by
Namat-Maria Majaess
B.Sc. (Hons), University of Victoria, 2010 A Thesis Submitted in Partial Fulfillment of the
Requirements for the Degree of MASTER OF SCIENCE
in the Department of Biochemistry and Microbiology
Namat-Maria Majaess, 2012 University of Victoria
All rights reserved. This thesis may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author.
Supervisory Committee
A Loss of the Fragile X Mental Retardation Protein Alters the Spatial and Temporal Expression of Glutamate Receptors in the Mouse Brain
by
Namat-Maria Majaess
B.Sc. (Hons), University of Victoria, 2010
Supervisory Committee
Dr. Brian R. Christie (Division of Medical Sciences; Department of Biology) Co-Supervisor
Dr. Perry L. Howard (Department of Biochemistry and Microbiology; Department of Biology) Co-Supervisor
Dr. Robert D. Burke (Department of Biochemistry and Microbiology) Departmental Member
Dr. Leigh Anne Swayne (Division of Medical Sciences; Department of Biology) Outside Member
Abstract
Supervisory CommitteeDr. Brian R. Christie (Division of Medical Sciences; Department of Biology) Co-Supervisor
Dr. Perry L. Howard (Department of Biochemistry and Microbiology; Department of Biology) Co-Supervisor
Dr. Robert D. Burke (Department of Biochemistry and Microbiology) Departmental Member
Dr. Leigh Anne Swayne (Division of Medical Sciences; Department of Biology) Outside Member
Fragile X Syndrome (FXS) is the leading cause of inherited intellectual disability. The disorder is caused by a trinucleotide expansion that silences the Fragile X Mental Retardation 1 (Fmr1) gene resulting in the loss of its protein product, the Fragile X Mental Retardation Protein (FMRP). FXS patients show broad clinical phenotypes including intellectual disability, as well as a number of cognitive and behavioral problems. The lack of FMRP is believed to be the direct cause of the deficits seen in FXS patients.
FMRP is an RNA-binding protein that is expressed in the brain and testes. This protein is believed to form a messenger ribonucleoprotein complex with mRNAs in the nucleus and
subsequently export them to polyribosomes in the cytoplasm, therefore influencing translation of its bound mRNAs. Importantly, FMRP has long been suspected to be involved in synaptic plasticity due to its ability to bind several mRNAs that encode for proteins important in synaptic plasticity. Such proteins include the GluN1, GluN2A and GluN2B subunits of the N-methyl-D-aspartate receptor (NMDAR).
FMRP is expressed in the hippocampus, a region of the brain involved in learning and memory processes. Recently, impaired NMDAR functioning in the dentate gyrus (DG)
subregion of the hippocampus has been observed in Fmr1 knockout (-/y) mice. This impairment also resulted in reduction in long-term potentiation (LTP) and long-term depression (LTD) of synaptic efficacy, two biological models of learning and memory. In the present study, I focused on the levels of the NMDAR GluN1, GluN2B and Glu2B subunits in order to determine the synaptic plasticity alterations seen in the DG of Fmr1-/y mice. Using Western blotting, I found
that there is a decrease in the GluN1, GluN2A and GluN2B subunits in the DG of young adult
Fmr1-/y mice, indicating that these mice have significantly lower amounts of total NMDARs.
These results could explain the altered LTP and LTD seen in Fmr1-/y mice at the molecular level
and might contribute to the intellectual impairments seen in these KO mice.
NMDARs appear to be important in the development and maturation of synapses. The GluN2A and GluN2B subunits are developmentally regulated, where GluN2B is predominantly expressed early in development and GluN2A in the adult brain. A dysregulation of GluN2A and GluN2B subunits has been proposed to affect the maturation and formation of synapses.
Intriguingly, FMRP is also believed to play a functional role in early brain development. Thus, this study also focused on the developmental expression of the GluN1, GluN2A and GluN2B subunits in the DG, Cornu Ammonis, prefrontal cortex and cerebellum of Fmr1-/y mice, all of which are brain regions implicated in FXS. We found that the developmental expression of these subunits is altered in Fmr1-/y mice in specific brain regions.
Together, these results demonstrate that the loss of FMRP differentially affects GluN1, GluN2A and GluN2B subunit expression both developmentally and spatially, further implicating NMDARs in the pathophysiology of FXS.
Table of Contents
Supervisory Committee ... ii
Abstract ... iii
Table of Contents ... v
List of Tables ... vii
List of Figures ... viii
List of Abbreviations ... ix Acknowledgements ... xiii Dedication ... xiv Chapter 1: Introduction ... 1 1.1 Fragile X Syndrome ... 1 1.1.1 History ... 1 1.1.2 Etiology ... 2 1.1.3 Manifestations ... 4
1.1.3.1 Development and Behavior ... 5
1.1.3.2 Physical features ... 6
1.1.3.3 Premutation ... 7
1.1.4 Neuroimaging Studies ... 8
1.2 The Fmr1 gene and FMRP ... 10
1.2.1 Structure ... 10 1.2.2 Homology ... 12 1.2.3 Expression ... 13 1.2.4 Function ... 14 1.3 Fmr1 knockout mouse ... 15 1.4 Synaptic Plasticity ... 16 1.4.1 NMDA receptor-dependent LTP ... 17 1.4.2 NMDA receptor-dependent LTD ... 18 1.4.3 mGluR-dependent LTD ... 19 1.5 NMDA receptors ... 20 1.5.1 Structure ... 20 1.5.2 Developmental expression ... 22 1.5.2.1 mRNA expression ... 22 1.5.2.2 Protein expression ... 23
1.6 Synaptic Plasticity alterations in FXS ... 24
1.6.1 The mGluR theory of FXS ... 24
1.6.2 Involvement of NMDARs in FXS ... 25
1.8 Purpose of this Thesis ... 26
1.9 Objectives and Hypotheses ... 27
Chapter 2: Materials and Methods ... 29
2.1 Animals ... 29
2.2 Genotyping ... 29
2.2.1 Tissue Digestion, Genomic DNA Extraction and Purification ... 29
2.2.2 PCR amplification ... 30
2.3 Biochemical Analysis ... 31
2.3.1 Tissue Preparation ... 31
2.3.2 Protein Quantification ... 32
2.3.3 Western blot analysis ... 33
2.3.4 Statistical Analysis ... 34
Chapter 3: Results ... 37
3.1 NMDAR subunit levels in the DG and CA of 2-4 month old Fmr1-/y mice ... 37
3.2 AMPAR phosphorylation states in the DG and CA of 2-4 month old Fmr1-/y mice . 40 3.3 Temporal and Spatial Expression of the GluN1, GluN2A and GluN2B subunits of the NMDAR in Fmr1-/y mice ... 42
3.3.1 Spatial expression of NMDAR subunits in P10 Fmr1-/y mice ... 42
3.3.2 Spatial expression of NMDAR subunits in P21 Fmr1-/y mice ... 47
3.3.3 Spatial expression of NMDAR subunits in 2-4 month old Fmr1-/y mice ... 51
3.3.4 Temporal Expression of GluN1, GluN2A and GluN2B in WT and Fmr1-/y mice ... 55
Chapter 4: Discussion ... 59
4.1 NMDAR Levels are Decreased in the DG of Fmr1-/y mice ... 59
4.2 Expression of the GluN1, GluN2A and GluN2B Subunits of the NMDAR in P10, P21 and 2-4 month old Fmr1-/y mice ... 63
4.2.1 P10 mice ... 63
4.2.2 P21 mice ... 64
4.2.3 2-4 month old mice ... 65
4.2.4 Developmental expression of GluN1, GluN2A and GluN2B subunits in Fmr1-/y mice ... ... 66
4.2.4.1 Temporal Expression Pattern in the DG ... 66
4.2.4.2 Temporal Expression Pattern in the CA ... 68
4.2.4.3 Temporal Expression Pattern in the PFC ... 69
4.2.4.4 Temporal Expression Pattern in the CB ... 70
Chapter 5: Conclusions and Future Research ... 72
5.1 Conclusions ... 72
5.2 Future Directions ... 73
References ... 75
List of Tables
Table 2.1 Summary of western blotting conditions used for 2-4 month old samples... 36 Table 2.2 Summary of western blotting conditions used for P10 and P21 samples... 36
List of Figures
Figure 1.1 The molecular and clinical effects of CGG repeat expansion on the Fmr1 gene and
FMRP expression. ... 4
Figure 1.2 Physical phenotypes present in FXS individuals. ... 7
Figure 1.3 Structure of the Fmr1 gene and FMRP. ... 12
Figure 1.4 Molecular Mechanisms Underlying NMDAR-dependent LTP. ... 18
Figure 1.5 Molecular Mechanisms Underlying NMDAR-dependent LTD. ... 19
Figure 1.6 NMDAR Structure. ... 22
Figure 2.1 Brain Regions of Interest for Microdissections. ... 32
Figure 3.1 Levels of GluN1 are decreased in the DG of 2-4 month old Fmr1-/y mice. ... 38
Figure 3.2 Levels of GluN2A and GluN2B are decreased in the DG of 2-4 month old Fmr1-/y mice... 39
Figure 3.3 Decreased total levels of the GluA1 subunit of the AMPAR in the CA and of phosphorylated S831 GluA1 in the DG of 2-4 month old Fmr1-/y mice. ... 42
Figure 3.4 Levels of GluN1 are decreased in the CA of P10 Fmr1-/y mice. ... 44
Figure 3.5 Levels of GluN2A are normal in the DG, CA, PFC and CB of P10 Fmr1-/y mice. .... 45
Figure 3.6 The levels of GluN2B are increased in the DG of P10 Fmr1-/y mice. ... 46
Figure 3.7 The levels of GluN1 are normal in the DG, CA, PFC and CB of P21 Fmr1-/y mice. . 48
Figure 3.8 Levels of GluN2A are normal in the DG, CA, PFC and CB of P21 Fmr1-/y mice. .... 49
Figure 3.9 Levels of GluN2B are increased in the CA of P21 Fmr1-/y mice. ... 50
Figure 3.10 Levels of GluN1 are decreased in the DG, PFC and CB of 2-4 month old Fmr1-/y mice. ... 52
Figure 3.11 Levels of GluN2A are decreased in the DG and PFC of 2-4 month old Fmr1-/y mice. ... 53
Figure 3.12 Levels of GluN2B are decreased in the DG of 2-4 month old Fmr1-/y mice. ... 54
Figure 3.13 Temporal pattern of expression of GluN1, GluN2A and GluN2B subunits in the DG and CA of Fmr1-/y and WT mice. ... 57
Figure 3.14 Temporal pattern of expression of GluN1, GluN2A and GluN2B subunits in the PFC and CB of Fmr1-/y and WT mice. ... 58
Figure 4.1 Mechanistic basis for the altered LTP seen in the DG of Fmr1-/y mice. ... 62
List of Abbreviations
ADHD Attention-deficit hyperactivity disorder AGE Agarose gel electrophoresis
AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid ASD Autism spectrum disorders
BCA bicinchoninic acid
bp base pairs
BSA bovine serum albumin
CA Cornu Ammonis
CaMKIIα Ca2+/calmodulin-dependent protein kinase II alpha
CB Cerebellum
CGG Cytidine-guanosine-guanosine CpG Cytidine-phosphate-guanosine
CYFIP1 Cytoplasmic FMRP Interaction Proteins 1 CYFIP2 Cytoplasmic FMRP Interaction Proteins 2
DAG Diacylglycerol
DG Dentate gyrus
dNTP deoxyribonucleotide triphosphate
DTT Dithiothreitol
E Embryonic day
ECL Enhanced chemiluminescence EDTA Ethylenediaminetetraacetic acid ELISA Enzyme-linked immunosorbent assay
ER Endoplasmic reticulum
FXS Fragile X syndrome
Fmr1 Fragile X Mental Retardation 1
Fmr1-/y Fmr1 knockout
FMRP Fragile X Mental Retardation Protein
FXR1P Fragile X Mental Retardation Syndrome-Related Protein 1 FXR2P Fragile X Mental Retardation Syndrome-Related Protein 2 FXTAS Fragile X-associated tremor/ataxia syndrome
GluA1 AMPA receptor subunit 1 GluA2 AMPA receptor subunit 2 GluN1 NMDA receptor subunit 1 GluN2A NMDA receptor subunit 2A GluN2B NMDA receptor subunit 2B GluN2C NMDA receptor subunit 2C GluN2D NMDA receptor subunit 2D GluN3 NMDA receptor subunit 3 G-quartet guanine-quartet HRP Horseradish peroxidase kDa kilodalton KH K Homology KH1 KH domain, type 1 KH2 KH domain, type 2 I304N isoleucine-304-asparagine iGluR ionotropic glutamate receptor
IQ Intelligence quotient
IP3 Inositol-1,4,5-triphosphate LTP Long-term potentiation
LTD Long-term depression
MAP1B Microtubule associated protein 1B mRNP Messenger ribonucleoprotein mGluR Metabotropic glutamate receptor MRI Magnetic resonance imaging mRNA messenger ribonucleic acid NES Nuclear export signal NLS Nuclear localization signal NMDA N-methyl-D-aspartic acid
NMDAR N-methyl-D-aspartic acid receptor
NTM Normal transmitting male p831 Phosphorylated serine 831 p845 Phosphorylated seine 845
P Postnatal day
PBS Phosphate-buffered saline PCR Polymerase chain reaction
PFC Prefrontal cortex
PKC Protein kinase C
POI Primary ovarian insufficiency
PP Protein phophatase
PSD-95 Post-synaptic density protein 95 PVDF Polyvinylidene fluoride
R138Q arginine-138-glutamine RCF relative centrifugal force RGG arginine-glycine-glycine RPM revolutions per minute
S831 serine at residue 831 S845 serine at residue 845 SEM standard error of the mean SDS sodium dodecyl sulfate
SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis STEP Striatal-enriched protein tyrosine phosphatase
TAE Tris-acetate-ethylenediaminetetraacetic acid
TBS Tris-buffered saline
TEMED Tetramethylethylenediamine
UTR Untranslated region
WT Wildtype
Acknowledgements
There are many people who helped and supported me throughout my Masters degree. First and foremost I would like to thank Dr. Christie for giving me the opportunity to participate as a volunteer and conduct research for my Honours degree, as well as my Masters degree. You gave me a chance when many others would not. You have been a very supportive and
encouraging supervisor and I really appreciate everything you have done for me.
The Christie laboratory was my second home and really became my second family. There were many highs and lows throughout my project, and I thank you all for the encouragement and support. I have truly made lifelong friends. I would first like to start by thanking Dr. Joana Gil-Mohapel, Dr. Patricia Brocardo and Dr. Mariana Vetrici, you are all amazing, smart individuals with great hearts. Not only were you there to discuss any research problems or concerns, but were also there as emotional support. Next, I would like to thank the Western blotting crew: Dr. Mariana Vetrici, Sarah de Rham and Kristin Morch. Thank you for all the discussions and for sharing the excitements and frustrations that are Western blots.
To all of the graduate students, you are all such great people and I wish you the best of luck. Crystal, you were a great colleague and friend. I will miss our lunch dates and our
ridiculous conversations with Timal that ended with one of us on the Quote Wall. Anna, you are such a sweet person and I really appreciate all the support you gave me during this thesis. Emily, I will miss our discussions in your room. Timal, the three things we share in common are: our love for loukoumades, our ridiculous French Immersion stories and the Quote Wall. Mohammed, I know you will do a great job carrying on the Fragile X torch.
I would also like to thank Jennifer Graham, Evelyn Wiebe and Karen Myers. Jen, I have no idea how anything got done around the lab without you. You are a great co-worker and friend. Evelyn, you are such an immense wealth of knowledge. Thank you for sharing it with me. Karen, thank you for your help and your jokes. I always left with a smile from yours and Evelyn’s room.
Dr. Dr. Brennan Eadie, thank you for being such a great supervisor and mentor. You are a big reason why I continued in research. Fanny Boehme, thank you so much for teaching me Western blots during my undergrad. All the skills I obtained were from you.
Dedication
I dedicate this thesis to my parents and to my sister and best friend, Gaby. You were all my rocks throughout my whole Masters degree. I could not have done this without your constant support
and unconditional love. I love you all very much.
Chapter 1: Introduction
1.1 Fragile X Syndrome
1.1.1 History
The observation of a sex-linked component to intellectual disability has been known for over a century. Some of the first noted cases were documented in prisons, where several
researchers found a larger number of males with intellectual disability when compared to females (Johnson, 1897; Penrose, 1938). At the time, researchers had believed that this finding was due to a greater likelihood of males being imprisoned due to their increased inclination for aggression. Studies from Australia, Europe, Canada and the United States of America further substantiated these observations and found approximately 30% more males than females with intellectual disability, agreeing with Penrose’s initial ratio (Baird & Sadovnick, 1985; Drillien, 1967; McLaren & Bryson, 1987; Stevenson et al., 1996; Stoller, 1965).
In 1943, two researchers named James Purdon Martin and Julia Bell performed a
pedigree analysis of a family with eleven severely intellectually disabled males and showed that intellectual disability was in fact inherited in an X-linked manner (Martin & Bell, 1943). In 1969, Herbert Lubs had also noticed a constriction near the end of the long arm of the X chromosome of intellectually disabled patients (Dunn et al., 1963; Losowsky, 1961; Lubs, 1969). Despite this interesting observation, many others were unable to reproduce this finding and Lubs’ discovery went largely unnoticed (Sherman, 2002).
Renewed interest in Lubs’ observation occurred a decade later, when Grant Sutherland noted that culture media lacking folic acid and thymidine was necessary for visualization of the constriction site, also known as the fragile X site due to its appearance as a break in the arm of the X chromosome (Sutherland, 1977). Using this protocol, reanalysis of chromosomes obtained from the affected males described by Martin and Bell in 1943 were found to contain the fragile X site (Richards et al., 1981).
In recognition of Martin and Bell’s original link between intellectual disability and the X chromosome, the syndrome was termed Martin-Bell syndrome. Today, Martin-Bell syndrome is more commonly known as Fragile X syndrome (FXS).
1.1.2 Etiology
Although FXS was thought to be an X-linked recessive disorder based on its association with the fragile site on the X chromosome (Dunn et al., 1963; Losowsky, 1961; Lubs, 1969; Martin & Bell, 1943), the typical inheritance was not observed. In a true X-linked recessive disorder, heterozygote female carriers are not affected, but all males are. However, 20% of male carriers were phenotypically normal, whereas 30% of heterozygote females were mildly affected (Sherman et al., 1984, 1985). These normal transmitting males (NTM) had a 40% chance of having a grandson with intellectual disability, whereas there was only a 9% chance that their brother would have intellectual disability (Sherman et al., 1984, 1985). The observation of increased penetrance for the disease in successive generations was termed the Sherman paradox (Opitz, 1986). The explanation for Sherman’s paradox was not clear until the discovery of the Fragile X Mental Retardation 1 (Fmr1) gene in 1991 (Verkerk et al., 1991) and was found to be due to a trinucleotide repeat expansion within this gene (Fu et al., 1991).
FXS is now known to be the most common form of inherited intellectual disability (Ashley et al., 1993a; Brown, 1990; Hornstra et al., 1993; Nussbaum & Ledbetter, 1986) and has a relatively high rate of incidence, affecting approximately 1 in 4000 males and 1 in 8000
females (Turner et al., 1996). In greater than 99% of the cases, FXS is caused by transcriptional repression of the Fmr1 gene, which normally encodes the Fragile X Mental Retardation Protein (FMRP) (Jin et al., 2004; Rousseau et al., 1995). In normal individuals, the 5’ untranslated region (UTR) found in exon 1 of the Fmr1 gene has about 30 cytidine-guanosine-guanosine (CGG) repeats (Ashley et al., 1993a; Brown et al., 1993; Fu et al., 1991; Pieretti et al., 1991; Snow et al., 1993). However, the trinucleotide sequence expands to greater than 200 repeats in FXS individuals with the full mutation (Ashley et al., 1993a; Fu et al., 1991). This leads to hypermethylation of the promoter and the 5’ UTR region causing transcriptional silencing of the
Fmr1 gene and subsequent loss of FMRP (Bell et al., 1991; Oberlé et al., 1991; Pieretti et al.,
The Fmr1 allele in humans has been subdivided into 4 categories based on the size of the CGG repeats (Maddalena et al., 2001) (Figure 1.1): normal, intermediate, premutation and full mutation. Normal alleles contain fewer than 45 CGG repeats and are usually transmitted in a stable manner (Fu et al., 1991; Kunst & Warren, 1994), whereas intermediate (also called ‘gray zone’ or ‘borderline’) alleles contain between 46 to 54 CGG repeats and are generally considered to be stable (Brown et al., 1993; Hagerman & Hagerman, 2004; McConkie-Rosell et al., 2005; Nolin et al., 1996, 2003; Reiss et al., 1994a; Sullivan et al., 2002; Zhong et al., 1996). In contrast, premutation alleles contain between 55-200 CGG repeats and are unstable upon transmission (Fu et al., 1991; Snow et al., 1993). A premutation can expand to a full mutation only through maternal transmission (Fu et al., 1991; Malter et al., 1997; Moutou et al., 1997; Oberlé et al., 1991; Rousseau et al., 1991; Snow et al., 1993), where the smallest premutation that has expanded into a full mutation upon transmission contained 59 repeats (Heitz et al., 1992; Nolin et al., 1996, 2003). Repeat alleles over 100 in a mother will expand to a full mutation in 100% of their offspring (Fu et al., 1991; Nolin et al., 1996; Snow et al., 1993), however alleles between 59-69 repeats, 70-79, 80-89 and 90-99 have a 37%, 65%, 70% and 93% chance, respectively to expand into a full mutation (Nolin et al., 1996). Lastly, full mutation alleles contain more than 200 CGG repeats. Females with the full mutation or the premutation have a 50% risk of passing their mutation to their offspring (Donnenfeld, 1998; Maddalena et al., 2001; Sherman, 2002). The expansion leads to hypermethylation of the cytidine-phosphate-guanosine (CpG) islands found in the promoter region as well as in the CGG repeats found in the 5’ UTR region, causing transcriptional silencing of the FMR1 gene and leading to the loss of FMRP (Ashley et al., 1993a; Bell et al., 1991; Fu et al., 1991; Oberlé et al., 1991; Pieretti et al., 1991; Sutcliffe et al., 1992).
Figure 1.1 The molecular and clinical effects of CGG repeat expansion on the Fmr1 gene and FMRP expression. The CGG repeat (yellow) located in the 5’ UTR region (green) of the Fmr1 gene expands to over 200 repeats in the full mutation, leading to hypermethylation ( ) of the promoter (blue) and the 5’ UTR region and subsequent transcriptional repression of the gene. This leads to loss of FMRP. The coding region of the gene is represented in purple.
1.1.3 Manifestations
The clinical phenotypes of FXS are quite broad, ranging from the most severe cases where it is associated with intellectual disability to milder cases where it can cause a number of cognitive and behavioral problems. These milder phenotypes include learning impairments, social phobia and attention-deficit hyperactivity disorder (ADHD) (Franke et al., 1998; Hagerman et al., 1999; Hagerman, 2002; Sobesky et al., 1996; Tassone et al., 2000a).
Almost all male individuals with the full mutation have some degree of intellectual disability (Cianchetti et al., 1991; Freund & Reiss, 1991; Jäkälä et al., 1997; Reiss et al., 1995a; Rousseau et al., 1991, 1995; Sobesky et al., 1994), where the average intelligence quotient (IQ) is about 40 (Merenstein et al., 1996). In contrast, about 50% of females with the full mutation have some degree of intellectual disability, with a mean IQ of 84, presumably due to X-inactivation (Cianchetti et al., 1991; Freund & Reiss, 1991; Jäkälä et al., 1997; Reiss et al., 1995a; Rousseau et al., 1991, 1995; Sobesky et al., 1994). In general, the clinical severity of FXS is correlated with the degree of FMR1 methylation, therefore influencing gene silencing and the level of FMRP expressed (Tassone et al., 1999; Taylor et al., 1994).
1.1.3.1 Development and Behavior
Developmental delays such as motor and language delays are common in FXS individuals. These include poor fine hand and finger motor coordination, low muscle tone (hypotonia), gross motor delays (e.g. walking, crawling, and motor skills dependent on large muscles), increased repetitive speech, stuttering (dysfluency), and articulation difficulties (Abbeduto & Hagerman, 1997; Friefeld & MacGregor, 1993; Fryns, 1984; Kau et al., 2002; Philofsky et al., 2004; Simko et al., 1989; Turner et al., 1980; Wolf-Schein et al., 1987).
Behavioral problems present in FXS individuals include hand flapping, poor eye contact, hand biting, hyperactivity, hyperarousal to sensory stimuli, aggression, shyness, anxiety and hypersensitivity to touch (tactile defensiveness) (Baumgardner et al., 1995; Chiu et al., 2007; Cohen et al., 1988; Cornish et al., 2004; Hagerman et al., 2009; Hagerman, 2002; Hatton et al., 2006; Kau et al., 2004; Merenstein et al., 1996; Miller et al., 1999; Reiss & Dant, 2003; Roberts
et al., 2001; Sullivan et al., 2006; Turk, 1998). Other features include ADHD, gaze aversion and
other social difficulties, mood disorders, autism spectrum disorders (ASD), visuospatial difficulties and executive dysfunction (Belser & Sudhalter, 2001; Borghgraef et al., 1987; Bregman et al., 1988; Cornish et al., 1999, 2001, 2004; Freund & Reiss, 1991; Hooper et al., 2008; Kaufmann et al., 2004; Merenstein et al., 1996; Munir et al., 2000a, b; Reiss & Freund, 1990; Tassone et al., 1999; Tsiouris & Brown, 2004; Wilding et al., 2002; Wolff et al., 1989). Of these features, autism-like behaviors are seen in 30% of male individuals affected with the full
mutation, ADHD behaviors are seen in 80% of the cases and 70-100% exhibit anxiety (Jacquemont et al., 2007; Rogers et al., 2001).
1.1.3.2 Physical features
Typical physical features associated with FXS are subtle in young children making the detection of the syndrome difficult (Figure 1.2). However, these become more evident later in childhood and early adolescence (Hersh & Saul, 2011; Lachiewicz et al., 2000). Characteristic physical features in FXS individuals include a long, narrow face, large prominent ears and prominent forehead and jaw (Fryns, 1984; Musumeci et al., 1988; Verma & Elango, 1994). Other features are associated with connective tissue defects (Opitz et al., 1984; Verma & Elango, 1994), such as a high-arched palate (Hagerman, 1987; Hagerman et al., 1983; Partington, 1984; Simko et al., 1989), hyperextensible joints (Hagerman et al., 1984), mitral valve prolapse (Loehr
et al., 1986), flat feet (Davids et al., 1990), and soft, velvet-like skin (Turner et al., 1980).
Additionally, abnormally large testes (macroorchidism) are present in about 90% of
postpubescent males (Butler et al., 1992; Flynn et al., 2002; Merenstein et al., 1996; Turner et
al., 1980).
Although the physical, behavioral and development features discussed above are common in FXS individuals with the full mutation, they can also be found in individuals carrying the premutation (Aziz et al., 2003).
Figure 1.2 Physical phenotypes present in FXS individuals. (A) Two male children with FXS (B) and the same individuals as adults (Adapted from http://raredisorders.wordpress.com). Notice how the typical physical phenotypes are subtle in these individuals as children but become more evident at the adult stage. (C) An X chromosome from a FXS individual, where constriction at the tip is very noticeable (red arrow) (Adapted from Harrison et al., 1983).
1.1.3.3 Premutation
Compared to the full mutation, the prevalence of the premutation is fairly high within the general population, occurring in 1 in 813 males and 1 in 259 females (Dombrowski et al., 2002; Rousseau et al., 1995). In general, most individuals with the premutation have a normal IQ, however mild cognitive and behavioral deficits can be present (Cornish et al., 2005; Kéri & Benedek, 2010; Moore et al., 2004). These include executive dysfunction, obsessive-compulsive disorder, anxiety, attention problems, ASD behaviors (hand flapping, poor eye contact, social deficits), and the presence of prominent ears (Aziz et al., 2003; Farzin et al., 2006; Franke et al., 1998; Goodlin-Jones et al., 2004; Hessl et al., 2005; Lachiewicz et al., 2006; Moore et al., 2004; Tassone et al., 2000a).
There are two forms of clinical presentations that can be seen in individuals carrying a premutation. The first occurs in females and is referred to as premature ovarian insufficiency
(POI), which is defined as an early onset of menopause before the age of 40 years (Coulam et
al., 1986). POI occurs in about 20% of premutation females as compared to approximately 1% of
the general population (Allingham-Hawkins et al., 1999; Coulam et al., 1986; Cronister et al., 1991; Marozzi et al., 2000; Partington et al., 1996; Schwartz et al., 1994; Sherman, 2000). The second form of clinical presentation is observed in about 30% of males over the age of 50 and consists of cognitive decline, brain atrophy, progressive cerebellar or intention tremor, poor muscle coordination (ataxia), parkinsonism, balance problems, frequent falls, and nerve damage (neuropathy) (Berry-Kravis et al., 2003; Hagerman et al., 2001; Hall et al., 2005; Jacquemont et
al., 2003, 2004, 2007; Leehey et al., 2003). This phenotype is termed fragile-X-associated
tremor/ataxia syndrome (FXTAS). Both POI and FXTAS are related to RNA toxicity due to increased FMR1 messenger ribonucleic acid (mRNA) levels (Kenneson et al., 2001; Tassone et
al., 2000b, c, d) (Figure 1.1). However, despite the observed increase in mRNA levels, some
individuals with the premutation present with lower than normal levels of FMRP (Aziz et al., 2003; Goodlin-Jones et al., 2004; Hessl et al., 2005; Kenneson et al., 2001; Tassone et al., 2000a).
1.1.4 Neuroimaging Studies
The loss of FMRP leads to a number of neurological symptoms, that may be explained by anatomical abnormalities in certain brain structures. Using magnetic resonance imaging (MRI), the first structure found to be affected in FXS individuals was the cerebellar vermis, a narrow, worm-like structure located between the left and right hemispheres of the cerebellum. This structure plays a role in regulating motor activity such as programming, control and balance, as well as executive function, working memory, and language and music processing (Andermann et
al., 1975; Desmond et al., 1997; Penhune et al., 1998; Ryding et al., 1993; Schmitt et al., 2001;
Takagi et al., 1998). Particularly, the size of the posterior segment of the cerebellar vermis is significantly decreased in these individuals, where the size of this region is negatively correlated with FMRP levels as well as IQ, but is positively correlated with autistic behaviors (Gothelf et
al., 2008; Guerreiro et al., 1998; Hoeft et al., 2008; Kaufmann et al., 2003; Lightbody & Reiss,
2009; Mazzocco et al., 1997; Mostofsky et al., 1998; Reiss et al., 1988, 1991a, b). The structural alteration of the cerebellar vermis might explain the motor, language and executive dysfunctions seen in FXS individuals.
A second brain region that has been shown to be affected in FXS is the temporal lobe, with a specific enlargement of the superior temporal gyrus (Reiss et al., 1994b). This area of the brain is important for the processing of complex auditory stimuli, such as language, speech and hearing (Binder et al., 1997; Howard et al., 2000; Rajarethinam et al., 2000). Additionally, this region has been implicated in the processing of facial perception, such as gaze direction, mouth movements, and emotional expressions (Calder et al., 2002; Calvert et al., 1997; Hoffman & Haxby, 2000; Puce et al., 1998; Wicker et al., 2003). This abnormal increase in the superior temporal gyrus may lead to the social cognitive deficits along with the language and speech delays seen in FXS individuals.
In addition, the caudate nucleus is significantly enlarged in FXS individuals, where the size of this region is negatively correlated with FMRP levels and IQ, but is positively correlated with aberrant sensory and language behaviors such as tactile defensiveness and repetitive speech (Eliez et al., 2001; Gothelf et al., 2008; Hazlett et al., 2009; Hoeft et al., 2008; Lee et al., 2007; Reiss et al., 1995b). This is an area that has several connections with the frontal lobes and thus is involved in executive functions as well as in planning and problem solving, verbal fluency, learning, short and long-term memory, motor planning, and goal-directed behavior planning, all of which are affected in FXS (Alexander & Crutcher, 1990; Alexander et al., 1986; Balleine et
al., 2007; Dagher et al., 1999; Davidson et al., 2004; Delgado et al., 2005; DeLong et al., 1983;
Fuh & Wang, 1995; Haruno & Kawato, 2006; Jueptner et al., 1997; Leh et al., 2007; Lewis et
al., 2004; Mendez et al., 1989; Monchi et al., 2001, 2006; Poldrack et al., 1999; Rogers et al.,
2000; Schmidtke et al., 2002; Turner et al., 2006).
Notably, the fusiform gyrus, fourth ventricle and lateral ventricles are all enlarged,
whereas the insula and amygdala are decreased (Eliez et al., 2001; Gothelf et al., 2008; Hazlett et
al., 2009; Hoeft et al., 2008; Kates et al., 1997; Mostofsky et al., 1998; Reiss et al., 1988, 1991a,
b, 1995b). The fusiform gyrus is a region that is involved in facial processing and colour perception. The fourth ventricle and lateral ventricles carry cerebrospinal fluid and play an important role in protecting the brain against traumatic injuries (Allison et al., 1993, 1994; Lueck
et al., 1989; Magendie, 1842; Novak et al., 2000; Puce et al., 1995). Of note, in FXS the lateral
ventricular volume is negatively correlated with IQ (Reiss et al., 1995b). The insula is involved in emotional processing, whereas the amygdala plays a role in emotional processing as well as social behavior (Maclean, 1955; Meunier et al., 1999; Phan et al., 2002; Rosvold et al., 1954;
Weiskrantz, 1956). The volumetric abnormalities of these regions may lead to region-specific behavioral and cognitive deficits, explaining the associated phenotypes in FXS individuals.
The hippocampus, a region involved in learning and memory and visuospatial ability, also appears to be affected in FXS (Hazlett et al., 2009; Hoeft et al., 2008; Jarrard, 1993; Kates
et al., 1997; Kesner et al., 1989; Morris et al., 1982; Reiss et al., 1994b; Schmajuk, 1990;
Scoville & Milner, 1957; Takase et al., 2005). However, a consensus as to how the disease affects this structure has not been reached, with some studies showing an increase in
hippocampal volume (Kates et al., 1997; Reiss et al., 1994b), an age-dependent decrease in the volume of this structure (Hoeft et al., 2008), or morphological alterations without a change in hippocampal volume (Jäkälä et al., 1997). Despite these discrepancies, the hippocampus is known to express high levels of Fmr1 mRNA (Abitbol et al., 1993; Hinds et al., 1993), perhaps suggesting that this region may be severely affected by the loss of FMR1 mRNA and FMRP in FXS individuals.
1.2 The Fmr1 gene and FMRP
1.2.1 Structure
The Fmr1 gene is a 38 kilobase (kb) gene consisting of 17 exons that encodes a 4.4 kb mRNA transcript (Eichler et al., 1993) and is located at the far end (27.3) on the long arm (q) of the X chromosome (X), collectively referred to as position Xq27.3 (Giraud et al., 1976; Harrison
et al., 1983; Harvey et al., 1977). In FXS, this position appears pale and thin under
folate-deficient medium giving this site a fragile appearance, hence the syndrome’s name (Sutherland, 1977) (Figure 1.2c). Due to alternative splicing through the exclusion of exons 12 and 14 and the choice of acceptor site in exons 15 and 17, about 12 FMRP isoforms are known (Ashley et
al., 1993a; Verkerk et al., 1993) (Figure 1.3). These isoforms range from 70-80 kilodaltons
(kDa), where the longest FMRP isoform is 631 amino acids long (Ashley et al., 1993a; Devys et
al., 1993; Pieretti et al., 1991; Verkerk et al., 1993). However, only 4-5 FMRP isoforms have
been detected and their functional significance is currently unknown (Devys et al., 1993; Lim et
al., 2005; Verkerk et al., 1993).
Encoded within the 17 exons of the Fmr1 gene are a nuclear localization signal (NLS), a nuclear export signal (NES), two coiled-coil domains as well as two K Homology (KH) domains
(KH1 and KH2), an arginine-glycine-glycine (RGG) box, and two tandem Tudor domains (labeled as Agenet 1 and 2 in Figure 1.3) (Adams-Cioaba et al., 2010; Ashley et al., 1993b; Eberhart et al., 1996; Maurer-Stroh et al., 2003; Siomi et al., 1993). The KH domains and the RGG box are RNA binding domains, which together can bind approximately 4% of the mRNA present in the mammalian brain (Ashley et al., 1993b; Brown et al., 1998; Siomi et al., 1993). A naturally occurring isoleucine-304-asparagine (I304N) mutation in the KH2 domain causes a severe form of FXS (Figure 1.3), altering the domain’s ability to associate with actively
translating polyribosomes, indicating the importance of this interaction for the proper function of FMRP (De Boulle et al., 1993; Brown et al., 1998; Feng et al., 1997a). The RGG box in FMRP binds to RNA-containing guanine-quartet (G-quartet) structures, which include Fmr1 mRNA, and many mRNA’s that encode proteins important for neuronal function, such as microtubule associated protein 1B (MAP1B) and post-synaptic density protein 95 (PSD-95) (Ashley et al., 1993b; Brown et al., 2001; Darnell et al., 2001; Schaeffer et al., 2001; Todd et al., 2003; Zalfa et
al., 2003) (Figure 1.3).
The NLS and NES are involved in the shuttling of FMRP between the nucleus and the cytoplasm, where FMRP has been shown to be primarily cytoplasmic (Devys et al., 1993; Feng
et al., 1997b). Indicating the importance of the NLS in the function of FMRP is an
arginine-138-glutamine (R138Q) mutation in a patient showing developmental delay (Collins et al., 2010) (Figure 1.3). The coiled-coil domains are involved in protein-protein interactions, where the first coiled coil domain, found in exon 7, is involved in binding to FMRP interacting proteins such as FMRP, its paralogs Fragile X Mental Retardation Syndrome-Related Protein 1 (FXR1P) and 2 (FXR2P), and Cytoplasmic FMRP Interaction Proteins 1 (CYFIP1) and 2 (CYFIP2) (Bardoni
et al., 2003; Schenck et al., 2001; Siomi et al., 1996; Zhang et al., 1995). The second coiled-coil
domain spans exons 13 and 14 and is believed to bind to ribosomes, specifically the 60S large ribosomal subunits, however this is contradicted by the ability of FMRP to bind directly to ribosomes through RNA (Siomi et al., 1996; Tamanini et al., 1996). Finally, the two tandem Tudor domains (labeled as Agenet 1 and 2 in Figure 1.3) located at the N-terminus of FMRP are involved in binding of trimethylated lysine residues (Adams-Cioaba et al., 2010; Maurer-Stroh et
Figure 1.3 Structure of the Fmr1 gene and FMRP. The coding exons (dark green), introns (black lines) and alternative splicing variants (blue lines) are depicted in the Fmr1 gene. FMRP and its functional domains are aligned with their corresponding amino acids (© Reproduced with permission of ANNUAL REVIEWS; Santoro et al., 7: 2012; permission conveyed through Copyright Clearance Center, Inc.).
1.2.2 Homology
The Fmr1 gene has been conserved during evolution indicating its importance. Orthologs to the human FMR1 gene have been identified in several mammalian vertebrates, including Mus
musculus (mouse), non-mammalian vertebrates such as Gallus gallus (chicken), Danio rerio
(zebrafish), Xenopus laevis (frog), and invertebrates such as Drosophila melanogaster (fruitfly) (Ashley et al., 1993b; Price et al., 1996; Siomi et al., 1995; Tucker et al., 2004; Wan et al., 2000).
The human and Mus Musculus Fmr1 gene homologues are 97% identical in amino acid sequence and show similar expression patterns (Ashley et al., 1993a; Bakker et al., 2000). Gallus
gallus, Xenopus laevis, Danio rerio, Drosophila melanogaster share 92%, 86%, 74%, and 35%
amino acid identity to human FMRP, respectively (van’t Padje et al., 2005; Price et al., 1996; Siomi et al., 1995; Zhang et al., 2001). Homologues for the Fmr1 gene in Caenorhabditis
The Fmr1 gene also possesses two paralogs, FXR1 and FXR2, which are 60% identical in their amino acid sequences (Siomi et al., 1995; Zhang et al., 1995).
1.2.3 Expression
The highest expression of Fmr1 mRNA in human adult tissue occurs in the brain and testes. This mRNA is also detectable in the placenta, lungs and kidneys but is almost absent in muscle and heart tissues (Hinds et al., 1993; Khandjian et al., 1995; Ladd et al., 2007). Within the brain, the granular layers of the hippocampus and the cerebellum have the highest expression of Fmr1 mRNA (Abitbol et al., 1993; Devys et al., 1993; Hinds et al., 1993).
Correlating with the RNA data, FMRP is also highly expressed in the brain and testes and is especially high within the hippocampus, cerebellum, cerebral cortex and piriform cortex, suggesting that the loss of FMRP may be especially detrimental to these regions (Bakker et al., 2000; Devys et al., 1993; Feng et al., 1997b; Tamanini et al., 1997). FMRP is highly expressed in neurons, however little or no expression is seen in glia (Abitbol et al., 1993; Bakker et al., 2000; Devys et al., 1993; Hinds et al., 1993). Within neurons, FMRP is found mainly in the cytoplasm, however it has also been found in both the nucleus and dendritic spines (Antar et al., 2004; Devys et al., 1993; Eberhart et al., 1996; Feng et al., 1997b; Verheij et al., 1993).
Importantly, FMRP is developmentally regulated in the murine brain, where the highest expression is seen shortly after birth, but diminishes quickly after the first week of postnatal life and remains low throughout life (Lu et al., 2004; Wang et al., 2004). Interestingly, PSD-95, a protein seen in mature synapses and whose mRNA is known to associate with FMRP (Brown et
al., 2001; Muddashetty et al., 2007; Zalfa et al., 2007), peaks one week following the peak
expression of FMRP, suggesting that the high levels of FMRP may play a functional role in early brain development (Wang et al., 2004).
As mentioned in section 1.2.1, there are a number of Fmr1 transcripts and thus FMRP isoforms. However, the same ratios of splice variants and isoforms exist in different human tissue, such as the kidney, liver and brain (Devys et al., 1993; Tamanini et al., 2000; Verkerk et
al., 1993). This could possibly indicate that there is no tissue-specific function for any of the
different isoforms or variants.
1.2.4 Function
Since the identification of the Fmr1 gene in 1991, a number of studies have tried to elucidate FMRP’s exact function (Verkerk et al., 1991). Although progress has been made, its precise physiological function is still unclear. However, proposed functions can be deduced from its structure and localization. Thus, the RNA-binding domains in FMRP suggest that it is an RNA-binding protein (Ashley et al., 1993b; Brown et al., 1998; Siomi et al., 1993). In addition, the NLS and NES suggest that FMRP shuttles between the nucleus and the cytoplasm, which is further supported by the observation that FMRP is primarily cytoplasmic, however it has also been seen in small amounts in the nucleus of neurons (Devys et al., 1993; Eberhart et al., 1996; Feng et al., 1997b; Verheij et al., 1993). Further evidence of FMRP’s function comes from the fact that FMRP has been found to associate with actively translating polyribosomes, which are highly localized beneath synapses (Antar et al., 2004; Corbin et al., 1997; Eberhart et al., 1996; Feng et al., 1997a, b; Khandjian et al., 1996; Steward & Fass, 1983; Steward & Levy, 1982; Steward, 1983). However, FMRP does not associate with polyribosomes on its own, but rather forms a messenger ribonucleoprotein (mRNP) complex composed of mRNAs and a number of proteins such as its autosomal paralogs, FXR1P and FXR2P (Ceman et al., 1999, 2000; Feng et
al., 1997a). Interestingly, thin, long spines have been observed in both FXS patients and Fmr1
knockout (Fmr1-/y) mouse brains, indicating that the loss of FMRP affects synaptic pruning and maturation (Comery et al., 1997; Cornish et al., 2004; Hinton et al., 1991; Irwin et al., 2001). Together, these results have indicated that FMRP may play a role in synaptic plasticity, the ability of synapses to change strength depending on whether the synapse is being used or not.
Additional evidence for FMRP’s involvement in synaptic plasticity comes from a number of studies looking at its bound mRNAs. Unexpectedly, one of the first mRNAs shown to bind to FMRP was its own (Ashley et al., 1993b; Schaeffer et al., 2001). Further studies have
investigated a number of mRNAs and found that a majority of the interacting mRNAs encode for proteins important in neuronal function. These include proteins found at postsynaptic sites such as MAP1B, PSD-95, Ca2+/calmodulin-dependent protein kinase II alpha (CaMKIIα), the N-methyl-D-aspartic acid (NMDA)-type glutamate receptor subunits GluN1, GluN2A and GluN2B, and the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type glutamate receptor subunits GluA1 and GluA2 (Brown et al., 2001; Edbauer et al., 2010; Muddashetty et al., 2007; Schütt et al., 2009; Zalfa et al., 2003, 2007). However, another study
has shown that the mRNAs that interact with FMRP do not include those that encode for the AMPA receptor subunits GluA1 and GluA2 (Darnell et al., 2011). Another indication that FMRP plays a role in synaptic activity is observed during activation of group I metabotropic glutamate receptors (mGluRs), where Fmr1 mRNA is translated near the synapse resulting in a subsequent increase in FMRP expression (Weiler et al., 1997).
Although a number of studies have shown that FMRP acts as a translational repressor (Antar et al., 2004; Huber et al., 2002; Krueger et al., 2011; Laggerbauer et al., 2001; Li et al., 2001; Wei et al., 2007; Zalfa et al., 2003; Zhang et al., 2001; Zhu et al., 2011), it can also act as a translational activator (Bechara et al., 2009). This is supported by the fact that an absence of FMRP can lead to both an upregulation and downregulation of its bound mRNA (Brown et al., 2001; Darnell et al., 2001; Miyashiro et al., 2003). In addition, a number of proteins are
increased (e.g. amyloid precursor protein) and decreased (e.g. MAP1B) in Fmr1-/y mice (Liao et
al., 2008; Schütt et al., 2009; Wei et al., 2007; Westmark & Malter, 2007). The exact mechanism
by which FMRP selectively represses proteins while activating others remains unclear. However, phosphorylation at serine 500 (serine 499 in murine) of FMRP is believed to play a role in its function. FMRP phosphorylated at this residue appears to be associated with stalled ribosomes and therefore acts as a translational repressor, whereas unphosphorylated FMRP appears to act as a translational activator (Ceman et al., 2003; Muddashetty et al., 2011).
Collectively, the observations discussed above have lead to the hypothesis that FMRP forms an mRNP complex with mRNAs in the nucleus and subsequently exports them to polyribosomes in the cytoplasm, therefore influencing translation of its bound mRNAs at the synapse (Jin & Warren, 2000).
1.3 Fmr1 knockout mouse
The understanding of the biochemical and synaptic alterations leading to FXS was greatly advanced in 1994 with the development of a mouse model of FXS (The Dutch-Belgian Fragile X Consortium, 1994). The Fmr1-/y mouse was generated by the insertion of a neomycin cassette
into exon 5 of the Fmr1 gene (Figure 1.3). These mice, like FXS individuals, do not express FMRP, showing similar levels and expression patterns of FMRP and Fmr1 mRNA (Hinds et al., 1993; The Dutch-Belgian Fragile X Consortium, 1994).
Importantly, Fmr1-/y mice mimic a number of physical, behavioral and cognitive
phenotypes seen in FXS individuals. Although facial abnormalities common to FXS individuals are not seen in Fmr1-/y mice, macroorchidism is observed in these mice (The Dutch-Belgian
Fragile X Consortium, 1994). Behavioral and cognitive abnormalities seen in Fmr1-/y mice include hyperactivity, as well as learning and memory impairments (Chen & Toth, 2001;
D’Hooge et al., 1997; Van Dam et al., 2000; Dobkin et al., 2000; Eadie et al., 2012; Huber et al., 2002; Kooy, 2003; Kooy et al., 1996; Li et al., 2002; Mineur et al., 2002; Nielsen et al., 2002; Paradee et al., 1999; The Dutch-Belgian Fragile X Consortium, 1994). Furthermore, Fmr1-/y mice show hyperactivity and hyper responsiveness to sensory stimuli, in compliance with sensory hyperarousal seen in FXS individuals (Chen & Toth, 2001; Miller et al., 1999; Musumeci et al., 2000). Additionally, Fmr1-/y mice and FXS individuals show dendritic
abnormalities such as an excess of long, thin immature spines and an increase in spine density in the cortex (Comery et al., 1997; Hinton et al., 1991; Irwin et al., 2001, 2002; Rudelli et al., 1985) and hippocampus (Bilousova et al., 2009; Grossman et al., 2006, 2010; Levenga et al., 2011).
Based on the high homology between mouse and human FMRP, comparable Fmr1 mRNA and FMRP expression levels, and emulation of clinical and pathological phenotypes seen in FXS, the Fmr1-/y mouse is considered to be a robust model to study this syndrome (The Dutch-Belgian Fragile X Consortium, 1994).
1.4 Synaptic Plasticity
Neurons have the ability to strengthen or weaken their synaptic connections in response to activity. This dynamic change in synapses in response to activity is termed synaptic plasticity (Eccles & McIntyre, 1951). One form of synaptic plasticity is long-term potentiation (LTP), which is described as the persistent strengthening of synapses based on recent patterns of activity that last greater than 30 minutes (Bliss & Lomo, 1973). In the late 1940s, Donald Hebb
hypothesized that learning and memory occurred in the brain through an increase in synaptic efficacy, which arose from the persistent stimulation of the postsynaptic cell (a phenomenon later known as LTP) (Hebb, 1949). This hypothesis is referred today as the Hebbian theory. In 1973, LTP was discovered when synaptic transmission was enhanced for days or weeks after
stimulation of the rabbit hippocampus with a few seconds of high-frequency stimulation (Bliss & Lomo, 1973). Furthermore, it was observed that some types of LTP are dependent on the
activation of the NMDA receptor and that addition of an NMDA receptor antagonist in the hippocampus of rats resulted in impairments in learning a spatial task in the Morris Water Maze test (Collingridge et al., 1988a, b; Davis et al., 1992).
A second form of synaptic plasticity discovered in the late 1980s is long-term depression (LTD), which is the opposite of LTP (Christie & Abraham, 1992a, b; Dudek & Bear, 1992; Lynch et al., 1977). LTD is defined as the persistent weakening of synapses based on recent patterns of activity lasting greater than 30 minutes (Abraham et al., 1994). Like LTP, some forms of LTD are also dependent on the activation of NMDA receptors (Christie & Abraham, 1992a, b; Desmond et al., 1991). Both LTP and LTD are believed to be neurobiological models for learning and memory, and have been shown to occur in several brain regions (Bliss & Collingridge, 1993; Goda & Stevens, 1996; Malenka & Bear, 2004).
1.4.1 NMDA receptor-dependent LTP
During LTP, glutamate is released from the presynaptic membrane, crosses the synaptic cleft and binds to AMPA and NMDA receptors (AMPAR and NMDAR) present on the postsynaptic membrane. This causes AMPAR channels to open, allowing the passage of Na+. When there is sufficient depolarization of the postsynaptic membrane, the Mg2+ ion in the NMDAR channel is expelled, allowing the flow of Ca2+ and Na+ through the NMDAR. Na+ further depolarizes the postsynaptic membrane, whereas Ca2+ acts as a secondary messenger and interacts with calmodulin, which in turn activates CaMKII. CaMKII phosphorylates existing AMPARs at serine amino acid residue 831 (S831) of the GluA1 subunit, increasing the ionic conductance of the channel. Additionally, CaMKII promotes the movement of non-synaptic AMPARs from the intracellular pool adjacent to the postsynaptic membrane to the postsynaptic membrane (Barria et al., 1997; Derkach et al., 1999; Mammen et al., 1997; Shi et al., 1999; Snyder et al., 2000) (Figure 1.4). This results in a larger subsequent postsynaptic response.
Figure 1.4 Molecular Mechanisms Underlying NMDAR-dependent LTP. Glutamate released by the presynaptic terminal binds and causes the opening of AMPARs and passage of Na+. Sufficient depolarization of the postsynaptic terminal causes the release of the Mg2+ ion found in the NMDAR channel, allowing the passage of Ca2+ and even more Na+. The high amounts of Ca2+ activate CaMKII, which in turn phosphorylates existing and intracellular AMPARs at S831 of the GluA1 subunit. This leads to an increase in ionic conductance and amount of AMPARs at the postsynaptic membrane. ℗ = phosphorylation.
1.4.2 NMDA receptor-dependent LTD
The cellular mechanism of LTD is very similar to that of LTP with two main differences. The first difference is that LTD occurs when there is a smaller rise in Ca2+ at the postsynaptic membrane. Secondly, rather than Ca2+ activating kinases as in LTP, it activates protein
phosphatases (PP) that dephosphorylate the serine amino acid residue 845 (S845) on the GluA1 subunit of the AMPAR. This causes internalization of AMPARs, decreasing the amount of
receptors present at the postsynaptic membrane and resulting in the weakening of synapses (Ehlers, 2000; Lee et al., 1998, 2000; Man et al., 2007) (Figure 1.5).
Figure 1.5 Molecular Mechanisms Underlying NMDAR-dependent LTD. Glutamate released by the presynaptic terminal binds and causes the opening of AMPARs and passage of Na+. Sufficient depolarization of the postsynaptic terminal causes the expulsion of the Mg2+ ion found in the NMDAR channel, allowing the passage of Ca2+ and even more Na+. The low amounts of Ca2+ activate PP, which in turn dephosphorylate AMPARs at S845 of the GluA1 subunit and results in their internalization. This leads to a reduction in the number of AMPARs at the postsynaptic membrane. -℗ = dephosphorylation.
1.4.3 mGluR-dependent LTD
A second form of LTD depends on mGluRs, specifically group I mGluRs (Bashir et al., 1993; Bolshakov & Siegelbaum, 1994; Ito et al., 1982; Kano & Kato, 1987). mGluRs are G-protein coupled receptors that appear to be linked to several biochemical pathways depending on
their location in the brain (Nicoletti et al., 1986a, b, c; Recasens et al., 1988; Sladeczek et al., 1985; Sugiyama et al., 1987). However, like NMDAR-dependent LTD, the end result is internalization of AMPA receptors and a reduction in the number of AMPARs at the synapse (Moult et al., 2006; Snyder et al., 2001; Steinberg et al., 2004; Wang & Linden, 2000). For example, during LTD in the cerebellum, activation of group I mGluRs leads to the formation of diacylglycerol (DAG) and inositol-1,4,5-triphosphate (IP3). IP3 binds to its receptor on the endoplasmic reticulum (ER), where it stimulates the release of Ca2+. Together, Ca2+ and DAG lead to activation of protein kinase C (PKC), which phosphorylates GluA2-containing AMPA receptors on serine 880 and causes their internalization (Chung et al., 2003; Pin & Duvoisin, 1995; Schoepp & Conn, 1993; Schoepp et al., 1994; Steinberg et al., 2006; Toms et al., 1995). In contrast, during LTD in the CA1 subregion of the hippocampus, activation of group I mGluRs leads to activation of a protein tyrosine phosphatase called striatal-enriched protein tyrosine phosphatase (STEP). STEP dephosphorylates AMPARs at tyrosine residues leading to
endocytosis of surface AMPARs (Gladding et al., 2009; Huang & Hsu, 2006; Moult et al., 2002, 2006; Schnabel et al., 1999; Zhang et al., 2008).
1.5 NMDA receptors
Ionotropic glutamate receptors (iGluRs) are the major excitatory synaptic receptors in the central nervous system (Collingridge & Lester, 1989; Mayer & Westbrook, 1987; Mayer, 2006; Watkins & Evans, 1981). This family of receptors allows passage of ions across a membrane in response to the binding of glutamate. The three types of receptors classified based on their pharmacological agonists are the NMDARs, AMPARs and kainate receptors (Dingledine et al., 1999). The NMDAR will be the focus of this thesis due to its role in learning and memory and its recent involvement in FXS (Bliss & Collingridge, 1993; Collingridge et al., 1988a, b; Danysz et
al., 1988; Davis et al., 1992; Morris et al., 1986; Ward et al., 1990).
1.5.1 Structure
NMDARs are heterotetrameric channels that are composed of two obligatory subunits, GluN1, and two regulatory subunits GluN2 (A-D) or GluN3 (A, B) (Furukawa et al., 2005;
Laube et al., 1998; Premkumar & Auerbach, 1997). Notably, the GluN3 subunit cannot form a functional receptor on its own, but needs to co-assemble with GluN1/GluN2 complexes (Cull-Candy et al., 2001; Das et al., 1998; Perez-Otano et al., 2001). Activation of NMDARs requires simultaneous binding of glycine (or D-serine) and glutamate (Johnson & Ascher, 1987; Kleckner & Dingledine, 1989; Lerma et al., 1990). The GluN1 and GluN3 subunits bind glycine, whereas the GluN2 subunits bind glutamate (Furukawa & Gouaux, 2003; Furukawa et al., 2005; Yao et
al., 2008).
Each subunit has an extracellular N-terminus, intracellular C-terminus, three
transmembrane regions (M1, M3 and M4) and one cytoplasmic re-entrant membrane loop (M2), where M2 from each of the four subunits make the channel pore (Bennett & Dingledine, 1995; Hollmann et al., 1994; Kuner et al., 1996; Wo & Oswald, 1994, 1995; Wood et al., 1995) (Figure 1.6). The ligand-binding domain is a clam-shaped structure formed from two
extracellular polypeptide stretches called S1 and S2. S1 is the polypeptide stretching from the N-terminus to M1, forming one half of the clamshell (D1). S2 is the polypeptide stretching between M3 and M4, which forms the other half of the clamshell (D2) (Figure 1.6). The ligand, either glutamate or glycine, binds between D1 and D2 (Furukawa et al., 2005; Traynelis et al., 2010).
The NMDAR differs from other iGluRs due to several unique properties, such as its high Ca2+ permeability and its voltage-dependent Mg2+ channel block (Davis & Linn, 2003;
MacDermott et al., 1986; Mayer et al., 1984). Despite its high permeability to Ca2+, the NMDAR is also selectively permeable to Na+ and K+ ions, albeit to a lesser extent (Mayer et al., 1987; Murphy et al., 1987; Sharma & Stevens, 1996). The high permeability of the NMDAR to Ca2+ is due to the presence of an asparagine residue in the asparagine/glutamine/arginine site located at the apex of the M2 loop (Burnashev et al., 1992; Mori et al., 1992; Moriyoshi et al., 1991). When this site is mutated from an asparagine to a glutamine, Ca2+ permeability and the Mg2+ block are reduced (Burnashev et al., 1992; Mori et al., 1992). However, mutation from an asparagine to arginine abolishes both Ca2+ permeability and the Mg2+ block (Burnashev et al., 1992; Sakurada et al., 1993).
Figure 1.6 NMDAR Structure. Diagram of the GluN1 (left) and GluN2 subunits (right). Each subunit contains an extracellular N-terminus domain, an intracellular C-terminus, three
transmembrane regions (M1, M3 and M4), and the pore-forming re-entrant loop (M2). The clam-shaped ligand-binding domain is formed by S1 and S2, where each forms half of the clamshell called D1 and D2. The ligand, either glycine or glutamate, can bind between D1 and D2.
1.5.2 Developmental expression
1.5.2.1 mRNA expression
GluN1, GluN2B and GluN2D mRNA expression begins as early as embryonic day (E) 14. By E17, all 3 mRNAs are increased, where GluN1 is highly and ubiquitously expressed and GluN2B is highly expressed in the cortex, thalamus and the spinal cord and to a lower extent in the hippocampus and hypothalamus. At this time, GluN2D is highly expressed in midbrain structures. Notably, GluN2A and GluN2D are not detectable throughout the whole embryonic period (Monyer et al., 1994).
There is persistent expression of GluN1, GluN2B and GluN2D at postnatal day (P) 0. Despite GluN1 remaining strong and ubiquitously expressed throughout the brain, GluN2B
becomes stronger in the hippocampus and low levels begin to appear in the cerebellum (CB) at this time. In addition, GluN2D becomes detectable in the cortex, hippocampus, and septum. Notably, GluN2A and GluN2C are faintly expressed in the hippocampus and cerebellum,
respectively (Monyer et al., 1994; Sheng et al., 1994; Watanabe et al., 1992; Zhong et al., 1995). All subunit mRNAs peak at around postnatal week 3 but decline to adult levels, with the exception of GluN2D which peaks earlier at around P7, but declines to barely detectable levels in adulthood. Specifically, GluN2A is upregulated during the 2nd week after birth, peaks in the hippocampus and cortex at postnatal week 3 and declines to adult levels. GluN2B is highly expressed during postnatal age, but peaks in the hippocampus and cortex during postnatal week 3, before declining to adult levels. Interestingly, GluN2C is found at very low levels in the cerebellum and forebrain at P7, but substantially increases in the cerebellum by P12, reaching a peak at postnatal week 3, and continuing to be expressed at high levels in the adult brain
(Monyer et al., 1994; Sheng et al., 1994; Watanabe et al., 1992; Zhong et al., 1995).
1.5.2.2 Protein expression
Like GluN1 mRNA, GluN1 is highly and ubiquitously expressed throughout the brain during development (Benke et al., 1995; Sheng et al., 1994). At P0, GluN2B is the predominant subunit throughout the brain, but becomes restricted to the cerebral cortex, and the CA1 and dentate gyrus (DG) subregions of the hippocampus in the adult brain. In the CB, GluN2B is expressed early until P10 but becomes undetectable at P21 and remains undetectable in
adulthood. In contrast, the GluN2A subunit is not detectable at birth but increases considerably in the cerebellum and the hippocampus (specifically the DG), where it is detectable at P10. GluN2A peaks during postnatal week 3, where the most dramatic increase occurs in the cerebellum. GluN2A is the predominant subunit in the adult brain. GluN2C is not expressed early, but increases dramatically in the cerebellum, where it remains restricted in the adult brain. Finally, GluN2D is expressed early on but continues being highly expressed in the thalamus and brainstem of the adult brain (Portera-Cailliau et al., 1996; Sans et al., 2000; Sheng et al., 1994; Wenzel et al., 1997).
1.6 Synaptic Plasticity alterations in FXS
1.6.1 The mGluR theory of FXS
Although FMRP regulates a number of different mRNAs and the expression of their protein products, the discovery of enhanced group I mGluR-dependent LTD in the Cornu
Ammonis (CA) 1 region of the Fmr1-/y mouse hippocampus has been a focal point in FXS research (Huber et al., 2002). This finding gave rise to ‘the mGluR theory of Fragile X’, which proposes that group I mGluRs and FMRP oppositely regulate translation of mRNA at the synapse. In the presence of FMRP, activation of mGluRs leads to protein synthesis, whereas FMRP suppresses protein synthesis (Bear et al., 2004). In FXS, the absence of FMRP leads to unregulated protein synthesis during activation of group I mGluRs resulting in excessive translation (Bear et al., 2004).
Additional evidence for this hypothesis is seen by the attenuation of enhanced mGluR-dependent LTD in Fmr1-/y mice by inhibition or reduction of group I mGluR signaling (Chuang
et al., 2005). Furthermore, reduction of one subtype of group I mGluR, mGluR5, decreases mGluR-mediated LTD to near normal levels (Dölen et al., 2007). Notably, Fmr1-/y mice possess deficits in mGluR-stimulated protein synthesis indicating that FMRP enables translation of synaptic proteins (Todd et al., 2003; Weiler et al., 2004). In contrast, these Fmr1-/y mice show
elevated protein synthesis rates along with elevated basal synaptic protein levels, indicating that there is a ceiling effect of protein synthesis where further stimulation of mGluRs does not lead to an increase in protein levels (Qin et al., 2005; Zalfa et al., 2003). Further proof of this idea comes from the observation that in the presence of a protein inhibitor, mGluR-dependent LTD persists (Nosyreva & Huber, 2006). This indicates that elevated synaptic protein levels are sufficient for the enhanced mGluR-mediated LTD observed in Fmr1-/y mice which, unlike wildtype (WT) controls, do not require de novo protein synthesis for mGluR-mediated LTD (Nosyreva & Huber, 2006).
Although the study of mGluRs has provided increasing evidence of synaptic dysfunction in FXS, there is now emerging evidence that dysfunction of another excitatory glutamate
receptor, the NMDAR, may play a role in FXS.
1.6.2 Involvement of NMDARs in FXS
Recent functional and molecular data has renewed the interest in studying the involvement of NMDARs in FXS. In the past, little focus has been given to NMDARs in FXS as previous studies have reported normal NMDAR-dependent LTP and LTD in the CA1 subfield of the hippocampus in Fmr1-/y mice (Godfraind et al., 1996; Huber et al., 2002). However, functional evidence shows that there is decreased NMDAR-mediated LTP and LTD in the DG subregion of the hippocampus in young adult Fmr1-/y mice, which can be attributed to an NMDAR
hypofunction (Eadie et al., 2012; Yun & Trommer, 2011). Interestingly, FMRP binds GluN1, GluN2A and GluN2B mRNAs, suggesting that the loss of FMRP, as in FXS patients and Fmr1-/y mice, may lead to dysregulated translation of these NMDAR subunits (Edbauer et al., 2010; Schütt et al., 2009). Confirming this hypothesis is an apparent reduction in the levels of GluN1, GluN2A and GluN2B in the medial prefrontal cortex of Fmr1-/y mice (Krueger et al., 2011). Despite this interesting finding, further regions implicated in FXS, such as the hippocampal subregions DG and CA, the CB and the prefrontal cortex (PFC) have yet to be investigated.
1.7 Brain Regions of Interest
The PFC, CB and the DG and CA subregions of the hippocampus of Fmr1-/y mice were selected for specific reasons in this study. The hippocampus is a region associated with learning and memory, and is composed of two subregions, the DG and CA (Bliss & Lomo, 1973; Scoville & Milner, 1957). FMRP is highly expressed in the DG, suggesting that a loss of this protein might have a drastic effect on this hippocampal subregion (Bakker et al., 2000; Feng et al., 1997b). Additionally, this region has just been associated with altered NMDAR-dependent LTP and LTD in 2-4 month old mice (Eadie et al., 2012; Yun & Trommer, 2011). On the other hand, the CA region (composed of the CA1-CA3 regions) has been highly studied in the FXS field, where a specific form of LTD dependent on the mGluR is altered in the CA1 of adult animals, but NMDAR-mediated synaptic plasticity is normal in this area (Godfraind et al., 1996; Huber et
al., 2002). However, juvenile mice show alterations in NMDAR-mediated LTP in the CA1,
suggesting an age-dependent alteration in NMDARs (Hu et al., 2008; Pilpel et al., 2009). Since FXS patients show alterations in learning and memory, the hippocampus is an important region to study with respect to FXS.
