spindle assembly checkpoint and chromosomal instability

Spindle Assembly Checkpoint And Chromosomal Instability

Ricardo João Silva e Sousa Mestrado de Bioquímica 2012 Orientador Sara Morais da Silva, PhD

FCUP/ICBAS Spindle Assembly Checkpoint And Chromosomal Instability

ii

Acknowledgments/Agradecimentos

Ao Professor Claudio Sunkel por me ter conferido a oportunidade de desenvolver a

minha tese no seu laboratório e pela preocupação demonstrada a cada passo do seu

desenvolvimento, um muito obrigado.

À Sara Morais da Silva um grande obrigado pela orientação, incentivo e principalmente

por me ter ajudado a desenvolver como investigador dando me os instrumentos para

os próximos passos e como pessoa aumentando o meu sentido de responsabilidade.

Obrigado também pela paciência demonstrada a cada passo menos conseguido por

minha parte. Por fim um obrigado por seres a única pessoa que também gosta de

misturar o português e inglês na mesma frase. Muito obrigado.

Agradeço a todos os restantes membros do GMM pela ajuda prestada, e pela

disponibilidade demonstrada.

Agradeço a quem me aturou diariamente na “sala das moscas”, nomeadamente ao

grupo de Biologia do Desenvolvimento com um especial agradecimento à Sara que

embora não tenha “apostado” em mim (ler com rancor a parte anterior) ainda me ouviu

diversas vezes, quando nem tudo correu pelo melhor, e me deu conselhos valiosos.

Quero também agradecer à Ana Cunha pela falta de “tento na língua” que gerou bons

momentos de diversão durante as pausas de trabalho. Muito obrigado a todos.

Aos amigos e colegas de Mestrado que fizeram esta viagem comigo, um grande

obrigado pelo apoio demonstrado.

Aos meus ex-companheiros de casa Leo e Stéphane que mesmo longe souberam

sempre guiar me no melhor sentido e ouvir me sempre que necessitei.

À Ana Lúcia, Pedro Albuquerque e Sofia Pinho pelas horas intermináveis de apoio e

conselhos e principalmente por terem sido o que eu considero, mentores não só para

a minha carreira científica mas também fora dela. “Miau miau” Um grande obrigado.

Ao André Cunha por constantemente congeminar nos bastidores para eu não me

distrair e escrever (sim eu “topava-te”). Obrigado “sócio”.


iii

Ao Sérgio Dias, o verdadeiro companheiro de escrita, que fez esta jornada comigo.

Um grande grande obrigado por ouvires e aconselhares e pelo característico humor

seco que tornavam horas de trabalho em.. bem, horas de trabalho na mesma porém

divertidas. “Estão duas gaivotas...” Um obrigado “a tender para o infinito” Sérgio.

Ao Paulo Almeida que poderia não ter apoiado minimamente durante esta tese e

estaria aqui mencionado na mesma. Não teria páginas suficiente para agradecer tudo

o que quereria mas quero apenas mencionar uma mítica frase dita alguns anos atrás

“por ti eu falt…” … “Palo” obrigado por esta amizade ser “para bom entendedor nem

meia palavra é necessária” Um grande grande obrigado.

À Rita Nobre que mais um bocado e teria que lhe dedicar esta tese. Um muito

obrigado não chega para compensar todo o trabalho, tempo, horas de sono perdidas e

paciência que tiveste durante a escrita desta tese. “Rita Nobre tem sempre….” Rita

Nobre…Obrigado… (“B.T.P.M.T.A.T.P.W.”)

À Cátia Mendes que acompanhou desde o início todas as peripécias, problemas, altos

e baixos desta tese. Da tua forma extremamente peculiar tiveste sempre a perícia de

me por no caminho certo. O teu apoio foi “para lá” de fundamental. Podia dissertar por

páginas intermináveis mas basta me agradecer por poder explicar 99% desta amizade

com “é a Cátia”. Cátia Mendes, obrigado…

À minha irmã pela paciência interminável. Sou suspeito eu sei mas dificilmente

conseguiria uma melhor. Já agradeci a paciência interminável? Mana muito obrigado.

À minha família que movem mundos e fundos para que eu seja bem sucedido. Queria

fazer um agradecimento especial aos meus avós. Cada um com seu ensinamento

ajudo-me a tornar a pessoa que sou hoje. Obrigado por me ensinarem como

descontrair e despreocupar quando necessário, o valor da inteligência e da cultura, o

valor da paciência e da diplomacia e o valor da família. Muito, muito obrigado.

Aos meus pais que por mais que eu me proteste é impossível, literalmente, alguém

fazer mais por mim do que vocês fizeram e continuam a fazer. Vocês são a base de

tudo o que eu me tornei e não consigo agradecer o suficiente. Mãe, Pai, “Velhote e

velhota”, obrigado.


iv

Abstract

The spindle assembly checkpoint (SAC) mechanism works through the

inhibition of APC/C (anaphase promoting complex/cyclosome) in order to delay in the

metaphase-anaphase transition. This mechanism is initiated in the presence of

unattached, unaligned and/or missattached chromosomes and therefore can prevent

abnormal chromosome segregation. Severe chromosome missegregation results in

aneuploidy and when this is very extensive cell death ensues. Given that chromosome

instability and aneuploidy have been closely associated with tumorigenesis, we

decided to analyze what happens to cells that do not have the SAC but at the same

time cannot activate apoptosis. Previous work in the lab using Drosophila had shown

that in these conditions, cells show all the signs of tumorigenesis including

hyperproliferation and chromosome instability. The purpose of this work was to apply

high throughput microarray technology, to determine the transcriptional profile of cells

in which the SAC was abolished and apoptosis was prevented. For this, we used wing

imaginal discs where a SAC gene (Bub3) expression was knocked down through RNA

interference and apoptosis is inhibited by the expression of the baculovirus protein p35.

The transcriptional profile of these cells was then compared to that of cells in which

only bub3 is knocked down, where we only inhibited apoptosis through p35 or in wild

type wing imaginal discs expressing only green fluorescent protein. As a result we

identified more that 200 genes that where specifically up or down regulated and

initiated a RNAi genetic screen using flies where Bub3 was knocked down and the

apoptosis blocked only in the wing imaginal discs. The preliminary results suggest that

the screen should be able to identify candidates that actively contribute for the

tumorigenic phenotype of cells in which the SAC is inactivated and apoptosis is at the

same time inhibited.

Keywords: SAC, apoptosis, tumorigenesis, Bub3, Drosophila, microarrays,

RNAi, screen.


v

Resumo

Através da inibição do APC/C (complexo promotor da anafase) o mecanismo

do “spindle assembly checkpoint” (SAC) consegue provocar um atraso na transição

metafase-anafase no caso da existência de cromossomas incorretamente ligados a

microtúbulos prevenindo assim a segregação anormal. A ocorrência da separação

incorreta dos cromossomas provoca aneuploidia e consequente morte celular. Uma

vez que tumorigénese se encontra associado com instabilidade cromossómica e

aneuploidia decidimos analisar o que acontece a células que não possuem SAC mas

que ao mesmo tempo não conseguem ativar morte celular. Experiências prévias

demonstraram que em Drosophila as condições mencionadas levam a tumorigénese e

instabilidade cromossómica. O objetivo do nosso trabalho foi aplicar tecnologia de

microarrays para determinar o perfil trancricional de células em que SAC e apoptose

foram inibidos. Para tal, utilizamos discos imaginais da asa onde silenciamos através

de RNA de interferência Bub3 (um gene do SAC) e inibimos apoptose com P35 uma

proteína de um baclovirus. Este perfil trancricional foi comparado com outros perfis de

diferentes genótipos, nomeadamente, quando apenas silenciávamos Bub3, quando

apenas inibíamos apoptose e com o fenótipo selvagem exprimindo apenas green

fluorescente protein. Como resultado identificamos mais de 200 genes que se

encontram sobreexpressos ou subexpressos que utilizamos para efetuar um screen

genético com RNA de interferência usando moscas onde silenciamos a expressão e

inibimos a apoptose em discos imaginais da asa. Resultados preliminares sugerem a

identificação de candidatos que contribuem ativamente para o fenótipo tumorigénico

de células onde SAC se encontra inativado e apoptose inibida.

Palavras-chave: SAC, apoptose, tumorigénese, Bub3, Drosofila, microarrays,

RNAi, screen


1

Index

Acknowledgments/Agradecimentos………………………………….……………… ii

Abstract……………………………………………………………………………….…… iv

Resumo……………………………………………………………………………………. v

Index……………………………………………………..………………………………… 1

Abbreviations List…………………………………………………………..…………….. 4

Chapter 1 - Introduction…………………………………………..……………………… 6

1.1. Model Organism……………………………………...……………………… 7

1.2. Tumorigenesis……..………………………………………………………… 8

1.3. Mitosis and Spindle Assembly Checkpoint (SAC) …………..…..………. 11

1.4. Apoptosis……………………...……………………………………………… 16

1.5. Main Technologies………...………………………………………………… 21

1.5.1.Microarrays………………..……………………………………………… 21

1.5.2. RNA Intereference…………….………………………………………… 23

1.6. Aims…………………………………………………………………………… 24

Chapter 2 - Materials and Methods……………………………………....…………….. 25

2.1. Fly stocks used……………………………………..…………………………… 26

2.2. RNA Extraction from Wing Imaginal Discs……………………………….….. 26

2.2.1.Collection of Wing Imaginal ……………………………………………. 26

2.2.3.RNA Extraction…………………………………………………………… 26

2.3. Microarrays……..……………………………………………………………….. 28

Chapter 3 - Results………………………..……………………………………………… 33


2

3.1. Introduction………………………………………………………………….…… 34

3.2. Selected Genotypes………………………………………….………………… 35

3.3. RNA Samples for Microarrays…………………………………………………. 37

3.3.1.Qualitative and Quantitative Analysis of the Extracted RNA..………. 37

3.3.1.1.Nanodrop Analysis………………..………………………….…… 37

3.3.1.2.Bioanalyzer Analysis……………………………………………… 38

3.4. Microarray Analysis…………………..………………………………………… 44

3.4.1. Microarray Data………………………………….………………………. 44

3.4.2.Mining the Microarray Data to Achieve Relevant Biological

Information………………………………………………………………………. 48

3.4.3.Validating Venn Diagrams through Bioinformatics Resources…….... 52

3.4.4.Mining the Microarray Data from Raw to Biologically Relevant.... 56

3.5.Screen for Enhancer/Suppression of Tumor Phenotype……………………. 58

Chapter 4 - Discussion……………………………………………...……………………. 65

4.1. Introduction………………………………………………………………………. 66

4.2. Reasons for Phenotype Selection…………………….………………………. 66

4.3. Microarray Comparison Reveals Significant Differences in Expression

Profiles Depending on the Genotypes…..…………………………..…………….. 67

4.4.Bioinformatic Methods Show Enrichment of Relevant Pathways from

Venn Diagram Selection………………………..…………………………………… 68

4.5.Legitimacy of Manual Selection from Venn Diagram………………………... 70

4.6.Bub3 has an Additional Role as a Tumor Suppressor………………………. 71

4.6.1.Genes Upregulated in Bub3rnai background Screen…………....…….. 71

4.6.2.Genes Downregulated in Bub3rnai background Screen……….……… 73


3

4.6.3.Genes Upregulated in Bub3rnai DIAP1 background Screen…….…… 75

4.6.4.Genes Upregulated in Bub3rnai DIAP1 background Screen….……… 80

Chapter 5 - References…………………………………..………………………………. 83


4

Abbreviations list

AcCoAS - Acetyl Coenzyme A synthase

alphaTub85E - alpha-Tubulin at 85E

Alr - Augmenter of liver regeneration

APC/C - Anaphase promoting complex/ cyclosome

Aplip1 - APP-like protein interacting protein 1

asf1 - anti-silencing factor 1

B-H1 - BarH1

Btl - breathless

Cam - Calmodulin

CanB - Calcineurin B

Damm - Death associated molecule related to Mch2

Ddc - Dopa decarboxylase

Dhc36C - Dynein heavy chain at 36C

Dhc62B - Dynein heavy chain at 62B

Drep-4 - DNA fragmentation factor-related protein 4

f-cup - flyers-cup

fz - frizzled

GFP - Green fluorescence protein

Gyc-89Da - Guanylyl cyclase at 89Da

InR - Insulin-like receptor

Irbp - Inverted repeat-binding protein

Kap3 - Kinesin associated protein 3

Mdr50 - Multi drug resistance 50

mre11 - meiotic recombination 11

mus205 - mutagen-sensitive 205

mus304- mutagen-sensitive 205

ncd - non-claret disjunctional

nimB1 - nimrod B1

nimC1 - nimrod C1

nimC4 - nimrod C4

os - outstretched

Osi6 - Osiris 6

Otu - ovarian tumor


5

Phlpp - PH domain leucine-rich repeat protein phosphatase

Phr - photorepair

phr6-4 - (6-4)-photolyase

phyl - phyllopod

Pkc53E - Protein C kinase 53E

PPP4R2r - Protein phosphatase 4 regulatory subunit 2-related protein

Prx5037 - Peroxiredoxin 3

Puc - puckered

Rep - Rab escort protein

RNAi - RNA interference

SAC - Spindle assembly checkpoint

Sns - sticks and stones

Socs36E - Suppressor of cytokine signaling at 36E

Spn100A - Serpin 100A

Spn28Da - Serpin 28Da

Spn28Db - Serpin 28Db

Spn43Aa - Serpin 43Aa

ST6Gal - Sialyltransferase

Top1 - Topoisomerase 1

Traf4 - TNF-receptor-associated factor 4

Tws - twins

UAS - Upstream Activating Sequence

Wnt2 - Wnt oncogene analog 2

Wun - Wunen

wun2 - Wunen - 2


6

CHAPTER 1

INTRODUCTION


7

1.1. MODEL ORGANISM

Drosophila melanogaster has proven to be a very influential model for genetic and

biology studies in human diseases (1). Thanks to genomic sequencing, it is possible to

state that 70% of all known human diseases genes have functional orthologs in

Drosophila and some of the components in major signal transduction pathways that

play essential roles in neural development and cancer, like EGFR/RTK-Ras, PI3K,

Notch, Wnt, Jak-STAT, Hedgehog, and TGF-b are conserved (2-9). Furthermore, many

core components of these pathways were first discovered in Drosophila since it was

one of the first experimental systems to demonstrate that cancer is linked to heritable

genetic mutations as one of the first recessive lethal mutations found in Morgan’s

laboratory, lethal 7, revealed to give rise to transplantable, malignant tumors (10).

Drosophila offers advantages regarding modeling human diseases mainly because

present techniques permit cell type specific manipulation approach of gene expression

even in a single cell in vivo within the CNS and other tissues (11). Also, when

identifying and characterizing new genes, the Drosophila model brings several

advantages through different genetic tools that are able to provide us with information,

such as: the well-annotated and sequenced fly genome, the availability of thousands of

loss-of-function mutants for the majority of genes, the possibility of knocking-out genes

with targeted mutagenesis, insertional mutagenesis, or RNA interference techniques.

Gene overexpression with temporal and spatial precision is also possible using

technology like the binary transgene expression system, Gal4/UAS system, which

allows the expression of multiple transgenes in the same organism (12-14). In this

system, we are able to express the target transgene that is cloned downstream of a

UAS sequence in the same tissue specific pattern as the Gal4 activator. Furthermore,

the ability to introduce the transcriptional activator and the target UAS transgene in

distinct parental lines allows for the viability of parental stocks even in the presence of

toxic transgenes. A transgenic Gal4 Drosophila line can be used as a general resource

due to the fact that, once generated, this line expressing Gal4 in a given spatial pattern

can be crossed to any UAS target line. The opposite is also true as, once generated, a

Drosophila UAS target line can be driven anywhere in the embryo, larvae or adult fly

depending on which Gal4 line was selected to cross (15, 16).


8

1.2. TUMORIGENESIS

Cancer is a leading cause of death worldwide and is described as an uncontrolled

malignant growth of cells although its lethality often comes as a result of secondary

tumors, normally spread from the original site through the blood or lymphatic system

becoming highly invasive and aggressive (17, 18). A significant number of cancer are

associated with chromosomal instability, however, whether tumorigensis is the cause

or the consequence of chromosome instability is not known. Extensive work relates

weakened mitotic checkpoint to the transformation process. In mice, Mad2

heterozygotes (19) or hypomorphic alleles of BubR1 (20) produce a weakened

checkpoint and usually result in spontaneous tumors with a frequency of 27% and 6%

respectively. Heterozygous mice for Bub3 (21) or BubR1 (22) exposed to a carcinogen

also showed high percentage of tumorigenesis (70% in Bub3 heterozygotes versus

50% in wild-type and 11% in BubR1 heterzygote versus 0.4% in wildtype). In humans,

Cahill et al. in 1998 reported that 2 out of 19 colorectal cancer cell lines exhibit

heterozygous point mutations in the Bub1 and BubR1 genes (23). This led to an

extensive search for mutations in mitotic checkpoint genes in several tumors and tumor

cell lines. As an example, human patients with a rare recessive disorder mosaic

variegated aneuploidy (MVA) show a raise in aneuploidy rate and childhood cancers

(24). Some MVA patients have mutations in both alleles of the BubR1 gene (25, 26).

On the other hand, not many mitotic checkpoint genes mutations have been identified

in human cancers, however, altered transcription levels of these proteins associated

with hypermethylated promoters are thought to be very frequent (27). Accordingly

human tumor cells and cell lines show a failure in maintaining a mitotic checkpoint

signaling. The underlying concept is that weakened checkpoint signaling yet sufficient

to maintain a viable population of cells leads to a missegregation of chromosomes in

cell division and the consequently aneuploidy and chromosomal instability (28).

Genetic studies in Drosophila have helped to identify many genes that when

mutated or dysregulated lead to tumorigenesis. For instance, some players in the

Hippo pathway, defined as hyperplastic tumor suppressors, are involved in increased

proliferation or survival although not impeding tissue structure or differentiation. On the

other hand, some neoplastic tumor suppressors like Lgl involved in the apical-basal cell

polarity are usually responsible for problems in differentiation and failure to exit cell

cycle (29). An important aspect has been the ability to study these effects in the context

of the whole organism possible through several genetic and cell biology techniques.

For instance, the ability to create clones of mutant tissue aided researchers in their


9

quest to understand the role of microenvironment in tumor development and cell

competition mechanisms that offer a malignant cell with growth advantage

characteristics (29). Moreover, adding research using other model organisms, the

knowledge about cancer initiation and progression has greatly increased, even

revealing unknown molecular components and pathways, and helped improve

mammalian model systems for cancer (29). Using Drosophila as a model organism it

has been possible to identify several relevant genes highly studied in humans. Genes

like Notch, Hedgehog (Hh), Wingless, Runt, and Trithorax among others, have been

extensively elucidated in animal models and they seem to play a major role in the

pathophysiology of human cancer (51). Thomas Hunt Morgan described the Notch

mutation almost a century ago when he observed mutant flies with what appeared to

be notches in their wingblades, however only much later molecular analysis revealed

the importance of this highly conserved pathway (30, 31). The phenotype of this

mutation exhibits wings incised at tips and often along edges in heterozygous females

and lethality in males and homozygous females (32). Another interesting feature is the

lethal Notch-like phenotypes shown by Loss-of function mutations in the Drosophila

presenilin gene (33). Notch is related to cancer in humans through NOTCH1, highly

homologous to the Drosophila gene, found in chromosome 9 upon the translocation

t(7;9)(q34;q34.3) in a case of acute T cell lymphoblastic leukemia (34). The Hedgehog,

Wingless and runt genes were discovered by Christiane Nüsslein-Volhard and Eric

Wieschaus through their famous screen looking for embryonic mutants in Drosophila

(35) Phenotypically, hh homozygous embryos show the posterior naked portion of the

ventral surface of each segment deleted being replaced by a mirror image of the

anterior denticle belts. The length of wild-type larvae is greatly diminished and appears

to have a hedgehog-like appearance with a lawn of denticles (32). The association of

this gene to human cancer was first identified through the mutation of patched, the

human homolog of Drosophila, in the Gorlins syndrome, a rare syndrome linked with a

number of skeletal, skin, and neural abnormalities and usually related to the

development of multiple skin basal cell carcinomas, and a considerable increase in risk

for medulloblastoma and rhabdomyosarcoma (36). Wingless was originally identified

through a recessive mutation that confers an absence of one or both wings and/or

halters (37). Observation of embryos with embryonic lethal alleles showed a mirror-

image duplication of the denticle belts at the cost of naked cuticle (32). In human

cancer, wingless is dysregulated in the event of the loss of an APC allele, associated

with most colon carcinomas (38). Drosophila embryonic patterning and segmentation

studies revealed the pair-rule gene Runt. Runt embryos are shorter due to possessing


10

half the normal number of denticle belts (35). Mirror image duplications substitute the

deleted regions of what is left of the more anterior pattern elements (32). Runt relation

with human cancer is associated with the discovery of cDNAs of the alpha subunit of

polyomavirus binding protein 2 that is considerably homologous to that of Drosophila

segmentation gene runt (run). In humans RUNX1/AML1 also shows a run homology

region and it is a gene associated with the translocation t(8;21) (39). Interestingly,

AML1 is usually a target of chromosome translocations related with leukemia (40). For

instance, t(3;21) translocation in cases of myelodysplasia and blast transformation of

chronic myelogenous leukemia there is a fusion between either AML1 and EVI1 gene

(zinc fingercontaining transcription factor) or AML1 with EAP or MDS1 which function is

still unkown (41). The identification of the trithorax gene was due to its part of

maintaining expression of Antennapedia and bithorax homeotic gene complexes in

Drosophila melanogaster (42). Analyzing mutants demonstrate transformations of the

first and the third thoracic segments to the second thoracic segment, and the abdomen

showed alterations (32). In humans, acute leukemias are associated with the

MLL/ALL1 gene present in the human chromosome 11 band q23 usually involved in

interstitial deletions or reciprocal translocations between this region and chromosomes

1, 4, 6, 9, 10, or 19. Regarding homology, three regions are considered homologous to

sequences of the Drosophila trithorax gene (43, 44). The existence of translocations

involving MLL in leukemia is usually associated with unique clinical and biologic

features and with poor prognosis. Interestingly, topoisomerase II inhibitors can lead to

this disease when used to treat other malignancies since it causes double strand

breaks in DNA, leading to chromosomal MLL fusions. This feature might indicate an

involvement of the DNA repair pathway although information on such is still scarce

(45). The longstanding question was how such a large amount of different proteins

could give rise to the same disease. At first the suggestion involved the importance of a

dimerization of the amino-terminus of MLL (46). Furthermore, proteins fused with MLL

can be divide in two categories: nuclear and cytomplasmic and only six frequent

partners were responsible for over 85% of the diseases (47, 48). However, with access

to gene array technology, expression profiles between MLL and non-MLL leukemic

samples could be compared and showed a different pattern of gene expression that

discriminate that group from other leukemias (49). Further observation showed that

MLL fusion Proteins loose the H3K4 methyltransferase activity and are able to

transform hematopoietic cells into leukemia stem cells in an efficient way (50).

The knowledge of key pathways and genes discussed previously resulted from

several fly genes with mutant phenotypes consistent with defects in signaling and/or


11

gene transcription and it has been increasing due to studies of the vertebrate

counterpart of these genes. Furthermore, a rising number of evidences from several

studies suggest that elements of these pathways, found in human cancer, might

reactivate key cell programs in embryonic development since pathways as Notch,

Hedghog and Wingless show atypical signaling without activation due to mutation. Like

previously mentioned Runx and Trithorax are essential in normal hematopoietic

development and are usually associated with translocations relevant in acute

leukemias. All this information and knowledge of these highly conserved pathways is

expected to be used in a near future in order to contribute with therapies and drugs

targeting these specific development pathways to fight against problems like resistance

to standard chemotherapy and the subsequent recurrence caused by cancer stem

cells. However, the effects of these pathways seems to be context dependent, so

without a clear understanding it will be hard to discover effective drugs and therapies.

Lastly, being able to know what pathway is altered in a particular cancer using

biomarkers would be a great improvement for choosing the correct drug administration

(51).

Defects in mitosis could play a role in initiating tumorigenesis and consequently,

drugs that target this process might represent a successful antitumor target. The use of

drugs capable of arresting mitosis has become a widespread treatment for several

human tumors like breast, ovarian non-small-cell lung cancer among others. Tao et al.

in 2005 showed some of the first evidence regarding the link between mitotic arrest, the

mitotic checkpoint and cell death. However, this approach is not ideal due to secondary

effects in other otherwise normal tissues (28).

1.3. MITOSIS AND SPINDLE ASSEMBLY CHECKPOINT

(SAC)

Cell division involves a series of temporally and spatially regulated events that

must be completed in strict order for a successful outcome, the generation of two

daughter cells with normal genome content. One of the major forces behind this orderly

progression is the periodic rise and fall of cyclin-dependent kinases (Cdk) activity. In

order to achieve this, an important mechanism known as the ubiquitin-proteasome

pathway is responsible for the timely degradation of cyclins, which by themselves

regulate Cdks (52). Multiple cell cycle transitions like metaphase-anaphase transition


12

and exit from mitosis are regulated by the anaphase-promoting complex or cyclosome

(APC/C) a multisubunit ubiquitin ligase (52, 53). APC/C relies mainly on two activators

that directly bind to it activating its ubiquitin ligase activity: Cdc20 is the mitotic activator

and Cdh1 executes its function during telophase and G1. The substrates of APC/C,

during mitosis are Securin and the mitotic cyclin (cyclin B). Degradation of securin

allows the activation of separase which then proceeds to cleave one subunit of the

cohesion complexes setting in motion sister-chromatid separation. Cyclin B

degradation also leads to mitotic exit by inhibiting the activity of Cdk1 (54). During S-

phase cells duplicate their chromosomes that are kept together through sister-

chromatid cohesion. During mitosis, the correct attachment of chromosomes to the

mitotic spindle is under the surveillance of the spindle assembly checkpoint ensuring

that anaphase onset does not take place before microtubule attachment and tension

between two opposing kinetochores of all pairs of sister chromatids (54, 55). Previous

experiments showed that even a single unattached kinetochore is enough to arrest cell

in mitosis and removal of this kinetochore was able to undo this metaphase arrest and

allow cells to enter anaphase. These experiments demonstrated the existence of a

conserved mitotic checkpoint (56, 57). These findings suggest that APC/C activation

and the consequence degradation of securing and cyclin B, separase activation and

sister-chromatid separation only occurs when all pairs of sister kinetochores are

correctly attached to spindles microtubules and are under tension (54, 55). Spindle

assembly checkpoint (SAC) proteins inhibit APC/CCdc20 by both stoichiometric

binding to and sequestration of Cdc20 by Mad2 and BubR1, and Cdc20

phosphorylation by Bub1 (58). It is possible to define APC/C as a multisubunit ubiquitin

ligase distantly related to the cullin-RING family of ubiquitin ligases. Ubiquitin-

conjugating enzymes are then activated and recruited by the subcomplexes of APC/C

namely, Apc2 (a cullin-like subunit) and Apc11 (a RING-contaning subunit) (59). Small

peptide motifs, APC/C degrons like the destruction box (D-box with a consensus

RXXLXXXXN), which bind to WD40 domains, and the KEN-box (with a consensus of

KEN), mediate recognition of APC/C substrates. Moreover, another protein called

Apc10 also recognizes the D-box motif (60-62). The inhibition of APC, occurs thorugh a

complex formed by Mad2, BubR1, Bub3, and Cdc20 identified as mitotic checkpoint

complex (MCC) (63, 64). Mad2 was subject of several studies including conformational

and genetic studies that reveal that this protein was in fact a two-state protein which

binding to Cdc20 is closely regulated by the spindle checkpoint and has a vital

importance in maintaining chromosome stability (65, 66). Mad2 appears in two native

folds: a latent open conformer (O-Mad2 or N1-Mad2) and a closed conformer more


13

active in Cdc20 binding and APC/C inhibition (C-Mad2 or N2-Mad2) (67, 68). For

activation, Mad2 binds to Mad1 during the cell cycle maintaining a Mad1-C-Mad2 core

complex that when spindle checkpoint is active it is targeted to unattached

kinetochores. Upon this event, Mad1-C-Mad2 core complex recruits cytosolic O-Mad2

through O–C Mad2 dimerization and this precise dimerization is responsible for a

conformational alteration of O-Mad2 turning into a transient conformation I-Mad2. I-

Mad2 can then dissociate from Mad1-Mad2 core turning into a C-Mad2 without a ligand

and being able to bind to Cdc20. However the mechanics behind APC/C inhibition

through Mad2 binding to Cdc20 are not clear yet (69). Another key player in the spindle

assembly checkpoint is BubR1 and this protein seems to induce a higher inhibition of

APC/CCdc20 than Mad2 in vitro (26, 70). Analysis of BubR1 structure revealed a

domain required for kinetochore targeting named N-terminal tetratricopeptide repeat

(TPR), a conserved Bub3 binding motif GLEBS and also a C-terminal kinase domain.

However, the kinase domain of BubR1 is not conserved in yeast and is not necessary

for APC/C inhibition (71). Studies in yeast showed that Mad3 (the yeast homologue of

BubR1) contains KEN-boxes and a D-box motif responsible for Mad3 binding to Cdc20,

competing with APC/C for its subtrates, resulting in the consequent checkpoint

function, highly conserved mechanism higher organisms like human, mouse and flies

(71, 72). BubR1/Mad3 seem to act as a competitive, pseudo-substrate inhibitor of

APC/CCdc20 as it does not suffer APC/CCdc20-dependent degradation. The reason

behind this might be the acetylation of BubR1 at K250 (73). What is clear is that BubR1

and Mad2 are able to inhibit APC/CCdc20 in vitro independently, however, most likely

they collaborate by forming MCC. Studies with unattached kinetochores showed an

enhanced inhibition of APC/C through purified MCC components (Mad2, BubR1, and

Bub3) and an enhanced binding of BubR1 to APC/CCdc20. This enhancement

depends on kinetochore-bound Mad1-Mad2 core complex and Mad2 dimerization (74).

The role of unattached kinetochores in the formation of MCC seems to be the

production of C-Mad2 and generate an initial Mad2–Cdc20 complex that allows further

interaction with BubR1–Bub3 (74). The binding of BubR1–Bub3 and Mad2 to Cdc20

changes how Cdc20 and APC/C interact. Therefore, it appears that the MCC inhibits

the activity of APC/CCdc20 by impeding substrate binding both competitively and non-

competitively (75). Inhibition of APC/C can also be performed by Cdc20

phosphorylation. However, it is unclear how this occurs (76). Through the course of

evolution, most checkpoint proteins remained conserved, however the knowledge

about the role of checkpoint proteins and their importance is still increasing, including

for Bub3. It is possible to find Bub3 in most eukaryotes (77-79). This protein contains


14

seven WD40 repeats, which are found in several distinct proteins and it is believed to

be involved in protein-protein interactions (80). Throughout mitosis, Bub3 is visible on

kinetochores during prophase, decreasing its levels by metaphase. If kinetochores

remain unattached the kinetochore-associated Bub3 antigen increases (79). Studies in

mice showed that one functional copy of the Bub3 gene is enough for normal

development since heterozygous Bub3 mice do not show significant abnormalities in

development or fertility (81). However, upon analyzing the progeny of their

heterozygous crosses they found no Bub3−/− pups, suggesting an embryonic lethal

phenotype. Therefore, Bub3 seems to be essential in mouse for normal mitosis and

early development (81). Likewise, depletion of Bub3 in Drosophila also leads to

embryonic lethality at the larval/pupal transition stage and cells from these mutants

larvae fail to arrest in mitosis in response to microtubule-depolymerizing drugs and

show severe mitotic abnormalities like premature chromatid separation, lagging

chromosomes and chromatin bridging (82, 83). It was shown that Bub3 forms two

independent complexes with Bub1 and BubR1 in mitotic mammalian and Xenopus cells

and seems to be required for the localization these proteins to the kinetochore. Yeast

has also provided some evidence of a complex with Mad1 and Bub1. Moreover, this

complex is dependent on Mad2 and seems essential for a checkpoint response (84).

Research in Drosophila S2 Bub3 depleted cells revealed an accelerated mitotic timing

when compared to normal cells (from NEBD to anaphase onset). It was also suggested

that there is a slower progression through early stages of mitosis specifically a delay in

prophase (84). Recently Niikura and co-workers linked Bub3 with cell death defined as

caspase independent mitotic death (CIMD). They demonstrated that when BUB3 is

dissociated from BUB1 it activates CIMD. BUB1 interacts with p73, a structural and

functional homolog of the p53 tumorsuppressor protein, in cells showing CIMD. It

seems that Y99 phosphorylation on p73 by c-Abl tyrosine kinase and S19

phosphorylation on BUB3 by BUB1 kinase are required for CIMD activation. These

observations point to the hypothesis that CIMD protects cells from aneuploidy, which

eventually leads to tumorigenesis (85).


15

Figure I.1 - Schematic representation of the stages of mitosis. Adapted from Weaver and Cleveland, 2005 (28).


16

1.4. APOPTOSIS

The ability to self-destruct by undergoing apoptosis is present in most animal

cells (86). Apoptosis is a genetically encoded process with characteristic morphological

features that serve to dispose of superfluous or unwanted cells and when this process

fails, it can be associated with many human diseases (87). In order for apoptosis to

occur, it is necessary to activate caspases, a family of cystein proteases that are

present in cells as weakly active zymogens (88). Caspases were initially identified in a

screen for C. elegans mutants unable to undergo developmental programmed cell

death (89). Different stimuli may be responsible for caspase activation such as

developmental signals as well as various forms of persistent cellular stress or injury

(86). Amongst some of these are DNA damage, viral infection, hypoxia, increased

presence of reactive oxygen species, loss of cellular adhesion, accumulation of

unfolded proteins, excitotoxicity, cytoskeletal damage as well as other insults (86).

Healthy cells contain caspase zymogens as inactive precursors (88). Caspase

zymogens possess a small amount of catalytic activity, however they are kept in check

by several regulatory molecules and only after receiving an apoptotic signal they are

able to perform its full activity upon undergoing proteolytic processing and generate two

subunits (9, 90). The cleavage of the zymogen is not always required for caspase

activation. Structural studies predict that the mature caspase is a heterotetramer

formed by two heterodimers originating from two precursor molecules (9, 90).

Moreover, beside the regions that give rise to the two subunits, procaspases include N-

terminal prodomains of different sizes and when analyzing these lengths one can

divide caspases into two groups: those with relatively long prodomain and those with a

shorter prodomain (9, 90). Several specific protein-protein interaction domains of

relevant importance in caspase activation are possible due to the existence of long

prodomains (9, 90). Procaspases are then recruited by these domains to specific death

signaling complexes which in turn lead to their autocatalytic activation by a mechanism

often designated “proximity-induced” activation (9, 90). The caspases recruited to

signaling complexes, therefore being activated, are known as the initiator caspases

due to providing a link between cell signaling and apoptotic execution (88). The main

initiator caspases present in mammals are caspase-2, -8, -9 and -10 and in Drosophila

is the Nedd2-like caspase (DRONC). Caspases -2, -9 and DRONC contain a caspase

recruitment domain (CARD) and caspase -8 and -10 contain a pair of death effector

domains (DEDs) (88). All these domains are able to bind to similar motifs in adaptor

proteins, however caspases lacking a long prodomain are not able to self-activate and


17

need to be cleavage by previously activated initiator caspases (9, 90). After being

activated these caspases cleave most of the cellular substrates upon apoptosis,

therefore being designated effector caspases. Some of the most important effector

caspases are caspases-3, -6 and 7 in mammals and drICE in Drosophila (88).

Activation of this caspases must be kept in check. A class of anti-apoptotic proteins

called inhibitor of apoptosis proteins IAPs accomplishes this task (91, 92). IAPs are

able to bind to both types of caspases through their BIR domains (93, 94) and they also

contain a RING motif that enable them to act as E3 ubiquitin ligases promoting

ubiquitination of caspases (95, 96). In Drosophila in order to prevent caspase activation

and apoptosis in virtually all somatic cells one key IAP must be present, DIAP1 (97-99).

DIAP1 exists in an auto-inhibited conformation being unable to inhibit the effector

caspase drICE (100). In order to disabled the auto-inhibition a caspase-mediated

cleavage of DIAP must occur (100). A cleaved DIAP is then able to bind to mature

drICE resulting on its inhibition and apparently targeting drICE for ubiquitylation (100).

If a cell is doomed to die, specific antagonists inactivate IAPs promoting apoptosis (98,

101, 102). In Drosophila it is possible to find three IAP antagonists clustered together in

the genome, Reaper (Rpr), Head involution defective (Hid) and Grim, and proceeding

to deletion of these genes gives rise to severe apoptosis inhibition (103-105). Another

IAP antagonist, Sickle (Skl), similar to Rpr was also found, however, since there is no

mutants and a thorough analysis of its role in apoptosis is still not possible (106). IAP

antagonists bind to IAPs through an evolutionary conserved feature, the N-terminal

IBM (IAP-binding motif) (94, 107), resulting in the displacement of IAP-bound caspases

(108-110). The active caspases generate a proteolytic cascade that proceeds to the

destruction of the cell. Moreover, IAP antagonists also stimulate IAP turnover by

proteasomal degradation (111, 112). Drosophila IAP antagonists expressed in human

cells also showed cell death inducing activity (113, 114). It is also important to mention

that IAP antagonists, Rpr, Hid and Grim may be induced by a several upstream factors

such as the p35 gene (115, 116) or the Jun N-terminal Kinase (JNK) pathway (117).

Apparently, these factors seem to direct developmental or environmental stimuli like

irradiation or cell competition to apoptosis (118-120). The current image of apoptosis in

Drosophila defines DIAP1 as a central brake on apoptosis performing the inhibition of

activated caspases. DIAP is able to bind to caspases resulting in its sequestration

impeding the caspase binding to its target or even in caspase degradation. Apoptosis

occurs when IBM proteins Rpr, Grim, Hid and Skl compete with caspases for binding to

DIAP1. Additionally Rpr and Grim are also able to suppress DIAP1 translation (108,

112). Several other IAP-like proteins in flies, DIAP2, deterin and dBRUCE seem to be


18

able to suppress apoptosis although their specific mechanisms are not completely

elucidated yet (121-124). Moreover, developmental and environmental inputs may

influence the apoptotic pathway either by changing the transcription and activity of IBM

proteins, or synthesis of some caspases and even the synthesis of some IAPs (103,

125-127).

Figure I.2 - Schematic representation of the understanding of apoptosis in Drosophila. Adapted from Kornbluth,

2005 (128).

There are two types of distinguishable apoptosis, a developmental regulated

(129) one and a stress induced one (130). The developmental regulated apoptosis is

important in processes like the cleft formed between the maxillary and mandibular

segments in early Drosophila embryos, which is possible through the activation of

Reaper by Deformed, a Hox protein (131). The elimination of the excess of

interomatidial cells during metamorphosis is also a good example of the developmental

apoptosis importance (132). In addition, another morphogenic role of apoptosis was

recently found, the formation of leg joints, resulting from a induced discontinuity of Dpp

signaling levels in the presumptive region providing with the required activation of

Reaper (133). The other type of apoptosis, the stress induced apoptosis is not

programmed however it has become a tool to dispose of cells damaged by insults like

irradiation, heat shocks, hypoxia among other examples (105, 134, 135). In Drosophila,

the wing imaginal disc is a suitable system to study stress-induced apoptosis since

developmental apoptosis is almost none existent therefore the majority of observed

apoptosis is caused by stress treatments (105, 134, 135). To induce cell death in


19

imaginal tissue there are several treatments available: irradiation of larvae with

moderate doses of X-Ray (eliminating a number close to 40-60% of wing imaginal cells

(136); heat-shock at 37ºC for 3 hours is also a mean to achieve high apoptotic levels

(135). Which genes are involved in the apoptotic stress response? The fly homolog of

the tumor suppressor gene p53, Dmp53 activates the apoptotic response after

irradiation and mutants or dominant negatives for this protein produce very little

apoptosis response to X-Rays (4, 116, 118, 120). It was shown that IAP antagonist like

Rpr, Hid and Sickle are transcriptional targets of Dmp53 and that X-ray induced

apoptosis in wing disc is highly dependent on Hid (115, 116). Upon γ-Ray irradiation,

the activity of the JNK pathway is altered giving rise to an up-regulation of the JNK

target gene puckered (puc, a phosphatase that negatively regulates the pathway

activity) present in the entire wing disc (117). Recently, the Hippo pathway and

apoptosis showed a connection (137) because γ-Rays are able to induce Hippo Kinase

activity in a Dmp53-dependent manner in cell culture and imaginal discs. Furthermore,

DIAP1 is also a target of the Hippo pathway (138). An important characteristic of the

stress-induced apoptosis is the non-lethality of some treatments even after the

production of large amounts of cell death. Even after more than 50% of the cells in

wings imaginal discs enter apoptosis, the flies recover and are able to restore organs of

normal shape and size (135, 136). Therefore, an additional proliferation by surviving

cells must compensate and originate a phenomenon called compensatory proliferation

(136, 139). Interestingly, Drosophila and its genetic tools allow us to interfere with

apoptosis at quite different stages of its progression: hinder the function of upstream

activators therefore meddling at the activation of the apoptotic program; the activation

of caspases; the function of effector caspases (140).

1) In order to prevent upstream activation it is possible to use p53 null mutants

or dominant negative forms of the gene (4, 116, 118, 120). Inhibition of JNK activity or

overexpression of puc are also effective (141, 142).

2) When the objective is to prevent caspase activation a deficiency in Rpr, Hid

and Grim functions can be used with the H99 deletion which impede DIAP1 inactivation

suppressing apoptosis (103). Another way to achieve prevention in caspase activation

is through the overexpression of DIAP1 using UAS-DIAP1 transgene, although it does

not inhibit apoptosis completely (111). At last, mutants or dominant negative forms of

the initiator caspase DRONC, do not allow the activation of the effector caspase drICE

(143).

3) In order to prevent effector caspase function, the usual method is to utilize

the baculovirus protein P35 (144) which blocks the function of drICE (unaffecting


20

DRONC) allowing however, its activation. It seems like P35 binds to drICE inhibiting its

protease activity (145).

Figure I.3 - A schematic view of Drosophila apoptotic machinery. Adapted from Martín, 2009 (140).

In order to prevent upstream activation it is possible to use p53 null mutants or

dominant negative forms of the gene (4, 116, 118, 120). Inhibition of JNK activity or

overexpression of puc are also effective (141, 142). When the objective is to prevent

caspase activation a deficiency in Rpr, Hid and Grim functions can be used with the

H99 deletion which impede DIAP1 inactivation suppressing apoptosis (103). Another

way to achieve prevention in caspase activation is through the overexpression of

DIAP1 using UAS-DIAP1 transgene, although it does not inhibit apoptosis completely

(111). At last, mutants or dominant negative forms of the initiator caspase DRONC, do

not allow the activation of the effector caspase drICE (143). Finally, in order to prevent

effector caspase function, the usual method is to utilize the baculovirus protein P35

(144) which blocks the function of drICE (not affecting DRONC) allowing however, its

activation. It seems that P35 binds to drICE inhibiting its protease activity (145).

The effects of arresting apoptosis at different stages were observed and

compared (140). Either using p53 mutants or overexpressing its the dominant negative,

therefore inhibiting apoptosis, does not seem to have a striking effect on the

development or morphology of the flies, however, there may be uncharacterized

consequences on longevity or fertility. An analogous outcome is observed upon


21

overexpressing DIAP1, impeding caspase activation. On the other hand, hinder the

activity of the effector caspase drICE activity using P35 produces great developmental

consequences (135). Without radiation, P35 by itself does not cause major alterations.

However if apoptosis is induced in cells expressing P35 using radiation, a group of

“undead cells” arise. These cells have their apoptosis program active but due to the

blockage of effector caspases they cannot die. These cells also keep their original

identity as they differentiate into wing structures (135) but retain all the molecular

markers of apoptosis such as low levels of DIAP1 and high levels of Hid, DRONC or

drICE proteins and even behave as normal apoptotic cells moving to the basal

membrane of epithelia (140). Genes expressed by undead cells were identified,

namely, decapentaplegic (dpp) and wingless (wg) which are defined by being major

players in patterning and growth in the wing disc (146). This feature allied with the

persistence of undead cells is probably responsible for an abnormal growth of the P35

containing compartments although additional proof is necessary (140).

1.5. MAIN TECHNOLOGIES

1.5.1. MICROARRAYS

Nowadays, aided by several “whole genome sequencing” projects, the

availability of genetic information is outstanding. RNA expression in cell and tissue at a

given moment can now be studied using high-throughput techniques due to the

emergence of microarray-based technologies in the transcriptomic field (147).

Microarray technology is becoming increasingly important for molecular biology mainly

because it analyzes the overall transcription profile of a cell or tissue instead of being

limited to “one gene at a time” (148). Microarray techonology was introduced 17 years

ago in a seminal paper by Schena et al. (149). This technology grants quantitative

monitoring of the expression profile of thousand of genes (150) and relies on base pair

complementary for identification of mRNA highly expressed in a sample of interest

(151). Microarray assays can be divided in five distinct experimental steps: biological

query, sample preparation, biochemical reaction, detection, data visualization and

modeling (152). Upon concluding a microarray assay, a researcher ends up with a

snapshot of genes expressed in a cell or tissue at the time of harvest and RNA

purification, which is a way to compare samples horizontally and obtain an expression


22

profile. For instance, in biomedicine the value of expression patterns is immense

yielding valuable insight on biological processes in diseased tissue, disease prediction,

diagnosis and treatment, cellular differentiation, development and drug discovery (153).

Currently DNA microarrays offer a way of exploring the genome in a systematic and

understandable way and they present themselves as oligonucleotide arrays and a

variety of cDNA arrays (154). Oligonucleotide arrays contain short fragments of DNA

spotted or synthesized in situ on solid supports (glass, coated glass, silicon or plastic)

in a way that knowing the location defines which oligonucleotide is present there (155).

They differ from cDNA, firstly because their probes are a set of 20 to 25 short

oligonucleotides, specific for each gene or exon, and are synthesized in situ by

photolithographic deposition. Secondly, the arrays are hybridized to a single

biotinylated amplified RNA sample, and an algorithm computes the result (156). It is

also important to refer to the general superiority of oligonucleotide-based microarrays

in terms of quality and reproducibility although there is a need for a larger capital input

as well as a larger amount of starting biological material (157, 158). cDNA arrays, on

the other hand, contain long fragments of DNA, reaching thousands of base pairs, and

are created using robotically spotting of individual samples of purified cDNA clones in a

solid support like glass or a membrane (159). Some of the most general applications of

cDNA arrays are comparative and functional genome analysis. Recently it was used in

drug resistance tests (159).

Following a microarray experiment, the amount information generated is

enormous and to be able to find the relevant information is still a quite challenge.

Therefore, bioinformatics methods are essential in order to extract pertinent biological

information (160). Over the last few years, these bioinformatic tools have indeed

contributed to several studies and large gene lists analysis from various high

throughput techniques. However, there is no rule or gold standard as far as these tools

go and there is still significant room for improvement (160). On the other hand as these

bioinformatics techniques evolve so does the difficulty to search and choose the right

way to perform, for instance, a microarray analysis (160). The complexity achieved by

the overabundance of tools led to several problems including which tools are most

fitting for a particular type of analysis (160). Because of the complexity of biological

data mining situations, we can define the analysis of large gene lists with the latest

bioinformatics techniques as a process of exploratory data-mining instead of a solution

based on statistics. Clearly, the current solutions depend on the aid of human

knowledge and thought, and this, supported with computing algorithms, integrated


23

annotations databases, and P-values derived from statistical methods is what allows

for a comprehensive and as best as possible analysis (160).

1.5.2 RNA interference

RNA interference (RNAi) was first seen by researchers trying to alter flower

colors in petunias (161). However, only when work on RNAi in Caenorhabditis elegans

was published (162) the mechanism became better understood. Later it was

demonstrated that RNAi works efficiently in many cell types including Drosophila and

can be used for large-scale gene silencing applications (163-165). Therefore, large-

scale RNAi-based screening is now a important tool for the identification of gene

products involved in different biological processes and pathways (166, 167). Research

centers like the Vienna Drosophila RNAi Center (VDRC) maintain and distribute to the

research community transgenic RNAi stocks constructed by Dickson and Keleman

groups (168). RNAi mechanism was unraveled through a series of genetic and

biological studies. The model starts with the conversion of the dsRNA into small RNAs

(siRNAs) by an RNase III family nuclease named Dicer (169). Then these small RNAs

join a multicomponent nuclease complex, and become the tool to identify homologous

mRNAs (170). Two of the major keys

players in this system are universally

conserved namely, Dicer and

Argonaute (Ago) gene family

members. Structure wise, dicer

shows a tandem repeat of RNase III

catalytic domains, a carboxylterminal

deRNA binding domain, an animo-

terminal DExH/DEAH RNA helicase

domain, and a PAZ domain (169,

171). Ago proteins, that take part of

the RNA-induced silencing complex

(RISC) include a PAZ domain and a

carboxyl-terminal PIWI domain

(172).

Figure I.4 - Schematic representation of a model of RNA interference in

mammalian cells. Adapted from Paddison, 2002 (172).


24

1.6 AIMS

In his project we established a few goals:

Use microarray technologies to determine the transcription profile of cells from

the wing disc of larvae expression an RNAi construct against Bub3 that at the

same time were prevented from undergoing apoptosis and showed clear sings

of transformation.

Apply a bioinformatic approach to identify possible genes or gene networks

associated with the transformation phenotype.

Perform a RNAi screen to validate candidate genes chosen through the

previous bioinformatic approaches

Eventually determine which which are the genes responsible for Bub3-related

tumorigenesis.


25

CHAPTER 2

MATERIALS AND METHODS


26

2.1. FLY STOCKS USED

Below are The Drosophila stocks used during this thesis:

Ap-Gal4>UAS-GFP/CyO;

Ap-Gal4>UAS-P35/CyO;

Ap-Gal4>UAS-GFP UAS-Bub3RNAi/CyO;

Ap-Gal4>UAS-GFP UAS-Bub3RNAi/UAS-P35;

Ap-Gal4>UAS-GFP UAS-Bub3RNAi UAS-DIAP1/CyO.

The RNA interference constructs used are mentioned in annex 1 and are

described in Flybase (http:/flybase.bio.indiana.edu/). Standard procedures were used

for the flies management.

2.2. RNA EXTRACTION FROM WING IMAGINAL DISCS

2.2.1. COLLECTION OF WING IMAGINAL DISCS

To obtain RNA we dissected in Phosphate Buffer Saline (PBS) 70 larvae (140

wing imaginal discs) of each genotype:

Ap-Gal4>UAS-GFP/CyO;

Ap-Gal4>UAS-P35/CyO;

Ap-Gal4>UAS-GFP UAS-Bub3RNAi/CyO;

Ap-Gal4>UAS-GFP UAS-BubRNAi /UAS-P35.

After dissection, the larvae were conserved frozen at -80ºC.

2.2.2. RNA EXTRACTION

Dissected wing discs were incubated in 500 µl TRIzol (ref. 15596-026,

Invitrogen) for 5 minutes (min) at room temperature (RT). Solution was cleared by

centrifugation at 12000g for 10 min at 4ºC.

The supernatant was transferred to a 2 ml microfuge tube and 100 µl of

chloroform (ref. 1.02445.2500, Merck) were added and vortexed for 1 min.


27

Samples were incubated 5 min at RT, centrifuged for 15 min at 4ºC at max

speed (13,2 rpm) and 250 µl of the aqueous phase were transferred to a 2 ml tube.

Afterwards we used RNeasy Mini Kit (ref. 74104, Qiagen) to proceed with the

extraction. We added 875 µl of RLT buffer and briefly vortex the mixture followed by the

addition of 625 µl absolute Etanol (ref. 1.00983.2511, Merck) and vortex again for a

few seconds.

The mix was loaded onto a RNeasy column and centrifuged for 30 seconds at

RT, max speed (13,2 rpm) and the flow through was discarded. This last procedure

was repeated two more times.

The columns were washed with 500 µl RPE buffer and centrifuged for 30

seconds at RT, max speed (13,2 rpm), and the flow through was discarded. This last

procedure was done three times. The column was spin dry for 2 min and placed in a

1ml RNAse free tube followed by the addition of 10 µl of water to the center of the

column.

After 2 min at RT the colum was centrifuged at max speed (13,2 rpm).


28

2.3. MICROARRAYS

Figure II.1 - Required equipment for microarray protocol. Adapted from One-Color Microarray-Based

Gene Expression Analysis Low Input Quick Amp Labeling, Agilent Technologies.


29

Figure II.2 - Required reagents for microarray protocol. Adapted from One-Color Microarray-Based

Gene Expression Analysis Low Input Quick Amp Labeling, Agilent Technologies.

Started by performing the sample preparation

Step 1. Preparation of the One-Color Spike Mix

- Equilibrated water baths to 37°C, 65°C, 80°C, 40°C and 70°C.

- Vigorously mixed the One-Color Spike Mix stock solution on a vortex mixer.

- Heated at 37°C for 5 minutes, and mixed on a vortex mixer once more.

- Briefly span in a centrifuge to drive contents to the bottom of the tube prior

to opening.

Step 2. Preparation of the labeling reaction (Time required: approximately 5,5 hours)

- Added 10 to 200 ng of total RNA to a 1,5-mL microcentrifuge tube in a final

volume of 1,5 µL.

- Added 2 µL of diluted Spike Mix to each tube.

- We prepared and added T7 Promoter Primer:

o Mixed the T7 Promoter Primer and water to prepare the T7 Promoter

Primer Master Mix

o Added 1.8 µL of T7 Promoter Primer Mix to the tube that contains 3.5 µL

of total RNA and diluted RNA spike-in controls.

o Denatured the primer and the template by incubating the reaction at

65°C in a circulating water bath for 10 minutes.

o Placed the reactions on ice and incubate for 5 minutes

- Prewarmed the 5X first strand buffer at 80°C for 3 to 4 minutes to ensure

adequate resuspensions of the buffer components.

- Prepared and add cDNA Master Mix:


30

o Immediately prior to use, we added the components for cDNA Master

Mix (description in original protocol)and kept it at room temperature.

o We then Briefly spin each sample tube in a microcentrifuge to drive

down the contents from the tube walls and the lid.

o Added 4.7 µL of cDNA Master Mix to each sample tube and mix by

pipetting up and down.

o Incubated samples at 40°C in a circulating water bath for 2 hours.

o Moved samples to a 70°C circulating water bath and incubate for 15

minutes.

o Moved samples to ice. Incubated for 5 minutes.

o Span the samples briefly in a microcentrifuge to drive down tube

contents from the tube walls and lid.

- Prepared and added Transcription Master Mix

Step 3. Purification of the labeled/amplified RNA (Time required: approximately 0,5

hours)

- Added 84 µL of nuclease-free water to your cRNA sample, for a total volume of

100 µL.

- Added 350 µL of Buffer RLT and mixed well by pipetting.

- Added 250 µL of ethanol (96% to 100% purity) and mixed thoroughly by

- pipetting.

- Transfered the 700 µL of the cRNA sample to an RNeasy mini column in a 2 mL

collection tube. Centrifuged the sample at 4°C for 30 seconds at 13,000 rpm.

Discarded the flow-through and collection tube.

- Transferred the RNeasy column to a new collection tube and add 500 µL of

buffer RPE (containing ethanol) to the column. Centrifuged the sample at 4°C

for 30 seconds at 13,000 rpm. Discarded the flow-through. Re-use the

collection tube.

- Added another 500 µL of buffer RPE to the column. Centrifuged the sample at

4°C for 60 seconds at 13,000 rpm. Discarded the flow-through and the

collection tube.

- Eluted the cleaned cRNA sample by transferring the RNeasy column to a new

1.5 mL collection tube. Added 30 µL RNase-free water directly onto the RNeasy

filter membrane. Waited 60 seconds, then centrifuged at 4°C for 30 seconds at

13,000 rpm.

- Maintained the cRNA sample-containing flow-through on ice. Discarded the

RNeasy column.


31

Hybridization

Step 1. Prepared the 10X Blocking Agent

- Added 500 µL of nuclease-free water to the vial containing lyophilized 10X

Blocking Agent supplied with the Agilent Gene Expression Hybridization Kit

- Mixed by gently vortexing.

- Drived down any material adhering to the tube walls or cap by centrifuging for 5

to 10 seconds.

Step 2. Prepared hybridization samples

- Equilibrated water bath to 60°C.

- For each microarray, added each of the components as indicated in the original

protocol

- Mixed well but gently on a vortex mixer.

- Incubated at 60°C for exactly 30 minutes to fragment RNA.

- Immediately cooled on ice for one minute.

- Add 2x GEx Hybridization Buffer HI-RPM to the 1-pack, 2-pack, 4-pack, and 8-

pack microarray formats at the appropriate volume to stop the fragmentation

reaction. Information detailed in the original protocol

- Mixed well by careful pipetting.

- Span for 1 minute at room temperature at 13,000 rpm in a microcentrifuge to

drive the sample off the walls and lid and to aid in bubble reduction.

- Placed sample on ice and loaded onto the array as soon as possible.

Microarray Wash

Step 1. Proceeded to add Triton X-102 to Gene Expression wash buffers

- Opened the cardboard box with the cubitainer of wash buffer and carefully

removed the outer and inner caps from the cubitainer.

- Used a pipette to add 2 mL of the provided 10% Triton X-102 into the wash

buffer in the cubitainer.

- Replaced the original inner and outer caps and mixed the buffer carefully but

thoroughly by inverting the container 5 to 6 times.

- Carefully removed the outer and inner caps and install the spigot provided

with the wash buffer.

Step 2. Prewarmed Gene Expression Wash Buffer 2 to 37°C

- Dispensed 1000 mL of Gene Expression Wash Buffer 2 directly into a sterile

1000-mL bottle.


32

- Capped the 1000-mL bottle and place in a 37°C water bath the night before

washing arrays.

Step 3. Prepared the equipment

- Added the slide rack and stir bar to the staining dish.

- Transfered the staining dish with the slide rack and stir bar to a magnetic stir

plate.

- Filled the staining dish with 100% acetonitrile.

- Turned on the magnetic stir plate and adjusted the speed to a setting of 4

(medium speed).

- Washed for 5 minutes.

- Discarded the acetonitrile

- Repeated from step 3

- Air-dried the staining dish in the vented fume hood.

- Proceeded to “Milli-Q water wash”

- Washed all dishes, racks, and stir bars with Milli-Q water.

- Ran copious amounts of Milli-Q water through the staining dish.

- Emptied out the water collected in the dish.

- Discarded the Milli-Q water.

Step 4. Washed the microarray slides with GE Wash Buffer 1 and 2

Finally, Scanned the microarray slides and extract the data.


33

CHAPTER 3

RESULTS


34

3.1. INTRODUCTION

Previous work in the laboratory addressed the role of checkpoint genes in

tumorigenesis. As homozygous mutants for Bub3 die during the third instar larval stage

(84), in vivo RNAi transgenic technology was used to knockdown Bub3 in the dorsal

region of the wing imaginal disc (173). Upon knockdown, the adult fly thoraxes show

loss of bristles and empty sockets when compared with controls. This phenotype was

the result from increased cellular death in the wing imaginal disc and it is in accordance

with reports for an increase apoptosis when SAC genes are removed (81, 84).

However, if the apoptosis inhibitor DIAP1 (174) is expressed where Bub3 is knockdown

the adult thorax becomes larger and deformed with missing bristles and empty sockets

and the wing disc shows a higher number of cells. However, if in the tissue where Bub3

is knockdown the stronger apoptotic inhibitor p35 is expressed, the larval wing imaginal

disc becomes highly hyperplastic (175). We used microarray and bioinformatics

approach to identify that were specifically misregulated in these hyperplastic cells.

Then we followed with an in vivo screen with RNAi technology of candidate genes.


35

Figure III.1 - images from the tumorigenesis

model adopted. a) Wing imaginal disc illustration

showing the division of the anterior and posterior

portions with a blue line (A - anterior portion, P -

posterior portion), the division of the ventral and

dorsal side through a red line (D - dorsal portion,

V - ventral portion). The green color represents

the portion that will generate the wing pouch, the

yellow color represents the portion that will

generate the hinge and the blue color represents

the portion that will generate the body wall. b)

Shows a larvae expressing GFP through a GAL4-

UAS system in its wing imaginal discs. c) Shows

a larvae expressing GFP, Bub3RNAi

and p35

through a GAL4-UAS system in its wing imaginal

discs. In both b)and c) the GFP is expressed in

the dorsal compartment of the wing disc under

the control of the apterous promoter (REF see

above). The size of this wing imaginal disc is

considerably bigger than wild type.

3.2. SELECTED

GENOTYPES

To determine the molecular basis for the

link between SAC genes and tumour formation

in the absence of apoptosis we used four

cellular genotypes for the microarray analysis.

The regulatory region of the apterous gene was

used to drive Gal4 expression in the dorsal

compartment of the wing disc (figure III.1) (176).

The non-expression in the ventral portion allows

for an internal control with a pool of cells not

expressing the RNAi of the candidate genes.

The first transgenic fly has Bub3 knocked down

through RNAi specifically on the dorsal

compartment of the wing disc and expresses

simultaneously GFP (Ap-Gal4>UASGFP

UASBub3RNAi/ CyO). The wing imaginal disc

shows high levels of apoptosis and is

significantly smaller than wild type (Figure

III.2.c). The adult fly presents loss of bristles

and empty sockets with altered wings that often

are necrotic (Figure III.2.d). In the second

transgenic fly, the apoptosis presented in the

previous genotype was blocked. We used two

different inhibitors, the previously mentioned

DIAP1 and P35. We crossed knockdown flies

for Bub3, expressing GFP with flies transgenic

for a UASDIAP1 construct and obtained a

recombinant fly expressing Bub3RNAi, GFP and

DIAP1 in the dorsal compartment of the wing

disc (Ap-Gal4>UASGFP UASBub3RNAi

UASDIAP1/ CyO). Wing imaginal discs from

this genotype are bigger with a higher number

of cells (Figure III.2.e). The adult fly presents a


36

larger, and generally deformed thorax, with some loss of bristles, empty sockets and

deformed wings showing necrosis (Figure III.2.f). To inhibit apoptosis with p35, flies

with the UASP35 construct (Ap-Gal4>UASP35/ CyO) were crossed with flies

expressing Bub3RNAi, and GFP in the dorsal compartment, and larvae for Bub3

expressing GFP and P35 (Ap-Gal4>UASGFP UASBub3RNAi/ UASP35). The wing

imaginal discs show overgrowth, loosing most of its structure with a significant

increased size (Figure III.2.g). This particular genotype is lethal as there are no viable

adult. As controls, we used flies expressing only GFP (Ap-Gal4>UASGFP/ CyO) as

their wing imaginal disc remains unaltered (Figure III.2.a). The adult fly, shows no sign

of abnormal phenotype revealing a full set of bristles, a normal thorax and wings

(Figure III.2.b). Finally, we used flies expressing only P35 (Ap-Gal4>UASP35/ CyO).

These flies are a control for the effects of inhibiting apoptosis. They show a normal

thorax although it has more bristles and some presented one black dot in each side on

the boundary between notum/scutellum (Figure III.2.h).

Figure III.2 - Selected genotypes. (a) Wing imaginal disc of Drosophila melanogaster expressing GFP through a

GAL4-UAS System in its dorsal region portion drived by the apterous promoter showing a wild type structure. (b) An

adult Drosophila melanogaster expressing GFP especially in the thorax and wings through a GAL4-UAS System driven

by the apterous promoter. It shows an essentially a wild type thorax. (c) Wing imaginal disc of Drosophila melanogaster

expressing GFP and Bub3RNAi

through a GAL4-UAS system in its dorsal region driven by the apterous promoter, (d) An

adult Drosophila melanogaster expressing GFP and Bub3RNAi

through a GAL4-UAS system driven by the apterous

promoter especially in the thorax and dorsal region of wings. It shows a thorax with empty sockets, missing bristles, and

deformed wings. (e) Wing imaginal disc of Drosophila melanogaster expressing GFP, Bub3RNAi

and DIAP1 through a

GAL4-UAS system in its dorsal region driven by the apterous promoter (f) An adult Drosophila melanogaster expressing

GFP, Bub3RNAi

and DIAP1 in the thorax and wings showing a larger and generally deformed thorax with empty sockets,

and deformed necrotic wings. (g) Wing imaginal disc of a Drosophila melanogaster larvae expressing GFP, Bub3RNAi

and p35 through a GAL4-UAS system in its dorsal region driven by the apterous promoter showing a large overgrowth.

(h) An adult Drosophila melanogaster expressing P35 through a GAL4-UAS system driven by the apterous promoter in

the thorax and wings. It shows a thorax resembling a wild type one although seeming hairier and sometimes showing a

black dot in each side on the boundary notum/scutellum.


37

3.3. RNA SAMPLES FOR MICROARRAYS

3.3.1. QUALITATIVE AND QUANTITATIVE ANALYSIS OF

THE EXTRACTED RNA

3.3.1.1. NANODROP ANALYSIS

In order to perform a microarray experiment we extracted RNA samples from

wing imaginal discs of several genotypes described above. Three separate

experiments were carried out for ach genotype. After extraction, we resorted to a

Nanodrop analysis to assess the quality of our samples for microarray analysis. All

samples provided an RNA concentration higher than 50ng/ul. The first three samples,

A, B and C (Table III.1) were very close in terms of concentration, however sample D

shows higher values due to the increased size, and consequent higher cell number, of

the wing imaginal discs in this genotype. The purity of the RNA shows was assessed

by Nanodrop showing an average ratio of 2,05 (260/280 nm). Further results using a

bioanalyzer helped to assure us of RNA quality despite the lower values of 260/230

ratio.

Table III.1 - Nanodrop quantifications in ng/ul of four different samples with three repetitions each, Absorbance at 260

nm, 280 nm, 230 nm, the ratio between 260 and 280 and the ratio between 260 and 230.

Genotype Sample

ID ng/ul A260 A280 260/280 260/230

Ap-Gal4>UAS-GFP/CyO A1 51,92 1,298 0,766 1,7 0,14



Ap-Gal4>UAS-P35/CyO B1 76,26 1,906 1,017 1,87 0,34

Ap-Gal4>UAS-P35/CyO B2 144,7 3,618 1,721 2,1 1,44

Ap-Gal4>UAS-P35/CyO B3 158,63 3,966 1,864 2,13 0,77

Ap-Gal4>UAS-GFP UAS-Bub3RNAi

/CyO C1 50,83 1,271 0,66 1,93 0,16


/CyO C2 125,2 3,13 1,492 2,1 2,06


/CyO C3 59,31 1,483 0,787 1,88 0,29


/UAS-P35 D1 310,31 7,758 3,533 2,2 2,21


/UAS-P35 D2 749,58 18,74 8,373 2,24 2,25


/UAS-P35 D3 894,69 22,367 9,551 2,34 2,97


38

3.3.1.2. BIOANALYZER ANALYSIS

The isolation of RNA from biological samples is a difficult process, hampered by

the action of ribonucleases within the samples. As such, RNA integrity is usually the

main concern upon RNA sample preparation (177). Generally, the certification of RNA

integrity is usually verified through an electrophoresis gel, monitoring the bands of

rRNA species since they include more than 90 % of the mass of total RNA. The

difficulty arises when there is a continued development and application of gene

expression studies based on quantitative, real-time RT-PCR and microarrays, which

require consistent and highly reproducible technologies (Comparing performance of the

Agilent 2100 Bioanalyzer bioanalyzer to traditional RNA analysis techniques, Agilent

Application Note, 2000). The method for these studies is a chip-based nucleic acid

analysis system, the Agilent 2100 bioanalyzer, where Pre-packed kits, standardized

sample preparation and automated analysis lessen manual interference yielding more

accurate and reproducible data requiring only 25 ng of total RNA (Comparing

performance of the Agilent 2100 Bioanalyzer bioanalyzer to traditional RNA analysis

techniques, Agilent Application Note, 2000). However, new technology is also available

that allows for a detection limit around 200 pg, specially designed for laser

microdissection technologies (Quality assurance of RNA derived from laser

microdissected tissue samples obtained by the PALM MicroBeam system using the

RNA 6000 Pico LabChip kit, Agilent Application Note, 2003). The very first indication of

good quality of total RNA samples when performing the evaluation of bioanalizer

electropherogram is the identification of well-defined ribosomal RNAs (rRNAs) peaks

(Comparing performance of the Agilent 2100 Bioanalyzer bioanalyzer to traditional

RNA analysis techniques, Agilent Application Note, 2000). These peaks however, are

species-dependent, therefore, they are prone to variance between them. Another

bioanalyzer feature involves the addition to all samples of a reference marker

compound with a fast migration so the software is able to align all electropherograms

within one LabChip run. Furthermore, the software can also automatically search and

identify peaks for the two most abundant large rRNAs (always labelled as 18S and

28S), and calculate peak area ratios. In Drosophila melanogaster however 28S rRNA is

usually split in two fragments, comigrating with 18S rRNA. At last, aside from rRNA

peaks the intensity of any other signal should be very low (Stringent RNA quality

control using the Agilent 2100 bioanalyzer, Agilent Technologies, 2005).

Regarding our analysis, the software available did not have the ability to recognize

Drosophila melanogaster rRNA peaks so we could only verify the definition of rRNA


39

peaks and concentration through the determination of the area under the entire RNA

electropherogram. The ladder supplies a concentration/area ratio that is applied to

convert the sample area into concentration values (Figure III.3). For the RNA extracted

from larvae expressing only GFP, the first sample electropherogram showed an area of

423,691 resulting in a concentration of 243,997 ng/µl, the second sample presented an

area of 466,238 resulting in a concentration of 268,499 ng/µl and the third an area of

20,896 resulting in a concentration of 12,034 ng/µl (Figure III.4). For the RNA extracted

from larvae expressing only p35, the first sample electropherogram demonstrated an

area of 676,874 resulting in a concentration of 389,802 ng/µl, the second showed an

area of 765,336 resulting in a concentration of 440,745 ng/µl and the third an area of

160,088 resulting in a concentration of 92,192 ng/µl (Figure III.5). For the RNA

extracted from larvae knocked down for Bub3 and expressing GFP, the first sample

electropherogram demonstrated an area of 672,095 resulting in a concentration of

387,049 ng/µl, the second showed an area of 193,792 resulting in a concentration of

111,602 ng/µl and the third an area of 116,453 resulting in a concentration of 67,063

ng/µl (Figure III.6). For RNA extracted from larvae knocked down for Bub3 expressing

GFP, and p35, the first repeat electropherogram presented an area of 2726,804

resulting in a concentration of 1570,324 ng/µl, the second an area of 2214,551

resulting in a concentration of 1275,326 ng/µl and the third area of 1119,473 resulting

in a concentration of 644,687 ng/µl (Figure III.7). Through this technology we also have

access to the gel photo where we ran our samples (Figure III.8).

Figure III.3 - Graphic representing the fluorescence peaks of the RNA ladder. [FU] represents fluorescence units

and [s] represents time in seconds.


40

Figure III.4 - Agilent Bioanalyzer electropherograms

showing rRNA peaks of RNA extract from wing

imaginal discs of 3rd

instar larvae expressing GFP.

[FU] represents fluorescence units and [s] represents

time in seconds. Graph a) represents the bioanalyzer

electropherograms of the first repeat of RNA extract from

wing imaginal discs of 3rd instar larvae expressing GFP. It

shows well-defined rRNA peaks (28s rRNA at

approximately 41,8 seconds and 18s rRNA at

approximately 43,2 seconds) with an area of 423,691

resulting in a concentration of 243,997 ng/µl. Graph b)

represents the bioanalyzer electropherograms of the

second repeat of RNA extract from wing imaginal discs of

3rd instar larvae expressing GFP. It shows well-defined

rRNA peaks (28s rRNA at approximately 41,8 seconds and

18s rRNA at approximately 43,2 seconds) with an area of

466,238 resulting in a concentration of 268,499 ng/µl.

Graph c) represents the bioanalyzer electropherograms of

the third repeat of RNA extract from wing imaginal discs of

3rd instar larvae expressing GFP. It shows well-defined


18s rRNA at approximately 43,2 seconds) however, with a

smaller area of 20,896 resulting in a lower concentration of

12,034 ng/µl.


41




instar larvae expressing P35. [FU]

represents fluorescence units and [s] represents time

in seconds. Graph a) represents the bioanalyzer

electropherograms of the first repeat of RNA extract from

wing imaginal discs of 3rd instar larvae expressing P35. It

shows well-defined rRNA peaks (28s rRNA at

approximately 41,8 seconds and 18s rRNA at

approximately 43,2 seconds) with an area of 676,874

resulting in a concentration of 389,802 ng/µl. Graph b)

represents the bioanalyzer electropherograms of the

second repeat of RNA extract from wing imaginal discs of

3rd instar larvae expressing P35. It shows well-defined


18s rRNA at approximately 43,2 seconds) with an area of

765,336 resulting in a concentration of 440,745 ng/µl.

Graph c) represents the bioanalyzer electropherograms of

the third repeat of RNA extract from wing imaginal discs of

3rd instar larvae expressing P35. It shows well-defined


18s rRNA at approximately 43,2 seconds) however, with a

smaller area of 160,088 resulting in a lower concentration

of 92,192 ng/µl.


42




instar larvae expressing GFP and

Bub3RNAi

. [FU] represents fluorescence units and [s]

represents time in seconds. Graph a) represents the

bioanalyzer electropherograms of the first repeat of RNA

extract from wing imaginal discs of 3rd instar larvae

expressing GFP, p35 and Bub3RNAi

. It shows well-defined

rRNA peaks (28s rRNA at approximately 41,8 seconds

and 18s rRNA at approximately 43,2 seconds) with an

area of 672,095 resulting in a concentration of 387,049

ng/µl. Graph b) represents the bioanalyzer

electropherograms of the second repeat of RNA extract

from wing imaginal discs of 3rd instar larvae expressing

GFP, p35 and Bub3RNAi

. It shows well-defined rRNA peaks

(28s rRNA at approximately 41,8 seconds and 18s rRNA

at approximately 43,2 seconds) with an area of 193,792

resulting in a concentration of 111,602 ng/µl. Graph c)

represents the bioanalyzer electropherograms of the third

repeat of RNA extract from wing imaginal discs of 3rd

instar larvae expressing GFP, p35 and Bub3RNAi

. It shows

well-defined rRNA peaks (28s rRNA at approximately 41,8

seconds and 18s rRNA at approximately 43,2 seconds)

with an area of 116,453 resulting in a concentration of

67,063 ng/µl.


43




instar larvae expressing GFP,

Bub3RNAi

and p35. [FU] represents fluorescence units

and [s] represents time in seconds. Graph a)

represents the bioanalyzer electropherograms of the first


instar larvae expressing GFP and Bub3RNAi

. It shows well-

defined rRNA peaks (28s rRNA at approximately 41,8



1570,324 ng/µl. Graph b) represents the bioanalyzer

electropherograms of the second repeat of RNA extract

from wing imaginal discs of 3rd instar larvae expressing

GFP and Bub3RNAi

. It shows well-defined rRNA peaks

(28s rRNA at approximately 41,8 seconds and 18s rRNA

at approximately 43,2 seconds) with an area of 2214,551

resulting in a concentration of 1275,326 ng/µl. Graph c)

represents the bioanalyzer electropherograms of the third


instar larvae expressing GFP and Bub3RNAi

. It shows well-

defined rRNA peaks (28s rRNA at approximately 41,8



644,687 ng/µl.


44

Figure III.8 - Electrophoresis gel with each repeat of each sample used in microarrays.

3.4. MICROARRAY ANALYSIS

3.4.1. MICROARRAY DATA

After verifying that the integrity and concentration of the samples, we proceeded to

perform a microarray protocol. Upon completion, validation and normalization (data not

shown) of the different hybridization, we resorted to the software MeV v4.6.1 to analyze

and compare the retrieved data. We did the following comparisons through t-test

(Figure III.9):

1) Expression profiles of flies expressing GFP were compared with flies

expressing GFP and Bub3RNAi. (annex 2)

2) Expression profiles of flies expressing GFP and Bub3RNAi were compared with

flies expressing GFP, Bub3RNAi and p35. (annex 3)


expressing GFP, Bub3RNAi and p35. (annex 4)


45

4) Expression profiles of flies expressing p35 were compared with flies expressing

GFP, Bub3RNAi and p35. (annex 5)


expressing p35. (annex 6)

Each comparison provided us a list with a large amount of genes differently expressed

between two genotypes. Therefore, to reduce the number of genes and obtain an

enhanced biological meaning we carried out t-tests choosing a p-value of 0,05,

allowing us to compare the distinct samples in terms of different expression profiles.

Analysis of these lists also required the choice of a cut-off fold in terms of differences in

expression. The selected fold was two-fold, meaning that we considered a gene

whenever the difference in expression in distinct samples was at least two fold, either

up or down-regulated. To illustrate the expression profiles and their cut-offs the

software provided a volcano plot for each comparison. Genes classified with the

desired cut off, 0,05 p-value and fold difference superior to two, are represented by

green dots. Red dots on the other hand represent genes with a p-value of at least 0,01

and a fold difference of less than two. This notation helps to understand the amount of

information lost and how it could be relevant. A closer look to the volcano plots

tendency reveals that flies expressing GFP and Bub3RNAi or flies expressing only p35

do not show a large amount of genes differently expressed when compared to the

control, flies expressing only GFP, with 1005 and 1096 genes respectively (Figure

III.10 and III.11 respectively). However, whenever the comparison involves flies with

the overgrowth phenotype, expressing simultaneously GFP, Bub3RNAi, and p35, the

amount of green dots and therefore differently expressed genes with a p-value of 0,05

and a fold superior to two increases drastically. This can be perceived when the

comparison involves flies expressing only GFP with 4556 genes (Figure III.12), flies

expressing only p35 with 5460 genes (Figure III.13) or flies expressing GFP and

Bub3RNAi with 3909 genes (Figure III.14).

Figure III.9 - Schematic representation of the t-test comparisons (represented by a black arrow) made with the

microarray output.


46

Figure III.10 - Volcano plot

comparing expression

profiles from microarray

output and selected genes

according to P-value

(represented in the vertical

axis by -log10(p-value) and

fold difference (represented in the horizontal axis). Expression profiles compared were Group B - tissue expressing

GFP and Bub3RNAi

(Ap-Gal4>UAS-GFP UAS-Bub3RNAi

) and Group A - tissue expressing GFP (Ap-Gal4>UAS-GFP).

Green dots represent genes with a fold induction superior equal to 2 and -2 and p-value superior or equal to 0,05. Red

dots represent genes with a fold induction difference lower to 2 and -2 and p-value superior or equal to 0,05.









P35 (Ap-Gal4>UAS-p35) Group A - tissue expressing GFP (Ap-Gal4>UAS-GFP). Green dots represent genes with a

fold difference higher or equal to 2 and -2 and p-value higher or equal to 0,05. Red represents genes with a fold

induction difference lower to 2 and -2 and p-value higher or equal to 0,05.


47









GFP, Bub3RNAi

and P35 (Ap-Gal4>UAS-GFP UAS- Bub3RNAi

UAS-p35) Group A - tissue expressing GFP (Ap-

Gal4>UAS-GFP). Green dots represent genes with a fold difference higher or equal to 2 and -2 and p-value higher or

equal to 0,05. Red represents genes with a fold induction difference lower to 2 and -2 and p-value higher or equal to

0,05.








fold difference

(represented in the horizontal axis). Expression profiles compared were Group B - tissue expressing GFP, Bub3RNAi


UAS-p35) Group A - tissue expressing P35 (Ap-Gal4>UAS-p35). Green

dots represent genes with a fold difference higher or equal to 2 and -2 and p-value higher/equal to 0,05. Red dots

represent genes with a fold induction difference lower to 2 and -2 and p-value higher or equal to 0,05.


48

Figure III.14. Volcano plot








GFP, Bub3RNAi


UAS-p35) Group A - tissue expressing GFP and Bub3RNAi

(Ap-Gal4>UAS-GFP UAS- Bub3RNAi

). Green dots represent genes with a fold difference higher or equal to 2 and -2 and

p-value higher or equal to 0,05. Red dots represent genes with a fold induction difference lower to 2 and -2 and p-value

higher or equal to 0,05.

3.4.2. MINING THE MICROARRAY DATA TO ACHIVE

RELEVANT BIOLOGICAL INFORMATION

The t-tests performed helped narrow the search for genes of biological interest,

however, the amount of genes is still too large to be manageable for a manual search

of genes of interest. Therefore, we resorted to another technique to reduce the list of

genes to a convenient yet relevant lower number. To achieve this we resorted to Venn

diagrams. The underlying principle behind the choices made lays in the selection of

genes that respond to a specific component. The focus was to understand what might

be influencing the overgrowth of the wing imaginal discs when the simultaneous

expression of Bub3RNAi and p35 occurs. Therefore, revolving around that genotype we

were able to distinguish three components. Genes either upregulated or downregulated

in the wing imaginal disc overgrowth could be a consequence of either the Bub3

knockdown component or p35 component or even the component revealing the

synergy between the two, i.e. the gene expression is only altered upon the synergistic


49

effect of Bub3RNAi and p35. Therefore, we made intersections with the lists using the

venn diagrams (Figures III.15 to III.20). The upregulated Bub3RNAi component (annex 7)

can be obtained using the upregulated group of genes in flies expressing Bub3RNAi, p35

and GFP that are not upregulated in the group of genes of flies expressing only GFP

(Figure III.15 blue circle). Subsequently, intersection with the upregulated gene group

of flies expressing Bub3RNAi, p35 and GFP that are not upregulated in the group of

genes of flies expressing only p35, removed this component (Figure III.15 green circle).

The red circle in Figure III.15 allows for the removal of the upregulated gene group

resulting from the effect of p35 alone by removing the group of genes upregulated in

flies expressing p35 that are not upregulated in flies expressing only GFP (figure III.15

Red circle). The amount of genes representing this upregulated component is 663 and

it is represented in Figure III.15 inside the yellow box. The downregulated component

(annex 8) rationale is exactly the same as for the upregulated one, using the

downregulated genes (Figure III.16). The amount of genes representing this

downregulated component is 351 and it is represented in Figure III.16 inside the yellow

box. The upregulated p35 component (annex 9) can be obtained using the upregulated

group of genes in flies expressing Bub3RNAi, p35 and GFP that are not upregulated in

the group of genes of flies expressing only GFP (Figure III.17 blue circle).

Subsequently, intersect it with the upregulated gene group of flies expressing Bub3RNAi,

p35 and GFP that are not upregulated in the group of genes of flies expressing only

Bub3RNAi, therefore removing this component (Figure III.17 green circle). The red circle

in Figure III.17 allows for the removal of the upregulated gene group resulting from the

effect of p35 alone by removing the group of genes upregulated in flies expressing p35

that are not upregulated in flies expressing only GFP. The amount of genes

representing this upregulated component is 540 and it is represented in Figure III.17

inside a yellow box. The downregulated component (annex 10) rationale is the exact

same as for the upregulated one but using instead downregulated genes (Figure III.18).

The amount of genes representing this downregulated component is 362 and it is

represented in Figure III.18 inside a yellow box.

The upregulated synergy component (annex 11) can be obtained using the

upregulated group of genes in flies expressing Bub3RNAi, p35 and GFP that are not

upregulated in the group of genes of flies expressing only GFP (Figure III.19 blue

circle). Afterwards, this list is intersected with the upregulated gene group of flies

expressing Bub3RNAi, p35 and GFP that are not upregulated in the group of genes of

flies expressing only p35, therefore removing this component (Figure III.19 green

circle). A final intersection with the upregulated gene group of flies expressing


50

Bub3RNAi, p35 and GFP that are not upregulated in the group of genes of flies

expressing only Bub3RNAi, removes this component (Figure III.19 red circle). The

remaining of genes are only differently expressed when Bub3RNAi and p35 are both

present. The amount of genes representing this upregulated component is 181 and it is

represented in Figure III.19 inside the yellow box. The downregulated component

(annex 12) rationale is the exactly the same as for the upregulated one using instead

the list for the downregulated genes (Figure III.20). The amount of genes representing

this downregulated component is 267 and it is represented in Figure III.20 inside the

yellow box.

Figure III.15 - Venn diagram. Blue circle represents genes

upregulated in tissue expressing GFP, Bub3RNAi

and P35 (Ap-

Gal4>UAS-GFP UAS-Bub3RNAi

UAS-p35) when compared with tissue

expressing GFP (Ap-Gal4>UAS-GFP); green circle represents genes


and P35 (Ap-



expressing P35 (Ap-Gal4>UAS-p35); red circle represents genes

upregulated in tissue expressing P35 (Ap-Gal4>UAS-p35) when

compared with tissue expressing GFP (Ap-Gal4>UAS-GFP); inside

the yellow box is the number of genes in common in the upregulated

portion of tissue expressing GFP, Bub3RNAi

and P35 (Ap-Gal4>UAS-

GFP UAS-Bub3RNAi

UAS-p35) when compared with tissue expressing GFP (Ap-Gal4>UAS-GFP) and the upregulated

portion of tissue expressing GFP, Bub3RNAi

and P35 (Ap-Gal4>UAS-GFP UAS-Bub3RNAi

UAS-p35) when compared

with tissue expressing P35 (Ap-Gal4>UAS-p35).


downregulated in tissue expressing GFP, Bub3RNAi

and P35 (Ap-


UAS-p35) when compared with

tissue expressing GFP (Ap-Gal4>UAS-GFP); green circle

represents genes downregulated in tissue expressing GFP,

Bub3RNAi


UAS-p35)

when compared with tissue expressing P35 (Ap-Gal4>UAS-p35);

red circle represents genes downregulated in tissue expressing P35

(Ap-Gal4>UAS-p35) when compared with tissue expressing GFP

(Ap-Gal4>UAS-GFP); inside the yellow box is the number of genes

in common in the downregulated portion of tissue expressing GFP, Bub3RNAi

and P35 (Ap-Gal4>UAS-GFP UAS-

Bub3RNAi

UAS-p35) when compared with tissue expressing GFP (Ap-Gal4>UAS-GFP) and the downregulated portion of

tissue expressing GFP, Bub3RNAi



expressing P35 (Ap-Gal4>UAS-p35).


51



and P35 (Ap-


UAS-p35) when compared with in

tissue expressing GFP (Ap-Gal4>UAS-GFP); green circle represents

genes upregulated in tissue expressing GFP, Bub3RNAi

and P35 (Ap-





); red

circle represents genes upregulated in tissue expressing P35 (Ap-

Gal4>UAS-p35) when compared with tissue expressing GFP (Ap-

Gal4>UAS-GFP); inside the yellow box is the number of genes in common in the upregulated portion in tissue

expressing GFP, Bub3RNAi



expressing GFP (Ap-Gal4>UAS-GFP) and the upregulated portion in tissue expressing GFP, Bub3RNAi

and P35 (Ap-


UAS-p35) when compared with tissue expressing GFP and Bub3RNAi

(Ap-Gal4\>UAS-

GFP UAS-Bub3RNAi

).



and P35 (Ap-


UAS-p35) when compared with in

tissue expressing GFP (Ap-Gal4>UAS-GFP); green circle represents

genes downregulated in tissue expressing GFP, Bub3RNAi

and P35 (Ap-





); red

circle represents genes downregulated in tissue expressing P35 (Ap-

Gal4>UAS-p35) when compared with tissue expressing GFP (Ap-

Gal4>UAS-GFP); inside the yellow box is the number of genes in

common in the downregulated portion in tissue expressing GFP, Bub3RNAi


UAS-p35) when compared with tissue expressing GFP (Ap-Gal4>UAS-GFP) and the downregulated portion in tissue

expressing GFP, Bub3RNAi





).

Figure III.19 - Venn diagram. Blue circle represents genes upregulated

in in tissue expressing GFP, Bub3RNAi

and P35 (Ap-Gal4>UAS-GFP

UAS-Bub3RNAi

UAS-p35) when compared with tissue expressing GFP

(Ap-Gal4>UAS-GFP); green circle represents genes upregulated in


and P35 (Ap-Gal4>UAS-GFP UAS-

Bub3RNAi

UAS-p35) when compared with tissue expressing GFP and

Bub3RNAi


); red circle represents

genes upregulated in tissue expressing GFP, Bub3RNAi

and P35 (Ap-



expressing P35 (Ap-Gal4>UAS-p35); inside the yellow box is the

number of genes of the unregulated portion in tissue expressing GFP,

Bub3RNAi


UAS-p35) when compared with tissue expressing GFP (Ap-

Gal4>UAS-GFP) however not present in neither the upregulated portion in tissue expressing GFP, Bub3RNAi

and P35


UAS-p35) when compared with tissue expressing GFP and Bub3RNAi

(Ap-


), nor the upregulated portion in tissue expressing GFP, Bub3RNAi

and P35 (Ap-


UAS-p35) when compared with in tissue expressing P35 (Ap-Gal4>UAS-p35).


52


downregulated in in tissue expressing GFP, Bub3RNAi

and P35 (Ap-



expressing GFP (Ap-Gal4>UAS-GFP); green circle represents genes


and P35 (Ap-





); red

circle represents genes downregulated in tissue expressing GFP,

Bub3RNAi


UAS-p35) when

compared with tissue expressing P35 (Ap-Gal4>UAS-p35); inside the

yellow box is the number of genes of the downregulated portion in




expressing GFP (Ap-Gal4>UAS-GFP) however not present in neither the downregulated portion in tissue expressing

GFP, Bub3RNAi


UAS-p35) when compared with tissue expressing GFP and

Bub3RNAi


), nor the downregulated portion in tissue expressing GFP, Bub3RNAi

and

P35 (Ap-Gal4>UAS-GFP UAS-Bub3RNAi

UAS-p35) when compared with in tissue expressing P35 (Ap-Gal4>UAS-p35).

3.4.3. VALIDATING VENN DIAGRAMS THROUGH

BIOINFORMATICS RESOURCES

Resorting to Venn Diagrams was the final bioinformatics approach to attempt to

turn the expression profiles more meaningful and manageable. Therefore, before

moving forward we attempted to understand if we were successful in our selections.

We decided to use DAVID bioinformatics resources 6.7. This resource refers to an

integrated biological knowledgebase and the analytic tools try to systematically extract

biological significance from large gene/protein lists (178). With DAVID, it is possible to

analyze gene lists resultant from high-throughput and integrated data-mining

environment. Initially, the gene list of interest is uploaded, holding any number of

common gene identifiers, and then analyzed with any of the desired pathway-mining

tool such as gene functional classification, functional annotation chart or clustering and

functional annotation table (178). However, due to the redundancy of biological

annotations, tools like the Functional Annotation Chart have the drawback of showing

analogous and significant annotations repeatedly. Such event weakens the focus on

biology. In order to diminish the redundancy, DAVID researchers have developed the

Functional Annotation Clustering tool that provides a clearer biology reading. The

hypothesis sustaining the grouping algorithm implies that closely related annotations

should have similar gene members. The Functional Annotation Clustering tool uses the


53

same techniques of Kappa statistics in order to measure the extent of common genes

between two annotations. Then with fuzzy heuristic clustering already used in Gene

Functional Classification Tool, it proceeds to classify the groups of similar annotations

according to the calculated kappa values (160). This technique leads to a higher

probability of groups being clustered together the more common genes annotations

they share. Annotation P-values integrated in each cluster possess the exact same

meaning as a p-value shown in a normal chart report for the same terms. However,

DAVID bioinformatics resources 6.7 uses a system called Group Enrichment Score

where they rank biological relevance according to the geometric mean in logarithm

scale of member’s p-values of the corresponding annotation cluster (160). Upon

uploading our gene lists obtained through Venn diagram, we proceeded by performing

a Functional Annotation Clustering. The obtained information emerged as very long

spreadsheets in form of an excel file showing large clusters of closely related

annotations with a particular enrichment score. The selected cut off for enrichment

score was 1,3 and it means that the geometric mean of the different cluster members is

approximately 0,05 providing us with a confidence interval of 95%. Regarding the gene

lists it is also important to refer that the program can only use the genes with an

annotated function, therefore, many of our choices of unknown CG numbers cannot be

included in this analysis. However, it can still give a good idea from the choices that

were annotated. We uploaded six different gene lists, the upregulated and

downregulated of each component: Bub3 knockdown component, the p35 component

and the synergy component and looked for enriched pathways related with anything

that could influence cell cycle, cell division and apoptosis in general. Regarding the

upregulated Bub3 knockdown component, we found thirteen different clusters above

the chosen enrichment score cut-off (annex 13). From these different clusters there is

one of particular interest with an enrichment score of 1,74. Cluster members are

related to DNA repair, response to DNA damage stimulus, telomere maintenance, and

telomere organization among other annotations (Figure III.21). Searching the

downregulated Bub3 knockdown component, eight distinct clusters (annex 14) were

identified with values of enrichment score higher than the cut-off. Upon analyzing them,

we were not able to extract any relevance to the purpose of our work. Nevertheless,

one of the clusters showing an enrichment score of 1,84 and its members are related to

to cell motility which might be important (Figure III.22). Investigating the upregulated

p35 component, we found twelve different clusters above the chosen enrichment score

cut-off (annex15). However, none seems to be related in any way to cell cycle, cell

division or apoptosis (Figure III.23). Inspecting the downregulated p35 component led


54

to 17 clusters whose values of enrichment score are higher than the cut-off (annex 16).

Thorough analysis revealed no significant cluster to the purpose of our work (Figure

III.24). Should put the figures next to the analysis.

Regarding the upregulated synergy component, we found one cluster above the

chosen enrichment score cut-off (annex 17). This particular cluster shows an

enrichment score of 1,30 and its members are closely related to cell cycle process,

mitosis, spindle components among other similar annotations (graph not shown). The

downregulated synergy component presents three clusters with an enrichment score

higher than the imposed cut-off (annex 18). Regarding cell cycle, we could not find any

cluster with members related to this event (Figure III.25).

Figure III.21 - Graphic showing annotation clusters from the upregulated gene list from Bub3 knockdown component

with an enrichment score superior to 1,3. Obtained through a functional annotation clustering analysis using DAVID

bioinformatics resources 6.7.

Figure III.22 - Graphic showing annotation clusters from the downregulated gene list from Bub3 knockdown component

with an enrichment score superior to 1,3. Obtained through a functional annotation clustering analysis using DAVID


3,33 2,96

2,23 2,22 2,09 2,09 2,02 2,01 1,83 1,74

1,56 1,49 1,42

0

1

2

3

4

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13

en

rich

me

nt

sco

re

Upregulated Bub3 component

4,28

3,24 3,03

2,62 2,28

1,84 1,49 1,47

0

1

2

3

4

5

B1 B2 B3 B4 B5 B6 B7 B8

en

rich

me

nt

sco

re

Downregulated Bub3 component


55

Figure III.23 - Graphic showing annotation clusters from the upregulated gene list from p35 component with an

enrichment score superior to 1,3. Obtained through a functional annotation clustering analysis using DAVID


Figure III.24 - Graphic showing annotation clusters from the downregulated gene list from p35 component with an



Figure III.25 - Graphic showing annotation clusters from the downregulated gene list from p35 component with an



3,22

2,11 2,09 2,03 1,79 1,67 1,61 1,57 1,50 1,47 1,41 1,41

0

0,5

1

1,5

2

2,5

3

3,5

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12

en

rich

me

nt

sco

re

Upregulated p35 component

4,99

4,12 3,67 3,61

2,56 2,45 2,31 2,02 2,01 1,95 1,90

1,60 1,60 1,59 1,58 1,42 1,38

0

1

2

3

4

5

6

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D11 D12 D13 D14 D15 D16 D17

en

rich

me

nt

sco

re

Downregulated p35 component

2,39

1,38 1,34

0

0,5

1

1,5

2

2,5

3

F1 F2 F3

en

rich

me

nt

sco

re

Downregulated synergy component


56

3.4.4. MINING THE MICROARRAY DATA FROM RAW TO

BIOLOGICALLY RELEVANT.

Resorting to Venn diagrammes helped to slim the list of relevant genes to a

much more manageable amount. However, as previously mentioned, there was still a

great quantity of background noise, resulting from the use of complex larval tissue

where many developmental pathways are highly represented. Therefore, it was

required human input to search for biological pertinence. Such action required the

following set of criteria. Firstly, we elected genes due to their biological relevance

taking particular attention to known function, pathways, protein-protein interactions and

the existence of homologous in humans using for that mostly NCBI and Flybase

databases. Secondly, genes with unknown function were selected if the difference of

expression was superior/equal to ten. Thirdly, an abnormally high fold difference was

also taken into account even if it was not an unknown sequence. As the candidates

seemed promising we decided that it was worth moving forward so we built a final list

comprising 210 genes, 139 upregulated (Table III.2) and 71 downregulated (Table

III.3).


57

Table III.2 - Table showing upregulated genes manually searched and selected from the Venn diagram components and their fold difference.

Upregulated genes

fold fold fold fold fold fold

Prx5037 1733,3 CG32625 35,74 CG4250 18,69 CG14355 12,95 CG34430 10,27 Irbp 3,58

Kap3 589,66 CG13403 34,72 CG5866 18,67 CG30485 12,9 CG13954 10,09 Dhc62B 3,55

PGRP-SB2 521,04 CG5646 32,63 CG18193 17,59 CG13116 12,52 Thor 10,08 Cp110 3,47

CG8620 382,58 CG32085 31,72 CG13659 17,35 AcCoAS 12,36 CG14907 10,06 otu 3,33

CG9641 250,14 CG2157 30,23 CG15142 17,23 CG12520 12,15 alphaTub85E 9,97 PNUTS 3,18

asf1 132,27 CG18522 28,9 CG13026 17,21 CG15497 12,11 CG14523 9,78 Socs36E 3,11

mus304 122,55 CG15047 28,86 CG14059 16,71 CG6836 12,02 CG32335 9,73 CG1951 3,09

CG4089 117,8 CG13283 28,68 CG42369 16,12 Ance-5 11,92 CG17124 8,26 CG6333 3,05

CG14695 109,48 CG12699 28,38 Ddc 15,88 CG32681 11,76 CG10859 8,11 mre11 2,94

nimC4 83,96 CG3635 26,3 CG32368 15,86 CG14253 11,44 CG14216 8,04 Ku80 2,93

CG6983 82,26 CG5783 26,17 CG15279 15,46 CG11750 11,34 CG6830 7,06 CG6191 2,88

CG13043 57,1 CG33920 25,44 CG31686 15,4 nimB1 11,26 nimC1 6,93 CG1572 2,78

CG6347 50,99 dpr18 24,77 CG18063 15,01 CG13465 11,21 ST6Gal 6,66 puc 2,73

Osi6 47,19 CG13250 24,74 CG4267 15,01 CG32313 11 phr 6,15 CG18745 2,67

CG31436 43,24 CG31533 24,45 CG3746 14,61 CG34301 10,98 wun2 6,13 Rep 2,54

CG34437 41,67 CG3884 24,07 CG31370 14,22 CG6495 10,9 CG12200 6 Dhc36C 2,51

CG33926 39,72 CG31279 23,11 CG33178 14,14 CG32668 10,86 Traf4 5,56 ncd 2,43

CG32155 39,52 CG31219 23,1 CG8160 14,04 CG6118 10,6 moody 5,36 rad50 2,19

CG13541 39,02 CG12911 22,79 CG4362 13,91 CG14636 10,6 rhea 5,36 Alr 2,15

CG4367 38,05 CG11073 22,52 CG15784 13,25 Neu3 10,53 os 5,18 CG15528 2,14

CG12519 37,77 CG8193 21,27 CG4763 13,18 CG17325 10,49 Aplip1 4,9 CG13559 2,13

CG9517 37,1 CG17108 20,07 CG5399 13,06 CG6579 10,49 zormin 4,67 f-cup 2,1

CG11893 35,75 CG14321 19,01 dpr6 13,04 CG7120 10,45 wun 3,86 mus205 2,04

CG7188 2,03

Table III.3 - Table showing downregulated genes manually searched and selected from the Venn diagram

components and their fold difference.

Downregulated genes

fold fold fold fold fold

Gyc-89Da 569,06 CG15905 18,35 Skeletor 11,1 CG12071 5,9 CG7231 2,65

phr6-4 299,38 CG5157 17,64 CG31875 10,65 CG14915 5,77 CG16947 2,65

CG2663 56,94 CG17140 17,04 CG12105 10,63 Pkc53E 5,61 CG34380 2,5

CG14356 39,98 CG14958 16,91 Mdr50 10,53 CG15544 5,36 Cam 2,41

CG3355 28,51 CG14395 15,87 CG31676 10,53 CG13138 4,37 CG14222 2,35

CG13494 28,32 Spn28Db 15 B-H1 10,35 Damm 4,24 CG3769 2,35

CG13036 26,16 CG7906 14,16 CG34342 9,69 Top1 4,22 tws 2,23

Spn43Aa 23,8 by 13,75 CG16758 8,94 CG2556 4,18 rab3-GEF 2,23

CG8419 22,62 CG9416 12,54 CG5023 8,42 Alk 4,12 Roc1a 2,22

CG10317 20,76 CG13081 12,46 Phlpp 6,99 phyl 3,61 Wnt2 2,2

CG9238 20,3 CG13101 12,09 CG6287 6,97 sns 3,14 Drep-4 2,18

Spn28Da 20,29 CG3823 11,96 CG4586 6,78 Ror 3,03 Sin3A 2,09

CG12912 19,57 CG13408 11,84 C15 6,46 btl 2,93 PPP4R2r 2,06

CG31790 19,25 Spn100A 11,47 CanB 6,01 InR 2,81 Rapgap1 2,05

fz 2


58

3.5. THE ENHANCER/SUPPRESSION OF PHENOTYPE

SCREEN

The purpose of using a microarray approach and all the subsequent

bioinformatics steps was to be able to select a manageable list of candidate genes for

a relevant biological role in our tumorigenesis model. Then, using RNAi technology, we

performed a screen with in vivo RNAi for each of the chosen genes. Flies carrying

individual RNAis were crossed with either a fly knockdown for Bub3 (Ap-Gal4>UAS-

GFP UAS-Bub3RNAi/ CyO) or a fly knockdown for Bub3 but also expressing the

apoptosis inhibitor DIAP1 (Ap-Gal4>UAS-GFP UAS-Bub3RNAi UAS-DIAP1/ CyO). We

chose to overexpress DIAP1 using a GAL4-UAS system instead of P35 since the

phenotype could be analyzed in the adult (Figure III.26) (140). In addition, the

“intermediate” phenotype of the DIAP1, a fly presenting a larger, and generally

deformed thorax, with some loss of bristles, empty sockets and deformed wings

showing necrosis, would also allow us to look for suppression, as well as enhancement

of the phenotype. It is important to note that the candidates genes are upregulated or

downregulated when comparing Ap-Gal4>UAS-GFP UAS-Bub3RNAi/ UAS-p35 flies with

Ap-Gal4>UAS-GFP/ CyO regardless of which Venn diagram they were chosen.

Figure III.26 - Schematic representation of a

cross that enables the expression of a gene

of interest through the use of Gal4-UAS

system

We classified the phenotypes found as “unaltered”, “voluptuous thorax”, “double

hunchback thorax” and “lethal”. Unaltered (Ua) phenotype defines flies presenting a

phenotype closely related to the phenotype of the driver fly used in the cross, either

expressing Bub3RNAi or expressing Bub3RNAi together with DIAP1. Voluptuous thorax

(VT) phenotype consists of a tall and swollen thorax when compared with the


59

phenotype of the driver fly used in the cross, either expressing Bub3RNAi or expressing

Bub3RNAi together with DIAP1 (Figure III.27 - a); b); c)). When comparing to either

driver phenotypes, Double hunchback thorax (DHT) phenotype shows a tall and

swollen thorax with a division in the middle like two separate hunchbacks (Figure III.27

- d); e); f)). Lethal (Lt) phenotype as the name implies reports to the crosses that do

not originate viable adult flies.

Figure III.27 - Image depicting the different phenotypes we found in the screen. a) and b) represent the phenotype

classified as “Voluptuous thorax” (VT) under visible light. c) represents the phenotype classified as “Voluptuous

thorax” (VT) under GFP excitatory light. d) and e) represent the phenotype classified as “Double hunchback thorax”

(DHT) under visible light. f) represents the phenotype classified as “Double hunchback thorax” (DHT) under GFP

excitatory light.

We classified each phenotype found for the 1) RNAi for upregulated genes

crossed to the Bub3RNAi driver (Table T1), 2) RNAi for downregulated genes crossed to

the Bub3RNAi driver (Table T2), 3) RNAi for upregulated genes crossed to the Bub3RNAi

expressed together with DIAP1 driver (Table T3) and 4) RNAi for upregulated genes

crossed to the Bub3RNAi expressed together with DIAP1 driver (Table T4).


60

Table III.4 - Table showing RNAi screen results from upregulated genes crossed with the Bub3

RNAi.

Upregulated genes crossed with Bub3RNAi

Ua VT DHT Lt

CG4267 CG32681 f-cup PGRP-SB2 CG9641 CG6830 Rep

os CG6347 CG5399 CG14321 CG14907 CG18745 rhea

CG13026 CG6191 CG14523 CG8620 CG2157 CG5866 CG32335

CG14636 CG13116 CG14216 CG13659 nimB1 CG4089 CG13465

CG31370 CG15497 CG13954 CG4763 Dhc62B Cp110 CG6579

CG10859 CG17108 CG15784 CG12520 CG34430 CG31686 Neu3

Dhc36C nimC4 CG33178 CG31436 CG11750 CG18193 CG32668

CG7120 CG4250 CG32313 CG13250 CG3746 CG33920 mus205

mre11 CG9517 zormin CG12699 wun2 CG30485 PNUTS

CG32085 CG13043 CG5783 CG6836 nimC1 CG12911 asf1

CG6983 CG13559 CG15142 CG13541 Alr CG31533 ncd

CG42369 CG6118 CG11893 otu Aplip1 CG8193

Osi6 CG15047 CG34301 CG12519 CG1572 CG6333

dpr18 alphaTub85E CG15528 CG3884 Ddc CG31279

AcCoAS CG32625 CG17325 CG13283

ST6Gal CG32368 Prx5037 CG34437

CG18522 rad50 CG32155 CG14355

Thor CG11073 CG8160 CG1951

CG14695 dpr6 CG13403 CG7188

Traf4

Table III.5 - Table showing RNAi screen results from downregulated genes crossed with the Bub3RNAi

.

Downregulated genes crossed with Bub3RNAi

Ua VT DHT Lt

Ror CG9238 fz CG13494 Top1 CG34380 CG14958

InR CG6287 CG10317 phr6-4 CG15905 Spn100A Cam

CG13081 CG5157 Roc1a Skeletor CG4586 CG31676 tws

phyl CG12071 CG5023 Spn28Da Rapgap1 CG14222

CG34342 CG3355 CG7231 sns Pkc53E Gyc-89Da

Spn28Db Wnt2 C15 CG16758 Drep-4

CG17140 CG31875 CG8419 CG3769 CG13408

Mdr50 B-H1 CG13101 CG12912 CG13036

CanB Spn43Aa CG3823 btl Damm

CG14356 PPP4R2r CG16947 Phlpp

CG2556 CG31790


61

Table III.6 - Table showing RNAi screen results from upregulated genes crossed with Bub3

RNAi also expressing DIAP1.

Upregulated genes crossed with Bub3RNAi also expressing DIAP1

Ua VT DHT Lt

CG4267 CG15047 CG34301 os nimB1 CG33920 Rep

CG13026 alphaTub85E CG15528 CG31370 CG18745 wun2 rhea

CG14636 CG32625 CG17325 CG10859 CG5866 Prx5037 CG32335

Dhc36C CG32368 CG32155 CG32085 AcCoAS nimC1 CG13465

CG7120 rad50 CG8160 CG14907 ST6Gal CG8620 CG32668

mre11 CG11073 CG13403 CG6830 Thor PNUTS mus205

CG6983 dpr6 CG14321 CG2157 CG4089 CG12911 CG30485

CG9641 f-cup CG13659 CG15497 Cp110 CG4763 ncd

CG42369 CG5399 CG31436 CG11750 Dhc62B CG31533

Osi6 CG14523 CG12699 CG31686 CG6347 CG12520

dpr18 CG14216 CG6836 PGRP-SB2 CG6191 CG13250

CG18522 CG13954 otu asf1 CG13116 CG8193

CG14695 CG15784 CG12519 CG13541 CG6579 CG6333

CG32681 CG33178 CG3884 CG34437 CG34430 Alr

nimC4 CG32313 CG13283 Ddc CG17108 Aplip1

CG9517 zormin CG14355 CG4250 CG1572

CG13043 CG5783 CG1951 Neu3 CG31279

CG13559 CG15142 Traf4 CG3746 CG7188

CG6118 CG11893 CG18193

Table III.7 - Table showing RNAi screen results from downregulated genes crossed with Bub3RNAi

also expressing

DIAP1.

Downregulated genes crossed with Bub3RNAi and DIAP1

Ua VT DHT Lt

Ror CG5157 CG10317 Skeletor InR Top1 CG14958

Pkc53E CG3355 Roc1a Spn28Da CG13081 Spn100A Cam

phyl CG13408 CG5023 sns CG15905 CG6287 tws

CG34380 CG13036 C15 CG16758 CG4586 CG12071 Damm

CG17140 Wnt2 CG8419 CG3769 Rapgap1 CG13101 CG14222

Mdr50 CG31875 CG3823 CG12912 Drep-4 Gyc-89Da

CanB B-H1 CG16947 btl CG34342

CG14356 Spn43Aa CG2556 Phlpp Spn28Db

CG9238 PPP4R2r CG13494 CG31790 CG7231

CG31676 fz phr6-4


62

We tested 175 out of 210 candidates. We decided to analyze the data starting

with the phenotype:

Hyperplasia - Could indicate that tumor suppressors are downregulated and

oncogenes upregulated, which means RNAi for candidates will suppress growth if they

are oncogenes, increase growth if they are tumour suppressors. High levels of

apoptosis - could indicate that apoptosis promoters are upregulated, and apoptotic

suppressors downregulated. On the other hand if the thorax was bigger it could point to

apoptosis promoters being downregulated or apoptotic suppressors upregulated.

Therefore, RNAi for candidates that increase apoptosis will reduce apoptosis, the ones

that block apoptosis will have apoptosis increased. The cross-talk between the two

screens is possible when we can identify a gene in the screen with Bub3RNAi that can

“replace the DIAP effect” by being either a pro-apoptic gene RNAi will suppress

apoptosis) and mimic the growth phenotype or by being a suppressor of a suppressor

of apoptosis therefore giving the RNAi the same phenotype. This is also important, as

we can’t assume that the candidate gene is directly downstream of Bub3. Therefore,

when we find an enhanced phenotype with the downregulation of a candidate gene

through RNAi we are expecting it to have one of the following functions (table III.8):

1) It could be an upregulated tumor suppressor that downregulation prevents its

function and leads to cell proliferation;

2) It could be a downregulated tumor suppressor that further downregulation could

make it reach levels of translation that allow for an increased cell proliferation;

3) It could be an upregulated suppressor of an oncogene that downregulation

prevents its function allowing the oncogene to act and promote cell proliferation;

4) It could be a downregulated suppressor of an oncogene that further

downregulation prevents its function even more leading increase of the

oncogene function and consequent of promotion of cell proliferation;

5) It could be an upregulated activator of a tumor suppressor that downregulation

prevents its function impeding the tumor suppressor function and promoting cell

proliferation;

6) It could be a downregulated activator of a tumor suppressor that further

downregulation prevents its function not allowing the tumor suppressor to act

and therefore promoting cell proliferation;

7) It could be an upregulated pro-apoptotic gene that downregulation impedes

apoptosis therefore leading to an enhanced phenotype

8) It could be a downregulated pro-apoptotic gene that further downregulation

blocks apoptosis even more therefore leading to an enhanced phenotype


63

9) It could be an upregulated suppressor of an anti-apoptotic gene that

downregulation impedes apoptosis therefore leading to an enhanced phenotype

10) It could be a downregulated suppressor of an anti-apoptotic gene that furher

downregulation blocks apoptosis therefore leading to an enhanced phenotype

11) It could be an upregulated activator of a pro-apoptotic gene that downregulation

impedes apoptosis therefore leading to an enhanced phenotype

12) It could be a downregulated activator of a pro-apoptotic gene that further

downregulation blocks apoptosis even more therefore leading to an enhanced

phenotype

On the other hand, when we find a suppressed phenotype with the

downregulation of a candidate gene through RNAi we are expecting it to have one of

the following functions (table III.8):

1) It could be an upregulated oncogene that downregulation prevents its function

and does not lead to cell proliferation;

2) It could be a downregulated oncogene that further downregulation could make it

reach levels of translation that would not allow for an increased cell

proliferation;

3) It could be an upregulated suppressor of a tumor suppressor that

downregulation prevents its function allowing the tumor suppressor to act and

prevent cell proliferation;

4) It could be a downregulated suppressor of a tumor suppressor that further

downregulation prevents its function even more leading increase of the tumor

suppressor function and consequent of inhibition of cell proliferation;

5) It could be an upregulated activator of an oncogene that downregulation

prevents its function impeding the oncogene function not leading to cell

proliferation;

6) It could be a downregulated activator of an oncogene that further

downregulation prevents its function not allowing the oncogene to act and

therefore not leading to cell proliferation;

7) It could be an upregulated anti-apoptotic gene that downregulation promotes

apoptosis therefore leading to a supressed phenotype;

8) It could be a downregulated anti-apoptotic gene that further downregulation

promotes apoptosis even more therefore leading to a suppressed phenotype;

9) It could be an upregulated suppressor of a pro-apoptotic gene that

downregulation leads to apoptosis therefore leading to an enhanced phenotype


64

10) It could be an downregulated suppressor of a pro-apoptotic gene that further

downregulation leads to apoptosis therefore leading to an enhanced phenotype

11) It could be an upregulated activator of an anti-apoptotic gene that

downregulation lead to an increased apoptosis therefore leading to a

suppressed phenotype

12) It could be an downregulated activator of an anti-apoptotic gene that further

downregulation increases apoptosis even more therefore leading to a

suppressed phenotype

Table III.8 - Possible functions of candidate genes.

We did not observe any RNAis that suppressed the phenotypes, therefore we

focused on the enhanced phenotype. We then analyze each hit of the most enhanced

phenotype found, the “Double hunchback thorax”, looking for connections with

tumorigenesis, cell proliferation, cell cycle and related annotations


65

CHAPTER 4

DISCUSSION


66

4.1. INTRODUCTION

Understanding tumorigenesis requires a systematic approach involving data

coming from different studies, from human cancer and from animal models. Regarding

the link between SAC, mitosis and tumorigenesis there seems to be a lot of valuable

data coming from in vitro studies. However, there is a need for in vivo studies using

animal models. Since little is known about Bub3 protein function and its connection with

tumorigenesis, we decided to investigate the link further by first using a high throughput

technology, microarrays. Resorting to microarrays allowed us to perform a large screen

for genes that might be involved in the hyperplastic growth transformation that is

observed when Bub3 is knockdown together with apoptosis inhibition.

4.2. REASONS FOR PHENOTYPE SELECTION

In order to move forward for a high throughput analysis we needed to establish

which flies we would collect wing imaginal discs from third instar larvae for RNA

extraction. This selection includes deciding which phenotypes are relevant for

understanding the overgrowth process and the required controls. We chose four

different phenotypes expressing the different transgenes in the wing imaginal disc. Our

main sample was the phenotype from Bub3 knockdown expressing GFP and p35 as it

has hyperplastic wing imaginal discs therefore serving as the main focus from which

we trimmed the large amount the genes differentially regulated into a manageable list

for further testing. With this phenotype, we wanted to answer one of the main questions

of the project regarding the consequences of impairing apoptosis using p35 when the

SAC protein Bub3 is depleted. To understand the role of Bub3 protein by itself and to

compare the expression profile it originates with the expression profile from the

overgrowth phenotype, we selected the Bub3 knockdown fly expressing GFP. This

particular genotype allowed us to follow the effects of having a fly with a SAC protein

depleted without interfering with apoptosis. Furthermore, comparing the transcriptional

profiles for these wing imaginal disc genotypes with and without p35, we expected to

decipher the changes in terms of transcription when introducing the apoptosis inhibition

factor. The presence of the heterologous P35 protein has been tested in the wing

imaginal disc and does not cause any significant alteration, except for a shortening of

wing vein 5 (135). The transcriptional profile from these flies allow for the identification


67

of what could be the effect due to preventing apoptosis or the effect of the synergy it

seems to have with SAC proteins. As we wanted to observe this, we also needed to

verify the consequences of inhibiting apoptosis alone. In order to accomplish that, we

selected a fly expressing the apoptosis inhibitor P35. We also needed an unaltered fly

in terms of SAC and apoptosis. For that purpose, we selected a fly expressing only

GFP. This fly allowed us to compare every other phenotype expression profile with a

normal condition profile and account for any change due to the expression of GFP

alone.

4.3. MICROARRAY OUTPUT REVEALS SIGNIFICANT

DIFFERENCE IN EXPRESSION PROFILES

DEPENDING ON GENOTYPES.

Data retrieved from microarrays from different genotypes was compared in

terms of p-value through t-test and fold difference using volcano plots. When choosing

a p-value for gene expression we considered and tested a p-value of 0,01, however the

amount of the information lost increased drastically. Therefore, we opted for lower 0,05

p-value granting us a confidence interval of 95% so we could reduce the data. Along

with the p-value, there was a need to define the fold difference between expression

profiles of different samples. The selected fold difference was 2 although we have seen

reports of lower values like 1,5 (160). We went for a fold difference of 2 since we were

not interested in very small changes and were expecting drastic changes to be the key.

However, the opposite might also be true since we are working with such important

cellular events, that a minimum change could influence, in theory, cells fate, and by

overlooking lower folds, we might have lost some valuable information. Also as this

was a first approach lower fold numbers would increase the number of genes

drastically and prevent a proper analysis. Observing the comparison between the

expression profiles of Bub3 knockdown fly with the control fly expressing only GFP we

detected one of the less spread volcano plot. Consequently, the number of genes

differently expressed are low. With this particular result, we can assume that the

depletion of Bub3 does not promote a major change in transcriptional profile when

compared to other genotypes. One possibility could be that RNAi is not uniform across

wing discs cells and therefore we are mostly analyzing those that survive and are not

significantly affected. The observation of a less spread volcano plot is also true when


68

comparing the expression profile of flies expressing p35 and control flies expressing

only GFP. Flies expressing p35 have their apoptosis inhibited however, SAC is still

functional, therefore the transcriptional profile is not dramatically altered. Moreover,

apoptosis in wild type discs is a rare phenomenon (approximately 1,4%) therefore one

would expect that prevent it does not alter the overall transcription profile (134). On the

other hand, when comparing the expression profile of Bub3 knockdown fly expressing

GFP and p35 with the other phenotypes transcription profiles we observed a dramatic

change in transcription. This means that when the two events occur together and as

cells undergo transformation, become immortal and tumorigenic, their transcription

profile is dramatically altered. Moreover, another trend is visible in the volcano plots for

this data, specifically the there are more upregulated genes than downregulated genes

in the overgrowth phenotype when comparing to any other phenotype. This might

indicate that gene upregulation has a larger role on the overgrowth phenotype,

validating our original assumtions.

4.4. BIOINFORMATIC METHODS SHOW ENRICHMENT

OF RELEVANT PATHWAYS FROM VENN DIAGRAM

SELECTION.

The microarray analysis goal was to find candidate genes responsible for the

tumorigenesis phenotype. The difficulty was to come up with a manageable list of

genes with biological relevance to the events and use them to perform further testing in

vivo. However, simple t-test comparison did not allow a selection of genes with a

manageable size. Therefore, we needed to resort to other methods to narrow the

search. With the visualization aid of Venn diagrams, we developed a rationale to

reduce the data. We separated our candidate genes in three classes and classified

them as components, the Bub3RNAi component, the p35 component, and the

component of the synergy effect of both components. The Bub3RNAi component is

expected to present genes responding to the lack of a functional SAC, as the p35

component should highlight genes responding to a block of apoptosis, and the synergy

component is expected to bring genes responding to the contribution of a lack of a

functional SAC and the apoptosis inhibition together. Although the same gene could

come from the same component we expected to find more often genes in the Bub3RNAi

component with a possible function as oncogenes or tumor suppressors or even their

activators or suppressors. On the other hand from the p35 component we were


69

expecting a list of genes more closely related with apoptosis either pro- or anti-

apoptotic genes or their activators or suppressors. We resorted to DAVID

bioinformatics resources 6.7 to evaluate the underlying principle beneath our selection.

We performed functional annotation clustering of each group of genes obtained from

the Venn diagrams. We chose a cut-off of approximately 1,3 enrichment score which

corresponds to the geometric mean of a p-value equal to 0,05 from all the members of

the particular cluster. This tool enrichment p-value is calculated based on, EASE

Score, a modified Fisher Exact Test (179). For further understanding I designed the

following example: assuming one submits a list of 400 genes and 20 hit the particular

category “cell cycle” and the background list from which the comparison was made has

300 genes in that category out of the total number of known genes for that organism

which could be 20000. With this information, we then ask: is the term “cell cycle”

enriched/over-represented in the list sent comparing with the number of hits that term

has in all the genome? The program then calculates one tail (right/greater tail) Fisher

Exact Test giving us a p-value and therefore the confidence interval of that particular

term being over-represented or not. It is important to notice that we are searching for

particular pathways that in theory could be related to an overgrowth phenotype which

represent a very small amount. Moreover, due to complexity and the unpredictable time

point of the tissue from where the data was extracted, it is likely that these pathways, if

altered, are being masqueraded beneath all the other pathways differently expressed

during development. As mention before we identified thirteen clusters above an

enrichment score of 1,3 in the upregulated Bub3RNAi component. From this thirteen we

classified one as relevant for the purpose of this thesis. This cluster is related to

several biological process GO terms like “DNA repair”, “telomere maintenance”,

“telomere organization” (180). Regarding the downregulated Bub3RNAi component we

did not classify any cluster as relevant however one of the clusters with cell motion as a

biological process GO term is worth referring. We did not define any cluster as

relevant for the p35 component either upregulated or down regulated. On the other

hand, the upregulated synergy component revealed only one cluster above the

specified enrichment value with 1,30 and some of its members are cellular component

GO terms like “spindle” or “microtubule cytoskeleton” and biological process “cell

cycle”. This is an expected result and supports the rationale used as the overgrowth

phenotype appears only in the case of the expression of p35 with Bub3 depletion. The

downregulated synergy component does not present any relevant cluster above the

specified enrichment value. Analysis of the previous results shows again, as referred

before in the volcano plots, the tendency for higher relevance in the upregulated genes.


70

However, downregulated genes should not be overlooked. Another important feature

from the functional annotation clustering is the loss of information due to some

members of clusters with higher p-values might alter the p-value geometric mean from

which the enrichment score is calculated

4.5. LEGITIMACY OF MANUAL SELECTION FROM VENN

DIAGRAM.

Biological data-mining has become increasingly complex and the present state

of enrichment tools impede a pure statistical solution allowing only for exploratory data-

mining methods. The biological knowledge of a researcher with aid from integrated

annotation databases, computing algorithms and the enrichment P-values is still crucial

for a valuable analytic conclusion (160). Our final goal was to perform an in vivo RNAi

screen. Therefore, we required a manageable list of genes differently expressed and

the Venn diagram approach did not confer such list. Additional selection with the help

of known biological data was the last logical step to produce a list small enough to work

with but that did not lose its biological significance. We began by defining selection

criteria. First, we investigated each candidate gene by collecting information that could

be in any way relevant to the work in progress. We started by looking for fold difference

(prioritizing higher values), annotated functions, pathways and reasons for being

involved in tumorigenesis, cell division, cell cycle, apoptosis, any other closely related

function and the existence of human homologs (in a positive case the priority of the

candidate would be higher). With this first approach, we were able to get rid of many

well-known genes that were known to perform in other non-related functions. In case of

doubt, we also paid attention to protein-protein interactions using NCBI database

looking for evidence of relations between the product of our selected genes and

proteins involved in events like cell cycle, apoptosis, microtubules, or spindle assembly

checkpoint among others. We did not want to neglect genes without an annotated

function since ultimately, we wanted to find key new players in the hyperplastic growth.

We decided to include a threshold of a 10-fold difference, imposed by the available

resources, means and maintaining a manageable list, where we would include any

unknown sequences with a superior or equal level.


71

4.6. BUB3 HAS AN ADDITIONAL ROLE AS A TUMOR

SUPPRESSOR.

We analyzed what could be each of genes that showed the most enhanced

phenotype in the performed screen. We also expect that whatever function they

perform should be influenced by the downregulation of Bub3 although further

investigation like testing the candidate genes by themselves is required. The

information depicted below is provided from Flybase.

4.6.1. GENES UPREGULATED IN BUB3RNAI

BACKGROUND SCREEN

Regarding the Bub3RNAi background the hits from the most enhanced phenotype

were:

CG6830 upregulated 7,06 fold in the Bub3 knockdown disc expressing GFP

and p35 when compared with the control disc expressing only GFP. It has ATPase

activity according to flybase and it interacts with HSP83 (181), which has an unknown

function in the mitotic spindle organization (182). Therefore, the gene CG6830 could

have a function of a suppressor of an oncogene or activator of a tumor suppressor.


and p35 when compared with the control disc expressing only GFP. It has no

annotated function according to flybase. However it seems to interact in an unknown

way with several cell cycle regulators like CDK4, CYCG, CYCK, MCM5 among others

(183). This interactions lead to the possible function as a tumor suppressor, suppressor

of an oncogene or an activator of a tumor suppressor.


and p35 when compared with the control disc expressing only GFP. This gene has

hardly any information available. So it could have any of the possible functions related

to an enhanced phenotype described in table III.8.


and p35 when compared with the control disc expressing only GFP. There are no


72

studies related to this gene therefore no annotated function. As such, it could perform

any of the possible functions related to an enhanced phenotype described in table III.8.

However, we can speculate that such a high fold could indicate a major role in the

tumorigenesis process.

Cp110 upregulated 3,47 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. It is linked with

centriole replication (184) it interacts with the Drosophila P53 and since its

phylogenetically related to the ancestor of the mammalian p53 family of tumor

suppressors (185), it could be an activator of the said tumor suppressor.







hardly any information available. It does seem to have a function related to electron

carrier activity inferred through electronic annotation. Therefore, it could have any of

the possible functions related to an enhanced phenotype described in table III.8.


and p35 when compared with the control disc expressing only GFP. There are no

studies related to this gene therefore no annotated function. As such, it could perform

any of the possible functions related to an enhanced phenotype described in table III.8.


and p35 when compared with the control disc expressing only GFP. This gene doesn’t

have any information available. Therefore, it could have any of the possible functions

related to an enhanced phenotype described in table III.8.


and p35 when compared with the control disc expressing only GFP. We found no

studies regarding this gene and as such , it could have any of the possible functions

related to an enhanced phenotype depicted in table III.8.


73






and p35 when compared with the control disc expressing only GFP. Through gene

similarity it has been associated with monophenol monooxygenase activity. In terms of

interactions it shows a relevant interaction with P53 and like previously mentioned,

since it is phylogenetically related to the ancestor of the mammalian p53 family of

tumor suppressors (185), it could be an activator of a tumor suppressor.


and p35 when compared with the control disc expressing only GFP. It has no studies

related to its function. However it interacts with P53 which could reveal an activator of a

tumor suppressor function (185). It also interacts with tribbles (trbl) a protein with

kinase activity and cells enters mitosis prematurely in tribbles mutant embryos (186),

which could mean the gene CG6333 might have a function as an activator of a tumor

suppressor.





4.6.2. GENES DOWNREGULATED IN BUB3RNAI

BACKGROUND SCREEN

Regarding the Bub3RNAi background the hits from the most enhanced phenotype

were:

CG34380, downregulated 2,5 fold in the Bub3 knockdown disc expressing GFP

and p35 when compared with the control disc expressing only GFP. This particular


74

gene has little information however, through electronic annotation it has been

connected to cell adhesion. It also has an interaction with smt3 (181), a Drosophila

gene involved in mitotic cell cycle event as it is involved in the SUMO pathway and

seems to be required for for the syncytial mitotic cycles, for cell cycle progression in S2

cells and larval tissue and it even seems to be connected with Ras signaling (187).

Although CG34380 does not have a large amount of data available, it has a human

homolog, discoidin domain receptor tyrosine kinase 2 (DDR2) from which some cancer

related studies were done. Steven Wall et al. showed that this cell surface-associated

tyrosine kinase receptor that binds to fibrillar collagen (FC), endow with a new

mechanism responsible for the G0/G1 cell cycle arrest observed when tumor cells are

in contact with FC (188). Downregulation of CG34380 could then interfere with the cell

proliferation process either by influencing signaling pathways like the Ras pathway

through its interaction with SMT3, or even through some similar function to its human

counterpart DDR2. Since its annotated function of cell adhesion is not closely

associated with cell proliferation per se, we expect it to probably have a function of a

suppressor of an oncogene or activator of a tumor suppressor, which correlates with its

interaction with smt3 as this gene seems to be involved in pathways that could lead to

tumorigenesis. According to the data from its human homolog of being responsible for

a cell cycle arrest, we could go further and hypothesize that the gene CG34380 could

be an activator of a tumor suppressor.

Serpin 100A (Spn100A), downregulated 11,47 fold in the Bub3 knockdown disc

expressing GFP and p35 when compared with the control disc expressing only GFP.

There are no studies regarding this gene being only classified with serine-type

endopeptidase inhibitor activity through gene or structural similarity. It could have any

of the possible functions related to an enhanced phenotype described in table III.8.

CG31676 downregulated 10,53 fold in the Bub3 knockdown disc expressing

GFP and p35 when compared with the control disc expressing only GFP. This gene

has hardly any information available. So it could have any of the possible functions



75

4.6.3. GENES UPREGULATED IN BUB3RNAI DIAP1

BACKGROUND SCREEN

Regarding the Bub3RNAiDIAP1 background the hits from the most enhanced phenotype

were:

The candidate genes CG18745, CG5866, CG4089, Cp110, CG18193, CG33920,

CG12911, CG31533, CG8193, CG6333 and CG31279 functions and previsions are

already depicted in the Bub3RNAi background screen.

Nimrod B1 (NimB1) found upregulated 11,26 fold in the Bub3 knockdown disc


This gene has hardly any information although it is linked with defense response to

bacterium through gene similarity (189). Therefore it could have any of the possible

functions related to an enhanced phenotype described in table III.8.

Acetyl Coenzyme A synthase (AcCoAS) found upregulated 12,36 fold in the Bub3

knockdown disc expressing GFP and p35 when compared with the control disc

expressing only GFP. This gene is linked with acetate-CoA ligase activity and its

closest relation with tumorigenesis is the interaction with supernumerary limbs

(SLMB) (190), which among other functions has been associated with regulation of

mitosis as loss of Slmb correlates with misregulation of mitotic cycles (191). Either

functions of tumor suppressor, suppressor of an oncogene or activator of a tumor

suppressor are a plausible hypothesis for this candidate gene.

Sialyltransferase (ST6Gal) found upregulated 6,66 fold in the Bub3 knockdown disc


This gene has little information annotated being associated oligosaccharide metabolic

process (192). However as no link to tumorigenesis was found it could have any of the

possible functions related to an enhanced phenotype described in table III.8.

Thor found upregulated 10,08 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. Its function is

annotated as an eukaryotic initiation factor (193). This gene interacts with Eukaryotic

initiation factor 4E (eIF-4E) (181) that has been associated with mitotic spindle


76

organization (182). As such, Thor could also have a function of a suppressor of an

oncogene or activator of tumor suppressor.

Dhc62B found upregulated 3,55 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. This gene has hardly

any information although it is linked with microtubule movements through gene

similarity (194). Therefore it could have any of the possible functions related to an

enhanced phenotype described in table III.8.

CG6347 found upregulated 50,99 fold in the Bub3 knockdown disc expressing GFP

and p35 when compared with the control disc expressing only GFP. There are not

many studies regarding this gene being primarily linked with proteolysis through

electronic annotation. Therefore, it could have any of the possible functions related to

an enhanced phenotype described in table III.8.

CG6191 found upregulated 2,88 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. Its function was

defined, through electronic annotation, as cyclin-dependent protein kinase regulator

activity. It seems to be closely related with regulation of cell cycle and shows several

interactions with cell cycle regulators like Cyclin-dependent kinase 5 (CDK5) Cyclin-

dependent kinase subunit 85A (CKS85A) or altered disjunction (ald) (183). There are

several articles pointing for Cdk5 and Abl enzyme substrate 1 (Cable5), the human

homolog, as associated with tumorigenesis (195). Therefore, the expected function of

the CG6191 gene could be either suppressing an oncogene, activating a tumor

suppressor or even a tumor suppressor itself.


and p35 when compared with the control disc expressing only GFP. We found no

studies regarding this gene and as such, it could have any of the possible functions

related to an enhanced phenotype depicted in table III.8.



many studies regarding this gene. Therefore, it could have any of the possible



77



hardly any information so it could have any of the possible functions related to an




hardly any information however, the interaction with supernumerary limbs (slmb) (190),

which among other functions has been associated with regulation of mitosis as loss of

Slmb correlates with misregulation of mitotic cycles (191). Either functions of tumor

suppressor, suppressor of an oncogene or activator of a tumor suppressor are a

plausible hypothesis for this candidate gene.





Neu3 found upregulated 10,53 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. There are not many

studies regarding this gene being primarily linked with proteolysis through electronic

annotation. Its human homolog ADAM metallopeptidase domain 12 (ADAM12) on the

other hand seems to have some relation with tumorigenesis as its expression is low in

most normal tissues but is significantly increased in numerous human cancers, like

breast carcinomas (196). This lead us to hypothesize that Neu3 might have a function

as a tumor suppressor, suppressor of an oncogene or activator of a tumor suppressor.





Wun2 found upregulated 6,13 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. This gene known

functions aren’t closely related to tumorigenesis, therefore, it could have any new

unknown function related to an enhanced phenotype described in table III.8.


78

Peroxiredoxin 3 (Prx5037) found upregulated 1733,3 fold in the Bub3 knockdown disc


Among other functions this gene seems to be related with negative regulation of

apoptotic process and its human homolog peroxiredoxin 3 (Prx III) has been implicated

in the tumorigenesis of various cancers and data suggests that Prx III has a relevant

role in cell cycle regulation and could be a potential proliferation marker in for example

breast cancer (197). Therefore this gene can influence either apoptosis or cell

proliferation and as such any of the function related to an enhanced phenotype

described in table III.8 are a plausible possibility.

NimC1 found upregulated 6,93 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP and its function is

annotated as a Wnt receptor signaling pathway and as such any of the function related

to an enhanced phenotype described in table III.8 are a plausible possibility.





PNUTS found upregulated 3,18 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. The main link of this

gene to tumorigenesis is its interaction with Spliceosomal protein on the X (SPX) which

seems to have a function in the mitotic spindle organization (182). With the information

available we can only speculate that it might be a tumor suppressor, suppressor of an

oncogene or activator of a tumor suppressor.








79







Augmenter of liver regeneration (Alr) found upregulated 2,15 fold in the Bub3

knockdown disc expressing GFP and p35 when compared with the control disc

expressing only GFP. This gene is linked to tissue regeneration (198) from which we

could speculate might confer him a function of either tumor suppressor, suppressor of

an oncogene or activator of a tumor suppressor.

Aplip1 found upregulated 4,9 fold in the Bub3 knockdown disc expressing GFP and

p35 when compared with the control disc expressing only GFP. This gene known

functions aren’t closely related to tumorigenesis, therefore, it could have any new

unknown function related to an enhanced phenotype described in table III.8.


p35 when compared with the control disc expressing only GFP. There are not many

studies regarding this gene. Therefore, it could have any of the possible functions



p35 when compared with the control disc expressing only GFP. This gene seems to be

related with apoptosis therefore we speculate that its function could be either of a pro-

apoptotic gene, a suppressor of an anti-apoptotic gene or even an activator of a pro-

apoptotic gene.


80

4.6.4. GENES DOWNREGULATED IN BUB3RNAI DIAP1

BACKGROUND SCREEN

Regarding the Bub3RNAiDIAP1 background the hits from the most enhanced

phenotype were:

CG6287 found downregulated 6,97 fold in the Bub3 knockdown disc expressing

GFP and p35 when compared with the control disc expressing only GFP. The gene

CG6287 has little information available, however it has phosphoglycerate

dehydrogenase activity annotated as function inferred from gene or structural

similarities and seems to be involved in L-serine biosynthetic process. Considering

interactions, we found two particular interesting ones, Lissencephaly-1 (LIS-1) (181)

and TSC1 (181). LIS-1 interacts with dynein through a WD40 repeat. Siller et al.

showed that LIS1 and dynactin are required for centrosome separation and spindle

assembly during prophase/prometaphase and by allowing an efficient mitotic

checkpoint inactivation it contribute to timely metaphase-to-anaphase transition (199).

TSC1 in junction with TSC2 seems the regulate Cell Cycle via Changes in Cyclin

Levels, TSC1 and TSC2 also seem to regulate Cell Cycle Exit as TSC mutant cells

aren’t able to sustain a developmentally induced G1 arrest (200).

Moreover, CG6287 shows homology in vertebrates. In humans its counterpart

is phosphoglycerate dehydrogenase (PHGDH) and it has been associated with cell

cycle processes through electronic annotation and oncogenesis (201, 202). Regarding

this data we expect this gene to have a function of either a tumor suppressor itself, a

suppressor of an oncogene or even an activator of a tumor suppressor.

Topoisomerase 1 (top1) found downregulated 4,22 fold in the Bub3 knockdown

disc expressing GFP and p35 when compared with the control disc expressing only

GFP. This gene is associated with functions in cell proliferation, chromosome

segregation and condensation (203). TOP1 is a monomeric protein and it works by

producing a transient single-strand break in DNA, allowing passage to another strand,

contributing to reduce DNA supercoiling. Top 1 function suggests that this protein might

act at several critical cellular events serving as a swivel for unwinding and rewinding of

DNA helices. The human homolog, besides having similar functions to the Drosophila

counterpart, it has been implicated in small cell lung cancer (204). A closer look to

TOP1 interactions reveal an association with a protein connected with mitotic cell cycle


81

G2/M transition DNA damage checkpoint, Topoisomerase I-interacting protein

(TOPORS) (205). By adding all the evidence it is possible to assume that there could

be a synergy work, implying that BUB3 and TOP1 might interact preventing

uncontrolled cell proliferation and their depletion when apoptosis is inhibited might

cause an increased cell proliferation that could eventually lead to an overgrowth

depending on levels of downregulation. This data also suggests that any of the

functions involving an enhanced phenotype, a tumor suppressing, a suppressing of an

oncogene or activatating of a tumor suppressor are viable possibilities to explain how

this protein works regarding tumorigenesis.

CG12071 found downregulated 5,9 fold in the Bub3 knockdown disc expressing

GFP and p35 when compared with the control disc expressing only GFP. There’s only

a reported role for this protein in phagocytosis (206). Although the downregulation of

expression for this candidate is below the threshold imposed for unknown genes, its

reference in a microarray screen for Notch and Ras signaling pathway led to the choice

of including it (207). With this data we can assume that this gene could have any of the

possible functions related to an enhanced phenotype described in table III.8.

CG13101 found downregulated 12,09 fold in the Bub3 knockdown disc


This is one of the unknown genes chosen due to being above the enforced threshold of

10 fold. Consequently, there is no studies regarding this particular gene. As such,

further research would be necessary to link this gene to any tumorigenesis or cell

proliferation event. As such, this gene could have any of the possible functions related


The hits found in the screen with the additional data from previous work send us to the

direction of Bub3 as possible tumor suppressor. Co-expression of Bub3 with p35, an

apoptosis inhibitor, in wing imaginal discs leads to an overgrowth (175). As such, the

RNAi enhancer/suppressor screen gave some insight on what might be downstream of

this process. However, further studies would have to be conducted in order to find if the

hits found are really linked to the Bub3 possible function as a tumor suppressor. These

studies could include using mutants for the most relevant candidates, test the

candidate genes by themselves or dive into molecular studies like pull down assays to

understand the exact role of each candidate in the hyperplastic phenotype and its

relation with Bub3. The screen itself has some flaws, as there is the possibility of many


82

false positives in which we might get the phenotype we are looking for although having

no connection with Bub3 and might simply act by themselves or by influencing another

cell cycle or apoptosis genes. However, a screen like this is one of the best

approaches to trim data and lead into the right direction.


83

CHAPTER 5

REFERENCES


84

1. F. Hirth, Drosophila melanogaster in the study of human neurodegeneration. CNS &

neurological disorders drug targets 9, 504 (Aug, 2010). 2. X. Gao, T. P. Neufeld, D. Pan, Drosophila PTEN regulates cell growth and proliferation

through PI3K-dependent and -independent pathways. Developmental biology 221, 404 (May 15, 2000).

3. F. D. Karim et al., A screen for genes that function downstream of Ras1 during Drosophila eye development. Genetics 143, 315 (May, 1996).

4. L. A. Lee, T. L. Orr-Weaver, Regulation of cell cycles in Drosophila development: intrinsic and extrinsic cues. Annual review of genetics 37, 545 (2003).

5. D. Pan, The hippo signaling pathway in development and cancer. Developmental cell 19, 491 (Oct 19, 2010).

6. M. A. Simon, Receptor tyrosine kinases: specific outcomes from general signals. Cell 103, 13 (Sep 29, 2000).

7. M. A. Simon, D. D. Bowtell, G. S. Dodson, T. R. Laverty, G. M. Rubin, Ras1 and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the sevenless protein tyrosine kinase. Cell 67, 701 (Nov 15, 1991).

8. M. G. Voas, I. Rebay, Signal integration during development: insights from the Drosophila eye. Developmental dynamics : an official publication of the American Association of Anatomists 229, 162 (Jan, 2004).

9. C. W. Wilson, P. T. Chuang, Mechanism and evolution of cytosolic Hedgehog signal transduction. Development 137, 2079 (Jul, 2010).

10. M. B. Stark, A Benign Tumor that is Hereditary in Drosophila. Proceedings of the National Academy of Sciences of the United States of America 5, 573 (Dec, 1919).

11. D. St Johnston, The art and design of genetic screens: Drosophila melanogaster. Nature reviews. Genetics 3, 176 (Mar, 2002).

12. H. J. Bellen et al., The BDGP gene disruption project: single transposon insertions associated with 40% of Drosophila genes. Genetics 167, 761 (Jun, 2004).

13. A. H. Brand, N. Perrimon, Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401 (Jun, 1993).

14. T. Lee, L. Luo, Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron 22, 451 (Mar, 1999).

15. G. Toba et al., The gene search system. A method for efficient detection and rapid molecular identification of genes in Drosophila melanogaster. Genetics 151, 725 (Feb, 1999).

16. P. Rorth, A modular misexpression screen in Drosophila detecting tissue-specific phenotypes. Proceedings of the National Academy of Sciences of the United States of America 93, 12418 (Oct 29, 1996).

17. A. F. Chambers, A. C. Groom, I. C. MacDonald, Dissemination and growth of cancer cells in metastatic sites. Nature reviews. Cancer 2, 563 (Aug, 2002).

18. G. P. Gupta, J. Massague, Cancer metastasis: building a framework. Cell 127, 679 (Nov 17, 2006).

19. L. S. Michel et al., MAD2 haplo-insufficiency causes premature anaphase and chromosome instability in mammalian cells. Nature 409, 355 (Jan 18, 2001).

20. D. J. Baker et al., BubR1 insufficiency causes early onset of aging-associated phenotypes and infertility in mice. Nature genetics 36, 744 (Jul, 2004).

21. J. R. Babu et al., Rae1 is an essential mitotic checkpoint regulator that cooperates with Bub3 to prevent chromosome missegregation. The Journal of cell biology 160, 341 (Feb 3, 2003).

22. W. Dai et al., Slippage of mitotic arrest and enhanced tumor development in mice with BubR1 haploinsufficiency. Cancer research 64, 440 (Jan 15, 2004).


85

23. D. P. Cahill et al., Mutations of mitotic checkpoint genes in human cancers. Nature

392, 300 (Mar 19, 1998). 24. S. Hanks et al., Constitutional aneuploidy and cancer predisposition caused by biallelic

mutations in BUB1B. Nature genetics 36, 1159 (Nov, 2004). 25. Y. Mao, A. Abrieu, D. W. Cleveland, Activating and silencing the mitotic checkpoint

through CENP-E-dependent activation/inactivation of BubR1. Cell 114, 87 (Jul 11, 2003).

26. Z. Tang, R. Bharadwaj, B. Li, H. Yu, Mad2-Independent inhibition of APCCdc20 by the mitotic checkpoint protein BubR1. Developmental cell 1, 227 (Aug, 2001).

27. X. Wang et al., Correlation of defective mitotic checkpoint with aberrantly reduced expression of MAD2 protein in nasopharyngeal carcinoma cells. Carcinogenesis 21, 2293 (Dec, 2000).

28. B. A. Weaver, D. W. Cleveland, Decoding the links between mitosis, cancer, and chemotherapy: The mitotic checkpoint, adaptation, and cell death. Cancer cell 8, 7 (Jul, 2005).

29. W. O. Miles, N. J. Dyson, J. A. Walker, Modeling tumor invasion and metastasis in Drosophila. Disease models & mechanisms 4, 753 (Nov, 2011).

30. K. A. Wharton, K. M. Johansen, T. Xu, S. Artavanis-Tsakonas, Nucleotide sequence from the neurogenic locus notch implies a gene product that shares homology with proteins containing EGF-like repeats. Cell 43, 567 (Dec, 1985).

31. T. H. Morgan, The Theory of the Gene. 51, 513 (1917). 32. D. L. Lindsley, G. G. Zimm, The Genome of Drosophila Melanogaster. (Academic Press,

Inc., 1992). 33. Y. Ye, N. Lukinova, M. E. Fortini, Neurogenic phenotypes and altered Notch processing

in Drosophila Presenilin mutants. Nature 398, 525 (Apr 8, 1999). 34. L. W. Ellisen et al., TAN-1, the human homolog of the Drosophila notch gene, is broken

by chromosomal translocations in T lymphoblastic neoplasms. Cell 66, 649 (Aug 23, 1991).

35. C. Nusslein-Volhard, E. Wieschaus, Mutations affecting segment number and polarity in Drosophila. Nature 287, 795 (Oct 30, 1980).

36. H. Hahn et al., Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 85, 841 (Jun 14, 1996).

37. R. P. Sharma, V. L. Chopra, Effect of the Wingless (wg1) mutation on wing and haltere development in Drosophila melanogaster. Developmental biology 48, 461 (Feb, 1976).

38. K. W. Kinzler et al., Identification of FAP locus genes from chromosome 5q21. Science 253, 661 (Aug 9, 1991).

39. E. Ogawa et al., PEBP2/PEA2 represents a family of transcription factors homologous to the products of the Drosophila runt gene and the human AML1 gene. Proceedings of the National Academy of Sciences of the United States of America 90, 6859 (Jul 15, 1993).

40. T. Okuda, J. van Deursen, S. W. Hiebert, G. Grosveld, J. R. Downing, AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 84, 321 (Jan 26, 1996).

41. C. H. Pui, M. V. Relling, J. R. Downing, Acute lymphoblastic leukemia. The New England journal of medicine 350, 1535 (Apr 8, 2004).

42. E. B. Lewis, A gene complex controlling segmentation in Drosophila. Nature 276, 565 (Dec 7, 1978).

43. M. Djabali et al., A trithorax-like gene is interrupted by chromosome 11q23 translocations in acute leukaemias. Nature genetics 2, 113 (Oct, 1992).


86

44. Y. Gu et al., The t(4;11) chromosome translocation of human acute leukemias fuses the

ALL-1 gene, related to Drosophila trithorax, to the AF-4 gene. Cell 71, 701 (Nov 13, 1992).

45. R. K. Slany, The molecular biology of mixed lineage leukemia. Haematologica 94, 984 (Jul, 2009).

46. M. E. Martin et al., Dimerization of MLL fusion proteins immortalizes hematopoietic cells. Cancer cell 4, 197 (Sep, 2003).

47. T. Burmeister et al., The MLL recombinome of adult CD10-negative B-cell precursor acute lymphoblastic leukemia: results from the GMALL study group. Blood 113, 4011 (Apr 23, 2009).

48. C. Meyer et al., New insights to the MLL recombinome of acute leukemias. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K 23, 1490 (Aug, 2009).

49. S. A. Armstrong et al., MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia. Nature genetics 30, 41 (Jan, 2002).

50. A. V. Krivtsov, S. A. Armstrong, MLL translocations, histone modifications and leukaemia stem-cell development. Nature reviews. Cancer 7, 823 (Nov, 2007).

51. K. Geissler, O. Zach, Pathways involved in Drosophila and human cancer development: the Notch, Hedgehog, Wingless, Runt, and Trithorax pathway. Annals of hematology 91, 645 (May, 2012).

52. J. M. Peters, The anaphase promoting complex/cyclosome: a machine designed to destroy. Nature reviews. Molecular cell biology 7, 644 (Sep, 2006).

53. B. R. Thornton, D. P. Toczyski, Precise destruction: an emerging picture of the APC. Genes & development 20, 3069 (Nov 15, 2006).

54. R. Bharadwaj, H. Yu, The spindle checkpoint, aneuploidy, and cancer. Oncogene 23, 2016 (Mar 15, 2004).

55. A. Musacchio, E. D. Salmon, The spindle-assembly checkpoint in space and time. Nature reviews. Molecular cell biology 8, 379 (May, 2007).

56. C. L. Rieder, R. W. Cole, A. Khodjakov, G. Sluder, The checkpoint delaying anaphase in response to chromosome monoorientation is mediated by an inhibitory signal produced by unattached kinetochores. The Journal of cell biology 130, 941 (Aug, 1995).

57. C. L. Rieder, A. Schultz, R. Cole, G. Sluder, Anaphase onset in vertebrate somatic cells is controlled by a checkpoint that monitors sister kinetochore attachment to the spindle. The Journal of cell biology 127, 1301 (Dec, 1994).

58. H. Yu, Cdc20: a WD40 activator for a cell cycle degradation machine. Molecular cell 27, 3 (Jul 6, 2007).

59. Z. Tang et al., APC2 Cullin protein and APC11 RING protein comprise the minimal ubiquitin ligase module of the anaphase-promoting complex. Molecular biology of the cell 12, 3839 (Dec, 2001).

60. C. W. Carroll, D. O. Morgan, The Doc1 subunit is a processivity factor for the anaphase-promoting complex. Nature cell biology 4, 880 (Nov, 2002).

61. C. W. Carroll, M. Enquist-Newman, D. O. Morgan, The APC subunit Doc1 promotes recognition of the substrate destruction box. Current biology : CB 15, 11 (Jan 11, 2005).

62. C. Kraft, H. C. Vodermaier, S. Maurer-Stroh, F. Eisenhaber, J. M. Peters, The WD40 propeller domain of Cdh1 functions as a destruction box receptor for APC/C substrates. Molecular cell 18, 543 (May 27, 2005).

63. V. Sudakin, G. K. Chan, T. J. Yen, Checkpoint inhibition of the APC/C in HeLa cells is mediated by a complex of BUBR1, BUB3, CDC20, and MAD2. The Journal of cell biology 154, 925 (Sep 3, 2001).

64. H. Yu, Regulation of APC-Cdc20 by the spindle checkpoint. Current opinion in cell biology 14, 706 (Dec, 2002).


87

65. M. Li, X. Fang, Z. Wei, J. P. York, P. Zhang, Loss of spindle assembly checkpoint-

mediated inhibition of Cdc20 promotes tumorigenesis in mice. The Journal of cell biology 185, 983 (Jun 15, 2009).

66. M. Yang et al., Insights into mad2 regulation in the spindle checkpoint revealed by the crystal structure of the symmetric mad2 dimer. PLoS biology 6, e50 (Mar 4, 2008).

67. M. Mapelli, A. Musacchio, MAD contortions: conformational dimerization boosts spindle checkpoint signaling. Current opinion in structural biology 17, 716 (Dec, 2007).

68. X. Luo, H. Yu, Protein metamorphosis: the two-state behavior of Mad2. Structure 16, 1616 (Nov 12, 2008).

69. S. Kim, H. Yu, Mutual regulation between the spindle checkpoint and APC/C. Seminars in cell & developmental biology 22, 551 (Aug, 2011).

70. G. Fang, Checkpoint protein BubR1 acts synergistically with Mad2 to inhibit anaphase-promoting complex. Molecular biology of the cell 13, 755 (Mar, 2002).

71. S. Elowe et al., Uncoupling of the spindle-checkpoint and chromosome-congression functions of BubR1. Journal of cell science 123, 84 (Jan 1, 2010).

72. M. Sczaniecka et al., The spindle checkpoint functions of Mad3 and Mad2 depend on a Mad3 KEN box-mediated interaction with Cdc20-anaphase-promoting complex (APC/C). The Journal of biological chemistry 283, 23039 (Aug 22, 2008).

73. E. Choi et al., BubR1 acetylation at prometaphase is required for modulating APC/C activity and timing of mitosis. The EMBO journal 28, 2077 (Jul 22, 2009).

74. A. Kulukian, J. S. Han, D. W. Cleveland, Unattached kinetochores catalyze production of an anaphase inhibitor that requires a Mad2 template to prime Cdc20 for BubR1 binding. Developmental cell 16, 105 (Jan, 2009).

75. F. Herzog et al., Structure of the anaphase-promoting complex/cyclosome interacting with a mitotic checkpoint complex. Science 323, 1477 (Mar 13, 2009).

76. Z. Tang, H. Shu, D. Oncel, S. Chen, H. Yu, Phosphorylation of Cdc20 by Bub1 provides a catalytic mechanism for APC/C inhibition by the spindle checkpoint. Molecular cell 16, 387 (Nov 5, 2004).

77. V. P. Efimov, N. R. Morris, A screen for dynein synthetic lethals in Aspergillus nidulans identifies spindle assembly checkpoint genes and other genes involved in mitosis. Genetics 149, 101 (May, 1998).

78. S. S. Taylor, E. Ha, F. McKeon, The human homologue of Bub3 is required for kinetochore localization of Bub1 and a Mad3/Bub1-related protein kinase. The Journal of cell biology 142, 1 (Jul 13, 1998).

79. M. J. Martinez-Exposito, K. B. Kaplan, J. Copeland, P. K. Sorger, Retention of the BUB3 checkpoint protein on lagging chromosomes. Proceedings of the National Academy of Sciences of the United States of America 96, 8493 (Jul 20, 1999).

80. E. J. Neer, C. J. Schmidt, R. Nambudripad, T. F. Smith, The ancient regulatory-protein family of WD-repeat proteins. Nature 371, 297 (Sep 22, 1994).

81. P. Kalitsis, E. Earle, K. J. Fowler, K. H. Choo, Bub3 gene disruption in mice reveals essential mitotic spindle checkpoint function during early embryogenesis. Genes & development 14, 2277 (Sep 15, 2000).

82. J. Basu et al., Mutations in the essential spindle checkpoint gene bub1 cause chromosome missegregation and fail to block apoptosis in Drosophila. The Journal of cell biology 146, 13 (Jul 12, 1999).

83. J. Basu et al., Localization of the Drosophila checkpoint control protein Bub3 to the kinetochore requires Bub1 but not Zw10 or Rod. Chromosoma 107, 376 (Dec, 1998).

84. C. S. Lopes, P. Sampaio, B. Williams, M. Goldberg, C. E. Sunkel, The Drosophila Bub3 protein is required for the mitotic checkpoint and for normal accumulation of cyclins during G2 and early stages of mitosis. Journal of cell science 118, 187 (Jan 1, 2005).


88

85. Y. Niikura, H. Ogi, K. Kikuchi, K. Kitagawa, BUB3 that dissociates from BUB1 activates

caspase-independent mitotic death (CIMD). Cell death and differentiation 17, 1011 (Jun, 2010).

86. A. Bergmann, H. Steller, Apoptosis, stem cells, and tissue regeneration. Science signaling 3, re8 (2010).

87. M. D. Piper, D. Skorupa, L. Partridge, Diet, metabolism and lifespan in Drosophila. Experimental gerontology 40, 857 (Nov, 2005).

88. S. Kumar, Caspase function in programmed cell death. Cell death and differentiation 14, 32 (Jan, 2007).

89. J. Yuan, S. Shaham, S. Ledoux, H. M. Ellis, H. R. Horvitz, The C. elegans cell death gene ced-3 encodes a protein similar to mammalian interleukin-1 beta-converting enzyme. Cell 75, 641 (Nov 19, 1993).

90. P. Fuentes-Prior, G. S. Salvesen, The protein structures that shape caspase activity, specificity, activation and inhibition. The Biochemical journal 384, 201 (Dec 1, 2004).

91. J. C. Reed, Apoptosis mechanisms: implications for cancer drug discovery. Oncology (Williston Park) 18, 11 (Nov, 2004).

92. E. N. Shiozaki, Y. Shi, Caspases, IAPs and Smac/DIABLO: mechanisms from structural biology. Trends in biochemical sciences 29, 486 (Sep, 2004).

93. A. Bergmann, A. Y. Yang, M. Srivastava, Regulators of IAP function: coming to grips with the grim reaper. Current opinion in cell biology 15, 717 (Dec, 2003).

94. Y. Shi, Mechanisms of caspase activation and inhibition during apoptosis. Molecular cell 9, 459 (Mar, 2002).

95. R. Wilson et al., The DIAP1 RING finger mediates ubiquitination of Dronc and is indispensable for regulating apoptosis. Nature cell biology 4, 445 (Jun, 2002).

96. T. Tenev, A. Zachariou, R. Wilson, M. Ditzel, P. Meier, IAPs are functionally non-equivalent and regulate effector caspases through distinct mechanisms. Nature cell biology 7, 70 (Jan, 2005).

97. S. L. Wang, C. J. Hawkins, S. J. Yoo, H. A. Muller, B. A. Hay, The Drosophila caspase inhibitor DIAP1 is essential for cell survival and is negatively regulated by HID. Cell 98, 453 (Aug 20, 1999).

98. L. Goyal, K. McCall, J. Agapite, E. Hartwieg, H. Steller, Induction of apoptosis by Drosophila reaper, hid and grim through inhibition of IAP function. The EMBO journal 19, 589 (Feb 15, 2000).

99. S. Lisi, I. Mazzon, K. White, Diverse domains of THREAD/DIAP1 are required to inhibit apoptosis induced by REAPER and HID in Drosophila. Genetics 154, 669 (Feb, 2000).

100. X. Li, J. Wang, Y. Shi, Structural mechanisms of DIAP1 auto-inhibition and DIAP1-mediated inhibition of drICE. Nature communications 2, 408 (2011).

101. D. Vucic, W. J. Kaiser, A. J. Harvey, L. K. Miller, Inhibition of reaper-induced apoptosis by interaction with inhibitor of apoptosis proteins (IAPs). Proceedings of the National Academy of Sciences of the United States of America 94, 10183 (Sep 16, 1997).

102. N. Yan et al., Structural, biochemical, and functional analyses of CED-9 recognition by the proapoptotic proteins EGL-1 and CED-4. Mol Cell 15, 999 (Sep 24, 2004).

103. K. White et al., Genetic control of programmed cell death in Drosophila. Science 264, 677 (Apr 29, 1994).

104. M. E. Grether, J. M. Abrams, J. Agapite, K. White, H. Steller, The head involution defective gene of Drosophila melanogaster functions in programmed cell death. Genes & development 9, 1694 (Jul 15, 1995).

105. P. Chen, W. Nordstrom, B. Gish, J. M. Abrams, grim, a novel cell death gene in Drosophila. Genes & development 10, 1773 (Jul 15, 1996).

106. S. M. Srinivasula et al., sickle, a novel Drosophila death gene in the reaper/hid/grim region, encodes an IAP-inhibitory protein. Current biology : CB 12, 125 (Jan 22, 2002).


89

107. D. Vucic, W. J. Kaiser, L. K. Miller, Inhibitor of apoptosis proteins physically interact

with and block apoptosis induced by Drosophila proteins HID and GRIM. Molecular and cellular biology 18, 3300 (Jun, 1998).

108. C. L. Holley, M. R. Olson, D. A. Colon-Ramos, S. Kornbluth, Reaper eliminates IAP proteins through stimulated IAP degradation and generalized translational inhibition. Nature cell biology 4, 439 (Jun, 2002).

109. J. Chai et al., Molecular mechanism of Reaper-Grim-Hid-mediated suppression of DIAP1-dependent Dronc ubiquitination. Nature structural biology 10, 892 (Nov, 2003).

110. A. Zachariou et al., IAP-antagonists exhibit non-redundant modes of action through differential DIAP1 binding. The EMBO journal 22, 6642 (Dec 15, 2003).

111. H. D. Ryoo, A. Bergmann, H. Gonen, A. Ciechanover, H. Steller, Regulation of Drosophila IAP1 degradation and apoptosis by reaper and ubcD1. Nature cell biology 4, 432 (Jun, 2002).

112. S. J. Yoo et al., Hid, Rpr and Grim negatively regulate DIAP1 levels through distinct mechanisms. Nature cell biology 4, 416 (Jun, 2002).

113. J. V. McCarthy, V. M. Dixit, Apoptosis induced by Drosophila reaper and grim in a human system. Attenuation by inhibitor of apoptosis proteins (cIAPs). The Journal of biological chemistry 273, 24009 (Sep 11, 1998).

114. W. N. Haining, C. Carboy-Newcomb, C. L. Wei, H. Steller, The proapoptotic function of Drosophila Hid is conserved in mammalian cells. Proceedings of the National Academy of Sciences of the United States of America 96, 4936 (Apr 27, 1999).

115. M. H. Brodsky et al., Drosophila melanogaster MNK/Chk2 and p53 regulate multiple DNA repair and apoptotic pathways following DNA damage. Molecular and cellular biology 24, 1219 (Feb, 2004).

116. N. Sogame, M. Kim, J. M. Abrams, Drosophila p53 preserves genomic stability by regulating cell death. Proceedings of the National Academy of Sciences of the United States of America 100, 4696 (Apr 15, 2003).

117. D. G. McEwen, M. Peifer, Puckered, a Drosophila MAPK phosphatase, ensures cell viability by antagonizing JNK-induced apoptosis. Development 132, 3935 (Sep, 2005).

118. M. H. Brodsky et al., Drosophila p53 binds a damage response element at the reaper locus. Cell 101, 103 (Mar 31, 2000).

119. E. Moreno, K. Basler, G. Morata, Cells compete for decapentaplegic survival factor to prevent apoptosis in Drosophila wing development. Nature 416, 755 (Apr 18, 2002).

120. M. Ollmann et al., Drosophila p53 is a structural and functional homolog of the tumor suppressor p53. Cell 101, 91 (Mar 31, 2000).

121. S. Y. Vernooy et al., Drosophila Bruce can potently suppress Rpr- and Grim-dependent but not Hid-dependent cell death. Current biology : CB 12, 1164 (Jul 9, 2002).

122. E. Arama, J. Agapite, H. Steller, Caspase activity and a specific cytochrome C are required for sperm differentiation in Drosophila. Developmental cell 4, 687 (May, 2003).

123. B. A. Hay, D. A. Wassarman, G. M. Rubin, Drosophila homologs of baculovirus inhibitor of apoptosis proteins function to block cell death. Cell 83, 1253 (Dec 29, 1995).

124. G. Jones, D. Jones, L. Zhou, H. Steller, Y. Chu, Deterin, a new inhibitor of apoptosis from Drosophila melanogaster. The Journal of biological chemistry 275, 22157 (Jul 21, 2000).

125. L. Zhou, Z. Song, J. Tittel, H. Steller, HAC-1, a Drosophila homolog of APAF-1 and CED-4 functions in developmental and radiation-induced apoptosis. Molecular cell 4, 745 (Nov, 1999).

126. C. Jiang, A. F. Lamblin, H. Steller, C. S. Thummel, A steroid-triggered transcriptional hierarchy controls salivary gland cell death during Drosophila metamorphosis. Molecular cell 5, 445 (Mar, 2000).


90

127. L. Dorstyn, P. A. Colussi, L. M. Quinn, H. Richardson, S. Kumar, DRONC, an ecdysone-

inducible Drosophila caspase. Proceedings of the National Academy of Sciences of the United States of America 96, 4307 (Apr 13, 1999).

128. S. Kornbluth, K. White, Apoptosis in Drosophila: neither fish nor fowl (nor man, nor worm). Journal of cell science 118, 1779 (May 1, 2005).

129. E. H. Baehrecke, How death shapes life during development. Nature reviews. Molecular cell biology 3, 779 (Oct, 2002).

130. Y. H. Song, Drosophila melanogaster: a model for the study of DNA damage checkpoint response. Molecules and cells 19, 167 (Apr 30, 2005).

131. I. Lohmann, N. McGinnis, M. Bodmer, W. McGinnis, The Drosophila Hox gene deformed sculpts head morphology via direct regulation of the apoptosis activator reaper. Cell 110, 457 (Aug 23, 2002).

132. T. Wolff, D. F. Ready, Cell death in normal and rough eye mutants of Drosophila. Development 113, 825 (Nov, 1991).

133. C. Manjon, E. Sanchez-Herrero, M. Suzanne, Sharp boundaries of Dpp signalling trigger local cell death required for Drosophila leg morphogenesis. Nature cell biology 9, 57 (Jan, 2007).

134. M. Milan, S. Campuzano, A. Garcia-Bellido, Developmental parameters of cell death in the wing disc of Drosophila. Proceedings of the National Academy of Sciences of the United States of America 94, 5691 (May 27, 1997).

135. A. Perez-Garijo, F. A. Martin, G. Morata, Caspase inhibition during apoptosis causes abnormal signalling and developmental aberrations in Drosophila. Development 131, 5591 (Nov, 2004).

136. J. L. Haynie, P. J. Bryant, The effects of X-rays on the proliferation dynamics of Cells in the imagina wing disc of Drosophila melanogaster. Wilhelm Roux’s Archives 183, 85 (1977).

137. J. Colombani, C. Polesello, F. Josue, N. Tapon, Dmp53 activates the Hippo pathway to promote cell death in response to DNA damage. Current biology : CB 16, 1453 (Jul 25, 2006).

138. L. J. Saucedo, B. A. Edgar, Filling out the Hippo pathway. Nature reviews. Molecular cell biology 8, 613 (Aug, 2007).

139. B. J. Graves, G. Schubiger, Cell cycle changes during growth and differentiation of imaginal leg discs in Drosophila melanogaster. Developmental biology 93, 104 (Sep, 1982).

140. F. A. Martin, A. Perez-Garijo, G. Morata, Apoptosis in Drosophila: compensatory proliferation and undead cells. The International journal of developmental biology 53, 1341 (2009).

141. Y. Xia, M. Karin, The control of cell motility and epithelial morphogenesis by Jun kinases. Trends in cell biology 14, 94 (Feb, 2004).

142. E. Martin-Blanco et al., puckered encodes a phosphatase that mediates a feedback loop regulating JNK activity during dorsal closure in Drosophila. Genes & development 12, 557 (Feb 15, 1998).

143. K. Mills, T. Daish, S. Kumar, The function of the Drosophila caspase DRONC in cell death and development. Cell Cycle 4, 744 (Jun, 2005).

144. B. A. Hay, T. Wolff, G. M. Rubin, Expression of baculovirus P35 prevents cell death in Drosophila. Development 120, 2121 (Aug, 1994).

145. B. A. Callus, D. L. Vaux, Caspase inhibitors: viral, cellular and chemical. Cell death and differentiation 14, 73 (Jan, 2007).

146. T. Tabata, Y. Takei, Morphogens, their identification and regulation. Development 131, 703 (Feb, 2004).


91

147. M. Gardiner-Garden, T. G. Littlejohn, A comparison of microarray databases. Briefings

in bioinformatics 2, 143 (May, 2001). 148. E. Georgii, L. Richter, U. Ruckert, S. Kramer, Analyzing microarray data using

quantitative association rules. Bioinformatics 21 Suppl 2, ii123 (Sep 1, 2005). 149. M. Schena, D. Shalon, R. W. Davis, P. O. Brown, Quantitative monitoring of gene

expression patterns with a complementary DNA microarray. Science 270, 467 (Oct 20, 1995).

150. C. A. Afshari, Perspective: microarray technology, seeing more than spots. Endocrinology 143, 1983 (Jun, 2002).

151. W. P. Kuo et al., Microarrays and clinical dentistry. J Am Dent Assoc 134, 456 (Apr, 2003).

152. M. Schena et al., Microarrays: biotechnology's discovery platform for functional genomics. Trends in biotechnology 16, 301 (Jul, 1998).

153. S. Michiels, S. Koscielny, C. Hill, Prediction of cancer outcome with microarrays: a multiple random validation strategy. Lancet 365, 488 (Feb 5-11, 2005).

154. P. O. Brown, D. Botstein, Exploring the new world of the genome with DNA microarrays. Nature genetics 21, 33 (Jan, 1999).

155. S. V. Tillib, A. D. Mirzabekov, Advances in the analysis of DNA sequence variations using oligonucleotide microchip technology. Current opinion in biotechnology 12, 53 (Feb, 2001).

156. M. Beier, J. D. Hoheisel, Production by quantitative photolithographic synthesis of individually quality checked DNA microarrays. Nucleic acids research 28, E11 (Feb 15, 2000).

157. W. Ji, W. Zhou, K. Gregg, K. Lindpaintner, S. Davis, A method for gene expression analysis by oligonucleotide arrays from minute biological materials. Analytical biochemistry 331, 329 (Aug 15, 2004).

158. R. Lucito et al., Representational oligonucleotide microarray analysis: a high-resolution method to detect genome copy number variation. Genome research 13, 2291 (Oct, 2003).

159. L. Guo et al., Gene expression profiling of drug-resistant small cell lung cancer cells by combining microRNA and cDNA expression analysis. Eur J Cancer 46, 1692 (Jun, 2010).

160. W. Huang da, B. T. Sherman, R. A. Lempicki, Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic acids research 37, 1 (Jan, 2009).

161. C. Napoli, C. Lemieux, R. Jorgensen, Introduction of a Chimeric Chalcone Synthase Gene into Petunia Results in Reversible Co-Suppression of Homologous Genes in trans. The Plant cell 2, 279 (Apr, 1990).

162. A. Fire et al., Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806 (Feb 19, 1998).

163. S. M. Hammond, E. Bernstein, D. Beach, G. J. Hannon, An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells. Nature 404, 293 (Mar 16, 2000).

164. J. R. Kennerdell, R. W. Carthew, Use of dsRNA-mediated genetic interference to demonstrate that frizzled and frizzled 2 act in the wingless pathway. Cell 95, 1017 (Dec 23, 1998).

165. M. Ramet, P. Manfruelli, A. Pearson, B. Mathey-Prevot, R. A. Ezekowitz, Functional genomic analysis of phagocytosis and identification of a Drosophila receptor for E. coli. Nature 416, 644 (Apr 11, 2002).

166. S. Cherry et al., Genome-wide RNAi screen reveals a specific sensitivity of IRES-containing RNA viruses to host translation inhibition. Genes & development 19, 445 (Feb 15, 2005).


92

167. R. DasGupta, A. Kaykas, R. T. Moon, N. Perrimon, Functional genomic analysis of the

Wnt-wingless signaling pathway. Science 308, 826 (May 6, 2005). 168. G. Dietzl et al., A genome-wide transgenic RNAi library for conditional gene

inactivation in Drosophila. Nature 448, 151 (Jul 12, 2007). 169. E. Bernstein, A. A. Caudy, S. M. Hammond, G. J. Hannon, Role for a bidentate

ribonuclease in the initiation step of RNA interference. Nature 409, 363 (Jan 18, 2001). 170. A. Nykanen, B. Haley, P. D. Zamore, ATP requirements and small interfering RNA

structure in the RNA interference pathway. Cell 107, 309 (Nov 2, 2001). 171. R. H. Nicholson, A. W. Nicholson, Molecular characterization of a mouse cDNA

encoding Dicer, a ribonuclease III ortholog involved in RNA interference. Mammalian genome : official journal of the International Mammalian Genome Society 13, 67 (Feb, 2002).

172. P. J. Paddison, G. J. Hannon, RNA interference: the new somatic cell genetics? Cancer cell 2, 17 (Jul, 2002).

173. V. Hartenstein, Atlas of Drosophila development. (Cold Spring Harbour Laboratory Press, 1993).

174. M. Orme, P. Meier, Inhibitor of apoptosis proteins in Drosophila: gatekeepers of death. Apoptosis : an international journal on programmed cell death 14, 950 (Aug, 2009).

175. S. Morais da Silva, personal communication. 176. M. Calleja, E. Moreno, S. Pelaz, G. Morata, Visualization of gene expression in living

adult Drosophila. Science 274, 252 (Oct 11, 1996). 177. B. Winnepenninckx, Y. Van de Peer, T. Backeljau, R. De Wachter, CARD: a drawing tool

for RNA secondary structure models. BioTechniques 18, 1060 (Jun, 1995). 178. W. Huang da, B. T. Sherman, R. A. Lempicki, Systematic and integrative analysis of

large gene lists using DAVID bioinformatics resources. Nature protocols 4, 44 (2009). 179. D. A. Hosack, G. Dennis, Jr., B. T. Sherman, H. C. Lane, R. A. Lempicki, Identifying

biological themes within lists of genes with EASE. Genome biology 4, R70 (2003). 180. L. Ciapponi, G. Cenci, Telomere capping and cellular checkpoints: clues from fruit flies.

Cytogenetic and genome research 122, 365 (2008). 181. L. Giot et al., A protein interaction map of Drosophila melanogaster. Science 302, 1727

(Dec 5, 2003). 182. G. Goshima et al., Genes required for mitotic spindle assembly in Drosophila S2 cells.

Science 316, 417 (Apr 20, 2007). 183. C. A. Stanyon et al., A Drosophila protein-interaction map centered on cell-cycle

regulators. Genome biology 5, R96 (2004). 184. J. Dobbelaere et al., A genome-wide RNAi screen to dissect centriole duplication and

centrosome maturation in Drosophila. PLoS biology 6, e224 (Sep 16, 2008). 185. A. Lunardi et al., A genome-scale protein interaction profile of Drosophila p53

uncovers additional nodes of the human p53 network. Proceedings of the National Academy of Sciences of the United States of America 107, 6322 (Apr 6, 2010).

186. J. Mata, S. Curado, A. Ephrussi, P. Rorth, Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating string/CDC25 proteolysis. Cell 101, 511 (May 26, 2000).

187. M. Nie, Y. Xie, J. A. Loo, A. J. Courey, Genetic and proteomic evidence for roles of Drosophila SUMO in cell cycle control, Ras signaling, and early pattern formation. PloS one 4, e5905 (2009).

188. S. J. Wall, E. Werner, Z. Werb, Y. A. DeClerck, Discoidin domain receptor 2 mediates tumor cell cycle arrest induced by fibrillar collagen. The Journal of biological chemistry 280, 40187 (Dec 2, 2005).

189. E. Kurucz et al., Nimrod, a putative phagocytosis receptor with EGF repeats in Drosophila plasmatocytes. Current biology : CB 17, 649 (Apr 3, 2007).


93

190. E. Formstecher et al., Protein interaction mapping: a Drosophila case study. Genome

research 15, 376 (Mar, 2005). 191. M. Muzzopappa, P. Wappner, Multiple roles of the F-box protein Slimb in Drosophila

egg chamber development. Development 132, 2561 (Jun, 2005). 192. K. Koles, K. D. Irvine, V. M. Panin, Functional characterization of Drosophila

sialyltransferase. The Journal of biological chemistry 279, 4346 (Feb 6, 2004). 193. M. Miron et al., The translational inhibitor 4E-BP is an effector of PI(3)K/Akt signalling

and cell growth in Drosophila. Nature cell biology 3, 596 (Jun, 2001). 194. L. S. Goldstein, S. Gunawardena, Flying through the drosophila cytoskeletal genome.

The Journal of cell biology 150, F63 (Jul 24, 2000). 195. L. R. Zukerberg et al., Loss of cables, a cyclin-dependent kinase regulatory protein, is

associated with the development of endometrial hyperplasia and endometrial cancer. Cancer research 64, 202 (Jan 1, 2004).

196. C. Frohlich et al., ADAM12 produced by tumor cells rather than stromal cells accelerates breast tumor progression. Molecular cancer research : MCR 9, 1449 (Nov, 2011).

197. P. J. Chua et al., Silencing the Peroxiredoxin III gene inhibits cell proliferation in breast cancer. International journal of oncology 36, 359 (Feb, 2010).

198. K. D. McClure, A. Sustar, G. Schubiger, Three genes control the timing, the site and the size of blastema formation in Drosophila. Developmental biology 319, 68 (Jul 1, 2008).

199. K. H. Siller, M. Serr, R. Steward, T. S. Hays, C. Q. Doe, Live imaging of Drosophila brain neuroblasts reveals a role for Lis1/dynactin in spindle assembly and mitotic checkpoint control. Molecular biology of the cell 16, 5127 (Nov, 2005).

200. N. Tapon, N. Ito, B. J. Dickson, J. E. Treisman, I. K. Hariharan, The Drosophila tuberous sclerosis complex gene homologs restrict cell growth and cell proliferation. Cell 105, 345 (May 4, 2001).

201. E. Mullarky, K. R. Mattaini, M. G. Vander Heiden, L. C. Cantley, J. W. Locasale, PHGDH amplification and altered glucose metabolism in human melanoma. Pigment cell & melanoma research 24, 1112 (Dec, 2011).

202. J. W. Locasale et al., Phosphoglycerate dehydrogenase diverts glycolytic flux and contributes to oncogenesis. Nature genetics 43, 869 (Sep, 2011).

203. C. X. Zhang, A. D. Chen, N. J. Gettel, T. S. Hsieh, Essential functions of DNA topoisomerase I in Drosophila melanogaster. Developmental biology 222, 27 (Jun 1, 2000).

204. M. Sereno et al., ERCC1 and topoisomerase I expression in small cell lung cancer: prognostic and predictive implications. International journal of oncology 40, 2104 (Jun, 2012).

205. S. Kondo, N. Perrimon, A genome-wide RNAi screen identifies core components of the G(2)-M DNA damage checkpoint. Science signaling 4, rs1 (2011).

206. S. L. Stroschein-Stevenson, E. Foley, P. H. O'Farrell, A. D. Johnson, Identification of Drosophila gene products required for phagocytosis of Candida albicans. PLoS biology 4, e4 (Jan, 2006).

207. R. Artero, E. E. Furlong, K. Beckett, M. P. Scott, M. Baylies, Notch and Ras signaling pathway effector genes expressed in fusion competent and founder cells during Drosophila myogenesis. Development 130, 6257 (Dec, 2003).

spindle assembly checkpoint and chromosomal instability

Documents