LETTERS
Transcriptional repression mediated by repositioning
of genes to the nuclear lamina
K. L. Reddy
1,2
, J. M. Zullo
1,2
, E. Bertolino
2
& H. Singh
1,2
Nuclear compartmentalization seems to have an important role
in regulating metazoan genes
1,2
. Although studies on immuno-
globulin and other loci have shown a correlation between posi-
tioning at the nuclear lamina and gene repression, the functional
consequences of this compartmentalization remain untested
2,3
.
We devised an approach for inducible tethering of genes to the
inner nuclear membrane (INM), and tested the consequences of
such repositioning on gene activity in mouse fibroblasts. Here,
using three-dimensional DNA-immunoFISH, we demonstrate
repositioning of chromosomal regions to the nuclear lamina that
is dependent on breakdown and reformation of the nuclear enve-
lope during mitosis. Moreover, tethering leads to the accumulation
of lamin and INM proteins, but not to association with pericentro-
meric heterochromatin or nuclear pore complexes. Recruitment of
genes to the INM can result in their transcriptional repression.
Finally, we use targeted adenine methylation (DamID) to show
that, as is the case for our model system, inactive immunoglobulin
loci at the nuclear periphery are contacted by INM and lamina
proteins. We propose that these molecular interactions may be used
to compartmentalize and to limit the accessibility of immuno-
globulin loci to transcription and recombination factors.
In mammalian nuclei, chromatin is organized into structural
domains by association with distinct nuclear compartments
2
.
Several studies have shown a correlation between the transcriptional
repression of mammalian genes and their positioning at the nuclear
periphery
3–7
. In yeast, the nuclear periphery is comprised of at least
two sub-compartments: a repressive compartment consisting of foci
of silencing factors, and a permissive compartment involving nuclear
pore complexes (NPCs) that facilitates gene expression
8–10
. However,
metazoan systems exhibit a greater complexity of nuclear compart-
ments and chromosome organization. The nuclear periphery in
mammalian cells is constituted by a distinct set of INM proteins,
such as LBR, LAP2 and emerin (EMD), as well as an underlying
nuclear lamina, which have been proposed to interact with transcrip-
tional repressors
11–14
. The ability of this nuclear compartment to
regulate gene activity has not been functionally tested in metazoan
cells
2
.
We designed a two-component inducible system that would relo-
calize an integrated reporter gene from the interior of a mammalian
nucleus to the INM (Fig. 1a). The reporter construct is comprised of
the herpes simplex virus thymidine kinase promoter and the hygro-
mycin resistance gene (Tk-hyg) as well as a nearby array of Lac opera-
tors (lacO) that constitute binding sites for the Escherichia coli Lac
repressor (LacI) (Fig. 1 and Supplementary Fig. 1a)
15
. The second
component is either a nucleoplasmic green fluorescent protein
(GFP)–LacI that binds lacO sites and enables visualization of the
reporter gene or a tethering protein GFP–LacI–DEMD that is tar-
geted to the INM by means of a carboxy-terminal segment of EMD
16
.
The GFP fusion proteins were stably expressed in NIH3T3 fibroblast
clones harbouring the reporter gene(s) integrated at single (S) or
multiple (M) chromosomal sites. Reporter gene visualization and/or
repositioning were controlled using the allosteric inhibitor IPTG
(isopropyl b-D-1-thiogalactopyranoside), which regulates LacI bind-
ing to lacO sites. The initial disposition of the integrated reporter
genes was analysed in cells stably expressing GFP–LacI. Up to four
bright GFP foci were visible in clone-M nuclei because these cells have
four integration sites, each containing multiple copies of the reporter
gene (Supplementary Fig. 1d, e). In contrast, clone-S nuclei exhibited
dimmer single GFP foci owing to a single site of insertion with fewer
copies (1–2) of the reporter (Supplementary Fig. 1c, d). We next
generated clone-M and clone-S derivatives stably expressing GFP–
LacI–DEMD. As anticipated, this tethering protein localized to the
INM. On removal of IPTG, large GFP foci were observed at the
nuclear periphery in clone-M but not in clone-S cells expressing
GFP–LacI–DEMD (Supplementary Fig. 1e). This suggested that the
reporter genes were being repositioned to the nuclear membrane in
clone-M cells.
Not all tethered reporter genes were expected to accumulate the
fusion protein at levels that are discernable as fluorescent signals
above the distribution in the INM. This was probably the case for
clone-S cells. Therefore, we undertook fluorescent DNA in situ
hybridization on three-dimensional preserved nuclei (3D DNA-
immunoFISH) to assess quantitatively the disposition of all Tk-hyg
integrations (Fig. 1b, c and Supplementary Fig. 2). Under control
conditions, the integrated reporter genes were distributed through-
out the nucleoplasm, with approximately 25–30% being positioned
near the nuclear periphery (Fig. 1d). This frequency represents the
initial sub-nuclear distribution and is similar to that observed for
endogenous genes that are not associated with the nuclear peri-
phery
17
. On withdrawal of IPTG, most Tk-hyg insertions were found
to be associated with the nuclear lamina in clone-M (70%) and clone-
S (90%) cells expressing GFP–LacI–DEMD. Moreover, in clone-M
cells, reporter genes residing on different chromosomes were
repositioned to distinct regions of the INM in a single nucleus
(Fig. 1b). In clone-S cells, repositioning was mediated by fewer copies
of the lacO segments (1–2) compared with in clone-M cells (,25
copies per integration site, Supplementary Fig. 1d). We note that
repositioning requires breakdown and reformation of the nuclear
envelope during mitosis (Supplementary Fig. 3 and Supplementary
Discussion). These data provide the first demonstration of directed
repositioning of chromosomal segments to the INM–lamina com-
partment, and suggest that an intervening cell cycle may be necessary
for such re-configuration.
We next analysed the consequences of accumulating GFP–LacI–
DEMD at sites of tethering on the disposition of other proteins at the
INM. Lamin A and B1, key components of the lamina, and the INM
protein LAP2 accumulated at sites of tethering (Fig. 2a and Supple-
mentary Fig. 4). No such interactions were observed on non-tethered
1
Howard Hughes Medical Institute,
2
Department of Molecular Genetics and Cell Biology, The University of Chicago, GCIS W522, 929 East 57th Street, Chicago, Illinois 60637, USA.
Vol 452 | 13 March 2008 | doi:10.1038/nature06727
243
Nature Publishing Group ©2008