Download Warren, ST and Ashley, CT: Triplet repeat expansion mutations: The example of fragile X syndrome. Annual Review of Neuroscience 18:77-99 (1995).

ANNUAL
REVIEWS
Further
Quick links to online content
Annu. Rev. Neurosci. 1995. 18:77-99
Copyright © 1995 by Annual Reviews inc. All rights reserved
TRIPLET REPEAT EXPANSION
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
MUTATIONS: The Example of
Fragile X Syndrome
Stephen T. Warren and Claude T. Ashley, Jr.
Howard Hughes Medical Institute and Departments of Biochemistry and Pediatrics,
Emory University School of Medicine, Atlanta, Georgia 30322
KEY WORDS:
FMRl gene, mental retardation. X chromosome. unstable DNA, DNA
methylation
INTRODUCTION
Mental retardation represents a deficiency in intelligence, as measured by IQ,
with limited adaptive behavior that is normally reflected in maturation, learn­
ing, or social adjustment (American Psychiatric Association 1987). Approxi­
mately 1 to 3% of the population, depending upon definitions of adaptive
behavior, is mentally retarded (Popper 1988). The etiologies and determinants
of mental retardation are diverse and include socioeconomic influences leading
to extreme malnutrition and/or inadequate prenatal care; toxic insults, such as
that leading to fetal alcohol syndrome; trauma and infection; and genetic
factors (Popper 1988). At least 300 genetic disorders include mental retardation
as part of the phenotype, and genetic components are considered important
influences on related disorders such as attention deficit disorder and learning
disability (Smith et al 1983, Biederman et al 1987, McKusick et al 1992).
Since the early twentieth century a male predominance at all levels of mental
retardation, ranging from 1.5 to 3 times the incidence in females, has been
acknowledged (Penrose 1938). Although a variety of explanations have been
put forth, including the probable ascertainment bias of mentally retarded males
being more frequently institutionalized because of uncontrollable or violent
behavior, there is good reason to believe X-linked loci contribute significantly
to this gender inequity (Opitz 1986). Perhaps as many as 95 genes have been
tentatively assigned to the X chromosome where mutations lead to mental
retardation as at least part of the phenotype (Schwartz 1993). As the vast
77
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
78
WARREN & ASHLEY
majority of these mutations appear recessive, females, protected by an addi­
tional X chromosome, are rarely affected, and hence this leads to the excess
male mental retardates. However, most of these disorders are exceedingly rare,
and as single loci, none would be expected to appear frequently in the popu­
lation. Thus, an aggregate of all X-linked mutations involving mental retarda­
tion appeared responsible for gender differences until a single locus, that
responsible for fragile X syndrome, became recognized as a clinical entity that
eventually would account for nearly half of the male predominance of mental
retardation (Opitz & Sutherland 1984) .
HISTORICAL AND CLINICAL ASPECTS
Fragile X syndrome, an X-linked dominant disorder with reduced penetrance,
is the most frequent form of familial mental retardation, with a prevalence
of 1 in 1500 males and 1 in 2500 females (Gustavson et al 1986, Webb et
al 1986). It is the single most common etiology of X-linked mental retarda­
tion (Neri et al 1992, Brown & Jenkins 1992), and ranks second only to
Down's syndrome, which is most often sporadic, as the most frequent genetic
cause of mental retardation overall. First described by Martin & Bell (Martin
& Bell 1943), the syndrome cosegregates with a marker X chromosome that
is observed cytogenetically as a nonstaining constriction in the long arm of
Figure 1
Partial metaphase spread from a patient with fragile X syndrome showing the iso­
chromatid gap (arrow) near the distal end of the X chromosome long arm, which is indicative of the
fragile X site (From Warren
& Nelson 1994. Used by permission).
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
79
the X chromosome near the distal tip (Lubs 1969). Sutherland (1977) later
identified this constriction as an inducible fragile site at Xq27.3 (Figure 1);
it is induced in tissue culture by perturbation of folic acid or thymidine pools,
which results in limited amounts of dTIP for DNA synthesis (Sutherland
1977, Glover 1981). Upon analysis by electron microscopy, the fragile site
was described as an isochromatid gap at Xq27.3 (Harrison et al 1983). In
fragile X males, levels of fragile site expression in metaphase spreads from
peripheral lymphocytes range from 5-50% of X chromosomes analyzed,
whereas the extent of expression in female heterozygotes is on the order of
1-30% (Sutherland 1977, Sutherland et al 1985). Levels of induced fragile
site expression in metaphase spreads of normal controls have been shown to
be less than 1% (Jenkins et al 1986). For many years, prenatal diagnosis and
carrier testing relied upon scoring the fragile X marker in metaphase spreads,
but recent advances in the field have rendered cytogenetic tests supplemen­
tary (Warren & Nelson 1994).
The major clinical manifestations of fragile X syndrome in adult males most
notably include: mental retardation, ranging from mild to severe; mild facial
dysmorphia, characterized by long, narrow facies, large dysmorphic ears, and
prominence of the jaw and forehead; and macroorchidism with testicular
volumes greater than 25 ml (Hagerman 1991, Hagerman et al 1991). With the
obvious exception of macroorchidism, all of the phenotypic features of fragile
X described above may be exhibited by adult carrier females, although they
tend to be less severe (Hagerman et al 1992). Additional clinical signs often
contributing to the fragile X phenotype include connective tissue abnormali­
ties, such as joint hyperextensibility, pes planus, pectus excavatum, and mitral
valve prolapse (Opitz et al 1984, Hagerman & Synhorst 1984, Hagerman et
al 1984, Loehr et al 1986). The presence of autistic-like behaviors is frequently
reported in fragile X males (and occasionally in female carriers). These be­
haviors include hand-flapping, rocking, or hand-biting (repetitive motor be­
haviors); gaze aversion; tactile defensiveness (decreased social and nonverbal
communication skills); and repetitive speech patterns, perseveration, and echo­
lalia (dysfunctional verbal communication) (Brown et al 1982, Reiss et al 1986,
Reiss & Freund 1992). Psychiatric disorders, such as chronic depression,
schizotypal personality disorder, social avoidance, social anxiety, and with­
drawal, are reportedly more prevalent in fragile X carrier females than in
multifeature-matched controls (Reiss et al 1988, Hagerman & Sobensky 1989,
Freund et aI1993). Hyperactivity and attention deficit disorder (ADHD) appear
to be common in fragile X males before puberty (Hagerman & Sobensky 1989),
although it is unclear if these features are more frequent in fragile X syndrome
than is normally encountered in general mental retardation (Fisch 1993). How­
ever, because of the subtlety of the fragile X phenotype prior to puberty,
hyperactivity represents the presenting illness in many fragile X males (Hager-
80
WARREN & ASHLEY
man & Sobensky 1989). Additionally, because the phenotype of fragile X
syndrome is quite pleiotropic and difficult to diagnose de novo prior to puberty,
detailed checklists involving precise anthropometric measurements have been
devised (Butler et al 1991a, Butler et al 1991b, Butler et al 1992).
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
INHERITANCE PATTERN
The inheritance pattern of fragile X syndrome is both complex and unusual
for a sex-linked Mendelian trait. Although the inclusion of affected heterozy­
gous females as well as hemizygous males would suggest an X-linked domi­
nant mode of inheritance, the occurrence of apparently unaffected males who
carry the fragile X chromosome has been well documented (Martin & Bell
1943, Dunn et al 1963, Howard-Peebles & Friedman 1985, Froster-Iskenius
et al 1986). In analyses of a total of 206 fragile X pedigrees by both classical
and com pl ex segregation analyses, Sherman et al (1984, 1985) reported a 20%
deficit in affected males in the absence of evidence of any sporadic cases, a
finding best explained by reduced penetrance in males of 80%. Given the fact
that affected males rarely reproduced, these unusual nonpenetrant carrier males
were termed transmitting males because they passed the fragile X chromosome
to all of their obligate carrier daughters. Interestingly, nonpenetrant transmit­
ting males were also cytogenetically negative for fragile site expression in
lymphocytes (Howard-Peebles & Friedman 1985, Froster-Iskenius et aI1986).
Penetrance of mental impairment in females was also found to be reduced,
with an overall penetrance of 35%, in the study by Sherman et al (1984), and
disease expression was again found to be in direct correlation with cytogenetic
expression of the fragile site. Thus, a model of an X-linked dominant gene
with reduced penetrance was imposed (Sherman et al 1984, 1985).
Sherman et al (1984, 1985) made a most peculiar observation that the risk
of mental impairment associated with fragile X syndrome was a function of
one's position within the pedigree. As mentioned, carrier males are both
nonpenetrant and cytogenetically negative for fragile X syndrome. Interest­
ingly, Sherman et al (1984, 1985) found that the risk of mental impairment in
obligate carrier daughters of carrier males approached 0%. Siblings of these
males were also at relatively low risks of 9% and 5% in males and females,
respectively. Offspring of normal carrier females, however, incurred risks of
38% (males) and 16% (females) for mental impairment, which corresponded
to penetrances of 76% and 32%, while offspring of mentally impaired carrier
females displayed penetrances of 100% (males) and 56% (females). Thus,
disease expression was dependent upon passage of the fragile X through a
female, which led some to hypothesize that a premutation in carrier males
causing the fragile X chromosome to be more susceptible to alteration in female
gametes might account for this observation (Pembrey et al 1985). This odd
FRAGILE X SYNDROME
81
pattern of increasing risk with vertical transcention through a pedigree, referred
to as the Sherman paradox, was unique in human genetics, and it was only
after the molecular basis of fragile X syndrome was elucidated that this peculiar
pattern of inheritance was fully understood.
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
IDENTIFICATION OF THE FRAGILE X MUTATION
In 1991, the combined approaches of genetic mapping, physical mapping, and
somatic cell genetics culminated in the positional cloning of the fragile X locus
and the identification of the mutation responsible for fragile X syndrome. Prior
to this time, the fragile X disorder had been mapped genetically through linkage
analysis to the same interval as the Xq27 fragile site between the factor VIII
(12 cM distal) and factor IX (5 cM proximal) loci, which narrowed the region
of interest to approximately 20 megabases (Goodfellow et al 1985, Oberle et
al 1986). Also localized to the region surrounding the fragile site were the
enzymes HPRT (Xq26) (Pai et a11980) and G6PD (Xq28) (Filippi et aI1983).
By using rodent cells deficient for both of these enzymes, Warren and cowork­
ers (Warren et al 1987, Warren et al 1990) successfully constructed somatic
cell hybrids containing either exclusively proximal or exclusively distal por­
tions of the human fragile X chromosome. They did this by selection either
for or against the human HPRT enzyme activity followed by histochemical
staining for G6PD activity. Because it had previously been shown that the
fragile X site was cytogenetically expressed in hybrid cells, chromosomal
breakage at or near the fragile site was expected for most of the hybrids
(Warren & Davidson 1984). This was confirmed through hybridization of
known proximal and distal X chromosome markers to the obtained hybrids
(Warren et aI1990). The availability of proximal and distal hybrids, as well
as other X-breakpoint hybrids, allowed a number of new polymorphic loci to
be mapped relative to the fragile site, which narrowed the gap to within 3
megabases (Suthers et al 1990, Hirst et al 1991, Suthers et al 1991, Rousseau
et al 1991b).
The newly identified battery of DNA markers bracketing the fragile X locus
allowed isolation of yeast artificial chromosomes (YACs) with inserts spanning
the fragile X hybrid breakpoints (Dietrich et al 1991, Heitz et al 1991, Verkerk
et al 1991, Kremer et al 1991b). Two groups independently identified an
aberrantly methylated CpG island within this region of DNA in fragile X
patients. Because of the absence of normal cleavage by methylation-sensitive
enzymes in fragile X cells, a band of anomalous mobility on pulsed-field gel
electrophoresis were used to identify the abnormal CpG island (Bell et al 1991,
Vincent et aI1991). Oberle et al (1991) identified a 9-kb DNA segment from the
YAC 209G4 that contained the CpG island hypermethylated in fragile X patients
and that detected an apparent insertion in fragile X patients on Southern blots. Yu
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
82
WARREN & ASHLEY
et al ( 1991) and Verkerk et al (1991) concurrently yet independently identified a
single restriction fragment of approximately 5 kb that contained both the
proximal and distal fragile X hybrid breakpoints and also detected length
variation in fragile X patients on Southern blots. Kremer et al ( 199 1a) further
localized this instability to a single 1.0-kb Pst I fragment, pfxa2, that was
sequenced and found to contain an unusual CGG trinucleotide repeat of 43
copies. By using the polymerase chain reaction (PCR) and primers flanking the
repeat, Kremer et al (1991a) mapped the region of instability in pfxa2 to within
the CGG trinucleotide repeat itself, thereby suggesting that the length variation
observed in fragile X patients could result from length variation within the
trinucleotide repeat. The translocation breakpoints of various hybrids were also
mapped to within the CGG repeat, thus colocalizing the fragile site and the fragile
X mutation to this novel trinucleotide repeat (Kremer et aI199 1a).
In order to identify expressed sequences in the fragile X breakpoint cluster
region, four cosmid subclones of the YAC 209G4, which spanned the region
encompassing the erroneously methylated CpG island, were used as probes
against a eDNA library from human fetal brain (Verkerk et al 1991). Two
cDNA clones, BC22 and BCn, comprising a contig of 3765 nucleotides, were
identified and the DNA sequence of each was obtained. Analysis of this
sequence revealed a single major open reading frame that remained open at
the 5' end and encoded a predicted polypeptide of 657 amino acids. The CGG
repeat potentially displaying length variation in fragile X patients was con­
tained within the 5' portion of the open reading frame where it was initially
thought to encode an uninterrupted stretch of 30 arginine residues. However,
later studies demonstrate that the CGG repeat is confined to the 5' untranslated
region and is, therefore, not translated (Ashley et a1 1993a) (see below). Upon
northern analysis, a 4.8-kb transcript was detected in RNA from human brain
and placenta, which suggested that approximately 1 kb of sequence remained
to be determined. A zoo blot containing genomic DNA from a number of
eukaryotes, including lower organisms such as nematode and yeast, displayed
band(s) in lanes of all organisms except Drosophila when probed with BC22
under high stringency, thus indicating strong evolutionary conservation of this
gene, referred to as FMRl (fragile X mental retardation 1), and implying
functional importance of its protein product, FMRP. Initial data base searches
at both the nucleotide and amino acid levels, however, produced no significant
homologies with any previously reported sequences and provided no immedi­
ate clues as to the normal function of this predicted protein (see below).
FRAGILE X TRINUCLEOTIDE REPEAT EXPANSION
The delimitation of the region of DNA instability to within the CGG repeat
suggested the possibility that length variation observed in fragile X carriers
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
83
might result from the inherent instability of this reiterated trinucleotide repeat,
which would be reminiscent of other tandemly repeated sequences such as the
highly polymorphic dinucleotide repeats (Weber 1990). Indeed, analysis of the
eGG repeat of FMRl across normal populations revealed that it was highly
polymorphic, with repeat sizes ranging from 6-54 triplets and a mode of 30
(Fu et al 1991, Brown et al 1993, Jacobs et al 1993, Snow et al 1993). In
analysis of repeat sizes in fragile X pedigrees, Fu et al (1991) discovered that
individuals nonpenetrant for fragile X syndrome, carrier males and their obli­
gate carrier daughters, displayed eGG repeat lengths ranging from 54-200
repeats, whereas affected individuals exhibited repeat sizes in excess of 200
and often greater than 1000 repeats. This finding was in agreement with the
previous reports of apparently smaller insertions in carriers and larger inser­
tions in affected individuals (Oberle et al 1991, Yu et al 1991, Kremer et al
1991a), which Oberle et al (1991) referred to as premutations and full muta­
tions, respectively, in order to relate the size of the insertion with penetrance
of the disease. This correlation between repeat size and penetrance not only
allowed resolution of the Sherman paradox (Fu et al 1991), but it also paved
the way toward identification of other trinucleotide repeat expansion mutations
by substantiating earlier claims (Harper 1989) that diseases displaying genetic
anticipation, increasing severity, or decreasing age of onset of disease from
one generation to the next might share a common form of mutation.
In addition to establishing an association between repeat size and penetrance
for fragile X syndrome, studies of the repeat length of FMRl also aided in
defining the boundaries of meiotic instability. In the study by Fu et al (1991), all
FMRl eGG repeats of 54 and greater displayed meiotic instability, and smearing
of the full mutation bands on Southern blots suggested mitotic instability as well.
In a recent analysis of 116 families referred for fragile X testing, a FMRl eGG
repeat of 51 was shown to be meiotically stable through five generations,
whereas a repeat of 57 displayed meiotic instability in the next generation (Snow
et al 1993). Also in this study, a eGG repeat of only 61 in a carrier female
expanded into the full mutation range in the following generation. However,
another premutation allele of 90 in a different carrier female only expanded to
115 upon transmission to her offspring. Therefore, repeats of 54 and above
clearly display meiotic instability and should be considered premutations;
however, the expansion of premutation alleles into the full mutation range must
be dependent upon other factors yet to be identified.
METHYLATION OF THE FMRl GENE
In addition to trinucleotide repeat expansion in fragile X patients, abnormal
methylation of a CpG island located 250 base pairs upstream of the CGG
repeat of FMRl was also found in individuals expressing the fragile X phe-
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
84
WARREN & ASHLEY
notype (Bell et al 1991, Heitz et al 1991, Oberle et al 1991, Vincent et al
1991). Like expansion of the repeat, the extent of CpG-island methylation is
directly correlated with disease expression because carrier males, which are
unaffected, display unmethylated CpG islands. Individuals with the full mu­
tation, however, have fully methylated CpG islands (Oberle et al 1991). Un­
affected carrier females display methylation only on the inactive X chrom­
osome, while affected females tend to be skewed toward excess methylation
(> 50%), although this is not always the case (Oberle et al 1991). The CGG
repeat, which is itself rich in CpG dinucleotides, is also fully methylated in
affected individuals with full mutations or at normal repeat copy number on
the inactive X (Hansen et al 1992, Hornstra et al 1993). The identification of
abnormal methylation associated with fragile X syndrome is in agreement with
Laird's (1987) hypothesis, which states that fragile X syndrome results from
persistence of an imprint that is due to failure of erasure of X inactivation in
a carrier female.
Methylation of the CpG island adjacent to the FMRI gene correlates with
transcriptional silencing of FMRI and expression of the fragile X phenotype.
In a study by Pieretti et al ".(1991), 16 of 20 affected males with expanded
repeats in the full mutation range displayed fully methylated CpG islands and
produced no FMRI transcript. Of the four affected males with FMRI expres­
sion, three were identified as mosaics for fully methylated full mutations and
unmethylated premutations, and one was a methylation mosaic with a partially
unmethylated full mutation. In a prenatal diagnosis of an affected fetus with
a full mutation, Sutcliffe et al (1992) reported an absence of methylation in
chorionic villus samples that correlated with normal levels of FMRI expres­
sion, while in fetal tissue a fully methylated CpG island was found but not
FMRI message. Recently, four carrier males with repeat sizes between 200
and 400 but with unmethylated CpG islands and repeats were found to be
intellectually and physically normal by observation (Loesch et aI1993). Con­
versely, two males exhibiting partially methylated CpG islands and CGG
repeats between 200 and 300 copies also displayed mild clinical and physical
manifestations of fragile X syndrome (McConkie-Rosell et al 1993).
Although many carrier females have premutation alleles, a significant pro­
portion of the normal carrier females carry full mutations with expansions well
beyond 200 repeats as well as the abnormal methylation of the FMRI gene.
Rousseau et al (1991a) demonstrated that approximately half (53%) of women
with full mutations are penetrant with IQ levels in the borderline and mentally
retarded range. The other half presumably escape the effect of the full mutation
because of X inactivation (lyonization) patterns. However, Taylor et al (1994)
showed that the pattern of X inactivation, i.e the proportion of fragile X
chromosomes active versus inactive, in peripheral lymphocytes of full mutation
females was not a predictor of intellectual function. Thus, it is most reasonable
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
85
to conclude that the protein encoded by FMRI is cell autonomous and that the
proportion of cells in key regions of the brain with the fragile X chromosome
in the active state (but still not expressing FMRI because of the abnormal
methylation) are responsible for penetrance in females.
Thus, the absence or reduction of FMRI expression as a function of meth­
ylation status of the upstream ep G island and the eGG repeat is thought to
be the basis for fragile X syndrome. The fact that at least four fragile X patients
lacking repeat expansion but possessing other mutations of the FMRI gene
itself, three large deletions and one point mutation, have been identified sup­
ports this hypothesis (Gideon et al 1992, Wohrle et al 1992, De Boune et al
1993, Tarleton et aI1993). The role of repeat expansion in abnormal methyl­
ation, or vice versa, remains to be determined.
DETECTION OF THE FRAGILE X MUTATION
Both the expansion of the eGG repeat and the abnormal methylation of the
FMRI gene are important diagnostic indicators. Methods have been developed
to assess these mutational parameters in patients (reviewed by Warren &
Nelson 1994). Detection can employ either Southern blotting (Figure 2a) or
the polymerase chain reaction (Figure 2b). Southern blotting most commonly
utilizes the digestion of genomic DNA with EcoR I and the methylation­
sensitive enzyme BssH II or Eag I. EcoR I digestion liberates a 5.2-kb fragment
containing the promoter and first exon of FMRl, which includes the eGG
repeat. Double digestion with BssH II or Eag I cleaves normal male DNA into
2.4-kb and 2.8-kb fragments; the latter fragment contains the trinucleotide
repeat. The active X chromosome of normal females also displays this pattern,
but the inactive normal X in a female is methylated, as is common for many
X-linked loci, and is resistant to BssH II or Eag I digestion. Thus, normal
females exhibit three bands as shown in lane two of Figure 2a.
Premutation alleles exhibit a shift in the 2.8-kb band to a slightly higher
molecular weight that reflects the increase in the CGO repeat to the 54-230
repeat range. Full mutations display a large, somewhat diffuse band that is
generally not cleaved by BssH II or Eag I. These bands reveal the expansion
of the eGG repeat into the range of many hundreds of triplets as well as the
abnormal methylation of the FMRI gene. Females with full mutations display
a similar pattern superimposed upon that of the normal X (in both the active
and inactive states).
Analysis by peR has allowed a much more precise estimation of repeat
length, particularly in normal and premutation alleles (Fu et aI1991). However,
amplification of full mutations is difficult, though not impossible (Pergolizzi
et al 1992). As shown in Figure 2b, PCR amplification of the eGG repeat by
using flanking primers of unique sequence demonstrates the remarkable insta-
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
86
Figure 2
WARREN & ASHLEY
Molecular detection of the FMRl gene. (A) Southern blot of genomic DNA digested with
EcoRI and BssHII, resolved through agarose, transferred to membrane, and probed with radiolabeled
FMRl gene fragment pE5.1 (Verkerk et al 1991). (8) peR analysis of the FMRl repeat by using
flanking primers as described by Pergolizzi et al (1993). Numbers next to pedigree symbols reflect
the number of FMRl triplet repeats (Both figures from Warren & Nelson 1994. Used by permission).
bility of the premutation alleles, where siblings often have alleles unique from
one another as well as distinct from the parent. Analysis by peR is useful for
defining accurate premutation repeat lengths and for demonstrating instability,
which may be of some utility in genetic counseling (Warren & Nelson 1994),
and is useful in situations where limited DNA is a constraint.
OTHER TRINUCLEOTIDE REPEAT EXPANSION
DISEASES
Since the emergence of the trinucleotide repeat expansion mutation responsible
for fragile X syndrome, six other diseases caused by trinucleotide expansions
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X S YNDROME
87
Ful Mula' on
have been identified (Table 1) (reviewed by Warren & Nelson 199 3, 1994):
spinal and bulbar muscular atrophy (SBMA) (La Spada et aI1991); myotonic
dystrophy (DM) (Aslanidis et a11992, Brook et a11992, Fu et a11992, Harley
et al 1992, Mahadevan et al 1992); Huntington's disease (HD) (Group 199 3);
spinal and cerebellar ataxia type 1 (SCA1) (Orr et al 199 3); FraX E mental
retardation (Knight et al 1993); and dentatorubral pallidoluysian atrophy
(DPA) (Koide et a1 1994, Nagafuchi et aI1994). Interestingly, all trinucleotide
repeat diseases described thus far are neurologic diseases, although there is
currently no good hypothesis to account for this pattern. Similarities between
the trinucleotide repeats of these diseases include that they are all normally
88
Table 1
WARREN & ASHLEY
Genetic diseases associated with unstable trinucleotide repeats
Change in
Repeat number
gene
Trinucleotide
Disease
repeat'
Normal
Fragile X syndrome
CGG 5'UTR
6-52
Myotonic dytrophy
CTG 3'UTR
5-37
Spinal & bulbar muscular
CAG coding
Huntington's disease
Spinocerebellar ataxia
Carrier
50-200
Affected
230 to >1000
function
Loss
50 to >1000
mRNA stability
12-34
40-62
Gain
CAG coding
11-36
42-100
Gainb
CAG coding
19-36
43-81
Gainb
CAG coding
7-23
49-75
Gainb
CCG-
6-25
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
atrophy
type 1
Dentatorubral
pallidoluy-
sian atrophy
FRAXE mental retardation
a
All repeat are exonic with the position shown;
VTR
116-133
200 to >800
Lossb
is untranslated region.
b Functional change is not established and is shown based upon similarity to other examples.
poly morphic, GC-rich, and expressed sequences, while di f ferences include the
exact nucleotide content of the repeat and the location of the repeats within
the various messages (Richards & Sutherland 1992). In the majority of these
diseases ( HD, SBMA, DPA, and probably SCAl), an unstable CAG repeat
within the coding region o f the respective genes undergoes expansion, which
produces an a mplified polyglutamine tract in the nascent protein. In the case
of X-linked SBMA, Warren & Nelson (199 3) have suggested that the expan­
sion of the polyglutamine tract in the androgen receptor must constitute a gain
of function, perhaps through allowing the receptor to interact with pro moters
that it does not normally recognize, because the absence o f mani festations o f
testicular fe minization make it unlikely that the receptor is nonfunctional.
Although the function o f the gene products in HD and SCAI are currently
unknown, the do minant inheritance pattern o f these autosomally inherited
disorders argues that expansion o f the polyglutamine tract in these two diseases
will constitute a gain-of- function as well. In myotonic dystrophy, a CTG repeat
in the 3' untranslated region (UTR) of the myotonin protein kinase gene is
enlarged, which results in altered message stability and aberrant intracellular
signaling due to abnormal dosage of this protein kinase (Fu et a1 199 3, Sabourin
et aI199 3). As previously discussed, expansion o f a CGG repeat in the 5' UTR
of the FMRI gene is associated with abnor mal methylation o f the repeat and
an upstrea m CpG island leading to transcriptional silencing of FMRI. No
infor mation concerning the mechanis m of disease in FraX E mental retardation
is currently available because the expanded CGG repeat has not yet been
ascribed to a gene.
Of all of these repeat expansion mutations, only the repeats of DM and
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
89
fragile X syndrome, which are within untranslated regions, have full mutations
with repeat expansions exceeding several hundred copies. Repeat expansions
of the other diseases are usually 2-4 ti mes normal, and maxi mu m repeat sizes
rarely exceed 100 copies. This i mplies that the mechanis m of expansion of
fragile X syndro me and myotonic dystrophy could be similar, although it is
puzzling that DM does not display the parent-of-origin requirement for expan­
sion observed in fragile X syndrome. However, reductions in trinucleotide
repeat sizes, which are almost never seen in fragile X syndrome, are o ften seen
in paternal DM alleles (Reyniers et al 1993). Nevertheless, given the fact that
the full mutations of DM or fragile X are not a menable to cloning, the other
trinucleotide diseases with more subtle trinucleotide expansions will most
likely be more infor mative with regard to the mechanis m o f trinucleotide
instability.
MECHANISM OF THE FRAGILE X REPEAT EXPANSION
The molecular mechanis m of trinucleotide repeat expansion still re mains an
enigma, although recent advances in the fragile X field have been enlightening.
In previous segregation analyses of over 200 fragile X pedigrees, no sponta­
neous mutations in fe male ga metes causing fragile X syndrome were identified
(Sherman et a11984, 1985). Given the fact that fragile X alleles are constantly
being re moved from the gene pool because of reduced fitness in affected
individuals, Sherman et al (1984, 1985) predicted an extraordinarily high
mutation rate in male gametes, on the order of 7.2 x 10-4, to account for the
high prevalence of fragile X syndro me within the population. However, sub­
sequent analyses of fragile X pedigrees have failed to detect any new mutations
producing fragile X syndrome (Jacobs et al 1986, Rousseau et al 1991a, Yu
et al 199 2, S mits et al 1993). Consistent with the absence of new mutations,
S mits et al (1993) recently found that five fragile X males related through
co m mon ancestors from the eighteenth century were all concordant for the
same allele at the DXS548 locus, a polymorphic marker locus located 15 0 kb
proximal to FMRl (Riggins et al 1992). This finding attests to the very old
nature of the fragile X mutation in this family and suggests the possibility of
linkage disequilibrium in the region enco mpassing the fragile X locus. Haplo­
type analyses of markers flanking the fragile X locus have, in fact, revealed
significant linkage disequilibrium between fragile X syndrome and certain
haplotypes, which indicates that a s mall group o f founder chromoso mes, per­
haps as few as six, might account for the vast majority of fragile X mutations
observed today (Richards et al 199 2, Oudet et al 1993). The observation of
linkage disequilibrium in the region of the fragile X locus is re miniscent of
the situation in individuals with myotonic dystrophy (DM), the gene of which
is in co mplete linkage disequilibrium with a single insertion/deletion poly mor-
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
90
WARREN & ASHLEY
phis m located 5 kb from the unstable CTG repeat ( I mbert et aI199 3). However,
un like the myotonic dystrophy gene, the fragile X locus appears to be in linkage
equi libriu m with a discrete subset of hap lotypes, which suggests that multip le
mutations of the pri mordia l FMRl repeat are responsib le for the fragile X
chro moso mes in today's popu lation (Smits et a l 1993).
A low preva lence of new mutations and marked linkage disequi libriu m
suggesting highly ancient mutations are unusual findings for X- linked muta­
tions (Chakravarti 1992), and this scenario is hard to reconcile in light of the
high preva lence of fragile X syndrome in the popu lation. One possib le expla­
nation for this pheno mena is an unexpectedly high nu mber of pre mutations in
the popu lation that go undetected in the absence of any obvious phenotype.
In favor of this hypothesis, Snow et a l (199 3) found the frequency of premuta­
tion a l le les in the nor ma l popu lation to be 0.8%. A mu ltistep model has been
proposed that invokes a discrete set of mutations of a nor ma l primordial a l le le
to a pre-pre mutation, or a predisposed a l le le, that is relatively stab le for many
years (thus accounting for the linkage disequilibriu m observed ) but has the
potential to convert to unstable premutations that rapidly expand to fu ll mu­
tations with concurrent disease expression (Morton & Macpherson 1992). In
light of this mode l and the above linkage disequi libriu m data, Oudet et a l
(199 3) asserted that the most co m mon 30 a l le le of the CGG repeat of FMRl
(Brown et al 1993, Snow et al 199 3) was probab ly the pri mordia l a l le le and
that larger a l le les in the range of 38-40 repeats represent the predisposed a l le les
that u lti mate ly account for a l l of the fragile X chro moso mes. In addition, it
has been suggested that a mutation of one of the interspersed AGG trip lets to
CGG within the nor ma l ly cryptic repeat of FMRl (Pieretti et a1 1991, Verkerk
et a1 1991) could account for the pri mordial mutation event ( Oudet et a1199 3)
because repeats that are purer in content are less stable ( Weber 1990 ). Final ly,
superimposed upon this mode l is the like lihood of poly merase s lippage during
replication of these reiterated sequences ( I mbert et a l 1993), which accounts
for the s ma l ler, inert variation in the nor ma l CGG repeat length contributing
to the nor mal ly poly morphic nature of the repeats. If this mode l is va lid, then
it fol lows that the primary new mutations responsible for the cases of fragile
X syndro me today probab ly occurred many generations ago, and any new
mutations identified today may not manifest the mselves phenotypica l ly for
many years to co me.
The apparent dearth of expansion of premutation a l le les into the fu l l muta­
tion range upon trans mission from carrier males to their obligate carrier daugh­
ters has also recently been addressed. Upon discovery of a rare mating event
by an affected male, Wi l lems et al (1992) fol lowed the segregation of the
fragile X mutation from an affected male through his nor ma l carrier daughter
to an affected grandson. At the mo lecular level, the daughter received only a
premutation a l le le, whereas the grandfather was mosaic for fu l l and premuta-
FRAGILE X S YNDROME
91
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
tion-size alleles, which suggested a priori that the grandfather displayed germ­
line mosaicis m as well as so matic mosaicism. However, subsequent analyses
o f a f fected males with full mutations in so matic tissues revealed only pre muta­
tion alleles in their sper m (Reyniers et al 1993). One interpretation o f these
findings is that partial trinucleotide repeat expansion occurs within the male
ga mete, with subsequent selection against sper m with fully expanded repeats.
Another interpretation is that trinulceotide repeat expansion occurs postzygoti­
cally following the sequestering o f primordial ger m cells that are spared the
massive expansion. However, a maternal i mprint must be invoked that leads
to so matic expansion of maternal pre mutations. With regard to this model o f
postzygotic expansion, i t will be crucial t o determine whether oocytes with
full mutations are observed in ovaries o f carrier fe males. To date, no such
analysis o f repeat sizes in oocytes has been reported.
A prediction o f such a postzygotic expansion model is that a variable degree
o f repeat length mosaicis m should be present in the so matic tissues o f a f fected
patients. Indeed, so matic mosaicism in cells o f fragile X patients has been well
documented ( Fu et al 1991, Pieretti et al 1991, Rousseau et al 1991a). In a
study o f 511 individuals from 63 fragile X fa milies, Rousseau et al (1991b)
reported that approxi mately 15% o f individuals with mutations within the fully
expanded range also displayed pre mutation alleles in a subset o f cells. In
individuals such as these, the pre mutation allele is predo minately un methylated
and transcriptionally active, whereas the full mutation allele is most often fully
methylated and transcriptionally silent (Pieretti et al 1991). In analysis o f a
13-week male fetus, nearly identical mosaic patterns were observed across all
tissues analyzed, and subsequent prolonged cell culture provided no evidence
o f mitotic instability o f the alleles initially present (Wohrle et aI1993). Inter­
estingly, chorionic villi obtained at 10 weeks gestation for prenatal diagnoses
displayed fully expanded, unmethylated FMRI repeats, thereby deli miting the
window for trinucleotide repeat expansion to be fore this ti me and methylation
o f expanded repeats to sometime afterwards. Thus, many data support the idea
that expansion occurs postconceptionally, and o f the two models described, it
is the most widely accepted at the present time.
One major shortcoming of the postzygotic model of repeat expansion is its
inability to satis factorily account for the strict parental influence observed in
fragile X pedigrees. I f this model is valid, then an i mprinting phenomenon
clearly must also occur to account for the absence of repeat expansion in
daughters o f carrier males. As previously mentioned, Laird (1987) has pro­
posed that fragile X syndro me results fro m the abnor mal persistence o f an
i mprint due to inco mplete erasure of X-inactivation on the fragile X chromo­
some in female gametes. Although the observation of abnormal methylation
associated with repeat expansion is consistent with this hypothesis, the occur­
rence o f mosaicis m for full and pre mutation alleles and the fact that pre muta-
92
WARREN & ASHLEY
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
tion alleles are unmethylated (Pieretti et al 1991) are dif ficult to reconcile with
regard to Laird's hypothesis. Further more, corroborating data from two labs
indicate that expansion precedes methylation in the embryo (Sutherland et al
1991, Sutclif fe et al 1992), which makes it questionable whether methylation
in any form contributes to the lack o f expansion in daughters of carrier males.
So, validation o f the postzygotic model o f repeat expansion is incumbent upon
identi fication o f the factor responsible for the strict maternal origin o f expanded
FMRI alleles.
CHARACTERIZATION OF FMRl AND THE GENE
PRODUCT
Since the identi fication o f the FMRI gene, much has been gleaned with regard
to its physical structure and pattern of expression. The FMRI gene includes
17 exons encompassing 38 kb o f the human genome (Eichler et al 1993 ). The
eGG repeat resides within the first exon of the gene and is positioned 25 0
basepairs downstream o f the CpG island that is aberrantly methylated in fragile
X patients ( Fu et al 1991, Eichler et al 1993 ) and may serve as the pro moter
o f the FMRI gene ( Hwu et al 1993 ). On Northern analysis, expression in a
nu mber o f human tissues both related and unrelated to the fragile X phenotype
is observed; these include brain, testes, placenta, lung, kidney, and heart ( Hinds
et al 1993 ). In situ R NA expression studies in adult mouse tissues expanded
the list of tissues by adding esophageal epithelium, thymus, spleen, ovary, eye,
colon, uterus, thyroid, and liver; expression in heart, aorta, or muscle was not
observed in the adult mouse ( Hinds et al 1993 ). In cross sections of mouse
brain, Hinds et al (1993 ) reported expression predominately in neurons as
opposed to white matter or glial celis, with highest levels o f expression in the
cerebellu m, habenula, and the granular layers o f the hippocampus and cerebral
cortex. In a recent in situ R NA expression analysis o f cross sections o f brain
from a 25-week-old human fetus (Abitbol et a11993), the authors report FMRI
expression mainly in neurons; the nucleus basalis magnoceliularis and areas
o f the hippocampus are most intense. These results are consistent with the
observations by Hinds et al (1993 ) in that predominant labeling is o bserved in
areas critical to limbic circuitry, which is a pathway in the brain involved in
cognition, memory, and behavior.
Im munohistochemical analysis with antibodies against FMRP con firmed the
predo minant neuronal and sparse glial distribution o f FMRP previously indi­
cated by in situ R NA studies (Devys et al 1993 ). In addition, two reports o f
intracellular immunolocalization o f FMRP concur that the protein is localized
predo minately to the cytosol, although nuclear localization in the a bsence o f
cytosolic expression was observed i n a small proportion o f cells (Devys e t al
1993, Verheij et aI1993 ). Interestingly, the a mino ter minal half o f the protein
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
93
displayed exclusive nuclear localization according to Devys et al (1993),
although this half of the protein displays no nuclear localization signal and
lacks the putative nuclear localization signal near the carboxy ter minus de­
scribed by Verkerk et al (1991 ). Thus, it re mains unclear whether FMRP might,
under certain conditions, have a functional role in the nucleus.
The mouse FMRI ho mologue has been identified and shown to be re mark­
ably conserved, with nucleotide and amino acid identity values of 95% and
9 7%, respectively, within the coding region (Ashley et al 1993a). The CGG
repeat responsible for the fragile X phenotype in hu mans is also conserved in
mouse both in its position and nucleotide sequence, although the repeat nu mber
is reduced to nine copies (8 CGGs and 1 CGA) (Ashley et al 1 993a). An
in-frame stop codon identified upstream of the human repeat and a conserved
ATG located 66 and 69 nucleotides downstream of the CGG repeat in mouse
and hu man, respectively, delimit the CGG repeat to the 5' untranslated region
( 5' UTR) of the FMRI message and predict a putative protein product of 6 9
kD (Ashley et a l 1993a ). Additionally, alternative splicing of the FMRI gene
has been docu mented. The results of this splicing predict at least twelve
potential isofor ms of the FMRI predicted protein, FMRP, in all tissues ana­
lyzed and suggest the potential for intracellular functional diversity of the
FMRI gene product (Verkerk et al 1 993, Ashley et al 1 993a). In favor of the
notion of functional heterogeneity is the fact that a subset of the alternative
splicing events in FMRI alter the normal reading frame and produce isofor ms
with novel carboxy ter mini and quite distinct hydropathy profiles (Ashley et
al 1993a).
Two copies of a 30-a mino acid do main conserved across evolution fro m
bac teri a to humans have been identified within the amino terminal half of the
FMRP predicted sequence (Ashley et al 1993b, Siomi et al 1993b). This
repeated do main, termed a KH do main (Siomi et al 1993a ), is also present in
a nu mber of proteins that have an interaction with R NA in co m mon, including
one member of the hnRNP class of ribonucleoproteins, hnRNP K. Also iden­
ti fied within the carboxy ter minal portion FMRP a mino acid sequence are two
copies of the RGG box (arginine-glycine-glycine) (Ashley et al 1 993b), an
a mino acid motif i mplicated in both R NA as well as DNA binding (Christensen
& Fuxa 1988). Nucleic acid binding assays of in vitro translated FMRP reveal
that it is an R NA binding protein that binds to its own R NA transcript with
high affinity (Kd
5.7 nM) and to a lesser extent to a pool of R NAs from
human fetal brain; weak binding of FMRP to both single-strand and double­
strand DNA is observed as well (Ashley et al 1 993b). Selective binding of
FMRP to individual human fetal brain transcripts occurs, and it is suggested
that approxi mately 4% of the messages present in human fetal brain might be
bound by FMRP, which is consistent with the highly pleiotropic nature of the
fragile X phenotype (Ashley et al 1993b ). Finally, a 2:1 stoichiometry of
=
94
WARREN & ASHLEY
binding of R NA:protein is reported, which i mplies that each protein molecule
has the potential to interact with two R NA molecules (Ashley et al 1993b).
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
CONCLUSIONS
Fragile X syndrome accounts for a major proportion of mental retardation in
hu mans. In addition to the intellectual deficit, these patients also exhibit a wide
range of behavioral problems, including autistic-like features. Thus, the mo­
lecular understanding of this syndrome should lead to further understanding
of the higher cognitive functioning in hu mans. Indeed, much has already been
learned from the cloning of the fragile X gene, FMR1. The mutational basis
of the syndro me is the extraordinary expansion of a trinucleotide repeat within
the 5' untranslated region of the gene. Since this discovery, six other genetic
diseases, all neurological in nature, have now been attributed to unstable triplet
repeats within exons of different genes. This novel mutational mechanis m will
likely be found responsible for even more genetic disorders. The mechanis m
of repeat expansion is poorly understood and is a fertile and i mportant area of
inquiry.
When the CGG repeat in the FMRl gene exceeds 23 0 repeats, the gene is
heavily methylated and transcriptionally suppressed. The absence of the en­
coded protein, FMRP, is therefore believed responsible for the mental retar­
dation and associated phenotype. This cytosolic protein has been de monstrated
to be a selective R NA-binding protein, interacting with as much as 4% of
human brain mRNA. Thus, the further understanding of this interaction and
the identification of those genes whose messages interact could lead to a great
deal of insight into the mechanis ms of cognitive function in hu mans.
ACKNOWLEDGMENTS
STW is an investigator with the Howard Hughes Medical Institute and CTA
is a predoctoral fellow with the March of Dimes Birth Defects Foundation.
Any Annual Review chap ter, as well as any article cited in an Annual Review chapter,
may be p urchased from the AnnulIl Reviews Preprints lind Reprints service.
1-800-347-8007; 415-259-5017; email: [email protected]
Literature Cited
Abitbol M, Menini C, Delezoide A-L, Rhyner
T, Vekemans M, Mallet 1. 1993. Nucleus
basalis magnolaris and hippocampus are the
major sites of FMR-l expression in the
human fetal brain. Nature Genet. 4:147-52
American Psychiatric Association. 1987. Diag­
nostic and Statistical Manual of Mental Dis-
orders, pp. 28-33. Washington DC: Am.
Psychiatr. Assoc. 567 pp. 3rd ed.
Ashley CT, Sutcliffe IS, Kunst CB, Leiner HA,
Eichler EE, et al. 1993a. Human and murine
FMR-l: alternative splicing and translational
initiation downstream of the CGG-repeat.
Nature Genet. 4:244-51
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
FRAGILE X SYNDROME
Ashley CT, Wilkinson KD, Reines D, Warren
ST. 1993b. FMRI protein: conserved RNP
family domains and selective RNA binding.
Science 262:563--66
Aslanidis C, Jansen G, Amemiya C, Shutler G,
Mahadevan M, et al. 1992. Cloning of the
essential myotonic dystrophy region and
mapping of the putative defect. Nature
355:548-51
Bell MV, Hirst MC, Nakahori Y, MacKinnon
RN, Roche A, et al. 1991. Physical mapping
across the fragile X: hypermethylation and
clinical expression of the fragile X syndrome.
Cell 64:861-66
Biederman J, Munir K, Knee D, Armentano M,
Autor S, et al. 1987. High rate of affective
disorders in probands with attention deficit
disorders and their relatives-a controlled
family study. Am. 1. Psychiatry 144:330-33
Brook JD, McCurrach ME, Harley HG, Buckler
AJ, Church D, et al. 1992. Molecular basis
of myotonic dystrophy: expansion of a
trinucleotide (CTG) repeat at the 3' end of a
transcript encoding a protein kinase family
member. Cell 68:799-808
Brown WT, Houck GE, Jeziorowska A, Levin­
son FN, Ding X, et al. 1993. Rapid fragile X
carrier screening and prenatal diagnosis
using a nonradioactive PCR test. JAMA.
270(13):1569-75
Brown WT, Jenkins EC. 1992. The fragile X
syndrome. Mol. Genet. Med. 2:39-65
Brown WT' Jenkins EC, Friedman E, Brooks
J, Wisniewski K, et al. 1982. Autism is as­
sociated with fragile X syndrome. 1. Autism
Dev. Dis. 12:303-8
Butler MG, Allen A, Haynes JL, Singh DN,
Watson MS, Breg WR. 1991a. Anthropomet­
ric comparison of mentally retarded males
with and without the fragile X syndrome.
Am. 1. Med. Genet. 38:260-68
Butler MG, Brunschwig A, Miller LK, Hager­
man RJ. 1992. Standards for selected an­
thropometric measurements in males with
the fragile X syndrome. Pediatrics 89: 105962
Butler MG, Mangrum T, Gupta R, Singh DN.
1991b. A 1 5-item checklist for screening
mentally retarded males for the fragile X syn­
drome. Clin. Gene t. 39 : 347-54
Chakravarti A. 1992. Fragile X founder effect.
Nature Genet. 1:237-38
Christensen ME, Fuxa KP. 1988. The nucleolar
protein, B-36, contains a glycine and dimeth­
ylarginine-rich sequence conserved in sev­
eral other nuclear RNA-binding proteins.
Biochem. Biophys. Res. Comm. 155(3):
1278-83
De Boulle K, Verkerk AJMH, Reyniers E, Vits
L, Hendrich J, et al. 1993. A point mutation
in the FMR-l gene associated with fragile X
mental retardation. Nature Genet. 3(1): 3135
95
Devys D, Lutz Y, Rouyer N, Bellocq J-P, Man­
del J-L. 1993. The FMR-l protein is cyto­
plasmic, most abundant in neurons, and
appears normal in carriers of the fragile X
premutation. Nature Genet. 4:335-40
Dietrich A, Kioschis P, Monaco AP, Gross B,
Kom B, et aI. 1991. Molecular cloning and
analysis of the fragile X region in man. Nucl.
Acids Res. 19:2567-72
Dunn HG, Renpenning H, Gerrard JW, Miller
JR, Tabata T, Federoff S. 1963. Mental re­
tardation as a sex-linked defect. Am. 1. Men­
tal De! 67:827-48
Eichler EE, Richards S, Gibbs RA, Nelson DL.
1993. Fine structure of the human FMRI
gene. Hum. Mol. Genet. 2:1 147-53
Filippi G, Rinaldi A, Archidiacono N, Rocchi
M, Balasz I, Sinas-Calco M. 1983. Brief re­
port: linkage between G6PD and fragile X
syndrome. Am. 1. Med. Genet. 15: 1 1 3-19
Fisch GS. 1993. What is associated with the
fragile X syndrome? Am. 1. Med. Genet.
48(2):112-2 1
Freund LS, Reiss AL, Abrams MT. 1993. Psy­
chiatric disorders associated with fragile X
in the young female. Pediatrics 91:321-29
Froster-Iskenius U, McGillivray BC, Dill FJ,
Hall JG, Herbst DS. 1986. Normal male car­
riers in the fra(X) form of X-linked mental
retardation (Martin-Bell syndrome). Am. J.
Med. Genet 232:619-32
Fu Y-H, Friedman DI, Richards S, Pearlman
JA, Gibbs RA, et al. 1993. Decreased expres­
sion of myotonin-protein kinase messenger
RNA and protein in adult form of myotonic
dystrophy. Science 260:235-38
Fu Y-H, Kuhl DP, Pizzuti A, Pieretti M,
Sutcliffe JS, et al. 1991. Variation of the
CGG repeat at the fragile X site results in
genetic instability: resolution of the Sherman
paradox. Cell 67:1047-58
Fu Y-H, Pizzuti A, Fenwock RG, King J,
Rajnarayan S, et al. 1992. An unstable triplet
repeat in a gene related to myotonic muscular
dystrophy. Science 255 : 1 256-58
Gideon AK, Baker E, Robinson H, Partington
MW, Gross B, et al. 1992. Fragile X syn­
drome without CCG amplification has an
FMRI deletion. Nature Genet. 1:341-44
Glover TW. 1981. FUdr induction of the X
chromosome fragile site: evidence for the
mechanism of folic acid and thymidine de­
privation. Am. J. Hum. Genet. 33:234-42
Goodfellow PN, Davies KE, Ropers HH. 1985.
Report of the committee on the genetic con­
stitution of the X and Y chromosomes.
Cytogenet. Cell. Genet. 40:296-352
Group THDCR. 1993. A novel gene containing
a trinucleotide repeat that is expanded and
unstable on Huntington's disease chromo­
somes. Cell 72:971-83
Gustavson K-H, Blomquist H, Holmgren G.
1986. Prevalence of fragile-X syndrome in
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
96
WARREN & ASHLEY
mentally retarded children in a Swedish
county. Am. J. Med. Genet 23:58 1 -88
Hagerman RJ. 1991. Physical and behavioral
phenotype. In Fragile X Syndrome, Diagno­
sis, Treatment and Research, ed. RJ Hager­
man, AC Silverman, pp. 3-68. Baltimore:
Johns Hopkins Univ. Press
Hagerman RJ, Amiri K, Cronister A. 1 99 1 .
Fragile X checklist. Am. J. Med. Genet 38:
283-87
Hagerman RJ, Jackson C, Amiri K, Silverman
AC, OConner R, Sobensky W. 1 992. Girls
with fragile X syndrome: physical and neu­
rocognative status and outcome. Pediatrics
89(3):395-400
Hagerman RJ, Sobensky WE. 1989. Psychopa­
thology in fragile X syndrome. Am. J. Ortho­
psychiatry 59: 142-52
Hagerman RJ, Synhorst DP. 1984. Mitral valve
prolapse and aortic dilitation in fragile X syn­
drome. Am. J. Med. Genet 1 7 : 1 23-31
Hagerman RJ, Van Housen K, Smith ACM,
McGavran L. 1984. Consideration of connec­
tive tissue dysfunction in fragile X syndrome.
Am. J. Med. Genet 17 : 1 1 1-2 1
Hansen RS, Gartler SM, Scott CR, Chen SoH,
Laird CD. 1 992. Methylation analysis of
CGG sites in the CpG island of the human
FMRI gene. Hum. Mol. Genet. 1 (8):571-78
Harley HG, Brook JD, Rundle SA, Crow S,
Reardon W, et a!. 1992. Expansion of an
unstable DNA region and phenotypic varia­
tion in myotonic dystrophy. Nature 355:54546
Harper PS. 1989. Myotonic dystrophy and re­
lated disorders. In Principles and Practice of
Medical Genetics, ed. AEH Emery, DL
Rimoin, pp. 579-97. Philadelphia: Saunders.
894 pp.
Harrison CJ, Jack EM, Allen TD, Harris R.
1983. The fragile X: a scanning electron
microscope study. J. Med. Genet. 20:28085
Heitz D, Rousseau F, Devys D, Saccone S,
Abderrahim H, et a!. 1 99 1 . Isolation of se­
quences that span the fragile X and identifi­
cation of a fragile X-related CpG island.
Science 25 1 : 1 236-39
Hinds HL, Ashley CT, Nelson DL, Warren ST,
Housman DE, Schalling M. 1993. Tissue
specific expression of FMR 1 provides evi­
dence for a functional role in fragile X syn­
drome. Nature Genet. 3:36-43
Hirst MC, Roche A, Flint n, MacKinnon RN,
Bassett JHD, et al. 199 1 . Linear order of new
and established markers around the fragile
site at Xq27.3. Genomics 10:243-49
Homstra IK, Nelson DL, Warren ST, Yang TP.
1993. High resolution methylation analysis
of the FMR 1 gene trinucleotide repeat region
in fragile X syndrome. Hum. Mol. Genet.
2(10) : 1 659-65
Howard-Peebles PN, Friedman JM. 1985. Un-
affected carrier males in families with fragile
X syndrome. Am. J. Hum. Genet. 37:956-64
Hwu W-L, Lee YoM, Lee SoC, Wang T-R.
1993. In vitro DNA methylation inhibits
FMRI promoter. Biochem. Biophys. Res.
Commun. 193:324-29
Imbert G, Kretz C, Johnson K, Mandel J-L.
1993. Origin of the expansion mutation in
myotonic dystrophy. Nature Genet. 4:72-76
Jacobs PA, Bullman H, Macpherson J, Youings
S, Rooney V, et a!. 1993. Population studies
of the fragile X: a molecular approach. J.
Med. Genet. 30:454-5 9
Jacobs PA, Sherman S, Turner G, Webb T.
1986. The fragile (X) syndrome: the mutation
problem. Am. J. Med. Genet 23:61 1-17
Jenkins ED, Brooks J, Duncan CJ, Sanz MM,
Silverman WP, et a!. 1986. Low frequencies
of apparently fragile X chromosomes in nor­
mal control cultures: a possible explanation.
Exp. Cell Bioi. 54:40-48
Knight SJL, Flannery A V, Hirst MC, Campbell
L, Christodoulou Z, et al. 1993. Trinucleotide
repeat amplification and hypermethylation of
a CpG island in FRAXE mental retardation.
Cell 74: 1 27-34
Koide R, Ikeuchi T, Onodera 0, Tanaka H,
Igarashi S, et al. 1 994. Unstable expansion
of CAG repeat in hereditary dentatorubral­
pallidoluysian atrophy (DRPLA). Nature
Genet. 1 :9-13
Kremer EJ, Pritchard M , Lynch M, Yu S , Hol­
man K, et a!. 1991a. Mapping of DNA insta­
bility at the fragile X to a trinucleotide repeat
sequence p(CCG)n. Science 252: 1 7 1 1-14
Kremer EJ, Yu S, Pritchard M, Nagaraja R,
Heitz D, et a!. 1991b. Isolation of a human
DNA sequence which spans the fragile X.
Am. J. Hum. Genet. 49:656-61
La Spada AR, Wilson EM, Lubahn DB, Har­
ding AE, Fischbeck KH. 1 99 1 . Androgen re­
ceptor gene mutations in X-linked spinal and
bulbar muscular atrophy. Nature 352:77-79
Laird CD. 1987. Proposed mechanism of inher­
itance and expression of the human fragile X
syndrome of mental retardation. Genetics
1 1 7:587-99
Loehr JP, Synhorst DP, Wolfe RR, Hagerman
RJ. 1986. Aortic root dilatation and mitral
valve prolapse in the fragile-X syndrome.
Am. J. Med. Genet. 23:1 89-94
Loesch DZ, Huggins R, Hay DA, Gedeon AK,
Mulley Ie, Sutherland GR. 1993. Genotype­
phenotype relationships in fragile X syn­
drome: a family study. Am. J. Hum. Genet.
5 3 : 1 064-73
Lubs JA Jr. 1 969. A marker X chromosome.
Am. J. Hum. Genet. 2 1 : 23 1-44
Mahadevan M, Tsilfidis C, Sabourin L, Shutler
G, Amemiya C, et a!. 1992. Myotonic dys­
trophy mutation: an unstable CTG repeat in
the 3 untranslated region of the gene. Sci­
ence 255 : 1 253-55
FRAGILE X SYNDROME
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
Martin JP, Bell J. 1943. A pedigree of mental
defect showing sex linkage. J. Neurol. Neu­
rosurg. Psychol. 6: 1 54-57
McConkie-Rosell A, Lachiewicz AM, Spiri­
digliozzi GA, Tarleton J, Schoenwald S, et
al. 1 993. Evidence that methylation of the
FMR-1 1ocus is responsible for variable phe­
notype expression of the fragile X syndrome.
Am. J. Hum. Genet. 53:800-9
McKusick V A, Francomano CA, Antonarakis
SE, eds. 1992. Mendelian Inheritance in
Man: Catalogs ofAutosomal Dominant, Au­
tosomal Recessive, and X-linked Phenotypes.
Baltimore: John Hopkins Univ. Press. 2028
pp. 1 0th ed.
Morton NE, Macpherso, IN. 1 992. Population
genetics of the fragile X syndrome: multialle­
lic model for the FMRI locus. Proc. Natl.
Acad. Sci. USA 89:4215-17
Nagafuchi S, Yanagisawa H, Sato K,
Shirayama T, Ohsaki E, et a1. 1 994. Denta­
torubral pallidoluysian atrophy expansion of
an unstable CAG trinucleotide on chromo­
some 12p. Nature Genet. 1 : 1 4- 1 8
Neri G, Chiurazzi P , Arena F , Lubs H A , Glass
IA. 1992. XLMR genes: update 1992. Am. J.
Med. Genet. 43:373-82
Oberle I, Heilig R, Moisan JP, Kloepfer C.
1986. Fragile-X mental retardation syndrome
with two flanking polymorphic DNA mark­
ers. Proc. Natl. Acad. Sci. USA 83: 1 0 1 6-20
Oberle I, Rousseau F, Heitz D, Kretz C, Devys
D, et a1. 1 99 1 . Instability of a 550-base pair
DNA segment and abnormal methylation in
fragile X syndrome. Science 252: 1097-102
Opitz JM. 1986. On the gates of hell and a most
unusual gene. Am. J. Med. Genet. 23:1-10
(erratum 1987. 26:37)
Opitz JM, Sutherland GR. 1984. International
workshop on the fragile X and X-linked men­
tal retardation. Am. J. Med. Genet. 17:5-94
Opitz 1M, Westphal 1, Daniel A. 1984. Discov­
ery of a connective tissue dysplasia in the
Martin-Bell syndrome. Am. J. Med. Genet.
1 7 : 101-9
Orr HT, Chung M, Banfi S, Kwiatkowski TJ.
Servadio A, et al. 1993. Expansion of an
unstable trinucleotide (CAG) repeat in spino­
cerebellar ataxia type 1 . Nature Genet. 4:
221-26
Oudet C, Mornet E, Serre lL, Thomas F.
Lentes-Zengerling SL, et al. 1993. Linkage
disequilibrium between the fragile X muta­
tion and two closely linked CA repeats sug­
gests that fragile X chromosomes are deri ved
from a small number of founder chromo­
somes. Am. J. Hum. Genet. 52:297-304
Pai GS, Sprenkle lA, Do IT, Mareni CE, Mi­
geon BR. 1980. Localization of loci for hy­
poxanthine phosphoribosyltransferase and
glucose-6-phosphate dehydrogenase and bio­
chemical evidence of nonrandom X chromo­
some expression from studies of a human
97
X-autosome translocation. Proc. Natl. Acad.
Sci. USA 77(5):2810-13
Pembrey ME, Winter RM, Davies KE. 1 985. A
premutation that generates a defect at cross­
ing over explains the inheritance pattern of
fragile X mental retardation. Am. J. Med.
Genet. 2 1 :709-17
Penrose LS. 1938. A Clinical and Genetic Study
of 1,280 Cases of Mental Defect. London:
MRC
Pergolizzi RG, Erster SH, Goonewardena P,
Brown WT. 1 992. Detection of full fragile X
mutations by polymerase chain reaction.
lAncet 339:27 1 -72
Pieretti M, Zhang F, Fu YH, Warren ST, Oostra
BA, et ai. 199 1 . Absence of expression of the
FMR-1 gene in fragile X syndrome. Cell 66:
8 1 7-22
Popper CWo 1988. Disorders usually first evi­
dent in infancy, childhood, or adolescence.
In Textbook ofPsychiatry, ed. JA Talbot, RE
Hales, SC Yudofsky, pp. 649-735. Washing­
ton, D.C: Am. Psychiatr.
Reiss AL, Feinstein C, Toomey KE, Goldsmith
B, Rosenbaum K, Caruso MA. 1986. Psychi­
atric disability associated with the fragile X
syndrome. Am. J. Med. Genet. 23:393-401
Reiss AL, Freund L. 1 992. Behavioral pheno­
type of fragile X syndrome. Am. J. Med.
Genet. 43:35-46
Reiss AL, Hagerman Rl, Vingradov S, Abrams
M, King RJ. 1988. Psychiatric disability in
carrier females of the fragile X syndrome.
Arch. Gen. Psychiatr. 45:25-30
Reyniers E, Vits L, De Boulle K, Van Roy B ,
Van Velzen D, e t al. 1993. The full mutation
in the FMR-l gene of male fragile X patients
is absent in their sperm. Nature Genet. 4:
143-46
Richards RI, Holman K, Friend K, Kremer E,
Hillen D, et al. 1 992. Evidence of founder
chromosomes in fragile X syndrome. Nature
Genet. 1 :257-60
Richards RI, Sutherland GR. 1992. Heritable
unstable DNA sequences. Nature Genet. I :
7-9
Riggins Gl, Sherman SL, Oostra BA, Sutcliffe
IS, Feitell D, et al. 1992. Characterization of
a highly polymorphic dinucleotide repeat
150 kb proximal to the fragile X site. Am. J.
Med. Genet. 43:237-43
Rousseau F, Heitz D, Biancalana V, Blu­
menfield S, Kretz C, et al. 1991a. Direct
diagnosis by DNA analysis of the fragile X
syndrome of mental retardation. N. Engl. J.
Med. 325 : 1 673-81
Rousseau F, Vincent A, Rivella S, Heitz D,
Triboli C, et al. 1991b. Four chromosomal
breakpoints and four new probes mark out a
l O-cM region encompassing the fragile-X
locus (FRAXA). Am. J. Hum. Genet. 48:
108-16
Sabourin LA, Mahadevan MS, Narang M, Lee
98
WARREN & ASHLEY
Dse, Surh LC, Komeluk R G. 1993. Effect
of the myotonic dystrophy (DM) mutation on
mRNA levels of the DM gene. Nature Genet.
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
4(3):233-38
Schwartz CEo 1993. Invited editorial: X-linked
mental retardation: in pursuit of a gene map.
Am. J. Hum. Genet. 52: 1025-3 1
Sherman SL, Jacobs PA. Morton NE, Froster­
Iskenius V. Howard-Peebles PN. et al . 1985.
Further segregation analysis of the frag ile X
syndrome with special reference to transmit­
ting males. Hum. Genet. 69:289-99
Sherman SL. Morton NE. Jacobs PA. Turner
G. 1 984. The marker (X) syndrome: a cyto­
genetic and genetic analysis. Ann. Hum.
Genet. 48:21-37
Siomi H, Matunis MJ. Michael WM. Dreyfuss
G. 1 993a. The pre-mRNA bi nding K protein
contains a n ovel evolutionarily conserved
motif. Nud. Acids Res. 21 (5): 1 1 93-98
Siomi H. Si omi MC, Nussbaum RL, Dreyfuss
G. 1993b. The protein product of the fragile
X gene, FMR 1 . has characteristics of an
RNA binding protein. Cell 74:29 1 -98
Smith SD, Kimberling WJ. Pennington SF,
Lubs HA. 1 983. Specific reading disability­
identification of an inherited fonn through
linkage analysis. Science 219:1 345-47
Smits APT, Dreesen JCFM, Post JG, Smeets
DFCM, de Die-Smulders C, et a!. 1993. The
fragile X syndrome: no evidence for any re­
cent mutati ons . l. Med. Genet. 30:94-96
Snow K, Doud LK. Hagennan R, Pergolizzi
RG, Erster SH, Thibodeau SN. 1993. A nal ­
ysis of a CGG sequence at the FMR-l l ocus
in fragile X families and in the general pop­
ulation. Am. J. Hum. Genet. 5 3 : 1 2 1 7-28
Sutcliffe JS. Nelson DL, Zhang F, Pi erelti M ,
Caskey CT, et a ! . 1 992. DNA methyl ation
represses FMR-l transcription in fragile X
syndrome. Hum. Mol. Genet. 1 :397-400
Sutherland GR. 1 977. Fragile sites on human
chromosomes; demonstration of the ir depen­
dence on the type of tissue culture medium.
Science 197:265-66
Sutherland GR, Gedeon A, Komman L,
Donnelly A. Byard RW. et a!. 1 99 1 . Prenatal
diagnosis of fragile X syndro me by direct
detection of the unstable DNA sequence. N.
Engl. J. Med. 325:1 720-22
Sutherland GR. Hecht F, M ulley JC, Glover
TW, Hech t BK. 1985. Fragile Sites 011
Humall Chromosomes. New York: Oxford
Univ. Pre�s. 280 pp.
Suthers GK. Hyland VJ, Callen DF, OberIe I,
Rocchi M, et al. 1 990. Physical mappi ng of
new DNA probes near the fragile X mutation
(FRAXA) by using a panel of ce l l lines. Am.
J. Hum. Genet. 47: 1 87-95
Suthers GK, Mulley JC . Voelckel MA, Dahl N.
Vaisanen ML, et al. 1 991 . Linkage heteroge­
n eity near the fragile X locus in normal and
fragile X families. Genomies 1 0:576--82
Tarleton J, Richie RX, Schwartz C, Rao K,
Aylsworth AS. Lachi ewi cz A. 1 993. An ex­
tensive de novo deletion removing FMRI jn
a patient with mental retardation and the
fragile X syn drome phenotype. Hum. Mol.
Genet. 2{U):1973-74
Taylor A, Safanda IF. Fall MZ, Quince C, Lang
KA, et al. 1 994. Molecular predictors of cog­
n itive involvement in female carriers of frag­
ile X syn drome . lAMA 271 (7):507-14
Verheij e, Bakker CE, deGraff E, Keulemans
J, Wil lemson R, et a1. 1993. Characterization
and localization of the FMR -I gene product
associated wi th fragile X syndrome. NaTUre
363:722-24
Verkerk AJMH, de Graff E, De BoulIe K,
Ei chler EE, Konecki DS, et al. 1993. Alter­
native splicing in the fragile X gene FMRI .
Hum. Mol. Genet. 2(4):399-404
Verkerk AJMH, Pieretti M. Sutcliffe JS, Pu
YH, Kuhl DPA. et al. 1 99 1 . Identification of
a gene (FMR-l) contai n i ng a eGG repeat
coincident with a breakpoint cluster region
exhibiting length variation in fragile X syn­
drome. Cell 65:905-14
Vincent A, Heitz D, Petit C. Kretz C, Oberl I.
Mandel JL. 1 99 1 . Abnormal pattern detected
in fragile-X patients by pulsed-field gel elec­
t rophore sis. Nature 349:624-26
Warren ST. Davidson RL. 1984. Expression of
fragile X chromosome in human-rodent so­
matic cell hybrids, Somal. Cell Mol. Genet.
1 0:409- 1 3
Warren ST, Knight SJL, Peters JF. Stayton CL,
Consalez GG, Zhang F. 1 990. Isolation of the
human chromosomal band Xq28 within so­
mati c cell hybrids by fragile X site cleavage.
Proc. Natl. Acad. Sci. USA 87:3856-60
Warren ST, Nelson DL. 1 993. Trinucleotide
repeat expansions i n neurologic disease .
Curro Opin. Neurobiol. 3:752-59
Warren ST, Nelson DL. 1 994. Advances in mo­
lecular analysis of fragile X syndrome .
lAMA 27 1 :536-42
Warren ST, Zhallg F, Licameli GR, Peters JF.
1987. The fragile X site in somatic cell hy­
brids : an approach for molecular cloning of
fragile sites. Science 237:420-23
Webb TP. Bundey SE, Thake AI, Todd J. 1 986.
Population incidence and segregation ratios
in Martin-Bell syndrome. Am. J. Med. Gellet.
23:573-80
Weber lL. 1 990. Infonnativeness of human
(dC-dA)n • (dG.dT) polymorphisms. Geno­
mies 7 :524-30
Willems PJ, Van Roy B, De Boulle K, Vi ts L,
Reyn icrs E, et a1. 1 992. Segregation of the
fragile X mutation from an affected male to
his normal daughter. Hum. Mol. Gellet. J (7):
5 1 1-17
Wohrle D, Hennig I, Vogel W, Steinbach P.
1 993. Mitotic stability of fragile X mutati ons
in differentiated cells indic ates early post-
FRAGILE X SYNDROME
conceptional trinucleotide repeat expansion.
Nature Genet. 4:140-42
Annu. Rev. Neurosci. 1995.18:77-99. Downloaded from arjournals.annualreviews.org
by EMORY UNIVERSITY on 01/04/10. For personal use only.
Wohrle D, Kotzot D, Hirst Me, Manca A, Korn
B, et al. 1992. A microdeletion of less than
250 kb, including the proximal part of the
FMR-l gene and the fragile-X site, in a male
with the clinical phenotype of fragile-X syn­
drome. Am. 1. Hum. Genet. 5 1 :299-306
99
Yu S , Mulley J, Loesch D, Turner G, Donnelly
A, et al. 1992. Fragile-X syndrome: unique
genetics of the heritable unstable element.
Am. J. Hum. Genet. 50:968-80
Yu S, Pritchard M, Kremer E, Lynch M, Nan­
carrow J, et al. 1 99 1 . Fragile X genotype
characterized by an unstable region of DNA.
Science 252: 1 1 79-81