Muscleblind-Like 2 in RNA Splicing Regulation and Disease

Permanent Link: http://ufdc.ufl.edu/UFE0043693/00001

Material Information

Title: Muscleblind-Like 2 in RNA Splicing Regulation and Disease
Physical Description: 1 online resource (136 p.)
Language: english
Creator: Charizanis, Konstantinos
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011


Subjects / Keywords: muscleblind
Genetics (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Myotonic dystrophy (DM) is a multi-systemic, late-onset and dominantly inherited neuromuscular disorder with characteristic disease features which include myotonia, muscle wasting, mental retardation, memory impairment, hypersomnolence, apathy and cerebral atrophy. DM is caused by two different microsatellite expansions (CUGexp, CCUGexp) in the untranslated regions of the DMPK and CNBP genes, respectively. Transcription of these repeats result in the synthesis of toxic RNAs which sequester, and functionally repress, the activities of proteins in the muscleblind-like splicing factor family. Previous studies have shown that Mbnl1 is essential for the switch from embryonic to adult splicing patterns in skeletal muscle and this switch is blocked in Mbnl1 knockout mice which develop muscle pathology characteristic of DM. In brain, loss of Mbnl1 does not recapitulate the splicing alterations seen in DM, suggesting a role for other Mbnl family members in disease pathogenesis. Based on this possibility, I hypothesized that Mbnl2 was the primary splicing factor sequestered by toxic RNAs in the DM brain and that loss of this factor underlies the neurological features associated with DM. To address this hypothesis, Mbnl2 knockout mice (Mbnl2?E2/?E2) were generated and examined for DM-relevant phenotypes as well as aberrant alternative splicing. Using these Mbnl2 knockouts and splicing microarrays combined with RNA-seq, the alternative exons mis-spliced following loss of Mbnl2 were identified. Using this splicing information, I discovered that Mbnl2 promoted the switch from embryonic to adult alternative splicing during postnatal brain development. Using high throughput sequencing and crosslinking-immunoprecipitation, the direct binding sites for Mbnl2 were identified and confirmed that Mbnl2 binds preferentially to UGCU motifs to regulate alternative splicing. Surprisingly, Mbnl2 knockouts showed impaired spatial memory formation and increased neuronal excitability which led to seizures upon treatment with a GABA antagonist. Epilepsy susceptibility has not been extensively reported in DM patients, but this result raises the issue of appropriate drug treatment and prolonged use. Importantly, Mbnl2 knockout mice are an important new resource for investigations focused on characterizing the molecular events which regulate normal brain development and DM-relevant brain pathology and a useful model for ongoing drug development.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Konstantinos Charizanis.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Swanson, Maurice S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043693:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043693/00001

Material Information

Title: Muscleblind-Like 2 in RNA Splicing Regulation and Disease
Physical Description: 1 online resource (136 p.)
Language: english
Creator: Charizanis, Konstantinos
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011


Subjects / Keywords: muscleblind
Genetics (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Myotonic dystrophy (DM) is a multi-systemic, late-onset and dominantly inherited neuromuscular disorder with characteristic disease features which include myotonia, muscle wasting, mental retardation, memory impairment, hypersomnolence, apathy and cerebral atrophy. DM is caused by two different microsatellite expansions (CUGexp, CCUGexp) in the untranslated regions of the DMPK and CNBP genes, respectively. Transcription of these repeats result in the synthesis of toxic RNAs which sequester, and functionally repress, the activities of proteins in the muscleblind-like splicing factor family. Previous studies have shown that Mbnl1 is essential for the switch from embryonic to adult splicing patterns in skeletal muscle and this switch is blocked in Mbnl1 knockout mice which develop muscle pathology characteristic of DM. In brain, loss of Mbnl1 does not recapitulate the splicing alterations seen in DM, suggesting a role for other Mbnl family members in disease pathogenesis. Based on this possibility, I hypothesized that Mbnl2 was the primary splicing factor sequestered by toxic RNAs in the DM brain and that loss of this factor underlies the neurological features associated with DM. To address this hypothesis, Mbnl2 knockout mice (Mbnl2?E2/?E2) were generated and examined for DM-relevant phenotypes as well as aberrant alternative splicing. Using these Mbnl2 knockouts and splicing microarrays combined with RNA-seq, the alternative exons mis-spliced following loss of Mbnl2 were identified. Using this splicing information, I discovered that Mbnl2 promoted the switch from embryonic to adult alternative splicing during postnatal brain development. Using high throughput sequencing and crosslinking-immunoprecipitation, the direct binding sites for Mbnl2 were identified and confirmed that Mbnl2 binds preferentially to UGCU motifs to regulate alternative splicing. Surprisingly, Mbnl2 knockouts showed impaired spatial memory formation and increased neuronal excitability which led to seizures upon treatment with a GABA antagonist. Epilepsy susceptibility has not been extensively reported in DM patients, but this result raises the issue of appropriate drug treatment and prolonged use. Importantly, Mbnl2 knockout mice are an important new resource for investigations focused on characterizing the molecular events which regulate normal brain development and DM-relevant brain pathology and a useful model for ongoing drug development.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Konstantinos Charizanis.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Swanson, Maurice S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043693:00001

This item has the following downloads:

Full Text




2 2011 Konstantinos Charizanis


3 To my family I left


4 ACKNOWLEDGMENTS I would like to thank my parents Ioannis and Paraskevi Charizanis and my sister Rozita Charizani s for the unconditional love, support and unbiased guidance throughout my life, a heritage that will always follow me anywhere I go. I owe my life, career and any future achievements to them. I would also like to thank my math teacher M r. Skordas who is not any longer with us, f or emphasizing the virtue of being humble and for challenging me in life mentally and socially. A special thank you to Dr. Swanson who cherished my scientific curiosity and struggled with me these 6 years of my P hD trying to sculpture and refine my thinkin g process and guide me through the pitfalls of lab work, ethics and beyond. I recognize him as my scientific father and will always seek his advice and help. Also, I would like to recognize all the labmates, Michael Poulos, Jason ang Y u ng Lee, Marina Scotti, and Marianne Goodwin, who had an input in this thesis either by executing experiments or by generating ideas My dear committee members Dr s Alfred Lewin, Rolf Renne and Harry Nick for providing constructive criticism on my exp erimental design and goals, as well as the numerous suggestions on my presentation skills and future directions. Lastly I would like to acknowledge our collaborators Lily Shiue, Man ny Ares, Yuan Yuan, Bob Darnell, Chaolin Zhang Guangbin Xia, Tetsuo Ashi zawa Takashi Kimura and Brent Clark who made this work possible


5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 Myotonic Dystrophy ................................ ................................ ................................ 16 Genetics ................................ ................................ ................................ ........... 17 Clinical Features of DM ................................ ................................ .................... 18 Myotonia ................................ ................................ ................................ .... 18 Muscle wasting ................................ ................................ .......................... 19 Cataracts ................................ ................................ ................................ .... 19 Heart problems ................................ ................................ .......................... 19 Endocrine system ................................ ................................ ...................... 20 Gastrointestinal track ................................ ................................ ................. 21 Mature balding and calcifying epithelioma ................................ ................. 21 Brain abnormalities ................................ ................................ .................... 21 The Molecular Etiology of Myotonic Dystrophy ................................ ................. 25 Muscleblind Like Protein Family ................................ ................................ ............. 27 Mouse DM Models ................................ ................................ ................................ .. 31 HSA LR ................................ ................................ ................................ ............... 32 Mbnl1 3/ 3 ................................ ................................ ................................ ...... 32 CELF1 Overexpression ................................ ................................ .................... 33 DM300, DM328XL and DMSXL ................................ ................................ ........ 33 M bnl2 Gene Trap Mouse Models ................................ ................................ ..... 34 Discussion ................................ ................................ ................................ .............. 35 2 MBNL2 IS EXPRESSED IN MURINE ADULT BRAIN BUT NOT MUSCLE ............ 38 Results ................................ ................................ ................................ .................... 38 Mbnl2 Is Widely Expressed In the Murine Brain ................................ ............... 38 Generation of an Mbnl2 Speci fic Antibody ................................ ........................ 39 Isoform analysis ................................ ................................ ......................... 40 Antibody production and testing ................................ ................................ 40 Generation of an Mbnl2 Knockout Mouse Model ................................ .............. 42 Generation of the Mbnl2 ES cell targeting vector ................................ ....... 42 ES cell targeting ................................ ................................ ......................... 44 ES cell blastocyst injections ................................ ................................ ....... 45


6 Mbnl2 conditional line generation and maintenance ................................ .. 45 Mbnl2 2/ 2 constitutive knockout mouse generation ................................ 46 Mbnl2 Protein Is Expressed and Mostly Nuclear In Brain, but Is Expressed at a Low Level in Muscle of Adult Mice ................................ ......................... 46 Discussion ................................ ................................ ................................ .............. 48 3 ANALYSIS OF THE Mbnl 2/ 2 KNOCKOUT MOUSE MODEL ............................. 60 Results ................................ ................................ ................................ .................... 61 Mbnl2 2/ 2 Mice Show Postnatal Growth Retardation ................................ .... 61 No Aberrant Muscle Pathology Due to Mbnl2 Loss ................................ .......... 61 Mbnl2 Mice Do Not Show Muscle Specific Chloride Channel Loss ........ 62 Mbnl2 2/ 2 Mice Do Not Recapitulate Aberrant Splicing of Clcn1 .................. 62 Loss of Mbnl2 Does Not Lead to an Overt Hippocampal Pathology ................. 63 Mbnl2 2/ 2 Mice Show Impaired Spatial Learning and Memory ..................... 63 Loss of Mbnl2 Results in Increased Seizure Susceptibility ............................... 65 Discussion ................................ ................................ ................................ .............. 68 4 MBNL2 REGULATES ALTERNATIVE SPLICING IN THE BRAIN ......................... 84 Results ................................ ................................ ................................ .................... 84 Alternative Splicing Dysregulation in Mbnl2 Knockout Brain ............................ 84 Intronic Enrichment of YGCY Clusters in Mbnl2 Knockout Mis Regulated Exons ................................ ................................ ................................ ............ 85 Direct Binding Sites for Mbnl2 on Mis Regulated Targets ................................ 86 Discussion ................................ ................................ ................................ .............. 87 5 CONCLUDING REMARKS AND FUTURE DIRECTIONS ................................ ...... 98 6 MATERIALS AND METHODS ................................ ................................ .............. 104 Lac Z Staining ................................ ................................ ................................ ....... 104 Generation of an Mbnl2 Polyclonal Antibody ................................ ........................ 104 Plasmid Transfections ................................ ................................ .................... 104 Western Blotting ................................ ................................ ............................. 105 Mbnl2 2/ 2 Mouse Generation ................................ ................................ ............ 106 ES Cell Targeting Construct ................................ ................................ ........... 106 ES Cell Targeting ................................ ................................ ........................... 108 Mating Scheme and Genotyping ................................ ................................ .... 109 RNA Analysis of Mbnl2 2 / 2 Mice ................................ ................................ ....... 109 Protein Expression Analysis of Mbnl2 2 / 2 Mice ................................ ................ 110 Muscle Immunofluorescence and Histology ................................ .......................... 110 Brain Immunofluorescence ................................ ................................ ................... 111 Nissl Staining ................................ ................................ ................................ ........ 111 Morris Water Maze Test ................................ ................................ ........................ 112 Animals ................................ ................................ ................................ ........... 112 Training ................................ ................................ ................................ .......... 112


7 Data Acquisition and Analysis ................................ ................................ ........ 113 PTZ Seizure Susceptibility Test ................................ ................................ ............ 114 Splicing Microarray ................................ ................................ ............................... 114 RT PCR Splicing Analysis of Mbnl2 Targets ................................ ......................... 115 HITS CLIP ................................ ................................ ................................ ............ 115 APPENDIX: MICROARRAY SPLICING RESULTS ................................ ..................... 118 LIST OF REFERENCES ................................ ................................ ............................. 121 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 136


8 LIST OF TABLES Table page 6 1 Mbnl2 target gene RT PCR analysis primer sets ................................ .............. 117


9 LIST OF FIGURES Figure page 1 1 Different models of Mbnl2 deficient mice show diverse phenotypes .................. 37 2 1 Mbnl2 is highly transcribed in the mouse brain ................................ ................... 51 2 2 Extensive alternative splicing of Mbnl2 occurs in the C terminus of the protein 52 2 3 Mbnl2 polyclonal anti body does not cross react with the other Mbnl protein isoforms ................................ ................................ ................................ ............. 53 2 4 The anti Mbnl2 polyclonal antibody d id not cross react with other endogenous expressed proteins in cell culture ................................ .................. 54 2 5 Generation of Mbnl2 KO targeting cons truct ................................ ..................... 55 2 6 Mbnl2 2/ 2 mice lack expression of exon 2 ................................ ...................... 56 2 7 Mbnl2 is highly expressed in mouse brain but not in muscle .............................. 57 2 8 Mbnl2 shows nuclear and cytoplasmic localization in cells of the frontal cortex ................................ ................................ ................................ .................. 58 2 9 Mbnl2 is mostly nuclear in cells of the hippocampal formation .......................... 59 3 1 Growth retardation in Mbnl2 2/ 2 mice ................................ ............................ 71 3 2 Mbnl2 loss does not le ad to muscle histopathology ................................ ........... 72 3 3 Mbnl2 loss does not lead to chloride channel loss due to Clcn1 aberrant splicing ................................ ................................ ................................ .............. 73 3 4 No overt histological changes in the hippocampal formation of Mbnl2 2/ 2 mic e ................................ ................................ ................................ ................... 74 3 5 Latency in learning and memory formation in the Morris water maze test .......... 75 3 6 Quadrant occupancy in the Morris water maze test ................................ ............ 76 3 7 Decreased spatial memory and preci sion in Mbnl2 2/ 2 mice ......................... 77 3 8 No significant difference in total path swim length between WT and Mbnl2 2/ 2 mice ................................ ................................ .............................. 78 3 9 No significant difference in swimming speed between WT and Mbnl2 2/ 2 mice ................................ ................................ ................................ ................... 79


10 3 10 Mbnl2 2/ 2 mice show increased thigmotaxis ................................ .................. 80 3 11 Latency to seiz ure onset is reduced in both Mbnl2 2/ 2 and Mbnl2 +/ 2 mice 81 3 12 Severity of convulsion is increased on both in Mbnl2 2/ 2 and Mbnl2 +/ 2 mice ................................ ................................ ................................ ................... 82 3 13 Reduced time to peak severity in both Mbnl2 2/ 2 and Mbnl2 +/ 2 mice .......... 83 4 1 Mbnl2 regulates neonatal to adult alternative splici ng changes in the hippocampus ................................ ................................ ................................ ..... 93 4 2 RNA seq and microa rray data show possible learning/memory and epileptic defects in Mbnl2 2/ 2 mice ................................ ................................ .............. 94 4 3 Mbnl2 regulates exon exclusion in the hippocampus for exons that show YG CY enrichment in the upstream introns ................................ ........................ 95 4 4 Mbnl2 RNA splicing map ................................ ................................ ................... 96 4 5 Mbnl2 binds directly to UGCU clusters ................................ .............................. 97 5 1 Functional diversion and tissue distribution among the three Mbnl proteins. .... 103 A 1 Top targets of splicing sensitive microarray with sepscore 2.38 1.00 ............... 118


11 LIST OF ABBREVIATIONS AV atrioventricular APP amyloid precurs or protein C3 complement 3 locus cAMP cyclic adenosin e monophosphate CDM congenital myotonic dystrophy CLCN1 muscle specific chloride channel CNS central nervous system cTNT cardiac troponin T CUGBP1 CUG binding protein 1 DM myotonic dystrophy DMPK dystrop hica myotonica protein kinase dsRNA double stranded RNA FAXTAS fragile x tremor associated syndrome FISH fluorescence in situ hybridization GI gastrointestinal GT gene trap H&E hematoxylin and eosin HAS human skeletal actin HSA LR human skeletal actin long repeats IR Insulin receptor Kb kilobase LTP long term potentiation Mbnl muscleblind like MEF mouse embryonic fibroblast


12 MRI magnetic resonance imaging mRNA messenger RNA NMDAR1 N Methyl D Aspartate Receptor 1 OCT optimal cutting temperature PKC protein kin ase C POMA paraneoplastic opsoclonus myoclonus ataxia PTB polypyrimidine tract binding protein PTZ pentylene tetrazol RRM RNA recognition motif SF1 splicing factor 1 ssRNA single stranded RNA TA tibialis anterior TFIIIA transcription factor III A TK Herpes Simplex Virus thymidine kinase gene UTR untranslated region


13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MUSCLEBLIND LIKE 2 IN R NA SPLICING REGULATION AND DISEASE By Konstantinos Charizanis December 2011 Chair: Maurice Swanson Major: Medical Sciences Genetics Myotonic dystrophy (DM) is a multi systemic, late onset and dominantly inherited neuromuscular disorder with characteri stic disease features which include myotonia, muscle wasting, mental retardation, memory impairment, hypersomnolence, apathy and cerebral atrophy DM is caused by two different microsatellite expansions (CUG exp CCUG exp ) in the untranslated regions of the DMPK and CNBP genes, respectively. Transcription of these repeats result in the synthesis of toxic RNAs which sequester, and functionally repress, the activities of proteins in the muscleblind like splicing factor family. Previous studies have shown that Mbnl1 is essential for the switch from embryonic to adult splicing patterns in skeletal muscle and this switch is blocked in Mbnl1 knockout mice which develop muscle pathology characteristic of DM In brain, loss of Mbnl1 does not recapitulate the splicing alterations seen in DM, suggesting a role for other Mbnl family members in disease pathogenesis. Based on this possibility, I hypothesized that Mbnl2 was the primary splicing factor sequestered by toxic RNAs in the DM brain and that loss of this factor un derlies the neurological features associated with DM. To address this hypothesis, Mbnl2 knockout mice ( Mbnl2 ) were generated and examined for DM relevant phenotypes as well as aberrant alternative


14 splicing. Using these Mbnl2 knockouts and splicing m icroarrays combined with RNA seq, the alternative exons mis spliced following loss of Mbnl2 were identified. Using this splicing information, I discovered that Mbnl2 promoted the switch from embryonic to adult alternative splicing during postnatal brain de velopment. Using high throughput sequencing and crosslinking immunoprecipitation the direct binding sites for Mbnl2 were identified and confirmed that Mbnl2 binds preferentially to UGCU motifs to regulate alternative splicing. Surprisingly, Mbnl2 knockout s showed impaired spatial memory formation and increased neuronal excitability which led to seizures upon treatment with a GABA antagonist. Epilepsy susceptibility has not been extensively reported in DM patients, but this result raises the issue of approp riate drug treatment and prolonged use. Importantly, Mbnl2 knockout mice are an important new resource for investigations focused on characterizing the molecular events which regulate normal brain development and DM relevant brain pathology and a useful mo d el for ongoing drug development.


15 CHAPTER 1 INTRODUCTION In the study of genetic disorders it is common to identify mutations that result in a non functional protein product (loss of function model) or a protein which has gained a deleterious function ( gain of function model). In both cases the affected gene leads to an abnormal phenotype usually associated with the normal function of the encoded protein. Many different types of mutations have been linked to known genetic disorders These mutations incl ude point mutations, deletions, insertions, duplications, inversions and microsatellite expansions A variation of the above scenario was recently introduced during studies on the molecular basis of m yotonic dystrophy ( dystrophia myotonica, DM), a disease caused by microsatellite expansions in untranslated regions of two unrelated genes, DMPK and CNBP In DM the expansions do not affect the function of the mut ant gene s or proximal genes, but rather act in an atypical fashion, now referred to as an RNA gain of function mutation According to this dominant negative RNA model, mutant RNAs accumulate in DM nuclei and sequester nuclear factors and impair a critical cellular pathway The unique nature of this pathogenic mechanism is highlighted by the phenotypic outcome of the se mutation s since this disease is multi systemic and leads to atrophy of a number of tissues The discovery of the RNA gain of function model has provided a new dimension to ongoing research on other diseases caused by nucleotide repeat exp ansions including sp inocerebellar ataxia s fragile X associated tremor/ataxia syndrome (FXTAS) and a recently identified form of amyloid lateral sclerosis ( ALS ) linked to chromosome 9


16 In this Introduction, I begin by discussing the main clinical features of myotonic dystrophy and the molecular events and mechanism that have been i mplicated in the disease progression Then I will focus on the current mouse models that have been generated to study the disease and point out advantages, disadvantages and DM cl inical features that are not phenocopied by these models. Finally I conclude by stating and justifying the hypothesis that I am addressing in this thesis. Myotonic Dystrophy Myotonic d ystrophy ( dystrophia myotonica or DM) is a common form of muscular dyst rophy with a n unusual mechanism of pathogenesis that affects more than 1:8000 adults around the world. DM was first described more than a century ago as a multisystemic disease with the name myotonia atrophica (Batten and Gibb, 1909) describing patients that developed different aspects of the disease including myotonia, muscle weakness and wasting, paralysi s of vocal cords (Adie, 1923; Fox, 1910) cataracts, mental defects including hypersomnol ence and cardiac defects (Evans, 1944) From the early days of the disease discovery it was suspected that DM is a genetically transmitted late onset disease that shifts i nto a more severe form with an earlier onset during vertical genetic transmiss ion, a phenomenon that is termed genetic anticipation. There were cases described in 1910 w h ere if they placed the child and the father next to each other, the diseased child sh owed similar symptoms and severity at an earlier age a clear indication of genetic anticipation. Due to the anticipation phenomenon the disease eventually appears in a pedigree as a congenital form of myotonic dystrophy (CDM) that affects newborns. To da te, DM is divided in three different types with a similar


17 phenotype and possibly a similar molecular pathogenesis, DM type 1 ( DM1 ), DM type 2 ( DM2 ) and CDM. Genetics The genetic locus responsible for DM1 was the subject of intense scientific research since the 1970s when scientists were focused on discovering a marker locus that was linked to the disease phenotype (D M locus) for indirect prenatal diagnosis. In 1971 J.H. Renwick and D.R. Bolling found that they could predict transmiss ion of the disease wit h a 92% chance, by a three locus linkage analysis of the D M locus, the ABH secretor locus and the Lutheran blood group locus (Renwick and Bolling, 1971) A decade later linkage of DM and complement 3 locus (C3) indicated that the locus that is responsible for the DM phenotype resides on chromosome 19 (Roses et al., 1986) Six years later (1992) a microsatellite repeat expansion was ident ified and correlated with the disease by several independent labs (Brook et al., 1992; Mahadevan et al., 1992) and the gene that carries the microsatellite expansion was identified to be a member of a protein serine /threonine kinase family. This finding in addition to th e finding that a CGG expansion is correlated with Fragile X syndrome contributed to the identification of the molecular etiology of a number of diseases that have been found since then to be caused by unstable microsatellite repeats such as Huntington disease, six spinocerebellar ataxias, FXTAS and spinobulbar muscular atrophy (SBMA) DM1 is now known to be an autosomal dominant late onset neuromuscular disorder that is caused by the expansion of a CTG microsatellite repeat located in the untranslated region (UTR) of the gene DMPK (dystrophica myotonica protein kinase) located on chromosome 19q13.3. The onset of the disease correlates with an expansion of the CTG repeats above a certain threshold, where normal individuals have


18 5 37 repeats individuals with mild DM phenotype that will develop mild myoton i a and cataracts carry 50 1 50 and individuals with classical DM phenotype carry 100 1000 repeats The severity and onset of the phenotype generall y correlates with repeat number. Clinical Features o f DM The most prevalent clinical features that DM1 patients develop are discussed below. Myotonia Myotonia is defined as the inability of muscle relaxation after a voluntary contraction. It is the hallmar k clinical feature of DM patients and an overt sign that allowed physicians to characterize and categorize the disease. Electric stimulation was used extensively to reveal prolonged involuntary contraction of muscles and a failure of muscle relaxation that was independent of a compromised nervous system. In a 1909 case report of a patient who suffered from DM the clinic i an reported : There is very marked reduction in irritability to the faradaic current amounting in the biceps and hand muscles almost to ext inction. In the long flexors of the fingers there is an imperfect myotonic reaction shown both on volitional movement and electric stimulation. The patient is able to relax his grasp at once, but only to an extent that leaves the fingers somewhat in the po sition of tetany; from this they gradually relax. On faradaic stimulation the same occurs. This case report not only describes myotonia characteristic of DM but also shows evidence that this patient was probably affected by myotonic dystrophy type 1 since the affected muscle groups were more distal to body axis. Since 1992, a cause of myotonia in muscle diseases like myotonia congenital was known to result from mutations in the muscle specific chloride channel ClC 1 or CLCN1. C L CN1 mutations caused either d ecreased levels of ClC 1 or alterations in the coding


19 sequence and the synthesis of channels with altered sensitivity. In 2002 myotonia in DM patients was found to be caused by CLCN1 gene defects (Mankodi et al., 20 02) However, in DM the CLCN1 gene is not directly affected and instead the splicing of CLCN1 pre mRNA is altered which leads to a premature termination codon, nonsense mediated decay of the transcript and decreased levels of the chloride channel. Muscle wasting Muscle wasting and weakness is the second most prominent clinical feature of DM. Specific muscle groups atrophy and thus cannot generate sufficient force required for certain tasks. Unlike DM2 where proximal muscles (e.g., hip flexors) are affected DM1 associated weakness and wasting is most prominent in distal muscles and also affects facial muscles such as sternomastoid as well as eyelid muscles which results in the characteristic facial weakness and eyelid ptosis respectively The molecular mec hanism of muscle wasting has not yet been identified. Cataracts Most DM1 patients also suffer from an unusual type of cataract (Fearnsides, 1915) Cataracts result in a loss of transparency of the eye lens but in the case of DM the cataracts are dust like opacities in the subcapsular region of the lens. T he gene ( s ) and molecular events that are responsible for th is eye phenotype are not understood. Heart problems As men tioned before the most prominent clinical feature of DM1 is myotonia and muscle wasting, but there is a group of clinical features unrelated to the skeletal muscle phenotype with severe consequences to DM1 patients. H eart conduction defects significantly a ffect the quality of life of DM patients and the severity of cardiac complications correlate with the length of the microsatellite repeat expansion. Cardiac


20 arrhythmias are common in this patient population and may result in cardiac arrest and sudden death (Bache and Sarosi, 1968) A peculiarity of DM heart disease is that the conduction system is selectively affected with first degree atrioventricular block (AV block) and PR prolongation often accompanied by cardiomyopathy. The cellular / molecular mechanism underlying this defect is un known but there is growing evidence that mis splicing of several genes, possible the cardiac troponin T gene TNNT2 is the main mechanism (Philips et al., 1998) TNNT2 mutations have been implicated in reduced heart efficiency and cardiomyopathy (Thierfelder et al., 1994 ) and more interestingly a shift in splicing isoforms similar to the shift observed in DM patients has been correlated with changes in myofilament calcium sensitivity and heart failure (Townsend et al., 1995) End ocrine system E ndocrine system defects are also present in DM Patients with DM type 1 or type 2 have been reported to develop hyperinsulinism (Tevaarwerk and Hudson, 1977) diabetes, testicular atrophy increased levels of luteinizing hormone (LT), estradiol and follicl e stimulating hormone (FSH), decreased levels of testosterone, dihydrotestosterone (DHT) (Mastrogiacomo et al., 1996) as well as some abnormalities in growth hormone (GH) secretion with lower levels of GH during sleep compared to control individuals (Barreca et al., 1980) The endocrine system abnormalities show decreased penetrance with 80% for testicular atrophy and 70% for muscle specific insulin resistance (Barbosa et al., 1974; Moxley et al., 1978) The muscle specific insulin r esistance is further supported by the finding that in skeletal muscle of DM1 patients alternative splicing of the insulin receptor pre mRNA is aberrantly regulated resulting in a non muscle isoform with lower signaling capabilities (Savkur et al., 2001)


21 Gastrointestinal track Gastrointestinal complaints from patients with DM are common. The symptoms re sult from the compromised gastrointestinal musculature. Patients suffer impaired low er bowel function leading to constipation, and delayed gastric emptying (Ronnblom et al., 2002) The majority ( 55% ) of DM1 patients suffer from abd ominal pain, 35% have upper GI tract movement inconsistencies leading to emesis, 33% have chronic or episodic diarrhea and 45% suffer from dysphagia (inability or difficulty to swallow) that rarely leads to parenteral feeding. In the DM patient community 25% of the patients find the GI track symptoms to have a profound effect on their quality of life (Ronnblom et al., 1996) M ature balding and calcifying epithelioma The first association of DM and calcifying epithelioma or pilomatrixomata was first described by Cantwell and Reed (Cantwell and Reed, 1965) and is strongly associated with DM1 (Geh and Moss, 1999) It is one of the most common pediatric superficial tumors with an early ons et indeed prior to the onset of other typical DM features. DM patients also develop alopecia or pattern baldness that affects men more than women. The etiology of the balding pattern is believed to be androgenic as these patients suffer from testicular at rophy and decreased levels of circulating testosterone, as well as defects in androgen receptors in the hair follicles. Brain abnormalities Although DM patients develop a variety of different symptoms the main concern is abnormal brain function, which dim inish es quality of life. The m ost common complaint s are altered sleep patterns early in disease progression and cognitive function due to cerebral atrophy at later stages P atients suffer from executive function deficits, age


22 related decline of frontal and temporal cognitive functions excessive daytime sleepiness ( hypersomnolence ) and memory problems as well as severe apathy (Meola et al., 2003; Modoni et al., 2004) T he central nervous system involvement in DM incl udes personality alterations, such as avoidant personality (Meola et al., 2003) obsessive compulsive, passive aggressive and schizotypic traits (Delaporte, 1998) as well as severe impairment of general intelligence and verbal fluency (Abe et al., 1994) Patients suffer from visuospatial, executive, arithmetic and attention ability deficits and score low on the general IQ test (converted Wechsler Adult Intelligence Scale Revised) (Turnpenny et al., 1994; Winblad et al., 2006) H ypersomnolence has not been attributed to hypercapnia (an increase in blood CO 2 due to hypoventilation ) (van der Meche et al., 1994) it is not cataplectic (Phillips et al., 1999) and is not described as narcolepsy, even though anti narcoleptic drugs such as Modafi nil have been employed and successfully reduce the abnormal sleep symptoms (Talbot et al., 2003) P atients with DM 1 fall asleep not due to loss of interest, but to their inability to remain engaged. O ne reason for DM sleep disorders may be a general atrop hy of the corpus callosum (Giubilei et al., 1999) MRI studies have revealed an extensive loss of brain matter, including both white and gray matter loss as well as cortical atrophy (Damian et al., 1993; Hashimoto et al., 1995a; Hashimoto et al., 1995b) In addition to brain tissue and cell body loss, patients have developed neuronal inclusion bodies, decreased myelin sheathing and increased neurofibrillary tangles loca ted mostly in the limbic system, brainstem, hippocampus, entorhinal cortex and temporal lobes (Ono et al., 1987; Oyamada et al., 2006)


23 In contrast to muscle, the molecular basis of brain related defects is mostly u nknown. It is well established that the pathogenic DMPK RNA in the brain form foci similar to the nuclear RNA foci present in myonuclei T hese foci have been detected by RNA FISH in a wide range of different regions of the brain and cell types, such as the dentate gyrus, hippocampus, cerebral cortex, thalamus, substantia nigra and the brainstem. The density and size of the RNA foci correlates with the size of the CTG repeat expansion in DM1 patients In the DM brain the molecular events that lead to clinic al phenotypes may be simpler than in muscle since there is no evidence for sequestration of transcription factors such as SP1, STAT1, STAT2 and the retinoic acid receptor gamma subunit. The only factors that have been found to bind strongly to CUG RNA foc i in the brain are the two members of the MBNL family, MBNL1 and MBNL2 suggesting a splicing mis regulation etiology that explains DM related brain deficits (Jiang et al., 2004) Sequestration of Mbnl1 and Mbnl2 in RNA foci is accompanied by a few well documented aberrant splicing events. In the current literature there are 14 alternative splicing events identified to be mis regulated in DM patients, correspond ing to 10 different genes throughout the tissues (Ranum and Cooper, 2006) As in muscle, ab errant splicing disrupts the developmental hierarchy in brain resulting in a reversion to an embryonic state. Despite the heterogeneity of the brain, only three genes have been found to be mis regulated so far, the N methyl D aspartate receptor 1 ( NMDAR1 ), the amyloid precursor protein ( APP ) and Tau ( MAPT ) (Jiang et al., 2004; Leroy et al., 2006a; Sergeant et al., 2001) The association of mis regulated events and disease is


24 yet to be established despite the fact tha t these genes have been already implicated in other brain diseases Exclusion of exon 2, 3 and 10 from MAPT transcripts generates alternative Tau isoforms with altered function. Exclusion of exon 2 results in an MAPT in which the N terminal domain interact s with the axonal membrane and possibly stabiliz es of microtubules as it was found to be enriched at the membrane of growth cones and distal axons of hippocampal neurons (Brandt et al., 1995) A similar role has been attributed to exon 10 that encodes an additional microtubule binding domain. MAPT exon 2 and 10 splicing has been reported to be regulated by an exonic silencer and enhancer respe ctively by providing binding sites for specific splicing factor s (Wang et al., 2005) Another very interesting exon whose splicing is mis regulated in DM patients is exon 6 that has been reported to lead to formati on of neurofibrillary tangles and axonopathy and also corresponds to the fetal splicing pattern (Ishihara et al., 2001; Leroy et al., 2006b) A gene that is mis regulated that may be linked to a DM clinical feature is NMDAR1. NMDAR1 is a glutaminergic receptor in the brain and is required for normal long term potentiation (LTP) in the hippocampus, an event that precedes memory formation and learning. (Brigman et al., 2010; Toneg awa et al., 1996) In patients with myotonic dystrophy there is a substantial increase of NMDAR1 exon 5 inclusion in the temporal cortex, an event that dictates alterations in gating, subcellular localization of the receptor and pharmacological behavior. been suggested as the main etiology of the memory impairment observed in DM (Traynelis et al., 1995)


25 The Molecular Etiology of Myotonic Dystrophy F or more than a decade we have known that the DMPK gene with a triplet repeat expansion is responsible for the autosomal dominant vertical transmission of DM1. The generation of polyCUG transgen ic mice, which carried a few hundred CTG repeats in a different gene (human skeletal actin, HSA), established the notion that the disease is caused by the repeat expansion and not by altered DMPK expression shown by the lack of classical DM manifestations in D mpk / mice (Mankodi et al., 2000) (Reddy et al., 1996) The HSA long repeat ( HSA LR ) mice showed severe myotonia and most of the muscle histopathological aspects of DM musc le, including centralized nuclei, split and pycnotic fibers, as well as fibrosis. These pathological features were only present in skeletal muscle because the HSA transgene is only expressed in this tissue. The fact that the se repeats cause a DM associated muscle phenotype when expressed in a different mRNA context suggested that DM pathogenesis results from the expression of the repeats alone and altered expression of the host gene is not an important aspect of the etiology M any hypothes e s were proposed t o explain how the se repeat expansions could lead to a multi systemic disease but a few elegant experiments established the RNA gain of function mechanism. In earlier studies, RNA FISH experiments using a hybridization probe against CTG repeats showed that the se repeats and its associated transcript are sequestered in distinct nuclear foci in DM patient cells and tissues (Mankodi et al., 2001) Label transfer and filter binding experiments showed that these RNA speci es have the capability of binding with high affinity to a specific set of muscleblind like or MBNL proteins (Miller et al., 2000; Yuan et al., 2007) These proteins belong to a family of proteins first described in a mutant line of Drosophila melanogaster (Miller et al., 2000) In this mutant line, the flies showed a n unusual late


26 developmental phenotype with defects in muscle and eye terminal differentiation, thus the gene wa s named Muscleblind or mbl (Artero et al., 1998; Begemann et al., 1997) Knocking out Mbnl1 in mice results in myotonia and muscle histopathology similar to DM, suggesting that loss of MBNL protein function by the C UG exp RNA gain of function and sequestration model is the main molecular event that leads to DM. Since the establishment of the RNA gain of function and protein sequestration model, it has been shown that loss o f function of the Mbnl1 protein leads to a m is regulation of alternative splicing and recapitulation of certain aspects of DM in a mouse Mbnl1 knockout model (Kanadia et al., 2003a) These clinical aspects of myotonic dystrophy were rescued in a classical g ene therapy/complementation experiment in HSA LR mice, substantiating the Mbnl loss of function theory (Kanadia et al., 2006) The finding that Mbnl1 plays a major role in alternative splicing regulation by switchin g from an embryonic to an adult splicing pattern indicated that DM is a spliceopathy Using a different approach, but involving a similar molecular mechanism, considerable research has been done on another factor that binds to CUG repeats and is mis regula ted in patients with DM. The CUG binding protein (CUGBP1 /CELF1 ) is an alternative splicing factor that was identified together with additional proteins from the CELF family of RNA binding proteins These proteins have a high affinity for UG dinucleotide re peats and CELF1 was the first factor to be considered as disease factor involved in DM1 pathogenesis (Timchenko et al., 1996) Although CELF1 binds to CUG repeats (Timchenko et al., 2001) this protein is not sequestered in RNA foci Surprisingly, CELF1 steady state levels are elevated in DM1 tissues and cells due to PKC mediated hyperphosphorylation of the CELF1 protein. This hyperphosphorylation


27 is due to activa tion of the PKC pathway via an unknown mechanism (Kuyumcu Martinez et al., 2007) Studies to date have shown that CELF1 acts as an antagonist to M BNL 1 since it promot es embryonic splicing pattern s In a mouse CELF1 overexpression model, myogenesis is inhibited as is the myofiber type switch from fast glycolytic to fast oxidative (Timchenko et al., 2004) The indirect relationship between CUG repeat expansion and CELF1 hyperac tivity supports the hypothesis that Mbnl loss of function play s an important role in the pathogenesis of DM. Muscleblind Like Protein Family The MBNL family consists of three known paralogous proteins with more than 80% sequence identity: MBNL1, MBNL2 and MBNL3. The proteins consist of an N terminus that contains two pairs of zinc finger like, or CCCH, motifs and a C terminus that is involved in homodimerization and other protein protein interactions (Yuan et al., 200 7) Proteins that bind RNA possess a variety of structural domains that have the ability to either bind to single stranded (ss) RNA by recognition of a n RNA sequence or to double stranded (ds) RNA or to other more complex RNA conformations by recognizing a c ombination of RNA sequence and higher order structures One of the most common RNA binding domains and one of the first to be identified i s the RNA Recognition Motif (RRM ) It consists of four beta strands packed between two alpha helixes ( fold ) and binds to ss RNA (Query et al., 1989) most likely via strand RNA interactions. Nova 1, is a well characteri zed RNA binding protein that binds to a UCAU tetranucleotide via its 3 rd KH domain (Buckanovich and Darnell, 1997) and regulate s alternative splicing (Jensen et al., 2000a) Nova 1 is implicated in paraneoplastic opsoclonus myoclonus ataxia (POMA) (Buckanovich et al., 1996) and is expressed


28 exclusive ly in specific regions of the brain (Buckanovich et al., 1993; Buckanovich et al., 1996) Another characteristic RNA binding domain that has been implicated in ds RNA binding is the zinc finger (ZF), or C2H2, motif. TFIIIA was the first protein to be described which consisted of repetitive ZFs that could bind selectively to DNA and regulate gene expression in Xenopus oocytes. These DNA binding domains have a finger like shape with the z inc ions stabilizing this confo rmation by binding to cy s tein e s and histidines, amino acids that were previously known to bind z inc. In the case of TFIIIA zinc ions are bound by nine 30 amino acid repetitive sequences that carry cystein e s and histidines at the base of the fingers, in a C2H2 conformation which is the most common DNA binding ZF domain (Miller et al., 1985) Later investigations demonstrated that the same ZF motif of TFIIIA (z inc finger 5) bound 5 S r RNA via direct interactions of the ZF with a double helical region of the RNA (Theunissen et al., 1998) A large number of different ZF motifs have been documented, including the CCCH, or C3H motif that is found in the Mbl/MBNL protei ns which bind to ds RNA. Mbl proteins in Drosophila and Mbnl1 in mammals have been shown by direct ed mutagenesis and synthetic CUG repeats to bind the ds RNA stem via recognition of pyrimidine mismatches bordered by a C G and G C base pairs (Goers et al., 2008; Yuan et al., 2007) The binding occurs via direct interaction of the protein with the G and mutation of this nucleotide eliminates binding. Most of the Mbnl splicing isoforms carry two pairs of C3H motifs that are separated from each other by a linker region Crystallographic analysis has shown that the second pair of ZFs promotes RNA folding into dsRNA or an anti parallel structure (Teplova and Patel, 2008)


29 The bin ding of RNA binding proteins to pre mRNA may regulate editing (Paul an d Bass, 1998) and alternative splicing (Ho et al., 2004) while binding to mRNA may alter m RNA function by either promoti ng translation (Cuchalova et al., 2010; Kessler and Sachs, 1998) increasing stability (Ruiz Echevarria and Peltz, 2000) directing localization (Adereth et al., 2005; R oss et al., 1997) or degradation (Lagos Quintana et al., 2001; Peltz et al., 1993) The major function of MBNL1 is alternative splicing regulation This protein promot es exon skipping when bound to conserved intro nic enhancers upstream of an exon or exon inclusion when bound downstream of an alternative exon (Grammatikakis et al., 2011; Ho et al., 2004; Sen et al., 2010) Using minigene splicing assays and mutagenesis analys is o f known DM mis spliced targets, it has been shown that all three M BNL protein isoforms regulate splicing in a similar fashion and independently of CELF1 or other CELF protein s. This finding argues against the hypothesis of a synergistic binding and reg ulat ion of alternative splicing from M BNL 1 and C ELF 1 on the same target (Kino et al., 2009) Mouse Mbnl1 an d Mbnl3 bind to the same intronic elements upstream or downstream of excluded or included exon s (Grammatikakis et al., 2011) Many of the mis spliced mRNA species that are found in DM patients, such as cardiac troponin T ( TNNT2 exon 5), insulin receptor (I NS R exon 11), chloride channel type 1 (CLC N 1 exon 7a) and Troponin T type 3 (TNNT3 F exon) are also mis regulated in Mbnl1 knockout mice (Kanadia et al., 2003a) The identification of mis spliced genes in Mbnl1 knockout mice, and the generation of an RNA splicing map, revealed that Mbnl1 binds pref erentially to YGCY cluster s, particularly UGCU Mbnl1 has been shown to directly bind to YGCY clusters to regulate


30 alternative splicing (Yuan et al., 2007) Interestingly YGCY clusters are a characteristic structura l feature of CUG expansion RNA hairpins (Du et al.; Ho et al., 2004) Most of the research on DM1 pathogenesis has been focused on Mbnl1 due to its relatively high expression level in muscle, a major affected tissu e in DM. The limited research that has been conducted on the two other family members, Mbnl2 and Mbnl3 was done in conjunction to Mbnl1 studies to test if all three Mbnl genes can function in a similar way A ll three family members regulate alternative sp licing by promoting adult splicing pattern s for same targets in cell culture splicing assays (Ho et al., 2004) Expression analysis shows a more uniform expression for mouse Mbnl2 throughout development and in different tissues, whereas Mbnl3 shows mostly embryonic expression with a few exceptions (Kanadia et al., 2003b; Lee et al., 200 7) Even though Mbnl3 promotes a n adult splicing pattern in transfected cells the fact that Mbnbl3 is primarily expressed during embr yogenesis argues against the hypothesis that this family member plays a role in postnatal splicing regulation Indeed, Mb nl3 is expressed in muscle progenitor cells (MPCs) including myoblasts and Mbnl 3 overexpression inhibits myogenesis (Lee et al., 2007) During myoge n ic differentiation in C2C12 myoblasts, Mbnl3 promotes exclusion of the beta exon of muscle transcription factor myocyte enhancer factor 2 (Mef2), leadi ng to a less active Mef2. Mbnl3 is only expressed in the muscle precursor cells where it is presumed to inhibit their differentiation in to mature skeletal myofibers (Lee et al., 2010) On the other hand, Mbnl2 is wi dely expressed in most tissues throughout development suggesting a more universal role in alternative splicing regulation. To evaluate the role of Mbnl2 loss of function in DM, two gene trap mice were developed


31 that gave contradictory results (Hao et al., 2008; Lin et al., 2006) and these will be further discussed in the DM mouse model section. Even though Mbnl proteins are paralogs, Mbnl2 has been implicated in non splicing functions that have not been addressed for the other two family members. Adereth et al. investigated the role of Mbnl2 assisted localization of integrin 3 in migrating tumor cells to understand how muscleblind loss leads to muscle detachment from the epidermis in Drosophila m bl mutants (Adereth et al., 2005; Artero et al., 1998) Mbnl1 or Mbnl3 were not included in t his study to test the possibility that this function may be a shared among all three family members. A function that may be unique to Mbnl2 is in circadian rhythm regulation. Mbnl2 level s show circadian rhythm fluctuations in the rat pineal gland with Mb nl2 RNA increasing 7 fold during the dark cycle and its upregulation may be adrenergic cAMP controlled (Kim et al., 2009) The pineal gland interacts with the suprachiasmatic nucleus of the hypothalamus to interpret visual input s in circadian rhythm regulation. (Falcon et al., 2009) The effect of Mbnl2 level fluctuation s in circadian rhythm maintenance and the downstream target RNA s that may be affected have not yet been defined but Mbnl2 loss of function could cause the hypersomnolence observed in DM patients. Mouse DM Models In an effort to understand the molecular basis of DM and design new therapeutic strategies, several mouse models have been created. Since DM is a genetic disease dependent on the expansion of mi crosatellite repeats several attempts have been made to generate mouse models that bear the same mutation. This is a genetic challenge since the CTG repeat expansions are unstable in mice


32 H SA LR A bout a decade ago Thornton and colleagues generated a tr ansgenic mouse model that expresses 250 CUG repeats behind the human skeletal actin gene promoter (Mankodi et al., 2000) This muscle specific DM model was generated to test the hypothesis that DM results from the C UG repeat expansion at the RNA level and not due to loss of function of DMPK function or mis regulatio n of neighboring genes. In HSA LR mice, the CUG repeat is expressed at very high levels only in skeletal muscle. In addition the trans gene is integrated at a site remote from the DMPK genetic locus. These non translated CTG repeats accumulated in nuclear foci and caused a similar muscle pathology including but not limited to splicing mis regulation, chloride channel loss, myotonia and centralized nuclei Mbnl1 protein was shown to be sequestered in these RNA foci but C ELF 1 levels are not up regulated suggesting a secondary role of CELF1 to the disease pathogenesis The severity of the phenotype was directly correlated to the expression level of the trans gene and control mice with short repeats ( HAS SR ) failed to reproduce the phenotype suggesting that the effect of the repeats is at the mRNA level. Mbnl1 3/ 3 To test the hypothesis that Mbnl1 sequestration and loss of function results in DM, a n Mbnl1 k nockout (KO) model was generated by deleting Mbnl1 exon 3 (Kanadia et al., 2003a) Mbnl1 3/ 3 mice recapitulat e faithfully a number of DM clinical features including myotonia muscle pathology cataracts cardiac c onduction defects and aberrant splicing of the muscle specific chloride channel, cardiac troponin T and insulin receptor,. Due to the severity of myotonia and movement impairment, a brain specific Mbnl1 KO mouse is necessary to address Mbnl1 functions in the brain. Importantly,


33 Mbnl1 knockout mice do not recapitulate many of the mis splicing patterns in the DM1 brain. C ELF 1 Overexpression As stated previously, a major factor in the regulation of the embryonic to adult alternative splicing switch is C ELF 1, which has been shown to be upregulated in DM patients. Increase of C ELF 1 leads to retention of an embryonic splicing pattern, causing a delay in muscle maturation and cardiac defects (Koshelev et al., 2010; Timchenk o et al., 2004) Two different transgenic mouse models have been generat ed which overexpress C ELF 1 in muscle and heart. In the doxycycline inducible muscle specific model, an 8 fold upregulation of C ELF 1 result s in severe muscle pathology including myoton ia, centralized myonuclei split fibers muscle wasting, a prominent DM feature absent from other model s, and splicing defects In a second C ELF 1 overexpressor model human C ELF 1 was tetracycline inducible using a heart specific reverse tetracycline trans activator transgene. These mice show dilated cardiomyopathy, widespread degeneration, necrosis and loss of myocardial fiber, leading to death after 14 days of doxycycline administration (Koshelev et al., 2010) Mos t of the known mis regulated alternative splicing targets in adults were also found to follow the embryonic splicing pattern. DM300 DM328XL and DMSXL In these transgenic mice, CTG repeats are contained in a 45 kb fragment encompassing the human DM1 locus which includes 300 (DM300), 800 (DM328XL) or 1 2 00 1800 (DMSXL) CUG repeats (Gomes Pereira et al., 2007; Seznec et al., 2000) The se mice were generated to reproduce a more human DM1 like expression pattern of the CTG repeats. The DM300 mice develop muscle pathology including muscle loss


34 and C ELF 1 upregulation, age dependent defects in insulin metabolism (Guiraud Dogan et al., 2007) and CNS pathology including the abnormal d istribution of tau protein isoforms (Seznec et al., 2001) found in DM patients (Sergeant et al., 2001) Intergenerational instability of the CUG repeats of the DM300 mice led to progeny that had 1 2 00 1800 C T G repeats with depressed expression levels compared to DM300 mice. In the homozygous state these mice also develop the clinical features listed above but with an onset as early as 1 month of age. Despite the useful informa tion that the above mice provided, there are concerns on the expression levels of the transgene due to random integration and epigenetic regulation of the locus ( e.g hypermethylatio n ). R esearch is now focused on resolving the molecular pathways involved in DM1 pathogenesis and possible therapeutic strategies. Mbnl2 G ene T rap M ouse M odels The success of the Mbnl1 KO model in recapitulating a DM relevant phenotype generated a controversy about the role of other MBNL proteins in DM. To verify a role of Mbnl 2 in DM disease two different knockout mice were generated by independent labs by introducing a gene trap in two different Mbnl2 introns, intron 2 and intron 4 (Hao et al., 2008; Lin et al., 2006) The gene trap co nsists of a cassette which contains neomycin selection and galactosidase markers downstream of an engrail ed 2 (EN2) splice site acceptor The EN2 splicing acceptor and a strong polyadenylation signal downstream of gal terminates translation and reduces production of the protein encoded by the targeted gene The difference between the two Mbnl2 gene traps is the number of C 3 H motifs that were present in the final fused protein. Integration of the gene trap in intron 4 generate d


35 a protein with fully funct ional pair s of C 3 H motifs and most of the Mbnl2 linker region whereas integration in intron 2 generate d only a small Mbnl2 fragment with one C 3 H pair Another difference between the two gene traps is that the GT2 mouse show ed gene trap leakiness with possi ble splicing of Mbnl2 exon 2 to exon 3, skipping the splicing acceptor of the gene trap, as indicated by the residual WT mRNA shown on Northern blot analysis of total RNA of Mbnl2 GT2/GT2 mice (Hao et al., 2008) A n alysis of the se gene trap mice showed different phenotype s ( Fig ure 1 1 ) The GT4 mice were healthy and lived a normal life span with no signs of DM clinical features, but a cross with Mbnl1 K O s did not yield double mutant mice suggesting that Mbnl2 express ion is essential following loss of Mbnl1 (unpublished data). On the other hand, the GT2 mouse showed a mild Mbnl1 3/ 3 related phenotype with very mild myotonia, mosaic loss of Clcn1 in muscle and very minor mis splicing of Clcn1 pre mRNA N o further ana lys e s have been performed on these mice and the fact that GT2 mice show only a mild DM like phenotype is intriguing suggesting that a low level of Mbnl2 protein is sufficient to maintain almost normal function. Discussion The missing DM relevant features o f the Mbnl1 KO model, the unlinked molecular cascade between CUG expansion and C ELF 1 upregulation and a more complete DM manifestation in CUG overexpressing mice points out possible role s for other proteins in DM pathogenesis. However, another possible i nterpretation is that recapitulation of the DM phenotype in mice requires coordinate sequestration of all three Mbnl proteins. In this study, I addressed this latter possibility by generating an Mbnl2 KO mouse model ( Fig. 1 1) which shows widespread mis re gulation of alternative splicing in the brain.


36 Two families of alternative splicing regulators have been proposed to be responsible for the splicing regulation of these genes, the MBNL and the CELF family of proteins. C ELF 1 has been shown to regulate the s plicing of at least one of the genes but a direct link between CUG expansion and C ELF 1 upregulation has not been established. On the other hand, loss of Mbnl1 in a mouse model failed to recapitulate the majority of mis splicing events in the brain suggesti n g a possible role for Mbnl2, or a novel unidentified RNA binding protein that is also sequestered by CUG repeats.


37 Fig ure 1 1 Different models of Mbnl2 deficient mice show diverse phenotypes A) Two different Mbnl2 gene traps were generated, one in in tron 4 (upper) that leads to a mouse with no apparent abnormal phenotype and one in intron 2 (lower) that gave rise to a mouse model with mild myotonia, muscle histopathology and mis splicing of Clcn1. B) Homologous recombination strategy to generate a nul l Mbnl2 mouse model targeting exon 2 that carries the only translation initiation codon.


38 CHAPTER 2 MBNL2 IS EXPRESSED IN MURINE ADULT BRAIN BUT NOT MUSCLE The high degree of conser vation between Mbnl2 and Mbnl1 the capability to bind and regulate alterna tive splicing of the same targets in vitro and expression pattern of Mbnl2 make this Mbnl protein a possible candidate to explain the brain re lated events in DM In this chapter I present a detailed Mbnl2 expression analysis in the mouse brain defining th e major splicing isoforms and the cellular and subcellular distribution. Results Mbnl2 Is Widely Expressed In the Murine Brain T o define the regions of the brain and the cell types that express Mbnl2 I took advantage of the previously generated Mbnl2 GT 4 mouse line that expresses galactosidase under the Mbnl2 promoter. galactosidase catalyze s the hydrolysis of lactose and other beta galactosides including the model substrate of 5 bromo 4 chloro 3 indolyl D galactoside The monomers of this indole d erivative are colorless but after oxidization they produce a stable blue precipitate (Holt and Sadler, 1958a; Holt and Sadler, 1958b) Using this approach, I assayed the expression pattern of the Mbnl2 gene. The galactosidase cassette is integrated in intron 4 of the Mbnl2 gene. A splicing acceptor upstream of the cassette results in splicing of the Mbnl2 exon 4 to the translation o galactosidase protein that is expressed where a full length Mbnl2 protein should be expressed. To visualize the expression pattern of Mbnl2 gene in the mouse brain I obtained coronal brain section from the GT4 mice and performed an X Gal staining.


39 The neuronal populations that stain ed intensely were the cerebellar Purkinje cells, cells in the deep nuclei of the cerebellum as well as hippocampal granular and pyramidal neurons and the habenular nucle i, thus providing evidence for relatively high Mbnl2 expression levels ( Fig ure 2 1). There were a plethora of other types of cells that show ed gal staining including glia within the white matter of the brain stem and cerebellum, ependymal cells in the ventricles as well as Golgi type 2 cells in the granular layer, but the signal was lower possibly due to lower levels of Mbnl2 transcription or due to the small size of the cells The nature of the gal staining technique does not address the issue of the subcellular localization of the Mbnl2 protein T o address th is issue we need to know the subcellular localization of Mbnl2, as well as the pre sence of D MPK CUG expanded RNA in the nuclei of these cells Generation of an Mbnl2 S pecific A ntibody T o evaluate the role of Mbnl2 in the brain and its implications in DM pathogenesis, the levels of Mbnl2 protein expression and its sub cellular localizati on need to be characterized by the development of a mono specific antibody The high degree of primary structure conservation among the three Mbnl paralogs did not allow the use of existing Mbnl antibodies to characterize the Mbnl2 protein expression patte rn. T o find a unique to Mbnl2 region that c ould be used to raise a homolog specific antibody, I first performed an Mbnl2 isoform analysis and then aligned all the known isoforms of the three Mbnl homologs at a protein level


40 I soform analysis In the curren t literature several Mbnl1 splicing isoforms have been studied which include or exclude exons 5, 7 and 8 as well as a truncated isoform that skips exon 1 2 and produces an translation init iation codon. T o verify the different Mbnl2 isoforms that are expressed in the brain, I produced a cDNA library using RNAs isolated from different regions of the brain and sequenced a number of clones with primers complementary to exon s 1 and 9. This analy sis identified t hree Mbnl2 alternative spliced exons, one of which was described for the first time and annotated as Mbnl2 exon 8a. The alternative exons include d : 1) exon 6 which is equivalent to Mbnl1 exon 5 ; 2) exon 8a equivalent to Mbnl1 exon 7 ; 3) e xon 8b equivalent to Mbnl1 exon 8 ( Figure 2 2 A ). The c ombination of the se three exons produce d 7 different splic ed isoforms and the majority include d exon 6. To define the major Mbnl2 isoform in brain and muscle the level of inclusion and exclusion of th e aforementioned exons was tested by RT PCR ( Fig ure 2 2 C ). In brain, a high percentage of the Mbnl2 mRNA population includes exon 6, as well as multiple isoforms that include only exon 8a, only exon8b or both. In muscle the majority of the isoforms skip al l three of these alternative exons. A ntibody p roduction and t esting Two regions that are unique to Mbnl2 were used to raise an isoform specific polyclonal antibody. The first region is part of the linker between the two pair s of C3H domains and is present in all Mbnl2 isoforms. The second region is isoform specific and generated by the skipping of exon 8b. Exon 8b skipping generate d a unique to C terminus domain with unknown function (Fig ure 2 2 B ) Both regions were used to


41 design small immunoreactive pepti des and generate polyclonal rabbit antibodi es against the murine Mbnl2 protein. The cross reactivity of the resulting antisera was characterized to e valuate the degree of antibody specificity for Mbnl2 To test the cross reaction to the other two members of the Mbnl family, plasmids that express one of the three Mbnl homologs fused to a C terminal Myc tag were transfected in to CosM6 cells. The specificity of the purified antibody was tested by both immunofluorescence (Fig ure 2 3) and western blotting (Figu re 2 4 A ). Finally to verify that the antibody recognize d endogenous Mbnl2, I employed C2C12 cells which have been reported to express high levels of Mbnl2 (Holt et al., 2009) Treatment of these cells with siRNA a gainst the endogenous Mbnl2 mRNA confirmed the specificity of the Mb2k antibody (Fig ure 2 4 B ) The purified antibody which was elicited against the linker region recognized protein expressed from the myc tagged M bnl2 expression ve ctor, but not from the myc tagged Mbnl1 and myc tagged Mbnl3 expression vectors ( Figure 2 4A, upper row ). To verify transfection efficiency and expression of the o ther two family member proteins in the cell lysates, western blot analysis was performed with an anti Myc antibody ( Fig u re 2 4A, middle row ). The results show similar expression levels for all three myc tagged proteins. For a loading control anti Gapdh antibody was used which demonstrated equal loading for all four cell lysates including a myc construct where the myc tag was out of frame (transfection control) and thus was not detected by the anti Myc antibody (Fig ure 2 4A, bottom lane) A s imilar detection pattern was observed by IF on these cells using the same transfection conditions (Fig ure 2 3) More importantly Mbnl 2 as well as the other Mbnl


42 proteins detected with the anti Myc antibody had the expected nuclear localization pattern It is worthwhile to note that all three M yc tagged Mbnl expressing constructs were constructed with the isoforms that includ e the exon responsible for the nuclear localization of the Mbnl protein. All three Mbnl paralogs were expressed in a cell culture system to test the cross reaction of the antibody with the other two Mbnl proteins, Mbnl1 and Mbnl3. A caveat this approach is that prot eins are ectopically expressed at a relatively high level so it is important to test whether these antibodies recognize endogenous Mbnl2 To test antibody cross reaction with other endogenously expressed proteins, siRNA knockdowns of Mbnl2 were performed i n C2C12 cells Western blot analysis of control versus siRNA treated cells showed that the purified anti Mbnl2 antibody recognized endogenous Mbnl2 protein (Fig ure 2 4B, upper lane). The results indicate that in a cell based system, these anti Mbnl2 antibo d ies s pecifically recognize Mbnl2 Generation of an Mbnl2 K nockout M ouse M odel Based on the hypothesis that loss of Mbnl2 in the brain resulted in the CNS deficits associated with DM I tested this hypothesis by generating Mbnl2 knockout mice. The Mbnl2 kn ockouts served also as a negative control in the Mbnl2 RNA and protein expression analysis. Generation of the Mbnl2 ES c ell t argeting v ector The Mbnl2 conditional knockout targeting vector was constructed using a new method of plasmid manipulation and gen etic engineering, termed recombineering to avoid complications associated with conventional cloning techniques. A major advantage of recombineering is that the introduc tion of extraneous sequences at the Mbnl2 genomic locus was not included except for a f ew sequence elements that


43 allowed correct targeting in ES cells A conditional targeted allele was generated to avoid a possible embryonic lethal phenotype due to loss of Mbnl2 during embryogenesis as well as to tease out potential tissue specific contrib utions of the protein to possible disease relevant phenotypes This strategy avoided a problem with the Mbnl1 3/ 3 mouse model which is a constitutive knock out model so that an unbiased analysis of potential CNS deficits is compromised All of the behavior al analysis tasks for mice require normal motor function. To generate the Mbnl2 conditional targeting con struct, a bacterial artificial chromosome (BAC) was obtained which included all of Mbnl2 exon 2 and ~ 10 kb of upstream and downstream genomic sequence. Exon 2 is the only exon that is utilized for translation initiation and there are annotated splicing eve nts which lead to exon 2 skipping. A 10 kb fragment of the Mbnl2 BAC was cloned in to a targeting vector (Fig ure 2 5, lane 1) that conta i ned all the necessary selection markers and elements for ES cell targeting including the negative selection marker thim idine kinase (TK). To generate a conditional KO gene Mbnl2 exon 2 was flanked by loxP sites. The loxP site is a 34 bp DNA element composed by two 13 bp inverted repeats that provide enzyme specificity and an 8 bp linker that provides directionality. The l oxP sites are recognized by Cre recombinase which, depending on the orientation of the linker element, either promotes inversion or circularization and deletion of the genomic sequence between two loxP sites by homologous recombination (Sauer and Henderson, 1988) I introduced the first loxP site about 250 bp upstream of the intron exon junction of exon 2 (Fig ure 2 5, lane 2), to prevent inference with cis acting splicing elements. In any DNA manipulation, a select ion method/marker is necessary to isolate and/or enrich for


44 carries a n eomycin selection marker (Neo) flanked by two loxP elements (floxed) was used. This cassette i s driven by two promoters, a eukaryotic Pgk, and a prokaryotic Em7 promoter to drive selection in both ES cells and E. coli After the correct targeting event and clonal selection, the selection cassette was deleted by introducing a lactose inducible cre expressing plasmid that upon induction recombined the two loxP sites (Fig ure 2 5, lane 3). This step is necessary before the introduction of the downstream same selection marker. bp downstream of Mbnl2 exon 2 and using a plasmid where Neo is flanked by two Frt ure 2 5, lane 4). The frt sites belo ng to a yeast site specific recombination system which promotes a balance between the levels of two different plasmid isoforms, plasmid A and B. The recombination is catalyzed by the enzyme Flp (Volkert and Broach, 1986) The final targeting construct was tested for sequence integrity by genomic sequencing of the introduced DNA elements, as well as loxP function. To test loxP functionality, I introduced the final targeting construc t to cells that express Cre recombinase and tested for exon 2 deletion (Fig ure 2 5, lane 5). ES c ell t argeting After complete verification the targeting construct was linearized and targeted to 129 SvlmJ ES cells following standard procedures (Kanadia et al., 2003a) Positive clones that survived G418 and Neo selection were picked and screened by Southern blotting to identify the correct targeting events. Two different probes where used for the screening,


45 one spanni ARM should yield a band size of 5 kb after Stu I digestion versus the 6.4 kb for the WT allele. A corre ARM should give a band size of 5 kb after Sca I digestion versus 31 kb for the WT allele. After selection, 432 colonies were picked and which 5 (1 To verify functional l oxP integration of the targeted Mbnl2 conditional allele, Cre was ectopically expressed in targeted ESCs and assayed for loss of exon 2 ES c ell b lastocyst i njection s ES cells from positive clo ne 173 were injected in C57BL/6 pseudopregnant female mice (University of Michigan Transgenic Core ) and chimeric male animals were acquired for mating and germline transmission of the conditional allele A fter two blastocycst injections, 5 chimeric animals were shipped and received by our lab at 3 weeks of age. The chimeric percentage and coat color shows the contribution of the 129 S vlmJ cells in the development of the embryo and t he highest percentage chimera we obtained was 95% and the lowest was 40%. Mb nl2 c onditional l ine g eneration and m aintenance C himeric mice carry a mixture of wild type cells and cells heterozygous for the conditional allele. The conditional allele is expected to contribute equally with the WT allele for Mbnl2 expression and the add itional genetic information should not result in an aberrant phenotype. To obtain heterozygous Mbnl2 +/con mice, all 5 chimeras were mated to C57BL/6 female mice and the progeny were genotyped for germline transmission of the conditional allele. From the 5 chimeric mouse matings only two


46 mated successfully (75% and 40%) and both liters from these matings gave germline transmission. Heterozygous conditional mice were further mated to obtain conditional allele homozygosity To minimize genetic drift mice we re backcrossed to either C57BL/6 or 129SvlmJ mice ( obtained from JAX). Mbnl2 2/ 2 c onstitutive k nockout m ouse g eneration In t he current model of DM pathogenesis toxi c CUG repeat RNAs sequester a number of proteins. T o determine if constitutive Mbnl2 loss was embryonic lethal I attempted to generate a line which lacked Mbnl2 ex pression in all the tissues. C onstitutive Mbnl2 knockout line s were generated by mating Mbnl2 +/con mice with B6.C Tg( CMV cre) 1Cgn/J mice that express Cre in all tissues d uring embryogenesis Under the CMV promoter, Cre is expressed in the germline which res ults in permanent deletion of Mbnl2 exon 2 in the resulting line Mbnl2 E2/ E2 were routinely intercrossed and then backcrossed every 10 generations to minimize genetic drift. Deletion of Mbnl2 exon 2 in Mbnl2 2/ 2 mice was confirmed by Southern blotting of WT, Mbnl2 + / 2 and Mbnl2 2/ 2 mice using a hybridization probe ou tside of the deleted region (Fig ure 2 6A) and RT PCR analysis with primers in Mbnl2 exon 2 and exon 3 (Fig ure 2 6B). Mbnl2 Protein Is E xpressed and M o stly Nuclear In B rain but Is Expressed at a Low Level in M uscle of A dult M ice The high expression of Mbnl 2 RNA in the brain suggested a more specialized role in the development and maturation of the neuronal network. To be a viable candidate for sequestration and involvement in the DM brain, Mbnl2 protein should be expressed in specific cell populations that are affected by DM disease and the sequestration hypothesis dictates that Mbnl2 should be predominantly nuclear.


47 After verification of the genotype, Mbnl2 2/ 2 mice were examined for Mbnl2 expression at both RNA and protein levels. To assess loss of exon 2 at the RNA level, RT PCR was performed using primer sets in exons 1 and 3, and 1 and 2, in and RNAs isolated from brain and muscle (Fig ure 2 7B). In both tissues, deletion of exon 2 was verified. During the analysis of Mbnl2 protein expression, the Mb 2k antibody described earlier in this thesis failed to detect Mbnl2 in mouse tissues likely due to relatively low affinity of this polyclonal for endogenous mouse Mbnl2. As an alternative, an anti Mbnl2 monoclonal antibody (mAb 3 B4), which was also elicite d against the same linker region as Mb2k and should recognize all Mbnl2 isoforms had recently become available (Holt et al., 2009) Therefore, we used mAb 3 B4 to examine the expression levels of Mbnl2 in six differe nt mouse tissues (hippocampus, cerebellum, heart, liver, lung, muscle, spleen) of WT and Mbnl2 knockout mice (Figure 2 7A. As expected, and in contrast to Mbnl1, Mbnl2 was not predominantly expressed in the muscle Indeed, mAb 3B4 detected very low levels of Mbnl2 protein in adult skeletal muscle in contrast to Mbnl2 RNA levels detected by RT PCR and Northern blot analyses. In contrast, Mbnl2 protein was readily detectable in the brain. Th e lack of Mbnl2 protein in skeletal muscle can be attributed to eithe r a translation al inhibitory mechanism, localization of the mRNA to regions of low translation al efficiency, or post translational modification of the protein at a region that masks the antibody epitope. A recent study also reported that Mbnl2 expression i s developmentally regulated with high expression levels early in embryogenesis and during muscle regeneration and decreased levels in adult muscle tissue (Holt et al., 2009) Most surprising was the finding that Mbn l2 protein levels are


48 the highest in multiple regions of the brain. No Mbnl2 protein was detected in Mbnl2 2/ 2 mice in any tissues confirming antibody specificity No difference in Mbnl1 protein levels was detected. In summary, all tissue s examined except skeletal muscle showed expression of Mbnl2 protein including the heart which is affected in DM patient s To identify the cell types which expressed Mbnl2 IHC was performed on coronal brain sections at the midbrain area and sagitt al cerebellar sections of both WT and Mbnl2 2/ 2 mice mAb 3B4. To evaluate the impact of the Mbnl2 loss on hippocampal area, Ni ssl staining was performed using both WT and Mbnl2 2/ 2 brains to assess for changes in gross morphology and integrity of the hippocampal formation. In agreement with the Mbnl2 LacZ staining patterns the IHC studies show ed an intense staining of cells i n the frontal cortex, dentate gyrus CA1 CA3 pyramidal region of the hippocampus and Purkinje cells of the cerebellum of WT mice N o staining was observed in Mbnl2 2/ 2 brain sections (Fig ure 2 8, 2 9). Mbnl2 protein co localiz ed with nuclear DAPI stain i ng indicating that the majority of Mbnl2 is nuclear. In cells of the cerebral cortex of the frontal lobe, the protein was localized in both the nucleus and the cytoplasm (Fig ure 2 8) supporting a prior suggestion that Mbnl2 has a role in integrin RNA traff icking and localization role in growing axons (de Andrade and Jansen, 2005) Discussion The lack of gal staining in other regions of the brain such as the neurons of the cerebellar molecular layer, the Bergmann glia and the choroid plexus epithelial cells suggest that Mbnl2 function(s) are important for a specific population of cells in the CNS and only these cells will be directly affected by loss of Mbnl2 function in the DM


49 brain DM patients do not suffer from ataxia which could be explained by the lack of DMPK expression in the cerebellum, Alternatively, MBNL2 protein may localize to the cytoplas m and is not trapped by CUG repeats Another possibility is functional complementation by another protein that is not sequestered by CUG repeats and can compensate for the loss of Mbnl2. On the other hand, Mbnl2 seems to be highly expressed in hippocampal neurons a region of the brain that is responsible for learning and memory which are functions affected in DM patients. The intense staining in the h a benular nuclei is of great importance as these nuclei are known to be involved in circadian rhythm regulat ion. The most common complaint of DM patients is that they cannot control their susceptibility to fall ing asleep. The h a benular nuclei are located above the thalamus at its posterior end proximal to the midline of the brain. This region has been considered to be a part of the epithalamus which interacts closely with the pineal body, which is the region where Mbnl2 RNA levels show circadian fluctuation s (Kim et al., 2009) Recently the h a benular nuclei have been impl icated in motivational control of behavior, such as reward based action selection (Matsumoto and Hikosaka, 2007) spatial reference memory and cognitive function (Lecourtier et al., 2004) and attention disturbances (Lecourtier and Kelly, 2005) During this study, several novel brain and muscle isoforms of Mbnl2 were discovered and a previous study indicated that alternative splicing of Mbnl1 exon 5 ( also referred to as exon 7) is developmentally regulated and plays role in the intracellular localization of the protein (Terenzi and Ladd, 2010) The high percentage of Mbnl2 exon 6 inclusion suggests that Mbnl2 localizes to the nucleus in neurons of the CNS. This


50 possibility supports the hypothesis that Mbnl2 plays a pivotal role in the regulation of alternative sp licing in the brain similar to the function of Mbnl1 in muscle A novel rabbit polyclonal antibody (Mb2k) was developed which can be used to detect Mbnl2 proteins. This antibody was used to evaluate Mbnl2 protein expression levels in different tissues by western blot and Mbnl2 localization by immunohistochemistry (IHC, data not shown). Unfortunately the antibody gave a non specific weak signal by IF and non specific bands by western blotting of mouse tissues that were present on Mbnl2 knockout mice. During the g eneration of the Mbnl2 knockout mouse, a new Mbnl2 specific monoclonal antibody became commercially available that was test ed in both WT and Mbnl2 knockouts and confirmed its specificity The predominan t nuclear localization of Mbnl2 and the high expressi on levels of this protein in all major neuronal cells of the hippocampus suggest that this Mbnl protein is the best candidate for a sequestered factor in DM The sequestration of Mbnl2 by toxic CUG repeats has been verified in human DM 1 fibroblast and myob last cells (Holt et al., 2009) CUG exp nuclear foci have been detected in neurons of autopsied DM1 brains in multiple regions including the, cerebr a l cortex cerebellar Purkinje cells, hippocampus, dentate gyrus, su bstantia nigra, subcortical white matter, corpus callosum, brain stem and oligodendrocytes of the centrum semiovale (Jiang et al., 2004) No characterized nuclear structures have been found to colocalize with these RNA foci, including PML bodies, nucleol i perinucleolar compartment s or nuclear splicing factor compartments ( SFCs or speckles )


51 Fig ure 2 1. Mbnl2 is highly transcribed in the mouse brain. LacZ staining performed in mouse brain coronal sections indica ting presence of Mbnl2 transcripts. Mbnl2 is highly expressed in Purkinje cells of the cerebellum (B) the hippocampal formation (C) and the medial habenula (D). More scattered expression is detected in the frontal cortex (A)


52 Fig ure 2 2. Extensive alte rnative splicing of Mbnl2 occurs in the C terminus of the protein. A) Mbnl2 gene alternative splicing map. Untranslated exonic regions are in white, coding exons in black and alternative spliced exons in red. B) Amino acid sequence of the Mbnl2 isoforms pr oduced by exon 8b skipping domain. C) RT PCR analysis of Mbnl2 alternative spliced exons 8a, 8b and 6 in different tissues (total brain, cerebellum hippocampus, tibialis anterior, gastrocnemius, soleus and quadriceps)


53 Fig ure 2 3. Mbnl2 polyclonal anti body does not cross react with the other Mbnl protein isoforms. Cosm6 cells were transiently transfected with myc tagged Mbnl2, Mbnl1, Mbnl3 and an empty Myc expression vector. Immunofluorescence staining was performed with the new Mbnl2 polyclonal antibod y (green) and an anti Myc antibody (red). In blue is the nuclear stain DAPI.


54 Fig ure 2 4. The anti Mbnl2 polyclonal antibody did not cross react with other endogenous expressed proteins in cell culture. A) Cosm6 cells were transiently transfected with m yc tagged Mbnl2, Mbnl1, Mbnl3 and an empty Myc expression vector. Western blot analysis was performed with the new Mbnl2 polyclonal antibody upper lane and an anti Myc antibody (middle lane). Gapdh was used as a loading control (lower lane). B) MEF cells w ere transfected with Mbnl2 specific siRNA (right) and scrambled siRNA as control (left). Western blot analysis was performed with antibodies against Mbnl2, Mbnl1, Cugbp1, and Gapdh as loading control.


55 Fig ure 2 5. Generation of Mbnl2 KO targeting cons truct. To verify the Mbnl2 construct 1ug of plasmid from different stages of the targeting construct generation were cut with XbaI. Lane 1) Mbnl2 retrieval in the backbone construct PL253. Lane 2) targeting of the upstream LoxP site. Lane 3) cre mediated n eomycin cassette excision. Lane 4) targeting of the downstream LoxP site. Lane 5) cre mediated deletion of exon 2. Digestion products were run in a 1% agarose gel and stained with e thidium b romide.


56 Fig ure 2 6. Mbnl2 2/ 2 mice lack expression of exon 2. A) DNA from WT, Mbnl2 + / 2 and Mbnl2 2/ 2 mouse tails was digested with Stu I and analyzed by Southern blotting. B) RNA from WT, Mbnl2 + / 2 and Mbnl2 2/ 2 mouse brain was analyzed by RT PCR for Mbnl2 exon 2 expres sion (fwd primer in e2 and rev primer in e3) as well as expression changes of Mbnl1 mRNA by analyzing levels of e3 (fwd primer in e3 and rev primer in e4). Ppia was used as loading control


57 Fig ure 2 7. Mbnl2 is highly expressed in mouse brain but not in muscle. A) WT and Mbnl2 2/ 2 m ouse protein lysates from cerebellum, hippocampus, frontal brain, heart, lung quad and spleen were analyzed by western blotting with a monoclonal anti Mbnl2 antibody (upper). Gapdh was used as a loading control (lower). B) WT and Mbnl2 2/ 2 m ous e RNA from total brain and quadriceps muscle was analyzed by RT PCR for exon 2 exclusion with two different primer sets (fwd e1 rev e3 on the left and fwd e2 rev e3 on the right)


58 Fig ure 2 8. Mbnl2 shows nuclear and cytoplasmic localization in cells of the frontal cortex. Frontal cortex brain sections from WT and Mbnl2 2/ 2 m ice were used for IF with an anti Mbnl2 antibody (green). DAPI served as a nuclear stain. Confocal imaging was taken at 20x (upper scale bar 0 ) and 250x (lower scale bar 0 ) zoom.


59 Fig ure 2 9. Mbnl2 is mostly nuclear in cells of the hippocampal formation. Hippocampal brain sections from WT and Mbnl2 2/ 2 m ice were used for IF with an anti Mbnl2 antibody (green). DAPI served as a nuclear stain. Confocal imaging was taken at 20x (upper scale bar 0 ) and 200x (lower scale bar 0 ) zoom.


60 CHAPTER 3 ANALYSIS OF THE M bnl 2 / 2 KNOCKOUT MOUSE MODEL In many cases, deletion of an essential gene that is implicated in multiple functions has a wide impact on the transcriptome and/or proteome of an organism and may lead to developmenta l defects, embryonic lethality or decreased life span or less severe phenotypes Gene duplications which result in paralogous gene families provide functional redundancy to complex organisms, which assure the survival of the organism in case of deleterious gene mutations. The three paralogous Mbnl genes possess a high degree of functional similarity and sequence conservation which suggests the possibility of functional complementation when another family member is compromised due to mutation. A prerequisite for functional complementation is that the spatial and temporal expression pattern of the paralogs should be similar. The presence of RNA foci in the nuclei of cells of the hippocampus and dentate gyrus indicates that toxic RNA expression could lead to in hibition Mbnl2 function in these cells. If Mbnl2 is vital for the finely tuned functional integrity of the brain, then we expect that loss of Mbnl2, if not embryonic lethal, it will compromise the computational capacity of the specific brain region/formati on and result in a distinct phenotype. Prior to testing mice for brain related phenotypes, we first confirmed that Mbnl2 knockouts did not show neuromuscular deficits that might impair their ability to complete a behavioral test. The previously reported Mb nl2 GT2 mouse model showed focal loss of the muscle specific chloride channel ClC 1 and muscle pathology with centralized nuclei and mild myotonia assessed by EMG (Hao et al., 2008) Thus, it was necessary to determ ine if Mbnl2 2/ 2 mice showed any muscle defects before any comprehensive behavioral analysis.


61 Results Mbnl2 2/ 2 M ice Show Postnatal Growth Retardation To test the hypothesis that Mbnl2 is essential for normal development of the mouse, the Mendelian ratios of the progeny of Mbnl2 + / 2 x Mbnl2 + / 2 crosses was compared to the normal 1:4 ratio for a homozygous Mbnl2 2/ 2 genotype. The genotype of mice was tested by genomic PCR amplification using primers in the region flanking exon 2. No statisticall y significant difference was found between the expected ratio and the actual ratio obtained from Mbnl2 + / 2 crosses (data not shown), suggesting that Mbnl2 is not an essential gene. Nevertheless, I noted enhanced death of Mbnl2 2/ 2 mice around the weani ng period. Only a few mice (n<10) that lived beyond weaning died prior to 5 months of age. Since constitutive deletion of Mbnl2 does no t result in embryonic lethality and the majority of the Mbnl2 knockouts survive post sexual maturity, it was not necessar y at this point to generate tissue specific Mbnl2 knockouts. During weaning an obvious difference in size was observed betwe en pups in the same litter (Figure 3 1, left). Genotyping of these mice showed that the smaller pups were Mbnl2 2/ 2 mice. To assess development of the Mbnl2 2/ 2 lineage, the body weights of Mbnl2 2/ 2 mice and littermates at several time points were obtained including weaning (P21) and P47 (Figure 3 1, right). At P21, Mbnl2 2/ 2 mice weighed <30% than WT l ittermates and this difference was more pronounced in males (data not shown) Both male and female Mbnl2 2/ 2 mice attain a normal weight by P29. N o A berrant M uscle P athology D ue to M bnl2 L oss The Mbnl1 3/ 3 mice showed severe muscle histopathology in cluding centralized nuclei, split fibers, as well as pycnotic nuclei but not muscle fiber loss. Mbnl2 + / GT2 mice developed a similar but milder phenotype when 9 month old mice were evaluated. To


62 examine Mbnl2 knockouts, muscles were obtained from mice at ~5 months of age to test for any of the above defects. In a representative transverse section of tibialis anterior (TA) from WT and Mbnl2 2 / 2 mice H&E staining was used to assess overt histolog ical changes in the mutant (Figure 3 2 ). Hematox y lin is a basic dye and stains the nuclei of cells (deep purple) and eosin is a basophilic dye that stains proteins (pink). In both WT and Mbnl2 2 / 2 mice, the muscle tissue appeared healthy with no obvious cytological differences between WT and Mbnl2 knockout muscle. Mbnl2 M ice D o N ot S how M uscle S pecific C hloride C hannel L oss In Mbnl1 mice, mis splicing of the muscle specific chloride channel Clcn1 le a d s to inclusion of a premature termination codon and loss of Clcn1 protein from the muscle fiber membrane (Kanadia et al., 2003a) Mbnl2 +/GT2 mice have been reported to show a similar but focal lo ss of Clcn1 expression but not the mis splicing of Clcn1 pre mRNA T o test if Mbnl2 mice show ed any loss of Clcn1 protein expression from the muscle fiber membrane, IHC was performed using transverse quadriceps muscle sections from 3 5 month old WT and Mbnl2 mice and an anti Clcn1 antibody (Fig ure 3 3 A ). The results showed no loss of Clcn1 from the muscle fiber membrane so at least up 5 months of age, Mbnl2 mice did not show any obvious muscle structural or functional changes Mbnl2 2 / 2 Mice Do Not Recapitulate Aberrant Splicing o f C lcn1 To verify that my Mbnl2 2 / 2 mic e do not show mis splicing of Clcn1 RT PCR analysis was performed with primers in the upstream and downstream exons of the fetal exon 7a (Figure 3 3 B). The analysis did not reveal an increase in the inclusion of Clcn1 exon 7a up to the age of 5 months in contrast to Mbnl1 3 / 3 mice.


63 In summary, Mbnl2 2 / 2 mice up to 5 months of age did not develop a movement disorder and thus it was possible to test for behavioral abnormalities using tests such as the Morris water maze (MWM), open field and sensitivity to seizure devel opment. Loss of M bnl2 Does Not Lead to an Overt Hippocampal Pathology Despite the high levels of Mbnl2 expression, loss of the protein did not result in any overt structural changes in the hippocampal formation (Figure 3 4). Both WT and Mbnl2 2/ 2 hippo campi displayed normal morphology without loss of any detectable cell population. Further analysis with more sensitive techniques is required to assess more subtle changes. Mbnl2 2/ 2 Mice Show Impaired Spatial Learning a nd Memory The integrity of the hi ppocampus is essential for spatial pattern establishment and learning. The Morris water maze test is an assay which tests the effects of hippocampal degeneration, or lesions, in rodents. Most studies have focused on using this assay to test for deficiencie s in the hippocampus, striatum, basal forebrain, cerebellum and neocortical areas and these as well as other CNS regions may influence the results. The Morris water maze tests the swimming pattern of mice and their ability to escape from the water onto a h idden platform. The test is based on the natural instinct of mice to avoid water and drowning. Two different versions of water maze test exist the hidden platform test (spatial version) and the visible (non spatial version) I performed t he hidden platfor m test (spatial version) assesses to assess the ability of the mouse to read spatial cues, learn and memorize a spatial map of the platform location and retrieve that map to evaluate the visual cues at any time point to find the platform. The Morris water maze consists of a round swimming pool filled with colored water. The color of the water depends on mouse coat color since an adequate contrast is required


64 by the software to detect the animal during the procedure. A platform is located under the water sur face so that mice cannot see it but can climb onto this stage once it is discovered. There are many training variations for the test but the most commonly used requires a 5 day training period during which the mouse learns to find the platform and 1 trial day when the platform is removed. During platform removal, a mouse that has been well trained and has no memory deficits should be able to remember where the platform was and search for it by crossing over the original platform position. The water maze is divided in four quadrants and the platform is located in the middle of one of the quadrants. If a mouse has a defect in learning and memory of spatial orientation then, in the most severe case, they will fail to restrict their platform search to the correc t quadrant and during the five training days they will fail to reach the platform. In less severe forms, the mouse will reach the platform during the training period, but during the trial day when the platform is removed, the number of cross overs over the platform is low. To test if loss of Mbnl2 leads to spatial learning and memory deficits, WT and Mbnl2 2 / 2 mice were tested in the hidden Morris water maze test. During the 5 day learning period the Mbnl2 2 / 2 mice showed a significant learning delay at Day 3, but recovered by Day 4 and performed as well as the WT at Day 5 (Figure 3 5). Mutant mice di d not show any difference in quadrant occupancy after removal of the platform from the target quadrant but Mbnl2 2 / 2 mice spent more time in the quadrant next to the target quadrant showing quadrant confusion in their searching path. In contrast, WT mic e showed an increase in target quadrant occupancy (Figure 3 6 ). Most importantly


65 on Day 6 after platform removal, Mbnl2 2 / 2 mice showed a significantly decreased number of platform crossovers compared to WT (Figure 3 7 ) To test if the decreased perform ance shown by the Mbnl2 knockout mice was due to impaired swimming speed or a decrease in the searching pathway, the average speed and total distance traveled to the platform of WT and Mbnl2 2 / 2 mice were compared (Figure 3 8 and 3 9 ) and no significant differences were found. To test if elevated stress response could affect the searching pathway, the time that mice spent in the thigmotaxis zone, defined as the perimeter region <10 cm from the tank wall, was used to assess stress level. Mice are introduc ed to a novel environment tend to display an elevated stress response due to their intrinsic escape behavior and rodents tend to limit their locomotor activity to an area close to the walls of a novel open space during the initial 5 10 minutes of initial p lacement. Subsequently, they slowly begin to explore the center of the open area. This induction of exploration and search for an escape is impaired by use of anxiolytic drugs (Simon et al., 1994) linking thigmotaxi s to stress response. When the platform was removed on Day 6, Mbnl2 2 / 2 mice remained in the thigmotaxi s zone compared to WT mice (Figure 3 10 ). Loss of Mbnl2 Results in Increased Seizure Susceptibility During normal bre e ding and maintenance of the colony, a few Mbnl2 2/ 2 mice developed a hyperexcitability phenotyp e near P21 followed by death with 24 hours. These mutants showed limited mobility movement and then nose twitching and sudden rapid/uncontrolled movement followed by tonic seizures. The etiology of the seizure onset and death was unclear but a number of he terozygous Mbnl2 +/ 2 and homozygous Mbnl2 2/ 2 animals died near weaning age.


66 Epilepsy and seizure induction results from abnormal neuronal network synchronization that has a multi factor pathogenesis that is attributed mainly to loss of inhibitory networks and less freq uent to a gain of excitatory networks in the brain. Over 70 different gene mutations and expression level changes have been implicated to an epileptic phenotype (Noebe ls, 2003) There are two main networks in the brain the glutaminergic or excitatory network and the GABAergic or inhibitory network. Neurons of the glutaminergic pathway release the neurotransmitter glutamate from the pre synaptic membrane to the synaptic cleft. Glutamate binds to post synaptic receptors and induces depolarization of the post synaptic membrane inducing membrane excitation and generation of axon potentials. Neurons of the GABAergic network release GABA, a neurotransmitter that causes hyper polarization or repolarization of the post synaptic membrane, thus inhibiting membrane excitation and generation of axon potentials. Disruption of the balanced excitation and inhibition of neurons can lead to hyper excitability and seizures. In most cases, synchronous neuronal excitation has been linked to channelopathies that disrupt normal synaptic inhibition (McCormick et al., 1999; Spampanato et al., 2001) but also to a constellation of perturbations in other me chanisms that lead to membrane excitability, such as vesicle docking and neurotransmitter synthesis (Kash et al., 1997; Rosahl et al., 1995) Cells that are not part of the main brain circuitry have also been implic ated in hyper excitability and seizures. Dec reased glutamate uptake by glia cells that lack expression of the glutamate transporter GLT 1 leads to seizures in GLT 1 knockout mice (Tanaka et al., 1997) Channelopathi es that result in presynaptic hyperexcitability can be caused by


67 mutations in either voltage or ligand gated channels that control the flow of potassium, sodium and calcium (Jouvenceau et al., 2001; Lau et al., 2000; Spampanato et al., 2001) The mutations interfere with the normal function of the channel by either altering the actual channel pore or any of the ligand binding domains and protein protein interaction sites. Even a small percentage change in the sensiti vity of these channels is capable to promote repetitive firing of neurons and seizure breakout. T o test the hypothesis that Mbnl2 2/ 2 mice were prone to seizure induction, the seizure inducing drug pentylenetetrazole (PTZ) was used. PTZ is a GABA antagonist and acts as an inhibitor of the inhibitory (GABAergic) pathways of the brain, exacerbating even modest changes and defects i n the excitatory pathways. Mbnl2 + / 2 mice were also included in this study due to the unexpected result that these mice showed higher susceptibility in seizure induction than WT mice. A pilot study was designed to test the effectiveness of different doses of PTZ in WT versus mutant Mbnl2 mice. This study determined that 40 mg PTZ per kg mouse body weight, injected intraperitoneally, was an ideal dose for this investigation since WT mice showed very limited seizure activity while Mbnl2 knockouts were profou ndly affected. Seizures were assayed using a modified Racine scale ( see methods for explanation ). PTZ injection in Mbnl2 2/ 2 homozygous k nockouts had a dramatic effect with an early onset (Level 1) with impaired movement and nose twitching during the first minute post injection. Mbnl2 + / 2 mice showed amore delayed response (~2 minutes post injection) while WT showed a m ore delayed response (Figure 3 11 ). A more profound difference was also observed between WT and mutants when the PTZ response was measured by the Racine scale (Fig ure 3 12). All Mbnl2 2/ 2 mice


68 reached Racine Level 6 and the majority of them died within 2 3 minutes whereas Mbnl2 + / 2 heterozygotes showed a wide range of response with few animals reaching the Level 6 followed by death while some only attained Level 2 3. WT mice remained at Level 1 (Fig ure 3 12 ). Another measure is the time interval between drug introduction and attainment of at least Level 5. In this case, Mbnl2 2/ 2 mice reached this threshold much earlier (~150 sec) compared to Mbnl2 + / 2 mice (>400 sec) (Fig ure 3 13 ). Discussion Growth retardation in Mbnl2 knockouts can result from mu ltiple molecular and behavioral changes which occur during early development and postnatal growth. Mbnl2 may be important for the splicing of genes responsible for normal development and growth, such as the IGF family, effectors that regulate pre and earl y postnatal growth (Liu et al., 1993) For example, Igf2 knockouts show a 40% decrease in birth weight but this difference remains as the mice age (DeChiara et al., 1990) Another pathway that regulates gr owth during the period between birth and P21 is the leptin pathway which controls the hypothalamic food intake circuitry (Cottrell et al., 2009) It is known that leptin normally inhibits the activity of orexigenic neurons that express neuropeptide Y, promote hunger and stimulates the activity of anorexigenic neurons that express pro opiomelanocortin (POMC) resulting in a robust decrease in food intake (Balthasar et al., 2004) Perinatally, leptin levels increase dramatically in the blood (Ahima et al., 1998) This increase in leptin has been implicated in the development and maturation of the food intake regulatory circuitry including r egions of the brain such as the hypothalamus and paraventricular nucleus which is progressively established between postnatal day P 6 and P 16 (Bouret et al., 2004) During the perinatal period the increased leptin le vel does not inhibit food intake since the body weight of young mice rapidly increases


69 (leptin resistance) (Mistry et al., 1999) .Disruption of the leptin resistance me chanism could potentially lead to decreased food intake and retarded growth in Mbnl2 2/ 2 mice. In addition, an early postnatal leptin blockage has been correlated to long term leptin resistance of mature animals leading to obesity in rats (Attig et al., 2008) Preliminary data (not shown) show a tendency of increase d body weight in 5 month and older Mbnl2 knockout mice which provides some supporting evidence for this hypothesis. The Mo rris water maze test showed spatial learning and memory impairment of the Mbnl2 knockout mice by a modest delay in learning during training and decreased spatial precision. A concern to this study is the increased thigmotaxis Mbnl2 knockouts showed as a re sult of platform removal. This result could explain the altered platform cross over performance of these mutants as the platform was placed away from the periphery of the water maze. Another potential explanation is that Mbnl2 knockouts never got accustome d to the new environment (pool arena) due to an elevated stress response. The Morris water maze is one of the most stressful tests for mice because they correlate water submersion with a danger of drowning. A different test to assess thigmotaxis and stress response such as open field should be also performed. The growth retardation could be attributed to similar behavioral abnormalities of Mbnl2 knockout mice for example an inability to compete for food, or even an inability to associate lactation with food intake. All of the above possible scenarios require further investigation. Seizure susceptibility has not been carefully studied in DM patients and very few unpublished incidents have been reported. There are several differences between the


70 Mbnl2 E2/ E2 a nd DM patients. First, DM requires the production of toxic C(C)UG RNAs so only those cells that transcribe either the DMPK or CNBP mutant genes will be affected in contrast to the Mbnl2 knockout model where Mbnl2 protein is ubiquitously absent Second, in DM there is partial loss of MBNL proteins from toxic RNA and the level of sequestration and loss depends on the expression and size of the repeats. Spontaneous seizures were a rare event in the Mbnl2 knockouts where Mbnl2 is absent. The genetic basis of t he seizure prone phenotype in the Mbnl2 2/ 2 mouse model is not known and further experiments should be conducted to identify the pathways which are responsible for this phenotype. Additionally, mouse models of DM that express CUG expansion repeats shoul d be tested for seizure susceptibility. However, existing poly(CUG) and poly(CCUG) models do not show sufficient expression of C(CTG expansions in the brain. If transgenic mice that express polyC(C)UG repeats in the brain show a similar seizure susceptibil ity phenotype then it is possible that seizure susceptibility is a common, but subclinical phenotype, in the DM population.


71 Fig ure 3 1. G rowth retardation in Mbnl2 2/ 2 mice 21 old WT and Mbnl2 2/ 2 day mice were photographed (left). Weight was recorded between P21 and P47.


72 Fig ure 3 2. Mbnl2 loss does not lead to muscle histopathology. TA muscle cross sections from WT (left) and Mbnl2 2 / 2 (right) mice we re stained with h ematoxylin (purple nuclear stain) and eosin (pink cytoplasmic stain).


73 Fig ure 3 3. Mbnl2 loss does not lead to chloride channel loss due to Clcn1 aberrant splicing. WT and Mbnl2 2 / 2 quadriceps muscle was used for IF (A) or alternat ive splicing RT PCR (B). Mbnl2 2 / 2 mice did not show any reduced staining for the muscle specific chloride channel protein Clcn 1 (red) compared to WT. Clcn1 fetal exon inclusion was not up regulated in Mbnl2 2 / 2 mice comparing to WT and as seen in p reviously characterized Mbnl1 3 / 3 mice (Kanadia et al., 2003a)


74 Fig ure 3 4. No overt histological changes in the hippocampal formation of Mbnl2 2/ 2 m ice. Coronal mid brain sections of the hippocampal form ation from WT and Mbnl2 2/ 2 m ice were treated with Nissl stain to visualize RNA (purple) and eosin to visualize protein (pink).


75 Fig ure 3 5. Latency in learning and memory formation in the Morris water maze test. Mbnl2 2/ 2 and WT mice were traine d for 5 consecutive days (4 trials/day) to reach a submerged platform. The average time to platform was recorded for each day; WT ( n =16), Mbnl2 2/ 2 ( n =13), p<0.005


76 Fig ure 3 6 Quadrant occupancy in the Morris water maze test. On Day 6, and after th e platform removal, mice were allowed to swim in the maze for 60 seconds. The total time they spend in each quadrant was measured Abbreviations are (T) target, (AR) adjacent right, (AL) adjacent left, (O) oposite quadrant using WT ( n =16) and Mbnl2 2/ 2 ( n =16). p<0.05


77 Fig ure 3 7 Decreased spatial memory and precision in Mbnl2 2/ 2 mice. On Day 6 and upon platform removal the mice were allowed to swim for 60 seconds in the maze. During the swim, the number of times that mice crossed over the platfo rm was counted : WT ( n =16) Mbnl2 2/ 2 ( n =13), p<0.05.


78 Fig ure 3 8 No significant difference in total path swim length between WT and Mbnl2 2/ 2 mice. WT and Mbnl2 2/ 2 mice were allowed to swim until they reached the platform, which was removed on Day 6. The total swimming path to reach the platform (< 60 seconds) on Day 6 was recorded. for WT ( n =16) and Mbnl2 2/ 2 ( n =13) mice.


79 Fig ure 3 9 No significant difference in swimming speed between WT and Mbnl2 2/ 2 mice. On Day 6, mice were a llowed to swim for 60 seconds without the platform and the average speed was calculated by dividing the total distance swam by total swim duration (60sec) : WT ( n =16), Mbnl2 2/ 2 ( n =13).


80 Fig ure 3 10 Mbnl2 2/ 2 mice show increased thigmotaxis A thi gmotaxis zone was defined as 10 cm from the periphery of the MWM wall. Mice were allowed to swim 60 seconds after platform removal and the time they spent in that zone was recorded: WT ( n =16) and Mbnl2 2/ 2 ( n =13), p<0.05.


81 Fig ure 3 11 Latency to s eizure onset is reduced in both Mbnl2 2/ 2 and Mbnl2 +/ 2 mice. WT, Mbnl2 +/ 2 and Mbnl2 2/ 2 mice where injected with 40 mg/kg PTZ and the latency scored on a modified Racine scale: WT ( n =4), Mbnl2 +/ 2 ( n = 6) and Mbnl2 2/ 2 ( n = 4 ) p<0.0005 (WT), ** p<0.005 ( Mbnl2 +/ 2 ).


82 Fig ure 3 12 Severity of convulsion is increased on both in Mbnl2 2/ 2 and Mbnl2 +/ 2 mice. WT, Mbnl2 +/ 2 and Mbnl2 2/ 2 mice were scored for seizure severity during a 1 hour observation period following PTZ injection: WT ( n = 4 ), Mbnl2 +/ 2 ( n = 6) and Mbnl2 2/ 2 ( n = 4) p<0. 005 ** p<0. 02


83 Fig ure 3 13 Reduced time to peak severity in both Mbnl2 2/ 2 and Mbnl2 +/ 2 mice. The time to maximum seizure score (>Level 5) was recorded for both Mbnl2 +/ 2 and Mbnl2 2/ 2 mice: Mbnl2 +/ 2 ( n = 6 for), Mbnl2 2/ 2 ( n = 4), p<0.05.


84 CHAPTER 4 MBNL2 REGULATES ALTERNATIVE SPLICING IN THE BRAIN Previous studies have shown that Mbnl1 acts as a developmental switch, by adapting the transcriptome for the demands of adulthood th rough the regulation of alternative splicing. Our discovery that Mbnl2 is a nuclear protein in neurons of the brain suggested that this Mbnl family member might serve a similar role in the CNS. To test the hypothesis that Mbnl2 regulates alternative splici ng in the brain, the WT and Mbnl2 2/ 2 transcriptomes were analyzed by splicing microarrays and RNA seq using hippocampal RNAs to detect all possible RNA targets affected by Mbnl2 loss. The reasons why I focused on the hippocampus were: 1) Mbnl2 is localized to the nucleus in the hippoc ampus supporting the hypothesis that Mbnl2 is an alternative splicing factor; 2) the hippocampus is a small region in the brain with well defined function in learning and memory. While this experimental strategy may have missed RNA processing events that o ccur in other regions of the brain, hippocampal cells have a closely related function which can be tested by behavioral assays To validate the targets that were found by both splicing microarrays and RNA seq, splicing sensitive PCR analysis was performed for the top targets. Results Alternative Splicing Dysregulation in Mbnl2 Knockout Brain The microarray analysis revealed hundreds of genes (Figure A 1 ) with alternative splicing abnormalities in the adult hippocampus but only a handful in the muscle and i n both arrays the number of genes with expression level changes was insignificant (data not shown). More than 38% of the genes that were revealed by the microarray analysis were also shown to be mis regulated by RNA seq analysis and subsequent RT PCR


85 ana lysis ( Figure 4 1 4 2 ) Interestingly, grouping the top microarray and RNA seq targets according to their function and known disease association by gene ontology ( Fig ure 4 2 ) revealed the possible functional implications of Mbnl2 loss in learning and memo ry formation as well as channel sensitivity that results in seizure formation Overall, this analysis shows that t he majority of the genes that are affected by loss of Mbnl2 are involved in axonal growth and neuronal plasticity Intronic Enrichment of YGC Y Clusters i n Mbnl2 Knockout Mis Regulated Exons Previously it has been shown that Mbnl1 regulates alternative splicing of muscle specific genes such as the muscle specific chloride channel Clcn 1 by direct binding to YGCY clusters (Goers et al., 2010) Other RNA binding proteins show similar short motif specificity such as NOVA that binds UCAY motifs (Jensen et al., 2000b) and the polypyrimidine track binding protein (PTB) that binds YCTY (Garcia Blanco et al., 1989) Mbnl2 shows a great degree of conservation with Mbnl1 and has a similar zinc finger domain structure so I hypothesized that Mbnl2 might also bind preferentially to YGCY motifs. To test this hypothesis, we searched for YGCY motif enrichment and other RNA binding motifs that have been previously found to be cis ac ting elements for other RNA binding proteins in the intronic regions upstream and downstream of the cassette exons mis regulated in Mbnl2 knockouts ( Figure 4 3 ). We found a significant increase o nly in YGCY clusters in the mis regulated genes providing su ppo rting evidence that Mbnl1 and Mbnl2 bind RNA via the YGCY motif In contrast only a few targets were found to be mis regulated and these did not enrichment for YGCY motifs ( M. Cline, personal communication).


86 Direct Binding Sites for Mbnl2 on Mis Regula ted Targets To test the hypothesis that Mbnl2 binds directly to target transcripts identified by microarray and RNA seq analyses, high throughput sequencing crosslinking immunoprecipitation (HITS CLIP) (Licatalosi et al., 2008) was used to identify direct binding targets. Briefly, hippocampi from WT and Mbnl2 2/ 2 mice were dissected, pulverized in liquid nitrogen and the resulting powder was crosslinked with UV light to generate covalent crosslinks between RNA and associated proteins. After labeling the RNA with 32 P and immunoprecipitation of Mbnl2 with mAb 3B4, the RNA was partially digested with RNAse A and then the protein was removed with Proteinase K. The released RNA was reverse transcribed, and subjected to high throughput sequencing. The sequencing reads (30 60 nt long), or CLIP tags were subjected to filtering for imperfect matches, exact duplicates and multiple genomic hits to achieve greater specificity. Mbnl2 2/ 2 reads were used as negative, non specific binding tags. Sequence tags were then mapped back to the mouse genome using Novoalign as described previously (Zhang and Darnell) The majority of the reads (>50%) were known genes (22% and 9% respectively ) (data not shown). To define sequences where Mbnl2 binds in protein cod ing genes, overlapping tag sequences (>2) were grouped to clusters using the UCSC genome browser ( http://genome.ucsc.edu/ ) as previously described (Yeo et al., 2009) ( Figure 4 4 A ). Unique clusters (20,441) were found to overlap with protein coding genes which under higher stringency (Bonferroni correction p<0.01) yielde and intronic regions of the genes verifying an alternative splicing role for Mbnl2. After combining the findings of splicing microarrays, RNA seq and CLIP, an RNA splicing


87 map was generat ed to correlate binding of Mbnl2 and splicing regulation (Licatalosi et al., 2008) ( Figure 4 4 B ). To generate the normalized Mbnl2 splicing map, all the CLIP tags for each transcript were normalized to 1.0 and the f raction of the normalized tags presented in 50 nucleotide windows upstream and downstream of the aberrant spliced transcripts were clustered summed and multiplied by the number of transcripts they were present This normalization will strengthen tags that occurred in multiple transcripts in the same region and resulted in a similar shift in alternative splicing. We exon skipping while binding 60 70 nt downstream of the 5 inclusion. To validate the hypothesis that Mbnl2 binds to YGCY clusters we employed crosslinking induced mutation site (CIMS) analysis of the peak CLIP targets as previously described (Zhang and Darnell) Clustering of 62,932 deletions in unique tags yielded 557 CIMS (FDR<0.0001) with deletions observed in at least 4 tags (n 4) with base composi tion 4%A, 33%C, 22%G, 41%U (Figure 4 5 A). To reveal motif enrichment, de novo motif analysis was performed by assessing 21 bp around the deletion site. The majority of CIMS clusters were located in a UCGU motif which showed 16 fold enr ichment completed to flanking sequences used as controls ( Figure 4 5 B ). Discussion The microarray and RNA seq data revealed a set of mis spliced RNAs encoding ion channels previously implicated in spatial memory and seizure prone phenotypes. Deletion of a different alternative splicing factor, Rbfox1, affects the alternative splicing of same genes in a similar manner and results mice that are susceptible to spontaneous and kainic acid induced seizures (Gehman et al., 2011) Two genes that are highly


88 altered in both mouse models are Cacna1d and Grin1, a calcium ion channel and an NMDA receptor respectively. Involvement of Grin1 in seizure formation has never been reported and will not be further discussed in this chapt er On the other hand, Canca1d is not only involved in seizure susceptibility but also in memory consolidation. According to RT PCR analysis, Canca1d shows the highest splicing mis regulation with >60% shift in alternative splicing of exon 12a ( Figure 4 1 ) Cacna1d is a fast voltage gated L type calcium channel that controls presynaptic glutamate release. Its role is to decode sound evoked depolarization of inner hair cells to increased calcium influx resulting in glutamate vesicle fusion at the presynaptic membrane, excitation of the postsynaptic membrane and propagation of the signal through the auditory pathway. Cacna1d is also highly expressed in atrial myocytes and cells of the sinoatrial and atrioventricular node, controlling the pacemaker activity of the heart (Mangoni et al., 2003) Recently a trinucleotide insertion in alternative exon 8b has been linked to deafness and irregularities in the sinoatrial node, the heart pacemaker of patients with SANDD syndrome (Baig et al., 2011) Individuals with SANDD syndrome suffer from hearing loss and bradycardia. In a mouse model study, Cacna1d / mice were viable and suffered from sinoatrial dysfunction and congenital deafness (Platzer et al., 2000) similar to humans with SANDD. In the hippocampus the role of Cacna1d has not been yet elucidated, despite its high expression levels and wide cellular distribution pattern. Some studies have rep orted protein protein interactions of the C termini of Cacna1d channels with Shank, a postsynaptic adaptor protein (Zhang et al., 2005) and the N termini with the ryanodine receptor type 2, which plays a role in Ca + + release from intracellular storage region (Kim


89 et al., 2007) Isoform expression level changes have been reported after pilocarpine induced status epilepticus involving Ca ++ channels in neuronal plasticity after e pilepsy (Xu et al., 2007) Two more genes with potentially high impact on the seizure prone phenotype are Dlg2 and Kcnma1 Dlg2 belongs to a family of intracellular scaffolding proteins known as membrane associated guanylate kinase (MAGUK) protein family. The role of the MAGUKs is to provide a postsynaptic scaffolding network that drives localization and specific density of different receptors, for example NMDA receptor NR2B ( Sans et al., 2003) and AMPA receptor subunits (Rumbaugh et al., 2003) to neuronal spines. To provide binding sites for protein protein interactions, MAGUKs are composed of several different binding domains, inclu ding a triple PDZ domain, and a non enzymatic guanylate kinase fragment. The diverse spatial and temporal expression of different MAGUK isoforms provides a mechanism to alter the cellular and sub cellular localization of receptors in neurons changing the s ensitivity and downstream pathway of signaling cascades. MAGUK protein change of function has been reported to be disease (Lacor et al., 2004) L DOPA induced diskynesia of Parkinson patients under L DOPA treatment (Gardoni et al., 2006) and Huntington disease (Sun et al., 2001) Specific isoforms of Digl2 has been found to bind the inwardly rectifying potassium channel Kir2.1 of the postsynaptic membrane (Leyland and Dart, 2004) The density and intracellular localization of Kir proteins determine the resting membra ne potential of neurons. Additional evidence also indicates alteration of Dlg2, and its binding partner NR2B, expression levels in temporal lobe related epilepsy (Liu et al., 2007) A change in


90 Dlg2 splicing isoform levels and potential change in their cellular localization or even binding preferences for NMDA, Kir and AMPA receptor subunits may underlie a change in postsynaptic sensitivity and hyperexcitability. Kcnma1, or Slo1, belongs to the Slo family of potassiu m channels characterized by high single channel conductance with a voltage range of activation that is modifiable in response to many factors and ions. Kcnma1 is expressed in numerous tissues including the mammalian CNS, pancreas, hair cells and smooth mus cle and can be activated by either a membrane depolarization, an increase in the intracellular calcium concentration or by a synergistic effect of both (Magleby, 2003) The gating characteristics are subject not only to changes due to extrinsic factors but also by pre and post translational modifications including alternative splicing, phosphorylation and heteromultimer f ormation (Joiner et al., 1998; Ramanathan et al., 1999; Schubert and Nelson, 2001) In our report, we show a change in alternative splicing of Kcnma1 in the hippocampus which could lead to an abnormal sensitivity to calcium levels and/or to the kinetics of the channel. Studies in C. elegans have shown that deletion of Kcnma1 leads to increased neurotransmitter release (Wang et al., 2001) whereas in a mouse Kcnma1 knockout mode l the phenotype was more diverse including high frequency hearing loss, erectile dysfunction, ataxia and vascular hypertension (Ruttiger et al., 2004; Sausbier et al., 2004; Werner et al., 2005) In humans, missense mutations lead to epilepsy (Du et al., 2005) The three most highly mis regulated genes in this study that influence learning and memory are Cacna1d, Tanc2 and Ndrg4. Tanc2 and Tanc1 are paralogous genes. Tanc1 is a scaffold/adaptor protein that interacts directly with PSD 95, a member of the


91 MAGUK family of proteins, and with other proteins via its multiple protein protein interaction domains (Suzuki et al., 2005) Tanc1 and Tanc2 overexpression in neurons in vitro results in an increase of spine density and excitatory synapses whereas Tanc1 knockout mice show impaired spatial memory and learning. Tanc2 deletion is embryonic lethal (Ha n et al.) Similar spatial memory impairment was shown in an Ndrg4 knockout mouse model. The Ndrg4 (N Myc downstream regulated gene 4) protein is the only member of the Ndrg family that is specifically expressed in brain and heart (Zhou et al., 2001) and undergoes extensive isoform regulation during development (Nakada et al., 2002) Downregulation of Ndrg4 results in decreased neurite growth and process length in an in vitro cell system (Ohki et al., 2002) and Ndrg4 levels are downregulated in (Zhou et al., 2001) suggesting a role in neuronal growth and development. Indeed, Ndrg4 deletion in mice resulted in inferior spatial learning and memory formation as well as increased neuronal damage and lesions following focal ischemia (Yamamoto et al.) Overall, the splicing microarray and RNA seq findings agree with the phenotype characterization of the Mbnl2 2/ 2 mouse model. Changes in the coding sequence of a protein isoform can alter its function and alter downstream molecular cascades. For instance, expression of a fetal isoform might lead to enhanced nonsense mediated decay and thus phenocopy a null mu tation. HITS CLIP and CIMS analyses validated the direct binding of Mbnl2 to aberrantly spliced targets These results are similar to the binding map of Nova (Licatalosi et al., 2008) which suggests a common mechanis m by which RNA binding proteins act to regulate alternative splicing Binding of a splicing


92 factor upstream of an exon promotes skipping of the exon and binding of the splicing factor downstream promotes in clusion.


93 Fig ure 4 1 Mbnl2 regulates neonatal to adult alternative splicing changes in the hippocampus. Five highly scored targets were tested by splicing sensitive RT PCR analysis with primers flanking the alternative exon. (Upper) For Mbnl2 alternative splicing, Ndrg4 and Tanc2 splicing was tested in WT, Mbnl1 3 / 3 and Mbnl2 2/ hippocampi, including the developmental splicing change in postnatal (P) day 6 and 42 of mouse for e brain (fb) and hindbrain (hb). (Lower) to assess possible splicing changes due to Mbnl2 haploinsufficiency alternative splicing of C anca1d, Kcnma1 and Dlg2 were tested in WT, Mbnl2 + / 2 and Mbnl2 2/ hippocampi ( c ourtesy of Kuang Yung Lee )


94 Fig ure 4 2 RNA seq and microarray data show possible learning/memory and epileptic defects in Mbnl2 2/ 2 mice. A) 104 splicing micro array targets with a sep score 0.7 were compared with 179 RNA seq targets (FDR<0.05, DI 0.1). B) Top RNA seq and microarray targets are grouped according to their function and gene ontology.


95 Fig ure 4 3 Mbnl2 regulates exon exclusion in the hippocampus for exons that show YGCY enrichment in the upstr eam introns. Enrichment of YGCY motifs was assayed for introns upstream (left) and downstream (right) of the misregulated exons in Mbnl2 2/ 2 hippocampi. Each point represents the average frequency of the motif s present at the specific site upstream of the intron exon junction (0). Error bars indicate +2/ 2 s.d. of the mean frequency distribution for the population of background exons ( C ourtesy of M. Cline ).


96 Fig ure 4 4 Mbnl2 RNA splicing map. (A) UCSC genom e browser representation of Ndrg4 alternative exon 14 showing CLIP tags (above) in different colors for different biological replicates and DNA conservation in blue bars for high and red bars for low (bellow). B) Normalized complexity Mbnl2 RNA map sh owing exon activation (red) and repression (blue) ( C ourtesy of Chaolin Zhang and Yuan Yuan )


97 Fig ure 4 5 Mbnl2 binds directly to UGCU clusters A) Alignment of 557 CIMS (FDR<0.001) with deletions observed in at least 4 tags ( n composition of t hese CIMS is 4% A, 33% C, 22%,G, 41 % U. B) UGCU enrichment around CIMS (blue line) and CLIP cluster peaks (orange) compared to surrounding sequences used as control ( C ourtesy of Chaolin Zhang and Yuan Yuan ).


98 CHAPTER 5 CONCLUDING REMARKS A ND FUTURE DIRECTIO NS Microsatellite repeat expansion mutations are responsible for many neurological abnormalities including certain types of spinocerebellar ataxias (SCAs), amyloid lateral pansion mutations can be toxic at a D NA, RNA or protein level depending on the ir location relevant to the affected gene A recently described RNA gain of function model implicates repeat expansion in untranslated regions of mRNAs, in sequestration of splicing factors lead ing to D M. Sequestrat ion of these alternative splicing factors causes aberrant splicing of genes, adding DM in the group of diseases termed spliceopathies. By s tudying the molecular events that lead to DM and similar diseases, we can create universal models of how expansion mu tations interfere with normal cell function. These universal models are useful in understanding the molecular basis of newly described disorders caused by expansion mutations, as well as a frame work to design better therapeutic strategies. A lso by identify ing and associating specific gene perturbations with certain clinical features of diseases can speed up narrowing down the genetic basis of many unidentified so far genetic disorders. In this study I report the functional role of Mbnl2 in the brain and a lso demonstrate the mechanism of action by determining the RNA binding sites. The Mbnl protein family shows remarkable genetic and protein similarity among the three paralogs and a high degree of functional domain conservation. The high degree of similarit y suggests functional redundancy in vivo but my studies indicate temporal or spatial functional specialization for the three family members. Mbnl2 appears to play a predominant role in developmentally regulated alternative splicing in the brain while Mbnl1 subserves the


99 same function in skeletal muscle For Mbnl2, the majority of the RNA targets identified by splicing microarrays and RNA seq are involved neuronal plasticity, synaptic and post synaptic function and neurogenesis. Thus, this study has identifi ed key elements that could reveal novel brain related defects in DM patients. Identification of the UGCU binding motif by HITS CLIP strengthens the hypothesis that Mbnl2 binds and is sequestered by toxic CUG repeat expansions. In this study, I also investi gated a possible role for Mbnl2 loss in the pathogenesis of brain related DM clinical features by of learning and memory defects in Mbnl2 knockouts. The results show that loss of Mbnl2 in mice affects their memory formation and spatial precision, a higher function of the brain suggesting possible executive function deficits. The unexpected discovery of seizure susceptibility in Mbnl2 knockouts raises an issue about drug administration to DM patients. Mbnl2 knockout mice show a dramatic increase in sensitivi ty to the GABAergic antagonist PTZ, suggestive of hyperexcitability in the DM nervous system. Although few reports of DM associated seizures exist, seizures could undermine the mental health of DM patients and also be induced by anti depressant or anti arr hythmic drugs. Despite the success of the Mbnl2 knockout model to recapitulate the DM brain aberrant splicing and reveal hundreds of new possible target genes implicated in the disease, there are certain drawbacks. There are phenotypes and molecular event s that could be unrelated to DM and the expression pattern of the human DMPK and CNBP which carry the expansion mutation. One possible scenario for the unexpected seizure express Dmpk or Cnbp. Also in the constitutive Mbnl2 knockouts the protein is

PAGE 100

100 completely absent during embryogenesis and adulthood. In DM we have partial but not complete sequestration of Mbnl proteins as shown by the strong correlation of repeat length a nd severity of the phenotype. A more relevant to the disease mouse model would be a conditional Mbnl2 mouse where exon 2 deletion is driven under either D mpk or C nbp promoters Another possible scenario and drawback of the Mbnl2 knockout and any other mous e model is the genetic difference between humans and mice. Genes that can be affected in mice may not be in humans due to sequence variation in coding regions and regulatory cis acting elements that regulate gene expression and alternative splicing. Future studies should focus on identifying the neuronal circuitry that is affected in the Mbnl2 KO mice by direct electrode stimulation of hippocampal slices to study changes in LTP formation and consolidation as well as NMDA and AMPA sensitivity changes. In add ition, the contribution of mis regulated targets identified during this study to seizure formation should be tested using blocking morpholinos to force fetal splicing patterns in adults. To establish a stronger connection between the seizure prone phenotyp e and myotonic dystrophy, the spatial expression C(C)UG repeats should be examined in further detail in the DM brain and in DM mouse models that express the repeats under DMPK or CNBP promoters. Beyond teasing out the molecular etiology of already identif ied behavioral effects of Mbnl2 loss, the Mbnl2 knockout mice are a great tool to address quest ions relevant to other clinical DM features. Another area of focus is DM associated hypersomnia. The microarray data revealed several candidate genes that are im plicated in circadian rhythm establishment and maintenance. Thus, Mbnl2 2/ mice should be tested for

PAGE 101

101 sleeping or circadian rhythm perturbations. Also DM is not only considered to be a late onset neuromuscular disease, but also a premature aging disease Most DM clinical features are natural signs of aging such as cataracts, loss of brain and muscle mass as well as heart conduction defects. In concordance with this notion the weak musc le defects present in the Mbnl2 GT2 mice were observed in 9 10 month o ld animals, suggesting a potential aging effect in disease progression due to Mbnl2 loss. For that reason i t is of great importance to study the aging process of the Mbnl2 2/ mice by analyzing the survival rate, late onset neuromuscular disorders, and possible declines in brain function in aged mice On the other hand, DM is thought to be caused by a combinatorial loss of function of all three Mbnl proteins. The generati on of each isoform specific Mbnl knockout m ouse line serves to understand the unique functions and contribution to disease pathogenesis of each Mbnl2 isoform but do es not mimic the DM sequestration model. To achieve a more complete DM mouse model all thr ee Mbnl knockouts should be mated and triple Mbnl knockout line s should be generated. A d rawback of this strategy is the likelihood that embryonic lethality will result from double or triple knockout generation The severity and onset of DM is correlated w ith the expansion of the repeats suggesting a partial protein sequestration. Complete ablation of these splicing factors is probably incompatible with life. To overcome this problem and recapitulate partial sequestration, tissue specific knockouts can be g enerated to avoid embryonic lethality and tripl y heterozygous knockouts could recapitulate the partial loss of functions of these proteins. Finally both isoform specific and triple knockout models could serve as a clinical tool for drug discovery and gene therapy testing Mbnl1 upregulation by gene therapy

PAGE 102

102 revert s most of the muscle related abnormalities in mice that express the toxic repeats in muscle (Kanadia et al., 2006) Mbnl2 and Mbnl3 upregulation has not bee n tested, but according to the structural and functional similarity between these proteins similar results are anticipated In conclusion, the muscleblind family of RNA binding proteins shows structural and functional similarity and each paralog is respon sible for the regulation of transcriptome plasticity either during different developmental periods or in different tissue s (Fig. 5). Mbnl1 and Mbnl2 regulate the embryonic to adult shift of specific alternative exons in muscle and brain, respectively, and are active during postnatal life. In contrast, Mbnl3 functions during the embryonic period. This functional distinction serves a s a robust paradigm for fine tuning the transcriptome by related RNA binding proteins during development.

PAGE 103

103 Fig ure 5 1 Functi onal diversion and tissue distribution among the three Mbnl proteins.

PAGE 104

104 CHAPTER 6 MATERIALS AND METHODS Lac Z Staining M ixed C57Bl6/129 male Mbnl2 +/GT4 mice (3 5 months of age) were obtained from Charles Thornton (University of Rochester) These mice were g enerated by integration of the GT vector pGTOpfs in Mbnl2 intron 4 B rains were dissected and incubated overnight at 4 0 C in 0.2% PFA in PBS followed by 30% sucrose solution in PBS overnight at 4 0 C and embedded in OCT (Tissue Tek) Transverse cryostat sect ions (15 m ) were dried (30 min, RT) and incubated overnight in X Gal staining s olution (0.1M sodium phosphate, pH 7.4 0.1% sodium deoxycholate, 2 mM MgCl, 0.2% NP 40, 1 mg/ml X Gal, 0.1 mM K 3 Fe(CN) 6 0.1 mM K 4 Fe(CN) 6 ) at 37 0 C in a humidified chamber. The stained sections were further fixed for 1 hr in 2% PFA for 10 min at RT For counterstaining the sections were incubated 30 s in water, 1min in 95% EtOH, 20 s in e osin solution (Fisher Richard A llan S cientific C at# 71311 Eosin Y with phloxine ) 3 times in 95% EtOH for 1 min each, 3 times in 100% EtOH for 3 min each, 3 times in Citrisolv (Fisher C at# 22 143975 Citrisolv) for 5 min each and finally mounted on a slide with P ermoun t. Generation of an Mbnl 2 P olyclonal A ntibody Plasmid Transfections COSM6 cel ls were grown in growth media containing DMEM (Invitrogen), 10% FBS (Invitrogen), 1% L glutamine (Invitrogen), and 1% penicillin/streptomycin (Invitrogen) in a humidified 37 C, 5% CO 2 incubator. The day before transfection cells were seeded in each well of a 6 well plate. We avoided antibiotics in the growth media prior to plasmid transfections. After 16 hr a mixture was added consisting of

PAGE 105

105 transfection reagent (6 L Fugene 6 [ Roche ] in 180 L OptiMEM I [ Invitrogen ] ) followed by vortex ing for 5 s, incubat ion at RT for 5 min, addition of 2 g of plasmid followed by vortex ing 2 sec and incubat ion at RT for 15min and expression vectors (empty vector pSP72, myc Mbnl1, my c Mbnl2 and myc Mbnl3) that were provided by Mike Poulos and described previously. After 24 hr, the cells were washed with PBS and the antibiotic free media was replaced. All wells were processed 48 hr post transfection for western blot analysis or immunof luorescence Western Blotting For western blot analysis each well was washed with PBS followed by cell lysis in H EPES buffer (20 mM Hepes, 100 mM KCl PH=8.0, PicD, PicW) at 4 0 C, centrifugation at 16,000 RCF at 4C 0 and collection of the supernatant. To tes t the specificity of the new polyclonal antibody, 50 from each transfection was fractionated on a 12.5 Tris glycine polyacrylamide gel, transfe rred to nitrocellulose membrane, blocked for 30 min with 5% skimmed milk in PBS and dried ov ernight. For immunobloting, the nitrocellulose membrane was incubated for 5 min in PBS followed by overnight incubation at 4 0 C with the new Mbnl2 specific antibody (1:1000) in 5% milk/0.05% Igepal/PBS while two more replicates were incubated with anti Myc antibody (1: 500 ) for expression levels of the myc tagged proteins and anti Gapdh (1:5000) (loading control). After washing 3 times with PBS/0.05% Igepal, the membranes were incubated with a horse radish peroxidase linked secondary anti rabbit (Mbnl2) or an ti mouse (anti Myc and anti Gapdh) antibody (1:5000 GE) in 5% milk/0.05% Igepal/PBS for 2 hr, washed three times with PBS/ 0.05% I gepal and processed with ECL reagents according to

PAGE 106

106 Bioscience ). Visualization of chemiluminescence was achieved using BioMax film ( Kodak ). Mbnl2 2/ 2 M ouse G eneration ES Cell Targeting Construct The genomic background of the Mbnl2 targeting vector sequence was obtained from a n SV 129 BAC clone and generation was performed with standard recombineering techniques using protocols 1 4 ( http://web.ncifcrf.gov/ research/brb/protocol.aspx ) and the following reagents: bacterial strains SW102, SW10. Plasmids PL253, PL451, PL452. Briefly isolated BAC plasmid was electroporated in SW102 cell s. The targeting backbone PL253 that carries the negative selection marker Herpes S i mplex virus thymidine kinase (HSV TK), was linearized by PCR amplification with primers that carry arms of homology (AH) to a 10 kb fragment that contains Mbnl2 exon 2 Gap re pair fwd: agcctgccgtgagagagtgaagtcatcagcctccagccacctgacttccgcggacagtggtctgctctccc a GCCAGGGTTTTCCCAGTCACGACGTTGT Gap repair rev: ttggctcacctcccacctttacctgctttatgtcattttccgtataaaaaggggaacgctcctccctc GTTACCC AACTTAATCGCCTTGCAGCACATCC B lack smal l letters indicate the homologous region to Mbnl2 and capital bold letters indicate the homologous region to the plasmid backbone. All targeting events to generate the construct were performed in SW102 after heat induction of the P1 phage recombination sys tem and correct ly targeted plasmid clones were verified with Xba I restriction analysis. The 10 kb Mbnl2 fragment retrieval via gap repair was also verified by PCR amplification of the PL253 Mbnl2 I

PAGE 107

107 targeted a neomycin selection marker floxed by two loxP sites, 250 bp upstream of Mbnl2 exon 2. This cassette was PCR amplified from plasmid PL45 2 with primers that have AH for the exon 2 upstream region. 1 st LoxP targeting fwd: tcctctaagtacagacgacgcaagtcgatgtca tactgttttAGGCCTataacttcgtataatgtatgcta tacga agttat CGACCTGCAGCCTGTTGA 1 ST LoxP targeting rev: ggtgccaatgagagtgatctccgggacaactttaaaccacaaacc tgcataacttcgtatagcatacattat acga agttat GTCGAGGCTGATCAGCGA C apital letters indicate a Stu I restriction site that we introduced and was used to by Southern blotting. Cre mediated recombination and excision of the neo cassette followed, by electroporat ion of the positively targeted plasmid into a rabino se inducible Cre expressing bacteria l PCR amplifying a neo cassette flanked by two Frt sites and one l oxP site directly from plasmid PL45 1 with the following primers 2 nd l oxP targeting fwd: aggtttatgttgtcttttggtttggttttcactgaaacttatttatac GAAGTTCCTATTCTCTAGA AAGTATAGGAACTTC AGG TCTGAAGAGGAGTT T 2 nd loxP targeting rev : ttaatttaaacacagatatggaaaaaaaaatttttgtatgcctaAGTACT ATAAC TTCGTAT AGCATACATTATACGAAGTTATATTATGTACCTGACTG

PAGE 108

108 In black capital letters is an exogenous Sca I site that was introduced to screen ES by Southern blotting. The functionality of both final l oxP and Frt sites were tested in arabinose inducible Cre or Flp expressi ng cells respectively (SW106 and SW105). ES Cell Targeting The targeting construct was linearized with Not I and electroporated in 129 SvlmJ ES cells. ES clones were cultured and selected as previously described. Positive clones that survived G418 and neo mycin selection were picked and screened by Southern blotting to identify correct ly ES cell genomic DNA was digested with Stu I used Sca a 300 base probe was used that was PCR amplified from genomic DNA AH C TCTCCTCCTCCAGT TTGGCTTTG GTGTTCAGACGTGAGCCAAGA CTGTAAC The targeted allele yields a 5 kb band and the WT allele a 6.4 kb band. used a 250 base probe that was PCR TCGACTTCCTCCATTCTGGGAGAAC GCATCCAGG GGACAATTCACATAGA The targeted allele gives a band of 5 kb and the WT yields a band of 31 kb. ES cells from a positive clone were injected in C57BL/6 pseudopregnant female mice (University of Michigan) and chimeric male animals where acquired and mated to C57BL/6 female s Germline transmission of the conditional allele should yield ed ago uti pups. To further test for correct germline transmission agouti pups from the F1 generation were genotyped with two forward primers (one in the neo CGCCTTCTATCGCCTTCTTGACGAGTTCTTCTGAG and one in Mbnl2 exon 2 GTAGGGCTCTCAAGGAGAGCACTGCAT TGAGC in 2:1 ratio) and one

PAGE 109

109 reverse primer in Mbnl2 AATGTCAAACCAGACCAGAAATACACCACCATG ). Amplification was performed for 3 3 cy cles (each cycle consisting of 94 0 C for 30 sec, 64 0 C for 30 sec and 70 0 C for 50 sec). Mating Scheme and Genotyp ing To create a constitutive Mbnl2 knockout mouse line, Mbnl2 +/con mice were mated with B6.C Tg( CMV cre) 1Cgn/J mice. F1 progeny was genotyped for CMV Cre by PCR using primers in the Cre cassette and for Mbnl2 exon 2 deletion with two forward primers (one u GTACCACCTTCCTTGTG ATACTGAAAGCTCTG AGGTC Mbnl2 GTAGGGCTCTCAAGGA GAGCACTGCATTG AGC Mbnl2 AATGTCAAAC CAGACCAGAAATACA CCCATG Amplification was performed in 3 3 cy cles (each cycle consisting of 94 0 C for 30 sec, 65 0 C for 30 sec and 72 0 C for 30 sec). To obtain Mbnl2 2/ 2 mice, heterozygous Mbnl2+ / 2 mice were crossed and the pups were genotyped with the same primers. RNA Analysis of Mbnl2 2 / 2 Mice Loss of Mbnl2 exon 2 was also as sayed at the mRNA level. M ale Bl6/129 Mbnl2 +/+ Mbnl2 2 and Mbnl2 / mice (5 months of age) were sacrificed and total RNA was isolated from the brain and TA muscles by homogenizing the tissue in Tri reagent (Sigma) ocol. First strand cDNA synthesis was performed using reverse transcription of 2.5 0 .5 mM dNTPs, 0.5 dT, 0.01 M DTT, 40 U RNasin, 200 U of Superscript III reverse polymerase (20 l total reaction volume) The RT mixt ure was incubated at 25 0 C for 5 min followed by 42 0 C for 1 hr and 72 0 C for 15 min. For Mbnl2 expression analysis of cDNA was used for PCR

PAGE 110

110 CAGTCAAGAGACTAGAACCCTGG GAGC GGCGTTCCTGGAAACA TAAA p ia was used as a loading control with f or w ar d GCGGCAGGTCCATCTACG GCCATCCAGCCATTCAGTCT Mbnl1 levels were also tested with a forward primer GTTAGTGTCACACCAATTCGGGACAC a reverse primer in exon GGGCATCATGGCATTGGCTAAC on a 1% agarose gel and visualized with e thidium b romide. Protein Expression Analysis of Mbnl2 2 / 2 Mice F emale Bl6/129 Mbnl2 +/+ and Mbnl2 / mice (2 3 months of age) were sacrificed and protein isolated by homogenizing the different tissues (cerebellum, hippocampus, frontal lobe, heart, lung, quadriceps and spleen) in H EPES buffer as described a bove Western blot analysis was performed with a commercial anti Mbnl2 antibody (1:1000 mAb 3B4 Santa Cruz Biotechnology 1:5000 secondary anti mouse antibody ) as described above Gapd h was used as a loading control ( 1:10000 mAb 6C5 primary, 1:5000 seconda ry anti mouse antibodies ) Muscle Immunofluorescence and Histology F emale Bl6/129 Mbnl2 +/+ and Mbnl2 / mice (2 3 months of age) were sacrificed and quadriceps muscles were dissected, mounted on wooden dowels with 10% gum tragacanth (in PBS) and frozen in a liquid nitrogen precooled isopentane bath. For histological examination of the muscle sections, I emplo y ed H&E staining. Briefly, 10 m cryosections mounted on microscope slides were air dried for 30 min, followed by a standard H&E staining protocol and mounted with Permount (Fisher cat# sp15 100 Permount) For all light image visualization of sectio ns and digital photography, we used

PAGE 111

111 a Leica DM 2000 inverted light microscope with a Qimaging MicroPublisher 5.0 RTV digital camera attached Brain Immunofluorescence B L 6/129 Mbnl2 +/+ and Mbnl2 / mice (5 months of age) were sacrificed and the brains di ssected and fixed at 4C 0 overnight in 4% paraformaldehyde in PBS. Sectioning was performed on a Vibratome at 4C 0 and different regions of the brain (sections were obtained at 70Hz vibration frequency, 0.2mm/sec cutti ng speed and sections were kept floating at 4 0 C in PBS). For immunostaining sections were washed briefly in water and microwaved in a 0.1M Urea solution 3 times for 15sec each followed by 2 min cooling down intervals in a conventional microwave oven. Sect ions were blocked overnight in a 0.3% T riton X 100 0.2% goat serum PBS solution and then incubated in primary anti Mbnl2 antibody (1:200 of mAb 3B4) in blocking solution for 3 days followed by washing ( X 20 min ) in PBS and incubation in secondary anti mo use antibody (1:400 A lexa 488) for 48 hr All incubations were performed at 4 0 C Sections were counterstained and mounted on slides with VectaShield mounting media with DAPI. Microscopy was performed on a Zeis Axioskope II inverted fluorescence microscope and a Leica TCS SP5 confocal microscope. Nissl Staining Vibratome sections were obtained as previously described from male Bl6/129 Mbnl2 +/+ and Mbnl2 / mice (5 months of age). Sections were dried on microscope slides for 15 min and then submerged in the following sequence of solutions: 95% EtOH 15 min, 70% EtOH 1 min, 50% EtOH 1 min, dH 2 O 2 min, dH 2 O 1 min, Cresyl violet stain 3.5 min (0.25 gm cresy l violet acetate, 170 l glacial acetic acid in 50 ml dH 2 O), dH 2 O 1

PAGE 112

112 min, 50% EtOH 1 min, 70% acid EtOH (1 ml glacial acetic acid in 100 ml 70% EtOH), 95% EtOH 2 min, 95% EtOH (few dips), 100% EtOH 1 min, Citrisolv 5 min and then mounted with Permount. Mor ris Water Maze Test Animals I used mixed WT and Mbnl2 2 / 2 mice (2 5 months of age) that were housed in cages with littermates in a room maintained at 22 0 C and on a normal 12/12 hr light/dark cycle. The genetic background is C57BL/6 and 129SvlmJ Water m aze: The water maze was a round galvanized trough, 1.22m in diameter silver color and 0.6 meters tall, filled with water at 25+ / 2 0 C W ater transparency was reduced with white tempera paint (RichArt ) for contrast purposes between the mouse body and the wa ter surface The maze was located in a dedicated test room with reduced noise where screens and 2D and 3D cues were strategically placed around it The platform was made from clear Plexiglas ( 10cm diameter ) and was submerged 1.5 c m below the water level s o that it is not visible to a viewer on the surface of the water. The surface of the platform was scraped so that mice could obtain grip and climb on it. A water pump was used to pump out water from the trough and add fresh water every day. Training For ea ch training day the mice were acclimatized before the test in the water maze room in their original cage for 30 min. Each mouse was placed in a clear b u cket with long handle to avoid direct interaction with the animals and placed in the tank facing the tan k wall The location that the mice were released was randomly chosen from one of the three opposite compass points from the quadrant where the platform was located.

PAGE 113

113 This point was matched for all the mice on the same trial. Mice were trained for 4 consecut ive days. Each mouse had 4 trials per day in sets of two consecutive trials with 1 min interval Each set was separated by 4 hours Each mouse was allowed to swim for 60 s per trial or until it reached the platform. Mice that did not reach the platform wit hin the first 60 s were guided to and allowed to rest on it for 10 s. After the end of each trial mice were allowed to sit on the platform and orient themselves for 10 s. After each trial mice were returned to a separate cage to dry under a heating bulb. The last day ( Day 6 ) the platform was removed and m ice were allowed to swim for 60 s and then removed from the maze. Data Acquisition and Analysis Data w as collected with a wired camera mounted on the ceiling above the tank. Data acquisition was recorded automatically by a computer The software (EthoVision 3 Noldus information Technology) began recording 1 s after the mouse was released in the tank and stop ped recording after the mouse was located on the platform for more than 1.5 s. The software recorde d the total path length the mouse swam and the latency time to the platform. The maze was divided in 4 quadrants (NW, SW, NE, and SE ) and a 10 cm thigmotactic zone was defined with a circle that was drawn 10 cm inside of the tank wall. The software zoning was used to calculate latency to platform, percentage of target quadrant occupancy, thigmotactic occupancy, swimming speed and total number of platform cross overs when the platform was removed. The platform was located in the SW quadrant.

PAGE 114

114 Statistical ana lysis : A two sided student T test was performed to obtain statistically significant difference between the two groups. The total number of mice that were used in this study is n =16 for WT and n =13 for Mbnl2 2 / 2 mice. PTZ Seizure Susceptibility Test M ixed WT, Mbnl2 + 2 and Mbnl2 2/ 2 mice (2 5 months of age) were tested for seizure susceptibility by injection with pentylenetetrazol (PTZ). Each mouse was weigh ed to determine the PTZ solution (4 mg/ml) in jection volume. Minimal PTZ dose concentration was determined by injecting 3 WT mice with different concentrations of PTZ (40, 60, 70 mg/kg). At 40 mg/kg, WT mice showed no reaction (Racine score=1) and all further injections were carried out under these c onditions. WT and Mbnl2 2/ 2 mice were injected at the same time. Mbnl2 + 2 mice were injected separately. Following each injection, the mice were placed in an observational area (cage) for a maximum time of 60 min and the time of onset of convulsive behavior and nature/sever ity of the convulsion was scored. Seizure severity was determined using a modified Racine 0 6 scoring scale and statistical significant values were determined with a Mann Whitney test. Racine scale : 0, no motor seizures; 1, freezing, staring, mouth or faci al movements; 2, head nodding or isolated twitches and rigid posture,; 3, tail extension unilateral bilateral forelimb clonus; 4, rearing and mice sit in an immobile state on their rear haunches with one or both forelimbs extended; 5: clonic seizures with loss of posture, jumping, falling; 6) tonic seizures with hindlimb extension and death. The observation period was immediately followed by euthanasia. Splicing Microarray Quadriceps muscle and hippocampi RNA was extracted from 3 month old WT and Mbnl2 2/ 2 mixed male mice as previously described. To identify differential alternative

PAGE 115

115 splicing between WT and Mbnl2 2/ 2 mice we processed the RNA samples for standards of the manufacturer (Sugnet et al., 2006) Data analysis: The analysis of the microarrays was performed as described previously (Sugnet et al., 2006) The separation score was calc ulated as the ratio between the skipped to included exon ratio of the mutant by the WT. Sepscore=log 2 [Mut(skip/include)/WT(skip/include)] Analysis of sequence motifs: the analysis of sequence motifs was performed as described previously (Du et al., 2010) Briefly, the 150 bp upstream and downstream intronic region of the differentially spliced exon was analyzed with Improbizer. As background sequence we used the upstream and downstream intronic regions of exons that did not show any difference between WT and KO animals. P value was set at <0.05. RT PCR Splicing Analysis of Mbnl2 Targets Hippocampi from 2 5 month old male or female WT, Mbnl 1 E3 / 3 and Mbnl2 2/ 2 mice was dissected and processed as described above to generate cDNA. The primer set that was used to test for splicing changes of mis regulated exons are shown in T able 6 1 To test developmental shifts in alternative exon utilizat ion, we used cDNA generated from WT and Mbnl2 2/ 2 forebrain and hindbrain at P6 and P42. HITS CLIP Hippocampi were dissected from WT mice (11 12 weeks of age) as described above and snap frozen in liquid nitrogen in a pre cooled metal mortar. Each biolo gical replicate (4 hippocampi from 2 mice) was processed separately on different days to account for diurnal changes in Mbnl2 expression. Hippocampi were ground to a fine

PAGE 116

116 powde r using a pre co oled pestle and the tissue slurry in liquid nitrogen was transfe rred to a glass petri dish kept on dry ice. After liquid nitrogen evaporation, the powdered 1800 (Stratagene). The biological replicates were saved at 80C. CLIP was performed as previously re ported (Jensen and Darnell, 2008; Licatalosi et al., 2008; Ule et al., 2005) with modifications described below. The anti Mbnl2 antibody mAb 3B4 was used Protei n A (Invitrogen) completely cleared the Mbnl2 from each biological replicate (equivalent to 2 mg total protein). Hippocampi from Mbnl2 mice were used as a control for the immunoprecipitation. RNase A concentrations were optimized and final concentra tions of 5000U/mL and 0.06U/mL were used as high and low RNase levels respectively The cDNA libraries were generated using RNA linkers and primers described for Ago CLIP (Chi et al., 2009) to enable better quantitation of CLIP tags. Library concentrations were estimated with Quant iT dsDNA assay kit, high sensitivity (Invitrogen, Catalog # Q33120). Libraries were diluted to 9 nM before submitting for clonal cluster generation with an Illumina cBot followed by sequencing using an Illumina Genome Analyzer IIx. Raw sequence reads were filtered to remove unmappable, rRNA and PCR duplicate reads. Unique reads were aligned to the mm9 mouse genome database on the UCSC Genome Browser for visualization.

PAGE 117

117 Table 6 1. Mbnl2 target gene RT PCR analysis primer sets Gene exon bp forward primer reverse primer Tanc2 23a 30 gccatgattgagcatgttgactacagt (in exon 22: 133 bp) cctcttccatcagctt gctcaaca (in exon 23b: 92 bp) Kcnma1 25a 81 gattcacacctcctggaatggacagat (in exon 24: 114 bp) gtgaggtacagctctgtgtcagggtcat (in exon 25b: 131 bp) Limch1 9 36 cggaagttgccagatgtgaagaaa (in exon 8: 253 bp) cctcctcacaccgcatgtcaaa (in exon 10: 127 bp) Clas p2 16a,16b 27_27 gttgctgtgggaaatgccaagac (in exon 15: 102 bp) gctccttgggatcttgcttctcttc (in exon 16c: 171 bp) Spna2 23 60 gattggtggaaagtggaagtgaatgac (in exon 22: 149 bp) tgatccactgctgtaactcgtttgct (in exon 24: 199 bp) St3gal3 3 48 gcctcttcctggtcctggga ttt (in exon 2: 145 bp) caggaggaagcccagcctatcatact (in exon 4: 43 bp) Ndrg4 14 39 cttcctgcaaggcatgggctaca (in exon 13: 52 bp) gggcttcagcaggacacctccat (in exon 15: 1942 bp) Csnk1d 9 63 gatacctctcgcatgtccacctcaca (in exon 8: 140 bp) gcattgtctgcccttcaca gcaaa (in exon 10: 2166 bp) Ppp1r12a 14 171 caagcaccacatcaacaccaacagtt (in exon 13: 168 bp) cttcgtccctaacaggagtgaggtatga (in exon 15: 91 bp) Cacna1d 12a 60 catgcccaccagcgagactgaa (in exon 11: 88 bp) caccaggacaatcaccagccagtaaa (in exon 13: 161 bp) Add1 15 37 ggatgagacaagagagcagaaagagaaga (in exon 14: 99 bp) ctgggaaggcaagtgcttctgaa (in exon 16: 1836 bp) Mbnl1 7 54 ggctgcccaataccaggtcaac (in exon 6: 258 bp) gggagaaatgctgtatgctgctgtaa (in exon 8: 154 bp) Grin1 4 63 tcatcctgctggtcagcgatgac (in exon 3:17 7 bp) agagccgtcacattcttggttcctg (in exon 5: 101 bp) Camk2d 14b, 15,16 33_60_42 cagccaagagtttattgaagaaaccaga (in exon 14a: 38 bp ) ctttcacgtcttcatcctcaatggtg (in exon 17: 49 bp) Mapt 3,4 87_87 aagaccatgctggagattacactctgc (in exon 2: 113 bp) ggtgtctccga tgcctgcttctt (in exon 5: 66 bp) Ryr2 4,5 21_15 cggacctgtctatctgcacctttgt (in exon 3: 105 bp) cataccactgtaggaatggcgtagca (in exon 6: 75 bp) Dlg2 17b 42 ccattctacaagaacaaggagcagagtga (in exon 17a: 100 bp) gcctcgtgacaggttcataggaaaga (in exon 18: 51 bp)

PAGE 118

118 A PPENDIX MICROARRAY SPLICING RESULTS Figure A 1. Top targets of splicing sensitive microarray with sepscore 2.38 1.00

PAGE 119

119 Figure A 1. Continued

PAGE 120

120 Figure A 1. Continued

PAGE 121

121 LIST OF REFERENCES Abe, K., H. Fujimura, K. Toyooka, S. Yorifuji Y. Nishikawa, T. Hazama, and T. Yanagihara. 1994. Involvement of the central nervous system in myotonic dystrophy. J Neurol Sci 127:179 85. Adereth, Y., V. Dammai, N. Kose, R. Li, and T. Hsu. 2005. RNA dependent integrin alpha3 protein localization regu lated by the Muscleblind like protein MLP1. Nat Cell Biol 7:1240 7. Adie, W.J. 1923. Dystrophia Myotonica (Myotonia Atrophica), an Heredo familial Disease with Cataract. Proc R Soc Med 16:36 44. Ahima, R.S., D. Prabakaran, and J.S. Flier. 1998. Postnatal leptin surge and regulation of circadian rhythm of leptin by feeding. Implications for energy homeostasis and neuroendocrine function. J Clin Invest 101:1020 7. Artero, R., A. Prokop, N. Paricio, G. Begemann, I. Pueyo, M. Mlodzik, M. Perez Alonso, and M. K. Baylies. 1998. The muscleblind gene participates in the organization of Z bands and epidermal attachments of Drosophila muscles and is regulated by Dmef2. Dev Biol 195:131 43. Attig, L., G. Solomon, J. Ferezou, L. Abdennebi Najar, M. Taouis, A. Gertler and J. Djiane. 2008. Early postnatal leptin blockage leads to a long term leptin resistance and susceptibility to diet induced obesity in rats. Int J Obes (Lond) 32:1153 60. Bache, R.J., and G.A. Sarosi. 1968. Myotonia atrophica: diagnosis in patient wi th complete heart block and Stokes Adams syncope. Arch Intern Med 121:369 72. Baig, S.M., A. Koschak, A. Lieb, M. Gebhart, C. Dafinger, G. Nurnberg, A. Ali, I. Ahmad, M.J. Sinnegger Brauns, N. Brandt, J. Engel, M.E. Mangoni, M. Farooq, H.U. Khan, P. Nurnb erg, J. Striessnig, and H.J. Bolz. 2011. Loss of Ca(v)1.3 (CACNA1D) function in a human channelopathy with bradycardia and congenital deafness. Nat Neurosci 14:77 84. Balthasar, N., R. Coppari, J. McMinn, S.M. Liu, C.E. Lee, V. Tang, C.D. Kenny, R.A. McGo vern, S.C. Chua, Jr., J.K. Elmquist, and B.B. Lowell. 2004. Leptin receptor signaling in POMC neurons is required for normal body weight homeostasis. Neuron 42:983 91. Barbosa, J., F.Q. Nuttall, W. Kennedy, and F. Goetz. 1974. Plasma insulin in patients w ith myotonic dystrophy and their relatives. Medicine (Baltimore) 53:307 23. Barreca, T., A. Muratorio, A. Sannia, L. Murri, B. Rossi, and E. Rolandi. 1980. Evaluation of twenty four hour secretory patterns of growth hormone and insulin in patients with my otonic dystrophy. J Clin Endocrinol Metab 51:1089 92.

PAGE 122

122 Batten, F.E., and H.P. Gibb. 1909. Two Cases of Myotonia Atrophica, showing a peculiar Distribution of Muscular Atrophy. Proc R Soc Med 2:32 3. Begemann, G., N. Paricio, R. Artero, I. Kiss, M. Perez A lonso, and M. Mlodzik. 1997. muscleblind, a gene required for photoreceptor differentiation in Drosophila, encodes novel nuclear Cys3His type zinc finger containing proteins. Development 124:4321 31. Bouret, S.G., S.J. Draper, and R.B. Simerly. 2004. Form ation of projection pathways from the arcuate nucleus of the hypothalamus to hypothalamic regions implicated in the neural control of feeding behavior in mice. J Neurosci 24:2797 805. Brandt, R., J. Leger, and G. Lee. 1995. Interaction of tau with the neu ral plasma membrane mediated by tau's amino terminal projection domain. J Cell Biol 131:1327 40. Brigman, J.L., T. Wright, G. Talani, S. Prasad Mulcare, S. Jinde, G.K. Seabold, P. Mathur, M.I. Davis, R. Bock, R.M. Gustin, R.J. Colbran, V.A. Alvarez, K. Na kazawa, E. Delpire, D.M. Lovinger, and A. Holmes. 2010. Loss of GluN2B containing NMDA receptors in CA1 hippocampus and cortex impairs long term depression, reduces dendritic spine density, and disrupts learning. J Neurosci 30:4590 600. Brook, J.D., M.E. McCurrach, H.G. Harley, A.J. Buckler, D. Church, H. Aburatani, K. Hunter, V.P. Stanton, J.P. Thirion, T. Hudson, and et al. 1992. Molecular basis of myotonic dystrophy: expansion of a trinucleotide (CTG) repeat at the 3' end of a transcript encoding a prot ein kinase family member. Cell 68:799 808. Buckanovich, R.J., and R.B. Darnell. 1997. The neuronal RNA binding protein Nova 1 recognizes specific RNA targets in vitro and in vivo. Mol Cell Biol 17:3194 201. Buckanovich, R.J., J.B. Posner, and R.B. Darnel l. 1993. Nova, the paraneoplastic Ri antigen, is homologous to an RNA binding protein and is specifically expressed in the developing motor system. Neuron 11:657 72. Buckanovich, R.J., Y.Y. Yang, and R.B. Darnell. 1996. The onconeural antigen Nova 1 is a neuron specific RNA binding protein, the activity of which is inhibited by paraneoplastic antibodies. J Neurosci 16:1114 22. Cantwell, A.R., Jr., and W.B. Reed. 1965. Myotonia atrophica and multiple calcifying epithelioma of Malherbe. Acta Derm Venereol 45:387 90. Chi, S.W., J.B. Zang, A. Mele, and R.B. Darnell. 2009. Argonaute HITS CLIP decodes microRNA mRNA interaction maps. Nature 460:479 86.

PAGE 123

123 Cottrell, E.C., R.L. Cripps, J.S. Duncan, P. Barrett, J.G. Mercer, A. Herwig, and S.E. Ozanne. 2009. Developme ntal changes in hypothalamic leptin receptor: relationship with the postnatal leptin surge and energy balance neuropeptides in the postnatal rat. Am J Physiol Regul Integr Comp Physiol 296:R631 9. Cuchalova, L., T. Kouba, A. Herrmannova, I. Danyi, W.L. Ch iu, and L. Valasek. 2010. The RNA recognition motif of eukaryotic translation initiation factor 3g (eIF3g) is required for resumption of scanning of posttermination ribosomes for reinitiation on GCN4 and together with eIF3i stimulates linear scanning. Mol Cell Biol 30:4671 86. Damian, M.S., G. Bachmann, D. Herrmann, and W. Dorndorf. 1993. Magnetic resonance imaging of muscle and brain in myotonic dystrophy. J Neurol 240:8 12. de Andrade, G.R., and R.P. Jansen. 2005. Moving with Muscleblind. Nat Cell Biol 7:1155 6. DeChiara, T.M., A. Efstratiadis, and E.J. Robertson. 1990. A growth deficiency phenotype in heterozygous mice carrying an insulin like growth factor II gene disrupted by targeting. Nature 345:78 80. Delaporte, C. 1998. Personality patterns in p atients with myotonic dystrophy. Arch Neurol 55:635 40. Du, H., M.S. Cline, R.J. Osborne, D.L. Tuttle, T.A. Clark, J.P. Donohue, M.P. Hall, L. Shiue, M.S. Swanson, C.A. Thornton, and M. Ares, Jr. Aberrant alternative splicing and extracellular matrix gene expression in mouse models of myotonic dystrophy. Nat Struct Mol Biol 17:187 93. Du, H., M.S. Cline, R.J. Osborne, D.L. Tuttle, T.A. Clark, J.P. Donohue, M.P. Hall, L. Shiue, M.S. Swanson, C.A. Thornton, and M. Ares, Jr. 2010. Aberrant alternative splici ng and extracellular matrix gene expression in mouse models of myotonic dystrophy. Nat Struct Mol Biol 17:187 93. Du, W., J.F. Bautista, H. Yang, A. Diez Sampedro, S.A. You, L. Wang, P. Kotagal, H.O. Luders, J. Shi, J. Cui, G.B. Richerson, and Q.K. Wang. 2005. Calcium sensitive potassium channelopathy in human epilepsy and paroxysmal movement disorder. Nat Genet 37:733 8. Evans, W. 1944. The Heart in Myotonia Atrophica. Br Heart J 6:41 7. Falcon, J., L. Besseau, M. Fuentes, S. Sauzet, E. Magnanou, and G. Boeuf. 2009. Structural and functional evolution of the pineal melatonin system in vertebrates. Ann N Y Acad Sci 1163:101 11. Fearnsides, E.G. 1915. Myopathy (? Myotonia Atrophica) with a Family History of Cataracts but no History of Familial Myopathy. P roc R Soc Med 8:39 43.

PAGE 124

124 Fox, H.C. 1910. Paralysis of the Right Vocal Cord in a Case of Myotonia Atrophica. Proc R Soc Med 3:20 1. Garcia Blanco, M.A., S.F. Jamison, and P.A. Sharp. 1989. Identification and purification of a 62,000 dalton protein that bind s specifically to the polypyrimidine tract of introns. Genes Dev 3:1874 86. Gardoni, F., B. Picconi, V. Ghiglieri, F. Polli, V. Bagetta, G. Bernardi, F. Cattabeni, M. Di Luca, and P. Calabresi. 2006. A critical interaction between NR2B and MAGUK in L DOPA induced dyskinesia. J Neurosci 26:2914 22. Geh, J.L., and A.L. Moss. 1999. Multiple pilomatrixomata and myotonic dystrophy: a familial association. Br J Plast Surg 52:143 5. Gehman, L.T., P. Stoilov, J. Maguire, A. Damianov, C.H. Lin, L. Shiue, M. Ares, Jr., I. Mody, and D.L. Black. 2011. The splicing regulator Rbfox1 (A2BP1) controls neuronal excitation in the mammalian brain. Nat Genet 43:706 11. Giubilei, F., G. Antonini, S. Bastianello, S. Morino, A. Paolillo, M. Fiorelli, C. Ferretti, and C. Fiesch i. 1999. Excessive daytime sleepiness in myotonic dystrophy. J Neurol Sci 164:60 3. Goers, E.S., J. Purcell, R.B. Voelker, D.P. Gates, and J.A. Berglund. 2010. MBNL1 binds GC motifs embedded in pyrimidines to regulate alternative splicing. Nucleic Acids R es 38:2467 84. Goers, E.S., R.B. Voelker, D.P. Gates, and J.A. Berglund. 2008. RNA binding specificity of Drosophila muscleblind. Biochemistry 47:7284 94. Gomes Pereira, M., L. Foiry, A. Nicole, A. Huguet, C. Junien, A. Munnich, and G. Gourdon. 2007. CTG trinucleotide repeat "big jumps": large expansions, small mice. PLoS Genet 3:e52. Grammatikakis, I., Y.H. Goo, G.V. Echeverria, and T.A. Cooper. 2011. Identification of MBNL1 and MBNL3 domains required for splicing activation and repression. Nucleic Acid s Res 39:2769 80. Guiraud Dogan, C., A. Huguet, M. Gomes Pereira, E. Brisson, G. Bassez, C. Junien, and G. Gourdon. 2007. DM1 CTG expansions affect insulin receptor isoforms expression in various tissues of transgenic mice. Biochim Biophys Acta 1772:1183 91. Han, S., J. Nam, Y. Li, S. Kim, S.H. Cho, Y.S. Cho, S.Y. Choi, J. Choi, K. Han, Y. Kim, M. Na, H. Kim, Y.C. Bae, and E. Kim. Regulation of dendritic spines, spatial memory, and embryonic development by the TANC family of PSD 95 interacting proteins. J Neurosci 30:15102 12.

PAGE 125

125 Hao, M., K. Akrami, K. Wei, C. De Diego, N. Che, J.H. Ku, J. Tidball, M.C. Graves, P.B. Shieh, and F. Chen. 2008. Muscleblind like 2 (Mbnl2) deficient mice as a model for myotonic dystrophy. Dev Dyn 237:403 10. Hashimoto, T., M. T ayama, M. Miyazaki, K. Murakawa, H. Kawai, H. Nishitani, and Y. Kuroda. 1995a. Neuroimaging study of myotonic dystrophy. I. Magnetic resonance imaging of the brain. Brain Dev 17:24 7. Hashimoto, T., M. Tayama, M. Miyazaki, K. Murakawa, H. Kawai, H. Nishit ani, and Y. Kuroda. 1995b. Neuroimaging study of myotonic dystrophy. II. MRI measurements of the brain. Brain Dev 17:28 32. Ho, T.H., B.N. Charlet, M.G. Poulos, G. Singh, M.S. Swanson, and T.A. Cooper. 2004. Muscleblind proteins regulate alternative splic ing. EMBO J 23:3103 12. Holt, I., V. Jacquemin, M. Fardaei, C.A. Sewry, G.S. Butler Browne, D. Furling, J.D. Brook, and G.E. Morris. 2009. Muscleblind like proteins: similarities and differences in normal and myotonic dystrophy muscle. Am J Pathol 174:21 6 27. Holt, S.J., and P.W. Sadler. 1958a. Studies in enzyme cytochemistry. II. Synthesis of indigogenic substrates for esterases. Proc R Soc Lond B Biol Sci 148:481 94. Holt, S.J., and P.W. Sadler. 1958b. Studies in enzyme cytochemistry. III. Relationship s between solubility, molecular association and structure in indigoid dyes. Proc R Soc Lond B Biol Sci 148:495 505. Ishihara, T., B. Zhang, M. Higuchi, Y. Yoshiyama, J.Q. Trojanowski, and V.M. Lee. 2001. Age dependent induction of congophilic neurofibrill ary tau inclusions in tau transgenic mice. Am J Pathol 158:555 62. Jensen, K.B., and R.B. Darnell. 2008. CLIP: crosslinking and immunoprecipitation of in vivo RNA targets of RNA binding proteins. Methods Mol Biol 488:85 98. Jensen, K.B., B.K. Dredge, G. Stefani, R. Zhong, R.J. Buckanovich, H.J. Okano, Y.Y. Yang, and R.B. Darnell. 2000a. Nova 1 regulates neuron specific alternative splicing and is essential for neuronal viability. Neuron 25:359 71. Jensen, K.B., K. Musunuru, H.A. Lewis, S.K. Burley, and R .B. Darnell. 2000b. The tetranucleotide UCAY directs the specific recognition of RNA by the Nova K homology 3 domain. Proc Natl Acad Sci U S A 97:5740 5. Jiang, H., A. Mankodi, M.S. Swanson, R.T. Moxley, and C.A. Thornton. 2004. Myotonic dystrophy type 1 is associated with nuclear foci of mutant RNA, sequestration of muscleblind proteins and deregulated alternative splicing in neurons. Hum Mol Genet 13:3079 88.

PAGE 126

126 Joiner, W.J., M.D. Tang, L.Y. Wang, S.I. Dworetzky, C.G. Boissard, L. Gan, V.K. Gribkoff, and L .K. Kaczmarek. 1998. Formation of intermediate conductance calcium activated potassium channels by interaction of Slack and Slo subunits. Nat Neurosci 1:462 9. Jouvenceau, A., L.H. Eunson, A. Spauschus, V. Ramesh, S.M. Zuberi, D.M. Kullmann, and M.G. Hann a. 2001. Human epilepsy associated with dysfunction of the brain P/Q type calcium channel. Lancet 358:801 7. Kanadia, R.N., K.A. Johnstone, A. Mankodi, C. Lungu, C.A. Thornton, D. Esson, A.M. Timmers, W.W. Hauswirth, and M.S. Swanson. 2003a. A muscleblind knockout model for myotonic dystrophy. Science 302:1978 80. Kanadia, R.N., J. Shin, Y. Yuan, S.G. Beattie, T.M. Wheeler, C.A. Thornton, and M.S. Swanson. 2006. Reversal of RNA missplicing and myotonia after muscleblind overexpression in a mouse poly(CUG) model for myotonic dystrophy. Proc Natl Acad Sci U S A 103:11748 53. Kanadia, R.N., C.R. Urbinati, V.J. Crusselle, D. Luo, Y.J. Lee, J.K. Harrison, S.P. Oh, and M.S. Swanson. 2003b. Developmental expression of mouse muscleblind genes Mbnl1, Mbnl2 and Mbn l3. Gene Expr Patterns 3:459 62. Kash, S.F., R.S. Johnson, L.H. Tecott, J.L. Noebels, R.D. Mayfield, D. Hanahan, and S. Baekkeskov. 1997. Epilepsy in mice deficient in the 65 kDa isoform of glutamic acid decarboxylase. Proc Natl Acad Sci U S A 94:14060 5 Kessler, S.H., and A.B. Sachs. 1998. RNA recognition motif 2 of yeast Pab1p is required for its functional interaction with eukaryotic translation initiation factor 4G. Mol Cell Biol 18:51 7. Kim, J.S., S.L. Coon, J.L. Weller, S. Blackshaw, M.F. Rath, M Moller, and D.C. Klein. 2009. Muscleblind like 2: circadian expression in the mammalian pineal gland is controlled by an adrenergic cAMP mechanism. J Neurochem 110:756 64. Kim, S., H.M. Yun, J.H. Baik, K.C. Chung, S.Y. Nah, and H. Rhim. 2007. Functional interaction of neuronal Cav1.3 L type calcium channel with ryanodine receptor type 2 in the rat hippocampus. J Biol Chem 282:32877 89. Kino, Y., C. Washizu, Y. Oma, H. Onishi, Y. Nezu, N. Sasagawa, N. Nukina, and S. Ishiura. 2009. MBNL and CELF proteins regulate alternative splicing of the skeletal muscle chloride channel CLCN1. Nucleic Acids Res 37:6477 90. Koshelev, M., S. Sarma, R.E. Price, X.H. Wehrens, and T.A. Cooper. 2010. Heart specific overexpression of CUGBP1 reproduces functional and molecular abnormalities of myotonic dystrophy type 1. Hum Mol Genet 19:1066 75.

PAGE 127

127 Kuyumcu Martinez, N.M., G.S. Wang, and T.A. Cooper. 2007. Increased steady state levels of CUGBP1 in myotonic dystrophy 1 are due to PKC mediated hyperphosphorylation. Mol Cell 28:68 78. Lacor, P.N., M.C. Buniel, L. Chang, S.J. Fernandez, Y. Gong, K.L. Viola, M.P. Lambert, P.T. Velasco, E.H. Bigio, C.E. Finch, G.A. Krafft, and W.L. Klein. 2004. Synaptic targeting by Alzheimer's related amyloid beta oligomers. J Neurosci 24:10191 200. Lagos Quintana, M., R. Rauhut, W. Lendeckel, and T. Tuschl. 2001. Identification of novel genes coding for small expressed RNAs. Science 294:853 8. Lau, D., E.C. Vega Saenz de Miera, D. Contreras, A. Ozaita, M. Harvey, A. Chow, J.L. Noebels, R. Paylor, J. I. Morgan, C.S. Leonard, and B. Rudy. 2000. Impaired fast spiking, suppressed cortical inhibition, and increased susceptibility to seizures in mice lacking Kv3.2 K+ channel proteins. J Neurosci 20:9071 85. Lecourtier, L., and P.H. Kelly. 2005. Bilateral l esions of the habenula induce attentional disturbances in rats. Neuropsychopharmacology 30:484 96. Lecourtier, L., H.C. Neijt, and P.H. Kelly. 2004. Habenula lesions cause impaired cognitive performance in rats: implications for schizophrenia. Eur J Neuro sci 19:2551 60. Lee, K.S., Y. Cao, H.E. Witwicka, S. Tom, S.J. Tapscott, and E.H. Wang. 2010. RNA binding protein Muscleblind like 3 (MBNL3) disrupts myocyte enhancer factor 2 (Mef2) {beta} exon splicing. J Biol Chem 285:33779 87. Lee, K.S., R.M. Squilla ce, and E.H. Wang. 2007. Expression pattern of muscleblind like proteins differs in differentiating myoblasts. Biochem Biophys Res Commun 361:151 5. Leroy, O., C.M. Dhaenens, S. Schraen Maschke, K. Belarbi, A. Delacourte, A. Andreadis, B. Sablonniere, L. Buee, N. Sergeant, and M.L. Caillet Boudin. 2006a. ETR 3 represses Tau exons 2/3 inclusion, a splicing event abnormally enhanced in myotonic dystrophy type I. J Neurosci Res 84:852 9. Leroy, O., J. Wang, C.A. Maurage, M. Parent, T. Cooper, L. Buee, N. Ser geant, A. Andreadis, and M.L. Caillet Boudin. 2006b. Brain specific change in alternative splicing of Tau exon 6 in myotonic dystrophy type 1. Biochim Biophys Acta 1762:460 7. Leyland, M.L., and C. Dart. 2004. An alternatively spliced isoform of PSD 93/ch apsyn 110 binds to the inwardly rectifying potassium channel, Kir2.1. J Biol Chem 279:43427 36.

PAGE 128

128 Licatalosi, D.D., A. Mele, J.J. Fak, J. Ule, M. Kayikci, S.W. Chi, T.A. Clark, A.C. Schweitzer, J.E. Blume, X. Wang, J.C. Darnell, and R.B. Darnell. 2008. HITS CLIP yields genome wide insights into brain alternative RNA processing. Nature 456:464 9. Lin, X., J.W. Miller, A. Mankodi, R.N. Kanadia, Y. Yuan, R.T. Moxley, M.S. Swanson, and C.A. Thornton. 2006. Failure of MBNL1 dependent post natal splicing transiti ons in myotonic dystrophy. Hum Mol Genet 15:2087 97. Liu, F.Y., X.F. Wang, M.W. Li, J.M. Li, Z.Q. Xi, G.M. Luan, J.G. Zhang, Y.P. Wang, J.J. Sun, and Y.L. Li. 2007. Upregulated expression of postsynaptic density 93 and N methyl D aspartate receptors subun its 2B mRNA in temporal lobe tissue of epilepsy. Biochem Biophys Res Commun 358:825 30. Liu, J.P., J. Baker, A.S. Perkins, E.J. Robertson, and A. Efstratiadis. 1993. Mice carrying null mutations of the genes encoding insulin like growth factor I (Igf 1) a nd type 1 IGF receptor (Igf1r). Cell 75:59 72. Magleby, K.L. 2003. Gating mechanism of BK (Slo1) channels: so near, yet so far. J Gen Physiol 121:81 96. Mahadevan, M., C. Tsilfidis, L. Sabourin, G. Shutler, C. Amemiya, G. Jansen, C. Neville, M. Narang, J Barcelo, K. O'Hoy, and et al. 1992. Myotonic dystrophy mutation: an unstable CTG repeat in the 3' untranslated region of the gene. Science 255:1253 5. Mangoni, M.E., B. Couette, E. Bourinet, J. Platzer, D. Reimer, J. Striessnig, and J. Nargeot. 2003. Fu nctional role of L type Cav1.3 Ca2+ channels in cardiac pacemaker activity. Proc Natl Acad Sci U S A 100:5543 8. Mankodi, A., E. Logigian, L. Callahan, C. McClain, R. White, D. Henderson, M. Krym, and C.A. Thornton. 2000. Myotonic dystrophy in transgenic mice expressing an expanded CUG repeat. Science 289:1769 73. Mankodi, A., M.P. Takahashi, H. Jiang, C.L. Beck, W.J. Bowers, R.T. Moxley, S.C. Cannon, and C.A. Thornton. 2002. Expanded CUG repeats trigger aberrant splicing of ClC 1 chloride channel pre mRN A and hyperexcitability of skeletal muscle in myotonic dystrophy. Mol Cell 10:35 44. Mankodi, A., C.R. Urbinati, Q.P. Yuan, R.T. Moxley, V. Sansone, M. Krym, D. Henderson, M. Schalling, M.S. Swanson, and C.A. Thornton. 2001. Muscleblind localizes to nucle ar foci of aberrant RNA in myotonic dystrophy types 1 and 2. Hum Mol Genet 10:2165 70. Mastrogiacomo, I., G. Bonanni, E. Menegazzo, C. Santarossa, E. Pagani, M. Gennarelli, and C. Angelini. 1996. Clinical and hormonal aspects of male hypogonadism in myoto nic dystrophy. Ital J Neurol Sci 17:59 65.

PAGE 129

129 Matsumoto, M., and O. Hikosaka. 2007. Lateral habenula as a source of negative reward signals in dopamine neurons. Nature 447:1111 5. McCormick, K.A., J. Srinivasan, K. White, T. Scheuer, and W.A. Catterall. 199 9. The extracellular domain of the beta1 subunit is both necessary and sufficient for beta1 like modulation of sodium channel gating. J Biol Chem 274:32638 46. Meola, G., V. Sansone, D. Perani, S. Scarone, S. Cappa, C. Dragoni, E. Cattaneo, M. Cotelli, C. Gobbo, F. Fazio, G. Siciliano, M. Mancuso, E. Vitelli, S. Zhang, R. Krahe, and R.T. Moxley. 2003. Executive dysfunction and avoidant personality trait in myotonic dystrophy type 1 (DM 1) and in proximal myotonic myopathy (PROMM/DM 2). Neuromuscul Disord 13:813 21. Miller, J., A.D. McLachlan, and A. Klug. 1985. Repetitive zinc binding domains in the protein transcription factor IIIA from Xenopus oocytes. EMBO J 4:1609 14. Miller, J.W., C.R. Urbinati, P. Teng Umnuay, M.G. Stenberg, B.J. Byrne, C.A. Thornto n, and M.S. Swanson. 2000. Recruitment of human muscleblind proteins to (CUG)(n) expansions associated with myotonic dystrophy. EMBO J 19:4439 48. Mistry, A.M., A. Swick, and D.R. Romsos. 1999. Leptin alters metabolic rates before acquisition of its anore ctic effect in developing neonatal mice. Am J Physiol 277:R742 7. Modoni, A., G. Silvestri, M.G. Pomponi, F. Mangiola, P.A. Tonali, and C. Marra. 2004. Characterization of the pattern of cognitive impairment in myotonic dystrophy type 1. Arch Neurol 61:1 943 7. Moxley, R.T., 3rd, R.C. Griggs, D. Goldblatt, V. VanGelder, B.E. Herr, and R. Thiel. 1978. Decreased insulin sensitivity of forearm muscle in myotonic dystrophy. J Clin Invest 62:857 67. Nakada, N., S. Hongo, T. Ohki, A. Maeda, and M. Takeda. 2002. Molecular characterization of NDRG4/Bdm1 protein isoforms that are differentially regulated during rat brain development. Brain Res Dev Brain Res 135:45 53. Noebels, J.L. 2003. The biology of epilepsy genes. Annu Rev Neurosci 26:599 625. Ohki, T., S. Ho ngo, N. Nakada, A. Maeda, and M. Takeda. 2002. Inhibition of neurite outgrowth by reduced level of NDRG4 protein in antisense transfected PC12 cells. Brain Res Dev Brain Res 135:55 63. Ono, S., K. Inoue, T. Mannen, F. Kanda, K. Jinnai, and K. Takahashi. 1 987. Neuropathological changes of the brain in myotonic dystrophy -some new observations. J Neurol Sci 81:301 20.

PAGE 130

130 Oyamada, R., M. Hayashi, Y. Katoh, K. Tsuchiya, T. Mizutani, I. Tominaga, and H. Kashima. 2006. Neurofibrillary tangles and deposition of oxi dative products in the brain in cases of myotonic dystrophy. Neuropathology 26:107 14. Paul, M.S., and B.L. Bass. 1998. Inosine exists in mRNA at tissue specific levels and is most abundant in brain mRNA. EMBO J 17:1120 7. Peltz, S.W., A.H. Brown, and A. Jacobson. 1993. mRNA destabilization triggered by premature translational termination depends on at least three cis acting sequence elements and one trans acting factor. Genes Dev 7:1737 54. Philips, A.V., L.T. Timchenko, and T.A. Cooper. 1998. Disruptio n of splicing regulated by a CUG binding protein in myotonic dystrophy. Science 280:737 741. Phillips, M.F., H.M. Steer, J.R. Soldan, C.M. Wiles, and P.S. Harper. 1999. Daytime somnolence in myotonic dystrophy. J Neurol 246:275 82. Platzer, J., J. Engel, A. Schrott Fischer, K. Stephan, S. Bova, H. Chen, H. Zheng, and J. Striessnig. 2000. Congenital deafness and sinoatrial node dysfunction in mice lacking class D L type Ca2+ channels. Cell 102:89 97. Query, C.C., R.C. Bentley, and J.D. Keene. 1989. A comm on RNA recognition motif identified within a defined U1 RNA binding domain of the 70K U1 snRNP protein. Cell 57:89 101. Ramanathan, K., T.H. Michael, G.J. Jiang, H. Hiel, and P.A. Fuchs. 1999. A molecular mechanism for electrical tuning of cochlear hair c ells. Science 283:215 7. Ranum, L.P., and T.A. Cooper. 2006. RNA mediated neuromuscular disorders. Annu Rev Neurosci 29:259 77. Reddy, S., D.B. Smith, M.M. Rich, J.M. Leferovich, P. Reilly, B.M. Davis, K. Tran, H. Rayburn, R. Bronson, D. Cros, R.J. Balic e Gordon, and D. Housman. 1996. Mice lacking the myotonic dystrophy protein kinase develop a late onset progressive myopathy. Nat Genet 13:325 35. Renwick, J.H., and D.R. Bolling. 1971. An analysis procedure illustrated on a triple linkage of use for pren atal diagnosis of myotonic dystrophy. J Med Genet 8:399 406. Ronnblom, A., S. Andersson, P.M. Hellstrom, and A. Danielsson. 2002. Gastric emptying in myotonic dystrophy. Eur J Clin Invest 32:570 4. Ronnblom, A., H. Forsberg, and A. Danielsson. 1996. Gast rointestinal symptoms in myotonic dystrophy. Scand J Gastroenterol 31:654 7.

PAGE 131

131 Rosahl, T.W., D. Spillane, M. Missler, J. Herz, D.K. Selig, J.R. Wolff, R.E. Hammer, R.C. Malenka, and T.C. Sudhof. 1995. Essential functions of synapsins I and II in synaptic ve sicle regulation. Nature 375:488 93. Roses, A.D., M.A. Pericak Vance, D.A. Ross, L. Yamaoka, and R.J. Bartlett. 1986. RFLPs at the D19S19 locus of human chromosome 19 linked to myotonic dystrophy (DM). Nucleic Acids Res 14:5569. Ross, A.F., Y. Oleynikov, E.H. Kislauskis, K.L. Taneja, and R.H. Singer. 1997. Characterization of a beta actin mRNA zipcode binding protein. Mol Cell Biol 17:2158 65. Ruiz Echevarria, M.J., and S.W. Peltz. 2000. The RNA binding protein Pub1 modulates the stability of transcripts containing upstream open reading frames. Cell 101:741 51. Rumbaugh, G., G.M. Sia, C.C. Garner, and R.L. Huganir. 2003. Synapse associated protein 97 isoform specific regulation of surface AMPA receptors and synaptic function in cultured neurons. J Neuros ci 23:4567 76. Ruttiger, L., M. Sausbier, U. Zimmermann, H. Winter, C. Braig, J. Engel, M. Knirsch, C. Arntz, P. Langer, B. Hirt, M. Muller, I. Kopschall, M. Pfister, S. Munkner, K. Rohbock, I. Pfaff, A. Rusch, P. Ruth, and M. Knipper. 2004. Deletion of t he Ca2+ activated potassium (BK) alpha subunit but not the BKbeta1 subunit leads to progressive hearing loss. Proc Natl Acad Sci U S A 101:12922 7. Sans, N., K. Prybylowski, R.S. Petralia, K. Chang, Y.X. Wang, C. Racca, S. Vicini, and R.J. Wenthold. 2003. NMDA receptor trafficking through an interaction between PDZ proteins and the exocyst complex. Nat Cell Biol 5:520 30. Sauer, B., and N. Henderson. 1988. Site specific DNA recombination in mammalian cells by the Cre recombinase of bacteriophage P1. Proc Natl Acad Sci U S A 85:5166 70. Sausbier, M., H. Hu, C. Arntz, S. Feil, S. Kamm, H. Adelsberger, U. Sausbier, C.A. Sailer, R. Feil, F. Hofmann, M. Korth, M.J. Shipston, H.G. Knaus, D.P. Wolfer, C.M. Pedroarena, J.F. Storm, and P. Ruth. 2004. Cerebellar at axia and Purkinje cell dysfunction caused by Ca2+ activated K+ channel deficiency. Proc Natl Acad Sci U S A 101:9474 8. Savkur, R.S., A.V. Philips, and T.A. Cooper. 2001. Aberrant regulation of insulin receptor alternative splicing is associated with insu lin resistance in myotonic dystrophy. Nat Genet 29:40 7. Schubert, R., and M.T. Nelson. 2001. Protein kinases: tuners of the BKCa channel in smooth muscle. Trends Pharmacol Sci 22:505 12.

PAGE 132

132 Sen, S., I. Talukdar, Y. Liu, J. Tam, S. Reddy, and N.J. Webster. 2010. Muscleblind like 1 (Mbnl1) promotes insulin receptor exon 11 inclusion via binding to a downstream evolutionarily conserved intronic enhancer. J Biol Chem 285:25426 37. Sergeant, N., B. Sablonniere, S. Schraen Maschke, A. Ghestem, C.A. Maurage, A. W attez, P. Vermersch, and A. Delacourte. 2001. Dysregulation of human brain microtubule associated tau mRNA maturation in myotonic dystrophy type 1. Hum Mol Genet 10:2143 55. Seznec, H., O. Agbulut, N. Sergeant, C. Savouret, A. Ghestem, N. Tabti, J.C. Will er, L. Ourth, C. Duros, E. Brisson, C. Fouquet, G. Butler Browne, A. Delacourte, C. Junien, and G. Gourdon. 2001. Mice transgenic for the human myotonic dystrophy region with expanded CTG repeats display muscular and brain abnormalities. Hum Mol Genet 10: 2717 26. Seznec, H., A.S. Lia Baldini, C. Duros, C. Fouquet, C. Lacroix, H. Hofmann Radvanyi, C. Junien, and G. Gourdon. 2000. Transgenic mice carrying large human genomic sequences with expanded CTG repeat mimic closely the DM CTG repeat intergenerational and somatic instability. Hum Mol Genet 9:1185 94. Simon, P., R. Dupuis, and J. Costentin. 1994. Thigmotaxis as an index of anxiety in mice. Influence of dopaminergic transmissions. Behav Brain Res 61:59 64. Spampanato, J., A. Escayg, M.H. Meisler, and A .L. Goldin. 2001. Functional effects of two voltage gated sodium channel mutations that cause generalized epilepsy with febrile seizures plus type 2. J Neurosci 21:7481 90. Sugnet, C.W., K. Srinivasan, T.A. Clark, G. O'Brien, M.S. Cline, H. Wang, A. Willi ams, D. Kulp, J.E. Blume, D. Haussler, and M. Ares, Jr. 2006. Unusual intron conservation near tissue regulated exons found by splicing microarrays. PLoS Comput Biol 2:e4. Sun, Y., A. Savanenin, P.H. Reddy, and Y.F. Liu. 2001. Polyglutamine expanded hunti ngtin promotes sensitization of N methyl D aspartate receptors via post synaptic density 95. J Biol Chem 276:24713 8. Suzuki, T., W. Li, J.P. Zhang, Q.B. Tian, H. Sakagami, N. Usuda, H. Kondo, T. Fujii, and S. Endo. 2005. A novel scaffold protein, TANC, p ossibly a rat homolog of Drosophila rolling pebbles (rols), forms a multiprotein complex with various postsynaptic density proteins. Eur J Neurosci 21:339 50. Talbot, K., J. Stradling, J. Crosby, and D. Hilton Jones. 2003. Reduction in excess daytime slee piness by modafinil in patients with myotonic dystrophy. Neuromuscul Disord 13:357 64.

PAGE 133

133 Tanaka, K., K. Watase, T. Manabe, K. Yamada, M. Watanabe, K. Takahashi, H. Iwama, T. Nishikawa, N. Ichihara, T. Kikuchi, S. Okuyama, N. Kawashima, S. Hori, M. Takimoto, and K. Wada. 1997. Epilepsy and exacerbation of brain injury in mice lacking the glutamate transporter GLT 1. Science 276:1699 702. Teplova, M., and D.J. Patel. 2008. Structural insights into RNA recognition by the alternative splicing regulator musclebl ind like MBNL1. Nat Struct Mol Biol 15:1343 51. Terenzi, F., and A.N. Ladd. 2010. Conserved developmental alternative splicing of muscleblind like (MBNL) transcripts regulates MBNL localization and activity. RNA Biol 7:43 55. Tevaarwerk, G.J., and A.J. H udson. 1977. Carbohydrate metabolism and insulin resistance in myotonia dystrophica. J Clin Endocrinol Metab 44:491 8. Theunissen, O., F. Rudt, and T. Pieler. 1998. Structural determinants in 5S RNA and TFIIIA for 7S RNP formation. Eur J Biochem 258:758 67. Thierfelder, L., H. Watkins, C. MacRae, R. Lamas, W. McKenna, H.P. Vosberg, J.G. Seidman, and C.E. Seidman. 1994. Alpha tropomyosin and cardiac troponin T mutations cause familial hypertrophic cardiomyopathy: a disease of the sarcomere. Cell 77:701 12 Timchenko, L.T., J.W. Miller, N.A. Timchenko, D.R. DeVore, K.V. Datar, L. Lin, R. Roberts, C.T. Caskey, and M.S. Swanson. 1996. Identification of a (CUG)n triplet repeat RNA binding protein and its expression in myotonic dystrophy. Nucleic Acids Res 24: 4407 14. Timchenko, N.A., Z.J. Cai, A.L. Welm, S. Reddy, T. Ashizawa, and L.T. Timchenko. 2001. RNA CUG repeats sequester CUGBP1 and alter protein levels and activity of CUGBP1. J Biol Chem 276:7820 6. Timchenko, N.A., R. Patel, P. Iakova, Z.J. Cai, L. Qu an, and L.T. Timchenko. 2004. Overexpression of CUG triplet repeat binding protein, CUGBP1, in mice inhibits myogenesis. J Biol Chem 279:13129 39. Tonegawa, S., J.Z. Tsien, T.J. McHugh, P. Huerta, K.I. Blum, and M.A. Wilson. 1996. Hippocampal CA1 region r estricted knockout of NMDAR1 gene disrupts synaptic plasticity, place fields, and spatial learning. Cold Spring Harb Symp Quant Biol 61:225 38. Townsend, P.J., P.J. Barton, M.H. Yacoub, and H. Farza. 1995. Molecular cloning of human cardiac troponin T iso forms: expression in developing and failing heart. J Mol Cell Cardiol 27:2223 36. Traynelis, S.F., M. Hartley, and S.F. Heinemann. 1995. Control of proton sensitivity of the NMDA receptor by RNA splicing and polyamines. Science 268:873 6.

PAGE 134

134 Turnpenny, P., C. Clark, and K. Kelly. 1994. Intelligence quotient profile in myotonic dystrophy, intergenerational deficit, and correlation with CTG amplification. J Med Genet 31:300 5. Ule, J., K. Jensen, A. Mele, and R.B. Darnell. 2005. CLIP: a method for identifying protein RNA interaction sites in living cells. Methods 37:376 86. van der Meche, F.G., J.M. Bogaard, J.C. van der Sluys, R.J. Schimsheimer, C.C. Ververs, and H.F. Busch. 1994. Daytime sleep in myotonic dystrophy is not caused by sleep apnoea. J Neurol Ne urosurg Psychiatry 57:626 8. Volkert, F.C., and J.R. Broach. 1986. Site specific recombination promotes plasmid amplification in yeast. Cell 46:541 50. Wang, Y., J. Wang, L. Gao, R. Lafyatis, S. Stamm, and A. Andreadis. 2005. Tau exons 2 and 10, which ar e misregulated in neurodegenerative diseases, are partly regulated by silencers which bind a SRp30c.SRp55 complex that either recruits or antagonizes htra2beta1. J Biol Chem 280:14230 9. Wang, Z.W., O. Saifee, M.L. Nonet, and L. Salkoff. 2001. SLO 1 potas sium channels control quantal content of neurotransmitter release at the C. elegans neuromuscular junction. Neuron 32:867 81. Werner, M.E., P. Zvara, A.L. Meredith, R.W. Aldrich, and M.T. Nelson. 2005. Erectile dysfunction in mice lacking the large conduc tance calcium activated potassium (BK) channel. J Physiol 567:545 56. Winblad, S., C. Lindberg, and S. Hansen. 2006. Cognitive deficits and CTG repeat expansion size in classical myotonic dystrophy type 1 (DM1). Behav Brain Funct 2:16. Xu, J.H., L. Long, Y.C. Tang, H.T. Hu, and F.R. Tang. 2007. Ca(v)1.2, Ca(v)1.3, and Ca(v)2.1 in the mouse hippocampus during and after pilocarpine induced status epilepticus. Hippocampus 17:235 51. Yamamoto, H., K. Kokame, T. Okuda, Y. Nakajo, H. Yanamoto, and T. Miyata. N DRG4 protein deficient mice exhibit spatial learning deficits and vulnerabilities to cerebral ischemia. J Biol Chem 286:26158 65. Yeo, G.W., N.G. Coufal, T.Y. Liang, G.E. Peng, X.D. Fu, and F.H. Gage. 2009. An RNA code for the FOX2 splicing regulator reve aled by mapping RNA protein interactions in stem cells. Nat Struct Mol Biol 16:130 7. Yuan, Y., S.A. Compton, K. Sobczak, M.G. Stenberg, C.A. Thornton, J.D. Griffith, and M.S. Swanson. 2007. Muscleblind like 1 interacts with RNA hairpins in splicing targe t and pathogenic RNAs. Nucleic Acids Res 35:5474 86.

PAGE 135

135 Zhang, C., and R.B. Darnell. Mapping in vivo protein RNA interactions at single nucleotide resolution from HITS CLIP data. Nat Biotechnol 29:607 14. Zhang, H., A. Maximov, Y. Fu, F. Xu, T.S. Tang, T. T katch, D.J. Surmeier, and I. Bezprozvanny. 2005. Association of CaV1.3 L type calcium channels with Shank. J Neurosci 25:1037 49. Zhou, R.H., K. Kokame, Y. Tsukamoto, C. Yutani, H. Kato, and T. Miyata. 2001. Characterization of the human NDRG gene family: a newly identified member, NDRG4, is specifically expressed in brain and heart. Genomics 73:86 97.

PAGE 136

136 BIOGRAPHICAL SKETCH Konstantinos Charizanis was born in Athens Greece in 1979. He is the younger of two children born to Ioannis and Paraskevi Charizani s. Konstantinos attended the University of Patras from 2000 2005 where he studied and earned a Bachelor of Science degree in b iology. After graduation, Konstantinos moved to Gainesville Florida and joined the Department of E xercise P hysiology and K inesiol ogy at the University of Florida and obtained the MS degree in H uman P erformance in 2006. H e then joined the I nterdisciplinary P rogram in B iomedical S ciences at the University o f Florida College o f Medicine. Konstantinos did his doctoral thesis studies in Dr. Maurice laboratory in the Department of Molecular Genetics and Microbiology and completed his Ph.D. dissertation in December 2011. Konstantinos plans to work in the pharmaceutical industry for the next two years and then pursue an MBA degree