Progressive aggregation despite chaperone
associations of a mutant SOD1-YFP in
transgenic mice that develop ALS
Jiou Wanga,b,1, George W. Farra,b,1, Caroline J. Zeissc, Diego J. Rodriguez-Gild, Jean H. Wilsonc, Krystyna Furtaka,b,
D. Thomas Rutkowskie,2, Randal J. Kaufmane, Cristian I. Rusef, John R. Yates IIIf, Steve Perring, Mel B. Feanyh,
and Arthur L. Horwicha,b,3
aHoward Hughes Medical Institute,bDepartment of Genetics,cSection of Comparative Medicine, anddDepartment of Neurosurgery, Yale University School
of Medicine, New Haven, CT 06510;eHoward Hughes Medical Institute and Department of Biological Chemistry, University of Michigan Medical Center,
Ann Arbor, MI 48109;fDepartment of Chemical Physiology, The Scripps Research Institute, La Jolla, CA 92037;gALS Therapeutic Development Institute,
Cambridge, MA 02142; andhDepartment of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115
Contributed by Arthur L. Horwich, December 20, 2008 (sent for review November 3, 2008)
Recent studies suggest that superoxide dismutase 1 (SOD1)-linked
amyotrophic lateral sclerosis results from destabilization and mis-
folding of mutant forms of this abundant cytosolic enzyme. Here,
we have tracked the expression and fate of a misfolding-prone
human SOD1, G85R, fused to YFP, in a line of transgenic G85R
SOD1-YFP mice. These mice, but not wild-type human SOD1-YFP
transgenics, developed lethal paralyzing motor symptoms at 9
months. In situ RNA hybridization of spinal cords revealed pre-
dominant expression in motor neurons in spinal cord gray matter
in all transgenic animals. Concordantly, G85R SOD-YFP was dif-
fusely fluorescent in motor neurons of animals at 1 and 6 months
of age, but at the time of symptoms, punctate aggregates were
observed in cell bodies and processes. Biochemical analyses of
spinal cord soluble extracts indicated that G85R SOD-YFP behaved
as a misfolded monomer at all ages. It became progressively
insoluble at 6 and 9 months of age, associated with presence of
soluble oligomers observable by gel filtration. Immunoaffinity
capture and mass spectrometry revealed association of G85R SOD-
YFP, but not WT SOD-YFP, with the cytosolic chaperone Hsc70 at all
ages. In addition, 3 Hsp110’s, nucleotide exchange factors for
Hsp70s, were captured at 6 and 9 months. Despite such chaperone
interactions, G85R SOD-YFP formed insoluble inclusions at late
times, containing predominantly intermediate filament proteins.
We conclude that motor neurons, initially ‘‘compensated’’ to main-
tain the misfolded protein in a soluble state, become progressively
unable to do so.
Hsc70 ? Hsp110 ? motor neuron ? neurodegeneration ? proteomic
uitin-positive aggregates in motor neurons in both sporadic and
inherited cases (1, 2). The nature of pathogenesis remains poorly
understood, but in at least one inherited form of the condition,
involving mutations in the abundantly expressed cytosolic ho-
modimeric enzyme superoxide dismutase (SOD1), there is in-
creasing evidence that protein misfolding is involved. In partic-
ular, disease-associated mutations affecting a variety of residues
in different regions of the SOD1 subunit all lead to destabili-
zation and misfolding (3–6), and in both humans and transgenic
mice carrying such alleles, the enzyme subunit itself is found
lodged in spinal cord aggregates (1, 2). This gain of function
behavior is consistent with the generally dominant inheritance
pattern of SOD1-linked ALS. By contrast, a loss of function
behavior that might be attributed to deficiency of free radical
scavenging function of the enzyme has not been supported: For
example, many ALS-linked SOD1 alleles retain enzymatic ac-
tivity (1, 2) and SOD1 knockout mice fail to develop motor
myotrophic lateral sclerosis (ALS) is a paralyzing neuro-
degenerative condition associated with production of ubiq-
The fate of mutant SOD1 protein is thus of major interest with
respect to disease pathogenesis, and tracking the protein in vivo
would seem desirable. Such an approach has recently been taken
in cultured cells, where ALS-associated mutant versions of
human SOD1 were fused with yellow fluorescent protein (YFP),
enabling optical tracking of SOD1 aggregation (8). Aggregation
of these fusion proteins was evidently a consequence of misfold-
ing of the mutant SOD1 moiety, because a wild-type SOD1-YFP
fusion did not form aggregates. We wished to address whether
a similar mutant SOD1-YFP fusion protein expressed in mice
would produce an ALS phenotype. If so, this could enable
several avenues of investigation: fluorescent reporting in vivo of
the distribution of expression and aggregation of the misfolded
protein; fluorescent monitoring to facilitate purification of
SOD1-YFP aggregates from spinal cord, allowing identification
of the protein constituents; and immunoaffinity capture of
soluble fusion protein species, including soluble oligomers,
through the fluorescent protein moiety, allowing identification
of associating proteins. We report here on such studies.
Transgenic Mice Expressing a G85R SOD1-YFP Fusion But Not WT
SOD1-YFP Develop ALS-Like Disease. We produced strains of trans-
genic mice expressing fusion proteins, either G85R SOD1-YFP
(G85R SOD-YFP) or wild-type SOD1-YFP (WT SOD-YFP),
formed by fusing the respective cloned human SOD1 gene
through its C-terminal coding sequence in exon 5 to the YFP
coding sequence (Fig. S1). G85R SOD1 is a misfolded mutant of
SOD1 that has been associated with ALS disease in humans and
transgenic mice (9, 10). Because SOD1 (hereafter referred to as
SOD) is broadly expressed, we were readily able to identify
transgenic SOD-YFP animals and genotype them based on the
fluorescence of the pups’ skin before growth of fur. Animals
heterozygous for G85R SOD-YFP transgene insertions did not
develop symptoms, but when homozygous, 2 lines with strong
Author contributions: J.W., G.W.F., C.J.Z., D.R.G., D.T.R., R.J.K., C.I.R., J.R.Y., M.B.F., and
A.L.H. designed research; J.W., G.W.F., C.J.Z., D.R.G., J.H.W., K.F., D.T.R., C.I.R., M.B.F., and
A.L.H. performed research; C.I.R., J.R.Y., and S.P. contributed new reagents/analytic tools;
J.W., G.W.F., C.J.Z., D.R.G., D.T.R., R.J.K., C.I.R., M.B.F., and A.L.H. analyzed data; and J.W.,
G.W.F., and A.L.H. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1J.W. and G.W.F. contributed equally to this work
2Present address: Department of Anatomy and Cell Biology, University of Iowa Carver
College of Medicine, Iowa City, IA 52242
3To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2009 by The National Academy of Sciences of the USA
February 3, 2009 ?
vol. 106 ?
fluorescence, referred to as 641 and 737, developed motor
neuron disease before a year of age. The age of onset of visible
motor dysfunction in homozygous 641 animals was consistently
9 months, whereas onset in homozygous 737 animals varied from
4 months to 12 months. This correlated in Southern blot analyses
with 641 carrying a single (multicopy) transgene insertion,
whereas 737 proved to carry multiple insertions, such that motor
of insertions were present (data not shown). WT SOD-YFP
transgenics, in contrast, did not exhibit disease out to beyond 2
years, even when homozygous for transgene insertions that
produced a higher steady-state level of fusion protein in spinal
cord than achieved with G85R SOD-YFP. Because of the
reproducible trajectory of disease, homozygous 641 animals
were studied in detail.
Fluorescent Aggregates in Motor Neuron Cell Bodies and Processes in
Spinal Cords of Symptomatic Animals. When the spinal cord from
a 1-month-old 641 G85R SOD-YFP animal was rapidly fixed and
examined in cross-section, strong but diffuse fluorescence was
observed in large cell bodies in the gray matter (Fig. 1A), most
at higher magnification to extend into processes (Fig. 1B). The
same pattern was also observed at 6 months of age (Fig. S2a).
When spinal cord from a symptomatic animal of 9 months age
was examined, an entirely different pattern was observed, with
intense punctate fluorescent staining now observed in the gray
matter (Fig. 1C), and with many neuron cell bodies containing
punctate green fluorescent aggregates (Fig. 1D). Additional
punctate aggregates localized in the neuropil (Fig. 1D), in some
cases in contiguous processes. In comparison with the spinal
cords from G85R SOD-YFP animals, cords from WT SOD-YFP
animals from multiple lines of different ages up to 1 year, some
exhibiting greater levels of total steady-state SOD-YFP protein
in Western blot analysis, exhibited only low levels of fluores-
cence (Fig. S2b) and, at higher magnification, exhibited only a
small amount of fluorescence in large neuron cell bodies (e.g.,
Ubiquitin and GFAP Reactivity. When spinal cords were fixed and
stained with anti-SOD antiserum, the same patterns as had been
observed with fluorescence were obtained (Fig. S3a). Cords
were also stained with anti-ubiquitin antibodies (Fig. S3b),
revealing no significant staining in cords of 6-month-old G85R
SOD-YFP animals or WT SOD-YFP animals but a distribution
pattern in symptomatic 9-month-old mutant cord corresponding
to that of the fluorescent aggregates, with punctate staining both
within motor neuron cell bodies and outside. Ubiquitin-positive
staining is classically observed in symptomatic conditions of both
sporadic and SOD-linked ALS (1, 2). In parallel with late
development of ubiquitin-positive staining, the gray matter of
symptomatic 9-month-old G85R SOD-YFP animals, but not of
6-month-old animals, stained positively for GFAP, reflecting a
glial response (Fig. S3c). In EM studies of a symptomatic G85R
SOD-YFP animal, aggregates were observed in many motor
neuron cell bodies (23/50) but not in astrocyte cell bodies (0/50)
(representative image of mutant in Fig. S4). In additional
fluorescence analysis of symptomatic animals, fluorescent ag-
gregates were observed in ventral nerve root axons (Fig. S5),
with a higher density observed proximally relative to distally, for
example, in sciatic nerve.
In Situ RNA Hybridization Reveals That SOD Transgenes Are Expressed
Predominantly in Motor Neurons in the Spinal Cord. The selective
fluorescence of large motor neuron cell bodies in the spinal cord
of G85R SOD-YFP transgenic animals raised a consideration of
whether neurons are the principal site of expression of the fusion
protein. To address this question, we carried out in situ RNA
symptoms. Cords were rapidly dissected, embedded in OCT, frozen, sectioned in a cryotome, and transferred to a glass slide. The slides were then examined by
fluorescence microscopy. Shown are low (A and C) and higher (B and D) magnification views of the same sections, showing diffuse fluorescence in large neuron
cell bodies in the young animal, and punctate fluorescence both within cell bodies and outside of them in the symptomatic animal.
Fluorescent images of spinal cord cross-sections from a 1-month-old G85R SOD-YFP mouse and from a 9-month-old G85R SOD-YFP animal with motor
Wang et al.
February 3, 2009 ?
vol. 106 ?
no. 5 ?
hybridization, using a YFP probe, which distinguishes transcrip-
tion of the transgene from that of endogenous SOD1. This
revealed prominent expression in large cell bodies in the gray
matter of both presymptomatic (data not shown) and symptom-
atic G85R SOD-YFP animals but also in WT SOD-YFP trans-
genic animals (Fig. S6). Minimal staining was observed in
smaller cell bodies. The strong expression of the RNA in
neurons, including motor neurons, is consistent with neurons as
the major site of SOD-YFP protein in the transgenic animals
(Fig. 1). This would imply that the fluorescent aggregates in the
neuropil of symptomatic animals are likely to lie within neuronal
Biochemical Fractionation Reveals Progressive Formation of Soluble
Oligomers and Aggregates. Solubility. The fate of the G85R SOD-
YFP protein over the course of disease progression was tracked
by steps of fractionation. First, homogenization of spinal cord
from either G85R SOD-YFP or WT SOD-YFP transgenic
animals was carried out under native conditions, followed by
centrifugation (17,000 ? g ? 15min). In Western blot analysis,
using a polyclonal anti-YFP antibody, we observed that all of the
fusion protein (43 kDa) was present in the soluble fraction of
both WT SOD-YFP animals and 1-month-old G85R SOD-YFP
animals (Fig. 2A Left). Similarly, when cords from wild-type
animals of any age were analyzed, the protein was entirely
old G85R SOD-YFP animals, ?25% of the protein was recov-
ered in the insoluble fraction, and from 9-month-old symptom-
atic animals, ?50% of the fusion protein was localized in the
insoluble fraction. Thus, the mutant fusion protein becomes
progressively less soluble with age.
G85R-YFP behaves as a misfolded monomer. To further evaluate the
behavior of the G85R SOD-YFP protein, the soluble fractions
from spinal cords of animals of the various ages were subjected
to gel filtration chromatography on Superose 6, monitored with
in-line fluorescence (Fig. 2B). First we evaluated a symptomatic
G85R SOD-YFP animal, comparing it with a WT SOD-YFP
one. Whereas SOD-YFP from wild-type cord eluted at the
position of a dimer (?80–90 kDa), corresponding to the ho-
modimeric state that is observed for nonfused native wild-type
SOD, mutant G85R SOD-YFP exhibited a major fluorescent
peak at a position corresponding to SOD-YFP monomer (?45
kDa), suggesting that dimerization of the mutant fusion protein
had failed to occur. This failure of assembly is consistent with
observations of Marklund and coworkers examining spinal cord
of transgenic mice with nonfused G85R SOD (11).
To address whether the unassembled G85R SOD-YFP was
properly folded, we assessed whether the disulfide bond nor-
mally present in native SOD (Cys-57–146) had been formed in
the fusion protein from mutant animals. Tissue extracts were
prepared in the presence of iodoacetamide to block free thiols,
then the migration of the fusion protein was analyzed in SDS/
PAGE in the absence or presence of DTT (Fig. 2C). As observed
in earlier studies, if the disulfide is present, then the protein
migrates more rapidly under nonreducing conditions (?DTT)
has not been formed, then the protein migrates at the ‘‘reduced’’
position in both the absence and presence of DTT. Whereas
SOD-YFP from cord of wild-type animals exhibited differential
migration reflecting the presence of the disulfide bond (Fig. 2C
Left), the fusion protein from mutant animals, including animals
at 1 month of age, exhibited relatively slower migration, corre-
sponding to the reduced protein, both in the absence and
presence of DTT (Fig. 2C Right). The lack of disulfide bond
formation in G85R SOD-YFP corresponds once again to be-
havior observed for nonfused G85R SOD from spinal cord of
transgenic mice (12). We conclude that G85R in either the
nonfused or YFP-fused context is a misfolded monomeric pro-
Soluble oligomers of G85R SOD-YFP at 6 and 9 months. We noticed that
a ‘‘shoulder’’ of fluorescence between 25 and 30 min of elution,
corresponding to higher molecular weight species, was present in
the gel filtration analysis of the symptomatic mutant animal (Fig.
2B, red trace), and analyzed these fractions by immunoblotting
with anti-YFP antibodies (Fig. 2D). In WT SOD-YFP animals
and in 1-month-old mutant animals, the fusion protein migrated
at lower molecular masses, corresponding to the sizes of the
dimer or monomer forms, respectively. Strikingly, however,
when mutant animals of 6 months or 9 months of age were
examined, the fusion was detected in higher molecular weight
fractions, with a portion of the material from the 9-month-old
found at the void volume of the column, corresponding to 4–6
MDa. Thus, there is an age-dependent progression in the
formation of soluble oligomers in parallel with progressive
production of insoluble protein. These two states likely comprise
a precursor-product relationship. These data raised questions
both about whether there are cellular components associating
with soluble misfolded G85R SOD-YFP and about the compo-
sition of the end-state insoluble aggregates. These were ad-
dressed using proteomic analyses.
Physical Associations of Soluble G85R SOD-YFP in Spinal Cord—Strong
Association with Hsc70 at All Ages and Association with 3 Hsp110
Proteins at Later Times. To detect physical association of soluble
G85R SOD-YFP with other proteins, the soluble fractions of
spinal cords from G85R SOD-YFP or WT SOD-YFP animals of
various ages were applied to anti-YFP antibody columns. After
extensive washing in buffer, columns were eluted with 10 M urea
in 20 mM HCl, and the products were analyzed by MudPIT,
involving proteolysis with trypsin, cation exchange chromatog-
raphy coupled with C18 reverse phase chromatography, and
mass spectrometric analysis in a Thermo LTQ Orbitrap XL (see
Experimental Procedures). As shown in the top row of Table 1,
the most abundant eluted species observed was the fusion
protein itself. Because equal amounts of tissue were extracted in
each case, it appears that a higher amount of the fusion protein
may be present in symptomatic G85R SOD-YFP cord (at 9
In addition to the fusion protein, we observed a small collec-
tive of coeluted species (Table 1). One of these proteins, the
copper chaperone CCS, which normally forms a heterodimer
with SOD monomer during biogenesis to deliver copper (13),
was equivalently coeluted with both the mutant and wild-type
fusion. In a second such partnering, endogenous mouse SOD1
was observed to be specifically captured by WT SOD-YFP,
apparently forming a heterodimer with it, but the endogenous
SOD1 was only marginally captured by G85R SOD-YFP, con-
sistent with the misfolded monomeric status of the latter protein,
indisposed to heterodimerize.
Soluble G85R SOD-YFP specifically captured a number of
additional components. Most prominent was the molecular
chaperone Hsc70, with a large number of spectral counts ob-
served to associate with the mutant fusion protein in animals of
1, 6, and 9 months of age (Table 1). By contrast, ?25% as many
spectral counts of Hsc70 were associated with WT SOD-YFP.
of soluble G85R SOD-YFP and WT SOD-YFP cord extracts
were directly analyzed by Western blot analysis with anti-Hsc70
antibodies, Hsc70 was exclusively observed in the G85R SOD-
YFP eluate (Fig. 3 Left). In an additional affinity capture
experiment, an SDS gel of the eluted material was directly
Coomassie-stained, allowing estimation that the amount of
Hsc70 recovered with G85R SOD-YFP corresponds to ?10%
the amount of the fusion protein (Fig. 3 Right).
In addition to association of G85R SOD-YFP with Hsc70,
www.pnas.org?cgi?doi?10.1073?pnas.0813045106 Wang et al.
specific elution with G85R SOD-YFP but not WT SOD-YFP
was observed for 3 additional chaperone proteins, the mouse
Hsp110 family members, Hspa4l, Hsph1, and Hspa4. Hsp110
proteins have recently been shown to function as nucleotide
exchange factors for Hsp70 class proteins (14–17), and these 3
components seem likely to function as the normal exchangers for
Hsc70. Interestingly, in mutant animals, the association of these
chaperones was negligible at 1 month of age but was appreciable
at 6 and 9 months. In an opposite pattern, the Hsp70 protein,
Hspa1b, showed significant spectral counts at early time but
lower counts at the later times. Available antibodies were not
sufficiently specific to allow detection of Hsp110 and Hsp70
components by Western analysis.
Insoluble Aggregates from Mutant Cord Contain Mutant SOD-YFP and
Intermediate Filament Proteins as the Major Constituents. To char-
acterize insoluble aggregates, spinal cords from symptomatic
G85R SOD-YFP animals and WT SOD-YFP animals were
identically fractionated using steps of centrifugation and deter-
gent extraction that enriched for the fluorescent aggregates from
mutant cord (see SI Methods). This would enrich for both
intracellular aggregates and fluorescent accretions from dead
cells. The final fractions were subjected to MudPIT. There were
34 proteins identified with ?30 spectral counts (Table S1). Most
of the strongly registering proteins in the aggregate fractions are
abundant cytoskeletal proteins, with the class of intermediate
filament proteins (NF-M, NF-H, GFAP, and vimentin) regis-
tering strongly and mutant-specifically (see SI Discussion for
Temporal Progression of Soluble Oligomer Formation and Aggrega-
tion of Misfolded G85R SOD-YFP. Experiments presented here with
G85R SOD-YFP and wild-type SOD-YFP transgenic mice
indicate that these fusion proteins, when expressed from the
motor neurons in the spinal cord, and that the G85R SOD-YFP
mice specifically develop a lethal motor disease that resembles
ALS, associated with late appearance of green fluorescent
aggregates in motor neuron cell bodies and processes. In bio-
chemical studies, the SOD G85R-YFP protein was observed to
behave as a misfolded monomer, and the constitutive chaperone
Hsc70 was observed to associate with it at all ages. Interestingly,
a concurrent study of ALS-affected transgenic mice expressing
a truncated SOD1 that was FLAG-tagged at its C terminus, also
behaving as a misfolded monomer, showed that Hsc70 was
brought down by anti-FLAG immunoprecipitation (18). To-
gether, these data suggest that this chaperone may comprise a
primary site of interaction.
A progression of misbehavior of the mutant protein in spinal
cord could be observed (Fig. 4). At 1 month it was entirely
soluble, associated to some degree with Hsc70, and displayed
diffuse green fluorescence in neuronal cell bodies and processes.
By 6 months of age it now recruited Hsp110 proteins along with
Hsc70, formed soluble oligomers observed in gel filtration, and
partitioned to a significant extent into the insoluble fraction. At
Hsc70/Hsp110 chaperone complexes and soluble oligomers con-
animal was also analyzed under nonreducing conditions and did not reveal
higher molecular weight species that would reflect intermolecular disulfide
bond formation.(D) Immunoblot analysis of gel filtration fractions of soluble
extracts from WT SOD-YFP and G85R SOD-YFP animals. Migration position of
size markers is indicated above the images. The void volume fraction (Vo)
corresponds to ?3 MDa.
and insoluble (P) fractions. Cords homogenized in nondenaturing buffer were
centrifuged at 17,000 ? g for 15 min and the supernatant directly solubilized in
SDS sample buffer. Pellets were washed in nondenaturing buffer and then
solubilized in SDS sample buffer. Identical aliquots of the fractions were loaded
onto an SDS/PAGE gel. (B) Fractionation of soluble extracts by gel filtration on a
Superose 6 column with in-line fluorescence. Note that G85R SOD-YFP migrates
of a dimer. A ‘‘shoulder’’ of fluorescence migrating between 25 and 30 min was
reproducibly present for the G85R SOD-YFP. (C) Analysis of soluble extracts
prepared in iodoacetamide under reducing (?DTT) and nonreducing (?DTT)
conditions. The migration positions of oxidized and reduced forms of WT SOD-
are indicated. As in other studies, the G85R SOD-YFP species, both reduced and
oxidized, migrate in SDS/PAGE slightly more rapidly than the respective WT
Biochemical fractionation of spinal cords from WT SOD-YFP and G85R
Wang et al.
February 3, 2009 ?
vol. 106 ?
no. 5 ?
substantially increased, associated with the morphologic appear-
ance of visible, green fluorescent aggregates in both neuronal
cell bodies and processes.
Nature of Progression to Oligomer Formation. The cause of the
progression of oligomerization and aggregation of the mutant
protein remains to be determined, with the transition to soluble
6 months likely to be particularly informative. Is this an effect of
increasing expression of the mutant protein itself as a response
of the SOD promoter? In additional in situ analyses similar to
Fig. S6, the amount of mRNA appears to increase with age, but
further quantitative studies will be needed to establish whether
Table 1. Anti-YFP-captured proteins
SOD1 - mouse
Hspa5 (Grp 78)
Dnaja1 (Dnaj A1)
Proteins identified by Mud-PIT analysis of affinity captured SOD-YFP with Spectral Counts above 20 for the G85R SOD-YFP 9 month-old sample are shown.
These spectral counts represent the number of times the analysis identified a peptide corresponding to the assigned protein and can be viewed as a relative
measure of protein abundance. Notably, when a peptide was common to two related proteins that were identified to be present by other unique peptides, the
peptide was scored with both proteins. The spectral counts that are unique to the identified protein are shown in parenthesis. Percent sequence coverage, the
percentage of the primary structure of the identified protein covered by peptides, is also reported. Dashes indicate the absence of any peptide for the indicated
(Right) of fractions eluted from anti-YFP chromatography of soluble extract
antibody to an Hsc70-specific peptide (SPA-816) was obtained from Assay
Immunoblot analysis with anti-Hsc70 (Left) and Coomassie staining
constitutive chaperone Hsc70. In a 1-month-old animal, the mutant protein is
entirely soluble as indicated by fluorescence and biochemical analyses. A
fraction is likely to be trafficked to the proteasome and degraded. At 6
months, there is oligomerization and insolubility, despite detectable associ-
ation at this point of Hsp110 exchange factors, presumably with Hsc70. The
time when animals are symptomatic, there are visible and readily detectable
insoluble aggregates, composed, as observed here by mass spectrometry (see
Table S1), principally of intermediate filaments and mutant SOD.
Model for temporal progression of behavior of G85R SOD-YFP. At all
www.pnas.org?cgi?doi?10.1073?pnas.0813045106Wang et al.
the level of mRNA progressively increases in the mutant setting,
potentially suggestive of a positive feedback effect of misfolding/
stress upon the SOD promoter itself. An alternative explanation
of the progression to aggregation is a potential effect of aging on
the chaperone/proteolytic system that is the likely cellular mech-
anism for handling the mutant protein. An age-related decline
in 1 or more such components could herald the decompensation.
Whatever the answer, it is striking that a stress response is not
mounted at any point against the mutant protein, at least as
judged by the lack of binding of stress-responsive proteins to it,
despite it being present at the level of a few percent of soluble
protein in motor neurons. (Typically ?10 ?g of fusion protein
was recovered per spinal cord of a total of ?800 ?g of soluble
protein). Although we observed a small level of associated
Hsp70 and Grp78 in 1-month-old mutant animals, the amounts
recovered were reduced at later times (Table 1). We note further
that a UPR response was not present in symptomatic mutant
animals (Fig. S7), contrasting with the reports in refs. 19 and 20.
Ubiquitination and Degradation of Misfolded G85R SOD-YFP.Thefate
of Hsc70-bound G85R SOD-YFP protein in 1-month-old ‘‘com-
pensated’’ animals is of interest to consider. A dynamic Hsc70-
DnaJ cycle could function to maintain the misfolded protein in
a soluble state. Presumably, however, the misfolded protein must
ultimately be turned over. How efficiently is it ‘‘handed off’’ to
the proteolytic apparatus as opposed to cycling on and off of
Hsc70? Interestingly, the level of ubiquitination of both the
G85R and the wild-type SOD-YFP fusion protein was relatively
low in 1- to 2-month-old animals but rose substantially in both
cases by 6–9 months of age, with ubiquitin modification occur-
ring at the later times principally on lysine 10 of the SOD moiety
(Table S2). With respect to turnover, we are currently measuring
rates for both the mutant and wild-type fusion proteins in vivo.
Notably, however, G85R SOD has been observed to turn over at
a more rapid rate than wild-type SOD in cell culture studies (21).
This seems inconsistent with the observations of significant
diffuse fluorescence in neuronal cell bodies in mutant transgenic
animals but not wild-type at 1 and 6 months of age. Perhaps,
however, the observed fluorescence in neuron cell bodies of our
mutant animals reflects slow diffusion of the mutant protein, as
opposed to slow turnover. The nature of neurotoxicity of the
mutant fusion protein, possibly conferred by soluble oligomers
(refs. 22 and 23, but see also ref. 8), also remains unclear.
Concerning cellular targets of toxicity, as mentioned, we failed
to observe a UPR response in symptomatic animals (Fig. S7),
and we also did not detect abnormalities of mitochondria, which
appeared morphologically normal upon EM inspection and did
not exhibit significant fluorescence when isolated (data not
shown). Thus, the nature of toxicity remains to be resolved.
In general, at least 4 animals were examined for each of the analyses pre-
sented, with the exception of studies of 1-month-old mutant animals, where
only 2 or 3 were used. Representative images and experiments are presented.
Tissue Microscopy. Fluorescent microscopy was performed on spinal cords
embedded in OCT (Sakura Finetek) and frozen in 2-methybutane cooled with
liquid nitrogen and stored at ?80 °C. Frozen tissue was sectioned (20 ?m),
using a Leica CM3000 Cryostat at ?20 °C and preserved with Vectashield
(VectorLabs). The tissue was observed using the green filter on an Axioskop2
Plus microscope (Zeiss).
Biochemical Analysis. Spinal cords from transgenic mice of various ages were
homogenized on ice in 0.5 mL of PBS containing 1 mM each EDTA, EGTA, and
TCEP, and 1 tablet/ml of Complete protease inhibitor mixture (Roche). After
centrifugation at 17,000 ? g for 15 min, the supernatant was fractionated on
were analyzed by SDS/PAGE and Western blot analysis, using an affinity-
purified rabbit polyclonal antibody to YFP.
to affinity-capture, using the anti-YFP antibody cross-linked with DSS to
Ultralink protein A (Thermo-Fisher). After extensive washing in PBS, affinity-
captured proteins were eluted using 20 mM HCl, 10 M urea, then neutralized
by addition of Tris (pH 8), to 200 mM. The eluted samples were reduced and
alkylated with 2 mM TCEP and 10 mM iodoacetamide. MudPIT analysis was
then carried out as described in ref. 24, using an LTQ Orbitrap XL mass
spectrometer (Thermo-Fisher). Data were analyzed using SEQUEST, and were
then inspected with DTASelect (25).
ACKNOWLEDGMENTS. We thank David Borchelt for a human SOD1 genomic
clone, Mikael Oliveberg for plasmids directing expression of human SOD1 in
E. coli, David Sarracino for help with initial proteomic studies, Gordon Ter-
was supported by Howard Hughes Medical Institute and NIH.
1. Bruijn LI, Miller TM, Cleveland DW (2004) Unraveling the mechanisms involved in
motor neuron degeneration in ALS. Annu Rev Neurosci 27:723–749.
2. Pasinelli P, Brown RH (2006) Molecular biology of amyotrophic lateral sclerosis: In-
sights from genetics. Nat Rev Neurosci 7:710–723.
3. Lindberg MJ, Bystro ¨m R, Bokna ¨s N, Andersen PN, Oliveberg M (2005) Systematically
perturbed folding patterns of amyotrophic lateral sclerosis (ALS)-associated SOD1
mutants. Proc Natl Acad Sci USA 102:9754–9759.
4. Valentine JS, Doucette PA, Potter SZ (2005) Copper-zinc superoxide dismutase and
amyotrophic lateral sclerosis. Annu Rev Biochem 74:563–593.
5. Hart PJ (2006) Pathogenic superoxide dismutase structure, folding, aggregation and
turnover. Curr Op Chem Biol 10:131–138.
6. SvenssonAK,BilselO,KondrashkinaE,ZitzewitzJA,MatthewsCR (2006)Mappingthe
folding free energy surface for metal-free human Cu,Zn superoxide dismutase. J Mol
7. Reaume AG, et al. (1996) Motor neurons in Cu/Zn superoxide dismutase-deficient mice
8. Matsumoto G, Stojanovic A, Holmberg CI, Kim S, Morimoto RI (2005) Structural prop-
erties and neuronal toxicity of amyotrophic lateral sclerosis-associated Cu/Zn super-
oxide dismutase 1 aggregates. J Cell Biol 171:75–85.
9. Rosen DR, et al. (1993) Mutations in Cu/Zn superoxide dismutase gene are associated
with familial amyotrophic lateral sclerosis. Nature 362:59–62.
10. Bruijn LI, et al. (1997) ALS-linked SOD1 mutant G85R mediates damage to astrocytes
and promotes rapidly progressive disease with SOD1-containing inclusions. Neuron
11. Zetterstro ¨m P, et al. (2007) Soluble misfolded subfractions of mutant superoxide
dismutase-1s are enriched in spinal cords throughout life in murine ALS models. Proc
Natl Acad Sci USA 104:14157–14162.
12. Jonsson PA, et al. (2006) Disulphide-reduced superoxide dismutase-1 in CNS of trans-
genic amyotrophic lateral sclerosis models. Brain 129:451–464.
13. Lamb AL, Torres AS, O’Halloran TV, Rosenzweig AC (2001) Heterodimeric structure of
superoxide dismutase in complex with its metallochaperone. Nat. Struct Biol 8:751–755.
14. Raviol H, Sadlish H, Rodriguez F, Mayer MP, Bukau B (2006) Chaperone network in the
yeast cytosol: Hsp110 is revealed as an Hsp70 nucleotide exchange factor. EMBO J
15. Dragovic Z, Broadley SA, Shomura Y, Bracher A, Hartl FU (2006) Molecular chaperones
of the Hsp110 family act as nucleotide exchange factors of Hsp70s. EMBO J 25:2519–
16. Schuermann JP, et al. (2008) Structure of the Hsp110:Hsc70 nucleotide exchange
machine. Mol Cell 31:232–243.
17. Polier S, Dragovic Z, Hartl FU, Bracher A (2008) Structural basis for the cooperation of
Hsp70 and Hsp110 chaperones in protein folding. Cell 133:1068–1079.
18. Watanabe Y, et al. (2008) Adherent monomer-misfolded SOD1. PLoS ONE
19. Kikuchi H, et al. (2006) Spinal cord endoplasmic reticulum stress associated with a
microsomal accumulation of mutant superoxide dismutase-1 in an ALS model. Proc
Natl Acad Sci USA 103:6025–6030.
20. Nishitoh H, et al. (2008) ALS-linked mutant SOD1 induces ER stress- and ASK1-
dependent motor neuron death by targeting Derlin-1. Genes Dev 22:1451–1464.
21. Borchelt DR, et al. (1995) Superoxide dismutase 1 subunits with mutations linked to
familial amyotrophic lateral sclerosis do not affect wild-type subunit function. J Biol
22. Bucciantini M, et al. (2002) Inherent toxicity of aggregates implies a common mech-
anism for protein misfolding diseases. Nature 416:507–511.
23. Shankar GM, et al. (2008) Amyloid-? protein dimers isolated directly from Alzheimer’s
brains impair synaptic plasticity and memory. Nature Med 14:837–842.
24. Washburn MP, Wolters D, Yates JR, III (2001) Large-scale analysis of the yeast
proteome by multidimensional protein identification technology. Nat Biotech
25. Tabb DL, McDonald WH, Yates JR, III (2002) DTASelect and Contrast: Tools for assem-
Wang et al.
February 3, 2009 ?
vol. 106 ?
no. 5 ?