Article
Reference
Positions of Potential: Nuclear Organization and Gene Expression
GASSER, Susan Margaret
GASSER, Susan Margaret. Positions of Potential: Nuclear Organization and Gene Expression.
Cell , 2001, vol. 104, no. 5, p. 639-642
DOI : 10.1016/S0092-8674(01)00259-8
Available at:
http://archive-ouverte.unige.ch/unige:123128
Disclaimer: layout of this document may differ from the published version.
1 / 1
Nuclear Organization and Gene Expression
genes inserted at telomeres or at theHMmating type loci become repressed through their association with a complex of three silent information regulators (Sir2p, 3p, and 4p), which can bind the N termini of histones H3 and H4 and propagate along nucleosomes (reviewed Susan M. Gasser*
Swiss Institute for Experimental Cancer Research CH-1066 Epalinges s/Lausanne
Switzerland
in Grunstein, 1998). Sir complexes may simply “coat”
nucleosomes, or possibly fold the repressed domain The patterns of gene expression that are established
into a higher-order structure. Double in situ and immuno- and maintained during cellular differentiation result not
fluorescence detection indicate that sites of Sir-medi- only from the targeting of stage-specific transcription
ated repression are found clustered in 6 to 8 foci, which, factors, but from the long-range organization of chroma-
compared to other genomic domains, remain relatively tin, which establishes “open” or “closed” states that are
immobile at the nuclear periphery until late G2 phase.
permissive or refractory to transcription. Typically, in The clustered telomeres provide a “sink” for repressor early embryonic development, pattern-forming mecha-
activator protein 1 (Rap1p), which binds roughly once nisms generate combinations of regulatory factors that every 18 bp in the terminal telomeric repeat. In turn, this define or characterize a given cell type. These, in turn, high concentration of Rap1p binding sites cooperates determine a specific pattern of gene expression, or a with the end binding protein yKu, to recruit Sir3p and pattern of potential gene expression, which is thereafter Sir4p and sequester the Sir complex from other sites in faithfully inherited through subsequent cell divisions (re- the genome at which it might act (reviewed in Cockell viewed in Cavalli and Paro, 1998). Replication mecha- and Gasser, 1999).
nisms capable of reforming higher-order chromatin It is clear that it is the clustering of yeast telomeric structures make use of DNA methylation, histone tail repeats that promotes the concentration of Sir proteins, modification, and special nucleosome assembly factors and not vice versa, since in a number of mutants, includ- (reviewed in Tyler and Kadonaga, 1999). In fly and mam- ing acetylation site mutations in histone tails, telomeric malian development, the “molecular memory” of higher- silencing is lost and Sir proteins are delocalized, without order chromatin structure is mediated at least in part disrupting the focal distribution of telomeric DNA (Gotta by members of the Polycomb group (PcG) of proteins, et al., 1996). On the other hand, mutations in structural which ensure proper formation of body structures during proteins, such as the myosin-like proteins (Mlp) 1 and differentiation (reviewed in Pirrotta, 1998). 2, were found to interfere with both telomere clustering The long-range and heritable type of gene regulation and silencing (Galy et al., 2000), providing strong confir- has many features in common with heterochromatin- mation that the organization of telomeres is indeed criti- induced silencing, or position effect variegation (PEV), cally important for repression. One of these proteins, a phenomenon in which active genes placed near repeti- Mlp2, binds yKu, which in turn provides the bridge to tive DNA succumb to heritable yet variegated states of telomeric DNA. Consistently, the absence of yKu com- repression. In such instances, it is thought that hetero- promises telomeric repression and telomere clustering chromatic satellite DNA nucleates a compact chromatin in yeast (Laroche et al., 1998). Ground-breaking studies structure that is refractory to transcription. Even in bud- on the nuclear myosin-like proteins give us our first ding yeast, which is especially poor in repetitive DNA, handle on conserved structural proteins that may orga- one observes a crucial role for a heterochromatin-like nize chromatin within the eukaryotic nucleus. The equiv- silencing in the maintenance of haploid cell mating type alent coiled-coil proteins in flies and mammals form and rDNA stability. The major components of this mech- fibers that extend inwards from nuclear pores, rather anism have been well-characterized in budding yeast than underlying the nuclear envelope as they do in yeast.
Nuclear lamins may provide alternative DNA attachment (called silent information regulators or Sir proteins), and
sites at the nuclear periphery in higher eukaryotes.
although Sir-mediated repression may reflect only a
In yeast, it could be shown that the tethering or tar- subset of the mechanisms at work in higher eukaryotes,
geting of a silencer-flanked reporter gene to the nuclear it has provided several paradigms for how nuclear orga-
envelope facilitates its repression (Andrulis et al., 1998).
nization can influence long-range silencing. In this re-
On the other hand, association with the nuclear periph- view, I summarize recent studies that address how the
ery, or even with a telomere cluster, is not sufficient to spatial organization of yeast, fly, and mammalian nuclei
silence a PolII promoter. Notably, if the telomeric pools can facilitate gene regulation, primarily acting through
of Sir factors are dispersed due to a Rap1 C-terminal chromatin binding factors that accumulate in regions
truncation, or if the reporter construct has no silencer rich in repetitive DNA.
element, localization to the nuclear periphery fails to Lessons from Yeast
promote repression (Andrulis et al., 1998, and D. Shore, Studies of yeast silencing have shown that both the
personal communication). In this samerap1C-terminal concentration of general chromatin factors and their
truncation mutant, silencer-mediated repression is pos- subnuclear distribution play crucial roles in chromatin-
sible at “internal” sites, whereas in wild-type cells, inter- mediated gene regulation. In S.cerevisiae, RNA PolII
nal Sir concentrations are insufficient to repress a si- lencer-flanked reporter located far from a telomere.
These results and others suggest a model in which
* E-mail: sgasser@eliot.unil.ch
Cell 640
Sir proteins and transcriptional activators compete for a given promoter once it is within a compartment that can confer repression, such as that formed by telomeric foci (Aparicio and Gottschling, 1994). Although Sir pro- teins are concentrated at telomeres, increased expres- sion can improve silencing efficiency, indicating that the normal protein levels are limiting for repression. Intrigu- ingly, strong overexpression of either Sir2p or Sir4p has the opposite effect, derepressing silent domains. Since this negative effect can be overcome by coordinately overexpressingSIR3andSIR4, or subdomains of Sir3p and Sir4p, the balance between these components ap- pears to be critical and is likely to be tightly regulated in normal cells.
From these studies, a few principles basic to hetero- chromatin-type repression are suggested. First, struc-
tural chromatin factors that promote extended regions Figure 1. Two Potential Modes of Action for Ikaros-Mediated Re- of repression are usually limiting in concentration within pression
the nucleus. Second, the binding of these complexes Both the silenced promoter and pericentric satellite DNA are shown can be nucleated by sequence-specific DNA binding to contain binding sites for Ikaros. Thus, gene a may be sequestered near heterochromatin by dimerization of Ikaros, as depicted in (A).
factors, that are themselves implicated in both transcrip-
Alternatively, the juxtaposition may be due to other factors. Then, tional activation and repression, depending on the con-
as shown in (B), high concentrations of Ikaros and enzymes involved text of their binding sites (examples include Rap1p in
in transcriptional remodeling and repression, such as the Mi-2-con- yeast and Ikaros in mouse, see below). Third, multiple
taining NuRD complex, may enhance the frequency with which gene binding sites for these transcription factors can be pres- a is repressed.
ent in repetitive DNA. Finally, the spatial distribution of repetitive DNA within the nucleus appears to form
subcompartments that favor the packaging of chromatin which tend to cluster together in interphase nuclei and to into a repressed state. Such compartments of repres-
nucleate a repressed chromatin structure that “spreads”
sion often contain hypoacetylated histones, a subclass stochastically into nearby genes, creating a variegated of histone variants, and in some species, CpG methyl-
pattern of transcription. However, several examples ation.
have demonstrated that gene expression can also be It is reasonable to predict that enzymes that catalyze
modified “in trans” if genes are placed in spatial proxim- the assembly of silent chromatin would also be enriched
ity to pericentric satellite DNA. Notably, inDrosophila in these compartments. Indeed, Sir2p itself has an NAD-
the insertion of a block of heterochromatin into one dependent histone deacetylase activity (Imai et al., 2000)
allele of the euchromaticbrown (bw)gene results in the that appears to be important for silencing. In mice, the
association of both this mutant locus and the wild-type methyl CpG binding protein 2 (MeCP2) is associated
allele, present on the homologous chromosome, with with satellite DNA and can recruit a deacetylase-con-
centromeric heterochromatin. The spatial juxtaposition taining complex to artificial promoters. Finally, in cultured
of the wild-type gene with the centromeric cluster corre- mammalian cells, the deacetylase-associated nucleo-
lates with its inactivation (reviewed in Cockell and Gas- some remodeling factor NuRD, can be found both at
ser, 1999). More recent data suggest that gene silencing the nuclear periphery and associated with centromeric
in mammalian cells can be mediated by the Zn-finger- satellite repeats, albeit in a cell type–dependent manner
containing, sequence-specific transcription factor Ikaros, (Kim et al., 1999; Cobb et al., 2000). Thus, along with
which becomes associated both with silent genes and structural components of heterochromatin, at least
centromeric heterochromatin in differentiating lympho- some enzymes capable of modifying chromatin states
cytes (Brown et al., 1997). Using in situ hybridization localize to zones that promote repression.
to identify the subnuclear position of several Ikaros- Analogies in Complex Organisms
regulated genes, it was shown that the sequences them- This simple view of chromatin-mediated repression
selves become juxtaposed in trans to centromeric het- gleaned from yeast is embellished in higher eukaryotic
erochromatin as they are silenced, particularly when organisms, most strikingly with respect to cell type–
primary B-cells switch from a resting to a cycling state specific regulation. For example, recent studies suggest
(Brown et al., 1999). In contrast, activated genes do not that genes that are specifically silenced during lympho-
become centromere associated.
cyte differentiation get “recruited” to heterochromatic
Further documentation of transcription-related chro- compartments when they are repressed, while genes
matin movement in mammalian cells comes from the that become activated move elsewhere (Brown et al.,
Groudine laboratory, which showed that the-globin 5⬘ 1997, 1999). One major advance in the last year has
HS2 enhancer actively prevents a transgene from being been to identify the order of events governing this mode
localized adjacent to heterochromatin and “recruits” it to of regulation. A second has been to demonstrate the
an active compartment, in which transcription is favored diversity of mechanisms that can function within a single
and stably inherited. Core enhancer motifs are required cell. Both are summarized below.
both to suppress transgene silencing and to relocate In multicellular eukaryotes, centromeric heterochro-
matin is composed of repetitive satellite sequences, the gene away from centromeric heterochromatin during
Figure 2. Cell Type–Specific Juxtaposition of Genes to Satellite DNA
A hypothetical situation in two differentiated higher eukaryotic cells, each containing three different types of repetitive elements (here arbitrarily labeled SatA, SatB, and SatC). Dif- ferent promoter-bound complexes (in blue or red, which may contain cell-type specific fac- tors) recruit target genes (X or Z) to the zones of heterochromatin in which they have bind- ing sites, depending on the cell type (com- pare A and B). Repression is likely to be achieved by a common mechanism involving histone modification and long-range chroma- tin compaction in both cases. Variations in spatial organization permit many different patterns of repression, making use of both cell type–specific transcription factors and general repressors of active chromatin (indi- cated as blue, red, or green spots within the heterochromatin).
interphase (Francastel et al., 1999). Further chromatin chromatin renders the expression of a constitutively ex- pressed gene stochastic, since the repressed chromatin studies show that it is not active transcription itself that
correlates with the “release” of the transgene from a state can efficiently compete for the promoter as cells divide.
heterochromatin environment, but the hyperacetylated
state of its associated histones. Specifically, localization To try to identify what mediates the large-scale move- ment of transcriptional domains to centromeric chroma- of a human-globin locus away from a satellite cluster
in mouse erythroleukemia cells correlates with hyper- tin such as those that occur during the differentiation of T and B cell lineages, Fisher and colleagues initially acetylation in the promoter region (Schu¨beler et al.,
2000). Since this occurs even in the absence of the-glo- suggested that Ikaros might bind members of the PcG group, which, at least in mammalian cells, also associate bin locus enhancer, or “locus control region,” which is
needed to induce transcription, the authors propose with pericentric heterochromatin (Saurin et al., 1998). A recent study, however, indicates that Ikaros’s innate that it is the potential to be expressed, and not active
transcription, that precedes movement. ability to recognize consensus binding sites both in tar- get promoters and within the␥satellite repeats allows A similar conclusion can be drawn from the study of
a transgene inserted directly into the major mouse␥ the recruitment of Ikaros, and presumably of its relevant target genes, to heterochromatin (Cobb et al., 2000;
satellite repeat. As expected, Lundgren and colleagues
found a variegated expression phenotype for the 5 Figure 1A). Although Ikaros can also bind Mi-2, a compo- nent of the deacetylase-containing chromatin remodel- transgene placed in a heterochromatic environment
(Lundgren et al., 2000). When totally inactive and inac- ing factor NuRD, Mi-2 does not colocalize significantly with Ikaros in the NIH3T3 cells used. In other cell types, cessible to DNase1, the promoter was shown by in situ
hybridization to be buried deep within a zone of hetero- Mi-2 does colocalize with pericentric heterochromatin, consistent with a role in repression, even though it is chromatin. Upon the binding of the basal E2A factors,
the promoter became accessible to DNase1 although unlikely to be the “landing pad” for Ikaros (Figure 1B).
Since not all Ikaros-regulated genes are centromeric the gene was not actively transcribed. This coincided
with movement of the 5 locus to the surface of the in a given cell type, it was proposed that the interaction of certain Ikaros isoforms with other cell lineage–specific centromeric heterochromatin, confirming that changes
in chromatin structure, rather than active transcription, partners might restrict the relocalization of the target genes to specific developmental pathways. This would precedes or promotes rearrangements, even within a
heterochromatic domain. When the transgene was ex- allow a coordinated sequestering and release of genetic domains into repressive compartments during the differ- amined in pre-B cells in which a strong transactivator
was expressed, the silent heterochromatic state was entiation process. Support for this hypothesis still re- quires direct proof that the Ikaros sites in promoters overcome and the gene was actively transcribed, yet it
remained on the surface of the heterochromatic domain. are required for gene repression, and positive evidence indicating that Ikaros binding precedes relocalization to This suggested that juxtaposition to repetitive DNA is
not incompatible with expression, but that a strong acti- centromeric domains.
Different Repeats, Different Sites, and Different vator is required to overcome the repressed state. This
is highly reminiscent of results obtained in yeast for a Types of Repression
An elegant study inDrosophilahas expanded the evi- subtelomericURA3reporter (Aparicio and Gottschling,
1994). Overexpression of the strong transcriptional acti- dence that different types of heterochromatin repeats have different characteristics by analyzing a number vator Ppr1p could override Sir-mediated repression,
particularly in late G2 phase, suggesting that open and of randomly inserted transgenes that landed either in telomeric associated sequences on chromosomes 2 and closed chromatin states compete in cycling cells. None-
theless, these studies indicate that proximity to hetero- 3, or in the repetitious transposable elements found at
Cell 642
Selected Reading the telomere of chromosome 4 (Cryderman et al., 1999).
As expected, these telomeric transgenes were subject
Andrulis, A.D., Neiman, A.M., Zappulla, D.C., and Sternglanz, R.
to variegated repression, but they showed different de-
(1998). Nature394, 592–595.
grees of repression and responded differently to muta-
Aparicio, O.M., and Gottschling, D.E. (1994). Genes Dev.8, 1133–
tions in genes known to affect PEV, namely, suppressor 1146.
2 of zeste (Su(z)2) andSu(var)2-5, which encodes HP1, a
Brown, K.E., Guest, S.S., Smale, S.T., Hahm, K., Merkenschlager, component of pericentric heterochromatin. Localization M., and Fisher, A.G. (1997). Cell91, 845–854.
of the transgenes within the nucleus by in situ hybridiza- Brown, K.E., Baxter, J., Graf, D., Merkenschlager, M., and Fisher, tion, showed that the transgenes that are sensitive to A.G. (1999). Mol. Cell3, 207–217.
HP1 mutations were spatially juxtaposed to centromeric Cavalli, G., and Paro, R. (1998). Curr. Opin. Cell Biol.10, 354–360.
heterochromatin, while those insensitive to HP1 were Cimbora, D.M., and Groudine, M. (2001). Cell 104, this issue, not. Reciprocal translocations allowed the authors to 643–646.
correlate changes in nuclear position with changes in Cobb, B.S., Morales-Alcelay, S., Kleiger, G., Brown, K.E., Fisher, A.G., and Smale, S.T. (2000). Genes Dev.14, 2146–2160.
the strength of the repression. However, the local “het-
erochromatin” context of the chromosome 2 transgene Cockell, M.M., and Gasser, S.M. (1999). Curr. Opin. Genet. Dev.9, 199–205.
was dominant over its nuclear position, i.e., the trans-
Cryderman, D.E., Morris, E.J., Biessmann, H., Elgin, S.C., and Wall- gene did not respond to changes in HP1 concentrations
rath, L.L. (1999). EMBO J.18, 3724–3735.
even when spatially associated with centromeric do-
Francastel, C., Walters, M.C., Groudine, M., and Martin, D.I. (1999).
mains (Cryderman et al., 1999). The fact that both local
Cell99, 259–269.
repeat context and position within the nucleus influence
Galy, V., Olivo-Marin, J.C., Scherthan, H., Doye, V., Rascalou, N., the character of gene repression, hints at the diversity
and Nehrbass, U. (2000). Nature403, 108–112.
that may be possible in mammalian cells, which seem
Gotta, M., Laroche, T., Formenton, A., Maillet, L., Scherthan, H., and to have numerous variants of repetitive elements. Due
Gasser, S.M. (1996). J. Cell Biol.134, 1349–1363.
to the inherent “stickiness” of repetitive DNA, and its
Grunstein, M. (1998). Cell93, 325–328.
capacity for self-recognition, it is possible that different
Imai, S.I., Armstrong, C., Kaeberlein, M., and Guarente, L. (2000).
satellites’ repeats will form distinct clusters, each able
Nature403, 795–800.
to offer a unique set of transcription factor recruitment
Kim, J., Sif, S., Jones, B., Jackson, A., Koipally, J., Heller, E., Wi- sites. Cell-type specific patterns of gene regulation nandy, S., Viel, A., Sawyer, A., Ikeda, T., et al. (1999). Immunity10, could then be mediated, at least in part, by a selective 345–355.
sequestering of genes into distinct heterochromatin Laroche, T., Martin, S.G., Gotta, M., Gorham, H.C., Pryde, F.E., Louis, compartments, which may respond differentially to mod- E.J., and Gasser, S.M. (1998). Curr. Biol.8, 653–656.
ifiers of chromatin structure, although they use the same Lundgren, M., Chow, C., Sabbattini, P., Georgiou, A., Minaee, S., general mechanism of repression (Figure 2). In view of and Dillon, N. (2000). Cell103, 733–743.
the data suggesting that it is largely chromatin structure Pirrotta, V. (1998). Cell93, 333–336.
and not active transcription that correlates with a posi- Saurin, A.J., Shiels, C., Williamson, J., Satijn, D.P., Otte, A.P., Sheer, D., and Freemont, P.S. (1998). J. Cell Biol.142, 887–898.
tion in the nucleus, it is perhaps more appropriate to
propose that relocated genes assume positions of po- Schu¨beler, D., Francastel, C., Cimbora, D.M., Reik, A., Martin, D.I., and Groudine, M. (2000). Genes Dev.14, 940–950.
tential, reflecting their potential for repression or expres-
Tyler, J.K., and Kadonaga, J.T. (1999). Cell99, 443–446.
sion, due to a position near heterochromatin or distant from it.
Naturally it is the inheritance of differentiated gene expression patterns that is of primary importance in the development of a complex organism. Thus, there may be need for a system for inheriting three-dimensional order. Since both repetitive DNAs and nuclear localiza- tion tend to influence the timing of replication, these pockets of potential repression or expression will tend to be replicated at specific times in S phase. One can speculate that the late replication of genes associated with heterochromatin propagates not only a local chro- matin organization, but a tag that maintains its localiza- tion within a subnuclear zone (see Cimbora and Grou- dine [this issue ofCell]). We should remember, however, that the nucleus is not hard wired, and that there is a great deal of movement of both DNA and proteins throughout the cell cycle. Perhaps in addition to the zones of heritable repression established by the Poly- comb group genes, the eukaryotic nucleus profits from the tendancy of repetitive, heterochromatic domains to cluster, to set up networks of interactions within the nucleus. It is tempting to speculate that these compart- ments and their associated genes embody the molecular memory of genetic expression.
