# 2002 Oxford University Press
Human Molecular Genetics, 2002, Vol. 11, No. 11
1263–1271
Genotype–phenotype correlations for EPM2A
mutations in Lafora’s progressive myoclonus
epilepsy: exon 1 mutations associate with an
early-onset cognitive deficit subphenotype
Subramaniam Ganesh1, Antonio V. Delgado-Escueta2,*, Toshimitsu Suzuki1, Silvana
Francheschetti3, Concetta Riggio3, Giuiliano Avanzini3, Adrian Rabinowicz4, Saeed Bohlega5,
Julia Bailey2, Maria E. Alonso6, Astrid Rasmussen6, Alfredo E. Thomson4, Adriana Ochoa6,
Aurelio J. Prado6, Marco T. Medina7 and Kazuhiro Yamakawa1
1
Laboratory for Neurogenetics, RIKEN Brain Science Institute, Wako-shi, Japan, 2Epilepsy Genetics/Genomics
Laboratories, Comprehensive Epilepsy Program, UCLA School of Medicine and VA GLAHS West Los Angeles
Medical Center, Los Angeles, CA, USA, 3Instituto Nazionale Neurologico, Besta, Milano, Italy, 4FLENI Medical Center,
Buenos Aires, Argentina, 5King Faisal Medical Center, Riyadh, Saudi Arabia, 6National Institute of Neurology and
Neurosurgery, Mexico City, Mexico and 7Direccion de Investigation Cientifica, Universidad National Autonoma de
Honguras, Tegulcigalpa, Honduras
Received February 4, 2002; Revised and Accepted March 17, 2002
Mutations in the EPM2A gene encoding a dual-specificity phosphatase (laforin) cause an autosomal
recessive fatal disorder called Lafora’s disease (LD) classically described as an adolescent-onset stimulussensitive myoclonus, epilepsy and neurologic deterioration. Here we related mutations in EPM2A with
phenotypes of 22 patients (14 families) and identified two subsyndromes: (i) classical LD with adolescentonset stimulus-sensitive grand mal, absence and myoclonic seizures followed by dementia and neurologic
deterioration, and associated mainly with mutations in exon 4 (P ¼ 0.0007); (ii) atypical LD with childhoodonset dyslexia and learning disorder followed by epilepsy and neurologic deterioration, and associated
mainly with mutations in exon 1 (P ¼ 0.0015). To understand the two subsyndromes better, we investigated
the effect of five missense mutations in the carbohydrate-binding domain (CBD-4; coded by exon 1) and three
missense mutations in the dual phosphatase domain (DSPD; coded by exons 3 and 4) on laforin’s
intracellular localization in HeLa cells. Expression of three mutant proteins (T194I, G279S and Y294N) in
DSPD formed ubiquitin-positive cytoplasmic aggregates, suggesting that they were folding mutants set for
degradation. In contrast, none of the three CBD-4 mutants showed cytoplasmic clumping. However, CBD-4
mutants W32G and R108C targeted both cytoplasm and nucleus, suggesting that laforin had diminished its
usual affinity for polysomes. Our data, thus, represent the first report of a novel childhood syndrome for LD.
Our results also provide clues for distinct roles for the CBD-4 and DSP domains of laforin in the etiology of
two subsyndromes of LD.
INTRODUCTION
Lafora’s progressive myoclonus epilepsy or Lafora’s disease
(LD; OMIM 254780) is an autosomal recessive, fatal disorder
with pathognomonic periodic acid–Schiff-positive (PASþ )
staining intracellular inclusion bodies (1). Besides the central
nervous system, the large basophilic PASþ inclusions, called
Lafora bodies, can also be found in the retina, heart, liver,
muscle and skin (2–4). In the medical literature, LD is
classically described as starting in early adolescence as
stimulus-sensitive grand mal tonic–clonic, absence, visual
and myoclonic seizures. Rapidly progressive dementia, psychosis, cerebellar ataxia, dysarthria, amaurosis, mutism, muscle
wasting and respiratory failure leads to death between 17 and
24 years of age (2,5,6), although improved nursing care has
commonly extended life now to 28 years. LD, although
*To whom correspondence should be addressed at: Epilepsy Genetics=Genomics Laboratories, Comprehensive Epilepsy Program, VA GLAHC Medical
Center, 11301 Wilshire Blvd., Los Angeles, CA 90073, USA, Tel: þ 1 310 268 3129; Fax: þ 1 310 268 4937; Email: [email protected]
1264
Human Molecular Genetics, 2002, Vol. 11, No. 11
relatively rare in the outbred populations of the USA and
Canada, is commonly encountered in the Mediterranean basin
of Spain, France and Italy, in restricted regions of Central Asia,
India, Pakistan, Northern Africa and the Middle East, and in
ethnic isolates from the southern USA and Quebec, Canada
(6–8). As with many other recessive disorders, consanguinity is
common in LD families.
In 1995, our laboratories localized a gene for LD to a 17 cM
span of chromosome 6q23–25, and, in 1997 and 1998, reduced
the region to 3 cM and then 300 kb (8–10). Subsequently,
Minassian et al. (10) and Serratosa et al. (11) independently
identified the LD gene EPM2A and showed that LD patients
were homozygous or compound heterozygous for presumably
loss-of-function mutations. To date, a total of 32 different
mutations in EPM2A have been described, including 12
missense, 4 nonsense, 5 frameshift, 1 insertion and 10 deletion
mutations of different sizes (10–14). The EPM2A gene is
expressed ubiquitously in humans, and its encoded protein,
laforin, is an active dual-specificity phosphatase, associated
with polyribosomes (15). Laforin also contains an N-terminal
carbohydrate-binding domain (CBD-4) (13,16) that targets
laforin to glycogen (17). Expression of the Epm2a gene in mice
brains is developmentally regulated and is predominantly in the
cerebellum, hippocampus, cerebral cortex and the olfactory
bulb (16). These results suggest that laforin is involved in
translational regulation and may play critical roles in the
growth and maturation of neural networks (6,15,16). In this
paper, we identify additional families with EPM2A mutations
and related the clinical manifestations of LD patients with
mutations. We also evaluate the effects of missense mutations
by intracellular targeting of laforin mutants in HeLa cells.
RESULTS
EPM2A mutation screening
We searched for sequence variation in the coding region of
EPM2A in individuals from 19 unrelated LD families and 74
control individuals of diverse ethnic background (Caucasian,
Table 1. New LD families with EPM2A mutations
Patient/
family
Origin
Consanguinity?
Affecteda
(tested/mutated)
Mutation
Genotype in affected
Predicted effect
LD44
LDN4
LDN5
LD40
Mexico
Spain
Spain
Argentina
No
Yes
Yes
No
1
3
1
2
LD2
LD12
LD35
USA
Spain
Italy
No
Yes
No
1 (1/1)
3 (1/1)
2 (2/2)
73T!C
258C!G
322C!T
364–365insGT and
721C!T
512G!A
721C!T
721C!T
Homozygous
Homozygous
Homozygous
Compound
heterozygous
Heterozygous
Homozygous
Homozygous
S25P (missense)
Y86X (nonsense)
R108C (missense)
Insertion (frameshift) and
R241X (nonsense)
R171H (missense)
R241X (nonsense)
R241X (nonsense)
(1/1)
(3/3)
(1/1)
(2/2)
a
Denotes the total number of affecteds in each family, and the number of affecteds screened for mutation (tested) and those who had mutations (mutated),
respectively.
Figure 1. Haplotype analysis for 6 LD families (LD12, LD13, LD15, LD16, LD35 and LD40) with the R241X mutation. Open square, unaffected male; open
circle, unaffected female; filled symbol, affected individuals. The four-microsatellite loci are arranged in physical order from centromere (top) to telomere (bottom).
Numbers indicate the genotypes for each of the above markers in the same order. The rectangular box denotes the disease haplotype that is common among affected
individuals [ þ , normal allele; m, R241X mutation allele; i, 364–365insGT mutation allele (family LD40)].
Human Molecular Genetics, 2002, Vol. 11, No. 11
Indian and Japanese). Mutation screening identified six different
mutations in seven families that are likely to disrupt the function
of laforin (Table 1). The S25P missense mutation identified in
family LD44 is a novel mutation that was not detected in 148
normal control chromosomes. LD40 is a non-consanguineous
family, and both affected children were compound heterozygotes for the disease mutations, 364–365insGT and R241X. A
sequence analysis confirmed that the two alleles were inherited
from the mother and father, respectively (Fig. 1). Whereas 364–
365insGT is a novel allele, the nonsense mutation R241X is
common among LD patients of Spanish origin (10,12,14). The
R241X mutation identified in three new families of our present
study (LD40, LD12 and LD35; Table 1) raise the number of LD
families that have the R241X mutation to 25. Gomez-Garre et al.
1265
(14) suggested that the high prevalence of the R241X mutation
in the Spanish population might be due to both a founder effect
and recurrent events. We therefore analyzed haplotypes around
the EPM2A gene in members of six families with the R241X
mutation, in order to determine whether families with this
mutation share a common ancestral haplotype or whether they
represent independent mutation events. The microsatellite
markers used, and their order on chromosome 6q24, are
centromere–D6S1010–D6S1703–D6S1042–D6S1649–telomere. D6S1703 is located in the first intron of the EPM2A gene
and the four markers span about 500 kb on 6q24 (10). As shown
in Figure 1, the disease haplotype (2–3–2–4) was found to be
associated with the mutated alleles in all six families, including
one family each from Italy and Argentina.
Table 2. Genotype–phenotype correlation: EPM2A mutations, clinical manifestations of Lafora disease and age during appearance of signs and symptoms
Family/
individuala
Mutation
(mutated exon)
Age (years) at seizure onsetb
(GM or MS or Abs or visual)
Age (years) at onset
of cognitive decline
Age (years) at onset of neurological deterioration
Ataxia and
Dementia and
Respiratory assist
spasticity
mutism
and gastrostomy
11
11
11
17
18; still alive
12
14 (MS), 19 (GM)
14
17
16
17
18
20
18; still alive
24; deceased
13
14
16
17
16
13
14
16
17
16
NI
NI
NI
NI
NI
NI; wheelchair-bound
NI
NI
NI
NI
15
15
15
14
16
15
15
16
20
20
20
Dementia
without mutism
22; deceased
20; deceased
20; deceased
Still alive at 20 with
normal respiration and
swallowing functions
Still alive at 18 with
normal respiration and
swallowing functions
I. Classical/typical LD
LD12-1
LD12-2
LD13 (10)
LD15 (10)
LD16 (10)
25 kb
deletion (2)
R171H (3)
60 kb
deletion (2)
R241X (4)
R241X (4)
R241X (4)
R241X (4)
R241X (4)
LD33-3 (10)
LD33-5 (10)
LD33-6 (10)
LD35-1
Q293L (4)
Q293L (4)
Q293L (4)
R241X (4)
10 (MS), 11 (GM)
11
12
10
11 (headaches), 13 (GM),
16 (Abs status)
13
11
11
13
LD35-2
R241X (4)
13
14
16
Dementia without
mutism
LDN4-3
LDN4-4
LDN4-5
LDN5-3
Y86X (1)
Y86X (1)
Y86X (1)
R108C (2)
5
4
4
4
(dyslexia, dysgraphia)
17
16
17
13
17
17
17
11
NI
NI
NI
NI
LD20 (13)
LD40-1
Q55X (1)
364–365insGT
þ R241X
(2 and 4)
364–365insGT
þ R241X
(2 and 4)
S25P (1)
5 (GM isolated) then 11
14 (Visual), 17 (Abs, GM)
4 (Abs), 11 (GM)
5 and 8 (isolated Abs,
MS and GM), 10 (MS
and convulsive status)
13
8 (isolated Abs) then 15
5
8
15
14
16
16
15; still alive
15 (deceased)
12
8
14
16
13
6
17
16
13
8
16
18
Still alive at 16 with
normal respiration and
swallowing functions
Still alive at 18 with
normal respiration and
swallowing functions
18 (deceased)
LD1-3 (12)
LD2-3
LD9-3
II. Atypical LD
LD40-2
LD44
LD9-16
a
80 kb deletion
(1 and 2)
Numbers in parentheses indicate the reference that originally described EPM2A mutations in these families.
GM, grand mal clonic tonic clonic or tonic clonic; MS, myoclonic seizure; Abs, absence. Unless indicated grand mal, myoclonias and absence appeared at the
same time.
NI, no information on present or final years of illness were obtained in these patients.
b
1266
Human Molecular Genetics, 2002, Vol. 11, No. 11
The significance of the missense mutation, R171H, found
heterozygously in the proband of family LD2 is presently
not known. Because the EPM2A exons sequenced were
PCR-amplified, hemizygous deletion for other exon(s) could
not be ruled out. Mutation R171H had been described previously in LD families (10,11). To make a genotype–phenotype
comparison (see below), we also confirmed the EPM2A
mutation status in seven other families (LD1, LD9, LD13,
LD15, LD16, LD20 and LD33) that had been reported
Figure 2. Map showing deletion breakpoints in the family LD9 as refined by
PCR analysis. LD9 is a large consanguineous Palestine family with four affecteds, of whom three are diseased (8). Mutation analysis was restricted to patient
LD9-3 and his diseased first cousin, LD9-16. Open square, unaffected male;
open circle, unaffected female; filled symbol, affected individual. The genomic
organization of exons 1 and 2 (open boxes) of EPM2A and the positions of STS
markers are indicated to scale. Primers were designed based on the genomic
sequence of the EPM2A locus (GenBank accession no. AL023806). Arrows
on the left define the maximum extent of homozygous deletion in the patients.
Control refers to PCR reaction without a template.
Figure 3. Schematic diagram showing domain organization of laforin and the
positions of various mutations found in the 14 LD families (top). The figure
also shows (bottom) the positions of 14 missense mutations known in laforin.
Missense mutations that have been investigated in the present study are denoted
by an asterisk and those that were studied earlier (15) by two asterisks. NH3, N
terminal; COOH, C terminal.
Human Molecular Genetics, 2002, Vol. 11, No. 11
previously (10,12,13) (Table 2). Amongst these, LD9 is a large
consanguineous Palestinian family with four affected members
(8), one of whom (LD9-16) was shown to have an approximately 80 kb homozygous deletion spanning exon 2 of the
EPM2A gene (12). We extended the deletion analysis to one
more affected member (LD9-3) who had a novel allele (Fig. 2).
Whereas LD9-16 had exons 1 and 2 deleted, homozygous
deletion ( 60 kb) in LD9-3 was limited to exon 2 (Fig. 2).
These results suggest that at least two different null alleles
segregate in this large consanguineous family. Figure 3 shows
the positions of 11 distinct mutations found in the 14 LD
families covered in the present study.
No mutation was found in the coding region of four exons of
the EMP2A gene in five LD families (LD10, LD36, LD38, LD39
and LD41) that had 6q24 haplotypes and homozygosities.
Genotype–phenotype correlations
We reviewed in detail the clinical manifestations of LD in 22
patients belonging to 14 families, and related their disease
phenotypes to EPM2A mutations found in their families. We
observed two subsyndromes (Table 2). The first subsyndrome,
which conforms to the description of typical or classic LD in
the literature, started with myoclonic seizures during adolescence or late childhood and was observed in 13 patients of
eight families. Of these 13 patients, 10 have missense
mutations in exon 4. One patient has a missense mutation in
exon 3 while two patients have microdeletions spanning exon
2. The second subsyndrome is atypical in its childhood-onset
educational difficulties. We observed such features in nine
patients of six families. Four patients have nonsense mutations
in exon 1, one patient has a deletion that occupies exons 1 and
2, two patients had missense mutations in exon 1 and 2
respectively, and two brothers were compound heterozygous
for mutations in exons 2 and 4.
There was significant difference between classic LD and
atypical LD for association with exon 4 as measured by the
Fisher exact test (P-value ¼ 0.0007) favoring typical LD and
exon 4 association. There was also significant difference
between typical LD and atypical LD for association with exon 1
(Fisher exact P-value ¼ 0.0015) favoring atypical LD and exon
1 association.
1267
Classic LD starts as epilepsy during early adolescence in an
otherwise neurologically normal individual (Table 2). Of 13
patients, 5 started epilepsy at 11 years of age. Eight others
started seizures at 10, 12, 13, and 14 years of age. Grand mal
clonic tonic clonic, absence, and myoclonic seizures were
present in all patients. Sound, light, or touch triggered bilateral
or segmental or facial myoclonias. Visual seizures were present
in 3 of 12 patients. Cognitive decline as evidenced by poor
school performance, poor memory and calculating abilities,
repetition of school years, and a drop in intelligence quotient
started at 11–17 years. Frequent bouts of status epilepticus,
corticospinal tract signs, slurred speech and ataxia disabled the
patient at 11–17 years. Patients became bedridden, demented
and mute between 17 and 20 years, and tracheostomy,
endotracheal intubation and gastrostomies were required by
18–22 years.
Atypical LD, in contrast, starts with childhood learning
problems and educational difficulties. Two patients were
diagnosed with dyslexia and dysgraphia. Three patients had
low intelligence quotient on formal psychometrics. All nine
patients had to repeat several school years and were enrolled in
special education (Table 2). Isolated seizures of absence or
grand mal appeared rarely at 4 and 5 years in three patients and
at 8 years in one patient. Frequent myoclonic, grand mal tonic
clonic and absences started at 8–13 years, rapidly becoming
daily and producing many bouts of convulsive and absence
status. Of eight patients, five had visual seizures. Dementia and
mutism were present at 16–18 years, but were observed
together with respiratory difficulties and swallowing problems
at 15–18 years. Hence, atypical LD starts earlier in childhood
as educational difficulties; in a few patients, isolated and rare
seizures also appear during childhood. More frequent and
enduring seizures follow in adolescence, and visual seizures
and hallucinations appear more often when compared with
typical LD. It is difficult to be sure if this subsyndrome
progresses more rapidly than typical LD, requiring tracheostomies and gastrostomies at an earlier age, because nursing and
chronic care for neurologically disabled patients vary from
country to country. Because of the atypical presentation of
these patients, LD was not suspected in four patients, and a
skin, muscle or liver biopsy was not requested until
adolescence when grand mal, absence, and myoclonic status
caused frequent hospitalizations.
Table 3. Effect of EPM2A missense mutation on the subcellular localization of laforin
Missense mutation
a
S25P
E28L
F88L
W32G
R108C
R171Hb
T194I
G279S
Q293Lb
Y294N
a
Mutated exon
Mutated domain
Subcellular localization
1
1
1
1
2
3
3
4
4
4
CBD-4
CBD-4
CBD-4
CBD-4
CBD-4
DSPD
DSPD
DSPD
DSPD
DSPD
No protein product detected
No difference when compared with wild-type laforin
No difference when compared with wild-type laforin
Cytoplasmic and nuclear localization; fractionate into soluble fraction
Cytoplasmic and nuclear localization; fractionate into soluble fraction
Form cytoplasmic clumps; larger ones are ubiquitin-positive
Form cytoplasmic clumps; larger ones are ubiquitin-positive
Form cytoplasmic clumps; larger ones are ubiquitin-positive
Form cytoplasmic clumps; larger ones are ubiquitin-positive
Form cytoplasmic clumps; larger ones are ubiquitin-positive
Several mutated clones were transfected and tested for expression by confocal microscopy and western blotting, but protein product could not be detected.
Reported in our previous study (15), but listed here for comparison.
b
1268
Human Molecular Genetics, 2002, Vol. 11, No. 11
Figure 4. Representative figures showing the effect of missense mutations on the intracellular targeting of laforin in HeLa cells. Cells expressing wild-type or
mutant laforin–EGFP chimeric fluorescent protein (Laforin; A–E) were double-stained with anti-QM (Ribosome; A0 –D0 ) or anti-ubiquitin antibodies (Ubiquitin;
E0 ) as indicated. A list of missense mutations created and their effect on the intracellular targeting of laforin are given in Table 3. Bar ¼ 10 mm).
Human Molecular Genetics, 2002, Vol. 11, No. 11
1269
G279S, or Y294N) formed punctate cytoplasmic aggregates
adjacent to the nucleus and were ubiquitinated (Fig. 4 and
Table 3). We had previously observed that two other missense
mutations (R171H and Q293L) in DSPD also led to the
formation of such cytoplasmic aggregates (15) (Table 3).
In contrast, none of the five CBD-4 mutants showed any
evidence of cytoplasmic clumping (Fig. 4 and Table 3). The
cellular localization of the CBD-4 mutants E28L and F88L had
not significantly changed and was identical to that of wild-type
laforin. Both CBD-4 mutants showed diffuse staining restricted
to cytoplasm. However, for the constructs W32G and R108C, a
significant quantity of the mutant protein was also targeted to the
nucleus (Fig. 4 and Table 3). For further confirmation of this
finding, we fractionated transiently transfected HeLa cells into
subcellular compartments by differential centrifugation (Fig. 5).
We had demonstrated earlier that laforin is enriched in the light
membrane fraction containing endoplasmic reticulum (ER) and
ribosomes, and associates with ploysome, possibly through
protein–protein interaction (15). The laforin mutants W32G and
R108C were found, as in wild-type laforin, in the microsomal
fraction (Fig. 5). However, in contrast with cells expressing
wild-type laforin, almost equal amount of mutant proteins also
fractionated into nuclear and soluble fractions (Fig. 5).
Figure 5. Western blot analysis of subcellular fractions derived from HeLa cells
expressing wild-type (WT) or mutant (E28L, W38G or R108C) laforin–EGFP
chimeric proteins. Differential centrifugation was performed to obtain nuclear
(N), heavy membrane (H), light membrane (L) and soluble (S) fractions, which
were western-blotted and probed with anti-GFP antibody. Note the difference in
the fractionation properties between wild-type laforin and the missense mutants
W32G and R108C. An almost-identical result was obtained when the N-terminal anti-laforin antibody (15) was used, confirming that they were full-length
fusion proteins (data not shown). A control probing with anti-QM protein antibody (a polyribosomal marker) is shown at the bottom (QM).
Of interest are the patients with mutations in exon 2. Three
patients (LDN5-3, LD40-2 and LD40-1) have mutations in
exon 2 and have manifestations of atypical LD. LDN5-3 started
illness with dyslexia and dysgraphia at 5 years, while two
brothers (LD40-1 and LD40-2) needed special education at
8 years. Both epilepsy (myoclonic, atonic, grand mal and
absence) and learning disorders (dyslexia, dysgraphia and poor
school performance) appeared at onset. Two other patients
(LD9-3 and LD1-3) have 60 and 25 kb deletions in exon 2 and
have clinical features of typical LD, with epilepsy starting at 14
and 11 years.
Subcellular targeting of laforin mutants
Each of 14 distinct missense mutations found so far in LD
affects a residue either in the dual-specificity phosphatase
domain (DSPD) or in the carbohydrate-binding domain (CBD4) that targets laforin to glycogen (17) (Fig. 3). Moreover, each
of the 14 missense mutations affects a residue that is conserved
between laforin orthologs (16). To understand the effects of
missense mutations on the subcellular localization of laforin,
we transfected HeLa cells with constructs expressing enhanced
green fluorescent protein (EGFP)-tagged laforin containing
each of the eight missense mutations created (five in CBD-4
and three in DSPD) (Fig. 3 and Table 3). Intriguingly, expression of all three constructs that had mutation in DSPD (T194I,
DISCUSSION
The results of our mutation screening as well as those reported
elsewhere (10–14) are consistent with the theory that EPM2A
mutations are inactivating and that LD occurs through a loss of
laforin function. The identification of two novel mutations
expands the spectrum of mutations known in the EPM2A gene
and emphasizes that they display marked allelic heterogeneity.
Of the total number of mutations so far detected, the common
mutation R241X appears to be more prevalent in the Spanish
population (present study, 10,14). Haplotype construction
showed that all six families tested in the present study shared
a common haplotype, suggesting a founder gene effect.
Interestingly, families LD35 (Italian) and LD40 (Argentinean)
also had the R241X mutation common to Spanish patients.
In transfection studies addressing the impact of missense
mutations, it was clear that all five mutations in the DSPD
resulted in the formation of cytoplasmic clumps. The majority
contained few large perinuclear cytoplasmic clumps that were
ubiquitin-positive. In some cells, numerous small fluorescent
specks were visible throughout the cytoplasm. Interestingly,
cytoplasmic clumps of each of these mutations did not show an
obvious effect on the distribution of ribosome or ER networks,
and these cells appeared to be otherwise normal (Fig. 4) (15),
suggesting that such large cytoplasmic clumps may not
negatively influence translational machinery. Missense mutations that result in cytoplasmic clumping of defective proteins
have also been identified in a few other recessive disorders
(18,19). Folding mutants of the cystic fibrosis transmembrane
conductance regulator (CFTR) protein result in the deposition
of a pericentriolar cytoplasmic inclusion that contains ubiquinated mutant protein (20). These inclusions, termed aggresomes, are thought to represent a general cellular response to
cytoplasmic accumulation of misfolded protein (20,21). It was
further suggested that the formation of aggresomes may not
1270
Human Molecular Genetics, 2002, Vol. 11, No. 11
inhibit vital cellular functions and that the pathology may
instead relate to unavailability of the aggregated protein (21).
Our findings that each of the five DSPD mutants, affects the
folding properties of laforin, leading to the formation of
enzymatically inactive ubiquitinated cytoplasmic clumps targeted for degradation, are consistent with the above reports. In
contrast to the DSPD mutants however, none of the five CBD-4
mutants showed any evidence of cytoplasmic clumping, but
two of them, W32G and R108C, targeted to the nucleus and
partitioned into the soluble fraction. We interpret these results
to mean that these mutations diminish the affinity of laforin for
polysomes, leading to the partition of mutant proteins into
nuclear and soluble fractions. While this interpretation warrants
further studies, especially on endogenous laforin in patients
carrying these missense mutations, our present set of results
indicates variable consequences of missense mutations on the
subcellular localization of laforin. Interestingly, a recent study
demonstrated that the mutation W32G might also affect the
laforin’s glycogen-binding property of laforin (17).
LD has long been considered as a clinically homogeneous
disorder that starts as epilepsy during adolescence in an
otherwise neurologically normal individual (2,5,6). A detailed
evaluation of the clinical histories of 22 patients recruited in
our study identified a novel subsyndrome not known previously
in the medical literature. This atypical LD is characterized by
childhood-onset educational difficulties. Grand mal, absence
and myoclonic seizures are present in both subsyndromes, but
visual seizure and hallucination occur more often in atypical
LD. Thus, atypical LD appears to be the more severe form,
with an earlier age at onset.
The EPM2A gene is composed of four exons that encode a
331-amino-acid laforin (15). Whereas CBD-4 is coded mainly
by exon 1, DSPD spans exons 3 and 4 (Fig. 3). Our observation
that classical LD associates mainly with mutations in exon 4 and
atypical LD mainly with mutations in exon 1 raises the
possibility of different functions for CBD-4 and DSPD, and
distinctive roles for these domains in the etiology of the two
subsyndromes. We therefore predicted the effect of each
mutation on the functions of the two domains of laforin and
compared it with the two clinical subtypes. It is highly likely that
four out of seven mutations found in atypical LD patients affect
the expression of laforin itself: the 80 kb deletion in LD9-16
removes the first exon critical for EPM2A expression, and the
transcripts harboring a premature termination codon (PTC)
(Q25X, Y86X and 364–365insGT) may be unstable as a result
of nonsense-mediated RNA decay (NMD) (22,23). Conversely,
the nonsense mutation R241X, localized in the fourth exon of
the EPM2A gene, predicted to produce a truncated laforin
lacking DSPD because of a PTC present in the last exon, may
escape NMD (22,23). Thus, for atypical LD, CBD-4 was
affected in seven out of nine patients; both CBD-4 and DSPD are
affected in five (families LDN4, LD20, LD9) and CBD-4 in two
(LDN5, LD44). In two patients (family LD40) the truncated
laforin may contain only CBD-4 (Table 2). In contrast, in 11 out
of 13 patients with classical LD, only DSPD was affected
(families LD2, LD12, LD13, LD15, LD16, LD33 and LD35).
This raises the interesting possibility that mutations affecting the
functions of DSPD, and not of CBD-4, lead to classical LD and
those that affect both domains might develop atypical LD. This
suggestion, however, does not explain why patients LD1-3 and
LD9-3 show typical LD, since homozygous deletion of exon 2 in
both patients is likely to result in NMD and no protein product.
Likewise, the results of our transfection studies for the DSPD
mutations R171H and Q293L contrast with the above suggestion: patients LD2 (mutation R171H) and LD33 (Q293L) of our
study group exhibited classical LD even though full-length
laforin (including CBD-4) remains inactive in the aggresome.
However, it should be noted that our results have relied on the
use of a transient overexpression cell culture system, and
therefore we do not know whether the mutant proteins are
completely inactive even though the majority have aggregated
and whether endogenous laforin having these mutations would
show similar properties. It is also of interest to note that families
bearing identical mutations show variable phenotypes. In both
LDN4 and LD40, only one affected in the respective families
(LDN4-3 and LD40-1) started cognitive decline and epilepsy in
childhood (Table 2). Thus, there appears to be additional genetic
factor(s) that modulate phenotypic variability. Taken together,
our data represent the first report of a novel childhood syndrome
for LD and also provide clues to explore distinctive roles for
CBD-4 and DSPD in the causation of two subsyndromes of LD.
The tentative correlations observed between mutation sites and
phenotypic subgroups need to be studied prospectively and
tested further in other LD families.
It is remarkable that we were unable to detect mutations in
the entire coding region and the relevant intron boundaries of
the EPM2A gene in five families that had 6q24 haplotypes and
homozygosities. Our study, however, did not exclude the
possibility that defects in the upstream (and uncharacterized)
regulatory region of the EPM2A gene or in critical enhancer
elements might produce the LD phenotype. Future efforts,
therefore, must be directed towards characterizing regulatory
elements of EPM2A and finding mutations in those regions.
MATERIALS AND METHODS
Patients, DNA specimens and mutation detection
We studied patients whose clinical, electroencephalographic and
biopsy data proved the diagnosis of LD. Our study was approved
by the Human Subjects Protection Committees at the UCLA
School of Medicine and the West Los Angeles Department of
Veterans Affairs Medical Center, and by the Institutional Review
Board at the RIKEN Brain Science Institute. Genomic DNA was
extracted from blood samples using a QIAamp blood DNA
purification kit (Qiagen, Inc., Valencia, CA), individual exons
were PCR-amplified using established primers (10,13,24),
directly sequenced using BigDye terminators (Applied
BioSystems, Foster City, CA) and detected with an ABI model
377 DNA sequencer. Haplotype analysis was performed with the
following fluorescently labeled markers from the 6q24 region
spanning the EPM2A locus: D6S1010, D6S1703, D6S1042 and
D6S1649. Amplified products were run on an ABI 377
automated sequencer using GENSCAN software and analyzed
using GENOTYPER software (Applied Biosystems) as
described elsewhere (24).
Human Molecular Genetics, 2002, Vol. 11, No. 11
Expression constructs, transfection and western blot
analyses
The expression construct pEGFP–LDH encoding the laforin–
EGFP chimeric protein (15) was used for the transfection
studies. Point mutations in the EPM2A coding region was
generated by using the Stratagene QuikChange site-directed
mutagenesis kit according to the manufacturer’s instructions
(Stratagene Inc., La Jolla, CA). Briefly, complementary primers
containing the desired single base change were used in the PCR
amplification of expression constructs. Following digestion with
DpnI, the PCR products were used to transform Escherichia coli
XL1-Blue cells (Stratagene), independent clones were isolated
and mutations were confirmed by sequencing. Transfection of
constructs expressing mutant proteins in HeLa cells were
performed in parallel, and subcellular localization was evaluated
as described elsewhere (15). In brief, HeLa cells were
transiently transfected with expression constructs using the
Lipofectamine Plus transfection kit (Life Technologies Inc.,
Grand Island, NY) and processed for examination at 48 hours
post transfection. For confocal microscopy, double staining was
done using a commercially available antibody for ubiquitin
(anti-Ubiquitin; Dako, Copenhagen, Denmark). Subcellular
fractionation by differential centrifugation was performed with
transfected HeLa cell homogenates by progressively increasing
the centrifugal force (15). At each step, the pellet was saved as a
designated fraction and the supernatant was carried on to the
next centrifugation step. The centrifugation steps were as
follows: 600g for 10 min to collect nuclei, 10 000g for 10 min to
collect the heavy membrane fraction and 100 400g for 60 min,
and the pellet and supernatant were used as the light membrane
and cytosolic fractions, respectively. Fractionated samples were
western-blotted and probed with anti-GFP antibody or antilaforin antibody as described elsewhere (15).
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
ACKNOWLEDGEMENTS
We gratefully acknowledge our LD patients and their parents,
families and respective physicians for their participation in this
study. We thank our colleagues Dr Ryoji Morita for his advice
on haplotype analysis and Ms Keiko Shoda for her excellent
technical support. This study was supported in part by NIH
Grant 5P01-NS21908 awarded to A.V.D.-E. and by contributions from the Quebec (Mrs Odette Malenfant) and Sweden
(Mrs Vera Faludi) Lafora’s Disease Associations.
REFERENCES
1. Lafora, G.R. and Glueck, B. (1911) Contribution to the histopathology and
pathogenesis of myoclonic epilepsy. Bull. Gov. Hosp. Insane., 3, 96–111.
2. Schwarz, G.A. and Yanoff, M. (1965) Lafora’s disease, distinct clinicopathologic form of Unverricht’s syndrome. Arch. Neurol., 12, 172–188.
3. Busard, H.L., Renie, W.O., Gabreels, F.J.M., Jaspar, H.H.J., van Haelst,
U.J.G. and Slooff, J.L. (1986) Lafora’s disease: comparison of inclusion
bodies in skin and in brain. Arch. Neurol., 43, 296–299.
4. Busard, H.L., Gabreels-Festen, A.A.W.M., Renier, W.O., Gabreels, F.J.M.
and Stadhouders, A.M. (1987) Axilla skin biopsy: a reliable test for the
diagnosis of Lafora’s disease. Ann. Neurol., 21, 599–601.
5. Van Heycop Ten Ham, M.W. (1975) Lafora disease, a form of progressive
myoclonus epilepsy. In Vinken, P.J. and Bruyn, G.W. (eds), The Epilepsies.
17.
18.
19.
20.
21.
22.
23.
24.
1271
Handbook of Clinical Neurology, 15. North-Holland, Amsterdam,
pp. 382–422.
Delgado-Escueta, A.V., Ganesh, S. and Yamakawa, K. (2001) Advances in
the genetics of progressive myoclonus epilepsy. Am. J. Med. Genet., 106,
129–138.
Acharya, J.N., Satischandra, P., Asha, T. and Shankar, S.K. (1993) Lafora’s
disease in South India: a clinical, electrophysiologic, and pathologic study.
Epilepsia, 34, 476–487.
Serratosa, J.M., Delgado-Escueta, A.V., Posada, I., Shih, S., Drury, I.,
Berciano, J., Zabala, J.A., Antunez, M.C. and Sparkes, R.S. (1995) The
gene for progressive myoclonus epilepsy of the Lafora type maps to
chromosome 6q. Hum. Mol. Genet., 4, 1657–1663.
Sainz, J., Minassian, B.A., Serratosa, J.M., Gee, M.N., Sakamoto, L.M.,
Iranmanesh, R., Bohlega, S., Baumann, R.J., Ryan, S., Sparkes, R.S. and
Delgado-Escueta, A.V. (1997) Lafora progressive myoclonus epilepsy:
narrowing the chromosome 6q24 locus by recombinations and homozygosities. Am. J. Hum. Genet., 61, 1205–1209.
Minassian, B.A., Lee, J.R., Herbrick, J.A., Huizenga, J., Soder, S., Mungall,
A.J., Dunham, I., Gardner, R., Fong, C.Y., Carpenter, S. et al. (1998)
Mutations in a gene encoding a novel protein tyrosine phosphatase cause
progressive myoclonus epilepsy. Nat. Genet., 20, 171–174.
Serratosa, J.M., Gomez-Garre, P., Gallardo, M.E., Anta, B., de Bernabe,
D.B., Lindhout, D., Augustijn, P.B., Tassinari, C.A., Malafosse, R.M.,
Topcu, M. et al. (1999) A novel protein tyrosine phosphatase gene is
mutated in progressive myoclonus epilepsy of the Lafora type (EPM2).
Hum. Mol. Genet., 8, 345–352.
Minassian, B.A., Ianzano, L., Delgado-Escueta, A.V. and Scherer, S.W.
(2000) Identification of new and common mutations in the EPM2A gene in
Lafora’s disease. Neurology, 54, 488–490.
Minassian, B.A., Ianzano. L., Meloche, M., Andermann, E., Rouleau, G.A.,
Delgado-Escueta, A.V. and Scherer, S.W. (2000) Mutation spectrum and
predicted function of laforin in Lafora’s progressive myoclonus epilepsy.
Neurology, 55, 341–346.
Gomez-Garre, P., Sanz, Y., de Cordoba, S.R. and Serratosa, J.M. (2000)
Mutational spectrum of the EPM2A gene in progressive myoclonus
epilepsy of Lafora: high degree of allelic heterogeneity and prevalence of
deletions. Eur. J. Hum. Genet., 8, 946–954.
Ganesh, S., Agarwala, K.L., Ueda, K., Akagi, T., Shoda, K., Usui, T.,
Hashikawa, T., Osada, H., Delgado-Escueta, A.V. and Yamakawa, K. (2000)
Laforin, defective in the progressive myoclonus epilepsy of lafora type, is a
dual-specificity phosphatase associated with polyribosomes. Hum. Mol.
Genet., 9, 2251–2261.
Ganesh, S., Agarwala, K.L., Amano, K., Suzuki, T., Delgado-Escueta, A.V.
and Yamakawa, K. (2001) Regional and developmental expression of
Epm2a gene and its evolutionary conservation. Biochem. Biophys. Res.
Commun., 283, 1046–1053.
Wang, J., Stuckey, J.A., Wishart, M.J., and Dixon, J.E. (2002) A unique
carbohydrate binding domain targets the Lafora disease phosphatase to
glycogen. J. Biol. Chem., 277, 2377–2380.
Cox, T.C., Allen, L.R., Cox, L.L., Hopwood, B., Goodwin, B., Haan, E.,
Suthers, G.K. (2000) New mutations in MID1 provide support for loss of
function as the cause of X-linked Opitz syndrome. Hum. Mol. Genet., 9,
2553–2562.
Harada, M., Sakisaka, S., Terada, K., Kimura. R., Kawaguchi, T., Koga, H.,
Kim, M., Taniguchi, E., Hanada, S., Suganuma, T. et al. (2001) A mutation
of the Wilson disease protein, ATP7B, is degraded in the proteasomes and
forms protein aggregates. Gastroenterology, 120, 967–974.
Johnston, J.A., Ward, C.L. and Kopito, R.R. (1998) Aggresomes: a cellular
response to misfolded proteins. J. Cell. Biol., 143, 1883–1898.
Garcia-Mata, R., Bebok, Z., Sorscher, E.J., Sztul, E.S. (1999) Characterization and dynamics of aggresome formation by a cytosolic GFP-chimera.
J. Cell. Biol., 146, 1239–1254.
Hentze, M.W., and Kulozik, A.E. (1999) A perfect message: RNA
surveillance and nonsense-mediated decay. Cell, 96, 307–310.
Byers, P.H. (2002) Killing the messenger: new insights into nonsensemediated mRNA decay. J. Clin. Invest., 109, 3–6.
Ganesh, S., Shoda, K., Amano, K., Uchiyama, A., Kumada, S., Moriyama,
N., Hirose, S., and Yamakawa, K. (2001) Mutation screening for Japanese
Lafora’s disease patients: identification of novel sequence variants in the
coding and upstream regulatory regions of EPM2A gene. Mol. Cell.
Probes, 15, 281–289.