The School of Molecular and Cellular Biology at the University of Illinois at Urbana-Champaign

The Fragile X Mental Retardation Protein (FMRP) is an mRNA binding protein that is required for normal cognition; its absence causes the most common form of inherited mental retardation, fragile X syndrome. FMRP binds a large collection of mRNAs and regulates their transport and translation. We are interested in understanding how these processes are regulated, which we will determine by addressing the following questions:

1. Does FMRP bind its mRNA cargoes in the nucleus and what are the molecular requirements for this association?
2. Does arginine methylation of the primary RNA binding motif in FMRP (the RGG box) modulate RNA binding and which enzymes mediate methylation?
3. How does FMRP regulate translation of its bound mRNAs through the microRNA pathway?

We published the first evidence that FMRP’s primary RNA binding motif, the RGG box is methylated in cells and that this methylation inhibits RNA binding in vitro (Stetler, et al. 2006). We now have evidence that protein arginine methyl transferase (PRMT) 1 and PRMT3 act on FMRP and that by reducing methylation of FMRP, there is more FMRP on polyribosomes suggesting that more mRNAs are bound and translated (Blackwell and Ceman, manuscript in preparation).

FMRP also functions in the microRNA pathway, associating with its component protein Dicer, as well as with precursor and mature microRNAs. We have evidence that phosphorylation of FMRP inhibits association with Dicer, resulting in an accumulation of precursor microRNAs with FMRP. We suspect that a paucity of mature microRNAs may lead to translational suppression, since microRNAs have shown to also be required for activation of translation (Cheever and Ceman, manuscript under revision).

Taken together, we propose a model in which FMRP enters the nucleus to bind its RNA cargoes and is methylated by PRMT1 to modulate its mRNA binding ability. We also propose that FMRP binds precursor microRNAs in the nucleus. RNA-bound FMRP is exported to the cytoplasm where phosphorylation regulates mRNA translation by modulating association with the nuclease Dicer. Like its autosomal paralog FXR1, we propose that FMRP requires microRNAs to activate translation. Since phosphorylation of FMRP leads to loss of Dicer association and a corresponding increase in precursor microRNAs, we propose that phosphorylation of FMRP leads to translation suppression through loss of activating microRNAs.

Fig.1. FMRP-RNA binding model. FMRP enters the nucleus where RGG box methylation by PRMT1 modulates mRNA binding. FMRP binds nascent transcripts and precursor microRNAs and is exported from the nucleus by Tap/NXF1 and exportin 5, respectively. In the cytoplasm, FMRP is methylated by PRMT3, which prevents further mRNA binding. FMRP associates with Dicer, cleaving precursors to microRNAs that activate translation. Phosphorylation of FMRP inhibits Dicer association.

In addition to understanding the molecular basis of FMRP regulation, we are interested in studying its role in learning and memory. Besides cognitive impairment, fragile X patients have significant speech delays, followed by significant speech and language deficits. To gain insight into the role of FMRP in vocal learning, we are studying FMRP expression in zebra finch – the only tractable lab model for learned vocalizations. Zebra finches, like humans, need to hear the vocalizations of an adult during their development in order to learn their own song properly. We cloned the FMR ortholog from zebra finch brain RNA and developed a specific antibody to it. We found FMRP expressed ubiquitously in the brain, including the regions important for song learning and production, i.e., the song nuclei. More importantly, FMRP was elevated in the premotor nucleus RA during a critical point in sensory-motor learning (Fig.2, right) (Winograd, Clayton and Ceman, manuscript in revision).

Fig. 2. Role of FMRP in the nucleus RA of zebra finch. (Left schematic) FMRP binds mRNAs and transports them to the dendrite where translation is activated by stimulation of metabotropic glutamate receptor 5. (Center schematic) Depiction of axonal presynaptic input from either song nuclei HVC or LMAN on the postsynaptic dendrite in RA containing FMRP (yellow). (Right) FMRP (red) is elevated in the neuropil of RA (immunofluorescent staining above the arrows) in neurons (green immunofluorescence staining) at a critical age during sensory-motor development.

To determine if FMRP indeed plays a role in sensorimotor learning, we will knock it down in the nucleus RA by delivering lentiviral encoded small hairpin RNAs by stereotactic injections. We have identified a small silencing hairpin RNA that reduces FMRP expression 85% in transfected cell. Using this information, we are now in the process of developing the lentiviral silencing hairpin RNAs and controls. We will deliver these stereotactically into the nucleus RA. Upon creating a TguFMRP-knockdown finch, we will examine his ability to learn song, which can be quantified by comparing the engineered bird’s mature song to the tutor’s song. We expect that loss of FMRP from the premotor nucleus RA will affect the ability to learn song. Such a bird will provide unique insights into the role of FMRP in sensory motor learning.