Common molecular signature in SOD1 for both

sporadic and familial amyotrophic lateral sclerosis

Arie Gruzman*, William L. Wood†, Evgenia Alpert*, M. Dharma Prasad*, Robert G. Miller‡, Jeffery D. Rothstein§,
Robert Bowser¶, Ronald Hamilton¶, Troy D. Wood†, Don W. Cleveland储**, Vishwanath R. Lingappa*, and Jian Liu**††
*Bioconformatics Laboratory and ††Department of Neuroscience, California Pacific Medical Center Research Institute, San Francisco, CA 94107;
†Department of Chemistry, Natural Sciences Complex, University at Buffalo, State University of New York, Buffalo, NY 14260; §Department

of Neurology, Johns Hopkins University, Baltimore, MD 21287; ‡Department of Neurology, California Pacific Medical Center, San Francisco, CA 94115;
储Ludwig Institute for Cancer Research, Departments of Medicine and Neuroscience, University of California at San Diego, La Jolla, CA 92093;

and ¶Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261

Contributed by Don W. Cleveland, June 4, 2007 (sent for review August 25, 2006)

Amyotrophic lateral sclerosis (ALS) is a devastating motor neuron
degenerative disease whose etiology and pathogenesis remain
poorly understood. Most cases of ALS (⬇90%) are sporadic (SALS),
occurring in the absence of genetic associations. Approximately
20% of familial ALS (FALS) cases are due to known mutations in the
copper, zinc superoxide dismutase (SOD1) gene. Molecular evi-
dence for a common pathogenesis of SALS and FALS has remained
elusive. Here we use covalent chemical modification to reveal an
attribute of spinal cord SOD1 common to both SOD1-linked FALS
and SALS, but not present in normal or disease-affected tissues
from other neurodegenerative diseases, including Alzheimer’s,
Parkinson’s, and Huntington’s diseases and spinal muscular atro-
phy, a non-ALS motor neuron disease. Biotinylation reveals a
32-kDa, covalently cross-linked SOD1-containing protein species
produced not only in FALS caused by SOD1 mutation, but also in
SALS. These studies use chemical modification as a novel tool for
the detection of a disease-associated biomarker. Our results iden-
tify a shared molecular event involving a known target gene and
suggest a common step in the pathogenesis between SALS and
FALS.
copper, zinc superoxide dismutase 兩 motor neuron 兩 neurodegeneration
A myotrophic lateral sclerosis (ALS) pathology and symptoms
in copper, zinc superoxide dismutase (SOD1)-linked famil-
ial ALS (FALS) patients closely resemble those of sporadic ALS
(SALS) patients. In light of the linkage of the SOD1 gene to
FALS (1), it has long been speculated that there may be a
common molecular basis for both forms of the disease, but no
such feature has been identified. A plethora of evidence has
demonstrated that mutant SOD1 causes motor neuron death by
gain of toxic properties (2, 3). However, mechanisms of mutant
SOD1-mediated toxicity to motor neurons remain elusive. Many
studies have investigated the possibility of structural differences
in SOD1 caused by mutations as a basis for such toxic effects (4,
5). Although several common features of mutant SOD1 proteins
such as instability, higher sensitivity to denaturing conditions,
and propensity to form aggregates are found to be associated
with mutant SOD1-linked FALS disease course and severity
(6–11), no single biochemical feature has been found to be
shared by all identified mutant SOD1s, nor has there been a
single molecular feature of SOD1 identified to be associated with
SALS.
We propose that SOD1 is more conformationally heteroge-
neous in vivo and that these ‘‘conformers’’ are relevant to ALS
but difficult to distinguish by conventional biochemical analyses.
To reveal such conformational differences especially those in-
trinsic to minor subspecies of SOD1 we applied a new approach
involving biotinylation of accessible lysine residues in proteins,
followed by SDS/PAGE and immunoblot analysis for SOD1.
Covalent chemical modification of amino acid side groups has
the advantage of rapidly proceeding to completion, being readily
quenched, and capable of potentially affecting antibody immu-
noreactivity (12). The extent of a chemical modification reaction
is strongly affected by the degree of the accessibility of a side
group to a modifying agent. This, in turn, is influenced by the
overall conformational structure of the protein, thereby provid-
ing a basis by which different conformers react differently to a
covalent modifying agent (13). Thus, a conformational differ-
ence in a small proportion of molecules or a subtle difference in
structure may be magnified. Using this approach, we now
provide direct evidence for a common underlying pathogenesis
of SALS and FALS through the identification of a unique,
crosslinked protein species of SOD1 that is common to both.
Results
A 32-kDa SOD1-Immunoreactive (IR) Protein Species Revealed by
Biotinylation of Spinal Cord Extracts of ALS Patients. Conventional
immunoblotting after protein denaturation under strongly re-
ducing conditions of spinal cord tissue extracts from ALS
patients or normal samples revealed only the expected 16-kDa
SOD1 polypeptide recognized by an antibody generated to a
peptide from human SOD1 (Fig. 1B, “⫺” lanes). After testing 12
commercially available chemical compounds that can modify
side-chains of amino acids of lysine, cysteine, aspartic acid,
glutamic acid, and arginine [see supporting information (SI)
Table 1], side-chain modification by covalent reaction with
N-hydroxysulfosuccinimide-long chain-biotin (Sulfo-NHS-LC-
Biotin), which covalently adds biotin to conformationally acces-
sible lysine side chains (Fig. 1 A), revealed a new 32-kDa species
in the SOD1-linked FALS sample [from a patient with a
dominant alanine to valine mutation in SOD1 at position 4
(SOD1A4V)], but not a normal sample (Fig. 1B). Surprisingly, a
similar biotinylation-dependent 32-kDa species was also seen in
a SALS sample (Fig. 1B).
Further characterization of the biotinylation reaction revealed
that the intensity of the 32-kDa species depended on time of
incubation (Fig. 1C), concentration of Sulfo-NHS-LC-Biotin,
the amount of the initial extract and incubation temperature
(Fig. 1 D–F). The reaction was highly SOD1 protein conforma-
Author contributions: A.G., W.L.W., T.D.W., V.R.L., and J.L. designed research; A.G., W.L.W.,
E.A., M.D.P., and J.L. performed research; R.G.M., J.D.R., R.B., R.H., and D.W.C. contributed
new reagents/analytic tools; A.G., W.L.W., T.D.W., D.W.C., V.R.L., and J.L. analyzed data;
and A.G., D.W.C., V.R.L., and J.L. wrote the paper.
Conflict of interest statement: V.R.L. is a founder and J.L. a consultant to Prosetta Corpo-
ration, an early stage biotechnology company.
Abbreviations: AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; FALS, familial
amyotrophic lateral sclerosis; HD, Huntington’s disease; Hip, anterior hippocampus; IR,
immunoreactive; PD, Parkinson’s disease; SALS, sporadic amyotrophic lateral sclerosis;
SMA, spinal muscular atrophy; SOD1, copper, zinc superoxide dismutase; Sulfo-NHS-LC-
Biotin, N-hydroxysulfosuccinimide-long chain-biotin.
**To whom correspondence may be addressed. E-mail: dcleveland@ucsd.edu or liuj@
cpmcri.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
0705044104/DC1.
© 2007 by The National Academy of Sciences of the USA

12524 –12529 兩 PNAS 兩 July 24, 2007 兩 vol. 104 兩 no. 30 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705044104

Fig. 1. A 32-kDa SOD1-IR protein species revealed by biotinylation of spinal
cord extracts of ALS patients. (A) Schematic diagram of the detection of a 32-kDa
SOD1-IR protein species by immunoblot analysis. Sulfo-NHS-LC-Biotin reacts with
structurally available NH2 groups (outside the rings) from the lysine residues in the
SOD1 protein resulting in the modification of the protein at the sites of the
side-chains (indicated by *). The biotinylated SOD1 proteins were separated on
SDS/PAGE gel and immunorecognized by the SOD1 antibody (darker ring). (B–G)
Spinal cord cytosolic proteins were incubated with (⫹) or without (⫺) Sulfo-NHS-
LC-Biotin, separated on SDS/PAGE gel, and immunoblotted with the SOD1 anti-
body. (B) The FALS sample was from an A4V mutation. Cytosolic proteins from
SALS-4 subject were used for C–G. (C) Time course of the reaction. One hundred
percent value was taken for the density of the 32 kDa at 150-min reaction time.
(D) Sulfo-NHS-LC-Biotin dose responses. One hundred percent value was taken
for the density of the 32 kDa at 100 mM biotin concentration. (E) Dose–response
of the total amount of proteins used. One hundred percent value was taken at a
total amount of 20 ␮g of proteins. (F) Temperature dependence of the reaction.
The optimal temperature for this reaction lies around 25°C. (G) Conformation-
specificity of the reaction. NHS-digoxigenin (1 mM) was added before the reac-
tion can block the biotinylation reaction resulting in the disappearance of the
32-kDa protein species (two middle lanes). All error bars represent SD from three
independent experiments.
tion-specific, because substrates containing partial or related
structural groups to Sulfo-NHS-LC-Biotin including sulfo-NHS,
biotin, long chain (LC)-biotin, iodoacetyl-LC-biotin, NHS-
Squarin carboxilate (data not shown), and NHS-digoxigenin.
(Fig. 1G) did not result in ALS-specific change of SOD1
proteins. Maximum intensity of the 32-kDa SOD1 species was
reached after 25 min at an optimal temperature of 25°C in the
presence of 10 mM biotin. This condition was used in all efforts
below.
The 32-kDa SOD1 Adduct Is Common to Both SALS and FALS. To
determine whether the appearance of the 32-kDa SOD1-IR
species in the initial SALS and SOD1-linked FALS samples after
reaction with Sulfo-NHS-LC-Biotin was a general feature of
FALS and SALS, a series of spinal cord extracts from 19 SALS
Gruzman et al.
and four FALS with mutations in the SOD1 gene were exam-
ined. After biotinyation, an increased intensity of the 32-kDa
protein species IR to SOD1 antiserum was found in all of the
ALS samples (Fig. 2A). In 14 of 16 normal samples, this species
was absent or at much reduced level, although the expected 16
kDa species representing monomeric SOD1 was found in com-
parable amounts before and after biotinylation (Fig. 2 A). The
signal intensities of the 32-kDa species were significantly higher
in spinal cord samples from ALS versus normal subjects when
determined by either combining FALS and SALS samples (Fig.
2B; 118.0 ⫾ 46.5, n ⫽ 23 for all ALS and 25.1 ⫾ 34.2, n ⫽ 16,
for normal; P ⬍ 0.0001) or treating them separately (105.6 ⫾
37.8, n ⫽ 19; P ⬍ 0.0001 for SALS and 176.7 ⫾ 81.0, n ⫽ 4; P ⬍
0.0001 for FALS). Taken together, there is an ⬇5-fold increase
in the 32-kDa signal intensity in all ALS compared with normal
control (Fig. 2B). In addition, when a sample that contains a
mutation in another ALS-linked gene, Senetaxin (14) (also
known as ALS4), was examined, the 32-kDa signal was also
present (Fig. 2C).
To further determine the area within the spinal cord that gives
rise to this 32-kDa signal, we separated the spinal cord into gray
and white matter and analyzed them separately. As shown in Fig.
2D, the 32-kDa signal was exclusively associated with the gray
matter where motor neuron, glia, and other cells reside in
contrast to the white matter containing different axonal tracks,
including the cortical-spinal track. Because these assays compare
constant amounts of protein before and after biotinylation, the
increased immunoreactivity is likely due to enhanced antigen
availability to the antiserum as a result of biotinylation.
SOD1 mutations are also observed in SALS, albeit at a very
low frequency (⬍5%) (15–18). To eliminate the possibility that
the 32-kDa signals from some SALS subjects could come from
subjects that might harbor SOD1 mutations, the coding se-
quences were determined for the SOD1 genes from spinal cord
autopsy samples from SALS-10, -11, -14, -15, and -17, samples
which had signal intensities of the 32 kDa that were comparable
to those from FALSs. No SOD1 mutations were found (data not
shown) in these five samples (Fig. 2 A). Although we cannot rule
out that there is no SOD1 mutation in the remaining 14 SALS
NEUROSCIENCE
(we have 19 total, minus 5 sequenced) samples analyzed, the
32-kDa signals from the SALS samples represent a typical
population of non-SOD1 mutant-linked SALS, rather than
SOD1-linked FALS.
The 32-kDa SOD1 Species Is Specific for Disease-Affected Tissues in
ALS but Not Those in Alzheimer’s Disease (AD), Parkinson’s Disease
(PD), Huntington’s Disease (HD), and Spinal Muscular Atrophy (SMA).
To understand whether the 32 kDa is unique to ALS pathology
as opposed to a reflection of a process derived simply from a
prolonged agonal state, we further examined this signal in similar
samples from three additional non-ALS neurodegenerative dis-
eases, including AD, PD, and HD. Two spinal cord samples were
initially examined from PD subjects. The intensities of the 32-
kDa signal were comparable to those from normal subjects and
were significantly lower than those from ALS (Fig. 3B). As spinal
cord is not the primary site of pathology in PD, we also examined
affected brain regions in AD, PD, and HD together with the
motor cortical area as the common central nervous tissue for all
diseases. Hippocampus is one of the areas severely affected in
AD. When examined, there was no significant increase of the 32-
kDa signal in this brain area compared with the much increased
level in ALS spinal cord (P ⬍ 0.0001, Fig. 3 A and D). Similar
results were found for putamen, an affected brain region in both
PD and HD (Fig. 3D, P ⬍ 0.0001 and P ⬍ 0.02, respectively,
compared with ALS spinal cord). For the motor cortexes, the
intensities of the 32-kDa signal in ALS, AD, PD, and HD were
significantly lower when compared with those from ALS spinal
cords (Fig. 3D, P ⬍ 0.001, P ⬍ 0.0001, P ⬍ 0.0001, and P ⬍ 0.001,
PNAS 兩 July 24, 2007 兩 vol. 104 兩 no. 30 兩 12525
Fig. 2. Prevalence of the 32-kDa SOD1-IR protein species in both SALS and FALS spinal cords after biotinylation. Immunoblot analysis of cytosolic proteins from
human spinal cord autopsy samples (A and B) incubated with (⫹) or without (⫺) Sulfo-NHS-LC-Biotin. The density of the 32-kDa SOD1-IR signal from SALS-1 spinal
cord autopsy sample (indicated as SALS-std) was chosen as a 100%, which was run on all gels as the standard for spinal cord samples. (A) Spinal cords are from
normal, SALS, and SOD1-linked FALS autopsy samples. Mutations in SOD1 are G93C, A4V, G127X, D101G for FALS-1,2,3,4, respectively. (B) Bar graphs of intensities
of the 32-kDa signal from spinal cords for all ALS (SALS ⫹ SOD1-linked FALS, n ⫽ 23) and normal (n ⫽ 16). Error bars represent SD. *, P ⬍ 0.0001. (C) Subject with
a mutation in the Senetexin gene (ALS4). (D) The spinal cord samples from one normal and one SALS subject were separated as the white and gray matter before
the analysis.

respectively). Furthermore, the difference in the intensity of the
32-kDa signal in the motor cortex between ALS and AD or PD
is also significant (Fig. 3D, P ⬍ 0.02). This motor cortex value
between ALS and HD did not quite reach statistical difference
(P ⫽ 0.05) because of the large variation and small sample size
(n ⫽ 4) for HD (Fig. 3D). The higher level of the 32 kDa in the
motor cortex of ALS but not AD or PD is consistent with the
known upper motor neuron pathology in ALS, demonstrating
the selective pathological role of the 32 kDa in ALS. Lastly, to
gain some insight on how widely the 32 kDa might be involved
in motor neuron damage in other motor neuron diseases, we
examined spinal cord tissues from subjects with Type 1 form of
SMA. The 32 kDa was not present in the four SMA spinal cords
examined (Fig. 3C and D, P ⬍ 0.0001). Motor neuron death in
Type I SMA differs from that of ALS by the fact that it is a much
accelerated process and devoid of aging factors. Together, these

Fig. 3. The 32-kDa SOD1-IR protein species is specific for ALS pathology. Immunoblot analysis of cytosolic proteins from different brain region autopsy samples
were incubated with (⫹) or without (⫺) Sulfo-NHS-LC-Biotin. The density of the 32-kDa SOD1-IR signal from spinal cord autopsy sample SALS-1 (indicated as
SALS-std) was chosen as a 100%, which was run on all gels as the standard for all samples. (A) The anterior hippocampus (Hip) and primary motor cortex (MC)
from AD subjects were analyzed. The spinal cord from two PD subjects (B) and four SMA patients (C) was analyzed. (D) Bar graphs of the relative intensities the
32-kDa signal from ALS spinal cords (SC) and MC, AD Hip and MC, PD putamen (PU) and MC, HD PU and MC, and SMA SC.

12526 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705044104 Gruzman et al.

Fig. 4. Antigen-specific peptide eliminates observed immunoreactivity to
SOD1 in the 32 kDa. Biotinylation and immunoblot were done as described.
Samples used are cytosolic proteins from muscle (SALS-2) and spinal cords
(SALS-7 and FALS-4). Peptides were incubated with the primary antibody for
30 min at room temperature in 500 ␮l of 5% milk in PBST before incubation
with the membranes. Antigen-specific peptide (DDLGKGGNEESTK) was used
at 1 ␮g and 4 ␮g, respectively. Nonspecific peptide (KESNGPVKVWGSIKGGC)
was used at 1 ␮g.
data demonstrate that the 32 kDa appears in the pathologic
process of age-related motor neuron degeneration such as that
in ALS but not involved in pathology of AD, PD, HD, or SMA.
The 32 kDa Is a Modified SOD1 Protein Species. To confirm that the
observed 32 kDa is an altered form of SOD1, peptide compe-
tition studies were performed. Antigen-specific peptide elimi-
nated reactivity to the 32-kDa (and the 16-kDa SOD1) signal in
a dose-dependent manner (Fig. 4). In contrast, a different
peptide from SOD1 had no such effect on the SOD1 immuno-
reactrivity when used at similar a concentration (Fig. 4). To
corroborate this conclusion, nanoelectrospray mass spectrome-
try was used to determine whether the 32 kDa in fact contained
SOD1 protein. The gel slice containing the 32 kDa was purified
as described (see SI Methods). Peptide sequence analysis based
on mass spectra acquired and peptide mapping (see SI Fig. 6 and
SI Table 2, respectively) provided 100% coverage of the human
SOD1 protein sequence. These results demonstrate that the 32
kDa does indeed contain the human SOD1 protein.
To further demonstrate that the SDS-resistant 32 kDa is a
modified SOD1 molecule in complex with itself or a similarly
sized protein (19), gel filtration and ion exchange chromatog-
raphy experiments were performed to characterize the pro-
tein(s) that gave rise to the 32-kDa species during biotinylation.
As expected, under nonreducing, nondenaturing conditions,
most SOD1 polypeptides in spinal cord cytosolic proteins were
present in dimers that eluted in gel filtration fractions corre-
sponding to ⬇32-kDa molecular mass, a range similar to the
elution position of commercially purified human SOD1 protein
(Fig. 5A Top and Middle). Biotinylation of gel filtration fractions
revealed that the 32-kDa species was observed exclusively in the
same molecular mass range (Fig. 5A Bottom), demonstrating the
substrate for biotinylation behaved as a 32-kDa protein species.
Although biotinylation was required to observe the 32-kDa
SOD1 species in spinal cord extracts, it was observable even
without biotinylation in muscle extracts from FALS or SALS
patients (Fig. 5B). We exploited this property to determine the
native characteristics of this 32-kDa SOD1 species without
biotinylation and to eliminate the possibility of noncovalent
dimeric SOD1 becoming covalently linked during biotinylation.
Cytosolic ALS protein extracts were fractionated before bioti-
nylation by Q-column ion exchange chromatography. The nor-
mally folded, non-covalently associated SOD1 dimers bound to
the matrix and eluted only at higher ionic strength (Fig. 5B,
eluted in fractions 5–7). A small amount of IR SOD1 did not bind
to the column. Subsequent immunoblotting after denaturation
under reducing conditions revealed that this unbound SOD1
exclusively migrated at a position corresponding to 32 kDa (Fig.
5 B, fractions 1–3; C Upper, fractions 3–6). As before, the
Gruzman et al.
Fig. 5. The 32 kDa is a modified SOD1 protein species. (A) The commercial
purified hSOD1 protein (Std, 40 ␮g) and SALS spinal cord cytosolic proteins
(SALS-4, 300 ␮g) were subjected to gel filtration without biotinylation (⫺Bi-
otin). Molecular mass standards were used during elution, and only the
fractions corresponding to ⬇32-kDa range (between ⬇45 kDa and 28 kDa, as
indicated) were shown in fractions 1–10. Fractions from the SALS-4 samples
were further subjected to biotinylation (⫹Biotin). (B) SALS muscle cytosolic
proteins (300 ␮g) were applied to the Q-type ion exchange column, and the
flow-through (fractions 1–3) and eluted fractions (4 –9) were analyzed. An
aliquot of the same sample (total input, 1/30) before the Q-Column is also
shown. (C) The Q-column flow-through material was subjected to gel filtra-
tions and analyzed directly (⫺Biotin) or after biotinylation (⫹Biotin). All
proteins were separated on SDS/PAGE and immunoanalyzed with the SOD1
antibody.
immunoreactivity of this 32- kDa species with the anti-peptide
SOD1 antibody was markedly enhanced after biotinylation (Fig.
5C Lower, fractions 2–7). These findings clearly demonstrate that
the 32-kDa protein species is an endogenously modified SOD1
that is covalently linked to another SOD1 polypeptide or to a
polypeptide of similar size and that this species has substantially
NEUROSCIENCE
divergent biochemical properties relative to the natural non-
covalently linked SOD1 homodimer.
Discussion
One mystery associated with not only ALS, but many other
neurodegenerative diseases, is how to reconcile disease mecha-
nisms at the molecular level between involvement of mutations
in a few molecules that are responsible for a small percentage of
familial forms of the disease and unknown molecules that
underlie the majority of the diseases that are of sporadic nature.
An assumption frequently made is that there is a common path
shared by both familial and sporadic forms of ALS that ulti-
mately leads to a common clinical manifestation of mixed lower
and upper motor neuron degeneration. Evidence supporting this
view has remained elusive until now. Our study demonstrates at
least one component of a shared pathway is an already known
molecule linked to ALS: SOD1.
Like other neurodegenerative diseases, ALS has been considered
a disease of protein misfolding (20). Demonstrations that mecha-
nisms of mutant SOD1-mediated toxicity are due to gain of toxic
properties (2) suggest that altered conformations of SOD1 proteins
because of mutation have different behaviors than the WT SOD1
(SOD1WT) protein. This is fully consistent with the more than 110
mutations in the SOD1 molecule linked to ALS (21) for which a
normally folded ‘‘WT’’ state and a misfolded one linked to toxicity
have frequently been found. Our observation of a covalently
crosslinked, biotinylation-dependent, SOD1-IR species (32 kDa)
within the spinal cord demonstrates that SOD1 exists in at least two
PNAS 兩 July 24, 2007 兩 vol. 104 兩 no. 30 兩 12527
forms, one of which is disease-associated and common to both
FALS and SALS. The increased level of the 32-kDa species in ALS
patients must originally come from the normally folded SOD1WT.
Our data demonstrate that an aberrant SOD1 species is present (or
associated with) human ALS pathology, suggesting a possible
involvement in that pathology. This conclusion extends previous
results from in vitro cellular and animal models of ALS indicating
that SOD1WT or its oxidized form can be toxic in ALS disease
mechanisms (22–28). Interestingly, the WT normal ␣-synuclein
proteins, when produced by an additional locus in human, can also
cause PD (29).
Our observation that the 32-kDa species is not present in
normal and three non-motor neuron diseases including AD, PD
and HD, nor in Type I SMA which involves motor neuron death
somewhat different from that in ALS, strongly suggests that the
32 kDa is ALS pathology-related. Compared with the vast
majority of SOD1 molecules that are almost certainly irrelevant
to disease, the low levels of this disease-specific SOD1-IR species
(⬇1% of total SOD1) are usually below detection limits, but can
be revealed after biotinylation. Thus, although chemical modi-
fication has been used to study overall differences in protein
conformation, the approach we have used here has uncovered a
molecular signature common to both familial and sporadic ALS.
SOD1 IR protein species of molecular weight approximately
the size of human SOD1 dimer are reported in refs. 30–32. More
recently, studies from transgenic mouse models of FALS have
suggested that the intermolecular disulfide-linked SOD1 aggre-
gates are associated with disease toxicity (8, 23). One study used
immunoprecipitation to conclude an apparent SOD1 species of
⬇35 kDa was a complex between SOD1 and Bcl-2 (19), albeit our
mass spectrometric analysis offered no support for any linkage
to Bcl-2, at least in SALS. The covalently crosslinked, partially
misfolded species we have identified here is different from these
previously proposed complexes. Its identification suggests a
common underlying pathogenic mechanism in inherited and
sporadic ALS, a finding that carries a substantial implication for
design of therapeutic approaches for sporadic disease.
Materials and Methods
Obtainment of Human Samples. Human autopsy samples were
obtained as described in ref. 24, from The Brain and Tissue Bank
for Developmental Disorders of the National Institute of Child
Health and Human Development (Baltimore, MD), Forbes
Norris MDA/ALS Research Center (San Francisco, CA), and
ALS Tissue Bank at University of Pittsburgh School of Medicine
(Pittsburgh, PA). Informed consents were obtained from all
subjects before sample collections. Handling of all human sam-
ples is in accordance with the approved protocols by California
Pacific Medical Center Internal Board Review, University of
Pittsburgh Internal Board Review, and Committee for Oversight
of Research Involving the Dead. Patients’ information is listed
in SI Table 3.
Subcellular Fractionation and Biotinylation. Cytosolic proteins from
the autopsy samples were prepared as described in ref. 33.
Protein concentrations were measured by using the BCA method
(Pierce Chemical, Rockford, IL). Biotinylation reaction, unless
indicated otherwise, was performed with a total of 10 ␮g of
cytosolic proteins incubated with 10 mM Sulfo-NHS-LC-Biotin
(Pierce Chemical) in PBS buffer, pH 7.4, for 25 min at 25°C. The
reaction was stopped by adding free lysine-HCl at a final
concentration of 20 mM for 20 min at 25°C. Control treatment
1. Rosen DR, Siddique T, Patterson D, Figlewicz DA, Sapp P, Hentati A,
Donaldson D, Goto J, O’Regan JP, Deng HX, et al. (1993) Nature 362:59–62.
2. Cleveland DW, Rothstein JD (2001) Nat Rev Neurosci 2:806–819.
3. Reaume AG, Elliott JL, Hoffman EK, Kowall NW, Ferrante RJ, Siwek DF, Wilcox
HM, Flood DG, Beal MF, Brown RH, Jr, et al. (1996) Nat Genet 13:43–47.
12528 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705044104
was carried out in identical procedure except omitting Sulfo-
NHS-LC-Biotin in the reaction.
Gel Filtration. Sephacryl 300 column (Amersham Biosciences, Pis-
cataway, NJ) was equilibrated with PBS buffer, pH 7.4, containing
150 mM sodium chloride. The protein sample was applied onto the
column at a ratio of 60 ␮g of cytosolic proteins per 1 ml of Sephacryl
300. Proteins were eluted under gravity in the same buffer at 4°C,
and a volume of 500 ␮l was collected for each fraction. All fractions
were dialyzed against deionizated water for 24 h at 4°C followed by
lyopholization. Lyopholized samples were resuspended in PBS for
biotinylation and/or immunoblot analysis.
Column Chromatography. Fast-flow Q Sepharose anion exchange
column was equilibrated with 20 mM triethanolamine buffer, pH
7.8. The protein sample was applied onto the column at a ratio
of 320 ␮g of cytosolic proteins per 1 ml of Q-Sepharose. Proteins
were eluted with the same buffer containing 0, 0.025, 0.05, 0.075,
0.1, 0.15, 0.2, 0.25, 0.35, and 0.4 M NaCl (corresponding to
fractions 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10, respectively) at a flow rate
of 0.5 ml/min. All fractions were dialyzed against deionized
water for 24 h followed by concentration in Speed Vac (Savant,
Holbrook, NY) at 25°C to a final volume of ⬇40 ␮l.
Immunoblot Analysis. Proteins were separated on 12% SDS/PAGE
gels and transferred onto PVDF membranes. Membranes were
incubated with a rabbit polyclonal anti-SOD1 antiserum (34)
followed by an HRP-conjugated anti-rabbit IgG secondary anti-
body before visualization with ECL plus solution (GE Biosciences,
Little Chalfont, U.K.). Protein signals were quantified on Typhoon
9410, using Image Quant program (GE Biosciences).
Analysis of Human SOD1 cDNA Sequence. Sequence analysis of
RT-PCR-generated SOD1 cDNAs from the spinal cords of five
individuals was performed. Total RNA was extracted by using
TRI-Reagent (Sigma–Aldrich, St. Louis, MO). Primers for human
SOD1 RT-PCR were synthesized based on GenBank cDNA com-
plete nucleotide sequence (GenBank accession no. AY049787)
as follows: forward 5⬘-ATGGCGACGAAGGCCGTGT-
GCGTGC-3⬘ and reverse 5⬘-TTATTGGGCGATCCCAATTA-
CACCACAAGCC-3⬘. The RNA was reverse-transcribed and
PCR-amplified by using the ImProm-II Reverse Transcription
System and Pfu DNA polymerase (Promega, Madison, WI). Syn-
thesis of SOD1 cDNA and PCR were carried in a PTC-100
programmable thermal controller. The PCR products were puri-
fied, the sequencing was performed on both strands, and the
sequence was determined at the DNA Analysis Unit of Integrated
DNA Technologies (Coralville, IA).
Statistical Analysis. Statistical significance was assessed between
groups, using a two-tailed Student’s t test. Significance is set at
P ⬍ 0.05.
We thank William J. Hansen, Yelena Mikityansky, William J. Santoni,
and David Ovadia for technical support and help with preparation of the
article and Jason Mass for coordinating tissue collections at The Forbes
Norris MDA/ALS Research Center. This work has been supported by
the California Pacific Medical Center Research Institute (V.R.L. and
J.L.), The Forbes Norris MDA/ALS Research Center (J.L.), the ALS
Association (J.L.), the Center for Excellence in Bioinformatics at
University at Buffalo (T.D.W.), and the National Institutes of Health
(R.B. and D.W.C.).
4. Bruijn LI, Miller TM, Cleveland DW (2004) Annu Rev Neurosci 27:723–749.
5. Valentine JS, Doucette PA, Potter SZ (2004) Annu Rev Biochem 74:563–593.
6. Borchelt DR, Lee MK, Slunt HS, Guarnieri M, Xu ZS, Wong PC, Brown RH,
Jr, Price DL, Sisodia SS, Cleveland DW (1994) Proc Natl Acad Sci USA
91:8292–8296.
Gruzman et al.
 7. Elam JS, Taylor AB, Strange R, Antonyuk S, Doucette PA, Rodriguez JA,
Hasnain SS, Hayward LJ, Valentine JS, Yeates TO, et al. (2003) Nat Struct Biol
10:461–467.
8. Furukawa Y, Fu R, Deng HX, Siddique T, O’Halloran T, V (2006) Proc Natl
Acad Sci USA 103:7148–7153.
9. Furukawa Y, O’Halloran TV (2005) J Biol Chem 280:17266–17274.
10. Lindberg MJ, Tibell L, Oliveberg M (2002) Proc Natl Acad Sci USA 99:16607–
16612.
11. Sato T, Nakanishi T, Yamamoto Y, Andersen PM, Ogawa Y, Fukada K, Zhou
Z, Aoike F, Sugai F, Nagano S, et al. (2005) Neurology 65:1954–1957.
12. Sugo T, Watanabe K, Naraki T, Matsuda M (1990) J Biochem (Tokyo)
108:382–387.
13. Lundblad RL, Bradshaw RA (1997) Biotechnol Appl Biochem 26(Pt 3):143–
151.
14. Chen YZ, Bennett CL, Huynh HM, Blair IP, Puls I, Irobi J, Dierick I, Abel A,
Kennerson ML, Rabin BA, et al. (2004) Am J Hum Genet 74:1128–1135.
15. Battistini S, Giannini F, Greco G, Bibbo G, Ferrera L, Marini V, Causarano
R, Casula M, Lando G, Patrosso MC, et al. (2005) J Neurol 252:782–788.
16. Corrado L, D’Alfonso S, Bergamaschi L, Testa L, Leone M, Nasuelli N,
Momigliano-Richiardi P, Mazzini L (2006) Neuromuscul Disord 16:800–804.
17. Garcia-Redondo A, Bustos F, Juan YSB, Del Hoyo P, Jimenez S, Campos Y,
Martin MA, Rubio JC, Canadillas F, Arenas J, et al. (2002) Muscle Nerve
26:274–278.
18. Gellera C, Castellotti B, Riggio MC, Silani V, Morandi L, Testa D, Casali C,
Taroni F, Di Donato S, Zeviani M, et al. (2001) Neuromuscul Disord 11:404–
410.
19. Pasinelli P, Belford ME, Lennon N, Bacskai BJ, Hyman BT, Trotti D, Brown
RH, Jr (2004) Neuron 43:19–30.
Gruzman et al.
20. Chiti F, Dobson CM (2006) Annu Rev Biochem 75:333–366.
21. Gros-Louis F, Gaspar C, Rouleau GA (2006) Biochim Biophys Acta 1762:956–
972.
22. Estevez AG, Crow JP, Sampson JB, Reiter C, Zhuang Y, Richardson GJ,
Tarpey MM, Barbeito L, Beckman JS (1999) Science 286:2498–2500.
23. Deng HX, Shi Y, Furukawa Y, Zhai H, Fu R, Liu E, Gorrie GH, Khan MS,
Hung WY, Bigio EH, et al. (2006) Proc Natl Acad Sci USA 103:7142–7147.
24. Ezzi SA, Urushitani M, Julien JP (2007) J Neurochem 102:170–178.
25. Jaarsma D, Haasdijk ED, Grashorn JA, Hawkins R, van Duijn W, Verspaget
HW, London J, Holstege JC (2000) Neurobiol Dis 7:623–643.
26. Jonsson PA, Graffmo KS, Brannstrom T, Nilsson P, Andersen PM, Marklund
SL (2006) J Neuropathol Exp Neurol 65:1126–1136.
27. Rakhit R, Crow JP, Lepock JR, Kondejewski LH, Cashman NR, Chakrabartty
A (2004) J Biol Chem 279:15499–15504.
28. Rakhit R, Cunningham P, Furtos-Matei A, Dahan S, Qi XF, Crow JP, Cashman
NR, Kondejewski LH, Chakrabartty A (2002) J Biol Chem 277:47551–47556.
29. Singleton AB, Farrer M, Johnson J, Singleton A, Hague S, Kachergus J,
Hulihan M, Peuralinna T, Dutra A, Nussbaum R, et al. (2003) Science 302:841.
30. Johnston JA, Dalton MJ, Gurney ME, Kopito RR (2000) Proc Natl Acad Sci
USA 97:12571–12576.
31. Jonsson PA, Ernhill K, Andersen PM, Bergemalm D, Brannstrom T, Gredal
O, Nilsson P, Marklund SL (2004) Brain 127:73–88.
32. Wang J, Xu G, Borchelt DR (2002) Neurobiol Dis 9:139–148.
33. Liu J, Lillo C, Jonsson PA, Vande Velde C, Ward CM, Miller TM, Subrama-
niam JR, Rothstein JD, Marklund S, Andersen PM, et al. (2004) Neuron
43:5–17.
34. Pardo CA, Xu Z, Borchelt DR, Price DL, Sisodia SS, Cleveland DW (1995)
Proc Natl Acad Sci USA 92:954–958.
NEUROSCIENCE
PNAS 兩 July 24, 2007 兩 vol. 104 兩 no. 30 兩 12529