Biology Reference
In-Depth Information
also proposes that a chromatin loop is formed, but includes
an active step in which protein complexes bound to the
regulatory element 'track' along the chromatin fiber, pull-
ing the regulatory element with them, until a target
promoter is reached.
Over the last decade 3C-based studies have confirmed
the formation of specific looping interactions between gene
promoters and enhancers and insulators. One of the best-
studied examples is the
the chromatin fiber is thus highly constrained by the
combined action of specific associations between two or
more loci. The fact that this class of chromatin interactions
displays considerable specificity does not mean that the
chromatin fiber will be folded in exactly the same 3D
structure in each cell. The occurrence of each looping
interaction is infrequent, across the cell population and
perhaps over time within a single cell, and some interac-
tions may be mutually exclusive. Thus, even at the sub-Mb
scale folding of the chromatin fiber will be variable.
Finally, the observation that chromatin loops are formed
does not provide insights into the processes that bring the
two interacting loci together. The tracking model would
propose a directed and possibly active process, but at the
length scale of hundreds of kb the chromatin fiber displays
significant Brownian motion so that two loci can readily
encounter one another by free diffusion. In addition, the
persistence length of the chromatin fiber is relatively short,
at several kb, compared to the relatively large chromatin
loops typically observed (
-globin locus. This locus contains
a cluster of related globin genes that encode the
b
-subunit
of hemoglobin. A single complex gene regulatory element,
the locus control region (LCR),
b
80 kb
upstream of the globin genes and is required for high levels
of gene expression. 3C assays have detected specific
looping interactions between the LCR and the globin genes,
but only in cells in which they are expressed
[61]
.
Furthermore, specific transcription factors that bind the
LCR and the globin promoters, e.g., GATA1 and EKLF1,
were shown to mediate this interaction and to be required
for globin gene activation
[62,63]
.
Since these initial studies a large number of examples
of chromatin looping interactions have been described that
are involved in gene expression. These include other
developmentally controlled gene clusters such as the
is located 20
e
10 kb), and thus the physical
stiffness of chromatin is not limiting the occurrence of
these interactions. Unraveling the molecular processes
leading to loop formation, and their dynamic properties, is
an exciting new challenge that may require the develop-
ment of new or
>
-
globin
[64,65]
, Th2 interleukin
[66]
, and Hox gene clus-
ters
[67
a
improved imaging and/or 3C-based
69]
, but also more broadly expressed single gene
loci such as BRCA1
[70]
,MYC
[71]
and CFTR
[72,73]
.
A recent large-scale study of looping interactions that
involve over 600 genes in the human genome performed in
the laboratory of J. Dekker found that the majority of them
in a given cell type are engaged in one or more chromatin
looping interactions with putative gene regulatory
elements located within several hundred kb
[54]
. Gene
promoters can interact with more than one enhancer or
insulator, but do not interact with all active enhancers
located around the gene, indicating a significant level of
specificity. How specificity is achieved is not known, but it
seems highly plausible that combinations of specific
DNA-binding proteins associating with the interacting
loci, as well as other chromatin-bound complexes, will
play a critical role.
technologies.
A network view of chromatin folding and nuclear
organization also provides a new lens through which to
view the regulation of genes through the combinatorial
action of multiple, possibly cell-type-specific regulatory
elements spread out over hundreds of kb. Current studies
are aimed at generating detailed maps of these networks of
gene-regulatory interactions. Over the next several years
we will undoubtedly witness tremendous progress in
genome-scale mapping of looping interactions, revealing
new principles of gene regulation and the roles of the
spatial organization of chromosomes in this process.
e
A STOCHASTIC INTERACTION-DRIVEN
MODEL FOR GENOME FOLDING
AND NUCLEAR ORGANIZATION
As outlined above, there are three main principles that drive
genome organization: polymer physics, anchoring of
chromosomes to nuclear scaffolds, and direct contacts
among chromosomal loci. We propose that a hierarchical
system of these defined, yet stochastic interactions along
and between chromosomes, and between chromosomes and
nuclear structures, determines the folding of chromosomes
and the organization of the interphase nucleus in general.
Below we describe this model in more detail and the
predictions the model makes, and then outline some of the
new questions that it invokes.
A Dynamic Network View of Chromosome
Folding
From these observations a picture of a highly looped and
folded chromatin fiber emerges, at least at the sub-Mb
scale. This structure is driven by a network of interactions,
with varying specificities, between genes, gene regulatory
elements, large chromosomal domains and between loci
and nuclear substructures such as the NL and the nucle-
olus. This network is cell-type specific and determined by
the set of active genes and regulatory elements present in
that cell. At the scale of several hundreds of kb the path of
