| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Review |
Department of Cell Biology and Cancer Center, University of Massachusetts Medical School, Worcester, Massachusetts 01655 [G. S. S., A. J. v. W., J. L. S., J. B. L.], and Departamento de Biología Molecular, Universidad De Concepcion, Concepcion, Chile [M. M.]
| ABSTRACT |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
The contributions by multiple levels of nuclear organization to control of gene expression will be evaluated. We will focus on the conceptual and experimental basis for the hypothesis that parameters of nuclear structure support cell growth and phenotypic properties of normal and tumor cells by facilitating the organization of genes, chromatin remodeling complexes, transcripts, and regulatory factors within the three-dimensional context of nuclear architecture. We will address perturbations in mechanisms that direct regulatory factors to subnuclear sites that may contribute to aberrations in control of transcription and posttranscriptional processing of gene transcripts.
| Components of Nuclear Organization Contributing to Control of Gene Expression |
|---|
|
|
|---|
From a biological perspective, each biochemical parameter of factor metabolism and activity requires control, and components of nuclear organization are linked to structure-function interrelationships that mediate the transcription and processing of gene transcripts. However, rather than representing regulatory obstacles, the complexities of nuclear biochemistry and morphology provide the required specificity for physiological responsiveness to a broad spectrum of signaling pathways to modulate transcription under diverse circumstances. It is therefore understandable why modifications in nuclear architecture and nuclear structure-function interrelationships accompany and appear to be causally related to compromised gene expression and DNA replication under pathological conditions.
Sequence Organization.
Appreciation is accruing for the high density of information in both
regulatory and mRNA coding sequences of cell growth and phenotypic
genes. The modular organizations of promoter elements provide
blueprints for responsiveness to a broad spectrum of regulatory cues
that support competency for transient developmental and homeostatic
control as well as sustained commitments to tissue-specific gene
expression. Overlapping recognition elements expand the options for
responsiveness to signaling cascades that mediate mutually exclusive
protein-DNA and protein-protein interactions. Splice variants for gene
transcripts further enhance the specificity of gene expression.
However, the linear order of genes and flanking regulatory elements is
necessary but insufficient to support expression in a biological
context. There is a requirement to integrate the regulatory information
at independent promoter elements and selectively utilize subsets of
promoter-regulatory information to control the extent to which genes
are activated and/or suppressed.
Chromatin Organization.
Chromatin structure and nucleosome organization provide architectural
linkages between gene organization and components of transcriptional
control. During the past two decades, biochemical and structural
analyses have defined the dimensions and conformational properties of
the nucleosome, the primary unit of chromatin structure. Each
nucleosome consists of approximately 200 bp of DNA wrapped in two turns
around an octameric protein core containing two copies each of histones
H2A, H2B, H3, and H4. A fifth histone, the linker histone H1, binds to
the nucleosome and promotes the organization of nucleosomes into a
higher-order structure, the 30-nm fiber. Nucleosomal organization
reduces distances between promoter elements, thereby supporting
interactions between the modular components of transcriptional control.
Higher-order chromatin structure further reduces nucleotide distances
between regulatory sequences. Folding of nucleosome arrays into
solenoid-type structures provides a potential for interactions that
support synergism between promoter elements and responsiveness to
multiple physiological regulatory signals.
It has been well established that the presence of nucleosomes generally blocks the accessibility of transcription factors to their cognate binding sequences (1) . Extensive analyses of chromatin structure have indicated that the most active genes exhibit increased nuclease hypersensitivity at promoter and enhancer elements. These domains generally reflect alterations in the classical nucleosomal organization and the binding of specific nuclear factors. Thus, DNase I digestion has been widely used to probe structures in vivo and in vitro based on the premise that chromatin accessibility to DNase I reflects chromatin access to regulatory molecules in the nucleus.
Changes in chromatin organization have been documented under many biological conditions in which modifications in gene expression are necessary for the execution of physiological control. Developmental and steroid hormone-related changes in the chromatin organization of the globin and ovalbumin genes served as the initial examples of chromatin remodeling linked to gene expression (2, 3, 4) . Although these studies predated the identification and characterization of promoter elements and cognate-regulatory factors, they provided the foundation for examining the control of nucleosome placement as a component of transcriptional control.
Transient changes in chromatin structure of a human cell cycle-regulated histone H4 gene illustrate remodeling of chromatin organization to support competency for gene expression during cell cycle progression at the G1-S-phase transition (5 , 6) . By combining DNase I digestion of isolated nuclei, indirect end labeling, and genomic sequencing, it was found that two regions in the proximal promoter of this gene exhibit cell cycle-dependent changes in chromatin structure. These regions contain key elements for both basal transcription and cell cycle regulation (7 , 8) .
Alterations in the chromatin organization of the bone tissue-specific and steroid hormone-responsive osteocalcin gene promoter during osteoblast differentiation provide a paradigm for remodeling chromatin structure and nucleosome organization that is linked to long-term commitment to phenotype-specific gene expression (9 , 10) . Transcription of the osteocalcin gene in late-stage, postproliferative osteoblasts (11 , 12) is controlled by a modularly organized promoter with proximal basal regulatory sequences and distal hormone-responsive enhancer elements (13, 14, 15, 16, 17, 18, 19, 20, 21, 22) . Modifications in chromatin structure and nucleosome organization at these two promoter regulatory domains parallel competency for transcription and the extent to which the osteocalcin gene is transcribed in response to physiological mediators of basal expression and steroid hormones. This remodeling of chromatin provides a basis for the involvement of nuclear architecture in growth factor- and steroid hormone-mediated control of osteocalcin gene expression during osteoblast differentiation. Basal expression and enhancement of osteocalcin gene transcription are accompanied by two changes in the structural properties of chromatin. DNase I hypersensitivity of sequences flanking the basal, tissue-specific element and the vitamin D enhancer domain are observed (9 , 23 , 24) . Together with changes in nucleosome placement (23) , a basis for accessibility of transactivation factors to basal and steroid hormone-dependent regulatory sequences can be explained. In early-stage proliferating osteoblasts, when the osteocalcin gene is repressed, nucleosomes are placed in the proximal basal domain and in the vitamin D-responsive enhancer promoter sequences. Nuclease-hypersensitive sites are not present in the vicinity of these regulatory elements. In contrast, when osteocalcin gene expression is transcribed postproliferatively and vitamin D-mediated enhancement of transcription occurs, the proximal basal and upstream steroid hormone-responsive enhancer sequences become nucleosome free, and these regulatory domains are flanked by DNase I-hypersensitive sites.
Among the studies that have established a role for nucleosomes regulating steroid hormone transcriptional activation, Archer et al. (25) have shown that in cell lines stably transfected with the MMTV3 constructs, positioned nucleosomes found at the long terminal repeat sequence prevent binding factors such as NF-1 to its cognate site. After ligand activation, glucocorticoid receptor can bind to a site located proximal to the NF-1 binding sequence. A hormone-dependent DNase I-hypersensitive site is generated that renders the NF-1 element available for occupancy and competent for interactions with components of the transcriptional initiation complex (26) .
Although it has been known for some time that regulatory activity of the vitamin D receptor requires chromatin remodeling to facilitate the accessibility of promoter-regulatory sequences, direct linkage of histone modifications with altered activities of steroid hormone- responsive promoter elements has been elusive. However, recent reports indicate that coactivators and repressors that interact with the vitamin D receptor include HATs and HDs (27) . Modifications have been demonstrated in the acetylation of histones in nucleosomes associated with vitamin D receptor promoter sequences (27) . These findings provide valuable insight into mechanisms linking changes in the placement and organization of nucleosomes with the control of transcription. Acetylation neutralizes positive charges of lysine residues and disrupts the association of HAT coactivator complexes with promoter-associated steroid hormone receptors. In vivo evidence is thereby provided for a key role of histone acetylation in steroid hormone-induced gene activation, and cofactor acetylation is implicated in hormonal signaling (27) .
The Nuclear Matrix.
As the intricacies of gene organization and regulation are elucidated,
the requirement to resolve a fundamental biological paradox becomes
increasingly evident. With a limited representation of gene-specific
regulatory elements and a low abundance of cognate transcription
factors, how can a threshold concentration for sequence-specific
interactions be attained to support the initiation of transcription
within nuclei of intact cells? Resolution is in part provided by
contributions of the nuclear matrix to transcriptional control. It was
this paradox, together with ultrastructural, biochemical, and molecular
genetic evidence for involvement of nuclear architecture, that prompted
the consideration of contributions by the nuclear matrix to control of
gene replication and expression (28, 29, 30, 31)
.
The anastomosing network of fibers and filaments that constitutes the nuclear matrix supports the structural properties of the nucleus as a cellular organelle and accommodates modifications in gene expression associated with proliferation and differentiation and changes necessary to sustain phenotypic requirements in specialized cells (30 , 32, 33, 34) . Regulatory functions of the nuclear matrix include but are by no means restricted to DNA replication (28) , gene localization (35) , imposition of physical constraints on chromatin structure that support the formation of loop domains, concentration and targeting of transcription factors (36, 37, 38, 39, 40, 41) , RNA processing and export of gene transcripts (42, 43, 44, 45, 46) , posttranslational modifications of chromosomal proteins, and modifications of chromatin structure (47) . Additional linkages between nuclear architecture and gene expression have recently been provided by biochemical and in situ immunofluorescence evidence showing that components of the RNA transcription complex are nuclear matrix associated and that nuclear matrix-associated sites of replication exhibit modified subnuclear distribution during the cell cycle (48) .
Subnuclear Domains.
An understanding of interrelationships between nuclear structure and
gene expression necessitates knowledge of the composition,
organization, and regulation of sites within the nucleus that are
dedicated to replication, transcription, and processing of gene
transcripts. During the past several years, there have been
developments in reagents and instrumentation to enhance the resolution
of nucleic acid and protein detection by in situ
hybridization and immunofluorescence analyses. The combined application
of isotopic and nonisotopic methods, together with a new generation of
high-resolution techniques for quantitation and three-dimensional
reconstruction from digitally captured images, is providing new
insights into the intranuclear distribution of genes and regulatory
factors (Fig. 1)
. We are beginning to make the transition from
descriptive in situ mapping of genes, transcripts, and
regulatory factors to visualization of gene expression from the
three-dimensional perspective of nuclear architecture. Initially,
in situ approaches were primarily used for intracellular
localization of nucleic acids and proteins that were first shown by
biochemical analyses to contribute to the control of gene expression.
We are now applying high-resolution in situ analyses for the
primary characterization of gene-regulatory mechanisms under in
vivo conditions.
|
| Mechanisms Mediating Nuclear Structure-Function Interrelationships |
|---|
|
|
|---|
Genomic Reconfiguration.
The most well-documented and mechanistically understood perturbations
in nuclear organization are modifications in genomic organization, gene
amplification, and rearrangements of gene loci that are prevalent in
cancer. Much remains to be established before the sequence of events
that controls gene amplification and recombination is fully defined.
However, significant progress has been made in identifying and
characterizing the enzymology of replication and recombination
(reviewed in Refs. 57
and 58
). Components of mechanisms that are
involved in cleavage, ligation, and editing are understood for events
that are associated with genomic, viral, and other episomal sequences.
Chromatin Remodeling.
During the past several years, there have been major advances in the
ability to experimentally address the molecular mechanisms that mediate
chromatin remodeling (Fig. 1)
. This is, to a significant extent,
attributable to an increased understanding of the enzymatic control of
nucleosome structure and organization. ATP-dependent chromatin
remodeling enzymes have been identified, and there is increased insight
into the activities of the enzymes that covalently modify histone
proteins.
A family of SWI/SNF-related proteins and protein complexes has been described in yeast and mammalian cells (1 , 59, 60, 61, 62) that promotes transcription by altering chromatin structure (61) . These ATP-dependent alterations render DNA sequences containing regulatory elements accessible for binding cognate transcription factors and mediate protein-protein interactions that influence the structural and functional properties of chromatin. Although the mechanisms by which these complexes function remain to be formally defined, there is general agreement that the increase in DNA sequence accessibility does not require the removal of histones (63, 64, 65) . Rather, multiple lines of evidence suggest that remodeling of the nucleosomal structure involves alterations in histone-DNA and/or histone-histone interactions. All chromatin remodeling complexes that have been reported to date include a subunit containing ATPase activity (59 , 60 , 66, 67, 68, 69, 70, 71, 72, 73) and have been shown to be critical for modifying nucleosomal organization. Because these subunits share significant homology, it has been suggested that they belong to a new family of proteins with a function that has been highly conserved throughout evolution (61) .
Posttranslational modifications of histones have been implicated in the physiological control of chromatin structure for the past three decades. However, recent findings have functionally linked histone acetylation and phosphorylation with changes in nucleosomal structure that alter accessibility to specific regulatory elements (1) . For example, acetylation of the amino-termini of nucleosomal histones has been directly correlated with transcriptional activation. Moreover, it has been observed that core histone hyperacetylation enhances the binding of most transcription factors to nucleosomes (74, 75, 76) . Nevertheless, there have been reports that chromatin hyperacetylation blocks steroid hormone transcriptional enhancement and steroid- dependent nucleosomal alterations (77, 78, 79) . Within this context, it was recently shown that hyperacetylation of nuclear proteins alters the chromatin organization of the bone tissue-specific osteocalcin gene promoter in a manner that prevents vitamin D-mediated transcriptional up-regulation. By combining nuclease accessibility, indirect end labeling, and ligation-mediated PCR analysis, it was demonstrated that protein-DNA interactions that promote the formation of a distal DNase I-hypersensitive site do not occur under conditions of hyperacetylation (79) . Similarly, Bresnick et al. (77) have reported that butyrate treatment, which results in chromatin hyperacetylation by inhibiting HD activities, abolished glucocorticoid hormone-dependent formation of the nuclease-hypersensitive site and blocked transcriptional induction of the MMTV long terminal repeat sequence. In contrast, Bartsch et al. (80) found that by decreasing the concentrations of HD inhibitors, only moderate acetylation can be obtained. Moderate acetylation leads to enhanced transcription from the MMTV promoter in the absence of hormone and potentiates transactivation by either glucocorticoids or progestins. Because these inducing inhibitor concentrations lead to a type of nucleosomal DNase I hypersensitivity similar to that caused by hormone treatment, it was suggested that moderate acetylation of core histones activated the stably integrated MMTV promoter by mechanisms involving chromatin remodeling similar to that generated by the inducing hormones.
A major breakthrough in experimentally addressing the physiological role of histone acetylation came with the recent purification and subsequent cloning of the catalytic subunits of yeast and mammalian nuclear HATs. Gcn5 encodes a 55-kDa protein in yeast that acetylates both histones H3 and H4 (81) . It has been reported that Gcn5 is part of a large (1.8-MDa) protein complex designated SAGA (82) , which includes proteins that are also present in complexes involved in transcriptional regulation (1) . Recruitment of SAGA by transcriptional activators results in localized acetylation of nucleosomal substrates in vivo and in vitro (82) . Importantly, the transcriptional stimulatory activity of the recruited SAGA complex is dependent on its HAT activity (82) .
Other proteins that contain nuclear HAT activity are p300 and its related homologue, CBP (83) . These two proteins function as transcriptional adaptors that interact with several transcription factors including cAMP-responsive element-binding protein, jun, Fos, Myb, and Myo D, and with nuclear steroid hormone receptors (84, 85, 86, 87, 88, 89, 90, 91, 92, 93) . In addition, human TAFII250 and yeast TAFII250 have HAT activity (94) . TAFII250 is part of the transcription factor IID complex that recognizes the TATA sequence at the promoter region of most genes and initiates the formation of transcription preinitiation complexes. The presence of HAT activity in this complex suggests that histone acetylation may be a requirement for transcription factor interaction with nucleosomal DNA.
Alterations in mechanisms that mediate the enzymology of histone modifications in cancer cells support involvement in the onset and/or progression of tumorigenesis. Relative examples are accumulating that implicate perturbations in both the acetylation and deacetylation of histones, and the known biochemical targets are increasing. Very importantly, in recent years, a direct correlation between abnormal histone acetylase activity and the potential for cell transformation has been found. Recent studies show the recruitment of CBP by Smad2/3 proteins in the transforming growth factor ß signaling pathway (95 , 96) . There is evidence that mutation of the players in this pathway (e.g., Smad2, Smad4, and p300) leads to the development of cancer (e.g., colorectal carcinomas; Refs. 96 and 97 ).
P/CAF, another protein that contains HAT activity, is highly homologous to both the yeast Gcn5 and the human homologue hGcn5. P/CAF interacts with p300 and CBP to form functional complexes (98) . These findings are consistent with the formation of complexes containing multiple HAT activities that can accommodate requirements for specificity of histone acetylation under different biological conditions.
Nuclear HAT activity appears to be critical during steroid hormone-dependent transcriptional activation. It has been reported that coactivation factors that include ACTR and SRC-1 recruit CBP/p300 and P/CAF to ligand-bound nuclear hormone receptors. SRC-1 and ACTR (and the related molecules RAC3, AIB1, and TRAM-1) have HAT activity. This is an example of multiprotein complexes containing different HAT activities that contribute to modifications of nucleosomal histones that are functionally linked to competency for chromatin remodeling that occurs during ligand-dependent transcriptional regulation (99) . Interestingly, AIB1 is expressed at high levels in breast cancer, but not in normal breast tissue. In addition, BRCA1 and BRCA2 are tumor suppressor genes involved in familial breast cancers. Both are thought to have roles in transcription, cell cycle control, and DNA repair. BRCA1 associates with CBP, whereas BRCA2 has HAT activity (100 , 101) . Thus, HAT activity may play a role in the function of these tumor suppressor proteins.
For histone acetylation to be a physiologically relevant component of transcriptional control, there is a requirement for a cellular mechanism to reverse this posttranslational modification. HDs that enzymatically remove acetate moieties from histone proteins have been studied extensively during the past several years. Multiple forms of this enzyme have been identified and characterized in several organisms (102, 103, 104, 105, 106) . The mammalian forms designated HDAC1, HDAC2, and HDAC3 were found to be homologous to the yeast form designated Rpd3 (107) . HDAC1 and HDAC2, but not HDAC3, are large multiprotein complexes containing corepressor molecules such as mSin3, N-CoR, or SMRT as well as the proteins SAP18, SAP30, RbAp48, and RbAp46 (107) . The Sin3A-N-CoR-/SMRT-HDAC1/2 complex and other complexes associated with HD activity can be recruited specifically to gene promoter-regulatory sequences by unliganded nuclear steroid receptors (108 , 109) . Thus, SAP30, which binds mSin3 and N-Cor, is required for N-CoR/mSin3-mediated repression of hydroxytamoxifen-bound estrogen receptor (110 , 111) but not unliganded retinoic acid receptor and thyroid receptor. It has been shown that decreasing the levels of intracellular N-CoR corepressors can lead to tamoxifen resistance in breast cancer (112) .
HDAC1 and HDAC2 are also found in another protein complex designated NuRD (113) . Thus, human NuRD complexes contain not only ATP-dependent nucleosome disruption activity but also HD activity that is usually associated with transcriptional repression. The deacetylation is stimulated by ATP on nucleosomal templates, suggesting that nucleosome disruption facilitates the access of the deacetylase to its substrates. One subunit of NuRD was identified as MTA1, a metastasis-associated protein with a region similar to N-CoR, indicating that ATP-dependent chromatin remodeling can participate in transcriptional repression by assisting repressors in gaining access to chromatin. It has been determined that the levels of MTA1 mRNA and the corresponding protein correlate with the metastatic potential of several cancer cells (114, 115, 116) . Although no direct evidence has been reported indicating that higher levels of MTA1 are required for metastasis to occur, HD complexes have been shown to interact with Rb to repress transcription (117, 118, 119) . The mSin3-associated deacetylases have also been demonstrated to be involved in acute promyelocytic leukemia (120, 121, 122) . These results led to the speculation that aberrant regulation of NuRD activity may alter the expression of its target genes, leading to metastatic growth potential.
It has been reported recently that HDAC1 and HDAC2 and the histone-binding proteins RbAp46 and RbAp48 form a core complex shared between NuRD and Sin-HD complexes (123) . A novel polypeptide highly related to MTA1, MTA2, and the methyl-CpG binding-domain-containing protein MBD3 were found to be subunits of the NuRD complex. MTA2 modulates the enzymatic activity of the HD core complex. MBD3 mediates the association of MTA2 with the core HD complex. MBD3 does not directly bind methylated DNA, but it is closely related to MBD2, a protein that binds to methylated DNA and has been reported to contain demethylase activity. MBD2 interacts with the NuRD complex and directs the complex to methylated DNA. These results indicate that NuRD may silence genes by DNA methylation (123) . Interestingly, MBD2 has also been identified as a colon cancer antigen (124) , suggesting a potential involvement of NuRD-like complexes in colon cell transformation.
Taken together, these findings indicate that in general, histone acetylation and deacetylation correlate with activation and suppression of gene expression, confirming that remodeling of chromatin structure and nucleosome organization is obligatory for physiological control of transcription. Alterations in mechanisms that mediate the enzymology of histone modifications can support changes in regulatory events that are involved in the onset and/or progression of tumorigenesis.
Nuclear Matrix-associated Transcriptional Domains.
The subnuclear distribution of transcription factors appears to be
important for the fidelity of transcriptional control. As mechanisms
that mediate the various components of transcription factor trafficking
are pursued, additional regulatory parameters of gene expression are
being defined. Nuclear import was the first aspect of transcription
factor trafficking to be addressed. Consequently, the biochemistry of
transcription factor entry to the nucleus is now understood in relation
to the structural and functional properties of the nuclear pore within
the context of linkages between morphology, biochemistry, and
regulatory activities. More recently, attention has turned to nuclear
retention and export of regulatory factors from the standpoints of
contributions to factor activity as well as the regulated and
regulatory aspects of subcellular factor distribution.
The AML transcription factors (also referred to as CBF
,
PEBP2
, and RUNX proteins) that support hematopoietic
(125, 126, 127, 128, 129, 130, 131, 132, 133, 134)
and bone tissue-specific (16
, 18)
gene expression have provided a paradigm for directly examining
mechanisms that target regulatory factors to subnuclear sites that
support transcription. Functional biochemical and in situ
immunofluorescence analyses of AML deletion and point mutations have
provided an indication of how these transcription factors are directed
to nuclear matrix-associated intranuclear domains. It has been shown
that: (a) sequences required for targeting AML factors to
the nuclear matrix reside in a 31-amino acid segment within the COOH
terminus that is physically distinct from the nuclear localization
signal; (b) nuclear matrix association of AML factors is
independent of DNA binding activity; (c) the principal
active and inactive splice variants of the AML transcription factors
are differentially localized within the nucleus; and (d) the
nuclear matrix targeting signal of AML factors functions autonomously.
These findings demonstrate that at least two trafficking signals are
required for subnuclear targeting of the AML transcription factors; the
first supports nuclear import, and the second mediates association with
the nuclear matrix. Recent results provide insight into the functional
consequences of directing transcripton factors to the nuclear matrix.
Invoking the rationale that guilt by association is biologically
relevant, it has been shown that 31-amino acid nuclear matrix targeting
sequence of the AML transcription factor targets the regulatory protein
to a subnuclear domain that supports transcription. Colocalization of
AML with transcriptionally active RNA polymerase II has been
demonstrated, as have requirements for a functional DNA binding domain
and ongoing transcription (135)
. Functional implications
of subnuclear localization of AML transcription factors are more
directly provided by studies that establish that targeting to the
nuclear matrix-associated sites is obligatory for maximal
transactivation activity (35)
.
From a general biological perspective, there is growing appreciation of sequence requirements for intranuclear targeting of steroid hormone receptors [estrogen receptor (136) and glucocorticoid receptor (137, 138, 139) ] as well as ubiquitous [YY1 (140) ] and selectively utilized [PIT1 (136) and PML (141) ] regulatory proteins. Evidence for a nucleolar targeting domain in parathyroid hormone-related protein (142) and YY1 (140) has been reported. Taken together, we are increasing our understanding of mechanisms that mediate the assembly of regulatory components to initiate and sustain transcription within the context of nuclear architecture.
Integration of Regulatory Cues.
We are gaining insight into the integration of regulatory signals that
control expression by mediating cross-talk between components of
signaling pathways that are operative within the three-dimensional
context of nuclear architecture. There is growing appreciation that
mechanisms modulating chromatin remodeling require the involvement of
higher-order nuclear structure (143)
. The human SWI/SNF
and mouse BAF complexes have been shown to be associated with the
nuclear matrix (144
, 145) . Functional implications are
provided by the observation that the BAF complex is only associated
with the nuclear matrix after mitogenic stimulation of T lymphocytes
when genes controlling competency for proliferation and cell cycle
progression are activated (145)
. In resting cells, the BAF
complex is primarily present in the soluble nuclear fraction. However,
immediately after the induction of proliferation (10 min, 90% of the
BAF complex is found tightly associated with the nuclear matrix
fraction (145)
. The specific parameters of chromatin
remodeling that are linked to nuclear matrix binding of BAF as well as
the cause and/or effect relationships between BAF activity and
parameters of nuclear organization will unquestionably be informative.
Further insight into linkages between nuclear architecture, cytoarchitecture, and the regulation of chromatin structure is provided by recent reports that actin-related proteins are components of chromatin remodeling complexes (62 , 145, 146, 147) . It has been suggested that these actin-related and actin-binding proteins may provide a basis for interactions between chromatin remodeling complexes and cytoskeletal structures involving actin. This suggestion is further supported by observations that both human SWI/SNF complexes and the Drosophila BRM complexes not only contain an actin-related protein (BAF 53 in human cells and BAP 55 in Drosophila) but also contain actin (145 , 147) . Furthermore, regions of BAF that contact myosin, proliferin, and other actin-binding proteins are similar to actin (145) . The possibility can therefore be considered that such interactions are important for SWI/SNF function.
Linkages of Aberrant Nuclear Organization with Modified Gene
Expression in Cancer.
Interrelationships of nuclear structure with gene expression are
illustrated by the modified subnuclear organization of genes and
regulatory factors in cancer (Fig. 1)
. Transformed and tumor cells
exhibit striking alterations in nuclear morphology as well as in the
representation and intranuclear distribution of nucleic acids and
regulatory factors. In both leukemias and solid tumors, there are
modifications in components of nuclear architecture that are involved
in control of gene expression. Examples include mutations of the AML,
ALL, and PML loci in leukemias that accompany changes in gene
expression and the subnuclear organization of encoded transcription
factors. In colon tumor cells, modifications in the subnuclear
distribution of the APC factor are observed (148)
. These
factors are associated with nuclear architecture, and the alterations
in relationships with nuclear architecture appear to be related to
changes in gene control. Identification of nuclear import signals in
transcription factors and the recent characterization of intranuclear
targeting signals that direct regulatory proteins to subnuclear domains
that support transcription reinforce linkages between nuclear structure
and aberrant transcriptional control. These observations provide an
opportunity to develop high-resolution in situ
immunofluorescence analysis to diagnose and stage tumors and to monitor
remission, relapse, and effectiveness of treatment. There is a
potential for developing therapeutics that are directed to subnuclear
sites that support specific components of gene expression.
Alterations in nuclear organization are the hallmarks of cancer cells. The gene locus encoding the AML transcription factor is frequently the target of chromosomal translocations in human leukemia. Replacement of the chromosome 21-encoded intranuclear trafficking signal by a targeting signal from chromosome 8 redirects the t(8;21) translocation-fusion protein to unique subnuclear sites. Thus, intranuclear targeting of the AML transcription factor may be abrogated because of gene rearrangements in leukemic cells. Fidelity of transcriptional control may involve the localization of gene-regulatory proteins to the correct subnuclear region (149) .
PML bodies are another example of nuclear structures that are associated with the nuclear matrix and modified in leukemia cells (52) . In normal cells, the PML protein resides in discrete PML bodies. However, in promyelocytic leukemic cells, the PML protein is genetically rearranged and dispersed throughout the nucleus (52 , 150) . A further example of chromosomal translocations involving a locus encoding a nuclear matrix-associated transcription factor occurs in acute lymphocytic leukemia (ALL/MLL). Recently, a translocation has been described in which the ALL/MLL protein is fused with a HAT. This chimeric protein may promote leukemia by modifying histone acetylation of specific genomic regions. Consequential modifications in the intranuclear distribution of factors encoded by the rearranged ALL locus occur (151, 152, 153) , although the chimeric transcription factors remain nuclear matrix associated (154) . Hence, these results suggest that perturbations in subnuclear location of regulatory proteins may be related to modifications in gene expression that are linked to leukemias. Additionally, tumor-associated modifications have been observed in nuclear domains that support the processing of transcripts, the intranuclear organization of Rb and DNA replication foci, and the nucleocytoplasmic shuttling of p53 that has been functionally linked to association of Mdm2 with the nucleolus (155, 156, 157) .
| Perspectives and Future Directions |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
1 Components of the work reported in this review
were supported by NIH Grants AR45688, AR39588, and AR45689. ![]()
2 To whom requests for reprints should be
addressed, at Department of Cell Biology and Cancer Center, University
of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA
01655. Phone: (508) 856-5625; Fax: (508) 856-6800; E-mail: gary.stein{at}umassmed.edu ![]()
3 The abbreviations used are: MMTV, mouse mammary
tumor virus; NF, nuclear factor; HAT, histone acetyltransferase; HD,
histone deacetylase; SAGA, Spt-Ada-Gcn5-acetyl transferase; NuRD,
nucleosome remodeling HD; CBP, CREB-binding protein; CREB,
cAMP-responsive element-binding protein; AML, acute myelogenous
leukemia; ALL, acute lymphocytic leukemia; PML, promyelocytic leukemia;
MLL, mixed lineage leukemia; ACTR, activator of retinoic acid receptor;
TRAM, thyroid receptor activator molecule; SRC, steroid receptor
coactivator; RAC, receptor-associated coactivator; P/CAF,
p300/CBP-associated factor; SMRT, silencing mediator of retinoic acid
and thyroid hormone receptor; CBF
, core binding factor
; PEBP
,
polyoma enhancer binding protein
; RUNX, runt-related. ![]()
Received 12/ 1/99. Accepted 3/ 2/00.
| REFERENCES |
|---|
|
|
|---|
transcription factors. Proc. Natl. Acad. Sci. USA, 94: 6746-6751, 1997.