rather than risk of penetrance, increases through 
generations. As discussed below, continued investi- 
gation of the fragile X mutation by Dr. Warren and 
his collaborators has demonstrated the molecular 
basis of the Sherman paradox and begun to illumi- 
nate the molecular mechanism (s) of the disease. 
Within the 5'-untranslated region of the FMR-1 
transcript is a CGG repeat that is normally polymor- 
phic (mean length of 29). In penetrant patients 
with fragile X syndrome, this repeat is massively ex- 
panded beyond 200 repeats, often exceeding 1 ,000 
triplets. When the repeat length exceeds approxi- 
mately 230 triplets, the FMR-1 gene, including its 
promoter and the repeat itself, becomes heavily 
methylated. This methylation leads to transcrip- 
tional silencing of the gene. The subsequent lack of 
FMR-1 gene product, whose function is currently 
unknown, presumably leads to the phenotype asso- 
ciated with the fragile X syndrome (mental retarda- 
tion with characteristic facies and macroorchidism 
in affected males) . 
Carrier individuals, who themselves are unaf- 
fected, have intermediate levels of CGG-repeat ex- 
pansion from 52 to 200 triplets, and alleles of this 
size in fragile X pedigrees exhibit marked meiotic 
instability. Each time the fragile X chromosome 
transmits from carrier to offspring, it changes to a 
repeat length different from the parental chromo- 
some (usually larger) and distinct from that of sib- 
lings. Thus the mutation rate of the fragile X allele 
is 1 . Female carriers, but significantly not male car- 
riers, have a chance with each meiosis of transmit- 
ting not just a slightly different allele size to a child, 
but rather a massively expanded repeat allele lead- 
ing to the fragile X syndrome. 
Dr. Warren and Dr. David L. Nelson (Baylor Col- 
lege of Medicine) with Dr. C. Thomas Caskey 
(HHMI, Baylor College of Medicine) established 
that the repeat length of the fragile X chromosome 
of carrier mothers influences this probability of 
massive expansion, such that a carrier with 60 re- 
peats has approximately an 8% chance of having an 
affected son, while another carrier mother with 1 00 
repeats has a 50% risk. Since NTMs (who always 
have less than 200 repeats) have mothers with less 
than 90 repeats, their siblings are less likely to be 
affected. As the mutant chromosome is transmitted, 
sometimes for a number of generations, it can con- 
tinually increase in size, such that descendant fe- 
males have an increasing risk of having an affected 
child. These data therefore resolve the Sherman par- 
adox and provide a molecular basis for the reduced 
penetrance in this disorder. 
To determine the normal function of the FMR-1 
gene product and deduce how its deficiency leads to 
fragile X syndrome. Dr. Warren's laboratory has 
cloned and sequenced the entire gene from the 
mouse. In turn, the murine /mr-i cDNA was used to 
identify both the yeast and Caenorhabditis elegans 
homologues of this highly conserved gene. The mu- 
rine gene has 95% amino acid identity to the human. 
It contains a CGG repeat in the 5'-untranslated re- 
gion, although smaller than human, with nine trip- 
lets. The gene is expressed in similar tissues in both 
human and mouse, primarily the brain and testes, as 
expected from the clinical picture. 
Working with Dr. David Housman (Massachusetts 
Institute of Technology), Dr. Warren and his col- 
leagues conducted in situ hybridization studies in 
various mouse developmental stages. Widespread 
expression of fmr- 1 was found throughout fetal de- 
velopment. Analysis of adult brain localized fmr-1 
expression to the cerebellum and hippocampus, 
with intermediate expression in the cerebral cortex 
selectively exhibited in neuronal rather than glial 
cells. Examination of testes was suggestive of Sertoli 
cell expression. 
Soon after the discovery of the fragile X mutation, 
two other disorders, myotonic dystrophy and spinal- 
bulbar muscular atrophy, were similarly found to be 
due to expansions of transcribed triplet repeats. To 
identify other such loci and determine the extent of 
such genes, Dr. Warren's laboratory screened hu- 
man cDNA libraries with repeat probes and searched 
databases for cDNAs with triplet repeats. Forty such 
genes were identified, and 14 have been character- 
ized to date. Five were found to contain transcribed 
triplet repeats, which are normally highly polymor- 
phic, a characteristic of the three-repeat expansion 
disorders. Four of these genes were similar to 
FMR-1, containing CGG repeats in 5'-untranslated 
regions. Several were of previously determined 
function, and examination of potentially related dis- 
orders is ongoing. 
Emery-Dreifuss Muscular Dystrophy 
Dr. Warren's laboratory has continued the refine- 
ment of the map position of Emery-Dreifuss muscu- 
lar dystrophy (EDMD), an X-linked dystrophy dis- 
tinct from the Duchenne or Becker forms. Using an 
exceptionally large family in north Georgia, more 
than 100 meioses have been analyzed for linkage. 
When compiled with other data, the results place 
the EDMD locus between the Xq28 loci, encoding 
the color vision genes, and coagulation factor 8. 
In collaboration with Dr. Jed Gorlin (Harvard 
Medical School), the gene for actin-binding protein 
280 (ABP 280) was placed within this map interval 
GENETICS 281 
