Analysis of oxidative events induced by expanded
polyglutamine huntingtin exon 1 that are differentially
restored by expression of heat shock proteins or treatment
with an antioxidant
Wance J. J. Firdaus
1
, Andreas Wyttenbach
2
, Chantal Diaz-Latoud
1
, R. W. Currie
1,3
and Andre
´
-Patrick Arrigo
1
1 Laboratoire Stress Oxydant, Chaperons et Apoptose, Centre de Ge
´
ne
´
tique Mole
´
culaire et Cellulaire, Universite
´
Claude Bernard Lyon-1,
Villeurbanne, France
2 Southampton Neuroscience Group, School of Biological Sciences, University of Southampton, UK
3 Department of Anatomy and Neurobiology, Dalhousie University, Halifax, Canada
Neuronal selective loss and formation of intraneuron-
al protein aggregates are characteristics of Hunting-
ton’s disease (HD), which is one of more than 10

´
dex, France
Fax: +33 472 440555
Tel: +33 472 432685
E-mail:
(Received 16 February 2006, revised
20 April 2006, accepted 12 May 2006)
doi:10.1111/j.1742-4658.2006.05318.x
We recently reported that the transient expression of polyglutamine tracts
of various size in exon 1 of the huntingtin polypeptide (httEx1) generated
abnormally high levels of intracellular reactive oxygen species that directly
contributed to cell death. Here, we compared the protection generated by
heat shock proteins to that provided by the antioxidant agent N-acetyl-l-
cysteine. In cells expressing httEx1 with 72 glutamine repeats (httEx1-72Q),
the overexpression of Hsp27 or Hsp70 plus Hdj-1(Hsp40) or treatment of
the cells with N-acetyl-l-cysteine inhibited not only mitochondrial mem-
brane potential disruption but also the increase in reactive oxygen species,
nitric oxide and protein oxidation. However, only heat shock proteins and
not N-acetyl-l-cysteine reduced the size of the inclusion bodies formed by
httEx1-72Q. In cells expressing httEx1 polypeptide with 103 glutamine
repeats (httEx1-103Q), heat shock proteins neither decreased oxidative
damage nor reduced the size of the inclusions. In contrast, N-acetyl-l-cys-
teine still efficiently decreased the oxidative damage induced by httEx1-
103Q polypeptide without altering the inclusions. N-Acetyl-l-cysteine was
inactive with regard to proteasome inhibition, whereas heat shock proteins
partially restored the caspase-like activity of this protease. These observa-
tions suggest some relationships between the presence of inclusion bodies
and the oxidative damage induced by httEx1-polyQ.
Abbreviations
DCFH-DA, 2¢,7¢-dichlorofluorescein diacetate; 2,4-DNPH, 2,4-dinitrophenyl hydrazine; EGFP, enhanced green fluorescent protein; FCCP,

motes the clearance of mutant protein by activating
autophagy [18,19].
Intracellular aggregates containing ubiquitylated
proteins are a prominent cytopathologic feature of
most neurodegenerative disorders. For example, aggre-
gated htt-polyQ in neuronal inclusions of HD mice
and HD patients appears to be ubiquitylated [20]. The
accumulation of ubiquitylated abnormal proteins
results in the formation of pathologic aggregates that
perturb the normal physiology of neurons and lead to
proteotoxicity. The ubiquitin-26S proteasome system
(UPS), which normally degrades short-lived and
abnormal proteins, is probably recruited to eliminate
the pathologic aggregates formed by ubiquitylated htt-
polyQ [21,22]. However, this degradation is likely to
be far from complete [23], because the proteasome can-
not digest polyglutamine sequences and release them
during degradation of polyglutamine-containing pro-
teins [24]. This may interfere with proteasome function
and help explain why long polyQ expansions promote
early disease onset.
Elevated levels of oxidative damage at the level of
DNA, lipids and proteins are evident in numerous neu-
rodegenerative disorders, including Alzheimer’s disease
and HD, suggesting that oxidative stress is inherent to
these neuronal degenerations [25–29]. Recently, we and
others reported that the expression of the expanded
httEx1-polyQ gene product generated mitochondrial
complex IV deficiency, elevated reactive oxygen species
(ROS) levels and elevated nitric oxide [15,30] levels

dative stress conditions) [45] or downregulates (anti-
oxidant conditions) [46] the chymotrypsin-like activity
of the 20S proteasome. Second, the 20S proteasome
appears to be responsible for the degradation of
oxidized proteins [47–51], probably without the need
for a ubiquitylation step [52,53]. Indeed, relatively
mild oxidative stress rapidly (but reversibly) inacti-
vates both the ubiquitin-activating ⁄ conjugating system
and 26S proteasome activity but does not affect 20S
proteasome activity [52,54,55]. Third, it has been
observed that proteasome inhibitors can mimic the
effects of oxidative stressors on mitochondrial mem-
brane potential and increase cell vulnerability to oxi-
dative injury [32]. Moreover, Hsps can confer
resistance to oxidative stress by preserving protea-
some function and attenuating the toxicity of protea-
some inhibition [31]. It is, however, not yet known if
the oxidative stress generated by polyglutamine-con-
taining httEx1 polypeptides is due to alterations in
proteasome activities.
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3077
The analysis presented here was performed in
COS-7 cells because of their very high transfection
efficiency. Using transiently transfected COS cells
expressing mutated Ex1 of htt (httEx1-polyQ), we have
compared the protective activity provided by Hsps and
the antioxidant agent N-acetyl-l-cysteine (NAC). For-
mation of inclusion bodies, mitochondrial membrane
potential ( DYm), ROS, protein oxidation, iron and

ing transfection with httEx1-103Q-EGFP polypeptide
(Fig. 1B,Cc,D). In both cases (72Q and 103Q), a
broad distribution of the size of the inclusions was
noticed. Moreover, the percentage of cells presenting
inclusions as well as the distribution of the size of the
inclusions were dependent on when the analysis was
performed after transfection. Therefore, all the follow-
ing analyses were performed 2 days after transfection.
At that time point, the size of the inclusions formed
by either httEx1-72Q-EGFP or httEx1-103Q-EGFP
polypeptide was heterogeneous but averaged around
10 lm.
We and others have already reported that the
expression of either Hsp70 ⁄ Hdj-1 or Hsp27 induces
protection against httEx1-polyQ-induced cell death
[15,56]. The Hsp70 ⁄ Hdj-1 chaperone machine acts by
decreasing htt aggregation [39,40], whereas Hsp27,
which is less effective than Hsp70 ⁄ Hdj-1 at reducing
aggregation, appears to interfere with cell death
through its antioxidant-related properties [15]. Hsp
overexpression in COS-7 cells was assessed by transient
transfection using vectors encoding either Hsp70, Hdj-
1, Hdj-2 or Hsp27. Hdj-2 is an isoform of Hdj-1 that
has been previously shown not to decrease htt inclu-
sion body formation in COS-7 cells [39]. Immunoblot
analysis of the intracellular level of Hsps revealed an
apparent large increase in the level of Hdj-1 and Hdj-
2, whereas the upregulation of Hsp27 and Hsp70 levels
was more modest (Fig. 1E). The effects mediated by
Hsp overexpression on the formation of inclusion bod-

httEx1-72Q
The expression of httEx1-polyQ is known to alter
mitochondrial activity, leading to mitochondrial
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3078 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
membrane potential (DYm) disruption and ROS
production [15,58]. The phenomenon was measured in
our cell system to compare the protective effects medi-
ated by Hsps and NAC. Analysis of DYm was per-
formed in COS-7 cells transiently transfected as
described above. Forty-eight hours after transfection,
AB
EF
CD
Fig. 1. (A–D) Characterization of httEx1-polyQ-EGFP expression in COS-7 cells. (A) Immunoblot analysis performed 48 h after transfection of
total protein extracts of COS-7 cells transfected with either the pCIneo-EGFP control vector (denoted EGFP) or the same vector bearing
either the httEx1-25Q-EGFP (denoted 25Q-EGFP), httEx1-72Q-EGFP (denoted 72Q-EGFP) or httEx1-103Q-EGFP (denoted 103Q-EGFP) coding
sequence. The immunoblots were probed with anti-EGFP and visualized with ECL as described in Experimental procedures. (B) Confocal
immunofluorescence analysis of transfected cell population. COS-7 cells were transfected with vectors encoding either (a) httEx1-25Q-EGFP,
(b) httEx1-72Q-EGFP) or (c) httEx1-103Q-EGFP. Forty-eight hours after transfection, cells were fixed and analyzed by confocal microscopy as
described in Experimental procedures. Bar, 100 lm. (C) As (B) but enlarged fields are shown. Note the presence of the granules in the cyto-
plasm of the cells. Bar, 20 lm. (D) The percentage of EGFP-containing cells displaying granules is shown. The average percentages,
including standard deviations calculated from three independent experiments, are shown. (E,F) Heat shock proteins (Hsps), but not N-acetyl-
L-cysteine (NAC), decrease the size of httEx1-72Q-EGFP inclusion bodies but are not efficient in decreasing the size of those containing
httEx1-103Q-EGFP. (E) Immunoblot analysis performed 48 h after transfection of total protein extracts of COS-7 cells transfected with
either (a) control vector (pCIneo) or (b) vectors bearing the Hsp70 ⁄ Hdj-1, Hdj-2 or Hsp27 coding sequence. The immunoblots were probed
with the corresponding antibodies. Control of gel loading was performed with anti-actin. Immunoblots were visualized by ECL as des-
cribed in Experimental procedures. (F) Confocal immunofluorescence analysis of COS-7 cells transfected with vectors encoding either
httEx1-72Q-EGFP (72Q-EGFP) or httEx1-103Q-EGFP (103Q-EGFP) together with pCIneo vector or vectors bearing the Hsp70 ⁄ Hdj-1
(+ Hsp40 + Hsp70) or Hsp27 (+ Hsp27) coding sequence. NAC (2 m

10
3
10
4
25Q
+NAC
72Q
72Q+NAC
103Q
103Q+NAC
04080
Counts
120 160 200
10
0
10
1
10
2
FL2-H
10
3
10
4
10
0
10
1
10
2

XRos and cytometry was performed as described in (A). The percentage of
the cell population with an FL2-H fluorescence greater than 2 · 10
1
was recorded during the FACS analysis. The percentage of decrease of
DYm was calculated as the ratio of the percentage of cells with FL2-H fluorescence greater than 2 · 10
1
in the samples to that observed in
control cells (transfected with pCIneo-EGFP). A representative experiment is presented. The data from three independent experiments were
used to perform statistical analysis (see Experimental procedures). (C) Electron microscopy analysis of mitochondrial morphology of COS-7
cells transfected with either pCIneo-EGFP vector (pCIneo), httEx1-72Q-EGFP vector (72Q) or httEx1-103Q-EGFP vector (103Q). Transfections
were performed with a combination of those encoding Hsp70 ⁄ Hdj-1 (Hsp70 + Hdj-1) and Hsp27 (Hsp27). Cells transfected with httEx1-
103Q-EGFP vector were also exposed to 2 m
M NAC before being analyzed (as described in the previous figures). Bar, 1 lm.
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3080 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
whereas it was slightly decreased by the transfection of
httEx1-72Q-EGFP vector. This suggests an alteration
of DYm as a consequence of the expression of httEx1-
72Q-EGFP polypeptide. A similar effect was noticed
when the experiment was carried out with httEx1-
polyQ-HA vectors [15] encoding httEx1 polypeptides
with 23 or 74 glutamine repeats not fused to EGFP,
but to a hemagglutinin (HA) tag (data not shown).
This control experiment suggests that no significant
green fluorescent spillover or cytotoxicity was induced
by EGFP expression. The mitochondrial depolariza-
tion mediated by httEx1-103Q-EGFP expression was
more drastic and was roughly similar to that induced
by a 15 min incubation of control cells with a 10 lm
solution of the mitochondrial depolarizer p-trifluoro-

the presence of NAC. In this respect, Hsps were active
only in the case of cells transfected with httEx1-72Q-
EGFP vector. In cells expressing httEx1-103Q-EGFP,
the presence of Hsps did not restore normal morphol-
ogy of the mitochondria (Fig. 2C).
This suggests that ROS are probably responsible for
the DYm disruption and damage to mitochondrial
morphology in httEx1-72Q-EGFP-expressing or
httEx1-103Q-EGFP-expressing cells. In contrast,
httEx1-25Q-EGFP expression did not alter DYm
(Fig. 2A) or the morphology of mitochondria (not
shown).
Comparative analysis of the protective effect of
NAC and Hsps against ROS, protein oxidation,
iron and nitric oxide level upregulation caused
by expanded httEx1 expression
DYm disruption usually causes an intracellular burst
of ROS [58] that induce oxidative damage, such as that
observed in cells expressing expanded httEx1 [15].
Recently, we showed, using different cell lines, inclu-
ding COS-7 cells, incubated with the fluorescent probe
DCFH-DA, that peroxide production was induced by
the expression of httEx1-polyQ-HA polypeptides [15].
An increase in the number of CAG repeats from 23 to
74 correlated with an increase in the oxidation process.
Here, we have performed similar experiments using the
httEx1-polyQ-EGFP vectors described above that con-
tain a broader range of polyQ repeats: 25, 72 and 103.
As seen in Fig. 3A, 48 h after transfection, the fluores-
cence of DCFH-DA (see Experimental procedures)

sing httEx1-103Q-EGFP. A similar observation was
made in the case of cells expressing httEx1-72Q-EGFP
(not shown). Hence, upregulation of ROS levels
appears to be an intrinsic property of living COS-7
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3081
cells expressing httEx1-72Q or httEx1-103Q polypep-
tides. Similar to the use of DCFH-DA, upregulated
fluorescence was detected using dihydroethidine (HE),
a probe that is preferentially oxidized to ethidium bro-
mide by superoxide anions O
2
•–
(data not shown) and
has a different fluorescent emission wavelength from
EGFP (EGFP, 510 lm; HE, 590 lm). Hence, despite
the fact that EGFP and DCFH-DA have quite sim-
ilar emission wavelengths, it is possible to detect an
NAC-sensitive increase in fluorescence that reflects
accumulation of intracellular ROS levels in COS-7
cells transiently transfected with httEx1-polyQ-EGFP
vectors.
To analyze the effects on ROS mediated by Hsp over-
expression, COS-7 cells were transiently transfected
with vectors encoding httEx1-72Q-EGFP and either
Hsp70 ⁄ Hdj-1 or Hsp27. In control cells transfected with
the EGFP vector, the overexpression of Hsp70 ⁄ Hdj-1
decreased ROS levels slightly but not significantly
(Fig. 3B). In contrast, overexpression of Hsp27 was
more efficient and induced a significant decrease

ROS burst generated by httEx1-72Q-EGFP expression,
and when all three Hsps were overexpressed, a more
intense decrease (P<0.001) in ROS levels was
observed, an effect that was reversed if Hsp27(C137A)
mutant was overexpressed instead of wild-type Hsp27.
Fig. 3. httEx1-polyQ expression enhances reactive oxygen species
(ROS) levels. (A) Analysis of ROS induced by httEx1-polyQ expres-
sion. COS-7 cells were transfected with either control pCIneo-EGFP
vector (EGFP vector), or vectors encoding httEx1-25Q-EGFP (25Q
EGFP), httEx1-72Q EGFP (72Q EGFP) or httEx1-103Q EGFP (103Q
EGFP). Forty-eight hours after transfection, cells were washed with
NaCl ⁄ P
i
, and incubated with 2¢,7¢-dichlorofluorescein (DCFH-DA),
and fluorescence was monitored by FACS cytometry as described
in Experimental procedures. The fluorescence index was
determined as the ratio of the fluorescence of cells expressing
httEx1-polyQ polypeptides to that of control pCIneo-EGFP vector-
transfected cells. A representative experiment is presented. As
control, 36 h after transfection, cells transfected with httEx1-103Q-
EGFP were exposed or not exposed to 20 l
M z-VAD-fmk before
being analyzed. Treatment with 2 m
M N-acetyl-L-cysteine (NAC)
was as previously described. (B) ROS induced by httEx1-72Q-EGFP
expression in COS-7 cells expressing different sets of heat shock
proteins (Hsps). COS-7 cells were transfected with either control
pCIneoEGFP vector (EGFP vector) or the vector encoding httEx1-
72Q EGFP (72Q EGFP). In addition, cotransfections were per-
formed using vectors encoding Hsp70 ⁄ Hdj-1, Hsp70 ⁄ Hdj-2, Hsp27

oxidative modifications mainly induced by OH
•
is the
formation of carbonyl residues on amino acid side
chains of proteins [43,64]. In order to explore the abil-
ity of httEx1-polyQ to oxidize cellular proteins, we
performed immunoblot detection of protein carbonyl
residues in 2,4-dinitrophenylhydrazine (2,4-DNPH)-
treated extracts of COS-7 cells expressing the different
httEx1-polyQ-EGFP polypeptides (see Experimental
procedures) (Fig. 4A). Quantitative analysis of the
oxyblots (in the 10–40 kDa molecular mass range) is
presented in Fig. 4B,C. As seen in Fig. 4A, httEx1-
polyQ-EGFP expression increased the detection of
protein carbonyl residues in cellular polypeptides in a
polyQ expansion size-dependent manner. No specific
oxidized protein bands corresponding to the gel migra-
tion of httEx1-polyQ-EGFP polypeptides were detec-
ted, suggesting that httEx1-polyQ-EGFP expression
mainly enhances the oxidation of cellular proteins
(particularly in the 10–40 kDa molecular mass range)
that already display a basal level of oxidation in con-
trol cells. Expression of Hsp70 ⁄ Hdj-1 or Hsp27 did
not significantly change the pattern and level of oxid-
ized proteins in control cells, whereas the overexpres-
sion of these chaperones correlated with a decreased
level of oxidized proteins in response to httEx1-72Q-
EGFP expression (Fig. 4A). Analysis of httEx1-103Q-
EGFP-expressing cells revealed that in this case
Hsp70 ⁄ Hdj-1 or Hsp27 were not efficient in counter-

httEx1-103Q-EGFP polypeptide is a consequence
rather than a cause of the deleterious effect generated
by the oxidative stress.
Another important parameter of oxidative stress is
nitric oxide (NO
•
). Indeed, elevated levels of NO
•
have
been observed in HD [65] and transgenic HD mice
(R6 ⁄ 2 and R6 ⁄ 1 model) that may contribute to patho-
genesis and precede neuronal cell death [66,67]. The
pathology of NO
•
results from its reaction with O
2
•–
to form peroxynitrite (ONOO
•–
), which can diffuse for
several micrometers before decomposing to form the
powerful and cytotoxic oxidants OH
•
and nitrogen
dioxide [68]. These observations prompted us to ana-
lyze NO
•
levels in COS-7 cells expressing httEx1-
polyQ-EGFP polypeptides and to test whether Hsps or
NAC could modulate NO

increase in NO
•
level generated by httEx1-72Q-EGFP
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3083
expression. Concerning the elevation of NO
•
induced
by httEx1-103Q-EGFP polypeptide, Hsp70 ⁄ Hdj-1
overexpression had no significant effects, whereas
Hsp27 reduced the increase in NO
•
level by more than
50% (P<0.001). It is of interest that NAC com-
pletely abolished the increase in NO
•
level generated
by httEx1-103Q-EGFP expression.
Analysis of httEx1-polyQ expression with regard
to the three major proteolytic activities of 20S
proteasome, a phenomenon partially restored
by Hsp expression but not by NAC
Proteasome inhibition is known to induce intracellular
protein aggregation and increased carbonyl formation
EGFP
vector
A
BC
Control
Hsp70/Hdj-1

M N-acetyl-L-cysteine (NAC)
before being analyzed. The arrow indicates the position of the more intensively oxidized polypeptide in the assay. The bracket underlines the
domain (molecular mass range 10–40 kDa) of the oxyblots that contains the greatest changes in protein oxidation. (B) Quantitative analysis
of the oxyblots presented in (A) (see Experimental procedures). The domains of the blots indicated by a bracket (see Fig. 4A) were scanned
and the signals quantified (see Experimental procedures). This approach was used to avoid the major oxidized protein (about 45 kDa), which
shows a rather unaltered signal throughout the experiment. The level of protein oxidation (arbitrary units) is presented. (C) Protein oxidation
index. The values in (B) were divided by the value determined for the control cells transfected with the EGFP vector. The results from a rep-
resentative experiment are shown.
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3084 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
in proteins [31,69]. Hence, we first investigated the pos-
sibility that proteasome inhibition could be responsible
for the oxidative stress mediated by the expression of
httEx1-polyQ-EGFP polypeptides. Control pCIneo-
EGFP-transfected COS-7 cells were exposed for 1 h to
10 lm of the proteasome inhibitor lactacystin. In these
cells, 80% inhibition of proteasome activities correla-
ted with a 50% increase in ROS levels and with a 1.7-
fold increase in the level of oxidized proteins (ranging
between 10 and 40 kDa, as defined above in Fig. 4)
(not shown). The oxidative stress induced by protea-
some inhibition therefore seems to be less intense than
that induced by the expression of httEx1-72Q-EGFP
and httEx1-103Q-EGFP polypeptides (see above;
Figs 3 and 4).
We next analyzed the effects mediated by the expres-
sion of the different httEx1-polyQ-EGFP polypeptides
and Hsps as well as those induced by NAC treatment
on the three major proteolytic activities of the 20S pro-
teasome. Indeed, Hsps (particularly, Hdj-1 ⁄ Hsp40) can

lowing centrifugation, pelleted cells were used to determine the
Fe(II) level as described in Experimental procedures. All samples
contained similar amounts of protein. Absorbance of the ferrozine–
Fe(II) complex (AU, arbitrary units) was read at 562 nm. The results
for cells treated, as described in the previous figures, with 2 m
M
N-acetyl-L-cysteine (NAC) is presented. The data from three inde-
pendent experiments were used to performed statistical analysis
(see Experimental procedures). *P < 0.05.
Fig. 6. Nitric oxide level determination. COS-7 cells were transfected
with either control pCIneo-EGFP vector (EGFP vector), or the same
vector encoding httEx1-25Q-EGFP (25Q EGFP), httEx1-72Q-EGFP
(72Q EGFP) or httEx1-103Q-EGFP (103Q EGFP). Cotransfections
were performed using the vectors encoding Hsp70 ⁄ Hdj-1 or Hsp27.
Forty-eight hours after transfection, cells were processed for nitric
oxide level determination as described in Experimental procedures.
Cells transfected with httEx1-103Q-EGFP were also exposed to
2m
M N-acetyl-L-cysteine (NAC) before being analyzed (as described
in the previous figures). The data from three independent experi-
ments were used to performed statistical analysis (see Experimental
procedures). The asterisks denote statistical significance when
compared with respective controls: *P < 0.05; **P < 0.001.
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3085
httEx1-polyQ expression (Fig. 7B). A similar observa-
tion was made in cells treated with NAC. In contrast,
the caspase-like activity was strongly altered by httEx1
polyQ-EGFP expression, and the level of this inhibi-
tion increased with polyQ expansion size, reaching

that the expression of either Hsp70 + Hdj-1 or Hsp27
reduced inclusion size in COS-7 cells expressing httEx1
with 72Q repeats. However, Hsps were ineffective
towards the inclusions formed by httEx1-103Q. No
significant modulation of inclusion size by NAC was
detected. These findings confirm the results of several
other studies performed in cellular models of polyQ
disease that showed no effect of antioxidant com-
pounds on inclusion formation [15]. An almost com-
plete reversal of httEx1-72Q-EGFP-mediated DYm
disruption and mitochondrial morphology alteration
was observed in cells that overexpress Hsp70 ⁄ Hdj-1 or
Hsp27. In contrast, the overexpression of these Hsps
did not attenuate the drastic mitochondrial defects
generated by httEx1-103Q-EGFP. Hence, despite the
fact that unaggregated mutant htt may already be
toxic [34,70], httEx1-polyQ toxicity towards mitochon-
dria increased in a CAG repeat expansion-dependent
manner to reach a point where it could not be restored
60
A
B
C
*
*
*
*
*
*
*

103Q EGFP
Hsp70/Hdj-1
Hsp27
NAC
+++ +
-+-
60
40
20
0
+++ +
-+-
+++

+

M N-acetyl-L-cysteine (NAC) before being analyzed
(as described in the previous figures). Treatment with lactacystin
abolished suc-LLVY-MCA fluorescence by more than 80% (not
shown). The data from three independent experiments were used
to perform statistical analysis (see Experimental procedures). The
asterisks denote statistical significance when compared with
respective controls: *P < 0.05; **P < 0.001.
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3086 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
by Hsp overexpression. The protective effect provided
by Hsp expression was then compared to that media-
ted by NAC. It is of interest that even in cells expres-
sing httEx1-103Q-EGFP, the presence of NAC
counteracted DYm disruption. Hence, mitochondrial
disfunction in httEx1-polyQ-expressing cells appears to
be ROS dependent. Taking into account that NAC did
not reduce the size of inclusion bodies, the difference
in protection mediated by Hsps and NAC suggests a
correlation between the ROS-dependent DYm disrup-
tion and the presence of inclusion bodies.
We have observed that polyQ-mediated DYm
disruption correlated, in a CAG repeat expansion-
dependent manner, with an increased level of intracel-
lular ROS and with increased formation of carbonyl
residues in proteins. The enhanced oxidation of pro-
teins observed in httEx1-polyQ-expressing cells was
not found to correlate with an increased intracellular
level of iron. Interestingly, a recent proteomic analysis
detecting protein carbonyl residues confirmed that,
in vivo, some proteins are indeed oxidized in the HD

•
that is pro-
duced in cells expressing mutated htt is toxic and pre-
cedes neuronal death [65–67]. We show here for the
first time in a cell culture model that an elevated level
of NO
•
is generated in COS-7 cells in response to
httEx1-polyQ-EGFP expression, indicating that this
effect is not restricted to neuronal cells. In httEx1-
72Q-EGFP-expressing cells, we have observed that
Hsp70 ⁄ Hsp40 or Hsp27 overexpression abolished the
increase in NO
•
levels. In contrast, Hsps were not effi-
cient in httEx1-103Q-EGFP-expressing cells, but NAC
did abolish the NO
•
increase induced by both httEx1-
72Q-EGFP and httEx1-103Q-EGFP polypeptides.
Hence, correlations again exist between NO
•
and ROS
upregulation and the presence of aggregated httEx1-
polyQ.
It has been reported that proteasome inhibitors
induce intracellular protein aggregation and stimulate
oxidative protein modifications such as increased car-
bonyl formation, similar to those seen with hydrogen
peroxide treatment [69]. Proteasome inhibition has

stress induced by this mutant protein. The effects
induced by Hsps in httEx1-72Q-EGFP-expressing cells
suggest that the interferences in proteasome activities
do not depend on inclusion body formation. More-
over, the comparison with the oxidative stress induced
by lactacystin suggests that the rather weak alterations
in proteasome activities induced by httEx1-polyQ
expression are not responsible for the oxidative stress
generated by these polypeptides.
Hence, the results presented in this study suggest the
following question: could httEx1-polyQ inclusion bod-
ies generate ROS by themselves? Intriguing observa-
tions may favor this hypothesis. For example, in vitro,
aggregated a-synuclein (Parkinson) and b-amyloid
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3087
(Alzheimer) polypeptides generate spin-trap detectable
ROS when incubated with redox-active transition
metals, such as iron or copper [75]. Since htt is an
iron-regulated protein, it will be of interest to test whe-
ther it shares the ability to produce ROS in the pres-
ence of metals. If the hypothesis is true, it will mean
that oxidative events are generated by httEx1-polyQ
inclusion bodies that alter the function of mitochon-
dria concentrated in the area surrounding the inclu-
sions [76], leading to a subsequent burst of ROS of
mitochondrial origin.
Several reports have proposed the provocative idea
that inclusion bodies are in fact beneficial to the cell
[17,19] by, for example, promoting the clearance of

mine repeats (called httEx1-72Q-HA) was also used [15].
The control vector (pCIneo) and Hsp27 wild-type-bearing
vector (pCIneohsp27) have already been characterized [77].
Vectors encoding human Hdj-1(Hsp40), Hdj-2 and Hsp70
have already been described [15], as well as vectors enco-
ding wild-type Hsp27 and mutant Hsp27(C137A) [44,78].
For transfection experiments, exponentially growing COS-7
cells were plated in 60 mm plates (5 · 10
5
cells per plate)
(TPP, Zu
¨
rich, Switzerland) 1 day before transfection. Each
transfection experiment was performed with 3 lgof
DNA encoding httEx1-polyQ or the various chaperones
(Hsp70 ⁄ Hdj-1 and Hsp27) using lipofectamine (Gibco, Invi-
trogen) according to the manufacturer’s instructions. In the
case of transfection with multiple vectors, we used identical
amounts of each vector (1–2 lg of DNA to a total of
4 lg). Forty-eight hours after transfection, the cell medium
was removed and replaced with DMEM supplemented with
10% FBS. z-VAD-fmk and NAC were from Sigma–Aldrich
(St-Quentin-Fallavier, France).
Immunoblot analysis
COS-7 cells were harvested and the cell pellet was solubi-
lized in boiling 2 · SDS sample buffer. Protein concentra-
tion was determined in aliquots using the Bradford protein
assay. Total protein samples were separated by 12%
SDS ⁄ PAGE before being analyzed on immunoblots probed
with either anti-EGFP (1 : 1000) (Molecular Probes ⁄ Inter-

i
before being analyzed in a
FACS calibur Cytometer (Becton Dickinson, Mountain
View, CA) equipped with an argon ion laser emitting at
488 nm. CM-H
2
XRos fluorescence was detected in the
FL2-H channel. The percentage of cells that displayed a
FL2-H fluorescence value higher than 2 · 10
1
was auto-
matically determined and used to calculate the percentage
of mitochondrial membrane potential disruption induced
by httEx1-polyQ or FCCP. For electron microscope ana-
lysis, transfected COS-7 cells were grown for 48 h in
60 mm diameter plates before being fixed for 30 min at
4 °C in a buffer composed of 2% glutaraldehyde in 0.1 m
Na-cacodylate ⁄ HCl buffer, pH 7.4. Cells were then rinsed
three times (overnight) at 4 °C in 0.1 m Na-cacody-
late ⁄ HCl buffer, pH 7.4, containing 0.2 m sucrose, before
being postfixed for 30 min at 4 °C in a buffer composed
of 1% osmium tetroxide and 0.15 m Na-cacodylate ⁄ HCl,
adjusted to pH 7.4. Cells were then dehydrated with gra-
ded ethanol, scraped and pelleted in 70% ethanol and
embedded in Epon as a cell pellet. After polymerization
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3088 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
at 60 °C for 3 days, ultrathin sections (60–80 nm) were
cut using an RMC MTX ultramicrotome (Ventana Medi-
cal Systems, Tucson, AZ, USA), collected on 200 mesh

HE and analyzed
by flow cytometry with an excitation wavelength of
488 nm. The emission filter was 610 nm bandpassed for
ethidium bromide fluorescence (FL2-H).
Determination of intracellular iron levels
The method used to estimate intracellular levels of total
Fe(II) was based on the Fish colorimetric assay [1]. In brief,
cells were washed and scraped off the culture dish in
NaCl ⁄ P
i
. Following centrifugation, pelleted cells were resus-
pended in 1 mL of water. An equivalent of 2 mg of total
cell protein was used per assay. After addition of 500 lL
of solution A (0.6 m HCl, 0.142 m KMnO
4
), the samples
were incubated for 2 h at 60 °C. During this incubation,
iron was released in soluble form. Subsequent to addition
of 100 lL of buffer B [6.5 mm ferrozine (disodium 3-(2-
pyridyl)-5,6-bis(4-phenylsulfonate)-1,2,4-triazine (Sigma
P5338), 2 m ascorbic acid, 5 m ammonium acetate] and a
further 60 min incubation, absorbance of the ferrozine–
Fe(II) complex was read at 562 nm.
Immunofluorescence of EGFP inclusions
Transfected COS-7 cells were grown on coverslips in
60 mm plates. Forty-eight hours after transfection, cells
were rinsed once in DMEM supplemented with 10% FBS
before being washed once with NaCl ⁄ P
i
devoid of calcium

NIH image 1.62 software (NIH, Bethesda, MD, USA).
Determination of NO
•
levels
The Nitric Oxide Quantification Kit from Active Motif
Europe (Rixensart, Belgium) was used. In brief, 48 h after
transfection, cells were trypsinized and lysed in the buffer
provided with the kit. Since large proteins can interfere
with the Griess reaction, cell lysates were filtered through
micropore filters (10 kDa cut) using centrifugal filter tubes
(Amicon-Milipore, St-Quentin en Yvelines, France). One
hundred microliters of the filtered lysates was then analyzed
according to the manufacturer’s instructions. Absorbance
was read on a Dynatech MR5000 microplate reader (Dyna-
tech, Chantilly, VA, USA) at 540 nm with a reference
wavelength of 620 nm.
Proteasome activities
20S proteasome activity was measured in cell extracts 48 h
after transfection. COS-7 cells were washed twice with cold
NaCl ⁄ P
i
and lyzed by a 30 min-incubation in 0.5 mm dithio-
threitol as already described [79]. Unlysed cells, membranes
and nuclei were eliminated by centrifugation at 14 000 g
(Biofuge Fresno centrifuge, rotor #3325, Heraeus SAS,
Courtabeouf, France). A typical proteasome assay was car-
ried out with 100 lg of protein cell extracts in a total vol-
ume of 200 lL of proteasome buffer (50 mm Tris ⁄ HCl,
pH 7.8, 20 mm KCl, 5 mm MgOAc, 0.5 mm dithiothreitol).
The fluorogenic substrates suc-LLVY-MCA, N-boc-LSTR-

gion Rhoˆ ne-Alpes (The
´
matique Cancer) (to
APA). Wance Firdaus was a French postgraduate
scholarship holder from CNOUS (Centre National des
Oeuvres Universitaires et Scolaires), Paris. Andreas
Wyttenbach thanks the HighQ Foundation and the
Medical Research Council (MRC) for financial sup-
port. William Currie was a Visiting Professor, from
Dalhousie University, Halifax, Canada and held a
CIHR ⁄ CNRS International Scientific Exchange Schol-
arship from the Canadian Institutes of Health
Research and the Centre National de la Recherche Sci-
entifique, France.
References
1 Ross CA (1997) Intranuclear neuronal inclusions: a
common pathogenic mechanism for glutamine-repeat
neurodegenerative diseases? Neuron 19, 1147–1150.
2 Hilditch-Maguire P, Trettel F, Passani LA, Auerbach
A, Persichetti F & MacDonald ME (2000) Huntingtin:
an iron-regulated protein essential for normal nuclear
and perinuclear organelles. Hum Mol Genet 9, 2789–
2797.
3 Kegel KB, Kim M, Sapp E, McIntyre C, Castano
JG, Aronin N & DiFiglia M (2000) Huntingtin
expression stimulates endosomal-lysosomal activity,
endosome tubulation, and autophagy. J Neurosci 20,
7268–7278.
4 Sherman MY & Goldberg AL (2001) Cellular defenses
against unfolded proteins: a cell biologist thinks about

intranuclear inclusions. Cell 95, 55–66.
12 Carmichael J, Chatellier J, Woolfson A, Milstein C,
Fersht AR & Rubinsztein DC (2000) Bacterial and yeast
chaperones reduce both aggregate formation and cell
death in mammalian cell models of Huntington’s dis-
ease. Proc Natl Acad Sci USA 97, 9701–9705.
13 Klement IA, Skinner PJ, Kaytor MD, Yi H, Hersch
SM, Clark HB, Zoghbi HY & Orr HT (1998) Ataxin-1
nuclear localization and aggregation: role in polygluta-
mine-induced disease in SCA1 transgenic mice. Cell 95,
41–53.
14 Marsh JL, Walker H, Theisen H, Zhu YZ, Fielder T,
Purcell J & Thompson LM (2000) Expanded polygluta-
mine peptides alone are intrinsically cytotoxic and cause
neurodegeneration in Drosophila. Hum Mol Genet 9,
13–25.
15 Wyttenbach A, Sauvageot O, Carmichael J, Diaz-Latoud
C, Arrigo A-P & Rubinsztein DC (2002) Heat shock
protein 27 prevents cellular polyglutamine toxicity and
suppresses the increase of reactive oxygen species caused
by huntingtin. Hum Mol Genet 11, 1137–1151.
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3090 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
16 Wyttenbach A (2004) Role of heat shock proteins dur-
ing polyglutamine neurodegeneration: mechanisms and
hypothesis. J Mol Neurosci 23 , 69–96.
17 Arrasate M, Mitra S, Schweitzer ES, Segal MR & Fink-
beiner S (2004) Inclusion body formation reduces levels
of mutant huntingtin and the risk of neuronal death.
Nature 431, 805–810.

oxidised proteins in neurodegenerative diseases. Lancet
351, 1510–1520.
28 Bogdanov MB, Andreassen OA, Dedeoglu A, Ferrante
RJ & Beal MF (2001) Increased oxidative damage to
DNA in a transgenic mouse model of Huntington’s dis-
ease. J Neurochem 79, 1246–1249.
29 Halliwell B (2001) Role of free radicals in the neurode-
generative diseases: therapeutic implications for anti-
oxidant treatment. Drugs Aging 18 , 685–716.
30 Giuliano P, De Cristofaro T, Affaitati A, Pizzulo GM,
Feliciello A, Criscuolo C, De Michele G, Filla A, Avv-
edimento EV & Varrone S (2003) DNA damage induced
by polyglutamine-expanded proteins. Hum Mol Genet
12, 2301–2309.
31 Ding Q & Keller JN (2001) Proteasome inhibition in
oxidative stress neurotoxicity: implications for heat
shock proteins. J Neurochem 77, 1010–1017.
32 Ding Q & Keller JN (2001) Proteasomes and protea-
some inhibition in the central nervous system. Free
Radic Biol Med 31, 574–584.
33 Sawa A, Wiegand GW, Cooper J, Margolis RL, Sharp
AH, Lawler JF Jr, Greenamyre JT, Snyder SH & Ross
CA (1999) Increased apoptosis of Huntington disease
lymphoblasts associated with repeat length-dependent
mitochondrial depolarization. Nat Med 5, 1194–1198.
34 Panov AV, Gutekunst CA, Leavitt BR, Hayden MR,
Burke JR, Strittmatter WJ & Greenamyre JT (2002)
Early mitochondrial calcium defects in Huntington’s
disease are a direct effect of polyglutamines. Nat Neuro-
sci 5, 731–736.

differentiation of rat olfactory neurons counteracts
apoptosis. Cell Death Differ 6, 227–233.
42 Wagstaff MJ, Collaco-Moraes Y, Smith J, de Belleroche
JS, Coffin RS & Latchman DS (1999) Protection of
neuronal cells from apoptosis by Hsp27 delivered with a
herpes simplex virus-based vector. J Biol Chem 274,
5061–5069.
43 Preville X, Salvemini F, Giraud S, Chaufour S, Paul C,
Stepien G, Ursini MV & Arrigo A-P (1999) Mammalian
small stress proteins protect against oxidative stress
through their ability to increase glucose-6-phosphate
W. J. J. Firdaus et al. Huntingtin inclusions and oxidation
FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS 3091
dehydrogenase activity and by maintaining optimal cel-
lular detoxifying machinery. Exp Cell Res 247, 61–78.
44 Arrigo A-P, Virot S, Chaufour S, Firdaus W, Kretz-
Remy C & Diaz-Latoud C (2005) Hsp27 consolidates
intracellular redox homeostasis by upholding glu-
tathione in its reduced form and by decreasing iron
intracellular levels. Antioxid Redox Signal 7, 414–422.
45 Ullrich O, Reinheckel T, Sitte N, Hass R, Grune T &
Davies KJ (1999) Poly-ADP ribose polymerase activates
nuclear proteasome to degrade oxidatively damaged his-
tones. Proc Natl Acad Sci USA 96, 6223–6228.
46 Kretz-Remy C & Arrigo A-P (2003) Modulation of the
chymotrypsin-like activity of the 20S proteasome by
intracellular redox status: effects of glutathione peroxi-
dase-1 overexpression and antioxidant drugs. Biol Chem
384, 589–595.
47 Grune T & Davies KJ (1997) Breakdown of oxidized

decreased CRE-mediated transcription and early gene
expression changes prior to cell death in an inducible
cell model of Huntington’s disease. Hum Mol Genet 10,
1829–1845.
57 Wacker JL, Zareie MH, Fong H, Sarikaya M &
Muchowski PJ (2004) Hsp70 and Hsp40 attenuate for-
mation of spherical and annular polyglutamine oligo-
mers by partitioning monomer. Nat Struct Mol Biol 11,
1215–1222.
58 Chen Q, Chai YC, Mazumder S, Jiang C, Macklis RM,
Chisolm GM & Almasan A (2003) The late increase in
intracellular free radical oxygen species during apoptosis
is associated with cytochrome c release, caspase activa-
tion, and mitochondrial dysfunction. Cell Death Differ
10, 323–334.
59 Mehlen P, Pre
´
ville X, Kretz-Remy C & Arrigo A-P
(1996) Human hsp27, Drosophila hsp27 and human
aB-crystallin expression-mediated increase in glutathione
is essential for the protective activity of these proteins
against TNFa-induced cell death. EMBO J 15, 2695–
2706.
60 Diaz-Latoud C, Buache E, Javouhey E & Arrigo A-P
(2005) Substitution of the unique cysteine residue of
murine hsp25 interferes with the protective activity of
this stress protein through inhibition of dimer forma-
tion. Antioxid Redox Signal 7, 436–445.
61 Graf E, Mahoney JR, Bryant RG & Eaton JW (1984)
Iron-catalyzed hydroxyl radical formation. Stringent

70 Panov AV, Burke JR, Strittmatter WJ & Greenamyre
JT (2003) In vitro effects of polyglutamine tracts on
Ca2+-dependent depolarization of rat and human
Huntingtin inclusions and oxidation W. J. J. Firdaus et al.
3092 FEBS Journal 273 (2006) 3076–3093 ª 2006 The Authors Journal compilation ª 2006 FEBS
mitochondria: relevance to Huntington’s disease. Arch
Biochem Biophys 410, 1–6.
71 Perluigi M, Poon HF, Maragos W, Pierce WM, Klein
JB, Calabrese V, Cini C, De Marco C & Butterfield DA
(2005) Proteomic analysis of protein expression and oxi-
dative modification in R6 ⁄ 2 transgenic mice ) a model
of Huntington’s disease. Mol Cell Proteomics 20, 20–30.
72 Solenski NJ, Kostecki VK, Dovey S & Periasamy A
(2003) Nitric-oxide-induced depolarization of neuronal
mitochondria: implications for neuronal cell death. Mol
Cell Neurosci 24, 1151–1169.
73 Kotamraju S, Tampo Y, Keszler A, Chitambar CR,
Joseph J, Haas AL & Kalyanaraman B (2003) Nitric
oxide inhibits H2O2-induced transferrin receptor-depen-
dent apoptosis in endothelial cells: role of ubiquitin-
proteasome pathway. Proc Natl Acad Sci USA 100,
10653–10658.
74 Holmberg CI, Staniszewski KE, Mensah KN, Matou-
schek A & Morimoto RI (2004) Inefficient degradation
of truncated polyglutamine proteins by the proteasome.
EMBO J 23, 4307–4318.
75 Tabner BJ, Turnbull S, El-Agnaf OM & Allsop D
(2002) Formation of hydrogen peroxide and hydroxyl
radicals from A(beta) and alpha-synuclein as a possible
mechanism of cell death in Alzheimer’s disease and Par-