<%BANNER%>

Minimal Sequence of the Prader-Willi Imprinting Center and Affected Genes

Permanent Link: http://ufdc.ufl.edu/UFE0024866/00001

Material Information

Title: Minimal Sequence of the Prader-Willi Imprinting Center and Affected Genes
Physical Description: 1 online resource (90 p.)
Language: english
Creator: Dubose, Amanda
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: angelman, atp10a, center, epigenetic, imprinting, prader, syndrome, ube3a, willi
Genetics (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are genetic disorders that can result from a loss of a cluster of oppositely imprinted genes. The orthologous region has been found in mouse, and mouse models have been used to examine imprinting in the PWS/AS region. The PWS/AS region is controlled by a bipartite imprinting center consisting of the PWS imprinting center (PWS-IC) and the AS imprinting center (AS-IC). Atp10a is located close to the PWS/AS region and there is some data to support the idea that Atp10a may also be imprinted and regulated as a part of the PWS/AS cluster of genes. Atp10a encodes a putative phospholipid translocase and is thought to possibly contribute to the AS phenotype. We investigated the imprinting status of Atp10a in mouse by examining differences in allelic gene expression using a single nucleotide polymorphism between mouse strains. We found that Atp10a is not imprinted in all examined regions of the mouse brain and the PWS-IC has no effect on Atp10a imprinting. We also examined CpG methylation in a CpG island associated with Atp10a and found no differential methylation supporting our result that Atp10a is not imprinted. Although the human PWS-IC has been narrowed down to 4.3 kb including Snrpn exon 1 by mapping microdeletions found in patients with PWS, the smallest region known to cause a complete imprinting defect in mouse is 35 kb. We have created a new mouse model that defines the limits of the PWS-IC to 6 kb, and allows for selective deletion of the PWS-IC using cre/loxp technology. Using crosses between this new mouse model, PWS-IC^delta6kb, and a ubiquitously expressing transgenic cre mouse line, CMV-cre, we were able to create mice with a paternally derived widespread deletion of the PWS-IC. These mice show a postnatal lethality phenotype and lack expression of normally paternally expressed PWS genes similar to what is observed in the 35 kb PWS-IC deletion mice. This new mouse model narrows the known boundaries of the mouse PWS-IC and opens up novel avenues for research on PWS and imprinting.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Amanda Dubose.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Resnick, James L.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024866:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024866/00001

Material Information

Title: Minimal Sequence of the Prader-Willi Imprinting Center and Affected Genes
Physical Description: 1 online resource (90 p.)
Language: english
Creator: Dubose, Amanda
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: angelman, atp10a, center, epigenetic, imprinting, prader, syndrome, ube3a, willi
Genetics (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are genetic disorders that can result from a loss of a cluster of oppositely imprinted genes. The orthologous region has been found in mouse, and mouse models have been used to examine imprinting in the PWS/AS region. The PWS/AS region is controlled by a bipartite imprinting center consisting of the PWS imprinting center (PWS-IC) and the AS imprinting center (AS-IC). Atp10a is located close to the PWS/AS region and there is some data to support the idea that Atp10a may also be imprinted and regulated as a part of the PWS/AS cluster of genes. Atp10a encodes a putative phospholipid translocase and is thought to possibly contribute to the AS phenotype. We investigated the imprinting status of Atp10a in mouse by examining differences in allelic gene expression using a single nucleotide polymorphism between mouse strains. We found that Atp10a is not imprinted in all examined regions of the mouse brain and the PWS-IC has no effect on Atp10a imprinting. We also examined CpG methylation in a CpG island associated with Atp10a and found no differential methylation supporting our result that Atp10a is not imprinted. Although the human PWS-IC has been narrowed down to 4.3 kb including Snrpn exon 1 by mapping microdeletions found in patients with PWS, the smallest region known to cause a complete imprinting defect in mouse is 35 kb. We have created a new mouse model that defines the limits of the PWS-IC to 6 kb, and allows for selective deletion of the PWS-IC using cre/loxp technology. Using crosses between this new mouse model, PWS-IC^delta6kb, and a ubiquitously expressing transgenic cre mouse line, CMV-cre, we were able to create mice with a paternally derived widespread deletion of the PWS-IC. These mice show a postnatal lethality phenotype and lack expression of normally paternally expressed PWS genes similar to what is observed in the 35 kb PWS-IC deletion mice. This new mouse model narrows the known boundaries of the mouse PWS-IC and opens up novel avenues for research on PWS and imprinting.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Amanda Dubose.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Resnick, James L.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024866:00001


This item has the following downloads:


Full Text

PAGE 1

1 MINIMAL SEQUENCE OF THE MURINE PRADER-WILLI IMPRINTING CENTER AND AFFECTED GENES By AMANDA JEANETTE DUBOSE 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 2009

PAGE 2

2 2009 Amanda J. DuBose

PAGE 3

3 To my family and friends

PAGE 4

4 ACKNOWLEDGMENTS First and forem ost, Id like to thank my family for their love and support. Im thankful for my Dad who sparked my interest in scie nce and the world around me (especially the entomological world). I have had many teach ers that have helped me on my journey, particularly my mentor and most supportive boss-man ever Jim Resnick, and Karen Johnstone, who knows everything even if she is Scottish and nobody can understand what shes saying. Im grateful for my all my labmates including my frosty buddy and the bisu lfite queen, Emily Smith and the most awesome undergrad ever, Ryan Hall ett. Im indebted to my friends for their encouragement and for keeping me sane through this purgatory called grad school, especially Amber Shatzer. This list of acknowledgments woul d not be complete without a grateful tip of the hat to Camilynn Brannan, whom I unfortunately never got to meet but whose presence is still in the lab today. Without her work, ideas and the wonderful people who worked with her, my work would not have been possible.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF FIGURES .........................................................................................................................7ABSTRACT ...................................................................................................................... ...............9 CHAP TER 1 INTRODUCTION .................................................................................................................. 11 The Discovery of Imprinting .................................................................................................. 11Imprinting: When and Why? .................................................................................................. 12Features of Imprinted Genes and Gene Clusters ....................................................................14Angelman Syndrome and Prader-Willi Syndrome ................................................................. 17The PWS/AS Locus .............................................................................................................. ..18Imprinting in the PWS/AS Locus ........................................................................................... 19The Evolution of the PWS/AS Locus .....................................................................................20Mouse Models ........................................................................................................................20Individual Gene Mutations of the PWS/AS Locus ..........................................................21PWS-IC Mouse Models ...................................................................................................23AS-IC Mouse Model .......................................................................................................25Transgenic Mouse Models ..............................................................................................26Other Mouse Models of PWS and AS ............................................................................. 26ATP10A ...................................................................................................................................27 2 ATP10A MATERIALS AND METHODS ............................................................................. 33Mouse lines and Tissues utilized ............................................................................................ 33Reverse Transcriptase PCR (RT-PCR) ................................................................................... 33Atp10a CpG Island Characterization ......................................................................................34 3 ATP10A RESULTS ................................................................................................................36Atp10a Expression in C57BL/6 and B6.CAST.c7 crosses ..................................................... 36Effect of the PWS-IC on Atp10a Expression .........................................................................37PWS-IC Deletion Effects ................................................................................................37PWS-ICHs Effects ............................................................................................................37Atp10a CpG Island Methylation Analysis ......................................................................39Atp10a Imprinting Status Summary ....................................................................................... 39 4 MOUSE MODELS MATERIALS AND METHODS ........................................................... 53Strategy ...................................................................................................................... .............53Targeting Construct Assembly ............................................................................................... 55

PAGE 6

6 Transformation of ES Cells ....................................................................................................57Analysis of ES Cell Clones..................................................................................................... 59PWS-IC 6kb .............................................................................................................................61PWS-ICflox6kb ..........................................................................................................................62PWS-ICflox6kb Analysis by RT-PCR ................................................................................ 63PWS-ICflox6kb Analysis by Northern Blot ........................................................................ 63PWS-ICflox6kb Analysis of Postnatal Phenotype .............................................................. 65 5 MOUSE MODELS RESULTS ............................................................................................... 74 6 DISCUSSION AND CONCLUSIONS ..................................................................................80The Imprinting Status of Atp10a ............................................................................................80The PWS-ICflox6kb Mouse .......................................................................................................81LIST OF REFERENCES ...............................................................................................................83BIOGRAPHICAL SKETCH .........................................................................................................90

PAGE 7

7 LIST OF FIGURES Figure page 1-1 Illustration of pattern s of gene expression.. ....................................................................... 301-2 Map of the murine PWS/AS locus (not to scale). ..............................................................311-3 Detailed maps of deletions near Snrpn exon 1...................................................................323-1 Atp10a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 crossed with C57BL/ 6 and reciprocal crosses. ..............................................423-2 Ube3a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 crossed with C57BL/ 6 and reciprocal crosses. ..............................................433-3 Atp10a sequencing chromatograms fo r 2 brain regions and whole brain from P1 mice from a B6.CASTc7 female and a PWS-IC 35kb male cross. .............................................. 443-4 Ube3a sequencing chromatograms for 2 brain regions and whole brain from P1 mice from a B6.CASTc7 female and a PWS-IC 35kb male cross. .............................................. 453-5 Atp10a s equencing chromatograms of 4 brain re gions of 8 week old mice from PWSICHs x B6.CASTc7. ............................................................................................................463-6 Ube3a s equencing chromatograms of 4 brain re gions of 8 week old mice from PWSICHs x B6.CASTc7.. ...........................................................................................................473-7 Atp10a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 x PWS-ICHs. ..................................................................................................483-8 Ube3a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 x PWS-ICHs. ..................................................................................................493-9 Atp10a sequencing chromatograms of whole brain from P1 mice from a PWS-ICHs x B6.CASTc7 and reciprocal cross. ......................................................................................503-10 Ube3a sequencing chromatograms for whole brain from P1 mice from a PWS-ICHs x B6.CASTc7 and reciprocal cross. ......................................................................................513-11 The methylation state the Atp10a CpG island in newborn cerebral cortex and olfactory bulb. ....................................................................................................................524-1 ES cell targeting schematic. ...............................................................................................664-2 3end screening for targeted cells. .....................................................................................674-3 Southern blot of the 3end of targeted ES cells. ................................................................68

PAGE 8

8 4-4 5end screening for targeted cells. .....................................................................................694-5 Southern blot of the 5end of targeted ES cells. ................................................................704-6 ES Cell PCR screening for the conditional allele. ............................................................. 714-7 ES Cell screening for the deletion allele. ...........................................................................724-8 Identification of ES cell contribution by coat color. ..........................................................735-1 Expression of normally paternally expresse d genes is missing in the brains of CMVcre tg/-, PWS-ICflox6kb +/flox pups.. ................................................................................... 765-2 CMV-cre tg/-, PWS-ICflox6kb +/flox pups are distinguish able from wild type. .................. 775-3 CMVcre tg/-, PWS-ICflox6kb +/flox pups typically weigh less at birth than their littermates. .................................................................................................................. ........785-4 Reduced survival of CMVcre tg/-, PWS-ICflox6kb +/flox pups compared to littermates. ... 79

PAGE 9

9 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 MINIMAL SEQUENCE OF THE MURINE PR ADER-WILLI IMPRINTING CENTER AND AFFECTED GENES By Amanda Jeanette DuBose August 2009 Chair: James Resnick Major: Medical Sciences Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are genetic disorders that can result from a loss of a cluster of oppositely imprinted genes. The orthologous region has been found in mouse, and mouse models have been used to examine imprinting in the PWS/AS region. The PWS/AS region is controlled by a bi partite imprinting center consisting of the PWS imprinting center (PWS-IC) and the AS imprinting center (AS-IC). Atp10a is located close to the PWS/AS region a nd there is some data to support the idea that Atp10a may also be imprinted and regulated as a part of the PWS/AS cluster of genes. Atp10a encodes a putative phospholipid translocase and is thought to possibly contribute to the AS phenotype. We investigated the imprinting status of Atp10a in mouse by examining differences in allelic gene expression using a single nucleotide polymorphism between mouse strains. We found that Atp10a is not imprinted in all examined regions of the mouse brain and the PWS-IC has no effect on Atp10a imprinting. We also examined CpG methylation in a CpG island associated with Atp10a and found no differential methylat ion supporting ou r result that Atp10a is not imprinted. Although the human PWS-IC has been narrowed down to 4.3 kb including Snrpn exon 1 by mapping microdeletions found in patients with PWS, the smallest region known to cause a

PAGE 10

10 complete imprinting defect in mouse is 35 kb. We have created a new mouse model that defines the limits of the PWS-IC to 6 kb, and allows for selective deletion of th e PWS-IC using cre/loxp technology. Using crosses between this new mouse model, PWS-IC^delta6kb, and a ubiquitously expressing transgenic cre mouse line, CMV-cre, we were able to create mice with a paternally derived widespread deletion of the PWS-IC. These mice show a postnatal lethality phenotype and lack expression of normally paternally expressed PWS genes similar to what is observed in the 35 kb PWS-IC deletion mice. This new mouse model narrows the known boundaries of the mouse PWS-IC and opens up novel avenues for research on PWS and imprinting.

PAGE 11

11 CHAPTER 1 INTRODUCTION The Discovery of Imprinting In mammalian cells, there are genes that are expressed and genes that are silent depending on several factors includ ing cell type and developmental st age. Normally, a gene will be silent or expressed from both copies, but in a subset of genes, only on e parental allele is expressed while the other is sile nced (Fig. 1-1). This phenomenon of expression or silence of a gene depending on parental origin ha s been termed genomic imprinting. The discovery of imprinting began with th e realization from nuclear transplantation experiments in the early 1980s that the pare ntal genomes are inequivalent. Nuclear transplantation entails the replacement of one of the pronuclei of a one cell stage embryo with another pronucleus making it possible to create diploid embryos with only paternal or only maternal genomic contribution1. Studies of these manipulated embryos yielded curious results. Parental contribution from both a maternal pronucleus and a pate rnal pronucleus is required for normal development to term. One cell stage embryos that have had a pronucleus removed and replaced by a pronucleus of the same parental origin as th e removed pronucleus can develop into normal mice. This is not the case for embryos that are given 2 pronuclei of the same parental origin. Both androgenetic em bryos (from 2 male pronuclei) and gynogenetic embryos (from 2 female pronuclei) fail to deve lop normally and usually die around the time of implantation. The most developed embryo obs erved using this technique was a gynogenetic embryo at the 11 somite stage, though it was ab out 1 day developmentally behind and half the size of control embryos2,3. The androgenetic embryos had more than normal extraembryonic tissues including the trophoblast but a poorly developed embryo proper. Gynogenetic embryos

PAGE 12

12 showed the opposite phenotype w ith a relatively well develope d embryo but poor development of the extraembyonic tissues.4,5 More evidence supporting the inequivalence of the parental genomes, at least for some chromosomes and chromosomal regions, came from research on mice with chromosomal rearrangements. Studies of mice with Robertsonian translocations and complementary chromosome loss displayed a normal phenotype in mice with uniparental disomy for chromosomes 1, 3, 4, 5, 6 (paternal disomy), 9, 18, 14 and 15. Though some uniparental disomies have no effect, noncomplementation lethality was found for chromosomes 2, 6 (maternal disomy), 7, and 8 and an abnormal phenotype was found for uniparental disomy of chromosome 11 and a region of chromosome 26. The inequivalence of the genomes explains several observations in mammals. Although parthenogenic reproduction (reproduction from an unf ertilized egg) is normal for some animals, artificially created mammalian parthenotes are no t viable. At the time, the cause of death for these embryos was thought to be due to a lack of genetic contribution, a lack of cytoplasmic contribution or possibly a lack of some sort of stimulus from the spermatozoa. The lack of paternal genetic contribution is the deciding fact or in the inviability of parthogenotes, because nuclear transplant embryos with a male and fema le pronucleus are viable, while embryos created with 2 female pronuclei are inviable despite initial fertilization by a spermatozoa3. Imprinting: When and Why? Recent sequ encing of marsupial and monotreme genomes has shed light on the timing and evolution of imprinting. Monotremes diverged from the ancestors of marsupials and placental (eutherian) mammals 210 MYA wh ile marsupials and placental mammals diverged from each other 180 MYA. Imprinting is found in mammals and marsupials but not monotremes, so imprinting must have evolved sometime between 180 and 210 MYA7. By silencing one copy of

PAGE 13

13 a gene, imprinting renders affected genes func tionally haploid. Given the advantages of diploidy, it is a mystery that imprinting evolved and persisted. Severa l theories have been proposed to explain why imprinti ng evolved and what kind of a dvantage imprinting may confer. One of the theories that have been propose d, the ovarian time bomb hypothesis, theorizes that imprinting serves to protect the female fr om malignant germ cell tumors that can develop from parthenogenesis in unfertilized oocytes. Teratomas do still develop in females but are almost always benign due to imprinting according to this hypothesis. By silencing genes required to form the trophoblast, the female is protected from maligna nt trophoblastic disease from parthenogenesis, though she is still at risk during pregnancy8. The most obvious problem with this hypothesis is that it does not explain why some genes ar e paternally sile nced. Another weak point is that it does not explain why imprinting occurs at several loci when one locus would be sufficient to suppre ss the parthenogenic trophoblast9. The parent-offspring conflict model is a much more commonly accepted theory in which the parental genomes are viewed as competing against each other for selective advantage. It proposes that in polyandrous species it is in the be st interests of the father to produce the biggest, most fit offspring possible even at the expense of the mother and her future offspring (which may be fathered by a different male and thus unrelated to him). It is in the best interests of the mother to conserve limited resources and distribute them to many offspring, all of which will be equal in relatedness to her. The paternally expr essed imprinted genes would promote growth enhancement and recruitment of maternal resources both before and after birth, and the maternal expressed imprinted genes would be exp ected to promote growth restriction10. A logical extension of this theory is that imprinting would be absent or relaxed in monogamous species because there would be no selec tive advantage to imprinting. Studies done with the closely

PAGE 14

14 related rodent species Peromyscus polionotus and Peromyscus maniculatus seem to support the parent-offspring conflict model. P. polionotus is a monogamous species while P. maniculatus is polyandrous11,12. Hybrids from crosses between these two species show parent of origin growth defects with small hybrids from P. polionotus x P. maniculatus crosses (mother lis ted first, father listed second) and oversized hybrids from P. maniculatus x P. polionotus crosses13. These defects suggested a relaxation of imprinting in the monogamous species, and molecular studies do indeed show a loss of imprinting in the hybrids of several norma lly imprinted genes14. Features of Imprinted Ge nes and Gene Clusters There are several features that are co mmon to imprinted genes. Imprinted genes are often found clustered together and this region is coordinately regulate d. Several imprinted clusters contain noncoding RNA transcripts th at play a role in gene silenc ing. There are differences that have been observed between the two parent al alleles including as ynchronous replication15. DNAse hypersensitive sites are diffe rent between the parental alle les as well as CpG (cytosine and guanine dinucleotide pairs) me thylation and histone modifications16. There has been much speculation about what is the epigenetic mark. We know that the epigenetic mark must be capable of being erased and reset during gametogenesis, and capable of being maintained and recognized in all somatic cells where there is imprinted expression16. The top candidates for the epigenetic mark are CpG methylation and hi stone modification (and variants). CpGs are capable of being methylated on th eir cytosine residues, and stretches of DNA with a high number of CpGs (called CpG is lands) have been found to be differentially methylated at some imprinted genes. Frequent ly found at promoter regions, these CpG islands are usually (but not always) methylated at the inactive parental allele. Islands that show differences in methylation between parental alleles are called diffe rentially methylated domains,

PAGE 15

15 or DMDs (also called imprint control regions (ICRs) or differentially methylated regions (DMRs)). The importance of DNA methylation to imprinting is apparent in mice harboring deletions of Dnmt1 Dnmt1 encodes a DNA methyltransferas e responsible for maintaining methylation by methylating hemi methylated DNA. Methylati on is lost through successive rounds of DNA replication without this maintena nce methyltransferase. Mice homozygous for a knockout of Dnmt1 show stunted growth and midg estation embryonic lethality. Dnmt1 homozygous knockout embryos also show biallelic expression of several normally imprinted genes17. Methylation at CpG dinucleotides can effect gene regulation in several ways. DNA methylation can cause silencing or activati on through methyl CpG bi nding proteins or by blocking binding proteins that are sensitive to D NA methylation. For example, in the case of the oppositely imprinted genes Igf2 and H19 the CpG island acts as an insulator element. CCCTCbinding factor (CTCF) binds the unmethylated maternal allele and blocks enhancers from interacting with the Igf2 promoter. The enhancers instead interact with the H19 promoter. On the paternal allele, DNA methyl ation blocks CTCF binding an d enhancers interact with Igf218. It has been theorized that the differential DNA methylation found in imprinted genes developed as an extension of the role of DNA methylati on in host defense19. Transposons insert themselves into the genome and can cause muta tions from insertion, pr omote chimeric mRNA transcription or promote transcription of antisense transcripts20. Efficient silencing of invading elements is essential, and methyl ation is the tool used for silenc ing. The differential methylation found between parental alleles of imprinted genes may have its root s in this ancient self-defense system against invading DNA.

PAGE 16

16 Histone tail modifications are also important epig enetic marks. The first step in packaging DNA into higher order structure is the formation of the nucle osome consisting of DNA wrapped around a core of 8 histone proteins. The histone core is made up of 2 each of H4, H3, H2A and H2B. The tails of the histones, particularly th e amino terminal tails, ar e subject to modification such as ubiquitination, acetylation, sumoylati on, methylation and phosporylation at specific residues21. These modifications and combinations of modifications make up the histone code22. Proteins are able to bind in a modification dependent (or lack of modification dependent) manner and are able to activate or repr ess transcription. For example, H3K9 (histone 3, lysine 9) methylation is associated with silencing while H3K4 (histone 3, lysi ne 4) methylation is associated with active chromatin23,24. In addition to histone tail modification, incorporation of histone variants have an impact on the transcriptional activity of bound DNA25. CpG methylation and histone m odification are known to be in terdependent processes. DNA methylation attracts methyl binding proteins, which in tu rn attract complexes that modify histone tails. For example, MeCP2 is a met hyl CpG binding protein th at binds methylated DNA and is associated with a corepresso r complex including histone deacetylases26. Noncoding RNAs are common in imprinted clusters, and of particular interest are long RNA transcripts with a role in gene silencing or activation. A well studied example of this can be found in X inactivation. It must be noted th at imprinted X chromosome inactivation has not been observed in humans, though it has been observed in mouse extraembryonic tissues. Xist is a large noncoding RNA that is transcribed fr om the future inactive X chromosome. Xist coats the future inactive X chromosome, while the active X chromosome produces Tsix which is an antisense RNA transcribed across the Xist gene Expression of Tsix on the active X chromosome prevents Xist transcription in cis27. There are several theories on exactly how an antisense

PAGE 17

17 transcript works to silence genes in cis It is possible that the an tisense works through an RNAi mechanism, or that transcription in and of itself is sufficient to cause silencing. Epigenetic reprogramming normally happens during early development and gametogenesis. After fertilization, the sper m chromatin is remodeled by replacement of protamines by acetylated histones. The pate rnal genome also undergoes genomewide active demethylation. The maternal genome and patern al genome are then passively demethylated through DNA replication. Imprinted genes that ar e methylated are spared from genomewide demethylation. Remethylation occurs around im plantation, although there are lower levels of remethylation in the extraembryonic tissues than in the embryo proper28. Dnmt 3a and Dnmt 3b are required for this de novo methylation that occu rs during embryogenesis29. Primordial germ cells, or PGCs, undergo a resetting of the parental methlylation pattern to the pattern that is appropriate for the sex of the embryo. PGCs undergo genome wide demethylation that is completed by E13-14 in mouse. At around the same time as the demethylation event, the PGCs enter the developing gonad. Remethylation takes pl ace at different times in the male and female germ cells. In male germ cells, remethylation occurs at E15-16 and up while in female germ cells, remethylation does not occur until after birth28. Angelman Syndrome and Prader-Willi Syndrome Angelm an Syndrome (AS) and Prader-Willi S yndrome (PWS) are related yet distinct neurodevelopmental disorders. Both disorders are caused by a lack of proper gene expression from an imprinted domain on 15q11-13. Across this 2 megabase region, there are several paternally expressed genes and at least one maternally expressed gene. PWS is caused by a lack expression of the paternally expressed genes in the imprinted domain, while AS is caused by a lack of maternal gene expression from the same region.

PAGE 18

18 AS is a disorder characterized by severe ment al retardation, developm ental delay, absent or nearly absent speech, ataxic gait and happy affect with inappropriate laughter30. Mutations in UBE3A are sufficient to cause AS31,32 but AS is most commonly caused by a large deletion of about 4 megabases of maternal chromosome 15 q11-q13. There are three common breakpoints which contain duplications of a gene named HERC 2, and it is thought that unequal recombination between these homologous sequences cause the large deletions. BP1 (breakpoint 1) and BP2 are on the centromeric side of the PW S/AS cluster and BP3 is on the telomeric side of the cluster. Deletions typically span from BP1 to BP3 or from BP2 to BP333. Other causes of AS include paternal uniparental disomy of chromoso me 15, microdeletions, or epimutations. Some patients with AS have no identifiable defect34. PWS patients have hypotonia, early failure to thrive followed by hyperphagia beginning at around 2 years of age, mild to moderate mental retardation, obesity and hypogonadism. Other features include small hands and feet, obsessive-compulsive behavior, and some patients have hypopigmentation or seizures33. The most common cause of PWS is a large deletion on paternal chromosome 15, the same region as the most common deletion causing AS, although in AS the deletion on the maternal chromosome 15. PWS can also be caused by paternal uniparental disomy or microdeletions. The PWS/AS Locus The PW S/AS locus includes several paternally expressed genes and at least one maternally expressed gene. The paternally expressed genes can be thought of as bein g in two clusters, the upstream cluster (upstream of SNURF/SNRPN ) and the downstream cluster. The upstream cluster consists of NECDIN, MKRN3 and MAGEL2. From a patient with a paternal deletion of the upstream gene cluster, is known that the loss of these genes are insufficient to cause PWS, though the patient did show some PWS-like features35. Both NECDIN and MAGEL2 encode

PAGE 19

19 proteins in the MAGE protein family, while MKRN3 encodes a RING zinc-finger protein33. There is a long transcript called SNURF/SNRPN that extends through the downstream genes and through UBE3A ( UBE3A-AST or antisense transcript). There are several alternative exons upstream of SNRPN exon 1 that splice to exon 2, and their role is still under investigation36. SNRPN encodes both SNURF ( SNRPN upstream reading frame) and SmN which is a part of snRNPs (small nuclear ribonucleoprotein particles) found in spliceosomes. There are several paternally expressed C/D box snoR NAs that are spliced from the SNURF/SNRPN transcript. They include HBII-85, HBII-52 (both with multiple copies in tandem) HBII-13, HBII-436, HBII437, and HBII-438a/438b37. HBII-52 has been of particul ar interest because HBII-52 is predicted to bind to serotonin receptor 2C pr e-mRNA and appears to regulate its alternative splicing, though deletion of the HBII-52 cluster of snoRNAs is insufficient to cause PWS38,39. There has been one PWS patient identifie d with a paternal deletion including HBII-85 indicating that missing expression from this gene is suffi cient to cause PWS, however the patient was atypical indicating that other genes c ontribute to the classical PWS phenotype40. UBE3A and ATP10A are the only genes in the PWS/AS locus that have been identified as being maternally expressed. UBE3A encodes ubiquitin ligase 3A and mutations in this gene are sufficient to cause AS, although other genes may be involved31,32. ATP10A has been thought to be another maternally expressed gene, although its imprin ted status has recently come into question41. Imprinting in the PWS/AS Locus Im printing in the 15q11-q15 cluster has been shown to be controlled by a bipartite imprinting center made up of the Angelman s yndrome imprinting center (AS-IC) and the PraderWilli syndrome imprinting center (PWS-IC)42. The imprinting centers have been identified by mapping the smallest region of deletion overlap (SRO ) in patients with imprinting defects. In humans, the PWS-SRO has been na rrowed down to 4.3 kb and includes SNRPN exon 143. The

PAGE 20

20 AS-SRO is 35 kb upstream of SNRPN exon 1 and is 880 bp long44. Our lab proposed the currently accepted "paternal only" model in which the two elements of the imprinting center act in cis42. The PWS-IC is a positive element that promotes both upstream and downstream gene expression on the paternally inherited chromoso me. The AS-IC acts as a negative element that silences the PWS-IC on the maternally inherited chromo some (Fig. 1-2). The Evolution of the PWS/AS Locus The PW S/AS region acquired imprinted stat us and its current human configuration recently in evolution. In marsupials (the ta mmar wallaby and the opossum) an ortholog to SNRPN was found in tandem with its homolog, SNRPNB Ube3a was also found in marsupials, but on a different chromosome from Snrpnb and Snrpn. In the platypus, the only ortholog found to genes in the PWS/AS region is Ube3a. Snrpnb was found in the platypus (a monotreme), but without the tandem duplication that created SNRPN Expression studies reveal non-imprinted expression of Snrpn in marsupials and non-imprinted expression of Ube3a in both marsupials and monotremes. NDN, MKRN3 and MAGEL2 each arose more recently in evolution by retrotransposition from their respective paralogs. All 3 of these genes are present in placental mammals, so must have existed in their current arrangement before place ntal radiation about 105 MYA45. Mouse Models The orthologous region to hum an 15q11-13 is located on mouse chromosome 7C; however there are a few differences between mouse and human. One notable difference is the addition in mouse of Frat3 to the paternally expressed upstream ge ne cluster indicating that chromosomal position is sufficient to impart imprinting46. Another difference is the presence of C15ORF2, an intronless gene biparentally expressed in the human testis, but not present in mouse47. Several of the snoRNAs, specifically HBII-437 and HBII-438a/438b, are also present in humans, but not in

PAGE 21

21 mouse48. Aside from these minor differences, the PWS/AS region has been largely conserved, making the mouse an excellent model for study. M ouse models have been used to discern the function of genes in the PWS/AS locus, to study th e imprinting centers and ev aluate their roles in regulating the region, and to model PWS and AS. Individual Gene Mutations of the PWS/AS Locus Individual gene m utations have been made in mice for several of the genes located in the PWS/AS locus. These deletions and null mutations have helped to identify what role (if any) the individual genes and gene products play, particularly in the PWS phenotype. Magel 2 is strongly expressed in the central nerv ous system and shows circadian rhythm in levels of expression. Mice with a paternal Magel 2 deletion show altered circadian rhythm, decreased food intake, and reduced male fertility. It is thought that Magel 2 may be involved in post translational processing of orexin: a hypothalamic neuropeptide49. Orexin positive neurons project into parts of the brai n involved in regulation of sleep feeding, wakefulness, and body temperature50. Mice inheriting a paternal Magel 2 deletion show a marked decrease in orexin levels and orexin positive neurons49. For the Snurf/Snrpn gene, two null mutations have been made that each eliminates one of the genes products while leaving the other intact. Mice that are null for SmN due to an insertion of a neomycin resistance cassette and deletion of Snrpn exon 6 and a portion of exons 5 and 7, show no obvious phenotype51. There is also no obvious phenot ype in mice with a null mutation of Snurf ( Snrpn upstream reading frame). These mice have a deletion of 6 kb including Snrpn exon 2 that leaves the downstream gene SmN in frame52. There have been several Necdin (or Ndn ) deletion or mutation mouse lines created with variable results. Of the 3 lines of Ndn mutation mice, 2 show PWS reminiscent phenotypes, while 1 shows no phenotype at all. Gerard et al. created a Ndn knockout that has a early

PAGE 22

22 postnatal lethality phenotype from respiratory failure. Though their diaphragms are normal, these mice fail to inflate their lungs properly. This is similar to one aspect of PWS in that PWS infants also show neonatal resp iratory distress. There was a higher incidence of postnatal lethality in male heterozygotes than in female heterozygotes (inheriting the deletion paternally) and a higher level of postnatal lethality on a C57/BL6 background than on an FVB background. Surviving animals that make it past the firs t 30 hours have normal weight and fertility53. The Ndn deficient mouse created by Muscatelli et al. is similar to the previously described Ndn mutant in that it too has a pos tnatal lethality phenotype, and surv ivors are fertile, viable and do not show obesity. They also found skin scraping behavior, improved spatia l learning, and a loss of oxytocin-expressing neurons in th e paraventricular nucleus; all findings that are observed in PWS patients54. In stark contrast, the Ndn null mouse created by Tsai et al. shows no phenotype. They did not observe any postnatal lethality or obesity55. It is possible th at there are strain specific differences between the various lines th at could account for differences in phenotype, but it is difficult to compare the mutants in part because Tsai et al. did not indicate the strain of mouse used for targeting or breeding56. There has been more interest recently into MBII-85 mutants because of one atypical PWS patient with a deletion including the HBII-85 cluster40. From mouse models, we have learned that the MBII-85 gene cluster is vital to postnatal growth. Ding et al. found that mice with a deletion of the over 40 copies of MBII-85 show postnatal growth reta rdation and hyperphagia (an increase of about 30% in food intake) although they did not show obesity. The MBII-85 deletion mice show an increase in ghrelin levels much like that observed in PWS individuals57. With a similar deletion mouse, Skryabin et al. also observed postnatal growth retardation. Unlike the deletion mouse just previously described, this MBII-85 deletion mouse has a low level of

PAGE 23

23 postnatal lethality depending on genetic backgr ound. No mention was made of the mouses weight or feeding behavior48. Null mutations of Ube3a have resulted in interesting mouse models of AS. One null mutation of Ube3a was found to have several AS-like f eatures upon maternal transmission such as seizures (induceable), impaired motor skills and a context-dependent learning. These mice also showed defects in long-term pot entiation and increased cytoplasmic p5358. Another mouse model for AS was created by a targeted inactivation of Ube3a by insertion of a lacZ reporter. This mouse shows increased beta galactocid ase activity in the hippocampal and cerebellar neurons upon maternal transmission of the allele. Maternal transmission of this allele results in impaired motor coordination and spatial learning, but not an observed increase in levels of p5359. Deletion of the other main cluster of snoRNAs, MBII-52 has no effect60. A lack of Mkrn3 also has no effect in mice49. PWS-IC Mouse Models There have been several m ouse lin es created with deletions around Snrpn exon 1 in efforts to model the imprinting center de letions found in humans (Fig.13). In 2001, the Beaudet lab created two deletion mice, one with a 0.9 kb dele tion and the other with a 4.8 kb deletion. The 0.9 kb deletion included the Snrpn promoter and exon 1, and the mouse showed no phenotype and no loss of imprinted gene expression upon pa ternal transmission. The 0.9 kb deletion mouse was important in that it allowed the discovery of the mouse Snrpn upstream exons, which splice into exon 2. The 4.8 kb deletion included the sa me region removed in the 0.9 kb deletion model and more of intron 1 and upstream sequence. Mi ce that inherited the 4.8 kb deletion paternally show 40-50% postnatal lethality and a partial imprinting defect61. Another deletion mouse missing Snrpn exon 2 through Ube3a exon 2 showed ~80 % postnatal lethality phenotype upon

PAGE 24

24 paternal transmission, but had no imprinting de fect indicating a gene in that region may contribute to the PWS phenotype, but the re gion is unnecessary for correct imprinting62. The only mouse model with a complete PWS -IC deletion phenotype has a deletion of 35 kb (originally published as a 42 kb deletion) extending through Snrpn exon 6. Mice inheriting the 35 kb deletion (called PWS-IC 35kb) paternally exhibit poor postnatal feeding and poor weight gain much like human PWS patients, an d die within 7 days on a C57/BL6 background. Survival can be achieved at lo w levels on other backgrounds, but curiously the mice inheriting the deletion paternally remain small a nd do not exhibit hyperphagia or obesity 51. The loss of the PWS-IC on the paternal allele has an effect on paternal expression on Ube3a. Pups inheriting this deletion paternally exhibit a loss of the Ube3a antisense transcript and an increase of Ube3a expression from the paternal allele63. Another mouse model that has been used to study imprinting in the region has a replacement of 6 kb of the murine PWS-IC region with a 6.9 kb re gion including the human PWS-IC. Upon maternal inheritance, this allele, PWS-ICHs, is able to acquire methylation in the oocyte but not able to maintain methylation and s ilencing in somatic cells. IC function seems to have diverged, and the human PWS-IC is unable to activate transcripti on of the upstream gene cluster. The human PWS-IC region does however retain SNRPN promoter function which drives transcription of the downstream genes. Mice inheriting PWS-ICHs maternally show expression of the normally paternally expressed downstr eam genes from their maternal chromosome, including the Ube3a antisense transcript. These mice lose imprinting at Ube3a exhibiting bialleleic Ube3a expression. They also show an obesity phenotype which is more prominent in the females. Mice inheriting the PWS-ICHs allele paternally have a postnatal lethality phenotype

PAGE 25

25 and growth retardation due to the lack of e xpression of the upstream cluster of paternally expressed genes64. AS-IC Mouse Model The AS-IC has proven to be m uch more difficult to locate in mice th an the PWS-IC region. The location of the AS-IC is not conserved between human and mouse. In humans, the AS-IC is located 35 kb upstream of SNRPN exon 1. Mice made by Edwin Peer y in our lab with deletions spanning from -37 kb to -24 kb or -37 kb to -29 kb upstream of Snrpn failed to show any phenotype upon maternal, paternal or biparental inheritance. At th e time of the creation of these mice the locations of the a lternative upstream exons of Snrpn were unknown65. The only targeted deletion mouse model fo r an AS-IC deletion shows incomplete penetrance. This mouse was engineered with a 80 kb deletion extending from -93 kb to -13 kb relative to Snrpn exon 1. Upon maternal inheritance of this allele, the mice showed variable levels of methylation at the Snrpn CpG island from a normal methyla tion pattern, to a partial loss of methylation, to a complete lo ss of methylation. One would e xpect to see a complete loss of methylation at the Snrpn CpG island in a AS-IC deletion mouse. The authors posit that one of the possible reasons for the variable penetrance is that the 80 kb deletion may not remove all of the elements of the AS-IC66. In the same article describing the 80 kb deleti on model, another mouse model is described with an insertion and duplica tion located 13 kb upstream of Snrpn The AS-ICan mouse has a 6 kb duplication at the site, and insertion of a pur omycin selectable marker, a lox p site, and a portion of an Hprt cassette. When the AS-ICan allele is inherited maternally, the mice show a loss of methylation at the Snrpn CpG island and a 2 fold increase in Snrpn expression. Maternal inheritance of the AS-ICan allele is also able to rescue paternal inheritance of a PWS-IC mutation (the 4.8 deletion described in the PW S-IC models section). The AS-ICan mutation fulfills several

PAGE 26

26 of the expectations one would have for an AS-IC deletion model, but how this mutation has this effect is unclear. It is po ssible that expression of the pur omycin cassette in the opposite orientation from Snrpn may disrupt imprinting at the PWS-IC66. Transgenic Mouse Models Transgenic m ice have been created to try to recapitulate and study AS-IC and PWS-IC function. Transgenes containing only the Snrpn differentially methyl ated region are not imprinted. Larger transgenes (75 kb) with Snrpn and 50 kb upstream are able to imprint normally with Snrpn methylation at the differentially methylated region upon maternal inheritance and a lack of methylation upon paternal inherita nce. Expression of normally paternally expressed genes also seems to be higher upon paternal inheritance of the transgene than upon maternal inheritance of th e transgene. It must be noted that this data is based only on high copy number of the transgene (possibly aff ecting imprinting). A mi nitransgene made up of 1 kb of human sequence including the AS -IC and 200 bp encompassing the murine Snrpn promoter region also seems to imprint properly, showing the appropriate methylation patterns and expression patterns and asynchronous replica tion (another feature of imprinted regions). This minitransgene was also stud ied in lines with high copy number67. A similar minitransgene was studied with mutations in the 160 bp Snrpn minimal promoter revealing 5 cis elements required for proper imprinting. These elements ar e 2 de novo methylation signals required for methylation during oogenesis, an allele discrimination signal necessary to prevent PWS-IC methylation in sperm, and 2 elements required to maintain the unmethylated state of the PWS-IC on the paternal allele68. Other Mouse Models of PWS and AS Mouse m odels of PWS and AS have been crea ted that bear either a uniparental disomy (UPD) or deletion of the entire PWS/AS locus. Partial UPD mice have been created using

PAGE 27

27 translocations of portions of mu rine chromosome 7. Maternal pa rtial UPD of a region including the PWS/AS locus causes postnatal lethality or sm all size with reduced viability similar to other mouse models of PWS. Mice with partial maternal UPD and no paternal contribution showed an absence of expression of the normally paternally expressed gene Snrpn69. Paternal partial UPD of the PWS/AS locus causes an AS like phenotype with growth retardation, ataxia and abnormal brainwaves suggestive of seizure activity. These mice also exhibit ex treme obesity which is more like PWS than AS, although some AS patients are also obese70. The fortuitous insertion of a transgene into murine chromosome 7 resulted in the creation of a mous e line bearing a deletion of the entire PWS/AS locus while surrounding loci we re left intact. Paternal transmission of this transgene deletion causes a lack of expression of normally paternally expressed PWS genes, and postnatal lethality while maternal transmission of the same deletion results in a lack of Ube3a expression71. ATP10A Although mutations in UBE3A are sufficient to cause AS differences in phenotype between the molecular classes of AS suggest othe r genes m ay contribute. For example, the large deletion class is the only class that shows severe intractable ep ilepsy. It is thought that this phenotype may be due to deletion of GABR B3, GABRA5 and GABRG3, all of which are subunits of the GABAA receptor72. These three unimprinted ge nes are outside of the 2 Mb imprinted region, but within the 44.5 Mb co mmon deletion region. Other differences between the various molecular classes include late ons et obesity, motor skill development, cognitive abilities and incide nce of microcephaly34. ATP10A ( ATP10C PFATP ) has been considered a likely candidate due to its proximity to UBE3A 73 and its possible maternal expression that has currently come into question. Meguro et al. demonstrated a strong alle le bias in several brain samples and a loss of ATP10A expression in cultured lymphoblasts from 8 AS patients 74.

PAGE 28

28 Herzig et al. also found evidence of imprinted ATP10A expression. They found a loss of ATP10A expression in samples from AS brain, and preferential expression of one allele in normal brain samples, although parent of origin c ould not be determined. Furthermore, a recent study shows variability in ATP10A allelic expression between individuals. Hogart et al. found a significant sex bias towards monoallelic expres sion in the female brain and a promoter polymorphism that improves th e likelihood of monoallelic ATP10A expression41. ATP10A has also been implicated in Autism-Spectr um Disorder (ASD). Individuals with autism or atypical autism have been identified who have a maternal duplication of the PWS/AS region, however paternal inheritance of the same duplication has no effect75. Association studies have yielded conflicting results. While some studies have found no evidence to support an association between ATP10A and autism, Kato et al. observed linkage dise quilibrium consistent with the existence of an autism susceptible variant of ATP10A76,77. Atp10a encodes an aminophospholipid translocase that is hypothe sized to be involved in modulating body fat. Mice with a deletion including Atp10a were identified among radiation induced mutants at Oak Ridge National Laborat ories. Mice inheriting a maternal deletion including Atp10a become obese with almost twice th e body fat of mice inheriting the same deletion paternally suggesting the possibility of imprinting78. Maternal inheritance of this deletion also causes type 2 diabetes and nonalcoholic fatty liver disease79. Despite evidence favoring Atp10a imprinted expression, molecular studies describing Atp10a imprinting in the mouse brain are contradictory. Kayashima et al. found bialleleic Atp10a expression in all tissues incl uding brain and concluded that the locus is not imprinted, but Kashiwagi et al. reported predominant maternal expres sion in the hippocampus and olfactory bulb80,81. The differences in the studies include the ages of the mice, the polymorphisms used to

PAGE 29

29 determine parent of origin of the transcripts, an d the strains of mice used. However, both studies used crosses to C57BL/6 suggesting that strain specific differences in imprinting are unlikely to be responsible for the variant results82. Despite these differences, both reports confirm that Atp10a is widely expressed with high expression in the brain and that Atp10a lacks DNA methylation at an associated CpG island. At the least, Atp10a is a potentially imprinted gene of interest that warrants further study.

PAGE 30

30 Figure 1-1. Illustration of patte rns of gene expression. The blue bar represents the paternal chromosome, while the pink bar represents the maternal chromosome. The arrows indicate transcription of the gene, and each white box represents a gene. For most genes, both parental copies are active and transcribed (Gene A) or both parental copies are silent (Gene B). In the case of imprinted genes, one parental copy is active while the other is silent. Gene C is patern ally expressed and maternally silent, while Gene D is paternally silent and maternally expressed.

PAGE 31

31 Figure 1-2. Map of the murine PWS/AS locus (not to scale). The male and female symbols on the right mark the paternally and maternal ly inherited copies of the region. The paternally expressed genes ar e in blue, while the materna lly expressed (and thought to be maternally expressed) genes are in pi nk. The PWS-IC and AS-IC are represented by the blue and red circles respectively. The location of the AS-IC is unknown and the imprinting status of Atp10a is investigated in this work.

PAGE 32

32 Figure 1-3. Detailed maps of deletions near Snrpn exon 1. A. This map shows the relative locations (not to scale) acco rding to restriction enzyme cut sites of the various mouse model deletions near Snrpn exon 1, as well as the locations of deletions attempted (8.7 kb) and accomplished (6 kb) in this di ssertation. The top horizontal bar shows the relative locations of the Snrpn exons (marked as blue boxes). Along the top bar, restriction enzyme cut sites for BamHI (B ), EcoRI (E), HindIII (H) and HpaI (Hpa) are shown as vertical bars. B. This ma p shows the locations of the mouse model deletions with respect to the Snrpn promoter. The numbers marked on each side of each deletion indicate the distance from the Snrpn promoter with negative numbers 5 of the Snrpn promoter and positive numbers 3 of the Snrpn promoter.

PAGE 33

33 CHAPTER 2 ATP10A MATERIALS AND METHODS Mouse lines and Tissues utilized Several different m ouse lines were used in analyzing Atp10a allelic expression. To look at the parent of origin of Atp10a transcripts, we made use of a congenic mouse, B6.CAST.c7, in which the PWS/AS syntenic region is congenic for Mus castaneus on the C57BL/6 background. For the initial analysis, crosses between B6.C AST.c7 and C57/BL6 were performed and the progeny analyzed. PWS-IC 35kb, the 35 kb PWS-IC dele tion model, and PWS-ICHs, bearing a replacement of the murine PWS -IC with the human PW S-IC, were also used in this study, and these three lines are on a C57BL/ 6 background (see Chapter 1 for a more detailed description). At the appropriate age (8 weeks for adult tissues at birth for neonate tissues) the mice were sacrificed and their brains were collected whole or dissected. Tissues were quick frozen in liquid nitrogen and stored at -800 C. Reverse Transcriptase PCR (RT-PCR) RNA was extracted per using R NA-Bee (TelTech, Inc.). For RNA extraction from neonate olfactory bulbs, samples were pooled from 3 neonate mice of the same genotype. RNA was DNase treated, then reverse transcribed with S uperScript II reverse transcriptase (Invitrogen) using standard procedures. PCR was perf ormed on the resultant cDNA (and no reverse transcriptase controls) under standard conditions. The primers used for the Atp10a RT-PCR were: Atp10a4F, 5-GATCATGCTGAC ATCATTGG-3 and Atp10a4R, 5GCTGGATCATGGTGAAGAG G-3. The primers used for the Ube3a RT-PCR were: Ube3a 5F, 5CACATATGATGAAGCTACGA -3 and Ube3a 6R, 5CACACTCCCTTCATATTCC -363. The PCRs were sequenced by the CMG sequenc ing core and analyzed using Sequencher v 4.2 (Gene Codes).

PAGE 34

34 For examining Atp10a allelic expression we made use of a previously unreported polymorphism. We found a polym orphism between C57BL/6 and B6.CAST.c7 in exon 13 of Atp10a, such that B6.CAST.c7 has a MspA11 site not present in C57BL/6 mice. This polymorphism was discovered by sequencing RT-P CR products from C57BL/6 and B6.CAST.c7 cDNA templates using Atp10a4F and Atp10a4R as prim ers (listed previously). All analysis of Ube3a allelic expression was done using a previous ly characterized polymorphism between the same two mouse lines63. RT-PCR reactions for both Atp10a and Ube3a were run on an agarose gel and the appropriate sized bands were exci sed, gel purified (QIAquick Gel Extraction Kit, QIAgen), and sequenced as previously describe d. The chromatograms were compared. When multiple samples from the same brain region from equivalent mice were analyzed as described, the resulting chromatograms were similar in relative peak height. Atp10a CpG Island Characteriz ation EMBOSS CpG plot (http://www.ebi.ac.uk/tools/emboss/cpgplot/index.html ) identif ies an approximately 730 base pair long CpG island encompassing most of Atp10a exon 1. The methylation state of this Atp10a CpG island was ascertained by sodium bisulfite methylation analysis. DNA for the methylation analysis was isolated from brain regions by incubating the mechanically homogenized tissue in lysis buffer (100 mM Tris pH 8, 5mM EDTA, 0.2% SDS, 200 mM NaCl ) with proteinase K (100 ug/mL) at 550 C overnight followed by PCIA extraction and ethanol precipitation. Olfactory bulb samples were pooled from 3 neonate mice of the same genotype for methylation analysis (p rogeny from a cross between PWS-IC 35kb and B6.CAST.c7 that did not inherit the PWS-IC deletion). Sodium bisulfite conversion was performed as described previously83. DNA samples were first sheared by vortexing, and then denatured using 0.3 M NaOH. The samples were kept in the dark at 550 C for 16-20 hours after adding 1.55M sodium bisulfite (pH 5) and 0.5 M hydroquinone. A desalting column was then used (Promega,

PAGE 35

35 Wizard DNA Clean-up) to remove free bisulfite. The samples were treated with 0.3 M NaOH, ethanol precipitated and dissolved in ddH2O. PCR was performed on the bisulfite converted samples using the following primer set: Atp10a bis EF2 (5-ATTTGGAAGTTTGGAT AGG-3) and Atp10a bis ER2 (5ACCAAATACTAAATACAACC-3). PCR products were run on a 2% agarose gel then gel purified (QIAquick Gel Extraction Kit, QIAgen). Gel purified PCR products were cloned with the TOPO TA cloning kit (Invitrogen). Sample s were sequenced by the CMG sequencing core and analyzed on Sequencher (v. 4.2). Sequences with <95% conversion we re discarded. Results stated are from at least 2 separate P CR reactions per brain region analyzed.

PAGE 36

36 CHAPTER 3 ATP10A RESULTS Research on Atp10a has disagreed on the matter of imprinting. Mice with deletions including Atp10a only show a phenotype of obesity and type 2 diabetes on maternal transmission which strongly suggests imprinting is at work79. While one study found no evidence of Atp10a imprinting in any tissue, another found predominant maternal expression in certain brain regions80,81. Given the conflicting research rega rding the imprinting status of Atp10a, we chose to examine the imprinting status of the gene in mi ce with no PWS-IC defect and in two lines of mice each with a different PWS-IC defect. We reasoned that if Atp10a is imprinted, then allelic expression patterns should be sensitive to the PW S-IC. The PWS-IC has been previously shown to have an impact on allelic expr ession of a gene that is normally maternally expressed in some tissues: Ube3a63. For each cross, we used a congenic mouse line, B6.CAST.c7, which has a portion of murine chromosome 7 from Mus musculus castaneus. This allowed us to take advantage of a single nucleoti de polymorphism (SNP) between Mus musculus castaneus and Mus musculus domesticus (the subspecies of most lab mice) to distinguish between maternal and paternal gene expression. Olfactory bulb and hi ppocampus were included in this study because of previous work indicating that Atp10a imprinting may be restricted to these brain regions81. Atp10a Expression in C 57BL/6 and B6.CAST.c7 crosses RT-PCR was performed for severa l brain regions collected fr om 8 week old mice derived from C57BL/6 and B6.CAST.c7 and reciprocal crosses. These RT-PCR reactions were sequenced and the chromatograms from the reciprocal crosses were compared. The relative peak at any given site in the sequen ce indicates how much of a partic ular base was present in the original sample. In the olf actory bulb, hippocampus and severa l other brain regions, there is

PAGE 37

37 roughly equal expression of Atp10a from the paternal and maternal allele. C57/BL/6 has a T at the polymorphic site while B6.CAST.c7 has a C base. In all of the samples, it is obvious that the maternal allele is not preferenti ally expressed. In both crosses a nd in all brain regions examined, there is an equal level of expression from each Atp10a allele (Fig.3-1). RT-PCR was also performed from the same cDNAs for Ube3a, a known preferentially maternally expressed gene of the PWS/AS locus. As expected, Ube3a shows the expected result of predominately maternal expression (Fig.3-2). Effect of the PWS-IC on Atp10a Expression PWS-IC Deletion Effects We found e qual levels of allelic expression of Atp10a in C57BL/6 and B6.CAST.c7 and reciprocal crosses, however Atp10a expression may still be se nsitive to the PWS-IC. To investigate this possibility, we crossed B6.CAST.c7 and PWS-IC 35kb (35 kb deletion) mice and collected brain samples from the newborn pups. Whole brain, as well as olfactory bulb and cerebral cortex were collected and Atp10a expression was analyzed for each. We performed RTPCR across the same polymorphism previously de scribed and compared the chromatograms of pups inheriting the deletion with their wild type littermates. There is no preference for maternal expression of Atp10a in either the deletion pups or the wild type pups in cerebral cortex, olfactory bulb or in the w hole brain (Fig.3-3). A cont rol RT-PCR chromatogram for Ube3a shows the expected preferential expression of the maternal alle le in the wild type pups and biallelic expression in the pups inheriting the PW S-IC deletion allele (Fig. 3-4). This is consistent with previous work showing the PWS-IC is necessary to repress paternal Ube3a 63. PWS-ICHs Effects The PWS-ICHs allele has a replacement of the murine PWS-IC with the human PWS-IC. Upon paternal transmission of the PWS-ICHs allele, the human PWS-IC is able to drive

PAGE 38

38 transcription of the downstream cl uster of PWS genes, but not the upstream cluster. When the PWS-ICHs allele is maternally inherited, it is unable to be silenced by the murine AS-IC. The PWS-ICHs activates transcription of the normally pate rnally expressed down stream cluster of PWS genes from the maternal allele including the Ube3a antisense transcript. Mice inheriting the PWS-ICHs allele maternally have biallelic expression of Ube3a, which is normally preferentially expressed from the maternal allele in some brain tissue. It is possible that Atp10a is regulated as Ube3a is regulated, and the PWS-IC has an effect on allelic expression. We looked at whether the PWS-ICHs allele changes the preferred parent of origin for Atp10a expression using reciprocal crosses between the congenic B6.CAST.c7 line and PWSICHs mice. Atp10a is not observed to be preferentially maternally expressed in either the wild type or in the mice inheriting the PWS-ICHs maternally (Fig. 3-5). When the PWS-ICHs allele is inherited maternally, there is an in crease of expression of maternal Ube3a antisense transcript ( Ube3a-ASTS ) and a decrease in maternal Ube3a expression. In the olfactory bulb, hippocampus, cerebral cortex and cerebell um of 8 week old mice inheriting PWS-ICHs allele maternally, Ube3a is no longer preferentially maternally e xpressed as it is in wild type mice (Fig. 3-6). When the PWS-ICHs allele is inherited paternally, th ere is a loss of expression of the upstream paternally expressed genes Necdin, Mkrn3, Magel2 and Frat3, but expression of the downstream paternally e xpressed genes including Ube3a-ASTS is preserved64. Mice inheriting the PWS-ICHs allele paternally show preferential maternal expression of Ube3a similar to wild type in all brain tissues (Fig. 3-7). In both wild type and in paternal PWS-ICHs allele mice, Atp10a shows roughly equal biallelic expression (Fig. 3-8).

PAGE 39

39 We also examined Atp10a and Ube3a expression in P1 mice inheriting the PWS-ICHs allele maternally and paternally. In whole brain, we found similar resu lts to the observations in the adult brains. Mice inheriting the PWS-ICHs allele maternally showed a loss of preferential maternal expression of Ube3a but no change in Atp10a expression compared to wild type (Fig. 3-9 and Fig. 3-10). Mice inheriting the PWS-ICHs allele paternally show preferential maternal expression of Ube3a and biallelic Atp10a expression (Fig. 3-9 and Fig. 3-10). Atp10a CpG Island Methylation Analysis EMBOSS CpG plot (http://www.ebi.ac.uk/tools/emboss/cpgplot/index.html ) identif ies an approximately 730 base pair long CpG island encompassing most of Atp10a exon 1. To examine the methylation status of this CpG island, bi sulfite conversion was performed on genomic DNA samples of cortex and olfactory bulb colle cted from newborn C57/BL/6 mice. PCR amplification of the converted DNA spanning a 362 bp portion of the CpG island was performed. The examined portion of the Atp10a CpG island exhibits nearly completely absent methylation in both the cortex and the olfactory bulb at all 47 CpGs analyzed (Fig. 3-11). Atp10a Imprinting Status Summary Atp10a is no t imprinted according to a lack of difference in al lelic expression of a SNP in several brain regions. For this analysis multiple informative crosses were performed between congenic B6.CAST.c7 and C57BL/6 mice. Equa l biallelelic expression was evident in the cerebral cortex, hippocampus, cereb ellum and olfactory bulb in 8 week old mice from C57BL/6 and B6.CAST.c7 reciprocal cro sses. From previous work, it is known that mice inheriting a PWS-IC deletion paternally s how a decrease in paternal Ube3a antisense transcript and an increase in paternal Ube3a expression63. Using the same PWS-IC deletion we did not detect changes in Atp10a expression. No antisense transcripts extending into Atp10a have been found, so it is possible that it may be regulated in a different way than Ube3a80. There is an increase in

PAGE 40

40 paternal expression of Ube3a but not of Atp10a in the olfactory bulb, ce rebral cortex and the whole brain of newborn mice inhe riting a PWS-IC deletion (PWS-IC 35kb) paternally. There is no difference in Atp10a allelic expression between the PWS-IC deletion mice and their wild type littermates; they all show equa l biallelic expression. Like wise, although mice inheriting PWSICHs maternally show disturbance of Ube3a allelic expression, there is no change in allelic expression of Atp10a. Imprinting center function seems to have diverged between mouse and human, as the human PWS-IC is una ble to activate transcription of the upstream gene cluster and is unable to be silenced when maternally tran smitted. The human PWS-IC region does promotes transcription of the downs tream genes including the Ube3a antisense transcript explaining the relative decrease in maternal Ube3a transcription when the hum an PWS-IC is inherited maternally64. The widespread imprinting of Ube3a seen throughout the brain was somewhat unexpected given previous brain region specifi c observation of maternal expr ession. Maternal expression of Ube3a has been identified in Purkinje cells in the cerebellum, in the hippocampus, and in mitral cells of the olfactory bulbs by in situ84. Preferential maternal expression of Ube3a in the cortex, and observable levels in neonatal whol e brain would not be expected if Ube3a imprinting were restricted solely to thos e regions previously identif ied. Significantly, Landers et al also observed more widespread imprinting of Ube3a in the brain than is obvious from in situ 85. Imprinted genes often have an associated Cp G island near their promoters and this CpG island is usually methylated on the silent allele. We found almost entirely absent methylation at the Atp10a CpG island in newborn pups, which is cons istent with results from other studies80,81. Allele specific methylation at the Atp10a CpG island would have been supported the notion that Atp10a is imprinted, however a lack of methylation at the Atp10a CpG island cannot prove that

PAGE 41

41 Atp10a is not imprinted, given that some imprinted genes (such as Ube3a) do not have a differentially methylated CpG island.

PAGE 42

42 Figure 3-1. Atp10a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 crossed with C57BL/6 and recipr ocal crosses. The polymorphic site is highlighted in blue. The mouse stra ins crossed are at the bottom of each chromatogram (mother listed first). The B6 .CASTc7 allele at this site is C, and the C57BL/6 allele is T.

PAGE 43

43 Figure 3-2 Ube3a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 crossed with C57BL/6 and recipr ocal crosses. The polymorphic site is highlighted in blue. The mouse stra ins crossed are at the bottom of each chromatogram (mother listed first). The B6 .CASTc7 allele at this site is G, and the C57BL/6 allele is A.

PAGE 44

44 Figure 3-3. Atp10a sequencing chromatograms for 2 brai n regions and whole brain from P1 mice from a B6.CASTc7 female and a PWS-IC 35kb male cross. The polymorphic site is highlighted in blue and the pup genot ype is noted at the bottom of the column. The maternal allele (B6.CASTc7) at this pol ymorphism is C, and the paternal allele (C57BL/6 which is the PWS-IC 35kb background strain) is T for this cross.

PAGE 45

45 Figure 3-4. Ube3a sequencing chromatograms for 2 brain regions and whole brain from P1 mice from a B6.CASTc7 female and a PWS-IC 35kb male cross. The polymorphic site is highlighted in blue and the pup genotype is noted at the bottom of the column. The maternal allele (B6.CASTc7) at this polym orphism is G, and the paternal allele (C57BL/6 which is the PWS-IC 35kb background strain) is A for this cross.

PAGE 46

46 Figure 3-5. Atp10a sequencing chromatograms of 4 brain regions of 8 week old mice from PWS-ICHs x B6.CASTc7. The polymorphic site is highlighted in blue. The mouse genotypes are at the bottom of each chromat ogram. The B6.CASTc7 allele (paternal for this cross) at this site is C, and the C57BL/6 allele is T.

PAGE 47

47 Figure 3-6. Ube3a s equencing chromatograms of 4 brain re gions of 8 week old mice from PWSICHs x B6.CASTc7. The polymorphic site is highlighted in blue. The mouse genotypes are at the bottom of each chromat ogram. The B6.CASTc7 allele (paternal for this cross) at this site is G, and the C57BL/6 allele is A.

PAGE 48

48 Figure 3-7. Atp10a sequencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 x PWS-ICHs. The polymorphic site is highl ighted in blue. The mouse genotypes are at the bottom of each chromatogram. The B6.CASTc7 allele (maternal for this cross) at this site is C, and the C57BL/6 allele is T.

PAGE 49

49 Figure 3-8. Ube3a s equencing chromatograms of 4 brain regions of 8 week old mice from B6.CASTc7 x PWS-ICHs. The polymorphic site is highl ighted in blue. The mouse genotypes are at the bottom of each chromat ogram. The B6.CASTc7 allele (maternal for this cross) at this site is G, and the C57BL/6 allele is A.

PAGE 50

50 Figure 3-9. Atp10a sequencing chromatograms of whole brain from P1 mice from a PWS-ICHs x B6.CASTc7 and reciprocal cross. The polymorphic site is highlighted in blue and the pup genotype is noted at the bottom of the chromatogram. A. PWS-ICHs x B6.CASTc7 pups. The maternal allele (C57BL/ 6) at this polymorphism is C, and the paternal allele (B6.CASTc7) is T for this cross. B. B6.CASTc7 x PWS-ICHs pups. The maternal allele (B6.CASTc7) at this polymorphism is T, and the paternal allele (C57BL/6) is C for this cross.

PAGE 51

51 Figure 3-10. Ube3a sequencing chromatograms for whole brain from P1 mice from a PWS-ICHs x B6.CASTc7 and reciprocal cr oss. The polymorphic site is highlighted in blue and the pup genotype is noted at the bott om of the chromatogram. A. PWS-ICHs x B6.CASTc7 pups. The maternal allele (C57BL/ 6) at this polymorphism is A, and the paternal allele (B6.CASTc7) is G for this cross. B. B6.CASTc7 x PWS-ICHs pups. The maternal allele (B6.CASTc7) at this polymorphism is G, and the paternal allele (C57BL/6) is A for this cross.

PAGE 52

52 Figure 3-11. The methylation state the Atp10a CpG island in newborn cerebral cortex and olfactory bulb. A) This diagram shows the relative positions of the Atp10a exon 1, the Atp10a CpG island and primers used for bi sulfite analysis. The CpG island is represented by a black bar, with the arrows show the positions of the primers (not to scale). B) The methylation status of the Atp10a CpG island. Each ci rcle represents a CpG with the white circles representing un methylated cytosine residues, and the black circles represent methylated cytosi ne residues. The numbers on the right indicate the number of clone s identified with the describe d patterns of methylation.

PAGE 53

53 CHAPTER 4 MOUSE MODELS MATERIALS AND METHODS Strategy Although the boundaries of the hum an PWS-IC have been narrowed down to 4.3 kb of sequence, the minimum deletion previously known to cause a complete imprinting defect in mice was 35 kb. Attempts to create mouse models with smaller deletions had not been successful in creating a complete imprinting defect. In order to identify key elements of the PWS-IC, we sought to create a new mouse model that narrow ed the known boundaries of the murine PWS-IC. We also sought to create model in which the PW S-IC could be selectively removed using the cre/loxp system. A floxed (flanked by loxp sites) PWS-IC mouse would be extremely useful and would make possible tissue specific or inducible deletion of the PWS-IC. The regions selected for creating a PWS-IC deletion and floxed (flanked by loxp sites) PWS-IC were chosen based on what was known a bout previous deletion m odels (see Fig. 1-3). Although it shows a strong postnatal lethality phenotype (~80%) and mo st likely contains at least one gene contributing to the PWS failure to thrive phenotype, the Snrpn-Ube3a deletion mouse has no defect in imprinting indicating that it probably does not contain the PWS-IC. The SnrpnUbe3a deletion begins at the Ec oRI site (2.4 kb 3 of the Snrpn promoter) located in Snrpn intron 1 and continues 3 into Ube3a. The Snrpn exon 2 deletion also begins at the same Eco RI site and shows no imprinting perturbation. Logically, the PWS-IC should be located 5 of this Eco RI site since deletions 3 of the site do not disturb imprinting. The 4.8 kb deletion model shows a partial imprinting defect with hypermethylation at both Ndn and Mkrn3 but only a partial reduction in Ndn expression. These results indicate that the 4.8 kb deletion removed part but not all of the PWS-IC when compared to the imprint defect of the PWS-IC 35kb mouse. The 4.8 kb deletion spans from the HindIII site (2.7 kb 5 of the Snrpn promoter) 5 of Snrpn exon 1 to the

PAGE 54

54 EcoRI site in Snrpn intron 1 (2.4 kb 3 of the Snrpn promoter). The regions selected for deletion and floxing with the goal of including the entire PWS-IC share the same 3 border as the 4.8 kb deletion but extend further 5. The initial region selected wa s 8.7 kb long and extended to the EcoRI site 5 of Snrpn exon 1. After multiple failed attempts to get properly targeted ES cells, it was determined that the 5 arm of the plas mid construct was in a repetitive region of DNA explaining why all ES cell clones identified as targeted at th e 3 end had insertions or rearrangements at the 5 end. To avoid the repetitive DNA region, we decide d to target a smaller region which we believed would still give us a complete PWS -IC deletion mouse. We chose a 6 kb region spanning from an HpaI site (3.7 kb 5 of the Snrpn promoter) to the same EcoRI site (2.4 kb 3 of the Snrpn promoter) in Snrpn intron 1. We chose this 6 kb because Karen Johnstone in our lab had previously made a mouse that swapped the same 6 kb for 6.9 kb of the human PWS IC. This mouse showed expression of the downstream cluste r of PWS genes, but li ttle to no expression of the upstream cluster of PWS genes including Necdin indicating that SNRPN promoter function is preserved between mouse and human but PWS-IC f unction is not. This is compelling evidence that the murine PWS-IC is w ithin that 6 kb of sequence. This 6 kb of sequence includes a DNase I hypersensitive site that could be essential to the full function of the PWS-IC. Two DNase I hypersensi tive sites have been found in at the 5 end of the human SNURF/SNRPN locus; one at the promoter and the other within exon 2. These DNase I hypersensitive sites are exclusively f ound on the paternal allele and these regions interact with regulatory proteins86. The murine Snurf/Snrpn locus also shows a parental allele specific DNase I hypersensitive site pattern with six hypersensitive sites on the paternal allele (Dr. Thomas Yang, unpublished data). DNase hypers ensitive site 6 (DHS6) is upstream of the

PAGE 55

55 Snrpn promoter, DHS1 at the promoter and DHS2 through DHS5 lie downstream of the promoter. Mice with either the Snrpn exon 2 deletion or the Snrpn exon 2 through Ube3a exon 2 deletion do not have a PWS-IC deletion phenot ype although they both re move DHS5 indicating that DHS5 is not essential for PWS-IC function. The 4.8 kb deletion mouse model removes DHS1 through DHS4, but only shows partial imprinting defect. The 6 kb chosen for the flox and deletion encompasses DHS6 as well as DHS1 through DHS4. There were also constructs made to create alle les with the reverse orie ntation of the same 2 regions (8.7 and 6 kb) to explor e whether the PWS-IC and AS-IC have directional requirements for interaction. There is evidence in a pair of siblings with AS with translocations separating and increasing the distance between th e ICs that relative direction of the ICs may be important for their appropriate interaction87. These efforts were discont inued after difficulty achieving targeted ES cells. Targeting Construct Assembly The targeting constructs were assem bled usi ng restriction digests, gel purification, and DNA ligation. The backbone for all of the constr ucts is pBluescript KS+ (Stratagene). All murine DNA used to create the cons truct originally came from a phage library of fragments from BAC (Bacterial Artificial Chromosome) 397F16 (Research Genetics). It is important to note that BAC 397F16 is of the same strain as the embryoni c stem (ES) cells used (129S1/Sv) since in is known that targeting efficiency decreases when th e targeting construct and the ES cells are from different strains of mice88. DH5 chemically competent E. coli cells and later Stbl 2 chemically competent E. coli cells were used to propagate the plas mids. Transformation of the cells was done according to standard protocol, and the ce lls were grown on ampicillin Luria broth (LB) plates (50ug/mL ampicillin). I ndividual colonies were grown in ampicillin LB, and plasmid

PAGE 56

56 DNA was extracted and purified using the QIAprep spin miniprep kit (QIAgen) or by alkaline lysis followed by PCIA extraction (phenol, chloroform, isoamyl alcohol in a 25:24:1 mixture respectively) and ethanol precipitation. Af ter each ligation, transf ormation, selection of individual colonies, and plasmid DNA extraction, restriction digest s were performed to check for proper insertion of each piece of DNA. Clones with the correct inser tion were stored as glycerol stocks (25% sterile glycerol, 75% culture media with the desired plasmid containing E. coli ) at 800 C. Restriction digests were performed acco rding to manufacturers instructions (New England Biolabs), and electropho resed on agarose gels and DNA visualized with ethidium bromide. When inserting a new component of th e construct, bands of the desired DNA were identified by size using a UV transilluminator, ex cised and purified usi ng either the QIAquick gel extraction kit (QIAgen) or GenElute agarose spin columns (Sigma). When performing ligations, the vector was sometimes dephosporylat ed using Antarctic phos phatase (New England Biolabs) or Alkaline Phosphatase, Calf Intestin al (New England Biolab s). Ligations were performed using either T4 DNA ligase (New England Biolabs) or DNA ligation kit Mighty Mix (Takara). For the insertion of the lox p site 5 of the IC region, a pair of internally complimentary oligonucleotides was designed with a lox p site and compatible ends for the desired insertion site. The pair of oligonucleotides (5 phos phorylated by the manufacturer) was annealed by heating to 950 C and slowly cooling to room temperature (500 ng of each oligo in 20 uL). Varying concentrations of the annealed lox p oligos (1 uL of straight, 10X, 100X or 1000X diluted) were ligated to appropr iately digested and dephosphorylated vector (25-50 ng vector). The pCAGneo cassette was inserted into all of th e targeting constructs as a positive selectable marker for selection in ES cells. Except for the first 8.7 kb flox and flip constructs, a viral

PAGE 57

57 thymidine kinase (TK) cassette was inserted for use as a negative selective marker selecting against ES cells with nonhomologous recombination. Transformation of ES Cells Cells con taining the targeting construct were grown overnight in baffled flasks containing ~300 mL of LB with ampicillin. Plasmid DNA was harvested by alkaline lysis, followed by purification in a cesium chloride equilibrium grad ient. The targeting construct was linearized by restriction enzyme digest, PCIA extracted twice and then etha nol precipitated. ~80 ug of linearized targeting construct DNA was used in each electroporati on (see Fig. 4-1 for targeting schematic). CJ.7 ES (embryonic stem) cells88 were grown on a feeder layer of mitomycin C (Moravek Biochemicals, Inc.) treated transgenic Neor mouse embryonic fibroblasts (MC MEFs) except for during the stage when the ES cells were grown on 10 cm plates after electroporation. Actively growing ES cells were trypsinized, triterated to a single cell suspensi on, washed twice with phosphate buffered saline (PBS), then 5x107 cells were resuspended in 0.8 mL PBS and placed in an electroporation cuvette with a 0.4 cm electr ode gap (Biorad). ~80 ug of linearized targeting construct were suspended in 100 uL of PBS and we re added to the cuvette. The contents of the cuvette were mixed by pipetting. Electroporation was carried out at 0.8 kV and 3 uF on a Gene Pulser (Biorad). After a recovery period of 10 minutes, the cells were transferred to 15 gelatinized 10 cm dish with 10 mL of 3M cell media (500mL Dulbeccos Modified Eagle Media, 5 mL penicillin/streptomycin, 5 mL nucleoside stock, 5 mL 100X non-essential amino acids, 75 mL heat inactivated fetal bovine serum (FBS), 0.9 mL beta-mercaptoethanol) with leukemia inhibitory factor (50 uL of 107 units of ESGRO LIF per bottle, Chemicon). The next day, the media was changed to 3M with LIF and Genetic in (G418 at 200 uL/mL, GibcoBRL). After feeding the ES cells with 3M with LIF and G418 for 3 days, the media was changed to 3M with

PAGE 58

58 LIF, G418, and 1-(2-Deoxy-2-flouro-beta-D-a rabinofuranosyl)-5-iodour acil (FIAU at 0.2-0.7 uM, Moravek Biochemicals, Inc.). After feed ing the ES cells for 4-5 days, the media was switched back to 3M with LIF and G418. 1-3 days later, the colonies we re ready to pick. The colonies of ES cells were examined under a microscope and any that seemed by morphology to be differentiating or looked like two merged colonies were not selected. Colonies were picked by drawing up into a pipette and placed in a 96 well pl ate, trypsinized, triterated by pipetting, then transferred to 24 well plates containing a feeder la yer of MC MEFs. When most of the wells on the 24 well plate reached 80-90% confluence, the plates were trypsinized and split with some of the cells getting mixed with 2X freeze media (50% 3M with LIF, 30% FBS, 20% DMSO) then frozen slowly at -800 C, and the other portion of cells left to grow for DNA. When the cells left for DNA changed the media color from red to yellow, the media was removed and lysis buffer (100 mM Tris pH 8.5, 5 mM EDTA, 0.2% SDS, 200 mM NaCl) with proteinase K (100 ug/mL) was added. After incubating at 370 C for at least 12 hours, the lysis buffer was collected and DNA was extracted by PC IA extraction and ethanol precipitation. ES cell DNA was resuspended in 100 uL of ddH20. For the transient cre transfection, a correctly ta rgeted ES clone (Kflox IC 6D1) was thawed onto MC MEFs and the cells were electroporated under the same conditions previously described with pCAG cre plasmid89 (non-linearized). Cells were grown and picked as described previously, except no G418 or FIAU was used in the media. ES colonies were split onto three 24 well plates; one for freezing, one for DNA and on e for G418 treatment. The day after splitting, the 24 well plates designated for G418 treatment were given media with G418 and continued to receive media with G418 for several days. Th e identities of the col onies sensitive to G418

PAGE 59

59 (because they lost the Neo cassette) were reco rded and DNA from those colonies (from the DNA plate) was further analyzed. Analysis of ES Cell Clones The initial analysis of the ES clones was done by Southern blot. ES cell D NA was digested with the appropriate enzyme or enzymes (Spe I for screening the 3 end and Hpa I and EcoRV for the 5end screening), then loaded onto a 0.7 % TAE agarose gel and run overnight at 35 V. The next day, the gel was photographed on a UV tr ansilluminator, and th e molecular weight marker was marked on the gel with a needle dipped in India ink. The gel was then treated with an acid solution (0.2 N HCl) for 10 minutes, and th en treated with alkali solution for 45 minutes and neutralizing solution for 90 mi nutes. For the Southern blot, the gel was sandwiched between a Whatman paper wick (3M) with both ends in 10X sodium chloridesodium citrate buffer (SSC) and a hybond-N+ membrane (GE Healthcare) on top. Two layers of Whatman paper were layered on top, along with paper towels and a wei ght. After overnight transfer of DNA onto the hybond membrane, the India ink loca tions were marked in pencil, and the hybond was rinsed in 10X SSC. The hybond membrane was then baked at 800 C for 2 hours. The blot was put in Dextran prehybridization solu tion for 2 hours rotating at 650 C, then in Dextran hybridization solution with the appropriate radioactive probe rotating at 650 C overnight. Radioactive probe was made using Prime-it II random primer labelin g kit (Stratagene). The DNA used for probes was isolated by restriction enzyme digest and gel purification (Qiagen) from plasmids 210 and 207 originally from a phage library of fragme nts from BAC (Bacterial Artificial Chromosome) 397F16 (Research Genetics). The next day, the bl ot was washed using 1 high stringency wash (2X SSC, 0.1% sodium dodecyl sulfate (SDS)) fo r 15 minutes, then 1-2 low stringency washes (0.1X SSC, 0.1% SDS) for 15 minutes each. The blots were wrapped in plastic wrap and exposed BioMax XAR film (Kodak) at -800C with intensifying screens. When necessary, the

PAGE 60

60 blots were stripped using boiling 0.5% SDS. Kf lox IC 2B5, Kflox IC 6D1 and Kflox IC 9A5 were all identified as properly targeted clone s by this method (Figures 4-2 through 4-5) After transient cre transfection, screening was done on the G418 sensitive clones by PCR under standard conditions. For finding colonies with the floxed (conditio nal) allele, primers were designed on each side of the 3 lox p site, one in the IC region and one in the 3 arm (Fig. 4-6). The primers are IC Recomb F1 5CTGGTGCTACGATAGCTGAG-3 and 3 Recom b R1 5CACCATCACCAATAATAGGTC-3. The endoge nous alleles PCR product is 328 bp long while the floxed alleles PCR product is 357 bp l ong. A colony with the IC deletion allele would show only the endogenous band, as one of the prim er sites is m issing on the deleted allele. For the IC deletion, the same primer in the 3 arm wa s used, but paired with a primer in the 5 arm (Fig. 4-7). The primers used are 5Recomb F1 5GTGGACATCAGTTGTAAATAGC-3 and 3Recom b R1 (listed previously). When an IC deletion is present, a 301 bp DNA fragment is amplified during PCR. The endogenous band would be 6.3 kb long but it doesnt show up during PCR, likely because it does not amplify as well. The floxed allele would also be large, and is also not amplified. Out of 192 colonies pi cked after transient cre transfection of Kflox IC 6D1, 13 were G418 resistant and of those coloni es, one clone tested positive by PCR for the floxed allele and 5 tested positive by PCR for the deletion. The PCR products from the 2 colonies selected for injection into blastocyst s (Kflox IC 6D1 7B1, and Kf lox IC 6D1 7C3) were sequenced by the CMG core and were found to cont ain the expected sequences. The 5 arms of Kflox IC 6D1 7B1, and Kflox IC 6D1 7C3 were also screened by Southern blot, and both contained the predicted bands (Fig.4-5). The se lected ES cells were injected into C57BL/6 blastocysts and transplanted in to pseudopregnant female mice by the UF Mouse Models Core.

PAGE 61

61 Genotyping of offspring from the resultant ch imeras was done with the same primers and PCR conditions as the genotyping of the ES cel ls. The offspring of the PWS-IC deletion chimera were also checked for 129S1/Sv DNA by PCR using a strain specific PCR product length difference between 129S1/Sv and C57B L/6. The length of the PCR product from 129S1/Sv DNA is 165 bp and the length of the PCR product from C57BL/6 is 210 bp. The primers used were D7Mit 178, 5ACCTCT GATTTCAGAACCCTTG -3 and D7Mit178 2, 5TAGAGAGCCACTAGCATATCATAACC -3. Primer sequences were obtained from Mouse Genome Informatics (www.informatics.jax.org). Tail tissue was digested in lysis buffer with Proteinase K, PCIA extracted and EtOH precipitated to obtain DNA for genotyping. PWS-IC 6kb The deletion allele has been designated PWS-IC 6kb (from the ES clone, Kflox IC 6D1 7B1). Seven chimeric mice were made from this ES line, 5 males and 2 females. They had high levels of ES cell contribution as estimated by coat color from a bout 80% to 99% (agouti fur from 129S1/Sv ES cells, black fur from C57BL/6 blas tocyst donor, see Fig.4-8). Offspring of the chimera were sacrificed at postnatal day 1 a nd tested by PCR for transmission of the PWS-IC 6kb allele and for 129 S1/Sv transmission. Over 60 pups were collected but none showed inheritance of 129 S1/Sv DNA or the deletion allele. With such a high contribution of ES cells to the chimeric animals (and a high likelihood of germline contribution) this is an unexpected result. The possibility that the deletion leads to a sper matogenesis defect seems unlikely given that the PWS-IC 35kb deletion mice show no such defect. The appearance of female chimeras was also unexpected because the ES cell line we used is male It is possible that early in the development of the female mice, the ES cells lost the Y chro mosome leading to XO or XX female mice. Both of these females were fertile, and gave birth to one litter each, but none of these pups inherited the PWS-IC 6kb allele or 129 S1/Sv DNA.

PAGE 62

62 PWS-ICflox6kb The floxed allele has been named PWS-ICflox6kb (the ES clones designation was Kflox IC 6D1 7C3). Four chimeric mice were created from this ES line, 3 males and 1 female, all with very high levels of ES cell contri bution (98% or more by coat colo r). One chimeric male had to be humanely euthanized because of a severely distended abdomen. Germline transmission was achieved for this line from one chimera to a male who was then used to propagate the PWSICflox6kb mouse line. CMV-cre transgenic mice (B6.C-Tg(CMV-cre)1Cgn/J stock number 006054 from the Jackson Laboratory) have widespread expres sion of cre recombinase from the human cytomegalovirus promoter in all tissues, including the germline90. CMV-cre mice were crossed to the PWS-ICflox6kb mouse to create pups with a deletion of the PWS-IC. Pups from these crosses were sacrificed at postn atal day 1 or monitored for phenotype. P1 mouse brains from CMV-cre and PWS-ICflox6kb crosses (female listed first) were collected along with the tails for genotyping. The pup brains were flash frozen in liquid nitrogen and stored at -800C. As described previously, tail tissue was digested in lysis buffer with Proteinase K, PCIA extracted and EtOH precipitated to obtain DNA for genotyping. The mice were genotyped by PCR for the PWS-ICflox6kb allele, CMV-cre transgene as well as for PWS-IC 6kb (a result of inheriting both the floxed allele and the CMV-cre transgene and carrying out recombination). The PWS-ICflox6kb and the PWS-IC 6kb were genotyped from tail DNA using th e same primers and conditions used in the analysis of the transiently cre transfect ed ES clones (see Analysis of ES Cell Clones subsection in this chapter). The primers used for CMV-cre genotyping were BcreF, GAGTGATGAGGTTCGCAAGAAC-3, and BcreR, TCGCCATCTTCCAGCAGG-3.

PAGE 63

63 PWS-ICflox6kb Analysis by RT-PCR RNA was extracted from whole brains using RNA-Bee (TelTech, Inc.). RNA was DNase treated, then reverse transcribe d with SuperScript II reverse tr anscriptase (Invitrogen) using standard procedures. PCR was performed on th e resultant cDNA (and no reverse transcriptase controls) under standard conditions. RT PCR primers used were: N2.1 5CCCCGAGTATTAAGGATCTTG-3 and N6.2 5-GCAACAGTGCCT CTTCCCTG-3for Snrpn, NLG2F 5-CAGTCCCCATCCTCACTAATACA-3 and NLG2R 5TCTCCAGACAGTATTTTACCGATG-3for Magel2 and Beta Actin 5 5GTGGGCCGCTCTAGGCACCAA -3 and Beta Actin 3 5CTCTTTGATGTCACGCACGATTTC-3for Beta Actin (as a control). PWS-ICflox6kb Analysis by Northern Blot As previously described, RNA was extracted from the brains using RNA-Bee (TelTech, Inc.). Expression of several PWS genes was determined by Northern blot. For detecting expression of PWS genes other than the snoRNAs, 10 ug of RNA for each sample was run on a 1% agarose gel with form aldehyde. The gel was run at 70 volts for 2.5 hours. The 18S and 28S ribosomal RNA bands were marked with a needle dipped in India ink. The gel was soaked in 20X SSC for 20 minutes then sandwiched betw een a Whatman paper wick (3M) with both ends in 10X sodium ch loridesodium citrate buffer (SSC) and a hybondN+ membrane (GE Healthcare) on top. Two layers of Whatman paper were layered on top, along with paper towels and a weight. Afte r overnight transfer of DNA onto the hybond membrane, the India ink locations were marked in pencil, and the hybond was rinsed in 2X SSC. The hybond membrane was then baked at 800 C for 2 hours. The blot was put in 25 mL of Church and Gilbert hybridization buffer for 2 hours rotating at 650 C, then in 5 mL Church and Gilbert hybridization buffer w ith the appropriate radioactive probe rotating at 650 C overnight.

PAGE 64

64 Radioactive probe was made using Prime-it II rand om primer labeling kit (Stratagene) and alpha 32P dCTP (Perkin Elmer). The DNA used to probe for Necdin was PCR amplified with the primers W43-Ndn 9F 5-GTATCCCAAATCCACAGTGC-3 and W44-Ndn 10R 5CTTCCTGTGCCAGTTGAAGT-3 (courtesy of Dr. Edwin Peery). The DNA used to probe for Beta Actin was amplified using the same pr imers as used for the RT-PCR (see previous subsection). After amplificati on, the PCR was run out on a 2% ag arose gel, the gel bands were excised, and then purified using a QIAquick ge l extraction kit (QIAge n). After hybridizing overnight, the blot was washed using 1 high stringency wash (2X SSC, 0.1% sodium dodecyl sulfate (SDS)) for 15 minutes, then 1-2 low st ringency washes (0.1X SSC, 0.1% SDS) for 15 minutes each. The blots were wrapped in plasti c wrap and exposed BioM ax XAR film (Kodak) at -800C with intensifying screens. The blots were stripped using boiling 0.1% SDS before hybridizing with other probes. To detect expression of the snoRNA MBII-85, Northern blots were also used, but a polyacrylamide gel instead of an agarose gel wa s used. After the 8% denaturing polyacrylamide gel with 7M urea and 1X TBE polymerized, it wa s run at 250V for 30 minutes. Next, heated samples of RNA (10 ug each) mixed with loading buffer were loaded onto the gel and the gel was run at 250V for an hour and a half. The polyacrylamide gel was equilibrated in .5X TBE for 15-20 minutes at 40 C, then electroblotted onto a hybondN+ membrane (GE Healthcare) for 1 hour at 20 V. The hybondN+ membrane was baked at 800C overnight. The blot was prehybridized and hybridized in Church and Gilb ert hybridization buffer as described previously in this subsection. The snoRNA Northern blots were washed, exposed to film, and stripped between probes as previously described in th is subsection. Probes were labeled using T4 DNA Kinase (Invitrogen) and gamma 32P ATP (Perkin Elmer). The probes used were 5-

PAGE 65

65 TTCCGATGAGAGTGGCG GTACAGA-3 for MBII-85 and 5TCCTGCAATTCACATTAATTCTCGC AGCTAGC-3 for 5.8S rRNA as a positive control. PWS-ICflox6kb Analysis of Postnatal Phenotype Transgenic CMV-cre females were crossed with PWS-ICflox6kb males and the pups were monitored for postnatal phenotype. The pups were marked for identification at birth with tattoo marks on their paw pads and weighed daily. Th e mice were genotyped for inheritance of the CMV-cre transgene, PWS-ICflox6kb allele, and the PWS-IC 6kb (from inheritance of the CMV-cre transgene and PWS-ICflox6kb allele) as previously described. The weights of the pups at birth were compared as well as the survival rates for each genotype.

PAGE 66

66 Figure 4-1. ES cell targeting schematic. The blue box with an arrow represents the Snrpn promoter and exon 1, and the blue box represents Snrpn exon 2. The black arrowheads show the locations of the lox p si tes. The Neomycin resistance cassette is represented by the white box marked neo, and the white box marked TK represents the viral thymadine kinase ca ssette. After homologous r ecombination within each of the targeting arms, the recombinant allele is created. After transient transfection with a cre expressing plasmid, several alleles are created including the conditional floxed allele and the IC deletion allele.

PAGE 67

67 Figure 4-2. 3end screening for targeted cells. This diagram show s the relative positions of the important enzyme cut sites and the locati on of probe hybridization for the endogenous and recombinant allele at the 3 end. The green box labeled 207 probe is the location just 3 of the 3 arm where the radioactiv e probe hybridizes. Th e red vertical bars indicate where SpeI restriction enzyme cuts Note the SpeI site introduced by the neo cassette. The endogenous band is 18.9 kb and the recombinant band is 12.6 kb in length.

PAGE 68

68 Figure 4-3. Southern blot of the 3end of target ed ES cells. Shown is a radiographic film of a Southern blot of SpeI digested ES DNA hybridized to 207 probe. Kflox IC 2B5, Kflox IC 6D1 and Kflox IC 9A5 ES cell lines have a 12.6 kb band indicating correct targeting at the 3 end. The last lane contains control ES cell DNA showing the endogenous band at 18.9 kb.

PAGE 69

69 Figure 4-4. 5end screening for targeted cells. This diagram show s the relative positions of the important enzyme cut sites and the locati on of probe hybridization for the endogenous and recombinant allele at the 5 end. The green box labeled 210 probe is the location just 5 of the 5 arm where the radioactive probe hybridizes. The or ange vertical bars indicate where EcoRV restriction enzyme cuts and the blue bars indicate where HpaI restriction enzyme cuts. Note that the reco mbinant allele is missing a HpaI site that has been replaced by a loxp site. The endogenous band is 7.2 kb and the recombinant band is 15.8 kb.

PAGE 70

70 Figure 4-5. Southern blot of the 5end of target ed ES cells. Shown is a radiographic film of a Southern blot of HpaI/EcoRV digested ES DNA hybridized to 210 probe. Kflox IC 2B5, Kflox IC 6D1 and Kflox IC 9A5 ES cell lines have a 15.8 kb band indicating correct targeting at the 5 end. Kflox IC 6D1 7B1 shows a 7.9 kb band consistent with a PWS-IC deletion and Kflox IC 6D1 7C3 shows a 14 kb band consistent with a floxed PWS-IC. The last lane contains control ES cell DNA showing the endogenous band at 7.2 kb.

PAGE 71

71 Figure 4-6. ES Cell PCR screening for the condi tional allele. The blue box with an arrow represents the Snrpn promoter and exon 1, the blue box represents Snrpn exon 2 and the pink arrows signifiy the primer locations. The black arrowheads show the locations of the lox p sites. The exp ected band sizes are 323 bp for the endogenous allele, 2.1 kb for the recombinant allele, and 357 bp for the conditional allele. No DNA amplification is expect ed on the deletion allele.

PAGE 72

72 Figure 4-7. ES Cell screening for the deletion alle le. The blue box with an arrow represents the Snrpn promoter and exon 1, the blue box represents Snrpn exon 2 and the pink arrows signify the primer locations. The black arrowheads show the locations of the lox p sites. The predicted band sizes are 6. 3 kb for the endogenous allele, 8 kb for the recombinant allele, 8 kb for th e conditional allele (all not expected to amplify due to length) and 301 bp for the deletion allele.

PAGE 73

73 Figure 4-8. Identification of ES cell contribution by coat color. Shown from left to right are a wild type C57BL/6 male, a PWS-IC 6kb chimeric female, and a PWS-IC 6kb chimeric male. The blastocyst donor for the chim eras was C57BL/6, while the ES cell line used was derived from 129 S1/Sv. Contribu tion by 129 S1/Sv is visually identifiable by the agouti (brown) coat color. Black coat color indicates contribution to the animal from C57BL/6.

PAGE 74

74 CHAPTER 5 MOUSE MODELS RESULTS Due to the inability to o btain transmission of the PWS-IC 6kb allele from the chimeras, it became necessary to create the de letion in another way. Analysis of the deletion is necessary in order to determine whether the en tire PWS-IC is contained within the 6 kb of sequence selected for study. By breeding the PWS-ICflox6kb mouse line to a CMV-cre transgenic mouse line, we were able to obtain widespread deletion of the 6 kb of floxed sequence. By RT-PCR, it is clear that pups that inherited the CMV-cre transgene maternally and the PWS-ICflox6kb allele paternally (referred to from here on as CMV-cre tg/-, PWS-ICflox6kb +/flox pups) are missing expression from the normally paternally expressed PWS genes Snrpn and Magel2 while expression of a control gene, Beta Actin is unaffected. Pups inheriting onl y the CMV-cre transgene or only the PWS-ICflox6kb allele have similar levels of expression to the wild type cont rol pups who inherited neither (Fig. 5-1). Expression of the normally pa ternally expressed genes Necdin and MBII-85 was analyzed by Northern Blot (Fig. 5-1). Although the wild type pups and pups with either the CMV-cre transgene or the PWS-ICflox6kb allele (but not both) express Necdin and MBII-85 the CMV-cre tg/-, PWS-ICflox6kb +/flox pups do not. This strongly sugge sts a complete imprinting defect, as previous mouse models with a partial impr inting center deletion showed expression of Necdin The CMV-cre tg/-, PWS-ICflox6kb +/flox pups exhibit a failure to thrive phenotype similar to the phenotype observed for the 35 kb deletion m odel. These pups are small compared to wild type, whereas there is no observa ble physical phenotype differen ce between wild type and pups inheriting only the CMV-cre transgene or only the PWS-ICflox6kb allele (Fig. 5-2). The CMV-cre tg/-, PWS-ICflox6kb +/flox pups generally weighed less than the other genotypes at birth. The median weight for CMV-cre tg/-, PWS-ICflox6kb +/flox pups is 1.2 grams at birth while the

PAGE 75

75 median weight of their littermates with other geno types is 1.4 grams at bi rth (Fig 5-3). CMV-cre tg/-, PWS-ICflox6kb +/flox pups are weak, with little to no milk observable in their bellies, and have a reduced survival rate (Fig. 5-4). We obs erved that on P1 (postnat al day 1), 1/3 of the CMV-cre tg/-, PWS-ICflox6kb +/flox pups were dead, 2/3 were dead by P2 and by P3 there were no surviving CMV-cre tg/-, PWS-ICflox6kb +/flox pups. The two CMV-cre tg/-, PWS-ICflox6kb +/flox pups that survived to P2 did not gain we ight as the other pups did. While the other pups gained 1/10th to 1/2 of a gram in weight between P1 and P2, one of the CMV-cre tg/-, PWSICflox6kb +/flox pups stayed the same weight while the other lost 1/10th of a gram. The PWS-ICflox6kb allele has narrowed the known limits of the PWS-IC in mouse from 35 kb to 6 kb. Using a CMV-cre transgene in combination the PWS-ICflox6kb allele, we were able to create mice with a widespread de letion of the PWS-IC. Mice inheri ting this deletion paternally show a lack expression of the normally patern ally expressed genes from the region, show postnatal failure to thrive with low birt h weight, and reduced viability. The PWS-ICflox6kb allele encompasses the PWS-IC and opens new possibilitie s for research into th e PWS-ICs function.

PAGE 76

76 Figure 5-1. Expression of normally paternally expressed genes is missing in the brains of CMVcre tg/-, PWS-ICflox6kb +/flox pups. The genotypes of the pups represented in each lane are shown at the top. The first three lanes show control pups that are wild type, inherited only the CMV cre transgene or only the PWS-ICflox6kb flox allele. The last three lanes show three pups that inherited both the CMV cre transgene and the PWSICflox6kb flox allele. For each gene examined by RT-PCR, the reverse transcriptase (+) and the no reverse transcriptase control (-) are shown.

PAGE 77

77 Figure 5-2. CMV-cre tg/-, PWS-ICflox6kb +/flox pups are distinguishab le from wild type. The CMV-cre tg/-, PWS-ICflox6kb +/flox pup (shown at top) is noticeably smaller and weaker than its wild type littermate. At postnatal day 2, this CMV-cre tg/-, PWSICflox6kb +/flox pup was only 2/3 the weight of its wild type littermate (1.2 grams and 1.8 grams respectively).

PAGE 78

78 Figure 5-3. CMVcre tg/-, PWS-ICflox6kb +/flox pups typically weigh less at birth than their littermates. The birth weights of pups i nheriting only the CMV cre transgene, only the PWS-ICflox6kb flox allele or neither have been combined into the other genotypes category (shown in red). The birth weights of the CMVcre tg/-, PWS-ICflox6kb +/flox pups (shown in blue) are usually lower than the other genotypes. 0 1 2 3 4 5 6 7 1.11.21.31.41.51.61.71.81.92Number of miceBirth weight in grams Other genotypes, N= 13 CMV cre tg/ PWS ICflox6kb +/flox, N= 6

PAGE 79

79 Figure 5-4. Reduced survival of CMVcre tg/-, PWS-ICflox6kb +/flox pups compared to littermates. The survival rates of pups inheriting only the CMV cre transgene, only the PWS-ICflox6kb flox allele or neither have been combined into the other genotypes category (shown in red). CMVcre tg/-, PWS-ICflox6kb +/flox pups (survival rate shown in blue) have reduced viability with only 2/3 of pups living on P1, 1/3 living on P2, and all dead by P3. 0 10 20 30 40 50 60 70 80 90 100 1234567Survival in percentageAge in postnatal days CMV-cre tg/-, PWSICflox6kb +/flox, N= 6 Other genotypes, N= 13

PAGE 80

80 CHAPTER 6 DISCUSSION AND CONCLUSIONS The Imprinting Status of Atp10a By com paring allelic expression of Atp10a, we have found that Atp10a is not imprinted in every region of the brain examined. Atp10a allelic expression is also insensitive to either a mutation of the PWS-IC or a deletion of the PWS -IC. This is somewhat surprising, given that mice with a maternal deletion including Atp10a show a phenotype while those inheriting the same deletion paternally do not. One possibility that could ex plain this difference and the discrepancies seen in prev ious allelic studies of Atp10a imprinting in the brain is strain specific differences. It is known that the imprinting status of the same gene can be different in different strains of mice91. The mouse strains used in each study are different although the research done by Kashiwagi et al. and Kayashima et al both used crosses with C57BL/680,81. In an attempt to rule out strain specific differences as a cause of the conflicting results, Kayashima et al. did an imprinting study with the same conditions as Kashiwagi et al. who had previously shown maternal expression of Atp10a in olfactory bulb and hippocampus Despite using the same strains of mice, the same ages of mice and th e same polymorphism for analysis as Kashiwagi et al. (all differences between the studies), Kayashima et al. found equal biallelic expression of Atp10a in frontal cortex, olfactory bulb, and hippocampus82. With this information, strain specific imprinting seems an unlikely answer as to why the studies disagree. Still, although our study shows that Atp10a is not imprinted in C57BL/6 and B6.CAST.c7 mice, it cannot be ruled out that Atp10a could be imprinted in other inbred strains. It is possible that Atp10a is not imprinted in mouse, but is imprinted in humans. Imprinting differences have been observed between the two species, although it is important to note that the imprinting status of ATP10A is not completely clear in humans either. Despite evidence from

PAGE 81

81 several sources supporting predomin antly maternal expression of ATP10A Hogart et al. observed variation in human monoallelic expression of ATP10A influenced by gender and a genetic variant. Even more curious is their di scovery of a patient with PWS due to uniparental disomy that exhibited monoallelic expression of ATP10A indicating monoallelic expression of the gene can be separa te from imprinting. The PWS-ICflox6kb Mouse The PWS-ICflox6kb mouse will be extremely useful to the future of PWS research. Aside from redefining the borders of the murine PWS -IC, there are several que stions that the PWSICflox6kb mouse will be essential to answering such as when the imprint is set and where PWS gene expression is most critical. The PWS-ICflox6kb mouse may even be the key to a new PWS model that more closely resembles the disease phenotype in humans. It is possible that the imprint set by the PWS-IC becomes locked in on the paternal chromosome and the PWS-IC is not required for ma intenance of the imprint. There is very little information available on that topic aside from one report of a man who is a mosaic for a PWS-IC deletion on his paternal chromosome Data from this individual indicates that the PWS-IC is required for postzygotic maintenance of the paternal imprint92. Whether the PWS-IC is necessary throughout life to maintain the paternal imprint or is only ne cessary to protect the paternal allele during the globa l deand remethylation cycle in the early embryo is unknown. It is possible to use the PWS-ICflox6kb mouse in conjunction with an i nducible cre transgenic mouse to ask if the role of the PWS-IC stops after the in itial setting of the imprint, or is necessary for mitotic maintenance of imprinted gene expression. The ability to delete the PWS-IC only in cer tain tissues while sparing others using a tissue specific cre opens up seve ral possibilities. There has be en a focus on the hypothalamus in PWS research because many PWS features incl uding hypogonadism and reduced satiety are due

PAGE 82

82 to hypothalamic insufficiency93. It is possible cross PWS-ICflox6kb mice to transgenic mice that express cre in the hypothalamus, another part of the brain, or any other tissue of interest to explore the role of the PWS genes specifically in that tissue. Although the post natal failure to thrive, the loss of PWS gene expression and the methlylation changes at CpG islands observed in the current 35 kb PWS-IC deletion mice are very similar to observations in PWS patients, th e mice never develop hyper phagia or obesity. It will be interesting to see whether a late onset deletion of the PWS-IC in mouse is able to abrogate the early postnatal fail ure to thrive, perhaps leading to a phenotype more similar to the human disease. If the deletion of the PWS-IC late in developmen t does indeed lead to a failure to maintain the paternal imprint and a loss of e xpression of paternally expressed genes, it should be possible to bypass the early failure to thrive phenotype by crossing the PWS-ICflox6kb mouse to a late expressing cre mouse. A mouse model with obesity and hyperphagia would be an enormously valuable resource to PWS research, particularly since obesity is the prime cause of morbidity and mortality in PWS patients93. These are just a few examples of the re search that is possible with the PWS-ICflox6kb mouse and an appropriate transgenic cre mouse. The PWS-ICflox6kb mouse will hopefully lead to new and exciting discoveries about PWS and impr inting, and ultimately aid in the understanding and treatment of PWS and other imprinting disorders.

PAGE 83

83 LIST OF REFERENCES 1. McGrath, J. & Solter, D. Nuclear transplant ation in the mouse embryo by microsurgery and cell fusion. Science 220, 1300-2 (1983). 2. Surani, M.A., Reik, W., Norris, M.L. & Barton, S.C. Influence of germline modifications of homologous chromosomes on mouse development. J Embryol Exp Morphol 97 Suppl 123-36 (1986). 3. McGrath, J. & Solter, D. Completion of mous e embryogenesis requires both the maternal and paternal genomes. Cell 37, 179-83 (1984). 4. Surani, M.A. & Barton, S.C. Developm ent of gynogenetic eggs in the mouse: implications for parthenogenetic embryos. Science 222, 1034-6 (1983). 5. Surani, M.A., Barton, S.C. & Norris, M.L. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308, 548-50 (1984). 6. Cattanach, B.M. Parental origin effects in mice. J Embryol Exp Morphol 97 Suppl 13750 (1986). 7. Hore, T.A., Rapkins, R.W. & Graves, J.A. C onstruction and evolution of imprinted loci in mammals. Trends Genet 23, 440-8 (2007). 8. Varmuza, S. & Mann, M. Genomic impr inting--defusing the ovarian time bomb. Trends Genet 10, 118-23 (1994). 9. Solter, D. Refusing the ovarian time bomb. Trends Genet 10, 346; author reply 348-9 (1994). 10. Haig, D. & Westoby, M. Parent-Specific Ge ne Expression and the Triploid Endosperm. The American Naturalist 134, 147-155 (1989). 11. Foltz, D.W. Genetic Evidence for Long-Term Monogamy in a Small Rodent, Peromyscus polionotus. The American Naturalist 117, 665-675 (1981). 12. Birdsall, D.A. & Nash, D. Occurrence of Su ccessful Multiple Insemi nation of Females in Natural Populations of Deer Mice (Peromyscus maniculatus). Evolution 27, 106-110 (1973). 13. Dawson, W.D. Fertility and Size Inherita nce in a Peromyscus Species Cross. Evolution 19, 44-55 (1965). 14. Vrana, P.B., Guan, X.J., Ingram, R.S. & T ilghman, S.M. Genomic imprinting is disrupted in interspecific Peromyscus hybrids. Nat Genet 20 362-5 (1998).

PAGE 84

84 15. Kitsberg, D. et al. Allele-specific rep lication timing of imprinted gene regions. Nature 364, 459-63 (1993). 16. Bartolomei, M.S. & Tilghman, S. M. Genomic imprinting in mammals. Annu Rev Genet 31, 493-525 (1997). 17. Li, E., Bestor, T.H. & Jaenisch, R. Target ed mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69 915-926 (1992). 18. Hark, A.T. et al. CTCF mediates methyla tion-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature 405, 486-9 (2000). 19. Barlow, D.P. Methylation and imprinting: from host defense to gene regulation? Science 260, 309-10 (1993). 20. Yoder, J.A., Walsh, C.P. & Bestor, T.H. Cytosine methylation and the ecology of intragenomic parasites. Trends in Genetics 13, 335-340 (1997). 21. Strahl, B.D. & Allis, C.D. The language of covalent histone modifications. Nature 403, 41-5 (2000). 22. Turner, B.M. Cellular memory and the histone code. Cell 111 285-91 (2002). 23. Zhang, Y. & Reinberg, D. Transcription re gulation by histone methylation: interplay between different covalent modifica tions of the core histone tails. Genes Dev 15, 2343-60 (2001). 24. Richards, E.J. & Elgin, S.C. Epigenetic codes for heterochromatin formation and silencing: rounding up the usual suspects. Cell 108 489-500 (2002). 25. Henikoff, S. & Ahmad, K. Assembly of variant histones into chromatin. Annu Rev Cell Dev Biol 21, 133-53 (2005). 26. Nan, X. et al. Transcriptional repressi on by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393 386-9 (1998). 27. Plath, K., Mlynarczyk-Evans, S., Nusinow D.A. & Panning, B. Xist RNA and the mechanism of X chromosome inactivation. Annu Rev Genet 36 233-78 (2002). 28. Morgan, H.D., Santos, F., Green, K., Dean, W. & Reik, W. Epigenetic reprogramming in mammals. Hum Mol Genet 14 Spec No 1 R47-58 (2005). 29. Okano, M., Bell, D.W., Haber, D.A. & Li E. DNA Methyltransf erases Dnmt3a and Dnmt3b Are Essential for De Novo Met hylation and Mammalian Developm ent. Cell 99, 247-257 (1999).

PAGE 85

85 30. Williams, C.A., Frias, J.L. & Opitz, J.M. The Angelman ("happy puppet") syndrome. American Journal of Medical Genetics 11, 453-460 (1982). 31. Kishino, T., Lalande, M. & Wagstaff, J. UBE3A/E6-AP mutations cause Angelman syndrome. Nat Genet 15, 70-3 (1997). 32. Matsuura, T. et al. De novo truncating mutations in E6-A P ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nat Genet 15, 74-7 (1997). 33. Nicholls, R.D. & Knepper, J.L. Genome organization, function, and imprinting in PraderWilli and Angelman syndromes. Annu Rev Genomics Hum Genet 2, 153-75 (2001). 34. Lossie, A.C. et al. Distinct phenotypes di stinguish the molecular classes of Angelman syndrome. J Med Genet 38, 834-45 (2001). 35. Kanber, D. et al. A paternal deletion of MKRN3, MAGEL2 and NDN does not result in Prader-Willi syndrome. Eur J Hum Genet 17, 582-90 (2009). 36. Landers, M. et al. Regulati on of the large (approximately 1000 kb) imprinted murine Ube3a antisense transcript by alterna tive exons upstream of Snurf/Snrpn. Nucleic Acids Res 32, 3480-92 (2004). 37. Peters, J. Prader-Willi and snoRNAs. Nat Genet 40, 688-9 (2008). 38. Kishore, S. & Stamm, S. The snoRNA HB II-52 regulates alternative splicing of the serotonin receptor 2C. Science 311, 230-2 (2006). 39. Burger, J., Horn, D., Tonnies, H., Neitzel, H. & Reis, A. Familial interstitial 570 kbp deletion of the UBE3A gene region causing Angelman syndrome but not Prader-Willi syndrome. Am J Med Genet 111, 233-7 (2002). 40. Sahoo, T. et al. Prader-Willi phenotype cause d by paternal deficiency for the HBII-85 C/D box small nucleolar RNA cluster. Nat Genet 40, 719-21 (2008). 41. Hogart, A., Patzel, K.A. & LaSalle, J.M. Gender influences m onoallelic expression of ATP10A in human brain. Hum Genet 124, 235-42 (2008). 42. Brannan, C.I. & Bartolomei, M.S. Mechanisms of genomic imprinting. Curr Opin Genet Dev 9, 164-70 (1999). 43. Ohta, T. et al. Imprinting-mutation mechanisms in Prader-Willi syndrome. Am J Hum Genet 64, 397-413 (1999). 44. Buiting, K., Lich, C., Cottrell, S., Barnicoa t, A. & Horsthemke, B. A 5-kb imprinting center deletion in a family with Angelman syndrome reduces the shortest region of deletion overlap to 880 bp. Hum Genet 105, 665-6 (1999).

PAGE 86

86 45. Rapkins, R.W. et al. Recent assembly of an imprinted domain from non-imprinted components. PLoS Genet 2, e182 (2006). 46. Chai, J.H., Locke, D.P., Ohta, T., Greally, J.M. & Nicholls, R.D. Retrotransposed genes such as Frat3 in the mouse Chromosome 7C Prader-Willi syndrome region acquire the imprinted status of their insertion site. Mamm Genome 12, 813-21 (2001). 47. Farber, C., Gross, S., Neesen, J., Buiting, K. & Horsthemke, B. Identif ication of a testisspecific gene (C15orf2) in the Prader -Willi syndrome region on chromosome 15. Genomics 65, 174-83 (2000). 48. Skryabin, B.V. et al. Deletion of the MB II-85 snoRNA gene cluster in mice results in postnatal growth retardation. PLoS Genet 3, e235 (2007). 49. Kozlov, S.V. et al. The imprinted gene Magel2 regulates normal circadian output. Nat Genet 39, 1266-72 (2007). 50. Hara, J., Yanagisawa, M. & Sakurai, T. Di fference in obesity phenotype between orexinknockout mice and orexin neuron-deficient mice with same genetic background and environmental conditions. Neurosci Lett 380, 239-42 (2005). 51. Yang, T. et al. A mouse model for Prader-Willi syndrome imprinting-centre mutations. Nat Genet 19 25-31 (1998). 52. Tsai, T.F., Jiang, Y.H., Bressler, J., Arms trong, D. & Beaudet, A.L. Paternal deletion from Snrpn to Ube3a in the mouse causes hypotonia, growth reta rdation and partial lethality and provides evidence for a gene contributing to Prader-Willi syndrome. Hum Mol Genet 8 1357-64 (1999). 53. Gerard, M., Hernandez, L., Wevrick, R. & St ewart, C.L. Disruption of the mouse necdin gene results in early post-natal lethality. Nat Genet 23, 199-202 (1999). 54. Muscatelli, F. et al. Disrup tion of the mouse Necdin gene results in hypothalamic and behavioral alterations reminiscent of the human Prader-Willi syndrome. Hum Mol Genet 9, 3101-10 (2000). 55. Tsai, T.F., Armstrong, D. & Beaudet, A.L. N ecdin-deficient mice do not show lethality or the obesity and infertility of Prader-Willi syndrome. Nat Genet 22, 15-6 (1999). 56. Nicholls, R.D. Incriminating gene suspects, Prader-Willi style. Nat Genet 23, 132-4 (1999). 57. Ding, F. et al. SnoRNA Snord116 (Pwcr1/MBI I-85) deletion causes growth deficiency and hyperphagia in mice. PLoS ONE 3, e1709 (2008).

PAGE 87

87 58. Jiang, Y.-h. et al. Mutation of the Angelman Ubiquitin Ligase in Mice Causes Increased Cytoplasmic p53 and Deficits of Contextual Learning and Long-Term Potentiation. Neuron 21, 799-811 (1998). 59. Miura, K. et al. Neurobehavioral an d Electroencephalographic Abnormalities in Ube3aMaternal-Deficient Mice. Neurobiology of Disease 9, 149-159 (2002). 60. Ding, F. et al. Lack of Pwcr1/MBII-85 snoRNA is critical for neonatal lethality in PraderWilli syndrome mouse models. Mamm Genome 16 424-31 (2005). 61. Bressler, J. et al. The SNRPN promoter is not required for genomic imprinting of the PraderWilli/Angelman domain in mice. Nat Genet 28, 232-40. (2001). 62. Tsai, T.F., Jiang, Y.H., Bressler, J., Arms trong, D. & Beaudet, A.L. Paternal deletion from Snrpn to Ube3a in the mouse causes hypotonia, growth retardation and partial lethality and provides evidence for a gene contributing to Prader-Willi syndrome. Hum Mol Genet 8 1357-64. (1999). 63. Chamberlain, S.J. & Brannan, C.I. The Prader-Willi syndrome imprinting center activates the paternally expressed murine Ube3a antisens e transcript but represses paternal Ube3a. Genomics 73, 316-22 (2001). 64. Johnstone, K.A. et al. A human imprinting centre demonstrates conserved acquisition but diverged maintenance of imprinting in a mouse model for Angelman syndrome imprinting defects. Hum Mol Genet 15 393-404 (2006). 65. Peery, E.G., Elmore, M.D., Resnick, J.L., Brannan, C.I. & Johnstone, K.A. A targeted deletion upstream of Snrpn does not result in an imprinting defect. Mamm Genome 18 255-62 (2007). 66. Wu, M.Y. et al. Mouse imprinting defect mutations that model Angelman syndrome. Genesis 44, 12-22 (2006). 67. Shemer, R. et al. The imprinting box of the Prader-Willi/Angelman syndrome domain. Nat Genet 26 440-3 (2000). 68. Kantor, B., Makedonski, K., Green-Finberg, Y., Shemer, R. & Razin, A. Control elements within the PWS/AS imprinting box and their function in the imprinting process. Hum Mol Genet 13, 751-62 (2004). 69. Cattanach, B.M. et al. A candidate mous e model for Prader-Willi syndrome which shows an absence of Snrpn expression. Nat Genet 2, 270-4 (1992). 70. Cattanach, B.M. et al. A candidate mode l for Angelman syndrome in the mouse. Mamm Genome 8, 472-8 (1997).

PAGE 88

88 71. Gabriel, J.M. et al. A transgene insertion creating a heritable chro mosome deletion mouse model of Prader-Willi and angelman syndromes. Proc Natl Acad Sci U S A 96, 9258-63 (1999). 72. Minassian, B.A. et al. Angelman syndrome: correlations between epilepsy phenotypes and genotypes. Ann Neurol 43, 485-93 (1998). 73. Herzing, L.B., Kim, S.J., Cook, E.H., Jr. & Ledbetter, D.H. The human aminophospholipid-transporting ATPase gene ATP10C maps adjacent to UBE3A and exhibits similar imprinted expression. Am J Hum Genet 68, 1501-5 (2001). 74. Meguro, M. et al. A novel maternally e xpressed gene, ATP10C, encodes a putative aminophospholipid translocase associ ated with Angelman syndrome. Nat Genet 28, 1920 (2001). 75. Cook, E.H., Jr. et al. Autism or atypical autis m in maternally but not paternally derived proximal 15q duplication. Am J Hum Genet 60, 928-34 (1997). 76. Kim, S.J. et al. Mutation screening and tr ansmission disequilibrium study of ATP10C in autism. Am J Med Genet 114, 137-43 (2002). 77. Kato, C. et al. Association study of th e 15q11-q13 maternal ex pression domain in Japanese autistic patients. Am J Med Genet B Neuropsychiatr Genet 147B 1008-12 (2008). 78. Dhar, M.S., Webb, L. S., Smith, L., Hauser, L. Johnson, D., West, D.B. A novel ATPase on mouse chromosome 7 is a candida te gene for increased body fat. Physiol Genomics 4, 93-100 (2000). 79. Dhar, M.S. et al. Mice heterozygous for At p10c, a putative amphipat h, represent a novel model of obesity and type 2 diabetes. J Nutr 134, 799-805 (2004). 80. Kayashima, T. et al. Atp10a, the mouse or tholog of the human imprinted ATP10A gene, escapes genomic imprinting. Genomics 81, 644-7 (2003). 81. Kashiwagi, A. et al. Predominant mate rnal expression of the mouse Atp10c in hippocampus and olfactory bulb. J Hum Genet 48, 194-8 (2003). 82. Kayashima, T., Ohta, T., Niikawa, N. & Kishino, T. On the conflicting reports of imprinting status of mouse ATP10a in th e adult brain: strain-background-dependent imprinting? J Hum Genet 48, 492-3; author reply 494 (2003). 83. Clark, S.J., Harrison, J., Paul, C.L. & Frommer, M. High sensitivity mapping of methylated cytosines. Nucl. Acids R es. 22, 2990-2997 (1994).

PAGE 89

89 84. Albrecht, U. et al. Imprinted expression of the murine Angelman syndrome gene, Ube3a, in hippocampal and Purkinje neurons. Nat Genet 17, 75-8 (1997). 85. Landers, M. et al. Maternal disruption of Ube3a leads to increased expression of Ube3aATS in trans. Nucleic Acids Res 33, 3976-84 (2005). 86. Rodriguez-Jato, S., Nicholls, R.D., Driscoll, D.J. & Yang, T.P. Characterization of cisand trans-acting elements in the imprinted human SNURF-SNRPN locus. Nucleic Acids Res 33, 4740-53 (2005). 87. Buiting, K. et al. Disruption of the bipartite imprinting center in a family with Angelman syndrome. Am J Hum Genet 68, 1290-4 (2001). 88. Swiatek, P.J. & Gridley, T. Perinatal lethal ity and defects in hindbrain development in mice homozygous for a targeted mutation of the zinc finger gene Krox20. Genes Dev 7, 2071-84 (1993). 89. Smith, A.J., Xian, J., Richardson, M., Johnstone, K.A. & Rabbitts, P.H. Cre-loxP chromosome engineering of a targeted dele tion in the mouse corresponding to the 3p21.3 region of homozygous loss in human tumours. Oncogene 21, 4521-9 (2002). 90. Schwenk, F., Baron, U. & Rajewsky, K. A cretransgenic mouse strain for the ubiquitous deletion of loxP-flanked gene segments including deletion in germ cells. Nucleic Acids Res 23, 5080-1 (1995). 91. Jiang, S., Hemann, M.A., Lee, M.P. & Feinbe rg, A.P. Strain-dependent developmental relaxation of imprinting of an endogenous mouse gene, Kvlqt1. Genomics 53, 395-9 (1998). 92. Bielinska, B. et al. De novo deletions of SNRPN exon 1 in early human and mouse embryos result in a paternal to maternal imprint switch. Nat Genet 25, 74-8 (2000). 93. Cassidy, S.B. Prader-Willi syndrome. J Med Genet 34, 917-23 (1997).

PAGE 90

90 BIOGRAPHICAL SKETCH Am anda DuBose grew up in Savannah, Georgia. She graduated valedictorian of the class of 1997 from Windsor Forest High School and graduated magna cum laude from Armstrong Atlantic State University with a bachelors degree in biology.