Numerous pedigrees reflecting this typical X-linked pattern of inheritance of MR have been described. Since patients seldom exhibit an obvious phenotype other than MR, they were grouped together as having nonspecific X-linked MR. Their condition, it was believed, was due to any number of X-linked mutations, each relatively rare. Work beginning in the early 1970s began to elucidate a single mutation among many families with nonspecific X-linked MR. This mutation causes what is now known as fragile X syndrome. It accounts for almost half of all nonspecific X-linked MR and, indeed, is the commonest form of inherited MR. Fragile X is named for a gap, observed on the X chromosomes of affected patients, that appears fragile when viewed under the microscope. Over the past decade our laboratory has been studying fragile X syndrome and in 1991, working with an international group of collaborators, we identified the FMR1 (fragile X mental retardation 1) gene, which is responsible for this syndrome.

Within the 5' untranslated portion of the FMR1 mRNA is an unusual tract of the repeated triplet CGG. Among normal individuals, this triplet repeat is polymorphic in length and content, with 7-52 triplets (mean of 30), often containing 1-3 AGG interruptions of the CGG array. Individuals with fragile X syndrome exhibit a massive expansion of this repeat beyond 230 triplets, usually exceeding 700 repeats. Unaffected, carrier males and many carrier females exhibit repeat lengths intermediate between normal and affected. These intermediate-length alleles are unstable when transmitted, tending to increase slightly in length. This form of mutation, due to unstable trinucleotide repeats, had not been previously observed in any other genetically studied organism. Since the repeat expansion in fragile X syndrome was discovered in 1991, over a dozen other genetic diseases have been attributed to this novel form of mutation.

The mechanism behind the repeat expansion for these disorders remains unknown. However, in fragile X syndrome some clues have emerged. Polymorphic markers within and near FMR1 can describe the regional "signature" of distinctive normal X chromosomes haplotypes. We and others have shown that certain haplotypes are more frequently found among fragile X chromosomes than would be expected based on the haplotype occurrence on normal X chromosomes. This is referred to as linkage disequilibrium and suggests that fragile X syndrome occurs only in a limited number of X chromosomes in the population.

When the CGG repeat is longer than approximately 230 triplets, sequences of and surrounding the repeat are concomitantly methylated. This abnormal methylation indirectly attracts the enzyme histone deacetylase, which alters the chromatin conformation of the FMR1 gene. The result is the transcriptional silencing of FMR1, and the absence of FMR1 protein (FMRP) is now accepted as the basis for the phenotype.

Much of our current work has focused upon the biochemical and neurobiological consequence of the loss of FMRP. We have previously shown that FMRP contains sequence attributes of RNA-binding proteins and indeed interacts with a subset of brain mRNA, even as a purified protein. We have also identified two additional functional domains within FMRPinvolving nuclear import and export of FMRPand have shown FMRP shuttles between the nucleus and the cytoplasm and incorporates into mRNP particles composed of other proteins and selective RNA molecules. These particles are associated with ribosomes in the cytoplasm of various cell types, including the somatodendritic compartments of neurons.

The importance of the association of FMRP with ribosomes is highlighted by our studies of an atypical patient with a mutation that changes an amino acid within FMRP. The mutant protein in this patient, who has unusually severe fragile X syndrome, binds RNA but no longer associates with ribosomes. We speculate that the mutant protein is sequestering the bound mRNAs, preventing even minimal translation.

The next major challenge in this research will be to identify those mRNAs that interact with FMRP and examine the consequence of FMRP loss on their translated proteins. We have identified several interacting messages by using mRNA isolated from the immunoprecipitated mouse brain Fmrp-containing mRNP complex, and interrogating some 40,000 genes on microarrays. Moreover, many of these same messages show an altered translational profile, which is the abundance of a message off or on polyribosomes, in the absence of Fmrp. These data are consistent with the model that the absence of FMRP may miscue translation of those messages normally bound to FMRP. In neurons, this may influence synaptic plasticity, as reflected by the delayed dendritic spine maturation we and others have observed in the Fmr1 knockout mouse. Understanding the function of these proteins could provide considerable insight into the pathophysiology of this disorder and may well illuminate molecular mechanisms important to human intelligence.