targets. Received December 16, 2003; revised manuscript received March 1, 2004; accepted March 1, 2004. Address correspondence and reprint requests to Dr Helen Ardley, Molecular Medicine Unit, Clinical Sciences Building, St James’s Uni- versity Hospital, Leeds LS9 7TF, UK. E-mail: [email protected]Abbreviations used: AD, Alzheimer’s disease; DUB, deubiquitylating enzyme; E1, ubiquitin-activating enzyme; E2, ubiquitin-conjugating enzyme; E3, ubiquitin-protein ligase; HA, influenza haemagglutinin; HD, Huntington’s disease; PD, Parkinson’s disease; UPS, ubiquitin/ proteasome system. Journal of Neurochemistry , 2004, 90, 379–391 doi:10.1111/j.1471-4159.2004.02485.x Ó 2004 International Society for Neurochemistry, J. Neurochem. (2004) 90, 379–391 379
13
Embed
UCH-L1 aggresome formation in response to proteasome impairment indicates a role in inclusion formation in Parkinson's disease
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
UCH-L1 aggresome formation in response to proteasome
impairment indicates a role in inclusion formation in Parkinson’s
disease
Helen C. Ardley, Gina B. Scott, Stephen A. Rose, Nancy G. S. Tan and Philip A. Robinson
Molecular Medicine Unit, University of Leeds, Clinical Sciences Building, St. James’s University Hospital, Leeds, UK
Abstract
Aggresomes are associated with many neurodegenerative
disorders, including Parkinson’s disease, and polyglutamine
disorders such as Huntington’s disease. These inclusions
commonly contain ubiquitylated proteins. The stage at which
these proteins are ubiquitylated remains unclear. A malfunc-
tion of the ubiquitin/proteasome system (UPS) may be asso-
ciated with their formation. Conversely, it may reflect an
unsuccessful attempt by the cell to remove them. Previously,
we demonstrated that overexpression of Parkin, a ubiquitin-
protein ligase associated with autosomal recessive juvenile
Parkinsonism, generates aggresome-like inclusions in UPS
compromised cells. Mutations in the de-ubiquitylating enzyme,
UCH-L1, cause a rare form of Parkinsonism. We now dem-
onstrate that overexpression of UCH-L1 also forms ribbon-like
aggresomes in response to proteasomal inhibition. Disease-
associated mutations, which affect enzymatic activities,
significantly increased the number of inclusions. UCH-L1
aggresomes co-localized with ubiquitylated proteins, HSP70,
c-tubulin and, to a lesser extent, the 20S proteasome and the
chaperone BiP. Similar to Parkin inclusions, we found UCH-L1
aggresomes to be surrounded by a tubulin rather than a
cin (N-glycosylation inhibitor, which induces the unfolded
protein response), brefeldin A (an inhibitor of movement
from the ER to Golgi apparatus) or carrier (DMSO or
ethanol). Similar to previous studies (reviewed by Lee and
Goldberg 1998), none of the treatments described appeared
to significantly affect cell viability in the time frame of our
experiments as visualized by the number of adherent cells
and the presence of intact nuclei. Furthermore, normal UCH-
L1 expression patterns were observed after exposure to
tunicamycin or brefeldin A. By contrast, UCH-L1 expression
patterns were significantly altered after inhibition of protea-
somal activity with either MG132 or lactacystin treatment. At
least 67% of MG132-treated and 28% of lactacystin-treated
cells contained inclusions (Fig. 1b, left-hand graph). These
inclusions were observed in COS-7 and SH-SY5Y cells with
both HA- and GFP-tagged UCH-L1 constructs (Fig. 1c).
Although inclusions were observed in UCHL1-GFP trans-
fected cells fixed without antibody staining, structures were
less distinct (data not shown). Mock or vector alone
transfected cells did not produce a significant number of
inclusions (data not shown). MG132 treatment caused
redistribution of endogenous UCH-L1, mainly to the
nucleus. However, protein aggregates or inclusions were
not observed in these cells (Fig. 1c, right-hand panel). To
confirm that the elevated number of inclusions observed in
MG132-treated cells compared with lactacystin-treated cells
was not due to inhibition of other proteases such as calpains,
we performed additional experiments using the calpain
inhibitor, calpeptin, in the presence or absence of lactacystin
(Fig. 1b, right-hand panel). Calpeptin treatment did not
produce inclusions within cells. Furthermore, addition of
calpeptin and lactacystin to cells did not increase the number
of inclusions observed (22% in lactacystin-treated cells
compared with 21% in cells treated with both inhibitors). As
the synthetic peptide inhibitor, MG132, and the natural
inhibitor, lactacystin, but not the calpain inhibitor, calpeptin,
caused UCH-L1 inclusion formation, the data suggest that
these protein aggregates resulted from an impairment of
proteasome function, and that the number of inclusions in
MG132 cells is unlikely to be a consequence of additional
protease inhibition.
We did not observe any alteration in the levels of either
UCHL1-HA protein or endogenous UCH-L1 after exposure
to any of the drugs tested (Fig. 1d, and data not shown).
The majority of MG132-induced UCH-L1 inclusions were
found in the cytoplasm (Fig. 2a, left-hand panel). However, a
small number of cells (< 10%) displayed either nuclear
(Fig. 2a, middle panel), or cytoplasmic and nuclear (Fig. 2a,
right-hand panel), staining patterns. Detailed analysis of
these inclusions revealed a ‘honeycomb’ or ‘chain-like’
structure (Fig. 2b).
The solubility of UCH-L1-HA inclusions in COS-7 cells
was also assessed (Fig. 2c). Prior to proteasome inhibition,
UCH-L1 was detected exclusively in the soluble fraction of
cell lysates (Fig. 2c). However, a substantial fraction of
UCH-L1 was redistributed to the insoluble fraction after the
inhibition of proteasomal activity (Fig. 2c). High molecular
weight species, characteristic of the presence of polyubiquit-
ylated products, were not observed in the insoluble fraction
(Fig. 2c).
The familial Parkinson’s disease-causing I93M mutation
produces inclusions without inhibition of proteasomal
activity
We next analysed whether UCH-L1 point mutations I93M
and S18Y caused changes to its cellular distribution. I93M is
proposed to be responsible for disease in a German PD
family (Leroy et al. 1998), whereas the S18Y polymorphism
may be protective against PD (Maraganore et al. 1999;
Zhang et al. 2000a; Elbaz et al. 2003).
In the absence of MG132, overexpression of the UCH-L1
I93M mutant produced significantly more inclusions than
either wild-type or the S18Y polymorphism (50% compared
with 7% and 11%, respectively, p < 0.001) (Fig. 3a and
Table 1). In the presence of MG132, both I93M and S18Y
gave higher levels of inclusions compared with those cells
overexpressing native UCH-L1. Indeed, inclusions were
observed in 99% of cells transfected with the I93M mutant.
The overwhelming majority of inclusions produced in cells
transfected with either I93M or S18Y were observed in the
cytoplasm (Table 1). Nuclear inclusions were not observed
with the S18Y polymorphism. No significant difference in
cell viability was observed with the different constructs.
Both S18Y and I93M UCH-L1 mutant proteins were
detected exclusively in the soluble fraction of cell lysates
(Fig. 3b). By contrast, both S18Y and I93M UCH-L1 mutant
proteins were detected in the insoluble fraction when
proteasomal activity was inhibited. These results mirror those
observed with wild-type UCH-L1. There was no obvious sign
of high molecular weight products either in the presence or
absence of proteasome inhibitor (data not shown).
As the S18Y polymorphism is thought to be protective, we
also investigated whether co-transfection of S18Y with the
I93M mutant would reduce the number of inclusions formed
in the co-transfected cells compared with I93M alone
UCH-L1 aggresome formation 383
� 2004 International Society for Neurochemistry, J. Neurochem. (2004) 90, 379–391
(Fig. 3c). The double transfection appeared to reduce
significantly the number of inclusions observed (p < 0.01),
although the numbers were not reduced compared with those
of S18Y alone.
Wild-type and mutant forms of UCH-L1 are
immunoprecipitated with polyubiquitylated
species when proteasomal activity is inhibited
We investigated whether insoluble UCH-L1 aggregates were
ubiquitylated as it had previously been noted that many
aggregation-prone proteins are ubiquitylated when protea-
somal activity is inhibited. HA-tagged, normal and mutant
UCH-L1 constructs were immunoprecipitated using HA
antibodies from cell lysates prepared from transfected cells
(Fig. 3d). High molecular weight ubiquitylated species were
observed after immunoprecipitation of all three HA-tagged
UCH-L1 constructs. By contrast, they were not observed
either in the absence of MG132, or in mock or vector alone
transfected cells (Fig. 3d, top panel left-hand side). Further-
more, monoubiquitylated UCH-L1 was not detected. The
presence of multiple ubiquitin species in total cell lysates
indicated that lack of monoubiquitylated species was not due
to a lack of ubiquitin availability in these samples (Fig. 3d,
top panel right-hand side).
Stripping and re-probing of the blot with antibodies to
UCH-L1 (Fig. 3d middle panels) or HA (Fig. 3c lower
panels) confirmed the specificity of the immunoprecipitation.
Elevated levels of the UCH-L1 constructs in the immuno-
precipitated samples after incubation with MG132 were
observed. This increase was not apparent in non-immuno-
precipitated lysates.
Co-localization of UPS and chaperone system
components to UCH-L1 inclusions
Many neurological-associated inclusions are ubiquitin-posit-
ive and stain for other components of the UPS and chaperone
system (Lowe et al. 1988; Pollen et al. 1993; Saudou et al.
1998; Cummings et al. 1999). Therefore, we analysed the
expression patterns of UCH-L1 and other components of the
UPS or chaperone system in cells overexpressing UCH-L1 in
which proteasomal activity was inhibited (Fig. 4). Prior to
proteasome inhibition, all UPS and chaperone components
localized to regions as previously described (Bush et al.
1997; Waelter et al. 2001; Ardley et al. 2003; Muqit et al.
2004). UCH-L1 co-localized with ubiquitin throughout the
inclusion (Fig. 4a). Indeed, consistent with our immunopre-
cipitation data, co-localization of UCH-L1 with an antibody
that recognizes only ubiquitin conjugates (Fujimuro et al.
1994) indicates that the inclusions comprise ubiquitylated
proteins rather than free ubiquitin moieties (Fig. 4b). A
predominant nuclear staining pattern was observed with anti-
20S proteasomal antibodies when proteasomal activity was
inhibited. Only weak 20S staining was localized within the
UCH-L1 inclusions (Fig. 4c).
Fig. 2 Characteristics of UCH-L1 inclusions. COS-7 cells were
transfected with UCHL1-HA. At 28 h post-transfection, cells were
incubated in the presence (a-c) or absence (c) of 5 lM MG132 for
16 h. (a) and (b) Cells were fixed and stained as described in Fig. 1.
(a) UCH-L1-induced inclusions appear to localize to the cytoplasm
(left-hand panel), nuclei (middle panel) or both (right-hand panel).
Arrows indicate large cytoplasmic inclusions; arrowheads, nuclear
inclusions. (b) Magnification of the aggregated UCH-L1 indicates a
‘honeycomb’-like structure. Boxed area in left-hand panel is magni-
fied ·3 in right-hand panel. Scale bars 20 lm. (c) Inhibition of pro-
teasomal activity leads to the formation of insoluble UCH-L1
aggregates. Protein lysates were prepared as described in Fig. 1. A
10 lg aliquot of each soluble (S) and insoluble (I) fraction was sep-
arated by SDS-PAGE and analysed by western blotting with either
mouse anti-HA (upper panel), or anti-b-Actin (lower panel), antibodies.
The positions of individual molecular mass markers are indicated.
384 H. C. Ardley et al.
� 2004 International Society for Neurochemistry, J. Neurochem. (2004) 90, 379–391
Proteasome inhibition can cause an increase in the level of
chaperones including BiP and HSP70 (Kuznetsov et al. 1996;
Bush et al. 1997; Ardley et al. 2003). Moreover, BiP and
HSP70 have been observed in PD-related inclusions (Junn
et al. 2002; Muqit et al. 2004; Tanaka et al. 2004). Therefore,
we investigated whether BiP or HSP70 were also found in
UCH-L1 inclusions. BiP co-localized with UCH-L1 within
inclusions but in discrete foci (Fig. 4d, arrowheads indicate
regions of co-localization). HSP70 co-localized with UCH-L1
within inclusions, although HSP70 was also observed within
the nucleus of the cells (Fig. 4e). HSP70 containing inclu-
sions were also observed in untransfected cells after MG132
treatment (Fig. 4e, cell indicated by an asterisk).
Parkin and a-synuclein co-localization in UCH-L1
inclusions
Both a-synuclein and Parkin are found in Lewy bodies in
idiopathic PD. Therefore, we investigated whether these
proteins co-localized within UCH-L1 inclusions. A signifi-
cant proportion of endogenous a-synuclein was found to co-localize with inclusions in cells overexpressing UCH-L1
(Fig. 5a).
Previously, we were unable to detect endogenous Parkin
in COS-7 cells (Ardley et al. 2003). Therefore, overexpres-
sion of UCHL1-HA and FLAG-Parkin was performed to
analyse any regions of co-localization within inclusions
(Fig. 5b). Most cells expressing both proteins generated
inclusions that contained both UCH-L1 and Parkin
co-localizing in the aggregate [Fig. 5b(i)]. However, a small
proportion of these cells contained inclusions where each
protein localized to different areas. For example, in
Fig. 5b(ii), UCH-L1 staining was found within the central
core region of the inclusion with partial Parkin
co-localization. By contrast, the majority of Parkin staining
was found at the periphery of the inclusion. Of interest, we
also noted the presence of a few cells that contained
aggregates of one protein only, despite the overexpression
of both. For example, in Fig. 5b(iii), Parkin expression was
observed throughout the cytoplasm despite the presence of
UCH-L1 inclusions.
Aggresome characteristics of UCH-L1 inclusions
We wished to establish whether UCH-L1 aggregates had
aggresome-like characteristics similar to those observed
previously with Parkin (Fig. 5b and Ardley et al. 2003).
First, we investigated whether UCH-L1 inclusions displayed
the microtubule-dependent properties characteristic of classi-
cal aggresomes (Fig. 6a) (Johnston et al. 1998; Garcia-Mata
et al. 1999). COS-7 cells overexpressing UCHL1-GFP were
treated with MG132, the microtubule disruption reagent,
nocodazole, or both, prior to fixation and immunostaining.
Co-staining with a-tubulin antibodies was performed in orderto confirm disruption of microtubules. Prior to drug treatment,
UCH-L1 displayed a mostly cytoplasmic localization. It did
not co-localize with a-tubulin. By contrast, in the presence ofMG132, bundling of a-tubulin around the aggregate was
observed in the majority of cells containing UCH-L1
(a)
(c)
(b)
(d)
Fig. 3 Effect of Parkinson’s disease-associated UCH-L1 mutations on
inclusion formation. COS-7 cells were transfected with UCHL1-HA
(WT) or HA tagged mutant protein (S18Y, I93M). At 28 h post-trans-
fection, cells were incubated in the presence or absence of 5 lM
MG132 for 16 h as indicated. (a) I93M overexpression causes inclu-
sion formation in the presence or absence of MG132. The expression
of UCH-L1 was analysed by immunofluorescence using anti-HA anti-
bodies. Cells were scored for the presence of inclusion bodies. Light
grey bars indicate untreated cells; dark grey bars indicate MG132-
treated cells. Error bars indicate the standard error from the mean.
Asterisk(s) indicate a significant difference between the number of
inclusions observed with each construct compared with wild-type
UCHL1-HA in the presence or absence of MG132: *p < 0.05;
**p < 0.001. (b) UCH-L1 and its mutants form insoluble inclusions only
in the presence of proteasome inhibitors. Cell lysates were prepared
and analysed as described for Fig. 2. Blots were probed with mouse
anti-HA (upper panel) and anti-b-Actin (lower panel) antibodies.
(c) Overexpression of S18Y with I93M may protect against inclusion
formation. The number of cells containing inclusions was scored as
described in (a) in cells expressing either S18Y, I93M or both (S18Y/
I93M). Error bars indicate the standard error from the mean. Aster-
isk(s) indicate a significant difference between the number of inclu-
sions observed with each construct compared with I93M in the
absence of MG132: *p < 0.01; **p < 0.001. (d) UCH-L1 and its
mutants associate with polyubiquitylated species in proteasomal
inhibited cells. Cell lysates were immunoprecipitated using anti-HA
antibodies and size fractionated by SDS PAGE. A 20 lL aliquot of
each total cell lysate was also analysed in parallel (no IP). The
membrane was probed with anti-ubiquitin antibodies (a-Ub). The
success of the immunoprecipitation was ascertained by western
blotting with anti-HA or anti-UCH-L1 antibodies (lower panels).
Endogenous UCH-L1 expression was observed in the total cell lysates
from the mock and vector alone samples. The positions of individual
molecular mass markers are indicated.
UCH-L1 aggresome formation 385
� 2004 International Society for Neurochemistry, J. Neurochem. (2004) 90, 379–391
inclusions (Fig. 6a, arrowheads). In nocodazole-treated cells,
UCH-L1 inclusions were not detected in the majority of cells.
Conversely, UCH-L1 inclusion bodies were observed in cells
overexpressing UCH-L1 after the addition of both MG132
and nocodazole (Fig. 6a).
Aggresome formation routinely causes disruption of
c-tubulin within cells (Johnston et al. 1998; Garcia-Mata
et al. 1999). UCH-L1 transfected cells were cultured in the
presence of MG132 and stained with c-tubulin antibodies inorder to assess the effect of UCH-L1 aggregates on c-tubulindistribution. UCH-L1 co-localized with c-tubulin at the
centrosome in some cells, but not others [Fig. 6b, compare
the lower cell where the yellow staining indicates that
UCH-L1 co-localizes with c-tubulin to the centrosome
(arrowheads) with the upper cell where UCH-L1 is found
around the centrosome as indicated by red c-tubulin stainingand green UCH-L1 staining].
Aggresomes (Johnston et al. 1998; Garcia-Mata et al.
1999) and aggresome-like inclusions (Meriin et al. 2001;
Lee et al. 2002) are often surrounded by a vimentin cage-like
structure. Co-immunostaining of UCH-L1 transfected cells
with anti-GFP and anti-vimentin antibodies (Fig. 6c)
revealed disruption of vimentin filament staining, but little
co-localization of UCH-L1 and vimentin.
Discussion
The purpose of this study was twofold. We wished first, to
address the question of whether mutants of UCH-L1 are
prone to aggregate into inclusions and second, to identify the
impairment or activation of which cellular function(s) may be
responsible for UCH-L1 inclusion formation. Such informa-
tion may help to explain its presence in Lewy bodies in
idiopathic and hereditary forms of PD.
Inhibition of ER to Golgi transport, or induction of the
unfolded protein response, did not result in UCH-L1
inclusion formation. By contrast, inclusions were formed
in cells overexpressing UCH-L1 when the proteasomal
activity was inhibited. These inclusions were predominantly
cytoplasmic, ubiquitylated and insoluble structures with a
chain-like appearance. MG132 produced far more inclusions
than treatment of cells with lactacystin. However, the
number of inclusions in lactacystin-treated cells did not
increase to levels observed with MG132 when together with
the calpain inhibitor, calpeptin. These data suggested that
the elevated number of inclusions with MG132 is unlikely
to be due to inhibition of other proteases such as calpain.
This is perhaps not surprising given that inhibition of such
protease activities requires at least a 10-fold higher
concentration of MG132 than for the proteasome (Tsubuki
et al. 1996). MG132 is the most potent inhibitor of the
proteasome (Bush et al. 1997; Kisselev and Goldberg
2001). This is likely to explain the apparent discrepancies
in the potency of MG132 and lactacystin to generate
inclusions (67% of cells contained inclusions compared
with 28%). MG132, unlike lactacystin, is able to enter cells
rapidly and it does not require transition to an active form.
These properties allow MG132 to be used at lower
concentrations and its effects are seen more rapidly than