Heterochromatin dynamics in Epithelial-to-Mesenchymal Transition

by user

Category: Documents





Heterochromatin dynamics in Epithelial-to-Mesenchymal Transition
Departament de Ciències Experimentals i de la Salut
Facultat de Ciències de la Salut i de la Vida
Universitat Pompeu Fabra
Tesi doctoral 2014
Heterochromatin dynamics in
Epithelial-to-Mesenchymal Transition
Dissertation presented by Alba Millanes Romero
for the degree of Doctor of Philosophy
Work carried out under the supervision of Drs. Sandra Peiró Sales and
Antonio García de Herreros Madueño in the Epithelial-to-Mesenchymal
Transition and Tumor Progression Group in the Cancer Research Program
in the Institut Hospital del Mar d’Investigacions Mèdiques (IMIM)
Dr. Sandra Peiró Sales
(thesis co-director)
Dr. Antonio García de Herreros Madueño
(thesis co-director)
Alba Millanes Romero
(PhD student)
Als meus pares, pel seu suport incondicional.
Al Manel, per ser-hi sempre.
“La humilitat més sincera per a un científic és acceptar
que res no és impossible”
Marc Levy, El primer dia.
Fotografia de Santiago Millanes
1. Chromatin structure and organization
2. Genome organization and transcription
2.1. Chromatin dynamics
2.2. Histone modifications and DNA methylation
3. Euchromatin vs. heterochromatin
3.1. Heterochromatin protein 1 (HP1)
3.2. Pericentromeric heterochromatin
3.2.1. Pericentromeric heterochromatin formation
3.2.2. Pericentromeric heterochromatin transcription
4. Cancer and tumor progression
4.1. Epithelial-to-Mesenchymal Transition (EMT)
4.1.1. Physiological EMT
4.1.2. Pathological EMT
4.1.3. Signaling EMT
4.1.4. EMT inducers
38 Snail1 transcription factor
38 Lysyl oxidase-like 2 (LOXL2)
1. Snail1 is essential for pericentromeric heterochromatin
maintenance and organization in mesenchymal cells
2. Snail1 regulates pericentromeric transcription
3. Snail1 is enriched in pericentromeric regions and interacts with
4. Pericentromeric transcription is linked to H3 oxidation
5. Pericentromeric transcription is tightly regulated during EMT
6. Pericentromeric transcription regulation is essential for a complete
Snail1 and heterochromatin organization
Snail1 and heterochromatin transcription
Major satellite regulation during EMT
HP1 dynamics during EMT
Snail1 pericentromeric regulation in other cell contexts
Chromatin dynamics and cancer
Cell Lines
Transfection procedures
Retroviral and lentiviral Infection
Cloning procedures
Micrococcal digestion
Salt Extraction Experiments
Subcellular Fractionation
Chromatin Immunoprecipitation (ChIP) experiments
Western Blot
Co-immunoprecipitation assays
Genomic DNA extraction
RNA extraction procedures
Real-time RT-PCR
Microarray gene expression analysis
Migration and Invasion assays
Table of figures
Figure I1. Atomic structure of the nucleosome. ........................................ 3
Figure I2. Chromatin organization in the mammalian nucleus .................. 4
Figure I3. Conserved role of insulators in nuclear organization ................ 8
Figure I4. Summary of histone modifications and histone modifying
enzymes.................................................................................................... 10
Figure I5. Map of main PTMs in histone tails ........................................... 11
Figure I6. Models for the functional outcomes of reading modified
histones .................................................................................................... 12
Figure I7. Properties of euchromatic and heterochromatic regions........ 15
Figure I8. Mouse HP1 domains and interaction partners ........................ 16
Figure I9. Examples of HP1 interacting partners ...................................... 18
Figure I10. Structural and functional elements of the centromere region
.................................................................................................................. 20
Figure I11. Comparative organization of centromeric (CT) and
pericentromeric (PCT) regions in fission yeast, mouse and human
chromosomes ........................................................................................... 21
Figure I12. Chromocenters in mouse cells ............................................... 22
Figure I13. Mechanisms for the initiation of heterochromatin assembly 24
Figure I14. Epithelial and mesenchymal cell features.............................. 32
Figure I15. Examples of primary EMT ...................................................... 33
Figure I16. Acquisition of the metastatic phenotype ............................... 35
Figure I17. Potential roles of EMT and MET in carcinoma progression ... 36
Figure I18. Overview of the molecular networks that regulate EMT....... 37
Figure I19. Structural domains of Snail1 transcription factor .................. 39
Figure I20. Model of Snail1 mediated transcriptional repression............ 40
Figure I21. Structural organization of lysyl oxidase family members ...... 42
Figure R1. Heterochromatin organization is compromised in the absence
of Snail1 .................................................................................................... 52
Figure R2. Snail1 is required for S-phase progression.............................. 54
Figure R3. Snail1 is required for pericentromeric heterochromatin
replication ................................................................................................ 55
Figure R4. Snail1 depletion induces genome instability .......................... 56
Figure R5.
Snail1 regulates pericentromeric heterochromatin
transcription ............................................................................................. 58
Figure R6. Snail1 binds pericentromeric regions in mouse and human
cells........................................................................................................... 60
Figure R7. Snail1 SNAG domain interacts with HP1α chromoshadow
domain ..................................................................................................... 62
Figure R8. LOXL2 favors nucleosomal compaction .................................. 63
Figure R9. There is less oxidized H3 in major satellite regions of Snail1 KO
iMEFs ........................................................................................................ 64
Figure R10. LOXL2 regulates major satellite transcription....................... 65
Figure R11. LOXL2 binds major satellite sequences and oxidizes H3....... 66
Figure R12. HP1α delocalizes from chromocenters after TGFβ treatment
.................................................................................................................. 67
Figure R13. HP1α is released from chromatin after 8h of TGFβ treatment
.................................................................................................................. 69
Figure R14. Release of HP1α from chromocenters is Snail1 and LOXL2
dependent ................................................................................................ 70
Figure R15. Snail1 up-regulation during EMT correlates with major
satellite repression ................................................................................... 71
Figure R16. Snail1 binding to major satellites after TGFβ treatment
correlates with increased H3 oxidation ................................................... 72
Figure R17. Snail1 is responsible for major satellite down-regulation
during EMT ............................................................................................... 72
Figure R18. LOXL2 regulates pericentromeric transcription during EMT
through H3 oxidation ............................................................................... 73
Figure R19. Major satellite overexpression blocks HP1α release from
chromatin during EMT ............................................................................. 75
Figure R20. Genes differently regulated by TGFβ in NMuMG-Major cells
are mainly associated to cancer and EMT related pathways................... 76
Figure R21. Genes involved in EMT are differently regulated by TGFβ in
NMuMG-Major cells. ................................................................................ 77
Figure R22. Major satellite cells have decreased migration and invasion
properties ................................................................................................. 78
Figure R23. HP1α and major satellite transcripts knock-down does not
affect mesenchymal genes induction after TGFβ treatment ................... 79
Figure D1. Snail1 ChIP-Seq predicted binding sites analysis .................... 90
Figure CR1. Model for pericentromeric heterochromatin reorganization
during EMT ............................................................................................. 106
3C: chromosome conformation capture
ATRx: alpha thalassemia/mental retardation syndrome
BER: base excision repair
BRG1: SWI/SNF related transcriptional activator
BSA: bovine serum albumin
CAF-1p150: chromatin assembly factor-1 p150 subunit
CBX: chromobox
CD: chromodomain
CDH1: E-cadherin gene
ChIP: chromatin immunoprecipitation
ChIP-Seq: ChIP sequencing
CLDN: claudin gene
Clr4: cryptic loci regulator 4
CSC: cancer stem cell
CSD: chromoshadow domain
CT: control
CTCF: CCCTC-binding factor
DAPI: 4',6-diamidino-2-phenylindole
DNA: deoxyribonucleic acid
Dnmt3a/Dnmt3b: DNA methyltransferase 3a and 3b
DSP: desmoplakin gene
ECM: extracellular matrix
EGF: epithelial growth factor
EMT: Epithelial-to-Mesenchymal Transition
ETaR: endothelin-A receptor
Ezh2: enhancer of zeste 2
FAK: focal adhesion kinase
FBS: fetal bovine serum
FGF: fibroblast growth factor
GFP: green fluorescent protein
GSK3β: glycogen-synthase kinase-3β
HAT: histone acetyltransferase
HDAC: histone deacetylase
HDM: histone demethylase
HGF: hepatocyte growth factor
HIF: hypoxia-inducible factor
HMT: histone methyltransferase
HP1: heterochromatin protein 1
HSF1: heat-shock factor 1
IAP1: intracisternal A particle 1
IgG: immunoglobulin
iMEF: immortalized mouse embryonic fibroblast
INCENP: inner centromere protein
Kap-1/Tif1β: Kruppel-associated box
/transcriptional intermediary factor 1β
Ki-67: cell proliferation antigen of monoclonal antibody Ki-67
KO: knock out
Ku70: 70K autoantigen
LAD: lamin-associated domain
LAP2: lamina-associated polypeptide 2
LEF-1/TCF4: lymphoid-enhancer-binding factor/T-cell factor-4
LINE: long interspersed nuclear domain
lncRNA: long non-coding RNA
LOCK: large organized chromatin K9-modifications
LOXL2: lysil oxidase-like 2
LSD1/2: lysine specific demethylase
MAPK: mitogen-activated protein kinase
MET: mesenchymal-to-epithelial transition
MITR: myocyte enhancer factor2 (MEF2) interacting transcription
MMP: matrix metalloproteinase
MSC: mesenchymal stem cell
NAD: nucleolus-associated domain
ncRNA: non-coding RNA
NER: nucleotide excision repair
NES: nuclear export signal
NF-κB: nuclear factor-κB
NL: nuclear lamina
OCLN: occluding gene
ORC1-6: origin recognition complex 1-6
PAR6: partitioning-defective protein-6
Pc: polycomb
PCNA: proliferating cell nuclear antigen
PCR1/2: polycomb repressive complex 1/2
PI3K: phosphatidyl-inositol 3-kinase
PIM1: proviral integration site 1 (pim-1) oncogene
PKB: protein kinase-B
pMEF: primary mouse embryonic fibroblast
Psc3: cohesion subunit Psc3
PTM: posttranslational modification
RA: retinoic acid
Rb: retinoblastoma protein
RITS: RNA-induced transcriptional silencing
RNA: ribonucleic acid
RNAi: interfering RNA
ROS: reactive oxygen species
satIII: human satellite III repeat
SD: standard deviation
SINE: short interspersed nuclear domain
siRNA: small interfering RNA
SP100: nuclear autoantigenSpeckled 100 kD
SRCR: scavenger receptor cysteine rich
Suv39h: suppressor of variegation 3-9 homolog
Suz12: suppressor of zeste 12
Swi6: switching gene 6. Refers to the S. pombe HP1 ortholog
TAD: topological associated domain
TAFII130: TATA-binding protein associated factor p130
TAK1: TGFβ- activated kinase-1
TF: transcription factor
TFGβ: transforming growth factor β
TGFβR: TGFβ receptor
tonEBP: tonicity enhancer-binding protein
Wnt: wingless-related integration site
WntR: Wnt receptor
WT: wild type
Although heterochromatin is enriched with repressive traits, it is actively
transcribed, giving rise to large amounts of non-coding RNAs. These
transcripts are responsible for the formation and maintenance of
heterochromatin, but little is known about how their transcription is
regulated. In this thesis we show that Snail1 transcription factor
represses mouse pericentromeric transcription and regulates
heterochromatin organization through the action of the H3K4 deaminase
LOXL2. Snail1 has a key role in epithelial-to-mesenchymal transition
(EMT). We show that, also during this process, Snail1 is responsible for
pericentromeric transcription regulation. At the onset of EMT, one of the
major structural heterochromatin proteins, HP1α, is transiently released
from heterochromatin foci in a Snail1/LOXL2 dependent manner,
concomitantly with a down-regulation of major satellite transcription.
Moreover, prevention of major satellite transcripts down-regulation
compromises the migratory and invasive behaviour of EMT resulting
mesenchymal cells. We propose that Snail1 and LOXL2 regulate
heterochromatin during this process, which may be crucial to allow the
genome reorganization required to complete EMT.
Tot i estar enriquida en marques repressores, l’heterocromatina es
transcriu activament i dóna lloc a grans quantitats d’ARNs no codificants.
Aquests trànscrits són responsables de la formació i el manteniment de
l’heterocromatina, però com es regula la seva transcripció segueix sent
quelcom poc clarificat. En aquesta tesi demostrem que el factor de
transcripció Snail1 reprimeix la transcripció pericentromèrica en cèl·lules
de ratolí i regula l’organització de l’heterocromatina a través de l’acció de
la LOXL2, que deamina l’H3K4. Snail1 té un paper clau en la transició
epiteli-mesènquima (EMT). Aquí demostrem que, també durant aquest
procés, Snail1 és responsable de la regulació de la transcripció
pericentromèrica. A l’inici de l’EMT, l’HP1α, una de les principals
proteïnes estructurals de l’heterocromatina, es desprèn de forma
transitòria de l’heterocromatina. Aquest esdeveniment està regulat per
Snail1 i LOXL2 i coincideix amb una disminució de la transcripció
pericentromèrica. El bloqueig de la baixada dels trànscrits durant l’EMT
compromet les capacitats migratòries i invasives de les cèl·lules
mesenchimals que en resulten. Així doncs, proposem que Snail1 i LOXL2
regulen l’heterocromatina durant aquest procés, i així permeten que
tingui lloc la reorganització genòmica que deu ser necessària per tal que
es completi la EMT.
In eukaryotic cells, DNA exists in the nucleus forming the chromatin, a
complex of DNA and a particular group of proteins called histones. The
basic repeating structural and functional unit of chromatin is the
nucleosome, which consists in 147 base pairs (bp) of DNA wrapped
around the nucleosome core particle that contains one (H3)2-(H4)2
tetramer and two H2A-H2B histone dimers 1 (Figure I1). The linker
histone H1 binds to the core particles and protects an additional ∼20 bp
of DNA.
Figure I1. Atomic structure of the nucleosome. Each strand of DNA is shown in
different colors. The DNA makes 1.7 turns around the histone octamer.
Nucleosome core particle is composed by a tetramer of H3 (green) and H4
(yellow), and two dimers of H2A (red) and H2B (pink) .
Nucleosomes are joined one to each other by the DNA that runs between
them, which is known as “linker” DNA. Its length ranges between 20 and
90 bp and varies among different species and tissues3; variation in the
length of linker DNA is important for the diversity of gene regulation4.
The DNA-nucleosome complex forms a 10 nm diameter fiber resembling
“beads on a string”5(Figure I2-e). Classically, it was proposed that the 10
nm chromatin fiber forms a higher order helical fiber 30 nm in diameter
1. Chromatin structure and organization
(Figure I2-d). However, although it has been extensively studied in vitro,
evidence for the existence of the 30 nm fiber in vivo is limited. In recent
years, an increasing amount of data suggests that chromatin organization
above the 10 nm fiber probably does not exist in most mammalian cells
and can only be found in very particular biological contexts, where
heterochromatic transcriptional repression and compaction prevail6.
Alternatively, the “polymer melt” model has been proposed to explain
chromatin organization inside the nucleus of mammary cells as an
alternative to the 30 nm fiber, and every day more evidences seem to
support it. In this model, chromatin is described as a dynamic disordered
state comparable with a “polymer melt”, where nucleosomes that are
not linear neighbors on the DNA strand interact within a chromatin
region7 (Figure I2-d) and organize into a series of small globules to form a
highly compact state termed “fractal globule”8(Figure I2-c).
Figure I2. Chromatin organization in the mammalian nucleus. (a) Chromosomes
are organized in chromosome territories. (b) Chromosomes territories are
comprised of fractal globules, and fractal globules from adjacent chromosome
territories can interdigitate. (c) Chromatin fibers interact (i) within a fractal
globule (frequent), (ii) between fractal globules of the same chromosome
territory (rare), or (iii) between adjacent chromosome territories (very rare). (d)
Chromatin may form a 30 nm fiber with a solenoid or zigzag structure, or it may
present “polymer melt” organization. (e) Chromatin is resolved as a 10 nm
“beads on a string” fiber consisting of nucleosomes .
However, how the chromatin is organized into a higher order structure
inside the nucleus and how and why the spatial organization of the
Historically, chromatin was divided into two domains: euchromatin and
heterochromatin, on the basis of differential compaction at interphase.
Heterochromatin was initially defined as the portion of the genome that
retains deep staining with DNA-specific dyes as the dividing cell returns
to interphase from metaphase. On the other hand, euchromatin was
defined as the actively transcribed chromatin9. However, apart from
being subdivided into functional domains, microscopic studies and other
type of experiments started to suggest the idea that, in the nucleus of
eukaryotic cells, interphase chromosomes occupy distinct chromosome
territories10(Figure I2-a).
Recent advances in genomic technologies, such as the development of
the Chromosome Conformation Capture (3C)11 and 3C-related genomewide techniques, and even more recently the development of Hi-C
technique12, have let to rapid advances in the study of three-dimensional
genome organization in vivo, based on the information they give about
cis and trans chromatin interactions. As a result, the idea that
chromosomes organize in chromosome territories has been recently
further reinforced by the tendency of distant loci on the same
chromosome to be near one another in space, as has been shown by
these long-range genome-wide chromatin interaction techniques8. This
can be easily understood if each chromosome, rather than being
intertwined, occupies its own distinct region in the nucleus6.
Moreover, Hi-C experiments have also shown that within the fractal
globule, chromatin is organized into large, megabase-sized chromatin
interaction domains, which are called “topological domains” or
“topological associated domains” (TAD). These domains correlate with
regions of the genome that constrain the spread of heterochromatin and
their boundaries are enriched for the insulator binding protein CTCF,
housekeeping genes, transfer RNAs and short interspersed element
(SINE) retrotransposons, indicating that these factors may have a role in
establishing the topological domain structure of the genome.
Interestingly, TADs are stable across different cell types and highly
genome is intimately linked to its biological function has been for many
years poorly understood.
conserved across species, indicating that they are an inherent property of
mammalian genomes13.
However, other types of domains have been lately described, and not all
of them seem to be conserved and maintained in different cell types.
Chromatin interactions within these domains have been shown to
change depending on the cell context, and seem to be associated to
specific transcriptional programs14,15.
2. Genome organization and transcription
Cell type specific transcriptional regulation is crucial in order to maintain
cell identity throughout the lifetime of an organism, but it must be
flexible enough to allow for responses to endogenous and exogenous
stimuli. This regulation is mediated not only by molecular factors such as
transcription factors and histone and DNA modifications, but also at the
level of chromatin and genome organization.
2.1. Chromatin dynamics
Chromatin dynamics and nuclear organization are recently arising as
novel transcriptional key regulators, since gene movement and
localization seem to be tightly linked with transcriptional activity.
Thus, although TADs are conserved in different cell types and even
between species, it has also been proposed that specific regions within
TADs may be dynamic, potentially taking part in cell-type specific
regulatory events13. Like this, although TADs as a whole do not change,
the internal TAD contacts rearrange upon ES cell differentiation, affecting
differentially regulated genes, supporting the link between chromatin
structure and transcription16.
Moreover, other types of sub-domains have also been described, which
are more dynamic and variable between cell types, and seem to be
Polycomb (Pc) group proteins and H3K27me3 form particular domains,
known as “Polycomb bodies” and are responsible for gene clusters
silencing during development. However, they are dynamic structures
that, depending on cell type, can be partially or entirely reexpressed15,17.
Furthermore, large organized chromatin K9-modifications (LOCKs) are
enriched in H3K9me2 and are highly conserved between human and
mouse, but are cell type and tissue specific. Interestingly, genes
differently located within or outside LOCKS in different tissues were
differently expressed, being largely silenced the ones within LOCKs and
showing a broad range of expression the ones outside them18.
Lamin-associated domains (LADs) are formed by genomic regions that
interact with the nuclear lamina (NL) and are characterized by low gene
density and transcriptional repression19. NL interaction pattern has been
shown to be in part cell type specific20, contributing again to the idea that
dynamic genomic subdomains reorganize based on cell specific
transcriptional requirements. Similar to LADs, nucleolus-associated
domains (NADs) preferentially contain repressed genes and show
enrichment for repressive histone marks21,22.
Interestingly, not only repressed domains have been described in
eukaryotic nuclei, but also transcription has been shown to be spatially
organized into discernible nuclear structures termed “transcriptional
factories”23, but how genes are dynamically targeted to them remains
poorly understood24,25.
Understanding how all these domains are formed and how do they
change in particular cell contexts is still something under study.
Insulators, which are multi-protein DNA complexes that mediate longrange physical interactions in the nucleus seem to be key elements in the
partition of chromatin into structural and functional domains25 (Figure
I3). However, further studies will be required to identify and characterize
new players involved in genome organization.
associated to active or silent transcription. Local and long-range contacts,
together with clustering of specific subsets of proteins are responsible
for the formation of these sub-domains.
Figure I3. Conserved role of insulators in nuclear organization. Insulators in
yeast (orange), Drosophila (green and yellow), and mammals (brown and
orange) mediate long-range inter- and intrachromosomal interactions
important for gene regulation; they are associated to Polycomb (Pc) body
repression (blue), to transcription factories (transcription factors in pink) and
to lamina-associated domains (LADs) .
The dynamism of all these chromatin sub-domains, the evidences for
clustering of active and inactive regions, as well as the looping-out from
chromosome territories of specifically actively transcribed genes suggests
that, beyond the organization of chromatin in chromosome territories,
the radial position of genes within the nucleus is clearly implicated in
gene regulation.
Although all cells have the same genome, cell identity is established by
the expression of a particular transcriptome, which means that each cell
type expresses and represses particular subsets of genes.
In a specific cell context, genes can be either actively transcribed or
silenced, which has been associated with different states of chromatin
compaction. In this way, “open” or decondensed chromatin is usually
associated to active transcription, whereas “closed” or condensed
chromatin is associated to silenced or repressed transcription.
In the last decades, epigenetics have arisen as one of the main regulators
of chromatin transcription. By definition, epigenetics enclose those
processes that ensure the inheritance of variation (“-genetic”) above and
beyond (“epi-”) changes in the DNA sequence. In other words, the study
of mitotically and/or meiotically heritable changes in gene function that
cannot be explained by changes in DNA sequence26. Two of the main
epigenetic mechanisms in mammals are histone modifications and DNA
methylation, which influence each other during development and during
other cell processes.
The ability of chromatin to condense can be regulated in part by
posttranslational modification (PTM) of the N-terminal tails of the
histones, which overhang from the DNA and become exposed to
chromatin modifying enzymes27. These enzymes are normally highly
specific and catalyze specific residue modifications at particular amino
acid positions in histone tails (Figure I4).
Since the epigenetic field appeared and started to expand at least 15
types of PTMs have been reported acting on 80 different amino acid
residues on histones28, and are either associated with gene activation or
2.2. Histone modifications and DNA methylation
Figure I4. Summary of histone modifications and histone modifying enzymes.
(a) Histone acetylation at lysines, generally an active mark, is catalyzed by
histone acetyltransferases (HATs) and removed by histone deacetlyases
(HDACs). (b) Histone methylation at lysine and arginine residues is catalyzed by
histone methyltransferases (HMTs) and removed by histone demethylases
(HDMs). Histone methylation has been linked to both transcriptional activation
and transcriptional repression. Methylation can occur in mono-, di-, or even trimethylated states. (c) Histone phosphorylation at serine residues, generally
linked to transcriptional activation, is catalyzed by protein kinases, whereas
phosphorylation marks are removed by protein phosphatases .
Methylation, acetylation and phosphorylation are the modifications that
have been more extensively described and characterized (Figure I5).
However, histones can suffer many other modifications including
ubiquitylation, proline isomerization, propionylation, butyrylation,
formylation, sumoylation, citrullination, hidroxilation, ADP ribosylation
and crotonylation28. Moreover, deamination of lysine has been recently
reported as a new PTM catalyzed by LOXL2, a lysine oxidase protein.30
Figure I5. Map of main PTMs in histone tails. Acetylation (blue), methylation
(red), phosphorylation (yellow) and ubiquitylation (green). The number in grey
under each amino acid represents its position in the sequence .
As previously mentioned, histone modifications are associated with both
gene silencing and activation, depending on the nature of the
modification and the specific amino acid modified. In this way, two major
mechanisms are thought to contribute to the regulation of chromatin
transcription by histone modifications32. On one hand, histone PTMs may
have a structural role based on the fact that resultant charge density in
histone tails may impact on their interactions with DNA. Thus, acetylated
histone tails would be expected to propagate a more open chromatin
state. Moreover, PTMs may alter inter-nucleosomal interactions, thereby
regulating chromatin structure and the access of DNA-binding proteins
such as transcription factors33.
On the other hand, accumulating evidence suggests that key function of
histone modification is to signal for recruitment or activity of
downstream effectors34. In this way, histone PTMs would be able to
regulate chromatin structure and function by recruiting PTM-specific
binding proteins, the so-called “histone readers”, which recognize
modified histones via specialized structural folds such as bromodomains
and chromodomains among others, which bind acetylated and
methylated lysine residues, respectively35,36. Alternatively, histone PTMs
can also function by inhibiting the interaction of specific binders with
chromatin. The functional readout of a particular PTM will depend on the
specific readers and effectors that are recruited there, and its particular
activity. Based on their functional outcome, readers are classified in four
groups34 (Figure I6).
Figure I6. Models for the functional outcomes of reading modified histones.
PTM readers are classified in four groups: architectural proteins, chromatin
remodelers, chromatin modifiers and adaptors .
Histone modifications induce changes in protein interactions between
chromatin and its binding partners. These changes contribute to the
establishment of a particular chromatin environment that correlates with
different transcriptional states, which in turn translate into different
biological outcomes.
Besides histone modifications, DNA methylation is one of the other main
epigenetic mechanisms in eukaryotic cells. In the mammalian genome,
DNA methylation occurs by covalent modification of the fifth carbon (C5)
The DNA methylation pattern is erased in the early embryo and then reestablished in each individual at approximately the time of
implantation39,40. De novo DNA methylation is carried out by the DNA
methyltransferase enzymes DNMT3A and DNMT3B41. On the other hand,
DNMT1 is considered the maintenance DNMT and is required to
methylate hemimethylated sites that are generated during
semiconservative DNA replication37. In contrast, DNA demethylation can
be achieved either passively, by simply not methylating the new DNA
strand after replication, or actively, by replication-independent processes
associated to base excision repair (BER) and nucleotide excision repair
(NER) pathways42.
DNA methylation does not seem to be as dynamic as histone
modifications. Therefore, the possible combinations between DNA
methylation and different types of repressive histone modifications will
establish several repressed chromatin states, which will be more or less
easily reversible depending on their epigenetic component.
Taking all this together, chromatin structure is characterized by the
degree of chromatin condensation, the location within the nuclear
architecture and the type of histone modifications. All this will contribute
to the generation of a particular chromatin state, directly linked with a
specific transcriptional program and biological outcome.
in the cytosine base and the majority of these modifications are present
exclusively at CpG dinucleotides37. DNA methylation associates with
stable long-term repression in comparison to histone tail methylation,
which is responsible for reversible local formation of heterochromatin38.
3. Euchromatin vs. heterochromatin
Euchromatin and heterochromatin have different patterns of histone
modifications, are associated with different modes of nucleosome
packaging and therefore, have differences in higher-order packaging and
nuclear organization.
Euchromatin is less condensed, more accessible and generally more
easily transcribed. It is mainly located at chromosome arms, is enriched
in gene sequences, replicates in early S-phase and suffers recombination
during meiosis (Figure I7).
On the other hand, heterochromatin is typically highly condensed and
more inaccessible43 and it can be subdivided into constitutive and
facultative heterochromatin. In higher eukaryotes, only regions
important for genome integrity such as telomeres and centromeres, as
well as repetitive and noncoding sequences are kept stably
heterochromatinized and referred to as constitutive heterochromatin,
which replicates in late S-phase and is protected against meiotic
recombination (Figure I7).
In contrast, facultative heterochromatin can be molecularly defined as
condensed transcriptionally silent chromatin regions that decondense
and allow transcription depending on developmental states, specific cellcycle stages or nuclear localization changes from the center to the
periphery or vice versa. This means that facultative heterochromatin is
transcriptionally silent but retains the potential to interconvert between
heterochromatin and euchromatin44.
At an epigenetic level, heterochromatin is distinguished by DNA
methylation, histone hypoacetylation and methylation of H3K9 and
H4K20. Heterochromatin is also enriched in heterochromatin protein 1
(HP1), which is essential for its formation and maintenance43.
Figure I7. Properties of euchromatic and heterochromatic regions. Cluster of
general important properties of euchromatin and heterochromatin are
specified, though there are exceptions in every instance. Not all centromeric and
telomeric domains exhibit these characteristics, but they are especially
consistently observed in pericentromeric heterochromatin, found in the regions
that flank the centromeres of many eukaryotic chromosomes. Adapted .
3.1. Heterochromatin protein 1 (HP1)
HP1 is a highly conserved protein, which has homologues in various
organisms, ranging from S. pombe (Swi6) to mammals, in which three
HP1 isoforms, HP1α, HP1β and HP1γ have been identified45. The HP1
family of proteins is encoded by a class of genes known as the
chromobox (CBX) genes: HP1α is encoded by the Chromobox homolog 5
(CBX5), HP1β by CBX1 and HP1γ by CBX346.
Although HP1 isoforms have similarities in their amino-acid sequences
and structural organization (Figure I8-a), there are some differences in
their localization. They are primarily associated with centromeric
heterochromatin, but HP1β and especially HP1γ also localize to
euchromatic sites47.
These proteins are small (around 25 kDa) and contain a conserved Nterminal region that is known as the chromodomain (chromatinorganization modifier), followed by a variable hinge region and a
conserved C-terminal chromoshadow domain (Figure I8-b).
All proteins containing chromodomains (CD) can characteristically alter
the structure of chromatin. In particular, the CD of HP1 binds specifically
to methylated lysine 9 of histone H3 (H3k9me3)48,49. The hydrophobic
pocket of the CD provides the appropriate environment for docking onto
this methylated residue50,51.
Figure I8. Mouse HP1 domains and interaction partners. (a) Amino-acid
sequence of mouse heterochromatin protein 1 (HP1) isoforms α, β and γ.
Asterisks indicate amino acids conserved among the three subtypes. Three
domains are distinguished: the N-terminal chromodomain (red), the central
hinge region (pale gree) with RNA-binding domain (mid green) and the
chromoshadow domain (blue). (b) HP1α domains of interaction with selected
partners that are potentially important for the stability of heterochromatin
domains .
Finally, the chromoshadow domain (CSD) is involved in homo- and/or
heterodimerization and interaction with other proteins. The overall
structure of the CSD domain is very similar to that of the CD. However, it
has a particular domain that is involved in dimerization, with conserved
residues that are unique to the CSD and localize at the dimer interface.
As a result, this dimer structure creates a nonpolar groove that can
accommodate HP1-interacting proteins containing the consensus
sequence PXVXL55. Like this, almost all HP1 partners described interact
through the CSD.
So far, many proteins have been shown to interact with HP1 family of
proteins (Figure I9). Some of them, like Suv39h1, Dnmt3a and Dnmt3b
are clearly involved in the most common of HP1 functions that is the
formation of heterochromatin. Suv39h1 is the histone methyltransferase
(HMT) that catalyzes H3K9 trimethylation56. Interestingly, Suv39h1 is able
to interact with HP1 through its CSD domain57,58. In this way, it is
proposed that, once Suv39h1 methylates H3K9, HP1 binds there and is
able to recruit more Suv39h1 establishing a ‘self-sustaining’ loop that
would explain how heterochromatin may propagate once an initiating
site has been established48,49. This model has been also extended to DNA
methylation, as both HP1 and Suv39h1 recruit DNMTs59.
However, HP1 is also able to interact with many other proteins involved
in cellular processes ranging from transcriptional regulation, chromatin
modification and replication to DNA repair, nuclear architecture and
chromosomal maintenance. Interestingly, these interactions can occur in
either a manner specific to one HP1 isoform or universally with all three
isoforms, which may be the clue to explain their different localizations
and functions.
The linker or hinge region is the least conserved domain of HP1 proteins,
which may explain why distinct HP1 isoforms are targeted to different
locations. This hinge region is responsible of HP1α RNA-binding
properties52, but it has no obvious homology to known RNA-binding
proteins. In addition, the hinge region of HP1 can also bind DNA and
chromatin without obvious sequence specificity53,54.
HP-1 variant
Transcriptional regulators of chromatin-modifying proteins
Histone H1
Histone H3
HP1, HP1 α, HP1 β, HP1 γ
Methyl K9 Histone H3
Swi6, HP1, HP1α, HP1β, HP1γ
Histone H4
HP1, HP1 α
HP1, HP1α, HP1β, HP1γ
HP1 α, HP1 γ
HP1 α
HP1α, HP1β
HP1α, HP1β, HP1γ
HP1 γ
HP1 α
HP1 α
HP1 α, HP1 β
HP1 α, HP1 γ
HP1 γ
HP1 α, HP1 γ
DNA replication and repair
CAF-1 p150
HP1α, HP1β
HP1 α, phosphoS83-HP1 γ
Other chromosome-associated proteins
HP1 α, HP1 γ
HP1 α, HP1 β, HP1 γ
HP1 α, HP1 β, HP1 γ
Nuclear structure proteins
Nuclear envelope
HP1 α, HP1 β, HP1 γ
Lamin B receptor
HP1 α, HP1 β, HP1 γ
Lamin B
HP1 β
HP1 β
CSD, Linker
Figure I9. Examples of HP1 interacting partners. “HP1” alone refers to
Drosophila HP1; HP1α, HP1β and HP1γ refer to both mouse and human unless
specified (Mm, mouse; Hs, human); Domain abbreviations: CD, chromodomain;
CSD, chromoshadow domain; ND, not determined. Adapted .
All this suggests that HP1 proteins are not only involved in
heterochromatin formation but may also have many other functions in
the cell. Accordingly, HP1 is found at telomeres, and telomere fusions
that occur in larvae lacking zygotic HP1 suggest that HP1 might function
to protect telomeres60.
Many of these modifications are found in both the CD and the CSD, as
well as in the linker domain, suggesting that they may have an important
role in modulating HP1 interactions or functions. In addition, different
modifications such as acetylation and methylation can occur in the same
residue, which suggests that specific residues of HP1 proteins may act as
“switches”, and may be directly involved in the regulation of HP1
3.2. Pericentromeric heterochromatin
Among the different constitutive heterochromatin domains, centric
heterochromatic, also known as pericentromeric heterochromatin, is the
most abundant in the genome.
Centromeres were originally defined as a cytologically visible “primary”
constriction in the chromosome, and later as chromosomal sites that
were essential for normal inheritance, which suffer greatly reduced or
absent meiotic recombination. Nowadays, centromere is defined as the
site of kinetochore formation, a proteinaceous structure on each
chromosome that is responsible for their attachment to and movement
along microtubules. The centromere is therefore essential for
chromosomal plateward prometaphase and poleward anaphase
movements63. However, in terms of terminology, the “centromere
region” also includes domains and functions present in the vicinity of a
centromere, such as the pericentromeric heterochromatin (Figure I10).
At chromatin level, the centromere region is characterized by the
presence of two different types of chromatin: centromeric chromatin and
pericentromeric heterochromatin. Both domains are enriched in
repetitive sequences, centromeric and pericentromeric repeats
respectively, and particular histone modifications and epigenetic traits64.
Lately, it has been shown that each HP1 isoform can suffer several PTMs,
such acetylation, phosphorylation, ubiquitilation and sumoylation, in a
similar way to histones61,62.
Figure I10. Structural and functional elements of the centromere region. The
centromere/inner kinetochore, outer kinetochore, centric heterochromatin and
chromosome arms, and associated functions, are shown .
Centromeric chromatin is mainly characterized by the incorporation of
the centromeric histone H3 variant (cenH3) within its nucleosomes, and
the presence of specific centromeric proteins required for kinetochore
formation. On the other hand, pericentromeric heterochromatin is
enriched in epigenetic repressive marks such as DNA methylation,
hypoacetylated histones and H3K9 and H4K20 methylation. Moreover,
pericentromeric heterochromatin is also enriched in the three isoforms
of HP1, especially α and β, and in an RNA component that seem to have a
crucial role in its formation and maintenance45. Interestingly,
pericentromeric heterochromatin also has a role in proper segregation of
sister chromatids, since its deregulation is associated to chromosome
mis-segregation and genome instability65–67.
Although centromeric and pericentromeric domains are very variable
between species, they share common features concerning their general
organization from yeast to mammals. Basically, all of them are composed
Figure I11. Comparative organization of centromeric (CT) and pericentromeric
(PCT) regions in fission yeast, mouse and human chromosomes. CT (green) and
PCT (pink) regions display a similar organization in all eukaryotes. CT regions are
enriched in cenH3 and play a direct role in spindle attachment. PCT regions,
which recruit the cohesion complex, are enriched in H3K9me3, swi6 (S. pombe)
and HP1 (mouse and human swi6 homolog). RNA molecules are also thought to
participate to the structure of CT and PCT regions. Adapted .
of repetitive elements, which can differ in sequence and size in different
species, but are enriched with the same type of proteins (Figure I11).
In S. pombe, centromeric regions are formed by a central core domain
which contains a unique AT rich sequence of ~4 kb (central core) flanked
by imperfect repeats (imr) of ~5 to 6 kb each. Pericentromeric regions
are made of two types of outer repeats: dh and dg repeats of ~5 kb each.
In mouse, centromeric and pericentromeric regions have not been fully
characterized. Centromeric regions of ~600 kb are made of a repetition
of AT-rich minor satellite motifs of 120 pb68, whereas pericentromeric
regions of ~6 Mb are made of a repetition of AT-rich major satellite
motifs of 234 bp.
In humans, centromeric regions of about ~240 kb to 5 Mb, depending on
the chromosome considered, are made of repetition of AT-rich alpha
satellite motifs of 171 bp. The size and structure of pericentromeric
regions, which also varies between chromosomes, are made of satellite
repeats of three types: type I (0.5% of the genome), type II (2% of the
genome) and type III (1.5% of the genome)69.
Pericentromeric heterochromatin of several chromosomes can cluster to
form chromocenters, a highly condensed structure characterized by
classical heterochromatin epigenetic traits. Chromocenters are especially
visible in mouse cells, were they become clearly defined just by DAPI
staining. In addition, they can also be easily visualized by HP1α, H3K9me3
or H4K20me3 immunofluorescence, among others, as well as with DNA
FISH against major satellite repeats (Figure I12).
Figure I12. Chromocenters in mouse cells. DAPI staining and endogenous
HP1α (red) and H3K9me3 (green) localization in wild-type MEFs. Adapted .
Therefore, different number of chromosomes may participate in the
formation of a chromocenter in different cell contexts. This implies a
certain degree of dynamism and flexibility to these highly condensed
domains, which may be able to reorganize under specific circumstances,
such as during the differentiation to a particular and specialized cell type.
3.2.1. Pericentromeric heterochromatin formation
The core of heterochromatin assembly pathway found in Drosophila and
mammals is conserved in S. pombe9. It involves posttranslational
modifications of histones and a common set of structural proteins. In
addition to histone deacetylation, heterochromatin assembly requires
methylation of H3K9 that provides binding site for the HP1 family of
chromodomain proteins73. In S. pombe, Chp1, Chp2, and Swi6 bind
through their chromodomain to H3K9me, which is methylated by Clr4
(homolog of human Suv39h), establishing a heterochromatic context.
However, the strategies used to initiate heterochromatin assembly may
differ depending on the chromosomal context (Figure I13)43. In particular,
constitutive heterochromatin formation is associated with the presence
of repetitive DNA elements. The repetitive nature of these elements,
rather than any specific primary DNA sequence, seems to be the trigger
and the target for heterochromatin formation. Moreover, several lines of
evidence suggest that there might be a conserved role for non-coding
RNAs (ncRNAs) and interfering RNAs (RNAi) in this context. How these
ncRNAs participate in heterochromatin formation has been partially
elucidated only in yeast74.
The number and size of chromocenters per cell varies between cell
types71,72. Interestingly, chromocenters tend to be smaller when they are
very abundant in a cell, and seem to be bigger when a lower number of
chromocenters is found per cell. For instance, lymphocyte nuclei show a
strong clustering of pericentromeric regions reflected by the small
number of chromocenters compared to fibroblasts. Accordingly,
chromocenters of lymphocytes appear to be much bigger than the ones
found in fibroblasts72.
Figure I13. Mechanisms for the initiation of heterochromatin assembly.
Heterochromatin structures can be nucleated by factors that recognize specific
DNA sequences, such as transcription factors (TF), or by the RNAi machinery that
targets repetitive DNA elements. Both mechanisms recruit histone-modifying
enzymes such HMTs and HDACs to nucleate heterochromatin at a specific site.
Swi6/HP1 binds to modified histone tails and allows heterochromatin to spread
to surrounding sequences. Red flag, H3K9me; blue flag, H3K4me .
Heterochromatin domains replicate during S-phase and the newly
synthetized DNA must conserve epigenetic heterochromatin traits. To do
so, pericentromeric regions may decondense to allow the entrance of the
DNA replication machinery but, once heterochromatin has been
replicated, it has to be re-silenced again. Interestingly, yeast
pericentromeric transcription increases during S-phase. These transcripts
are processed by the RNAi machinery and originate siRNA that target
RITS complex and other heterochromatin factors such as Clr4 to
pericentromeric domains, which will proceed to re-establish
heterochromatin structure and silencing75,76.
In mammals, pericentromeric repeats are also transcribed and some
evidences suggest that the transcripts generated may be involved in
heterochromatin formation and silencing by a mechanism similar to the
once described in yeast. In this way, mouse cells express RNA transcripts
of nearly unknown function that are homologous to both strands of the
major satellite repeats, as well as an RNA component (“structural” RNA)
that is implicated in the maintenance of pericentromeric
heterochromatin organization45. However, how these transcripts are
regulated is still not well understood, since not every component of the
fission yeast machinery has been identified in mammalian cells.
3.2.2. Pericentromeric heterochromatin transcription
Although classically it was thought that heterochromatin is completely
repressed, nowadays it has become evident that, despite its “silenced –
like” state, it is actively transcribed.
In particular, transcripts derived from pericentromeric repeats have been
detected in a wide variety of cell types and they represent a new subtype
of ncRNAs. Indeed, they should be considered as long ncRNAs (lncRNAs)
at least in mammalian cells, where long pericentromeric transcripts of up
to 8Kb have been detected64. However, little is known about their
function in the cell, beyond its role in the formation and maintenance of
heterochromatin and genome stability.
Interestingly pericentromeric transcripts are differently expressed in
different cell types and their expression is modulated in response to
specific stimuli, which suggests that tightly regulation of these transcripts
may be essential in particular cell contexts.
One of the most dramatic examples of transcriptional activation of
pericentromeric specific sequences is that which occurs in response to
cell stress. In numerous normal primary and cancer human cell lines,
heat-shock has been shown to induce satellite III sequence (satIII)
transcription primarily from the 9q12 locus, although transcription from
other pericentromeric regions has also been observed particularly in
tumor cells, independent of the cell cycle77,78. In addition, induction of
satIII expression is also triggered by a wide range of stress conditions
including cellular exposure to DNA damaging agents and oxidative stress
Nevertheless, it has been recently shown that de novo targeting of HP1α
in mouse pericentromeric regions is mediated by major satellite
transcripts. SUMOylated HP1α interacts with forward major RNAs
providing specificity to the initial targeting of HP1α to these domains.
Then, HP1α is stabilized by the recognition of H3K9me3, and together
with Suv39h is able to spread by a “self-enforcing” loop70.
and, interestingly, transcript levels and orientation vary according to the
nature of the stress signal79.
Sense and antisense pericentromeric transcription is also spatially and
temporally regulated throughout mouse embryonic development. As
early as during the 2-cell stage, there is a strand-specific burst in major
satellite transcription that rapidly decreases by the 8-cell stage.
Interestingly, both up- and down-regulation of major satellite transcripts
are necessary events for proper chromocenter organization and
developmental progression80.
Major satellite transcripts are also detected in more advanced stages of
mouse embryonic development. In 11.5-15.5 dpc (day post coitum)
embryos, pericentromeric transcripts in sense orientation, and in
antisense in minor extent, are ubiquitously distributed in various tissues,
especially in the central nervous system. In adult tissues, sense
expression of pericentromeric sequences is detected in liver and testis
but not in other tissues such as brain, colon, spleen, heart and lung, thus
revealing a specific pattern of expression, not only with regard to
embryonic stage, but also with regard to cell type81. Expression of
pericentromeric transcripts has also been described in human testis, and
it is thought that they might play an essential role in the differentiation
of germinal cells82.
In agreement with this, transcription from major satellite regions is
significantly increased during neuronal differentiation, both in vitro and
in vivo, and is accompanied by an enrichment of H3K4me3 in
heterochromatin domains, as well as increased nuclease sensitivity83.
This suggests that the structural and transcriptional state of major
satellite regions changes dramatically during neuronal differentiation.
During terminal muscle differentiation, there is also an increase in major
and minor satellite transcripts but, in this case, it is accompanied by
enhanced histone H3K9me3 and H4K20me3 across major satellite
regions and global reorganization and spatial clustering of constitutive
In concordance, retinoic acid (RA) treatment causes different effects on
major satellite transcription depending on cell type. Major satellites were
strongly repressed by RA in HeLa and P19 cells81, but were up-regulated
in ES cells induced to form embryonic bodies by RA treatment85.
So, although the function of major satellite transcripts is still rather
enigmatic, strong evidences point that there is a link between major
satellite transcription and differentiation and cell type specification.
However, many efforts remain to be done to further understand how
major satellite transcripts may be able to specify or determine cell fate
Major satellite transcription has also been shown to be cell cycle
dependent. In proliferative mouse cells, two different populations of
major satellite transcripts accumulate at different times during cell cycle.
Small RNA species are exclusively synthesized during mitosis and are
rapidly eliminated during mitotic exit, suggesting they may be involved in
reinforcing the heterochromatin structure during the late stages of
mitosis or assisting the reloading of HP1 during anaphase. Alternatively, a
more abundant population of large, heterogeneous transcripts is induced
late in G1-phase and their synthesis decreases during mid S-phase, which
is coincident with pericentromeric heterochromatin replication86. This
observation strongly suggests that, similar to what is described in S.
pombe75, pericentromeric transcripts may also play a role in the reformation of pericentromeric heterochromatin after replication in
mammalian cells64.
The complex expression pattern of mouse and human pericentromeric
repeats suggests the existence of equally complex transcriptional and
post-transcriptional regulation of these transcripts. Controversy exists
concerning the size of these pericentromeric transcripts, since both
mouse and human pericentromeric transcripts are either detected as
very long transcripts or as short species, which raises the possibility that
These conflicting results suggest that major satellite transcription
regulation may be more complex than initially thought, and that it may
not only depend on changes in histone modifications in pericentromeric
regions, but also on cell specific additional factors.
the latter could be generated from long precursors through posttranscriptional mechanisms involving splicing.
Furthermore, the mechanism through which pericentromeric
transcription is ultimately regulated is still poorly understood. The
impossibility of defining conventional promoter regions in
heterochromatin domains, has classically nearly discarded the possibility
of transcription factors regulating pericentromeric transcription.
However, several transcription factors have been shown to bind
pericentromeric heterochromatin and regulate its transcription.
The first transcription factors identified to be involved in human
pericentromeric transcription were Heat-Shock Factor 1 (HSF1)87 and
tonicity Enhancer-Binding Protein (tonEBP)79. Upon stress, HSF1 directly
binds to satIII sequences at the 9q12 locus and promotes transcription,
since its absence prevents accumulation of satIII transcripts in heatshocked cells. TonEBP also accumulates at the 9q12 locus and is required
for induction of SatIII RNAs during hyper-osmotic stress.
More recently, it has been shown that the transcription factors Pax3 and
Pax9 act as redundant regulators of mouse heterochromatin, as they
repress major satellite repeats by associating with DNA within
pericentromeric heterochromatin. Interestingly, this study also shows
that all HMT Suv39h–dependent heterochromatic repeat regions in the
mouse genome present a high concordance with the presence of
transcription factor binding sites, and suggests transcription factor
dependent heterochromatin formation as a general mechanism 88.
This new role for transcription factors in pericentromeric transcription
may give some light into how major satellite transcripts are regulated in
different cell contexts, since transcription factors themselves are also
differently expressed depending on cell type.
Today, cancer is a leading cause of death worldwide. The word “cancer”
includes a broad group of diseases, since more than 100 different types
of cancers have been described. All of them are characterized by the
following alterations in cell physiology: self-sufficient growth,
insensitivity to anti-growth signals, evasion of apoptosis and immune
response, and ability for sustained angiogenesis, tissue invasion and
metastasis, among others89. Moreover, a highly heterogeneity among the
cancer cells within a tumor is also a common hallmark of all cancer
Carcinoma, originated in epithelial cells, is the most prevalent form of
cancer, and it includes colorectal, breast, prostate and lung cancer
among others91. Although diagnosis and prognosis of these cancers is
improving every day, metastasis is still responsible for as much as 90% of
cancer-associated mortality. So further understanding of this process is
essential to progress in the fight against cancer.
Still nowadays, it is not clear how tumors initiate and progress. Two
current ideas describe the establishment and maintenance of tumor
heterogeneity, which are the clonal evolution model and the cancer stem
cell hypothesis. Both of them propose that tumors originate from a single
cell that has acquired multiple mutations and has gained unlimited
proliferative potential, but through different mechanisms92.
The clonal evolution model states that, when cancer cells acquire genetic
alterations, each of them confers one or another form of increased
fitness that triggers clonal expansion. This genetic drift and stepwise
natural selection favors the most aggressive cells and drives progression.
Based in this model, any tumor cell that acquires the capacity of selfrenew has the potential to contribute to tumor progression.
On the other hand, the cancer stem cell hypothesis states that a
particular subset of tumor cells with stem cell-like properties, called
“cancer stem cells” (CSC), are the ones that accumulate mutations and
drive tumor initiation, progression, and recurrence. In this way, their self29
4. Cancer and tumor progression
renewal and differentiation capability is responsible for tumor
Both paradigms of tumor propagation are likely to exist in human cancers
since they are not mutually exclusive, and CSC themselves may undergo
clonal expansion92.
As previously mentioned, carcinomas arise in epithelial tissues. Usually,
normal epithelial cells are tightly bound to neighboring cells and to
underlying basement membranes by a complex junctional network.
However, as the tumor progresses, carcinoma cells lose these
associations, acquire mesenchymal traits and increased migration and
invasion capability, being able to dissolve extracellular matrix and move
through the surrounding tissue towards blood and lymphatic vessels,
establishing the first step for metastasis progression93.
The process by which epithelial cells undergo all this morphological and
molecular changes is referred as epithelial-to-mesenchymal transition
(EMT). Its importance in cancer is now widely accepted by pathologists
and the cancer research field considers this process as one of the most
important to understand.94
4.1. Epithelial-to-Mesenchymal Transition (EMT)
The EMT program describes a series of events during which epithelial
cells lose many of their epithelial characteristics and take on properties
that are typical of mesenchymal cells. Accordingly, cells engaged in the
EMT program undergo complex changes in cell architecture and
behavior95. The reverse process, known as mesenchymal to epithelial
transition (MET) has also been reported96.
In a typical epithelial layer, epithelial cells develop adhesive structures
between adjacent cells, such as adherens junctions, desmosomes and
tight junctions, to establish robust intercellular adhesions and facilitate
intercellular communication, thus restricting motility, preserving tissue
integrity, and permitting individual cells to function as a cohesive unit91.
E-cadherin represents the best characterized molecular marker
expressed in epithelial cells. It is a transmembrane protein localized to
the adherens junctions, with an extracellular domain that mediates
calcium-dependent homotypic interactions with E-cadherin molecules on
adjacent cells, and an intracellular domain that binds cytosolic catenins
and provides a link to the actin cytoskeleton. Loss of E-cadherin is by
itself a hallmark of EMT since it is associated with loss of epithelial
phenotype. However, many other epithelial markers are down-regulated
during EMT, such as claudins and occludins, located in tight junctions and
acting as barriers that inhibit lateral diffusion of lipids and proteins
between the apical and basolateral plasma membrane domains97,98.
Mesenchymal cells, on the other hand, do not have stable intercellular
junctions and possess an elongated morphology with front-to-back
asymmetry that facilitates motility and locomotion. Upon EMT, many
mesenchymal markets become expressed, such as integrin family
receptors that localize in the filapodial extensions at the leading edge of
the mesenchymal cells and interact with the extracellular matrix99, or
matrix metalloproteinases (MMP) that digest basement membranes and
promote invasion100.
Mesenchymal cells also have increased expression of mesenchymal
proteins, such as the intermediate filament protein vimentin, or other
cytoskeletal proteins, including smooth muscle actin, as well as
extracellular matrix components such as fibronectin and collagen
precursors101(Figure I14).
These changes in protein expression, and many others, are associated
with changes in transcription101,102. Therefore, activation and repression
of specific sets of genes during this process must be tightly regulated to
establish the particular transcriptome of mesenchymal cells, which
clearly differs from the one typical of epithelial cells.
Since there is a clear link between chromatin and transcription, it is
tempting to speculate that changes in chromatin dynamics and
Epithelial cells are apico-basal polarized, with the apical and basal
surfaces serving different functions.
organization may be occurring during EMT. Activation and repression of
transcription has been associated to gene movement, but is still
unknown whether clusters of epithelial and mesenchymal genes move
and re-localize to specific sub-nuclear domains during EMT. What seems
obvious, although not yet demonstrated, is that nuclear architecture and
organization may notably differ in mesenchymal cells compared to
epithelial cells, since physiological requirements of each type of cell is
completely different.
Figure I14. Epithelial and mesenchymal cell features. Epithelial cells are
characterized by their apico-basal polarity and their cell-cell and cell-ECM
contacts (adherens junctions, tight junctions and desmosomes). Upon EMT,
epithelial cells acquire mesenchymal properties as loss of cell contacts,
expression of mesenchymal markers such as vimentin and fibronectin, and
increased cell motility.
4.1.1. Physiological EMT
EMT has long been recognized by developmental biologists as a crucial
process for the generation of tissues and organs during embryogenesis of
The earliest example of EMT during embryonic development is the
generation of the mesoderm, which marks the beginning of gastrulation.
After invagination of the epithelial cells from the epiblast (primitive
ectoderm) around the primitive streak, the basement membrane
breaches locally and cells lose their tight cell-cell adhesions and remain
attached to neighboring cells only by disperse focal contacts. After
completing EMT they migrate along the narrow extracellular space
underneath the ectoderm to form the mesoderm104(Figure I15).
Figure I15. Examples of primary EMT. The first EMT after implantation is that
undergone by the mesendodermal progenitors during gastrulation, whereas the
delamination of neural crest cells from the dorsal neural tube is a later event .
After gastrulation, the epidermal and neural territories are progressively
defined along the rostro-caudal axis and the neural crest forms at the
boundary between these two territories. Neural crest cells within the
dorsal neural epithelium also undergo EMT, and individual cells migrate
before giving rise to different derivatives, including craniofacial
both vertebrates and invertebrates. Indeed, several rounds of EMT and
MET are necessary for the final differentiation of specialized cell types
and the acquisition of the complex three-dimensional structure of
internal organs. Accordingly, these sequential rounds are referred to as
primary, secondary, and tertiary EMT. Examples of primary EMT include
mesoderm formation during gastrulation and neural crest delamination,
whereas liver and pancreas result from secondary and heart from tertiary
structures, most of the peripheral nervous system, some endocrine cells,
and melanocytes103(Figure I15).
Although EMT has been largely described during development, similar
processes occur in adult tissues. The classical example is skin wound
healing, during which keratinocytes at the border of the wound migrate
to close it. To do so, they acquire a “metastable” state, characterized by
rearrangement of their actin cytoskeleton, extended lamellipodia and
loss of both cell–cell contacts and hemidesmosomes. In addition, they
alter the expression of integrin receptors and express various proteases
to degrade connective tissue. But they do no undergo a complete EMT
since they retain some intercellular junctions and express epidermal
keratins but not vimentin105.
4.1.2. Pathological EMT
EMT has an important role in the development of many tissues during
embryogenesis, but similar cell changes are recapitulated during
pathological processes, such as fibrosis and cancer.
During progression to metastatic competence, carcinoma cells acquire
mesenchymal gene-expression patterns and properties. This results in
changed adhesive properties and activation of proteolysis and motility,
which allows the tumor cells to metastasize and establish secondary
tumors at distant sites95.
Activation of an EMT program during tumorigenesis often requires
signaling between cancer cells and neighboring stromal cells104. Islands of
cancer cells in advanced primary carcinomas are thought to recruit a
variety of cell types into the surrounding stroma, such as fibroblasts,
myofibroblasts, granulocytes, macrophages, mesenchymal stem cells,
and lymphocytes. These recruited cells create an inflammatory
microenvironment or “reactive” stroma, which appears to result in the
release of EMT-inducing signals. Carcinoma cells respond to these signals
by activating expression of certain transcription factors that proceed to
orchestrate EMT programs within these cells. In this way, EMT may only
Furthermore, EMT can induce non-CSC to enter into a CSC-like state. As
such, the EMT confers on epithelial cells precisely the set of traits
(invasion, apoptosis resistance…) that would empower them to
disseminate from primary tumors and seed metastases. This implies that
not only intrinsic CSC but also induced subtypes of CSC within a tumor
would be able to acquire a metastatic phenotype (Figure I16)93.
Figure I16. Acquisition of the metastatic phenotype. Intrinsic CSCs are thought
to exist in primary tumors from the very early stages of tumorigenesis and may
be the oncogenic derivatives of normal-tissue stem or progenitor cells. Induced
CSCs may arise as a consequence of the EMT. In this case, carcinoma cells
initially recruit a variety of stromal cells that create a reactive microenvironment
that releases factors that cause the neighboring cancer cells to undergo the EMT
and acquire CSC-like characteristics .
Interestingly, although EMT is often depicted as a bi-stable switch that
causes cells to flip from one state into the other, the biological reality is
likely to be more subtle. In many tumors, epithelial carcinoma cells
appear to advance only partially toward the mesenchymal state,
acquiring a “metastable phenotype”, characterized by the co-expression
of epithelial and mesenchymal markers106. Furthermore, this metastable
phenotype has been associated to enhanced metastatic potential96. Thus,
operate in a small fraction of cancer cells that are in intimate contact
with the adjacent reactive stroma.
maintenance of some epithelial properties may somehow favor
metastasis. In addition, observation that metastases appear histologically
similar to the primary tumor from which they are derived suggests that
migrating cells may undergo MET once they reach a secondary site to
establish a micrometastasis. Thus, acquisition of mesenchymal
characteristics may be transitory and favor invasion and intravasation of
cancer cells, but may undergo a reversal during later tumorigenesis to
allow establishment and development of metastasis at distal organs
(Figure I17).
Figure I17. Potential roles of EMT and MET in carcinoma progression. In
primary tumor invasion (1), mesenchymal cells (CSC or non-CSC that have
undergone EMT (2)) may be better equipped to escape the tumor mass. These
cells may migrate and avoid surgical removal and radiotherapy (3) to form a
local recurrence through MET (4). Mesenchymal forms may be best equipped to
intravasate (5), and a mixture of tumor cell phenotypes is present in the
circulation, including epithelial (brown), hybrid and possibly mesenchymal cells
(6). Extravasation may be enhanced by mesenchymal attributes (7), after which
the tumor cells may persist as micrometastases (8) till re-epithelization via MET
establishes macrometastasis (9), which can be favored by autocrine (10) and
paracrine (11) pathways .
Many signaling pathways trigger EMT in both embryonic development
and in normal and transformed cell lines. These signaling pathways
include those triggered by soluble factors such as different members of
the TGFβ superfamily, Wnts, Notch, other tyrosine kinase receptors
family members (EGF, HGF and FGF), HIF, and many others103,104. Also
components of the extracellular matrix (ECM), such as collagen and
hyaluronic acid108 are able to induce EMT95(Figure I18).
Figure I18. Overview of the molecular networks that regulate EMT. Selection of
signaling pathways activated by regulators of EMT and limited representation of
their crosstalk. Activation of receptor tyrosine kinases (RTKs) induces EMT but it
often requires co-activation of integrin receptors. TGFβ triggers EMT, but there
is also a mutual regulation of the TGFβ and NOTCH pathways during EMT. Other
signaling pathways could have an important role in EMT, including G-proteincoupled receptors and matrix metalloproteinases (MMPs) that can also trigger
EMT through as-yet-undefined receptors .
The vast majority of the signaling pathways known to trigger EMT
converge at the induction of the E-cadherin repressors. In particular, in
4.1.3. Signaling EMT
response to TGFβ, Smad2 and Smad3 are activated and form complexes
with Smad4, which then regulate transcription of target genes. The
expression of three families of transcription factors (Snail, ZEB and bHLH
families) is induced in response to TGFβ, either through a Smaddependent mechanism (in the case of Snail proteins) or indirectly
through activation of other transcription factors or relief of repression.
Upon activation, these transcription factors repress epithelial markers
gene expression and activate mesenchymal gene expression109.
A central target of these transcriptional regulators is the repression of
the E-cadherin gene (CDH1), an important caretaker of the epithelial
phenotype. Its down-regulation abolishes E-cadherin-mediated
sequestering of β-catenin in the cytoplasm and as a result, β-catenin
localizes to the nucleus and feeds into the Wnt signalling pathway by
activating transcription of LEF-1/TCF4 target genes95, which in turn favors
4.1.4. EMT inducers
E-cadherin repressors can be classified into two groups depending on
their effects on the CDH1 promoter: Snail/Slug, Zeb1/Zeb2, E47, and
KLF8 bind to and repress the activity of the CDH1 promoter111–117,
whereas factors such as Twist, Goosecoid, E2.2, and FoxC2 repress CDH1
transcription indirectly118–121. Snail1 transcription factor
Snail1 is one of most widely studied effectors of EMT and CDH1
repression. Besides CDH1, Snail1 represses several other epithelial genes
such as occluding (OCLN) and claudin (CLDN), but it is also involved in the
regulation of genes related to cell cycle progression or apoptosis122.
Interestingly, Snail1 is able to repress itself, limiting its own expression123.
Moreover, Snail1 also induces the expression of mesenchymal genes and,
Snail1 belongs to the Snail superfamily of transcription factors, which is
subdivided into the Snail and Scratch families. Three members of the
Snail family have been described in vertebrates to date: Snail (SNAI1),
Slug (SNAI2) and Smuc (SNAI3). Members of the Snail family are zincfinger transcription factors that share a common organization: a highly
conserved C-terminal region that contains from four to six zinc C2H2 type
fingers and a divergent N-terminal region125.
In particular, Snail1 has 4 zing fingers in its C-terminal domain, which
function as DNA binding domains and bind to specific sequences called Eboxes: 5’-CACCTG-3’ or 5’-CAGGTG-3’126,127 located in the promoters of
its target genes (Figure I19).
Figure I19. Structural domains of Snail1 transcription factor. Snail1 has an Nterminal SNAG domain involved in co-repressor interaction. The central region
comprises a serine-proline domain, a destruction box and a nuclear export signal
domain (NES) and is important for protein localization and stability. The Cterminal domain has 4 zing fingers responsible for direct binding to E-boxes in
the DNA. Adapted .
The central region of Snail1 comprises a nuclear export signal (NES), a
destruction box domain, and a serine-proline rich region128. This central
region can be widely posttranslational modified and is involved in the
regulation of protein stability and localization. In this way,
phosphorylation of Snail1 by GSK3β in the nucleus makes the NES
accessible to the CRM1 transporter and facilitated translocation of the
protein to the nucleus129, where is further phosphorylated by GSK3β and
at least in certain conditions, it may work as a direct activator together
with p65-NFκB124.
targeted to proteasome degradation by β-TrCP1130. By contrast,
phosphorylation of Snail1 by Pak1 retains it in the nucleus enhancing its
repressive function131. Fbxl14, another ubiquitin ligase is also able to
ubiquitylate Snail1 in this central region and target it to degradation132.
Finally, Snail1 has a SNAG (Snail/Gfi-1) domain in its N-terminal domain
that is important for co-repressor interaction and repressive capacity126.
Thus, Snail1 binds the promoter of its target genes through its zing finger
domain, and interacts with histone modifying enzymes through the SNAG
domain. In this way, Snail1 recruits the co-repressor Sin3A together with
HDAC1/2 and induces deacetylation of histones H3 and H4133. It also
recruits Polycomb repressive complex 2 (PRC2) which trimethylates
H3K27134 and LSD1, which is found as a component of the HDAC1/ 2containing co-repressor complex CoREST and removes mono- and
dimethyl marks on H3K4135(Figure I20).
Figure I20. Model of Snail1 mediated transcriptional repression. Snail1 binds
LSD1 through its SNAG domain. The ternary Snail1–LSD1–CoREST complex is
recruited to E-boxes of Snail target gene promoters, in which LSD1 binds the tail
of histone H3 and demethylates lysine 4 of histone H3 (H3K4). Histone
deacetylases 1 and 2 (HDAC1/2) deacetylate histones 3 and 4 (H3/H4), and the
polycomb repressor complex 2 (PRC2) trimethylates lysine 27 of histone 3
(H3K27). Together, Snail1-targeted chromatin modifications lead to repression
of Snail1 target genes .
40 Lysyl oxidase-like 2 (LOXL2)
LOXL2 belongs to the LOX family of proteins, which encodes genes for
copper-dependent amine oxidases. These enzymes catalyze the covalent
cross-link of the component side chains of collagen and elastin, and in
this way stabilize the extracellular matrix (ECM). There are five members
in the LOX family: LOX, LOXL1, LOXL2, LOXL3, and LOXL4. They are
differently expressed during development and in adult tissues, what
suggests each one of them may have individual and specific functions136.
The C-terminal end of LOX proteins encodes the enzyme domain and is
highly conserved between the family members. The N-terminal end
encodes for variable pro-peptide regions with variable sequences, which
may determine individual roles (Figure I21)137. LOXL2 in particular has
four scavenger-receptor cysteine-rich domains (SRCR), which are
supposed to interact with other proteins and in this way modulate its
amine oxidase activity138.
LOX proteins have classically been described as extracellular proteins
involved in ECM remodeling, but several reports show that this proteins
are also expressed inside the cell, where they may have many other
functions139. Accordingly, besides their function in ECM stabilization,
some other functions for LOX proteins are starting to be described, such
as their involvement in cell proliferation140,141 and bone formation142.
Interestingly, LOX and LOXL2 are arising as key players in cancer
progression since they promote tumor cell invasion and metastasis143.
Their expression is induced by hypoxia and TGFβ, and they contribute to
CDH1 repression during EMT144,145. Indeed, LOXL2 is able to induce EMT
by itself. Moreover, it interacts and stabilizes Snail1, and in this way
collaborates with Snail1 mediated CDH1 repression146.
Recently, we described that a new histone modifying enzyme, LOXL2, is
also recruited by Snail to the CDH1 promoter and induces repression by
the deamination of H3K4me330.
Figure I21. Structural organization of lysyl oxidase family members. C-terminal
end of LOX family of proteins encodes a conserved cooper containing catalytic
domain. N-terminal end is more variable among different members but it
contains a signal peptide in all cases (Si). LOX has a PP domain, LOXL1 a prolinerich domain, and LOXL2 four scavenger-receptor cysteine-rich domains (SRCR).
LOXL3 and LOXL4 have similar composition seen in LOXL2 .
LOXL2 also promotes invasion by regulating the expression and activity of
the tissue inhibitor of metalloproteinase-1 (TIMP1) and matrix
metalloproteinase-9 (MMP9)147.
The mechanism by which LOXL2 regulates transcription was until recently
poorly understood. However, we have shown that LOXL2 is recruited to
CDH1 promoter and deaminates H3K4me3, a newly characterized LOXL2
substrate. In this way, it removes an active histone mark, establishing a
new repressive histone modification that explains how LOXL2 represses
CHD1 together with Snail130.
Moreover, deamination of H3K4me3 generates a highly reactive
aldehyde group in the resulting oxidized histone H3. So far, it is still
unknown how or whether this lysine deamination can be reversed, since
none histone aminotransferase enzymes have been described to date.
This suggests that lysine deamination may be a highly stable repressive
mark, since histone replacement or histone tail exchange seem to be the
best candidates to erase this mark. Thus, it may be enriched in
heterochromatin domains. However, no genome-wide analyses have
Interestingly, this opens the question whether other LOX family members
may also function as histone modifying enzymes and gives a new
perspective on the role that this family of proteins have in gene
expression regulation and cancer progression.
been done to characterize the distribution of this modification to date.
Therefore, it is still unknown whether its distribution correlates with
highly condensed and repressed chromatin domains.
Thus, the objective of this PhD thesis is to investigate the potential roles
of Snail1 and the H3K4me3-remodeling enzyme LOXL2 in the regulation
of heterochromatin reorganization during EMT.
Chromatin reorganization is highly likely to occur during EMT since
important changes in the cellular characteristics and gene expression
occur during this process. However, only few pieces of evidence support
this idea148. Snail1 transcription factor plays an essential role in EMT
through its interaction with LOXL2146 and heterochromatin is
characterized by the absence of H3K4me343.
1. Snail1 is essential for pericentromeric heterochromatin
maintenance and organization in mesenchymal cells.
In order to check whether Snail1 could have a role in pericentromeric
heterochromatin, we first studied the effect of Snail1 depletion in
chromocenter organization in mesenchymal cells. Since Snail1 KO mice
are embryonically lethal150 we used Snail1 conditional KO mice151 to
isolate primary mouse embryonic fibroblasts (pMEFS) from day E12.5
(Snai1F/F/Cre- and Snai1F/F/Cre+). Snail1 was efficiently ablated by Cremediated recombination after tamoxifen treatment, as shown by
Western blot (Figure R1-A) giving rise to Snai1F/F/Cre- (CT) and Snai1F/F/Cre+
(KO) pMEFs.
Heterochromatin was initially analyzed in CT and KO pMEFs by DAPI
staining (Figure R1-B) and the number of foci per cell was quantified. This
analysis revealed an important decrease in the average number of
heterochromatin foci per nucleus upon Snail1 depletion (Figure R1-C), as
well as clear changes in the distribution of heterochromatin foci number
per nucleus in CT and KO pMEFs (Figure R1-D).
Pericentromeric heterochromatin has a particular epigenetic signature
characterized by specific histone modifications such as H3K9me3 and
enrichment in HP1α43. We decided to further study heterochromatin
organization by H3K9me3 and HP1α immunostaining that also allows us
to visualize chromocenters. Staining patterns for H3K9me3 and HP1α in
CT and KO pMEFs were similar to those observed for DAPI stained nuclei
(Figure R1-B). The average number of foci per nucleus was significantly
lower in Snail1 depleted cells not only for H3K9me3 but also for HP1α
staining (Figure R1-C). These observations suggest that pericentromeric
heterochromatin loses its normal organization in pMEFs in the absence
of Snail1.
In mouse cells, pericentromeric heterochromatin clusters into
chromocenters, which can be easily visualized as foci by DAPI staining.
Different subgroups of chromosomes arrange to give rise to several
chromocenters, the number and organization of which can differ in
distinct cell types and developmental stages149.
Figure R1. Heterochromatin organization is compromised in the absence of
and Snail1
Snail1. (A) Western blot for Snail1 in pMEFs Snail1
F/F/Crewith tamoxifen. (B) A representative image of pMEFs Snail1
(CT) and
(KO) stained with DAPI and immunolabeled with anti-HP1α and antiH3K9me3 after tamoxifen treatment. (C) Average of foci number per nucleus in
CT and KO pMEFs detected with DAPI, H3K9me3 and HP1α. (D) Distribution of
heterochromatin foci number per nucleus in CT and KO pMEFs stained with DAPI
and anti-HP1α respectively. Error bars indicate standard deviation (SD) for at
least three independent experiments; in all of them at least 100 nuclei were
analyzed. Scale bar, 10 µm. The two asterisks indicate a p<0.01.
Replication timing has been widely characterized in mouse cells and it is
a well-defined spatio-temporal process with euchromatin replication in
early S-phase and heterochromatin replication during mid-late Sphase154. In particular, major satellite sequences in pericentromeric
regions first replicate during mid S-phase followed by minor satellite
sequences in centromeric regions, which replicate during late S-phase.
We used PCNA, which is a replication marker, to visualize replicating cells
in immortalized MEFs (iMEFs) CT and KO for Snail1 (Figure R2-A). Diffuse
staining of PCNA in the nucleus correlates with early S-phase whereas
spotted staining is associated to mid-late S-phase. The percentage of
early and late replicating cells was quantified for both CT and KO iMEFs.
As shown in Figure R2-B, there is a decrease in mid-late replicating cells
in KO iMEFs compared to CT, suggesting these cells have a delay in the
progression beyond early S-phase in the absence of Snail1.
Accordingly, quantification of the total number of replicating cells in CT
and KO iMEFs shows an accumulation of replicating cells in KO iMEFs
compared to CT cells (Figure R3-A), which further confirms that, in the
absence of Snail1, replication is compromised.
To further characterize these replication defects, we used real-time
quantitative PCR (qRT-PCR) to measure the relative amounts of major
and minor satellite using DNA extracted from CT and KO iMEFs. As an
early-replicating control DNA sequence, we used the α-globin gene155.
An identical number of PCR cycles was required for amplification of the
α-globin gene from DNA extracted from CT and KO iMEFS (Figure R3-B),
consistent with the absence of a defect in replication fork progression
during early S-phase.
A clear link exists between heterochromatin organization and replication.
Such a complex structure as is pericentromeric heterochromatin has to
be maintained and accurately reproduced throughout multiple cell
divisions. Defects in heterochromatin replication may be either the cause
or the consequence of heterochromatin disorganization152,153. Therefore,
we decided to check whether Snail1 depletion affected heterochromatin
Figure R2. Snail1 is required for S-phase progression. (A) A representative
image of iMEFs CT and KO for Snail1 stained with DAPI and PCNA (replication
marker) to show typical early replication staining (PCNA diffused in the nucleus)
and mid-late replication staining (dotted pattern of PCNA in the nucleus). (B) The
graphs show the percentage of early and mid-late replicating cells in iMEFs CT
and iMEFs KO. Three independent experiments were quantified; in all of them at
least 100 nuclei were analyzed.
In contrast, amplification of the major and minor satellite sequences
required additional cycling when using DNA from cells lacking Snail1
(Figure R3-B). In particular, we observed a 30% reduction in major and
minor satellite DNA sequences in KO iMEFs compared to CT iMEFs (Figure
R3-B). These data suggests that depletion of Snail1 does not affect
replication of an early-replicating gene (α-globin) but affects replication
of both pericentromeric and centromeric satellite DNA, which is
consistent with the decrease of mid-late replicating cells observed in KO
cells (Figure R2-B).
Defects in pericentromeric heterochromatin replication have been
associated to genome instability since proper heterochromatin structure
and organization seem to be crucial for correct chromosome segregation
during mitosis67. Thus, we did karyotype analysis for CT and KO pMEFs at
passage 4. Although the number of chromosomes per cell was not
significantly changing in the absence of Snail1 (Figure R4-A), KO pMEFs
showed an increase in chromosomal anomalies compared to CT such as
ring and acentric chromosomes, among others. Most of these anomalies
were associated to heterochromatin defects that lead to chromosomal
instability (Figure R4-B,C).
All this data shows that Snail1 is necessary for proper chromocenter
organization in mesenchymal cells, and that it also has a role in
pericentromeric heterochromatin replication, which in turn is essential
for heterochromatin organization and maintenance.
Figure R3. Snail1 is required for pericentromeric heterochromatin replication.
(A) The graph shows the percentage of cells in S-phase in CT and KO iMEFs after
quantification of immunostaining for PCNA (shown in Figure R2-A). (B) qRT-PCR
shows the relative amount of α-globin (early replicating gene) and major and
minor satellite DNA (mid-late replicating sequences) in KO iMEFs compared to
CT iMEFs, which was set as 1. Error bars indicate SD of at least three different
experiments; one asterisk indicates a p<0.05.
Figure R4. Snail1 depletion induces genome instability. (A) Distribution of
number of chromosomes per cell observed in CT and KO pMEFs methaphases.
(B) A representative image of CT and KO pMEFs methaphase spreads; some
alterations (ring chromosomes) can be observed in the KO condition. (C)
Quantification of the total number of chromosomal alterations observed in
methapase spreads of CT and KO pMEFs, as well as quantification of those
associated with heterochromatin defects, such as ring chromosomes and
acentric chromosomes. A total of 30 methaphases were analyzed both for CT
and KO pMEFs.
2. Snail1 regulates pericentromeric transcription
To determine if the observed heterochromatin structural deficiencies
and replication defects were associated to pericentromeric transcription,
we analyzed major satellite transcription by qRT-PCR in iMEFs CT and KO
for Snail1. We observed an increase in major satellite transcription of up
to 50-fold in KO iMEFs compared with CT iMEFs (Figure R5-A). Other
interspersed repetitive elements such as minor satellites, L1 LINE, SINE1
and IAP1 presented lower changes (Figure R5-A). Moreover, although the
levels of major satellite transcripts were elevated in the absence of
Snail1, they were partially restored when Snail1 was reintroduced into
KO iMEFs (Figure R5-A).
Similar results were obtained in pMEFs, where the absence of Snail1
caused up-regulation of not only major satellite transcripts but also
minor satellite transcripts. However, transcription of other interspersed
repetitive elements was not affected (Figure R5-B).
This, together with the previous results, shows an inverse correlation
between Snail1 and pericentromeric and centromeric transcripts,
confirming that Snail1 is able to repress heterochromatin transcription.
Snail1 is a transcription factor that has been mainly associated to
repression. Precise regulation of major satellite repeats transcription has
been associated with proper heterochromatin organization and
replication43. Therefore, we asked whether Snail1 transcription factor
could regulate heterochromatin organization by repression of major
satellite transcripts.
Figure R5. Snail1 regulates pericentromeric heterochromatin transcription. (A)
Left panel: qRT-PCR shows the changes in expression of major satellite and other
repetitive sequences in CT and KO iMEFs. The levels of major satellite transcripts
were also analyzed by qRT-PCR after Snail1 transfection in Snail1 KO iMEFs
(KO+Sna). Right panel: qRT-PCR showing the re-expression of Snail1 mRNA in KO
iMEFs. Gene expression was normalized against an endogenous control (HPRT or
Pumilio) and presented as RNA levels over those obtained in CT iMEFs in the left
panel, and KO iMEFs in the right panel, which were set as 1. (B) qRT-PCR shows
the changes in expression of major satellite and other sequences in CT and KO
pMEFs. Normalization was done same way as in (A). Error bars indicate SD of at
least three different experiments; two asterisks indicate a p<0.01.
3. Snail1 is enriched in pericentromeric regions and interacts
with HP1α.
In human cells, pericentromeric regions are enriched in satellites II and
III, which have a less conserved sequence than major satellite repeats
and varies not only between the satellites in the same chromosome, but
also among the satellites in different chromosomes69.
Due to the complexity of human pericentromeric sequences and the
impossibility to predict in silico the Snail1 binding sites in human
heterochromatin, we analyzed whether Snail1 also binds to
pericentromeric regions in human cells by ChIP-Seq analysis. We used
SW620 cell line, derived from human colorectal adenocarcinoma, which
express high levels of endogenous Snail1. We covered an average depth
of 1 million reads that mapped to unique sites in the genome. From the
identified binding sites, more than 50% mapped to intergenic regions.
Exhaustive analysis of these intergenic sequences revealed putative
binding sites of Snail1 to pericentromeric and centromeric regions of 14
human chromosomes (Figure R6-B).
We confirmed that Snail1 was bound to the pericentromeric regions of
three different human chromosomes (chr4, chr7 and chr20) by ChIP
assay (Figure R6-C). We further confirmed by ChIP that these binding
sites were indeed pericentromeric regions by checking HP1α enrichment
(Figure R6-C).
Snail1 binds to the promoter of its target genes and mediates repression
by the recruitment of repressive histone modifying enzymes122. Thus, we
checked by chromatin immunoprecipitation (ChIP) whether Snail1 binds
to mouse pericentromeric regions. Indeed, endogenous Snail1 was
enriched in major satellite regions in CT iMEFs. This enrichment was not
observed when an irrelevant sequence (α-globin gene) was used as a
negative control (Figure R6-A).
Figure R6. Snail1 binds pericentromeric regions in mouse and human cells. (A)
Snail1 binding to major satellite regions was determined by ChIP in CT and KO
iMEFs. The α-globin gene was used as a negative control in the ChIP assay. Data
from qRT-PCR amplifications of the major satellite regions and α-globin gene in
CT iMEFs was normalized to the input and expressed as fold enrichment over the
data obtained in KO MEFs, which was set as 1. (B) Whole human chromosomes
view of Snail1 ChIP-Seq predicted binding sites (rectangles) in pericentromeric
regions. (C) Snail1 binding to pericentromeric regions of human chromosomes 4,
7 and 20 was determined by ChIP in SW620 cells. HP1α was used as positive
control in the same regions. qRT-PCR data of pericentromeric regions was
normalized to the input and expressed as fold-change over the data obtained
when an irrelevant IgG was used (dark bars), that was set as 1. Error bars
indicate SD in at least three independent experiments. One asterisk indicates a
p<0.05, while two indicate p<0.01.
Moreover, Snail1 interacted by co-immunoprecipitation (CoIP) with HP1α
in mouse (NIH-3T3) and in human (MiaPaCa-2 and SW620) cell lines
(Figure R7-A), which suggests that, besides being both located in
In addition, we transfected several Snail1 and HP1α mutants in HEK293T
cells to further characterize this interaction. The mutant Snail1-SD has all
serine residues in the serine rich domain replaced with aspartates, and
Snail1-SA with alanines. Snail-SA remains in the nucleus, whereas Snail1SD is located in the cytoplasm because its mutation mimics
phosphorylation and induces nuclear export129. On the other hand,
Snail1-P2A has a point mutation in the SNAG domain and it is neither
able to repress CDH1 transcription111 nor able to recruit histone
modifying enzymes to its promoter134,156. The upper panel of Figure R7-B
shows that wild type (WT) and Snail-SA interacted with HP1α whereas
Snail-SD and Snail-P2A did not. This suggests that the interaction with
HP1α takes place though Snail1 SNAG domain and that it only occurs in
the nucleus, where HP1α is located.
To determine the domain of HP1α involved in the interaction with Snail1,
three different mutants were used: HP1α-V21M has a mutation in the
chromodomain so it does not interact with K9me3; HP1α-I165K has a
mutation in the chromoshadow domain that disrupts dimerization;
HP1α-W174A has a mutation in the chromoshadow domain, not critical
for dimerization but involved in the binding to other proteins157. As
shown in the lower panel of Figure R7-B, only WT and HP1α-V21M were
able to interact with Snail1, indicating that chromoshadow domain is
required for the interaction.
This demonstrates that the interaction between these two proteins is
taking place through the Snail1 SNAG domain and the HP1α
chromoshadow domain (Figure R7-B). However, this interaction may be
indirect since the consensus PXVXL sequence involved in HP1α
interaction is not found in Snail1 SNAG domain. Other proteins described
to interact with HP1α through its chromoshadow domains are recruited
to and involved in heterochromatin formation or regulation46. Thus, the
fact that Snail1 also interacts with HP1α through this domains further
suggests a role for Snail1 in heterochromatin regulation.
pericentromeric domains, there is a function link between these two
Figure R7. Snail1 SNAG domain interacts with HP1α chromoshadow domain.
(A) Endogenous HP1α co-immunoprecipitated with endogenous Snail1 both in
mouse cells (NIH-3T3) and human cells (MiaPaCa-2 and SW620). (B) Extracts
from HEK293T cells, transiently transfected with HA-tagged Snail1 WT and
mutants (upper panel) or Flag-tagged HP1α WT and mutants (lower panel), were
immunoprecipitated with anti-flag and the immunocomplexes were analyzed by
Western blotting using the indicated antibodies. Extracts from cells transfected
only with Snail1 but without HP1α were used as negative control (first lane).
4. Pericentromeric transcription is linked to H3 oxidation
In addition, LOXL2 induced global chromatin compaction when
overexpressed in cultured cells, as shown by MNase digestion (Figure
R8). This assay consists in digesting chromatin with Micrococcal, an
enzyme that makes double-stranded cuts between nucleosome particles.
The compaction status of chromatin determines the enzyme accessibility
to the DNA and its digestion efficiency. Thus, decreased sensitivity to
MNase, as observed in LOXL2 overexpressing cells, has been associated
to increased chromatin compaction.
Figure R8. LOXL2 favors nucleosomal compaction. Nuclei isolated from
HEK293T transfected with LOXL2 wild type (wt), inactive mutant (mut) or an
empty vector (Ø) were digested with micrococcal nuclease (MNase) for 5 or 10
min. Total genomic DNA was analyzed using agarose gel electrophoresis.
Positions of the nucleosomes are indicated (mono-, di-, tri-, tetra-).
Enrichment of repressive histone marks and lack of H3K4 methylation is a
common feature of pericentromeric domains43. As previously shown,
Snail1 represses major satellite transcription. Moreover, Snail1 interacts
with LOXL2146 and together, they cooperate in the repression of target
genes by oxidation of histone H3 in lysine430.
Therefore, we investigated whether LOXL2 is also involved in the
regulation of pericentromeric transcription. First of all, we checked the
levels of oxidized H3 in major satellite regions in CT and Snail1 KO iMEFs
(Figure R9). We found that the previously observed up-regulation of
pericentromeric transcription in Snail1 KO MEFs correlated with a
decrease in the enrichment of oxidized H3 in major satellites.
Figure R9. There is less oxidized H3 in major satellite regions of Snail1 KO
iMEFs. H3 oxidation in major satellite domains was determined by Re-ChIP in CT
and KO iMEFs. Lysates were incubated with biotinhydrazide (an activated biotin
that reacts with oxidized H3) before Re-ChIP. Extracts were sequentially
immunoprecipitated with anti-H3 and streptavidin-beads. DNA binding was
quantified by qRT-PCR. Data were normalized to the total amount of H3
immunoprecipitated and to the input, and expressed as fold enrichment over
the data obtained when an irrelevant IgG was used, that was set as 1. Error bars
indicate SD in at least three independent experiments; two asterisks indicate
To assess the contribution of LOXL2 in pericentromeric silencing, we
infected Snail1 KO iMEFs with a retrovirus vector expressing LOXL2
(Figure R10-A, right panel). As expected, we observed that
overexpression of LOXL2 led to a down-regulation of major satellite
transcription, similar to that observed when overexpressing Snail1
(Figure R10-A, left panel).
On the other hand, we also knocked down LOXL2 in CT iMEFs to see the
opposite effect. Accordingly, LOXL2 knock-down (shLOXL2) in CT iMEFs
induced an increase in major satellite transcription analyzed by RT-qPCR
(Figure R10-B). However, RNA levels of other repetitive sequences, such
as minor satellites, L1 LINE, SINE1, and IAP1, were not affected by the
absence of LOXL2 (Figure R10-B).
Figure R10. LOXL2 regulates major satellite transcription. (A) Left panel: qRTPCR shows the changes in major satellite transcription when LOXL2 is
overexpressed in Snail1 KO iMEFs, after retrovirus infection. Right panel: qRTPCR shows the overexpression of LOXL2 mRNA in Snail1 KO iMEFs after
infection. Expression levels were normalized to an endogenous control and are
shown relative to the KO iMEFs values, which were set as 1. (B) qRT-PCR shows
the changes in expression of major satellite and other repetitive sequences after
LOXL2 knock-down using shRNAs in CT iMEFs. Expression levels were normalized
to an endogenous control and expressed relative to the shRNA control infected
cells, set as 1. The lower panel shows a representative Western blot for LOXL2 of
shControl or shLOXL2 MEFs.
Moreover, LOXL2 bound to major satellite sequences as shown by ChIP
assay (Figure R11, left panel). In addition, loss of LOXL2 in CT iMEFs after
shRNA infection resulted in a reduction of oxidized H3 levels in
pericentromeric regions (Figure R11, middle panel). Accordingly, there
was an enrichment of H3K4me3 in major satellite regions in shLOXL2
condition compared to shControl (Figure R11, right panel). All these
results confirm that LOXL2 is also involved in the regulation of
pericentromeric transcription, specifically by the repression of major
satellite transcripts through oxidation of H3.
Figure R11. LOXL2 binds major satellite sequences and oxidizes H3. CT iMEFs
were infected with a shRNA control or a shRNA against LOXL2, and LOXL2 ChIP
(left panel), H3K4me3 ChIP (right panel) and oxidized H3 Re-ChIP conducted as in
Figure R9 (middle panel), were performed. qRT-PCR data from the major
satellite region in MEFs treated with either shRNA control or shRNA against
LOXL2 were normalized to the input and expressed as fold enrichment relative
to the data obtained in shRNA control, which was set as 1. Error bars indicate SD
in at least three independent experiments. One asterisk indicates a p< 0.05,
while two indicate a p < 0.01.
5. Pericentromeric transcription is tightly regulated during EMT
Since Snail1 is a transcription factor that plays a key role in EMT, we
investigated whether it also regulates heterochromatin organization
during this process. For this, we used the well-established model of
mouse mammary epithelial NMuMG cells, which undergo EMT after
TGFβ treatment158.
We treated NMUMG cells with TGFβ and visualized chromocenters by
DAPI and anti-HP1α staining at different time points. HP1α foci were
clearly visible in control cells (untreated) and 24 hours after TGFβ
treatment. However, after 2 hours, and even more evident after 8 hours,
the number of HP1α foci per cell clearly decreased (Figure R12).
Figure R12. HP1α delocalizes from chromocenters after TGFβ treatment. Upper
panel: NMuMG cells were treated with TGFβ and stained with DAPI or
immunolabeled with anti-HP1α antibody at different time points. Scale bar, 10
µm. Inset shows magnified representative nucleus, scale bar 5 µm. Lower panel:
the graph depicts the average of DAPI and HP1α foci number per nucleus at
different time points after TGFβ treatment.
Interestingly, HP1α staining appeared diffused in the nucleus of most of
the cells after 8h of treatment but the total fluorescence intensity
remained the same in all cases. This suggests that, rather than being
degraded or repressed, HP1α is being delocalized from chromocenters
but still remains in the nucleus. Accordingly, expression of CBX5 mRNA
did not change after TFGβ treatment (data not shown), nor HP1α protein
levels did (Figure R13).
Moreover, HP1α was found back at chromocenters 24 hours after
treatment. This fits perfectly with the idea that maintenance of stable
heterochromatin domains in living cells involves the transient binding
and dynamic exchange of HP1α from chromatin159. Interestingly, the
number of HP1α foci per cell 24 hours after TFGβ treatment, when cells
already look mesenchymal-like, increased compared to untreated
epithelial cells (Figure R12) The fact that HP1α foci were not only more
abundant but also smaller 24 hours after treatment suggests that
heterochromatin reorganization is occurring during EMT and that new
clustering and organization of chromocenters may be a hallmark of EMT
resulting mesenchymal cells.
The delocalization of HP1α from chromocenters after TGFβ treatment
suggests that it may be released from chromatin and accumulated in the
nucleoplasm. To confirm this idea, we did salt extraction assays and
subcellular fractionation.
Salt extraction of isolated nuclei released higher levels of endogenous
HP1α 8 hours after TGFβ treatment, compared with other time points
(Figure R13-A). This suggests that a pool of HP1α was loosely bound to
chromatin at that time point. Total HP1α levels were constant in this
fraction, as shown in the input by Western blot (Figure R13-A).
Consistently, subcellular fractionation showed a decrease in HP1α levels
in the chromatin fraction 8 hours after TGFβ treatment (Figure R13-B).
These results suggest that TGFβ leads to the transient loss of HP1α at
pericentromeric heterochromatin during the first steps of the EMT
Based on the fact that Snail1 depletion was affecting chromocenter
organization in mesenchymal cells, and that it is one of the main EMT
inducers, we next analyzed whether heterochromatin reorganization
after TGFβ treatment was Snail1 dependent. NMuMG cells were
transfected with a siRNA control or siRNA against Snail1 in the presence
of TGFβ. As expected, Snail1 protein levels were up-regulated 8 hours
after TGFβ treatment in siRNA control cells but not in those transfected
with Snail1 siRNA, as shown by immunofluorescence (Figure R14-A). The
number of HP1α foci per cell was quantified in these conditions (Figure
R14-B). Strikingly, although in siControl condition the number of HP1α
foci decreased 8 hours post TFGβ, the number of HP1α foci was
maintained 8 hours after TGFβ treatment in the absence of Snail1. This
confirms that HP1α released from chromatin is Snail1 dependent.
As expected, same results were obtained in LOXL2 knock-down
conditions (Figure R14-C), which confirm that HP1α release from
chromatin is also LOXL2 dependent. Moreover, they suggest that
heterochromatin reorganization, as observed by following HP1α
localization during EMT, depends on the action of Snail1 transcription
factor through LOXL2.
Figure R13. HP1α is released from chromatin after 8h of TGFβ treatment. (A)
Purified nuclei from NMuMG cells treated with TGFβ were extracted with NaCl
(0.5 M) and the extracted fractions or the inputs were analyzed by Western blot.
Quantification is shown at the bottom. (B) Release of HP1α from the chromatin
fraction of NMuMG cells treated with TGFβ was monitored by Western blot.
HP1α levels were standardized using H3 levels. The quantification shows the
increase or decrease in the amount of HP1α of treated compared to untreated
cells (time point 0 hours), which was set as 1.
Figure R14. Release of HP1α from chromocenters is Snail1 and LOXL2
dependent. (A) NMuMG cells were transfected with siRNA control or siRNA
against Snail1. After 24 hours cells were treated with TGFβ for 8 hours. Snail1
induction was detected by immunofluorescence. Scale bar 50 µm. (B)
Alternatively, immunofluorescence against HP1α was performed in the same
conditions. Total number of HP1α foci per nucleus was quantified in siControl
and siSnail1 cells, before and 8 hours after treatment. (C) NMuMG cells were
transfected with shRNA control or shRNA against LOXL2. After 24 hours, cells
were treated with TGFβ. The number of HP1α foci per cell was quantified prior
and 8 hours after TGFβ addition. Error bars indicate SD in at least three
independent experiments. One asterisk indicates p<0.05; two asterisks indicate
p < 0.01.
To investigate whether the heterochromatin reorganization that occurs
during EMT is linked with pericentromeric transcription regulation, we
analyzed major satellite transcription after TGFβ treatment in NMuMG
cells. Interestingly, we observed an inverse correlation between Snail1
protein levels (Figure R15-A) and major satellite transcripts expression
(Figure R15-B). In this way, the up-regulation of Snail1 protein after TGFβ
treatment was accompanied by a down-regulation of major satellite
Figure R15. Snail1 up-regulation during EMT correlates with major satellite
repression. (A) Snail1 induction upon TGFβ treatment is shown by Western blot
in NMuMG cells at different time points. (B) qRT-PCR shows major satellite
transcripts levels after normalization versus an endogenous control in NMuMG
cells treated with TGFβ. Error bars indicate SD in at least three independent
experiments. Two asterisks indicate p < 0.01.
In addition, ChIP experiments showed that Snail1 binds to major satellite
regions (Figure R16-A) in a TGFβ-dependent manner, and according to
Snail1 protein kinetics. Moreover, Snail1 binding to major satellite
sequences correlates with an increase in the levels of oxidized H3 in
pericentromeric regions (Figure R16-B), that perfectly fits with major
satellite transcripts down-regulation (Figure R15-B).
To confirm the role of Snail1 in major satellite transcripts repression
during EMT, we depleted Snail1 in NMuMG by transfection of siRNAs
against Snail1 (Figure R17-A). As expected, the repression of major
satellite transcription after TGFβ treatment was impaired in Snail1 knockdown conditions (Figure R17-B), confirming that Snail1 is responsible for
this repression.
Figure R16. Snail1 binding to major satellites after TGFβ treatment correlates
with increased H3 oxidation. (A) ChIP of Snail1 in major satellite regions in
NMuMG cells treated with TGFβ. (B) Re-ChIP for oxidized H3 in major satellite
regions in NMuMG cells after TGFβ treatment. qRT-PCR data from major satellite
regions were normalized to the input and expressed as fold enrichment relative
to the data obtained in the untreated condition (TGFβ 0 hours), which was set as
1. Error bars indicate SD in at least three independent experiments. One asterisk
indicates p< 0.05.
Figure R17. Snail1 is responsible for major satellite down-regulation during
EMT. (A) Western blot showing Snail1 induction upon TGFβ treatment in
siControl NMuMG cells, and absence of Snail1 up-regulation in siSnail1 cells. (B)
qRT-PCR shows major satellite transcription in NMuMG cells transfected with
either siRNA control or siRNA against Snail1 after EMT induction by TGFβ. Data
were normalized versus an endogenous control and data from untreated (TGFβ
0 hours) siControl NMuMG cells was set as 1. Error bars indicate SD in at least
three independent experiments. One asterisk indicates p< 0.05.
We next decided to determine whether LOXL2 was also involved in
pericentromeric transcription repression during EMT, based on the
previous data and the enrichment of oxidized H3 in major satellite
sequences during this process. With this purpose, LOXL2 was depleted in
NMuMG cells by shRNA infection (Figure R18-C). In the absence of
LOXL2, pericentromeric transcription kinetic that normally occurs during
EMT was deregulated (Figure R18-A), confirming its role in major satellite
transcripts regulation during this process.
Figure R18. LOXL2 regulates pericentromeric transcription during EMT through
H3 oxidation. (A) qRT-PCR shows major satellite transcription in NMuMG cells
infected with a shRNA against LOXL2 upon EMT induction by TGFβ. (B) ChIP
analysis of oxidized H3 under the same conditions. Error bars indicate SD for at
least three independent experiments. One asterisk indicates a p < 0.05, while
two indicate p < 0.01. (C) Western blot shows the levels of LOXL2 in Control and
LOXL2 knock-down conditions.
In agreement, the levels of oxidized H3 did not change after TGFβ
treatment in LOXL2-depleted cells (Figure R18-B), which explains why
major satellite transcripts were not repressed in the absence of LOXL2.
These results suggest that Snail1 works together with LOXL2 to actively
regulate major satellite transcription during EMT induction.
6. Pericentromeric transcription regulation is essential for a
complete EMT.
Upon TGFβ treatment and during the consequent EMT induction, major
satellite transcription is tightly regulated and suffers a transient
repression induced by Snail1 and LOXL2. At this point, we decided to
address whether this down-regulation of pericentromeric transcription
was relevant and necessary for the EMT process to occur.
Major satellite transcripts were ectopically expressed in NMuMG cells
under a constitutive promoter to prevent the endogenous downregulation that takes place after TGFβ treatment (Figure R19-C). To
characterize heterochromatin reorganization in these conditions, control
(NMuMG-Control) and stable major satellite overexpressing NMuMG
cells (NMuMG-Major) were treated with TGFβ, and DAPI staining and
HP1α immunofluorescence were performed at different time points
(Figure R19-A). We observed that ectopic expression of major satellite
transcripts blocked the release of HP1α during EMT (Figure R19-B),
suggesting that pericentromeric repression during this process is
involved in HP1α dynamics.
To fully understand how major satellite overexpression was affecting
EMT, we decided to compare the global transcriptome of NMuMGControl and NMuMG-Major cells 8 hours after TGFβ treatment.
Microarray experiments revealed that 1269 genes were differently
regulated by TGFβ in NMuMG-Major cells respect to NMuMG-Control
cells. Interestingly, gene ontology analysis indicated that those genes
were mainly associated to pathways related with cell cycle, cancer, cell
death and survival, and cellular movement and assembly (Figure R20-A).
Moreover, we observed that classical mesenchymal genes activated
during EMT, such as FN1, ZEB1, ZEB2, SNAI2, as well as several MMPs,
presented higher expression levels upon TGFβ treatment in NMuMGControl compared to NMuMG-Major cells (Figure R20-B). Also several
epithelial genes were differently regulated (Figure R20-B). This suggests
that NMuMG-Major cells fail to induce mesenchymal genes and to
repress epithelial genes in the same way that NMuMG-Control cells do.
Figure R19. Major satellite overexpression blocks HP1α release from chromatin
during EMT. (A) Control or NMuMG cells ectopically expressing major satellites
were treated with TGFβ and stained with DAPI or immunolabeled for HP1α at
different time points. Scale bar, 10 µm. (B) The average number of HP1α foci per
cell in NMuMG-Control and NMuMG-Major is shown in the graph. (C) qRT-PCR
shows major satellite expression levels in NMuMG-Control and NMuMG-Major.
Values represent fold change with respect to untreated NMuMG-Control. Error
bars indicate SD in at least three independent experiments. One asterisk
indicates p<0.05; two asterisks indicate p < 0.01.
Figure R20. Genes differently regulated by TGFβ in NMuMG-Major cells are
mainly associated to cancer and EMT related pathways. (A) Gene ontology
analysis of differentially expressed genes in NMuMG-Major cells compared with
NMuMG-Control cells 8 hours after TGFβ treatment. (B) List of selected
mesenchymal and epithelial genes that are differentially expressed in NMuMGMajor cells compared with NMuMG-Control cells upon TGFβ treatment. LogFC
indicates Log2 fold changes.
From all genes that were differently regulated in the two conditions, we
chose some of the ones reported to be associated to EMT and validated
the observed changes in the microarray by qRT-PCR (Figure R21-A). In
addition, the impaired up-regulation of other classical mesenchymal
markers such as Fibronectin and N-cadherin was also validated by
Western blot (Figure R21-B).
Although many classical EMT related genes appeared to be differently
regulated in the microarray after TGFβ treatment, CDH1 mRNA downregulation was not statistically different in the two cell populations
(Figure R21-C). However, typical loss of E-cadherin protein was not
observed in NMuMG-Major cells upon 24 hours of TGFβ treatment
(Figure R21-D), suggesting major satellite RNAs may regulate EMT by
different mechanisms besides transcriptional regulation.
Thus, these results show that the TGFβ dependent conversion of
epithelial cells to mesenchymal cells was affected when major satellite
were overexpressed, and this suggests that tightly regulation of
pericentromeric transcription is crucial for this process to take place.
Figure R21. Genes involved in EMT are differently regulated by TGFβ in
NMuMG-Major cells. (A) Validation by qRT-PCR of selected genes differently
expressed in NMuMG-Control and NMuMG-Major after TGFβ treatment
according to microarray data. Gene expression was normalized against an
endogenous control and presented as fold enrichment respect untreated
NMuMG-Control cells, which were set as 1. Error bars indicate SD in at least
three independent experiments. One asterisk indicates p<0.05; two asterisks
indicate p< 0.01. (B) Western blot showing expression of Fibronectin and Ncadherin in NMuMG-Control and NMuMG-Major cells prior and after TGFβ
treatment. (C) qRT-PCR of CDH1 mRNA in NMuMG-Control and NMuMG-Major
at different time points after TGFβ treatment. Gene expression was normalized
against an endogenous control and presented as fold enrichment respect
untreated NMuMG-Control cells, which were set as 1. (D) Expression of Ecadherin protein levels in NMuMG-Control and NMuMG-Major prior and after
24 hours of TGFβ treatment was determined by immunofluorescence.
To detect if these changes in gene expression also affect functional
properties of mesenchymal cells, we did migration and invasion assays
with NMuMG-Control and NMuMG-Major cells. We used 10% FBS as
chemoattractant and, as previously reported160, NMuMG-Control cells
treated with TGFβ showed an increase in migration and invasion capacity
compared to non-treated cells (Figure R22). However, this increase in
migration and invasion properties was much smaller in NMuMG-Major
cells (Figure R22). Similar results were obtained with transient major
satellite overexpressing cells (data not shown). This confirms that major
satellite transcription regulation is essential for the acquisition of
functional properties in mesenchymal cells.
Figure R22. Major satellite cells have decreased migration and invasion
properties. NMuMG-Control and NMuMG-Major cells were either non treated
or treated with TGFβ and, after 24 hours, cells were reseeded on transwell
chambers and incubated for 10 hours (left panel, migration), or placed in
Matrigel-coated transwells and incubated for 24 hours (right panel, invasion).
Non-migrating and non-invading cells were removed from the upper surface of
the membrane, while cells present at the lower surface were fixed and stained
with DAPI. The DAPI-stained nuclei were counted in four different fields per filter
by ImageJ software. Error bars indicate SD for at least three independent
experiments. One asterisk indicates p < 0.05.
Finally, we further assessed the relevance of HP1α and major satellites in
EMT by knocking-down these two elements. HP1α was efficiently downregulated by siRNA transfection in NMuMG cells. However, the up78
On the other hand, pericentromeric transcription was knocked-down in
NMuMG cells by transfection of gapmers against major satellite
transcripts. The up-regulation of mesenchymal genes upon TGFβ
treatment was not affected by major satellite depletion (Figure R23-B).
Figure R23. HP1α and major satellite transcripts knock-down does not affect
mesenchymal genes induction after TGFβ treatment. NMuMG cells were
transfected with siRNA control or siRNA against HP1α (A) or with LNA-DNA
control or major gapmers to deplete major satellite expression (B). After
transfection, cells were treated with TGFβ for 24 hours when indicated. Levels of
fibronectin, Zeb1/Zeb2, Slug (Snail2), CD44, HP1α and major satellites were
determined by qRT-PCR. RNA levels were normalized against and endogenous
control, and presented as fold enrichment respect untreated control cells, set as
1. Error bars indicate SD in at least three independent experiments.
regulation of mesenchymal markers observed 24 hours after TGFβ
treatment in NMuMG cells was not significantly affected by depletion of
HP1α (Figure R23-A). This suggests that HP1α is not required for the
induction of mesenchymal markers during the first hours after TGFβ
treatment. Moreover, it also shows that HP1α release from
chromocenters, mimicked by HP1α depletion, is necessary but not
sufficient for EMT induction, since TGFβ untreated HP1α depleted cells
did not induce mesenchymal markers (Figure R-23A).
Chromocenters result from the clustering of pericentromeric regions of
several chromosomes in the nucleus71. Interestingly, specific
chromocenter organization pattern can be found in different cell types,
as shown by the different amount and size of chromocenters present in
their nucleus72. Thus, clustering of pericentromeric regions to generate a
particular number of chromocenters seems to be a regulated and cell
type dependent process. However, it is still under study whether
interchomosomal interactions found in each chromocenter are always
the same. In this context, two scenarios may be plausible:
pericentromeric domains of specific chromosomes tend to cluster
together in a particular cell type, or pericentromeric clustering of several
chromosomes occurs randomly. In the nucleus, chromatin organizes into
chromosome territories, which are spatially defined domains
characterized by non-random interactions between chromosome
fibers10. Thus, the idea that chromocenters clustering is synchronized and
chromosome specific fits better in this scenario.
We have shown that, in mesenchymal cells, absence of Snail1 associates
with heterochromatin organization alterations. In particular, Snail1
depletion in pMEFs causes a decrease in the number of chromocenters
per cell, which tend have increased size compared to control cells. In
addition, chromocenter organization also changes upon TGFβ treatment
after Snail1 up-regulation. In this case, epithelial cells appear to have less
and bigger chromocenters compared to those cells that have undergone
EMT. Since different cell types are characterized by specific
chromocenter organization, it is not surprising that epithelial cells and
the mesenchymal cells that result from an EMT process differ in their
chromocenter organization, but further exemplifies the wide differences
that exist between these two cell types.
Moreover, the fact that different chromocenter organization seem to be
a hallmark of cell identity and that dynamic reorganization of
chromocenters takes place during EMT suggests that it may be a tightly
regulated process.
Snail1 and heterochromatin organization
Nevertheless, how chromocenters are organized and which key
regulators are involved in orchestrating dynamics changes in this
organization during physiological processes is poorly understood. During
the first stages of mouse embryonic development, major satellite
transcripts are involved in de novo establishment of chromocenters after
the epigenetic reprograming that takes place after fertilization80.
Additionally, in neuronal and muscle differentiation, changes is histone
modifications in pericentromeric regions seem to modulate major
satellite transcription and in this way determine its clustering. Although
there is still some controversy about whether active or repressive marks,
as well as condensed or decondensed chromatin state is favoring major
satellite transcription, it is usually associated to loss of repressive marks
. Regardless, the link between major satellite transcription and
chromocenter organization seems evident.
Recently, it was shown that the transcription factors Pax3 and Pax9 are
required for H3K9 trimethylation and major satellite repression, although
the effect on chromocenter clustering was not analyzed88. Interestingly,
here we have shown for the first time that Snail1, another transcription
factor, is not only involved in heterochromatin organization in
mesenchymal cells, but also in its reorganization during a physiological
process as is EMT.
In Snail1 KO pMEFs, the decrease in the number of chromocenters and
their bigger size suggest an increased clustering of pericentromeric
regions into chromocenters compared to control cells. Accordingly,
epithelial cells, that do not express Snail1, have less and bigger
chromocenters than mesenchymal cells. This is not only observed in
epithelial NMuMG cells and the EMT resulting mesenchymal cells, but
also in other epithelial and mesenchymal cell lines (data not shown). All
this suggests that the presence of Snail1 is limiting the clustering of
chromocenters since when it is expressed cells tend to have an increased
number of chromocenters. Although the mechanism by which Snail1 is
able to regulate chromocenter clustering has to be further characterized,
it is tempting to speculate that pericentromeric transcription repression
by this transcription factor could be involved. Increased major satellite
transcription has been associated with increased heterochromatin
condensation and HP1α binding70, which could be the cause of increased
clustering. Accordingly, different cell types not only have different
average number of chromocenters but also differ in major satellite RNA
expression. Therefore, it would be interesting to determine whether
major satellite transcripts levels do always inversely correlate with the
average number of chromocenters per cell.
Loss of Snail1 not only causes changes in chromocenter organization but
also up-regulation of major satellite transcription in pMEFs. This upregulation is also observed when the transcription factors Pax3 and Pax9
are depleted and, in this case, it correlates with loss of H3K9me3 at
chromocenters and diffuse staining of this histone modification in the
nucleus88. However, despite the increase in major satellite transcription,
no changes in H3K9me3 localization were observed in Snail1 depleted
pMEFs and chromocenters visualized by DAPI staining were positive for
H3K9me3. This suggests that other histone modifications might be
involved in this regulation. Interestingly, we have partially elucidated the
epigenetic mechanism by which Snail1 regulates transcription of these
heterochromatin domains. Snail1 may recruit the histone-modifying
enzyme LOXL2 that removes the trimethylated amino group in lysine 4.
The absence of the amino group suggests that the demethylated histone
cannot be methylated again since none histone aminotransferase
enzymes have been identified to date, and this may establish a highly
repressive mark. Thus, the presence of LOXL2 and oxidized H3 in
heterochromatin explains the absence of H3K4me3 in these regions.
Interestingly, oxidized H3 is not a static modification since Snail1 and
LOXL2 change its levels in pericentromeric regions and in this way
repress major satellite transcription in mesenchymal cells. Thus, other
mechanisms, such as histone exchange162 or histone tail clipping163, may
be involved in the removal of this new modification to allow dynamic
changes of oxidized H3 levels in different cell contexts and processes.
Snail1 and heterochromatin transcription
We also show that Snail1 is responsible for tightly regulation of major
satellite transcription during a physiological process as EMT is. Little is
known about the precise role that major satellite RNAs have in the cell
and how can they function differently depending on cell context.
However, the accurate kinetics of major satellite transcription during
EMT reinforces the idea that, rather than being unspecifically expressed,
major satellite transcription is a controlled and physiologically relevant
event. Lately, several functions related to nuclear organization have been
described for lncRNAs. They are involved in cis and trans gene regulation,
in the formation or remodeling of nuclear domains through recruitment
or sequestration of nuclear proteins, and they participate or favor
interactions between different chromosomal regions164. Since major
satellites are also lncRNAs, these mechanisms may contribute to their
function in the cell. Indeed, it has already been reported that major
satellite RNAs interact with and target HP1α to pericentromeric regions
to establish heterochromatin domains.
Moreover, knock-down experiments show that not only Snail1 but also
LOXL2 is involved in the regulation of major satellite transcription during
EMT, and that it depends on LOXL2 mediated H3k4me3 deamination
since the levels of oxidized H3 in pericentromeric domains also change
during EMT. This indicates that, also during EMT, Snail1 binds to
pericentromeric regions with LOXL2 after TGFβ treatment, and together
they repress pericentromeric transcription by H3 oxidation.
The recruitment of LOXL2 by Snail1 to specific DNA sequences not only
occurs in heterochromatin domains. Indeed, we previously published
that Snail1 recruits LOXL2 to CDH1 promoter to induce gene silencing30.
Other histone modifying enzymes such as Polycomb and LSD1 have also
been shown to be recruited by Snail1 to CDH1 promoter and cooperate
in the repression134,135. Thus, it is tempting to speculate that Snail1 could
recruit other histone modifying enzymes, besides LOXL2, to
pericentromeric regions, and in this way further contribute to the
Several reports show the presence of Polycomb repressive histone marks
on repetitive elements165. Both murine leukemia virus (MLV) and IAP
Heterochromatin is characterized by hypomethylation of H3K4. We have
shown that LOXL2 oxidases H3 in pericentromeric regions and in this way
contributes to the H3K4 hypomethylated status of heterochromatin.
LSD1 and LSD2 also demethylate H3K4. Whereas LOXL2 deaminates
trimethylated H3K4, LSD1 and LSD2 act on mono- and dimethylated
H3K4170,171. Unpublished data from Rutenberg and colleagues shows that
LSD2 localizes to pericentromeric heterochromatin and is essential to
maintain it stability172. In addition, LSD1 interacts and cooperates with
Snail1 in CDH1 repression156 and its deletion in S. pombe results in the
propagation of pericentromeric heterochromatin beyond its physiological
boundaries173. Thus, these histone demethylases may also contribute to
the establishment or maintenance of heterochromatin epigenetic
Snail1 is not the only transcription factor involved in CDH1 repression,
but several others have been described to contribute to its silencing103.
The identification of Snail1 as a regulator of major satellite transcription
raises the question whether other transcription factors previously
described to repress CDH1 could be also involved in major satellite
retroelements are targets of Polycomb complexes, and loss of PRC1 and
PRC2 components causes a strong increase in expression of LTR
retrotransposons, which in turn induces their active mobilization166.
Human Polycomb proteins also associate to large pericentromeric
heterochromatin domains167. Accordingly, our ChIP experiments show
that both Ezh2 and Suz12 bind the human pericentromeric regions in chr
4, 7 and 20 identified in the Snail1 ChIP-Seq, which are also enriched in
H3K27me3 (not shown). Moreover, in embryonic stem cells,
pericentromeric heterochromatin acquires H3K27me3 in the absence of
the Suv39h168. In mouse zygotes, only maternal constitutive
heterochromatin is labelled by H3K9me3 and H4K20me3, and PRC1 is
responsible for paternal major satellite repression169. Thus, although
methylation of H3K9 appears to be the main mechanism by which major
satellite domains are repressed since Suv39h dominates over PRC2,
Polycomb proteins may arise in the future as key heterochromatin
transcription regulators in other specific conditions, which might be
regulated by Snail1.
silencing. Zeb1 and Zeb2 are zinc finger proteins that share some of
Snail1 target genes and bind to the same DNA sequences, the E-boxes174.
During EMT, Zeb1 and Zeb2 cooperate with Snail1 in the repression of
CDH1. Indeed, Zeb1 induction occurs upon Snail1 activation and it has
been proposed that, whereas Snail1 has a more important role in the
initiation of EMT and CDH1 repression, Zeb1 downstream activation is
more relevant for the maintenance of this repression. Moreover, Zeb1
depletion causes up-regulation of major satellite transcription in MEFs88.
Thus, it seems plausible that, as happens in CDH1 repression, not only
Snail1 but also Zeb proteins may be involved in major satellite regulation
during EMT.
Interestingly, Snail1 and Zeb1 use different epigenetic mechanisms in
CDH1 repression. Whereas Snail1-associated epigenetic regulators have
been extensively studied, Zeb1 has been functionally associated only
with the CtBP co-repressor complex that contains CtBp1/2, HDAC1/2,
G9a and CoREST proteins among others175. Thus, it can be proposed that
Snail1 and Zeb1 share those interactors involved in initiating the
repression process (HDACs, CoREST and LSD1), but contribute in a
different way to the establishment of a repressive mark: H3K27me3 via
Snail1 through PCR2 and H3K9me3 via Zeb1 through G9a. This scenario
also fits perfectly in a pericentromeric heterochromatin context. Indeed,
depletion of Zeb1 in MEFs causes loss of H3K9me3 in chromocenters88,
which could be explained by loss of G9a recruitment by Zeb1 to those
domains. Moreover, Zeb1/2 and LOXL2 mRNA are induced upon TGFβ
treatment at the same time (unpublished data), which suggests they may
be required in the same context. Thus, it would be interesting to
determine whether LOXL2 also interacts and cooperates with Zeb
proteins in this context.
All this suggests that different transcription factors may be involved in
major satellite transcription regulation by recruitment of different
histone modifying enzymes. These, in turn, may be responsible for the
establishment of different types of repressive marks, and their
combination may result in different degrees of heterochromatin
Interestingly, Snail1 ChIP-Seq also showed that 40% of predicted binding
sites were located in intergenic regions (Figure D1-B). This suggests that,
besides major satellite transcription regulation, Snail1 could also be
involved in the regulation of other ncRNAs. Accordingly, Slug
transcription factor, which shares some of Snail1 functions, binds and
regulates the insulator activity of B1 SINE retrotransposons176. In
addition, we observed an up-regulation of minor satellite transcripts in
pMEFs after Snail1 depletion. However, this does not occur in Snail1 KO
iMEFs. Thus, rather than being directly regulated by Snail1, minor
satellite transcripts up-regulation may be a consequence of major
satellite deregulation, that immortalized cells have already overcome.
Transcription factors bind to specific sequences in the DNA, such as Eboxes in the case of Snail1. We have shown that Snail1 binds
pericentromeric regions in mouse cells. However, although there are
binding sites for other transcription factors, none canonical E-boxes are
found in the consensus sequence for major satellite repeats. Despite
being highly conserved, slight differences exist between the hundreds of
major satellite repeats found in the genome88. Thus, it might be possible
that Snail1 binds to a particular subgroup of major satellite repeats,
characterized by the presence of E-boxes, and in this way regulates
transcription. We have also shown that Snail1 interacts with the
heterochromatin structural protein HP1α. Thus, Snail1 could also bind
major satellite indirectly through interaction with HP1α. However,
preliminary EMSA assays show that Snail1 is able to bind directly major
satellite sequences in vitro despite the lack of canonical E-boxes (data not
shown). This suggests that Snail1 binding to DNA may be more
promiscuous than previously thought. Accordingly, a Snail1 mutant that
lacks the last zinc finger is unable to bind E-boxes in CDH1 promoter, but
still binds major satellite sequences (data not shown), suggesting Snail1
binding to major satellite may mechanistically differ from the classical
binding to E-boxes. In agreement, only 20% of predicted binding sites
identified in the Snail1 ChIP-Seq experiment contain E-boxes (Figure D1A). Indeed, from the three pericentromeric sequences were Snail1
binding was validated by ChIP, only the one located at chromosome 7
contains an E-box. This further supports the idea that Snail1 binding to
DNA may be more flexible and not strictly limited to canonical E-boxes.
Figure D1. Snail1 ChIP-Seq predicted binding sites analysis. (A) Presence of Eboxes across the predicted binding sites, where canonical Snail1 E-boxes are 5’CACCTG-3’ and 5’-CAGGTG-3’, non-canonical Snail1 E-boxes are 5’-CACGTG-3’
and 5’-CAGCTG-3’ and Snail1-box-like is 5’-CAGGTT-3’. (B) Classification of all
binding sites according to where they map: Expressed Sequence Tags (ESTs),
genes or intergenic regions. (C) Classification of the binding sites which overlap a
gene in relation to the gene region to which they align. RefSeq genes were used
as reference and the 5 kb region upstream of the transcription start site was
defined as the promoter.
Furthermore, only 4% of binding sites located in genes were in
promoters, while most of them were in exons and especially in introns
(Figure D-3). Thus, besides the classical gene repression by binding to
promoters, Snail1 may be involved in other types of gene transcriptional
regulation such as splicing. Indeed, Snail1 regulates the levels of Zeb2
Natural Antisense Transcript (NAT)177 and is involved in splicing
regulation of LEF-1 NAT (M. B, E. A-P and A. GH; submitted). However,
the precise mechanism by which Snail1 is regulating these lncRNAs has to
be further characterized.
In mouse cells, major satellite transcripts have been detected as long and
short species in different phases of cell cycle64. Snail1 may be involved in
regulating the expression of long species of major satellite RNAs since
those are the ones expressed during interphase. However, further
experiments will be required to confirm this, as well as to determine
whether the length of these ncRNAs may also be important for their
regulation or function. Moreover, several reports demonstrate that
forward and reverse major satellite transcripts can be differently
expressed and regulated and that they are involved in different
functions70,80. Thus, it would be interesting to determine whether Snail1
is regulating sense specific major satellites RNAs or whether forward and
reverse transcripts are differently regulated during EMT.
We have shown that, during TGFβ induced EMT, major satellite
transcription is characterized by tightly regulated kinetics. Knock-down
experiments show that both Snail1 and LOXL2 are responsible for the
transient down-regulation of major satellite observed during this
process. However, after the initial repression, major satellite transcripts
tend to be up-regulated again later during the EMT process despite the
presence of Snail1 in the resulting mesenchymals cells. Several
explanations may account for this observation.
The existence of a mechanism that regulates transient binding of Snail1
to pericentromeric regions after TGFβ would explain why Snail1
mediated major satellite repression takes place especially at the onset of
EMT. However, the fact that Snail1 is also found in pericentromeric
regions in MEFs makes this possibility rather unlikely. Two other possible
explanations may fit better this scenario.
On one hand, Snail1 protein levels fluctuate during EMT. In particular,
there is a peak of Snail1 expression 2 hours after TGFβ treatment, but
Snail1 protein levels progressively decrease subsequently. Thus, the
amount of Snail1 molecules available to bind major satellites differs in
both contexts, and changes in transcription may be consequence of
changes in the availability of transcription factors.
On the other hand, binding sites of several transcription factors are
found in major satellite repeats88. This suggests that different
transcription factors may be able to bind and regulate pericentromeric
transcription. Interestingly, expression of transcription factors is usually
cell type and context dependent. Although most transcription factors
Major satellite regulation during EMT
involved in major satellite transcription regulation are repressors, it is
presumably that others may contribute to transcription activation. Thus,
in a particular cell context such as the mesechymal cells resulting from
EMT, Snail1 may not be the only transcription factor bound to major
satellite. Epithelial cells treated with TGFβ undergo important
transcriptome and proteome changes when they are converted to a
mesenchymal state. Therefore, transcription factors that cooperate with
Snail1 in the repression may be silenced or, alternatively, transcription
factors that induce major satellite transcription may be up-regulated and
compensate Snail1 mediated repression. In this way, the combination of
activation and repression activities of several transcription factors could
regulate and determine the final major satellite transcriptional output.
As previously mentioned, major satellite transcripts are first downregulated and partially recovered upon TGFβ treatment. These kinetics
are crucial for EMT to take place since ectopically expression of major
satellite transcripts that block the endogenous down-regulation prevents
many of the changes in gene expression that occur during EMT.
Moreover, major satellite overexpressing cells present lower invasion
and migration capability than control cells upon TGFβ treatment, which
further indicates an incomplete EMT.
It has been previously suggested that an epigenetic reprogramming
occurs during EMT since some histone modifications are globally
modified in this process. Specifically, there is a global reduction of
H3K9me2 in LOCKs and an increase in H3K4me3 and H3K36me3148, which
suggests reactivation reorganization of large facultative heterochromatin
Although few transcription factors have been implicated in genome
organization so far88,178, we propose that global chromatin reorganization
takes place during EMT to allow the acquisition of a new transcriptional
state, and that this depends on major satellite down-regulation by Snail1
and LOXL2. Our data provide new evidence that transcription factors not
only regulate specific gene transcription but may also have a critical role
in establishing a functional nuclear genome organization during the
conversion of epithelial to mesenchymal cells.
We have shown that major satellite overexpression prevents many of the
transcriptional changes that occur during EMT. Thus, it is tempting to
speculate that when major satellite transcripts are overexpressed
blocking the endogenous down-regulation, chromatin reorganization is
compromised and this affects the typical activation and repression of
genes that takes place during EMT.
During interphase, centromeres act as boundaries that decrease
interaction between chromatin arms15. These interactions may change
during EMT, when new clustering of active or inactive genes may need to
occur. In this way, reorganization of pericentromeric domains may allow
euchromatin reorganization and the establishment of new and different
interactions between chromosome arms. Interestingly, when major
satellite transcripts are overexpressed, HP1α release from
heterochromatin is blocked and this may compromise chromocenters
reorganization, which may difficult further reorganization of euchromatin
Major overexpression affects both activation of mesenchymal and
repression of epithelial genes during EMT. However, although activation
of the most classical mesenchymal genes such as SNA2, ZEB1/2 and FN1
seem to be compromised in these conditions, repression of classical
In general, nuclear organization of chromatin reflects its active or
inactive state. Euchromatin occupies the internal nucleoplasm, whereas
heterochromatin preferentially localizes at the nuclear and nucleolar
periphery179,180. In addition, nuclear machineries are not uniformly
distributed in the nucleoplasm, but are organized in functional subcompartments181–183. For instance, in “transcription factories” different
genes, localized on distant chromosomal loci, can associate to the same
active foci to be co-transcribed23. On the other hand, silenced genes tend
to cluster in LOCKs, LADs or other repressed domains. These nuclear subcompartments are dynamic since transcriptional requirements vary in
different cell types. Thus, some genomic regions may need to move from
one location to another during EMT to allow or block gene transcription.
Likes this, we would expect that epithelial genes move to LADs to be
silenced, and mesenchymal genes migrate to transcription factories.
epithelial genes such as CDH1, CLDN, OCLN and DSP is not affected.
Interestingly, relocation to the periphery is not always necessary for
silencing as many inactive loci are located within the nucleoplasm away
from the nuclear periphery. In addition, movement is not always
required for gene activation6. Thus, transcriptional changes occurring
during EMT may be regulated by different mechanisms, which may or not
involve gene movement. This may explain why not all transcriptional
changes occurring during EMT are compromised when major satellite
transcripts are overexpressed. Presumably, the genes affected would be
those whose activation or repression depends on gene movement and
genomic reorganization.
On the other hand, major satellite knock-down did not alter the TGFβ
induced EMT. This suggests that major satellite down-regulation is
important in the initial steps of the transition, but has no effect later on
in this process. Cells undergo complete EMT in terms of gene expression
despite low levels of major satellite transcripts even 24 hours after
induction, when major satellite transcripts are up-regulated in control
In any case, further work will be necessary to determine which chromatin
domains and how do they move during EMT. However, it seems evident
that the heterochromatin regulation that takes place in a specific window
of time during TFGβ induced EMT is essential for the genome
reorganization required to acquire mesenchymal traits.
Interestingly, microarray experiments showed that not only genes but
also other ncRNAs were differently regulated after TGFβ treatment when
major satellite RNAs were overexpressed. Indeed, more than 30% of
differentially regulated transcripts corresponded to an RNA component,
including ncRNas and nuclear RNAs. This suggests that major satellite
transcripts, which are themselves ncRNAs, regulate the transcription of
other ncRNAs.
HP1 dynamics during EMT
We have shown that during EMT there is a transient delocalization of
HP1α from chromocenters that is Snail1 dependent. Nevertheless, HP1α
is subsequently reloaded to chromocenters despite the fact that Snail1 is
still expressed. In addition, although Snail1 depletion causes decreased
number of HP1α foci in pMEFs, no changes in HP1α localization are
observed and it remains permanently associated to chromocenters in
these cells.
Release of HP1α from chromocenters during EMT is not only dependent
on Snail1 but also on LOXL2 and major satellite transcription downregulation, since both shLOXL2 and major satellite RNA overexpression
block HP1α delocalization. HP1α directly interacts with major satellite
transcripts and it has been proposed that these transcripts are involved
at least in de novo targeting of HP1α to pericentromeric
heterochromatin70. Thus, major satellite repression could lead to HP1α
delocalization due to loss or decreased recruitment by major satellite
On the other hand, several posttranslational modifications have been
described in HP1α and they seem to regulate HP1α affinity for chromatin
as well as the interaction with other proteins62. In particular, methylation
of HP1 increases affinity for chromatin and, consequently, demethylation
may cause release. However, none HP1α demethylase has been reported
so far. Interestingly, HP1α interacts with LOXL2, which presents a PXVSL
HP1α interacting domain. In addition, some new nuclear LOXL2
substrates are being identified in our laboratory besides histone H3.
Thus, oxidation of methylated HP1α by LOXL2 could be a possible
mechanism to explain HP1α release from chromocenters upon TGFβ
This suggests that HP1α delocalization from chromocenters during EMT
may depend on particular elements activated or repressed by Snail1,
specifically upon TGFβ induction. Moreover, the fact that HP1α is
subsequently reloaded to chromocenters suggests that the mechanism
that regulates this process may suffer a negative feedback regulation,
which may also be TGFβ dependent, and in this way allow new binding of
HP1α to major satellite.
treatment. Understandably, other histone modifying enzymes with
demethylase activity such as LSD1 or LSD2 may also be good candidates
to modify HP1α and in this way cause its release from chromatin.
Reloading of HP1α into chromocenters observed 24 hours after TGFβ
treatment could also depend on HP1α posttranslational modifications. Nterminal phosphorylation of HP1α plays a central role in its targeting to
chromatin since it enhances HP1α affinity for H3K9me3184. Interestingly,
phosphorylation of HP1α is likely to be mediated by CK2, which has been
proposed to regulate EMT in cancer cells185,186. As previously mentioned,
major satellite transcripts are involved in de novo targeting of HP1α to
heterochromatin. Specifically, forward major satellite transcripts interact
with SUMO-HP1α and guide it to pericentromeric heterochromatin70.
Thus, up-regulation of major satellite observed after HP1α delocalization,
together with Snail1 dependent sumoylation of HP1α could also
contribute to HP1α reloading into heterochromatin.
Besides HP1α posttranslational modifications, other mechanisms have
been described to cause HP1α release from chromatin. For instance,
phosphorylation of serine 10 in histone H3 (H3S10) by Aurora B kinase
causes HP1α dissociation from heterochromatin during mitosis187. While
both Aurora A and B can phosphorylate H3S10 in vitro, only Aurora B colocalizes with phosphorylated H3S10 during early stage of mitosis188.
However, little is known about the role they may have in interphase cells.
Aurora kinases are overexpressed in a wide range of human cancers189. In
particular, Aurora A overexpression can lead to cell transformation190,191
and its inhibition suppresses EMT and invasion in carcinoma cells192. In
addition, Aurora B is stabilized during the first hours after TGFβ
treatment in epithelial cells193. This stabilization takes place at the time
when HP1α begins to be released from heterochromatin in our TGFβ
induced model. Thus, stabilization of Aurora B after TGFβ might cause
transient phosphorylation of H3S10 in major satellites and in this way
induce HP1α delocalization. In agreement, both Aurora A and B mRNAs
are significantly down-regulated 24 hours after TGFβ treatment in
NMuMG cells, when HP1α is perfectly located at chromocenters in
almost 100% of cells (data not shown). In addition, mRNA of DUSP1,
which dephosphorylates H3S10194, is significantly up-regulated at that
time point (not shown). Thus, phosphorylation and dephosphorylation of
H3S10 seems a plausible mechanism for the regulation of HP1α dynamics
during EMT.
These different mechanisms proposed are not mutually exclusive.
Actually, they may coexist and together contribute to HP1α spatial
regulation. In any case, HP1α release from chromatin seems necessary to
permit the heterochromatin reorganization that takes place during the
EMT since when it is blocked, cells fail to undergo a complete EMT.
However, HP1α knock-down did not alter TGFβ induced EMT, suggesting
that HP1α release is important in the initial steps of the transition, but
that its reloading to chromocenters is not necessary for the acquisition of
the mesenchymal phenotype.
We have shown that major satellite transcripts are tightly regulated by
Snail1 during EMT and that this regulation is crucial for a complete EMT
to take place. Although EMT is being widely studied in the context of
carcinogenesis and tumor progression, it also takes place in physiological
situations, especially during development. Therefore, regulation of major
satellite transcripts may be essential in several developmental processes
such as gastrulation or neural crest delamination. Interestingly, EMT is
not the only context were Snail1 protein levels are modulated. Thus, it is
tempting to speculate that Snail1 could also regulate major satellite
transcription in other physiological situations besides EMT.
Major satellite transcripts of variable size accumulate during G1 reaching
a peak of expression at G1-S phase64. However, transcription of these
long species ceases before replication of heterochromatin domains
during mid-late S phase. Interestingly, expression of Snail1 is cell cycle
dependent with a peak of maximal expression at early S-phase
(unpublished data; A.V and A. GH), which suggests that also in this
context Snail1 could be repressing major satellite transcription.
Accordingly, Snail1 KO iMEFs, which express higher levels of major
satellite transcripts compared to control cells, have problems with
Snail1 pericentromeric regulation in other cell contexts
heterochromatin replication. The proportion of cells in S-phase is higher
compared to control iMEFs and there is a delay in the progression from
early to mid-late S-phase.
Although euchromatin replication is a process that was well described a
long time ago, there are still many questions that have to be answered
concerning heterochromatin replication. Regions in the genome with
complex chromatin organization could potentially impose structural
constraints that, if not relieved, prevent replication fork progression
without possible bypass. In this way, repression of major satellite
transcription in S-phase may be required to allow the replication fork
progression through heterochromatin domains. Recently, new players
such as CAF-1 or Np95 have been identified to play a role in
heterochromatin replication155,195. The loss of both CAF-1 and Np95
causes accumulation of cells in early S-phase, similar to what is observed
in Snail1 KO iMEFs, further supporting the implication of Snail1 in
pericentromeric replication. In addition, p150 subunit of CAF-1 has been
proposed to be involved in HP1 displacing from the chromatin ahead of
the replication fork and in its transferring onto newly replicated DNA
behind the fork. Interestingly, LOXL2 interacts both with HP1α and CAF130, which suggests it could also have a role in replication. Up-regulation
of Snail1 protein levels during S-phase could result in recruitment of
LOXL2 to pericentromeric regions, oxidation of histone H3 and major
satellite repression to allow replication.
Pericentromeric heterochromatin domains are crucial for proper
chromosome segregation and genome stability. Therefore, maintenance
and accurate reproduction of such a complex heterochromatin structure
throughout multiple cell divisions is essential to ensure its stability.
During replication, old and newly synthetized histones are incorporated
to the nascent DNA. To maintain a particular epigenetic status, histone
modifications present in the parental DNA must be preserved in the
replicated DNA. Accordingly, in human fibroblast cells, PRC2 is able to
bind H3K27me3, its own methylation site, and localizes with replication
sites throughout S-phase, suggesting that it may be implicated in
maintaining the mark at sites of DNA replication196. In addition, Mi2/NuRD complex associates with pericentromeric heterochromatin
Pericentromeric transcription varies in different cell types, and it has
been shown to be repressed in undifferentiated states and activated
during specific differentiation programs64. Members of our lab recently
published that Snail1 is required for the maintenance of mouse
mesenchymal stem cells (MSCs)151. In addition, Snail1 expression
decreases during MSCs differentiation and ectopic expression prevents
their conversion into osteoblasts or adipocytes. Thus, major satellite
transcription regulation by Snail1 during differentiation seems also
Heterochromatin is characterized by its condensed status, which
represents a significant barrier to efficient detection and repair of DNA
damage198. DNA-repair machinery must be able to detect DNA damage,
remodel the local chromatin architecture to provide access to the site of
damage and repair it, and restore the local chromatin organization found
prior to damage. Like this, early chromatin-based events after DNA
damage promote the formation of open, relaxed chromatin structures at
DSBs to allow the DNA-repair machinery to access the spatially confined
region surrounding the DSB. Accordingly, depletion of HP1 proteins can
decondense heterochromatin and promote repair of DSBs199–201 and it
has been proposed that HP1 proteins are actively ejected from the
chromatin during DNA repair202. On the other hand HP1α is rapidly
recruited to euchromatin DSBs within seconds to minutes after damage
induction, which may rapidly “heterochromatinize” the DSB region,
during S-phase in rapidly proliferating lymphoid cells, suggesting a
possible role for Mi-2/NuRD complex in the maturation and maintenance
of heterochromatin after replication197, but it is still not known how Mi2/NuRD is recruited to heterochromatin replicating foci. All this data
suggests that other chromatin repressing complexes may be recruited at
the replication fork during heterochromatin replication in order to
maintain its repressed state. Thus, Snail1 and LOXL2 could also be
involved in the oxidation of histone H3 in the daughter DNA strands to
re-establish the characteristic H3K4 hypomethylated status of
heterochromatin domains after replication. Increased amount of
pericentromeric transcripts in the absence of Snail1 could be enhanced
by improper maturation of heterochromatin in a post-replicative phase.
preventing transcription and stabilizing the chromatin structure203,204.
Importantly, HP1α recruitment is transient, and it dissociates from the
break few minutes after damage induction203. Thus, HP1α seems to play a
dual role in DNA damage repair being first recruited to DNA damage sites
to silence and protect chromatin till the recruitment of DNA-repair
machinery. At that point, HP1α would be displaced to induce chromatin
decondensation and favor DNA damage repair. Therefore, also in a DNA
damage context, HP1α dynamics seem to regulate changes in chromatin
organization. However little is known about major satellite transcripts
regulation during DNA damage or how could they contribute to HP1α
dynamics in this context. Thus, it would be very interesting to determine
whether Snail1 and LOXL2 could be also involved in the regulation of
major satellite transcription during heterochromatin DNA damage repair.
Chromatin dynamics and cancer
During last years, a connection between heterochromatin and cancer has
clearly been established. General loss of heterochromatin is common to
many cancer types and it is often associated with high malignancy and
poor prognosis205. For instance, the Barr body, which result from
chromatin condensation after X chromosome inactivation and is easily
visualizes by light microscopy in normal cells, is frequently lost in
cancer206. Sequences within LADs and LOCKs are also often
hypomethylated and decondensed in cancer207, which may cause
transcriptional hypervariability of genes relevant for tumor progression
and contribute in this way to tumor cell heterogeneity favoring drug
resistance208. In addition, pericentromeric heterochromatin is also
associated to cancer since loss of its condensed structure leads to
genome instability64. Accordingly, mouse and human pericentromeric
repeats are repressed by the tumor suppressor BRCA1 to maintain
heterochromatin structure by H2A ubiquitylation209, and major satellite
transcripts are aberrantly overexpressed in mouse epithelial tumors as
well as different subtypes of satellite repeats in human carcinomas210.
Interestingly, we showed that, after an initial down-regulation of major
satellite transcripts, their levels are partially recovered 24 hours after
Our results also show that Snail1 depletion causes alterations in
heterochromatin organization in mesenchymal cells, and that this
associates with increased chromosomal alterations. Moreover, we also
show for the first time a link between major satellite transcription and
EMT, a process that has been widely associated to metastasis and tumor
progression. These results also suggest that major satellite RNAs may be
involved in heterochromatin reorganization during the transition. It has
been proposed that organization and clustering of pericentromeric
heterochromatin sequences in normal cells may provide anchor points
for genomic organization since nuclear matrix/scaffold proteins have
uncovered a preferential association with satellite and other repeat-rich
heterochromatin may reflect or promote instability of overall nuclear
organization, which in turn may affect the epigenetic landscape and gene
expression pattern genome wide. Thus, elucidating the precise role that
these major satellite transcripts have in the cell, how do they contribute
to nuclear organization, as well as how do they become deregulated in
cancer cells may open a new door in the understanding of tumor
formation and progression.
TGFβ treatment. Indeed, preliminary data (not shown) indicate that
mesenchymal cell lines tend to express higher levels of major transcripts
compared to epithelial cell lines. Thus, rather than establishing a
mesenchymal state with lower major satellite expression, the initial
down-regulation of major satellite may be necessary to allow
pericentromeric domains reorganization during early EMT. After this
reorganization, major satellite transcripts tend to be up-regulated again.
Indeed, based on the differences in major satellite levels observed in
mesenchymal and epithelial cell lines, it seems plausible that major
satellite transcripts continue to be up-regulated later on during EMT, and
that the resulting mesenchymal cells express higher levels compared
with the epithelial cells from which they are derived. This could partially
explain the up-regulation of major satellite transcripts observed in
epithelial tumors.
Although classically it was thought that heterochromatin was
permanently silenced and repressed, nowadays it is unquestionable that
it is indeed transcribed and that its transcriptional regulation is essential
for cell homeostasis maintenance. Heterochromatin epigenetic features
and structure have been extensively characterized. However, many
questions about this process remain unanswered, such as the
identification of specific key regulators and the mechanisms through
which they regulate heterochromatin transcription.
Moreover, Snail1 is one of the main inducers of EMT, a process by which
epithelial cells change completely their morphology and properties and
acquire a mesenchymal phenotype211. This may be accompanied by
genome reorganization, since different cell types seem to be
characterized by distinct higher-order chromatin structure212. In fact, it
has been previously suggested that an epigenetic reprogramming occurs
during EMT148. Accordingly, we propose that a general chromatinreorganization is required for a complete EMT transition and that Snail1
and LOXL2 play a key role by repressing pericentromeric transcription.
Epithelial cells do not express Snail1 and have a particular clustering of
pericentromeric regions into chromocenters, with active transcription of
major satellites. Upon TGFβ treatment, Snail1 protein levels are upregulated and stabilized. In the early steps of EMT, epithelial cells go
through an intermediate state in which Snail1 binds to pericentromeric
regions and recruits LOXL2. Together, they participate in major satellite
repression by oxidation of H3 in these domains. Additionally, H3K4me3
deamination and major satellite repression cause transient release of
heterochromatin domains and allows chromatin movement and
reorganization. After 24 hours, when cells already have spindle
In this thesis we show that Snail1 transcription factor is one of these key
regulators, and that it is involved in the repression of major satellite
transcription through the co-repressor LOXL2. In this way, Snail1 binds
major satellite sequences at pericentromeric regions together with
LOXL2, which deaminases H3K4me3 oxidizing H3 and establishing a
repressive mark that leads to major satellite repression.
morphology and express mesenchymal genes, HP1α has been reloaded
to chromocenters and major satellite transcription has been recovered.
Therefore, for the conversion of an epithelial cell to a mesenchymal cell
during TGFβ induced EMT, there is a window of time during which major
satellite transcription is down-regulated due to the actions of Snail1 and
LOXL2 through the oxidation of H3 and HP1α is released from
heterochromatin. Altogether, this may allow chromatin reorganization
and acquisition of mesenchymal traits (Figure CR1).
However, future work on advanced genomic technologies such as Hi-C
will be required to confirm and further study the three-dimensional
genome reorganization that may take place during EMT.
Figure CR1. Model for pericentromeric heterochromatin reorganization during
EMT. Upon TGFβ induction of EMT, Snail1 is rapidly up-regulated, binds to
pericentromeric regions and recruits LOXL2 to oxidize H3 and repress major
satellite transcription. As consequence, HP1α is released from heterochromatin,
enabling chromatin reorganization and acquisition of mesenchymal traits. This
effect is transient and, after 24 hours, major satellite levels and HP1α binding to
heterochromatin have been recovered.
Cell Lines:
All cells were grown and maintained in Dulbecco’s modified Eagle’s
medium (Invitrogen) supplemented with 100 units/mL penicillin,
100µg/mL streptomycin, 2 mM glutamine, and 10% FBS (Invitrogen) at
37⁰C in 5% CO2.
The cell lines used during this study were the following:
SW620: human colon adenocarcinoma cells with epithelial
MiaPaCa-2: human pancreas carcinoma derived cells. They grow as
attached epithelial cells with floating rounded cells.
HEK293T: human embryonic kidney cells, which are very easy to
grow and transfect.
HEK293T Phoenix: second-generation retrovirus producer cell line. It
is derived from HEK293T but stably expresses gag-pol and envelope
proteins for amphotropic viruses (Phoenix-ampho cells).
NMuMG: mouse epithelial mammary gland cells. They grow forming
colonies and undergo EMT when treated with TGFβ.
NIH-3T3: mouse embryonic fibroblasts cell line, with mesenchymal
MEF: mouse embryonic fibroblasts, derived in this case from control
and conditional Snail1 knockout mice. They have mesenchymal
NMuMG cells were also supplemented with insulin (10 µg/mL). A final
concentration of 1-5 ng/ml TGFβ was used to stimulate NMuMG cells and
induce EMT, and cells were collected after 0, 2, 8, and 24 hours.
Primary MEFs (pMEFs) were obtained from conditional knock-out mice,
resulting from crossing mice from The Jackson Laboratory in which Snail1
is floxed and Cre-recombinase expressing mice (Snai1F/F/Cre- and
Snai1F/F/Cre-). Snai1 deletion was induced by at least three doses of
tamoxifen treatment to a final concentration of 1 μM.
Immortalized MEFs (iMEFs) were obtained following the 3T3 protocol213.
MEFs were transfected with the pcDNA3 empty vector (CT iMEFs) or
pcDNA3-CreGFP to induce Snai1 deletion (KO iMEFs) and selected with
0.5 µg/mL G418.
Transfection procedures:
MEFs were transfected with pBabe-mSnail1–HA and pBabe-empty
vectors using lipofectamine 2000 reagent (Invitrogen), following
manufacturer’s instructions, and stably selected with puromycin (1μ/mL).
For coimmunoprecipitation experiments, Snail1 and HP1α vectors were
transfected into HEK293T cells by adding, drop-wise, a mixture of NaCl,
the specified DNA and polyetherimide polymer (PEI), previously
incubated for 15min at room temperature. Cells were lysed 48 hours
To down-regulate Snail1 in NMuMG cells, mmSnail siRNA (QIAGEN,
Target sequence: 5’-CACCTGTTTCACAGCAGTTTA-3’) was reverse
transfected in NMuMG cells (96-well plates) at a final concentration of 30
nM siRNA using a 3.5% solution of HiPerFect transfection reagent
(QIAGEN). The Cy3-labelled siGLO cyclophilin B siRNA (Dharmacon) was
used to monitor transfection efficiency, and the AllStars scrambled siRNA
and siRNA targeting GFP (QIAGEN) served as negative controls. Media
was replaced 10 hours after transfection, and TGFβ (5 ng/ml) was added
24 hours later. Cells were fixed 8 hours after TGFβ addition with PFA 4%
and then were stained by immunofluorescence.
Alternatively, Snail1 was also down-regulated in NMuMG using siRNAs to
obtain protein and RNA at the different time point after TGFβ treatment
previously indicated, following the same procedure as specified below
for HP1α silencing.
To down-regulate HP1α in NMuMG cells, mouse CBX5 siRNA (Dharmacon
L- 040799-01) was transfected in NMuMG cells at a final concentration of
5 nM siRNA with RNAiMax transfection reagent (Invitrogen) and OptiMEM transfection media (Invitrogen) following manufacturer’s
instructions. siGENOME Non-Targting siRNA#3 (Dharmacon D-00121003-20) was used as control. Media was replaced 24h later and cells were
treated with TGFβ overnight.
To down-regulate major satellite transcripts, NMuMG cells were
transfected with LNA-DNA gapmers using RNAiMax transfection reagent
(Invitrogen) and Opti-MEM transfection media (Invitrogen) following
manufacturer’s instructions. TGFβ was added at the time of transfection
and RNA was collected after 24h.
Gapmers’ sequences used were the following:
• Maj1 Gapmer: acatTCCACTTGACCGActtg
• Maj2 Gapmer: tattTCACGTCCTAAagtg
• Control Gapmer: tatctGCATACGATACggtt
Retroviral and lentiviral Infection:
For retroviral infections, HEK293 Phoenix cells were transfected (day 0)
using Lipofectamine 2000 reagent (Invitrogen) with the indicated vectors.
The transfection medium was replaced with fresh medium after 24 hours
(day 1), and the cell-conditioned medium at day 2 was filtered and used
to infect target cells with 8 μg/ml polybrene. HEK293T Phoenix cells were
incubated with fresh medium for a further 24 hours and on day 3, a
second infection with the conditioned medium and polybrene was
NMuMG cells and MEFs were infected with lentiviral vectors containing
shLOXL2 (TRCN0000076711-12) or shControl.
NMuMG cells were infected with retroviruses using pBabe-MajorSat (see
cloning procedures) and pBabe-empty vectors whereas MEFs were
infected with retrovirus using pCMV-LOXL2 or pCMV-empty vectors. Both
cell types were selected with puromycin 1 μg/ml after infection.
For lentiviral infection, HEK293T cells were used to produce viral
particles. Cells were grown to 90% confluence and then transfected (day
0) by adding, drop-wise, a mixture of NaCl, DNA (50% of indicated vector,
10% pCMV-VSVG, 30% pMDLg/Prre and 10% pRSV rev) and PEI that had
been pre-incubated for 15 min at room temperature. The transfection
medium was replaced with fresh medium after 24 hours (day 1), and the
cell-conditioned medium at day 2 was filtered and used to infect target
cells with 8 μg/ml polybrene. HEK293 cells were incubated with fresh
medium for further 24 hours, and a second infection with the
conditioned medium and polybrene was performed on day 3.
Cloning procedures:
Generation of pBabe-mSnail1-HA, pcDNA3-hLOXL2wt/mut-Flag and
pCMV-LOXL2 was previously described30. pcDNA3-Snail-HA vector was
generated as previously specified111, as well as pcDNA3 vectors
containing Snail1 mutants SA, SD and P2A123,129. Plasmids containing
human HP1α WT and mutants157 were kindly provided by Albert Jordan.
pBabe-MajorSat plasmids were generated by subcloning PCR4Maj-9-2
plasmid214 into a pBabe-empty vector after EcoRI digestion. Two plasmids
were generated, one with the insert orientated in the sense direction
and the other in the antisense. In major satellite overexpression
experiments, both vectors were co-infected.
The following antibodies were used: α-H3K9me3 (Millipore, 17-625), αH3K4me3 (Millipore 04-745), α-HP1α (H2164, Sigma), α-LOXL2 (Abcam
ab60753; ab55470), α-H3 (Abcam ab1791), α-Snail1215, α-Tubulin (T9026,
Sigma), α-Pyruvate kinase (Chemicon ab 1235), α-Ecadherin
(Transduction Labs 610182), α-N-cadherin (Abcam ab76011), α-PCNA
(Abcam ab2426; Santa Cruz sc-25280), α-Fibronectin (DAKO A0245).
Alternatively, fluorescence images corresponding to DAPI, Snail1, Ecadherin, HP1α and H3K9me3 were acquired for each condition (10 fields
per well) with the InCell 2000 automated epifluorescence microscope
(GE) at 40X magnification. Images were analysed using In Cell
investigator Software (GE) by first defining the cell nuclei with DAPI
staining. The nuclei were then segmented using top-hat segmentation
defining a minimum nucleus area of 100 μm2. To define the expression of
Snail1 protein by cell, the average intensity of pixels in the reference
channel (Alexa488) within the defined nuclear region was measured.
HP1α foci within the nuclei (Alexa594) were segmented using multiscale
top-hat to define a granule size of 1–2 μm. Once each cell was assigned a
nuclear intensity for the specific expression of Snail1 and a number for
HP1α foci, a threshold filter defining positive and negative expressing
cells was set. Threshold filter uses a histogram for data visualization. The
cut-off of the filters was set as follows: Snail1 expression was measured
in knockout MEFs or NMuMG untreated cells to define the negative
population. Positive controls (e.g. TGFβ-treated cells) were analysed to
define the highly-expressing population. The opposite criterion was
followed to specify the filter settings for HP1α foci. Once the cut-off was
set up, the analysis was carried out. The program thus assigns a
Cells were washed and fixed with 4% PFA for 15 min at room
temperature (RT). For PCNA staining, soluble fraction of nucleus was
removed prior to fixation with CSK-T buffer (CSK + 0.5% Triton) for 5’
followed by 3 washes with CSK buffer (20 mM HEPES [pH 7.8], 100 mM
NaCl, 3 mM MgCl2, 300 mM sucrose and proteinase inhibitors). After
fixation, cells were blocked for 1 hour with 1% PBS-BSA and then
incubated at RT for 2 hours with primary antibody, followed by 3 washes
with PBS and 1 hour incubation at RT with the secondary antibody. Cells
were washed again three times with PBS and incubated for 5 min with
DAPI (0.25 µg/mL) to stain cell nuclei before mounting them with
fluoromount. Fluorescence images were capture with a LEICA confocal
microscope and DAPI, K9me3 and HP1α foci were counted manually.
definition of positive or negative to each cell and generates a percentage
of both cell populations per well.
Micrococcal digestion:
HEK293T cells were cotransfected (10:1) with pcDNA3-hLOXL2wt/mutFlag or an empty pcDNA3 and pEGF-C1. Cells expressing high levels of
GFP were sorted by a fluorescence-activated cell sorter 36 hours after
transfection. Cell pellets (1.5x106 cells per condition) were lysed in 500 µL
of Buffer A (10mM Tris [pH 7.4], 10 mM NaCl, 3 mM MgCl2 and 0.3 M
sacarose supplemented with protease and phosphatase inhibitors) for 10
min at 4⁰C. After NP-40 addition to a final 0.2% (v/v) and 10 min of
incubation at 4⁰C, the lysate was centrifuged for 10 min at 1200 rpm at
4⁰C. The resulting pellet was resuspended in 100 µL of Buffer A (CaCl2
was added to a final concentration of 10 mM) and digestion with MNase
(0.08 units) was carried out for 5 or 10 min. Enzyme was inactivated with
50 mM EDTA. After treatment with Rnase A (2 min at RT) and Proteinase
K (10 min at 56⁰C), DNA was purified (GFXTM PCR DNA Purification Kit, GE
Healthcare). Digestion products were analyzed by agarose gel
electrophoresis. Total DNA was visualized by SYBR® safe staining.
Cultured cells were treated with colcemid for 2 hours at 37⁰C and
centrifuged 10 min at 2000 rpm after trypsinisation. Pellets were resuspended in 7 mL of 67 mM KCL and incubated at 37⁰C for 5-10 min.
After addition of 2 mL of CARNOY solution (methanol: acetic acid 3:1),
cells were centrifuged 10 min at 2000 rpm. Pellets were further resuspended in 10 mL of CARNOY solution and incubated 1 hour at -20⁰C.
After centrifugation and two more washes with CARNOY solution,
methaphase spreads were performed by dropping CARNOY re-suspended
pellets onto the slides and drying at 45⁰C, followed by overnight
incubation at 65⁰C. Bands were stained with Wright colorant after two
washes with 2x SSC at 65⁰C, and methaphase alterations were analyzed.
Salt Extraction Experiments:
Salt extraction experiments were performed as described previously216.
Briefly, cells were washed in PBS, dounced in buffer A (0.32 M sucrose,
15 mM HEPES [pH 7.9], 60 mM KC, 2 mM EDTA, 0.5 mM EGTA, 0.5% BSA,
0.5 mM spermidine, 0.15 mMspermine, and 0.5mMDTT), layered over a
cushion of high-sucrose Buffer A (30% sucrose), and centrifuged (15 min,
3000 g). Pelleted nuclei were resuspended in buffer B (15 mM HEPES [pH
7.9], 60 mM KCl, 15 mM NaCl, 0.34 mM sucrose, 10% glycerol) and
incubated with different 500 mM NaCl at 4⁰C for 30 min. Supernatants
were analysed by Western blot.
For subcellular fractionation, cells were lysed in buffer A (10 mM HEPES,
[pH 7.8], 10 mM KCl, 1.5m M MgCl2, and 0.5 mM DTT) supplemented
with protease and phosphatase inhibitors for 10 min at 4⁰C. One-third
volume of 10% Triton X-100 was added, and samples were mixed for 30
sec. Samples were then centrifuged at 15000 g at 4⁰C for 1 min.
Supernatant was collected, and 0.11 volume of buffer B (0.3 M HEPES
[pH 7.8], 1.4 M KCl, and 30 mM MgCl2) was added. After 30 min rotation
at 4⁰C, samples were centrifuged for 15 min at 15000 g at 4⁰C, and
supernatant was collected again (cytoplasmic fraction). Pelleted nuclei
were resuspended in one-fifth the original volume in buffer C (20 mM
HEPES [pH 7.8], 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA,
and 0.5 mM DTT) and rotated for 30 min at 4⁰C. After 15 min
centrifugation at 15000 g at 4⁰C, supernatant was collected (soluble
nucleus) and separated from the pellet (chromatin fraction), which was
resuspended in Laemmli buffer. Histone H3 and α-tubulin were used as
chromatin and cytoplasmic markers, respectively.
Subcellular Fractionation:
Chromatin Immunoprecipitation (ChIP) experiments:
Cells were crosslinked in 1% formaldehyde for 10 min at 37⁰C.
Crosslinking was stopped by adding glycine to a final concentration of
0.125 M for 2 min at room temperature. Cell monolayers were scraped in
cold Soft Lysis Buffer (50 mM Tris [pH 8.0], 10 mM EDTA, 0.1 % NP-40,
and 10% glycerol) and centrifuges 15 min at 3000 rpm. Nuclei pellets
were lysed with SDS Lysis Buffer (1% SDS, 10 mM EDTA, and 50 mM Tris,
pH 8.0) and cell extracts were sonicated. Sonication was performed tentwelve times at 40% for 10 seconds (Branson) to generate 200- to 500-bp
DNA fragments. After 20 min of incubation on ice, sonicated extracts
were centrifuged 10 min at 13000 rpm.
Supernatants were diluted 1:10 with Dilution Buffer (0.01% SDS, 1.1%
Triton X-100, 1.2 mM EDTA, 16.7 mM Tris [pH 8], 167 mM NaCl), and
immunoprecipitation was done by rotation overnight at 4⁰C with primary
antibody or irrelevant IgGs. Immunocomplexes were captured by
incubation with proteinG/A magtenic beads, previously blocked with 1
mg/mL salmon sperm (Ambion), for 1 hour at 4⁰C. Immunocomplexes
were then sequentially washed at 4⁰C with Low Salt Buffer (0.1 % SDS, 1
% Triton X-100, 2 mM EDTA, 20 mM Tris [pH 8.0], 150 mM NaCl), High
Salt Buffer (0.1 % SDS, 1 % Triton X-100, 2 mM EDTA, 20 mM Tris [pH
8.0], 500 mM NaCl) and LiCl buffer (250 mM LiCl, 1 % NP-40, 1 % NaDOC,
1 mM EDTA, 10 mM Tris [pH 8.0]), before eluting them with Elution
Buffer (1 % SDS, 100 mM Na2CO3).
Crosslinking was reverted by addition of 200mM NaCl and overnight
incubation at 65⁰C. Finally, samples were treated for 1 hour at 55⁰C with
Proteinase K solution (0.4 mg/mL proteinase K (Roche), 50mM EDTA and
200mM Tris [pH 6.5]). DNA was purified with GFX kit from GE Healthcare
and eluted in MilliQ water. The amount of immunoprecipitated DNA was
analised by qRT-PCR and ChIP results were quantified in relation to the
input amount as described217.
To detect H3 oxidation in specific promoters, ChIP assays were
performed as described above, but extracts were lysed in presence of
biotin hydrazide reagent (5 mM) and samples were incubated at 25⁰C for
2 hours. Immunoprecipitation was carried out with α-histone H3
antibody. Precipitates were re-extracted with SDS Lysis Buffer and reimmunoprecipitated with streptavidin-magnetic beads for 30 min at 4⁰C.
Samples were then treated with Elution Buffer and incubated at 65⁰C to
reverse formaldehyde crosslinking. Results were quantified in relation to
the input and the total amount of H3 immunoprecipitated in each
The primers used in ChIP experiments were the following:
Oligonucleotide sequence
For ChIP-Seq analysis, the Solexa library was prepared as recommended
by the manufacturer (www.illumina.com). The size selection of this
library was performed by gel electrophoresis, with subsequent excision
and purification of DNA in the 150–300 bp range. The positional
resolution of ChIP-Seq was improved by reducing the size, and narrowing
the size range of DNA collected from gel purification218. DNA sequencing
of each sample was performed by the Solexa/Illumina protocol. Solexa
ChIP and control reads were analyzed jointly to identify regions with an
over-representation of reads in the ChIP sample versus the control
sample. Reads obtained from the samples were aligned to the human
reference genome using the GEM mapper, allowing up to 3 mismatches.
Reads that were unambiguously mapped were selected, and site
Primer name
Mapping to mouse sequences
Major (pair I)
Major (pair II)
Mapping to human sequences
Chr 4
Chr 7
Chr 20
identification of short sequence reads (SISSRs), a peak-finding algorithm,
was used to identify significantly enriched genomic locations219.
Candidates for enriched regions were identified as aggregations of five or
more ChIP reads that were neither separated by more than 100 bp nor
present in the control sample. It was also determined whether flanking
genes and promoter regions were associated with these peaks in the
human genome (assembly March.2006).
Western Blot:
Cells were lysed in SDS buffer (2% SDS, 50mM Tris [pH 8.0]) and protein
was quantified by Lowry. Western blots were performed according to
standard procedures. Briefly, protein was loaded in sodium dodecyl
sulphate polyacrylamide gel electrophoresis (SDS-PAGE) at different
percentages and gels were run in TGS buffer. Proteins were transferred
to a nitrocellulose membrane in Transfer Buffer and blocked with 5%
non-fat milk in TBS-T buffer (25 mM Tris-HCl [pH 7.5], 137 mM NaCl, 0.1%
Tween20) for 1 hour. Primary antibody was added to fresh blocking
solution and incubated overnight at 4⁰C. After three washes of 10 min
with TBS-T, membranes were incubated for 1h at RT with secondary
antibodies peroxidase-combined (HRP). After a new round of washes,
membranes were developed by incubation with substrate for HRP
Enhanced ChemiLuminiscence (ECL) and exposure to autoradiographic
Co-immunoprecipitation assays:
MiaPaCa-2, SW-620 or NIH-3T3 cells were lysed in RIPA buffer (50 mM
Tris-HCl [pH 7.3], 150 mM NaCl, 1 mM EDTA, 1 % NP-40, 0.25 % sodium
deoxycholate) for 20 minutes on ice. Samples were centrifuged at 3000
rpm for 15 min and the supernatant was collected and pre-cleared with
proteinA-agarose beads (Roche) blocked with 1 % BSA, for 3 h at 4 ⁰C.
Immunoprecipitation with α-HP1α polyclonal antibodies or irrelevant
rabbit immunoglobulin were carried out overnight at 4⁰C. ProteinAagarose beads were added for 1 h at 4⁰C. Immunocomplexes were
washed 4 times with the RIPA buffer and then re-suspended in Laemmli
buffer. Proteins were resolved in SDS-PAGE and analysed with α-Snail1
and α-HP1 antibodies by Western blot. After transfection of HEK-293T
cells, a similar co-immunoprecipitation protocol was used to characterize
the interaction. Immunoprecipitations were carried out with anti-Flag M2
Affinity Gel (Sigma), and anti-Flag and anti-HA antibodies were used in
the Western Blot.
Genomic DNA extraction:
Genomic DNA was extracted from cultured cells using the GenElute
Mammalian Genomic DNA Miniprep kit (Sigma-Aldrich), and was eluted
in MilliQ water. Afterwards, qRT-PCR was performed to quantify the
relative amount of major and minor DNA compared to α-globin gene, as
RNA extraction procedures:
Genomic DNA digestion was performed with DNase Turbo (Ambion) at
37⁰C in this way: a total of 10 µg of RNA per condition was digested with
2.5 µL of enzyme for 1.5 hours, followed by 1.5 hours further digestion
after adding 1.5 µL of extra enzyme. RNAs were purified using RNeasy
Minikit (Qiagen) and a second round of DNase digestions was performed
as previously described. After re-purification with the Qiagen kit, RNAs
were eluted in DEPC water.
For RNA extraction, 800 µL of TRIzol® reagent (Invitrogen) was used to
lyse cultured cells. After addition of 200 µL of chloroform, samples were
centrifuged for 20 min at 4⁰C at top speed, and the upper aqueous phase
was recovered. RNA was precipitated by the addition of 500 µL of
isopropanol. After at least 10 min of incubation at RT, samples were
centrifuged and pellets washed in ethanol 75% before re-suspending
them in DEPC water.
Real-time RT-PCR:
RNA was reverse transcribed using Transcription First Strand cDNA
Synthesis Kit (Roche) and either oligo-dT or random hexamer primers.
Quantitative determination of RNA levels was performed in triplicate, by
real-time quantitative PCR (Roche LightCycler), using the LightCycler 480
SYBR Green Kit from Roche. In all cases, values were normalized to the
expression of housekeeping genes (HPRT or Pumilio). For the
quantification of major satellite transcripts levels, an additional RT
reaction was performed as negative control, in which no retrotranscriptase was added, to confirm the absence of genomic DNA
The following primers were used to quantify RNA expression levels:
Primer name
Mouse RNA amplification
Major (pair I)
Major (pair II)
Oligonucleotide sequence
Human RNA amplification
Both mouse and human RNA amplification
We measured gene-expression levels of NMuMG-Control and NMuMGMajor cells. For microarrays analysis, amplification, labeling and
hybridizations were performed according to protocols from Ambion WT
Expression Kit (Ambion). Samples were labeled using the WT Terminal
Labeling Kit (Affymetrix), and then hybridized to GeneChip Mouse Gene
1.0 ST Array (Affymetrix) in a GeneChip Hybridization Oven 640. Washing
and scanning were performed using the Hybridization Wash and Stain Kit
and the GeneChip System of Affymetrix (GeneChip Fluidics Station 450
and GeneChip Scanner 3000 7G). After quality control of raw data, they
were background corrected, quantile-normalized and summarized to a
gene-level using the robust multi-chip average (RMA) (Irizarry 2003)
obtaining a total of 28822 transcript clusters, excluding controls, which
roughly correspond to genes. Linear Models for Microarray (LIMMA)
(Smyth 2004), a moderated t-statistics model, was used for detecting
differentially expressed genes between the conditions. Correction for
multiple comparisons was performed using false discovery rate. Genes
with an adjusted p-value less than 0.05 and with an absolute fold change
(FC) value above 1.5 were selected as significant. Ingenuity Pathway
Microarray gene expression analysis:
Analysis v 9.0 (Ingenuity® Systems, www.ingenuity.com) was used to
functionally analyze the results.
Accession Numbers
The GEO accession number for the data referred to this thesis is
Migration and Invasion assays:
For the migration experiments, NMuMG-Control and NMuMG-Major
cells were treated with TGFβ. After 24 hours, 50.000 cells were resuspended in DMEM 0.2%FBS-0.1%BSA, reseeded on a transwell filter
chamber (Costar 3422) and incubated for 6-8 hours. For invasion assays,
cells were placed in matrigel-coated transwell filter (BD356234) and
incubated for 24 hours. In both cases DMED 10%FBS was placed in the
lower chamber and used as chemoattractant. Non-migrating and noninvading cells were removed from the upper surface of the membrane,
while cells that adhered to the lower surface were fixed with PFA4% for
20 min and nuclei stained with DAPI. The DAPI-stained nuclei were
counted in four fields per filter by ImageJ software.
Kornberg, R. D. & Lorch, Y. Twenty-Five Years of the Nucleosome,
Fundamental Particle of the Eukaryote Chromosome. Cell 98,
285–294 (1999).
Khorasanizadeh, S. The nucleosome: from genomic organization
to genomic regulation. Cell 116, 259–72 (2004).
Szerlong, H. J. & Hansen, J. C. Nucleosome distribution and linker
DNA: connecting nuclear function to dynamic chromatin
structure. Biochem. Cell Biol. 89, 24–34 (2011).
Spadafora, C., Bellard, M., Lee Compton, J. & Chambon, P. The
DNA repeat lengths in chromatins from sea urchin sperm and
gastrula cells are markedly different. FEBS Lett. 69, 281–285
Olins, A. L. & Olins, D. E. Spheroid chromatin units (v bodies).
Science 183, 330–2 (1974).
Hübner, M. R., Eckersley-Maslin, M. A. & Spector, D. L. Chromatin
organization and transcriptional regulation. Curr. Opin. Genet.
Dev. 23, 89–95 (2013).
Maeshima, K., Hihara, S. & Eltsov, M. Chromatin structure: does
the 30-nm fibre exist in vivo? Curr. Opin. Cell Biol. 22, 291–7
Lieberman-Aiden, E. et al. Comprehensive mapping of long-range
interactions reveals folding principles of the human genome.
Science 326, 289–93 (2009).
Grewal, S. I. S. & Elgin, S. C. R. Transcription and RNA interference
in the formation of heterochromatin. Nature 447, 399–406
Cremer, T. & Cremer, M. Chromosome territories. Cold Spring
Harb. Perspect. Biol. 2, a003889 (2010).
Dekker, J., Rippe, K., Dekker, M. & Kleckner, N. Capturing
chromosome conformation. Science 295, 1306–11 (2002).
Dekker, J. et al. Hi–C: A comprehensive technique to capture the
conformation of genomes. Methods 58, 268–276 (2012).
Dixon, J. R. et al. Topological domains in mammalian genomes
identified by analysis of chromatin interactions. Nature 485, 376–
80 (2012).
Hübner, M. R., Eckersley-Maslin, M. a & Spector, D. L. Chromatin
organization and transcriptional regulation. Curr. Opin. Genet.
Dev. 23, 89–95 (2013).
Bickmore, W. A. & van Steensel, B. Genome architecture: domain
organization of interphase chromosomes. Cell 152, 1270–84
Nora, E. P. et al. Spatial partitioning of the regulatory landscape of
the X-inactivation centre. Nature 485, 381–5 (2012).
Schwartz, Y. B. et al. Alternative epigenetic chromatin states of
polycomb target genes. PLoS Genet. 6, e1000805 (2010).
Wen, B., Wu, H., Shinkai, Y., Irizarry, R. A. & Feinberg, A. P. Large
histone H3 lysine 9 dimethylated chromatin blocks distinguish
differentiated from embryonic stem cells. Nat. Genet. 41, 246–50
Guelen, L. et al. Domain organization of human chromosomes
revealed by mapping of nuclear lamina interactions. Nature 453,
948–51 (2008).
Peric-Hupkes, D. et al. Molecular maps of the reorganization of
genome-nuclear lamina interactions during differentiation. Mol.
Cell 38, 603–13 (2010).
Németh, A. et al. Initial genomics of the human nucleolus. PLoS
Genet. 6, e1000889 (2010).
Van Koningsbruggen, S. et al. High-resolution whole-genome
sequencing reveals that specific chromatin domains from most
human chromosomes associate with nucleoli. Mol. Biol. Cell 21,
3735–48 (2010).
Osborne, C. S. et al. Active genes dynamically colocalize to shared
sites of ongoing transcription. Nat. Genet. 36, 1065–71 (2004).
Chakalova, L. & Fraser, P. Organization of transcription. Cold
Spring Harb. Perspect. Biol. 2, a000729 (2010).
Van Bortle, K. & Corces, V. G. Nuclear organization and genome
function. Annu. Rev. Cell Dev. Biol. 28, 163–87 (2012).
Bonasio, R., Tu, S. & Reinberg, D. Molecular signals of epigenetic
states. Science 330, 612–6 (2010).
Black, J. C. & Whetstine, J. R. Chromatin landscape: methylation
beyond transcription. Epigenetics 6, 9–15 (2011).
Tan, M. et al. Identification of 67 histone marks and histone lysine
crotonylation as a new type of histone modification. Cell 146,
1016–28 (2011).
Day, J. J. & Sweatt, J. D. Epigenetic treatments for cognitive
impairments. Neuropsychopharmacology 37, 247–60 (2012).
Herranz, N. et al. Lysyl oxidase-like 2 deaminates lysine 4 in
histone H3. Mol. Cell 46, 369–76 (2012).
Portela, A. & Esteller, M. Epigenetic modifications and human
disease. Nat. Biotechnol. 28, 1057–68 (2010).
Margueron, R., Trojer, P. & Reinberg, D. The key to development:
interpreting the histone code? Curr. Opin. Genet. Dev. 15, 163–76
Ruthenburg, A. J., Li, H., Patel, D. J. & Allis, C. D. Multivalent
engagement of chromatin modifications by linked binding
modules. Nat. Rev. Mol. Cell Biol. 8, 983–94 (2007).
Yun, M., Wu, J., Workman, J. L. & Li, B. Readers of histone
modifications. Cell Res. 21, 564–78 (2011).
Zeng, L. & Zhou, M. M. Bromodomain: an acetyl-lysine binding
domain. FEBS Lett. 513, 124–8 (2002).
Tajul-Arifin, K., Teasdale, R., Ravasi, T., Hume, D. A. & Mattick, J. S.
Identification and analysis of chromodomain-containing proteins
encoded in the mouse transcriptome. Genome Res. 13, 1416–29
Kim, J. K., Samaranayake, M. & Pradhan, S. Epigenetic
mechanisms in mammals. Cell. Mol. Life Sci. 66, 596–612 (2009).
Cedar, H. & Bergman, Y. Linking DNA methylation and histone
modification: patterns and paradigms. Nat. Rev. Genet. 10, 295–
304 (2009).
Kafri, T. et al. Developmental pattern of gene-specific DNA
methylation in the mouse embryo and germ line. Genes Dev. 6,
705–714 (1992).
Monk, M., Boubelik, M. & Lehnert, S. Temporal and regional
changes in DNA methylation in the embryonic, extraembryonic
and germ cell lineages during mouse embryo development.
Development 99, 371–82 (1987).
Ooi, S. K. T. et al. DNMT3L connects unmethylated lysine 4 of
histone H3 to de novo methylation of DNA. Nature 448, 714–7
Chen, Z. & Riggs, A. D. DNA methylation and demethylation in
mammals. J. Biol. Chem. 286, 18347–53 (2011).
Grewal, S. I. S. & Jia, S. Heterochromatin revisited. Nat. Rev.
Genet. 8, 35–46 (2007).
Trojer, P. & Reinberg, D. Facultative Heterochromatin: Is There a
Distinctive Molecular Signature? Mol. Cell 28, 1–13 (2007).
Maison, C. & Almouzni, G. HP1 and the dynamics of
heterochromatin maintenance. Nat. Rev. Mol. Cell Biol. 5, 296–
304 (2004).
Lomberk, G., Wallrath, L. & Urrutia, R. The Heterochromatin
Protein 1 family. Genome Biol. 7, 228 (2006).
Minc, E., Courvalin, J. C. & Buendia, B. HP1gamma associates with
euchromatin and heterochromatin in mammalian nuclei and
chromosomes. Cytogenet. Cell Genet. 90, 279–84 (2000).
Bannister, a J. et al. Selective recognition of methylated lysine 9
on histone H3 by the HP1 chromo domain. Nature 410, 120–4
Lachner, M., O’Carroll, D., Rea, S., Mechtler, K. & Jenuwein, T.
Methylation of histone H3 lysine 9 creates a binding site for HP1
proteins. Nature 410, 116–20 (2001).
Nielsen, P. R. et al. Structure of the HP1 chromodomain bound to
histone H3 methylated at lysine 9. Nature 416, 103–7 (2002).
Jacobs, S. A. & Khorasanizadeh, S. Structure of HP1
chromodomain bound to a lysine 9-methylated histone H3 tail.
Science 295, 2080–3 (2002).
Muchardt, C. et al. Coordinated methyl and RNA binding is
required for heterochromatin localization of mammalian
HP1alpha. EMBO Rep. 3, 975–81 (2002).
Sugimoto, K., Yamada, T., Muro, Y. & Himeno, M. Human homolog
of Drosophila heterochromatin-associated protein 1 (HP1) is a
DNA-binding protein which possesses a DNA-binding motif with
weak similarity to that of human centromere protein C (CENP-C).
J. Biochem. 120, 153–9 (1996).
Meehan, R. R., Kao, C.-F. & Pennings, S. HP1 binding to native
chromatin in vitro is determined by the hinge region and not by
the chromodomain. EMBO J. 22, 3164–74 (2003).
Cowieson, N. P., Partridge, J. F., Allshire, R. C. & McLaughlin, P. J.
Dimerisation of a chromo shadow domain and distinctions from
the chromodomain as revealed by structural analysis. Curr. Biol.
10, 517–25 (2000).
Rea, S. et al. Regulation of chromatin structure by site-specific
histone H3 methyltransferases. Nature 406, 593–9 (2000).
Aagaard, L. et al. Functional mammalian homologues of the
Drosophila PEV-modifier Su(var)3-9 encode centromereassociated proteins which complex with the heterochromatin
component M31. EMBO J. 18, 1923–38 (1999).
Yamamoto, K. & Sonoda, M. Self-interaction of heterochromatin
protein 1 is required for direct binding to histone
methyltransferase, SUV39H1. Biochem. Biophys. Res. Commun.
301, 287–292 (2003).
Fuks, F., Hurd, P. J., Deplus, R. & Kouzarides, T. The DNA
methyltransferases associate with HP1 and the SUV39H1 histone
methyltransferase. Nucleic Acids Res. 31, 2305–12 (2003).
Fanti, L., Giovinazzo, G., Berloco, M. & Pimpinelli, S. The
heterochromatin protein 1 prevents telomere fusions in
Drosophila. Mol. Cell 2, 527–38 (1998).
Lomberk, G., Bensi, D., Fernandez-Zapico, M. E. & Urrutia, R.
Evidence for the existence of an HP1-mediated subcode within
the histone code. Nat. Cell Biol. 8, 407–15 (2006).
LeRoy, G. et al. Heterochromatin protein 1 is extensively
decorated with histone code-like post-translational modifications.
Mol. Cell. Proteomics 8, 2432–42 (2009).
Sullivan, B. a, Blower, M. D. & Karpen, G. H. Determining
centromere identity: cyclical stories and forking paths. Nat. Rev.
Genet. 2, 584–96 (2001).
Eymery, A., Callanan, M. & Vourc’h, C. The secret message of
heterochromatin: new insights into the mechanisms and function
of centromeric and pericentric repeat sequence transcription. Int.
J. Dev. Biol. 53, 259–68 (2009).
Allshire, R. C., Nimmo, E. R., Ekwall, K., Javerzat, J. P. & Cranston,
G. Mutations derepressing silent centromeric domains in fission
yeast disrupt chromosome segregation. Genes Dev. 9, 218–33
Kellum, R. & Alberts, B. M. Heterochromatin protein 1 is required
for correct chromosome segregation in Drosophila embryos. J.
Cell Sci. 108 ( Pt 4, 1419–31 (1995).
Peters, a H. et al. Loss of the Suv39h histone methyltransferases
impairs mammalian heterochromatin and genome stability. Cell
107, 323–37 (2001).
Wong, A. K. & Rattner, J. B. Sequence organization and cytological
localization of the minor satellite of mouse. Nucleic Acids Res. 16,
11645–61 (1988).
Prosser, J., Frommer, M., Paul, C. & Vincent, P. C. Sequence
relationships of three human satellite DNAs. J. Mol. Biol. 187,
145–55 (1986).
Maison, C. et al. SUMOylation promotes de novo targeting of
HP1α to pericentric heterochromatin. Nat. Genet. 43, 220–7
Cristina, M. et al. Original article Organization of complex nuclear
domains in somatic mouse cells. 55–65 (1980).
Mayer, R. et al. Common themes and cell type specific variations
of higher order chromatin arrangements in the mouse. BMC Cell
Biol. 6, 44 (2005).
Jenuwein, T. & Allis, C. D. Translating the histone code. Science
293, 1074–80 (2001).
Cam, H. P., Chen, E. S. & Grewal, S. I. S. Transcriptional scaffolds
for heterochromatin assembly. Cell 136, 610–4 (2009).
Chen, E. S. et al. Cell cycle control of centromeric repeat
transcription and heterochromatin assembly. Nature 451, 734–7
Grewal, S. I. RNAi-dependent formation of heterochromatin and
its diverse functions. Curr. Opin. Genet. Dev. 20, 134–41 (2010).
Jolly, C., Morimoto, R., Robert-Nicoud, M. & Vourc’h, C. HSF1
transcription factor concentrates in nuclear foci during heat
shock: relationship with transcription sites. J. Cell Sci. 110 ( Pt 2,
2935–41 (1997).
Denegri, M. et al. Human chromosomes 9, 12, and 15 contain the
nucleation sites of stress-induced nuclear bodies. Mol. Biol. Cell
13, 2069–79 (2002).
Valgardsdottir, R. et al. Transcription of Satellite III non-coding
RNAs is a general stress response in human cells. Nucleic Acids
Res. 36, 423–34 (2008).
Probst, A. V et al. A strand-specific burst in transcription of
pericentric satellites is required for chromocenter formation and
early mouse development. Dev. Cell 19, 625–38 (2010).
Rudert, F., Bronner, S., Garnier, J.-W. & Doll�, P. Transcripts from
opposite strands of ? satellite DNA are differentially expressed
during mouse development. Mamm. Genome 6, 76–83 (1995).
Jehan, Z. et al. Novel noncoding RNA from human Y distal
heterochromatic block (Yq12) generates testis-specific chimeric
CDC2L2. Genome Res. 17, 433–40 (2007).
Kishi, Y., Kondo, S. & Gotoh, Y. Transcriptional activation of mouse
major satellite regions during neuronal differentiation. Cell Struct.
Funct. 37, 101–10 (2012).
Terranova, R., Sauer, S., Merkenschlager, M. & Fisher, A. G. The
reorganisation of constitutive heterochromatin in differentiating
muscle requires HDAC activity. Exp. Cell Res. 310, 344–56 (2005).
Martens, J. H. A. et al. The profile of repeat-associated histone
lysine methylation states in the mouse epigenome. EMBO J. 24,
800–12 (2005).
Lu, J. & Gilbert, D. M. Proliferation-dependent and cell cycle
regulated transcription of mouse pericentric heterochromatin. J.
Cell Biol. 179, 411–21 (2007).
Jolly, C. et al. Stress-induced transcription of satellite III repeats. J.
Cell Biol. 164, 25–33 (2004).
Bulut-Karslioglu, A. et al. A transcription factor-based mechanism
for mouse heterochromatin formation. Nat. Struct. Mol. Biol. 19,
1023–30 (2012).
Hanahan, D., Weinberg, R. A. & Francisco, S. The Hallmarks of
Cancer Review University of California at San Francisco. 100, 57–
70 (2000).
Ichim, C. V & Wells, R. A. First among equals: the cancer cell
hierarchy. Leuk. Lymphoma 47, 2017–27 (2006).
Christiansen, J. J. & Rajasekaran, A. K. Reassessing epithelial to
mesenchymal transition as a prerequisite for carcinoma invasion
and metastasis. Cancer Res. 66, 8319–26 (2006).
Campbell, L. L. & Polyak, K. Breast Tumor Heterogeneity. 6, 2332–
2338 (2007).
Chaffer, C. L. & Weinberg, R. a. A perspective on cancer cell
metastasis. Science 331, 1559–64 (2011).
Nieto, M. A. The ins and outs of the epithelial to mesenchymal
transition in health and disease. Annu. Rev. Cell Dev. Biol. 27, 347–
76 (2011).
Thiery, J. P. & Sleeman, J. P. Complex networks orchestrate
epithelial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol. 7,
131–42 (2006).
Chaffer, C. L. et al. Mesenchymal-to-epithelial transition facilitates
bladder cancer metastasis: role of fibroblast growth factor
receptor-2. Cancer Res. 66, 11271–8 (2006).
Van Meer, G. & Simons, K. The function of tight junctions in
maintaining differences in lipid composition between the apical
and the basolateral cell surface domains of MDCK cells. EMBO J. 5,
1455–64 (1986).
Anderson, J. M. & Van Itallie, C. M. Tight junctions and the
molecular basis for regulation of paracellular permeability. Am. J.
Physiol. 269, G467–75 (1995).
Li, S., Guan, J.-L. & Chien, S. Biochemistry and biomechanics of cell
motility. Annu. Rev. Biomed. Eng. 7, 105–50 (2005).
Egeblad, M. & Werb, Z. New functions for the matrix
metalloproteinases in cancer progression. Nat. Rev. Cancer 2,
161–74 (2002).
LaGamba, D., Nawshad, A. & Hay, E. D. Microarray analysis of
gene expression during epithelial-mesenchymal transformation.
Dev. Dyn. 234, 132–42 (2005).
Venkov, C. et al. Transcriptional networks in epithelialmesenchymal transition. PLoS One 6, e25354 (2011).
Thiery, J. P., Acloque, H., Huang, R. Y. J. & Nieto, M. A. Epithelialmesenchymal transitions in development and disease. Cell 139,
871–90 (2009).
Yang, J. & Weinberg, R. a. Epithelial-mesenchymal transition: at
the crossroads of development and tumor metastasis. Dev. Cell
14, 818–29 (2008).
Schäfer, M. & Werner, S. Cancer as an overhealing wound: an old
hypothesis revisited. Nat. Rev. Mol. Cell Biol. 9, 628–38 (2008).
Lee, J. M., Dedhar, S., Kalluri, R. & Thompson, E. W. The epithelialmesenchymal transition: new insights in signaling, development,
and disease. J. Cell Biol. 172, 973–81 (2006).
Thompson, E. W. & Haviv, I. The social aspects of EMT-MET
plasticity. Nat. Med. 17, 1048–9 (2011).
Zoltan-Jones, A., Huang, L., Ghatak, S. & Toole, B. P. Elevated
hyaluronan production induces mesenchymal and transformed
properties in epithelial cells. J. Biol. Chem. 278, 45801–10 (2003).
Xu, J., Lamouille, S. & Derynck, R. TGF-beta-induced epithelial to
mesenchymal transition. Cell Res. 19, 156–72 (2009).
Kriegl, L. et al. LEF-1 and TCF4 expression correlate inversely with
survival in colorectal cancer. J. Transl. Med. 8, 123 (2010).
Batlle, E. et al. The transcription factor snail is a repressor of Ecadherin gene expression in epithelial tumour cells. Nat. Cell Biol.
2, 84–9 (2000).
Cano, A. et al. The transcription factor snail controls epithelialmesenchymal transitions by repressing E-cadherin expression.
Nat. Cell Biol. 2, 76–83 (2000).
Hajra, K. M., Chen, D. Y.-S. & Fearon, E. R. The SLUG zinc-finger
protein represses E-cadherin in breast cancer. Cancer Res. 62,
1613–8 (2002).
Eger, A. et al. DeltaEF1 is a transcriptional repressor of E-cadherin
and regulates epithelial plasticity in breast cancer cells. Oncogene
24, 2375–85 (2005).
Comijn, J. et al. The two-handed E box binding zinc finger protein
SIP1 downregulates E-cadherin and induces invasion. Mol. Cell 7,
1267–78 (2001).
Perez-Moreno, M. A. et al. A new role for E12/E47 in the
repression of E-cadherin expression and epithelial-mesenchymal
transitions. J. Biol. Chem. 276, 27424–31 (2001).
Wang, X. et al. Krüppel-like factor 8 induces epithelial to
mesenchymal transition and epithelial cell invasion. Cancer Res.
67, 7184–93 (2007).
Yang, J. et al. Twist, a master regulator of morphogenesis, plays
an essential role in tumor metastasis. Cell 117, 927–39 (2004).
Hartwell, K. A. et al. The Spemann organizer gene, Goosecoid,
promotes tumor metastasis. Proc. Natl. Acad. Sci. U. S. A. 103,
18969–74 (2006).
Sobrado, V. R. et al. The class I bHLH factors E2-2A and E2-2B
regulate EMT. J. Cell Sci. 122, 1014–24 (2009).
Mani, S. A. et al. Mesenchyme Forkhead 1 (FOXC2) plays a key
role in metastasis and is associated with aggressive basal-like
breast cancers. Proc. Natl. Acad. Sci. U. S. A. 104, 10069–74
Christofori, G. Snail1 links transcriptional control with epigenetic
regulation. EMBO J. 29, 1787–9 (2010).
Peiró, S. et al. Snail1 transcriptional repressor binds to its own
promoter and controls its expression. Nucleic Acids Res. 34, 2077–
84 (2006).
Stanisavljevic, J., Porta-de-la-Riva, M., Batlle, R., de Herreros, A. G.
& Baulida, J. The p65 subunit of NF-κB and PARP1 assist Snail1 in
activating fibronectin transcription. J. Cell Sci. 124, 4161–71
Nieto, M. A. The snail superfamily of zinc-finger transcription
factors. Nat. Rev. Mol. Cell Biol. 3, 155–66 (2002).
Chiang, C. & Ayyanathan, K. Snail/Gfi-1 (SNAG) family zinc finger
proteins in transcription regulation, chromatin dynamics, cell
signaling, development, and disease. Cytokine Growth Factor Rev.
24, 123–31 (2013).
Chiang, C. & Ayyanathan, K. Characterization of the E-box binding
affinity to SNAG-zinc finger proteins. Mol. Biol. 46, 809–816
Peinado, H., Olmeda, D. & Cano, A. Snail, Zeb and bHLH factors in
tumour progression: an alliance against the epithelial phenotype?
Nat. Rev. Cancer 7, 415–28 (2007).
Domínguez, D. et al. Phosphorylation Regulates the Subcellular
Location and Activity of the Snail Transcriptional Repressor
Phosphorylation Regulates the Subcellular Location and Activity of
Zhou, B. P. et al. Dual regulation of Snail by GSK-3beta-mediated
phosphorylation in control of epithelial-mesenchymal transition.
Nat. Cell Biol. 6, 931–40 (2004).
Yang, Z. et al. Pak1 phosphorylation of snail, a master regulator of
epithelial-to-mesenchyme transition, modulates snail’s subcellular
localization and functions. Cancer Res. 65, 3179–84 (2005).
Viñas-Castells, R. et al. The hypoxia-controlled FBXL14 ubiquitin
ligase targets SNAIL1 for proteasome degradation. J. Biol. Chem.
285, 3794–805 (2010).
Peinado, H., Ballestar, E., Esteller, M. & Cano, A. Snail mediates Ecadherin repression by the recruitment of the Sin3A/histone
deacetylase 1 (HDAC1)/HDAC2 complex. Mol. Cell. Biol. 24, 306–
19 (2004).
Herranz, N. et al. Polycomb complex 2 is required for E-cadherin
repression by the Snail1 transcription factor. Mol. Cell. Biol. 28,
4772–81 (2008).
Lin, Y. et al. The SNAG domain of Snail1 functions as a molecular
hook for recruiting lysine-specific demethylase 1. EMBO J. 29,
1803–16 (2010).
Molnar, J. et al. Structural and functional diversity of lysyl oxidase
and the LOX-like proteins. Biochim. Biophys. Acta - Proteins
Proteomics 1647, 220–224 (2003).
Nishioka, T., Eustace, A. & West, C. Lysyl oxidase: from basic
science to future cancer treatment. Cell Struct. Funct. 37, 75–80
Kim, M. S. et al. Expression and purification of enzymatically
active forms of the human lysyl oxidase-like protein 4. J. Biol.
Chem. 278, 52071–4 (2003).
Cano, A., Santamaría, P. G. & Moreno-Bueno, G. LOXL2 in
epithelial cell plasticity and tumor progression. Future Oncol. 8,
1095–108 (2012).
Baker, A.-M. et al. The role of lysyl oxidase in SRC-dependent
proliferation and metastasis of colorectal cancer. J. Natl. Cancer
Inst. 103, 407–24 (2011).
Saad, F. A., Torres, M., Wang, H. & Graham, L. Intracellular lysyl
oxidase: effect of a specific inhibitor on nuclear mass in
proliferating cells. Biochem. Biophys. Res. Commun. 396, 944–9
Pischon, N. et al. Lysyl oxidase (lox) gene deficiency affects
osteoblastic phenotype. Calcif. Tissue Int. 85, 119–26 (2009).
Brekhman, V. & Neufeld, G. A novel asymmetric 3D in-vitro assay
for the study of tumor cell invasion. BMC Cancer 9, 415 (2009).
Schietke, R. et al. The lysyl oxidases LOX and LOXL2 are necessary
and sufficient to repress E-cadherin in hypoxia: insights into
cellular transformation processes mediated by HIF-1. J. Biol.
Chem. 285, 6658–69 (2010).
Voloshenyuk, T. G., Landesman, E. S., Khoutorova, E., Hart, A. D. &
Gardner, J. D. Induction of cardiac fibroblast lysyl oxidase by TGFβ1 requires PI3K/Akt, Smad3, and MAPK signaling. Cytokine 55,
90–7 (2011).
Peinado, H. et al. A molecular role for lysyl oxidase-like 2 enzyme
in snail regulation and tumor progression. EMBO J. 24, 3446–58
Barker, H. E. et al. LOXL2-mediated matrix remodeling in
metastasis and mammary gland involution. Cancer Res. 71, 1561–
72 (2011).
McDonald, O. G., Wu, H., Timp, W., Doi, A. & Feinberg, A. P.
Genome-scale epigenetic reprogramming during epithelial-tomesenchymal transition. Nat. Struct. Mol. Biol. 18, 867–74 (2011).
Almouzni, G. & Probst, A. V. Heterochromatin maintenance and
establishment: lessons from the mouse pericentromere. Nucleus
2, 332–8 (2011).
Carver, E. A., Jiang, R., Lan, Y. U., Oram, K. F. & Gridley, T. The
Mouse Snail Gene Encodes a Key Regulator of the EpithelialMesenchymal Transition. 21, 8184–8188 (2001).
Batlle, R. et al. Snail1 controls TGF-β responsiveness and
differentiation of mesenchymal stem cells. Oncogene 32, 3381–9
Li, Q. & Zhang, Z. Linking DNA replication to heterochromatin
silencing and epigenetic inheritance. Acta Biochim. Biophys. Sin.
(Shanghai). 44, 3–13 (2012).
Li, F., Martienssen, R. & Cande, W. Z. Coordination of DNA
replication and histone modification by the Rik1-Dos2 complex.
Nature 475, 244–8 (2011).
Quivy, J.-P. et al. A CAF-1 dependent pool of HP1 during
heterochromatin duplication. EMBO J. 23, 3516–26 (2004).
Quivy, J.-P., Gérard, A., Cook, A. J. L., Roche, D. & Almouzni, G. The
HP1–p150/CAF-1 interaction is required for pericentric
heterochromatin replication and S-phase progression in mouse
cells. Nat. Struct. Mol. Biol. 15, 972–979 (2008).
Lin, T., Ponn, a, Hu, X., Law, B. K. & Lu, J. Requirement of the
histone demethylase LSD1 in Snai1-mediated transcriptional
repression during epithelial-mesenchymal transition. Oncogene
29, 4896–904 (2010).
Lechner, M. S., Schultz, D. C., Negorev, D., Maul, G. G. & Rauscher,
F. J. The mammalian heterochromatin protein 1 binds diverse
nuclear proteins through a common motif that targets the
chromoshadow domain. Biochem. Biophys. Res. Commun. 331,
929–37 (2005).
Miettinen, P. J., Ebner, R., Lopez, A. R. & Derynck, R. TGF-beta
induced transdifferentiation of mammary epithelial cells to
mesenchymal cells: involvement of type I receptors. J. Cell Biol.
127, 2021–36 (1994).
Cheutin, T. et al. Maintenance of stable heterochromatin domains
by dynamic HP1 binding. Science 299, 721–5 (2003).
Lamouille, S. & Derynck, R. Cell size and invasion in TGF-betainduced epithelial to mesenchymal transition is regulated by
activation of the mTOR pathway. J. Cell Biol. 178, 437–51 (2007).
Enukashvily, N. I., Donev, R., Waisertreiger, I. S.-R. & Podgornaya,
O. I. Human chromosome 1 satellite 3 DNA is decondensed,
demethylated and transcribed in senescent cells and in A431
epithelial carcinoma cells. Cytogenet. Genome Res. 118, 42–54
Kaufman, P. D., Das, C. & Tyler, J. K. Histone exchange and histone
modifications during transcription and aging. Biochim. Biophys.
Acta - Gene Regul. Mech. 1819, 332–342 (2012).
Santos-Rosa, H. et al. Histone H3 tail clipping regulates gene
expression. Nat. Struct. Mol. Biol. 16, 17–22 (2009).
Batista, P. J. & Chang, H. Y. Long noncoding RNAs: cellular address
codes in development and disease. Cell 152, 1298–307 (2013).
Casa, V. & Gabellini, D. A repetitive elements perspective in
Polycomb epigenetics. Front. Genet. 3, 199 (2012).
Leeb, M. et al. Polycomb complexes act redundantly to repress
genomic repeats and genes. Genes Dev. 24, 265–76 (2010).
Saurin, A. J. et al. The human polycomb group complex associates
with pericentromeric heterochromatin to form a novel nuclear
domain. J. Cell Biol. 142, 887–98 (1998).
Peters, A. H. F. M. et al. Partitioning and plasticity of repressive
histone methylation states in mammalian chromatin. Mol. Cell 12,
1577–89 (2003).
Puschendorf, M. et al. PRC1 and Suv39h specify parental
asymmetry at constitutive heterochromatin in early mouse
embryos. Nat. Genet. 40, 411–20 (2008).
Shi, Y. et al. Histone demethylation mediated by the nuclear
amine oxidase homolog LSD1. Cell 119, 941–53 (2004).
Karytinos, A. et al. A novel mammalian flavin-dependent histone
demethylase. J. Biol. Chem. 284, 17775–82 (2009).
Rutenberg, M. The characterization of lysine specific demethylase
2 (LSD2), a novel histone demethylase. 2, (2007).
Lan, F. et al. S. pombe LSD1 homologs regulate heterochromatin
propagation and euchromatic gene transcription. Mol. Cell 26, 89–
101 (2007).
Vandewalle, C. et al. SIP1/ZEB2 induces EMT by repressing genes
of different epithelial cell-cell junctions. Nucleic Acids Res. 33,
6566–78 (2005).
Shi, Y. et al. Coordinated histone modifications mediated by a
CtBP co-repressor complex. Nature 422, 735–8 (2003).
Román, A. C. et al. Dioxin receptor and SLUG transcription factors
regulate the insulator activity of B1 SINE retrotransposons via an
RNA polymerase switch. Genome Res. 21, 422–32 (2011).
Beltran, M. et al. A natural antisense transcript regulates
Zeb2/Sip1 gene expression during Snail1-induced epithelialmesenchymal transition. Genes Dev. 22, 756–69 (2008).
Hakim, O. et al. Spatial congregation of STAT binding directs
selective nuclear architecture during T-cell functional
differentiation. Genome Res. 23, 462–72 (2013).
Kosak, S. T. et al. Subnuclear compartmentalization of
immunoglobulin loci during lymphocyte development. Science
296, 158–62 (2002).
Shopland, L. S., Johnson, C. V, Byron, M., McNeil, J. & Lawrence, J.
B. Clustering of multiple specific genes and gene-rich R-bands
around SC-35 domains: evidence for local euchromatic
neighborhoods. J. Cell Biol. 162, 981–90 (2003).
Lamond, A. I. & Spector, D. L. Nuclear speckles: a model for
nuclear organelles. Nat. Rev. Mol. Cell Biol. 4, 605–12 (2003).
Hall, L. L., Smith, K. P., Byron, M. & Lawrence, J. B. Molecular
anatomy of a speckle. Anat. Rec. A. Discov. Mol. Cell. Evol. Biol.
288, 664–75 (2006).
Meaburn, K. J. & Misteli, T. Cell biology: chromosome territories.
Nature 445, 379–781 (2007).
Hiragami-Hamada, K. et al. N-terminal phosphorylation of
HP1{alpha} promotes its chromatin binding. Mol. Cell. Biol. 31,
1186–200 (2011).
Zou, J. et al. Protein kinase CK2α is overexpressed in colorectal
cancer and modulates cell proliferation and invasion via
regulating EMT-related genes. J. Transl. Med. 9, 97 (2011).
Kim, J. & Hwan Kim, S. CK2 inhibitor CX-4945 blocks TGF-β1induced epithelial-to-mesenchymal transition in A549 human lung
adenocarcinoma cells. PLoS One 8, e74342 (2013).
Hirota, T., Lipp, J. J., Toh, B.-H. & Peters, J.-M. Histone H3 serine
10 phosphorylation by Aurora B causes HP1 dissociation from
heterochromatin. Nature 438, 1176–80 (2005).
Crosio, C. et al. Mitotic phosphorylation of histone H3: spatiotemporal regulation by mammalian Aurora kinases. Mol. Cell. Biol.
22, 874–85 (2002).
Fu, J., Bian, M., Jiang, Q. & Zhang, C. Roles of Aurora kinases in
mitosis and tumorigenesis. Mol. Cancer Res. 5, 1–10 (2007).
Bischoff, J. R. et al. A homologue of Drosophila aurora kinase is
oncogenic and amplified in human colorectal cancers. EMBO J. 17,
3052–65 (1998).
Zhou, H. et al. Tumour amplified kinase STK15/BTAK induces
centrosome amplification, aneuploidy and transformation. Nat.
Genet. 20, 189–93 (1998).
Wan, X.-B. et al. Inhibition of Aurora-A suppresses epithelialmesenchymal transition and invasion by downregulating MAPK in
nasopharyngeal carcinoma cells. Carcinogenesis 29, 1930–7
Lee, J., Choi, J.-H. & Joo, C.-K. TGF-β1 regulates cell fate during
epithelial-mesenchymal transition by upregulating survivin. Cell
Death Dis. 4, e714 (2013).
Kinney, C. M. et al. Histone H3 as a novel substrate for MAP
kinase phosphatase-1. Am. J. Physiol. Cell Physiol. 296, C242–9
Papait, R. et al. Np95 is implicated in pericentromeric
heterochromatin replication and in major satellite silencing. Mol.
Biol. Cell 18, 1098–106 (2007).
Hansen, K. H. et al. A model for transmission of the H3K27me3
epigenetic mark. Nat. Cell Biol. 10, 1291–300 (2008).
Helbling Chadwick, L., Chadwick, B. P., Jaye, D. L. & Wade, P. A.
The Mi-2/NuRD complex associates with pericentromeric
heterochromatin during S phase in rapidly proliferating lymphoid
cells. Chromosoma 118, 445–57 (2009).
Price, B. D. & D’Andrea, A. D. Chromatin remodeling at DNA
double-strand breaks. Cell 152, 1344–54 (2013).
Chiolo, I. et al. Double-strand breaks in heterochromatin move
outside of a dynamic HP1a domain to complete recombinational
repair. Cell 144, 732–44 (2011).
Goodarzi, A. A. et al. ATM signaling facilitates repair of DNA
double-strand breaks associated with heterochromatin. Mol. Cell
31, 167–77 (2008).
Goodarzi, A. A., Kurka, T. & Jeggo, P. A. KAP-1 phosphorylation
regulates CHD3 nucleosome remodeling during the DNA doublestrand break response. Nat. Struct. Mol. Biol. 18, 831–9 (2011).
Ayoub, N., Jeyasekharan, A. D., Bernal, J. A. & Venkitaraman, A. R.
HP1-beta mobilization promotes chromatin changes that initiate
the DNA damage response. Nature 453, 682–6 (2008).
Baldeyron, C., Soria, G., Roche, D., Cook, A. J. L. & Almouzni, G.
HP1alpha recruitment to DNA damage by p150CAF-1 promotes
homologous recombination repair. J. Cell Biol. 193, 81–95 (2011).
Luijsterburg, M. S. et al. Heterochromatin protein 1 is recruited to
various types of DNA damage. J. Cell Biol. 185, 577–86 (2009).
Carone, D. M. & Lawrence, J. B. Heterochromatin instability in
cancer: from the Barr body to satellites and the nuclear periphery.
Semin. Cancer Biol. 23, 99–108 (2013).
Pageau, G. J., Hall, L. L., Ganesan, S., Livingston, D. M. & Lawrence,
J. B. The disappearing Barr body in breast and ovarian cancers.
Nat. Rev. Cancer 7, 628–33 (2007).
Hansen, K. D. et al. Increased methylation variation in epigenetic
domains across cancer types. Nat. Genet. 43, 768–75 (2011).
Reddy, K. L. & Feinberg, A. P. Higher order chromatin organization
in cancer. Semin. Cancer Biol. 23, 109–15 (2013).
Zhu, Q. et al. BRCA1 tumour suppression occurs via
heterochromatin-mediated silencing. Nature 477, 179–84 (2011).
Ting, D. T. et al. Aberrant overexpression of satellite repeats in
pancreatic and other epithelial cancers. Science 331, 593–6
Lim, J. & Thiery, J. P. Epithelial-mesenchymal transitions: insights
from development. Development 139, 3471–86 (2012).
Meshorer, E. & Misteli, T. Chromatin in pluripotent embryonic
stem cells and differentiation. Nat. Rev. Mol. Cell Biol. 7, 540–6
Nilausen, K. & Green, H. Reversible arrest of growth in G1 of an
established fibroblast line (3T3). Exp. Cell Res. 40, 166–8 (1965).
Lehnertz, B. et al. Suv39h -Mediated Histone H3 Lysine 9
Methylation Directs DNA Methylation to Major Satellite Repeats
at Pericentric Heterochromatin. 13, 1192–1200 (2003).
Francí, C. et al. Expression of Snail protein in tumor-stroma
interface. Oncogene 25, 5134–44 (2006).
Meshorer, E. et al. Hyperdynamic plasticity of chromatin proteins
in pluripotent embryonic stem cells. Dev. Cell 10, 105–16 (2006).
Frank, S. R., Schroeder, M., Fernandez, P., Taubert, S. & Amati, B.
Binding of c-Myc to chromatin mediates mitogen-induced
acetylation of histone H4 and gene activation. Genes Dev. 15,
2069–82 (2001).
Johnson, D. S., Mortazavi, A., Myers, R. M. & Wold, B. Genomewide mapping of in vivo protein-DNA interactions. Science 316,
1497–502 (2007).
Jothi, R., Cuddapah, S., Barski, A., Cui, K. & Zhao, K. Genome-wide
identification of in vivo protein-DNA binding sites from ChIP-Seq
data. Nucleic Acids Res. 36, 5221–31 (2008).
Loyola, A. et al. The HP1alpha-CAF1-SetDB1-containing complex
provides H3K9me1 for Suv39-mediated K9me3 in pericentric
heterochromatin. EMBO Rep. 10, 769–75 (2009).
Most of the work presented in this thesis is published in:
Millanes-Romero A., Herranz N., Perrera V., Iturbide A., LoubatCasanovas J., Gil J., Jenuwein T., García de Herreros A., Peiró S.
Regulation of Heterochromatin Transcription by Snail1/LOXL2 during
Epithelial-to-Mesenchymal Transition. Molecular Cell (2013).
Independently of the work published in this thesis, the author worked in
other projects as reflected:
Herranz N., Dave N., Millanes-Romero A., Morey L., Díaz V.M., LórenzFonfría V., Gutierrez-Gallego R., Jerónimo C., Di Croce L., García de
Herreros A., Peiró S. Lysyl oxidase-like 2 deaminates lysine 4 in histone
H3. Molecular Cell (2012).
Resulta curiós mirar enrere i recordar l’inici d’aquesta petita aventura.
Suposo que tot va anar així gràcies a Lídia Bardia, que em va acabar
d’empentar, amb les seves bones referències, a triar el grup que més
m’agradava per fer les pràctiques de final de carrera.
Encara recordo la primera entrevista amb tu Antonio, quan em vas oferir
la possibilitat de treballar en un projecte in vivo, que em va cridar molt
l’atenció, o ajudar la Sandra i el Niko (qui?) en un projecte d’epi... què?
Afortunadament, quan vaig començar al lab, la colònia de ratolins KO
encara no estava llesta, així que em va tocar esbrinar el significat de la
paraula epigenètica. Em va caldre menys d’un mes per saber que volia
seguir descobrint aquest món misteriós, que fins al dia d’avui encara em
segueix fascinant.
El trajecte recorregut durant aquests 5 anys ha estat molt enriquidor, tan
a nivell professional com a nivell personal. I és que han passat un munt
de coses durant tot aquest temps! Cadascun de vosaltres n’heu estat
partícips i m’heu aportat petits i grans moments que mai oblidaré.
En primer lloc, volia donar-te les gràcies a tu, Sandra. Per tot el que
m’has ensenyat sobre ciència i per sempre estar disponible i
predisposada a ajudar-me. Però també per totes les bones coses que
m’has aportat durant aquests anys... la teva contagiosa motivació i ànsies
de coneixement, el teu constant positivisme i la manera de sempre veure
un resultat positiu per desastrós que hagi estat l’experiment, la teva
capacitat per arriscar-te i llençar-te de cap a la piscina o la teva confiança
en les meves idees i suggeriments. Però sobretot, valoro moltíssim la
infinitat d’hores que hem passat discutint i divagant sense pressa sobre
ciència o treballant colze a colze durant els períodes de revisions. M’ho
he passat molt bé treballant amb tu al lab i també liant-la fora! No
tothom té l’oportunitat de veure la seva “jefa” cantant a l’antikaraoke
després de 15 dies d’estar treballant per ella...
Una de les millors coses de treballar en equip és que pots aprendre
quelcom de totes les persones que t’envolten, sobretot d’aquelles que
tenen més experiència. De l’Antonio, amb els teus Journals i les seves
introduccions magistrals, les teves correccions de l’anglès i, sobretot, els
teus comentaris, idees i aportacions als seminaris o, en comptades però
enriquidores ocasions, al teu despatx. Gràcies també per les barbacoes i
les partides de Trivial a Mieres! O el Jepi i el Víctor, que sempre heu
intentat resoldre tots els dubtes que us he formulat, i en el teu cas en
particular Víctor, m’has ensenyat i refrescat un munt de coses sobre
bioquímica! M’ho vaig passar genial fent d’ajudant de pràctiques i
sempre recordaré amb molt d’amor la teva manera de compensar els
incomptables informes que vaig haver de corregir. I tampoc m’oblidaré
dels teus “bailoteos” Jepi...
Per descomptat, qui més m’ha ensenyat en aquest laboratori has estat tu
Niko. Em vas intentar transmetre tot el que sabies, incloent els teus
“truquillos” i les teves paranoies i em sento privilegiada d’haver estat la
teva “mindu” i haver compartit tot el que hem compartit, tant a nivell
professional com a nivell personal. I és que m’has aportat tantes coses...
els teus consells sobre la vida, les teves observacions i la teva manera de
veure les coses m’han ajudat enormement en moments clau, així que
Però el pilar indispensable que ha vetllat per mantenir-me en peu davant
qualsevol situació durant aquests anys has estat tu Estel. Qui ens ho
hagués dit quan vam començar la carrera que acabaríem compartint
tant! Gràcies per mil i una coses, sobretot per escoltar-me sempre, per
les teves reconfortants abraçades, els teus riures contagiosos, els teus
comentaris carregats d’ànims i positivisme i les tardes de “Vull cap de
setmana” i Shakira al lab! Una tesi aporta moltes coses bones però
també hi ha moments durs, gràcies per endolcir els meus!
Y muchas gracias especialmente a ti Lorena. Quién habría dicho que esa
chica un poco borde y reservada y yo tendríamos tantas coses en común
y llegaríamos a ser tan amigas! Gracias por tus consejos, tu
compañerismo, tu actitud protectora hacia las personas que quieres, tu
compromiso y fidelidad, por aportar un poco de racionalidad en los
momentos oportunos y por tener un punto de vista tan similar al mío en
tantas cosas. Al fin y al cabo las dos somos virgo, qué se le va a hacer…
Pero sobretodo, gracias por llenar de tan buenos momentos estos dos
últimos años, por comprenderme tan bien y por hacerme sentir alguien
tan especial.
Em sento molt afortunada d’haver començat el doctorat en un laboratori
on tothom estava disposat a ajudar-me amb els meus dubtes i incerteses,
però també a fer pinya i passar-s’ho bé. Mai oblidaré les nits de Bitàcora,
Magic, Tres Flores i totes les aventures i records associats. Gràcies a tots!
A la Montse, per ensenyar-me bones pràctiques al lab; a la Patri, per
aportar un punt crític als meus experiments i fer-me partícip de moments
molt especials que recordo amb molta estima; al Manolo, per haver
rescatat els meus RNAs una oblidada vesprada de ja fa molt temps, així
com per les infinites hores compartides ballant tango, parlant sobre
literatura, filosofant sobre la vida o passejant per Barcelona; a Susana,
por tu predisposición a ayudar incluso trabajando en otro lab; a la
Natàlia, per aportat tanta vitalitat i dinamisme al laboratori; al David
Casagolda, per tenir una resposta per a tot; a l’Ali, per ser tan feliç i
contagiar aquesta felicitat a la gent que t’envolta, i també per fer-me
riure tant; a la Clara, per iniciar-me en el món dels ratolins i l’estabulari; a
Raul, por estar siempre dispuesto a buscarme o pedirme cualquier
reactivo, ayudarme con la burocracia y también a decidirme en mis
momentos de máxima indecisión; a Jelena, por tu sarcasmo, tus
sugerencias científicas y por ayudarme a mejorar mi inglés; a la Rosa, per
donar-me un cop de mà en un nombre indefinit de moments, per tots els
favors que m’has fet i pels cafès compartits; a l’Àlex, pel teu peculiar
sentit de l’humor i per la teva actitud “don’t worry, be happy”; i a la
Jordina, per tot el temps que has invertit en ajudar-me a sobreviure a
l’estabulari, però sobretot per la teva confiança en mi, la nostra
complicitat i totes les copes de vi compartides que han anat teixint, a poc
a poc però de forma robusta, aquest vincle tan especial que ara ens
I per descomptat, a tots els nous companys i companyes que s’han
incorporat o que han passat pel laboratori durant aquests anys.
A la Núria, pels teus somriures, la teva energia i empenta, els teus
moments de “Uh! Oh! No tinc por!” i per estar sempre disposada a
resoldre qualsevol consulta mèdica. I naturalment, pels moments viscuts
amb la petita Estel i el Guillem, gràcies per compartir-los amb mi.
Y a Rocco, por tu espontaneidad y transparencia, por tu buen humor, tu
vitalidad, por las clases de swing, el rissoto y por aportar un agradable
soplo de aire fresco.
I als nous membres del Chromatin Team. A Ane, por tu sutil toque de
humor vasco, por traer queso al lab después de las visitas a tu tierra y por
tus recomendaciones culinarias; al Joan Pau, pel teu bon humor, per
deixar-te abraçar quan necessito “mimitos”, per la teva predisposició a
fer favors i pel teu interès en compartir la ciència i les teves idees; i a la
Laura, per la teva motivació i positivisme, que fa que resulti gratificant i
engrescador invertir el temps en ensenyar a algú que comença.
I a la Jessica, pel curt però agradable temps en què vam compartir poiata
i pels caramels! A la Sílvia, pels teus comentaris als “Chromatin
meetings”; al Josu, per ser tan divertit i esbojarrat però a la vegada tan
sweet i proper; a Fani, por alegrar mis días, por llenarme de energía, por
las cenas en tu casa y los “bailoteos” dentro y fuera del lab, pero
sobretodo por ser alguien tan especial; a l’Alba Azagra, per ser tan dolça i
encantadora; al Pere, per involucrar-te en tot malgrat la teva curta
estada; i a la Cristina, per la teva vitalitat i per transmetre i infondre tan
bones vibracions.
Sembla mentida tota la gent que s’arriba a conèixer al llarg d’una tesi, i
encara resulta més sorprenent que n’hi hagi tantes que t’acabin aportant
alguna cosa positiva. I es que fora del lab també hi ha molta gent a qui
voldria agrair que hagin fet el meu dia a dia una mica més agradable,
regalant somriures pels passadissos, prestant-me reactius o material de
laboratori en moments crítics, amenitzant les llargues jornades a cultius,
compartint dinars, jugant a vòlei, xerrant a les Beer Sessions, facilitantme protocols o simplement per les vostres mostres d’afecte.
Neus, un plaer haver compartit tot aquest temps amb tu, la teva
predisposició a ajudar tothom, la teva dolçor i il·lusió, el teu detallisme i
la teva manera perfecta de fer les coses. Gràcies també Conchi i Silvia,
per les converses, les vostres recomanacions tan pràctiques i tots els
riures compartits. I al Joan, per fer-me companyia al lab a hores
intempestives durant les revisions i per ser tan encantador; i a la Mireia,
per salvar-me la vida a l’estabulari quan no hi era la Jordina.
Gràcies també a la Tània, la Laura, l’Héctor, el Xavier i la Piedad pels
petits favors. I al David i l’Elena, per les converses a l’hora de dinar, y en
concreto a ti Elena, gracias por el apoyo y la hospitalidad recibida en mis
visitas a Londres. A la Mercè, per la complicitat que vam anar generant
poc a poc, a la Marta Garrido, per ajudar-me amb els blocs de parafina; la
Laura Ortet i la Marta Moreno, pels riures compartits; a Alex, por
prestarme anticuerpos, i al Marc, la Lara i la Kathi per tots els bons
moments viscuts a la platja jugant a vòlei. Gracias también Dani y Raul
por formar parte del equipo este último año.
I per descomptat al Biga’s Lab. Erika, que sepas que aunque últimamente
no pasemos mucho tiempo juntas aun trabajando una al lado de la otra,
encontrarte por los pasillos y por la galería siempre me reconforta, sobre
todo cuando el encuentro va acompañado de uno de tus cálidos abrazos.
Júlia, merci per dedicar el teu temps a ajudar-me sempre que t’ho he
demanat. I gràcies també a la resta de membres del lab: Leonor, Cris,
Pol, Vero, Vane, Jordi, Teresa, Jessi, Eva i Roshani per compartir el dia a
dia i pels moments de “cachondeo” viscuts fora del lab.
I de fora del Programa de Càncer, m’agradaria recordar el Thomas,
l’Amado i la Laura Cutando, m’ho he passat teta amb vosaltres tot aquest
temps; l’Eulàlia, pel teu somriure incondicional i pels arrays; a l’Heleia,
l’Aina i el David per les miniconverses als passadissos; a Maribel, por tu
ayuda en el estabulario; a Carlos, por entregar los paquetes siempre de
tan buen humor; y a Montse, por hacernos compañía mutua durante tu
turno de limpieza.
Ernest, tampoc m’oblido de tu! A més dels bons moments viscuts al PRBB
i de les interessants converses sobre la vida, ha estat genial compartir
trucades, idees, impressions i una fantàstica caminata por la Sierra
durant aquest últim temps.
Durant la tesi, també he tingut l’oportunitat de col·laborar amb gent
d’altres laboratoris, la qual cosa sempre és engrescadora i motivant.
Gràcies al David Torrent i la Bàrbara per ensenyar-me i ajudar-me amb
l’anàlisi del ChIP-Seq; a la Maria pels cariotips ; a l’Andreu Casali, la Kyra i
la Yolanda pels embrions de Drosophila i per ensenyar-me a extreure els
discs de l’ala de les larves; and thanks also to Thomas Jenuwein and
Valentina Perrera for their help and advices on the pericentromeric
heterochromatin field. Valentina, thank you so much for solving my
problems with genomic DNA contaminations, the project could never
have gone ahead without your help. And thank you above all for the
great time we had here in the lab and also in Freiburg, I will never forget
També volia agrair als membres del comitè de tesi les recomanacions i
suggeriments aportats quan el projecte tot just s’estava iniciant, com
també als membres del tribunal de tesi, per haver acceptar formar-ne
part i pel temps invertit en la lectura del manuscrit i la defensa.
I finalment agrair a l’AGAUR per concedir-me la beca que m’ha permès
fer el doctorat durant aquests últims tres anys, així com a l’IMIM, per
l’ajuda de publicació de tesi doctoral.
Hi ha molta gent de fora del món de la ciència que també m’ha
acompanyat i ajudat enormement durant aquesta etapa.
Un suport molt important han estat els amics de la universitat. Des que
vam acabar, els nostres camins han anat divergint, però és molt
satisfactori i gratificant descobrir que l’amistat que ens unia no era
circumstancial i que ens seguim tenint els uns als altres.
Volia agrair a la Colla Pessigolla tots els bons moments viscuts des
d’aleshores. Escapades a Camprodon, aniversaris, tardes de guitarra,
caps d’any, castanyades, un gran viatge al Sud Est asiàtic, mil converses,
recomanacions, ànims i suport quan més ho he necessitat i les més
sinceres felicitacions quan hi ha hagut coses a celebrar.
I a les Mixines, que tot i estar repartides per mitja Europa sempre ens ho
fem per reunir-nos de tant en tant i posar-nos al dia. Helena i Mercè,
companyes de penes i alegries associades al PhD, ha estat un plaer poder
compartir tot això amb vosaltres malgrat la distància. M’he sentit molt
acompanyada en aquest sentit, i us ho agraeixo enormement.
D’altra banda, aquest doctorat ha estat clarament marcat pel ball i la
dansa, que m’han omplert d’optimisme, energia i vitalitat, i m’han ajudat
a sentir-me millor en aquells moments que em faltaven els ànims.
Gràcies Danzarinas per acompanyar-me durant aquests últims 4 anys.
Per les classes, els somriures, els sopars, els nervis i les emocions
compartides i sobretot, per la complicitat aclaparadora que curiosament
s’ha generat en un grup tan heterogeni. Johanna, gracias por
descubrirme el mundo de la danza oriental y por hacerlo llegar a un
rincón tan profundo de mi corazón. Y por fortalecer a través de la danza
mi seguridad, feminidad y autoestima. Pero sobre todo, gracias por tus
nubes rosas, arcoíris y estrellas en días grises.
Però d’entre totes les ballarines del món, ningú m’ha aportat ni
m’aportarà tant com les Smiling, o hauria de dir Mahsheed? Gràcies
Gemma, Esther i Carla per tot! Podria omplir llargues pàgines recordant
mil anècdotes i destacant allò que més aprecio de vosaltres, però crec
que les paraules de Virgina Satir que ens va fer arribar l’Esther ho
resumeixen prou bé: “Crec que el millor regal que puc rebre d’algú és que
em vegi, m’escolti, m’entengui i em toqui. El millor regal que puc donar és
veure, escoltar, entendre i tocar a l’altra persona. Quan això s’ha fet,
sento que s’ha establert contacte”. Merci per haver establert aquest càlid
i reconfortant contacte.
Tampoc no em vull oblidar de la Salsa, menys profunda i espiritual que la
dansa del ventre però no menys important, i que també m’ha descobert
un munt de persones encantadores. Oscar, merci per totes les salses,
batxates i kizombes; m’encanta ballar amb tu i amb poques persones
m’ho passo tan bé. Però gràcies també per tots els petits moments, les
converses, els sopars, i la complicitat que s’ha anat teixit darrerament. I
gràcies de nou a la Gemma, l’Esther, la Carla però també a la Marta, la
Montse, el Dani, el Carlos, Luis, Siro i tants altres per les incomptables
nits de salsa i diversió.
I a la gent del CEC! Fa només un any que vaig posar per primera vegada
els peus dins l’emblemàtic edifici del carrer Paradís, però ja sento que
formo part d’una gran família. Especialmente interesante ha sido nuestra
conexión Miki, siento como si nos conociéramos de hace un montón.
Muchas gracias por editar la portada de esta tesis, pero sobre todo
gracias por ser una persona tan cercana y transmitir tanta confianza, por
abrirme tu casa, y por convencer a Lorena que Sants mola! Víctor,
coincidir amb tu també ha estat un plaer. Tots els sopars, les converses,
les bromes picants, els riures, “ametlles”… però un dels records que
guardo amb més amor és la sessió pràctica de salsa a Antilla, m’ho vaig
passar molt bé i espero que es repeteixi.
I finalment queda la família, què faríem sense ella? Em sento molt
afortunada de tenir la família que tinc i voldria agrair-vos a tots i
cadascun de vosaltres el suport que m’heu donat des de sempre.
Especialment vull donar les gràcies als meus pares, per tantíssimes
coses... per poder xerrar amb vosaltres sobre tot el que em passa pel cap,
per escoltar-me i tenir en compte les meves opinions, per donar-me
suport amb cada una de les meves decisions, encara que de vegades
penseu que no són les més adequades, o per la infinitat de vegades que
m’heu esperat amb el plat a taula quan vivia a casa i s’allargaven els
experiments. Però, per sobre de tot, per fer-me saber i sentir que sempre
sóc la vostra prioritat. Us estimo molt.
I Manel, què dir de tu? Simplement no hi ha una persona en aquest
planeta a qui estimi tant, que em conegui tan bé i amb qui tingui tanta
confiança i complicitat. T’estimo moltíssim i això mai canviarà.
Com diu Daniel a "Si tu em dius vine, ho deixo tot... però digue'm vine", a
la vida trobes perles i diamants. Les perles son les persones que
t'acompanyen a la vida, mentre que els diamants són les persones que et
marquen a la vida: les que estimes, les que t'estimen... les que et deixen
una empremta que no se s’anirà mai. Aquelles persones que, per molt
que passi el temps i per molt que deixis de veure, sempre recordaràs. Així
doncs, puc estar contenta d’haver coincidit amb tantes perles durant
aquests anys, i fins i tot d’haver conegut algun dels meus diamants.
Fly UP