Analysis of the Barr body with super-resolution microscopy

X chromosome inactivation (XCI) in female mammalian cells is an
ideal model system to study the relationship of epigenetic
regulation and higher-order chromatin structure. However, light
microscopic studies of chromosomal organization have long been
limited by the diffraction barrier of optical resolution.
Super-resolution 3D-structured illumination microscopy (3D-SIM) –
one of several recent techniques that circumvent this limitation –
enables multicolor optical sectioning of entire cells with
eightfold-improved volumetric resolution compared to conventional
fluorescence imaging methods. In the present work, 3D-SIM has been
applied to analyze higher-order chromatin structure of the Barr
body in mammalian nuclei, a characteristic hallmark of XCI, with
yet unprecedented detail. First, the increased resolution prompted
to reappraise the potential detrimental effect of the DNA-FISH
procedure on chromatin structure. Comparative analyses revealed
slight deteriorations at the resolution level of 3D-SIM, especially
within more decondensed euchromatin sites within the nuclear
interior. In contrast, overall nuclear morphology and the nuclear
envelope as well as heterochromatic sites in general maintained
well preserved. The results suggest that DNA-FISH studies can
benefit from a combination with super-resolution microscopy. In
particular, when keeping in mind the current developments of the
FISH technique with increasingly small and higher-complexity
probes. The compact shape of the Barr body led to the assumption of
a contribution of this special higher-order chromatin structure to
the establishment and maintenance of the silenced state in the
inactive X chromosome (Xi). However, a confirmation of this view
has always been hampered by the restrictions of conventional light
microscopy. In this work, the 3D chromosomal organization of the Xi
and autosomes has been investigated with 3D-SIM in various human
and mouse somatic cells and in mouse embryonic stem cell (ESC)
lines. The precise subchromosomal localization of a variety of
factors involved in XCI in different developmental states was
qualitatively and quantitatively assessed utilizing combined
immunofluorescence, EdU- pulse and RNA-/DNA-FISH labeling protocols
and novel data analysis tools customized for the special
requirements of 3D-SIM. The results demonstrate that all autosomes
are made of a three-dimensional interconnected network of chromatin
domains (CDs, or topology associated domains, TADs) of
highly-variable shape and dynamics. CDs/TADs are comprised of a
compacted chromatin core enriched with repressive marks, which is
collectively proposed to be the functionally passive chromatin
compartment (PNC). This PNC is surrounded by a 50 – 150 nm locally
defined, less compacted perichromatin region (PR) that is enriched
with active histone modifications and pervaded by a
three-dimensional interchromatin (IC) network. The PR and the IC
are collectively referred to as being the functionally relevant
active nuclear compartment (ANC) that harbors all major nuclear
processes, including transcription and replication. 3D-SIM data
revealed that the Barr body maintains this principle
compartmentalization and that it is still pervaded by a narrow ANC
network, which is able to fulfill its functional role as a hub for
replication or rarely occurring expression of XCI-escape genes.
Live-cell super-resolution imaging on HeLa H2B-GFP cells confirmed
that the observed chromatin features do not reflect fixation
artifacts. Xist RNA, the key factor of XCI, has been found to be
preferentially located as distinct discernible foci within the ANC
throughout the entire volume of the Barr body. Here, it is tightly
associated with a Xi-specific form of the nuclear matrix protein
SAF-A, which confirms a previously suggested role for this
Xi-enriched protein in Xist RNA spreading. In contrast, Xist RNA
shows no spatial correlation with repressive Xi-enriched histone
marks that are found within compacted chromatin sites. This
specific localization of Xist RNA reflects an intrinsic feature as
it is already present during early spreading in differentiating
female ESCs, where it precedes chromatin compaction concomitant
with RNA Polymerase II exclusion. Its localization is further
confirmed in a male ESC line carrying an inducible Xist transgene
on an autosome, but where Xist RNA fails to form a true autosomal
Barr body, which is less compacted and maintains transcriptional
activity. Last, Xist RNA shows no direct association with PRC2, the
mediator of H3K27me3, which is in contrast to the generally
believed direct recruitment model of PRC2 to the Xi by Xist RNA.
The data collected in this work reflects further support and a
refinement of the not unequivocally accepted CT-IC (chromosome
territory - interchromatin compartment) model of higher-order
chromosome architecture. In addition, a first attempt has been made
to integrate these findings with a recently growing number of
studies using chromosome conformation capturing (3C)-based
techniques and to complement them on the single-cell level.
Finally, a novel model for Xist RNA function in XCI is presented,
which proposes a sequence-independent structural role for gene
silencing and the formation of a repressive chromatin compartment.