Studies have revealed a significant decrease in the expression of FMR1 mRNA in fragile X cases as a result of the hypermethylation of the CpG island (35).  An important feature of repressing of this particular gene is the presence of FMR1in fetal tissue.  The varying levels of expression between normal and fragile X fetal tissue reflects a very important role in development.  This data also suggests early acquisition of the methylation, probably during embryogenesis (38,40).

Early acquisition of methylation points out the importance of the loss of FMR1 expression.   The human FMR1 gene, being highly conserved among species (16), has been shown to display alternative splicing properties suggesting the presence of many isoforms (4).  Mouse models have been developed to study fragile X syndrome due to the allelic similarities (16).  Three different levels of repeat within the gene have been reported: normal (7-50 repeats), permutation (50-200 repeats), and full mutation (> 200 repeats).  The class of permutation is interesting because it only becomes affective in successive generations when the repeat region has a chance to expand leading to Shermans’s paradox.  The permutation allele is unstable when transmitted from generation to generation. 

All males with the full mutation display the fragile X phenotype where only approximately 50% of females with the mutation display the fragile X phenotype (4, 40).  Taken together with data from our more recent studies this proves that the germ line is susceptible to full expansion (32).  There seem to be no new mutations appearing to cause fragile X syndrome. 

The only known mutation of resulting in fragile X syndrome interestingly shows increases in the transcription of FMR1 have been proportionally linked to the CGG tri-nucleotide repeat length and the repeat number approaches 200 in permutations.  As the number of repeats increases the number of FMR1 mRNA levels increase.  Unexpectedly, the FMR1 protein products decrease in relation to the CGG repeat length in permutation (25).  This finding shows the development of the disease at various points of repeat length during the permutation stage of fragile X syndrome.    

Upon extending from the permutation to the full mutation, the FMR1 gene is silenced.  The silencing of the FMR1 gene is interesting because it remains unclear as to how the tri-nucleotide repeat expansion occurs in an organism.  Moreover, questions still remain as to how the loss of the FMR1 protein, FMRP can have such broad effects as those observed in fragile X syndrome (21, 40).  Studies have started to reveal the role and importance of FMRP in patient and mouse models uncovering mRNA binding properties (5,29,17). 

Within the past few years we have uncovered information linking FMRP to protein synthesis (30) and a micro-RNA dependent translational suppression pathway (23).  In this review we will focus on explaining the neuronal functions and importance of FMRP in model systems.  We will also propose a mechanism for both translational and transcriptional repression through a micro-RNA pathway.

Functions of FMRP  

In fetal tissue, FMRP is found to be most abundantly and universally expressed in the testis and the brain (21).  The differentiation of neuronal stems cells related to the fragile X phenotype has shown alteration in fragile X patients linking the disease to development (10). FMRP was discovered to assert properties of RNA-binding proteins in areas of expression (40).  The protein itself encompasses three RNA binding domains, two K domains and an RGG box (40, 15).  Importantly, the FMRP protein maintains both a nuclear location signal (NLS) at the amino terminus end of the FMRP and a nuclear export signal (NES) encoded by exon 14 of the FMRP.  These signals help to confirm RNA binding properties of FMRP by hinting at the possible translocation of various mRNAs via the FMRP pathway (17).

In order to bind mRNAs, FMRP forms an RNP complex containing FMRP, FRX1P, FRX2P, nucleolin, and three other proteins.  The particles making up human RNPs are conserved in mice as well. (11). This RNP binds mRNA selectively by associating the G-quartet structure of the mRNA to the RGG box of the FMRP (5, 15).  The G-quartet structure acts as a target for FMRP explaining the selective binding properties and the importance of transcriptional regulation via FMRP of the targeted mRNAs (15).  FMRP binds directly to mRNAs via this interaction (9). 

This direct interaction with mRNA allows FMRP to act as a translational control factor for many of these targeted mRNAs.  Evidence for this was found because a large majority of FMRP is associated with ribosomes, translational machinery (38).  Because the absence of the FMR1 gene was associated with fragile X syndrome, a functional role of FMRP was investigated (38).  FMRP was found to be associated with ribosomes in the dendritic structures of neurons elucidating a possible role of the proteins related to dendritic structure and neuronal plasticity (18). The association of FRMP with polyribosomes was eliminated in I304N mutations of extreme fragile X syndrome (19). The tissue specific expression of FMR1 connects the protein to important developmental structures in the brain such as the hippocampus (21).  A set of FMR1 knockout mouse models were used to understand the relationship between drendritic formation and FMRP. Knock out mice were found to posses irregular dendritic spines (14).  The importance of altered neuronal formation is discussed later.

The next important link connecting neuronal alterations and the function of FMRP is the evidence pointing to negative regulation of mRNA translation.  By removing the binding site of FMRP we found a lack of translational inhibition (30, 31).  Specifically, FMRP has been shown to regulate the production of MAP1B, a protein.  In Futsch mutant mice show inverse regulation of MAP1B and altered synaptic growth (31, 45).  This shows that FMRP possibly plays a major role in development of neuronal structure throughout the body.

mRNA ligands associated with FMRP

Because FMRP is shown to play significant roles in both mRNA transportation and translation, the characteristics of the specific mRNAs associated with FMRP have been explored.  A variety of mRNAs are associated with FRMP.  One of our recent studies examined 13 potential candidate mRNAs associated with FMRP and found that at least ten encode proteins involved in synaptic plasticity or neuronal development.  This data helped to identify some of the specific mRNA translation that is affected in mental retardation (15, 33).  One specific mRNA was the mRNA coding for MAP1B.  MAP1B is negatively regulated in the Drosophila nervous system.  In the absence of FMRP, MAP1B experiences no inhibition and altered dendrite and axon development is noted (45).  

We also interestingly found that FMRP associates with the non-translatable BC1 RNA.  In addition to binding directly to FMRP, BC1 can associate with the FMRP target mRNAs in the absence of FMRP (43, 44, 45).  This data suggests that BC1 is involved in specificity of FMRP to mRNAs and BC1 helps to inhibit translation of some mRNAs by blocking the initiation codon (43, 44).  Large numbers of mRNAs have been recently associated with FMRP.  Of these mRNAs, many have been found to differ in expression and distribution among wild type and FMR1 knockout mice (31).  These recent studies identifying specific mRNAs associated with FMRP and their coded proteins have shown the importance of FMRP in neuronal development.

Altered neuronal development in Fragile X Syndrome

FMRP regulates the translation of many mRNAs which code for proteins involved in neuronal development.  FMRP has been specifically found to be highly expressed in neurons.  Areas controlling cognition are commonly affected in fragile X syndrome.  A mouse model study shows deficient amygdale and hippocampal functions in FMR1 knockout mice during fear and conditioning tests (36).  Specifically, FMRP is highly expressed in the dendritic formations of non-fragile X organisms (18).  The dendritic functions encoded by the target mRNAs of FMRP are deficient in fragile X patients thus leading to altered synaptic function as observed in the fragile X  phenotype (18).

Further proving the neuronal effects of FMR1 silencing mouse models have elucidated interesting findings in the area of altered synaptic plasticity related to fragile X mental retardation (22).  Long term depression (LTD) dependent on glutamate receptors was found to be significantly altered in the hippocampus cells of knockout mice.  By using DHPG to induce the glutamate dependent LTD, enhanced results were found in the knockout mice (22).  This supports the earlier hypothesis that FMRP is important in regulating protein production in the synapse.  

In mice lacking the expression of the FMR1 gene, irregular dendritic spines were observed (26) connecting the FMRP protein with synaptic growth.  Dendritic spines in the visual cortex were compared among FMR1 knock out mice and wild type mice to explore the specific differences in dendritic growth.  The dendrites of the knockout mice show a high incidence of long thin dendritic spines as well as dendritic spines of higher density (13).  This data combined with the mRNAs targeted by FMRP that regulate neuronal structure connect fragile X syndrome with altered synaptic plasticity.

FMRP influences synaptic growth by normally inhibiting translation of functional mRNAs in synaptosomes (18).  Two methods of the release of inhibition in normal models have been discovered:  the first is dephosphorylation (12) and the second is excitation by glutamate receptors (3, 42).  FMRP associated with stalled polyribosomes was consistently phosphorylated where the FMRP associated with actively translating polyribosomes was consistently dephorphorylated in both brian cells and cultured cells (12).  This is possibly the mechanism by which FMRP releases its inhibitory effect on targeted mRNAs. 

Recent models have associated the release of translational inhibition with the presence of mRNA granules and mGluR-induced translation (3).  FMRP normally inhibits mRNA translation, but induction of translation has been associated with glutamate receptors.  In fragile X syndrome mGluR-induced translation is heightened due to the lack of inhibition normally provided by FMRP.  The decrease in mRNA granules in knockout mice supports this finding.  Furthermore, upon excitation with an mGluR5 agonist increased granule levels were observed (3).  The lack of rapid protein translation in sites mediated by neurotransmitters such as glutamate has been proposed to cause several abnormalities observed in fragile X syndrome (41).  

The micro-RNA silencing of FMR1

The FMR1 gene is silenced by methylation induced by the CGG tri-nucleotide repeat expansion alone, but also by interaction with micro-RNAs (23, 24).  FMRP interacts with Argonaute  (AGO; 33) and components of the micro-RNA pathway such as Dicer (33, 24).  In Drosophila melanogaster models the AGO ortholog was found to be suppressed in the presence of FMRP.  When FMRP was removed the models experienced a loss of AGO suppression leading to a rough eye phenotype.  Upon induced suppression of AGO, the rough eye phenotype was significantly reduced to levels of almost normal (24).  AGO was found to be important to the biological functions of synapses, but not totally dependent on FMRP.

FMRP interacts with many molecules like AGO to influence translation.  FMRP also has recently been associated with translational regulation through a micro-RNA pathway.  Transcripts are produced from the expanded fragile X allele at some point early in development before complete methylation of the CpG island (23, 24, 34).  These transcripts form structures referred to as hairpins that are cleaved by the enzyme Dicer (23) resulting in small mRNAs approximately 20 nucleotides in length.  These small mRNAs communicate with the RITS complex, a transcriptional silencer of genes.  The small RNAs direct the RITS complex to homologous mRNAs through complementary base pairing.  The RITS complex recruits methylation machinery eventually leading to the suppression of the FMR1 gene as observed in fragile X syndrome (24).  The RITS complex also may change the acetylation of specific histones (7, 24).

The FMR1 promoting region has shown to possess qualities of varying chromatin conformations linked to altered histones (20).  When FMR1 is inactive as in fragile X cells the chromatin was displayed uniform conformation.  In normal cells the interactions are far lesser than in fragile X cells (20).  Histones acetylation is associated with normally expressed FMR1, but not the absence of FMR1 as in fragile X syndrome (7, 8, 20).  Recent studies have treated fragile X cells with 5-aza-2-deoxycytidine and observed re-acetylation of histones 3 and 4.  The re-acetlyation of these histones resulted in FMR1 transcriptional reactivation (8).  This finding presents an alley of treatment investigation for fragile X patients in the future.

Discussion

FMRP has shown tissue specific expression in areas of the brain and other parts of the body associated with the observed phenotype of fragile X syndrome such as the hippocampus and the testis.  Specific binding of FMRP to various mRNAs has been elucidated also linking FMRP to an inhibitory role in the translation of respective mRNAs.  Translation inhibition and mRNA targeting may be influence by non-translatable mRNAs such as BC1.  The silencing of the FMR1 gene is the central link in fragile X syndrome, however, recent studies have shown gene silencing by the binding of transcription factors to the promoter regions of FMR1 (28).  This is data another example of the mechanism by which methylation silences the FMR1 gene.

The specific mRNAs associated with FMRP have proven to reveal a wealth of information explaining the gap of knowledge between the silencing of the FMR1 gene and the fragile X phenotype.  MAP1B is an example of one such mRNA.  MAP1B encodes for microtubule structural functions, thereby influencing physical features in the development of neurons.  The absence of FMRPs regulatory influence leads to severely altered neurons in fragile X syndrome.  Figure 2 proposes an over all view of the FMRP interactions in a normal neuron.