Embed Size (px)
Citation preview
UNIVERSIDADE ESTADUAL DE CAMPINAS
Instituto de Biologia
RAPHAEL RICON DE OLIVEIRA
Regulação epigenética do neogene Qua-Quine Starch durante o desenvolvimento de
Arabidopsis thaliana e o impacto no metabolismo de amido
CAMPINAS
2015
RAPHAEL RICON DE OLIVEIRA
Regulação epigenética do neogene Qua-Quine Starch durante o desenvolvimento de
Arabidopsis thaliana e o impacto no metabolismo de amido
Tese apresentada ao Instituto de Biologia
da Universidade Estadual de Campinas
como parte dos requisitos exigidos para
obtenção do Título de Doutor em Genética
e Biologia Molecular na área de Genética
Vegetal e Melhoramento.
Orientador: PROF. DR. MICHEL GEORGES ALBERT VINCENTZ
ESTE EXEMPLAR CORRESPONDE À VERSÃO
FINAL DA TESE DEFENDIDA PELO ALUNO
RAPHAEL RICON DE OLIVEIRA E ORIENTADA
PELO PROF. DR. MICHEL GEORGES ALBERT
VINCENTZ.
CAMPINAS
2015
COMISSÃO EXAMINADORA
Membros Titulares:
Prof. Dr. Michel Georges Albert Vincentz (Orientador)
Prof. Dr. Juan Armando Casas-Mollano
Prof. Dr. Marcio Alves Ferreira
Profa. Dra. Camila Caldana
Prof. Dr. Jörg Kobarg
Membros Suplentes:
Prof. Dr. Paulo Cavalcanti Gomes Ferreira
Prof. Dr. Fábio Tebaldi Silveira Nogueira
Profa. Dra. Katlin Brauer Massirer
A Ata da Defesa assinada pelos membros da Comissão Examinadora consta no processo
de vida acadêmica do aluno.
DEDICATÓRIA
A todos aqueles que não tiveram a oportunidade concreta de estudar, dedico.
AGRADECIMENTOS
Hoje tenho a convicção de que não somos um ponto discreto, mas parte de uma
linha contínua e por vezes emaranhada. Essa tese não seria possível sem essa percepção, sem
o pensamento focado nas pessoas que amo e no ideal de mundo melhor. Algumas pessoas
entendem essa mensagem e caminham ao seu lado. Com essas pessoas gostaria de
compartilhar essa conquista e agradecer pela companhia. Começo agradecendo ao professor
Michel pela apresentação do empolgante tema e pela orientação. Agradeço à FAPESP, assim
como à população do estado de São Paulo, pelo financiamento do projeto. Aos membros da
banca e também aos colegas de laboratório pela disponibilidade, discussões e conhecimentos
compartilhados. A toda estrutura da Universidade Estadual de Campinas, funcionários e
professores. Em especial, agradeço minha esposa Thaís, pela paciência, companheirismo,
conselhos e carinho em todos os momentos durante esses últimos anos. Aos meus avós pelas
posturas corajosas e exemplos de garra perante as adversidades impostas pela vida. A minha
mãe, uma fonte de ideias e bondade que irá sempre me inspirar. Ao meu pai, pelo exemplo de
fortaleza e tranquilidade. Aos meus irmãos e amigos (talvez não haja distinção entre essas
duas palavras e incluo os cachorros Churrasco e Pinga nesse grupo) pelo prazer da
convivência e boas risadas para sempre guardadas no coração. Termino agradecendo a todos
os envolvidos de alguma forma nessa aventura e reforço o que não pode ser descrito, numa
tentativa tosca de descrever algo que apenas pode ser sentido, minha gratidão.
RESUMO
Metilação do quinto carbono da citosina do DNA é uma marca epigenética que pode afetar a
expressão gênica. A regulação de perfis de metilação é importante para o desenvolvimento
normal das plantas sendo crítica para o silenciamento de transposons, imprinting,
gametogênese e o desenvolvimento inicial do embrião. Entretanto, como metilações de
sequências regulatórias em cis afetam a interação entre o DNA e fatores em trans associados
ao desenvolvimento para modular a expressão gênica é ainda pouco compreendida. Qua-
Quine Starch (QQS) é um gene órfão de Arabidopsis thaliana que existe sob diversas formas
epialélicas estavelmente herdadas e com níveis de expressão correlacionados inversamente
com os níveis de metilação em seu promotor e 5’UTR. Por meio de análises da expressão do
gene marcador GUS sob o controle da sequência do promotor e da região 5’UTR de QQS em
linhagens transgênicas de Arabidopsis, inferimos o potencial de expressão de QQS nos vários
órgãos e várias fases do desenvolvimento. A atividade GUS foi detectada em meristemas,
folhas jovens de roseta e no pólen onde uma reprogramação epigenética deste gene foi
descrita. Epialelos contrastantes de QQS apresentaram expressão diferencial ao longo do
desenvolvimento. O epialelo QQS metilado (QQSme
) possui diferenças marcantes de
expressão entre folhas, tecidos da inflorescência e fruto, enquanto o epialelo demetilado
Col*3-2 apresenta uma expressão mais homogênea entre esses vários órgãos. Estes resultados
indicam que o grau de metilação das sequências regulatórias de QQS interage com fatores
associados ao desenvolvimento para estabelecer o perfil de expressão deste gene. A expressão
de QQSme
aumenta durante o processo de envelhecimento das folhas e foi correlacionada a
uma demetilação no lócus. Levantamos a hipótese de que o perfil de metilação de QQSme
poderia sofrer uma reprogramação durante o desenvolvimento das folhas a partir do
meristema vegetativo. Tal reprogramação poderia ocorrer de maneira ativa, mediada pela
DNA glicosilase ROS1, ou passiva, como consequência da ausência ou falha do mecanismo
de manutenção do padrão de metilação durante ciclos de replicação celular. De acordo com
isso, mostramos que expressão de QQSme
é baixa em amostras enriquecidas com células
meristemáticas e aumenta gradativamente com a idade das folhas e uma redução dos níveis de
metilação de QQSme
foi observada em folhas mais velhas. Em alguns órgãos foram
encontradas variações nos níveis de expressão de QQS não correlacionadas a variações nos
níveis de metilação. Assim, os resultados também mostram que além da metilação outros
fatores trans associados ao desenvolvimento são necessários para modular a expressão de
QQS. O processo de demetilação de QQSme
em folhas velhas não é totalmente dependente de
ROS1. Porém, os níveis e padrão de expressão de QQS em mutantes ros1 são diferentes de
plantas wild-type e uma análise da expressão de QQS numa população F2 resultante da
autofecundação de um híbrido entre wild-type e o mutante ros1-4 revelaram que ROS1
aparentemente regula a expressão global de QQS. Dessa forma, as metilações do DNA
parecem atuar como sinalizadores de sítios de ligação para fatores trans e ROS1 pode ser
importante para regular e/ou definir estados epialélicos de QQS. Além disso, durante o
desenvolvimento reprodutivo a expressão de QQS é detectada no pólen e correlaciona com a
demetilação ativa de seu lócus mediada por DEMETER reforçando o fato de que QQS é um
alvo de DNA glicosilases. Durante o desenvolvimento vegetativo QQS é descrito como um
regulador negativo do metabolismo de amido. De acordo com isso, verificamos que plantas
carregando epialelos QQS contrastantes acumulam diferentes quantidades de amido.
ABSTRACT
DNA methylation of the fifth carbon of cytosine is an epigenetic mark that may affect gene
expression. The regulation of methylation profiles is important for normal plant development
being critical for transposon silencing, imprinting, gametogenesis and early embryo
development. However, how methylation of cis regulatory sequences affect the interaction
between DNA and developmental associated trans factors to modulate gene expression is
poorly understood. Qua-Quine Starch (QQS) is an orphan gene of Arabidopsis thaliana which
exhibits several epiallelic forms stably-inherited and their expression levels correlate
inversely with the methylation levels at the promoter and 5’UTR. Through the expression
analyses of the GUS reporter gene under the control of a QQS promoter sequence and QQS
5’UTR in Arabidopsis transgenic lines, we inferred the potential of QQS expression in various
organs and at several developmental stages. The GUS activity was detected in meristems,
young rosette leaves and pollen where an epigenetic reprogramming of this gene has been
described. Contrasting QQS epialleles presented different expression patterns during
development. The methylated QQS epiallele (QQSme
) showed more pronounced expression
differences between leaves, inflorescence tissues and fruit, whereas the demethylated Col*3-2
epiallele presented a more homogeneous expression between this various organs. These
results indicate that the methylation level of QQS cis regulatory sequence interacts with
developmental-related factors to establish the expression profile of this gene. QQSme
expression increases during aging of leaves and was correlated with demethylation. We
hypothesized that the QQSme
methylation profile could be reprogrammed during leaf
development starting from the shoot apical meristem. Such process could be controlled by the
DNA glycosilase ROS1, or be passive, as a consequence of the absence or failure in the
maintenance mechanisms of the methylation pattern during cell replication. Accordingly, we
showed that the QQSme
expression is low in samples enriched in meristematic cells and
increases progressively with leaf aging and reduction of QQSme
methylation levels was
observed in older leaves. In some organs, QQS expression variations not correlated with
variation in the methylation levels were found. Thus, the results indicate that in addition to
DNA methylation developmental associated trans factors are also necessary to modulate QQS
expression. The demethylation process of QQSme
in old leaves is not totally dependent upon
ROS1. However, the levels and pattern of QQS expression in ros1 mutants are different from
those in wild-type plants and the analysis of a segregating F2 population coming from selfing
of a hybrid between the wild-type and ros1-4 mutant revealed that ROS1 regulates the global
expression level of QQS. Therefore, DNA methylation seems to act as traffic lights of binding
sites to the trans factors and ROS1 may be important to regulate and/or define QQS epiallelic
states. Furthermore, during the reproductive development the expression of QQS is detected
in the pollen and correlates with the active demethylation of its locus mediated by
DEMETER, which reinforces the fact that QQS is a target of DNA glycosilases. During
vegetative development QQS has been described as a negative regulator of starch metabolism.
We verified that plants carrying contrasting QQS epialleles accumulate different starch
quantities.
LISTA DE ABREVIATURAS E SIGLAS
5’UTR: região não traduzida da extremidade 5’ do RNA mensageiro
CTAB: cetyltrimethylammonium bromide
DEPC: diethylpyrocarbonate
epiRILS: Epigenetic recombinant inbred lines
mC: methylated cytosine
McrBC: endonuclease capaz de clivar DNA contendo citosina metilada
PcG: Polycomb Group
Pol: RNA polimerase
QQS:GUS: regiões promotora e 5’UTR de QQS fusionadas ao gene repórter GUS
QQSme: epialelo de QQS contendo altos níveis demetilação de DNA
qRT-PCR: quantitative real-time polymerase chain reaction
RdDM: RNA-directed DNA methylation
RNAi: RNA de interferência
ros1-4: mutante para DNA glicosilase ROS1
SAM: Shoot Apical Meristem
siRNA: small interfering RNA
TE: Transposable Elements
TF: Transcription Factors
TFBS: Transcription Factors Binding Sites
TrxG: Trithorax Group
wt: wild-type
SUMÁRIO
INTRODUÇÃO ...................................................................................................................... 13
O que é epigenética? .......................................................................................................... 13
Mecanismos de regulação epigenética .............................................................................. 16
Importância das metilações do DNA para o desenvolvimento e adaptação vegetal ......... 27
QQS, um gene órfão e provocador .................................................................................... 30
OBJETIVO ............................................................................................................................. 33
RESULTADOS E DISCUSSÃO ........................................................................................... 34
Manuscrito “Developmental expression pattern of QQS is influenced by DNA
methylation: involvement of ROS1 and impact on starch metabolism”. .......................... 34
Material suplementar do manuscrito ................................................................................. 66
CONCLUSÃO ......................................................................................................................... 81
PERSPECTIVAS .................................................................................................................... 82
REFERÊNCIAS BIBLIOGRÁFICAS ................................................................................. 84
ANEXOS ................................................................................................................................. 93
Anexo I – Primeira página do artigo publicado “An efficient method for simultaneous
extraction of high-quality RNA and DNA from various plant tissues” ............................ 93
Anexo II – Declaração de Bioética e Biossegurança......................................................... 94
Anexo III – Declaração referente aos direitos autorais ..................................................... 95
13
INTRODUÇÃO
O que é epigenética?
No início do século XX as disciplinas biologia do desenvolvimento e genética
eram estudadas separadamente. A grande pergunta vigente na época entre os pesquisadores
dessas áreas era como uma célula (zigoto) contendo o mesmo material genético poderia dar
origem a diferentes tipos celulares durante o desenvolvimento para formar o organismo. A
solução conceitual desta questão foi dada por Conrad Hal Waddington que imaginou a
existência de um mecanismo além ou acima dos genes controlando a leitura da sequência
codificada no DNA (Waddington, 1942). Tal mecanismo originaria produtos gênicos e
funções variáveis entre as células determinando o desenvolvimento e o fenótipo. Assim, ele
foi o primeiro a associar as duas disciplinas em questão cunhando o termo epigenética, que
significa literalmente acima (epi) da genética, se referindo à clássica teoria da epigenesis
proposta por Aristóteles para explicar o desenvolvimento do organismo a partir de um
embrião indiferenciado (Aristotle, 1979).
Originalmente, o termo epigenética foi cunhado por Waddington para se referir ao
“estudo dos mecanismos causais pelos quais os genes que compõem o genótipo trazem à tona
efeitos fenotípicos” (Waddington, 1942). Essa definição traz a essência do que é epigenética,
a conexão entre genótipo e fenótipo como um fenômeno que modifica o resultado final de um
lócus ou cromossomo sem mudar a sequência do DNA subjacente (Goldberg et al., 2007).
Assim, Waddington previu que a diferenciação celular seria um fenômeno epigenético
amplamente governado pelo contexto, tais como, informações fornecidas pelos genes,
comunicações entre células e sinais do ambiente (Goldberg et al., 2007; Kinoshita e Jacobsen,
2012), e usou a metáfora das “paisagens epigenéticas” para ilustrar essa ideia (Waddington,
1957). Tais paisagens montanhosas representam o percurso de uma célula ao longo do
desenvolvimento o que leva a diferentes destinos celulares (Fig. 1A). A superfície dessas
paisagens seria distorcida por fatores epigenéticos, além dos genes e suas interações,
formando vales e morros atrativos para a célula e resultando em estados fenotípicos, um
processo que foi chamado de canalização (Fig. 1B).
14
Figura 1. Paisagem epigenética descrita por Waddington (1957). A) Paisagens
montanhosas representando os possíveis percursos de uma célula indiferenciada (esfera na
figura) durante o desenvolvimento o que leva a diferentes destinos celulares ou diferenciação.
B) Canalização, processo no qual a superfície da paisagem é distorcida por fatores
epigenéticos, genéticos e suas interações (pontos e linhas interligadas na figura) formando os
vales e morros atrativos para a célula e representando os estados fenotípicos. Figuras
modificadas de Waddington, 1957.
Com o avanço das pesquisas surgiram outras definições para o termo epigenética,
sendo este um motivo de grande debate dependendo da área em que é empregado (Hollyday,
2006). Pesquisadores das áreas de biologia do desenvolvimento e câncer costumam dar ênfase
ao critério da herdabilidade mitótica, por considerarem a fidelidade do fenótipo celular após a
divisão uma questão crítica para diferenciação de tecidos (Sweatt et al., 2013). Assim,
preferem definir epigenética como o estudo de alterações na atividade gênica herdáveis
mitótica e/ou meioticamente que não podem ser explicadas por mudanças na sequência de
DNA (Russo et al., 1996). Entretanto, o termo epigenética também é utilizado para descrever
processos não herdáveis, por exemplo, modificações transientes de histonas, e por isso pode
ser definido num sentido mais amplo como a adaptação estrutural de regiões da cromatina
para registrar, sinalizar ou manter estados de atividade alterados (Bird, 2007). O tema comum
entre essas definições é a existência de um mecanismo para armazenar e perpetuar uma
memória ao nível celular (Sweatt et al., 2013), sendo a definição de Russo et al. (1996) mais
amplamente aceita.
Fisicamente, alterações epigenéticas são marcas químicas no DNA ou histonas
que contribuem para compactação, estruturação, organização e atividade da cromatina. Dentre
essas marcas, uma das mais importantes e estudadas é a metilação, sendo a ligação covalente
15
de um grupo metil (-CH3), principalmente encontrada no quinto carbono da citosina no caso
do DNA, ou em aminoácidos específicos formando caudas no caso de histonas (Fig. 2).
Entretanto, metilações também são observadas em adenina e no domínio globular de histonas.
Tais marcas alteram a força de ligação entre DNA e histonas que formam o nucleossomo,
bem como a interação entre histonas e nucleossomos, participando na compactação da
cromatina (Sarma e Reinberg, 2005; Klose e Bird, 2006; Kouzarides, 2007). Além disso, as
metilações são reconhecidas por fatores associados à transcrição e remodelação da cromatina
e regulam a acessibilidade em lócus específicos do DNA (Schuettengruber et al., 2007;
Zemach e Grafi, 2007). Dessa maneira, enquanto o genoma é definido pelo conjunto de genes
codificados pelo DNA, o epigenoma é definido pela soma das histonas associadas à cromatina
com o padrão de metilação do DNA, que definem a estrutura tridimensional dentro do núcleo
e interconecta os genes ao ambiente (Fig. 2; Sweatt et al., 2013).
Figura 2. Metilação do DNA e histonas. As metilações do quinto carbono da citosina DNA
ou de histonas são marcas epigenéticas que contribuem para formação da cromatina e
influenciam a atividade dos genes ao longo do desenvolvimento e em resposta a sinais
ambientais. Figura modificada de http://commonfund.nih.gov/epigenomics/figure.
16
Mecanismos de regulação epigenética
Análises do perfil global de metilação do genoma (metiloma) de Arabidopsis
indicaram que 6-7% das citosinas são metiladas e ocupam cerca de 20% das regiões
cromossômicas (Cokus et al., 2008; Lister et al., 2008). As metilações ocorrem em três
diferentes contextos de sequência, CG, CHG e CHH (onde H é igual a A, T ou C),
distribuídas principalmente em regiões hererocromáticas ricas em transposons e elementos
repetitivos (Fig. 3). Além disso, as metilações também afetam cerca de um terço dos genes
ocorrendo em 5% das sequências promotores e em 20-30% das regiões transcritas, chamada
de corpo do gene (Zhang et al., 2006; Zilberman et al. 2007). Em geral, a metilação no corpo
dos genes apresenta um padrão bell shape ausente nas extremidades 5’e 3’o que sugere
interferência na expressão e seleção negativa (Zhang et al., 2006; Zilberman et al. 2007).
Dois padrões principais de metilação são observados em plantas. Um deles é
caracterizado por altos níveis de metilação em todos os contextos de sequência ocorrendo
principalmente em transposons e genes localizados em regiões heterocromáticas, enquanto o
outro é caracterizado por metilações do tipo CG no corpo de genes localizados em regiões
eucromáticas (Feng e Jacobsen, 2011). Esses padrões estão relacionados a funções distintas,
pois, o primeiro está associado à repressão transcricional constituindo o principal mecanismo
de controle da atividade de transposons para garantir a integridade do genoma (Zemach et al.,
2010). Já o segundo está associado à ativação da expressão ao invés de silenciamento devido
sua preferência por genes moderada ou constitutivamente expressos (Zilberman et al., 2007).
De acordo com isso, plantas mutantes para enzimas envolvidas na regulação da metilação
apresentam reativação transcricional e aumento da atividade de elementos transponíveis
(Zhang and Jacobsen, 2006; Slotkin and Martiessen, 2007; Slotkin et al., 2009; Zemach et al.,
2010). Análises recentes do metiloma de outras espécies revelaram a mesma tendência
sugerindo conservação da função da metilação entre as angiospermas (Regulski et al., 2013;
Zhong et al., 2013; Schmitz et al., 2013). Entretanto, as proporções e distribuição das
metilações ao longo do genoma variam de acordo com a espécie, em geral, aumentando com o
tamanho do genoma em consequência da maior quantidade de elementos transponíveis (Fig.
3; Mirouze e Vitte, 2014). Isso evidencia que a organização do genoma e a composição de
elementos repetitivos e de transposição modelam em parte o metiloma e vice-versa.
17
Figura 3. Perfis de metilação do DNA em diferentes espécies de angiospermas. A)
Porcentagem de metilcitosinas presentes no genoma de diferentes espécies em cada contexto
de sequência, CG, CHG, CHH (H = A, T ou C). Os valores foram calculados em relação ao
número total de citosinas em cada contexto. B) Distribuição das metilcitosinas nos diferentes
contextos ao longo do cromossomo 3 e 12 de Arabidopsis thaliana e Oriza sativa japonica,
respectivamente. C) Densidade de elementos repetitivos nos mesmos cromossomos
apresentados em B. Em A. thaliana, a região de maior pico corresponde ao centrômero.
Figuras modificadas de Mirouze e Vitte, 2014.
Marcas epigenéticas podem ser colocadas e remodeladas durante a determinação
do destino e identidade celular servindo como um sistema de armazenamento da informação
para perpetuar essa identidade ao longo da vida. Em plantas, o padrão de metilação do DNA é
inicialmente estabelecido sobre transposons e sequências repetitivas em todos os contextos
pela atividade de novo da DNA metiltransferase DOMAINS REARRANGED
METHYLTRANSFERASE2 (DRM2) em uma via conhecida como RNA-directed DNA
methylation (RdDM; Matzke et al., 2009). A RdDM é uma via complexa e não totalmente
compreendida que envolve diversas enzimas sendo elas, principalmente, componentes da
maquinaria de RNA interferente (RNAi), metiltransferases e remodeladores de cromatina
(Fig. 4; Pikaard et al. 2012; Matzke e Mosher, 2014). O modelo canônico proposto por
Pikaard et al. (2012) e adaptado por Matzke e Mosher (2014) para via RdDM descreve que,
inicialmente, transcritos gerados por RNA polimerase IV (Pol IV) são sintetizados em RNAs
de fita dupla (dsRNAs) pela enzima RNA-DEPENDENT RNA POLYMERASE 2 (RDR2)
com o auxílio do remodelador de cromatina CLASSY 1 (CLSY1). Os dsRNAs são
processados em pequenos RNAs de 24 nucleotídeos (siRNAs) pela enzima DCL3 (DICER-
18
LIKE3) no citoplasma e metilados em sua extremidade 3’ pela enzima HUA ENHANCER 1
(HEN1). Uma das fitas metiladas de siRNA é incorporada a ARGONAUTA 4 (AGO4) e o
complexo reimportado para o núcleo, onde reconhece por complementariedade transcritos
recém originados pela RNA polimerase V (Pol V). AGO4 é recrutada para o locus por meio
de interações com regiões da subunidade maior de Pol V e KOW DOMAIN-CONTAINING
TRANSCRIPTION FACTOR 1 (KTF1). A enzima RNA-DIRECTED DNA
METHYLATION 1 (RDM1), capaz de se ligar a fita simples de DNA (Gao et al., 2010),
conecta AGO4 com DRM2 que, por sua vez, catalisa a metilação de novo do DNA.
Outras enzimas também participam da via RdDM. Por exemplo, SUVH2 e
SUVH9 se ligam ao DNA metilado e contribuem para o recrutamento de Pol V (Johnson et
al., 2014; Liu et al., 2014). Já a transcrição pela Pol V é promovida pela ação do complexo
DDR, constituído pelas enzimas DEFECTIVE IN RNA-DIRECTED DNA METHYLATION
1 (DRD1), que ajuda na separação das fitas de DNA e, RDM1 e DEFECTIVE IN
MERISTEM SILENCING 3 (DMS3) que auxiliam na estabilização do lócus (Zhong et al.,
2012). Além disso, transcritos Pol V auxiliam o posicionamento dos nucleossomos por se
ligarem ao complexo INVOLVED IN DE NOVO 2/IDN2 PARALOGUE (IDN2/IDP) que,
por sua vez, interage com o complexo remodelador da cromatina SWI/SNF e, assim,
influenciam na formação da heterocromatina (Fig. 4; Finke et al., 2012; Zhu et al., 2013).
Metilações do DNA interagem com histonas modificadas em um mecanismo auto-regulatório
que garante o reforço das marcas epigenéticas e a estrutura da cromatina (Bernatavichute et
al. 2008; Ito et al. 2015).
19
Figura 4. Modelo canônico da via RdDM (RNA-directed DNA methylation). O modelo
apresenta a via RdDM para metilação de novo de regiões específicas do DNA. A via conta
com diversas enzimas e uma complexa rede de interações. Em resumo, o modelo descreve
que transcritos gerados pela Pol IV são copiados em RNA fita dupla pela enzima RDR2 e
processados em pequenos RNAs de 24 nucleotídeos (siRNAs) pela enzima DCL3 (painel da
esquerda). Uma das fitas do siRNA é incorporada a AGO4 e o complexo se liga por
complementariedade a transcritos recém originados pela Pol V. RDM1 conecta AGO4 a
DRM2 que, por sua vez, catalisa a metilação de novo do DNA qualquer contexto de sequência
(painel do meio). Transcritos da Pol V auxiliam o posicionamento dos nucleossomos por se
ligarem ao complexo IDN2/IDP que interage com o complexo remodelador da cromatina
SWI/SNF e, assim, influenciam na formação da heterocromatina (painel da direita). Maiores
detalhes das enzimas envolvidas e suas funções estão disponíveis no texto. Figura modificada
de Matzke e Mosher, 2014.
Após estabelecimento, o padrão de metilação nos contextos simétricos CG e CHG
é mantido de maneira semiconservativa e independente de siRNAs durante os ciclos
subsequentes de replicação celular (Fig. 5; Law e Jacobsen, 2010). As enzimas responsáveis
por esse processo são chamadas de metiltransferases de manutenção. Elas atuam na fita de
DNA hemimetilada de maneira preferencial em cada contexto restabelecendo as modificações
presentes em uma das fitas para a fita recém-formada. No contexto simétrico CG, a
perpetuação da metilação é realizada pela METHYLTRANSFERASE1 (MET1; Kankel et al.,
2003) em um processo que depende da remodelação da cromatina mediada por DECREASED
IN DNA METHYLATION1 (DDM1) para acesso ao DNA (Zemach et al., 2013). A
manutenção das metilações CG em plantas também requer a atividade de proteínas da família
Variant In Methylation (VIMs; Liu et al. 2007; Kraft et al., 2008). VIMs se ligam
preferencialmente ao DNA hemimetilado in vitro e possuem domínios SET- e RING-
associated (SRA) que estão relacionados ao reconhecimento de metilcitosinas (Yao et al.,
2012). Assim, foi sugerido que VIMs atuem no recrutamento de MET1 para sítios CG, uma
vez que MET1 não apresenta um módulo de ligação a metilcitosinas (Shook e Richards,
2014). VIMs e MET1 regulam a transcrição de alvos similares indicando fortemente que
VIMs regulam a expressão genica exclusivamente atuando como cofatores de MET1 (Shook e
Richards, 2014).
20
Já no contexto CHG, a metilação é mantida pela enzima
CHROMOMETHYLASE 3 (CMT3; Cao e Jacobsen, 2002) e no contexto CHH, que não
conta com simetria de sequência, a manutenção depende de siRNAs e da via RdDM para seu
reestabelecimento de novo (Chan et al., 2005). Entretanto, estudos recentes indicam que
metilação CHH também pode ser mantida na ausência de siRNAs pela ação da enzima
CHROMOMETHYLASE 2 (CMT2). As ações de CMT2 e CMT3 (CMTs) são favorecidas
pela metilação da lisina 9 da histona H3 (H3K9me; Zemach et al., 2013; Stroud et al., 2014)
catalisada pela metiltransferase SUVH4, também conhecida como KRYPTONITE (KYP;
Jackson et al., 2002). SUVH4 por sua vez é recrutada para cromatina pelas metilações CHG e
CHH (non-CG) evidenciando um mecanismo auto-regulatório positivo entre metilação de
histonas e metilação do DNA para manutenção de marcas epigenética (Du et al., 2012; Stroud
et al., 2014). Além disso, tem sido demonstrado que múltiplas metiltransferases e enzimas
relacionadas contribuem para preservação da metilação nos três contextos de sequência,
atuando de maneira auto-regulatória e reforçando a atividade umas das outras (Zemach et al.,
2013; Stroud et al., 2013).
Figura 5. Manutenção dos padrões de metilação nos três contextos de sequência após
replicação celular. Após replicação do DNA, as fitas filhas hemimetiladas são alvo da ação
das metiltransferases de manutenção que restabelecem as marcas na nova fita sintetizada. As
metilações de citosinas no contexto simétrico CG e CHG são mantidas por MET1 e CMT3,
respectivamente, enquanto metilações CHH, que não contam com a simetria de sequência, são
restabelecidas de novo pela DRM2 e a via RdDM. Outras enzimas também participam desse
processo conforme descrito em maiores detalhes no texto. Figura modificada de Teixeira e
Colot, 2010.
21
A metilação do DNA é considerada uma marca epigenética mitótica e
meióticamente estável, entretanto, níveis reduzidos são observados durante a gametogênese e
embriogênese de plantas e animais, demonstrando que o perfil de metilação pode ser
reprogramado (Law e Jacobsen, 2010). Atuando em oposição aos mecanismos de
estabelecimento e manutenção da metilação no DNA, existem mecanismos de demetilação
que moldam de forma dinâmica escapes epigenômicos e permitem a expressão de genes (Feng
et al., 2010). Tais processos são importantes, por exemplo, para imprinting parental e reseting
de marcas epigenética adquiridas ao longo da vida. A perda de metilação pode ocorrer
passivamente, por replicação celular na ausência ou deficiência de vias de manutenção, ou
ativamente pela remoção de citosinas metiladas realizada por enzimas específicas (Zhu et al,
2009; Law e Jacobsen, 2010).
Em plantas, a demetilação ativa é promovida pela atividade de DNA glicosilases
associadas à via de reparo de bases excisadas (Zhu, 2009). DEMETER (DME) e
REPRESSOR OF SILENCING1 (ROS1) são os membros fundadores da família de DNA
glicosilases encontradas em A. thaliana (Choi et al., 2002; Gong et al., 2002), que também
incluem DML2 e DML3 (Fig. 6A; Penterman et al., 2007a). As DNA glicosilases são
enzimas bifuncionais que removem citosinas metiladas e causam uma quebra na fita de DNA
(Agius et al., 2006). Essa lacuna no DNA é catalisada pelas AP endonucleases APE1L e ZDP,
gerando uma extremidade 3’-OH que permite a atividade de uma DNA polimerase e ligase
ainda desconhecidas para incorporação da nova citosina (Fig. 6B; Li et al. 2015c). APE1L e
ROS1 interagem in vitro e co-localizam in vivo, sugerindo fortemente a formação de
complexos que coordenam suas atividades (Li et al., 2015c). De acordo com isso, os mutantes
ros1 e rdd (o triplo mutante ros1, dml2 e dml3), assim como ape1l e zdp, apresentam aumento
da metilação em todos os contextos de citosinas, um fenômeno descrito como hipermetilação
(Gong et al., 2002; Agius et al., 2006; Penterman et al., 2007a; Li et al. 2015c).
DME e ROS1 apresentam papeis biológicos distintos, sendo DME
preferencialmente expresso na célula central do gametófito feminino e na célula vegetativa do
pólen (Choi et al., 2002; Schoft VK et al., 2011), enquanto ROS1 é expresso em quase todos
os tecidos vegetais (Gong et al., 2002). Em Arabidopsis, DME demetila o alelo maternal de
genes imprintados no endosperma, por exemplo, FWA (FLOWERING WAGENINGEN), MEA
(MEDEA) e FIS2 (FERTILIZATION-INDEPENDENT SEED 2), entre outros (Hsieh et al.,
2009; Ikeda, 2012). Essa demetilação é necessária para expressão do alelo maternal e o
desenvolvimento normal do embrião, já que mutantes dme resultam na letalidade da semente
22
(Choi et al., 2002). Entretanto, estudos recentes indicam que a função basal de DME seja a
reativação de transposons em células companheiras tanto do gametófito feminino quanto
masculino (Calarco et al., 2012; Ibarra et al., 2012). Esse mecanismo atuaria como um
sistema imune para ativar a produção de siRNAs e reforçar o silenciamento de transposons
nos gametas. As demais DNA glicosilases também parecem contribuir nesse processo
(Calarco et al., 2012).
Em contrapartida, ROS1 atua na demetilação de regiões que flanqueiam os genes
evitando a dispersão da metilação a partir de transposons próximos e um possível
silenciamenteo ocasional (Penterman et al., 2007b; Zhu et al., 2007). DML2 e DML3 atuam
no mesmo sentido, mas ficam à margem de ROS1 uma vez que são bem menos expressos
(Zhu et al., 2009). Diversos estudos indicam que a demetilação ativa ocorre em loci
eucromáticos e seria um mecanismo para equilibrar a ação da via RdDM (Penterman et al.,
2007b; Zhang e Zhu, 2012; Matzke e Mosher, 2014). De acordo com isso, muitos
componentes da via RdDM estão desregulados em mutantes ros1 e, analogamente, mutantes
relacionados a RdDM apresentam expressão reduzida de ROS1 (Zhang e Zhu, 2012).
Recentemente, foi revelado que RdDM e ROS1 interagem para balancear os níveis de
metilação do epigenoma (Lei et al., 2015; Williams et al., 2015). Estes trabalhos mostraram
que as vias de RdDM e de demetilação ativa atuam fortemente em transposons próximos ao
lócus de ROS1 e que a expressão de ROS1 é promovida pela metilação e antagonizada por
demetilação, em contraste a maioria dos alvos típicos metilados (Fig. 6C). Dessa forma, o
lócus de ROS1 funciona como um reostato epigenético sintonizando a atividade de
demetilação em resposta a alteração dos níveis de metilação para, assim, garantir a
estabilidade do genoma. Também foi demonstrado que essa expressão de ROS1 sensível a
metilação é conservada em outras espécies, sugerindo ser uma característica importante para
adaptação (Williams et al., 2015).
23
Figura 6. DNA glicosilases e mecanismos de funcionamento. A) Esquema mostrando a
estrutura dos genes que compõem a família de DNA glicosilases (DMLs) de Arabidopsis
thaliana. Regiões retangulares correspondem aos éxons e linhas aos introns. Éxons azuis
codificam o domínio helix–hairpin–helix, rosa e amarelo codificam outros domínios
conservados enquanto pretos domínios não conservados entre as DMLs. Os locais de inserção
de T-DNA ros1–3, ros1–4, ros1–5, dml2–1, e dml3–1 são indicados pelos triângulos. Figura
modificada de Penterman et al., 2007a. B) Esquema mostrando a via de demetilação ativa.
Nesse modelo ROS1 é bifuncional removendo citosina metilada e clivando o sítio abásico do
esqueleto de DNA por meio da β ou βδ eliminação. Esse processo resulta numa lacuna com
terminação PUA ou fosfato (P) em uma das fitas de DNA que é catalisada por APE1L or ZDP
e substituída por uma hidroxila (-OH). Assim, essa lacuna pode ser preenchida por uma
citosina não metilada pela ação de uma DNA polymerase e DNA ligase ainda não
identificadas (representadas por “?” na figura). Figura modificada de Li et al., 2015c. C)
Esquema mostrando que a regulação de ROS1 por metilação do DNA atua como um reostato
epigenético. O esquema resume o papel de RdDM e ROS1 no lócus de ROS1 (auto-regulação)
e em um outro alvo genômico metilado. Ao contrário de genes típicos, cuja diminuição dos
níveis de metilação está relacionada ao aumento de expressão, a demetilação no lócus de
ROS1 reduz sua própria expressão antagonizando a via RdDM. Isso garante que um balanço
das atividades de metilação e demetilação seja mantido no genoma. Os símbolos “+” e “-”
24
denotam um efeito positivo e negativo respectivamente na expressão dos genes. Figura
modificada de Williams et al., 2015.
Todas as DNA glicosilases contém uma região N-terminal rica em lisina sendo
demonstrado em ROS1 que essa região é necessária para ligação ao DNA e ocorre de maneira
independente da metilação (Ponferrada-Marin et al., 2010; Ponferrada-Marin et al., 2012). A
hipermetilação dos mutantes ros1 e rdd ocorre em loci específicos, não sendo observado
alterações substanciais nos níveis globais (Penterman et al., 2007a; Lister et al., 2008), o que
sugere a atuação de mecanismos direcionando as DNA glicosilases (Zhu et al., 2007; Zhu,
2009). Em mutantes de Arabidopsis para as polimerases Pol IV e V, a produção de siRNAs
heterocromáticos de 24 nucleotídeos é reduzida, coincidindo com maiores níveis de metilação
em alguns loci correspondentes (Mosher et al., 2008). Isso indica que esse tipo de siRNA guia
a demetilação ativa e que Pol IV e V estão envolvidas nesse processo (Zhu, 2009). De acordo
com isso, foi descrita uma associação de ROS1 com ROS3, uma enzima RNA-binding
envolvida na demetilação ativa e capaz de incorporar siRNAs que reconheceriam regiões do
DNA por complementariedade (Zheng et al., 2008). Mecanismos similares têm sido descritos
em animais, no qual siRNAs direcionados ao promotor parecem ativar a expressão de certos
genes (Janowski et al., 2007). Este processo foi relacionado à demetilação e referido como
RNA-directed DNA demethylation (Zhu, 2009). Interessantemente, siRNAs também são
necessários para atividade RdDM, reforçando que mecanismos de metilação e demetilação
compartilham componentes e estão estreitamente relacionados.
Embora seja bem estabelecida a interdependência entre metilação do DNA e
transcrição (Zhang et al., 2006; Zilberman et al., 2007), como as metilações afetam a
acessibilidade de fatores de transcrição (FTs) ao DNA é um tema pouco compreendido. Ainda
não se sabe quando a alteração da metilação é causa ou consequência da alteração da
expressão (Teixeira e Colot, 2009; Medvedeva et al., 2014). Se as metilações são a causa,
então poderiam afetar diretamente a afinidade de FTs aos seus sítios de ligação ou,
alternativamente, atrair fatores remodeladores de cromatina interferentes. Se a metilação é
uma consequência da transcrição, então modificações da cromatina poderiam reduzir o acesso
dos FTs e da maquinaria de transcrição ao DNA, e a metilação serviria para fixar o estado da
cromatina (Medvedeva et al., 2014). Nesse caso, a metilação do DNA seria acumulada
passivamente como resultado da ausência de ligação de FTs (Wang et al., 2012) ou
25
diretamente pelo recrutamento de DNA metiltransferases por proteínas associadas a
modificações de histonas específicas (Vire et al., 2006).
Estudos recentes mostram que a ação direta e seletiva das metilações do DNA
prevenindo a ligação de FTs é restrita a poucos exemplos e, ao invés disso, as metilações
parecem atuar como sinalizadores atraindo complexos ativadores ou repressores (Medvedeva
et al., 2014; Li et al., 2015b). Nesse contexto, Methyl-Binding Proteins (MBPs) são descritas
como proteínas que interpretam a metilação do DNA e controlam a expressão de genes por
atraírem fatores remodeladores da cromatina (Zemach e Grafi, 2007, Baubec et al., 2013). Em
Arabidopsis são encontradas 13 MBPs com diferentes funções (Zemach e Grafi, 2007).
MBD7, por exemplo, é descrita como um fator anti-silenciamento que reconhece sítios CG
metilados e se associa a proteínas chamadas Increased DNA Methylation (IDMs) para
facilitar a demetilação do lócus por ROS1 (Fig. 7A; Qian et al., 2012; Qian et al., 2014; Lang
et al., 2015).
Além disso, metilações do DNA regulam a metilação de histonas em um processo
inter-relacionado e auto-regulatório necessário para reforço das marcas epigenéticas e que
também afeta a transcrição (Cedar e Bergman, 2009; Rose e Klose, 2014; Ito et al., 2015).
Conforme mencionado anteriormente, CMTs, responsáveis pela manutenção das metilações
CHG e CHH (non-CG), são recrutadas ao lócus pela marca H3K4me e, de maneira recíproca,
SUVH4, responsável por metilar H3K4, é atraída por metilações non-CG (Du et al., 2012;
Stroud et al., 2014). Outros exemplos de modificações de histona incluem a trimetilação da
lisina 27 da histona H3 (H3K27me3) adicionada por complexos do grupo Polycomb (PcG;
Reddington et al. 2013) e trimetilação da lisina 4 da histona H3 (H3K4me3) adicionada por
complexos do grupo Trithorax (trxG; Klymenko e Muller, 2004). H3K27me3 e H3K4me3
estão envolvidas no silenciamento e ativação transcricional, respectivamente (Schwartz e
Pirrotta, 2007; Schuettengruber et al., 2011).
Recentemente foi descrito um modelo que pode explicar como FTs interagem
com modificações de histona para promover a transcrição (Fig. 7B; Song et al., 2015). Nesse
modelo, FTs induzíveis são capazes de recrutar o complexo trxG para o lócus atraindo outros
fatores e formando o complexo COMPASS que, por sua vez, facilita a interação das
metiltransferases com a histona H3 gerando H3K4me3 e permitindo a transcrição pela
polimerase II (Pol II). De acordo com isso, diversos grupos de FTs e non-coding RNA
parecem recrutar proteínas dos complexos PcG e trxG (Pu e Sung, 2015). Além disso,
26
COMPASS é conhecido por facilitar a formação do complexo de pré-iniciação (PIC) e gerar
H3K4me3 durante a elongação da transcrição (Ding et al., 2012), sendo essa modificação
associada a atividade de Pol II (Cedar e Bergman, 2009). Assim, esse modelo juntamente com
outros dados explica em parte a inter-relação previamente observada entre metilação do DNA,
metilação de histonas e a expressão gênica (Li et al., 2008; Cedar e Bergman, 2009).
Fig. 7. Modelos de interação entre fatores em trans e marcas epigenéticas. A) Esquema
mostrando a função anti-silenciamento de MBD7 e IDMs em um lócus metilado.
Inicialmente, MBD7 se liga a metilação (1) e recruta proteínas IDM para o lócus (2). IDM1
acetila caudas de histonas facilitando acesso de ROS1 (3) que, por sua vez, previne a
hipermetilação e silenciamento gênico (4). Figura modificada de Lang et al., 2015. B)
Esquema mostrando um modelo hipotético para o envolvimento de fatores de transcrição e
modificações de histona. Durante a formação do complexo de pré-iniciação (PIC), fatores de
transcrição específicos (FT) e gerais (GTFs) são recrutados para o promotor juntamente com
proteínas de ligação TATA-box (TBP) e mediadores. Ao mesmo tempo, FTs interagem com
componentes do complexo COMPASS (Ash2, WDR5a e RbBP5) para facilitar sua interação
com metiltransferases de histona e o domínio carboxyl-terminal (CTD) da polimerase II (Pol
II). Tal interação promove a trimetilação da lisina 4 da histona H3 (H3K4me3) no promotor
do gene que está relacionada a alongamento da transcrição (representado por “+1” na figura).
Figura modificada de Song et al., 2015.
27
Importância das metilações do DNA para o desenvolvimento e adaptação vegetal
O desenvolvimento apropriado dos organismos depende de uma rede intrincada e
precisa de regulação da expressão gênica. A metilação do DNA, por sua vez, interfere na
transcrição sendo assim também importante para definição de identidades celulares e
coordenação do desenvolvimento (Zilberman et al., 2007; Zhang et al., 2010). De acordo com
isso, mutantes para enzimas relacionadas à regulação epigenética apresentam múltiplas
anormalidades (Fig. 8A; Zhang e Jacobsen, 2006) e linhagens isogênicas de plantas com
epigenoma variante (Epigenetic recombinant inbred lines ou epiRILs) respondem
diferencialmente a estímulos ambientais (Cortijo et al., 2014).
Além disso, foram descritos diversos exemplos em que versões de genes
metilados (epialelos) afetam características morfológicas. Em Linaria vulgaris, o aumento dos
níveis de metilação do gene Lcyc, um ortólogo de CYCLOIDEA, reprime sua expressão
modificando a simetria das flores (Fig. 8B; Cubas et al. 1999). No melão (Cucumis melo L.),
alterações epigenéticas no promotor de CmWIP1 influencia na determinação sexual das flores
(Martin et al., 2009). FLOWERING LOCUS WAGENINGEN (FWA) não é expresso durante o
desenvolvimento vegetativo, entretanto, perdas de metilação no promotor induzem sua
expressão e atrasam o florescimento de Arabidopsis (Soppe et al., 2000; Fujimoto et al.,
2011). Muitas variações epialélicas naturais têm sido descritas relacionadas ao florescimento
(Fornara et al. 2010; Andres e Coupland 2012), além de outras afetando características
agronômicas, tais como, amadurecimento e produção de vitamina E no fruto de tomate
(Zhong et al., 2013; Quadrana et al., 2014). Assim, os dados sugerem que mudanças
fenotípicas causadas por alterações na metilação de DNA poderiam contribuir para processos
adaptativos.
28
Fig. 8. Alterações epigenéticas afetam características morfológicas de plantas. A) Duplos
mutantes para enzimas envolvidas em vias regulatórias da metilação do DNA, por exemplo,
met1/cmt3 e drm1/met, apresentam múltiplas anormalidades durante o desenvolvimento.
Figura modificada de Zhang e Jacobsen, 2006. B) Variação da simetria floral em Linaria
vulgaris (pelórico) causada pelo aumento dos níveis de metilação de Lcyc e consequente
repressão de sua expressão. Figura modificada de Cubas et al, 1999.
Diferente de mamíferos, cujo padrão de metilação do DNA é apagado e
restabelecido de novo durante a gametogênese e embriogênese, em plantas as metilações são
em grande parte mantidas através da meiose (Feng et al., 2010). Plantas se desenvolvem a
partir de meristemas que estabelecem suas linhagens germinais tardiamente e, dessa maneira,
alterações epigenéticas adquiridas ao longo da vida podem ser memorizadas e herdadas
(Henderson e Jacobsen, 2007). Entretanto, a descoberta de que siRNAs estão envolvidos na
correção de perdas acidentais da metilação (Teixeira et al., 2009), bem como remetilações
observadas no epigenoma de epiRILs (Johannes et al., 2009; Reinders et al., 2009), revelou
que nem todas as alterações epigenéticas são mantidas, pois alguns loci são remetiláveis
enquanto outros podem permanecer alterados e originar epialelos.
Análises em A. thaliana revelaram que, no decorrer das gerações, as taxas de
alteração espontânea na metilação do DNA são 1.000 vezes maiores do que mutações
genéticas (Becker et al. 2011; Schmitz et al. 2011). Entretanto, a taxa de alteração epigenética
herdável em grupos de citosinas (Differential Methylated Regions ou DMRs), funcionalmente
mais relevantes do que individuais (Weigel e Colot, 2012), são similares às taxas de mutações
na sequência de DNA (Becker et al. 2011; Schmitz et al. 2011). De qualquer forma, uma vez
que as DMRs afetam a expressão e podem ser estavelmente herdadas, esses dados indicam
que a variação epigenética pode ser selecionada e ter um papel adaptativo similar às variações
genéticas. Tais modificações epigenéticas são frequentemente revertidas e os epialelos
formados estão associados à produção de siRNAs, o que sugere a participação de
metiltrasferases e DNA glicosilases atuando dinamicamente no controle das metilações
(Matzke e Mosher, 2014; Lei et al., 2015; Williams et al., 2015).
Enzimas envolvidas no controle da metilação são responsivas a estresses
modelando o epigenoma e regulando a expressão de genes (Sahu et al., 2013; Le et al., 2014).
Por exemplo, plantas de Arabidopsis infectadas por patógenos apresentam regiões
29
diferencialmente metiladas referentes a genes requeridos para resistência (Dowen et al.,
2012). De acordo com isso, mutantes relacionados à via de regulação da metilação de DNA,
tais como, met1, ago4, polV e ros1, têm apresentado diferenças na resistência ou
susceptibilidade à infecção (Yu et al., 2013). Em geral, a ativação ou repressão desses genes
está envolvida com a regulação de transposons inseridos em regiões vizinhas. Certo nível da
atividade de transposons é mantido pela ação das DNA glicosilases, sugerindo um efeito
positivo em processos de adaptação a estresses (Zhu et al., 2009; Mirouze e Vitte, 2014).
Dessa forma, o controle dinâmico das metilações de DNA é interpretado como um
mecanismo importante para manter o epigenoma variável e adaptável, capaz de responder de
maneira rápida e eficiente a sinais do ambiente durante o desenvolvimento (Kooke et al.,
2015).
O impacto das modificações epigenéticas no desenvolvimento de plantas
cultivadas sob condições de estresse é complicado de ser estudado, principalmente porque
grande parte da variação fenotípica se deve a variação genética entre populações naturais
(Kooke et al., 2015). Entretanto, análises de diferentes populações de epiRILs têm revelado
que a variação epigenética contribui para variações morfológicas sendo essas características
altamente herdáveis (Johannes et al., 2009; Reinders et al., 2009; Latzel et al., 2012; Zhang et
al., 2013). Recentemente foi demonstrado por Cortijo et al. (2014) que regiões
diferencialmente metiladas atuam como loci epigenéticos de traços quantitativos (Epigenetic
quantitative trait loci ou epiQTL) e são responsáveis pela maioria da variação herdável no
tempo de florescimento e tamanho da raíz observada entre epiRILs. Importantemente, nesse
mesmo trabalho também foi demonstrado que os efeitos fenotípicos não são causados por
inserções de transposons indicando ser um efeito da própria diferença de metilação do DNA.
De acordo com isso, genes metilados e não próximos a transposons apresentam maior
variação da expressão quando perdidas as metilações do que genes não metilados (Zilberman
et al., 2007). Isso indica que variação da metilação do DNA pode causar variação fenotípica,
sendo uma característica herdável e selecionável.
Plantas mutantes de vias relacionadas à metilação do DNA apresentam variações
morfológicas pleiotrópicas suprimidas em plantas selvagens, indicando que a metilação é
capaz de regular a plasticidade fenotípica (Zhang e Jacobsen, 2006; Mirouze e Paszkowski,
2011). Por plasticidade fenotípica entende-se a habilidade de um genótipo para expressar
fenótipos alternativos em diferentes ambientes (Schlichting, 1986). Essa característica
30
governada por variações do epigenoma tem sido interpretada como uma compensação ao
modo de vida séssil das plantas e importante para adaptação às condições ambientais
oscilantes e variáveis (Feng e Jacobsen, 2011). Assim, a metilação do DNA tem sido
interpretada como um fator atenuante da variação morfológica e, quando desregulada, poderia
ser valiosa para melhoramento do desempenho em condições desfavoráveis (Mirouze e
Paszkowski, 2011). De acordo com isso, estados hipometilados induzidos aumentam a
sensibilidade de resposta do genoma ao ambiente, como observado para epiRILs, e por serem
herdáveis essas características poderiam ser alvo de seleção (Zhang et al., 2013).
Entretanto, mecanismos de reforço da metilação do DNA são descritos em
meristemas para silenciamento da transcrição via RdDM indicando que modificações
epigenéticas em tecidos somáticos sejam mais comuns do que em células germinativas
(Baubec et al., 2014). Isso sugere que alterações na metilação do DNA podem participar na
adaptação local e momentânea, mas não seja fixada, pois, poderia ser desnecessária em outros
contextos. Detalhes de como é memorizada essa variação epigenética adquirida ao longo do
desenvolvimento e herdada transgeneracionalmente são aspectos ainda pouco compreendidos.
Tais modificações, quando recorrentes, poderiam ocorrer diretamente em linhagens germinais
ou, alternativamente, gerar pequenos RNAs migratórios a partir da região induzida (tecidos
somáticos) para meristemas, formando um conduite de informação conforme proposto por
Dowen et al. (2012). De qualquer modo, variação na metilação do DNA claramente afeta a
expressão de genes em populações naturais e, existem indicações de que isso possa ser
predominante para genes recentes originados de novo (Silveira et al. 2013).
QQS, um gene órfão e provocador
O estudo de mutantes de Arabidopsis thaliana para enzimas relacionadas à via
RdDM revelou a existência de alguns loci candidatos ao controle epigenético (Kurihara et al.,
2008). Um desses lócus, identificado como At3g30720, foi caracterizado contendo o gene
Qua-quine Starch (QQS) que codifica uma proteína de 59 aminoácidos sem domínios
catalíticos ou motifs estruturais descritos e que está envolvida no metabolismo de amido (Li et
al., 2009). QQS não apresenta homologia com qualquer sequência de outros organismos
sendo, portanto, considerado um gene órfão. Genes órfãos são definidos como genes com
sequências codificadoras de proteínas novas e únicas a uma espécie ou táxon (Gollery et al.,
31
2006), sendo por isso também chamados de genes espécie-específicos, genes novos ou
neogenes. Essa classificação inclui genes recém-formados a partir de sequências não gênicas,
bem como, descendentes de genes ancestrais cujas sequências mudaram além da capacidade
de reconhecimento, mas rejeita genes transferidos horizontalmente e genes duplicados que
assumiram novas funções (Arendsee et al., 2014). Posteriormente, foi demonstrado que QQS
é regulado epigenéticamente, pois, sua expressão está inversamente correlacionada aos seus
níveis de metilação de DNA e, além disso, que é propenso à variação epialélica natural
sugerindo participação em processos adaptativos (Silveira et al. 2013).
A estrutura de QQS é formada por três éxons e dois introns inseridos um na região
5’UTR e outro na região codificadora e as metilações se distribuem principalmente na região
promotora e 5'UTR sendo ausente na região traduzida (Fig. 9; Lister et al., 2008; Cokus et al.,
2008). Uma notável característica de sua vizinhança cromossômica são os múltiplos
pseudogenes e transposons homólogos a família CACTA-like, gypsi-like e retroelementos,
revelando seu posicionamento em uma ilha heterocromática (Fig. 9). A região cromossômica
que cerca QQS parece ter sido um sítio de alta atividade de transposons durante a evolução e,
como esperado, é altamente metilada (Lister et al., 2008). A expressão de QQS varia durante
o desenvolvimento e em condições de estresse, além de mutantes para genes de enzimas
envolvidas em processos de metilação e entre acessos naturais, estando sempre correlacionada
a alteração dos níveis de metilação (Hruz et al., 2008; Li et al., 2009; Seo et al., 2011;
Silveira et al. 2013). As diferenças de expressão não são relacionadas a mutações na
sequência de DNA ou em fatores em trans ou a atividade de transposons vizinhos, sugerindo
que a regulação se deve a variações na metilação do DNA (Silveira et al. 2013).
32
Fig. 9. Localização, estrutura e distribuição de metilações do DNA de QQS. O esquema
mostra o posicionamento de QQS em uma ilha heterocromática do cromossomo 3 de
Arabidopsis thaliana (acesso Columbia-0) repleta de transposons e repetições. A estrutura de
QQS é formada por três éxons e dois íntrons, inseridos um na região 5’UTR e outro na região
codificadora. As metilações de citosinas estão distribuídas na região promotora e 5'UTR e
ausentes na região transcrita e 3’UTR. Metilações no contexto CG estão representadas por
traços vermelhos, enquanto CHH e CHG, por traços azuis e verdes, respectivamente.
Pequenos RNAs (siRNAs) de diferentes tamanhos e o número de cópias encontradas de cada
região também estão presentes na figura (setas). Setas vermelhas representam siRNAs de 24
nucleotídeos (nt), enquanto setas verdes e azuis, siRNAs de 22-23 nt e 20-21 nt,
respectivamente. Figura modificada de Silveira et al., 2013.
A existência de epialelos naturais de QQS sugere que regulações epigenéticas
atuantes sobre o lócus são bastante flexíveis e versáteis, embora não se saiba o quanto estes
epialelos influenciam no vigor ou adaptação da planta. Entretanto, diversos trabalhos mostram
que a função de QQS durante o desenvolvimento vegetativo é regular o metabolismo de
amido. Li et al. (2009) foi a primeira a demonstrar a correlação inversa entre a expressão de
QQS e acúmulo de amido nas folhas de Arabidopsis. Outros trabalhos confirmaram essa
associação e mostraram que QQS é afetado por condições de estresse impactando o
crescimento das plantas. Por exemplo, sob condições de frio uma isoforma do fator de
transcrição IDD14 é ativada diminuindo a expressão de QQS e, assim, a quantidade de amido
aumenta para maior eficiência do desenvolvimento durante noites frias (Seo et al., 2011). A
expressão de QQS também é afetada por patógenos como Pseudomonas syringae pv. tomato
(Pst) sendo correlacionada a demetilação ativa do loci (Dowen et al., 2012). Além disso,
recentemente foi descrito que mutantes para ARA6, uma proteína que intermedia o tráfego a
partir dos endossomos para membrana plasmática, é responsável pela homeostase de amido e
açúcar por meio da regulação de QQS e tais plantas apresentam resistência a Pst (Tsutsui et
al., 2015).
Esses resultados indicam que QQS integra uma diversa gama de estresses bióticos
e abióticos para sintonizar o desenvolvimento vegetal, contribuindo para adaptação e
constituindo um modelo para estudos de genes órfãos conforme sugerido por Arendsee et al.
(2014). Além disso, QQS continua apresentando funções quando transferido para outras
espécies distantes como a soja, sendo sugerida uma interação com parceiros e vias
33
conservadas (Li e Wurtele, 2014). De fato, foi demonstrado recentemente por Li et al. (2015a)
que que a proteína QQS é capaz de se ligar ao regulador transcricional NF-YC4 (Nuclear
factor Y, subunit C4) conservado entre eucariotos e, assim, regular processos metabólicos que
afetam o particionamento de carbono e nitrogênio entre proteínas e carboidratos modulando a
composição de folhas e sementes de Arabidopsis e soja (Glycine max). Dessa forma, QQS é
um gene que apresenta características únicas e extremamente provocadoras para uma série de
questões relevantes a biologia.
A taxa de produção de transcritos num certo lócus é determinada pela interação
entre reguladores em trans e suas sequências alvo em cis. Entretanto, a capacidade de
interação entre esses fatores cis e trans depende do perfil estrutural da cromatina que é
influenciada por marcas epigenéticas, tais como, a metilação do DNA (Zilberman et al.,
2007). Em plantas superiores, uma correlação estreita entre metilação e silenciamento da
transcrição foi estabelecida (Law e Jacobsen, 2010). Porém, o quanto e como as metilações do
DNA afetam a expressão de genes ainda é um aspecto pouco relatado e compreendido.
A expressão de QQS está sob controle epigenético e apresenta um padrão de
regulação específico ao longo do desenvolvimento e em resposta a alterações do ambiente.
Portanto, QQS poderia funcionar como um sensor e ser usado como um modelo para melhor
entendimento da interação entre mecanismos de controle epigenético e sinais do
desenvolvimento/ambiente que regulam a expressão gênica. A regulação epigenética é
considerada atualmente um importante aspecto da interação entre genótipo e ambiente para
definir o fenótipo. Estudar como esses tópicos interagem é essencial para compreender o
funcionamento dos genomas e para programas de melhoramento vegetal mais refinados.
OBJETIVO
Estudar como os perfis de metilação de citosinas no promotor e região 5’UTR de
epialelos Qua-Quine Starch (QQS) são alterados no decorrer do desenvolvimento
contribuindo para estabelecimento do padrão de expressão deste gene e os efeitos sobre o
acúmulo de amido.
34
RESULTADOS E DISCUSSÃO
Manuscrito “Developmental expression pattern of QQS is influenced by DNA
methylation: involvement of ROS1 and impact on starch metabolism”.
Oliveira, RR1; Viana, AJC
1; Matiolli, CC
1; Silveira, AB
2; Vincentz, M
1.
1 Centro de Biologia Molecular e Engenharia Genética, Universidade Estadual de Campinas,
Cidade Universitária Zeferino Vaz, Campinas, São Paulo, Brazil.
2 Ecole Normale Supérieure, Institut de Biologie (IBENS), Paris, France.
Abstract
Cytosine methylation (mC) is an epigenetic mark that affects gene expression. In plants, the
developmentally induced re-programming of mC profiles is critical for transposon silencing
and gene expression regulation. Yet, the extent to which and how the changes of mC pattern
may be modulated during vegetative development is poorly understood. Qua-Quine Starch
(QQS) is an orphan gene of Arabidopsis thaliana which present several spontaneous stably-
inherited epialleles and their expression levels correlate inversely with the mC levels in the
promoter and 5’UTR. We show that the mC level in QQS epialleles modulates its expression
during vegetative development. The QQS expression was increased in leaf and correlated with
a demethylation which is not associated with the action of the DNA glycosilase ROS1.
Genetic analysis using ros1-4 plants, which carry a strongly methylated QQS epiallele,
showed that ROS1 is necessary to define QQS expression level, most likely, by regulating
directly the methylation status of QQS. Plants carrying contrasting QQS epialleles showed
different starch contents in agreement with its described function. Together, this data offer
insights into how mC modulates gene expression during plant development and how QQS
epialleles could be formed and affect starch metabolism.
35
Introduction
Cytosine-5 methylation (mC) is an epigenetic mark associated with transcriptional
regulation, in most cases promoting silencing of transcription (Law and Jacobsen, 2010). In
plants, mC is associated with heterochromatin formation necessary to control the activity of
transposable elements (TEs) that could threaten genome integrity and influences expression of
nearby genes (Goll and Bestor, 2005; Feng and Jacobsen, 2011). Development of all
organisms is dependent on an intricate and precise regulation of gene expression which is
interdependent of mC profiles (Zilberman et al., 2007; Zhang et al., 2010). Accordingly,
multiple abnormalities are observed in mutants for enzymes involved in regulatory pathways
of DNA methylation (Zhang and Jacobsen, 2006). Moreover, the developmentally induced re-
programming of mC profiles are critical for gene imprinting in the germline and for early
embryo development (Choi et al., 2002; Hsieh et al., 2009; Ibarra et al., 2012). However, how
mC at cis regulatory sequences interacts with trans factors to modulate gene expression and
vegetative development is not well known.
How mC interacts with transcription has been widely debated and a main issue is
related to whether altered gene expression is a cause or a consequence of mC (Teixeira and
Colot, 2009; Medvedeva et al., 2014). If mC is the cause of gene repression, then mC may
directly affect the affinity of transcription factors (TFs) towards their binding sites (TFBSs)
or, alternatively, attract remodeling chromatin factors, such as, methylation-binding proteins
(MBPs) or Polycomb (PcG) factors (Schuettengruber et al., 2007; Li et al., 2015c) which
affect chromatin accessibility. If mC is the consequence, then chromatin modification reduce
the access of TFs and transcriptional machinery to the DNA and mC serves to fix this
repressed state of the chromatin. In this later case, mC could accumulate passively as a result
of the absence of TF binding (Wang et al., 2012) or direct DNA methyltransferase
recruitment by associated PcG protein (Vire et al., 2006). However, mC of TBFS preventing
TF binding does not appear to be a general transcriptional regulatory mechanism since very
few cases have been described (Medvedeva et al., 2014; Li et al., 2015b). Instead, mC at
genes seems to acts largely as traffic-lights at TBFS attracting mainly repressor but also
activator complexes such as MBPs and PcG factors (Medvedeva et al., 2014).
The regulation of mC levels is a dynamic process involving pathways that are
dedicated to the incorporation or excision of methylated cytosines (Law and Jacobsen., 2010).
In plants, the mC is found in all contexts of sequence (CG, CHG or CHH, where H = A, T or
36
C). This mC is established by the de novo DNA metiltransferase DRM2 through a pathway
known as RdDM (RNA-directed DNA methylation), which is guided by 24 nt siRNA
(Matzke et al., 2009). Then, posteriorly, the mC patterns are maintained in a semi-
conservative manner during DNA replication by the action of the DNA methyltransferases,
which present variable affinities depending on the mC sequence context (Teixeira and Colot,
2010). In the asymmetric CHH context, mC is maintained by RdDM and also CMT2
(Henderson and Jacobsen, 2007; Zemach et al., 2013).
DNA methylation can be also excised by active mechanisms or lost passively
(Law and Jacobsen, 2010). Passive demethylation occurs in the absence of DNA
methyltransferases during DNA replication, whereas active demethylation occurs via DNA
glycosilases able to excise mC independent of the sequence context and important to avoid
genome hypermethylation and protect genes near to heterochromatic regions (Zhu, 2009;
Feng et al., 2010). DEMETER (DME) and REPRESSOR OF SILENCING1 (ROS1) are the
major active demethylation players found in Arabidopsis thaliana and show distinct
biological roles (Penterman et al., 2007). DME is required for gene imprinting during
gametogenesis whereas ROS1 acts at specific loci during vegetative development (Morales-
Ruiz et al., 2006). A relationship between RdDM and ROS1 was described showing that the
regulation of mC levels is at a dynamic equilibrium (Lei et al., 2015; Williams et al., 2015)
which partly explains the interplay between methylation and demethylation processes. In
agreement, ros1 mutant plants accumulate hypermethylation at specific loci throughout the
genome (Gong et al., 2002; Agius et al., 2006; Zhu et al., 2007).
Plants present high plasticity along development, probably, to compensate for
their sessile lifestyle and the necessity to deal with frequent environmental alterations.
Recently, isogenic Arabidopsis thaliana lineages with different epigenomes, were shown to
respond differently along development to environmental changes suggesting that epigenetic
components modulate plant developmental plasticity and contribute to local adaptation (Turck
and Coupland 2014; Cortijo et al., 2014; Kooke et al., 2015). In general, it has been
demonstrated that global induced hypomethylation states increases plant sensibility to
environmental conditions and cause phenotypic variation that, in turn, could be inherited and
selected contributing to plant adaptation (Cortijo et al. 2014; Kooke et al. 2015; Dubin et al.
2015). Accordingly, multiples aspects of development, such as, morphology, vernalization
and flowering, stress responses and others, are direct or indirectly regulated by epigenetic
marks (Feng and Jacobsen, 2011). However, the developmental variation was only revealed
37
for large epigenetic QTLs induced in mutant backgrounds or by extreme stress conditions and
the impact of individual epialleles is poorly understood Thus, elucidating how mC interacts
with developmental signals and modulate gene expression constitute a crucial step to
understand how the genome interacts with its environment and generates phenotypical
variability.
Qua-Quine Starch (QQS, At3g30720) is an orphan gene of A. thaliana prone to
natural epiallelic variation and stably inherited over generations (Silveira et al. 2013).
Epiallelic variation means here different cis methylation levels in the QQS promoter and
5’UTR which correlates inversely with its expression levels. During gametogenesis QQS is
demethylated by DEMETER (DME; Ibarra et al., 2012) at the vegetative nucleus (VN) of the
pollen grain showing mC lost mainly at CG sites, whereas remains highly methylated at the
spermatic nucleus (Calarco et al., 2012). An epigenetic regulation was described in the pollen,
in which, Athila transposons are activated in the vegetative nucleus and accumulates
correspondent 21 nt siRNAs in the spermatic cells, though to reinforce silencing in the germ
cells (Slotkin et al., 2009). Moreover, targets loci of DNA glycosilases, as epialleles and
imprinted genes, lost CG methylation in the vegetative nucleus being suggested that this
reprogramming in pollen contributes to epigenetic inheritance TE silencing and imprinted by
siRNAs dependent pathways (Calarco et al., 2012). Altogether, these results indicate that
QQS is under an epigenetic regulation.
QQS expression varies during vegetative development and in response to stress
conditions, such as, cold and biotic stress (Hruz et al., 2008; Li et al., 2009; Seo et al., 2011;
Tsutsui et al., 2015). Thus, we used QQS as a probe to explore the interplay between
developmental programs and DNA methylation in establishing a specific expression pattern.
We compared expression and mC levels of QQS contrasting epialleles during Arabidopsis
development and found that contrasting epialleles showed different behaviors indicating that
mC interacts with developmental pathways to modulate QQS expression. The global
expression of QQS is regulated by ROS1 and, moreover, we verified that contrasting QQS
epialleles show different starch accumulation in agreement with the documented negative
regulation of starch metabolism by QQS (Li et al., 2009; Seo et al., 2011; Tsutsui et al.,
2015). Epigenetic regulation is currently considered an important aspect of the interactions
between genotype and environment which defines the phenotype.
38
Results
DNA methylation of QQS shapes its expression pattern during development
QQS is epigenetically regulated by cytosine DNA methylation (mC) and prone to
epiallelic variation (Silveira et al., 2013). In order to investigate the potential of QQS
expression along Arabidopsis development, we analyzed transgenic lines containing the GUS
reporter gene under the control of a region of the QQS promoter and QQS 5’UTR sequence
(QQS:GUS; Fig. 1). GUS activity was detected at all stages of Arabidopsis life cycle starting
in the cotyledons in mature embryos (Fig. 1A). Then, it appears in the shoot and root apical
meristems (SAM), beyond cotyledons, in seedlings (Fig. 1B), in the meristematic region and
rosette young leaves (Fig. 1C), mainly occurring in veins and guard cells (Fig. 1D), and also
in specific regions of the flower (Fig. 1E) and the silique wall (Fig. 1F). GUS activity is
maintained in vegetative SAMs (Fig. 1G and H), especially in the L1 layer of 21 day-old
SAM (Fig. 1H), and in the whole early reproductive SAM (Fig. 1I) and lateral meristems of
the late reproductive SAM (Fig. 1 J). GUS was also detected in pollen, where the vegetative
nucleus is less compacted than the two spermatic ones (Fig. 21-O) which is in agreement with
the QQS expression detected by mRNA-seq analyses (Loraine et al., 2013). These GUS
pattern were very similar to that reported earlier by identical QQS:GUS analyses (Li et al.,
2009), except for a few differences. Thus, promoter and 5’UTR of QQS present a specific
expression potential along plant development, which is in good agreement with public
available expression data of QQS (Fig. S1; Hruz et al. 2008; Li et al., 2009; Seo et al., 2011).
39
Figure 1 – In situ staining of GUS activity in transgenic lines expressing the GUS
reporter gene under the control of QQS promoter and 5’UTR. To evaluate the expression
potential of QQS we analyzed transgenic lines containing the GUS gene fused to the QQS
promoter and 5’UTR (QQS:GUS). Similar results were obtained for six lines and the pictures
show in situ GUS activity from at representative transgenic line, R1-1-3. A) Mature embryos;
B) Seedlings 12 days after germination (DAG); C) Plants with 21 DAG; D) Detail of the
fourth rosette leaf from plants of 21 DAG plants showing GUS activity in veins and guard
cells; E) Flower (60 DAG); F) Histological section of the silique (60 DAG); G) Detail of the
shoot apical meristem (SAM) 12 DAG; H) Histological section of the SAM 21 DAG; I)
Histological section of an early reproductive meristem; J) Histological section of a late
reproductive meristem; K) Detail of pistil and anthers with pollen; L) Histological section of
the anther containing pollen grains; M) GUS activity in pollen; N) DAPI staining of the
vegetative nucleus (VN) and the two spermatic nucleus (SNs) of the pollen; O) Merged M and
N images. Figures A to E, K and L captured in magnifying glass and F to J and M to O in
optic microscope. Scale bars: A = 200 µm; B, E and K = 1 mm; C = 5 mm; D, G, H, I and L=
100 µm; F and J = 500 µm; M = 25 µm.
Next, a quantitative evaluation of the impact of mC on QQS expression profile
during development was undertaken. Since in situ GUS activity appears to vary between
different leaves, with stronger blue staining found in younger leaves than older ones (Fig. 1C
and S2), QQS expression of six contrasting QQS epialelles in young and old rosette leaves
sampled from 40 day-old plants was measured (Fig. 2 and S3-A for details of plant material).
In agreement with previous data (Silveira et al., 2013), an inverse correlation between QQS
mC level and its expression level was observed (Fig. 2A and B). The expression of QQSme
increases significantly as the leaves get older (1,6 fold; Fig. 2A), but parallel changes in DNA
methylation which were quantified by McrBC assays were not detected (Figure 2B). This
expression pattern was progressively inverted with the reduction of methylation of QQS (Fig.
2A and B). For instance, a 40% reduction of mC in the 5’UTR of Col*3-10 as compared to
QQSme
resulted in an 4-7 fold increase of QQS expression and also a reduction of expression
with leaf age (Fig. 2A and B).
40
41
Figure 2 – Expression and DNA methylation profiles of contrasting QQS epialleles in
different organs. A) Absolute expression of QQS epialleles (QQSme
, Col*3-10, Col*3-14,
Col*3-15, Col*3-2, Col*3-16) in internal rosette leaves, considered as younger rosette leaves
(YRL; leaves 13 and 14th
), and external older ones (RL; leaves 6 and 7th
) of 40 day-old plants
(plant material details in Fig. S3A). The absolute QQSme
expression in YRL and RL is
indicated in the figure and were significantly different between them (*, p-value<0.05, t-
Student test). B) McrBC-qPCR analysis showing the DNA methylation level of QQS
promoter and 5’UTR of the different epialleles and in the respective organs used in A. C)
Relative expression of two contrasting epialleles QQSme
and Col*3-2 in different organs of 60
day-old plants. The QQS expression values were normalized separately for each epiallele in
relation to its respective value in silique. Expression of QQS in Col*3-2 is 29-fold larger than
in QQSme
as indicated. The expression of QQSme
in rosette leaf, bract and fruit are
significantly different of the value in flower (*, p<0.05, t-Student test). D) McrBC-qPCR
analysis showing the DNA methylation level of QQS promoter and 5’UTR of two contrasting
epialleles, QQSme
and Col*3-2, in the respective organs used in C. The methylation levels of
QQSme
promoter and 5’UTR in rosette leaf are significantly different from the respective
values in flower (a, p-value<0.05, t-Student test) whereas significant difference for
methylation in the 5’UTR was only found between bract and fruit in relation to flower (b, p-
value<0.05, t-Student test). The data shown in panels A and B represent the mean of
expression and methylation levels from three individual plants while panels C and D the data
represent the mean of three biological replicates each consisting a pool of organs from five
plants (fifteen plants in total). The expression of PDF2 and UBI were used to normalize
expression analyses and values below 40% in McrBC-qPCR analyses were excluded. YRL -
Younger rosette leaves, RL - older rosette leaves, In - internode of the flowering stem. E and
F) DNAs extracted from rosette leaves and flower of 60 day-old plants carrying QQSme
or
Col*3-2 epiallele (same material used in fig. 2A) were used for bisulfite sequencing of QQS.
For each sample, at least 10 clones were sequenced and the cytosine methylation determined
and analyzed by the Kismeth program (Gruntman et al., 2008). G) QQS structure and
methylation profile available from Lister et al. (2008) in which the position of primers used
for previous McrBC-qPCR analyses (red and blue arrows) and bisulfite sequencing (green
arrows) were included. H) The figure shows the position and methylation context of each
cytosine of the QQSme
and Col*3-2 sequences in the analyzed organs. Colored circles
represent methylated cytosines, whereas rings non-methylated cytosines. The colors identify
the cytosin context, CG (red), CHG (blue) and CHH (green) being H any nucleotide except G.
42
The asterisks indicate cytosine positions considered more variable. The sizes of QQS
promoter sequence (256 pb) and 5’UTR (257 pb) analyzed are indicated and correspond the
positions 12.348.789 until 12.349.046 pb and 12.349.047 until 12.349.304 kb (257 pb) of the
third Arabidopsis chromosome, respectively. Details of the plant material used are shown in
fig. S3B. RL – rosette leaves, Fl – flowers.
A more detailed analysis was then performed, in which the expression and mC
levels between organs of 60 day-old plants carrying two contrasting epialleles were compared
(Fig. 2C and D and S3-B for plant material details). As expected, the expression of QQSme
was 29 fold lower than the demethylated Col*3-2 epiallele and more variable between organs
(Figure 2C). The higher expression of QQSme
in rosette leaves and bracts compared to
inflorescence organs (node, internode and flower) was correlated with a lower level of mC at
QQSme
locus in these organs (Fig. 2C and D). In contrast to the QQSme
epiallele, the less
methylated epiallele Col*3-2 presented an almost equal expression level in all organs (Fig.
2C). Together, these data indicate that the level of methylation in the promoter region and
5’UTR are involved in defining the QQS expression pattern. These results were supported by
similar analysis carried out with other demethylated epialleles (Fig.S4-A and B).
Although a negative correlation between methylation and expression levels of
QQS appears to be the rule, this trend was not confirmed for siliques, where QQSme
epialleles
expression was the lowest as compared to other organs and yet, the respective methylation
level in siliques was lower than in flower for instance (Figure 2C and D). Similarly, during
the early stages of seedlings development despite the increase of QQSme
expression, the global
methylation level did not change (Fig S5-A and B). Moreover, in young seedlings, the
expression level was higher in leaves than in cotyledons although the methylation level was
lower in cotyledons (Fig S6-A and B). These results indicate that in addition to mC profiles,
trans regulatory transcriptional factors (TF) may also play a role in defining the expression
pattern of QQS. Until now, IDD14 is the only TF known to regulate QQS and, not
surprisingly, it is weakly expressed at siliques and cotyledons as compared to leaf (Seo et al.,
2011).
Bisulfite sequencing analyses of the genomic DNA from leaves and flower of 60
day-old plants carrying the QQSme
or Col*3-2 epiallele confirmed that methylation is reduced
in older leaves, mainly at CG positions, as compared to inflorescence of 60 day-old QQSme
43
plants (Fig. 2E and F). A qualitatively similar lower methylation essentially in the promoter
sequence was also observed in older leaves relatively to flowers of plants carrying the Col*3-
2 epiallele (Figure 2D and 2F). Col*3-2 was mainly demethylated in the 5’UTR sequence and
the loss of cytosine methylation in its promoter sequence in older leaves occurred at specific
CG and CHG sites different from those occurring in the QQSme
epiallele in older leaves
(Figure 2G and H). For example, Col*3-2 is demethylated at the end of promoter sequence in
contrast to QQSme
(second to sixth asterisk on fig. 2H). More importantly, reduction of
methylation in the promoter of Col*3-2 epiallele in older leaves did not affect QQS
expression when compared to flowers (Figure 2C and D). The data shows that methylation in
the 5’UTR sequence of QQS is crucial in determining the QQS developmental regulation,
while the overall expression level involves both 5’UTR methylation and methylation in the
promoter right upstream of the transcription start site. Moreover, the observation that QQSme
and Col*3-2 present a lower methylation levels in leaves than in organs formed later, such as
flowers, suggests that an active and/or passive demethylation reprograming may occurs from
SAM towards organ formation.
QQSme
expression increases along leaf development and methylation levels decrease
In order to study in more detail QQSme
regulation during leaf development, we
determined QQSme
expression and its methylation profile in samples enriched for shoot apical
meristem (SAM) cells and leaves of 21, 31 and 43 day-old plants, where leaves 1-2 are the
first ones (older leaves) and 11-15 the last ones (younger leaves) to emerge in the rosette (Fig.
3A; details of plant material in Fig. S7). QQSme
expression does not vary significantly
between meristems samples at different ages (p<0.05, t-Student test) but increases
progressively during leaf development. The QQSme
methylation profile determined by
bisulfite sequencing of the respective genomic DNAs (Fig. 3B), revealed that globally the
high mC levels found in QQSme
loci decreases in leaves of 43 day-old plants in comparison to
leaves of 31 day-old plants.
44
Figure 3 – QQSme
expression and methylation profile in enriched samples of SAM and
along leaf development. Meristem samples containing SAM cells and rosette leaves (RL)
numbered according to the order of emergence (RL1-2 the first ones and 11-15 the last ones)
were sampled from plants with 21, 31 and 43 days after germination (DAG; see details in fig.
S7). Total RNA was extracted and QQS expression was measured by RT-qPCR. A) Absolute
expression of QQSme
along leaf development. B) Bisulfite sequencing of one biological
replicate of the respective DNA sample showed in A. For each sample at least 10 clones were
sequenced and the cytosine methylation percentage obtained from Kismeth program
(Gruntman et al., 2008). C to E) Absolute expressions analyses of meristem-specific genes
used as controls of meristem-enrichment cells. AINTEGUMENTA (ANT) is expressed in
meristems being involved in organogenesis (Mizukami and Fischer, 2000), At5g59330 is
exclusively expressed in vegetative meristems (Hruz et al., 2008), whereas LEAFY is more
expressed in reproductive meristems (Siriwardana e Lamb, 2012). These genes present an
opposite expression pattern in accordance with plant transition to a reproductive state 43
DAG. The endogenous controls PDF2 and UBI were used to normalize expression analyses.
For each of the three biological replicate, dissected organs of ten plants were pooled. Similar
result was obtained by two other experiments (Fig. 6 and S11).
45
To complement the analyses of mC impact on QQS expression during leaf
development we also evaluated demethylated QQS epialleles and compared with QQSme
(Fig.
4). As expected, Col*3-2 is always more expressed and possess an opposite behavior than
QQSme
. This result was confirmed in a second experiment and was also observed for other
demethylated epiallele (Fig. S8). Similarly to QQSme
, the expression of Col*3-2 is maintained
at the same levels in SAM enriched tissues of different ages and increase in younger leaves
(Fig. 4). These results show that QQS expression also increases in leaf comparative to SAM
enriched tissues for plants carrying less methylated epialleles. The QQS promoter present at
least nine different types of cis regulatory motifs responsive to light which could be involved
in this expression and methylation regulation in leaf (Fig. S9; Lescot et al., 2002).
Figure 4 – Expression analyses of the demethylated QQS epiallele Col*3-2 along leaf
development. Col*3-2 expression was measured by qRT-PCR in dissected leaves and
meristematic region of plants at different ages (21 and 31 day-old). A) Absolute expression of
Col*3-2 in pools of plant dissected organs. B) Absolute expression of the meristem-specific
gene AINTEGUMENTA used as controls of meristem-enrichment cells. The endogenous
controls PDF2 and UBI were used to normalize expression analyses. The initials RL followed
by numbers means the order of leaf formation in the rosette, being RL1-2 the older ones.
These results support the hypothesis that the QQSme
expression increases and the
methylation levels decreases during leaf development. This demethylation of QQS in leaves
could be an active process through the action of DNA glycosilases (Penterman et al., 2007)
and/or occur passively in the absence of the methylation maintenance machinery (Law and
Jacobsen, 2010). Thus, this aspect was therefore investigated into more details.
46
ROS1 regulates QQSme
expression but is not essential for demethylation in leaves
ROS1 is one of four DNA glycosilases found in Arabidopsis and the major player
acting in excision of mC at specific loci during vegetative development (Penterman et al.,
2007). Thus, we analyzed if ROS1 is involved in the QQSme
demethylation process observed
along leaf development. Firstly, we compared the expression and methylation levels of QQS
in isolated organs of wild-type plants carrying the QQSme
epialelle (wt) and the ros1-4 mutant
(Fig. 5; details of the plant material used in fig. S3, details of ROS1 structure and ros1
mutants in fig. S10). QQS was less expressed in ros1-4 than in wt, ranging between 2-15 fold
decrease in expression depending on organ, and lower expression in ros1-4 was correlated
with higher methylation (Fig. 5). This hypermethylation observed for QQS in ros1-4 mutants
agrees with the data published by Qian et al. (2012) also showing higher methylation levels of
QQS for the same mutant compared to wild-type. Both genotypes presented the highest QQS
expression in leaves (Fig. 5A and B). Both genotypes also exhibited lower levels of
methylation in leaves and bracts as compared to other inflorescence organs (Figure 5B).
These results suggest that DNA demethylation at QQS in older leaves and bracts is partially
independent of ROS1. Other demethylases or passive demethylation may also be involved in
QQS demethylation (Penterman et al., 2007; Law and Jacobsen, 2010), but ROS1 seems to
modulate the overall expression level of QQS.
Figure 5 – Comparison of QQS expression and DNA methylation in different organs of
wild-type plants carrying the QQSme
epiallele and of the ros1-4 mutant. A) Absolute
expression of QQS in 60 day-old plants of wt and ros1-4. Fold change of QQS expression in
each organ between genotypes is indicated. B) Methylation percentage of QQS in the same
organs used in A. Methylation levels of QQSme
promoter in wt rosette leaf are significantly
different from the respective values in flower (a, p-value<0.05, t-Student test), whereas ros1-4
47
rosette leaf show significant difference in both QQS promoter and 5’UTR (b, p-value<0.05, t-
Student test) and ros1-4 bract only in the QQS 5’UTR (c, p-value<0.05, t-Student test) as
compared to the respective values in flower. The samples were collected in five biological
replicates each one constituted by dissected organs from one single plant, similarly to other
experiments (details in Fig. S3). The endogenous controls PDF2 and UBI were used to
normalize expression analyses. Values below 40% in McrBC-qPCR analyses were not
considered. Rosette leaves (RL) and inflorescence stem (IS).
To further evaluate whether QQS expression is dependent upon ROS1, we
determined QQS expression in samples enriched in meristematic tissues and leaves of 21 and
31day-old wt and two ros1 T-DNA insertion mutant, ros1-3 and ros1-4 (Fig. 6; details of
plant material in fig. S7 and details of ros1 mutants in fig. S10). If ROS1 is involved in
demethylation and consequently expression of QQS during leaf development, a different
expression pattern in the mutant as compared to the wt pattern would be expected. The ros1-3
and ros1-4 plants showed lower expression of QQS than in wt plants possibly reflecting a
global higher methylation status of QQS in the mutants. However, QQS expression pattern
was similar between all three genotypes increasing from meristematic regions into leaf
development (Fig. 6A). Curiously, the expression of QQS in 31 day-old rosette leaves relative
to meristem appears to be more pronounced in ros1-3 and ros1-4 than in the wt (Fig. 6A). The
experiment was repeated once with similar results (Fig. S11). This analysis suggests that
ROS1 plays a role in defining global QQS expression but does not participate in defining the
expression pattern during leaf development. This conclusion is in agreement with the notion
that in addition to the methylation profile trans factors are necessary to establish QQS
expression during leaf development.
48
Figure 6 – Comparison of QQS expression in enriched samples of shoot apical meristem
and along leaf development of wild-type, ros1-3 and ros1-4 mutants. Samples enriched in
shoot apical meristem (SAM) cells and leaves 1-2 (first to emerge) and 3-4 (younger leaves)
were sampled from 21 and 31 day-old plants (see details in fig. S7). The samples were
collected in three biological replicates each one constituted by dissected organs of ten plants
pooled. Total RNA was extracted and QQS expression was measured by RT-qPCR. A)
Relative expression of QQS in the different genotypes. QQS expressions in leaves were
normalized for each genotype in relation to the respective value in the meristem. QQS
expression in ros1-3 and ros1-4 was 0.6 and 0.05 lower than the wild-type (wt carrying
QQSme
epiallele), respectively. B) Absolute expression of ROS1 showing null or altered
expression of ros1 mutant plants in relation to wt. C and D) Absolute expression of the
meristem-specific genes AINTEGUMENTA and SHOOTMERISTEMLESS used as controls.
The meristem samples were, as expected, enriched in higher expression of AINTEGUMENTA
and SHOOTMERISTEMLESS, in accordance to their described function (Clark et al, 1996;
Mizukami e Fischer, 2000). The expression of PDF2 and UBI genes were used to normalize
expression analyses. The assay was repeated once as showed in Fig. S11.
49
To further analyze the role of ROS1 in defining QQS expression in leaves, we
crossed wt plants carrying the QQSme
epiallele with ros1-4 to generate double heterozygous
plants (F1; QQSme
/QQSros1-4
and ROS1/ros1-4; where QQSros1-4
is the QQS epiallele present in
ros1-4) and analyzed the QQS expression in the F2 population resulting from selfing of the F1
heterozygous in comparison to the QQS expression in the parental genotypes (Fig. 7). If
ROS1 is not involved in QQS expression in leaves then we should observe the same average
of QQS expression in leaves of the three possible ROS1 genotypes (i.e. ROS1/ROS1,
ROS1/ros1-4 and ros1-4/ros1-4) as a consequence of the random combination of each
parental QQS epialelle. The results showed that QQS expression does not follow this
expectation, instead ros1-4 plants presented QQS expression level of the parental ros1-4,
whereas, the heterozygote ROS1/ros1-4 showed and intermediate level between parents and
the homozygotes wt ROS1 in which the expression levels was higher and closer to the female
wt parental leaves (Fig. 7 and Fig. S12).
This result clearly demonstrates that ROS1 determines the overall expression level
of QQS in leaves and suggest that ROS1 acts in a dynamic equilibrium, perhaps, protecting
the loci against hypermethylation. Since ros1-4 always present lower levels of QQS
expression than wt plants (Fig. 5 and 6), it seems that ROS1 regulate QQS expression in the
whole organism. Furthermore, we found that QQS is also regulated by DEMETER (DME)
during gametogenesis (Fig. S15; Ibarra et al., 2012) since its locus shows lower DNA
methylation levels in the vegetative nucleus of the pollen in comparison with sperm cells and
this demethylation is partially resumed in dme/+ heterozygous pollen. These results indicate
that QQS is under epigenetic regulation during both vegetative and reproductive development
and reinforce that it is a target of DNA glycosilases.
50
Figure 7 – Analyses of QQS expression in leaves of a F2 segregating population resulting
from selfing of ROS1/ros1-4 heterozygote. Box plot representation of QQS expression in
different related genotypes. In order to evaluate the influence of ROS1 in QQS expression,
wild-type plants (wt, female parent) was crossed with homozygous ros1-4 (male parent),
generating the double heterozygous QQSme
/QQSros1-4
and ROS1/ros1-4, and the 6 to 9th
rosette
leaves of 69 plants of the segregating F2 population were analyzed in comparison to the
parental genotypes (F0). The analysis shows that QQS expression is different between the
different ROS1 genotypes (ROS1/ROS1, ROS1/ros1-4 and ros1-4/ros1-4) indicating that
ROS1 controls QQS expression. The number of each plant genotype analyzed is given in
parentheses. The ros1-4 mutant carry a T-DNA insertion (Penterman et al., 2007), thus, the
plants genotypes (ROS1/ROS1, ROS1/ros1-4 and ros1-4/ros1-4) were determined by PCR
(Fig. S10-C). The proportion of the different ROS1 genotypes fits 1:2:1 proportion (χ² =
1,1739, p<0.05). QQS expression was measured by RT-qPCR with RNA extracted from
rosette leaves 6 to 9 of individual plants (31 day-old). The parental wt and ros1-4 (F0) were
analyzed in four biological replicates and the endogenous controls PDF2 and UBI were used
to normalize expression analyses. The expression values were calculated in relation to ros1-4
and grouped according the ROS1 genotype. The outlier values of the box plot analysis were
represented by ж. The individual expression value of QQS in each plant is shown in Fig. S12.
51
Plants carrying contrasting QQS epialleles show differences in starch accumulation
Contrasting epialleles present differential expression behaviors during plant
development, thus, we revisited the effect of QQS epigenetic variation on starch
accumulation. Based on Arabidopsis transgenic lines overexpressing or silencing QQS, as
well as, transgenic lines overexpressing QQS in other species, this gene has been shown to be
a negative regulator of starch accumulation (Li et al., 2009; 2015a; Li and Wurtele, 2014; Seo
et al., 2011; Tsutsui et al., 2015). We also demonstrated in here that several QQS epialleles
present increased expression correlating with lower methylation, which is in agreement with
previous results including natural accessions (Silveira et al., 2013).
Plants carrying the demethylated and highly expressed QQS epialleles, Col*3-2
and Col*3-16, the methylated and lower expressed QQSme
epiallele and the T-DNA/TE
insertion qqs-1 were compared for their starch accumulation at the end of the day and end of
night under long day conditions (16h light/8h dark; Fig. 8). The two highly expressed
demethylated epialleles accumulated around two times less starch than the QQSme
genotype at
the end of the day (Figure 8). The qqs-1 mutant plants presented similar starch content than
plants carrying QQSme
(p-value<0.05). Two other experiments carried out with 9 day-old
seedlings and 42 day-old plants cultivated in the same long day conditions showed similar
results (Fig. S13). At night, starch is degraded in such a way as to sustain growth until dawn
(Graf et al., 2010) and it is therefore likely that the demethylated epialleles which have less
starch to be used at night end up with lower starch amounts at dawn than the QQSme
a
prediction which actually happened (Fig. 8). Together, the data clearly indicate that QQS
epiallelic status can modulate starch accumulation.
52
Figure 8 – Starch accumulation in plants carrying contrasting QQS epialleles and the
mutant qqs-1. The figure shows that plants carrying the less expressed QQS epiallele
(QQSme
) or the qqs-1 T-DNA/TE mutant insertion accumulate more starch at the end of day
and end of the night than plants carrying more expressed epialleles Col*3-2 and Col*3-16.
The significant differences in relation to QQSme
(p-value<0.05, t-Student test) are marked
with asterisk (*). Plants were cultivated for 31 days under long day conditions (16 hour of
light and 8 of night) and the starch analysis was made with 5 biological replicates containing
6 plants for each genotype (30 plants in total).
Discussion
In general, mC is considered a repressive mark of gene transcription but this
relationship is complex and sometimes controversial, since gene body methylation, for
example, does not seems to affect expression (Zilberman et al., 2007; Zhang et al., 2010).
Thus, a relevant question would be if methylation reprogramming occurs during development
and affects gene expression. In this context, few examples are described, such as,
FLOWERING WAGENINGEN (FWA), MEDEA and FERTILIZATION-INDEPENDENT
SEED 2, being mainly observed in the gametogenesis and important for imprinting processes
(Hsieh et al., 2009; Ikeda, 2012). However, although examples of epialleles affecting
morphological traits are also described, such as, Lcyc and CmWIP1 (Cubas et al, 1999; Martin
et al., 2009), how mC interacts with vegetative developmental signals and modulates gene
expression is an aspect poorly understood.
In order to obtain new insights about this, we compared the expression and
methylation levels of contrasting QQS epialleles during A. thaliana development. We showed
that the profile of mC at the QQS locus is a key element to define the expression pattern of
this gene at different stages of the vegetative development. For example, demethylated
epialleles presented a more homogeneous expression pattern between organs as compared to
the methylated epialleles (Fig. 2C and D) and, while the expression of the demethylated
epialleles decreases during leaf development an opposite profile holds for the methylated
epiallele (Fig. 2A and B, 3A and 4A). The main differences of mC levels between epialleles
were found in the CG context of the 5’UTR and proximal to the transcription start (Fig. 2)
suggesting that methylations of these regions are important in defining alteration of the QQS
expression pattern.
53
The methylated QQS epiallele (QQSme
) is more expressed and less methylated in
rosette leaves and bracts compared to other organs formed later during development (Fig. 2C
and D). This result suggested that loss of methylation occurs during leaf development.
Furthermore, demethylation of QQSme
is dependent on plant age. Lower levels of methylation
correlated with higher expression levels were only observed in older leaves of 43 and 60 day-
old plants (Fig. 3B and 2E, respectively). In older leaves of 40 day-old plants the expression
of QQS increases in relation to younger leaves but no parallel demethylation was observed
(Fig. 2A and B). Thus, in addition to mC, trans regulatory factors are also important to define
QQS expression pattern. Other evidences for this conclusion include the facts that QQS
expression is altered in siliques (Fig. 2C and D), seedlings (Fig. S5-A and B) and cotyledons
(Fig.S6-A and B) without correspondent expected changes in DNA methylation. A potential
transcription factor (TF) is IDD14 and the expression pattern of QQS and IDD14 are
correlated (Hruz et al. 2008; Seo et al., 2011).
Demethylation of QQS in leaves of older plants could be the result of the DNA
glycosilase activity of ROS1, which is known to act mainly in vegetative tissues (Penterman
et al., 2007). We showed that ros1-4 plants, present a reduction of the QQS methylation levels
in leaves and bracts in comparison to flower of 60 day-old plants (Fig. 5). This result was
similar to wt plants, although the methylation of QQS in ros1-4 was slightly higher compared
to wt, thus, we concluded that QQS demethylation in leaves is, at least, partially independent
of ROS1. Nevertheless, the QQS expression level was strongly reduced in ROS1 mutant
background during leaf development (Fig. 6) suggesting that ROS1 is an important
determinant of the overall expression level of QQS.
Through the crossing of wt and ros1-4 plants (F1 heterozygote) and analyses of
the F2 segregating population originated from selfing of the F1 (Fig. 7), we demonstrate that
indeed QQS regulation is targeted by ROS1. If ROS1 would not be involved in the regulation
of QQS, a mendelian proportion of parental QQS epialelles combinations should be observed
to each ROS1 genotype (i.e. ROS1/ROS1, ROS1/ros1-4, ros1-4/ros1-4) or, in other words, the
same expression average between the genotypes should be found. As expected, the mendelian
proportion was achieved for the ros1 alleles. However, the QQS expression did not follow this
rule, since different averages of expression were found in the different ros1 genotypes
suggesting that different epigenetic alterations affect QQS expression according to the ros1
genotype. The F2 ros1-4 homozygotes led to a drastic reduction of QQS expression (15 folds
less), since QQS expression of the parental ros1-4 was completely resumed and little variation
54
was observed among the ros1-4 individuals. The F2 ROS1 homozygotes showed higher QQS
expression averages and also more variability among the different individuals compared to
ros1-4 mutants indicating that the levels of QQS expression are not readily re-established to
the wt parental level when wild-type ROS1 is resumed. Moreover, QQS expression in
heterozygous plants (ROS1/ros1-4) is lower than wt pointing a dosage-dependent effect of
ROS1 in QQS regulation. Thus, to explain these results we suggest that the loss of ROS1
wild-type activity leads to a QQS hypermethylation, probably, as a consequence of the
antagonistic effect of the RdDM (Williams et al., 2015), and consequently, it alter QQS
expression. This hypermethylation is not readily decreased with the resume of the wild-type
ROS1 copies, which could indicate that the methylation pathways are more efficient than the
demethylation pathways. Other methylation analyses by bisulfite sequencing are being
performed to confirm this suggestion.
Additionally, QQS shows hypermethylation in the triple ros1, dml2 and dml3
(rdd) mutant plants (Cokus et al., 2008; Lister et al., 2008) and also in the single ros1-4
mutant plants (Gong et al., 2002; Agius et al., 2006; Zhu et al., 2007; Qian et al., 2012).
Moreover, a differential methylated region (DMR) is localized at the QQS locus in plants
infected with Pseudomonas syringae pv. Tomato (Pst; in Table S3 of Dowen et al., 2012) and
ROS1 removes methylation and activates expression of a set of genes in response to Pst (Yu
et al., 2013). Together, this data strongly suggest that QQS is a target of ROS1, whose
activity, probably, avoids hypermethylation in the locus.
To reconcile the ROS1 independent demethylation of QQS in leaves and the fact
that ROS1 regulates QQS expression, three hypotheses can be considered. First, it is possible
that other DNA glycosilases, such DML2 and 3, acts redundantly with ROS1 (Penterman et
al., 2007). Second, the demethylation could be achieved passively as a consequence of
reduced maintenance by MET1 and/or other methyltransferases. Accordingly, it was shown
that the expression of methylation maintenance genes is decreased in leaves during
senescence (i.e. aging) and, also, in comparison to SAM (Fig. S14). Third, the loci occupancy
by TFs could preclude by competitiveness the action of enzymes involved in de novo or
maintenance methylation and cause demethylation (Lin et al., 2000; Wang et al., 2012). Thus,
the QQSme
demethylation in leaves could be a consequence of the induced transcription,
maybe in response to light (Fig. S9). In agreement, Zilberman et al. (2007) reported that small
transcribed genes are highly biased to lose methylation.
55
Recently, an antagonistic relationship between ROS1 and RdDM to balance
cellular mC levels was described (Lei et al., 2015; Williams et al., 2015). ROS1 and RdDM
share components, such as small RNAs dependent on Pol IV and V, showing that these
pathways are intrinsically related (Mosher et al., 2008; Zheng et al., 2008; Matzke and
Mosher, 2014). Thus, ROS1 could act in a dynamic equilibrium protecting specific loci
against hypermethylation and how this equilibrium is disrupted constitute an important issue
to understand how epialleles are formed. Methyl binding proteins (MBPs) could be an
important part of this process, since they are able to bind to CG methylated sites and recruit
chromatin remodeler factors to facilitate ROS1 action (Zemach and Grafi, 2007; Lang et al.,
2015). Thus, DNA methylation could be interpreted as traffic lights attracting other activator
or repressor complexes of gene transcription (Medvedeva et al., 2014; Li et al., 2015c). This
could explain why the methylated QQS epiallele and ros1 mutants-associated QQS epialleles
share the same expression pattern during leaf development but this pattern is lost in
demethylated QQS epialelles (Fig. 2A, 3A, 4 and 6A). Since, our results indicate that ROS1
influences QQS expression, ROS1 could be involved in generating QQS epiallelic states,
possibly, throughout de disturbance of this methylation dynamic equilibrium and regulation of
the methylation status at early stages after fertilization. The newly acquired QQS epigenetic
state could then be stably maintained and inherited, an issue that have to be addressed in
future studies.
Besides the QQS regulation during vegetative development by ROS1, QQS is also
regulated by DEMETER (DME) during gametogenesis (Fig. S15; Ibarra et al., 2012). QQS is
demethylated at the vegetative nucleus (VN) of the pollen grain showing mC lost mainly at
CG sites, whereas remains highly methylated at the spermatic nucleus (Fig. S15; Calarco et
al., 2012). This demethylation at the VN are partially recovered in demeter (dme/+) mutant
plants (Ibarra et al., 2012) indicating the participation of DME in this process. GUS activity
found in pollen (Fig. 1M to O) is in agreement with RNA-seq analysis showing QQS
expression in pollen (Loraine et al., 2013). Thus, QQS seems to be under a similar epigenetic
regulation during the reproductive development.
Although the meaning of QQS demethylation in pollen and its biological function
is unknown, during vegetative development QQS seems to impact starch accumulation (Fig.
8), which is in agreement with other studies showing that QQS act as a negative regulator of
starch metabolism (Li et al., 2009; Seo et al., 2011; Tsutsui et al., 2015). In addition to the
fact that starch was shown to be crucial to sustain growth at night (Stettler et al., 2009), an
56
inverse correlation between starch accumulation and biomass has been uncovered by the
analyses of a set of Arabidopsis accessions (Sulpice et al., 2009). These observations
emphasize the importance of starch in modulating growth and biomass acquisition. Arendsee
et al. (2014) suggest that QQS integrates a range of environmental signals to regulate plant
growth under biotic and abiotic stress conditions via starch regulation. Thus, QQS epiallelic
variation may indeed play a role in the control of growth and consequently may affect fitness.
Material and methods
Plant material
The analyses were performed in Arabidopsis thaliana Columbia-0 (Col-0) ecotype containing
contrasting QQS epialleles, independent transgenic linages of QQS:GUS, the T-DNA
insertion WiscDsLoxHs077 (in which QQS is disrupted at the third exon and referred here as
qqs-1, as previously described by Silveira et al., 2013) and the T-DNA mutants ros1-3 and
ros1-4 (SALK_45303, described by Penterman et al., 2007). Transgenic linages and mutant
plants were confirmed by PCR and sequencing of the T-DNA insertions in comparison to
wild-type plants as following described. The list of primers used is shown in table S1. For the
diverse experiments the seeds were planted in soil (2:1 Plantmax HT® and vermiculite)
stratified for 72 hours at 4ºC in darknessAlternatively, some experiments used seedlings
grown in MS/2 medium as indicated in the respective figure. In this case, approximately 5 mg
of seeds were sterilized (1’ in 70% etanol, 20’ in 1% sodium hypochlorite under shaking and
followed by four washes in milli-Q autoclaved water), maintained for 72 hours at 4ºC in
darkness and then cultivated in liquid MS/2 (Murashige and Skoog, 1962) supplemented with
0.5% sucrose. Embryo tissue was obtained from macerated seeds as described by Raissig et
al. (2013). All the experiments were performed in long day conditions (16 hours of light and 8
of darkness) at 22º C and at least three biological replicates were collected directly in liquid
nitrogen.
Nucleic acid extractions
Total RNA was extracted by the lithium chloride method (Oñate-Sánches and Vicente-
Carbajosa, 2008) and the DNA by de CTAB method (Doyle and Doyle, 1990) or, in
57
substitution, using the method described by Oliveira et al. (2015), in which RNA and DNA
can be obtained simultaneously from different tissues without any differences in comparison
with the previous cited protocols. RNAs of plants cultivated in MS/2 culture medium were
isolated by the guanidine extraction method (Logemann et al., 1987). All the nucleic acids
extractions were verified by agarose gel electrophoresis, TAE buffer for DNA and MOPS for
RNA (Sambrook et al, 1989), and showed no degradation signals.
Genotyping of T-DNA mutant plants
The T-DNA insertion present in the mutant plants ros1-3, ros1-4 and qqs-1 were verified by
PCR DNA amplification (primers listed in table S1) using the GoTaq® DNA Polymerase
(Promega) followed by DNA separation on agarose gel electrophoresis and comparison with
wild-type plants. Purification of agarose gel bands were made by Wizard® SV Gel and PCR
Clean-Up System kit (Promega), cloned into pGEM T-easy vector (Promega) and sequencing
with BigDye® Terminator v3.1 (ThermoFisher Scientific). All the procedures followed the
manufacturer’s instructions. ROS1 structure, T-DNA insertion ros1-4 and the agarose gel
containing the PCR genotyping are available in in fig. S10.
Arabidopsis transgenic lineages and GUS activity analyses
Amplification of QQS promoter and 5’UTR was made by PCR using the Platinum® Taq
DNA Polymerase High Fidelity (ThermoFisher Scientific) and genomic DNA as template
(primers listed in table S1). The promoter sequence length (320 bp) to be cloned was defined
by the presence of a transposons and the analyses were similar to Li et al. (2009) allowing
comparison. The fragment was purified from agarose gel, cloned and sequenced as described
above. The 35S:GUS:tNOS insert was obtained from the pBI121 vector (Chen et al., 2003)
and cloning into the binary vector pFP101-HA (http://www.isv.cnrs-
gif.fr/jg/alligator/vectors.html). The final QQS:GUS cassette was obtained from the enzyme
digestion and substitution of 35S promoter present in the 35S:GUS:tNOS of the pFP101
vector by the QQS fragment extracted from pGEM T-easy vector (details in fig. S16). All the
plasmids were grown in Escherichia coli (DH5α) and the minipreps extractions, enzymes
digestion, Klenow enzyme treatment and T4 ligation necessary for the cloning procedures was
performed following Sambrook et al. (1989) instructions. The fragments of each step of the
58
cloning plasmids were confirmed by enzyme digestions and PCRs. The binary vector was
transfected into Agrobacterium tumefaciens GV3101::pMP90 (Koncz et al., 1986) following
the method described by Brasileiro and Carneiro (1998) and then, used for in planta
transformation method (Bechtold et al., 1993) modified by Silveira et al. (2007) of the A.
thaliana Col-0 ecotype carrying the QQSme
epiallele. Homozygous linages were selected by
seed GFP fluorescence in the F3 generation. GUS activity analyses in transgenic plant organs
were performed as described by Vitha et al. (1995). Samples were directly photographed
using a magnifying glass or optical microscopy as described by Oliveira et al. (2014).
RT-qPCR expression and McrBC DNA methylation analyses
The RT-qPCR and McrBC DNA methylation analyses were performed as described by
Silveira et al. (2013) using nucleic acids extracted as described above. The primers used for
RT-qPCR were designed in an exon-exon region avoiding DNAse treatment (Sambrook et al.,
1989; primers listed in Table S1). Real time PCR reactions were run on an Applied Biosystem
7500 Real-Time PCR System (ThermoFisher Scientific) or CFX384 Touch™ Real-Time PCR
Detection System (BIO-RAD). Gene expression levels were normalized in relation to the
reference genes PDF2/PP2A (At1g13320) and ACTIN2 (Act2, At3g18780) and/or
UBIQUITIN E2 (Ubq, At2g36060) as suggested by Czechowski et al. (2005). The absolute
expression levels were calculated by the 2-∆CT
*100 formula (Pfaffl, 2004) and the relative
expression by the 2-∆∆CT
formula (Livak et al., 2001).
DNA sodium bisulfite sequencing
Primers were designed using Methyl Primer Express® software v1.0 (Applied Biosystems)
and/or MethPrimer (Li and Dahiya, 2002) and are listed in table S1. DNA samples were
treated with the EZ DNA Methylation-Lightning™ or -Gold™ kit (Zymo Research) and used
for PCR reactions with Platinum® Taq DNA polymerase (ThermoFisher Scientific). Primers
designed for genomic amplification were also used in the PCR with the bisulfite treated DNA
as a negative control and non-treated DNA as a positive control. After separation on a 1%
agarose gel, the DNA amplifications were purified using the Wizard® SV Gel and PCR
Clean-Up System kit (Promega) and cloned into pGEM T-easy vector (Promega). The vectors
were transformed into E. coli (DH5α) and at least ten clones were selected for miniprep
59
extractions and BigDye® Terminator v3.1 (ThermoFisher Scientific) sequencing. All the
procedures followed Sambrook et al. (1989) and the reagents manufacturer’s instructions.
Sequencing results were analyzed with the KisMeth software (Gruntman et al., 2008). As
internal controls, we analyzed the genes At5g13440 absent for cytosine methylation and FWA
always presenting around 30% of cytosine methylation in vegetative tissues (Fujimoto et al.,
2008). Thus, the bisulfite DNA conversion efficiency was measured by the non-methylated
cytosines levels of At5g13440 and values below 99% were not considered.
Starch quantification
Starch levels were measured in the aerial part (rosette leaves) of 31 day-old plants grown in
long day conditions. The analysis was performed with 5 biological replicates containing 6
plants for each genotype (30 plants in total). It was used the method described by Stitt et al.
(1989) and modified by Hendriks et al. (2003) with 20-25 mg of each sample macerated in
liquid nitrogen and.
References
Agius F et al. (2006). Role of the Arabidopsis DNA glycosylase lyase ROS1 in active DNA
demethylation. PNAS, 103(31):11796–11801.
Arendsee Z et al. (2014). Coming of age: orphan genes in plants. Trends in Plant Science,
19(11): 698–708.
Bechtold N et al. (1993). In plat Agrobacterium mediated gene transfer by infiltration of adult
Arabidopsis thaliana plants. C. R. Acad. Sci., 316:1194-1199.
Berretta J and Morillon A (2009). Pervasive transcription constitutes a new level of
eukaryotic genome regulation. EMBO Rep., 10(9):973–982.
Brasileiro ACM and Carneiro VTC. (1998). Manual de transformação genética de plantas.
Brasília: Embrapa-SPI, p. 75-92.
Calarco JP et al. (2012) Reprogramming of DNA Methylation in Pollen Guides Epigenetic
Inheritance via Small RNA. Cell, 151:194–205
60
Chen P et al. (2003) Complete sequence of the binary vector pBI121 and its application in
cloning T-DNA insertion from transgenic plants. Molecular Breeding, 11:287–293.
Choi et al. (2012). DEMETER, a DNA Glycosylase Domain Protein, is Required for
Endosperm Gene Imprinting and Seed Viability in Arabidopsis. Cell, 110:33–42.
Clark SE et al. (1996). The CLAVATA and SHOOT MERISTEMLESS loci competitively
regulate meri-stem activity in Arabidopsis. Development, 122:1567–1575.
Cortijo et al. (2014). Mapping the epigenetic basis of complex traits. Science,
343(6175):1145-8.
Cubas P et al. (1999). An epigenetic mutation responsible for natural variation in floral
symmetry. Nature 401:157–161.
Czechowski T et al. (2005) Genome-wide identification and testing of superior reference
genes for transcript normalization in Arabidopsis. Plant Physiology, 139:5-17.
Dowen et al. (2012). Widespread dynamic DNA methylation in response to biotic stress. Proc
Natl Acad Sci USA, 109(32):E2183-91.
Doyle JJ and Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12:13-15.
Dubin et al. (2015). DNA methylation in Arabidopsis has a genetic basis and shows evidence
of local adaptation. Elife, 4:e05255.
Feng, S et al. (2010). Epigenetic Reprogramming in Plant and Animal Development. Science,
330(6004):622-627.
Feng, S et al. (2011). Epigenetic modifications in plants: an evolutionary perspective. Curr.
Opinion in Plant Biology, 14(2):179-86.
Fujimoto R et al. (2008). Evolution and control of imprinted FWA genes in the genus
Arabidopsis. Plos Genetics, 4:1-8.
Goll MG and Bestor TH (2005). Eukaryotic cytosine methyltransferases. Annu. Rev.
Biochem., 74:481–514.
Gong Z et al. (2002). ROS1, a repressor of transcriptional gene silencing in Arabidopsis,
encodes a DNA glycosylase/lyase. Cell, 111(6):803–814.
61
Gruntman E et al. (2008). Kismeth: Analyzer of plant methylation states through bisulfite
sequencing. BMC Bioinformatics, 9:371.
Hauben et al. (2009). Energy use efficiency is characterized by an epigenetic component that
can be directed through artificial selection to increase yield. Proc Natl Acad Sci USA.
106(47):20109-14.
Henderson IR and Jacobsen SE (2007). Epigenetic inheritance in plants. Nature, 447:418-424.
Hendriks JH et al. (2003). ADPglucose pyrophosphorylase is activated by posttranslational
redox modification in response to light and to sugars in leaves of Arabidopsis and other
plant species. Plant Physiology, 133:838–849.
Hsieh T et al. (2009). Genome-Wide Demethylation of Arabidopsis Endosperm. Science,
324(5933):1451-1454.
Hruz T et al. (2008). Genevestigator V3: a reference expression database for the meta-
analysis of transcriptomes. Advances in Bioinformatics, 420747.
Ibarra et al. (2012). Active DNA Demethylation in Plant Companion Cells Reinforces
Transposon Methylation in Gametes. Science, 337:1360.
Ikeda Y (2012). Plant imprinted genes identified by genome-wide approaches and their
regulatory mechanisms. Plant Cell Physiol., 53: 809–816.
Koncz C et al. (1986) High-frequency T-DNA-mediated gene tagging in plants. Proc. Natl.
Acad. Sci. USA, 86:8467-8471.
Kooke et al. (2015). Epigenetic basis of morphological variation and phenotypic plasticity in
Arabidopsis thaliana. Plant Cell, 27(2):337-48.
Lang Z et al. (2015). The methyl-CpG-binding protein MBD7 facilitates active DNA
demethylation to limit DNA hyper-methylation and transcriptional gene silencing. Mol
Cell, 57(6):971-83.
Law J and Jacobsen SE (2010). Establishing, maintaining and modifying DNA methylation
patterns in plants and animals. Nature Reviews Genetics, 11(3): 204-220.
Le T-N et al. (2014). DNA demethylases target promoter transposable elements to positively
regulate stress responsive genes in Arabidopsis. Genome Biology, 15:458.
62
Lei M et al. (2015). Regulatory link between DNA methylation and active demethylation in
Arabidopsis. Proc. Natl. Acad. Sci., 112(11):3553-7.
Lescot et al., (2002). PlantCARE, a database of plant cis-acting regulatory elements and a
portal to tools for in silico analysis of promoter sequences. Nucleic Acids Res.,
30(1):325-7.
Li LC and Dahiya R (2002). MethPrimer: designing primers for methylation PCRs.
Bioinformatics. 2002 Nov;18(11):1427-31.
Li L and Wurtele ES (2014). The QQS orphan gene of Arabidopsis modulates carbon and
nitrogen allocation in soybean. Plant Biotechnol J., 13(2):177–187.
Li L et al. (2009) Identification of the novel protein QQS as a component of the starch
metabolic network in Arabidopsis leaves. Plant Journal, 58:485-498.
Li L et al. (2015a). QQS orphan gene regulates carbon and nitrogen partitioning across
species via NF-YC interactions. Proc. of the Nat. Acad. of Sci., 112(47:14734-14739.
Li Q et al. (2015b). Examining the causes and consequences of context-specific differential
DNA methylation in maize. Plant Physiol., 168(4):1262-74.
Li Q et al. (2015c). Regulation of Active DNA Demethylation by a Methyl-CpG-Binding
Domain Protein in Arabidopsis thaliana. PLoS Genet., 11(5):e1005210.
Lin IG et al. (2000) Modulation of DNA binding protein affinity directly affects target site
demethylation. Mol Cell Biol., 20:2343–2349.
Livak KJ and Schmittgen TD (2001). Analysis of Relative Gene Expression Data Using Real-
Time Quantitative PCR and the 2-∆∆CT
Method. METHODS, 25:402–408.
Loraine AE et al. (2013) RNA-Seq of Arabidopsis Pollen Uncovers Novel Transcription and
Alternative Splicing. Plant Physiol., 162(2):1092–1109.
Martin et al. (2009). A transposon-induced epigenetic change leads to sex determination in
melon. Nature, 461(7267):1135-8.
Matzke M et al. (2009). RNA-mediated chromatin-based silencing in plants. Curr. Opin. Cell
Biol., 21:367–376.
63
Matzke MA and Mosher RA (2014). RNA-directed DNA methylation: an epigenetic pathway
of increasing complexity. Nature Reviews Genetics, 15(6):394–408.
Medvedeva et al. (2014). Effects of cytosine methylation on transcription factor binding sites.
BMC Genomics, 15:119.
Mirouze M and Paszkowski J (2011). Epigenetic contribution to stress adaptation in plants.
Current Opinion in Plant Biology, 14:267–274.
Mizukami Y and Fischer RL (2000). Plant organ size control: AINTEGUMENTA regulates
growth and cell numbers during organogenesis. PNAS, 97(2):942-947.
Morales-Ruiz et al. (2006). DEMETER and REPRESSOR OF SILENCING 1 encode 5-
methylcytosine DNA glycosylases. PNAS, 103(18):6853–6858.
Mosher RA et al. (2008). PolIVb influences RNA-directed DNA methylation independently
of its role in siRNA biogenesis. Proc. Natl. Acad. Sci. USA, 105:3145–50.
Mutskov and Felsenfeld (2004). Silencing of transgene transcription precedes methylation of
promoter DNA and histone H3 lysine 9. EMBO J, 23(1):138–149.
Oliveira et al. (2014). Flower development in Coffea arabica L.: new insights into MADS-box
genes. Plant Reprod., 27(2):79-94.
Oliveira et al. (2015). An efficient method for simultaneous extraction of high-quality RNA
and DNA from various plant tissues. Genet. Mol. Res., 14(4):18828-18838.
Oñate-Sánchez L and Vicente-Carbajosa J (2008). DNA-free RNA isolation protocols for
Arabidopsis thaliana, including seeds and siliques. BMC Research Notes, 1:93.
Penterman J et al. (2007). DNA demethylation in the Arabidopsis genome. PNAS, 104(16):
6752–6757.
Pfaffl MW (2004). Quantification strategies in real-time PCR. Chaper 3: 87-112 in: A-Z of
quantitative PCR (Editor: S.A. Bustin), International University Line (IUL), La Jolla,
CA, USA.
Qian W et al. (2012). A Histone Acetyltransferase Regulates Active DNA Demethylation in
Arabidopsis. Science, 336(6087):1445–1448.
64
Raissig MT et al. (2013). Efficient and rapid isolation of early-stage embryos from
Arabidopsis thaliana seeds. J. Vis. Exp.. 76:e50371.
Sambrook J et al. (1989) Molecular Cloning: A Laboratory Manual, vol. I. 2° edition. Cold
Spring Harbor Laboratory Press. ISBN 0-87969-309-6.
Schuettengruber B et al. 2007. Genome Regulation by Polycomb and Trithorax Proteins. Cell,
128:735–745.
Seo PJ et al. (2011). Two splice variants of the IDD14 transcription factor competitively form
nonfunctional heterodimers which may regulate starch metabolism. Nature
communications, 2:303.
Schoft VK et al. (2011). Function of the DEMETER DNA glycosylase in the Arabidopsis
thaliana male gametophyte. Proc. Natl. Acad. Sci. USA. 108:8042–8047.
Silveira AB et al. (2013). Extensive Natural Epigenetic Variation at a De Novo Originated
Gene. PLoS Genetics, 9(4):e1003437.
Slotkin RK et al. (2009) Epigenetic reprogramming and small RNA silencing of transposable
elements in pollen. Cell, 136(3):461-72.
Stitt M et al. (1989). Metabolite levels in specific cells and subcellular compartments of plant
leaves. Methods Enzymology, 174:518–552.
Stroud H et al. (2013). Comprehensive Analysis of Silencing Mutants Reveals Complex
Regulation of the Arabidopsis Methylome. Cell, 152:352–364.
Teixeira FK and Colot V (2009) Gene body DNA methylation in plants: a means to an end or
an end to a means? The EMBO Journal, 28(8):997-8.
Teixeira FK and Colot V (2010). Repeat elements and the Arabidopsis DNA methylation
landscape. Heredity, 105:14–23.
Tsutsui et al. (2015). The Plant-Specific RAB5 GTPase ARA6 is Required for Starch and
Sugar Homeostasis in Arabidopsis thaliana. Plant Cell Physiol., 56(6):1073-83.
Turck and Coupland (2014). Natural variation in epigenetic gene regulation and its effects on
plant developmental traits. Evolution, 68 (3):620–631.
65
Vire et al. (2006). The Polycomb group protein EZH2 directly controls DNA methylation.
Nature, 439(7078):871-4.
Vitha S et al. (1995). Histochemical GUS Analysis. In Agrobacterium Protocols, K.M.A.
Gartland and M.R. Davey, eds (Totowa, NJ: Humana Press), pp. 185-193.
Wade JT and Grainger DC (2014). Pervasive transcription: illuminating the dark matter of
bacterial transcriptomes. Nature Reviews Microbiology, 12:647–653.
Wang H et al. (2012). Widespread plasticity in CTCF occupancy linked to DNA methylation.
Genome Res., 22(9):1680–1688.
Williams BP et al. (2015). Methylation-Sensitive Expression of a DNA Demethylase Gene
Serves As an Epigenetic Rheostat. PLoS Genet., 11(3):e1005142.
Yu A et al. (2013). Dynamics and biological relevance of DNA demethylation in Arabidopsis
antibacterial defense. Proc Natl Acad Sci U S A. 2013 Feb 5; 110(6):2389–2394.
Zemach A and Graf G. (2007). Methyl-CpG-binding domain proteins in plants: interpreters of
DNA methylation. Trends Plant Sci., 12(2):80-5.
Zemach et al. (2013). The Arabidopsis nucleosome remodeler DDM1 allows DNA
methyltransferases to access H1-containing heterochromatin. Cell, 153(1):193-205.
Zhang et al. (2010). DNA cytosine methylation in plant development. J Genet Genomics,
37(1):1-12.
Zheng et al. (2008). ROS3 is an RNA-binding protein required for DNA demethylation in
Arabidopsis. Nature, 455(7217):1259–1262.
Zhu J et al. (2007). The DNA glycosylase/lyase ROS1 functions in pruning DNA methylation
patterns in Arabidopsis. Curr. Biol., 17:54–59.
Zhu JK (2009). Active DNA demethylation mediated by DNA glycosylases. Annu. Rev.
Genet., 43:143–166.
Zilberman et al. (2007). Genome-wide analysis of Arabidopsis thaliana DNA methylation
uncovers an interdependence between methylation and transcription. Nat. Genet.,
39:61–69.
66
Material suplementar do manuscrito
Supplementary material
Table S1 – Primer list. List of primers used for different assays. References of already
described primers (Ref.): a - Silveira et al. (2013); b - Modified from Zhu et al. (2007); c -
Penterman et al. (2007); d - Salk Primer L (Isect tool).
Gene (AGI)
Primer
direction Primer sequence Assay Ref.
QQS (At3g30720) Forward AAGACCAATAGAGAGCAGGAA
RT-qPCR
a
Reverse CCTGATGTAGAAGTGTGAGG a
GUS reporter Forward TCATTTTCTCCACAGCGACC
Reverse TGTAACGCGCTTTCCCACCA
ROS1
(AT2G36490)
Forward TGCTATATGGACGCCAGGTGAG
Reverse AGAAACAAGTCTCCTCGTCAC
ANT (AT4G37750) Forward TACCTTGGAACTTTTGGAACC
Reverse CAGACAAGAGTGTGTTACTAG
STM (At1g62360) Forward GGCCTTACCCTTCGGAGCA
Reverse GTAATGGTGAGGATGTGTTGC
LEAFY
(AT5G61850)
Forward AATGCCCCACCAAGGTGAC
Reverse CAGTGAACGTAGTGTCGCAT
At5g59330 Forward GGATCTGTGTTGTGAAGGAG
Reverse GACGTTGTCGCAGTTAGTG
ACTIN (At3g18780)
Forward GTACAACCGGTATTGTGCTGG a
Reverse CAAGGTCAAGACGGAGGATG a
PDF2/PP2A
(At1g13320)
Forward CATGTTCCAAACTCTTACCTG a
Reverse GTTCTCCACAACCGCTTGGT a
UBQ(At2g36060) Forward TCCTCACAACGTAACTGTAC
Reverse GACACAAGTCATGTTGATACG
QQS promoter Forward TCACTCGGATTGATGTCGTG
McrBC
a
Reverse AGGAGACGAAACAGACAAATC a
QQS 5'UTR Forward ATTAGAACATCTCAGAAGAAGC
Reverse TTGATGAGAGGTAATGAGAGG
At5g13440 Forward ACAAGCCAATTTTTGCTGAGC a
Reverse ACAACAGTCCGAGTGTCATGGT a
QQS Forward TTTGAAAATGGTATGTAAAAAATATTT
Bisulfite
Reverse AAAACATATAACACAACCCARATC
67
At5g13440 Forward TTAAGGAAAGGAAATTTTTTTATAAA
Reverse AAAATCCTACACCATACAACCTAA
FWA (At4G25530) Forward GGTTTTATATTAATATTAAAGAGTTATGGG b
Reverse AAACCAAAATCATTCTCTAAACAAAATA b
ros1-3 - wild-type Forward TGGAAGGGATCCGTCGTGGATTCT
PCR
c
Reverse CCCGCGACTCTTGATTGTTTCAGCAACTT
ros1-3 - mutant (JL202)
Forward TGGAAGGGATCCGTCGTGGATTCT c
Reverse CATTTTATAATAACGCTGCGGACATCTAC
ros1-4 - wild-type Forward ATTGATTGGGTTTAGGCTGGATG c
Reverse ACATGTCCAGCGCTTTAGTTG
ros1-4 - mutant
(LBb1.3)
Forward ATTGATTGGGTTTAGGCTGGATG d
Reverse ATTTTGCCGATTTCGGAAC
qqs-1 - wild-type Forward ATATAGCGGTTTGGGTTCTGC d
Reverse AAGAAGCCTCCTTTCGATCTG
qqs-1 - mutant
(WISK_LB)
Forward ATATAGCGGTTTGGGTTCTGC d
Reverse TGATCCATGTAGATTTCCCGGACATGAAG
QQS
promoter+5'UTR
Forward CTGCAGAAAAATCTGCAATTATG PCR
(Transgene)
Reverse GGATCCGAATCTAAGAACCAAAC
68
Supplementary analyses and figures
Figure S1 – Expression analyses of QQS in different organs of Arabidopsis. QQS expression is found in
several organs being very variable in most of them, which suggest influence of developmental and
environmental growth conditions. Data obtained from the GENEVESTIGATOR data mining tool (Hruz et al.,
2008), which gathers a lot of microarray data and provides a normalized and statistical analysis.
69
Figure S2 – Analysis of in situ GUS activity during leaf development of six independent Arabidopsis
transgenic lines for the GUS gene fused to QQS promoter region and 5’UTR (QQS:GUS). 21 day-old plants
were analyzed. In all six lines, GUS activity was mainly detected in the vegetative meristematic region and
younger leaves and was reduced in older leaves. GUS activity reflects the potential of QQS expression. Wild-
type genotype segregated from heterozygous transgenic lines, thus without the transgenic insertion, was named
as background and used as negative controls (i.e., not showing GUS activity). Bars equal 0.5 cm except H with
0.2 cm.
Figure S3 – Plant material used in the experiments of quantitative expression and methylation analysis of
QQS shown in Fig. 2. A) 40 day-old plants showing the 6 and 7th
rosette leaves (RL) and the 13 and 14th
rosette
(YRL) leaves which were collected for the analyses showed in the figure 3A and B. Leaves from individual
plants were used as single biological replicate. B) 60 day-old plants from which the different organs and kind of
leaves (mature rosette leaves*) were collected for the analysis undertaken in figure 2C and D are shown. A pool
of five plants was used for each organ as single biological replicate. All the plants were cultivated under long
70
day conditions (16 hours of light and 8 of darkness) at 22°C. RL – rosette leaves and YRL – younger rosette
leaves.
Figure S4 – Analysis of QQS expression and methylation in different organs. Contrasting QQS epialleles
present different expression and methylation patterns in different organs. The epialleles QQSme
and Col*3-10,
which are more methylated, exhibited a more variable expression during development compared to the others
less methylated epialleles. For QQSme
a correlation between increase of expression and decrease of methylation
was found in leaves compared to organs formed later, such as, flowers. A) Relative expression of five
contrasting QQS epialleles in different organs of 60 day-old plants. The expressions values were normalized for
each epiallele in relation to the respective expression value in silique fixed to 1 and the expression fold in
relation to QQSme
is indicated. B) McrBC-qPCR analysis showing the DNA methylation percentage in the QQS
promoter and 5’UTR of the two epialles QQSme
and Col*3-2 in the respective organs used in A. The growth
conditions were long day (16 hours of light and 8 of darkness) and 22°C in soil. The data shown in the panels A
and B represent the mean of expression and methylation levels, respectively, from three individual plants. The
endogenous controls PDF2 and UBI were used to normalize the expression analyses. Values below 40% in
McrBC-qPCR analyses were not considered by technical limitations. RL – Rosette leaves, In – internode of the
inflorescence stem, Fl – Flowers, Sil - Siliques.
71
Figure S5 – Analyses of QQS expression and methylation during early stages of development. Contrasting
QQS epialleles present different expression and methylation patterns. The epialleles QQSme
and Col*3-10, which
are more methylated, exhibited a more variable expression although no difference of methylation was found
during seedlings development. A) Relative expression of contrasting QQS epialleles in whole seedlings sampled
3, 6, 9 and 12 days after germination. The expressions values were normalized for each epiallele in relation to
the respective expression value at 3 days which was fixed to 1 and the fold in relation to QQSme
is indicated. B)
McrBC-qPCR analysis showing the DNA methylation percentage in the QQS promoter and 5’UTR of the
respective seedlings containing the epialleles QQSme
, Col*3-15 e Col*3-2. Seedlings were cultivated in liquid
MS/2 supplemented with 0.7% sucrose. The growth conditions were long day (16 hours of light and 8 of
darkness) and 22°C. The data represent the mean of three biological replicates each consisting of a pool of
seedlings (5 mg of seeds). C) Relative expression of contrasting QQS epialleles in mature embryos. The
endogenous controls ACTIN and PDF2 were used to normalize expression analyses. Values below 40% in
McrBC-qPCR analyses were not considered by technical limitations.
72
Figure S6 – Profile of QQSme
expression and methylation in root, cotyledon, meristematic region and
leaves of seedlings. The expression and methylation levels from dissected organs of 12 day-old seedlings
cultivated in agar MS/2 medium supplemented with 0.7% sucrose and long day conditions (16 hours of light/8
hours of dark). A) QQSme
absolute expression normalized with the endogenous controls ACTIN and PDF2. B)
McrBC-qPCR analysis showing the DNA methylation percentage in the QQS promoter and 5’UTR in the
respective organs described in A.
Figure S7 – Details of the plant material used for the analysis of QQS expression and methylation during
leaf development. Figures A, B and C show, respectively, 21, 31 and 43 old plants from which sampling of
organs used in experiment presented in figure 3 were realized. All the plants were cultivated under long day
conditions (16 hours of light/8 hours of dark) at 22° and for each of the three biological replicate, dissected
organs of ten plants were pooled.
73
Figure S8 – Expression analyses of the demethylated QQS epialleles Col*3-2 and Col*3-16 during leaf
development. QQS expression was measured by qRT-PCR from total RNA extracted of dissected leaves and
meristematic region of 31 day-old plants. The samples were collected in three biological replicates each one
constituted by dissected organs of ten plants pooled. A) Absolute expression of Col*3-2 and Col*3-16 in pools
of dissected organs. B) Absolute expression of the meristem-specific gene AINTEGUMENTA and
SHOOTMERISTEMLESS (STM) used as controls of meristem-enrichment cells. The endogenous controls PDF2
and UBI were used to normalize expression analyses. The initials RL followed by numbers means the order of
leaf formation in the rosette, being RL1-2 the older ones.
aaaacattcttatgccttcatgaaatagccaattagaatgtttcactcggattgatgtcgtggcgaaatctcagc
ctttcctcatgccatcttaaataaatgtccaagcttgccaaaacgatccaatcaccatatatcactcggatacat
gacgtggcgagatattggcctttgatttgtctgtttcgtctccttttcgataatttattcacttctttcgatcca
aaaaattagaacatctcagaagaagcctcctttcgatctgtcagccattgaagaaacctcctttcgatctgtcag
ccattgaagatcagaagaaacaagactcacacggtcagccattgaagaagcctcctctcattacctctcatcaaa
catctagatctgtacccaaaccttatccctttttccttatttctcgctttgtctattcttaatctgattaatact
tgttgttgttccaggttatagaagatctgggttgtgttatatgcttcattttctccacaggtaatattcttttca
tcctttcaaaaataatttccgattcgtcagagtgatattgataagatgcgagtaccattgattttcttttgtgta
tattgaatgattcttatatttcagcgaccagttggtgtttggttcttagattcATGAAGACCAATAGAGAGgtaa
acgattagccattagtaacttatttgtaattcaaacgactaactttctttagtttactgagattaaatctttcgt
tttgtagCAGGAAATTTACGTTGAAAGAAGCTTCAAACCAAACAATTCAACAATTCAGAATTTGATGGACATTGA
AAGGTTCATTTTGCCTCACACTTCTACATCAGGTGTCGCAAGGCTCAAAATGAGGGTCATATCATGGGTCGGGCT
TCAGTTCTACAACTACTGAtattgggccttatcacaaattagttatagggccattgtatccaatatttaatatct
ctgtaaacttgtttaatggttattttgttctaatgcccattacaactagacttaatatcaatataaatactttgt
aaaatactttttttttgttaaagagtcaccaaatagtataaaataaggcataaaatacagtaattatatggatac
tctaaatacagtatccaacgactaaatacagtaattatacggtataaattacaatataaaatatggtataaaaaa
cagtataaattacagtattaaaatacaatatccagcgcttaaatacagtaattatacggatactctaagaaagct
attgttcagcaacccttaaactatcaattattacataatttaaatgctattataaatttaaagtactgctatagt
tacgcaatttaatagcaaatagtaattgctattgtaaaattaacaataacacacaacaaaacgctatattagttg
aaatttttatagcagaacccaaaccgctatataaagtgttcgacctactataacaagcgcatacatagtgtttac
tacaacgctataatatcttataatattgtttttcctgtgctattgatactcacaattcctgtagtgaattcTTTG
TCGCATAGTCTCATTCCG
XXX – QQS UTRs.
XXX – QQS protein coding sequence.
XXX - AAGAA-motif: “cis-acting elements involved in photoregulation of an
oat phytochrome promoter in rice”.
XXX – AE-box: part of a system composed of 3 to 4 Gap-boxes and 2 AE-boxes;
conferring light responsiveness.
XXX – AT-rich sequence: element for maximal elicitor-mediated activation.
74
XXX – AT1-motif: part of a light responsive module.
XXX – Box 4: part of a conserved DNA module involved in light
responsiveness.
XXX – Box I: light responsive element.
XXX – G-box: cis-acting regulatory element involved in light
responsiveness.
XXX – GATA-motif & I-box: part of a light responsive element.
XXX – GT1-motif: light responsive element.
XXX – HSE: cis-acting element involved in heat stress responsiveness.
XXX – Circadian: cis-acting regulatory element involved in circadian
control.
Figure S9 – Analyses of the cis-acting regulatory elements present in the QQS genomic loci. We used the
PlantCARE tool (Lescot et al., 2002) to search for regulatory motifs at the genomic and adjacent sequence of
QQS. The figure shows several motifs considered important to expression regulation in response to light (colored
blocks). A brief description of each motif is available. Moreover, a lot of CAAT-box 7 TATA-box motifs, a
common cis-acting element in promoter and enhancer regions, were found but not indicated in the figure.
Figure S10 – ROS1 structure, expression along leaf development and ros1 mutant genotyping. A) Figure
modified from Penterman et al. (2007) showing the structure of the ROS1 glycosilases and the T-DNA insertions
(arrows). Exons are boxed regions and introns are lines. Blue exons encode the helix–hairpin–helix DNA
glycosylase domain, and pink and orange exons encode conserved domains of unknown function. B) ROS1
expression in the same material used in fig. 6. Meristematic regions containing shoot apical meristem (Mer) cells
and rosette leaves 1-2 (first to emerge, RL 1-2) and 3-4 (younger leaves, RL 3-4) were sampled from 21 and 31
day-old plants (plant details on fig. S7). Total RNA was extracted and QQS expression was measured by RT-
qPCR. ROS1 showed null or altered expression in ros1 mutants in relation to wt. The expression of PDF2 and
UBI genes were used to normalize expression analyses. C) Genotyping of wt and ros1-4 plants by conventional
PCR. The figure shows an agarose gel with different bands containing the products of amplification and it was
75
used to identify the genotypes. L - 1kb ladder, wt - wild-type plants carrying QQSme
; ros1 – ros1-4 mutants and
Neg. – the PCR negative control.
Figure S11 – Analysis of QQS expression during leaf development of wild-type and ros1 mutant. This
experiment constitutes a repetition and confirmation of the data shown in figure 4. A) Relative QQS expression
in dissected leaves and meristematic region of plants of different ages. QQS expressions in leaves were
normalized for each genotype in relation to the respective expression value in the meristem. The folds of QQS
expression in ros1-3 and ros1-4 relative to wild-type (wt carrying QQSme
epiallele) are shown and QQS is less
expressed in both ros1 mutants. B) and C) Absolute expression analyses of meristem-specific genes
(AINTEGUMENTA and SHOOTMERISTEMLESS) used as controls of the degree of enrichment in meristem
cells. The endogenous controls PDF2 and UBI were used to normalize expression analyses. The initials RL
followed by numbers means the order of leaf formation in the rosette, being RL1-2 the older ones.
76
Figure S12 – QQS expression in the F2 segregating population resulting from selfing of the heterozygous
ROS1+/-
genotype. In order to evaluate the influence of ROS1 in QQS expression, wild-type plants (wt, female
parent) was crossed with ros1-4 (male parent), generating the double heterozygous QQSme
/QQSros1-4
and
ROS1/ros1-4, and the 69 plants of the segregating F2 population were analyzed in comparison to the parental
genotypes (F0). The ros1-4 mutant carries a T-DNA insertion (Penterman et al., 2007) and the plants genotypes
(ROS1+/+
, ROS1+/-
and ROS1-/-
) were determined by PCR (see Fig. S10-C) and the proportion of the genotypes
fits the 1:2:1 proportion (χ² = 1,1739 and p = 0,5560). QQS expression was conducted after RNA extraction of
the rosette leaves 6 to 9 of individual plants (31 day-old) and measured by RT-qPCR. The parental wt and ros1-4
(F0) were analyzed in four biological replicates and the endogenous controls PDF2 and UBI were used to
normalize expression analyses.
77
Figure S13 – Comparative analysis of starch accumulation between plants carrying the contrasting
epialleles QQSme
and Col*3-16. The figure shows that plants carrying the less expressed QQS epiallele (QQSme
)
accumulate more starch at the end of day and end of night than plants carrying the more expressed epiallele
Col*3-16. A) Starch analyses of 9 day-old seedlings cultivated in MS/2 media supplemented with 0.5% sucrose.
B) Expression analyses in the same tissues used in A and normalized with the endogenous controls ACTIN and
PDF2. C) Starch analyses of 42 day-old plants cultivated in soil. Plants were cultivated in long day conditions
(16 hour of light and 8 of night) and the starch analysis was made with 5 biological replicates containing 6 plants
for each genotype (30 plants in total). This result is in accordance with Fig. 10 and the notion that QQS is a
negative regulator of starch accumulation (Li et al., 2009; Seo et al., 2011).
78
Figure S14 – Expression analyses of genes involved DNA methylation regulation. QQS and several genes
related to demethylation, de novo methylation (RdDM) and methylation maintenance pathways were analyzed
regarding its expression at plant organs. Expression levels of enzymes involved in DNA methylation
maintenance were lower in leaves in relation to meristems supporting the hypothesis of a passive demethylation
of QQS throughout leaf development. L – Leaf and M – meristem. Data collected from GENEVESTIGATOR
tool (Hruz et al., 2008), which gathers a lot of microarray data and provides a normalized and statistical analysis
Figure S15 - QQS methylation profile in the vegetative spermatic nucleus of the pollen of wt and dme/+
plants. QQS methylation is lost in the vegetative nucleus (VN) in relation to the spermatic ones and partially
79
restored by the dme/+ mutation. This result points the participation of DEMETER in this active demethylation
process. The image corresponds to the region 12348429-12350350 pb of the third Arabidopsis chromosome
(DMR n° 46) kindly analyzed and provided by Filipe Borges (prof. Martienssen’s lab, Cold Spring Harbor,
EUA) from the combined data generated by Calarco et al., 2012 and Ibarra et al., 2012.
Figure S16 – A schematic figure of the binary vector containing the QQS:GUS cassette. QQS promoter and
5’UTR (pro+5’UTR QQS) was amplified by PCR and cloned into pGEM T-easy vector (left figure). GUS
sequence was digested and transferred from pBI121 plasmids to the pFP101-HA plasmid (right figure) forming
the base vector pFP-GUS. After enzyme digestions and DNA agarose purification, pro+5’UTR QQS was cloned
in the pFP-GUS binary vector substituting CaMV35S-Pro. The final plasmid was inserted into Agrobacterium
tumefaciens GV3101::pMP90 (Koncz et al., 1986) and used for in planta Arabidopsis transformation.
References of the supplementary material
Calarco JP et al. (2012) Reprogramming of DNA Methylation in Pollen Guides Epigenetic Inheritance via Small
RNA. Cell, 151:194–205
80
Clark SE et al. (1996). The CLAVATA and SHOOT MERISTEMLESS loci competitively regulate meri-stem
activity in Arabidopsis. Development, 122:1567–1575.
Gruntman E et al. (2008). Kismeth: Analyzer of plant methylation states through bisulfite sequencing. BMC
Bioinformatics, 9:371.
Hruz T et al. (2008). Genevestigator V3: a reference expression database for the meta-analysis of transcriptomes.
Advances in Bioinformatics, 420747.
Ibarra et al. (2012). Active DNA Demethylation in Plant Companion Cells Reinforces Transposon Methylation
in Gametes. Science, 337:1360.
Koncz C et al. (1986) High-frequency T-DNA-mediated gene tagging in plants. Proc. Natl. Acad. Sci. USA, 86:
8467-8471.
Lescot et al. (2002). PlantCARE, a database of plant cis-acting regulatory elements and a portal to tools for in
silico analysis of promoter sequences. Nucleic Acids Res, 30(1):325-7.
Li L et al. (2009) Identification of the novel protein QQS as a component of the starch metabolic network in
Arabidopsis leaves. Plant Journal, 58:485-498.
Mizukami Y and Fischer RL (2000). Plant organ size control: AINTEGUMENTA regulates growth and cell
numbers during organogenesis. PNAS, 97(2):942-947.
Murashige T and Skoog F (1962) A revised medium for rapid growth and bio-asays with tobacco tissue culture.
Physiol. Plant, 15:473-497.
Penterman J et al. (2007). DNA demethylation in the Arabidopsis genome. PNAS, 104 (16):6752–6757.
Seo PJ et al. (2011). Two splice variants of the IDD14 transcription factor competitively form nonfunctional
heterodimers which may regulate starch metabolism. Nature communications, 2:303.
Silveira AB et al. (2007) The Arabidopsis AtbZIP9 protein fused to VP16 transcriptional activation domain
alters leaf and vascular development. Plant Sci., 172:1148–1156.
Silveira AB et al. (2013) Extensive Natural Epigenetic Variation at a De Novo Originated Gene. PLoS Genetics,
9 (4):e1003437.
Siriwardana NS and Lamb RS. The poetry of reproduction: the role of LEAFY in Arabidopsis thaliana flower
formation. Int J Dev Biol., 56(4):207-21.
Zhu J et al. (2007). The DNA glycosylase/lyase ROS1 functions in pruning DNA methylation patterns in
Arabidopsis. Curr. Biol., 17:54–59.
81
CONCLUSÃO
O gene órfão QQS apresenta um padrão de expressão específico nos órgãos de
Arabidopsis thaliana que é parcialmente determinado pelos níveis de metilação presentes em
seu lócus. Mostramos neste trabalho que o perfil de metilação de QQS é um elemento
importante para definir o padrão de expressão durante o desenvolvimento vegetativo. Por
exemplo, epialelos demetilados apresentam expressão mais homogênea entre órgãos e um
padrão decrescente com a idade das folhas enquanto o epialelo metilado apresenta expressão
mais variável e oposto com a idade das folhas. As principais diferenças nos níveis de
metilação ocorrem na região próxima ao sítio de início da trasncrição e 5’UTR sugerindo que
a metilação dessas regiões são responsáveis pela alteração do padrão de expressão. Em folhas,
o epialelo metilado QQSme
é mais expresso e menos metilado do que em outros órgãos
formados posteriormente e, assim, levantamos a hipótese de que uma reprogramação ocorra a
partir do meristema. De acordo com isso, mostramos que a expressão de QQSme
aumenta
progressivamente ao longo do desenvolvimento foliar, entretanto uma demetilação somente
foi observada em folhas mais velhas indicando ser este um processo dependente da idade.
Foram observadas também variações na expressão de QQS em órgãos independente de
variações nos níveis de metilação. Isso indica que outros fatores trans também são
importantes para definir a expressão de QQS e um potencial candidato seria o fator de
transcrição IDD14 que foi descrito como um regulador de QQS (Seo et al., 2011). Avaliamos
se a demetilação de QQSme
em folhas seria mediada pela DNA glicosilase ROS1 e os
resultados indicam que, embora ROS1 não seja necessária para demetilação de QQS durante o
envelhecimento foliar, ROS1 está envolvida no estabelecimento do nível de expressão de
QQS. Adicionalmente, dados da literatura indicam que o lócus de QQS apresenta
hipermetilaçao em diversos mutantes ros1 (Gong et al., 2002; Cokus et al., 2008; Lister et al.,
2008; Agius et al., 2006; Zhu et al., 2007; Qian et al., 2012) e também é demetilado em
plantas infectadas com Pseudomonas syringae pv. Tomato (Pst; in Table S3 of Dowen et al.,
2012) cuja parte da resposta imune parece ser desencadeada por ROS1 (Yu et al., 2013; Le et
al., 2014). Esses resultados indicam que QQS é um alvo de ROS1. Recentemente, foi descrita
uma relação antagônica entre ROS e RdDM para balancear os níveis de metilação celulares
(Lei et al., 2015; Williams et al., 2015). Assim, sugerimos que ROS1 poderia atuar em
equilíbrio dinâmico para proteger lócus específicos contra hipermetilaçao e o desbalanço
desse equilíbrio por influência de fatores ambientais poderia contribuir para geração de
82
epialelos. Proteínas capazes de reconhecer sítios metilados podem ser importantes nesse
processo, pois, recrutam remodeladores de cromatina e facilitam a ação de ROS1 (Zemach
and Grafi, 2007; Lang et al., 2015). Dessa forma, a metilação do DNA pode ser interpretada
como um sinalizador atraindo ou excluindo outros fatores ativadores e repressores da
transcrição (Medvedeva et al., 2014; Domcke et al., 2015; Li et al., 2015b). Isso poderia
explicar porque epialelos QQS metilados apresentam padrões similares de expressão, mas
esse padrão é modificado qualitativamente com a diminuição dos níveis de metilação.
QQS também apresenta regulação epigenética durante o desenvolvimento
reprodutivo onde uma reprogramação epigenética foi descrita (Slotkin et al., 2009) e esse
processo é dependente de DEMETER na gametogênese (Calarco et al., 2012; Ibarra et al.,
2012). Entretanto, o significado biológico dessa regulação não foi descrito. Por outro lado,
diversos trabalhos demonstram que QQS é um regulador negativo da síntese de amido (Li et
al., 2009; Tsutsui et al., 2015) e impacta o crescimento das plantas em condições de frio (Seo
et al., 2011). De acordo com isso, nossos resultados indicam diferenças entre epialelos QQS
contrastantes para o acúmulo de amido. Arendsee et al. (2014) sugere que QQS integra
diversos sinais ambientais para regular o crescimento sob diferentes condições via regulação
do amido. Assim, a variação epialélica natural de QQS (Silveira et al., 2013) poderia de fato
desempenhar um papel no controle do crescimento vegetal e, consequentemente, afetar o
fitness da planta. Entretanto, para comprovar essa sugestão, o acúmulo de amido deve ser
analisado em maiores detalhes envolvendo outros genótipos, tais como, epiRILS e diferentes
condições de crescimento, nos quais a expressão de QQS seria alterada. QQS apresenta função
na regulação do amido mesmo quando transferido para outra espécie onde a princípio não está
presente (Li e Wurtele, 2014). Isso expande ainda mais as aplicações agronômicas dos genes
órfãos e reforça a importância da conservação da biodiversidade já que cada espécie
apresenta, em geral, de 5 a 15% de genes órfãos (Arendsee et al., 2014).
PERSPECTIVAS
Comparar a quantidade de amido entre linhagens EpiRILS de Arabidopsis que diferem
quanto ao nível de metilação em QQS e, também em outros genótipo de interesse para
83
reforçar a idéia de que o estado epigenético de QQS pode influenciar no crescimento da planta
e contribuir para processos adaptativos.
Avaliar o fenótipo de crescimento de plantas carregando epialelos QQS contrastantes
em diferentes condições de fotoperíodo para verificar o envolvimento de QQS em processos
adaptativos.
No intuito de entender como epialelos QQS são formados, poderia ser avaliado se a
expressão de QQS é afetada por estresses abióticos e os efeitos sobre a metilação do DNA. A
hipótese a ser testada é de que o excesso de luz induz transcrição de QQS e promove sua
demetilação que, por sua vez, poderia regular o excesso de amido produzido nas folhas sob
essa condição garantindo maior eficiência na relação metabolismo versus
desenvolvimento/envelhecimento.
Qual seria a tendência da alteração dos níveis de metilação em epialelos com valores
entre QQSme
e Col*3-2, aumentar, não alterar ou diminuir ao longo das gerações? Para
responder essa questão, deverá ser analisada a herança transgeneracional dos epialelos QQS
intermediários. Essas análises poderão ser combinadas com estresses induzindo enzimas da
via RdDM ou ROS1 para verificar se o equilíbrio dinâmico é desbalanceado para aumento ou
diminuição da metilação.
Avaliar a hipótese de que a perda de ROS1 leva a remetilação de QQS por meio da
introgressão da mutação ros1-4 em uma planta carregando um epialelo QQS demetilado. A
ausência de ROS1 poderia desequilibrar a regulação do lócus e permitir remetilação?
Similarmente, avaliar se mutações para enzimas da via RdDM quando introgredidas em
plantas carregando um epialelo QQS metilado induzem a perda de metilação.
Testar a hipótese de que o fator de transcrição IDD14 impede a remetilação do lócus de
QQS ou, em outras palavras, testar se a transcrição promove a demetilação por ocupação do
sítio de início da transcrição.
Análise da via de controle epigenético que atua sobre epialelos QQS definindo padrões
de expressão, em especial, a participação da enzima DICER-like 4 na manutenção da
metilação.
Analisar como o lócus do epialelo metilado QQSme
é reconhecido e regulado por ROS1.
Dentre as hipóteses, pretende-se avaliar a participação de ROS3 na regulação de QQS, uma
enzima que pode se associar a pequenos RNAs e capaz de interagir com ROS1.
84
Investigar a possibilidade de QQS desempenhar um papel no desenvolvimento do pólen
por meio da análise comparativa de entre o mutante T-DNA/TE qqs-1 (WiscDsLoxHs077) e
plantas wild-type.
REFERÊNCIAS BIBLIOGRÁFICAS
Agius F et al. (2006). Role of the Arabidopsis DNA glycosylase lyase ROS1 in active DNA
demethylation. PNAS, 103(31):11796–11801.
Andres F e Coupland G (2012). The genetic basis of flowering responses to seasonal cues.
Nat. Rev. Genet., 13:627–639.
Arendsee ZW et al. (2014). Coming of age: orphan genes in plants. Trends in Plant Science,
19(11):698–708.
Aristotle (1979). Generation of Animals, translated by A. L. Peck, Cambridge, MA: Harvard
University Press.
Baubec T et al. (2013). Methylation-dependent and-independent genomic targeting principles
of the MBD protein family. Cell, 153(2):480-492.
Baubec T et al. (2014). Meristem-specific expression of epigenetic regulators safeguards
transposon silencing in Arabidopsis. EMBO Reports, 15(4):446–52.
Becker C et al. (2011). Spontaneous epigenetic variation in the Arabidopsis thaliana
methylome. Nature, 480(7376):245–249.
Bernatavichute YV et al. (2008). Genome-wide association of histone H3 lysine nine
methylation with CHG DNA methylation in Arabidopsis thaliana. PLoS ONE, e3156.
Bird A (2007). Perceptions of epigenetics. Nature, 447:396–398.
Calarco JP et al. (2012). Reprogramming of DNA methylation in pollen guides epigenetic
inheritance via small RNA. Cell, 151:194–205.
Cao X e Jacobsen SE (2002). Locus-specific control of asymmetric and CpNpG methylation
by the DRM and CMT3 methyltransferase genes. Proc. Natl Acad. Sci. USA, 99 (suppl.
4):16491–16498.
Cedar H e Bergman Y (2009). Linking DNA methylation and histone modification: patterns
and paradigms. Nat Rev Genet., 10(5):295-304.
Chan SW et al. (2005). Gardening the genome: DNA methylation in Arabidopsis thaliana.
Nature Rev. Genet., 6:351–360.
85
Choi Y et al. (2002). DEMETER, a DNA glycosylase domain protein, is required for
endosperm gene imprinting and seed viability in arabidopsis. Cell, 110:33–42.
Cokus SJ et al. (2008). Shotgun bisulphite sequencing of the Arabidopsis genome reveals
DNA methylation patterning. Nature, 452:215-219.
Cortijo et al. (2014). Mapping the epigenetic basis of complex traits. Science,
343(6175):1145-8.
Cubas P et al. (1999). An epigenetic mutation responsible for natural variation in floral
symmetry. Nature, 401:157–161.
Ding Y et al. (2012). ATX1-generated H3K4me3 is required for efficient elongation of
transcription, not initiation, at ATX1-regulated genes. PLoS Genet, 8(12):e1003111.
Domcke S et al. (2015). Competition between DNA methylation and transcription factors
determines binding of NRF1. Nature, 528(7583):575-9.
Dowen RH et al. (2012). Widespread dynamic DNA methylation in response to biotic stress.
Proc Natl Acad Sci USA, 109(32):E2183-91.
Du J et al. (2012) Dual binding of chromomethylase domains to H3K9me2-containing
nucleosomes directs DNA methylation in plants. Cell, 151:167–180.
Feng S e Jacobsen SE (2011). Epigenetic modifications in plants: an evolutionary perspective.
Curr. Opinion in plant biology, 14(2):179-86.
Feng S et al. (2010). Epigenetic Reprogramming in Plant and Animal Development. Science,
330(6004):622-627.
Finke A et al. (2012). IDN2 has a role downstream of siRNA formation in RNA-directed
DNA methylation. Epigenetics, 7:950–960.
Fornara F et al. (2010). SnapShot: control of flowering in Arabidopsis. Cell, 141(3):550,
550.e1-2.
Fujimoto R et al. (2011). Epigenetic variation in the FWA gene within the genus Arabidopsis.
Plant J. 66:831–843.
Gao ZH et al. (2010). An RNA polymerase II- and AGO4-associated protein acts in RNA-
directed DNA methylation. Nature, 465:106–109.
Goldberg AD et al. (2007). Epigenetics: A Landscape Takes Shape. Cell, 128(4):635–638.
Gollery M et al. (2006). What makes species unique? The contribution of proteins with
obscure features. Genome Biol., 7: R57.
Gong Z et al. (2002). ROS1, a repressor of transcriptional gene silencing in Arabidopsis,
encodes a DNA glycosylase/lyase. Cell, 111:803–814.
86
Henderson IR e Jacobsen SE (2007). Epigenetic inheritance in plants. Nature, 447:418-424.
Holliday R (2006). Epigenetics: A Historical Overview. Epigenetics, 1(2):76–80.
Hruz T et al. (2008). Genevestigator V3: a reference expression database for the meta-
analysis of transcriptomes. Advances in Bioinformatics, 420747.
Hsieh TF et al. (2009). Genome-wide demethylation of Arabidopsis endosperm. Science, 324:
1451–1454.
Ibarra CA et al. (2012). Active DNA demethylation in plant companion cells reinforces
transposon methylation in gametes. Science, 337:1360–1364.
Ikeda Y (2012). Plant imprinted genes identified by genome-wide approaches and their
regulatory mechanisms. Plant Cell Physiol., 53:809–816.
Ito T et al. (2015). Genome-wide negative feedback drives transgenerational DNA
methylation dynamics in Arabidopsis. PLoS Genet., 11(4):e1005154.
Jackson JP et al. (2002). Control of CpNpG DNA methylation by the KRYPTONITE histone
H3 methyltransferase. Nature, 416:556–560.
Janowski BA et al. (2007). Activating gene expres- sion in mammalian cells with promoter-
targeted duplex RNAs. Nat. Chem. Biol., 3:166–73
Johannes F et al. (2009). Assessing the impact of transgenerational epigenetic variation on
complex traits. PLoS Genet., 5:e1000530.
Johnson LM et al. (2014). SRA- and SET-domain-containing proteins link RNA polymerase
V occupancy to DNA methylation. Nature, 507:124–128.
Kankel MW et al. (2003). Arabidopsis MET1 cytosine methyltransferase mutants. Genetics,
163:1109–1122.
Kinoshita T e Jacobsen SE (2012). Opening the door to epigenetics in PCP. Plant Cell
Physiol., 53(5):763-5.
Klose RJ e Bird AP (2006). Genomic DNA methylation: the mark and its mediators. Trends
Biochem Sci., 31:89–97.
Klymenko T e Muller J (2004) The histone methyltransferases Trithorax and Ash1 prevent
transcriptional silencing by Polycomb group proteins. EMBO Rep., 5:373–377.
Kooke et al. (2015). Epigenetic basis of morphological variation and phenotypic plasticity in
Arabidopsis thaliana. Plant Cell, 27(2):337-48.
Kouzarides T (2007). Chromatin modifications and their function. Cell, 128:693–705.
87
Kraft E et al. (2008). ORTH/VIM proteins that regulate DNA methylation are functional
ubiquitin E3 ligases. Plant J., 56:704–15.
Kurihara Y et al. (2008). Identification of the candidate genes regulated by RNA-directed
DNA methylation in Arabidopsis. Biochem. Biophys. Res. Commun., 376:553–557.
Lang Z et al. (2015). The methyl-CpG-binding protein MBD7 facilitates active DNA
demethylation to limit DNA hyper-methylation and transcriptional gene silencing. Mol
Cell, 57(6):971-83.
Latzel V et al. (2012). Epigenetic variation in plant responses to defence hormones. Ann. Bot.
(Lond.), 110:1423–1428.
Law J and Jacobsen SE (2010). Establishing, maintaining and modifying DNA methylation
patterns in plants and animals. Nature Reviews Genetics, 11(3):204-220.
Le T-N et al. (2014). DNA demethylases target promoter transposable elements to positively
regulate stress responsive genes in Arabidopsis. Genome Biology, 15:458.
Lei M et al. (2015). Regulatory link between DNA methylation and active demethylation in
Arabidopsis. Proc. Natl. Acad. Sci., 112(11):3553-7.
Li L e Wurtele ES (2014). The QQS orphan gene of Arabidopsis modulates carbon and
nitrogen allocation in soybean. Plant Biotechnol. J., 13(2):177–187.
Li L et al. (2009). Identification of the novel protein QQS as a component of the starch
metabolic network in Arabidopsis leaves. Plant Journal, 58:485-498.
Li L et al. (2015a). QQS orphan gene regulates carbon and nitrogen partitioning across
species via NF-YC interactions. Proc. of the Nat. Acad. of Sci., 112(47):14734-14739
Li Q et al. (2015b). Regulation of Active DNA Demethylation by a Methyl-CpG-Binding
Domain Protein in Arabidopsis thaliana. PLoS Genet., 11(5):e1005210
Li X et al. (2008). High-resolution mapping of epigenetic modifications of the rice genome
uncovers interplay between DNA methylation, histone methylation, and gene
expression. Plant Cell 20(2):259–276.
Li Y et al. (2015c). An AP Endonuclease Functions in Active DNA Dimethylation and Gene
Imprinting in Arabidopsis. PLoS Genetics, 11(1):e1004905.
Lister R. et al. (2008). Highly integrated single-base resolution maps of the epigenome in
Arabidopsis. Cell, 133:523–536.
Liu S et al. (2007). Plant SET- and RING-associated domain proteins in
heterochromatinization. Plant J., 52:914–26.
88
Liu ZW et al. (2014). The SET domain proteins SUVH2 and SUVH9 are required for Pol V
occupancy at RNA-directed DNA methylation loci. PLoS Genet., 10:e1003948.
Martin et al. (2009). A transposon-induced epigenetic change leads to sex determination in
melon. Nature, 461(7267):1135-8.
Matzke M et al. (2009). RNA-mediated chromatin-based silencing in plants. Curr. Opin. Cell
Biol., 21:367–376.
Matzke MA e Mosher RA (2014). RNA-directed DNA methylation: an epigenetic pathway of
increasing complexity. Nature Reviews Genetics, 15(6):394–408.
Medvedeva YA et al. (2014). Effects of cytosine methylation on transcription factor binding
sites. BMC Genomics, 15:119.
Mirouze M and Paszkowski J (2011). Epigenetic contribution to stress adaptation in plants.
Current Opinion in Plant Biology, 14:267–274.
Mirouze M e Vitte C. (2014). Transposable elements, a treasure trove to decipher epigenetic
variation: insights from Arabidopsis and crop epigenomes. Journal of Experimental
Botany, 65(10):2801–2812.
Mosher RA et al. (2008). PolIVb influences RNA-directed DNA methylation independently
of its role in siRNA biogenesis. Proc. Natl. Acad. Sci. USA, 105:3145–50.
Penterman J et al. (2007a). DNA demethylation in the Arabidopsis genome. PNAS, 104(16):
6752–6757.
Penterman J et al. (2007b). Genetic Interactions between DNA Demethylation and
Methylation in Arabidopsis. Plant Physiology, 145:1549–1557.
Pikaard CS et al (2012). A transcription fork model for Pol IV and Pol V‑dependent RNA-
directed DNA methylation. Cold Spring Harb. Symp. Quant. Biol., 77:205–212.
Ponferrada-Marin MI et al. (2010). Methylation-independent DNA binding modulates
specificity of Repressor of Silencing 1 (ROS1) and facilitates demethylation in long
substrates. J. Biol. Chem., 285:23032−23039.
Ponferrada-Marin MI et al. (2012). Demethylation initiated by ROS1 glycosylase involves
random sliding along DNA. Nucleic Acids Research, 40(22):11554-62.
Pu L e Sung ZR (2015). PcG and trxG in plants – friends or foes. Trends in Genetics,
31(5):252–262.
Qian W et al. (2012). A histone acetyltransferase regulates active DNA demethy- lation in
Arabidopsis. Science, 336:1445–1448.
Qian W et al. (2014). Regulation of active DNAdemethylation by an a-crystal- lin domain
protein in Arabidopsis. Mol. Cell, 55:361–371.
89
Quadrana L et al. (2014). Natural occurring epialleles determine vitamin E accumulation in
tomato fruits. Nat. Commun., 5:3027.
Reddington JP et al. (2013). Redistribution of H3K27me3 upon DNA hypomethylation
results in de-repression of Polycomb target genes. Genome Biol., 25;14(3):R25.
Regulski M et al. (2013). The maize methylome influences mRNA splice sites and reveals
widespread paramutation-like switches guided by small RNA. Genome Research,
23:1651–1662.
Reinders J et al. (2009). Compromised stability of DNA methylation and transposon
immobilization in mosaic Arabidopsis epigenomes. Genes Dev., 23: 939–950.
Rose e Klose (2014). Understanding the relationship between DNA methylation and histone
lysine methylation. Biochim Biophys Acta, 1839(12):1362-72.
Russo VEA et al. (1996). Epigenetic mechanisms of gene regulation. Cold Spring Harbor
Laboratory Press, Woodburry.
Sahu et al. (2013). Epigenetic mechanisms of plant stress responses and adaptation. Plant Cell
Reports, 32(8):1151–1159.
Sarma K e Reinberg D (2005). Histone variants meet their match. Nat. Rev. Mol. Cell Biol.,
6:139–149.
Schlichting CD (1986). The evolution of phenotypic plasticity. Annu. Rev. Ecol. Syst.,
17:667–693.
Schmitz et al. (2011). Trans- generational epigenetic instability is a source of novel
methylation variants. Science, 334:369–373.
Schmitz et al. (2013). Patterns of population epigenomic diversity. Nature, 495:193–198.
Schoft VK et al. (2011). Function of the DEMETER DNA glycosylase in the Arabidopsis
thaliana male gametophyte. Proc. Natl. Acad. Sci. USA, 108:8042–8047.
Schuettengruber B et al. (2007). Genome Regulation by Polycomb and Trithorax Proteins.
Cell, 128:735–745.
Schuettengruber B et al. (2011). Trithorax group proteins: switching genes on and keeping
them active. Nat. Rev. Mol. Cell Biol., 12:799–814.
Schwartz YB e Pirrotta V (2007). Polycomb silencing mechanisms and the management of
genomic programmes. Nat. Rev. Genet., 8(1):9-22.
Seo PJ et al. (2011). Two splice variants of the IDD14 transcription factor competitively form
nonfunctional heterodimers which may regulate starch metabolism. Nature
communications, 2:303.
90
Shook MS e Richards EJ (2014). VIM proteins regulate transcription exclusively through the
MET1 cytosine methylation pathway. Epigenetics, 9(7):980-986.
Silveira AB et al. (2013). Extensive Natural Epigenetic Variation at a De Novo Originated
Gene. PLoS Genetics, 9(4):e1003437.
Slotkin RK e Martienssen R (2007). Transposable elements and the epigenetic regulation of
the genome. Nat. Rev. Genet. 8:272–285.
Slotkin RK et al. (2009). Epigenetic reprogramming and small RNA silencing of transposable
elements in pollen. Cell, 136(3):461-72.
Song et al. (2015). Transcription factor interaction with COMPASS-like complex regulates
histone H3K4 trimethylation for specific gene expression in plants. Proceedings of the
National Academy of Sciences, 112(9):2900–2905.
Soppe et al. (2000). The late flowering phenotype of fwa mutants is caused by gain-of-
function epigenetic alleles of a homeodomain gene. Molecular Cell, 6:791–802
Stroud H et al. (2013). Comprehensive analysis of silencing mutants reveals complex
regulation of the Arabidopsis methylome. Cell, 152:352–364.
Stroud H et al. (2014). Non-CG methylation patterns shape the epigenetic landscape in
Arabidopsis. Nature Struct. Mol. Biol., 21:64–72.
Sweatt et al. (2013). An Overview of the Molecular Basis of Epigenetics. In Epigenetic
Regulation in the Nervous System (First Edit, pp. 3–33). Elsevier.
Teixeira et al. (2009). A role for RNAi in the selective correction of DNA methylation
defects. Science, 323:1600–1604.
Teixeira FK e Colot V (2009). Gene body DNA methylation in plants: a means to an end or
an end to a means? The EMBO journal, 28(8):997-8.
Teixeira FK e Colot V (2010). Repeat elements and the Arabidopsis DNA methylation
landscape. Heredity, 105:14–23.
Tsutsui et al. (2015). The Plant-Specific RAB5 GTPase ARA6 is Required for Starch and
Sugar Homeostasis in Arabidopsis thaliana. Plant Cell Physiol., 56(6):1073-83.
Vire et al. (2006). The Polycomb group protein EZH2 directly controls DNA methylation.
Nature, 439(7078):871-4.
Waddington CH (1942). The epigenotype. Endeavor 1:18-20.
Waddington CH (1957). The Strategy of the Genes (Geo Allen & Unwin, London, 1957).
Wang H et al. (2012). Widespread plasticity in CTCF occupancy linked to DNA methylation.
Genome Res., 22(9):1680–1688.
91
Weigel D e Colot V. (2012). Epialleles in plant evolution. Genome Biol., 13:249.
Williams BP et al. (2015). Methylation-Sensitive Expression of a DNA Demethylase Gene
Serves As an Epigenetic Rheostat. PLoS Genet., 11(3):e1005142.
Yao Q et al. (2012). Heterologous expression and purification of Arabidopsis thaliana VIM1
protein: in vitro evidence for its inability to recognize hydroxymethylcytosine, a rare
base in Arabidopsis DNA. Protein Expr. Purif., 83:104–11.
Yu A et al. (2013). Dynamics and biological relevance of DNA demethylation in Arabidopsis
antibacterial defense. Proc Natl Acad Sci USA,110(6):2389–2394.
Zemach A e Grafi G (2007). Methyl-CpG-binding domain proteins in plants: interpreters of
DNA methylation. Trends Plant Sci., 12(2):80-5.
Zemach A et al. (2010). Genome-Wide Evolutionary Analysis of Eukaryotic DNA
Methylation. Science, 328(5980):916-919.
Zemach A et al. (2013). The nucleosome remodeler DDM1 allows DNA methyltransferases
to access H1-containing heterochromatin. Cell, 153(1):193–205.
Zhang e Jacobsen (2006). Genetic Analyses of DNA Methyltransferases in Arabidopsis
thaliana. Cold Spring Harbor Symposia on Quantitative Biology, 71(0):439–447.
Zhang H e Zhu JK (2012). Active DNA demethylation in plants and animals. Cold Spring
Harb. Symp. Quant. Biol., 77:161–173.
Zhang M et al. (2010). DNA cytosine methylation in plant development. J Genet Genomics,
37(1):1-12.
Zhang X et al. (2006). Genome-wide high-resolution mapping and functional analysis of
DNA methylation in Arabidopsis. Cell, 126:1189–1201.
Zhang Y-Y et al. (2013). Epigenetic variation creates potential for evolution of plant
phenotypic plasticity. New Phytol., 197:314–322.
Zheng X et al. (2008). ROS3 is an RNA-binding protein required for DNA demethylation in
Arabidopsis. Nature, 455(7217):1259–1262.
Zhong X et al. (2012). DDR complex facilitates global association of RNA polymerase V to
promoters and evolutionarily young transposons. Nature Struct. Mol. Biol., 19:870–875.
Zhong S et al. (2013). Single-base resolution methylomes of tomato fruit development reveal
epigenome modifications associated with ripening. Nature Biotechnology, 31:154–159.
Zhu J et al. (2007). The DNA glycosylase/lyase ROS1 functions in pruning DNA methylation
patterns in Arabidopsis. Curr. Biol., 17:54–59.
Zhu JK (2009). Active DNA demethylation mediated by DNA glycosylases. Annu. Rev.
Genet., 43:143–166.
92
Zhu, Y et al. (2013). SWI/SNF chromatin-remodeling complex acts in noncoding RNA-
mediated transcriptional silencing. Mol. Cell, 49:298–309.
Zilberman D et al. (2007). Genome-wide analysis of Arabidopsis thaliana DNA methylation
uncovers an interdependence between methylation and transcription. Nat. Genet.,
39:61–69.
93
ANEXOS
Anexo I – Primeira página do artigo publicado “An efficient method for simultaneous
extraction of high-quality RNA and DNA from various plant tissues”
94
Anexo II – Declaração de Bioética e Biossegurança
95
Anexo III – Declaração referente aos direitos autorais
