Citation
Using Mouse Models to Study the Mechanism of Imprinting Involved in Prader-Willi and Angelman Syndromes

Material Information

Title:
Using Mouse Models to Study the Mechanism of Imprinting Involved in Prader-Willi and Angelman Syndromes
Creator:
PERRY, EDWIN G. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Alleles ( jstor )
Chromosomes ( jstor )
DNA ( jstor )
Exons ( jstor )
Gels ( jstor )
Genes ( jstor )
Genetic mutation ( jstor )
Methylation ( jstor )
Mice ( jstor )
Plasmids ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Edwin G. Perry. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/31/2005
Resource Identifier:
71302163 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

USING MOUSE MODELS TO STUDY THE MECHANISM OF IMPRINTING INVOLVED IN PRADER-WILLI AND ANGELMAN SYNDROMES By EDWIN G. PEERY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

This document is dedicated to the memory of Cami Brannan and Betsy Peery.

PAGE 3

iii ACKNOWLEDGMENTS I would like to thank my mother, father, and brother for their love, support, and encouragement over the last 30 years. Their examples of kindness, humor, hard work, and determination will always provide a very important guiding influence. I am extremely grateful to my mentors, both Cami Brannan and Jim Resnick, for their scientific guidance, friendship, and stre ngth of character. No one could ask for better advisors. In particular , I thank Cami Brannan for demonstrating an effective and successful approach to life, and for being so patient and understandi ng. I also thank Jim Resnick for having the courage to battle thr ough some truly tough times, for taking on the burden of looking out for the students in both labs, and for doing it all with incredible grace and humor. I have especially appreci ated the chance to enjoy JimÂ’s delightfully cynical and sarcastic wit over the last 7 years. Many thanks are also due to friends and cowo rkers for their help in the lab and also for their good cheer and comradery. Present and past members of the Brannan and Resnick labs include Todd Adamson, Jo anne Anderson, Susan Blaydes, Stormy Chamberlain, Amanda Dubose, Michael Elmore , Steve Filippelli, Chris Futtner, Karen Johnstone, Lori Kellam, Steve Kushert, Da nielle Maatouk, Edith Orozco, Missy Shelley, Tom Simon, Jessica Walrath, Lindsey Williams, and Tao Yang. I am especially grateful to Karen Johnstone, for all of her advice and help with constructs and karyotyping; to Danielle Maatouk, for her assistance with bisulfite sequencing and dissertation

PAGE 4

iv formatting; to Michael Elmore, for his mous e making skills; and to Stormy Chamberlain, for innumerable laughs at her expense. I thank Christine Mione Kiefer, for gene rously providing unpublished primers and protocols for bisulfite sequencing; Ahmad Khalil and Sung-Hae Lee Kang, for their assistance with ES cell karyotyping; a nd Laurence Morel and Ward Wakeland, for providing the castaneous chromosome 7 mice. Finally, I thank Henry Baker, Jorg Bunge rt, Dan Driscoll, Gerry Shaw, and Tom Yang for taking time out of their busy sc hedules to provide advice at committee meetings. Thanks also go to Marisa Bartol omei for helpful advice and for the role that she has played in maintaining funding for Brannan/Resnick lab projects.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF FIGURES..........................................................................................................vii ABSTRACT....................................................................................................................... ix CHAPTER 1 INTRODUCTION........................................................................................................1 Genomic Imprinting......................................................................................................1 Definition of Genomic Imprinting.........................................................................1 Discovery of Genomic Imprinting.........................................................................1 Evolution of Genomic Imprinting.........................................................................3 Extent of Genomic Imprinting...............................................................................6 Characteristics of Imprinted Genes.......................................................................8 Prader-Willi and A ngelman Syndromes.....................................................................12 Mouse Model for PWS-IC Deletions.........................................................................15 Specific Aim: Identifying the Murine AS-IC.............................................................18 Summary.....................................................................................................................19 2 MATERIALS AND METHODS...............................................................................29 Cloning.......................................................................................................................2 9 Restriction Digest................................................................................................29 Phenol-Chloroform Extraction............................................................................30 Ethanol Precipitation...........................................................................................30 Removal of Phosphate Groups from Vector.......................................................31 Gel Purification...................................................................................................32 QIAquick gel-extraction kit.........................................................................32 Wizard-Prep gel-extraction kit.....................................................................33 Electro-elution..............................................................................................35 DNA Ligation......................................................................................................36 Transformation....................................................................................................37 Culturing Embryonic Stem Cells................................................................................37 Southern Blot..............................................................................................................39

PAGE 6

vi Northern Blot..............................................................................................................40 Sodium Bisulfite Analysis..........................................................................................40 3 DELETION STRATEGY...........................................................................................42 Strategy for Deletions.................................................................................................42 Plasmid Creation.........................................................................................................47 Building the 12 kb Deletion Plasmid...................................................................47 Building the 7 kb Deletion Plasmid.....................................................................49 ES Cell Targeting.......................................................................................................52 Targeting the 12 kb Deletion...............................................................................52 Targeting the 7 kb Deletion.................................................................................52 Establishment of Breeding Colonies..........................................................................53 Establishing the 12 kb Deletion Line..................................................................53 Establishing the 7 kb deletion line.......................................................................54 4 RESULTS...................................................................................................................57 Introduction.................................................................................................................57 Overview of Analysis.................................................................................................57 Overview of DNA Methylation and Gene Expression Studies...........................58 Cursory Examination of Phenotype.....................................................................59 Genetic Background of Mice...............................................................................59 DNA Methylation Studies..........................................................................................60 Southern blot analysis of DNA methylation.......................................................60 Sodium bisulfite analysis of DNA methylation without polymorphisms...........62 Sodium bisulfite analysis of DN A methylation with polymorphisms.................67 Gene Expression Studies............................................................................................71 Summary.....................................................................................................................75 5 DISCUSSION...........................................................................................................102 Introduction...............................................................................................................102 Analysis of 12 kb and 7 kb Deletions.......................................................................102 Methylation Analysis.........................................................................................103 Gene Expression Analysis.................................................................................106 Future Analysis of the 12 kb and 7 kb Deletions......................................................108 Methylation Analysis.........................................................................................108 Gene Expression Analysis.................................................................................108 Phenotype Analysis...........................................................................................109 Relevant Findings from Other Laboratories.............................................................110 Future Directions......................................................................................................113 LIST OF REFERENCES.................................................................................................117 BIOGRAPHICAL SKETCH...........................................................................................131

PAGE 7

vii LIST OF FIGURES Figure page 1-1 Life cycle of the imprint........................................................................................20 1-2 Molecular classes of PWS and AS.........................................................................21 1-3 Imprinted genes on human chromosome 15 and mouse chromosome 7...............22 1-4 Effects of imprinting mutations.............................................................................23 1-5 Location and bipartite nature of the imprinting center..........................................24 1-6 Inheritance patterns of IC deletions.......................................................................25 1-7 The 35 kb PWS-IC deletion mouse model............................................................26 1-8 The “Paternal-Only” model to explain the function of the IC...............................27 1-9 Deletions upstream of Snrpn to locate the murine AS-IC.....................................28 3-1 Deletions upstream of Snrpn to locate the murine AS-IC.....................................55 3-2 Southern blot genotyping of the 12 kb and 7 kb deletion mice.............................56 4-1 Deletions upstream of Snrpn to locate the murine AS-IC.....................................76 4-2 Effects of imprinting mutations.............................................................................77 4-3 Southern blot examinat ion of DNA methylation...................................................78 4-4 Bisulfite analysis of the Snrpn promoter in wild-type mice..................................79 4-5 Bisulfite analysis of the Snrpn promoter in 12 kb deletion mice...........................80 4-6 Bisulfite analysis of the Snrpn promoter in 7 kb deletion mice............................ 81 4-7 Bisulfite analysis of Snrpn intron 8 in wild-type mice..........................................82 4-8 Bisulfite analysis of Snrpn intron 8 in 12 kb deletion mice...................................83 4-9 Bisulfite analysis of Snrpn intron 8 in 7 kb deletion mice.....................................84

PAGE 8

viii 4-10 B6.Cast.c7 mice provide polymorphi sms for allele-specific analysis...................85 4-11 Bisulfite analysis of the Snrpn promoter in wild-type mice..................................86 4-12 Bisulfite analysis of the Snrpn promoter in 12 kb deletion mice...........................87 4-13 Bisulfite analysis of the Snrpn promoter in 7 kb deletion mice.............................88 4-14 Summary of bisulfite analysis of the Snrpn promoter...........................................89 4-15 Bisulfite analysis of Snrpn intron 1 in wild-type mice..........................................90 4-16 Bisulfite analysis of Snrpn intron 1 in 12 kb deletion mice...................................91 4-17 Bisulfite analysis of Snrpn intron 1 in 7 kb deletion mice.....................................92 4-18 Summary of bisulfite analysis of Snrpn intron 1...................................................93 4-19 Bisulfite analysis of the Mkrn3 promoter in wild-type mice.................................94 4-20 Bisulfite analysis of the Mkrn3 promoter in 12 kb deletion mice..........................95 4-21 Bisulfite analysis of the Mkrn3 promoter in 7 kb deletion mice............................96 4-22 Summary of bisulfite analysis of the Mkrn3 promoter..........................................97 4-23 Crossing with the PWS-IC deletion mouse to measure gene expression..............98 4-24 Northern blot analysis of Ndn and Srnpn expression.............................................99 4-25 Northern blot analysis of snoRNA expression.....................................................100 4-26 Allele-specific expression of Ube3a measured by RT-PCR analysis..................101 5-1 Insertion 11 kb upstream of Snrpn creates an imprinting mutation.....................116

PAGE 9

ix Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy USING MOUSE MODELS TO STUDY THE MECHANISM OF IMPRINTING INVOLVED IN PRADER-WILLI AND ANGELMAN SYNDROMES By Edwin G. Peery December 2004 Chair: James Resnick Major Department: Molecular Genetics and Microbiology Although most genes in mammals are expr essed from both the maternally and paternally inherited alleles, a small number of genes are selectively expressed from only one parental allele. These genes are subject to genomic imprinting, meaning that each allele is modified during gametogenesis so that it shows parental-origin-specific expression in the offspring. The most extensively investigated examples of imprinting disorders in humans are Prader-Willi syndrome (PWS) and Angelman Syndrome (AS). Characteristics of PWS include neonatal failure to thrive, hypotonia, and feeding difficulties with subsequent hyperphagia and extreme obesity, hypogonadism, small hands and feet, short stature, mild to moderate mental retardation, a nd obsessive-compulsive like behaviors. Characteristics of AS include severe mental retardation, lack of speech, ataxia, seizures, hyperactivity, and a happy disposition.

PAGE 10

x Both disorders result from the disturba nce of imprinted gene expression from human chromosome 15q11-q13. AS is due to a lack of maternal gene expression, and PWS is due to a lack of paternal gene expres sion from this region. Some cases of AS and PWS are caused by microdeletions near the SNRPN gene that disrupt a regulatory element termed the imprinting center (IC). Th is regulatory element appears to have two parts, with one part at the promoter of SNRPN involved in PWS (PWS-IC) and another part 35 kilobases (kb) upstream of SNRPN involved in AS (AS-IC). To further understand the func tion of the IC, we sought to create a mouse model for AS-IC mutations. Given that the human AS-IC is locate d roughly 35 kb upstream of SNRPN , we focused our attention on the corres ponding region of the mouse genome. Using homologous recombination in embryonic stem cells, we generated three nested deletions upstream of Snrpn . From our analysis of these deletions, we concluded that the murine AS-IC resides outside of the 12 kb region investigated in our study. This implies that the location of the AS-IC relative to Snrpn is not conserved between human and mouse. This finding is significant, because it raises the possibility that the mechanism of imprinting for this region has di verged between human and mouse.

PAGE 11

1 CHAPTER 1 INTRODUCTION Genomic Imprinting Definition of Genomic Imprinting As a consequence of sexual reproduction, ma mmals inherit two copies (alleles) of each gene: one copy from the mother, and one copy from the father. Although most genes in mammals are expressed from both the maternally and paternally inherited alleles, a small number of gene s are selectively expressed from only one parental allele. These genes are subject to genomic imprinti ng, meaning that each allele is modified during gametogenesis (i.e., imprinted) so that it shows parental-origin-specific expression in the offspring [1]. Genomic imprinting repr esents a striking exception to the principles of Mendelian inheritance, given that imprinted genes clearly ha ve a history – they retain some information about the germline through which they have passed, be it male or female, and this information alters their f unction in the next ge neration. Deciphering exactly how the two parental a lleles can reside in the same nucleus and yet behave very differently is one of the most intriguing mysteries in biology. Discovery of Genomic Imprinting The phenomenon of genomic imprinting was di scovered in the early 1980s with the observation that normal mammalian developmen t requires a geneti c contribution from both the mother and the father. Nuclear tr ansplantation experiments were employed to generate uniparental mouse embryos that, alt hough diploid, were derived from either two male pronuclei (androgenetic) or two fema le pronuclei (gynogenetic). As the

PAGE 12

2 androgenetic and gynogenetic embryos invariably died during the early postimplantation period, these experiments demonstrated that the maternal and paternal genomes are not equivalent and that both genomes are re quired to complete embryogenesis [2, 3]. Interestingly, the androgenetic embryos were composed almost entirely of extraembryonic tissue, whereas the g ynogenetic embryos displayed a lack of extraembryonic development. This sugge sted that the patern al contribution was necessary for the extraembryonic developmen tal program to proceed but that the maternal contribution was required for development of the actual embryo. In a similar situation to the mouse pronuc lear transplant experiments described above, evidence that normal human development requires genetic information from both parents was derived from studies of hydatid iform moles and ovarian teratomas [4]. Hydatidiform moles occur when an egg lacki ng a nucleus is either fertilized by two sperm or is fertilized by a single sperm w ith a subsequent doubling of chromosome number. The result is a di ploid androgenetic zygote [5]. As was seen for the androgenetic mouse embryos, hydatidiform mole s appear to be diso rganized placental tissue lacking embryonic development [6]. In contrast, ovarian teratomas are diploid female germ-cell tumors that contain e ndodermal, ectodermal, and mesodermal tissues but no extraembryonic tissue [7]. The observation that uniparental embryos fa il to develop normally gave rise to the idea that there must be certain genes that are expressed exclusively from either the maternal or the paternal allele. The first step in identifying the location of imprinted genes was made by using mice with chromo somal rearrangements to determine the effects of inheriting both copies of a specific chromosome from one parent, referred to as

PAGE 13

3 uniparental disomy (UPD) [8, 9]. Mice hete rozygous for Robertsonian translocations were bred to produce uniparental disomy for an entire chromosome, and mice heterozygous for reciprocal translocations were bred to produce partial uniparental disomy for certain portio ns of chromosomes. (A Robertsonian translocation occurs when the long arms of two non-homologous chromosome s are joined at a single centromere. A reciprocal translocation is produced when two non-homologous chromosomes exchange segments.) In this manner, it was possible to construct a rough map of the chromosomes, or regions of chromosomes, that contain impr inted genes. Eleven imprinted regions of the mouse genome were identified [10]. Si milarly, cases of uniparental disomy in humans have revealed numer ous chromosomes that contain imprinted genes [11]. From this type of analysis, not all chromosomes appear to contain imprinted genes [4]. Evolution of Genomic Imprinting Two obvious questions surrounding the field of genomic imprinting are how such a bizarre system of gene regulation first e volved, and why some ge nes are imprinted and others are not. The dogma of genetics has been that an advantage of diploidy is the masking of deleterious recessive alleles [12]. Given this vi ew, it is puzzling why certain genes should be maintained in a functionally haploid state. One theory, termed the “ovarian time bomb” hypothesis, proposed that genomic imprinting serves the function of reducing the danger of ovarian teratomas by s ilencing genes necessary for the formation of trophoblastic tissue that coul d allow the tumor to become more invasive [13]. This theory received little support, given that very few imprinted genes identified so far seem to play a role in trophoblast development and th at this theory was una ble to explain genes that are silenced on the paternally inherited allele. Others suggested that imprinting was necessary to prevent chromosome lo ss or parthenogenesis [14, 15].

PAGE 14

4 The only theory to gain much widespread s upport to date, however, is the parental conflict model, also referred to as the “tugof-war” hypothesis or the kinship theory [16]. This theory maintains that im printing evolved in response to conflicting forces of natural selection that act differently on alleles tran smitted by the mother as compared to alleles transmitted by the father. The conflict model presumes that imprinting arose in a polygamous species in which the female is primarily responsible for supporting the growth of the young. The idea here is patern al genes that modulate embryonic and neonatal growth or nutrient transfer shoul d act to maximize the amount of maternal resources to the current offspring at the expense of potential future progeny. This strategy is meant to increase the size and strength of t hose offspring and improve the chances that they will survive to reprodu ce and pass on the genes from the father. Therefore, the paternal genome is under select ive pressure to ensu re the expression of genes that promote growth and to silence ge nes that inhibit growth. If the lifetime reproductive fitness of the female is optim ized by spreading maternal resources among multiple pregnancies, however, this shoul d force the maternal genome to express growth-inhibiting genes and to silence gr owth-promoting genes. Indeed, initial examination of several imprinted genes seemed to support this hypothesis [17]. Futhermore, imprinting was disrupted in F1 hybrids between Peromyscus maniculatus, a “polygamous” mouse species; and Peromy scus polionotus, a “monogamous” mouse species [18]. The simplicity of the “tug-of-war” hypot hesis and its early success in explaining observations of how several imprinted genes could influence growth quickly thrust the theory into prominence. Not everyone, how ever, was convinced. The results of the

PAGE 15

5 Peromyscus experiment, in addition to the an alysis of uniparental disomy mice, have been interpreted by some as violating the pred ictions of the conflict theory [19-21]. For now, it is difficult to discern just how well the kinship theory will stand up to the accumulation of information regarding the role of imprinted genes in physiology relevant to growth, placental function in nutrient tran sfer, and neonatal feeding behavior. It should be noted that the descri ption of the kinship theory pu t forth here is based on the original, much simpler version [16]. Over th e last decade, Haig has continued to refine and elaborate the hypothesis [22]. In fact, he now boasts that one formulation of the kinship theory “is not vu lnerable to empirical falsification” [23] (page 367) . This would seem to limit its usefulness in guiding scientif ic investigation. So perhaps the jury will forever be out on the kinship theory and the evolution of imprinting. Several others have recently commente d on the origins of genomic imprinting. Drawing on the parallels between X chromoso me inactivation and genomic imprinting, Ohlsson and colleagues stated that genes that are expressed at a low level are frequently and randomly inactivated, and that this propert y could have provided the potential for the emergence of imprinted gene expression and X chromosome inactivation [24]. Taking a different stance entirely, Carmen Sapien za’s group advanced the hypothesis that monoallelic expression was not the subject of selective pressure that led to the evolution of genomic imprinting, but rather that impr inted gene expression is the byproduct of chromatin structure modifications whose prim ary functions are to enhance chromosome pairing and segregation during meiosis [25]. In a third approach, Beaudet and Jiang proposed the “rheostat” model, which suggest ed that the primary useful function of genomic imprinting is to allow a state of tem porary and reversible functional haploidy for

PAGE 16

6 a given allele that can (for a short time) shelte r the silenced allele from the constraints of selective forces, and thereby provide an advantage to a populati on by allowing greater potential to adapt to a rapidl y changing environment [26]. Although it is interesting to contemplate the evolution of genomic imprinting, there is certainly no consensus or ev en substantial experimental ev idence in favor of any of the theories described above. Given that di scussion of evolutionary hypotheses beyond this brief introduction would proba bly not contribute directly to the understanding of the molecular mechanisms of imprinting (our prim ary focus), we now turn our attention to examining the scope of genomic imprinting a nd the common characteristics of imprinted genes. Extent of Genomic Imprinting After the discovery of the phenomenon of genomic imprinting in the 1980s, several years passed before the first imprinted gene was identified. Some initial clues regarding the imprinting process were garnered from studi es of an imprinted tr ansgene [27]. Major progress in understanding the mechanis m of imprinting, however, depended on identifying endogenous imprinted loci. Finally, in 1991, DeChiara and colleagues fortuitously discovered that a targeted mutati on of the insulin-like growth factor 2 gene ( Igf2 ) produced a growth defect in the he terozygous state when the mutation was inherited from the father, but not when the mu tation was inherited from the mother [28]. Subsequent identification of additional imprin ted genes, such as the insulin-like growth factor 2 receptor gene ( Igf2r ) and the small nuclear ribonucleoprotein N gene ( Snrpn ), largely relied upon a positional-candidate approa ch that involved testing for monoallelic expression of genes that mapped to known im printed chromosomal regions in mouse and man [29-34].

PAGE 17

7 In addition to the positional candidate a pproach, numerous screening methods were employed to identify genes that showed monoallelic expression or differential methylation [35, 36]. (Differential methyl ation, which is a property shared by many imprinted genes, will be explained below.) As of 2004, over 60 imprinted genes have been identified. The Harwell imprinting webpa ge gives an up-to-da te list of imprinted genes [37]. (source: http://www.mgu.har.mrc.ac.uk/research/imprinting/ , last accessed December 3, 2004) Based on the number of imprinted regions identified in uniparental disomy mice, the number of imprinted domains recognized in humans, and the result from screens for imprinted genes, several authors have estim ated that there might be 100-200 imprinted genes in total [38-41]. Othe r estimates, however, range anywhere from O.1% to 1.0% of genes in the mammalian genome [42-44]. These estimates might need to be revised in light of information coming from the human genome project, which has been interpreted to suggest that there are roughly 30,000 human genes [45]. If indeed there are only 30,000 mammalian genes, then it seems that imprinted genes must make up more than 0.1% of these, as the curr ent number of imprinted gene s (about 60) would represent 0.2% of the genome. It is clearly difficult to accurately predic t the total number of imprinted genes, as indicated by the broad range of the estimates listed above. N onetheless, it is apparent that although there are certainly more imprin ted genes waiting to be identified, genomic imprinting affects only a very small proporti on of mammalian genes overall. Despite the fact that genomic imprinting affects perhaps only a few dozen genes, it is an attractive field of research for several reasons. Not only is imprin ting an extremely interesting

PAGE 18

8 phenomenon, worthy of investigation in its own right, but information gained from studies of imprinting should provide more ge neral insights into the mechanisms that regulate gene expression. Furthermore, the ro le of imprinting in human disease provides a compelling purpose for research in this area. These points will be elaborated next. Characteristics of Imprinted Genes One approach to determining the factors involved in the mechanism of imprinting has been to examine the features that ar e shared among imprinted genes. A striking theme is that about 80% of imprinted genes are located together with other imprinted genes within imprinted chromosome domains [ 46]. In some cases, these domains can be as large as one to two megabase s (Mb) [47]. Some clusters of imprinted genes appear to be under the influence of a cis-acting regulat ory element, termed an imprinting control element (ICE), imprinting control region (ICR) , or imprinting center (IC), that can act over long distances to coordina tely regulate the imprinted ex pression of genes throughout the domain. Although in most cases the precis e mechanism of imprinting appears to be different for each cluster of imprinted gene s, it is possible to identify some common aspects shared by most imprinted domains. Some of these features include silencing elements, chromatin boundary or insulator elements, and non-coding or antisense RNA transcripts [47]. Another important property that is comm on among imprinted genes (and that is indisputably relevant to the imprinting mechan ism) is DNA methylation. Almost 90% of imprinted genes harbor CpG islands, which ar e regions that contain a high density of CpG dinucleotides [46]. The cytosine of a CpG dinucleotides can be modified by the addition of a methyl group, and the presen ce of this modification at CpG islands overlapping the promoters of genes has been associated with transcriptional silencing

PAGE 19

9 [48]. It is important to recognize, however, that CpG islands are not always located at the promoters of genes, but can be found within the body of a gene or outside of a gene. Furthermore, methylation of a gene is not alwa ys associated with si lencing of that gene [49]. A key aspect of DNA methylation found in im printed genes is that in most cases the methylation is confined to one allele. In fact, this characte ristic of differential methylation is so widespread among imprinted genes that it has been used in screens to identify novel imprinted genes [40, 50]. Regi ons that show this property of allelespecific methylation have been termed diffe rentially methylated domains (DMDs) or differentially methylated regi ons (DMRs). For most imprin ted genes it is the silenced allele that is methylated, a lthough this is not always the cas e [51]. The DMRs are often associated with DNA sequences that are impor tant for imprinting, and the identification of regions of allele-specific methylation ha s often revealed the presence of imprinting control elements [52]. The importance of DNA methylation in imprinting was clearl y confirmed by the results of an experiment involving a dele tion of the DNA methyltransferase 1 gene (Dnmt1). Mice homozygous for the Dnmt1 gene died by the 9th day of embryonic development, and total levels of DNA methyl ation were vastly reduced [53]. Further analysis of these embryos revealed that the no rmal patterns of gene expression for several imprinted genes were disrupted, demonstrating that DNA methylation pl ays a critical role in the imprinting process [54]. Methylation of DNA provides a good candi date for the imprinting mark, as it seems to fulfill the criteria expected of such a mark. Any modification that could

PAGE 20

10 function as the imprint that distinguishes th e parental alleles s hould be erased during gametogenesis, applied to the gametes of onl y one parent, and then faithfully maintained throughout development [46] (Figure 1-1). Although differential methylation for many imprinted genes is inherited from the game tes and maintained throughout development, other imprinted genes acquire th eir allele-specific methylatio n after fertilization [1, 52]. At least for this latter group of genes, as well as the minority of imprinted genes that do not exhibit allele-speci fic methylation at all, somethi ng other than DNA methylation must serve as the primary imprinting mark. So what other modifications could serve as the imprinting mark? Given that the alleles of imprinted genes can be distinguished in the nucleus even when those alleles are identical in sequence (as shown by the exis tence of imprinting in inbred mice), the imprint must be epigenetic in nature. Ep igenetics has been defined as “the study of heritable changes in gene expression that occur without a change in DNA sequence” [55] (page 481). In addition to DNA methylation, other epigenetic modifications include allele-specific histone modifications, such as acetylation or methylation, or other differences in chromatin structure. These have, in fact, been s hown for a number of imprinted genes [56-62]. It also is con ceivable that mammalian DNA binding proteins could be involved in a system of transcriptional memory that locks in a state of silencing or activation at a locus based on its expressi on pattern early in development. Such a system could function in a similar manner to the Polycomb and trithorax method of regulating homeotic gene expres sion patterns in the fruit fly Drosophila melanogaster [63]. Although there is no evid ence that such a mechanism ex ists in precisely the same form in mammals, the mouse Polycomb ho molog embryonic ectoderm development gene

PAGE 21

11 ( Eed ) interacts with histone deacetylase and histone methyltransferase enzymes and has been shown to be involved in regula ting a subset of imprinted genes [64]. Although good evidence links the clustering of imprinted genes, allele-specific methylation, and perhaps other chromatin m odifications directly to the mechanism of imprinting, some other features common among imprinted genes may or may not play a role in the imprinting process. For instan ce, it has been noted that imprinted genes contain fewer and smaller introns than most genes [65]. Another curious and unusual property is that several imprint ed genes contain direct rep eats associated with the CpG islands of those genes, and it has been proposed that these repeats might play a role in establishing DNA methylati on patterns [66]. Some characteristics that seem common among imprinted genes may be related to the organization of imprinted domains. The first of these is asynchronous replication timing, where for most imprinted domains the region on the pate rnal chromosome replicates earlier than the region on the maternal chromoso me during the S phase of the cell cycle as measured by fl uorescent in situ hybridization [67-70]. However, one study that used a different technique to assess replication asynchrony (the bromodeoxyuridine incorporation assay) did not de tect allele-specific replicat ion timing at several of the regions previously identified as showing as ynchronous replication [71]. Furthermore, some have cautioned that the results of both approaches might not be an actual measure of replication asynchrony but ra ther may simply reflect differe nces in chromatin structure of the alleles [71, 72]. In addition to differences in replicat ion timing, some imprinted domains also display differences in the rate of recombin ation during meiosis, with the recombination

PAGE 22

12 rate being greater during male meiosis than in female meiosis [ 73, 74]. As with the studies of replication asynchrony, however, the functional significance of this observation is not clear. Although differences in repli cation timing or meiotic recombination could play some role in the imprinting process, it seems just as likely that these phenomena are not directly involved in the imprinting mechanism, but are instead secondary manifestations of some more fundamental di fferential modification of imprinted regions. In another intriguing observation, one study of the Prader-Willi/Angelman imprinted region on human chromosome 15 revealed that the two homologs of chromosome 15 appeared to associate late in the S phase of the cel l cycle [75]. This raised the possibility that such an associ ation of the oppositely imprinted domains might be involved in maintaining the pattern of im printing throughout the region. Subsequent examination of a mouse model that carried a deletion encompassing the entire imprinted region revealed that the impri nt was established and mainta ined properly on the intact chromosome in mice heterozygous for the de letion [76]. This provided strong evidence that the association of opposite ly imprinted domains is not a critical component of the imprinting process. Prader-Willi and Angelman Syndromes The most extensively investigated exampl es of imprinting disorders in humans are Prader-Willi syndrome (PWS) and Angelman syndrome (AS). The neonatal phase of PWS involves hypotonia , feeding difficulties, and a poorly understood metabolic defect that combine to cause a “failure to thrive.” This phase of the diseas e can last up to a year, but as the PWS patient grows older the symptoms change to include hyperphagia and extreme obesity, hypogonadism, small hands and feet, short stature, moderate mental retardation, and obsessive-compulsive like behavi or [77]. Characteristics of AS include

PAGE 23

13 severe mental retardation, lack of speech, ataxia, seizures, hyperactivity, and a happy disposition [78]. Each disease occurs w ith a frequency of about 1/15,000 births [79]. Although these diseases are clinically very different, they share an underlying genetic basis. PWS is caused by a lack of paternal gene expression from chromosome 15q11-q13, whereas AS is due to a lack of ma ternal gene expression from the same region (Figure 1-2). This loss of imprinted ge ne expression can occur in several different ways. The most common cause of PWS and AS is a de novo dele tion that physically removes all of the imprinted genes in the re gion. Roughly 70% of PWS cases are due to a 4 megabase deletion on the paternally inhe rited chromosome [8083]. Similarly, about 70% of AS cases are due to the same 4 mega base deletion; however for AS, the deletion is on the maternally inheri ted chromosome [80, 84-86]. Both PWS and AS can also be caused by unipa rental disomy (UPD). Inheritance of two copies of chromosome 15 from the moth er results in PWS [82, 83, 87]. Conversely, inheritance of both copies of chromosome 15 from the father results in AS [88, 89]. In addition to a large deletion or UPD, AS can also result from mutations in the UBE3A gene [90-96]. Interestingly, no PWS pati ents have been iden tified as having an intragenic point mutation or truncation, s uggesting that PWS is a contiguous gene syndrome that is caused by the loss of two or more paternally expressed genes [97]. Numerous imprinted genes have been iden tified within a 2 megabase imprinted domain (Figure 1-3). Genes that are expr essed exclusively from the paternal allele include MKRN3 (formerly ZNF127 ), MAGEL2 , NDN , SNRPN , IPW , multiple snoRNA genes, and an antisense transcript to the UBE3A gene [32, 33, 98-107]. The genes listed here that are transcribed from the paternal alle le will be referred to as “paternal-specific”

PAGE 24

14 or “PWS candidate” genes. Two other genes, UBE3A and ATP10C , are expressed biallelically in most tissues but are expressed exclusively from the maternal allele in the brain [108-111]. In addition to the more common molecula r classes of PWS and AS, rare examples are due to a defect in the imprinting proce ss (Figure 1-4). Thes e patients inherit a chromosome 15 from each parent, but one chromosome displays an incorrect DNA methylation and gene expression pr ofile, so that it appears as if it had been inherited from the opposite parent. In cases of PWS imprin ting mutations, some genes that are normally unmethylated and expressed on the paternally inherited chromosome become methylated and silenced. In cases of AS imprinting muta tions, at least some genes that are normally methylated and silenced on the maternally inherited chromosome show a loss of DNA methylation and become in appropriately activated. For roughly half of PWS and AS imprin ting mutation cases, Southern blotting, DNA sequencing, and other methods of muta tional analysis have not revealed a molecular lesion within the regions examin ed on chromosome 15 [112-114]. These cases appear to be due to a stochastic failure of the imprinting machinery [115]. The other half of PWS and AS imprinting mutations, many of which are familial, are caused by microdeletions that disrupt a regulatory element called the imprinting center (IC), that acts in cis to coordinately control the imprinted expression of the surrounding genes [116] (Figure 1-5). By mapping the deleti on breakpoints and comparing the extents of the deletions in several PWS patients, it ha s been shown that the smallest region of deletion overlap (SRO) involved in PWS is a 4.3 kb region spanning the promoter and exon1 of the SNRPN gene [112, 117-120]. In AS patients, the SRO is 880 basepairs, and

PAGE 25

15 is located 35 kb upstream of SNRPN [112, 119, 121]. It therefor e appears that the IC has a bipartite structure, with once critical element involved in PWS (PWS-IC) and a distinct, more centromeric component involved in AS (AS-IC). Clues to the function of the IC have b een revealed by the striking inheritance pattern of IC mutations [119] (Figure 1-6). Deletions of the PWS-IC appear to originate in the female germline and are silent in th e next generation. Dis ease results only upon subsequent transmission of the PWS-IC dele tion through a male germline, which has the effect of silencing numerous genes throughout the 2 megabase imprinted domain that are normally expressed exclusively from the pa ternal chromosome. In contrast, AS-IC mutations are benign when passed through a male germline, and result in AS only upon maternal transmission. These observations led to the suggestion that th e IC is involved in “resetting” the imprint during gametogenesis [116]. Mouse Model for PWS-IC Deletions An equivalent imprinted domain in mouse is located in the central portion of mouse chromosome 7 in a segment that is syntenic to human chromosome 15q11-q13 (Figure 1-5). All of the genes described above are conserved and imprinted in mouse [30, 31, 100, 101, 103, 104, 122-125]. It should be not ed, however, that there are several differences between the human and mouse with regard to the genes located within the imprinted domain. One distinction is that the brain-specific patte rn of imprinting for Ube3a is even more restricted in mouse as compared to human, as the murine Ube3a gene shows maternal-specific expression onl y in Purkinje cells, hippocampal neurons, and mitral cells of the olfactory bulb [124].

PAGE 26

16 Another difference between human and mouse is the imprinted status of Atp10A . There are conflicting reports regarding the imprinting of Atp10A in mouse, and it appears that this gene may be imprinted in some strains of mice but not in others [125-127]. A third difference is the presence of an additional paternally-expressed gene, Frat3 , that is located near Mkrn3 in mouse but is absent in hum ans [128, 129]. Finally, there is a gene that shows testis-specific expression, C15orf2 , that resides between NDN and SNRPN in human but that is not pr esent in mouse [130]. (The C15orf2 gene is not shown in Figure 1-3.) Despite these relatively minor differences, the large amount of conservation overall between the Prader-Willi/Angelman region in human and the syntenic region on mouse chromosome 7 provide s the opportunity to use the mouse as an experimental tool to investigat e the imprinted control of this cluster of genes. In fact, mouse models for the large deletion and UPD classes of patients faithfully recapitulate many of the molecular and phenotypic aspects of PWS and AS [31, 76, 131]. To investigate the function of the PWS-IC , Dr. Tao Yang, a former postdoc in our lab, created a mouse model for PWS-IC deletio ns [132]. A 35 kb deletion involving most of the Snrpn gene and extending 16 kb upstream was engineered in embryonic stem (ES) cells (Figure 1-7). Chimeric males derived from these cells were bred with wild-type females, and it was found that offspring from these matings that inherited the PWS-IC deletion did not express the PWS candidate genes Mkrn3 , Ndn , or Ipw . All of the mutant mice displayed a failure to thrive and died w ithin several days afte r birth. These results confirmed that the position and function of the PWS-IC is conserved between human and mouse and established the mouse as a valid model system to study the mechanism of imprinting involved in PWS.

PAGE 27

17 To address the possibility that the PWS-IC is required to main tain paternal gene expression in somatic tissues, chimeric mice ge nerated from ES cells that had sustained a PWS-IC deletion on the paternal allele were examined for signs of an imprinting defect. Southern blot DNA methylation analysis of the Mkrn3 and Ndn loci in the chimeric mice revealed hypermethylation of the paternal ch romosome. Parallel to this investigation, similar results were found for a human mosaic for a paternal PWS-IC deletion [133]. A subsequent study showed that a PWS-IC dele tion does not cause abnormal methylation of the paternal-specific genes in sperm [134]. Th ese results point to a role for the PWS-IC in maintaining the paternal imprint in so matic cells. Because the PWS-IC mutations described above occurred very early in de velopment, however, it is not clear if the PWS-IC is only required duri ng early development or if it is continuously required to stabilize the imprint during m itotic cell division throughout all stages of development and adulthood. Based on the results of these experi ments, Dr. Brannan suggested the “Paternal-Only” model to explain the function of the IC [135] (Figure 1-8). This model proposes that the PWS-IC functions as a positive element to promote paternal gene expression in somatic tissues. Paternal tran smission of a PWS-IC deletion leads to the loss of paternal gene expression in the offs pring and results in PWS. The AS-IC is described as a negative element operating in th e female germline to suppress the paternal expression pattern by inhibiting the acti on of the PWS-IC. Regulation of the tissue-specific imprinting of UBE3A is predicted to be an indirect result of a brain-specific pate rnally expressed UBE3A -antisense transcript whose expression is regulated by the PWS-IC. Maternal transmission of an AS-IC deletion leads to

PAGE 28

18 inappropriate activation of th e normally paternal-specific genes on the maternal allele, including the UBE3A -antisense transcript. This results in AS due to a loss of UBE3A expression in critical regions of the brain. Th is model conveniently e xplains the fact that many PWS-IC deletions extend upstream of SNRPN far enough to include the AS-SRO, however all known examples of AS-IC deletions stop short of the PWS-SRO. If the AS-IC serves merely to negatively regulate the PWS-IC as predicted by the model, then a mutation affecting both elements would simply prevent the expression of the paternal-specific genes and lead to PW S upon paternal transmission. Any deletion involving the PWS-IC would prevent the inappr opriate activation of the paternal-specific genes and therefore would NOT lead to an AS phenotype upon maternal transmission. Specific Aim: Identifying the Murine AS-IC Although the location of the AS-IC in huma n has been revealed by studies of IC deletion patients, the location of this elemen t in mouse is not known. Our goal for this project was to identify and characterize the murine AS-IC. At the outset of this work, before the completion of the mouse genome se quencing effort, the sequence of the region upstream of the Snrpn gene was not available. The onl y information that could be used as a guide to finding the mu rine AS-IC was the position of the AS-SRO in human. Considering that the human AS-S RO is located 35 kb upstream of SNRPN , we focused our analysis on the corresponding region upstream of the mouse Snrpn gene. It was decided that the mo st direct and feasible me thod of locating the murine AS-IC would be to create deletions upstream of Snrpn . If mice inheri ting a deletion on the maternal chromosome showed signs of an Angelman imprinting mutation, this would provide evidence that the AS-IC was located within that deletion. Using homologous recombination in embryonic stem cells, we ge nerated three nested deletions upstream of

PAGE 29

19 Snrpn (Figure 1-9). Deletion A was 7 kb, extend ing from -30 kb to -37 kb relative to Snrpn exon 1. Deletion B was 12 kb, extending fr om -25 kb to -37 kb. Deletion C was 58 kb, extending from -20 kb to -78 kb. Summary This introductory chapte r began with a discussion of genomic imprinting. Descriptions of the first experiments that revealed the existence of genomic imprinting were followed by the presentation of theori es regarding the evolution or purpose of genomic imprinting. Topics including th e extent of genomic imprinting, common characteristics of imprinted genes, and ca ndidates for the imprinting mark were then examined so as to provide background inform ation for a more detailed analysis of the Prader-Willi Syndrome (PWS) and Angelman Sydrome (AS) region of human chromosome 15. The second section of the introductory chapter presented the PWS/AS region as an example of imprinting relevant to human disease and as an opport unity to dissect the mechanism of imprinting. Following a brie f overview of the clinical features and molecular classes of PWS and AS, the disc ussion turned to studies of imprinting mutations in PWS and AS patients that led to the delineation of an imprinting control element termed the Imprinting Center (IC). A mouse model for PWS-IC deletions was introduced, and the results of experiments invol ving that mouse prove to be relevant to investigations of human PWS -IC deletions. Important aspects of both the human and mouse IC studies were expl ained by Dr. Brannan’s “Paternal-Only” model of IC regulation. Finally, the specific aim of the dissertation project to map the murine AS-IC was briefly described. The next chapter of this dissertation will catalogue the materials and

PAGE 30

20 methods involved in producing and analyzing the deletions in tended to reveal the location of the murine AS-IC. Chapter 3 will list th e details of the design and production of the deletions and will document the establishment of breeding colonies of mice for the 12 kb and 7 kb deletions. Chapter 4 will then repor t the results of DNA methylation and gene expression studies of the 12 kb and 7 kb deletion mice. The fifth and final chapter of the dissertation will discuss the results desc ribed in chapter 4 and will conclude by suggesting future experiments that could shed further light on the function of the IC. Figure 1-1. Life cycle of the imprint. During gametogenesis, the imprints on the maternal and paternal alleles are erased and then new imprints are established. In this manner, all alleles contained in sperm will carry a paternal imprint and all alleles contained in eggs will carry a maternal imprint. The imprints must be faithfully maintained on each allele in somatic tissues through many rounds of mitotic cell division. The process then repeats itself in the next generation, with the imprints once again being er ased and newly established during the next round of gametogenesis.

PAGE 31

21 Figure 1-2. Molecular classes of PWS and AS . Chromosomes are shown as lines, with the maternal chromosome labeled with “M” and the paternal chromosome labeled with “P”. The numbers beneat h each set of chromosomes indicate the proportion of all cases of the disease that fall into that particular molecular class. The classes of PWS are shown in the top row, and the classes of AS are shown in the bottom row. The first column shows that for both PWS and AS, roughly 70% of cases are due to a 4 mb deletion. The deletion is always on the paternally-inherited chromosome in PWS and the maternally-inherited chromosome in AS. The second column shows that 25% of PWS cases are due to UPD, but only 2% of AS cases are due to UPD. The third column shows that 25% of AS cases are caused by a mutation in the UBE3A gene. The single gene mutation class appears to be absent in PWS. Less that 5% of PWS and AS cases are caused by imprinting mutations. Rare cases of translocations and inversions for bot h PWS and AS are not shown. Also, a fifth molecular class exists for AS in which the cause of the disease has not been ascertained.

PAGE 32

22 Figure 1-3. Imprinted genes on human chromosome 15 and mouse chromosome 7. The top panel shows the imprinted genes found within human chromosome 15q11q13. The bottom panel shows the homol ogous region of mouse chromosome 7. The paternal and maternal chromo somes are shown by horizontal lines and are labeled with “P” and “M,” respectively. Genes are depicted by boxes, and transcription of each gene is indicated by an arrow. An oval at each gene represents DNA methylation. Vertical lines indicate clusters of snoRNA genes. The dashed horizont al line extending from the SNRPN gene represents a 460 kb transcript that c ontains the snoRNA transcri pts and that overlaps the UBE3A gene. “Paternal-specific” genes include MKRN3 , MAGEL2 , NDN , SNRPN , snoRNA genes, and an antisense transcript to UBE3A . The UBE3A and ATP10A genes are expressed biallelica lly in most tissues but are transcribed preferentially from the mate rnal allele in the brain. The bottom panel shows the existence of an additional paternal-specific gene, Frat3 , in mouse.

PAGE 33

23 Figure 1-4. Effects of imprinting mutations. The top panel shows the gene expression and DNA methylation profile found in a normal individual, as described in Figure 1-3. The middle panel shows the results of a PWS imprinting mutation, and the bottom panel shows the results of an AS imprinting mutation. In a PWS imprinting mutati on, the “paternal-specific” genes are inappropriately methylated and silenced. Loss of these gene products leads to the symptoms of PWS. Biallelic expression of UBE3A and ATP10A in the brain is speculative. In an AS im printing mutation, the “paternal-specific” genes are inappropriately demethylat ed and activated on the maternal chromosome. Activation of the UBE3A -antisense transcript on the maternal allele is predicted to cause the loss of UBE3A expression in crit ical regions of the brain. There is evidence for the loss of methylation at SNRPN and inappropriate activation of this gene on the maternal chromosome in response to an AS imprinting mutation. The cha nges in methylation and expression of other genes shown here is speculative.

PAGE 34

24 Figure 1-5. Location and bipartite nature of the imprinting center. The boxes represent exons 1 to 10 of the SNRPN gene. The small horizontal lines indicate the size of various deletions causing AS or PWS, and the vertical rectangles show the smallest regions of deletion overlap (SRO) for deletions. Some imprinting mutations are caused by microdeletions near the SNRPN gene that disrupt a regulatory element called the imprinti ng center (IC). By comparing the extents of the deletions in several PWS patients, the SRO involved in PWS has been narrowed to a 4.3 kb region spanning the SNRPN promoter. The SRO in AS patients has been narrow ed to an 880 bp region located 35 kb upstream of SNRPN . It therefore appears that the IC has a bipartite structure. One critical element is involved in PWS (the PWS-IC), and a distinct component is involved in AS (the AS-IC).

PAGE 35

25 Figure 1-6. Inheritance pattern s of IC deletions. The panel on the left (A) shows a pedigree for PWS-IC deletions. The pa nel on the right (B) shows a pedigree for AS-IC deletions. Boxes depict male s and circles depict females. Dots indicate unaffected carriers, and filled boxes or circles indicate affected individuals. Deletions of the PWS-IC most often appear to originate in the female germline and are silent in the next generation. Disease results only upon transmission of the deletion through a male germline. In contrast, AS-IC deletions are benign when passed through th e male germline, and result in AS only upon maternal transmission.

PAGE 36

26 Figure 1-7. The 35 kb PWS-IC deletion mouse model. The boxes represent exons 1 to 10 of the SNRPN gene. The lower horizontal line indicates the deletion extending from 16 kb upstream to 19 kb downstream of SNRPN exon 1. The deletion removed exons 1-6 of SNRPN . Mice inheriting the deletion on the paternal chromosome showed a failure to thrive and died within 7 days after birth. Affected mice did not express Mkrn3 , Ndn , or Snrpn . These results confirmed that the position and general function of the PWS-IC is conserved between human and mouse.

PAGE 37

27 Figure 1-8. The “Paternal-Only” model to ex plain the function of the IC. This model proposes that the PWS-IC functions as a positive element to promote paternal gene expression in somatic tissues. Paternal transmission of a PWS-IC deletion leads to the loss of paternal-s pecific gene expression and results in PWS. The AS-IC is described as a ne gative element operating in the female germline to suppress the paternal patter n of gene expressi on by inhibiting the action of the PWS-IC. Regulation of the tissue-specific imprinting of UBE3A is predicted to be the indirect result of a brain-specific paternally expressed UBE3A -antisense transcript whose expre ssion is regulated by the PWS-IC. Maternal transmission of an AS-IC dele tion leads to inappropriate activation of the normally paternal-specific genes on the maternal allele, including the UBE3A -antisense transcript. This results in AS due to loss of UBE3A expression in critical regions of the brain.

PAGE 38

28 Figure 1-9. Deletions upstream of Snrpn to locate the murine AS-IC. In an effort to locate the murine AS-IC, deletions were created upstream of Snrpn . If mice inheriting a deletion on the maternal ch romosome showed signs of an AS imprinting mutation, this would provide evidence that the AS-IC was located within that deletion. Homologous recombination in ES cells was used to engineer 3 nested deletions. Dele tion A was 7 kb, extending from –30 kb to – 37 kb relative to Snrpn exon 1. Deletion B was 12 kb, extending from –25 kb to –37 kb. Deletion C was 58 kb, extending from –20 kb to –78 kb.

PAGE 39

29 CHAPTER 2 MATERIALS AND METHODS Cloning To construct the targeting vectors used to produce the 12 kb, 7 kb, and 58 kb deletions in ES cells, several steps were requi red to add additional pieces of DNA to each plasmid. This involved shuffling DNA frag ments from one plasmid to another using restriction enzyme digestion of the DNA to linearize a circular plasmid or to cut a plasmid into more that one fragment. Phenol -chloroform extraction was used to clean the DNA by removing proteins. Ethanol precipitati on was used to remove residual traces of phenol-choroform and to concentrate DNA by resuspending in a smaller volume. Ligations were performed to join fragment s of DNA together, and the ligation products were used to introduce circularized plasmids into chemically competent E. coli . Restriction Digest For a typical restriction digest, typica lly 0.5 to 1.0 micrograms (ug) of plasmid DNA or 10 ug of genomic DNA was cut. (For plasmid DNA to be gel purified, 10 ug was cut.) Most often the DNA was digested in a 1.5 milliliter (mL) microfuge tube (Sarstedt), but large numbers of digests were sometimes done in 96-well microtitre plates (Fisher) sealed with microplate adhesive film (USA Scientific, Inc.). Most digests were done in a volume of 30 uL, with the total volume of DNA added being dependent on the concentration of each sample. The amount of water (H2O) added to each 30 uL reaction was dependent on the volume of the DNA being cu t. (For this and a ll other applications listed in this section, only au toclaved Millipore-treated H2O was used.) For example,

PAGE 40

30 10 uL of DNA, 15 uL of water, 1 uL of bovine serum albumin (BSA), 3 uL of the appropriate 10X restriction enzyme buffer (NEB ), and 1 uL of the a ppropriate restriction enzyme (NEB) would be added to the reaction. The tube or 96-well plate would then be incubated at the temperature appropriate for maximal activity of the restriction enzyme (in most cases 37oC). Plasmid DNA was cut for at le ast 2 hours and sometimes as long as overnight. Genomic DNA was cut overnight, and then an extra 1 uL of restriction enzyme was added and the sample was cut for another 2 to 4 hours. Phenol-Chloroform Extraction Phenol-chloroform extraction was used to remove proteins from DNA. The volume of DNA to be extracted was br ought up to at least 200 uL with H2O in a 1.5 mL microfuge tube, although volumes as great as 50 0 uL could be extracted in a single tube. An equal volume of phenol-chloroform solution was added, and the sample was mixed by vortexing. The phenol-chlorofor m mixture contained 25 mL of phenol (USB), 24 mL of chloroform (Fisher) and 1 mL of isoamyl alc ohol (Fisher). After vor texing, the tube was spun at 13,000 rotations per minute (rpm) in an Eppendorf table-top centrifuge, and the upper aqueous layer containing the DNA was transf erred to a clean tube. In cases that a sample had to be especially clean, the sample was extracted 1 or 2 times with straight phenol, 2-3 times with phenol-chloroform, and 1 to 2 times with straight chloroform. These extra steps were typically done for DNA that was to be used for electroporating embryonic stem (ES) cells. Ethanol Precipitation After being extracted with phenol-chlorof orm, a sample was always subjected to ethanol precipitation to remove traces of the phenol-chloroform mixture. A 1/10 volume of 3 molar (M) sodium acetate (NaOAc) was added to the DNA solution, and the sample

PAGE 41

31 was mixed by vortexing. Then 2 volumes of 100% ethanol were added, and the sample was again mixed by vortexing. In cases of la rge amounts of DNA, the tube was spun in a microcentrifuge immediately. In order to maximize recovery of DNA, however, most samples were placed at –80 oC for at least 20 minutes and as long as overnight. The sample was then spun at 13,000 rpm for between 2 and 20 minutes in the microcentrifuge to pellet the DNA. Longer spin times were used maximize the recovery of small amounts of DNA. The supernatant was then poured off, and 1.0 mL of 70% ethanol (35 mL 100% ethanol and 15 mL H2O) was added to the tube. The sample was vortexed and then spun again at 13,000 rpm for between 2 and 10 minutes. The supernatant was poured off and this process of washing with 70% ethanol was repeated. Following the second wash, the supernatant was poured off and residual ethanol at the bottom of the tube was carefully removed by pipetting. The sample was then dried, either by placing in a speedvac for between 2 and 10 minutes or by air-drying for between 2 and 40 minutes. The DNA was then resuspended in an appropriate volume of H2O. In some cases this required vortexing the sample and/or heating it at 55 oC. The concentration of DNA could then be determined by measuring a 40-fold dilution in the spectrophotometer (Bio-Rad) or by running a small amount on an agarose gel and comparing against a standard of HindIII-cut lambda phage DNA marker. Removal of Phosphate Groups from Vector In order to reduce the frequency of a linea rized plasmid re-circularizing due to an intramolecular ligation reaction, the phosphate groups were removed from the ends of the DNA. This was done by treating with calf intestinal alkaline phosphatase (CIAP) (Invitrogen). The DNA was resuspended in 16 uL in a 1.5 mL microfuge tube. Then 2 uL of CIAP “reaction” buffer, 2 uL of CI AP “dilution” buffer, and 0.5 uL of CIAP

PAGE 42

32 enzyme were added to the sample. The sample was mixed gently by flicking and spinning down, and was placed in a 37 oC water bath for 30 minutes. The sample was then transferred to a 55 oC water bath for 10 minutes to denature the enzyme. Following this, another 0.5 mL of CIAP was added and the sample was again mixed, spun down, and incubated at 37 oC for 30 minutes. The sample was then transferred to 55 oC for 10 minutes. Following this procedure, the sa mple was then ready for gel purification. Gel Purification The fragment of DNA to be purified was separated by resolving on an 0.8 agarose gel containing 4 uL of ethidium bromide pe r 100 mL of gel (Bio-R ad). If the Wizard Prep kit (Promega) was to be used for pur ification, the gel was made with “NuSieve GTG” low-melt agarose (Cambrex Bioscience Rockland, Inc.). If the QIAquick kit (Qiagen) or electro-elution was to be used for purification, the gel was made with regular agarose (Invitrogen). Once the fragment of DNA was separated from other fragments on the gel, the appropriate band was visualized on an ultraviolet (UV) trans-illuminator and was excised from the gel with a razor bl ade as quickly as possible to minimize UV nicking of the DNA. The DNA was then re covered from the gel slice by using the Qiaquick gel extraction kit, the Wizard Pr ep gel extraction kit, or electro-elution. QIAquick gel-extraction kit The QIAquick kit was used according to the manufacturer’s recommendations. Once the band of DNA was excised from the ge l, the gel slice was placed in a 1.5 mL microfuge tube. The gel slice was first wei ghed to determine to volume of “QG” solution required. After the scale was tared using an empty microfuge tube, the tube containing the gel slice was then placed on a scale to determine the weight of the gel slice. For each 100 milligrams (mg) of gel, 300 uL of “QG” buffer was adde d. If a gel slice weighed

PAGE 43

33 more than 300 mg, it was divided in two and each half was processed in parallel in two separate tubes. The sample was then incubated in a 55 oC water bath for 10 minutes and vortexed intermittently throughout this incuba tion. Once the gel slice was dissolved, the color of the solution was examined as an i ndication of pH. If the color of the solution had turned from yellow to orange, 10 uL of 3 M NaOAc pH 5.0 was added, and the tube was vortexed. (It was extremely rare for a samp le to require such a pH adjustment.) For each 100 mg of gel that was dissolved, 100 uL of 100% isopropanol was added and the tube was vortexed. The sample was then tran sferred to a QIAquick spin column in a 2 mL collection tube and centrifuged at 13,000 rpm in an Eppendorf table top centrifuge for 1 minute, and the flow through was discarded. The capacity of the spin column was less than 800 uL, so samples exceeding this volume sometimes requi red 2 or 3 rounds to load onto and spin through the column. On ce the flow through was discarded, the sample was washed by adding 0.75 mL of buffer “P E” and spinning the column at 13,000 rpm for 1 minute. The flow through was dis carded, and the column was spun again at 13,000 rpm for 1 minute to remove residual etha nol from buffer “PE” remaining on the column. The column was then placed in a clean 1.5 mL microfuge tube from which the cap had been removed. Between 30 and 50 uL of buffer “EB” was added to the column, and the column was allowed to stand for 1 minute before eluting the DNA by spinning the column at 13,000 rpm for 1 minute. Since some residual ethanol in the sample was a concern, the DNA was often ethanol precipitated and resuspende d prior to ligation or use for other applications. Wizard-Prep gel-extraction kit The Wizard-Prep kit was used according to the manufacturer’s recommendations. In preparation for the final step of the procedure, 1 mL of H2O was placed in a 1.5 mL

PAGE 44

34 microfuge tube and incubated at 65 oC for the duration of the protocol. Once the band of DNA was excised from the low-melt gel, the gel slice was placed in a 1.5 mL microfuge tube and incubated in a 65 oC water bath for 10 minutes to liquefy the gel slice. The Promega Wizard DNA cleanup resin was resusp ended and mixed by swirling, and then 1 mL of this solution was added to the micr ofuge tube containing the molten agarose. A filter unit was labeled to identify the sample, a nd the filter unit was attached to the end of a Leur-lock syringe. The molten agarose/clea nup resin solution was pipetted into the empty syringe and pushed through the syringe and filter unit using a plunger, and the eluent was discarded. The filter unit was detached, the plunger removed, and the filter unit was then reattached to the syringe. The cartridge was then washed by passing 3 mL of 80% isopropanol (40 mL 100% isopropanol plus 10 mL H2O) through the syringe with the plunger. The filter cartridge was then detached and placed onto a 1.5 mL microfuge tube from which the cap had been removed. The filter was centrifuged at 13,000 rpm in an Eppendorf table top centrifuge for 25 s econds to remove residual isopropanol. The filter cartridge was then transferred to a clean microfuge tube from which the cap had been removed, and 30 uL of H2O at 65 oC was pipetted onto the center of the filter cartridge. After waiting for 1 minute, the cartridge was spun at 13,000 rpm for 20 seconds to elute the DNA. If any resin came through the column along with the DNA, the resin was removed by spinning the sample at 13,000 rpm for 10 minutes and transferring the supern atant to a fresh 1.5 mL microfuge tube. Since some residual isopropanol in the sample was a concern, th e DNA was often ethanol precipitated and resuspended prior to ligation or use for other applications.

PAGE 45

35 Electro-elution Once the band of DNA was excised from th e gel, the gel slice was placed in a 1.5 mL microfuge tube while di alysis tubing was prepared by washing thoroughly inside and out with distilled H2O. The dialysis tubing was cut ge nerously to allow for weighted, locking alligator clips to be placed on either si de of the gel slice. The dialysis tubing was smoothed until entirely flat and the fist clip wa s affixed to one end of the tubing. The gel slice was then pushed into the tubing using a small spatula. The gel slice was positioned snugly against the first clip and one side of the tubing, and then 1 mL of 0.5 X TBE was added to the tubing. The tubing was squeezed starting from the closed end and working towards the open end to push out any air bubbles and to minimize the volume of TBE remaining in the tubing. It was found that mi nimizing the volume of liquid remaining in the tubing was a critical step in the protocol. The second alligator clip was then fastened to the end of the tubing to seal the gel sli ce inside. The tubing containing the gel slice was then entirely submerged in a standard gel-running electrophoresis apparatus and subjected to 100 volts for 30 to 90 minutes. Small fragments of DNA (for example, a 1.8 kb neomycine resistance cassette) would elute in roughly 30 mi nutes, but larger fragments (for example, a 19 kb linear plasmi d) would require much longer to elute and would elute with less efficien cy. The advantage of using weighted clips was to help prevent the tubing from floating around during the elution, but unweighted clips could be substituted as needed. The extent of elut ion of the DNA was visualized on a UV transilluminator as a smear of ethidium bromide stai n on one edge of the dialysis tubing. This was done as quickly as possible so as to avoid UV-nicking the DNA. It was found that an estimate of ethidium bromide staining at this step was a very good predictor of the final yield of DNA. Very little DNA was typi cally recovered for sa mples that were not

PAGE 46

36 visible by ethidium bromide staining. (No et hidium bromide was added to the solution in the dialysis tubing, but enough of it was already intercalated into th e DNA in the gel slice to allow for visualization of the DNA in the tubing.) Once th e DNA was eluted, the current to the electrophoresis apparatus was reversed for 20 seconds to help back the DNA off of the edge of the dialysis tubi ng. The tubing was then removed from the electrophoresis apparatus, the outside of the t ubing was dried with a paper towel, one clip was carefully removed, and the liquid contai ning the DNA was removed with a pipettor and transferred to a 1.5 mL microfuge tube . Before being removed, the liquid was pipetted up and down to help wash the DNA off of the edge of the dialysis tubing. An additional 200 uL of liquid was pipetted up and down inside the tubing and then combined with the liquid in the microfuge tube. The total final volume was typically less than 0.5 mL. The sample was then subj ected to phenol extraction and ethanol precipitation. DNA Ligation To ligate fragments of DNA together, a pproximately 100 nanograms (ng) of each fragment was placed in a 1.5 mL microfuge tube. The total volume of both samples combined was typically less than 5 uL. A variable amount of H2O was added to bring the volume up to 7 uL, and then 1 uL of ATP (Sig ma), 1 uL of 10 X Buffer (NEB), and 1 uL of ligase enzyme (NEB) were added to the t ube. Addition of ATP was optional only only really necessary when the NEB 10 X ligase buffer had been used extensively and subjected to many freeze-thaw cycles. The tube was gently flicked to mix the contents, spun down, and placed at 14 oC overnight. If ligations did not work on the first try, the ratio of the amount of vector DNA to insert DNA was varied by 2, 5, and 10 fold in either direction

PAGE 47

37 Transformation A 1.5 mL microfuge tube containing a 30 uL aliquot of chemically-competent DH5-alpha cells (Invitrogen) was removed from –80 oC and the cells were allowed to thaw on ice. The entire 10 uL ligation reac tion was added to the tube, and the contents were mixed by gentle flicking. The ligation reaction and competent cells were incubated together on ice for at least 20 minutes and as long as 60 minutes. The cells were then heat shocked by incubating the tube at 42 oC for 45 seconds and placed on ice for 2 minutes. The cells were allowed to recove r by adding 1.0 mL of Luria Broth (LB) to the tube and shaking at 200 rpm in an orbital shaker incubator at 37 oC for at least 20 and at most 40 minutes. The tube was then spun at 7,000 rpm for 1 minute, most of the liquid was decanted, and the bacteria were resusp ended by pipetting up and down in a volume of 100 uL of LB. The transformation was then spread onto a LB plate containing ampicillin, and the plate was placed in a 37 oC incubator overnight. The next day, colonies were picked and cultured individua lly in 3 mL cultures. The cultures were shaken at 200 rpm at 37 oC overnight in an orbital shak er. The next day, plasmid DNA was isolated from each colony by standard alka line lysate and screened by restriction digest. Larger DNA preparations for correct clones were prepared using the Qiagen plasmid maxi kit. Culturing Embryonic Stem Cells Embryonic stem (ES) cells were grown on a feeder layer of mito tically-inactivated mouse embryonic fibroblasts (MEFs) on gelatin coated plates. Mitomycin C was used to prepare the mitotically inactivated feeder laye r. ES cells were fed once or twice a day with prewarmed 3M media and were always fed one hour before splitting or feeding. MEFs were seeded onto six gelatinized 10 cm plates at a density of 3.5 x 106 cells per

PAGE 48

38 plate. One vial of passage 8 CJ7 ES cells was then thawed and di stributed onto one of the six MEF-coated 10 cm plates. The ES cells were expanded for three days, and on the fourth day they were trypsinized and split onto the five remaining MEF-coated 10 cm plates. The ES cells were grown for a pproximately two more days until about 70% confluent, and then they were trypsiniz ed for electroporation. In preparation for plating the cells following th e electroporation, fifteen 10 cm plates were gelatinized and then filled with 10 mL of 3M media. The trypsinized ES cells were counted, and 5 x 107 cells were resuspended in a total volume of 0.8 mL PBS and transferred to an electroporation cuvette. A to tal of 100 uL of the lineari zed targeting vector DNA was added to the electroporation cuvette, and the cells were gently mi xed. A BioRad gene pulser was used to electroporate at 0.8 kV a nd 3 uF. The time constant was usually 0.1. After 10 minutes, the cells were resuspended in 14.5 mL of 3M media and 1 mL of the media containing the electroporated cells was th en added to each of fifteen gelatinized 10 cm plates. The next day, the cells we re fed with 3M containing 200 ug/mL G418. The cells were fed daily and kept under G 418 selection for 7 or 8 days until colonies grew to a suitable size. Ten 24-well plat es were gelatinized and seeded with 2 x 105 MEFs per well in 1 mL of 3M plus G418 per well. The 10 cm dishes were washed and covered with 7 mL of PBS, and a P200 pipettor was used to pick individual colonies in a volume of 30 uL of PBS. Each colony was placed in a separate well in a 96-well plate, and once 96 colonies had been picked a multichannel pipettor was used to aliquot 30 uL of trypsin into each well. The cells were tritura ted by pipetting up and down and then transferred to 24-well plates. The cells were fed once or twice daily for 4 or 5 days until most of the wells in the 24 well plate appeared to be approximately

PAGE 49

39 70% confluent. A parallel set of gelatinized 24-well plates was prepared for the purpose of expanding and making DNA from each clone. The ES cells were treated with 200 uL trypsin and triturated. A total of 500 uL of quench media was added to each well, and half of each clone was transf erred to the fresh 24-well plat es. A total of 500 uL of 2x freeze media was then aliquoted into each we ll, and the ES cells were frozen in the plates at –80 oC. Between 4 and 7 days later, DNA was isolated for each clone and used for Southern blot analysis to screen for targeted clones. Southern Blot For genomic DNA, 10 ug of DNA was digested in a total volume of 40 uL. The DNA was resolved on a 0.8% agarose gel, the DNA was UV-nicked by exposing the gel on a trans-illuminator for 5 minutes, the la mda DNA markers were identified by stabbing the bands with a needle dipped in India ink, and the gel was placed in alkali solution for 45 minutes to denature the DNA. The gel was then soaked in neutralizing solution for 90 minutes, and the gel was placed on top of a Whatman paper wick overlaying a plexiglass plate suspended on the edges of a pyrex dish containing 400 mL of 10 X SSC. A piece of Hybond membrane (Amersham) was placed directly onto the gel, and then two layers of Whatman paper and a stack of paper towels was placed on top. The DNA was allowed to transfer to the Hybond membra ne over night by capillary transfer. The next day, the position of the lamda DNA bands were marked onto the Hybond with India ink, the membrane was baked at 80 oC for 2 hours to fix the DNA to the membrane, and then the membrane was hybridized overni ght with a radioactively-labeled probe overnight at 65 oC. The following day, the membrane was subjected to two 15 minute washes in 0.2 X SSCP and 0.1% SDS at 65 oC. Kodac XAR film was then exposed to the membrane at –80 oC.

PAGE 50

40 Northern Blot For total RNA, 10 ug of RNA was load ed onto a 1.0% agarose gel containing formaldehyde. A piece of Hybond membrane (Amersham) was placed directly onto the gel, and then two layers of Whatman paper and a stack of paper towels was placed on top. The DNA was allowed to transfer to the Hybond membrane over night by capillary transfer. The next day, the position ribos omal RNA bands were marked onto the Hybond with India ink, the membrane was baked at 80 oC for 2 hours to fix the RNA to the membrane, and then the membrane was hybridi zed overnight with a radioactively-labeled probe overnight at 65 oC. The following day, the membrane was subjected to two 15 minute washes in 0.2 X SSCP and 0.1% SDS at 65 oC. Kodac XAR film was then exposed to the membrane at –80 oC. Sodium Bisulfite Analysis Protocols for sodium bisulfite analysis were kindly provide d by Danielle Maatouk and Christine Mione Kiefer. Between 5 a nd 10 ug of DNA in a total volume of 37.8 uL of H2O was denatured by adding 4.2 uL of fres hly prepared 3M NaOH and incubating at 37 oC for 30 minutes. Next, 198 uL of freshly prepared 2.0M sodium metabisulfite and 10mM hydroquinone was added, and the sample was incubated in the dark at 55 oC for 16 to 20 hours. The bisulfite was removed us ing the Promega Wizard-Prep kit, and the sample was eluted from the column in 45 uL of H2O. To denature the DNA, 5 uL of 3M NaOH was added, and the sample was incubated at 37 oC for 15 minutes. The DNA was ethanol precipitated, wash ed, dried, and resupended in 100 uL H2O. Between 2 and 10 uL of the sample was used for a PCR amplification using primers specific for converted DNA. Either an aliquot of the PCR reaction was used directly for cloning with the Invitrogen Topo TA cloning kit, or the product was gel purified for cloning with the

PAGE 51

41 Promega PGEM-T Easy kit. Colonies we re screened by restriction digest, and DNA from clones containing the correct insert was phenol extracted, ethanol precipitated, washed, dried, and resuspended in 25 uL H2O. The concentration of each sample was determined by spectrophotometry (BioRad), and all samples were adjusted to a concentration of 0.5 ug/uL. Sequencing reac tions were prepared using the Big Dye Terminator kit (Perkin Elmer) and processed by the University of Florida Center for Mammalian Genetics DNA sequencing core faci lity. Sequences were analyzed using the Sequencher 4.1 program (Gene Codes).

PAGE 52

42 CHAPTER 3 DELETION STRATEGY Strategy for Deletions At the outset of this work, before the completion of the mouse genome sequencing effort, the sequence of th e region upstream of the Snrpn gene was not available. The only information that could be used as a guide to finding the murine AS-IC was the position of the AS-SRO in human. At that time, nine AS-IC deletions had been reported, and comparison of the extents of those deleti ons had defined a 880 bp SRO located 35 kb upstream of SNRPN [121]. Interestingly, the 88 0 bp region contained the u5 exon, previously shown to be part of transcripts originating upstream of and splicing into the SNRPN gene [136]. As the u5 exon was found to be deleted in every AS patient reported to have an IC deletion, this ra ised the possibility that the “upstream” or “IC” transcripts might play some role in the imprinting pr ocess. On the other hand, it also seemed possible that a cis-acting elemen t critical to the function of the AS-IC could be located nearby exon u5. In either case, the inform ation available at the beginning of this dissertation project indicated that the lo cation of the human AS-IC is about 35 kb upstream of the SNRPN gene. Given the lack of sequence information, a former graduate student in the lab, Dr. Stormy Chamberlain, attempted to use a probe for the human AS-SRO to locate the murine AS-IC by Southern blot. Unfortunate ly the human probe failed to cross-hybridize with mouse genomic or cloned DNA, so this approach was unsuccessful.

PAGE 53

43 It was decided that the mo st direct and feasible me thod of locating the murine AS-IC would be to create deletions upstream of Snrpn and then ascertain if mice that inherited these deletions maternally showed signs of an Angelman imprinting mutation (Figure 3-1). Dr. Cami Brannan began laying the groundwork for the deletion mapping project by screening a ph age library of fragments from a Bacterial Artificial Chromosome (BAC) which spanned the region. The BAC, labeled 397F16, was isolated by Dr. Chamberlain from a library purchased fr om Research Genetics. The advantage of using this particular library was that it had been construc ted from the 129S1/Sv strain of mouse, the same strain of mouse from wh ich the CJ.7 ES cell line had been isolated [137]. The source of DNA used for the arms of homology in the gene targeting construct was an important consideration, as this could greatly affect the targ eting frequency. By matching the source of DNA with the ES cell li ne to be targeted, it was possible to maximize the chance of recovering targeted ES cell clones by eliminating the potentially negative effects on the frequency of ho mologous recombination due to sequence divergence between the arms of homology and the endogenou s locus of the ES cell. Dr. Brannan then used a series of phage clones to create an EcoRI restriction map extending almost 50 kb upstream of Snrpn . The inserts from several phage clones were subcloned into the pBluescript KS vector so that they could be maintained in E. coli. In order to determine the orientation of the fr agments in pBluescript, the ends of some inserts were sequenced using primers for th e T7, T3, M13-Forward, or M13-Reverse sites flanking the pBluescript multiple cloning site. In some cases, DNA from BAC397F16 was digested with other restriction enzymes a nd fragments of interest were gel-purified and cloned directly into pBluescript. As the mouse genome project progressed and bits

PAGE 54

44 of sequence became available for parts of the region upstream of Snrpn , it became possible to design oligos and PCR amplify ar ms of homology or probes directly from BAC DNA. It was during this phase that the 58 kb deleti on construct was designed and created by a postdoc in the la b, Dr. Karen Johnstone, as desc ribed in the next section. There were two main considerations in de signing the deletion constructs. The first was a desire to extend the de letions as far upstream of Snrpn as possible so as to maximize the chances that the AS-IC would be encompassed by one of the deletions. The second major consideration was to make sure that the deletions did not encroach upon the PWS-IC. According to the predictions of the Paternal-Only model, as described in the introductory chapter, any deletion that involved the PWS-IC should result in PWS upon paternal transmission. And indeed, if the sole function of the AS-IC is to negatively regulate the PWS-IC during oogenesis, any de letion affecting both elements should be benign upon maternal transmission. Operating under the predictions of this model, we decided to leave a large buffer between the deletions and the Snrpn gene. Although the 3-prime ends of the deletions were somewhat arbitrary, it was determined that they should not extend any closer than 16 kb upstream of Snrpn , which was the 5-prime end of the PWS-IC deletion previously generated in the lab [123, 132]. Granted, this approach ran the risk of “missing” the AS-IC if it were located much closer to Snrpn in mouse than in human, but this risk was deemed to be smaller than the risk of impinging upon the PWS-IC. A major practical consideration in the design of the deletions was to isolate flanking probes and arms of homology that we re free of repeats. Unfortunately, this precluded extending the deletions all the way to the end of the phage restriction map

PAGE 55

45 roughly 50 kb upstream of Snrpn because the restriction map ended in a Line element. Given this constraint, a 3.9 kb EcoRI frag ment located –40.9 kb to –37.0 kb upstream of Snrpn exon 1 was selected as the 5-prime ar m of homology for both the 12 kb and the 7 kb deletion constructs so as to reserv e enough upstream non-repetitive DNA to use as a flanking probe. The idea behind designing two constructs to generate nested deletions was to set up several different scenarios th at, regardless of the outcome of the experiment, would enable either the identification of a sm aller interval containing the AS-IC or, alternatively, the exclusion of a larger area. If the smaller of the two deletions revealed an imprinting mutation, then it would serve to define a 7 kb region containing the murine AS-IC. If the larger deleti on showed indications of an imprinting mutation but the smaller deletion did not, howev er, it would then be possible to identify a 5 kb region within the larger deletion that harbored an element critical to the function of the AS-IC. If neither deletion showed an effect, then it would be possible to determine that the entire 12 kb region is NOT critical to the imprinting process. Once the mouse genome project began generating and releasing sequence information for the region upstream of Snrpn , in became possible to design a larger deletion in addition to the 12 kb and 7 kb deleti ons. Dr. Brannan was able to sort through the numerous unordered contigs from BAC 306 D24 and find a region of repeat-free DNA that we initially estimated was between 100 and 150 kb upstream of Snrpn . It should be noted that more recent searches of the upda ted sequence database by Dr. Johnstone and Mr. Chris Futtner, a graduate student in the la b, have revealed that this estimate of the

PAGE 56

46 location of the upstream loxP element was inco rrect. Current information indicates that the larger deletion extents to –78 kb relative to Snrpn exon 1. Dr. Johnstone used this upstream sequence information to order oligos and PCR-amplify arms of homology and flanking pr obes directly from CJ.7 ES cell genomic DNA. It was decided that in order to create a larger deletion it would be necessary to employ the CreloxP system for chromosome engin eering. Cre is a site-specific recombinase that recognizes asymmetric 34-nucleotide loxP sites. When Cre catalyses recombination between two loxP sites that are in the same orientation on a chromosome, the reaction results in the excision of any DNA that is flanked by the loxP sites. In order to generate a large deletion, it was necessary to produce two different targeting constructs that could be used sequentially to introduce loxP sites at two differe nt locations upstream of Snrpn . One construct was designed to insert a loxP site 20 kb upstream of Snrpn. Dr. Johnstone then created a second construct to insert a loxP element 78 kb upstream of Snrpn . By isolating ES cells that contained both loxP sites and transiently transfecting these cells with a plasmid encoding the Cre recombinase, it was possible to generate a 58 kb deletion. Although ES cells carrying the 58 kb deletion were used in numerous attempts to generate chimeras, none of the chimeras produced from these ES cells showed germline transmission. Karyotype an alysis of the 58 kb deletion ES cells performed by Dr. Johnstone revealed what app eared to be a transloc ation. As ES cells that carry a chromosomal abnormality are unlikely to be germline competent, we abandoned our attempts to produce chimeras from the 58 kb deletion ES cells. Because a breeding colony was not established and theref ore no analysis of this deletion could be

PAGE 57

47 done, a detailed discussion of the plasmid design and ES cell targeting will not be included in this chapter.. Plasmid Creation Building the 12 kb Deletion Plasmid The first step in creating the 12 kb de letion construct invol ved digesting BAC 397F16 with EcoRI, resolving on a 0.8% agarose gel, excising the 3.9 kb band, gel-purifying the fragment, and cloning it in to the EcoRI site of pBluescriptKS to produce plasmid number 387. In this plasmid the 5-prime end of the 3.9 kb fragment is closer to the T3 site of Bl uescript. The 3.9 kb fragment, whose ends are located at – 40.9 kb and –37.0 kb relative to Snrpn exon 1, serves as the 5-prime arm of homology in the deletion construct. The second step involved PCR amplifying a 4.8 kb fragment using primers flanking BglII sites located at –25.0 kb and –20.2 kb relative to Snrpn exon 1. The oligos used were “W53-4.8-BglII-Forward” (5’CTGAATTTGTTCCTAAAAGCACA-3’) and “W54-4.8-BglII-Reverse” (5’-GAAAATGCTCTA CAGCCTTGC-3’). The PCR product was cloned into the pGEM-T Easy vector to produce plasmid 388. The 4.8 kb fragment serves as the 3-prime arm of ho mology in the deletion construct. The third step involved cutting plasmid 387 with SmaI to linearize the vector at the SmaI site in the Bluescript multiple cloni ng site, phenol-chloroform extracting, ethanol precipitating, treating with CIAP, resolvi ng on a 0.8% agarose gel, excising the 6.8 kb band, and gel-purifying the vector frag ment. Concurrent with this, a PGK promoter-driven neomycin (neo ) resistance cassette flanked by loxP sites in direct orientation was isolated from plasmid #350. This involved cutting plasmid 350 with SmaI, resolving on a 0.8% agarose gel, ex cising the 1.8 kb band, and gel-purifying the

PAGE 58

48 neo fragment. The 6.8 kb vector fragment and the 1.8 kb neo fragment were ligated together to produce plasmid 389 containing the 3.9 kb 5-prime arm in the EcoRI site and the 1.8 kb neo fragment in the SmaI site of Bluescript. In plasmid 389 the PGK promoter of the neo fragment is closer to the T7 site of Bluescript, and th e polyadenylation signal is closer to the T3 site and the 5-prime ar m of homology fragment. In this orientation, the loxP sites are pointing towards the T7 site and away from the 5-prime arm of homology. The fourth step involved cu tting plasmid 389 with BamHI to linearize the vector at the BamHI site in the Bluescript multiple cloning site, phenol-chloroform extracting, ethanol precipitating, treating with CIAP, re solving on a 0.8% agarose gel, excising the 8.6 kb band, and gel-purifying the fragment. Concurrent with this, the 4.8 kb 3-prime arm of homology in plasmid 388 was liberated by cutting with BglII, resolved on a 0.8% agarose gel, excised from the gel, and gel-purified. The 8.6 kb vector fragment and the 4.8 kb fragment were ligated together to produce plasmid 390 containing the 3.9 kb 5-prime arm in the EcoRI site, the 1.8 kb neo fragment in the SmaI site, and the 4.8 kb 3-prime arm in the BamHI site of Bluescript . In this plasmid the 5-prime end of the 4.8 kb BglII fragment is closer to the neo fragment and the 3-prime end of the 4.8 kb BglII fragment is closer to the T7 site of Bluescript. The fifth step involved cutting plasmid 390 w ith SalI to linearize the vector at the SalI site in the Bluescript multiple cloning site, resolving on a 0.8% agarose gel, excising the 13.4 kb band, and gel-purifying. (In this case the lineariz ed vector was not treated with CIAP.) Concurrent with this, a PGK promoter-driven thymidine kinase (TK) gene was isolated from plasmid 155. This involve d cutting plasmid 155 w ith SalI, resolving

PAGE 59

49 on a 0.8% agarose gel, excising the 2.9 kb band, and gel-purifying the TK fragment. The 13.4 kb vector fragment and the 2.9 kb TK frag ment were ligated together to produce plasmid 391 containing the 3.9 kb 5-prime arm in the EcoRI site, the 1.8 kb neo fragment in the SmaI site, the 4.8 kb 3-prime arm in th e BamHI site, and the 2.9 kb TK gene in the SalI site of Bluescript . This 16.3 kb plasmid was linearized with NotI, phenol-chloroform extracted, a nd ethanol precipitated prior to ES cell electroporation. Building the 7 kb Deletion Plasmid The first step in creating the 7 kb dele tion construct involved digesting BAC 397F16 with EcoRI, resolving on a 0.8% ag arose gel, excising the 3.9 kb band, and gel-purifying and cloning it into the EcoRI s ite of pBluescriptKS to produce plasmid 392. In this plasmid the 5-prime end of the 3.9 kb fragment is closer to the T7 site of Bluescript. The 3.9 kb fragment, whose ends are located at –40.9 kb and –37.0 kb relative to Snrpn exon 1, serves as the 5-prime arm of homology in the deletion construct. The second step involved digesting BA C 397F16 with HincII, resolving on a 0.8% agarose gel, excising the 8.0 kb band, a nd gel-purifying and cloning the fragment into the HincII site of Bluescript to produce plasmid 393. The 8.0 kb fragment, whose ends are located at –30.0 kb to –22.0 kb relative to Snrpn exon 1, serves as the 3-prime arm of homology in the deletion construct. The third step involved cutting plasmid 392 w ith EcoRV to linearize the vector at the EcoRV site in the Bluescript multiple cloning site, phenol-chloroform extracting, ethanol precipitating, treating with CIAP, re solving on a 0.8% agarose gel, excising the 6.8 kb band, and gel-purifying the vector fragment. Concurrent with this, a PGK promoter-driven neomycin (neo ) resistance cassette flanked by loxP sites in direct orientation was isolated from plasmid 350. This involved cutting plasmid 350 with SmaI,

PAGE 60

50 resolving on a 0.8% agaros e gel, excising the 1.8 kb ba nd, and gel-purifying the neo fragment. The 6.8 kb vector fragment and the 1.8 kb neo fragment we re ligated together to produce plasmid 394 containing the 3.9 kb 5-prime arm in the EcoRI site and the 1.8 kb neo fragment in the EcoRV site of Bl uescript. (In this case the 1.8 kb SmaI fragment was able to be liga ted into the EcoRV site of Bl uescript because both enzymes produce blunt ends.) In plasmid 394 the PGK pr omoter of the neo fragment is closer to the T7 site in Bluescript, and the polyadenylatio n signal is closer to the T3 site. In this orientation, the loxP sites are pointing towards the T7 site and towards the 5-prime arm of homology. The fourth step involved cu tting plasmid 394 with SalI to linearize the vector at the SalI site in the Bluescript multiple cloni ng site, phenol-chloroform extracting, ethanol precipitating, resolving on a 0.8% agar ose gel, excising the 8.6 kb band, and gel-purifying the fragment. (In this case the linearized vector was not treated with CIAP.) Concurrent with this, the 8.0 kb 3prime arm of homology in plasmid 393 was liberated by cutting with HincII, phenol-chloro form extracted, and precipitated. Linkers containing an XhoI site (NEB S1072S) were resuspended to a concentration of 120 pmol/ul and treated with T4 kinase. The kinase was heat inactivated at 65 oC for 20 minutes, and the 8.0 kb HincII fragment and XhoI linkers were ligated together overnight. The ligase was th en heat inactivated at 65oC for 20 minutes, and the linkers were “trimmed back” by cutting with XhoI. The 8.0 kb HincII/XhoI fragment was then resolved on a 0.8% agarose gel, excised, gel-pu rified, and ligated to the 8.6 kb SalI-cut vector. This produced plasmid 395 containi ng the 3.9 kb 5-prime arm of homology in the EcoRI site, the 1.8 kb neo fragment in the EcoRV site, and the 8.0 kb 3-prime arm of

PAGE 61

51 homology in the SalI site of Bluescript. In this plasmid the 5-prime end of the 8.0 kb HincII fragment is closer to the neo frag ment and the 3-prime end of the 8.0 kb BglII fragment is closer to the T3 site of Bluescri pt. It should be note d that the XhoI oligo linker approach was adopted because a HincII site was present in the 3.9 kb 5-prime arm of homology. Initial efforts to clone the 8.0 kb HincII fragme nt directly into the HincII site of Bluescript by employing a partial di gest of plasmid 394 were unsuccessful. The linker approach took advantag e of the fact that the SalI and XhoI enzymes produce compatible cohesive ends. Ligation of th e SalI-XhoI-HincII fragments resulted in the “destruction” of all three restriction sites so that they could not be re-cut by those enzymes. The fifth step involved cutting plasmid 395 with XhoI to linearize the vector at the XhoI site in the Bluescript multiple cl oning site, resolving on a 0.8% agarose gel, excising the 13.4 kb band, and gel-purifying the fragment. (In this case the linearized vector was not treated with CIAP.) Conc urrent with this, a PGK promoter-driven thymidine kinase (TK) gene was isolated from plasmid 155. This involved cutting plasmid 155 with SalI, resolving on a 0.8% gel, excising the 2.9 kb band, and gel-purifying the TK fragment. The 16.6 kb v ector fragment and the 2.9 kb TK fragment were ligated together to produce plasmid 396 containing the 3.9 kb 5-prime arm in the EcoRI site, the 1.8 kb neo fragment in the EcoR V site, the 8.0 kb 3-prime arm in the SalI site, and the 2.9 kb TK gene in the XhoI si te of Bluescript. This 19.5 kb plasmid was linearized with NotI, phenol-c hloroform extracted, and ethano l precipitated prior to ES cell electroporation.

PAGE 62

52 ES Cell Targeting Targeting the 12 kb Deletion To produce the 12 kb deletion, CJ.7 ES ce lls were electroporated with plasmid 391 that had been linearized by NotI digest ion. A total of 307 neomycin-resistant colonies were individually cultured and then preserved by storage at –80oC. To screen for recombinants, DNA isolated from individua l clones was digested with SacI, resolved on 0.8% agarose gels, transferred to Hybond me mbranes. The membranes were probed with a 980 bp PCR fragment (probe Q) that recognizes an 18.0 kb wild-type band and a 12.4 kb recombinant band. A single recombin ant clone, number 178, was identified and expanded. (The targeting frequency was 1 in 307.) Clone 178 was electroporated with a plasmid encoding the Cre recombinase in order to remove the neomycin resistance cassette. Clones that had lost the neom ycin cassette included 178-8, 178-20, 178-23, and 178-31. These clones were expanded and confirmed by Southern blot. Targeting the 7 kb Deletion To produce the 7 kb deletion, CJ.7 ES cells were electroporated with plasmid 396 that had been linearized by NotI digestion. A total of 280 neomycin-resistant colonies were individually cultured and th en preserved by storage at –80 oC. To screen for recombinants, DNA isolated from individual cl ones was digested with SacI, resolved on 0.8% agarose gels, transferred to Hybond memb ranes. The membranes were probed with a 980 bp PCR fragment (probe Q) that r ecognizes an 18.0 kb w ild-type band and a 12.4 kb recombinant band. Two recombinant clone, numbers 29 and 89, were identified and expanded. (The targeting frequency was 1 in 140.) Clones 29 and 89 were electroporated with a plasmid encoding the Cre recombinase in order to remove the

PAGE 63

53 neomycin resistance cassette. Clones that had lost the neomycin cassette included 29-5, 29-7, 89-3, and 89-14. These clones were e xpanded and confirmed by Southern blot. Establishment of Breeding Colonies Establishing the 12 kb Deletion Line In order to create a mouse line for the 12 kb deletion, Michael Elmore, a former technician in our lab, injected cells of cl one 178-20 into wild-type C57BL/6J blastocysts to produce two chimeric males. One of thes e males transmitted the 12 kb deletion allele when mated to a wild-type C57BL/6J female, and a breeding colony of 12 kb deletion mice was established. At weaning, mice we re genotyped by Southern blot. Tail DNA was digested with EcoRV, and the membra ne was probed with a 980 bp PCR fragment (Probe Q) that recognizes a 12.0 kb wild-t ype band and a 9.8 kb mutant band. Probe Q was amplified using the primers “W55-Q Forward” (5’-TTTAAGAAAGAGTATTGGACAGAGA G-3’) and “W56-Q Reverse” (5’-TCGTTAGTTCATCACAGTGCTTGCA-3’). In some cases, the genotype of brain DNA samples to be used in experiments were confirmed by Southern bl ot. (Figure 3-2) Mice were also routinely ge notyped by PCR using the primers “W57-29 LoxP Forward” (5’-CTACATCCAAATAGAGTGC TAATATC-3’), “W58-4.8BglII LoxP Reverse” (5’-ATAGAAAATCAGAGTCATAGTAGGG-3’) to produce a 349 bp product for mutant mice. (With this primer set th ere was no wild-type co ntrol product.) PCR genotyping was improved by using th e primers “W59-AS-IC 7 Forward” (5’-ACTATCAGAAAATGGGGTTCAGAG-3’), “W60-AS-IC 12 Reverse” (5’-TGGAGGTTAGAAATGATATAAGTGG-3’), and “W61-AS-IC 12 WT Control” (5’-TATACAGTCAGGTACTATACACCG-3’) to produce a 401 bp wild-type band and a 189 bp mutant band.

PAGE 64

54 Establishing the 7 kb deletion line In order to create a mouse line for the 7 kb deletion, Michael Elmore injected cells of clone 89-3 into wild-type C 57BL/6J blastocysts to produce two chimeric males. One of these males transmitted the 7 kb deletion al lele when mated to a wild-type C57BL/6J female, and a breeding colony of 7 kb deleti on mice was established. At weaning, mice were genotyped by Southern blot. Tail DNA was digested with EcoRV, and the membrane was probed with a 980 bp PCR frag ment (Probe Q) that recognizes a 12.0 kb wild-type band and a 14.4 kb mutant band. In some cases, the genotype of brain DNA samples to be used in experiments were conf irmed by Southern blot. (Figure 3-2) Mice were also routinely genotyped by PCR us ing the primers “W57-29 LoxP Forward” (5’-CTACATCCAATAGAGTGCTAATATC-3”) and “W62-29 LoxP Reverse” (5’-AACAAGAGAGAGGTTTTCTGTTTAC-3’) to produce a 432 bp product for mutant mice. (With this primer set there was no wild-type control product.) PCR genotyping was improved by using the primers “W57-29 LoxP Forward” (5’-CTACATCCAATAGAGTGCTAATATC-3’), “W63-AS-IC 7c Reverse” (5’-GACAGTCTTACATTTCTTTCTTC ATC-3’), and “W26” (5’-TCACTTCTTGTGATATTCTC TCCC-3’) to produce a 282 bp wild-type band and a 363 bp mutant band.

PAGE 65

55 Figure 3-1. Deletions upstream of Snrpn to locate the murine AS-IC. In an effort to locate the murine AS-IC, deletions were created upstream of Snrpn . If mice inheriting a deletion on the maternal ch romosome showed signs of an AS imprinting mutation, this would provide evidence that the AS-IC was located within that deletion. Homologous recombination in ES cells was used to engineer 3 nested deletions. Dele tion A was 7 kb, extending from –30 kb to – 37 kb relative to Snrpn exon 1. Deletion B was 12 kb, extending from –25 kb to –37 kb. Deletion C was 58 kb, extending from –20 kb to –78 kb.

PAGE 66

56 Figure 3-2. Southern blot genotyping of the 12 kb and 7 kb deletion mice. Genomic DNA was digested with EcoRV, resolved on a 0.8% agarose gel, Southern blotted, and hybridized with a radiolab eled 980 bp 5-prime flanking probe. The wild-type samples in lanes 1, 3, 5, and 7 display only the endogenous 12.0 kb band. Samples from mice inheriting the 12 kb deletion shown in lanes 2 and 4 display an additional 9.8 kb mu tant band. Samples from mice inheriting the 7 kb deletion shown in lanes 6 and 8 di splay an additional 14.4 kb mutant band.

PAGE 67

57 CHAPTER 4 RESULTS Introduction In an effort to locate the murine AS-IC, three different deletions were engineered upstream of Snrpn . (Figure 4.1) Although attempts to establish a mouse line for the 58 kb deletion were unsuccessful, breeding colonies of mice were established for both the 12 kb and the 7 kb deletions. This chapter wi ll present the results from a variety of assays used to test the effects of the 12 kb and 7 kb deletions on DNA methylation and gene expression patterns within the PWS/ AS imprinted domain on mouse chromosome 7. Overview of Analysis In order to determine if the murine AS-IC was located within either the 12 kb or the 7 kb deletion, mice harboring each deletion were tested for signs of an imprinting defect. (Figure 4-2) Given that imprinting mutations in human Angelman syndrome patients cause a loss of DNA methylation and inappropriate activation of Snrpn and perhaps several other PWS-candidate genes on the maternal chromosome, it seemed obvious to look for these same changes as indicators of an Angelman imprinting mutation in mice [119, 138]. Also, considering that AS-IC mu tations in humans are benign when passed through a male germline but resu lt in AS upon maternal transm ission, we focused most of our analysis on mice inheriting the 12 kb and 7 kb deletion on the maternal chromosome. To evaluate if the 12 kb or 7 kb deletion produced an Angelman imprinting phenotype, the DNA methylation and gene expression prof iles of several imprinted genes were examined.

PAGE 68

58 Overview of DNA Methylation and Gene Expression Studies First, DNA methylation at thr ee paternally-expressed genes, Mkrn3 , Ndn , and Snrpn , was investigated by Southern blot to determine if maternal inheritance of the 12 kb or 7 kb deletion caused a loss of methyla tion at these genes. This was followed by the higher-resolution sodium bisulfite genomic sequencing analysis of the Snrpn promoter and intron 8 region. An improvement over these initial at tempts at bisulfite sequencing came by crossing females carrying the 12 kb or 7 kb deletion with males of the castaneus chromosome 7 line. This allowed the parental alleles to be distinguished by sequence polymorphisms and provided a much more straightforward and robust analysis of DNA methylation levels at the Snrpn promoter and intron 1 regions as well as the promoter of Mkrn3 . Second, the expression of several patern al-specific genes was examined to determine if maternal transmission of either deletion would cause inappropriate activation of these genes on the maternally inherited ch romosome harboring the deletion. This was accomplished by breeding females carrying the 12 kb or 7 kb deletion with males of the 35 kb PWS-IC deletion line. The advantage of employing the PWS-IC deletion mice was that inheritance of this dele tion on the paternal allele produc es a loss of paternal-specific gene expression [132]. We reasoned that th e reduction of expression from the paternal allele in these mice would provide a low “b ackground” of transcription over which it would be easy to detect inappropriate activation of the maternal allele. In addition to testing for inappropriate act ivation of the patern al-specific genes on the maternal chromosome, expression of the Ube3a gene was investigated as another means of assessing an impr inting mutation. Although Ube3a is expressed from both alleles in most tissues, it is expressed preferen tial from the maternal allele in certain cell

PAGE 69

59 types the brain. Therefore, an Angelman imprinting mutation should cause a loss of Ube3a expression from specific cells, such as hippocampal and Purkinje neurons, in which the paternal copy of Ube3a is silenced [124]. Instead of attempting to document a complete loss of Ube3a expression in specific cell types, we examined whole brain RNA from mice inheriting either the 12 kb or 7 kb deletion maternally for signs of a decrease in Ube3a expression from the maternal allele. Cursory Examination of Phenotype In should be noted that the 12 kb or 7 kb deletion mice did not show obvious signs of seizures or obesity regardless of whethe r the deletion was inherited on the maternal or paternal allele. The mice have not yet been subjected to more detailed analysis designed to reveal subtle phenotypes such as ataxia or learning defects that ha ve been detected in other mouse models for Angelman s yndrome [131, 139, 140]. Mice homozygous for either the 12 kb or the 7 kb deletion were viable and fertile but have not been investigated for changes in DNA methylation or gene expression. Genetic Background of Mice All the mice investigated in these expe riments were of a mixed 129S1/Sv and C57BL/6J genetic background. The 129S1/Sv genetic contribution was provided by the CJ.7 ES cells. Chimeras were derived from the introduction of targeted CJ.7 ES cells into host C57BL/6J blastocysts. Male chim eras were mated with wild-type C57BL/6J females to produce F1 mice that were bred to establish the 12 kb and 7 kb deletion lines. These mutations were continuously back crossed onto the C57BL/6J background for several generations. The brain DNA samples used in the Southern blot and initial sodium bisulfite examinations of DNA methylation were derived from adult animals of the N2 generation. The brain DNA samples used in the later allele-speci fic sodium bisulfite

PAGE 70

60 assays which employed the B6.Cast.c7 pol ymorphisms were derived from newborn animals. These pups were produced by mati ng females that carried the 12 kb or 7 kb deletion at either the N5 or N6 generation with males of th e inbred congenic B6.Cast.c7 line. The brain RNA samples used in Northe rn blot examinations of gene expression were derived from newborn animals. Thes e pups were produced by mating females that carried the 12 kb or 7 kb deletion at either the N5 or N6 generation with males that carried the 35 kb PWS-IC deletion on an inbred C57BL/6J bac kground. The brain RNA samples used in RT-PCR analysis of Ube3a expression were derived from newborn animals. These pups were produced by mati ng females that carried the 12 kb or 7 kb deletion at either the N5 or N6 generation with males of th e inbred congenic B6.Cast.c7 line. Pups were also produced from the reci procal matings of B6.Cast.c7 females with males that carried either the 12 kb or 7 kb dele tion. All of the wild -type C57BL/6J mice used in this work were obtai ned from The Jackson Laboratory. DNA Methylation Studies Southern blot analysis of DNA methylation The first step in evaluating the 12 kb and 7 kb deletion mice for an Angelman imprinting defect was to look for a loss of DN A methylation at three paternally expressed genes that are normally methylated specifically on the silenced maternal allele. Several different methods were employed to assess me thylation levels. The first method used methylation-sensitive restric tion enzymes to examine DNA met hylation by Southern blot. High molecular weight DNA was isolated from the brains of 7 kb and 12 kb deletion mice, and this DNA was digested with methyl ation-sensitive restri ction enzymes. A methylation-sensitive restriction enzyme recogni zes a restriction site that contains one or more CpG dinucleotides. The enzyme can cu t the DNA only if the cytosines within the

PAGE 71

61 restriction site are unmethylated. The dige sted DNA was then Southern blotted, and the membrane was hybridized with a radiolabeled probe that could r ecognize both the uncut (methylated) fragment and the cut (unmet hylated) fragment. (Figure 4-3) This analysis was applied to three genes, Mkrn3 , Ndn , and Snrpn , that normally lack methylation on the expressed paternally-i nherited allele but th at are methylated on the silent maternally-inherited allele at the CpG residues examined in this Southern blot assay. The details of which enzymes and pr obes were employed in the analysis of each gene are listed below. These methods were previously described in published reports [132-134]. The methylation level at Mkrn3 was examined by digesting the DNA with BamHI and NotI and probing with a 554 bp PCR fragment amplified from the 3-prime end of the gene using the primers “W39” (5 ’-GTTCTTCCTTCTCTGATGAC-3’) and “W40” (5’-CACAAGTTAACAAGTGCAC-3’). The unme thylated fraction of DNA that was cut by NotI produced an 18 kb band, and the met hylated fraction that was not cut by NotI produced a 20 kb band. The methylation level at Ndn was examined by digesting the DNA with BglII and SacII and probing with a 1.5 kb BamHI-EcoRI fragment from the 5-prime end of the gene. The unmethylated fr action of DNA that was cut by SacII produced a 3 kb band, and the methylated fraction that wa s not cut by SacII produced a 5 kb band. The methylation level at Snrpn was examined by digesting the DNA with TaqI and SacII and probing with a 234 bp PCR fragment amplified from exon 1 of Snrpn using the primers “W41” (5’CCGCAGT AGGAATGCTC -3’) and “W42” (5’-GCGTTGCAAATCACTCCTCAG-3’). The unm ethylated fraction of DNA that was

PAGE 72

62 cut by SacII produced a 1.8 kb band, and the me thylated fraction that was not cut by SacII produced a 5.6 kb band. Although this method did not provide a means of assessing the methylation of each allele separately, any overall gain or lo ss of methylation at each locus could be determined by comparing the intensity of th e larger band representing the methylated fraction to the intensity of the smaller ba nd representing the unme thylated fraction. A potential Angelman imprinting mutation that might result in the demethylation of the maternal allele could therefore be inferred by a reduction in the intensity of the larger band that was not cut by the methylati on-sensitive restriction enzyme and a corresponding increase in the intensity of th e smaller band that was cut by the enzyme. The results of this analysis were not quantified, but examination of Mkrn3 , Ndn , and Snrpn did not reveal an obvious loss of methyl ation at these imprinted genes in the brains of mice inheriting the 12 kb or 7 kb dele tion maternally as compared to their wildtype siblings or to mice inhe riting the deletions paternall y. Although these results were not definitive, they suggeste d that the murine AS-IC was not located within the 12 kb deleted region. It should be noted, however, th at this technique might not be sensitive enough to detect a subtle cha nge in DNA methylati on reflective of a partial or mosaic imprinting defect. Another caveat to testing DNA methylation levels by Southern blot is that in this case only one restriction site wa s examined for each gene. To obtain a higher resolution picture of DNA methyl ation levels, it was necessary to turn to the sodium bisulfite genomic sequencing method. Sodium bisulfite analysis of DNA me thylation without polymorphisms A major advantage of looking at DNA met hylation with sodium bisulfite genomic sequencing is that it provides information re garding the level of me thylation at numerous

PAGE 73

63 CpG dinucleotides within a region. This me thod, hereafter referred to simply as the “bisulfite” method, involves treating genomic DNA with sodium bisulfite [141]. This chemical converts unmethylated cytosines to ur acils. Methylated cytosines, however, are protected from the conversion. Following the so dium bisulfite treatment, the region to be analyzed is PCR amplified, cloned, and seque nced. Unmethylated cytosines that were converted appear as thymines in the seque nce, and methylated cytosines that were protected from the conversion remain as cy tosines. The sequence for each clone represents the methylation pr ofile of a single molecule of DNA, and by comparing the methylation status of each CpG dinucleotide among nume rous sequenced clones it is possible to measure the overall level of methylation at that locus. The methylation levels of two different regions of the Snrpn gene were examined. The first region investigated was the Snrpn promoter and exon 1 region, which is embedded in a CpG island that displays stri ct maternal-specific methylation in both human and mouse [117, 142-144]. Th e second region investigated was Snrpn intron 8. There is some controversy as to whether the body of the Snrpn gene displays allelespecific methylation in mouse [143, 144]. Despite this un certainty, however, the region within Snrpn seemed to warrant further attention because a similar group of several CpG dinucleotides within human SNRPN intron 7 are methylated exclusively on the active paternal allele [33]. For the Snrpn promoter region, primers “W3” (5’-AATTATATTTATTATTTTAGATTG ATAGTGAT-3’) and “W4” (5’-TTTACAAATCACTCCTCAAAACCAA-3’) were used to amplify a 289 bp fragment containing 14 CpG di nucleotides. In a previ ous publication, the region

PAGE 74

64 amplified by these primers was found to be enti rely methylated in eggs and completely unmethylated in spermatozoa. [134]. For the Snrpn intron 8 region, primers “W13” (5’-GTGTAAGTTTGGTAAAATAATAT-3’) and “W14” (5’-AATTAAAAAAATAAACCAACAATAACA-3’) were used to amplify a 604 bp product containing 5 CpG dinucleotides. These primers were originally used to examine primordial germ cells at 11.5 dpc, 12.5 dpc, a nd 13.5 dpc for differences in methylation levels, and among 20 clones amplified from 11.5 dpc samples the average level of methylation at each of the five CpG sites was found to be about 40 percent as compared to a near absence of methylation in 10 clones amplified from 13.5 dpc samples. [145]. For both regions of Snrpn , brain DNA was used as the starting material for the bisulfite conversion. In each case the brain DNA samples were isolated from mice wild-type mice, mice inheriti ng the 12 kb deletion materna lly, and mice inheriting the 7 kb deletion maternally. For the primer sets employed in this anal ysis there was no way to identify which clones were amplified from the maternal allele and which clones were amplified from the paternal allele. The interpretation of the results, therefore, was based on comparing the number of clones that showed very low levels of methylation to the number of clones that showed very high levels of methylation. This ratio was used as an assessment of the overall level of methylation at that locus for a given sample. It was expected that for the Snrpn promoter region roughly half of the clones from wild-type samples would show an “unmethylated” profile representing the patern al allele and half would show a “methylated” profile representi ng the maternal allele. If the ratio of unmethylated to methylated clones was mu ch higher in mice inheriting the 12 kb or 7 kb

PAGE 75

65 deletion on the maternal allele, this would be interpreted as an indication of an Angelman syndrome imprinting mutation. Initial bisulfite sequencing analysis of the Snrpn promoter CpG island region produced information for 14 clones from th e wild-type sample, 19 clones from the maternally inherited 12 kb deletion sample, and 20 clones from the maternally inherited 7 kb deletion sample. For the wild-type sample, 29% of the clones were determined to be “unmethylated” and 71% were determined to be “methylated.” (Figure 4-4) For the 12 kb deletion sample, 79% of the clones were determined to be “unmethylated” and 21% were determined to be “methylated.” (Figure 4-5) For the 7 kb deletion sample, 55% of the clones were determined to be “unmethylated” and 45% of the clones were determined to be methylated. (Figure 4-6) For the wild-type sample, it was surprising to see that the ratio was heavily skewed in favor of methylated clones. Given that for the 12 kb deletion sample the ratio was even more skewed in the opposite direction towards unmethylated clones, this coul d be interpreted as indicating a loss of methylation due to an imprinting defect produced by the 12 kb de letion. One major factor, however, argues for caution in interpreting these results. Th e fact that the wild-type sample did not produce anything close to an even number of unmethylated and methylated clones calls into doubt the initial pr esumptions of the analysis descri bed in the preceding paragraph. One possibility is that some of the paternal alleles amplified from the wild-type sample were highly methylated. This is doubtful in light of previous studies that clearly demonstrated that the Snrpn promoter and exon 1 region shows maternal-specific methylation [143, 144]. It seems much more likely that the skewed results for the wild-type sample, as well as the 12 kb dele tion sample, are due to a PCR amplification

PAGE 76

66 bias. Indeed, all of the closed examined in this analysis were derived from a single PCR-amplification. One means of combating the potential for PCR amplification bias would be to sequence a large number of clone s from several different PCR amplification reactions. So although the maternal 12 kb de letion sample did indeed produce a much higher ratio of unmethylated to methylated clones as compared to the wild-type sample, these data might not necessarily indicate an imprinting defect in the 12 kb deletion mice. Bisulfite sequencing analysis of Snrpn intron 8 produced information for 22 clones from the wild-type sample, 11 clones from the maternally inherited 12 kb deletion sample, and 17 clones from the maternally inherited 7 kb deletion sample. For the wildtype sample, 18% of the clones were methyl ated at 3 out of 5 CpG dinucleotides, 41% were methylated at 4 out of 5 CpGs, and 41% were methylated at all 5 CpGs. (Figure 47) For the maternal 12 kb deletion sample, 10% of the clones were methylated at 3 out of 5 CpG dinucleotides, 45% were methylated at 4 out of 5 CpGs, and 45% were methylated at all 5 CpGs. (Figure 4-8) For the mate rnal 7 kb deletion sample, 18% of the clones were methylated at 3 out of 5 CpG dinucleot ides, 29% were methyl ated at 4 out of 5 CpGs, and 53% were methylated at all 5 CpGs. (Figure 4-9) These data did not reveal any significant differences in methylation levels between the wild-type, maternally inherited 12 kb deletion, and maternally inhe rited 7 kb deletion samples and provided no indication of an imprinting defect in the 12 kb and 7 kb deletion mice. Since there is no clear evidence that this region possesses alle le-specific methylation anyway, this cursory analysis of the 5 CpG sites within Snrpn intron 8 fell short of pr oviding a clear-cut test for an imprinting defect.

PAGE 77

67 Sodium bisulfite analysis of DNA methylation with polymorphisms A considerable improvement over these in itial attempts at bisulfite sequencing analysis of Snrpn came by crossing females carrying the 12 kb or 7 kb deletion with males of the castaneus chromosome 7 (B6.Cast.c7) lin e [146]. (Figure 4-10) The B6.Cast.c7 strain is a congenic line that contains a region of Mus musculus castaneus chromosome 7 on a Mus musculus domesticus C57BL/6J background. This cross allowed the parental a lleles to be distinguished by se quence polymorphisms and provided a much more straightforward and robust anal ysis of DNA methyla tion levels at the Snrpn promoter and intron 1 regions . For both regions of Snrpn , brain DNA was used as the starting material for the bisulfite conversi on. In each case the brain DNA samples were isolated from wild-type mice, mice inheri ting the 12 kb deletion maternally, and mice inheriting the 7 kb dele tion maternally. The products of PCR amplification react ions from two independent bisulfite conversion reactions were separately cloned and sequenced. The goal of performing two rounds of analysis on each region of Snrpn was to ensure that the results from the first round could be repeated and th ereby provide greater confiden ce in the final analysis of the data. The results presented here are th e combination of information from both rounds of bisulfite conversion, amplificatio n, and cloning for each region of Snrpn . In addition to the promoter and intron 1 regions of Snrpn , the promoter of Mkrn3 was also subjected to bisulfite sequencing analysis. For Mkrn3 , however, the clones that were sequenced were derived from a single bisulfite convers ion and PCR amplification from brain DNA, and so this information represents only an initial examination for this gene. Primer sequences and information rega rding polymorphisms were generously provided by Mrs. Christine Mione Kief er and Dr. Thomas Yang. For the Snrpn

PAGE 78

68 promoter/exon1 region, primers “W18–TY657” (5’-GTAGTAGGAATGTTTAAGTATTTTTTTTGG-3’) and “W19-TY658” (5’-CCAATTCTCAAAAATAAAAATATCTAAATT-3’) were used to amplify a 364 bp fragment containing 14 CpG dinucleotides. The polymorphism for this primer set is located at base 69 of this pr oduct (closer to the W18 primer), with a “T” at this position identifying the castaneus allele and a “G” identifying the domesticus allele. The 5-prime end of the W18 primer is located at base –175 relative to Snrpn exon 1, and the 3-prime end of the W19 primer is located at base +189. For the Snrpn intron 1 region, primers “W16-TY655” (5’-AATTTAGATATTTTTATTTTTGAGAATTGG-3’) and “W17-TY656” (5’-TCTACAAATCCCTACAACAACAACAA-3’) were used to amplify a 365 bp product containing 13 CpG dinucleotides. The polymorphism for this primer set is located at bases 165 and 166 of the product (clo ser to the W16 primer), with “AA” at these positions identifying the castaneus allele and “CG” identifying the domesticus allele. The 5-prime end of the W16 prim er is located at base +160 relative to Snrpn exon 1, and the 3-prime end of the W17 primer is located at base +524. For the Mkrn3 promoter, primers “W33” (5’-GATTTAAAGAATTTTGTTAGATTTATAAGT-3’) and “W28-TY822” (5’-ACCTCAATAAAAACTATAAACTCTTCCAT -3 ’) were used to amplify a 625 bp product containing 21 CpG dinucleotides. The polymorphism for this primer set is located at bases 209 of the produc t (closer to the W33 primer), with a “G” at this position indentifying the castaneus allele and a “T” identifying the domesticus allele.

PAGE 79

69 The 5-prime end of the W33 primer is located at base –483 relative to the transcriptional start of Mkrn3 , and the 3-prime end of the W28 primer is located at base +145. Bisulfite sequencing analysis of the Snrpn promoter/exon 1 region produced information for 12 clones representing the maternal allele from the wild-type sample and 31 clones representing the paternal allele from the wild-type sample. (Figure 4-11) For the 12 kb deletion sample, 16 clones represen ting the maternal allele and 25 clones representing the paternal allele were seque nced. (Figure 4-12) For the 7 kb deletion sample, 8 clones representing the maternal alle le and 30 clones representing the paternal allele were sequenced. (Figure 4-13) In summary, about 90% of the CpG dinucleotides in clones representing the maternal alleles were methylated in all three samples. (Figure 4-14) In contrast, clones representing the paternal alleles were almost entirely unmethylated for all three samples. These da ta did not reveal any demethylation of the maternal alleles in the maternally inherite d 12 kb deletion and maternally inherited 7 kb deletion samples. These results clearly indicate the absence of an imprinting defect in the 12 kb and 7 kb deletion mice and suggest that the murine AS-IC resides outside of the 12 kb deleted region. Bisulfite sequencing analysis of the Snrpn intron 1 region produced information for 3 clones representing the maternal allele from the wild-type sample and 30 clones representing the paternal alle le from the wild-type sample. (Figure 4-15) For the 12 kb deletion sample, 1 clone representing the ma ternal allele and 28 clones representing the paternal allele were sequen ced. (Figure 4-16) For the 7 kb deletion sample, 2 clones representing the maternal al lele and 27 clones representing the paternal allele were sequenced. (Figure 4-27) In summary, about 90% of the CpG dinucleotides in clones

PAGE 80

70 representing the maternal alleles were methylat ed in all three samples. (Figure 4-18) In contrast, clones representing th e paternal alleles were almost entirely unmethylated for all three samples. Similar to what was seen for the Snrpn promoter/exon 1 region, these data did not reveal any demethylation of the matern al alleles in the mate rnally inherited 12 kb deletion and maternally inherited 7 kb deleti on samples. Considering the small number of clones derived from the maternal alleles that were sequenced for the Snrpn intron 1 region, this analysis should be viewed as pr eliminary and will need to be repeated in order to have absolute confidence in the data . The results are, however, in line with the analysis of the Snrpn promoter/exon 1 region and support the conclusion th at there is no imprinting defect in the 12 kb and 7 kb deletion mice. Bisulfite sequencing analysis of the Mkrn3 promoter produced information for 5 clones representing the maternal allele from the wild-type sample and 5 clones representing the paternal alle le from the wild-type sample. (Figure 4-19) For the 12 kb deletion sample, 2 clones representing the ma ternal allele and 8 clones representing the paternal allele were sequen ced. (Figure 4-20) For the 7 kb deletion sample, 11 clones representing the maternal al lele and 18 clones representing the paternal allele were sequenced. (Figure 4-21) The average le vel of methylation for all of the CpG dinucleotides summed together for the clones derived from the maternal alleles was 43% in the wild-type sample, 67% in the 12 kb deletion sample, and 58% in the 7 kb deletion sample. (Figure 4-22) In cont rast to what was seen for clones derived from the maternal alleles, clones representing the paternal alle les were almost entirely unmethylated for all three samples. Although the extent of me thylation found on the maternal alleles for Mkrn3 was across the board lower much lower th an the roughly 90% methylation seen at

PAGE 81

71 the Snrpn promoter or intron 1, these data did not reveal any demethylation of the maternal alleles in the maternally inherite d 12 kb deletion and maternally inherited 7 kb deletion samples as compared to the wild-typ e sample. Because only a small number of clones derived from the maternal alleles were sequence for the Mkrn3 promoter, this analysis must be repeated before strong conc lusions can be made regarding the level of methylation at this gene. N onetheless, these preliminary results are in agreement with what was seen for Snrpn and do not indicate an imprinting defect in the 12 kb and 7 kb deletion mice. Gene Expression Studies In addition to looking for changes in DNA methylation as an indicator of an imprinting mutation, the expression of severa l paternal-specific genes was examined to determine if maternal transmission of either deletion would cause inappropriate activation of these genes on the maternally inherited ch romosome harboring the deletion. This was accomplished by breeding females carrying the 12 kb or 7 kb deletion with males of the 35 kb PWS-IC deletion line. (Figure 4-23) The advantage of employing the PWS-IC deletion mice was that paternal inheritanc e of this deletion produces a loss of paternal-specific gene expre ssion [132]. We reasoned that the reduction of expression from the paternal allele in these mice woul d provide a low “background” of transcription over which it would be easy to detect ina ppropriate activation of the maternal allele. The RNAzol B method was used to isolat e total RNA from the brains of newborn mice with the following six genotypes: (1) wild -type, (2) a 12 kb deletion on the maternal allele, (3) a 7 kb deletion on the maternal al lele, (4) a PWS-IC deletion on the paternal allele, (5) a 12 kb deletion on th e maternal allele plus a PW S-IC deletion on the paternal allele, and (6) a 7 kb deletion on the matern al allele plus a PWS-IC deletion on the

PAGE 82

72 paternal allele. These samples were subjecte d to Northern blot analysis to examine the expression of several paternal-specific genes. For Ndn and Snrpn Northern blot analysis, 10 ug of each sample was loaded onto a 1.0% agarose gel containing formaldehyde and resolved at 70 volts for 2 hours. The RNA was transferred to a Hybond membrane and sequentially probed with three fragments. The expression level of Ndn was examined by probing with the “Ndn 1060” PCR fragment produced usi ng the primers “W43-Ndn 9F” (5’-GTATCCCAAATCCACAGTGC-3’) and “W44-Ndn 10R” (5’-CTTCCTGTGCCAGTTGAAGT-3’). The expression level of Snrpn was examined by probing with a 480 bp ClaI-PstI fragment isolated from a cDNA clone spanning the 5-prime end of the gene. The expression of Actin was examined as a loading control by probing with a 539 bp PCR fragment produced using the primers “W45-Actin Forward” (5’-GTGGGCCGCTCTAGGCACCAA-3 ’) and “W46-Actin Reverse” (5’-CTCTTTGATGTCACGCACGATTTC-3’). For snoRNA Northern blot analysis, 10 ug of each sample was loaded in the order listed above and resolved on an 8.0% acrylam ide gel containing 7 molar urea. The RNA was transferred to a Hybond membrane and sequentially probed with four oligonucleotides. (The MBII-85 and MBII52 oligos were used in the same hybridization.) The sequences of the oligonucleotides used as probes were as follows: “W47-MBII-13” (5’-GAGAGAGAGAGA GAGAGA-3’), “W48-MBII-85” (5’-TTCCGATGAGAGTGGCGGTACAGA-3’), “W49-MBII-52” (5’-CCTCAGCGTAATCCTA TTGAGCATGAA-3’), and “W50-5.8S rRNA” (5’-TCCTGCAATTCACATTAATTCTCGCAGCTAGC-3’).

PAGE 83

73 Northern blot analysis of Ndn and Snrpn showed that these transcripts were expressed at approximately equal levels in the wild-type, mate rnal 12 kb deletion, and maternal 7 kb deletion brain RNA samples. (Figure 4-24) As expected, Ndn and Snrpn were not detected in the pa ternal PWS-IC sample. These transcripts were also not detected in the “maternal 12 kb deletion plus the paternal PWS-IC deletion” or the “maternal 7 kb deletion plus the patern al PWS-IC deletion” samples. Northern blot analysis of the s noRNA genes MBII-13, MBII-85, and MBII-52 showed that these transcripts were expre ssed at approximately equal levels in the wild-type, maternal 12 kb de letion, and maternal 7 kb dele tion brain RNA samples. (Figure 4-25) A low level of expression for MBII-85 and MBII-52 was detected in the paternal PWS-IC sample. The expression of these paternal-specific genes was much lower in the paternal PWS-IC sample than in the wild-type sample, and this was consistent with prior studies of low-le vel “leaky” expression for some of the paternal-specific genes in the brains of mice inheriting the PWS-IC deletion on the paternal allele [147]. A lo w level of the MBII-85 and MBII-52 transcripts was also detected in the “maternal 12 kb deletion plus the paternal PWS-IC deletion” and the “maternal 7 kb deletion plus the paternal PWS -IC deletion” samples, but in neither case did the expression of these transcripts appear to be higher than in the paternal PWS-IC deletion. In summary, Northern blot analys is for the paternal-specific genes Ndn , Snrpn , MBII-13, MBII-85, and MBII-52 did not reveal any evidence for activation of these genes on a maternal chromosome carrying e ither the 12 kb or the 7 kb deletion. These

PAGE 84

74 results did not indicate an imprinting def ect in the 12 kb or 7 kb deletion mice and strongly suggested that the murine AS-IC is located outside of the 12 kb deleted region. In addition to testing for inappropriate act ivation of the patern al-specific genes on the maternal chromosome, expression of the Ube3a gene was investigated as another means of assessing an impr inting mutation. Although Ube3a is expressed from both alleles in most tissues, it is expressed preferen tial from the maternal allele in certain cell types the brain. Therefore, an Angelman imprinting mutation should cause a loss of Ube3a expression from cells, such as hippocam pal and Purkinje neurons, in which the paternal copy of Ube3a is silenced [124]. In order to examine the expression of Ube3a in an allele-specific manner, we crossed females carrying the 12 kb or 7 kb deletion with males of the castaneus chromosome 7 (B6.Cast.c7) line. (Figure 4-10 ) The reciprocal crosses of B6.Cast.c7 females bred with 12 kb or 7 kb deletion ma les were also performed. As mentioned previously for the sodium bisulfite an alysis, these crosse s provided sequence polymorphisms by which the maternal and paternal alleles could be dis tinguished. In this case, a polymorphism identified in Ube3a exon 5 created a Tsp509I re striction site in the Mus musculus domesticus allele that was not present in the Mus musculus castaneus allele. To measure the expression of Ube3a derived from each parental allele, we employed an RT-PCR assay developed by Dr. Stormy Chamberlain, a former graduate student in the lab [123]. Whole brain RNA was isolated from hybrid offspring produced by the matings described above, and the R NA was converted to cDNA by treating with reverse transcriptase in the presence of random primers. Amplification of a 246 bp PCR

PAGE 85

75 product was performed using the primers “W51-Ube3a exon 5F” (5’-CACATATGATGAAGCTACGA-3’) and “W52-Ube3a exon 6R” (5’-CACACTCCCTTCATATTCC-3’). The PCR products were digested with Tsp509I overnight, resolved on a 4.8% agarose gel, Sout hern blotted, and then hybridized with a radiolabeled 246 bp PCR fragment gene rated using the “Ube3a exon 5F” and “Ube3a exon 6R” primers. (Figure 4-26) The results of this analys is did not reveal any decrease in Ube3a expression from the maternal allele in the 12 kb or 7 kb deletion samples as compared to the wild-type sample. Consistent with the data from studies of DNA methylation and gene expr ession levels for several paternal-specific genes, the investigation of Ube3a expression levels did not indicate an Angelman imprinting defect in the 12 kb or 7 kb deletion mice. Summary Toward the goal of identifying the loca tion of the murine AS-IC, three deletions were made upstream of the Snrpn gene. The first deletion was 12 kb in size, extending from –37 kb to –25 kb relative to Snrpn exon 1. The second deletion was 7 kb in size, extending from –37 kb to –20 kb. The third dele tion was 58 kb in size, extending from – 78 kb to –20 kb. (Figure 4-1) The two smaller deletions were each made in a single step, but the 58 kb deletion required two sequential loxP targeting events followed by Cremediated recombination. ES cells harboring each of the three dele tions were isolated, however attempts to produce a mouse line for the 58 kb deletion were unsuccessful. Breeding colonies were established for the 12 kb and 7 kb deletions, and analysis of DNA methylation and gene expressi on patterns in these mice did not reveal evidence for an AS imprinting mutation. These results will be discussed further in the next chapter.

PAGE 86

76 Figure 4-1. Deletions upstream of Snrpn to locate the murine AS-IC. In an effort to locate the murine AS-IC, deletions were created upstream of Snrpn . If mice inheriting a deletion on the maternal ch romosome showed signs of an AS imprinting mutation, this would provide evidence that the AS-IC was located within that deletion. Homologous recombination in ES cells was used to engineer 3 nested deletions. Dele tion A was 7 kb, extending from –30 kb to – 37 kb relative to Snrpn exon 1. Deletion B was 12 kb, extending from –25 kb to –37 kb. Deletion C was 58 kb, extending from –20 kb to –78 kb.

PAGE 87

77 Figure 4-2. Effects of imprinting mutations. The top panel shows the gene expression and DNA methylation profile found in a normal individual, as described in Figure 1-3. The middle panel shows the results of a PWS imprinting mutation, and the bottom panel shows the results of an AS imprinting mutation. In a PWS imprinting mutati on, the “paternal-specific” genes are inappropriately methylated and silenced. Loss of these gene products leads to the symptoms of PWS. Biallelic expression of UBE3A and ATP10A in the brain is speculative. In an AS im printing mutation, the “paternal-specific” genes are inappropriately demethylat ed and activated on the maternal chromosome. Activation of the UBE3A -antisense transcript on the maternal allele is predicted to cause the loss of UBE3A expression in crit ical regions of the brain. There is evidence for the loss of methylation at SNRPN and inappropriate activation of this gene on the maternal chromosome in response to an AS imprinting mutation. The cha nges in methylation and expression of other genes shown here is speculative.

PAGE 88

78 Figure 4-3. Southern blot examination of DNA methylation. Br ain DNA was digested with methylation-sensitive restrictio n enzymes, resolved on a 0.8% agarose gel, Southern blotted, and hybridized with radiolabeled probes for Mkrn3 , Ndn , and Snrpn . For Mkrn3 , the DNA was cut with BamHI and NotI. The methylated fraction of DNA produced a 20 kb band, and the unmethylated DNA produced an 18 kb band. For Ndn , the DNA was cut with BglII and SacII. The methylated fraction of DNA produced a 5 kb band, and the unmethylated DNA produced a 3 kb band. For Snrpn , the DNA was cut with TaqI and SacII. The methylated fr action of DNA produced a 5.6 kb band and the unmethylated DNA produced a 1.8 kb band. The extent of DNA methylation at each locus was judged by co mparing the intensity of the larger band to the intensity of the smaller ba nd for a given sample. A reduction in the intensity of the upper band and an in crease in the intensity of the lower band would indicate a loss of methylation. Maternal inheritance of the 12 kb deletion (shown in lane 2) and matern al inheritance of the 7 kb deletion (shown in lane 6) did not reveal an o bvious loss of methylation as compared to the wild-type controls.

PAGE 89

79 Figure 4-4. Bisulfite analysis of the Snrpn promoter in wild-type mice. Brain DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Out of 14 clones, 71% were heavily methylated and 29% were mostly unmethylated.

PAGE 90

80 Figure 4-5. Bisulfite analysis of the Snrpn promoter in 12 kb de letion mice. Brain DNA was treated with sodium bisulfite and then PCR am plification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line repres ents an individual clone that was sequenced. Filled circles indicate met hylated CpGs, and open circles indicate unmethylated CpGs. Out of 19 clones, 21% were heavily methylated and 79% were mostly unmethylated.

PAGE 91

81 Figure 4-6. Bisulfite analysis of the Snrpn promoter in 7 kb de letion mice. Brain DNA was treated with sodium bisulfite and then PCR am plification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line repres ents an individual clone that was sequenced. Filled circles indicate met hylated CpGs, and open circles indicate unmethylated CpGs. Out of 20 clones, 45% were heavily methylated and 55% were mostly unmethylated.

PAGE 92

82 Figure 4-7. Bisulfite analysis of Snrpn intron 8 in wild-type mice. Brain DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for Snrpn intron 8. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Out of 22 clones, 18% were methylated at 3 of 5 CpGs, 41% were methylated at 4 of 5 CpGs, a nd 41% were methylated at all 5 CpGs.

PAGE 93

83 Figure 4-8. Bisulfite analysis of Snrpn intron 8 in 12 kb deletion mice. Brain DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for Snrpn intron 8. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Out of 11 clones, 10% were methylated at 3 of 5 CpGs, 45% were methylated at 4 of 5 CpGs, a nd 45% were methylated at all 5 CpGs.

PAGE 94

84 Figure 4-9. Bisulfite analysis of Snrpn intron 8 in 7 kb deleti on mice. Brain DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for Snrpn intron 8. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Out of 17 clones, 18% were methylated at 3 of 5 CpGs, 29% were methylated at 4 of 5 CpGs, a nd 53% were methylated at all 5 CpGs.

PAGE 95

85 Figure 4-10. B6.Cast.c7 mice provide polymor phisms for allele-specific analysis. The B6.Cast.c7 line is a congenic line th at contains a region of chromosome 7 from Mus musculus castaneus on a C57BL/6J background. Females harboring the 12 kb or 7 kb deletion (A ) were mated with males of the B6.Cast.c7 line (B) to produce F1 hybr id offspring (C). The maternallyinherited and paternally-inhe rited alleles in the F1 mice could be distinguished by using sequence polymorphisms between the domesticus and castaneus alleles. Total brain DNA was isolated fr om newborn F1 mice and subjected to sodium bisulfite genomic sequencing anal ysis to measure methylation levels on each allele at Snrpn and Mkrn3 .

PAGE 96

86 Figure 4-11. Bisulfite analysis of the Snrpn promoter in wild-type mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 1 and 2.

PAGE 97

87 Figure 4-12. Bisulfite analysis of the Snrpn promoter in 12 kb de letion mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 1 and 2.

PAGE 98

88 Figure 4-13. Bisulfite analysis of the Snrpn promoter in 7 kb deletion mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Snrpn promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 1 and 2.

PAGE 99

89 Figure 4-14. Summary of bi sulfite analysis of the Snrpn promoter. The frequency at which each CpG residue was methylated among all the clones from a given sample was determined, and the results for all 14 CpG residues within the region analyzed were combined to generate a single index of the total level of methylation measured in each sample. The results for the maternal and paternal alleles for each sample are shown on the horizontal axis, with the number of clones sequenced for each allele listed below. The percent methylation index for each allele is shown on the vertical axis.

PAGE 100

90 Figure 4-15. Bisulfite analysis of Snrpn intron 1 in wild-type mice. DNA was treated with sodium bisulfite and then P CR amplification was performed using primers specific for Snrpn intron 1. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 7 and 8.

PAGE 101

91 Figure 4-16. Bisulfite analysis of Snrpn intron 1 in 12 kb deletion mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for Snrpn intron 1. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 7 and 8.

PAGE 102

92 Figure 4-17. Bisulfite analysis of Snrpn intron 1 in 7 kb deletion mice. DNA was treated with sodium bisulfite and then P CR amplification was performed using primers specific for Snrpn intron 1. The PCR product was then cloned and sequenced. Each line represents an individual clone that was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 7 and 8.

PAGE 103

93 Figure 4-18. Summary of bisulfite analysis of Snrpn intron 1. The frequency at which each CpG residue was methylated among all the clones from a given sample was determined, and the results for all 13 CpG residues within the region analyzed were combined to generate a single index of the total level of methylation measured in each sample. The results for the maternal and paternal alleles for each sample are shown on the horizontal axis, with the number of clones sequenced for each allele listed below. The percent methylation index for each allele is shown on the vertical axis.

PAGE 104

94 Figure 4-19. Bisulfite analysis of the Mkrn3 promoter in wild-type mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Mkrn3 promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 5 and 6.

PAGE 105

95 Figure 4-20. Bisulfite analysis of the Mkrn3 promoter in 12 kb de letion mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Mkrn3 promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 5 and 6.

PAGE 106

96 Figure 4-21. Bisulfite analysis of the Mkrn3 promoter in 7 kb deletion mice. DNA was treated with sodium bisulfite and then PCR amplification was performed using primers specific for the Mkrn3 promoter. The PCR product was then cloned and sequenced. Each line represents an individual clone th at was sequenced. Filled circles indicate methylated CpGs, and open circles indicate unmethylated CpGs. Clones are grouped by parental origin (labeled “M” for maternal and “P” for paternal) as de termined by the Cast.c7 polymorphism located between CpGs 5 and 6.

PAGE 107

97 Figure 4-22. Summary of bi sulfite analysis of the Mkrn3 promoter. The frequency at which each CpG residue was methylated among all the clones from a given sample was determined, and the results for all 21 CpG residues within the region analyzed were combined to generate a single index of the total level of methylation measured in each sample. The results for the maternal and paternal alleles for each sample are shown on the horizontal axis, with the number of clones sequenced for each allele listed below. The percent methylation index for each allele is shown on the vertical axis.

PAGE 108

98 Figure 4-23. Crossing with the PWS-IC de letion mouse to measure gene expression. Paternal inheritance of the 35 kb PWS -IC deletion causes a loss of paternalspecific gene expression. We reas oned that this reduction of normal expression from the paternal allele would provide a low background over which it would be possible to detect in appropriate activation of the maternal allele. Females harboring the 12 kb or 7 kb deletion (A) were mated with males carrying the 35 kb PWS-IC deleti on (B) to produce offspring in which expression of the paternal-specific ge nes was greatly reduced (C). Whole brain RNA was isolated from offspring at birth, and Ndn , Snrpn , and snoRNA gene expression was measured by Northern blot.

PAGE 109

99 Figure 4-24. Northern blot analysis of Ndn and Srnpn expression. Whole brain RNA was isolated from newborn mice produ ced by crossing females harboring the 12 kb or 7 kb deletion with males of the 35 kb PWS-IC deletion strain. The following six genotypes were examined: (1), wild-type; (2), maternal 12 kb deletion; (3), maternal 7 kb deletion; (4 ), paternal 35 kb PWS-IC deletion; (5), maternal 12 kb deletion and paternal PWS -IC deletion; and (6), maternal 7 kb deletion and paternal PWS-IC deleti on. For each sample, 10 ug of RNA was resolved on a 1% formaldehyde-containing agarose gel, Northern blotted, and hybridized sequentially with radiolabeled probes for Ndn , Snrpn , and Actin . ( Actin was examined as a loading control. ) The top panel shows the results for Ndn , the middle panel shows the results for Snrpn , and the bottom panel shows the results for Actin . To the right of each panel, the size of the transcript is indicated in kilobases.

PAGE 110

100 Figure 4-25. Northern blot analysis of snoRNA expr ession. Whole brain RNA was isolated from newborn mice produced by crossing females harboring the 12 kb or 7 kb deletion with males of th e 35 kb PWS-IC deletion strain. The following six genotypes were examined: (1), wild-type; (2), maternal 12 kb deletion; (3), maternal 7 kb deletion; (4 ), paternal 35 kb PWS-IC deletion; (5), maternal 12 kb deletion and paternal PWS -IC deletion; and (6), maternal 7 kb deletion and paternal PWS-IC deleti on. For each sample, 10 ug of RNA was resolved on an 8% acrylamide gel cont aining 7 M urea, Nort hern blotted, and hybridized sequentially with radiolabeled probes for MBII-13, MBII-85, MBII-52, and 5.8S rRNA. (The 5.8S ri bosomal RNA was examined as a loading control.) The top panel shows the results for MBII-13, the middle panel shows the results for MBII-85 and MBII-52, and the bottom panel shows the results for 5.8S rRNA. To the right of each panel, the size of the transcript is indica ted in base pairs.

PAGE 111

101 Figure 4-26. Allele-specific expression of Ube3a measured by RT-PCR analysis. Brain cDNA samples were generated from mi ce produced by the following matings: (1), B6.Cast.c7 female crossed with w ild-type male; (2), B6.Cast.c7 female crossed with 12 kb deletion male; (3), B 6.Cast.c7 female crossed with 7 kb deletion male; (4), wild-type female crossed with B6.Cast.c7 male; (5), 12 kb deletion female crossed with B6.Cast.c7 male; and (6), 7 kb deletion female crossed with B6.Cast.c7 male. An RT-PCR product for Ube3a was amplified using primers for Ube3a exons 5 and 6, and 2/3 of each product was digested with Tsp509I. This enzyme cuts th e B6.Cast.c7 allele once and the domesticus allele twice. The sample s were resolved on a 4.8% agarose gel, Southern blotted, and hybridized with a radiolabeled probe derived from Ube3a exons 5 and 6. The proportion of Ube3a expression derived from the B6.Cast.c7 allele was indicated by th e intensity of the 225 bp band, and the expression derived from the domesticus al lele was indicated by the intensity of the 138 and 87 bp bands. No differences in Ube3a expression were detected in the 12 kb and 7 kb deletion samples as compared to the wild-type samples.

PAGE 112

102 CHAPTER 5 DISCUSSION Introduction Our goal was to produce a mouse model for Angelman sydrome imprinting mutations by generating a series of nested deletions upstream of the Snrpn gene. As described in chapter 4, analysis of mice ha rboring a 12 kb deletion or a 7 kb deletion did not provide evidence that either of these de letions produced an imprinting defect. This chapter will discuss the results of the DNA methylation and ge ne expression analysis of these mice and will suggest other studies that could be performed to provide a more complete investigation of the effects (or l ack thereof) resulting from the 12 kb and 7 kb deletions. The results from the experiments de scribed in this dissertation will then be placed in the context of findings from other la boratories that raise interesting questions about the nature and location of the IC. Th e chapter will conclude with an outline of further experiments that could be conducted to gain greater insight into the mechanism and function of the Prader-Willi/Angelman imprinting center. Analysis of 12 kb and 7 kb Deletions To determine if the murine AS-IC was lo cated within either the 12 kb or the 7 kb deletion, mice harboring each deletion were tested for signs of an imprinting defect. Because imprinting mutations in human A ngelman syndrome patients cause a loss of DNA methylation and inappropriate activation of at least the Snrpn gene and perhaps other PWS-candidate genes on the maternal chromosome, the 12 kb and 7 kb deletion

PAGE 113

103 mice were examined for these same changes as indicators of an Angelman imprinting mutation in mice. Methylation Analysis First, DNA methylation at thr ee paternally-expressed genes, Mkrn3 , Ndn , and Snrpn , was investigated by Southern blot to determine if maternal inheritance of the 12 kb or 7 kb deletion caused a loss of methylatio n at these genes. This analysis did not reveal any obvious loss of methylation, and th erefore suggested that the AS-IC in mouse was not located within the 12 kb deleted regi on. The Southern blot data alone, however, hardly represented a comprehensive analys is of DNA methylation levels. One major limitation of this approach was a lack of sensitivity. A modest change in methylation levels might not easily be detected by Southe rn blot. This could be addressed to some extent by repeating the analysis with vari ous quantities of DNA for each sample and by taking a range of exposures for each blot to achieve an optimal intensity that might be quantified using densitometry software. Even going to such lengths, however, it would still be difficult to detect a small change in methylation. Furthermore, a second major limitation to the Southern blot approach was that only a si ngle methylation-sensitive restriction site was examined for each gene. Even if additional methylation-sensitive restriction sites showing allele -specific methylation were identified for these genes, such examination would still be confined to a handf ul of CpG sites. Despite these limitations, however, the Southern blot analysis served its purpose as a quick and inexpensive preliminary test for an imprinting mutation. In order to gain a more complete pi cture of DNA methyla tion levels at the Snrpn gene, the Southern blot analysis was follo wed by the higher-resolution sodium bisulfite genomic sequencing analysis of the Snrpn promoter and intron 8 regions. The chief

PAGE 114

104 advantage of the bisulfite technique was th at it provided a means of investigating the methylation level at numerous CpG sites – in this case 14 CpG residues in the promoter and 5 CpG residues within intron 8 of Snrpn . Unfortunately, the data for these intial attempts at bisulfite sequencing were inconclusive. The results for the Snrpn promoter, which are discussed in greater detail below, pr oved to be difficult to interpret. On the other hand, the results for the Snrpn intron 8 region did not indi cate any clear differences between the wild-type, 12 kb deletion, or 7 kb deletion samples. It was difficult to make any statement regarding an imprinting phenotype from the Snrpn intron 8 data, however, considering that there was no prior eviden ce that an Angelman imprinting mutation would alter methylation levels at this region. For the Snrpn promoter, 71% of the clones seque nced for the wild-type sample were heavily methylated as compared to 21% of the clones from the 12 kb deletion sample and 45% of the clones from the 7 kb deletion sample that showed extensive methylation. Previous studies had dem onstrated that the promoter region of Snrpn was only methylated on the maternal allele in soma tic cells and that this maternal methylation imprint was inherited from the oocyte [ 134, 143, 144, 148]. Therefore, it was surprising that our bisulfite sequencing produced such a large proportion of methylated clones for the wild-type sample. It was expected that roughly half of the clones from wild-type samples would show an unmethylated profile representing the paternal allele and half would show a methylated profile representing the maternal allele. One explanation for the discrepancy between the expectations for a nd the actual result of the bisulfite analysis is that the expectations were based on a fals e assumption. The data from the studies cited above, however, clearly indicate di fferential methylation of the Snrpn promoter. A more

PAGE 115

105 likely explanation is that the highly skewed re sults for the wild-type sample were due to PCR amplification bias. This called into question the results for the 12 kb deletion sample that by direct comparison to the re sults for the wild-type sample might have seemed to indicate a reduced amount of DNA methylation in the 12 kb deletion sample. One way to address the problem of PCR bias would be to sequence a large number of clones from multiple amplifications of more than one bisulfite conversion for each DNA sample. A more effective solution to this problem, however, came by using intersubspecific crosses to allow the methyla tion level at each allele to be assessed independently as described below. An improvement over the initial attempts at bisulfite seque ncing came by crossing females carrying the 12 kb or 7 kb deletion with males of the castaneus chromosome 7 line (B6.Cast.c7). This allowed the parent al alleles to be distinguished by sequence polymorphisms and provided a much more st raightforward and r obust analysis of DNA methylation levels at the Snrpn promoter and intron 1 regions as well as the promoter of the Mkrn3 gene. For all three genes, clones derive d from the maternal allele were heavily methylated and clones derived from the paternal allele were almost entirely unmethylated. These data did not provide a ny indication of a loss of methylation on the maternal allele in the 12 kb deletion or 7 kb deletion samples as compared to the wild-type samples. For the Snrpn intron 1 and Mkrn3 promoter, only a small number of clones that were sequenced represented the mate rnal alleles. In or der to draw any firm conclusions regarding the level of methylation at these region s, it will be necessary to repeat this analysis. For the Snrpn promoter, however, the results were clear and

PAGE 116

106 convincing. These data indicated that there was no Angelman imprinting mutation in the 12 kb deletion or 7 kb deletion mi ce as judged by DNA methylation. Gene Expression Analysis In addition to looking for changes in DNA methylation as an indicator of an imprinting mutation, the expression of severa l paternal-specific genes was examined to determine if maternal transmission of either deletion would cause inappropriate activation of these genes on the maternally inherited chromosome harboring the deletion. Crossing females carrying the 12 kb or 7 kb deletion w ith males of the 35 kb PWS-IC deletion line proved to be an effective means of reduci ng the normal expression from the paternal allele so as to provide a way of more eas ily detecting inappropri ate activation of the maternal allele. Northern blot anal ysis for the paternal-specific genes Ndn , Snrpn , MBII-13, MBII-85, and MBII-52 did not reveal any evidence for activation of these genes on a maternal chromosome carrying e ither the 12 kb or the 7 kb deletion. These results did not indicate an imprinting def ect in the 12 kb or 7 kb deletion mice and strongly suggested that the murine AS-IC is located outside of the 12 kb deleted region. In addition to testing for inappropriate act ivation of the patern al-specific genes on the maternal chromosome, expression of the Ube3a gene was investigated as another means of assessing an imprinting mutation. Whole brain RNA was isolated from offspring of 12 kb and 7 kb deletion mice cro ssed with B6.Cast.c7 mice, and an RT-PCR assay was employed to examine allele-specific expression of Ube3a . The results of this analysis did not reve al any decrease in Ube3a expression from the maternal allele in the 12 kb or 7 kb deletion samples as compared to the wild-type sample. Consistent with the data from studies of DNA methylation and gene expression levels for several

PAGE 117

107 paternal-specific genes, the investigation of Ube3a expression levels did not indicate an Angelman imprinting defect in the 12 kb or 7 kb deletion mice. One major shortcoming common to all of the results reported here is the absence of a positive control for any of the assays. Although other mouse models of Angelman syndrome do exist and perhaps could be used as controls for some phenotypic measures, none of the models that have been publishe d would provide an adequate positive control for the molecular tests that we have app lied to the 12 kb and 7 kb deletion mice. The chromosome 7 paternal unipa rental disomy mouse model produced by Cattanach and colleagues, the several megabase deletion m odel characterized by the Nicholls lab, and the two Ube3a knockout models generated independe ntly by the Beaudet and Wagstaff labs would not serve as suitable controls fo r our analysis of DNA methylation and gene expression [76, 131, 139, 140]. Although currently we do not possess a positive control for most of the DNA methylation and gene expressi on assays, two mouse models engineered in our lab could be used as controls specifically in the Ube3a RT-PCR assay. One control would be a sample from a hybrid “PWS-IC deletion” m ouse, produced by crossing a B6.Cast.c7 female with a male carrying a 35 kb PWS-IC deletion. The PWS-IC deletion sample should show an increase in Ube3a expression from the paternal chromosome harboring the PWS-IC deletion as compared to a wild-t ype sample [123]. Another control sample could be produced by crossing a female of th e “human PWS-IC subs titution” line with a B6.Cast.c7 male. The “human PWS-IC s ubstitution” line was created by Dr. Karen Johnstone, a research assistant professor in our lab. Mice that inhe rit this mutation on the

PAGE 118

108 maternal chromosome show a reduced level of Ube3a expression from the maternal allele in the brain (Dr. Karen Johnstone, personal communication). Future Analysis of the 12 kb and 7 kb Deletions Methylation Analysis Although the results of the bisulfit e analysis for the promoter of Snrpn are conclusive, the examinations of DNA methylation levels at Snrpn intron 1 and the Mkrn3 promoter need to be repeated to confirm th e findings. Future work could also involve identifying polymorphisms between the domesticus (C57BL/6J or 129S1/Sv) and castaneus alleles within the CpG islands of othe r paternal-specific genes. This would allow the bisulfite analysis to be extended to Frat3 and Magel2 to provide a better understanding of the methylation levels of each allele for these genes in both wild-type and deletion mice. If any restriction frag ment length polymorphisms are found within these genes, it might be possible to employ an allele-specific Southern blot assay for methylation. Since convincing results are al ready in hand for the bi sulfite analysis of Snrpn , however, these additional investigations are a low priority. Gene Expression Analysis In order to gain an idea of the level of expression for Ube3a and Atp10a , these genes will be subjected to Northern blot an alysis in order to determine if there are differences in the total leve l of expression for each gene between wild-type and deletion mice. If Northern blot anal ysis is inconclusive, Western blots could be performed to determine the amount of protein produced for each gene. It is possible that Atp10a is not imprinted in the mouse strain we are using, so assessing the expression for this gene is a lower priority than exam ining the expression of Ube3a .

PAGE 119

109 An alternative approach to measure gene expression for both the paternal-specific and maternal-specific genes will take advantage of expressed polymorphisms between the domesticus and castaneus alleles. Such polymorphisms ha ve already been identified for Mkrn3 , Ndn , Snrpn , and Ube3a . These polymorphisms will allow the expression level for each allele of a gene to be determ ined by performing RT-PCR and sequencing the product. Similar to the procedure used to investigate expression of Ube3a that was described in the previous chap ter, an AvaII restriction frag ment length polymorphisms in Ndn provides a method of distinguishing each allele by cutting the RT-PCR product with the AvaII enzyme, resolving on an agarose ge l, and Southern blotting [147]. Also, the Tsp509I polymorphism RT-PCR assay for Ube3a should be repeated so as to include appropriate controls such as the “PWS -IC deletion” and the “human PWS-IC substitution” mouse models. Phenotype Analysis As another means of assessing an imprin ting defect, the 12 kb deletion and 7 kb deletion mice could be examined for signs of obe sity, ataxia, or learni ng defects that have been detected in other mouse models for Angelman syndrome [131, 139, 140]. So far, no indications of obesity or atax ia have been observed in the 12 kb or 7 kb deletion mice, but this has not yet been carefully documented. The weights of the deletion mice will be compared to the weights of wild-type litte rmates to determine if the deletion mice become obese. Ataxia could initially be asse ssed by a bar crossing test, but a full picture of coordination would require a more in-dep th workup on a rota-rod apparatus that, along with testing for electroencepha lographic and learning abnormali ties, would best be done in collaboration with a lab that specializes in analysis of mouse behavior. Considering that the 12 kb and 7 kb deletion mice do not sh ow any molecular aspects of an Angelman

PAGE 120

110 imprinting mutation, it seems unlikely that furt her testing of these mice will reveal the murine equivalent of an Angelman syndrome phenotype. Relevant Findings from Other Laboratories At the 2002 American Society of Huma n Genetics Meeting, the Beaudet lab reported that a 100 kb deletion extending from –11 kb to –111 kb relative to Snrpn exon 1 did not produce an obvious phenot ype or imprinting defect regardless of whether the deletion was inherited on the pate rnal or the maternal chromosome [149]. The deletion was engineered using a puromycin cassette to select fo r the targeting of a loxP site 11 kb upstream of Snrpn and a neomycin cassette to se lect for the targeting of a loxP site 111 kb upstream of Snrpn . Cre-mediated recombination between the loxP sites was employed to create the 100 kb deleti on. Considering that the 12 kb and 7 kb deletions produced in our lab fall within this larger deletion interv al, it is not surprising that the 12 kb and 7 kb deletion mice also did not show signs of an imprinting defect. (Figure 5-1) Taken together, the results of the 12 kb, 7 kb, and 100 kb deletions are consistent with the notion that the location of the AS-IC is not conserved between human and mouse. Interestingly, the Beaudet lab found th at mice inheriting the 3-prime proximal anchor site for the large deletion on the mate rnal chromosome showed a partial loss of methylation at Ndn and a complete loss of methylation at Snrpn as measured by Southern blot. It is not clear if this effect was due to the expressi on of the puromycin selectable marker located 11 kb upstream of Snrpn or if the orientation of the puromycin cassette, that was transcribed in the direction opposite of Snrpn , was relevant to the imprinting mutation. One interpretation of these results is that the insertion event 11 kb upstream of Snrpn or the expression of the selectable mark er at this location interfered with the

PAGE 121

111 function of a nearby element required to esta blish or maintain the proper imprint on the maternally inherited chromsome. This presen ted the possibility that the AS-IC is located much closer to Snrpn in mouse than in human. An a lternative interpretation is that transcription from the puromycin casse tte disrupted the function of upstream “IC transcripts” that could be involved in the imprinting process. This second interpretation will be discussed more extensively below. Dr. Bernhard Horsthemke’s group has iden tified alternative transcripts of the SNRPN gene that span the region upstream of SNRPN containing the Angelman syndrome imprinting center and that have been proposed to play a role in the function of the IC [150, 151]. These transcripts, which ar e reported to be expr essed specifically from the paternal chromosome and at le vels much lower than the main SNRPN transcript, originate from two different promoters located roughly 150 kb and 115 kb upstream of SNRPN exon 1. Both of these transcriptional st art sites are methylated on the maternal chromosome and unmethylated on the paternal chromosome. There are a wide variety of IC transcript splice variants, but they all lack SNRPN exon 1 and instead splice into SNRPN exon 2. The IC transcripts have been observed at highest levels in brain and heart but are detectable at lower levels in testis, ovary, and numerous other tissues. Several of the IC exons are very similar in sequence, and it has b een suggested that the IC exons arose through multiple duplication events. Horsthemke and colleagues have suggested that the IC transcripts may be involved in imprint switching, but the precise functi on of the transcripts is uncertain. The upstream IC exons do not appear to have protei n coding potential, and it is unclear if the RNA molecules themselves are functional. Indee d, the fact that some IC transcript splice

PAGE 122

112 variants contain fragments of Alu sequences might suggest th at the transcripts are not functional. Perhaps it is not the transcripts themselves but ra ther the act of transcription through the locus that is important. On th e other hand, it might be a specific splice variant or a subset of splice variants that carry out the im printing task. The finding that one of the IC exons, the “u5” exon, is cont ained within the AS-SRO (the 880 bp region located 35 kb upstream of SNRPN that is common to all AS-IC deletions) suggests that transcripts containing this exon might play a role in the imprinting process [121, 151]. It could be a coincidence, however, that th e u5 exon is located within the AS-SRO. Perhaps these cases of Angelman syndrome imprinting mutations are not due to the disruption of the IC transcript but instead are caused by th e deletion of an unidentified regulatory element located within the SRO. So although the transcription upstream of SNRPN and/or the transcripts th at are produced are suspected to participate in the function of the IC, it is also possible that the IC transcript phenomenon is entirely irrelevant to the imprinting process. Exons upstream of the Snrpn gene have also been iden tified in mouse [152, 153]. Although it was not possible to te st all varieties of these ups tream transcripts for allelespecific expression, transcript s including upstream exons that contained informative polymorphisms were demonstrated to be expres sed exclusively from th e paternal allele. Similar to what was seen for the IC transc ripts in human, many of the mouse upstream transcripts splice into Snrpn exon 2. In contrast to the hum an IC transcripts, however, the upstream transcripts in mouse are detected only in brain. The 11 known mouse upstream exons show considerable sequence similarity to each other and to 36 analogous exons identified in rat, but they do not show homo logy to the human IC exons. Interestingly,

PAGE 123

113 some transcripts originating from the upstream exons in mouse do not include Snrpn exons 2-10 but instead splice into exons located downstream of the Snrpn locus. In humans, exons lo cated downstream of SNRPN are thought to be part of a large transcript spanning a 460 kb region between SNRPN and UBE3A [106]. It has been proposed that these long transcripts not only serve as a host for the processing of numerous snoRNA genes but also might be involved in the regul ation of imprinted UBE3A expression. The observation of transcri pts that contain both exons that are upstream of Snrpn as well as exons that are downstream of Snrpn raises the possibility that the upstream exons are part of an antisense transcript to the Ube3a gene [153]. Despite the observation of “IC” or upstream transcripts in both human and mouse and the suggestion that these may play a role in the imprinting process, evidence in favor of the AS-IC operating as a cisacting element comes from studies of transgenic mice. It was found that a minitransgene composed of a 1 kb sequence containing the human AS-SRO joined to a 200 bp fragment containing the mouse Snrpn minimal promoter and exon 1 was properly imprinted [154]. The re sults of these experi ments showing that upon maternal transmission the 1 kb AS-SRO fragment was able to confer a methylated and silenced status on the Snrpn promoter suggest the existe nce of an element residing entirely within the AS-SRO that is sufficien t to carry out the func tion of establishing the maternal imprint. The transgenic experime nts described here are somewhat artificial, however, and a more complete understandi ng of the mechanism by which the AS-IC functions will require further investigation. Future Directions In order to gain greater insight into the function of the AS-IC, it would be useful to have an AS-IC deletion mouse model. As the results from the 12 kb and 7 kb deletions

PAGE 124

114 produced in our lab taken together with th e results of the 100 kb deletion produced in Dr. BeaudetÂ’s lab suggest that the location of the AS-IC is not conserved between human and mouse, this raises the obvious question of where to look for the murine AS-IC. Perhaps two different interpretations of th e Angelman imprinting mutation produced by the insertion 11 kb upstream of Snrpn will provide a logical starting point. Considering that this insertion mutation might exert its influence by disrupting the function of a nearby element that resides close to Snrpn , it could be possible to identify this hypothetical element by engineering deletions starting 11 kb upstream of Snrpn and extending very close to the gene itself. The success of this project might depend on making several nested deletions to guar d against the danger of impinging upon the PWS-IC. If the AS-IC in mouse is located very close to the PWS-IC, however, it might not be possible to distinguish the two elem ents using a traditional targeted deletion approach. On the other hand, if the insertion 11 kb upstream of Snrpn produced an Angelman imprinting mutation by interfering with the ups tream transcripts, it might be possible to recapitulate this imprinting mutation by deleting the promoters of the upstream transcripts. Given the complex nature of the upstream transcripts that show extensive variation in splicing and which are initiated from multiple promoters, perhaps the most straightforward way of accomplishing this goa l would be to generate a large deletion extending even further than 111 kb upstream of Snrpn so as to include all of the promoters for the upstream transcripts. Anot her means of addressing this issue would be to target a transcriptional terminator upstream of Snrpn . This strategy would be risky, because it might be difficult to efficiently truncate the upstream transcripts. But if

PAGE 125

115 termination of the upstream transcripts produ ced an Angelman imprinting mutation, this would provide strong evidence for the involveme nt of these transcripts in the imprinting process. In conclusion, it appears that there is stil l the opportunity for ge netic experiments to contribute useful information about the function of the Prader-Willi/Angelman imprinting center. It seems that the field of study is at a turn ing point, however, and soon biochemical studies of cis-acting sequences and the trans-acting f actors that bind them will provide a more detailed understanding of the mechanism of imprinting involved in the Prader-Willi and Angelman syndromes.

PAGE 126

116 Figure 5-1. Insertion 11 kb upstream of Snrpn creates an imprinting mutation. Dr. Art BeaudetÂ’s lab created a 1 00 kb deletion upstream of Snrpn using CreloxP technology. One loxP site was introduced 11 kb upstream of Snrpn , and a second loxP site was inserted 111 kb upstream of Snrpn . Cre-mediated recombination between the loxP sites was employed to create the 100 kb deletion. Dr. BeaudetÂ’s lab also cr eated a second deletion starting 111 kb upstream of Snrpn and extending to include the Ndn gene located roughly 1.5 mb upstream of Snrpn . (This second deletion is not shown.) The small filled squares indicate Snrpn exons 1 to 10. The 100 kb deletion created by the Beaudet lab is shown by a line above, with a triangle indicating the loxP site at one end of the deletion. The 3-prime selectable marker/ loxP anchor site for the deletion is indicated by the larg e rectangle. The 7 kb, 12 kb, and 58 kb deletions are shown by lines below. The 100 kb deletion did not produce an obvious phenotype or imprinting defect, but maternal inheritance of the 3prime proximal anchor site for the deletion showed an imprinting mutation as revealed by a partial lo ss of methylation at Ndn and a complete loss of methylation at Snrpn as measured by Southern blot.

PAGE 127

117 LIST OF REFERENCES 1. M.S. Bartolomei and S.M. Tilghman, Genomic imprinting in mammals. Annu. Rev. Genet., 1997. 31 : p. 493-525. 2. S.C. Barton, M.A. Surani, and M.L. Norris, Role of paternal and maternal genomes in mouse development. Nature, 1984. 311 : p. 374-376. 3. J. McGrath and D. Solter, Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell, 1984. 37 : p. 179-183. 4. D.J. Driscoll, Genomic imprinting in humans. Mol. Genet. Med., 1994. 4 : p. 3777. 5. P.A. Jacobs, C.M. Wilson, J.A. Sprenkle, N.B. Rosenshein, and B.F. Migeon, Mechanism of origin of co mplete hydatidiform moles. Nature, 1980. 286 : p. 714716. 6. N.M. Lindor, J.A. Ney, T.A. Gaffey, R. B. Jenkins, S.N. Thibodeau, and G.W. Dewald, A genetic review of complete and partial hydatidiform moles and nonmolar triploidy. Mayo Clin. Proc., 1992. 67 : p. 791-799. 7. K. Ohama, K. Nomura, E. Okamoto, Y. Fukuda, T. Ihara, and A. Fujiwara, Origin of immature teratoma of the ovary. Am. J. Obstet. Gynecol., 1985. 152 : p. 896900. 8. A.G. Searle and C.V. Beechey, Complementation studies with mouse translocations. Cytogenet. Cell. Genet., 1978. 20 : p. 282-303. 9. B.M. Cattanach and M. Kirk, Differential activity of maternally and paternally derived chromosome regions in mice. Nature, 1985. 315 : p. 496-498. 10. B. Cattanach, Chromosome imprinting and its significance for mammalian development. Genome Anal., 1991. 2 : p. 41-71. 11. D.H. Ledbetter and E. Engel, Uniparental disomy in humans Development of an imprinting map and its implica tions for prenatal diagnosis. Hum. Mol. Genet., 1995. 4 : p. 1757-1764. 12. S.P. Otto and D.B. Goldstein, Recombination and the evolution of diploidy. Genetics, 1992. 131 : p. 745-751.

PAGE 128

118 13. S. Varmuza and M. Mann, Genomic imprinting--defusi ng the ovarian time bomb. Trends Genet., 1994. 10 : p. 118-123. 14. J.H. Thomas, Genomic imprinting proposed as a surveillance mechanism for chromosome loss. Proc. Natl. Acad. Sci. USA, 1995. 92 : p. 480-482. 15. S. Varmuza, Gametic imprinting as a speci ation mechanism in mammals. J. Theor. Biol., 1993. 164 : p. 1-13. 16. T. Moore and D. Haig, Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet., 1991. 7 : p. 45-49. 17. P.A. Leighton, J.R. Saam, R.S. Ingram, and S.M. Tilghman, Genomic imprinting in mice: Its function and mechanism.. Biol. Reprod., 1996. 54 : p. 273-278. 18. P.B. Vrana, X.J. Guan, R.S. Ingram, and S.M. Tilghman, Genomic imprinting is disrupted in interspecific Peromyscus hybrids. Nat. Genet., 1998. 20 : p. 362-365. 19. L.D. Hurst, Peromysci, promiscuity and imprinting. Nat. Genet., 1998. 20 : p. 315316. 20. L.D. Hurst and G.T. McVean, Growth effects of unipar ental disomies and the conflict theory of genomic imprinting. Trends Genet., 1997. 13 : p. 436-443. 21. L.D. Hurst and G.T. McVean, Do we understand the evolution of genomic imprinting? Curr. Opin. Genet. Dev., 1998. 8 : p. 701-708. 22. D. Haig, The kinship theory of genomic imprinting. Annu. Rev. Ecol. Syst., 2000. 31 : p. 9-32. 23. J.F. Wilkins and D. Haig, What good is genomic impr inting: the function of parent-specific gene expression. Nat. Rev. Genet., 2003. 4 : p. 359-368. 24. R. Ohlsson, A. Paldi, and J.A. Graves, Did genomic imprinting and X chromosome inactivation arise from stochastic expression? Trends Genet., 2001. 17 : p. 136-141. 25. F. Pardo-Manuel de Villena, E. de La Casa-Esperon, and C. Sapienza, Natural selection and the function of genomic imprinting: b eyond the silenced minority. Trends Genet., 2000. 16 : p. 573-579. 26. A.L. Beaudet and Y.H. Jiang, A rheostat model for a rapid and reversible form of imprinting-dependent evolution. Am. J. Hum. Genet., 2002. 70 : p. 1389-1397.

PAGE 129

119 27. J.R. Chaillet, T.F. Vogt, D.R. Beier, and P. Leder, Parental-specific methylation of an imprinted transgene is establishe d during gametogenesis and progressively changes during embryogenesis. Cell, 1991. 66 : p. 77-83. 28. T.M. DeChiara, E.J. Robertson, and A. Efstratiadis, Parental imprinting of the mouse insulin-like growth factor II gene. Cell, 1991. 64 : p. 849-859. 29. D.P. Barlow, R. Stoger, B.G. Herrmann, K. Saito, and N. Schweifer, The mouse insulin-like growth factor type-2 receptor is imprin ted and closely linked to the Tme locus. Nature, 1991. 349 : p. 84-87. 30. S.E. Leff, C.I. Brannan, M.L. Reed, T. Ozcelik, U. Francke, N.G. Copeland, and N.A. Jenkins, Maternal imprinting of the mouse Snrpn gene and conserved linkage homology with the hum an Prader-Willi syndrome region. Nat. Genet., 1992. 2 : p. 259-264. 31. B.M. Cattanach, J.A. Barr, E.P. Evans, M. Burtenshaw, C.V. Beechey, S.E. Leff, C.I. Brannan, N.G. Copeland, N.A. Jenkins, and J. Jones, A candidate mouse model for Prader-Willi syndrome whic h shows an absence of Snrpn expression. Nat. Genet., 1992. 2 : p. 270-274. 32. T. Ozcelik, S. Leff, W. Robinson, T. Donlon, M. Lalande, E. Sanjines, A. Schinzel, and U. Franke, Small nuclear ribonucleoprotein polypeptide N (SNRPN), an expressed gene in the Prader-Willi syndrome critical region. Nat. Genet., 1992. 2 : p. 265-269. 33. C.C. Glenn, K.A. Porter, M.T. Jong, R.D. Nicholls, and D.J. Driscoll, Functional imprinting and epigenetic modification of the human SNRPN gene. Hum. Mol. Genet., 1993. 2 : p. 2001-2005. 34. M.L. Reed and S.E. Leff, Maternal imprinting of human SNRPN, a gene deleted in Prader-Willi syndrome. Nat. Genet., 1994. 6 : p. 163-167. 35. G. Kelsey and W. Reik, Analysis and identifica tion of imprinted genes. Methods, 1998. 14 : p. 211-234. 36. R.J. Oakey and C.V. Beechey, Imprinted genes: identi fication by chromosome rearrangements and post-genomic strategies. Trends Genet., 2002. 18 : p. 359366. 37. C.V. Beechey, B.M. Cattanach, and A. Blake, Genetic and physical imprinting map of the mouse. 2001. http://www.mgu.har.mrc.ac.uk/research/imprinting/ , last accessed December 3, 2004. 38. D.P. Barlow, Gametic imprinting in mammals. Science, 1995. 270 : p. 1610-1613.

PAGE 130

120 39. M.S. Bartolomei, The search for imprinted genes. Nat. Genet., 1994. 6 : p. 220221. 40. Y. Hayashizaki, H. Shibata, S. Hirotsune, H. Sugino, Y. Okazaki, N. Sasaki, K. Hirose, H. Imoto, H. Okuizumi, M. Murama tsu, H. Komatsubara, T. Shiroishi, K. Moriwaki, M. Katsuki, N. Hatano, H. Sasa ki, T. Ueda, N. Mise, N. Takagi, C. Plass, and V.M. Chapman, Identification of an imprinted U2af binding-protein related sequence on mouse chromosome-11 using the RLGS method. Nat. Genet., 1994. 6 : p. 33-40. 41. J.V. Schmidt, P.G. Matteson, B.K. Jones, X.J. Guan, and S.M. Tilghman, The Dlk1 and Gtl2 genes are linked and reciprocally imprinted. Genes Dev., 2000. 14 : p. 1997-2002. 42. D.P. Barlow, Competition a common motif for the imprinting mechanism? EMBO J., 1997. 16 : p. 6899-6905. 43. F. Sleutels, D.P. Barlow, and R. Lyle, The uniqueness of the imprinting mechanism. Curr. Opin. Genet. Dev., 2000. 10 : p. 229-233. 44. M.L. Hanel and R. Wevrick, The role of genomic imprinting in human developmental disorders: lessons from Prader-Willi syndrome. Clin. Genet., 2001. 59 : p. 156-164. 45. J.M. Claverie, What if there are only 30,000 human genes? Science, 2000. 16 : p. 1255-1256. 46. W. Reik and J. Walter, Genomic imprinting: parental influence on the genome. Nat. Rev. Genet., 2001. 2 : p. 21-32. 47. R.I. Verona, M.R. Mann, and M.S. Bartolomei, Genomic imprinting: intricacies of epigenetic regul ation in clusters. Annu. Rev. Cell Dev. Biol., 2003. 19 : p. 237259. 48. P.A. Jones and D. Takai, The role of DNA methylation in mammalian epigenetics. Science, 2001. 293 : p. 1068-1070. 49. P.A. Jones, The DNA methylation paradox. Trends Genet., 1999. 15 : p. 34-37. 50. J. Peters, S.F. Wroe, C.A. Wells, H.J. Miller, D. Bodle, C.V. Beechey, C.M. Williamson, and G. Kelsey, A cluster of oppositely imprinted transcripts at the Gnas locus in the distal imprin ting region of mouse chromosome 2. Proc. Natl. Acad. Sci. USA, 1999. 96 : p. 3830-3835.

PAGE 131

121 51. A. Wutz, O.W. Smrzka, N. Schweifer, K. Schellander, E.F. Wagner, and D.P. Barlow, Imprinted expression of the Igf2r gene depends on an intronic CpG island. Nature, 1997. 389 : p. 745-749. 52. F. Sleutels and D.P. Barlow, The origins of genomic imprinting in mammals. Adv. Genet., 2002. 46 : p. 119-163. 53. E. Li, T.H. Bestor, and R. Jaenisch, Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell, 1992. 69 : p. 915-926. 54. E. Li, C. Beard, and R. Jaenisch, Role for DNA methyl ation in genomic imprinting. Nature, 1993. 366 : p. 362-365. 55. A.P. Wolffe and M.A. Matzke, Epigenetics: regulati on through repression. Science, 1999. 286 : p. 481-486. 56. C. Fournier, Y. Goto, E. Ballestar, K. Dela val, A. Hever, M. Esteller, and R. Feil, Allele-specific histone lysine methylati on marks regulatory regions at imprinted mouse genes. EMBO J., 2002. 21 : p. 6560-6570. 57. Z. Xin, C.D. Allis, and J. Wagstaff, Parent-specific complementary patterns of histone H3 lysine 9 and H3 lysine 4 methylation at the Prader-Willi syndrome imprinting center. Am. J. Hum. Genet., 2001. 69 : p. 1389-1394. 58. Z. Xin, M. Tachibana, M. Guggiari, E. Heard, Y. Shinkai, and J. Wagstaff, Role of histone methyltransferase G9a in CpG methylation of the Prader-Willi syndrome imprinting center. J. Biol. Chem., 2003. 278 : p. 14996-15000. 59. R.I. Gregory, T.E. Randall, C.A. Johnson, S. Khosla, I. Hatada, L.P. O'Neill, B.M. Turner, and R. Feil, DNA methylation is linked to deacetylation of histone H3, but not H4, on the imprin ted genes Snrpn and U2af1-rs1. Mol. Cell. Biol., 2001. 21 : p. 5426-5436. 60. S.B. Fulmer-Smentek and U. Francke, Association of acetyl ated histones with paternally expressed genes in th e Prader-Willi deletion region. Hum. Mol. Genet., 2001. 10 : p. 645-652. 61. J. Schweizer, D. Zynger, and U. Francke, In vivo nuclease hypersensitivity studies reveal multiple sites of parental origin-dependent differential chromatin conformation in the 150 kb SNRPN transcription unit. Hum. Mol. Genet., 1999. 8 : p. 555-566. 62. T. Watanabe, A. Yoshimura, Y. Mishima, Y. Endo, T. Shiroishi, T. Koide, H. Sasaki, H. Asakura, and R. Kominami, Differential chromatin packaging of genomic imprinted regions between expressed and non-expressed alleles. Hum. Mol. Genet., 2000. 9 : p. 3029-3035.

PAGE 132

122 63. N.J. Francis and R.E. Kingston, Mechanisms of transcriptional memory. Nat. Rev. Mol. Cell Biol., 2001. 2 : p. 409-421. 64. J. Mager, N.D. Montgomery, F. Pardo-Manuel de Villena, and T. Magnuson, Genome imprinting regulated by the mouse Polycomb group protein Eed. Nat. Genet., 2003. 33 : p. 502-507. 65. L.D. Hurst, G. McVean, and T. Moore, Imprinted genes have few and small introns. Nat. Genet., 1996. 12 : p. 234-237. 66. B. Neumann, P. Kubicka, and D.P. Barlow, Characteristics of imprinted genes. Nat. Genet., 1995. 9 : p. 12-13. 67. D. Kitsberg, S. Selig, M. Brandeis, I. Simon, I. Keshet, D.J. Driscol, R.D. Nicholls, and H. Cedar, Allele-specific replication timing of imprinted gene regions. Nature, 1993. 364 : p. 459-463. 68. J.H. Knoll, S.D. Cheng, and M. Lalande, Allele specificity of DNA replication timing in the Angelman/Prader-Willi sy ndrome imprinted chromosomal region. Nat. Genet., 1994. 6 : p. 41-46. 69. P.H. Gunaratne, M. Nakao, D. Ledbetter, J.S. Sutcliffe, and A.C. Chinault, Tissue-specific and allele-spe cific replication timing c ontrol in the imprinted human Prader-Willi syndrome region. Genes Dev., 1995. 9 : p. 808-820. 70. J.M. LaSalle and M. Lalande, Domain organization of allele-specific replication timing within the GABRB3 gene cl uster requires a biparental 15q11-13 contribution. Nat. Genet.., 1995. 9 : p. 386-394. 71. H. Kawame, S.M. Gartler, and R.S. Hansen, Allele-specific replication timing in imprinted domains absence of asynchrony at several loci. Hum. Mol. Genet., 1995. 4 : p. 2287-2293. 72. W.A. Bickmore and A.D. Carothers, Factors affecting the timing and imprinting of replication on a mammalian chromosome. J. Cell Sci., 1995. 108 : p. 28012809. 73. A. Paldi, G. Gyapay, and J. Jami, Imprinted chromosomal regions of the human genome display sexspecific meio tic recombination frequencies. Curr. Biol., 1995. 5 : p. 1030-1035. 74. W.P. Robinson and M. Lalande, Sex-specific meiotic recombination in the Prader-Willi/Angelman syndrome imprinted region. Hum. Mol. Genet., 1995. 4 : p. 801-806.

PAGE 133

123 75. J.M. LaSalle and M. Lalande, Homologous association of oppositely imprinted chromosomal domains. Science, 1996. 272 : p. 725-728. 76. J.M. Gabriel, M. Merchant, T. Ohta, Y. Ji, R.G. Caldwell, M.J. Ramsey, J.D. Tucker, R. Longnecker, and R.D. Nicholls, A transgene insertion creating a heritable chromosome deletion mous e model of Prader-Willi and angelman syndromes. Proc. Natl. Acad. Sci. USA, 1999. 96 : p. 9258-9263. 77. V.A. Holm, Cassidy, S.B., Butler, M.G., Hanchett, J.M., Greenswag, L.R. Whitman, B.Y. and Greenberg, F., Prader-Willi syndrome: consensus diagnostic criteria. Pediatrics, 1993. 91 : p. 398-402. 78. J. Clayton-Smith and M.E. Pembrey, Angelman syndrome. Med. Genet., 1992. 29 : p. 412-415. 79. J. Clayton Smith, D.J. Driscoll, M.F. Wa ters, T. Webb, T. Andrews, S. Malcolm, M.E. Pembrey, and R.D. Nicholls, Difference in methylation patterns within the D15S9 region of chromosome 15q11-13 in fi rst cousins with Angelman syndrome and Prader-Willi syndrome. Am. J. Med. Genet., 1993. 47 : p. 683-686. 80. J.H. Knoll, R.D. Nicholls, R.E. Magenis, J.M. Graham, Jr., M. Lalande, and S.A. Latt, Angelman and Prader-Willi syndromes share a common chromosome 15 deletion but differ in paren tal origin of the deletion. Am. J. Med. Genet., 1989. 32 : p. 285-290. 81. R.D. Nicholls, J.H. Knoll, K. Glatt, J.H. Hersh, T.D. Brewster, J.M. Graham, Jr., D. Wurster Hill, R. Wharton, and S.A. Latt, Restriction fragment length polymorphisms within proximal 15q and th eir use in molecular cytogenetics and the Prader-Willi syndrome. Am. J. Med. Genet., 1989. 33 : p. 66-77. 82. W.P. Robinson, A. Bottani, Y.G. Xie, J. Balakrishman, F. Binkert, M. Machler, A. Prader, and A. Schinzel, Molecular, cytogenetic, and clin ical investigations of Prader-Willi syndrome patients. Am. J. Hum. Genet., 1991. 49 : p. 1219-1234. 83. M.J. Mascari, W. Gottlieb, P.K. Rogan, M. G. Butler, D.A. Waller, J.A. Armour, A.J. Jeffreys, R.L. Ladda, and R.D. Nicholls, The frequency of uniparental disomy in Prader-Willi syndrome. Implic ations for molecular diagnosis. N. Engl. J. Med., 1992. 326 : p. 1599-1607. 84. J.H. Knoll, R.D. Nicholls, and M. Lalande, On the parental origin of the deletion in Angelman syndrome. Hum. Genet., 1989. 83 : p. 205-207.

PAGE 134

124 85. R.E. Magenis, S. Toth Fejel, L.J. Allen, M. Black, M.G. Brown, S. Budden, R. Cohen, J.M. Friedman, D. Kalousek, J. Zonana, and et al., Comparison of the 15q deletions in Prader-Willi and Angelman sy ndromes: specific regions, extent of deletions, parental origin, and clinical consequences. Am. J. Med. Genet., 1990. 35 : p. 333-349. 86. C.A. Williams, R.T. Zori, J.W. Stone, B.A. Gray, E.S. Cantu, and H. Ostrer, Maternal origin of 15q11-13 deletions in Angelman syndrome suggests a role for genomic imprinting. Am. J. Med. Genet., 1990. 35 : p. 350-353. 87. R.D. Nicholls, J.H. Knoll, M.G. Butler, S. Karam, and M. Lalande, Genetic imprinting suggested by maternal he terodisomy in nondeletion Prader-Willi syndrome. Nature, 1989. 342 : p. 281-285. 88. S. Malcolm, J. Clayton Smith, M. Nichol s, S. Robb, T. Webb, J.A. Armour, A.J. Jeffreys, and M.E. Pembrey, Uniparental paternal disomy in Angelman's syndrome. Lancet, 1991. 337 : p. 694-697. 89. R.D. Nicholls, G.S. Pai, W. Gottlieb, and E.S. Cantu, Paternal uniparental disomy of chromosome 15 in a child with Angelman syndrome. Ann. Neurol., 1992. 32 : p. 512-518. 90. T. Kishino, M. Lalande, and J. Wagstaff, UBE3A/E6-AP mutations cause Angelman syndrome. Nat. Genet., 1997. 15 : p. 70-73. 91. T. Matsuura, J.S. Sutcliffe, P. Fang, R. -J. Galjaard, Y.-H. Jiang, C.S. Benton, J.M. Rommens, and A.L. Beaudet, De novo truncating mutations in E6-AP ubiquitinprotein ligase gene (UBE 3A) in Angelman syndrome. Nat. Genet., 1997. 15 : p. 74-77. 92. P. Malzac, H. Webber, A. Moncla, J.M. Graham, M. Kukolich, C. Williams, R.A. Pagon, L.A. Ramsdell, T. Kishino, and J. Wagstaff, Mutation analysis of UBE3A in Angelman syndrome patients. Am. J. Hum. Genet., 1998. 62 : p. 1353-1360. 93. P. Fang, E. Lev-Lehman, T.F. Tsai, T. Matsuura, C.S. Benton, J.S. Sutcliffe, S.L. Christian, T. Kubota, D.J. Halley, H. Meij ers-Heijboer, S. Langlois, J.M. Graham, J. Beuten, P.J. Willems, D.H. Ledbetter, and A.L. Beaudet, The spectrum of mutations in UBE3A causing Angelman syndrome. Hum. Mol. Genet., 1999. 8 : p. 129-135. 94. A. Moncla, P. Malzac, M.O. Liver, M.A. Voelckel, J. Mancini, J.C. Delaroziere, N. Philip, and J.F. Mattei, Angelman syndrome resulting from UBE3A mutations in 14 patients fom eight families: clinical manifestations and genetic counselling. J. Med. Genet., 1999. 36 : p. 554-560.

PAGE 135

125 95. S. Russo, F. Cogliati, M. Viri, F. Cavalleri , A. Selicorni, L. Turolla, S. Belli, A. Romeo, and L. Larizza, Novel mutations of ubiquitin protein ligase 3A gene in Italian patients with Angelman syndrome. Hum. Mutat., 2000. 15 : p. 387. 96. A. Baumer, D. Balmer, and A. Schinzel, Screening for UBE3A gene mutations in a group of Angelman sydrome patients se lected according to non-stringent clinical criteria. Hum. Genet., 1999. 105 : p. 598-602. 97. R.D. Nicholls, New insights reveal complex mechanisms involved in genomic imprinting. Am. J. Hum. Genet., 1994. 54 : p. 733-740. 98. M.T. Jong, T.A. Gray, Y. Ji, C.C. Gl enn, S. Saitoh, D.J. Driscoll, and R.D. Nicholls, A novel imprinted gene , encoding a RING zinc-finger protein, and overlapping antisense transcript in th e Prader-Willi syndrome critical region. Hum. Mol. Genet., 1999. 8 : p. 783-793. 99. I. Boccaccio, H. Glatt-Deeley, F. Watrin, N. Roeckel, M. Lalande, and F. Muscatelli, The human MAGEL2 gene and its mouse homologue are paternally expressed and mapped to the Prader-Willi region. Hum. Mol. Genet., 1999. 8 : p. 2497-2505. 100. S. Lee, Kozlov, S., Hernandez, L. Chamberlain, S.J., Brannan, C.I., Stewart, C.L., and Wevrick, R., Expression and imprinting of MA GEL2 suggest a role in Prader-Willi syndrome and the homologous murine imprinting phenotype. Hum. Mol. Genet., 2000. 9 : p. 1813-1819. 101. H.R. MacDonald and R. Wevrick, The necdin gene is deleted in Prader-Willi syndrome and is imprinted in human and mouse. Hum. Mol. Genet., 1997. 6 : p. 1873-1878. 102. R. Wevrick, J.A. Kerns, and U. Francke, Identification of a novel paternally expressed gene in the Prader-Willi syndrome region. Hum. Mol. Genet., 1994. 3 : p. 1877-1882. 103. J. Cavaille, K. Buiting, M. Kiefmann, M. Lalande, C.I. Brannan, B. Horsthemke, J.P. Bachellerie, J. Brosius, and A. Huttenhofer, Identification of brain-specific and imprinted small nucleolar RNA genes exhibiting an unusual genomic organization. Proc. Natl. Acad. Sci. USA, 2000. 97 : p. 14311-14316. 104. T. de los Santos, J. Schweizer, C.A. Rees, and U. Francke, Small evolutionarily conserved RNA, resembling C/D box small nucleolar RNA, is transcribed from the PWCR1, a novel imprinted gene in th e Prader-Willi deleted region, which is highly expressed in the brain. Am. J. Hum. Genet., 2000. 67 : p. 1067-1082.

PAGE 136

126 105. M. Meguro, K. Mitsuya, N. Nomura, M. K ohda, A. Kashiwagi, R. Nishigaki, H. Yoshioka, M. Nakao, M. Oishi, and M. Oshimura, Large-scale evaluation of imprinting status in the Pr ader-Willi syndrome region: an imprinted direct repeat cluster resembling small nucleolar RNA genes. Hum. Mol. Genet., 2001. 10 : p. 383-394. 106. M. Runte, A. Huttenhofer, S. Gross, M. Kiefmann, B. Horsthemke, and K. Buiting, The IC-SNURF-SNRPN transcript serves as a host for multiple small nucleolar RNA species and as an antisense RNA for UBE3A. Hum. Mol. Genet., 2001. 10 : p. 2687-2700. 107. C. Rougeulle, C. Cardoso, M. Fontes, L. Colleaux, and M. Lalande, An imprinted antisense RNA overlaps UBE3A and a s econd maternally expressed transcript. Nat. Genet., 1998. 19 : p. 15-16. 108. T.H. Vu and A.R. Hoffman, Imprinting of the Angelman syndrome gene, UBE3A, is restricted to brain. Nat. Genet., 1997. 17 : p. 12-13. 109. C. Rougeulle, H. Glatt, and M. Lalande, The Angelman syndrome candidate gene, UBE3A/E6-AP, is imprinted in brain. Nat. Genet., 1997. 17 : p. 14-15. 110. L.B. Herzing, S.J. Kim, E.H. Cook, Jr., and D.H. Ledbetter, The human aminophospholipid-transporting ATPase gene ATP10C maps adjacent to UBE3A and exhibits similar imprinted expression. Am. J. Hum. Genet., 2001. 68 : p. 15011505. 111. M. Meguro, A. Kashiwagi, K. Mitsuya, M. Nakao, I. Kondo, S. Saitoh, and M. Oshimura, A novel maternally expressed gene, ATP10C, encodes a putative aminophospholipid translocase asso ciated with Angelman syndrome. Nat. Genet., 2001. 28 : p. 19-20. 112. K. Buiting, S. Saitoh, S. Gross, B. Dittri ch, S. Schwartz, R.D. Nicholls, and B. Horsthemke, Inherited microdeletions in the Angelman and Prader-Willi syndromes define an imprinting centre on human chromosome 15. Nat. Genet., 1995. 9 : p. 395-400. 113. J. Burger, K. Buiting, B. Dittrich, S. GroB, C. Lich, K. Sperling, B. Horsthemenke, and A. Reis, Different mechanisms and recurrence risks of imprinting defects in Angelman syndrome. Am. J. Hum. Genet., 1997. 61 : p. 8893.

PAGE 137

127 114. K. Buiting, B. Dittrich, S. Gross, C. Lic h, C. Farber, T. Buchholz, E. Smith, A. Reis, J. Burger, M.M. Nothen, U. Bart h-Witte, B. Janssen, D. Abeliovich, I. Lerer, A.M.W. van den Ouweland, D.J.J. Halley, C. Schrander-Stumpel, H. Smeets, P. Meinecke, S. Malcolm, A. Gardner, M. Lalande, R.D. Nicholls, K. Friend, A. Schulze, G. Matthijs, H. Kokkonen, P. Hilbert, L. Van Maldergerm, G. Glover, P. Carbonell, P. Willems, G. Gillessen-Kaesbach, and B. Horsthemke, Sporadic imprinting defects in Prader-W illi syndrome and Angelman syndrome: Implications for imprin t-switch models, genetic counseling, and prenatal diagnosis. Am. J. Hum. Genet., 1998. 67 : p. 170-180. 115. R.D. Nicholls, S. Saitoh, and B. Horsthemke, Imprinting in Prader-Willi and Angelman syndromes. Trends Genet., 1998. 14 : p. 194-200. 116. B. Horsthemke, B. Dittrich, and K. Buiting, Imprinting mutations on human chromosome 15. Hum. Mutat., 1997. 10 : p. 329-337. 117. J.S. Sutcliffe, M. Nakao, S. Christ ian, K.H. Orstavik, N. Tommerup, D.H. Ledbetter, and A.L. Beaudet, Deletions of a differential ly methylated CpG island at the SNRPN gene define a putat ive imprinting control region. Nat. Genet., 1994. 8 : p. 52-58. 118. A. Reis, B. Dittrich, V. Greger, K. Bu iting, M. Lalande, G. Gillessen Kaesbach, M. Anvret, and B. Horsthemke, Imprinting mutations suggested by abnormal DNA methylation patterns in familia l Angelman and Prader-Willi syndromes. Am. J. Hum. Genet., 1994. 54 : p. 741-747. 119. S. Saitoh, K. Buiting, P.K. Rogan, J.L. Buxton, D.J. Driscoll, J. Arnemann, R. Konig, S. Malcolm, B. Horsthemke, and R.D. Nicholls, Minimal definition of the imprinting center and fixation of chromosome 15q11-q13 epigenotype by imprinting mutations. Proc. Natl. Acad. Sci. USA, 1996. 93 : p. 7811-7815. 120. T. Ohta, T.A. Gray, P.K. Rogan, K. Buiting, J.M. Gabriel, S. Saitoh, B. Muralidhar, B. Bilienska, M. Krajewska-Walasek, D.J. Driscoll, B. Horsthemke, M.G. Bulter, and R.D. Nicholls, Imprinting-mutation mec hanisms in Prader-Willi syndrome. Am. J. Hum. Genet., 1999. 64 : p. 397-413. 121. K. Buiting, C. Lich, S. Cottrell, A. Barnicoat, and B. Horsthemke, A 5-kb imprinting center deletion in a fami ly with Angelman syndrome reduces the shortest region of deletion overlap to 880 bp. Hum. Genet., 1999. 105 : p. 665666. 122. M.T. Jong, A.H. Carey, K.A. Caldwell, M.H. Lau, M.A. Handel, D.J. Driscoll, C.L. Stewart, E.M. Rinchik, and R.D. Nicholls, Imprinting of a RING zinc-finger encoding gene in the mouse chromosome region homologous to the Prader-Willi syndrome genetic region. Hum. Mol. Genet., 1999. 8 : p. 795-803.

PAGE 138

128 123. S.J. Chamberlain and C.I. Brannan, The Prader-Willi syndrome imprinting center activates the paternally expressed mu rine Ube3a antisense transcript but represses paternal Ube3a. Genomics, 2001. 73 : p. 316-322. 124. U. Albrecht, J.S. Sutcliffe, B.M. Cattanach, C.V. Beechey, D. Armstrong, G. Eichele, and A.L. Beaudet, Imprinted expression of the murine Angelman syndrome gene, Ube3a, in hippocampal and Purkinje neurons. Nat. Genet., 1997. 17 : p. 75-78. 125. A. Kashiwagi, M. Meguro, H. Hoshiya, M. Haruta, F. Ishino, T. Shibahara, and M. Oshimura, Predominant maternal express ion of the mouse Atp10c in hippocampus and olfactory bulb. J. Hum. Genet., 2003. 48 : p. 194-198. 126. T. Kayashima, K. Yamasaki, K. Joh, T. Yamada, T. Ohta, K. Yoshiura, N. Matsumoto, Y. Nakane, T. Mukai, N. Niikawa, and T. Kishino, Atp10a, the mouse ortholog of the human imprin ted ATP10A gene, escapes genomic imprinting. Genomics, 2003. 81 : p. 644-647. 127. T. Kayashima, T. Ohta, N. Niikawa, and T. Kishino, On the conflicting reports of imprinting status of mouse ATP10a in the adult brain: strain-backgrounddependent imprinting? J. Hum. Genet., 2003. 48 : p. 492-493. 128. J.H. Chai, D.P. Locke, T. Ohta , J.M. Greally, and R.D. Nicholls, Retrotransposed genes such as Frat3 in the mouse chro mosome 7C PraderWilli syndrome region acquire the imprinted status of their insertion site. Mamm. Genome, 2001. 12 : p. 813-821. 129. S. Kobayashi, T. Kohda, H. Ichikawa, A. Ogura, M. Ohki, T. Kaneko-Ishino, and F. Ishino, Paternal expression of a novel im printed gene, Peg12/Frat3, in the mouse 7c region homologous to th e Prader-Willi syndrome region. Biochem. Biophys. Res. Commun., 2002. 290 : p. 403-408. 130. C. Farber, S. GroB, J. Neesen, K. Buiting, and B. Horsethemke, Identification of a testis-specific gene (C15orf2) in the Prader-Willi syndrome region on chromosome 15. Genomics, 2000. 65 : p. 174-183. 131. B.M. Cattanach, J.A. Barr, C.V. Beechey, J. Martin, J. Noebels, and J. Jones, A candidate model for Angelman syndrome in the mouse. Mamm. Genome, 1997. 8 : p. 472-478. 132. T. Yang, T.E. Adamson, J.L. Resnick, S. Leff, R. Wevrick, U. Francke, N.A. Jenkins, N.G. Copeland, and C.I. Brannan, A mouse model for Prader-Willi syndrome imprinting-centre mutations. Nat. Genet., 1998. 19 : p. 25-31.

PAGE 139

129 133. B. Bielinska, S.M. Blaydes, K. Buiting, T. Yang, M. Krajewska-Walasek, B. Horsthemke, and C.I. Brannan, De novo deletions of SNRPN exon 1 in early human and mouse embryos result in a pat ernal to maternal imprint switch. Nat. Genet., 2000. 25 : p. 74-78. 134. O. El-Maarri, K. Buiting, E.G. Peery, P.M. Kroisel, B. Balaban, K. Wagner, B. Urman, J. Heyd, C. Lich, C.I. Brannan, J. Walter, and B. Horsthemke, Maternal methylation imprints on human chromosome 15 are established during or after fertilization. Nat. Genet., 2001. 27 : p. 341-344. 135. C.I. Brannan and M.S. Bartolomei, Mechanisms of genomic imprinting. Curr. Opin. Genet. Dev., 1999. 9 : p. 164-170. 136. C. Färber, B. Dittrich, K. Buiting, and B. Horsethemke, The chromosome 15 imprinting centre (IC) region has unde rgone multiple duplication events and contains an upstream exon of SNRPN that is deleted in all Angelman syndrome patients with an IC microdeletion. Hum. Mol. Genet., 1999. 8 : p. 337-343. 137. P.J. Swiatek and T. Gridley, Perinatal lethality and defects in hindbrain development in mice homozygous for a target ed mutation of the zinc finger gene Krox20. Genes Dev., 1993. 7 : p. 2071-2084. 138. M. Runte, P.M. Kroisel, G. GillessenKaesbach, R. Varon, D. Horn, M.Y. Cohen, J. Wagstaff, B. Horsthemke, and K. Buiting, SNURF-SNRPN and UBE3A transcript levels in patient s with Angelman syndrome. Hum. Genet., 2004. 114 : p. 553-561. 139. Y.H. Jiang, D. Armstrong, U. Albrecht, C. M. Atkins, J.L. Noebels, G. Eichele, J.D. Sweatt, and A.L. Beaudet, Mutation of the Angelman ubiquitin ligase in mice causes increased cytoplasmic p53 and de ficits of context ual learning and longterm potentiation. Neuron, 1998. 21 : p. 799-811. 140. K. Miura, T. Kishino, E. Li, H. Webber, P. Dikkes, G.L. Holmes, and J. Wagstaff, Neurobehavioral and electroencephalographi c abnormalities in Ube3a maternaldeficient mice. Neurobiol. Dis., 2002. 9 : p. 149-159. 141. S.J. Clark, J. Harrison, C.L. Paul, and M. Frommer, High sensitivity mapping of methylated cytosines. Nucl. Acids Res., 1994. 22 : p. 2990-2997. 142. C.C. Glenn, S. Saitoh, M.T.C. Jong, M.M. Filbrandt, U. Surti, D.J. Driscoll, and R.D. Nicholls, Gene structure, DNA methylation, and imprinted expression of the human SNRPN gene. Am. J. Hum. Genet., 1996. 58 : p. 335-346. 143. S.M. Blaydes, M. Elmore, T. Yang, and C.I. Brannan, Analysis of murine Snrpn and human SNRPN gene imprinting in transgenic mice. Mamm. Genome, 1999. 10 : p. 549-555.

PAGE 140

130 144. R. Shemer, Y. Birger, A.D. Riggs, and A. Razin, Structure of the imprinted mouse Snrpn gene and establishment of its par ental-specific methylation pattern. Proc. Natl. Acad. Sci. USA, 1997. 94 : p. 10267-10272. 145. P. Hajkova, S. Erhardt, N. Lane, T. H aaf, O. El-Maarri, W. Reik, J. Walter, and M.A. Surani, Epigenetic reprogramming in m ouse primordial germ cells. Mech. Dev., 2002. 117 : p. 15-23. 146. E.K. Wakeland, L. Morel, K. Achey, M. Yui, and J. Longmate, Speed congenics: A classic technique in the fast lane (relatively speaking). Immunol. Today, 1997. 18 : p. 472-477. 147. S.J. Chamberlain, K.A. Johnstone, A.J. DuBose, T.A. Simon, M.S. Bartolomei, J.L. Resnick, and C.I. Brannan, Evidence for genetic modifiers of postnatal lethality in PWS-IC deletion mice. Hum. Mol. Genet., 2004. 13 : p. 2971-2977. 148. D. Lucifero, C. Mertineit, H.J. Cl arke, T.H. Bestor, and J.M. Trasler, Methylation dynamics of imprinted gene s in mouse germ cells. Genomics, 2002. 79 : p. 530538. 149. K.S. Chen, T.F. Tsai, and A.L. Beaudet, An insertion/duplication mutation 11 kb upstream of Snurf-Snrpn produces a m ouse model of an Angelman syndrome imprinting mutation. Am. J. Hum. Genet., 2002. 71 : p. S175. 150. B. Dittrich, K. Buiting, B. Korn, S. Rick ard, J. Buxton, S. Saitoh, R.N. Nicholls, A. Poustka, A. Winterpacht, B. Zabel, and B. Horsthemke, Imprint switching on human chromosome 15 may involve alterna tive transcripts of the SNRPN gene. Nat. Genet.., 1996. 14 : p. 163-170. 151. C. Farber, B. Dittrich, K. Buiting, and B. Horsethemke, The chromosome 15 imprinting centre (IC) region has unde rgone multiple duplication events and contains an upstream exon of SNRPN that is deleted in all Angelman syndrome patients with an IC microdeletion. Hum. Mol. Genet., 1999. 8 : p. 337-343. 152. J. Bressler, T. Tsai, M. Ramir ez, D. Armstrong, and A. Beaudet, The SNPRN promoter is not required for genomic imprinting of the Prader-Willi/Angelman domain in mice. Nat. Genet., 2001. 28 : p. 232-240. 153. M. Landers, D.L. Bancescu, E. Le Me ur, C. Rougeulle, H. Glatt-Deeley, C. Brannan, F. Muscatelli, and M. Lalande, Regulation of the large (approximately 1000 kb) imprinted murine Ube3a antisense transcript by alternative exons upstream of Snurf/Snrpn. Nucleic Acids Res., 2004. 32 : p. 3480-3492. 154. R. Shemer, A.Y. Hershko, J. Perk, R. Mostoslavsky, B. Tsuberi, H. Cedar, K. Buiting, and A. Razin, The imprinting box of the Prader-Willi/Angelman syndrome domain. Nat. Genet., 2000. 26 : p. 440-443.

PAGE 141

131 BIOGRAPHICAL SKETCH Edwin Peery was born in Richmond, Vi rginia, in 1974. He graduated from Douglas Freeman High School in Richmond, Virg inia, in 1992. After receiving a BA in Biology and Environmental Sciences from the Un iversity of Virginia in Charlottesville, Virginia, in 1996, he enrolled as a graduate st udent in the Interdisciplinary Program in Biomedical Sciences at the University of Florida in 1997.