Biology Reference
In-Depth Information
DamID and ChIP mapping studies in worms, fly, mouse
and human cells have revealed remarkable patterns of
genome
Genes associated with the NL are almost invariably
expressed at very low levels. The converse is not always
true: not every inactive gene is associated with the NL.
During differentiation of mouse embryonic stem cells,
hundreds of genes shift their position relative to the NL.
Detachment of genes from the NL coincides with activation
of transcription, or with priming for activation at a later
stage
[31]
.
At least to some degree, positioning of genes at the NL
plays a causal role in their repression. Artificial tethering of
a reporter gene to the NL often leads to repression of this
gene, and can also lead to downregulation of neighboring
genes, although the magnitude of this repressive effect
appears to depend on the reporter gene and its integration
context
32]
. In each species,
hundreds of large genomic domains contact the NL. Mouse
and human genomes harbor more than 1000 of these
lamina-associated domains (LADs), distributed over all
chromosomes. Mammalian LADs have a median size of
about 0.5 Mb and together cover nearly 40% of the genome.
A similar fraction of the genome was found to interact with
the NL in worms and flies. How does so much DNA fit at
the periphery? FISH microscopy of individual LADs
suggests that the interactions of LADs with the NL do not
take place in every cell at all times, but rather have
a significant stochastic component. Thus, effectively only
a subset of LADs may contact the NL at any given time.
This is not to say that LADs are 'softly' defined. On the
contrary, genome-wide maps indicate that most mammalian
LADs have remarkably sharp borders, which tend to be
marked by specific sequence elements, such as binding sites
for the insulator proteinCTCF
[30]
. This indicates that LADs
are at least in part 'hard-coded' in the genome itself. While
some LADs are cell-type specific, the overall pattern of NL
interactions is strikingly similar in a variety of cell types
[31]
.
This raises the interesting possibility that a basal spatial
organization of chromosomes is shared among all cell types.
It is not firmly established whether LADs are actively
anchored to the NL, or instead are passively pushed to the
nuclear periphery as a consequence of forces that drive inter-
LAD regions (genomic regions not associated with the NL,
and located inbetween LADs) to in the nuclear interior. Both
mechanisms may contribute to the positioning of loci rela-
tive to the periphery. However, two lines of circumstantial
evidence indicate that active anchoring of LADs occurs.
First, lamins and other proteins of the NL are known to bind
in vitro to DNA, nucleosomes, and various other chromatin
components
[33]
. Second, the sharp demarcation of LAD
borders argues against a passive 'brushing' of LADs against
the NL. At present, molecular mechanisms that anchor
LADs to the NL in vivo are not known.
Worms, fruit flies andmammals share the striking domain
organization of NL interactions. Nevertheless, there are
interesting differences. LADs in D. melanogaster are about
five times smaller than their mammalian counterparts. This
may be related to the smaller genome size and themuch closer
spacing of genes in the fly genome. Interestingly, the average
number of genes per LAD is strikingly similar between
human and fly, suggesting a role of LAD organization in the
regulation of gene expression (see below). NL interactions in
C. elegans occur primarily in the distal parts of the chromo-
somes, suggesting a spatial organization of chromosomes
where central parts of the chromosomes are positioned in the
nuclear interior and the chromosome ends at the periphery.
A common theme among the four investigated species
is that NL interactions are tightly linked to gene repression.
NL interaction
[29
e
e
37]
. Furthermore, knockout of lamin in
worms and fruit flies caused relocation of NL-associated
genes towards the nuclear interior, concomitant with
increased expression (in Drosophila) or stochastic activa-
tion (in Caenorhabditis) of these genes
[38,39]
. These
results do not rule out an inverse causal relationship, i.e.,
that repression of genes may cause their targeting to the
NL. In fact, it is quite possible that a positive feedback loop
exists between NL association and gene inactivation, which
could contribute to stable gene silencing. In any case, these
observations indicate that spatial organization of
[34
e
the
genome is tightly linked to gene regulation.
Yeasts do not have an NL. Nevertheless, there are some
interesting parallels with metazoans. In budding yeast the
32 telomeres cluster into 4
8 foci, and these foci are
tethered to the periphery of the nucleus. This tethering
is mediated by two nuclear envelope proteins which
interact with components of telomeric chromatin. Like
NL-associated chromatin in metazoans, telomeric chro-
matin in yeast represses transcription; moreover, detailed
mechanistic studies have indicated that the peripheral
positioning facilitates the silencing of telomeric sequences.
Thus, very similar principles apply, even though the
proteins involved differ.
e
Gene Attachment to Nuclear Pores
Nuclear pores (NPs) form another anchoring site for the
genome at the nuclear envelope. NPs are large multi-
subunit transport channels that perforate the nuclear
membranes. They occupy positions in the nuclear envelope
that are depleted of lamins. Electron microscopy indicates
that the NL is generally in close contact with relatively
condensed chromatin (consistent with the repressed state of
LADs), whereas nuclear pores are surrounded by less
condensed chromatin. Indeed, ChIP and DamID mapping
studies in yeast and fly have revealed that NP proteins
interact preferentially with a subset of active genes. There
is a caveat, however: most NP proteins are not only located
at NPs, but also roam the nucleoplasm, where they can
